Poliovirus RNA-dependent RNA polymerase (3D(pol)). Assembly of stable, elongation-competent complexes by using a symmetrical primer-template substrate (sym/sub).

Detailed studies of the kinetics and mechanism of nucleotide incorporation catalyzed by the RNA-dependent RNA polymerase from poliovirus, 3D(pol), have been limited by the inability to assemble elongation complexes that permit activity to be monitored by extension of end-labeled primers. We have solved this problem by employing a short, symmetrical, heteropolymeric RNA primer-template that we refer to as "sym/sub." Formation of 3D(pol)-sym/sub complexes is slow owing to a slow rate of association (0.1 microM(-1) s(-1)) of 3D(pol) and sym/sub and a slow isomerization (0. 076 s(-1)) of the 3D(pol)-sym/sub complex that is a prerequisite for catalytic competence of this complex. Complex assembly is stoichiometric under conditions in which competing reactions, such as enzyme inactivation, are eliminated. Inactivation of 3D(pol) occurs at a maximal rate of 0.051 s(-1) at 22 degrees C in reaction buffer lacking nucleotide. At this temperature, ATP protects 3D(pol) against inactivation with a K(0.5) of 37 microM. Once formed, 3D(pol)-sym/sub elongation complexes are stable (t((1)/(2)) = 2 h at 22 degrees C) and appear to contain only a single polymerase monomer. In the presence of Mg(2+), AMP, 2'-dAMP, and 3'-dAMP are incorporated into sym/sub by 3D(pol) at rates of 72, 0.6, and 1 s(-1), respectively. After incorporation of AMP, 3D(pol)-sym/sub product complexes have a half-life of 8 h at 22 degrees C. The stability of 3D(pol)-sym/sub complexes is temperature-dependent. At 30 degrees C, there is a 2-8-fold decrease in complex stability. Complex dissociation is the rate-limiting step for primer utilization. 3D(pol) dissociates from the end of template at a rate 10-fold faster than from internal positions. The sym/sub system will facilitate mechanistic analysis of 3D(pol) and permit a direct kinetic and thermodynamic comparison of the RNA-dependent RNA polymerase to the other classes of nucleic acid polymerases.

Indeed, inhibitors of viral DNA polymerases and reverse transcriptase (RT) are used clinically to treat infection by herpesviruses and human immunodeficiency virus (1,2). Of course, cellular polymerases are also potential targets for antiviral chemotherapeutics that inhibit viral polymerases. Therefore, target specificity is of central importance to the development of clinically useful compounds. In some cases, drug activation only occurs in virus-infected cells thereby increasing specificity and diminishing toxicity (1). In others, however, infected cellspecific, activation mechanisms are not an option, and other approaches are required (2).
A comprehensive understanding of the differences between viral and cellular polymerases facilitates the design of specific inhibitors. In the case of human immunodeficiency virus RT, a combination of detailed kinetic (3)(4)(5)(6)(7)(8)(9)(10)(11)(12) and structural (13)(14)(15)(16)(17)(18) information was essential to provide an understanding of the molecular basis for RT function and the mechanism of action of non-nucleoside inhibitors of RT (9,10). In fact, a description of the nucleotide addition cycle catalyzed by all classes of polymerases studied to date has benefited from both kinetic and structural analyses (19 -28). The "structural era" for the RdRP has begun with the recent solution of the x-ray crystal structure for the RdRP from poliovirus, 3D pol (29). For this reason, poliovirus 3D pol is the ideal model system to use in the study of RdRP function. Unfortunately, detailed, mechanistic information regarding 3D pol -catalyzed nucleotide incorporation is sparse. The existing gap derives from the inability to assemble stoichiometrically 3D pol -primer-template complexes that permit nucleotide incorporation to be followed by extension of end-labeled primers.
Previous studies of 3D pol -catalyzed nucleotide incorporation that utilized short, heteropolymeric RNA primer-templates as substrates demonstrated that this enzyme binds to primertemplate such that some enzyme has the 3Ј-OH from the "primer" strand in the catalytic center ("correct" orientation) and some enzyme has the 3Ј-OH from the "template" strand in the catalytic center ("incorrect" orientation) (30). When bound in the "incorrect" orientation, 3D pol adds non-templated nucleotides to the blunt end of primer-template (30). Of course, heterogeneous binding of 3D pol to primer-template further complicates mechanistic analysis of this enzyme. In this report, we demonstrate that the use of a symmetrical primer-template substrate (sym/sub) permits the assembly of stable, elongationcompetent 3D pol -primer-template complexes. The sym/sub system will facilitate quantitative analysis of the kinetics and mechanism of 3D pol -catalyzed nucleotide incorporation. 32 P]ATP (Ͼ7,000 Ci/mmol) was from ICN; nucleoside 5Ј-triphosphates, 2Ј-deoxynucleoside 5Ј-triphosphates, and 2Ј,3Ј-deoxynucleoside 5Ј-triphosphates (all nucleotides were ultrapure solutions) were from Amersham Pharmacia Biotech; 3Ј-deoxyadenosine 5Јtriphosphate (cordycepin) was from Sigma; all RNA oligonucleotides were from Dharmacon Research, Inc. (Boulder, CO); T4 polynucleotide kinase was from New England Biolabs, Inc.; All other reagents were of the highest grade available from Sigma or Fisher.

Materials-[␥-
Expression and Purification of 3D pol -Expression and purification of 3D pol were performed as described previously (31).
Purification of sym/sub-The RNA oligonucleotide was purified by denaturing PAGE. Gels consisted of 21.3% acrylamide, 1.7% bisacrylamide, 7 M urea, and 1ϫ TBE (89 mM Tris base, 89 mM boric acid, and 2 mM EDTA). The oligonucleotide ladder was visualized by UV shadowing. A gel slice containing only the full-length oligonucleotide was removed, and the nucleic acid was electroeluted from the gel in 1ϫ TBE by using an Elutrap apparatus (Schleicher & Schuell). Oligonucleotides were desalted on Sep-Pak columns (Millipore) as specified by the manufacturer. Oligonucleotides were hydrolyzed to remove the 2Ј-OH protecting groups as specified by the manufacturer. The concentrated, hydrolyzed RNA stocks (typically 1 mM) contained 330 mM Tris acetate, pH 7.0. Oligonucleotides were diluted in T 10 E 1 (10 mM Tris, 1 mM EDTA, pH 8.0), to the indicated concentrations. Oligonucleotides were stored at Ϫ80°C until use. Concentrations were determined by measuring the absorbance at 260 nm in 50 mM K 2 PO 4 , pH 7.0, 6 M guanidine by using a calculated extinction coefficient of 101,600 M Ϫ1 cm Ϫ1 (32).
5Ј-32 P Labeling of sym/sub-The RNA oligonucleotide was end-labeled by using [␥-32 P]ATP and T4 polynucleotide kinase from New England Biolabs essentially as specified by the manufacturer. Reactions, typically 30 l, contained 11 M [␥-32 P]ATP, 10 M RNA oligonucleotide (sym/sub), 1ϫ Kinase Buffer and 0.4 units/l T4 polynucleotide kinase. Reactions were incubated at 37°C for 60 min. Unincorporated nucleotide was removed by passing the sample over two consecutive 1-ml Sephadex G-25 (Sigma) spun columns. We assumed the concentration of end-labeled RNA remained constant.
Annealing of sym/sub-1 M end-labeled sym/sub was mixed with 99 M unlabeled sym/sub in T 10 E 1 , pH 8.0, and heated to 90°C for 1 min and slowly cooled to 10°C at a rate of approximately 5°C/min in a Progene Thermocycler. 3D pol Assays-Reactions contained 50 mM HEPES, pH 7.5, 10 mM 2-mercaptoethanol, 5 mM MgCl 2 , 60 M ZnCl 2 , 500 M NTP, sym/sub, and 3D pol . Reactions were quenched by addition of EDTA to a final concentration of 50 mM. Specific concentrations of primer-template and 3D pol , along with any deviations from the above, are indicated in the appropriate figure legend. 3D pol was diluted immediately prior to use in 50 mM HEPES, pH 7.5, 10 mM 2-mercaptoethanol, 60 M ZnCl 2 , and 20% glycerol. The volume of enzyme added to any reaction was always less than or equal to one-tenth the total volume.
Rapid Chemical Quench-flow Experiments-Rapid mixing/quenching experiments were performed by using a model RQF-3 chemical quench-flow apparatus (KinTek Corp., State College, PA). Experiments were performed at 30°C by using a circulating water bath. 3D pol -sym/ sub complex in 50 mM HEPES, pH 7.5, 10 mM 2-mercaptoethanol, 60 M ZnCl 2 , and 5 mM MgCl 2 was rapidly mixed with the nucleotide substrate in 50 mM HEPES, pH 7.5, 10 mM 2-mercaptoethanol, 60 M ZnCl 2 , and 5 mM MgCl 2 , and the reactions were quenched by addition of 0.5 M EDTA to a final concentration of 0.3 M.
Product Analysis, Denaturing PAGE-An equal volume of loading buffer, 10 l (90% formamide, 0.025% bromphenol blue, and 0.025% xylene cyanol), was added to 10 l of the quenched reaction mixtures and heated to 70°C for 2-5 min prior to loading 5 l on a denaturing polyacrylamide gel of the appropriate percentage containing 1ϫ TBE and 7 M urea. Highly cross-linked gels contained 1.7% bisacrylamide. Electrophoresis was performed in 1ϫ TBE at 75 watts. Gels were visualized by using a PhosphorImager and quantitated by using the ImageQuant software (Molecular Dynamics). Data were fit by nonlinear regression using the program, KaleidaGraph (Synergy Software, Reading, PA).
Kinetic Simulation-Kinetic simulations were performed by using KinTekSim version 2.03 (KinTek Corp., State College, PA). All rate constants were determined experimentally, except where noted. The agreement between the experimental data and kinetic simulations was determined by visual inspection.

RESULTS
A Symmetrical Primer-Template Substrate (sym/sub) to Study 3D pol -catalyzed Nucleotide Incorporation-We designed the symmetrical, heteropolymeric RNA primer-template substrate shown in Fig. 1A to study the kinetics and mechanism of 3D pol -catalyzed nucleotide incorporation. We refer to this substrate as "sym/sub" throughout. This substrate is a 10-nt, selfcomplementary RNA that forms a 6-base pair duplex flanked by two, 4-nt 5Ј-overhangs. Each overhang is capable of templating incorporation of a unique nucleotide, thus permitting the evaluation of single and multiple cycles of nucleotide incorporation. The calculated T m value is 55°C for this duplex when 1 M RNA strands (0.5 M duplex) is employed (33). Based upon previous studies of 3D pol , this T m value is in the optimal range for maximal rates of nucleotide incorporation (34).
Incorporation of the First Nucleotide into sym/sub by 3D pol Is Slow-Utilization of sym/sub by 3D pol is shown in Fig. 1B. In this experiment, enzyme was in excess (5 M) relative to sym/ sub (0.5 M duplex); ATP (500 M) was supplied as the sole nucleotide substrate; and either Mn 2ϩ or Mg 2ϩ was employed as divalent cation cofactor. In the presence of both divalent cations, AMP incorporation into sym/sub was observed. AMP incorporation was very error-prone in the presence of Mn 2ϩ but very specific in the presence of Mg 2ϩ . Quantitative evaluation of the kinetics of AMP incorporation into sym/sub showed that in the presence of Mn 2ϩ , incorporation was biphasic. During the first phase, 0.52 Ϯ 0.02 M sym/sub was utilized at a rate too fast to measure by quenching the reaction manually. During the second phase, an additional 0.34 Ϯ 0.04 M sym/sub was utilized at a rate of 0.012 Ϯ 0.001 s Ϫ1 . It should be noted that in order to extend greater than 0.5 M RNA strands, the "flip side" of an extended sym/sub duplex must be utilized, thus mandating dissociation of a previously existing 3D pol -sym/sub product complex. In the presence of Mg 2ϩ , AMP incorporation into sym/sub was monophasic; 0.36 Ϯ 0.02 M sym/sub was extended at a rate of 0.041 Ϯ 0.004 s Ϫ1 . The measured value of extended sym/sub (0.36 M) was lower than the theoretical value (0.5 M) for a single, stoichiometric turnover.
Incorporation of the Second Nucleotide into sym/sub by 3D pol Is Fast-If the slow observed rate of nucleotide incorporation in the presence of Mg 2ϩ reflected the true rate of 3D pol -catalyzed nucleotide incorporation, then each subsequent cycle of incorporation should occur at approximately the same rate. However, if the true rate of incorporation was masked by the slow assembly of 3D pol -sym/sub complexes, then each subsequent cycle of nucleotide incorporation may be faster. In order to address this issue, an experiment was performed in which 3D pol , sym/sub, and ATP were incubated at 30°C for 90 s, a time sufficient to accumulate 97% of the total possible 11-mer product ( Fig. 2A). At this point, either buffer was added as a control for misincorporation, or the second nucleotide, UTP, was added ( Fig. 2A). The kinetics of 11-mer conversion to 12-mer product were monitored from 15 s to 5 min. During the 5-min incubation period in the absence of UTP, misincorporation of AMP to form 12-mer product was not observed (Fig. 2B, panel Buffer). However, in less than 15 s, 94% of 11-mer was converted to products 12-14-nt in length in the presence of UTP (Fig. 2B, panel UTP; Fig. 2C). Products of misincorporation represented 6% of the total product at 15 s and increased linearly to 27% of the total product during the 5-min reaction time. These data are consistent with a model in which assembly of reactive ternary complexes limits the rate of single nucleotide incorporation when reactions are initiated by addition of 3D pol .
Assembly of 3D pol -sym/sub Complexes Prior to Addition of the First Nucleotide Results in Fast Incorporation-In order to determine whether a preferred order of assembly existed, the experiment shown in Fig. 3 was performed. In each experiment, the final concentrations of 3D pol (5 M), sym/sub (0.5 M duplex), and ATP (500 M) remained constant. The only difference was in the order of mixing/incubation at 30°C. The results from a reaction initiated by addition of enzyme ( Fig. 3A) were consistent with the analysis presented in Fig. 1 (k obs was 0.06 Ϯ 0.01 s Ϫ1 ; amplitude was 0.38 Ϯ 0.01 M). Incubation of 3D pol with sym/sub for 90 s prior to addition of ATP ( Fig. 3B) completely removed the slow rate of incorporation. However, a 2-fold reduction in the amplitude (0.21 M) was noted. Incubation of 3D pol and ATP prior to addition of sym/sub (Fig. 3C) reduced the observed rate of AMP incorporation by 3-fold (k obs was 0.02 s Ϫ1 ) with very little effect on the end point of the reaction (amplitude was 0.33 M). Incubation of 3D pol at 30°C prior to addition of sym/sub and ATP (Fig. 3D) showed significant thermal inactivation of 3D pol .
Stability of 3D pol -Thermal inactivation of 3D pol will contribute significantly to the outcome of kinetic experiments performed with this enzyme; therefore, a rigorous analysis of this reaction was performed. In the presence of 20% glycerol, 3D pol was stable for at least 1 week at temperatures less than or equal to 4°C (data not shown). When frozen in glycerol-containing buffer at Ϫ80°C, the enzyme was stable indefinitely and lost less than 10% activity after a single cycle of freezing and thawing (data not shown). However, upon dilution of 3D pol A, sym/sub and ATP were incubated at 30°C for 90 s, at which time 3D pol was added to initiate the reaction. Reactions were quenched at the indicated times by addition of EDTA (50 mM). The kinetics of AMP incorporation could be fit to a single exponential with an observed rate of 0.06 s Ϫ1 and an amplitude of 0.38 M. B, 3D pol and sym/sub were incubated at 30°C for 90 s, at which time ATP was added to initiate the reaction. Reactions were quenched at the indicated times by addition of EDTA (50 mM). C, 3D pol and ATP were incubated at 30°C for 90 s, at which time sym/sub was added to initiate the reaction; the reaction was quenched at the indicated times by addition of EDTA (50 mM). The kinetics of AMP incorporation could be fit to a single exponential with an observed rate of 0.02 s Ϫ1 and an amplitude of 0.33 M. D, 3D pol was incubated at 30°C for 90 s, at which time sym/sub and ATP were added to initiate the reaction. Reactions were quenched at the indicated times by addition of EDTA (50 mM). The kinetics of AMP incorporation could be fit to a single exponential with an observed rate of 0.004 s Ϫ1 and an amplitude of 0.18 M.
into our standard reaction buffer which lacks glycerol (see "Experimental Procedures"), 3D pol activity was lost at a finite rate at temperatures greater than or equal to 0°C (Fig. 4A). Consistent with previous results (35), nucleotide, in this case ATP, protected 3D pol against inactivation (Fig. 4B). Protection of 3D pol by ATP was apparent at all temperatures analyzed (Table I) and was saturable ( Fig. 4C and Table II). The maximal rates of inactivation varied over a 20-fold range between 0 and 30°C. Interestingly, the concentration of ATP sufficient to decrease the observed rate of inactivation to 50% of the maxi-mal value (K 0.5(ATP) ) also showed a temperature dependence, ranging from 17 M at 0°C to 84 M at 30°C. In addition to ATP, GTP, GDP, UTP, and CTP were also capable of protecting 3D pol against thermal inactivation (Fig. 5). As observed previously, GTP protects 3D pol against thermal inactivation better than any other nucleotide (35). We also noted a linear, 3-fold decrease in k inact(max) when the concentration of 3D pol was varied between 2 and 20 M (data not shown).
Stability of 3D pol -sym/sub Complexes-Although the rate of formation of the initial 3D pol -sym/sub complex was slow, this complex was stable enough to incorporate quantitatively the second nucleotide. In order to determine directly the rate constant for dissociation of this complex, we performed the experiment that is illustrated in Fig. 6A. 3D pol and 5Ј-32 P-labeled sym/sub were given 90 s to assemble, at which point a 1000-fold molar excess of unlabeled sym/sub was added as a trap for free enzyme. At various times after addition of trap, the amount of complex remaining was determined by following production of 11-mer (Fig. 6B). At 30°C, the 3D pol -sym/sub complex had a half-life of ϳ20 min, whereas at 22°C, this complex had a half-life of ϳ2 h (Fig. 6C). A similar experiment was performed to determine the stability of the 3D pol -sym/sub product complex (Fig. 6D); however, in this case, the ability to form 12-mer was monitored (Fig. 6E). Again, there was a temperature dependence to the stability of this complex (Fig. 6F). The product complexes had half-lives of 4.5 and 8.5 h at 30 and 22°C, respectively.
Kinetics of Ribo-and Deoxyribonucleotide Incorporation by 3D pol -By using a rapid mixing/quenching device, we were able to evaluate the kinetics of AMP incorporation into sym/sub from preassembled 3D pol -sym/sub complexes (Fig. 7). In the presence of 600 M ATP, incorporation of AMP was fast, occurring at a rate of ϳ70 s Ϫ1 . Under the same conditions, both 2Ј-dAMP and 3Ј-dAMP were also incorporated, albeit at rates 70 -120-fold slower than for AMP (Table III). The loss of each hydroxyl appeared to be additive as incorporation of 2Ј,3Ј-ddAMP was reduced by 35,000-fold relative to AMP incorporation (Table III).
Steady-state kinetic analysis of AMP incorporation into sym/ sub showed a burst of product formation followed by a linear accumulation of product (Fig. 8). The steady-state rate constant, k cat , for AMP incorporation was 0.0001 s Ϫ1 , and the y intercept was 0.8 M. In the presence of all four NTPs, the k cat value was increased by 10-fold relative to that in the presence of ATP only (Fig. 8). There was not a significant difference in the y intercept (Fig. 8). The k cat value for AMP incorporation did not change by increasing the concentration of 3D pol to 2.5    M; the y intercept was 2.0 M (data not shown).
Kinetic Mechanism for Assembly of Stable, Elongation-competent 3D pol -sym/sub Complexes-In order to understand better the kinetic mechanism for assembly of 3D pol -sym/sub complexes, we evaluated the kinetics of assembly indirectly by monitoring AMP incorporation into sym/sub under conditions in which either the concentration of 3D pol or sym/sub was varied. The results of this analysis are presented in Fig. 9. The data were most consistent with the mechanism presented in Scheme 1. The validity of this model was tested by kinetic simulation of this mechanism using the parameters shown in Table IV. The simulated data are represented by the solid lines through the data shown in Fig. 9. The concentration of active 3D pol (0.8 M) was from the y intercept of the steady-state kinetic data (Fig. 8). A K d value for ATP of 100 M was employed. 2 This value is consistent with reported K m values for nucleotides (30). This K d value was inconsequential to the outcome of the simulation as this value could be varied by an order of magnitude in either direction without any effect on the simulated kinetics of assembly. The rate constant for the conformational change was constrained by the observed rate of assembly at 5 and 10 M sym/sub. A value of 0.076 s Ϫ1 was obtained from a single exponential fit of these individual time courses. Therefore, the only simulated parameters were k ϩ1 and k -1 . Given the constraints discussed above, values of 0.1 M Ϫ1 s Ϫ1 and 0.1 s Ϫ1 for k ϩ1 and k -1 , respectively, were shown to fit the data best. The simulated data were quite sensitive to these rate constants as a 25% change in either altered the goodness of fit to the data. A good fit of the data acquired by varying 3D pol concentration was obtained only when the fraction of inactive enzyme (20% of the total upon initiation) was allowed to bind to sym/sub. This unproductive complex was kinetically significant at all concentrations of 3D pol , even more so at low concentrations of sym/sub (Յ1 M). At 30°C in the presence of sym/sub and absence of glycerol, 3D pol concentrations in the 10 -25 M range showed evidence of precipitation; therefore, these data were not included in the analysis. Given this problem with precipitation, we did not force a fit of the mechanism to the 5 M 3D pol data (Fig. 9A). It should be noted, however, that at 5 M 3D pol we were unable to observe any precipitation even after centrifugation.

DISCUSSION -
We have used a symmetrical primer-template substrate (sym/sub) to assemble stable, elongation-competent 3D pol -sym/ sub complexes suitable for analysis of the kinetics and mecha-nism of nucleotide incorporation catalyzed by this enzyme. Our current model for the mechanism of 3D pol -sym/sub complex assembly is presented in Scheme 1. 3D pol undergoes iterative cycles of binding and release of sym/sub until the complex isomerizes into a catalytically competent elongation complex. After each cycle of dissociation, enzyme inactivation can occur. 3D pol inactivation results in substoichiometric assembly of elongation complexes. The fraction of enzyme that will be inactivated during the assembly process depends upon the partitioning between the two pathways. Maximizing the rate of 3D pol -sym/sub complex formation and minimizing the rate of 3D pol inactivation will permit stoichiometric assembly of elongation complexes. By using concentrations of sym/sub in a range greater than or equal to 10 M, the rate of assembly will be maximized (Fig. 9B). By assembling complexes at lower temperatures in the presence of nucleotide, the rate of 3D pol inactivation will be minimized (Fig. 4). It may also be possible to assemble complexes efficiently by using Mn 2ϩ as the divalent cation cofactor and then diluting these mixtures into Mg 2ϩ -containing buffer to prevent error-prone RNA synthesis ( Fig. 1 and Ref. 30). Once complexes form, they are very stable (Fig. 6A), support a fast rate of nucleotide incorporation (Fig.  7), and produce complexes that are significantly more stable than the initial complex (Fig. 6B). The rate of complex disassembly limits the rate of primer utilization (Fig. 8). The rate of complex disassembly is faster from the end of template than from internal positions (Fig. 8). Therefore, the number of primers utilized during the course of a reaction can be increased by following incorporation of all four nucleotides instead of a single nucleotide.
We were also able to fit the data by using an alternative model with a slow association step without a conformational change step. In this case, the rate constant for association was 0.035 M Ϫ1 s Ϫ1 , and the rate constant for dissociation was that measured directly, 0.0006 s Ϫ1 (Fig. 6). However, in order to obtain a good fit, it was necessary to reduce artificially the concentration of sym/sub employed from 0.5 to 0.4 M. We observe stoichiometric (0.5 M) utilization of sym/sub in the presence of Mn 2ϩ (Fig. 1C). This observation confirms that all of the substrate is competent for primer extension. Kinetic simulation should always be constrained by parameters determined empirically. Therefore, this model was not considered further. Attempts were also made to add a monomer-oligomer equilibrium to the mechanism in order to fit the kinetics of assembly in the presence of excess (5 M) 3D pol (Fig. 9A). These attempts failed owing to the lack of empirical data (and corresponding constraints) to describe this reaction. Direct analysis of 3D pol oligomerization should provide information necessary to explore this possibility more rigorously.
Whether the mechanism of assembly described herein can be related to initiation of transcription or replication in vivo remains to be determined. These studies employ an artificial primer-template; therefore, the kinetic scheme and associated parameters may only be relevant for the model system. The kinetic properties may be different for a reaction of the polymerase with a natural template in the presence of other protein factors. Replication of the poliovirus genome initiates within the poly(A)-tract at the 3Ј-end of genomic RNA from a complex comprising viral factors 3AB, 3B (VPg), 3CD, and 3D pol (Fig.  10) (36). Once 3D pol initiates RNA synthesis from the protein primer, an elongation-competent complex is formed. As indicated in Fig. 10, a single template is thought to support multiple, continuous cycles of initiation. The nascent RNA chain is only annealed to template over a short distance, and the RNA is spooled through a channel formed by the polymerase (37). Whether or not a helicase is required to disrupt base pairing 2 J. J. Arnold and C. E. Cameron, unpublished results. between template and nascent RNA is not known.
We hypothesize that 3D pol -sym/sub complexes mimic complexes formed during elongation of poliovirus plus-and minusstrand RNAs by 3D pol (Fig. 10). In support of this hypothesis is the observation that a 3D pol derivative with an amino acid substitution in conserved structural motif B that has less than a 2-fold reduction in poly(rU) polymerase activity but a 15-fold reduction in the observed rate of elongation measured by using the sym/sub system produces virus with significantly impaired growth properties relative to wild-type virus. 3 Moreover, the rate of nucleotide incorporation into sym/sub and stability of the 3D pol -sym/sub complex predicts processive replication of the complete genome by a single polymerase molecule in approximately 100 s. This value is in agreement with the value of 45 s measured in vivo (37). As long as 3D pol can proceed along the template unimpeded by RNA secondary structure, additional factors may not be required to assist in elongation.
Within infected cells, replication occurs in spherical compartments formed by virus-induced vesicles that are 0.5 m in diameter (38). This diameter predicts a volume of 0.065 m 3 . Enclosure of as few as 10 RNA molecules within this structure would correspond to a concentration in the micromolar range. Nonstructural proteins, such as 3D pol , are present in a 1:1 stoichiometry with structural proteins that are required in 60-fold molar excess of RNA for virion assembly (37). Therefore, the concentration of 3D pol within replication complexes will also be in the micromolar range. However, metabolic labeling experiments have shown that in virus-infected cells the

Rates of incorporation of ribo-and deoxyribonucleotides
Experiments were performed as described in the legend to Fig. 7 for AMP, 2Ј-dAMP, and 3Ј-dAMP incorporation. For 2Ј,3Ј-ddAMP incorporation, reactions were performed as follows. Reactions contained 5 M 3D pol , 1 M sym/sub (0.5 M duplex), 5 mM MgCl 2 , and 500 M 2Ј,3Ј-ddATP. Reactions were initiated by addition of 3D pol , incubated at 30°C, and quenched by addition of EDTA to a final concentration of 50 mM. The k obs was obtained from the fit of the data to a single exponential. polymerase subunit can be found in precursor (P3 and 3CD) and processed forms (3D pol and 3DЈ). 3DЈ is a proteolytic fragment of 3D pol that presumably lacks polymerase activity. The ratio of P3:3CD:3D pol :3DЈ is 2:4:1:1 after a 10-min labeling period, thus reducing the concentration of 3D pol available for transcription and replication. We conclude that the concentration of sym/sub and 3D pol employed in this study is in the biologically relevant range. In addition, the calculated K d value Table IV) of 1 M for sym/sub reported herein along with the micromolar K m and K d values for nucleic acid reported previously by us (30,34) and others (39) are in a physiological range. Despite the fact that polymerase is present in excess of RNA and both are in the micromolar range, given an association rate constant of 0.1 M Ϫ1 s Ϫ1 , it is possible that formation of the initial 3D pol -sym/sub complex could be the rate-limiting step for initiation.
In vivo, the density of polymerase molecules on any template has been determined to be ϳ6 -8 (37). If it takes 100 s for the polymerase to traverse the entire genome, then this would set a lower limit of 0.06 -0.08 s Ϫ1 on the rate of initiation. This value is remarkably close to that measured for the rate of the conformational change step (0.076 s Ϫ1 ) in our model for assembly. This conformational change after binding of sym/sub to 3D pol may reflect repositioning of the primer strand into the "exit channel" for nascent RNA. Such reorganization could maximize polymerase-nucleic acid interactions. Indeed, this isomerization greatly enhances the stability of this complex; the rate constant for dissociation is decreased from 0.1 s Ϫ1 to 0.0006 s Ϫ1 . This difference corresponds to at least 5 kcal/mol, consistent with the formation of 5-10 hydrogen bonds. A conformational change upon complete occupancy of the polymerase nucleic acid-binding site by single-stranded RNA has also been proposed by Kirkegaard and colleagues (39). Perhaps a conformational change that positions the protein primer into the exit  Table IV. B, concentration dependence of sym/sub. Reactions contained 1 M 3D pol , 500 M ATP, 5 mM MgCl 2 , and either 0.25 (q), 0.5 (E), 1 (f), 2 (Ⅺ), 5 (OE), or 10 (‚) M sym/sub (duplex). Reactions were initiated by addition of 3D pol and incubated at 30°C; reactions were quenched at the indicated times by addition of EDTA to a final concentration of 50 mM. The solid lines represent the kinetic simulation of the mechanism shown in Scheme 1 with the kinetic parameters shown in Table IV. SCHEME 1. Minimal kinetic mechanism for assembly of and nucleotide incorporation by 3D pol -sym/sub complexes. E(3D pol ); X (inactive 3D pol ); R n (sym/sub); *ER n (active elongation complex).  channel limits and regulates the rate of initiation in vivo. An alternative possibility for the conformational change step observed kinetically comes from structural studies of the RdRP from hepatitis C virus (40). Extensive interactions between the fingers and thumb subdomains exist in this enzyme which completely encircle the catalytic center of the enzyme. Similar interactions should be possible for 3D pol ; therefore, the observed conformational change step may reflect "closing" of the enzyme. The increased stability of this complex would derive from the additional interactions between the fingers and thumb subdomains.
The steady-state kinetic experiments (Fig. 8) were performed by using excess sym/sub; therefore, the y intercept reflects the active concentration of enzyme used in the reaction. The value measured in all cases was 80% of the value expected based on the concentration determined by absorbance at 280 nm (31). The value of 80% was also corroborated by the assembly experiments shown in Fig. 9B. That 20% was inactive was supported by the necessity to consider this species in our modeling of the kinetics of assembly in reactions that vary 3D pol concentration (Fig. 9A). Taken together, these data suggest that a 3D pol monomer is an active form of the enzyme. Oligomerization of this enzyme clearly can occur in the presence and absence of nucleic acid (29,39). Mutations in 3D pol -coding sequence that are predicted to shift the monomer-oligomer equilibrium in favor of monomer render the virus inviable (41). Our data suggest that sym/sub binding to 3D pol oligomers produces catalytically inactive complexes, thus precluding quantitative assembly of active complexes when a high concentration of 3D pol is employed (Fig. 9A). In addition, it is possible that the inactive fraction (20%) observed at 1 M 3D pol may represent an oligomeric species. The role of 3D pol oligomerization in vivo remains elusive and warrants continued investigation. Consistent with previous results (30), 3D pol incorporates 2Јand 3Ј-dNMPs (Table III). During the first cycle of incorporation, NTP/2Ј-dNTP selection is low (121-fold) relative to 2Ј-dNTP/NTP selection by Klenow fragment of DNA polymerase I (10 3 -10 6 -fold) from Escherichia coli (42) and reverse transcriptase (10 5 -fold) from human immunodeficiency virus (7). However, NTP/2Ј-dNTP selection is similar to that observed for the DNA-dependent RNA polymerase (80-fold) from bacteriophage T7 (43). Incorporation of 2Ј-dNMPs in vivo should be minimized by the low cytosolic availability of dNTPs relative to rNTPs. Of course, active mechanisms that preclude single and/or multiple cycles of dNMP incorporation cannot be ruled out at this time (30). Whether the observed NTP/2Ј-dNTP selection is a ground state, transition state, or combined effect requires another study but can be determined by using the sym/sub system. In addition, elucidation of the structural basis for NTP/dNTP selection should be facilitated by using the sym/ sub system. The use of the sym/sub system to study 3D pol will advance our understanding of the RdRP to the level of the other classes of nucleic acid polymerase. The sym/sub system should be useful for the analysis of other RdRPs, especially members of the supergroup I family (44). We have used this substrate to identify and characterize the active form of the feline calicivirus polymerase. 4 In contrast, the RdRPs from bovine viral diarrhea virus and hepatitis C virus, members of the super-group II family (44), do not assemble stable, elongation-competent complexes by using sym/sub. 2 Whether a universal primertemplate design exists that is suitable for mechanistic analysis of all RdRPs remains to be determined.