Page 2
Page 3
Page 4
Page 5
Page 6
Page 7
Page 8
Page 9
Page 10
Page 11
Page 12
Page 13
Page 14
Page 15
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBPIO0 binds remarkably de. enerate octamer moUfs through specmc interactions with flanking sequences Thomas Baumruker, Richard Sturm, and Winship Herr Cold Spring Harbor Laboratory, Cold Spring Harbor, New York 11724 USA We have used the 100-kD HeLa cell octamer-binding protein OBP100 as a model to study flexible DNA sequence recognition by promoter-binding proteins. OBPI00 binds to the conserved octamer motif ATGCAAAT found in numerous promoters and additionally to two degenerate octamer motifs (sites I and II) within the SV40 enhancer region. We show here that OBP100 binds the herpes simplex virus immediate early promoter TAATGARAT (R = purine) motif itself, extending the flexibility of OBP100 sequence recognition to sequences that bear very little resemblance (four matches over a 14-bp region). Nevertheless, a progression of OBP100- binding sites can be established that links the sequences of these two apparently unrelated binding sites by incremental steps. Mutational and chemical modification interference analyses of a degenerate octamer binding site (SV40 site II) show that specific sequences, which are not normally conserved but flank the degenerate octamer motif, can compensate for the degeneracy in the octamer core sequence. Thus, different regions of the binding site sequence (core or flanking) can diverge separately but not independently of one another. These results suggest that flexible DNA sequence recognition arises because there are few obligatory contact sites for OBP100 binding, but, rather, specific binding reflects the sum of many independent interactions. [Key Words: SV40 enhancer; herpes simplex virus; TAATGARAT motif; transcription; replication] Received August 3, 1988; revised version accepted September 22, 1988. Transcriptional regulation is largely mediated by the in- camer motif ATGCAAATNA (Falkner and Zachau teraction of DNA-binding proteins with specific pro- 1984; Falkner et al. 1986)] was characterized originally moter elements. In eukaryotes, this process is not only as part of a well-conserved 13-bp sequence of histone complicated by the relatively large number of cis-acting H2B promoters (Harvey et al. 1982) but has been found elements and protein factors but also by their frequent since in a large number of different promoters. The oc- overlapping organization and combinatorial usage (see tamer motif has been shown to interact with both ubiq- Ondek et al. 1988). Important cis-acting elements com- uitous and lymphoid-specific cellular proteins (Landolfi monly have been identified as either conserved sequence et al. 1986; Singh et al. 1986; Staudt et al. 1986; Rosales motifs (e.g., CCAAT box; Benoist et al. 1980; Efstra- et al. 1987; Scheidereit et al. 1987) and is implicated in tiadis et al. 1980) or as conserved transcription-factor- S-phase regulation of H2B genes (LaBella et al. 1988), binding sites (e.g., Spl; Dynan and Tjian 1983). The ex- adenovirus DNA replication (Pruijn et al. 1986; Rosen- istence of conserved sequence motifs and transcription- feld et al. 1987), regulation of small nuclear RNA tran- factor-binding sites has favored a model in which gene scription (Ares et al. 1985; Mattaj et al. 1985), and cell- expression is regulated by relatively strict sequence rec- type-specific regulation of both immunoglobulin genes ognition by DNA-binding proteins. Recent studies, how- (Gerster et al. 1987; Wirth et al. 1987) and the SV40 en- ever, have complicated the picture because some pro- hancer (Davidson et al. 1986). Unlike the appearance of moter-binding proteins can recognize very different se- octamer elements in many different regulatory regions, quences (Johnson et al. 1987; Pfeifer et al. 1987; Sturm et the TAATGARAT sequence has only been implicated in al. 1987; Costa et al. 1988; Michael et al. 1988). transcriptional activation of HSV early genes by the viral Two well-characterized but apparently unrelated con- gene product VP16 (or Vmw65) (Campbell et al. 1984; sensus sequences are the octamer motif ATGCAAAT, Kristie and Roizman 1984; Preston et al. 1984). The initially described in immunoglobulin gene upstream VP16 gene product does not bind to DNA directly promoter and enhancer regions (Falkner and Zachau (Marsden et al. 1987), but one or more cellular factors 1984; Parslow et al. 1984), and the TAATGARAT present in uninfected cells have been shown to interact (R = purine) motif in herpes simplex virus (HSV) imme- specifically with the TAATGARAT element, suggesting diate early gene promoters (Mackem and Roizman that these may be cellular vehicles for VP 16 transactiva- 1982b). The octamer motif [also referred to as the de- tion through the TAATGARAT element (Kristie and 1400 GENES & DEVELOPMENT 2:1400-1413 © 1988 by Cold Spring Harbor Laboratory ISSN 0890-9369/88 $1.00 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBPIO0 sequence recognition Roizman 1987; McKnight et al. 1987; O'Hare and with guanosine residues. We therefore extended the Goding 1988; Preston et al. 1988; Triezenberg et al. DEPC interference analysis of OBP100 to other well- 1988b). characterized octamer sites. Recently, we have reported the purification from HeLa Figure 1 characterizes the interaction of OBP100 with cells of a 100-kD octamer-binding protein (OBP), called four different octamer motifs: three perfect octamer OBP100, and described its interaction with the SV40 and motifs from the murine IgH enhancer, the human U2 immunoglobulin heavy-chain (IgH) gene enhancer re- small nuclear RNA (snRNA) promoter, and the chicken histone H2B promoter and one imperfect octamer (6-bp gions (Sturm et al. 1987). This protein probably repre- match; ATGATAAT)within the adenovirus 2 (Ad2) in- sents the ubiquitous octamer-binding factor, first called NF-A1 (Staudt et al. 1986) and NF III (Pruijn et al. 1986). verted terminal repeat (ITR). Figure 1A shows that The IgH enhancer contains a single octamer motif, and OBP100, which has been purified extensively by DNA this motif is the only IgH enhancer-binding site for affinity and SDS-polyacrylamide gel electrophoresis OBP100. In contrast, the SV40 enhancer region lacks a from a heparin-agarose fraction (Sturm et al. 1987), perfect octamer sequence but, instead, contains the binds to each of the four octamer sequences (lanes 1-4) closely related sequence ATGCAAAG (Falkner and Za- in a gel retardation assay. This result is consistent with chau 1984). OBP100 binds to this octamer-related motif, OBP100 representing the ubiquitous octamer-binding referred to here as site I, and also binds a flanking site activities that have been shown to bind to these octamer lsite II) that was not noted by sequence comparison. Site elements: NF-A1 (Singh et al. 1986; Staudt et al. 1986) II was shown by chemical modification interference to and NF III (Pruijn et al. 1986; Rosenfeld et al. 1987). extend over a 13-bp sequence that contains two overlap- Consistent with the ubiquitous nature of OBP100, nu- ping octamer-related sequences with six (ATGCATCT) clear extracts from human 293 cells, monkey kidney or five (ATCTCAAT) matches (the matches are under- CV-1 cells, and mouse NS-1 and 70Z/3 lymphocytes all lined) to the octamer consensus sequence. These results form the same sized gel retardation complex as the suggested that OBP100 can interact with degenerate oc- OBP100-specific complex (data not shown). tamer motifs, but that flanking sequences are important For the OBP100 DNA-binding studies described for these interactions. However, the significance of these below, the source of OBP100 was the heparin-agarose flanking sequences was not known. fraction. This source of OBP100 has the advantage that To understand the mechanism of degenerate sequence OBP 100 has not been denatured by SDS-polyacrylamide interaction, we have used mutagenesis and chemical gel electrophoresis. Even though this fraction is not modification interference in combination with gel retar- pure, we believe that OBP100 is the protein responsible dation to examine the sequence requirements for for the binding activity being analyzed because (1) OBP100 binding to the two SV40 octamer-related sites. OBP100 is the only OBP we can detect in this partially Also, we have extended the repertoire of OBP100- purified extract (Sturm et al. 1987, and see below); (2) all binding sites. Although the OBP100-binding site II is a of the binding activities can be competed by a single se- degenerate octamer motif, it shares considerable se- quence (SV40 site I) that we used to purify OBP100; (3) quence similarity to the TAATGARAT motif (seven/ expression of a human cDNA clone of OBP100 in Esche- nine match). We show that OBP100 can bind to a richia coli produces an OBP with a DNA-binding speci- TAATGARAT motif that matches the SV40 site I at ficity that parallels the activity in the heparin-agarose- only 4 out of 14 positions. Together, the results reported purified fraction (R. Sturm and W. Herr, unpubl.); and (4) here indicate that normally nonconserved sequences antisera raised against the E. coli-expressed OBP100 flanking an octamer motif can stabilize the interaction completely inhibits octamer-specific complex formation of a degenerate octamer core to OBP100, thereby greatly by the heparin-agarose fraction (R. Sturm and W. Herr, extending the flexibility of sequence recognition by unpubl.). OBP100. The results of the DEPC interference assays on both strands of the four octamer sequences are shown in Figure 1B and are summarized in Figure 1C. At each site, Results interfering modifications map over the 8-bp octamer se- quence as well as flanking residues. Nearly all of the oc- Base-specific interaction of OBP1 O0 with different tamer motif modifications affect complex formation, octamer elements with one notable exception: Modification of the central In our previous studies, the base-specific interaction of adenine on the top strand (ATGCAAAT) does not inter- OBP100 with the two SV40 enhancer octamer-related fere with complex formation [see IgH, U2 snRNA, and sequences was deduced by chemical modification inter- H2B in Fig. 1, SV40 site I in Sturm et al. (1987), and Fig. 4 ference, in combination with a gel retardation assay below]. The interference by modification of flanking res- (Sturm et al. 1987). In addition to dimethylsulfate idues has been noted also by DMS modification of (DMS), we used diethylpyrocarbonate (DEPC) to modify flanking guanines (Ares et al. 1987; Pruijn et al. 1987; the DNA. DEPC modification proved superior to DMS Rosales et al. 1987). The detrimental effect of flanking for characterization of the two SV40-binding sites be- modifications within nonconserved residues suggests a cause these sites are AT rich and DEPC reacts three- to role for these sequences in OBP100 binding but without fivefold more strongly with adenosine residues than strict sequence requirement. GENES & DEVELOPMENT 1401 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBPIO0 Interactions U2 H2B IgH Ad [ I r i I OBPIO0 Octamer 0 + O- 0 + O- 0 + O- 0 + O- I I • o[o • • • 1 £D £D CD (.9 £D (_9 T ~o 4- ~ 4- H GAAATGCAAATT AC 4- ~-4 4. ,--4 ~I,C_) ~iiO t,-i IgH CTTTACGTTTAATG • • • • ii-" AITIT;ITI~TGAG Ad ITR TTATACTATTACTC ii!i: t[ It, G CAT CAAATTCG U2 snRNA CCGTACGTTTAAGC • O • • CTTATGCAAATGAG H2B GAATACGTTTACTC • • • • ,[ I 2 3 45 789 I0 II 12 131415161718 192021222:324 123 456 Figure 1. OBP100 interaction with four different octamer motifs. (A) Gel retardation assay using highly purified OBP100 shows specific interaction with each of the four sepa- rate octamer motifs. Five different DNA restriction fragments were prepared and radiolabeled as described in Methods and are as follows: (IgH) Murine IgH enhancer~ (Ad) adeno- virus 2 ITR; (U2) human U2 small nuclear RNA distal element~ (H2B) chic'ken histone H2B promoter; (pUC) negative control fragment from the pUC119 polylinker containing an SV40 coreA/coreA motif. Labeled fragments were combined with the extensively purified OBP100 described previously (Sturm et al. 1987) and analyzed by electrophoresis through a 6% polyacrylamide gel in Tris-glycine buffer (see Methods). (B) DEPC modification interference analyses. Both strands of the four-octamer containing fragments de- scribed in A were individually end labeled, modified with DEPC, and used in a preparative gel retardation assay with HeLa cell OBP100 partially purified by heparin-agarose column chromatography. Free probes (F) and probes comigrating with the OBP100 complex (CII were recovered and cleaved by treatment with piperidine. The cleavage products were displayed by electrophoresis through 16% denaturing polyacrylamide gels, along with a purine-specific ladder (A + G)generated by acid treatment of unmodified probe DNA. The strands shown represent either the top strand (O+i ATGCAAAT) or the bottom strand (O-i ATTTGCAT), as shown in C. The brackets alongside the interference pattems identify the position of the octamer element, and the arrows identify the 5' ---* 3'direction. (C) Summary of the DEPC modification interference patterns. Modifications that interfere with complex formation are identified by dots. Large dots indicate a reduction of threefold or greater in band intensity in the complexed sample, and small dots indicate a reduction less than threefold. Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBP100 sequence recognition Boundaries of the SV40 OBPlOO-binding sites I and H lated sequences, Octal, Octa2, and Octa3, which lie within the OBP100-binding site region, are identified. Figure 2A shows a diagram of the SV40 early promoter These octamer-related sequences contain seven (Octal), with the sequence of the OBP100-binding site region six (Octa2), and five (Octa3) matches to the 8-bp octamer displayed below. Over the sequence, three octamer-re- consensus sequence. Below the sequence, the dashed 72bp 21bpl 21bp - i ori 0/524:5 A C J B Octa 2 6/8 oct6rner x x~ Octo I Octa :3 7/8 octamer 5/8 octamer Xq r X X X i GCAGAAGTATGCAAAGCATGCATCTCAATTAGTCAGCA CGTCTTCATACGTTTCGTACGTAGAGTTAATCAGTCGT F ......... qF .......... i I II i SV40/OBPIO0 SV40/OBPIO0 siteT siteTr siteI site]-[ I ] I B C 125 125 I 2:5456 C I ..... CI- :" [ 2 : F-- F- F- GCA] ~ GCAI FCTA GCA] i--x x x I ETA J TATGCAAAGC - I ATGCATCTCAATT - I ATGCATCTCAATTAG + 2 ATGCAAAGCA - 2 GCATGCATCTCAATT - 2 TA (-) :3 TATGCAAAGCA + :3 ATGCATCTCAATTAG + 5 T + t___×J t--.x ×J 4 6C +/- 7/8 Octal 6/80cta 2 5 C + L---xx x I 6 C +/- 5/80cta 3 Figure 2. Definition of the two SV40 enhancer OBP100-binding sites I and II. (A) Diagram of the SV40 early promoter with one copy of the 72-bp repeat. Represented from right to left are the early transcription start sites {wavy arrows), the origin (ori) of replication, the TATA-like sequences (A/T), and the 21-bp repeats. Below the 72-bp repeat region is shown the position of the A, B, and C enhancer elements, as defined by Herr and Clarke (1986). Below the diagram is shown a part of the 72-bp sequence spanning the OBP100- binding sites. Above the sequence, three octamer-related sequences (Octal-Octa3) are bracketed with mismatches to the octamer consensus shown by Xs. Below the sequence are shown the two SV40 OBP100-binding sites as originally defined by chemical interfer- ence analysis (Sturm et al. 1987) (dashed lines) and as defined in this paper by oligonucleotides and mutagenesis (solid lines). (B-D) Gel retardation assays of wild-type and mutant SV40 OBP100-binding sites I and II. Gel retardation assays were performed with HeLa cell nuclear extract partially purified for OBP100 by heparin-agarose column chromatography. The probes were EcoRI-HindIII-di- gested fragments continuously labeled on one strand (see Methods). The free (F) and complexed (CI) probes are indicated. The lane numbers in each panel correspond to the numbered sequences shown below that were used as probe, along with the flanking poly- linker nucleotides shown boxed. As diagramed, the pUC119 HindIII site lies to the left. The octamer-related sequences Octal-Octa3 are bracketed, indicating the positions that differ from the octamer consensus sequence ATGCAAAT. A qualitative measure of the relative levels of complex formation with each probe is indicated by the + and - designations to the right of each sequence. (B) Site I definition; IC} Site II definition; (D) Effect of mutations at the boundaries of site II. Mutations are shown in boldface type, and the straight line indicates identity with the wild-type site II sequence shown above. GENES & DEVELOPMENT 1403 liT!~i:.!~i::ii:~i~{i'i{~:i:::i~il f.i:~!/.:~!i:::{::.i::4:i,::,i:!l Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press Baumruker et al. line shows the limits of the two SV40 OBP100-binding To test this hypothesis we created three point mutations sites, as defined previously by DEPC and DMS modifica- within the degenerate Octa3 sequence, creating a perfect tion interference (Sturm et al. 1987). From the chemical octamer motif (ATGCATGCAAATTAG) and then com- interference studies, OBPl00-binding site II spanned pared (1) the effect of these mutations on OBP100 af- finity for site II and (2) the effect of mutations and chem- both octamer-related sequences Octa2 and Octa3, of which only the former was noted in our previous study ical modifications in sequences flanking the perfect and (Sturm et al. 1987). To study the SV40 sites I and II sepa- degenerate site-II octamer sequences. Figure 3 shows a competition experiment in which rately, we used synthetic oligonucleotides to create indi- the binding affinity of OBP100 to the wild-type SV40 vidual binding sites within the pUC119 polylinker, and site II and perfect octamer mutant has been compared to these were used in gel retardation assays (see Methods). the affinity for the perfect H2B octamer and SV40 site I. Unexpectedly, neither of the two sites defined by chem- In each case, the same restriction fragment, containing ical interference bound to OBP100 in a gel retardation 14 copies of the SV40 site I (14xB17; Ondek et al. assay (Fig. 2B, C, lanes 1). To elucidate the boundaries of binding sites I and II, 1987), was used as competitor in 3.3-fold serial dilu- the synthetic oligonucleotide sequences were extended tions. The relative rate at which the OBP100 complex disappears is an indication of relative affinity. The re- in either direction. Figure 2B shows that shifting the en- tire binding site I by 1 bp, to contain the decamer-related sequence ATGCAAAGCA, does not recreate the COMPETITOR OBP100-binding site (lane 2), but addition of the 5'- 14xBI7 flanking T residue to the decamer motif restores t- o') OBP100 binding (lane 3). Thus, these studies define site I as an 11-bp binding site. Site II was defined similarly, as ~6 shown in Figure 2C; extension of the sequences 5' of the o o o -- r~ r~ rr Octa2 sequence did not restore octamer-binding activity (lane 2), but two additional residues 3' of the Octa3 motif did restore activity (lane 3). The exact boundaries 1.00 H2B of site II were defined more precisely by mutagenesis of the external residues of the active binding site II (Fig. 2D). These studies show that the 5' A residue of the Perf. Octa Octa2 sequence is not essential for OBP100 binding to 0.94 Site 13" site II (Fig. 2D, lanes 3 and 5), but when the two 5' A and T residues are mutated by transitions (lane 4) or by transversions (lane 2), the site II binding activity is im- paired. These results, in combination with the effect of SV40 ~ 0.75 Site I mutating the 3'-terminal G residue (lane 6), indicate that site II extends over the 14-bp sequence TGCATCT- CAATTAG (octamer-related residues are identified by underlining), in which the degenerate Octa3 motif is SV40 Siterf ...... 0.24 centrally located. These mutagenesis studies indicate that the more closely octamer-related Octa2 sequence does not bind to OBP100 in these assays (see Fig. 2C, Figure 3. Gel retardation competition experiments to assay lane 2); this result emphasizes the difficulty in identi- the relative affinity of two perfect (H2B and the perfect site II octamer) and two imperfect (wild-type SV40 sites I and II) oc- fying protein-binding sites by sequence similarity. Fur- tamer motifs for complex formation. The gel retardation assays thermore, the differences between the mutagenesis re- were performed with continuously labeled probes and the hep- sults described here and our prior results with chemical arin-agarose-purified HeLa cell nuclear extract containing modification (cf. the dashed line and solid line in Fig. OBP 100. Only the retarded complexes for each series of compe- 2A) show that chemical interference patterns are insuffi- tition are shown. The specific competitor, 14xB17 restriction cient to define an entire binding site. fragment, or nonspecific competitor, HaeIII-digested k phage DNA, were mixed together before addition of the protein ex- tract. The specific competitor was added in 3.3-fold serial dilu- Flanking residues stabilize OBPIO0 binding to a tions starting with a 100-fold greater number of competitor degenerate octamer binding sites to probe binding sites (3.3 x ). The nonspecific competitor was at the same concentration of competitor DNA The interaction with SV40 site II shows that OBP100 as in the 3.3 x lane with the 14xB17 fragment. The exposures can bind to very degenerate octamer motifs, but because differ so that the bands representing the complexes are of more of the large size (14 bp) of this binding site, appropriate equal intensity. To quantitate the disappearance of the com- flanking residues appear to be necessary for OBP100 to plexes, the films were scanned with a densitometer, and the bind such a degenerate octamer motif. It is unlikely, midpoint of disappearance of the complex identified. The rela- however, that specific flanking interactions are essential tive concentration of competitor necessary to reach this mid- for binding to a perfect octamer motif because these res- point is shown as the relative affinity with respect to the H2B idues are not conserved between different octamer sites. probe. 1404 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBP100 sequence recognition sults show that the 3-bp change that creates a perfect tion destroys one of the five matches between Octa3 and the octamer consensus, it becomes more similar to the site II octamer sequence creates a considerably better TAATGARAT motif (see below.)] Thus, changes within OBP100-binding site, having nearly as high an affinity as a binding site for OBP100 can affect the requirement for the perfect H2B octamer site and better than site I. The both 5'- and 3'-flanking residues. Nevertheless, the high high affinity of the perfect octamer in site II is partially a affinity of OBP100 for the perfect octamer sequence result of its context because not all perfect octamer (or within site II (Fig. 3) indicates that residues flanking decamer) motifs bind OBP effectively. For example, in a even a perfect octamer motif still contribute to the different context, mutating the Octa2 sequence to a per- binding of OBP100 to DNA. In the experiments de- fect octamer (decamer) sequence created an OBP100- scribed in Figure 4, the effect of the residues flanking the binding site with only the same affinity as the wild-type perfect site II octamer probably has gone undetected be- degenerate site II sequence (data not shown). cause the mutations and modifications that have a large Figure 4 shows that interactions with flanking res- idues that are essential for effective binding of OBP100 effect on wild-type site II binding are too subtle to have a to the site II degenerate octamer are not critical for large effect on binding to the perfect site II octamer. binding when the site II octamer is changed into a per- OBP binds to the HSV TAATGARAT motif fect octamer. First, when the double AT ~ TA point Kristie and Roizman (1987) noted a sequence similarity mutation that affects site II interactions (Fig. 4A, lanes 1 between the SV40 OBP100 site II sequence and, in in- and 2) is incorporated into the perfect octamer site II se- verse orientation, the TAATGARAT motif of HSV im- quence, there is little effect on OBP100 binding (cf. lanes mediate early promoters. The degenerate Octa3 se- 3 and 4 in the shorter exposure). Second, Figure 4, B and quence is very similar to the inverse orientation of the C, shows that DMS and DEPC modifications of residues TAATGARAT motif [ATCTCAAT (Octa3) compared to flanking the wild-type site II Octa3 sequence that inter- ATYTCATTA (TAATGARAT);Y = pyrimidine], but fere with OBP100 binding have little effect on binding to the TAATGARAT motif itself is only a 4-bp match with the perfect octamer site II. Thus, whereas modification the octamer motif (ATYTCATTI. A host cell factor(s), of either strand of the 5'-flanking TGC sequence inter- feres with binding to the wild-type degenerate site II, variously called H1 (Kristie and Roizman 1987), HC3 these modifications interfere little with binding to the (Preston et al. 1988), and TRF (O'Hare and Goding 1988), perfect octamer sequence (see arrowheads in Fig. 4B). has been identified in uninfected HeLa cell extracts; it This difference does not appear to result from a shift in binds specifically to the TAATGARAT motif. Given the the OBP100-binding site in the perfect site II octamer high degree of sequence similarity between the SV40 sequence because the interference pattern remains the OBP100 site II sequence and the TAATGARAT motif, same on the 3' side of the octamer sequence and does we tested the ability of OBP100 to bind to the TAAT- not include the flanking polylinker sequences. Although GARAT motif, thereby possibly expanding the reper- flanking modifications do not prevent binding to the toire of highly degenerate OBP 100 octamer-binding sites perfect site II octamer, modifications within the perfect to a new class of regulatory elements. octamer sequence have a large effect on binding. Indeed, To accomplish this goal, however, it was necessary to the very strong DEPC modification interference pattern choose carefully the TAATGARAT motif for analysis on the top strand of the perfect octamer (lanes 11 and because many TAATGARAT motifs contain overlap- 12), compared to the weaker interference pattern over ping sequences with a much higher degree of sequence similarity to the octamer motif. For example, Pruijn et the imperfect octamer (lanes 8 and 9), suggests that con- al. (1986) and O'Hare and Goding (1988) have noted tacts within the octamer sequence become more critical overlapping octamer motifs in several TAATGARAT when it is a perfect match to the consensus. motifs but on the same strand [e.g., the seven out of Together, these results indicate that the interaction of eight octamer sequence ATGCTAATGATAT in a OBP100 with the degenerate site II octamer sequence is nearly perfect TAATGARAT motif within the ICP0 the result of interactions with flanking residues that are (IE110K) promoter]. Therefore, to show that OBP100 can not essential for binding to a perfect octamer. The com- bind directly to the conserved TAATGARAT motif and plementary view is that mutations within a binding site not to overlapping octamer motifs, we chose to study can relieve the requirement for specific flanking res- the TAATGARAT motif located 260 bp upstream of the idues. We have observed this effect with other binding transcriptional initiation site for the ICP4 (IE175K)gene. sites as well. For example, if the incomplete SV40 site I The sequence of this site is shown in Figure 5C, and sequence TATGCAAAGC, which is missing the 3'- flanking A residue required for OBP100 to bind (cf. Fig. whereas it is a perfect match to the TAATGARAT motif 2B, lanes 1 and 3), is mutated to the perfect octamer se- (identified by the arrow on the lower strand in Fig. 5C) quence TATGCAAATC, this flanking A residue is not and is functional for VP16 transactivation (Triezenberg required (data not shown). We also found that a muta- et al. 1988b), it only contains at best 4-bp matches to the tion that changed the nonbinding site II sequence ATG- octamer motif. The match that corresponds to the oc- CATCTCAATT (Fig. 2C, lane 1) to ATGCATCT- tamer motif similarity in the SV40 Octa3 sequence is CATTT (called clone 12) created a binding site in the shown by the bracket above the TAATGARAT sequence absence of the 3'-flanking AG sequence required for in Figure 5C. binding to site II (Fig. 2C, lane 3). [Although this muta- When an 87-bp SphI/NcoI restriction fragment from GENES & DEVELOPMENT 1405 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press Baumruker et al. the ICP4 promoter region containing the -260 TAAT- II, indicating that the same factor(s) is responsible for GARAT motif was used in a gel retardation assay with both complexes and has a similar affinity for both sites. the heparin-agarose-purified HeLa cell fraction, a single A DEPC modification interference assay, shown in retarded complex comigrating with the OBP100 com- Figure 5B and summarized in Figure 5C, shows that the plex formed (data not shown). Figure 5A shows that for- interaction responsible for complex formation is indeed mation of this complex can be inhibited by competition with the TAATGARAT motif. The interfering modifica- with unlabeled SV40 site I DNA as readily as SV40 site tions include all but two of the residues in the TAAT- 8h exposure 2h exposure l I [ I I 2 3 4 5 3 4 5 GCAI ICTA CI-- i X X X I ATGCATCTCAATTAG + 2 TAGCATCTCAATTAG (-) [ TIT 3 ATGCATGCAAA AG 4-4- 4 TAGCATGCAAATTAG +4- i 5 non specific control - F -- m i Top Strand Bottom Strand I I I DMS DEPC DMS DEPC I I 1 [ 1 I" co co co Co CO CO CO CO 4. F.4 4. H + ~_~ ,~u_o <[ LL C) <[ LL C) <~ LL (..) iwl F !iiii iii . "! 1 ....... ~ ~,i~i ¸ ........ ..... jT I :,o i 2 3 4 5 6 7 8 9 I0 II 12 Siterr perf. Octo SiteTf perf. Octa : = X X X , • • 0@ • SV40 G C~ATGC ATCT C AATT AGICT A ....... Site][ CG TACGTAGAGTTAA TCGAT ; i • A • • • • t3 t4 ~5 t6 17 18 ,9 ao a 22 23 24 SiteT[ perf. Octa , Site][ perf. Octo I 0 . • oO perf. Octa GCTA GCAIATGCATGOAAATTAcG SiteTr CGTTACGTACGTTTAAT AT • A oo Figure 4. {See facing page for legend.) 1406 GENES & DEVELOPMENT .......... Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBP100 sequence recognition Figure 5. OBP binds to the HSV TAATGARAT motif. (A) Gel COMPETITOR TAAT TAAT+ retardation competition analysis. The competition experiments "1 (_9 ~9 + a + H were performed as described in the legend to Fig. 3, using the 14xBI7 ~u_o I ] ~.. <~ u_ o 14 x B17 restriction fragment as unlabeled competitor in 3.3- t- fold serial dilutions and continuously labeled probes of the TAATGARAT motif (top) and wild-type SV40 site II (bottom) as cloned oligonucleotides (see Methods) after excision from O -- rO (~ rO z c50 O O -- rd rd the pUC119 polylinker by digestion with EcoRI and HindIII. The source of OBP100 was heparin-agarose-fractionated HeLa T] cell nuclear extract. Only the retarded complexes are shown. ~'~::~:!~ The highest level of competitor represents a 100-fold greater ~i number of competitor binding sites to probe binding sites. The nonspecific competitor was HaelII-digested h DNA in equal TAAT concentration to the highest level of 14 x B17 fragment. (B) i' DEPC modification interference analysis of the TAATGARAT ,~i: . motif. The 8 7-bp NcoI-SphI restriction fragment spanning the -260 TAATGARAT motif of the ICP4 promoter was either 3' ..... (TAAT-) or 5' (TAAT +) end labeled at the NcoI site with the ......... SV40 Site rr ~:: Klenow fragment of DNA polymerase or T4 polynucleotide ki- ....... i nase, respectively. The end-labeled DNAs were modified with i DEPC, used in a preparative gel retardation with heparin- F X X X~X-~ I agarose-purified OBP100 and cleaved with piperidine, and the !!~!i : : CATGGC" resulting fragments fractionated by electrophoresis through a ATCTCATTACCGCC + ......... 16% denaturing polyacrylamide gel as described in Methods. GTACCGTAGAGTAATGGCGG - • I,, • O11 (A + G) Purine ladder; (F) free probe; (CI) retarded complex. ~;~:,:: : The boundaries of the TAATGARAT motif are identified by { TAAT I 4 5 6 the dashes with the arrow, indicating the 5' to 3' direction of the sequence. (C) Summary of chemical interference experi- ments. Shown is the double-stranded sequence of the -248 to -267 (left to right) region of the ICP4 promoter with the TAATGARAT motif on the bottom strand (arrow and dashes) and the octamer-related sequence shown on the top strand (bracket with x s for mismatches to the octamer consensus). The + (top strand) and - (bottom strand) designation correspond to TAAT + and TAAT- in B). The large dots indicate adenine modifications that result in a decrease in band intensity greater than threefold in the CI samples, and the small dots indicate an interference by DEPC modification of guanosines of two- to threefold. GARAT motif, and extend 2 bp to the left of the motif as Octa3 motif in SV40 site II (see Fig. 4). shown in Figure 5C. The 2-bp extension is the same as Given the low degree of sequence similarity between the GC dinucleotide interaction just upstream of the the TAATGARAT and octamer motifs, it is difficult to Figure 4. Changes within a core octamer motif affect the requirement for flanking sequences. (AI Gel retardation assay with contin- uously labeled DNA probes (see Methods and legend to Fig. 2) and partially purified OBP100 from a HeLd cell nuclear extract. The nucleotide sequence of the variable portion of each probe is shown with the designations corresponding to the lane numbers. The nonspecific control {lane 51 is a pUC119 polylinker with the coreA/coreA fragment sequence. Differences between the nucleotide sequence of the wild-type SV40 OBPl00-binding site II (lane 1) and the sequence of site II with flanking mutations (lane 21 or the perfect octamer in site II with (lane 4) or without (lane 31 flanking mutations are shown in boldface type. The boxed sequences shown above represent adjacent polylinker sequences, the brackets indicate the sequences similar (1 and 2) or identical (3 and 4) to the octamer consensus, and the + and - designations represent a qualitative assessment of OBP-binding affinity. The shorter 2-hr exposure of lanes 3-5 is of the same gel as the 8-hr exposure shown. {F) Free probe, (CII complexed probe. (BI DMS and DEPC modification interference analysis of the wild-type (site III and perfect octamer (perf. octal site II probes shown in A (probes 1 and 3, respectively). DNAs modified with DMS or DEPC were used in a preparative gel retardation with partially purified OBP100, and subsequently cleaved with piperidine and displayed on 16% (top strandl or 20% (bottom strandl denaturing polyacrylamide gels, as described in Methods. The top strand was displayed by 3' end labeling at the pUC119 polylinker EcoRI site, and the bottom strand by 3' end labeling at the HindlII site. (A + GI Purine ladders for each fragment generated by acid treatment. The free (FI and complexed (CI) probes are shown in pairs for each site. The brackets identify the positions of the octamer related Octa3 sequence (with Xsl or the perfect octamer, and the arrows show the relationship of the modification pattern to the sequence shown in C in a left-to-right direction. The arrowheads identify modifications of residues flanking the octamer motifs that have different effects between wild-type SV40 site II and the perfect octamer site II. (C) Summary of the chemical interference patterns. The guanosine DMS (AI and adenosine DEPC (e I modifications that interfere with complex formation are shown above and below the double-stranded sequence of the two binding sites. The brackets over the octamer motifs, boxed polylinker sequences, and mutations in boldface type are as described in A. The large triangles and circles indicate a reduction in band intensity threefold or greater in the complexed DNA probes; the small circles indicate a reduction of two- to threefold; effects less than twofold have not been included. The relative band intensities were assayed by densitometry. GENES & DEVELOPMENT 1407 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press Baumruker et al. know how the octamer and TAATGARAT motifs the presence of nonspecific competitor; the proteins that should be aligned. We favor the alignment shown in are bound specifically to the biotinylated probe are then Figure 5C, with the degenerate octamer motif centered purified by precipitation of the DNA probes with strep- over the opposite strand from the TAATGARAT motif tavidin-agarose beads and analyzed by gel electropho- sequence because of the patterns of DEPC interference, resis. which are very similar between this TAATGARAT Figure 6 shows a silver-stained SDS-polyacrylamide motif and SV40 site II (cf. Figs. 4 and 5). It should be gel resulting from a DNA affinity-precipitation assay. noted, however, that the strand containing the TAAT- The starting heparin-agarose fraction (lane 1) is acom- GARAT motif itself also has a 4-bp match {ATGA- GATG) to the octamer motif within the region of inter- ferring modifications. Indeed, even the other strand oi a Frog. Oligo. perfect octamer motif (ATTTGCAT) has a 4-bp match to the octamer consensus sequence. The TAATGARAT in- terference pattern shown in Figure 5B indicates, how- ever, that there is no obvious OBP100 interaction with • ° the bases 1-4 bp 5' of this TAATGARAT motif (GCGG on the bottom strand), which correpond to the region of lw octamer similarity noted in other TAATGARAT motifs 200 -- by Pruijn et al. (1986) and O'Hare and Goding (1988). Thus, OBP can bind to the TAATGARAT motif itself and not just flanking octamer motifs. In conclusion, the 116 -- interaction with the TAATGARAT motif described here 97-- indicates that octamer-binding protein can exhibit re- ~ (-- OBPIO0 markably flexible sequence recognition to sequences that are important regulatory elements in vivo. Our interpretation is likely to be controversial, how- 66-- ever, because O'Hare and Goding (1988) and Gerster and Roeder (1988) have emphasized the overlapping octamer motifs as the site of interaction by OBPs. But unlike the experiment shown in Figure 5, these studies did not ana- lyze the actual site of OBP interaction by chemical mod- ification interference analyses, which can identify the 4:3-- actual protein-binding site. It seems likely that when a TAATGARAT motif is overlapped by an octamer motif, the octamer motif will be the primary OBP100-binding site. But if the octamer motif is mutated without af- I 2 5 4 5 6 7 fecting the TAATGARAT motif [as in the TAAT9 and Figure 6. Microscale DNA affinity precipitation of OBP100 TAAT10 mutants described by O'Hare and Goding with the octamer-related SV40 site I and the dissimilar TAAT- (1988)], then the OBP100-binding site is likely to shift to GARAT motif. DNA affinity precipitation with biotinylated the TAATGARAT motif and interact in a manner analo- DNA fragments was performed with the heparin-agarose-frac- gous to the interaction with the solitary TAATGARAT tionated HeLa nuclear extract as described in Methods. Shown motif described here. As described below, VP16 may be is a silver-stained 8% polyacrylamide-SDS gel. The position of involved in generating just such a shift in binding site the molecular weight markers is shown to the left. (Lane 1) preference. 1/250 portion of the starting material used in DNA affinity pre- cipitation shown in lanes 2-7. (Lane 2) No biotinylated DNA in the precipitation. The DNA probes used were 14 × B17 (lane 3) and 6 × B17 dpm8 (lane 4) HindIII-PstI restriction fragments OBPIO0 is responsible for the interaction with the and coreA/coreA (lane 5), sphII/MsphI (wt SV40 site I; lane 6), TAA TGARA T motif and TAATGARAT (lane 7) ligated double-stranded synthetic The SV40 site I and -260 ICP4 TAATGARAT motif oligonucleotides. The coreA/coreA oligonucleotide (Ondek et share very little sequence similarity (4 out of 14 bp) but al. 1988) is a sequence unrelated to the octamer motif, and dpm8 is a double-point mutation (AAGGAAAG) within the form the same gel retardation complex and compete SV40 site I octamer motif. The position of OBP100 is marked. with one another for the same factor(s) in the partially The slightly faster mobility of OBP100 compared to the 97.4- purified HeLa cell extract. To test further whether kD rabbit muscle phosphorylase B marker, as opposed to its OBP100 was responsible for the interaction with the previous apparent molecular weight of 100 kD (Sturm et al. TAATGARAT motif in the heparin-agarose-purified 1987), is due to changes in polyacrylamide gel electrophoresis fraction and to test whether other polypeptides might be conditions (e.g., bis-acrylamide to acrylamide ratio in the gel). involved in complex formation, we used a DNA affinity- This polyacrylamide gel-specific shift in apparent molecular precipitation technique. This technique (Franza et al. weight may explain some of the differences in the reported mo- 1987) is a simple microscale assay in which biotinylated lecular weight of OBPs [e.g., the 90-kD OTF-1 (Fletcher et al. DNA probes are allowed to incubate with the extract in 1987) and the 92-kD NFIII (O'Neill and Kelly 1988)]. 1408 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBPIO0 sequence recognition plex mixture of proteins which, after precipitation in the core sequences, have been examined. Figure 7 shows an absence of biotinylated probe, generates a weak back- alignment of many of the binding sites described here in ground protein pattern (lane 2). When, as described pre- which either the similarity (Fig. 7A) or dissimilarity (Fig. viously (Sturm et al. 1987), reiterated and biotinylated 7B) between binding sites is emphasized. copies of the wild-type SV40 site I sequence (14xB17) In Figure 7, a 14-nucleotide-long sequence is shown are included in the precipitation assay, a single new for each binding site because this is the full size of the band representing OBP100 appears (lane 3). A pair of largest OBP100-binding site we have analyzed--the point mutations (dpm8) within SV40 site I, which pre- SV40 site II sequence (see Fig. 2C,D). To the right of each vents complex formation (Sturm et al. 1987), also pre- sequence comparison is shown the number of sequence vents precipitation of OBP100 without affecting the matches within the 14-bp stretch between the two adja- background pattern of precipitated proteins (lane 4). This cent sequences. These numbers overemphasize the dis- result further emphasizes that although OBP100 can similarity between sites that closely match the octamer display remarkably flexible sequence recognition, this consensus (i.e., SV40 site I, H2B, IgH, and U2) because recognition is still sequence specific and can be affected these sites are not very sensitive to changes in the se- by simple point mutations. quences flanking the octamer sequence (e.g., SV40 site I To assay the proteins that bind to the TAATGARAT is only 11 bp long; see Fig. 2B). Nevertheless, all of the motif in the heparin-agarose-purified HeLa cell frac- binding sites shown are identical to the two adjacent tion, we used ligated, double-stranded, synthetic oligon- ucleotides instead of restriction fragments. As expected, ligated synthetic oligonucleotides for an unrelated se- quence (SV40 coreA, lane 5) did not precipitate OBP100, I I SV40 but wild-type SV40 site I oligonucleotides (lane 6) did AGTATGCAAAGCAT Site I ,..°.,,. precipitate OBP100. The background pattern of nonspe- ,..°°.,. 9/14 cific proteins was very similar, but not identical, H2B CTTATGCAAATGAG :::::::: : whether restriction fragments (lanes 3 and 4) or syn- 9/14 thetic oligonucleotides (lanes 5-7) were used. The pat- IgH GAAATGCAAATTAC tern of proteins after DNA affinity precipitation with lO/14 the TAATGARAT motif containing oligonucleotide U2 GGCATGCAAATTCG (lane 7) is indistinguishable from the pattern generated 12/14 Perf. TGCATGCAAATTAG by the wild-type SV40 site I oligonucleotides (lane 6). Site II • °.°° .°.°.. °°.°, .,°,.. 11/14 Thus OBP100 is responsible for the TAATGARAT inter- SV40 TGCATCTCAATTAG actions, and in these experiments no ancillary proteins Site II 11/14 appear to be involved in OBP100 binding to the TAAT- Clone 12 TGCATCTCATTTCT GARAT motif. Together with the similar migration of • ....°.... . • . ...... ., 11/14 the SV40 site I and TAATGARAT motif gel regardation TAAT GGCATCTCATTACC complexes, this result suggests that OBP100 can bind I I independently of other factors to both of these two seemingly very different sequences. F X7 SV40 AGTATGCAAAGCAT Discussion Site I TCATACGTTTCGTA : ": : 4/14 Extreme degeneracy in DNA sequence recognition by GGCATCTCATTACC eukaryotic trans-acting factors has become evident in TAAT CCGTAGAGTAATGG recent studies of promoter-binding proteins (Johnson et al. 1987; Pfeifer et al. 1987; Sturm et al. 1987; Costa et Figure 7. (A) Sequence comparison of different OBPI00- al. 1988; Michael et al. 1988; this paper). A system that binding sites. Shown are six natural and two synthetic has been well characterized genetically is the yeast OBP100-binding sites as single-stranded sequence over a 14-bp HAP1 activator interaction with a binding site in the region that was determined by the largest binding site (SV40 II). upstream activator sequence UAS1 of the CYC1 gene Except for clone 12 (a relatively weak binding site), data for the and a binding site in the UAS of the CYC7 gene (Pfeifer interaction of OBP100 with these sites in the form of gel retar- dation and chemical modification interference analyses are rep- et al. 1987). Although HAP1 binding to the CYC1 UAS1 resented in this paper. The sequences are ordered to achieve the and CYC7 UAS shares many general features (e.g., major greatest degree of similarity between sites. The numbers to the and minor groove contacts), the sequences do not ex- right between sequences indicate the number of sequence hibit any obvious similarity (seven matches in a 23-bp matches (out of 14) between each pair. Base identity between region). Our previous (Sturm et al. 1987) and current sequences is shown by the colons. (B) Direct comparison of studies show that OBP 100, which binds paradoxically to SV40 site I and the TAATGARAT sequence. Double-stranded an extremely well-conserved motif, also displays re- sequence of 14 bp spanning each site is shown. The positions of markably degenerate sequence recognition. Through the octamer-related sequences are outlined by brackets above these studies, a relatively large number of OBP100- (SV40 site I) or below (TAATGARAT) the sequence. (x) Mis- binding sites, both with and without perfect octamer matches with the consensus octamer motif. GENES & DEVELOPMENT 1409 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press Baumruker et al. sequences shown at 9 or more positions out of 14 bind to a degenerate octamer motif (Octa3) within the (>60%) positions. In Figure 7A six of the eight se- SV40 enhancer (Sturm et al. 1987). However, the experi- quences are natural OBP100-binding sites, whereas two ments presented here show that the sequence context of of the sites (perfect octamer SV40 site II and clone 12) a binding site is critical and a mutation that has a dele- are mutants of SV40 site II sequences. By including these terious effect in one context may have little effect in a two latter binding sites, the transition between the per- different context. fect octamer motif in the U2 snRNA promoter-binding The context-dependent effect of mutations within a site and the TAATGARAT motif is shown with >75% DNA-binding site may also explain the ability of muta- similarity at each position. tions within the DNA-binding protein itself to have a The provocative result of the sequence comparison differential effect on binding to one site versus another. shown in Figure 7A is that although a clear progression For example, Pfeifer et al. (1987) have described a mu- can be observed between each adjacent sequence, the tant HAP1 protein (HAPI-18) that affects binding to the two sites at the end of the progression, SV40 site I and CYC1 UAS1 but not the CYC7 UAS. Such an effect the TAATGARAT motif, match at <30% of the posi- could be explained if the HAP 1-18 mutation interferes tions (Fig. 7B), which is little more than a random with a binding site contact that is essential for CYC1 match. Thus, these two apparently unrelated sequences interaction but not CYC7. By analogy, a mutation in bind the same factor OBP100, and their relationship can OBP100 that affects the ability to recognize residues be uncovered by the larger number of binding sites that flanking the octamer motif is likely to have a large effect have been analyzed (Fig. 7A). In contrast, in the studies on SV40 site II binding, where contacts with flanking of HAP1 (Pfeifer et al. 1987) and other degenerate se- residues are important, but not on binding to a perfect quence-binding proteins (Johnson et al. 1987; Costa et octamer motif, where flanking residue contacts are less important. These interpretations suggest that modifica- al. 1988; Michael et al. 1988), it has not been possible to tions of transcription factors may also have subtle ef- establish the relationship between very different binding fects on the patterns of sequence recognition. sites. The sequence comparison shown in Figure 7, to- gether with the loss of requirement for flanking se- The identification of many degenerate octamer motifs quences when a degenerate octamer motif (site II) is that can bind to ubiquitous OBP raises the question of converted to a perfect octamer motif, argues that why the octamer motif has been so well conserved in OBP100 can bind to very dissimilar sequences because many upstream promoter and enhancer regions. The an- few, if any, residues are obligatory for binding. Instead, swer may lie in that the degenerate motifs are almost all the sum of many independent contacts result in effec- exclusively of viral origin and that many of the degen- tive binding. erate motifs appear to bind OBP100 with a lower affinity It is not possible from the experimental approach de- than the perfect octamer motifs (see Fig. 3). Except for a scribed here to determine whether the effects on few cases [e.g., the human U1 promoter (Mangin et al. OBP100 binding by sequence alterations or chemical 1986)], the octamer motif is highly conserved in cellular modifications are due to changes in base contacts be- promoters, particularly the large number of immuno- tween OBP100 and the DNA-binding site or changes in globulin upstream promoter regions where the octamer DNA structure. For example, OBP100 is likely to bind to motif was identified first (Falkner and Zachau 1984; the octamer motif with the DNA in a particular confor- Parslow et al. 1984). This strongly suggests that a good mation (e.g., a certain width of the minor groove, curva- match to the octamer consensus is required for effective ture of the DNA helix, etc.), and this conformation may promoter activity. The preponderance of degenerate oc- be stabilized by flanking sequences. Thus, if putative tamer motifs in DNA viruses such as the SV40 early pro- base-specific contacts are debilitated by mutations moter and HSV immediate early gene promoters may in- within the core sequence, then the OBP100 interaction dicate a mechanism by which viruses can recruit host may be stabilized by those flanking sequences that favor cellular factors for viral transcription. For example, the correct DNA conformation for OBP100 binding. these viruses may express factors that are able to alter The influence of flanking sequences to compensate for the relative affinity of OBP 100 for perfect and degenerate deleterious mutations within a perfect octamer motif octamer motifs favoring the degenerate viral motifs. shows how difficult it is to draw conclusions from mu- Such a mechanism would allow a virus to reprogram the tational analyses of a functional protein-binding site. For DNA-binding specificity of cellular factors such that the example, Staudt et al. (1986) performed an extensive cellular binding sites do not compete for the host tran- mutational analysis of a consensus octamer/decamer scriptional apparatus. motif (ATGCAAATNA) cloned into the pUC18 poly- A good candidate for a factor that can reprogram the linker and showed that transversions in positions 1-7 DNA-binding specificity of OBP100 is the HSV late pro- all severely affected the interaction of the ubiquitous oc- tein VP16 (Vmw65). VP16 has been shown to activate tamer-binding factor NF-A1, which probably represents the HSV immediate early region genes, in part, through OBP100, as well as the lymphoid-specific factor NF-A2. the TAATGARAT motif but probably in conjunction These results suggested that nearly every position of the with a cellular intermediate (see Triezenberg et al. 1988a). Recently, O'Hare and Goding (1988)have shown highly conserved octamer motif was essential for inter- action with the octamer-binding proteins. Partially, for that VP16 expression can result in trans-activation of this reason, we were surprised to find that OBP100 could promoters containing octamer motifs and that octamer 1410 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBP100 sequence recognition Smith 1984; Kunkel 1985). Direct cloning was performed for and TAATGARAT motifs can compete for the same cel- the clones of the SV40 site I ~hown in Figure 2B (lanes 1 and 2) lular factor(s) in a HeLa cell nuclear extract. Our results and the SV40 site II clones shown in Figure 2C. The SV40 site I on direct interaction of OBP100 with the TAATGARAT clone in Figure 2B (lane 3) was generated by oligonucleotide-di ~ motif suggest that OBP 100 itself may be the cellular me- rected insertion of the 3' A, whereas the clones containing a diator of VP16 trans-activation. The interaction of mutated SV40 site II in Figures 2D and 4A were generated by OBP100 with the TAATGARAT motif analyzed here, oligonucleotide-directed mutagenesis of the wild-type SV40 however, is weaker than the interaction with the perfect site II sequence shown in Figure 2C (lane 3). Direct cloning was octamer motifs. Thus, OBP100 sequence recognition performed into the plasmid p[3e-A, which contains the human may be altered by interaction with VP16 to recognize B-globin gene (Ondek et al. 1987), with an upstream deletion the TAATGARAT motif more favorably. In the cases from -37 to -127 of the [3-globin promoter (T. Baumruker, where the TAATGARAT motif is overlapped by an oc- unpubl.), using single-stranded oligonucleotides with PstI- and XbaI-compatible termini. A 100-fold molar excess of single- tamer motif, VP16 may shift the binding site preference stranded phosphorylated oligonucleotide was ligated overnight from the octamer motif to the TAATGARAT motif or to PstI-digested and dephosphorylated pBe-A vector. Excess may facilitate dimerization of OBP100 at these sites. oligonucleotide was removed by two sequential ethanol precip- itations in the presence of 2 M ammonium acetate, and the plasmid redigested with XbaI. Ligation was then performed ei- Methods ther after 50-fold dilution of the DNA or after purification of With only a few modifications, the methods used in this paper the resulting large vector fragment on a low-melting-point for HeLa cell extract preparation, heparin-agarose chromatog- agarose gel. After ligation, the DNAs were transfected into raphy, gel retardation and chemical modification interference MV1193 cells. Loss of the 2.2-kb 8-globin fragment was used to assays, and sequence-specific DNA affinity-precipitation assays screen rapidly for successful cloning of oligonucleotides. Point are described by Sturm et al. (1987). mutations were introduced using suitable single-stranded pa- rental plasmid DNAs prepared from the ung- dut- E. coli strain BW313 (Kunkel 1985) as templates for oligonucleotide- Origin and cloning of octamer and TAATGARAT probes directed mutagenesis (Zoller and Smith 1984). The nucleotide The murine IgH enhancer, Ad2 ITR, human U2 snRNA distal sequence of all probes was confirmed by complete DNA se- element, and chicken histone H2B octamer motifs were ob- quencing using the dideoxy method (Sanger et al. 1977). tained as restriction fragments from wild-type sequences or cloned synthetic oligonucleotides: The IgH enhancer probe was the HindIII-PstI fragment from pBIgH + described previously Preparation of probes for gel retardation and chemical {Sturm et al. 1987), 5' or 3' end labeled at the HindIII site; the modification interference assays Ad2 ITR probe was a 72-bp EcoRI-SalI fragment derived from Fragments used for chemical modification interference were ei- the left end of the Ad2 genome in the plasmid pLAS 108 (Ta- ther 3' end labeled with the large fragment of DNA polymerase manoi and Stillman 1983), 5' or 3' end labeled at the EcoRI I or 5' end labeled with T4 polynucleotide kinase, modified site; the U2 snRNA probe was an EcoRI-AvaI fragment from with DEPC or DMS, and subsequently purified by preparative the clone NE (Ares et al. 1987)i 3' end labeled at either the polyacrylamide gel electrophoresis, as described previously EcoRI or AvaI sites; and the H2B promoter fragment was an (Sturm et al. 1987). For gel retardation assays with wild-type EcoRI-PstI fragment, either 5' or 3' end labeled at the EcoRI and mutant SV40 binding sites I and II or the TAATGARAT site, from a pUG119 clone of the paired synthetic oligonucleo- motif, we used continuously labeled probes generated by de tides GATCCTAGCCCCTCTATGCAAATGAGAAGCATT- novo synthesis of double-stranded DNA from a single-stranded CCTTTCGAATT (top strand only) in the SmaI site. template. This labeling procedure ensured that all probes were The HSV ICP4 promoter fragment containing the -260 labeled reproducibly to the same specific activity. To generate TAATGARAT motif and used for DEPC modification interfer- these continuously labeled probes, the 'universal' sequencing ence was prepared as an SphI-NcoI fragment from the plasmid primer was annealed to single-stranded pUGll9 DNA, gener- pIE3CAT (DeLuca and Schaffer 1985), which was generously ated by superinfection with the helper phage M13K07 (Vieira provided by Dr. N. DeLuca. For the competition assay and and Messing 1987), and extended by incubation with the large DNA affinity-precipitation assay, we used paired oligonucleo- fragment of DNA polymerase I and nucleotide triphosphates in tides that derived from the -265 to -245 region of the ICP4 which one nucleotide was s2p labeled. After 10 rain incubation promoter (numbering system from Mackem and Roizman at ambient temperature, the reaction was terminated by addi- 1982a) {GCGGTAATGAGATGCCATGCGGCTCGA and tion of ammonium acetate to a final concentration of 2 M and AGCCGCATGGCATCTCATTACCGCTCG; XhoI linker se- subsequent phenol/chloroform {1:1) extraction and ethanol quences are underlined) and that generated 3-bp 3' overhangs precipitation. After digestion with EcoRI and HindIII, the re- after annealing. For the competition assay, end-repaired an- sulting probes were purified by preparative polyacrylamide gel nealed oligonucleotides were cloned as one copy into the Sinai electrophoresis and elution into 10 mM Tris-HCl (pH 8.0), 1 mM site of pUG119. For DNA affinity precipitation the synthetic EDTA at 37°C. The resulting eluate was collected, and the oligonucleotides themselves were concatenated by ligation (see DNA concentration adjusted to 10,000 cpm (Cerenkov)/~l for below). direct use in the gel retardation. Generation of wild-type and mutant SV40 OBPlOO-binding sites I and H OBP1 O0 purification, gel retardation, and chemical modification interference Individual single-stranded synthetic oligonucleotides corre- sponding to SV40 binding sites I and II either were cloned di- Except for the highly purified OBP100 used in Figure 1A, which rectly, as described (Derbyshire et al. 1986), or used for site-di- was the same purified OBP100 protein described previously rected in vitro mutagenesis of existing binding sites (Zoller and (Sturm et al. 1987), the source of OBP activity was the 0.1-0.2 GENES & DEVELOPMENT 1411 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press Baumruker et al. N KC1 heparin-agarose column fraction from nuclear HeLa cell Campbell, M.E.M., J.W. Palfreyman, and C.M. Preston. 1984. extracts prepared as described previously (Sturm et al. 1987), Purification of herpes simplex virus DNA sequences which but the Bio-Gel P-10 column chromatography was replaced by encode a trans-acting polypeptide responsible for stimula- dialysis in some cases. We modified the binding buffer for the tion of immediate early transcription. ]. Mol. Biol. 180: 1- preparative gel retardations of chemically modified fragments by removal of MgC12 and addition of 10 mM EDTA. This modi- Costa, R.H., D.R. Grayson, K.G. Xanthopoulos, and J.E. Dar- fication inhibited nucleases in the extract but did not affect the nell, Jr. 1988. A liver-specific DNA-binding protein recog- gel retardation patterns themselves. Otherwise, the procedures nizes multiple nucleotide sites in regulatory regions of were as described previously. The exact DEPC modification re- transthyretin, al-antitrypsin, albumin, and simian virus 40 sponsible for the interference patterns is unknown but probably genes. Proc. Natl. Acad. Sci. 85: 3840-3844. reflects carbethoxylation at the N7 position. This is because Davidson, I., C. Fromental, P. Augereau, A. Wildeman, M. even though DEPC can modify other positions on adenines (i.e., Zenke, and P. Chambon. 1986. Cell-type specific protein N1 and N3) these are probably not sensitive to the piperidine binding to the enhancer of simian virus 40 in nuclear ex- cleavage step that displays sites of modification. Modification tracts. Nature 323: 544-548. at the N7 position is consistent with the patterns of DEPC reac- DeLuca, N.A. and P.A. Schaffer. 1985. Activation of imme- tivity to left-handed Z DNA (see Herr 1985). diate-early, early, and late promoters by temperature-sensi- tive and wild-type forms of herpes simplex virus type 1 pro- tein ICP4. Mol. Cell. Biol. 5: 1997-2008. DNA affinity precipitation Derbyshire, K.M., I.l- Salvo, and N.D.F. Grindley. 1986. A simple and efficient procedure for saturation mutagenesis The microscale biotin/avidin DNA affinity-precipitation assay using mixed oligodeoxynucleotides. Gene 46: 145-152. was carried out as described (Franza et al. 1987; Sturm et al. Dynan, W.S. and R. Tjian. 1983. The promoter-specific tran- 1987), either using biotinylated 14xB17 (Ondek et al. 1987) or scription factor Spl binds to upstream sequences in the 6xB17 dpm8 (W. Herr, unpubl.) restriction fragments, or oligon- SV40 early promoter. Cell 35: 79-87. ucleotides containing the SV40 OBP100 site I (sphII/MsphI; Efstratiadis, A., J.W. Posakony, T. Maniatis, R.M. Lawn, C. Ondek et al. 1988), the SV40 coreA motif (coreA/coreA; Ondek O'Connell, R.A. Spritz, J.K. DeRiel, B.G. Forget, S.M. et al. 1988), or the TAATGARAT motif separated by XhoI rec- Weissman, J.L. Slightom, A.E. Blechl, O. Smithies, F.E. ognition sites were kinased, annealed, and ligated. The extent Barale, C.C. Shoulders, and N.J. Proudfoot. 1980. The struc- of otigonucleotide ligation was determined by gel electropho- ture and evolution of the human I]-globin gene family. Cell resis (average, more than four), and the ligated products were biotinylated, after phenol/chloroform extraction and ethanol 21: 653-668. Falkner, F.G. and H.G. Zachau. 1984. Correct transcription of precipitation. Approximately 0.4 ~g of biotinylated DNA, ei- an immunoglobulin K gene requires an upstream fragment ther as restriction fragments or ligated oligonucleotides, was containing conserved sequence elements. Nature 310: 71- used to directly identify OBP100 protein in 350 ixl of heparin- agarose-fractionated HeLa cell extract (Sturm et al. 1987). Pro- 74. teins precipitated after attachment of the biotinylated DNAs to Falkner, F.G., R. Mocikat, and H.G. Zachau. 1986. Sequences streptavidin-agarose beads and several washes were resolved closely related to an immunoglobulin gene promoter/en- by SDS-polyacrylamide gel electrophoresis and visualized by hancer element occur also upstream of other eukaryotic and silver staining. of prokaryotic genes. Nucleic Acids Res. 14: 8819-8827. Fletcher, C., N. Heintz, and R.G. Roeder. 1987. Purification and characterization of OTF-1, a transcription factor regulating Acknowledgments cell cycle expression of a human histone H2b gene. Cell 51: 773-781. We thank J. Clarke for aiding in heparin-agarose purification of Franza, B.R., Jr., S.F. losephs, M.Z. Gilman, W. Ryan, and B. OBP100 and establishment of the DNA affinity-precipitation Clarkson. 1987. Characterization of cellular proteins recog- technique; N. DeLuca for kindly providing the plasmid nizing the HIV enhancer using a microscale DNA-affinity pIECAT; M. Zoller for synthetic oligonucleotides; R. Duffy and precipitation assay. Nature 330: 391-395. M. Goodwin for help in preparing the manuscript, D. Greene, J. Gerster, T. and R.G. Roeder. 1988. A herpesvirus trans-acti- Duffy, and M. Ockler for artwork; and B.R. Franza, M. Gilman, rating protein interacts with transcription factor OTF-1 and N. Hernandez, and B. Stillman for critical readings of the manu- other cellular proteins. Proc. Natl. Acad. Sci. 85: 6347- script. T.B. is a recipient of an EMBO long-term fellowship; R.S. is a recipient of a Cancer Research Institute postdoctoral fel- Gerster, T., P. Matthias, M. Thali, I. Jiricny, and W. Schaffner. lowship, New York; W.H. is a Rita Allen Foundation Scholar. 1987. Cell type-specificity elements of the immunoglobulin This work was supported by U.S. Public Health Services grant heavy chain gene enhancer. EMBO ]. 6: 1323-1330. CA- 13106 from the National Cancer Institute. Harvey, R.P., A.J. Robins, and J.R.E. Wells. 1982. Independently evolving chicken histone H2B genes: identification of a References ubiquitous H2B-specific 5' element. Nucleic Acids Res. 10: 7851-7863. Ares, M., Jr., M. Mangin, and A.M. Weiner. 1985. Orientation- Herr, W. 1985. Diethyl pyrocarbonate: A chemical probe for dependent transcriptional activator upstream of a human secondary structure in negatively supercoiled DNA. Proc. U2 snRNA gene. Mol. Cell. Biol. 5: 1560-1570. Natl. Acad. Sci. 82: 8009-8013. Ares, M. Jr., J.-S. Chung, L. Giglio, and A.M. Weiner. 1987. Dis- Herr, W. and J. Clarke. 1986. The SV40 enhancer is composed of tinct factors with Spl and NF-A specificities bind to adja- multiple functional elements that can compensate for one cent functional elements of the human U2 snRNA gene en- another. Cell 45: 461-470. hancer. Genes Dev. 1: 808-817. Benoist, C., K. O'Hare, R. Breathnach, and P. Chambon. 1980. Johnson, P.F., W.H. Landschultz, B.J. Graves, and S.L. The ovalbumin gene-sequence of putative control regions. McKnight. 1987. Identification of a rat liver nuclear protein Nucleic Acids Res. 8: 127-142. that binds to the enhancer core element of three animal vi- 1412 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBPIO0 sequence recognition ruses. Genes Dev. 1: 133-146. Preston, C.M, M.G. Cordingley, and N.D. Stow. 1984. Analysis Kristie, T.M. and B. Roizman. 1984. Separation of sequences of DNA sequences which regulate the transcription of a defining the basal expression from those conferring c~ gene Herpex simplex virus immediate early gene. J. Virol. recognition within the regulatory domains of Herpes sim- 50: 708-716. plex virus 1 e~ genes. Proc. Natl. Acad. Sci. 81: 4065-4069. Preston, C.M., M.C. Frame, and M.E.M. Campbell. 1988. A • 1987. Host cell proteins bind to the cis-acting site re- complex formed between cell components and an HSV quired for virion-mediated induction of herpes simplex virus structural polypeptide binds to a viral immediate early gene 1 a genes. Proc. Natl. Acad. Sci. 84: 71-75. regulatory DNA sequence. Cell 52: 425-434. Kunkel, T.A. 1985. Rapid and efficient site-specific muta- Pruijn, G.I.M., W. van Driel, and P.C. van der Vliet. 1986. Nu- genesis without phenotypic selection. Proc. Natl. Acad. Sci. clear factor III, a novel sequence-specific DNA-binding pro- 82: 488-492. tein from HeLa cells stimulating adenovirus DNA replica- LaBella, F., H.L. Sire, R.G. Roeder, and N. Heintz. 1988. Cell- tion. Nature 322: 656-659. cycle regulation of a human histone H2b gene is mediated Pruijn, G.I.M., W. van Driel, R.T. van Miltenburg, and P.C. van by the H2b subtype-specific consensus element. Genes Dev. der Vliet. 1987. Promoter and enhancer elements containing 2: 32-39. a conserved sequence motif are recognized by nuclear Landolfi, N.F., J.D. Capra, and P.W. Tucker. 1986. Interaction of factor III, a protein stimulating adenovirus DNA replica- cell-type-specific nuclear proteins with immunoglobulin tion. EMBO J. 6: 3771-3778. Vtt promoter region sequences. Nature 323: 548-551. Rosales, R., M. Vigneron, M. Macchi, I. Davidson, ].H. Xiao, Mackem, S. and B. Roizman. 1982a. Differentiation between and P. Chambon. 1987. In vitro binding of cell-specific and promoter and regulator regions of herpes simplex virus 1: ubiquitous nuclear proteins to the octamer motif of the The functional domains and sequence of a movable e, regu- SV40 enhancer and related motifs present in other pro- lator. Proc. Natl. Acad. Sci. 79: 4917-4921. moters and enhancers. EMBO J. 6: 3015-3025. • 1982b. Structural features of the herpes simplex virus c~ Rosenfeld, P-I., E.A. O'Neill, R.I. Wides, and T.]. Kelly. 1987. gene 4, 0, and 27 promoter-regulatory sequences which Sequence-specific interactions between cellular DNA- confer ~ regulation on chimeric thymidine kinase genes. J. binding proteins and the adenovirus origin of DNA replica- Virol. 44: 939-949• tion. Mol. Cell. Biol. 7: 875-886. Mangin, M., M. Ares, Jr., and A.M. Wiener. 1986. Human U2 Sanger, F., S. Nicklen, and A.R. Coulson. 1977. DNA se- small nuclear RNA genes contain an upstream enhancer. quencing with chain-terminating inhibitors. Proc. Natl. EMBO J. 5: 987-995• Acad. Sci. 74: 5463-5467. Marsden, H.S., M.E.M. Cambpell, L. Haarr, M.C. Frame, D.S. Scheidereit, C., A. Heguy, and R.G. Roeder. 1987. Identification Parris, M. Murphy, R.G. Hope, M.T. Muller, and C.M. and purification of a human lymphoid-specific octamer Preston. 1987. The 65,000-Mr DNA binding and virion binding protein (OTF-2) that activates transcription of an trans-inducing proteins of Herpes simplex virus type 1. J. immunoglobulin promoter in vitro• Cell 51: 783- 793. Virol. 61: 2428-2437. Singh, H., R. Sen, D. Baltimore, and P.A. Sharp• 1986. A nuclear Mattaj, I.W., S. Lienhard, J. Jiricny, and E.M. De Robertis. 1985. factor that binds to a conserved sequence motif in transcrip- An enhancer-like sequence within the Xenopus U2 gene tional control elements of immunoglobulin genes. Nature promoter facilitates the formation of stable transcription 319: 154-158. complexes. Nature 316: 163-167. Staudt, L.M., H. Singh, R. Sen, T. Wirth, P.A. Sharp, and D. McKnight, J.L.C., T.M. Kristie, and B. Roizman. 1987. Binding Baltimore. 1986. A lymphoid-specific protein binding to the of the virion protein mediating a gene induction in herpes octamer motif of immunoglobulin genes. Nature 323: 640- simplex virus 1-infected cells to its cis site requires cellular proteins. Proc. Natl. Acad. Sci. 84: 7061-7065. Sturm, R., T. Baumruker, B.R. Franza, Jr., and W. Herr. 1987. A Michael, N., D. Spector, P. Mavromara-Nazos, T.M. Kristie, 100-kD HeLa cell octamer binding protein {OBP100) in- and B. Roizman. 1988. The DNA-binding properties of the teracts differently with two separate octamer-related se- major regulatory protein ~4 of herpes simplex viruses. quen~ces within the SV40 enhancer. Genes Dev. 1: 1147- Science 239: 1531-1534. 1160: O'Hare, P. and C.R. Goding. 1988. Herpes simplex virus regula- Tamanoi, F. :and B.W. Stillman. 1983. Initiation of adenovirus tory elements and the immunoglobulin octamer domain DNA replication in vitro requires a specific DNA sequence. bind a common factor and are both targets for virion tran- Proc. Natl. Acad. Sci. 80: 6446-6450. sactivation. Cell 52: 435-445. Triezenberg, S.J., R.C. Kingsbury, and S.L. McKnight. 1988a. Ondek, B., L. Gloss, and W. Herr. 1988. The SV40 enhancer Functional dissection of VP16, the transactivator of herpes contains two distinct levels of organization• Nature simplex virus immediate early gene expression. Genes Dev. 333: 40-45. 2: 718-729. Ondek, B., A. Shepard, and W. Herr. 1987. Discrete elements Triezenberg, S.J., K.L. LaMarco, and S.L. McKnight. 1988b. Evi- within the SV40 enhancer region display different cell spe- dence of DNA: protein interactions that mediate HSV-1 im- cific enhancer activities. EMBO J. 6: 1017-1025. mediate early gene activation by VP16. Genes Dev. 2: 730- O'Neill, E.A. and T.J. Kelly. 1988. Purification and character- ization of nuclear factor III (origin recognition protein C) a Vieira, J. and J. Messing. 1987. Production of single-stranded sequence-specific DNA binding protein required for effi- plasmid DNA. Methods Enzymol. 153: 3-11. cient initiation of adenovirus DNA replication. J. Biol. Chem. 263: 931-937. Wirth, T., L. Staudt, and D. Baltimore. 1987. An octamer oli- gonucleotide upstream of a TATA motif is sufficient for Parslow, T.G., D.L. Blair, W.J. Murphy, and D.K. Granner. 1984. lymphoid-specific promoter activity. Nature 329:174-178. Structure of the 5' ends of immunoglobulin genes: A novel conserved sequence. Proc. Natl. Acad. Sci. 81: 2650-2654. Zoller, M.J. and M. Smith. 1984. Oligonucleotide-directed mu- Pfeifer, K., T. Prezant, and L. Guarente. 1987. Yeast HAP1 acti- tagenesis: A simple method using two oligonucleotide vator binds to two upstream activation sites of different se- primers and a single-stranded DNA template. DNA 3: 479- quence. Cell 49: 19-27. 488. GENES & DEVELOPMENT 1413 Downloaded from genesdev.cshlp.org on December 2, 2021 - Published by Cold Spring Harbor Laboratory Press OBP100 binds remarkably degenerate octamer motifs through specific interactions with flanking sequences. T Baumruker, R Sturm and W Herr Genes Dev. 1988, 2: Access the most recent version at doi:10.1101/gad.2.11.1400 This article cites 55 articles, 25 of which can be accessed free at: References http://genesdev.cshlp.org/content/2/11/1400.full.html#ref-list-1 License Receive free email alerts when new articles cite this article - sign up in the box at the top Email Alerting right corner of the article or click here. Service Copyright © Cold Spring Harbor Laboratory Press
Genes & Development – Unpaywall
Published: Nov 1, 1988