Microbial Engineering for Aldehyde Synthesis
by
by
ARMt'ES
MASSACHUSETT INSTIrJTE
OF rECH.ULOLGY
Aditya Mohan Kunjapur
JU
JUN 2 2 2015
Bachelor of Science in Chemical Engineering
University of Texas at Austin, Austin, TX, USA, May 2010
LIBRARIES
Submitted to the Department of Chemical Engineering
In Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Chemical Engineering
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2015
2015 Massachusetts Institute of Technology. All rights reserved.
Signature redacted
Signature of Author..................
.........
.................
.
........
Aditya M. Kunjapur
Department of Chemical Engineering
Mav 19. 2015
Signature redacted
C e rtified by ..........................................................................
..
...........
.................................................
Kristala L. Jones Prather
Associate Professor of Chemical Engineering
Thesis Supervisor
.Signature redacted
A cce pted by............................................................................
.. ....................
.............................
........
Richard D. Braatz
Professor of Chemical Engineering
Chairman, Committee for Graduate Students
Microbial Engineering for Aldehyde Synthesis
by
Aditya Mohan Kunjapur
Submitted to the Department of Chemical Engineering on May 19, 2015
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy in Chemical Engineering
Abstract
Microbes have been engineered to produce many useful classes of chemicals from
renewable carbon sources instead of from finite petroleum reserves. Aldehydes represent a class
of chemicals that has been challenging to obtain using microbes given the rapid conversion of
aldehydes into their corresponding alcohols that occurs naturally. Microbes are thought to have
evolved numerous endogenous enzymes responsible for catalyzing these conversions in order to
alleviate the negative effect of many aldehydes on cellular processes.
In this thesis, we investigate several aspects of microbial aldehyde synthesis. Driven first
by the hypothesis that targeted gene deletions could decrease endogenous aldehyde reduction
in a model E. coli host strain, we demonstrate that benzaldehyde accumulation occurs upon
deletion of a combination of genes encoding enzymes known to have benzaldehyde reductase
activity in vitro. Using deletion subset studies and quantitative real-time PCR, we discover that
deletion of many, but not all, of these genes is required to curtail endogenous reduction. We also
show that the same engineered strain has a significantly decreased rate of reduction of other
aromatic aldehydes. As an added benefit, cell growth rate is unaffected by these deletions. We
demonstrate the utility of this strain for two applications: (i) conversion of glucose into vanillin,
which is the most widely used flavoring additive; and, (ii) conversion of benzaldehyde and glucose
into L-phenylacetylcarbinol, which is a chiral pharmaceutical intermediate.
We next explore the ability to produce and retain non-aromatic aldehydes with the
specific objective of studying the conversion of fatty aldehydes into gasoline-range alkanes. We
find that a carboxylic acid reductase (Car) from Nocardia iowensis achieves biosynthesis of
aldehydes from free fatty acid substrates ranging in carbon chain length from C4 to C10. The use
of Car, the engineered host strain, and previously elucidated pathways to free fatty acids enables
production of alkanes ranging from C3 to C9. Although alcohol byproduct formation significantly
decreases, it does not significantly increase alkane titer because of poor aldehyde decarbonylase
kinetics.
Additional work presented in this thesis seeks to identify and surmount limitations in
aldehyde biosynthesis in vitro and in E. coli de novo vanillin biosynthesis.
Thesis Supervisor: Kristala L. Jones Prather
Title: Associate Professor of Chemical Engineering
3
Dedication
My mother, Shaku Kunjapur, once ran a pediatric clinic in Hyderabad, India, after
completing her medical studies there. Soon after marrying my father, she closed down her
practice and followed him to America, where he had begun pursuing a Master's Degree in
Engineering. Although my father's interest and perseverance in earning an American education
were vital to later providing me with opportunities that few of my kin in India could dream of,
this short story remains about my mother's sacrifice. Once she passed the medical residency
exams required of foreign educated doctors in order to ultimately practice in the U.S., she had a
newfound constraint in the form of me. This became especially challenging when my father's
engineering career took us to Houston, Texas, and my mother reluctantly began her residency
four hours away in the city of Corpus Christi. I was a toddler at the time and one of the earliest
memories that I still retain was a weekend trip to Corpus Christi to visit her. In a few months, my
mother quit her residency and became a stay-at-home mom. Even as I grew older and remained
an only child, she maintained an immensely loving and nurturing environment at home. The
privilege of receiving so much attention from her imparted a sense of responsibility in me to
make a difference in society on behalf of her and me. These were the foundations that fostered
the lofty aspirations that drive me today. And that is why this doctoral thesis is dedicated to my
mom, the real Dr. Kunjapur.
4
Acknowledgements
Countless individuals have supported me throughout my journey at MIT. Foremost on this
list is my advisor, Professor Kris Prather, who took me into her lab and patiently guided me along
as I began with minimal working knowledge of biology. She displayed tremendous understanding
over the years as my growing appreciation for experimentation strayed beyond my initial
objectives and even occasionally beyond the laboratory bench. I also benefited immensely from
her participation with other members of the synthetic biology community and from her
recruitment of many talented and kindhearted individuals who I can call labmates.
I had the opportunity to engage with other world-class faculty members during my time
at MIT and am especially grateful for the perspectives shared by Professors Charlie Cooney, Amy
Keating, and John Dueber. Besides serving on my thesis committee, these professors cared to
meet with me individually and to provide valuable constructive criticism. I am a better scientist
because of their interest in my project and in my development.
I mentioned labmates earlier, but several individuals from the Prather Lab deserve
personal recognition. Dr. Micah Sheppard taught me most of what I know about so many aspects
of research, including metabolic pathway design, controlled experimentation, and even effective
scientific illustration. In addition to nurturing my intellectual growth, Micah's selfless interest in
regularly discussing diverse aspects and implications of my project ultimately led to exciting
collaborations on biofuel production. With regard to collaborations, I am also grateful to Kat
Tarasova for her interest in the aromatic aldehyde work and for developing a method for their
detection, replicating my initial encouraging results, and helping develop and edit the potential
story as it unfolded. I also had many fruitful scientific discussions with Dr. Kevin Solomon and Dr.
Himanshu Dhamankar that increased my understanding of the field, and I acquired important lab
skills with the help of Dr. Matt McMahon, Dr. Chris Reisch, Dr. Eric Shiue, Dr. Hsien-Chung Tseng,
and Dr. Diana Ritz. I was privileged to experience my MIT journey alongside my classmate and
labmate Irene Brockman and especially enjoyed the summers that we synthesized and shared
our budding knowledge of metabolic engineering principles to high school students together.
Finally, I would like to thank the undergraduate research assistants who I worked with for their
help in achieving diverse research objectives and for the many things I learned from them:
Bernardo Cervantes, Spencer Wenck, Nikita Khlystov, Asmamaw Wassie, and Jason Hyun.
Several organizations made invaluable contributions to my professional development at
MIT. I extend my deepest gratitude to everyone I worked with in the MIT Energy Club and in the
Synthetic Biology Engineering Research Center (SynBERC) for providing me with great
opportunities, for inspiring me, and for trusting me to help manage initiatives that hopefully
aided the professional development of other students. I would also like to thank my SynBERC
Industrial Mentor, Todd Peterson, for volunteering his time towards being a terrific mentor and
providing perspectives on research that are rarely found in academic contexts.
This research would not have been possible without significant financial support, and for
that I would like to thank the National Science Foundation, SynBERC, and Dr. Pete Bixler. I also
thank the following agencies for providing more focused funding: MITei, Dow, and the GSC.
There are so many mentors and friends left to thank, but so little space. However, I would
be remiss without expressing my appreciation for Dr. Bruce Eldridge and for Nancy McBride, who
have each provided exceptional support for my academic pursuits since my days in Texas.
5
Table of Contents
Abstract...........................................................................................................................................
3
Dedication .......................................................................................................................................
4
Acknow ledgem ents.........................................................................................................................5
Table of Contents............................................................................................................................
6
List of Figures..................................................................................................................................
9
List of Tables .................................................................................................................................
13
Chapter 1: Introduction ................................................................................................................
14
1.1. Biological aldehydes........................................................................................................
14
1.2. Engineering aldehyde biosynthetic reactions and pathways ........................................
15
1.3. Minimizing endogenous conversion of aldehydes to alcohols ......................................
25
1.4. Enhancing bioconversion of aldehydes to other chemical classes ................................
26
1.5. Addressing aldehyde toxicity .........................................................................................
27
1.6. Thesis organization.........................................................................................................
30
Chapter 2: Engineering synthesis and accumulation of aromatic aldehydes in E. coli............. 32
2.1. Introduction........................................................................................................................
32
2.2. M aterials and M ethods.................................................................................................
34
2.2.1. Strains and plasm ids.................................................................................................
34
2.2.2. Chem icals.....................................................................................................................
39
2.2.3. Culture conditions ...................................................................................................
39
2.2.4. M etabolite analysis..................................................................................................
41
2.2.5. Quantitative Reverse Transcription PCR (qRT-PCR) ................................................
42
2.3. Results................................................................................................................................43
2.3.1. A combination of rationally targeted gene deletions enables benzaldehyde
accum ulation in E. coli.................................................................................................
.. 43
2.3.2. In vanillate-feeding experiments, the RARE strain eliminates conversion of vanillin into
vanillyl alcohol.......................................................................................................................
49
2.3.3. The RARE strain enables production of vanillin from glucose in E. coli.................. 51
6
2.3.4. The RARE strain enables the synthesis of L-PAC in E. coi....................................... 54
2.3.5. The deletion of all targeted genes to form the RARE strain has no effect on growth 57
2.4. Discussion ...........................................................................................................................
59
Chapter 3: Decreasing endogenous reduction of aliphatic aldehydes and the effect on
biosynthesis of gasoline-range n-alkanes ................................................................................
64
3.1. Introduction........................................................................................................................
64
3.2. M aterials and M ethods.................................................................................................
69
3.2.1. Strains and M odules .................................................................................................
69
3.2.2. Chem icals.....................................................................................................................
72
3.2.3. Culture Conditions....................................................................................................
73
3.2.4. M etabolite Analysis .................................................................................................
73
3.3. Results ................................................................................................................................
74
3.3.1. Modules 1-Ma, 3-Oc, and 4-LA Result in Heptane and Nonane Biosynthesis ........ 74
3.3.2. Substitution of an RBO Module (Module 2-MCC) Results in Biosynthesis of Pentane as
the Sole Alkane Product ...................................................................................................
78
3.3.3. M odules 1-Pr, 2-M CC, and 4-SA Enable Butane Biosynthesis................................. 79
3.3.4. M odules 2-BC and 4-SA Enable Propane Biosynthesis............................................ 80
3.4. Discussion ...........................................................................................................................
81
Chapter 4: Enhancing in vitro aldehyde biosynthesis by pairing carboxylic acid reductase with
inorganic pyrophosphatase ......................................................................................................
84
4.1. Introduction........................................................................................................................
84
4.2. M aterials and M ethods .................................................................................................
86
4.2.1. Plasm id construction ..............................................................................................
86
4.2.2. Chem icals.....................................................................................................................
88
4.2.3. Enzym e purification .................................................................................................
88
4.2.4. Kinetic studies.........................................................................................................
90
4.3. Results ................................................................................................................................
92
4.4. Discussion ...........................................................................................................................
98
Chapter 5: Towards improving de novo vanillin biosynthesis in E. coli by deregulating Sadenosylmethionine biosynthesis ..............................................................................................
101
7
5.1. Introduction......................................................................................................................
101
5.2. M aterials and M ethods ....................................................................................................
105
5.2.1. Strains and plasm ids..................................................................................................
105
5.2.2. Chem icals...................................................................................................................
110
5.2.3. Culture conditions .....................................................................................................
110
5.2.4. M etabolite analysis....................................................................................................112
5.2.5. SDS-PAGE analysis .........................................................................................................
113
5.3. Results ..............................................................................................................................
114
5.3.1. Focusing on central carbon m etabolism first ............................................................
114
5.3.2. Understanding why conversion of protocatechuate to vanillate was limiting: SAM 118
5.3.3. Investigating potential bottlenecks in SAM biosynthesis.......................................... 120
5.3.4. Improving vanillate production by deregulating SAM biosynthesis ......................... 123
5.4. Discussion .........................................................................................................................
Chapter 6: Lessons Learned and Future Directions....................................................................
129
132
6.1. Sum m ary ..........................................................................................................................
132
6.2. Future Directions..............................................................................................................
135
6.2.1. M etabolite sensors for the vanillin pathw ay.............................................................
136
6.2.2. Use of CoA-dependent pathways to generate novel aliphatic aldehydes................137
6.2.3. M icrobial aldehyde toxicity .......................................................................................
References ..................................................................................................................................
138
140
8
List of Figures
Figure 1 - 1. Overview of natural metabolic pathways that can be harnessed for the conversion
of glucose to valuable aromatic and aliphatic aldehydes through carboxylic acid intermediates
based on E. coli metabolism. Aldehydes can also be obtained from the 2-keto acid pathway,
terpenoid pathways, and other pathways................................................................................
19
Figure 1- 2. Potential biocatalytic and metabolic engineering opportunities that could be enabled
by, or enhanced by, microbial aldehyde accumulation.......................................................... 26
Figure 2 - 1. A combination of rational gene deletions enables benzaldehyde accumulation in E.
coli. (A) Scheme depicting intracellular formation of benzaldehyde from benzoate and
endogenous conversion to the byproduct benzyl alcohol. (B) Conversion of 5 mM benzoate after
24 hours in strains transformed with pETDuet-1 and pACYC-car-sfp. (C) Conversion of 5 mM
benzoate after 24 hours in RARE strains transformed with pACYC-car-sfp and a pET plasmid
harboring the gene indicated below the x axis.........................................................................
47
Figure 2 - 2. Estimated copies of reverse transcribed mRNA per nanogram of total RNA based on
quantitative reverse transcription PCR (qRT-PCR). ..................................................................
48
Figure 2 - 3. In vanillate-feeding experiments, the RARE strain eliminates conversion of vanillin to
vanillyl alcohol. (A) Scheme depicting intracellular formation of vanillin from vanillate and
endogenous conversion to the byproduct vanillyl alcohol. (B) Conversion of 5 mM vanillate after
48 hours in strains transformed with pETDuet-1 and pACYC-car-sfp. (C) Conversion of 5 mM
vanillate after 24 hours in RARE strains transformed with pACYC-car-sfp and a pET plasmid
harboring the gene indicated below the x axis......................................................................... 50
+
Figure 2 - 4. The RARE strain enables production of vanillin from glucose in E. coli. (A) Scheme
depicting complete pathway from glucose to vanillin with overexpressed E. coil aroG* indicated
in bold typeface. [PYR = pyruvate, F6P = fructose 6-phosphate, G3P = glyceraldehyde 3phosphate, X5P = xylulose 5-phosphate, PEP = phosphoenolpyruvate, E4P = erythrose 4phosphate, DAHP = 3-deoxy-D-arabinoheptulosonate 7-phosphate, DHQ = 3-dehydroquinate,
DHS = 3-dehydroshikimate] (B) Scheme illustrating heterologous portion of pathway with two
possible undesired alcohol byproducts. (C) Concentration profiles of the six heterologous
metabolites of interest (including vanillin) when wild-type and RARE hosts are transformed with
pET-OMT-asbF and pACYC-car-sfp-aroG* plasmids and grown in LB + 1.2% glucose for 48 hours.
(D) Concentration profiles of the six heterologous metabolites of interest (including vanillin)
produced from glucose as a sole carbon source after 60 hours in the same strains grown in M9
1.2% gluco se . ................................................................................................................................
52
Figure 2 - 5. Concentration of isovanillin at final time points from experiments (either 48 hours
for LB or 60 hours for M9) that examined the production of vanillin from glucose............... 54
Figure 2 - 6. The RARE strain enables the synthesis of L-phenylacetylcarbinol (L-PAC). (A) Scheme
depicting the synthesis of L-PAC from the condensation of exogenously supplied benzaldehyde
and metabolized pyruvate, catalyzed by either PDC or PDC_E473Q. (B) Concentration profiles of
9
PAC, benzaldehyde, and benzaldehyde oxidation/reduction products 24 hours after addition of 5
mM benzaldehyde to RARE and wild-type host strains. (C) Time course of benzaldehyde reduction
using the wild-type host transformed with pRSF/PDCE473Q and pACYC/Car/Sfp................ 56
+
Figure 2 - 7. OD600 measurements of MG1655 and RARE strains when grown in (A) LB medium
1.2% glucose or (B) M9 minimal media + 1.2% glucose. (C) Specific growth rates of each strain
calculated from experiments in the different media listed above........................................... 58
Figure 2 - 8. OD 6oo measurements at final time points from reported experiments: (A)
benzaldehyde from benzoate; (B) vanillin from vanillate; (C) vanillin from LB + 1.2% glucose; (D)
vanillin from M 9 + 1.2% glucose.............................................................................................. 59
Figure 3 - 1. Relative activity of the carboxylic acid reductase from Nocardia iowensis (CarNi) on
straight and branched aliphatic acids ranging in carbon chain length from C2 to C8. ............ 65
Figure 3 - 2. Biochemical pathway illustration depicting fatty aldehydes as a precursor to either
an alkane or an alcohol.............................................................................................................
66
Figure 3 - 3. (A) Composition of typical regular unleaded gasoline displayed in weight percent (wt.
%) based on the average of Refs. 146 and 147. Single asterisk indicates that compounds below
0.5 wt. %are not reported in Ref. 147. Double asterisks indicate that wt. % includes contribution
from trace compounds in Ref. 146. (B) Modular pathway design used for selective synthesis of
key gasoline-range alkanes in engineered E. coli. Genes in gray within Modules 1-Pr and 1-Ma are
native and were not overexpressed, whereas genes in black were overexpressed. Module names
are abbreviations for the following: "Pr" = Propionate; "Ma" = Malonyl-ACP; "BC" = Butyrl-CoA;
"MCC" = Medium-Chain-CoA; "Oc" = Octanoate; "SA" = Short Alkanes; "LA" = Long Alkanes. .. 68
Figure 3 - 4. Selective production of heptane and nonane using FAS for carbon chain extension.
(A) C8 and C10 FFA titers resulting from Module 3-Oc or Modules 1-Ma and 3-Oc in WTAfadD. (B)
Gas-phase titers of octanal observed 24 hours after supplying octanoate to WT and RARE
expressing CarNi. (C) Illustration of octanal as a branch-point metabolite to heptane or octanol.
(D) Alkane titers resulting from Modules 1-Ma, 3-Oc, and 4-LA in WTsAfadD and RAREnfadD.
Experiments performed in triplicate with averages as reported values and standard deviation as
error bars. All alkane titers are gas-phase................................................................................
76
Figure 3 - 5. Selective production of pentane using RBO for carbon chain extension. (A) Liquidphase titers of hexanoate and downstream metabolites observed 24 hours after supplying
hexanoate to WT and RARE expressing CarNm. (B) Alkane titers resulting from Modules 2-MCC and
either 4-LA or 4-SA in WT and RARE. (C) Liquid-phase titers of butanol and hexanol in WT and
RARE containing Modules 2-MCC and 4-SA. Experiments performed in triplicate with averages as
reported values and standard deviation as error bars. All alkane titers are gas-phase. ......... 77
Figure 3 - 6. Alternative modules enable synthesis of butane and propane. ..........................
80
Figure 3 - 7. Intermediate and byproduct profiles associated with propane synthesis. (A) Relative
butyraldehyde concentrations in the headspace of cultures containing Modules 2-BC and 4-SA.
An increased concentration of butyraldehyde was observed in the gas phase using RARE. (B)
Liquid-phase concentrations of butyraldehyde and butanol in cultures containing Modules 2-BC
10
and 4-SA. Increased levels of butyraldehyde and decreased levels of butanol were observed in
the liquid phase using RARE ......................................................................................................
81
Figure 4 - 1. Effect of varying MgC 2 concentration or adding commercial inorganic
pyrophosphatase (Ppa) from New England Biolabs on Car-catalyzed conversion of the substrate
benzoate. The concentration of Car used was 224 nM. The units of Ppa added was 0.1, where
one unit is as defined by NEB (The amount of enzyme that will generate 1 pmol of phosphate per
minute from inorganic pyrophosphate under standard reaction conditions [a 10 minute reaction
at 250 C in 20 mM Tris-HCI, pH 8.0, 2 mM MgC1 2 and 2 mM PPi]). Experiment performed in
duplicate. Data points shown are averages with error bars representing standard deviations.. 94
Figure 4 - 2. Effect of MgC 2 concentration and addition of an "in-house" Ppa (896 nM) on an in
vitro reaction pathway involving Car (224 nM) and a heterologous aldo-keto reductase, YtbE
(1422 nM). To ensure that no other component of the commercial Ppa mixture was responsible
for the reaction enhancement, we expressed and purified the E. coli ppa gene product. We
included an aldo-keto reductase that catalyzes the conversion of benzaldehyde into benzyl
alcohol to investigate whether the reaction catalyzed by Car would be enhanced simply by
creating a sink for the product. Subsequent experiments showed that the higher concentration
of MgC2 slightly reduced the activity of the second enzyme. Experiment performed in duplicate.
.......................................................................................................................................................
95
Figure 4 - 3. Addition of Ppa enables an in vitro pathway involving Car and an aldo-keto reductase
to be modeled with far greater accuracy using Michaelis-Menten kinetics and parameters
obtained from initial rate measurements. Model parameters: KM,Car-GBD = 0.35 mM; KM,YtbE-SH3 = 2
mM; kcat, Car-GBD 216 min-; kcat,YtbE-SH3 = 96 min 1 . Data points represent the average of duplicate
experim ental values......................................................................................................................
96
Figure 4 - 4. Effect of Ppa addition on the Car-catalyzed conversions of two substrates that result
in aldehydes valuable as flavors. X represents the conversion of substrate C (X = C,/Cio).
Experim ent perform ed in duplicate. ........................................................................................
97
Figure 4 - 5. Effect of the molar ratio of Ppa to Car on conversion of benzoate. The concentration
of Car was fixed at 224 nM. The purpose of this experiment was to help determine the minimum
amount of Ppa required to add relative to Car in order to achieve saturating levels of
enhancement. All ratios tested achieved saturating enhancement. Experiment performed in
d u p licate. ......................................................................................................................................
98
Figure 5 - 1. The engineered vanillin pathway in E. coli. (A) Endogenous portion of the vanillin
pathway. (B) Heterologous portion of the vanillin pathway, with reactions catalyzed by CarNi
shaded in gray. Genes corresponding to enzymes labeled in red are overexpressed in experiments
investigating improvement of vanillate production. Enzymes written without subscripts are
native to E. coli. The heterologous pathway portion in the engineered yeast vanillin pathway
contains identical metabolites and enzymes with the exception of AsbFBt............................... 103
11
Figure 5 - 2. Effect of perturbations in central metabolism intended to increase PEP and E4P
availability on heterologous metabolite titers and specific yields. (A) Deletion of PTS- glu' did not
improve titers of either protocatechuate or vanillate. (B) Overexpression of ppsA and tktA in the
PTS~ glut RARE' host resulted in an increase in protocatechuate titer and specific yield compared
to expression of the pathway without ppsA and tktA. (C) Bioreactor culture of PTS- glu* RARE'
host expressing the pathway (without ppsA and tktA overexpression) leads to increased
protocatechuate titers without a concomitant increase in vanillate titers, indicative of room for
improvement in the conversion of protocatechuate to vanillate. Host in blue text, overexpressed
genes in red text. ........................................................................................................................
118
Figure 5 - 3. Identification of the bottleneck in vanillate production. (A) SDS-PAGE result showing
robust expression of OMTHs. (B) Effect of 10 mM L-methionine supplementation at peak
productivity (24 h) on vanillate titers, with and without overexpression of metK. (C) Pathway
illustrating the reaction catalyzing conversion of protocatechuate into vanillate in the context of
SAM biosynthesis and recycling. (D) Effect of 2.5 mM L-homocysteine supplementation at peak
productivity (24 h) on vanillate titers and specific yield. In both pathway experiments shown here
(B and D), the PTS~ glut RARE' host overexpressing aroG*, ppsA, tktA, asbF, and OMTwas tested.
.....................................................................................................................................................
1 19
Figure 5 - 4. Effect of meti deletion (A) in different host strains and (B) in the presence of amino
acid supplementation. For these experiments, the following genes were overexpressed: aroG*,
ppsA, tktA, asbF, and OMT. For the amino acid supplementation experiment (B), 10 mM of amino
acid w as added at induction.......................................................................................................
124
Figure 5 - 5. Effect of metA* and cysE* overexpression. (A) Effect of overexpressing feedbackdesensitized variants of metA and/or cysE along with usual pathway constructs in the RARE Ametj
host. The control represents co-transformation with an empty pCOLADuet-1 plasmid. (B) Effect
of methionine supplementation level and timing on vanillate titers in metA*-cysE* cultures. (C)
Kinetics of vanillin production without overexpression of metA*-cysE*. (D) Kinetics of vanillin
production with overexpression of metA*-cysE*. Although final titers achieved in (C) and (D) are
similar, the metA*-cysE* cultures grow more slowly, produce vanillin more slowly, but display
greater conversion of protocatechuate to vanillate................................................................... 126
Figure 5 - 6. Images of plates testing for potential loss of ampicillin-resistant plasmid. No
significant plasmid loss was observed for samples taken at 24, 48, and 72 h. .......................... 128
Figure 5 - 7. The activated methyl cycle in E. coli (in black), along with an alternative SAH recycling
route featuring a heterologous SAH hydrolase (sahH, in blue).................................................. 131
12
List of Tables
Table 1 - 1. Relevant published aldehyde biosynthesis patent applications. ..........................
22
Table 2 - 1. Strains and plasmids used in this study.................................................................
35
Table 2 - 2. Oligonucleotides used in this study......................................................................
37
Table 2 - 3. Reported activities of E. co/i gene products on benzaldehyde in vitro. ................ 45
Table 2 - 4. Results of protein BLAST sequence alignments used to organize deletion targets. 45
Table 2 - 5. E. coli strains featuring different combinations of gene deletions ("X" indicates
d e letio n )........................................................................................................................................
46
Table 3 - 1. Performance and separation metrics for select gasoline alternatives and constituents.
.......................................................................................................................................................
67
Table 3 - 2. Strains and modules used in this study. ................................................................
71
Table 3 - 3. Oligonucleotides used in this study......................................................................
72
Table 4 - 1. Oligonucleotides used in this study......................................................................
88
Table 4 - 2. Combinatorial testing of in vitro components for formation of precipitate.......... 93
Table 5 - 1. Strains and plasm ids used in this study ...................................................................
107
Table 5 - 2. Synthesized gene sequences used in this study...................................................... 109
Table 5 - 3. Oligonucleotides used in this study.........................................................................
110
13
Chapter 1: Introduction
Portions of this chapter are adapted from thefollowing manuscript: Kunjapur A.M. and Prather
K.L.J. (2015), AppL. Environ. Microbiol., 81 (6)1892-1901.
1.1. Biological aldehydes
The word "aldehyde" was coined in the early
1 9 th
century by Justin von Liebig, who
formed a contraction using the Latin words "alcohol dehydrogenatus," or "alcohol deprived of
hydrogen" (1). Aldehydes have a variety of industrial uses, but they are perhaps most familiar for
their effects on two of the mammalian senses: olfaction and gustation. Numerous aldehyde
odorants are known to bind to G-protein-coupled receptors, triggering reaction cascades that
ultimately result in mammalian perception (2-5). At dilute concentrations, fatty aldehydes such
as hexanal, octanal, decanal, and dodecanal offer apple, citrus, orange peel, and violet scents,
respectively (6). Aromatic aldehydes, such as benzaldehyde, anisaldehyde, vanillin, and
cinnamaldehyde, are responsible for the natural fragrances of almond, sweet blossom, vanilla,
and cinnamon, respectively (6, 7). Notable terpenoid aldehydes include citral, which provides
lemon scent (6), and safranal, which is one of the primary molecules responsible for saffron
aroma (8). Aldehydes play a role in other animal phyla as well. Certain aldehydes, such as trans2-hexenal, phenylacetaldehyde, and nonanal, evoke responses in insects by serving as
pheromones or attractants (9-11). The high reactivity of the carbonyl group of aldehydes enables
many industrial uses beyond flavors and fragrances, such as precursors to pharmaceuticals (1215). However, the high reactivity of aldehydes also contributes to their increased toxicity in
microorganisms. Given the high-value applications and large markets for several aldehydes,
14
commercial focus on microbial aldehyde synthesis has surged in recent years (16). This chapter
summarizes published efforts towards microbial engineering for aldehyde synthesis, with an
emphasis on de novo aldehyde synthesis, attempts at engineering aldehyde accumulation in E.
coli, and the challenge of aldehyde toxicity.
1.2. Engineering aldehyde biosynthetic reactions and pathways
Because most microbes do not naturally accumulate aldehydes, microbial production of
these molecules from simple carbon sources requires at least two parallel approaches: pathway
construction for product generation and strain engineering for product accumulation. A starting
point for pathway construction is consideration of enzymatic reactions that can produce desired
aldehydes from cellular metabolites. Carboxylic acids are found throughout cellular metabolism
and many can be converted to aldehydes with the aid of a single enzyme. Prior to the detailed
characterization and cloning of enzymes capable of broadly catalyzing aldehyde formation,
various natural organisms ranging from actinomycetes to white rot fungi were tested for innate
ability to convert carboxylic acids into their corresponding aldehydes or alcohols (17-21). A
significant advance occurred roughly one decade ago, when a carboxylic acid reductase (CarNi)
from Nocardia iowensis was cloned into Escherichia coli and shown to be active on several
aromatic carboxylic acids in vitro (22). Later publications from Rosazza and colleagues
demonstrated that CarNi requires one-time activation by a phosphopantetheinyl transferase and
that CarNi has activity in vitro on a broader range of substrates that includes several citric acid
cycle dicarboxylic acids (23, 24). A homolog of CarNi from Mycobacterium marinum was
demonstrated to have activity on straight-chain aliphatic acids ranging from C6-C18 (25). A recent
review describes a larger number of carboxylic acid reductases that could be harnessed for
15
biosynthesis of a variety of aldehydes (26). The general stoichiometry for reactions catalyzed by
carboxylic acid reductases is as follows (where "e" represents a reducing equivalent):
R-COOH + e- + ATP
-
R-CHO + AMP + PPi
Aliphatic aldehydes across a broad range of carbon lengths can also be formed by using
fermentative aldehyde reductases or by using enzymes that act on activated forms of carboxylic
acids (acyl-CoA or acyl-ACP). During anaerobic cultivation of E. coli, conversion of acetyl-CoA to
acetaldehyde is catalyzed by a CoA-dependent acetaldehyde dehydrogenase (also known as
acetaldehyde CoA dehydrogenase) (27). However, the same protein, encoded by adhE, has a
second catalytic site that converts acetaldehyde into ethanol (28). In solvent-producing clostridial
strains, acetaldehyde and butyraldehyde can be produced by CoA-acylating aldehyde
dehydrogenases that are found as individual enzymes or as bifunctional enzymes (29-32). The
conversion of acyl-CoA to aldehyde is as follows (for acyl-ACP substrates instead of acyl-CoA
substrates, replace "S-CoA" and "CoASH" with "ACP"):
R-CO-S-CoA + e- 4 R-CHO + CoASH
Synthesis of longer carbon-chain aliphatic aldehydes from acyl-ACP precursors can occur
using enzymes from luminescent bacteria. In these bacteria, the multienzyme fatty acid
reductase complex consisting of luxCDE is used to produce aldehydes that are immediate
substrates for the light emission reaction (33). Note that the aldehyde biosynthetic reactions
discussed so far use similar chemistries that primarily differ in the source of reducing equivalents
and whether the carboxylic acid molecule or the reductase enzyme is activated first. In either
16
case, activation requires the conversion of ATP to AMP and pyrophosphate and occurs because
the energetics of converting a carboxylic acid to an aldehyde are ordinarily unfavorable.
Another set of non-oxidative aldehyde biosynthetic routes utilizes decarboxylation of 2keto acid substrates. In these cases, no ATP is required because the irreversibility of CO 2
formation provides the driving force for aldehyde formation. However, one carbon atom is lost
per molecule of 2-keto acid substrate, which reduces the theoretical maximum yield. Two wellknown enzymes in this category are pyruvate decarboxylase (PDC) and 2-ketoisovalerate
decarboxylase (KivD). The native role of PDCs are to convert pyruvate to acetaldehyde, but their
promiscuity and capability to catalyze carboligation side reactions has led to their use in synthesis
of chiral carboligation products (12). KivD is also promiscuous and has been utilized for synthesis
of numerous non-natural alcohols derived from amino acid intermediates (34). The 2-keto acid
decarboxylation reaction is as follows:
R-CO-COOH 4 R-CHO + CO 2
Oxidative reactions can also be used for aldehyde synthesis, starting from either
carboxylic acid substrates or primary alcohol substrates. C,, fatty acids can be converted to C,fatty aldehydes, as was shown using E. coli resting cells that expressed an a-dioxygenase from
Oryza sativa (rice) (35). In this case, spontaneous decarboxylation of a C, hydroperoxy fatty acid
intermediate provides a driving force for aldehyde generation. The dioxygenase-catalyzed
reaction is as follows:
R-CH 2-COOH
+ 02
4 R-CHO + CO 2 + H 2 0
17
In addition, aldehydes can be obtained by enzymatic oxidation of primary alcohols (3639). From a de novo aldehyde synthesis perspective, these reactions are less relevant given that
alcohols are typically produced via aldehyde intermediates. However, biocatalytic conversion of
primary alcohols to aldehydes may provide an array of new opportunities for alcohols as starting
materials and will be revisited later in this chapter. Oxidation of alcohols to aldehydes generates
a reducing equivalent as follows:
R-CH 2-OH -* R-CHO + e-
Natural and engineered pathways could be used to produce useful aldehydes from simple
carbon sources via their corresponding carboxylic acids. Pathway selection leading to the relevant
carboxylic acid precursor depends on the category of target aldehyde. Fig. 1-1 illustrates known
aromatic and aliphatic acid biosynthesis pathways that can be engineered to result in several
familiar flavors and fragrances. In the case of vanillin, which has the largest annual market volume
of any flavor compound, previous reports have described engineered heterologous pathways
that use the natural aromatic amino acid precursor 3-dehydroshikimate as a branch-point
metabolite to the heterologous reactions (40, 41). Frost and coworkers constructed a system to
produce vanillin from glucose that used an engineered strain of E. coli to produce vanillate from
glucose, followed by extraction and reduction of vanillate to vanillin in vitro using purified
carboxylic acid reductase from Neurospora crassa (40). De novo biosynthesis of vanillin and
vanillin-p-D-glucoside was first demonstrated
in both Saccharomyces cerevisiae and
Schizosaccharomyces pombe and has since been optimized using flux balance analysis (41-43). In
initial reports, titers of de novo vanillin-p-D-glucoside were roughly 50 mg/L in batch flask cultures
18
(41) and 500 mg/L in 1.5 L continuous cultures (42). The company Evolva has improved and
commercialized this process (16).
Glucose
Pentose Phosphate Pathway
Glycolysis
Aromatic Amino Acid Synthesis
Arom 3tic aldehydes
A~%~
TCA
Cycle
I
Omo
Fatty Acid Synthesis
Aliphatic aldehydes
Figure 1 - 1. Overview of natural metabolic pathways that can be harnessed for the conversion
of glucose to valuable aromatic and aliphatic aldehydes through carboxylic acid intermediates
based on E. coli metabolism. Aldehydes can also be obtained from the 2-keto acid pathway,
terpenoid pathways, and other pathways.
19
Among flavor compounds, benzaldehyde has the second largest annual market volume
after vanillin (44). Aromatic amino acid biosynthesis could also be used to engineer a microbial
pathway to benzaldehyde, potentially from phenylalanine as the starting endogenous
metabolite. Formation of benzaldehyde was reported after phenylalanine addition to cell extract
of Lactobacillus plantarum (45). In plants, benzaldehyde is derived from phenylalanine,
potentially from 1-oxidative and non-1-oxidative pathways (46). Recent work has uncovered key
steps in the P-oxidative pathway that can lead to synthesis of benzoate, which could serve as the
precursor to benzaldehyde in an engineered microbial pathway (47).
Aliphatic aldehydes can be obtained using pathways that result in free fatty acids (FFAs).
Although microbial FFAs have been produced for decades, recent work has demonstrated the
potential for obtaining advanced fuels or valuable chemicals as derivatives of FFAs (48-51).
Addition of suitable carboxylic acid reductases could potentially result in production of C4-C18
aliphatic aldehydes. Microbial synthesis of other valuable aldehyde classes, such as terpenoid
aldehydes, could potentially occur in E. coli using variations of previously engineered terpenoid
pathways (52).
As mentioned earlier, commercial entities have actively pursued aldehyde biosynthesis
routes using engineered microbes. Table 1-1 contains an overview of relevant published
aldehyde biosynthesis patent applications during the past 30 years. These patents were grouped
into three types of dominant routes of aldehyde biosynthesis. Although the third category (i.e.,
engineered microbes) pertains most to the topic of this thesis, the other two categories of
processes were included to provide context and perspective into chronological trends. For
example, during the 1980s and 1990s, industry patents on biotransformation processes featured
20
either isolated microbes or fruit homogenates. Commercial processes featuring fully de novo
aldehyde synthesis using engineered microbes appear to emerge only within the last decade. Of
course, an overview of patent literature does not account for industrial advances that were
retained as trade secrets.
21
NMI
Table 1 - 1. Relevant published aldehyde biosynthesis patent applications.
Dominant
Aldehyde
Applicant
Publicatio
n Date
Publication
Number
Patent Name
Relevant Claims
Grant (G) or
Application (A)
Takasago
Perfumery
Sep 6,
1988
US 4769243
A
Method for preparing
green aroma compounds
Use of ground soybeans to convert unsaturated
fatty acids to aliphatic aldehydes and alcohols
G
General Foods
Corporation
Feb 21,
1989
US 4806379
A
Process for producing a
green leaf essence
Use of strawberry homogenate to convert
linolenic acid to cis-3-hexanal and related
aldehydes
G
BASF
Oct 17,
1989
US 4874701
A
Preparation of
coniferylaldehyde by a
microorganism
Use of Arthrobacter globiformisDSM 3597 to
convert n-eugenol to coniferylaldehyde
G
Haarmann &
Reimer Gmbh
May 21,
1991
US 5017388
Process for the
preparation of vanillin
Use of certain species from the genera Serratia,
Klebsiella, or Enterobacter to convert eugenol or
isoeugenol to vanillin
G
Kraft General
Foods
Jul 7, 1992
US 5128253
A
Bioconversion process for
the production of vanillin
Use of ferulic acid degrading microorganisms
such as Aspergillus niger, Rhodotorula glutinis, or
Corynebacterium glutamicum to convert ferulic
acid to vanillin
G
Firmenich
Nov 7,
1995
US 5464761
A
Process for the enzymatic
preparation of aliphatic
alcohols and aldehydes
from linoleic acid or a
natural precursor
Use of lipoxygenase-containing soya flour and
lyase-containing guava homogenate to convert
linoleic acid to hexanal and related aldehydes
G
BASF
May 19,
1998
US 5753471
A
Biotechnological
preparation of alcohols,
aldehydes, and carboxylic
acids
Use of isolated microorganisms capable of
converting alkyl, alkenyl, aryl, and related
compounds to their oxidized forms, including
aldehydes
G
Biosynthesis Route
Biotransformation
using
homogenates or
natural
microorganisms
_________________________________________________ J___________________________________________________________________________________________
L
I._____________________________
22
using purified
carboxylic acid
reductases
De novo synthesis
using engineered
microbes
harboring
recombinant
aldehyde
biosynthetic genes
(e.g., car, aar, kivD)
University of
Iowa
Aug 18,
1998
US 5795759
A
Carboxylic acid reductase,
and methods of using
same
A purified carboxylic acid reductase (Car) enzyme
from Nocardia iowensis, and use of it to convert
Synthesis of vanillin from
a carbon source
Use of an engineered microbe expressing
recombinant DHSD and COMT as part of a
metabolic pathway from glucose to vanillic acid,
followed by reduction of vanillic acid to vanillin
using a purified Car
G
G
vanillic acid to vanillin
Michigan State
University
Apr 16,
2002
US 6372461
University of
Iowa
Sep 16,
2008
US 7425433
B2
Carboxylic acid reductase
polypeptide, nucleotide
sequence encoding same
and methods of use
Use of Car to convert aromatic, aliphatic, and
acyclic carboxylic acids to corresponding
aldehydes
G
Archer-Daniels-
Feb 17,
US 7491854
2009
B2
Enzymatic method of
making aldehydes from
fatty acids
Use of Car to convert fatty acids ranging from C6C32 to corresponding aldehydes
G
Midland
DuPont
Aug 29,
2006
US 7098000
B2
Method for production of
Use of an engineered microorganism to convert
fermentable carbon sources to
diaponeurosporene monoaldehyde,
diapocarotene monoaldehyde, or diapocarotene
dialdehyde
G
Jan 17,
US 8097439
Methods and
Use of engineered microbes containing
G
2012
B2
compositions for
recombinant Car homologues to convert
producing fatty aldehydes
carbohydrates to aliphatic aldehydes
LS9
B1
C30-aldehyde
carotenoids
LS9
Sep 18,
2012
US 8268599
B2
Method for producing a
fatty alcohol or fatty
aldehyde
Use of acyl-ACP reductases to convert acyl-ACPs
to aliphatic aldehydes
G
International
Flavors
Fragrances,
Evolva
Feb 14,
WO 201302
2881 Al
Compositions and
methods for the
biosynthesis of vanillin or
Use of a microbe expressing recombinant AROM
and/or COMT to convert glucose to vanillin or
vanillin-beta-d-glucoside
A
2013
&
In vitro conversion
of acid substrates
vanillin-beta-d-glucoside
23
-
University of
California
T
Dec 27,
2013
T
WO 201319
T
2237 Al
Escherichia coli
engineered for
isobutyraldehyde
7
Use of an E. coli strain with reduced
isobutyraldehyde reductase activity to
accumulate isobutyraldehyde
A
production
Easel
Biotechnologies
Jan 9,
2014
US
2014001123
1 Al
Microbial synthesis of
aldehydes and
corresponding alcohols
Use of an engineered microbe to convert glucose
to short fatty aldehydes, followed by removal of
aldehydes from the fermentation medium and
conversion to alcohols ex vivo
A
Genomatica
Apr 24,
2014
WO
2014062564
Al
Microorganisms and
methods for production
of specific length fatty
alcohols and related
compounds
Use of a microbe expressing malonyl-CoA
independent (or dependent) fatty acyl-CoA
elongation pathways to produce fatty acids,
aldehydes, and alcohols
A
Evolva
Sep 4,
US
2014024866
Methods and materials
for recombinant
production of saffron
compounds
Use of a microorganism expressing recombinant
pathways to convert glucose to picrocrocin,
safranal, crocin, crocetin, or crocetin esters
A
2014
8 Al
I ________
I _________
I.
L __________________________________
___________
24
1.3. Minimizing endogenous conversion of aldehydes to alcohols
Despite known routes to a variety of aldehydes, microbial aldehyde production is
hindered by the rapid endogenous conversion of nearly all aldehydes to their corresponding
alcohols. For example, when expression of recombinant CarNi was first reported in E. coli,
aromatic acids supplied to culture media were rapidly converted into aromatic alcohols (22). Even
in E. coli, the most genetically well-understood organism, numerous uncharacterized genes were
thought to contribute to this activity. It is worth highlighting here that although oxidation of an
aldehyde to a carboxylic acid is thermodynamically more favorable than reduction of a carboxylic
acid to an aldehyde, endogenous aldehyde oxidation does not appear to be significant for most
aldehydes of interest in model microbes. On the other hand, endogenous aldehyde reduction has
been thoroughly documented in the literature.
In 2012, Rodriguez and Atsumi reported accumulation of isobutyraldehyde in E. coli by
sequentially deleting eight genes encoding putative isobutyraldehyde reductases (yqhD, adhP,
eutG, yiaY, yjgB (now ahr), betA, fucO, and eutE) (53). When individually overexpressed, five of
these genes displayed activity toward isobutyraldehyde. The engineered deletion strain
increased isobutyraldehyde production from 0.14 g/L/OD 6oo to 1.5 g/L/OD6oo and decreased
isobutanol production from 1.5 g/L/OD6oo to 0.4 g/L/OD6oo. Although isobutanol formation still
occurred, this study suggested that the number of gene deletions required to mitigate conversion
of a particular aldehyde may be a manageable quantity. A major goal of the research documented
in this thesis is to elucidate the number of relevant aldehyde reductases for other model
aldehydes, starting with benzaldehyde.
25
..
............
1.4. Enhancing bioconversion of aldehydes to other chemical classes
Microbial aldehyde accumulation is expected to enable biosynthesis of several previously
problematic compounds that can be derived enzymatically from aldehyde intermediates (Fig. 12). One aldehyde-derived molecule of interest is L-phenylacetylcarbinol (L-PAC), a chiral
precursor to the pharmaceutical ephedrine (12-15). Although whole cell catalysts have been used
for L-PAC synthesis for a long time, significant benzyl alcohol byproduct formation occurs from
their use, resulting in low yields (12). In addition to PDC, other enzymes capable of catalyzing
chiral carboligations of aldehyde substrates have been discussed (54).
0
R
R
H
aldehydes
as products
alkanes
0
0
reductionl
----- -----
reductiot
+
cellular
metabolites
OH
R)
R
oxidation
carboxylic acids
R
aldehydes
OH
oxidato
alcohols
OH
R
R
R
NH2
primary amines
0
chiral condensations
other potential aldehyde-derived products:
- nitroalcohols;
-
OH
R
esters
- 13-amino-carbonyls
chiral cyanohydrins
Figure 1- 2. Potential biocatalytic and metabolic engineering opportunities that could be enabled
by, or enhanced by, microbial aldehyde accumulation.
26
A similar challenge of limiting unwanted flux from aldehyde intermediates to alcohol
byproducts has been encountered in the context of alkane production. The final step to alkane
biosynthesis features the conversion of a C,, aldehyde to a C,,- alkane catalyzed by an aldehyde
decarbonylase or aldehyde deformylating oxygenase (25, 55-59). Although the problem of
alcohol byproduct formation has been described extensively, very few reports of alkane
biosynthesis have used strains engineered with deletions of aldehyde reductases. A later chapter
in this thesis will investigate the ability to reduce alcohol byproduct formation and the resulting
effect on alkane titers.
In addition to chiral carboligations and decarbonylations, aldehyde substrates can
participate in numerous other enzyme-catalyzed reactions (Fig. 1-2); for example, transamination
to form primary amines (60, 61), hydrocyanation to form chiral cyanohydrins (62), Henry
reactions to form nitroalcohols (63), Baeyer-Villager oxidation to form esters (64), and Mannich
reactions to form 0-amino-carbonyl compounds (65, 66). Some of the aforementioned reactions
have already been demonstrated to be functional in a cellular context using resting E. coli cells
(62, 67). Microbial aldehyde accumulation would enable potential synthesis of these compounds
using metabolically-active cells that can supply and regenerate expensive cofactors. Synthesis of
some of these products may also be achieved using glucose or other simple sugars as the sole
carbon source. In addition, biocatalytic oxidation of exogenously supplied alcohols (37-40, 64,
72) would be more effective in the absence of aldehyde reduction. In theory, any of the classes
of aldehyde-derived compounds enabled in the absence of aldehyde reduction could also be
obtained directly from the corresponding primary alcohols using a single engineered microbe.
1.5. Addressing aldehyde toxicity
27
If microbial aldehyde accumulation could be engineered, the next impediment to
consider would be aldehyde toxicity. Observable toxicity is manifested by inhibition of microbial
growth in the presence of aldehydes, but morphological changes have also been reported (68).
In most cases, the extent of toxicity seems to depend on the aldehyde but may also depend on
the choice of microorganism. Cinnamaldehyde, for example, is known to be a potent
antimicrobial (69). In the case of vanillin, Zaldivar et. aL. found that 1.5 g/L of vanillin completely
inhibited growth of the E. coli strains examined (68). The same study investigated the effect of
exposing E. colito several representative aromatic aldehyde products of hemicellulose hydrolysis
and found that toxicity was directly related to the hydrophobicity of the aldehyde. The
relationship with hydrophobicity suggested that a hydrophobic target, such as the cell
membrane, may be involved. However, none of these aldehydes caused sufficient membrane
damage to allow the leakage of intracellular magnesium (68). Another study investigated the
toxicity of four aldehydes (furfural, 5-hydroxymethylfurfural, vanillin, and syringaldehyde) on
Candida tropicalis and found that vanillin was the most toxic, followed by syringaldehyde,
furfural, and 5-hydroxymethylfurfural (70). The influence of the structural elements of vanillin
and related compounds on antifungal activity has also been examined and differences in
antifungal activity were found (71). However, when the effect of five aldehydes on the growth of
the oleaginous yeast Trichosporon fermentans was investigated, no relationship was found
between the hydrophobicity and toxicity of the aldehyde (72).
The E. coli strains investigated by Zaldivar et. aL. were not engineered to have minimal
aldehyde reductase activity, and later studies from the same group suggested that growth
inhibition may be caused by NADPH consumption resulting from aldehyde reduction (73, 74).
28
Two genes (dkgA and yqhD) were found to be silenced in an evolved furfural-resistant strain.
Expression of these genes, which encode enzymes with low KM values for NADPH, decreased
furfural tolerance (73). In a separate investigation, transcriptome data was analyzed before and
after exposure to furfural. Several lines of evidence suggested that cysteine and methionine
biosynthesis was upregulated in order to combat a limitation in sulfur assimilation due to NADPH
depletion (74).
A deeper understanding of precisely how aldehydes cause harm to cells may enable
engineering strategies to surmount particular modes of toxicity. Certain aldehydes may be
involved in far more detrimental mechanisms of toxicity than others. For example, acetaldehyde
has been shown to induce single-strand and double-strand breaks in DNA (75). Several aliphatic
aldehydes are products of lipid peroxidation and have been implicated in forming adducts on a
variety of biological macromolecules and as second messengers of reactive oxygen species (ROS)
(76-78). However, the precise relationship between aldehydes and ROS is unclear. For example,
it was recently shown that resistance of E. coli to exogenous methylglyoxal is conferred by
decreased expression of sodC(79). This is a surprising result given that sodCencodes a superoxide
dismutase, which breaks down ROS (80). There are numerous other potential mechanisms of
aldehyde toxicity. Given the importance of lignocellulose utilization, potential mechanisms of
toxicity for furfural in particular have been extensively reviewed and include mechanisms not
described here (81, 82).
Until precise mechanisms of aldehyde toxicity are elucidated, there are some general
engineering strategies that can be employed. Some bacteria naturally evolved solutions to
aldehyde toxicity beyond rapid reduction of aldehydes, such as protein microcompartments that
29
feature aldehyde intermediates (83, 84). If control of selective metabolite transport through the
protein shells were achieved, then the engineering of these compartments for biosynthesis of
new aldehyde-derived products may aid in limiting the pool size of free aldehyde intermediates
(85). Independent of the mode of toxicity, in situ separation using stripping (86), two-phase
systems (87), or selective resins (88) may result in increased production of aldehydes as end
products. Many aldehydes of interest are hydrophobic and volatile, which are properties that aid
separation from aqueous-based fermentation processes. In the event that precise mechanisms
of aldehyde toxicity become known and prove to be insurmountable, then efforts should shift
towards microbial engineering of aldehyde intermediates for synthesis of aldehyde-derived
products. In addition, the issue of aldehyde toxicity can be circumvented entirely with the use of
a cell-free or in vitro biosynthetic process, which is a topic that Chapter 4 of this thesis will
explore.
1.6. Thesis organization
The work documented in this thesis sought to answer a relatively orderly set of questions
motivated by Chapter 1 and by results from later chapters. Chapter 2 begins by exploring the
question of whether endogenous reduction of two simple aromatic aldehydes in E. coli can be
decreased using a series of gene deletions. If so, then what might the contributions of these genes
be towards the overall native level of endogenous aromatic aldehyde reduction? In addition,
Chapter 2 asks whether the decreased endogenous aromatic aldehyde reduction could be useful
for metabolic engineering or biocatalytic applications in the flavor and pharmaceutical industries,
such as biosynthesis of vanillin from glucose or biosynthesis of L-phenylacetylcarbinol from
glucose and exogenously supplied benzaldehyde. Given the ability of the engineered host strain
30
to accumulate both aromatic aldehydes tested, Chapter 3 poses the question of whether the
same genes targeted for deletion mightgovern the reduction of structurally unrelated aldehydes,
such as aliphatic aldehydes. If so, then could the same engineered host strain be used to reroute
aliphatic aldehydes to n-alkanes instead of primary alcohols, given that the former class of
biofuels are more compatible with current petroleum and automobile infrastructure? Next,
Chapter 4 examines how effectively aldehyde biosynthesis can occur outside of the cell using
purified enzymes. Previously, the use of carboxylic acid reductases to produce aldehydes in vitro
has been hampered by limited turnover. Chapter 4 provides some insight into why this
phenomenon occurs and explores the possibility of in vitro aldehyde biosynthesis as an
alternative to microbial aldehyde biosynthesis given that it completely circumvents the problem
of product toxicity. Providing a foundation for future metabolic engineering efforts, Chapter 5
returns to de novo biosynthesis of the model aldehyde vanillin using E. coli and seeks to
understand what biochemical reaction steps in the engineered pathway limit vanillin production.
Finally, Chapter 6 synthesizes lessons learned along the graduate research journey and
documents potential areas for further investigations.
31
Chapter 2: Engineering synthesis and accumulation of aromatic aldehydes in E.
coli
Portions of this chapter are adapted from the following manuscript: Kunjapur et al (2014), J. Am.
Chem. Soc., 136 (33) 11644-11654. Results presented were obtained with the help of Yekaterina
Tarasova.
2.1. Introduction
Chapter 1 discussed applications of numerous aldehydes and aldehyde-derived products.
Based on ease of detection and commercial relevance, aromatic aldehydes were initially focused
on for this thesis. Two model aromatic aldehydes are benzaldehyde, which is structurally the
simplest aromatic aldehyde, and vanillin. Biotechnological production of vanillin is of special
interest given that less than 1% of the 16,000 tons of vanillin sold annually originates from vanilla
beans (89). Due to the limited supply and high price of extract from the vanilla bean, most of the
market consists of vanillin that is chemically synthesized from either lignin or petroleum.
Consumer preference for flavors in which no chemicals are used [i.e., natural flavors (90)] has led
to a price differential of $1,200-$4,000/kg for natural vanillin compared to $15/kg for artificial
vanillin (89).
As discussed in Chapter 1, the primary barrier to overproduction of aromatic aldehydes
in engineered microorganisms is the rapid conversion of desired aldehydes into undesired
alcohols by numerous endogenous enzymes (41, 91). Several factors motivate the selection of E.
coli for identification and deletion of genes that encode aromatic aldehyde reductases. The
superior tools and know-how established with this organism enable rapid evaluation of gene
32
targets. Strategies to increase flux from central metabolism to the biosynthesis of aromatics in E.
coli are well-documented (92-95) and thus expected to facilitate swift improvements in aromatic
aldehyde production if their accumulation were realizable. Furthermore, E. coli K-12 strains have
been used to produce food additives designated as GRAS (Generally Recognized As Safe), such as
chymosin (96). Knowledge of the E. coli genes responsible for aromatic aldehyde reductase
activity can also help inform efforts to engineer other microbial hosts.
In this chapter, we describe how we began this endeavor by exploring serial deletions of
different combinations of genes that were reported to act on benzaldehyde, which is structurally
the simplest aromatic aldehyde. Our search initially focused on aldo-keto reductases (AKRs),
which form a superfamily of enzymes that have broad substrate specificity and convert aldehydes
and ketones to alcohols in the presence of NADPH (97). In the few previously reported attempts
at constructing microbial pathways to produce aldehydes, only genes encoding alcohol
dehydrogenases (ADHs) were targeted for deletion (41, 53). Unlike AKRs, ADHs typically use
NADH as their co-factor, and under anaerobic conditions they perform the important function of
recycling co-factors. We hypothesized that rational deletion of AKRs with activity on
benzaldehyde, in addition to the deletion of select ADHs, would be a promising route to
engineering the accumulation of aromatic aldehydes in E. coli. After pursuing this strategy, we
report the construction of an E. coli MG1655 strain with reduced aromatic aldehyde reduction
(RARE) that can serve as a platform for the synthesis of aromatic aldehydes with minimal or no
conversion to their corresponding alcohols. This chapter concludes with the use of the RARE
strain to enable the synthesis of vanillin and L-PAC, demonstrating the utility of this particular
33
engineered strain and the general approach of rationally combining gene knockouts to overcome
a highly redundant endogenous activity.
2.2. Materials and Methods
2.2.1. Strains and plasmids
E. coli strains and plasmids used in this study are listed in Table 2-1. Molecular biology
techniques were performed according to standard practices (98) unless otherwise stated.
Molecular cloning and vector propagation were performed in DH5a. All targeted genes were
deleted from E. co/i K-12 MG1655(DE3). The genes dkgB, yeaE, yahK, yjgB, endA, and recA were
deleted using donor strains from the Keio collection (99) and P1 transduction (100). P1
bacteriophage was obtained from ATCC (25404-B1). The operon encoding yqhC-dkgA was
deleted using the A Red system (101). To generate homology, three pairs of oligonucleotides were
used as PCR primers. These and other oligonucleotides are shown in Table 2-2. Oligonucleotides
were purchased from Sigma. Q5 High Fidelity DNA Polymerase (New England Biolabs) was used
for DNA amplification. In all cases of gene deletions, pCP20 was used to cure the kanamycin
resistance cassette (101).
34
Table 2 - 1. Strains and plasmids used in this study.
Invitrogen
DH1OB
I- WULIaCAM15 A(IaCZYA-argF) U169 recAl endAl fSGR17 (rK-,
mK+) phoA supE44 X- thi-1 avrA96 reIA1
F-mcrA A(mrr-hsdRMS-mcrBC) (D80acZAM15 AIacX74 recAl endAl
araDl39A(ara,leu)7697 galU galK k rpsL nupG
MG1655
MG1655(DE3)
F k ilvG- rfb-50 rph-1
F k iIvG- rfb-50 rph-1 (DE3)
ATCC 700926
Ref. (102)
MG endA- recA-
MG1655(DE3) AendA ArecA
Ref. (103)
JW0197-1
JW1770-5
JW0317-1
JW5761-1
JW2912-1
BW26547
Subset 1
(AMKO01)
Subset 2
(AMKO02)
RARE endArecA- (AMK003)
AMKO04
AMKOO5
AMKO06
AMKO07
AMKO08
AMKO09
AMK010
AMK011
AMK012
AMKO13
AMK014
Subset 3
(AMKO16)
AMKO17
RARE
(AMKO18)
AMKO19
Subset 4
(AMK035)
AMK036
RARE endA- recAvanillin
MG endA- recAvanillin
AdkgB726::kan
AyeaE778::kan
AyahK767::kan
AyjgB740::kan
ArecA635::kan, recA+
MG1655(DE3) AdkgB AyeaEAyahKAyjgB
CGSC 12026
CGSC 9486
CGSC 8516
CGSC 11992
CGSC 10253
CGSC 7652
This study
MG1655(DE3) A(yqhC-dkgA)::kan
This study
dendA720::kan
MG1655(DE3) AdkgB AyeaEA(yqhC-dkgA) AyahK AyjgB AendA
Invitrogen
ArecA
This study
AMK0O3 harboring pETDuet-1 and pACYC-car-sfp
AMKO01 harboring pETDuet-1 and pACYC-car-sfp
AMKO02 harboring pETDuet-1 and pACYC-car-sfp
AMKO03 harboring pET-dkgB and pACYC-car-sfp
AMKO03 harboring pET-yeaE and pACYC-car-sfp
AMKO03 harboring pET-yahK and pACYC-car-sfp
AMKO03 harboring pET-yjgB and pACYC-car-sfp
AMKO03 harboring pET-yqhC and pACYC-car-sfp
AMK003 harboring pET-yqhD and pACYC-car-sfp
AMK003 harboring pET-dkgA and pACYC-car-sfp
MG1655(DE3) AendA ArecA harboring pETDuet-1 and pACYC-car-sfp
MG1655(DE3) AdkgB AyeaE A(yqhC-dkgA)::kan
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
This study
MG1655(DE3) AdkgB AyeaE
MG1655(DE3) AdkgB AyeaE A(yqhC-dkgA) AyahK AyjgB
This study
This study
AMKO16 harboring pETDuet-1 and pACYC-car-sfp
MG1655(DE3) A(yqhC-dkgA) AyahKAyjgB::kan
This study
This study
AMK035 harboring pETDuet-1 and pACYC-car-sfp
AMKO03 harboring pET-OMT-asbF and pACYC-car-sfp-aroG*
This study
This study
MG1655(DE3) AendA ArecA harboring pET-OMT-asbF and pACYCcar-sfp-aroG*
This study
35
RARE
pRSF/PDCE473Q
RARE
pRSF/PDCE473Q
pACYC/Car/Sfp
MG
pRSF/PDCE473Q
pACYC/Car/Sfp
AMKO03 harboring pRSF/PDCE473Q
This study
AMKO03 harboring pRSF/PDCE473Q and pACYC-car-sfp
This study
MG1655(DE3) AendA ArecA harboring pRSF/PDCE473Q and pACYCcar-sfp
This study
pCP20
pKD13
Xc1857 (ts), X pr Repts, AmpRJ
CGSC 7629
CGSC 7633
pKD46
pETDuet-1
CGSC 7739
Novagen
pRSFDuet-1
oriR101, repA1O1, AmpR, araC, araBp-AyAe-Aexo
AmpR, lad, T7/ac
CmR, lac, T7iac
KanR, lad, T7/ac
pACYC-car-sfp
pACYCDuet-1 harboring caropt (carboxylic acid reductase from
This study
pACYCDuet-1
CmR, A pr FLP
oriRy, AmpR, kan
Novagen
Novagen
Nocardia iowensis, codon optimized for expression in E. coli) and
sfpopt (phosphopantetheinyl transferase from Bacillus subtilis, codon
optimized for expression in E. coli)
pET-dkgB
pET-yeaE
pET-yahK
pET-yjgB
pET-yqhC
pET-yqhD
pET-dkgA
pET-OMT
pETDuet-1
pETDuet-1
pETDuet-1
pETDuet-1
pETDuet-1
pETDuet-1
pETDuet-1
pETDuet-1
harboring dkgB from E. coli MG1655
harboring yeaE from E. coli MG1655
harboring yahK from E. coll MG1655
harboring yjgB from E. coli MG1655
harboring yqhCfrom E. coli MG1655
harboring yqhD from E. coli MG1655
harboring dkgA from E. coli MG1655
harboring Hs-S-COMTopt (catechol O-methyltransferase
from Homo sapiens, codon
pET-OMT-asbF
optimized
for expression
This study
This study
This study
This study
This study
This study
This study
This study
in E. coli)
pETDuet-1 harboring Hs-S-COMTopt and asbFopt (dehydroshikimate
dehydratase from Bacillus thuringiensis, codon optimized for
This study
expression in E. coli)
pS4
Plasmid containing the shikimate module, version 4, kindly provided
by the Keasling Lab at UC Berkeley. (Source of aroG*)
Ref. (104)
This study
pACYC-car-sfp-
pACYC-car-sfp plasmid harboring feedback-resistant aroG*from E.
aroG*
coli
pRSF/PDC
pRSFDuet-1 harboring the gene encoding PDC from Zymomonas
mobilis
pRSFDuet-1 harboring the gene encoding the mutant PDCE473Q
pRSF/PDCE473Q
This study
This study
36
Table 2 - 2. Oligonucleotides used in this study.
(yqhC-dkgA)::kan A F
(yqhC-dkgA)::kan A R
(yqhC-dkgA)::kan B F
TTCCCCGTTCCCGGTTGCTGTACCGGGAACGTATrAGTGTAGGCTGGAGCTGCTTC
GGTAGCGGAACATTACCGCCACCGGGAGAATTGCATGTTATCCGTCGACCTGCAGTT
GTTAACCGCTGGCTTGTTAGGCACGCTGTTTGTGGTGATTAAAAAAAAATACTGTAACGCCTGA
CGATTTCCCCGTTCCCGG
(yqhC-dkgA)::kan B R
GAACAAGGAAGAGTAACAACGGGCGGGACGCGAGGGGAATAAATGATTTCTGAAAAGTCCGG
TAGCGGAACATTACCGCC
(yqhC-dkgA)::kan C F
GTTAAACGCCATGAAGATCAGGTAATGACGTTCCTGATGATCCTGCCAATTGCCTTGTTAACCG
CTGGCTTGTTAGGC
(yqhC-dkgA)::kan C R
GTTC1TVUAAGAGCTTCCGGCTCTGCATGATGATGTCCTTATATTTGGCATTCCTGAACAAGGAA
GAGTAACAACGGGC
pET PIPE F
pET PIPE R
dkgB PIPE F
dkgB PIPE R
yeaE PIPE F
yeaE PIPE R
yahK PIPE F
yahK PIPE R
yjgB PIPE F
yjgB PIPE R
yqhC PIPE F
yqhC PIPE R
yqhD PIPE F
yqhD PIPE R
dkgA PIPE F
dkgA PIPE R
dkgB KO ver F
dkgB KO ver R
yeaE KO ver F
yeaE KO ver R
yahK KO ver F
yahK KO ver R
yjgB KO ver F
yjgB KO ver R
yqhC-dkgA KO ver F
yqhC-dkgA KO ver R
endA KO ver F
endA KO ver R
recA KO ver F
recA KO ver R
Sfp F
Sfp R
AroG* F
AroG* R
PDC WT F
PDC WT R
PDCE473Q F
TGCTTAAGTCGAACAGAAAGTA
CATGGTATATCTCCTTCTTAAAGTTAAAC
CTTTAAGAAGGAGATATACCATGGCTATCCCTGCATTTG
CTTTCTGTTCGACTTAAGCATTAATCCCATTCAGGAGCC
CTTTAAGAAGGAGATATACCATGCAACAAAAAATGATTCAATTTAGTG
CTTTCTGTTCGACTTAAGCATCACACCATATCCAGCG
CTAAGAAGGAGATATACCATGAAGATCAAAGCTGTTGGTG
CTTTCTGTTCGACTTAAGCATCAGTCTGTTAGTGTGCGATTA
CTTTAAGAAGGAGATATACCATGTCGATGATAAAAAGCTATGC
C1TCTGTTCGACTTAAGCATCAAAAATCGGCMTTCAACAC
CTTTAAGAAGGAGATATACCATGCTACAAAATTGCGCAC
CTCTGTTCGACTrAAGCATTAATTCCCCTGCATCGC
CTTTAAGAAGGAGATATACCATGAACAACTTTAATCTGCACAC
CTTTCTGTTCGACTTAAGCATTAGCGGGCGGCTTC
CTTTAAGAAGGAGATATACCATGGCTAATCCAACCGTTATTAAG
CTTTCTGTTCGACTTAAGCATTAGCCGCCGAACTG
AATGCGGAAGAGATAAGTGCTGAA
GCCA1TT1GTTTCGGTCGTC
CAGTAACGCTAAATTCATTTGGCTG
GACTTCGGTCGCTC1T111TUTAC
CCTCGACACCATGTTCCAG
CTGCACTCTATTAGATATCCATTCATTTAATC
CACTGCATAGCGCATGATG
CGATAACTTCATGACCTAACACCATC
GGTAATTCTTCAAATACTGCAACGG
GGTCAGCGTAAAACGAACATG
GTTGGTTTGCCGCCAAA
CAGGCAGTACGGTACCGG
CATGGCTCCGTTATCGCA
GTGCGGAACAGGTCGATG
AAAAAAGCGGCCGCTAATAAAAGGAGATATACCATGAAAATCTATGGCATTTACAT
AAATTTCTTAAGTTACAGCAGTTCTTCGTAGCT
AAAAAAAGATCTGATGAATTATCAGAACGACGATTTAC
AAAAAACCTAGGCCTCCTTTAGATCCTTACCC
AAAAAACATATGAGTTATACTGTCGGTACCTATTTAGCG
AAAAAACCTAGGATTAGAGGAGCTTGTTAACAGGCTTACG
GTTACACCATCCAAGTTATGATCCATGATGGTCCGTAC
37
PDCE473Q R
dkgA-f qRT
dkgA-rqRT
dkgB-f qRT
dkgB-r qRT
yea E-fLqRT
yeaE-r qRT
yahK-f qRT
yahK-r qRT
yjgB-fqRT
yjgB-r qRT
yqhD-fqRT
yqhD-r qRT
GATCATAACTTGGATGGTGTAACCATAGTTATTGATCAAGAAG
GCTAATCCAACCGTTATTAAG
CGGTGATTACTTCCTCATTAC
TGATAACGAAGCCGCAGTAG
GGCTCTCTTTCAGACTTGGG
AACAGAAGTTGCTGCACTAC
GCAGACCGGTTAATGCT
AAGATCAAAGCTGTTGGTG
CACAGTAAGCGATTTCGATT
GGCGAACTGGAAGTTTATGAG
AATCCACCTGCACTTCAACA
CGAACAAATTCCTCACGATG
CTTTCAGGGCATCCAGAACT
The car gene from Nocardia iowensis (22), the sfp gene from Bacillus subtilis (105), and
the Hs-S-COMT gene from Homo sapiens (41) were synthesized and codon-optimized for
expression in E. co/i (GenScript). The asbF gene from Bacillus thuringiensis (106) was synthesized
as a DNA string codon-optimized for expression in E. coli (GeneArt, Invitrogen). The aroG* gene
was kindly provided by Professor Jay D. Keasling at the University of California, Berkeley (USA).
The gene encoding the pyruvate decarboxylase (PDC) enzyme from Zymomonas mobilis was
amplified from genomic DNA (ATCC 39676). The gene encoding the mutant PDC_E473Q enzyme
was generated using Polymerase Incomplete Primer Extension (PIPE) cloning (107). All codon-
optimized and mutant gene sequences are included in SI Text. E. coli AKR and ADH gene targets
were amplified from MG1655(DE3) genomic DNA using PCR and cloned into the Duet vector
system (Novagen) using PIPE cloning. Unless otherwise specified, all other genes were cloned
into the Duet vector system (Novagen) using restriction digest-based cloning. Restriction
enzymes and T4 DNA ligase were purchased from New England Biolabs. Propagated constructs
were purified using a QlAprep Miniprep Kit (Qiagen) and agarose gel fragments were purified
38
using a Zymoclean Gel DNA Recovery Kit (Zymo Research). All constructs were confirmed to be
correct by nucleotide sequencing (Genewiz).
2.2.2. Chemicals
The following compounds were purchased from Sigma: sodium benzoate, benzaldehyde,
benzyl alcohol, vanillic acid, vanillin, isovanillin, vanillyl alcohol, 3,4-dihydroxybenzoic acid, 3,4dihydroxybenzaldehyde, and dimethyl sulfoxide (DMSO). 3,4-Dihydroxybenzyl alcohol was
purchased from TCI America.
Biotechnology. Isopropyl
L-phenylacetylcarbinol was purchased from Santa Cruz
-D-1-thiogalactopyranoside (IPTG) was purchased from Denville
Scientific. Ampicillin sodium salt, chloramphenicol, and kanamycin sulfate were purchased from
Affymetrix.
2.2.3. Culture conditions
With the exception of growth rate experiments, all experiments were performed in 50 ml
Pyrex VISTA screw-cap culture tubes (Sigma), which contained 5 ml of culture in order to maintain
aerobic conditions and limit evaporation of volatile metabolites. Experimental cultures were
initiated using 1% (v/v) inoculum volumes of overnight culture that were transferred into either
LB medium or M9 minimal medium containing 1.2% glucose, incubated at 30*C, and agitated at
250 rpm. Overnight cultures were grown in 3 ml of the same medium in 14 ml round-bottom
tubes (Corning). In general, experiments were performed in triplicates, and results are presented
as averages with error bars representing one standard deviation.
For all substrate-feeding experiments excluding glucose as a substrate, filter-sterilized
and pH-neutralized substrates were added to the cultures upon induction with IPTG. During 5
39
mM benzoate-feeding experiments, cultures were induced with 1 mM IPTG between optical
densities (OD60 0) of 0.7-1.0. During 5 mM vanillate-feeding experiments, cultures were induced
between OD600 of 1.0-1.3. Culture medium was supplemented with 50 mg/L ampicillin and 17
mg/L chloramphenicol to provide selective pressure for plasmid maintenance.
For experiments testing the production of vanillin directly from glucose, 1.2% (w/v)
glucose was added prior to inoculation and cultures were induced with 0.5 mM IPTG between
OD600
of 0.8-1.1. Screw-caps remained tightly closed until sampling at final time points, but mass
balances in the liquid phase did not always close, indicating some evaporation of volatile
products. Either LB medium or M9 minimal medium was used. Culture medium was
supplemented with 50 mg/L ampicillin and 17 mg/L chloramphenicol to provide selective
pressure for plasmid maintenance.
For experiments testing the synthesis of L-phenylacetylcarbinol, 1.2% (w/v) glucose was
added prior to inoculation and cultures were induced with 0.5 mM IPTG between OD600 of 0.81.1. At induction, a filter-sterilized solution of 50 mM benzaldehyde in 10% DMSO was added to
the cultures, resulting in an initial concentration of 5 mM benzaldehyde and 1% DMSO. For these
experiments, culture medium was supplemented with either only 50 mg/L kanamycin or 25 mg/L
kanamycin and 17 mg/L chloramphenicol depending on whether strains were expressing Car.
Growth rates for wild-type and RARE strains were determined from 50 ml shake flask
cultures. Cultures were initiated with 1% inoculum volumes of overnight culture that were
transferred into either LB medium containing 1.2% glucose or M9 minimal medium containing
1.2% glucose. Overnight cultures were grown in the same media. In both cases, cultures were
40
incubated at 30*C and agitated at 250 rpm. The OD 6oo was measured regularly during exponential
growth using a DU800 UV/Vis spectrophotometer (Beckman Coulter).
2.2.4. Metabolite analysis
Culture samples were pelleted by centrifugation and aqueous supernatant was collected
for HPLC analysis using either an Agilent 1100 series or 1200 series instrument equipped with a
diode array detector. Wavelengths of 223, 242, and 192 nanometers were used to detect benzoic
acid, benzaldehyde, and benzyl alcohol, respectively. The benzoate family of analytes was
separated using an Aminex HPX-87H anion-exchange column (Bio-Rad Laboratories), with a
mobile phase consisting of 70% 5 mM H 2SO4 and 30% acetonitrile. All three compounds eluted
within 35 minutes at a flow rate of 0.4 ml/min. Column temperature was maintained at 300 C. All
chemicals reported in figures were quantified using calibration of standards on the HPLC
instrument and linear interpolation.
Compounds used in vanillin experiments were separated using a Zorbax Eclipse XDB-C18
column (Agilent) and detected using a wavelength of 280 nm. A gradient method used the
following solvents: (A) 50% acetonitrile + 0.1% trifluoroacetic acid (TFA); (B) water + 0.1% TFA.
The gradient began with 5% Solvent A and 95% Solvent B. The setting at 20 minutes was 60%
Solvent A and 40% Solvent B. The program restored the original ratio at 22 minutes and ended
at 25 minutes. The flow rate was 1.0 ml/min and all compounds of interest eluted within 15
minutes. Column temperature was maintained at 30*C.
Phenylacetylcarbinol was detected using a Zorbax Eclipse XDB-C18 column (Agilent) and
detected using a wavelength of 210 nm. A gradient method used the following solvents: (A) 100%
41
acetonitrile + 0.1% trifluoroacetic acid (TFA); (B) water + 0.1% TFA. The gradient began with 5%
Solvent A and 95% Solvent B. The setting at 20 minutes was 60% Solvent A and 40% Solvent B.
The program restored the original ratio at 22 minutes and ended at 25 minutes. Column
temperature was maintained at 30*C. The flow rate was 1.0 ml/min and the retention time of
phenylacetylcarbinol was 8.3 minutes. Peak area corresponding to a co-eluting and static
background peak in LB medium was subtracted in order to quantify concentrations of
phenylacetylcarbinol produced. Although enantiomeric excess (ee) was not determined, it has
been shown previously using chiral HPLC and near-UV circular dichroism spectroscopy that both
wild-type PDC and PDCE473Q catalyze the formation of the R stereoisomer in 98-99% ee (15).
2.2.5. Quantitative Reverse Transcription PCR (qRT-PCR)
For isolation of RNA and generation of cDNA in biological duplicate, two cultures of
MG1655(DE3) AendA ArecA were grown overnight in 3 ml of LB medium contained in 14 ml
round-bottom tubes. Each overnight culture was used to inoculate two cultures in 4.5 ml of LB
medium (1% v/v inoculum) contained in 50 ml PYREX VISTA tubes. Once cultures reached an
OD6oo of 0.7-0.9, benzaldehyde was added to two out of the four cultures to a final concentration
of 5 mM. After one hour of further incubation, 0.5 ml of cells were harvested for RNA isolation.
RNA protect bacterial reagent (Qiagen) was added to cells prior to centrifugation and lysis. Total
RNA was isolated using the illustra RNAspin Mini Isolation Kit (GE healthcare) with an on-column
DNasel treatment according to protocol. Turbo DNA-free reagents (Ambion) were used to further
treat isolated RNA for removal of genomic DNA. Next, QuantiTect Reverse Transcriptase (Qiagen)
was used to generate cDNA from 500 ng of total RNA for each isolation. Concentrations of RNA
and DNA were measured using a NanoDrop 2000 (Thermo Scientific).
42
Quantitative PCR was performed using an ABI 7300 Real Time PCR System Instrument
(Applied Biosystems). All samples analyzed by qPCR were performed in triplicate. 2 I of cDNA
from each RNA isolation was added to Brilliant 11 SYBR Green High ROX QPCR Mix (Agilent
Technologies) and 0.5 pM of appropriate primers to a final volume of 25 pl per well. Amplification
was performed according to the following program: an initial step of 950 C for 10 min, followed
by 40 cycles of 950 C for 30 s and 60*C for 1 min. The number of cycles to reach the threshold
(CT
value) was measured for each primer pair in triplicate samples of each cDNA. The "Auto
CT"
option in the 7300 System SDS RQ Study Software was used to determine the threshold values.
Sequences of primers used for analysis of dkgA, dkgB, yeaE, yqhD, yahK, and yjgB are listed in
Table 2-2. Primer sequences were designed to have melting temperatures ranging from 56-600 C
and to generate roughly 100 base pair amplicons. The specificity of primers was verified using gel
electrophoresis using gDNA from MG1655(DE3) AendA ArecA as a template. No-template and
no-RT controls confirmed that primer dimer formation was absent or negligible (i.e., CT values
greater than 33). Plasmids containing gene deletion targets were diluted to concentrations
ranging from roughly 10-4 to 10-8 ng/pI and analyzed in triplicate as standards during each
respective run. Using linear standard curves, CT values were used to quantify absolute
concentrations of reverse transcribed mRNA corresponding to each gene of interest, with and
without benzaldehyde treatment. All PCR efficiencies were similar.
2.3. Results
2.3.1. A combination of rationally targeted gene deletions enables benzaldehyde accumulation
in E. col
43
Deletions of AKR genes were guided by literature reported activities (108) of gene
products on benzaldehyde (Table 2-3). Protein BLAST was used to organize E. coli AKRs based on
sequence similarity to DkgA, the AKR with the greatest reported activity on benzaldehyde (Table
2-4). The gene encoding DkgA is located downstream of yqhD in the same operon, and
transcription of both genes is activated by the product of yqhC, which is located immediately
upstream of the operon (109). Given that YqhD is a broad substrate ADH that is also reported to
act on benzaldehyde (73, 110, 111), the entire operon was deleted. Benzaldehyde reductase
activity was unaffected by the deletion of this operon and by the subsequent deletion of the two
genes encoding AKRs with the greatest protein sequence similarity to DkgA (dkgB and yeaE). A
recent report revealing the activity of two E. coli cinnamyl alcohol dehydrogenases (YahK and
YjgB) on benzaldehyde was then discovered (112). A few months prior to the publication of that
report, overexpression of the yahK and yjgB genes had also been reported to improve the
production of aromatic alcohols in E. coil (113). After deleting these two additional genes, we
observed a significant change in the ability of E. coli to accumulate benzaldehyde. The final
engineered strain (RARE AendA ArecA) consists of nine gene deletions: dkgB, yeaE, yqhC, yqhD,
dkgA, yahK, yjgB, endA, recA. The genes endA and recA, which encode an endonuclease and a
recombinase, were deleted to increase plasmid stability. The deletion of yqhCforms a convenient
but nonessential control. The remaining six genes constitute the rationally targeted set.
44
Table 2 - 3. Reported activities of E. coli gene products on benzaldehyde in vitro.
DkgA (formerly YqhE)
3880
100a
Ref. 105
DkgB (formerly YafB)
2790
40a
Ref. 105
YeaE
520
30a
Ref. 105
YjgB
1305
Ref. 109
YahK
2 6 .7b
Ref. 109
YqhC
Not applicable (N/A)
N/A
YqhD
Not reported
N/A
Table 2 - 4. Results of protein BLAST sequence alignments used to organize deletion targets.
DkgB
2e-45
Y
YeaE
7e-22
Y
YdjG
2e-10
N
Target: YahK
YjgB
n/a
le-46
Y
Y
AdhP
3e-23
N
Target: YqhD
n/a
Y
FucO
le-21
N
To compare the ability of the RARE strain to accumulate aromatic aldehydes, four
additional strains containing complementary subsets of the total set of gene deletions were
constructed (Table 2-5). In short, the "Subset 1" strain contained the yqhC-yqhD-dkgA operon
intact, the "Subset 2" strain contained all gene targets that were not in the operon, the "Subset
3" strain contained only the targeted ADHs that were not in the operon, and the "Subset 4" strain
contained only the targeted AKRs that were not in the operon. All strains were also built to
45
express a recombinant and activated carboxylic acid reductase (Car). Car, from Nocardia
iowensis, has broad substrate specificity and was previously used to catalyze the formation of
benzaldehyde and vanillin from their corresponding acids in vivo in E. coli and in vitro (22).
Expression of Car was paired with expression of Sfp from Bacillus subtilis. Sfp is a
phosphopanthetheine transferase that has been shown to activate Car (23, 105). In these
experiments, pH-neutralized acid substrates were added to the medium to obtain greater
solubility and mimic product formation in an engineered pathway. The corresponding aldehydes
were then generated intracellularly by the action of Car on the acid substrate.
Table 2 - 5. E. coli strains featuring different combinations of gene deletions ("X" indicates
deletion).
dkgB
AKR
X
X
X
yeaE
AKR
X
X
X
dkgA
AKR
X
X
X
X
yqhC
activator
X
X
X
X
yqhD
ADH
X
X
X
X
yahK
ADH
X
X
X
yjgB
ADH
X
X
X
Accumulation of benzaldehyde was investigated in the six different strains (Figure 2-1A).
The wild-type strain and three out of the four strains containing subsets of gene knockouts
converted all of the supplied 5 mM benzoate into benzyl alcohol within 24 hours, with no
accumulation of benzaldehyde. On the other hand, the RARE strain accumulated 3.3
0.1 mM
benzaldehyde and displayed less than 12% conversion of benzaldehyde to benzyl alcohol (Fig. 246
1B). Similarly, the "Subset 4" strain also enabled the accumulation of benzaldehyde. The "Subset
4" strain contained the dkgB and yeaE genes intact in the genome, indicating that the deletions
of dkgB and yeaE were not necessary for benzaldehyde accumulation under these conditions.
A
Carm
Endegwnom
Emymes
0
KYOH
bzoabenzoate
NbA"d Nhd
PA,.
ATP
benzaldehyde
P
B
benzyl alcohol
C
5-
5-
Benzoate
Benzaldehyde
Benzyl alcohol
4.
2
4.
E
.
3
3Eu
S2-
2
52
011
MG
Subset 1
endA- recA-
Subset 2
Subset 3
Subset 4
RARE
endA- recA-
dkgB
yeaE
AKRs
dkgA
yqhC
yqhD
yahK
yjgB
ADHs
Figure 2 - 1. A combination of rational gene deletions enables benzaldehyde accumulation in E.
coli. (A) Scheme depicting intracellular formation of benzaldehyde from benzoate and
endogenous conversion to the byproduct benzyl alcohol. (B) Conversion of 5 mM benzoate after
24 hours in strains transformed with pETDuet-1 and pACYC-car-sfp. (C) Conversion of 5 mM
benzoate after 24 hours in RARE strains transformed with pACYC-car-sfp and a pET plasmid
harboring the gene indicated below the x axis.
Deleted genes were overexpressed individually alongside car in the RARE strain to
determine whether each gene could contribute to benzaldehyde reductase activity in vivo (Fig.
2-iC). Benzoate was supplied and formation of aldehyde and alcohol products was monitored
as before. Individual overexpression of each of the six target genes prevented detectable
accumulation of benzaldehyde. Conversely, benzaldehyde accumulated when the control gene
yqhC was overexpressed. For some strains, lower conversion rates were observed in this
47
-
..........
- -- --. - I..........
-
experiment relative to the previous experiment, which may have been due to lower expression
of car in the presence of native gene overexpression. Interestingly, individual overexpression of
dkgB and yeaE also restored the inability to accumulate benzaldehyde, even though the deletion
strain experiment demonstrated that these deletions were not required under these conditions.
The results from these two experiments suggest that native expression of dkgB and yeaE may be
minimal under these conditions and that, in general, overexpression experiments alone may
mislead efforts to determine the significance of gene deletions. qRT-PCR results provide further
support for our hypothesis of low levels of baseline expression of dkgB and yeaE compared to
the other targeted genes (Fig. 2-2). Furthermore, expression of these two genes is not
significantly different in the presence or absence of benzaldehyde, whereas expression of all
other targeted genes increases upon benzaldehyde addition.
-
500000
400000 -
*
300000 -
dkgA
dkgB
yeaE
yahK
yjgB
yqhD
-
200000
120000
C
100000
-
-
0
20000
0
No Benzaldehyde Added
Benialdehyde
Added
Figure 2 - 2. Estimated copies of reverse transcribed mRNA per nanogram of total RNA based on
quantitative reverse transcription PCR (qRT-PCR).
48
2.3.2. In vanillate-feeding experiments, the RARE strain eliminates conversion of vanillin into
vanillyl alcohol
We next investigated the utility of the engineered strain by attempting to produce and
accumulate other aromatic aldehydes. The RARE strain and subset deletion strains were next fed
vanillate to assess the effect of the same set of gene deletions on the undesired conversion of
vanillin to vanillyl alcohol (Fig. 2-3A). Our hypothesis was that a strain that displays minimal
reductase activity on benzaldehyde would also display low reductase activity on vanillin.
Although the probable physiological substrates of AKRs are smaller aldehydes such as
methylglyoxal (108), benzaldehyde and vanillin are structurally similar, with vanillin differing only
by the presence of additional hydroxyl and methoxy groups distant from the aldehyde group.
49
A
Carl#
vanillate
ATP
Endoge*0m4
vanliin
"W0
B_
Vanillate
V
5.
Vanillin
Vanillyl alcohol
vanilly alcoWh
OR
C
5
2
4.
Ewymes
4-
3-
3.
e
.
2.
1.
0.
MG
endA- recA-
Subset 1
Subset 2
Subset 3
Subset 4
RARE
endA- recA-
dkgB
yeaE
AI(Rs
dkgA
yqhD
yahK
yjgB
ADI~s
Figure 2 - 3. In vanillate-feeding experiments, the RARE strain eliminates conversion of vanillin to
vanillyl alcohol. (A) Scheme depicting intracellular formation of vanillin from vanillate and
endogenous conversion to the byproduct vanillyl alcohol. (B) Conversion of 5 mM vanillate after
48 hours in strains transformed with pETDuet-1 and pACYC-car-sfp. (C) Conversion of 5 mM
vanillate after 24 hours in RARE strains transformed with pACYC-car-sfp and a pET plasmid
harboring the gene indicated below the x axis.
Cultures were supplied with vanillate and compared 48 hours after induction due to the
slower kinetics of Car and endogenous enzymes on vanillate and vanillin relative to benzoate and
benzaldehyde (22). After 48 hours, the RARE and "Subset 4" strains were the only strains that
resulted in no detectable formation of vanillyl alcohol (Fig. 2-3B). As before, in order to
investigate whether each gene could contribute to vanillin reductase activity in vivo, individual
overexpression of deleted genes in the RARE strain was examined for the presence of the alcohol
after 24 hours. All gene products were active on vanillin, but surprisingly the overexpression of
targeted AKRs (DkgB, YeaE, and DkgA) resulted in significantly more vanillyl alcohol production
compared to the overexpression of targeted ADHs (YahK, YjgB, and YqhD) (Fig. 2-3C). As with
50
benzaldehyde, these results suggest that dkgB and yeaE expression may remain minimal in the
presence of vanillin and that overexpression results can mislead gene deletion efforts. These
results also validate that the RARE strain can be used for the bioconversion of multiple acid
substrates into their corresponding aldehydes.
2.3.3. The RARE strain enables production of vanillin from glucose in E. coli
To investigate whether the RARE strain could enable the production of an aromatic
aldehyde directly from glucose, a non-optimized pathway from glucose to vanillin was assembled
in the RARE and wild-type strains. As mentioned in Chapter 1, previous reports established a
route from glucose to vanillate in E. coil (40), and a route featuring the same metabolites but
different enzymes was also assembled in yeast (41). In both cases, the native branch-point
metabolite was 3-dehydroshikimate, which is part of the aromatic amino acid biosynthesis
pathway (Fig. 2-4A). A feedback-resistant form of E. coli aroG (aroG*) (104), which encodes a 3deoxy-D-arabinoheptulosonate 7-phosphate (DAHP) synthase, was included to ensure that flux
enters this endogenous pathway.
51
A
Aromatic
Anmno Acids
PPAA
*~PYR
PEP
MA6G
Glucose
-
AroB
F6P
DAHP
AroD
DHQ
-
A, oE
OHS
Shikimate
E4P
G3 P
B
3-deh
~
-----
Vanilin
-xr----
-
hikirnat
(OHS)
protacafchuate
protacatechusaldehyda
protwatchu c alcohol
OMT
Wanilate
C
llin
1.0
ailll
D
0.8
A0.4
0.5
0.4
I
0.6
S
alcohol
Protocatechuate
Protocatechualdehyde
Protocatechuic alcohol
g
c
0.3
SVanillate
0.2
Vanillin
Vaniltyl alcohol
0.2
0.1
0.0
0.01
MG endA- recA-
RARE endA-
recA-
MG endA- recA-
RARE enA-
recA-
+
Figure 2 - 4. The RARE strain enables production of vanillin from glucose in E. coli. (A) Scheme
depicting complete pathway from glucose to vanillin with overexpressed E. coli aroG* indicated
in bold typeface. [PYR = pyruvate, F6P = fructose 6-phosphate, G3P = glyceraldehyde 3phosphate, X5P = xylulose 5-phosphate, PEP = phosphoenolpyruvate, E4P = erythrose 4phosphate, DAHP = 3-deoxy-D-arabinoheptulosonate 7-phosphate, DHQ = 3-dehydroquinate,
DHS = 3-dehydroshikimate] (B) Scheme illustrating heterologous portion of pathway with two
possible undesired alcohol byproducts. (C) Concentration profiles of the six heterologous
metabolites of interest (including vanillin) when wild-type and RARE hosts are transformed with
pET-OMT-asbF and pACYC-car-sfp-aroG* plasmids and grown in LB + 1.2% glucose for 48 hours.
(D) Concentration profiles of the six heterologous metabolites of interest (including vanillin)
produced from glucose as a sole carbon source after 60 hours in the same strains grown in M9
1.2% glucose.
The heterologous pathway constructed for our experiments consists of three genes: asbF
from Bacillus thuringiensis (106), Hs-S-COMT from Homo sapiens (41, 114), and car (Fig. 2-4B).
Together with aroG* and sfp, a total of five genes were overexpressed. The asbF gene encodes a
52
3-dehydroshikimate dehydrogenase, which efficiently converts 3-dehydroshikimate into
protocatechuate. The Hs-S-COMT gene encodes a soluble O-methyltransferase (OMT) that has
activity on catechols and related compounds. Depending on the relative enzyme kinetics and
availability of co-factors, protocatechuate can either be converted into protocatechualdehyde by
Car or be converted into vanillate by OMT. The final step in the pathway is either the conversion
of protocatechualdehyde to vanillin by OMT or the conversion of vanillate to vanillin by Car.
Because this pathway can lead to the production of two possible alcohol byproducts
(protocatechuic alcohol and vanillyl alcohol), we expected vanillin titers to be greater when using
the RARE strain host rather than the wild-type host.
Fig. 2-4C displays the concentrations of the six metabolites of interest produced as a result
of the vanillin pathway 48 hours after induction in wild-type and RARE strains. As expected, the
dominant products generated using the wild-type host are the two alcohol byproducts, with
minimal formation of vanillin (0.014
0.001 mM). Conversely, the dominant products made by
the RARE strain are vanillin and its precursor aldehyde, protocatechualdehyde. No detectable
protocatechuic alcohol formed, revealing that the RARE host strain is capable of accumulating at
least three different aromatic aldehydes (benzaldehyde, vanillin, and protocatechualdehyde).
Although some vanillyl alcohol was detected, it represents less than 14% conversion of vanillin
formed. Using the RARE strain, the average vanillin titer was 0.78
0.02 mM (119
3 mg/L),
representing more than a 55-fold increase in production over the wild-type strain.
Finally, the same strains were cultivated in M9 minimal medium + 1.2% glucose instead
of LB medium + 1.2% glucose to determine whether vanillin could be produced from glucose as
a sole carbon source. Concentrations of the six metabolites of interest demonstrate that the
53
RARE strain enables the accumulation of vanillin in minimal medium (Fig. 2-4D). Under these
conditions, vanillin is the dominant product of the six metabolites of interest even after 60 hours.
Compared to results obtained from growth in LB medium, the ratios of protocatechuate to
protocatechualdehyde and vanillate to vanillin sharply increase in M9 minimal medium. Although
the build-up of protocatechualdehyde in the cultures grown in LB indicated that the 0methyltransferase
was
limiting,
the
greater
pool
size
of
vanillate
relative
to
protocatechualdehyde suggests that this enzyme was no longer limiting in M9. Given the
documented lack of specificity of the O-methyltransferase (41), the byproduct isovanillin was also
produced in these experiments (Fig. 2-5).
0.20
0.106
0.0
LB+1.2%Glucose:
MG endA- recA-
LB+1.2%Glucose:
RARE endA- recA-
M9+1.2%Glucose:
MG endA-recA-
M9+12%Glucose:
RARE enidA- recA-
Figure 2 - 5. Concentration of isovanillin at final time points from experiments (either 48 hours
for LB or 60 hours for M9) that examined the production of vanillin from glucose.
2.3.4. The RARE strain enables the synthesis of L-PAC in E. coil
We next sought to demonstrate the utility of the RARE strain as a platform for the
biocatalysis of products derived from aromatic aldehyde intermediates. Given the reactivity of
54
aldehyde functional groups, there are numerous enzymatic chemistries that may be enabled by
the accumulation of aromatic aldehydes. We were particularly interested in carboligations
because these reactions are known to yield chiral products. We hypothesized that use of the
RARE host strain could enhance synthesis of L-PAC. Wild-type and RARE strains were transformed
to express a recombinant mutant pyruvate decarboxylase (PDCE473Q) with improved kinetic
properties over the wild-type enzyme from Z. mobilis. Specifically, PDCE473Q displays inverted
partitioning between aldehyde release and carboligation compared to the wild-type PDC, with
an up to 100-fold preference for carboligation (15). In our case, the PDCE473Q enzyme was
expected to catalyze the condensation of benzaldehyde supplied exogenously and pyruvate
resulting from metabolism of glucose (Fig. 2-6A).
Expression of only PDCE473Q resulted in the synthesis of 2.59
0.04 mM PAC in the
RARE strain 24 hours after benzaldehyde addition, along with less than 4% reduction of
benzaldehyde to benzyl alcohol (Fig. 2-6B). We also observed roughly 8% oxidation of
benzaldehyde to benzoate. To prevent net oxidation of benzaldehyde, we expressed both Car
and PDCE473Q. In this case, no benzoate was detected, and similar levels of PAC and benzyl
alcohol were produced in the RARE strain. Surprisingly, wild-type strains transformed with the
same constructs resulted in no synthesis of PAC (Fig. 2-6B). This result suggested that the
timescale of benzaldehyde reduction is much shorter than 24 hours. To investigate this further,
a time course study was performed to monitor the conversion of benzaldehyde to benzyl alcohol
using the wild-type strain expressing Car and PDCE473Q (Fig. 2-6C). Within just 2 hours of
supplying 5 mM benzaldehyde, all of the benzaldehyde was either reduced to benzyl alcohol (-4
mM) or lost to the headspace ('1 mM). This result explains why the wild-type E. coli host strain
55
failed to produce any PAC under these conditions. The RARE strain enables the synthesis of PAC
in E. coli by extending the duration of benzaldehyde availability more than 10-fold, to the relevant
timescale of PDCE473Q kinetics.
A
0
H
-~
OH
benzaldehyde
0
Glycolys
glucose........--
OH
CO 2
(L)-phenylacetylcarbinol
0
pyruvate
B
5-
U,
4-
CU
C
5
Benzoate
Benzaldehyde
Phenylacetylcarbinol
Benzyl alcohol
4
09
0
-C
-C
3-
-C
C
W
2-
to
0
U
2-
C
0
0
M
U
0RARE
endA- recApRSF/PDCQE473Q
RARE
endA- recApRSF/PDCE473Q
pACYC/Car-Sfp
MG
endA- recApRSF/PDCE473Q
pACYC/Car-Sfp
0.0
0.5
1.0
1.5
Time (hours)
a
Benzaldehyde
Benzyl alcohol
)K
2.0
Figure 2 - 6. The RARE strain enables the synthesis of L-phenylacetylcarbinol (L-PAC). (A) Scheme
depicting the synthesis of L-PAC from the condensation of exogenously supplied benzaldehyde
and metabolized pyruvate, catalyzed by either PDC or PDCE473Q. (B) Concentration profiles of
PAC, benzaldehyde, and benzaldehyde oxidation/reduction products 24 hours after addition of 5
mM benzaldehyde to RARE and wild-type host strains. (C) Time course of benzaldehyde reduction
using the wild-type host transformed with pRSF/PDCE473Q and pACYC/Car/Sfp.
56
2.3.5. The deletion of all targeted genes to form the RARE strain has no effect on growth
To test whether the full set of gene deletions affects cell growth, the RARE and wild-type
strains were grown in LB medium + 1.2% glucose and in M9 minimal medium + 1.2% glucose. In
the absence of any particular stress, no effect on specific growth rate was observed under either
condition (Fig. 2-7). The average specific growth rates were p = 1.07 h- 1 and p = 0.37 h- 1 in LB and
M9, respectively. Furthermore, final OD600 measurements were taken in the presence of
aldehydes synthesized from experiments featuring the RARE and wild-type strains (Fig. 2-8).
These measurements demonstrate that the presence of vanillin at concentrations explored in
this study does not significantly affect the growth rate of the RARE strain relative to the wild-type
strain. These results reveal the non-essential nature of the complete set of targeted genes and
increase the industrial relevance of the engineered strain.
57
4.0-
-- WT
RARE
35.
3.0-
2.5
C
0
2.0-
1.0-
0.50.0
0
1
2
3
4
5
6
7
8
9
Time (hours)
B
.RARE
-
1.5
0.5
0.0
0
3
6
9
12
15
18
Time (hours)
C
-
p(h*
MG1655(DE3)
RARE
LB+1.2%Glu
1.07
1.07
M9+1.2%Glu
0.37
0.37
+
Figure 2 - 7. OD6 00 measurements of MG1655 and RARE strains when grown in (A) LB medium
1.2% glucose or (B) M9 minimal media + 1.2% glucose. (C) Specific growth rates of each strain
calculated from experiments in the different media listed above.
58
A
2.5.
2.0
2.0-
-
Ii
,,B
.51.5
C
0
D
i
1.0
0.0
1
MG
endA- r
1.0
0.01
RARE
tc fn e adA- reCAp endA- recAp endA- recA-
V
I
MG
RARE
MG
endA- recA-
RARE
endA- reca-
1.5
35n
.4 s s2.51.0
P
0.5-
0.5
01
0.0
MG
en~dA- recA-
RARE
endA- recA-
Figure 2 S. OD 600 measurements at final time points from reported experiments: (A)
benzaldlehyde from benzoate; (B) vanillin from vanillate; (C) vanillin from LB + 1.2% glucose; (D)
vanillin from M9 + 1.2% glucose.
2.4. Discussion
AKRs are found in organisms ranging from vertebrates to archaebacteria (115) and are
believed to be responsible for catalyzing the conversion of methylglyoxal and related reactive
metabolites into less toxic compounds (108). Microbial AKRs belong to ten families: AKR2, AKR3,
AKR5, and AKR8-14 (97). Because AKRs consume the co-factor NADPH to reduce aldehydes, they
are expected to be physiologically relevant under aerobic conditions. A total of nine open reading
frames that encode AKRs in E. coli have been identified using sequence similarity searches (116).
Unlike many other E. coli AKRs, DkgA has been characterized and a crystal structure has been
59
resolved (117). Activity assays performed in vitro constitute the majority of published data on E.
coli AKRs and demonstrate that most of these gene products act on overlapping sets of substrates
(108). Perhaps because of the known redundancy of AKR activity in E. coli, no one has previously
reported the intentional deletion of AKRs for the purpose of building up aldehyde pools. To our
knowledge, there are no previous publications even describing the simultaneous deletion of two
or more AKRs in E. coli. In our case, although we initially focused on AKRs as rational targets for
deletion, we found that two out of the three targeted AKRs were not contributing to
benzaldehyde or vanillin reduction under the conditions tested. Nevertheless, we also found that
all AKRs were active on these substrates when overexpressed, and in the case of vanillin, they
were significantly more active than targeted ADHs. From our experiments, it is still unclear what
conditions, if any, may lead to significant expression of dkgB and yeaE, but there was no
reduction in growth rate or other disadvantage incurred by their deletion.
Previously, as briefly highlighted in Chapter 1, yields of some aldehydes have been
increased by deleting ADH genes responsible for reductase activity in model organisms such as
S. cerevisiae and E. coli. ADHs are generally classified under the short-chain (118-120) or mediumchain (121, 122) dehydrogenase/reductase families and, like AKRs, are known to have broad and
redundant substrate specificity. Specifically, the ADH6 gene in S. cerevisiae was deleted in
another study in order to produce 45 mg/L of vanillin from glucose (41). However, other genes
known to convert aldehydes to alcohols in S. cerevisiae remained intact in that study and no
further deletions have been reported to our knowledge. Another report described an effort to
improve the production of isobutyraldehyde, which is a bulk chemical feedstock (53). In that
study, eight genes were deleted (yqhD, adhP, eutG, yiaY, yjgB, betA, fucO, eutE). Unlike in our
60
study, only five of these targeted genes (yqhD, adhP, eutG, yiaY, yjgB) were found to be capable
of reducing isobutyraldehyde to isobutanol when individually overexpressed. Overall, the
combination of all eight deletions resulted in an improvement in the ratio of isobutyraldehyde to
isobutanol produced from 0.14 gaidehyde/L/OD600 and 1.5 galcohol/L/OD600 to 1.5 galdehyde/L/OD600
and 0.4 galcohol/L/OD600 (53). However, the deletion of the yqhD gene alone led to a 1:1 ratio of
aldehyde to alcohol, indicating a diminishing marginal return on the deletions of the four other
genes that may contribute to isobutryaldehyde reduction. Neither of the aforementioned studies
described an attempt to delete genes encoding AKRs nor did they culture their strains under
aerobic conditions.
By rationally evaluating three AKR deletions in combination with three ADH deletions, we
constructed an E. coli host strain that displayed a significant step-change in the ability to
accumulate aromatic aldehydes relative to several engineered strains containing subsets of these
deletions. The brute-force method of rationally targeting and combining several knockouts to
overcome a redundant endogenous activity should become more accessible given significant
advances in genome engineering and an increasing rate of functional gene annotation.
Techniques that enable prompt construction of rational combinations of gene knockouts, such
as Multiplex Automated Genome Engineering (MAGE) (123), or gene expression knockdowns,
such as RNA or CRISPR interference (124), can be harnessed in future studies like this.
However, selection of the correct target genes to eliminate a highly redundant activity is
not straightforward. Kinetic data from purified enzyme assays may be misleading given that
enzyme activity and gene expression have no correlation but together affect endogenous activity.
Similarly, we have shown that gene products displaying undesired activity in vivo when
61
overexpressed may not need to be deleted in order to eliminate the undesired endogenous
activity. A further complication is that not all of the necessary target genes may be fully
characterized. Current genome modeling approaches for determining knockouts such as
OptStrain would not predict these deletions. In fact, OptStrain has already been used in an
attempt to optimize vanillin production in E. coli and did not include any of the aldehyde
reductases targeted in our study (125). The alternative of transcriptional profiling is limited by
the inability to account for redundant or constitutively expressed genes, as well as by off-target
effects and secondary responses. Finally, combinatorial approaches to gene knockouts (126, 127)
have led to the generation of numerous strain improvements but are unlikely to surmount
sufficiently redundant activities, especially in the absence of an effective selection strategy.
Although the creation of an E. coli strain that has minimal conversion of vanillin to vanillyl
alcohol has been previously sought after (22, 24, 40), some groups have reported alternative
strategies to use E. coli for vanillin production. At least one report documents the use of resting
E. coli cells for the bioconversion of ferulic acid into vanillin with minimal vanillyl alcohol
formation under these conditions (128). Under non-growing conditions, it is likely that the AKR
and ADH genes identified in this study are not expressed to a high degree. However, the use of
resting E. coli cells is not a viable strategy for producing vanillin from glucose, which is an order
of magnitude more affordable as a starting substrate than ferulic acid ($0.3/kg versus $5/kg) (41).
Additionally, microbial cells in a non-growing state lose desired activity more rapidly than cells in
a growing state, frequently resulting in lower overall productivity. Another group has published
the use of growing E. coli cells for the bioconversion of ferulic acid into vanillin, asserting that E.
coli has no degradation pathway from vanillin to vanillyl alcohol (129-131). It may be that no
62
attempt to detect vanillyl alcohol was made in these particular studies. As illustrated throughout
this report, even when performing simple bioconversions using vanillate as a substrate, we
observe significant vanillyl alcohol formation. Whether for the purpose of bioconversions or
utilization of glucose as a sole carbon source, the engineered strain reported in this study is
expected to be a superior E. coli host strain for producing aromatic aldehydes.
As envisioned in Chapter 1, the ability of the RARE strain to accumulate aromatic
aldehydes has broad implications for the biosynthesis of products derived from aldehyde
intermediates. The demonstration of L-PAC synthesis illustrates that aldehydes can be
biologically converted into new chemical classes if they can avoid the fate of rapid reduction. This
previously unattainable option may be more attractive than harvesting resting cells or using
purified enzymes for many biocatalytic processes. Furthermore, it opens up the potential for
synthesis of end products directly from glucose as a sole carbon source.
63
Chapter 3: Decreasing endogenous reduction of aliphatic aldehydes and the
effect on biosynthesis of gasoline-range n-alkanes
A portion of the introduction to this chapter is adapted from thefollowing manuscript: Sheppard
et a/ (2014), Nat. Commun., 5, 5031. The remaining portion of this chapter is adapted from a
manuscript in preparation in which Dr. Micah Sheppard contributed equally as an author in every
aspect. Much of the latter manuscript was incorporated into the Ph.D. thesis of Dr. Sheppard
(2014, http://hdl.handle.net/721.1/91064).
3.1. Introduction
In Chapter 2, accumulation of benzaldehyde and vanillin was demonstrated using the
engineered E. coli RARE strain, which was initially designed with the objective of minimizing
endogenous reduction of benzaldehyde. The ability of the RARE strain to accumulate vanillin
raised an intriguing question: would this engineered strain also display decreased endogenous
reduction of structurally unrelated aldehydes, such as aliphatic aldehydes? If so, then we would
have discovered that reduction of a broad range of aldehydes is governed by a small number of
genes, and the potential utility of this strain for metabolic engineering and biocatalysis would
increase. If not, then perhaps a wider or separate set of endogenous genes would be responsible
for the conversion of other aldehydes into their corresponding alcohols in E. coli.
As results presented in Chapter 2 were being obtained, a related question was emerging
from the doctoral work of Micah Sheppard, who was another student in the Prather Lab. Could
an enzyme be identified that would selectively catalyze the conversion of medium-chain
branched aliphatic carboxylic acids into their corresponding aldehydes? To determine whether
CarNi had this desired attribute, we purified the enzyme and assayed for relative activity in vitro
64
on 13 straight and branched aliphatic acid substrates from C2-C8. CarNi displayed activity on most
substrates, peaking around a primary chain-length of 5-6 carbons (Fig. 3-1). Combined with
previous work in the Prather Lab, the promiscuity of CarN supported the hypothesis that we could
construct complete pathways from glucose to several aliphatic aldehydes of varying chain length.
This provided added motivation to investigate whether we could decrease endogenous reduction
of aliphatic aldehydes upon potential de novo biosynthesis.
1.00.90.80.7-
E 0.60
c 0.5
0
Z: 0.4t0.3-
? 0.2('~0.1
-
0.0
W
00
Figure 3 - 1. Relative activity of the carboxylic acid reductase from Nocardia iowensis (CarNi) on
straight and branched aliphatic acids ranging in carbon chain length from C2 to C8.
Beyond simply an intellectual interest in aliphatic aldehydes, the most compelling
motivation for the work presented in this chapter was the societal relevance of alkanes, the
alternative biofuels that we sought to make from aliphatic aldehydes instead of alcohols (Figure
65
3-2). The United States relies heavily on gasoline to fulfill transportation needs. The U.S.
consumes roughly 4 billion barrels of gasoline annually, which amounts to 40% of total annual
domestic petroleum usage and 47% of all gasoline produced worldwide (132, 133). Vast gasoline
infrastructure exists to facilitate its usage, and prevailing renewable liquid fuel alternatives have
limited compatibility. Ethanol requires a different distribution system than gasoline because of
its hygroscopicity, corrosivity, and biodegradability (134). Due to dissimilar fuel performance
characteristics such as energy density and research octane number (RON) (Table 3-1), renewable
gasoline alternatives are blended with gasoline for use in conventional automobile engines. One
approach towards addressing the compatibility of renewable fuels is to metabolically engineer a
microbe that converts sugars into a product that mimics the composition of gasoline. An added
benefit of such a process would be streamlined product separation from the aqueous phase due
to increased product hydrophobicity (Table 3-1).
R
O
OH
NAD(P)H
HO
yEndogenous
Aldehyde
Reductases
Carboxylic Acid
0
HO\\H*Reductase
OH
glucose
OH
fatty alcohol
NAD(P)+
R
OHR
free fatty acid
NAD(P)H
NAD(P)+
ATP
AMP+PPi
fatty aldehyde
Heterologous
AD
C02
alkane
Figure 3 - 2. Biochemical pathway illustration depicting fatty aldehydes as a precursor to either
an alkane or an alcohol.
66
Table 3 - 1. Performance and separation metrics for select gasoline alternatives and constituents.
Compound
Enthalpy of
Combustion
(kJ/mol)
Ref.
Research
Octane
Number
(RON)
Ethanol
-1370
(135)
109
Butanol
-2670
(135)
Propane
-2220
Butane
-2880
Pentane
Ref.
(136)
Solubility at
Room
Temperature
(mass %)
fully miscible
Ref.
-
Henry's Law
Constant kH
(kPa m 3 mol-1)
-
96
(137)
10.4
(138)
<0.001
(139)
(135)
-
-
6.7E-03
(140)
71.6
(141)
(135)
113
(142)
7.2E-03
(140)
95.9
(141)
-3510
(135)
62
(142)
4.1E-03
(143)
128
(144)
Heptane
-4820
(135)
0
(142)
3.OE-04
(143)
230
(141)
Nonane
-6120
(135)
-17
(142)
1.7E-05
(145)
333
(144)
-
Ref.
To engineer a proof of concept pathway in Escherichia coli, we first looked at published
chemical composition studies of gasoline. Typical regular unleaded gasoline is a blend of over 30
aliphatic and aromatic hydrocarbons (Fig. 3-3A) (146-148). Isopentane, pentane, and butane are
three alkanes among the most common components in gasoline. As briefly highlighted in Chapter
1, these and other alkanes can be obtained from aliphatic aldehydes using aldehyde
decarbonylases. A key question this chapter seeks to investigate is the effect of decreased
endogenous aldehyde reduction on alkane titers.
67
.....
.......
A
0a1% Pr"Opae
3.7% .-Sutane
9.3%
Cycioalkef"
6.5% A
Isopene
7.2% A-meityn
Length
42% 2-Methylpentane
2.7% 34Aettrytpentane4.0%
n4l4exane"
1 A% 2.Methyihexane"
IS5% 3-Metylhaxana
2.0% aHeptmew
0.7% n-#Onane
11.0% Other Akatnes
B
GLYCOLYSIS
4
0,
0- 0.........
2A~)~L
2%
0
b Ityrat
Popano
valerag
n-Sutane
hexanosa.
n.Pentane
0
cctoe
fl.H*Am
0
~
decanoate
n44oane
Figure 3 - 3. (A) Composition of typical regular unleaded gasoline displayed in weight percent (wt.
%) based on the average of Refs. 146 and 147. Single asterisk indicates that compounds below
0.5 wt. % are not reported in Ref. 147. Double asterisks indicate that wt. % includes contribution
from trace compounds in Ref. 146. (B) Modular pathway design used for selective synthesis of
key gasoline-range alkanes in engineered E. coli. Genes in gray within Modules 1-Pr and 1-Ma are
native and were not overexpressed, whereas genes in black were overexpressed. Module names
are abbreviations for the following: "Pr" = Propionate; "Ma" = Malonyl-ACP; "BC" = Butyrl-CoA;
"MCC" = Medium-Chain-CoA; "Oc" = Octanoate; "SA" = Short Alkanes; "LA" = Long Alkanes.
68
In this chapter, we report selective production of propane, butane, pentane, heptane,
and nonane using a modular pathway approach in metabolically engineered E. coli (Fig. 3-3B).
Modules represent convenient gene groupings that collectively convert one easily detectable
metabolite into another. Module 1 variants enable generation and activation of precursors for
carbon chain extension using either Acyl Carrier Protein (ACP) or Coenzyme A (CoA). Module 2
variants perform carbon chain extension either by FAS or by Reverse -Oxidation (RBO). Module
3 variants result in termination of chain extension and generation of free fatty acids (FFAs)
ranging from C4 to C10. Finally, Module 4 variants convert Cn fatty acids into corresponding C(n1) alkanes via a Cn fatty aldehyde intermediate.
3.2. Materials and Methods
3.2.1. Strains and Modules
E. coli strains and modules used in this study are listed in Table 3-2. Molecular biology
techniques were performed according to standard practices (98) unless otherwise stated.
Molecular cloning and new vector propagation were performed in DH5a. Previously constructed
E. coli K-12 MG1655(DE3)AendAArecA and the reduced aromatic aldehyde reduction ("RARE")
AendAArecA strains (103, 149) were used as hosts for experiments testing biosynthesis of C3-C5
alkanes. ThefadD gene was deleted in MG1655(DE3) and RARE strains using a donor strain from
the Keio collection (99) and the method of P1 transduction (100). The vector pCP20 was used to
cure the kanamycin resistance cassette (101). MG1655(DE3)AfadD and RARE1fadD strains were
used as hosts for experiments testing biosynthesis of C7 and C9 alkanes. Oligonucleotides (Sigma,
The Woodlands, TX) used as PCR primers are shown in Table 3-3. Q5 High Fidelity DNA
69
Polymerase (New England Biolabs, Beverly, MA) was used for DNA amplification. A codonoptimized sequence for P. marinus MIT9313 AD (PMT1231) was purchased as a DNA String
(GeneArt, Regensburg, Germany). Once codon-optimized PMT1231 was cloned into pACYC-carsfp-PMT1231, the AD_A134Fpm mutant sequence was generated using Polymerase Incomplete
Primer Extension (PIPE) cloning (107). The new sequence replaced the GCA codon at positions
400-402 with a TTT codon. Codon-optimized sequences for the N. punctiforme wild-type and
mutant ADs were also purchased as DNA Strings. All DNA Strings were digested with Ndel and
AvrIl restriction enzymes and cloned into the second multiple cloning site of the pACYC-car-sfp
plasmid. Construction of pACYC-car-sfp was previously reported (149, 150). Restriction enzymes
and T4 DNA ligase were purchased from New England Biolabs. A codon optimized version of the
full open reading frame of C. hookeriana FATB2 was purchased from GenScript (Piscataway, NJ).
The FATB2 gene and variants were cloned into the pETDuet-1 vector (Novagen, Darmstadt,
Germany) using BamHI and Notl.
70
Table 3 - 2. Strains and modules used in this study.
Strain Name
Genotype
Abbreviation
MG1655(DE3) AendA ArecA
A(DE3 [lacl lacUV5-T7 gene 1 indl sam7 nin5])
AendAArecA
WT
MG1655(DE3) AfadD
A(DE3 [lacl
AfadD
RARE AendA ArecA
A(DE3 [lacl IacUV5-T7 gene 1 indi sam7 nin5])
AdkgB AyeaE AyahK AyjgB AyqhC AdkgA AyqhD
AendA ArecA
RARE
RARE AfadD
A(DE3 [lad lacUV5-T7 gene 1 indi sam7 nin5])
AdkgB AyeaE AyahK AyjgB AyqhC AdkgA AyqhD
AfadD
RAREAfadD
Plasmid Name
Module
pET-FatB2m2Ch
3-Oc
pET-FatB2m2Ch-accABCDE
1-Ma, 3-Oc
pACYC-(carNI-sfpP)-PMT1 231 Pm
4-1-A
pET-(bktBCn-terTd)-(phaBcn-phaJ4bcn)
2-MCC
pACYC-(carNi-sfp)-PMT1231_Ai34Fpem
4-SA
pCOLA-(thrA fr BC)-ilvA
f
lacUV5-T7 gene 1 indi sam7 nin5])
WTAfadD
1-Pr
pET-(thla-terTd)-(phaBn-phaJ4bcn)
2-BC
pACYC-(carm-sfpB)-ADNp
4-LA'
pACYC-(carN-sfp5-ADA122FNP
4-SA'
71
Table 3 - 3. Oligonucleotides used in this study.
Plasmid name
Primer name
Primer sequence
pACYC-(car-sfpip)-PMT1231pm
pACYC-(carwsIps)-PMT1231,.mut
pET-4h
eru
Notes
Codon optimized PMT1231pm
was digested and cloned into
pACYC-(carNsfIpfs) using
Ndel/Avrll
mutADup
mutADdn
TTTGCAATTAGCTTTTATCATACCTATATTCCGG
TAGGTATGATAAAAGCTAATTGCAAATGCTTC
PMT1231i, was subcloned into
pCDFDuet-i and the plasmid was
PCR amplified with the mutAD
primer set. Once sequenced, the
mutant version was suboloned back
into pACYC-(carN-sfp.,)
terTd(thl)up
AAAAAACATATGATTGTGAAACCGATGG
AAAAAACCTAGGTCAAATACGGTCAAAGCG
The terTd gene was amplified from
pET-(bktBcerTd)-(phaBc,-phaJ4bfl)
and cloned with Ndel/Avrl
The thic, gene was subcloned from
an existing pRSF-thC6 plasmid with
Ncol/EcoRI
terTd(thl)_dn
pET-FstB2,,
Subcloned from pUC57
plasmid using Ndel/Pac
pET-Hi-FstB2C
fatB2hisjup
fatB2hisdn
AAAAGGATCCGATGGTGGCTGCAGCC
AAAAGCGGCCGCTTAGGAAACACTGTTGCCATTG
BamHI/Notl
pET-Fat92mc1
fatB2mlhisup
fatB2mhis_dn
AAAAGGATCCACTGCCGGATTGGTC
AAAAGCGGCCGCTTAGGAAACACTGTTGCC
BamHVNotl
pET-Fats2m2a,
fatB2m2his-up
AAAAGGATCCACTGGTTGA1TCCTTTGG
Downstream primer same
as for FatB2mic
BamHl/Notl
pACYC-accABCD~c
accA-up
accA-dn
accBC_up
accBCdn
accD_up,
accDdn
ATATACATATGAGTCTGAATTTCCTTGATTT
GAGTGGGTTCCGTACTTACGCGTAACCGTAGCTC
GTACGGAACCCACTCATGGATATTCGTAAGATTAAAAAACTG
TAGGGACCTTTCTGTCTTATTTrTCCTGAAGACCGAGT
GACAGAAAGGTCCCTAATGAGCTGGATTGAACGAA
ATATAGACGTCTCAGGCCTCAGGTTCC
pET-FatB2m2ch-acmABCD~c
Assembled using Splicing
by Overlap Extension (SOE)
PCR, Ndel/Aatil
accABCDr, operon
subcloned into
pET-FatB2m2ch
Ndel/Aatl
pACYC-(car.sfp,,)-ADP
Ndel/Avrll
pACYC-(CarMsp,).ADA122F
Ndel/Avrll
3.2.2. Chemicals
The following compounds were purchased from Sigma: pentane, heptane, nonane,
butyraldehyde, butanol, sodium hexanoate, hexanal, hexanol, sodium octanoate, octanoic acid,
octanal, and decanoic acid. Isopropyl -D-1-thiogalactopyranoside (IPTG) was purchased from
Denville Scientific (South Plainfield, NJ). Ampicillin sodium salt, chloramphenicol, and kanamycin
sulfate were purchased from Affymetrix (Santa Clara, CA).
72
3.2.3. Culture Conditions
For all production experiments, 3 mL LB overnight seed cultures in 14 mL round-bottom
tubes were used as inocula. All 2 mL production cultures were inoculated with overnight culture
at 1% by volume and grown directly in 10 mL GC vials with PTFE Silica Septa screw caps (Supelco,
Bellefonte, PA, Cat.#: SU860099 and SU860103). The production medium was LB with either 1.2%
(w/v) glucose or 1.2% (v/v) glycerol. Depending on modules used, culture medium was
supplemented with 50 mg/L ampicillin (for Modules 3-Oc, 1-Ma, 2-MCC, and 2-BC), 17 mg/L
chloramphenicol (for all Module 4 variants), and/or 25 mg/L kanamycin (for Module 1-Pr).
Culture vials were placed in tube racks at 45* angles and incubated with agitation at 30*C and
250 rpm. Cultures were induced with 0.5 mM IPTG (final concentration) at OD 60 0 values between
0.7 and 0.9. Cultures were incubated for 22-26 hours after induction prior to metabolite analysis.
For aldehyde accumulation experiments, 2 mL cultures were similarly inoculated and grown in LB
medium supplemented with 34 mg/L chloramphenicol in GC vials. At induction, culture medium
was supplemented with either 721 mg/L octanoate or 580 mg/L hexanoate (both equivalent to 5
mM). Cultures were incubated for 22-26 hours after induction prior to metabolite analysis.
3.2.4. Metabolite Analysis
Liquid chromatography: Culture samples were pelleted by centrifugation and supernatant
was removed for HPLC analysis with an Agilent 1200 series instrument with a refractive index
detector. Analytes were separated using the Aminex HPX-87H anion exchange column (Bio-Rad
Laboratories, Hercules, CA) with a 5 mM sulfuric acid mobile phase at 35*C and a flowrate of 0.6
mL/min.
73
Gas chromatography: At 24 hours post induction, vial cultures were placed in a 42*C
incubator for 20 minutes in order to drive the volatile alkanes into the gas phase. Gas headspace
samples were then taken using a 10 mL gas-tight syringe (SGE Ringwood, Victoria, Australia). In
order to mix the gas sample and prevent formation of a vacuum, one syringe was used to inject
a 1 mL volume of air as 1 mL of sample was drawn. The process was repeated until a 9 mL volume
was taken from the vial and an additional 9 mL of air was injected into the vial. The concentration
injected into the GC was thus diluted 2-fold. A Shimadzu GC (GC-2014) with a RT-Q bond column
(30 m length, 530 im ID, 20 pm film thickness) and flame ionization detector (FID) was used for
the analysis. A 5 ptL sample loop was used for sample injection. The method oven conditions were
as follows: a 40*C hold for 1 minute followed by a 25*C/min ramp up to 280*C with a 5 minute
final hold. Quantification of propane, butane, and pentane was based on a one point calibration
using a standard gas mixture purchased from AIRGAS. A separate standard curve for heptane was
created by adding known volumes to a 1.127 L glass bottle fitted with a septum cap. A heptane
standard and the 1.127 L bottle were chilled to 40C and different known volumes of heptane were
added. The bottle was then warmed to room temperature allowing the heptane to fully vaporize.
A gas tight syringe was used to inject 8 ml of gas from the bottle into the GC. A similar curve was
generated for nonane.
3.3. Results
3.3.1. Modules 1-Ma, 3-Oc, and 4-LA Result in Heptane and Nonane Biosynthesis
We began by investigating whether native FAS could be used to produce alkanes shorter
than nonane using a selective acyl-ACP thioesterase. Microbial production of heptane from
74
glucose was realized by developing and using Modules 1-Ma, 3-Oc, and 4-LA (Fig. 3-3B). A
truncated version of the FatB2ch from Cuphea hookeriana (FatB2m2ch) was used as Module 3-Oc
to achieve synthesis of octanoate and decanoate (Fig. 3-4A). Combined overexpression of
Module 1-Ma, which consists of the E. coli acetyl-CoA carboxylase complex, and Module 3-Oc led
to increased titers of both FFAs (72
9 mg/L octanoate and 3.2
0.2 mg/L decanoate) (Fig. 3-
4A).
After obtaining selective carbon chain termination, we next focused on conversion of
octanoate to heptane (Module 4-LA) (Fig. 3-3B). We hypothesized that the RARE host strain may
be able to accumulate octanal given that several genes deleted in the RARE strain encode
enzymes capable of reducing both aromatic and aliphatic aldehydes (112). We grew wild-type
(WT) and RARE host strains expressing CarN in medium supplemented with octanoate directly in
gas chromatography (GC) vials and then measured octanal concentrations in the headspace after
24 hours (Fig. 3-4B). When RARE was used, 37
12 mg/L of octanal was detected in the
headspace, whereas no octanal was detected in the headspace using the WT host. This suggested
that we would be elevating the substrate pool for ADpm when we would express the complete
Module 4 (Fig. 3-4C).
75
A
100-
B
octanoate
so.
Decanoate
50-
80EJ
C
0.6-
Octanal
Car,,
0.5AC
Ack
AyeaE
30.
octanal
Heptane
Nonane
octanoate
I
60-
D
0
-J
0.4-
0.3-
AYqAD
I-.
M..20-
P~
-C
40420-
'
Module
Modules
3-Oc
1-Ma &3-Oc
0.
1=
AD,,
10-
octanol
heptane
WT
RARE
0.20.1
-
GJ
4-.
0.0WTAfadD
RAREAfadD
Figure 3 - 4. Selective production of heptane and nonane using FAS for carbon chain extension. (A) C8 and CIO FFA titers resulting
from Module 3-Oc or Modules 1-Ma and 3-Oc in WTAfadD. (B) Gas-phase titers of octanal observed 24 hours after supplying octanoate
to WT and RARE expressing CarNi. (C) Illustration of octanal as a branch-point metabolite to heptane or octanol. (D) Alkane titers
resulting from Modules 1-Ma, 3-Oc, and 4-LA in WTAfadD and RAREnfadD. Experiments performed in triplicate with averages as
reported values and standard deviation as error bars. All alkane titers are gas-phase.
76
A
B
C
2.0 1
-
500
Hexanoate
140-
Pentane
I
Hexanal
400.
Hexanol
1.6-
120100-
j
300-
1.2-
-
E
200.
I
I
''
60.
0.8-
80
-
im
Butanol
Hexanol
-6
:2
= 100-
40-
0.41
20-
0-
_r 0.0
~1~~
WT + Car
RARE + Car
WT
4-LA
WT
4-SA
RARE
4-LA
RARE
4-SA
0o
WT + 2-MCC +4-SA
RARE + 2-MCC + 4-SA
Figure 3 - 5. Selective production of pentane using RBO for carbon chain extension. (A) Liquid-phase titers of hexanoate and
downstream metabolites observed 24 hours after supplying hexanoate to WT and RARE expressing CarNj. (B) Alkane titers resulting
from Modules 2-MCC and either 4-LA or 4-SA in WT and RARE. (C) Liquid-phase titers of butanol and hexanol in WT and RARE
containing Modules 2-MCC and 4-SA. Experiments performed in triplicate with averages as reported values and standard deviation as
error bars. All alkane titers are gas-phase.
77
We then tested production of heptane from glucose in the WTAfadD and RAREAfadD host
strains containing Modules 1-Ma, 3-Oc, and 4-LA. We sampled culture headspace 24 hours after
induction for formation of possible C3-C9 alkanes (Fig. 3-4D). In both cases, we observed heptane
0.03 mg/L versus 0.32
+
synthesis. RAREnfadD produced more heptane than WTAfadD (0.52
0.05 mg/L, respectively). Given that decanoate was also observed previously (Fig. 3-4A), roughly
0.2 mg/L of nonane was also produced using each strain (Fig. 3-4D). No other alkanes were
detected. Based on low alkane titers relative to FFA titers and activity of CarN on octanoate (150),
the AD-catalyzed reaction was the apparent bottleneck, consistent with other recent studies
featuring aldehyde decarbonylases (151, 152). The low observed AD activity in vivo suppressed
the benefit of increasing aldehyde substrate pool size using RAREnfadD.
3.3.2. Substitution of an RBO Module (Module 2-MCC) Results in Biosynthesis of Pentane as
the Sole Alkane Product
To achieve selective synthesis of more common gasoline-range alkanes, we next
substituted carbon chain extension modules for an RBO module tailored for C6 fatty acids. We
recently reported a modular pathway framework selective for synthesis of 4-methylvalerate,
which is a branched C6 acid (150). In the absence of isobutyryl-CoA generating modules, use of
the extension module described in that report (listed here as 2-MCC) results in the production of
hexanoate directly from glucose. As demonstrated in the synthesis of 4-methylvalerate, the use
of RBO eliminates the need for a termination module due to endogenous thioesterase activity
(150). In addition, engineered RBO has resulted in higher titers of C4-C6 fatty acids than FAS in
previous studies and is a more efficient metabolic route given decreased ATP utilization for
precursor generation (50, 150, 153-156). Although CarN can efficiently convert hexanoate to
78
hexanal (150), we faced two potential challenges with hexanal. The first was whether we could
decrease endogenous conversion of hexanal to hexanol. When we cultured WT and RARE
expressing CarNi in medium supplemented with hexanoate, we observed that only RARE
accumulated hexanal (Fig. 3-5A). The second obstacle was the potential inability of ADs to act on
aldehydes as short as C6. Based on in vitro activity data discussed earlier (59), we examined
whether use of CarNi with AD_A134Fpm (Module 4-SA) or ADpm (Module 4-LA) would lead to higher
pentane titers. We tested alkane production using Module 2-MCC and either Module 4-SA or 4LA in either WT or RARE (Fig. 3-5B). All four strains produced pentane, and the highest titer (1.6
0.3 mg/L) was observed for the pairing of RARE with Module 4-SA. No other alkanes were
detected. RARE eliminated butanol and hexanol byproduct formation in the liquid-phase,
whereas WT produced 31
6 mg/L butanol and 115
16 mg/L hexanol (Fig. 3-5C). Despite the
presence of butyraldehyde and the possibility of AD-catalyzed conversion to propane, the
absence of propane in these cultures likely results from the preference of both Module 4
enzymes (Car and AD) for C6 over C4 substrates. The absence of longer alkanes stems from the
highly selective carbon-chain extension module (Module 2-MCC).
3.3.3. Modules 1-Pr, 2-MCC, and 4-SA Enable Butane Biosynthesis
To investigate whether alkane chain lengths could be further shortened, we added a
module previously used for pentanol synthesis (listed here as Module 1-Pr) (Fig. 3-3B) (153). We
hypothesized that pentanal, the immediate precursor to pentanol, could be converted to butane
instead. In this case, we chose to use glycerol rather than glucose as a carbon source given that
previously reported pentanol titers were 6-fold higher using glycerol (153). When we combined
Module 1-Pr with Modules 2-MCC and 4-SA, we observed synthesis of butane and pentane from
79
0.12 mg/L and 0.46
glycerol (Fig. 3-6). Butane titers were similar in WT and RARE (0.35
0.15
0.08 mg/L pentane using
mg/L, respectively). Pentane titers exceeded butane, with up to 1.27
RARE. Pentane production was expected given the preference of Module 4 enzymes for longer
substrates and the ability of Module 2-MCC to generate butyryl-CoA and hexanoyl-CoA. A route
to propionyl-CoA that does not rely on the PDH complex may offer greater selectivity for butane.
2.0
1.81.6-
1.4
E
Propane
Butane
Pentane
Hexane
Heptane
1.2
1.0
P~
0.8-
0.6-
OA
0.20.0
il ml
WT-C3
Gkwcose
WI
RARE-C3
I
WVT-C4
I
I
RARE-C4
I
I
RARE-C5
WiT-C5
i
++
2-MCC
-+
4-%
+
-
+
+
1-Pr
-
Glycerol
Figure 3 - 6. Alternative modules enable synthesis of butane and propane.
3.3.4. Modules 2-BC and 4-SA Enable Propane Biosynthesis
Although propane is a minor constituent of gasoline, we were curious about the lower
bound of our selective platform and next tested for propane synthesis by attempting to limit
80
carbon chain extension beyond C4. We created Module 2-BC by replacing BktBcn from
Cupriavidus necator in Module 2-MCC with Thlca from Clostridium acetobutylicum based on
increased specificity of ThIca for condensation of two acetyl-CoA (157). WT and RARE, each
harboring Modules 2-BC and 4-SA, produced propane from glucose (0.17
0.04 mg/L and 0.13
0.02 mg/L, respectively) (Fig. 3-6). Although propane titers were similar, the two strains displayed
contrasting intermediate and byproduct profiles (Fig. 3-7). Surprisingly, RARE produced more
pentane (0.41 0.09 mg/L) than propane, whereas propane titers exceeded pentane titers in WT.
A
B
24-
-100
G)
tm
20
_
$
16
;
C
12
u0
__
41M
Butyraldehyde
Butanol
80
20
8-
0
-~"0.
~C
4-*
20-
0
0
WT
RARE
WT
RARE
Figure 3 - 7. Intermediate and byproduct profiles associated with propane synthesis. (A) Relative
butyraldehyde concentrations in the headspace of cultures containing Modules 2-BC and 4-SA.
An increased concentration of butyraldehyde was observed in the gas phase using RARE. (B)
Liquid-phase concentrations of butyraldehyde and butanol in cultures containing Modules 2-BC
and 4-SA. Increased levels of butyraldehyde and decreased levels of butanol were observed in
the liquid phase using RARE.
3.4. Discussion
Initial efforts to produce microbial fuels as gasoline alternatives focused on production of
ethanol, which remains the dominant biofuel (158). In recent decades, biotechnological advances
81
enabled the design and construction of de novo biosynthetic pathways, several of which have
resulted in the production of next-generation or advanced biofuels that more closely
approximate properties of gasoline (159). Unlike many previous attempts to produce biofuels
with similar properties as that of gasoline, this study aimed to produce a representative set of
gasoline constituents, thereby naturally conferring properties characteristic of gasoline. Although
only the production of low titers of straight chain n-alkanes have been demonstrated here, this
platform could be engineered to incorporate branched precursor-generating modules and
further optimized to achieve higher titers. In particular, we provide a foundation for the eventual
customization of alkane content for desired properties in diverse fuel applications. Production of
any one of these alkanes alone would likely be insufficient for most transportation uses given the
diversity of gasoline constituents.
Overall, we have demonstrated that E. coli can be metabolically engineered to selectively
produce key gasoline-range alkanes (propane, butane, pentane, heptane, and nonane) from
simple and renewable carbon sources. We achieved this in a modular fashion such that product
distribution can be easily tuned by substituting or adding independent pathway modules. Along
the way, we have demonstrated the use of a thioesterase from C. hookeriana in E. coli for
selective termination of fatty acid synthesis at C8 and have utilized this enzyme for selective
production of heptane. We also showed that extension by RBO, which is more ATP-efficient,
could be substituted for extension by FAS. Using a common RBO extension module, we were able
to achieve production of pentane, blends of pentane/butane, and blends of pentane/propane.
We established that use of the A134F mutant of ADpm resulted in higher titers of the model
gasoline-range alkane pentane than the wild-type ADpm. Finally, we demonstrated the ability to
82
decrease the endogenous reduction of aliphatic aldehydes, thus lowering alcohol byproduct
formation. If aldehyde decarbonylase variants displaying greater catalytic efficiency are found or
engineered, then we expect the decreased competition with endogenous reduction to improve
alkane titers. Biosynthesis of gasoline-range alkanes represents a step forward towards the
bridging of biofuels research with existing assets of the petroleum and automotive industries.
83
Chapter 4: Enhancing in vitro aldehyde biosynthesis by pairing carboxylic acid
reductase with inorganic pyrophosphatase
4.1. Introduction
Chapters 2 and 3 revealed how a variety of aldehydes could be produced in small
quantities using engineered E. coli. In all of the cases of aldehyde biosynthesis, a recombinant
carboxylic acid reductase from Nocardia lowensis was expressed in E. coli in order to catalyze the
conversion of carboxylic acids into aldehydes. In this chapter, we take a closer look at this
reaction by purifying the enzyme and observing the kinetics in vitro. By performing aldehyde
biosynthesis in a noncellular environment, we are also studying an alternative to microbial
aldehyde synthesis that circumvents the problem of aldehyde toxicity described in Chapter 1.
As discussed previously in this thesis, many aldehydes find uses in the flavors and
fragrances industries. For the flavor industry in particular, it is important that aldehyde flavoring
agents be produced without the use of harsh chemicals or severe processing conditions. It is
often the case, however, that the desired aldehyde obtained naturally from plant extracts is
either expensive, scarce, or both. Given that, there are two intriguing opportunities for
biotechnological processes to advance aldehyde production from more abundant natural
precursors: (1) microbial conversion processes, or (2) cell-free enzymatic conversion processes.
Advantages of using microbial conversion include the ability to use more inexpensive inputs and
carbon sources, as well as the ability to achieve economies of scale. Until recently, however,
aldehydes have not been stable in microbial cultures due to rapid reduction by the microbes of
the aldehydes into their corresponding alcohols. Although this problem has essentially been
84
solved in laboratory scale cultures of E. coli (149, 151, 160), aldehyde toxicity remains a problem
that is sufficiently potent for this class of molecules to warrant simultaneous investigation of
alternative biosynthetic routes. Decades before the emergence of advances in our ability to
engineer microbes, industry made frequent use of homogenates or purified enzymes to enable
gentler aldehyde biosynthetic processes with limited byproduct formation (160). Such cell-free
or in vitro biosynthetic routes have the advantage of entirely circumventing the issue of aldehyde
toxicity, with the minor exception of aldehyde adducts that may form on, and inhibit, aldehyde
biosynthetic enzymes. An additional benefit of an in vitro route is that the dilute solution and
high purity streamlines recovery of the desired product. However, use of aldehyde biosynthetic
enzymes in vitro will introduce additional costs in the form of enzyme purification and any cofactors required for the reaction.
In recent years, carboxylic acid reductases have shown outstanding promise for their use
in aldehyde biosynthesis in vitro or in vivo (160). The carboxylic acid reductase (Car) from
Nocardia iowensis is a model enzyme that has undergone detailed in vitro characterization and
has been expressed in E. coli to produce diverse aldehydes ranging from aromatic to aliphatic
(21-24, 149, 150). Many of the carboxylic acid precursors to desired aldehydes can be found more
inexpensively in nature than their corresponding aldehyde. In addition, although the expensive
co-factors ATP and NADPH are required, there are known enzymatic regeneration schemes for
both co-factors (24, 161-167) that may make an in vitro reaction scheme practical depending on
the value of the aldehyde and the quantity desired.
A major obstacle, however, is that in vitro reactions catalyzed by this enzyme or by
homologs (25) have often been limited to low conversions of the carboxylic acid substrates (22).
85
After initial studies expressing Car in E. coli, it was discovered that the recombinant enzyme
contained
lower
than
wild-type
activity
because
Car
requires
activation
by
a
phosphopantetheinyl transferase (PPTase) and the endogenous PPTase activity for Car in E. coli
was lower than in its native host (23). However, as soon as we began studying activated Car in
vitro, we observed that the rate of the reaction decreased dramatically within an hour at 30*C or
200 C. We also observed gradual formation of precipitate in spectroscopic reaction cuvettes.
In this chapter, we investigate why Car is subject to apparently limited turnover in vitro.
Ultimately, we identify that a known byproduct of the reaction, pyrophosphate (PP1 i), is inhibitory.
Addition of inorganic pyrophosphatase (Ppa) solves the problem of limited turnover, improving
conversion by as much as 2-fold under conditions tested and improving the accuracy of modeled
in vitro pathway kinetics. Based on titration of the molar ratios of Ppa to Car, a fusion of Car-Ppa
would be worthwhile for preparative-scale in vitro reactions.
4.2. Materials and Methods
4.2.1. Plasmid construction
Escherichia coli DH1OB (Invitrogen, Carlsbad, CA) was used for plasmid cloning
transformations and plasmid propagation. In order to potentially study the effect of enzyme colocalization in a subsequent study, C-terminal peptide tags corresponding to synthetic protein
scaffold domains appended to flexible glycine-serine linkers were added to Car and to YtbE prior
to expression and purification. As previously described, the codon-optimized gene encoding Car
was first cloned to generate the pET/His-Car-RBS2-Sfp vector (149, 150). Next, the gene encoding
Car was amplified by PCR using the two sets of oligonucleotides shown in Table 4-1 in order to
86
add a sequence encoding the GBD domain cognate peptide to the open reading frame (168). This
amplicon was then cloned into the same site as the original Car using the restriction enzymes
BamHI and Notl. Note that the untagged version of Car from Nocardia iowensis, which we
purified and assayed previously (150), has very similar activity and is subject to the same limited
turnover phenomenon. The gene encoding YtbE was amplified from Bacillus subtilis PY79
genomic DNA (gDNA) by PCR and then cloned into the pCDFDuet vector (Novagen, Madison, WI)
using the restriction enzymes BamHl and Sall. Next, the gene encoding YtbE was amplified by PCR
in order to add a sequence encoding the SH3 domain cognate peptide to the open reading frame
(168). This amplicon was then cloned into the same site as the original YtbE using the restriction
enzymes BamHl and Sall. The gene encoding Ppa was amplified from E. coli MG1655 gDNA and
cloned into the pTEV5 vector using the restriction enzymes Ndel and Noti. B. subtilis and E. coli
gDNA were prepared using the Wizard Genomic DNA purification kit (Promega, Madison, WI).
PCR amplification was performed using custom oligonucleotides (Sigma-Genosys, St. Louis, MO)
and Q5 High-Fidelity DNA polymerase (New England Biolabs, Ipswich, MA). Restriction enzymes
were also obtained from New England Biolabs.
87
Table 4 - 1. Oligonucleotides used in this study.
Car-GBD-f
CATCACCATCATCACCAC
Car-GBD-rl
YtbE-f
GTGGATGGCTCTGCTTCTCTTCTGCATCACGTGCATCAGGGCACCCACCAGTCCAGAGCCACTACCGTTGCAG
CAGTTCCA
AAAAAAGCGGCCGCTCAATCTTCATCTTCATCGCCAGCCTGGTCCTCCCCTTCGTCGGAGGAGTGGATGGCTC
TGCT
AAAAAAGGATCCAATGACAACACATTTACAAG
YtbE-r
AAAAAAGTCGACATTAAAAATCAAAGTTGTC
YtbE-SH3-f
CTAATAAGGAGATATACCAT
Car-GBD-r2
YtbE-SH3-r
I I I I I iGTCGACTCACCCCGGACGGCGACGTFGGCGGAAGAGCTGGCGGAGGGCCAGAACCGCTACCGAA
ATCAAAGTTGTCCG
Ppa-f
AAAAAACATATGAGCTTACTCAACGTCCCT
Ppa-r
AAAAAAGCGGCCGCTTATTTATTCTTTGCGCGCT
4.2.2. Chemicals
Commercial inorganic pyrophosphatase (E. coli) was obtained from New England Biolabs.
The following compounds were purchased from Sigma: benzoic acid, benzaldehyde, benzyl
alcohol, vanillic acid, vanillin, magnesium chloride, dithiothreitol (DTT), ATP disodium salt
hydrate, AMP disodium salt, NADPH tetrasodium salt, NADP' sodium salt, and sodium
pyrophosphate tetrabasic. Isopropyl P-D-1-thiogalactopyranoside (IPTG) was purchased from
Denville Scientific. Ampicillin sodium salt and streptomycin sulfate were purchased from
Affymetrix.
4.2.3. Enzyme purification
All proteins in this study were overproduced using Escherichia coli BL21 Star (DE3)
obtained from Invitrogen. All proteins were purified using two-step purification techniques to
ensure high purity. His-Car-GBDtag and His-YtbE-SH3tag (henceforth referred to as Car and YtbE,
respectively) were purified using sequential affinity and anion exchange chromatography.
88
Overnight cultures harboring either pET/His-Car-GBDtag-RBS2-Sfp or pCDF/His-YtbE-SH3tag was
used as 10% (v/v) inoculum in two liters of LB Broth containing either 100 mg/L ampicillin or 50
mg/L streptomycin. Cultures were incubated at 30*C and 250 rpm, and expression was induced
using a final concentration of 1 mM IPTG at an OD6oo of 0.6. Cells were harvested after 20 hours
using centrifugation and resuspended in in Buffer A (100 mM MOPS-NaOH [pH 7.0], 300 mM
NaCl, and 10% glycerol). Cells were subsequently lysed using sonication. The supernatant was
collected, supplemented with imidazole (5 mM) and batch bound at 4*C for 2 h to 1 mL of Ni-NTA
resin (Qiagen, Germantown, MD). The resin was washed with Buffer A containing 7.5 mM
imidazole and subsequently poured into a column. Affinity chromatography was performed using
step-wise increasing concentrations of imidazole (20, 40, 60, 100, and 250 mM). Fractions
containing purified His-tagged enzyme were pooled and dialyzed overnight at 4*C into Buffer B
(100 mM MOPS-NaOH [pH 7.0], 50 mM NaCl, 1 mM DTT, and 10% glycerol).
For subsequent anion exchange chromatography, dialyzed fractions were loaded onto a
5x5 mL HiTrap Q HP anion exchange column (GE Life Sciences, Piscataway, NJ) via a superloop,
which were integrated into an AKTApurifier with a UNICORN control system v5.20 and a Frac-950
collector (GE Life Sciences). The purification was performed at a flow rate of 1 ml/minute at 4 *C.
An initial wash of 25 ml was followed by a linear gradient from 50 mM NaCl to 500 mM NaCl for
100 ml elution volume. Fractions of 2 ml were collected and absorbance at 280 nm was used to
determine desired fractions. Desired fractions were pooled and dialyzed once again in Buffer B
to reduce salt content. Dialyzed enzyme was then flash frozen using liquid nitrogen and stored at
-800 C.
89
The gene encoding Ppa was inserted into the pTEV5 vector for protein purification,
leading to an enzyme product containing an N-terminal hexahistidine (His) tag removable by
treatment with TEV protease. One liter of cells harboring pTEV5/Ppa was grown at 30*C and 250
rpm in LB medium containing 100 mg/liter of ampicillin. Expression was induced using a final
concentration of 1 mM IPTG at an OD600 of 0.6. Cells were harvested after 20 hours using
centrifugation and resuspended in Buffer A. Cells were subsequently lysed using sonication. The
supernatant was collected, supplemented with imidazole (5 mM) and batch bound at 4*C for 2 h
to 1 mL of Ni-NTA resin (Qiagen, Germantown, MD). The resin was washed with Buffer A
containing 7.5 mM imidazole and subsequently poured into a column. Affinity chromatography
was performed using step-wise increasing concentrations of imidazole (20, 40, 60, 100, and 250
mM). Fractions containing purified His-Car were pooled along with 0.5 mg His-tagged TEV
protease and dialyzed overnight at 4*C into Buffer B. The dialyzed and TEV-digested protein was
passed through 1 ml of Ni-NTA resin to remove His-tagged Ppa and TEV protease. The untagged
Ppa that did not bind to the resin was collected, flash frozen using liquid nitrogen, and stored at
-80*C. Qualitative purity of all protein fractions were determined using SDS-PAGE (Bio-Rad,
Hercules, CA). All protein concentrations were determined using the Bradford assay with bovine
serum albumin as a standard (169).
4.2.4. Kinetic studies
Michaelis-Menten parameters for Car on benzoate were determined by measuring
changes in absorbance at 340 nm for up to 5 minutes. Reactions were prepared as follows: 100
mM MOPS-NaOH [pH 7.0], 10 mM MgCl2, 0.6 mM NADPH, 1 mM ATP, 224 nM Car, and at 6
different concentrations of pH neutralized benzoic acid (0.1, 0.5, 1, 5, 10, and 25 mM). All
90
concentrations were assayed in triplicate. MOPS was used instead of Tris to buffer our reactions
given the propensity of Tris to react with aldehydes (170, 171).
Michaelis-Menten parameters for YtbE on benzaldehyde were determined by measuring
changes in absorbance at 340 nm for up to 5 minutes. Reactions were prepared as follows: 100
mM MOPS-NaOH [pH 7.0], 10 mM MgC 2 , 0.6 mM NADPH, 1 mM ATP, 1422 nM YtbE, and at 6
different concentrations of benzaldehyde (1, 5, 10, 15, 25, and 35 mM). All concentrations were
assayed in triplicate.
For all in vitro experiments excluding initial rate measurements, samples were quenched
using 1% TFA and then subject to centrifugation. Aqueous supernatant was collected for HPLC
analysis using either an Agilent 1100 series or 1200 series instrument equipped with a diode array
detector. Wavelengths of 223, 242, and 192 nanometers were used to detect benzoic acid,
benzaldehyde, and benzyl alcohol, respectively. The benzoate family of analytes was separated
using an Aminex HPX-87H anion-exchange column (Bio-Rad Laboratories), with a mobile phase
consisting of 70% 5 mM H 2 SO4 and 30% acetonitrile. All three compounds eluted within 35
minutes at a flow rate of 0.4 ml/min. Column temperature was maintained at 300 C. All chemicals
reported in figures were quantified using calibration of standards on the HPLC instrument and
linear interpolation. All experiments were performed in duplicate. Data points shown are
averages with error bars representing standard deviations.
Conversion of vanillate to vanillin was determined using a Zorbax Eclipse XDB-C18 column
(Agilent) and detected using a wavelength of 280 nm. A gradient method used the following
solvents: (A) 50% acetonitrile + 0.1% trifluoroacetic acid (TFA); (B) water + 0.1% TFA. The gradient
91
began with 5% Solvent A and 95% Solvent B. The setting at 20 minutes was 60% Solvent A and
40% Solvent B. The program restored the original ratio at 22 minutes and ended at 25 minutes.
The flow rate was 1.0 ml/min and all compounds of interest eluted within 15 minutes. Column
temperature was maintained at 30*C.
4.3. Results
From initial in vitro experiments, we observed three related and unexpected phenomena:
(i) reaction progress terminated prematurely and did not progress further upon addition of more
Car; (ii) the final concentration of benzoate observed was sensitive to the concentration of Mg 2+,
which is required for formation of the acyl-adenylate intermediate; and, (iii) a precipitate
gradually formed in the reaction cuvettes that contained all components necessary for reaction.
With regard to the latter observation, upon testing all possible pairs of substrates, products, and
co-factors, we found that Mg 2+and PPR were responsible for precipitate formation (Table 4-1). To
investigate further, we decided to test the use of a commercial inorganic pyrophosphatase (Ppa),
which dissociates PPi into inorganic phosphate. These enzymes are abundant in living cells and
are essential for growth of E. coli (172-177).
92
Table 4 - 2. Combinatorial testing of in vitro components for formation of precipitate.
As shown in Figure 4-1, an increase in the concentration of Mg 2+from 20 mM to 100 mM
increases the final conversion but not the rate of reaction within the first 15 minutes after
addition of Car. On the other hand, addition of Ppa increases both the final conversion and the
initial rate of reaction. Under conditions tested, addition of Ppa was required in order for the
relatively low starting concentration of 1 mM benzoate to be fully converted to benzaldehyde.
93
,
1.0
0.9*
0.8
*- .
.
-
0.4
CO
*
0.3
--
0.2
-- A. 100 mM MgC2
0.1
--
0.0
---
0
.
0.5.
20mM MgC2
20 mM MgCI2 and Commercial Ppa Added
r
5
10
15
20
25
30
35
40
45
Time (minutes)
Figure 4 - 1. Effect of varying MgC2 concentration or adding commercial inorganic
pyrophosphatase (Ppa) from New England Biolabs on Car-catalyzed conversion of the substrate
benzoate. The concentration of Car used was 224 nM. The units of Ppa added was 0.1, where
one unit is as defined by NEB (The amount of enzyme that will generate 1 pimol of phosphate per
minute from inorganic pyrophosphate under standard reaction conditions [a 10 minute reaction
at 25 0 C in 20 mM Tris-HCI, pH 8.0, 2 mM MgC 2 and 2 mM PPi]).
To investigate the effect of Ppa addition further, we next modified the system in two
ways. First, we overexpressed the E. colippa gene and purified it ourselves, thereby ensuring that
no other component from the commercial Ppa mixture was responsible for the observed
enhancement. Second, we added an enzyme that would function as a sink for benzaldehyde. We
purified YtbEBS, which is a heterologous aldo-keto reductase known to catalyze the conversion of
benzaldehyde into benzyl alcohol. Figure 4-2 illustrates first that the problem of early termination
and slow kinetics is not simply solved when there is a sink for benzaldehyde. This suggests that
the aromatic product is not inhibitory. However, the use of Ppa coupled to Car enhances flux
94
through this in vitro pathway. At first, it appeared as though higher Mg2 + concentrations might
counter the beneficial effect of Ppa addition. However, later experiments showed that the higher
Mg2+ concentration slightly reduced YtbE activity. For all following experiments, 10 mM MgC2
was used.
Nocardia iowensis Car
0
BacV Uts 3Ubtlls
1.00
0.90
S0.80
OH
TVN
--
Oro
-3
benzaldehyde
benzoic acid
YtbE
C K*
OH
benzyl alcohol
0.70
C
0.60
0.50
0.40
0.30
0.20
0.10
0.00
10 mM MgCI2,
No Ppa
100 mM Mgcl2,
No Ppa
U Benzoate
U Benzaldehyde
10 mM MgCI2,
In-House Ppa
100 mM Mgcl2,
In-House Ppa
U Benzyl alcohol
Figure 4 - 2. Effect of MgC 2 concentration and addition of an "in-house" Ppa (896 nM) on an in
vitro reaction pathway involving Car (224 nM) and a heterologous aldo-keto reductase, YtbE
(1422 nM). To ensure that no other component of the commercial Ppa mixture was responsible
for the reaction enhancement, we expressed and purified the E. coli ppa gene product. We
included an aldo-keto reductase that catalyzes the conversion of benzaldehyde into benzyl
alcohol to investigate whether the reaction catalyzed by Car would be enhanced simply by
creating a sink for the product. Subsequent experiments showed that the higher concentration
of MgC 2 slightly reduced the activity of the second enzyme.
We next sought to understand whether the enhancement observed by Ppa addition
allowed reaction kinetics to be predictable based solely on initial rate measurements and
Michaelis-Menten kinetic equations. We performed initial rate measurements for CarN with
95
respect to benzoate and YtbEBs with respect to benzaldehyde. We then tested the performance
of the two-step in vitro pathway with and without addition of Ppa. Figure 4-3 depicts how the
kinetics of Car and YtbE, when modeled in simple Michaelis-Menten form using parameters
obtained from initial rate measurements, soon deviate from the model predictions in the absence
of pyrophosphatase. When pyrophosphatase is included, however, the in vitro pathway kinetics
perform as expected.
No Ppa Added
0.8-
0.8
X
0.7-
0.7
0.6-
E
0
0
C
0.5-
E30.5
C:
0.4-
)
(D0.4-
.
i
Ppa Added
0.9
-
0.9-
0.3-.
0.3 -
jI
0.2-
0.2-
0.1-
0
Xj
5
0.1
10
15
20
Time (Minutes)
25
30
5
10
15
20
Time (Minutes)
25
30
Figure 4 - 3. Addition of Ppa enables an in vitro pathway involving Car and an aldo-keto reductase
to be modeled with far greater accuracy using Michaelis-Menten kinetics and parameters
obtained from initial rate measurements. Lines represent simulated concentrations, whereas "x"
and "o" symbols represented observed concentrations. Model parameters: KM, Car-GBD = 0.35 mM;
KM, YtbE-SH3
=
2 mM; kcat, car-GBD
=
216 min-; kcat, YtbE-SH3 = 96 min-1 . Enzyme concentrations: Car = 224
nM; YtbE = 1422 nM. Error bars omitted here for clarity.
Figure 4-4 most clearly demonstrates the benefit provided by coupling Ppa to Car for the
in vitro reduction of two substrates that lead to formation of valuable flavor compounds. The
starting concentrations of 5 mM were also chosen to be larger to demonstrate utility for
96
preparative chemistry. Figure 4-4 illustrates that the pairing of Ppa to Car can more than double
the final conversion of substrate.
1.00
0.90
0.80
0.70
0.60
x
C/0
5 mM, CE
1.1 pM
0.50
0.40
0.30
0.20
0.10
0.00
0
20
40
60
80
100
120
Time (minutes)
-U--Benzoate: With Ppa
--
Benzoate: No Ppa
-M-Vanillate: With Ppa
-A-Vanillate: No Ppa
Figure 4 - 4. Effect of Ppa addition on the Car-catalyzed conversions of two substrates that result
in aldehydes valuable as flavors. x represents the conversion of substrate C (X = C/Co).
Finally, after establishing that Ppa addition is beneficial for in vitro reactions involving Car,
we wanted to determine how much Ppa might be required. In order to gain insight into the
relationship between Ppa concentration and Car concentration, we varied the molar ratio of Ppa
to Car. Figure 4-5 explores the relationship between the concentration of Ppa in the assay
solution relative to the concentration of Car and shows that lowest molar ratio tested (1:4) is
97
sufficient to see maximal enhancement. Assuming kinetics would be unaffected, this suggests
that a direct fusion of Car-Ppa would display maximally enhanced in vitro kinetics while
simplifying and reducing the cost of protein purification.
1.00
0.90
E
0.80
0.70
E2 0.60
0.50
T
tsC 0.40
0
0.30
0.20
0.10
0.00
1:4
1:1
4:1
Molar Ratio of Ppa to Car
U Benzoic Acid
1 Benzaldehyde
Figure 4 - 5. Effect of the molar ratio of Ppa to Car on conversion of benzoate. The concentration
of Car was fixed at 224 nM. The purpose of this experiment was to help determine the minimum
amount of Ppa required to add relative to Car in order to achieve saturating levels of
enhancement. All ratios tested achieved saturating enhancement.
4.4. Discussion
Inorganic pyrophosphatase has been used alongside pyrophosphate-generating reactions
to assay enzyme kinetics with increased sensitivity based on formation of two moles of
phosphate for every one mole of pyrophosphate (178). In a manner analogous to its use here,
Ppa has also been used to enhance in vitro transcription reactions since the 1990s (179, 180). A
98
thermodynamic justification for why pyrophosphate hydrolysis would benefit in vitro RNA
synthesis was documented even earlier, in 1975 (181). In addition to increasing the yield of RNA
produced, Ppa was also shown to minimize the effect of the Mg" concentration on product yields
(179). Interestingly, the addition of Ppa increased synthesis of transcripts roughly twofold, and
the lowest concentration of Ppa tested provided the full effect. Although we were initially
unfamiliar with in vitro transcription, these observations are very consistent with the results
reported here for pairing Ppa with Car.
Based on the long history of Ppa use for in vitro transcription, it is somewhat surprising
that Ppa is not a standard addition to other preparative-scale reactions or in vitro reconstituted
pathways that are known to involve adenylate-forming enzymes (182). For example,
reconstitution of nonribosomal peptide synthesis has occurred without the addition of Ppa
although it features enzymes that contain adenylation domains (183). In the case of carboxylic
acid reductases, there have been examples within roughly the past decade attempting largerscale in vitro aldehyde synthesis (22) or in vitro reconstitution of a multi-enzyme pathway
featuring Car (25, 26) without Ppa addition. In such scenarios and any others in which more data
than initial rate measurements is desired, Ppa should be included in the reaction to avoid early
termination. Furthermore, this study demonstrates that reaction progress can be modeled
predictability when Ppa is present.
Interestingly, when one of the first enzymes of this class from Neurospora crassa was
characterized by Gross in 1972 and demonstrated to form an acyl-adenylate intermediate, Ppa
was used in the ATP exchange assay (184). Three years prior to that result, Gross and Zenk had
formulated the reaction as involving the formation of ADP and inorganic phosphate rather than
99
AMP and PPR, and thus they had no rationale to include Ppa in the original activity assay (185).
When Kato and colleagues reported characterization of a related aromatic acid reductase from
Nocardia asteroides in 1991, they assayed activity using the method of Gross and Zenk (without
Ppa), although they independently confirmed the presence of an acyl-adenylate intermediate
using Ppa (186). To the best of our knowledge, the use of Ppa alongside carboxylic acid reductases
has not been reported since then and never reported in a preparative reaction context.
The results from this study provide an explanation for the discrepancy in performance of
Car observed in vitro versus in vivo. Because pyrophosphatases are essential and abundant in E.
coli and other organisms, inhibitory byproduct formation is not expected to occur in an in vivo
context. This may mean that cell-free aldehyde biosynthetic processes that utilize cellular lysate
would not suffer from this drawback as well. Of course, introduction of many other cellular
components along with pyrophosphatase would decrease the ease of product purification, a
supposed advantage of an in vitro biosynthetic process. Overall, given currently limited
understanding of microbial aldehyde toxicity and high market values for numerous aldehydes,
we expect these results to aid in informing the development of in vitro preparative aldehyde
biosynthesis for applications in flavor and related industries.
100
Chapter 5: Towards improving de novo vanillin biosynthesis in E. coil by
deregulating S-adenosylmethionine biosynthesis
5.1. Introduction
As mentioned at the outset of the previous chapter, Chapters 2 and 3 revealed how a
variety of aldehydes could be produced in small quantities using engineered E. coli. Beyond
synthesis of novel compounds, an important component of the discipline of chemical engineering
is the framework required to analyze a process, to determine its limitations, and to ultimately
improve the process. The field of metabolic engineering has developed an enormous toolkit of
approaches that aid these objectives.
In this chapter, we describe our efforts to better understand the pathway responsible for
production of the model aromatic aldehyde vanillin. Now that E. coli has been shown to produce
and retain vanillin, this study aims to understand how to increase the productivity (g/L.h) and
titer (g/L) of vanillin produced from glucose as a sole carbon source in E. coli using metabolic
engineering approaches. The focus on metrics such as productivity and titer rather than yield
(gproduct/gsubstrate) reflects vanillin's current status as a high-value molecule. Natural vanillin is
roughly priced between $1000-$2000/kg, whereas artificial vanillin produced using chemical
conversion processes is priced around $15/kg (89). In comparison, the substrate glucose is priced
at approximately $0.5/kg (41).
In the example of vanillin production previously engineered in E. coli (149), the
engineered
vanillin
pathway relied
on
virtually identical
heterologous
enzymes
(a
dehydroshikimate dehydratase, an O-methyltransferase, and a carboxylic acid reductase) as were
101
reported for engineering de novo biosynthesis of vanillin in yeast (Fig. 5-1) (41). When de novo
biosynthesis was first reported in both organisms, vanillin titers achieved were similarly low
under somewhat comparable flask conditions (65 mg/L in Schizosaccharomyces pombe supplied
with yeast extract-based media versus 119 mg/L or 56 mg/L in E. coli supplied with glucose in LB
or M9 minimal media, respectively). Production of vanillin or vanillin-0-D-glucoside in yeast has
since been enhanced and even commercialized by the Swiss start-up Evolva. Production of the
glucoside form of vanillin was motivated by reduced toxicity, increased product secretion in
yeast, and prolonged flavor retention when consumed.
102
A
T
PYR
AP A
PykF I
PyFADP
.-
-. ---- --
Aromatic
Amino Acids
T
PpsA
AP+P
PEP
AroG*
Glucose
AroB
AroD
DHQ
DAHP
F6P
-- - --
+-
DHS
AroE
1
Shikimate
TktA
E4P
P.
P.
G3P
--
vanillin
Heterologous
Pathway
XsP
B
0
0
3-dehydroshikimate
(DHS)
Car
AsbF
protocatechuate
protoc
th
y
OMTO
0
Heterologous
Pathway
vanillate
Figure 5 - 1. The engineered vanillin pathway in E. coli. (A) Endogenous portion of the vanillin
pathway. (B) Heterologous portion of the vanillin pathway, with reactions catalyzed by CarNi
shaded in gray. Genes corresponding to enzymes labeled in red are overexpressed in experiments
investigating improvement of vanillate production. Enzymes written without subscripts are
native to E. coli. The heterologous pathway portion in the engineered yeast vanillin pathway
contains identical metabolites and enzymes with the exception of AsbFot.
Two academic studies, one of which was in conjunction with Evolva, were published after
the initial demonstration of the engineered yeast vanillin pathway and describe improvements
obtained using Flux Balance Analysis (FBA) model-guided optimization (42,43). In the first report,
103
OptGene was used to identify target reactions that, if deleted, would increase vanillin production.
Different deletion targets were suggested depending on the reference flux distribution in the
Minimization of Metabolic Adjustment (MOMA) biological objective function. In general, the
predicted benefit from these modifications was increased availability of the co-factors ATP and
NADPH required by the carboxylic acid reductase from Nocardia iowensis (CarNm). Genes related
to pyruvate metabolism, ammonium metabolism, the pentose phosphate pathway, and central
carbon metabolism were identified. The only three modifications tested experimentally were
deletion of one of the pyruvate decarboxylases (PDC1), deletion of the most active glutamate
dehydrogenase (GDH1), and overexpression of GDH2 to ensure sufficient nitrogen uptake in the
absence of GDH1. These modifications led to an overproducer strain that, when cultivated in a
low dilution rate continuous fermentation, resulted in vanillin--D-glucoside titers of 500 mg/L
(42). As of that study, identification and overexpression of a potential rate-limiting enzyme in the
pathway had not yet been described.
In the second report, which was a short communication, the O-methyltransferase from
Homo sapiens (OMTHS) and CarNi were overexpressed in the highest-producing strain obtained
from the previous study (43) based on the observed accumulation of two heterologous
intermediates (protocatechuate and protocatechualdehyde). CarNi overexpression did not lead
to an increase in production, whereas OMTHs overexpression led to a 30% increase in titer (now
380 mg/L vanillin--D-glucoside compared to a different baseline than referenced above).
However, OMTHs overexpression in the parental strain from the first study did not alter titer. The
authors concluded that this was likely because CarNi was limiting due to low availability of ATP
and NADPH without the model-guided modifications made in the first study.
104
Despite our increased level of understanding of the engineered yeast vanillin pathway,
and despite the presence of generally similar challenges in the engineered E. coli vanillin pathway
(e.g., low initial titers of all heterologous metabolites, accumulation of heterologous
intermediates, and vanillin toxicity), E. coil is a fundamentally different host than yeast for the
pathway and we believed it was likely that pathway limitations would vary. In this report, we
describe metabolic engineering experiments used to identify key areas for improvement in the
production of vanillin from glucose as a sole carbon source using E. coli. We employed a general
strategy that targeted upstream improvements first and then focused on changes in downstream
heterologous metabolite titers to better identify pathway bottlenecks. The novelty of insights
presented increases as further downstream pathway components are investigated. Overall, our
findings indicate that limited availability of S-adenosylmethionine (SAM), which is regulated very
differently in E. coli versus yeast, represents a major hurdle to achieving improved vanillin titers
from glucose as a sole carbon source in E. co/i. Although this observation has been made
previously with respect to vanillate production in E. coli (40), this study takes the additional steps
of deregulating SAM biosynthesis and examining the efficacy of those interventions.
5.2. Materials and Methods
5.2.1. Strains and plasmids
E. coli strains and plasmids used in this study are listed in Table 5-1. Molecular biology
techniques were performed according to standard practices (98) unless otherwise stated.
Molecular cloning and vector propagation were performed in DH5a. All host strains used for
production experiments were derived from E. coli K-12 MG1655(DE3). In order to construct new
105
host strains, two methods were used. The first was P1 transduction (100) using donor strains
from the Keio collection (99) and P1 bacteriophage from ATCC (25404-B1). P1 transduction was
used for all deletions of single genes. The second method was recombineering using the X Red
system (101). Recombineering was used to delete the yqhC-dkgA operon and to upregulate ga/P
expression by promoter substitution, as previously described (149). Oligonucleotides were
purchased from Sigma. Q5 High Fidelity DNA Polymerase (New England Biolabs) was used for
DNA amplification. In all cases of host strain modifications, pCP20 was used to cure the
kanamycin resistance cassette (101).
106
Table 5 - 1. Strains and plasmids used in this study
UHdat
MG1655
MG1655(DE3)
1- Wt5jaCZaMi5 atacZYA-argF) U1b9 recAi enaA nsaKiI jrr-,
mK+) phoA supE44 X- thi-1 gyrA96 re/A1
F mcrA A(mrr-hsdRMS-mcrBC) $80/acZAM15 AlacX74 recAl endAl
araD139A(ara, leu)7697 ga/U gaK k rpsL nupG
F- k i/vG- rfb-50 rph-1
F k ilvG- rfb-50 rph-1 (DE3)
ATCC 700926
Ref. (102)
RARE
MG1655(DE3) AdkgB AyeaE A(yqhC-dkgA) AyahK AyjgB
Ref. (149)
RARE Ameti
MG1655(DE3) AdkgB AyeaEA(yqhC-dkgA) AyahKAyjgBAmetl
This study
PTS- glu'
MG1655(DE3) AptsHlcrr PgIk::Pcon* galPq
This study,
but based on
Ref. (187)
PTS- glu* RARE'
MG1655(DE3)
PTS- glu* RAR E'
MG1655(DE 3) AptsH/crrPglk::Pcon- galPpq(yqhC-dkgA) AyahK AyjgB Ametd
This study
pCP20
Xc1857 (ts), X pr Repts, AmpR, CMR, X p, FLP
CGSC 7629
pKD13
oriRy, AmpR, kan
oriR101, repA101P, A mp', araC, araBp-Av-Ae-Aexo
CGSC 7633
AmpR, lac/, T7ac
CmR, lac/, T7/ac
KanR, lacd, T7/ac
StrR, lad, T7/ac
pACYCDuet-1 harboring caropt (carboxylic acid reductase from Nocardia
iowensis, codon optimized for expression in E. coli) and sfpopt
(phosphopantetheinyl transferase from Bacillus subtilis, codon optimized
for expression in E. coli)
pETDuet-1 harboring Hs-S-COMTpt (catechol O-methyltransferase from
Homo sapiens, codon optimized for expression in E. coli) and asbFopt
(dehydroshikimate dehydratase from Bacillus thuringiensis, codon
optimized for expression in E. coli)
Plasmid containing the shikimate module, version 4, kindly provided by the
Keasling Lab at UC Berkeley. (Source of aroG*-ppsA-tktA artificial operon)
pACYCDuet-1 harboring the feedback-resistant aroG* from E. coli
pACYCDuet-1 harboring three E. coli genes in an artificial operon: aroG*,
ppsA, and tktA
pCOLADuet-1 harboring the E. coli metK gene
pETDuet-1 harboring Hs-S-COMTopt
pCOLADuet-1 harboring a feedback-desensitized version of E. coli metA
Novagen
Novagen
Novagen
DH10B
Ametd
pK D46
pETDuet-1
pACYCDuet-1
pCOLADuet-1
pCDFDuet-1
pACYC-car-sfp
pET-OMT-asbF
pS4
pACYC-aroG*
pACYC-aroG*ppsA-tktA
pCOLA-metK
pET-OMT
pCOLA-metA*
AptsH/crr PgIk::Pcon* galP A(yqhC-dkgA) AyahK AyjgB
invitrogen
Invitrogen
This study
CGSC 7739
Novagen
Ref. (149)
Ref. (149)
Ref. (104)
This study
This study
This study
This study
This study
(27 Arg->Cys, 296_Ile->Ser, and 298 Pro->Leu, "metA*")
pCOLA-cysE*
pCOLADuet-1 harboring a feedback-desensitized version of E. coli cysE
This study
(95 Val->Arg, and 96 Asp->Pro, "cysE*")
107
pCOLA-metA*cysE*
pCDF-car-sfp
pCOLADuet-1 harboring an artificial operon consisting of the metA* and
cysE* genes
pCDFDuet-1 harboring an artificial operon containing caropt and sfpopt
This study
This study
The aroG*, ppsA, and tktA genes were kindly provided by Professor Jay D. Keasling at the
University of California, Berkeley (USA). The genes encoding metA* and cysE* were synthesized
as gBlocks (IDT, San Jose, CA) and their sequences are included in Table 5-2. The E. colimetK gene
was amplified from MG1655(DE3) genomic DNA using PCR amplication and the oligonucleotides
shown in Table 5-3. All genes of interest were cloned into the Duet vector system (Novagen) using
restriction digest-based cloning. Restriction enzymes and T4 DNA ligase were purchased from
New England Biolabs. Propagated constructs were purified using a QlAprep Miniprep Kit (Qiagen)
and agarose gel fragments were purified using a Zymoclean Gel DNA Recovery Kit (Zymo
Research). All constructs were confirmed to be correct by nucleotide sequencing (Genewiz).
108
Table 5 - 2. Synthesized gene sequences used in this study.
MetA*
(Ndel/Aatil)
AAAAAACATATGCCGATTCGTGTGCCGGACGAGCTACCCGCCGTCAATTTCTTGCGTGAAGAAAACGTCT
TTGTGATGACAACTTCTTGTGCGTCTGGTCAGGAAATTCGTCCACTTAAGGTTCTGATCCTTAACCTGATG
CCGAAGAAGATTGAAACTGAAAATCAGTTTCTGCGCCTGCTTTCAAACTCACCTTTGCAGGTCGATATTCA
GCTGTTGCGCATCGA17CCCGTGAATCGCGCAACACGCCCGCAGAGCATCTGAACAACTTCTACTGTAACT
TTGAAGATATTCAGGATCAGAACTrGACGGTTTGATTGTAACTGGTGCGCCGCTGGGCCTGGTGGAGTT
TAATGATGTCGCTTACTGGCCGCAGATCAAACAGGTGCTGGAGTGGTCGAAAGATCACGTCACCTCGAC
GCTGTTTGTCTGCTGGGCGGTACAGGCCGCGCTCAATATCCTCTACGGCATTCCTAAGCAAACTCGCACC
GAAAAACTCTCTGGCGTTTACGAGCATCATATTCTCCATCCTCATGCGCTTCTGACGCGTGGCTTTGATGA
TTCATTCCTGGCACCGCATTCGCGCTATGCTGACTTTCCGGCAGCGTTGATTCGTGATTACACCGATCTGG
AAATTCTGGCAGAGACGGAAGAAGGGGATGCATATCTGTTTGCCAGTAAAGATAAGCGCATTGCCTTTG
TGACGGGCCATCCCGAATATGATGCGCAAACGCTGGCGCAGGAA1TTTCCGCGATGTGGAAGCCGGAC
TAGACCCGGATGTACCGTATAACTATTTCCCGCACAATGATCCGCAAAATACACCGCGAGCGAGCTGGCG
TAGTCACGGTAATTTACTGTTTACCAACTGGCTCAACTATTACGTCTACCAGAGCACGCTATACGATCTAC
GGCACATGAATCCAACGCTGGATTAAGACGTCAAAAAA
CysE*
(AatlI/Xhol)
AAAAAAGACGTCTAATAAAAGGAGATATACCATGTCGTGTGAAGAACTGGAAATTGTCTGGAACAATAT
TAAAGCCGAAGCCAGAACGCTGGCGGACTGTGAGCCAATGCTGGCCAGT1TrTACCACGCGACGCTACTC
AAGCACGAAAACCTTGGCAGTGCACTGAGCTACATGCTGGCGAACAAGCTGTCATCGCCAATTATGCCTG
CTATTGCTATCCGTGAAGTGGTGGAAGAAGCCTACGCCGCTGACCCGGAAATGATCGCCTCTGCGGCCTG
TGATATTCAGGCGGTGCGTACCCGCGACCCGGCAAGACCCAAATACTCAACCCCGTTGTTATACCTGAAG
GGTTTTCATGCCTTGCAGGCCTATCGCATCGGTCACTGGTTGTGGAATCAGGGGCGTCGCGCACTGGCAA
TCTTTCTGCAAAACCAGGTTTCTGTGACGTTCCAGGTCGATATTCACCCGGCAGCAAAAATTGGTCGCGGT
ATCATGCTTGACCACGCGACAGGCATCGTCGTTGGTGAAACGGCGGTGATTGAAAACGACGTATCGATTC
TGCAATCTGTGACGCTTGGCGGTACGGGTAAATCTGGTGGTGACCGTCACCCGAAAATTCGTGAAGGTG
TGATGATTGGCGCGGGCGCGAAAATCCTCGGCAATATTGAAGTTGGGCGCGGCGCGAAGATTGGCGCA
GGTTCCGTGGTGCTGCAACCGGTGCCGCCGCATACCACCGCCGCTGGCGTTCCGGCTCGTATTGTCGGTA
AACCAGACAGCGATAAGCCATCAATGGATATGGACCAGCATTTCAACGGTATTAACCATACATTTGAGTA
TGGGGATGGGATCTAACTCGAGAAAAAA
109
Table 5 - 3. Oligonucleotides used in this study.
MetJ-verify-f
MetJ-verify-r
PtsHICrr-verify-f
PtsHICrr-verify-r
AroG*-f (BgIII)
AroG*-r (AvrIl)
AroG*-PpsA-TktA-f (BgIll)
AroG*-PpsA-TktA-r (AvrIl)
MetK-f (Ndel)
MetK-r (Avri)
TCTTTAGCAATCACCACG
GGAATATTCTTGCCGTAAC
GAAAGGCGCAATCCAA
CGATTTGACTGCCAGAAT
AAAAAAAGATCTGATGAATTATCAGAACGACGATTTAC
AAAAAACCTAGGCCTCCTTTAGATCCTTACCC
AAAAAAAGATCTGATGAATTATCAGAACGACGATTTAC
AAAAAACCTAGGTTACAGCAGTTCTTTTGCTTTC
AAAAAACATATGGCAAAACACC11TFUAC
AAAAAACCTAGGTTACTTCAGACCGGCAG
5.2.2. Chemicals
The following compounds were purchased from Sigma: vanillic acid, vanillin, 3,4dihydroxybenzoic acid (otherwise known as protocatechuic acid), 3,4-dihydroxybenzaldehyde
(protocatechualdehyde), L-methionine, L-homocysteine, L-cysteine, and L-aspartate. Isopropyl 1D-1-thiogalactopyranoside (IPTG) was purchased from Denville Scientific. Ampicillin sodium salt,
chloramphenicol, streptomycin sulfate, and kanamycin sulfate were purchased from Affymetrix.
5.2.3. Culture conditions
In experiments described in this chapter, a IX M9 salt medium (Sigma, Aldrich) containing
6.78 g/L Na2HPO 4 -7H20, 3 g/L KH 2PO 4, 1 g/L NH 4C, and 0.5 g/L NaCl, supplemented with 2 mM
MgSO 4 , 0.1 mM CaC1 2, glucose, trace elements, and antibiotics was used as the culture medium.
2
,
The trace element solution (10OX) used contained 5 g/L EDTA, 0.83 g/L FeCI3-6H 2 0, 84 mg/L ZnC
10 mg/L CoC12-6H20, 13 mg/L CuCI2-2H 20, 1.6 mg/L MnC12-2H20 and 10 mg/L H3 B0 3 dissolved in
water. This was added to a concentration of LX to supplement the M9-glucose medium. This
110
medium will be henceforth referred to as "M9-glu-trace." For most experiments, the initial
glucose concentration was 1.8%, or otherwise it was 1.2%.
With the exception of bioreactor experiments, all experiments were performed in 250 ml
baffled PYREX shake flasks that contained 50 ml culture volumes. Overnight cultures were grown
in 3 ml in 14 ml round-bottom tubes (Corning). Experimental cultures were initiated as follows:
1% (v/v) inoculum volumes of overnight culture in LB medium were first transferred into
overnight culture in M9-glu-trace medium, and then 1% (v/v) inoculum volumes of overnight
culture in M9-glu-trace were transferred into 50 mL M9-glu-trace medium, incubated at 30*C,
and agitated at 250 rpm. The OD 6oo was measured regularly during exponential growth using a
DU800 UV/Vis spectrophotometer (Beckman Coulter). Depending on the experiment, culture
medium was supplemented with either 50 mg/L ampicillin, 17 mg/L chloramphenicol, 25 mg/L
streptomycin, 25 mg/L kanamycin, or combinations of the previous antibiotics to provide
selective pressure for plasmid maintenance. All experiments were performed in biological
triplicate, and results are presented as averages with error bars representing one standard
deviation.
For experiments in which S-adenosylmethionine precursors were supplemented, flask
cultures were set up as mentioned before but with culture volumes adjusted to achieve final
concentrations as follows: 10 mM L-methionine, 2.5 mM L-homocysteine, 10 mM cysteine, or 10
mM aspartate. Stocks of supplemented metabolites were pH-neutralized and sterile filtered. In
these experiments, control cultures received an equal volume of sterile deionized water instead
of the metabolic precursors at the time of supplementation. Supplementation times varied from
at induction (0 h) to twenty four hours after induction.
111
Bioreactor experiments were performed using a Labfors 3 bioreactor (Infors, Bottmingen,
Switzerland) with a maximum working volume of 2.3 L. A D140 OxyProbe dissolved oxygen sensor
(Broadley-James, Irvine, CA) and an F-695 FermProbe pH electrode (Broadley-James) were used
to monitor the dissolved oxygen and pH, respectively. M9-glu-trace medium was used to grow
the bioreactor cultures. Cultures were inoculated to an initial OD6oo of 0.1 from cells obtained
from a 50 mL overnight M9-glu-trace seed culture.
To set up the bioreactor, 1.5 L of M9 salt medium was autoclaved in the reactor. On the
day of inoculation, the medium was supplemented with glucose, CaC1 2, MgSO 4, appropriate
antibiotics, and trace elements. The dissolved oxygen setpoint was controlled at 35% of the
saturation value using a cascade to agitation (250 rpm to 850 rpm), and air was provided at a
constant flow rate of 1 vvm. pH was controlled at the desired setpoint using 4 M NaOH and 2 M
H 3PO 4. Online data was logged using IRIS fermenter log and control software (Infors). Antifoam
was manually added in 0.1 mL increments as needed. Samples were taken periodically to
measure OD60 0 offline using a DU800 Spectrophotometer (Beckman Coulter, Brea, CA) and to
collect culture supernatants for metabolite analysis.
5.2.4. Metabolite analysis
Culture samples were pelleted by centrifugation and aqueous supernatant was collected
for HPLC analysis using an Agilent 1100 series instrument equipped with a diode array detector.
Heterologous compounds produced in vanillin experiments were separated using a Zorbax
Eclipse XDB-C18 column (Agilent) and detected using a wavelength of 280 nm. A gradient method
+
used the following solvents: (A) 50% acetonitrile + 0.1% trifluoroacetic acid (TFA); (B) water
112
0.1% TFA. The gradient began with 5% Solvent A and 95% Solvent B. The setting at 20 minutes
was 60% Solvent A and 40% Solvent B. The program restored the original ratio at 22 minutes and
ended at 25 minutes. The flow rate was 1.0 ml/min and all vanillin pathway compounds of
interest eluted within 15 minutes. Column temperature was maintained at 300 C.
5.2.5. SDS-PAGE analysis
To determine qualitative protein expression level of OMT in the absence of other pathway
gene overexpression, E. coli MG1655(DE3) was transformed with empty pETDuet-1 or pET-OMT.
Single colonies from plates of each transformation were grown overnight in 3 ml of LB with
appropriate antibiotic. Cells were passaged into second overnight cultures by inoculating 3 ml of
M9 + 1.8% glucose with 100 uL of the overnight LB cultures. Shake flask cultures containing 50
ml M9 + 1.8% glucose were inoculated at 1% inoculum from overnight M9 cultures and incubated
with agitation at 30*C and 250 rpm. Shake flasks were induced with 0.5 mM IPTG at OD6oo values
between 0.8-1.0. Twenty four hours after induction, 5 ml of each culture were sampled and
pelleted by centrifugation. Cell pellets were resuspended in 1 ml of 10 mM Tris-HCI at pH 8.0 and
lysed using sonication. After lysis, samples were pelleted by centrifugation (6,000g, 40 C, 10 min)
and the supernatant was removed as soluble lysate. The remaining pellet was resuspended in 10
mM Tris-HCI and deemed the insoluble fraction.
To determine qualitative protein expression level of OMT in the presence of other
pathway gene overexpression along with other pathway constructs, 5 ml samples were taken at
regular time intervals from a representative bioreactor experiment. Cells were then pelleted,
resuspended, and lysed as mentioned above.
113
Total protein was quantified by the Bradford assay method (169) using Bio-Rad Protein
Assay Dye Reagent (Cat #500-0006) and a bovine serum album (BSA) standard. A Bio-Rad 10%
Mini-PROTEAN TGX gel (Cat #456-1034) was run using the Mini-PROTEAN
Tetra Cell
electrophoresis apparatus. Bio-Rad Precision Plus Protein All Blue Standard (Cat #161-0373) and
10 pg of total protein for each sample was loaded on the gel. After running at 200 volts for 33
minutes, the gel was washed with deionized water before staining with Bio-Rad Bio-Safe
Coomassie Stain (Cat #161-0786).
5.3. Results
5.3.1. Focusing on central carbon metabolism
As a starting point for understanding pathway limitations, an investigation of metabolic
perturbations upstream of the heterologous pathway was motivated by three prior results: (i)
the reported kinetics of AsbF, which efficiently catalyzes protocatechuate from endogenous 3dehydroshikimate (106); (ii) the low titers of all heterologous metabolites reported during initial
de novo vanillin production in E. coli from glucose as a sole carbon source (149); and, (iii) the
accumulation of protocatechuate observed in the same study (149). Although these results
indicated that steps downstream of protocatechuate synthesis were not performing optimally,
we hypothesized that if increased flux could first enter the heterologous pathway, then it would
be easier to identify bottlenecks based on changes in titer (mg/L) and specific yields (g/gDcw) of
intermediate metabolites.
The vanillin pathway relies on endogenous aromatic amino acid biosynthesis, and efforts
to improve aromatic amino acid biosynthesis in E. coli have been well-documented (93-95).
114
Previous studies demonstrated two strategies that were successfully used to increase availability
of two key aromatic precursor metabolites that thereby increased titers of aromatic products.
These endogenous metabolites are phosphoenolpyruvate (PEP) from glycolysis and erythrose-4phosphate (E4P) from the pentose phosphate pathway, which condense to form the first
committed step towards aromatic amino acid biosynthesis. The first strategy to increase their
availability is deletion of the phosphotransferase system (PTS), which is the primary means for
glucose import and consumes one molecule of PEP per molecule of glucose. Growth of a PTSstrain on glucose as a sole carbon source can be made viable by upregulating the gene encoding
galactose permease (gaiP), which allows glucose entry independent of PEP consumption (PTSglu'). A second documented strategy is to overexpress the genes encoding PEP synthase (ppsA)
and transketolase (tktA) (93, 95). PEP synthase catalyzes the conversion of PEP into pyruvate, and
transketolase catalyzes the reversible formation of E4P and xylulose 5-phosphate from fructose
6-phosphate and glyceraldehyde 3-phosphate (Fig. 5-1).
When we engineered a PTS- glu' variant of the RARE strain (PTS- glu+ RARE'), we observed
that use of this host compared to the RARE host did not improve titers of either protocatechuate
or vanillate (Fig. 5-2A). However, because the PTS- glu' modification had been pursued by many
others, we decided to continue using that host until we would reevaluate its performance against
the RARE host with final plasmid constructs or if further host engineering was required. The
"RARE prime" designation indicates that the deletion of two potentially inconsequential genes
(dkgB and yeaE) did not occur in this strain. Next, ppsA and tktA were overexpressed in the PTSglu' RARE' host, and protocatechuate titer increased by 50% to 300 mg/L (Fig. 5-2B). In addition,
the specific yield of protocatechuate noticeably increased around 24 hours after pathway
115
induction when ppsA and tktA were overexpressed. However, no change in the kinetics of
vanillate formation was observed, suggesting that the reaction catalyzed by the 0methyltransferase was limiting. In a parallel experiment, the PTS- glut RARE' host expressing the
pathway was cultured in a bioreactor and displayed an increasing difference in protocatechuate
and vanillate titers throughout the time course (Fig. 5-2C). In all of these experiments, additional
protocatechuate and vanillate was no longer produced after roughly 36-48 hours. To better
investigate this phenomenon, remaining experiments were performed at flask scale in biological
triplicate in order to increase experimental throughput.
116
Protocatechuate
I' Vanillate
-
500
RARE
aroG*, asbF, OMT
50
-
A
400-
300
300-
-
400-
0
200
-
P
e
100-
*
*
PTS- glu+ RARE'
aroG*, asb OMT
200-
0
12
24
36
48
C
*
C
48
60
72
84
0 I
60
72
84
96
0
12
24
36
Time (h)
B
C
100-
0
500
Protocatechuate
Vanillate
96
Time (h)
PTS- glu+ RARE'
*
*
aroG*, asbF OMT
500-
Protcxatechuate
Vanilate
400
PTS- glu+ RARE'
aroG*, asbF OMT ppsA, rktA
Protocatechuate
Vaniltate
*
e
400-
300
-
300
0
-.
200
-
200
-4
100
100-
0
0
0
48
60
72
0
0
12
24
36
48
60
72
12
C
24
Time (h)
36
Time (h)
0.5-
0.5
*
*
Protocatechuate
Vanillate
0.4-
0.3
03
-
0.4
0.2
*
*
Protocatechuate
Vanillate
0.2
-
0.1
0.1
*
0
*
C
48
60
0.0
0
12
24
36
48
60
72
12
24
Time (h)
36
72
Time (h)
C
500
-
-
500
*
400-
Protocatechuate
Vanillate
*
*
Protocatechuate
Vanillate
400-
PTS- glu+ RARE'
aroG*, asbF, OMT
300-
300-
200-
*
100
100-
-
0
200-
*
e
*
0
24
32
*
*
0
*
*
0
40
48
56
64
72
80
g
0
8
16
24
32
40
48
Time (h)
56
64
72
80
8
16
Time (h)
117
Figure 5 - 2. Effect of perturbations in central metabolism intended to increase PEP and E4P
availability on heterologous metabolite titers and specific yields. (A) Deletion of PTS- glu' did not
improve titers of either protocatechuate or vanillate. (B) Overexpression of ppsA and tktA in the
PTS- glu' RARE' host resulted in an increase in protocatechuate titer and specific yield compared
to expression of the pathway without ppsA and tktA. (C) Bioreactor culture of PTS- glut RARE'
host expressing the pathway (without ppsA and tktA overexpression) leads to increased
protocatechuate titers without a concomitant increase in vanillate titers, indicative of room for
improvement in the conversion of protocatechuate to vanillate. Host in blue text, overexpressed
genes in red text.
5.3.2. Understanding why conversion of protocatechuate to vanillate was limiting: SAM
To investigate whether a higher level of OMTHs would result in greater vanillate formation,
a second plasmid harboring the gene encoding OMTHs was initially introduced along with our
original pathway constructs and did not result in an improvement. Although the gene encoding
OMTHs was codon-optimized for expression in E. coli, we next wondered whether OMTHS may be
expressing poorly or whether it may have low activity in E. coli given its human origin. SDS-PAGE
results suggested that OMTHs expressed well, both when expressed on its own at flask scale and
when expressed in conjunction with the other pathway constructs in the bioreactor culture (Fig.
5-3A).
118
A
Ladder
No
OMT
OMT
No
OMT
OMT
i
24 h
B
Bioreactor Samples
32h
40 h
48 h
200
Control
Control +Met
_ merK overexpresslon
-Y- metK overexpression + Met
-U
-
kD
*
150
-
100
75
.5
E
so
10mM
r'
Met
-4
Addition
100
.1-
-
37
OMT
24.5 kD 25
-- 0 20
50-
0 4W,
Soluble
Fraction
insoluble
Fraction
-
15
10
0
Soluble
Fraction
12
24
48
36
60
72
Time (h)
C
HO
t
HrOt
HOe
Protocatechuate
Vanillate
SAM
SAH
H~o
-
Other Methylations,
Polyamine Synthesis,
Autoinducer-1 Synthesis
Adenim
Autoinducer-2 Synthesis
SRH
ATP
Protein Synthesis
- ---- Met
Hr si
THF
- ----
Cystelne and
Aspartate Biosynthesis
CHF-THP
D
0.0
.
-
250
SControl
a Control
Control + Hcys
Control + Hcys
200-
o .15-
E
0.10Stoo.
000.05*
so-
50-12
0.00
24
36
Time (h)
48
60
72
12
24
36
Time
48
80
72
(h)
Figure 5 - 3. Identification of the bottleneck in vanillate production. (A) SDS-PAGE result showing
robust expression of OMTHs. (B) Effect of 10 mM L-methionine supplementation at peak
productivity (24 h) on vanillate titers, with and without overexpression of metK. (C) Pathway
illustrating the reaction catalyzing conversion of protocatechuate into vanillate in the context of
SAM biosynthesis and recycling. (D) Effect of 2.5 mM L-homocysteine supplementation at peak
productivity (24 h) on vanillate titers and specific yield. In both pathway experiments shown here
(B and D), the PTS- glu* RARE' host overexpressing aroG*, ppsA, tktA, asbF, and OMT was tested.
119
Given that OMTHs seemed to be expressed in cells at times during which conversion of
protocatechuate into vanillate was not occurring, we next considered that co-factor availability
may be limiting. We monitored vanillate titers in cultures with and without supplementation of
10 mM L-methionine at peak vanillate productivity (24 hours after induction). L-Methionine is
endogenously converted to SAM by the E. coli methionine adenosyltransferase encoded by metK.
L-methionine addition may indirectly perturb SAM availability in vivo, whereas exogenously
supplied SAM does not enter E. coli. To simultaneously test whether conversion of L-methionine
to SAM was limiting, metK gene overexpression was investigated in the presence and absence of
L-methionine supplementation. Overexpression of metK was not required to see an
improvement in vanillate titer from L-methionine addition (Fig. 5-3B). Interestingly, although Lmethionine supplementation did not result in immediate changes in vanillate titer, final vanillate
titer was increased 2-fold, and the duration of vanillate synthesis was extended to the final
sampling time of 72 hours. This suggested that a limitation in SAM and methionine pools later in
the culture may have been responsible for the limited conversion of protocatechuate into
vanillate.
5.3.3. Investigating potential bottlenecks in SAM biosynthesis
To better understand the contributions of methionine biosynthesis to the vanillin
pathway, a second supplementation experiment was performed, this time with the direct
precursor to methionine, L-homocysteine. The methylation of L-homocysteine to form Lmethionine is reported to be problematic under oxidative conditions due to the inactivation of
catalytic residues in the cobalamin-independent methionine synthase (MetE) (188-191). Lhomocysteine is also an intermediate in the potential recycling pathway from SAM back to L120
methionine (Fig. 5-3C). As before, cultures at peak vanillate productivity (24 h) were
supplemented with and without L-homocysteine. Because L-homocysteine is reported to be toxic
for E. coli, we added 2.5 mM rather than 10 mM. Once again, supplemented cultures displayed
an increase in vanillate titer (to 200 mg/L) and an increase in duration of vanillate production
consistent with what was observed for L-methionine addition (Fig. 5-3D). The average rate of
vanillate synthesis doubled from 1.3 mg/L-h to 2.6 mg/L-h. This indicated to us that reactions
upstream of L-homocysteine synthesis in the methionine biosynthesis pathway needed to be
improved in order to achieve an increase in vanillate production from glucose as a sole carbon
source.
Unlike aromatic amino acid biosynthesis, methionine biosynthesis in E. coli and other
bacteria is intricately regulated and not well understood. As a result, until very recently,
methionine was the only essential amino acid that was not commercially produced using
fermentative processes (192, 193). In the academic literature, titers of 910 mg/L were achieved
using an E. coli strain that was constructed by mutagenesis with nitrosoguanidine along with
selection based on resistance to L-methionine-analogs. Among other potential mutations, this
strain had a mutation in the meU gene, a global regulator of methionine biosynthesis, that
rendered it inactive at repressing much of the pathway (194). In recent years, published patent
applications assigned to the French startup Metabolic Explorer reveal significant progress in Lmethionine overproduction, with titers in an engineered E. coli strain reaching upwards of 30 g/L
from a fed-batch process (193, 195). Though sparse in details, these disclosures are encouraging
because they establish the potential for improvement of methionine biosynthesis in E. coli. In
particular, patent literature related to the proprietary Metabolic Explorer strain suggests several
121
gene targets that may be important, and among those are met, metA, and cysE. MetA
(homoserine succinyltransferase) catalyzes the first committed step in methionine biosynthesis
and is reported to be inhibited by both L-methionine and SAM (196). CysE (L-serine 0acetyltransferase) catalyzes the first step of L-cysteine biosynthesis and is reported to display
significant inhibition by L-cysteine (197). Fortunately, academic literature describes specific
variants of MetA and CysE that have been engineered to display desensitization to feedback
inhibition (MetA* and CysE*) (196, 197). Other gene targets highlighted in the Metabolic Explorer
patents include metE, metF, and glyA. However, these steps are all downstream of Lhomocysteine formation, and the L-homocysteine supplementation results obtained in this study
suggest that endogenous processes convert L-homocysteine into SAM at a rate sufficient to
improve vanillate production.
The relevance of known L-methionine overproducers to our study is complicated by the
fact that a greater pool size of SAM, not L-methionine itself, is ultimately needed to improve
vanillate formation. This distinction is not trivial as SAM fulfills a variety of cellular roles (e.g., the
primary methyl donor for all cellular methylations) that compete directly with the engineered
vanillin pathway. To our knowledge, there is only one other report of metabolic engineering in E.
coli involving a methyltransferase reaction that relies on SAM (198). In that report, a novel
bacterial fatty acid methyltransferase is used to catalyze the formation of fatty acid methyl esters
using free fatty acids and SAM. The authors of that study note that SAM availability strongly
regulates methyl ester production. By deleting the metU gene mentioned earlier, and by
overexpressing a gene encoding methionine adenosyltransferase from rat, the authors achieved
an improvement in methyl ester production. However, the normalized titers of methyl esters in
122
the supernatant achieved in their study increased from below 1 pM/OD to roughly 2.5 pM/OD.
The corresponding amount of SAM required to achieve such conversion is orders of magnitude
below what drives vanillate production, and thus this study ventures into uncharted territory in
the extent to which it seeks to improve SAM availability.
5.3.4. Improving vanillate production by deregulating SAM biosynthesis
Given our ultimate goal of improving vanillin production from glucose as a sole carbon
source, we were curious to know whether we could achieve an improvement in vanillate
production by modifying methionine biosynthesis in three ways: (i) deleting meU; (ii)
overexpressing metA*; and, (iii) overexpressing cysE*. If these three modifications did not result
in an improvement in vanillate production, then we would likely redefine our objectives to
include methionine supplementation because of the potential recalcitrance of the methionine
biosynthetic pathway to improvements. When meti was deleted in both the PTS- glut RARE' and
RARE host strains, a slight decrease in protocatechuate and vanillate titers was observed (Fig. 54A). However, as mentioned before, meU is only one of many simultaneous modes of methionine
biosynthesis regulation. Given feedback-resistance at the entrance to the pathway, one could
envision the meti deletion strain performing slightly worse because of increased expression of
downstream genes with minimal flux entering the pathway. From here on, given that the RARE
h-ost again resulted in higher titers than the PTS- glu' variant, we decided to continue with the
RARE AmetJ host.
To further understand the limitations in methionine biosynthesis, we next supplemented
the RARE Ameti host harboring the pathway with three different amino acids: L-methionine, L-
123
cysteine, and L-aspartate. L-cysteine and L-aspartate are both precursors to methionine, and our
goal was to investigate whether reaction steps downstream of their biosynthesis were
problematic. In this case, we added 10 mM of each amino acid at the time of induction (0 h) to
see whether the time of supplementation would affect pathway kinetics. In this experiment, we
also included the RARE host (with meti intact) expressing the pathway, though results should not
be compared to the previous L-methionine supplementation experiment (where the PTS~ glu'
RARE' host was used instead). Performance of the RARE and RARE Amet hosts supplemented
with L-methionine was similar, with roughly 280 mg/L vanillate produced in just 24 hours (Fig. 54B). However, little additional vanillate formed after the first 24 hours. The decrease in the rate
of vanillate formation suggested that all of the L-methionine added initially had been depleted
within 24 hours. Addition of L-cysteine or L-aspartate did not improve vanillate titers, which
supported the notion of next focusing on the reactions catalyzed by MetA and CysE. Furthermore,
addition of 10 mM L-cysteine significantly decreased titers and biomass formation.
A
B
500 -500Protocatechuate
Vanillate
400
400
300
300
S200
20
-C .2E
100
Protocatechuate, 24 h
Vanillate, 24 h
Protocatechuate, 48 h
Vanillate, 48 h
100-
0
0o-011
RARE
metL::kanR
PTS- glu+
met::kanR
RARE
+ Met
RARE AmeU
+ Met
RARE AmetJ
+ Cys
RARE AmeU
+ Asp
Figure 5 - 4. Effect of metJ deletion (A) in different host strains and (B) in the presence of amino
acid supplementation. For these experiments, the following genes were overexpressed: aroG*,
ppsA, tktA, asbF, and OMT. For the amino acid supplementation experiment (B), 10 mM of amino
acid was added at induction.
124
We next cloned the metA* and cysE* genes into a separate plasmid (pCOLADuet-1) to
determine whether overexpression of these genes would improve conversion of vanillate to
protocatechuate (Figure 5-5A). Like previous experiments, this would provide indirect insight into
whether SAM availability had increased. The performance of a control strain harboring typical
vanillate pathway constructs and an empty pCOLADuet-1 plasmid was consistent with previous
results in both the magnitude of titers achieved from glucose as well as the greater mass
concentration of protocatechuate relative to vanillate after the first 24 hours of the reaction.
Interestingly, in the cases in which metA* and cysE* were overexpressed separately or together,
the vanillate titer now exceeded the protocatechuate titer at all time points sampled. Although
metA*-cysE* overexpression had registered a measureable effect, the increase in final vanillate
titer compared to the control was less than 100 mg/L. In order to gain insight into the potential
upper bound of vanillate titer obtainable simply by increasing flux through methionine
biosynthesis, we took cultures overexpressing metA*-cysE* and supplemented them with 10 mM
L-methionine 24 hours after induction or at induction and 24 hours after induction (Fig. 5-5B). As
expected, in all of these cases vanillate titers continued to exceed protocatechuate titers at all
time points sampled. However, in both cases addition of methionine at 24 hours only led to
marginal (< 100 mg/L) increases in vanillate titers during the remainder of the experiment. In
contrast, addition of methionine at induction led to nearly 350 mg/L of vanillate produced in the
first 24 hours. This suggested that further increased SAM availability within the first 24 hours
could still improve vanillate production. The results also suggested that something else may be
limiting after 24 hours.
125
A
B
*
400-
2
500-
Protocatechuate, 24 h
Vanillate, 24 h
Protocatechuate, 48 h
Vanillate, 48 h
Protocatechuate, 72 h
Vanillate, 72 h
300-
I
Protocatechuate, 24 h
VaWn1ate, 24 h
Protocatechuate, 48 h
VanIllate, 48 h
Protocatechuate. 72 h
Vaniliate, 72 h
400-
a
I
T
300
-
500-
200-
200-
100-
100-
0-
0Control
metA*
cysE*
0 h and 24 h
24 h
metA*-cysE*
Time of 10 mM Met addition
D
500450
0
400-
V
-
C
500-
Protocatechuate
Vanillate
Protocatechualdehyde
Vanillin
450400V
350
-
3502
250
CA
-
300-
200
-
*
-
iv
100
/
0
~
V
0
50
A
A
t
12
-Y-------
iv
a
A
507
250
1501
a
100-
300
200
U
150-
0,
Protocatechuate
Vanillate
Protocatechualdehyde
Vanillin
-F
24
.
36
Time (h)
-P
48
- --
0-
. jp
60
72
0
12
24
36
48
60
72
Time (h)
Figure 5 - 5. Effect of metA* and cysE* overexpression. (A) Effect of overexpressing feedbackdesensitized variants of metA and/or cysE along with usual pathway constructs in the RARE AmeU
host. The control represents co-transformation with an empty pCOLADuet-1 plasmid. (B) Effect
of methionine supplementation level and timing on vanillate titers in metA*-cysE* cultures. (C)
Kinetics of vanillin production without overexpression of metA*-cysE*. (D) Kinetics of vanillin
production with overexpression of metA*-cysE*. Although final titers achieved in (C) and (D) are
similar, the metA*-cysE* cultures grow more slowly, produce vanillin more slowly, but display
greater conversion of protocatechuate to vanillate.
To understand the effect of metA*-cysE* overexpression in the context of the entire
vanillin pathway, we transformed the strain with another plasmid (either empty pCDFDuet-1 or
pCDF-Car-Sfp). By including the reaction catalyzed by Car in the pathway, we enable generation
126
of both vanillin and protocatechualdehyde. We monitored titers of these two additional
heterologous metabolites as well as vanillate and protocatechuate during the course of the
reaction (Figs. 5-5C and 5-5D). Overexpression of metA*-cysE* improved the kinetics of
protocatechuate methylation as evidenced by the improved ratio of vanillate to protocatechuate
during the first half of the reaction. However, final vanillin titers remained essentially the same
for both strains.
Given continued observations of slowing protocatechuate conversion after 24 hours, we
determined that potential loss of the ampicillin-resistant plasmid harboring OMT (pET-AsbFOMT) was worth investigating. From the previously described vanillin production experiment, we
sampled 1 mL of cells from cultures overexpressing metA*-cysE* at 24, 48, and 72 hours. Each
sample was serially diluted in sterile deionized water to the range of 104-107 fold and then
immediately plated by pipetting 10 uL droplets on plates containing LB or. LB and 100 mg/L
carbenicillin. Carbenicillin was used instead of ampicillin for a more stringent selection. Based on
the similar number of colonies appearing on plates taken at the same time point (Fig. 5-6), we
could rule out plasmid loss as an explanation for the limited conversion observed.
127
Figure 5 - 6. Images of plates testing for potential loss of ampicillin-resistant plasmid. No
significant plasmid loss was observed for samples taken at 24, 48, and 72 h.
128
5.4. Discussion
This chapter has presented a detailed perspective on de novo vanillin biosynthesis in
engineered E. coli. First, genetic perturbations were made upstream of the heterologous pathway
in order to improve titers and better determine which reaction steps may be limiting. Upon
identification of a limitation in the reaction catalyzed by the O-methyltransferase,
supplementation experiments were performed and provided indirect evidence that availability
of the co-factor S-adenosylmethionine was responsible for the limitation. Given that endogenous
conversion of L-homocysteine appeared to be robust, attention was next focused on deregulating
methionine biosynthesis by targeting the global regulator MetJ and the feedback-sensitive
enzymes MetA and CysE. When deletion of metU was coupled with expression of feedbackdesensitized variants of either MetA or CysE (MetA* and CysE*, respectively), modest
improvement in the conversion of vanillate to protocatechuate was observed. However, final
vanillin titers were not significantly improved with overexpression of metA*-cysE*, indicating
that there are remaining issues with this pathway. Given the status of this project as both ongoing
and future work, the remainder of this discussion will delve into hypotheses that warrant testing
in the near future. Chapter 6 will take a longer-term view of future opportunities with the vanillin
pathway and other work presented in this thesis.
Although much of the work presented in Chapter 5 focused on the relationship between
SAM availability and protocatechuate conversion, little discussion has been devoted thus far to
what happens to SAM after its consumption by the OMT. S-adenosylhomocysteine (SAH) is a coproduct of the reaction along with vanillate, and SAH is a potent inhibitor of SAM-dependent
methyltransferases (199, 200). The L-homocysteine supplementation experiment suggested that
129
recycling of SAH back to L-homocysteine is not occurring at a rate sufficient to maintain SAM
concentrations. In addition, slow or limited recycling of SAH is likely to inhibit OMT activity. This
may be especially relevant in an E. coli host because one of the two reactions required to recycle
SAH is catalyzed by LuxS and coupled to the production of autoinducer AI-2, which is a quorum
sensing molecule (201, 202). In E. coil and other bacteria, AI-2 has been implicated as an inducer
of biofilm formation (203, 204) and as an attractant for chemotaxis (205). Because recycling of
SAH is committed to the formation of AI-2, this pathway (also known as the activated methyl
cycle) is expected to be subject to immense regulation.
In eukaryotes, archaea, and non-LuxS-containing bacteria, SAH hydrolase (SAHase)
converts SAH to homocysteine and adenosine (202). It is possible that the overexpression of a
heterologous SAHase may improve protocatechuate conversion both by decreasing the pool size
of the inhibitory byproduct of the reaction and by increasing the pool size of SAM (Fig. 5-7).
Furthermore, use of a SAHase may mitigate whatever global regulatory effects may be occurring
due to elevated concentrations of AI-2. Candidate SAHases include one from Corynebacterium
efficiens, which has been expressed previously in E. coli (206). In addition to expressing a
heterologous SAHase, it may be important to reduce or delete the background rate of SAH
recycling in E. coli. As shown in Figure 5-7, the two genes responsible for these reactions are mtn
and luxS. Neither of these genes are essential, and a recent LC-MS study profiles the effect of
their individual deletion on intracellular concentrations of SAM, SAH, SRH, Hcys, and Met at
different OD6 00values for E. coli MG1655 (207).
130
N
N
H
%
QMTN
OMT
N
H2N
N'N
H%.~
O
H
N
2
0
N
N
0
OH
NH 2
NH
SAM
2
SAH
Pi + PPi
NAD
metK
-H20
mtn
ATP
0
Adenine
sahH
Adenosine
+ NADH-
S
"0
OH
OH
NH 2
HO
NH 2
Met
O
2
HO
metE/metH
IuxS
SRH
0
THF
CH 3-THF
OH
DHPD
NHH
Hcys
AI-2
Figure 5 - 7. The activated methyl cycle in E. coli (in black), along with an alternative SAH recycling
route featuring a heterologous SAH hydrolase (sahH, in blue).
Overall, based on results obtained for vanillin pathway experiments thus far, further
investigation of the SAH recycling pathway is justified. If increased SAH recycling and decoupling
to AI-2 biosynthesis were to improve vanillate formation, then the potential utility of this result
would be broader than merely vanillin production. SAM-dependent methyltransferases
constitute a broad class of enzymes that could find use in numerous engineered metabolic
pathways, and thus we hope to have exciting results to share soon after completion of this thesis.
131
Chapter 6: Lessons Learned and Future Directions
6.1. Summary
The field of metabolic engineering has successfully enabled production of a variety of
chemical classes using engineered microbes that utilize renewable inputs rather than petroleum
inputs. Aldehydes represent one valuable class of chemicals that had been elusive for metabolic
engineers not because of a lack of known aldehyde biosynthetic pathways but rather because of
the rapid reduction of aldehydes into their corresponding alcohols catalyzed by microbial hosts.
At the time that the work presented in this thesis began, the potential number of endogenous
genes encoding aldehyde reductases and the perceived magnitude of aldehyde toxicity made the
concept of microbial engineering for the synthesis of any one aldehyde initially seem daunting.
While few aspects of the research outlined in this thesis progressed as envisioned, we are pleased
to report that we did in fact succeed in determining and sharing a number of original findings
related to aldehyde biosynthesis in vivo and in vitro. Furthermore, our results relate to
applications across industries spanning flavors, pharmaceuticals, and fuels. Summarized below
are key findings from this investigation.
In Chapter 2, we investigated the contributions of members of two enzyme superfamilies,
aldo-keto reductases and alcohol dehydrogenases, towards endogenous aromatic aldehyde
reductase activity. Upon serial deletion of six targeted genes, we observed a marked decrease in
endogenous reduction of our model aromatic aldehyde, benzaldehyde. This important result
allowed us to then characterize the contributions of individual genes when overexpressed in the
engineered host strain. One of the most valuable original findings from this thesis was our
132
demonstration that individual overexpression of aldehyde reductases would lead to false
positives, which was supported by complementary analyses that included deletion subset strains
and qRT-PCR. These results showed that it would not be necessary to delete every gene encoding
an enzyme capable of aldehyde reductase activity in order to dramatically decrease endogenous
aldehyde reductase activity while maintaining wild-type cell growth rate. It was fascinating to
discover that E. coli had evolved several redundant enzymes that all appeared to be unnecessary
under lab conditions. After characterizing the engineered "RARE" host strain, we went on to
demonstrate its utility for de novo vanillin biosynthesis and for production of L-PAC from glucose
and exogenously supplied benzaldehyde. Another significant finding occurred when we observed
that the wild-type host strain was unable to produce any L-PAC and subsequently investigated
the kinetics of benzaldehyde reduction. Although we expected a high level of wild-type
endogenous benzaldehyde reduction, the magnitude of this activity (leading to full conversion of
5 mM benzaldehyde within 2 hours) exceeded our expectations and was on an entirely different
timescale than our previous daily measurements. Of course, it also explained why a metabolically
active wild-type E. coli host strain was unable to produce L-PAC, whereas the RARE host strain
could.
In Chapter 3, we sought to learn more about the potential synthesis and accumulation of
non-aromatic aldehydes. We were initially encouraged by the activity of CarNi on a number of
aliphatic acid substrates ranging in carbon chain length from C3 to C8. We next used carbon chain
extension pathways that utilized either fatty acid synthesis or reverse 1-oxidation (or CoAdependent extension) to demonstrate the synthesis of free fatty acids of a variety of chain
lengths from glucose. By coupling CarNito these pathways, and by using the RARE host strain, we
133
were also able to successfully produce and accumulate aliphatic aldehydes. Although the
remaining endogenous reductase activity for these compounds was higher than for aromatic
aldehydes, it was intriguing to discover that the deletion of a small number of genes encoding
enzymes known to act on benzaldehyde could result in significantly decreased reductase activity
across such a broad range of aldehydes. Our ultimate objective in this study was to take
advantage of decreased alcohol formation by instead producing alkanes, which are exact
constituents in gasoline. Interestingly, although we observed elevated concentrations of the
aldehyde intermediates, the limited activity of the aldehyde decarbonylase led to little
improvement in alkane titers using the RARE host strain. However, if aldehyde decarbonylases
with improved kinetics are identified or engineered, then we expect the decreased endogenous
alcohol byproduct formation that we achieved to contribute to higher alkane titers.
In Chapter 4, we sought a greater understanding of in vitro aldehyde biosynthesis as a
potential alternative to microbial aldehyde biosynthesis given that it circumvents the issue of
microbial aldehyde toxicity. In addition, we wanted to better understand the kinetics of CarNi,
which initially appeared to have in vitro activity for only a limited time. Our observation of
pyrophosphate precipitation led us to form the hypothesis that pyrophosphatase addition might
improve CarNi performance. Indeed, Ppa addition improved CarNi-catalyzed conversions more
than two-fold under conditions tested. Perhaps more importantly, Ppa addition to an in vitro
pathway featuring CarNi enabled accurate modeling of reaction kinetics on the timescale of hours
based simply on initial rate measurements. Although straightforward in hindsight and when
considering standard practices in preparative in vitro transcription, the coupling of Ppa to
134
carboxylic acid reductases for enhanced in vitro aldehyde biosynthesis was an original insight that
explains the previously observed discrepancy in CarNi performance in vitro versus in vivo.
In Chapter 5, we revisited the de novo vanillin biosynthesis pathway in E. coli with the goal
of improving pathway performance. Given that titers of all heterologous metabolite titers were
low when we began, we initially focused on improving flux from central carbon metabolism into
aromatic amino acid biosynthesis using previously documented genetic perturbations. We
quickly observed an accumulation of the first heterologous metabolite, protocatechuate. Further
investigation using supplementation experiments provided indirect evidence that the conversion
of protocatechuate to vanillate was limited by availability of S-adenosylmethionine (SAM). SAM
biosynthesis is heavily regulated in E. coli, but we attempted to deregulate the pathway by
deleting the gene encoding the global regulator MetJ and by expressing feedback-desensitized
variants of MetA and CysE, which both catalyze the formation of SAM precursors. When we
deleted meti and overexpressed metA* or cysE*, we observed improvements in the conversion
of protocatechuate to vanillate. However, these modifications did not lead to appreciable gains
in vanillin titer, and it may be because of additional challenges in the regulation of these
pathways. In particular, Chapter 5 discusses the role that the co-product of the reaction (Sadenosylhomocysteine) may have on the O-methyltransferase-catalyzed reaction as well as on
general cell physiology. This remains the subject of active investigation.
6.2. Future Directions
Based on the findings presented in this thesis, there are at least two distinct research
areas that may benefit from further investigation. These are the use of metabolite sensors to
135
improve the vanillin pathway and the expansion of CoA-dependent carbon chain extension
pathways to generate novel aliphatic aldehydes or to use them as intermediates. These two areas
will be discussed separately, but more attention will be directed towards the metabolite sensors
portion given that it is more thoroughly conceived and that some preliminary efforts have already
been taken towards this aim.
6.2.1. Metabolite sensors for the vanillin pathway
The results from Chapter 5 indicated that the vanillin pathway suffers from a major
limitation in the conversion of protocatechuate to vanillate. Once this limitation were to be
overcome, then metabolite sensors could be considered to further improve pathway
performance. In the field of metabolic engineering, several flagship studies have established that
pathways can be improved using biosensors that implement dynamic control of key metabolic
intermediates (208-211). In addition, sensors can enable evolution-guided optimization of entire
pathways (212). The challenge encountered in most engineered pathways is the lack of natural
transcription factors that respond specifically to any one metabolite, let alone responding to the
metabolites that matter (i.e., key rate-limiting intermediates or the product). In the case of the
vanillin pathway, a vanillate-inducible system exists naturally in Caulobacter crescentus (213).
Furthermore, this repressor-operator system (VanR-VanO) has been extended to control gene
expression in organisms ranging from Myxococcus (214), Sphingomonas (215), and even mice
(216). In the mice study, it was shown that VanR was unresponsive to 16 closely related
compounds to vanillate. By turning on expression of CarNi specifically in response to a threshold
concentration of vanillate, we could potentially delay the formation of a toxic product and limit
formation of the toxic intermediate protocatechualdehyde.
136
The notion that C. crescentus had evolved a vanillate-inducible consumption pathway
because vanillate is a natural lignin degradation product led us to hypothesize that a natural
system might also exist for sensing protocatechuate, which is another lignin degradation product.
Given the inherent delay of gene expression in response to transcription factors, we thought that
it was worth investigating whether delayed CarNi expression could instead be activated by a
threshold protocatechuate concentration. We found that Acinetobacter and related bacteria
indeed contain protocatechuate responsive gene expression systems (217, 218). Surprisingly, a
report published in 2014 already demonstrated the development of a synthetic protocatechuate
sensor in E. coli using the protocatechuate responsive transcriptional activator PcaU and the
PcaU-PcaI intergenic sequence (219). In the same study, the protocatechuate sensor was further
optimized by generating libraries of intergenic sequences and selecting for improved induction
ratios using FACS. It was also shown to be unresponsive to vanillate and other related
compounds, though a high concentration of vanillin (10 mM) did seem to cause a response.
Overall, these sensors for two key vanillin pathway intermediates may serve as valuable tools for
dynamically regulating and/or evolving the vanillin pathway.
6.2.2. Use of CoA-dependent pathways to generate novel aliphatic aldehydes
Among the many things I learned while working with Dr. Micah Sheppard was greater
knowledge of the CoA pathway platform developed in the Prather Lab. In conjunction with Car
expression and the RARE host strain, this platform enables the production of aliphatic aldehydes
across a variety of carbon chain lengths ranging from C3 to at least C8. The volatilities of several
of these aldehydes are so high that pure aldehydes could be stripped out from fermentation
much more readily than corresponding carboxylic acids or alcohols. In fact, it is likely that some
137
aldehyde loss is observed in aerobic cultures even without intentional stripping. As Chapter 3
outlined, these carbon chain lengths are in the range of gasoline constituents; however,
biosynthesis of fuels using aldehyde decarbonylases is currently limited by poor enzyme kinetics.
If fuels are the goal, whether they are alcohols, alkanes, or olefins, it may make sense to take
advantage of the ease of separation of aldehydes from the fermentation of renewable substrates
and to then reduce them using inorganic catalysis.
Perhaps more interesting than generating reduced products from aliphatic aldehydes is
expanding the portfolio of potential microbial aldehyde products. The CoA pathway contains
several acyl-CoA intermediates, and previous work in the Prather Lab has sought to identify
thioesterases that may be more selective for acyl-CoAs that are not saturated (220). Similarly,
other work in the Prather Lab has focused on developing novel 3-hydroxyacids using the CoA
pathway (221). This past body of work suggests that the Prather Lab would be well poised to
potentially produce diverse novel aldehydes ranging from unsaturated
aldehydes to
hydroxyaldehydes at several carbon chain lengths. From a technical perspective, remaining
hurdles to achieving this include identification of appropriate thioesterases, determination of
carboxylic acid reductase activity on these free acids, and detection of the products and key
intermediates. Beyond technical issues, there is also the question of whether any of these
aldehyde targets are worthwhile from an academic or commercial perspective.
6.2.3. Microbial aldehyde toxicity
One other remaining area of future work that represents the most logical next step from
this thesis is microbial aldehyde toxicity. This problem appears to be interesting from both an
138
academic and industrial perspective given that aldehyde toxicity mechanisms are not wellelucidated and yet aldehyde products are being produced commercially via fermentation. There
are potential opportunities to employ strategies at both the process and cellular levels, and
synthetic biology advances may enable some novel solutions for the latter. However, as
suggested in Chapter 1, what is needed first is a better understanding of cell physiology. For
example, what specific cellular processes or components are negatively interacting with aldehyde
molecules? Is an indirect effect such as a stress response primarily responsible for the decrease
in growth rate? How does the phase of cell growth influence susceptibility to aldehydes? These
types of questions may be better suited for investigators with a microbiology background rather
than a chemical engineering background. Fortunately, the Prather Lab has had graduate students
and post-docs from both disciplines, which increases the odds that we could successfully
engineer solutions to microbial aldehyde toxicity in the future. Ultimately, this would enable
higher titers of microbial aldehyde products.
139
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
M. P. Crosland, Historical Studies in the Language of Chemistry. (Dover Publications, Inc.,
Mineola, New York, 2004).
H. Zhao et al., Functional Expression of a Mammalian Odorant Receptor. Science 279, 237-242
(1998).
R. C. Araneda, A. D. Kini, S. Firestein, The molecular receptive range of an odorant receptor. Nat
Neurosci 3, 1248-1255 (2000).
R. C. Araneda, Z. Peterlin, X. Zhang, A. Chesler, S. Firestein, A pharmacological profile of the
aldehyde receptor repertoire in rat olfactory epithelium. The Journal of Physiology 555, 743-756
(2004).
E. R. Soucy, D. F. Albeanu, A. L. Fantana, V. N. Murthy, M. Meister, Precision and diversity in an
odor map on the olfactory bulb. Nat Neurosci 12, 210-220 (2009).
K.-G. Fahlbusch et al., in Ullmann's Encyclopedia of Industrial Chemistry. (Wiley-VCH Verlag
GmbH & Co. KGaA, 2000).
S. Hagedorn, B. Kaphammer, Microbial Biocatalysis in the Generation of Flavor and Fragrance
Chemicals. Annual Review of Microbiology 48, 773-800 (1994).
P. A. Tarantilis, M. G. Polissiou, Isolation and Identification of the Aroma Components from
Saffron (Crocus sativus). Journal of Agricultural and Food Chemistry 45, 459-462 (1997).
A. K. Raina, T. G. Kingan, A. K. Mattoo, Chemical Signals from Host Plant and Sexual Behavior in a
Moth. Science 255, 592-594 (1992).
J. C. Dickens, E. B. Jang, D. M. Light, A. R. Alford, Enhancement of insect pheromone responses
by green leaf volatiles. Naturwissenschaften 77, 29-31 (1990).
Z. Syed, W. S. Leal, Acute olfactory response of Culex mosquitoes to a human- and bird-derived
att ractant. Proceedings of the National Academy of Sciences 106, 18803-18808 (2009).
C. M. Tripathi, S. C. Agarwal, S. K. Basu, Production of I-phenylacetylcarbinol by fermentation.
Journal of Fermentation and Bioengineering 84,487-492 (1997).
V. B. Shukla, P. R. Kulkarni, L-Phenylacetylcarbinol (L-PAC): biosynthesis and industrial
applications. World Journal of Microbiology and Biotechnology 16:6,499-506 (2007).
H. Yun, B.-G. Kim, Enzymatic production of (R)-phenylacetylcarbinol by pyruvate decarboxylase
from Zymomonas mobilis. Biotechnology and Bioprocess Engineering 13, 372-376 (2008).
D. Meyer et al., Conversion of Pyruvate Decarboxylase into an Enantioselective Carboligase with
Biosynthetic Potential. J. Am. Chem. Soc. 133, 3609-3616 (2011).
E. C. Hayden, Synthetic-biology firms shift focus. Nature 505, 598 (2014).
1. Jeio, J. Zemek, Enzymatische reduktion einiger aromatischer carboxysauren. Chem. Pap 40,
279-281 (1986).
N. Kato, H. Konishi, K. Uda, M. Shimao, C. Sakazawa, Microbial Reduction of Benzoate to Benzyl
Alcohol. Agricultural and Biological Chemistry 52, 1885-1886 (1988).
J. Casey, R. Dobb, Microbial routes to aromatic aldehydes. Enzyme and Microbial Technology 14,
739-747 (1992).
H.-A. Arfmann, W.-R. Abraham, Microbial reduction of aromatic carboxylic acids. Z. Naturforsch.
Teil C 48, 52-57 (1993).
T. Li, J. P. Rosazza, Purification, characterization, and properties of an aryl aldehyde
oxidoreductase from Nocardia sp. strain NRRL 5646. The Journal of Bacteriology 179, 3482-3487
(1997).
140
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
A. He, T. Li, L. Daniels, I. Fotheringham, J. P. N. Rosazza, Nocardia sp. Carboxylic Acid Reductase:
Cloning, Expression, and Characterization of a New Aldehyde Oxidoreductase Family. Applied
and Environmental Microbiology 70, 1874-1881 (2004).
P. Venkitasubramanian, L. Daniels, J. P. N. Rosazza, Reduction of Carboxylic Acids by Nocardia
Aldehyde Oxidoreductase Requires a Phosphopantetheinylated Enzyme. Journal of Biological
Chemistry 282,478-485 (2007).
P. Venkitasubramanian, L. Daniels, S. Das, A. S. Lamm, J. P. N. Rosazza, Aldehyde oxidoreductase
as a biocatalyst: Reductions of vanillic acid. Enzyme and Microbial Technology 42, 130-137
(2008).
M. K. Akhtar, N. J. Turner, P. R. Jones, Carboxylic acid reductase is a versatile enzyme for the
conversion of fatty acids into fuels and chemical commodities. Proc. Nat. Acad. Sci. U.S.A. 110,
87-92 (2013).
K. Napora-Wijata, G. A. Strohmeier, M. Winkler, Biocatalytic reduction of carboxylic acids.
Biotechnology Journal 9, 822-843 (2014).
D. P. Clark, The fermentation pathways of Escherichia coli. FEMS Microbiology Letters 63, 223234 (1989).
P. E. Goodlove, P. R. Cunningham, J. Parker, D. P. Clark, Cloning and sequence analysis of the
fermentative alcohol-dehydrogenase-encoding gene of Escherichia coli. Gene 85, 209-214
(1989).
N. R. Palosaari, P. Rogers, Purification and properties of the inducible coenzyme A-linked
butyraldehyde dehydrogenase from Clostridium acetobutylicum. Journal of Bacteriology 170,
2971-2976 (1988).
R. V. Nair, G. N. Bennett, E. T. Papoutsakis, Molecular characterization of an aldehyde/alcohol
dehydrogenase gene from Clostridium acetobutylicum ATCC 824. Journal of Bacteriology 176,
871-885 (1994).
J. Toth, A. A. Ismaiel, J.-S. Chen, The aid Gene, Encoding a Coenzyme A-Acylating Aldehyde
Dehydrogenase, Distinguishes Clostridium beijerinckii and Two Other Solvent-Producing
Clostridia fromClostridium acetobutylicum. Applied and Environmental Microbiology 65, 49734980 (1999).
L. Fontaine et al., Molecular Characterization and Transcriptional Analysis of adhE2, the Gene
Encoding the NADH-Dependent Aldehyde/Alcohol Dehydrogenase Responsible for Butanol
Production in Alcohologenic Cultures of Clostridium acetobutylicum ATCC 824. Journal of
Bacteriology 184, 821-830 (2002).
E. A. Meighen, Molecular biology of bacterial bioluminescence. Microbiological Reviews 55, 123142 (1991).
S. Atsumi, T. Hanai, J. C. Liao, Non-fermentative pathways for synthesis of branched-chain higher
alcohols as biofuels. Nature 451, 86-89 (2008).
F. Kaehne, M. Buchhaupt, J. Schrader, A recombinant a-dioxygenase from rice to produce fatty
aldehydes using E. coli. Applied Microbiology and Biotechnology 90,989-995 (2011).
R. Gandolfi, N. Ferrara, F. Molinari, An easy and efficient method for the production of
carboxylic acids and aldehydes by microbial oxidation of primary alcohols. Tetrahedron Letters
42, 513-514 (2001).
D. Romano, R. Villa, F. Molinari, Preparative Biotransformations: Oxidation of Alcohols.
ChemCatChem 4, 739-749 (2012).
V. C. Corberen, M. E. Gonz lez-Perez, S. Martfnez-Gonzdlez, A. G6mez-Aviles, Green oxidation of
fatty alcohols: Challenges and opportunities. Applied Catalysis A: General 474, 211-223 (2014).
141
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
S. J. B. Duff, W. D. Murray, Oxidation of benzyl alcohol by whole cells of Pichia pastoris and by
alcohol oxidase in aqueous and nonaqueous reaction media. Biotechnology and Bioengineering
34, 153-159 (1989).
K. Li, J. W. Frost, Synthesis of Vanillin from Glucose. Journal of the American Chemical Society
120, 10545-10546 (1998).
E. H. Hansen et al., De Novo Biosynthesis of Vanillin in Fission Yeast (Schizosaccharomyces
pombe) and Baker's Yeast (Saccharomyces cerevisiae). Applied and Environmental Microbiology
75, 2765-2774 (2009).
A. Brochado et al., Improved vanillin production in baker's yeast through in silico design.
Microbial Cell Factories 9, 84 (2010).
A. R. Brochado, K. R. Patil, Overexpression of O-methyltransferase leads to improved vanillin
production in baker's yeast only when complemented with model-guided network engineering.
Biotechnology and Bioengineering 110, 656-659 (2013).
U. Krings, R. G. Berger, Biotechnological production of flavours and fragrances. Applied
Microbiology and Biotechnology 49:1, 1-8 (1998).
M. N. Nierop Groot, J. A. M. de Bont, Conversion of Phenylalanine to Benzaldehyde Initiated by
an Aminotransferase in Lactobacillus plantarum. Applied and Environmental Microbiology 64,
3009-3013 (1998).
N. Dudareva, A. Klempien, J. K. Muhlemann, I. Kaplan, Biosynthesis, function and metabolic
engineering of plant volatile organic compounds. New Phytologist 198, 16-32 (2013).
A. V. Qualley, J. R. Widhalm, F. Adebesin, C. M. Kish, N. Dudareva, Completion of the core 1oxidative pathway of benzoic acid biosynthesis in plants. Proceedings of the National Academy
of Sciences 109, 16383-16388 (2012).
E. J. Steen et al., Microbial production of fatty-acid-derived fuels and chemicals from plant
biomass. Nature 463, 559-562 (2010).
C. Dellomonaco, C. Rivera, P. Campbell, R. Gonzalez, Engineered Respiro-Fermentative
Metabolism for the Production of Biofuels and Biochemicals from Fatty Acid-Rich Feedstocks.
Applied and Environmental Microbiology 76, 5067-5078 (2010).
C. Dellomonaco, J. M. Clomburg, E. N. Miller, R. Gonzalez, Engineered reversal of the 0-oxidation
cycle for the synthesis of fuels and chemicals. Nature 476, 355-359 (2011).
P. Handke, S. A. Lynch, R. T. Gill, Application and engineering of fatty acid biosynthesis in
Escherichia coli for advanced fuels and chemicals. Metabolic Engineering 13, 28-37 (2011).
V. J. J. Martin, D. J. Pitera, S. T. Withers, J. D. Newman, J. D. Keasling, Engineering a mevalonate
pathway in Escherichia coli for production of terpenoids. Nat Biotech 21, 796-802 (2003).
G. Rodriguez, S. Atsumi, Isobutyraldehyde production from Escherichia coli by removing
aldehyde reductase activity. Microbial Cell Factories 11, 90 (2012).
M. Pohl, B. Lingen, M. Muller, Thiamin-Diphosphate-Dependent Enzymes: New Aspects of
Asymmetric C-C Bond Formation. Chemistry -A European Journal 8, 5288-5295 (2002).
A. Schirmer, M. A. Rude, X. Li, E. Popova, S. B. del Cardayre, Microbial Biosynthesis of Alkanes.
Science 329, 559-562 (2010).
C. Andre, S. W. Kim, X.-H. Yu, J. Shanklin, Fusing catalase to an alkane-producing enzyme
maintains enzymatic activity by converting the inhibitory byproduct H202 to the cosubstrate
02. Proc. Nat. Acad. Sci. U.S.A. 110, 3191-3196 (2013).
T. P. Howard et al., Synthesis of customized petroleum-replica fuel molecules by targeted
modification of free fatty acid pools in Escherichia coli. Proc. Nat. Acad. Sci. U.S.A. 110, 76367641 (2013).
M. Harger et al., Expanding the Product Profile of a Microbial Alkane Biosynthetic Pathway. ACS
Synth. Biol. 2, 59-62 (2012).
142
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
B. Khara et al., Production of Propane and Other Short-Chain Alkanes by Structure-Based
Engineering of Ligand Specificity in Aldehyde-Deformylating Oxygenase. ChemBioChem 14,
1204-1208 (2013).
M. Fuchs et al., Amination of benzylic and cinnamic alcohols via a biocatalytic, aerobic,
oxidation-transamination cascade. RSCAdvances 2, 6262-6265 (2012).
E. Park, M. Kim, J. Shin, Molecular determinants for substrate selectivity of w-transaminases.
Appl.Microbiol.Biotechnol. 93, 2425-2435 (2012).
K. E. Scholz et al., Synthesis of Chiral Cyanohydrins by Recombinant Escherichia coli Cells in a
Micro-Aqueous Reaction System. Applied and Environmental Microbiology 78, 5025-5027
(2012).
T. Purkarthofer et al., A Biocatalytic Henry Reaction-The Hydroxynitrile Lyase from Hevea
brasiliensis Also Catalyzes Nitroaldol Reactions. Angewandte Chemie International Edition 45,
3454-3456 (2006).
M. J. H. Moonen, A. H. Westphal, I. M. C. M. Rietjens, W. J. H. van Berkel, Enzymatic BaeyerVilliger Oxidation of Benzaldehydes. Advanced Synthesis & Catalysis 347, 1027-1034 (2005).
T. He et al., Utilization of biocatalytic promiscuity for direct Mannich reaction. Journal of
Molecular Catalysis B: Enzymatic 67, 189-194 (2010).
K. Li et al., Lipase-catalysed direct Mannich reaction in water: utilization of biocatalytic
promiscuity for C-C bond formation in a "one-pot" synthesis. Green Chemistry 11, 777-779
(2009).
M. D. Mihovilovic et a/., Biooxidation of ketones with a cyclobutanone structural motif by
recombinant whole-cells expressing 4-hydroxyacetophenone monooxygenase. Journal of
Molecular Catalysis B: Enzymatic 32, 135-140 (2005).
J. Zaldivar, A. Martinez, L. 0. Ingram, Effect of selected aldehydes on the growth and
fermentation of ethanologenic Escherichia coli. Biotechnology and Bioengineering 65, 24-33
(1999).
J. Visvalingam, J. D. Hernandez-Doria, R. A. Holley, Examination of the Genome-Wide
Transcriptional Response of Escherichia coli 0157:H7 to Cinnamaldehyde Exposure. Applied and
Environmental Microbiology 79, 942-950 (2013).
L. Wang, P. Tang, X. Fan, Q. Yuan, Effect of selected aldehydes found in the corncob
hemicellulose hydrolysate on the growth and xylitol fermentation of Candida tropicalis.
Biotechnology Progress 29, 1181-1189 (2013).
D. J. Fitzgerald, M. Stratford, M. J. Gasson, A. Narbad, Structure-Function Analysis of the Vanillin
Molecule and Its Antifungal Properties. Journal of Agricultural and Food Chemistry 53, 17691775 (2005).
C. Huang, H. Wu, Q.-p. Liu, Y.-y. Li, M.-h. Zong, Effects of Aldehydes on the Growth and Lipid
Accumulation of Oleaginous Yeast Trichosporon fermentans. Journal of Agricultural and Food
Chemistry 59, 4606-4613 (2011).
E. N. Miller et al., Silencing of NADPH-Dependent Oxidoreductase Genes (yqhD and dkgA) in
Furfural-Resistant Ethanologenic Escherichia coli. Applied and Environmental Microbiology 75,
4315-4323 (2009).
E. N. Miller et al., Furfural Inhibits Growth by Limiting Sulfur Assimilation in Ethanologenic
Escherichia coli Strain LY180. Applied and Environmental Microbiology 75, 6132-6141 (2009).
N. P. Singh, A. Khan, Acetaldehyde: genotoxicity and cytotoxicity in human lymphocytes.
Mutation Research/DNA Repair 337, 9-17 (1995).
H. Esterbauer, R. J. Schaur, H. Zollner, Chemistry and biochemistry of 4-hydroxynonenal,
malonaldehyde and related aldehydes. Free Radical Biology and Medicine 11, 81-128 (1991).
143
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
J. D. West, L. J. Marnett, Endogenous Reactive Intermediates as Modulators of Cell Signaling and
Cell Death. Chemical Research in Toxicology 19, 173-194 (2006).
P. A. Grimsrud, H. Xie, T. J. Griffin, D. A. Bernlohr, Oxidative Stress and Covalent Modification of
Protein with Bioactive Aldehydes. Journal of Biological Chemistry 283, 21837-21841 (2008).
J. R. Warner, P. J. Reeder, A. Karimpour-Fard, L. B. A. Woodruff, R. T. Gill, Rapid profiling of a
microbial genome using mixtures of barcoded oligonucleotides. Nat Biotech 28, 856-862 (2010).
A. S. Gort, D. M. Ferber, J. A. Imlay, The regulation and role of the periplasmic copper, zinc
superoxide dismutase of Escherichia coli. Molecular Microbiology 32, 179-191 (1999).
T. Mills, N. Sandoval, R. Gill, Cellulosic hydrolysate toxicity and tolerance mechanisms in
Escherichia coli. Biotechnol Biofuels 2, 1-11 (2009).
J. M. Almeida, M. Bertilsson, M. Gorwa-Grauslund, S. Gorsich, G. Liden, Metabolic effects of
furaldehydes and impacts on biotechnological processes. Applied Microbiology and
Biotechnology 82, 625-638 (2009).
E. M. Sampson, T. A. Bobik, Microcompartments for B12-Dependent 1,2-Propanediol
Degradation Provide Protection from DNA and Cellular Damage by a Reactive Metabolic
Intermediate. Journal of Bacteriology 190, 2966-2971 (2008).
V. J. Starai, J. Garrity, J. C. Escalante-Semerena, Acetate excretion during growth of Salmonella
enterica on ethanolamine requires phosphotransacetylase (EutD) activity, and acetate recapture
requires acetyl-CoA synthetase (Acs) and phosphotransacetylase (Pta) activities. Microbiology
151, 3793-3801 (2005).
E. Y. Kim, D. Tullman-Ercek, Engineering nanoscale protein compartments for synthetic
organelles. Current Opinion in Biotechnology 24, 627-632 (2013).
H. Zhu, R. Gonzalez, T. A. Bobik, Coproduction of Acetaldehyde and Hydrogen during Glucose
Fermentation by Escherichia coli. Applied and Environmental Microbiology 77, 6441-6450
(2011).
A. N. Jain, T. R. Khan, A. J. Daugulis, Bioproduction of benzaldehyde in a solid-liquid two phase
partitioning bioreactor using Pichia pastoris. Biotechnology Letters 32:11, 1649-1654 (2010).
D. Hua et al., Enhanced vanillin production from ferulic acid using adsorbent resin. Applied
Microbiology and Biotechnology 74:4, 783-790 (2007).
N. J. Walton, M. J. Mayer, A. Narbad, Vanillin. Phytochemistry 63, 505-515 (2003).
G. A. Burdock, Fenaroli's Handbook of Flavor Ingredients. (CRC Press, Boca Raton, Florida, 2005),
vol. Fifth Edition.
T. Li, J. P. N. Rosazza, Biocatalytic Synthesis of Vanillin. Applied and Environmental Microbiology
66, 684-687 (2000).
K. M. Draths, J. W. Frost, Conversion of D-glucose into catechol: the not-so-common pathway of
aromatic biosynthesis. Journal of the American Chemical Society 113, 9361-9363 (1991).
K. M. Draths et al., Biocatalytic synthesis of aromatics from D-glucose: the role of transketolase.
Journal of the American Chemical Society 114, 3956-3962 (1992).
K. D. Snell, K. M. Draths, J. W. Frost, Synthetic Modification of the Escherichia coli Chromosome:
Enhancing the Biocatalytic Conversion of Glucose into Aromatic Chemicals. Journal of the
American Chemical Society 118, 5605-5614 (1996).
A. Berry, Improving production of aromatic compounds in Escherichia coli by metabolic
engineering. Trends in Biotechnology 14, 250-256 (1996).
Food, D. o. H. Drug Administration, S. Human. (2012).
E. M. Ellis, Microbial aldo-keto reductases. FEMS Microbiology Letters 216, 123-131 (2002).
J. Sambrook, D. W. Russell, Molecular cloning: a laboratory manual. (Cold Spring Harbor
Laboratory Press, Cold Spring Harbor, N.Y., 2001).
144
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
T. Baba et al., Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the
Keio collection. Mol Syst Biol 2, 2006.0008 (2006).
L. C. Thomason, N. Costantino, D. L. Court, in Current Protocols in Molecular Biology. (John Wiley
& Sons, Inc., 2001).
K. A. Datsenko, B. L. Wanner, One-step inactivation of chromosomal genes in Escherichia coli K12 using PCR products. Proc NatlAcad Sci USA 97, 6640-6645 (2000).
H.-C. Tseng, C. H. Martin, D. R. Nielsen, K. L. J. Prather, Metabolic Engineering of Escherichia coli
for Enhanced Production of (R)- and (S)-3-Hydroxybutyrate. Applied and Environmental
Microbiology 75, 3137-3145 (2009).
H.-C. Tseng, C. Harwell, C. Martin, K. Prather, Biosynthesis of chiral 3-hydroxyvalerate from
single propionate-unrelated carbon sources in metabolically engineered E. coli. Microbial Cell
Factories 9, 96 (2010).
D. Juminaga et al., Modular Engineering of I-Tyrosine Production in Escherichia coli. Applied and
Environmental Microbiology 78, 89-98 (2012).
M. M. Nakano, N. Corbell, J. Besson, P. Zuber. (Springer Berlin / Heidelberg, 1992), vol. 232, pp.
313-321.
D. T. Fox, K. Hotta, C.-Y. Kim, A. T. Koppisch, The Missing Link in Petrobactin Biosynthesis: asbF
Encodes a (-)-3-Dehydroshikimate Dehydratase. Biochemistry 47, 12251-12253 (2008).
H. E. Klock, S. A. Lesley, in High Throughput Protein Expression and Purification, S. A. Doyle, Ed.
(Humana Press, 2009), vol. 498, pp. 91-103.
J. Ko et al., Conversion of Methylglyoxal to Acetol by Escherichia coli Aldo-Keto Reductases.
Journal of Bacteriology 187, 5782-5789 (2005).
P. C. Turner et al., YqhC regulates transcription of the adjacent Escherichia coli genes yqhD and
dkgA that are involved in furfural tolerance. J Ind Microbiol Biotechnol 38,431-439 (2011).
G. Sulzenbacher et al., Crystal Structure of E. coli Alcohol Dehydrogenase YqhD: Evidence of a
Covalently Modified NADP Coenzyme. Journal of Molecular Biology 342,489-502 (2004).
J. M. Perez, F. A. Arenas, G. A. Pradenas, J. M. Sandoval, C. C. Vasquez, Escherichia coli YqhD
Exhibits Aldehyde Reductase Activity and Protects from the Harmful Effect of Lipid Peroxidationderived Aldehydes. Journal of Biological Chemistry 283, 7346-7353 (2008).
A. Pick, B. Ruhmann, J. Schmid, V. Sieber, Novel CAD-like enzymes from Escherichia coli K-12 as
additional tools in chemical production. AppI Microbiol Biot 97:13, 5815-5824 (2013).
D. Koma, H. Yamanaka, K. Moriyoshi, T. Ohmoto, K. Sakai, Production of Aromatic Compounds
by Metabolically Engineered Escherichia coli with an Expanded Shikimate Pathway. Applied and
Environmental Microbiology 78, 6203-6216 (2012).
J. Zhang, J. P. Klinman, Enzymatic Methyl Transfer: Role of an Active Site Residue in Generating
Active Site Compaction That Correlates with Catalytic Efficiency. Journal of the American
Chemical Society 133, 17134-17137 (2011).
J. M. Jez, T. M. Penning, The aldo-keto reductase (AKR) superfamily: an update. Chemicobiological interactions 130-132, 499-525 (2001).
A. J. Lapthorn, X. Zhu, E. M. Ellis, The diversity of microbial aldo/keto reductases from
Escherichia coli K12. Enzymology and Molecular Biology of Carbonyl Metabolism 16 202, 168177 (2013).
S. Jeudy, V. Monchois, C. Maza, J.-M. Claverie, C. Abergel, Crystal structure of Escherichia coli
DkgA, a broad-specificity aldo-keto reductase. Proteins: Structure, Function, and Bioinformatics
62, 302-307 (2006).
H. J6rnvall et al., Short-chain dehydrogenases/reductases (SDR). Biochemistry 34, 6003-6013
(1995).
145
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
138.
139.
140.
U. Oppermann et al., Short-chain dehydrogenases/reductases (SDR): the 2002 update.
Enzymology and Molecular Biology of Carbonyl Metabolism - 11th International Workshop. 143144, 247-253 (2003).
K. L. Kavanagh, H. J6rnvall, B. Persson, U. Oppermann, Medium- and short-chain
dehydrogenase/reductase gene and protein families. Cellular and Molecular Life Sciences 65,
3895-3906 (2008).
B. Persson, J. S. Zigler, H. Jornvall, A Super-Family of Medium-Chain Dehydrogenases/Reductases
(MDR). European Journal of Biochemistry 226, 15-22 (1994).
E. Nordling, H. J6rnvall, B. Persson, Medium-chain dehydrogenases/reductases (MDR). European
Journal of Biochemistry 269, 4267-4276 (2002).
H. H. Wang et al., Programming cells by multiplex genome engineering and accelerated
evolution. Nature 460,894-898 (2009).
L. S. Qi et al., Repurposing CRISPR as an RNA-Guided Platform for Sequence-Specific Control of
Gene Expression. Cell 152, 1173-1183 (2013).
P. Pharkya, A. P. Burgard, C. D. Maranas, OptStrain: A computational framework for redesign of
microbial production systems. Genome Research 14, 2367-2376 (2004).
H. Alper, K. Miyaoku, G. Stephanopoulos, Construction of lycopene-overproducing E. coli strains
by combining systematic and combinatorial gene knockout targets. Nat Biotech 23, 612-616
(2005).
C. N. S. Santos, W. Xiao, G. Stephanopoulos, Rational, combinatorial, and genomic approaches
for engineering L-tyrosine production in Escherichia coli. Proceedings of the National Academy
of Sciences 109, 13538-13543 (2012).
P. Barghini, D. Di Gioia, F. Fava, M. Ruzzi, Vanillin production using metabolically engineered
Escherichia coli under non-growing conditions. Microbial Cell Factories 6, 13 (2007).
S.-H. Yoon et al., Production of Vanillin by Metabolically Engineered Escherichia coli.
Biotechnology Letters 27, 1829-1832 (2005).
S.-H. Yoon et aL., Enhanced Vanillin Production from Recombinant E.coli Using NTG Mutagenesis
and Adsorbent Resin. Biotechnology Progress 23, 1143-1148 (2007).
E.-G. Lee et al., Directing vanillin production from ferulic acid by increased acetyl-CoA
consumption in recombinant Escherichia coli. Biotechnology and Bioengineering 102, 200-208
(2009).
Annual Energy Review 2011, United States Energy Information Administration (2012).
International Energy Statistics, United States Energy Information Administration (2014).
B. Strogen, A. Horvath, Greenhouse Gas Emissions from the Construction, Manufacturing,
Operation, and Maintenance of U.S. Distribution Infrastructure for Petroleum and Biofuels. J.
Infrastruct. Syst. 19, 371-383 (2013).
D. R. Burgess, in NIST Chemistry WebBook, NISTStandard Reference Database Number 69, Eds.
P.J. Linstrom and W.G. Mallard. (National Institute of Standards and Technology, Gaithersburg,
MD).
1. Hunwartzen, Modification of CFR Test Engine Unit to Determine Octane Numbers of Pure
Alcohols and Gasoline-Alcohol Blends. SAE Technical Paper 820002, (1982).
C. Jin, M. Yao, H. Liu, C.-f. F. Lee, J. Ji, Progress in the production and application of n-butanol as
a biofuel. Renewable and Sustainable Energy Reviews 15, 4080-4106 (2011).
Solubility Data Series, International Union of Pure and Applied Chemistry. (Pergamon Press,
Oxford, 1982), vol. 15.
Soil Screening Guidance, US Environmental Protection Agency (9355.4-23, 1996).
L. Gevantman, in CRC Handbook of Chemistry and Physics. (CRC Press, Boca Raton, FL, 1996).
146
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
156.
157.
158.
159.
160.
161.
D. Mackay, W. Y. Shiu, A Critical Review of Henry's Law Constants for Chemicals of
Environmental Interest. J. Phys. Chem. Ref. Data 10, 1175-1199 (1981).
C. Morley, A Fundamentally Based Correlation Between Alkane Structure and Octane Number.
Combustion Science and Technology 55, 115-123 (1987).
Solubility Data Series, International Union of Pure and Applied Chemistry. (Pergamon Press,
Oxford, 1988), vol. 37.
D. Mackay, W. Y. Shiu, K. C. Ma, Illustrated Handbook of Physical-Chemical Properties and
Environmental Fatefor Organic Chemicals. (Lewis Publishers/CRC Press, Boca Raton, FL, 1993),
vol. 3.
Solubility Data Series, International Union of Pure and Applied Chemistry. (Pergamon Press,
Oxford, 1988), vol. 38.
W. N. Sanders, J. B. Maynard, Capillary gas chromatographic method for determining the C3-C12
hydrocarbons in full-range motor gasolines. Anal. Chem. 40, 527-535 (1968).
N. G. Johansen, L. S. Ettre, R. L. Miller, Quantitative Analysis of Hydrocarbons by Structural
Group Type in Gasolines and Distillates: 1. Gas Chromatography. J. Chromatogr. A 256, 393-417
(1983).
P. V. Cline, J. J. Delfino, P. S. C. Rao, Partitioning of aromatic constituents into water from
gasoline and other complex solvent mixtures. Environ. Sci. Technol. 25, 914-920 (1991).
A. M. Kunjapur, Y. Tarasova, K. L. J. Prather, Synthesis and Accumulation of Aromatic Aldehydes
in an Engineered Strain of Escherichia coli. J. Am. Chem. Soc. 136, 11644-11654 (2014).
M. J. Sheppard, A. M. Kunjapur, S. J. Wenck, K. L. J. Prather, Retro-biosynthetic screening of a
modular pathway design achieves selective route for microbial synthesis of 4-methyl-pentanol.
Nat Commun 5, 5031 (2014).
G. M. Rodriguez, S. Atsumi, Toward aldehyde and alkane production by removing aldehyde
reductase activity in Escherichia coli. Metab. Eng. 25, 227-237 (2014).
P. Kallio, A. Pasztor, K. Thiel, M. K. Akhtar, P. R. Jones, An engineered pathway for the
biosynthesis of renewable propane. Nat Commun 5, (2014).
H.-C. Tseng, K. L. J. Prather, Controlled biosynthesis of odd-chain fuels and chemicals via
engineered modular metabolic pathways. Proc. Nati. Acad. Sci. U.S.A. 109, 17925-17930 (2012).
Y. Dekishima, E. 1. Lan, C. R. Shen, K. M. Cho, J. C. Liao, Extending Carbon Chain Length of 1Butanol Pathway for 1-Hexanol Synthesis from Glucose by Engineered Escherichia coli. J. Am.
Chem. Soc. 133, 11399-11401 (2013).
J. P. Torella et al., Tailored fatty acid synthesis via dynamic control of fatty acid elongation. Proc.
Natl. Acad. Sci. U.S.A. 110, 11290-11295 (2013).
B. Bond-Watts, R. J. Bellerose, M. C. Y. Chang, Enzyme mechanism as a kinetic control element
for designing synthetic biofuel pathways. Nat Chem Biol 7, 222-227 (2011).
S. Slater et al., Multiple 1-Ketothiolases Mediate Poly(P-Hydroxyalkanoate) Copolymer Synthesis
in Ralstonia eutropha. Journal of Bacteriology 180, 1979-1987 (1998).
J. L. Fortman et al., Biofuel alternatives to ethanol: pumping the microbial well. Trends in
Biotechnology 26, 375-381 (2008).
P. P. Peralta-Yahya, F. Zhang, S. B. del Cardayre, J. D. Keasling, Microbial engineering for the
production of advanced biofuels. Nature 488, 320-328 (2012).
A. M. Kunjapur, K. L. J. Prather, Microbial Engineering for Aldehyde Synthesis. Applied and
Environmental Microbiology 81, 1892-1901 (2015).
K. Ishige, T. Noguchi, Inorganic polyphosphate kinase and adenylate kinase participate in the
polyphosphate:AMP phosphotransferase activity of Escherichia coli. Proceedings of the National
Academy of Sciences 97, 14168-14171 (2000).
147
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
175.
176.
177.
178.
179.
180.
S. M. Resnick, A. J. B. Zehnder, In Vitro ATP Regeneration from Polyphosphate and AMP by
Polyphosphate:AMP Phosphotransferase and Adenylate Kinase from Acinetobacter johnsonii
210A. Applied and Environmental Microbiology 66, 2045-2051 (2000).
W. A. van der Donk, H. Zhao, Recent developments in pyridine nucleotide regeneration. Current
Opinion in Biotechnology 14,421-426 (2003).
H. Zhao, W. A. van der Donk, Regeneration of cofactors for use in biocatalysis. Current Opinion in
Biotechnology 14, 583-589 (2003).
H. Itoh, T. Shiba, Polyphosphate Synthetic Activity of Polyphosphate:AMP Phosphotransferase in
Acinetobacter johnsonii 210A. Journal of Bacteriology 186, 5178-5181 (2004).
T. Shiba et al., Polyphosphate:AMP Phosphotransferase as a Polyphosphate-Dependent
Nucleoside Monophosphate Kinase in Acinetobacter johnsonii 210A. Journal of Bacteriology
187, 1859-1865 (2005).
W. Liu, P. Wang, Cofactor regeneration for sustainable enzymatic biosynthesis. Biotechnology
Advances 25, 369-384 (2007).
J. E. Dueber et al., Synthetic protein scaffolds provide modular control over metabolic flux. Not
Biotech 27, 753-759 (2009).
M. M. Bradford, A rapid and sensitive method for the quantitation of microgram quantities of
protein utilizing the principle of protein-dye binding. Analytical Biochemistry 72, 248-254 (1976).
J. W. Ogilvie, S. C. Whitaker, Reaction of tris with aldehydes: Effect of tris on reactions catalyzed
by homoserine dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase. Biochimica et
Biophysica Acta (BBA) - Enzymology 445, 525-536 (1976).
W. A. Bubb, H. A. Berthon, P. W. Kuchel, Tris Buffer Reactivity with Low-Molecular-Weight
Aldehydes: NMR Characterization of the Reactions of Glyceraldehyde-3-Phosphate. Bioorganic
Chemistry 23, 119-130 (1995).
J. Josse, Constitutive Inorganic Pyrophosphatase of Escherichia coli : 1. PURIFICATION AND
CATALYTIC PROPERTIES. Journal of Biological Chemistry 241, 1938-1947 (1966).
J. Josse, R. A. D. With an appendix by, Constitutive Inorganic Pyrophosphatase of Escherichia coli
: 11. NATURE AND BINDING OF ACTIVE SUBSTRATE AND THE ROLE OF MAGNESIUM. Journal of
Biological Chemistry 241, 1948-1954 (1966).
S. C. K. Wong, D. C. Hall, J. Josse, Constitutive Inorganic Pyrophosphatase of Escherichia coli : Ill.
MOLECULAR WEIGHT AND PHYSICAL PROPERTIES OF THE ENZYME AND ITS SUBUNITS. Journal of
Biological Chemistry 245, 4335-4341 (1970).
P. M. Burton, D. C. Hall, J. Josse, Constitutive Inorganic Pyrophosphatase of Escherichia coli : IV.
CHEMICAL STUDIES OF PROTEIN STRUCTURE. Journal of Biological Chemistry 245, 4346-4352
(1970).
J. Chen et al., Pyrophosphatase is essential for growth of Escherichia coli. Journal of Bacteriology
172, 5686-5689 (1990).
A. A. Baykov et al., Catalysis by Escherichia coli Inorganic Pyrophosphatase: pH and Mg2+
Dependence. Biochemistry 35, 4655-4661 (1996).
R. H. Upson, R. P. Haugland, M. N. Malekzadeh, R. P. Haugland, A Spectrophotometric Method
to Measure Enzymatic Activity in Reactions That Generate Inorganic Pyrophosphate. Analytical
Biochemistry 243, 41-45 (1996).
P. R. Cunningham, J. Ofengand, Use of inorganic pyrophosphatase to improve the yield of in vitro
transcription reactions catalyzed by T7 RNA polymerase. (Eaton, Natick, MA, ETATS-UNIS, 1990),
vol. 9, pp. 2.
I. D. Pokrovskaya, V. V. Gurevich, In Vitro Transcription: Preparative RNA Yields in Analytical
Scale Reactions. Analytical Biochemistry 220, 420-423 (1994).
148
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199.
200.
201.
L. Peller, In vitro RNA synthesis should be coupled to pyrophosphate hydrolysis. Biochemical and
Biophysical Research Communications 63, 912-916 (1975).
S. Schmelz, J. H. Naismith, Adenylate-forming enzymes. Current Opinion in Structural Biology 19,
666-671 (2009).
E. S. Sattely, M. A. Fischbach, C. T. Walsh, Total biosynthesis: in vitro reconstitution of polyketide
and nonribosomal peptide pathways. Natural Product Reports 25,757-793 (2008).
G. G. Gross, Formation and Reduction of Intermediate Acyladenylate by Aryl-Aldehyde.
European Journal of Biochemistry 31, 585-592 (1972).
G. G. Gross, M. H. Zenk, Reduktion aromatischer Sauren zu Aldehyden und Alkoholen im
zellfreien System. European Journal of Biochemistry 8, 413-419 (1969).
N. Kato et al., Purification and Characterization of Aromatic Acid Reductase from Nocardia
asteroides JCM 3016. Agricultural and Biological Chemistry 55, 757-762 (1991).
K. V. Solomon, T. S. Moon, B. Ma, T. M. Sanders, K. L. J. Prather, Tuning Primary Metabolism for
Heterologous Pathway Productivity. ACS Synthetic Biology 2, 126-135 (2013).
E. R. Hondorp, R. G. Matthews, Oxidative Stress Inactivates Cobalamin-Independent Methionine
Synthase (MetE) in Escherichia coli. PLoS Biol 2, e336 (2004).
L. 1. Leichert, U. Jakob, Protein Thiol Modifications Visualized In Vivo. PLoS Biol 2, e333 (2004).
E. R. Hondorp, R. G. Matthews, Oxidation of Cysteine 645 of Cobalamin-Independent
Methionine Synthase Causes a Methionine Limitation in Escherichia coli. Journal of Bacteriology
191, 3407-3410 (2009).
E. A. Mordukhova, J.-G. Pan, Evolved Cobalamin-Independent Methionine Synthase (MetE)
Improves the Acetate and Thermal Tolerance of Escherichia coli. Applied and Environmental
Microbiology 79, 7905-7915 (2013).
D. Kumar, J. Gomes, Methionine production by fermentation. Biotechnology Advances 23, 41-61
(2005).
T. Willke, Methionine production-a critical review. Applied Microbiology and Biotechnology 98,
9893-9914 (2014).
S. Nakamori, S. Kobayashi, T. Nishimura, H. Takagi, Mechanism of I-methionine overproduction
by Escherichia coli: the replacement of Ser-54 by Asn in the MetJ protein causes the
derepression of 1-methionine biosynthetic enzymes. Applied Microbiology and Biotechnology 52,
179-185 (1999).
W. F. Dischert, R. (Metabolic Explorer, 2013), vol. W02013/190343(A1).
Y. Usuda, 0. Kurahashi, Effects of Deregulation of Methionine Biosynthesis on Methionine
Excretion in Escherichia coli. Applied and Environmental Microbiology 71, 3228-3234 (2005).
Y. Kai et al., Engineering of Escherichia coli 1-serine O-acetyltransferase on the basis of crystal
structure: desensitization to feedback inhibition by I-cysteine. Protein Engineering Design and
Selection 19, 163-167 (2006).
P. Nawabi, S. Bauer, N. Kyrpides, A. Lykidis, Engineering E. coli for biodiesel production utilizing a
bacterial fatty acid methyltransferase. Applied and Environmental Microbiology 77, 8052-8061
(2011).
J. K. Coward, M. D'Urso-Scott, W. D. Sweet, Inhibition of catechol-O-methyltransferase by Sadenosylhomocysteine and S-adenosylhomocysteine sulfoxide, a potential transition-state
analog. Biochemical Pharmacology 21, 1200-1203 (1972).
J. K. Coward, E. P. Slisz, F. Y. H. Wu, Kinetic studies on catechol 0-methyltransferase, Product
inhibition and the nature of the catechol binding site. Biochemistry 12, 2291-2297 (1973).
N. Parveen, K. A. Cornell, Methylthioadenosine/S-adenosylhomocysteine nucleosidase, a critical
enzyme for bacterial metabolism. Molecular microbiology 79, 7-20 (2011).
149
202.
203.
204.
205.
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
K. B. Xavier, B. L. Bassler, LuxS quorum sensing: more than just a numbers game. Current
Opinion in Microbiology 6, 191-197 (2003).
A. F. Gonzalez Barrios et al., Autoinducer 2 Controls Biofilm Formation in Escherichia coli
through a Novel Motility Quorum-Sensing Regulator (MqsR, B3022). Journal of Bacteriology 188,
305-316 (2006).
J. Li et al., Quorum Sensing in Escherichia coli Is Signaled by Al-2/LsrR: Effects on Small RNA and
Biofilm Architecture. Journal of Bacteriology 189, 6011-6020 (2007).
M. Hegde et al., Chemotaxis to the Quorum-Sensing Signal AI-2 Requires the Tsr Chemoreceptor
and the Periplasmic LsrB Al-2-Binding Protein. Journal of Bacteriology 193, 768-773 (2011).
J. D. Lozada-Ramirez, I. Martinez-Martinez, F. Garcia-Carmona, A. Sanchez-Ferrer, Cloning,
Overexpression, Purification, and Characterization of S-Adenosylhomocysteine Hydrolase from
Corynebacterium efficiens YS-314. Biotechnology Progress 24, 120-127 (2008).
N. M. Halliday, K. R. Hardie, P. Williams, K. Winzer, D. A. Barrett, Quantitative liquid
chromatography-tandem mass spectrometry profiling of activated methyl cycle metabolites
involved in LuxS-dependent quorum sensing in Escherichia coli. Analytical Biochemistry 403, 2029 (2010).
W. R. Farmer, J. C. Liao, Improving lycopene production in Escherichia coli by engineering
metabolic control. Nat Biotech 18, 533-537 (2000).
F. Zhang, J. M. Carothers, J. D. Keasling, Design of a dynamic sensor-regulator system for
production of chemicals and fuels derived from fatty acids. Nat Biotech 30, 354-359 (2012).
R. H. Dahl et al., Engineering dynamic pathway regulation using stress-response promoters. Nat
Biotech 31, 1039-1046 (2013).
P. Xu, L. Li, F. Zhang, G. Stephanopoulos, M. Koffas, Improving fatty acids production by
engineering dynamic pathway regulation and metabolic control. Proceedings of the National
Academy of Sciences 111, 11299-11304 (2014).
S. Raman, J. K. Rogers, N. D. Taylor, G. M. Church, Evolution-guided optimization of biosynthetic
pathways. Proceedings of the National Academy of Sciences 111, 17803-17808 (2014).
M. Thanbichler, A. A. Iniesta, L. Shapiro, A comprehensive set of plasmids for vanillate- and
xylose-inducible gene expression in Caulobacter crescentus. Nucleic Acids Research 35, e137e137 (2007).
A. A. Iniesta, F. Garcia-Heras, J. Abell6n-Ruiz, A. Gallego-Garcia, M. Elias-Arnanz, Two Systems
for Conditional Gene Expression in Myxococcus xanthus Inducible by lsopropyl-fl-DThiogalactopyranoside or Vanillate. Journal of Bacteriology 194, 5875-5885 (2012).
A. Kaczmarczyk, J. A. Vorholt, A. Francez-Charlot, Synthetic vanillate-regulated promoter for
graded gene expression in Sphingomonas. Sci. Rep. 4, (2014).
M. Gitzinger et al., The food additive vanillic acid controls transgene expression in mammalian
cells and mice. Nucleic Acids Research 40, e37-e37 (2012).
U. Gerischer, A. Segura, L. N. Ornston, PcaU, a Transcriptional Activator of Genes for
Protocatechuate Utilization in Acinetobacter. Journal of Bacteriology 180, 1512-1524 (1998).
S. Y. Siehler, S. Dal, R. Fischer, P. Patz, U. Gerischer, Multiple-Level Regulation of Genes for
Protocatechuate Degradation in Acinetobacter baylyi Includes Cross-Regulation. Applied and
Environmental Microbiology 73, 232-242 (2007).
R. K. Jha, T. L. Kern, D. T. Fox, C. E. M. Strauss, Engineering an Acinetobacter regulon for
biosensing and high-throughput enzyme screening in E. coli via flow cytometry. Nucleic Acids
Research 42, 8150-8160 (2014).
M. D. McMahon, K. L. J. Prather, Functional Screening and In Vitro Analysis Reveal Thioesterases
with Enhanced Substrate Specificity Profiles That Improve Short-Chain Fatty Acid Production in
Escherichia coli. Applied and Environmental Microbiology 80, 1042-1050 (2014).
150
221.
C. H. Martin et al., A platform pathway for production of 3-hydroxyacids provides a biosynthetic
route to 3-hydroxy-y-butyrolactone. Nat Commun 4, 1414 (2013).
151