Post-Translational Formation of Aminomalonate by a Promiscuous Peptide-Modifying Radical SAM Enzyme

Article abstract of DOI:10.1002/anie.202107192

Aminomalonate (Ama) is a widespread structural motif in Nature, whereas its biosynthetic route is only partially understood. In this study, we show that a radical S-adenosylmethionine (rSAM) enzyme involved in cyclophane biosynthesis exhibits remarkable catalytic promiscuity. This enzyme, named three-residue cyclophane forming enzyme (3-CyFE), mainly produces cyclophane in vivo, whereas it produces formylglycine (FGly) as a major product and barely produce cyclophane in vitro. Importantly, the enzyme can further oxidize FGly to produce Ama. Bioinformatic study revealed that 3-CyFEs have evolved from a common ancestor with anaerobic sulfatase maturases (anSMEs), and possess a similar set of catalytic residues with anSMEs. Remarkably, the enzyme does not need leader peptide for activity and is fully active on a truncated peptide containing only 5 amino acids of the core sequence. Our work discloses the first ribosomal path towards Ama formation, providing a possible hint for the rich occurrence of Ama in Nature.

Full text of DOI:10.1002/anie.202107192

Angewandte
Research Articles
Chemie
How to cite:
International Edition: doi.org/10.1002/anie.202107192
German Edition: doi.org/10.1002/ange.202107192
Biosynthesis Hot Paper
Post-Translational Formation of Aminomalonate by a Promiscuous
Peptide-Modifying Radical SAM Enzyme
+
+
Suze Ma , Heng Chen , He Li, Xinjian Ji, Zixin Deng, Wei Ding, and Qi Zhang*
Abstract: Aminomalonate (Ama) is a widespread structural
motif in Nature, whereas its biosynthetic route is only partially
understood. In this study, we show that a radical S-adenosyl-
methionine (rSAM) enzyme involved in cyclophane biosyn-
thesis exhibits remarkable catalytic promiscuity. This enzyme,
named three-residue cyclophane forming enzyme (3-CyFE),
mainly produces cyclophane in vivo, whereas it produces
formylglycine (FGly) as a major product and barely produce
cyclophane in vitro. Importantly, the enzyme can further
oxidize FGly to produce Ama. Bioinformatic study revealed
that 3-CyFEs have evolved from a common ancestor with
anaerobic sulfatase maturases (anSMEs), and possess a similar
set of catalytic residues with anSMEs. Remarkably, the enzyme
does not need leader peptide for activity and is fully active on
a truncated peptide containing only 5 amino acids of the core
sequence. Our work discloses the first ribosomal path towards
Ama formation, providing a possible hint for the rich
occurrence of Ama in Nature.
adenosylmethionine (rSAM) superfamily enzymes are exten-
sively involved in RiPP biosynthesis and catalyze strikingly
[16,17]
diverse reactions.
The rSAM-dependent RiPPs include
[
18–20]
sactipeptide (a-thioether linkage),
ranthipeptide (b or g-
streptide (CÀC crosslink between Lys
rotapeptide (C-O crosslink between Thr and
[
21–23]
thioether linkage),
[
24]
and Trp),
[
25]
[26]
Gln) and ryptide (CÀC crosslink between Arg and Tyr),
among others. In these reactions, the rSAM enzymes utilize
a strictly conserved [4Fe-4S] cluster to reductively cleave S-
adenosylmethionine (SAM), and the resulting 5’-deoxyade-
nosyl (dAdo) radical abstracts a hydrogen atom from the
[
27,28]
substrate to initiate subsequent oxidation reactions.
In
addition, the RiPP-modifying rSAM enzymes contain a char-
acteristic C-terminal extension called SPASM domain, which
accommodates one or more additional [4Fe-4S] clusters, with
[
29]
roles suggestive of peptide binding or electron transfer.
In silico analysis of novel rSAM-dependent RiPPs with
the aid of co-expression in E. coli proves to be fruitful in
exploring novel rSAM enzymes and their substrate specific-
[
30]
Introduction
ity. A recent study showed that a group of rSAM enzymes
catalyze the crosslinking between an aromatic residue (W1)
and a non-aromatic residue (X3) in a W1-X2-X3 motif to
Aminomalonate (Ama), the structural congener of aspar-
tate and glutamate, is a long known metabolite in Nature and
[31]
produce a cyclophane moiety. These enzymes are hence
defined as three-residue cyclophane-forming enzymes (3-
CyFEs). 3-CyFEs are abundant in diverse bacterial phyla and
can be further classified into different clades based on the
characteristic motif of substrate, such as Xye (TIGR04495),
[
1,2]
was proposed to be an intermediate in glycine biosynthesis.
Ama has been found in several natural products,
[
3–7]
and can
also serve as an extender unit in the biosynthesis of
[8–11]
polyketides and/or non-ribosomal peptides (Figure S1).
[
31]
Ama was also found as an important constituent in the
Grr (TIGR04260) and Fxs (TIGR04268). Xye is named for
its wide distribution in strains of the genera Xenorhabdus,
Yersinia and Erwinia. Grr system denotes that the precursor
peptide contains multiple glycine-rich repeats. Fxs contains
a highly conserved substrate tail F-X-S-X-X-COOH. These 3-
CyFE-derived RiPPs were designated as triceptide (three-
Escherichia coli (E. coli) protein hydrolysate, and is likely
[
12,13]
widespread in various proteomes.
modifications, such as a-carboxylation of glycine, and over-
Post-translational
[13]
oxidation of Ser, have been hypothesized to produce Ama
in a ribosomally synthesized polypeptide, but thus far none of
these feasibilities have been experimentally validated.
Ribosomally synthesized and post-translationally modi-
fied peptides (RiPPs) afford fascinating opportunities to
[
31]
residue in cyclophane peptides). In this study, we report
detailed in vitro characterization of a 3-CyFE involved in the
biosynthesis of a Fxs-type triceptide. We show that unexpect-
edly, the enzyme did not catalyze cyclophane formation in
vitro. Instead, it catalyzes oxidation of the Ser in the W1-X2-
S3 motif to produce formylglycine (FGly) and Ama. We
further show that 3-CyFEs have evolved from a common
ancestor with anaerobic sulfatase maturases (anSMEs), and
both groups of enzymes possess a similar set of catalytic
residues. In constrast to most RiPP biosynthetic enzymes, 3-
CyFE does not require leader peptide for activity.
[14,15]
discover novel biochemistries.
Particularly, the radical S-
[
+]
[+]
[
*] S. Ma, H. Chen, H. Li, Dr. X. Ji, Prof. Dr. Q. Zhang
Department of Chemistry, Fudan University
Shanghai, 200433 (China)
E-mail: qizhang@sioc.ac.cn
Prof. Dr. Z. Deng, Prof. Dr. W. Ding
State Key Laboratory of Microbial Metabolism, School of Life
Sciences & Biotechnology, Shanghai Jiao Tong University
Shanghai, 200240 (China)
+
[
] These authors contributed equally to this work.
Supporting information and the ORCID identification number(s) for
the author(s) of this article can be found under:
https://doi.org/10.1002/anie.202107192.
&
&&&
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Results and Discussion
terminal sequence FQSSI is characteristic of the FxS sub-
strate motif. The cluster also encodes a putative 3-CyFE SjiB.
We first set out to validate the enzyme activity by a co-
expression approach, similar to the recent report by Morinaka
The sji Biosynthetic Cluster
[
31]
RiPPs are derived from gene-encoded precursor peptides
consisting of an N-terminal leader and a C-terminal core. In
most cases, the leader peptide is recognized by a small
structural domain called RiPP recognition element (RRE),
which directs the interaction between substrate and the
et al. To this end, SjiA was co-expressed with the rSAM/
SPASM domain of SjiB (hereafter termed SjiB-399) in E. coli,
2
+
and the product was purified by Ni affinity chromatography.
Liquid chromatography with high-resolution mass spectrom-
etry (LC-HRMS) analysis showed the resulting SjiA is 2 Da
[
32,33]
8+
modifying enzymes.
RRE is indispensable for most RiPP-
less ([M + 8H] = 819.31) compared to the unmodified SjiA
[34–37]
8+
modifying rSAMs known thus far,
which can be found
obtained by expressing SjiA alone ([M + 8H] = 819.56)
either as an N-terminal extension domain or as a standalone
polypeptide encoded in the biosynthetic gene cluster (BGC)
(Figure 1B). The in vivo modified peptide was then digested
by trypsin and analyzed by high-resolution tandem MS (HR-
MS/MS), and the result clearly revealed that the tryptic
(
Table S2).
In an in silico analysis (Table S2), we occasionally noted
+
fragment SjiA51–60 ([M + H] = 1034.51) contains a crosslink
the RRE domain seems to be absent in tricepeptide
biosynthesis, as it is neither found in 3-CyFEs, nor encoded
in the corresponding BGCs. Intrigued by this observation, we
focused on a putative Fxs gene cluster (hereafter termed sji,
Figure 1A and Note S1) from Streptacidiphilus jiangxiensis,
between the conserved Phe and Ser as previously character-
ized (Figure 1C and Figure S2). We hereafter refer to the
crosslinked SjiA as 1, and the tryptic fragment was accord-
ingly termed SjiA51–60-1.
[39]
an actinomycete from acidic rhizosphere soil in China. This
gene cluster encodes a 60-aa precursor peptide SjiA, whose C-
In Vitro Activity of SjiB-399
After validating the activity of SjiB-399 in vivo, we
prepared SjiA and SjiB-399 separately for in vitro character-
ization. SjiB-399 was purified to near homogeneity under
strictly anaerobic condition, followed by chemical reconsti-
tution and gel filtration (Figure S9). The ultraviolet-visible
(UV/Vis) absorption spectrum of the chemically reconstitut-
ed SjiB-399 revealed a broad feature at 410 nm (Figure 2A),
2
+
suggestive of the presence of one or more [4Fe-4S] clusters.
Quantification analysis showed that each enzyme contains
7
.5 Æ 0.4 iron and 7.3 Æ 0.5 labile sulfide, suggesting it harbors
two [4Fe-4S] clusters. We next anaerobically incubated SjiB-
99 with SAM and sodium dithionite (DTH) in the absence of
3
SjiA. This assay revealed a time-dependent production of 5’-
deoxyadenosine (dAdoH) (Figure 2B and Figure S10), indi-
cating that SjiB-399 is indeed a rSAM protein.
To reconstitute the enzyme activity in vitro, we treated
SjiA with SjiB-399, SAM and DTH under strictly anaerobic
condition, and the reaction mixture was analyzed by LC-
HRMS. In this analysis, we observed a major product with the
8
+
expected À2 Da monoisotopic peak ([M + 8H] = 819.31),
and two minor products with isotopically mixed mass signals
8
+
at + 14 and + 16 Da (i.e. [M + 8H] = 821.31 and 821.56,
Figure 2C). Trypsin digestion ascertained the three products,
which exhibit À2, + 14, + 16 Da mass change, respectively
+
(
correspond to [M + H] = 1034.51, 1050.51, and 1052.52,
respectively) (Figure S12). Unexpectedly, extracted ion chro-
matography (EIC) showed that the À2 Da species produced
in vitro (hereafter referred to as 2) is apparently different
from 1 obtained in vivo, as the focused tryptic fragment of 2
(hereafter SjiA51–60-2) is eluted ꢀ 20 sec latter in LC-MS
analysis (Figure 2D). Careful re-examination showed that 2
was also produced in vivo, with a yield ꢀ 10-fold lower than
that of 1 (Figure 2D, Figure S11 and S17).
Figure 1. The sji gene cluster and its cyclophane product. (A) The sji
biosynthetic gene cluster encodes a precursor peptide SjiA and a FxsB-
type 3-CyFE SjiB. As the boundary between leader and core in SjiA is
currently unknown, we only showed the C-terminus of SjiA, with the
tryptic sequence highlighted in bold. See Note S1 for detailed
sequences. SjiB consists of an N-terminal SPASM family of rSAM
domain and a C-terminal HExxH domain; the latter domain was
[
31,38]
suggestive of metalloprotease activity.
(B) HR-MS characterization
of SijA modified by SjiB-399 in vivo (1). The loss of 2 Da is a result of
CÀC crosslink in cyclophane formation. (C) HR-MS/MS characteriza-
The apparent difference of 2 and 1 indicates 2 contains
a distinct modification. Comparative HR-MS/MS analysis of
tion of the tryptic sequence SjiA5 –61. The y ions in red correspond
1–60
to the À2 Da mass change resulting from the CÀC crosslink.
the two tryptic fragments showed that, in contrast to SjiA51–60
-
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Figure 2. In vitro characterization of SjiB-399 reveals the production of FGly and Ama-containing peptide. A. UV/Vis spectrum of the chemically
reconstituted SjiB-399. B. The extracted ion chromatograms (EICs) of SjiB-399-catalyzed dAdoH production. C. HR-MS analysis of SjiA peptide
treated with purified SjiB-399 in vitro. The full reaction was carried out with 50 mM SjiA, 5 mM reconstituted SjiB-399, 2 mM SAM, and 10 mM
DTH; the three control assays omit SAM, DTH, and SjiB-399, respectively. The full reaction catalyzed by SjiB-399 produce a major product with
À2 Da, and two minor products with +14 Da and +16 Da. D. EICs for the tryptic products obtained from (i) in vivo co-expression and (ii) in vitro
reaction, as well as NH OH-treatment of the (iii) in vivo and (iv) in vitro products. ox denotes the oxime derivative of 2 and 3. E. HR-MS/MS
2
dissection of the FGly-containing peptide 2. Note that FGly can be hydrated to yield the gem-diol counterpart 3. F. HR-MS/MS dissection of the
Ama-containing peptides 4, showing a CID-mediated decarboxylation-type of neutral loss and the corresponding fragmentation.
1
, SjiA51–60-2 exhibited a series of b and y ions within the to-
reduced or converted to the oxime-derivative, whereas the
SjiA51–60-1 remain unchanged (Figure 2D, Figure S6-S8).
These results confirmed FGly production in the SjiB-399-
catalyzed reaction.
be-cyclized FxS motif (Figure 2E and Figure S3), indicating
that the FxS motif was not cross-linked in 2. More impor-
tantly, the À2 Da modification was observed exclusively on
the Ser residue within the FxS motif (Figure 2E), which is
reminiscent of formylglycine (FGly), a structural motif can be
formed by anaerobic sulfatase maturation enzyme
rSAM-Dependent Ama Formation in SjiA
[40–42]
(
anSME).
This proposal is consistent with the observa-
+
+
tion of the + 16 Da (i.e. [M + H] = 1052.52) product,
a putative gem-diol (3) formed via hydration of 2. Indeed,
SjiA51-60-3 is eluted simultaneously with SjiA51–60-2 in the LC-
MS analysis (Figure 2D), and displayed exactly the same MS/
MS fragmentation pattern as that of SjiA51–60-2 (Figure S4).
To further verify the structure of 2 and 3, we treated the
tryptic products from both in vivo and in vitro studies with
sodium borohydride (NaBH ) and hydroxylamine (NH OH).
For the + 14 Da product (i.e. [M + H] = 1050.51, SjiA
5
1–
60-4), HR-MS/MS analysis revealed a set of b and y ions that
do not match the SjiA51–60 sequence. However, we noticed the
monoisotopic peak at m/z = 1006.52 is the most intensive
peak in the spectrum, which is 44 Da less compared to the
parent ion (i.e. m/z = 1050.52), suggesting a decarboxylation-
type neutral loss in MS/MS analysis (Figure 2F and Fig-
ure S5). This type of neutral loss has not been observed for
peptidyl moiety produced by common post-translational
4
2
In this analysis, SjiA51–60-2 and SjiA51–60-3 were completely
&
&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
[
43]
modifications,
containing amino acid.
60-4 contains a Ser-derived Ama, which was decarboxylated to
but is reminiscent of Ama, a malonate-
[
44,45]
We hence speculated that SjiA51–
produce a Gly during collision-induced dissociation (CID) in
HR-MS/MS analysis. Consistent with this proposal, the HR-
MS/MS spectrum of SjiA51–60-4 matches well to a Ser-to-Gly
+
mutant of SjiA51–60 (i.e. the ion with [M+H] = 1006.52), in
which the Ser residue in the FxS motif is changed to a Gly.
To validate the CID-induced decarboxylation of Ama, we
[
13]
obtained Ama according to a previous report
and per-
formed MS/MS analysis with the same instrumental setting.
This analysis clearly shows that Ama underwent decarbox-
ylation efficiently (Figure S13), which is consistent with the
fragmentation pattern observed for SjiA51–60-4. To unambig-
uously characterize the Ama in SjiA51–60-4, we carried out the
SjiB-399-catalyzed in vitro reaction in large scale, and the
product was subjected to alkaline hydrolysis, followed by
phenyl isothiocyanate (PITC) derivatization. LC-HRMS
analysis showed the expected product was co-eluted with
the authentic Ama derivative (Figure S14), confirming the
production of Ama in the assay.
We hypothesized the Ama is likely produced by over-
oxidation of the Ser residue by SjiB-399, which likely involves
the gem-diol intermediate 3 (Figure 2E). To test this proposal,
1
8
we ran the reaction in 70% H2 O, and the resulting product
was treated with trypsin and analyzed by LC-HRMS. This
1
8
analysis clearly indicated that up to two O atoms can be
incorporated into SjiA5 -4, showing two newly formed
1–60
+
+
species at [M + H] = 1052.51 and [M + H] = 1054.52,
respectively (Figure 3A). Selected ion monitoring (SIM)
followed by HR-MS/MS analysis of the two species showed
Figure 3. Enzymatic Ama formation in SjiA. A. HR-MS analysis of the
Ama product derived from in vitro enzymatic assays in H O or 70%
H2 O, showing up to two O atoms were incorporated into the Ama
product. B. Reaction Scheme of the SjiB-399-catalyzed Ama formation,
highlighting the oxidation proceeds in a solvent-exchangeable manner
(dashed box).
2
1
8
18
that the HR-MS/MS spectra display an invariant daughter ion
+
(
i.e. [M + H] = 1006.52) (Figure S15) as well as correspond-
1
8
ing b and y ions. The enriched O signal is present in the
parent ion but is absent in the daughter ion, which corrob-
1
8
orates the loss of O-incorporated CO during CID-induced
3-CyFEs are Phylogenetically Closely Related to anSMEs
2
decarboxylation. This analysis confirms that, after first
oxidation of Ser to FGly, SjiB-399 can further oxidize the
hydrated FGly (i.e. gem-diol) to produce Ama (Figure 3B).
The fact that SjiB-399 also serves as a robust serine
oxidase intrigued us to carry out extensive bioinformatic
analysis. Using SjiB-399 as a sequence query, we performed
five rounds of position specific iterated (PSI)-BLAST in
RefSeq database. The resulting homologs (n = 5000) were
1
8
Incorporation of two O atoms into Ama is apparently owing
to solvent exchange via the equilibrium of aldehyde and gem-
[
46]
diol, and the latter is oxidized to Ama.
[
47]
Re-examination of in vivo data revealed Ama was also
produced in the co-expression study, which is ꢀ 30-fold lower
than that of the cyclophane product (Figure S11 and S17). We
also carried out a set of in vivo and in vitro experiments in
different conditions. This analysis showed that although
increasing co-expression time did not significantly change
the ratio of the products, apparently increased yield of the
Ama product was observed with prolonged reaction time in
vitro; in this condition, the cross-linked cyclophane product
was still not observed (Figure S17). The factors that contrib-
ute to the distinct reaction outcomes between in vivo and in
vitro assays remain unknown and is under investigation in our
laboratory. Despite this, the data presented herein clearly
demonstrated that Ama can be produced by posttranslational
oxidation of Ser, shedding lights on the elusive origin of Ama
in Nature.
analyzed by RODEO
to retrieve Pfam and TIGRFAM
annotation of surrounding genes. Surprisingly, in addition to
ꢀ 3800 3-CyFEs that are encoded within the apparent
triceptide BGCs, we noticed ꢀ 1000 proteins belonging to
the anSME protein family (TIGR03942). Most of these
putative anSMEs (n = 801) are encoded next to a ꢀ 600 aa
protein annotated as “sulfatase” (PF00884) or “chol sulfa-
tase: choline-sulfatase” (TIGR03417) (Supporting Informa-
tion). These sulfatase-like sequences possess the character-
istic substrate motif (C/S)x(P/A)xR as described previously
[42]
(Supporting Information), suggesting that they are bona
fide anaerobic sulfatases that could be post-translationally
modified by the corresponding anSMEs. This survey indicates
3-CyFEs bear significant sequence homology to anSMEs
involved in anaerobic sulfatase maturation. We noted that the
rSAM enzymes involved in the biosynthesis of other RiPP
classes were not found in this BLASTp search.
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
To further gain insights into the phylogenetic relationship
between 3-CyFE and anSMEs, we constructed a rooted
maximum likelihood (ML) phylogenetic tree using selected
sequences from the dataset (Supporting Information). The
known rSAMs involved in the biosynthesis of ranthipeptide,
sactipeptide and shp/rgg-associated cyclic peptides were
manually incorporated into this analysis. The result showed
that anSMEs and 3-CyFEs fall into the same major clade in
the phylogenetic tree (Figure 4A). Sequence similarity net-
work (SSN) analysis verifies that 3-CyFEs are more related to
anSMEs, in stark contrast to other known RiPP-modifying
rSAMs that form CÀO, CÀC, CÀ-S crosslinks (Figure S18).
These analyses indicated that anSMEs and 3-CyFEs likely
have evolved from the same ancestor; both groups of enzymes
are phylogenetically distantly related to other known RiPP-
modifying rSAM enzymes.
Extensive efforts to crystalize SjiB-399 prove to be
[
42]
unsuccessful. However, we noted that, anSMEcpe,
an
anSME that modifies a Cys-type sulfatase, displays apparent
sequence homology to SjiB-399 (26% identity and 41%
similarity, Figure S19), and its crystal structure is available
(PDB: 4K39). We hence constructed a homology model of
SjiB-399, which consists of an N-terminal rSAM domain and
a C-terminal SPASM domain (Figure S20). C323, C329, C335
and C365 in the SPASM domain are highly conserved among
3
-CyFEs. These four Cys residues are likely involved in
coordinating an auxiliary [4Fe-4S] cluster (referred to as 3-
CyFE Aux cluster), which is equivalent to the AuxII cluster of
anSMEcpe. We noted that SjiB-399 and other 3-CyFEs lack
the other four Cys residues in anSMEcpe for ligating the AuxI
cluster, and this finding is consistent with the in vitro Fe/S
quantification analysis showing that SjiB-399 only has two
[
4Fe-4S] clusters (i.e. a rSAM cluster and an auxiliary cluster).
We respectively changed C19, C329 and C364 to Ala and co-
expressed each of these mutants with SjiA. The results
showed that the enzyme activity was completely abolished in
all the three mutants (Figure S16), indicating both [4Fe-4S]
clusters are strictly essential for enzyme activity.
Careful inspection of the putative catalytic site of SjiB-399
reveals that H72, H106 and D282 (Figure 5C and Figure S21)
appear to be equivalent to the Q64, Q98 and D277 found in
anSMEcpe (Figure 5B). In anSMEcpe, D277 likely serves as
a general base to deprotonate Cys thiol (Figure 5D), while
Q64 and Q98 are suggested to stabilize the catalytic D277,
[
42]
and participate in substrate binding and orientation. We
constructed the H72A, H106A, D282A and D282N mutants
and co-expressed each mutant with SjiA. The result (Fig-
ure 5E) clearly shows that both D282A and D282N are
completely inactive, demonstrating the strictly essential role
of Asp282 in catalysis (Figure S22). In contrast, H72A and
H106A are still able to produce both cyclophane and FGly
product, albeit with diminished activity, suggesting these two
His residue is not strictly essential. Further inspection into the
multiple sequence alignment of Fxs-type 3-CyFEs also
reveals a highly conserved Lys residue K285 (Figure S19
and Figure S21), but changing this residue to Ala only
partially decreased enzyme activity. Ama production was
not observed in all these analyses.
Figure 4. 3-CyFEs are close homologs of anSMEs. A. Maximum like-
lihood (ML) phylogenetic tree of 3-CyFEs, anSMEs and other known
RiPP-modifying rSAM enzymes. The tree is rooted using tryptophan
Substrate-Tuned Partition of Divergent Pathways in SjiB-
3
99 Catalysis
[
51]
lyase NosL. Key bootstrap values are indicated along the branch.
B. The active site of anSMEcpe (PDB: 4K39), showing the rSAM
cluster, SAM, catalytic residues and the substrate peptide with Cys
highlighted in sticks. C. The homology structure of SjiB-399 using
anSMEcpe as the structure template. Shown is the same view as that
in B. D. The proposed catalytic mechanism of anSMEcpe, in which the
conserved D277 serves as a base to deprotonate substrate Cys. Aux,
auxiliary [4Fe-4S] clusters. E. Site-directed mutagenesis and the result-
ing activity for SjiB-399. The activity is assessed by LC-HRMS and
quantified by the relative EIC area of all products versus the unmodi-
fied precursor peptide. See Figure S21C for the detailed ratio of each
product.
Given the remarkable catalytic promiscuity of SjiB-399,
we envisioned that proper variation of the substrate motif W1-
X2-S3 may alter the reaction paths (Figure 5A). To test this
hypothesis, we first conducted a thorough mining of the
putative substrates of the FxsB-type 3-CyFEs. By aligning to
the profile hidden Markov model (HMM) of FxsA
(
TIGR04268), we generated a comprehensive list of substrate
candidates containing ꢀ 1000 sequences (Supporting Infor-
mation), which can be classified into 23 major groups based
on SSN analysis (Figure S23-S24). Careful inspection of these
&
&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
gem-diol (Figure S29-S30), whereas no Ama product is
observed, indicating His at the W1 position is less favorable
in the sji system. In contrast, the F56Y mutant was fully
modified, affording ꢀ 5% Ama product and ꢀ 95% FGly and
gem-diol products (Figure S26-S28). The F56Y mutant was
also co-expressed with SjiB-399 D282N; however, no modi-
fication occurs, which further supports the indispensable role
of D282 in SjiB-399 catalysis (Figure S22).
We further changed Phe56 to Ala and conducted co-
expression analysis. No modification of the F56A mutant was
observed (Figure S44), indicating the aromatic W1 residue is
indispensable for enzyme processing. We also constructed the
S58C and S58T mutants, but no modification was observed for
these two mutants (Figure S44), indicating the sji system can
only specifically modify a Ser residue within the W1-X2-S3
motif.
Because aromatic residues are not found in the X2
position (Figure S25), we constructed a Q57F mutant to
interrogate whether the enzyme can accept aromatic residues
at this position. Moreover, the resulting mutant has a FFSS
motif in the C-terminus, which is actually an overlap of two
FxS motifs. Subsequent analysis showed that the Q57F
mutant was mostly converted to the FGly product, and the
cyclophane product is insignificant (Figure 5B, Figure S33-
S34). Notably, although two overlapped FxS motifs are
present in FFSS, HR-MS/MS analysis showed that modifica-
tion occurs exclusively on the first Ser residue (i.e. the native
substrate Ser58) (Figure S34), indicating the enzymatic pro-
cess is strictly site-specific. Similarly, the Q57W mutant
peptide is completely modified to the FGly product whereas
the cyclophane product is not observed (Figure 5B, Fig-
ure S35-S36). These results indicated that the big steric
aromatic residue at the X2 position can block the cyclo-
phane-forming pathway between W1 and S3, thereby only
allow for Ser oxidation. We also changed Gln57 to Gly, and
the Q57G mutant was found to be roughly equally partitioned
into both cyclophane and FGly pathways (Figure 5B, Fig-
ure S37-S38).
Figure 5. Substrate-tuned and leader-independent peptide modifica-
tion by SjiB-399. A. A general reaction Scheme showing that the
modification outcome of SjiB-399 is highly dependent on the substrate
motif W1-X2-S3. B. Product ratio of various SjiA precursor mutants co-
expressed with SjiB-399. Data are mean Æ s.d. of biological replicates
(
n=3). For product titer (mg per liter culture), see Table S3. Note the
formylglycine product was calculated incorporating its hydrated form
i.e. the gem-diol counterpart). C. The leader peptide-independent
We also investigated the modification in the double
mutants. These analyses showed that cyclophane formation
was fully abolished in the F56Y/Q57G mutant, which was
completely converted to the FGly/gem-diol product (Fig-
ure 5B, Figure S39-S40). Intriguingly, we observed signifi-
cantly enhanced production of Ama in the F56Y/Q57A
mutant, which accounts for more than 10% of conversion
(
reaction of SjiB-399, showing the co-expression analysis with various
truncated SjiA peptides. The minimal substrate that can be processed
SjiB-399 is FQSSI and the EIC traces for the three products are
indicated (the Ser subjected to modification is underlined). Note the
GGG in gray is a linker released from maltose binding protein (MBP)
tag by TEV protease digestion (Supporting Information).
(Figure 5B, Figure S41-43). On the contrary, Ama production
sequence patterns revealed that the conserved Phe in the FxS
motif can be substituted with Trp, Tyr and His, while the
substrate Ser can be substituted with Asn and Thr (Fig-
ure S25). The X2 residue is more divergent, including Asn,
Asp, Gln, and Glu, although other residues such as Ala, Val
are also found.
is barely observed for the F56Y/Q57G mutant.
SjiA Modification by SjiB-399 is Leader Peptide-Independent
Because anSMEs are active on short oligopeptides
[
40]
We next changed the Phe in SjiA to its natural substitutes
derived from the native substrate,
the close phylogeny
(
i.e. Trp, Tyr and His), and the mutants were each co-
between 3-CyFEs and anSMEs suggests that 3-CyFEs may
also do not require leader peptide for activity. To test this
hypothesis, we constructed a series of SjiA truncation mutants
by removing the 15, 25, 30, 40, 45, 49 and 55 amino acids at the
N-terminus, respectively, which were then co-expressed with
SjiB-399. Remarkably, we found that all the truncated
expressed with SjiB-399 in E. coli. Interestingly, although for
F56W, the cyclophane product remained predominant (Fig-
ure S27-S28), the F56Y and F56H mutants were all switched
to the Ser oxidation pathways (Figure 5B, Figure S26-S32). In
F56H, only ꢀ 50% of substrate was converted to FGly and
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
mutants were fully modified by the enzyme, even for the
pentapeptide FQSSI, which lacks the N-terminal 55 aa
[2] H. Nagayama, M. Muramatsu, K. Shimura, Nature 1958, 181,
17 – 418.
4
[
3] I. Ciciliato, E. Corti, E. Sarubbi, S. Stefanelli, L. Gastaldo, N.
(
Figure 5C, Figure S45). These results are consistent with
Montanini, M. Kurz, D. Losi, F. Marinelli, E. Selva, J. Antibiot.
observation that the 23 groups of precursor peptides mined in
this study lack the overall similarity in the N-terminal region
2
004, 57, 210 – 217.
4] A. W. Sangster, S. E. Thomas, N. L. Tingling, Tetrahedron 1975,
1, 1135 – 1137.
5] H. B. Park, J. M. Crawford, J. Nat. Prod. 2015, 78, 1437 – 1441.
[
[
(
Figure S19-S20). Importantly, these results are consistent
3
with the fact that no RRE was found in relation to 3-CyFEs,
which are potentially interesting for future endeavors in
protein engineering and peptide modification.
[6] J. G. Batelaan, J. W. F. K. Barnick, J. L. van der Baan, F. Bick-
elhaupt, Tetrahedron Lett. 1972, 13, 3107 – 3110.
[
7] B. J. C. Law, Y. Zhuo, M. Winn, D. Francis, Y. X. Zhang, M.
Samborskyy, A. Murphy, L. J. Ren, P. F. Leadlay, J. Micklefield,
Nat. Catal. 2018, 1, 977 – 984.
Conclusion
[
8] A. O. Brachmann, C. Garcie, V. Wu, P. Martin, R. Ueoka, E.
Oswald, J. Piel, Chem. Commun. 2015, 51, 13138 – 13141.
9] Y. A. Chan, M. G. Thomas, Biochemistry 2010, 49, 3667 – 3677.
In this study, we showed that a FxS-type 3-CyFE exhibits
remarkable catalytic promiscuity, which not only produces the
cyclized cyclophane product but also carries out Ser oxidation
to produce FGly and Ama. Catalytic promiscuity has recently
been reported for several rSAM proteins, and in most cases,
the promiscuous activities are associated with unnatural
[
[
10] Y. A. Chan, M. T. Boyne 2nd, A. M. Podevels, A. K. Klimowicz,
J. Handelsman, N. L. Kelleher, M. G. Thomas, Proc. Natl. Acad.
Sci. USA 2006, 103, 14349 – 14354.
[11] Y. A. Chan, A. M. Podevels, B. M. Kevany, M. G. Thomas, Nat.
Prod. Rep. 2009, 26, 90 – 114.
[12] J. J. Van Buskirk, W. M. Kirsch, D. L. Kleyer, R. M. Barkley,
T. H. Koch, Proc. Natl. Acad. Sci. USA 1984, 81, 722 – 725.
13] S. D. Copley, E. Frank, W. M. Kirsch, T. H. Koch, Anal.
Biochem. 1992, 201, 152 – 157.
[
48–56]
[57–59]
substrates.
The tryptophan lyase NosL
and the
[
60]
anaerobic coproporphyrinogen III oxidase HemN
are
[
exceptional examples in which promiscuous activity were
observed with native substrates. By characterization of the
Ser oxidation activity of SjiB-399, our study adds a new
example to these exceptions. These findings highlight the
remarkable catalytic plasticity of the rSAM enzymes and the
great potential to engineer these catalysts for novel activities.
Our study also represents the first report of Ama
production in a ribosomal peptide. It has long been proposed
that Ama could be produced by post-translational modifica-
[
14] P. G. Arnison, M. J. Bibb, G. Bierbaum, A. A. Bowers, T. S.
Bugni, G. Bulaj, J. A. Camarero, D. J. Campopiano, G. L. Challis,
J. Clardy, P. D. Cotter, D. J. Craik, M. Dawson, E. Dittmann, S.
Donadio, P. C. Dorrestein, K. D. Entian, M. A. Fischbach, J. S.
Garavelli, U. Goransson, C. W. Gruber, D. H. Haft, T. K.
Hemscheidt, C. Hertweck, C. Hill, A. R. Horswill, M. Jaspars,
W. L. Kelly, J. P. Klinman, O. P. Kuipers, A. J. Link, W. Liu,
M. A. Marahiel, D. A. Mitchell, G. N. Moll, B. S. Moore, R.
Muller, S. K. Nair, I. F. Nes, G. E. Norris, B. M. Olivera, H.
Onaka, M. L. Patchett, J. Piel, M. J. Reaney, S. Rebuffat, R. P.
Ross, H. G. Sahl, E. W. Schmidt, M. E. Selsted, K. Severinov, B.
Shen, K. Sivonen, L. Smith, T. Stein, R. D. Sussmuth, J. R. Tagg,
G. L. Tang, A. W. Truman, J. C. Vederas, C. T. Walsh, J. D.
Walton, S. C. Wenzel, J. M. Willey, W. A. van der Donk, Nat.
Prod. Rep. 2013, 30, 108 – 160.
15] M. Montalbꢀn-Lꢁpez, T. A. Scott, S. Ramesh, I. R. Rahman,
A. J. van Heel, J. H. Viel, V. Bandarian, E. Dittmann, O.
Genilloud, Y. Goto, M. J. Grande Burgos, C. Hill, S. Kim, J.
Koehnke, J. A. Latham, A. J. Link, B. Martꢂnez, S. K. Nair, Y.
Nicolet, S. Rebuffat, H. G. Sahl, D. Sareen, E. W. Schmidt, L.
Schmitt, K. Severinov, R. D. Sꢃssmuth, A. W. Truman, H. Wang,
J. K. Weng, G. P. van Wezel, Q. Zhang, J. Zhong, J. Piel, D. A.
Mitchell, O. P. Kuipers, W. A. van der Donk, Nat. Prod. Rep.
[
13]
tion,
but this hypothesis remains only speculative. Our
study validates the possibility of the post-translationally
produced Ama and sheds lights on the biosynthetic origin of
this widespread structural moiety in biological systems. Our
work also demonstrates the rich chemistry in RiPP biosyn-
thesis and the fascinating opportunity to expand chemical
diversity by engineering RiPP biosynthetic pathways.
[
Acknowledgements
This work is supported in part by grants from the National
Key Research and Development Program (2016 Y F
A0501302 and 2018Y F A0900402), from National Natural
Science Foundation of China (21822703, 21921003, and
2020, 38, 130 – 239.
[
[
16] K. Yokoyama, E. A. Lilla, Nat. Prod. Rep. 2018, 35, 660 – 694.
17] N. Mahanta, G. A. Hudson, D. A. Mitchell, Biochemistry 2017,
3
2070050), and from Innovative research team of high-level
56, 5229 – 5244.
local universities in Shanghai.
[
18] T. Mo, X. Ji, W. Yuan, D. Mandalapu, F. Wang, Y. Zhong, F. Li,
Q. Chen, W. Ding, Z. Deng, S. Yu, Q. Zhang, Angew. Chem. Int.
Ed. 2019, 58, 18793 – 18797; Angew. Chem. 2019, 131, 18969 –
18973.
19] L. B. Bushin, B. C. Covington, B. E. Rued, M. J. Federle, M. R.
Seyedsayamdost, J. Am. Chem. Soc. 2020, 142, 16265 – 16275.
20] S. Chiumento, C. Roblin, S. Kieffer-Jaquinod, S. Tachon, C.
Lepretre, C. Basset, D. Aditiyarini, H. Olleik, C. Nicoletti, O.
Bornet, O. Iranzo, M. Maresca, R. Hardre, M. Fons, T. Giardina,
E. Devillard, F. Guerlesquin, Y. Coute, M. Atta, J. Perrier, M.
Lafond, V. Duarte, Sci. Adv. 2019, 5, eaaw9969.
Conflict of Interest
[
The authors declare no conflict of interest.
[
Keywords: anaerobic sulfatase maturase · gem-diol · oxidation ·
posttranslational modification · RiPP
[
21] G. A. Hudson, B. J. Burkhart, A. J. DiCaprio, C. J. Schwalen, B.
Kille, T. V. Pogorelov, D. A. Mitchell, J. Am. Chem. Soc. 2019,
141, 8228 – 8238.
[
1] K. Shimura, H. Nagayama, A. Kikuchi, Nature 1956, 177, 935 –
36.
9
&
&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!

Angewandte
Research Articles
Chemie
[
[
[
[
[
[
[
[
22] Y. Chen, Y. Yang, X. Ji, R. Zhao, G. Li, Y. Gu, A. Shi, W. Jiang,
[44] J. W. Thanassi, Biochemistry 1970, 9, 525 – 532.
[45] P. Wheelan, W. M. Kirsch, T. H. Koch, J. Org. Chem. 1989, 54,
1364 – 1370.
[46] J. Wang, S. Ma, W. Ding, T. Chen, Q. Zhang, Chin. J. Chem. 2021,
39, 1923 – 1926.
[47] J. I. Tietz, C. J. Schwalen, P. S. Patel, T. Maxson, P. M. Blair, H. C.
Tai, U. I. Zakai, D. A. Mitchell, Nat. Chem. Biol. 2017, 13, 470 –
478.
[48] W. Ding, X. Ji, Y. Zhong, K. Xu, Q. Zhang, Curr. Opin. Chem.
Biol. 2020, 55, 86 – 95.
[49] N. A. Bruender, T. A. Grell, D. P. Dowling, R. M. McCarty, C. L.
Drennan, V. Bandarian, J. Am. Chem. Soc. 2017, 139, 1912 –
1920.
[50] A. Silakov, T. L. Grove, M. I. Radle, M. R. Bauerle, M. T. Green,
A. C. Rosenzweig, A. K. Boal, S. J. Booker, J. Am. Chem. Soc.
2014, 136, 8221 – 8228.
Q. Zhang, Biotechnol. J. 2020, e2000136.
23] A. Caruso, L. B. Bushin, K. A. Clark, R. J. Martinie, M. R.
Seyedsayamdost, J. Am. Chem. Soc. 2019, 141, 990 – 997.
24] K. R. Schramma, L. B. Bushin, M. R. Seyedsayamdost, Nat.
Chem. 2015, 7, 431 – 437.
25] K. A. Clark, L. B. Bushin, M. R. Seyedsayamdost, J. Am. Chem.
Soc. 2019, 141, 10610 – 10615.
26] A. Caruso, R. J. Martinie, L. B. Bushin, M. R. Seyedsayamdost,
J. Am. Chem. Soc. 2019, 141, 16610 – 16614.
27] H. J. Sofia, G. Chen, B. G. Hetzler, J. F. Reyes-Spindola, N. E.
Miller, Nucleic Acids Res. 2001, 29, 1097 – 1106.
28] W. E. Broderick, B. M. Hoffman, J. B. Broderick, Acc. Chem.
Res. 2018, 51, 2611 – 2619.
29] T. A. Grell, P. J. Goldman, C. L. Drennan, J. Biol. Chem. 2015,
290, 3964 – 3971.
[
[
30] D. H. Haft, M. K. Basu, J. Bacteriol. 2011, 193, 2745 – 2755.
31] T. Q. N. Nguyen, Y. W. Tooh, R. Sugiyama, T. P. D. Nguyen, M.
Purushothaman, L. C. Leow, K. Hanif, R. H. S. Yong, I. Agatha,
F. R. Winnerdy, M. Gugger, A. T. Phan, B. I. Morinaka, Nat.
Chem. 2020, 12, 1042—1053.
[51] M. Dong, M. Horitani, B. Dzikovski, M. E. Pandelia, C. Krebs,
J. H. Freed, B. M. Hoffman, H. Lin, J. Am. Chem. Soc. 2016, 138,
9755 – 9758.
[52] M. Dong, M. Horitani, B. Dzikovski, J. H. Freed, S. E. Ealick,
B. M. Hoffman, H. Lin, J. Am. Chem. Soc. 2017, 139, 5680 – 5683.
[53] S. Joshi, N. Mahanta, D. Fedoseyenko, H. Williams, T. P. Begley,
J. Am. Chem. Soc. 2017, 139, 10952 – 10955.
[54] X. Ji, D. Mandalapu, J. Cheng, W. Ding, Q. Zhang, Angew.
Chem. Int. Ed. 2018, 57, 6601 – 6604; Angew. Chem. 2018, 130,
6711 – 6714.
[55] S. Ghosh, E. N. G. Marsh, J. Biol. Chem. 2020, 295, 11513 –
11528.
[
[
[
[
32] B. J. Burkhart, G. A. Hudson, K. L. Dunbar, D. A. Mitchell, Nat.
Chem. Biol. 2015, 11, 564 – 570.
33] A. M. Kloosterman, K. E. Shelton, G. P. van Wezel, M. H.
Medema, D. A. Mitchell, mSystems 2020, 5, e00267-20.
34] T. W. Precord, N. Mahanta, D. A. Mitchell, ACS Chem. Biol.
2019, 14, 1981 – 1989.
35] E. L. Morinaka, M. Verest, M. J. Helf, T. Scalvenzi, A. L.
Vagstad, J. Sims, S. Sunagawa, M. Gugger, J. Piel, Science 2018,
[56] D. Y. Yin, X. Ji, Q. Zhang, Chin. J. Chem. 2021, https://doi.org/
10.1002/cjoc.202100304.
359, 779 – 782.
[
[
36] L. B. Bushin, K. A. Clark, I. Pelczer, M. R. Seyedsayamdost, J.
Am. Chem. Soc. 2018, 140, 17674 – 17684.
37] J. A. Latham, A. T. Iavarone, I. Barr, P. V. Juthani, J. P. Klinman,
J. Biol. Chem. 2015, 290, 12908 – 12918.
[57] D. M. Bhandari, D. Fedoseyenko, T. P. Begley, Methods Enzy-
mol. 2018, 606, 155 – 178.
[58] W. Ding, X. Ji, Y. Li, Q. Zhang, Front. Chem. 2016, 4, 27.
[59] G. Sicoli, J. M. Mouesca, L. Zeppieri, P. Amara, L. Martin, A. L.
Barra, J. C. Fontecilla-Camps, S. Gambarelli, Y. Nicolet, Science
2016, 351, 1320 – 1323.
[60] X. Ji, T. Mo, W. Q. Liu, W. Ding, Z. Deng, Q. Zhang, Angew.
Chem. Int. Ed. 2019, 58, 6235 – 6238; Angew. Chem. 2019, 131,
6301 – 6304.
[
38] N. M. Hooper, FEBS Lett. 1994, 354, 1 – 6.
[
39] Y. Huang, Q. Cui, L. Wang, C. Rodriguez, E. Quintana, M.
Goodfellow, Z. Liu, Antonie Van Leeuwenhoek 2004, 86, 159 –
165.
[
[
[
40] T. L. Grove, K. H. Lee, J. St Clair, C. Krebs, S. J. Booker,
Biochemistry 2008, 47, 7523 – 7538.
41] A. Benjdia, S. Subramanian, J. Leprince, H. Vaudry, M. K.
Johnson, O. Berteau, FEBS J. 2010, 277, 1906 – 1920.
42] P. J. Goldman, T. L. Grove, L. A. Sites, M. I. McLaughlin, S. J.
Booker, C. L. Drennan, Proc. Natl. Acad. Sci. USA 2013, 110,
Manuscript received: May 29, 2021
8
519 – 8524.
Revised manuscript received: June 23, 2021
Accepted manuscript online: June 24, 2021
[
43] T. B. Moyer, N. C. Parsley, P. W. Sadecki, W. J. Schug, L. M.
Hicks, Nat. Prod. Rep. 2021, 38, 489 – 509.
Version of record online: && &&
,
&&&&
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
ꢀ 2021 Wiley-VCH GmbH
www.angewandte.org
&&&&
These are not the final page numbers!

Angewandte
Research Articles
Chemie
Research Articles
Biosynthesis
We report biochemical study of a radical
S-adenosylmethionine (rSAM) enzyme
involved in the biosynthesis of a riboso-
mally synthesized and post-translation-
ally modified peptide (RiPP). This rSAM
enzyme exhibits remarkable catalytic
promiscuity, which produces cyclophane
in vivo but formylglycine and amino-
malonate (Ama) in vitro. Unlike most
enzymes involved in RiPP biosynthesis,
this enzyme does not require leader
peptide for activity.
S. Ma, H. Chen, H. Li, X. Ji, Z. Deng,
W. Ding, Q. Zhang*
&&&& — &&&&
Post-Translational Formation of
Aminomalonate by a Promiscuous
Peptide-Modifying Radical SAM Enzyme
&
&&&
www.angewandte.org
ꢀ 2021 Wiley-VCH GmbH
Angew. Chem. Int. Ed. 2021, 60, 2 – 10
These are not the final page numbers!