Improved Chemical Syntheses of 1- and 5-Deazariboflavin

Article abstract of DOI:10.1021/jo049859f

The cofactor flavin adenine dinucleotide (FAD) is required for the catalytic activity of a large class of enzymes known as flavoenzymes. Because flavin cofactors participate in catalysis via a number of different mechanisms, isoalloxazine analogues are valuable for mechanistic studies. We report improved chemical syntheses for the preparation of the two key analogues, 5-deazariboflavin and 1-deazariboflavin.

Full text of DOI:10.1021/jo049859f

Im p r oved Ch em ica l Syn th eses of 1- a n d
5-Dea za r ibofla vin
Erin E. Carlson† and Laura L. Kiessling*,†,‡
Departments of Chemistry and Biochemistry,
University of Wisconsin, Madison, 1101 University Avenue,
Madison, Wisconsin 53706
kiessling@chem.wisc.edu
Received J anuary 24, 2004
F IGURE 1. Flavin analogues are useful in the elucidation of
the cofactor’s role in catalysis.
Abstr a ct: The cofactor flavin adenine dinucleotide (FAD)
is required for the catalytic activity of a large class of
enzymes known as flavoenzymes. Because flavin cofactors
participate in catalysis via a number of different mecha-
nisms, isoalloxazine analogues are valuable for mechanistic
studies. We report improved chemical syntheses for the
preparation of the two key analogues, 5-deazariboflavin and
1-deazariboflavin.
The exceptional chemistry of the flavin cofactor allows
flavoenzymes to play a wide variety of roles in vivo. These
proteins participate in many biological processes includ-
ing nitric oxide generation, photosynthesis, soil detoxi-
fication, and even apoptosis.1 The diversity of functions
performed by flavoproteins is due to the ability of the
isoalloxazine ring system to participate in an assortment
of catalytic mechanisms. It can perform redox and radical
chemistry, act as a nucleophile or electrophile, or simply
play a structural role.1,2 With this range of potential
functions, discerning the catalytic role of a flavin cofactor
is often difficult.
Flavin analogues are invaluable tools for elucidating
the role of the cofactor.1,3 Most flavin adenine dinucleotide
(FAD) analogues have been obtained through the chemi-
cal synthesis of the corresponding riboflavin derivatives,
followed by coupling to the nucleotide diphosphate medi-
ated by FAD synthetase to yield the full FAD analogue.4
Previously, both 5-deazariboflavin 15,6 and 1-deazaribo-
flavin 26,7 have been synthesized3,8 and utilized in chemi-
cal and biological settings (Figure 1). Some analogues,
however, such as 1-deazariboflavin, are seldom employed
because of the difficulty of their synthesis. We found key
F IGURE 2. Coupling of ribose and 3 provides precursors to
both analogues 1 and 2. Previous syntheses required high
pressure and temperature hydrogenation to perform this
coupling.
steps within the routes to 1 and 2 to be irreproducible.
To overcome these limitations, we developed new routes
that afford efficient and reproducible syntheses of both
analogues.
The syntheses of both 1 and 2 proceed via intermedi-
ates 4 and 10, which possess common features (Figure
2). In previous routes, syntheses of these intermediates
required high temperature and pressure hydrogenation
steps, which are difficult to carry out on larger scales,
and require specialized equipment (Figure 2).3 To develop
a more convenient route, we surveyed a series of reduc-
tive amination conditions. The use of NaCNBH3 in
refluxing methanol was found to be optimal, giving the
ribitylated aniline 4 in a 90% yield and 10 in a 92% yield.
The published methods for conversion of 4 (or 10) to
compound 1 (or 2) involved other problematic steps.
Although several syntheses of compound 1 were pub-
lished around 1970,8 a number of alternative pathways
have since appeared in the literature9 because the key
coupling step between intermediate 4 and 6-chlorouracil10
was irreproducible (Figure 3). Typically, this reaction is
run at high temperature and without any type of base
or catalyst. We determined that a catalytic amount of
† Department of Chemistry.
‡ Department of Biochemistry.
(1) Massey, V. Biochem. Soc. Trans. 2000, 28, 283-296. Chemistry
and Biochemistry of Flavoenzymes; Muller, F., Ed.; CRC Press: Boca
Raton, 1992.
(2) Niemz, A.; Rotello, V. Acc. Chem. Res. 1999, 32, 44-53. Sheng,
D.; Ballou, D. P.; Massey, V. Biochemistry 2001, 40, 11156-11167.
Lario, P. I.; Sampson, N.; Vrielink, A. J . Mol. Biol. 2003, 326, 1635-
1650.
(3) Ashton, W. T.; Graham, D. W.; Brown, R. D.; Rogers, E. F.
Tetrahedron Lett. 1977, 30, 2551-2554 and references therein.
(4) Spencer, R. F. J .; Walsh, C. Biochemistry 1976, 15, 1043-1053.
(5) Some recent examples: Poinas, A.; Gaillard, J .; Vignais, P.;
Doussiere, J . Eur. J . Biochem. 2002, 269, 1243-1252. Sevrioukova, I.
F.; Hazzard, J . T.; Tollin, G.; Poulos, T. L. Biochemistry 2001, 40,
10592-10600. Huang, Z.; Zhang, Q.; Liu, H. Bioorg. Chem. 2003, 31,
494-502.
(8) O’Brien, D. E.; Weinstock, L. T.; Cheng, C. C. J . Heterocycl.
Chem. 1970, 7, 99-105. Smit, P.; Stork, G. A.; van der Plas, H. C.
Recl. Trav. Chim. Pays-Bas 1986, 105, 538-544. O’Brien D. E.;
Weinstock L. T.; Cheng C. C. Chem. Ind. (London) 1967, 48, 2044-
2045.
(9) Ashton, W. T.; Brown, R. D.; Tolman, R. L. J . Heterocycl. Chem.
1978, 15, 489-491. Frier, C.; De´cout, J .-L.; Fontecave, M. J . Org. Chem.
1997, 62, 3520-3528. Yoneda, F.; Sakuma, Y.; Ichiba, M.; Shinomura,
K. J . Am. Chem. Soc. 1976, 98, 830-835.
(10) Ishikawa, I.; Itoh, T.; Melik-Ohanjanian, R. G.; Takayanagi,
H.; Mizuno, Y.; Ogura, H, Kawahara, N. Heterocycles 1990, 31, 1641-
1646.
(6) Payne, G.; Wills, M.; Walsh, C.; Sancar, A. Biochemistry 1990,
29, 5706-5711.
(7) Kekelidze, T. N.; Edmondson, D. E.; McCorkmick, D. B. Arch.
Biochem. Biophys. 1994, 315, 100-103.
10.1021/jo049859f CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/03/2004
2614
J . Org. Chem. 2004, 69, 2614-2617

F IGURE 3. Development of reliable conditions for the key
coupling reaction involved in the synthesis of 1 has been an
elusive goal.
F IGUR E 4. Previous methods used to introduce an ortho-
amino group into 4.
SCHEME 2
SCHEME 1
malononitrile, which has been reported to catalyze
couplings with conjugated vinyl halides,11 in refluxing
methanol is required for efficient and reproducible cou-
pling to form the functionalized uracil, 5 (Scheme 1).
Purification of 5 proved difficult as a result of its high
polarity and acid sensitivity. To circumvent this problem,
the crude material 5 was acetylated, and the product was
isolated by extraction to give protected uracil 6 in a 65%
yield over two steps. The last two steps of the synthesis
of compound 1 were carried out as previously reported.12
Treatment of compound 6 with the Vilsmeier reagent
followed by removal of the acetate protecting groups
provided crude 5-deazariboflavin, 1. Purification by high
performance liquid chromatography (HPLC) provided the
yellow solid 1 as the trifluoroacetic acid (TFA) salt, in
50% yield (19% overall).
The synthesis of compound 2 also presented several
challenges. As previously mentioned, the reductive ami-
nation conditions developed for the synthesis of 1 were
effective en route to 2. However, established strategies
employed aniline 4, which can be converted to the
ribitylated intermediate 10 via further aromatic func-
tionalization (Figure 4).13 To avoid this late-stage func-
tionalization, the commercially available diamine 8 was
utilized. Initial attempts to monoribitylate the free
aniline proved problematic. When compound 8 was
desymmetrized with tert-butyloxycarbonyl anhydride
(Boc2O), however, the protected aniline 9 was obtained
in 66% yield (Scheme 2). Next, D-ribose was coupled with
aniline 9 by reductive amination to afford the desired
product in a 92% yield. The t-Boc protecting group was
removed using HCl-dioxane to give intermediate 10 in
near quantitative yield.
In previous syntheses, the formation of bicyclic inter-
mediate 12 from compounds 10 and 11 was extremely
low yielding (19%).9 These routes utilized potassium
carbonate as a base, which is rather insoluble in organic
solvents. When cesium carbonate in a mixture of dim-
ethylformamide and dichloromethane was employed,
yields up to 55% of 12 were obtained. The isoalloxazine
ring system was completed by stirring intermediate 12
in methanolic ammonia to give 1-deazariboflavin, 2. The
crude material was purified by HPLC to give the TFA
salt of 2 as a purple solid in 33% yield (11% overall).
The routes described here address each of the prob-
lematic steps in previous syntheses of flavin analogues
1 and 2. The resulting strategies are more efficient and
alleviate the need for specialized equipment. Thus, they
can readily be carried out on a multigram scale. Most
importantly, the reactions are reproducible. We antici-
pate that the results presented here will facilitate the
syntheses of flavin analogues for chemical and biological
studies.
(11) Safar, P.; Povazanec, F.; Cepec, P.; Pronayova, N. Collect.
Czech. Chem. Commun. 1997, 62, 1105-1113.
(12) J anda, M.; Hemmerich, P. Angew. Chem., Int. Ed. Engl. 1976,
15, 443-444.
(13) Tishler, M.; Wellman, J . W.; Ladenburg, K. J . Am. Chem. Soc.
1945, 67, 2165-2168. Tishler, M.; Pfister, K.; Babson, R. D.; Laden-
burg, K.; Fleming, A. J . J . Am. Chem. Soc. 1947, 69, 1487-92.
J . Org. Chem, Vol. 69, No. 7, 2004 2615

bicarbonate (0.62 g, 7.3 mmol) were dissolved in dioxane (50 mL)
and water (50 mL). The mixture was stirred for 5 h at room
temperature. The reaction was diluted with water (150 mL) and
then extracted with CH2Cl2 (3 × 50 mL). The combined organic
layers were washed with an aqueous solution of saturated
bicarbonate and then brine, dried over magnesium sulfate, and
concentrated under reduced pressure. The resulting red oil was
purified by flash chromatography (silica, 4:1 hexanes/ethyl
acetate) to yield 9 (1.1 g, 66%) as a pink solid: 1H NMR (300
MHz, CDCl3) δ 7.00 (s, 1 H), 6.54 (s, 1 H), 6.28 (br s, 1 H), 3.57
(br s, 2 H), 2.13 (s, 3 H), 2.12 (s, 3 H), 1.50 (s, 9 H) ppm; 13C
NMR δ 153.9, 137.5, 134.3, 127.5, 125.7, 122.3, 118.7, 80.1, 28.3,
19.2, 18.7 ppm; EMM (m/z) [M + H] calcd for C13H20N2O2
237.1603, found 237.1613.
Syn th esis of 10. Boc-protected aniline 9 (940 mg, 3.98 mmol),
D-ribose (1.79 g, 11.9 mmol), and sodium cyanoborohydride (500
mg, 7.96 mmol) were dissolved in methanol (50 mL). The solution
was heated to 65 °C degrees for 48 h. Solvent was removed under
reduced pressure, and the residue was dissolved in 1 M HCl (10
mL) and swirled until gas evolution ceased. The solution was
carefully neutralized using saturated sodium bicarbonate and
extracted with ethyl acetate (3 × 50 mL). The combined organic
layers were washed with brine (20 mL), dried with magnesium
sulfate, and the solvent was removed under reduced pressure
to give1.35 g, a 92% yield as a yellow/orange solid: 1H NMR
(300 MHz, d3-MeOD) δ 6.87 (s, 1 H), 6.61 (s, 1 H), 3.97-3.62
(m, 5 H), 3.44 (dd, 1 H, J ) 12.9, 3.3 Hz), 3.15 (dd, 1 H, J )
12.6, 7.8 Hz), 2.19 (s, 3 H), 2.13 (s, 3 H), 1.50 (s, 9 H) ppm; 13C
NMR δ 157.4, 142.8, 136.2, 128.9, 126.4, 123.1, 115.4, 81.1, 74.9,
74.5, 72.2, 64.8, 28.9, 20.0, 18.9 ppm; EMM (m/z) [M + Na] calcd
for C18H30N2O6 371.2182, found 371.2170. This compound (238
mg, 0.642 mmol) was dissolved in 4 M HCl in dioxane (11.6 mL),
and the mixture was stirred at room temperature for 5 h. The
dioxane was removed under reduced pressure, the residue was
dissolved in water (100 mL), and the aqueous layer was washed
with ether (3 × 20 mL). The water layer was removed by
lyophilization to yield 10 as a brown foam in a near quantitative
yield: 1H NMR (300 MHz, d3-MeOD) δ 6.77 (s, 1 H), 6.67 (s, 1
H), 3.37-3.43 (m, 1 H), 2.87-3.16 (m, 6 H), 1.58 (s, 3 H), 1.57
(s, 3 H) ppm; 13C δ129.5, 120.2, 116.9, 75.0, 74.5, 72.1, 64.8, 19.5,
19.1 ppm; ESI (m/z) [M + Na] calcd for C13H22N2O4 293.2, found
293.2.
Exp er im en ta l Section
Syn th esis of 4. Aniline 3 (3.0 g, 26 mmol), D-ribose (11.2 g,
74.3 mmol), and sodium cyanoborohydride (3.11 g, 49.5 mmol)
were dissolved in methanol (150 mL), and the mixture was
heated to 65 °C for 48 h. Solvent was removed under reduced
pressure, and the residue was dissolved in 1 M HCl (50 mL)
and swirled until gas evolution ceased. The solution was
carefully neutralized using aqueous saturated sodium bicarbon-
ate solution, and then the mixture was extracted with ethyl
acetate (6 × 50 mL). The combined organic layers were washed
with brine and dried with magnesium sulfate, and the solvent
was removed under reduced pressure to yield 4 (5.67 g, 90%) as
a white solid: 1H NMR (300 MHz, d3-MeOD) δ 6.88 (d, 1 H, J )
7.8 Hz), 6.55 (d, 1 H, J ) 2.4 Hz), 6.47 (dd, 1 H, J ) 7.8, 2.4 Hz),
3.82-3.91 (m, 1 H), 3.72-3.81 (m, 2 H), 3.60-3.67 (m, 2 H), 3.43
(dd, 1 H, J ) 12.6, 3.6 Hz), 3.09 (dd, 1 H, J ) 12.6, 8.1 Hz), 2.17
(s, 3 H), 2.12 (s, 3 H) ppm; 13C NMR δ 148.1, 138.1, 131.2, 126.9,
116.9, 112.6, 74.9, 74.5, 72.3, 64.8, 20.9, 20.2, 18.9 ppm; ESI (m/
z) [M + H] calcd for C13H21NO4 256.2, found 256.2.
Syn th esis of 6. Ribitylated aniline 4 (300 mg, 1.18 mmol),
6-chlorouracil (207 mg, 1.41 mmol), and malononitrile (23 mg,
0.35 mmol) (a higher mol % of catalyst was necessary for gram-
scale reactions) were suspended in dry methanol (5 mL) and
heated at reflux for 48 h. Solvent was removed under reduced
pressure, the crude material 5 was dissolved in pyridine (5 mL)
and acetic anhydride (554 µL, 5.90 mmol), and the mixture was
allowed to stir at room temperature for 1 h. The solvent was
removed under reduced pressure, and the resulting residue was
dissolved in dichloromethane (20 mL) and washed with water
(10 mL) and brine (10 mL). The organic layer was dried with
magnesium sulfate and filtered, and the solvent was removed
under reduced pressure. The residue was recrystallized from
hexanes and dichloromethane, resulting in the orange solid 6
(408 mg, 65% over two steps): 1H NMR (300 MHz, CDCl3) δ
9.90 (br s, 1 H), 8.08 (br s, 1 H), 7.10-7.30 (m, 1 H), 6.86-7.04
(br s, 3 H), 5.09-5.42 (m, 3 H), 4.90 (s, 1 H), 4.29 (dd, 1 H, J )
12.3, 3.0 Hz), 4.00-4.12 (m, 2 H), 3.75 (dd, 1 H, J ) 15.3, 2.7
Hz), 2.29 (s, 3 H), 2.28 (s, 3 H), 2.10 (s, 3 H), 2.02 (s, 3 H), 2.01
(s, 3 H), 1.91 (s, 3 H) ppm; 13C NMR δ 170.4, 170.2, 169.9, 169.5,
145.4, 137.1, 130.2, 125.8, 114.8, 110.3, 70.9, 70.2, 69.6, 61.8,
43.9, 20.7, 20.7, 20.6, 20.5, 19.8, 18.5 ppm; ESI (m/z) [M + Na]
calcd for C25H31N3O10 556.2, found 556.2.
Syn th esis of 11. Diethyl 3-oxoglutarate (0.63 g, 3.1 mmol)
was placed in a three necked round-bottom equipped with a
condenser and two septa. Carbon dioxide gas was bubbled
through the neat solution while it was heated at 65 °C. Bromine
(320 µL, 6.30 mmol) was added slowly, and the solution was
allowed to stir for 30 min. The reaction was cooled to room
temperature and diluted with CH2Cl2 (20 mL). The organic
solution was washed with 10% aqueous Na2SO3 (until color
disappeared), then saturated sodium sulfate (10 mL), saturated
sodium bicarbonate (10 mL), and brine (10 mL). The organic
layer was dried with magnesium sulfate, and the solvent was
removed under reduced pressure to yield a red oil. Crude 11
was used immediately, as purification or heat caused decomposi-
tion: 1H NMR (mixture of mono and dibrominated products, 300
MHz, d3-MeOD) δ 5.33-5.37 (m, 1 H), 4.28 (m, 4 H), 1.32 (m, 6
H) ppm; ESI (m/z) [M + Na] calcd for C9H13BrO5 303.0, found
303.0.
Syn th esis of 7. Bicyclic compound 6 (360 mg, 0.680 mmol)
was dissolved in DMF (1.2 mL) to which phosphorus oxychloride
(119 µL, 1.28 mmol) was added dropwise. This solution was
allowed to stir at room temperature for 30 min and then heated
at 100 °C for 15 min. Ice was added, and the solution was
adjusted to pH ∼6 with ammonium hydroxide. This solution was
stirred at room temperature for 30 min, resulting in precipitation
of product. Filtration resulted in the orange solid 7 (234 mg,
64%): 1H NMR (300 MHz, CDCl3) δ 8.76 (s, 1 H), 8.46 (s, 1 H),
7.57 (s, 2 H), 5.58-5.64 (m, 1 H), 5.39-5.42 (m, 2 H), 4.40 (dd,
1 H, J ) 12.0, 2.7 Hz), 4.22 (dd, 1 H, J ) 12.0, 5.4 Hz), 2.51 (s,
3 H), 2.38 (s, 3 H), 2.26 (s, 3 H), 2.18 (s, 3 H), 2.04 (s, 3 H), 1.71
(s, 3 H) ppm; 13C NMR δ 170.4 170.1, 169.8, 169.5, 161.6, 158.1,
155.0, 142.1, 139.3, 134.5, 131.4, 119.7, 116.4, 113.8, 70.4, 69.7,
69.1, 61.8, 44.7, 21.2, 20.8, 20.6, 20.5, 20.1, 19.0 ppm; ESI (m/z)
[M + Na] calcd for C26H29N3O10 566.2, found 566.2.
Syn th esis of 5-Dea za r ibofla vin (1). Compound 7 (32 mg,
0.059 mmol) was dissolved in methanolic ammonia (3 mL) and
allowed to stir overnight. Solvent was removed under reduced
pressure, and the residue was purified by using HPLC (100%
f 50% A:B over 20 min, retention time about 16 min) to give
11 mg, a 50% yield of the yellow solid 1 as the trifluoroacetic
acid (TFA) salt: mp > 300° (dec); 1H NMR (300 MHz, d6-DMSO)
δ 11.06 (s, 1 H), 8.81 (s, 1 H), 7.93 (s, 1 H), 7.83 (s, 1 H), 4.90
(m, 1 H), 4.63 (d, 1 H, J ) 13.5 Hz), 4.22 (m, 1 H), 3.72-3.59
(m, 3 H), 3.49-3.43 (m, 1 H), 2.45 (s, 3 H), 2.32 (s, 3 H) ppm;
13C NMR δ 162.2, 157.5, 156.3, 146.0, 141.0, 140.0, 133.7, 130.6,
119.7, 117.9, 113.7, 73.7, 72.9, 69.6, 63.5, 47.3, 21.0, 18.7 ppm;
ESI (m/z) [M + Na] calcd for C18H29N3O10 398.1, found 398.1.
Syn th esis of 9. Aniline 8 (1.0 g, 7.3 mmol), tert-butyl
benzyloxy carbonyl anhydride (1.6 g, 7.3 mmol), and sodium
Syn th esis of 12. Compounds 10 (179 mg, 0.583 mmol) and
11 (342 mg, 1.22 mmol) were dissolved in DMF (5 mL) and CH2-
Cl2 (6 mL). Cesium carbonate (570 mg, 1.75 mmol) was added
to the reaction. After stirring at room temperature for 24 h, the
solution was filtered and diluted with water (50 mL), and the
aqueous layer was extracted (3 × 20 mL) with ethyl acetate.
The organic layers were dried with magnesium sulfate, and the
solvent was removed under reduced pressure. The resulting
brown oil was purified with flash chromatography (silica, 3%
MeOH in CHCl3) to yield 12 (144 mg, 55%) as a brown oil: 1H
NMR (300 MHz, CDCl3) δ 8.55 (s, 1 H), 8.24 (s, 1 H), 7.16 (br s,
1 H), 5.55 (br s, 1 H), 4.88-3.80 (m, 7 H), 4.54 (q, 2 H, J ) 6.9
Hz), 4.20 (q, 2 H, J ) 6.9 Hz), 2.64 (s, 3 H), 2.58 (s, 3 H), 1.46 (t,
3 H, J ) 7.2 Hz), 1.23 (t, 3 H, J ) 7.2 Hz) ppm; 13C NMR δ
168.6, 162.9, 161.3, 153.0, 146.5, 145.6, 142.9, 131.3, 118.7, 117.7,
2616 J . Org. Chem., Vol. 69, No. 7, 2004

113.9, 73.0, 72.2, 71.6, 63.8, 62.8, 56.0, 37.7, 21.7, 20.1, 13.9,
13.7 ppm; ESI (m/z) [M + Na] calcd for C22H30N2O8 473.2, found
473.2.
Ack n ow led gm en t. We are grateful to the NSF
(CHE-9357093) and the NIH (GM49975) for support of
this research. E.E.C. was supported by the NIH Bio-
technology Training Program (GM08349). The UW
Chemistry NMR facilities are supported in part by NSF
(CHE-9208463) and the NIH (1S10RR08389). We thank
R. Owen and J . Pontrello for helpful discussions.
Syn th esis of 1-Dea za r ibofla vin (2). Compound 12 (25 mg,
0.056 mmol) was dissolved in ammonia-saturated methanol (3
mL), and the resulting solution was allowed to stir at room
temperature for 48 h. Solvent was removed under reduced
pressure. Crude material was purified by HPLC (100% f 60%
A:B over 20 min, retention time about 18 min). This resulted in
the purple solid 2 (7 mg, 33%) as the TFA salt: mp > 300° (dec);
1H NMR (300 MHz, d6-DMSO) δ 11.1 (s, 1 H, exchange D2O),
7.61 (s, 1 H), 7.58 (s, 1 H), 7.22 (b s, 1H, D2O exchange) 5.51 (s,
1 H), 5.20-3.95 (m, 7 H), 2.33 (s, 3 H), 2.23 (s, 3 H) ppm; 13C
NMR δ 164.4 (TFA), 159.9, 144.4, 142.1, 140.9, 133.1, 133.0,
132.5, 131.0, 115.9 (TFA), 86.7, 73.5, 72.9, 63.4, 48.2, 20.6, 18.6
ppm; ESI (m/z) [M + Na] calcd for C18H21N3O6 398.1, found
398.0.
Su p p or tin g In for m a tion Ava ila ble: General experimen-
1
tal procedures and H and 13C NMR spectra for the complete
syntheses of compounds 1 and 2. This material is available
free of charge via the Internet at http://pubs.acs.org.
J O049859F
J . Org. Chem, Vol. 69, No. 7, 2004 2617