A Substitution in Rous Sarcoma Virus Integrase Tha
http://www.100md.com
病菌学杂志 2005年第8期
Department of Microbiology and Immunology
Department of Medicine, The Milton S. Hershey Medical Center, Pennsylvania State University College of Medicine, Hershey, Pennsylvania
ABSTRACT
Retroviral integrase prepares viral DNA for integration by removing 2 nucleotides from each end of unintegrated DNA in a reaction referred to as processing. However, it has been known since the processing assay was first described that avian integrases frequently nick 3 nucleotides, as well as 2 nucleotides, from viral DNA ends when reaction mixtures contain Mn2+. We now report that specificity for the biologically relevant "–2" site is enhanced when the serine at amino acid 124 of Rous sarcoma virus (RSV) integrase is replaced by alanine, valine, glycine, lysine, or aspartate. The protein with a serine-to-aspartate substitution exhibited especially high fidelity for the correct site, as evidenced by a ratio of –2 nicks to –3 nicks that was more than 40-fold greater than that for the wild-type enzyme in reactions with Mn2+. Even with Mg2+, the substituted proteins exhibited greater specificity than the wild type, especially the S124D protein. Moreover, this protein was more efficient than the wild type at processing viral DNA ends. Unexpectedly, however, the S124D protein was significantly impaired at catalyzing the insertion of viral DNA ends in reactions with Mn2+ and joining was undetectable in reactions with Mg2+. Thus, the S124D protein has separated the processing and joining activities of integrase. Similar results were found for human immunodeficiency virus integrase with the analogous substitution. No proteins with comparable properties have been described. Moreover, RSV virions containing integrase with the S124D mutation were unable to replicate in cell cultures. Together, these data suggest that integrase has evolved to have submaximal processing activity so that it can also catalyze DNA joining.
INTRODUCTION
Integrase is the viral enzyme responsible for inserting the reverse-transcribed DNA copy of the retroviral RNA genome into the chromosomal DNA of a recently infected cell. This recombination event is necessary for retrovirus replication and also makes retroviral infections permanent. To integrate retroviral DNA into cellular DNA, integrase must catalyze two ordered endonuclease reactions, both of which can be modeled in vitro by using purified enzymes and oligonucleotide substrates. First, integrase prepares the newly synthesized viral DNA for integration by cleaving the phosphodiester bond on the 3' side of conserved CA nucleotides that are near the 3' ends of all retroviral DNA; this site-specific nicking reaction shortens each DNA strand by 2 nucleotides and is referred to as processing (Fig. 1A) (17). Subsequently, integrase inserts the recessed 3'-OH groups at the two processed viral DNA ends into each strand of cellular DNA across a short stagger of a few base pairs; this sequence-independent nicking reaction is referred to as DNA joining or strand transfer (Fig. 1B models the joining reaction into one strand of DNA) (5, 14). These two assays were initially described by using enzymes from avian and murine retroviruses, but have since been used to study the integrases of a large number of retroviruses and retrotransposons, including human immunodeficiency virus type 1 (HIV-1) (16, 29). Other oligonucleotide-based assays have also been useful for studying the biochemistry of various integrases (Fig. 1C and D) (4, 20), but only the in vitro processing and joining reactions have obvious biological counterparts during retroviral integration in vivo.
Purified integrases generally exhibit greater processing and joining activity when reactions are conducted with Mn2+ as the divalent metal cofactor (25), even though Mg2+ is presumed to be the relevant metal cation in vivo. Interestingly, it has been known since the processing assay using avian myeloblastosis virus (AMV) integrase was first described that the avian enzymes frequently nick 3 nucleotides, as well as 2 nucleotides, from the viral DNA ends when reactions are conducted with Mn2+ (17). Under these conditions, individual substrates are nicked either between the conserved C and A nucleotides (an incorrect site) or 3' to the CA nucleotides (the correct site). Recently, however, we noticed that the related Rous sarcoma virus (RSV) integrase with a serine-to-alanine substitution at amino acid residue 124 had enhanced specificity for the biologically relevant site 2 nucleotides from the viral DNA ends (11). To investigate this observation further, we have now made additional mutations that alter the size or charge of the side chain at this position relative to alanine. In particular, we replaced serine-124 by amino acids that were larger or smaller than alanine (valine or glycine, respectively) or basic or acidic (lysine or aspartate, respectively). Unexpectedly, these studies revealed that placement of an acidic amino acid at this position yields an enzyme with improved processing activity but markedly impaired joining activity when either divalent cation is used. No protein that so clearly separates the processing and joining activities of integrase in this way has ever been described. Moreover, the biological relevance of these results was suggested by virus infection studies.
MATERIALS AND METHODS
Integrase coding sequences. The sequences that encode the RSV integrase used for these studies came from a derivative of plasmid pBH-RCAN-HiSV (6) and were amplified and cloned into plasmid pQE-30 (QIAGEN, Inc., Chatsworth, Calif.) as described previously (19, 30). The protein expressed from this cloned sequence is identical to that of the Schmidt-Ruppin D strain of RSV (GenBank accession number D10652) except that it has a valine in place of isoleucine at residue 209. Compared to the integrase of the Prague C strain of RSV (clone pATV-8; GenBank accession number J02342), there are five amino acid differences: alanine in place of valine at 101, lysine in place of arginine at 166, alanine in place of threonine at 182, and valine in place of isoleucine at 209 and 269. Site-directed mutations were created by overlap extension PCR (12), and all constructs were confirmed by sequencing the entire integrase-coding regions.
Protein expression and purification. Proteins were expressed in Escherichia coli M15(pREP4) (QIAGEN) by induction with isopropyl-?-D-thiogalactopyranoside. Full-length proteins were purified under native conditions from 10 ml of bacterial cultures by metal affinity chromatography using magnetic Ni2+-nitrilotriacetic acid (NTA) beads (QIAGEN), as described previously (15). Core domain fragments were similarly purified but from 250 ml of cultures with Ni2+-NTA agarose (19). The purified proteins were stored in a solution that contained 33 mM Tris-HCl (pH 7.6), 0.67 M NaCl, 0.7 mM dithioerythritol, 0.07 mM EDTA, 0.07% Triton X-100, and 40% glycerol. All of the proteins were readily purified under native conditions at relatively high concentrations (range, 4 to 10 μg/μl), as measured by comparison to a series of Coomassie blue-stained standards following sodium dodecyl sulfate-polyacrylamide gel electrophoresis and densitometry of dried gels. For the experiments shown in Fig. 2 and 3, all proteins were diluted 25-fold (to approximately 200 to 400 ng/μl) before being added to the assay mixtures, and for subsequent experiments equimolar amounts of the proteins were used.
Oligonucleotides. Processing assays used 18-mer oligonucleotides derived from the U3 end of RSV DNA (the sequence of the minus strand, which was radioactively labeled and for which the conserved CA bases are underlined, was 5'ATTGCATAAGACTACATT3'). The sequence of the plus strand of a similar substrate from the U5 end was 5'GAAGCAGAAGGCTTCATT3'. Some joining assays used a 5'-labeled, longer, preprocessed 30-mer derived from the terminal 32 bp of the U3 end (5'AAGACTACAAGAGTATTGCATAAGACTACA3'). The four oligonucleotides that were annealed for the disintegration assay were 5'GAGCTACGGATCCTCG3' (a nonviral 16-mer that was 5' labeled), 5'ATTGCATAAGACTACAAGCTCGAGGTCGACG3' (which modeled the final 16 nucleotides of processed viral U3 DNA, ending with the underlined CA, attached to 15 nucleotides of nonviral DNA), 5'CGTCGACCTCGAGCTCGAGGATCCGTAGCTC3' (the 31-nucleotide nonviral complementary strand), and 5'AATGTAGTCTTATGCAAT3' (the viral 18-mer complementary strand). The 23-mer of nonviral sequence for the nonspecific alcoholysis assay was 5'GAGACTACGTTCGAGGATCCGAG3'. All oligodeoxynucleotides used as assay substrates were gel purified following synthesis and again after being 5' end labeled with [-32P]ATP by T4 polynucleotide kinase, as described previously (19).
Integrase assays. Double-stranded DNA substrates for the processing, joining, and alcoholysis assays were prepared by annealing the labeled strand with a fourfold excess of unlabeled complementary oligonucleotide by sequentially incubating the DNA at 95°C for 5 min, 37°C for 30 min, and 4°C for at least 10 min. The substrate for the disintegration assay was prepared as described previously (18) except that the annealed complex was not gel purified because other experiments (not shown) verified that results were indistinguishable whether or not this step was included. Standard 10-μl reaction mixtures contained 0.5 pmol of double-stranded DNA (or 0.05 pmol of the disintegration substrate), 25 mM Tris-HCl (pH 8.0), 10 mM dithiothreitol, 10 mM MnCl2 or 5 mM MgCl2, and 1.0 μl of integrase or protein storage buffer. Reaction mixtures were incubated for 90 min at 37°C, and the reactions were then stopped by addition of 10 μl of loading buffer (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol) and heating at 95°C for 5 min. Aliquots were loaded onto 20% polyacrylamide (acrylamide-to-methylene-bisacrylamide ratio, 19:1)-7 M urea denaturing gels, followed by electrophoresis at 75 W until the bromophenol blue dye had migrated 28 cm (or 14 cm for the disintegration assay). Wet gels were autoradiographed at –70°C.
Quantitation of results. Results were quantified at the Hershey Medical Center Macromolecular Core Facility by measuring the radioactivity of bands in wet gels with a PhosphorImager (Molecular Dynamics, Sunnyvale, Calif.) or by measuring the intensities of bands on autoradiograms with a laser densitometer (Molecular Dynamics). To maximize the accuracy of quantitation, measurements were almost always made from gels in which the reaction lanes were separated by empty lanes. To calculate the yields of various products, phosphorimager or densitometry units for the entire gel lane were used as the denominator and corrections were made for appropriate control reactions without integrase or at zero time (25). Primary data (or logarithms in the case of ratios) were compared by using the Student's t test function included in the Microsoft Excel program (with a two-tailed distribution assuming equal variances). Results in graphs are presented as the means ± standard errors of replicate reactions.
Virus infections. Viruses were prepared from a derivative of the RCASBP(A) proviral construct in which the src gene was replaced by the gene encoding green fluorescent protein (kindly provided by Mark Federspiel of the Mayo Clinic College of Medicine) (28). This vector was modified by introducing a silent mutation to destroy a BglII site within the capsid region of the gag gene, and the modified construct was considered the wild type for our experiments. Mutations were introduced at integrase residue 124 by replacing the BglII/KpnI fragment of the modified construct with a similar fragment from the corresponding pQE-30-integrase plasmid. DH5 bacteria were transformed with the proviral constructs, and the integrase-coding regions of the plasmids purified from two colonies were sequenced to confirm that there were no other mutations in integrase; two clones for each mutant were used to reduce the possibility that any observed virus phenotype was due to extraneous mutations outside of integrase. QT6 quail cells (26), which were maintained in Ham's F-10 medium supplemented with 8% tryptose phosphate broth, 5% fetal bovine serum, 1% chicken serum, and antibiotics, were transfected with 15 μg of wild-type or mutant DNA by the calcium phosphate method. After 48 h, virus-containing supernatants were collected and clarified, and virus was quantified by measuring reverse transcriptase activity. DF-1 chicken cells (28), which were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and antibiotics, were infected with equal counts per minute of virus for 16 h, after which the medium was replaced. Cell cultures were passaged every 3 days, at which time the spread of infection was assessed by using an automated fluorescence-activated cell sorter (FACS) to measure the percentage of cells that were fluorescent.
RESULTS
Effects of amino acid substitutions on specificity of processing. The wild-type RSV integrase and all of the proteins with substitutions at residue 124 efficiently processed double-stranded 18-mer substrates designed to represent the U3 end of unintegrated RSV DNA, as demonstrated by the shortening of the 5'-labeled strand (Fig. 2A, lanes 2 to 7 and 9 to 14). However, as previously established, the wild-type RSV integrase created prominent 15-mer products in addition to 16-mer products when reactions were conducted in the presence of Mn2+ (Fig. 2A, lane 2). In contrast, the S124A protein exhibited greater selectivity for the biologically relevant "–2" site compared to the –3 site in reactions that used Mn2+ (Fig. 2A, lane 3), consistent with our previous observation (11) that was the basis for this project. In fact, all of the proteins with substitutions exhibited greater specificity than the wild-type protein for the –2 site when Mn2+ was used (Fig. 2A, lanes 3 to 7 compared to lane 2), especially that with the serine-to-aspartate substitution (lane 6). Quantitation of triplicate reactions that used Mn2+ (Fig. 2B, left graph) showed that the ratio of 16-mers to 15-mers averaged 1.4 for the wild-type enzyme; between 8 and 13 for the proteins with A, K, V, or G substitutions; and 57 for the S124D protein (P < 0.0002 for each log-transformed value versus that for the wild type). Stated another way, the proteins with A, K, V, or G substitutions were six to nine times more specific, and the S124D protein was greater than 40-fold more specific, than the wild-type integrase in reactions that used Mn2+.
When similar reactions were conducted with Mg2+ as the metal cofactor, the wild-type RSV integrase, as expected, exhibited greater specificity for the biologically relevant site after the CA nucleotides to yield a predominant 16-mer product (Fig. 2A, lane 9). However, even in reactions with Mg2+, all of the proteins with substitutions were more specific than the wild-type enzyme, as reflected by minimal amounts of 15-mer products (Fig. 2A, lanes 10 to 14 compared to lane 9). Quantitation of triplicate reactions that used Mg2+ (Fig. 2B, right graph) showed that the ratio of 16-mers to 15-mers averaged 16 for the wild-type enzyme, reflecting the greater specificity of RSV integrase with Mg2+ than with Mn2+. However, this ratio was further elevated to between 24 and 80 for the proteins with A, K, V, or G substitutions and averaged 228 for the S124D protein (P < 0.05 for each log-transformed value versus the that for the wild type). Stated another way, the proteins with A, K, V, or G substitutions were 1.5 to 5 times more specific, and the S124D protein averaged 14 times greater specificity, than the wild-type integrase in reactions that used Mg2+. Thus, all five proteins with amino acid substitutions at position 124, especially the S124D protein, had enhanced specificity for the biologically relevant site that follows the CA bases in reactions that used either metal cofactor.
Effects on efficiency of processing. To calculate the total amount of processing at the –2 site, we measured the amount of substrate converted to 16-mer or joined products in triplicate reactions; the joined products were included because they are known to derive from correctly processed viral DNA ends (5, 14). The data showed that the total amount of processing was increased for all of the proteins with substitutions, especially the S124D protein, in reactions that used Mn2+ (Fig. 3A, left graph). In particular, the wild-type protein correctly processed an average of 33% of the substrate, proteins with the A, K, V, or G substitutions averaged 47 to 55% processing, and the S124D protein processed 83% of the substrate (P < 0.006 for each substituted protein compared to the wild type). If the 15-mer products (i.e., nicks at the –3 site) are included in the calculations, the amount of processing by the wild-type protein would increase to 53% and amounts by proteins with the other substitutions would be similar to the wild type at 52 to 59% but the efficiency of the S124D protein would remain elevated at 84% (P < 0.001 for the S124D protein compared to the wild type).
Even with Mg2+, the S124D protein appeared to have increased efficiency for processing, as suggested by the lesser amount of remaining substrate in Fig. 2A, lane 13. Quantitation of triplicate reactions (Fig. 3A, right graph) showed that the amount of processing (including only the 16-mer and joined products) averaged 55% for the wild type and between 56 and 64% for the proteins with other substitutions but was 75% for the S124D protein; however, these differences were not statistically significant, and there was much overlap for the Mg2+-dependent reactions. It is important that no proteins were present at significantly higher concentration than the wild type in any of the experiments to this point. In particular, the protein/DNA molar ratios during these reactions were 23:1 for the wild type, 18:1 for the S124D protein, and between 10:1 and 24:1 for proteins with the other substitutions. Thus, we conclude that all of the proteins with substitutions were at least as efficient for processing as the wild type with either divalent cation and that the S124D protein was significantly more efficient with Mn2+.
Effects on DNA joining. Inspection of the upper parts (which are not shown) of the autoradiograms in the processing reactions described above suggested that the proteins with A, K, V, or G substitutions catalyzed joining at levels comparable to wild-type levels, especially in reactions with Mn2+. Unexpectedly, however, the amount of joining catalyzed by the S124D protein was consistently diminished in reactions with Mn2+ (Fig. 3B, left graph). Whereas with this metal the wild-type integrase converted 6% of substrate to joined products and the proteins with the other substitutions converted an average of 5 to 7% to products, the S124D protein converted only 1% of substrate to joined products (P < 0.002 for S124D compared to the wild type). This result cannot be attributed to differences in the amounts of processed ends available because, compared to the other proteins, the S124D protein created more processed ends (Fig. 3A, left graph), which should have been available for the subsequent joining reaction. Even more striking was the complete lack of any detectable joined products for the S124D protein in the reactions that used Mg2+ (Fig. 3B, right graph). The difference between the S124D and wild-type proteins was statistically significant (P < 0.05), even though the amount of joined products as a percentage of initial substrate typically is low in reactions that use Mg2+ (note the scale of the y axis in the right graph of Fig. 3B).
To confirm that the S124D protein was defective at joining, we used a preprocessed substrate representing the 32 bp the U3 end of RSV DNA but with the 2 nucleotides 3' to the conserved CA already removed. We also fixed the protein/DNA molar ratio at 8:1 for each protein. The results showed that, for reactions done with Mn2+, products greater than 30 nucleotides in length were evident at comparable amounts for the wild-type protein and other proteins (Fig. 4, lanes 1 to 5 and 7) but were clearly diminished for the S124D protein (lane 6). Moreover, in reactions with Mg2+, joined products were readily apparent with the wild-type protein and other proteins (Fig. 4, lanes 8 to 12 and 14) but were not detected for the S124D protein (lane 13). Quantitation of replicate reactions similar to those in Fig. 4 showed that, with Mn2+, the wild-type protein and other proteins converted 7 to 14% of preprocessed substrate to joined products whereas the S124D protein converted only 2% (P < 0.0003 for S124D versus the wild type). In contrast, in reactions with Mg2+, the wild-type protein and other proteins converted 0.2 to 0.7% of substrate to joined products but such products were not observed for the S124D protein (P < 0.007 for S124D versus the wild type). Thus, despite its enhanced processing activity, the S124D protein is significantly impaired for joining with either divalent metal cation.
The S124D protein acts similarly across a range of reaction conditions. To examine the unusual activity profile of the S124D protein further, we performed time course experiments at high and low enzyme/substrate ratios for the wild-type and S124D proteins (Fig. 5). We tested a protein/DNA molar ratio of 40:1 because there are an estimated 50 to 100 integrase molecules but only two ends of viral DNA per infectious virion in vivo. We also tested a 1:1 ratio because current evidence reveals only one active site per integrase monomer and whether any enzyme turnover occurs in vivo is unknown (13). The results showed that, in Mn2+-dependent reactions, the extent of processing by each protein, whether at the high or low protein/DNA ratio, reached a plateau by 30 min (Fig. 5A and B). At either protein/DNA ratio, however, the S124D protein was always more efficient than the wild type at processing viral DNA. Moreover, whether the relative amount of protein was high or low, the S124D protein had enhanced specificity for the –2 site (Fig. 5A, lanes 11 to 20), as confirmed by measuring the ratios of 16-mer to 15-mer products (Fig. 5C). The S124D protein also created fewer joined products than did the wild-type protein at either protein/DNA ratio (not shown), despite having created more processed DNA that should have been available for the joining reaction.
Interestingly, the wild-type integrase exhibited its greatest specificity for the –2 site at early time points, as reflected by ratios of 16-mers to 15-mers that were approximately 10 after 2 min of reaction time but fell to 2 by 30 min (Fig. 5C). This finding is consistent with the observation in the original description of the processing assay that the selectivity of the related AMV integrase for the –2 site in Mn2+-dependent reactions improved when the extent of the reaction was limited by shorter incubation times (17). The wild-type RSV protein also had improved specificity for the –2 site when the concentration of Mn2+ was less than 1 mM (data not shown). Nonetheless, the S124D protein was more specific than the wild-type protein across a range of Mn2+ concentrations from 0.1 to 40 mM even when reactions were limited to 5 min. Similarly, the S124D protein exhibited impaired joining activity compared to the wild-type protein across this 400-fold range of Mn2+ concentrations (not shown).
Time course experiments with Mg2+ showed that the extent of processing took longer to reach a plateau than it took with Mn2+ for both proteins, whether at the high or low protein/DNA ratio (Fig. 5D and E). At either ratio, however, the S124D protein was always more efficient at processing than was the wild-type protein. Moreover, although the wild-type protein was highly specific for the –2 site with Mg2+ (as expected), the S124D protein was even more specific. In fact, specificity could not be calculated for the S124D protein because it did not create any measurable 15-mer products in these experiments (Fig. 5D, lanes 7 to 12); the only detectable 15-mer products were created by the wild-type protein at the longest time point (lanes 3 and 6). In addition, joined products were not detected for the S124D protein in these or other reactions that used protein/DNA ratios from 40:1 to 1:1 or Mg2+ concentrations from 0.5 to 15 mM (data not shown).
Other activities of the S124D protein. Given the impaired joining by the S124D protein, we tested the proteins in a different assay that also reflects nonspecific nicking of nonviral DNA (20, 21). It was previously shown that various integrases can use water or other nucleophilic molecules, such as glycerol, to nick DNA at any internal site (Fig. 1C). Under the conditions used in this nonspecific alcoholysis assay, wild-type RSV integrase nicked more than one-half of the substrate but the S124D protein exhibited little DNA nicking (Fig. 6A, lanes 2 and 3, respectively). Quantitation of six reactions with three different nonviral DNA substrates showed that the wild-type protein nicked an average of 59% of substrate DNA, whereas the S124D protein averaged only 7% nicking (Fig. 6B; P < 0.00002 versus the wild type). For comparison, the proteins with A, K, V, or G substitutions averaged 22 to 53% nicking of these substrates (not shown). Thus, the S124D protein was particularly inefficient in this nonspecific nicking assay.
We also tested these proteins for disintegration activity, which reflects the ability of integrase to resolve a complex of four oligonucleotides that mimic the product of a joining reaction (Fig. 1D). The wild-type and S124D proteins were very active in this assay (Fig. 6C, lanes 2 and 3), as were the other proteins with substitutions (not shown). However, although the wild-type protein (and proteins with the other substitutions) created additional bands between the 16-mer substrate and 31-mer product (Fig. 6C, lane 2), the S124D protein yielded few or none of these additional bands (Fig. 6C, lane 3). These products have been shown to result from two mechanisms: (i) reinsertion of the processed viral DNA ends that are released by the disintegration reaction (33) and (ii) nonspecific nicking of the 31-mer disintegration product (22, 34). Thus, the minimal amounts of these additional products with the S124D protein provide further support for the conclusion that this protein is defective both for joining and for nonspecific DNA nicking.
Activities of core domain fragments. Because the central (or core) domain of integrase is sufficient to catalyze nonspecific alcoholysis and disintegration (2, 20), we purified protein fragments representing residues 52 to 207 of the RSV wild-type and S124D proteins. Under the conditions used to test for these activities, the central fragment of the wild-type protein acted similarly to the full-length protein in the nonspecific alcoholysis assay (Fig. 6A, lane 4) and in the disintegration assay (Fig. 6C, lane 4). In contrast, the core domain of the S124D protein did not nick the nonviral DNA substrate in the nonspecific nicking assay (Fig. 6A, lane 5). Moreover, the S124D core fragment exhibited negligible or no disintegration activity in repeated experiments (e.g., Fig. 6C, lane 5), despite the high activity of its full-length protein (lane 3). In addition to catalyzing nonspecific alcoholysis and disintegration, the core domain of RSV integrase was previously shown to have another nicking activity. In the presence of Mn2+, this protein fragment can nick viral DNA ends at several sites, especially 3 nucleotides from the end, but not at the important site 2 nucleotides from the end (22); thus, this action can be considered to reflect nonspecific nicking. We confirmed these observations with our preparation of the wild-type RSV IN52-207 (Fig. 6D, lane 2). However, the S124D fragment from 52 to 207 was inactive in this assay (Fig. 6D, lane 3), supporting the conclusion that the S124D mutation caused a defect in nonspecific nicking.
Viruses with the S124D mutation do not replicate. Given the unusual activity profile of the S124D protein, we examined the effect of this substitution on virus replication by cloning this mutation into an infectious clone of RSV that encodes green fluorescent protein. Supernatants of cells transfected by wild-type or S124D plasmids were equalized for reverse transcriptase activity and used to infect a susceptible chicken cell line. Spread of infection was monitored by FACS analysis of the cultures at each time point. The results showed that the wild-type virus rapidly spread through the culture, infecting essentially 100% of the cells by 10 days. In contrast, the virus containing integrase with the S124D mutation was unable to replicate (Fig. 7). Although the replication defect has not yet been localized to a particular step in the virus life cycle, extraneous defects at a step other than integration are less likely because other experiments indicate that RSV virions with the S124A integrase mutation replicate similarly to the wild type (11).
DISCUSSION
In vitro assays that model aspects of retroviral integration have been very useful for studying the mechanism of integrase and for preclinical screening of potential antiretroviral agents. When the oligonucleotide-based processing assay was described more than a decade ago (17), it was noted that AMV integrase frequently nicked between the conserved C and A nucleotides, as well as at the expected site after the CA nucleotides, when reaction mixtures contained Mn2+. This finding was consistent with previous data that the AMV and RSV integrases nicked both of these sites in the presence of Mn2+ when supercoiled DNA with covalently linked viral long terminal repeats was used as the substrate (9, 31), even though such molecules are not relevant to integration (1, 8). These observations may have become of less interest in light of the higher specificity for the correct site after the CA nucleotides when these enzymes are tested in the presence of Mg2+ (9, 17, 31). However, many purified integrases are inactive with Mg2+ and require Mn2+ for in vitro activity. A striking characteristic of the avian retroviral integrases is that they readily exhibit Mg2+-dependent activity without needing other reagents, such as dimethyl sulfoxide, to augment this activity (25). Nonetheless, the avian enzymes are more active with Mn2+ and are often studied with this cation, despite the frequent placement of nicks 1 nucleotide away from the biologically relevant site under these conditions.
Our recent work with patient-derived HIV-1 integrase variants identified residue 119 as having an important role in the selection of insertion sites in nonviral DNA (10). Unexpectedly, when extending this finding to other viruses, we noticed that RSV integrase with a serine-to-alanine substitution at the analogous residue 124 had enhanced specificity during processing (11). With this stimulus for the present report, we have now found that several substitutions at this residue improve specificity during processing using Mn2+ or Mg2+ (Fig. 2). In particular, the protein with aspartic acid at this position had especially high fidelity for the –2 site (Fig. 2) and was more efficient than the wild type at processing viral DNA ends (Fig. 3A). Even more unexpectedly, the S124D protein inefficiently catalyzed DNA joining in reactions that used Mn2+, and joined products were not detected in reactions that used Mg2+. The impaired joining was demonstrated with preprocessed substrates (Fig. 4), as well as with substrates that first had to be processed by integrase and theoretically would have interacted with the enzyme in a more natural fashion (Fig. 3B). Moreover, both the enhanced processing and impaired joining by the S124D protein were robust findings across a range of conditions, including varied Mn2+ and Mg2+ concentrations, protein/DNA ratios, and incubation times (Fig. 5).
Although the experiments reported here used DNA substrates derived from the U3 end of RSV DNA, which are known to be more susceptible to avian integrases than substrates from the U5 end (36), we also obtained very similar results with U5 substrates. In particular, the S124D protein had improved specificity and efficiency during processing but was impaired for joining (data not shown). The RSV S124D protein is the first enzyme to exhibit an activity profile that separates the processing and joining activities of integrase in this way. Although mutations at other positions in HIV or avian integrases were reported to diminish processing more than joining or joining more than processing and although some mutations may have affected only one of these activities (3, 23, 24, 27, 32), to our knowledge there are no reports of any mutations that catalyze one activity better than the wild type while being impaired for the other activity. Thus, the RSV S124D protein offers a new perspective for studying integrase.
Several mechanisms could explain the activity profile of the S124D protein. For example, its ability to catalyze DNA joining could be impaired because the complex that it makes with processed viral DNA is excessively tight (such that it cannot release viral DNA during joining) or unusually weak (such that it releases the viral DNA too soon). Other possibilities are that this mutation interferes with a conformation or oligomerization state needed to catalyze joining. However, none of these ideas would explain the observed defect for nonspecific alcoholysis (Fig. 6A and B). Indeed, the limited amount of nonviral DNA nicking by the S124D protein in this reaction, which does not use any viral DNA, suggests a problem in binding or orienting nonviral (i.e., cellular) DNA. This idea is very consistent with our recent work showing that the amino acid at this residue in RSV integrase (and the analogous residues in the HIV-1 and visna virus integrases) strongly influences which nonviral target DNA sites are chosen (11). That the S124D protein was impaired at creating bands between the substrate and major product in the disintegration assay (Fig. 6C) also indicates a defect in nonspecific nicking of nonviral DNA (22, 33, 34). The lack of nicking by the S124D core fragment on viral DNA ends, especially at the –3 site (Fig. 6D), also can be interpreted as a defect in nonspecific nicking (22). Moreover, it has been suggested that nicking at sites that are 3 nucleotides from each viral DNA end, as first demonstrated on substrates that contain linked copies of the viral long terminal repeats, may reflect the ability of avian integrases to make the 6-bp staggered nicks in cellular DNA that occur during RSV integration (9, 31). Thus, the inactivity of the S124D core fragment in the assay in Fig. 6D also can be interpreted as reflecting a defect in interacting with cellular DNA. A weaker interaction with cellular DNA also is supported by our preliminary data which indicate that joining and nonviral DNA nicking catalyzed by the S124D protein are more sensitive to increased NaCl concentrations than these reactions catalyzed by the wild-type protein (data not shown).
The improved processing by the S124D protein also likely reflects an altered interaction with DNA (but in this case with viral DNA) because we found no difference between this protein and the wild type in their interactions with the attacking nucleophilic substrate that is used to release the terminal two nucleotides (the ROH in Fig. 1A) (7, 35). In particular, all of the proteins with substitutions at residue 124 had the same preference as the wild type for using water as the nucleophile during Mg2+-dependent processing (to release linear dinucleotide products) and for using the 3' end of viral DNA as the nucleophile for Mn2+-dependent processing (to release cyclic dinucleotide products; data not shown). Although the improved processing by the S124D protein may reflect more-specific or tighter binding to viral DNA ends, we found that processing by the S124D protein and that by the wild-type protein were equally sensitive to increased salt concentrations (data not shown). Thus, further studies are needed to explain the activities of this protein. Useful information is also likely to be gained by further analysis of the discordance for the S124D protein in the nonspecific-alcoholysis assay (in which it had minimal activity; Fig. 6A and B) and the disintegration assay (in which it had high activity; Fig. 6C), because we have not found a discordance between these assays for other proteins. Similarly, understanding the discordance between the full-length S124D protein and its core domain in the disintegration assay (Fig. 6C) should also shed light on how integrase interacts with its DNA substrates.
Whatever mechanism underlies the enhanced specific nicking of viral DNA but impaired nonspecific nicking of nonviral DNA exhibited by the S124D protein, the acidic side chain at residue 124 is likely to be important. This conclusion is based on our finding that RSV integrase with glutamic acid at this position (i.e., S124E; data not shown) and HIV-1 integrase with an aspartic acid substitution at the analogous position 119 (i.e., S119D; Fig. 8) had the same activity profile as the RSV S124D protein. In particular, compared to wild-type HIV-1 integrase, the HIV-1 S119D protein had enhanced specificity during processing with Mn2+ or Mg2+ (Fig. 8A; note the greater amount of 16-mers versus 15-mers in lane 3 compared to lane 2 and in lane 6 compared to lane 5) and enhanced processing efficiency with Mn2+ (Fig. 8A, lane 3). Furthermore, the HIV-1 S119D protein was impaired for joining when either divalent cation was used, whether joining followed processing (Fig. 8A, upper part of the gel) or used preprocessed substrates (Fig. 8B, lane 3 compared to lane 2 and, even more strikingly with Mg2+, lane 6 compared to lane 5). Thus, an acidic side chain at this position has similar effects in two viral systems. That RSV virions containing the S124D mutation were replication defective also suggests the biological relevance of these findings. Moreover, these data suggest that integrase maintains such a delicate balance between its specific and nonspecific nicking activities that it has evolved to have submaximal processing activity so that it can also catalyze joining.
ACKNOWLEDGMENTS
This work was supported by a grant from the G. Harold and Leila Y. Mathers Charitable Foundation (to M.K.). S.B. was supported by the Penn State Huck Institutes of the Life Sciences Summer Undergraduate Research Program. M.K. is an MD Research Facilitation Award Scholar of the Penn State College of Medicine. This project was funded, in part, under a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds; the Department specifically disclaims responsibility for any analyses, interpretations or conclusions.
We thank Anna Marie Skalka and Richard Katz of the Fox Chase Cancer Center for suggesting the use of the assay shown in Fig. 6D and Vernon M. Chinchilli of the Department of Health Evaluation Sciences at the Penn State College of Medicine for helpful discussions about statistics.
REFERENCES
Brown, P. O., B. Bowerman, H. E. Varmus, and J. M. Bishop. 1987. Correct integration of retroviral DNA in vitro. Cell 49:347-356.
Bushman, F. D., A. Engelman, I. Palmer, P. Wingfield, and R. Craigie. 1993. Domains of the integrase protein of human immunodeficiency virus type 1 responsible for polynucleotidyl transfer and zinc binding. Proc. Natl. Acad. Sci. USA 90:3428-3432.
Calmels, C., V. R. de Soultrait, A. Caumont, C. Desjobert, A. Faure, M. Fournier, L. Tarrago-Litvak, and V. Parissi. 2004. Biochemical and random mutagenesis analysis of the region carrying the catalytic E152 amino acid of HIV-1 integrase. Nucleic Acids Res. 32:1527-1538.
Chow, S. A., K. A. Vincent, V. Ellison, and P. O. Brown. 1992. Reversal of integration and DNA splicing mediated by integrase of human immunodeficiency virus. Science 255:723-726.
Craigie, R., T. Fujiwara, and F. Bushman. 1990. The IN protein of Moloney murine leukemia virus processes the viral DNA ends and accomplishes their integration in vitro. Cell 62:829-837.
Dong, J., J. W. Dubay, L. G. Perez, and E. Hunter. 1992. Mutations within the proteolytic cleavage site of the Rous sarcoma virus glycoprotein define a requirement for dibasic residues for intracellular cleavage. J. Virol. 66:865-874.
Engelman, A., K. Mizuuchi, and R. Craigie. 1991. HIV-1 DNA integration: mechanism of viral DNA cleavage and DNA strand transfer. Cell 67:1211-1221.
Fujiwara, T., and K. Mizuuchi. 1988. Retroviral DNA integration: structure of an integration intermediate. Cell 54:497-504.
Grandgenett, D. P., and A. C. Vora. 1985. Site-specific nicking at the avian retrovirus LTR circle junction by the viral pp32 DNA endonuclease. Nucleic Acids Res. 13:6205-6221.
Harper, A. L., L. M. Skinner, M. Sudol, and M. Katzman. 2001. Use of patient-derived human immunodeficiency virus type 1 integrases to identify a protein residue that affects target site selection. J. Virol. 75:7756-7762.
Harper, A. L., M. Sudol, and M. Katzman. 2003. An amino acid in the central catalytic domain of three retroviral integrases that affects target site selection in nonviral DNA. J. Virol. 77:3838-3845.
Horton, R. M., S. N. Ho, J. K. Pullen, H. D. Hunt, Z. Cai, and L. R. Pease. 1993. Gene splicing by overlap extension. Methods Enzymol. 217:270-279.
Jones, K. S., J. Coleman, G. W. Merkel, T. M. Laue, and A. M. Skalka. 1992. Retroviral integrase functions as a multimer and can turn over catalytically. J. Biol. Chem. 267:16037-16040.
Katz, R. A., G. Merkel, J. Kulkosky, J. Leis, and A. M. Skalka. 1990. The avian retroviral IN protein is both necessary and sufficient for integrative recombination in vitro. Cell 63:87-95.
Katzman, M., A. L. Harper, M. Sudol, L. M. Skinner, and M. E. Eyster. 2001. Activity of natural HIV-1 integrases recovered from subjects with varied rates of disease progression. J. Acquir. Immune Defic. Syndr. 28:203-210.
Katzman, M., and R. A. Katz. 1999. Substrate recognition by retroviral integrases. Adv. Virus Res. 52:371-395.
Katzman, M., R. A. Katz, A. M. Skalka, and J. Leis. 1989. The avian retroviral integration protein cleaves the terminal sequences of linear viral DNA at the in vivo sites of integration. J. Virol. 63:5319-5327.
Katzman, M., and M. Sudol. 1994. In vitro activities of purified visna virus integrase. J. Virol. 68:3558-3569.
Katzman, M., and M. Sudol. 1995. Mapping domains of retroviral integrase responsible for viral DNA specificity and target site selection by analysis of chimeras between human immunodeficiency virus type 1 and visna virus integrases. J. Virol. 69:5687-5696.
Katzman, M., and M. Sudol. 1996. Nonspecific alcoholysis, a novel endonuclease activity of human immunodeficiency virus type 1 and other retroviral integrases. J. Virol. 70:2598-2604.
Katzman, M., M. Sudol, J. S. Pufnock, S. Zeto, and L. M. Skinner. 2000. Mapping target site selection for the non-specific nuclease activities of retroviral integrase. Virus Res. 66:87-100.
Kulkosky, J., R. A. Katz, G. Merkel, and A. M. Skalka. 1995. Activities and substrate specificity of the evolutionarily conserved central domain of retroviral integrase. Virology 206:448-456.
Leavitt, A. D., L. Shiue, and H. E. Varmus. 1993. Site-directed mutagenesis of HIV-1 integrase demonstrates differential effects on integrase functions in vitro. J. Biol. Chem. 268:2113-2119.
Moreau, K., C. Faure, G. Verdier, and C. Ronfort. 2002. Analysis of conserved and non-conserved amino acids critical for ALSV (avian leukemia and sarcoma viruses) integrase functions in vitro. Arch. Virol. 147:1761-1778.
Morgan, A. L., and M. Katzman. 2000. Subterminal viral DNA nucleotides as specific recognition signals for human immunodeficiency virus type 1 and visna virus integrases under magnesium-dependent conditions. J. Gen. Virol. 81:839-849.
Moscovici, C., M. G. Moscovici, H. Jimenez, M. M. C. Lai, M. J. Hayman, and P. K. Vogt. 1977. Continuous tissue culture cell lines derived from chemically induced tumors of Japanese quail. Cell 11:95-103.
Parissi, V., A. B. Caumont, V. R. de Soultrait, C. Calmels, S. Pichuantes, S. Litvak, and C. H. Dupont. 2000. Selection of amino acid substitutions restoring activity of HIV-1 integrase mutated in its catalytic site using the yeast Saccharomyces cerevisiae. J. Mol. Biol. 295:755-765.
Schaefer-Klein, J., I. Givol, E. V. Barsov, J. M. Whitcomb, M. VanBrocklin, D. N. Foster, M. J. Federspiel, and S. H. Hughes. 1998. The EV-O-derived cell line DF-1 supports the efficient replication of avian leukosis-sarcoma viruses and vectors. Virology 248:305-311.
Sherman, P. A., and J. A. Fyfe. 1990. Human immunodeficiency virus integration protein expressed in Escherichia coli possesses selective DNA cleaving activity. Proc. Natl. Acad. Sci. USA 87:5119-5123.
Skinner, L. M., M. Sudol, A. L. Harper, and M. Katzman. 2001. Nucleophile selection for the endonuclease activities of human, ovine, and avian retroviral integrases. J. Biol. Chem. 276:114-124.
Terry, R., D. A. Soltis, M. Katzman, D. Cobrinik, J. Leis, and A. M. Skalka. 1988. Properties of avian sarcoma-leukosis virus pp32-related pol-endonucleases produced in Escherichia coli. J. Virol. 62:2358-2365.
van Gent, D. C., A. A. M. Oude Groeneger, and R. H. A. Plasterk. 1992. Mutational analysis of the integrase protein of human immunodeficiency virus type 2. Proc. Natl. Acad. Sci. USA 89:9598-9602.
Vincent, K. A., V. Ellison, S. A. Chow, and P. O. Brown. 1993. Characterization of human immunodeficiency virus type 1 integrase expressed in Escherichia coli and analysis of variants with amino-terminal mutations. J. Virol. 67:425-437.
Vink, C., A. A. M. Oude Groeneger, and R. H. A. Plasterk. 1993. Identification of the catalytic and DNA-binding region of the human immunodeficiency virus type 1 integrase protein. Nucleic Acids Res. 21:1419-1425.
Vink, C., E. Yeheskiely, G. A. van der Marel, J. H. van Boom, and R. H. A. Plasterk. 1991. Site-specific hydrolysis and alcoholysis of human immunodeficiency virus DNA termini mediated by the viral integrase protein. Nucleic Acids Res. 19:6691-6698.
Vora, A. C., M. L. Fitzgerald, and D. P. Grandgenett. 1990. Removal of 3'-OH-terminal nucleotides from blunt-ended long terminal repeat termini by the avian retrovirus integration protein. J. Virol. 64:5656-5659.(Wesley M. Konsavage Jr., )
Department of Medicine, The Milton S. Hershey Medical Center, Pennsylvania State University College of Medicine, Hershey, Pennsylvania
ABSTRACT
Retroviral integrase prepares viral DNA for integration by removing 2 nucleotides from each end of unintegrated DNA in a reaction referred to as processing. However, it has been known since the processing assay was first described that avian integrases frequently nick 3 nucleotides, as well as 2 nucleotides, from viral DNA ends when reaction mixtures contain Mn2+. We now report that specificity for the biologically relevant "–2" site is enhanced when the serine at amino acid 124 of Rous sarcoma virus (RSV) integrase is replaced by alanine, valine, glycine, lysine, or aspartate. The protein with a serine-to-aspartate substitution exhibited especially high fidelity for the correct site, as evidenced by a ratio of –2 nicks to –3 nicks that was more than 40-fold greater than that for the wild-type enzyme in reactions with Mn2+. Even with Mg2+, the substituted proteins exhibited greater specificity than the wild type, especially the S124D protein. Moreover, this protein was more efficient than the wild type at processing viral DNA ends. Unexpectedly, however, the S124D protein was significantly impaired at catalyzing the insertion of viral DNA ends in reactions with Mn2+ and joining was undetectable in reactions with Mg2+. Thus, the S124D protein has separated the processing and joining activities of integrase. Similar results were found for human immunodeficiency virus integrase with the analogous substitution. No proteins with comparable properties have been described. Moreover, RSV virions containing integrase with the S124D mutation were unable to replicate in cell cultures. Together, these data suggest that integrase has evolved to have submaximal processing activity so that it can also catalyze DNA joining.
INTRODUCTION
Integrase is the viral enzyme responsible for inserting the reverse-transcribed DNA copy of the retroviral RNA genome into the chromosomal DNA of a recently infected cell. This recombination event is necessary for retrovirus replication and also makes retroviral infections permanent. To integrate retroviral DNA into cellular DNA, integrase must catalyze two ordered endonuclease reactions, both of which can be modeled in vitro by using purified enzymes and oligonucleotide substrates. First, integrase prepares the newly synthesized viral DNA for integration by cleaving the phosphodiester bond on the 3' side of conserved CA nucleotides that are near the 3' ends of all retroviral DNA; this site-specific nicking reaction shortens each DNA strand by 2 nucleotides and is referred to as processing (Fig. 1A) (17). Subsequently, integrase inserts the recessed 3'-OH groups at the two processed viral DNA ends into each strand of cellular DNA across a short stagger of a few base pairs; this sequence-independent nicking reaction is referred to as DNA joining or strand transfer (Fig. 1B models the joining reaction into one strand of DNA) (5, 14). These two assays were initially described by using enzymes from avian and murine retroviruses, but have since been used to study the integrases of a large number of retroviruses and retrotransposons, including human immunodeficiency virus type 1 (HIV-1) (16, 29). Other oligonucleotide-based assays have also been useful for studying the biochemistry of various integrases (Fig. 1C and D) (4, 20), but only the in vitro processing and joining reactions have obvious biological counterparts during retroviral integration in vivo.
Purified integrases generally exhibit greater processing and joining activity when reactions are conducted with Mn2+ as the divalent metal cofactor (25), even though Mg2+ is presumed to be the relevant metal cation in vivo. Interestingly, it has been known since the processing assay using avian myeloblastosis virus (AMV) integrase was first described that the avian enzymes frequently nick 3 nucleotides, as well as 2 nucleotides, from the viral DNA ends when reactions are conducted with Mn2+ (17). Under these conditions, individual substrates are nicked either between the conserved C and A nucleotides (an incorrect site) or 3' to the CA nucleotides (the correct site). Recently, however, we noticed that the related Rous sarcoma virus (RSV) integrase with a serine-to-alanine substitution at amino acid residue 124 had enhanced specificity for the biologically relevant site 2 nucleotides from the viral DNA ends (11). To investigate this observation further, we have now made additional mutations that alter the size or charge of the side chain at this position relative to alanine. In particular, we replaced serine-124 by amino acids that were larger or smaller than alanine (valine or glycine, respectively) or basic or acidic (lysine or aspartate, respectively). Unexpectedly, these studies revealed that placement of an acidic amino acid at this position yields an enzyme with improved processing activity but markedly impaired joining activity when either divalent cation is used. No protein that so clearly separates the processing and joining activities of integrase in this way has ever been described. Moreover, the biological relevance of these results was suggested by virus infection studies.
MATERIALS AND METHODS
Integrase coding sequences. The sequences that encode the RSV integrase used for these studies came from a derivative of plasmid pBH-RCAN-HiSV (6) and were amplified and cloned into plasmid pQE-30 (QIAGEN, Inc., Chatsworth, Calif.) as described previously (19, 30). The protein expressed from this cloned sequence is identical to that of the Schmidt-Ruppin D strain of RSV (GenBank accession number D10652) except that it has a valine in place of isoleucine at residue 209. Compared to the integrase of the Prague C strain of RSV (clone pATV-8; GenBank accession number J02342), there are five amino acid differences: alanine in place of valine at 101, lysine in place of arginine at 166, alanine in place of threonine at 182, and valine in place of isoleucine at 209 and 269. Site-directed mutations were created by overlap extension PCR (12), and all constructs were confirmed by sequencing the entire integrase-coding regions.
Protein expression and purification. Proteins were expressed in Escherichia coli M15(pREP4) (QIAGEN) by induction with isopropyl-?-D-thiogalactopyranoside. Full-length proteins were purified under native conditions from 10 ml of bacterial cultures by metal affinity chromatography using magnetic Ni2+-nitrilotriacetic acid (NTA) beads (QIAGEN), as described previously (15). Core domain fragments were similarly purified but from 250 ml of cultures with Ni2+-NTA agarose (19). The purified proteins were stored in a solution that contained 33 mM Tris-HCl (pH 7.6), 0.67 M NaCl, 0.7 mM dithioerythritol, 0.07 mM EDTA, 0.07% Triton X-100, and 40% glycerol. All of the proteins were readily purified under native conditions at relatively high concentrations (range, 4 to 10 μg/μl), as measured by comparison to a series of Coomassie blue-stained standards following sodium dodecyl sulfate-polyacrylamide gel electrophoresis and densitometry of dried gels. For the experiments shown in Fig. 2 and 3, all proteins were diluted 25-fold (to approximately 200 to 400 ng/μl) before being added to the assay mixtures, and for subsequent experiments equimolar amounts of the proteins were used.
Oligonucleotides. Processing assays used 18-mer oligonucleotides derived from the U3 end of RSV DNA (the sequence of the minus strand, which was radioactively labeled and for which the conserved CA bases are underlined, was 5'ATTGCATAAGACTACATT3'). The sequence of the plus strand of a similar substrate from the U5 end was 5'GAAGCAGAAGGCTTCATT3'. Some joining assays used a 5'-labeled, longer, preprocessed 30-mer derived from the terminal 32 bp of the U3 end (5'AAGACTACAAGAGTATTGCATAAGACTACA3'). The four oligonucleotides that were annealed for the disintegration assay were 5'GAGCTACGGATCCTCG3' (a nonviral 16-mer that was 5' labeled), 5'ATTGCATAAGACTACAAGCTCGAGGTCGACG3' (which modeled the final 16 nucleotides of processed viral U3 DNA, ending with the underlined CA, attached to 15 nucleotides of nonviral DNA), 5'CGTCGACCTCGAGCTCGAGGATCCGTAGCTC3' (the 31-nucleotide nonviral complementary strand), and 5'AATGTAGTCTTATGCAAT3' (the viral 18-mer complementary strand). The 23-mer of nonviral sequence for the nonspecific alcoholysis assay was 5'GAGACTACGTTCGAGGATCCGAG3'. All oligodeoxynucleotides used as assay substrates were gel purified following synthesis and again after being 5' end labeled with [-32P]ATP by T4 polynucleotide kinase, as described previously (19).
Integrase assays. Double-stranded DNA substrates for the processing, joining, and alcoholysis assays were prepared by annealing the labeled strand with a fourfold excess of unlabeled complementary oligonucleotide by sequentially incubating the DNA at 95°C for 5 min, 37°C for 30 min, and 4°C for at least 10 min. The substrate for the disintegration assay was prepared as described previously (18) except that the annealed complex was not gel purified because other experiments (not shown) verified that results were indistinguishable whether or not this step was included. Standard 10-μl reaction mixtures contained 0.5 pmol of double-stranded DNA (or 0.05 pmol of the disintegration substrate), 25 mM Tris-HCl (pH 8.0), 10 mM dithiothreitol, 10 mM MnCl2 or 5 mM MgCl2, and 1.0 μl of integrase or protein storage buffer. Reaction mixtures were incubated for 90 min at 37°C, and the reactions were then stopped by addition of 10 μl of loading buffer (95% formamide, 20 mM EDTA, 0.05% bromophenol blue, 0.05% xylene cyanol) and heating at 95°C for 5 min. Aliquots were loaded onto 20% polyacrylamide (acrylamide-to-methylene-bisacrylamide ratio, 19:1)-7 M urea denaturing gels, followed by electrophoresis at 75 W until the bromophenol blue dye had migrated 28 cm (or 14 cm for the disintegration assay). Wet gels were autoradiographed at –70°C.
Quantitation of results. Results were quantified at the Hershey Medical Center Macromolecular Core Facility by measuring the radioactivity of bands in wet gels with a PhosphorImager (Molecular Dynamics, Sunnyvale, Calif.) or by measuring the intensities of bands on autoradiograms with a laser densitometer (Molecular Dynamics). To maximize the accuracy of quantitation, measurements were almost always made from gels in which the reaction lanes were separated by empty lanes. To calculate the yields of various products, phosphorimager or densitometry units for the entire gel lane were used as the denominator and corrections were made for appropriate control reactions without integrase or at zero time (25). Primary data (or logarithms in the case of ratios) were compared by using the Student's t test function included in the Microsoft Excel program (with a two-tailed distribution assuming equal variances). Results in graphs are presented as the means ± standard errors of replicate reactions.
Virus infections. Viruses were prepared from a derivative of the RCASBP(A) proviral construct in which the src gene was replaced by the gene encoding green fluorescent protein (kindly provided by Mark Federspiel of the Mayo Clinic College of Medicine) (28). This vector was modified by introducing a silent mutation to destroy a BglII site within the capsid region of the gag gene, and the modified construct was considered the wild type for our experiments. Mutations were introduced at integrase residue 124 by replacing the BglII/KpnI fragment of the modified construct with a similar fragment from the corresponding pQE-30-integrase plasmid. DH5 bacteria were transformed with the proviral constructs, and the integrase-coding regions of the plasmids purified from two colonies were sequenced to confirm that there were no other mutations in integrase; two clones for each mutant were used to reduce the possibility that any observed virus phenotype was due to extraneous mutations outside of integrase. QT6 quail cells (26), which were maintained in Ham's F-10 medium supplemented with 8% tryptose phosphate broth, 5% fetal bovine serum, 1% chicken serum, and antibiotics, were transfected with 15 μg of wild-type or mutant DNA by the calcium phosphate method. After 48 h, virus-containing supernatants were collected and clarified, and virus was quantified by measuring reverse transcriptase activity. DF-1 chicken cells (28), which were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and antibiotics, were infected with equal counts per minute of virus for 16 h, after which the medium was replaced. Cell cultures were passaged every 3 days, at which time the spread of infection was assessed by using an automated fluorescence-activated cell sorter (FACS) to measure the percentage of cells that were fluorescent.
RESULTS
Effects of amino acid substitutions on specificity of processing. The wild-type RSV integrase and all of the proteins with substitutions at residue 124 efficiently processed double-stranded 18-mer substrates designed to represent the U3 end of unintegrated RSV DNA, as demonstrated by the shortening of the 5'-labeled strand (Fig. 2A, lanes 2 to 7 and 9 to 14). However, as previously established, the wild-type RSV integrase created prominent 15-mer products in addition to 16-mer products when reactions were conducted in the presence of Mn2+ (Fig. 2A, lane 2). In contrast, the S124A protein exhibited greater selectivity for the biologically relevant "–2" site compared to the –3 site in reactions that used Mn2+ (Fig. 2A, lane 3), consistent with our previous observation (11) that was the basis for this project. In fact, all of the proteins with substitutions exhibited greater specificity than the wild-type protein for the –2 site when Mn2+ was used (Fig. 2A, lanes 3 to 7 compared to lane 2), especially that with the serine-to-aspartate substitution (lane 6). Quantitation of triplicate reactions that used Mn2+ (Fig. 2B, left graph) showed that the ratio of 16-mers to 15-mers averaged 1.4 for the wild-type enzyme; between 8 and 13 for the proteins with A, K, V, or G substitutions; and 57 for the S124D protein (P < 0.0002 for each log-transformed value versus that for the wild type). Stated another way, the proteins with A, K, V, or G substitutions were six to nine times more specific, and the S124D protein was greater than 40-fold more specific, than the wild-type integrase in reactions that used Mn2+.
When similar reactions were conducted with Mg2+ as the metal cofactor, the wild-type RSV integrase, as expected, exhibited greater specificity for the biologically relevant site after the CA nucleotides to yield a predominant 16-mer product (Fig. 2A, lane 9). However, even in reactions with Mg2+, all of the proteins with substitutions were more specific than the wild-type enzyme, as reflected by minimal amounts of 15-mer products (Fig. 2A, lanes 10 to 14 compared to lane 9). Quantitation of triplicate reactions that used Mg2+ (Fig. 2B, right graph) showed that the ratio of 16-mers to 15-mers averaged 16 for the wild-type enzyme, reflecting the greater specificity of RSV integrase with Mg2+ than with Mn2+. However, this ratio was further elevated to between 24 and 80 for the proteins with A, K, V, or G substitutions and averaged 228 for the S124D protein (P < 0.05 for each log-transformed value versus the that for the wild type). Stated another way, the proteins with A, K, V, or G substitutions were 1.5 to 5 times more specific, and the S124D protein averaged 14 times greater specificity, than the wild-type integrase in reactions that used Mg2+. Thus, all five proteins with amino acid substitutions at position 124, especially the S124D protein, had enhanced specificity for the biologically relevant site that follows the CA bases in reactions that used either metal cofactor.
Effects on efficiency of processing. To calculate the total amount of processing at the –2 site, we measured the amount of substrate converted to 16-mer or joined products in triplicate reactions; the joined products were included because they are known to derive from correctly processed viral DNA ends (5, 14). The data showed that the total amount of processing was increased for all of the proteins with substitutions, especially the S124D protein, in reactions that used Mn2+ (Fig. 3A, left graph). In particular, the wild-type protein correctly processed an average of 33% of the substrate, proteins with the A, K, V, or G substitutions averaged 47 to 55% processing, and the S124D protein processed 83% of the substrate (P < 0.006 for each substituted protein compared to the wild type). If the 15-mer products (i.e., nicks at the –3 site) are included in the calculations, the amount of processing by the wild-type protein would increase to 53% and amounts by proteins with the other substitutions would be similar to the wild type at 52 to 59% but the efficiency of the S124D protein would remain elevated at 84% (P < 0.001 for the S124D protein compared to the wild type).
Even with Mg2+, the S124D protein appeared to have increased efficiency for processing, as suggested by the lesser amount of remaining substrate in Fig. 2A, lane 13. Quantitation of triplicate reactions (Fig. 3A, right graph) showed that the amount of processing (including only the 16-mer and joined products) averaged 55% for the wild type and between 56 and 64% for the proteins with other substitutions but was 75% for the S124D protein; however, these differences were not statistically significant, and there was much overlap for the Mg2+-dependent reactions. It is important that no proteins were present at significantly higher concentration than the wild type in any of the experiments to this point. In particular, the protein/DNA molar ratios during these reactions were 23:1 for the wild type, 18:1 for the S124D protein, and between 10:1 and 24:1 for proteins with the other substitutions. Thus, we conclude that all of the proteins with substitutions were at least as efficient for processing as the wild type with either divalent cation and that the S124D protein was significantly more efficient with Mn2+.
Effects on DNA joining. Inspection of the upper parts (which are not shown) of the autoradiograms in the processing reactions described above suggested that the proteins with A, K, V, or G substitutions catalyzed joining at levels comparable to wild-type levels, especially in reactions with Mn2+. Unexpectedly, however, the amount of joining catalyzed by the S124D protein was consistently diminished in reactions with Mn2+ (Fig. 3B, left graph). Whereas with this metal the wild-type integrase converted 6% of substrate to joined products and the proteins with the other substitutions converted an average of 5 to 7% to products, the S124D protein converted only 1% of substrate to joined products (P < 0.002 for S124D compared to the wild type). This result cannot be attributed to differences in the amounts of processed ends available because, compared to the other proteins, the S124D protein created more processed ends (Fig. 3A, left graph), which should have been available for the subsequent joining reaction. Even more striking was the complete lack of any detectable joined products for the S124D protein in the reactions that used Mg2+ (Fig. 3B, right graph). The difference between the S124D and wild-type proteins was statistically significant (P < 0.05), even though the amount of joined products as a percentage of initial substrate typically is low in reactions that use Mg2+ (note the scale of the y axis in the right graph of Fig. 3B).
To confirm that the S124D protein was defective at joining, we used a preprocessed substrate representing the 32 bp the U3 end of RSV DNA but with the 2 nucleotides 3' to the conserved CA already removed. We also fixed the protein/DNA molar ratio at 8:1 for each protein. The results showed that, for reactions done with Mn2+, products greater than 30 nucleotides in length were evident at comparable amounts for the wild-type protein and other proteins (Fig. 4, lanes 1 to 5 and 7) but were clearly diminished for the S124D protein (lane 6). Moreover, in reactions with Mg2+, joined products were readily apparent with the wild-type protein and other proteins (Fig. 4, lanes 8 to 12 and 14) but were not detected for the S124D protein (lane 13). Quantitation of replicate reactions similar to those in Fig. 4 showed that, with Mn2+, the wild-type protein and other proteins converted 7 to 14% of preprocessed substrate to joined products whereas the S124D protein converted only 2% (P < 0.0003 for S124D versus the wild type). In contrast, in reactions with Mg2+, the wild-type protein and other proteins converted 0.2 to 0.7% of substrate to joined products but such products were not observed for the S124D protein (P < 0.007 for S124D versus the wild type). Thus, despite its enhanced processing activity, the S124D protein is significantly impaired for joining with either divalent metal cation.
The S124D protein acts similarly across a range of reaction conditions. To examine the unusual activity profile of the S124D protein further, we performed time course experiments at high and low enzyme/substrate ratios for the wild-type and S124D proteins (Fig. 5). We tested a protein/DNA molar ratio of 40:1 because there are an estimated 50 to 100 integrase molecules but only two ends of viral DNA per infectious virion in vivo. We also tested a 1:1 ratio because current evidence reveals only one active site per integrase monomer and whether any enzyme turnover occurs in vivo is unknown (13). The results showed that, in Mn2+-dependent reactions, the extent of processing by each protein, whether at the high or low protein/DNA ratio, reached a plateau by 30 min (Fig. 5A and B). At either protein/DNA ratio, however, the S124D protein was always more efficient than the wild type at processing viral DNA. Moreover, whether the relative amount of protein was high or low, the S124D protein had enhanced specificity for the –2 site (Fig. 5A, lanes 11 to 20), as confirmed by measuring the ratios of 16-mer to 15-mer products (Fig. 5C). The S124D protein also created fewer joined products than did the wild-type protein at either protein/DNA ratio (not shown), despite having created more processed DNA that should have been available for the joining reaction.
Interestingly, the wild-type integrase exhibited its greatest specificity for the –2 site at early time points, as reflected by ratios of 16-mers to 15-mers that were approximately 10 after 2 min of reaction time but fell to 2 by 30 min (Fig. 5C). This finding is consistent with the observation in the original description of the processing assay that the selectivity of the related AMV integrase for the –2 site in Mn2+-dependent reactions improved when the extent of the reaction was limited by shorter incubation times (17). The wild-type RSV protein also had improved specificity for the –2 site when the concentration of Mn2+ was less than 1 mM (data not shown). Nonetheless, the S124D protein was more specific than the wild-type protein across a range of Mn2+ concentrations from 0.1 to 40 mM even when reactions were limited to 5 min. Similarly, the S124D protein exhibited impaired joining activity compared to the wild-type protein across this 400-fold range of Mn2+ concentrations (not shown).
Time course experiments with Mg2+ showed that the extent of processing took longer to reach a plateau than it took with Mn2+ for both proteins, whether at the high or low protein/DNA ratio (Fig. 5D and E). At either ratio, however, the S124D protein was always more efficient at processing than was the wild-type protein. Moreover, although the wild-type protein was highly specific for the –2 site with Mg2+ (as expected), the S124D protein was even more specific. In fact, specificity could not be calculated for the S124D protein because it did not create any measurable 15-mer products in these experiments (Fig. 5D, lanes 7 to 12); the only detectable 15-mer products were created by the wild-type protein at the longest time point (lanes 3 and 6). In addition, joined products were not detected for the S124D protein in these or other reactions that used protein/DNA ratios from 40:1 to 1:1 or Mg2+ concentrations from 0.5 to 15 mM (data not shown).
Other activities of the S124D protein. Given the impaired joining by the S124D protein, we tested the proteins in a different assay that also reflects nonspecific nicking of nonviral DNA (20, 21). It was previously shown that various integrases can use water or other nucleophilic molecules, such as glycerol, to nick DNA at any internal site (Fig. 1C). Under the conditions used in this nonspecific alcoholysis assay, wild-type RSV integrase nicked more than one-half of the substrate but the S124D protein exhibited little DNA nicking (Fig. 6A, lanes 2 and 3, respectively). Quantitation of six reactions with three different nonviral DNA substrates showed that the wild-type protein nicked an average of 59% of substrate DNA, whereas the S124D protein averaged only 7% nicking (Fig. 6B; P < 0.00002 versus the wild type). For comparison, the proteins with A, K, V, or G substitutions averaged 22 to 53% nicking of these substrates (not shown). Thus, the S124D protein was particularly inefficient in this nonspecific nicking assay.
We also tested these proteins for disintegration activity, which reflects the ability of integrase to resolve a complex of four oligonucleotides that mimic the product of a joining reaction (Fig. 1D). The wild-type and S124D proteins were very active in this assay (Fig. 6C, lanes 2 and 3), as were the other proteins with substitutions (not shown). However, although the wild-type protein (and proteins with the other substitutions) created additional bands between the 16-mer substrate and 31-mer product (Fig. 6C, lane 2), the S124D protein yielded few or none of these additional bands (Fig. 6C, lane 3). These products have been shown to result from two mechanisms: (i) reinsertion of the processed viral DNA ends that are released by the disintegration reaction (33) and (ii) nonspecific nicking of the 31-mer disintegration product (22, 34). Thus, the minimal amounts of these additional products with the S124D protein provide further support for the conclusion that this protein is defective both for joining and for nonspecific DNA nicking.
Activities of core domain fragments. Because the central (or core) domain of integrase is sufficient to catalyze nonspecific alcoholysis and disintegration (2, 20), we purified protein fragments representing residues 52 to 207 of the RSV wild-type and S124D proteins. Under the conditions used to test for these activities, the central fragment of the wild-type protein acted similarly to the full-length protein in the nonspecific alcoholysis assay (Fig. 6A, lane 4) and in the disintegration assay (Fig. 6C, lane 4). In contrast, the core domain of the S124D protein did not nick the nonviral DNA substrate in the nonspecific nicking assay (Fig. 6A, lane 5). Moreover, the S124D core fragment exhibited negligible or no disintegration activity in repeated experiments (e.g., Fig. 6C, lane 5), despite the high activity of its full-length protein (lane 3). In addition to catalyzing nonspecific alcoholysis and disintegration, the core domain of RSV integrase was previously shown to have another nicking activity. In the presence of Mn2+, this protein fragment can nick viral DNA ends at several sites, especially 3 nucleotides from the end, but not at the important site 2 nucleotides from the end (22); thus, this action can be considered to reflect nonspecific nicking. We confirmed these observations with our preparation of the wild-type RSV IN52-207 (Fig. 6D, lane 2). However, the S124D fragment from 52 to 207 was inactive in this assay (Fig. 6D, lane 3), supporting the conclusion that the S124D mutation caused a defect in nonspecific nicking.
Viruses with the S124D mutation do not replicate. Given the unusual activity profile of the S124D protein, we examined the effect of this substitution on virus replication by cloning this mutation into an infectious clone of RSV that encodes green fluorescent protein. Supernatants of cells transfected by wild-type or S124D plasmids were equalized for reverse transcriptase activity and used to infect a susceptible chicken cell line. Spread of infection was monitored by FACS analysis of the cultures at each time point. The results showed that the wild-type virus rapidly spread through the culture, infecting essentially 100% of the cells by 10 days. In contrast, the virus containing integrase with the S124D mutation was unable to replicate (Fig. 7). Although the replication defect has not yet been localized to a particular step in the virus life cycle, extraneous defects at a step other than integration are less likely because other experiments indicate that RSV virions with the S124A integrase mutation replicate similarly to the wild type (11).
DISCUSSION
In vitro assays that model aspects of retroviral integration have been very useful for studying the mechanism of integrase and for preclinical screening of potential antiretroviral agents. When the oligonucleotide-based processing assay was described more than a decade ago (17), it was noted that AMV integrase frequently nicked between the conserved C and A nucleotides, as well as at the expected site after the CA nucleotides, when reaction mixtures contained Mn2+. This finding was consistent with previous data that the AMV and RSV integrases nicked both of these sites in the presence of Mn2+ when supercoiled DNA with covalently linked viral long terminal repeats was used as the substrate (9, 31), even though such molecules are not relevant to integration (1, 8). These observations may have become of less interest in light of the higher specificity for the correct site after the CA nucleotides when these enzymes are tested in the presence of Mg2+ (9, 17, 31). However, many purified integrases are inactive with Mg2+ and require Mn2+ for in vitro activity. A striking characteristic of the avian retroviral integrases is that they readily exhibit Mg2+-dependent activity without needing other reagents, such as dimethyl sulfoxide, to augment this activity (25). Nonetheless, the avian enzymes are more active with Mn2+ and are often studied with this cation, despite the frequent placement of nicks 1 nucleotide away from the biologically relevant site under these conditions.
Our recent work with patient-derived HIV-1 integrase variants identified residue 119 as having an important role in the selection of insertion sites in nonviral DNA (10). Unexpectedly, when extending this finding to other viruses, we noticed that RSV integrase with a serine-to-alanine substitution at the analogous residue 124 had enhanced specificity during processing (11). With this stimulus for the present report, we have now found that several substitutions at this residue improve specificity during processing using Mn2+ or Mg2+ (Fig. 2). In particular, the protein with aspartic acid at this position had especially high fidelity for the –2 site (Fig. 2) and was more efficient than the wild type at processing viral DNA ends (Fig. 3A). Even more unexpectedly, the S124D protein inefficiently catalyzed DNA joining in reactions that used Mn2+, and joined products were not detected in reactions that used Mg2+. The impaired joining was demonstrated with preprocessed substrates (Fig. 4), as well as with substrates that first had to be processed by integrase and theoretically would have interacted with the enzyme in a more natural fashion (Fig. 3B). Moreover, both the enhanced processing and impaired joining by the S124D protein were robust findings across a range of conditions, including varied Mn2+ and Mg2+ concentrations, protein/DNA ratios, and incubation times (Fig. 5).
Although the experiments reported here used DNA substrates derived from the U3 end of RSV DNA, which are known to be more susceptible to avian integrases than substrates from the U5 end (36), we also obtained very similar results with U5 substrates. In particular, the S124D protein had improved specificity and efficiency during processing but was impaired for joining (data not shown). The RSV S124D protein is the first enzyme to exhibit an activity profile that separates the processing and joining activities of integrase in this way. Although mutations at other positions in HIV or avian integrases were reported to diminish processing more than joining or joining more than processing and although some mutations may have affected only one of these activities (3, 23, 24, 27, 32), to our knowledge there are no reports of any mutations that catalyze one activity better than the wild type while being impaired for the other activity. Thus, the RSV S124D protein offers a new perspective for studying integrase.
Several mechanisms could explain the activity profile of the S124D protein. For example, its ability to catalyze DNA joining could be impaired because the complex that it makes with processed viral DNA is excessively tight (such that it cannot release viral DNA during joining) or unusually weak (such that it releases the viral DNA too soon). Other possibilities are that this mutation interferes with a conformation or oligomerization state needed to catalyze joining. However, none of these ideas would explain the observed defect for nonspecific alcoholysis (Fig. 6A and B). Indeed, the limited amount of nonviral DNA nicking by the S124D protein in this reaction, which does not use any viral DNA, suggests a problem in binding or orienting nonviral (i.e., cellular) DNA. This idea is very consistent with our recent work showing that the amino acid at this residue in RSV integrase (and the analogous residues in the HIV-1 and visna virus integrases) strongly influences which nonviral target DNA sites are chosen (11). That the S124D protein was impaired at creating bands between the substrate and major product in the disintegration assay (Fig. 6C) also indicates a defect in nonspecific nicking of nonviral DNA (22, 33, 34). The lack of nicking by the S124D core fragment on viral DNA ends, especially at the –3 site (Fig. 6D), also can be interpreted as a defect in nonspecific nicking (22). Moreover, it has been suggested that nicking at sites that are 3 nucleotides from each viral DNA end, as first demonstrated on substrates that contain linked copies of the viral long terminal repeats, may reflect the ability of avian integrases to make the 6-bp staggered nicks in cellular DNA that occur during RSV integration (9, 31). Thus, the inactivity of the S124D core fragment in the assay in Fig. 6D also can be interpreted as reflecting a defect in interacting with cellular DNA. A weaker interaction with cellular DNA also is supported by our preliminary data which indicate that joining and nonviral DNA nicking catalyzed by the S124D protein are more sensitive to increased NaCl concentrations than these reactions catalyzed by the wild-type protein (data not shown).
The improved processing by the S124D protein also likely reflects an altered interaction with DNA (but in this case with viral DNA) because we found no difference between this protein and the wild type in their interactions with the attacking nucleophilic substrate that is used to release the terminal two nucleotides (the ROH in Fig. 1A) (7, 35). In particular, all of the proteins with substitutions at residue 124 had the same preference as the wild type for using water as the nucleophile during Mg2+-dependent processing (to release linear dinucleotide products) and for using the 3' end of viral DNA as the nucleophile for Mn2+-dependent processing (to release cyclic dinucleotide products; data not shown). Although the improved processing by the S124D protein may reflect more-specific or tighter binding to viral DNA ends, we found that processing by the S124D protein and that by the wild-type protein were equally sensitive to increased salt concentrations (data not shown). Thus, further studies are needed to explain the activities of this protein. Useful information is also likely to be gained by further analysis of the discordance for the S124D protein in the nonspecific-alcoholysis assay (in which it had minimal activity; Fig. 6A and B) and the disintegration assay (in which it had high activity; Fig. 6C), because we have not found a discordance between these assays for other proteins. Similarly, understanding the discordance between the full-length S124D protein and its core domain in the disintegration assay (Fig. 6C) should also shed light on how integrase interacts with its DNA substrates.
Whatever mechanism underlies the enhanced specific nicking of viral DNA but impaired nonspecific nicking of nonviral DNA exhibited by the S124D protein, the acidic side chain at residue 124 is likely to be important. This conclusion is based on our finding that RSV integrase with glutamic acid at this position (i.e., S124E; data not shown) and HIV-1 integrase with an aspartic acid substitution at the analogous position 119 (i.e., S119D; Fig. 8) had the same activity profile as the RSV S124D protein. In particular, compared to wild-type HIV-1 integrase, the HIV-1 S119D protein had enhanced specificity during processing with Mn2+ or Mg2+ (Fig. 8A; note the greater amount of 16-mers versus 15-mers in lane 3 compared to lane 2 and in lane 6 compared to lane 5) and enhanced processing efficiency with Mn2+ (Fig. 8A, lane 3). Furthermore, the HIV-1 S119D protein was impaired for joining when either divalent cation was used, whether joining followed processing (Fig. 8A, upper part of the gel) or used preprocessed substrates (Fig. 8B, lane 3 compared to lane 2 and, even more strikingly with Mg2+, lane 6 compared to lane 5). Thus, an acidic side chain at this position has similar effects in two viral systems. That RSV virions containing the S124D mutation were replication defective also suggests the biological relevance of these findings. Moreover, these data suggest that integrase maintains such a delicate balance between its specific and nonspecific nicking activities that it has evolved to have submaximal processing activity so that it can also catalyze joining.
ACKNOWLEDGMENTS
This work was supported by a grant from the G. Harold and Leila Y. Mathers Charitable Foundation (to M.K.). S.B. was supported by the Penn State Huck Institutes of the Life Sciences Summer Undergraduate Research Program. M.K. is an MD Research Facilitation Award Scholar of the Penn State College of Medicine. This project was funded, in part, under a grant with the Pennsylvania Department of Health using Tobacco Settlement Funds; the Department specifically disclaims responsibility for any analyses, interpretations or conclusions.
We thank Anna Marie Skalka and Richard Katz of the Fox Chase Cancer Center for suggesting the use of the assay shown in Fig. 6D and Vernon M. Chinchilli of the Department of Health Evaluation Sciences at the Penn State College of Medicine for helpful discussions about statistics.
REFERENCES
Brown, P. O., B. Bowerman, H. E. Varmus, and J. M. Bishop. 1987. Correct integration of retroviral DNA in vitro. Cell 49:347-356.
Bushman, F. D., A. Engelman, I. Palmer, P. Wingfield, and R. Craigie. 1993. Domains of the integrase protein of human immunodeficiency virus type 1 responsible for polynucleotidyl transfer and zinc binding. Proc. Natl. Acad. Sci. USA 90:3428-3432.
Calmels, C., V. R. de Soultrait, A. Caumont, C. Desjobert, A. Faure, M. Fournier, L. Tarrago-Litvak, and V. Parissi. 2004. Biochemical and random mutagenesis analysis of the region carrying the catalytic E152 amino acid of HIV-1 integrase. Nucleic Acids Res. 32:1527-1538.
Chow, S. A., K. A. Vincent, V. Ellison, and P. O. Brown. 1992. Reversal of integration and DNA splicing mediated by integrase of human immunodeficiency virus. Science 255:723-726.
Craigie, R., T. Fujiwara, and F. Bushman. 1990. The IN protein of Moloney murine leukemia virus processes the viral DNA ends and accomplishes their integration in vitro. Cell 62:829-837.
Dong, J., J. W. Dubay, L. G. Perez, and E. Hunter. 1992. Mutations within the proteolytic cleavage site of the Rous sarcoma virus glycoprotein define a requirement for dibasic residues for intracellular cleavage. J. Virol. 66:865-874.
Engelman, A., K. Mizuuchi, and R. Craigie. 1991. HIV-1 DNA integration: mechanism of viral DNA cleavage and DNA strand transfer. Cell 67:1211-1221.
Fujiwara, T., and K. Mizuuchi. 1988. Retroviral DNA integration: structure of an integration intermediate. Cell 54:497-504.
Grandgenett, D. P., and A. C. Vora. 1985. Site-specific nicking at the avian retrovirus LTR circle junction by the viral pp32 DNA endonuclease. Nucleic Acids Res. 13:6205-6221.
Harper, A. L., L. M. Skinner, M. Sudol, and M. Katzman. 2001. Use of patient-derived human immunodeficiency virus type 1 integrases to identify a protein residue that affects target site selection. J. Virol. 75:7756-7762.
Harper, A. L., M. Sudol, and M. Katzman. 2003. An amino acid in the central catalytic domain of three retroviral integrases that affects target site selection in nonviral DNA. J. Virol. 77:3838-3845.
Horton, R. M., S. N. Ho, J. K. Pullen, H. D. Hunt, Z. Cai, and L. R. Pease. 1993. Gene splicing by overlap extension. Methods Enzymol. 217:270-279.
Jones, K. S., J. Coleman, G. W. Merkel, T. M. Laue, and A. M. Skalka. 1992. Retroviral integrase functions as a multimer and can turn over catalytically. J. Biol. Chem. 267:16037-16040.
Katz, R. A., G. Merkel, J. Kulkosky, J. Leis, and A. M. Skalka. 1990. The avian retroviral IN protein is both necessary and sufficient for integrative recombination in vitro. Cell 63:87-95.
Katzman, M., A. L. Harper, M. Sudol, L. M. Skinner, and M. E. Eyster. 2001. Activity of natural HIV-1 integrases recovered from subjects with varied rates of disease progression. J. Acquir. Immune Defic. Syndr. 28:203-210.
Katzman, M., and R. A. Katz. 1999. Substrate recognition by retroviral integrases. Adv. Virus Res. 52:371-395.
Katzman, M., R. A. Katz, A. M. Skalka, and J. Leis. 1989. The avian retroviral integration protein cleaves the terminal sequences of linear viral DNA at the in vivo sites of integration. J. Virol. 63:5319-5327.
Katzman, M., and M. Sudol. 1994. In vitro activities of purified visna virus integrase. J. Virol. 68:3558-3569.
Katzman, M., and M. Sudol. 1995. Mapping domains of retroviral integrase responsible for viral DNA specificity and target site selection by analysis of chimeras between human immunodeficiency virus type 1 and visna virus integrases. J. Virol. 69:5687-5696.
Katzman, M., and M. Sudol. 1996. Nonspecific alcoholysis, a novel endonuclease activity of human immunodeficiency virus type 1 and other retroviral integrases. J. Virol. 70:2598-2604.
Katzman, M., M. Sudol, J. S. Pufnock, S. Zeto, and L. M. Skinner. 2000. Mapping target site selection for the non-specific nuclease activities of retroviral integrase. Virus Res. 66:87-100.
Kulkosky, J., R. A. Katz, G. Merkel, and A. M. Skalka. 1995. Activities and substrate specificity of the evolutionarily conserved central domain of retroviral integrase. Virology 206:448-456.
Leavitt, A. D., L. Shiue, and H. E. Varmus. 1993. Site-directed mutagenesis of HIV-1 integrase demonstrates differential effects on integrase functions in vitro. J. Biol. Chem. 268:2113-2119.
Moreau, K., C. Faure, G. Verdier, and C. Ronfort. 2002. Analysis of conserved and non-conserved amino acids critical for ALSV (avian leukemia and sarcoma viruses) integrase functions in vitro. Arch. Virol. 147:1761-1778.
Morgan, A. L., and M. Katzman. 2000. Subterminal viral DNA nucleotides as specific recognition signals for human immunodeficiency virus type 1 and visna virus integrases under magnesium-dependent conditions. J. Gen. Virol. 81:839-849.
Moscovici, C., M. G. Moscovici, H. Jimenez, M. M. C. Lai, M. J. Hayman, and P. K. Vogt. 1977. Continuous tissue culture cell lines derived from chemically induced tumors of Japanese quail. Cell 11:95-103.
Parissi, V., A. B. Caumont, V. R. de Soultrait, C. Calmels, S. Pichuantes, S. Litvak, and C. H. Dupont. 2000. Selection of amino acid substitutions restoring activity of HIV-1 integrase mutated in its catalytic site using the yeast Saccharomyces cerevisiae. J. Mol. Biol. 295:755-765.
Schaefer-Klein, J., I. Givol, E. V. Barsov, J. M. Whitcomb, M. VanBrocklin, D. N. Foster, M. J. Federspiel, and S. H. Hughes. 1998. The EV-O-derived cell line DF-1 supports the efficient replication of avian leukosis-sarcoma viruses and vectors. Virology 248:305-311.
Sherman, P. A., and J. A. Fyfe. 1990. Human immunodeficiency virus integration protein expressed in Escherichia coli possesses selective DNA cleaving activity. Proc. Natl. Acad. Sci. USA 87:5119-5123.
Skinner, L. M., M. Sudol, A. L. Harper, and M. Katzman. 2001. Nucleophile selection for the endonuclease activities of human, ovine, and avian retroviral integrases. J. Biol. Chem. 276:114-124.
Terry, R., D. A. Soltis, M. Katzman, D. Cobrinik, J. Leis, and A. M. Skalka. 1988. Properties of avian sarcoma-leukosis virus pp32-related pol-endonucleases produced in Escherichia coli. J. Virol. 62:2358-2365.
van Gent, D. C., A. A. M. Oude Groeneger, and R. H. A. Plasterk. 1992. Mutational analysis of the integrase protein of human immunodeficiency virus type 2. Proc. Natl. Acad. Sci. USA 89:9598-9602.
Vincent, K. A., V. Ellison, S. A. Chow, and P. O. Brown. 1993. Characterization of human immunodeficiency virus type 1 integrase expressed in Escherichia coli and analysis of variants with amino-terminal mutations. J. Virol. 67:425-437.
Vink, C., A. A. M. Oude Groeneger, and R. H. A. Plasterk. 1993. Identification of the catalytic and DNA-binding region of the human immunodeficiency virus type 1 integrase protein. Nucleic Acids Res. 21:1419-1425.
Vink, C., E. Yeheskiely, G. A. van der Marel, J. H. van Boom, and R. H. A. Plasterk. 1991. Site-specific hydrolysis and alcoholysis of human immunodeficiency virus DNA termini mediated by the viral integrase protein. Nucleic Acids Res. 19:6691-6698.
Vora, A. C., M. L. Fitzgerald, and D. P. Grandgenett. 1990. Removal of 3'-OH-terminal nucleotides from blunt-ended long terminal repeat termini by the avian retrovirus integration protein. J. Virol. 64:5656-5659.(Wesley M. Konsavage Jr., )