Menaquinone biosynthesis: Formation of aminofutalosine requires a unique radical SAM enzyme

Article abstract of DOI:10.1021/ja408594p

Menaquinone (MK, vitamin K₂) is a lipid-soluble molecule that participates in the bacterial electron transport chain. In mammalian cells, MK functions as an essential vitamin for the activation of various proteins involved in blood clotting and

Full text of DOI:10.1021/ja408594p

Communication
pubs.acs.org/JACS
Menaquinone Biosynthesis: Formation of Aminofutalosine Requires
a Unique Radical SAM Enzyme
†
†
‡
,†
Nilkamal Mahanta, Dmytro Fedoseyenko, Tohru Dairi, and Tadhg P. Begley*
†Department of Chemistry, Texas A&M University, College Station, Texas 77843, United States
‡Graduate School of Engineering, Hokkaido University, Sapporo 060-8628, Japan
*
S Supporting Information
ABSTRACT: Menaquinone (MK, vitamin K ) is a lipid-
2
soluble molecule that participates in the bacterial electron
transport chain. In mammalian cells, MK functions as an
essential vitamin for the activation of various proteins
involved in blood clotting and bone metabolism. Recently,
a new pathway for the biosynthesis of this cofactor was
discovered in Streptomyces coelicolor A3(2) in which
chorismate is converted to aminofutalosine in a reaction
catalyzed by MqnA and an unidentified enzyme. Here, we
reconstitute the biosynthesis of aminofutalosine and
demonstrate that the missing enzyme (aminofutalosine
synthase, MqnE) is a radical SAM enzyme that catalyzes
the addition of the adenosyl radical to the double bond of
3
-[(1-carboxyvinyl)oxy]benzoic acid. This is a new
reaction type in the radical SAM superfamily.
enaquinone (8, vitamin K) is a lipid-soluble quinone
Figure 1. Futalosine-dependent pathway to menaquinone, 8.
M
that shuttles electrons between membrane-bound redox
enzymes in the respiratory chain in bacteria. It is also an
1
essential vitamin in humans, where it functions as a cosubstrate
in the carboxylation of glutamic acid residues in proteins
2
3
involved in blood clotting and bone formation. Until recently,
it was believed that the bacterial biosynthesis of menaquinone
4
,5
from chorismate was well understood. The observation that
some strains of Streptomyces (e.g., S. coelicolor A3(2) and S.
avermitilis) do not possess orthologs of the known
menaquinone biosynthetic genes, however, suggested the
existence of an alternative pathway. Using a combination of
bioinformatics, labeling studies, gene deletions, and in vitro
biochemical assays, the general features of this new pathway in
6
−8
S. coelicolor A3(2) were recently determined
(Figure 1). In
this biosynthetic pathway, MqnA (SCO4506) plays an
undefined role in catalyzing the coupling of chorismate (1),
adenosine (2), and pyruvate (3) to form aminofutalosine (4)
6
using unknown chemistry. Aminofutalosine (4) is then
converted to compound 5 by aminofutalosine hydrolase
Figure 2. Three possible pathways for the formation of amino-
futalosine (4) from chorismate (1).
(
MqnB) or by a combination of aminofutalosine deaminase
9
,10
11
and MqnB. Cyclization of 5 by MqnC gives compound 6,
pyrophosphate dependent ketoacid decarboxylase, while path C
requires a radical SAM enzyme. To determine which path was
used, we analyzed the gene neighborhood of the identified
menaquinone biosynthetic genes for additional genes encoding
which is then converted to 1,4-dihydroxy-6-naphthoic acid (7)
by MqnD. Bioinformatics analysis suggests that 7 is converted
to menaquinone 8 in S. coelicolor A3(2) in reactions catalyzed
by SCO4491 (prenylation), SCO4556 (methylation), and
SCO4490 or SCO4492 (decarboxylation).
Three possible pathways for the formation of 4 were
considered (Figure 2). Paths A and B require a thiamin
Received: August 19, 2013
Published: October 1, 2013
©
2013 American Chemical Society
15318
dx.doi.org/10.1021/ja408594p | J. Am. Chem. Soc. 2013, 135, 15318−15321

Journal of the American Chemical Society
Communication
Figure 3. HPLC analysis of the reaction catalyzed by MqnA. Blue
trace: chromatogram of the full reaction mixture (MqnA +
chorismate). Red trace: chromatogram of the control reaction mixture
lacking MqnA. The inset shows the structure of the identified MqnA
reaction product, 12, which was also characterized by 1D and 2D
NMR (Figures S2−S6) and MS analysis (Figure S7).
Figure 6. Mechanistic proposal for the MqnE-catalyzed formation of
aminofutalosine (4).
Figure 4. UV−visible spectra of MqnE (120 μM in each case). Blue
trace: protein “as-isolated”. Red trace: dithionite (excess) treatment of
the “as-isolated” sample. Green trace: aerobic oxidation of the “as-
isolated” sample.
Figure 7. (A) Adenosyl radical addition to a double bond generates a
new, potentially versatile biosynthon enabling chemistry at both C1
and C5 of ribose. (B) Griseolic acid biosynthesis presents a possible
example of this chemistry.
named as mqnE), encoding a putative radical SAM enzyme that
frequently clustered with menaquinone biosynthetic genes
(SCO4494 in S. coelicolor, TTHA0804 in Thermus thermophilus
HB8). As we were unable to find a similarly clustered putative
ketoacid decarboxylase gene, we focused our attention on
experimentally testing path C as the most likely route to 4. In
this Communication, we describe the successful reconstitution
of aminofutalosine formation and demonstrate that it occurs by
a new motif in radical SAM enzymology.
Figure 5. HPLC analysis of the MqnE reaction mixture. Blue trace:
chromatogram of the full reaction mixture (MqnE + 12 + SAM +
dithionite). Red trace: chromatogram of the control reaction mixture
lacking MqnE. Green trace: chromatogram of the control reaction
mixture lacking substrate 12, showing low levels of the uncoupled
production of 5′-deoxyadenosine. No signals at 20.2 and 19.5 min
were obtained in the control reactions individually lacking SAM or
dithionite (data not shown). The inset shows the structure of the
identified MqnE reaction product, 4, which was also characterized by
Genetic studies previously identified MqnA as essential for
6
aminofutalosine formation. To identify the reaction catalyzed
1
D and 2D NMR (Figures S19−S23) and MS analysis (Figure S24).
by MqnA, the mqnA gene from T. thermophilus HB8
(
TTHA0803) was amplified by PCR, cloned into the p28nT
putative ketoacid decarboxylases or a putative radical SAM
enzyme using the SEED database (http://theseed.uchicago.
edu/FIG/). This analysis yielded one promising gene (now
expression vector, overexpressed in E. coli BL21 (DE3), and
purified by Ni-NTA affinity chromatography (Figure S1).
Incubation of chorismate with MqnA resulted in the formation
1
5319
dx.doi.org/10.1021/ja408594p | J. Am. Chem. Soc. 2013, 135, 15318−15321

Journal of the American Chemical Society
Communication
1
5−17
of a new product, which was identified as 3-[(1-carboxyvinyl)-
oxy]benzoic acid (12) by NMR and MS analysis (Figures 3 and
where all previously characterized enzymes
except for
18
diphthamide synthase use the adenosyl radical to abstract a
hydrogen atom from the cosubstrate or from an active-site
glycine. The MqnE-catalyzed reaction represents an efficient
way to add adenosyl or ribose units to a double bond to
generate an intermediate with rich biosynthetic potential, and
other examples of this efficient catalytic motif are likely to
emerge (Figure 7A). Griseolic acid biosynthesis presents one
possible example in which intermediate 22 may be formed by
S2−S7). The steady-state kinetic parameters for MqnA are k
cat
−1
=
1.27 ± 0.02 s , K = 161.6 ± 7.1 μM, and k /K = 7.8 ×
m cat m
3
−1
−1
10 s
M
(Figure S17). To the best of our knowledge, MqnA
is the first characterized enzyme catalyzing the dehydration of
chorismate to give 12.
To test the hypothesis that aminofutalosine (4) is formed via
path C (Figure 2), mqnE (TTHA0804) from T. thermophilus
HB8 was PCR amplified, cloned into the pQE-30 expression
vector, and overexpressed in E. coli M15 (pREP4). The
2
0,21
addition of the adenosyl radical to fumaric acid
7B).
(Figure
resulting His -tagged MqnE protein was purified by Ni-NTA
6
ASSOCIATED CONTENT
Supporting Information
affinity chromatography under anaerobic conditions (Figure
S18). The purified protein had an absorbance maximum at 415
nm, which disappeared upon treatment with excess dithionite
or on exposure to oxygen, consistent with the presence of the
predicted [4Fe-4S] cluster (Figure 4). Iron and sulfide analysis
yielded 2.4 irons and 2.8 sulfides per monomer of MqnE,
indicating incomplete reconstitution of the cluster (Supporting
Information).
To investigate the involvement of MqnE (TTHA0804) in
the formation of aminofutalosine, 3-[(1-carboxyvinyl)oxy]-
benzoic acid (12) was prepared by incubating chorismate
with MqnA or by chemical synthesis (Figures S8−S16).
Compound 12 (2 mM) was then incubated anaerobically
with MqnE (200 μM) in the presence of S-adenosylmethionine
■
*
S
Detailed procedure for constructing the p28nT-mqnA; over-
expression and purification protocols for MqnA, MqnE,
SCO5662, and Hp0089; NMR and LC-MS characterization
of 3-((1-carboxyvinyl)oxy)benzoic acid and aminofutalosine;
and procedures for iron and sulfur determination in the purified
TTHA0804. This material is available free of charge via the
Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
begley@chem.tamu.edu
Notes
■
(13, 2 mM) and excess sodium dithionite. After 10 h, the
The authors declare no competing financial interest.
reaction was quenched with 8 M guanidine-HCl, protein was
removed by ultrafiltration, and the reaction mixture was
analyzed by reverse-phase HPLC. This analysis showed the
formation of a new compound (Figure 5), which was identified
as aminofutalosine 4 by 1D and 2D NMR (Figures S19−S23)
and MS analysis (Figure S24). This identification was further
confirmed by enzymatic conversion of the reaction product to
ACKNOWLEDGMENTS
■
We thank Cynthia L. Kinsland from Cornell University for
construction of p28nT-mqnA and Dr. Howard Williams of
Texas A&M University for assistance with the NMR analysis.
This research was supported by a grant from the National
Institutes of Health (DK44083) and by the Robert A. Welch
Foundation (A-0034).
10
futalosine using aminofutalosine deaminase (Figures S25 and
10
S26) and to compound 5 using MqnB (Figures S27 and S28).
The MqnE reaction mixtures were also analyzed to
determine if MqnE was catalytic and to determine the fate of
the carboxylate of compound 12. With the current enzyme
preparations, 1 equiv of enzyme gives 14 equiv of the reaction
product 4, clearly demonstrating that MqnE is catalytic. NMR
REFERENCES
■
(
1) Nowicka, B.; Kruk, J. Biochim. Biophys. Acta, Bioenerg. 2010, 1797,
587.
2) Cranenburg, E. C. M.; Schurgers, L. J.; Vermeer, C. Thromb.
Haemost. 2007, 98, 120.
3) Plaza, S. M.; Lamson, D. W. Altern. Med. Rev. 2005, 10, 24.
(4) Bentley, R.; Meganathan, R. Microbiol. Rev. 1982, 46, 241.
(5) Meganathan, R. Vitam. Horm. 2001, 61, 173.
(6) Hiratsuka, T.; Furihata, K.; Ishikawa, J.; Yamashita, H.; Itoh, N.;
Seto, H.; Dairi, T. Science 2008, 321, 1670.
7) Seto, H.; Jinnai, Y.; Hiratsuka, T.; Fukawa, M.; Furihata, K.; Itoh,
N.; Dairi, T. J. Am. Chem. Soc. 2008, 130, 5614.
8) Dairi, T. Methods Enzymol. 2012, 515, 107.
9) Li, X.; Apel, D.; Gaynor, E. C.; Tanner, M. E. J. Biol. Chem. 2011,
86, 19392.
10) Arakawa, C.; Kuratsu, M.; Furihata, K.; Hiratsuka, T.; Itoh, N.;
Seto, H.; Dairi, T. Antimicrob. Agents Chemother. 2011, 55, 913.
(11) Cooper, L. E.; Fedoseyenko, D.; Abdelwahed, S. H.; Kim, S.-H.;
Dairi, T.; Begley, T. P. Biochemistry 2013, 52, 4592.
12) Bietti, M.; Calcagni, A.; Cicero, D. O.; Martella, R.; Salamone,
M. Tetrahedron Lett. 2010, 51, 4129.
1
(
1
3
analysis of the reaction mixture, run using [ C]-12,
demonstrated the formation of bicarbonate and not formate,
suggesting that the carboxy group of 12 is lost in a
decarboxylation reaction (Figure S29).
(
A mechanistic proposal for the aminofutalosine synthase
(
(MqnE)-catalyzed reaction is outlined in Figure 6. Addition of
the adenosyl radical 11, generated by the reductive cleavage of
SAM 13, to the vinyl double bond of compound 12 gives the
captodative radical 15. A similar radical addition has been
reported in a pyruvate formate lyase-derived peptide in which
(
(
2
(
19
the active-site glycine was replaced with dehydroalanine.
Rearrangement of this radical via 16 gives 17. Such rearrange-
ments have previously been proposed but not mechanistically
1
2−14
(
characterized.
Decarboxylation of 17, facilitated by the
4Fe-4S]2 cluster completes the formation of aminofutalosine
4).
alkoxy radical, gives 18. A final electron transfer from 18 to the
+
(13) Baroudi, A.; Alicea, J.; Flack, P.; Kirincich, J.; Alabugin, I. V. J.
Org. Chem. 2011, 76, 1521.
14) Baroudi, A.; Flack, P.; Alabugin, I. V. Chem.Eur. J. 2010, 16,
2316.
15) Frey, P. A.; Hegeman, A. D.; Ruzicka, F. J. Crit. Rev. Biochem.
Mol. Biol. 2008, 43, 63.
(16) Roach, P. L. Curr. Opin. Chem. Biol. 2011, 15, 267.
(17) Vey, J. L.; Drennan, C. L. Chem. Rev. 2011, 111, 2487.
[
(
(
1
(
In summary, we have reconstituted the formation of
aminofutalosine (4) in the menaquinone biosynthesis pathway.
This reaction involves the addition of an adenosyl radical to the
enol ether double bond of chorismate-derived 12 and
represents a new catalytic motif in radical SAM enzymology,
1
5320
dx.doi.org/10.1021/ja408594p | J. Am. Chem. Soc. 2013, 135, 15318−15321

Journal of the American Chemical Society
Communication
(
18) Zhang, Y.; Zhu, X.; Torelli, A. T.; Lee, M.; Dzikovski, B.;
Koralewski, R. M.; Wang, E.; Freed, J.; Krebs, C.; Ealick, S. E.; Lin, H.
Nature 2010, 465, 891.
(
19) Wagner, A. V. F; Demand, J.; Schilling, G.; Pils, T.; Knappe, J.
Biochem. Biophys. Res. Commun. 1999, 254, 306.
20) Miyakoshi, S.; Haruyama, H.; Shioiri, T.; Takahashi, S.;
Torikata, A.; Yamazaki, M. J. Antibiot. 1992, 45, 394.
21) Cooper, R.; Mierzwa, R.; Patel, M.; Pramanik, B.; Puar, M. S.;
Troyanovich, J.; Gullo, V. J. Nat. Prod. 1988, 51, 1207.
(
(
1
5321
dx.doi.org/10.1021/ja408594p | J. Am. Chem. Soc. 2013, 135, 15318−15321