Metabolic coupling of two small-molecule thiols programs the biosynthesis of lincomycin A

Article abstract of DOI:10.1038/nature14137

Low-molecular-mass thiols in organisms are well known for their redox-relevant role in protection against various endogenous and exogenous stresses. In eukaryotes and Gram-negative bacteria, the primary thiol is glutathione (GSH), a cysteinyl-containing tripeptide. In contrast, mycothiol (MSH), a cysteinyl pseudo-disaccharide, is dominant in Gram-positive actinobacteria, including antibiotic-producing actinomycetes and pathogenic mycobacteria. MSH is equivalent to GSH, either as a cofactor or as a substrate, in numerous biochemical processes, most of which have not been characterized, largely due to the dearth of information concerning MSH-dependent proteins. Actinomycetes are able to produce another thiol, ergothioneine (EGT), a histidine betaine derivative that is widely assimilated by plants and animals for variable physiological activities. The involvement of EGT in enzymatic reactions, however, lacks any precedent. Here we report that the unprecedented coupling of two bacterial thiols, MSH and EGT, has a constructive role in the biosynthesis of lincomycin A, a sulfur-containing lincosamide (C8 sugar) antibiotic that has been widely used for half a century to treat Gram-positive bacterial infections. EGT acts as a carrier to template the molecular assembly, and MSH is the sulfur donor for lincomycin maturation after thiol exchange. These thiols function through two unusual S-glycosylations that program lincosamide transfer, activation and modification, providing the first paradigm for EGT-associated biochemical processes and for the poorly understood MSH-dependent biotransformations, a newly described model that is potentially common in the incorporation of sulfur, an element essential for life and ubiquitous in living systems.

Full text of DOI:10.1038/nature14137

LETTER
doi:10.1038/nature14137
Metabolic coupling of two small-molecule thiols
programs the biosynthesis of lincomycin A
Qunfei Zhao1*, Min Wang1*, Dongxiao Xu1, Qinglin Zhang2 & Wen Liu1,2
Low-molecular-mass thiols in organisms are well known for their
redox-relevant role in protection against various endogenous and
exogenous stresses1–3. In eukaryotes and Gram-negative bacteria, the
primary thiol is glutathione (GSH), a cysteinyl-containing tripep-
tide. Incontrast,mycothiol(MSH),acysteinylpseudo-disaccharide,
is dominant in Gram-positive actinobacteria, including antibiotic-
producing actinomycetes and pathogenic mycobacteria. MSH is equi-
valent to GSH, either as a cofactor or as a substrate, in numerous
biochemical processes4, most of which have not been characterized,
largely due to thedearth of information concerning MSH-dependent
proteins. Actinomycetesareabletoproduceanother thiol, ergothio-
neine (EGT), a histidine betaine derivative that is widely assimilated
by plants and animals for variable physiological activities5. The involve-
ment of EGT in enzymatic reactions, however, lacks any precedent.
Here we report that the unprecedented coupling of two bacterial
thiols, MSH and EGT, has a constructive role in the biosynthesis of
lincomycinA, asulfur-containinglincosamide(C8sugar)antibiotic
that has been widely used for half a century to treat Gram-positive
bacterial infections6–9. EGT acts as a carrier to template the molecu-
lar assembly, and MSH is the sulfur donor for lincomycin matura-
tion after thiol exchange. These thiols function through two unusual
S-glycosylations that program lincosamide transfer, activation and
modification, providing the first paradigm for EGT-associated bio-
chemical processes and for the poorly understood MSH-dependent
biotransformations, a newly described model that is potentially com-
mon in the incorporation of sulfur, an element essential for life and
ubiquitous in living systems.
As anticipated, the DlmbE mutant strain accumulated a MSH-associated
lincomycin analogue, 1 (Fig. 2). In this molecule, the C8-sugar unit is
appended to MSH via an a-S-linkage (Supplementary Text), identical
to that of lincomycin A in configuration. The DlmbE mutant strain still
produced lincomycin A, albeit in a lower yield, indicating that an addi-
tional mca orthologue is present outside the lmb cluster and partially
a
OH
SH
O
O
HO
HO
SH
O
N
NH
O
HN
H
OH
OH
O
HOOC
N
COOH
HO
N
OH
OH
H
HS
N
NH2
NHAc
COOH
GSH
MSH
EGT
b
OH
O
OH
OH
HO
HO
O
O
AcHN
HO
HO
HO
OUDP
HO
AcHN
H2N
OPO3H2
OH
OH
OH
MshA
O
O
+
HO
HO
OH
OH
OH
OH
UDP
OPO3H2
OH
HO
HO
GlcN-Ins
GlcNAc-Ins-3-P
OH
OH
c
N
NH2
MSH
N
H
Electrophilic toxin
OH
COOH
O
HO
HO
EgtD
HN
O
OH
OH
HO
N
OH
OH
Toxin
S
O
N
Mycothiol (MSH, Fig. 1a) mediated detoxification typically relies on
the conjugation of MSH to an electrophilic toxin. An amidase, Mca,
then hydrolyses the resulting MSH S-conjugate to produce a pseudo-
disaccharide unit, 1-O-glucosamine-D-myo-inositol (GlcN-Ins), and a
mercapturic acid derivative, which is an excretive N-acetyl-cysteinyl pro-
duct common in GSH-mediated metabolism10–12 (Fig. 1b). In actino-
mycetes, mcaorthologues have beenfound in several biosynthetic gene
clustersofantibiotics4,11, includingthatoflincomycinA, suggesting that
MSH S-conjugation is associated with the production of these anti-
biotics. Lincomycin A consists of an N-methylated 4-propyl-L-proline
(PPL)moietyandlincosamide, anunusualeight-carbonsugardecorated
withamethylmercaptogroupatC-1(Fig.1d). Cellprotectionagainstthe
activity of lincomycin A depends largely on methylation of the bacterial
ribosome, whereby the molecule mimics the 39 end of (de)acetyl-tRNA
NHAc
N
H
MSH S-conjugate
COOH
Mca
OH
Toxin
S
O
NHAc
EGT
Mercapturic acid
d
Me
N
Lincosamide
8
Me
Me
8
N
Me
Me
OH
OMe
N
OH
OH
O
O
H
N
O
O
H
1
HO
2
HO
N-methyl-
4-propyl-L-proline
O
HO
1
2
HO
S
O
SMe
Methylmercapto group
Lincomycin A
HO
Celesticetin
and blocks protein synthesis at the initial stage of the elongation cycle13,14
.
Figure 1
|
Representative low-molecular-mass thiols, relevant metabolic
This fact, in combination with the methylmercapto group found in the
structure, leads to the proposal that MSH S-conjugation has a con-
structive role in lincomycin biosynthesis by supplying sulfur rather
than a protective role in antibiotic detoxification. To validate this hypo-
thesis, we inactivated lmbE, a mca orthologue that is located in the
lincomycin biosynthetic gene cluster (lmb) in Streptomyces lincolnen-
sis15,16 (Extended Data Fig. 2a).
pathways, and associated lincosamide natural products. a, Structures of the
thiols GSH, MSH and EGT. b, Biosynthetic pathway of MSH and its typical
associated detoxification process (shown in the dashed rectangle). GlcN-Ins is
recyclable as an intermediate or product. c, Biosynthetic pathway of EGT.
EgtD is an S-adenosyl methionine-dependent protein that catalyses the first
reaction. d, Structures of the lincosamide antibiotics lincomycin A and
celesticetin.
1State Key Laboratory of Bioorganic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 345 Lingling Road, Shanghai 200032, China. 2Huzhou Center
of Bio-Synthetic Innovation, 1366 Hongfeng Road, Huzhou 313000, China.
*These authors contributed equally to this work.
0 0 M O N T H 2 0 1 5
| V O L 0 0 0 | N A T U R E | 1
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
a
after a 5 day cultivation period, representing a ,4-fold increase in pro-
duct concentration compared with the precursor. This increase could
result from the regeneration of intermediate 1 by restoring the MSH
cycle, becauseGlcN-Ins,the productoriginatingfrom1through LmbE-
catalysed hydrolysis, is the intermediate in MSH biosynthesis4,11 (Fig. 1b).
The cell homogenate of the DmshAlin mutant strain was capable of
transforming ,0.20 mM 2 into a nearly equal amount of lincomycin A
(Fig. 2b), further confirming the essentiality of the LmbE-catalysed re-
action in the lincomycin biosynthetic pathway, which hydrolyses 1 to
generate recyclable GlcN-Ins and intermediate 2. Processing 2, includ-
ing C–S bond cleavage of the cysteinyl group (mechanistically similar
to thatin the GSH-associatedbiosynthesis of gliotoxin17)and subsequent
S-methylation of the exposed sulfhydryl group, may eventually afford
the sulfur appendage of lincomycin A. Notably, the generation of ex-
cretivethiomethylproductshaspreviouslybeenfoundasanalternative
to the mercapturate pathway for xenobiotic detoxification in GSH-
mediated metabolism18.
Lincomycin A
1
i
ΔlmbE
ESI m/z [M+H]+ modes for HPLC-MS:
831.0
342.0
1
407.0
508.0
Lincomycin A
ii
ΔlmbE-E3457
ΔmshAlin
2
GlcN-Ins
iii
iv
b
Lincomycin A
ΔmshAlin + 1
2
i
4h
1h
0h
ii
v
Wild type
iii
6
8
10 12 14 16 18 min
2
4
6
8
10 12 14 16 18 20 min
We next focused on how and when MSH is incorporated to generate
S-conjugate 1. Sequence analysis of the lmb cluster revealed two closely
linked, functionally unassigned genes (Extended Data Fig. 2a), lmbT
and lmbV. lmbT encodes aproteinbelonging tothe GTase superfamily.
Recently, lincosamide formation has proven to involve the generation
of GDP-D-a-D-octose and its associated modifications19–21. The phos-
phonucleotidyl group on the resulting product, GDP-D-a-D-lincosamide
(3) (Fig. 3), facilitatestheattackbynucleophilessuchasMSH, probably
requiring a GTase activity; however, direct transfer of lincosamide from
GDP onto MSH appears unlikely, because GTase-catalysed glycosyla-
tion is often envisioned proceeding through SN2 nucleophilic displace-
ment, which would not explain the same a-linkage that is predicted in
3 and found in intermediate 1. lmbV encodes a protein classified into
the DinB-2 superfamily (Extended Data Fig. 5), which now includes
over ten thousand members with activities presumably related to vari-
ous low-molecular-mass thiols22. These proteins share a conserved DinB-
2-like domain combined with variable functional domain(s), indicating
that thiols potentially act in different biochemical processes. Notably, a
cladephylogeneticallyrelevanttoLmbVcontainstheMSH-maleylpyr-
uvate isomerase (Extended Data Fig. 5), one ofthe few MSH-dependent
proteins that have been biochemically characterized for the isomeriza-
tion of maleylpyruvate to fumarylpyruate23,24. Accordingly, LmbV was
proposed to catalyse a MSH-dependent reaction, although the role of
MSH remained unknown.
We next established the relevance of LmbT and LmbV to lincomycin
biosynthesis, as the inactivation of lmbT or lmbV completely abolished
lincomycinproduction, whichwaspartiallyrestoredbycomplementing
each of the genes in the corresponding mutant strain (Fig. 3a). Surpris-
ingly, alincomycin analogue, EGT S-conjugate 4,wasfoundintheDlmbV
mutant (Fig. 3 and Supplementary Text). Re-examination of the DmshAlin
mutantstrainshowedasimilarproductprofileinwhich4wasthemajor
product(Fig.3a);evidently, EGTS-conjugationisindependentofMSH.
To correlate this thiol with lincomycin production, we surveyed the
S. lincolnensisgenomeandidentifiedtheEGTbiosyntheticgenes25 (Ex-
tended Data Fig. 2c), in which egtDlin, responsible for triple N-methylation
of histidine in EGT biosynthesis (Fig. 1c), was chosen for inactivation.
The DegtDlin mutant strain produced lincomycin A, the yield of which,
however, was significantly reduced at the indicated time points (Fig. 3b).
Analysisofthemutantcells revealedatraceamountof EGT (withayield
,16.8% of that in the wild-type cells, Extended Data Fig. 4), which
probably resulted from the exogenous assimilation of the culture med-
iumthat includedplant-derived, EGT-containingcomponents4,5. Over
time, the production of lincomycin A from the mutant became much
lower thanthat fromthewild-type strain (Fig. 3b), presumably because
of the limiting amount of EGT. Altogether, EGT is necessary for lin-
comycin biosynthesis, whereby its S-conjugate 4 is an intermediate
instead of a shunt product in the pathway.
c
1
H
N
OH
Me
Me
HO
HO
– LmbE
OH
O
OH
N
O
O
H
HO
HN
O
HO
HO
S
O
NHAc
OH
OH
HO
OH
1
2
Hydrolysis
GlcN-Ins
+ LmbE
H
N
Me
Me
GlcN-Ins
OH
OH
O
N
O
H
HO
O
HO
S
OH
NHAc
2
1
2
3
4
5 min
Figure 2
|
Characterization of LmbE as a pathway-specific Mca protein to
process 1, the MSH S-conjugated intermediate, in the biosynthesis of
lincomycin A. For high-performance liquid chromatography-mass
spectrometry (HPLC-MS) analysis, the electrospray ionization (ESI) m/z
[M 1 H]1 modes are indicated in the dashed rectangle. a, In vivo product
profiles of S. lincolnensis strains, including the mutants (i, for DlmbE; ii, for
DlmbE-E3457; iii, for DmshAlin; and iv, for DmshAlin supplemented with
,0.20 mM intermediate 1) and the wild-type control (v). b, Time-dependent
biotransformation of mercapturic acid 2 (,0.20 mM) into lincomycin A using
the cell homogenate of the DmshAlin mutant strain. c, In vitro hydrolysis of
intermediate 1 to generate 2 and GlcN-Ins (left) in the absence (top right)
and presence (lower right) of LmbE.
compensates for the loss of lmbE. Sequencing the genome of S. lincolnensis
revealed three lmbE homologues: lmbE80, lmbE447 and lmbE3457. The
individual inactivation of these genes was performed in the DlmbE mutant
strain, but only the DlmbE-E3457 double mutant strain completely lost
lincomycin producibility and had a concomitant increase in 1 (Fig. 2a
and Extended Data Fig. 3). Moreover, we expressed and purified LmbE
from Escherichia coli. This recombinant protein rapidly converted 1
into two products (Fig. 2c): GlcN-Ins and 2, a mercapturic acid deriv-
ative (Supplementary Text). Thus, the involvement of LmbE as a specific
Mca protein in lincomycin biosynthesis was established.
MSH biosynthesis begins with the glycosyltransferase (GTase) MshA,
which catalyses the formation of 1-O-(2-N-acetyl)-glucosamine-D-myo-
inositol-3-phosphate (GlcNAc-Ins-3-P) to affordthe essential pseudo-
disaccharide unit12 (Fig. 1b). To validate the necessity of MSH for
lincomycin biosynthesis, we identified a mshA orthologue (Extended
Data Fig. 2b), mshAlin, from the S. lincolnensis genome and inactivated
it in the wild-type strain. The DmshAlin mutant strain failed to produce
MSH along with lincomycin A (Fig. 2a and Extended DataFig. 4). MSH
S-conjugate 1 was then added to this mutant, leading to the restoration
oflincomycinproduction(Fig. 2a). Unambiguously, 1 isa keyinterme-
diaterather thanadetoxifiedantibiotic in thebiosynthetic pathway. No-
Notably, in EGT S-conjugate 4, lincosamide is attached to the thiol
tably, adding ,0.20 mM 1 to the cells yielded ,1.08 mM lincomycin A via a b-S-linkage (Fig. 3 and Supplementary Text). If 1 is the product of
2
| N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
a
Despite our numerous attempts, LmbV was highly refractory to various
methods of expression. We then selected CcbV, a homologue of LmbV
(57% identity) in celesticetin biosynthesis26, as the catalyst. Celesticetin
is a naturally occurring analogue of lincomycin A that possesses the
same a-S-linkage (Fig. 1d). Heterologous expression of the gene ccbV
in the DlmbV mutant strain partially restored the production of linco-
mycin A (Fig. 3a), confirming that CcbV functionally substitutes for
LmbV. The in vitro conversion of 4 occurred in the presence of CcbV,
resulting in the production of 1 along with the release of EGT (Fig. 3c
and Extended Data Fig. 6). This reaction is irreversible, as MSH S-conjugate
1 was not converted back to EGT S-conjugated 4 in the presence of
EGT. Therefore, LmbV and its counterpart CcbV form a new type of
GTasethat catalyses a unique S-glycosylation forthiol exchangebetween
EGT and MSH.
We thus considered that the other GTase, LmbT, could transfer lin-
cosamine from GDP onto EGT via SN2 nucleophilic displacement. LmbT
was expressed and purified, and to determine its activity, the reverse
glycosylation reaction was conducted according to a method established
previously using saturated GDP as the nucleophilic agent27. Because
LmbT did not convert EGT S-conjugate 4 to the corresponding GDP-
mediated sugar product (Extended Data Fig. 7a), we re-examined the
DlmbT mutant strain and found that the PPL moiety had accumulated
extensively (Extended Data Fig. 8a). This finding suggested that the
S-glycosylationofEGT precedesPPL incorporation. Consistently, inac-
tivation of lmbC, which encodes an adenylation protein characterized
in PPL activation28, resulted in the production of PPL and an EGT
S-conjugated lincosamine, 5 (Extended Data Fig. 8a and Supplemen-
tary Text). Compared with 4, 5 lacks the PPL moiety. A similar result
was found in the DlmbN and DlmbD mutant S. lincolnensis strains (Ex-
tended Data Fig. 8a), supporting the corresponding proteins, LmbC,
LmbN and LmbD, being responsible for incorporating PPL to trans-
form5into4. LmbNisabi-functionalprotein, andits1,2-isomerization
activity has recently been shown in lincosamine formation20. Careful
analysis of the protein sequence revealed a peptidylcarrier protein (PCP)
domain present at its amino terminus. LmbC is thus considered to
activate PPL with ATP and transfer it onto this PCP domain, followed
by LmbD-catalysed condensation with EGT S-conjugate 5 to afford 4
(Extended Data Fig. 8b). Consequently, LmbC, LmbN (or LmbN-PCP,
the N-terminal PCP domain) and CcbD (the homologue in celesticelin
biosynthesis26 thatisfunctionallyidenticaltoLmbD)wereexpressedand
purified, and in vitro assays showed that these proteins indeed coord-
inate PPL attachment to generate 4 (Extended Data Fig. 8). Clearly,
functionalization of the lincosamine unit by PPL occurs in an EGT
S-conjugated form, and the PPL-lacking compound 5 probably serves
as the product of LmbT-catalysed S-glycosylation.
4
Lincomycin A
i
ΔlmbV
ESI m/z [M+H]+ modes for HPLC-MS:
407.0
LincomycinA 831.0
1
ii
ΔlmbT
487.0
649.0
443.5
MSH
3
230.0
574.0
435.0
EGT
4
iii
iv
ΔmshAlin
ΔegtDlin
GDP
5
b
100
Wild type
v
ΔlmbV+PermE*-lmbV
ΔlmbV+PermE*-ccbV
Δ
egtDlin
vi
50
0
ΔlmbT+PermE*-lmbT
vii
6
1
2
3
4
5
6
7
d
8
10 12 14 16 18 min
c
H
N
Me
Me
4
OH
OH
N
O
O
H
H
N
S
HO
– CcbV
+ CcbV
OH
N
N
4
MSH
HOOC
MSH
Thiol exchange
1
EGT
H
OH
4
N
Me
Me
HO
HO
EGT
2
OH
O
OH
O
N
H
O
HO
HN
O
HO
HO
S
O
OH
OH
HO
NHAc
OH
1
1
3
4
5 min
5
d
Me
OH
OH
O
– LmbT
H2N
HO
GDP
N
+
H
S
N
OH
N
GDP
5
HOOC
Reverse glycosylation
+ LmbT
EGT
Me
OH
3
OH
H2N
HO
O
EGT
+
HO
OGDP
3
2
3
4 min
As anticipated, in the presence of LmbT and the co-substrate GDP,
5 was efficiently transformed to GDP-D-a-D-lincosamide 3, accompan-
ied by EGT release (Fig. 3d and Extended Data Fig. 9). We then vali-
dated the forward activity of LmbT to produce 5 and GDP using the
substrates 3 and EGT (Extended Data Fig. 7b). The LmbT-catalysed
reaction exhibited an equilibrium constant Keq of 1.94 (Extended Data
Fig. 9d), indicating that the reverse and forward conversions are com-
parable; however, the activities of downstream enzymes may drive the
pathwayforwardtoproducelincomycinA. Consequently, LmbTrepre-
sentsanewenzymethatemploys EGTasasugar acceptor and catalyses
S-glycosylation with the naturally rare C8 sugar to generate EGT
S-conjugate 5.
The DlmbC, DlmbN or DlmbD mutant strain produced a minor pro-
duct, 6 (Extended Data Fig. 8a), which is a MSH S-conjugated linco-
samide lacking the PPL moiety (Supplementary Text), suggesting that
LmbV tolerated 5 as a substrate for thiol exchange. CcbV, the homo-
logue of LmbV, converted 5 into 6 in vitro (Extended Data Fig. 10a);
Figure 3
|
Characterization of MSH- and EGT-associated
biotransformations. For HPLC-MS analysis, the ESI m/z [M 1 H]1 modes
areindicatedin the dashedrectangle. a, In vivo product profilesofS. lincolnensis
strains, including the mutants DlmbV (i), DlmbT (ii), DmshAlin (iii) and
DegtDlin (iv); the DlmbV mutant harbouring homologous lmbV (v) or
heterologous ccbV (vi); and the DlmbT mutant harbouring homologous
lmbT (vii). PermE* is the constitutive promoter for expressing the
erythromycin-resistance gene in Saccharopolyspora erythraea. b, Time-course
analysis of lincomycin production in the wild-type (white) and DegtDlin mutant
(black) strains. Mean 6 s.e.m. (n 5 3) are shown. c, In vitro thiol exchange
for converting EGT S-conjugate 4 into MSH S-conjugate 1 (left) in the absence
(top right) and presence of CcbV (lower right). d, In vitro reverse glycosylation
for converting EGT S-conjugate 5 into GDP-mediated C8 sugar 3 (left) in
the absence (top right) and presence of LmbT (lower right).
LmbV, MSH isprobably the agent thatperformsthe nucleophilic attack
at C-1 of the lincosamide unit, which is activated by EGT rather than
(d)NDP, and the resulting inversion of configuration is consistent with
the glycosylation reaction that proceeds via an SN2 displacement to however, the rate of this reaction was much lower than that for the
generate the a-S-linkage in MSH S-conjugate 1. Therefore, we chem- transformation of 4 into 1. The Mca protein LmbE was not active on 6
ically synthesized MSH (Supplementary Methods), making use of the andfailedtoproducemercapturicacidforfurtherprocessing(Extended
scaled-up precursor GlcN-Ins, for an in vitro assay of LmbV activity. DataFig. 10b). Clearly, lincomycinbiosynthesisinvolvestheEGT-mediated
0 0 M O N T H 2 0 1 5
| V O L 0 0 0 | N A T U R E | 3
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
H
N
OH
Me
Me
Me
Me
LmbC,
OH
OH
OH
OH
OH
OH
OH
O
HO
HO
HO
N
H
H2N
H2N
O
O
O
O
H
N
H
N
LmbT
N and D
S
S
HO
HO
HO
HO
HO
OH
OH
N
N
OGDP
OGDP
EGT GDP
N
N
PPL
4
3
5
HOOC
HOOC
Ribose-5-P + fructose-6-P or sedoheptulose-7-P
L-Tyr
Me
Recycle
H
N
H
N
OH
O
C–S bond cleavage
Me
Me
Me
LmbE
HO
HO
OH
OH
S-methylation
OH
OH
O
N
H
N
......
EGT
O
O
O
H
LmbV
MSH
Lincomycin A
HO
HO
HN
O
O
HO
HO
S
HO
S
O
OH
OH
1
2
GlcN-Ins
HO
OH
NHAc
OH
NHAc
Regeneration
MSH
Figure 4
|
The EGT (pink) andMSH (blue) programmed biosynthetic pathway of lincomycin A. The recycling of EGT andthe regenerationof MSHareshown
as dashed lines.
3. Newton, G. L. et al. Bacillithiol is an antioxidant thiol produced in Bacilli. Nature
Chem. Biol. 5, 625–627 (2009).
4. Rawat, M. & Av-Gay, Y. Mycothiol-dependent proteins in actinomycetes. FEMS
Microbiol. Rev. 31, 278–292 (2007).
5. Cheah, I. K. & Halliwell, B. Ergothioneine; antioxidant potential, physiological
function and role in disease. Biochim. Biophys. Acta 1822, 784–793 (2012).
6. Mason, D. J., Dietz, A. & DeBoer, C. Lincomycin, a new antibiotic. I. Discovery and
biological properties. Antimicrob. Agents Chemother. 554–559 (1962).
assembly of lincosamide with the building blocks, PPL and MSH, and
the MSH-associated post-modifications in a highly ordered process.
We uncovered a constructive role of two different bacterial thiols in
S. lincolnensis and clarified the biosynthetic pathway of lincomycin A,
in which EGT and MSH-associated metabolism creates an extraord-
inary biochemical strategy for the formation of active molecule (Fig. 4
and Extended Data Fig. 1). Lincomycin biosynthesis involves a con- 7. Herr, R. R. & Bergy, M. E. Lincomycin, a new antibiotic. II. Isolation and
characterization. Antimicrob. Agents Chemother. 560–554 (1962).
8. Lewis, C., Clapp, H. W. & Grady, J. E. In vitro and in vivo evaluation of lincomycin, a
vergent pathway to synthesize the L-Tyr-derived PPL moiety and the
GDP-activated C8sugarlincosamide. The thiolEGTservesas acarrier,
new antibiotic. Antimicrob. Agents Chemother. 570–582 (1963).
via the first S-glycosylation (reversible), to channel the lincosamine unit
and mediates its condensation with PPL. The thiol MSH goes through
the second S-glycosylation (irreversible) to associate with the lincomy-
cin intermediate resulting from EGT and then acts as the sulfur donor
for affording the methylmercapto group. Both thiols are recyclable or
reproducible, thus maintaining the biosynthetic route to lincomycin A.
The characterization of the low-molecular-mass thiol-programmed
biosynthetic pathway largely expands our knowledge regarding the
intrinsic, versatile functions of thiols, which are apparently not limited
to a protective role against oxidative stress and the neutralization of
electrophilic toxins. The lincosamide antibiotic celesticetin probably
shares this biosynthetic strategy based on the thiols EGT and MSH,
despite their differences in processing the MSH appendage26. Theinvolve-
ment of EGT in C8-sugar transfer and activation represents the first
biochemical evidence of this thiol in enzymatic reactions and generates
interest in nucleotide-independent sugar modifications and associated
glycosylations, which have been less appreciated to date29. Sulfur is one
of the most abundant elements in living organisms and contributesto a
large number of biologically active natural products; however, incorp-
oration of this atom has not been well established. Complementing the
recentadvanceofco-optingthesulfur-deliverysystemofprimarymeta-
bolism for thiosugar formation30, we demonstrated that MSH serves as
a different source for sulfur incorporation (Fig. 4). This could be a
general paradigm because the biosynthetic pathways of several sulfur-
containing natural products involve homologues of Mca, the protein
responsible for processing the MSH S-conjugate4,11. The findings reported
here represent a key step towards elucidating the biochemical mechan-
isms of numerous MSH and EGT-dependent but poorly understood
proteins and exploring new features of thiols with regard to their cur-
rently unknown associated biochemical processes.
9. Mason, D. J. & Lewis, C. Biological activity of the lincomycin related antibiotics.
Antimicrob. Agents Chemother. 7–12 (1964).
10. Newton, G. L., Buchmeier, N. & Fahey, R. C. Biosynthesis and functions of
mycothiol, the unique protective thiol ofActinobacteria. Microbiol. Mol. Biol. Rev. 72,
471–494 (2008).
11. Jothivasan, V. K. & Hamilton, C. J. Mycothiol: synthesis, biosynthesis and biological
functions of the major low molecular weight thiol in actinomycetes. Nature Prod.
Rep. 25, 1091–1117 (2008).
12. Fan, F., Vetting, M. W., Frantom, P. A. & Blanchard, J. S. Structures andmechanisms
of the mycothiol biosynthetic enzymes. Curr. Opin. Chem. Biol. 13, 451–459
(2009).
13. Reusser, F. Effect of lincomycin and clindamycin on peptide chain initiation.
Antimicrob. Agents Chemother. 7, 32–37 (1975).
14. Wilson, D. N. Ribosome-targeting antibiotics and mechanisms of bacterial
resistance. Nature Rev. Microbiol. 12, 35–48 (2014).
15. Peschke, U., Schmidt, H., Zhang, H.-Z. & Piepersberg, W. Molecular
characterization of the lincomycin-production gene cluster of Streptomyces
lincolnensis 78–11. Mol. Microbiol. 16, 1137–1156 (1995).
ˇ
´
16. Koberska, M. et al. Sequence analysis and heterologous expression of the
lincomycin biosynthetic cluster of the type strain Streptomyces lincolnensis ATCC
25466. Folia Microbiol. 53, 395–401 (2008).
17. Scharf, D. H. et al. Epidithiol formation by an unprecedented twin carbon-sulfur
lyase in the gliotoxin. Angew. Chem. Int. Ed. 124, 10211–10215 (2012).
18. Cooper, A. J. L. et al. Cysteine S-conjugate b-lyases: important roles in the
metabolism of naturally occurring sulfur and selenium-containing compounds,
exnobiotics, and anticancer agents. Amino Acids 41, 7–27 (2011).
19. Brahme, N. M. et al. Biosynthesis of the lincomycins. 2. Studies using stable
isotopes on the biosynthesis of methylthiolincosaminide moiety of lincomycin A.
J. Am. Chem. Soc. 106, 7878–7883 (1984).
20. Sasaki, E., Lin, C.-I., Lin, K.-Y. & Liu, H.-w. Construction of the octose 8-phosphate
intermediate in lincomycin A biosynthesis: characterization of the reactions
catalyzed by LmbR and LmbN. J. Am. Chem. Soc. 134, 17432–17435 (2012).
21. Lin, C.-I., Sasaki, E., Zhong, A. & Liu, H.-w. In vitro characterization of LmbK and
LmbO: identification of GDP-D-erythro-a-D-gluco-octose as a key intermediate in
lincomycin A biosynthesis. J. Am. Chem. Soc. 136, 906–909 (2014).
22. Newton, G. I., Leung, S. S., Wakabayashi, J. I., Rawat, M. & Fahey, R. C. The DinB
superfamily includes novel mycothiol, bacillithiol, and glutathione S-transferases.
Biochemistry 50, 10751–10760 (2011).
23. Feng, J. et al. The gene ncgl2918 encodes a novel maleylpyruvate isomerase that
needs mycothiol as cofactor and links mycothiol biosynthesis and gentisate
assimilation in Corynebacterium glutamicum. J. Biol. Chem. 281, 10778–10785
(2006).
Online Content Methods, along with any additional Extended Data display items
andSource Data, are available in the onlineversion of the paper; references unique
to these sections appear only in the online paper.
24. Wang, R. et al. Crystal structures and site-directed mutagenesis of a mycothiol-
dependent enzyme reveal a novel folding and molecular basis for mycothiol-
mediated maleylpyruvate isomerization. J. Biol. Chem. 282, 16288–16294 (2007).
25. Seebeck, F. P. In vitro reconstitution of mycobacterial ergothioneine biosynthesis.
J. Am. Chem. Soc. 132, 6632–6633 (2010).
Received 18 July; accepted 3 December 2014.
Published online 14 January 2015.
ˇ
´
26. Cermak, L. et al. Hybridization analysis and mapping of the celesticetin gene
cluster revealed genes shared with lincomycin biosynthesis. Folia Microbiol. 52,
457–462 (2007).
1. Fahey, R. C. Glutathione analogs in prokaryotes. Biochim. Biophys. Acta 1830,
3182–3198 (2013).
2. Deponte, M. Glutathione catalysis and the reaction mechanisms of glutathione-
dependent enzymes. Biochim. Biophys. Acta 1830, 3217–3266 (2013).
27. Zhang, C. et al. Exploiting the reversibility of natural product glycosyltransferase-
catalyzed reactions. Science 313, 1291–1294 (2006).
4
| N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 5
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
ˇ´
28. Kadlcık, S. et al. Adaptation of an
L
-proline adenylation domain to use 4-propyl-
L-
Author Contributions Q.Zhao, M.W. andQ.Zhangperformed in vivo investigations. M.W.
and D.X. prepared and characterized the chemical compounds. Q.Zhao and W.L.
conducted sequence analysis. Q.Zhao and M.W. performed in vitro enzymatic
investigations. Q.Zhao, M.W. and W.L. analysed the data and wrote the manuscript. W.L.
directed the research.
proline in the evolution of lincosamide biosynthesis. PLoS ONE 8, e84902 (2013).
29. Gantt, R. W., Peltier-Pain, P. & Thorson, J. S. Enzymatic methods for glyco
(diversification/randomization) of drugs and small molecules. Natural Prod. Rep.
28, 1811–1853 (2011).
30. Sasaki, E. et al. Co-opting sulphur-carrier proteins from primary metabolic
pathways for 2-thiosugar biosynthesis. Nature 510, 427–431 (2014).
Author Information The sequences of the genes lmbE80, lmbE447, lmbE3457, mshAlin
and egtDlin are deposited in GenBank with the NCBI accession numbers KJ958528,
KJ958529, KJ958530, KJ958531 and KJ958532, respectively. Reprints and
permissions information is available at www.nature.com/reprints. The authors declare
no competing financial interests. Readers are welcome to comment on the online
version of the paper. Correspondence and requests for materials should be addressed
to W.L. (wliu@mail.sioc.ac.cn).
Supplementary Information is available in the online version of the paper.
Acknowledgements This work was supported in part by grants from NSFC
(81302674, 31430005, 91213303, 21472231 and 91413101), STCSM
(14JC1407700 and 13XD1404500) and MST (2012AA02A706) of China.
0 0 M O N T H 2 0 1 5
| V O L 0 0 0 | N A T U R E | 5
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
obtained using primers lmbE3457-R-for and lmbE3457-R-rev was digested by
XbaI and HindIII and cloned into the same site of pLL1007, yielding the recom-
binant plasmid pLL1008. Following the procedure described above, pLL1008 was
introduced into LL1001 for double-crossover recombination, yielding the recom-
binant strain LL1004, in which the 402-bp internal fragment of lmbE3457 was deleted
in frame. Primers lmbE3457-gt-for and lmbE3457-gt-rev were used for genotype
validation by PCR amplification. Further sequencing of this PCR product con-
firmed the genotype of LL1004, in which lmbE3457 was also in-frame deleted.
For mshAlin deletion, the 2.25-kb fragment obtained using primers mshA-L-for
and mshA-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1009. The 2.18-kb fragment obtained using primers
mshA-R-for and mshA-R-rev was digested by XbaI and HindIII and cloned into
the same site of pLL1009 to yield the recombinant plasmid pLL1010, in which a
633-bp in-frame coding region of mshAlin was deleted. Following the procedure
described above, pLL1010 was introduced into the S. lincolnensis wild-type strain
for double-crossover recombination. The resulting strain LL1005 was then sub-
jected to PCR amplificationto examinethe genotype, as judged by the formationof
the desired 1.09-kb product when using primers mshA-gt-for and mshA-gt-rev.
Further sequencing of this PCR product confirmed the genotype of LL1005, in
which mshAlin was in-frame deleted.
For lmbT deletion, the 2.56-kb fragment obtained using primers lmbT-L-for
and lmbT-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1011. The 2.56-kb fragment obtained using primers
lmbT-R-for andlmbT-R-rev was digested by XbaI and HindIII andclonedinto the
same site ofpLL1011toyieldthe recombinant plasmidpLL1012, inwhicha462-bp
in-frame coding region of lmbT was deleted. Following the procedure described
above, pLL1012 was introduced into the S. lincolnensis wild-type strain for double-
crossover recombination. The resulting strain LL1006 was then subjected to PCR
amplification to examine the genotype, as judged by the formation of the desired
1.04-kb product when using primers lmbT-gt-for and lmbT-gt-rev. Further sequen-
cing ofthis PCR product confirmedthe genotype ofLL1006, inwhichlmbT wasin-
frame deleted.
For lmbV deletion, the 2.29-kb fragment obtained using primers lmbV-L-for
and lmbV-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1013. The 2.30-kb fragment obtained using primers
lmbV-R-for and lmbV-R-rev was digested by XbaI and HindIII and cloned into
the same site of pLL1013 to yield the recombinant plasmid pLL1014, in which a
399-bp in-frame coding region of lmbV was deleted. Following the procedure
described above, pLL1014 was introduced into the S. lincolnensis wild-type strain
for double-crossover recombination. The resulting strain LL1007 was then sub-
jected to PCR amplificationto examinethe genotype, as judged by the formationof
the desired 1.23-kb product when using primers lmbV-gt-for and lmbV-gt-rev.
Further sequencing of this PCR product confirmed the genotype of LL1007, in
which lmbV was in-frame deleted.
For egtDlin deletion, the 1.98-kb fragment obtained using primers egtD-L-for
and edtD-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1019. The 1.75-kb fragment obtained using primers
egtD-R-for and egtD-R-rev was digested by XbaI and HindIII and cloned into the
same site ofpLL1019toyieldthe recombinant plasmidpLL1020, inwhicha465-bp
in-frame coding region of egtDlin was deleted. Following the procedure described
above, pLL1020 was introduced into the S. lincolnensis wild-type strain for double-
crossover recombination. The resulting strain LL1010 was then subjected to PCR
amplification to examine the genotype, as judged by the formation of the desired
2.05-kb product when usingprimers egtD-gt-for and egtD-gt-rev. Further sequen-
cing of this PCR product confirmed the genotype of LL1010, in which egtDlin was
in-frame deleted.
METHODS
General materials and methods. Biochemicals and media were purchased from
Sinopharm Chemical Reagent Co., Ltd (China), Oxoid Ltd (UK) or Sigma-Aldrich
Corporation (USA) unless otherwise stated. Restriction endonucleases were pur-
chased from Thermo Fisher Scientific Co. Ltd (USA). Chemical reagents were pur-
chased from standard commercial sources.
DNAisolationandmanipulationinEscherichia coli orStreptomycesstrainswere
carried out according to standard methods31,32. PCR amplifications were carried
out on an Applied Biosystems Veriti Thermal Cycler using either Taq DNA poly-
merase (Vazyme Biotech Co. Ltd, China) for routine genotype verification or Phanta
Max Super-Fidelity DNA Polymerase (Vazyme Biotech Co. Ltd, China) for high
fidelityamplification. PrimersynthesiswasperformedatShanghaiSangonBiotech
Co. Ltd (China), and DNAsequencing was performed at Shanghai Majorbio Biotech
Co. Ltd or Shenzhen BGI in China.
High performance liquid chromatography (HPLC) analysis was carried out on
the Agilent 1200 HPLC system (Agilent Technologies Inc., USA). HPLC Electrospray
ionization MS (HPLC-ESI-MS)analysis was performed on the ThermoFisher LTQ
FleetESI-MS spectrometer (ThermoFisher Scientific Inc., USA), andthe datawere
analysed using Thermo Xcalibur software. ESI-high resolution MS (ESI-HR-MS)
analysis was carried out on the 6230B Accurate-Mass TOF LC/MS System or 6530
Accurate-Mass Q-TOF LC/MS System (Agilent Technologies Inc., USA) and the
data were analysed using Agilent MassHunter Qualitative Analysis software. NMR
data were recorded on the Bruker DRX400 and Bruker AV500 spectrometers (Bruker
Co. Ltd, Germany), or onthe Agilent 500 MHz PremiumCompact1 NMR spectro-
meter (Agilent Technologies Inc., USA).
No statistical methods were used to predetermine sample size.
Gene inactivation and complementation. The inactivation of each gene in S.
lincolnensis was performed by in-frame deletion to exclude polar effects on down-
stream gene expression. For complementation in trans, the target gene was under
the control of PermE*, the constitutive promoter for expressing the erythromycin-
resistance gene in Saccharopolyspora erythraea. The genomic DNA of the S. lin-
colnensis wild-type strain served as the template for PCR amplification unless
otherwise stated.
For lmbE deletion, the 2.52-kb fragment obtained using primers lmbE-L-for
and lmbE-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1001. The 2.34-kb fragment obtained using primers
lmbE-R-for and lmbE-R-rev was digested by XbaI and HindIII and cloned into the
same site ofpLL1001toyieldthe recombinant plasmidpLL1002, inwhicha549-bp
in-framecoding region of lmbEwas deleted. To transfer pLL1002 into the lincomycin-
producing strain S. lincolnensis, conjugation between E. coli ET12567-Streptomyces
was carried out following the standard procedure1. The colonies that were apra-
mycin resistant at 37 uC were identified as integrating mutants, in which a single-
crossover homologous recombination event took place. These mutants were cultured
for several rounds in the absence of apramycin, and the resulting apramycin-sensitive
isolates were subjected to PCRamplification to examine the genotype, as judged by
the formation of the desired 1.1-kb product when using primers lmbE-gt-for and
lmbE-gt-rev. Further sequencing of this PCR product confirmed the genotype of
LL1001, in which lmbE was in-frame deleted.
For lmbE-E80 double deletion, the 1.84-kb fragment obtained using primers
lmbE80-L-for and lmbE80-L-rev was digested by EcoRI and XbaI, and cloned into
the same sites of pKC1139 to yield plasmid pLL1003. The 2.06-kb fragment obtained
using primers lmbE80-R-for and lmbE80-R-rev was digested by XbaI and HindIII
and cloned into the same site of pLL1003, yielding the recombinant plasmid pLL1004.
Following the procedure described above, pLL1004 was introduced into LL1001 for
double-crossover recombination, yielding the recombinant strain LL1002, in which
the 447-bp internal fragment of lmbE80 was deleted in frame. Primers lmbE80-gt-
for and lmbE80-gt-rev were used for genotype validation by PCR amplification.
Further sequencing of this PCR product confirmed the genotype of LL1002, in
which lmbE80 was also in-frame deleted.
For lmbE-E447 double deletion, the 1.81-kb fragment obtained using primers
lmbE447-L-for and lmbE447-L-rev was digested by EcoRI and XbaI, and cloned
into the same sites of pKC1139 to yield plasmid pLL1005. The 1.82-kb fragment
obtained using primers lmbE447-R-for and lmbE447-R-rev was digested by XbaI
and HindIII and cloned into the same site of pLL1005, yielding the recombinant
plasmid pLL1006. Following the procedure described above, pLL1006 was intro-
duced into LL1001 for double-crossover recombination, yielding the recombinant
strain LL1003, in which the 459-bp internal fragment of lmbE447 was deleted in
frame. Primers lmbE447-gt-for and lmbE447-gt-rev were used for genotype valid-
ation by PCR amplification. Further sequencing of this PCR product confirmed
the genotype of LL1003, in which lmbE447 was also in-frame deleted.
For lmbC deletion, the 1.69-kb fragment obtained by using primers lmbC-L-for
and lmbC-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
pKC1139 to yield plasmid pLL1023. The 1.82-kb fragment obtained by using pri-
mers lmbC-R-for and lmbC-R-rev was digested by XbaI and HindIII and cloned
into the same site of pLL1023 to yield the recombinant plasmid pLL1024, in which
a 708-bp in-frame coding region of lmbC was deleted. Following the procedure
described above, pLL1024 was introduced into the S. lincolnensis wild-type strain
for double-crossover recombination. The resulting strain LL1012 was then sub-
jected to PCR amplificationto examinethe genotype, as judged by the formationof
the desired 0.82-kb product when using primers lmbC-gt-for and lmbC-gt-rev.
Further sequencing of this PCR product confirmed the genotype of LL1012, in
which lmbC was in-frame deleted.
For lmbD deletion, the 2.54-kb fragment obtained using primers lmbD-L-for
and lmbD-L-rev was digested by EcoRI and XbaI, and cloned into the same sites of
For lmbE-E3457 double deletion, the 1.79-kb fragment obtained using primers pKC1139 to yield plasmid pLL1031. The 2.51-kb fragment obtained using primers
lmbE3457-L-for and lmbE3457-L-rev was digested by EcoRI and XbaI, and cloned lmbD-R-for and lmbD-R-rev was digested by XbaI and HindIII and cloned into
into the same sites of pKC1139 to yield plasmid pLL1007. The 1.89-kb fragment the same site of pLL1031 to yield the recombinant plasmid pLL1032, in which a
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
459-bp in-frame coding region of lmbD was deleted. Following the procedure period as follows: t 50min, 10%B;t5 9 min, 10% B; t517 min, 60% B; t 518 min,
described above, pLL1032 was introduced into the S. lincolnensis wild-type strain 60%B; t 5 20 min, 10%B; andt 5 22 min, 10%B (mAU at 210nM). The data were
for double-crossover recombination. The resulting strain LL1016 was then sub- analysed using Thermo Xcalibur software. The concentration of lincomycin A was
jected to PCR amplificationto examinethe genotype, as judged by the formationof estimated by HPLC using the commercially available product as the standard.
the desired 0.60-kb product when using primers lmbD-gt-for and lmbD-gt-rev.
Further sequencing of this PCR product confirmed the genotype of LL1016, in
which lmbD was in-frame deleted.
For biotransformation of 2 (the mercapturic acid derivative) into loncomycin A
by the cell homogenate of the DmshAlin mutant strain, LL1005, the mycelia from
50 ml of the 6-day fermentation culture broth was collected, washed twice, and
Forsite-specificmutationoflmbN, the 2.01-kbfragment obtainedusingprimers then re-suspended in 10 ml of 50 mM Tris-HCl buffer (pH 7.5). Sonication for
lmbN-L-for and lmbN-L-rev was digested by EcoRI and NehI, and cloned into the 20 minonice, followed by centrifugation at 4 uCto remove the cellular debris, resulted
same sites of pKC1139 to yield plasmid pLL1033. The 1.94-kb fragment obtained in the supernatant that was used for the conversion of 2. Each biotransformation
using primers lmbN-R-for and lmbN-R-rev was digested by NheI and HindIII and was conducted in 100 ml of the mixture, which contained 1 ml of 2 (giving a final
cloned into the same site of pLL1033 to yield the recombinant plasmid pLL1034, concentration of 0.20 mM) and 99 ml of the supernatant of LL1005, and incubated
which encoded a S37A mutation that replaced the conserved motif XXXDSL with at 30 uC for 0, 1 or 4 h. After quenchingthe reactionby addition of an equal volume
XXXDAL and had a mutation that replaced CTC (encoding L38) with CTA (result- of methanol, 20 ml of each biotransformation mixture was centrifuged and subjected
ing no coding change) to introduce theNheI site. Following the procedure described to HPLC-ESI-MS analysis of 2 consumption and of lincomycin A production.
above, pLL1034 was introduced into the S. lincolnensis wild-type strain for double-
crossover recombination. The resulting strain LL1017 was then subjected to PCR
amplification to give a 0.86-kb product (using primers lmbN-gt-for and lmbN-gt-
rev). NheI digestion was then carried out on this PCR product to determine the
genotype, as judged by the release of the desired 0.11-kb and 0.75-kb fragments.
Further sequencing of this PCR product confirmed the genotype of LL1017, in
which lmbN was site-specifically mutated.
For complementation of lmbT in LL1006, the 1.5-kb lmbT-containing fragment
was amplified by PCR using primers lmbT-C-for and lmbT-C-rev, and then
cloned into pMD19-T to yield pLL1015. After digestion with BamHI and XbaI, this
1.5-kb DNA fragment was recovered from pLL1015 and ligated to a 0.45-kb EcoRI/
BamHII fragment from pWHM79, and the resulting product was ligated into the
EcoRI/XbaI site of pKC1139, yielding the recombinant plasmid pLL1016, in which
lmbT was under the control of the constitutive promoter PermE*. pLL1016 was
then introduced into LL1006 (DlmbT mutant) by conjugation, generating the cor-
responding recombinant strain LL1008 that expressed lmbT in trans.
For feeding of the MSH S-conjugate 1 to LL1005 (DmshAlin), 1 was added into
the culture broth of LL1005 on the 4th day of fermentation (giving a final con-
centration of approximately 0.20 mM). In a further 5-day incubation period, the
production of lincomycin A was monitored daily by HPLC and HPLC-ESI-MS as
described above.
Production and analysis of MSH and EGT in S. lincolnensis. Approximately
1 cm2 of the sporulated agar of S. lincolnensis or its derivative was cut, chopped,
and inoculated into 25 ml of YEME medium and then incubated at 28 uC and
220rpm for 36 h. 5 ml of the resulting culture broth was added into 100ml of the
same medium for scale-up and then further incubated at 28 uC and 220rpm for
3 days. The mycelia were then harvested by centrifugation at 4 uC and 5,000 rpm
for 15 min.
For thiol extraction, derivatization and detection, the procedure described by
Fahey and Newton33 was used with slight modifications. Approximately 200mg
(wet weight) of the freshly harvested mycelia was weighed in a 5 ml microcentri-
fuge tube and then re-suspended in 2 ml of a mixture of 50% CH3CN and 50%
2 mM monobromobimane (mBBr) dissolved in 20 mM Tris-HCl (pH 8.0) buffer
for sonication. The mixture was incubated in the dark at 60 uC for 15 min, acidified
with 5 ml of methanesulfonic acid (5N) and then centrifuged at 12,000rpm for
For complementation of lmbV in LL1007, a 1.03-kb lmbV-containing fragment
was amplified byPCRusing primerslmbV-C-forand lmbV-C-rev, and thencloned
intopMD19-TtoyieldpLL1017.AfterdigestionwithBamHIandXbaI, this1.03-kb
DNA fragment was recovered from pLL1017 and ligated to a 0.45-kb EcoRI/ 10 min to remove the debris before storing at 280 uC. HPLC-ESI-HR-MS analysis
BamHII fragment from pWHM79, and the resulting product was then ligated into ofthe resultingthiol-mBBr derivativeswascarriedoutona AgilentZorbaxcolumn
the EcoRI/XbaI site of pKC1139, yielding the recombinant plasmid pLL1018, in (Extend-C18, 1.8 mm, 2.13 50 mm, Agilent Technologies Inc., USA) by gradient
elution of solvent A (H2O) and solvent B (CH3CN) at a flow rate of 0.2ml min21
which lmbV was under the control of the constitutive promoter PermE*. pLL1018
was then introduced into LL1007 (DlmbV mutant) by conjugation, generating the over a 15 min period as follows: t5 0min, 5% B; t55 min, 5%B; t510 min, 50% B;
corresponding recombinant strain LL1009 that expressed lmbV trans.
t 5 12 min, 50% B; t 5 13 min, 5% B; and t 5 15 min, 5% B (mAU at 370nm).
For heterologous complementation of ccbV in LL1007, the 1.15-kb ccbV-containing Protein expression and purification. The recombinant proteins LmbE, CcbV,
fragment wasamplifiedfromthe genomicDNA ofthe celesticetin-producingstrain LmbT, LmbC, LmbN, LmbN-PCP and CcbD, all of which were in an N-terminal
S. caelestis NRRL2418 by PCR using primers ccbV-C-for and ccbV-C-rev, and 63 His-tagged form, were overproduced in E. coli BL21(DE3) and purified by Ni-
then cloned into pMD19-T to yield pLL1021. After digestion with BglII and XbaI, affinity followed by desalting.
this 1.03-kbDNA fragmentwas recovered from pLL1021 and subsequently ligated
to a 0.45-kb EcoRI/BamHII fragment from pWHM79; the resulting product was
ligated into the EcoRI/XbaI site of pKC1139, yielding the recombinant plasmid
pLL1022, in which ccbV was under the control of the constitutive promoter PermE*.
pLL1022 was then introduced into LL1007 (DlmbV mutant) by conjugation, gen-
erating thecorresponding recombinant strain LL1011 that heterologously expressed
ccbV in trans.
The genes of recombinant proteins LmbE, CcbV, LmbT, LmbC, LmbN, LmbN-
PCP and CcbD were PCR amplified from S. lincolnensis or S. caelestis genomic
DNA using primers with engineered NedI and HindIII restriction sites. The PCR-
amplified genefragmentswere purified, digestedwith NdeIand HindIIIand ligated
into a pET28a(1) vector (Novagen) that had been digested with the same enzymes.
The resultant plasmids were used to transform E. coli BL21(DE3) for protein over-
expression. The E. coli transformant cultures were incubated in Luria-Bertani (LB)
Production and analysis of lincomycin A, intermediate or shunt product. The mediumcontaining 50 mgml21 kanamycinat 37 uC and 250 rpm until the celldensity
S. lincolnensis wild-type strain or its derivative was spread on agar plates, which reached 0.6–0.8 at OD600 nm. To induce protein expression, IPTG (0.1 mM) was
was composed of 19 g of starch, 5 g of soybean meal, 0.5 g of K2HPO4, 0.5 g of added to the cultures, which were further incubated at 16 uC for 40–48 h. The cells
MgSO4N7H2O, 1.0 g of KNO3, 0.5 g of NaCl, 0.01 g of FeSO4N7H2O and 20.0 g of were harvested by centrifugation and stored at 280 uC before lysis. The thawed cells
agar per litre (pH 7.0,7.5), and then incubated at 28 uC for sporulation and were re-suspended in lysis buffer containing 50 mM Tris-HCl (pH 7.6), 300 mM
growth. Approximately 1 cm2 of the sporulated agar of S. lincolnensis was cut, NaCl, 10 mM imidazole and 10% (v/v) glycerol. After disruption by a low-temperature
chopped, and inoculated into 25 ml of the seed medium, which was composed of ultra-high-pressure cell disrupter (FB-110X, Shanghai Litu Mechanical Equipment
20 g of starch, 10 g of glucose, 10 g soybean of meal, 30 g of corn steep liquor, 1.5 g Engineering Co., Ltd, China or JN-02HC, JNBIO, China), the soluble fraction was
of (NH4)2SO4 and 5 g of CaCO3 per litre (pH7.0). After incubation at 28 uC and collected, subjected to purification of each target protein by using a HisTrap FF
220rpm for 36 h, 5 ml of the seed culture broth was transferred into 50 ml of the column (GE Healthcare, USA) and then desalted using a PD-10 Desalting Column
fermentation medium, which was composed of 100g of glucose, 25 g of soybean (GE Healthcare, USA) according to the manufacturers’ protocols. The resulting
meal, 2 g of corn steep liquor, 8 g of(NH4)2SO4, 0.2 g of KH2PO4, 8 g of NaNO3, 5 g proteins were concentrated and stored at 280 uC for in vitro assays. The purity of
of NaCl and 8 g of CaCO3 per litre (pH7.0). Further incubation was carried out at the proteinswasexaminedby12%SDS–PAGEanalysis, andtheconcentrationwas
28 uC and 220rpm for 7 days.
determined by Bradford assay using bovine serum albumin (BSA) as the standard.
For LmbN-PCP, HPLC-HR-ESI-Q-TOF MS (Agilent Technologies Inc., USA)
analysis indicated that the recombinant protein purified from E. coli BL21(DE3)
was fully phosphopantetheinylated into a holo form (m/z [M 1 H]1 calculated
11,842.94, found 11,843.42).
For product examination, 500 ml of each fermentation broth was mixed with an
equal volume of methanol, and after centrifugation to remove the residue, the super-
natant was subjected to HPLC analysis on an Agilent Zorbax column (SB-C18,
5 mm, 4.63 250mm, Agilent Technologies Inc., USA) by isocratic elution of 40%
5 mM ammonium acetate (NH4Ac) and 60% MeOH for 20 min at a flow rate of Determination of the amidase activity of LmbE in vitro. The assayswere carried
0.6 ml min21. HPLC-ESI-MS analysis of the supernatant diluted five times with out at 30 uC for 5h in a 50ml reaction mixture containing 50 mM Tris-HCl (pH7.5)
50% MeOH was carried out on the same column by gradient elution of solvent A and 2 mM substrate (MSH S-conjugate) in the presence of 2 mM (for 1) or 20 mM
(10 mM NH4Ac) and solvent B (CH3CN) ata flowrateof1 mlmin21 overa 22-min (for6)LmbE.Reactionsintheabsenceoftheenzymewereusedasnegativecontrols.
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
An equal volume of methanol was added into each mixture to terminate the and 2.00 mM). All assays were performed in triplicate, and each conversion was
reaction. After removal of the denatured protein by centrifugation, the reaction analysed by HPLC for EGT production as described above. The resulting initial
mixtures were subjected to HPLC-ESI-MS analysisona Phenomenex Luna C18(2) velocities were then fitted to the Michaelis–Menten equation using GraphPad Prism5
column (5 mm, 4.63 250mm, USA) by isocratic elution of 90% 10 mM NH4Ac
software (GraphPad Software, Inc., USA) to extract the Km and kcat parameters.
To determine the equilibrium constant (Keq) of the LmbT-catalysed reaction,
theexperimentwasperformedaccordingtothemethoddescribedpreviously34. Keq
was measured by performing a series of saturated reactions, in which the concen-
tration ratio of [GDP]/[3] varied from 1/3 to 5 on the premise that the ratio of [5]/
[EGT]wasfixedat1. The total concentrationsof[3]1[GDP]and[5]1[EGT]were
both kept at 4 mM. The reaction was performed in a 25 ml mixture containing
50 mM Tris-HCl (pH 7.5), 2 mM MgCl2 and 4 mM LmbT at 30 uC for 3 h to reach
equilibrium. The change of [EGT] was monitored by HPLC as described previously
and plotted against the ratio of [GDP]/[3]. The equilibrium constant was subse-
quently determined from the equation Keq 5 ([GDP]/[3])3 ([5]/[EGT]).
Characterization of PPL incorporation in vitro. To convert holo-LmbN-PCP
into PPL-acylated LmbN-PCP, the reaction was carried out at 30 uC for 3 h in a
50 ml reaction mixture containing 50 mM Tris-HCl (pH 7.5), 2 mM LmbC, 2 mM
PPL, 5 mM ATP, 10 mM MgCl2, 2 mM TCEP, and 1 mM CoA in the presence of
100 mM LmbN-PCP. Reactions in the absence of LmbC or ATP were used as nega-
tive controls. For product examination, the reaction mixtures were subjected to
HPLC-HR-ESI-Q-TOF MS (Agilent TechnologiesInc., USA) analysis on anAgilent
Zorbax column by gradient elution of solvent A (H2O containing 0.1% formic acid)
and solvent B (CH3CN containing 0.1% formic acid) at a flow rate of 0.2 ml min21
over a 45-min period as follows: t 5 0 min, 10% B; and t 5 45 min, 95% B (mAU at
210nm).
For PPL incorporation, the assays were carried out at 30 uC for 3 h in a 50 ml
reaction mixture containing 50 mM Tris-HCl (pH 7.5), 2 mM LmbC, 2 mM PPL,
5 mM ATP, 10 mM MgCl2, 2 mM TCEP, 1 mM CoA, and 2 mM substrate 5 (MSH
S-conjugate) in the presence of 50 mM LmbN or LmbN-PCP and 10 mM CcbD.
Reactions in the absence of each enzyme were used as negative controls. The ter-
minationofeachreactionandHPLC-ESI-MSanalysiswascarriedoutaccordingto
the methods described above for LmbE-catalysed conversion.
Compound isolation and purification. For compounds 1 and 4 (from the DlmbE
and DlmbV mutant strain, respectively), 100g of Amberlite XAD-2 resin (Rohm
and Haas Co., USA) was incubated with 1 l of each fermentation culture broth over-
night to absorb the target compound. After filtration, the resin was backwashed
with 2 l of water, and then eluted with 3 l of MeOH. The eluent was evaporated in vacuo
to a crude extract. The resultant residue was re-dissolved in 5 ml of water, and then
loaded onto a Sephadex LH20 column (3.53 200cm, GE Healthcare, USA) by
eluting 500 ml of MeOH at a flow rate of 0.5 ml min21. According to ESI-MS analysis,
the fractions containingthe target compound were combined, evaporated in vacuo
and then loaded onto an Agilent Zorbax column (SB-C18, 5 mm, 9.43 250mm,
AgilentTechnologiesInc.,USA)byisocraticelutionof40%5 mMNH4Acand60%
MeOHfor12 minataflowrateof2 ml min2l (mAU210nm). Afterasimilarwork-
up for fractionation and concentration, further purification was carried out on a
COSMOSIL HILIC Packed Column (5 mm, 103 250 mm, Nacalai Tesque Inc.,
Japan) by isocratic elution of 21% 10 mM NH4Ac and 79% CH3CN for 70 min at a
flow rate of 2 ml min21 (mAU 210nm).
and 10% CH3CN for 15 min period at a flow rate of 1 ml min21
.
Characterization of the CcbV-catalysed reaction in vitro. The assays were
carried out at 30 uC for 2 h in a 50 ml reaction mixture. For substrate 4, the mixture
contained50mMTris-HCl(pH8.0),1mMMSH,2mMTCEP,1mMEGTS-conjugate
(4) and 40 mM CcbV. For substrate 5, the mixture contained 50 mM Tris-HCl
(pH8.0), 2 mMMSH, 2 mMTCEP, 2 mMEGTS-conjugate (5) and100mM CcbV.
Reactions in the absence of the enzyme were used as negative controls. The ter-
mination of each reaction and analysis of the resulting MSH S-conjugate and EGT
were carried out according to the methods described above for LmbE-catalysed
conversion. HPLC analysis of EGT production was performed on a COSMOSIL
HILIC Packed Column (5mm, 4.63250 mm, Nacalai Tesque Inc., Japan) by iso-
cratic elution of 30% 10 mM NH4Ac and 70% CH3CN for 20 min at a flow rate of
1 mlmin21 (mAUat260 nM). ThecommerciallyavailableEGT(EnzoLifeSciences
Inc., USA) was used as the standard.
To evaluate the pH dependence, each reaction was performed in triplicate at
30 uC for 1 h in a 25 ml reaction mixture containing 1 mM MSH, 2 mM TCEP,
1 mM 4 and 20 mM CcbV in 50 mM PIPES (pH6.0–7.0) or 50 mM Tris-HCl
(pH7.5–9.0) buffer.
For the kinetic analysis, a time course was carried out to determine the initial
rate conditions in a 25 ml of reaction mixture containing 50 mM Tris-HCl buffer
(pH8.0), 1 mM MSH, 1 mM TCEP, 1 mM 4 and 20 mM CcbV. The reactions were
initiated by the addition of CcbV, incubated at 30 uC, and then terminated by 25 ml
of methanol at 3, 5, 7, 10, 16, 25, 45 and 60 min. The samples were subjected to the
same work-ups and HPLC analysis of EGT production as described above. The
production of EGT, linear with respect to time during 0–20 min, was fitted into a
linear equation to obtain the initial velocity. To determine the kinetic parameters
for substrate 4, the reactions were carried out at 30 uC for 20 min, each in 25 ml of
mixture containing 50 mM Tris-HCl buffer (pH8.0), 2 mM MSH, 2 mM TCEP,
and 20 mM CcbV, and varying the concentration of substrate 4 (0.02, 0.05, 0.10,
0.20, 0.50, 1.00 and 2.00 mM). To determine the kinetic parameters for the sub-
strate MSH, the reactions were carried out at 30 uC for 20 min, each in 25 ml of
mixture containing 50 mM Tris-HCl buffer (pH8.0), 5 mM 4, 2 mM TCEP, and
20 mM CcbV, and varying the concentration of substrate MSH (0.02, 0.05, 0.10,
0.20, 0.50 and 1.00 mM). All assays were performed in triplicate, and each con-
version was analysed byHPLCfor EGT production as described above. The resulting
initialvelocitieswerethenfittedtotheMichaelis–MentenequationusingGraphPad
Prism5 software (GraphPad Software, Inc., USA) to extract the Km and kcat parameters.
Characterization of the LmbT-catalysed reaction in vitro. The reverse glyco-
sylationassaysofLmbTwerecarriedoutat30 uCfor2 hina50 mlreactionmixture.
For substrate 5, the mixture contained 50 mM Tris-HCl (pH7.5), 2 mM MgCl2,
1mM5, 2 mMGDPand4 mM LmbT. For substrate 4, themixturecontained50mM
Tris-HCl (pH7.5), 2 mM MgCl2, 2 mM 4, 4 mM GDP and 20 mM LmbT. The
forward glycosylation assays were carried out at 30 uC for 2 h in a 50 ml reaction
mixture containing 50 mM Tris-HCl (pH7.5), 2 mM MgCl2, 1 mM 3, 1 mM EGT
and 4 mM LmbT. Reactions in the absence of the enzyme were used as negative
controls. The termination of each reaction was carried out according to the meth-
ods described above for LmbE-catalysed conversion. The reaction mixtures con-
taining 5 were subjected to HPLC-ESI-MSanalysison a Phenomenex Luna C18(2)
column by isocratic elution of 95% 10 mM NH4Ac and 5% CH3CN for 15 min at a
ForPPL(fromtheDlmbTmutantstrain), after asimilarprocedure with Amberlite
XAD-2 resin, further HPLC semi-preparation purification was carried out twice on
anAgilentZorbaxcolumnbyisocraticelutionof20%5 mMNH4Ac and 80% MeOH
for 30 min at a flow rate of 2 ml min21 (mAU at 210nm).
flow rate of 1 ml min21
.
For compounds 5 and 6 (from the DlmbC mutant strain), 100g of Amberlite
XAD-16 resin (Rohm and Haas Co., USA) was incubated with 1 l of fermentation
culture broth overnight to remove most of the impurity. After filtration and con-
centration, the resultantresidue wasthen loaded onto a SephadexLH20columnby
eluting 500ml of H2O:MeOH (1:1) at a flow rate of 0.5 ml min21. According to
ESI-MS analysis, the fractions containing compound 5 or 6 were combined, eva-
porated in vacuo and then loaded onto a COSMOSIL HILIC Packed Column by
isocratic elution of 30% 10 mM NH4Ac and 70% CH3CN for 30 min at a flow rate
of2 ml min21 (mAU at210nm). For compound6, further purification wascarried
out on a Sephadex G10 column (1.5 3 120cm, GE Healthcare, USA) by eluting
To evaluate the pH dependence, each reaction was performed in triplicate at
30 uC for 1 h in 25 ml of reaction mixture containing 2 mM MgCl2, 4 mM GDP,
2 mM 5 and 4 mM LmbT in 50 mM PIPES (pH6.0–7.0) or 50 mM Tris-HCl
(pH7.5–9.0) buffer.
For the kinetic analysis, a time course was carried out to determine the initial
rate conditions in 25 ml of reaction mixture containing 50 mM Tris-HCl (pH7.5),
2 mM MgCl2, 1 mM 5, 2 mM GDP and 4 mM LmbT. The reactions were initiated
by the addition of LmbT, incubated at 30 uC, and then terminated by 25 ml of
methanol at 3, 5, 7, 10, 16, 25, 45 and 60 min. The samples were subjected to the
samework-upsandHPLCanalysisofEGTproductionasdescribedabove.Thepro-
ductionofEGT, linear with respect totimeduring0–20min, wasfittedinto alinear
equation to obtain the initial velocity. To determine the kinetic parameters for
substrate 5, the reactions were carried out at 30 uC for 20 min, each in 25 mlofmixture
containing 50 mM Tris-HCl buffer (pH 7.5), 2 mM MgCl2, 2mM GDP, and 4mM
LmbT, and varying the concentration of substrate 5 (0.02, 0.05, 0.10, 0.20, 0.50,
1.00, 2.00 and 4.00 mM). To determine the kinetic parameters for substrate GDP,
the reactions were carried out at 30 uC for 20 min, each in 25 ml of mixture con-
taining 50 mM Tris-HCl buffer (pH7.5), 2 mM MgCl2, 4 mM 5, and 4 mM LmbT,
H2O at a flow rate of 0.2 ml min21
.
For compounds 2 and GlcN-Ins (from the LmbE-catalysed reaction), the assay
was scaled up and carried out at 30 uC for 2 h in 20 ml of mixture containing 200 mM
LmbE, 1 mMpure compound1and50 mMTris-HClbuffer(pH 7.5). After filtration
with an ultra-filtration membrane (Amicon YM-30, Millipore) to remove protein,
the solution was loaded onto an Agilent Zorbax column by gradient elution of
solvent A (H2O)andsolvent B (CH3CN)ata flow rate of2 ml min21 over a 18-min
period as follows: t 5 0 min, 5% B; t 5 5 min, 5% B; t 5 10 min, 55% B; t 5 16 min,
55% B; and t 5 18 min, 5% B (mAU at 210nm). The fraction containing GlcN-Ins
and varying the concentration of substrate GDP (0.02, 0.05, 0.10, 0.20, 0.50, 1.00 was concentrated and then loaded onto a COSMOSIL HILIC Packed Column by
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
isocratic elution of 25% 10 mM NH4Ac and 75% CH3CN for 40 min at a flow rate
of 2 ml min21. GlcN-Ins was examined using a refractive index (RI) detector. For
compound 2, further purification was carried out on the same column by the same
isocratic elution for 60 min (mAU at 210nm).
31. Kieser, T. et al. Practical Streptomyces Genetics (John Innes Foundation,
2000).
32. Green, M. R. & Sambrook, J. Molecular Cloning: A Laboratory Manual 4th edn (Cold
Spring Harbor Laboratory Press, 2012).
33. Fahey, R. C. & Newton, G. L. Determination of low molecular weight
thiols using monobromobimane fluorescent labeling and
high-performance liquid chromatography. Methods Enzymol. 143, 85–96
(1987).
34. Zhang, C. et al. Exploiting the reversibility of natural product glycosyltransferase-
catalyzed reactions. Science 313, 1291–1294 (2006).
35. Lee, S. & Rossaza, J. P. First total synthesis of mycothiol and mycothiol disulfide.
Org. Lett. 6, 365–368 (2004).
36. Ajayi, K., Thakur, V. V., Lapo, R. C. & Knapp, S. Intramolecular
a-glucosaminidation: synthesis of mycothiol. Org. Lett. 12, 2630–2633
(2010).
For compound 3 (from the LmbT-catalysed reverse glycosylation reaction), the
assay was scaled up and carried out at 30 uC for 5 h in 10 ml of mixture containing
4 mM LmbT, 2 mM pure compound 5, 2 mM GDP, 1 mM MgCl2 and 50 mM Tris-
HCl buffer (pH7.5). After filtration with an ultra-filtration membrane to remove
protein, the solution was concentrated and then subjected to a Sephadex G10 col-
umn by eluting H2O at a flow rate of 0.1ml min21
.
Chemical synthesis of MSH. The synthesis was carried out according to the
procedures described previously35,36 with slight modifications. For details, see Sup-
plementary Methods.
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
Extended Data Figure 1
|
Constructive role of MSH (orange) and EGT (green) in lincomycin biosynthesis (shown as cartoon models).
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
Extended Data Figure 2
|
Genes and/or clusters relevant to lincomycin
genes lin5590, lin5591, lin5593 and lin5594 share sequence homology with
biosynthesis in S. lincolnensis. a, Biosynthetic gene cluster of lincomycin A. genes encoding the methylase (WP_026151264.1) from S. prunicolor, the
The mca homologue lmbE and two GTase-encoding genes, lmbV and lmbT, are DUF899-like protein (WP_004004056.1) from S. viridochromogenes, the
shown in yellow, blue and green, respectively. The genes responsible for PPL
incorporation (lmbC, lmbD and lmbN) are shown in grey. b, The location of
mshAlin (shown in purple), which is not clustered with the other genes
YbjN-like protein (WP_020130332.1) from Streptomyces sp. 303MFCol5.2
and the hypothetical protein (WP_004000160.1) from S. viridochromogenes,
respectively. c, The biosynthetic gene cluster of EGT. The gene egtDlin is shown
responsibleforMSHbiosynthesis, in thegenome of S. lincolnensis. The flanking in red.
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
Extended Data Figure 3
lincomycin A in various S. lincolnensis strains. The strains include the
mutants (i, for DlmbE; ii, for DlmbE-E80; iii, for DlmbE-E447; and iv, for
|
The production of MSH S-conjugate 1 and
DlmbE-E3457) and the wild-type control (v). For HPLC-MS analysis, the ESI
m/z [M 1 H]1 modes are indicated in the dashed rectangle.
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
Extended Data Figure 4
|
Analysis of MSH and EGT production in the
[M 1 H]1 677.2340) and EGT-mBBr (calculated for C19H26N5O4S1
DmshAlin and DegtDlin mutant S. lincolnensis strains. a, Derivatization of
MSH and EGT with mBBr to generate the corresponding S-conjugates.
b, HPLC-HR-MS analysis of MSH-mBBr (calculated for C27H41N4O14S1
[M 1 H]1 420.1706) in the wild-type control (i), LL1005 (ii, DmshAlin mutant)
and LL1010 (iii, DegtDlin mutant). c, The HR-ESI-MS spectra of EGT-mBBr
(top) and MSH-mBBr (below).
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
Extended Data Figure 5
|
Phylogenetic analysis in the DinB-2 superfamily. which is found in eukaryotic proteins required for post-translational
a, LmbV (from S. lincolnensis) and CcbV (from S. caelestis), shown in red, with modification to produce sulfatases, which are essential for the degradation and
selected DinB-2-like proteins (from various actinomycetes, in which the thiol remodelling of sulfate esters. Clade iii features the N-terminal zf-HC2 domain,
MSH is dominant) in the phylogenetic tree. The evolutionary distances were
computed using thep-distance method. The support for grouping clades i, ii, iii, anti-sigma factor proteins. Clade iv features the C-terminal SCP2 domain
iv and v (shaded in different colours) is indicated by bootstrap values. The involved in binding sterols. Clade v features the C-terminal wyosine_f domain.
known MSH-maleylpyruvate isomerase Ncgl2918 is shown in blue. b, Typical Some proteins containing this domain appear to be important in wyosine base
domain organization of the DinB-2-like proteins. The conserved DinB-2 formation in a subset of phenylalanine-specific tRNAs. ‘Others’ indicate a
which contains a putative zinc-finger binding motif and is found in some
domain is shown in green. Clade i features the C-terminal MDMPI-C domain. number of DinB-2-like proteins that possess an unknown domain(s) either at
Proteins containing this domain include the MSH-maleylpyruvate isomerases, the C terminus or at both the C and N termini of the proteins.
such as Ncgl2918. Clade ii features the C-terminal FGE-sulfatase domain,
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
Extended Data Figure 6
|
Kinetic analysis of CcbV-catalysed thiol exchange. steady-state kinetic parameters for substrate 4 and for MSH, respectively. The
a, pH dependence. The activity of CcbV in 50 mM PIPES (pH 6.0-7.0) or
50 mM Tris-HCl (pH 7.5–9.0) buffer was measured. b, c, Determination of the
error bars are standard error of mean (n 5 3).
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
Extended Data Figure 7
|
Characterization of LmbT-catalysed reverse and
GDP-D-a-D-sugar 7 (left) in the absence (top right) and in the presence
forward glycosylation. For HPLC-MS analysis, the ESI m/z [M 1 H]1 modes (lower right) of LmbT. b, Characterization of LmbT-catalysed forward
are indicated in dashed rectangle. a, Examination of the acylated C8-sugar
transfer in the presence of LmbT, which showed that LmbT was unable to
utilize 4 as a substrate for reverse glycosylation to generate the predicted
glycosylation. LmbT used 3 as a substrate for glycosylation to generate the
GDP-D-a-D-sugar 5 (left) in the absence (top right) and in the presence
(lower right) of LmbT.
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
Extended Data Figure 8
|
Characterization of PPL incorporation in
examination of LmbC-catalysed conversion of holo-LmbN-PCP
lincomycin A biosynthesis. For HPLC-MS analysis, the ESI m/z [M 1 H]1
modes are indicated in the dashed rectangle. a, In vivo product profiles of
(m/z [M 1 H]1 calculated 11,842.94, found 11,843.42) into PPL-acylated
LmbN-PCP (m/z [M 1 H]1 calculated 11,982.01, found 11,983.21) in the
S. lincolnensis strains, including the mutants (i, for DlmbC; ii, for DlmbN; iii, for absence (left) and in the presence (right) of ATP. d, In vitro analysis of the
DlmbD; and iv, for DlmbT) and the wild-type control (v). b, Process of the
incorporation of PPL (with EGT S-conjugate 5) into intermediate 4. PCP
(blue), peptidyl carrier protein; and I (grey), isomerase. c, HPLC-MS
condensation between PPL and 5 to generate 4. The catalyst systems included
LmbC 1 CcbD (i), LmbC 1 LmbN (ii), LmbC 1 LmbN-PCP (iii),
LmbC 1 LmbN 1 CcbD (iv), and LmbC 1 LmbN-PCP 1 CcbD (v).
Macmillan Publishers Limited. All rights reserved
©2015

RESEARCH LETTER
Extended Data Figure 9
glycosylation. a, pH dependence. The activity of LmbT in 50 mM PIPES
(pH 6.0–7.0) or 50 mM Tris-HCl (pH 7.5–9.0) buffer was measured.
|
Kinetic analysis of LmbT-catalysed reversible
GDP, respectively, in LmbT-catalysed reverse glycosylation. The error bars
are standard error of mean (n 5 3). d, Determination of the equilibrium
constant (Keq) of LmbT-catalysed glycosylation. Keq 5 ([GDP]/[3])3 ([5]/
b, c, Determination ofthesteady-statekinetic parameters forsubstrate 5 andfor [EGT]) 5 1.943 1 5 1.94.
Macmillan Publishers Limited. All rights reserved
©2015

LETTER RESEARCH
Extended Data Figure 10
|
Validation of the highly ordered process
(bottom right) of CcbV, showing that CcbV was able to convert 5 to MSH
involving EGT-mediated assembly and MSH-associated post-modifications S-conjugate 6 (left). b, Determination of the hydrolysis reaction in the absence
in lincomycin biosynthesis. For HPLC-MS analysis, the ESI m/z [M 1 H]1 (top right) and in the presence (bottom right) of LmbE, showing that this
modes are indicated in the dashed rectangle. a, Examination of the thiol
enzyme was unable to utilize 6 as a substrate and convert it to the predicted
exchange using 5 as a substrate in the absence (top right) and in the presence mercapturic acid 8 (left).
Macmillan Publishers Limited. All rights reserved
©2015