Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171, available online at http:www.idealibrary.com on
REVIEW
Hybrid PeptidePolyketide Natural Products: Biosynthesis and
Prospects toward Engineering Novel Molecules
Liangcheng Du, Cesar Sanchez, and Ben Shen 1
Department of Chemistry, University of California, One Shields Avenue, Davis, California 95616
Received February 7, 2000; accepted October 3, 2000; published online December 29, 2000
The order and number of the modules on an NRPS 2 protein
dictate the sequence and number of amino acids in the resultant peptide product. A typical NRPS module consists minimally of an adenylation (A) domain responsible for amino
acid activation, a thiolation (T) domain, also known as peptidyl carrier protein (PCP), for thioesterification of the
activated amino acid, and a condensation (C) domain for
transpeptidation between the aligned peptidyl and amino
acyl thioesters to elongate the growing peptide chain
(Fig. 1A)(Cane, 1997; Cane et al., 1998; Cane and Walsh,
1999; Konz and Marahiel, 1999; von Dohren et al., 1999).
Additional domains have also been identified for the
modification of the amino acyl andor peptidyl substrates
during this process, such as an epimerization (E) domain
for the conversion of an l- to d- configuration of an amino
acid (Marahiel et al., 1997), a methylation (MT) domain for
N-methylation of the amide nitrogen (Marahiel et al.,
1997), a cyclization (Cy) domain for the formation of
heterocyclic rings (Konz et al., 1997), a reduction (R)
domain for reductive release of an aldehyde product
(Ehmann et al., 1999), and oxidation domains for the conversion of a thiazoline to a thiazole (Ox) (Du et al., 2000a,b;
Julien et al., 2000; Molnar et al., 2000; Shen et al., 1999) or
for :-hydroxylation of the incorporated amino acid (Ox$)
(Silakowski et al., 1999).
Polyketides are one of the largest groups of natural
products and are derived from sequential condensations of
The structural and catalytic similarities between modular nonribosomal peptide synthetase (NRPS) and polyketide synthase
(PKS) inspired us to search for hybrid NRPSPKS systems. By
examining the biochemical and genetic data known to date for the
biosynthesis of hybrid peptidepolyketide natural products, we show
(1) that the same catalytic sites are conserved between the hybrid
NRPSPKS and normal NRPS or PKS systems, although the
ketoacyl synthase domain in NRPSPKS hybrids is unique, and (2)
that specific interpolypeptide linkers exist at both the C- and N-termini
of the NRPS and PKS proteins, which presumably play a critical
role in facilitating the transfer of the growing peptide or polyketide
intermediate between NRPS and PKS modules in hybrid NRPS
PKS systems. These findings provide new insights for intermodular
communications in hybrid NRPSPKS systems and should now be
taken into consideration in engineering hybrid peptidepolyketide
biosynthetic pathways for making novel ``unnatural'' natural products.
2001 Academic Press
INTRODUCTION
Nonribosomal peptides refer to linear, cyclic, or branched
peptides, mostly consisting of less than 20 amino acid
residues, which are often modified by acylation, glycosylation, epimerization, heterocyclization, or N-methylation of
the amide nitrogen. Many of the nonribosomal peptides are
clinically important drugs, such as cyclosporin A (1),
penicillin, and vancomycin. Nonribosomal peptides are synthesized by nonribosomal peptide synthetases (NRPSs)
that can incorporate into the peptide products both
proteinogenic and nonproteinogenic amino acidsover 300
different amino acids are known to date (Kleinkauf and von
Dohren, 1990). NRPS possesses a modular structure, and
each module is a functional building block responsible for
the incorporation and modification of one amino acid unit.
2
Abbreviations used: A, adenylation; ACP, acyl carrier protein;
AdoMet, S-adenosylmethionine; AL, acyl CoA ligase; AMT, amino transferase; AT, acyltransferase; Bmt, (4R)-4-[(E)-2-butenyl]-4-methyl-lthreonine; C, condensation; CoA, coenzyme A; Cy, cyclization; DEBS,
6-deoxyerythronolide B synthase; DH, dehydratase; E, epimerization; ER,
enoyl reductase; FAS, fatty acid synthase, KR, ketoreductase; KS, ketoacyl
synthase; MT, methyltransferase; NRPS, nonribosomal peptide synthetase; O-MT, O-methyltransferase; Ox and Ox$, oxidation; PCP, peptidyl carrier protein; PKS, polyketide synthase; PPTase, 4$-phosphopantetheinyl transferases; R, reduction; T, thiolation; TE, thioesterase.
1
To whom correspondence and reprint requests should be addressed.
Fax: (530) 752-8995. E-mail: shenchem.ucdavis.edu.
1096-717601 35.00
Copyright 2001 by Academic Press
All rights of reproduction in any form reserved.
78
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 1. Modular organization of NRPS and PKS and comparison to hybrid NRPSPKS and PKSNRPS. (A) CN bond formation for peptide
biosynthesis catalyzed by two hypothetical NRPS modules. (B) CC bond formation for polyketide biosynthesis catalyzed by two hypothetical PKS
modules. (C) Posttranslational modification of apo-ACP or apo-PCP into holo-ACP or holo-PCP by a PPTase. (D) CC bond formation for hybrid
peptidepolyketide biosynthesis catalyzed by a hypothetical NRPSPKS hybrid. (E) CN bond formation for hybrid polyketidepeptide biosynthesis
catalyzed by a hypothetical PKSNRPS hybrid. Abbreviations are defined in a footnote.
79
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
from the proteins, as catalyzed by a thioesterase (TE)
domain that is usually found at the distal C-terminus of the
NRPS or PKS enzymes (Cane, 1997; Cane et al., 1998). An
additional discrete TE has also been found to associate with
several NRPS and PKS gene clusters, which has been
implicated in liberating the mischarged NRPS or PKS when
the latter is blocked by an unspecific thioesterification at the
PCP (de Ferra et al., 1997; Marahiel et al., 1997) or ACP
domain (August et al., 1998; Butler et al., 1999), respectively.
The modular structure of NRPS and PKS has greatly
facilitated rational engineering of metabolic pathways for
both nonribosomal peptide and polyketide biosynthesis.
Numerous novel peptide and polyketide metabolites with
predicted structural alterations have been produced by
targeted domain substitution, deletion, addition, reposition,
or by introduction of other mutations into NRPSs and
PKSs, as well as by exploiting intermodular communications to facilitate the transfer of biosynthetic intermediates
between unnaturally linked PKS modules (Cane, 1997;
Cane et al., 1998; Cane and Walsh, 1999; Gokhale et al.,
1999). One particularly successful system is the 6-deoxyerythronolide B synthase (DEBS), from which most of our
current understandings of structure and mechanism of PKS
were derived. Thus, on one hand, polyketides with defined
size and functional groups can be designed by specific
engineering of the DEBS proteins (McDaniel et al., 1999).
On the other hand, genetic engineering of the DEBS
proteins in a combinatorial manner by a multiplasmid
approach can result in the production of large libraries of
polyketides with diverse structures (Xue et al., 1999).
While genetic engineering of PKS currently holds the
most promise in generating novel structures, NRPS is
emerging to offer equally promising potential for making
novel metabolites (Cane, 1997; Cane et al., 1998; Cane and
Walsh, 1999; Konz and Marahiel, 1999; von Dohren et al.,
1999). In their 1995 seminal work, Marahiel and co-workers
first demonstrated that novel peptides with altered amino
acid sequence can be made by targeted substitution of an A
domain in the surfactin (3) synthetase SrfA of Bacillus subtilis with A domains of other NRPSs of either bacterial or
fungal origin (Stachelhaus et al., 1995). More recently, these
researchers defined general rules, also known as nonribosomal codes, for the structural basis of substrate
recognition in the A domain and showed that the amino
acid specificity of A domain could be altered rationally by
site-directed mutation of the nonribosomal code of the A
domain, rather than A domain substitution, providing a
fundamentally new strategy for engineered biosynthesis of
novel peptides (Stachelhaus et al., 1999; Challis et al., 2000).
Alternatively, Walsh and co-workers demonstrated that the
editing function of an A domain can be bypassed by using
short carboxylic acids. Many of the polyketides are clinically valuable drugs, such as daunorubicin, erythromycin,
lovastatin, and rapamycin (2). Polyketide biosynthesis is
catalyzed by polyketide synthases (PKSs). Two types of
microbial PKSs are known. Type II PKSs are multienzyme
complexes that carry a single set of iteratively used activities
and consist of several monofunctional proteins for the synthesis of aromatic polyketides, such as daunorubicin and
tetracycline. Type I PKSs are multifunctional proteins that
harbor sets of noniteratively used distinct active sites, termed modules, for the catalysis of each cycle of polyketide
chain elongation in biosynthesis of reduced polyketides,
such as the macrolide and polyene antibiotics. A typical
PKS module consists minimally of an acyltransferase (AT)
domain for extender unit selection and transfer, an acyl
carrier protein (ACP) for extender unit loading, and a
ketoacyl synthase (KS) domain for decarboxylative condensation between the aligned acyl thioesters to elongate
the growing polyketide chain (Fig. 1B) (Cane, 1997; Cane et
al., 1998; Cane and Walsh, 1999; Staunton and Wilkinson,
1998; Shen, 2000). Additional domains have also been identified for the modification of the initial ;-carbonyl group,
such as a ketoreductase (KR) domain for a ;-hydroxyl
group (Cane, 1997), a methyl transferase domain (O-MT)
for O-methylation of the ;-hydroxyl or enol group
(Silakowski, 1999), a dehydratase (DH) domain for an
alkene moiety (Cane, 1997), an enoyl reductase (ER)
domain for an alkane moiety (Cane, 1997), an MT domain
for the introduction of a methyl branch into : position (Du
et al., 2000b; Kennedy et al., 1999; Molnar et al., 2000;
Pelludat et al., 1998), and an acyl CoA ligase (AL) domain
for the priming of the loading module with an unusual starter unit (Duitman et al., 1999). An amino-transferase
(AMT) domain for the conversion of an activated fatty acid
into an amino acid has also been noted, which is uniquely
located between a PKS and a NRPS module (Duitman et
al., 1999; Kaebernick et al., 2000; Tillett and Neilan, 1999;
Tillett et al., 2000).
NRPSs and PKSs apparently use a very similar strategy
for the biosynthesis of two distinct classes of natural
products. In addition to sharing a modular organization,
both systems use carrier proteinsPCP for NRPS and
ACP for PKSto tether the growing chain. Both
PCP and ACP are posttranslationally modified by a
4$-phosphopantetheine prosthetic group, and this modification is catalyzed by a family of 4$-phosphopantetheinyl
transferases (PPTases) (Fig. 1C) (Lambalot et al., 1996;
Walsh et al., 1997). During the entire elongation process,
the growing intermediates remain covalently attached to the
carrier proteins, in a thioester linkage via the sulfhydryl
group of the 4$-phosphopantetheine group. Once reaching
its full length, the peptide or polyketide product is released
80
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
bone is assembled via other mechanisms that do not require
direct functional hybridization between NRPS and PKS
proteins (Fig. 2A). While the hybrid NRPSPKS system is
the focus of this paper, we would like to discuss the systems
that do not involve direct functional hybridization between
NRPS and PKS proteins briefly first, since they serve as
excellent examples to demonstrate nature's versatility in
biosynthesizing hybrid peptidepolyketide products.
a PPTase to directly transfer the aminoacyl phosphopantetheine group from an amino acyl coenzyme A (CoA) to
the apo-PCP. Subsequent C domain-catalyzed peptide
bond formation between the aligned aminoacyl-S-PCPs
yielded several ``unnatural'' dipeptides (Belshaw et al.,
1999). We recently described a type II PCP that can be
aminoacylated by a totally unrelated A domain, providing
yet another approach to bypass the editing function of A
domains for engineered biosynthesis of novel peptides (Du
and Shen, 1999).
These striking structural and catalytic similarities
between NRPS and PKS (Figs. 1A1C) have inspired us
(Du et al., 2000b; Shen et al., 1999) and others (Cane and
Walsh, 1999; Gehring et al., 1998a,b; Quadri, 2000) to
search for hybrid NRPSPKS systems integrating both
NRPS and PKS modules (Figs. 1D and 1E). The fact that
individual domains and modules of both NRPS and PKS
are considerably tolerant toward genetic engineering supports the wisdom of combining individual NRPS and PKS
modules for combinatorial biosynthesis. It is imagined that
these hybrid NRPSPKS systems will result in the production of novel metabolites by incorporating biosynthetic
building blocks of both amino acids and short carboxylic
acids and that the genetic tools developed for engineering
NRPS and PKS should be directly applicable for engineering hybrid NRPSPKS systems. A great challenge will then
be to understand the mechanism by which an NRPS-bound
growing peptidyl intermediate is further elongated by a
PKS module (Fig. 1D) or vice versa (Fig. 1E). Since natural
products of hybrid peptidepolyketide origin are known,
the goal of this article is to examine the biosynthesis of these
compounds to shed light on intermodular communications
between NRPS and PKS modules and to illustrate the
feasibility of constructing hybrid NRPSPKS systems for
expanding the size and diversity of ``unnatural'' natural
product libraries.
1.1. Systems Not Involving Direct
Functional Hybridization between NRPS and PKS
Coronatine biosynthesis is an example in which the
amino acid and polyketide moieties are synthesized by
NRPS and PKS enzymes individually and coupled into a
hybrid polyketideamino acid metabolite by a discrete
ligase (Fig. 3). The phytotoxin coronatine (4), produced by
many pathovars of Pseudomonas syringae, contains two distinct componentsthe polyketide moiety of coronafacic
acid (14), and the amino acid moiety of coronamic acid
(15). Variation at the amino acid moiety constitutes the
other naturally occurring analogs of 4 (Bender et al., 1999).
While acetate, butyrate and pyruvate as precursors for 14,
and l-isoleucine as a precursor for 15 were established early
by feeding experiments (Parry et al., 1994, 1996), it is the
recent cloning and molecular characterization of the
biosynthesis gene cluster for 4 that confirm that 14 and 15
are biosynthesized by PKS and NRPS, respectively
(Rangaswamy et al., 1998). The gene cluster for 4 is composed of two loci, encoding PKS and NRPS enzymes,
respectively, separated by a regulatory region. Both the
NRPS and PKS loci have their own thioesterase, supporting the hypothesis that both 14 and 15 are released from the
NRPS and PKS enzymes before being coupled together to
form 4. The latter reaction is most likely catalyzed by a
ligase, and two candidates for the latter function, Cfa5 and
Cfl, were indeed identified in the cloned gene cluster. The
fact that various amino acids have been identified in analogs
of 4 is consistent with this model, indicative that the ligase
appears to have relaxed amino acid specificity.
Cyclosporin biosynthesis is an example in which a
polyketide intermediate is first converted into an amino acid
that is subsequently incorporated into the natural product
by an NRPS enzyme (Fig. 4). The cyclic peptide 1 contains
an unusual amino acid of (4R)-4-[(E)-2-butenyl]-4methyl-l-threonine (Bmt, 16) that is of polyketide origin.
Although its gene is yet to be characterized, the Bmt
PKS has been partially purified and extensively studied
(Offenzeller et al., 1996). The Bmt PKS appears to contain
all the enzymatic activities in a single protein. In vitro
incubation of acetyl CoA, malonyl CoA, NADPH, and
S-adenosylmethionine (AdoMet) in the presence of Bmt
1. BIOSYNTHESIS OF HYBRID PEPTIDE
POLYKETIDE NATURAL PRODUCTS
Hybrid peptidepolyketide metabolites refer to natural
products that are biosynthetically derived from amino acids
and short carboxylic acids, and a few examples are depicted
in Fig. 2. Based on the biosynthetic mechanisms by which
the amino acid, or peptide, and carboxylic acid, or
polyketide, moieties are incorporated into these products,
these hybrid peptidepolyketide natural products could be
divided into two classesthose whose hybrid peptidepolyketide backbone is assembled by a hybrid NRPS
PKS system that mediates the direct elongation of a NRPSbound peptidyl intermediate by a PKS module or vice versa
(Fig. 2B) and those whose hybrid peptidepolyketide back81
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 2. Examples of hybrid peptidepolyketide natural products cyclosporin A (1), rapamycin (2), surfactin (3), coronatine (4), mycosubtilin (5),
microcystin (6), bleomycins (7), epothilones (8), myxothiazol (9), pristinamycin II B (10), TA (11), yersiniabactin (12), and nostopeptolides (13). The
junctions between the peptide and polyketide moieties are shaded. (A) The biosynthesis of these compounds does not require direct functional hybridization between NRPS and PKS proteins. (B) The biosynthesis of these compounds involves hybrid NRPSPKS systems.
82
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 3. Biosynthetic pathway for coronatine (4) in P. syringae. The polyketide moiety coronafacic acid (14) and amino acid moiety coronamic acid
(15) are coupled into 4 by a ligase.
PKS resulted in the synthesis of (3R)-hydroxy-(4R)-methyl(6E)-octenoic acid as a CoA thioester (17), clearly establishing the polyketide origin of 16. Further transformations
by additional enzymes convert 17 into 16, which is subsequently incorporated into 1 by the CssA NRPS enzyme
(Fig. 4). Cloning and molecular characterization of the cssA
gene from Tolypocladium niveum has revealed that CssA
contains 11 NRPS modules, the fifth of which is predicted to
be responsible for the incorporation of 16 into 1 (Weber et
al., 1994).
The fatty acid chains of lipopeptides, such as syringomycins, 3, and fengycins, have long been believed to be
incorporated into the peptide products by direct transfer of
the acyl group to the amino group of an NRPS-activated amino acid. Subsequent condensations between the
acylated aminoacyl-PCP with the rest of NRPS-activated
amino acids result in the synthesis of lipopeptides. Several
gene clusters for lipopeptide biosynthesis have been cloned
and characterized recently (Cosmina et al., 1993; Guenzi et
al., 1998; Konz and Marahiel, 1999; Tosato et al., 1997).
Biochemical investigations of both SrfA-A (Vollenbroich et
al., 1994) and SyrE1 (Guenzi et al., 1998) showed that the
first module of the Srf and Syr NRPS complexes activates
Glu and Ser, respectively, a fact that strongly supports the
hypothesis that acylations in lipopeptide biosynthesis occur
after the activation of the amino acid. Although enzymes
catalyzing the coupling reaction between the fatty acid and
activated amino acid are yet to be identified, a C domain,
preceding the first NRPS module, has been observed in all
lipopeptide NRPSs identified so far, serving as a good
candidate for this activity.
In contrast to the aforementioned lipopeptides that contain a ;-hydroxy fatty acid chain, mycosubtilin (5) and
microcystin (6) are members of a lipopeptide family that
carry a ;-amino fatty acid modification, and their
biosyntheses represent another variation in which a fatty
acid or polyketide moietyafter its conversion into an
amino acidis incorporated into the hybrid peptide
polyketide product by an NRPS enzyme. It is important to
notice that, although an enzyme containing both FAS or
PKS and NRPS modules is present in these systems, there
is not a direct transfer of a PKS-bound polyketide intermediate to a NRPS module; instead the transfer requires an
intermediate step for the conversion of the polyketide intermediate into an amino acid. The biosynthesis gene cluster
for 5 has been recently cloned and characterized from B.
subtilis ATCC6633, revealing features unique for FAS and
NRPS (Duitman et al., 1999). The mycosubtilin synthase
FIG. 4. Biosynthetic pathway for cyclosporin (1) in T. niveum. The polyketide moiety (3R)-hydroxy-(4R)-methyl-(6E)-octenoyl CoA (17) is
converted into the corresponding ;-amino acid 16 before its incorporation into 1 by an NRPS enzyme.
83
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
such as P-3A (18), P-4 (19), and P-6m (20), from fermentation cultures led them to propose the biosynthetic pathway
for 7 as shown in Fig. 6. According to a hybrid NRPSPKS
NRPS model, we could easily envisage the biosynthesis
of the bleomycin aglycone in three stages(I) NRPSmediated formation of 18 from Ser, Asn, Asn, and His, (II)
PKS-mediated elongation of 18 with malonyl CoA and
AdoMet to 19, and (III) NRPS-mediated elongation of 19
with ;-Ala, Cys, and Cys to yield 20involving functional
hybridizations between NRPS and PKS for the transition
III and between PKS and NRPS for the transition IIIII
(Du et al., 2000b; Shen et al., 1999). While the processive
assembly of the bleomycin aglycone by a hybrid NRPS
PKSNRPS system is evident from the isolation of the
linear peptide (18), peptidepolyketide (19), and peptidepolyketidepeptide (20) intermediates, such intermediates are rarely accumulated among most other NRPS
and PKS systems studied so far (Yu et al., 1999).
Pristinamycins are members of the streptogramin A
family of metabolites, whose biosynthesis from amino acids
and short carboxylic acids was established by extensive
feeding experiments (Kingston et al., 1983). According to a
hybrid NRPSPKS model, the pristinamycin II B backbone
(10) could be assembled by a hybrid NRPSPKS system
that mediates the transfer of the growing peptide or
polyketide intermediate between NRPS and PKS modules
three times (Fig. 7A). Bamas-Jacques and co-workers
(1997) have cloned and partially characterized the gene
cluster for the biosynthesis of 10 in Streptomyces
pristinaespiralis and revealed that snaD encodes the NRPS
enzyme catalyzing the activation and incorporation of the
Ser residue into 10. Inactivation of snaD yielded a
pristinamycin-nonproducing S. pristinaespiralis mutant that
accumulated a series of linear polyketide and polyketide
peptidepolyketide intermediates, such as 21, 22, and 23.
The structures of the latter metabolites were determined and
are shown in Fig. 7B, providing direct evidence to support
a hybrid PKSNRPSPKSNRPS system for the biosynthesis
of 10 (Bamas-Jacques et al., 1997).
While the feeding experiments and isolation of biosynthetic intermediates and shunt metabolites certainly support the hybrid NRPSPKS hypothesis, it is the recent
cloning and characterization of multiple gene clusters
encoding hybrid peptidepolyketide metabolite biosynthesis
that provide the genetic and biochemical basis for
investigating functional hybridization between NRPS and
PKS proteins. To date, biosynthetic gene clusters that have
been shown to involve direct functional hybridization
between NRPS and PKS include the Blm synthetase for 7
from S. verticillus (Fig. 8A) (Du and Shen, 1999; Du et al.,
2000b; Shen et al., 1999), the EPOS synthetase for
epothilones (8) from Sorangium cellulosum (Fig. 8B) (Julien
subunit A (MycA) combines functional domains of both
FAS and NRPS, as well as an unprecedented AMT domain
(Fig. 5A). It is proposed that myristate, activated as acyl
CoA by the AL domain, and malonyl CoA are first loaded
to the two ACP domains of MycA. Condensation between
myristyl-ACP 1 and malonyl-ACP 2 , catalyzed by the KS
domain, results in a ;-ketoacyl-ACP 2 intermediate, which is
subsequently converted into a ;-aminoacyl-ACP 2 by the
AMT domain. In contrast to normal NRPS, the extra C
domain between ACP 2 and PCP 1 seems to act as an aminoacyl transferase, catalyzing the transfer of the ;-aminoacyl group from ACP 2 to PCP 1 to yield the ;-aminoacyl-PCP 1 intermediate, setting the stage for peptide
elongation. Subsequent condensations between ;-aminoacyl-S-PCP 1 and the remaining PCP-activated amino acids
are catalyzed by the MycA, MycB, and MycC NRPS
enzymes, to finally furnish 5 (Fig. 5B). For the biosynthesis
of 6, although only the NRPS region has been published
(Meissner et al., 1996; Nishizawa et al., 1999), the DNA
sequence of the entire gene cluster has been determined from
the marine cyanobacterium Microcystis aeruginosa
PCC7806 and deposited directly in GenBank (Kaebernick
et al., 2000; Tillett and Neilan, 1999; Tillett et al., 2000).
On the basis of the functions of similar domains in NRPS
and PKS, as well as the unique AMT domain in the mycosubtilin NRPS (Duitman et al., 1999), a similar strategy for
incorporating the polyketide-derived ;-amino acid (see
Fig. 2A) by an NRPS enzyme could be envisaged for the
biosynthesis of 6.
1.2. Systems Involving Direct Functional Hybridization
between NRPS and PKS
Most of the hybrid peptidepolyketide metabolites whose
biosynthesis has been examined so far (Fig. 2B) are assembled by hybrid NRPSPKS systems that mediate direct
transfer of a NRPS-bound peptidyl intermediate to a PKS
module or vice versa (Figs. 1D and 1E). On the basis of
feeding experiments with isotope-labeled precursors and
isolation of biosynthetic intermediates and shunt metabolites, such a model in fact was implicated long before the
characterization of the modular structure of either NRPS
or PKS. For example, by feeding 14C- and 13C-labeled
biosynthetic precursors, Fujii, Umezawa, Takita, and coworkers (Fujii, 1979; Nakatani et al., 1980; Takita, 1984;
Takita and Muroka, 1990) showed that the aglycone of
bleomycin (7) was derived from a Ser, two Asn, a His, an
Ala, an acetate, a Thr, a ;-Ala, and two Cys in Streptomyces
verticillus ATCC15003, a fact that supports a hybrid peptidepolyketide biogenesis. Subsequent isolation and structural determination of a series of biosynthetic intermediates,
84
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 5. Biosynthetic pathway for mycosubtilin (5) in B. subtilis. (A) Domain organization of the MycA protein. (B) MycA-mediated biosynthesis
of the ;-amino acid moiety and its incorporation into 5 by NRPS enzymes. Abbreviations are defined in the footnote.
FIG. 6. Biosynthetic pathway for bleomycin (7) in S. verticillus involved a hybrid NRPSPKSNRPS system. The growing hybrid peptidepolyketide
biosynthetic intermediates P-3A (18), P-4 (19), and P-6m (20) were isolated from the wild-type S. verticillus fermentation and their structures were determined.
FIG. 7. Biosynthetic pathway for pristinamycin II B (10) in S. pristinaespiralis involved a hybrid PKSNRPSPKSNRPS system. The growing polyketide
intermediate 21 and hybrid peptidepolyketide intermediates 22 and 23 were isolated from a snaD-inactivated mutant and their structures were determined.
85
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 8. Domain and modular organization and classification of known hybrid NRPSPKS systems for (A) bleomycin (7) biosynthesis from
S. verticillus, (B) epothilone (8) biosynthesis from So. cellulosum, (C) and (D) myxothiazol (9) biosynthesis from St. aurantiaca, (E) rapamycin (2)
biosynthesis from S. hygroscopicus, (F) TA (11) biosynthesis from M. xanthus, (G) yersiniabactin (12) biosynthesis from Y. enterocolitica, and (H)
nostopeptolide (13) biosynthesis from Nostoc sp. GSV224. Abbreviations are defined in a footnote. References are given in text. The Ox domain in C
is defined in comparison with similar domains identified in the belomycin and epothilone gene clusters.
86
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
et al., 2000; Molnar et al., 2000; Tang et al., 2000), the Mta
synthetase for myxothiazol (9) from Stigmatella aurantiaca
DW43-1 (Figs. 8C and 8D) (Silakowski et al., 1999), the
Rap synthetase for 2 from Streptomyces hygroscopicus
(Fig. 8E) (Konig et al., 1997), the Ta1 synthetase for
antibiotic TA (11) from Myxococcus xanthus (Fig. 8F)
(Paitan et al., 1999), the HMWPs for yersiniabactin (12)
from Yersinia enterocolitica and Y. pestis (Fig. 8G)
(Pelludat et al., 1998; Gehring et al., 1998a,b), and the Nos
synthetase for nostopeptolides (13) from Nostoc sp.
GSV224 (Hoffmann et al., 1999). Analogous to the
classification system used in FAS, NRPS, and PKS, the
functional hybridization between NRPS and PKS could be
grouped into two classestype I and type II. As depicted in
Fig. 8, the interacting NRPS and PKS modules in type I
hybrids are covalently linked with all domains arranged in
a linear order on the same protein, while the interacting
NRPS and PKS modules in type II hybrids are physically
located on separate proteins. For a given hybrid peptidepolyketide biosynthetic pathway, the functional
hybridization between the NRPS and PKS modules could
be either type I, such as the one for 11 (Fig. 8F), type II,
such as the ones for 7 (Fig. 8A), 8 (Fig. 8B), 9 (Figs. 8C and
8D), 2 (Fig. 8E), and 13 (Fig. 8H), or a combination of both
types, such as the ones for 9 (Fig. 8C) and 12 (Fig. 8G).
known if the acyl group from the upstream peptidyl-S-PCP
forms a transient intermediate with the C domain before it
reacts with its cognate aminoacyl-S-PCP to form the peptide linkage. Similarly, in a hybrid PKSNRPS system, the
C domain should mediate the same CN bond formation
but by catalyzing the nucleophilic substitution between the
acyl group of the growing polyketide intermediate of acyl-SACP from the upstream PKS module and the amino group
of its cognate aminoacyl-S-PCP, resulting in the elongation
of the polyketide chain with an amino acid (Fig. 1E).
While the above discussion provides the biochemical
basis for hybrid NRPSPKS and PKSNRPS systems, it
also suggests that the critical domains for functional
hybridizations should be the PCP domain of NRPS and the
KS domain of PKS in a NRPSPKS hybrid, and the ACP
domain of PKS and the C domain of NRPS in a
PKSNRPS hybrid, respectively. Therefore, it is imagined
that detailed comparison of these domains between hybrid
and non-hybrid systems could shed light on the mechanism
for the altered catalytic activities and intermodular communications between the interacting NRPS and PKS
modules to constitute hybrid enzymes.
2.1. Hybrid PKSNRPS Systems
Among all known PKSNRPS hybrids, the ACP domain
of PKS and the C domain of NRPS show no unique features
compared to usual NRPS and PKS domains, with the
exception of the ACPs from BlmVIII and NosB PKSs
(Figs. 8A and 8H), which are more similar to PCPs than to
ACPs. This is despite the fact that C domains in PKSNRPS
hybrids must elongate an acyl-S-ACP, instead of an
aminoacyl-S-PCP, from the upstream module with their
cognate aminoacyl-S-PCP. Actually this is in accord with
the recent findings that the C domain of tyrocidin synthetase shows low selectivity toward upstream aminoacylS-PCPknown as donor site, and high selectivity toward
its cognate aminoacyl-S-PCPknown as acceptor site
(Belshaw et al., 1999). If this finding could be confirmed as
a general characteristic of the C domain, it will explain why
there is no evolutionary pressure to dramatically rearrange
C domains in the PKSNRPS hybrids. Physical proximity
of the active sites in combination with subtle changes in the
C domain primary structure may be enough for it to accept
the growing polyketide intermediate of acyl-S-ACP, instead
of aminoacyl-S-PCP, as a donor substrate.
Intermodular communications in PKS has been attributed to either the intermodular linkers, which exist between
modules within a protein, or the interpolypeptide linkers,
which exist between modules residing on two separate
proteins (Gokhale et al., 1999; Gokhale and Khosla, 2000).
2. GENETIC EVIDENCE FOR FUNCTIONAL
HYBRIDIZATION BETWEEN NRPS AND PKS
In a PKS system, the elongation step is the CC bond formation mediated by the KS domain that catalyzes (1) the
transfer of the growing polyketide intermediate of acyl-SACP from the upstream PKS module to the active site Cys
of KS, and (2) the decarboxylative condensation between
the resulting acyl-S-KS and its cognate malonyl-S-ACP
(Fig. 1B). In a hybrid NRPSPKS system, however, the KS
domain should mediate the similar CC bond formation by
catalyzing the transfer of the growing peptide intermediate
of peptidyl-S-PCP from the upstream NRPS module to the
active site Cys of KS to form an peptidyl-S-KS species,
followed by similar decarboxylative condensation with the
cognate malonyl-S-ACP, resulting in the elongation of the
peptide chain with a short carboxylic acid (Fig. 1D).
In a NRPS system, the elongation step is the CN bond
formation mediated by the C domain that catalyzes the
nucleophilic substitution between the acyl group of the
growing peptide intermediate of peptidyl-S-PCP from the
upstream NRPS module and the amino group of its cognate
aminoacyl-S-PCP (Fig. 1A). Although the active site of the
C domain has been mapped to a His residue (Stachelhaus et
al., 1998), unlike KS in polyketide biosynthesis, it is not
87
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
embedded in the primary structure of the NRPS or PKS
NRPS proteins, respectively, when the interacting modules
are physically arranged in the same protein. The local
vicinity of two active sites of ACP and C at the junction of
a type I PKSNRPS hybrid may be sufficient to shift and
integrate the chemistry of hybrid polyketidepeptide
biosynthesis. For such systems, there is probably little
pressure for the fusion protein to evolve a specific molecular
recognition or intermodular communication mechanism
between the active sites of the PKS and NRPS modules.
Interpolypeptide linkers, on the other hand, are identified
for both NRPS (Figs. 9B and 9C) and hybrid PKSNRPS
systems (type II) (Figs. 9A and 9B), suggesting that the
correct pairing of two interacting NRPS proteins or PKS
and NRPS proteins in type II PKSNRPS hybrids appears
to require specific protein-protein recognition. The N-terminal linkers for NRPS proteins in PKSNRPS hybrids
(boxed in Fig. 9B) are 2376 amino acids long, rich in basic
and acidic residues, such as Arg, Glu, and Asp, and very
hydrophilic. Intriguingly, the C-terminal linkers for the
PKS proteins in the PKSNRPS hybrids (boxed in Fig. 9A)
are 1554 amino acids long, rich in acidic residues, such as
Asp and Glu, and also generally very hydrophilic. It is,
therefore, tempting to propose that these interpolypeptide
linkers may play a critical role in protein-protein recognition, possibly by electrostatic interactions, to constitute a
functional PKSNRPS hybrid.
The regions of the interpolypeptide linkers were in fact
noticed in early sequence analysis of both the rapamycin
(Aparicio et al., 1996) and rifamycin PKSs (Tang et al.,
1998). These regions were recognized to have the potential
to form amphipathic helices, which could mediate coiledcoil interactions between the interacting PKS enzymes
(Aparicio et al., 1996). However, it is Khosla and coworkers who very recently put forth the linker hypothesis
and suggested that the chain transfer between modules is
permissive as long as these evolutionarily optimized linkers
can provide the connectivity between the adjacent modules
(Gokhale et al., 1999; Gokhale and Khosla, 2000). By
appropriate engineering of either the intermodular linker
between module 1 and 2 of DEBS and the interpolypeptide
linker between module 2 and 3 of DEBS, these researchers
demonstrated that it is possible to facilitate the transfer of
polyketide intermediates of acyl-S-ACP between heterologous PKS modules derived from both the DEBS and the
rifamycin PKSs (Gokhale et al., 1999), providing a
fundamentally new strategy of using modules as building
blocks for combinatorial biosynthesis.
Inspired by the linker hypothesis for PKS, we set out to
do parallel sequence analyses for both NRPS and hybrid
NRPSPKS systems to search for the molecular basis
of intermodular communications in these systems. We
reasoned that identification of similar linkers in the latter
systems should be viewed as an evidence supporting the
linker hypothesis and that linkers may be a general solution
to provide suitable module connectivity for other multimodular enzyme systems. We further imagined that any difference of the linkers between the NRPS or PKS and the
hybrid NRPSPKS systems may provide insight into the
evolution of modular NRPS and PKS into NRPSPKS
systems.
To search for the linkers, we use the C-terminal boundaries of the ACP and PCP domains, defined according to
the conserved active site Ser (Figs. 9A and 9C), and the
N-terminal boundaries of the KS and C domains, defined
according to the conserved active site Cys and core motif
C-1 (Konz et al., 1997), respectively (Figs. 9D and 9B). In
contrast to PKS, no apparent intermodular linker could be
identified between both NRPSNRPS modules and PKS
NRPS modules (type I hybrids). Since the intermodular
linkers for PKS (shaded in Fig. 9E) are very short (between
17 and 21 amino acid residues) and poorly conserved (with
Pro as the only conserved residue) (Gokhale et al., 1999),
our inability to identify any intermodular linker for both
NRPS and hybrid PKSNRPS systems does not necessarily
exclude its existence. On the other hand, lack of intermodular linker in the latter systems may reflect that the
mechanism for intermodular communications between the
two aligned NRPS modules or PKS and NRPS modules is
2.2. Hybrid NRPSPKS Systems
In NRPSPKS hybrids, the PCP of NRPS seems to have
no relevant difference compared to other PCPs, while the
KS domain of PKS is unique in comparison with other KSs.
The latter finding contrasts to the apparent lack of special
features shown by the C domain of PKSNRPS hybrids.
During our analysis of the biosynthesis gene cluster for 7 in
S. verticillus (Du and Shen; 1999, Du et al., 2000b; Shen et
al., 1999), we noticed that the KS domain of the BlmVIII
PKS protein (see Fig. 8A) displayed lower similarity to all
known streptomycete KSs than to a particular set of KS
domains from other bacteria. Intrigued by this observation,
we decided to set a phylogenetic analysis of KS domains
from bacterial multidomain PKSs, with special interest in
those sequences with higher similarity to KS of BlmVIII. All
the analyzed KSs for gene clusters encoding macrolide
biosynthesis in Streptomyces and Saccharopolyspora fall
into two distinct clustersKS domains from loading
modules, also known as KS Q (Bisang et al., 1999), and KS
domains from the extending modules, abbreviated as
``MACRO'' in Fig. 10. The only known Streptomyces KS
domain from a NRPSPKS hybrid system, the KS of
88
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 9. Amino acid sequences of the interpolypeptide linkers for type II PKSNRPS hybrids (A and B) and type II NRPSPKS hybrids (C and D)
and amino acid sequences of the intermodular linkers for type I NRPSPKS hybrids (E). Known intermodular linkers and interpolypeptide linkers for
PKS are shaded. Newly identified intermodular linkers and interpolypeptide linkers for hybrid NRPSPKS systems are boxed. The conserved residues,
such as S for ACP or PCP, Cys for KS, and SxxQ for C, used to establish domain boundaries, are shaded. References for all sequences are given in text.
89
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
FIG. 10. Phylogenetic relationships among KSs from bacterial multifunctional PKSs. Domains in the shaded box are involved in NRPSPKS interactions. Total number of analyzed domains was 107, boundaries of which were defined as in Aparicio et al. (1996). The E. coli KASII sequence was
included as an outgroup sequence of known crystal structure. Multiple sequence alignment and phylogenetic analysis using the bootstrapping method
(1000 resampled data sets) were performed by using the CLUSTAL X program (Thompson et al., 1997), and the resulting tree was displayed using
TREEVIEW (Page, 1996). For simplicity, those nodes with a bootstrap support <500 were collapsed in the final tree. The number of amino acid substitutions is proportional to the length of the horizontal lines. The word ``MACRO'' represents KS sequences from Streptomyces or Saccharopolyspora
macrolide PKSs (except KSQ domains Nid-Q, Pik-Q, and Tyl-Q): avermectin (Ikeda et al., 1999), erythromycin (Donadio and Katz, 1992), ``hyg'' (Ruan
et al., 1997), niddamycin (Kakavas et al., 1997), pikromycin (Xue et al., 1998), rapamycin, rifamycin (Tang et al., 1998), and tylosin (Accession
No. U78289) gene clusters. AviM, avilamycin (Streptomyces viridochromogenes; Gaisser et al., 1997); BlmVIII, bleomycin (S. verticillus); EPOSA-D,
epothilone (Sorangium cellulosum); HMWP1, yersiniabactin (Yersinia enterocolitica); MAS, mycocerosic acid (Mycobacterium tuberculosis; Mathur and
Kolattukudy, 1992); McyDEG, mycrocystins (Mycrocystis aeruginosa); MtaBDEF, myxothiazol (Stigmatella aurantiaca); MycA, mycosubtilin
(Bacillus subtilis); NosB, nostopeptolides (Nostoc sp. GSV224); PksCEF and PpsABCD, phthiocerolphenolphthiocerol (M. leprae and M. tuberculosis, respectively); PksKP, pksX locus (unknown product, B. subtilis); Ta1, antibiotic TA (Myxococcus xanthus). See text for references not included
here.
90
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
an unusual eventas far as we know, it only happened once
(as represented by the starred node in the phylogenetic tree
of Fig. 10), and all KSs interacting with a NRPS module are
derived from that single event. On the other hand, evolution
in the opposite direction (a KS belonging to this family
adapting back to PKSPKS interactions) is possible, as
exemplified by PksF and PksE. This is interesting for future
genetic engineering of hybrid peptidepolyketide biosynthetic pathways, raising the question of whether any KS
domain from type I PKS will be able to function in a
chimeric NRPSPKS hybrid, provided the proper intermodular or interpolypeptide linkers (see discussion below),
or if only those KSs from the aforementioned family can be
used to contrast new hybrids.
Finally, similar to the PKSNRPS hybrids, the intermodular linkers in type I NRPSPKS hybrids are not
apparent. In reference to the intermodular linkers of PKS
(shaded in Fig. 9E) (Gokhale et al., 1999; Gokhale and
Khosla, 2000), a putative region of variable sequence
between 22 and 62 amino acid residues is identified for type
I NRPSPKS hybrids (boxed in Fig. 9E). Although it is not
possible to draw any conclusion based on the limited data
currently available, the sequences in this region are clearly different from those of the intermodular linkers in PKS, implying that they may play a role in facilitating the transfer of a
growing peptidyl intermediate, instead of a polyketide intermediate, to the PKS module in type I NRPSPKS hybrids.
On the other hand, putative interpolypeptide linkers for
type II NRPSPKS hybrids are readily identified at the
C-termini of the NRPS proteins (Fig. 9C) and the N-termini
of the PKS proteins (Fig. 9D). Intriguingly, the latter
sequences (boxed in Fig. 9D) are much shorter (626 amino
acids long) in comparison with the corresponding interpolypeptide linkers in PKS (shaded in Fig. 9D), and are rich
in acidic residues such as Asp and Glu. In contrast, the
C-termini of NRPS in the type II NRPSPKS hybrids are
generally longer than those of normal NRPS proteins
(Fig. 9C). As boxed in Fig. 9C, the sequences in this region
are 2550 amino acids long and very rich in basic residues
such as Arg. We would like to propose that both the
C-terminal region of the NRPS proteins (boxed in Fig. 9C)
and the N-terminal region of the PKS proteins (boxed in
Fig. 9D) act together as interpolypeptide linkers in type II
NRPSPKS hybrids. The fact that the linkers at the C-termini of NRPS are rich in basic residues and those at the
N-termini of PKS are rich in acidic residues and that both
types of linkers are hydrophilic suggests once again that
they may play a critical role in protein-protein recognition
by electrostatic interactions to correctly pair the NRPS and
PKS proteins in type II NRPSPKS hybrids.
It should be emphasized that these sequence-based
speculations will have to be experimentally assessed in the
BlmVIII, does not fall into either of the aforementioned
groups, but belongs to a family that includes sequences from
different gene clusters of diverse bacterial species. Strikingly,
close examination of the latter family of sequences revealed
that it contains all the KS domains known to interact
with NRPS modules in both type I and type II NRPS
PKS hybrids, including the KSs of BlmVIII (Fig. 8A),
EPOS B (Fig. 8B), MtaD (Fig. 8C), Ta1 (Fig. 8F),
HMWP1 (Fig. 8G), and NosB (Fig. 8H), as well as the first
KS domains of PksK and PksP. Although the latter two
PKS genes belong to two gene clusters of unknown function
from B. subtilis (Albertini et al., 1995; Kunst et al., 1997),
they are respectively located on multifunctional proteins
downstream of an NRPS module that most likely activates
Gly (sequence analysis not shown), a genetic organization
strongly suggesting that they could be part of a type I
NRPSPKS hybrid. However, it should be pointed that this
family also includes a few KSs that unlikely belong to
NRPSPKS hybrids, such as the KS of McyG (Kaebernick
et al., 2000; Tillett and Neilan, 1999; Tillett et al., 2000),
PksF (GenBank Accession No. U00023), and PpsE (also
known as Pps5; Azad et al., 1997).
On the basis that the KS domains in all known NRPS
PKS hybrids contain the highly conserved catalytic residues
Cys-163, His-303, and His-340 [numbering follows the
Escherichia coli KASII sequence (Huang et al., 1998)], we
propose that the transfer of the peptidyl intermediate from
the aminoacyl-S-PCP of the upstream NRPS module to the
Cys residue of the KS domain of the PKS module and the
subsequent decarboxylative condensation with its cognate
malonyl-S-ACP are catalyzed by the KS domain in a similar
mechanism as in normal PKS. On the other hand, we have
noticed that sequence dissimilarities among KS are especially concentrated in regions which, in the dimeric KASII
enzyme, are involved in monomer-monomer interaction
and substrate binding. X-ray crystal structural analysis of
the KS enzymes, KASI, KASII, and KASIII, of the FAS
complex from E. coli has unveiled recently that the use of
dimer interface to modulate substrate specificity indeed
seems to be a conserved feature in condensing enzymes
(Huang et al., 1998; Moche et al., 1999; Olsen et al., 1999;
Qiu et al., 1999). It is, therefore, tempting to propose that
the KS in NRPSPKS hybrids, while using the same
catalytic sites conserved in all condensing enzymes, may
alter its substrate binding site to adapt the peptidyl intermediate in hybrid peptidepolyketide biosynthesis. However,
a few considerations must be taken into account as deduced
from our phylogenetic analysis, although we are certainly
aware that the conclusions are based on the analysis of
sequences from a very limited number of hybrid systems,
and they must be taken cautiously. It seems that the evolution of a KS domain to function in a hybrid NRPSPKS is
91
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
should provide a selection of platforms to engineer novel
metabolites with tailored structural features. Second, it
should be possible to construct chimeric PKSNRPS
hybrids by the use of appropriately designed linkers.
Although we are unable to identify the intermodular linkers
for type I PKSNRPS hybrids, several interpolypeptide
linkers are defined and available for such investigations.
Without knowing the specificity of individual linkers, the
natural pair of interpolypeptide linkers, such as those identified at the C-terminus of BlmVIII and N-terminus of
BlmVII, should be used in the initial experiments. On the
other hand, in strategic analogy to the PKS interpolypeptide linker engineering, the N-terminal interpolypeptide
linkers alone could also be tested and may be sufficient to
promote proteinprotein recognition and to facilitate intermediate transfer in the resultant chimeric PKSNRPS
hybrid. Third, chimeric NRPSPKS hybrids could also be
constructed by the use of appropriately engineered linkers
with or without the inclusion of the unique KS domains
identified in the natural NRPSPKS hybrids. Since both
intermodular and interpolypeptide linkers for NRPSPKS
hybrids have been identified, it is conceivable that the
chimeric NRPSPKS hybrids could be constructed in either
a type I or a type II structure, depending on the choice of the
linkers used. For the chimeric type II NRPSPKS hybrid, it
would be wise to use the natural pair of interpolypeptide
linkers in the initial experiments, such as those identified at
the C-terminus of BlmIX and N-terminus of BlmVIII.
Should the linkers along prove to be insufficient, the unique
KS domain from natural NRPSPKS hybrids should then
be included, which may provide the needed selectivity for
the elongation of a peptidyl intermediate by a PKS module
in NRPSPKS hybrids.
future. However, they are very compelling based on the
available genetic and biochemical data and should be taken
into consideration in experimental designs to investigate the
molecular basis for intermodular communications in hybrid
NRPSPKS systems and in attempts to engineer hybrid
peptidepolyketide biosynthetic pathways for making novel
unnatural natural products.
3. PROSPECTS TOWARD ENGINEERING HYBRID
PEPTIDEPOLYKETIDE METABOLITES
The field of hybrid peptidepolyketide biosynthesis has
witnessed an exponential growth, with multiple biosynthesis
gene clusters cloned and characterized in the past 2 years,
and we anticipate many more clusters to be cloned and
characterized in the next few years. Although biochemical
and mechanistic characterizations of hybrid NRPSPKS
systems are clearly lacking, sequence analysis of the cloned
hybrid peptidepolyketide biosynthesis genes and functional comparison of the deduced domains and modules
with the better-characterized NRPS and PKS systems are
starting to shed light on hybrid peptidepolyketide
biosynthesis. While nature certainly exhibits its versatility in
making hybrid peptidepolyketide metabolites, it is the
hybrid NRPSPKS systems that are most likely amenable
for combinatorial biosynthesis. A great challenge in studying the biosynthesis of hybrid peptidepolyketide natural
products, therefore, lies at revealing the basic catalytic and
molecular recognition features and structurefunction relationships of these remarkable systems, without which the
potential of combinatorial biosynthesis for the production
of novel peptidepolyketide metabolites cannot be fully
realized. However, based (1) on the similar domain and
module functions and similar modular organization of
individual NRPS and PKS modules between known hybrid
NRPSPKS systems and NRPS and PKS, (2) on the
demonstrated role of the intermodular linkers and interpolypeptide linkers played in PKS, and (3) on the proposed
functions of the putative intermodular linkers and interpolypeptide linkers identified for hybrid NRPSPKS
systems, we could envisage future endeavor in engineering
hybrid NRPSPKS systems for the production of novel
structures along the following directions. First, the methodologies developed for NRPS and PKS engineering should be
directly applicable to hybrid NRPSPKS engineering, as
long as the natural intermodular or interpolypeptide linkers
are maintained in the resultant NRPSPKS hybrids. This
strategy should allow the introductions of perturbations
into both the peptide and polyketide moieties of a hybrid
peptidepolyketide metabolite. The ever-growing inventory
of hybrid peptidepolyketide biosynthesis gene clusters
ACKNOWLEDGMENTS
Studies on peptide and polyketide biosynthesis in our laboratories have
been supported in part by an IRG grant from the American Cancer Society
and the School of Medicine, University of California, Davis; National
Science Foundation Grant MCB9733938; National Institutes of Health
Grants AI40475 and CA78747; a University of California BioSTAR grant;
and the Searle Scholars ProgramChicago Community Trust.
REFERENCES
Albertini, A. M., Caramori, T., Scoffone, F., Scotti, C., and Galizzi, A.
(1995). Sequence around the 159 degree region of the Bacillus subtilis
genome: The pksX locus spans 33.6 kb. Microbiology 141, 299309.
Aparicio, J. F., Molnar, I., Schwecke, T., Konig, A., Haydock, S. F., Khaw,
L. E., Staunton, J., and Leadley, P. F. (1996). Organization of the
biosynthetic gene cluster for rapamycin in Streptomyces hygroscopicus:
Analysis of the enzymatic domains in the modular polyketide synthase.
Gene 169, 916.
92
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
Duitman, E. H., Hamoen, L. W., Rembold, M., Venema, G., Seitz, H.,
Saenger, W., Bernhard, F., Reinhardt, R., Schmidt, M., Ullrich, C.,
Stein, T., Leenders, F., and Vater, J. (1999). The mycosubtilin synthetase of Bacillus subtilis ATCC6633: A multifunctional hybrid
between a peptide synthetase, an amino transferase, and a fatty acid
synthase. Proc. Natl. Acad. Sci. USA 96, 1329413299.
Ehmann, D. E., Gehring, A. M., and Walsh, C. T. (1999). Lysine biosynthesis in Saccharomyces cerevisiae: Mechanism of a-aminoadipate
reductase (Lys2) involves posttranslational phosphopantetheinylation
by Lys5. Biochemistry 38, 61716177.
Fujii, A. (1979). Biosynthetic aspects of bleomycinphleomycin group
antibiotics. In ``Bleomycin: Chemical, Biochemical, and Biological
Aspects'' (S. M. Hecht, Ed.), pp. 7591, Springer-Verlag, New York.
Gaisser, S., Trefzer, A., Stockert, S., Kirschning, A., and Bechthold, A.
(1997). Cloning of an avilamycin biosynthetic gene cluster from
Streptomyces viridochromogenes Tu57. J. Bacteriol. 179, 6271
6278.
Gehring, A. M., Mori, I., Perry, R. D., and Walsh, C. T. (1998a). The nonribosomal peptide synthetase HMWP2 forms a thiozoline ring during
biogenesis of yersiniabactin, an iron-chelating virulence factor of
Yersinia pestis. Biochemistry 37, 1163711650.
Gehring, A. M., DeMoll, E., Fetherston, J. D., Mori, I., Mayhew, G. F.,
Blattner, F. R., Walsh, C. T., and Perry, R. D. (1998b). Iron acquisition
in plague: Modular logic in enzymatic biogenesis of yersiniabactin by
Yersinia pestis. Chem. Biol. 5, 573586.
Gokhale, R. S., and Khosla, C. (2000). Role of linkers in communication
between protein modules. Curr. Opin. Chem. Biol. 4, 2227.
Gokhale, R. S., Tsuji, S. Y., Cane, D. E., and Khosla, C. (1999). Dissecting
and exploiting intermodular communication in polyketide synthases.
Science 284, 482485.
Guenzi, E., Galli, G., Grgurina, I., Gross, D. C., and Grandi, G. (1998).
Characterization of the syringomycin synthetase gene cluster. A link
between prokaryotic and eukaryotic peptide synthetases. J. Biol. Chem.
273, 3285732863.
Hoffmann, D., Hevel, J. M., and Moore, R. E. (1999). Characterization of
the nostopeptolide biosynthetic gene cluster of Nostoc sp. GSV224,
GenBank Accession No. AF204805.
Huang, W., Jia, J., Edwards, P., Dehesh, K., Schneider, G., and Lindqvist,
Y. (1998). Crystal structure of ;-ketoacyl-acyl carrier protein synthase
II from E. coli reveals the molecular architecture of condensing enzymes.
EMBO J. 17, 11831191.
Ikeda, H., Nonmiya, T., Usami, M., Ohta, T., and Omura, S. (1999).
Organization of the biosynthetic gene cluster for the polyketide
anthelmintic macrolide avermectin in Streptomyces avermitilis. Proc.
Natl. Acad. Sci. USA 96, 95099514.
Julien, B., Shah, S., Ziermann, R., Goldman, R., Katz, L., and Khosla, C.
(2000). Isolation and characterization of the epothilone biosynthetic
gene cluster from Sorangium cellulosum. Gene 249, 153160.
Kaebernick, M., Neilan, B. A., Borner, T., and Dittmann, E. (2000). Light
and the transcriptional response of the microcystin biosynthetic gene
cluster. Appl. Environ. Microbiol. 66, 33873392.
Kakavas, S. J., Katz, L., and Stassi, D. (1997). Identification and characterization of the niddamycin polyketide synthase genes from Streptomyces caelestis. J. Bacteriol. 179, 75157522.
Kennedy, J., Auclair, K., Kendrew, S. G., Park, C., Vederas, J. C., and
Hutchinson, C. R. (1999). Modulation of polyketide synthase activity
by accessory proteins during lovastatin biosynthesis. Science 284,
13681372.
Kingston, D. G. I., Kolpak, M. X., LeFever, J. W., and BorupGrochtmann, I. (1983). Biosynthesis of antibiotics of the virginiamycin
family. 3. Biosynthesis of virginiamycin M 1 . J. Am. Chem. Soc. 105,
51065110.
August, P. R., Tang, L., Yoon, Y. J., Ning, S., Muller, R., Yu, T.-W.,
Taylor, M., Hoffmann, D., Kim, C.-G., Zhang, X., Hutchinson, C. R.,
and Floss, H. G. (1998). Biosynthesis of the ansamycin antibiotic
rifamycin: Deductions from the molecular analysis of the rif
biosynthetic gene cluster of Amycolatopsis mediterranei S699. Chem.
Biol. 5, 6979.
Azad, A. K., Sirakova, T. D., Fernandes, N. D., and Kolattukudy, P. E.
(1997). Gene knockout reveals a novel gene cluster for the synthesis of
a class of cell wall lipids unique to pathogenic mycobacteria. J. Biol.
Chem. 272, 1674116745.
Bamas-Jacques, N., Lorenzon, S., Massey, F., de Swetschin, C., Decourty,
I., Sezonov, G., Pernodet, J. L., Friedmann, A., Vuilhorgne, M.,
Couder, M., Thibaut, D., and Desnotes, J. F. (1997). Identification of
peptide synthetase gene from Streptomyces pristinaespiralis involved in
the biosynthesis of pristinamycin II. In ``Xth International Symposium
on Biology of Actinomycetes, Abstract 2P11,'' Beijing.
Belshaw, P. J., Walsh, C. T., and Stachelhaus, T. (1999). Aminoacyl-CoAs
as probes of condensation domain selectivity in nonribosomal peptide
synthesis. Science 284, 486489.
Bender, C. L., Alarcon-Chaidez, F., and Gross, D. C. (1999). Pseudomonas
syringae phytotoxins: Mode of action, regulation, and biosynthesis by
peptide and polyketide synthetases. Microbiol. Mol. Biol. Rev. 63,
266292.
Bisang, C., Long, P. F., Cortes, J., Westcott, J., Crosby, J., Matharu, A.,
Cox, R. J., Simpson, T. J., Staunton, J., and Leadley, P. F. (1999). A
chain initiation factor common to both modular and aromatic
polyketide synthases. Nature 401, 502505.
Butler, A. R., Bate, N., and Cundliffe, E. (1999). Impact of thioesterase
activity on tylosin biosynthesis in Streptomyces fradiae. Chem. Biol. 6,
287292.
Cane, D. E. (1997). A special thematic issue on polyketide and nonribosomal polypeptide biosynthesis. Chem. Rev. 97, 24632706.
Cane, D. E., Walsh, C. T., and Khosla, C. (1998). Harnessing the
biosynthetic code: Combinations, permutations, and mutations. Science
282, 6368.
Cane, D. E., and Walsh, C. T. (1999). The parallel and convergent universes of polyketide synthases and nonribosomal peptide synthetases.
Chem. Biol. 6, R319R325.
Challis, G. L., Ravel, J., and Townsend, C. A. (2000). Predictive, structurebased model of amino acid recognition by nonribosomal peptide synthetase adenylation domains. Chem. Biol. 7, 211224.
Cosmina, P., Rodriguez, F., de Ferra, F., Grandi, G., Perego, M., Venema,
G., and van Sinderen, D. (1993). Sequence and analysis of the genetic
locus responsible for surfactin synthesis in Bacillus subtilis. Mol.
Microbiol. 8, 821831.
de Ferra, F., Rodriguez, F., Tortora, O., Tosi, C., and Grandi, G. (1997).
Engineering of peptide synthetases. J. Biol. Chem. 272, 2530425309.
Donadio, S., and Katz, L. (1992). Organization of the enzymatic domains
in the multifunctional polyketide synthase involved in erythromycin
formation in Saccharopolyspora erythraea. Gene 111, 5160.
Du, L., Chen, M., Sanchez, C., and Shen, B. (2000a). An oxidation domain
in the BlmIII non-robosomal peptide synthetase probably catalyzing
thiazole formation in the biosynthesis of the anti-tumor drug bleomycin
in Streptomyces verticillus. FEMS Microbiol. Lett. 189, 171175.
Du, L., Sanchez, C., Chen, M., Edwards, D. J., and Shen, B. (2000b). The
biosynthetic gene cluster for the antitumor drug bleomycin from Streptomyces verticillus ATCC15003 supporting functional interactions
between nonribosomal peptide synthetases and a polyketide synthase.
Chem. Biol. 7, 623642.
Du, L., and Shen, B. (1999). Identification and characterization of a type
II peptidyl carrier protein from the bleomycin producer Streptomyces
verticillus ATCC15003. Chem. Biol. 6, 507517.
93
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
Kleinkauf, H., and von Dohren, H. (1990). Nonribosomal biosynthesis of
peptide antibiotics. Eur. J. Biochem. 192, 115.
Konig, A., Schwecke, T., Molnar, I., Bohm, G. A., Lowden, P. A.,
Staunton, J., and Leadlay, P. F. (1997). The pipecolate-incorporating
enzyme for the biosynthesis of the immunosuppressant rapamycinNucleotide sequence analysis, disruption and heterologous expression of rapP from Streptomyces hygroscopicus. Eur. J. Biochem. 247,
526534.
Konz, D., Doekel, S., and Marahiel, M. A. (1999). Molecular and
biochemical characterization of the protein template controlling
biosynthesis of the lipopeptide lichenysin. J. Bacteriol. 181, 133140.
Konz, D., Klens, A., Schorgendorfer, K., and Marahiel, M. A. (1997). The
bacitracin biosynthesis operon of Bacillus licheniformis ATCC 10716:
Molecular characterization of three multi-modular peptide synthetases.
Chem. Biol. 4, 927937.
Konz, D., and Marahiel, M. A. (1999). How do peptide synthetases
generate structural diversity? Chem. Biol. 6, R39R48.
Kunst, F., Ogasawara, N., Moszer, I., Albertini, A. M., Alloni, G.,
Azevedo, V., Bertero, M. G., Bessieres, P., Bolotin, A., Borchert, S.,
Borriss, R., Boursier, L., Brans, A., Braun, M., Brignell, S. C., Bron, S.,
Brouillet, S., Bruschi, C. V., Caldwell, B., Capuano, V., Carter, N. M.,
Choi, S. K., Codani, J. J., Connerton, I. F., Danchin, A., et al. (1997).
The complete genome sequence of the gram-positive bacterium Bacillus
subtilis. Nature 390, 249256.
Lambalot, R. H., Gehring, A. M., Flugel, R. S., Zuber, P., LaCelle, M.,
Marahiel, M. A., Reid, R., Khosla, C., and Walsh, C. T. (1996). A new
enzyme superfamilyThe phosphopantetheinyl transferases. Chem.
Biol. 3, 923936.
Marahiel, M. A., Stachelhaus, T., and Mootz, H. D. (1997). Modular peptide synthetases involved in nonribosomal peptide synthesis. Chem. Rev.
97, 26512673.
Mathur, M., and Kolattukudy, P. E. (1992). Molecular cloning and
sequencing of the gene for mycocerosic acid synthase, a novel fatty acid
elongating multifunctional enzyme, from Mycobacterium tuberculosis,
var. bovis Bacillus Calmette-Guerin. J. Biol. Chem. 267, 1938819395.
McDaniel, R., Thamchaipenet, A., Gustafsson, C., Fu, H., Betlach, M., and
Ashley, G. (1999). Multiple genetic modifications of the erythromycin
polyketide synthase to produce a library of novel ``unnatural'' natural
products. Proc. Natl. Acad. Sci. USA 96, 18461851.
Meissner, K., Dittmann, E., and Boerner, T. (1996). Toxic and non-toxic
strains of the cyanobacterium Microcystis aeruginosa contain sequences
homologous to peptide synthetase genes. FEMS Microbiol. Lett. 135,
295303.
Moche, M., Schneider, G., Edwards, P., Dehesh, K., and Lindqvist, Y.
(1999). Structure of the complex between the antibiotic cerulenin and its
target, ;-ketoacyl-acyl carrier protein synthase. J. Biol. Chem. 274,
60316034.
Molnar, I., Schupp, T., Ono, M., Zirkle, R. E., Milnamow, M., NowakThompson, B., Engel, N., Toupet, C., Stratmann, A., Cyr, D. D.,
Gorlach, J., Mayo, J. M., Hu, A., Goff, S., Schmid, J., and Ligon, J. M.
(1999). The biosynthetic gene cluster for the microtubule-stabilizing
agents epothilones A and B from Sorangium celluosum So ce90. Chem.
Biol. 7, 97109.
Nakatani, T., Fujii, A., Naganawa, H., Takita, T., and Umezawa, H.
(1980). Chemistry of bleomycin. XXVI biosynthetic study using
13
C-enriched precursors. J. Antibiot. 33, 717721.
Nishizawa, T., Asayama, M., Fujii, K., Harada, K., and Shirai, M. (1999).
Genetic analysis of the peptide synthetase genes for a cyclic heptapeptide microcystin in Microcystis spp. J. Biochem. 126, 520529.
Offenzeller, M., Santer, G., Totschnig, K., Su, Z., Moser, H., Traber, R.,
and Schneider-Scherzer, E. (1996). Biosynthesis of the unusual amino
acid (4R)-4-[(E)-2-butenyl]-4-methyl-l-threonine of cyclosporin A:
Enzymatic analysis of the reaction sequence including identification of
the methylation precursor in a polyetide pathway. Biochemistry 35,
84018412.
Olsen, J. G., Kadziola, A., von Wettstein-Knowles, P., Siggaard-Andersen,
M., Lindquist, Y., and Larsen, S. (1999). The X-ray structure of
;-ketoacyl [acyl carrier protein] synthase I. FEBS Lett. 460,
4652.
Page, R. D. M. (1996). TREEVIEW: An application to display phylogenetic trees on personal computers. Comput. Appl. Biosci. 12, 357358.
Paitan, Y., Alon, G., Orr, E., Ron, E. Z., and Rosenberg, E. (1999). The
first gene in the biosynthesis of the polyketide antibiotic TA of
Myxococcus xanthus codes for a unique PKS module coupled to a
peptide synthetase. J. Mol. Biol. 286, 465474.
Parry, R. J., Jiralerspong, S., Mhaskar, S., Alemany, L., and Willcott, R.
(1996). Investigations of coronatine biosynthesis. Elucidation of the
mode of incorporation of pyruvate into coronafacic acid. J. Am. Chem.
Soc. 118, 703704.
Parry, R. J., Mhaskar, S. V., Lin, M-T., Walker, A. E., and Mafoti, R.
(1994). Investigations of the biosynthesis of the phytotoxin coronatine.
Can. J. Chem. 72, 8699.
Pelludat, C., Rakin, A., Jacobi, C. A., Schubert, S., and Heesemann, J.
(1998). The yersiniabactin biosynthetic gene cluster of Yersinia
enterocolitica: Organization and siderophore-dependent regulation.
J. Bacteriol. 180, 538546.
Qiu, X., Janson, C. A., Konstantinidis, A. K., Nwagwu, S., Silverman, C.,
Smith, W. W., Khandekar, S., Lonsdale, J., and Abdel-Meguid, S. S.
(1999). Crystal structure of ;-ketoacyl-acyl carrier protein synthase III.
J. Biol. Chem. 274, 3646536471.
Quadri, L. E. N. (2000). Assembly of acyl-capped siderophores by modular
peptide synthetases and polyketide synthases. Mol. Microbiol. 37, 112.
Rangaswamy, V., Jiralerspong, S., Parry, R., and Bender, C. L. (1998).
Biosynthesis of the Pseudomonas polyketide coronafacic acid requires
monofunctional and multifunctional polyketide synthase proteins. Proc.
Natl. Acad. Sci. USA 95, 1546915474.
Ruan, X., Stassi, D., Lax, S. A., and Katz, L. (1997). A second type-I PKS
gene cluster isolated from Streptomyces hygroscopicus ATCC 29253, a
rapamycin-producing strain. Gene 203, 19.
Shen, B. (2000). The biosynthesis of aromatic polyketides. Top. Curr.
Chem. 209, 151.
D. J. (1999). Bleomycin biosynthesis in Streptomyces verticillus
ATCC15003: A model of hybrid peptide and polyketide biosynthesis.
Bioorg. Chem. 27, 155171.
Silakowski, B., Schairer, H. U., Ehret, H., Kunze, B., Weinig, S., Nordsiek,
G., Brandt, P., Blocker, H., Hofle, G., Beyer, S., and Muller, R. (1999).
New lessons for combinatorial biosynthesis from myxobacteria.
The myxothiazol biosynthetic gene cluster of Stigmatella aurantiaca
DW43-1. J. Biol. Chem. 274, 3739137399.
Stachelhaus, T., Mootz, H. D., Bergendahl, V., and Marahiel, M. A. (1998).
Peptide bond formation in nonribosomal peptide biosynthesis.
Catalytic role of the condensation domain. J. Biol. Chem. 273,
2277322781.
Stachelhaus, T., Mootz, H. D., and Marahiel, M. A. (1999). The specificityconferring code of adenylation domains in nonribosomal peptide
synthetases. Chem. Biol. 6, 493505.
Stachelhaus, T., Schneider, A., and Marahiel, M. A. (1995). Rational
design of peptide antibiotics by targeted replacement of bacterial and
fungal domains. Science 269, 6972.
Staunton, J., and Wilkinson, B. (1998). The biosynthesis of alphatic
polyketides. Top. Curr. Chem. 195, 4992.
Takita, T. (1984). The bleomycins: Properties, biosynthesis, and fermentation. Drugs Pharm. Sci. 22, 595603.
94
Review
Metabolic Engineering 3, 7895 (2001)
doi:10.1006mben.2000.0171
Tang, L., Shah, S., Chung, L., Carney, J., Katz, K., Khosla, C., and Julien,
B. (2000). Cloning and heterologous expression of the epothilone gene
cluster. Science 287, 640642.
Tang, L., Yoon, Y. J., Choi, C., and Hutchinson, C. R. (1998). Characterization of the enzymatic domains in the modular polyketide synthase
involved in rifamycin B biosynthesis by Amycolatopsis mediterranei.
Gene 216, 255269.
Takita, T., and Muroka, Y. (1990). Biosynthesis and chemical synthesis of
bleomycin. In ``Biochemistry of Peptide Antibiotics: Recent Advances in
the Biotechnology of ;-Lactams and Microbial Peptides'' (H. Kleinkauf
and H. von Dohren, Eds.), pp. 289309, de Gruyter, New York.
Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F., and Higgins,
D. G. (1997). The CLUSTAL X windows interface: Flexible strategies
for multiple sequence alignment aided by quality analysis tools. Nucleic
Acids Res. 25, 48764882.
Tillett, D., and Neilan, B. A. (1999). Structure of the microcystin synthetase
gene cluster, GenBank Accession No. AF183408.
Tillett, D., Dittmann, M., Erhard, M., von Doren, H., Burner, T., and
Neilan, A. (2000). Structural organization of microcystin biosynthesis in
M. aeruginosa PCC7806: An integrated peptidepolyketide synthetase
system. Chem. Biol. 7, 753764.
Tosato, V., Albertini, A. M., Zotti, M., Sonda, S., and Bruschi, C. V. (1997).
Sequence completion, identification and definition of the fengycin
operon in Bacillus subtilis 168. Microbiology 143, 34433450.
Vollenbroich, D., Mehta, N., Zuber, P., Vater, J., and Kamp, R. M. (1994).
Analysis of surfactin synthetase subunits in srfA mutants of Bacillus
subtilis OKB105. J. Bacteriol. 176, 395400.
von Duhren, H., Dickmann, R., and Pavela-Vrancic, M. (1999). The nonribosomal code. Chem. Biol. 6, R273R279.
Walsh, C. T., Gehring, A. M., Weinreb, P. H., Quadri, L. E. N., and
Flugel, R. S. (1997). Post-translational modification of polyketide
and nonribosomal peptide synthases. Curr. Opin. Chem. Biol. 1,
309315.
Weber, G., Schorgendorfer, K., Schneider-Scherzer, E., and Leitner, E.
(1994). The peptide synthetase catalyzing cyclosporin production in
Tolypocladium niveum is encoded by a giant 45.8-kilobase open reading
frame. Curr. Genet. 26, 120125.
Xue, Q., Ashley, G., Hutchinson, C. R., and Santi, D. V. (1999). A multiplasmid approach to preparing large libraries of polyketides. Proc. Natl.
Acad. Sci. USA 96, 1174011745.
Xue, Y., Zhao, L., Liu, H-W., and Sherman, D. (1998). A gene cluster for
macrolide antibiotic biosynthesis in Streptomyces venezuelae: Architecture of metabolic diversity. Proc. Natl. Acad. Sci. USA 95, 12111
12116.
Yu, T-W., Shen, Y., Doi-Katayama, Y., Tang, L., Park, C., Moore, B. S.,
Hutchinson, C. R., and Floss, H. G. (1999). Direct evidence that the
rifamycin polyketide synthase assembles polyketide chains processively.
Proc. Natl. Acad. Sci. USA 96, 90519056.
Printed in Belgium
95