X. He et al.
a transamination reaction is necessary. A homologue of BlsH,
MilD, was identified in the MIL cluster, and both enzymes were
similar to DegT/DnrJ/EryC1/StrS (VI) aminotransferases that use
l-Glu as amine donor and various keto sugars as amine accept-
ors. MilD also showed weak homology to CDP-6-deoxy-l-threo-
d-glycero-4-hexulose-3-dehydrase (22% identity over the full
length), involved in the biosynthesis of 3,6-dideoxyhexoses.[23]
To address the function of MilD, milD was disrupted by double
crossover and the resulting mutant failed to produce the char-
acteristic halo circle caused by MIL on the bioassay plate (Fig-
ure S6A). No intermediates were identified in the HPLC profile
for the fermentation broth extract of the mutant (Figure S6B).
Formation of the 5-guanidino-2,4-dihydroxyvalerate side chain:
In the 13C-Arg incorporation experiment, we established that
the guanidino side chain was derived from l-Arg and that the
attachment to the pyranoside ring was coupled with decarbox-
ylation (C6 of the hexose). The presence of the tertiary hydroxy
group in the side chain suggested that l-Arg was converted
first into a-keto-d-guanidino-valerate, which could be carried
out by an aminotransferase. In the MIL biosynthesis gene clus-
ter, there are two ORFs, MilM and MilD, predicted to be amino-
transferases. As it was believed that MilD catalyzed the transa-
mination of the pyranoside ring, it was proposed that the
other transamination reaction was catalyzed by MilM. MilM
belongs to the aspartate aminotransferase superfamily, the
common amine acceptor of which is 2-ketoglutarate. The
newly formed l-Glu might then serve as the amino donor in
other transamination reactions, such as the reaction catalyzed
by MilD.
duction of LL21 (Figure S16B) was reduced to an extremely
low level, which indicated that milK might function as a regula-
tory gene. Moreover, we analyzed every new peak appearing
in the HPLC analysis of every mutant fermentation broth by
LC-MS. Unfortunately, no intermediate associated with MIL bio-
synthesis was detected. According to the in silico analysis, the
genes in the MIL biosynthesis gene cluster can be classified as
biosynthesis genes (milA, B, C, D, G, H, I, J, M and N), regulatory
genes (milO and K), resistance genes (milE, P and Q), and
a gene of unknown function (milF).
Conclusions
In our previous work, we demonstrated that MilA and MilB
were responsible for the catalysis of the initial steps in the bio-
synthesis pathway of MIL. The work presented here revealed
that there were 14 additional genes involved in this pathway.
The incorporation experiments with 13C6-labeled l-Arg and
[guanidino-13C]-4-hydroxyarginine established the origin of the
side chain, the timing of hydroxylation, and the fact that the
carbon in the CÀC bond that formed between the carboxy-
guanidine-butyl group and the dihydropyranoside was derived
from l-arginine. The formation of HM-CGA was clearly revealed
by in vivo and in vitro assays of MilC, as well as of MilA and
MilB. Although the function of MilG, the only oxidoreductase
out of the six homologous genes shared by the BLS and MIL
biosynthesis gene clusters, was not very clear, the HM-CGA
accumulated in the milG disruption mutant implied that MilG
was involved in the formation of the C2=C3 double bond of
the pyranoside ring. The other genes were individually disrupt-
ed through double crossover. However, except for LL2 and LL8
(the DmilA and DmilG mutants), none of the other mutants
accumulated any intermediate. The combined data with bio-
informatic analysis allow us to propose the whole biosynthetic
pathway of MIL (Scheme 2).
MilN is similar to dihydrodipicolinate synthetases and might
therefore couple the a-keto-d-guanidine-valerate with the nu-
cleoside. Both milM and milN were disrupted by double cross-
over; the consequent lack of MIL production indicates they are
indispensable in the pathway (Figure S7A and S8A). Unfortu-
nately, no intermediates were detected either by bioassay or
by HPLC (Figures S7B and S8B).
In the MIL biosynthesis gene cluster, only MilG and MilJ are
predicted to be oxidases. However, MilG shows high similarity
with BlsE, an essential gene in the BLS gene cluster. BLS does
not have a hydroxylated arginine side chain, so it is not very
likely that MilG catalyzes the insertion of the hydroxy group.
On the other hand, no homologue of MilJ, which showed
weak homology (identities=29/100, positives=44/100) to
Enoyl-CoA hydratase (ECH), which facilitates the syn-addition
of a water molecule across the double bond of a trans-2-enoyl-
CoA thioester to form a b-hydroxyacyl-CoA thioester,[23,24] was
present in the BLS biosynthesis gene cluster. MilJ is thus as-
sumed to catalyze the hydroxylation of the guanidine-hydroxy-
valerate side chain. Although no intermediate was accumulat-
ed in the milJ disruption mutant, MilJ was found to be essen-
tial for the MIL pathway (Figure S9).
Experimental Section
Strains and culture conditions: Key strains and plasmids are listed
in Table S3. Sv remofaciens ZJU5119 and its derivatives were grown
at 308C on SFM agar plates for sporulation or in TSBY [10.3%, TSB
liquid medium supplemented with sucrose (10.3%, w/v) and yeast
extract (1%, w/v)] for growth of mycelia. Conjugation was per-
formed as described by Kieser et al.,[25] and exoconjugants were
cultivated in SFM agar plates containing appropriate antibiotic
[spectinomycin (150 mgmLÀ1) or thiostrepton (15 mgmLÀ1)] or in
TSBY (10.3%) containing appropriate antibiotic [spectinomycin
(50 mgmLÀ1) or thiostrepton (5 mgmLÀ1)]. For secondary metabolite
analysis, the seed culture was prepared by inoculating a loop of
stock culture into a 250 mL baffled flask containing seed medium
[glucose (20 g), soybean flour (30 g), yeast extract (5.0 g), (NH4)2SO4
(1.0 g), MgSO4·7H2O (0.5 g), CaCO3 (3.0 g), 25 mL per L], and the
flask was incubated in a rotary shaker for 40 h (308C, 200 rpm).
Seed culture (5 mL) was then transferred into a 500 mL baffled
flask containing fermentation medium [glucose (80 g), soybean
flour (40 g), (NH4)2SO4 (1.0 g), MgSO4·7H2O (0.5 g), CaCO3 (10.0 g),
FeSO4 (0.05 g), K2HPO4 (0.4 g), 50 mL per 1 L] for another six days
under the same conditions. For LC-MS analysis of the production
Analysis of other genes in the MIL biosynthesis gene cluster:
Open reading frames other than milA, B, C, G, J, M, and N were
disrupted individually and analyzed by HPLC and bioassay to
identify their involvement in the MIL biosynthesis (Figures
S10–S17). Except for LL21 (milK disruption mutant), all the
other mutants abolished the production of MIL. The MIL pro-
&6
&
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
ÝÝ
These are not the final page numbers!
Wu, Jun
