The Bacteriophage T4 Transcriptional Activator MotA Accepts Various Base-pair Changes within its Binding Sequence

Article No. jmbi.1998.2373 available online at http://www.idealibrary.com on

J. Mol. Biol. (1999) 285, 931±944

The Bacteriophage T4 Transcriptional Activator MotA Accepts Various Base-pair Changes within its Binding Sequence Philip Marshall, Mridula Sharma and Deborah M. Hinton* Laboratory of Molecular and Cellular Biology, National Institute of Diabetes and Digestive and Kidney Diseases National Institutes of Health Bethesda, MD 20892, USA

During infection, bacteriophage T4 regulates three sets of genes: early, middle, and late. The host RNA polymerase is capable of transcribing early genes, but middle transcription requires the T4-encoded transcriptional activator, MotA protein, and the T4 co-activator, AsiA protein, both of which bind to the sigma 70 (s70) subunit of RNA polymerase. MotA also binds a DNA sequence (a MotA box), centered at position ÿ30. The identi®cation of more than 20 middle promoters suggested that a strong match to the MotA box consensus sequence (t/a)(t/a)TGCTT(t/c)A was critical for MotA activation. We have investigated how speci®c base changes within the MotA box sequence affect MotA binding and activation in vitro, and we have identi®ed seven new middle promoters in vivo. We ®nd that an excellent match to the s70 ÿ10 consensus sequence, rather than an excellent match to the MotA box consensus sequence, is an invariant feature of MotA-dependent promoters. Many single base changes in the MotA box are tolerated in binding and activation assays, indicating that there is more ¯exibility in the sequence requirements for MotA than was previously appreciated. We also ®nd that using the natural T4 DNA, which contains glucosylated, 5-hydoxymethylated cytosine residues, affects the ability of particular MotA box sequences to activate transcription. We suggest that MotA and AsiA may function like certain eukaryotic TAFs (TATA binding protein (TBP) associated factors) whose binding to TBP results in transcription from new core promoter sequences. # 1999 Academic Press

*Corresponding author

Keywords: activation; AsiA; bacteriophage; MotA; transcription

Introduction Transcription from bacteriophage T4 middle promoters (reviewed by Stitt & Hinton, 1994) is accomplished by the host RNA polymerase, but it requires both an activator, the T4 MotA protein (Mattson et al., 1974, 1978), and a co-activator, the T4 AsiA protein (Ouhammouch et al., 1994). The MotA activator binds to middle promoter DNA (Hinton, 1991; Schmidt & Kreuzer, 1992; MarchPresent addresses: P. Marshall, Ottawa General Hospital Research Institute, 501 Smythe Road, Ottawa, Ontario, Canada K1H 8L6; M. Sharma, AgResearch, Ruakura, East Street, Hamilton, New Zealand. Abbreviations used: s70, sigma 70 subunit of RNA polymerase; TBP, TATA binding protein; TAFs, TBP associated factors; BSA, bovine serum albumin. E-mail address of the corresponding author: [email protected] 0022-2836/99/030931±14 $30.00/0

Amegadzie & Hinton, 1995), but it also binds to the sigma 70 (s70) subunit of RNA polymerase (Gerber & Hinton, 1996). AsiA also binds to s70 (Stevens, 1972, 1976; Stevens & Rhoton, 1975), and recent work has shown that this binding inhibits the interaction of s70 with its consensus ÿ35 sequence (Adelman et al., 1997; Colland et al., 1998). The identi®cation of more than 20 T4 middle promoters (Nivinskas et al., 1989; Stitt & Hinton, 1994) revealed two highly conserved sequence elements in the ÿ10 and ÿ30 regions of these promoters. The ÿ10 element matches the ÿ10 consensus sequence for s70, TAnnnT (Siebenlist et al., 1980; Harley & Reynolds, 1987). This suggests that T4 middle promoters, like typical Escherichia coli promoters, require an interaction between domain 2.4 of s70 and the ÿ10 region of the promoter (Siegele et al., 1989; Waldburger et al., 1990; Dombroski, 1997). Footprints of the polymerase/ # 1999 Academic Press

932 AsiA/MotA complex at the middle promoter PuvsX support this idea (Hinton et al., 1996a). However, PuvsX, like other middle promoters, lacks a match to the s70 ÿ35 consensus sequence, and footprints of polymerase/AsiA/MotA at PuvsX indicate that the presence of MotA and AsiA alters the proteinDNA contacts made by polymerase in the ÿ35 region (Hinton et al., 1996a; M.S., P.M. & D.M.H., unpublished results). Instead of a s70 ÿ35 consensus sequence, middle promoters have a different consensus sequence of (t/a)(t/a)TGCTT(t/c)A (a MotA box) centered at position ÿ30 (Brody et al., 1983; Guild et al., 1988). Within the set of middle promoters, this MotA box sequence is highly conserved. This conservation has implied that there might be rigid sequence requirements within the MotA box sequence for MotA binding and activation. Here, we have examined how changes in the MotA box sequence affect MotA binding and transcriptional activation in vitro and in vivo. We have found that the set of acceptable base-pairs in the MotA box sequence is broader than was previously inferred from the consensus sequence. In addition, we have found that for the middle promoter, P46 (Guild et al., 1988; Hsu & Karam, 1990), cytosine modi®cation, which is present in natural T4 DNA, affects MotA-dependent transcription.

Results Deletion and mutational analyses of the T4 middle promoter PuvsX Analyses of the T4 middle promoter PuvsX showed that the sequences between ÿ18 and ÿ38 were important for MotA activation. Although sequences upstream of position ÿ38 stimulated transcription under certain conditions, they were not required for MotA activation (MarchAmegadzie & Hinton, 1995). However, MotA protected large regions of PuvsX from DNase I digestion: from ÿ59 to ÿ45, from ÿ40 to ÿ25 and from ‡40 to ‡57 on the template strand (MarchAmegadzie & Hinton, 1995; designated as regions I, I, and III, respectively, in Figure 1). These areas include matches to the MotA box consensus sequence centered at ÿ51 (region I) and at ÿ30 and ÿ35 (region II) as well as a downstream region which lacks a match to the MotA box sequence (region III). We had found that mutation of the overlapping MotA boxes at ÿ35 did not affect MotA activation of PuvsX (Hinton et al., 1996a). To explore the contributions of other regions for transcription and binding, we constructed plasmids in which these sequences were mutated or deleted (Figure 1). We found that the T4 sequences between ÿ34 and ‡3 were suf®cient for MotA activation, and only the MotA box at ÿ30 was critical. Mutation, inversion, or deletion of the box at ÿ51, deletion of the ‡40 to ‡57 region, or increasing the spacing between the boxes at ÿ51 and ÿ30 by 5 bp

Sequence Requirements for T4 MotA Activation

(half a helical turn) did not signi®cantly affect transcription. DNase I footprint analyses (results not shown) showed that mutation of the ÿ35 MotA boxes had no effect on the footprint, but mutation of the MotA box at ÿ30 speci®cally eliminated protection of region II (see Figure 1 for the positions of regions I, II, and III). Thus, the MotA box at ÿ30 is critical both for MotA binding to region II and for activation of transcription. Although the MotA box at ÿ51 and positions ‡40 to ‡57 were not necessary for MotA-dependent transcription, these sequences were required for DNase I protection of regions I and III, respectively. Furthermore, inserting 5 bp between positions ÿ41 and ÿ42 simply increased the spacing between protected regions I and II by 5 bp. This indicates that DNA phasing is not important for MotA binding to regions I and II. Taken together, the DNase I footprints indicate that MotA binds independently to each of the three protected regions of PuvsX. Sequence requirements for MotA binding and activation in vitro The MotA box at ÿ30 of PuvsX has an excellent match to the highly conserved consensus sequence of (t/a) (t/a) TGCTT (t/c) A (Brody et al., 1983; Guild et al., 1988). The CTT at the center of the consensus sequence (positions ÿ30 to ÿ28 of PuvsX; see Figure 1) is invariant among the original set of T4 middle promoters (Nivinskas et al., 1989; Stitt & Hinton, 1994). However, MotA protected region III of PuvsX (Figure 1) from DNase I digestion, and this region lacks a CTT (March-Amegadzie & Hinton, 1995). In addition, there was a twofold activation by MotA when the CTT of the ÿ30 MotA box of PuvsX was changed to ACT (Figure 1). Therefore, we analyzed the sequence requirements for MotA binding and transcriptional activation in more detail. Table 1 summarizes the ability of various mutant versions of the MotA box sequence to compete with a wild-type MotA box for binding to MotA. In assays with 50-fold excess competitor, mutation of the G at position ÿ31 to an A or T or mutation of the C at position ÿ30 to a T was deleterious. However, other single base changes at ÿ33, ÿ32, ÿ30, ÿ29, ÿ28, or ÿ26 resulted in reasonable competitors. In addition, the duplex that contained positions ‡39 to ‡53 downstream of the start of PuvsX transcription (region III in Figure 1) was a good competitor in this assay. In the presence of 300-fold excess competitor, every mutant MotA box, even ones with multiple changes, could compete with the wild type MotA box sequence for binding. We also examined the effect of the MotA box mutations on MotA activation of transcription in vitro. In the reactions shown in Figure 2 and Table 2, multiple rounds of transcription were performed using T4-modi®ed RNA polymerase isolated from infected cells. In other assays, single and multiple round transcriptions were performed

Figure 1. PuvsX and mutant variants. Top, PuvsX sequence is shown from ÿ59 to ‡57 with ‡1 designating the start of transcription. The match to the MotA box consensus sequence at ÿ30 is shown by the ®lled arrow. Other matches to the MotA box sequence (centered at ÿ35 and ÿ51) are shown as broken arrows. Regions on the template strand that are protected by MotA from DNase I digestion (March-Amegadzie & Hinton, 1995) are designated as I, II, and III. Underneath are shown the sequences of various PuvsX promoter mutants; sequence changes are in bold and underlined. The sequences corresponding with the ÿ10 and ÿ30 regions in the various mutant promoters are boxed along with the s70 ÿ10 sequence and the ÿ30 MotA box sequence present in PuvsX. Mutant/wild-type represents the amount of RNA seen in the presence of MotA from the mutant promoter relative to that from a PuvsX wild-type promoter, containing positions ÿ94 to ‡83, present in the same reaction mixture. Fold activation is calculated as (amount of RNA synthesized by modi®ed polymerase plus MotA)/(amount of RNA synthesized by modi®ed polymerase alone). For the wild-type promoter, the fold activation was 14(3).

934

Sequence Requirements for T4 MotA Activation Table 1. Binding of MotA to mutant MotA box sequences How effective is the competitorb at an excess of 50-fold 300-fold

MotA box present as competitora PuvsX Mutants ÿ33T ! G ÿ32T ! A ÿ31G ! A ÿ31G ! T ÿ30C ! T ÿ30C ! G ÿ30C ! A ÿ29T ! C ÿ28T ! G ÿ26A ! G ‡39 ! ‡ 53 Multiple mutant 1 Multiple mutant 2

ÿ34 ÿ 26 TTTGCTTAA

100%

100%

TGTGCTTAA TTAGCTTAA TTTACTTAA TTTTCTTAA TTTGTTTAA TTTGGTTAA TTTGATTAA TTTGCCTAA TTTGCTGAA TTTGCTTAG AAAGATTAA TTATACTAA

81% 30% 4% 0% 0% 65% 92% 60% 56% 52% 65% 0%

100% 70% 59% 44% 66% 94% 101% 88% 84% 85% 92% 19%

TTATACTTA

0%

13%

a Except for the duplexes ‡39 ! ‡53 and multiple mutant 2, the MotA box sequence (shown underlined below) was present within the double-stranded oligomer:

50 GATCCTATTTGCTTAATAATCCA 30 30 GATAAACGAATTATTAGGTAT 50 Numbers refer to positions within the PuvsX sequence; the base change(s) from the PuvsX sequence is shown in bold. The ‡39 ! ‡53 oligo contains positions from ‡39 to ‡53 downstream of the start of PuvsX (see Figure 1),which are present (shown underlined below) within the sequence: 50 GATCCATTGAAAAGATTAATAATCCA 30 GTAACTTTTCTAATTATTAGGTAT 50 Multiple mutant 2 is present (underlined) within the sequence: 50 AATTATTATACTTAGATTA 30 30 TTAATAATATGAATCTAAT 50 b Competitive effectiveness was calculated by the formula given in Materials and Methods. For experiments with 50-fold excess of competitor to DNA, the fraction bound without competitor was 28(6)% and the fraction bound with wild-type competitor was 4.8(2.0)%. For experiments with 300-fold excess of competitor to DNA, the fraction bound without competitor was 29(8)% and the fraction bound with wild-type competitor was 1.6(0.8)%. Data represent averages of two to eight experiments, except for ‡39 ! ‡53 competitor at 300-fold excess which is from a single experiment.

with polymerase modi®ed only by the presence of AsiA protein (results not shown). Similar results were obtained with these various transcription conditions. The promoter with the wild-type MotA box yielded 14-fold more RNA with modi®ed polymerase plus MotA than with modi®ed polymerase alone. All of the single base mutations, except ÿ31G ! A, resulted in promoters that were activated eightfold or higher in the presence of MotA and yielded levels of MotA-dependent RNA that were at least 25 % of that produced by the wild-type (Table 2). The ÿ30C ! A mutation resulted in a promoter that was indistinguishable from wild-type (Figure 2, lanes 5 and 6; Table 2), and replacement of the ÿ30 MotA box with the ‡39 to ‡53 region resulted in a promoter that was activated by MotA (Figure 2, lanes 15 and 16; Table 2). Thus, a MotA box with a center motif of GATT works well in both the in vitro binding and transcription assays. We conclude that the protection of region III by MotA (positions ‡40 to ‡57) arises from the presence of a MotA box with the

sequence AAAGATTAA (positions ‡45 to ‡53 downstream of PuvsX; Figure 1). Of the mutants we tested, the only single base mutant promoter that was not substantially activated by MotA was ÿ31G ! A (Figure 2, lanes 3 and 4; Table 2). In the absence of MotA, this promoter yielded a high level of RNA, but addition of MotA increased the level of RNA only about twofold. Thus, the ÿ31G ! A mutation converted PuvsX from a promoter that is primarily dependent on MotA into one that is only marginally dependent. Identification of new T4 middle promoters To investigate sequence diversity among natural middle promoters, we performed primer extension analyses of RNA isolated from T4 motA‡ and T4 motAÿ infections. In some cases, we selected primers to scan regions of the T4 genome which were known to express prereplicative genes, but which lacked nearby middle promoters. In other cases,

935

Sequence Requirements for T4 MotA Activation

Figure 2. In vitro transcription from promoters having a wild-type MotA box or a MotA box with speci®c base changes. Reactions contained both the wild-type MotA box (pMotbox digested with SspI) and mutant MotA box (pMotbox derivative digested with BsaAI) templates and modi®ed RNA polymerase isolated from a T4 infection without (odd lanes) or with (even lanes) MotA. See Table 2 for MotA box sequences and Materials and Methods for reaction conditions. The positions of wild-type and mutant promoter generated transcripts are indicated.

we performed computer searches to locate ÿ10 regions and various MotA box motifs which had not been previously assigned as promoters. We then selected primers to target those sequences for testing. We identi®ed seven new MotA-dependent transcripts (Figure 3). The sequence at the start of each of these transcripts (Figure 4) contains an excellent match to the s70 ÿ10 consensus sequence,

positioned from 5 to 6 bp upstream of the mapped MotA-dependent ends. Thus, we conclude that these RNAs initiate at legitimate transcription start sites downstream from unidenti®ed MotA-dependent promoters. We have designated these new middle promoters as P42, P61, P55, P55.9, P55.8, Ptd, and PnrdA (Figure 4). P61 and P42 each contain a MotA box at ÿ30 that matches the consensus

Table 2. Transcription from promoters with mutant MotA box sequences MotA Boxa wild-type (PuvsX) Mutants ÿ32T ! A ÿ31G ! A ÿ30C ! A ÿ30C ! G ÿ30C ! T ÿ29T ! C ÿ28T ! G ÿ26A ! G ‡39 ! ‡53 Multiple mutant 1

Level of RNA mutant/ wild-typeb

TTTGCTTAA TTAGCTTAA TTTACTTAA TTTGATTAA TTTGGTTAA TTTGTTTAA TTTGCCTAA TTTGCTGAA TTTGCTTAG AAAGATTAA TTATACTAA

Fold activation by MotAc 144

25 % 57 %d 96 % 52 % 40 % 35 % 33 % 39 % 30 % 10 %

8 2 17 15 12 20 14 12 28 e

a MotA box sequences are present within the plasmid pMotbox which contains the PuvsX sequences from ÿ34 to ‡83. b Level of RNA observed in the presence of MotA from the mutant promoter relative to that from the wild-type promoter present in the same reaction mixture. c Calculated as (amount of RNA synthesized by modi®ed polymerase plus MotA)/(amount of RNA synthesized by modi®ed polymerase alone). Wild-type value represents the average of ten experiments. d Only half of this RNA is MotA-dependent. e Level of RNA from mutant promoter too low to calculate ratio accurately.

936 sequence well. However, the other promoters contain changes within the MotA box center motif that had not been seen before. P55, P55.9, P55.8, and Ptd have the sequences GGTC, GCTA, GCAT, and GGTT, respectively, while PnrdA has multiple changes with its ÿ30 region, resulting in no obvious match to the MotA box motif. An excellent match to the s 70 ÿ10 consensus sequence is a necessary, but insufficient requirement for MotA-dependent transcription An examination of the 31 identi®ed T4 middle promoters (Figure 4; Nivinskas et al., 1989; Stitt & Hinton, 1994) reveals that all but one (see Discussion) have the s70 ÿ10 consensus sequence TAnnnT. The sequence of P55 demonstrates the importance of this match in the ÿ10 region. Two other possible MotA boxes occur upstream of the ÿ30 region: a perfect match to the consensus MotA box sequence (shown by the ®lled arrow in Figure 4) and a match to a MotA box with a center motif of GATT (shown by the broken arrow). However, each of these possible MotA boxes lacks a good match to the s70 ÿ10 sequence at the appropriate distance, and in the primer extension assay, only the 50 ends corresponding with P55 were observed. This result suggests that the need

Sequence Requirements for T4 MotA Activation

for an excellent match to the s70 ÿ10 consensus outweighs the need for a perfect match to the MotA box consensus sequence. Given that some MotA-activated transcription occurred in vitro from promoters lacking a recognizable MotA box (Figure 1) and in vivo from PnrdA (Figure 3), we wondered if any perfect ÿ10 sequence (TATAAT) by itself could be used as a MotA-dependent promoter. The overwhelming majority of T4 early and middle genes are oriented in the same direction (counterclockwise) on the T4 genome (Guha & Szybalski, 1968). Therefore, we performed computer searches of the appropriate strand of the T4 genome (Kutter et al., 1994) for the sequence TATAAT. From this analysis, we selected three sequences that had not been identi®ed as early or middle promoters: Pvs, P5R0 , and Pcd (Figure 5). All three candidates have ideal ÿ10 regions. In fact, Pvs has the same sequence as PuvsX from positions ÿ16 to ÿ6. However, all contain multiple mismatches to the MotA box consensus sequence. Primer extension analyses of RNA from T4 motA‡ and T4 motAÿ infections failed to detect any MotA-dependent 50 ends coming from these possible promoters (results not shown). Thus, an excellent s70 ÿ10 sequence by itself is not suf®cient to de®ne a middle promoter.

Figure 3. Primer extension analyses of T4 middle promoters. Primer extension analyses for each indicated promoter were performed as described in Materials and Methods using RNA isolated from suppressing (‡) or non-suppressing (ÿ) E. coli NapIV infected with T4amG1 (amber mutation of motA). Except for P42(‡), analyses are shown in duplicate. In each case, the MotA-dependent 50 end(s) are indicated. GATC lanes show the sequence obtained with wild-type T4 DNA, the same labeled primer, and ddGTP, ddATP, ddTTP, and ddCTP, respectively.

Sequence Requirements for T4 MotA Activation

937

Figure 4. Sequences of T4 middle promoters identi®ed by primer extension analyses. Top, Original Consensus Sequence refers to the inferred consensus sequence from the previously identi®ed set of 24 middle promoters (listed by Stitt & Hinton, 1994, except for P31 which is given by Nivinskas et al., 1989). The size of the base represents its conservation within this data set; letters in bold were invariant. Underneath are shown the sequences of the seven new middle promoters identi®ed in this study and the sequence of P46 (reported by Guild et al., 1988; Hsu & Karam, 1990). Underlined base denotes the RNA 50 end(s) determined by primer extension analyses (see Figure 3). A shadowed base indicates a base not seen in that position in the original set. Filled and broken arrows denote matches to MotA boxes with a center motif of GCTT or GATT, respectively, found at a positions other than ÿ30 in P55.

Although a ÿ30C!A MotA box is activated by MotA in vitro, it is not activated when present in the T4 genome The ÿ30C ! A mutation, which produces a MotA box with a center motif of GATT, resulted in a very active promoter in vitro (Figure 2, lanes 5 and 6; Table 2). Because no such naturally occurring promoter was identi®ed previously or in our screen, we searched the complete T4 sequence for this type of promoter and examined seven possible candidates: Pcd.3, Pe.1, P5R, Panti-late, Ppset.1, P550 , and P40 (Figure 5). Except for the ÿ30C ! A change, all of these potential promoters match the derived MotA box consensus sequence quite well. Furthermore, Pcd.3 has a MotA box sequence that is nearly identical with the ‡39 to ‡53 promoter that was active in vitro (Figure 2, lanes 15 and 16; Table 2). Pcd.3, Panti-late, Pe.1, and P5R also contain excellent ÿ10 regions that match those found in naturally occurring middle promoters. However, primer extension analyses failed to detect 50 ends that would be expected if these promoters were active (results not shown). It is possible that these promoters appeared inactive because their transcripts

were rapidly degraded. However, the fact that no apparent promoter activity was found in seven different cases strongly argues that a MotA box with the center motif of GATT cannot be used for MotA-dependent transcription during T4 infection. MotA-dependent activation of the middle promoter P46 is affected by the presence of modified cytosine residues in the DNA Of all the single base changes that we tested in vitro, only the ÿ31G ! A promoter, which has a center motif of ACTT, was poorly activated by MotA (Figure 2, lanes 3 and 4; Table 2) This result was puzzling because one naturally occurring T4 middle promoter, P46, contains the ACTT motif. Primer extension analysis (Figure 3) con®rmed the earlier designation of P46 as a MotA-dependent promoter in vivo (Guild et al., 1988; Hsu & Karam, 1990). Because there are sequence differences between P46 and the ÿ31G ! A promoter that we used in Figure 2, we tested the in vitro transcriptional activity of two templates containing the P46 sequence. One template contained just the core

938

Sequence Requirements for T4 MotA Activation

Figure 5. Tested promoter sequences that failed to yield detectable MotA-dependent 50 ends in vivo. Designations are the same as described in the legend to Figure 4.

sequences of this promoter (from ÿ34 to ‡1; Figure 6, lanes 1 and 2), while another contained additional sequences both upstream and downstream of the P46 start site (positions ÿ124 to ‡201; Figure 6, lanes 3-4). Like the ÿ31G ! A promoter, these templates also failed to yield substantial MotA-dependent transcription in vitro. However, unlike the ÿ31G ! A template, both of these P46 templates were poor substrates for polymerase in the absence of MotA, indicating that the ability of the ÿ31G ! A promoter to give a high level of MotA-independent transcription was due to other sequence differences. Failure to observe MotA-dependent transcription from P46 in our in vitro system could re¯ect the need for other transcription factors. To test this, we transcribed the natural T4 template using either modi®ed polymerase isolated after T4 infection (Figure 7, lanes 5-8) or polymerase modi®ed only by the addition of AsiA (Figure 7, lanes 2-4) and then assayed P46 transcription by primer extension analysis. Signi®cant MotA activation of transcription was observed in both cases, indicating that transcription from P46 can be accomplished with a minimal system of polymerase, AsiA, and MotA. The amount of RNA observed in lanes 5-8 is much less than that seen in lanes 2-4, because modi®ed

polymerase from a T4 infection has low activity (Hinton et al., 1996b, and references cited therein). Natural T4 DNA is modi®ed by the presence of glucosylated, hydroxymethylated residues at position 5 of the cytosine base (Carlson et al., 1994). To ask whether this modi®cation is important for P46 activation, we transcribed the wild-type T4 template and two different deoxycytosine (dC)-containing T4 DNA. T4 phage that produce (dC) DNA must contain multiple mutations to be viable (Carlson & Miller, 1994). The two dC DNAs used in this experiment were obtained from phages whose mutations should be identical except that T4 JW2095 has a large deletion that removes T4 gene denB, while T4 JW819 has point mutations within denB (see Materials and Methods for details). Using the T4 dC DNA from JW819 and JW2095, we observed a 2.3-fold or 9.5-fold activation of P46 by MotA, respectively, while a 23fold activation was seen with the wild-type template (Table 3). In contrast, MotA activation of PuvsX and PsegA, two middle promoters with GCTT motifs, was similar or greater when using either of the T4 (dC) DNAs than when using the wild-type DNA (Table 3). Taken together, these results suggest that the cytosine modi®cation present in

Sequence Requirements for T4 MotA Activation

Figure 6. In vitro transcription from the T4 middle promoter P46. Multiple round transcription reactions contained both a PuvsX template (pMotbox digested with SspI, lanes 1-2; PuvsX PCR product, lanes 3-4) and a P46 template (pP46 digested with BsaAI, lanes 1-2; P46 PCR product, lanes 3-4), MotA, and either T4-modi®ed polymerase or polymerase modi®ed with AsiA, as indicated. Reaction conditions were as detailed in Materials and Methods, except that lanes 3-4 contained 0.03 pmol of DNA, 0.7 pmol of unmodi®ed polymerase modi®ed with 2 pmol of AsiA and as indicated, 1.5 pmol MotA. The PuvsX and the P46 transcripts (or their expected positions) are identi®ed.

wild-type T4 DNA affects MotA activation of transcription from P46.

Discussion What defines a T4 middle promoter? MotA activation of bacteriophage T4 middle promoters provides a simple system for investigating the actions of a DNA binding activator (MotA), a co-activator (AsiA), and E. coli RNA polymerase. At the start of this work, the list of more than 20 MotA-dependent promoters (Nivinskas et al., 1989; Stitt & Hinton, 1994) had revealed two highly conserved regions: the s70 consensus sequence of TAnnnT centered at ÿ10 and a MotA box consensus sequence of (t/a)(t/a)TGCTT(t/c)A centered at ÿ30 and located from 11 to 13 bp upstream of the ÿ10 region (Brody et al., 1983; Guild et al., 1988). Every promoter in this set contained CTT at the center of this MotA box, and all but one contained a center motif of GCTT. Thus, it seemed likely that strong matches to the consensus sequences in both the ÿ10 and ÿ30 regions were essential for middle promoter activity. However, our analyses of MotA binding and activation using mutant MotA box sequences show that MotA can tolerate changes at

939 the most highly conserved positions of the MotA box sequence, and that the CTT trinucleotide is not an invariant feature of a MotA box. Our in vivo studies con®rm that there are several natural T4 middle promoters whose ÿ30 regions lack a perfect match to the GCTT MotA box motif. If these promoters are analogous to PuvsX, then they do not require sequences upstream of ÿ34. However, we do not yet know if sequences outside of the ÿ30 region are needed for some of these promoters to compensate for a poorer match to the MotA box, or if the modi®ed cytosine residues present in wild-type T4 DNA provide additional speci®city for binding and activation. In contrast with the results with the MotA box sequence, our analyses reinforce the idea that middle promoters must have a strong match to the s70 ÿ10 consensus sequence. Of the 31 T4 middle promoters now identi®ed (Nivinskas et al., 1989; Stitt & Hinton, 1994; this paper), only P43 has an assigned sequence which is not TAnnnT. P43 was assigned the sequence GAGTAT (Guild et al., 1988; Hsu & Karam, 1990) because it lies 12 bp downstream of a MotA box with a GCTT motif. However, a perfect ÿ10 sequence of TATAAT is found just 3 bp downstream of this assigned ÿ10. In the light of our ®ndings, we would assign the ÿ10 region of P43 to the TATAAT sequence and then accept a ÿ30 region that deviates from the MotA box consensus sequence. Cytosine modification of the DNA can affect MotA transcription Although previous studies have shown that T4 modi®cation of its cytosine residues is not required for MotA binding or activation of many promoters with a center motif of GCTT (Hinton, 1991; Schmidt & Kreuzer, 1992; Ouhammouch et al., 1995; Adelman et al., 1997), we found discrepancies with MotA boxes having a center motif of ACTT or GATT when using (dC)-containing DNA versus modi®ed T4 DNA. Two different promoters with center motifs of GATT, the ÿ30C ! A promoter and the ‡39 ! ‡ 53 promoter, were active when present in plasmid DNA, and duplexes with these sequences bound MotA well. However, in vivo screens failed to detect RNA ends initiating from GATT motif sequences. The opposite result was seen with the ACTT motif. The duplex DNA with the ACTT motif bound MotA poorly, and there was little to modest activation from the ACTT motif promoter P46 when present in plasmid DNA, PCR-generated templates or T4 (dC)-containing DNA. However, P46 was very active when present in modi®ed T4 DNA. These results suggest that T4 modi®cation of its cytosine residues affects MotA action at certain MotA box sequences. However, because this phenomenon has not been seen with promoters having the GCTT motif, it seems unlikely that the glucosyl, 5-hydroxymethyl moiety affects MotA binding directly by providing a required contact for MotA. Perhaps cytosine modi-

940

Sequence Requirements for T4 MotA Activation

Figure 7. Primer extension analyses of P46 RNA generated from T4 DNA in vivo and in vitro. Primer extension analyses were performed as described in Materials and Methods using RNA isolated in vivo from suppressing E. coli NapIV infected with T4amG1 (amber mutation of motA) or RNA generated in vitro with wild-type T4 DNA as a template and as indicated, MotA, AsiA, and T4-modi®ed polymerase or unmodi®ed polymerase. Lanes 7 and 8 represent an overexposure of lanes 5 and 6, respectively. The position of the end that represents the P46 transcript is indicated.

®cation in¯uences whether a particular structure can be attained. By this reasoning, the modi®cation would promote the structure in the ACTT motif, but inhibit the structure in the GATT motif. Further studies are needed to determine the contribution of this modi®cation to MotA activation. Does MotA activation represent a novel type of prokaryotic activation? In E. coli, there are many promoters which lack good ÿ35 sequences and require activators (for reviews, see Busby & Ebright, 1994; Hochschild & Dove, 1998). Class I activators bind the DNA at sequences located at ÿ60 or upstream, and simultaneously contact an a subunit. Class II activators bind to the DNA just upstream of the ÿ35 region, making contact with s70 and/or a. In both cases, it has been proposed that the presence of the activator stabilizes the interaction of s70 with sequences in the ÿ35 region that match the s70 ÿ35 consensus poorly. Although T4 middle promoters lack a Table 3. Transcription from wild-type T4 and T4 dCcontaining DNA T4 template

P46

Wild-type (dC) JW819 (dC) JW2095

23 2.3 9.5

Fold activation by MotAa at PuvsX PsegA 19 14 17

38 38 64

a Level of primer extended ends observed in the reaction with MotA, AsiA, and RNA polymerase relative to that observed in the reaction with AsiA and polymerase (see Materials and Methods for details).

s70 ÿ35 consensus sequence and both MotA and AsiA bind directly to s70 (Gerber & Hinton, 1996), MotA-dependent activation of PuvsX does not ®t either of these paradigms. All of the sequences needed for MotA-dependent transcription from PuvsX are between ÿ34 and ‡3 and important elements lie within the center of the ÿ30 MotA box (positions ÿ29 to ÿ32). Thus, the sequences needed for MotA activation reside within the core promoter sequences. Furthermore, footprints of polymerase/AsiA/MotA at PuvsX indicate that there is a major rearrangement of the protein-DNA contacts in the ÿ30 region from what is observed with promoters that use class I, class II, or no activators (Hinton et al., 1996a; M.S., P.M. & D.M.H., unpublished results). Thus, at middle promoters there appears to be a switch to new core promoter sequences rather than a stabilization of the interaction of s70 with a poor match to its ÿ35 consensus. Although MotA and AsiA appear to have no prokaryotic equivalents, their behavior is somewhat similar with that of particular eukaryotic TAFs (TATA binding protein (TBP) associated factors; for a review, see Verrijzer & Tijan, 1996). TAFs bind directly to TBP, a component of each of the three eukaryotic polymerases (pol I , pol II, and pol III). In pol II, TBP is needed to contact the TATA box sequence found within the core of most pol II promoters, just as s70 makes direct contact with the ÿ10 and ÿ35 elements of E. coli promoters. However, TBP is also required for transcription from TATA-less pol II promoters as well as transcription from pol I and pol III promoters which lack TATA boxes. A set of TAFs (human

Sequence Requirements for T4 MotA Activation

TAFIs 110, 63, and 48), which are responsible for the switch in promoter preference from TATA-containing pol II promoters to rRNA pol I promoters (Beckman et al., 1995), share functional similarities with MotA and AsiA. TAFI 48 inhibits TBP recognition of the pol II TATA box sequence (Beckman et al., 1995) while AsiA inhibits the ability of s70 to recognize its ÿ35 consensus sequence (Adelman et al., 1997; Colland et al., 1998). TAFIs 110/63 bind DNA (although weakly), and their binding to TBP directs the TAFI48-TBP complex toward pol I promoter sequences. MotA binding to the DNA and to s70 is necessary for the switch to middle promoter sequences. Thus, the T4 MotA/AsiA model may represent a prokaryotic version of TAFmediated promoter switching that is employed by the eukaryotic polymerases.

Materials and Methods Proteins and reaction buffers MotA, AsiA, and T4 modi®ed polymerase were obtained as described (Hinton et al., 1996b). Unmodi®ed E. coli RNA polymerase was purchased from USB of Amersham Life Science. Kglu transcription buffer (Zou & Richardson, 1991) contains 40 mM Tris-acetate (pH 7.9), 150 mM potassium glutamate, 4 mM magnesium acetate, 0.1 mM EDTA, 0.1 mM DTT, and 100 mg/ml BSA (bovine serum albumin). DNA pDKT90, which contains the PuvsX sequence from ÿ94 to ‡83 inserted into the SmaI site of the vector pTZ19U (Amersham Life Science), and pDKT90
HD and pDKT90
NH, in which the PuvsX sequence in pDKT90 has been deleted from ÿ94 to ÿ38 and ÿ18, respectively, have been described (March-Amegadzie & Hinton, 1995). The plasmid harboring the ÿ35 mutation (p(ÿ35 mutant); Figure 1) has been described (designated as pMS1403 by Hinton et al., 1996b). The plasmids p(ÿ30 mutant), p(ÿ51 mutant), p(ÿ51 inverse), and p(5 bp insert) were constructed by in vitro mutagenesis of pDKT90. Plasmid p(ÿ56 to ‡18) was generated by insertion of a double-stranded oligonucleotide with the PuvsX sequence from ÿ56 to ‡18 into the SmaI site of pTZ19U (Figure 1). Plasmid p(ÿ56 to ‡83) was constructed as follows (Figure 1): a polymerase chain reaction (PCR) with AmpliTaq DNA polymerase (Perkin Elmer), pDKT90 as template, and oligomers that annealed 135 bases upstream and 167 bases downstream of the start of PuvsX, generated a 302 bp fragment. This fragment was digested with both MboII, which cleaves at position ÿ50 within PuvsX, and EcoRI, which cuts within the vector sequence downstream of PuvsX. This fragment was then ligated to pTZ19U, which had been cleaved with XbaI and EcoRI, in the presence of a double-stranded oligomer that recreated the PuvsX sequence to ÿ56 and provided an XbaI site for ligation with the vector. Oligodeoxyribonucleotides were either synthesized using a 381A DNA synthesizer (Applied Biosystems, Inc.) or were purchased from Cruachem, Inc., Midland Certi®ed Reagent Co., or Genosys Biotechnologies, Inc. Oligomers were puri®ed by reverse-phase chromatography. Molar concentrations (per molecule of oligonucleotide) were calculated from A260 measurements.

941 Labeled wild-type MotA box DNA, used for gel retardation assays, was generated by ®rst labeling one of the single strands with T4 polynucleotide kinase and [g-32P]ATP. This strand was then hybridized to the complementary strand by heating both oligomers to 90-95  C for two minutes in a solution containing 20 mM Tris-HCl (pH 7.5), 2 mM MgCl2, 50 mM NaCl, and then slowly cooling to room temperature. Unincorporated nucleotides were removed by chromatography through a G-25 spin column (Pharmacia, LKB). Other non-radioactive double-stranded oligomers were used directly after hybridization. Recombinant plasmids were engineered by introducing the various MotA box duplexes between the BamHI and NdeI sites of pDKT90. Thus, pMotbox contains the wild-type sequence of PuvsX from ÿ34 to ‡84 while the mutant plasmids contain the changes indicated in Table 2. Plasmid p(ÿ34 to ‡3) was constructed by replacing the NdeI to EcoRI fragment of pMotbox, which contains PuvsX sequences from ÿ19 to ‡83, with a doublestranded oligomer that contains PuvsX sequences from ÿ19 to ‡3. Plasmid pP46 was constructed by inserting a double-stranded oligomer containing sequences that extend from ÿ34 to ‡1 of the T4 middle promoter P46 between the BamHI and EcoRI sites of pDKT90, thus replacing the sequences of PuvsX with those of P46. Dideoxy sequencing reactions (Sanger et al., 1977) were performed to con®rm the expected mutations in the pDKT90 and pMotbox derivatives. Transcription templates were generated from these plasmids by restriction with either BsaAI or SspI enzymes which cleave 490 or 698 bp, respectively, downstream of the start site of PuvsX in pDKT90. Transcription templates for PuvsX and P46 were also obtained as PCR products of wild-type T4 DNA, primers, and pfu polymerase (Stratagene). Template DNA, isolated after electrophoresis on agarose gels, were puri®ed by using DNA puri®cation kits from Qiagen or Promega. DNA substrates for DNase I footprinting, labeled on the 50 end of the bottom (template) strand, were obtained as polymerase chain reaction products of pDKT90 or the pDKT90 derivatives and oligonucleotides spaced to yield products of 199 to 264 bp (described by MarchAmegadzie & Hinton, 1995). After electrophoresis on 4 % (w/v) polyacrylamide gels, the desired products were excised, isolated by electroelution, and puri®ed by using the Wizard PCR prep kit (Promega). The T4 strains JW819 and JW2095, which produce (dC) T4 DNA, were gifts of J. Wiberg (Pittsford, New York). JW2095 is T4 56amE51, 42amN55, denAnd28, alc10,
H23 (
H23 is a several kb deletion which extends from within rII into ac, and thus removes denB). JW819 was derived from JW2095 and its phenotype is consistent with a T4 56amE51, 42amN55, denAnd28, alc10, and denBD2a2 phage (see Warner et al., 1970; Vetter & Sadowski, 1974; Kutter et al., 1975; Carlson & Wiberg, 1983; Snyder & Jorissen, 1986, for information about these mutations). Thus, JW819 should differ from JW2095 only by the presence of the denBD2a2 double point mutations in denB rather than the
H23 deletion. Wild-type T4 DNA and (dC) containing DNA from JW819 and JW2095 were isolated and puri®ed as described (Hinton et al., 1985). DNA sequence analyses of wild-type T4, JW819, and JW2095 DNAs indicated that all three have the same sequence surrounding the P46 start site.

942 DNase I footprints DNase I digestions were performed and products separated as described (March-Amegadzie & Hinton, 1995). Densitometer scans of autoradiograms Films were scanned using either a Desktop Plus scanner from Protein Databases, Inc. or an Arcus II scanner from AGFA and quanti®ed using Diversity One software from Protein-Databases, Inc. Competition binding assays Labeled wild-type MotA box duplex (0.04 pmol) was incubated at 4  C for 30 minutes in a reaction (10 ml) containing 1 pmol MotA, Kglu transcription buffer, and 0, 2, or 12 pmol of the short duplexes containing the various mutations within the MotA box sequence. Sucrose (1 ml of a 60 % (w/v) solution) was added and the reaction loaded on a native 12 % (w/v) polyacrylamide (37.5, acrylamide:1,bis-acrylamide). Protein-DNA complexes were resolved from free DNA by electrophoresis at 10 V/cm for three hours at 4  C. Competitive effectiveness was calculated as follows: [(fraction bound without competitor - fraction bound with mutant competitor )/(fraction bound without competitor - fraction bound with wild type as competitor)] X 100. In vitro transcription of plasmid and PCR templates Unless otherwise indicated, reactions (5 ml) containing DNA (0.04 pmol total), 1 pmol MotA, Kglu transcription buffer, 50 mM each of ATP, GTP, and CTP, 2.5 mM [a-32P]UTP, and 0.35 pmol T4-modi®ed RNA polymerase were incubated at 37  C for 7.5 minutes. A solution of cold rNTPs (0.5 ml of 6 mM each NTP) was added, and the reaction incubated for an additional 7.5 minutes at 37  C. The RNA was then isolated and separated by electrophoresis as described (Hinton, 1991). In vitro transcription of T4 DNA Reactions (100 ml) containing 6.3 mg DNA (either wildtype DNA, JW819 (dC) DNA, or JW2095 (dC) DNA), Kglu transcription buffer, 285 mM each ATP, GTP, and CTP, 14 mM UTP, and as indicated, 50 pmol MotA, 50 pmol AsiA, and either 7 pmol T4-modi®ed RNA polymerase or 10 pmol unmodi®ed RNA polymerase were incubated at 37  C for eight minutes. A 10 ml solution of 6 mM each rNTP was added, and the reactions were incubated for an additional eight minutes. RNA was isolated and puri®ed as described for the labeled RNA products, except that the RNA was treated as described (Hinton, 1989), dissolved in 12 ml of 10 mM Tris-Cl (pH 7.9), 1 mM EDTA, and then used for primer extension analyses. Computer search for possible T4 middle promoter sequences The T4 sequence database (detailed by Kutter et al., 1994) contains the complete sequence of the T4 genome. This database was searched for speci®c sequences (the s70 ÿ10 consensus sequence, TATAAT or TAnnnT, and various MotA box motifs) using the Find patterns command of the Wisconsin Package of the Genetics Compu-

Sequence Requirements for T4 MotA Activation ter Group, Version 8, running on the VAX/VMS Cluster, Frederick Biomedical Supercomputing Center, NCIFCRDC, Frederick, Maryland. Isolation of in vivo RNA and primer extension analyses RNA was isolated from NapIV suppressing or nonsuppressing cells (Nelson et al., 1982; denoted by K. Kreuzer, Duke University, Durham, NC) that had been infected for four minutes with T4D‡ or the motA amber mutant T4amG1 (Mattson et al., 1978; denoted by K. Kreuzer) using method II described by Hinton (1989). Primer extension analyses were performed as described (Hinton, 1991) using AMV reverse transcriptase (AMV RT XL from Life Sciences, Inc.), all four dNTPs or all four dNTPs plus ddGTP, and the following (50 -32P)labeled oligonucleotides that annealed approximately 100 bases downstream from the predicted 50 end of the RNA (T4 sequence coordinates (Kutter et al., 1994) are given in parentheses; sequences of primers are available upon request): P42 (26195-26214); P61 (19021-19040); P55 (40116-40137); P55.9 (42712-42732); P55.8 (42712-42732); Ptd (145036-145058); PnrdA (142620-142641); P46 (3491134935); PuvsX and PsegA (23663-23681); Panti-late (108973108994); Pcd.3 (133667-133689); Pe.1 (67841-67859); P5R (79218-79240); Ppset.1 (135320-135335); P55' (40116-40137); P40 (23383-23414); Pvs (61829-61852); P5R' (79218-79240); Pcd (132909-132931). (Except for Panti-late, promoter names are based on the name of the ®rst downstream gene. Panti-late would express RNA starting at position 109100 that is antisense to the late gene 24.) Dideoxy sequencing reactions (Sanger et al., 1977) were performed with wildtype T4 DNA, the labeled primers, and Sequitherm polymerase (Epicentre Technologies) to generate sequence ladders for the primer extension analyses.

Acknowledgments We thank John Wiberg for the gift of T4 strains JW819 and JW2095, Sonja Bockenhauer for constructing the plasmids p(ÿ56 to ‡18) and p(5 bp insert), and Nancy Nossal, Anthony Furano, and Tom Schneider for helpful discussions.

References Adelman, K., Orsini, G., Kolb, A., Graziani, L. & Brody, E. N. (1997). The interaction between the AsiA protein of bacteriophage T4 and the s70 subunit of Escherichia coli RNA polymerase. J. Biol. Chem. 272, 27435-27443. Beckman, H., Chen, J.-L., O'Brien, T. & Tijan, R. (1995). Coactivator and promoter-selective properties of RNA polymerase I TAFs. Science, 270, 1506-1509. Brody, E., Rabussay, D. & Hall, D. H. (1983). Regulation of transcription of prereplicative genes. In Bacteriophage T4 (Mathews, C. K., Kutter, E. M., Mosig, G. & Berget, P. B., eds), pp. 174-183, American Society for Microbiology, Washington, DC. Busby, S. & Ebright, R. (1994). Promoter structure, promoter recognition, and transcription activation in prokaryotes. Cell, 79, 743-746. Carlson, K. & Wiberg, J. (1983). In vivo cleavage of cytosine-containing bacteriophage T4 DNA to geneti-

Sequence Requirements for T4 MotA Activation cally distinct, discretely sized fragments. J. Virol. 48, 18-30. Carlson, K. & Miller, E. S. (1994). Preparing phage with unmodi®ed DNA. In Molecular Biology of Bacteriophage T4 (Karam et al., eds), pp. 433-434, American Society for Microbiology, Washington, DC. Carlson, K., Raleigh, E. A. & Hattman, S. (1994). Restriction and modi®cation. In Molecular Biology of Bacteriophage T4 (Karam et al., eds), pp. 369-381, American Society for Microbiology, Washington, DC. Colland, F., Orsini, G., Brody, E. N., Buc, H. & Kolb, A. (1998). The bacteriophage T4 AsiA protein: a molecular switch for sigma 70-dependent promoters. Mol. Microbiol. 27, 819-829. Dombroski, A. J. (1997). Recognition of the ÿ10 promoter sequence by a partial polypeptide of s70 in vitro. J. Biol. Chem. 272, 3487-3494. Gerber, J. S. & Hinton, D. M. (1996). An N-terminal mutation in the bacteriophage T4 motA gene yields a protein that binds DNA but is defective for activation of transcription. J. Bacteriol. 178, 6133-6139. Guha, A. & Szybalski, W. (1968). Fractionation of the complementary strands of coliphage T4 DNA based on the asymmetric distribution of the poly U and poly U,G binding sites. Virology, 34, 608-616. Guild, N., Gayle, M., Sweeney, T., Hollingsworth, T., Modeer, T. & Gold, L. (1988). Transcriptional activation of bacteriophage T4 middle promoters by the motA protein. J. Mol. Biol. 199, 241-258. Harley, C. B. & Reynolds, R. P. (1987). Analysis of E. coli promoter sequences. Nucl. Acids Res. 15, 2343-2361. Hinton, D. M. (1989). Transcript analyses of the uvsX ÿ 40-41 region of bacteriophage T4. Changes in the RNA as infection proceeds. J. Biol. Chem. 264, 14432-14439. Hinton, D. M. (1991). Transcription from a bacteriophage T4 middle promoter using T4 motA protein and phage-modi®ed RNA polymerase. J. Biol. Chem. 266, 18034-18044. Hinton, D. M., Silver, L. L. & Nossal, N. G. (1985). Bacteriophage T4 DNA replication protein 41. Cloning of the gene and puri®cation of the expressed protein. J. Biol. Chem. 260, 12851-12857. Hinton, D. M., March-Amegadzie, R., Gerber, J. S. & Sharma, M. (1996a). Characterization of pre-transcription complexes made at a bacteriophage T4 middle promoter: involvement of the T4 MotA activator and the T4 AsiA protein, a s70 binding protein, in the formation of the open complex. J. Mol. Biol. 256, 235-248. Hinton, D. M., March-Amegadzie, R., Gerber, J. & Sharma, M. (1996b). The bacteriophage T4 middle transcription system: T4-modi®ed RNA polymerase, AsiA (sigma-70 binding protein), and the transcriptional activator MotA. Methods Enzymol. 274, 43-57. Hochschild, A. & Dove, S. L. (1998). Protein-protein contacts that activate and repress transcription. Cell, 92, 597-600. Hsu, T. & Karam, J. D. (1990). Transcriptional mapping of a DNA replication gene cluster in bacteriophage T4. Sites for initiation, termination and mRNA processing. J. Biol. Chem. 265, 5303-5316. Kutter, E., Beug, A., Sluss, R., Jensen, L. & Bradley, D. (1975). The production of undegraded cytosine-containing DNA by bacteriophage T4 in the absence of dCTPase and endonucleases II and IV, and its effects on T4-directed protein synthesis. J. Mol. Biol. 99, 591-607.

943 Kutter, E., Stidham, T., Guttman, B., Kutter, E., Batts, D., Peterson, S., Djavakhishvili, T., Arisaka, F., Mesyanzhinov, V., RuÈger, W. & Mosig, G. (1994). Genomic map of bacteriophage T4. In Molecular Biology of Bacteriophage T4 (Karam et al., eds), pp. 491-519, American Society for Microbiology, Washington, DC. March-Amegadzie, R. & Hinton, D. M. (1995). The bacteriophage T4 middle promoter PuvsX: analysis of regions important for binding of the T4 transcriptional activator MotA and for activation of transcription. Mol. Microbiol. 15, 649-660. Mattson, T., Richardson, J. & Goodin, D. (1974). Mutant of bacteriophage T4D affecting expression of many early genes. Nature, 250, 48-50. Mattson, T., van Houwe, G. & Epstein, R. H. (1978). Isolation and characterization of conditional lethal mutations in the mot gene of bacteriophage T4. J. Mol. Biol. 126, 551-570. Nelson, M. A., Ericson, M., Gold, L. & Pulitzer, J. F. (1982). The isolation and characterization of TabR bacteria: hosts that restrict bacteriophage T4 rII mutants. Mol. Gen. Genet. 188, 60-68. Nivinskas, R. G., Raudonikiene, A. & Guild, N. (1989). A new early gene upstream of the middle gene 31 in bacteriophage T4: cloning and expression. Mol. Biol. (Moscow), 23, 739-749. Ouhammouch, M., Orsini, G. & Brody, E. N. (1994). The asiA gene product of bacteriophage T4 is required for middle mode RNA synthesis. J. Bacteriol. 176, 3956-3965. Ouhammouch, M., Adelman, K., Harvey, S. R., Orsini, G. & Brody, E. N. (1995). Bacteriophage T4 MotA and AsiA proteins suf®ce to direct Escherichia coli RNA polymerase to initiate transcription at T4 middle promoters. Proc. Natl Acad. Sci. USA, 92, 1415-1419. Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA, 74, 5463-5467. Schmidt, R. P. & Kreuzer, K. N. (1992). Puri®ed motA protein binds the ÿ30 region of a bacteriophage T4 middle-mode promoter and activates transcription in vitro. J. Biol. Chem. 267, 11399-11407. Siebenlist, U., Simpson, R. B. & Gilbert, W. (1980). E. coli RNA polymerase interacts homologously with two different promoters. Cell, 20, 269-281. Siegele, D. A., Hu, J. C., Walter, W. A. & Gross, C. A. (1989). Altered promoter recognition by mutant forms of the s70 subunit of Escherichia coli RNA polymerase. J. Mol. Biol. 206, 591-603. Snyder, L. & Jorissen, L. (1986). Molecular proof that bacteriophage T4 alc and unf genes are the same gene. J. Bacteriol. 168, 833-838. Stevens, A. (1972). New small polypeptides associated with DNA-dependent RNA polymerase of Escherichia coli after infection with bacteriophage T4. Proc. Natl Acad. Sci. USA, 69, 603-607. Stevens, A. (1976). A salt-promoted inhibitor of RNA polymerase isolated form T4 phage-infected E. coli. In RNA Polymerase (Losick, & Chamberlin, , eds), pp. 617-627, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Stevens, A. & Rhoton, J. C. (1975). Characterization of an inhibitor causing potassium chloride sensitivity of an RNA polymerase from T4 phage-infected Escherichia coli. Biochemistry, 14, 5074-5079. Stitt, B. & Hinton, D. (1994). Regulation of middle-mode transcription. In Molecular Biology of Bacteriophage T4

944

Sequence Requirements for T4 MotA Activation

(Karam et al., eds), pp. 142-160, American Society for Microbiology, Washington, DC. Verrijzer, C. P. & Tijan, R. (1996). TAFs mediate transcriptional activation and promoter selectivity. Trends Biochem. Sci. 21, 338-342. Vetter, D. & Sadowski, P. D. (1974). Point mutants in the D2a region of bacteriophage T4 fail to induce T4 endonuclease IV. J. Virol. 14, 207-213. Waldburger, C., Gardella, T., Wong, R. & Susskind, M. M. (1990). Changes in conserved region 2 of

Escherichia coli s70 affecting promoter recognition. J. Mol. Biol. 215, 267-276. Warner, H. R., Snustad, D. P., Jorgensen, S. E. & Koerner, J. F. (1970). Isolation of bacteriophage T4 mutants defective in the ability to degrade host deoxyribonucleic acid. J. Virol. 5, 700-708. Zou, L. & Richardson, J. P. (1991). Enhancement of transcription termination factor rho activity with potassium glutamate. J. Biol. Chem. 266, 10201-10209.

Edited by M. Yaniv (Received 2 June 1998; received in revised form 23 October 1998; accepted 28 October 1998)