An HIV reverse transcriptase-selective nucleoside chain terminator

Article abstract of DOI:10.1021/ja020639y

The synthesis of a 2-,3-dideoxynucleoside cytidine analogue, but one that lacks the O2-carbonyl, is described from 2-aminopyridine in an overall yield of 60%. The synthesis of the 2-pyridone C-nucleoside relies upon the use of a Heck-type coupling between an appropriately protected sugar glycal and the 5-iodo derivative of 2-aminopyridone. Upon conversion of the dideoxynucleoside to the corresponding 5-triphosphate, the analogue ddNTP is observed to be a reasonable substrate with HIV reverse transcriptase (for a template dG residue), but is not a substrate for calf thymus DNA polymerase α or for human DNA polymerase β. With the human mitochondrial DNA polymerase the analogue functions as a poor substrate. The observed polymerase selectivities appear to arise from the absence of the O2-carbonyl, which either results in a destabilized Watson-Crick base pair or represents a critical contact for some polymerases. Copyright

Full text of DOI:10.1021/ja020639y

Published on Web 12/21/2002
An HIV Reverse Transcriptase-Selective Nucleoside Chain Terminator
Andrew W. Fraley,† Dongli Chen,† Kenneth Johnson,‡ and Larry W. McLaughlin*,†
Department of Chemistry, Boston College, 2609 Beacon Street, Chestnut Hill, Massachusetts 02467-3801,
Institute for Cellular and Molecular Biology, The UniVersity of Texas at Austin, Louise and James,
Robert Moffett Molecular Biology Building, 2500 Speedway, A4800, MBB 3.122. Austin, Texas 78712-1095
Received May 3, 2002 ; E-mail: larry.mclaughlin@bc.edu
The mode of action of anti-retroviral (e.g., anti-HIV) nucleoside
chain terminators such as AZT, ddI, and ddC involves their in vivo
conversion to the corresponding 5′-triphosphates, which enables
them to function as substrates for retroviral reverse transcriptases
(RT). In the absence of a 3′-hydroxyl, chain termination results
after the incorporation of such derivatives and DNA synthesis of
the viral cDNA genome ceases. A common problem with antiviral
therapies involving chain-terminating nucleosides is both their
immediate and delayed toxicity effects. Delayed toxicity most
commonly takes the form of peripheral neuropathy, myopathy, or
pancreatitis, but the nature of toxicity depends on the specific
analogue.1-4 A variety of mechanisms can account for toxicity
effects, but the most obvious involve substrate activity with host
polymerases. In proliferating cells, toxicity effects (such as the bone
marrow toxicity of AZT) may be due to the actions of human
chromosomal polymerases such as Pol â involved in DNA repair.5
In nonproliferating cells toxicity appears to be correlated with
mitochondrial DNA synthesis (the role of Pol γ).5-10 In fact, many
of the symptoms resulting from antiviral nucleosides are similar to
the dysfunctions reported for mitochondrial genetic disorders.6-9
Human mitochondrial polymerase γ appears to be the most sensitive
to the presence of some antiviral nucleosides.10-12
Our design of an HIV-selective nucleoside chain terminator relied
upon recent work suggesting that some mammalian polymerases
make contacts to the minor groove functional groups of base
residues in both the primer and template strands.13,14 Similarly,
studies of dNTP building blocks designed on the basis of geometric
shape have suggested that the loss of functionality can result in
reduced rates of incorporation by some mammalian polymerases.15-17
The described nucleoside analogue (dd2APy, Figure 1) incorporated
two design features; one was a modification of the nucleobase
residue to eliminate the O2-carbonyl, and the second was the use
of a 2′,3′-dideoxycarbohydrate to eliminate the 3′-hydroxyl. In other
respects the derivative was designed to form a bidentate Watson-
Crick base pair with dG (Figure 1), but to maintain the correct
tautomeric form for base pairing, the derivative was prepared as a
C-nucleoside. Nearly all of the current nucleoside antivirals contain
essentially native nucleobases coupled to appropriately modified
carbohydrate residues. The described modifications were imple-
mented to result in a derivative in which the heterocyclic nucleobase
could provide the desired polymerase selectivity (e.g., HIV vs
mammalian polymerases), while the 2′,3′-dideoxy sugar would
ensure chain termination of the elongating strand.
Figure 1. Structure of dd2APy (left). Watson-Crick base-pairing scheme
for an incoming dd2APy triphosphate analogue and template dG (right).
of the nucleoside to generate 5, the carbonyl was eliminated by
conversion to the p-toluylsulfonyl hydrazone (6) and then smoothly
reduced to the 2′,3′-dideoxynucleoside (7). This latter transformation
is based upon earlier work19,20 with R,â-unsaturated steroids. A
small portion of 7 was converted to the corresponding 5′-
triphosphate (dd2APyTP) by first protection of the exocyclic amino
group as the N,N-dimethylamidine followed by conventional
phosphorylation/pyrophosphate treatment and finally removal of
the amidine in aqueous ammonia.
The dideoxy nucleoside 7 as the 5′-triphosphate derivative
(dd2APyTP) was incubated with either calf thymus DNA poly-
merase R or human DNA polymerase â over a range of concentra-
tions (50-1000 µM) for a 45-min period with a primer/template
complex. No significant elongation of the primer strand was
observed (Figure 2). In a control reaction both enzymes effectively
use dCTP as a substrate. In related control reactions, incorporation
of dCTP by DNA polymerase â at low (2 µΜ) or relatively high
(100 µM) concentrations was not substantially inhibited by the
presence of dd2APyTP (e1 mM). By comparison at low concentra-
tions of dCTP (2 µM), DNA polymerase R could be inhibited by
0.5 and 1.0 mM concentrations of dd2APyTP. This experiment
suggests that in the absence of the O2-carbonyl, the dd2APyTP
derivative does not have significant binding activity with either of
these mammalian enzymes, except at very high concentrations.
Scheme 1a
The 2′,3′-dideoxy nucleoside chain terminator (7) could be
prepared in 60% overall yield from a simple five-step synthesis
(Scheme 1). After iodonation of 2-aminopyridine (1), the product
(2) was coupled to the silyl-protected glycal (3) by a Heck-type
reaction18 to generate solely the â-nucleoside (4). After deprotection
a Conditions: i. I2/HIO4/H2SO4/CH3COOH; ii. (dba)2Pd0/Ph3P/i-Pr2EtN/
CH3CN; iii. n-Bu4N+F-/CH3COOH/THF; iv. p-toluylsulfonylhydrazide; v.
Na(OAc)3BH, CH3COOH/CH3CN, H2O quench.
† Boston College.
‡ The University of Texas at Austin.
9
616
J. AM. CHEM. SOC. 2003, 125, 616-617
10.1021/ja020639y CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
polymerase. HIV RT is the enzyme in this study that exhibits the
least dependence upon the O2-carbonyl and will incorporate the
analogue dideoxy chain terminator much more efficiently than any
of the other three enzymes.
Effective antiviral activity requires that the analogue be converted
to the triphosphate. Initial experiments with deoxycytidine kinase
indicate that the dd2APy derivative is a poor substrate for this
enzyme, but this initial step in the pathway to the triphosphate can
be circumvented by preparation of neutral phosphate analogues
designed to permit intracellular release of nucleoside monophos-
phates.24,25
Although the analogue described here is not as effective a
substrate as is dCTP, incorporation of a single chain-terminating
derivative during the course of viral genome replication is in
principle sufficient for chain termination. The absence or low
activity with human polymerases could result in reduced toxicity
effects for this class of derivatives.
Figure 2. Primer extension reactions with dd2APy triphosphate were
performed for the enzymes, HIV RT, calf thymus DNA polymerase R,
human DNA polymerase â, and human mitochondrial DNA polymerase γ,
each under its optimal pH conditions. For Pol γ the primer contains a
phosphorothioate diester at the 3′-d(ApA).
This observation is consistent with the results reported for
hydrophobic bases that also lack minor groove functionality and
are reported to be inactive with these enzymes.15 Initial work with
human mitochondrial DNA polymerase γ21 resulted largely in
primer degradation by the exonuclease activity of the enzyme. To
better assess analogue incorporation, we altered the primer to
contain a 3′-phosphorothioate diester, a derivative known to be
refractory to nuclease activity.22 Under these conditions we could
observe moderate elongation by the mitochondrial enzyme. In a
time-course assay (100 µM dd2APy triphosphate) roughly 11% of
the primer was elongated with the analogue by human mitochondrial
polymerase γ, while under the same conditions HIV RT elongated
50% of the primer.
Acknowledgment. We thank Dr. Dutschman and Professor
Cheng, Yale University, for a sample of human deoxycytidine
kinase. This work was supported by NSF MCB-0066776.
Supporting Information Available: Synthetic procedures for the
analogue nucleoside, procedures for enzyme assays and selected gels
(PDF). This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Faulds, D.; Brogden, R. N. Drugs 1992, 44, 94-116.
(2) Whittington, R.; Brogden, R. N. Drugs 1992, 44, 656-683.
(3) Wilde, M. I.; Langtry, H. D. Drugs 1993, 46, 515-578.
(4) Adkins, J. C.; Peters, D. H.; Faulds, D. Drugs 1997, 53, 1054-1080.
(5) Parker, W. B.; Cheng, Y.-C. J. NIH Res. 1994, 6, 57-61.
(6) Horn, D. M.; Neeb, L. A.; Colacino, J. M.; Richardson, F. C. AntiViral
Res. 1997, 34, 71-74.
Analogue triphosphates capable of selective termination of the
replication process catalyzed by HIV RT have the potential to
become effective antivirals. On the other hand, similar termination
of replication of the mitochondrial genome has been linked to
toxicity effects. Limited toxicity may be related in part to the extent
of exonucleolytic removal of the incorporated dideoxy derivative.11
To determine whether incorporation of dd2APy by the mitochon-
drial polymerase could be reversed, we incubated the dd2APy-
elongated primer with the human mitochondrial polymerase. Over
a period of 12 h the analogue was removed (kobs ) 0.02 min-1).
This rate is commensurate with the removal rate for the dideoxy
derivative 3TC and significantly faster than that observed for ddC.11
The observed polymerase selectivity can be explained by two
possible mechanisms: (i) Duplexes containing dG-d2APy base pairs
appear to be destabilized more than what would be expected for
the simple loss of an interstrand hydrogen bond.21 A mispairing
effect may be present and may alter the relative positions of the
R-phosphate of the incoming dNTP and 3′-hydroxyl of the 3′-primer
residue. Such relative changes in geometry may be more significant
for the mammalian polymerases than for the HIV RT. (ii) Simple
hydrophobic isosteric triphosphates in the absence of a 3′-minor
groove functional group are generally poor substrates13 for mam-
malian polymerases. These reports suggest that a minor groove
functional group on the incoming dNTP is a prerequisite for
effective substrate activity,13 although this requirement varies with
the specific polymerase.15 The critical minor groove functional
group of dCTP is the O2-carbonyl, a residue absent in the dd2APy
derivative (Figure 1). Its absence may be most significant for the
mammalian polymerases R and â where essentially no incorporation
of the dd2APy analogue is observed (Figure 2). Human mitochon-
drial polymerase γ exhibits somewhat less dependence upon the
presence of the minor groove O2-carbonyl, but the analogue
dd2APyTP is still at best a relatively poor substrate for this
(7) Lewis, W.; Dalakas, M. C. Nat. Med. 1995, 1, 417-422.
(8) McKenzie, R.; Fried, M. W.; Sallie, R.; Conjeevaram, H.; Dibisceglie,
A. M.; Park, Y.; Savarese, B.; Kleiner, D.; Tsokos, M.; Luciano, C.; Pruett,
T.; Stotka, J. L.; Straus, S. E.; Hoofnagle, J. H. N. Engl. J. Med. 1995,
333, 1099-1105.
(9) Simon, D. K.; Johns, D. R. Annu. ReV. Med. 1999, 50, 111-127.
(10) Chen, C. H.; Vazquezpadua, M.; Cheng, Y. C. Mol. Pharmacol. 1991,
39, 625-628.
(11) Feng, J. Y.; Johnson, A. A.; Johnson, K. A.; Anderson, K. S. J. Biol.
Chem. 2001, 276, 23832-23837.
(12) Johnson, A.; Feng, J.; Anderson, K.; Johnson, K. FASEB J. 2001, 15,
A520-A520.
(13) Morales, J. C.; Kool, E. T. J. Am. Chem. Soc. 1999, 121, 2323-2324.
(14) Osheroff, W. P.; Beard, W. A.; Wilson, S. H.; Kunkel, T. K. J. Biol.
Chem. 1999, 274, 20749-20752.
(15) Morales, J. C.; Kool, E. T. Biochemistry 2000, 39, 1279-12988.
(16) Tae, E. L.; Wu, Y.; Xia, G.; Schultz, P. G.; Romesberg, F. E. J. Am.
Chem. Soc. 2001, 123, 7439-7440.
(17) Wu, Y.; Ogawa, A. K.; Berger, M.; McMinn, D. L.; Schultz, P. G.;
Romesberg, F. E. J. Am. Chem. Soc. 2000, 122, 7621-7632.
(18) Daves, G. D. Acc. Chem. Res. 1990, 23, 201-206.
(19) Hutchins, R. O.; Natale, N. R. J. Org. Chem. 1978, 43, 2299-2304.
(20) Kabalka, G. W.; Hutchins, R.; Natale, N. R.; Yang, D. T. C.; Broach, V.
Org. Synth. 1988, 6, 293-299.
(21) Johnson, A. A.; Tsai, Y. C.; Graves, S. W.; Johnson, K. A. Biochemistry
2000, 39, 1702-1708.
(22) Vorsberg, H. P.; Eckstein, F. J. Biol. Chem. 1982, 257, 6595-6599.
(23) Guckian, K. M.; Morales, J. C.; Kool, E. T. J. Org. Chem. 1998, 63,
9652-9656.
(24) Qiu, Y.-L.; Ptak, R. G.; Breitenbach, J. M.; Lin, L.-S.; Cheng, Y.-C.;
Drach, J. C.; Kern, E. R.; Zemlicka, J. AntiViral Res. 1999, 43, 37-53.
(25) Meier, C. Mini-ReV. Med. Chem. 2002, 2, 219-234.
JA020639Y
9
J. AM. CHEM. SOC. VOL. 125, NO. 3, 2003 617