DNA binding by Kaposi's sarcoma-associated herpesvirus lytic switch protein is necessary for transcriptional activation of two viral delayed early promoters

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV; also known as human herpesvirus-8) establishes latent and lytic infections in both lymphoid and endothelial cells and has been associated with diseases of both cell types. The KSHV open reading frame 50 (ORF50) protein is a transcriptional activator that plays a central role in the reactivation of lytic viral replication from latency. Here we identify and characterize a DNA binding site for the ORF50 protein that is shared by the promoters of two delayed early genes (ORF57 and K-bZIP). Transfer of this element to heterologous promoters confers on them high-level responsiveness to ORF50, indicating that the element is both necessary and sufficient for activation. The element consists of a conserved 12-bp palindromic sequence and less conserved sequences immediately 3' to it. Mutational analysis reveals that sequences within the palindrome are critical for binding and activation by ORF50, but the presence of a palindrome itself is not absolutely required. The 3' flanking sequences also play a critical role in DNA binding and transactivation. The strong concordance of DNA binding in vitro with transcriptional activation in vivo strongly implies that sequence-specific DNA binding is necessary for ORF50-mediated activation through this element. Expression of truncated versions of the ORF50 protein reveals that DNA binding is mediated by the amino-terminal 272 amino acids of the polypeptide.

FIG. 1

Deletion analysis of the ORF57 promoter reveals the minimal sequences necessary for transactivation by ORF50 (50RE). This schematic depicts the features of the ORF57 promoter lying between the transcriptional start site (genomic position 82003 [12]) and 602 bp upstream of it [here named -602(WT)]. Each deletion of the promoter is depicted by the grey bars and named Δ1 to Δ5K (described in Materials and Methods); the endpoint of each deletion is also described by its distance in base pairs from the start site, listed by a negative number next to the name of the deletion. Each was cotransfected into CV-1 cells with increasing amounts of pcDNA3-FLg50 or empty pcDNA3, using Superfect, as described in Materials and Methods, and fold transactivation was calculated. The maximal point of transactivation in each titration curve is listed (as well as the standard deviation [SD], in parentheses) for each deletion. The black bar denotes promoter positions -106 to -54, referred to as 50RE. The abbreviations above the wt promoter indicate consensus binding sites for cellular transcription factors as predicted by TRANSFAC (32). Abbreviations: SREBP-1, sterol regulatory element binding protein 1; CDP, CAAT displacement protein; SRY, sex-determining region Y protein; GATA, GATA family factors; OCT, octamer proteins; CREB, cyclic AMP-responsive binding element protein; AP-1, activating protein 1; SOX, SRY-like high-mobility-group box-containing protein family; AML-1a, acute myeloid leukemia 1a protein; TATA, TATA box; ISRE, interferon-stimulated response element.

FIG. 2

Enhancer-like function of 50RE and conservation with the K-bZIP promoter. (A) 50RE confers ORF50 responsiveness on a heterologous TATA box. 50RE was fused to the hsp70 TATA box (30) in forward and reverse orientations to generate the two reporter plasmids depicted. Each was cotransfected into CV-1 cells with various amounts of ORF50 expression vector using Fugene-6, and fold activation was calculated as for Fig. 1. Maximal activation of each reporter is given at the right, followed by standard deviations (S.D.) in parentheses. (B) The TATA-proximal promoters of the ORF57 and K-bZIP genes share three conserved sequences. The -106/-5 sequence of the ORF57 promoter is given at the top, the sequence of the TATA-proximal promoter of K-bZIP is shown at the bottom, and the consensus between them is in the center. The grey boxes denote the conserved functional blocks shared by the two promoters. The convergent arrows denote the 12-bp palindromic sequence shared by the promoters, and the open box denotes the extended, palindromic flanking sequences found only in p57. The numbers above and below the sequences refer to the genomic positions of the first nucleotide depicted in each promoter sequence. Kpn I refers to the KpnI restriction endonuclease site which demarcates the right-hand side of 50RE at nt -54 from the ORF57 start site.

FIG. 3

Time course of ORF50 protein expression in Sf9 cells infected with recombinant baculovirus. A recombinant baculovirus in which ORF50 is expressed under control of the polyhedrin promoter (Bac-50) was constructed and grown to high titer as described in Materials and Methods. Sf9 cells were incubated with medium alone (Mock) or infected with Bac-50 at a multiplicity of infection of 10; whole-cell extracts were prepared at the indicated hours postinfection and analyzed by immunoblotting with anti-ORF50 antibody. Extract from Sf9 cells infected with a recombinant baculovirus expressing an irrelevant, non-KSHV protein (“WT”; a gift from David Morgan) and mock extract were harvested at 48 hpi and similarly analyzed. ORF50 was also expressed by transcription and translation in vitro in RRL and detected together with the Sf9 proteins or with a shorter exposure to film (short).

FIG. 4

Both the palindrome and flanking sequences contribute to DNA binding and transactivation by ORF50. (A) Sequences of four 26-bp oligonucleotides used as probes in EMSA and cloned into reporter plasmids. The sequence of the top strand of 50RE is shown at top, with the grey box depicting the 12-bp palindrome and the open box depicting the extended, flanking palindromic sequences. Each sequence below that represents the top strand of each of four oligonucleotides named 5A to 5D, which were annealed to complementary oligonucleotides as described in Materials and Methods. The resultant products were radiolabeled for use in EMSA or cloned as dimers upstream of the hsp70 TATA box in plasmid hsp-luc for use in transactivation assays (generating plasmids p57-5Ahsp-luc, p57-5Bhsp-luc, p57-5Chsp-luc, and p57-5Dhsp-luc; see Materials and Methods). The palindromic and flanking sequences which are included in each oligonucleotide are indicated by the boxes. (B) Four DNA-protein complexes are formed by incubation of Bac-50-infected Sf9 cell extracts with the EMSA probes. ORF50 was partially purified from Bac-50-infected Sf9 cells as described in Materials and Methods, and 0, 2.5, 5.0, and 10.0 μg, of extract (containing ca. 0, 50, 100, and 200 ng of His₆-tagged ORF50 [see Materials and Methods]) were analyzed for interaction with probes 5A to 5D by EMSA. Each set of four binding reactions is indicated above the autoradiogram according to the EMSA probe used, and resulting complexes are indicated by the numbers to the right of the autoradiogram (see text). (C) Transactivation of the 5A to 5D sequences by ORF50. Each reporter plasmid (A) was cotransfected into CV-1 cells with increasing amounts of pcDNA3-FLg50 or empty expression vector, and fold activation was assayed as for Fig. 2A. The entire titration curve is depicted for each reporter. (D) ORF50 is a component of complexes 1 and 1*. Partially purified extract from Bac-50-infected Sf9 cells (250 μg) was depleted with Ni-NTA-agarose or mock depleted with glutathione-Sepharose (glutathione-seph.) as described in Materials and Methods. The Ni-NTA-depleted extract was further depleted by a second incubation with a fresh aliquot of beads; 5 μg of untreated extract or an equivalent volume of each depleted extract was then subjected to EMSA with probe 5D as for panel B (top). The migration of complexes 1 to 3 is indicated to the right. Alternatively, 1/10 volume of each extact was analyzed by immunoblotting for the presence of ORF50 (bottom). (E) Competition EMSAs demonstrate a hierarchy of binding preferences for ORF50 binding to the 5A to 5D sequences. Partially purified extract from Bac-50-infected Sf9 cells (6 μg) was incubated with a 10-, 20-, 40-, or 80-fold molar excess of each nonradiolabelled EMSA probe before incubation with a 1-fold molar equivalent of radiolabeled EMSA probe 5D. These preparations were subject to EMSA for panel B, and each set of four binding reactions is indicated above the autoradiogram. Included for comparison are an EMSA of a binding reaction lacking protein (No protein) and a reaction containing labeled 5D probe and ORF50 without specific competitor (No competitor); the migration of complexes 1 to 3 is indicated to the right.

FIG. 5

ORF50 binding to sequences within the palindrome is necessary for transcriptional activation. (A) Sequences of four 26-bp oligonucleotides used as probes in EMSA and cloned into reporter plasmids. The sequence of the top strand of p57-5Dwt is shown at the top, with the grey box depicting the 12-bp palindrome and the open box depicting the extended, flanking palindromic sequences. Each sequence below that represents the top strand of each of four mutant oligonucleotides named 5Dm1 to 5Dm3, in which the 6-bp linker substitution represented by the hatched box was substituted for the wt sequence indicated in each probe. These were annealed to a complementary oligonucleotide and then radiolabeled for use in EMSAs; 5Dwt, 5Dm2, and 5Dm3 were also cloned to produce the reporter vectors p57-5Dwthsp-luc, p57-5Dm2hsp-luc, and p57-5Dm3hsp-luc (see Materials and Methods). (B) The right half of the palindrome and flanking sequences are required for binding of ORF50 to p57-5D. Aliquots of 0, 2.5, 5.0, and 10.0 μg of partially purified extract (containing ca. 0, 50, 100, and 200 ng of His₆-tagged ORF50 [see Materials and Methods]) from Bac-50-infected Sf9 cells were analyzed for interaction with probes 5Dwt and 5Dm1 to 5Dm3 by EMSA as for Fig. 4B. Each set of four binding reactions is indicated above the autoradiogram according to the EMSA probe used. (C) Competition EMSAs confirm the sequence requirements for ORF50 binding to p57-5D. Six micrograms of partially purified extract from Bac-50-infected Sf9 cells was incubated with a 50-, 100-, or 200-fold molar excess of each nonradiolabeled EMSA probe before incubation with a 1-fold molar equivalent of radiolabeled EMSA probe 5Dwt. These preparations were subject to EMSA as for Fig. 4E, and each set of four binding reactions is indicated above the autoradiogram. Included for comparison are an EMSA of a binding reaction lacking protein (No protein) and a reaction containing labeled 5Dwt probe and ORF50 without specific competitor (No competitor). (D) DNA binding of the 5D sequence by ORF50 is required for transcriptional activation. Each reporter plasmid (A) was cotransfected into CV-1 cells with increasing amounts of pcDNA3-FLg50 or empty expression vector, and fold activation was assayed and depicted as in Fig. 4C.

FIG. 6

Fine mutational analysis of 50RE. A second series of nine linker substitution mutants across the 5D element was generated as EMSA probes and cloned into hsp-luc using oligonucleotides (as described in Materials and Methods). All of the mutants are named at the left; the new series of mutants, containing 1- to 4-bp changes of the wt sequence, are listed in lines 5 to 13. The specific base pairs changed in each clone of both series are listed in the center column underneath the sequence of 5Dwt. Each was tested both for DNA binding by the Bac-50-generated ORF50 and for transcriptional activation as in Fig. 4 and 5. DNA binding data were semiquantitated for each mutant by comparison to the binding of ORF50 to the 5Dwt probe, which was assigned the value of +++; similarly, activation was quantitated for each mutant by comparison to ORF50's activation of the 5Dwt reporter construct, which was assigned the value of 100% (the first four lines summarize the data from Fig. 5).

FIG. 7

The palindromic sequences within the K-bZIP promoter are required for full activation by ORF50. A 12-bp linker substitution mutation was introduced to replace the 12-bp palindromic sequence of the wt, full-length ORF57 (A) or K-bZIP (B) promoter (see text). Each mutant promoter was compared to its corresponding wt promoter in cotransfections of CV-1 cells as described for Fig. 5D.

FIG. 8

ORF50 binds the TATA-proximal K-bZIP promoter with specificity that corresponds with ORF50's ability to activate transcription. (A) Sequence of the top strand of two 28-bp oligonucleotides used as probes in EMSA. Both were annealed to their complementary oligonucleotides as described in Materials and Methods. The box depicts the palindrome shared with the ORF57 promoter, and the grey nucleotides correspond to positions which are homologous to the ORF57 promoter (as in Fig. 2B). (B) Partially purified extract from Bac-50-infected Sf9 cells (6 μg) was incubated with a 50- or 200-fold molar excess of each nonradiolabeled 5Dwt or 5Dm2 EMSA probe (Fig. 4A) or 5- or 10-fold molar excess of each nonradiolabeled ZIPwt or ZIPm1+2 EMSA probe (A) and then incubated with a 1-fold molar equivalent of radiolabeled EMSA probe ZIPwt. These preparations were subject to EMSA as for Fig. 4B, and each set of four binding reactions is indicated above the autoradiogram. (C) Partially purified ORF50 protein was incubated with a 50- or 200-fold molar excess of nonradiolabeled ZIPwt or ZIPm1+2 EMSA probe and then incubated with 1-fold molar equivalent of radiolabeled 5Dwt probe. Included for comparison are an EMSA of a binding reaction lacking protein (No protein) and a reaction containing labeled 5Dwt probe and ORF50 without specific competitor (No competitor).

FIG. 9

The amino terminus of ORF50 generated in E. coli binds to the response element. (A) Expression of N- and C-terminal truncations of ORF50 in E. coli. The top displays a schematic of full-length ORF50 (FL50). Below are schematics of the truncations of ORF50 expressed as fusions to the His₆ epitope tag: aa 1 to 272 (N50) and aa 525 to 691 (C50 [16]). The bottom shows Western blot analysis of the proteins expressed and purified from E. coli transfected with pET28b-N50 (N50) and pRSET 0.8 (C50 [16]), detected using anti-His₆ (Clontech) and anti-ORF50 (16) antibodies, respectively. Positions of migration of protein molecular weight standards (Life Technologies) are shown to the left and right. Abbreviations: NLS, putative nuclear localization signal; Basic, highly basic domain (described in the text); LZ, leucine repeats; ST, serine/threonine-rich region; AD, activation domain (15). (B) N50 but not C50 binds stably to the 5D sequence. N50 and C50 were expressed and purified from E. coli as described in Materials and Methods and then diluted to equal final concentrations; 0, 0.5, 1.0, and 2.0 μg of each protein (containing ca. 0, 0.1, 0.2, and 0.4 μg of truncated ORF50 polypeptide; see Materials and Methods) was incubated with radiolabeled 5Dwt probe and analyzed by EMSA as described in Materials and Methods. Each set of four binding reactions is indicated above the autoradiogram according to the polypeptide used. (C) N50 binds to the ORF57 5D but not 5B sequence. EMSA was performed as for panel B using the indicated radiolabeled probes and 0, 0.5, 1.0, and 2.0 μg of N50 polypeptide. (D) Unlabeled competitors confirm N50's binding specificity. One microgram of N50 polypeptide was incubated with a 50-, 100- or 200-fold molar excess of nonradiolabeled 5B or 5Dwt EMSA probe and then incubated with a 1-fold molar equivalent of radiolabeled 5Dwt probe. Included for comparison are an EMSA of a binding reaction lacking protein (B, No Protein) and a reaction containing labeled 5Dwt probe and ORF50 without specific competitor (D, No Protein).