Insights into the role of an active site aspartate in Ty1 reverse transcriptase polymerization

LTR-retrotransposons encode reverse transcriptases (RTs) that replicate their RNA into integratable, double-stranded DNA. A mutant version of the RT from Saccharomyces cerevisiae retrotransposon Ty1, in which one of the three active site aspartates has been changed to asparagine (D 211 N), is still capable of in vitro polymerization although it is blocked for in vivo transposition. We generated recombinant WT and D 211 N Ty1 RTs to study RT function and determine specific roles for the D 211 residue. Pre-steady-state kinetic analysis of the two enzymes shows that the D 211 N mutation has minimal effect on nucleotide binding, but reduces the k pol by ~230 fold. The mutation reduces binding affinity for both Mn 2+ and Mg 2+ , indicating that the D 211 side chain helps create a tight metal binding pocket. Although both enzymes are highly processive and tend to remain bound to their initial substrate, each shows distinctive patterns of pausing, attributable to interactions between metal ions and the active site residue. These results provide insights to specific roles for the D 211 residue during polymerization and indicate unusual enzymatic properties that bear on the Ty1 replication pathway.


Introduction
Reverse transcriptases (RTs) are DNA polymerases that can use either RNA or DNA as their template. These enzymes are essential in multiple biological settings, ranging from retroviral and retrotransposon replication to telomere maintenance. A characteristic feature of the primary amino acid sequences of RTs is the presence of the "YXDD motif" (1)(2)(3). The two invariant D residues, as well as a third D that is amino terminal to this motif, are positioned within the active site of the three RTs whose crystal structures have been solved (4)(5)(6)(7)(8). By analogy with other polymerases, which contain either two or three invariant Ds, it has been suggested that these residues function to coordinate the two catalytic metal ions involved in the polymerization reaction. Polymerization is a complex process and the architecture of the active site must properly position incoming dNTPs and the primer 3' OH, aid in the nucleophilic attack, stabilize the transition state, assist in the removal of the pyrophosphate leaving group, and allow translocation of the primer end by one base. The active site D residues may influence all of these processes either directly or indirectly.
Among the myriad of RTs, HIV-1 RT has received the lion's share of biochemical and genetic attention. Early mutagenesis studies indicated that substitution of any of the three aspartates led to loss of both infectivity and in vitro polymerase activity (9,10). Because of the severe biochemical defects associated with these substitutions, the consequences of modifications of these residues have not been extensively studied. Only one report measured steady-state biochemical parameters for mutant HIV RTs in which the first D (110) was changed to A or S, the second D (185) was changed to A, E or N, and the third D (186) was changed to A, that this altered enzyme can be used to better understand the catalytic function of the active site D residues in RTs.
While VLPs are excellent sources of Ty1 intermediates, they are too impure to use for detailed biochemical comparisons of RT, as they contain multiple other activities at stoichiometric amounts. We have therefore purified recombinant forms of WT and D 211 N RT and compared their pre-steady-state or steady-state parameters associated with a number of steps in polymerization, including nucleotide binding, single base addition, divalent metal ion binding, and dissociation of the enzyme from the template. Since most of these parameters have not previously been studied for the WT enzyme, this allowed us to compare the retrotransposon enzyme to retroviral RTs, as well as to look at the specific effects of this novel active site mutation. A previous study comparing a different recombinant form of WT and D 211 N Ty1 RT, concluded that the WT enzyme binds Mg 2+ with high affinity but Mn 2+ with very low affinity, while the mutant enzyme behaves in a reciprocal manner (15). On the contrary, we find that the mutant enzyme has reduced affinity for both catalytic metals, and that the two different metals affect polymerization of the two enzymes in idiosyncratic concentration-dependent ways. By single base turnover analysis, we find that the D 211 N substitution has an effect on the rate of polymerization, indicating a role in the chemistry of polymerization, but that it is not involved in either dNTP binding or dissociation of enzyme from its primer/template. Finally, the D 211 N mutant results in distinctive changes in the pattern of pausing during polymerization, suggesting a role for this residue in pyrophosphate release and/or translocation.

6
The plasmid p6H Ty1 IN-RT-RH (construct 2 in (16)) contained WT Ty1 RT-RH plus a 115 amino acid contiguous C-terminal portion of Ty1 integrase fused to the N-terminus of the RT-RH domain, all preceded by 6 histidine residues. It was kindly provided by Dr. F. X. Wilhelm (IBMC, Strasbourg) and designated AGE2186 in our laboratory. These extra 115 amino acids are part of a C-terminal extension found in Ty1 integrase, which is far downstream of all potential metal binding domains of Ty1 integrase (17). The fusion construct was used because equivalent recombinant proteins beginning at the N terminus of the RT-RH domain totally lack in vitro polymerase activity ((16) and A. Gabriel, unpublished results). We had previously constructed the D 211 N polymerase active site mutation and expressed it in plasmid pGTy1D 211 Nmhis3AI (AGE1603) (12). To make the analogous mutant expression plasmid, constructs AGE1603 and AGE2186 were cut with Sph I and Afl II and the 386 bp fragment of the former was ligated to the 5.016 kb fragment of the latter, to generate AGE2352 in strain XL1-

Determination of the K d of the divalent ions
Reactions were carried out in the presence of various concentrations of divalent ions, 0 -230 mM Mg 2+ (as MgCl 2 ) and 0 -49 mM Mn 2+ (as MnCl 2 ), in a 10 minute extension assay using the 5' 32 P-end labeled 14-mer/28-mer substrate, with 8.0 micromolar each dNTP and equivalent amounts of active WT or D 211 N enzymes. Extended samples were resolved on gels and analyzed, as above. Amount of metal ion bound to dNTP at each total metal ion concentration was calculated using a quadratic equation (Eqn. 3, see supplementary information). For this estimate, we used a K d Mg•dATP of 20 micromolar (20) and a K dMn•dATP of 9.77 micromolar (21). The amount of product formed versus the free metal ion concentration was fitted to the Hill equation

Pausing experiments
We used the same 5' 32 P-end labeled 14-mer/28-mer DNA/DNA substrate (40 nM) for this experiment. All reactions were performed in extension buffer at the specified concentrations of Mg 2+ or Mn 2+ using WT or D 211 N enzymes. The reactions were divided into two, without or with an excess of an activated calf thymus DNA trap at 2.63 mg/ml, final concentration. Reactions were initiated by adding all four dNTPs (8.0 micromolar each), to the pre-warmed enzyme and labeled DNA substrate. The reactions were incubated at 22°C for 10 minutes. All reactions were stopped with loading buffer, and denatured samples were analyzed by electrophoresis as described. A trap effectiveness control was also carried out where the reaction was initiated by adding dNTPs with Mg 2+ or Mn 2+ to a mix containing enzyme, trap and primer/template (where enzyme was added to the labeled substrate in the presence of trap and then pre-warmed along with the two). For the DNA/RNA substrate, we used the 5' labeled 14-mer DNA primer as above and a 28-mer RNA template, whose sequence was equivalent to the above DNA template: (5'-AUU ACA UUA UGG GUG GUA UGU UGG AAU A -3'). These reactions were also carried out in the presence and the absence of trap. However, here we used the unlabeled DNA/DNA substrate as a trap at 200-fold excess to the labeled substrate concentration.
Reactions were carried out at 10 or 30 mM Mg 2+ or at 2 mM Mn 2+ .

Processivity Assay
Poly r(A) (500 -600 nt) or poly d(A) (350 -400 nt) primed with 5' end labeled oligo d(T) [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] oligonucleotide (0.0894 ng/microliter final concentration) was pre-incubated with WT or D 211 N Ty1 RT for 10 minutes at 22°C. The extension reactions were initiated by adding a mixture containing 24 micromolar dTTP along with 10 mM or 30 mM Mg 2+ or 2 mM Mn 2+ with or without an unlabeled 14-mer/28-mer DNA/DNA substrate trap (at a final concentration of 200 fold more than the labeled primer/template). Reactions were carried out in the extension buffer at 22°C for 30 minutes and then terminated with loading buffer. Control reactions were carried out using Klenow polymerase under the same reaction conditions, with 10 mM Mg 2+ or 2 mM Mn 2+ . Trap effectiveness controls were carried out as above. The terminated processivity reaction and control reactions were resolved by 6% polyacrylamide-urea denaturing gel electrophoresis and analyzed as above. Trap effectiveness reactions were performed by adding the enzyme to the mix of labeled and unlabeled substrates, pre-incubated for 10 minutes, then the reaction was started by adding dATP and Mg 2+ for 3 minutes. For the 0 time samples, dATP and Mg 2+ were added along with the trap and reaction was carried out for 3 minutes. The reactions were terminated with loading buffer and the samples were analyzed on denaturing gels. The amounts of the extension products (15-mer and greater) were determined relative to total labeled product (extended and un-extended 14mer). These values were normalized for 0 seconds, and then the ratio of intensity of extension products/total intensity were plotted against the time in minutes. The obtained curves were fitted to a single exponential equation using a three parameter fit for the decay (Eqn. 5, see supplementary information).

Pre-steady-state kinetics of dATP incorporation of WT and mutant D 211 N Ty1 RT
Recombinant versions of both WT and D 211 N mutant Ty1 RT enzymes were purified by nickelaffinity chromatography (Fig. 1). Both WT and D 211 N mutant enzymes were active in a standard homopolymer assay, using α-[ 32 P] dGTP (data not shown), as expected from previous studies (12,15). The homopolymer assay is, however, extremely crude and not designed to measure subtle differences between enzymes.
As a baseline for comparing the polymerization properties of the WT and D 211 N enzymes, we therefore used a model substrate and measured the K d and k pol for single nucleotide incorporation for the two enzymes, using pre-steady-state kinetics, based on the following minimal scheme for single base addition: To this end, a 5' end-labeled 14-mer primer was annealed to a 28-mer template, mixed with RT,

Primer extension patterns differ for WT and D 211 N mutant RTs on a model substrate
We next compared the extension pattern of the two enzymes on the same model 14-mer/28-mer DNA/DNA substrate, in the presence of all four dNTPs. First we examined the pre-steady-state rates of extension for multiple base additions, using saturating concentrations of dNTPs. The results were similar to those obtained with our single base addition studies (data not shown).
Next we compared the extension patterns of the two enzymes in the presence of either MgCl 2 or MnCl 2 . The two enzymes generated distinct extension patterns under these conditions (Fig. 3).
In the presence of 10 mM Mg 2+ , the WT enzyme was able to fully extend the primer within 10 minutes (Fig. 3, lane 5), and further extension consisted of non-templated base addition beyond the 28-mer length (Fig. 3, lane 7). For the D 211 N RT, extension was slower (Fig. 3 This suggests either stalling or dissociation of enzyme from primer/template (Fig. 3, lanes 9 and 11). We refer to this phenomenon as "pausing". The D 211 N RT carried out more complete extension in Mn 2+ than in the presence of Mg 2+ , and has distinct pause sites compared with Mg 2+ (compare Fig. 3, lanes 8 and 12). Nontemplated base addition was much more apparent in the presence of Mn 2+ , particularly for the WT enzyme, where up to 4 bases were added beyond the template.  extension patterns for the WT enzyme were unaffected by the presence of the trap (Fig. 5 a, lanes 1 through 6). For the mutant enzyme, a small decrease in full-length products was observed at 10 and 20 mM Mg 2+ (Fig. 5 a, lanes 9 through 12). When 2 mM Mn 2+ was used in the reactions both enzymes had the same extension patterns in the presence or absence of the trap (Fig. 5 b).

Stalling on-versus dissociation from-the template
Thus it appears that the observed pausing patterns are not primarily due to dissociation of the enzyme from its primer/template. Note that we used a Mn 2+ concentration in the millimolar range for this experiment, after we discovered that the trap chelated the metal in the more optimal micromolar range (see Fig. 4).
Since a DNA/RNA substrate is a natural substrate for RTs, we also compared the extension and pausing patterns of our 14-mer/28-mer DNA/DNA substrate with a sequenceequivalent DNA primer/RNA template. This allowed us to determine if pausing is a consequence of the chemical make-up of the template or depended more on the specific base sequence of the template. As shown in Fig. 6

Reduced extension capacity for the D 211 N mutant enzyme with long templates
Since extension by 14 bases is not a particularly long template upon which to judge processive polymerization, we next examined extension patterns using substrates consisting of poly r(A) or poly d(A) templates, primed with a 5' 32 P-labeled oligo d(T) [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] oligonucleotide. Fig. 7 shows the resolved extension products in the presence or absence of trap. In the presence of Mg 2+ , the WT enzyme was much more processive than Klenow (compare Fig. 7 As seen with the 14-mer/28-mer substrate, the WT enzyme extends well in the presence of Mg 2+ , but has a relative extension defect in the presence of Mn 2+ , particularly for DNA/DNA substrates. In Mg 2+ , the WT enzyme appears extremely processive, extending over 600 bases in the presence of a trap (Fig. 7, lanes 1 to 8)

Rate of Dissociation of the Ty1 RT•DNA complex is low
To see if the mutant enzyme's low k pol and unexpectedly high processivity are related to the rate of dissociation of enzyme from its DNA substrate, we measured the k off for both enzymes from the 5' labeled 14-mer/28-mer DNA/DNA substrate, by incubating substrate and enzyme in the presence of a trap for various times before initiating extension. The resolved extension products were quantified, and the decreasing fraction of enzyme still bound to the DNA template (I extended /I total ) was plotted versus time and k off for each enzyme was determined (Fig. 8). As shown in the

Discussion
We have carried out the first pre-steady-state analysis of a retrotransposon RT, to better understand the biochemical properties of this enzyme and the defects associated with an unusual polymerase mutant. Although the D 211 N substitution renders the mutated Ty1 element incapable of in vivo transposition, in vitro polymerization with homopolymeric templates is still observed (12,15). Further, second site substitutions in the RNase H domain restore 5 -10 % of the transposition activity (12). Similar active site substitutions for HIV-1 RTs have yielded severely crippled enzymes with no infectivity (9)(10)(11)22). Second site suppressors of equivalent HIV-1 mutant have not been reported. Among other polymerases, substitution of active site carboxylates is also generally inactivating (23,24), but in a few cases, some level of activity can be restored in the presence of alternate cations, e.g. Mn 2+ (25,26).
Using pre-steady-state single turnover kinetic analysis we determined the dissociation constant (K d ) for dATP and the rate of the phosphoryl transfer reaction (k pol ), for both WT and mutant RTs. Both kinetic constants for the WT Ty1 enzyme (Table) are similar to those obtained for HIV-1 RT, by either pre-steady-state or steady-state calculations (11,27).
Interestingly, although the k pol for the D 211 N mutant was significantly decreased, the K d was similar to WT. Current models of polymerization suggest that dNTPs enters the active site already bound to a divalent metal ion (referred to as the B metal), and that this metal could also aid in expelling the PP i leaving group after the chemical step of catalysis (28,29). Our result suggests that the D 211 side chain is not crucial for entry of the dNTP into the binding site or for the formation of the coordination shell around the entering metal, consistent with the crystal structure of HIV-1 RT and other polymerases (7,28,30), as well as biochemical analysis of HIV-primer end translocation. The D 211 side chain is likely to interact with metal ions, either directly or indirectly through a water molecule, given the very different response to Mg 2+ and Mn 2+ seen with the WT and mutant enzymes, which differ by only the single negatively charged side chain carboxylate versus the uncharged carbonyl amide. Mg 2+ has a slightly smaller ionic radius than Mn 2+ and is a "harder" metal than Mn 2+ , much less likely to coordinate with an uncharged nitrogen compared with a negatively charged oxygen. Both metals coordinate readily with water or with carboxylate side chains, and prefer a coordination number of 6 (32,33). We found that both enzymes have a much lower K d for Mn 2+ relative to Mg 2+ (Table). Further, both metals bind tighter to the WT enzyme than to the mutant enzyme. Interpretation is complicated by the fact that two metal ions bind cooperatively in the active site, but the implication is that the D 211 side chain plays a role in creating a tight binding pocket for whichever metal is present.
The metal ion titrations (Fig. 4)  the range where polymerization is blocked beyond the first base addition, well above the optimal range in which polymerization occurs. The concentration of dGTP that they used is 10 fold lower than our estimated K d for nucleotide binding, which could be another limiting factor.
Further, Bolton et al measured % incorporation with a homopolymer substrate and did not resolve their products, whereas we measured the amount of extension using a specific substrate and a known amount of active enzyme in a time frame before reagents became exhausted, and visually determined the lengths of extension.
An unexpected finding of our study was the distinct patterns of pausing seen with the two enzymes under different metal conditions. Using traps we demonstrated that intermediate extension products were not due to enzyme dissociation. Both enzymes bind their primer/templates much more stably than either AMV or MMLV RTs, and had off rates similar to the RT from the non-LTR retrotransposon R2Bm (35). We also inferred tight binding by the processivity of the yeast enzymes (Fig. 7) as well as the relative lack of turnover seen during time courses (Figs. 2, 3 and data not shown). Instead of dissociating from the full-length extension product and associating with additional substrate, the major consequence over time for full-length WT enzyme extensions was an increase in nontemplated base additions (Fig. 3). This        Sizes are marked according to labeled marker.  The samples were then saved in the same buffer at -20 o C with 50% glycerol. 5'-32 P

Labeling of oligonucleotides and DNA markers
DNA oligonucleotides and 1kb and 100 bp marker ladders (New England Biolabs) were 5' end-labeled using [g- 32

Primer Extension Assay with DNA Templates
In the usual assay, a DNA/DNA template/primer was prepared by annealing a 28-mer plus-stand sequence from the polypurine tract region of Ty1 RT (5'-ATT ACA TTA TGG GTG GTA TGT TGG AAT A -3', where the polypurine tract is underlined) with a complementary 14-mer (5' T ATT CCA ACA TAC C 3')] ( template/primer ratio is 0.85/1), whose 5'-end is 32 P-labeled, to generate the following model substrate: † polyacrylamide gels (or as specified time in Materials and Methods section). Using a phosphorimager, the amounts of the extension products (15-mer and greater) were determined relative to total labeled products seen (extended and un-extended 14-mer).

Eqn. 1: †
[P n +1 ] t = y 0 + [P n +1 ] max (1-e (-k¥time ) where k is the observed rate per second, [P n+1 ] t is the amount of product at a given time, t, [P n+1 ] max is the maximum amount of product formed is in nM and y o is the y intercept. where n is the Hill constant, and product formed is in nM.