Immucillins in custom catalytic-site cavities

Article abstract of DOI:10.1016/j.bmcl.2008.08.047

Neighboring-group participation in the reaction catalyzed by purine nucleoside phosphorylase involves a compression mode between the 5′- and 4′-ribosyl oxygens, facilitated by His257. The His257Gly mutant opens a space in the catalytic site. Hydrophobic 5′-substituted Immucillins are transition-state analogue inhibitors of this mutant enzyme. Dissociation constants as low as 2 pM are achieved, with K_m/K_d as high as 400,000,000.

Full text of DOI:10.1016/j.bmcl.2008.08.047

Bioorganic & Medicinal Chemistry Letters 18 (2008) 5900–5903
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Immucillins in custom catalytic-site cavities
Andrew S. Murkin a, Keith Clinch b, Jennifer M. Mason b, Peter C. Tyler b, Vern L. Schramm a,
*
a Department of Biochemistry, Albert Einstein College of Medicine of Yeshiva University, 1300 Morris Park Avenue, Bronx, NY 10461, USA
b Carbohydrate Chemistry Team, Industrial Research Limited, PO Box 31310, Lower Hutt, New Zealand
a r t i c l e i n f o
a b s t r a c t
Article history:
Neighboring-group participation in the reaction catalyzed by purine nucleoside phosphorylase involves a
compression mode between the 50- and 40-ribosyl oxygens, facilitated by His257. The His257Gly mutant
opens a space in the catalytic site. Hydrophobic 50-substituted Immucillins are transition-state analogue
inhibitors of this mutant enzyme. Dissociation constants as low as 2 pM are achieved, with Km/Kd as high
as 400,000,000.
Received 13 June 2008
Revised 11 August 2008
Accepted 12 August 2008
Available online 19 August 2008
Ó 2008 Elsevier Ltd. All rights reserved.
Keywords:
PNP
Immucillin
ImmH
Mutant
Binding
Transition-state analogue
Purine nucleoside phosphorylase (PNP)1 catalyzes the phospho-
rolytic cleavage of inosine and guanosine as well as their 20-deoxy
analogues to (deoxy)ribose 1-phosphate and hypoxanthine or
guanine (Fig. 1).
Inhibition of human PNP activity causes accumulation of deoxy-
guanosine, which in turn leads to downstream inhibition of cell
division and apoptosis specifically in T-lymphocytes.2–4 Thus,
PNP has been identified as a target for the treatment of T-cell
lymphoma, rheumatoid arthritis, psoriasis, multiple sclerosis, and
other T-cell mediated disorders.5
The human PNP-catalyzed reaction has been shown to proceed
through a dissociative transition state characterized by a ribo-
oxacarbenium ion with a cationic C-10, which is separated from
the nucleobase by >3 Å.6 These features were incorporated into a
family of transition-state analogue inhibitors called Immucillins
(Fig. 2), whose members typically possess low nanomolar to
picomolar affinity for human PNP. The potent inhibitors exhibit
slow-onset behavior, where a time-dependent conformational
change converts the initial enzyme–inhibitor complex (EꢀI), with
dissociation constant Ki, to an even more stable complex (E*ꢀI), with
dissociation constant Ki*. The first-generation inhibitor, Immucil-
lin-H (ImmH, 1),7 has a Ki of 3.3 nM and a Ki* of 58 pM.8 Modifica-
tion of ImmH produced second- and third-generation Immucillins
represented by DADMe-ImmH (4)9 and SerMe-ImmH (9),10 respec-
tively, with Ki* values of 11 and 5.2 pM, respectively.
Mutation of His257 has shown that this residue serves an
important role in transition-state formation by hydrogen-bonding
with the 50-OH, orienting it into an electron-rich ‘oxygen stack’
with O-40 and OP from the phosphate nucleophile (Fig. 3).8 This
interaction is reasoned to provide electron density to weaken the
ribosidic bond and stabilize the developing cationic transition
state. Though mutation adversely affected the steady-state proper-
ties of PNP, mutants were capable of binding ImmH and DADMe-
ImmH with reasonably good affinity. Structural analysis revealed
that one of the mutants, His257Gly, was capable of binding these
inhibitors nearly identically to the native protein, despite side-
chain removal (Fig. 4). We envisioned that the active-site cavity
introduced in the His257Gly mutant could be exploited in the
binding of bulkier Immucillin derivatives.
50-Methylthio-ImmH (MeS-ImmH, 2) and 50-phenylthio-ImmH
(PhS-ImmH, 3)11 were originally developed as specific inhibitors
of PNP from Plasmodium falciparum, exhibiting 112- and 2-fold
binding preference over human PNP, respectively.12 With native
human PNP, 2 and 3 bind with weaker affinity than 1, showing
no slow-onset behavior and yielding Ki values of 101 and
160 nM, respectively (Table 1). In comparison to the substrate ino-
sine (Km = 40 l
M), modification of the 50-hydroxyl of ImmH re-
sulted in a drop in relative affinity (Km/Ki) from 690,000 to 400
and 250, respectively. Mutation of His257 to glycine abolished
the slow-onset character of ImmH (1), resulting in a Ki of
11.0 nM. Taken in light of the elevated Km (750 lM) with this mu-
tant, the relative affinity decreased 10-fold relative to the native
enzyme [(mutant Km/Ki)/(native Km/Ki) = 0.099]. In the case of the
bulkier derivatives 2 and 3, however, not only was increased
* Corresponding author. Tel.: +1 718 430 2813; fax: +1 718 430 8565.
E-mail address: vern@aecom.yu.edu (V.L. Schramm).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.08.047

A. S. Murkin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5900–5903
5901
Figure 1. Phosphorolysis of inosine catalyzed by PNP.
Figure 2. Three-generations of Immucillins, potent PNP transition-state analogue inhibitors. ImmH, DADMe-ImmH, and SerMe-ImmH are shown along with 50-alkylthio and
arylthio derivatives used in this study.
absolute affinity observed (Ki = 4.9 and 6.0 nM, respectively),
but also the relative affinity increased to 150,000 and 120,000,
respectively, up to 500-fold over native PNP.
(PrS-DADMe-ImmH, 6)12 maintained strong potency with the na-
tive enzyme with Ki* values of 19.6 pM (Km/Ki = 2,000,000) and
9.8 pM (Km/Ki = 4,100,000), respectively. These inhibitors were
found to be strikingly effective with the His257Gly mutant, exhib-
iting slow-onset inhibition and yielding Ki* values of only 2.8 and
1.9 pM, respectively. Not only are these values lower than the most
potent inhibitor tested to date with native human PNP, but also the
relative affinities (Km/Ki) of 270,000,000 and 400,000,000 are the
largest ever reported for any enzyme–inhibitor system. Inhibitors
5 and 6 are therefore bound with 130- and 97-fold preference,
respectively, over the native protein.
The second-generation transition-state analogue DADMe-
ImmH (4) was designed to mimic the human PNP transition state
by increasing the leaving group distance through introduction of
a methylene bridge between the pseudoribosidic bond and by
moving the cationic ring nitrogen to the 10-position, where signif-
icant positive-charge character is developed at the transition
state.13 With the native enzyme, DADMe-ImmH was found to be
a more potent inhibitor than ImmH, giving a Ki* of 10.7 pM.8 Unlike
the case with ImmH, the His257Gly mutant bound DADMe-ImmH
in a slow-onset manner, with only slightly lowered relative affinity
[(mutant Km/Ki)/(native Km/Ki) = 0.74]. 50-Methylthio-DADMe-
ImmH (MeS-DADMe-ImmH, 5) and 50-propylthio-DADMe-ImmH
Further derivatization of 5 to ( )-50-deoxy-40-fluoro-50-methyl-
thio-DADMe-ImmH (40-F-MeS-DADMe-ImmH, 7) was accom-
plished by functional group exchange from the fluorinated diol
precursor 11, followed by Mannich reaction of 13 with 9-deaza-

5902
A. S. Murkin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5900–5903
Scheme 1. Reagents and conditions: (a) Bu2SnO, toluene, reflux, then MsCl, 81%; (b)
NaSMe, DMF, 68%; (c) 6 N HCl; (d) 9-deazahypoxanthine, CH2O, NaOAc, water–
dioxane, 100 °C, 30%.
Figure 3. Proposed role of His257 in formation of the transition state, featuring
dynamic compression of the O50-O40-OP ‘oxygen stack’. The ‘oxygen stack’ is
represented by hashed bonds connecting bolded atoms, and arrows indicate
promoting vibrational modes. Dashed bonds represent hydrogen bonds and partial
bonds. Dynamic compressive motions (larger solid arrows) from the enzyme push
O-50 and the phosphate oxygen toward the ring oxygen, providing electron density
to stabilize ribooxacarbenium-ion development and promote ribosidic bond fission.
Me-ImmH (40-OH-MeS-DADMe-ImmH, 8)12 resulted in the loss of
slow-onset inhibition with native and mutant enzymes. These
compounds bound nearly equally well to native PNP with relative
affinities of 7500 and 5100, respectively. As in the case with 5 and
6, when compounds 7 and 8 were tested with the glycine mutant,
the dissociation constants dropped, resulting in enhanced relative
affinities of 540,000 and 600,000, respectively, corresponding to
78- and 120-fold improvements over the native enzyme.
The third generation of PNP transition-state inhibitors consist of
acyclic analogues of DADMe-ImmH. SerMe-ImmH (9), despite its
lack of stereocenters, binds to human PNP with a Ki* of 5.2 pM.10
As was observed with ImmH, mutation of His257 resulted in the
loss of slow-onset behavior, lowering the relative affinity from
7,700,000 to 260,000. The corresponding ratio of the relative affin-
ities is 0.033, which is the lowest among the 10 inhibitors tested.
SerMe-ImmH was also derivatized to the methylthio analogue 10
in a manner similar to that outlined for compound 7 above
(Scheme 2).15 As observed for compounds 7 and 8, incubation of
MeS-SerMe-ImmH (10) with native and mutant PNP lacked the
slow-onset property of its underivatized analogue, yielding a lower
Km/Ki with the native protein (9300) but a larger value (680,000)
with His257Gly. The discrimination for 10 by His257Gly is
therefore 73-fold greater than that by native PNP.
This study has confirmed that human PNP tolerates substitution
of the 50-hydroxyl of the transition-state analogues ImmH, DAD-
Me-ImmH, and SerMe-ImmH with alkylthio and arylthio groups;
however, except for compounds 5 and 6, the slow-onset nature
of inhibition is lost, and inhibitor dissociation constants increase
by two to three orders of magnitude. When the imidazole moiety
of residue 257 is removed by mutation, the loss of an H-bond part-
ner for the 50-OH likely accounts for the observed decreases
in binding affinity for the unmodified analogues 1, 4, and 9.
Figure 4. Overlay of the crystal structures of native human PNP and His257Gly
complexed with ImmH and phosphate (PO4). Side chains of selected active-site
residues within 3.2 Å of ImmH have been included. Carbon atoms of these residues
and of ImmH are green in the native enzyme and cyan in His257Gly. PO4 is colored
yellow and black in the native and mutant PNPs, respectively. The H-bond between
His257 and the 50-OH is indicated.
hypoxanthine and formaldehyde (Scheme 1).14 Both the fluori-
nated derivative 7 and 50-deoxy-40-hydroxy-50-methylthio-DAD-
Table 1
Dissociation constants of transition-state analogues with native human PNP and His257Glya
Inhibitor
Native human PNP (inosine Km = 40
l
M)b
Km/Kid
690,000
His257Gly (inosine Km = 750
Ki* (nM)c
l
M)b
Km/Kid
Mutant Km /Ki
Native Km/Ki
Ki (nM)
3.3 0.2
101
Ki* (nM)c
Ki (nM)
ImmH (1)b
MeS-ImmH (2)
PhS-ImmH (3)
0.0579 0.0015
n/a
11.0 0.9
4.9 0.5
6.0 0.3
e
n/a
n/a
n/a
0.27 0.02
68,000
150,000
120,000
2,800,000
270,000,000
400,000,000
540,000
600,000
260,000
0.099
390
500
0.74
130
97
4
400
250
160 15
n/a
DADMe-ImmH (4)b
MeS-DADMe-ImmH (5)
PrS-DADMe-ImmH (6)
40-F-MeS-DADMe-ImmH (7)
40-OH-MeS-DADMe-ImmH (8)
SerMe-ImmH (9)
1.10 0.12
0.10 0.02
0.117 0.015
5.8 0.4
0.0107 0.0011
0.0196 0.0012
0.0098 0.0007
n/a
3,700,000
2,000,000
4,100,000
7500
0.11 0.02
0.066 0.006
1.40 0.04
1.26 0.11
2.93 0.04
1.10 0.03
0.0028 0.0003
0.0019 0.0002
n/a
n/a
n/a
n/a
78
7.9 1.1
n/a
5100
120
0.033
73
0.11 0.02
4.3 0.2
0.0052 0.0004
n/a
7,700,000
9300
MeS-SerMe-ImmH (10)
680,000
a
All inhibition constants reported here were determined by the reported xanthine-oxidase-coupled assay (Ref. 12) using native or mutant enzyme from the same
preparation.
b
c
Values are from Ref. 8. All other values in this table are either newly determined or have been redetermined in this study to ensure reliable comparisons.
Ki* is the final, equilibrium dissociation constant for the slow-onset, tight-binding phase of inhibition. ‘n/a’ indicates that no slow-onset phase was observed.
This ratio is Km/Ki* in cases where slow-onset inhibition occurs.
d
e
The weak inhibition phase (Ki) was observed but too short to accurately quantitate, so only Ki* is reported.

A. S. Murkin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 5900–5903
5903
81%), which was taken up in DMF (7 ml) and stirred with sodium
thiomethoxide (0.33 g, 4.7 mmol) for 2 h. The mixture was partitioned
between water and diethyl ether, and the organic phase was concentrated
under reduced pressure. Column chromatography (silica gel, 33% EtOAc in
hexanes) gave the title compound 13 (0.28 g, 68% from 12) as a yellow oil. 1H
NMR (300 MHz, CDCl3, TMS) d 4.23–4.16 (br s, 1H), 3.89–3.41 (br m, 3H), 3.24
(m, 1H), 3.02–2.71 (br m, 3H), 2.23 (s, 3H), 1.46 (s, 9H). 13C NMR (75.5 MHz,
CDCl3) (as mixture of invertomers, referenced to the center line of CDCl3 at
77.0 ppm) d154.3/154.1 (C), 101.7/100.9 (d, J = 180 Hz, C), 80.1 (C), 72.8/72.3
(d, J = 19 Hz, CH), 53.2/52.5, d, J = 27 Hz, CH2), 50.2/49.3 (CH2), 36.5 (d,
J = 26 Hz, CH2), 28.4 (CH3), 17.8 (CH3). ESI-MS C11H20FNNaO3
Calcd 288.1046, found 288.1033.( )-cis-1-((9-Deazahypoxanthin-9-yl)methyl)-
4-fluoro-4-(methylthiomethyl)pyrrolidin-3-ol (40-F-MeS-DADMe-ImmH, 7):
S
[M+Na]+;
A
solution of 13 (43 mg, 0.16 mmol) in methanol (1 ml) and conc. HCl (0.5 ml)
was evaporated to dryness under reduced pressure. The resulting deprotected
pyrrolidinol was taken up in water (1.8 ml) and dioxane (0.2 ml). Sodium
acetate (27 mg, 0.32 mmol), 9-deazahypoxanthine (33 mg, 0.24 mmol), and
formaldehyde (37%, 0.026 ml, 0.32 mmol) were added and the solution was
heated at 100 °C for 2.5 h. The solvents were removed under reduced pressure
and the residue was chromatographed on a column of silica gel eluted with
20% methanolic ammonia (7 M) in CH2Cl2. Fractions containing compound 4
were concentrated under reduced pressure. Trituration of this residue with
MeOH gave the title compound 4 (15 mg, 30%) as a white amorphous solid. 1H
NMR (DMSO-d6, TMS) d 11.96 (br s, 1H), 11.83 (br s, 1H), 7.80 (s, 1H), 7.30 (s,
1H), 5.03 (d, J = 6.8 Hz, 1H), 3.82 (m, 1H), 3.69 (m, 2H), 3.11–2.59 (m, 5H), 2.48
(m, 1H), 2.10 (s, 3H). 13C NMR (DMSO-d6, referenced to the solvent center line
at 39.9 ppm) d 154.0 (C), 143.8 (C), 141.7 (CH), 127.4 (CH), 118.0 (C), 113.0 (C),
101.7 (d, J = 186 Hz, CH), 73.1 (d, J = 17 Hz, CH2), 60.4 (d, J = 24 Hz, CH2), 57.9
(CH2), 47.7 (CH2), 38.9 (d, J = 26 Hz, CH2), 17.1 (CH3). ESI-MS C13H18FN4O2S
[M+H]+; Calcd 313.1135, found 313.1136. Dissolution in methanolic HCl and
concentration to dryness gave the hydrochloride salt.
Scheme 2. Reagents and conditions: (a) Boc2O, MeOH; (b) 1 equiv NaH, TBDMSCl,
THF, 71% (2 steps); (c) MsCl, Et3N, CH2Cl2; (d) NaSMe, DMF, 78% (2 steps); (e) HCl,
MeOH, H2O; (f) 9-deazahypoxanthine, CH2O, NaOAc, H2O, 80 °C, 28% (2 steps).
Nevertheless, the active-site cavity that is created accommodates
bulkier functionalities which lack H-bonding opportunities. In all
cases, the His257 mutant exhibited enhanced binding affinities rel-
ative to substrate for bulkier analogues over native PNP by factors
ranging from 73 to 500. Unprecedented selective binding was ob-
served with 5 and 6, which associate up to 400 million times more
tightly with His257Gly than inosine.
15. Experimental
procedure.
( )-tert-Butyl
1-(tert-butyldimethylsilyloxy)-3-
hydroxypropan-2-ylcarbamate (14): Serinol (0.80 g, 8.78 mmol) and Boc
anhydride (2.11 g, 9.66 mmol) were stirred together in MeOH (10 ml) at rt
for 1 h, and the solvent was evaporated. The solid residue was dried over P2O5
under vacuum and then added in portions to a suspension of NaH (60%, 0.35 g,
8.78 mmol) in dry THF (10 ml) with cooling in an ice bath (McDougal, P. G.;
Rico, J. G.; Oh, Y.-I.; Condon, B. D. J. Org. Chem., 1986, 51, 3388). The mixture
was stirred for 45 min, then TBDMSCl (1.32 g, 8.78 mmol) was added. After 2 h,
H2O (4 ml) was added and the mixture was diluted with Et2O (60 ml), washed
with brine, dried (MgSO4), and the solvent was evaporated. Chromatography
on silica gel (EtOAc/hexanes, 2:8) gave 14 as a colorless oil (1.89 g, 70.5%). 1H
NMR (300 MHz, CDCl3, TMS) d 5.15 (br s, partly exchanged to D2O, 1H), 3.88–
3.57 (m, 5H), 2.86 (br s, exchanged to D2O, 1H), 1.45 (s, 9H), 0.90 (s, 9H), 0.08 (s,
6H). 13C NMR (75.5 MHz, CDCl3, referenced to the center line of CDCl3 at
77.0 ppm) d 156.0 (C), 79.6 (C), 64.1 (2 x CH2), 52.6 (CH), 28.4 (CH3), 25.8 (CH3),
18.2 (C), ꢁ5.6 (CH3). ESI-MS C14H31NNaO4Si [M+Na]+ Calcd 328.1920, found
328.1913.( )-tert-Butyl 1-(tert-butyldimethylsilyloxy)-3-(methylthio)propan-2-
ylcarbamate (15). Methanesulfonyl chloride (0.57 ml, 7.31 mmol) was added
to a stirred solution of 14 (1.86 g, 6.09 mmol) and Et3N (1.28 ml, 9.13 mmol) in
CH2Cl2 (15 ml) cooled in an ice bath. The mixture was warmed and stirred at rt
for 30 min, then diluted with CH2Cl2 (60 ml) and washed with sat. aq NaHCO3
(3ꢂ 20 ml), dried (MgSO4), and the solvent was evaporated to give the crude
mesylate. It was dissolved in DMF (10 ml), sodium thiomethoxide (0.85 g,
12.18 mmol) added, and the mixture was stirred at rt for 3 h. Et2O (50 ml) was
added and the mixture was washed with H2O (4ꢂ 10 ml), and brine (10 ml),
dried (MgSO4), and the solvent was evaporated. Chromatography on silica gel
Acknowledgments
We thank Drs. G. Painter and G. Evans of Industrial Research Ltd
for providing inhibitors used in this study.
References and notes
1. Abbreviations: PNP, purine nucleoside phosphorylase; ImmH, Immucillin-H;
DADMe-ImmH, 40-deaza-10-aza-20-deoxy-10-(9-methylene)-ImmH; SerMe-
ImmH, serinol-N-(9-methylene)-ImmH.
2. Giblett, E. R.; Ammann, A. J.; Wara, D. W.; Sandman, R.; Diamond, L. K. Lancet
1975, 1, 1010.
3. Markert, M. Immunodef. Rev. 1991, 3, 45.
4. Oliver, F.; Collins, M.; López-Rivas, A. Experientia 1996, 52, 995.
5. Schramm, V. L. Annu. Rev. Biochem. 1998, 67, 693.
6. Lewandowicz, A.; Schramm, V. L. Biochemistry 2004, 43, 1458.
7. ImmH (1) was synthesized as described in Evans, G. B.; Furneaux, R. H.;
Hutchison, T. L.; Kezar, H. S.; Morris, P. E., Jr; Schramm, V. L.; Tyler, P. C. J. Org.
Chem. 2001, 66, 5723.
8. Murkin, A. S.; Birck, M. R.; Rinaldo-Matthis, A.; Shi, W. X.; Taylor, E. A.; Almo, S.
C.; Schramm, V. L. Biochemistry 2007, 46, 5038.
(EtOAc/hexanes, 5:95) gave 15 as
a
colorless oil (1.59 g, 78%). 1H NMR
9. DADMe-ImmH (4) was synthesized as described in Evans, G. B.; Furneaux, R.
H.; Tyler, P. C.; Schramm, V. L. Org. Lett. 2003, 5, 3639.
(300 MHz, CDCl3, TMS) d 4.89 (bd, partly exchanged to D2O, J = 7.0 Hz, 1H), 3.86
(dd, J = 9.9, 2.9 Hz, 1H), 3.75 (br s, 1H), 3.64 (dd, J = 9.9, 4.1 Hz, 1H), 2.65 (d, J =
7.0 Hz, 2H), 2.14 (s, 3H), 1.45 (s, 9H), 0.90 (s, 9H), 0.06 (s, 6H). 13C NMR
(75.5 MHz, CDCl3, referenced to the center line of CDCl3 at 77.0 ppm) d 155.3
(C), 79.4 (C), 62.7 (CH2), 50.8 (CH), 35.2 (CH2), 28.4 (CH3), 25.9 (CH3), 18.3 (C),
15.9 (CH3), ꢁ5.5 (CH3). ESI-MS C15H33NNaO3SSi [M+Na]+ Calcd 358.1848, found
358.1846. ( )-10-Methylthio-SerMe-ImmH (10): Compound 15 (0.40 g,
1.19 mmol) was dissolved in a 1:1 mixture of MeOH/37% aq HCl (4 ml) and
left at rt for 1.5 h. The solvent was evaporated and the residue was dissolved in
H2O (2 ml) and NaOAc (0.11 g, 1.31 mmol), aq formaldehyde solution (37%,
0.32 ml, 3.99 mmol) and 9-deazahypoxanthine (0.16 g, 1.19 mmol) were
added. The mixture was heated and stirred at 80 °C for 16 h, then
concentrated onto silica gel. Chromatography (CH2Cl2/7 M NH3–MeOH,
9:1 ? 85:15) gave 10 as a colorless solid, which was converted to the HCl
salt by evaporation from HCl (5% aq). Recrystallization (H2O–EtOH) gave the
HCl salt of 10 as a colorless, hygroscopic solid (0.103 g, 28.4%). Mp 224–225 °C.
1H NMR (300 MHz, D2O, referenced to internal acetone at 2.23 ppm) d 8.74 (s,
1H), 7.86 (s, 1H), 4.61 (d, J = 14.2 Hz, 1H), 4.55 (d, J = 14.2 Hz, 1H), 4.07 (dd,
J = 12.8, 3.6 Hz, 1H), 3.96 (dd, J = 12.8, 4.8 Hz, 1H), 3.58 (m, 1 H), 2.98–2.85 (m,
2H), 2.07 (s, 3H). 13C NMR (75.5 MHz, D2O, referenced to internal CH3CN at
1.47 ppm) d 153.7 (C), 144.9 (CH), 135.5 (C), 132.9 (CH), 118.7 (C), 104.4 (C),
58.9 (CH2), 58.0 (CH), 39.2 (CH2), 31.8 (CH2), 15.2 (CH3). ESI-MS C11H17N4O2S
[M+H]+ Calcd 269.1072, found 269.1069.
10. SerMe-ImmH (9) was prepared by reductive amination between serinol and 7-
benzyloxymethyl-6-methoxy-9-carbaldehyde-9-deazahypoxanthine
using
sodium cyanoborohydride, followed by protecting-group removal with conc.
HCl. Identity and purity were confirmed by NMR, ESI-MS, and elemental
analysis. Full experimental details will be published elsewhere: Clinch, K.;
Evans, G. B.; Furneaux, R. H.; Kelly, P. M.; Legentil, L.; Murkin, A. S.; Li, L.;
Schramm, V. L.; Tyler, P. C.; Woolhouse, A. D., in preparation.
11. Miles, R. W.; Tyler, P. C.; Furneaux, R. H.; Bagdassarian, C. K.; Schramm, V. L.
Biochemistry 1998, 37, 8615.
12. Lewandowicz, A.; Taylor Ringia, E. A.; Ting, L.-M.; Kim, K.; Tyler, P. C.; Evans, G.
B.; Zubkova, O. V.; Mee, S.; Painter, G. F.; Lenz, D. H.; Furneaux, R. H.; Schramm,
V. L. J. Biol. Chem. 2005, 280, 30320.
13. Lewandowicz, A.; Tyler, P. C.; Evans, G. B.; Furneaux, R. H.; Schramm, V. L. J.
Biol. Chem. 2003, 278, 31465.
14. Experimental
procedure.
( )-cis-1-tert-Butoxycarbonyl-4-fluoro-4-
methylthiomethylpyrrolidin-3-ol (13): Racemic pyrrolidinol 11 (Mason, J. M.;
Murkin, A. S.; Li, L.; Schramm, V. L.; Gainsford, G. J.; Skelton, B. W. J. Med. Chem.
in press) (0.52 g, 2.2 mmol) and dibutyltin oxide (0.66 g, 2.7 mmol) were
heated to reflux in toluene under a Dean–Stark trap for 1 h. After cooling to rt,
methanesulfonyl chloride (0.21 ml, 2.7 mmol) was added, and the resulting
solution was stirred for 8 h. The solution was then applied to a column of silica
gel. Elution with CH2Cl2/EtOAc (1:1) gave mesylate 12 (0.49 g, 1.78 mmol,