Demonstrating biocompatibility with supercritical C0 2: Biphasic cultivation
of Bacillus spp. and probing acclimation mechanisms through proteome and
lipid analysis
MACHSTSINTTT
MASSACHUSETTS INSTITUTE
OF TECHNOLOLGY
by
MAY 0 5 2015
Kyle Creighton Peet
LIBRARIES
B.S. Worcester Polytechnic Institute (2008)
Submitted to the Department of Civil and Environmental Engineering
in partial fulfillment of the requirements for the Degree of
Doctor of Philosophy in Environmental Biology
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
February 2015
C 2015 Massachusetts Institute of Technology. All rights reserved.
Signature redacted
A uth o r .............................................................................
...........
Department of Cii1 and Environmental Engineering
2015
Signature redacted
C ertified by ................................................
Doherty Assistant Professor in Oce
. ... .. .........................................
... . .. ..
Utilization in Civil and Environmental Engineering
Signature redacted
A ccepted by..............................
............
/7 anelle R. Thompson
Thesis Supervisor
.................................
Heidi Nepf
Donald and Martha Harleman Professor of Civil and Environmental Engineering
Chair, Graduate Program Committee
2
Demonstrating biocompatibility with supercritical C0 2 : Biphasic cultivation
of Bacillus spp. and probing acclimation mechanisms through proteome and
lipid analysis
by
Kyle Creighton Peet
Submitted to the Department of Civil and Environmental Engineering
On January 29', 2015, in partial fulfillment of the requirements for the Degree of
Doctor of Philosophy in Environmental Biology
Abstract
Supercritical (sc) CO 2 usage is increasing globally with applications as a sterilizing agent,
as a non-toxic solvent, and as the form of the greenhouse gas CO 2 injected underground for
geologic carbon sequestration (GCS). In this thesis I have described the isolation of
microorganisms from three different carbon sequestration pilot sites through a novel method of
successive scCO 2 enrichments. I show that microorganisms of the genus Bacillus, including GCS
site isolates, are resistant to the bactericidal properties of scCO 2 , and can germinate and grow in
an aqueous phase incubated under scCO 2 (Chapter 2). Bacterial resistance to scCO 2 challenges
the efficacy of scCO 2 based sterilization and indicates that microbial activity may be harnessed
in engineered environments containing scCO2 (e.g. biochemical catalysis involving scCO 2 as a
solvent or biofilm/biomineralized barriers to scCO 2 leakage from GCS sites). In an effort to
understand the physiology of acclimation to scCO 2 , I have sequenced and analyzed the genomes
of two GCS-site isolates, B. cereus MIT0214 and B. subterraneusMITOTI (Chapter 3). I have
used genome-enabled analysis of the proteome combined with analysis of membrane lipids to
ask whether cellular macromolecules are differentially represented in cells grown under different
headspace and pressure conditions including CO 2 and scCO 2 (Chapters 4). In this chapter I have
examined the following three hypotheses regarding the mechanisms employed by Bacilli to resist
scCO 2 : (1): Resistance to CO 2 stress is governed by a similar response as acclimation to low pH
stress. (2): Cell wall and membrane alterations promote bacterial growth under scCO 2 by
modulating the cell's microenvironment. (3): Global expression of proteins mediating cellular
homeostasis in viable but non-growing (stationary-phase) populations acclimated to scCO 2
resembles a generalized profile of anaerobic growth, with notable exceptions of individual
protein(s) that mediate acclimation. The results from this thesis enhance understanding of
bacterial resistance to scCO 2 , enabling improved strategies for scCO 2-based sterilization and
accelerating biotechnological applications of scCO 2-biocompatible organisms.
Thesis Supervisor: Janelle R. Thompson
Title: Doherty Assistant Professor in Ocean Utilization in Civil and Environmental Engineering
3
4
Acknowledgements
I have many people and agencies to thank for this thesis and I am sure that there are some
who I forgetten to mention. Major funding sources for the research in this thesis include MIT and
the MIT's Energy Initiative (MITEI), which funded both the research and my first year of
graduate school through the BP-MITEI fellowship. The US Department of Energy National
Energy Technology Lab (DOE NETL) has also provided essential grants for this work. The
Ralph M. Parsons Lab in the Department of Civil and Environmental Engineering has enabled
this research to be conducted, and the MIT Ippen Fund has funded travel for presentation of
portions of this work at American Geophysical Union (AGU) conferences.
My advisor, Janelle R. Thompson, has been the most important factor in the completion
of this work, and without her mentorship, it would not have been possible. I'm forever indebted
to her for taking me on and I have learned so much over the past 6.5 years. She has helped guide
this research throughout many successes and failures, and has helped me see through perceived
failures to see that positive results can be obtained where I did not initially see any. She has
helped greatly with enhancing the presentation of these results, both in written form, and in
presentation at conferences and my thesis defense.
Members of the Thompson lab past and present have been friends, mentors and great
people to conduct science with over the years. Samodha Fernando and Hector Hernandez helped
get me started in the lab and teach me many things in the beginning of graduate school. Adam
Freedman has been a great lab and project mate over the years and I think we've both learned
many things from each other as we struggled to grow microbes under supercritical CO 2 . Kevin
Penn has also been very helpful in giving advice over the past few years. Other members of the
lab who I must thank include Jia Yi Har, Jia Wang, Jean Pierre Nshimyimana, Tim Helbig,
Hanny Rivera, Ju Young Lim, Luciane Chimetto, Eric Hill, and Carolina Bastidas. I enjoyed
working with several UROPs who helped with parts of this work (Vanya Britto, Joseph Aboki,
Holly Josephs, Tzipora Wagner, and Matthew Archer) and I hope I was able to teach them useful
skills through their work.
My committee members, Martin Polz, Penny Chisholm, and Roger Summons have also
provided essential advice for guiding the direction of my research and crafting the story of this
thesis. All of my committee members have always welcomed me to use their lab facilities
throughout the years, and members of their labs have taught me many techniques and methods.
Particular members of their labs who were of great help include Kelden Pehr, Florence Shubotz,
Carolyn Colonero, Michael Cutler, Jan Hendrick, Alison Takemura, Allison Coe, and Jessie
Thompson. In addition to committee members and their lab members, many collaborators and
other scientists have provided essential advice and assistance for this work, including: Jonathan
Ajo-Franklin, Chris Boreham, Tommy Phelps, Susan Pfiffner, Mike Timko, Pete Wishnok, and
Ravi Kodihalli.
The Parsons community also deserves recognition as a great place to work, with great
people. Sheila Frankel, Jim Long, Darlene Strother, Vicky Murphy have all helped me and all
help keep Parsons running. I have met numerous people who have helped me through friendship
and discussing science. Anthony, James, Ben, Jeff, Gaj, Sean, Mitul, Ryan, Amy, Matt, Jessie,
Teresa, Alison T, Jim, Robin, Mason, Chris M, Mike S, Ruby, Alex, Kesley, Jen, Sarah Jane,
Dave G, Dave W, Sarah, Fatima, Katya, Illana, Alison H, Chris L, Jon, Patricia, Ali P, Mark S,
Chris S, Jesse S, and I'm sure others who I have forgotten to mention. Most of all among the
5
Parsons community, I must thank Kelly who has helped me throughout graduate school, both in
science and in life. I am excited for many more adventures with Kelly.
Finally I must thank friends and family for their support outside of MIT. Joey, Duncan,
Sergey, James, Ande, Andreas, George, Corey, Rachel, and Adam have all been great friends
over the years. All of my aunts, uncles, and cousins have been positive parts of my life,
encouraging me in whatever I have pursued. Especially Helene, Duncan, Nico, Janny, Jonnie,
Jacob, Rebecca, and Hannah. Ling-Se has been the best sister and I'll always look up to her. Joey
and my niece Helene are the best additions to my loving family. Lastly my Mom and Dad raised
me, guided me, and inspired me and I'll always be in their debt.
6
Table of contents
Chapter 1: Review of supercritical CO 2 industries that may involve microbial
activity and possible survival strategies to scCO 2
1.1
9
Microbial impacts of Geologic Carbon Sequestration ....................
9
1.2 Microbial activity in high pressure, C0 2-rich environments ...............
1.3
Microbial sterilization and survival under scCO 2
1.4
Biocatalysis under supercritical CO 2 ...
. . . . . . . . . . . . . . . . . . . . .
12
..
. . . . . . . . . . . . .. . . . . . . . . . . . . . . . .
13
15
1.5 Adaption mechanisms allowing survival in extreme environments, implications
for organisms growing under sCO 2 ...
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
17
1.6 Acclimation mechanisms allowing survival in extreme environments,
implications for growing under scCO 2 .
. .. . .. . . . . . . . . . . . . . . . . . . . . .
1.7 Research questions .............................................
...
18
22
Chapter 2: Microbial growth under supercritical CO 2
27
2.1 Introduction ....................................................
30
2.2 M ethods .......................................................
32
2.3 Results ........................................................
42
2.4 D iscussion .....................................................
53
2.5 Acknowledgements ............................................
59
2.6 Supplemental Figures ............................................
60
Chapter 3: Draft genome sequences of the supercritical CO 2 tolerant bacteria
Bacillus subterraneus MITOTI and Bacillus cereus MIT0214
70
3.1 M ain text .......................................................
72
3.2 Acknowledgements ............................................
74
Chapter 4: Changes in lipid and proteome composition accompany growth of
Bacillus spp. under supercritical CO 2 and may promote acclimation to
75
associated stresses
4.1 Introduction ....................................................
7
78
4.2 M ethods .......................................................
82
R esults ........................................................
91
4.4 D iscussion .....................................................
112
4.5
119
4.3
Supplemental Figures and Tables ..................................
Chapter 5: Conclusions
133
References
136
8
Chapter 1: Review of supercritical CO 2 industries that may involve microbial activity and
possible survival strategies to scCO 2
Growth and survival of microorganisms under high pressure CO 2 has become an
increasingly important research topic with the use of supercritical phase CO 2 (scCO 2 ) as a
sterilizing agent, bioengineering in geologic carbon sequestration (GCS), and interest in carrying
out biocatalysis and product extraction with a scCO 2 solvent. Supercritical phase fluids are space
filling, similar to a gas, but they are denser and able to solvate similar to a liquid. These
properties help make scCO 2 an effective sterilizing agent, and an industrially important, nontoxic solvent, with growing applications involving scCO 2 phase biocatalysis. Large
anthropogenic point sources of CO 2 have led to increasing studies in reducing these greenhouse
gas emissions by capturing emitted CO 2 followed by injection underground for GCS, creating
subsurface environments where it is not yet known what the effects of microbial activity will be
on the fate of injected CO 2 . The potential for microbial activity in these environments opens up
possibilities for utilizing microorganisms in bioengineering solutions to enhance the permanence
of CO 2 storage. This thesis is the first demonstration and examination of microbial growth in the
aqueous phase under a scCO 2 headspace and includes analysis of potential acclimation
mechanisms that mediate growth under scCO2.
1.1 Microbial impacts on Geologic Carbon Sequestration
Geologic carbon sequestration has been increasingly cited as one solution to reduce
-
atmospheric emissions from large point sources of CO 2 . In GCS, carbon dioxide from large C0 2
producing sources is separated from other gases, pressurized, and injected into permeable
9
geologic formations where it remains trapped beneath an impermeable rock layer or cap-rock.
After injection, CO 2 remains a separate supercritical phase in most formations targeted for GCS
(Benson et al., 2005), with supercritical CO 2 present at temperatures above 31 'C and pressures
above 72.8 atm. The major fates of injected CO 2 are: (i) remain as a supercritical phase between
the aqueous phase and cap-rock or in residually trapped pockets, (ii) dissolve into the aqueous
phase and (iii) mineralize into carbonates and precipitate (Benson et al., 2005). The idea of
geologic CO 2 storage is not without precedent, as there are stable, natural geologic
accumulations of CO 2 (Watson et al., 2004), but leakage rates from the GCS sites will need to
remain below 1% per thousand years to effectively reduce CO 2 emissions (Shaffer, 2010).
However, there are legitimate concerns of leakage through cap-rock fractures, especially from
injection-triggered earthquakes (Zoback and Gorelick, 2012) and along improperly sealed
wellbores (Kutchko et al., 2007; Watson et al., 2009). There is a need for more research into
microbial influences on CO 2 permanence and the potential for bioengineering solutions to
address leakage scenarios (Oldenburg et al., 2008).
The subsurface is one of the largest reservoirs of microbial biomass (Whitman et al.,
1998), and microbial communities are both present and active in deep subsurface environments
(Onstott, 2005; Kieft et al., 2005; Chapelle et al., 2002; Lavalleur and Colwell, 2013). These
active populations raise questions of how microbial activity may affect human activities like
enhanced oil recovery (EOR), fracking, or GCS in the subsurface, and whether microbes can be
harnessed for bioengineered applications within these contexts.
While geochemical modeling suggests biological CO 2 fixation is likely to be a negligible
factor relative to the massive quantities of injected CO 2 (Onstott, 2005), microbial activity can
affect CO 2 permanence in other ways. Microbial activity has been harnessed in in subsurface
10
bioremediation applications (Cunningham et al., 2003; Williams et al., 2005), and with the
possibility of CO 2 leakage from fractures and wellbores, there is a need for engineering solutions
to address these leakage pathways. One bioengineering proposal to enhance structural trapping is
the biofilm-barrier, which has been demonstrated to reduce to the flow of scCO 2 in a sandstone
core with pre-grown microbial biofilms (Mitchell et al., 2009). CO 2 permanence would also be
increased by the precipitation of CO 2 into carbonate minerals, and microbial activity can effect
the rate of precipitation by microbial mineral weathering which liberates metal cations necessary
for incorporation of CO 2 into carbonates (Mcmahon and Chapelle, 1991; Ferris et al., 1996;
Barker et al., 1998). Microbially induced precipitation of carbonates is well documented (Wright
and Oren, 2005; Mitchell and Ferris, 2006), which has led to the application of microbial mineral
plugging (Ferris and Stehmeier, 1992), to the context of GCS (Cunningham et al., 2009).
Engineered microbial mineral plugging (or biomineralization) could reduce rock porosity and
permeability through the addition of urea to injected CO 2 as an energy source for
microorganisms, which will hydrolyze urea, resulting in increased pH and subsequent increased
CaCO 3 precipitation (Mitchell et al., 2010; Cunningham et al., 2011; Phillips et al., 2012;
Cunningham et al., 2013). While the use of biofilms and biomineralization to impede scCO 2 flow
shows promise, current literature is limited to lower, sub-critical CO 2 pressures (Mitchell et al.,
.
2010; Phillips et al., 2012), which do not have the same inhibitory effects as scCO 2
Additionally, the greatly reduced growth rates in these deep subsurface environments (Phelps et
al., 1994), may present a major obstacle in developing bioengineering solutions in situ.
11
1.2 Microbial activity in high-pressure, C0 2-rich environments
For bioengineering solutions to be realized, microorganisms will need to remain active in
these deep formations after CO 2 injection. Theoretical studies indicate microbial processes are
thermodynamically favorable in deep subsurface formations both before and after CO 2 injection
(Onstott, 2005; Kirk, 2011). A study of the Ketzin CO 2 sequestration site in Germany observed
that after an initial decrease in cell numbers (> 2 log orders) and a shift toward Archaea for
several months after CO 2 injection, the total cell numbers rebounded and microbial community
composition shifted to be dominated by sulfate reducing bacteria, suggesting the presence of an
active community acclimating to near-critical levels of CO 2 (Morozova et al., 2011). A recent
study of microbial diversity from the Otway Basin CO 2 sequestration site also found changes in
community composition in formation water after CO 2 injection, with the population shifting
from Firmicutes to Proteobacteria, again suggesting that at least a subset of these microbial
populations are acclimating to (or differentially surviving) CO 2 injection, although changes in
biomass were not reported in this study (Mu et al., 2014).
In addition to studies of carbon sequestration sites, marine seeps may form CO 2 rich
environments with active microbial populations. The Okinawa Trough off the coast of Japan and
Taiwan is a hydrothermal system with active CO 2 seeps in sediments, which include liquid and
hydrate CO 2 . Community composition at the CO 2 - sediment interface was primarily composed
of methanotrophic Archaea and chemolithotrophic Epsilonproteobacteria (Inagaki et al., 2006).
A more recent study of the Okinawa Trough verified that microbial communities in these CO 2
rich sediments are active, through reverse transcribed 16S ribosomal RNA, with increasing
fractions of Deltaproteobacteria and Euryarchaeota in deeper sediments with higher CO 2
concentrations (Yanagawa et al., 2012). Microbial abundance decreased sharply in sediments at
12
the interface with the liquid CO 2 (from 109 to 107 cm- 3 direct cell counts), with decreasing
diversity in sediments with CO 2 concentration. These studies on CO 2 rich environments indicate
that high concentrations of CO 2 will alter the community composition, and in some cases result
in decreased biomass, suggesting that some microbial population may not easily acclimate to
.
high concentrations of CO 2
1.3 Microbial sterilization and survival under scCO 2
It is well documented that high pressure C0 2 , particularly in supercritical phase is an
effective sterilizing agent, which has been utilized by food and biomedical industries for over a
decade as a lower temperature alternative to traditional heat-based sterilization methods.
Supercritical CO 2 is effective at sterilizing a diverse range of vegetative cells, including fungi,
gram-positive, and gram-negative bacteria (Garcia-Gonzalez et al., 2009; Dillow et al., 1999;
Zhang et al., 2006). In yeast strains, cell death in response to CO 2 follows a sigmoidal curve with
increasing CO 2 pressure, with major decreases in viability occurring at phase changes of CO 2
(from gas to liquid and liquid to supercritical fluid) (Isenschmid et al., 1995). The mechanism by
which scCO 2 inactivates cells is theorized to result from a combination of factors where the
initial step is the influx of high concentrations of dissolved CO 2 which will permeabilize cell
membranes (Tamburini et al., 2014; Spilimbergo et al., 2008; Zhang et al., 2006; Hong and
Pyun, 1999), in part due to the lipid disordering effects of high concentrations of dissolved gases
(Chin et al., 1976). Following permeabilization, cytoplasmic acidification will occur (Tamburini
et al., 2014; Spilimbergo et al., 2008; Zhang et al., 2006; Hong and Pyun, 1999), and some
microorganisms (especially gram negative species) will experience cell wall collapse (Oule et al.,
2006; Dillow et al., 1999). Finally, inactivation of enzymes and leakage of intracellular contents
13
via scCO 2 extraction (Bertoloni et al., 2006; Kim et al., 2008), and in some cases physical cell
rupture will occur (Oule et al., 2006). It is important to note that some effects of scCO 2
sterilization may be partly effects of decompression, as extremely rapid depressurization (from
260 atm to 1 atm) will lyse cells; however depressurization over 5 min time scales results in
much lower lysis rates (Park and Clark, 2002). Figure 1 is a schematic of the mechanisms of
sCCO 2 sterilization that combines membrane permeabilization, cytoplasm acidification, scCO 2
extraction, and cell rupture.
SC
2
Co 2
Aqueous
0
00
..-0 0
.0
-.
.-
-
-
*
*
..
.~
HCO~ + H+ = H 2CO~
0
0
0..
*0
~CO2 + H2
je
.
.
0
1t
4W
0
Cytosol
Itceulrpoin
Figure 1. Conceptual model of scCO 2 sterilization documenting processes of membrane
permeabilization, cytoplasm acidification, supereritical extraction, and cell rupture. CO2 is a
small, uncharged molecule that readily diffuses across the cell lipid membrane. High cytoplasmic
CO2 may inhibit metabolic processes and lead to acidification due to carbonate chemistry. Nonpolar components of cells may be extracted by scCO2 , and partition into the C0 2 -phase.
14
While vegetative cells show high degrees of sterilization upon scCO 2 exposure,
microorganisms residing in biofilms (Mitchell et al., 2008) and spores show resistance to scCO 2
sterilization (Kamihira et al., 1987; Dillow et al., 1999; Ballestra and Cuq, 1998; Enomoto et al.,
1997; Watanabe et al., 2003; Zhang et al., 2006b). Sterilization of spores by scCO 2 often requires
additional methods including extended exposure time, higher temperatures, pressure cycling, and
addition of co-solvents or oxidizing agents (Ishikawa et al., 1997; Watanabe et al., 2003; Shieh et
al., 2009; Zhang et al., 2006a,b). Recent work suggest that additional factors such as mineral
matrices may enhance microbial survival to scCO 2 exposure by providing substrates for biofilm
formation and/or by creating buffered microenvironments (Wu, 2010; Santillan, 2013). However,
these studies were conducted on shorter timescales (less than 20 hours) and most experiments
.
used sub-supercritical CO 2
In an effort to simulate CO 2 sequestration in a lab environment Frerichs et al. (2014)
incubated formation fluids from a natural gas well under a scCO 2 headspace and found that
during scCO 2 exposure, microbial populations were static with no apparent activity, but upon
removal of scCO 2 , there was an outgrowth of spore-forming Clostridialeswith active sulfate
reduction. These studies, and the extensive sterilization literature indicate that certain
physiologies, including organisms capable of spore-formation will be more resistant to initial
scCO 2 exposure.
1.4 Biocatalysis under supercritical CO 2
Part of the mechanism that makes scCO 2 an effective sterilizing agent is its solvating
property, and this solvent capability also makes scCO 2 important industrially. ScCO 2 has been
15
used for some time as an environmentally friendly solvent for extracting chemicals (especially
from plants), but enzymes can often be inactivated due to effects of both scCO2 and
depressurization (Kasche et al., 1988). Despite this, research efforts have developed scCO 2 (and
other solvents) as non-aqueous solvents for enzyme catalysis, using both purified enzymes as
well as whole cells (Baiker, 1999). ScCO 2 enzyme catalysis allows for interesting applications
like the modification of compounds that are not soluble in the aqueous phase, as the exact
solvent properties of scCO 2 can be altered by changing temperature and pressure, and continuous
removal of catalysis products with immobilized enzymes (Wimmer and Zarevucka, 2010).
Examples of scCO 2 -phase catalysis include esterification reactions with lipase (Nakamura et al.,
1986; Marty et al., 1991), carboxylation of pyrrole with Bacillus megaterium cells (Matsuda et
al., 2001), reduction of ketones with Geotrichum candidum cells (Matsuda et al., 2008) and
hydrolysis of carboxymethyl cellulose with cellulase (Paljevac et al., 2007). These reactions are a
few of the multitude of enzymatic reactions demonstrated under supercritical CO 2, with many
more possibilities including alkylation, amination, isomerization, oxidation, and dehydrogenation
reactions (Baiker, 1999; Wimmer and Zarevucka, 2010). One major limitation to biocatalysis
under scCO 2 is the variable stability of enzymes under scCO 2 , and even more stable enzymes
eventually need to be replaced after they lose activity. Use of microorganisms capable of growth
in reactors containing scCO 2 , may help reduced these limitations by protecting the stability of
enzymes and allowing enzyme renewal through growth. The development of enzymatic catalysis
under scCO 2 , combined with apparent biologic activity under high pressure, high CO 2
environments as demonstrated in this thesis suggest that new possibilities are on the horizon for
supercritical catalysis.
16
1.5 Adaption mechanisms allowing survival in extreme environments, implications for
organisms growing under scCO 2
For microorganisms to be utilized in carbon sequestration bioengineering solutions or
industrial biocatalysis, we need a better understanding of the adaptations and acclimation
mechanisms employed by cells to survive and grow under stresses associated with supercritical
CO 2. Adaptations, or evolved changes in an organism, are frequently part of the explanation for
how a microbe survives and grows in a specific environment. Some physiological adaptations are
more specific to a subset of microorganisms such as the ability to form endospores, which is
largely observed in Firmicutes. The adaptation of many Firmicutes to form spores allows those
organisms to survive in a dormant state, withstanding high temperatures, desiccation, and many
other stresses. While Frerichs et al. (2014) suggested that the ability to form endospores was
crucial to the enrichment of Clostridialesafter scCO 2 incubations, Mu et al. (2014) observed an
increase in Proteobacteria after sCCO 2 injection into the deep subsurface, suggesting that non
spore forming organisms may also survive scCO 2 exposure.
Other adaptations are more broadly distributed across phyla, including the adaptive
regulation by Sigma factors, which regulate general transcription and transcriptional responses to
a variety of specific stresses including heat, cold, acid, starvation, nitrogen limitation and
pressure (Merrick et al., 1993; Gaidenko et al., 1998; Wemekamp-Kamphuis et al., 2004).
Specialized adaptations can also often be conferred through mobile genetic elements, such as the
ability to degrade xenobiotics that may be transferred by plasmids and transposons (Top and
Springael, 2003), or resistance to phage through CRISPRs (Barrangou et al., 2007).
Adaptations may also include changes that do not manifest in additional gene content.
Differences have been observed in the amino acid composition between halophiles and non-
17
halophiles (Paul et al., 2008), high and low pressure adapted microorganisms (Di Giulio, 2005;
Simonato et al., 2006), and mesophiles and thermophiles (McDonald et al., 1999). While specific
amino acids may be selected for during adaptation to different environments, there are also
specific gene changes that have been identified as adaptations to certain environments. For
example, coding changes in specific regions of malate dehydrogenase (Saito et al., 2006) and
16S rRNA changes (Lauro et al., 2007) have been observed in high-pressure adapted
microorganisms. With coding changes occurring throughout genomes, differences in the ratio of
nonsynonymous (coding) changes to synonymous (non-coding) changes (dN/dS or Ka/Ks), can
be used to determine specific genes that are under selection (McDonald and Kreitman, 1991).
With this method, Campanaro et al. (2008) identified numerous genes involved in solute
transport and nucleotide transport and metabolism are under selection in deep-sea bacteria.
Further investigations in using dN/dS have developed Selective Signature Analysis, which
improves on the traditional dN/dS to estimate genes under selection by controlling for mutation
rate variation in different genomes and gene families (Shapiro and Alm, 2008). However, while
novel gene content, changing amino acid content and coding changes may confer adaptations to
different environments, many survival strategies rely on non-heritable acclimation mechanisms.
1.6 Acclimation mechanisms allowing survival in extreme environments, implications for
organisms growing under scCO 2
The high pressure, solvent effects, and pH decrease accompanying scCO 2 exposure
suggest that microbes will need to adjust aspects of their physiology to acclimate. It is well
documented that changes in temperature, salinity, pH, and headspace all force cells to alter their
membrane lipid composition to maintain a liquid crystalline membrane that can regulate
18
osmolarity, intracellular pH, and membrane protein folding (Beales, 2004; Beranova et al., 2010;
Guerzoni et al., 2001; Kieft et al., 1994; Mukhopadhyay et al., 2006). Short duration exposures
to scCO 2, before inactivation occurs, of S. enterica and E coli results in few changes to lipid
acyl chains (Kim et al., 2009; Tamburini et al., 2014), but . coli does shows changes in lipid
head groups, with a reduction in phosphatidylglycerol lipids (Tamburini et al., 2014). However,
there are limited studies on lipid changes under scCO 2, as growth under scCO 2 has not been
documented until this thesis. Thus, examining previous studies that have focused on conditions
with some similarities to scCO 2 (e.g. pressure and acid stress) may help elucidate acclimation
mechanisms necessary for growth under scCO 2 . High-pressure gases (not just C0 2) have been
documented to alter membrane phospholipid ordering (Chin et al., 1976), providing evidence
that cells may alter their membranes in the presence of scCO 2 . High pressure conditions tend to
have the effect of compressing and decreasing fluidity of membranes, which cause bacteria to
compensate by producing more unsaturated lipids in order to maintain the fluidity of their
membranes (Kato and Hayashi, 1999). Similar to high pressures, cold temperatures decrease
membrane fluidity, and microbes acclimate to this by increasing the proportion of unsaturated
and/or branched lipids (particularly the anteiso form) that have lower melting points due less
dense packing (Miladi et al., 2013; Russell and Fukunaga, 1990; Beales, 2004; Klein et al.,
1999). Cells under cold conditions may also respond by decreasing the average chain length
(McGibbon and Russell, 1983).
The alteration of pH further complicates cell membrane adjustments. Acid stressed B.
subtilis will increase the rigidity of their membranes by producing fewer branched and
unsaturated lipids (Petrackova et al., 2010). The previous study also observed that cells
acclimated to acid stressed conditions had lower proton permeability, although is not possible to
19
determine if this was due to membrane diffusion or proton pumps. Clostridiumacetobutylicum
responds to pH reduction with a similar reduction in unsaturated lipids, but also increases the
lipids containing cyclopropane rings (LePage et al., 1987). Acid stressed Streptococcus mutans
have a different strategy, as they shift from shorter chain saturated fatty acids to longer chain
monounsaturated fatty acids, which would seem to increase membrane fluidity (Fozo and
Quivey, 2004). B. subtilis also alters membrane lipids under anaerobic conditions, which result
in increased chain length and increases in the anteiso to iso ratio when compared to aerobically
grown cells (Beranova et al., 2010). Another important class of membrane compounds is
hopanoids, which are 5 membered rings that may be involved in stress response. A
Rhodopseodomonaspalustris mutant that could not produce hopanoids was highly sensitive to
acid and alkali stress (Welander et al., 2009).
While membrane changes in response to scCO 2 exposure seem likely, microbial stress
responses include numerous changes in expression. Among the most common expression
changes across stresses is the upregulation of general and specific sigma factors (Browne and
Dowds, 2002; Foster, 1999; Ferreira et al., 2003; Gaidenko and Price, 1998). In B. subtilis, both
the general stress response transcription factor (sigmaB) and the sporulation transcription factor
(sigmaH) enhance survival under acid and alkaline stress (Gaidenko and Price, 1998). Listeria
monocytogenes also expresses sigmaB (Ferriera et al., 2003), while Salmonella typhimurium
expresses sigmaS in response to acid (Foster, 1995). Another acid tolerance mechanism
demonstrated in Streptococci is the increased expression of FoF1 ATPase, which pumps protons
out of the cell (Martin-Galiano et al., 2001). Additional transporters are also involved such as
glutamate and arginine transporters and decarboxylases, which transport amino acids into cells
20
and subsequently decarboxylate them, consuming protons in the process (Richard and Foster,
2004; Cotter et al., 2001). Amino acids may also be subject to deiminase enzymes that remove
carboxyl and amino groups resulting in ammonia and CO 2 which act to buffer the pH (Foster,
1999). Similarly, enzymes like arginase and urease produce the alkaline products ornithine and
urea, and ammonia, respectively which can aid in buffering intracellular pH (Chen et al., 1998;
Casiano-Colon et al., 1988; Curran et al., 1995; McGee et al., 1999). A recent study of D.
vulgaris exposed to high pressures of CO 2 observed upregulation of genes involved in
production of leucine and isoleucine before cells were inactivated (Wilkins et al., 2014),
indicating that amino acid metabolism may be important in a scCO 2 stress response, either for
neutralization of pH, or for use as compatible solutes for osmotic regulation (Csnonka, 1989).
Similar to acid stress, high-pressure stress also results in induction of sigma factors (Abe
et al., 1999), and various proteins with chaperone activity (Welch et al., 1993; Ishii et al., 2005).
Not surprisingly, organisms adapted to high pressures (barophiles), do not show elevated levels
of chaperone expression under high pressure, as those organisms are adapted to growing under
high pressure (Boonyaratanakornkit et al., 2007). However, elevated pressure is a different
physical stress than acidity, with different expression patterns, aside from general stress
responses like sigma factors and chaperone proteins. In a study of Lactobacillus sanfranciscensis
proteomes from heat, cold, acid, salt, starvation, and pressure stressed cells, high-pressure stress
shared more common expression patterns with cold and osmotic stresses than acid stress
(Hormann et al., 2006). In experiments with E coli under pressure stress, transcription,
translation, and nucleotide metabolism all show upregulation under pressure stress (Ishii et al.,
2005). Interestingly, barophilic organisms appear to show evidence of pressure-regulated
21
metabolism, as multiple barophiles have shown increases in certain respiratory proteins when
grown under pressure (Vezzi et al., 2005; Abe et al., 1999).
1.7 Research Questions
While it is clear that supercritical CO 2 presents a unique combination of stresses to
microorganisms, existing studies of high pressure, C0 2-rich environments suggest that
microorganisms are active in close proximity to pure-phase CO 2. I sought to determine if
microorganisms can in fact grow under a scCO 2 headspace, and to understand how
microorganisms survive and acclimate to grow under scCO 2 . The questions I address in this
thesis are:
1) Are microorganisms isolated from GCS sites capable of growth in bioreactors consisting of an
aqueous phase under a supercritical CO 2 headspace? What are the taxonomic identities and
physiological characteristics of recovered scCO 2-tolerant organisms (Chapter 2)
2) What are the genomic characteristics of scCO 2 tolerant isolates B. cereus MIT0214 and B.
subterraneusMITOT1? (Chapter 3)
3) What are the changes in membrane lipids and expressed proteins that encompass the
acclimation response to growth under scCO 2? (Chapter 4)
Regarding my first question (are isolates from GCS sites capable of growth in an aqueous
phase under a scCO 2 headspace?), I hypothesized that certain microorganisms will be able to
grow in the aqueous phase under a scCO 2 headspace. This hypothesis has been supported by
recent studies of CO 2 sequestration pilot sites that show changes in microbial community after
scCO 2 exposure (Mu et al., 2014; Morozova et al., 2011), and deep ocean CO 2 seeps that contain
active microbial populations in direct proximity to liquid and hydrate CO 2 (Yanagawa et al.,
22
2012). Second, while CO 2 does impose a range of stresses on vegetative cells, vegetative cells
tested in current literature (Tamburini et al., 2014; Wilkins et al., 2014) are not acclimated to
high concentrations of C0 2 , and the resistance of spores opens up the possibility that some
spores may eventually germinate. Third, microorganisms are incredibly diverse and able to grow
under a variety of stresses including: considerably higher temperatures (Takai et al., 2008) and
pressures (Boonyaratanakornkit et al., 2007) and far lower pH's (Johnson, 1998) than associated
with the critical point of CO 2 . Finally, modeling predicts that growth in scCO 2 containing
environments is thermodynamically favorable (Onstott, 2005; Kirk, 2011). To test this
hypothesis, I incubated microorganisms that I isolated through scCO 2 enrichments (notably only
Bacilli were isolated) and surface isolated strains in growth media under a scCO 2 headspace and
demonstrated that a variety of Bacilli (both subsurface and surface isolates) are able to grow
under a scCO 2 headspace. Additionally, I determined that microbial growth from spores under
sCCO 2 is stochastic and varies as a function of inocula density and incubation time.
To address my second research question, (what are the genomic characteristics of scCO 2
tolerant isolates B. cereus MIT0214 and B. subterraneusMITOTI) I hypothesized that
supercritical CO 2 tolerant organisms MIT0214 and MITOTI will show characteristics and
evidence of adaptations that would enhance their fitness in the deep subsurface. As both of these
organisms are capable of sporulation and were isolated from samples collected from the deep
subsurface through successive passages of scCO 2 enrichment passages, it stands to reason they
tolerate growth at different pressures, and can survive drastic environmental changes as a spore.
Additionally, the anoxic nature of the subsurface and the enrichment method in this thesis
suggests these strains may be able to grow anaerobically. These microorganisms may also
contain other adaptations of novel genes or differentially evolving genes unique to slow growth
23
in the nutrient poor subsurface (Phelps et al., 1994). To examine this hypothesis I sequenced and
annotated the genomes of scCO 2 tolerant isolates B. cereus MIT0214 and B. subterraneus
MITOT 1. I did not observe substantial differences between MIT0214 and closely related B.
cereus isolates from surface environments, as their genomes are highly similar. MITOTI is more
distantly related to genome-sequenced strains and contains more unique gene content than
MIT0214, with some content involved in respiratory processes that may enable the strain to
access alternative electron acceptors for anaerobic respiration.
To address my final research question (what are the changes in membrane lipids and
expressed proteins that encompass the acclimation response to growth under scCO 2 ?), I
hypothesized that MIT0214 and MITOT 1 will show changes in membrane lipids in response to
headspace conditions and in response to different pressures, but that these responses will not
necessarily manifest similarly for pressure and headspace. Among the acclimation responses
described above, one method that acid stressed cells use to acclimate is through reducing
membrane fluidity, via reduced branching (Petrackova et al., 2010). As the referenced study was
performed with B. subtilis, and both MIT0214 and MITOT 1 are Bacilli, I would hypothesize that
a similar response may be observed to CO 2 . As the response to pressure and cold is generally to
increase membrane fluidity, through increased unsaturated lipids (Kato and Hayashi, 1999) or
increased branching (Miladi et al., 2013), I would hypothesize that MIT0214 and MITOTI may
increase the fluidity of their membranes under elevated pressure. Based on the present literature,
the effects of CO 2 and pressure may be somewhat opposing each other, leading to a scCO 2
phenotype that is intermediate between only CO 2 and only high pressure.
To acclimate to scCO 2 , I hypothesized that protein expression will also change in
response to scCO 2 , in addition to membrane lipids. One clear intersection of lipids and proteins
24
is the cell membrane region, which is affected by both acid and pressure, albeit in different
manners. I would hypothesize that some cell membrane localized proteins will be involved in
acclimation to scCO 2 . Additionally, the effects of pressure and high CO 2 concentration are both
stresses that may upregulate general stress response mechanisms like sigma factors and
chaperone proteins (Browne and Dowds, 2002; Welch et al., 1993). Aside from those responses,
I would expect that other acclimation mechanisms will not necessarily be related to both CO 2
and pressure, as CO 2 stress (similar to acid stress) often involves various mechanisms to
neutralize the acid or pump protons out of the cell (Cotter and Hill, 2003). Depending on the
pressure tolerance of MIT0214 and MITOTI, the effects of elevated pressure may not be a stress
to these strains, as they were isolated from higher pressure, deep subsurface environments. To
examine this hypothesis, I analyzed lipids from MITOTI and MIT0214 and proteomes from
MITOTI under 1 atm and 100 atm pressures of N 2 and CO 2 headspaces. I observed that both
MIT0214 and MITOT 1 respond to CO 2 by reducing branched lipids and increasing average acyl
chain lengths, similar to patterns observed in Bacilli under acid stress. In proteomes of MITOTI
under these conditions, I observed similar profiles to other Bacillus proteomes. Proteomes
separated by headspace and pressure along the first two components of principal component
analysis and amino acid metabolic proteins including proteins in the glycine cleavage system
were enriched in CO 2 headspace samples.
My findings suggest that microbial growth in close proximity with supercritical CO 2 is
indeed possible, but that injection of scCO 2 into the subsurface may initially select for
organisms, particularly spore forming organisms, that can survive the associated stresses. For
bioengineering solutions (e.g. biofilm and biomineralized barriers to scCO 2 flow) to be realized
in deep subsurface environments, it will require that either native communities grow, or that non-
25
native scCO 2 tolerant microorganisms be injected. My results suggest that it may be
,
advantageous to select spore forming strains for bioengineering applications involving scCO 2
and Bacillus strains I have isolated may be potentially useful in these applications. The
importance of microbial growth under scCO 2 also extends to other areas of biotechnology (e.g.
biocatalysis) where scCO 2 is an important industrial solvent. Finally this thesis provides insight
into the physiological plasticity of the Bacillus genus, which might extend to closely related
spore-forming organisms, and expands the known range of conditions under which these
organisms can grow.
26
Chapter 2: Microbial growth under supercritical CO 2
Manuscript submitted to Applied and Environmental Microbiology (currently in review).
27
Microbial growth under supercritical CO 2
Kyle C. Peet', Adam J.E. Freedman, Hector H. Hernandez'*, Vanya Britto', Chris Boreham 2,3
Jonathan B. Ajo-Franklin 4 and Janelle R. Thompson, **
Department of Civil and Environmental Engineering, Massachusetts Institute of Technology,
Cambridge, MA, 02139
2 Geoscience
3 CO2CRC,
4 Earth
Australia, GPO Box 378, Canberra, ACT 2601, Australia.
GPO Box 463, Canberra,ACT 2601, Australia
Science Division, Lawrence Berkeley National Laboratory, #1 Cyclotron Rd. MS
74R0 120 Berkeley, CA 94720
* Microbial and Environmental Chemical Engineering Laboratory (MECEL), iEnergy Center,
Masdar Institute of Science and Technology, PO Box 54224, Abu Dhabi, United Arab Emirates.
** Corresponding
author:
Janelle R. Thompson. Department of Civil and Environmental Engineering, Massachusetts
Institute of Technology, Cambridge, MA, 02139. Telephone: 617.324.5268. Fax: 617.258.8850
Email: jthompson@mit.edu
Keywords: CO 2 Sequestration, GCS, supercritical C0 2, Bacillus
28
Abstract
Growth of microorganisms in environments containing CO 2 above its critical point is unexpected
due to a combination of deleterious effects including cytoplasmic acidification and membrane
destabilization. Thus, supercritical (sC) CO 2 is generally regarded as a sterilizing agent. We
report isolation of bacteria from three sites targeted for geologic carbon dioxide sequestration
(GCS) that are capable of growth in pressurized bioreactors containing scCO 2. Analysis of 16S
rRNA genes from scCO 2 enrichment cultures revealed microbial assemblages of varied
complexity including representatives of the genus Bacillus. Propagation of enrichment cultures
under scCO 2 headspace led to isolation of six strains corresponding to B. cereus, B.
subterraneus, B. amyloliquefaciens, B. safensis, and B. megaterium. Isolates are spore-forming,
facultative anaerobes, and capable of germination and growth under a scCO 2 headspace. In
addition to these isolates, several Bacillus type strains grew under scCO 2 , suggesting this may be
a shared feature of spore-forming Bacilli. Our results provide direct evidence of microbial
activity at the interface between scCO 2 and an aqueous phase. Since microbial activity can
influence the key mechanisms for permanent storage of sequestered CO 2 (i.e. structural, residual,
solubility, and mineral trapping), our work suggests that during GCS microorganisms may grow
and catalyze biological reactions that influence the fate and transport of CO 2 in the deep
subsurface.
29
2.1 Introduction
Geologic carbon dioxide sequestration (GCS) is an emerging strategy to abate CO 2
emissions associated with the burning of fossil fuels by capture, compression, and subsurface
injection of generated CO 2 (Lal, 2008; Metz et al., 2005). Although many subsurface geologic
formations targeted for storage of compressed CO 2 are known to be biologically active
environments (Chapelle et al., 2002; Kieft et al., 2005; Lavalleur and Colwell, 2013; Onstott,
2005), the extent to which biological processes may play a role in the fate and transport of CO 2
remains unknown (Gaus, 2010). CO 2 exists as a supercritical fluid (scCO 2 ) at the temperature
and pressures of the vast majority of reservoirs targeted for sequestration (i.e. >31 'C and 72.9
atm). Aqueous fluids in contact with scCO 2 may have dissolved CO 2 concentrations on the order
of 3 M (Table 1). ScCO 2 has generally been regarded as a microbial sterilizing agent due to a
combination of factors including cytoplasm acidification, increased CO 2 anion concentration,
osmotic stress, membrane permeabilization and leakage via CO 2 extraction, and physical cell
rupture (Bertoloni et al., 2006; Dillow et al., 1999; Hong and Pyun, 1999; Kamihira et al., 1987;
Spilimbergo et al., 2008; Wu et al., 2010; Zhang et al., 2006).
While there has been no direct evidence that microorganisms can sustain metabolic
activity and grow in environments containing scCO 2 , previous work indicates this possibility.
Survival of spores and biofilms after short-term sCCO 2 exposure (i.e., minutes to hours)
(Ballestra and Cuq, 1998; Dillow et al., 1999; Mitchell et al., 2008; Zhang et al., 2006) is welldocumented, and recent studies show that mineral matrices may enhance microbial survival to
scCO 2 exposure by providing substrates for biofilm formation and/or by creating buffered
microenvironments (Santillan et al., 2013; Wu et al., 2010). Biogeochemical models also suggest
30
that diverse forms of microbial metabolism are thermodynamically favorable under reservoir
conditions post-CO 2 injection (Kirk, 2011; Onstott, 2005). Furthermore, high-pressure
incubations to simulate reservoir conditions with elevated (but not supercritical) CO 2 have
)
demonstrated activity of acetoclastic methanogens (under 49.3 atm pressure with 86.4 mM C0 2
)
(Mayumi et al., 2013) and homo-acetogens (under 395 atm total pressure with 126 mM C0 2
(Ohtomo et al., 2013). Recently, field studies at the Ketzin CO 2 sequestration site in Germany
and the Otway Basin site in Australia provide evidence that changes in microbial community
composition occur following CO 2 injection, suggesting that a combination of processes including
differential survival, and possibly growth, occur in the subsurface after exposure to near- and
super-critical levels of CO 2 (Morozova et al., 2011; Mu et al., 2014).
Whether microorganisms survive and remain active post-CO 2 injection is relevant for
predicting the fate and stability of the injected CO 2 . Microbial activity can influence the various
trapping mechanisms that are crucial to permanent storage of sequestered CO 2 . Prior work has
documented that microbial biofilms can be employed to plug pore spaces and impede the flow of
sCCO 2 through sandstone cores, providing a means of "structural trapping" for a mobile CO 2
phase (Mitchell et al., 2008; 2009). Trapping of CO 2 residuals in pore spaces by capillary forces
(residual trapping) may be affected by biosurfactant effects on wetting (Jenneman et al., 1983).
Bacterial surfaces may provide sites for carbonate mineral nucleation, while bacterial activity
can increase the rate of mineral weathering and therefore liberate the metal cations necessary for
incorporation of CO 2 into carbonate minerals (mineral trapping) (Barker et al., 1998; Ferris et al.,
1996; McMahon and Chapelle, 1991). Finally, increased dissolution of CO 2 into an aqueous
31
phase (solubility trapping) has been demonstrated by pH increases induced by bacterial ureolysis
under high pCO 2 (Mitchell et al., 2010).
In this study we tested whether environmental microbes could be isolated with the ability
to survive and exhibit microbial activity (growth) during exposure to scCO 2 . We performed a
series of experimental enrichment cultures inoculated with subsurface fluid filtrate or well core
samples from three subsurface environments targeted for CO 2 sequestration: the Frio-2 site near
Liberty, TX (Hovorka et al., 2006), the Otway Basin site in Southeastern Australia (Dance et al.,
2009; Mu et al., 2014; Sharma et al., 2011), and the King Island site near Stockton, CA (Downey
and Clinkenbeard, 2010). These three sites are geologically attractive as prospective CO 2
injection zones because they consist of high porosity/permeability sandstone formations overlaid
by low-permeability sealing layers capable of structurally trapping buoyant scCO 2 in the
underlying zone. Enrichment cultivation was followed by isolation and characterization of strains
able to grow under scCO 2. Microbial growth under sCCO 2 is surprising given its inhibitory
properties, and indicates the possibility that microbial activities will influence CO 2 trapping
during geologic carbon dioxide sequestration.
2.2 Methods
Subsurface sample collection and storage. Samples from GCS sites were utilized as inocula for
microbial enrichment cultures using scCO 2 as the selective agent. Samples from the Frio-2 site
were collected as part of the Frio-2 project and shared courtesy of Dr. Tommy Phelps (Oak
Ridge National Laboratory). For sample collection, 10 to 20 L of formation fluids were collected
32
by U-tube from the Frio-2 CO 2 sequestration site near Liberty, TX before, during and after CO 2
injection (Freifeld et al., 2005; Hovorka et al., 2006). Frio-2 formation fluids from 1,528 to 1,534
m depth were filtered through borosilicate glass filters (nominal pore size 0.8 pm) and frozen on
site. Three samples screened for this study correspond to samples collected prior to CO 2
injection, 7.5 hours after injection and 372 days post-injection. Otway Basin samples consisted
of rock cores from 929 to 1,530 m depth from the Pemble, Paaratte and Skull Creek formations
(Sharma et al., 2011). Otway cores were collected as part of the CO2CRC project in Southeast
Australia (www.co2crc.com.au). The King Island core sample was obtained from the
Mokelumne River Formation at ~1447 m depth during drilling of the Citizen Green #1 deep
characterization well by the West Coast Regional Sequestration Partnership (WESTCARB), San
Joaquin County, CA. Rock cores were kept refrigerated at 4'C prior to analysis.
Enrichment cultivation. Inocula for enrichment cultures were prepared in an anaerobic glove
bag with 02 and H 2 monitor (95% CO 2 / 5% H 2) and added to 4 or 10 ml pressure vessels
containing a 50% volume of growth media (below). For Frio-2 samples, inocula consisted of 10
pL of hydrocarbon and particulate residue from the surface of glass fiber filters, collected with a
sterile scalpel. For Otway Basin cores, drilling fluid tracer penetration data was used to guide
collection from the core interior where contamination was least likely. The selected regions of
Otway cores were pulverized with a stainless steel mortar and pestle and 1 gram of crushed rock
was used as inoculum. For the King Island formation core, no tracer-free interior could be
identified, as the sediment was highly permeable and unconsolidated. Thus, a representative
sample from the center of the King Island core was used as inoculum and processed in the same
manner as Otway cores.
33
Media for enrichment cultivation of Frio-2 samples was modified GYP Sodium Acetate
Mineral Salts Broth (GYP) consisting of (in g/l) 2.0 glucose, 1.0 yeast extract, 1.0 tryptic
peptone, 1.0 sodium acetate, 0.2 MgSO 4 .7H 20, 0.01 NaCl, 0.01 MnSO 4 .4H 20, 0.01
FeSO 4 .7H 2 0. Both GYP and MS media were used for enrichment cultivation from the Otway
Basin and King Island cores with supplements targeting different microbial functional groups
added to MS medium (Colwell et al., 1997). MS medium consisted of (in g/l) 0.5 yeast extract,
0.5 tryptic peptone, 10.0 NaCl, 1.0 NH 4 Cl, 1.0 MgCl 2 .6H 2 0, 0.4 K2 HPO 4 , 0.4 CaCl 2, 0.0025
EDTA, 0.00025 CoCl 2 .6H20, 0.0005 MnCl 2 .4H20, 0.0005 FeSO 4.7H2 0, 0.0005 ZnCl 2, 0.0002
,
AlCl 3.6H 20, 0.00015 Na2 WoO 4.2H 20, 0.0001 NiSO 4.6H 2 0, 0.00005 H 2 SeO 3 , 0.00005 H 3 B0 3
and 0.00005 NaMoO 4 .2H2 0. MS medium supplements consisted of: 0.5 g/l glucose for
fermenters; 1.3 g/l MnO 2 , 2.14 g/l Fe(OH) 3 , and 1.64 g/l sodium acetate for metal reducers; 0.87
g/l K 2 SO 4 , 0.83 g/l FeSO 4, 0.82 g/l sodium acetate for sulfate reducers; or 1.3 ml trimethylamine
and 0.82 g/l sodium acetate to target methanogens (Colwell et al., 1997). Culture media were
added to serum bottles and degassed with a stream of 100% CO 2 or 100% N2 gas for 30 minutes
prior to pressurization. Na2 S (at 0.25 g/l), a reducing agent to maintain anaerobicity, and
resazurin (at 0.001 g/l), a visual redox indicator, were added to culture media. Following
inoculation, vessels were pressurized and incubated for 2-4 weeks at 37 'C for Frio-2 samples
and 37 and 60 'C for Otway Basin and King Island samples. In addition to the above media,
Luria-Bertani Broth (LB) (Difco) and LB agar were used for strain cultivation at 37C under
ambient aerobic conditions.
The pH of ambient and CO 2 saturated media was measured at 1 atm and 21 'C using an
Orion model 520A pH meter. The pH of media under a scCO 2 headspace was measured by
visualization of a pH indicator strip (EMD Chemicals) through the sapphire window of a 25 ml
34
view cell (Thar Technologies, 05422-2). In addition, PHREEQC Version 2 was used to predict
the equilibrium pH and potential precipitation of chemical species in the growth media (Table 1)
under a CO 2 or N 2 atmosphere and as a function of temperature and pressure. Thermodynamic
data was obtained from the Lawrence Livermore National Library (LLNL) database.
Table 1. Observed and predicted pH, and predicted CO2 concentration as a function
a
of GYP culture medium composition , headspace as, and pressure at 370C.
Predicted [CO 2]
Headspace
CO 2
c
b
CO 2
N2
Pressure (atm)
1
100
1
N2
100
Predicted pH
5.3
3.6
7.0
7.0
Observed pH
5.1
3.9 to 4.5
7.0
ND
b
(M)
0.026
2.7
0
0
ND: Not determined
2+
2+
+
a
Growth media components (g/kg): Acetate 0.72, Na 0.43 1, Cl 0.607, Fe 0.002, Mn 0.0025,
2+
22Mg 0.0196, SO4 0.08575, and S 0.1028. Buffering of pH by the acetate system in GYP
maintained the final pH above the value predicted for deionized water alone.
b
geochemical modeling performed in PhreeqC calling the LLNL database.
C
pH measured by Orion model 520A pH meter (P = 1 atm) or visualized at (P=120 atm) via
indicator strip (EMD Chemicals).
High-pressure incubation. Vessels for high-pressure growth (Supplemental Fig. 10) were 316
stainless steel HPLC column bodies or 316 stainless steel tubing (4 and 10 ml capacity). Vessels
were fitted with ball valves (Supelco) or quarter turn plug valves (Swagelok or Hylok). Vessels
were filled to
1/2
capacity (2 or 5 ml) with cultivation media, and following inoculation, the
headspace of the stainless steel culture vessels, representing 50% of the total vessel volume, was
pressurized at a rate of 2-3 atm min' to 100-136 atm with industrial grade N 2 gas (Airgas) or
with extraction grade CO 2 gas (Airgas) with a helium (He) head pressure such that the final gas
mixture was 97-99% CO 2 . Pressurized vessels were incubated in a 37'C warm room, shaking at
100 rpm to increase mixing of media and subsurface inocula. Following incubation, the vessels
35
were connected with 316 stainless steel tubing and fittings to Swagelok pressure gauges to
measure the final vessel pressure before samples were depressurized at a rate of 3-5 atm min
over approximately 30 min. Generally, culture vessels with initial headspace pressures of>100
atm lost between 5-25 atm of pressure, due to slow leakage through fittings, over the course of a
multiple-week incubation with greater losses associated with longer incubations. Unless
specifically noted, all vessel incubation data reported herein maintained scCO 2 headspace
pressures of >72.9 atm, the critical pressure for CO 2 mixed with <3% inert Helium at 37'C
(Roth, 1998), or for incubations at lower pressures Pfinal was > 70% of Pinitial. All pressures were
measured at room temperature (21 'C). Based on the ideal gas law we can estimate a maximum
pressure increase of 6% associated with incubation of reactors at 37'C, although since scCO 2 is a
non-ideal gas this may be an overestimate. Following depressurization, cultures were transferred
to an anaerobic chamber (Coy lab products) containing a 95% CO 2/5% H 2 atmosphere for subsampling and passaging. All pressure vessels and valves were cleaned and autoclaved between
uses, and high-pressure tubing was flushed before use with 10% bleach for 30 minutes, rinsed
with milliQ-H 20, rinsed with 100% ethanol, and dried with CO 2 gas.
Enrichment cultures from the Frio-2, Otway and King Island sites were serially passaged
by diluting 10% v/v of the previous culture in fresh growth media under a 95% CO 2 / 5% H 2
atmosphere, pressurizing to 120 atm with C0 2 , followed by incubation at 37'C. The contents of
enrichment vessels were analyzed for cell abundance at the end of passages and in the inocula
prior to incubations. Frio-2 passages 1-3 incubated for 15 days each, while passage 4 incubated
for 60 days; subsequent passages were incubated for 9 to 15 days (Table 2). Otway and King
Island passages were incubated for longer time periods (i.e. 1-2 months) to increase the
likelihood of cellular growth based on earlier observations. Samples from each passage were
36
subjected to microscopic enumeration and archived as a glycerol stocks at -80 0 C.
Enumeration of cell density. To quantify biomass, we used a combination of methods including
direct cell counts, viable cell counts and optical density. Cell density at the beginning and end of
incubations was determined by microscopic epifluorescent cell staining using 4',6-diamidino-2phenylindole (DAPI, Sigma) or SYTO9 (Invitrogen) with shaking for 10 minutes in the dark.
500 pl to 1 ml of sample was then filtered onto 25 mm, 0.2 pm pore size, black polycarbonate
filters (Nucleopore), followed by 2 washes with 1 ml of Phosphate Buffered Saline (PBS). PBS
was incubated on the filtered sample for 1 minute to help wash off excess dye. Filters were laid
on slides under microscope immersion oil with a cover slip (Thermo Scientific), and were stored
at 4'C in the dark until counting. The cell density (in cells/ml) was calculated by multiplying the
mean cell counts (in one 10 x 10 microscope grid) by the dilution factor and then by 3.46x10 4 (as
one 10 x 10 grid at 1000X magnification corresponds to 1/3.46x10 4 of a 25 mm filter). Samples
were visualized on a Zeiss Axioplan fluorescent microscope. Images were captured on a Nikon
D100 camera using the NKRemote live imaging software. Viability counts, i.e. Colony Forming
Unit (CFU) plating was performed using Luria Broth Agar. Viable spore counts were carried out
by heating aliquots to 80'C for 10 minutes to kill vegetative cells (Setlow, 2006) prior to plating
on Luria Broth Agar. Culture optical density (600nm) was measured on a Bausch and Lomb
spectrophotometer (1 cm path length) or via 96-well microplate reader (BioTek Synergy 2) (200
pI per well). Optical density was not measured for incubations using metal reducer medium due
to confounded readings from solid particulate content.
For MIT0214, growth was defined by increased cell density and evidence of vegetative
cell morphologies by microscopy. Growth-positive cultures had at least 45-fold increased direct
37
cell counts relative to initial inocula less than lx 106 spores/ml, or at least 5-fold increase in direct
cell counts for cultures with inocula greater than lx106 spores/ml, as MIT0214 final cell
densities generally varied between 1x10 7 and 1x10 8 cells/ml. For MITOT1, growth was defined
by observation of culture turbidity accompanied by at least a 4-fold increase in viable cell counts
(CFU/ml) above the initial spore density (on average, increases were >50-fold) and at least 25fold higher viable counts compared to replicates without observed turbidity since samples
without evident growth all showed a decline in viable counts relative to the initial viable counts
of the inoculated cultures, presumably due to a loss of spore viability during the incubation
period.
Extraction of DNA and analysis of 16S rRNA gene diversity in enrichment cultures. DNA
extraction from Frio-2 sample enrichment passages was performed using a protocol modified for
gram-positive bacteria (Lessard et al., 2004). DNA extraction from Otway Basin project
passages was performed using two methods, the Qiagen Blood and Tissue DNA extraction kit
protocol for gram-positive cells (Qiagen), or the MoBio Soil DNA extraction kit (MoBio).
Amplification of 16S rRNA genes from extracted DNA was performed using universal Bacterial
primers 515F 5'- GTG CCA GCM GCC GCG GTA A- 3' and 1406R 5'-ACG GGC GGT GWG
TRC AA- 3' (Frio-2 passages 1, 2 and 7) and 27F 5'- AGA GTT TGA TCM TGG CTC AG- 3'
and 1492R 5'-TAC GGY TAC CTT GTT ACG ACT T- 3' (Frio-2 passage 9, Otway Passage 3,
and colony-purified isolates). PCR mixtures (20 gL per reaction) contained 25 to 75 ng of
genomic DNA, IX Phusion Polymerase buffer, 0.4 gM of each primer (IDT), 0.4 pM
deoxynucleotide mixture and 1 U Phusion Polymerase (New England Biolabs). Thermal cycling
conditions consisted of an initial 3 minutes at 95 C followed by 35 cycles of 95C for 30 sec,
38
52'C for 30 see, and 72'C for 90 sec; followed by a final extension time of 5 min. Every PCR
reaction included negative and positive controls.
Amplified 16S rRNA gene fragments from Frio-2 samples were gel purified (Qiagen gel
extraction kit) and ligated into the pJET1.2 vector (Fermentas) according to manufacture's
protocol. Ligation products were transformed into E. coli DH5a or . coli Top 10 cells and
clones were selected for sequencing (using LacZ/IPTG plating). For the Frio-2 enrichment,
sequencing reactions were prepared using Big Dye Terminator 3.1 according to the
manufacturer's instructions and sequencing was performed on an ABI 3130 platform. Otway
project 16S rRNA gene fragments were gel purified (Qiagen gel extraction kit) and sequenced
commercially (Genewiz, Cambridge, MA). Removal of vector and primer sequences, and
manual editing and clustering of operational taxonomic units (OTUs) at 99% nucleotide identity
was performed using Sequencher 4.5 (Gene Codes Corp). Chimeric sequences were identified by
Chimera Check 2.7 (RDP II Database) software and removed from analysis. 16S rRNA gene
sequences obtained from passages of the scCO 2 enrichments from the Frio-2 and Otway Basin
sites and from all colony-purified isolates were uploaded to the Ribosomal Database Project
(RDP) server (Cole et al., 2009) for multiple sequence alignment using their weighted neighborjoining tree-building algorithm. Stability of the groupings was estimated by bootstrapping on 100
trees, and phylogenetic tree files were downloaded and visualized with MABL (Dereeper et al.,
2008).
Isolation of strains and preparation of spores. Samples from passage 2, 5, and 3 of the
enrichment cultures from Frio-2 (sample 9-26-1039-20L), Otway Basin (core 3), and the King
39
Island core, respectively were plated on LB agar and incubated aerobically at ambient pressure at
37'C to obtain colonies, followed by colony purification by re-streaking on LB agar and
identification using DNA extraction and 16S rRNA gene sequencing. Since isolates had 16S
rRNA sequence types matching cloned sequences corresponding to endospore forming bacteria,
spores were prepared as described in (Kim and Goepfert, 1974) to serve as inocula for
subsequent characterization. For spore preparation, overnight stationary-phase cultures grown
under aerobic ambient conditions in LB medium were diluted 1:50 into Modified G Medium
which consists of the following (in g/l): yeast extract 2.0, CaCI 2 .2H20, 0.025, K 2HPO 4 0.5,
MgSO 4.7H 20 0.2, MnSO 4 .4H2 0 0.05, ZnSO 4 .7H20 0.005, CuSO 4 .5H2 0 0.005, FeSO 4 .7H 20
0.0005, (NH 4 )2 SO4 2.0, adjusted to pH 7.1 after autoclaving. Cells were incubated shaking for 72
hours to induce sporulation, then centrifuged for 15 minutes at 4000 X g. Samples were
resuspended and centrifuged 5 times in autoclaved wash buffer containing 0.058 g/l
NaH 2PO 4 .H20 and 0.155 g/l Na 2HPO 4.7H 20 with 0.0 1% (v/v) Tween20 to prevent aggregation
of spores. Spore preparations were heat treated at 80'C for 10 minutes to kill residual vegetative
cells. Spores for each isolated strain were stored in wash buffer at 4'C until further use.
Physiological characterization. Physiological tests for isolates displayed in Table 3 were
conducted in triplicate. LB media was used for assaying temperature, pH, and salinity ranges,
with positive growth scored by an OD 600nm of greater than 0.05. Media pH was adjusted with
NaOH or HCl followed by incubation at 37'C. Salinity was adjusted with NaCl in LB medium
followed by incubation at 37'C. LB agar was used for colony morphology determination after 20
hours incubation at 37'C. Spore formation was deternined by confirming viability after heat-
40
killing cultures grown in Modified G sporulation medium, and microscopic evaluation after
staining for spores (Ashby, 1938).
Measuring growth and survival in anaerobic CO 2 and N 2 atmospheres at variable pressure.
The dynamics of growth and survival of strains isolated from CO 2 sequestration sites and the
type strains Bacillus subtilis PY79, Bacillus mojavensis JF-2 (ATCC 39307) and Bacillus cereus
(ATCC 14579) were investigated under variable headspace gas composition and pressures.
Inocula consisted of either cells or spores; vegetative cell cultures were grown overnight to
stationary phase under aerobic conditions and diluted approximately 100-fold to 10 7 cells/ml in
fresh growth media while spores maintained in wash buffer were diluted to approximately 105
spores/ml unless otherwise specified. Media for MIT0214 and MITOT 1 were GYP and MS
media with the metal reducer supplement, respectively. Incubations at pressures from 3 to 136
atm were conducted in 316 stainless steel pressure vessels at 37'C, shaking at 100 rpm.
Pressurization was achieved by regulating backpressure from the C0 2/He tank. Determination of
growth dynamics under 1 atm N 2 or 1 atm 95%CO 2 / 5%H 2 was conducted in 100 ml serum
bottles at 37'C, shaking at 100 rpm.
To test whether Bacillus spores could tolerate indirect or direct exposure to scCO 2 , spores
of B. cereus MITO214 (Frio-2 isolate), B. subtilis PY79 and B. mojavensis JF-2 were aliquoted
into Durham vials in equal volumes of spore storage buffer corresponding to 5x 1 07, 7x1 05 and
108 spores, respectively, and dried for 2-3 days at 70'C to achieve a desiccated state. Dried
spores were either covered in 2 ml of fresh spore storage buffer or left desiccated, then incubated
under scCO 2 (100 atm, 37C) in a IL High Pressure Equipment Company (HIP) vessel for two
41
weeks. After depressurization, samples were plated for viable counts on LB media. Desiccated
samples were resuspended in 2 ml of buffer before plating.
Antibiotic-amended control experiments were used to confirm growth under scCO 2. We
amended reactors containing spores with both Kanamycin and Chloramphenicol antibiotics at
100 and 10 ptg/ml respectively in parallel with reactors without antibiotics. Additionally, we
included no cell controls for all experiments. Presence or absence of growth was assessed with a
combination of metrics including direct cell counts, changes in cell morphology from spores to
vegetative cells, viable cell counts, and optical density.
Statistical methods. Logistic regression analyses were performed using JMP Pro v. 10 where
growth outcome (growth/no-growth) was the dependent variable and incubation time and
inoculating spore density were the independent variables. Student's T-Tests were performed in
Microsoft Excel.
2.3 Results
Enrichment cultivation under scCO 2
Enrichment cultivation of subsurface samples in bioreactors containing nutrients, water
and scCO 2 yielded isolates from all three GCS sites examined in this study (Fig. 1; Table 2).
Enrichment cultures inoculated with filtered formation fluid (Frio-2) or crushed rock cores
(Otway or King Island) were evaluated for growth under a scCO 2 headspace after 14-30 days as
evidenced in some cases by an increase in the turbidity of culture media. All samples were
evaluated by epifluorescence microscopy to confirm biomass in turbid cultures and to screen for
42
-
,
Figure 1. Epiflourescent
microscopy of enrichment
cultures grown under scCO 2
filter concentrated onto 25 mm
0.2 prm pore size membranes,
stained with DAPI (A) or SYTO9
(B-F). Scale bars correspond to
10 pm. (A) Cells from Frio-2
initial enrichment. (B) Cells from
Frio-2 Passage 1 counter stained
with Propidium Iodide (red) to
identify membrane-compromised
cells). (C) Cells from Otway core
3, Passage 3 showing larger
vegetative cells and smaller cells
that may be spores observed
within the same sample. Black
particles (denoted by white
arrowheads) correspond to metal
solids from the growth medium
for metal reducers, also
observable in panel F. (D) King
Island initial enrichment, Metal
Reducer media. Images (E-H)
correspond to cultures of
MIT0214, MITOTl, WMe2, and
WG1 respectively. E-H were
inoculated with spores, with a
mixture of mostly vegetative cells
and some spores visible after 30
days growth under scCO
2
microbial cells in non-turbid samples (Fig. 1). Microscopy and extraction of total nucleic acids
confirmed biomass generation after successive rounds of dilution and growth (Table 2).
43
Table 2. ScCO 2 enrichment cultivation summary
Enrichment/Passage
I
I
Initial
1
1
1
2
37 1 60 I1 37 1 60 137 1 60 1 37 3 60
Temperature (*C)
Frio 2; Formation water particles; 1500 m; [days: 15, 14, 15, 15]
GYP
Ii+++IIND
+++
+++
+++ IIL Ii.I.IIM MIT0214
Otway Core 2; 1164.12-1164.33 m; Silty claystone [days: 21, 33, 61, 41]
+
MS + Fermenter
+
I
+
MS + Methanogen
-
+
GYP
-
+
+
-
+
+
Otway Core 3; 1193.59-1193.69 m; Sandstone [days: 14, 28, 81, 40]
+
+
+
+
MS + Fermenter
-
C,)
-
-
Cu
MS + Metal Reducer
MS + Sulfate Reducer
MS + Methanogen
GYP
0
-
-
+
.++
MITOT1
+++
-
-
-
Otway Core 4; 1239.31-1239.48 m; Silty claystone [days: 28, 81, 41, NDJ
+
+
+
+
MS + Fermenter
+
+
+
MS + Metal Reducer
+
+
MS + Sulfate Reducer
-
C.)
+
+
-
a)
E
MS + Metal Reducer
MS + Sulfate Reducer
-
Ci)
Cu)
CL
0)
0
CL
MS + Methanogen
-
+
0
GYP
-
-
C:
Cu
-
-
~0
Otway Core 5; 1273.86 - 1273.95 m; Sandstone [days, 29, 42, 49, ND]
+
+
MS + Fermenter
E
MS + Metal Reducer
MS + Sulfate Reducer
MS + Methanogen
CO)
-
ECu
GYP
(9
King Island Core; 1447 m; Unconsolidated sandstone [days: 29, 38, 49, 37]
++
++
+
+
+
MS + Fermenter
MS + Metal Reducer
MS + Sulfate Reducer
+
-
-
-
0
MS + Methanogen
+
-
+
+++
+++
WMR1
+
+++
+++
WMe1,
+
I
GYP
_WMe2
WG1
+
* Growth media consisted of GYP or basal MS media with supplementation targeting various microbial
metabolic groups (see Methods).
** Strain identification by 16S rRNA gene BlastN: MIT0214, B. cereus; MITOTI, B. subterraneus;
WMR I, B. safensis; WMe 1, B. safensis; WMe2, B. megaterium; WG 1, B. amyloliquefaciens.
Abbreviations: ND - Not determined
- cells not observed above microscopy detection limit (<1700 cells/ml)
+ biomass observed at <I E4 cells/ml
++ biomass observed at 1E4 to I E5 cells/ml
++ biomass observed at >1 E5 cells/ml
44
The composition of microbial assemblages enriched under a scCO 2 atmosphere was
determined by analysis of cloned 16S rRNA gene sequences for enrichments from the Frio-2 and
Otway samples achieving cell biomass greater than approximately 105 cells/ml. The libraries
from the initial enrichment and first passage of the Frio-2 sample revealed a mixed community
of microbes dominated by members of the Bacillus genera (Supplemental Fig. 1). DNA isolated
from Frio-2 passages 7 and 9 revealed a single dominant Bacillus sequence type sharing 99.8%
16S rRNA sequence identity (1463/1465 nucleotides) with B. cereus ATCC 14579
(Supplemental Fig. 1).
Biomass was detected in enrichment cultures inoculated with four different core samples
from the Otway Basin under a variety of media and environmental conditions (Table 2).
However, sparse cell densities (<104 cells/ml) did not yield PCR-amplifiable DNA. Subsequent
rounds of dilution and incubation in scCO 2 bioreactors were performed leading to increased
biomass consisting of a mixture of vegetative and spore-like cells for core #3 in MS medium
with an oxidized metal supplement (Fig. 1 C, Table 2). The 16S rRNA gene was amplified and
cloned from DNA extracted from passage 3 of Otway core #3, yielding 26 clones of a single
ribotype matching B. subterraneuswith 98.5% identity (1502/1521 nucleotides) (Table 3,
Supplemental Fig. 2). ScCO 2 enrichment cultures inoculated with material from the King Island
core revealed positive growth in 5 out of 8 incubations (Table 2). 16S rRNA clone libraries of
King Island enrichments were not pursued as the core was unconsolidated and thus fully
permeated by drilling tracer fluid which likely contaminated the microbial diversity. The 16S
rRNA gene of isolates from the scCO 2-enrichment of King Island core material was sequenced
following colony purification (Table 3).
45
Table 3. ScCO 2 enrichment isolate physiology
Bacillus sp. Bacillus sp. Bacillus sp. WG1
MITOT1
MIT0214
16S rRNA
B.
BLAST ID subterraneus
BLAT_ID (98.6%)
(98.6%)
Sandstone
B. cereus
B.
(99.8%) amyloliquefaciens
(99.2%) (99.1%) (99.4%)
Formation
Isolation
core, 1193.6
water,
Environment m, 39'C,
1528-1534
O Aut Basin, m,o2, TX
02
Facultative Facultative
anaerobe
requirement anaerobe
Bacillus sp.
WMR1
Bacillus sp.
WMe1
Bacillus sp.
WMe2
B. safensis
B. safensis
B. megaterium
(9.%(91)
(94)
Unconsolidated
Unconsolidated Unconsolidated Unconsolidated
sandstone core,
sandstone
sandstone
sandstone
1447 m, King
core, 1447 m, core, 1447 m, core, 1447 m,
King Island, CA King Island, CA King Island, CA
Island, CA
Facultative
anaerobe
Facultative
anaerobe
Facultative
anaerobe
Facultative
anaerobe
Spore
formation
Temperature
range ( C)
a
pH rangeb
Salinity
range (g/L)
c
Colony
morphology
23-37
23-45
23-55
23-55
23-45
23-45
4-8
4-10
2-10
4-12
4-12
4-10
0.5-50
0-40
0-60
0-60
0-60
0-60
0.2 mm,
circular,
entire,
raised,
smooth,
grayish
5 mm,
circular,
undulate,
raised,
smooth,
off-white
4 mm, circular, 2 mm, circular, 2 mm, circular, 2 mm, circular,
undulate,
entire, raised, entire, raised, entire, convex,
smooth, yellow-smooth, yellow- smooth yellowumbonate,
wrinkly, off-white
white
white
white
Colony purification and strain characterization
Since the Bacilli identified in clone libraries are known to be predominantly aerobic and
heterotrophic/copiotrophic, we attempted to isolate Bacillus strains as colonies on LB agar plates
from scCO 2 enrichment passages of all samples with significant growth (>105 cells/ml). When
growth under scCO 2 was observed, depressurized cultures were used as inocula and subsequent
colony formation under ambient aerobic conditions (37C) was observed after 1-3 days.
Individual colonies with distinct morphologies were colony-purified and identified by 16S rRNA
gene sequencing and corresponded to B. cereus (Frio-2 sample, passage 2 and 7, isolate
MITO214); B. subterraneus(Otway core 3, passage 5, isolate MITOT 1); and B.
amyloliquefaciens, B. megaterium and B. safensis (2 strains) (King Island core, passage 3,
46
isolates WGl, WMe2, WMel, and WMRl respectively) (Table 3, Supplemental Fig. 2). Isolates
were subjected to physiological characterization and were facultative anaerobes with variable
ranges of pH, temperature and salinity similar to previously characterized Bacilli (Table 3).
.
Isolates were subjected to re-growth experiments to confirm bacterial growth under scCO 2
Growth of Bacillus strains MIT0214 and MITOTI and Bacillus type strains under scCO 2
Based on our observations of biomass increases in enrichment cultures, spore-formation
in isolated strains, and the lethal effects of scCO 2 on vegetative cells of isolates and type strains
grown under ambient conditions (Supplemental Figs. 3 and 4), we hypothesized that spores but
not vegetative cells would germinate and grow under scCO 2. To test our hypothesis we
inoculated growth media with a range of spore concentrations (2.4x10 4 to 9.9x10 6 spores/ml for
isolate MIT0214 and 6.9x10 3 to 5.6x106 spores/ml for isolate MITOTi), followed by incubation
under scCO 2 from 1 week to 49 days (Fig. 2), accompanied by antibiotic (Supplemental Fig. 5)
and no-cell controls. As we observed highly variable outcomes (i.e. growth/no growth) in
replicate cultures of MIT0214 and MITOTI incubated under a scCO 2 headspace, we used
logistic regression (Fig. 2A and B) to determine the relationship between frequency of growth
under scCO 2 , incubation time, and initial spore density. For >20 day incubations of strain
MIT0214, 26/78 cultures exhibited increased cell density (average final cell density of 5.6x1 07
cells/ml), while 0/15 no cell controls and 0/30 antibiotic controls displayed growth (Fig. 2A,
Supplemental Fig. 5). For >20 day incubations of strain MITOTl, 32/58 cultures exhibited
increased cell density (average final cell density of 1.2x1 07 cells/ml), while no increase in cell
density was seen for no cell controls (0/15) or antibiotic controls (0/4) (Fig. 2B, Supplemental
Fig. 5). For logistic regression analysis, no cells controls were input at half the detection level
47
A
B
7-
7
6-
6
1?
9
N
0
5
0
E
0
0
Os
0
0 0
S
4
0
5
006
E
01
00
0.7
0
aL
0
4-
0
0
0
3
0
3-
0
C:
0
o E] 0
So
2
2
6
4
C)
D
culture
76
2-
control
o
E E
1 3 5
8
1
10.2
0.0
0.8
0.6
0.4
Frequency of growth
1.0
20
25
0.0
0
0
10
5
15
30
I
35
n I
0
0.2
E
]
E
7
9
11
I
0.8
0.4
0.6
Frequency of growth
I
I
1
30
20
10
0
I
1.0
1
40
50
Incubation time (days)
Incubation time (days)
Z=
control
1
1
Z
8.505-0.121X-0.936Y
Figure 2. Logistic regressions of MIT0214 (A) and MITOTI (B) growth outcomes under scCO 2
show a significant increase in the frequency of observed growth with increasing incubation time
(MIT0214 p=0.0005; MITOTI p=0.023) and increasing density of spore inocula (MIT0214
p=0.045; MITOTI p=0.0014). Model results predicting frequency of growth are represented by
contour lines. Experimental data (MIT0214 N=78; MITOTI N=73) are overlaid on top of the
contours with symbol shade indicating the frequency of growth and symbol size proportional to
number of replicates represented by each point. Circles are cell-containing samples, and squares
are no cell controls that were entered into the model at half the detection limit (i.e. at 870
cells/ml).
(i.e. at 870 cells/ml), while antibiotic controls were excluded from the model. As expected,
growth under scCO 2 is significantly affected by both initial inocula concentration and incubation
time (Fig. 2A: p=0.045 and p=0.0005 respectively for MIT0214, N=78; Fig. 2B: p=0.0014 and
p=0.023 respectively for MITOTI, N=73). We noted that growth of MITOTI consistently
accompanied more reduced conditions in growth medium that contained the resazurin redox
indicator dye and oxidized metal supplements (Supplemental Fig. 6). The incubation times
48
associated with a 50% likelihood of growth in cultures inoculated with at least lx104 spores/ml is
33 days for MIT0214 and 36 days for MITOTI.
Growth of Bacillus type strains and King Island isolates under scCO 2
After observing growth under scCO 2 of isolates MIT0214 and MITOT 1, we investigated
whether other Bacillus strains, including subsurface isolates from the King Island site and
additional Bacillus type strains could germinate and grow under scCO 2. Spores of three isolates
from the King Island core (WG1, WMel, and WMe2) and three Bacillus type strains B.
mojavensis JF-2, B. subtilis PY79, and B. cereus ATCC 14579 were inoculated into growth
media and incubated under scCO 2 for 30 days. King Island isolates and Bacillus type strains
yielded increased cell densities after incubation (Fig. 3A and B, respectively).
However, growth patterns were variable between replicates as previously observed during
experiments with strains MIT0214 and MITOTi (Fig. 2). A conceptual model for observed
growth variability of Bacilli cultures under scCO 2 is considered in the Discussion section. This
unexpected result appears to indicate that the ability to grow in environments containing scCO 2
may be widespread among Bacilli.
49
1.E+08
A
B
1.E+08
E
U)1.E+07
0
-E
1. E+07
L
-61.E+06
1.E+06
1
E+05
(-3'0
--
-
1. E+05
WWW
G)
B. subtilis PY79
B. mojavensis JF-2
B. cereusATCC 14579
Figure 3. Microscopic enumeration of cell abundance in cultures grown under scCO 2 for (A)
King Island isolates and (B) three Bacillus type strains (B. subtilis PY79, B. mojavensis JF-2 and
B. cereus ATCC # 14579). Data shown is total direct cell counts after 30 days incubation, with
5
horizontal lines representing initial cell counts. Initial spore densities ranged from 3.2x10 to
5
9.5x10 spores/ml.
Variable survival of cells and spores after exposure to scCO 2
In contrast to the observed germination and growth of spores in bioreactors containing
sCCO 2 , vegetative cells grown under aerobic conditions and ambient pressure were not able to
acclimate to scCO 2 exposure and grow in our experimental system. Vegetative cells from
overnight stationary phase cultures of isolate B. cereus MIT0214, and type strains B. cereus
ATCC 14579, B. subtilis PY79, and B. mojavensis JF-2, were exposed to scCO 2 for 6 hours at
370 C. 4-8 orders of magnitude of reduction in viable cell counts were observed for all strains
exposed to scCO2 relative to controls incubated at ambient conditions (Supplemental Fig. 3 and
50
cells surviving heat-kill treatments (Supplemental Fig. 4).
Spores from B. cereus MIT0214, B. subtilis PY79, and B. mojavensis JF-2 were robust to
scCO 2 exposure under both aqueous and desiccated conditions. Spores that were dried and then
resuspended in aqueous buffer and exposed to scCO 2 showed no significant change in viability
after 2 weeks relative to the initial viable count or controls incubated under ambient conditions
(p>0.05) (Fig. 4A). While dried spores directly exposed to scCO 2 were more susceptible to the
10
A
A scCO 2
10
o Ambient
1-
0
1
o0
-
A scCO 2
0 Ambient
00
o
E
E
CU
B
0.1
moavnis
C
P79moaeni
P7
.
.1(\.-
4). Cells surviving exposure to scCO 2 were most likely spores based on similar proportions of
U-6
0.01
0.01
B.
mojavensis
JF-2
B. subtilis
PY79
MIT0214
B.
mojavensis
JF-2
B. subtilis
PY79
MIT0214
Figure 4. Change in viable cell count (CFU/ml final/initial) in spores from three Bacillus species
exposed to scCO 2 for 2 weeks. (A) Spores resuspended in spore-preparation buffer (dark
triangles) revealed no decrease in viability upon exposure to scCO 2 relative to starting spore
counts and ambient atmosphere incubated controls (open circles). B. subtilis spores may have
germinated and grown during the incubation periods under ambient conditions. (B) Spores
exposed to dry scCO 2 (dark triangles) decreased in viability by 66-86% relative to ambient
atmosphere incubated controls (open circles).
killing effect, 14-24% of spores remained viable after two weeks relative to ambient-incubated
51
controls for B. mojavensis and B. cereus (p<0.05) and for B. subtilis (p=0.08) (Fig. 4B). Strain B.
cereus MITO214, isolated early in the project, was used as a model scCO 2 -tolerant isolate for
most analyses. Anecdotally, spores of strain B. subterraneusMITOTI appeared to be less robust
than spores of other Bacilli tested. In the spore-tolerance experiment the pre-treatment protocol
to desiccate the spores resulted in total loss in viability for MITOTI and thus no data is available
for this strain. Overall, this experiment indicated that spores from diverse Bacillus strains can
withstand both direct and indirect exposure to scCO 2, with enhanced survival of spores under
aqueous conditions (Fig. 4, Supplemental Fig. 7).
Growth of MIT0214 under variable CO 2 and N 2 pressure
We incubated strain MIT0214 under subcritical pressures (<72.9 atm) of N 2 and CO 2 to
observe whether more consistent growth outcomes (growth/no growth) would accompany more
permissive growth conditions. At 1 atm pressure, anaerobic growth under CO 2 or N 2 was
reproducible among triplicate incubations with increases in turbidity and viable cell counts
revealing lag, log, stationary and decline phases over a I week time frame (Supplemental Fig. 8A
and B). Considerably lower germination frequencies were observed under 1 atm CO 2 (<19%)
than 1 atm N 2 (>99%) (Supplemental Fig. 8A). In contrast to reproducible growth observed at
ambient pressure, growth at all pressures above 1 atm under either N 2 or CO 2 revealed positive
growth, but variability in outcome (i.e. growth/no-growth) between replicates over 1 and 2 week
time frames respectively, similar to the variability observed for cultures incubated under scCO 2
(Supplemental Fig. 9).
52
2.4 Discussion
We demonstrated that bacterial growth under a scCO2 atmosphere is possible using
strains isolated through enrichment cultivation of samples from three sites targeted for geologic
sequestration of C0 2 : filtrate samples from the Frio-2 project in Texas, consolidated rock cores
from the Paaratte formation in the Otway Basin Australia, and an unconsolidated core from the
Mokelumne formation at the King Island site in California. In addition, we observed germination
and growth under scCO 2 of three Bacillus type strains isolated from subsurface (B. mojavensis
JF-2) and surface (B. subtilis PY79 and B. cereus ATCC 14579) environments. These
observations of microbial growth under scCO 2 have several implications for geological carbon
dioxide sequestration (GCS): (1) Microbial survival and activity is possible, and perhaps likely,
at the CO 2 plume / water interface, even in areas previously exposed to pure-phase scCO 2 . (2)
Engineering biofilm barriers or stimulating biomineralization in situ at the C0 2-plume / water
interface may be feasible, despite previously documented lethal properties of scCO 2 to nonacclimated microorganisms. (3) Organisms likely to survive and proliferate after CO 2 injection
include spore-forming microbes that can withstand the initial CO 2 stresses.
Phylogenetic and physiological analyses reveal that strains isolated from the three GCS
sites are similar to previously described Bacillus isolates. Although enrichment media
formulations were designed to target diverse metabolic groups, all appeared to select for
facultative heterotrophs in the genus Bacilli and the form of anaerobic metabolism (fermentation
or respiration) has not been determined. Strain MIT0214 is most closely related to B. cereus - a
widely distributed bacterium that has been isolated from a diverse range of environments
including sites characterized by heavy metals, deep oil reservoirs and/or hypersaline conditions
(Pandey et al., 2011; Singh et al., 2010; Sriram et al., 2011; Xiong et al., 2009). Strain MITOTI
53
is closely related to strains B. subterraneusand B. infernus, both of which were isolated from
deep subsurface environments and are capable of metal respiration (Boone et al., 1995; Kanso et
al., 2002). King Island isolates are closely related to strains of B. megaterium, B. safensis and B.
amyloliquefaciens, which have all been isolated from diverse terrestrial environments (Idriss et
al., 2002; Mori et al., 1996; Raja and Omine, 2012; Satomi et al., 2006; Vary et al., 2007;
Yoshida et al., 2001). In light of the diversity of Bacillus strains able to grow under scCO 2 , we
hypothesize that the ability to grow under high pCO 2 is a widespread feature of Bacilli and
possibly other organisms.
All Bacilli isolated in this study were aerobic spore-formers capable of facultative
anaerobic growth under both N 2 and CO2 atmospheres. Notably, we have observed that all strains
exhibit fastest growth under aerobic conditions and atmospheric pressure, but are also able to
grow anaerobically at elevated pressures beyond the critical point for CO 2 (greater than 37'C and
73 atm). Growth dynamics of strain MIT0214 at 1 atm are consistent with anaerobic growth via
fermentation where end-product inhibition leads to a reduction of cell and spore viability in
stationary phase (Supplemental Fig. 8A and B). In strain MITOTI the observed change in culture
oxidation state that accompanied cell growth (as demonstrated by the resazurin indicator)
suggests that MITOTI may be capable of Fe(III) or Mn(IV) reduction (Supplemental Fig. 6)
although whether oxidation state changes occur via fermentation or respiration has not been
established in this study. We note that several close phylogenetic relatives of strain MITOT1 are
anaerobic metal-reducers capable of Fe(III), Mn(IV), Se(VI) and As(V) respiration (Boone et al.,
1995; Kanso et al., 2002; Yamamura et al., 2007).
Spores of three Bacillus strains including isolate B. cereus MIT0214 are tolerant of both
direct and indirect exposure to sCCO 2 . Direct exposure of dried spores to scCO 2 proved to be a
54
more severe stress resulting in a 66-86% loss in spore viability over a 2-week interval, while no
significant loss was observed in spores indirectly exposed via aqueous phase (Fig. 4,
Supplementary Fig. 7). The decrease in viability of spores directly exposed to scCO 2 may be due
to its solvating and desiccating properties, as spores in the aqueous phase do not experience the
same degree of contact with scCO 2 and are protected from dehydration. While it remains to be
determined whether the viability decrease is time-dependent, our observation of <1 order of
magnitude decrease in survival of desiccated spores exposed to scCO 2 suggests that spores in
subsurface environments will likely be able to tolerate at least brief periods of direct exposure to
pure-phase scCO 2 . This finding is especially relevant in the GCS context because as buoyant
sCCO 2 migrates toward the cap-rock interface, spores may experience fluctuations in scCO 2
contact in areas where formation fluids have been displaced, creating local areas of more
extreme and desiccating conditions. Persistence of a subpopulation of spores after direct scCO 2
exposure suggests that scCO 2 injection during GCS may not effectively sterilize the subsurface,
enabling the resumption of microbial activities upon return of an aqueous phase. Our results, and
previous demonstrations of spore resistance to scCO 2 sterilization (Zhang et al., 2006), suggest
that spore-forming microbial taxa may be able to survive and acclimate to in situ conditions postCO 2 injection during geologic carbon dioxide sequestration.
Previous work suggests that several features of Bacillus physiology are likely responsible
.
for the variable, but predictable, time- and density-dependent growth observed under scCO 2
First, germination frequencies under CO 2 headspaces are known to be significantly lower than
under aerobic or anaerobic nitrogen headspaces, with only 10-30% of B. cereus spores
germinating under 1 atm CO 2 (56). Growth dynamics for isolate B. cereus MIT0214
55
(Supplemental Fig. 8A) were consistent with these findings, displaying a germination frequency
at 1 atm CO 2 no greater than 19%, in contrast to a 99% germination frequency under 1 atm N 2
(Supplemental Fig. 8A). Secondly, it has been shown that certain stresses, such as pH and
oxidative, significantly extend the lag time for germination of Bacillus spores (Lee et al., 2003;
Setlow et al., 2013). We have observed extended lag phases for growth of MITO214 under 1 atm
CO 2 compared to 1 atm N 2 (Supplemental Fig. 8A and B) and have also observed extended lag
times for growth of all isolates at low pH (pH<6) (data not shown) during physiological tests
(Table 3). The third mechanism that may explain the observed variability in growth has been
termed the "Microbial Scout Hypothesis" where activation of cells from dormant bacterial
populations (e.g. spores) is stochastic, but occurs at a constant low frequency (i.e. ~0.01% per
day) independent of environmental cues. This behavior would allow a clonal population to sense
and respond to permissive growth conditions without risking activation of the entire dormant
population (Buerger et al., 2012). Finally, it has been shown that modulation of the sporulation
conditions has a significant effect on spore physiology and resistance to stress (Condon et al.,
1992; Russell, 1990) and we suggest that a population of spores may contain a range of this
physiological plasticity where some, but not all, spores are able to grow following germination
under stressful conditions. Furthermore, heterogeneity within mixed cultures due to factors
including adherent biofilms may lead to variable resistance to stresses associated with scCO 2
(Stewart and Franklin, 2008). Thus, variability in stress-tolerance of spores in combination with
extended lag times under conditions of high pCO 2 and low pH, and inherently stochastic and
suppressed germination frequencies, may account for observed variability of growth under
.
scCO 2
56
Microbial biomass obtained from ongoing GCS projects supports the notion that sporeforming Firmicute populations, such as Bacillus spp., persist and potentially grow in the
subsurface after sCCO 2 injection. Recovery of PLFA from Frio-2 groundwater samples indicated
the presence of viable microorganisms after CO 2 injection, especially Firmicutes, as indicated by
the observation of characteristic terminally branched saturated fatty acids. In particular, i15:0,
i 17:0, and al 7:0, which are prominent fatty acids in B. cereus (Haque and Russell, 2004;
Kaneda, 1972), were observed in the sample from the Frio-2 project from which B. cereus strain
MIT0214 was isolated, and in samples following CO 2 breakthrough (Personal Communication,
Susan Pfiffner). Microbial communities recovered in formation waters of the Otway Basin,
Australia and Ketzin, Germany GCS sites (10-12 days and 2 days-10 months post-CO 2 injection
respectively) include Firmicute taxa (e.g. Bacilli and Clostridia) among the detectable diversity.
Finally, in experimental incubation of formation fluids under scCO 2, enrichment of Firmicutes
(Clostridiales) was observed after removal of scCO 2 , demonstrating resilience of native
Firmicute populations to scCO 2 exposure (Frerichs et al., 2014).
While some Bacilli populations are well-established as members of subsurface biospheres
(Boone et al., 1995; Kanso et al., 2002), others may be introduced to the subsurface through
drilling activities. Although sampling methods for Frio-2 formation water and Otway cores were
designed to reduce the possibility of drilling fluid contamination through use of a U-tube
sampling device (Freifeld et al., 2005), or fluorescein drilling fluid penetration data, respectively,
drilling activities remain a potential source for some of the strains recovered in this study. Indeed
the King Island core, which yielded the highest number of scCO 2 -tolerant isolates, was
57
unconsolidated sandstone and fully penetrated by drilling fluid based on tracer data. Notably,
drilling fluid characterized by Mu et al. (Mu et al., 2014) was dominated by Bacilli sequences.
High concentrations of local dissolved organic carbon at the edges of injected CO 2
plumes, as observed during the Frio-2 project (Hovorka et al., 2006; Kharaka et al., 2006), may
stimulate growth of native or introduced bacteria. Enrichment of dissolved organic carbon near
the plume may be due to extraction of the subsurface organic matrix by the nonpolar solvent-like
properties of scCO2. We suggest these enrichments of dissolved organic carbon at the leading
edge of the CO 2 plume may serve as potential hot spots for microbial activity fueled by
anaerobic metabolism. Previous observations that DNA and lipid biomass from Firmicute
populations persist in scCO 2-exposed environments (Morozova et al., 2011; Mu et al., 2014),
coupled with the presented evidence that a diverse set of Bacillus strains and isolates may grow
in the presence of scCO 2 , water, and nutrients, suggests that microbially-mediated processes may
continue to occur up to the scCO 2 plume-water interface during geologic carbon dioxide
sequestration. Understanding how CO 2 injection shifts the balance of microbial community
structure, metabolic potential, and associated biogeochemical processes is therefore necessary to
.
predict the long-term fate of injected CO 2
58
2.5 Acknowledgements
The authors would like to thank Tommy Phelps and Susan Pfiffner for providing access to Frio-2
samples, sharing associated data, and for helpful comments throughout the project. We are
grateful to Peter Cook for arranging access to Otway samples through the CO2CRC project and
to Joanne Emerson for collection of core materials from the King Island site. In addition, we'd
like to thank Martin Polz for feedback on the manuscript, Eric D. Hill for advice on statistical
analyses, and Roger Summons and Mike Timko for project advice. Funding for experimental
work was provided to JRT by the Department of Energy Office of Fossil Energy under award
number DE-FE0002128 and by the MIT Energy Initiative. CJB published with the permission of
the CEO, Geoscience Australia. Drilling and coring activities were carried out through the Frio-2
project (U.S. Department of Energy), CO2CRC project (Australian Government), and the
WESTCARB project at King Island (U.S. Department of Energy, under contract number DEAC02-05CH1 1231; secondary sampling supported by the Center for Nanoscale Control of
Geologic C0 2 , an Energy Frontier Research Center, funded by the U.S. Department of Energy
under Award Number DE-AC02-05CHI 1231).
Disclaimer: "This publication was prepared as an account of work sponsored by an agency of the
United States Government. Neither the United States Government nor any agency thereof, nor
any of their employees, makes any warranty, express or implied, or assumes any legal liability or
responsibility for the accuracy, completeness, or usefulness of any information, apparatus,
product, or process disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade name,
trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement,
recommendation, or favoring by the United States Government or any agency thereof. The views
and opinions of authors expressed herein do not necessarily state or reflect those of the United
States Government or any agency thereof."
The authors declare no conflict of interest with this work.
59
2.6 Supplemental Figures
Aquifex-aeolicusVF5
OTU_24 010 0
_ OTU_33 0 10 0
100 Limnobacterspe8
S
9
Fo
alcalifaciensC1714
Proteobacteria
Mr1 1 iMM2
79
100
90Eschew
100)
e
500
cha '06i KG11
Yokenefla regentsburgei
53
Micrococcaceae bacteriumIMMIBL 1656
mucifaciens_97-0160
OTU23 01O0
Corynebacterium tuberculostearicumMedalleX
62
81
-orynebacterium
100
81
Actinobacteria
Dermabacter hominisRA7
Uncultred b ateium JPL-1_C21
OTU_023200
96 PrpionibacteriumspRB9
OTU 150100
Firmicutes
M 100
90
CL 80
CL 70
0.
60
50
40
30
20
10
0
,
4,
Proteobacteria
Actinobacteria
Firmicutes
(), 0N
Supplemental Figure 1. (A) Weighted neighbor-joining phylogenetic tree of 16S rRNA
sequences (positions 515 to 1406 E. coli) from clone libraries prepared from biomass from the
initial enrichment, and passages 1, 7 and 9. Numbered columns after OTU names indicate
frequency of sequence type by libraries respectively. Type strains are indicated by species name
and Genbank sequence accession number wherever possible. Bootstrap values of over 50
(expressed as percentages of 100 replications) are shown at branch points. Colors represent
bacterial phyla. The scale bar on upper left represents the unit length of the number of nucleotide
substitutions per site. (B) The bar chart represents the distribution of sequence types among the
three phyla associated with enrichments and passages.
60
Geobacillusthermode nitrificansNG80-2
Geobacillus-kaustophilu s_HTA426
AnoxybacillusflavithermusWKI
BacillusselenitireducensMLS 10
BacioushalodurnsC-125
Oceanobacillus_iheyensisHTE831
Bacilluscoagulans_36Dl
BacillusinfernusTH-22
BacillusboroniphilusCM25
97 Bacillus-jeotgaliYKJ- 10
BacillusselenatarsenatisKNUC9070
11
98 BacillussubterraneusHWG-A
9 BacillusspMITOT1
<
BacillusithioparansKU23
BacillusfirmusIAM12464
BacilluscereussubspcytotoxisNVH391-98
76
1100
100
Bacillus weihenstephanensisKBAB4
BacilluscereusATCC10987
BacilluscereusATCC 14579
Bacillus-spMlT0214
BacilluslicheniformisDSM13
69
Bacillus_subtilissubsp.subtilis168
55 Bacillusamyloliquefaciens_FZB42
3 Bacillus_spWGl 4
91
%
70
BacillusspWMe1
65 Bacillus_safensisFO-036b
8
Bacilluspumilus_ATCC7O61
Bacilluspumilus_SAFR-032
BacillususpWMRl
Bacillus_spWMe2 4Bacillusmegaterium_DSM3 19
100 Bacillus-megateriumQMB1551
*
9
0.02
Supplemental Figure 2. Phylogenetic tree of strains isolated from scCO 2 enrichments and related
type strains. Strain MIT0214 from Frio-2 is denoted with a blue arrow, strain MITOTI from
Otway Basin is denoted with an orange arrow, and strains WG1, WMel, WMR1, and WMe2
from King Island are denoted with green arrows. Alignment was performed with RDP and
visualized with MABL. Bootstrap values greater than 50% are displayed at branch points.
61
B. subtilis
1.E+01
1 atm
Ambient
1 atm
120 atm
CO 2
CO
2
B. mojavensis
B. cereus
1 atm
1 atm 120 atm 1 atm 1 atm
120 atm
Ambient
CO2
CO
Ambient
CO2
CO 2
2
-,-I-
.~1.E+00
1.E-03
0-
_0
0 1.E-04
U)
1 .E-05 2Strain
I
I
and Growth Conditions
Supplemental Figure 3. Change in viable cell counts for Bacillus type strain cultures grown
under ambient pressure and atmosphere after 6 hours exposure to ambient conditions, 1 atm CO 2
6
6
6
or 120 atm CO 2. Initial cell concentrations were 7.3x10 , 8.3x10 and 8.5x10 cells/ml for B.
subtilis PY79, B. mojavensis JF-2 and B. cereus ATCC 14579, respectively.
62
MIT0214
1 atm
Ambient
1 atm
100 atm
CO2
CO2
B. cereus 14579
Initial
1 atm
Heat-kill Ambient
100 atm
CO2
CO2
Initial
Heat-kill
I
I
I
1.E+01
1.E+00
1 atm
I
I
T
1.E-01
_E
Z-
0
1.E-02
1.E-03
0
aI)
1. E-04
CU
a)
c,)
C
Cu
0
U-L
1.E-05
1.E-06
1.E-07
1.E-08
Supplemental Figure 4. Change in viable cell counts for B. cereus MIT0214 (blue) and B. cereus
ATCC 14579 (orange) grown under ambient pressure and atmosphere after 6 hours exposure to
ambient conditions, 1 atm CO 2 or 100 atm CO 2.An aliquot of the initial inoculum was also
exposed to a "heat killing" treatment i.e. incubation of cells at 80'C for 10 minutes to select for
spores. The fraction of cells surviving the heat kill matches the fraction that survive 100 atm CO
2
exposure indicating that the cells surviving scCO 2 exposure are spores and pointing to a higher
fraction of spores in MIT0214 cultures relative to B. cereus 14579 after the same growth interval.
63
1.0
B
A
0.8
EMIT0214 OMITOT1
With Ab No Ab
MIT0214
7 trials
i 0.6
0
~-
0.
Cr
U)
MITOTI
(n=33)
(n=35)
0/5
0/4
0/5
0/5
0/5
0/5
0/4
0/4
1/5
0/5
4/7
2/4
2/5
2/5
I
I-
1/4
2/41
0.2
0.0
+ Antibiotics
No Antibiotics
Supplemental Figure 5. Growth inhibition of MITO214 (dark gray) and MITOT 1 (light gray) by
antibiotics (100 pg/ml Kanamycin and 10 ig/ml Chloramphenicol). Spores were incubated in
parallel with or without antibiotics under a scCO 2 headspace and growth was assessed after 30
days. (A) & (B) All antibiotic amended samples showed no evidence of growth as assessed by a
variety of metrics: cell density, cell morphology, viability counts and turbidity. (B) Growth of
MIT0214 with and without antibiotics was evaluated in seven independent trials consisting of 4
to 7 replicates, while (A) growth of MITOTI with and without antibiotics was evaluated in one
trial of four replicate cultures. We note that the final pressure of one of the four antibiotic
treatments for MITOTI was 61% of the initial pressure, but have included this datum point as
the stringency of the assay was not influenced by the more growth-permissive final pressure.
64
Fold Change Viable counts (CFU/ml) [final/initial]
0.0001
*
0.001
0
0.01
000
0.1
1
10
100
1000
OOOUHO)0
a,)
0
U)
*
0***fee
410
0
0
I0
Supplemental Figure 6. Comparison of the growth outcome for isolate MITOTI under scCO 2
with the final medium redox state as indicated by the color of the resazurin dye dissolved in the
media. The pink color of resazurin indicates a more oxidized redox state, and cultures containing
oxidized metals are pink at the start of incubation. A colorless indicator after incubation indicates
a shift in redox state to more reduced conditions.
65
MIT0214 1 atm CO2
-
14579 1 atm CO2
MIT0214 100 atm CO2
+
14579 100 atm CO2
1.E+07
E
L..
1 1.E+06
0
1.E+05
0
50
100
150
200
250
300
350
Time (hours)
Supplemental Figure 7. Viable counts of spores from MIT0214 (blue shades) and its relative, B.
cereus ATCC 14579 (orange shades) incubated in GYP media under 1 atm CO 2 and 100 atm
CO2 headspace at 370 C. Triplicate cultures were inoculated with prepared spores. Reactors were
pressurized, incubated for 1, 7 or 14 days followed by plating on LB to obtain viable cell counts
(CFU/ml). No significant loss in viability was detected after 2 weeks incubation (p>0.05). Spores
were not quantified directly in this experiment (i.e. through a heat-kill incubation) therefore it is
not possible to distinguish whether the viable counts represent spores or vegetative cells that
germinated in the reactor.
66
A
1atm N2
--
- 1 atm N 2 spores -*-
C1 atm CO 2
1 atm
CO2 spores
1.E+08
1.E+07
U-
0
1. E+06
1. E+05
M,
0;
ir
1.E+04
1.E+03
TI
1.E+02
0
50
100
150
200
250
300
350
400
350
400
ime (hours)
--
P
-. r- 1 atm CO2
1 atm N2
0.12
0.1
E0.08
0.06
0.04
0.02
0
0
50
100
150
200
250
300
Time (hours)
Supplemental Figure 8. Growth of MIT0214 under 1 atm N 2 and 1 atm CO 2 headspace over 412
hours in GYP medium. Triplicate cultures were incubated in serum bottles shaking at 100 rpm at
370 C. (A) CFU counts for MIT0214 grown under N 2 (red line) or CO 2 (blue line), with heatkilling used to determine spore numbers (N 2, pink line and C0 2, light blue line). (B) Optical
.
density (600nm) of cultures measured in part (A). The percent of spores germinating under N 2
was >99% based on decreased spore counts at the onset of the growth phase, while in CO 2 no
more than 19% of the spores disappear by the onset of growth phase (based on plating
uncertainty), indicating a lower germination frequency under CO 2 than N 2
67
N2
99% 1
-I
0
0)
66%
6
3
3
3
3
10
20
25
30
50
-
A
4-
0
Cr
33%
-
9)
L-
-
0%
1
Initial Pressure (atm)
Co 2
99% 1
B
CY)
3
3
1N=
3
66%
'4-
0
C.)
Cr
33%
U-
__
- -I
0%
1
5
3
Initial Pressure (atm)
10
Supplemental Figure 9. MIT0214 growth in pressurized reactors as a function of variable
pressure headspace (A) 100% N 2 (B) 100% CO2. Cultures under N 2 were incubated and observed
after 1 week, while cultures under CO 2 that have an extended lag-phase under 1 atm pressure
were observed after 2 weeks. Dark Gray corresponds to proportion of reactors showing increased
culture turbidity (OD 600nm > 0.1) and light gray corresponds to the proportion without evident
growth. Number of replicates analyzed are shown on each bar.
68
3
0
T
7
Temperature
Control
4
2
1
Co 2
6|
1.
2.
3.
4.
5.
6.
7.
Gas tank
Needle valve
Pressure gauge
4 or 10 mL pressure vess el
1 L Pressure vessel
Heating tape
Temperature control
5
Supplemental Figure 10. Schematic of high pressure cultivation system. All pressure vessels are
pressurized gradually with gas tank pressure through control of a needle valve. After
pressurization, 4 and 10 mL vessels are closed and disconnected for incubation in a 37 0 C
temperature controlled room. The temperature of the 1 L pressure vessel is controlled with a
thermocouple and heating tape. All pressure vessels are gradually depressurized with a needle
valve.
69
Chapter 3: Draft genome sequences of the supercritical CO 2 tolerant bacteria Bacillus
subterraneusMITOTI and Bacillus cereus MIT0214
In preparation for submission to Genome Announcements
70
Draft genome sequences of the supercritical CO 2 tolerant bacteria Bacillus subterraneus
MITOTI and Bacillus cereus MIT0214
Kyle C. Peet' and Janelle R. Thompson'
I Department of Civil and Environmental Engineering, Massachusetts Institute of Technology,
Cambridge, MA, 02139
Abstract
We report draft genome sequences of Bacillus subterraneusMITOTI and Bacillus cereus
MIT0214 isolated through enrichment of samples from geologic sequestration sites in
pressurized bioreactors containing a supercritical (sc) CO 2 headspace. Their genome sequences
expand the phylogenetic range of sequenced Bacilli and allow characterization of molecular
mechanisms of scCO 2 tolerance.
71
3.1 Main text
During geologic carbon sequestration (GCS) large quantities of CO 2 are captured,
compressed to supercritical (sc) state, and injected underground. Whether microbial activities
transform injected CO 2 is not well constrained due to toxic effects of scCO 2 (Bertoloni et al.,
2006; Hong and Pyun, 1999; Spilimbergo et al., 2008; Tamburini et al., 2014; Zhang et al.,
2006). Samples from GCS sites at Otway Basin, Australia and Frio-2, Texas, were used as
inocula for serial enrichment cultures in bioreactors containing scCO 2, yielding strains Bacillus
subterraneusMITOTI and Bacillus cereus MIT0214, respectively (Peet et al., In Review).
Tolerance of scCO 2 was confirmed by growth of spores in pure cultures and was time and
inocula density dependent (Peet et al., In Review). To investigate molecular mechanisms of
growth under scCO 2 , genomic DNA was isolated for genome sequencing.
MITOTI was sequenced on the Illumina HiSeq 2000 platform at the Beijing Genomics
Institute while MIT0214 was sequenced on the Illumina GAIIx platform at the MIT
Biomicrocenter. Paired-end 100 bp reads were trimmed based on quality scores (removing 10
starting and 20 trailing bases) and assembled de novo with CLC Genomic Workbench (CLCBio) with automatic Kmer-sizes of 23 and 21, yielding 185 and 238 contigs >500bp,
respectively. The draft genome of MITOTI is 3.9 Mbp with 42.1% GC, while the MIT0214 draft
genome is 5.6 Mbp with 34.9% GC. Automated annotation using the Rapid Annotation using
Subsystem Technology (RAST) server (Aziz et al., 2008) predicted 4021 (with 1235
hypothetical) and 5640 (with 1399 hypothetical) total coding sequences in MITOT1 and
MIT0214, respectively.
Phylogenetic analysis of the 16S rRNA gene placed MITOT1 within a clade of Bacilli
isolated from diverse environments including deep subsurface, soil, manufacturing effluent, and
72
fermented seafood (Ahmed et al., 2007; Deepa et al., 2010; Kanso et al., 2002; Yamamura et al.,
2007; Yoon et al., 2001), some of which are capable of anaerobic reduction of Fe(III), Mn(IV),
Se(VI), and As(V) (Kanso et al., 2002; Yamamura et al., 2007). The closest relative by BLASTN
of the 16S rRNA gene was B. subterraneusHWG-Al 1 at 98.6% identity. The nearest genome
sequenced strain at 98.1% 16S rRNA identity was B. boroniphilusDSM 17376 isolated from soil
with high concentrations of boron (gil et al., 2014), sharing 83.3% Average Nucleotide Identity
(ANI) (Goris et al., 2007), and 2600 sequence homologs (>60% identity). RAST functional
comparison of the MITOT1 genome and B. boroniphilus 17376 with other closely related
sequenced Bacilli (1NLA3E, B. infantis NRRL B-1491 1, B. megaterium DSM319, and B.
coagulans 36D1) predicted multiple anaerobic respiratory reductases and terminal cytochrome C
oxidases unique to MITOTI and B. boroniphilus 17376, pointing to diverse catabolic potential
for this group of Bacilli (Lovley and Chapelle, 1995; Lovley, 1991).
Strain MIT0214 was most similar to B. cereus ATCC 14579 by BLASTN of 16S rRNA
(99.8% identity) and shared 98.5% ANI and 4858 sequence homologs (>60% identity). B. cereus
strains have been isolated from diverse environments, including strain Q1 (92.5% ANI; 4617
sequence homologs) from a deep subsurface oil reservoir (Xiong et al., 2009). Comparisons
among genomes of MITOT 1, MIT0214, and the most closely related sequenced genomes, did
not reveal clear signatures associated with scCO 2 tolerance, which is not surprising in light of
recent observations that this property is wide-spread among Bacilli (Peet, et al. In Review).
Availability of draft genome sequences for Bacillus subterraneusMITOTI and Bacillus cereus
MIT0214 from two GCS sites will facilitate future work targeting gene/protein expression to
advance mechanistic insights into sCCO 2 tolerance.
73
3.2 Acknowledgements
Funding for experimental work was provided to JRT by the Department of Energy Office of
Fossil Energy under award number DE-FE0002128 and by the MIT Energy Initiative. Drilling
and coring activities were carried out through the Frio-2 project (U.S. Department of Energy),
CO2CRC project (Australian Government). Samodha Fernando and Hector Hernandez assisted
with genome sequencing for strain MIT0214.
74
Chapter 4. Changes in lipid and proteome composition accompany growth of
Bacillus spp. under supercritical C02 and may promote acclimation to associated stresses.
In preparation
75
Chapter 4: Changes in lipid and proteome composition accompany growth of
Bacillus spp. under supercritical C02 and may promote acclimation to associated stresses.
Kyle C. Peet' and Janelle R. Thompson'
1Department
of Civil and Environmental Engineering, Massachusetts Institute of Technology,
Cambridge, MA, 02139
Abstract
Recent demonstration of the ability of multiple Bacillus strains to grow in batch
bioreactors containing supercritical (sc) CO 2 is surprising given the recognized roles of scCO2 as
a sterilant and solvent; however growth under scCO2 opens new possibilities for harnessing
diverse biological processes in two-phase reactor systems or during geologic carbon
sequestration (GCS). We hypothesized that Bacilli may alter cell wall and membrane
composition and function to maintain homeostasis in response to scCO2 stresses, which include
cytoplasmic acidification and membrane destabilization. In this study, protein expression (of B.
subterraneusMITOT 1) and membrane lipids (of B. subterraneusMITOT 1 and B. cereus
MITO214) were profiled in stationary phase cultures grown under headspaces of 1 and 100 atm
of CO 2 and N 2 . Strain MITOT 1 grown under scCO2 revealed significantly decreased fatty acid
branching and increased lipid chain lengths relative to low pressure cultures. Strain MIT0214
displayed similar effects from C0 2: decreased fatty acid branching and increased chain length is
significant under 1 atm CO 2 headspace relative to cultures grown under 1 atm N 2 and ambient
headspaces, with high pressures displaying similar but not significant patterns. Proteomes of
76
MITOTI under high and low pressures of CO 2 and N 2 reveal similar profiles over all conditions
(Spearman R > 0.65), with proteins characteristic of stationary phase cultures. Principal
component analyses revealed that proteomes varied with headspace condition, where the first
two principal components corresponded to headspace gas (CO 2 or N 2) and pressure (1 atm and
100 atm), respectively. A highly expressed S-layer protein may be a crucial feature of this
strain's physiology, and numerous citric acid cycle and electron transport chain proteins suggest
that MITOTI may be anaerobically respiring. Ribosomal proteins are enriched in high pressure
samples, and amino acid metabolic proteins are enriched under C0 2, including the glycine
cleavage system, previously documented as upregulated in acid stress response. These results
provide insights into the stationary phase physiology of strains grown under scCO 2 , suggesting
modifications of cell membranes and shifted amino acid metabolism may be involved in
.
response to the acidic, high CO 2 conditions of growth under sCCO 2
77
4.1 Introduction
Supercritical (sc) phase CO 2 has been developed as an industrial solvent and sterilizing
agent, but the bactericidal properties of scCO 2 have been a limiting factor for development of
two phase systems (Knutson et al., 1999) and for harnessing microbial processes during geologic
carbon sequestration (GCS) (Mitchell et al., 2010). However, recent demonstration of microbial
growth under scCO 2 (Peet et al., In Review) opens new prospects in biotechnology and GCS
applications where scCO 2 is present, while challenging the efficacy of scCO 2 sterilization of
spores in food and medical industries. Understanding microbial growth under scCO 2 will help
advance fundamental knowledge of how natural microbial populations may respond when
.
exposed to scCO 2
ScCO 2 is a complicated stress for cells, and is generally regarded as a microbial
sterilizing agent due to a combination of factors including cytoplasm acidification, increased
CO 2 anion concentration, membrane permeabilization, leakage via CO 2 extraction, and physical
cell rupture (Dillow, 1999; Zhang, 2006; Wu, 2010; Bertoloni, 2006; Hong, 1999; Spilimbergo,
2008; Kamihira, 1987; Tamburini, 2014). It is well documented that changes in temperature,
salinity, pH, pressure, and headspace all force cells to alter their membrane lipid composition to
maintain a liquid crystalline membrane that can regulate osmolarity, intracellular pH, and
membrane protein folding (Beales, 2004; Beranova et al., 2010; Guerzoni et al., 2001; Kieft et
al., 1994; Mukhopadhyay et al., 2006). While short duration exposure (less than 30 minutes) of
Salmonella enterica and Eschericiacoli to scCO 2 result in few changes to lipid acyl chains (Kim
et al., 2009; Tamburini et al., 2014), E. coli does show changes in lipid head groups, with a
reduction in phosphatidylglycerol lipids (Tamburini et al., 2014). However, it is important to
note that these short scCO 2 exposures are of unacclimated vegetative cells, with little time to
78
alter membrane lipids before the cells are killed. The stresses imposed by scCO 2 suggest cell
.
membrane and protein expression changes may be necessary to acclimate and grow under scCO 2
Current studies of scCO 2 effects on microbial cells have examined changes as cells are
rapidly inactivated, thus revealing stress responses linked to cell death (Wilkins et al., 2014;
Tamburini et al., 2014; Kim et al., 2009). As we have only recently described populations
growing with a scCO 2 headspace (Peet et al., In Review), how cells acclimate to these conditions
is unknown. Insights from previous work on pressure and acid stress responses may help tease
apart acclimation mechanisms necessary for growth under scCO 2 . High pressure gas exposure
initially compresses membrane lipids, resulting in less fluid membranes, followed by some
disordering caused by dissolution of gases into membrane bilayers (Chin et al., 1976). The
effects of high pressure cause bacteria to compensate by producing more unsaturated lipids in
order to maintain membrane fluidity (Kato and Hayashi, 1999). Similarly, decreased temperature
will decrease membrane fluidity which cells may acclimate to through increased production of
unsaturated and/or branched lipids (particularly the anteiso form) that have lower melting points
due less dense packing (Miladi et al., 2013; Russell and Fukunaga, 1990; Beales, 2004; Klein et
al., 1999) or by decreasing the average chain length (McGibbon and Russell 1983).
While pressure and temperature impact the fluidity of membranes, bacteria will also alter
membrane lipids to acclimate to environmental changes such as pH and headspace composition.
Bacillus subtilis alters membrane lipids under anaerobic conditions, which result in increased
chain length and increases in the anteiso to iso ratio when compared to aerobically grown cells
(Beranova et al. 2010). Acid stressed B. subtilis will increase membrane rigidity through
decreasing branched and unsaturated lipids, which has been suggested to decrease the proton flux
across membranes (Petrackova et al., 2010). Clostridiumacetobutylicum responds to pH
79
reduction with a similar reduction in unsaturated lipids, and also produces more lipids containing
cyclopropane rings (LePage et al., 1987). Among eukaryotes, Saccharomyces cerevisiae cultures
with a larger fraction of straight chain lipids showed increased tolerance to heat and oxidative
stresses (Steels et al., 1994). Since pressure and acidity have opposing effects on membrane
fluidity, it is not yet apparent how cells exposed to the high pressure and high acidity of scCO 2
will balance production of more fluid membrane lipids expected under high pressure with more
rigid membranes that acid stressed cells produce. One hypothesis is that cells acclimated to
scCO 2 may display a lipid profile that is intermediate between acid stressed and pressure stressed
phenotypes. Alternatively, they may have a profile more similar to one of those conditions (i.e.
acid or high pressure) if one aspect of scCO 2 is a more severe stress.
In addition to changes imposed on the cell wall and membrane, stresses associated with
scCO 2 (e.g. acid, pressure, and high CO 2 concentration) may affect expression of proteins.
Upregulation of general and specific Sigma factors and proteins with chaperone function is a
common response to a variety of stresses including acid, pressure, heat, and osmotic (Browne
and Dowds, 2002; Foster, 1999; Ferreira et al., 2003; Gaidenko and Price, 1998; Welch et al.,
1993; Ishii et al., 2005; Abe et al., 1999). Mechanisms for acclimation to low pH include
upregulation of ATP transporters to pump protons out of cells (Martin-Galiano et al., 2001), and
transport/ metabolism of amino acids to buffer intracellular pH (Richard and Foster, 2004; Cotter
and Hill, 2003). Amino acid decarboxylating enzymes can be part of acid stress responses
through consumption of intracellular protons (Cotter et al., 2001), while various enzymes (e.g.
urease and the arginine deiminase system) can produce alkaline products from amino acids and
other compounds to buffer intracellular pH (Chen et al., 1998; Casiano-Colon et al., 1988;
Curran et al., 1995; McGee et al., 1999). Desulfovibrio vulgaris expression patterns indicate
80
increased production of leucine and isoleucine after high pressure CO 2 exposure but before cells
are inactivated (Wilkins et al., 2014), indicating these amino acids may be involved in a scCO 2
stress response, which is not surprising given that various amino acids have been documented as
compatible solutes for osmotic regulation (Csnonka, 1989).
Elevated hydrostatic pressure may also prove stressful for microorganisms, as biomass
and protein expression may be reduced under high pressure (Welch et al., 1993; Bothun et al.,
2004; Ishii et al., 2005). Responses to high pressure share similarities with the general responses
to a broad array of stresses, e.g. upregulation of sigma factors and chaperone proteins (Hormann
et al., 2006; Ishii et al., 2005; Welch et al., 1993). Responses specific to high pressure may also
include upregulation of transcription, translation, and nucleotide metabolism in E. coli (Ishii et
al., 2005), and upregulation of Ribokinase, Clp protease (which may have chaperone function),
and proteins of unknown function in L. sanfranciscensis,(Hormann et al., 2006). In contrast,
high pressure may be a more optimal growth condition for barophilic organisms, which show
upregulation of additional respiratory genes under high pressure (Vezzi et al., 2005; Abe et al.,
1999), and down-regulation of chaperone proteins (Boonyaratanakornkit et al., 2007).
Isolates B. subterraneusMITOTI and B. cereus MIT0214 were enriched from deep
subsurface samples through successive rounds of cultivation under supercritical CO 2 (Peet et al.,
In Review). To better understand how strains MITOTI and MIT0214 acclimate to scCO 2 , and to
separate the effects of headspace gas and pressure, we have examined how the lipid content of
MITOT 1 and MIT0214 and the proteome of MITOT 1 vary in stationary phase cultures as a
function of N 2 and CO 2 headspaces at ambient pressure or at 100 atm pressure. Our results
suggest several modifications of lipid and proteome profiles that may facilitate growth under
sCCO 2, but also point to a high degree of similarity across all anaerobic conditions, suggesting
81
the homeostasis of stationary phase cultures is maintained by similar mechanisms independent of
responses to scCO2 exposure.
4.2 Methods
Cell growth and collection for lipid and protein analyses
All cultures were started from spore inocula of 10 5/ml (direct counts). Growth media for
B. subterraneusMITOTI consisted of MS media with a supplement for metal reducers (Colwell
et al., 1997) and LB media (Difco) was used for strain B. cereus MIT0214. Aerobic cultures of
both MITOT1 and MIT0214 were incubated for 72 hours with shaking at 37'C to target late
stationary phase (OD > 5), before sporulation occurred. Anaerobic cultures of MITOTI were
grown in 316 stainless steel vessels as described in Peet et al. (In Review) in triplicate, shaking at
37 'C in the following four conditions: 1) 1 atm, 100% N 2 headspace; 2) 1 atm, 95% CO2 , 5% H2
,
headspace (referred to as 1 atm C0 2); 3) 100 atm, 100% N 2 headspace; 4) 100 atm, 97% CO2
3% He headspace (referred to as 100 atm CO 2). Incubation time to sample pressurized
bioreactors was 30 days, corresponding to the time predicted by logistic regression where
approximately 47% of bioreactors would show positive growth for strain MITOT1
(Supplemental Fig. 1) (Peet et al., In Review). Bioreactors at 1 atm were sampled at 21 days due
to faster growth of cultures at low pressure, based on growth curves at 1 atm N 2 and 1 atm CO 2
(Supplemental Fig. 2). Sampling times of 21 and 30 days for MITOTl cultures grown under 1
atm and 100 atm, respectively, were selected to target similar durations of time spent in
stationary phase (>7 days) in order to minimize variability associated with growth phase in
cultures. Sampling times were estimated based on observed dynamics under 1 atm and the
82
conditional probability of growth under 100 atm based on application of Bayes theorem to a
logistic regression model for growth outcome (observed/not observed) for MITOTI as a function
of time with an inocula of 105 spores per ml (Peet et al., In Review). Final cell densities of
reactors with positive growth under the four conditions were measured by direct counts and
compared by ANOVA. Anaerobic cultures of B. cereus MIT0214 were incubated in hungate test
tubes (1 atm pressure) or 316 stainless steel vessels (100 atm pressure) under conditions 1- 4
above, and were sampled at 96 hours for conditions 1 and 2 and at 30 days for conditions 3 and
4. Both strains were used for lipid analyses in the 4 conditions above and aerobic controls, while
only strain MITOT1 was used for protein analyses, with proteins extracted from conditions 1- 4
above. All culture media for anaerobic conditions was degassed with the respective N 2 / CO 2
headspace for 30 minutes, followed by addition of Na2 S in an anaerobic chamber to further purge
any residual oxygen.
After completion of incubation, high-pressure samples were depressurized at a rate of 3-5
atm 1 minute over approximately 30 minutes. Cells were collected by centrifugation at 14,000 X
g for 6 minutes, followed by resuspension in sterile filtered PBS and re-centrifugation. Cell
pellets were frozen in -80 'C until further analysis. Aliquots of biomass were taken for direct cell
counts and viable cell counts.
Lipid extraction and construction of fatty acid methyl esters (FAMEs)
Triplicate samples for aerobic stationary phase cultures and each of the four anaerobic
growth conditions were extracted by a modified Bligh-Dyer extraction for polar lipids (Bligh and
Dyer, 1959). All vials, glass pipettes and foil were combusted before use. Syringes were triple
washed in each of the following solvents before and between pipetting: methanol,
83
dichloromethane (DCM), and hexane. 1 ml of a solvent mixture containing methanol: DCM:
phosphate buffered saline (PBS) (10:5:4) was added to the 1.5 ml centrifuge tube containing
frozen biomass and the pellet was resuspended and transferred to a 50 ml glass centrifuge tube.
The original 1.5 ml centrifuge tube was washed twice more with this solvent mixture to collect
all cells, and a total of approximately 7 ml of solvent was added to the 50 ml glass tube. 1 tg of
3-Methylheneicosane was added as an internal standard to each sample at the beginning of
extractions. Cells in this solvent mixture were vortexed for 5 minutes, followed by 15 minutes of
sonication to further extract lipids and then centrifuged for 10 minutes at 720 X g.
Approximately 90% of the solvent mixture was aspirated to a new glass vial without disturbing
cells, and this extraction process was repeated once with the same solvent mixture, twice with a
solvent mixture containing methanol: DCM: 2.5 % trichloroacetic acid (10:5:4), and once with
DCM: methanol (3:1), pooling all the extractions in a separate vial. Phase separation was
conducted by adding 5 ml of PBS and 5 ml of DCM with gentle shaking, followed by removal of
the lower (DCM) phase to a new vial. 5 ml of DCM was added twice more to reextract aqueous
phase, and the pooled DCM phases were evaporated in a Turbovap at 37 'C under a stream of
100% ultrapure N 2 . The concentrated samples (of approximately 1 ml volume) were transferred
to 4 ml vials and then dried down completely. These were labeled as total lipid extracts (TLE)
and stored at 4 'C. Intact polar diacylglycerols were converted to FAMEs by methanolysis.
Briefly, TLE's were resuspended in 200 ptl of DCM: methanol (9:1), removing 140 PI to a 2 ml
vial with insert, and then drying down the 140 pA. Once dried, 100 pl of methanoic HC was
added, samples were capped and heated at 60'C for greater than 1.5 hours. Samples were
evaporated, followed by addition of 200 pl of DCM: methanol (9:1), evaporation, addition of
200 pl of methanol, evaporation, and then resuspension in hexane for analysis via GC/MS.
84
Determination of unsaturated double-bond positions
Monounsaturated fatty acid double-bond positions were determined by the method of
Yamamoto et al. (1991). An aliquot of FAMEs for each sample with detectable unsaturated fatty
acids based on GC/MS analysis of TLE's was transferred to a new 4 ml vial and dried down. 100
pL of dimethyl disulfide (DMDS) and 50 tL of iodine was added to each vial and vials were
heated on a heating block at 35'C for 30 minutes. 1 ml of hexane was added to each vial and
Na2 S 2O 3 was added to vials 20 p at a time until mixture turned clear, with vigorous vortexing
between each addition. The organic layer was removed and then re-extracted with DCM:
methanol (9:1). Organic layers were combined and run through a Na2 SO4 column to remove
residual water, followed by rinsing the column with DCM and hexane, before concentrating
under N2 gas. Samples were resuspended in hexane and analyzed via GC/MS.
Analysis of lipids
Samples were analyzed on an Agilent 7890A gas chromatograph attached to an Agilent
5975C inert XL mass selective detector (MSD) equipped with a programmable temperature
vaporization (PTV) injector. 1 pl of sample dissolved in hexane was injected into an Agilent
J&W DB-5ms column (60 meter X 0.25 mm internal diameter, with 0.25 pm coating). The PTV
injector was held at 60'C for 2 minutes, then ramped up to 150'C at 10 C per minute, followed
by ramping up to 315'C at 3C per minute. Total run time was 90 minutes per sample. GC/MS
peaks and spectra were acquired with Agilent GC/MSD software and peak areas were manually
integrated with Enhanced Data Analysis software. Lipids were determined by searching the mass
spectra of integrated peaks against the National Institute of Standards and Technology (NIST)
85
database and matching retention times and peak elution order to NIST predictions. To search for
hopanoids, m/z 191 was used to indicate the potential presence of hopanoids (Sessions et al.,
2013). Ion 191 was extracted from chromatograms and any spectra with large 191 ions were
searched against the NIST database to compare mass spectra. Lipids were normalized to the
internal standard for comparison between samples. The normalized means of triplicate extraction
blanks were subtracted from all ni 6:0 and n1 8:0 peak areas in all samples (Supplemental Fig. 3).
The uncertainty from extraction blanks was propagated into error bars shown in all figures.
Average lipid chain length was calculated by multiplying each lipid's fractional abundance by
the length of the acyl chain and summing those weighted values. Microsoft Excel was used for
T-tests and analysis of variance (ANOVA) was calculated with JMP Pro 11 (SAS software).
Protein extraction and purification
Following growth of cultures, cells were collected by centrifugation (14,000 X g for 6
min), washed in PBS and frozen at -80'C until extraction. Whole cell proteomes were extracted
through adding 200 pul of lysis buffer (Huang et al. 2012) containing 8 M urea, 4% (w/v)
CHAPS, 40 mM Tris and 65 mM DTT to each frozen cell pellet. 100 p of sterile 0.1 mm
zirconia beads was added to each tube and samples were bead beat for 1 min at maximum speed
(MOBio vortexer), followed by 30 seconds on ice. Bead beating and ice bath cooling was
repeated 10 times, with the final removal of beads and cell debris by centrifugation for 3 min at
14,000 X g. The protein-containing supernatant was aspirated, placed in a new tube, and frozen
at -80'C until digestion. An aliquot of each sample was used for protein quantification via
BioRad Protein Assay with bovine serum albumin standard.
86
Proteins from 8 samples (2 from each condition), were purified using optimized
conditions for proteome analysis. Cell lysates were added to Vivaspin columns (Sartorius) with a
3000 Da size cutoff to remove urea and extraction buffer reagents. For all samples, an additional
500 pl of water was added to each column to dilute the sample and columns were spun for 10
min at 13,000 X g, room temperature. This was repeated, with the flow-through discarded each
time. The remaining proteins in the column were resuspended and added to a new centrifuge tube
-
with four volumes of cold acetone (-20'C). Proteins were precipitated by incubating tubes at
80'C for 30 minutes and then centrifuging at 16,000 X g for 10 minutes at 4'C. The supernatant
was discarded and the protein pellet was washed with 300 [d acetone and spun again, followed
by air drying for 5 minutes to allow residual acetone to evaporate.
An additional set of replicates (1 from each condition) was processed in the manner as
described above, with the exception of an initial precipitation step (with acetone in order to
remove urea and extraction buffer reagents from cell lysates) that reduced the efficiency of
product recovery. These samples are noted in figures and tables and are included for qualitative
or univariate comparison to other replicates.
Protein digestion and peptide fractionation
Proteins were re-suspended in a 15 pl of 8M urea (dissolved in 50 mM ammonium
bicarbonate) followed by adding 20 pl of 0.2% ProteaseMAXiM (Promega) surfactant, 50 pd of
ammonium bicarbonate, and 2.12 pl of 400 mM DTT to reduce disulfide bonds. Samples were
incubated at 56'C for 30 minutes, and then alkylated by addition of 6 pl of 550 mM
iodoacetamide, followed by incubation for 30 minutes at room temperature in the dark. To
prevent alkylation of trypsin, excess iodoacetamide was inactivated by addition of 2.12 p of
87
DTT and incubated for an additional 30 minutes in the dark. Proteins were digested by adding
3.7 d of 0.5 tg/ptl trypsin (1:27 trypsin: protein ratio) and 1 pl of 1% ProteaseMAXTM followed
by 3 hours incubation at 37'C. After digestion, trypsin was inactivated by addition of 20%
trifluoroacetic acid to a final concentration of 0.5%. Digested proteins were concentrated and
desalted with OMIX tips (Agilent technologies, Part No. A5 7003 100) according to manufacturer
instructions, and dehydrated to dryness in a SpeedVac*.
To fractionate peptides by isoelectric point, samples were re-suspended in 3.6 ml of IX
off-gel buffer and then loaded onto an Agilent Off-gel fractionator with IPG strips (pH 3-10)
according to manufacturer instructions. For the first 4 samples, the 24 fractions were pooled into
19 fractions (combining 1 and 24, 2 and 23, 3 and 22, 4 and 21, 5 and 20, 6 and 19, without
combining fractions 7-18). As protein concentrations were relatively low and LC-MS runs did
not show a large number of peptides, we pooled the 24 fractions down to 12 fractions in the
second set of 8 samples. All fractions were dried in a SpeedVack prior to re-suspension in 20 pl
of 98% H 20, 2% acetonitrile, and 0.1% formic acid for LC-MS analysis as described below.
LC-MS parameters and protein profiling
An Agilent 6530 quadrupole time-of-flight (QTOF) mass spectrometer equipped with an
electrospray ionization (ESI) source was used. All samples were analyzed using an Agilent 1290
series Ultra Performance Liquid Chromatography (UPLC) (Agilent Technologies, Santa Clara,
CA, USA) containing a binary pump, degasser, well-plate auto-sampler with thermostat, and
thermostatted column compartment. Mass spectra were acquired in the 3200 Da extended
dynamic range mode (2 GHz) using the following settings: ESI capillary voltage, 3800 V;
fragmentor, 150 V; nebulizer gas, 30 psig; drying gas, 8 L/min; drying temperature, 380'C. Data
88
were acquired at a rate of 6 MS spectra per second and 3 MS/MS spectra per second in the mass
ranges of m/z 100-2000 for MS, and 50-2500 for MS/MS and stored in profile mode with a
maximum of five precursors per cycle. Fragmentation energy was applied at a slope of 3.0 V/100
Da with a 3.0 offset. Mass accuracy was maintained by continually sprayed internal reference
ions, m/z 121.0509 and 922.0098, in positive mode. Agilent ZORBAX 300SB-C18 RRHD
column 2.1 x 100 mm, 1.8pm (Agilent Technologies, Santa Clara, CA) was used for all analyses.
LC parameters: autosampler temperature, 4'C; injection volume, 20 pl; column temperature,
40'C; mobile phases were 0.1% formic acid in water (phase A) and 0.1% formic acid in
acetonitrile (phase B). The gradient started at 2% B at 400 pl/min for 1 min, increased to 50% B
from I to 19 min with a flow rate of 250 pl/min, then increased to 95% B from 19 to 23 min with
an increased flow rate of 400 pl/min and held up to 27 min at 95%B before decreasing to 2% B
at 27.2, ending at 30 min and followed by a 2 minute post run at 2% B.
Protein data processing
Raw data was extracted and searched using the Spectrum Mill search engine
(B.04.00.127, Agilent Technologies, Palo Alto, CA). "Peak picking" was performed within
Spectrum Mill with the following parameters: signal-to-noise was set at 25:1, a maximum charge
state of 7 is allowed (z = 7), and the program was directed to attempt to "find" a precursor charge
state. During peptide searching the following parameters were applied: peptides were searched
against the genome of B. subterraneusMITOTI, carbamidomethylation was added as a fixed
modification, and the digestion was set to trypsin. Additionally, a maximum of 2 missed
cleavages, a precursor mass tolerance +/- 20 ppm, product mass tolerance +/- 50 ppm, and
maximum ambiguous precursor charge = 3 were applied. Data were evaluated and protein
89
identifications were considered significant if the following confidence thresholds were met:
protein score > 13, individual peptide scores of at least 10, and Scored Peak Intensity (SPI) of at
least 70%. The SPI provides an indication of the percent of the total ion intensity that matches
the peptide's MS/MS spectrum. A reverse (random) database search was simultaneously
performed and manual inspection of spectra was used to validate the match of the spectrum to
the predicted peptide fragmentation pattern, hence increasing confidence in the identification.
Standards were run at the beginning of each day and at the end of a set of analyses for quality
control. Protein expression values (spectrum counts) were calculated in Scaffold software with
the imported peptide hits from Spectrum Mill. The threshold for including a protein was a
minimum of two distinct peptides with 95% confidence. To compare between samples, spectrum
counts for each protein were divided by the sum of spectrum counts in respective samples,
resulting in protein expression values as a percent of total.
Data analysis and statistics were conducted with Microsoft Excel, JMP Pro 11, Simca and
Primer 6 software. We used a nonparametric test, the Kruskal-Wallis test, to determine if
individual proteins were differentially expressed under different conditions. The Kruskal-Wallis
test is based on ranks and is more robust with non-normal distributions, which is important
considering that the protein expression values varied considerably across samples (especially due
to the first experiment's reduced total peptide counts compared to the second experiment).
Proteins with 2 or more values below the detection limit were not considered for significance
with the Kruskal-Wallis test. Clustering was performed with Spearman rank correlation and
Principal Component Analysis (PCA) was performed in order to determine in an unsupervised
manner how samples vary. Partial least squares discriminant analysis (PLS-DA) is a supervised
90
method, which maximizes the separation of variables, and was used to identify proteins that are
most responsible for acclimation to different conditions. Gene set enrichment analysis (GSEA)
was performed with GSEA software from the Broad Institute (Subramanian et al., 2005), and
was used to determine if pathways are differentially expressed in response to different
conditions. Samples grown under a CO 2 headspace were compared with the N 2 samples, and
low-pressure samples were compared with high-pressure samples to see what gene sets
correspond to those four variables. The MITOTI genome (Peet et al., In Prep) was annotated
with the Kyoto encyclopedia of genes and genomes (KEGG) automatic annotation server
(KAAS) (Moriya et al., 2007) to obtain KEGG IDs for proteins to be used in conjunction with
KEGG gene sets in GSEA. Proteins that could not be annotated with KEGG were excluded from
this analysis. Gene set size was set to a minimum of 5 proteins, and 1000 permutations were
performed.
4.3 Results
Growth of MITOTI and MIT0214 in bioreactors under different headspace and pressures
Anaerobic cultures of MITOT1 were grown in 316 stainless steel bioreactors as described
previously (Peet et al., In Review). Cultures grown under 1 atm CO 2 or N 2 were harvested at day
21. 1 atm cultures appear to reach maximum cell density after 9-14 days (Supplemental Fig. 2)
and were thus expected to have aged in stationary phase for at least 7 days. High pressure
cultures were harvested from 100 atm batch reactors at 30 days. Pressurized samples exhibit
highly variable growth outcomes (observed/not observed) that is time and initial spore densitydependent (Peet et al., In Review), thus the conditional probability of prior growth of cultures
91
with observed biomass at day 30 was calculated (Supplemental Fig. 1) (Peet et al., In Review).
Reactors with observed biomass at day 30 have a 49% chance of growth by day 23 and thus
having aged for >7 days. Thus, it is expected that cultures grown at 1 atm and 100 atm will
represent a mixture of culture ages but that these cultures are likely to have entered stationary
phase to minimize variability associated with growth phase and to enable comparisons among
the different headspace conditions. We observed growth in 5 of 6 bioreactors containing 1 atm
CO 2 headspace, 3 of 3 bioreactors containing 1 atm N 2 headspace, 4 of 11 bioreactors under 100
atm CO 2 , and 7 of 7 bioreactors under 100 atm N 2. MIT0214 yielded growth in 4 of 6 bioreactors
under 100 atm C0 2 , and 3 of 3 bioreactors under 100 atm N 2. When more than 3 replicates were
available for analysis, samples with higher biomass were selected in order to maximize material
for protein and lipid analyses. Direct cell counts and viable cell counts of triplicates cultures used
for lipid and proteomic characterization are displayed in Supplemental Fig. 4. Final cell densities
are similar to those in Peet et al. (In Review), as final direct counts under anaerobic conditions
are generally greater than 10 7 cells/ml, and always less than 2x10 8 cells/ml, for both strains.
Viable counts were consistent with stationary phase cultures, when compared to growth curves
under 1 atm N 2 and 1 atm CO 2 (Supplemental Fig. 2 for MITOTI; (Peet et al., In Review) for
MIT0214). Viable counts between conditions were not significantly different in MITOTI (twofactor ANOVA with headspace and pressure), however direct counts varied significantly, with
greater counts in CO 2 samples (p= 0.005, F-ratio= 15.1), and in high pressure samples (p= 0.003,
F-ratio= 18.1), but the interaction was not significant. It is unknown whether these differences in
biomass (based on direct counts) significantly affected downstream analyses. For strain
MIT0214 samples (only used in lipid analyses), no significant differences were observed in
viable counts or direct counts (2-way ANOVA with headspace and pressure).
92
,
B. subterraneus MITOTI lipids display significant changes under supercritical C0 2
appearing similar to an acid stressed phenotype
Variation in lipid composition, branching, saturation, and chain length were examined to
determine potential acclimation patterns in response to different headspace gases and pressures.
Lipids from B. subterraneusMITOTI grown under an aerobic, ambient pressure atmosphere to
stationary phase are composed of primarily saturated, branched lipids (67% of total), consistent
with lipid content in a wide range of Bacilli that vary from 44-99% branched (Kaneda, 1991).
The lipid profile of strain MITOTI is similar to the closely related strains B. boroniphilus,B.
selenatarsenatis,and B. jeotgali and consists of the major lipids il5:0, ail5:0, i16:0, n16:0,
i17:0, ail7:0 and n18:0, composing 83% of total lipids, with a larger percentage of i16:0, n16:0,
and n18:0, and a lower percentage of i15:0 compared to its close relatives (Table 1).
Table 1. Fractional abundance of major lipids of B. subterraneus MITOT1 and close phylogenetic
relatives
B.
B. subterraneus
selenatarsenatis
MITOTI
B. boroniphilus
SF1
B. jeotgali
YKJ-10
B. jeotgali
YKJ-11
i14
n14
3.7
1.0
0.7
0.6
4.5
1.9
1.3
4.3
0.6
i15
15.8
44.8
47.3
49.3
46.0
a15
n15
i16
n16
Sum n16
unsaturated
i17
a17
n17
Sum n17
unsaturated
n18
Sum n18
unsaturated
13.0
0.5
17.9
9.2
12.5
0.0
1.6
1.5
4.2
0.0
2.9
0.8
8.8
0.0
2.3
3.2
6.4
0.9
5.4
3.9
4.8
4.8
12.1
0.3
5.4
5.0
9.3
0.0
11.2
2.6
1.4
0.0
9.5
4.1
3.7
0.0
13.8
4.3
3.3
0.0
5.6
10.6
6.9
0.0
10.1
0.0
7.5
0.0
5.7
0.0
0.9
0.0
0.0
0.0
0.0
93
1.6
Changes in branched/straight lipid content relative to aerobic cultures are observed when
MITOT1 is grown under N 2 or CO 2 to stationary phase, although since the onset of stationary
phase occurs at different times for cultures grown under the various conditions (i.e. 24 hours for
aerobic cultures, 9-14 days for 1 atm-grown cultures (Supplemental Fig. 2) and 21-30 days for a
50%-100% probability of growth for 100 atm grown cultures (Supplemental Fig. 1)) the role of
culture age in modulating lipid trends cannot be determined with the current data. While
branched lipids dominate the aerobic stationary phase culture (67% branched), the four anaerobic
conditions tested show varying degrees of reduction in branching in both iso and anteiso lipids,
,
with branched lipids composing 29% of total lipids under 100 atm CO 2, 42% under 100 atm N 2
59% under 1 atm CO 2 , and 58% under 1 atm N 2 (Fig. 1). This overall reduction in branching and
N 100atmCO2
latmCO2
U 100atmN2
latmN2
U latmAmbient
1.4
1.2
0
'6
0.8
0.6
0.4
0.2
Iso branched
Anteiso
branched
Total branched
Straight
Saturated
Unsaturated
Figure 1. Major lipid classes of MITOT1 sampled in stationary phase under 5 headspace and
pressure conditions. Significance described in the text is denoted with (*). Iso and anteiso
branched lipids are summed to make up 'Total branched' lipids. Total branched lipids and
straight lipids and summed to make up saturated lipids. Saturated and unsaturated lipids sum to 1
for each sample.
increase in straight lipids in anaerobic cultures is particularly noticeable in a subset of lipids, e.g.
i16:0 is the most abundant lipid under aerobic conditions (18% of total), but it is only present
from 4-7% of total in anaerobic conditions (Supplemental Fig. 5 A). The branched lipid i 16:0
94
varies significantly among the five headspace and pressure combinations with higher levels
under aerobic conditions than all anaerobic condition (ANOVA p< 0.0003; F-ratio= 15.8,
Tukey's post-hoc test alpha > 0.05). In contrast, the straight lipid, n16:0 is more abundant under
all anaerobic conditions but this difference only met significance between aerobic and both high
pressure samples (ANOVA p< 0.02, F-ratio= 5.6, Tukey's post-hoc test alpha > 0.05).
Significant variation was observed for branched lipid i 15:0 among the four anaerobic conditions
with decreased abundance in high-pressure conditions (Two-factor ANOVA p= 0.04; Fratio=5.6). Significantly more saturated, straight lipids were present in scCO 2 grown biomass
when compared to both 1 atm CO 2 and 1 atm ambient atmosphere grown samples (t-test p= 0.04,
p= 0.02, respectively). Lipid n16:0 is also significantly elevated under 100 atm CO 2 relative to 1
atm CO 2 (t-test p= 0.01). However, between all anaerobically grown samples, most individual
lipids do not show significant differences with respect to headspace and/or pressure by ANOVA
or pairwise t-test. The high concentrations of CO 2 (associated with a supercritical headspace)
will result in a significant pH reduction, and the response of MITOTI to substantially reduce its
branched lipid content under scCO 2 is similar to observations of B. subtilis acclimating to acid
stress (Petrackova et al., 2010).
Changes in the average acyl chain length of MITOTl lipids are also observed in response
to different headspaces and pressures, with higher average chain lengths in higher pressure and
CO 2 headspace cultures (Fig. 2). While these trends in headspace and pressure are not
statistically significant by two-factor ANOVA, pairwise t-tests reveal a significant increase in the
chain length of 100 atm CO 2 compared to 1 atm N2 and 1 atm ambient headspace samples (t-test
p< 0.04 for both comparisons). In addition to chain length and branching patterns, unsaturated
lipid content may be altered in response to various stresses. Several unsaturated lipids were only
95
16.4
detected in aerobic samples (two unsaturated n16 and two unsaturated n17 lipids). However, no
significant differences in total unsaturated lipids were observed between the different anaerobic
headspace and pressure combinations, by either 2-factor ANOVA or pairwise t-tests. Hopanoids
may also be important to tolerating stresses, e.g. acid stress (Welander et al., 2009), however we
did not find any evidence of hopanoid production (presence of m/z 191 ion with spectral
comparison to the NIST database) for strain MITOTi.
17
*
16.8
16.6
156
chain lengths are significantly
16.2
16
U1O0atmCO22
0
latmC02
100atmN2
Figure 2. Average lipid chain
lengths of MITOT 1 sampled in
stationary phase under 5 headspaces
and pressure conditions.
Significance described in the text is
denoted with (*). 100 atm CO2
greater than 1 atm N2 and 1 atm
ambient samples (p<0.04).
9 latmN2
> latmAmbient
15.8
15.6
15.4
B. cereus MIT0214 lipids reveal similar trends as strain MITOT1 when incubated under
identical anaerobic conditions
In stationary-phase cultures grown under anaerobic headspaces, significant decreases
were observed in the proportion of branched lipids under CO 2 when compared to N 2 . Total
branched lipids were significantly lower in 1 atm CO 2 samples when compared to 1 atm N 2
samples (t-test p<0.003, Supplemental Fig. 6), and individual branched lipids i15:0, a15:0, i16:0,
il7:0, and al 7:0 decreased significantly in 1 atm CO 2 when compared to 1 atm N 2 (t-test p<0.05
for all) (Supplemental Fig. 5B). Significant variation was observed in al 7:0, with higher
96
abundance in both N 2 headspace conditions (two-factor ANOVA p=0.001; F-ratio=25.5, Tukey's
post-hoc test alpha >0.05). Significant variation was also found in n17:0, with higher abundance
in CO 2 headspace conditions (two-factor ANOVA p=O.004; F-ratio=15.7, Tukey's post-hoc test
alpha >0.05). These reductions in MIT0214 branched lipids under CO 2 headspaces are
qualitatively similar to patterns observed in MITOT 1.
In addition to altered branching patterns between aerobic and anaerobic samples, average
chain lengths in MIT0214 lipids were longer under all four anaerobic conditions compared to
aerobic samples, although only 1 atm CO 2 and 1 atm N 2 samples were significantly longer than
ambient (t-test p=0.001, p=0.0006, respectively) (Supplemental Fig. 7). Increased chain lengths
in anaerobic conditions are consistent with observations of other Bacilli (Beranova et al., 2010).
In contrast to MITOTI, chain lengths did not vary significantly between anaerobic conditions,
but were qualitatively similar to B. subterraneusMITOTi. ScCO 2 conditions were associated
with the longest average chain lengths in MIT0214 (Supplemental Fig. 7).
We hypothesized that high pressure grown cultures might increase production of
unsaturated lipids to balance the membrane compressing effects of pressure, but we did not
observe significant increases in unsaturated lipids in high-pressure samples for either strain (Fig.
1, Supplemental Fig. 6). However, in strain MIT0214, only high pressure grown samples
produced detectable unsaturated lipids; two of three scCO 2 replicates and 1 of 3 high-pressure N 2
replicates (Supplemental Fig. 5B, Supplemental Fig. 6). Increasing unsaturated lipids under
elevated pressure would be consistent with documented high-pressure acclimation mechanisms
(Kato and Hayashi, 1999), but it is important to note that many studies of high-pressure
acclimation use substantially higher pressures than 100 atm used in this study.
(Boonyaratanakomkit et al., 2007). We did not find any evidence of hopanoid production by
97
strain MIT0214. Similar to MITOTI, CO 2 appears to have a strong influence on membrane lipid
phenotype in MIT0214. Both of these strains indicate that membrane lipid adjustments may be
.
part of the acclimation response enabling growth under supercritical CO 2
Proteomes of B. subterraneus MITOT1 vary by headspace composition and pressure
Proteomes from MITOTl cultures under high and low pressures of CO 2 and N2 resulted
in 623 distinct proteins in total (with at least 2 peptide hits), corresponding to 15% of total
proteins predicted by RAST annotation (Peet et al., In Prep). Duplicate samples for each
condition yielded 347 to 598 distinct proteins per sample, with an additional set of replicates,
processed as a separate batch yielding 68 to 180 proteins per sample (Supplemental Fig. 8). With
the biomass constraints of growing MITOTI in limited volume stainless steel culture vessels (5
ml of culture, 5 ml headspace), and the reduced biomass obtained from anaerobic cultivation,
these results compare reasonably well with other Bacillus proteomes (less than a factor of 2
difference in total proteins) (Eymann et al., 2004; Huang et al., 2012).
In order to determine whether different headspace and pressure conditions resulted in
different proteome profiles, we performed clustering and principal component analysis (PCA) on
MITOTI proteomes. Spearman clustering analysis revealed significant differences between the
two batches of proteomes (Supplemental Fig. 9), thus multivariate analysis was restricted to the 8
proteomes prepared by identical methods. These 8 proteomes display high correlation between
all proteomes (Spearman R> 0.65), with the highest correlation between duplicates of 1 atm N 2
and 1 atm CO 2 (Spearman R> 0.80). PCA on the remaining 8 proteomes (2 from each condition)
(Fig. 3), resulted in 79.7% of sample variation explained by the first two principal components
(PCi: 40.4% and PC2: 3 9.2%). These two principal components (Fig. 3) resolved samples by
98
condition, with N 2 and CO2 headspaces separated along PCI and high and low pressure samples
separated along PC2. This result supports our hypothesis that differential protein expression will
be necessary in order to acclimate to different pressure and headspace conditions.
0.02
100 atm CO 2 J
100 atm N 2 C
A
I
0.01-100 atm N 2 A
()
100atmCO 2 D
0--------------------
----------------I
V
1 atm C02 B
-0.01 --
1 atm N 2 B
1 atmCO 2 C
V
1 atm N 2 A
-
-0.02 -0.02
-0.01
0
0.01
PCi (40.4 %)
0.02
Figure 3. Principal component analysis of duplicate MITOTI proteomes. Samples separate by
headspace along principal component 1, and by pressure along principal component 2, with the
first two principal components explaining 79.6% of variability. Samples grown under 1 atm CO
2
and 1 atm N 2, were sampled at 21 days; Samples grown under 100 atm CO 2 (scCO 2 ) and 100 atm
N2 were sampled at 30 days.
MITOT1 proteomes show similarity in highly represented proteins across conditions
The most abundant proteins in MITOTI proteomes are consistent with highly expressed
proteins present in other Bacilli across multiple growth phases (Eymann et al., 2004; Huang et al.
2012; Bernhardt et al., 2003). In all conditions, the most highly expressed proteins include
proteins that are involved in: translation (e.g. tranlation elongation factor Tu and G), energy
generation (e.g. ATP synthase, and citric acid cycle proteins such as aconitase hydratase and
99
succinyl-CoA ligase), and general stress response (e.g. heat shock protein 60 GroEL and
chaperone protein DnaK). These proteins are abundant in all 4 of the conditions tested; Tables 2
- 5 summarize rank values for highly expressed proteins in each of the conditions tested (1 atm
CO 2 : Table
2, 100 atm CO 2 : Table 3, 1 atm N 2 : Table 4, 100 atm N 2 : Table 5). The expression of
multiple proteins involved in the citric acid cycle and the electron transport chain (Supplemental
Fig. 10) suggests that MITOT1 may be anaerobically respiring. This is notable as other closely
strains are capable of metal respiration (Kanso et al., 2002; Yamamura et al. 2007), and the
growth media for these experiments contained potential electron acceptors (the oxidized metals
FeIl and MnIV). In addition to the common, highly expressed proteins listed, MITOTI
proteomes also show a number of proteins involved in amino acid metabolism, transcription,
ribosomal proteins, glycolysis/ gluconeogenesis, and flagellar proteins. Relatively few
sporulation proteins were present; consistent with microscopic observations that did not find any
evidence that cultures had sporulated.
Among the most highly expressed proteins across conditions, gene 2630, is a predicted Slayer protein (BLASTX results: Supplemental Table 1), by Phyre2 structural prediction (42% of
the protein predicted with greater than 90% confidence) (Kelley and Sternberg, 2009) (Fig. 4 A).
The structural prediction showed highest homology to N-terminal domains of S-layer proteins
from other Bacilli, including the SLH motif, which are responsible for anchoring S-layer proteins
to membranes (Mesnage et al., 1999). Expression values varied from 1.7 - 24.2% of total
spectrum counts under CO 2 and from below detection to 3.6% of total spectrum counts under N 2
headspaces (Fig. 4 B). A second predicted S-layer protein, gene 2635, was present with average
expression (for samples above detection) elevated under CO 2 (0.66 % of total spectrum counts)
when compared to N 2 (0.13% of total spectrum counts) (Phyre2 structural prediction: 27% of the
100
protein predicted with greater than 90% confidence). This protein was not detected in samples
from the third set of replicates with lower protein yields. While existing studies have
documented S-layer proteins are upregulated under elevated CO 2 and low pH conditions
(Chitlaru et al., 2006; Passalacqua et al., 2009; Sara and Sleytr, 2000), we find S-layer proteins
do not vary significantly across anaerobic headspace or pressure variables (nonparametric
Kruskal-Wallis test).
B
A
Putative S-layer protein
0.25
(Gene No. 2630)
0
0.24
0.23
0.22
0.21
0
0.2
0.05
'-I
CL
0.
0.04
0.03
000
E
z
0
0
0.02
0.01
0
8
0
0
0
1 atm
N2
100
atm
N2
1 atm
C02
100
atm
C02
Figure 4. A highly expressed putative S-layer protein (gene 2630) based on structural prediction
(A), and the corresponding expression values of gene 2630 in 12 MITOTI proteomes (B).
Samples from the third set of proteomes are shaded in gray. Structural prediction with Phyre2
(Kelley and Sternberg, 2009) predicted gene 2630 to be a S-layer protein with 42% of the protein
predicted with greater than 90% confidence. The N-terminal region, pictured in blue and teal
shades, shows high homology to the membrane-bound regions of S-layer proteins in other
Bacilli.
101
Table 2. Highly expressed proteins under 1 atm CO2
Rank
1 atm
C02
#2
1
1
10
2
Rank
1 atm
C02
#3
Rast Annotated Function
Subsystems
1
Hypothetical S-layer protein 2630
- none
2
Aconitate hydratase (EC 4.2.1.3) @ 2methylisocitrate dehydratase (EC
4.2.1.99)
Glyoxylate bypass; <br>Serine-glyoxylate
cycle; <br>Serine-glyoxylate cycle - GJO;
<br>TCA Cycle
-
Rank
1 atm
C02
#1*
3
4
3
Heat shock protein 60 family chaperone
GroEL
32
5
4
Translation elongation factor G
2
14
5
Translation elongation factor Tu
GroEL GroES; <br>ar-427-EC
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factor G family; <br>Translation
elongation factors bacterial
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factors bacterial
4
9
6
ATP synthase beta chain (EC 3.6.3.14)
FOF1-type ATP synthase
17
6
7
Chaperone protein DnaK
36
3
8
Heat shock dnaK gene cluster extended;
<br>Protein chaperones
CBSS-349161.4.peg.2427;
<br>Glycolysis and Gluconeogenesis;
<br>Pyruvate metabolism 1: anaplerotic
Pyruvate,phosphate dikinase (EC 2.7.9.1) reactions, PEP
Arginine and Ornithine Degradation;
<br>Proline, 4-hydroxyproline uptake and
utilization; <br>ar-245-EC
25
11
9
Delta-1 -pyrroline-5-carboxylate
dehydragenase (EC 1.5.1.12)
34
16
10
Inosine-5&#39;-monophosphate
dehydrogenase (EC 1.1.1.205)
Purine conversions; <br>Purine salvage
cluster
23
13
11
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>TCA Cycle
6
15
12
Succinyl-CoA ligase [ADP-forming] beta
chain (EC 6.2.1.5)
Unspecified monosaccharide ABC
transport system, substrate-binding
component / CD4+ T cell-stimulating
antigen, lipoprotein
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>Succinate
dehydrogenase; <br>TCA Cycle
-
- none
31
13
26
24
14
Alkyl hydroperoxide reductase protein C
(EC 1.6.4.-)
Thioredoxin-disulfide reductase
39
8
15
2-oxoglutarate oxidoreductase, alpha
subunit (EC 1.2.7.3)
- none
*Gray filled column is from the third set of proteomes
+Protein expression is displayed as rank expression
102
-
8
Succinate dehydrogenase flavoprotein
subunit (EC 1.3.99.1)
Table 3. Highly expressed proteins under 100 atm CO
2
14
Rank Rank
100
100
atm
atm
CO2
2
CO2
#3
1
1
2
2
Rast Annotated Function
Hypothetical S-layer protein 2630
Aconitate hydratase (EC 4.2.1.3) @ 2methylisocitrate dehydratase (EC
4.2.1.99)
Subsystems
- none
Glyoxylate bypass; <br>Serine-glyoxylate
cycle; <br>Serine-glyoxylate cycle - GJO;
<br>TCA Cycle
-
Rank
100
atm
CO2
#1*
6
Heat shock protein 60 family chaperone
2
3
3
GroEL
1
4
4
Translation elongation factor Tu
13
5
5
6
6
13
Translation elongation factor G
ATP synthase alpha chain (EC 3.6.3.14)
GroEL GroES; <br>ar-427-EC
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factors bacterial
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factor G family; <br>Translation
elongation factors bacterial
FOF1-type ATP synthase
Heat shock dnaK gene cluster extended;
7
5
4
8
12
<br>Protein chaperones
CBSS-349161.4.peg.2427;
<br>Glycolysis and Gluconeogenesis;
<br>Pyruvate metabolism 1: anaplerotic
Pyruvate,phosphate dikinase (EC 2.7.9.1) reactions, PEP
Chaperone protein DnaK
Succinyl-CoA ligase [ADP-forming] beta
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>TCA Cycle
FOF1-type ATP synthase
8
3
9
10
8
11
16
11
9
34
12
18
chain (EC 6.2.1.5)
ATP synthase beta chain (EC 3.6.3.14)
Unspecified monosaccharide ABC
transport system, substrate-binding
component / CD4+ T cell-stimulating
antigen, lipoprotein
Inosine-5&#39;-monophosphate
dehydrogenase (EC 1.1.1.205)
10
13
29
Succinate dehydrogenase flavoprotein
subunit (EC 1.3.99.1)
45
14
14
Translation initiation factor 2
- none
Purine conversions; <br>Purine salvage
cluster
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>Succinate
dehydrogenase; <br>TCA Cycle
CBSS-138119.3.peg.2719; <br>CBSS350688.3.peg.1509; <br>NusA-TFiI
Cluster; <br>Translation initiation factors
bacterial
-
12
oxidoreductase of aldo/keto reductase
15
15
- none
family, subgroup 1
-
21
*Gray filled column is from the third set of proteomes
+Protein expression is displayed as rank expression
103
Table 4. Highly expressed proteins under 1 atm N
2
Rank Rank Rank
1 atm 1 atm 1 atm
N 2 #1* N 2 #2 N2 #3 Rast Annotated Function
Subsystems
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factors bacterial
1
1
1
2
2
2
Translation elongation factor Tu
Heat shock protein 60 family chaperone
GroEL
26
3
3
Hypothetical S-layer protein 2630
- none
5
4
7
ATP synthase alpha chain (EC 3.6.3.14)
FOF1-type ATP synthase
11
5
6
38
6
5
20
7
12
Chaperone protein DnaK
Succinyl-CoA ligase [ADP-forming] beta
chain (EC 6.2.1.5)
Aconitate hydratase (EC 4.2.1.3) @ 2methylisocitrate dehydratase (EC
4.2.1.99)
Heat shock dnaK gene cluster extended;
<br>Protein chaperones
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>TCA Cycle
Glyoxylate bypass; <br>Serine-glyoxylate
cycle; <br>Serine-glyoxylate cycle - GJO;
<br>TCA Cycle
3
8
13
ATP synthase beta chain (EC 3.6.3.14)
27
9
8
12
10
14
Translation elongation factor G
Unspecified monosaccharide ABC
transport system, substrate-binding
component / CD4+ T cell-stimulating
antigen, lipoprotein
28
11
10
29
12
9
14
13
16
21
14
15
4
15
4
-
GroEL GroES; <br>ar-427-EC
F0F1-type ATP synthase
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factor G family; <br>Translation
elongation factors bacterial
-
- none
Purine conversions; <br>Purine salvage
I nosine-5&#39;-monophosphate
cluster
dehydrogenase (EC 1.1.1.205)
Tungsten-containing aldehyde:ferredoxin
oxidoreductase (EC 1.2.7.5)
Tungstate strays
DNA structural proteins, bacterial
Arginine and Ornithine Degradation;
<br>Proline, 4-hydroxyproline uptake and
Delta-1 -pyrroline-5-carboxylate
utilization; <br>ar-245-EC
dehydrogenase (EC 1.5.1.12)
CBSS-349161.4.peg.2427;
<br>Glycolysis and Gluconeogenesis;
<br>Pyruvate metabolism 1: anaplerotic
Pyruvate,phosphate dikinase (EC 2.7.9.1) reactions, PEP
DNA-binding protein HBsu
*Gray filled column is from the third set of proteomes
+Protein expression is displayed as rank expression
104
Table 5. Highly expressed proteins under 100 atm N
2
1
B.D.
Rank
100
atm
N 2 #2
Rank
100
atm
N 2 #3 Rast Annotated Function
Heat shock protein 60 family chaperone
1
2 GroEL
2
1 Hypothetical S-layer protein 2630
Subsystems
GroEL GroES; <br>ar-427-EC
- none
-
Rank
100
atm
N 2 #1*
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
2
3
4 Translation elongation factor Tu
factors bacterial
5
4
5 Chaperone protein DnaK
Heat shock dnaK gene cluster extended;
<br>Protein chaperones
4
5
7
6
11 ATP synthase alpha chain (EC 3.6.3.14)
Succinyl-CoA ligase [ADP-forming] beta
3 chain (EC 6.2.1.5)
3
7
8 ATP synthase beta chain (EC 3.6.3.14)
16
8
7 Translation elongation factor G
FOF1-type ATP synthase
Serine-glyoxylate cycle; <br>Serineglyoxylate cycle - GJO; <br>TCA Cycle
FOF1-type ATP synthase
Mycobacterium virulence operon involved
in protein synthesis (SSU ribosomal
proteins); <br>Translation elongation
factor G family; <br>Translation
elongation factors bacterial
Unspecified monosaccharide ABC
transport system, substrate-binding
component / CD4+ T cell-stimulating
9
10 antigen, lipoprotein
- none
Tungsten-containing aldehyde:ferredoxin
9 oxidoreductase (EC 1.2.7.5)
Tungstate strays
12
10
6
11
8
12
37
13
Inosine-5&#39;-monophosphate
12 dehydrogenase (EC 1.1.1.205)
Aconitate hydratase (EC 4.2.1.3) @ 2methylisocitrate dehydratase (EC
64.2.1.99)
oxidoreductase of aldo/keto reductase
13 family, subgroup 1
74
14
NAD-dependent glyceraldehyde-3Glycolysis and Gluconeogenesis;
28 phosphate dehydrogenase (EC 1.2.1.12) <br>Pyridoxin (Vitamin B6) Biosynthesis
10
15
16 LSU ribosomal protein L21p
-
35
Purine conversions; <br>Purine salvage
cluster
Glyoxylate bypass; <br>Serine-glyoxylate
cycle; <br>Serine-glyoxylate cycle - GJO;
<br>TCA Cycle
-
- none
Ribosome LSU bacterial
*Gray filled column is from the third set of proteomes
+Protein expression is displayed as rank expression
105
Growth phase biomarkers across MITOTI cultures are consistent with stationary phase
All bioreactors with observed biomass were within a factor of 5 of the maximum cell
density observed in all previous experiments carried out under anaerobic conditions, consistent
with these cultures being in, or entering, stationary phase (Supplemental Fig. 4 and 11).
Furthermore, no significant differences in viable counts were observed between either headspace
or pressure, which could have indicated differences in growth phase. As the percent of ribosomal
proteins expressed may vary with growth phase, we examined ribosomal protein abundance in
different headspace and pressures. Suggesting similar growth phase, the proportion of ribosomal
proteins varied by a factor of 2 or less across all MITOTi proteomes (Fig 5. A), while ribosomal
content in F. coli at 10 hours of growth is more than 5-10 fold higher than at 40 hours of growth
(Arnold et al., 1999). However, the ribosomal protein gene set was enriched in high pressure
incubated samples (GSEA p=0.00 6 , FDR<0.25) (Fig. 5 B), which may indicate that high
pressure samples were generally collected at an earlier point in stationary phase than low
pressure samples. Repeating PCA after removing ribosomal proteins from the dataset still
showed sample separation primarily by headspace and pressure along the first two principle
components (Supplemental Fig. 12). Additionally, it is important to note that while most
ribosomal proteins are more highly expressed in log phase, the L7/L12 ribosomal protein is more
highly expressed in stationary phase (Ramagopal, 1976), and this protein is expressed in all 12
MITOTI proteomes, but did not vary significantly by headspace or pressure (Kruskal-Wallis
test).
106
B
1 atm 100 atm
CO 2 N 2 CO 2 N 2 KEGG ID
A
K2*latmCO2
14%
-
12%
2
10%
Mlatm N2
100 atmCO2
Initial
Final
A
8%
u27
0
(
.0
*100atm N2
6%
4%
U
2%
5.E+04
0
5.E+05
5.E+06
5.E+07
Direct Counts (cells/ml)
Figure 5. The fraction of total expressed proteins that is composed of ribosomal proteins as a
function of cell counts per sample in (A). Ribosomal proteins are enriched in high pressure
samples based on GSEA (p= 0.006, FDR<0.25) (B). Samples from the third set of proteomes are
outlined in black. Higher direct counts were observed for high pressure samples (two-factor
ANOVA, p= 0.003, F-ratio= 18.1).
There are several other proteins that suggest these proteome samples are in the targeted
stationary phase (Supplemental Table 2). The codY regulon and related proteins are observed to
increase in stationary phase (Bernhardt et al., 2003), and we have observed codY in 11 of 12
samples (the exception was a replicate of I atm CO 2 with the fewest total proteins). We found
107
three different proteins involved in acetoin metabolism in these proteomes, which is notable as
acetoin is often utilized as a secondary carbon source during stationary phase (Xiao and Xu,
2007). Huang et al. (2012) observed multiple acetoin metabolism proteins in stationary phase
proteomes of B. thuringiensis,but no acetoin metabolism proteins in log phase cultures.
Additional proteins that indicate stationary growth phase include a carbon starvation protein
present in 10 of 12 samples, and the cell division protein, FtsH, present in 9 of 12 samples. FtsH
is induced in stationary phase, and also as a stress response protein (Fischer et al., 2002). Stress
response proteins are frequently involved in stationary phase and nutrient limitation, and two
cold shock proteins that are induced under stationary phase (Graumann and Marahiel, 1999)
were expressed in 10 of 12 and 8 of 12 samples.
Glycine cleavage system is enriched in CO 2 headspaces
To control for the potential effects of variability in culture age within stationary phase,
we focused on differences between gas headspace (i.e. pooling 1 atm and 100 atm samples from
each headspace). To examine differential expression at a pathway level, we used gene set
enrichment analysis (GSEA) of proteins classified by KEGG. We annotated 476 proteins (of 623
total) with KEGG and 43 gene sets (or pathways) were analyzed by GSEA. Six of 43 gene sets
were significantly enriched with respect to either headspace or pressure (p<0.05). The glycine,
serine, and threonine metabolism gene set was enriched under CO 2 headspace, while no gene sets
were significantly enriched under N 2 (Supplemental Table 3). Five of the proteins in the gene set
for glycine, serine, and threonine metabolism (Figure 6) comprise the glycine cleavage system,
which is involved in glycine and serine catabolism or synthesis, depending on which direction
the pathway operates in. The protein most highly correlated with CO 2 samples through PLS-DA
108
CO 2
1
.-
mn
100
N2
1
100
Glycine dehydrogenase [decarboxylating] (glycine cleavage system P2 protein) 1939
Dihydrolipoamide dehydrogenase of pyruvate dehydrogenase complex 929
Low-specificity L-threonine aldolase 1700
Glycine dehydrogenase [decarboxylating] (glycine cleavage system P1 protein) 1940
Aminomethyltransferase (glycine cleavage system T protein) 1941
2,3-bisphosphoglycerate-independent phosphoglycerate mutase 753
Serine hydroxymethyltransferase 175
2-amino-3-ketobutyrate coenzyme A ligase 3020
Aspartate-semialdehyde dehydrogenase 2880
Phosphoserine aminotransferase 885
Aspartokinase 320
Tryptophan synthase beta chain 3917
Figure 6. The KEGG gene set for glycine, serine and threonine metabolism is enriched under
CO 2 headspace samples via GSEA (nominal p = 0.018). This gene set includes proteins involved
in the Glycine Cleavage System (highlighted in gray). These 5 proteins are all positively
correlated with CO 2 headspace, (PLS-DA p(corr) of 0.98 for Glycine dehydrogenase P2 protein,
0.90 for Dihydrolipoamide dehydrogenase, 0.85 for Glycine dehydrogenase P1 protein 0.42 for
Aminomethyltransferase, and 0.40 for Serine hydroxymethyltransferase).
is glycine dehydrogenase (decarboxylating) enzyme, which is noteworthy as it can either
produce or consume CO 2 depending on which direction the reaction proceeds, and has been
previously shown to be upregulated in acid stressed E. coli (House et al., 2009). All 5 proteins
involved in the glycine cleavage system were positively correlated with CO 2 headspaces.
PLS-DA also revealed that several other proteins involved in amino acid metabolism,
including deblocking aminopeptidase and arginase, were highly correlated with CO 2 grown
cultures (Table 6 A). Amino acid metabolism is particularly noteworthy in relation to
acclimation to acid (or C0 2 ) stress, as there are a number amino acid metabolizing pathways that
can aid in acid neutralization through production of neutralizing compounds (Cotter, 2003;
Foster, 1999) or through consumption of protons (Cotter et al., 2001). A deblocking
aminopeptidase is potentially significant as it is involved in amino acid metabolism and it is
upregulated during heat and oxidative stress (Jia, 2011). Arginase is particularly notable as it is
involved in the H. pylori acid stress response by producing urea from arginine (McGee et al.,
109
1999). Additional proteins highly correlated with CO 2 from PLS-DA include stress response and
respiratory proteins (Table 6 A).
Table 6. Highly correlated proteins from PLS-DA: CO headspace (A), N 2 headspace
2
(B), low pressure (C), and hiqh pressure (D)
A.
RAST Annotated function
Glycine dehydrogenase [decarboxylating] (glycine cleavage
system P2 protein) (EC 1.4.4.2) 1939
Correlation
with CO 2
0.98
General Function
Amino acid
metabolism
Amino acid
metabolism
Methionyl-tRNA formyltransferase (EC 2.1.2.9) 1282
0.97
Hypothetical protein SAV1845 865
0.93
Deblocking aminopeptidase (EC 3.4.11.-) 2488
Heat shock protein GrpE 2140
FMN reductase (EC 1.5.1.29) 3000
3-ketoacyl-CoA thiolase [isoleucine degradation] (EC 2.3.1.16)
140
Cytochrome c oxidase polypeptide I (EC 1.9.3.1) 3243
Thioredoxin 3890
0.93
0.92
0.92
0.91
0.91
0.90
Amino acid
metabolism, protease
Stress response
Riboflavin metabolism
Amino acid
metabolism
Respiration
Respiration
Cold-shock DEAD-box protein A 4022
0.90
Stress response
Correlation
with N 2
General Function
0.97
Respiration
Transition state regulatory protein AbrB 2351
Diaminopimelate decarboxylase (EC 4.1.1.20) 1329
LSU ribosomal protein L23p (L23Ae) 916
LSU ribosomal protein L4p (Lie) 917
LSU ribosomal protein L17p 890
Succinyl-CoA ligase [ADP-forming] beta chain (EC 6.2.1.5) 3171
Menaguinone-cytochrome C reductase iron-sulfur subunit 2822
Universal stress protein family 1540
0.96
0.96
0.95
0.95
0.95
0.95
0.93
0.93
Cell cycle control
Amino acid metabolism
Ribosomal
Ribosomal
Ribosomal
TCA cycle
Respiration
Stress response
hypothetical protein 2109
0.93
B. RAST Annotated function
Menaquinone-cytochrome c reductase, cytochrome B subunit
2823
110
Correlation
C.
with low
pressure
RAST Annotated function
General Function
Fatty acid/ non-polar
Amino acid
metabolism
Respiration
Glycolysis
Fatty acid biosynthesis
Butyryl-CoA dehydrogenase (EC 1.3.99.2) 137
NADH-ubiquinone oxidoreductase chain B (EC 1.6.5.3) 157
Glucokinase (EC 2.7.1.2) 337
3-oxoacyl-[acyl-carrier protein] reductase (EC 1.1.1.100) 1297
oxidoreductase domain protein 1836
0.97
0.93
0.93
0.93
0.92
Transamidase GatB domain protein 3830
Amino acid ABC transporter, periplasmic amino acid-binding
protein 115
0.92
Enoyl-CoA hydratase (EC 4.2.1.17) 571
Molybdopterin biosynthesis protein MoeA / Periplasmic
molybdate-binding domain 694
Phosphate transport system regulatory protein PhoU 353
0.90
0.90
0.90
Fatty acid metabolism
molybdenum cofactor
biosynthesis
Transport, phosphate
D.
Correlation
with high
pressure
General Function
RAST Annotated function
0.91
Amino acid
metabolism
Transport, Amino
acids
Heat shock protein 60 family chaperone GroEL 3377
oxidoreductase of aldo/keto reductase family, subgroup 1 2597
LSU ribosomal protein L21p 251
Unspecified monosaccharide ABC transport system, substratebinding component / CD4+ T cell-stimulating antigen, lipoprotein
2889
Translation elongation factor P 802
SSU_ ribosomal protein S p 2243
0.90
Ribosomal
LSU ribosomal protein L24p (L26e) 907
0.96
0.93
0.92
Stress response,
0.92
0.90
Transport, sugar
Translation
0.90
S-adenosylmethionine synthetase (EC 2.5.1.6) 2128
GTP-sensing transcriptional pleiotropic repressor codY 3163
hypothetical protein 976
0.89
Ribosomal
Amino acid
metabolism,
secondary metabolites
Cell cycle control
0.89
Ribosomal
0.88
Several proteins show differential expression between N 2 and CO 2 conditions. Chaperone
Dnak, ATP-dependent Clp protease ATP-binding subunit ClpX, ATP-dependent hsl protease
ATP-binding subunit, and a Universal stress protein are all involved in stress response and/or
protein folding, and all show significantly lower expression under CO 2 (Kruskal-Wallis test,
p<0.05) (Supplemental Fig. 13 A-D). This is a puzzling observation, as normally these proteins
would be increased during stress response, for example DnaK and ATP-dependent Clp proteases
111
have been shown to be elevated in acid stressed Streptococcus mutans cells (Len et al., 2004).
Clp proteases have also been found to be elevated under high-pressure stress (Hormann et al.,
2006), and while this Clp protease is slightly elevated under pressure, it is not significantly so.
Additional proteins with differential expression between N 2 and CO 2 (Kruskal-Wallis
test, p<0.05) include several that are involved in energy generation pathways (Supplemental Fig.
13 E-G). Aconitase hydratase is a TCA cycle protein that is abundant in all conditions, although
it is significantly higher under CO 2 headspaces. However, other TCA cycle enzymes do not show
significantly different expression under CO 2 or N 2 . An electron transfer flavoprotein, beta
subunit is significantly reduced under CO 2 , with another electron transfer flavoprotein showing
similar patterns of expression, but not significantly so. One interesting protein that shows
significantly lower expression under CO 2 is a tungsten-containing aldehyde:ferredoxin
oxidoreductase. This protein is oxygen sensitive and is frequently found in both archaea and
anaerobic bacterial species, and while its exact substrate is unknown, it is suggested to be
involved in carbon utilization (Kercher and Oesterhelt, 1982).
4.4 Discussion
Much of the current literature on microbial response to supercritical CO 2 involves
shocking vegetative cells with rapid increases in CO 2 content, pressure, and acidity which leads
to a loss in cell viability (Garcia-Gonzalenz et al., 2009; Tamburini et al., 2014; Zhang et al.,
2006; Hong and Pyun et al., 1999; Dillow et al., 1999). In contrast, the model of growth we have
studied here entails inoculation with spores, where germination occurs after scCO 2 addition, and
resulting protein and lipid profiles reflect acclimated growth. Global protein expression across
headspace and pressure conditions generally indicate that cells in all conditions are in stationary
112
phase, and appear to have acclimated to growth under their respective environment. Proteomes
are similar to other stationary phase Bacillus proteomes, which is notable as it suggests that B.
subterraneusMITOTI can acclimate to a relatively 'normal' lifestyle and maintain
housekeeping metabolic processes under a scCO 2 headspace. In this study we have examined
how proteins and lipids vary in different headspace and presssue conditions to gain insight into
.
how cells acclimate to growth under sCCO 2
Analysis of acclimation to scCO 2 through lipid profiling supports the hypothesis that
resistance to CO 2 stress is similar to acid stress response (Petrackova et al., 2010), and that
increasing membrane rigidity is important for acclimating to growth under scCO 2 . These changes
manifested in a decreased proportion of branched lipids and increased average acyl chain lengths
under scCO 2 in strain MITOTI, with qualitatively similar trends observed in MIT0214.
However, MITOTI showed a stronger response to high pressure C0 2, with scCO 2 grown
samples showing significantly fewer branched lipids than 1 atm CO 2 grown samples, while
MIT0214 did not show a reduction in branching in scCO 2 samples compared to 1 atm CO 2. Both
strains showed the longest average lipid chain lengths under 100 atm C0 2, with MITOTI
showing significantly longer chain lengths under scCO 2 when compared to 1 atm N 2 or 1 atm
ambient samples. These membrane changes observed in cells grown under sCCO 2 suggest that
modulation of membrane lipids to form less fluid membranes may be an important acclimation
.
mechanism in response to the membrane permeabilizing effects of scCO 2
Proteomic analysis of B. subterraneusMITOT1 revealed an abundance of expressed
TCA cycle and electron transport chain proteins across conditions suggest that it may be
anaerobically respiring, and the probable electron acceptors are oxidized metals (FeIII and MnIV
are provided in the MS growth media used for this study). MITOTI is also most closely related
113
(via 16S rRNA) to a clade of Bacilli capable of anaerobic metal-reduction of Fe(III), Mn(IV),
Se(VI) and As(V) (Kanso et al. 2002; Yamamura et al. 2007; Boone et al., 1995), suggesting this
ability may be conserved in strain MITOTI. Other highly expressed protein functions across
variable headspace and pressures include general stress response, translation, ribosomal, amino
acid metabolism, glycolysis / gluconeogenesis, and putative S-layer proteins.
While our data do not reveal significant differences in the abundance of predicted S-layer
proteins between N 2 and CO 2, uniformly high expression under C0 2, slightly lower expression
under N 2, and existing literature indicate that S-layers may facilitate acclimation to sCCO 2 . The
abundance of S-layer proteins may be influenced by environmental conditions. Certain S-layer
proteins are increased during oxygen limitation, in B. stearothermophilus(Sara et al., 1996), and
in studies of aerobic growth in B. anthracisand B. cereus group strains, S-layer protein
expression increased when CO 2 concentrations were raised greater than 300-fold above the
ambient background (i.e. to a 14-15% CO 2 atmosphere) (Chitlaru et al., 2006; Passalacqua et al.,
2009). Additionally, S-layer proteins are elevated under other stressed conditions (heat and acid
stress) in Lactobacillus acidophilus (Khaleghi et al., 2012), indicating that S-layer production is
not solely a function of headspace composition. S-layers also play a role in cell adhesion,
virulence, and membrane stability as they form a protein lattice on the cell surface (Sara and
Sleytr, 2000; Schuster et al., 2008). Given the stresses that scCO 2 imposes (e.g. cytoplasm
acidification and membrane permeabilization), and the changes observed in membrane lipids, it
stands to reason that universally high S-layer production may predispose MITOTI to surviving
and growing under scCO 2 containing conditions.
Principal component and clustering analyses indicated that both headspace and pressure
help explain variability in MITOTI proteomes, with samples separating by headspace along the
114
first principal component and by pressure along the second principal component. Interestingly
the protein profiles of cells grown under scCO 2 appears to be intermediate between low pressure
CO 2 and high pressure N 2, emphasizing that acidity and pressure may have some opposing
effects.
Observed cell densities and the presence of numerous stationary phase proteins across
conditions indicate that MITOTI lipids and proteomes were sampled in stationary phase,
consistent with the targeted relative culture age of 7-12 days post entry into stationary phase
based on growth curves for MITOTI under 1 atm of pressure and the conditional probability of
growth under 100 atm pressure. Ribosomal protein-content differed by 2-fold or less in
proteomes sampled from the four tested conditions, also pointing to similar growth phases across
cultures, as ribosomal proteins are known to vary significantly (5-10 fold) between exponential
and stationary phases (Arnold et al., 1999). Despite this low magnitude, enrichment of ribosomal
proteins in cultures incubated under high pressure, may suggest variability associated with
absolute culture age (i.e. 21 versus 30 days), growth phase, or acclimation to elevated pressure.
To control for potential variability associated with culture age high and low pressure samples
were pooled to evaluate differences associated with each headspace.
The protein most highly correlated with the CO 2 headspace condition was glycine
dehydrogenase, a component of the glycine cleavage system, which has been demonstrated to be
upregulated in acid stressed E. coli (House et al., 2009). Enrichment of amino acid metabolism,
in the CO 2 condition, in particular four additional proteins mediating the glycine cleavage
system, suggests this system may play a role in CO 2 acclimation. Two other proteins enriched
under CO 2 conditions, arginase and deblocking aminopeptidase, have been previously shown to
be pH or stress-responsive. Arginase is associated with pH neutralization (McGee et al., 1999)
115
and deblocking aminopeptidase is upregulated during heat and oxidative stress (Jia et al., 2011).
These amino acid metabolic proteins represent future targets for understanding how cells respond
to elevated CO 2 stress.
The stationary phase cultures examined in this study reflect cells that are no longer
growing, and share some similarities with natural settings where a large portion of cells are static
i.e. in dormant or very slow growing states (Brock, 1971; Koch, 1997). However, stationary
phase cultures have some fundamental differences from static cultures. In stationary phase
cultures, growth ceases due to exhaustion of nutrients and build-up of metabolic end products as
no removal of biomass or culture media occurs, but in static populations the influx of nutrients is
too low to support growth and division, while the environment is continually diluted, preventing
build-up of toxic end products. Bacteria grown under continuously-diluted culture conditions
mimicking static conditions reveal lower expression of DNA and protein repair than stationary
phase cultures, which is suggested to be a result of conditions in static cultivation that do not
degrade cells like in stationary phase (Overkamp et al., 2014). However, many similarities still
exist, as both non-growing populations must downregulate ribosomal proteins and energy
generation pathways as growth rate slows, suggesting that stationary phase cultures can
approximate some aspects of static populations.
While doubling times for static subsurface microbial populations are estimated to exceed
hundreds of years (Phelps et al., 1994), these populations can influence the biogeochemistry of
the subsurface and may respond to perturbations in their environment. Stimulation of static
microbial populations may be relevant in the context of various subsurface industrial activities
including enhanced oil recovery, hydraulic fracturing, and geologic carbon sequestration. For
example, active subsurface communities can affect variables such as porosity and permeability
116
through mineral dissolution and nucleation (Mitchell et al. 2010; Barker et al. 1998; Cunningham
et al. 2009). Additionally, cells in dormant or stationary phase are more resistant to stresses
(Ferrerira et al., 2003; Hong, 1999; Watanabe et al; 2003), thus environments targeted for scCO 2
injection may have populations that are predisposed to survive and grow after scCO 2 exposure.
Indeed, it has been inferred that microbial communities in the deep subsurface may be
acclimating to influxes of sCCO 2 through changes in community composition (Morozova et al.,
2011; Mu et al., 2014).
The demonstration of microbial growth under a scCO 2 headspace calls into question the
use of scCO 2 as a food and medical sterilizing agent where spores may be present, but it is
encouraging for the development of applications involving scCO 2 including bioengineering in
GCS and biocatalysis under scCO 2 . Biocatalysis under scCO 2 is currently conducted with
enzymes or inactivated cells (Wimmer and Zarevucka, 2010). Biphasic reactors containing
scCO 2 and an aqueous (or other solvent) phase have been explored for extraction of various
biologically produced compounds that partition into the scCO 2 phase (Baiker, 1999; Knutson et
al., 1999). This study provides insights into how live cells maintain biocompatibility with scCO 2
and provides candidate targets to improve the growth of non-biocompatible strains. Membrane
lipid modifications to create a more rigid membrane and activity of the glycine cleavage system
may help cells acclimate to scCO 2. In addition, we hypothesize that introduction of cells as
spores to scCO 2 systems, followed by germination of acclimated cells may help Bacilli tolerate
the complex stresses associated with scCO 2. New opportunities for biotechnology development
in biofuels, pharmaceuticals, and other industries will be possible with microbes that are able to
grow in an aqueous phase that is concurrently being extracted with scCO 2. For any of these
applications to be realized in biphasic reactors or in natural environments targeted for GCS, the
117
continued development of supercritical CO 2 tolerant microorganisms like B. subterraneus
MITOTI or B. cereus MIT0214 is crucial.
118
4.5 Supplemental Figures and Tables
i
1.0
Original Probability
Conditional probability
-_
-
0.9
0.8
-_-
-_
0.6
0.5
0.4
0.3
0.2
0.1
0.0
5
10
15
20
25
30
Time (days)
Supplemental Figure 1. Predicted probability of growth for strain MITOTI under a scCO
2
headspace. The gray line displays the logistic regression model probability of growth (from Peet
et al., In Review), while the black line displays the conditional probability of growth at an earlier
time point, based on observed growth at 30 days.
119
A 1.OE+06
1.0E+05
1.0E+04
0
10
5
15
20
25
20
25
Time (days)
B
Of
5.OE+06
lop
'
5. OE+05
5.0E+04
5.0E+03
0
5
10
15
Time (days)
Supplemental Figure 2. Growth curves of B. subterraneusMITOTI under 1 atm N 2 (A) and 1
atm CO 2 (B). Anaerobic germination and growth of MITOT1 is variable, with maximum viable
counts obtained between 9 and 14 days.
120
0.6
0.5
I
PC
0.4
0.3
N
0
0
0.2
I
T-
T
a
0.1
0
MITOT1
Spores
MIT0214
Spores
I
I
LB Media
Extraction
Blanks
8 n16
n18
i
MR media
Supplemental Figure 3. Triplicate normalized values of two lipids (n 16 and n 18) that were
present in all samples, including extraction blanks. These lipids do not appear to be a result of
trace amounts of growth media or from residual spores in samples that did not show growth, as
extraction blanks of solvents resulted in similar amount of these two lipids. The peak areas of
these contaminants were normalized to the internal standard and subtracted from n 16 and n 18
peaks in MIT0214 and MITOTI samples.
* MITOT1 Cell/ml
MITOT1 CFU/mI
A MIT0214 Cell/ml
MIT0214 CFU/mI
1.OE+09
0
1.OE+08
A.?
LL
1.OE+07
1.QE+06
E
1.OE+05
o1.OE+04
CN
E
E
E
CN
0
E
0C
E
0
(O
CO
0
0
Supplemental Figure 4. Direct and viable counts for MITOT 1 (blue) and MIT0214 (orange)
samples used in proteome and lipid analyses.
121
0 100atmC02
M100atmN2
latmCO2
latmN2
latmAmbient
A
*
0.5
0.4
**
VI
0.3
0
U
ILL
0.2
T
i14
n
i
n14
i15
* *
-
0 .1
al5 ni5 n16:1 116 n16:1 n16 n17:1n17:1
M100atmN2
latmCO2
U 100atmCO2
|17
latmN2
a17
*
0
n17 n18:1n18:1n18:1 ni 8
S latmAmbient
B
0.7
I
0.6
T
0.5
0.
0.4
0.3
**
*
0
0
0
*
4-0
T
*
0.1
*
0.2
0
113
114
n14
i15
a15
16
n16
i17
a17
n17
n18:1 n18:1
n18
Supplemental Figure 5. Total lipid profiles of MITOTI (A) and MIT0214 (B). Significance
described in the text is denoted with (*). Iso and anteiso branched lipids are denoted with 'i' and
'a' respectively, while non branched lipids are denoted with 'n', and the number of unsaturations
is noted after the ':' (if unsaturated).
122
Supplemental Figure 5. (continued) B. cereus MIT0214's lipid profile is dominated by branched
lipids (92% of lipids are iso or anteiso branched), with i15:0, i17:0, and i16:0 comprising 46%,
16%, and 11% of total lipids, respectively (Fig. 1 B, Supplemental Fig. 4 B). However,
anaerobically grown cultures revealed a marked decrease in total branched lipids under all
conditions, containing 13-32% branched lipids (t-test: p<0.007 for each of the 4 anaerobic
conditions) (Fig. 1 B). Analysis of individual lipids reveals significantly fewer of the branched
lipids i14:0, il5:0, i16:0, and i17:0 under the 4 anaerobic conditions tested (t-test p<0.03 for each
lipid). Additionally, a15:0 was significantly reduced under all anaerobic conditions except 100
atm N 2 (t-test p<O.04) and a17:0 was significantly reduced under all anaerobic conditions except
1 atm N 2 (t-test p<0.02). The reductions in branched lipids in anaerobically grown MIT0214
resulted in a significant increase in the fraction of straight lipids: n16:0 under all 4 anaerobic
conditions (t-test p<O.004), and n18:0 under 1 atm N2 and 1 atm CO2 (t-test p<0.05)
(Supplemental Fig. 4 B).
High pressure results in limited differences in MITO214 branched lipid patterns. Under
high-pressure anaerobic headspaces, the pattern in decreased branching from N2 to CO is
2
generally consistent with several lipids also showing reductions under high pressure. Significant
variation was observed in i17:0, a17:0, and n17:0 with lower values in high pressure samples for
each lipid (two-factor ANOVA il7:0 p=0.01, F-ratio=11.5; a17:0 p<0.0001, F-ratio= 111.8; n17:0
p<O.0001, F-ratio=423.9) (Supplemental Fig. 4 B). While we might expect an increase in the
branched lipids i17:0 and a17:0 under high pressure to counteract membrane compressing effects
of pressure, we did not observe this, although the presence of CO2 and acidity is a confounding
factor for 100 atm CO2 samples.
E100atmCO2
M100atmN2
-latmCO2
(OlatmN2
NlatmAmbient
1.4
1.2
1
0.8
0
-*
0.6
0.4
*
4
-
0.2
0"
Iso branched
Anteiso
branched
Total branched
Straight
Saturated
Unsaturated
Supplemental Figure 6. Major lipid classes of MIT0214 sampled in stationary phase under 5
headspace and pressure conditions. Significance described in the text is denoted with (*). Iso and
anteiso branched lipids are summed to make up 'Total branched' lipids. Total branched lipids and
straight lipids and summed to make up saturated lipids. Saturated and unsaturated lipids sum to 1
for each sample.
123
17
I *4
16.8
16.6
S100atmCO2
16.4
llatmCO2
16.2
* 100atmN2
16
15.8
4
llatmN2
15.6
E latmAmbient
15.4
15.2
15
Supplemental Figure 7. Average lipid chain lengths of MIT0214 sampled in stationary phase
under 5 headspaces and pressure conditions. Significance described in the text is denoted with
(*). 1 atm CO2 and 1 atm N 2 chain lengths are significantly greater than 1 atm ambient samples
(p=0.00I and p=0.0006, respectively).
*1 atm CO2
* 1 atm N2
L100 atm CO2
100 atm N2
700
600
4-a
500
400
0
0
300
200
100
0
0
1000
2000
3000
4000
5000
6000
Total Spectral Counts
Supplemental Figure 8. Sampling coverage for 12 MITOTI proteome samples: Total proteins
detected per sample vs. Total spectral counts (i.e. protein expression based on Scaffold software
quantification). Samples from the third set of proteomes are outlined in black. Only two
replicates grown under 1 atm CO 2 appear to be leveling off in depth of proteins obtained.
124
Group average
Resemblance Spearman rank correlation
0.5--
0-6-
0.7-
-
-:
U_
0
090
E
E
E
E
0
0
0
C
Samples
Supplemental Figure 9. Clustering analysis of MITOTI proteomes based on Spearman
E9
9EE U by '*'are the third set of proteomes processed.
E
correlations. Samples
denoted
003
M
00
MA
c
X
LU
0
z
0
OCC0
0
Z
o
NNADH
jNADH-ubiquinone
I6
CU
0~
~2-annotated
AUC0 function
Aconitate hydratase (EC 4.2.1.3) @ 2-methylisocitrate dehydratase (EC 4.2.1.99) 3012
Succinyl-CoA ligase [ADP-forming] beta chain (EC 6.2.1.5) 3171
Succinate dehydrogenase flavoprotein subunit (EC 1.3.99.1) 317
Isocitrate dehydrogenase [NADP] (EC 1.1.1.42) 533
Citrate synthase (si) (EC 2.3.3.1) 532
Succinate dehydrogenase iron-sulfur protein (EC 1.3.99.1) 316
Malate dehydrogenase (EC 1.1.1.37) 534
Succinyl-CoA ligase [ADP-forming] alpha chain (EC 6.2.1.5) 3170
Succinyl-CoA ligase [ADP-forming] beta chain (EC 6.2.1.5) 3171
Dihydrolipoamide dehydrogenase of pyruvate dehydrogenase complex (EC 1.8.1.4) 929
"Fumarate hydratase class I, aerobic (EC 4.2.1.2) 2149"
Fumarate hydratase class I (EC 4.2.1.2) 855
Dihydrolipoamide succinyltransferase component (E2) of 2-oxoglutarate dehydrogenase complex (EC 2.3.1.61)
2-oxoglutarate dehydrogenase El component (EC 1.2.4.2) 3921
NADH-ubiquinone oxidoreductase chain D (EC 1.6.5.3) 2902
dehydrogenase (EC 1.6.99.3) 2704
oxidoreductase chain C (EC 1.6.5.3) 2901
NADH-ubiquinone oxidoreductase chain H (EC 1.6.5.3) 2903
NADH-ubiquinone oxidoreductase chain L (EC 1.6.5.3) 2907
NADH dehydrogenase (EC 1.6.99.3) 2710
NADH-ubiquinone oxidoreductase chain B (EC 1.6.5.3) 157
Cytochrome c oxidase polypeptide I (EC 1.9.3.1) 3243
Menaquinone-cytochrome C reductase iron-sulfur subunit 2822
Cytochrome c oxidase polypeptide II (EC 1.9.3.1) 3242
"Menaquinone-cytochrome c reductase, cytochrome B subunit 2823"
M"Menaouinone-cvtochrome C oxidoreductase. cvtochrome C subunit 2824"
Supplemental Figure 10. Respiratory proteins are highly expressed across all headspace and
pressure conditions. Both TCA cycle and electron transport chain proteins are expressed.
125
*100atmCO2-proteome
* 100atmCO2
M100atmN2-proteome
X latmCO2-proteome
A 100atmN2
0 latmN2-proteome
1.OE+07
U
1.OE+06
9
O1.OE+05
1.OE+04
1.OE+04
1.OE+05
1.OE+06
1.OE+07
1.OE+08
Direct counts (cells/ml)
Supplemental Figure 11. Viable vs. direct counts of MITOT1. Typical final direct counts vary
between 20 and 100 fold fold more than viable counts, however direct counts may be as much as
1000 fold higher than viable counts, which may potentially be linked to culture age. Samples
analyzed in proteomes are noted in the legend, while other samples show counts from other
experiments. Initial spore inocula direct and viable counts are also displayed for reference.
126
2
-
1 atm N 2 B
1 atm N 2 A
-
1 atm CO 2 C
1
-
0-
-
V
100 atm N2 A
A
1atmCO 2 B
V
100 atm CO 2 D
0
a_
-1
100 atm N 2 C
A
-2-
100 atmCO 2 J
-
-3-
-2
-2
-1
0
I
2
3
PC1
Supplemental Figure 12. Principal component analysis of duplicate MITOTi proteomes with
ribosomal proteins removed from analysis. Samples grown under 1 atm CO2 and 1 atm N 2, were
sampled at 21 days; Samples grown under 100 atm CO2 (scCO 2) and 100 atm N2 were sampled
at 30 days.
127
B
0.03
ATP-dependent Cip protease ATPbinding subunit CIpX 295
-
Chaperone protein DnaK 2139
0.012
C
0.025
0
V
o
0.015
.
0
0.008
0
.t
-0
z0 0.005
z
N2
1 atm
C02
D
3
0.004
0
7
a
0.01
0
0
0.0c
+----
--
,--
E 0.005
z
I
atm
C02
0
0
N2
F
0
0
1 atm
C02
100 atm
C02
Electron transfer flavoprotein, beta
subunit 1188
0.016
0.025
0.014
0
0.012
0.02
o00
CL
S0.015
0.01
0.008
S
0.006
00
0
0.004
0.005
z
a
1 atm N2 100 atm
1 0 atm
C02
Aconitate hydratase (EC 4.2.1.3) @
2-methylisocitrate dehydratase (EC
4.2.1.99) 3012
0
0
0
0
-*--
1 atm N2 100 atm
N2
0
10
0.015
0.0c
E
0.02
0
U
3
0
100 atm
C02
0.025
0
.
U
1 atm
C02
Universal stress protein family 1540
binding subunit HsIU 3164
CL
-~Q
1 atm N2 100 atm
N2
100 atm
C02
ATP-dependent hsl protease ATP0.009
0.008
0.007
0.006
0.005
8
a
0.002
0
1 atm N2 100 atm
IA
0
0.004
0-
C
0
U
a5 0.006
0.01
-
0
0.01
0
0.02
-
A
0
z
0
0
a
0.002
0
0
1 atm N2 100 atm
N2
1 atm
C02
1 atm N2 100 atm
100 atm
C02
N2
1 atm
C02
100 atm
C02
Supplemental Figure 13. Proteins with significantly different expression proteins with respect to
headspace (A-G) and pressure (H-J). Samples from the third set of proteomes are shaded in gray.
128
Tungsten-containing
aldehyde:ferredoxin oxidoreductase
(EC 1.2.7.5) 1557
H
Succinyl-CoA ligase [ADP-forming]
beta chain (EC 6.2.1.5) 3171
0.025
-
G
0
0.018
0.016
0 0.014
0.012
0.01
U) 0.008
CL 0.006
0.004
0.002
zG
0
0
0.02
In
M
8
0
0
0
'a0.01
atm N2 100 atm
N2
1 atm
100 atm
C02
C02
Translation elongation factor G 592
0.
1 atm N2 100 atm
N2
J
-
0
U
:.-
(U
0.014
0.012
0.01
0
-I
0
I
M
0.008
0.
0.006
0.004
z
0.002
18
IA
100 atm
C02
C02
0.014
0.012
0
0
0.01
0.008
0.006
-1
4
1 atm
Delta-1-pyrroline-5-carboxylate
dehydrogenase (EC 1.5.1.12) 3105
0.018
0.016
8
0
0
I
I
0.015
CL
-
4..
0.
0
z
0.004
0*O
8
8
0.002
0
1 atm N2 100 atm
N2
1 atm
C02
100 atm
C02
1 atm N2 100 atm
N2
1 atm
100 atm
C02
C02
Supplemental Figure 13 (continued). Proteins with significantly different expression proteins
with respect to headspace (A-G) and pressure (H-J). Samples from the third set of proteomes are
shaded in gray.
129
Supplemental Table 1. Top 20 BLASTx resuIts for gene 2630
Query
Cover E value % ID
Accession No.
Function
Species
ref|WP_026078567.11
hypothetical protein
Geobacillus caldoxylosilyticus
92 4.E-158
44
reflWP 020961287.11
hypothetical protein
Geobacillus sp. JF8
98 4.E-146
41
reflWP_009792711.11
surface layer (S-layer)
glycoprotein
Bacillus infantis NRRL B-14911
98 4.E-143
44
refIWP 025785866.11
hypothetical protein
Sporosarcina sp. D27
95 4.E-93
34
refjWP 013779879.11
cell wall binding repeat 2Mahella australiensis
containing proteins
63 4.E-85
41
refIWP_028396020.11
hypothetical protein
Bacillus sp. FJAT-14578
91
2.E-84
35
dbjIGAE48302.1I
S-layer protein precursor
Bacillus boroniphilus JCM
21738
75
2.E-63
35
refIWP 023613649.11
hypothetical protein
Bacillus sp. 17376
75
8.E-63
35
reflWP_018394794.11
hypothetical protein
Bacillus sp. 37MA
73
3.E-58
36
refIWP_016740221.11
hypothetical protein
Brevibacillus brevis
71
3.E-56
36
gbIEWG09306.1|
hypothetical protein
PBF 20073
Bacillus firmus DS1
74
1.E-55
34
refIWP_015893584.11
hypothetical protein
Brevibacillus brevis
69
7.E-55
35
refIWP 008407408.11
Parasporal protein
Bacillus isronensis
83
2.E-52
34
reflWP_002017639.11
hypothetical protein
Bacillus cereus
93
3.E-49
28
refiWP_006904189.11
S-layer domain containing
Thermaerobacter subterraneus
protein
83
5.E-47
31
dbjIGAJ39301.11
hypothetical protein
GCA01S 016 00260
Geobacillus caldoxylosilyticus
NBRC 107762
26
2.E-46
53
ref|WP_015865152.11
S-layer protein
Geobacillus sp. WCH70
26
1.E-44
51
refIWP 016428085.11
hypothetical protein
Paenisporosarcina sp.
HGH0030
71
2.E-43
34
gbIAAX46285.11
SgtA precursor
Anoxybacillus tepidamans
69
8.E-43
33
gblAAL46630.1
surface layer glycoprotein
SgsE precursor
Geobacillus stearothermophilus
62
1.E-42
33
130
Supplemental Table 2. Ex ressed proteins that correlate with stationary phase
100
I atm 1 atm 1 atm 1 atm
atm
C02 C02 C02 N2 #1 1 atm 1 atm C02
Function
LSU ribosomal
protein L7/L12
(P1/P2)
GTP-sensing
transcriptional
pleiotropic
repressor codY
Carbon
starvation
protein A
Cell division
protein FtsH
(EC 3.4.24.-)
Acetoin
dehydrogenase
El component
beta-subunit
(EC 1.2.4.-)
Acetoin
dehydrogenase
El component
beta-subunit
(EC 1.2.4.-)
Hypothetical
protein AcuB,
not involved in
acetoin
utilization
cold-shock
protein
cold-shock
protein
#1*
#2
#3
*
12
106
101
18
31
ND
107
84
44
66
306
283
ND
142
ND
N2#2 N2#3
100
atm
C02
100
100
atm
atm
C02 N2 #1
100
atm
100
atm
#1*
#2
#3
*
40
11
44
41
14
26
23
26
39
95
33
32
58
19
17
37
309
274
86
247
ND
46
ND
342
61
63
110
96
ND
81
188
ND
88
79
516
261
ND
ND
407
ND
ND
ND
ND
ND
262
ND
320
559
ND
ND
ND
ND
ND
324
ND
ND
263
ND
439
420
ND
284
213
ND
216
260
ND
292
313
ND
318
465
ND
208
307
134
257
143
142
237
228
ND
441
482
ND
294
332
ND
468
314
ND
325
330
*Gray filled columns are the third set of proteomes with lower protein abundances
+Protein expression is displayed as rank expression
131
N2#2 N2#3
Supplement Table 3. GSEA summary of ene sets with p<0.05
KEGG Gene Set #
Enriched under CO
2
260
Enriched under N
2
# genes
p-value
FDR
12
0.0179558
0.5784708
Gene Set
Glycine Serine and Threonine
metabolism
None with p-val < 0.05
Enriched under High Pressure
3010
40
0.00617284
0.1278369
Ribosomal
Starch and sucrose
metabolism
Enriched under Low Pressure
500
7
0.007716049
0.3099797
670
7
0.018376723
0.2805447
330
11
0.025641026
0.2643317
260
12
0.03003003
0.2964778
132
One carbon pool by Folate
Arginine and Proline
metabolism
Glycine Serine and Threonine
metabolism
Chapter 5. Conclusions
This thesis is the first demonstration of microbial growth in biphasic systems with a
supercritical CO 2 headspace. Through successive enrichment cultivation under scCO 2 , I have
isolated six Bacillus strains capable of growth under a scCO 2 headspace, corresponding to the
species: B. cereus, B. subterraneus, B. amyloliquefaciens, B. safensis (two isolates), and B.
megaterium. These isolates, and three Bacillus type strains (B. subtilis PY79, B. cereus ATCC
14579, and B. mojavensis JF-2) are all capable of growth under scCO 2 , suggesting that the ability
to grow under scCO 2 may be a widespread trait among Bacilli, extending the known range of
growth of Bacilli. Moreover, as these strains are all spore forming organisms, this suggests that
other spore forming organisms may be capable of growth under sCCO 2 . This is supported by a
study that conducted incubations of subsurface formation fluids under scCO 2 and observed
enrichment of Clostridiales after release of the scCO 2 headspace (Frerichs et al., 2014). These
results have implications for what organisms may be suitable for use in biocatalysis and
bioengineering in geologic carbon sequestration, as non-spore forming organisms may not
survive the initial stresses that accompany scCO 2 exposure.
Analysis of acclimation mechanisms to scCO 2 through lipid profiling confirms my
hypothesis that resistance to CO 2 stress is similar to the low pH stress response documented in
literature (Petrackova et al., 2010). These changes manifest in decreased branching and increased
average acyl chain lengths of lipids under CO 2 in both MIT0214 and MITOTI. However,
MITOT 1 shows a stronger response to pressure, with scCO2 grown samples showing fewer
branched lipids than low pressure CO 2 grown samples. These membrane changes suggest that
alteration membrane lipids to form less fluid membranes may be an important acclimation
133
mechanism in response scCO 2. A highly expressed putative S-layer protein in the proteome of
MITOT 1 further highlights the importance of the cell membrane in response to scC02, as these
proteins are elevated in CO 2 and acid stress in other Bacilli (Passalacqua et al., 2009; Khaleghi et
al., 2012). These findings support our hypothesis that modification of cell membrane and cell
wall components are crucial to resisting the acidic, membrane permeabilizing stresses associated
.
with scCO 2
Global protein expression of MITOTI across headspace and pressure conditions
generally indicate that cells in all conditions are in stationary phase, and appear to have
acclimated to growth under their respective environment. Principal component and clustering
analyses indicate that expression profiles from scCO 2-grown cells are not substantially different
from other conditions, and that both headspace and pressure help explain variability in MITOT 1
proteomes, with samples separating by headspace and pressure. High and low pressure samples
in each headspace were pooled for analysis to control for some potential variability in growth
phase, as proteomes from high pressure samples showed enrichment of ribosomal proteins.
Proteomes from CO 2 headspaces showed enrichment of amino acid metabolic proteins (e.g. the
glycine cleavage system), including several proteins that have previously been identified as
upregulated in responses to acid stress. These findings are notable as they suggest that MITOTI
can acclimate to a relatively 'normal' lifestyle and carry out traditional metabolic processes
under a scCO 2 headspace. Furthermore, these protein and lipid analyses provide targets for
improving the growth of strains under scCO 2 headspaces.
These results of growth and acclimation support the idea that natural microbial
communities will be able to acclimate to scCO 2 , which has been inferred in recent studies from
changes in microbial communities in deep subsurface environments after influxes of scCO 2
134
(Morozova et al., 2011; Mu et al., 2014). Microbial growth under a scCO 2 headspace raises
important concerns regarding the efficacy of using scCO 2 as a food and medical sterilizing agent,
especially whenever spores may be present. However it is encouraging for the development of
biotechnology applications involving scCO 2. Biocatalysis under scCO 2 is currently conducted
with enzymes or cell extracts, and biphasic reactors containing scCO 2 and an aqueous (or other
solvent) phase are used for extraction of various compounds, but neither have been able to utilize
living cells due to the sterilizing properties of scCO 2 . However, many new possibilities in
biofuels, pharmaceuticals, and other industries will be possible with microbes that are able to
grow in an aqueous phase that is concurrently being extracted with scCO 2 . For any
bioengineering applications to be realized in reactors or natural environments targeted for GCS,
the continued development of supercritical CO 2 tolerant microorganisms is crucial.
135
References
Abe F, Kato C, Horikoshi K. 1999. Pressure-regulated metabolism in microorganisms. Trends in
Microbiology 7:447-453.
Ahmed I, Yokota A, Fujiwara T. 2007. A novel highly boron tolerant bacterium, Bacillus
boroniphilus sp. nov., isolated from soil, that requires boron for its growth. Extremophiles
11:217-224.
Arnold RJ, Karty JA, Ellington AD, Reilly JP. 1999. Monitoring the growth of a bacteria culture
by MALDI-MS of whole cells. Anal Chem 71:1990-1996.
Aziz R, Bartels D, Best A, DeJongh M, Disz T, Edwards R, Formsma K, Gerdes S, Glass E,
Kubal M. 2008. The RAST Server: rapid annotations using subsystems technology. BMC
genomics 9:75.
Baiker A. 1999. Supercritical Fluids in Heterogeneous Catalysis. Chem Rev 99:453-474.
Ballestra P, Cuq J-L. 1998. Influence of pressurized carbon dioxide on the thermal inactivation
of bacterial and fungal spores. LWT-Food Science and Technology 31:84-88.
Barker W, Welch S, Chu S, Banfield J. 1998. Experimental observations of the effects of
bacteria on aluminosilicate weathering. American Mineralogist 83:1551-1563.
Barrangou R, Fremaux C, Deveau H, Richards M, Boyaval P, Moineau S, Romero DA, Horvath
P. 2007. CRISPR Provides Acquired Resistance Against Viruses in Prokaryotes. Science
315:1709-1712.
Beales N. 2004. Adaptation of microorganisms to cold temperatures, weak acid preservatives,
low pH, and osmotic stress: A review. Comprehensive Reviews in Food Science and Food Safety
3:1-20.
Benson S, Cook P. 2005. Intergovernmental Panel on Climate Change, Special Report on Carbon
dioxide Capture and Storage: Underground Geological Storage. Chapter.
Beranova J, Mansilla MC, de Mendoza D, Elhottova D, Konopasek I. 2010. Differences in Cold
Adaptation of Bacillus subtilis under Anaerobic and Aerobic Conditions. Journal of Bacteriology
192:4164-4171.
Bernhardt J, Weibezahn J, Scharf C, Hecker M. 2003. Bacillus subtilis during feast and famine:
visualization of the overall regulation of protein synthesis during glucose starvation by proteome
analysis. Genome Research 13:224-237.
Bertoloni G, Bertucco A, De Cian V, Parton T. 2006. A study on the inactivation of microorganisms and enzymes by high pressure C02. Biotechnol. Bioeng. 95:155-160.
136
Bligh EG, Dyer WJ. 1959. A rapid method of total lipid extraction and purification. Can J
Biochem Physiol 37:911-917.
Boone D, Liu Y, Zhao Z, Balkwill D, Drake G, Stevens T, Aldrich H. 1995. Bacillus infernus sp.
nov., an Fe (III)-and Mn (IV)-reducing anaerobe from the deep terrestrial subsurface.
International Journal of Systematic and Evolutionary Microbiology 45:441.
Boonyaratanakornkit BB, Miao LY, Clark DS. 2007. Transcriptional responses of the deep-sea
hyperthermophile Methanocaldococcus jannaschii under shifting extremes of temperature and
pressure. Extremophiles 11:495-503.
Bothun G, Knutson B, Berberich J, Strobel H, Nokes S. 2004. Metabolic selectivity and growth
of Clostridium thermocellum in continuous culture under elevated hydrostatic pressure. Applied
Microbiology and Biotechnology 65:149-157.
Brock TD. 1971. Microbial growth rates in nature. Bacteriol Rev 3 5:39-58.
Browne N, Dowds B. 2002. Acid stress in the food pathogen Bacillus cereus. Journal of Applied
Microbiology 92:404-414.
Buerger S, Spoering A, Gavrish E, Leslin C, Ling L, Epstein SS. 2012. Microbial scout
hypothesis, stochastic exit from dormancy, and the nature of slow growers. Applied and
Environmental Microbiology 78:3221-3228.
Campanaro S, Treu L, Valle G. 2008. Protein evolution in deep sea bacteria: an analysis of
amino acids substitution rates. BMC Evol Biol 8:313.
Casiano-Col6n A, Marquis RE. 1988. Role of the arginine deiminase system in protecting oral
bacteria and an enzymatic basis for acid tolerance. Applied and Environmental Microbiology
54:1318-1324.
Chapelle FH, O'Neill K, Bradley PM, Methe BA, Ciufo SA, Knobel LL, Lovley DR. 2002. A
hydrogen-based subsurface microbial community dominated by methanogens. Nature 415:312315.
Chatterjee I, Schmitt S, Batzilla CF, Engelmann S, Keller A, Ring MW, Kautenburger R,
Ziebuhr W, Hecker M, Preissner KT, Bischoff M, Proctor RA, Beck HP, Lenhof H-P,
Somerville GA, Herrmann M. 2009. Staphylococcus aureus ClpC ATPase is a late growth phase
effector of metabolism and persistence. Proteomics 9:1152-1176.
Chen YY, Hall TH, Burne RA. 1998. Streptococcus salivarius urease expression: involvement of
the phosphoenolpyruvate:sugar phosphotransferase system. FEMS Microbiology Letters
165:117-122.
137
Chitlaru T, Gat 0, Gozlan Y, Ariel N, Shafferman A. 2006. Differential proteomic analysis of
the Bacillus anthracis secretome: distinct plasmid and chromosome C02-dependent cross talk
mechanisms modulate extracellular proteolytic activities. Journal of Bacteriology 188:3551.
Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS, Megarrell
DM, Marsh T, Garrity GM, Tiedje JM. 2009. The Ribosomal Database Project: improved
alignments and new tools for rRNA analysis. Nucleic Acids Research 37:D141-D145.
Colwell F, Onstott T, Delwiche M, Chandler D, Fredrickson J, Yao Q, McKinley J, Boone D,
Griffiths R, Phelps T, Ringelberg D, White D, LaFreniere L, Balkwill D, Lehman R, Konisky J,
Long P. 1997. Microorganisms from deep, high temperature sandstones: Constraints on
microbial colonization. FEMS Microbiology Reviews 20:425-435.
Condon S, Bayarte M, Sala FJ. 1992. Influence of the Sporulation Temperature Upon the HeatResistance of Bacillus-Subtilis. J. Appl. Bacteriol. 73:251-256.
Cotter PD, Hill C. 2003. Surviving the acid test: Responses of gram-positive bacteria to low pH.
Microbiology and Molecular Biology Reviews 67:429-.
Cotter PD, Gahan CG, Hill C. 2001. A glutamate decarboxylase system protects Listeria
monocytogenes in gastric fluid. Molecular Microbiology 40:465-475.
Csonka LN. 1989. Physiological and genetic responses of bacteria to osmotic stress. Microbiol
Rev 53:121-147.
Cunningham AB, Gerlach R, Spangler L. 2009. Microbially enhanced geologic containment of
sequestered supercritical C02. Energy Procedia 1:3245-3252.
Cunningham A, Gerlach R, Spangler L, Mitchell A. 2011. Reducing the risk of well bore leakage
of C02 using engineered biomineralization barriers. Energy Procedia 4:5178-5185.
Cunningham A, Sharp R, James G. 2003. Subsurface biofilm barriers for the containment and
remediation of contaminated groundwater. Bioremediation journal 7:151-164.
Cunningham A, Lauchnor E, Eldring J, Esposito R, Mitchell A, Gerlach R, Phillips A, Ebigbo A,
Spangler L. 2013. Abandoned well C02 leakage mitigation using biologically induced
mineralization: current progress and future directions. Greenhouse Gases: Science and
Technology 3:40-49.
Curran TM, Lieou J, Marquis RE. 1995. Arginine deiminase system and acid adaptation of oral
streptococci. Applied and Environmental Microbiology 61:4494-4496.
Dance T, Spencer L, Xu J-Q. 2009. Geological characterisation of the Otway project pilot site:
What a difference a well makes. Energy Procedia 1:2871-2878.
138
Deepa CK, Dastager SG, Pandey A. 2010. Plant growth-promoting activity in newly isolated
Bacillus thioparus (NII-0902) from Western ghat forest, India. World J Microbiol Biotechnol
26:2277-2283.
Dereeper A, Guignon V, Blanc G, Audic S, Buffet S, Chevenet F, Dufayard J-F, Guindon S,
Lefort V, Lescot M, Claverie J-M, Gascuel 0. 2008. Phylogeny.fr: robust phylogenetic analysis
for the non-specialist. Nucleic Acids Research 36:W465-W469.
Derr6 I, Rapoport G, Msadek T. 1999. CtsR, a novel regulator of stress and heat shock response,
controls clp and molecular chaperone gene expression in gram-positive bacteria. Molecular
Microbiology 31:117-131.
Dillow A, Dehghani F, Hrkach J, Foster N. 1999. Bacterial inactivation by using near-and
supercritical carbon dioxide. Proceedings of the National Academy of Sciences 96:10344-10348.
Enfors S, Molin G. 1978. The influence of high concentrations of carbon dioxide on the
germination of bacterial spores. Journal of Applied Microbiology 45:279-285.
Enfors SO, Molin G. 1980. Effect of high concentrations of carbon dioxide on growth rate of
Pseudomonas fragi, Bacillus cereus and Streptococcus cremoris. Journal of Applied
Microbiology 48:409-416.
Enomoto A, Nakamura K, Hakoda M, Amaya N. 1997. Lethal effect of high-pressure carbon
dioxide on a bacterial spore. Journal of Fermentation and Bioengineering 83:305-307.
Eymann C, Dreisbach A, Albrecht D, Bernhardt J, Becher D, Gentner S, Tam LT, Btttner K,
Buurman G, Scharf C, Venz S, V61ker U, Hecker M. 2004. A comprehensive proteome map of
growing Bacillus subtilis cells. Proteomics 4:2849-2876.
Falkowski PG, Fenchel T, Delong EF. 2008. The Microbial Engines That Drive Earth's
Biogeochemical Cycles. Science 320:1034-1039.
Feng L, Wang W, Cheng J, Ren Y, Zhao G, Gao C, Tang Y, Liu X, Han W, Peng X, Liu R,
Wang L. 2007. Genome and proteome of long-chain alkane degrading Geobacillus
thermodenitrificans NG80-2 isolated from a deep-subsurface oil reservoir. Proc Natl Acad Sci
USA 104:5602-5607.
Ferreira A, Sue D, O'Byrne C, Boor K. 2003. Role of Listeria monocytogenes {sigma} B in
survival of lethal acidic conditions and in the acquired acid tolerance response. Applied and
Environmental Microbiology 69:2692.
Ferris FG, Stehmeier LG, Kantzas A, Mourits FM. 1996. Bacteriogenic mineral plugging.
Journal of Canadian Petroleum Technology 35.
139
Fischer B, Rummel G, Aldridge P, Jenal U. 2002. The FtsH protease is involved in development,
stress response and heat shock control in Caulobacter crescentus. Molecular Microbiology
44:461-478.
Foster J. 1999. When protons attack: microbial strategies of acid adaptation. Current Opinion in
Microbiology 2:170-174.
Foster J. 1995. Low pH adaptation and the acid tolerance response of Salmonella typhimurium.
Critical reviews in microbiology 21:215-237.
Fozo E, Quivey R Jr. 2004. Shifts in the membrane fatty acid profile of Streptococcus mutans
enhance survival in acidic environments. Applied and Environmental Microbiology 70:929.
Freifeld BM, Trautz RC, Kharaka YK, Phelps TJ, Myer LR, Hovorka SD, Collins DJ. 2005. The
U - tube: A novel system for acquiring borehole fluid samples from a deep geologic C02
sequestration experiment. J Geophys Res-Sol Ea 110.
Frerichs J, Rakoczy J, Ostertag-Henning C, KrUger M. 2014. Viability and Adaptation Potential
of Indigenous Microorganisms from Natural Gas Field Fluids in High Pressure Incubations with
Supercritical CO 2. Environ. Sci. Technol 48:1306-1314.
Gaidenko T, Price C. 1998. General stress transcription factor sigma(B) and sporulation
transcription factor sigma(H) each contribute to survival of Bacillus subtilis under extreme
growth conditions. Journal of Bacteriology 180:3730-3733.
Galloway JN, Dentener FJ, Capone DG, Boyer EW, Howarth RW, Seitzinger SP, Asner GP,
Cleveland CC, Green PA, Holland EA, Karl DM, Michaels AF, Porter JH, Townsend AR,
Vorosmarty CJ. 2004. Nitrogen cycles: past, present, and future. Biogeochemistry 70:153-226.
Garcia-Gonzalez L, Geeraerd AH, Elst K, Van Ginneken L, Van Impe JF, Devlieghere F. 2009.
Influence of type of microorganism, food ingredients and food properties on high-pressure
carbon dioxide inactivation of microorganisms. Int. J. Food Microbiol. 129:253-263.
Gaus I. 2010. Role and impact of C02-rock interactions during C02 storage in sedimentary
rocks. International Journal of Greenhouse Gas Control 4:73-89.
Di Giulio M. 2005. A comparison of proteins from Pyrococcus furiosus and Pyrococcus abyssi:
barophily in the physicochemical properties of amino acids and in the genetic code. Gene 346:16.
Goris J, Konstantinidis KT, Klappenbach JA, Coenye T, Vandamme P, Tiedje JM. 2007. DNADNA hybridization values and their relationship to whole-genome sequence similarities.
International Journal of Systematic and Evolutionary Microbiology 57:81-91.
Graumann PL, Marahiel MA. 1999. Cold shock proteins CspB and CspC are major stationaryphase-induced proteins in Bacillus subtilis. Archives of Microbiology 171:135-138.
140
Gross M, Jaenicke R. 1994. Proteins under pressure. The influence of high hydrostatic pressure
on structure, function and assembly of proteins and protein complexes. Eur. J. Biochem.
221:617-630.
Guerzoni ME, Lanciotti R, Cocconcelli PS. 2001. Alteration in cellular fatty acid composition as
a response to salt, acid, oxidative and thermal stresses in Lactobacillus helveticus. Microbiology
147:2255-2264.
Guinebretibre M-H, Thompson FL, Sorokin A, Normand P, Dawyndt P, Ehling-Schulz M,
Svensson B, Sanchis V, Nguyen-The C, Heyndrickx M, De Vos P. 2008. Ecological
diversification in the Bacillus cereus Group. Environ Microbiol 10:851-865.
Haque M, Russell N. 2004. Strains of Bacillus cereus vary in the phenotypic adaptation of their
membrane lipid composition in response to low water activity, reduced temperature and growth
in rice starch. Microbiology 150:1397.
Hong S, Pyun Y. 1999. Inactivation Kinetics of Lactobacillus plantarum by High Pressure
Carbon Dioxide. Journal of Food Science 64:728-733.
H6rmann S, Scheyhing C, Behr J, Pavlovic M, Ehrmann M, Vogel RF. 2006. Comparative
proteome approach to characterize the high-pressure stress response ofLactobacillus
sanfranciscensis DSM 20451 T. Proteomics 6:1878-1885.
House B, Kus JV, Prayitno N, Mair R, Que L, Chingcuanco F, Gannon V, Cvitkovitch DG,
Barnett Foster D. 2009. Acid-stress-induced changes in enterohaemorrhagic Escherichia coli
0157 : H7 virulence. Microbiology 155:2907-2918.
Hovorka SD, Benson SM, Doughty C, Freifeld BM, Sakurai S, Daley TM, Kharaka YK, Holtz
MH, Trautz RC, Nance HS, Myer LR, Knauss KG. 2006. Measuring permanence of C02 storage
in saline formations: the Frio experiment. Environ. Geosci. 13:105-121.
Huang S, Ding X, Sun Y, Yang Q, Xiao X, Cao Z, Xia L. 2012. Proteomic analysis of Bacillus
thuringiensis at different growth phases by using an automated online two-dimensional liquid
chromatography-tandem mass spectrometry strategy. Applied and Environmental Microbiology
78:5270-5279.
Idriss EE, Makarewicz 0, Farouk A, Rosner K, Greiner R, Bochow H, Richter T, Borriss R.
2002. Extracellular phytase activity of Bacillus amyloliquefaciens FZB45 contributes to its plantgrowth-promoting effect. Microbiology 148:2097-2109.
Inagaki F, Kuypers MMM, Tsunogai U, Ishibashi J-I, Nakamura K-I, Treude T, Ohkubo S,
Nakaseama M, Gena K, Chiba H, Hirayama H, Nunoura T, Takai K, Jorgensen BB, Horikoshi
K, Boetius A. 2006. Microbial community in a sediment-hosted C02 lake of the southern
Okinawa Trough hydrothermal system. Proc Natl Acad Sci USA 103:14164-14169.
141
Isenschmid A, Marison IW, Stockar von U. 1995. The influence of pressure and temperature of
compressed C02 on the survival of yeast cells. Journal of Biotechnology 39:229-237.
Ishii A, Oshima T, Sato T, Nakasone K, Mori H, Kato C. 2005. Analysis of hydrostatic pressure
effects on transcription in Escherichia coli by DNA microarray procedure. Extremophiles 9:6573.
Ishikawa H, Shimoda M, Tamaya K, Yonekura A, Kawano T, Osajima Y. 1997. Inactivation of
Bacillus spores by the supercritical carbon dioxide micro-bubble method. Biosci. Biotechnol.
Biochem. 61:1022.
Jenneman G, McInerney M, Knapp R. 1985. Microbial Penetration through Nutrient-Saturated
Berea Sandstone. Applied and Environmental Microbiology.
Jia B, Lee S, Pham BP, Kwack JM, Jin H, LI J, Wang Y, Cheong G-W. 2011. Biochemical
characterization of deblocking aminopeptidases from the hyperthermophilic archaeon
Thermococcus kodakarensis KODI. Biosci. Biotechnol. Biochem. 75:1160-1166.
Johnson D, Hallberg K. 2003. The microbiology of acidic mine waters. Research in
Microbiology 154:466-473.
Johnson D. 1998. Biodiversity and ecology of acidophilic microorganisms. FEMS Microbiology
Ecology 27:307-317.
Kamihira M, Taniguchi M, Kobayashi T. 1987. Sterilization of Microorganisms with
Supercritical Carbon Dioxide. Agricultural and Biological Chemistry 51:407-412.
Kaneda T. 1991. Iso- and anteiso-fatty acids in bacteria: biosynthesis, function, and taxonomic
significance. Microbiol Rev 55:288-302.
Kaneda T. 1977. Fatty acids of the genus Bacillus: an example of branched-chain preference.
Bacteriol Rev 41:391-418.
Kanso S, Greene A, Patel B. 2002. Bacillus subterraneus sp. nov., an iron-and manganesereducing bacterium from a deep subsurface Australian thermal aquifer. International Journal of
Systematic and Evolutionary Microbiology 52:869-874.
Kasche V, Schlothauer R, Brunner G. 1988. Enzyme denaturation in supercritical C02:
Stabilizing effect of S-S bonds during the depressurization step. Biotechnology letters 10:569574.
Kato M, Hayashi R. 1999. Effects of high pressure on lipids and biomembranes for
understanding high-pressure-induced biological phenomena. Biosci. Biotechnol. Biochem.
63:1321-1328.
142
Kelley LA, Sternberg MJE. 2009. Protein structure prediction on the Web: a case study using the
Phyre server. Nat Protoc 4:363-371.
Kerscher L, Oesterhelt D. 1982. Pyruvate: ferredoxin oxidoreductase-new findings on an
ancient enzyme. Trends in Biochemical Sciences 7:371-374.
Khaleghi M, Kermanshahi RK. 2012. Effect of Environmental Stresses on S-Layer Production in
Lactobacillus acidophilus ATCC 4356. Advances in Applied Biotechnology.
Kieft TL, McCuddy SM, Onstott TC, Davidson M, Lin L-H, Mislowack B, Pratt L, Boice E,
Lollar BS, Lippmann-Pipke J, Pfiffner SM, Phelps TJ, Gihring T, Moser D, Heerden AV. 2005.
Geochemically Generated, Energy-Rich Substrates and Indigenous Microorganisms in Deep,
Ancient Groundwater. Geomicrobiology J. 22:325-335.
Kieft T, Ringelberg D, White D. 1994. Changes in ester-linked phospholipid fatty acid profiles
of subsurface bacteria during starvation and desiccation in a porous medium. Applied and
Environmental Microbiology 60:3292.
Kim H, Goepfert J. 1974. A sporulation medium for Bacillus anthracis. Journal of Applied
Microbiology 37:265-267.
Kim S, Park H, Yim D, Kim H, Choi I, Kim K. 2008. Analysis of survival rates and cellular fatty
acid profiles of Listeria monocytogenes treated with supercritical carbon dioxide under the
influence of cosolvents. Journal of Microbiological Methods 75:47-54.
Kim SR, Kim HT, Park HJ, Kim S, Choi HJ, Hwang G-S, Yi JH, Ryu DH, Kim KH. 2009. Fatty
acid profiling and proteomic analysis of Salmonella enterica serotype Typhimurium inactivated
with supercritical carbon dioxide. Int. J. Food Microbiol. 134:190-195.
Kirk MF. 2011. Variation in Energy Available to Populations of Subsurface Anaerobes in
Response to Geological Carbon Storage. Environ. Sci. Technol 45:6676-6682.
Klein W, Weber M, Marahiel M. 1999. Cold shock response of Bacillus subtilis: isoleucinedependent switch in the fatty acid branching pattern for membrane adaptation to low
temperatures. Journal of Bacteriology 181:5341.
Knutson BL, Strobel HJ, Nokes SE, Dawson KA. 1999. Effect of pressurized solvents on ethanol
production by the thermophilic bacterium Clostridium thermocellum. The Journal of
supercritical fluids 16:149-156.
Koch AL. 1997. Microbial physiology and ecology of slow growth. Microbiology and Molecular
Biology Reviews 61:305-318.
Krumholz L. 2000. Microbial communities in the deep subsurface. Hydrogeology Journal.
143
Kutchko BG, Strazisar BR, Dzombak DA, Lowry GV, Thaulow N. 2007. Degradation of Well
Cement by CO 2under Geologic Sequestration Conditions. Environ. Sci. Technol 41:4787-4792.
Lauro F, Chastain R, Blankenship L, Yayanos A, Bartlett D. 2007. The unique 16S rRNA genes
of piezophiles reflect both phylogeny and adaptation. Applied and Environmental Microbiology
73:838-845.
Lavalleur HJ, Colwell FS. 2013. Microbial characterization of basalt formation waters targeted
for geological carbon sequestration. FEMS Microbiology Ecology 85:62-73.
Len ACL, Harty DWS, Jacques NA. 2004. Stress-responsive proteins are upregulated in
Streptococcus mutans during acid tolerance. Microbiology 150:1339-1351.
Lepage C, Fayolle F, Hermann M, Vandecasteele JP. 1987. Changes in Membrane Lipid
Composition of Clostridium acetobutylicum during Acetone-Butanol Fermentation: Effects of
Solvents, Growth Temperature and pH. Microbiology 133:103-110.
Madern D, Ebel C, Zaccai G. 2000. Halophilic adaptation of enzymes. Extremophiles 4:91-98.
Martin Galiano A, Ferra'ndiz M, La Campa De A. 2001. The promoter of the operon encoding
the FOFI ATPase of Streptococcus pneumoniae is inducible by pH. Molecular Microbiology
41:1327-1338.
Matsuda T, Marukado R, Mukouyama M, Harada T, Nakamura K. 2008. Asymmetric reduction
of ketones by Geotrichum candidum: immobilization and application to reactions using
supercritical carbon dioxide. Tetrahedron: Asymmetry 19:2272-2275.
Matsuda T, Ohashi Y, Harada T, Yanagihara R, Nagasawa T, Nakamura K. 2001. Conversion of
pyrrole to pyrrole-2-carboxylate by cells of Bacillus megaterium in supercritical C02. Chem.
Commun. 2194-2195.
Mayumi D, Dolfing J, Sakata S, Maeda H, Miyagawa Y, Ikarashi M, Tamaki H, Takeuchi M,
Nakatsu CH, Kamagata Y. 2013. Carbon dioxide concentration dictates alternative methanogenic
pathways in oil reservoirs. Nat Commun 4:1998.
McDonald JH, Grasso AM, Rejto LK. 1999. Patterns of temperature adaptation in proteins from
Methanococcus and Bacillus. Molecular Biology and Evolution 16:1785-1790.
McDonald JH, Kreitman M. 1991. Adaptive protein evolution at the Adh locus in Drosophila.
Nature 351:652-654.
McGibbon L, Russell NJ. 1983. Fatty-Acid Positional Distribution in Phospholipids of a
Psychrophilic Bacterium During Changes in Growth Temperature. Current Microbiology 9:241244.
144
McMahon P, Chapelle F. 1991. Microbial production of organic acids in aquitard sediments and
its role in aquifer geochemistry. Nature 349:233-235.
Merrick MJ. 1993. In a Class of Its Own - the Rna-Polymerase Sigma-Factor Sigma(54)
(Sigma(N)). Molecular Microbiology 10:903-909.
Mesnage S, Tosi-Couture E, Fouet A. 1999. Production and cell surface anchoring of functional
fusions between the SLH motifs of the Bacillus anthracis S-layer proteins and the Bacillus
subtilis levansucrase. Molecular Microbiology 31:927-936.
Mikucki JA, Pearson A, Johnston DT, Turchyn AV, Farquhar J, Schrag DP, Anbar AD, Priscu
JC, Lee PA. 2009. A Contemporary Microbially Maintained Subglacial Ferrous "Ocean."
Science 324:397-400.
Miladi H, Bakhrouf A, Ammar E. 2013. Cellular Lipid Fatty Acid Profiles of Reference and
Food Isolates Listeria Monocytogenes as a Response to Refrigeration and Freezing Stress.
Journal of Food Biochemistry 37:136-143.
Mitchell A, Dideriksen K, Spangler L, Cunningham A, Gerlach R. 2010. Microbially Enhanced
Carbon Capture and Storage by Mineral-Trapping and Solubility-Trapping. Environmental
Science & Technology 94-100.
Mitchell A, Ferris F. 2006. The Influence of Bacillus pasteurii on the Nucleation and Growth of
Calcium Carbonate. Geomicrobiology J. 23:213-226.
Mitchell AC, Phillips AJ, Hiebert R, Gerlach R, Spangler LH, Cunningham AB. 2009. Biofilm
enhanced geologic sequestration of supercritical C02. International Journal of Greenhouse Gas
Control 3:90-99.
Mitchell AC, Phillips AJ, Hamilton MA, Gerlach R, Hollis WK, Kaszuba JP, Cunningham AB.
2008. Resilience of planktonic and biofilm cultures to supercritical C02. The Journal of
Supercritical Fluids 47:318-325.
Mori T, Sakimoto M, Kagi T, Sakai T. 1996. Isolation and characterization of a strain of Bacillus
megaterium that degrades poly (vinyl alcohol). BioScience 60:330-332.
Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M. 2007. KAAS: an automatic genome
annotation and pathway reconstruction server. Nucleic Acids Research 35:W182-5.
Morozova D, Zettlitzer M, Let D, Wtirdemann H, Group TC. 2011. Monitoring of the microbial
community composition in deep subsurface saline aquifers during C02 storage in Ketzin,
Germany. Energy Procedia 4:4362-4370.
Mu A, Boreham C, Leong HX, Haese RR, Moreau JW. 2014. Changes in the deep subsurface
microbial biosphere resulting from a field-scale C02 geosequestration experiment. Front
Microbiol 5:209.
145
Mukhopadhyay A, He Z, Alm EJ, Arkin AP, Baidoo EE, Borglin SC, Chen W, Hazen TC, He Q,
Holman H-Y, Huang K, Huang R, Joyner DC, Katz N, Keller M, Oeller P, Redding A, Sun J,
Wall J, Wei J, Yang Z, Yen H-C, Zhou J, Keasling JD. 2006. Salt Stress in Desulfovibrio
vulgaris Hildenborough: an Integrated Genomics Approach. Journal of Bacteriology 188:40684078.
Nakamura K, Chi YM, Yamada Y, Yano T. 1986. Lipase activity and stability in supercritical
carbon dioxide. Chemical Engineering Communications 45:207-212.
Ohtomo Y, Ijiri A, Ikegawa Y, Tsutsumi M, Imachi H, Uramoto G-I, Hoshino T, Morono Y,
Sakai S, Saito Y, Tanikawa W, Hirose T, Inagaki F. 2013. Biological C02 conversion to acetate
in subsurface coal-sand formation using a high-pressure reactor system. Front Microbiol 4:361.
Oldenburg C, Oldenburg C, Torn M. 2008. Biologically Enhanced Carbon Sequestration:
Research Needs and Opportunities. Lawrence Berkeley National Laboratory.
Onstott TC. 2005. Impact of C02 Injections on Deep Subsurface Microbial Ecosystems and
Potential Ramifications for the Surface Biosphere. Elsevier 1217-1249.
Oulk MK, Tano K, Bernier A-M, Arul J. 2006. Escherichia coli inactivation mechanism by
pressurized C02. Canadian Journal of Microbiology 52:1208-1217.
Overkamp W, Ercan 0, Herber M, van Maris AJA, Kleerebezem M, Kuipers OP. 2014.
Physiological and cell morphology adaptation of Bacillus subtilis at near-zero specific growth
rates: a transcriptome analysis. Environ Microbiol n/a-n/a.
Park CB, Clark DS. 2002. Rupture of the cell envelope by decompression of the deep-sea
methanogen Methanococcus jannaschii. Applied and Environmental Microbiology 68:14581463.
Passalacqua K, Varadarajan A, Byrd B, Bergman N. 2009. Comparative transcriptional profiling
of Bacillus cereus sensu lato strains during growth in C02-bicarbonate and aerobic atmospheres.
PLoS ONE 4:e4904.
Paul S, Bag SK, Das S, Harvill ET, Dutta C. 2008. Molecular signature of hypersaline
adaptation: insights from genome and proteome composition of halophilic prokaryotes. Genome
Biol.
Peet, K. C, Freedman, A. J. E., Hernandez, H. H., Britto, V., Boreham, C., Ajo-Franklin, J. B.
and Janelle R. Thompson J. R.: Microbial growth under supercritical C0 2, In Review at Applied
and Environmental Microbiology.
Petrackova D, Vecer J, Svobodova J, Herman P. 2010. Long-term adaptation of Bacillus subtilis
168 to extreme pH affects chemical and physical properties of the cellular membrane. J. Membr.
Biol. 233:73-83.
146
Phelps T, Murphy E, Pfiffner S, White D. 1994. Comparison between geochemical and
biological estimates of subsurface microbial activities. Microbial Ecology 28:335-349.
Phillips AJ, Lauchnor E, Eldring J. 2012. Potential C02 leakage reduction through biofilminduced calcium carbonate precipitation. Environmental science & technology 47:142-149.
Poralla K, Hartner T, Kannenberg E. 1984. Effect of Temperature and Ph on the Hopanoid
Content of Bacillus-Acidocaldarius. FEMS Microbiology Letters 23:253-256.
Raja CE, Omine K. 2012. Arsenic, boron and salt resistant Bacillus safensis MS1 1 isolated from
Mongolia desert soil. Afr. J. Biotechnol. 11.
Ramagopal S. 1976. Accumulation of free ribosomal proteins Si, L7, and L 12 in Escherichia
coli. Eur. J. Biochem. 69:289-297.
Rasko DA, Worsham PL, Abshire TG. 2011. Bacillus anthracis comparative genome analysis in
support of the Amerithrax investigation. In.
Rhee MS, Moritz BE, Xie G, Glavina del Rio T, Dalin E, Tice H, Bruce D, Goodwin L,
Chertkov 0, Brettin T, Han C, Detter C, Pitluck S, Land ML, Patel M, Ou M, Harbrucker R,
Ingram LO, Shanmugam KT. 2011. Complete Genome Sequence of a thermotolerant sporogenic
lactic acid bacterium, Bacillus coagulans strain 36D1. Stand. Genomic Sci. 5:331-340.
Richard H, Foster J. 2004. Escherichia coli glutamate-and arginine-dependent acid resistance
systems increase internal pH and reverse transmembrane potential. Journal of Bacteriology
186:6032.
Roth M. 1998. Helium head pressure carbon dioxide in supercritical fluid extraction and
chromatography: thermodynamic analysis of the effects of helium. Anal Chem 70:2104-2109.
Russell AD. 1990. Bacterial spores and chemical sporicidal agents. Clinical Microbiology
Reviews.
Russell NJ, Fukunaga N. 1990. A Comparison of Thermal Adaptation of Membrane-Lipids in
Psychrophilic and Thermophilic Bacteria. FEMS Microbiology Letters 75:171-182.
Saito R, Kato C, Nakayama A. 2006. Amino acid substitutions in malate dehydrogenases of
piezophilic bacteria isolated from intestinal contents of deep-sea fishes retrieved from the abyssal
zone. J Gen Appl Microbiol 52:9-19.
Sara M, Sleytr UB. 2000. S-layer proteins. Journal of Bacteriology.
Sara M, Kuen B, Mayer HF, Mandl F, Schuster KC, Sleytr UB. 1996. Dynamics in oxygeninduced changes in S-layer protein synthesis from Bacillus stearothermophilus PV72 and the S-
147
layer-deficient variant T5 in continuous culture and studies of the cell wall composition. Journal
of Bacteriology 178:2108-2117.
Satomi M, La Duc MT, Venkateswaran K. 2006. Bacillus safensis sp. nov., isolated from
spacecraft and assembly-facility surfaces. International Journal of Systematic and Evolutionary
Microbiology 56:1735-1740.
Saw JH, Mountain BW, Feng L, Omelchenko MV. 2008. Encapsulated in silica: genome,
proteome and physiology of the thermophilic bacterium Anoxybacillus flavithermus WKl.
Genome Biology.
Schuster B, Pum D, Sleytr UB. 2008. S-layer stabilized lipid membranes (Review).
Biointerphases.
Sessions AL, Zhang L, Welander PV, Doughty D, Summons RE, Newman DK. 2013.
Identification and quantification of polyfunctionalized hopanoids by high temperature gas
chromatographydE"mass spectrometry. Organic Geochemistry 56:120-130.
Setlow B, Yu J, Li YQ, Setlow P. 2013. Analysis of the germination kinetics of individual
Bacillus subtilis spores treated with hydrogen peroxide or sodium hypochlorite. Letters in
applied microbiology 57:259-265.
Shaffer G. 2010. Long-term effectiveness and consequences of carbon dioxide sequestration.
Nature Geosci 3:464-467.
Shapiro BJ, Alm EJ. 2008. Comparing Patterns of Natural Selection across Species Using
Selective Signatures. PLoS Genet 4:e23.
Sharma S, Cook P, Jenkins C, Steeper T, Lees M, Ranasinghe N. 2011. The CO2CRC Otway
Project: Leveraging experience and exploiting new opportunities at Australia's first CCS project
site. Energy Procedia 4:5447-5454.
Shieh E, Paszczynski A, Wai CM, Lang Q, Crawford RL. 2009. Sterilization of Bacillus pumilus
spores using supercritical fluid carbon dioxide containing various modifier solutions. Journal of
Microbiological Methods 76:247-252.
Shively JM, English RS, Baker SH, Cannon GC. 2001. Carbon cycling: the prokaryotic
contribution. Current Opinion in Microbiology 4:301-306.
Simonato F, Campanaro S, Lauro FM, Vezzi A, D'Angelo M, Vitulo N, Valle G, Bartlett DH.
2006. Piezophilic adaptation: a genomic point of view. Journal of Biotechnology 126:11-25.
Slepecky RA, Hemphill HE. 2006. The genus Bacillus-nonmedical. The prokaryotes.
Sleytr UB, Messner P. 1988. Crystalline surface layers in procaryotes. Journal of Bacteriology
170:2891-2897.
148
Slonczewski JL, Fujisawa M, Dopson M, Krulwich TA. 2009. Cytoplasmic pH Measurement
and Homeostasis in Bacteria and Archaea. Adv. Microb. Physiol. 55:1-79.
Spilimbergo S, Mantoan D, Quaranta A, Mea G. 2008. Real-time monitoring of cell membrane
modification during supercritical C02 pasteurization. The Journal of Supercritical Fluids 5.
Spilimbergo S, Elvassore N, Bertucco A. 2002. Microbial inactivation by high-pressure. The
Journal of Supercritical Fluids 22:55-63.
Steels EL, Learmonth RP, Watson K. 1994. Stress tolerance and membrane lipid unsaturation in
Saccharomyces cerevisiae grown aerobically or anaerobically. Microbiology 140 ( Pt 3):569576.
Stewart PS, Franklin MJ. 2008. Physiological heterogeneity in biofilms. Nat Rev Micro 6:199210.
Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, Paulovich A,
Pomeroy SL, Golub TR, Lander ES, Mesirov JP. 2005. Gene set enrichment analysis: a
knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci
USA 102:15545-15550.
Swiatlo E, Champlin F, Holman S, Wilson W, Watt J. 2002. Contribution of choline-binding
proteins to cell surface properties of Streptococcus pneumoniae. Infection and immunity 70:412.
Takai K, Nakamura K, Toki T, Tsunogai U, Miyazaki M, Miyazaki J, Hirayama H, Nakagawa S,
Nunoura T, Horikoshi K. 2008. Cell proliferation at 122 degrees C and isotopically heavy CH4
production by a hyperthermophilic methanogen under high-pressure cultivation. Proc Natl Acad
Sci USA 105:10949-10954.
Takami H, Takaki Y, Chee G, Nishi S, Shimamura S, Suzuki H, Matsui S, Uchiyama I. 2004.
Thermoadaptation trait revealed by the genome sequence of thermophilic Geobacillus
kaustophilus. Nucleic Acids Research 32:6292.
Takami H, Takaki Y, Uchiyama I. 2002. Genome sequence of Oceanobacillus iheyensis isolated
from the Iheya Ridge and its unexpected adaptive capabilities to extreme environments. Nucleic
Acids Research 30:3927.
Takami H, Nakasone K, Takaki Y, Maeno G, Sasaki R, Masui N, Fuji F, Hirama C, Nakamura
Y, Ogasawara N, Kuhara S, Horikoshi K. 2000. Complete genome sequence of the alkaliphilic
bacterium Bacillus halodurans and genomic sequence comparison with Bacillus subtilis. Nucleic
Acids Research 28:4317-4331.
Tamburini S, Anesi A, Ferrentino G, Spilimbergo S, Guella G, Jousson 0. 2014. Supercritical
C02 Induces Marked Changes in Membrane Phospholipids Composition in Escherichia coli
K12. J. Membr. Biol. 247:469-477.
149
Top EM, Springael D. 2003. The role of mobile genetic elements in bacterial adaptation to
xenobiotic organic compounds. Current Opinion in Biotechnology 14:262-269.
Vary PS, Biedendieck R, Fuerch T, Meinhardt F, Rohde M, Deckwer W-D, Jahn D. 2007.
Bacillus megaterium-from simple soil bacterium to industrial protein production host. Applied
Microbiology and Biotechnology 76:957-967.
Vezzi A, Campanaro S, D'angelo M, Simonato F, Vitulo N, Lauro FM, Cestaro A, Malacrida G,
Simionati B, Cannata N, Romualdi C, Bartlett DH, Valle G. 2005. Life at Depth: Photobacterium
profundum Genome Sequence and Expression Analysis. Science 307:1459-1461.
Watanabe T, Furukawa S, Hirata J, Koyama T, Ogihara H, Yamasaki M. 2003. Inactivation of
Geobacillus stearothermophilus spores by high-pressure carbon dioxide treatment. Applied and
Environmental Microbiology 69:7124-7129.
Watson M, Zwingmann N, Lemon N. 2004. The Ladbroke Grove-Katnook carbon dioxide
natural laboratory: A recent C02 accumulation in a lithic sandstone reservoir. Energy 29:14571466.
Watson T, Bachu S. 2009. Evaluation of the Potential for Gas and C02 Leakage along
Wellbores. SPE Drilling & Completion 24:115-126.
Welander PV, Hunter RC, Zhang L, Sessions AL, Summons RE, Newman DK. 2009. Hopanoids
play a role in membrane integrity and pH homeostasis in Rhodopseudomonas palustris TIE-1.
Journal of Bacteriology 191:6145-6156.
Welch TJ, Farewell A, Neidhardt FC, Bartlett DH. 1993. Stress response of Escherichia coli to
elevated hydrostatic pressure. Journal of Bacteriology 175:7170-7177.
Wemekamp-Kamphuis HH, Wouters JA, de Leeuw PPLA, Hain T, Chakraborty T, Abee T.
2004. Identification of sigma factor sigma B-controlled genes and their impact on acid stress,
high hydrostatic pressure, and freeze survival in Listeria monocytogenes EGD-e. Applied and
Environmental Microbiology 70:3457-3466.
Whitman W, Coleman D, Wiebe W. 1998. Prokaryotes: The unseen majority. Proceedings of the
National Academy of Sciences 95:6578-6583.
Wilkins MJ, Hoyt DW, Marshall MJ, Alderson PA, Plymale AE, Markillie LM, Tucker AE,
Walter ED, Linggi BE, Dohnalkova AC, Taylor RC. 2014. C02 exposure at pressure impacts
metabolism and stress responses in the model sulfate-reducing bacterium Desulfovibrio vulgaris
strain Hildenborough. Front Microbiol 5:507.
Williams K, Ntarlagiannis D, Slater L, Dohnalkova A, Hubbard S, Banfield J. 2005. Geophysical
imaging of stimulated microbial biomineralization. Environ. Sci. Technol 39:7592-7600.
150
Wimmer Z, Zarevncka M. 2010. A Review on the Effects of Supercritical Carbon Dioxide on
Enzyme Activity. IJMS 11:233-253.
Wright D, Oren A. 2005. Nonphotosynthetic Bacteria and the Formation of Carbonates and
Evaporites Through Time. Geomicrobiology J. 22:27-53.
Wu B, Shao H, Wang Z, Hu Y, Tang Y, Jun Y. 2010. Viability and Metal Reduction of
Shewanella oneidensis MR- 1 under C02 Stress: Implications for Ecological Effects of C02
Leakage from Geologic C02 Sequestration. Environmental Science & Technology 44:92139218.
Xiao Z, Xu P. 2007. Acetoin Metabolism in Bacteria. Critical reviews in microbiology 33:127140.
Xiong Z, Jiang Y, Qi D, Lu H, Yang F, Yang J, Chen L, Sun L, Xu X, Xue Y, Zhu Y, Jin Q.
2009. Complete genome sequence of the extremophilic Bacillus cereus strain Qi with industrial
applications. Journal of Bacteriology 191:1120-1121.
Yamamoto K, Shibahara A, Nakayama T. 1991. Determination of double-bond positions in
methylene-interrupted dienoic fatty acids by GC-MS as their dimethyl disulfide adducts.
Chemistry and physics of ... 60:39-50.
Yamamura S, Yamashita M, Fujimoto N, Kuroda M, Kashiwa M, Sei K, Fujita M, Ike M. 2007.
Bacillus selenatarsenatis sp. nov., a selenate- and arsenate-reducing bacterium isolated from the
effluent drain of a glass-manufacturing plant. International Journal of Systematic and
Evolutionary Microbiology 57:1060-1064.
Yanagawa K, Morono Y, De Beer D, Haeckel M, Sunamura M, Futagami T, Hoshino T, Terada
T, Nakamura K-I, Urabe T, Rehder G, Boetius A, Inagaki F. 2012. Metabolically active
microbial communities in marine sediment under high-C02 and low-pH extremes. The ISME
Journal 7:555-567.
Yayanos AA, Pollard EC. 1969. A study of the effects of hydrostatic pressure on
macromolecular synthesis in Escherichia coli. Biophysical journal 9:1464-1482.
Yoon J, Kang S, Lee K, Kho Y, Choi S, Kang K, Park Y. 2001. Bacillus jeotgali sp. nov.,
isolated from jeotgal, Korean traditional fermented seafood. International Journal of Systematic
and Evolutionary Microbiology 51:1087-1092.
Yoshida S, Hiradate S, Tsukamoto T, Hatakeda K, Shirata A. 2001. Antimicrobial Activity of
Culture Filtrate of Bacillus amyloliquefaciens RC-2 Isolated from Mulberry Leaves.
Phytopathology 91:181-187.
Zhang J, Davis T, Matthews M, Drews M, LaBerge M, An Y. 2006. Sterilization using highpressure carbon dioxide. The Journal of Supercritical Fluids 38:354-372.
151
Zhang J, Burrows S, Gleason C, Matthews M, Drews M, LaBerge M, An Y. 2006. Sterilizing
Bacillus pumilus spores using supercritical carbon dioxide. Journal of Microbiological Methods
66:479-485.
Zoback MD, Gorelick SM. 2012. Earthquake triggering and large-scale geologic storage of
carbon dioxide. Proceedings of the National Academy of Sciences 109:10164-10168.
152