Proton relay mechanism of general acid catalysis by DNA topoisomerase IB.

Type IB topoisomerases cleave and rejoin DNA through a DNA-(3'-phosphotyrosyl)-enzyme intermediate. A constellation of conserved amino acids (Arg-130, Lys-167, Arg-223, and His-265 in vaccinia topoisomerase) catalyzes the attack of the tyrosine nucleophile (Tyr-274) at the scissile phosphodiester. Previous studies implicated Arg-223 and His-265 in transition state stabilization and Lys-167 in proton donation to the 5'-O of the leaving DNA strand. Here we find that Arg-130 also plays a major role in leaving group expulsion. The rate of DNA cleavage by vaccinia topoisomerase mutant R130K, which was slower than wild-type topoisomerase by a factor of 10(-4.3), was stimulated 2600-fold by a 5'-bridging phosphorothiolate at the cleavage site. The catalytic defect of the R130A mutant was also rescued by the 5'-S modification (190-fold stimulation), albeit to a lesser degree than R130K. We surmise that Arg-130 plays dual roles in transition state stabilization and general acid catalysis. Whereas the R130A mutation abolishes both functions, R130K permits the transition state stabilization function (via contact of lysine with the scissile phosphate) but not the proton transfer function. Our results show that the process of general acid catalysis is complex and suggest that Lys-167 and Arg-130 comprise a proton relay from the topoisomerase to the 5'-O of the leaving DNA strand.


Berit Olsen Krogh and Stewart Shuman ‡
Here we find that Arg-130 also plays a major role in leaving group expulsion. The rate of DNA cleavage by vaccinia topoisomerase mutant R130K, which was slower than wild-type topoisomerase by a factor of 10 ؊4.3 , was stimulated 2600-fold by a 5-bridging phosphorothiolate at the cleavage site. The catalytic defect of the R130A mutant was also rescued by the 5-S modification (190-fold stimulation), albeit to a lesser degree than R130K. We surmise that Arg-130 plays dual roles in transition state stabilization and general acid catalysis. Whereas the R130A mutation abolishes both functions, R130K permits the transition state stabilization function (via contact of lysine with the scissile phosphate) but not the proton transfer function. Our results show that the process of general acid catalysis is complex and suggest that Lys-167 and Arg-130 comprise a proton relay from the topoisomerase to the 5-O of the leaving DNA strand.
Type IB DNA topoisomerases relax DNA supercoils via a reaction pathway entailing noncovalent binding of the enzyme to duplex DNA, cleavage of one DNA strand with formation of a covalent DNA-(3Ј-phosphotyrosyl)-protein intermediate, strand passage, and strand religation (1,2). Tyrosine recombinases use a similar transesterification mechanism to form and resolve Holliday junctions. The catalytic domains of topo 1 IB and tyrosine recombinases adopt a common fold composed of eight ␣ helices and a three-stranded antiparallel ␤ sheet (3)(4)(5)(6)(7)(8)(9). The constituents of the active site occupy similar positions in the topo IB and recombinase tertiary structures.
Four conserved amino acid side chains (e.g. Arg-130, Lys-167, Arg-223, and His-265 in the vaccinia topoisomerase) cat-alyze the attack of the active site tyrosine nucleophile (Tyr-274) on the scissile phosphodiester (10 -12). Mutational, stereochemical, and structural data for vaccinia and nuclear topoisomerase IB and tyrosine recombinases suggest that the two arginines and the histidine contact the nonbridging oxygens of the scissile phosphodiester and that these interactions serve to stabilize a proposed pentacoordinate phosphorane transition state (3, 5, 9, 10 -13).
Recently we used 5Ј-bridging phosphorothiolate-modified DNAs to implicate Lys-167 of vaccinia topoisomerase as a general acid catalyst of the DNA cleavage reaction (14). The hypothesis was that if the expulsion of the 5Ј-oxygen of the leaving DNA strand was indeed catalyzed by a general acid on the topoisomerase, then the requirement for the general acid ought to be alleviated by introducing a 5Ј-bridging phosphorothiolate at the scissile phosphodiester because the pK a of the DNA 5Ј-SH is ϳ5 log units lower than that of the DNA 5Ј-OH. Accordingly the diagnostic feature of a topoisomerase mutant defective in general acid catalysis would be a significant increase in the cleavage rate on a 5Ј-S substrate compared with a 5Ј-O substrate. The "sulfur/oxygen ratio" ought not to increase for topoisomerase mutants that are impaired in other aspects of transesterification chemistry (e.g. transition state stabilization). Our findings were that that the 5Ј-S group restored activity to the catalytically defective K167A mutant, whereas there was no positive thiolate effect for mutants R223A and H265A (14). Thus, we concluded that Lys-167 functions in proton transfer to the leaving strand. Lys-167 is located in a flexible interstrand hairpin loop of vaccinia topoisomerase and is conserved in all known type IB topoisomerases and tyrosine recombinases.
Our earlier experiments did not address whether Arg-130 plays any role in leaving strand expulsion. Arg-130, which is invariant among type IB topoisomerases and tyrosine recombinases, enhances the transesterification rate of vaccinia topoisomerase by 5 orders of magnitude as gauged by the effect of the R130A mutation (10). Because a conservative R130K substitution had little restorative effect, it was surmised that multivalent contacts of the guanidinium nitrogens of Arg-130 with the scissile phosphodiester are essential for catalysis. The occurrence of either one or two hydrogen bonding contacts between this conserved arginine side chain and the DNA backbone in various crystal structures of topoisomerase IB and tyrosine recombinases (3,5,9) was broadly consonant with the mutational data for the vaccinia enzyme (11). Here we report the surprising observation that Arg-130 plays a major role in proton transfer to the 5Ј-leaving DNA strand.

EXPERIMENTAL PROCEDURES
Topoisomerase Mutants-Recombinant wild-type vaccinia topoisomerase and mutants K167A, R130A, and R130K were produced in Escherichia coli and purified by phosphocellulose chromatography as described previously (15). The double mutant R130K/K167A was constructed for the present study and was produced and purified using the same protocol.
Equilibrium Cleavage Assay-Reaction mixtures containing (per 20 l) 50 mM Tris-HCl (pH 7.5), 0.3 pmol of 34-mer/60-mer DNA (Fig. 1), and 75 ng of purified topoisomerase were incubated at 37°C. Aliquots (20 l) were withdrawn at the times indicated and quenched immediately with SDS. The mixtures were digested with 10 g of proteinase K for 60 min at 37°C, then adjusted to 50% formamide, and heat-denatured. The samples were analyzed by electrophoresis through a 17% polyacrylamide gel containing 7 M urea in 90 mM Tris borate, 2.5 mM EDTA. The cleavage product, a 32 P-labeled 12-mer oligonucleotide bound to a short peptide, was resolved from the 34-mer substrate. The extent of strand cleavage was quantitated by scanning the gel with a PhosphorImager. Rate constants for approach to equilibrium on the phosphodiester DNA were determined by normalizing the data to the end point values (redefined as 100) and fitting to the equation 100 Ϫ %Cleavage (norm) ϭ 100e Ϫkt . The cleavage equilibrium constants (K cl ) were calculated as %covalent/(100 Ϫ %covalent). Single-turnover cleavage rate constants (k cl ) on the phosphorothiolate-modified DNA were determined by normalizing the data to the end point values (redefined as 100) and fitting to the equation 100 Ϫ %Cleavage (norm) ϭ 100e Ϫkt .

Involvement of Arg-130 in General Acid
Catalysis-The DNA cleavage reactions of wild-type vaccinia topoisomerase and mutant protein R130A were examined using equilibrium substrates containing a 34-mer scissile strand with either a phosphodiester Tp2A (5Ј-O) or a 5Ј-bridging phosphorothiolate Tp2(s)A (5Ј-S) at the CCCTT2 cleavage site (Fig. 1). We refer to this as an equilibrium substrate because the 22-mer leaving strand generated upon cleavage at CCCTT remains associated with the topoisomerase-DNA complex via base pairing to the nonscissile strand. The yield of covalent adduct at equilibrium in reactions containing saturating levels of the wild-type vaccinia topoisomerase was increased from 26% for the 5Ј-O substrate (K cl ϭ 0.35) to 96% for the 5Ј-S substrate (Fig. 1A). An explanation for the observed altered equilibrium is that the 5Ј-sulfhydryl leaving strand in the cleavage reaction is an extremely poor nucleophile in the religation step (16,17). The reaction rates of wild-type topoisomerase on the 5Ј-O and 5Ј-S equilibrium substrates were identical within our limits of detection; both reactions were effectively complete in 5 s (the earliest time examined).
The rate of approach to equilibrium by R130A on the 5Ј-O duplex substrate was extremely slow (k obs ϭ 7.9 ϫ 10 Ϫ6 s Ϫ1 , K cl ϭ 0.54) (Fig. 1B). Solving the equations k obs ϭ k cl ϩ k rel and K cl ϭ k cl /k rel yielded a value for k cl of 2.8 ϫ 10 Ϫ6 s Ϫ1 , which is 5 orders of magnitude slower than the rate of single-turnover cleavage by wild-type topoisomerase (k cl ϭ 0.3 s Ϫ1 ) (10,18). The instructive finding was that the rate constant for single-turnover cleavage of the 5Ј-S DNA by R130A (k cl ϭ 5.3 ϫ 10 Ϫ4 s Ϫ1 ) was 190 times faster than k cl with the 5Ј-O substrate. Thus, the simple replacement of the 5Ј-O leaving group by the less nucleophilic 5Ј-S diminished the requirement for Arg-130 in catalysis. Nonetheless, the R130A cleavage rate on the 5Ј-S substrate was still much slower (by a factor of ϳ570) than the single-turnover cleavage rate of wild-type topoisomerase (0.3 s Ϫ1 ). These results provide evidence that Arg-130 participates in the step of proton donation to the 5Ј-O of the leaving strand, but they also suggest additional functions of Arg-130 in transesterification chemistry.
We compared the thiolate effects for R130A with those for the K167A mutant assayed in parallel (Fig. 1C). The approach to equilibrium by K167A on the 5Ј-O duplex substrate (k obs ϭ 6.8 ϫ 10 Ϫ5 s Ϫ1 , K cl ϭ 1.04) was an order of magnitude faster than that of R130A. As reported previously (11,14), the K167A mutation increased the cleavage equilibrium constant compared with wild-type topoisomerase. We calculated a value for k cl of 3.5 ϫ 10 Ϫ5 s Ϫ1 . The reaction of K167A with 5Ј-S DNA (k cl ϭ 1.3 ϫ 10 Ϫ2 s Ϫ1 ) was 370 times faster than the reaction with the 5Ј-O substrate and was within a factor of 25 of the k cl of wild-type topoisomerase. These results agree with kinetic data reported previously for K167A (14), and they underscore that the 5Ј-S modification boosted the activity of K167A closer to the wild-type level than it did the activity of R130A. Again, the implication is that Arg-130 plays two roles in catalysis.
5Ј-S Effect on the R130K Mutant-Insights to the dual function of Arg-130 emerged from an analysis of the thio effects on equilibrium cleavage reaction of the conservative mutant R130K (Fig. 1D). From the observed kinetic parameters of the approach to equilibrium by R130K on the 5Ј-O duplex substrate (k obs ϭ 3.5 ϫ 10 Ϫ5 s Ϫ1 , K cl ϭ 0.67), we calculated a value for k cl of 1.4 ϫ 10 Ϫ5 s Ϫ1 . This rate is 5-fold faster than that of R130A but slower by at least 4 orders of magnitude than wild-type topoisomerase. Reaction of R130K with the 5Ј-S DNA was much faster than the reaction with the 5Ј-O substrate (Fig. 1D). 93% of the input 5Ј-S DNA was bound covalently by R130K, and the apparent k cl was 3.7 ϫ 10 Ϫ2 s Ϫ1 . This value was within a factor of 10 of the k cl of wild-type topoisomerase and was 2600 times faster than k cl of the reaction of R130K with the 5Ј-O substrate. Thus, the gain of function by R130K in response to the bridging thiolate modification was comparable to that of the K167A mutation, i.e. R130K phenocopied K167A.
We surmise that Arg-130 and Lys-167 jointly participate in a pathway of proton transfer to the 5Ј-leaving strand. We also invoke a role for Arg-130 in transition state stabilization entailing a contact to one of the nonbridging oxygens. Whereas the R130A mutation abolishes both functions, the introduction of lysine permits the transition state stabilization function (via contact of lysine with the scissile phosphate) but not the proton transfer function.
5Ј-S Effect on a R130K/K167A Double Mutant-If the guanidinium group at position 130 and the primary amine at position 167 are collaborating in a proton transfer relay, then the catalytic defect of an R130K/K167A double mutation should still be rescued by the 5Ј-S modification, and indeed that is what we observed (Fig. 2). Cleavage of the 5Ј-O DNA by purified recombinant R130K/K167A was slower than either of the single mutants. The covalent adduct accumulated steadily over Catalytic Mechanism of DNA Topoisomerase IB 5712 a 4-day incubation period, after which 10% of the input 5Ј-O DNA had been cleaved (longer times were not tested). Absent an end point, we could not reliably calculate a cleavage rate constant for R130K/K167A on the 5Ј-O substrate. Nonetheless, it was clear that cleavage by R130K/K167A displayed a pronounced 5Ј-bridging thiolate effect (Fig. 2). 91% of the input 5Ј-S DNA was bound covalently by R130K/K167A, and the apparent k cl of 4.3 ϫ 10 Ϫ3 s Ϫ1 was within a factor of 3 of that of the single K167A mutant on the 5Ј-S substrate. DISCUSSION Lys-167 and Arg-130 enhance the rate of transesterification by factors of ϳ10 4 and ϳ10 5 , respectively. The substantial alleviation of the requirements for Lys-167 and Arg-130 during scission of a 5Ј-bridging phosphorothiolate-modified substrate indicates that the process of general acid catalysis is more complex than was anticipated initially (14). The present data suggest that Lys-167 and Arg-130 comprise a proton relay from the topoisomerase to the 5Ј-O of the leaving DNA strand. The order of the proton relay steps (e.g. Arg-130 3 Lys-167 3 5Ј-O versus Lys-167 3 Arg-130 3 5Ј-O) cannot be surmised from the biochemical data alone. Moreover, none of the crystal structures of topoisomerase IB or tyrosine recombinases are directly illuminating on this issue. The structures provide snapshots of the protein-DNA interface before and after transesterification, but the atomic contacts during the chemical step, and especially in the transition state, are unclear and likely to differ from those seen in the pre-and postcleavage states. Some of the key side chains may interact only transiently with the scissile phosphate.
In particular, it is evident that the ␤ loop in which Lys-167 is situated displays considerable conformational flexibility in the various tyrosine recombinase and topoisomerase IB structures (see Ref. 14 for discussion). We suggested a model whereby the ␤ loop and Lys-167 initially lie outside the circumference of the DNA helix, and a conformational change upon cleavage site recognition triggers movement of the loop such that Lys-167 interacts transiently with the 5Ј-bridging oxygen in the transition state, donating its proton to form the 5Ј-OH of the leaving strand and subsequently reorienting its side chain to interact with the ϩ1 base in the minor groove (14). Recent structures of Flp recombinase and human topoisomerase IB are at least consistent with the idea that Lys is a direct proton donor to the leaving strand, insofar as (i) the Lys is located within hydrogen bonding distance (3.16 Å) of the 5Ј-OH of the cleaved strand in one of the protomer pairs of the Flp-DNA cocrystal and (ii) the Lys is located 4.0 Å from the 5Ј-O of the scissile phosphodiester in a new DNA cocrystal of human to-poisomerase IB (9,19). Still we do not exclude the possibility that Arg-130 is the direct proton donor to the 5Ј-O leaving group; indeed an arginine is proposed to play such a role in the hydrolysis of phosphodiesters by staphylococcal nuclease (20). In the Flp-DNA cocrystal, the Lys is closer to the 5Ј-OH than is the Arg side chain, but the opposite situation holds for human topoisomerase IB where the Arg is within hydrogen binding distance (2.96 Å) of the 5Ј-O of the scissile phosphodiester (9,19).
Arg-130 is located between helices ␣2 and ␣3 of the catalytic domain; this segment is disordered in the crystal structure of free vaccinia topoisomerase and is sensitive to proteolysis in the free state in solution but resistant in the DNA-bound state (4,21). Thus, it is suggested that Arg-130 also undergoes a conformational change upon DNA binding that incorporates it into the active site. Although the Arg-130 equivalents are typically in contact with the scissile phosphate in the DNA cocrystal structures of topoisomerase IB or tyrosine recombinases, a recent structure of Cre recombinase bound to a Y-junction showed that the Arg-173 side chain (equivalent to vaccinia Arg-130) was rotated away from the scissile phosphate and that the ␤ loop containing Lys-201 (equivalent of vaccinia Lys-167) was disordered, i.e. there was no visible electron density (22). Taken together these observations suggest that the assembly of a catalytically competent active site in the DNAbound topoisomerase IB and recombinase enzymes may involve conformational changes of both Arg-130 and Lys-167.
There are several ways in which Arg-130 might assist proton donation by Lys-167: (i) Arg-130 could help lower the pK a of Lys-167 by contributing to a local positive electrostatic environment; (ii) Arg-130 could engage in a hydrogen bond to Lys-167; or (iii) Arg-130 could transfer a proton to Lys-167, conceivably coupled to its acquisition of a proton from an "upstream" component in a proton relay. The first model, based on purely electrostatic effects, does not account for the observed mutational and 5Ј-bridging phosphorothiolate effects on transesterification, insofar as the R130K substitution should have restored a positive electrostatic environment, yet the R130K mutant remains grossly defective in expulsion of the leaving DNA strand as gauged by the 2600-fold thiolate effect.
We speculate that the Arg-130 side chain contacts both the scissile phosphate and Lys-167 in the transition state. The former interaction would help neutralize the developing negative charge on the proposed pentacoordinate phosphorane transition state, while the latter would donate either a hydrogen bond (proton sharing) or fully transfer a proton to Lys-167 as it gives up a proton to the 5Ј-O of the leaving strand. A model of proton relay from Arg-130 to Lys-167 can account for the 5Јbridging phosphorothiolate effects observed for the R130K and K167A mutants, and it raises the prospect that Arg-130 in turn accepts a proton from an upstream donor, which either directly or indirectly leads to abstraction of a proton from the active site tyrosine.
An answer to the question of how the tyrosine is activated for nucleophilic attack on the scissile phosphodiester has been extremely elusive. A scheme has been proposed whereby a general base on the enzyme accepts a proton from the attacking tyrosine nucleophile during covalent adduct formation (23,24); however, the available structures of type IB topoisomerases provide no clues to the identity of a putative general base, i.e. there is no "conventional" proton-accepting side chain (e.g. His, Asp, or Glu) in the immediate vicinity of the tyrosine nucleophile that can be shown by mutagenesis to be functionally relevant. Comprehensive alanine scanning of His, Asp, Glu, and Cys residues of vaccinia topoisomerase has failed to uncover any other than His-265 that contributes significantly to Catalytic Mechanism of DNA Topoisomerase IB 5713 transesterification chemistry; moreover, it has been established by conservative mutagenesis that His-265 is not a general base catalyst (10). Alternatively it has been suggested based on the recent structural data for human topo IB that an ordered water molecule in the active site could act as a specific base to accept a proton from the Tyr O-4 atom (19). We speculate that Arg-130 might accept a proton from the active site tyrosine either directly or via an intervening water molecule as part of a wider proton relay network (conceivably Tyr-274 3 Arg-130 3 Lys-167 3 5Ј-O or Tyr-274 3 water 3 Arg-130 3 Lys-167 3 5Ј-O) that links in a concerted fashion the deprotonation of the attacking Tyr nucleophile and the protonation of the 5Ј-O leaving group during the cleavage reaction. The religation reaction, entailing attack of the 5Ј-OH of the noncovalently held DNA strand on the covalent DNA-3Ј-phosphotyrosyl-enzyme intermediate, would presumably entail a reversal of the proposed proton relay whereby Lys-167 accepts a proton from the 5Ј-OH, transfers its to Arg-130, and then (directly or via water or some other intermediary) to the oxygen of the Tyr-274 leaving group.