A tRNA-dependent two-enzyme pathway for the generation of singly and doubly methylated ditryptophan 2,5-diketopiperazines

Article abstract of DOI:10.1021/bi4004827

A large number of bioactive natural products containing a 2,5-diketopiperazine (DKP) moiety have been isolated from various microbial sources. Especially tryptophan-containing cyclic dipeptides (CDPs) show great structural and functional diversity, while little is known about their biosynthetic pathways. Here, we describe the bioinformatic analysis of a cyclodipeptide synthase (CDPS)-containing gene cluster from Actinosynnema mirum spanning 2.9 kb that contains two putative DKP-modifying enzymes. We establish the biosynthetic pathway leading to two methylated ditryptophan CDPs through in vivo and in vitro analyses. Our studies identify the first CDPS (Amir-4627) that shows high substrate specificity synthesizing only one main product, cyclo(Trp-Trp) (cWW). It is the first member of the CDPS family that can form ditryptophan DKPs and the first prokaryotic CDPS whose main product constituents differ from the four amino acids (Phe, Leu, Tyr, and Met) usually found in CDPS-dependent CDPs. We show that after cWW formation a S-adenosyl-l-methionine- dependent N-methyltransferase (Amir-4628) conducts two successive methylations at the DKP-ring nitrogens and additionally show that it is able to methylate four other phenylalanine-containing CDPs. This makes Amir-4628 the first identified DKP-ring-modifying methyltransferase. The large number of known modifying enzymes of bacterial and fungal origin known to act upon Trp-containing DKPs makes the identification of a potent catalyst for cWW formation, encoded by a small gene, valuable for combinatorial in vivo as well as chemoenzymatic approaches, with the aim of generating derivatives of known CDP natural products or entirely new chemical entities with potentially improved or new biological activities.

Full text of DOI:10.1021/bi4004827

Article
pubs.acs.org/biochemistry
A tRNA-Dependent Two-Enzyme Pathway for the Generation of
Singly and Doubly Methylated Ditryptophan 2,5-Diketopiperazines
Tobias W. Giessen, Alexander M. von Tesmar, and Mohamed A. Marahiel*
Department of Chemistry and LOEWE-Center for Synthetic Microbiology, Philipps-University, Hans-Meerwein-Strasse-4, 35032
Marburg, Germany
S
* Supporting Information
ABSTRACT: A large number of bioactive natural products
containing a 2,5-diketopiperazine (DKP) moiety have been
isolated from various microbial sources. Especially tryptophan-
containing cyclic dipeptides (CDPs) show great structural and
functional diversity, while little is known about their
biosynthetic pathways. Here, we describe the bioinformatic
analysis of a cyclodipeptide synthase (CDPS)-containing gene
cluster from Actinosynnema mirum spanning 2.9 kb that
contains two putative DKP-modifying enzymes. We establish
the biosynthetic pathway leading to two methylated
ditryptophan CDPs through in vivo and in vitro analyses. Our studies identify the first CDPS (Amir_4627) that shows high
substrate specificity synthesizing only one main product, cyclo(Trp-Trp) (cWW). It is the first member of the CDPS family that
can form ditryptophan DKPs and the first prokaryotic CDPS whose main product constituents differ from the four amino acids
(Phe, Leu, Tyr, and Met) usually found in CDPS-dependent CDPs. We show that after cWW formation a S-adenosyl-L-
methionine-dependent N-methyltransferase (Amir_4628) conducts two successive methylations at the DKP-ring nitrogens and
additionally show that it is able to methylate four other phenylalanine-containing CDPs. This makes Amir_4628 the first
identified DKP-ring-modifying methyltransferase. The large number of known modifying enzymes of bacterial and fungal origin
known to act upon Trp-containing DKPs makes the identification of a potent catalyst for cWW formation, encoded by a small
gene, valuable for combinatorial in vivo as well as chemoenzymatic approaches, with the aim of generating derivatives of known
CDP natural products or entirely new chemical entities with potentially improved or new biological activities.
large number of 2,5-diketopiperazine (DKP)-containing
natural products have been isolated from microbial sources,
6) ring system (Figure 1A).6 A large majority of DKP natural
products that contain two Trp residues are annulated, while only
a small number of monocyclic ditryptophan DKPs could be
isolated so far, namely, the fellutanins from Penicillium
fellutanum,11 Sch 725418 from Micromonospora sp.,12 and a
number of okaramine congeners from Penicillium simplicissimum
(Figure 1A).13
From a biosynthetic standpoint, it has long been known that
nonribosomal peptide synthetases (NRPSs) are able to generate
DKP-containing compounds, either through a dedicated path-
way or through the premature release of peptidyl intermediates
via cyclization.14−16 With the discovery of the albonoursin
biosynthetic gene cluster in Streptomyces noursei and the
characterization of the first cyclodipeptide synthase (CDPS)
AlbC, a second unrelated pathway exclusively dedicated to the
formation of CDPs could be identified.17,18 CDPSs are able to
form two successive peptide bonds in an ATP-independent
fashion by hijacking aminoacyl-tRNAs (aa-tRNAs) from the
ribosomal machinery and directly using them as substrates for the
generation of a DKP-ring system (Figure 1B).19 Putative CDPS-
A
and many have received an increased level of attention in recent
years because of their diverse and interesting biological activities,
ranging from antibacterial1,2 and immunosuppressive3 to
antineoplastic.4,5 In general, the DKP-ring system is formed
through the condensation of two amino acids resulting in the
smallest possible cyclic peptide and has long been recognized as a
privileged structural scaffold in the search for new synthetic or
semisynthetic pharmacologically useful compounds.6 With
regard to the physiological function of CDPs in the respective
producing organisms, only a limited amount of information is
available that indicates a function in biochemical communication
phenomena like intra- and intercellular signaling7,8 and quorum
sensing.9,10
Tryptophan (Trp)-containing CDPs are among the most
prevalent DKP-containing natural products and have been
isolated from a diverse range of microbial sources, among them
many different Aspergillus and Penicillium species.6 In contrast,
comparatively few Trp-containing CDPs of bacterial origin are
known.6 Tryptophan residues that are part of a DKP can exist in a
monocyclic form like most other amino acids or in an annulated
form in which an additional five-membered ring is generated
through formation of a bond between the α-nitrogen and the
indole C2 of the Trp residue resulting in a tetracyclic (6−5−5−
Received: April 17, 2013
Revised: May 24, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 1. Overview of ditryptophan DKPs and CDPS-catalyzed reactions. (A) General structures of monocyclic, annulated, and diannulated
ditryptophan DKPs. Additionally, all known monocyclic ditryptophan DKP natural products are shown highlighted with a gray background. (B) CDPSs
use aa-tRNAs as substrates to generate a 2,5-DKP that is usually further modified by at least one tailoring enzyme found in the respective CDPS gene
cluster.
encoding genes have been found through the use of iterative PSI-
BLAST searches in many prokaryotic and a few eukaryotic
species, amounting to more than 50 putative gene clusters in
total.20,21 To date, all biochemically characterized CDPSs have
been shown to possess a certain degree of substrate promiscuity,
usually generating a number of different CDP products. In
prokaryotes, all main products consist of different combinations
of the four amino acids phenylalanine, leucine, tyrosine, and
methionine, while the only characterized eukaryotic CDPS
mainly generates cyclo(Trp-Xaa) (cWX), with Xaa being one of
the four previously mentioned amino acids.21,22 In the direct
surroundings of all putative prokaryotic CDPS genes, at least
one, but oftentimes more than one, putative DKP-modifying
enzyme can be found, including cytochrome P450s, non-heme
FeII/α-ketoglutarate-dependent oxygenases and different kinds
of transferases among others.21
Recently, a putative CDPS gene cluster could be identified in
the genome of the actinobacterium Actinosynnema mirum using
known CDPS sequences as BLAST queries.20 Here, we describe
the bioinformatic analysis of this CDPS gene cluster and the
elucidation of the biosynthetic pathway leading to two Sch
725418 analogues. Our studies identify Amir_4627 as the first
CDPS that shows high substrate specificity generating only one
main product. Additionally, it is the first CDPS family member
that can form ditryptophan DKPs and the first prokaryotic CDPS
whose main product constituents differ from the four amino
acids that usually make up CDPS-dependent CDPs (Phe, Leu,
Tyr, and Met). We show that after cWW formation a S-adenosyl-
L-methionine (SAM)-dependent N-methyltransferase
(Amir_4628) conducts two successive methylations at the
DKP-ring nitrogens and additionally show that it is able to
methylate four other CDPs. This makes Amir_4628 the first
identified DKP-ring-modifying methyltransferase.
MATERIALS AND METHODS
■
General Methods and Instruments. High-performance
liquid chromatography−tandem mass spectrometry (HPLC−
MS2) analyses of M9 fermentations were conducted using an
Agilent 1100 HPLC system equipped with a Nucleodur C-18 ec
column (125 mm × 4.6 mm, particle size of 3 μm, pore size of
110 Å, Macherey and Nagel, flow rate of 0.3 mL/min, solvent A
consisting of 0.1% (v/v) TFA and dd-H2O, solvent B consisting
of 0.1% (v/v) TFA and acetonitrile, linear gradient from 0 to 95%
over 50 min) connected to a Finnigan LTQ-FT Ultra instrument
(Thermo Fischer Scientific). MS fragmentation experiments
were conducted using collision-induced dissociation (CID)
fragmentation. Preparative HPLC purification of scaled-up
Amir_4628 assays was performed employing an Agilent 1100
Series HPLC system equipped with a Nucleodur C-18 ec column
(250 mm × 21 mm, particle size of 5 μm, pore size of 110 Å,
Macherey and Nagel, flow rate of 18 mL/min, solvent A
consisting of 0.1% (v/v) TFA and dd-H2O, solvent B consisting
of 0.1% (v/v) TFA and acetonitrile, linear gradient from 0 to 95%
over 50 min) and connected to a preparative scale fraction
collector. HPLC−MS analyses of in vitro assays were conducted
on an Agilent ESI-Quad 1100(A) Series MSD system equipped
with a Nucleodur C-18 ec column (125 mm × 4.6 mm, particle
size of 3 μm, pore size of 110 Å, Macherey and Nagel, flow rate of
0.3 mL/min, solvent A consisting of 0.1% (v/v) TFA and dd-
H2O, solvent B consisting of 0.1% (v/v) TFA and acetonitrile,
linear gradient from 0 to 95% over 40 min). Nuclear magnetic
resonance (NMR) spectra were recorded on a Bruker Avance
300 instrument with dimethyl sulfoxide-d6 (Sigma-Aldrich) as
B
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
the solvent. Kinetic parameters were determined by using
experimentally obtained starting velocities and the Enzyme
Kinetic Module for Sigma Plot version 8.0 (SPSS).
°C to an OD600 of 0.6 and subsequently induced using IPTG
(final concentration of 0.05 mM). After incubation at 18 °C for
18 h, cultures were harvested through centrifugation (17000 rpm
for 20 min at 4 °C), filtered through a 0.2 μm Filtropur S filter,
and subjected to HPLC−MS2 analysis. Produced amounts of
cWW were estimated using an authentic standard (Bachem).
Determination of the cWW Stereoconfiguration. The
absolute stereoconfiguration of cWW was determined by acid
hydrolysis of the DKP followed by chiral HPLC analysis and
comparison with authentic standards of D-/L-Trp; 500 μg of
cWW was hydrolyzed by the addition of 200 μL of 6 M HCl and
incubation at 95 °C for 24 h. After neutralization using 10 M
NaOH, the sample was subjected to chiral HPLC analysis
(Phenomenex, Chirex3126, 250 mm × 4.6 mm, particle size of 5
μm, pore size of 110 Å, flow rate of 1 mL/min, solvent A
consisting of 2 mM CuSO4 in dd-H2O, solvent B consisting of
methanol) using isocratic conditions (30% B).
In Vitro Characterization of Amir_4627. In a typical
Amir_4627 assay, recombinant Amir_4627 (10 μM) was
incubated with 50 μM tRNA mix from Saccharomyces cerevisiae
(Sigma-Aldrich), E. coli aminoacyl-tRNA synthetase mix (250
units, Sigma-Aldrich), and different amino acids (1 mM) at 30
°C. Additionally, the assay contained ATP (500 μM), MgCl2 (10
mM), KCl (30 mM), and DTT (2 mM) in Tris buffer [50 mM
Tris and 300 mM NaCl (pH 7.5)]. The reactions were stopped
by the addition of TCA [final concentration of 5% (v/v)] and
subjected to HPLC−MS analysis.
In Vivo Characterization of Amir_4628. M9 fermenta-
tions of E. coli BL21(DE3) cells transformed with both
Amir_4627 and Amir_4628 or AlbC and Amir_4628 were
conducted in a manner analogous to that of the Amir_4627
fermentations described above. To ensure reproducibility of
Amir_4627/Amir_4628 and AlbC/Amir_4628 fermentations,
the expression constructs of Amir_4627 and Amir_4628 were
constructed in a way that they carried both compatible selection
marker genes and compatible origins of replication (Amir_4627/
AlbC, kanamycin, f1; Amir_4628, spectinomycin, CDF).
In Vitro Characterization of Amir_4628. For time course
measurements, recombinantly produced Amir_4628 (20 μM)
was incubated with 1 mM S-adenosyl-L-methionine and 1 mM
cWW in Tris buffer [50 mM Tris and 300 mM NaCl (pH 7.5)].
Reactions were stopped by the addition of TCA [final
concentration of 5% (v/v)] and mixtures subjected to HPLC−
MS analysis.
The kinetic parameters of the second Amir_4628 methylation
step were determined by varying the concentration of purified
cWW-Me from 0 to 1 mM while keeping the SAM and enzyme
concentrations constant. Assays were initiated by the addition of
Amir_4628 and stopped by adding TCA [final concentration of
5% (v/v)] after 8 min, which was determined to be an
appropriate reaction time being in the linear conversion range.
Samples were then subjected to HPLC−MS analysis.
Generation of Expression Constructs for Amir_4627,
Amir_4628, and AlbC. The genes encoding Amir_4628 and
AlbC were amplified by polymerase chain reaction (PCR) from
genomic DNA of A. mirum (DSM 43827) and S. noursei (DSM
40635), respectively, obtained from DSMZ. The sequence of the
putative CDP cluster found in A. mirum, containing the
Amir_4626, Amir_4627, and Amir_4628 genes, was optimized
for Escherichia coli expression, synthesized, and cloned into the
pMK-RQ vector by GeneArt (Life Technologies). The pMK-RQ
vector was subsequently used for PCR amplification of different
constructs. All PCR primers are listed in Table S1 of the
Supporting Information. The Amir_4628 gene was cloned into
the first multiple cloning site of the pCDFDuet-1 vector, while
Amir_4627 and albC were cloned into the pET28a and pET41a
vectors, respectively. After transformation of E. coli TOP10 cells
(Invitrogen, Life Technologies), the resulting constructs were
confirmed by sequencing (GATC-Biotech).
Production and Purification of Recombinant Proteins.
The resulting expression constructs of Amir_4627 and
Amir_4628 were used to transform E. coli BL21(DE3) cells
(Novagen), which were grown at 37 °C to an OD600 of 0.6 and
subsequently induced using isopropyl thiogalactopyranoside
(IPTG, final concentration of 0.05 mM). Then the cultures were
grown for 18 h at 18 °C and harvested via centrifugation (7000
rpm for 20 min at 4 °C). Cells were suspended in Tris buffer A
[100 mM Tris, 300 mM NaCl, 10 mM β-mercaptoethanol, and
5% glycerol (pH 8)] and lysed with a French press (SLM
Aminco, Thermo French press). The soluble lysate was
separated from cell debris by centrifugation (17000 rpm for 30
min at 4 °C), filtered through a 0.2 μm Filtropur S filter
(Sarstedt), and then subjected to Ni-NTA affinity purification
̈
using an Akta Prime system (GE Healthcare Life Sciences). After
equilibration of the Ni-NTA column with Tris buffer A and
application of the lysate, the bound protein was eluted using
gradient elution with an increasing imidazole concentration
(from 2 to 95% over 30 min). Protein-containing fractions were
identified by sodium dodecyl sulfate−polyacrylamide gel electro-
phoresis (SDS−PAGE), combined, and in the case of
Amir_4627 subjected to heparin affinity chromatography using
a HiTrap Heparin HP column (GE Healthcare Biosciences).
Amir_4627 was eluted employing a linear salt gradient (from 0.3
to 2 M NaCl). Designated protein fractions were identified by
SDS−PAGE and combined. The chromatographic step using a
Heparin HP column was omitted from the purification protocol
of Amir_4628. After the one-step (Amir_4628) or two-step
(Amir_4627) purification procedure, protein-containing frac-
tions were subjected to buffer exchange using Tris dialysis buffer
[100 mM Tris, 50 mM NaCl, 10 mM β-mercaptoethanol, and 5%
glycerol (pH 7.5)] with a HiPrep desalting column (GE
Healthcare Life Sciences). Protein-containing fractions were
pooled, concentrated, and flash-frozen using liquid N2 before
being stored at −80 °C until further use.
Purification of cWW-Me for Use in Amir_4628 Kinetic
Analysis. Amir_4628 assays were scaled up, stopped after 20
min, which was determined to be the most appropriate reaction
time for maximizing the cWW-Me yield, and subsequently
subjected to preparative HPLC purification. Product fractions
were combined, flash-frozen using liquid N2, and lyophilized.
The resulting products were stored at −20 °C until further use
(Figure S1 of the Supporting Information).
In Vivo Characterization of Amir_4627. Preparatory
cultures of cells containing the Amir_4627 expression construct
were used to inoculate [1:50 (v/v)] M9 minimal medium [17 g/
L Na2HPO4·12H2O, 3 g/L KH2PO4, 0.5 g/L NaCl, 1 g/L
NH4Cl, a 1 mL/L MgSO4 solution (2 M), a 0.2 mL/L CaCl2
solution (0.5 M) (pH 7.0), a 10 mL/L glucose solution (40%, w/
v), and a 0.2 mL/L vitamin mix (Table S2 of the Supporting
Information) after autoclaving]. M9 cultures were grown at 37
Antibacterial Activity Assays of cWW, cWW-Me, and
cWW-Me2. Antibacterial activities were evaluated against
bacterial strains E. coli BL21(DE3), Bacillus subtilis MR168,
C
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 2. Bioinformatic analysis of the A. mirum CDPS gene cluster. (A) Multiple-sequence alignment of known CDPSs and Amir_4627. The secondary
structural elements of AlbC are colored green above the alignment. Residues conserved in all characterized CDPSs are highlighted with a red
background, while residues conserved in all prokaryotic members are highlighted with a yellow background. The positively charged residues in helix α4
proposed to mediate protein−tRNA interactions are colored blue. (B) Phylogenetic tree based on the multiple-sequence alignment shown in panel A
with the respective CDPS main products. The product shown for Ndas_1148 is based on studies currently conducted by our group. (C) A. mirum CDPS
gene cluster with the truncated IS4 transposase gene located directly downstream of Amir_4628. Abbreviations: ACS, acyl-CoA ligase; MT,
methyltransferase; TP, transposase.
and Micrococcus flavus ATCC 10240 using spot-on-lawn assays
on LB (lysogeny broth). Soft agar was prepared by adding 0.7%
agar to liquid LB. Soft agar (10 mL) was then inoculated using
100 μL of an overnight culture of the respective organism and
spread evenly on an LB agar plate. Next, 10 μL of the respective
compound and dilution (1, 10, and 100 μg/mL cWW, cWW-Me,
and cWW-Me2) was spotted on the soft agar plate followed by
overnight incubation at 30 °C.
sequence identity and similarity of characterized CDPSs among
each other. To further investigate if Amir_4627 is a functional
CDPS, a Clustal Omega23,24 multiple-sequence alignment with
known CDPSs was generated. This analysis shows that the
active-site nucleophile Ser36 (β3), 11 other positions usually
conserved in prokaryotic CDPSs, and the positively charged helix
(α4) proposed to mediate tRNA binding during enzyme catalysis
are all well-preserved in Amir_4627 (Figure 2A).
Using the protein threading and fold recognition server I-
TASSER,25,26 a structural model of Amir_4627 was generated on
the basis of the crystal structure of the CDPS YvmC from Bacillus
licheniformis as a template (Figure S2 of the Supporting
Information).27 The quality of a predicted structure can generally
be judged by inspecting the respective TM score of a structural
model and the root-mean-square deviation (rmsd) of the Cα
atoms between the generated model and template. The TM
RESULTS
■
Bioinformatic Analysis of a Putative CDP Biosynthetic
Gene Cluster in A. mirum. A bioinformatic survey in 2010 first
identified a putative CDPS gene (Amir_4627) in A. mirum using
AlbC (the prototypical CDPS) and other characterized CDPS
family members as queries in iterative PSI-BLAST searches.20
The sequence of the identified gene product, annotated as a
“hypothetical protein”, is 33% identical and 46% similar to that of
AlbC, which is in accordance with the only moderate level of
score for the Amir_4627 model is 0.93
0.06, indicating a
correctly predicted overall topology and a high degree of
D
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 3. HPLC−MS analysis and product identification of the CDPS Amir_4627. (A) UV traces for Amir_4627 and the empty vector control of M9
fermentations. (B) Reaction catalyzed by Amir_4627 identified as a functional CDPS family member. (C) MS2 fragmentation spectrum of cWW (m/z
373). Relevant fragment ions and neutral losses used to identify the CDPS Amir_4627 reaction product are shown.
structural homology to the template, which is additionally
confirmed by the low rmsd of 2.6 1.9 Å (over 215 Cα atoms).
Taken together, those in silico results suggest that Amir_4627 is
indeed a CDPS family member.
gene from A. mirum into pET28a were successful, but after IPTG
induction, no protein production could be detected under
various expression conditions. To circumvent this problem, an E.
coli optimized synthetic version of the Amir_4627 gene that
could again be successfully cloned into pET28a and subsequently
introduced into E. coli BL21(DE3) was used. Using this
optimized expression construct, the encoded CDPS could be
successfully produced, and subsequent HPLC−MS analysis of
overnight M9 fermentation supernatants and comparison with
an empty vector control confirmed the presence of one
additional major peak in the Amir_4627 UV trace (Figure 3A).
HR-MS analysis showed that the identified peak belongs to a
compound with an m/z of 373.1658. Checking against the
masses of all possible DKPs resulting from the condensation of
two proteinogenic amino acids, we found the experimentally
determined mass corresponds well to the mass of cWW (m/z
373.1659) (Figure 3B). This assumption could be confirmed by
detailed HPLC−MS2 studies (Figure 3C) and NMR spectros-
copy (Table S3 and Figures S3−S8 of the Supporting
Information). The fragmentation spectra clearly show the
characteristic fragmentation pattern of a 2,5-DKP moiety with
successive neutral losses of 28 Da (-CO) and 45 Da (-HCONH2)
or vice versa.28 Additionally, five fragments characteristic of Trp,
including the immonium ion (m/z 159), could be identified. The
absence of any further fragment peaks in the MS2 spectra
characteristic of other amino acids and the comparison of 1H and
13C NMR spectra with literature data12 confirm that the isolated
compound is indeed cWW (yield of 29 mg/L of M9
supernatant). For the determination of the cWW stereo-
configuration, acid hydrolysis of the isolated compound followed
by chiral HPLC analysis was conducted. As expected for a CDPS-
dependent DKP, only L-Trp could be identified (Figure S9 of the
Supporting Information). To investigate if the CDPS in question
produces any other minor DKP products, we checked the
HPLC−MS chromatograms for all possible DKPs resulting from
proteinogenic amino acids, but no additional differential peaks
could be identified.
On the basis of the constructed multiple-sequence alignment,
a phylogenetic tree was generated using the neighbor joining
method (Figure 2B). The tree shows that all CDPSs that form
cLL as a main product cluster together, while all other
characterized CDPSs can be found in a separate clade. This
expectedly reflects the evolutionary relationship of the respective
producing organisms with cLL-producing species belonging to
the phyla proteobacteria and firmicutes (except for JK0923
originating from the actinobacterium Corynebacterium jeikeium)
and the remaining organisms being part of the phylum
actinobacteria (except for Nvec-CDPS2 originating from the
eukaryotic cnidarian Nematostella vectensis). The CDPS in
question, Amir_4627, can be found in the lower main clade,
interestingly with Nvec-CDPS2 as its closest relative.
With respect to the genetic surroundings of Amir_4627, a gene
annotated as an “AMP-dependent synthetase and ligase”
belonging to the family of long-chain fatty acyl-CoA ligases
(Amir_4626) is located directly upstream of Amir_4627 while a
gene encoding a “CheR-type methyltransferase” (Amir_4628)
can be found directly downstream of Amir_4627. It is possible
that one or both of those genes are part of a dedicated CDP
biosynthetic pathway. Additionally, a truncated transposase of
the IS4 family (Amir_4629) is located directly downstream of
Amir_4628, which could indicate that the A. mirum CDPS gene
cluster described above was acquired through horizontal gene
transfer (Figure 2C).
Amir_4627 Represents a Highly Specific CDPS Gen-
erating Exclusively cWW. To investigate the ability of
Amir_4627 to synthesize DKPs, a heterologous expression
approach was used in which E. coli, harboring an Amir_4627
expression construct, was cultivated in M9 minimal medium.
This approach followed previous studies reporting that when a
functional CDPS is heterologously produced in E. coli the
generated DKPs can be identified directly in the culture
supernatant.17 Initial attempts to clone the native Amir_4627
Subsequently, in vitro studies using purified Amir_4627 were
conducted to further investigate DKP formation with regard to
E
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
substrate specificity. Following a three-step purification
procedure, active CDPS could be obtained with a yield of 0.7
mg/L of culture (Figure S10 of the Supporting Information). To
provide aa-tRNAs as substrates in in vitro assays, they had to be
generated in situ by using the respective amino acids, a tRNA mix
isolated from S. cerevisiae, and a mixture of E. coli aminoacyl-
tRNA synthetases. Different combinations of the amino acids
usually found in DKPs produced by CDPSs were tested, but only
when Trp was present could DKP formation be detected with
cWW being the only observed product (Figure 4).
confirm this assumption and to clarify the positions at which the
methylations take place, detailed HPLC−MS2 studies were
conducted (Figure 6A). The resulting fragmentation spectra
clearly show that the two new compounds both contain a 2,5-
DKP moiety because of the characteristic neutral losses discussed
above. Additionally, the presence of characteristic Trp fragments
indicates that the two compounds are indeed cWW derivatives.
To elucidate the positions at which methylation took place first,
the increase and the decrease in the intensity of fragments
corresponding to Trp (m/z 159) and putatively methylated Trp
(m/z 173) immonium ions were analyzed. Via comparison of the
fragmentation spectra of cWW, putative cWW-Me, and putative
cWW-Me2, it becomes obvious that in the case of cWW only
unmethylated immonium ions (m/z 159) can be detected, while
for putative cWW-Me, both methylated and unmethylated
immonium ions can be found in equal amounts (m/z 173 and
159, respectively). In the case of putative cWW-Me2, as expected,
only methylated immonium ions (m/z 173) are present. Upon
closer examination of the characteristic DKP fragmentation
pattern, in the case of putative cWW-Me and cWW-Me2 not only
can the expected neutral losses of 28 (-CO) and 45 Da
(-HCONH2) be detected, but an additional neutral loss of 59 Da
can be observed, which completely substitutes the neutral loss of
45 Da (-HCONH2) in the case of putative cWW-Me2. This
neutral loss of 59 Da corresponds to a HCONHCH3 fragment,
indicating that a methylation took place in the DKP ring itself.
Combining this fact with the detection of methylated immonium
ions for cWW-Me and cWW-Me2 and the NMR spectroscopic
data leads to the conclusion that the only reasonable methylation
positions are the two α-nitrogens of the 2,5-DKP moiety (Figure
6B).
Figure 4. In vitro time course analysis of the CDPS-catalyzed formation
of cWW.
The SAM-Dependent Methyltransferase Amir_4628 Is
Able To Conduct N-Methylations at the DKP Core of
Different CDPs. To investigate the potential of the putative
methyltransferase Amir_4628 to modify DKPs in general and
cWW in particular, we successfully cloned the respective native
gene into pCDFDuet-1 and used it in M9 fermentation
experiments. Subsequent HPLC−MS analysis of the M9 culture
supernatants and comparison with empty vector and Amir_4627
controls confirmed the presence of two new differential peaks at
later retention times (Figure 5). HR-MS analysis revealed the
masses of the two compounds to correspond to m/z 387.1815
and 401.1971. Those are in good agreement with the masses
expected for singly (cWW-Me; m/z 387.1815) and doubly
(cWW-Me2; m/z 401.1972) methylated cWW, respectively. To
To test the antibacterial activities of the three identified
compounds (cWW, cWW-Me, and cWW-Me2), we subjected
them to spot-on-lawn assays using E. coli Bl21(DE3), B. subtilis
MR168, and M. flavus ATCC 10240 at concentrations of ≤100
μg/mL. None of the tested compounds showed any significant
antibacterial activity against the tested indicator strains.
For subsequent in vitro studies, the methyltransferase
Amir_4628 was recombinantly produced in E. coli and purified
with a yield of 2.5 mg/L of culture. A time course analysis of the
methylations of cWW was conducted showing that two
successive separate methylation steps take place where the first
methylation product, cWW-Me, is released from the enzyme
after the initial methylation reaction (Figure 7A). This can be
deduced by the fact that large amounts of free cWW-Me can be
observed during the Amir_4628-catalyzed reaction with a
maximal relative amount of cWW-Me after an incubation time
of ∼20 min. After incubation for 60 min, formation of the final
methylation product cWW-Me2 has almost come to a halt,
probably because of enzyme degradation because increased levels
of enzyme precipitate could be observed using longer incubation
times and methylation activity could be restored by adding fresh
enzyme. To further characterize Amir_4628, methylation assays
were scaled up, incubated for 20 min, at which time the largest
amount of cWW-Me is present in the reaction mixture, and
subjected to preparative HPLC purification. cWW-Me was then
used as a substrate to determine the kinetics of the second
methylation step (Figure 7B). For the methylation of cWW-Me,
an apparent Km of 51 4 μM and a kcat of 4.4 0.13 min−1 were
determined.
Figure 5. HPLC analysis of M9 culture supernatants resulting from
cotransformation experiments using Amir_4627 and Amir_4628. UV
traces of M9 supernatants from cultures transformed with both
constructs and an empty vector control are shown.
To further test the substrate specificity of the methyltransfer-
ase Amir_4628, M9 fermentations were conducted using E. coli
BL21(DE3) cells transformed with both albC and Amir_4628.
F
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
Figure 6. Elucidation of the regiochemistry of the investigated methylation reactions. (A) MS2 fragmentation comparison of cWW, cWW-Me, and
cWW-Me2. On the left, the decrease and increase of unmethylated and methylated Trp immonium ions are shown. On the right, the differences in the
characteristic 2,5-DKP fragmentation patterns between methylated and unmethylated cWW are shown. Series of neutral losses that are observed only in
methylated cWW derivatives are highlighted in cyan. Neutral losses containing a newly installed methyl group are colored red. (B) Reaction sequence
resulting in the formation of cWW-Me2 catalyzed by Amir_4628. The newly installed methyl groups are colored red.
Figure 7. In vitro analysis of the methyltransferase Amir_4628. Error bars indicate standard errors of three independent experiments. (A) Time course
for the two successive methylation reactions. (B) Kinetic characterization of the second methylation step and calculated kinetic parameters.
The CDPS AlbC has been previously characterized and was
shown to produce all possible DKPs using the amino acids Phe,
Leu, Tyr, and Met.17 With this approach, it was possible to test
Amir_4628 substrate specificity for 10 different DKPs in one
experiment. Four DKPs could be shown to be valid substrates of
Amir_4628 (cFF, cFL, cFY, and cFM), all containing at least one
phenylalanine residue, with cFF-Me and cFF-Me2 being the
major methylation products (Table 1 and Figure S11 of the
Supporting Information). The only other DKP besides cFF for
which double methylation could be observed was cFY.
G
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
CDPS Amir_4627 produces only a single product contradicts the
notion that the characteristic product profiles observed for most
CDPSs result from a high specificity for the first aminoacyl-tRNA
substrate and a relaxed specificity for the second aminoacyl-
tRNA. On the contrary, it seems that in the case of Amir_4627
the second step is also highly specific. With regard to the
specificity of the first incorporated amino acid, it was previously
proposed that by comparing the residues lining the CDPS
substrate pocket some general rules governing substrate
specificity in CDPS catalysis could be deduced. For the cWX-
producing CDPS Nvec-CDPS2 from N. vectensis, it was
suggested that the specificity for Trp resulted from the presence
of a large number of aromatic residues in the active site pocket
(F61, Y194, and F209), which would supposedly facilitate Trp
binding through π-interactions with the indole side chain.22
Using a structure-based alignment, a comparison of the active site
residues of all characterized CDPSs was generated (Table S4 of
the Supporting Information) showing that at the corresponding
three positions in Amir_4627 only one aromatic amino acid can
be found (V65, F184, and N199). Via comparison of this active
site with those of CDPSs of other specificities, it seems obvious
that no easy prediction about the identity of the first incorporated
amino acid, solely based on the active site residues, can be made.
The high specificity of the second reaction step in Amir_4627
catalysis seems to be governed by an as yet unknown mechanism.
This topic could be addressed in future studies through
cocrystallization experiments with CDPS and an appropriate
aminoacyl-tRNA analogue.
Turning now to the two putative DKP-modifying enzymes
located in the A. mirum CDPS cluster, we could not assign a
function to Amir_4626 annotated as a fatty acyl-CoA ligase.
Neither in vivo fermentation experiments using all three genes
(Amir_4626, Amir_4627, and Amir_4628) in their native
organization nor in vitro studies of the recombinant protein led
to the detection of any reaction products of Amir_4626. This
could be explained by the absence of the correct fatty acid
substrate in the heterologous host and the in vitro assays (C12:0,
C14:0, C16:0, and C18:0 fatty acids were the substrates tested).
Additionally, the absence of a dedicated acyl-CoA transferase in
the A. mirum CDPS gene cluster makes it possible that even if the
correct fatty acid substrate was present in E. coli, the Amir_4626
product could not be transferred onto its dedicated nucleophile.
The second putative tailoring enzyme located directly down-
stream of the Amir_4627 gene is Amir_4628 annotated as a
CheR-type methyltransferase. With respect to the occurrence of
different putative DKP-tailoring enzymes in CDPS gene clusters,
by far the most common group consists of various kinds of
oxidoreductases, including cytochrome P450s, non-heme FeII/α-
ketoglutarate-dependent oxygenases, and other flavin- or
pyridoxal phosphate-dependent oxidases.21,29,30 On the other
hand, only four organisms besides A. mirum in which
methyltransferase genes are closely associated with putative
CDPS genes are known, namely Nocardiopsis dassonvillei (two
MTs), Nocardiopsis alba (two MTs), the organism encoding a
CDPS cluster highly similar to that found in A. mirum mentioned
above [Streptomyces sp. AA4 (one MT)] and Kutzneria sp. 744
(two MTs). The four Nocardiopsis MTs (currently under
investigation in our group) all belong to the PF00891 family
that includes various small molecule O-methyltransferases,
among them DnrK (Streptomyces peucetia) and RdmB
(Streptomyces purpurascen), which conduct different O-methyl-
ations at the anthracycline scaffold,31,32 as well as various caffeic
acid and isoflavone O-methyltransferases. One of the Kutzneria
Table 1. Substrate Specificities of the Methyltransferase
Amir_4628
a
b
b
cXX
cXX-Me (relative amount)
cXX-Me2 (relative amount)
cWW
cFF
68%
31%
4%
22%
11%
<1%
−
cFY
cFM
cFL
2%
2%
−
a
Values obtained for relative production amounts are based on 18 h
b
fermentations in M9 medium. Relative amount compared to the
respective cXX calculated using UV trace integrals.
DISCUSSION
■
Trp-containing CDPs are among the most common natural
products harboring a 2,5-DKP moiety with a large majority of
known compounds isolated from fungal sources. In bacteria,
CDPs containing one Trp residue are also quite common, while
only one example of a monocyclic ditryptophan DKP could be
isolated so far, Sch 725418 (Micromonospora sp.).12 Until now,
neither a biosynthetic gene cluster nor a biochemical pathway for
the generation of a monocyclic ditryptophan CDP has been
reported.
Starting with the hypothesis that Amir_4627, recently
identified in a bioinformatic survey, is a putative CDPS, we set
out to further analyze Amir_4627 and the surrounding gene
products initially using different in silico methods. The putative
CDPS gene cluster contains two genes encoding candidate DKP-
modifying enzymes (Amir_4626 and Amir_4628) located
directly up- and downstream of Amir_4627 (Figure 2C).
Together, those three genes span approximately 2.9 kb, not
including the truncated IS4 transposase located directly
downstream of the gene encoding the methyltransferase
Amir_4628. The presence of a transposase gene may indicate
that A. mirum has acquired this particular CDPS gene cluster via
horizontal gene transfer, which is supported by the fact that a
highly similar gene cluster with an identical genetic organization
is present in its close relative Streptomyces sp. AA4 (Figure S12 of
the Supporting Information). We turn now to the two putative
CDP tailoring enzymes Amir_4626 and Amir_4628, which are
annotated as a long-chain fatty acyl-CoA ligase and a CheR-type
methyltransferase, respectively. The fatty acyl-CoA ligase family
catalyzes a two-step process in which a fatty acid is initially
activated as an adenylate, which is then attacked by CoA to form
a fatty acyl-CoA product. Generally, this activated fatty acid is
then transferred onto a specific nucleophile through the action of
an acyl-CoA transferase. No such transferase is encoded in the
vicinity of Amir_4626, raising the question of the function it
could play in CDP biosynthesis, which will be further addressed
below.
Our initial assumption that Amir_4627 belongs to the CDPS
family was proven to be correct through a combination of in vivo
and in vitro analyses. It could be shown that Amir_4627 is able to
produce a single DKP product, cWW (Figures 3 and 4). This
high substrate specificity as well as the identity of the product is
unusual for CDPSs. So far, all reported CDPS family members
possess relaxed substrate specificities producing between 2 and
13 different DKPs.17,22 Additionally, the identity of the DKP
constituents of the Amir_4627 product is unusual because all
known prokaryotic CDPSs produce DKPs consisting of different
combinations of the amino acids Phe, Leu, Tyr, and Met, while
the only characterized CDPS of eukaryotic origin additionally
generates Trp-containing CDPs, but not cWW. The fact that the
H
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
MTs belongs to the largely uncharacterized PF04672 family that
includes human histamine N-methyltransferase,33 while the
other mentioned MTs, including Amir_4628, are part of the
PF01739 family. Among the members of the PF01739 family,
different protein glutamate O-methyltransferases can be found as
well as ToxA (Burkholderia glumae), which conducts N-
methylations in toxoflavin biosynthesis.34 The structure of
toxoflavin consists of fused pyrimidinedione and 1,2,4-triazine
rings in which one nitrogen in each ring is methylated through
the action of ToxA. This specificity may indicate that Amir_4628
also acts as a N-methyltransferase on a six-memberd nitrogen-
containing heterocyclic substrate. This hypothesis could indeed
be proven right. In this study, we were able to show that
Amir_4628 catalyzes two successive N-methylations at the 2,5-
DKP core of five different CDPs. The position of methylation
was investigated via HPLC−MS2 studies in which both the
characteristic neutral losses of the 2,5-DKP scaffold and
methylated and unmethylated immonium ions were used to
elucidate the exact regiospecificity of the Amir_4628-catalyzed
reactions. The native substrate of Amir_4628 is cWW because it
is the only DKP produced by the associated CDPS Amir_4627.
The four other substrates accepted by Amir_4628, discovered
through in vivo fermentation experiments using the formerly
characterized promiscuous CDPS AlbC, are cFF, cFY, cFM, and
cFL, listed in order of methylation efficiency. This indicates that
Amir_4628 prefers DKPs containing large aromatic amino acids.
Additionally, only substrates containing two aromatic amino
acids could be doubly methylated (cWW, cFF, and cFY),
suggesting that only α-nitrogens of aromatic amino acids are
valid substrates for Amir_4628. It seems that the planar and
aromatic nature of those substrates plays an important role in
substrate recognition and/or catalysis. Kinetic analysis of the
second methylation step catalyzed by Amir_4628 shows that its
apparent Km is approximately an order of magnitude larger than
that for other CheR-type methyltransferases (protein-glutamate/
aspartate O-methyltransferases), probably reflecting the differ-
ence between the interaction with a small molecule substrate and
a larger protein surface. However, the turnover numbers are
comparable.35−37
A large number of different modifying enzymes are known to
originate from bacterial and especially fungal sources that use
Trp-containing DKPs as substrates. Those include various
varieties of prenyltransferases with different chemo- and
regiospecificities, hydroxylases and other oxidoreductases
modulating the oxidation state of the indole side chain, and
various other enzymes triggering, for example, the formation of
annulated Trp-containing DKPs. Many of the modifications
introduced by those enzymes significantly influence the
biological activity of the respective CDPs through the alteration
of the hydrophobicity, shape, or rigidity of the natural product
scaffold. The discovery of a potent catalyst for cWW formation
encoded by a small gene opens the possibility for combinatorial
in vivo as well as chemoenzymatic approaches in which different
tailoring enzymes known to act upon Trp-containing DKPs can
be used to generate derivatives of known natural products or
entirely new chemical entities with potentially improved or new
biological activities.
Amir_4627 gene, comparison of CDPS active site residues,
purification of cWW-Me via HPLC, structural models of
Amir_4627, AlbC, and Nvec-CDPS2, chiral HPLC chromato-
gram of cWW acid hydrolysis, SDS−PAGE gels of Amir_4627
and Amir_4628, HPLC−MS analysis of AlbC/Amir_4628 M9
culture supernatants, and comparison of CDPS clusters in A.
mirum and Streptomyces sp. AA4. This material is available free of
charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
■
*Department of Chemistry, Philipps-University Marburg, 35043
Marburg, Germany. E-mail: marahiel@staff.uni-marburg.de.
Phone: +49 (0) 6421 28 25722. Fax: +49 (0) 6421 28 22191.
Funding
This work has been supported by the Deutsche Forschungsge-
meinschaft and the LOEWE Center for Synthetic Microbiology
(SYNMIKRO).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
We thank Dr. Uwe Linne for support regarding mass
spectrometric analyses.
■
REFERENCES
■
(1) Cain, C. C., Lee, D., Waldo, R. H., III, Henry, A. T., Casida, E. J., Jr.,
Wani, M. C., Wall, M. E., Oberlies, N. H., and Falkinham, J. O., III
(2003) Synergistic antimicrobial activity of metabolites produced by a
nonobligate bacterial predator. Antimicrob. Agents Chemother. 47, 2113−
2117.
(2) Magyar, A., Zhang, X., Kohn, H., and Widger, W. R. (1996) The
antibiotic bicyclomycin affects the secondary RNA binding site of
Escherichia coli transcription termination factor Rho. J. Biol. Chem. 271,
25369−25374.
(3) Waring, P., and Beaver, J. (1996) Gliotoxin and related
epipolythiodioxopiperazines. Gen. Pharmacol. 27, 1311−1316.
(4) Yamazaki, Y., Tanaka, K., Nicholson, B., Deyanat-Yazdi, G., Potts,
B., Yoshida, T., Oda, A., Kitagawa, T., Orikasa, S., Kiso, Y., Yasui, H.,
Akamatsu, M., Chinen, T., Usui, T., Shinozaki, Y., Yakushiji, F., Miller, B.
R., Neuteboom, S., Palladino, M., Kanoh, K., Lloyd, G. K., and Hayashi,
Y. (2012) Synthesis and structure-activity relationship study of
antimicrotubule agents phenylahistin derivatives with a didehydropiper-
azine-2,5-dione structure. J. Med. Chem. 55, 1056−1071.
(5) Kanzaki, H., Yanagisawa, S., Kanoh, K., and Nitoda, T. (2002) A
novel potent cell cycle inhibitor dehydrophenylahistin: Enzymatic
synthesis and inhibitory activity toward sea urchin embryo. J. Antibiot.
55, 1042−1047.
(6) Borthwick, A. D. (2012) 2,5-Diketopiperazines: Synthesis,
reactions, medicinal chemistry, and bioactive natural products. Chem.
Rev. 112, 3641−3716.
(7) Ortiz-Castro, R., Diaz-Perez, C., Martinez-Trujillo, M., del Rio, R.
E., Campos-Garcia, J., and Lopez-Bucio, J. (2011) Transkingdom
signaling based on bacterial cyclodipeptides with auxin activity in plants.
Proc. Natl. Acad. Sci. U.S.A. 108, 7253−7258.
(8) Li, J., Wang, W., Xu, S. X., Magarvey, N. A., and McCormick, J. K.
(2011) Lactobacillus reuteri-produced cyclic dipeptides quench agr-
mediated expression of toxic shock syndrome toxin-1 in staphylococci.
Proc. Natl. Acad. Sci. U.S.A. 108, 3360−3365.
(9) Degrassi, G., Aguilar, C., Bosco, M., Zahariev, S., Pongor, S., and
Venturi, V. (2002) Plant growth-promoting Pseudomonas putida
WCS358 produces and secretes four cyclic dipeptides: Cross-talk with
quorum sensing bacterial sensors. Curr. Microbiol. 45, 250−254.
(10) Holden, M. T., Ram Chhabra, S., de Nys, R., Stead, P., Bainton, N.
J., Hill, P. J., Manefield, M., Kumar, N., Labatte, M., England, D., Rice, S.,
Givskov, M., Salmond, G. P., Stewart, G. S., Bycroft, B. W., Kjelleberg, S.,
ASSOCIATED CONTENT
* Supporting Information
Primers used in this study, M9 minimal medium vitamin mix,
table of chemical shifts, and 1H and 13C NMR spectra of cWW,
cWW-Me, and cWW-Me2, sequence of the E. coli-optimized
■
S
I
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX

Biochemistry
Article
and Williams, P. (1999) Quorum-sensing cross talk: Isolation and
chemical characterization of cyclic dipeptides from Pseudomonas
aeruginosa and other Gram-negative bacteria. Mol. Microbiol. 33,
1254−1266.
(28) Guo, Y. C., Cao, S. X., Zong, X. K., Liao, X. C., and Zhao, Y. F.
(2009) ESI-MSn study on the fragmentation of protonate cyclic-
dipeptides. Spectroscopy 23, 131−139.
(29) Belin, P., Le Du, M. H., Fielding, A., Lequin, O., Jacquet, M.,
Charbonnier, J. B., Lecoq, A., Thai, R., Courcon, M., Masson, C.,
Dugave, C., Genet, R., Pernodet, J. L., and Gondry, M. (2009)
Identification and structural basis of the reaction catalyzed by CYP121,
an essential cytochrome P450 in Mycobacterium tuberculosis. Proc. Natl.
Acad. Sci. U.S.A. 106, 7426−7431.
(30) Gondry, M., Lautru, S., Fusai, G., Meunier, G., Menez, A., and
Genet, R. (2001) Cyclic dipeptide oxidase from Streptomyces noursei.
Isolation, purification and partial characterization of a novel, amino acyl
α,β-dehydrogenase. Eur. J. Biochem. 268, 1712−1721.
(31) Jansson, A., Koskiniemi, H., Mantsala, P., Niemi, J., and Schneider,
G. (2004) Crystal structure of a ternary complex of DnrK, a
methyltransferase in daunorubicin biosynthesis, with bound products.
J. Biol. Chem. 279, 41149−41156.
(11) Kozlovsky, A. G., Vinokurova, N. G., Adanin, V. M., Burkhardt, G.,
Dahse, H. M., and Grafe, U. (2000) New diketopiperazine alkaloids
from Penicillium fellutanum. J. Nat. Prod. 63, 698−700.
(12) Yang, S. W., Chian, T. M., Terracciano, J., Loebenberg, D., Chen,
G., Patel, M., Gullo, V., Pramanik, B., and Chu, M. (2004) Structure
elucidation of a new diketopiperazine Sch 725418 from Micromonospora
sp. J. Antibiot. 57, 345−347.
(13) Shiono, Y., Akiyama, K., and Hayashi, H. (1999) New Okaramine
Congeners, Okamines J, K, L, M and Related Compounds, from
Penicillium simplicissimum ATCC90288. Biosci., Biotechnol., Biochem. 63,
1910−1920.
(14) Gardiner, D. M., and Howlett, B. J. (2005) Bioinformatic and
expression analysis of the putative gliotoxin biosynthetic gene cluster of
Aspergillus fumigatus. FEMS Microbiol. Lett. 248, 241−248.
(15) Healy, F. G., Wach, M., Krasnoff, S. B., Gibson, D. M., and Loria,
R. (2000) The txtAB genes of the plant pathogen Streptomyces
acidiscabies encode a peptide synthetase required for phytotoxin
thaxtomin A production and pathogenicity. Mol. Microbiol. 38, 794−
804.
(16) Robbel, L., Knappe, T. A., Linne, U., Xie, X., and Marahiel, M. A.
(2010) Erythrochelin: A hydroxamate-type siderophore predicted from
the genome of Saccharopolyspora erythraea. FEBS J. 277, 663−676.
(17) Gondry, M., Sauguet, L., Belin, P., Thai, R., Amouroux, R., Tellier,
C., Tuphile, K., Jacquet, M., Braud, S., Courcon, M., Masson, C., Dubois,
S., Lautru, S., Lecoq, A., Hashimoto, S., Genet, R., and Pernodet, J. L.
(2009) Cyclodipeptide synthases are a family of tRNA-dependent
peptide bond-forming enzymes. Nat. Chem. Biol. 5, 414−420.
(18) Lautru, S., Gondry, M., Genet, R., and Pernodet, J. L. (2002) The
albonoursin gene Cluster of S noursei biosynthesis of diketopiperazine
metabolites independent of nonribosomal peptide synthetases. Chem.
Biol. 9, 1355−1364.
(32) Jansson, A., Niemi, J., Lindqvist, Y., Mantsala, P., and Schneider,
G. (2003) Crystal structure of aclacinomycin-10-hydroxylase, a S-
adenosyl-L-methionine-dependent methyltransferase homolog involved
in anthracycline biosynthesis in Streptomyces purpurascens. J. Mol. Biol.
334, 269−280.
(33) Rutherford, K., Parson, W. W., and Daggett, V. (2008) The
histamine N-methyltransferase T105I polymorphism affects active site
structure and dynamics. Biochemistry 47, 893−901.
(34) Shingu, Y., and Yoneyama, K. (2004) Essential regulator gene
toxR for toxoflavin biosynthesis of Burkholderia glumae. J. Gen. Plant
Pathol. 70, 108−114.
(35) Schubert, H. L., Blumenthal, R. M., and Cheng, X. (2003) Many
paths to methyltransfer: A chronicle of convergence. Trends Biochem. Sci.
28, 329−335.
(36) Djordjevic, S., and Stock, A. M. (1997) Crystal structure of the
chemotaxis receptor methyltransferase CheR suggests a conserved
structural motif for binding S-adenosylmethionine. Structure 5, 545−
558.
(37) Simms, S. A., and Subbaramaiah, K. (1991) The kinetic
mechanism of S-adenosyl-L-methionine: Glutamylmethyltransferase
from Salmonella typhimurium. J. Biol. Chem. 266, 12741−12746.
(19) Sauguet, L., Moutiez, M., Li, Y., Belin, P., Seguin, J., Le Du, M. H.,
Thai, R., Masson, C., Fonvielle, M., Pernodet, J. L., Charbonnier, J. B.,
and Gondry, M. (2011) Cyclodipeptide synthases, a family of class-I
aminoacyl-tRNA synthetase-like enzymes involved in non-ribosomal
peptide synthesis. Nucleic Acids Res. 39, 4475−4489.
(20) Aravind, L., de Souza, R. F., and Iyer, L. M. (2010) Predicted class-
I aminoacyl tRNA synthetase-like proteins in non-ribosomal peptide
synthesis. Biol. Direct 5, 48.
(21) Belin, P., Moutiez, M., Lautru, S., Seguin, J., Pernodet, J. L., and
Gondry, M. (2012) The nonribosomal synthesis of diketopiperazines in
tRNA-dependent cyclodipeptide synthase pathways. Nat. Prod. Rep. 29,
961−979.
(22) Seguin, J., Moutiez, M., Li, Y., Belin, P., Lecoq, A., Fonvielle, M.,
Charbonnier, J. B., Pernodet, J. L., and Gondry, M. (2011)
Nonribosomal peptide synthesis in animals: The cyclodipeptide
synthase of Nematostella. Chem. Biol. 18, 1362−1368.
(23) Sievers, F., Wilm, A., Dineen, D., Gibson, T. J., Karplus, K., Li, W.,
Lopez, R., McWilliam, H., Remmert, M., Soding, J., Thompson, J. D.,
and Higgins, D. G. (2011) Fast, scalable generation of high-quality
protein multiple sequence alignments using Clustal Omega. Mol. Syst.
Biol. 7, 539.
(24) Goujon, M., McWilliam, H., Li, W., Valentin, F., Squizzato, S.,
Paern, J., and Lopez, R. (2010) A new bioinformatics analysis tools
framework at EMBL-EBI. Nucleic Acids Res. 38, W695−W699.
(25) Roy, A., Kucukural, A., and Zhang, Y. (2010) I-TASSER: A unified
platform for automated protein structure and function prediction. Nat.
Protoc. 5, 725−738.
(26) Zhang, Y. (2008) I-TASSER server for protein 3D structure
prediction. BMC Bioinf. 9, 40.
(27) Bonnefond, L., Arai, T., Sakaguchi, Y., Suzuki, T., Ishitani, R., and
Nureki, O. (2011) Structural basis for nonribosomal peptide synthesis
by an aminoacyl-tRNA synthetase paralog. Proc. Natl. Acad. Sci. U.S.A.
108, 3912−3917.
J
dx.doi.org/10.1021/bi4004827 | Biochemistry XXXX, XXX, XXX−XXX