SAR analysis of adenosine diphosphate (hydroxymethyl)pyrrolidinediol inhibition of poly(ADP-ribose) glycohydrolase

Article abstract of DOI:10.1021/jm020541u

Polyadenosine diphosphoribose glycohydrolase (PARG) catalyzes the intracellular hydrolysis of adenosine diphosphoribose polymers. Because structure-activity data are lacking for PARG, the specific inhibitor adenosine diphosphate (hydroxymethyl)pyrrolidinediol (ADP-HPD) was utilized to determine the effects of structure on inhibitor potency using PARG isolated from bovine thymus (bPARG) and recombinant bovine PARG catalytic fragment (rPARG-CF). Both enzymes were strongly inhibited by submicromolar levels of ADP-HPD, but ADP and the phosphorylated pyrrolidine displayed no activity. Utilizing ADP-HPD analogues containing 2-, N⁶, or 8-adenosyl substituents or guanine instead of adenine, the importance of adenine ring recognition as well as a correlation between loss of PARG inhibition and the length and bulkiness of 8-adenosyl substituents was shown. Utilization of ADP-HPD analogues lacking one or both pyrrolidine cis-hydroxyls demonstrated their importance for inhibitor binding. Last, the similarity between naturally occurring bPARG and heterologously expressed rPARG-CF was demonstrated. Therefore, readily available rPARG-CF is suitable for use in future studies to determine the structural aspects of PARG.

Full text of DOI:10.1021/jm020541u

4322
J . Med. Chem. 2003, 46, 4322-4332
SAR An a lysis of Ad en osin e Dip h osp h a te (Hyd r oxym eth yl)p yr r olid in ed iol
In h ibition of P oly(ADP -r ibose) Glycoh yd r ola se
David W. Koh,† Donna L. Coyle,‡ Nimish Mehta,§ Sushma Ramsinghani,§ Hyuntae Kim,‡ J ames T. Slama,§ and
Myron K. J acobson*,‡
Division of Pharmaceutical Sciences, College of Pharmacy, University of Kentucky, Lexington, Kentucky 40536, Department of
Pharmacology & Toxicology, College of Pharmacy, and Arizona Cancer Center, University of Arizona, Tucson, Arizona 85724,
and Department of Medicinal and Biological Chemistry, College of Pharmacy, University of Toledo, Toledo, Ohio 43606
Received December 3, 2002
Polyadenosine diphosphoribose glycohydrolase (PARG) catalyzes the intracellular hydrolysis
of adenosine diphosphoribose polymers. Because structure-activity data are lacking for PARG,
the specific inhibitor adenosine diphosphate (hydroxymethyl)pyrrolidinediol (ADP-HPD) was
utilized to determine the effects of structure on inhibitor potency using PARG isolated from
bovine thymus (bPARG) and recombinant bovine PARG catalytic fragment (rPARG-CF). Both
enzymes were strongly inhibited by submicromolar levels of ADP-HPD, but ADP and the
phosphorylated pyrrolidine displayed no activity. Utilizing ADP-HPD analogues containing
2-, N6, or 8-adenosyl substituents or guanine instead of adenine, the importance of adenine
ring recognition as well as a correlation between loss of PARG inhibition and the length and
bulkiness of 8-adenosyl substituents was shown. Utilization of ADP-HPD analogues lacking
one or both pyrrolidine cis-hydroxyls demonstrated their importance for inhibitor binding. Last,
the similarity between naturally occurring bPARG and heterologously expressed rPARG-CF
was demonstrated. Therefore, readily available rPARG-CF is suitable for use in future studies
to determine the structural aspects of PARG.
In tr od u ction
tial to treat telomere-associated immortality,8 cardiac
or cerebral ischemia,9 diabetes,10 and Parkinson’s dis-
ease.11
The metabolism of adenosine diphosphoribose (ADP-
ribose) polymers is of importance to the maintenance
of genomic integrity of the multicellular organism and
in regulating the response to DNA damage. Polymers
are synthesized by the poly(ADP-ribose) polymerase
(PARP) family of enzymes, which utilize nicotinamide
adenine dinucleotide (NAD+) as substrate to covalently
modify acceptor proteins with ADP-ribose polymers.1-3
The typical characteristics of these polymers, which can
attain lengths of 200 ADP-ribose residues and contain
several branches,1 include a high negative charge and
the ability to mimic DNA for noncovalent interactions.4
PARP-1, the original and best studied PARP, is a DNA
damage sensing protein specifically activated by DNA
strand breaks.5 ADP-ribose polymer metabolism follow-
ing PARP-1 activation facilitates the recovery of dividing
cells from DNA damage.6 Accordingly, inhibition of the
metabolism of ADP-ribose polymers was recognized as
a strategy for treatment of malignancy.6 However, two
significant discoveries led to the expansion of this
therapeutic potential. First, PARP-1 inhibition was
shown to prevent massive cell necrosis following expo-
sure of cells to oxidative damage.7 Second, the discovery
of a diverse family of PARP isozymes suggests novel
functions for ADP-ribose polymer metabolism.2 Target-
ing these cycles offers the additional therapeutic poten-
Poly(ADP-ribose) glycohydrolase (PARG) catalyzes
the hydrolysis of the O-linked R(1′′f2′) or R(1′′′f2′′)
ribosyl-ribose linkages of ADP polymers to produce free
ADP-ribose.3,12 The efficient action of PARG results in
the rapid turnover of ADP-ribose polymers,13-15 thereby
reversing ADP ribosylation.
Bovine PARG (bPARG) has been shown to originate
from a single copy gene,16 and to date, no other PARG
homologues have been identified within the mammalian
genome. Therefore, preliminary evidence suggests that
PARG is required to complete the ADP-ribose polymer
cycles initiated by the entire PARP family of enzymes.
While therapies targeting the metabolism of ADP-ribose
polymers have primarily focused on PARP-1 inhibi-
tion,17,18 PARG is also predicted to be a therapeutic
target due to its essential function, its aforementioned
uniqueness, its presence at low levels, and the unique
structure of its substrate within the cell.
Our laboratory recently deduced the cDNA sequence
encoding full-length bPARG, which allowed the elucida-
tion of the domain organization of bPARG and enabled
the bacterial expression and production of the recom-
binant PARG catalytic fragment (rPARG-CF).16 In ad-
dition, prior studies identified adenosine diphosphate
(hydroxymethyl)pyrrolidine diol (1, ADP-HPD), a nitro-
gen in the ring analogue of ADP-ribose, as a potent and
specific inhibitor of PARG with a partial noncompetitive
mechanism of inhibition.19 Thus, the ability to specifi-
cally and effectively inhibit PARG has allowed ADP-
HPD to be utilized as a versatile tool to further study
* To whom correspondence should be addressed. Tel: (520)626-5957.
Fax: (520)626-8567. E-mail: mjacobson@pharmacy.arizona.edu.
† University of Kentucky.
‡ University of Arizona.
§ University of Toledo.
10.1021/jm020541u CCC: $25.00 © 2003 American Chemical Society
Published on Web 08/29/2003

SAR Analysis of Poly(ADP-ribose) Glycohydrolase
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4323
Sch em e 1. Synthesis of ADP-HPD Analogues
this enzyme. Because structure-activity relationships
(SARs) of PARG are poorly understood, we investigated
the SAR of naturally occurring bPARG and rPARG-CF
obtained from heterologous bacterial expression of the
cDNA encoding the bPARG catalytic fragment using
ADP-HPD as the lead inhibitor. This paper presents the
results of a detailed investigation into the molecular
interactions necessary for optimal binding to the PARG
active site. We first examined the ability of the major
partial structures of 1 to function as PARG inhibitors
by determining the potency of ADP and pyrrolidines 2
and 3 as PARG inhibitors. Next, the effect of deoxygen-
ation of the pyrrolidine ring on inhibitor potency is
determined using deoxy compounds 4 and 5. Last, the
effect of modification of the purine ring, particularly
with respect to 8-substitution, is evaluated using com-
pounds 6-13. Our results offer insight into the interac-
tion between ADP-HPD and its binding site. Further-
more, we show that the active sites of both enzymes are
similar as demonstrated by SAR analysis. Therefore, the
readily available rPARG-CF may be used in future
studies designed to determine the structural aspects of
PARG.
to the alcohol 15 by borohydride reduction. The phos-
phate is introduced by reaction of the protected alcohol
with N,N-diisopropyldibenzylphosphoramidite, followed
by oxidation producing phosphoester 16 in protected
form. The dibenzyl-protected phosphate ester 16 was
purified easily and in large quantity using flash chro-
matography on silica gel. It was quantitatively depro-
tected yielding 17 and converted to its tributylammo-
nium salt for coupling with nucleotides. The dideoxypyr-
rolidine derivative 19 was similarly available from
t-Boc-L-proline (Scheme 2) through the intermediacy of
the dibenzyl phosphate 18.
Compounds 8-12 required substituted AMP deriva-
tives that are not commercially available. The derivative
required for the production of both 8 and 9 was
synthesized from commercial 8-(6-aminohexyl)amino-
AMP by protection of the amine as a trifluoroacetyl
derivative. The synthesis of the 5′-nucleotides required
for synthesis of 10-12 began with the commercially
available nucleosides, 8-bromoadenosine and 6-chloro-
purine ribofuranoside (Scheme 3). The 6- or 8-substitu-
ent was introduced by nucleophilic displacement of the
halogen, and the phosphate was introduced by selective
phosphorylation of the 5′-hydroxyl using POCl3 in
trimethyl phosphate.24 The final products 10-12 were
obtained by coupling of the phosphate ester to the
protected pyrrolidine phosphate and subsequent depro-
tection as stated above and previously described in
Scheme 1.23,25
Resu lts
Ch em istr y. The syntheses of ADP-HPD (1), pyrro-
lidine (2), and pyrrolidine phosphate (3) were performed
as described in our previous papers,20,21 with some
modification as noted in the Experimental Section. The
synthesis of the novel ADP-HPD analogues 4-13 fol-
lowed the strategy depicted in Scheme 1. 5′-Mononu-
cleotides obtained commercially or through synthesis
are converted to their n-butylammonium salts, and the
phosphate is activated. Activation was accomplished by
treating the phosphate anion with carbonyldiimidazole,
converting it to the phosphoimidazolide,22 or by treating
it with diphenylchlorophosphate, converting it to the
mixed anhydride.23 The phosphate-activated 5′-mono-
nucleotide was condensed with the butylammonium salt
of a pyrrolidine phosphate in which the nitrogen was
protected with either a t-Boc or a Cbz group. The initial
coupling products in which the pyrrolidine nitrogen is
protected were purified using anion exchange chroma-
tography and completely characterized. Removal of
protection using either acid treatment or hydrogenolysis
resulted in the production of analogues 4-13.
SAR Stu d ies. The rPARG-CF and bPARG utilized
were similar in size and antigenicity to polyclonal anti-
bPARG antibodies,26 and each was catalytically active
(see below). Because of these similarities, each was used
to characterize the pattern of PARG catalytic activity
inhibition in order to interpret the molecular require-
ments needed to produce optimal interaction between
substrate and enzyme by SAR analysis.
Synthesis of the monodeoxygenated ADP-HPD de-
rivative 4 required the production of a new 4-monohy-
droxylated pyrrolidine, 17 (Scheme 2). This substance
was readily available from commercial 4-hydroxy-L-
proline using the sequence of reactions depicted in
Scheme 2. In this scheme, the proline is first completely
protected as 14, and the methyl ester 14 is converted
E lu cid a t ion of P AR G Act ive Sit e Molecu la r
In ter a ction s w ith Su bstr a te. Several PARG inhibi-

4324 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Sch em e 2. Synthesis of Modified Pyrrolidines
Koh et al.
Sch em e 3. Synthesis of Modified Adenosines
tors have been reported,27,28 but the most specific and
potent inhibitor reported to date is ADP-HPD (1). ADP-
HPD was a potent inhibitor of bPARG with an ap-
proximate IC50 of 0.33 µM (Figure 2A, open squares),
which is in agreement with the previously reported
value.20 Several compounds were then assayed to pro-
vide initial insight into the structural requirements
needed for optimal interaction of ADP-HPD with its
PARG binding site. The first series of compounds
utilized were the constitutive portions of ADP-HPD. The
inability of ADP to inhibit bPARG enzymatic activity,
even up to a concentration of 1 mM (Figure 2A, closed
squares), indicated the need for the pyrrolidine region
of ADP-HPD in order to optimally bind with PARG.
Likewise, the inability of HPD 2 (Figure 2A, open
circles) or P-HPD 3 (Figure 2A, closed circles), com-
pounds that each constitute the remaining portion of
ADP-HPD, to inhibit bPARG activity demonstrated the
requirement for the ADP portion as well. Thus, only the
sum of these parts displayed inhibition of bPARG
activity. The same potencies for these compounds were
observed when they were tested as inhibitors of rPARG-
CF (Figure 2B), where an IC50 ) 1.4 µM was measured
for the inhibition of rPARG-CF by 1. Because of the
inability of the constitutive portions of ADP-HPD to
independently inhibit the activity in either enzyme, it
was concluded that affinity for the PARG active site
requires major binding interactions with both the ADP
and the HPD regions of the ADP-HPD molecule.
Sign ifica n ce of th e P yr r olid in e cis-Hyd r oxyls in
th e In ter a ction of Su bstr a te w ith th e bP ARG a n d
r P ARG-CF Active Sites. The PARG active site was

SAR Analysis of Poly(ADP-ribose) Glycohydrolase
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4325
Ta ble 1. Activities of Test Compounds as Inhibitors of bPARG and rPARG-CF
bPARG
rPARG-CF
logIC50 ( error (µM)
compd
logIC50 ( error (µM)
IC50 (µM)
IC50 (µM)
1, ADP-HPD
4, ADP-HPM
5, ADP-HP
6, 8-N3-ADP-HPD
7, 2-N3-ADP-HPD
13, GDP-HPD
-0.48 ( 0.03
+0.48 ( 0.05
+1.28 ( 0.13
-0.41 ( 0.02
+2.45 ( 0.04
>3
0.33
3.07
19.2
0.39
290
+0.148 ( 0.07
+0.623 ( 0.08
+1.80 ( 0.18
-0.035 ( 0.02
+2.17 ( 0.07
+2.99 ( 0.06
1.4
4.2
63.
0.44
148
970
>1000
F igu r e 1. Structures of ADP-HPD and related compounds
tested as PARG inhibitors.
investigated further by utilizing compounds designed
to determine the importance of the pyrrolidine cis-
hydroxyls of ADP-HPD. ADP-HPM (4) (Table 1), identi-
cal to ADP-HPD but devoid of the pyrrolidine 3-hydroxyl
group, was a less potent inhibitor of bPARG catalytic
activity as ADP-HPD, since the IC50 of this compound
was approximately 3 vs 0.3 µM for ADP-HPD (1) (Table
1). A more profound effect on inhibition of bPARG
activity was seen by removing both pyrrolidine hy-
droxyls of ADP-HPD to produce 5, since the IC50 value
observed for this compound was 19 µM, 6-fold higher
than 4 and 60-fold higher than 1. A similar pattern of
inhibition was seen with rPARG-CF, where the IC50
value of 4 was 4 µM while the same value for 5 was
63 µM, 15-fold higher. Because the results demonstrated
significant differences in inhibition by comparing the
effect of 1 vs 4 and 5 on the enzymatic activity of bPARG
and rPARG-CF, it was concluded that the proximal cis-
hydroxyls of ADP-HPD are key structural requirements
for interactions with the PARG active site.
Sign ifica n ce of th e Ad en in e Rin g in th e In ter a c-
tion of Su bstr a te w ith th e bP ARG a n d r P ARG-CF
Active Sites. To investigate the contribution that the
adenine ring of ADP-HPD makes to inhibitor binding,
the next group of ADP-HPD-related compounds con-
tained variations in the purine ring. ADP-HPD (1) and
8-N3-ADP-HPD (6) were approximately equipotent as
inhibitors of bPARG (Table 1). Interestingly, 8-N3-ADP-
HPD (6) displayed even greater potency for inhibition
of rPARG-CF activity than did ADP-HPD. Both of these
observations suggested that limited modification of the
F igu r e 2. Elucidation of bPARG and rPARG-CF active site
molecular interactions. SAR analyses of bPARG (A) and
rPARG-CF (B) were performed by adding varying amounts of
the inhibitor to the enzyme assay conducted as described in
the Experimental Section. Inhibitor concentrations of 0.025,
0.05, 0.1, 1, 10, 100, and 1000 µM were used. The compounds
tested as inhibitors were ADP-HPD (open squares), ADP
(closed squares), HPD (open circles), and P-HPD (closed
circles).
8-adenosyl position would not decrease the inhibitory
potency of ADP-HPD. The IC50 measured for 2-N3-ADP-
HPD (7) was approximately 300 µM (Table 1) or 100-
fold higher than the IC50 value for ADP-HPD. Similar
results were seen using rPARG-CF, where an IC50 value
of 150 µM was determined for 2-N3-ADP-HPD (7).
Clearly, this substitution at the 2-position had a pro-
found negative effect on inhibition of bPARG and
rPARG-CF activity. The effect of replacing adenine with
guanine in ADP-HPD to produce GDP-HPD (13) re-
sulted in the least potent inhibitor in this group of
compounds. No significant inhibition of bPARG or
rPARG-CF catalytic activity was observed using GDP-
HPD concentrations up to 100 µM, while the highest
concentration utilized (1 mM) inhibited bPARG activity
by 10% and rPARG-CF activity by 50% (data not
shown). However, the inability of GDP-HPD to inhibit
catalytic activity greater than 50% at high concentra-
tions suggested that this compound did not efficiently
bind the bPARG or rPARG-CF active site with high
affinity. In summary, the results demonstrated the
requirement for an adenine moiety to facilitate optimal

4326 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Koh et al.
Ta ble 2. Activities of Purine-Substituted Compounds on
rPARG-CF
since the magnitude of bPARG inhibition by ADP-HPD
and analogues resulted in the following succession:
1 > 4 . 5. In the ADP portion of ADP-HPD, the bPARG
active site was shown to selectively recognize the
adenine moiety of ADP-HPD since GDP-HPD (13) did
not inhibit bPARG catalytic activity until the highest
concentration of compound was utilized. If the mecha-
nism of PARG hydrolysis of ADP-ribose polymers is
similar to that of other glycosyl hydrolases or even a
reverse of the PARP-1 catalytic mechanism, then the
possibility exists that the 3′′′- and 4′′′-ribosyl hydroxyls
are critical sites of hydrogen bond acceptors/donors
within the active site in order to properly hold and
orient the substrate for catalysis.29,30 In addition, the
demonstrated importance of the adenine ring may
suggest its involvement in initial substrate recognition
by PARG, with either a hydrophobic or a hydrogen-
bonding interaction with an active site residue(s). A
direct interaction between the adenine ring of ADP-HPD
and the Tyr796 of PARG is supported by the results of
our recent photoaffinity labeling study.31 However,
further high resolution studies are needed to substanti-
ate these assertions.
rPARG-CF
compd
logIC50 ( error (µM)
IC50 (µM)
1, ADP-HPDa
8
9
10
11
12
-0.00 ( 0.08
+0.97 ( 0.13
+2.75 ( 0.07
+2.07 ( 0.25
>3
1.
9.5
570
120
>1000
+2.71 ( 0.27
510
a
The IC50 for ADP-HPD was remeasured with the determina-
tion of the IC50 values for 8-12.
interaction with the PARG active site, while the sub-
stituents in the 8-adenosyl position did not adversely
affect this interaction.
Effect of 8-Su bstitu en ts on th e In h ibition of
r P ARG-CF Ca ta lytic Activity by ADP -HP D. The
next set of compounds was designed to determine the
tolerance of PARG to modifications at the 8-position of
the adenine ring. Three additional 8-substituted ADP-
HPD analogues, 8-10, were assayed for their ability to
inhibit rPARG-CF enzymatic activity, since bPARG and
rPARG-CF were demonstrated to be similar in all
previous SAR analyses. All were less potent inhibitors
of rPARG-CF activity than was either ADP-HPD (1) or
8-N3-ADP-HPD (6), but the most profound reduction
was observed with 9 and 10 (Table 2). The IC50 value
of ADP-HPD (1) was 1 µM, that for 8 was 10 µM, the
IC50 value for 10 approached 120 µM, and that for 9
was almost 600-fold higher at 570 µM. In summary,
although 8-N3-ADP-HPD (6) did not significantly affect
the inhibition of rPARG-CF activity by parent com-
pound, 8-10 decreased the inhibition to varying de-
grees.
Effect of N6-Su bstitu en ts on th e In h ibition of
r P ARG-CF Ca ta lytic Activity by ADP -HP D. The
last set of compounds was designed to investigate yet
another site on the adenine ring of ADP-HPD to
determine their effects on inhibition of rPARG-CF. Two
lipophilic substituents were coupled to the N6-position
of adenine to produce 11 and 12. Both significantly
reduced the inhibition of rPARG-CF activity by ADP-
HPD since a dose of 100 µM, the highest dose utilized
for each compound, only produced 33-34% inhibition
of enzymatic activity (Table 2). The results provide
additional evidence of the requirement for optimal
interaction of adenine with the PARG active site, with
substituents at the N6-position significantly affecting
this interaction.
Interesting effects were observed when substitutions
were introduced into the adenine ring. Both 11 and 12
were shown to greatly affect the inhibition of PARG
catalytic activity by ADP-HPD. However, differing ef-
fects were seen when 2- and 8-adenosyl substituents
were utilized. The difference of bPARG inhibition
observed between 2-N3-ADP-HPD (7) and 8-N3-ADP-
HPD (6) possibly reflects a conformational requirement
in order to efficiently bind the active site, since purine
nucleotides containing substituents in the 8-position
favor the syn configuration32 while those in the 2-posi-
tion favor the anti.33 However, lack of potent inhibition
exhibited by 2-N3-ADP-HPD may also reflect a decrease
in active site access due to the steric effects elicited after
cyclization of the azido group with the adenosyl-N3
nitrogen at physiologic pH to produce a tetrazolo
isomer.33 Nevertheless, it was concluded that substit-
uents in the N6- and 2-adenosyl positions of ADP-HPD
greatly decrease access to the PARG active site, prob-
ably by obstructing adenine ring binding.
A further investigation of substituents at the 8-
adenosyl position of ADP-HPD (1) provided useful
information. Loss of inhibitor potency for both rPARG-
CF and bPARG is associated with the introduction of
large substituents at the 8-adenosyl as in compounds
8-10. Because 8 contains a functionalized spacer arm
8-atoms in length vs 3-atoms in 6, loss of potency seems
to be associated with introduction of large substituents.
Because a greater loss of potency was seen with 10, it
seems that large, bulky substituents in the 8-adenosyl
position produce an even larger decrease in inhibition
potency. However, because the greatest effect on rPARG-
CF inhibition was demonstrated by 9, it was concluded
that the combination of a long and bulky substituent
at the 8-adenosyl position of ADP-HPD would most
significantly decrease the inhibition of PARG catalytic
activity by parent compound. Compound 10 was syn-
thesized for the explicit purpose of utilization as a cell-
permeable inhibitor of PARG as previously reported for
other nucleotides with this modification.34-36 However,
whenever a spacious aliphatic group was directly at-
Discu ssion
To date, there have been no detailed papers on
structure-activity characterizations of PARG. In this
paper, SAR studies utilized ADP-HPD (1) as the lead
inhibitor, not only because of its high potency but also
because of its demonstrated high specificity.20 The
results of the SAR analysis provided significant insight
into the substrate structural requirements for optimal
affinity for the PARG active site. Both the ADP and the
HPD portions were required for high affinity binding
to the bPARG active site since ADP, HPD (2), or P-HPD
(3) alone did not significantly inhibit bPARG activity.
More specifically, the cis-diols of the pyrolidine moiety
were shown to be important for high potency binding,

SAR Analysis of Poly(ADP-ribose) Glycohydrolase
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4327
tached to the adenine ring (10) or linked via an 8-atom
spacer arm (9), profound reductions were seen in the
inhibition of rPARG-CF activity as compared to the
parent compound ADP-HPD. Although the exact mech-
anism of this effect on ADP-HPD inhibition of rPARG-
CF was not known, it is possible that the steric
hindrance created by these large aliphatic groups
obstructed binding of the compound to the active site,
thereby interfering with its ability to inhibit PARG
activity. On the other hand, the ability of 8-N3-ADP-
HPD (6) to inhibit PARG activity comparable to ADP-
HPD suggested that the azido group did not create any
hindrance for the compound to optimally bind the active
site. Therefore, it was concluded that limited modifica-
tions of the 8-adenosyl position of ADP-HPD should not
extensively affect the high affinity binding of inhibitor
to the PARG active site. Although the IC50 of 8 is lower
than that of the parent compound 1, it is still sufficiently
potent at 10 µM to enable 8 to be successfully used as
a ligand for developing biospecific absorbents for PARG.
substrate using SAR analysis. Furthermore, we dem-
onstrate the similarity of naturally isolated and recom-
binant enzymes. The information presented here should
facilitate both structural and biological studies of PARG
in the future in order to elucidate the structural biology
of PARG, establish its molecular mechanism of cata-
lyzing ADP-ribose polymer hydrolysis, facilitate the
rational structure-based inhibitor design for this poten-
tially novel therapeutic target, and determine the exact
biological role of PARG within the cell.
Exp er im en ta l Section
Ma ter ia ls. Recombinant PARP-1 was prepared as previ-
ously described,38 as was rPARG-CF16 and bovine thymus
PARG,39 except that steps nos. 6 (DNA-agarose) and 7 (Hep-
arin-Sepharose) were omitted and a Red-Sepharose step was
added.40 [R-32P]NAD+ at high specific activity was purchased
from ICN. ADP was purchased from Sigma. ADP-HPD (1),
2-N3-ADP-HPD (7), and 8-N3-ADP-HPD (6) were synthesized
as previously described.20,25 Compounds 8 and 9 were produced
starting from 8-(6-aminohexyl)amino-AMP (Sigma) by convert-
ing it to 8-(6-trifluoroacetylaminohexyl)amino-AMP using S-
ethyltrifluorothioacetate treatment and coupling the resulting
AMP derivative to t-Boc-protected pyrrolidine (3). A detailed
synthetic procedure will be published separately.
Gen er a l Meth od s. 1H NMR spectra were determined at
300 mHz at ambient probe temperature, at a concentration of
ca. 10 mg/mL. 13C NMR spectra were determined at 75 mHz.
31P NMR spectra were acquired on a Varian Gemini-200
spectrometer at 80.95 mHz using 5 mm tubes. Chemical shifts
are referenced to an external standard of 85% aqueous H3-
PO4. Mass spectral analysis (electron impact, fast atom
bombardment (FAB), or MALDI) was performed at the Uni-
versity of Kentucky Mass Spectrometry Facility, Lexington,
KY.
High-performance liquid chromatography (HPLC) was per-
formed using a high-pressure gradient system and a detector
operating at 260 nm. Reversed-phase separations were per-
formed using a 3.9 mm × 300 mm reversed-phase column (C18
Bondapak,15-20 µm, 125 Å; Waters, Milford, MA) using an
ion-pairing technique at a constant flow rate of 1.5 mL/min.
The solvents were (A) 20 mM NaH2PO4‚H2O + 2 mM Bu4NH2-
PO4, pH 6, and (B) 50% (v/v) solvent A and acetonitrile. A
mobile phase of 80% A and 20% B was changed linearly to
50% A over 5 min and thereafter maintained isocratically for
25 min. Under these conditions, AMP eluted at 2.7 min and
GMP eluted at 3.8 min.
(2R,3R,4S)-1-(Ben zyloxyca r bon yl)-2-(h yd r oxym eth yl)-
p yr r olid in e-3,4-d iol-3,4-O-isop r op ylid in e Aceta l. The pre-
cursor for 2 and 3 was synthesized as described by Goli et al.21
(2R,3R,4S)-2-Hyd r oxym et h ylp yr r olid in e-3,4-d iol h y-
d r och lor id e (HP D, 2). 1-Benzyloxycarbonyl-2-(hydroxym-
ethyl)pyrrolidine-3,4-diol 3,4-O-isopropylidene acetal (250 mg,
0.81 mmol) was dissolved in 4 mL of water at 35 °C, 1 mL of
trifluoroacetic acid (TFA) was added, and the reaction was
stirred at 35 °C for 3 h. Thin-layer chromatography (TLC)
(silica gel, ethyl acetate/methanol 99:1) verified the complete
conversion of the starting material to a lower Rf product. The
solvent was evaporated in vacuo and chased by 5 mL of
toluene. The product was purified by flash column chroma-
tography (silica gel 35-70 µm, 100 g, ethyl acetate/methanol
98:2, 225 mL followed by ethyl acetate/methanol 95:5, 200 mL)
to obtain a yellow-colored oil. The benzyloxycarbonyl group was
removed by dissolving the oil in 15 mL of methanol, adding
50 mg of 5% palladium on carbon, and hydrogenating at
30 psi for 4 h. TLC (silica gel, ethyl acetate/methanol 95:5)
verified complete consumption of the starting material. The
catalyst was removed by filtration over Celite filter aid.
Evaporation of solvent in vacuo yielded a black residue. The
product was stirred in 2 mL of 1 M HCl for 30 min and
lyophilized to obtain a gray-colored, flaky solid (80 mg, 58%).
TLC (silica gel, ethyl acetate/methanol 95:5): Rf ) 0.07. 1H
Structural studies of PARG are lacking due to the
limited availability of the enzyme isolated from tissue.
Furthermore, the rPARG-CF obtained from heterolo-
gous bacterial expression in other published papers has
been utilized only qualitatively in order to verify that
the protein translated from the cloned PARG cDNA
from various sources exhibited PARG catalytic activ-
ity.16,37 Because it is now clear that PARG isolated from
tissue is derived from proteolysis of a 111 kDa precursor,
rPARG-CF may further differ both structurally and in
its enzymatic activity from the enzyme isolated from
natural sources. The present study not only provides a
detailed analysis of substrate structural requirements
needed in order to properly bind PARG, but it also
demonstrated the similarity between the conventionally
purified and recombinant enzyme. Because similar
patterns of inhibition, or lack thereof, utilizing ADP-
HPD and related compounds were observed for bPARG
and rPARG-CF (Figure 1, Table 1), it was concluded
that the inhibitor binding sites were similar as demon-
strated by SAR analysis. Close similarity between
bPARG and rPARG-CF is further supported by similar-
ity of kcat and Km values for the proteins. For bPARG,
kcat ) 31 µmol/min/mg and Km ) 0.38 µM (in ADP-ribose
residues), while for rPARG-CF kcat ) 10 µM/min/mg and
Km ) 0.52 µM. rPARG-CF is readily available, so it will
be utilized in future PARG structural studies, such as
the mapping of its active site, the identification of
essential catalytic residues within its active site, and
the determination of the three-dimensional structure of
its active site by X-ray crystallographic analysis.
Biological studies regarding PARG are lacking due
in part to the absence of cell permeable specific PARG
inhibitors. ADP-HPD is highly polar and therefore is
not expected to readily traverse the plasma membrane.
Therefore, its use in determining the in vivo effects of
intracellular PARG inhibition is not warranted. The
information presented in this study, which suggests that
an appropriate hydrophobic 8-substituent might be
tolerated, should prove useful when designing potential
cell permeable PARG inhibitors using ADP-HPD, the
most specific inhibitor of PARG known, as a template.
In summary, we present the first insight into the
molecular interactions of the PARG active site with

4328 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Koh et al.
NMR (D2O, referenced to HDO at δ 4.67 ppm): δ 3.20-3.25
(dd, 1 H, CH2N), 3.34-3.38 (dd, 1 H, CH2N), 3.34-3.35 (m,
1 H, CHN), 3.65-3.71 (dd, 1 H, CH2OH), 3.80-3.86 (dd, 1 H,
CH2OH), 4.04-4.08 (m, 1 H, CHOH), 4.22-4.26 (m, 1 H,
CHOH). 13C NMR (D2O, referenced to CD3OD as external
reference δ 49.0 ppm): δ 50.56 (CH2N), 58.94 (CH2OH), 62.74
(CHN), 70.37 (CHOH), 72.12 (CHOH). FAB mass spectrum
(positive ion): calcd for (M+) (C5H12O3N), m/z 134; found, m/z
134.
(2R,3R,4S)-1-(Ben zyloxyca r b on yl)-2-[(p h osp h ooxy)-
m eth yl]pyr r olidin e-3,4-diol (Cbz-3). Freshly distilled POCl3
(0.71 mL, 7.6 mmol) was cooled in an ice bath, and water (68
µL, 3.8 mmol) followed by pyridine (0.6 mL, 7.6 mmol) was
added with stirring. The reaction mixture solidified but
liquified upon the addition of acetonitrile (0.85 mL). The
starting alcohol, (2S,3R,4S)-1-(benzyloxycarbonyl)-2-(hydroxy-
methyl)pyrrolidine-3,4-diol-3,4-O-isopropylidine acetal (0.59 g,
1.9 mmol), was added as a solution in 0.85 mL of acetonitrile.
The mixture was stirred at 4 °C for 2 h, at which time several
small pieces of ice were added to destroy the excess POCl3.
Four milliliters of 1 M HCl was added, and the reaction was
stirred at ambient temperature for 45 min to remove the
acetonide protecting group. The product was desalted by
chromatography on a column of Amberchrome CG 71ms resin
(1.5 cm × 45 cm), a reversed-phase type adsorbent (Toso Haas
Inc.). The chromatography was developed with water. A large
peak of pyridine preceded two smaller peaks. The test for
inorganic phosphate (Ames, 1966) showed that most of the
inorganic phosphate had coeluted with pyridine. The latter two
UV absorbing peaks were pooled, and the pH was adjusted to
7.5.
aqueous NaHCO3 solution (3 × 20 mL) and mixed with the
aqueous layer. The combined aqueous layer was acidified to
approximately pH 1.5 with 10% H2SO4 solution while main-
taining the temperature to 0-5 °C (ice bath) and extracted
with ether (4 × 50 mL). The combined ether extracts were
washed with water (2 × 30 mL) and brine (1 × 50 mL) and
dried (Na2SO4). Evaporation of the solvent gave a sticky,
colorless oil (11 g, 62%). 1H NMR (CDCl3): δ 1.45 (d, J ) 24.8
Hz, 9H), 2.2 (m, 1H), 2.4 (m, 1H), 3.52 (d, J ) 3.2 Hz, 1H),
3.58 (m, 1H), 4.41 (m, 1H), 4.47 (br, 2H). 13CNMR (CDCl3): δ
28.27, 38.35, 52.17, 54.68, 69.46, 80.21, 152.84, 172.19.
N-ter t-Bu t yloxyca r b on yl-tr a n s-4-h yd r oxy-L-p r olin e
Meth yl Ester . To a solution of N-tert-butyloxycarbonyl-trans-
4-hydroxy-L-proline (11.00 g, 47.6 mmol) in ether containing
5% CH3OH (200 mL) was added a solution of diazomethane
(CH2N2, prepared from 21.5 g Diazald according to the
manufacturer’s directions) slowly with swirling, until a yellow
color persisted. The yellow color disappeared with a few drops
of glacial acetic acid (1 mL), and the ethereal solution was
extracted with saturated NaHCO3 solution (2 × 50 mL). It was
washed with water (1 × 100 mL) and saturated brine and dried
(Na2SO4). Evaporation of the solvent gave a colorless oil (11.00
g, 95%); Rf ) 0.07 (silica gel, 7:3 hexane:ethyl acetate); Rf )
0.28 (silica gel, 1:1 hexane:ethyl acetate). 1H NMR (CDCl3): δ
1.42 (d, J ) 19.6 Hz, 9H), 1.76 (s, br, 1H), 2.06 (m, 1H), 2.28
(m, 1H), 3.50 (dd, J ) 42 Hz, 11.6 Hz, 1H), 3.62 (dd, J ) 11.6
Hz, 3.6 Hz, 1H), 3.72 (d, J ) 4.4 Hz, 3H), 4.4 (s, 1H). 13C NMR
(CDCl3): δ 28.24, (38.48, 39.14), (52.02, 52.21), 54.71, (57.45,
57.86), (69.54, 70.26), 80.33, 153.92, 173.59. MS (MALDI): m/e
268 (MNa+). Anal. (C11H19NO5) C, H, N.
(2S,4R)-N-ter t-Bu tyloxyca r bon yl-4-[(ter t-bu tyld ip h en -
ylsilyl)oxy]p yr r olid in e-2-ca r boxylic Acid , Meth yl Ester
(14). A solution of N-tert-butyloxycarbonyl-trans-4-hydroxy-
L-proline methyl ester (3.50 g, 14.28 mmol) in anhydrous
dimethyl formamide (DMF; 10 mL), which had been flushed
with N2, was treated with tert-butylchlorodiphenylsilane (5.88
g, 21.4 mmol) with stirring followed by imidazole (2.43 g, 35.7
mmol), and the mixture was allowed to stir at room temper-
ature overnight under N2. TLC indicated the presence of both
of the starting materials; hence, DMAP (12 mg) was added,
and the mixture was stirred for another 4 h. A complete
disappearance of the silane was observed by TLC (although
some alcohol was still present); hence, the reaction was stopped
by partitioning between KHSO4 (5%, 100 mL) and CHCl3 (100
mL). The layers were separated, the aqueous layer was
extracted with CHCl3 (2 × 100 mL), and the combined organic
phase was washed with water (100 mL) and saturated brine
(100 mL) and dried (Na2SO4). The solvent was then evaporated
to give the crude silyl ether, which upon flash column chro-
matography on silica gel using hexane:ethyl acetate (90:10)
as the elutant gave 3.07 g of the pure product (45%) as a sticky
oil; Rf ) 0.52 (silica gel, 7:3 hexane:ethyl acetate); Rf ) 0.78
(silica gel, 1:1 hexane:ethyl acetate). 1H NMR (CDCl3): δ 1.05
(s, 9H), 1.44 (d, J ) 10.4 Hz, 9H), 1.85 (m, 1H), 2.2 (m, 1H),
3.45 (m, 2H), 3.68 (d, J ) 9.6 Hz, 3H), 4.42 (m, 2H), 7.40 (m,
6H), 7.63 (m, 4H). 13C NMR (CDCl3): δ 26.67, 28.10, (38.10,
39.52), (51.43, 51.90), (54.52, 55.00), (57.62, 58.10), (70.48,
71.19), 79.28, 127.62, 130.00, (133.33, 133.81), 135.71, (152.86,
The monophosphate was purified by anion exchange chro-
matography on a 1.5 cm × 43 cm column of DE-52 cellulose
(Whatman), developed by the application of a linear gradient
formed between 400 mL of 0.01 M NH4HCO3 (pH 7.5) and 400
mL of 0.2 M NH4HCO3 (pH 7.5). Fractions (7.5 mL) were
collected, and absorbance at 254 nm was measured. A single
major peak eluting approximately midway through the gradi-
ent was pooled and lyophilized. After several lyophilizations,
the monophosphate was obtained as a white amorphous solid
(0.51 g, 77% yield). The structure was confirmed by comparison
to authentic material20 using TLC, HPLC, and NMR.
(2R,3R,4S)-2-[(P h osp h ooxy)m eth yl]p yr r olid in e-3,4-d i-
ol (3). Cbz-3 (0.27 g, 0.77 mmol) was dissolved in 10 mL of
water and 30 mL of methanol and subjected to catalytic
hydrogenation (50 psi, 5% Pd on carbon, 340 mg) for 4 h. TLC
verified that the starting material had been completely
consumed at this time. The catalyst was removed by filtering
through a Celite filter aid. The solvent was evaporated under
reduced pressure to obtain an oil in quantitative yield that
was dissolved in water and lyophilized. TLC (silica gel,
2-propanol/concd NH4OH/water, 6:3:1): Rf ) 0.13. 1H NMR
(D2O): δ 3.33-3.38 (d, 1H, CH2N), 3.45-3.50 (dd, 1H, CH2N),
3.68-3.73 (m, 1H, CHN), 3.97-4.06 (m, 1H, CH2OP), 4.19-
4.27 (m, 1H, CH2OP), 4.37-4.41 (m, 2H, 2CHOH). 13C NMR
(D2O): δ 52.20 (C-5), 62.95 and 63.01 (C-6), 64.02 and 64.08
(C-2), 72.59 and 73.86 (C-3 and C-4). FAB mass spectrum
(positive ion): calcd for (M + H)+ (C5H13O6NP), m/z 214; found,
m/z 214.
153.81), 172.62. MS (MALDI): m/e 506 (MNa+). Anal. (C27H37
-
NO5 Si) C, H, N.
Syn th esis of Ad en osin e Dip h osp h a te (Hyd r oxym eth -
yl)p yr r olid in e Mon oa lcoh ol (ADP -HP M, 4). N-ter t-Bu tyl-
oxyca r bon yl-tr a n s-4-h yd r oxy-L-p r olin e. A three-necked
(2S,4R)-N-ter t-Bu tyloxyca r bon yl-4-[(ter t-bu tyld ip h en -
ylsilyl)oxy]-2-h yd r oxym eth yl P yr r olid in e (15). To a solu-
tion of N-tert-butyloxycarbonyl-4-[(tert-butyl diphenyl silyl)-
oxy]pyrrolidine-2-carboxylic acid (methyl ester, 3.00 g, 6.21
mmol) in anhydrous THF (30 mL), LiBH4 (0.27 g, 12.4 mmol)
was added with stirring under N2. The mixture was allowed
to stir at room temperature for 2 h, at which point a complete
disappearance of the starting material was verified by TLC
(hexane:ethyl acetate 80:20). The excess LiBH4 was then
inactivated with glacial acetic acid, and the mixture was
partitioned between NaHCO3 (5%, 100 mL) and CHCl3 (100
mL). The layers were separated, the aqueous layer was
extracted with CHCl3 (2 × 100 mL), and the combined organic
phase was washed with water (100 mL) and saturated brine
round-bottom flask (500 mL), equipped with
a magnetic
stirring bar, a dropping funnel, and a reflux condenser, was
charged with a solution of NaOH (3.4 g, 84 mmol) in water
(85 mL). Stirring was then initiated, and trans-4-hydroxy-L-
proline (10 g, 76.2 mmol) was added at ambient temperature
and diluted with tert-butyl alcohol (50 mL). A solution of di-
tert-butyl dicarbonate (16.6 g, 76.2 mmol) was added to the
stirred solution of the amino acid (pH 12-12.5), and the
mixture was allowed to stir at room temperature overnight.
The pH of the solution at this point was 8.5. The reaction
mixture was then extracted with pentane (2 × 50 mL), and
the combined organic phase was back-extracted with saturated

SAR Analysis of Poly(ADP-ribose) Glycohydrolase
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4329
(100 mL) and dried (Na2SO4). Stripping of the solvent gave
pure product as a colorless oil (2.86 g, 100%); Rf ) 0.22 (silica
gel, 8:2 hexanes:ethyl acetate); Rf ) 0.57 (silica gel, 1:1
N-tert-Butyloxycarbonyl-4-hydroxy-2-(dibenzylphosphoxy-
methyl) pyrrolidine (300 mg, 0.62 mmol) was dissolved in
methanol (10 mL) with stirring, and 5% Pd-C (40 mg) was
added. The reaction flask was purged with H2 and allowed to
stir overnight at room temperature under H2. It was filtered
through Celite filter aid, and the solvent was evaporated to
give a white solid. It was suspended in methanol (2 mL), and
tributylamine (116.6 mg, 150 µL, 0.62 mmol) was added. The
mixture was stirred for 5 min, and the methanol was evapo-
rated in vacuo and chased once with DMF (2 mL). This solid
was suspended in DMF and added to the AMP-imidazolide
solution obtained as above. The mixture was allowed to stir
for 5 days at room temperature until the completion of the
reaction was verified by TLC.
1
hexanes:ethyl acetate). H NMR (CDCl3): δ 1.05 (s, 9H), 1.45
(s, 9H), 2.02 (m, 1H), 3.15 (dd, J ) 12 Hz, 1.7 Hz, 1H), 3.50
(dd, J ) 12 Hz, 1.7 Hz, 2H), 3.71 (m, 2H), 4.29 (s, br, 2H),
4.95 (d, J ) 9 Hz, 1H), 7.42 (m, 6H), 7.65 (m, 4H). 13C NMR
(CDCl3): δ 19.06, 26.79, 28.41, 37.68, 55.83, 59.19, 67.34, 70.84,
80.41, 127.75, 129.84, 133.60, 135.59, 135.64, 157.51. MS
(MALDI): m/e 478 (MNa+). Anal. (C26H37NO4 Si) C, H, N.
(2S,4R)-N-ter t-Bu tyloxyca r bon yl-4-[(ter t-bu tyld ip h en -
ylsilyl)oxy]-2-(d iben zylp h osp h oxym eth yl) P yr r olid in e
(16). 1-H-Tetrazole (0.65 g, 9.32 mmol) was added to a solution
of N-tert-butyloxycarbonyl-4-[(tert-butyl diphenyl silyl)oxy]-2-
hydroxymethyl pyrrolidine (2.82 g, 6.21 mmol) in CH2Cl2 (80
mL), and the suspension was stirred under N2 for 5 min. N,N-
Diisopropyldibenzylphosphoramidite (3.22 g, 9.32 mmol) was
added, and the reaction mixture was stirred at room temper-
ature for 3 h under N2. The reaction mixture was cooled to
-40 °C (in dry ice-acetonitrile bath), and m-chloroperoxy-
benzoic acid (mCPBA) (2.14 g, 12.4 mmol) in CH2Cl2 (20 mL)
was added. The bath was replaced by an ice bath, and the
mixture was stirred at 0 °C for 1 h, diluted with CH2Cl2 (100
mL), washed with NaSO3 (10%, 2 × 50 mL), NaHCO3 (satu-
rated, 2 × 50 mL), water (100 mL), and saturated brine (100
mL), and dried (MgSO4). The solvent was evaporated to give
crude product, which on flash column chromatography on silica
gel using hexane:ethyl acetate (70:30) as the elutant, gave 4.4
g of pure product (100%) as a colorless oil; Rf ) 0.22 (silica
gel, 7:3 hexane:ethyl acetate); Rf ) 0.49 (silica gel, 1:1 hexane:
ethyl acetate). 1HNMR (CDCl3): δ 1.05 (s, 9H), 1.45 (s, 9H),
1.80 (m, 1H), 2.03 (m, 2H), 3.09 (m, 1H), 3.50 (m, 1H), 3.95
(m, 1H), 4.30 (m, 1H), 4.93 (m, 4H), 5.00 (m, 1H), 7.29 (m,
16H), 7.62 (d, J ) 6 Hz, 4H). 13CNMR (CDCl3): δ 19.05, 26.79,
28.40, (36.80, 37.74), (54.85, 55.42), (55.54, 55.65), (67.72,
68.60), 69.22, (70.67, 71.33), (79.57, 79.91), 127.61, 127.73,
127.92, 128.53, 129.81, 133.74, 135.43, 135.57, 135.63, 155.20.
Anal. (C40H50NO7PSi) C, H, N.
(2S,4R)-N-ter t-Bu tyloxyca r bon yl-4-h yd r oxy-2-(d iben -
zylp h osp h oxym eth yl) P yr r olid in e (17). N-tert-Butyloxy-
carbonyl-4-[(tert-butyl diphenyl silyl)oxy]-2-(dibenzylphos-
phoxymethyl) pyrrolidine (4.2 g, 5.87 mmol) was dissolved in
THF (40 mL) in a 500 mL round-bottomed flask, and a 1 M
solution of tetra-n-butylammonium fluoride in THF (23.5 mL,
23.5 mmol) was added with stirring. The mixture was allowed
to stir for 2 h at room temperature, and the solvent was
evaporated in vacuo to give crude product appearing as a
colorless oil. Flash column chromatography on silica gel (using
hexane:ethyl acetate 40:60 as the elutant) gave a colorless,
clear oil (2.60 g, 92.8%); Rf ) 0.24 (silica gel, 2:8 hexanes:ethyl
acetate); Rf ) 0.35 (silica gel, ethyl acetate). 1H NMR
(CDCl3): δ 1.45 (s, 9H), 2.0 (m, 2H), 2.48 (br, 1H, D2O
exchangeable), 3.40 (m, 2H), 4.18 (m, 4H), 5.03 (d, J ) 9.9 Hz,
4H), 7.35 (s, 10H). 13CNMR (CDCl3): δ 28.36, (36.54, 37.12),
54.94, (55.32, 55.44), (67.67, 68.05), 69.30, 80.10, 127.91,
127.95, 128.56, 135.70, 154.55. 31P NMR (CDCl3, external
reference, H3PO4): δ -0.10, -0.32. FAB mass spectrum
(positive-ion): calcd m/z 477; found, m/z 478 (MH+). Anal.
(C24H32NO7P) C, H, N.
The reaction was quenched by adding methanol (25 mL),
and the solvents were evaporated in vacuo. The mixture was
diluted with water (100 mL), the pH was adjusted to 7.5 with
1 M NH4OH, and the sample was applied to a benzyl DEAE
cellulose column (40 mL; 1.2 cm × 35 cm) previously equili-
brated with 10 mM NH4HCO3. The column was developed with
a linear gradient formed between 400 mL of 10 mM NH4HCO3
and 400 mL of 400 mM NH4HCO3, pH 7.5. Fractions (10 mL)
were collected, and the product was eluted in fractions 61-
79. This was found to be a mixture of three components as
detected by analytical reversed-phase HPLC (C18 Bondapak).
The combined fractions were evaporated in vacuo at 35 °C,
and the white solid obtained was dissolved in a minimum
volume of water and acidified to pH 1.7 with dilute HCl. The
mixture was applied to a reversed-phase column (Amber-
chrome CG-71 md, 1.2 cm × 47 cm) and eluted with deionized
water. Fractions containing the product were pooled and
analyzed by reversed-phase HPLC, which showed the presence
of a small amount of AMP. Hence, the solution was concen-
trated and reapplied to the Amberchrome column. The frac-
tions containing the pure product were pooled, adjusted to pH
7.2 with 1 M NH4OH, and lyophilized to give the pure product
(50 mg, 40%). HPLC (reversed-phase C18 Bondapak, ion pair),
retention time ) 12.5 min, retention time of AMP ) 2.7 min;
Rf ) 0.52 (silica gel, 6:3:1 2-propanol:concentrated ammonia:
water). 1H NMR (D2O): δ 1.25 (s, 9H, tert-butyl), 1.86 (m, 1H,
pyrrolidine H-3′), 2.00 (m, 1H, pyrrolidine H-3′), 3.21 (m, 2H,
pyrrolidine H-1′), 3.83 (m, 3H, pyrrolidine H-4′ and H-5′), 4.05
(s, 2H, adenosine H-5′), 4.22 (s, 2H, adenosine H-4′ and
pyrrolidine H-2′), 4.35 (m, 1H, adenosine H-3′), 4.61 (m, 1H,
adenosine H-2′), 5.97 (d, J ) 5.6 Hz, 1H, adenosine H-1′), 8.10
(s, 1H, adenine H-2), 8.37 (s, 1H, adenine H-8). 13CNMR
(D2O): δ 28.60 (tert-butyl), 37.11 (pyrrolidine C-3′), 55.02
(pyrrolidine C-1′), 56.60 (pyrrolidine C-4′), 66.05 (pyrrolidine
(C-5′), 67.42 (adenosine C-5′), 69.84 (pyrrolidine C-2′), 71.30
(adenosine C-3′), 75.27 (adenosine C-2′), 82.57 (tert-butyl 4°
carbon), 84.97 (adenosine C-4′), 88.07 (adenosine C-1′), 119.57
(adenine C-5), 140.38 (adenine C-8), 148.68 (adenine C-4),
155.54 (adenine C-2), 157.30 (adenine C-6). 31P NMR (D2O,
external reference, H3PO4): δ -10.40 (multiplet). FAB mass
spectrum (positive-ion): calcd m/z 626; found, m/z 627 (MH+).
Ad en osin e Dip h osp h a te (Hyd r oxym eth yl)p yr r olid in e
Mon oa lcoh ol (ADP -HP M, 4). t-Boc-ADP-HPM (40 mg, 0.064
mmol) was mixed with TFA (5 mL) and stirred at room
temperature for 30 min. TFA was then evaporated in vacuo,
the residue was dissolved in water (100 mL), and the pH was
adjusted to basic (9.7) with 1 M NH4OH. The sample was then
applied as a dilute solution to an 83 mL column of benzyl
DEAE cellulose (1.5 cm × 47 cm), previously equilibrated with
10 mM NH4HCO3 and eluted with a linear gradient formed
between 10 mM NH4HCO3 and 300 mM NH4HCO3 (400 mL
each) at a flow rate of 1 mL/min. Fractions of 10-11 mL each
were collected. The chromatography produced two well-
separated peaks. Fractions 52-66 were pooled together and
lyophilized twice to give pure, salt free product (30 mg, 90%).
HPLC (reversed phase C18 Bondapak, ion pair), retention
time ) 5.5 min, retention time of AMP ) 2.7 min. 1H NMR
(D2O): δ 1.91 (m, 2H, pyrrolidine H-3′), 3.14 (d, J ) 12.51 Hz,
1H, pyrrolidine H-1′), 3.25 (dd, J ) 12.5 Hz, 3.32 Hz, 1H,
pyrrolidine H-1′), 3.88 (m, 1H, pyrrolidine H-4′), 4.07 (m, 4H,
Ad en osin e Dip h osp h a te N-(ter t-Bu tyloxyca r bon yl)-
h yd r oxym eth yl P yr r olid in e Mon o-ol (t-Boc-4). Adenosine-
5′-monophosphate (AMP) (70 mg, 0.2 mmol) was suspended
in methanol (2 mL), and tri-n-butylamine (48 µL, 0.2 mmol)
was added. The mixture was heated gently in a water bath to
obtain a clear solution (5 min). Methanol was removed in
vacuo, and DMF (2 mL) was added. The clear solution obtained
was evaporated in vacuo at 35 °C to yield a white solid. This
was suspended in another 2 mL portion of DMF, and the
solvent was evaporated. The resulting white solid was sus-
pended in DMF (2 mL), and 1,1′-carbonyl diimidazole (160 mg,
1 mmol) in DMF was added with stirring. The mixture was
allowed to stir at room temperature overnight. The mixture
was then treated with methanol (66 µL, 1.6 mmol) to quench
the reaction and stirred at room temperature for 30 min.

4330 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Koh et al.
pyrrolidine H-5′ and adenosine H-5′), 4.23 (s, 1H, adenosine
H-4′), 4.35 (t, J ) 3.75 Hz, 1H, adenosine H-3′), 4.47 (s, 1H,
pyrrolidine H-2′), 4.57 (t, J ) 5.2 Hz, 1H, adenosine H-2′), 5.92
(d, J ) 5.4 Hz, 1H, adenosine H-1′), 7.97 (s, 1H, adenine H-2),
8.28 (s, 1H, adenine H-8). 13C NMR (D2O) δ 35.34 (pyrrolidine
C-3′), 53.67 (pyrrolidine C-1′), 59.19 (pyrrolidine C-4′), 65.13
(pyrrolidine C-5′), 66.28 (adenosine C-5′), 70.71 (pyrrolidine
C-2′), 71.24 (adenosine C-3′), 75.23 (adenosine C-2′), 84.65
(adenosine C-4′), 88.02 (adenosine C-1′), 114.84 (adenine C-5),
140.32 (adenine C-8), 150.38 (adenine C-4), 153.41 (adenine
C-2), 156.21 (adenine C-6). 31P NMR (D2O, external reference,
H3PO4): δ -10.38 (multiplet). FAB mass spectrum (positive-
ion): calcd m/z 526; found, m/z 527 (MH+). Anal. (C15H24N6O11P2‚
1/4NH4+‚1/2H2O) C, H, N.
external reference, H3PO4): δ -0.30 (d, J ) 15 Hz). MS: m/e
461 (M+). Anal: (C24H32NO6 P) C, H, N.
(2S)-N-(ter t-Bu tyloxyca r bon yl)-2-(p h osp h oxym eth yl)
P yr r olid in e (19). To a solution of N-(tert-butyloxycarbonyl)-
2-(dibenzylphosphoxymethyl) pyrrolidine (0.63 g, 1.36 mmol)
in methanol (25 mL), 5% Pd-C catalyst (60 mg) was added
with stirring, and the flask was purged with H2. The mixture
was then allowed to stir under H2 for 16 h at which point the
TLC indicated a complete absence of starting material. The
mixture was filtered through a Celite pad and washed with
methanol, and the combined filtrate and washings were
1
evaporated in vacuo to yield a colorless oil (0.63 g, 100%). H
NMR (CDCl3): δ 1.20 (s, 9H), 1.72 (m, 4H), 3.07 (m, 2H), 3.71
(m, 3H).
Syn th esis of Ad en osin e Dip h osp h a te (Hyd r oxym eth -
yl)p yr r olid in e (ADP -HP , 5). N-(ter t-Bu tyloxyca r bon yl)-
L-p r olin e, Meth yl Ester . N-(tert-Butyloxycarbonyl)-L-proline
(2.60 g, 12.00 mmol) was dissolved in a mixture of ether (50
mL) and methanol (5 mL) in a wide-mouthed conical flask,
and a solution of diazomethane (16 mmol) in ether (made from
Diazald, according to the manufacturers instructions) was
added with swirling. The solvents were then evaporated to give
a colorless oil (2.78 g, 100%). An analytically pure sample was
obtained by bulb-to-bulb distillation; Rf ) 0.35 (silica gel, 70:
30 hexane:ethyl acetate); Rf ) 0.58 (silica gel, 50:50 hexane:
ethyl acetate). 1H NMR (CDCl3): δ 1.44 (d, J ) 15 Hz, 9H),
1.93 (m, 3H), 2.21 (m, 1H), 3.48 (m, 2H), 3.74 (s, 3H), 4.28
(ddd, J ) 30 Hz, 8.2 Hz, 4.1 Hz, 1H). 13C NMR (CDCl3): δ
(23.35, 24.01), 28.08, (29.59, 30.54), (45.98, 46.22), (51.54,
51.66), (58.41, 58.76), 79.39, (153.424, 154.06), (173.11, 173.37).
MS: m/e 229 (M+). Anal. (C11H19NO4) C, H, N.
(2S)-N-(ter t-Bu tyloxyca r bon yl)-2-h yd r oxym eth ylp yr -
r olid in e. N-(tert-Butyloxycarbonyl)-L-proline (methyl ester,
2.76 g, 12 mmol) was dissolved in dry THF (20 mL), and the
solution was flushed with N2. Lithium borohydride (0.53 g,
24 mmol) was then added under a blanket of N2, and the
mixture was stirred under N2 for 4 h at room temperature.
TLC indicated that the reaction was complete, so excess LiBH4
was destroyed with glacial acetic acid (5 mL) and the solvents
were evaporated in vacuo. The residue was partitioned be-
tween NaHCO3 (5%, 100 mL) and CHCl3 (100 mL), the
aqueous layer was extracted with CHCl3 (2 × 100 mL), and
the combined organic layers were backwashed with water (100
mL), saturated brine (100 mL), and dried (Na2SO4). Evapora-
tion of the solvent gave a colorless oil (2.03 g, 84%); Rf ) 0.22
(silica gel, 70:30 hexane:ethyl acetate). 1H NMR (CDCl3): δ
1.48 (s, 9H), 1.81 (m, 3H), 2.02 (m, 1H), 3.30 (m, 1H), 3.45 (m,
1H), 3.59 (m, 2H), 3.95 (m, 1H), 4.76 (s, br, 1H). 13C NMR
(CDCl3): δ 23.90, 28.34, 28.54, 47.40, 60.01, 67.33, 80.03,
156.95. MS: m/e 201 (M+). Anal. (C10H19NO3) C, H, N.
(2S)-N-(ter t-Bu tyloxyca r bon yl)-2-(d iben zylp h osp h oxy-
m eth yl)p yr r olid in e (18). To a solution of N-(tert-butyloxy-
carbonyl)-2-(hydroxymethyl)pyrrolidine (0.500 g, 2.5 mmol) in
CH2Cl2 (50 mL), 1-H-tetrazole (0.525 g, 7.5 mmol) was added
and the suspension was stirred for 5 min. N,N-Diisopropyl-
dibenzylphosphoramidite (1.3 g, 3.75 mmol) was then added
under N2, and the reaction mixture was stirred at room
temperature for 3 h. The reaction mixture was cooled to
-40 °C in dry ice-acetonitrile bath, mCPBA (0.85 g, 5 mmol,
57-86% grade) in CH2Cl2 (20 mL) was added all at once, the
cooling bath was removed, and the mixture was allowed to
stir at 0 °C for 1 h. The reaction mixture was diluted with
CH2Cl2 (100 mL), washed with Na2SO3 solution (10%, 2 × 50
mL), saturated NaHCO3 solution (2 × 50 mL), water (100 mL),
and saturated brine (100 mL), and dried (MgSO4). Evaporation
of the solvent gave the crude product as a colorless oil (1.5 g),
which was purified by flash column chromatography on silica
gel using hexanes:ethyl acetate (70:30) as the elutant to yield
a product appearing as a colorless oil (0.63 g, 55%); Rf ) 0.35
(silica gel, 50:50 hexane:ethyl acetate). 1H NMR (CDCl3): δ
1.45 (s, 9H), 1.77 (m, 2H), 1.89 (m, 2H), 3.31 (m, 2H), 3.96 (m,
3H), 5.04 (d, J ) 8.1 Hz, 4H), 7.34 (s, 10H). 13C NMR (CDCl3):
δ (22.60, 23.72), 27.44, 28.60, (46.42, 47.16), 56.50, 67.67, 69.77,
(79.53, 80.00), 127.90, 128.60, 135.80, 154.20. 31P NMR (CDCl3,
Ad en osin e Dip h osp h a te N-(ter t-Bu tyloxyca r bon yl)-2-
(h yd r oxym eth yl)p yr r olid in e (t-Boc-ADP -HP , t-Boc-5).
Adenosine-5′-monophosphate (AMP) (70 mg, 0.2 mmol) was
suspended in methanol (2 mL), and tri-n-butylamine (48 µL,
0.2 mmol) was added. The mixture was heated gently in a
water bath to yield a clear solution (5 min). Methanol was
removed in vacuo, and DMF (2 mL) was added. The clear
solution obtained was evaporated in vacuo at 35 °C to yield a
white solid. This was then suspended in another 2 mL portion
of DMF, and the solvent was evaporated. The resulting white
solid was suspended in DMF (2 mL), and 1,1′-carbonyl diim-
idazole (81 mg, 0.500 mmol) in DMF was added with stirring.
The mixture was allowed to stir at room temperature over-
night. The mixture was then treated with methanol (33 µL,
0.8 mmol) to destroy excess carbonyl diimidazole and stirred
at room temperature for 30 min.
(2S)-N-(tert-Butyloxycarbonyl)-2-(phosphoxymethyl)pyrroli-
dine (monotributylamine salt, 232.5 mg, 0.5 mmol) in DMF (5
mL) was added to the reaction mixture, and the mixture was
stirred at room temperature for 24 h. Methanol (5 mL) was
added, and the solvents were evaporated in vacuo until
dryness.
The product was purified by anion exchange chromatogra-
phy. The residue was dissolved in water (100 mL), and the
pH was adjusted to 7.5 with NH4OH solution and applied to a
50 mL column of benzyl DEAE cellulose (1.2 cm × 45 cm)
previously equilibrated with 10 mM NH4HCO3, pH 7.5. The
column was then developed with a linear gradient formed
between 400 mL of 10 mM NH4HCO3 and 400 mL of 400 mM
NH4HCO3 buffered at pH 7.5. Fractions (10 mL) were collected,
and the product eluted into fractions 70-90. Pooled fractions
were evaporated in vacuo at 35 °C and lyophilized to give the
dinucleotide, with very little contamination of AMP. This was
then repurified by preparative anion exchange HPLC to yield
pure dinucleotide (92 mg, 74.8%); Rf ) 0.56 (silica gel, 6:3:1
2-propanol:NH4OH:H2O); Rf ) 0.59 (silica gel, 7:1:2 2-propanol:
triethylamine:H2O); HPLC (reversed phase C18 Bondapak, ion
pair), retention time ) 12.5 min, retention time of AMP ) 2.7
min. 1H NMR (D2O): δ 1.10 (s, 9H, tert-butyl), 1.59 (m, 4H,
pyrrolidine H-2′ and H-3′), 2.97 (m, 2H, pyrrolidine H-1′), 3.58
(m, 3H, pyrrolidine H-4′ and H-5′), 3.95 (s, 2H, adenosine H-5′),
4.12 (s, 1H, adenosine H-4′), 4.26 (m, 1H, adenosine H-3′), 4.53
(dd, J ) 11 Hz, 5.6 Hz, 1H, adenosine H-2′), 5.86 (d, J ) 5.7
Hz, 1H, adenosine H-1′), 7.97 (s, 1H, adenine H-2), 8.24 (s,
1H, adenine H-8). 13C NMR (CDCl3): δ 24.58, 26.56, 28.69,
30.56, 46.63, 60.66, 65.42, 66.17, 71.23, 75.60, 82.29, 85.05,
85.16, 88.95, 119.54, 149.35, 150.95, 157.37. 31P NMR (D2O,
external reference, H3PO4): δ -10.50 (multiplet). MS (MAL-
DI): m/e 609 (M - H)-.
Ad en osin e Dip h osp h a te (Hyd r oxym eth yl)p yr r olid in e
(ADP -HP , 5). Adenosine diphosphate N-(tert-butyloxycarbon-
yl)-2-(hydroxymethyl) pyrrolidine (t-boc-ADP-HP) (80 mg, 0.13
mmol) was mixed with TFA (5 mL) and stirred at room
tempearature for 30 min. TFA was evaporated in vacuo, and
the mixture was dissolved in water (20 mL) and purified by
preparative anion exchange HPLC (Waters, Accell Plus,
quaternary methylamine [QMA] anion exchange media, 500
Å, 1.2 cm × 10 cm, volume ) 10 mL; elutant buffers, 10 mM
NH4HCO3 and 400 mM NH4HCO3). The fractions containing
the product were pooled and lyophilized to give pure ADP-HP

SAR Analysis of Poly(ADP-ribose) Glycohydrolase
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20 4331
(60 mg, 90%); Rf ) 0.30 (silica gel, 6:3:1 2-propanol:NH4OH:
H2O); Rf ) 0.21 (silica gel, 7:1:2 2-propanol:triethylamine:H2O).
1H NMR (D2O): δ 1.61 (m, 1H, pyrrolidine H-2′), 1.88 (m, 3H,
pyrrolidine H-2′ and H-3′), 3.15 (t, J ) 7 Hz, 2H, pyrrolidine
H-1′), 3.74 (m, 1H, pyrrolidine H-4′), 3.86(m, 1H, pyrrolidine
H-5′), 4.07 (m, 3H, pyrrolidine H-5′ and adenosine H-5′), 4.24
(s, 1H, adenosine H-4′), 4.37 (m, 1H, adenosine H-3′), 4.59 (dd,
J ) 11 Hz, 5.6 Hz, 1H, adenosine H-2′), 5.97 (d, J ) 5.8 Hz,
1H, adenosine H-1′), 8.07 (s, 1H, adenine C-2), 8.32 (s, 1H,
adenine H-8). 13CNMR (D2O): δ 24.28 (pyrrolidine C-2′), 27.05
(pyrrolidine C-3′), 46.76 (pyrrolidine C-1′), 60.62 (pyrrolidine
C-4′), 65.95 (pyrrolidine C-5′), 67.10 (adenosine C-5′), 71.38
(adenosine C-3′), 75.00 (adenosine C-2′), 84.76 (adenosine C-4′),
87.81 (adenosine C-1′), 119.80 (adenine C-5), 140.02 (adenine
C-8), 149.82 (adenine C-4), 155.46 (adenine C-2), 159.13
(adenine C-6). 31P NMR (D2O, external reference, H3PO4): δ
-10.5 (multiplet). MS (FAB mass spectrum, positive ion):
calcd m/z 510; found, m/z 511 (MH+).
8-Ch lor op h en ylt h ioa d en osin e 5′-Dip h osp h a t e (H y-
d r oxym eth yl)p yr r olid in ed iol (10). 8-Bromoadenosine (2 g,
5.8 mmol), p-chlorobenzenethiol (2 g, 14 mmol), and LiOCH3
(1.3 g) were reacted in 40 mL of anhydrous DMF at 60 °C for
4 h. The solvent was evaporated in vacuo, and the residue was
purified by chromatography on silica gel developed in meth-
ylenechloride/methanol (10:1) to obtain 8-CPT-adenosine.
8-CPT-AMP was prepared by the method of Yoshikawa et al.24
8-CPT-adenosine (2 g) was dissolved in 20 mL of trimethyl
phosphate followed by the addition of 25 mmol of POCl3. The
mixture was left at 0 °C overnight. Purification was obtained
on a DEAE-cellulose column using a linear gradient of 0-0.25
M NH4HCO3. (2R,3R,4S)-1-(tert-Butyloxycarbonyl)-2-[(phos-
phooxy)methyl]pyrrolidine-3,4-diol was prepared according to
Ramsinghani et al.25 8-CPT-AMP was coupled to (2R,3R,4S)-
1-(tert-butyloxycarbonyl)-2-[(phosphooxy)methyl]pyrrolidine-
3,4-diol by a method based on the procedure of Michelson23
and purified by DEAE-cellulose anion exchange chromatog-
raphy. The tert-butyloxycarbonyl group was removed by react-
ing in 10% TFA in methylenechloride at room temperature
for 30 min to obtain the desired product. TLC (silica gel,
2-propanol/concd NH4OH/water 6:3:1) Rf ) 0.57. 1H NMR
(D2O, referenced to HDO at δ 4.67): δ 3.20-3.3.25 (d, 1H,
CH2N), 3.37-3.47 (d, 1H, CH2N), 3.65 (broad s, 1H, CHN),
4.08-4.28 (m, 7H), 4.48-4.53 (t, 1H,CHOH), 5.14-5.22 (t, 1H,
CHOH), 6.10-6.22 (d, 1H, anomeric H), 7.16-7.43 (d, 2H,
PhH), 7.45-7.61 (d, 2H, PhH), 8.19 (s, 1H, adenosyl H).
(2 mL) was added. The clear solution obtained was evaporated
in vacuo at 35 °C to give a white solid. This was then
suspended in another 2 mL portion of DMF, and the solvent
was evaporated. The resulting white solid was suspended in
DMF (2 mL), and 1,1′-carbonyldiimidazole (160 mg, 1 mmol)
in DMF was added with stirring. The mixture was allowed to
stir at room temperature overnight. The mixture was treated
with methanol (66 µL, 1.6 mmol) to destroy excess carbonyl-
diimidazole and stirred at room temperature for 30 min.
+
To this mixture, a solution of Cbz-HPD phosphate (NH4
salt) in DMF (2 mL) was added, and the mixture was allowed
to stir at room temperature for 7 days, at which time TLC
indicated the completion of the reaction. Methanol (2 mL) was
then added, and the solvents were evaporated in vacuo at 35
°C, and the residue was dissolved in water and acidified with
dilute HCl. This mixture was then loaded onto a reversed-
phase column (Amberchrome CG-71md, 1.5 cm × 100 cm, 175
mL) with deionized water elution. Fractions (10-11 mL) were
collected, and fractions 18-40 were pooled, which appeared
as three peaks on analytical reversed-phase HPLC (Bondapak
C18) analysis. The combined fractions were made basic (pH
9.0) with 1 M NH4OH and purified by anion exchange
chromatography (benzyl-DEAE cellulose, 1.5 cm × 47 cm,
volume ) 83 mL) and developed using with a linear gradient
formed between 10 mM NH4HCO3 (400 mL) and 300 mM NH4-
HCO3 (400 mL). The chromatography produced well-separated
peaks, from which the fractions containing the dinucleotide
product were pooled and lyophilized to give a chromatographi-
cally pure product.
The product was dissolved in water (20 mL), catalyst 5%
Pd-C was added (10 mg), and the mixture was allowed to stir
at room temperature under H2 overnight. The mixture was
then filtered through a Celite pad, the solvent was evaporated,
and the residue was dissolved in water (100 mL) with the pH
adjusted to 7.5 and purified by anion exchange chromatogra-
phy using benzyl-DEAE cellulose (1.5 cm × 47 cm, 83 mL).
The column was developed with a linear gradient of 10 mM
NH4HCO3 and 250 mM NH4HCO3 (400 mL each). Lyophiliza-
tion of the fractions containing the product yielded the pure
dinucleotide (70 mg, 62%). HPLC (reversed-phase C18 Bonda-
pak, ion pair), retention time ) 2.4 min, retention time of
1
GMP ) 3.8 min. H NMR (D2O): δ 3.18 (d, J ) 11.5 Hz, 1H,
pyrrolidine H-1′), 3.32 (dd, J ) 12 Hz, 2 Hz, 1H, pyrrolidine
H-1′), 3.57 (m, 1H, pyrrolidine H-4′), 4.06 (m, 3H), 4.20 (m,
4H), 4.36 (t, J ) 4 Hz, 1H, guanosine H-3′), 4.60 (m, 1H,
guanosine H-2′), 5.78 (d, J ) 4.8 Hz, 1H, guanosine H-1′), 7.95
(s, 1H, guanine H-8).
N6-Ben zyla d en osin e 5′-Dip h osp h a te (Hyd r oxym eth yl)-
p yr r olid in ed iol (11). N6-Benzyladenosine was phosphory-
lated to obtain N6-benzyl-AMP as described above. N6-Benzyl-
AMP was coupled to Cbz-3 as described above. The benzyloxy-
carbonyl group was removed by catalytic hydrogenation to
obtain the desired product. TLC (silica gel, 2-propanol/
concentrated NH4OH/water 6:3:1) Rf ) 0.57. 1H NMR (D2O,
1% w/w DSS): δ 8.39 (s,1H), δ 8.16 (s,1H), δ 7.31-7.23 (m,5H),
δ 6.04 (d,1H), δ 4.26-4.12 (m, 4H), δ 3.68-3.42 (m,4H), δ
3.38-3.22 (m, 4H).
SAR Stu d ies. PARG inhibition assays were performed in
at least triplicate as described previously,42 using 0.25 ng of
bPARG or 5 ng of rPARG-CF with selected inhibitors included
in reaction mixes (except for control) in the presence of 10 µM
[R-32P]ADP-ribose polymers in a total volume of 30 µL. The
catalytic activity of PARG was measured in the presence of
varying concentrations of inhibitors. Data were fit to the
equation describing a sigmoidal dose response relation between
% activity (100 to 0%) and log[inhibitor] using GraphPad prism
version 3.02 for Windows, GraphPad Software, San Diego, CA,
www.graphpad.com. The procedure resulted in calculated
values for logIC50 and for the standard error in logIC50, which
are reported.
N6-Hexyla d en osin e 5′-Dip h osp h a te (Hyd r oxym eth yl)-
p yr r olid in ed iol (12). 6-Chloropurine riboside was phospho-
rylated to obtain 6-chloropurine riboside 5′-monophosphate as
described above. N6-Hexyl-AMP was prepared by nucleophilic
aromatic substitution with hexylamine at 60 °C.41 N6-Hexyl-
AMP was coupled to Cbz-3 as described above. The benzyl-
oxycarbonyl group was removed by catalytic hydrogenation to
obtain the desired product. TLC (silica gel, 2-propanol/concd
Ack n ow led gm en t . Supported by research grants
from the Ohio Division of the American Cancer Society
and de Arce Memorial Endowment Fund to J .T.S. and
the National Institutes of Health Grant CA 43894 to
M.K.J .
1
NH4OH/water 6:3:1) Rf ) 0.59. H NMR (D2O, 1% w/w DSS):
δ 8.32 (s,1H), δ 8.11(s,1H), δ 6.02(d,1H), δ 4.58-4.65 (m,1H),
δ 4.52-3.90 (m, 3H), δ 3.28-3.25 (m, 4H), δ 2.91-2.75 (m,
4H), δ 1.68-1.62 (m, 4H), δ 1.17-1.12 (m, 4H),δ 0.73-0.52
(m, 3H).
Refer en ces
Gu a n osin e 5′-Dip h osp h a t e (H yd r oxym et h yl)p yr r o-
lid in ed iol (GDP -HP D, 13). Guanosine-5′-monophosphate
(GMP, 72.64 mg, 0.2 mmol) was suspended in methanol (2 mL),
and tri-n-butylamine (37 mg, 48 µL, 0.2 mmol) was added. The
mixture was heated gently in a water bath to yield a clear
solution (5 min). Methanol was removed in vacuo, and DMF
(1) Alvarez-Gonzalez, R.; J acobson, M. K. Characterization of
polymers of adenosine diphosphate ribose generated in vitro and
in vivo. Biochemistry 1987, 26, 3218-3224.
(2) J acobson, M. K.; J acobson, E. L. Discovering new ADP-ribose
polymer cycles: protecting the genome and more. Trends Bio-
chem. Sci. 1999, 24, 415-417.

4332 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 20
Koh et al.
(3) Ame´, J . C.; J acobson, E. L.; J acobson, M. K. ADP-ribose polymer
metabolism. From DNA Damage and Stress Signaling to Cell
Death: Poly-ADP-Ribosylation Reactions; Oxford University
Press: New York, 2000; pp 1-34.
(4) Althaus, F. R. Poly ADP-ribosylation: a histone shuttle mech-
anism in DNA excision repair. J . Cell Sci. 1992, 102, 663-670.
(5) de Murcia, G.; Menissier de Murcia, J . Poly(ADP-ribose) polym-
erase: a molecular nick-sensor. Trends Biochem. Sci. 1994, 19,
172-176.
(6) J acobson, E. L.; Smith, J . Y.; Wielckens, K.; Hilz, H.; J acobson,
M. K. Cellular recovery of dividing and confluent C3H10T1/2
cells from N-methyl-N′-nitro-N-nitrosoguanidine in the presence
of ADP-ribosylation inhibitors. Carcinogenesis 1985, 6, 715-718.
(7) Pieper, A. A.; Verma, A.; Zhang, J .; Snyder, S. H. Poly (ADP-
ribose) polymerase, nitric oxide and cell death. Trends Pharma-
col. Sci. 1999, 20, 171-181.
(8) Smith, S.; de Lange, T. Tankyrase promotes telomere elongation
in human cells. Curr. Biol. 2000, 10, 1299-1302.
(9) Szabo, C.; Dawson, V. Role of Poly(ADP-ribose) Synthetase in
Inflammation and Ischaemia-Reperfusion. Trends Pharmacol.
Sci. 1998, 19, 287-298.
(10) Burkart, V.; Wang, Z. Q.; Radons, J .; Heller, B.; Herceg, Z.; et
al. Mice Lacking the Poly(ADP-ribose) Polymerase Gene are
Resistant to Pancreatic â-Cell DEatruction and Diabetes De-
velopment Induced by Streptozocin. Nat. Med. 1999, 5, 314-
319.
(11) Mandir, A. S.; Przedborski, S.; J ackson-Lewis, V.; Wang, Z. Q.;
Simbulan-Rosenthal, C. M.; et al. Poly(ADP-ribose) polymerase
activation mediates 1-methyl-4-phenyl-1, 2,3,6-tetrahydropyri-
dine (MPTP)-induced parkinsonism. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 5774-5779.
(12) Miwa, M.; Tanaka, M.; Matsushima, T.; Sugimura, T. Purifica-
tion and properties of glycohydrolase from calf thymus splitting
ribose-ribose linkages of poly(adenosine diphosphate ribose). J .
Biol. Chem. 1974, 249, 3475-3482.
(13) J uarez-Salinas, H.; Sims, J . L.; J acobson, M. K. Poly(ADP-ribose)
levels in carcinogen-treated cells. Nature 1979, 282, 740-741.
(14) Wielckens, K.; Schmidt, A.; George, E.; Bredehorst, R.; Hilz, H.
DNA fragmentation and NAD depletion. Their relation to the
turnover of endogenous mono(ADP-ribosyl) and poly(ADP-ribo-
syl) proteins. J . Biol. Chem. 1982, 257, 12872-12877.
(15) Alvarez-Gonzalez, R.; Althaus, F. R. Poly(ADP-ribose) catabolism
in mammalian cells exposed to DNA-damaging agents. Mutat.
Res. 1989, 218, 67-74.
(24) Yoshikawa, M.; Kato, T.; Takenishi, T. A novel method for
phosphorylation of nucleosides to 5′-nucleotides. Tetrahedron
Lett. 1967, 50, 5065-5068.
(25) Ramsinghani, S.; Koh, D. W.; Ame´, J . C.; Strohm, M.; J acobson,
M. K.; et al. Syntheses of photoactive analogues of adenosine
diphosphate (hydroxymethyl)pyrrolidinediol and photoaffinity
labeling of poly(ADP-ribose) glycohydrolase. Biochemistry 1998,
37, 7801-7812.
(26) Ame´, J . C.; J acobson, E. L.; J acobson, M. K. Molecular hetero-
geneity and regulation of poly(ADP-ribose) glycohydrolase. Mol.
Cell Biochem. 1999, 193, 75-81.
(27) Aoki, K.; Nishimura, K.; Abe, H.; Maruta, H.; Sakagami, H.; et
al. Novel inhibitors of poly(ADP-ribose) glycohydrolase. Biochim.
Biophys. Acta 1993, 1158, 251-256.
(28) Tavassoli, M.; Tavassoli, M. H.; Shall, S. Effect of DNA inter-
calators on poly(ADP-ribose) glycohydrolase activity. Biochim.
Biophys. Acta 1985, 827, 228-234.
(29) Davies, G.; Henrissat, B. Structures and mechanisms of glycosyl
hydrolases. Structure 1995, 3, 853-859.
(30) Ruf, A.; Rolli, V.; de Murcia, G.; Schulz, G. E. The mechanism
of the elongation and branching reaction of poly(ADP-ribose)
polymerase as derived from crystal structures and mutagenesis.
J . Mol. Biol. 1998, 278, 57-65.
(31) Koh, D. W.; Patel, C. N.; Ramsinghani, S.; Slama, J . T.; Oliveira,
M. A.; et al. Identification of an inhibitor binding site of poly-
(ADP-ribose) glycohydrolase. Biochemistry 2003, 42, 4855-4863.
(32) Ikehara, M.; Uesugi, S.; Yoshida, K. Studies on the conformation
of purine nucleosides and their 5′- phosphates. Biochemistry
1972, 11, 830-836.
(33) Czarnecki, J . J . Tautomerism of 2-azidoadenine nucleotides.
Effects on enzyme kinetics and photoaffinity labeling. Biochim.
Biophys. Acta 1984, 800, 41-51.
(34) Miller, J . P.; Boswell, K. H.; Muneyama, K.; Simon, L. N.; Robins,
R. K.; et al. Synthesis and biochemical studies of various
8-substituted derivatives of guanosine 3′,5′-cyclic phosphate,
inosine 3′,5′-cyclic phosphate, and xanthosine 3′,5′-cyclic phos-
phate. Biochemistry 1973, 12, 5310-5319.
(35) Sandberg, M.; Butt, E.; Nolte, C.; Fischer, L.; Halbrugge, M.; et
al. Characterization of Sp-5,6-dichloro-1-â-D-ribofuranosylbenz-
imidazole-3′,5′-monophosphorothioate (Sp-5,6-DCl-cBiMPS) as
a potent and specific activator of cyclic-AMP-dependent protein
kinase in cell extracts and intact cells. Biochem. J . 1991, 279,
521-527.
(36) Butt, E.; Nolte, C.; Schulz, S.; Beltman, J .; Beavo, J . A.; et al.
Analysis of the functional role of cGMP-dependent protein kinase
in intact human platelets using a specific activator 8-para-
chlorophenylthio-cGMP. Biochem. Pharmacol. 1992, 43, 2591-
2600.
(16) Lin, W.; Ame´, J . C.; Aboul-Ela, N.; J acobson, E. L.; J acobson,
M. K. Isolation and characterization of the cDNA encoding
bovine poly(ADP-ribose) glycohydrolase. J . Biol. Chem. 1997,
272, 11895-11901.
(17) Bowman, K. J .; White, A.; Golding, B. T.; Griffin, R. J .; Curtin,
N. J . Potentiation of anti-cancer agent cytotoxicity by the potent
poly(ADP-ribose) polymerase inhibitors NU1025 and NU1064.
Br. J . Cancer 1998, 78, 1269-1277.
(18) Bowman, K. J .; Newell, D. R.; Calvert, A. H.; Curtin, N. J .
Differential effects of the poly (ADP-ribose) polymerase (PARP)
inhibitor NU1025 on topoisomerase I and II inhibitor cytotoxicity
in L1210 cells in vitro. Br. J . Cancer 2001, 84, 106-112.
(19) Slama, J . T.; Aboul-Ela, N.; J acobson, M. K. Mechanism of
inhibition of poly(ADP-ribose) glycohydrolase by adenosine
diphosphate (hydroxymethyl)pyrrolidinediol. J . Med. Chem.
1995, 38, 4332-4336.
(20) Slama, J . T.; Aboul-Ela, N.; Goli, D. M.; Cheesman, B. V.;
Simmons, A. M.; et al. Specific inhibition of poly(ADP-ribose)
glycohydrolase by adenosine diphosphate (hydroxymethyl)pyr-
rolidinediol. J . Med. Chem. 1995, 38, 389-393.
(21) Goli, D. M.; Cheesman, B. V.; Hassan, M. E.; Lodaya, R.; Slama,
J . T. Synthesis of (2R,3R,4S)-2-hydroxymethylpyrrolidine-3,4-
diol from (2S)-3,4-dehydroproline derivatives. Carbohydr. Res.
1994, 259, 219-241.
(37) Shimokawa, T.; Masutani, M.; Nagasawa, S.; Nozaki, T.; Ikota,
N.; et al. Isolation and cloning of rat poly(ADP-ribose) glycohy-
drolase: presence of a potential nuclear export signal conserved
in mammalian orthologs. J . Biochem. (Tokyo) 1999, 126, 748-
755.
(38) Giner, H.; Simonin, F.; de Murcia, G.; Menissier-de Murcia, J .
Overproduction and large-scale purification of the human poly-
(ADP-ribose) polymerase using a baculovirus expression system.
Gene 1992, 114, 279-283.
(39) Thomassin, H.; J acobson, M. K.; Guay, J .; Verreault, A.; Aboul-
Ela, N.; et al. An affinity matrix for the purification of poly-
(ADP-ribose) glycohydrolase. Nucleic Acids Res. 1990, 18, 4691-
4694.
(40) Hatakeyama, K.; Nemoto, Y.; Ueda, K.; Hayaishi, O. Purification
and characterization of poly(ADP-ribose) glycohydrolase. Dif-
ferent modes of action on large and small poly(ADP-ribose). J .
Biol. Chem. 1986, 261, 14902-14911.
(41) Fleysher, M. H.; Hakala, M. T.; Bloch, A.; Hall, R. Synthesis
and Biological Activity of Some N6-alkyladenosines. J . Med.
Chem. 1968, 11, 717-720.
(42) Me´nard, L.; Poirier, G. G. Rapid assay of poly(ADP-ribose)
glycohydrolase. Biochem. Cell Biol. 1987, 65, 668-673.
(22) Hoard, D. E.; Ott, D. G. Conversion of Mono- and Oligodeoxyri-
bonucleotides to 5′-triphosphates. J . Am. Chem. Soc. 1965, 87.
(23) Michelson, A. M. Synthesis of Nucleotide Anhydrides by Anion
Exchange. Biochim. Biophys. Acta 1964, 91, 1-13.
J M020541U