Synthesis of a 'twisted' transition-state analogue of biotin

Article abstract of DOI:10.1016/j.tetlet.2003.12.087

A facile, racemic synthesis of a 'twisted' transition-state analogue of biotin is described. A key reaction is the electronically assisted ring-closure of the sulfur containing ring by displacement of an in situ generated mesylate by a suitably positioned 4-methoxybenzyl sulfide. The crystal structure of tricyclic compound 6 shows the AB ring system to indeed be twisted. The 'twist' was introduced to examine the possible involvement of sulfur participation in biotin biochemistry.

Full text of DOI:10.1016/j.tetlet.2003.12.087

Tetrahedron
Letters
Tetrahedron Letters 45 (2004) 1777–1780
Synthesis of a ÔtwistedÕ transition-state analogue of biotin
Andrew S. Grant,a,* Kleem Chaudhary,b Lisa Stewart,b Andrea Peters,b
Spencer Delislea and Andreas Deckenc
aDepartment of Chemistry, Mount Allison University, Sackville, NB, Canada E4L 1G8
bDepartment of Chemistry, University of Winnipeg, Winnipeg, MB, Canada R3B 2E9
cDepartment of Chemistry, University of New Brunswick, Fredericton, NB, Canada E3B 6E2
Received 5 July 2003; revised 11 December 2003; accepted 12 December 2003
Abstract—A facile, racemic synthesis of a ÔtwistedÕ transition-state analogue of biotin is described. A key reaction is the electron-
ically assisted ring-closure of the sulfur containing ring by displacement of an in situ generated mesylate by a suitably positioned
4-methoxybenzyl sulfide. The crystal structure of tricyclic compound 6 shows the AB ring system to indeed be twisted. The ÔtwistÕ
was introduced to examine the possible involvement of sulfur participation in biotin biochemistry.
ꢀ 2003 Elsevier Ltd. All rights reserved.
The question of whether sulfur participates in biotin
biochemistry via a transannular interaction has attracted
the interest of several research groups over the years.1;2
The idea is that such an interaction could (a) increase the
nucleophilicity of N1 thereby facilitating attack by N1
onto the ÔCO2 equivalentÕ; either carboxyphosphate or
CO2 itself and (b) avoid the formation of a ureide anion
intermediate. Despite considerable effort to find experi-
mental evidence for such an interaction, the general
consensus is that it probably doesnÕt exist.3
stabilized a conformation of biotin that broke the
symmetry of the bicyclic portion of the molecule by
pulling the sulfur off of the centerline of ring A. This
distortion, or ÔtwistÕ, results in net overlap, in the form
of a 2-orbital/4 electron destabilizing interaction,
between sulfur and the ureido moiety.
To find experimental evidence for this through-bond/
through-space interaction, what was needed was a
molecule whose AB ring structure was permanently
twisted in its ground state. Examination of hand held
models and some preliminary modeling suggested that
incorporation of a six-membered ring as in 1 below
would induce the appropriate twist. As 1 has structural
characteristics of potential interest to a wider audience,
we wish to report an efficient synthesis of racemic 1. An
analysis of the crystal structure of the thiooxazolidone
precursor 6 shows the molecule to indeed be twisted.
N1
O
H
H
N
N
A
H
H
B
S
HO2C
H
biotin
O
We recently reported a molecular orbital (MO) study
that suggested that during the first half-reaction of
biotin-dependent enzyme-catalyzed carboxylations, sul-
fur could enhance the nucleophilicity of N1 via a com-
bined through-bond/through-space interaction between
sulfur and the urea.4 This interaction could occur if, in
the transition-state for biotin carboxylation, the enzyme
H
O
N
H
S
H
1
The synthesis of 1 was fashioned after VolkmannÕs
synthesis of biotin.5 Hence, the oxazolidone ring would
be formed via the addition of the anion of ethyl iso-
thiocyanatoacetate to a suitably a-thio-functionalized
Keywords: Organic synthesis; Bioorganic.
* Corresponding author. Tel.: +1-506-364-2368; fax: +1-506-364-2313;
e-mail: agrant@mta.ca
0040-4039/$ - see front matter ꢀ 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2003.12.087

1778
A. S. Grant et al. / Tetrahedron Letters 45 (2004) 1777–1780
cyclohexanone. Unmasking of the thiol and suitable
manipulation of the ester would hopefully lead to clo-
sure of ring-B, assuming addition of the isothiocyanate
to the ketone would yield at least some of the desired
diastereomer. The choice of a suitable protecting group
for the thiol then became the main issue.
5a showed that the sulfide is ideally positioned to dis-
place the mesylate as soon as it is formed. It seemed
reasonable that an electron donating group on the
aromatic group would accelerate this intramolecular
ring closure. We therefore decided to run through the
synthesis again with a methoxy group in the para posi-
tion.
We eventually settled on the use of a benzyl group, more
out of convenience than foresight, but the choice proved
to be a useful one. We initially synthesized a-S-benzyl-
cyclohexanone 2a via displacement of bromine from
a-bromocyclohexanone with the anion of benzyl
mercaptan. However we found the synthesis of
a-bromocyclohexanone troublesome, and as shown in
Scheme 1 we adopted the methodology of Scholz,6
wherein a-S-benzylcyclohexanone is prepared in one
step from the enolate of cyclohexanone and the thio-
lating agent S-benzyl 4-methylbenzenethiosulfonate 3a.
Reagent 3a is prepared from potassium thiotosylate and
benzyl chloride in DMF. With a convenient preparation
of a key starting material in hand, the synthesis pro-
ceeded as outline in Scheme 1.
The requisite thiolating agent 3b was obtained by stir-
ring 4-methoxybenzyl chloride with potassium thio-
tosylate in DMF for 12 h at room temperature.
Treatment of the enolate of cyclohexanone with 3b
yielded 2b, which had the advantage over 2a in that it
crystallized from ether, thus avoiding a chromato-
graphic step.8 The anion of ethyl isothiocyanatoacetate
added to ketone 2b again yielding a 60:40 mixture of
diastereomers.9 Fortuitously, the major desired isomer
4b crystallized from an ether solution of the crude
worked-up reaction in 50% overall yield. The remaining
10% can be recovered by column chromatography.
The stereochemistry of the minor isomer was eventually
shown by X-ray analysis of a single crystal to be 4c as
indicated in Scheme 1. This result shows that the iso-
thiocyanate added with complete facial selectivity to the
hydrogen side of ketone 2b.
Treatment of ethyl isothiocyanatoacetate with an
equivalent of lithium bis-(trimethylsilyl)amide in THF
at )78 ꢁC generated the anion, which added smoothly to
a-S-benzylcyclohexanone to give a 60:40 separable
mixture of stereoisomers, the major one proving even-
tually to be the desired diastereomer 4a. The ester group
of 4a was reduced with NaBH4 in THF/MeOH (9:1) at
room temperature over a period of several hours to give
alcohol 5a in moderate yield. Treatment of 5a with a
slight excess of methanesulfonyl chloride in pyridine as
solvent yielded the tricyclic thiooxazolidone 6 directly,
without isolation of the mesylate, again in moderate
yield, over a period of several hours.7
Reduction of ester 4b with NaBH4 again seemed slug-
gish, but fresh LiBH4 in THF/MeOH (9:0) reduced the
ester cleanly to the alcohol 5b within minutes.10 Treat-
ment of 5b in pyridine with a slight excess of methane-
sulfonyl chloride gave the desired ring-closed product 6
cleanly within 10 min, thus confirming our hunch that
the 4-methoxy group would accelerate the reaction.11
The sulfur/oxygen exchange was again taken from
Volkmann.5 Heating thiooxazolidone 6 with 2-bromo-
ethanol in DMF at 100 ꢁC for 2 h, followed by addition
of 6 M KOH and stirring at 50 ꢁC for an additional 2 h
We were delighted at this one step ring closure, but
overall yields were low. Hand held models of the alcohol
S
S
H
H
O
H
N
O
H
1) LiBTMSA
S
N
O
CO2Et
O
S
O
C
THF, -78C
N
CO2Et
OEt
+
H
S
S
2)
H
O
THF, -78C
CH3C6H4
S
S
X
O
X
X
OCH3
4a X=H
4b X=OCH3
2a X=H
3a X=H
4c
2b X=OCH3
3b X=OCH3
LiBH4
THF, MeOH
S
S
O
H
O
H
H
N
H
BrCH2CH2OH
DMF
O
MsCl, pyr
CH2Cl2
N
O
N
CH2OH
KOH
S
S
S
H
H
6
1
X
5a X=H
5b X=OCH3
Scheme 1.

A. S. Grant et al. / Tetrahedron Letters 45 (2004) 1777–1780
1779
goal is to examine the possibility that sulfur could par-
ticipate in biotin biochemistry via a combined through-
bond/through-space interaction between sulfur and N1
that operates at the transition-state for carboxylation.4
Acknowledgements
Financial support for this work from the University of
Winnipeg, NSERC of Canada and Mount Allison
University is gratefully acknowledged.
Figure 1. Crystal structure of thiooxazolidone 6.
yielded 1 in variable yields ranging from 40–100%.13
Starting material and product are conveniently sepa-
rated on silica gel using CH2Cl2:acetone (20:1) as eluent.
Yields based on recovered starting material were greater
than 90%.
References and notes
1. Mildvan, A. S.; Scrutton, M. C.; Utter, M. F. J. Biol.
Chem. 1966, 241, 3488.
We were unable to grow suitable crystals of 1, but
crystals of thiooxazolidone
2. For a recent review of biotin chemistry see: Attwood,
P. V.; Wallace, J. C. Acc. Chem. Res. 2002, 35, 113–120.
3. (a) Berkessel, A.; Breslow, R. Bioorg. Chem. 1986, 249–
261; (b) Perrin, C. L.; Dwyer, T. J. J. Am. Chem. Soc.
1987, 109, 5163–5167; (c) Knowles, J. R. Annu. Rev.
Biochem. 1989, 58, 195–221; (d) Berkessel, A. J. Org.
Chem. 1989, 54, 1685–1692.
6
(mp ¼ 223 ꢁC) were
obtained via slow evaporation of chloroform at room
temperature. As shown in Figure 1, X-ray analysis12
confirmed the overall structure and revealed the precise
nature of the twist.
4. Grant, A. S. J. Mol. Struct. (THEOCHEM) 1998, 422,
79–87.
5. Volkmann, R. A.; Davis, J. T.; Meltz, C. N. J. Am. Chem.
Soc. 1983, 105, 5946–5948.
6. Scholz, D. Liebigs. Ann. Chem. 1984, 259–263.
7. For a similar ring-closure see: Eames, J.; Kuhnert, M.;
Jones, R. V. H.; Warren, S. Tetrahedron Lett. 1998, 39,
1247–1250.
A key indicator of the twist is the dihedral angle N(10)–
C(3)–C(3A)–O(8) ¼ )13ꢁ obtained from the X-ray data.
The equivalent dihedral in biotin17 (N(1)–C(4)–C(3)–
N(3)) is essentially 0ꢁ. In dethiobiotin18 it is +22ꢁ and in
oxybiotin19 it is )12ꢁ. It has been suggested that a pos-
sible role for sulfur in biotin is to maintain the planarity
of the 2-imidazolidone ring18 for the purpose of pre-
venting N1-carboxybiotin from spontaneously decarb-
oxylating.20 Compound 6 then has sulfur in ring-B but is
twisted (presumably 1 is as well), which satisfies our
definition4 of a transition-state analogue for the car-
boxylation of biotin. We are interested to know whether
that twist translates into unique chemistry at nitrogen.
8. Synthesis of 2b: Lithium bis(trimethylsilyl)amide (17 mL)
was added to a stirred solution of cyclohexanone (1.5 g,
17 mmol) in dry THF (20 mL) at )78 ꢁC under argon.
After 5 min, a solution of toluene-4-thiosulfonic acid S-(4-
methoxybenzyl)ester 3b (2.0 g, 16 mmol) in dry THF
(10 mL) was added to the enolate over a period of
2–3 min. Stirring was continued at )78 ꢁC for 5 min, the
ice bath was removed, and after an additional 5 min, the
reaction was quenched with 15 mL of saturated NH4Cl.
The quenched reaction mixture was transferred to a
separatory funnel, diluted with EtOAc (75 mL) and the
aqueous layer drained off. The organic layer was washed
with 10% HCl (3 · 25 mL), saturated NaHCO3 (1 · 30 mL)
and brine (1 · 25 mL), dried with MgSO4, filtered and
evaporated to dryness. The crude oil was taken up in
3–4 mL of ether and placed in the fridge over night to yield
1 g of white crystalline product. Mp 60 ꢁC; 1H NMR
(270 MHz, CDCl3) d ¼ 7:24 (d, 2H), 6.84 (d, 2H), 3.81 (s,
3H), 3.63 (s, 2H), 3.74 (t, 1H), 3.0 (m, 1H), 2.24 (m, 1H),
2.10 (m, 1H), 1.95 (m, 2H), 1.70 (m, 3H); 13C NMR
(270 MHz, CDCl3) d ¼ 208:12, 158.78, 130.30, 129.62,
113.95, 55.32, 50.77, 37.77, 35.12, 32.81, 26.94, 21.93; IR
A consequence of the twist can be appreciated from a
comparison of the S(1)–O(8), S(1)–C(9), and S(1)–N(10)
bond distances obtained from the X-ray data of 6
ꢀ
(3.1043, 3.3905, 3.5289 A, respectively) with the corre-
sponding distances obtained from the energy minimized
structure (6-31Gꢀ) of the nontwisted bicyclic analogue
that lacks the six-membered ring (3.2920, 3.3605,
ꢀ
3.6501 A, respectively). The S(1)–O(8) and S(1)–N(10)
distances are similar in the nontwisted structure. The
same distances in the X-ray structure show that S(1) has
moved closer to the oxygen and further from nitrogen.
Also, S(1) has moved closer to the thiocarbonyl carbon;
ꢀ
ꢀ
3.5289 A versus 3.6541 A for the nontwisted structure.
The N–C–C–O dihedral angle in the nontwisted struc-
ture is essentially 0ꢁ as in biotin. Therefore, the effect of
adding the six-membered ring is to twist the AB ring
portion of the molecule ()13ꢁ dihedral) moving the
sulfur in ring-B up and over to the left relative to ring-A.
(CHCl3) 2940(s), 1697(s), 1610(s), 1511(s) cmꢁ1
.
9. Synthesis of 4b: A THF (20 mL) solution of ethyl
isothiocyanatoacetate (3.0 g, 20 mmol) was added to a
dry THF (20 mL) solution of lithium bis(trimethyl-
silyl)amide (20 mL) at )78 ꢁC over a period of 2 min. To
the resultant anion was added compound 2b (5.1 g,
20 mmol) dissolved in dry THF (20 mL) over a period of
5 min. The reaction was stirred for 10 min at )78 ꢁC, the
ice bath was removed, stirred for an additional 15 min, and
the reaction quenched with saturated NH4Cl. The
quenched reaction mixture was poured into a separatory
funnel and diluted with EtOAc (200 mL). The aqueous
Assuming the structures of 1 and 6 are similar, we are
currently in the process of studying 1 and some related
compounds to determine what effect the ÔtwistÕ has in
terms of chemical reactivity at nitrogen, and these
results will be reported shortly. As mentioned, the overall

1780
A. S. Grant et al. / Tetrahedron Letters 45 (2004) 1777–1780
layer was drained off, and the organic layer was washed
12. Crystals of 6 were grown by slow evaporation of chloro-
form at room temperature. Single crystals were coated
with Paratone-N oil, mounted using a glass fiber and
frozen in the cold nitrogen stream of the goniometer. A
hemisphere of data was collected on a Bruger AXS P4/
SMART 1000 diffractometer using x and h scans with a
scan width of 0.3ꢁ and 30 s exposure times. The dector
distance was 5 cm. The data were reduced (SAINT)14 and
corrected for absorption (SADABS).15 The structure was
solved by direct methods and refined by full-matrix least
squares on F2(SHELXTL).16 All nonhydrogen atoms were
refined anisotropically. Hydrogen atoms were found in
Fourier difference maps and refined isotropically.
with 10% HCl (3 · 35 mL), saturated NaHCO3 (1 · 35 mL)
and brine (1 · 35 mL). Organic layer was dried (MgSO4),
filtered, and concentrated down to a volume of about
30 mL and placed in the fridge for 12 h. The excess EtOAc
was pipetted off to yield 2.3 g of crystalline 4b. An
additional 1.3 g of crystalline 4b was obtained from the
EtOAc layer by concentrating and seeding it. Compound
4b recrystallizes from EtOAc. Isolated yield 50%. Mp
153 ꢁC. 1H NMR (270 MHz, CDCl3) d ¼ 7:71 (s, 1H), 7.22
(m, 2H), 6.82 (m, 2H), 4.21 (s, 1H), 4.20 (m, 2H), 3.84 (d,
1H), 3.80 (s, 3H), 3.69 (d, 1H), 3.10 (dd, 1H), 2.25 (m, 1H),
1.85 (m, 6H), 1.28 (t, 3H); 13C NMR (CDCl3) d ¼ 187:77,
167.06, 158.73, 130.25, 129.55, 113.89, 95.04, 66.56, 62.39,
55.28, 47.69, 39.96, 34.23, 30.11, 25.44, 21.57, 13.99; IR
(CHCl3) 3445, 2989, 2942, 1733, 1495; IR (CHCl3)
13. Synthesis of compound 1: 2-Bromoethanol (100 mg) was
added to a DMF (5 mL) solution of thioanalogue 6
(65 mg, 0.3 mmol) and the resultant mixture was heated to
100 ꢁC in an oil bath for 2 h. The reaction was allowed to
cool to 50 ꢁC, a 6 M solution of KOH (1 mL) was added
and the reaction stirred at that temperature for an
additional 2 h. After cooling to room temperature, the
reaction was diluted with EtOAc (50 mL) and washed with
water (3 · 20 mL) and brine (1 · 20 mL). The aqueous
washings were re-extracted with EtOAc (25 mL), the
EtOAc layer was washed with water (1 · 20 mL) followed
by brine (1 · 20 mL). The two organic layers were com-
bined, dried (MgSO4) filtered and evaporated to yield a
solid. TLC of the organic layer in 20:1 CH2Cl2:acetone
shows starting material (rf ¼ 0.6) and product (rf ¼ 0.3).
Column chromatography of the crude reaction mixture
with the above solvent mixture and isolation of the lower
spot gave product in a yield ranging from 40–100%
(always >95% based on recovered starting material),
which was re-crystallized (white filaments) from CHCl3.
3446(m), 2942(m), 1735(s), 1494(s), 1251(s) cmꢁ1
.
10. Synthesis of 5b: Thiooxazolidone ester 4b (2.3 g, 5.8 mmol)
was dissolved in a 9:1 mixture of /MeOH (20 mL). To this
solution was added solid LiBH4 (500 mg) and the resulting
(heterogeneous) solution was stirred for 15 min. (TLC, 2:1
hexane:EtOAc) showed reaction to be complete after 5–
6 min. The reaction was quenched by drop-wise addition
of 10% HCl (10 mL) over a period of 30 min. The reaction
was then partitioned between ethyl acetate (100 mL) and
water (50 mL). The organic layer was washed with
saturated bicarbonate (25 mL) then brine (25 mL), dried
over MgSO4, filtered and evaporated to dryness. Recrys-
tallization from EtOAc yields 2.0 g (95% yield) of a white
solid, mp 145 ꢁC; 1H NMR (270 MHz, CDCl3) d ¼ 8:33 (s,
1H, NH), 7.29 (m, 2H), 6.87 (m, 2H), 3.95 (m, 5H), 3.84 (s,
3H), 3.14 (t, 1H), 2.85 (s, 1H), 2.24 (m, 1H), 1.95 (m, 1H),
1.75 (m, 4H), 1.40 (m, 2H); 13C NMR (d6-DMSO) d
187.17, 158.76, 130.75, 130.46, 114.34, 92.01, 66.07, 50.10,
55.60, 45.77, 35.21, 31.38, 23.43, 22.37; IR (CHCl3)
1
Mp 228 ꢁC; H NMR (270 MHz, CDCl3) d ¼ 6:2 (s, 1H,
NH), 3.97 (d, 1H, NCH), 2.88 (2m, 2H, SCHÕs), 2.7 (d,
1H, SCH), 2.3 (d, 1H), 1.9 (m, 4H), 1.7 (m, 1H), 1.4 (m,
2H); 13C NMR (CDCl3) d ¼ 159:06, 91.83, 62.53, 57.12,
40.80, 34.13, 25.42, 24.99, 21.10; IR (CHCl3) 3461(m),
3441(m), 3355(m), 2944(s), 1610(w), 1511(s), 1178(s) cmꢁ1
.
11. Synthesis of compound 6: Thiooxazolidone alcohol 5b
(0.95 g, 2.9 mmol) was dissolved in pyridine (5 mL) and to
this solution was added methanesulfonyl chloride (350 mg,
3.0 mmol). The reaction is stirred at room temperature for
2 h. TLC in hexane/EtOAc (1:1) shows lose of starting
material at rf ¼ 0.4 and the appearance of product at
rf ¼ 0.3 (iodine). The reaction was quenched by the
addition of EtOAc (50 mL) followed by 10% HCl
(100 mL). The organic layer was washed with saturated
bicarbonate (1 · 25 mL) and brine (1 · 30 mL). The organic
layer is dried (MgSO4) filtered and evaporated to yield a
solid. The product recrystallizes from EtOAc: mp 223 ꢁC;
1H NMR (270 MHz, CDCl3) d ¼ 8:80 (s, 1H, NH), 4.19 (s,
1H, NCH), 2.95 (m, 3H, SCHÕs), 2.25 (d, 1H), 2.0 (m, 4H),
1.80 (m, 1H), 1.60 (m, 1H) 1.40 (m, 1H); 13C NMR
(CDCl3) d ¼ 188:57, 99.72, 66.02, 57.34, 40.50, 33.62,
25.30, 24.93, 20.99; IR (CHCl3) 3446(w), 3185(w),
3259(w), 2944(s), 1752(s), 1402(w), 1371(w) cmꢁ1
.
14. SAINT 6.02, 1997–1999, Bruker AXS, Madison, Wiscon-
sin, USA.
15. SADABS George Sheldrick, 1999, Bruker AXS, Madison,
Wisconsin, USA.
16. SHELXTL 5.1, George Sheldrick, 1997, Bruker AXS,
Madison, Wisconsin, USA.
17. DeTitta, G. T.; Edmonds, J. W.; Stallings, W.; Donohoue,
J. J. Am. Chem. Soc. 1976, 98, 1920–1926.
18. Chen, C.-S.; Parthasarathy, R.; DeTitta, G. T. J. Am.
Chem. Soc. 1976, 98, 4983–4990.
19. DeTitta, G. T.; Parthasarathy, R.; Blessing, R. H.;
Stallings, W. C. Proc. Natl. Acad. Sci. USA 1980, 77,
333–337.
20. Tipton, P. A.; Cleland, W. W. J. Am. Chem. Soc. 1988,
110, 5866–5869.
2946(m), 1494(s), 1365(w), 1290(w), 1186(s) cmꢁ1
.