Angewandte
Communications
Chemie
is likely responsible for appending a methyl group on the C4
of the indolyl moiety during nosiheptide maturation.[4]
The fact that NosN is a methyltransferase is further
supported by the observation that this enzyme shares
significant sequence similarity with other known methyltrans-
ferases, such as PbtM2 and PbtM3, which are involved in
GE2270 biosynthesis,[7] and YtkT, which is involved in
yatakemycin biosynthesis.[8] NosN, YtkT, PbtM2, PbtM3,
and some other putative methyltransferases are homologous
to the coproporphyrinogen III oxidase HemN (Figure 1B),
and are classified as class C RSMTs.[2a,c] Notably, HemN
catalyzes two sequential steps of oxidative decarboxylation of
coproporphyrinogen III,[9] a reaction that is totally unrelated
to methylation. Although the methyltransferase activity of
YtkT has been successfully reconstituted in vitro by Tang and
co-workers,[8] and ChuW has very recently been reconstituted
by Lanzilotta and co-workers,[10] the catalytic mechanisms of
the class C RSMTs remain largely unknown.
To interrogate the function and mechanism of NosN, the
nosN gene was amplified from the genomic DNA of S.
actuosus and expressed in Escherichia coli with an N-terminal
hexa-histidine tag, and the protein was purified by Ni2+
affinity chromatography under strictly anaerobic conditions.
After chemical reconstitution of the [4Fe-4S] cluster, followed
by gel-filtration, the protein was found to contain 4.5 Æ
0.2 molFe and 4.8 Æ 0.3 molS per mol protein. UV/Vis
spectroscopy showed that the protein solution had a broad
absorption around 415 nm (Figure S4), a feature character-
istic of [4Fe-4S]-containing proteins. Analysis of the reaction
mixture containing SAM, sodium dithionite, and the recon-
stituted NosN showed that 5’-deoxyadenosine (dAdoH) was
produced in the assay mixture (Figure 2A, trace ii), thus
suggesting that NosN is a radical SAM enzyme.
Figure 2. Mechanistic investigation of NosN. A) HPLC analysis of the
NosN assay mixture, using a supernatant with boiled NosN as
a negative control (i), and reconstituted NosN with the other required
components in the absence (ii) or presence (iii) of 4. Because of the
low yield and the similar retention time to that of dAdoH, tAdoH is
not visible in this analysis. B) Structures of the indolyl compounds in
relation to our biochemical analysis. C) MS spectra of 5 (i) and 5d (ii)
produced in the NosN-catalyzed reaction with SAM or d3-SAM,
respectively. D) Chemical structures of the MTA and tAdoH produced
in the NosN reaction. E) LC–MS analysis of NosN reaction mixtures,
showing the extracted ion chromatograms (EICs) of [M+H]+ =284.1
(corresponding to tAdoH) for the control reaction with boiled NosN
(i), the NosN reaction (ii), and a synthetic tAdoH standard (iii).
A possible substrate of NosN is 3-methyl-2-indolic acid
(MIA, 3), a nosiheptide biosynthetic intermediate produced
from l-tryptophan by the radical SAM enzyme NosL (Fig-
ure 1A).[11] To test whether MIA is the NosN substrate, we
performed the NosN assays with MIA and carefully examined
the assay mixture by liquid chromatography high resolution
mass spectrometry (LC–HRMS). No methylated product was
found in the assay mixture, thus suggesting that MIA is
probably not the substrate of NosN. Another possible
substrate is 2, which is produced by the nosN-knockout
mutant (Figure 1A). We obtained the purified compound 2
from the nosN-knockout mutant of a nosiheptide high-
producing strain (Methods and Figure S1–S3 in the Support-
ing Information), and performed the assays with 2 and the
other required components. Again, no methylated product
was observed in the assay mixtures, suggesting that 2 is likely
not the NosN substrate but rather an off-pathway product
produced by the mutant strain.
Because the indolic acid moiety in nosiheptide is attached
to a cysteine thiol as a thioester (Figure 1A), it is possible that
MIA is first converted into an activated form before being
methylated by NosN and incorporated into the nosiheptide
scaffold.[4] Thioester linkage to the cysteamine group of
a phosphopantetheinyl cofactor is arguably the most common
strategy used by nature for the activation of a carboxylate,
and in biochemical studies, N-acetylcysteamine (SNAC)
thioesters (Figure 2B) have been widely used as surrogates
of phosphopantetheinylated substrates.[12] We therefore syn-
thesized SNAC thioester 4 (Figure 2B) and used it as
a potential substrate in the NosN assays. In this analysis,
a product with a protonated molecular ion at m/z = 291.1161
in LC–HRMS analysis was observed (Figure 2C and Fig-
ure S5), which is absent in all the three negative control
reactions (Figure S6). The suggested molecule formula
C15H18N2O2S ([M+H]+ calc. 291.1167, 2.0 ppm error) is
consistent with the methylated product 5 (Figure 2B), the
identity of which was further supported by comparative
HRMS/MS analysis (Figure S7). To further confirm the
production of 5 in the reaction, we treated the deproteinized
fraction of the assay mixture with NaOH, which leads to
hydrolysis of any thioesters, and the resulting mixture was
subjected to LC–HRMS analysis. As expected, a product
consistent with 3,4-dimethyl-2-indolic acid (6) was observed
([MÀH]À calc. 188.0711, obs. 188.0710, 0.5 ppm error), and its
identity was further supported by comparison with two
synthetic dimethylindolic acid isomers (Figure S8). These
results clearly indicate that NosN is a novel radical SAM
methyltransferase that is responsible for appending a methyl
group onto C4 of the indolyl moiety in nosiheptide biosyn-
thesis.
We next synthesized S-adenosyl-l-[methyl-2H3]-methio-
nine (d3-SAM) and used it in the NosN assay. LC–HRMS
2
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2017, 56, 1 – 6
These are not the final page numbers!