Deciphering the Biosynthetic Origin of L-allo-Isoleucine

Article abstract of DOI:10.1021/jacs.5b11380

The nonproteinogenic amino acid l-allo-isoleucine (l-allo-Ile) is featured in an assortment of life forms comprised of, but not limited to, bacteria, fungi, plants and mammalian systems including Homo sapiens. Despite its ubiquity and functional importance, the specific origins of this unique amino acid have eluded characterization. In this study, we describe the discovery and characterization of two enzyme pairs consisting of a pyridoxal 5′-phosphate (PLP)-linked aminotransferase and an unprecedented isomerase synergistically responsible for the biosynthesis of l-allo-Ile from l-isoleucine (l-Ile) in natural products. DsaD/DsaE from the desotamide biosynthetic pathway in Streptomyces scopuliridis SCSIO ZJ46, and MfnO/MfnH from the marformycin biosynthetic pathway in Streptomyces drozdowiczii SCSIO 10141 drive l-allo-Ile generation in each respective system. In vivo gene inactivations validated the importance of the DsaD/DsaE pair and MfnO/MfnH pair in l-allo-Ile unit biosynthesis. Inactivation of PLP-linked aminotransferases DsaD and MfnO led to significantly diminished desotamide and marformycin titers, respectively. Additionally, inactivation of the isomerase genes dsaE and mfnH completely abolished production of all l-allo-Ile-containing metabolites in both biosynthetic pathways. Notably, in vitro biochemical assays revealed that DsaD/DsaE and MfnO/MfnH each catalyze a bidirectional reaction between l-allo-Ile and l-Ile. Site-directed mutagenesis experiments revealed that the enzymatic reaction involves a PLP-linked ketimine intermediate and uses an arginine residue from the C-terminus of each isomerase to epimerize the amino acid β-position. Consequently, these data provide important new insight into the origins of l-allo-Ile in natural products with medicinal potential and illuminate new possibilities for biotool development.

Full text of DOI:10.1021/jacs.5b11380

Subscriber access provided by The Chinese University of Hong Kong
Article
Deciphering the biosynthetic origin of L-allo-isoleucine
Qinglian Li, Xiangjing Qin, Jing Liu, Chun Gui, Bo Wang, Jie Li, and Jianhua Ju
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.5b11380 • Publication Date (Web): 15 Dec 2015
Downloaded from http://pubs.acs.org on December 16, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
Deciphering the biosynthetic origin of L-allo-isoleucine
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Qinglian Li, Xiangjing Qin, Jing Liu, Chun Gui, Bo Wang, Jie Li, and Jianhua Ju*
CAS Key Laboratory of Tropical Marine Bioꢀresources and Ecology, Guangdong Key Laboratory of Marine Materia Mediꢀ
ca, RNAM Center for Marine Microbiology, South China Sea Institute of Oceanology, Chinese Academy of Sciences,
Guangzhou 510301, China
Supporting Information
ABSTRACT: The nonproteinogenic amino acid Lꢀalloꢀisoleucine (LꢀalloꢀIle) is featured in an assortment of life forms comprised
of, but not limited to, bacteria, fungi, plants and mammalian systems including homo sapiens. Despite its ubiquity and function
importance, the specific origins of this unique amino acid have eluded characterization. In this study, we describe the discovery and
characterization of two enzyme pairs consisting of a pyridoxal 5’ꢀphosphate (PLP)ꢀlinked aminotransferase and an unprecedented
isomerase synergistically responsible for the biosynthesis of
from the desotamide biosynthetic pathway in Streptomyces scopuliridis SCSIO ZJ46, and MfnO/MfnH from the marformycin bioꢀ
synthetic pathway in Streptomyces drozdowiczii SCSIO 10141 drive ꢀalloꢀIle generation in each respective system. In vivo gene
inactivations validated the importance of the DsaD/DsaE pair and MfnO/MfnH pair in ꢀalloꢀIle unit biosynthesis. Inactivation of
PLPꢀlinked aminotransferases DsaD and MfnO led to significantly diminished desotamide and marformycin titers, respectively.
Additionally, inactivation of the isomerase genes dsaE and mfnH completely abolished production of all ꢀalloꢀIleꢀcontaining meꢀ
tabolites in both biosynthetic pathways. Notably, in vitro biochemical assays revealed that DsaD/DsaE and MfnO/MfnH each cataꢀ
lyze a bidirectional reaction between ꢀalloꢀIle and ꢀIle. Siteꢀdirected mutagenesis experiments revealed that the enzymatic reacꢀ
tion involves a PLPꢀlinked ketimine intermediate and uses an arginine residue from the Cꢀterminus of each isomerase to epimerize
LꢀalloꢀIle from Lꢀisoleucine (LꢀIle) in natural products. DsaD/DsaE
L
L
L
L
L
the amino acid βꢀposition. Consequently, these data provide important new insight into the origins of LꢀalloꢀIle in natural products
with medicinal potential and illuminate new possibilities for biotool development.
and relevant mechanisms involved in
have eluded characterization.
LꢀalloꢀIle production
ꢀ INTRODUCTION
Isoleucine contains two asymmetric centers and can thereꢀ
We recently discovered two sets of cyclic peptide natural
products from two deep South China Seaꢀderived Streptomyꢀ
ces. The desotamides7 display antibacterial activities against
pathogenic Gramꢀpositive bacteria, and the marformycins8 are
bioactive against Propionibacterium acnes hinting at their
potential for treating acne vulgaris (Figure 1B); most imꢀ
fore exist as any one of four possible stereoisomers, (2S, 3S)
isoleucine, (2R, 3R) ꢀisoleucine, (2S, 3R) ꢀalloꢀisoleucine,
and (2R, 3S) ꢀalloꢀisoleucine (Figure 1A). The latter three
nonꢀproteinogenic amino acids have been found in nature. Of
Lꢀ
D
L
D
particular interest is
L
ꢀalloꢀIle which was originally reported in
1958.1 The unique
L
ꢀalloꢀIle has been detected in plants,2
portantly, the rare LꢀalloꢀIle moiety is represented in both clasꢀ
found as a building block that is resistant to enzymatic digesꢀ
tion within cyclic peptide antibiotics [i.e., aureobasidin A,3
cordyheptapeptides,4 aspergillicin E5 from fungi, and gloꢀ
bomycin,6 desotamides7 and marformycins8 from actinomyꢀ
cetes], and has been used as a precursor to synthesize the phyꢀ
totoxin coronatine9 from Pseudomonas syringae (Figure S1).
ses of new natural products. We also have identified the gene
clusters responsible for producing the desotamides in Streptoꢀ
myces scopuliridis SCSIO ZJ46,18 and the marformycins19
from Streptomyces drozdowiczii SCSIO 10141; these efforts
set the stage for deciphering the means by which LꢀalloꢀIle is
biosynthesized and subsequently used as a natural product
building block. We report herein two enzyme pairs
DsaD/DsaE and MfnO/MfnH that are functionally equivalent
and drive the biosynthesis of LꢀalloꢀIle during production of
the desotamides and marformycins, respectively.
Interestingly, LꢀalloꢀIle is typically present in human plasma at
barely detectable concentrations.10–13 However, in people sufꢀ
fering from the inherited autosomal recessive disease maple
syrup urine disease (MSUD) LꢀalloꢀIle accumulates in plasma
and urine, the result of a deficiency in mitochondrial
branchedꢀchain 2ꢀketoacid dehydrogenase activity.13–17 Conseꢀ
ꢀ RESULTS AND DISCUSSION
quently,
(>5 ꢀM in blood) in diagnosing MSUD patients.13,17 Despite
its structural similarity to ꢀIle and ubiquity in various life
kingdoms, the precise means of ꢀalloꢀIle generation in living
systems has long been a mystery; the biosynthetic enzymes
LꢀalloꢀIle has been used as a pathognomonic marker
Bioinformatics Analysis of DsaD/DsaE and MfnO/MfnH
Enzyme Pairs in the Desotamides (DSA) and Mar-
formycins (MFN) Biosynthetic Gene Clusters. Stereochemiꢀ
L
L
cal comparisons between
LꢀIle and its diastereomer LꢀalloꢀIle,
revealed the only difference to be that of the βꢀcarbon stereoꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 10
1
2
3
4
5
6
7
8
comprehensive HRMS2, NMR data, and Xꢀray diffraction data
analyses (Figures S8–S13). Fermentations of the ꢁmfnO muꢀ
chemical configuration. The βꢀcarbon stereochemistry was
assigned as 3S for ꢀIle and 3R for ꢀalloꢀIle (Figure 1A). This
realization inspired us to hypothesize that ꢀalloꢀIle might
originate from the natural amino acid ꢀIle, and that the transꢀ
formation might involve an ꢀIle 2ꢀdeaminization process
L
L
L
tant afforded compounds 3 (MFN C), 5 (MFN E) and 7 (MFN
G) (Figure 1B and Figure 2A, trace ii). The structures of 3 and
7 were confirmed following large scale fermentation, subseꢀ
quent isolation, and HRMS2 and NMR data comparisons with
authentic samples. Although the ꢁmfnO mutant retains the
L
L
catalyzed by an aminotransferase followed by an isomeraseꢀ
catalyzed 3ꢀepimerization process. Bioinformatics analysis
revealed that, dsaD in the DSA gene cluster, and mfnO in the
MFN gene cluster, each encodes a typical aminotransferase.
Sequence alignment revealed that these two proteins showed
high sequence homology to the wellꢀstudied branchedꢀchain
aminotransferases (BCATs), and featured the conserved
EXGXXNLFXnLXTXnLXGVXR signature motif found in all
classꢀIV aminotransferases. DsaD and MfnO also both possess
the conserved catalytic lysine residue covalently linked to
pyridoxal 5’ꢀphosphate (PLP) (Figure S2),20–23 strongly sugꢀ
gesting their PLPꢀdependent aminotransferase activities. Durꢀ
ing efforts to identify isomerase candidates, BLASTP analyses
failed to reveal clear candidates in either the DSA or MFN
gene clusters. Subsequent structural homology searching was
then performed using HHpred24 revealing that the folding arꢀ
chitecture of DsaE, a small protein (124 aa) coded for immeꢀ
diately upstream of the dsaD aminotransferase gene in the
DSA gene cluster (Figure 1C), is shared with the nuclear
transport factor 2 (NTF2) superfamily. The NTF2 superfamily
contains thousands of functionally divergent proteins bearing
low sequence homology to each other; members include the
delta5ꢀ3ꢀketosteroid isomerases (KSI) responsible for isomerꢀ
izing delta5ꢀ3ꢀketosteroid to delta4ꢀ3ꢀketosteroid.25 Accordingꢀ
ability to produce 3 containing the LꢀalloꢀIle moiety, its imꢀ
9
paired biosynthetic capability, relative to that of the wildꢀtype
producer, may be explained by the presence of a compensatory
aminotransferase elsewhere within the S. drozdowiczii SCSIO
10141 genome.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A
O
O
(S)
(R)
(S)
(R)
OH
OH
NH2
NH2
L
-Ile
D
-Ile
O
O
(S)
(R)
(S)
(R)
OH
OH
NH2
NH2
L allo
-
D allo
-
-Ile
-Ile
B
L allo
L
-
-Ile / -Val
NH
O
R1
R2
O
H
N
O
O
O
N
H
H
N
N
H
H
N
O
NH
HN
N
H
O
O
H
N
O
O
O
O
N
O
NH
HN
O
R3
H
N
HN
O
R1
R2
O
L allo
L
-
-Ile / -Val
ly, we envisioned that DsaE may be involved in the
L
ꢀIle βꢀ
O
1
2
6
3
4
5
7
8 DSA A, R1 = CH3, R2 = NH2
9 DSA B, R1 = H, R2 = NH2
10 DSA G, R1 = CH3, R2 = OH
11 DSA H, R1 = H, R2 = OH
MFN A, R1 = H, R2 = CH3, R3 = H
MFN B, R1 = CH3, R2 = CH3, R3 = H
MFN F, R1 = H, R2 = H, R3 = H
MFN C, R1 = CH3, R2 = H, R3 = OH
MFN D, R1 = CH3, R2 = CH3, R3 = OH
MFN E, R1 = H, R2 = H, R3 = OH
MFN G, R1 = H, R2 = CH3, R3 = OH
carbon epimerization en route to ꢀalloꢀIle. Notably, an
L
orthologue of DsaE (identity: 42%; similarity: 56%), MfnH,
was found in the MFN gene cluster. In contrast to dsaE which
is located immediately upstream of dsaD within the DSA clusꢀ
ter, mfnH is located six open reading frames upstream of the
aminotransferase mfnO gene within the MFN gene cluster
(Figure 1C). The identification of aminotransferase/isomerase
enzyme pairs in both the DSA and MFN gene clusters indiꢀ
cates a conserved biosynthetic logic in generating the Lꢀalloꢀ
Ile precursor for these two sets of cyclic peptide natural prodꢀ
ucts.
C
Figure 1. Structures of (A) the four stereoisomers of Ile and (B)
marformycins and desotamides in which the ꢀalloꢀIle moieties
L
In vivo Gene Inactivations of DsaD/DsaE and
MfnO/MfnH and Resultant Metabolite Profiles in Deso-
tamide and Marformycin Biosynthetic Pathways. To probe
the exact roles of the DsaD/DsaE and MfnO/MfnH enzyme
pairs in the biosynthesis of desotamides and marformycins, we
inactivated mfnO and mfnH in S. drozdowiczii using estabꢀ
lished λꢀREDꢀmediated PCRꢀtargeting mutagenesis methods26
(Figures S4 and S5). The ꢁmfnH mutant failed to produce
compounds 3 and 4 (MFNs C and D) both of which contain
are shown in red. (C) The locations of the aminotransferꢀ
ase/isomerase pairs DsaD/DsaE and MfnO/MfnH in the DSA and
MFN gene clusters, respectively. The aminotransferases
DsaD/MfnO are shaded in red, and the isomerases DsaE/MfnH
are shaded in green.
To explore the roles of dsaD and dsaE, each gene was inꢀ
frame deleted within the DSA cluster and heterologously exꢀ
pressed in the host strain Streptomyces coelicolor M1152
(Figures S6 and S7). S. coelicolor M1152 was selected as the
host for these experiments because the natural DSA producer
S. scopuliridis SCSIO ZJ46 is recalcitrant to genetic engineerꢀ
ing.18 The results of these heterologous expression experiꢀ
ments paralleled those obtained with mfnO and mfnH inactivaꢀ
tions in S. drozdowiczii. The deletion of dsaE fully abolished
the
However, fermentations of the ꢁmfnH mutant did afford comꢀ
pounds 5 and 7 (MFNs E and G) in which the ꢀalloꢀIle moieꢀ
ty of the MFN scaffold was replaced by a corresponding ꢀVal
LꢀalloꢀIle moiety (Figure 1B and Figure 2A, trace iii).
L
L
moiety, as revealed by HPLCꢀUV, HPLCꢀMS and NMR data
analyses. We noted that compound 7 (MFN G) has a retention
time almost identical to that of 4 (MFN D) (Figure 2A, traces i
– iii). Largeꢀscale fermentation of the ꢁmfnH mutant enabled
isolation of quantities of 7 sufficient to support thorough strucꢀ
ture elucidation and 7 was unambiguously assigned as a new
production of
LꢀalloꢀIleꢀcontaining compounds 8 and 10
(DSAs A and G) (Figure 1B and Figure 2B, trace iv). Howevꢀ
er, the ꢁdsaE mutant did produce 9 (DSA B) and 11 (DSA H);
both 9 and 11 contain LꢀVal instead of LꢀalloꢀIle. The deletion
compound containing a pendant LꢀVal moiety on the basis of
ACS Paragon Plus Environment

Page 3 of 10
Journal of the American Chemical Society
1
2
of dsaD correlated to production of the same four products 8–
sole DsaD or DsaE in the presence of αꢀketoglutarate (αꢀKG)
and additional PLP in phosphate buffer (50 mM, pH 8.0). Unꢀ
der these conditions, ꢀIle remained unchanged and none of
the expected 3ꢀmethylꢀ2ꢀoxylꢀpentanoic acid product was obꢀ
served upon chiral phase HPLC analysis of the reaction mixꢀ
ture (Figure S15A, traces iii and iv). DsaD, in isolation, failed
to display classical aminotransferase activity. However, when
such reactions contained both DsaD and DsaE, a new HPLC
11. However, the yields of 8 and 10 were significantly deꢀ
creased and the yields of 9 and 11 were increased relative to
those noted with the heterologously expressed full DSA pathꢀ
way (Figure 1B and Figure 2B, traces ii and iii). This finding,
as with those related to the MFN system, suggests that the loss
of aminotransferase function by DsaD was partially compenꢀ
sated for by another enzyme within the S. coelicolor M1152
genome. The results of these in vivo gene inactivations clearly
confirm the necessity of the DsaD/DsaE and MfnO/MfnH
3
4
5
6
7
8
9
L
peak with a retention time corresponding to that of LꢀalloꢀIle
standard was observed (Figure S15A, trace v). We propose
that the PLP coꢀpurified with DsaD might be sufficient for the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
enzyme pairs in LꢀalloꢀIle biosynthesis.
DsaD/DsaEꢀcatalyzed conversion of LꢀIle to LꢀalloꢀIle. To test
this hypothesis and to ascertain the possible requirement of
another cofactor, αꢀkG, for the DsaD/DsaEꢀcatalyzed converꢀ
sion of LꢀIle to LꢀalloꢀIle, exogenous PLP, αꢀkG and PLP/ αꢀ
kG were removed from the reaction, respectively. The results
of these efforts revealed that the intrinsic PLP coꢀpurified with
DsaD is, in fact, sufficient to enable the LꢀIle ꢀ LꢀalloꢀIle
conversion, and that αꢀkG is not required for the DsaD/DsaEꢀ
coupled reaction (Figure 3A and Figure S16A).
A
1.0
3
m/z: 887.5421
[M + H]+
0.5
S. drozdowiczii
SCSIO 10141
4
i
5
5
0
0.5
m/z: 873.5086
[M + H]+
3
ii
mfnO mutant
mfnH mutant
7
0
m/z: 873.5086
0.5
[M + H]+
7
iii
0
5
13
14
15
16
17
18
10
11
12
Time (min)
B
1.0
i
S. coelicolor M1152
0
1.0
8
S. coelicolor M1152
10
10
ii
9
11
11
/07-6H
0
1.0
S. coelicolor M1152
9
9
8
iii
iv
/07-6H-DKO
0
1.0
S. coelicolor M1152
11
/07-6H-EKO
0
15
Time (min)
12
13
14
16
17
18
10
11
Figure 3. In vitro characterization of DsaD/DsaE. (A) Production
of LꢀalloꢀIle proceeds only upon coꢀincubation of LꢀIle with DsaD
and DsaE (v). Enzyme assays with DsaD alone (iii), DsaE alone
(iv), or control lacking DsaD/DsaE (vi) afforded only starting
Figure 2. HPLC analyses of fermentation broths. (A) (i) marꢀ
formycin wildꢀtype producer S. drozdowiczii SCSIO 10141; (ii)
ꢁmfnO mutant strain; (iii) ꢁmfnH mutant strain; (B) (i) negative
control of the host strain S. coelicolor M1152; (ii) S. coelicolor
M1152/07ꢀ6H harboring cosmid 07ꢀ6H with the intact DSA gene
cluster; (iii) S. coelicolor M1152/07ꢀ6HꢀDKO harboring cosmid
07ꢀ6H with dsaD deleted; (iv) S. coelicolor M1152/07ꢀ6HꢀEKO
harboring cosmid 07ꢀ6H with dsaE deleted.
substrate
incubation of
with DsaD alone (iii), DsaE alone (iv), or control lacking
DsaD/DsaE (vi) afforded only starting substrate ꢀalloꢀIle.
LꢀIle. (B) Production of
LꢀIle proceeds only upon coꢀ
LꢀalloꢀIle with DsaD and DsaE (v). Enzyme assays
L
To further confirm the structure of the enzymatic product as
ꢀalloꢀIle, the enzymatic reaction containing ꢀIle, DsaD and
L
L
DsaE was performed on largeꢀscale (10 mL); the product was
purified by chiralꢀphase HPLC followed by heptanesꢀdi(2ꢀ
ethylhexyl)phosphonic acid (7:3) extraction, and finally, silica
gel column chromatography (isocratic elution with butanolꢀ
acetic acidꢀH2O, 4:1:5). The purified product, had the same
In vitro Biochemical Characterization of DsaD/DsaE-
and MfnO/MfnH-catalyzed Interconversion between
and -allo-Ile. To validate the in vitro biochemical activities
of DsaD/DsaE and MfnO/MfnH enzyme pairs in ꢀalloꢀIle
L-Ile
L
L
biosynthesis, we overexpressed and purified DsaD, DsaE,
MfnO and MfnH as soluble Nꢀterminus His6ꢀtagged proteins
from E. coli (Figure S14A). Both purified DsaD and MfnO
proteins were yellow in color. The UV spectrum of DsaD exꢀ
hibited an absorption band at 430 nm (Figure S14B), characꢀ
teristic of PLPꢀdependent aminotransferases such as HspAT.27
Denaturation of DsaD by boiling led to precipitation of the
yellow protein and a colorless supernatant indicating the covaꢀ
lent nature of the PLP association with the catalytic lysine
residue of DsaD; this association most likely involves a Schiff
base linkage.
molecular formula as that of
LꢀalloꢀIle as revealed by
1
HRESIMS (Figure S17) and was found to yield H NMR data
identical to those previously reported for an authentic sample
of
effect data for the newly generated species were also found to
be identical to those of previously reported
ꢀalloꢀIle.3 That
generation of ꢀalloꢀIle is contingent on the presence of both
DsaD and DsaE suggests that these enzymes work synergistiꢀ
cally to produce ꢀalloꢀIle from Ile.
In the enzymatic assays, we found that the DsaD/DsaEꢀ
catalyzed conversion of ꢀIle to ꢀalloꢀIle proceeded in ≈ 60%
yield no matter how long the reaction was allowed to proceed
L
ꢀalloꢀIle (Figure S18).3 Optical rotation and CD cotton
L
L
L
L
L
The activity of the DsaD/DsaE pair using
substrate was first explored. Initially, we incubated
LꢀIle as a possible
LꢀIle with
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 10
1
2
3
4
5
6
7
8
and regardless of moderate changes to reaction conditions
such as pH and buffers. This indicated that the DsaD/DsaEꢀ
catalyzed reaction might be reversible. To test this idea,
alloꢀIle was coꢀincubated with DsaD/DsaE under standard
conditions, and a new peak with a retention time matching that
activity to either mutant enzyme pairing (Figure S25). These
data highlight that K198 in DsaD and K206 in MfnO play
essential, and likely similar, roles in their respective enzymatic
transformations as reported for classꢀIV aminotransferases.
Lꢀ
of the
demonstrating that, indeed, DsaD/DsaE catalyzed a reversible,
bidirectional reaction. Furthermore, the production of ꢀIle
LꢀIle standard was observed (Figure 3B, trace v),
L
9
was observed in the simultaneous presence of DsaD/DsaE,
demonstrating that DsaD/DsaE act synergistically to generate
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
L
ꢀIle from
LꢀalloꢀIle.
We then determined the equilibrium constant for the
DsaD/DsaEꢀcatalyzed bidirectional reaction using equivalent
concentrations of the enzyme pair. The resulting data show
that the Keq for the reaction with LꢀIle as the substrate is 1.37
(Figure S19).
To validate the putative DsaD/DsaE substrate specificity,
Ile and ꢀalloꢀIle were evaluated as potential substrates. Coꢀ
incubation of ꢀIle or ꢀalloꢀIle with the DsaD/DsaE enzyme
pair under standard assay conditions afforded no new species
(Figure S20), indicating the stereospecificity of DsaD/DsaEꢀ
Dꢀ
D
D
D
Figure 4. (A) Activity assays of DsaD/MfnO mutants (iv and vi)
and the MfnH mutant (viii) using LꢀIle as substrate by coupling
with DsaE/MfnH or MfnO, respectively. (B) Activity assays of
the DsaD/MfnO mutants (iv and vi) and the MfnH mutant (viii)
using LꢀalloꢀIle as substrate by coupling with their respective
catalyzed chemistry and validating its relegation to
ꢀalloꢀIle interconversion.
The biochemical activity of the MfnO/MfnH enzyme pair
LꢀIle and
L
DsaE/MfnH or MfnO, respectively.
was evaluated in a fashion similar to that applied to
DsaD/DsaE. As anticipated, the MfnO/MfnH pair displayed a
high degree of functional similarity with the DsaD/DsaE pair
(Figures S15B and S21; Figure S16, C and D). We additionalꢀ
ly carried out experiments using the crossed enzyme pairs
DsaD/MfnH and MfnO/DsaE. Importantly, both "crossed"
pairs retained the ability to catalyze interconversion between
Although the folding architectures for isomerases DsaE and
MfnH are shared with KSIs, these enzymes display low KSI
sequence homology. As validated by biochemical and mutaꢀ
genesis studies, as well as, Xꢀray crystallography, Tyr14,
Asp99 and Asp38 represent the three catalytic residues in the
KSI from Comamonas testosteroni (KSICt); Tyr14 and Asp99
are involved in direct Hꢀbonding interactions with the C3 oxyꢀ
anion of the bound substrate and intermediate. The Asp38 is
an essential catalytic residue that functions as a general base to
intramolecularly transfer a proton from C4β to C6β.25 Seꢀ
quence alignment revealed that, within MfnH, DsaE, and their
homologues, in the same corresponding positions, Tyr14 and
Asp99 of KSICt appear to be conserved (Tyr10 and Asp95 as
seen in MfnH). However, the most important catalytic Asp38
of KSICt is replaced with an arginine (Arg34) in MfnH (Figure
S3A), indicating that MfnH and DsaE might utilize a mechaꢀ
LꢀIle and LꢀalloꢀIle suggesting cross compatibility between the
relevant aminotransferase homologs and isomerase homologs
(Figure S22). It is also notable that two highly homologous
characterized BCATs from primary metabolism, EcBCAT
from E. coli (NCBI gi number: 30749295)21 and MsBCAT
from Mycobacterium smegmatis MC2155 (NCBI gi number:
399233140)22 can also functionally substitute for DsaD (MfnO)
(Figure S23); thus primary metabolism may, in some cases,
play a compensatory role in LꢀalloꢀIle generation.
Site-directed Mutagenesis of DsaD/MfnO and
DsaE/MfnH and a Proposed Catalytic Model for the Inter-
conversion between L-Ile and L-allo-Ile. To probe the cataꢀ
nism different from that of traditional KSIs to carry out LꢀIle
βꢀcarbon epimerization.
lytic residues of aminotransferases DsaD/MfnO and the isoꢀ
merases DsaE/MfnH, a series of siteꢀdirected mutagenesis
experiments were performed. For the aminotransferases
DsaD/MfnO, K198 in DsaD and K206 in MfnO were each
converted to leucine. This mutation was inspired by the fact
that residues homologous to K198 and K206 in DsaD/MfnO
homologs identified by multiple sequence alignments are
known to make Schiff bases with PLP and are responsible for
elimination of the αꢀproton of the PLPꢀlinked amino acid inꢀ
termediate. The resultant K198L DsaD and K206L MfnO varꢀ
iants were overexpressed in E. coli BL21(DE3), purified (Figꢀ
ure S24A) and employed in enzymatic assays coupled with
To interrogate the catalytic residues of isomerases
MfnH/DsaE, we aligned the two proteins with 10 closely reꢀ
lated proteins with undefined functions on the basis of BLAST
results from the NCBI database. The purpose of this alignment
was to unveil potentially conserved amino acid residues (Figꢀ
ure S3B) across the 12 enzymes. In addition to the three resiꢀ
dues Tyr10, Arg34, Asp95 (corresponding to the three catalytꢀ
ic residues Tyr14, Asp38 and Asp99 of KSICt), another 10
polar amino acids Tyr11, Asp15, Phe26, Tyr32, Phe49, Tyr50,
Arg54, His61, Phe79, Arg107 of MfnH, conserved in
MfnH/DsaE and their homologues, were mutated to nonpolar
or nonactive residues. Consequently, 13 MfnH variants were
their respective isomerases DsaE or MfnH using LꢀIle or Lꢀ
alloꢀIle as substrates. Not surprisingly, neither the DsaD nor
the MfnO mutant displayed any activity (Figure 4A, traces iv
and vi and Figure 4B, traces iv and vi). Moreover, supplemenꢀ
tation of each reaction with additional PLP failed to return
ACS Paragon Plus Environment

Page 5 of 10
Journal of the American Chemical Society
1
2
MfnO-Lys206
MfnH- Arg107
O
/DsaD-Lys198
/DsaE- Arg115
3
4
5
6
7
8
9
(S)
O
(S)
O
O
H
OH
N
O
H
NH2
(S)
L-Ile
OH
OH
OH
(S)
OH
N+
N+
N
H
N
MfnO-Lys206
/DsaD-Lys198
MfnO-Lys206
/DsaD-Lys198
MfnH/
DsaE
MfnH/
DsaE
H
O
N
H
O
H
O
+
H
-
-
-
OPO3
HO
OPO3
OPO3
HO
-
-
OPO3
OPO3
N
H
N
N
H
N
N
H+
I
10
11
12
13
14
15
16
17
18
19
MfnO-Lys206
/DsaD-Lys198
-
OPO3
O
O
MfnH- Arg107
/DsaE- Arg115
MfnO-Lys206
/DsaD-Lys198
OH
N
OH
O
MfnO/
DsaD
MfnH/
DsaE
MfnH/
DsaE
N+
H
NH
+HN
HO
(S)
(R)
HO
N
OH
-
H2O
OPO3
+
O-
H
O
NH2
L-allo-Ile
-
-
OPO3
OPO3
N
O
OH
I
N
H
N
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Proposed mechanism of DsaD/DsaE or MfnO/MfnH.
constructed and overexpressed in E. coli BL21(DE3). Eight of
the 13 MfnH variants were readily purified to homogeneity as
soluble proteins (Figure S24B); and two MfnH variants (F26A
and F79A) were overexpressed and partially purified due to
the low yield of the specific proteins in soluble form (Figure
S24C); the remaining three MfnH variants (D15L, H61F and
D95L) were overexpressed but resisted even partial purificaꢀ
tion. Additionally, the D22, H68 and D102 residues of DsaE,
corresponding to D15, H61 and D95 of MfnH, were also muꢀ
tated and the resulting mutant enzymes were partially purified
(Figure S24D). The purified or partially purified mutant proꢀ
teins were employed in enzymatic assays involving MfnO or
the PLPꢀsubstrate complex yielding a quinonoid intermediate.
The third step in the envisioned pathway invokes the use of
Rl07 in MfnH (or R115 in DsaE) to deprotonate the Cβꢀproton;
this is enabled, in part, by acidification of the Cβ proton resultꢀ
ing from earlier Cα deprotonation. Subsequent reprotonation of
the transient enamine (Cα,β olefin) at Cβ on the face opposite
that involved in Cβ deprotonation installs the inverted Cβ steꢀ
reochemistry. Penultimately, the ammonium salt of Lys206
within MfnO (or K198 of DsaD) can reprotonate the Cαꢀ
position on the same stereochemical face as had been used for
Cβ reprotonation by R107 (MfnO) or R115 (DsaD) affording
the Cβ epimerized product LꢀalloꢀIle. Finally, product release
DsaD and using LꢀIle or LꢀalloꢀIle as substrates. HPLC analꢀ
is envisioned to occur by PLP scission and reassociation with
MfnO or DsaD through standard Schiff base formation with
either K206 (MfnO) or K198 (DsaD), respectively (regeneratꢀ
ing I). Notably, the mechanism envisioned here is clearly reꢀ
versible as is reflected by Figure 5 as well as our observations
yses of the reaction mixtures revealed that the R107L MfnH
mutant was completely devoid of enzymatic activity (Figure
4A, trace viii and Figure 4B, trace viii). The R34L MfnH muꢀ
tant, together with the D22L, H68F and D102L DsaE mutants,
displayed sharply reduced enzymatic activity; relative to wildꢀ
type MfnH/DsaE these mutants were only 8% active (Figure
S26 C and D). The remaining eight mutations introduced to
MfnH exerted no apparent effects upon enzymatic activity
(Figure S26 A and B). Consequently, it was realized that R107
in MfnH, and the corresponding R115 in DsaE, represent esꢀ
sential active site residues for this family of isomerases. Imꢀ
portantly, the active sites of MfnH/DsaE, identified on the
basis of these mutagenesis studies, differ from those of KSICt,
thereby demonstrating that MfnH/DsaE represent a new famiꢀ
ly of isomerase.
that LꢀIle serves not only as a substrate but also a product of
Mfn and Dsa enzyme pair chemistries.
Phylogenetic analysis of DsaE/MfnH Unveils a Family of
Related Proteins Potentially Functioning as Isomerases
Biosynthesizing
L-allo-Ile. Phylogenetic analyses of DsaE
and MfnH with related proteins in the Genbank database reꢀ
vealed that they belong to a family of proteins with unknown
functions represented across two domains, the Bacteria and
Archaea (Figure S27). Homologous proteins from the Bacteria
domain include proteins from actinomycetes such as Streptoꢀ
myces, Nocardiopsis, and Salinispora that have a demonstratꢀ
ed record of producing bioactive natural products, but also
contain proteins from actinobacteria such as Actinokineospora,
Dermacoccus, and Actinoplanes, as well as Proteobacteria
such as Burkholderia and Oceanospirillaceae (Osedax symbiꢀ
ont), Deinococcus, and members of Clostridiales whose secꢀ
ondary metabolites have only been marginally investigated. It
is noteworthy that DsaE and MfnH also have a large number
of homologous proteins from Archaea including Halobacteriꢀ
aceae and Thermoplasmatales, secondary metabolites from
which have been largely unexplored. Thus, DsaD/DsaE or
MfnO/MfnH pairs might serve as effective genetic markers
On the basis of these siteꢀdirected mutagenesis experiments,
we envision a rational model for the two enzymeꢀcatalyzed
bidirectional reaction between LꢀIle and LꢀalloꢀIle (Figure 5).
The K206 of MfnO (or K198 of DsaD) initially forms a PLPꢀ
linked Schiff base (Figure 5, PLPꢀE complex I) followed by
enzyme displacement (with retention of close nonꢀcovalent
associations between the enzyme and newly formed substrateꢀ
PLP complex) upon presentation with substrate LꢀIle (or Lꢀ
alloꢀIle). The new PLPꢀsubstrate complex, like the preceding
PLP species, is a Schiff base. Secondly, the terminal amino
group of newly liberated K206 then removes the Cαꢀproton of
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 10
1
2
3
4
5
6
7
8
enabling genome miningꢀbased identification of natural prodꢀ
ucts containing the ꢀalloꢀIle moiety.
were introduced at the following concentrations: apramycin
L
(Apr, 50 ꢁg/mL), kanamycin (Kan, 50 ꢁg/mL), ampicillin
(Amp, 100 ꢁg/mL), spectinomycin (Spec, 50 ꢁg/mL), chloꢀ
ramphenicol (Chl, 25 ꢁg/mL), and trimethoprim (TMP, 50
ꢁg/mL). LꢀalloꢀIle and Ile were purchased from SigmaꢀAldrich
Co. LLC. Other chemicals and solvents were purchased from
standard commercial sources.
ꢀ CONCLUSION
LꢀalloꢀIle is a crucial building block for bioactive natural
products from various microorganisms and exists in the free
form within assorted life forms including homo sapiens; its
functional role in many such scenarios remains the topic of
NMR data were acquired using an Avance 500 MHz specꢀ
trometer instrument (Bruker). High resolution mass spectra
data were determined with a Maxis quadrupoleꢀtimeꢀofꢀflight
mass spectrometer (Bruker). Optical rotation was obtained
with an MCPꢀ500 polarimeter (Anton Paar). The UV spectrum
was recorded with a UVꢀ2600 spectrometer (Shimadzu). CD
spectra were obtained with a Chirascan circular dichroism
spectrometer at 25 °C. The crystal structure data for compound
7 were recorded with an Oxford Xcalibur Onyx Nova singleꢀ
crystal diffractometer with Cu Kα radiation (λ = 1.5418 Å);
the structure was solved by direct methods (SHELXSꢀ97) and
refined using fullꢀmatrix leastꢀsquares difference Fourier techꢀ
niques.
9
speculation. To decipher and understand the origin of Lꢀalloꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Ile we studied the biosynthetic machineries for two groups of
antiꢀinfective natural products, the desotamides and marꢀ
formycins; representatives of both natural product classes conꢀ
tain the unique
LꢀalloꢀIle. Bioinformatics analyses of both
biosynthetic pathways revealed two conserved enzyme pairs
DsaD/DsaE and MfnO/MfnH as candidates involved in generꢀ
ating the LꢀalloꢀIle unit found in the desotamides and marꢀ
formycins. In vivo gene inactivation trials coupled with
metabolomics and in vitro biochemical characterizations reꢀ
vealed that the DsaD/DsaE and MfnO/MfnH enzyme pairs
drive bidirectional interconversion of
directed mutagenesis experiments revealed that this interconꢀ
version proceeds through a PLPꢀlinked ꢀIle (or ꢀalloꢀIle)
LꢀIle and LꢀalloꢀIle. Siteꢀ
Gene Inactivation of mfnO, mfnH, dsaD, and dsaE. The
genes mfnO and mfnH were inactivated in wildꢀtype strain S.
drozdowiczii SCSIO 10141 using λꢀRED mediated PCRꢀ
targeting technology. First, a 1425 bp spectinomycin reꢀ
sistance gene as a PCR template was obtained by digestion of
plasmid pIJ778 by HindIII/EcoRI. Then, each of the specꢀ
tinomycin resistance gene cassettes (oriTꢀaadA) was amplified
by PCR using the obtained template with designed primers
listed in Table S1. Subsequently, the PCRꢀtargeting and conꢀ
jugation processes were carried out as described previously.19
Doubleꢀcrossover mutant strains ꢂmfnO and ꢂmfnH posꢀ
sessing the kanamycinS and spectinomycinR phenotypes were
confirmed by PCR using primers listed in Table S1.
L
L
intermediate; a lysine residue of aminotransferases DsaD or
MfnO removes the αꢀproton, and an arginine residue from the
Cꢀterminus of isomerases DsaE or MfnH, respectively deproꢀ
tonate and subsequently reprotonate the ꢀalloꢀIle βꢀposition.
L
Following this epimerization by either the DsaE or MfnH arꢀ
ginine, the lysine residue of aminotransferase DsaD or MfnO
previously used for Cαꢀdeprotonation, stereospecifically reproꢀ
tonates the αꢀposition, leading to retention of the original Cα
stereochemistry. Phylogenetic analyses validate DsaE and
MfnH as representatives of a new isomerase family within
Bacteria and Archaea domains whose exact functions have not
yet been assigned. In addition to shedding light on the role of
this isomerase family, our work here explains the mysterious
For inꢀframe deletion of dsaD and dsaE within the desotamꢀ
ide gene cluster from S. scopuliridis SCSIO ZJ46, the Superꢀ
Cos 1 vector based cosmid 07ꢀ6H harboring the intact desoꢀ
tamide gene cluster was used. Each of the apramycin reꢀ
sistance gene cassettes (aac(3)IVꢀoriT) was amplified by PCR
using the obtained template with designed primers (with SpeI
restriction site between the 19 or 20 nt matching the right or
left end of the disruption cassette and the 39 nt homologous to
the gene to be inactivated) listed in Table S1. Each apramycin
resistance cassette was introduced into E. coli
BW25113/pIJ790/07ꢀ6H (harboring the intact desotamide
gene cluster) by electroporation to disrupt the homologous
gene region in the cosmid 07ꢀ6H. The mutated cosmids 07ꢀ
6HꢀDKO and 07ꢀ6HꢀEKO were digested with SpeI and then
selfꢀligated. The resulting cosmids 07ꢀ6HꢀDKOꢀIF and 07ꢀ6Hꢀ
EKOꢀIF lost the (aac(3)IVꢀoriT) cassette and were screened
with primers listed in Table S1. For heterologous expression
of cosmids 07ꢀ6HꢀDKOꢀIF and 07ꢀ6HꢀEKOꢀIF in S. coelicolꢀ
or M1152, they were modified by replacing each kanamycin
resistance gene within the SuperCos 1 vector with a
pSET152ABꢀderived fragment. The pSET152AB vector conꢀ
structed in our laboratory using RED recombination approachꢀ
es contains an apramycin resistance gene and elements for
conjugation and site specific recombination (oriT, integrase
gene and intψC31 site).28 The resulting cosmids, designated
07ꢀ6HꢀDKOꢀAB and 07ꢀ6HꢀEKOꢀAB were transferred into S.
coelicolor M1152 by conjugation and stable integration into
biosynthesis of
LꢀalloꢀIle in actinomyces. These findings lay
the foundation for further studies into the origins of
LꢀalloꢀIle
in assorted living systems. The findings disclosed herein also
point the way to discovery (likely through genome mining)
and development of other
with possible therapeutic and/or biomedical applications. The
wellꢀestablished elevated ꢀalloꢀIle levels in humans suffering
LꢀalloꢀIle containing compounds
L
from maple syrup urine disease, in particular, gives cause for
significant excitement in the biomedical and diagnostics arena.
ꢀ EXPERIMENTAL SECTION
Bacterial strains, Plasmids, Culture Conditions, and Re-
agents. The marformycin producer Streptomyces drozdowiczii
SCSIO 1014119 and desotamide producer Streptomyces
scopuliridis SCSIO ZJ4618 have been previously described. S.
coeliecolor M1152 was used as the host strain for heteroloꢀ
gous expression of the desotamide gene cluster and has also
been described. Escherichia coli DH5α was used as the host
for cloning purposes. E. coli ET12567/pUZ8002 was used to
transfer DNA into the Streptomyces from E. coli by conjugaꢀ
tion. E. coli BW25113/pIJ790 was used as the host for
Red/ETꢀmediated recombination for inactivation of dsaD,
dsaE, mfnO and mfnH. All E. coli strains were grown in Luꢀ
ria−Bertani medium at 37°C or 30°C. When used, antibiotics
ACS Paragon Plus Environment

Page 7 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
their chromosomes via attB/attpꢀsiteꢀspecific recombination,
to yield S. coelicolor M1152/07ꢀ6HꢀDKO, S. coelicolor
M1152/07ꢀ6HꢀEKO strains, respectively.
SCSIO 10141, respectively. And the genes for EcBCAT and
MsBCAT were PCR amplified from genomic DNA of E. coli
DH5α and Mycobacterium smegmatis MC2155, respectively.
The gelꢀpurified PCR products were digested with NdeI/EcoRI
(for dsaD, dsaE, mfnO and mfnH) or NheI/EcoRI (for Ecꢀ
BCAT and MsBCAT) and then ligated into the same sites of
pET28a (+) to yield the expression plasmids pET28a/dsaD,
Metabolite Analyses of Wild-type and Mutant strains.
The fermentations and subsequent HPLC analyses of the ferꢀ
mentation broths of wildꢀtype S. drozdowiczii SCSIO 10141
and its mutant strains were carried out as described previousꢀ
ly.19 Fermentations and subsequent HPLC analyses of the ferꢀ
mentation broths of the recombinant heterologous DSA gene
cluster (intact gene cluster or the cluster with dsaD or dsaE
deleted) expression strains were also carried out as described
previously.18
pET28a/dsaE,
pET28a/mfnO,
pET28a/mfnH,
9
pET28a/EcBCAT and pET28a/MsBCAT; these were conꢀ
firmed for correctness by DNA sequencing. The resulting exꢀ
pression plasmids were each transformed into E. coli BL21
(DE3) for protein expression. Expression and purification of
all proteins was carried out following the general procedure
previously described.29 The fractions containing proteins of
interest were pooled, desalted by passage through a PDꢀ10
column (GE Healthcare, USA), and concentrated by ultrafilꢀ
tration (Millipore membrane, 3 kDa cutꢀoff). The purified
proteins were finally stored in 10% glycerol, 50 mM sodium
phosphate buffer, pH 8.0 for further experiments at ꢀ80°C.
Protein concentrations were determined by Bradford assays
using bovine serum albumin as a standard.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Isolation and Structural Elucidation of Compounds 3, 5
and 7. The ꢂmfnH mutant was fermented using 1 L Erlenꢀ
meyer flasks each containing 250 mL production medium as
described previously.19 After fermentation, the culture (8 L)
was centrifuged to yield supernatant and a mycelial cake. The
supernatant was extracted three times with equal volumes of
butanone. The mycelial cake was extracted three times with
acetone. The two extracts were combined and the organic solꢀ
vent was removed under reduced pressure to afford 10.2 g of
residue which was further subjected to normal phase silica gel
column chromatography (100–200 mesh). The residue was
eluted using gradient elution with CHCl3 and CH3OH mixtures
(100:0, 98:2, 96:4, 94:6, 96:4, 90:10, v/v, each solvent combiꢀ
nation used was 100 mL in volume) to afford six fractions
(Fr.1–Fr.6). Fr.3 was further subjected to normal phase silica
gel column chromatography (100–200 mesh), and eluted with
CHCl3 and CH3OH mixtures (100:0, 98:2, 96:4, 94:6, 96:4,
90:10, v/v, each solvent combination in 50 mL volume) afꢀ
fording six subꢀfractions (Fr.3ꢀ1–Fr.3ꢀ6). Then Fr.3ꢀ3 was
separated by semiꢀpreparative HPLC (YMCꢀPack ODSꢀA
column, 250 × 10 mm, 5 ꢁm) eluted with a linear gradient
from 40% to 100% B (solvent A: H2O，solvent B：CH3CN)
over 30 min and using UV detection at 225 nm and 275 nm to
give compounds 5 (6.5 mg) and 7 (8.0 mg).
The siteꢀdirected mutations of DsaD, MfnO, MfnH and
DsaE were performed following the Fast Mutagenesis System
(Transgen, Beijing, China) manual protocol. The expression
and purification of mutant proteins was carried out using the
same procedures described above. All primers used here are
listed in Table S1.
In vitro Enzymatic Assays.
(a) The enzymatic assays investigating standꢀalone DsaD,
DsaE, MfnO, MfnH activities and DsaD/DsaE or
MfnO/MfnHꢀcatalyzed coupled reactions in the presence of αꢀ
ketoglutarate (αꢀKG) and additional exogenous PLP were
performed in a volume of 100 ꢁL, at 30°C for a period of 4 h.
Reaction mixtures contained 5 ꢁM enzyme, 1 mM substrate
Ile or ꢀalloꢀIle, 1 mM αꢀKG, and 0.1 mM PLP in 50 mM
sodium phosphate buffer (pH 8.0).
Lꢀ
L
The ꢂmfnO mutant was similarly fermented (8 L) and proꢀ
cessed. The butanone extract of the supernatant and acetone
extract of the mycelia were combined and subjected to normal
phase silica gel column chromatography (100–200 mesh), and
eluted with CHCl3 and CH3OH mixtures (100:0, 98:2, 96:4,
94:6, 96:4, 90:10, 80:20, 50:50, v/v, 100 mL for each solvent
combination used) to afford six fractions (Fr.1–Fr.8). Fr.3 was
repeatedly subjected to normal phase silica gel column chroꢀ
matography (100–200 mesh), and eluted with CHCl3 and
CH3OH mixtures (100:0, 98:2, 96:4, 94:6, 96:4, 90:10, v/v,
each solvent combination in 50 mL volume) affording six
fractions (Fr.3ꢀ1–Fr.3ꢀ6). Fr.3ꢀ2 and Fr.3ꢀ3 were combined
and further isolated by semiꢀpreparative HPLC (YMCꢀPack
ODSꢀA column, 250 × 10 mm, 5 ꢁm) eluted with a linear graꢀ
dient from 40% to 100% B (A: H2O, B: CH3CN) over 30 min
and with UV detection at 225 nm and 275 nm to yield comꢀ
pounds 3 (5.3 mg)，5 (8.0 mg) and 7 (14.4 mg). Compound 7,
white powder; [α] D25 ꢀ72o, (c 0.98, MeOH); HRꢀESIꢀMS (m/z):
873.5086 [M + H]+ (calc. for C43H69N8O11, 873.5080).
(b) Assays for determination of the necessity of the possible
cofactor αꢀKG and additional cofactor PLP for the DsaD/DsaE
or MfnO/MfnHꢀcatalyzed interconversion between
LꢀIle and
LꢀalloꢀIle, each cofactor was omitted from the reaction mixture
containing 5 ꢁM aminotransferase DsaD (or MfnO), 5ꢁM
isomerase DsaE (or MfnH), 1 mM αꢀKG, 0.1 mM PLP and 50
mM sodium phosphate buffer (pH 8.0) at a time.
(c) The enzymatic assays assessing standꢀalone DsaD, DsaE,
MfnO, MfnH activities or the DsaD/DsaE, MfnO/MfnH,
DsaD/MfnH and MfnO/DsaEꢀcatalyzed coupled reactions in
the absence of αꢀKG and additional exogenous PLP were perꢀ
formed in a volume of 100 ꢁL, 30oC for 4 h. Each reaction
mixture contained 5 ꢁM enzyme/s, and 1 mM substrate
LꢀIle
or ꢀalloꢀIle in 50 mM sodium phosphate buffer (pH 8.0).
L
These reaction conditions were used as the standard assay
conditions for DsaD/DsaE activity assays.
(d) Assays used to investigate substrate specificities for
DsaD/DsaE or MfnO/MfnH were performed in a volume of
o
100 ꢁL, 30 C for 4 h. Each reaction mixture contained 5 ꢁM
Cloning, Expression, and Purification of DsaD, DsaE,
MfnO, MfnH, EcBCAT, MsBCAT, DsaD mutant, MfnO
mutant, MfnH mutants and DsaE mutants. The genes for
dsaD, dsaE, mfnO, mfnH were PCR amplified from genomic
DNA of S. scopuliridis SCSIO ZJ46 and S. drozdowiczii
aminotransferase DsaD or MfnO, 5 ꢁM isomerase DsaE or
MfnH, and 1 mM substrate
DꢀIle or DꢀalloꢀIle in 50 mM sodiꢀ
um phosphate buffer (pH 8.0).
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 10
1
2
3
4
5
6
7
8
(e) Assays to investigate the activity of aminotransferase
L
-allo-Ile standard: 1H NMR (500 MHz, D2O), δ 0.86 (3H,
t, J = 8.0 Hz), 0.84 (3H, d, J = 7.0 Hz), 1.20 ~1.40 (2H, m),
DsaD/MfnO mutants as coupled with their respective isomerꢀ
ases DsaE or MfnH were performed in a volume of 100 ꢁL,
30°C for 4 h. Each reaction mixture contained 5 ꢁM DsaD or
MfnO mutant, 5 ꢁM isomerase DsaE or MfnH, and 1 mM
25
1.98 (1H, m), 3.64 (1H, m). [α]
+23.2o (c 0.1, aq HCl, pH
D
2.5). CD [θ]202 +333.3o (c 0.1, aq HCl, pH 2.5).
Equilibrium Constant of DsaD/DsaE. Assays to establish
equilibrium constants were performed in 50 ꢁL sodium phosꢀ
phate buffer (50 mM, pH8.0) containing 2.5 ꢁM DsaD, 2.5
substrate
(pH 8.0).
LꢀIle or LꢀalloꢀIle in 50 mM sodium phosphate buffer
ꢁM DsaE, 0.1 mM PLP and 5 mM of substrate LꢀIle or Lꢀalloꢀ
(f) The enzymatic assays to evaluate activities of
MfnH/DsaE mutants as coupled with their respective amiꢀ
notransferase pair MfnO or DsaD were performed in a volume
of 100 ꢁL, 30°C for 4 h. The reaction mixtures contained 5
ꢁM MfnH or DsaE mutant, 5 ꢁM aminotransferase MfnO or
9
Ile at 30°C in triplicate. Each reaction was stopped by addition
of 150 ꢁL of methanol at times of 2.5, 5, 10, 15, 25, 30, 40, 60,
120, 180, and 240 min. Subsequent treatment of reaction mixꢀ
tures and HPLC analyses were conducted using the same
methods applied to the standard DsaD/DsaE activity assays.
The final concentrations of substrates and products were calꢀ
culated on the basis of fraction peak integrals. The Keq was
calculated using the equation Keq = ([product]/[substrate])
when the reaction achieved equilibrium.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
DsaD, 0.1 mM PLP and 1 mM substrate
50 mM sodium phosphate buffer (pH 8.0).
LꢀIle or LꢀalloꢀIle in
(g) Assays to investigate the activity of EcBCAT and
MsBCAT coupled with the isomerases DsaE or MfnH were
o
performed in a volume of 100 ꢁL, 30 C for 4 h. The reaction
mixtures contained 5 ꢁM MsBCAT or EcBCAT, 5 ꢁM isoꢀ
merases DsaE or MfnH, and 1 mM substrate
in 50 mM sodium phosphate buffer (pH 8.0).
LꢀIle or LꢀalloꢀIle
ꢀ ASSOCIATED CONTENT
Supporting Information.
Supporting figures and tables. This material is available free of
charge via the Internet at http://pubs.acs.org.
All reactions were quenched by addition of 200 ꢁL MeOH.
After incubation for 20 min at room temperature, the precipiꢀ
tated proteins were removed by centrifugation (12,000 × g for
20 min at room temperature). Supernatants were dried in vacꢀ
uo and the resulting residues dissolved in 40 ꢁL 2 mM CuSO4
aqueous solution and 25 ꢁL of each was subjected to chiralꢀ
phase HPLC analyses performed on an Agilent Technologies
1260 Infinity system using an MCI GEL CRS10W column
(Mitsubishi, 50 × 4.6 mm, 3 ꢁm) eluted with 2 mM CuSO4
aqueous solution over the course of 30 min at a flow rate of 1
mL/min and with UV detection at 254 nm.
ꢀ AUTHOR INFORMATION
Corresponding Author
jju@scsio.ac.cn
Author Contributions
Notes
The authors declare no competing financial interest.
Enzymatic Synthesis and Isolation of
duction and isolation of ꢀalloꢀIle, a largeꢀscale enzymatic
reaction was performed in 10 mL sodium phosphate buffer (50
mM, pH 8.0) containing 2 mM ꢀIle, 20 ꢁM DsaD, 20 ꢁM
L-allo-Ile. For proꢀ
L
ꢀ ACKNOWLEDGMENT
L
This study was supported, in part, by the National Natural Science
Foundation of China (31400072, 81425022, 41306146), the Naꢀ
tional High Technology Research and Development Program of
China (2012AA092104), the Programs of Chinese Academy of
Sciences (XDA11030403, KGZDꢀEWꢀ606), and a special finanꢀ
cial fund for innovative developments of the Marine Economic
Demonstration Project (GD2012ꢀD01ꢀ001). Additionally, we
thank the analytical facility center (Ms. Aijun Sun, Dr. Zhihui
Xiao and Mr. Chuanrong Li) of the South China Sea Institute of
Oceanology for recording NMR and MS data.
DsaE at 30°C for 4 h. The reaction was quenched by addition
of 20 mL MeOH to precipitate the proteins and the solvent
was removed in vacuo. The resulting residues were dissolved
in 2 mM CuSO4 aqueous solution, and purified by chiralꢀphase
HPLC using an MCI GEL CRS10W column (Mitsubishi, 50 ×
4.6 mm, 3 ꢁm) eluted with 2 mM CuSO4 aqueous solution at a
flow rate of 1 mL/min and employing UV detection at 254 nm.
The fraction with a retention time of 13 min was pooled, ethꢀ
ylene diamine tetraacetic acid (EDTA) was added to a final
concentration of 2 mM and the pH of the subsequent mixture
adjusted to 4.0. The mixture was extracted using heptaneꢀdi(2ꢀ
ethylhexyl)phosphonic acid (7:3) twice, and the combined
organic layer was further extracted by equal volumes of 5%
HCl solution twice. The combined aqueous layer was evapoꢀ
rated to dryness, and finally purified by silica gel column
chromatography (isocratic elution with the upper layer of buꢀ
ꢀ REFERENCES
(1) Van DamꢀBakker A. W. Nature 1958,181, 116.
(2) Ali, H. S.; Alhaj, O. A.; AlꢀKhalifa, A. S.; Brückner, H. Amino
Acids 2014, 46, 2241.
(3) Ikai, K.; Takesako, K.; Shiomi, K.; Moriguchi, M.; Umeda, Y.;
Yamamoto, J.; Kato, I.; Naganawa, H. J. Antibiot. (Tokyo) 1991, 44,
925.
(4) Chen, Z.; Song, Y.; Chen, Y.; Huang, H.; Zhang, W.; Ju, J. J. Nat.
Prod. 2012, 75, 1215.
(5) Capon, R. J.; Skene, C.; Stewart, M.; Ford, J.; O'Hair, R. A.; Wilꢀ
liams, L.; Lacey, E.; Gill, J. H.; Heiland, K.; Friedel, T. Org. Biomol.
Chem. 2003, 1, 1856.
(6) Kogen, H.; Kiho, T.; Nakayama, M.; Furukawa, Y.; Kinoshita, T.;
Inukai, M. J. Am. Chem. Soc. 2000, 122, 10214.
(7) Song,Y.; Li, Q.; Liu, X.; Chen, Y.; Zhang, Y.; Sun, A.; Zhang, W.;
Zhang, J.; Ju, J. J. Nat. Prod. 2014, 77, 1937.
tanolꢀacetic acidꢀH2O (4:1:5)), yielding pure
white solid.
LꢀalloꢀIle as a
Enzymatically produced
L
-allo-Ile: HRꢀESIꢀMS [M + H]+
1
=132.1038 (calc. for C6H14NO2, 132.1019); H NMR (500
MHz, D2O), δ 0.86 (3H, t, J = 7.5 Hz), 0.83 (3H, d, J = 7.0
25
Hz), 1.19 ~1.40 (2H, m), 1.96 (1H, m), 3.62 (1H, m). [α] D
+
23.4o (c 0.0575, aq HCl, pH 2.5). CD [θ]202 + 165.7° (c 0.0575,
aq HCl, pH 2.5).
ACS Paragon Plus Environment

Page 9 of 10
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
(8) Zhou, X.; Huang, H.; Li, J.; Song, Y.; Jiang, R.; Liu, J.; Zhang, S.;
Hua, Y.; Ju, J. Tetrehedron 2014, 70, 7795.
(9) Vaillancourt, F. H.; Yeh, E.; Vosburg, D. A.; O'Connor, S.E.;
Walsh, C. T. Nature 2005, 436, 1191.
(10) Schadewaldt, P.; BodnerꢀLeidecker, A.; Hammen, H. W.;
Wendel, U. Pediatr. Res. 2000, 47, 271.
(11) Schwarz, E. L.; Roberts, W. L.; Pasquali, M. Clin. Chim. Acta.
2005, 354, 83.
(12) Sowell, J.; Pollard, L.; Wood, T. J. Sep. Sci. 2011, 34,631.
(13) Strauss, K. A.; Puffenberger, E. G.; Morton, D. H. Maple Syrup
Urine Disease. In GeneReviews; Pagon, R. A., Adam, M. P., Ardingꢀ
er, H. H., Bird, T. D., Dolan, C. R., Fong, C. T., Smith, R. J. H., Steꢀ
phens, K., Eds; University of Washington: Seattle, 2006; pp 71ꢀ88.
[updated 2013 May 09].
(14) Wendel, U.; Langenbeck, U.; Seakins, J. W. Pediatr. Res. 1989,
25, 11.
(15) Schadewaldt, P.; Wendel, U. Pediatr. Res. 1987, 22, 591.
(16) Schadewaldt, P.; Wendel, U. Biochem. Med. Metab. Biol. 1989,
41, 105.
(17) Schadewaldt, P.; BodnerꢀLeidecker, A.; Hammen, H. W.;
Wendel, U. Clin. Chem. 1999, 45, 1734.
(18) Li, Q.; Song, X.; Qin, X.; Zhang, X.; Sun, A.; Ju, J. J. Nat. Prod.
2015, 78, 944.
(19) Liu, J.; Wang, B.; Li, H.; Xie, Y.; Li, Q.; Qin, X.; Zhang, X.; Ju,
J. Org. Lett. 2015, 17, 1509.
(20) Chen, C. D.; Lin, C. H.; Chuankhayan, P.; Huang, Y. C.; Hsieh,
Y. C.; Huang, T. F.; Guan, H. H.; Liu, M. Y.; Chang, W. C.; Chen, C.
J. J. Bacteriol. 2012, 194, 6206.
(21) Okada, K.; Hirotsu, K.; Sato, M.; Hayashi, H.; Kagamiyama, H.
J. Biochem. 1997, 121, 637.
(22) Castell, A.; Mille, C.; Unge, T. Acta. Crystallogr. D Biol. Crysꢀ
tallogr. 2010, 66, 549.
(23)Yennawar, N. H.; Islam, M. M.; Conway, M.; Wallin, R.; Hutson,
S. M. J. Biol. Chem. 2006, 281, 39660.
(24) Soding, J.; Biegert, A.; Lupas, A. N. Nucleic. Acids. Res. 2005,
33, W244.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(25) Ha, N. C.; Choi, G.; Choi, K. Y.; Oh, B. H. Curr. Opin. Struct.
Biol. 2001,11, 674.
(26) Gust, B.; Challis, G. L.; Fowler, K.; Kieser, T.; Chater, K. F.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 1541.
(27) Haruyama, K.; Nakai, T.; Miyahara, I.; Hirotsu, K.; Mizuguchi,
H.; Hayashi, H.; Kagamiyama, H. Biochemistry 2001, 40, 4633.
(28) Zhang, Y.; Huang, H.; Chen, Q.; Luo, M.; Sun, A.; Song, Y.; Ma,
J.; Ju, J. Org. Lett. 2013, 15, 3254.
(29) Gui, C.; Li, Q.; Mo, X.; Qin, X.; Ma, J.; Ju, J. Org. Lett. 2015, 17,
628.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 10
1
2
3
4
Table of Contents artwork
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment