Synthesis of (±)-cyclic dehypoxanthine futalosine, the biosynthetic intermediate in an alternative biosynthetic pathway for menaquinones

Article abstract of DOI:10.1016/j.tetlet.2011.07.061

The first synthesis of (±)-cyclic dehypoxanthine futalosine (cyclic DHFL), a biosynthetic intermediate in the futalosine pathway for menaquinones operating in microorganisms, has been achieved. Efficient growth of the Streptomyces coelicolor mutant, which lacks the cyclic DHFL synthetase gene (mqnC gene) was observed in the presence of synthetic (±)-cyclic DHFL.

Full text of DOI:10.1016/j.tetlet.2011.07.061

Tetrahedron Letters 52 (2011) 4934–4937
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
Synthesis of ( )-cyclic dehypoxanthine futalosine, the biosynthetic intermediate
in an alternative biosynthetic pathway for menaquinones
Arata Yajima a, , Saki Kouno a, Tohru Dairi b, Manami Mogi a, Ryo Katsuta a, Haruo Seto c, Tomoo Nukada a
⇑
a Department of Fermentation Science, Faculty of Applied Biological Science, Tokyo University of Agriculture (NODAI), 1-1-1 Sakuragaoka, Setagaya-ku, Tokyo 156-8502, Japan
b Division of Biotechnology and Macromolecular Chemistry, Laboratory of Applied Biochemistry, Faculty of Engineering, Hokkaido University, Kita 13, Nishi 8, Kita-ku,
Sapporo, Hokkaido 060-8628, Japan
c Department of Applied Biology and Chemistry, Faculty of Applied Biological Science, Tokyo University of Agriculture (NODAI), 1-1-1 Sakuragaoka, Setagaya-ku,
Tokyo 156-8502, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
The first synthesis of ( )-cyclic dehypoxanthine futalosine (cyclic DHFL), a biosynthetic intermediate in
the futalosine pathway for menaquinones operating in microorganisms, has been achieved. Efficient
growth of the Streptomyces coelicolor mutant, which lacks the cyclic DHFL synthetase gene (mqnC gene)
was observed in the presence of synthetic ( )-cyclic DHFL.
Received 14 June 2011
Revised 6 July 2011
Accepted 13 July 2011
Available online 22 July 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Cyclic DHFL
Menaquinone
Futalosine pathway
Helicobacter pylori
Campylobacter jejuni
Many bacteria use menaquinone (MK) as carriers of electrons in
their electron transport chains.1,2 Escherichia coli uses ubiquinone
(Q) under aerobic conditions, but MK under anaerobic conditions
(Fig. 1). On the other hand, Bacillus subtilis uses only MK as a
lipid-soluble molecule that shuttles electrons between mem-
brane-bound protein complexes. In mammalian cells, MK plays
multiple roles in the activation of a family of proteins involved in
blood coagulation3 and bone metabolism.4 Thus, MK is a critical
molecule for survival of both microorganisms and mammals and
is classified as a vitamin.
The biosynthesis of MK has been extensively studied in E. coli.
MK is derived from chorismic acid (or its anionic form, chorismate)
by a series of transformations utilizing eight enzymes, MenA to
MenH, and generating various intermediates, including o-succinyl-
benzoic acid and 1,4-dihydroxy-2-naphthoic acid (Scheme 1).2
However, Borodina et al. reported that the whole-genome analysis
of Streptomyces coelicolor A3(2) and Streptomyces avermitilis re-
vealed the absence of the known menaquinone biosynthetic path-
way genes, menB to menF.5 Furthermore, some pathogenic
microorganisms, such as Helicobacter pylori and Campylobacter
jejuni, which are known to cause gastric carcinoma and diarrhea,
respectively, have also been reported to lack men gene homologs,
even though they synthesize MKs.6,7 These results suggest that
there is a novel pathway for the biosynthesis of MK in some micro-
organisms. Recently, Seto, Dairi, et al. reported evidence for the
presence of a new biosynthetic pathway acquired by extensive tra-
cer experiments with 13C-labeled glucose.8 They succeeded in out-
lining the pathway by a combination of bioinformatics and
biochemical experiments (Scheme 1).9 Compared to the classical
one, the new pathway involves unique transformations from
chorismate, the common starting material. Namely, futalosine
(1), which was previously isolated from the fermentation broth
of Streptomyces sp. MK359-NF1,10 was derived from chorismic acid
and then was converted to dehypoxanthine futalosine (DHFL, 2) by
the release of hypoxanthine. Six-membered ring formation gave
cyclic dehypoxanthine futalosine (cyclic DHFL, 3), followed by
the loss of a three-carbon unit to afford 1,4-dihydroxy-6-naphthoic
acid (DHNA, 4). It should be noted that this compound is very sim-
ilar to an intermediate of the classical pathway in that they are
⇑
Corresponding author. Tel./fax: +81 3 5477 2264.
E-mail address: ayaji@nodai.ac.jp (A. Yajima).
Figure 1. Structures of menaquinones and ubiquinones.
0040-4039/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2011.07.061

A. Yajima et al. / Tetrahedron Letters 52 (2011) 4934–4937
4935
be an intermediate in the futalosine pathway in H. pylori and C.
jejuni. These methods enabled us to provide a sufficient amount
of 1 and its derivatives for studies on menaquinone biosynthesis
and the development of inhibition assays for MqnB.13 This commu-
nication describes the synthesis and biological activity of ( )-cyclic
DHFL.
Scheme 2 shows our retrosynthetic analysis of ( )-cyclic DHFL.
Because cyclic DHFL has diverse functional groups which, in com-
bination, include multiple oxidation states (i.e., hydroxyl, keto, and
carboxylic groups and lactol) in a relatively small molecule, the key
to the synthesis of 3 is in the order of introduction of the oxygen
functional groups. In view of the instability of cyclic DHFL, the car-
boxyl and keto groups that might cause side reactions should be
installed in the later stages of the synthesis. The two hydroxyl
groups would be installed by dihydroxylation of the double bond
of B, and spiro-c-butenolide (B) would be synthesized from com-
mercially available tetralone (C).
Scheme 3 summarizes our synthesis of ( )-3. According to Car-
da’s protocol,14 the known tetralone derivative 515 was converted
to spiro-
6 with Grubbs second-generation catalyst was followed by allylic
oxidation to give spiro- -butenolide ( )-7 in good yield. Next,
c-butenolide 7. Namely, ring closing metathesis of diene
c
OsO4-catalyzed dihydroxylation of 7 gave the corresponding diol
exclusively (>10:1) in moderate yield (ca. 40%). The relative stereo-
chemistry of the major isomer was confirmed by NOESY experi-
ments after derivatization into 8. This stereoselectivity can be
rationalized by the steric effect of H-12 shielding the b-face of
the double bond of 7. Attempts to improve the yield of this trans-
formation were unsuccessful. The best result was 57% yield
achieved using a relatively large amount (0.6 equiv) of OsO4 with
trimethylamine-N-oxide (TMNO) and methanesulfonamide. Lac-
tone 8 was then carefully reduced by LAH at a low temperature
to afford the corresponding lactol. Various reducing agents such
as DIBAL were not effective in this reaction. Initially, lactol was
protected as a simple methyl group, but we found that deprotec-
tion of methyl acetal was difficult in the final stage of the synthesis.
We therefore selected silyl acetal (TIPS) for the lactol protecting
group. Treatment of lactol with TIPSOTf gave the corresponding si-
lyl acetal. The resulting anomeric isomers were separated after the
Scheme 1. Menaquinone biosynthetic pathways. The four carbon units highlighted
by red lines were derived from erythrose-4-phosphate. The units were located at
different positions in the final menaquinones.
hydrogenolysis of the benzyl group (9,
istry of the separated major -isomer was determined by a NOESY
experiment on 9, and the -isomer was used for the subsequent
a:b = 3:1). The stereochem-
a
a
transformations. The corresponding triflate of 9 was subjected to
palladium-catalyzed methoxycarbonylation16 to afford 10. The
benzylic position of 10 was oxidized with PhI(OAc)2 and TBHP in
the presence of K2CO3,17 and then the methyl ester was hydrolyzed
to give 11. The deprotection of the acetonide and TIPS groups with-
positional isomers. Dairi et al. named this new pathway the futalo-
sine pathway.9 Because mammals and some beneficial intestinal
bacteria, such as lactobacilli, lack this futalosine pathway, it would
be a promising target for the development of chemotherapeutics
against H. pylori and C. jejuni.11 Although enzymes for the conver-
sion of DHFL to cyclic DHFL and cyclic DHFL to DHNA have been
identified, detailed mechanisms of these unique transformations
have not been disclosed. Recently, Yokoyama et al. reported the
crystal structure of MqnD (TTHA1568) from Thermus thermophilus
HB8 as a co-crystal with L
-(+)-tartaric acid.12 They proposed a bind-
ing site for the target compound of MqnD and a putative reaction
mechanism for the transformation of cyclic DHFL to DHNA (vide
infra). However, further studies are necessary to establish a de-
tailed reaction mechanism for the development of effective chemo-
therapeutics. Although there is an urgent need to provide samples
of cyclic DHFL for this purpose, it is very difficult to isolate cyclic
DHFL from the bacterial culture broth because of its extreme scar-
city and instability.9 We became interested in synthesizing cyclic
DHFL to confirm the structure of the natural product and provide
samples for bioorganic experiments. Very recently, Dairi et al.
and Tanner et al. independently reported an efficient synthesis of
futalosine and aminodeoxyfutalosine, which has an adenine moi-
ety instead of the hypoxanthine in 1 and confirmed the latter to
Scheme 2. Retrosynthetic analysis of cyclic DHFL (3).

4936
A. Yajima et al. / Tetrahedron Letters 52 (2011) 4934–4937
Scheme 4. Putative reaction mechanisms for the conversion of cyclic DHFL to
DHNA catalyzed by MqnD.
gested a putative mechanism of this reaction based on the depro-
tonation of H-5 by His-145 of TthMqnD.12 Indeed His-145 would be
essential, but there is a reason to reconsider this mechanism. The
precursor of DHNA would be the corresponding 1,4-diketone, be-
cause keto–enol tautomerism can easily give a hydroquinone moi-
ety. The 1,4-diketone could be derived from either of the two
different pathways (A or B in Scheme 4). Because the resulting side
products, propanedial or pyruvaldehyde, are different in the two
pathways, an analysis of the byproduct should reveal the enzy-
matic reaction mechanism.
In summary, we achieved the first synthesis of ( )-cyclic DHFL.
Synthetic ( )-3 enabled the survival of the mutant that lacks MqnC.
Improvement of the synthesis and the establishment of the enzy-
matic reaction mechanisms using the synthetic sample are cur-
rently underway in our laboratory.
Scheme 3. Synthesis of ( )-cyclic DHFL (3). Reagents, conditions and yields: (a)
vinylmagnesium chloride, THF, À78 °C (94%); (b) NaH, allyl bromide, DMF, 0 °C to rt
(86%); (c) Grubbs’ second gen. cat., CH2Cl2 (98%); (d) CrO3, 3,5-DMP, CH2Cl2 À40 °C
(99%); (e) OsO4, TMAO, MeSO2NH2, t-BuOH, acetone, 0 °C (57%); (f) 2,2-dimethoxy-
propane, PPTS, reflux (73%); (g) LiAlH4, THF, 0 °C (90%); (h) TIPSOTf, 2,6-lutidine,
CH2Cl2, 0 °C (quant.); (i) Pd/C, H2, MeOH, EtOAc (73%); (j) Tf2O, pyridine 0 °C (96%);
(k) Pd(Ph3P)4, Et3N, CO, MeOH, DMSO, THF, 60 °C (quant.); (l) PhI(OAc)2, TBHP,
K2CO3, n-butyl butanoate À20 °C (66%); (m) LiOH, THF, H2O (83%); (n) BiCl3, H2O,
CH3CN, THF (20%).
Acknowledgments
We thank Professor Hidenori Watanabe (The University of To-
kyo) for his suggestions on the synthetic plan. We thank Dr. T. Tas-
hiro (RIKEN) for the collection of MS spectra. We also thank Mr. N.
Kawanishi for his early contributions to the synthetic study. This
work was supported by JSPS KAKENHI (22310139).
out degradation of the resultant cyclic DHFL was very difficult un-
der standard acidic conditions. We obtained 12 as a degradation
product of cyclic DHFL, which was produced by an intramolecular
aldol reaction under neutral, basic, and strongly acidic conditions.
However, we ultimately found that BiCl3-mediated hydrolysis18 of
the acetonide and TIPS groups of 11 gave ( )-cyclic DHFL (3). The
HPLC retention time of the synthetic 3 was identical to that of nat-
ural cyclic DHFL, and the 1H and 13C NMR spectra were also in good
accordance, except for the 13C NMR chemical shifts of C-10 and C-
14.19 The small differences for these chemical shifts can be as-
cribed to differences in the pH or concentration of the samples.
Next, the biological activity of synthetic ( )-3 was examined by
the same method as described by Dairi et al.9 Synthetic ( )-3 was
added to an agar plate, and the growth of the S. coelicolor mutant,
which lacks the cyclic DHFL synthetase gene (mqnC gene)9 was
examined. Although, this mutant cannot grow in the absence of
MK, efficient growth was observed in the presence of synthetic
References and notes
1. Bentley, R.; Maganathan, R. Microbiol. Rev. 1982, 46, 241–280.
2. Maganathan, R. Vitam. Horm. 2001, 61, 173–218.
3. Dahlbäck, B.; Villoutreix, B. O. Arterioscler. Thromb. Vasc. Biol. 2005, 25, 1311–
1320.
4. Adams, J.; Pepping, J. Am. J. Health Syst. Pharm. 2005, 62, 1311–1320.
5. Borodina, I.; Krabben, P.; Nielsen, J. Genome. Res. 2005, 15, 802–827.
6. Marcelli, S. W.; Chang, H.-T.; Chapman, T.; Chalk, P. A.; Miles, R. J.; Poole, R. K.
FEMS Microbiol. Lett. 1996, 138, 59–64.
7. Moss, C. W.; Lambert-Fair, M. A.; Nicholson, M. A.; Guerrant, G. O. J. Clin.
Microbiol. 1990, 28, 395–397.
8. Seto, H.; Jinnai, Y.; Hiratsuka, T.; Fukawa, M.; Furihata, K.; Itoh, N.; Dairi, T. J.
Am. Chem. Soc. 2008, 130, 5614–5615.
9. Hiratsuka, T.; Furihata, K.; Ishikawa, J.; Yamashita, H.; Itoh, N.; Seto, H.; Dairi, T.
Science 2008, 321, 1670–1673.
( )-3 (10 lg/ml).
In the final stage of our synthesis, the degradation products of
cyclic DHFL (12) were obtained during the hydrolysis of the aceto-
nide and lactol protecting groups (Scheme 3). These compounds
were produced by an intramolecular aldol reaction. This indicates
that the open form of 3 is easily formed in aqueous media. On
the basis of this observation, we formulated potential enzymatic
reaction mechanisms for the conversion of cyclic DHFL into DHNA
as shown in Scheme 4. As described above, Yokoyama et al. sug-
10. Hosokawa, N.; Naganawa, H.; Kasahara, T.; Hattori, S.; Hamada, M.; Takeuchi,
T.; Yamamoto, S.; Tsuchiya, K. S.; Hori, M. Chem. Pharm. Bull. 1999, 47, 1032–
1034.
11. (a) Dairi, T. J. Antibiot. 2009, 62, 347–352; (b) Kurosu, M.; Begari, E. Molecules
2010, 15, 1531–1553.
12. Arai, R.; Murayama, K.; Uchikubo-Kamo, T.; Nishimoto, M.; Toyama, M.;
Kuramitsu, S.; Terada, T.; Shirouzu, M.; Yokoyama, S. J. Struct. Biol. 2009, 168,
575–581.
13. (a) Arakawa, C.; Kuratsu, M.; Furihata, K.; Hiratsuka, T.; Itoh, N.; Seto, H.; Dairi,
T. Antimicrob. Agents Chemother. 2011, 55, 913–916; (b) Li, X.; Tanner, E.

A. Yajima et al. / Tetrahedron Letters 52 (2011) 4934–4937
4937
Tetrahedron Lett. 2010, 51, 6463–6465; (c) Li, X.; Apel, D.; Gaynor, E. C.; Tanner,
M. E. J. Biol. Chem. 2011, 286, 19392–19398.
14. Carda, M.; Castillo, E.; Rodríguez, S.; Uriel, S.; Marco, J. A. Synlett 1999, 1639–
1641.
15. Durden, J. A., Jr. J. Agric. Food Chem. 1971, 19, 432–440.
16. Cacchi, S.; Morera, E.; Ortar, G. Tetrahedron Lett. 1985, 26, 1109–1112.
17. Zhao, Y.; Yeung, Y.-Y. Org. Lett. 2010, 12, 2128–2131.
18. Swamy, N. R.; Venkateswarlu, Y. Tetrahedron Lett. 2002, 43, 7549–7552.
19. Properties of the selected synthetic compounds (8): 1H NMR (400 MHz,
CDCl3):d 1.39 (s, 3H), 1.54 (s, 3H), 1.95–2.05 (m, 2H), 2.10 (m, 1H), 2.46 (m, 1H),
2.76–2.92 (m, 2H), 4.63 (d, J = 5.5 Hz, 1H), 5.05 (brs, 2H), 5.10 (d, J = 5.5 Hz, 1H),
6.80 (d, J = 10.5 Hz, 1H), 6.80 (d, J = 10.5 Hz, 1H), 6.82 (s, 1H), 7.30–7.45 (m,
5H); 13C NMR (100 MHz, CDCl3):d 19.2, 26.0, 26.8, 29.1, 29.8, 69.9, 76.3, 81.8,
86.1, 113.4, 113.9, 115.2, 125.7, 127.3, 128.0, 128.5, 128.6, 136.5, 139.0, 158.6,
174.0. Anal. Calcd for C23H24O5: C, 72.01, H, 6.36. Found: C, 72.08, H, 6.49. (11):
1H NMR (400 MHz, CDCl3):d 1.00–1.12 (m, 21H), 1.28 (brs, 1H), 1.35 (s, 3H),
1.60 (s, 3H), 2.46 (m, 2H), 2.69 (m, 2H), 4.81 (brs, 2H), 5.65 (s, 1H), 8.07 (d,
J = 8.2 Hz, 1H), 8.21 (dd, J = 1.8, 8.2 Hz, 1H), 8.61 (d, J = 1.8 Hz, 1H); 13C NMR
(100 MHz, CDCl3):d 11.8, 17.66, 17.69, 25.1, 26.6, 31.4, 35.3, 85.8, 86.5, 89.3,
104.3, 113.4, 128.2, 129.1, 129.3, 131.4, 134.4, 151.4, 169.5, 196.7; HR-MS-ESI
(m/z): [M+Na]+ calcd for C26H38NaO7Si, 513.2285; found, 513.2302. Synthetic
( )-cyclic DHFL [anomeric mixture (ca. b:a
= 6:4]: colorless powder; 1H NMR
(400 MHz, CD3OD):d 2.18 (ddd, J = 5.0, 13.3, 13.3 Hz, 0.4H), 2.27 (ddd, J = 5.0,
13.3, 13.3 Hz, 0.6H), 2.66 (m, 0.4H), 2.70 (m, 0.6H), 2.83 (m, 0.6H), 2.86 (m,
0.4H), 3.07 (ddd, J = 5.0, 13.7, 18.8 Hz, 0.6H), 3.15 (ddd, J = 5.2, 13.5, 18.5 Hz,
0.4H), 3.99 (d, J = 5.3 Hz, 0.6H), 4.06 (dd, J = 3.3, 4.5 Hz, 0.4H), 4.18 (d, J = 4.5 Hz,
0.4H), 4.32 (dd, J = 4.6, 5.3 Hz, 0.6H), 5.43 (d, J = 3.3 Hz, 0.4H), 5.56 (d, J = 4.6 Hz,
0.6H), 7.55 (d, J = 7.6 Hz, 0.6H), 8.07 (d, J = 8.0 Hz, 0.4H), 8.21 (d, J = 7.6 Hz,
0.6H), 8.22 (d, J = 8.0 Hz, 0.4H), 8.54 (s, 0.4H), 8.56 (s, 0.6H); 13C NMR
(100 MHz, CD3OD):d 32.7 (C-5,
(C-2, b), 77.7 (C-2, ), 77.8 (C-3, b), 79.7 (C-3,
(C-1, b), 102.7 (C-1, ), 127.0 (C-12, b), 128.8 (C-12,
9, b), 132.3 (C-8, b), 132.8 (C-8, ), 133.9 and 134.1 (C-10,
and 131.9), 135.4 (C-11, ), 135.4 (C-11, b), 153.1 (C-13, b), 153.9 (C-13,
a
), 33.4 (C-5, b), 36.4 (C-6,
), 84.8 (C-4,
), 129.0 (C-9,
and b lit.9 d 131.6
),
a
a
), 36.6 (C-6, b), 71.7
), 86.0 (C-4, b), 98.5
), 129.5 (C-
a
a
a
a
a
a
a
a
a
167.3 and 169.0 (C-14,
a
and b, lit.9 d 168.8 and 168.8), 198.9 (C-7, b), 199.5 (C-
7,
a
); HR-MS-ESI (m/z): [MÀH]+ calcd for C14H13O7, 293.06613; found,
293.06671. 1H and 13C NMR spectra are in good accordance with those of the
reported natural product.