`
`Infectious Disorders - Drug Targets 2006, 6, 17-29
`
`17
`
`S. S. Carroll and D. B. Olsen*
`
`Department ofAntiviral Research, Merck Research Laboratories, West Point, PA, 19486
`
`infections, 15 are
`Abstract: Of the 30 compounds currently marketed in the United States for treatment of viral
`nucleoside analogs, demonstrating the utility of this class of compound as a source ofantiviral drugs. The success of
`nucleoside analogs in treating other viral
`infections provides a compelling rationale for the significant effort that is
`currently being devoted to the discovery and development of nucleoside analogs to treat infection by hepatitis C virus
`(HCV) that may lead to improvements in response rates compared to currently available therapies. Several different
`approaches have been adopted to identify promising analogs, including the use of surrogate viruses in cell culture assays,
`screening in the cell-based bicistronic HCV replicon assay, and screening nucleoside triphosphates for the ability to
`inhibit the activity of the HCV RNA-dependent RNA polymerase in vitro. Several classes of ribonucleoside analogs with
`modifications of the ribose inhibit HCV replication. Nucleoside analogs incorporating a 2’-C-methyl modification are
`potent inhibitors in the replicon assay in the absence of cytotoxicity, and appear to exert their inhibition by acting as
`functional chain terminators of RNA synthesis. NM283, a prodrug of 2’-C-methylcytidine, has entered clinical trials and
`demonstrated viral load reductions in subjects infected with genotype 1 HCV, a genotype knownto be difficult to treat
`effectively with currently approved therapies. Overall, results to date offer encouragement that improved therapiesto treat
`HCYinfection including newly developed nucleoside analogs may becomeavailable within the next few years.
`
`Keywords: Nucleoside analog, replicon,structure-activity relationships, RNA polymerase, chain terminator.
`
`HCV BACKGROUND
`
`the negative strand copy of the viral genome and subsequent
`positive strand copies that serve as the genomes of progeny
`HCV'was recognizedasthe infectious agent responsible
`virus. Since it is absolutely required for viral infectivity [6],
`for community-acquired non-A non-B hepatitis in 1989 [1].
`HCV RdRp is a validated and attractive target
`for the
`This discovery made possible the developmentof diagnostic
`development of new treatments for HCV infection based on
`tests for HCV that have reduced the risk of infection through
`administration of compounds
`that directly inhibit viral
`blood transfusion. Yet, estimates of the total number of
`enzyme function.
`infected individuals are currently 170-200 million worldwide
`Extensive structural
`information is now available for
`
`[2]. HCV_infection is the leading cause of liver
`
`
`
`
`HCV RdRpincluding co-crystals with some non-nucleoside
`transplantation in the United States. Currently preferred
`inhibitors [7-11]. As with other polymerases such as the
`therapies to treat HCV infection consist of six to twelve
`Klenow fragment of E. coli DNA polymerase J and HIV
`month courses of combinations of pegylated interferon a and
`reverse transcriptase, the overall structure of HCV RdRp has
`ribavirin, which result in sustained viral response (SVR, no
`been compared to a right hand with fingers, palm and thumb
`detectable virus six months after cessation of therapy) in 40-
`subdomains. Unique to the HCV RdRp,though, are extended
`60% of treated patients [3, 4]. The SVR rates with patients
`finger domains that come into contact with the thumb
`infected with genotype | virus, the predominant genotype in
`leading to a completely encircled active site, which includes
`most western countries, are lower than with genotype 2
`catalytic aspartic acids 318 and 319. The dynamics of the
`infections [3, 4]. The low SVR coupled with the frequency
`of
`side
`effects
`associated with
` interferon-ribavirin
`contacts between the fingertips and the thumb have been
`implicated in the mechanism of action of some non-
`combination therapies necessitates
`the development of
`nucleoside inhibitors [12]. The C-terminal 21 amino acids
`improved therapies to treat HCV infection.
`are highly hydrophobic and likely form a membrane anchor
`[13]. Deletion of the C-terminal
`tail gives rise to a more
`soluble enzyme whichis utilized in many in vitro studies
`[14].
`
`HCV is a positive-stranded RNA virus of the family
`Flaviridae, which also includes human pathogens yellow
`fever, West Nile, and dengue viruses. Replication of the
`HCV genomeis catalyzed by a complex of virally-encoded
`and potentially cellular [5] proteins. At
`the heart of the
`replication complex is
`the HCV RNA-dependent RNA
`polymerase (RdRp) which is
`responsible for catalyzing
`ribonucleotide incorporation leading to the formation of both
`
`“Address correspondence to this author at
`Research Laboratories, West Point, PA 19486;
`E-mail: david_olsen@merck.com
`
`the WP26A-3000, Merck
`
`NUCLEOSIDES AS THE BASIS FOR ANTIVIRAL
`THERAPIES
`
`Nucleoside analogs are successfully employed to treat
`infections with HIV, hepatitis B virus, and herpes viruses,all
`of which encode a polymerase whose primary activity is
`DNA synthesis. All of the nucleoside drugs used to treat
`these infections can be considered deoxynucleoside analogs.
`
`1871-5265/06 $50.00+.00
`
`© 2006 Bentham Science Publishers Ltd,
`
`GIL2001
`I-MAK, INC. V GILEAD PHARMASSETLLC
`IPR2018-00120
`
`1
`
`GIL2001
`I-MAK, INC. V GILEAD PHARMASSET LLC
`IPR2018-00120
`
`
`
`18
`
`Infectious Disorders - Drug Targets 1006, Vol. 6, No. 1
`
`Carroll and Olsen
`
`HCV RdRp as an RNA polymerase might be expected to
`exhibit different
`structural
`requirements
`for nucleoside
`analogs that would function as inhibitors, particularly at the
`2’-position of the ribose where inclusion of a hydroxyl
`would likely be advantageous for recognition by RdRp. The
`presence of a 2’-hydroxyl group on the nucleoside offers a
`chance for greater selectivity of inhibition of the HCV RdRp
`over cellular DNA polymerases. Inhibition of the mitochon-
`drial DNA polymerase gamma by deoxynucleoside analog
`triphosphates is thought to be the cause of mitochondrial
`toxicity associated with the administration of some nucleo-
`side analogs,
`for example, ddC [15, 16]. Mitochondrial
`toxicity is likely the basis for the hyperlactatemia,
`lactic
`acidosis,
`steatosis,
`peripheral
`neuropathy, myopathy,
`cytopenias, pancreatitis, and lipoatrophy that are associated
`with long term use of nucleoside inhibitors of HIV reverse
`transcriptase [17]. However the use of a ribonucleoside
`analog could also open the doorto toxicities associated with
`interfering with one or more of the many other roles that
`ribonucleosides play in vivo — in inter- or
`intracellular
`signaling, energy storage and use, protein modification, or
`cellular RNA synthesis.
`
`chain-
`effective
`an
`for
`requirements
`overall
`The
`terminating nucleoside analog inhibitor for oral treatment of
`HCY infection are numerous. First,
`the analog must be
`absorbed efficiently from the gastrointestinal tract. Nucleo-
`side transporters expressed in the intestinal epithelium that
`are responsible for the uptake of most nucleoside analogs
`
`In contrast, passive diffusion
`have been described [18].
`appears to be important for acyclovir. Prodrug approaches to
`improving the oral bioavailability of nucleosides have made
`use of amino acid ester modifications,
`for example
`valacyclovir,
`that
`appear
`to
`allow transport
`by the
`oligopeptide transporter pepT! and others [18, 19].
`
`Additionally, the analog must have acceptable pharmaco-
`kinetic (PK) parameters, have a long half-life sufficient
`preferably for once or twice a day dosing, and achieve
`significant concentrations in the target organ, liver.
`In this
`sense standard plasma PK analysis,
`though useful
`in
`determining the overall exposure to a compound, can
`actually be misleading. A compound that can achieve
`excellent uptake into the liver may appear to have poor
`exposure in plasma in a short-term analysis, since little
`remains in the circulation. In other words, a high first-pass
`uptake into liver is desirable in this case. Balancing the need
`for high concentrations of the nucleoside in liver is the
`general requirement
`that since drug exposure is typically
`monitored in plasma in the clinical setting, some plasma
`exposure is also required. Optimally, the ratio of compound
`concentration in liver to that
`in plasma will be constant
`across time and species, so that the level of compound in
`plasma will be predictive of liver concentrations when the
`compoundis dosed in human subjects.
`
`Once it has been absorbed into blood the analog must
`enter hepatocytes,
`either
`through nucleoside or other
`transporters or by passive diffusion, as shown in Fig (1). The
`
`B
`
`oO=
`
`“HOH “.
`, HO
`
`S
`
`9
`
`8
`
`Base Modifications
`
`~.
`
`B'
`tOH
`
`B
`
`S
`X
`
`aiQOH
`
`=
`
`:
`x
`
`G
`
`X
`
`?
`
`ss
`
`&
`2
`oi?
`“ho
`41QH
`
`=
`OH
`
`0/0,
`Oa’
`YO oO
`HO” ‘y-
`
`HO.
`
`4.
`N%
`hy
`RNR \ %
`+
`%
`Ne
`Oy Pp
`P,
`Oo ‘o
`ce
`0.0 ©
`oA
`o oO
`
`B
`
`wi
`
`y
`=
`x
`
`Fig. (1). Intracellular metabolism of nucleoside analogs. A ribonucleoside analog with a modified ribose enters a cell via nucleoside
`transporters, passive diffusion, or peptide transporters, In the intracellular environment, the nucleoside can either undergo 5’-phosphorylation
`to the monophosphate by the action of a nucleoside kinase, conversion to a base modified form (B’) via nucleoside metabolizing enzymes
`such as adenosine deaminase, or “base-swapping” via purine nucleoside phosphorylase which reversibly catalyzes the cleavage of the
`glycosidic bond. The monophosphate is converted to the diphosphate via nucleotidyl kinase. The diphosphate may be converted to a
`deoxyribose diphosphate by ribonucleotide reductase, which may thenbe incorporated into cellular DNA. Alternately the disphosphate may
`be converted to the triphosphate by nucleoside diphosphate kinase.
`
`2
`
`
`
`Nucleoside Analog Inhibitors ofHepatitis C Virus Replication
`
`Infectious Disorders - Drug Targets 2006, Vol. 6, No. I
`
`19
`
`DISCOVERY OF—2’-C-METHYL
`
`MODIFIED
`roles of equilibrative and concentrative transporters in the
`ADENOSINE ANALOGS
`uptake and efflux of nucleoside analogs from cells have been
`recently reviewed [20]. The analog must
`then serve as a
`substrate for three intracellular kinases to be convertedto the
`active
`5’-triphosphate,
`in general without undergoing
`unwanted metabolism through
`the
`activity of other
`nucleos(t)ide metabolizing enzymes. Modification of an
`adenine base by adenosine deaminase, for example, would
`form the inosine analog, which may not be inhibitory to
`polymerase activity. Cleavage of
`the glycosidic bond
`through the activity of purine nucleoside phosphorylase
`(PNP) can result in scrambling of a modified ribose with
`other nucleobases. In particular, conversion of the diphos-
`phate to a deoxynucleoside diphosphate via ribonucleotide
`reductase is undesirable for nucleobase-modified analogs,
`since that would allow for possible incorporation into
`cellular nuclear DNA which may lead to mutagenesis and
`carcinogenesis. For nucleoside analogs intended to treat
`hepatitis, an important consideration is the expression pattern
`of transporters and the nucleoside metabolizing enzymes
`specific to hepatic tissue which is characterized by high
`catabolic rates [21].
`
`In the past, the inability to propagate HCV robustly in
`cell culture necessitated the use of surrogate assays to screen
`for inhibitors of viral replication. Bovine viral diarrhea virus
`(BVDV)is a pestivirus within the family Flaviviridae, with
`homology to HCV that has been used as a surrogate, sinceit
`is easily propagated in cultures of MDBKcells, and does not
`infect humans. The use of BVDV antiviral assays for
`investigating inhibitors has primarily been replaced with
`assays based on the HCV replicon [27]. Recently, cell
`culture systems capable of propagating HCV have been
`developed [28-30]
`that will
`likely replace the replicon
`system for routine compoundscreening.
`
`Once the 5’-triphosphate of the analog is generated,it
`must serve specifically as a substrate for the HCV RdRp and
`not as a substrate for cellular RNA polymerases and be
`incorporated into HCV genomic RNA. Incorporation of the
`nucleotide analog must take place against a background of
`competing natural nucleoside triphosphates that are present
`at concentrations that range from hundreds of micromolar for
`CTP, UTP and GTP to millimolar concentrations for
`intracellular ATP [22].
`In contrast
`the concentrations of
`deoxynucleoside triphosphates are generally 10-fold lower.
`Thus, effective inhibition through incorporation of a chain
`terminator may be more difficult
`to
`achieve
`in
`the
`intracellular environment with an RdRp than with a DNA
`polymerase, Once incorporated, the analog should no longer
`support further RNA synthesis by preventing the addition of
`more nucleotides,
`In the absence of a mechanism for
`excision of the incorporated chain terminator, the truncated
`viral genome would be unable to support further rounds of
`RNA synthesis. Although
`it
`has
`not
`been
`directly
`demonstrated, since HCV RdRp catalyzes the incorporation
`of ribonucleotides with concomitant release of pyrophos-
`phate,
`it must
`also
`catalyze
`the
`reverse
`reaction,
`pyrophosphorolysis, It remains to be seen whether such an
`excision mechanism operates
`in cells and under what
`conditions. These multiple
`requirements
`for
`effective
`nucleoside analog inhibitors present a real challenge to drug
`development.
`
`Despite these challenges, several different nucleoside
`analogs that
`inhibit HCV replication as measured in the
`replicon assay have been discovered. The majority of the
`published information is available for the class of nucleoside
`analogs
`containing a
`2°-C-methyl
`substituent. Recent
`reviews of progress in the discovery and development of
`other classes of inhibitors of HCV replication are available
`[23-26]. The proposed mechanism of ribavirin notwith-
`standing, this review will focus on novel nucleoside analogs
`that inhibit HCV RdRp.
`
`level of homology between the RNA
`The overall
`polymerases of HCV and BVDV, even when ignoring an N-
`terminal domain of BVDV RdRpthat is not found in the
`HCV enzyme, is quite low. However, within the active sites
`of the polymerases,
`the homology is much stronger,
`suggesting that BVDV might be a useful surrogate for HCV
`for discovery of active-site directed inhibitors, such as
`nucleoside analogs. Confounding this approach,
`though,
`would be the use of MDBKcells to propagate BVDV which
`might have different capabilities for converting nucleoside
`analogs to the active 5’-triphosphates than human hepatic or
`hepatoma cells [31]. The low overall homology between
`HCV and BVDV RdRpsuggests that searches for non-active
`site directed compounds analogous to the non-nucleoside
`reverse
`transcriptase inhibitors of HIV might not be
`successful. This lack of correspondence has been borne out
`experimentally by the
`identification of non-nucleoside
`inhibitors (NNIs) that are BVDV- or HCV-specific [10, 32-
`37).
`
`identified as potent
`2’-C-Methyl nucleosides were
`inhibitors of BVDV replication in cell culture by screening
`libraries of nucleoside analogs
`(R. LaFemina, personal
`communication; [38]). Subsequently it was shown that the
`5’-triphosphates of 2’-C-Me-nucleosides can inhibit
`the
`catalytic activity of both BVDV and HCV RdRpinvitro. In
`parallel
`to screening nucleosides in BVDV antiviral cell
`culture assays, chemically synthesized nucleoside triphos-
`phates were screened for inhibition of purified HCV RdRp.
`The in vitro screens for HCV RdRp inhibitors aided in the
`understanding of some aspects of the structure-activity
`relationships for inhibition of the target enzyme.
`
`STRUCTURE-ACTIVITY RELATIONSHIPS OF 2’-
`MODIFIED NUCLEOSIDE ANALOGS
`
`Structural modifications to the ribose and nucleobase
`have delineated some aspects of the requirements
`for
`efficient inhibition in both the RdRp and cell-based replicon
`assay. In general the results indicate that a very narrow range
`of substituents gives rise to potent inhibition, particularly in
`the replicon assay, owing in large part
`to the multiple
`structural requirements for efficient uptake of the nucleoside
`into the cell, conversion to the 5’-triphosphate, and the
`absence of unwanted metabolic conversion to inactive
`analogs,
`that are necessary in order to inhibit viral RNA
`replication in the cellular environment.
`
`3
`
`
`
`20
`
`Infectious Disorders - Drug Targets 1006, Vol. 6, No, 1
`
`Modifications to the ribose have investigated the size,
`hydrophobicity,
`electronics,
`and
` regiospecificity
`of
`substituents, with the focus on modifications at the 2’ and 3’
`positions, as shown in Fig. (2). Conversion of the 2’-C-
`methyl substituent of the adenosine analog (1) to 2’-C-ethyl
`(2) completely abolishes inhibition by the nucleoside in the
`replicon assay, and by the corresponding 5’-triphosphate in
`the enzyme assay [39]. Modeling of the ethyl substituted
`analog into the enzyme active site in the position that
`correspondsto the substrate nucleosidetriphosphate suggests
`that where
`the
`2’-C-methyl modification
`can
`be
`accommodated into the enzyme active site, the larger ethyl
`modification would clash sterically with the side chain of
`Ser282, analogously to the steric clash between the 2’-C-
`methyl substituent with the methyl group of the resistance
`mutation Ser282Thr, as discussed below.
`
`The importance of the stereoelectronic character of the
`2’-C-methyl to inhibition was investigated by conversion of
`the methyl group to either CH»F (3) or to CF ; (4) [39]. In the
`case of CHF, a 16-fold reduction in inhibitory potency in
`the
`replicon assay was observed, whereas
`the CF,
`substitution led to abolishment of cell-based activity. The
`stereospecificity for inhibition was examined by inverting
`
`Carroll and Olsen
`
`the configuration at the 2°-position. The inverted analog (5)
`was not inhibitory either as the triphosphate in the enzyme
`assay or as the nucleoside in the replicon assay, suggesting
`the importance of maintaining the ribo configuration at the
`2’-carbon [39]. The critical
`importance of retaining the 3’-
`hydroxyl to inhibitory potency is demonstrated by the lack of
`activity of the 3°-deoxy-2’-C-methyladenosine analog (6) in
`both the enzyme and replicon assays. Though HCV RdRp
`does not make use of the 3’-hydroxyl of 2’-C-methy]
`nucleotides as a nucleophile during extension as discussed
`below, the 3’-hydroxyl likely serves as an important binding
`determinant for the initial incorporation event that leads to
`formation of the analog-terminated primer. Moving the
`methyl substituent from the 2’-carbon to the 3’-carbon (7)
`also leads to complete loss of inhibitory potency [39].
`
`Exploration of the structure-activity relationships of the
`nucleobase reveals modifications that enhance inhibitory
`potency. 2’-C-Methylcytidine (8) and 2’-C-methylguanosine
`(9) were both active in the replicon assay though with
`reduced
`potency
`relative
`to
`2’-C-methyladenosine,
`suggesting that
`the identity of the base is not critical
`to
`inhibition [40]. The reduced inhibitory potency relative to
`the adenosine analog is a consequence of deleterious
`
`
`
`és,werj=
`
`v
`
`%
`In
`OH NS
`HO
`1, replicon ECs) = 0.25 uM
`enzyme ICsq = 1.9 uM
`
`pe
`
`Hp
`
`NH:
`
`Ho(Y ven nia -_
`
`.
`¢
`OH
`HS
`2, replicon ECs9> 50M
`enzyme ICc, > 50 uM
`
`—~
`
`Z
`Soy
`HO
`3, replicon ECs) = 4 uM
`enzyme ICog = 4.5uM
`
`Hp
`
`—N
`
`IT
`Ee NS,
`OH
`
`2
`
`5, replicon ECsq > 50 uM
`enzyme IC5q > 50 uM
`
`6, replicon EC59 > 50 uM
`enzyme IC;, > 50M
`
`=
`,
`i=
`wry © AONwih
`SY
`
`N
`
`—
`-
`CF
`:
`NOY
`Dy
`HO
`4, replicon ECsp > 50 uM
`
`~
`om _ |
`a
`HS
`OH
`7, replicon ECs9 > 50 uM
`enzyme ICs59 > 50 uM
`
`~~.
`
`0
`
`NH: I
`3
`CH
`
`‘
`:
`OH
`HO
`8, repiconECso =5 uM
`enzymelC.;, = 0.2 uM
`qd
`
`H,
`
`HO
`
`<
`
`NH2
`replicon ECsy = 2.7 1M
`enzyme ICs = 0.15 uM
`
`9,
`
`
`
`Ho
`’
`HO
`
`ses
`
`=
`ii a4
`3
`Mas NNN
`
`Ae”\_Lcu,)_
`
`|
`
`sy ~
`HO
`OH
`
`10,
`
`replicon ECs) = 0.25 uM
`enzyme |Cg, =0.12 uM
`
`41,R,=Me repiconEC., =29 4M
`12, Ry = C(O)NH2 cytotoxic
`13,Ri1= CN cytotoxic
`
`14,R,=Cl cytotoxic
`15, Ro=Br
`cytotoxic
`16,R,=F replicon EC.) = 0.07 nM
`
`Fig. (2). Structure activity relationships for inhibition of HCV RdRp by nucleoside analogs.
`
`4
`
`
`
`Nucleoside Analog Inhibitors of Hepatitis C Virus Replication
`
`Infectious Disorders - Drug Targets 2006, Vol. 6, No.1
`
`21
`
`metabolism to the inactive uridine analog in the case of 2’-C-
`methylcytidine, and a reduction in cellular uptake and/or
`efficiency of kinasing in the case of the guanosine analog.
`One of the most useful modifications of the base that has
`been identified is the 7-deaza-purine substitution. The 7-
`deaza modification was
`first
`identified as
`a potency-
`enhancing substitution within the context of the enzyme
`assay. By comparison of ICs) values for inhibition in the
`enzyme assay for pairs of nucleoside triphosphates with
`either nitrogen or carbon at the 7-position of either adenine
`or guanine-containing analogs and with several different
`ribose modifications, a consistent 10-20-fold increase in
`inhibitory potency was observed for the 7-deaza version,
`demonstrating that the 7-deaza modification is beneficial to
`inhibitory potency independent of the nature of the ribose
`modification [41]. In fact, the substrate efficiency of ATP
`compared to that of 7-deaza-ATP (tubercidin triphosphate)
`also demonstrated a 10-fold improvement for the 7-deaza-
`modified
`triphosphate,
`suggesting
`that
`the
`improved
`recognition by the enzyme for the 7-deaza substitution does
`not require any ribose modification (Carroll, S., unpublished
`observation).
`
`The physical basis for the improved recognition of the 7-
`deaza substitution by HCV RdRpis not clear but could be a
`result of changes in glycosidic bond angle or length, or
`changes in the electronic character of the purine base, or
`possibly suggestive of a direct
`interaction between a
`hydrophobic region of the RdRp active site and the 7-
`position of the purine base. A numberofsubstituents at the
`7-deaza position are tolerated by HCV RdRp, including 7-
`methyl (11), 7-carboxamide (sangivamycin analog, 12), 7-
`cyano (toyocamycin analog, 13) with little change in
`inhibitory potency in the enzyme assay. However, in general,
`substitutions at the 7-position give rise to cytotoxicity in the
`cell-based assay and are therefore not useful [40, 42]. A
`trend towards decreasing cytotoxicity was noted for the 7-
`halogen substituted series, 7-Br > 7-C] > 7-F (compounds
`14-16). The 7-F-7deaza-2’-C-methyladenosine analog was
`found to be essentially non-cytotoxic by MTS assay in the
`replicon cells at 100 tM, but
`to have the most potent
`replicon ECs) measured for a nucleoside analog (EC 59 = 0.07
`uM) [40]. Further investigations of 16 are underway.
`
`The 7-deaza-adenosine modification is beneficial not just
`in terms of target enzyme potency but also in terms of
`stabilization of the analog to unwanted modification by
`adenosine-metabolizing enzymes. Most notably, the 7-deaza-
`modification essentially eliminates the ability of adenosine
`deaminase (ADA) to convert the 2’-C-methyladenosine to
`the corresponding inosine analog which is inactive in the
`either the replicon or enzyme assays. Under conditions
`where 2’-C-methyl adenosine is quantitatively converted to
`2’-C-methylinosine in in vitro ADA-catalyzed reactions, 2’-
`C-methy!-7-deazaadenosine is left
`intact [40]. The lack of
`conversion of the 7-deaza-adenosine analog is related to the
`mechanism of catalysis by ADA, wherein the 7-nitrogen is
`protonated to aid in the delocalization of charge that occurs
`during formation of a tetrahedral intermediate that results in
`displacementof the 6-amino group by hydroxyl anion [43].
`
`purine
`the
`phosphorolysis of
`the
`PNP catalyzes
`glycosidic bond, resulting in removal of the modified ribose,
`
`and allowing for subsequent metabolism of the ribose. The
`2’-C-methyl substitution itself reduces the ability of PNP to
`catalyze glycosidic bond breaking [39]. However,
`the 7-
`deaza modification further stabilizes the glycosidic bond to
`PNP-catalyzed breakdown, further increasing the metabolic
`stability of 7-deaza-substituted nucleoside analogs
`[40].
`These results demonstrating increased stability to adenosine-
`metabolizing enzymesin vitro correlate well with increased
`pharmacokinetic stability in vivo as discussed below.
`
`MECHANISM OF CHAIN TERMINATION BY 2?-C-
`METHYL NUCLEOTIDES
`
`The physical mechanism of the inhibition of RdRp
`activity by 2’-C-methyl-modified nucleotides appears to
`involve disruption of the growth of the RNA strand after
`incorporation of
`the modified
`nucleotide. Typically
`nucleoside analogs used as antiviral therapeutics act as chain
`terminators due to the absence of the 3’-hydroxyl group that
`functions as the nucleophile during extension of the DNA
`chain by incorporation of a nucleotide. Thus, analogs such as
`AZT, 3TC, ddC, d4T and the active forms of didanosine, and
`abacavir are incorporated onto the 3’-end of a growing
`primer strand but are unable to support further elongation of
`the growing strand. The family of 2’-C-methyl modified
`nucleotides appears to act via a similar mechanism despite
`the fact
`that
`the 3’-hydroxyl
`is still present
`in the 2°-C-
`methyl
`nucleotides. The 2’-C-methyl nucleotides
`are
`therefore non-obligate chain terminators.
`
`the
`supporting chain termination as
`evidence
`The
`mechanism of inhibition by 2’-C-methyl nucleotides comes
`from analysis of polyacrylamide gel electrophoresis — based
`nucleotide incorporation experiments. Hairpin RNAs formed
`from synthetic oligoribonucleosides have been found to
`serve as efficient templates for RNA synthesis catalyzed by
`HCV RdRp[44]. An example of the designed structure of an
`RNAhairpin template is shown in Fig. (3), which has been
`labeled with **P at its 5’-end via polynucleotide kinase. The
`RNAtemplate and products resulting from the incorporation
`of nucleotides or nucleotide analogs are then separated via
`gel electrophoresis and visualized via Phosphorlmaging. In
`the case of this RNA template sequence, the first nucleotide
`to be incorporated would be an adenosine or adenosine
`analog, which would be followed by incorporation of a
`uridine or uridine analog, as shown in the control reactions
`on the left side of the gel. HCV RdRp is capable of
`incorporating 3°-deoxy-adenosine monophosphate, but
`is
`incapable of extending the 3’-deoxy-adenosine monophos-
`phate once it has been incorporated, as expected for this
`obligate chain terminating analog, as shownin lanes 6 and 7.
`HCV RdRp is also capable of incorporating 7-deaza-
`adenosine monophosphate (tubercidin monophosphate) into
`this RNA hairpin, and is also capable of extending the
`incorporated 7-deaza-adenosine by addition of the next
`nucleotide, as shownin lanes 8 and 9. However even though
`HCV RdRp can incorporate either 2’-C-methyladenosine or
`2’-C-methyl-7-deaza-adenosine into the RNA template, the
`enzyme is incapable of adding the next nucleotide onto the
`end of the analog-terminated template. Thus the 2’-C-methy]
`nucleoside analogs act as chain terminators,
`
`5
`
`
`
`22
`
`Infectious Disorders - Drug Targets 2006, Vol. 6, No. 1
`
`Carroll and Olsen
`
`U
`
`GG CCC GGU AGA 5’
`CC GGG CC 3”
`
`U U
`
` E
`
`+
`-
`-
`10.
`ATP(nM)
`10 -
`UTP(UM)
`-
`-
`3°dATP (UM)
`-
`-
`7-deuza-ATP (UM)
`-
`-
`2°-C- Me-ATP (UM)
`-
`-
`7-deaza-2°-C- Mc-ATP (uM)
`Fig. (3). Gel based incorporation/extension assays. The hairpin RNA oligoribonucleotide with the sequence shown wasradiolabeledat the
`5’-end with *P-phosphate via polynucleotide kinase. Reactions were catalyzed by purified HCV RdRp in the presence of the indicated
`concentrations of nucleoside triphosphates. Reactions were quenched by addition of loading buffer and reaction products were separated on
`20% polyacrylamide-8 M urea sequencing gels. Radiolabeled products were visualized with Phosphorlmaging. The sequence was chosen so
`that the first nucleotide to be incorporated onto the 3’-end of the hairpin is AMP (third lane of the gel from the left), followed by
`incorporation of UMP ca" lane), 3’-deoxy-AMPis incorporated in place of AMP ot lane), but the product of incorporation of 3’-dAMP
`cannot be extended by addition of UTP. Tubercidin triphosphate serves as a substrate for HCV RdRp, and can beincorporated in place of
`AMP andthe product of the incorporation can be extended by addition of UMP (lanes 7 and 8). Both 2’~C-Me-ATP and 2’-C-Me-7-deaza-
`ATP can be incorporated onto the end of the RNA,but neither is efficiently extended by addition of UTP. A trace of extended product is
`evident with 2’-C-Me-ATPthatis absent with the 7-deaza analog.
`
`1010 -
`10
`
`-
`10
`
`The physical basis for the chain termination mechanism
`is unknown but a potential explanation is illustrated in the
`model shown in Fig. (4). This model was constructed using
`the initiation complex from a co-crystal of the bacteriophage
`phi-6 RdRp[45], consisting of two nucleoside triphosphates
`hydrogen bonded to the 3’-end of a single stranded RNA
`template bound to the phi-6 RdRp. One of the nucleoside
`triphosphates corresponds to the nucleophile or primer 3’-
`terminal nucleotide (the one on the left in Figure (4)), and
`the
`other
`corresponds
`to
`a
`“substrate”
`nucleoside
`triphosphate, the pyrophosphate of which is displaced during
`catalysis. The initiation complex was copied into the HCV
`RdRp active site [8]. For
`illustrative purposes both the
`nucleophilic nucleoside triphosphate and the
`substrate
`nucleoside triphosphate are shown as 2’-C-methyl-adenosine
`triphosphates in the Figure. In the case of chain termination
`the RdRp is capable of incorporating the 2’-C-methy]
`nucleotide analog. We suppose that the enzyme can then
`translocate forward on the template strand such that
`the
`nucleotide
`analog now occupies
`the primer-terminus
`position, corresponding to the nucleophile during catalysis.
`When the substrate nucleoside triphosphate then binds (as
`shown in Fig.
`(4)
`the substrate is another 2’-C-methyl
`adenosine triphosphate, but
`it could be any nucleoside
`triphosphate)
`the 2’-C-methyl group of the incorporated
`analog is now in proximity to both the ribose ring oxygen,
`and the 8-position of the nucleobase of the incoming
`
`nucleoside triphosphate, as represented by the yellow and
`red crosshatching. Thus the methyl group may provide a
`steric clash that either prevents binding or prevents optimal
`alignment of the substrate nucleoside triphosphate to inhibit
`efficient nucleophilic attack of the 3’-hydroxyl on the alpha-
`phosphorous, thus preventing chain extension of the 2’-C-
`methyl! analog-terminated primer,
`
`RESISTANCE TO INHIBITION BY 2’-C-METHYL
`NUCLEOSIDES
`
`Oneof the most convincing lines of evidence to establish
`the intracellular target of inhibition by a given antiviral
`compound is the generation of resistance to inhibition in a
`cell culture setting. Mapping the mutation that creates
`resistance to a target gene, and correlating changes in the
`inhibitory potency of the compound in biochemical assays of
`the mutated enzyme demonstrates the validity of the target.
`Additionally,
`in vitro resistance
`studies
`can provide
`information regarding the likelihood of the appearance of
`resistance
`during
`clinical
`use of
`an_
`investigational
`compound,
`as well
`as
`information on
`the potential
`deleterious effects of the mutation on the replicative capacity
`of the mutatedvirus.
`
`Resistance to 2’-C-methyl nucleosides in the bicistronic
`replicon assay is engendered by a single amino acid change,
`Ser282Thr, within the active site of the RdRp [46]. A 38-fold
`
`6
`
`
`
`23
`
`Nucleoside Analog Inhibitors ofHepatitis C Virus Replication
`
` Infectious Disorders - Drug Targets 2006, Vol. 6, No. 1
`
`Fig. (4). A model ofthe active site of HCV RdRp. The model was constructed using the structure of a de novoinitiation complex structure of
`the phi-6 bacteriophage RdRp [45] containing a bound template RNA coordinated to two nucleoside triphosphates. The protein structure was
`replaced with the crystal structure of the HCV RdRp[8]. In place of the nucleoside triphosphates were modeled two molecules of 2’-C-
`methyl-7-deazaadenosine triphosphate. The crosshatching represents a van der Waals radius of the 2’methyl group of one nucleoside analog
`and the ribose ring oxygen and 8-position of the other. The blue numbers represent distances in Angstroms from the methyl carbon to the
`ting oxygen and 8-positions of the other nucleoside analog. The purple spheres represent Mg” ions required for catalysis.
`
`loss of inhibitory potency was determined for 2’-C-methyl
`adenosine in the resistant replicon. The mutation is quite
`specific for
`the 2’-C-methyl modification as shown by
`essentially no change in inhibitory potency for 2’-O-
`methylcytidine, where the methyl group is now present as an
`ether
`linkage. Likewise a 10-fold loss of potency was
`observed for inhibition by 2’-C-methyl-guanosine with the
`Ser282Thr
`replicon, suggesting that
`the identity of the
`nucleobase was less important for resistance than the methyl
`modification of the ribose. The RdRp containing the
`Ser