Purification of the pets factor. A nuclear protein that binds to the inducible TG-rich element of the human immunodeficiency virus type 2 enhancer.

The peri-ets (pets) site is a TG-rich element found immediately adjacent to two binding sites for the ets family member Elf-1 in the human immunodeficiency virus type 2 (HIV-2) enhancer. Enhancer activation in response to T cell stimulation by phorbol myristate acetate, phytohemagglutinin, soluble or cross-linked antibodies to the T cell receptor, or antigen is mediated through this site in conjunction with its two adjacent Elf-1 binding sites, PuB1 and PuB2, and a κB site. Site-specific mutation of the pets element significantly reduces inducible activation of this enhancer but does not affect its transactivation by HIV-2 tat or other viral transactivators. Similar TG-rich sequences adjacent to ets-binding sites have also been found to be functionally important in the human T-cell leukemia virus type I and murine Moloney leukemia virus enhancers. As the cellular factor binding to the pets site plays a significant role in regulating the HIV-2 enhancer in both T cells and monocytes, we have purified this protein from bovine spleens and demonstrate that it is 43 kDa in size. In addition, using glycerol gradient centrifugation, Southwestern blotting, electrophoretic mobility shift assays employing purified protein eluted from a gel, and a new in solution UV cross-linking competitive assay, we show that the dominant protein binding to the pets site is 43 kDa in size. These results indicate that a nuclear protein of 43 kDa binds specifically to the pets site of the HIV-2 enhancer and may mediate transcriptional activation of this important human pathogen in response to T cell stimulation. As retroviruses generally expropriate important human regulatory proteins for their own use, the 43-kDa pets factor is also likely to play a significant role in signal transduction in T cells and in other cellular processes.

The peri-ets (pets) site is a TG-rich element found immediately adjacent to two binding sites for the ets family member Elf-1 in the human immunodeficiency virus type 2 (HIV-2) enhancer. Enhancer activation in response to T cell stimulation by phorbol myristate acetate, phytohemagglutinin, soluble or cross-linked antibodies to the T cell receptor, or antigen is mediated through this site in conjunction with its two adjacent Elf-1 binding sites, PuB1 and PuB2, and a B site. Sitespecific mutation of the pets element significantly reduces inducible activation of this enhancer but does not affect its transactivation by HIV-2 tat or other viral transactivators. Similar TG-rich sequences adjacent to ets-binding sites have also been found to be functionally important in the human T-cell leukemia virus type I and murine Moloney leukemia virus enhancers. As the cellular factor binding to the pets site plays a significant role in regulating the HIV-2 enhancer in both T cells and monocytes, we have purified this protein from bovine spleens and demonstrate that it is 43 kDa in size. In addition, using glycerol gradient centrifugation, Southwestern blotting, electrophoretic mobility shift assays employing purified protein eluted from a gel, and a new in solution UV cross-linking competitive assay, we show that the dominant protein binding to the pets site is 43 kDa in size. These results indicate that a nuclear protein of 43 kDa binds specifically to the pets site of the HIV-2 enhancer and may mediate transcriptional activation of this important human pathogen in response to T cell stimulation. As retroviruses generally expropriate important human regulatory proteins for their own use, the 43-kDa pets factor is also likely to play a significant role in signal transduction in T cells and in other cellular processes.
The human immunodeficiency virus type 2 (HIV-2) 1 is found in West Africa and, increasingly, in other parts of the world (1).
Like HIV-1, HIV-2 can cause AIDS, but typically the asymptomatic period following HIV-2 infection is much longer than that following HIV-1 infection. Genetically, HIV-2 is quite different from HIV-1 and only shares ϳ40% nucleic acid sequence similarity (1,2). Different transcriptional enhancer elements which respond to T cell stimulation in the enhancers of these two viruses may explain, in part, the clinical differences observed between persons infected with HIV-1 and HIV-2. Unlike HIV-1, in which the two B sites play the dominant role in regulating inducible enhancer function in activated T cells (3)(4)(5)(6), HIV-2 enhancer activation in T cells is regulated by at least four distinct cis-acting elements: two purine-rich sites (PuB1 and PuB2), which bind the ets proto-oncogene family member Elf-1; the peri-ets, or pets site, which binds a protein described below; and a single B site, which binds the well described components of NF-B (3,(7)(8)(9)(10)(11). A fifth element, the peri-B site, mediates HIV-2 enhancer induction in monocytic cells but not in T cells (12). While these five sites are not present in HIV-1, they are conserved in a wide range of HIV-2 isolates. Mutation of any of these elements markedly diminishes the response of the enhancer to cellular activation but does not affect the response to tat or to other viral transactivating proteins. Therefore, these elements specifically mediate enhancer stimulation in activated T cells or monocytes and serve as the final common mediators in signal transduction pathways.
The pets site is a TG-rich element (TTGGTCAGGG) found between the two Elf-1 binding sites in the HIV-2 enhancer (Fig.  1). We have demonstrated the functional importance of this site in the activation of the HIV-2 enhancer in both T cells and monocytes (8,9). The pets element mediates activation whether the enhancer is stimulated by phorbol myristate acetate alone, phytohemagglutinin alone, phorbol myristate acetate plus phytohemagglutinin, soluble antibodies to the T cell receptor, immobilized antibodies to the T cell receptor, or by antigen (3,(7)(8)(9)(10)(11). We and others have also demonstrated that a similar TG-rich pets-like site, again adjacent to an Elf-1 binding site, plays a significant role in mediating activation of the human T cell leukemia virus type 1 (HTLV-1) enhancer in stimulated T cells (13,14). TG-rich sites adjacent to the etsbinding sites are also found in murine retroviruses, and alteration of these pets-like sites can change the type of malignancy seen in mice following infection (15). Therefore pets-like elements adjacent to ets-binding sites appear to play an important role in the enhancers of retroviruses and serve as a final common mediator for signal transduction pathways in monocytes and especially T cells.
In this report, we describe the biochemical purification of the pets factor and show that it is a 43-kDa protein. This conclusion is based on glycerol gradient sedimentation, Southwestern blotting, DNase footprinting using highly purified extracts obtained through chromatographic separation, and on a newly developed competitive UV cross-linking assay. To characterize the native molecular mass of the pets factor, we purified the protein from bovine spleen nuclear extracts using DNA affinity chromatography and separated the purified protein using glycerol gradient ultracentrifugation. Bovine spleen was selected because a high level of pets binding activity was found in this abundantly available, inexpensive tissue. In addition, preliminary data, confirmed in the experiments described below, showed the bovine and human pets factors to be the same size and hence, presumably, the same protein. SDS-PAGE and electrophoretic mobility shift assay (EMSA) analysis of the glycerol gradient fractions revealed that the pets factor co-sedimented with ovalbumin, a 43-kDa internal control standard. We also performed UV cross-linking in solution using both crude bovine spleen nuclear extracts and highly purified pets factor to show that the predominant protein binding specifically to the pets site in vitro is a 43-kDa protein. Southwestern blotting experiments show that in cultured human T cells, as well as bovine spleen nuclear extracts, the predominant protein binding specifically to the pets site is 43 kDa in molecular mass. Having extensively characterized the molecular mass of the pets factor, we purified it to homogeneity from bovine spleen and showed that the purified pets factor, which is capable of protecting the HIV-2 pets site in a DNase I footprinting assay, is a 43-kDa protein. Furthermore, the 43-kDa purified protein, after band extraction and elution from SDS-PAGE, is capable of site specific binding to the pets element upon renaturation. Taken together, these results, from many different types of experimental assays, indicate that a specific cellular protein of 43 kDa is the major protein which binds to the pets site of the HIV-2 enhancer. This protein is therefore likely to be the pets factor, which mediates transcriptional activation and signal transduction pathways.

EXPERIMENTAL PROCEDURES
Preparation of Bovine Spleen Nuclear Extracts-Fresh bovine spleens (ϳ1 kg each) were obtained from a slaughter house and approximately one-half of one spleen was used per purification run. The following steps were performed at 4°C: the spleens were finely diced and homogenized in a blender for 30 s on highest speed in 4 liters of buffer H (250 mM sucrose, 50 mM HEPES, pH 6.5, 25 mM KCl, 10 mM ␤-mercaptoethanol, 5 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride). The nuclei and connective tissue were collected by centrifugation at 2,000 ϫ g for 10 min. The nuclear proteins were then extracted by hard shaking in 1.6 liters of buffer E (250 mM sucrose, 400 mM NaCl, 50 mM HEPES, pH 7.5, 10 mM ␤-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride) followed by incubation for 30 min. The nuclear extract was then clarified by centrifugation at 12,000 ϫ g for 20 min. Proteins in the supernatant were precipitated by the gradual addition of ammonium sulfate to 50% saturation. The precipitate was collected by centrifugation at 10,000 ϫ g for 30 min and stored in aliquots at Ϫ70°C.
Chromatographic Purification-The following steps were performed at 4°C: for each column run, the ammonium sulfate precipitate from half of a spleen was resuspended in buffer A (20 mM Tris-HCl, pH 7.0, 100 mM KCl, 1 mM EDTA, 0.01% Nonidet P-40, 10% glycerol, 0.5 mM DTT, 0.5 mM phenylmethylsulfonyl fluoride) and dialyzed overnight against 100 times its volume of buffer A. The dialysate was centrifuged at 10,000 ϫ g for 60 min, and successively filtered through 2.7-and 1.6-m glass fiber filters (Whatman) to clarify the protein solution before loading onto a column packed with 250 ml of Q-Sepharose Fast Flow resin (Pharmacia) according to the manufacturer's guidelines. Following loading, the column was washed with 2 liters of buffer A or until the absorbance at 280 nm returned to baseline. Proteins bound to the column were eluted with 10 bed volumes of a 100 mM KCl stepwise gradient. The binding activity, as measured by EMSA, typically eluted at 400 -500 mM KCl. Positive fractions were then pooled and dialyzed against buffer A before loading onto a 40-ml heparin-agarose (Bio-Rad) column. The procedures for subsequent washing and elution were the same as for the Q-Sepharose chromatography. The binding activity elutes at around 300 mM KCl. DNA affinity chromatography was performed using a DNA-Sepharose column and the method of Kadonaga and Tjian (16), with the pets oligonucleotide used in EMSAs. The positive fractions eluted from the heparin column were pooled, dialyzed against buffer A, and loaded onto a 2-ml DNA affinity column at gravity flow. During the first pass over the DNA affinity column, 12 g of calf thymus DNA was added as a nonspecific competitor per mg of protein applied to the column. The column was washed with 20 ml of buffer A and bound proteins were eluted with 10 ml of buffer A containing 1 M KCl. The eluate was diluted to 100 mM KCl in buffer A and reapplied to a newly prepared 1-ml DNA affinity column. To precipitate the eluted proteins, trichloroacetic acid was added to 10% final concentration in the presence of 200 g/ml sodium deoxycholate. After incubation on ice for 60 min, the precipitate was collected by centrifugation at 13,000 ϫ g for 10 min and washed once with acetone before it was resolubilized in 100 l of 3 M guanidine HCl (or 6 M urea for SDS-PAGE), 0.05 M Tris, pH 8.5, 25 mM DTT. 5 l were analyzed on SDS-PAGE, and the remainder was injected onto a C 4 reverse phase column (Vydac). Proteins were eluted using a 0 to 100% acetonitrile gradient in 0.1% trifluoroacetic acid and the 220-nm absorbance peaks were collected and analyzed by silver staining after SDS-PAGE.
Glycerol Gradient Sedimentation-Density gradient centrifugation was performed according to the method of Martin and Ames (17), with the following modifications. After 2 successive rounds of DNA affinity chromatography, 100 l of the eluate was mixed with 50 g each of lysozyme, bovine serum albumin, ovalbumin, and alcohol dehydrogenase and layered onto a 7-ml 10 -25% glycerol gradient in buffer A without DTT. After 21 h of centrifugation at 45,000 rpm in a Beckman SW-41 rotor, a Beckman fractionator was used to collect 40 aliquots. Each fraction was assayed for binding activity (EMSA) and silver staining after SDS-PAGE. The specific binding activity was measured by counting the 32 P cpm in the EMSA shifted bands using a ␤-scanner (Betagen).
DNase I Protection Assay-Jurkat nuclear extract was prepared by a modification of the method of Dignam et al. (20). Purified pets factor was prepared as described above. A radiolabeled probe corresponding to nucleotides Ϫ107 to Ϫ189 (Fig. 1) was prepared, and DNase I protection assays were performed as described previously (3,8,21,22). Purified proteins were concentrated in Centricons (Amicon) before use in DNase I footprinting.
Electrophoretic Mobility Shift Assays-Oligonucleotides corresponding to both strands of the HIV-2 sequence from Ϫ162 to Ϫ131 but with the PuB2 site mutated (Fig. 1) were synthesized with an Applied Biosystems 300B synthesizer. The sequence of the pets site oligonucleotide used in EMSA was: 5Ј-GATCCAGCTATACTTGGTCAGGGCGAA-TTCTAACTA. The sequence of the mutated pets oligonucleotide was: 5Ј-GATCCAGCTATACTAGATCTGGGCGAATTCTAACTA. Equimolar amounts of the two strands were combined, boiled for 1 min in 0.5 M NaCl, and allowed to cool gradually. The double-stranded oligonucleotide was then radiolabeled with T 4 polynucleotide kinase in the pres- ence of [␥-32 P]ATP and unincorporated nucleotide was then removed using a Sephadex G-50 column. The binding reactions were performed as described previously (8).
Southwestern Hybridization-200 g of Jurkat or bovine extract was separated by 8% SDS-PAGE and electroblotted onto polyvinylidine difluoride membrane with a semi-dry electroblotter (Fisher Scientific) in 16 mM Tris, 120 mM glycine, and 20% methanol. The blot was blocked in buffer A containing 5% dry milk (Carnation) at room temperature for 60 min with gentle shaking before incubation in buffer A containing 6 M urea and 50 mM DTT for 60 min at room temperature. The blot was then incubated overnight at 4°C in buffer A and was then reblocked as above for 30 min, rinsed, and hybridized to an 8x-pets concatemeric probe (see below) at 1 ϫ 10 6 cpm/ml in buffer A with a range of concentrations of calf thymus DNA (Fig. 6) for 60 min at room temperature. To remove excess probe, the blot was washed for 30 min at room temperature with 3 changes of 100 ml of buffer A before autoradiography.
Renaturation of Gel-purified pets Factor-Purified proteins (see above) were precipitated and separated by SDS-PAGE and the region corresponding to the 43-kDa prestained molecular mass marker was excised. The 43-kDa protein was extracted from the gel slice and renatured by the method of Hager and Burgess (23), except that the protein was renatured by dialysis against buffer A on a 0.025-m filter (Millipore).
Photo-affinity Cross-linking Assay-UV cross-linking was performed in solution with crude spleen extracts or purified pets factor using a bromouridine-substituted probe as described elsewhere (24) with modifications. UV cross-linking was performed on ice for 40 min. DNA polymerase I (Klenow fragment) and T 4 DNA polymerase (New England Biolabs) were also cross-linked to the same probe in other reactions and then used in the assay to estimate the effect of the cross-linked probe on migration through SDS-PAGE of a protein of known size. Specific and nonspecific oligonucleotide competitors were added to each reaction as indicated in Fig. 3. When purified proteins were used, poly(dI-dC) was excluded from the binding reactions.
Construction of a Multimeric pets Plasmid-All enzymes used for cloning were purchased from New England Biolabs. The pets site oligonucleotide used for EMSA is flanked by BamHI and BglII ends. A BglII restriction site was inserted into Puc 18 immediately 3Ј to the BamHI cutting site. The 1x-pets oligonucleotide was inserted into this plasmid at the BamHI-BglII sites so that restriction with BamHI and BglII yielded a fragment 36 base pairs long. This 1x clone was cut with XmnI and BamHI or XmnI and BglII and fragments were gel purified before being ligated to make a 2x-pets clone. This procedure was repeated to obtain clones as large as a 32x-pets construct. All clones were verified by DNA sequencing. To generate a 32 P-labeled probe for South-western hybridization, polymerase chain reaction primers flanking the multimeric insert were synthesized and used to amplify from a linearized 8x-pets construct using standard polymerase chain reaction procedures but with the inclusion of 5 l of [␣-32 P]dCTP (Amersham, 6000 Ci/mmol, 10 mCi/ml). Unincorporated nucleotides were removed after polymerase chain reaction, using Ultrafree-MC filter units (Millipore).

Determination of the Molecular Mass of the pets Factor-
Bovine spleen nuclear extracts were found to have a high level of specific binding activity for the HIV-2 pets site, 2 and was chosen as a starting material for purification because of the large quantity that was readily available. To estimate the native molecular mass of the pets factor, purified pets factor (see below) was applied to a glycerol gradient and protein complexes were separated on the basis of their sedimentation coefficients using ultracentrifugation. Individual fractions were collected and assayed for sequence-specific DNA binding activity using EMSA. In our glycerol gradient sedimentation experiments, we have included four internal standards in each tube: alcohol dehydrogenase (150 kDa), bovine serum albumin (68 kDa), ovalbumin (43 kDa), and lysozyme (17 kDa). The location of these standards in the glycerol gradient fractions was determined by SDS-PAGE of 10 l of each fraction followed by silver staining. The fraction that contained the peak DNA binding activity was also the fraction with the peak ovalbumin concentration (Fig. 2, lane 14).
As determination of molecular mass by glycerol gradient ultracentrifugation may be misleading due to aggregation of the protein of interest with itself or co-purified proteins, we used crude spleen extracts in UV cross-linking studies to find out the molecular mass of the pets factor under conditions where all nuclear proteins are present. Proteins which are capable of close contact with the bromouridine pets probe were cross-linked to the 32 P-labeled probe with UV irradiation. The protein-DNA complexes were then separated by SDS-PAGE to estimate the molecular mass of the protein-DNA complex. Since many nuclear proteins in the crude extract will exhibit nonspecific affinity for the labeled DNA, we improved the specificity of this often used technique by including poly(dI-dC) at 1 g/reaction and oligonucleotide competitors as indicated in Fig.  3. As expected, crude Jurkat nuclear extracts contain numerous proteins that cross-link to the pets probe (lane 3). Inclusion of 500 M excess of an unlabeled pets oligonucleotide competed out two bands: one indicated by the arrow in Fig. 3, and the other at 70 kDa (lane 4). However, inclusion of a 500 M excess of an oligonucleotide with an unrelated mutant HIV-2 B (lane 6) or wild-type HIV-2 B (lane 7) oligonucleotide did not compete away the two bands. The mutant pets oligonucleotide, in which only three nucleotides have been changed from the wildtype sequence, only weakly competed with the pets probe for binding (lane 5). In lane 8, purified bovine pets factor (after one round of DNA affinity chromatography) was used in the crosslinking reaction and a more highly purified fraction (after two rounds of DNA affinity chromatography) of bovine pets factor was used in lane 9. The cross-linking bands created by the purified pets factor appear to be identical in mobility to the two bands that competed out specifically in the experiments using crude extracts. As the effect of the cross-linked pets probe on protein migration in an SDS-PAGE was not known, we exploited the affinity for DNA of two commercially available enzymes to estimate this effect. DNA polymerase I (Klenow fragment, New England Biolabs) and T 4 DNA polymerase (Boehringer Mannheim) were cross-linked to the pets site probe (lanes 1 and 2). From these two DNA-binding proteins, the effect of a cross-linked pets probe on protein mobility in SDS-PAGE was estimated to be an addition of about 10 kDa to the proteins apparent molecular mass, assuming that the dynamics of DNA-protein cross-linking did not vary considerably. From these estimates, the dominant band from the purified protein represents a protein of about 43 kDa. A lighter band at 70 kDa capable of specific binding was observed in the experiments using crude spleen nuclear extracts. This band was also seen in lanes using highly purified pets factor (Fig. 3, lanes 8  and 9). As Southwestern blotting experiments using purified pets factor (Fig. 6C) did not show any additional bands other than the 43-kDa protein, this suggests that either some other protein may be cross-linked to the pets probe through its affinity for the 43-kDa protein or that two or more proteins that do not bind to the pets site by themselves dimerize to bind to the pets site. The improvements we made to the commonly used UV cross-linking assay gives this technique the additional ability to test the sequence specificity of proteins binding to DNA, as well as a means to estimate the effect of a particular DNA probe in mobility of protein-DNA complexes through SDS-PAGE.
Purification of the pets Factor-The pets factor was purified from bovine spleen using a five-step purification scheme (Fig.  4). Initial studies done with EMSA using nuclear extracts from mice and bovine heart, lung, thymus, and brain tissues indicated that a high pets-site specific binding activity was detected in spleen extracts (not shown). We therefore obtained spleens from freshly slaughtered cows and processed them as described under "Experimental Procedures." A Q-Sepharose Fast Flow column was chosen as the first chromatography step because of its high binding capacity and high flow rates. For each chromatography procedure used, a KCl salt gradient was used to elute the bound proteins from the column. Two l from each fraction was used in EMSA to test for the presence of the pets factor. Two rounds of DNA-affinity chromatography purified the pets factor more than 3000-fold over crude extract with an overall yield of 4.7% (Table I). Using approximately 6000 times less protein than that used with crude Jurkat nuclear extracts, the purified pets extracts bound to the HIV-2 pets site in DNase I footprinting assays (Fig. 5, lane A). Of note, the pets footprint using either crude or purified extract extended somewhat more 5Ј than in our previous experiments ( Fig. 1 and Ref.  Table I. While this scheme represents a typical purification run, sometimes the column sizes were increased (see "Experimental Procedures") or decreased depending on the amount of starting material available. 8). The 3Ј extension of the footprint obtained with the crude Jurkat extract (lane J) corresponds to the PuB2 site, which our previous work has shown binds Elf-1 (7,9). Silver staining of the highly purified extract shows only one dark staining band at 43 kDa (Fig. 7A, lanes 3-6), although other much lighter staining bands were also detected.

The 43-kDa Polypeptide Is Sufficient for Binding to the pets
Site-To determine whether or not the purified 43-kDa protein itself was capable of binding to the pets site, we performed Southwestern blotting experiments. Our initial concatemeric pets site probe generated by ligation of oligonucleotides containing a single pets-binding site yielded inconsistent results in Southwestern blotting experiments and we were unable to detect any definite binding. In contrast to Southwestern experiments using a probe generated through ligation in vitro of 1 ϫ oligonucleotides, we found that an 8 ϫ concatemeric probe with the binding sites all cloned in the same orientation gave more well defined bands and much more consistent results from one experiment to another. This may be because a "flip-flop" orientation of the binding sites can lead to conflicts in protein-DNA interactions. Southwestern experiments performed using this probe and crude Jurkat nuclear extracts (Fig. 6A), or crude spleen extracts (Fig. 6B), demonstrate that the molecular mass of the protein binding specifically to the pets site in the presence of increasing amounts of a nonspecific competitor is a 43-kDa protein. Southwestern blotting of highly purified pets factor also shows binding to a protein of 43 kDa (Fig. 6C). Thus, both bovine and human pets factor are 43 kDa and are likely to be the same protein.
To further purify the 43-kDa protein and confirm its ability to bind to the pets site, we injected proteins eluted from the final DNA affinity column onto an HPLC reverse phase column. The protein eluted (peak absorbance) from the reverse phase column was separated by SDS-PAGE and the band corresponding to 43 kDa was extracted from the gel slice. After a denaturation-renaturation step, this 43-kDa protein was shown to bind to the pets probe (Fig. 7B, lane 1) and can be competed out by the addition of unlabeled pets oligonucleotide but not by the unrelated HIV-2 B oligonucleotide (lane 2 versus lane 3).

DISCUSSION
In this report, we describe the purification of a 43-kDa cellular factor that binds to an element of the HIV-2 enhancer that we have previously shown to be functionally important in mediating transcriptional activation in response to T cell stimulation (8,10,11). We have also shown using glycerol gradient ultracentrifugation, DNase I footprinting, UV cross-linking, and Southwestern blotting that, in both crude bovine spleen FIG. 5. Purified pets factor protects the HIV-2 pets site in DNase footprinting studies. The lane labeled "G" is a GϩA ladder generated from the probe, which corresponds to nucleotides Ϫ107 to Ϫ189 (Fig. 1). The lane labeled "F" shows the result of DNase digestion when no nuclear protein is present. The lane labeled "A" shows the footprint obtained when purified pets factor is added. The lane labeled "J" demonstrates the footprint obtained when crude Jurkat T cell nuclear extract is added. Approximately 6000 times more protein is added in lane J than is added in lane A. The radioactive probe used was labeled on the non-coding strand and, therefore, the more 5Ј the sequence, the closer to the top of the gel it is found. The pets footprint is indicated by the brackets. extracts and Jurkat nuclear extracts, the dominant protein binding to the pets site in vitro is 43 kDa in molecular mass. Several proteins of other apparent molecular masses were seen in our less purified fractions, but at much lower concentrations, and it is possible that these may eventually turn out to contribute to the regulation of gene expression mediated by the pets element. The data we present here, based on renaturation after protein elution from SDS-PAGE, DNase I footprinting, Southwestern hybridization, glycerol gradient ultracentrifugation, reverse phase HPLC, and UV cross-linking, taken together point to the 43-kDa polypeptide as the pets-site specific DNA-binding protein. This 43-kDa protein is also seen using purification schemes that slightly differ from the scheme described above. The use of a Superose FPLC (Pharmacia) or a Mono-Q column (Pharmacia) also results in the purification of a 43-kDa protein. 2 In this work, we have made improvements to the widely used UV cross-linking and Southwestern blotting techniques. To our knowledge, UV cross-linking studies so far described have not included the use of competitors to test for DNA binding speci-ficity. Traditional UV cross-linking studies also do not include an internal control to estimate the effect of a particular bound probe on protein mobility in SDS-PAGE. We describe here improvements to the method which we have used successfully for our studies on the HIV-2 pets site. During the course of our Southwestern blotting experiments, we noticed that the previously described means of generating an effective concatemeric probe for hybridization to immobilized proteins separated by SDS-PAGE produced inconsistent results. We find that concatemerization through ligation of single copy oligonucleotides to each other often did not produce a satisfactory probe because of the conflicting flip-flop orientation of the binding site. While it is possible to synthesize a multimeric oligonucleotide, this alternative is not cost effective. For our Southwestern blotting experiments, we have constructed clones containing the petsbinding site in a head-to-tail orientation, so that all binding sites are in the same orientation. The insert can be excised and end-labeled, or used as a template for polymerase chain reaction-mediated labeling reactions, but probes generated from random-primed labeling gave rise to much higher backgrounds, presumably due to the heterogeneity of labeled probe. We have also found that probes up to 16x-pets proved to be extremely sensitive for Southwestern blotting, but an 8x-pets probe was sufficient for most experiments. In addition, we have also used another multimeric construct containing a different DNA-binding site to successfully show the applicability of this approach to other DNA-binding proteins as well. 3 Using the improvements described above, we demonstrate that the bovine protein binding specifically to the pets site appears to have the same 43-kDa molecular mass as the human counterpart. The pets site consists of a TG-rich element found between two Elf-1 binding sites on the HIV-2 enhancer. This arrangement of a TG-rich element immediately adjacent to ets-binding sites is also seen in the HTLV-1, MLV, and human T-cell receptor ␤-chain (TCR␤) enhancers (13,15,(25)(26)(27). We have previously shown that when the HIV-2 pets site is mutated, enhancer activation in response to T cell stimulation is greatly reduced (8). The HIV-2 pets site appears to be a final common element in signal transduction, as this site mediates activation whether the enhancer is stimulated by phorbol myristate acetate alone, phytohemagglutinin alone, phorbol myristate acetate plus phytohemagglutinin, soluble antibodies to the T cell receptor, immobilized antibodies to the T cell receptor, or by antigen (8,10,11). Not only is this site important for transcriptional activation of the HIV-2 enhancer, but mutation of the TG-rich sequence of the HTLV-1 enhancer also markedly inhibits inducible enhancer function (13). An intact TG-rich sequence found adjacent to the ets-binding site is again required for transactivation of both the MLV and TCR␤ enhancers (25). It has been shown that factors (termed core binding factors) binding to this TG-rich element cooperate in vivo to regulate transcription from the MLV and TCR␤ enhancers (25). Furthermore, point mutations at the site where core binding factors bind to the MLV enhancer cause a shift in disease specificity of the virus, resulting in erythroid leukemia rather that T cell leukemia (15). These observations suggest that the pets/ets motif, which has been conserved among widely divergent retroviruses, is likely to be broadly important. As the pets factor binds DNA constitutively (8), it may need to interact with other proteins or be post-translationally modified in order to regulate transcription. Elf-1, an ets family member which binds both immediately upstream and downstream of the pets site in the HIV-2 enhancer, is very similar to the Drosophila developmental factor E74 (7,8). Perhaps the interaction of 3 N. Clark and D. Markovitz, unpublished observations. FIG. 7. A, purified pets was precipitated with trichloroacetic acid, separated by SDS-PAGE, and stained with silver. Lanes 3-6 are from four different purification runs and represent the fractions after two rounds of DNA affinity chromatography. Lanes 1 and 2 are duplicates which represent the eluate from the Q-Sepharose column, before DNAaffinity chromatography. The apparent molecular mass of the dominant band seen here as 43 has been determined to be 43.3 kDa by mass spectrometry of 1 g of purified protein. B, the purified 43-kDa protein binds to the HIV-2 pets probe when eluted from a gel slice after SDS-PAGE separation (lane 1). Binding of this 43-kDa protein can be efficiently competed out with the addition of unlabeled pets oligonucleotide (lane 2) but is only weakly competed for with the unrelated HIV-2 B oligonucleotide (lane 3).
Elf-1 with the pets factor will prove necessary for regulating transcriptional activation of the HIV-2 enhancer. Indeed, many ets family members, including Elf-1, require cofactors in order to bind DNA or to activate transcription (25, 26, 28 -30).
While the HIV-2 enhancer arrangement is similar to the MLV enhancer with regards to ets sites found immediately adjacent to a TG-rich sequence, the factors previously described to bind to the MLV TG-rich sequence appear to be distinct from the HIV-2 pets factor, as they are much smaller than 43 kDa, ranging from 19 to 35 kDa (26,27). Experiments to test the requirement of the HIV-2 pets element for the 43-kDa factor in mediating enhancer activation, either alone or in combination with Elf-1 or other proteins, will be facilitated by the identification and/or cloning of this purified pets factor. The elucidation of the mechanism by which the pets factor mediates HIV-2 transcriptional activation will provide us with a better understanding of this important human pathogen and of signal transduction in T cells. While previous experience suggests that more than one cloned protein is likely to bind to a given enhancer site in vitro (8,9), the studies presented here show that the dominant protein recognizing the pets site in crude and highly purified nuclear extracts is 43 kDa in size. Therefore, this is likely to be the size of the biologically relevant pets factor.