вход по аккаунту


Nonribosomal Peptide Biosynthesis Point Mutations and Module Skipping Lead to Chemical Diversity.

код для вставкиСкачать
DOI: 10.1002/anie.200503737
Nonribosomal Peptide Biosynthesis: Point
Mutations and Module Skipping Lead to
Chemical Diversity**
Silke C. Wenzel, Peter Meiser, Tina M. Binz,
Taifo Mahmud, and Rolf Mller*
Many pharmaceutically important peptides are nonribosomally biosynthesized by multifunctional protein complexes
termed nonribosomal peptide synthetases (NRPSs).[1, 2]
NRPSs are the largest enzymes known in nature, they can
exceed 1000 kilodaltons (kDa) in size, for example, the 1700kDa cyclosporin synthetase. As NRPSs are not bound by the
ribosome*s rulebook, the structural diversity of their products
is remarkably large. This is at least in part owing to their
ability to synthesize and incorporate numerous unusual
nonproteinogenic amino acids. To further increase chemical
diversity, NRPSs can also interact with a second class of
multimodular enzyme systems, the polyketide synthases
(PKSs).[3, 4] The resultant hybrid PKS/NRPS megasynthetase
is capable of inserting short-chain carboxylic acids as monomers into the product, thereby further increasing the chemical
diversity.[5] Since the mid-1980s, an increasing number of
genes that encode NRPSs and PKSs of bacterial and fungal
origin have been cloned, and subsequent biochemical and
genetic studies have led to a (general) model of how these
multimodular systems work.[1–4] The gained molecular understanding of multimodular megasynthetases has facilitated
efforts to redesign NRPS and PKS systems to create new
products with altered and/or increased activities.[6, 7]
Herein, we focus on the biosynthesis of myxochromides,
lipopeptides that are produced by a hybrid PKS/NRPS from
several myxobacteria.[8] Recently, we described the structures
[*] Dipl.-Chem. S. C. Wenzel, Apotheker P. Meiser, Dipl.-Biol. T. M. Binz,
Prof. Dr. R. M*ller
Pharmaceutical Biotechnology, Saarland University
P.O. Box 151150, 66041 Saarbr*cken (Germany)
Fax: (+ 49) 681-302-5473
Prof. Dr. T. Mahmud
College of Pharmacy, Oregon State University
Corvallis, OR 97331-3507 (USA)
[**] The authors would like to thank B. Arbogast and L. Barofsky (Mass
Spectrometry Core of the Environmental Health Sciences Center,
Oregon State University, NIH grant P30 ES00210) for the ESI/ICR
and MALDI-TOF analysis and B. Hinkelmann and H. Sch*ler for
performing a large-scale fermentation. We are also grateful to
Prof. Dr. T. M. Zabriskie, Dr. H. B. Bode, and B. Lelarge for the
critical review of this manuscript. This work was supported by
grants from the BMB + F and the DFG. S.C.W. and R.M. would like
to acknowledge a KekulD fellowship of the Fonds der Chemischen
Industrie and a DAAD fellowship.
Supporting information, including experimental data, for this article
is available on the WWW under or from
the author.
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
and respective biosynthetic megasynthetase of myxochromides S3 from Stigmatella aurantiaca.[9] A detailed analysis of
the involved megasynthetase led to a model for myxochromide S biosynthesis that implied several striking deviations
from standard PKS and NRPS biochemistry.[9, 10] The PKS
involved in the formation of the unsaturated myxochromide
side chain was shown to be an iteratively acting enzyme with
the intrinsic capacity to produce polyketide chains of varying
length.[9, 11] Iterative PKSs, capable of conducting multiple
rounds of chain extension, were formerly thought to be
employed exclusively by fungi and as such were falsely
classified as “fungal PKS”.[4, 10, 12]
These investigations led to the assumption that the
(penta)peptide core of myxochromide S is formed through a
unique module-skipping process during assembly by a hexamodular NRPS. Interestingly, myxochromides A, produced
by Myxococcus virescens,[8] are structurally similar lipohexapeptides (Figure 1) that incorporate the same five amino acids
(plus a proline) in a different sequence pattern. Intriguingly,
the in silico analysis of the gene cluster for the biosynthesis of
myxochromide S should theoretically coincide with that of
the hexapeptide myxochromide A.[9]
To establish an evolutionary link between the biosyntheses of the two myxochromide compounds and to analyze the
evolutionary connection to the NRPS system, we set out to
identify the myxochromide A biosynthesis genes. As myxochromide A is known to be produced from several Myxococcus species,[8] we scanned the genome sequence of M. xanthus
DK1622[28] for the presence of myxochromide biosynthetic
genes. By using the sequence from the myxochromide S gene
cluster (mchS[9]) as a probe, we identified a very similar threegene operon, which was later termed mchA (see the Supporting Information). The encoded megasynthetase (MchABCA)
directs the biosynthesis of myxochromides A, as shown by
gene-inactivation experiments (see the Supporting Information). To our knowledge, this finding correlates, for the first
time, one of the approximately 18 gene clusters for secondary-metabolite biosynthesis found in the genome sequence
of M. xanthus DK1622[13] to a specific product.
After isolation of myxochromides A from extracts of M.
xanthus DK1050 (a derivative of DK1622 that is stable in the
production of a yellow pigment[14]), structure elucidation was
carried out with myxochromide A3 (for the isolation procedure and structural data see the Supporting Information).
Based on the NMR data, the original myxochromide A
structure[8] was revised such that the a-carboxyl group of the
glutamine formes an ester linkage with the threonine hydroxy
group. This linkage was suggested in our previous work[9]
based on theoretical biosynthetic assumptions. Determination
of the absolute configuration of the amino acids correlates
well with the results from Trowitzsch-Kienast and co-workers[8] and provides evidence that one of the two alanines has
the d configuration. HPLC analysis of crude extracts from M.
xanthus DK1050 indicates the presence of at least two further
myxochromides (A2 and A4). HPLC/diode-array MS/MS
analysis of an extract from a feeding experiment employing
Figure 1. The myxochromide biosynthetic assembly line from M. xanthus DK1622 and S. aurantiaca DW4/3-1 with the corresponding products
myxochromides A and myxochromides S. The amino acids incorporated into the mchA and mchS pathway are indicated below the corresponding
NRPS modules. The structural differences in the peptide cores of myxochromides A and myxochromides S, caused by point mutations in modules
2–4, are highlighted in gray. KS = ketoacyl synthase, AT = acyl transferase, DH = b-hydroxyacyl dehydratase, ER = enoyl reductase, KR = b-ketoacyl
reductase, ACP = acyl carrier protein, C = condensation domain, A = adenylation domain, PCP = peptidyl carrier protein, MT = methyl transferase,
E = epimerization domain, TE = thioesterase.
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
[U-13C4,15N1]threonine was performed and indicated that
myxochromides A2–4 have the same peptide core but differ
in the structures of their side chains (see the Supporting
Information). In the MS/MS analysis, a specifically labeled
threonine–polyketide side-chain fragment could be observed
for each of the three types of myxochromide A, which is in
agreement with the structures shown in Figure 1.
On first inspection, the previously identified myxochromide S biosynthetic proteins from S. aurantiaca and their
counterparts from the myxochromide A pathway in M.
xanthus DK1622 seem to be almost identical. Sequence
analysis indicates a three-gene operon (mchABCS or
mchABCA) encoding a bacterial type I PKS (MchAS and
MchAA) and two nonribosomal peptide synthetases (MchBCS
and MchBCA). In general, these megasynthetases are composed of numerous domains bundled together into functional
modules that carry out biosynthetic steps involving monomer
selection and activation, covalent binding as thioesters, and
chain elongation and subsequent release.[1, 3] Commonly, the
order and the number of modules within a NRPS or PKS
dictate the number and the sequence of amino acids (NRPSs)
or carboxylic acids (PKSs) in the resultant product, which is
why the enzymes have also been called “protein templates”.[1, 3] Comparison of the myxochromide S and A megasynthetases revealed an identical arrangement of modules
and domains (Figure 1). However, there are intriguing and
significant structural differences in the peptide cores of
myxochromides S and A. In the myxochromide S structure,
the sequence of two amino acids has changed from d-Ala/lLeu to l-Leu/l-Ala, and the amino acid proline is absent in
myxochromide S.
The myxochromide peptide backbones are synthesized by
two NRPSs (MchBC) that employ adenylation (A) domains
to recruit amino acids for the growing peptide chain. The
specificity, order, and placement of the A domain within the
megasynthetase provide the instructions for thiotemplated
peptide biosynthesis.[1] Initially, we speculated that the
structural differences in the peptide cores from the two
types of myxochromides were caused by genetic exchange of
the respective A domains or complete modules. Therefore,
we expected to find a high degree of genetic similarity
between modules or A domains specific for identical amino
acids and more variation between the modules or domains
that correspond, in terms of their arrangement, in the
MchBCA and MchBCS megasynthetase. However, the amino
acid and nucleotide identities between the corresponding
domains and modules encoded in each gene cluster are much
higher (approximately 70 % nucleotide identity) than those
between domains or modules that incorporate the same
amino acids (approximately 58–59 % nucleotide identity; see
the Supporting Information).
As an alternative to the speculation above, a change in the
nonribosomal code might be caused by point mutations in the
gene cluster resulting in amino acid substitutions in the
substrate-binding pocket of the A domains. To analyze this,
the theoretical binding specificities of the A domain binding
pocket were determined through the identification of the
critical residues responsible for substrate recognition.[15, 16]
Next, these residues were compared with consensus sequen-
ces available in databases (Table 1 and Reference [29]). This
shows that the observed point mutations in the A domains
from module 2 and 3 of the mchS and mchA pathways lead to
increased chemical diversity in myxochromide biosynthesis.
Table 1: Comparison of the putative specificity-conferring codes of the
A domains from modules 2–4 from the mchA and mchS biosynthetic gene
clusters with consensus sequences.[a]
Activated Residue according to GrsA Phe numbering[b]
235 236 239 278 299 301 322 330
module 2 mchA
module 2 mchS
module 3 mchA
module 3 mchS
module 4 mchA
module 4 mchS
[a] Residues identical to the consensus sequences are shown in bold.
[b] The numbering of the amino acids corresponds to the GrsA
gramicidine synthetase. The GrsA phenylalanine substrate binding
pocket has been crystallized,[27] which lead to the establishment of the
nonribosomal code.[15, 16]
Condensation (C) domains are thought to discriminate
against elongation with “false” incoming aminoacyl-S-4’phosphopantheine (Ppant) residues, and as a result, one
would assume that the respective C domains in MchBC would
disable further assembly of the intermediate.[7, 17, 18] However,
in this case the C domains involved must inherit a promiscuous substrate specificity that allows for the production of
novel compounds which might represent an evolutionary
strategy to create chemical diversity. Although striking
similarities could be observed between the C domains of
modules 2 and 3 from both gene clusters (see the Supporting
Information), further investigations are needed to clarify
their role during myxochromide assembly.
Another interesting feature of the megasynthetase is that
the l-leucine incorporated by module 2 into myxochromides S is not epimerized even though it contains an epimerization (E) domain. From sequence analysis, the respective
E domain does not seem to be inactive but rather it seems
likely that the epimerization domain is nonfunctional when
the peptidyl carrier protein (PCP) domain from module 2 is
charged with leucine instead of alanine, which is then
epimerized to the d form in myxochromides A.
The myxochromide S NRPS, which is composed of six
modules, does not match the five biosynthetic units incorporated into the peptide core, leading to the assumption that one
module is skipped during the assembly process.[9] Structural
comparison with the myxochromide A peptide core revealed
that the fourth amino acid (l-proline) is not present in
myxochromide S (Figure 1). Examination of the critical
residues of module 4 from MchCS and MchCA reveals a
high similarity to the proline-specific binding pockets
(Table 1), which is in good agreement with the myxochromide A structure and also suggests that module 4 is skipped
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
during myxochromide S assembly.[9] Module skipping has
recently been described for an engineered PKS system[19] and
is assumed to be responsible for the generation of minor byproducts of PKS assembly lines.[10, 12, 20–22] To date, no such
process has been described in the biochemistry of multimodular NRPSs.
Two different mechanisms can be considered (Figure 2).
First, the complete module might be skipped as a result of
direct transfer of the biosynthetic intermediate from
module 3 to 5 (mechanism 1). Alternatively, a skipping
process that involves PCP-to-PCP chain transfer (mechanism 2) requires that the PCP of the skipped module be
posttranslationally activated but not primed with an extender
aminoacyl moiety. In this case, the A domain in the skipped
module must be inactive. Indeed, in the core regions of the
A domain from module 4 of MchCS, several variations from
the normally conserved amino acids are conspicuous (see the
Supporting Information). Nevertheless, variations could also
be found in the PCP-domain core motif of this module. Most
notably, the highly conserved serine residue (S), which is
required for the posttranslational activation,[23] is not in the
expected position (EDFFQMGGNPS compared to the core
motif DxFFxLGGHSL; see the Supporting Information). In
contrast to the A and PCP domains from module 4, inspection
of the C domain core motifs does not indicate any specific
deviations. We therefore set out to investigate the biochemical properties of the A and PCP domains of this module.
To answer the question if the A domain from module 4 is
able to activate amino acids, the corresponding domains from
the mchS and mchA gene clusters were expressed as glutathione-S-transferase (GST) fusion proteins in Escherichia
coli. After affinity, purification, and cleavage of the GST
fusion protein, an amino acid dependent ATP/PPi exchange
assay[24] showed that both domains are capable of specifically
activating proline (Figure 3 a). It can thus be concluded that
the specificity of the A domain of module 4 from MchCA
coincides with the primary structure of myxochromides A.
The finding that the A domain from module 4 of MchCS is
active was unexpected because of mutations in highly
conserved core motifs[2] (see the Supporting Information).
For example, the exchange of the strictly conserved aspartate
(D) for a glycine (G) residue in the core motif A4 (FDxS;
shown to interact with the a-amino group of the substrate).
Furthermore, the conserved asparagine (N) was exchanged
for serine (S) as part of the core motif A5 (NxYGPTE), which
interacts with the nucleotide ribose moiety.[2] Two additional
Figure 2. a) Alternative proposed mechanisms for module skipping during myxochromide S biosynthesis; b) Regular assembly of the myxochromide A peptide core.
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 3. Biochemical investigations to elucidate the module-skipping process in myxochromide S biosynthesis. a) Relative substrate specificities
of the adenylation domain of module 4 (A4) from MchCS and MchCA determined by ATP-PPi exchange reactions with different amino acids and a
control without any amino acid. The highest activities were set at 100 %. b) Activation of PCPs by the PPTase MtaA from the myxochromide
producer S. aurantiaca DW4/3-1; SDS-PAGE of purified PCP-GST fusion proteins from module 4 (PCP4-GST) from MchCs and MchCA and
subsequent mass analysis by HPLC-ESI/ICR. The PCPs were expressed with and without the PPTase MtaA.[26] The arrows indicate mass signals
derived from GST fusions, whereas “ L ” marks signals occurring because of GST-fusion glutathione adducts. Expected masses for the PCP4-GSTfusion proteins: apo-PCP4-GST from MchCS = 34 580 Da, holo-PCP4-GST from MchCS = 34 920 Da, apo-PCP4-GST from MchCA = 34 688 Da, holoPCP4-GST from MchCA = 35 028 Da.
mutations were also found in the core motif A7
(Y(RK)TGDL to FQTGDL), which is highly conserved in
A domains and various ATPases (see the Supporting Information).
As the inactivity of the A domain is clearly not the reason
for the module-skipping process, we next investigated
whether the PCP from module 4 is posttranslationally activated. Posttranslational modification of carrier proteins
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
implies the transfer of the 4’-Ppant moiety from coenzyme A
to a highly conserved serine residue. 4’-Phosphopantetheinyl
transferases (PPTases[25]) catalyze the conversion from the
inactive apo form to the active holo form[23] of the enzyme
(Figure 3 b). The PPTase MtaA from the myxochromide S
producer S. aurantiaca DW4/3-1 was previously identified and
characterized.[26] Inactivation of mtaA in S. aurantiaca DW4/
3-1 leads to a complete loss of natural product formation
indicating that MtaA is responsible for the activation of all
secondary-metabolite biosynthetic proteins in this strain
including MchABCS. Heterologous coexpression of MtaA
with the PCP domain from the skipped module 4 from MchCS
was performed in E. coli to investigate whether this domain
can be activated. The PCP from module 4 of MchCA was
included as a positive control. As negative controls, both
domains were expressed in the absence of mtaA. After
affinity purification of these proteins (as GST fusion proteins), their masses were measured by using Fourier transform
ion cyclotron resonance mass spectrometry (FTICR-MS)
(Figure 3 b). This revealed a Ppant transfer to the PCP from
MchCA, whereas no mass shift could be observed when the
PCP from module 4 of the mchS pathway was analyzed. It can
thus be concluded that this domain is not present in the active
holo form and therefore module 4 is completely skipped
during the biosynthetic process (see mechanism 1, Figure 2).
To date, no such process has been described in the biochemistry of multimodular NRPSs, and of the skipping processes
reported for polyketide biosynthesis,[10, 12, 19–21] only one has
been analyzed biochemically.[19] For an engineered PKS
system it was shown that the skipping process involves
passage of the growing polyketide through the skipped
module by direct ACP-to-ACP transfer.[19] In contrast, the
module skipping during myxochromide S biosynthesis represents the first verified example that multimodular megasynthetases are able to skip a complete module, which requires
structural flexibility within these giant assembly lines.
Recent biochemical investigations of NRPS- and PKScatalyzed assembly processes clearly show that these multimodular enzymes do not always follow textbook biochemistry
to generate natural products. This indicates that these
biosynthetic systems are more flexible than originally
thought.[10, 12]
Herein, two highly similar myxochromide megasynthetases are compared revealing the genetic and biochemical
basis for the generation of chemical diversity. A unique
module-skipping process is caused by a “loss of function
mutation” in the core motif of the respective PCP domain.
Furthermore, it can be deduced that point mutations in the
encoding DNA lead to amino acid substitutions in the
specificity pocket of A domains resulting in different peptide
core sequences. Therefore, the general assumption of a
gatekeeper function for all C domains in NRPS has to be
questioned. These findings are of significant importance for
combinatorial approaches and should be taken into consideration in engineering NRPS and/or PKS biosynthetic pathways for generating novel “unnatural” natural products.
Keywords: biosynthesis · module skipping · myxobacteria ·
natural products · peptides
[1] S. A. Sieber, M. A. Marahiel, Chem. Rev. 2005, 105, 715.
[2] M. A. Marahiel, T. Stachelhaus, H. D. Mootz, Chem. Rev. 1997,
97, 2651.
[3] J. Staunton, K. J. Weissman, Nat. Prod. Rep. 2001, 18, 380.
[4] D. Hopwood, Chem. Rev. 1997, 97, 2465.
[5] L. Du, B. Shen, Curr. Opin. Drug Discovery Dev. 2001, 4, 215.
[6] S. Donadio, M. Sosio, Comb. Chem. High Throughput Screening
2003, 6, 489.
[7] U. Keller, F. Schauwecker, Comb. Chem. High Throughput
Screening 2003, 6, 527.
[8] W. Trowitzsch Kienast, K. Gerth, H. Reichenbach, G. HJfle,
Liebigs Ann. Chem. 1993, 1233.
[9] S. C. Wenzel, B. Kunze, G. HJfle, B. Silakowski, M. Scharfe, H.
BlJcker, R. MKller, ChemBioChem 2005, 6, 375.
[10] S. C. Wenzel, R. MKller, Curr. Opin. Chem. Biol. 2005, 9, 447.
[11] S. C. Wenzel, F. Gross, Y. Zhang, J. Fu, F. A. Stewart, R. MKller,
Chem. Biol. 2005, 12, 349.
[12] S. J. Moss, C. J. Martin, B. Wilkinson, Nat. Prod. Rep. 2004, 21,
[13] H. B. Bode, R. MKller, Angew. Chem. 2005, 117, 6988; Angew.
Chem. Int. Ed. 2005, 44, 6828.
[14] A. Martinezlaborda, M. Elias, R. Ruizvazquez, F. J. Murillo,
Mol. Gen. Genet. 1986, 205, 107.
[15] G. L. Challis, J. Ravel, C. A. Townsend, Chem. Biol. 2000, 7, 211.
[16] T. Stachelhaus, H. D. Mootz, M. A. Marahiel, Chem. Biol. 1999,
6, 493.
[17] U. Linne, M. A. Marahiel, Biochemistry 2000, 39, 10 439.
[18] P. J. Belshaw, C. T. Walsh, T. Stachelhaus, Science 1999, 284, 486.
[19] I. Thomas, C. J. Martin, C. J. Wilkinson, J. Staunton, P. F.
Leadlay, Chem. Biol. 2002, 9, 781.
[20] I. H. Hardt, H. Steinmetz, K. Gerth, F. Sasse, H. Reichenbach,
G. HJfle, J. Nat. Prod. 2001, 64, 847.
[21] B. Beck, Y. Yoon, K. Reynolds, D. Sherman, Chem. Biol. 2002, 9,
[22] C. C. Aldrich, B. J. Beck, R. A. Fecik, D. H. Sherman, J. Am.
Chem. Soc. 2005, 127, 8441.
[23] C. T. Walsh, A. M. Gehring, P. H. Weinreb, L. E. Quadri, R. S.
Flugel, Curr. Opin. Chem. Biol. 1997, 1, 309.
[24] T. Stachelhaus, M. A. Marahiel, J. Biol. Chem. 1995, 270, 6163.
[25] R. H. Lambalot, A. M. Gehring, R. S. Flugel, P. Zuber, M.
LaCelle, M. A. Marahiel, R. Reid, C. Khosla, C. T. Walsh, Chem.
Biol. 1996, 3, 923.
[26] N. Gaitatzis, A. Hans, R. MKller, S. Beyer, J. Biochem. 2001, 129,
[27] E. Conti, T. Stachelhaus, M. A. Marahiel, P. Brick, EMBO J.
1997, 16, 4174.
[28] The genome sequence is available at the Institute of Genomic
Research, Rockville, MD, USA or from the website http://
[29] For databases of consensus sequences see,, ravel/nrps//
Received: October 21, 2005
Published online: February 28, 2006
Angew. Chem. Int. Ed. 2006, 45, 2296 –2301
2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Без категории
Размер файла
368 Кб
point, skipping, modul, chemical, mutation, diversity, leads, nonribosomal, peptide, biosynthesis
Пожаловаться на содержимое документа