Evidence against a simple tethering model for enhancement of herpes simplex virus DNA polymerase processivity by accessory protein UL42

Abstract

The DNA polymerase holoenzyme of herpes simplex virus type 1 (HSV-1) is a stable heterodimer consisting of a catalytic subunit (Pol) and a processivity factor (UL42). HSV-1 UL42 differs from most DNA polymerase processivity factors in possessing an inherent ability to bind to double-stranded DNA. It has been proposed that UL42 increases the processivity of Pol by directly tethering it to the primer and template (P/T). To test this hypothesis, we took advantage of the different sensitivities of Pol and Pol/UL42 activities to ionic strength. Although the activity of Pol is inhibited by salt concentrations in excess of 50 mM KCl, the activity of the holoenzyme is relatively refractory to changes in ionic strength from 50 to 125 mM KCl. We used nitrocellulose filter-binding assays and real-time biosensor technology to measure binding affinities and dissociation rate constants of the individual subunits and holoenzyme for a short model P/T as a function of the ionic strength of the buffer. We found that as observed for activity, the binding affinity and dissociation rate constant of the Pol/UL42 holoenzyme for P/T were not altered substantially in high- versus low-ionic-strength buffer. In 50 mM KCl, the apparent affinity with which UL42 bound the P/T did not differ by more than twofold compared to that observed for Pol or Pol/UL42 in the same low-ionic-strength buffer. However, increasing the ionic strength dramatically decreased the affinity of UL42 for P/T, such that it was reduced more than 3 orders of magnitude from that of Pol/UL42 in 125 mM KCl. Real-time binding kinetics revealed that much of the reduced affinity could be attributable to an extremely rapid dissociation of UL42 from the P/T in high-ionic-strength buffer. The resistance of the activity, binding affinity, and stability of the holoenzyme for the model P/T to increases in ionic strength, despite the low apparent affinity and poor stability with which UL42 binds the model P/T in high concentrations of salt, suggests that UL42 does not simply tether the Pol to DNA. Instead, it is likely that conformational alterations induced by interaction of UL42 with Pol allow for high-affinity and high-stability binding of the holoenzyme to the P/T even under high-ionic-strength conditions.

FIG. 1.

Purity of UL42, Pol, and Pol/UL42 preparations. Proteins were expressed in insect cells infected with recombinant baculoviruses and purified by column chromatography as described in Materials and Methods. Concentrated preparations (2 μg) of UL42 (lane 1), Pol (lane 2), or Pol/UL42 (lane 3) were subjected to electrophoresis through a denaturing gradient polyacrylamide gel (5 to 15%) followed by silver staining. Preparations were estimated to be at least 95% pure.

FIG. 2.

Electrophoretic mobility shift analysis of protein binding to model P/T. Purified Pol (lanes 3 to 6), UL42 (lanes 9 to 12), or Pol/UL42 heterodimer (lanes 15 to 20) preparations were incubated with ³²P-end-labeled P/T (100 nM), and the formation of complexes was analyzed by electrophoresis through nondenaturing 6% polyacrylamide gels. The final concentrations of Pol and UL42 tested were 100, 200, 300, and 400 nM, while those tested for the Pol/UL42 heterodimer were 25, 50, 100, 200, 300, and 400 nM (concentration shown by the thickness of the triangle over the lanes). The mobility of the P/T probe (P) alone (lanes 1, 7, and 13) or in the presence of 400 nM BSA (B) (lanes 2, 8, and 14) also is shown. In some preparations of probe, nonspecific (ns) higher-mobility bands were noted. Specifically shifted complexes are indicated by arrows.

FIG. 3.

Effect of ionic strength on polymerase activities of purified Pol and Pol/UL42. Polymerase activities were measured by the incorporation of [³H]dTTP into acid-insoluble form using activated calf thymus DNA as the template essentially as described previously (12) in buffer containing the indicated concentrations of KCl. The relative activities of 1.2 nM Pol (•) or 2.5 nM Pol/UL42 (▪) were calculated as a percentage of the activities in 50 mM KCl, corresponding to 1,200 and 1,050 units, respectively. A unit of activity was defined as the number of femtomoles of [³H]dTTP incorporated in 20 min at 37°C.

FIG. 4.

Nitrocellulose filter-binding assays. (A to C) Binding isotherms. Purified Pol with an active concentration of 52 nM (A) or purified Pol/UL42 complexes with active concentrations of 10 nM (B) and 11 nM (C) were incubated with increasing concentrations of ³²P-labeled P/T in buffer C containing either 50 mM (A and B) or 125 mM (C) KCl and applied in duplicate to a filter stack containing nitrocellulose and DEAE to trap DNA-protein complexes and free DNA, respectively. The data were fitted to equation 3 to predict the maximum stoichiometry of DNA-protein complexes and the DNA concentration at half-maximum binding (K_d). (D) Scatchard plots of the same data fitted by linear regression analysis. The negative values of the reciprocals of the slopes give estimates of P/T binding K_ds of 11.5 nM for Pol (•) and 1.7 nM for Pol/UL42 (▪) in 50 mM KCl and 1.0 nM for Pol/UL42 in 125 mM KCl (▴).

FIG. 5.

Stability of protein-DNA complexes. Purified Pol (52 nM) (A) or Pol/UL42 (11 nM) (B and C) was equilibrated with 6 nM labeled P/T in buffer C containing either 50 (A and B) or 125 mM KCl (C). At time zero, excess unlabeled activated calf thymus DNA trap was added and the amount of labeled DNA-protein complexes remaining as a function of time was determined by nitrocellulose filter-binding activity as described in the legend to Fig. 4. The data were transformed to produce linear plots of the natural logarithm of the proportion remaining versus time.

FIG. 6.

Sensorgrams of binding of purified proteins to model P/T. Binding of Pol (A and D), UL42 (B and E), and Pol/UL42 heterodimer (C and F) to biosensor SA-gold chips to which 120 RU of biotinylated P/T was bound was measured using the BIAcore 2000 system as described in Materials and Methods. Curves of increasing thickness denote binding of 25, 50, 150, 300, and 600 nM protein. Binding and dissociation were performed in buffer A containing either 50 mM KCl (A to C) or 125 mM KCl (D to F). Sensorgrams shown are representative of at least two independent experiments for each protein and the concentrations indicated.

FIG. 7.

Effect of salt concentration on binding affinity of UL42 for P/T. The apparent K_d for binding of UL42 to P/T was determined by BIAcore analysis and global curve fitting as described in the legend to Fig. 6 except that binding and dissociation buffers contained the KCl concentrations indicated on the x axis.

FIG. 8.

Conformational model for mechanism by which UL42 increases Pol processivity. Pol and Pol/UL42 are represented schematically before (left) and after (right) binding to P/T. (Top) Although Pol alone displays some change in conformation upon DNA binding, it dissociates rapidly from the P/T, leading to low processivity. (Bottom) Binding of UL42 to the Pol C terminus alters the conformation of Pol and in the presence of DNA, the two proteins together serve as a clamp, decreasing the dissociation rate and increasing processivity.