* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download DNA Specificity of the Bicoid Activator Protein Is Determined by
Biochemistry wikipedia , lookup
Western blot wikipedia , lookup
Real-time polymerase chain reaction wikipedia , lookup
Paracrine signalling wikipedia , lookup
Metalloprotein wikipedia , lookup
Molecular cloning wikipedia , lookup
Biosynthesis wikipedia , lookup
Zinc finger nuclease wikipedia , lookup
Transformation (genetics) wikipedia , lookup
DNA supercoil wikipedia , lookup
Nucleic acid analogue wikipedia , lookup
Protein–protein interaction wikipedia , lookup
Deoxyribozyme wikipedia , lookup
Non-coding DNA wikipedia , lookup
Community fingerprinting wikipedia , lookup
Signal transduction wikipedia , lookup
Gene regulatory network wikipedia , lookup
Proteolysis wikipedia , lookup
Expression vector wikipedia , lookup
Promoter (genetics) wikipedia , lookup
Point mutation wikipedia , lookup
Gene expression wikipedia , lookup
Vectors in gene therapy wikipedia , lookup
Endogenous retrovirus wikipedia , lookup
Transcriptional regulation wikipedia , lookup
Silencer (genetics) wikipedia , lookup
Cell, Vol. 57, 1275-1283, June 30, 1989, Copyright 0 1989 by Cell Press DNA Specificity of the Bicoid Activator Protein Is Determined by Homeodomain Recognition Helix Residue 9 Steven D. Hanes and Roger Brent Department of Molecular Biology Massachusetts General Hospital Boston, Massachusetts 02114 and Department of Genetics Harvard Medical School Boston, Massachusetts 02115 Summary Formation of anterior structures in the Drosophila embryo requires the product of the gene bicoid. The bi. coid protein contains a homeodomain and may exert its effects in early development by regulating transcription of the gap gene, hunchback (hb). Consistent with this view, we have demonstrated that DNA-bound Bicoid fusion proteins stimulate gene expression. We used the gene activation phenotype in yeast to study DNA recognition by the Bicoid homeodomain. We found that a single amino acid replacement at position 9 of the recognition helix was sufficient to switch the DNA specificity of the Bicoid protein. The altered specificity Bicoid mutants recognized DNA sites bound by Ulfrabithorax, fushi taram, and other related homeodomain proteins. Our results suggest that DNA specificity in Bicoid and Anfennapedia class proteins is determined by recognition helix residue 9. Introduction Pattern formation in Drosophila embryos appears to be initiated by a small number of maternal-effect genes. One of these genes, bicoid, is necessary for establishing anterior-posterior polarity (Ntisslein-Volhard et al., 1987). Mutants that lack functional bicoid product fail to develop head and thoracic structures. bicoid mRNA is synthesized in nurse cells during oogenesis and transported into the developing oocyte where it is localized in the anterior (Berleth et al., 1988). After egg deposition, bicoid mRNA is translated and the protein (here called Bicoid) accumulates in the anterior of the egg (Driever and NtissleinVolhard, 1988a). Bicoid is present in embryonic nuclei where it may stimulate zygotic expression of hunchback (Driever and Niisslein-Volhard, 1988b, 1989), and other gap genes whose expression is necessary to establish pre-segmental domains in the immature egg (reviewed in Ingham, 1988). Like many other proteins important for development, Bicoid contains a conserved 60 amino acid sequence known as the homeodomain (reviewed in Gehring, 1987). Sequence similarity of the homeodomain to the helix-turnhelix DNA binding motif of many bacterial proteins (Pabo and Sauer, 1984), suggests that the homeodomain may also bind DNA (Laughon and Scott, 1984). Indeed, many Drosophila homeodomain proteins bind to specific DNA sequences in vitro (Desplan et al., 1985; Hoey and Levine, 1988; Beachy et al., 1988; Driever and Ntisslein-Volhard, 1989); in some cases, DNA binding has been shown to require the homeodomain (Desplan et al., 1988; Miller et al., 1988b; Hoey et al., 1988). Similarly, some mammalian transcription factors also contain homeodomain sequences (Scheidereit et al., 1988; Mtiller et al., 1988a) that are involved in DNA binding (Ko et al., 1988; Sturm and Herr, 1988). Many bacteriophage repressor proteins use the second a-helix of the helix-turn-helix motif (the recognition helix) to contact DNA. Sequence specificity is determined primarily by hydrogen bonding of side chains of residues 1 and 2 of the recognition helix to functional groups of base pairs in the major groove (Lewis et al., 1983; Wharton and Ptashne, 1985; Hochschild and Ptashne, 1986; Wharton and Ptashne, 1987; Jordan and Pabo, 1988). Additional contributions to specificity are made by polar interactions or hydrogen bonds by residues 5 and 6 (as in h repressor, Hochschild et al., 1986; or as in 434 repressor, Wharton and Ptashne, 1987; Aggarwal et al., 1988). Studies of h and 434 repressor-DNA cocrystals reveal hydrogen bonds between the amino-terminal residues of the two helices in the motif that may help position the recognition helix correctly in the major groove (Aggarwal et al., 1988; Jordan and Pabo, 1988). Several nonspecific contacts with the sugar-phosphate backbone, including one made by the amino acid at recognition helix position 9, may contribute to the stability of the repressor-operator complex. The amino acid sequence of the Bicoid homeodomain differs from that in proteins of the Anfennapedia (An@) sequence class (Scott et al., 1989). Several of these differences lie in positions which, by analogy with the bacterial helix-turn-helix proteins, might be important for DNA recognition. Amino acids 1 and 2 of the putative Bicoid recognition helix (homeodomain helix 3) differ from those found at equivalent positions in Antp class proteins, and might be predicted to make site-specific contact with DNA. Other residues that are important in bacteriophage repressors and that differ between Bicoid and Antp class homeodomain proteins include recognition helix residue 9, which could make nonspecific DNA contacts, and a residue near the amino terminus of the first helix of the motif (Bicoid residue 127). The position of residue 127 suggests that it might interact with the first residue in the recognition helix, and thus help position the recognition helix for favorable contact with DNA. Bicoid binds to variants of a 9 bp consensus sequence, TCTAATCCC, that are found upstream of the transcription start site of the zygotically expressed form of hunchback (Driever and Niisslein-Volhard, 1989). This hunchback element (or Bicoid binding site) is distinct from sites bound by other Drosophila homeodomain proteins. These sites, TCAATTAAAT (Desplan et al., 1988) and (TAA)” (Beachy et al., 1988) are bound by proteins encoded by engrailed, even-skipped, and the Antp class genes Ultrabithorax, Abdominal-A, and fushi tarazu (Desplan et al., 1988; Hoey and Levine, 1988; Beachyetal., 1988; Karsch and Bender, Cdl 1276 personal communication; Samson et al., 1989). These sites are not recognized by the Bicoid protein (Samson et al., 1989; this paper). It seemed possible that Bicoid and other homeodomain proteins distinguish between these sites using residues in the recognition helix which differ, and are in positions corresponding to those important for repressor-DNA recognition. In this study, we employed fusion proteins that contain the DNA binding domains of the bacterial LexA repressor or the yeast GAL4 activator (Brent and Ptashne, 1985; Ma and Ptashne, 1987a). Whenever examined, fusion proteins that carry a DNA binding domain from the native protein in addtion to the LexA or GAL4 domain stimulate expression of target genes that carry binding sites for the native protein (Brent and Ptashne, 1985; Hope and Struhl, 1988; Fitzpatrick and Ingles, 1989; Samson et al., 1989). We took advantage of this fact to examine gene regulation and DNA binding by Bicoid. Here we show that DNA-bound LexA-Bicoid and GAL4Bicoid fusion proteins stimulate gene expression in yeast. Bicoid fusion proteins activated target genes that carried upstream LexA operators, GAL4 sites, or Bicoid binding sites. We used the gene activation phenotype to establish which residues in the Bicoid homeodomain are required for DNA recognition. Unlike bacterial helix-turn-helix proteins, amino acids in positions 1,2, or 5 of the putative recognition helix do not appear to be important. Instead, “helix swap” experiments showed that Bicoid’s DNA specificity is determined by recognition helix amino acid 9. Our results also suggest that for Antp class and other closely related homeoproteins, DNA specificity inheres in residue 9. Results DNA-Bound Bicoid Activates Gene Expression We tested whether Bicoid regulates gene expression by producing, in yeast, Bicoid fusion proteins that contained the DNA binding domains of either LexA or GALA Hybrid genes were constructed by fusing a bicoid cDNA (Berleth et al., 1988) to sequences encoding the amino-terminal 87 residues of LexA or the amino-terminal 147 residues of GAL4 (Figure 1). The fusion proteins were expressed from an ADHl promoter on 2 urn-based plasmids carrying the yeast HIS3 gene. All but the first two amino acids of Bicoid were included in the fusions. To test whether DNA-bound Bicoid activates gene expression, yeast cells were cotransformed with “maker” plasmids that expressed one of the fusion proteins, and “target” plasmids that contained a 2 urn replicator, the yeast MA3 gene, and a GALl-IacZ reporter gene that carried an appropriate binding site positioned upstream of the transcription start site (Figure 2). Changes in GAL&IacZ expression were measured by assaying 8-galactosidase activity of cells grown in liquid culture. Bicoid fusion proteins stimulated gene expression in yeast (Tables la and lb). Target genes lacking upstream binding sites were not stimulated by Bicoid fusion proteins, while those carrying a LexA operator were stimulated about 50-fold by LexA-Bicoid, and those carrying a GAL4 binding site were stimulated about 20-fold by GALC a ADHI promoter LexA (I- / . /. f Blcold ima cDNA I TAG BamHl (Hhal) . . . EcoR’ I 1 Sail b GAL4Bicoid Figure [I 1. Expression 147 of Hybrid 3 bicoid 489 Proteins in Yeast (A) Plasmid pSHP-I encodes the amino-terminal 87 amino acids of LexA under control of the alcohol dehydrogenase I (ADHl) promoter. Fusions to a bicoidcDNA were made at the BamHl cloning site. Other unique sites are EcoRl and Sall. The pSHP-1 vector backbone and polylinker sequence were derived from pMA424 (Ma and Ptashne, 1987b). The GAU-bicoidfusion construct was made at the BamHl site of pMA424. The DNA sequences and encoded amino acids of the LexA-polylinker-bicoid fusion junction, and the GAL4-polylinker-bicoid fusion junction are given in Experimental Procedures. (B) Schematic of the LexA-Bicoid and GAU-Bicoid hybrid proteins. The shaded areas show the approximate location of the homeodomain. Numbers refer to the amino acid residues in the native proteins. Bicoid. We interpret these increases in activity of the reporter gene to reflect increases in transcription due to DNA-bound proteins (Guarente and Ptashne, 1981; Yocum et al., 1984). The strong transcription activator LexA-GAL4 stimulated gene expression about 15-fold better than LexA-Bicoid, and native GAL4 stimulated gene expression about 8-fold better than GAL6Bicoid. Bicoid Activates Genes Containing Upstream hunchbacksequences To test whether the Bicoid moiety of the LexA- and GALC fusion proteins recognized specific DNA sites in vivo to activate gene expression, we used sequences found upstream of hunchback that are protected by purified Bicoid in a DNAase I footprint assay (Driever and NiissleinVolhard, 1989). Multiple copies of this TCTAATCCC consensus Bicoid binding site were inserted upstream of GALWacZ (Figure 2). Target genes bearing only two copies of the site were not activated by Bicoid fusion proteins rJJ;;pected DNA Contact I / I by Bicoid Recognition TATA , Helix GAL I LacZ Table 1. Gene Maker C-167) Activation Plasmids Target a LexA Op GTACTGTATGTACATACAGTAC ALexA <l <l <l LexA(l-a7) LexA-Bicoid LexA-GAL4 b AGAL GAL+-atv) TCTAATCCC(TA) Bicoid site Antp class site TCAATTAAATCGA) Figure 2. Binding Sites Used in This Study Autonomously replicating target plasmids contained the indicated sequences inserted at a unique Xhol site upstream of reporter gene GALMacZ. Bases in parentheses were included in these sites, but are not an integral part of the core sequences. The LexA operator and GAL4 binding site were present in single copy, the hunchback (hb) element was present in 10 copies, and the Antp class binding site was present in 12 copies. The target plasmids were derivatives of pLR1 Al (West et al., 1984) except for the GAL4 site target plasmid, which was a derivative of pLR1 A20 (West et al., 1984). The Xhol sites in pLR1 At and pLRlA20 are positioned at 167 bp and 128 bp upstream of the GALMacZ transcription start site, respectively. (data not shown), whereas target genes bearing ten copies of the Bicoid site were activated more than 500-fold by LexA-Bicoid and more than 300-fold by GALCBicoid (Tables la and lb). Bicoid target genes were not activated by the amino termini of LexA or GAL4 alone. Since both fusion proteins activated a Bicoid target, we presumed that Bicoid site recognition depended on some region common to both proteins, possibly the Bicoid homeodomain. Helix 3 in the Bicoid Homeodomain Is Required for DNA Recognition Amino acid substitutions were made within the putative recognition helix of the Bicoid homeodomain at positions we thought were likely to make base-specific contacts with DNA. To destroy DNA recognition, these residues were changed to alanines, small residues that do not hydrogen bond to base pairs in DNA, and have been shown not to disrupt the recognition helices of bacteriophage repressors (Hochschild et al., 1986; Wharton and Ptashne, 1987; Hochschild and Ptashne, 1986). By analogy with repressor proteins, the Bicoid recognition helix has been predicted to encompass, minimally, residues 138-147 in the native protein. Here, these ten amino acids will be referred to as positions 1 through 10. To confirm that the mutant LexA-Bicoid proteins were produced in yeast, we tested their ability to stimulate gene expression of LexA operator-containing targets (Table lc). Only one mutant protein did not satisfy this criterion (mutant A,As&A,); we assume this protein was unstable or defective in its activation function. As expected, none of the mutant proteins activated expression of target genes that did not contain upstream binding sites. The mutant proteins were then assayed for activation of target genes that carried Bicoid binding sites (Table lc). C LexA-Bicoid LexA-Bicoid LexA-Bicoid LexA-Bicoid A, As As A1A5A.& Op Site ALexA <I <l <l <l Bicoid Fusion Proteins Plasmids <l <l <l GAL+-147) GALcBicoid GAL4 site CGGAAGACTCTCCTCCG by DNA-Bound Op LexA Op <l 64 844 GAL4 Site 7 19 158 LexA Op 65 38 94 4 Bicoid Site <l 564 Bicoid Site 1 365 1 Bicoid Site 585 187 <l <l Results are given in units of 8-galactosidase activity. Yeast cell cultures cotransformed with “maker” and “target” plasmids were grown to a density of 0.8-2.0 x IO’ cells per ml in glucose-supplemented complete minimal medium lacking uracil and histidine. LexA-Bicoid de rivatives were tested in yeast strain MGLD4-4a; GALCBicoid derivatives were tested in a gal- strain, RBY52 (Agal4) in glucosesupplemented medium. 8-galactosidase assays were performed on at least three individual transformants of a given plasmid combination (activities varied up to 20%). Assays performed on different days were normalized to values obtained for LexA-GAL4 with a LexA operator target plasmid. LexAo..sr) includes the first 87 amino acids of LexA plus additional residues encoded by the pSHP-1 polylinker (RPEFPGIRRPAAKLIPAEFLMIYDFYY). GAL4rr-143 includes the first 147 amino acids of GAL4 plus additional residues encoded by the pMA424 polylinker (PEFPGIRRPAAKLIPAEFLMIYDFYY). Target plasmids were all derivatives of pLRlA1 except for the GAL4 site target, which was a derivative of pLRlA20. LexA-Bicoid mutant A, contained a Thr-Ala substitution at position 1 and still activated a Bicoid site target. LexABicoid mutant A5 contained a Lys+Ala substitution at position 5 and also activated a Bicoid site target but at reduced levels. However, activation by this mutant was also reduced using a LexA site target, perhaps because the protein is less stable or impaired in its activation function. Finally, a LexA-Bicoid mutant that contained a Lys-Ala substitution at position 9 did not activate a Bicoid site target. This mutant still activated LexA site targets. These experiments indicate that position 9 of the Bicoid recognition helix is critical for recognition of DNA. The data do not, however, distinguish between specific and nonspecific interactions with DNA. Altered Specificity Bicoid Mutants To identify those residues in Bicoid required for DNA specificity, we carried out “helix swap” experiments similar to those described by Wharton and Ptashne (1985). In these experiments, we changed the predicted solvent exposed residues in the Bicoid recognition helix to those found in the recognition helix of Anfennapedia class proteins (Figure 3). The mutant proteins were tested for their ability to distinguish between two sites known to bind different classes of homeodomain proteins in vitro. As before, our assay measured expression of target genes bearing these upstream sites. Bicoid target genes contained 10 copies of the hunchback element (TCTAATCCC), and Antp class Cell 1270 Activation UQ I23456789 IO Antp class site Wild-type Bicoid + + - AAQVKIWFKN Al + + - TAQVAIWFKN AS + + - TAQVAIWFAN As + - RHZ-E I RzQ9 + - + E 1R2Qg + - + + + - + - + + - + Al AsA6Aq EBQVKlWFgN ~IEBQVKI 3. Schematic Bicoid sil? TAQVKIWFKN AAQVAAWFAN Figure of Gene Expression EIRZ WFKNO TAQVKIWFQN Qs ERQ I KIWFQN Antp Showing Amino Acid Residues class* in the Putative Recognition Helix (helix 3) of Wild-Type LexA-Bicoid and Mutant Derivatives The corresponding residues of Uhbithorax are shown for comparison, and are representive of the consensus sequence for Antp class homeoproteins. The barrels identify the putative helix-turn-helix motif of these proteins. The residues that differ from wild-type LexA-Bicoid sequence are indicated. Activation data are summarized from Tables 1 and 2 and our unpublished work. (‘) Results derived from Samson et al. (1989) and Fitzpatrick and lngles (1989). target genes contained 12 copies of an Antp class binding site (TCAATTAAATGA). Results are given in Table 2. In the yeast strain used here, wild-type LexA-Bicoid and the various mutant proteins stimulated expression of LexA target genes to comparable levels, although these levels were uniformly lower than those obtained previously (see legend to Table 2). As expected, the mutant proteins did not stimulate target genes lacking upstream binding sites. Wild-type LexABicoid activated Bicoid site targets but did not activate Anto class target8 above background levels. LexA-Bicoid Table Maker 2. Gene Activation Plasmids Lex$87) LexA-Bicoid LexA-Bicoid LexA-Bicoid LexA-Bicoid LexA-Bicoid LexA-Bicoid b-‘S R&g - El R&g - E,RP - - - Qs - - - A9 by Mutant Bicoid mutant ErRa, which contains a Thr+Glu substitution at position 1 and an Ala+Arg substitution at position 2, retained its ability to stimulate expression of target genes containing Bicoid sites and did not stimulate target genes containing Art@ class sites. However, LexA-Bicoid mutant Rk2 -ElR2Q9 which contained, in addition to the changes at positions 1 and 2, a Leu+Arg substitution at position 3 in helix 2 and a Lys+Gln substitution at position 9, lost the ability to stimulate target genes carrying Bicoid sites but gained the ability to stimulate target genes carrying An@ class sites. A LexA-Bicoid mutant bearing only three Derivatives Target Plasmids ALexA <l <l Op LexA Op <1 6 Bicoid Site <l 530 Antp Class 35 26 Site ALexA <l < <l <l <l Op LexA Op 7 11 5 8 6 Bicoid Site <I <l 750 <l <l Antp Class 610 970 25 650 35 Site Results are given in units of 8-galactosidase activity. Experimental conditions are described in the legend to Table 1. Target plasmids (see Figure 2) were all pLR1 Al derivatives and produced no detectable background activity, except for the Antp class binding site target plasmid which regularly produced about 30 U of background 6-galactosidase activity. In these experiments yeast strain YM709 was used throughout because of very high background levels of 8-galactosidase activity produced in strain MGLD4-4a when transformed with the Antp site target plasmid. Unexpected DNA Contact by Bicoid Recognition Helix 1279 of these changes, Elf+&, exhibited a similar phenotype. This mutant also stimulated expression of target genes that carried the Anfp type l(TAA), site (data not shown) recognized by Ulfrabifhorax protein and other Antp class proteins (Beachy et al., 1988; Samson et al., 1989). The results above implicate the importance of residue 9 in DNA specificity. This was tested directly by constructing LexA-Bicoid mutant Q,, which bears a single amino acid substitution of Lys*Gln at position 9 in the recognition helix. This mutant did not stimulate expression of Bicoid site targets but strongly stimulated expression of targets carrying Antp class sites. For each mutant bearing this Lys-Gln substitution at postion 9, levels of activation of Bicoid targets were decreased more than 500-fold, while levels of activation of Antp class targets were increased more than 20-fold (Table 2). If one subtracts background levels obtained with the Antp class target (~30 units), then postition 9 mutants activated Antp class targets more than 800-fold relative to wild-type LexA-Bicoid. Discussion Bicoid Is a DNA Binding-Dependent Gene Activator Bicoid is a gene activator. LexA-Bicoid and GAL4-Bicoid activated expression of target genes that contained LexA binding sites or GAL4 binding sites, and did not stimulate gene expression from target genes that lacked binding sites. The fusion proteins also activated expression of targets genes carrying Bicoid binding sites. Since, in yeast, LexA sites and GAL4 sites are only bound by proteins that contain the cognate binding domains, we conclude that stimulation of gene expression by these Blcoid fusion proteins requires their binding to DNA (Brent and Ptashne, 1985; Giniger et al., 1985; Keegan et al., 1988). These results support the notion that activation of targets with Bicoid sites in our yeast system and in Drosophila cells (Driever and Nijsslein-Volhard, 1989) is a consequence of binding by Bicoid to these sites. We do not know how Bicoid turns on transcription. Bicoid contains an acidic region (residues 345-390, net charge -9), a glutamine-rich region (or M-repeat, McGinnis et. al., 1984), and a region that contains alternating histidines and prolines (or PRD repeat, Frigerio et al., 1988). Acidic regions, glutamine-rich regions, and proline-rich regions have all been shown to function as transcription activation domains (Hope and Struhl, 1988; Ma and Ptashne, 1987a, 1987b; Courey and Tjian, 1988; Williams et al., 1988). We observed activation by a GAL4Bicoid derivative that contained only Bicoid residues 3-251. This protein lacked the acidic region and the polyglutamine tract but retained the PRD repeat. It is possible, therefore, that transcription activation by native Bicoid is due to the PRD repeat. Alternatively, it is possible that gene activation by this Bicoid derivative is due to phosphate groups. Bicoid is likely to be differentially phosphorylated during development (Driever and NiissleinVolhard, 1989); it is possible that the phosphate groups may subsitute for acidic amino acids by providing equiva- lent negative surface charge and Pelham, 1988). (Lech et al., 1988; Sorger Oligomerization of Bicoid The DNA bound form of LexA-containing proteins is a dimer (Brent and Ptashne, 1981, Brent, 1982). As previously noted for other fusion proteins, the amino terminus of LexA does not dimerize efficiently, and the fact that LexA-Bicoid recognizes LexA operators suggests that the Bicoid moiety contributes to dimerization (Brent and Ptashne, 1985; Hope and Struhl, 1988; Lech et al., 1988; Samson et al., 1989). Whether dimerization occurs on or off the DNA, the results with LexA-Bicoid thus suggest that in Drosophila embryos, native Bicoid binds DNA as an oligomer. Our data suggest Bicoid oligomers may exhibit another form of cooperativity. In yeast, Bicoid strongly stimulated targets that carried ten hunchback elements, but did not detectably stimulate targets that carried two hunchback elements. This result, which is similar to results observed with a number of homeodomain proteins in transient expression experiments (Desplan et al., 1988; Jaynes and O’Farrell, 1988; Driever and Niisslein-Volhard, 1989), suggests that either binding to the hunchback elements or gene activation by the DNA-bound proteins is highly cooperative. LexA-Bicoid dimers activate a single LexA site target more strongly than they activate a target with two Bicoid sites, which suggests that the cooperativity facilitates binding to the Bicoid sites rather than gene activation once the protein is bound to DNA. Clusters of Bicoid binding sites occur upstream of hunchback (Berleth et al., 1988); cooperative binding of Bicoid oligomers to these sites provides a mechanism by which the continuous Bicoid gradient might be converted into the threecell-wide boundary observed for hunchback expression (Tautz, 1988). Unexpected DNA Contact by the Bicoid Recognition Helix While nothing so simple as a code has emerged, a large body of work in recent years has suggested that interaction of helix-turn-helix proteins with DNA is governed by some common rules (for review, see Pabo and Sauer, 1984; Ptashne, 1988; Schleif, 1988). Protein monomers typically associate into dimers, in which each symmetrically disposed monomeric subunit interacts with symmetric or near symmetric operator half-sites that lie in adjacent major grooves. In the case of lambdoid phage repressors and cro proteins, and the E. coli CAP protein, the second a-helix of the helix-turn-helix motif (the recognition helix) protrudes into the major groove of DNA, and the amino-terminal end of the helix makes base-specific contacts. Sequence alignments (Laughon and Scott, 1984; Scott et al., 1989) and NMR studies (Otting et al., 1988) have suggested that Drosophila homeodomain proteins contain a helix-turn-helix motif. If the Bicoid homeodomain interacts with DNA, as do bacteriophage repressors, we reasoned that base-specific contacts would be made by amino acids 1 and 2 of the recognition helix, with possible contributions from residues at positions 5, 6, and 9. Cell 1280 To test this idea, we made a series of mutant proteins in which residues in the putative Bicoid recognition helix were replaced by alanines. The LexA-Bicoid mutants were examined for their ability to stimulate expression of target genes that carried upstream Bicoid binding sites. These experiments benefitted from the fact that the Bicoid derivatives were bifunctional for DNA recognition. By comparing activation of target genes that carried LexA sites or Bicoid sites with those of “wild-type” LexA-Bicoid, we could ascertain whether a given mutation affected Bicoid DNA specificity, or affected some other property of the protein (such as nuclear transport or protein stability). Results indicated that positions 1, 2, and 5 were not important for recognition of Bicoid sites. However, a mutant that borealanine at position 9 no longer activated Bicoid targets. These experiments suggested that residue 9 in the recognition helix is important for either specific or nonspecific interaction with DNA. To probe the function of individual amino acids in sitespecific DNA recognition, we performed “helix swap” experiments (Wharton et al., 1984; Wharton and Ptashne, 1985). First, we established that Bicoid does not recognize An@ class sites and that proteins of the Antp class genes, Ultrabirhorax and Abdominal-A, do not recognize Bicoid sites (this paper; Samson et al., 1989). Based on these facts, we attempted to change the DNA specificity of Bicoid by replacing amino acids in its recognition helix with those found at equivalent positions in the Antp class proteins. A series of LexA-Bicoid derivatives that contained changes at recognition helix residues 1, 2 and 9, and Bicoid residue 127 (in helix 2) were tested for recognition of Bicoid and Antp class sites. Proteins that contained changes at positions 1 and 2 still recognized Bicoid sites, whereas all proteins with changes that included a LysGln substitution at position 9 no longer recognized Bicoid sites, but recognized Antp class sites. Most dramatically, a mutant protein in which the only change was a Lys-+Gln substitution at residue 9 exhibited a complete switch of specificity; it did not recognize Bicoid sites but did recognize Antp class sites. These results show that Lys at position 9 of the recognition helix is critical for DNA specificity by Bicoid. A single Lys+Gln change, in the context of a Bicoid recognition helix, confers Antp class DNA recognition on the Bicoid protein. This implies that Gln at position 9 is also critical for DNA specificity by Antennapedia class proteins and that Bicoid and Antp class proteins contact DNA in an analogous manner. The simplest interpretation is that, unlike the bacteriophage repressor proteins, amino acid 9 of Bicoid and Antp class recognition helices make basespecific contacts with DNA, while residues at the the amino terminus of the recognition helix do not. This model suggests that the recognition helix of these homeodomain proteins is oriented differently in the major groove than the recognition helix of bacteriophage repressor proteins. It is not difficult to imagine how changing a single lysine in Bicoid’s position 9 would result in a dramatic change in specificity. Most Drosophila homeodomain proteins, including Antp class and related proteins encoded by the genes engrailed, even-skipped, and zerkniillt, contain a glutamine at this position (Scott et al., 1989) and recognize the type 1 (TAA)” and type 2 (TCAATTAAAT) sites used here (Beachy et al., 1988; Beachy, personal communication; Karsch and Bender, personal communication; Desplan et al., 1988; Hoey and Levine, 1988; Samson et al., 1989). It is possible that the side-chain amide of the Gln donates a hydrogen bond to N7 of an Adenine in the type 1 and type 2 sites, and accepts a hydrogen bond from N8 of the same base (Seeman et al., 1978). Bicoid has a Lys at position 9; and by analogy with models proposed for h cro protein, it is possible that the amino group of the Lys donates hydrogen bonds to the 06 and N7 of a Guanine in ?he Bicoid site (Ohlendorf et al., 1982). In this view, each LexA-Bicoid monomer that carries a Gln-+Lys substitution at position 9 loses two hydrogen bonds to the Bicoid site and gains two hydrogen bonds to the Antp class sites, a difference in energy sufficient to account for the altered DNA specificity of LexA-Bicoid with Gln at position 9. Based on the identity of position 9, there are at least three other classes of homeodomain proteins: proteins such as paired product, which has a Ser at this position, proteins such as cut product, which has His at this position, and POU class proteins, which have Cys at this position (Frigerio et al., 1986; Blochlinger et al., 1988; Finney et al., 1988; Herr et al., 1988). If these proteins contact DNA like the proteins studied here, then all members of a class should bind similar sequences, members of different classes should recognize different sequences, and in each case some of the DNA specificity should derive from contacts between residue 9 and the binding site. Experimental Procedures Strains, Media, and p-Galactosidase Assays Escherichia coli strains were JMIOI (supE thi A(/ac-proAB) [PtraDSB pra4B lacP ZAM75]), BW313 (HfrKL16 PO/45 [/ysA(61-62)] dutl ungl tbil m/Al), CSH50 (A(/ac-pro) thi ara s&I/f (proA Mac/q ZAM75 traD36), BMH 71-18 (A[/ac-proAs] thi supE F’/ac/q ZAM75 proAB), BMH 71-18 mutS (A[/ac-proM] thi supE F’lacilqAM75 proAB mutS275:;TnlO), MK30-3 (A[/ac-pfa46] mcd ga/E strA Wacis ZAM75 proAB). Saccharomyces cerevisiae strains used were MGLD4-4a (a ura3-52 leu2 his3 trp7 lys2 cyrr), RBY52 (a Dgal4 ura3-52 leu2 hisd), and YM709 (a ura3-52 his3-200 ade2-707 /ys2-807 trp7-907 tyrl-507 met canR ga/4-542 ga/80-538). Bacteriological work was done using standard techniques (Miller, 1972; Ausubel et al., 1987). Yeast media preparation and cell growth were carried out as described in Sherman et al. (1978). Yeast transformations were done using the LiOAc procedure (Ito et al., 1983). fi-galactosidase assays of liquid cultures were carried out as described by Harshman (1988) using at least three individual yeast transformants. Cells were grown in complete minimal medium (lacking uracil and histidine) to mid-logarithmic stage and one ml was collected by centrifugation. The cell pellet was resuspended in 200 ~1 of 0.1 M I-is-HCI (pH 75), 0.05% Triton X-160 and frozen at -800 C. After thawing slowly on ice, B-galactosidase assays were carried out on permeablized cells, and cell debris was then removed by centrifugation. Units are expressed as (lOOO)(&~)/(minutes of reaction)(cell volume)(OD& and varied up to about 20% between individual isolates. Plasmid Construction DNA manipulations were carried out by standard methods (Maniatis et al., 1982; Ausubel et al., 1967). Oligonucleotides were synthesized by the phosphate triester method using a Biosearch Series 8ooO DNA Synthesizer and purified using Applied Biosystems oligonucleotide purification cartridges. A GAL4-fusion plasmid, pMA424 (Ma and Ptashne, 1987b) was ob- Unexpected 1281 DNA Contact by Bicoid Recognition Helix tained from Jun Ma. A LexA-fusion plasmid, pSHP-1, was constructed by insertion of an EcoRI-Sacl(*3 kb) fragment of pMA424 containing a polylinker, ADHl terminator, and yeast 2 pm replicator into the EcoRI-Sac1 backbone of pMA457 (Ma and Ptashne, 1987b). The ADHl promoter-LexA fragment in pMA457 was originally derived from pRBlO27 (Brent and Ptashne, 1985). The second BamHl site upstream of the ADHl promoter in this plasmid (pSHI-1) was then destroyed creating pSHP-I. Plasmid pSH2-1 is essentially identical to pMA424 except that the DNA binding region of GAL4 (residues l-147) has been replaced by the DNA binding region of LexA (residues i-87). Both plasmids contain the HIS3 gene for selection in yeast. Unique sites in these plasmids for fusions to LexA or GAL4 are EcoRI, BamHI, and Sail. The sequence of the polylinker from EcoRl to Sall is WTTCCCGGGGATCCGTCGAC. The reading frame in the polylinker in each of the plasmids (pSH2-1; pMA424) is the same, and is indicated by the underline. Fusion of a nearly full length bicoid cDNA (Berleth et al., 1988) to LexAo-rr,) and GAL4r1-t4r) was accomplished in several steps. First, the cDNA (in EcoRl site of a Bluescript vector) was digested with Hhal, made flush with T4 DNA polymerase, and then digested with Sall. The 743 bp fragment was then ligated to pMA424 that had been previously digested with BamHI, Klenow treated, and Sall digested. The resulting plasmid was then digested with Sall, and the remainder of the bicoid cDNA sequence (including more than 1 kb of 3’ flanking sequences) was inserted as a Sall-Sal1 (ml.8 kb) fragment. The final plasmid, pSHlO-1, encodes an in-frame fusion of GAL4 ,,-,,r, to residue 3 (Gln) of the bicoid coding sequence separated by an 18 bp polylinker sequence. A BamHI-Sac1 fragment containing bicoid and about 2.5 kb of adjacent vector sequences was removed and inserted into the BamHI-Sac1 backbone of pSHP-1. The resulting plasmid, pSHll-1. contains an in-frame fusion of LexA(I-sr, to residue 3 of Bicoid via a 21 bp linker. The top strand sequence across the junction, and the corresponding amino acid sequence of the LexABicoid fusion is as follows: CCT(LexAsr)-CGACCGGAATTCCCGGGGATC-CAA(Bicoid& P-RPEFPGI-Q. The top strand sequence across the junction, and corresponding amino acid sequence of the GAL4Bicoid fusion is as follows: TCG(GAL4&-CCGGAATTCCCGGGGATC-CAA(Bicoid& S-PEFPGI-G. The fusion junctions were confirmed by DNA sequence analysis. All target plasmids (except pSV15) were derivatives of LRI Al (West et al., 1985) and contain the yeast URA3 gene, 2 urn replicator, and a GALI-/acZfusion gene downstream of a GAL7 promoter from which the UASo had been removed. Plasmid ~I840 (Brent and Ptashne, 1985) contains a single 22 bp LexA operator (Brent and Ptashne, 1984) in the Xhol site at -167 bp from the GALI-/acZ transcription start site. pSV15 is an LRl A20 derivative (West et al., 1984) and contains a 17 bp GAL4 consensus binding site (Giniger et al., 1985) in the Xhol site at approximately -128 bp from the GAD/acZ transcription start site. LRl A20 is otherwise identical to LRl Al except that the Xhol site is positioned 39 bp closer to the transcription start site. Target plasmids containing multiple copies of the hunchback consensus sequence TCTAATCCC (Driever and Niisslein-Volhard, 1989) were constructed by ligation of complementary 28 bp oligonucleotides of sequence TCGAATCTAATCCCTATCTAATCCCT and TCGAAGGGATTAGATAGGGATTAGAT into the Xhol site (-167) of LRl Al. Plasmid pHB9 contains five 26 bp hunchback oligonucleotide inserts (ten consensus binding sites). Target plasmids containing multiple copies of the Anfp class binding site consensus sequenceTCAATTAAAlGA(Desplan et al., 1988) were constructed by ligation of complementary 54 bp oligonucleotides of sequence TCGAGTCAATTAAATGATCAATTAAATGATCAATTAAATGATCAAT TAAATGAC and TCGAGTCATTTAATTGATCATTTAATTGATCATTTAATTGATCATTTAATTGAC into the Xhol site (-167) of LRll. Plasmid pFtz-7 contains three 54 bp Antp sequence class binding site oligonucleotide inserts (12 consensus binding sites). Mutagenesis of Bicoid Helix 3 Oligonucleotidedirected mutagenesis was carried out by the methods of Kunkel et al. (1987) or Kramer et al. (1984). For creation of the A, and EIRP mutants (see Figure 3). the bicoid-containing EcoRl fragment of pSHll-1 was removed and inserted into the EcoRl site of phagmid pUC119 (Vieira and Messing, 1987). DNA was transformed into strain E. coli strain BW313 (dufl ungl) and single-stranded DNA prepared (after superinfection with phage M13K07) from recombinants containing bicoid inserts in the desired orientation. Plus strand muta- genie oligonucleotides were treated with T4 polynucleotide kinase, annealed in excess to bicoid template DNAs (minus strands), and the mixtures extended with T4 polymerase and ligated according to the method of Kunkel et al. (1987) as described in Ausubel et al. (1987). The DNA was then transformed into CSH50 cells (dut+ ung+) for selection of mutant plasmids. Initial screening was for a newly created restriction site (88111) introduced by the mutant oligonucleotides. The sequences of the oligonucleotides used for creating the A, and E,Rz mutations were GCCCTGGGCGCAGCCCAGGTGAAGATCTGGTTTAAGAACCG and CTAGCCCTGGGCGAACGCCAGGTGAAGATCTGGTTTAAGAACCG, respectively. The bases changed from wild-type are underlined. For creation of mutants As, As, AtA&&,, Rh2-E,R.&, E,RzQs, and 09 (Figure 3) the method of Kramer et al. (1984) was employed using commercially available reagents (Boehringer Mannheim). bicoidcontaining EcoRl fragments from the LexA-Bicoid expression plasmids were subcloned into phage M13mp9 am 16 and propagated in strain BMH 71-18 (supE). Single-stranded DNA of minus strand polarity (with respect to bicoid sequences) was annealed to linearized M13mp9 rev 86 and to excess mutagenic oligonucleotide. The mixture was extended with Klenow, ligated, and transformed into BMH 7l-18 mut.9. After passaging of phage through BMH 71-18 mut.9, recombinant phage were selected on a nonsuppressing strain, MK30-3. The desired phage were identified by screening of RF DNAs for new restriction sites created by silent mutations introduced by mutagenesis. Oligonucleotides used to create As, Ae, A,A&&,, R,,z -E,R&, El R2Q9, and Q are listed in order with the altered bases underlined (newly created restriction sites are indicated): ACAGCCCAGGTCGCGATATGGTTTAAG (Nrul) GATATGGTTTGCGAACCGTCGACGTCGTCAC (Sall) CAGCCCAGGTGGCGGCCTGGTTBGAACCGTCGACGTCGTCA CAAG (Nrul, Sall) CTCACAGCCCCACGACGTGCGGATCTG (Dralll) ATCTGGTTTCAGACCCGTCGACGTCGTCAC (Bglll. Sall) ATATGGTTTCAGAACCGTCGACGTCGTCAC (Sall) Note that the “template” bicoid single-stranded DNAs used for mutagenesis with these six oligonucleotides were wild-type, wild-type, A,, E1R2Q9, E,Rz, and wild-type, respectively. Once obtained, the desired bicoid derivatives were recloned as EcoRl fragments into pSH2-1. All base substitutions introduced by these procedures were confirmed by DNA sequence analysis. Acknowledgments We are grateful to Jen Sheen for her expertise on mutagenesis procedures, Marie-Laure Samson for plasmids and communication of unpublished data, and Joanne Kamens and David Eisenmann for plasmid vectors and helpful advice. We thank Aneel Aggarwal, Thomas Burglin, David Eisenmann, Ann Ephrussi, Mike Finney, Erica Golemis, Gary Ruvkun, and Karen Lech for helpful discussions and/or comments on the manuscript. This work was supported by Hoechst AG and an award to R. B. from the Pew Scholars Program. S. D. H. is a fellow of the Anna Fuller Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Received February 24, 1989; revised May 23, 1989. References Aggarwal. A. K., Rodgers, D. W.. Drotter, M., Ptashne, M.. and Harrison, S. C. (1988). Recognition of DNA operator by the repressor of phage 434: a view at high resolution. Science 242, 899-907. Ausubel, F. M., Brent, R.. Kingston, R., Moore, D., Seidman, J., Smith, J., and Struhl, K. (1987). Current Protocols in Molecular Biology (New York: John Wiley 8 Sons). Beachy, f? A., Krasnow, M. A., Gavis, L. R., and Hogness, D. S. (1988). An Ultrabithorax protein binds sequences near its own and the Antennapedia Pl promoters. Cell 55, 1069-1081. Cell 1282 Berleth, T, Burri, M., Thoma, G., Bopp, D., Richstein, S., Frigerio, G., Nell, M., and Nusslein-Volhard, C. (1988). The role of localization of bicoid RNA in organizing the anterior pattern of the Drosophila embryo. EMBO J. 7, 1749-1758. Hochschild, A., and Ptashne, M. (1986). Homologous interactions repressor and I cro with the 1 operator. Cell 44, 925-933. Blochlinger, K., Bodmer, Ft., Jack, J., Jan, L. Y., and Jan, Y. N. (1988). Primary structure and expression of a produt from cut, a locus involved in specifying sensory organ identitiy in Drosophila. Nature 333, 629-635. Hoey, T, and Levine, M. (1988). Divergent homeobox proteins recognize similar DNA sequences in Drosophila. Nature 332, 858-861. Brent, Ft. (1982). Regulation chimie 64, 565-569. and autoregulation by IexA protein. Brent, R., and Ptashne, M. (1981). Mechanism of action gene product. Proc. Natl. Acad. Sci. USA 78, 4204-4208. Bio- of the /exA Brent, R., and Ptashne, M. (1984). A bacterial repressor protein yeast transcriptional terminator can block upstream activation yeast gene. Nature 372, 812-815. Brent, R.. and Ptashne, M. (1985). A eukaryotic transcriptional tor bearing the DNA specificity of a prokaryotic repressor. 729-736. or a of a activaCell 40, Courey, A. J., and Tjian, R. (1988). Analysis of Spl in vivo reveals multiple transcriptional domains, including a novel glutamine-rich activation motif. Cell 55, 887-898. Desplan, C., Theis, J., and O’Farrell, P H. (1985). velopmental gene, engrailed, encodes a sequence ing activity. Nature 378, 630-635. The Drosophila despecific DNA bind- Desplan, C., Theis, J.. and O’Farrell, f? H. (1988). The sequence ficity of homeodomain-DNA interaction. Cell 54, 1081-1090. Driever, W., and Nusslein-Volhard, C. (1988a). tein in Drosophila embryos. Cell 54, 83-93. Driever, W., and Nusslein-Volhard, mines position in the Drosophila manner. Cell 54, 95-104. A gradient speci- of bicoidpro- C. (1988b). The bicoidprotein deterembryo in a concentration-dependent Hochschild, A., Douhan. sor and A cro distinguish of 1 J., III, and Ptashne, M (1986). How 1 represbetween On1 and 0~3. Cell 47, 807-816. Hoey, T., Warrior, R.. Manak, J., and Levine, M. (1988). DNA-binding activities of the Drosophila melanogaster even-skipped protein are mediated by its homeo domain and influenced by protein context. Mol. Cell. Biol. 8, 4598-4607. Hope, I. A., and Struhl, K. (1986). Functional dissection of a eukaryotic transcriptional activator, GCN4 of yeast. Cell 46, 885-894. Ingham, P W. (1988). The molecular genetics mation in Drosophila. Nature 335, 25-34. of embryonic pattern for- Ito, H., Fukada, Y., Murata, K., and Kimura, A. (1983). Transformation of intact yeast cells treated with alkali cations. J. Bacterial. 53, 163-188. Jaynes, J. B., and C’Farrell, P H. (1988). Activation transcription by homeodomain-containing proteins mon site. Nature 336, 744-749. and repression of that bind a com- Johnson, A. D., and Herskowitz, I. (1985). A repressor (MATo product) and its operator control expression of a set of cell type specific genes in yeast. Cell 42, 237-247. {ordan, S. R., and Pabo, C. 0. (1988). Structure A resolution: details of the repressor-operator 242, 893-899. of the A complex at 2.5 interactions. Science Keegan, L., Gill, G., and Ptashne, M. (1986). Separation of DNA binding from the transcription activation function of a eukaryotic regulatory protein. Science 231, 699-704. Ko, H.-S., Fast, P, McBride W., Staudt, L. M. (1988). A human specific for the immunoglobulin octamer DNA motif contains tional homeobox domain. Cell 55, 135-144. protein a func- Driever, W.. and Nusslein-Volhard, C. (1989). The bicoid protein is a positive regulator of hunchback transcription in the Drosophila embryo. Nature 337, 138-143. Kramer, W., Drutsa, V., Jansen, H.-J., Kramer, B., Pflugfelder, M., and Fritz, H.-J. (1984). The gapped duplex DNA approach to oligonucleotide-directed mutation construction. Nucl. Acids Res. 72, 9441-9456. Finney, M., Ruvkun, G., and Horvitz, H. R. (1988). The C. elegans cell lineage and differentiation gene encodes a protein with a homeodomain and extended similarity to transcription factors. Cell 55, 757-769. Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1987). Rapid and efficient site-specific mutagenesis without phenotypic selection. Meth. Enzymol. 154, 367-382. Fitzpatrick V. D., and Ingles. C. J. (1989). The Drosophila fushi farazu polypeptide is a DNA-binding transcriptional activator in yeast cells. Nature 337, 666-668. Laughon, A., and Scott, M. P (1984). Sequence of a Drosophila mentation gene: protein structure homology with DNA binding teins. Nature 370, 25-31. Frigerio, G., Burri, M., Bopp, D., Baumgartner, S., and Nell, M. (1986). Structure of the segmentation gene paired and the Drosophila PRD gene set as part of a gene network. Cell 47; 735-746. Lech, K., Anderson, activate transcription Gaul, U., and Jackie, H. (1987). Pole region-dependent repression of the Drosophila gap gene Kriippel by maternal gene products. Cell 57, 549-555. Gehring, Science W., J. (1987). Homeo 236, 1245-1252. boxes in the study of development. Giniger, E., Varnum, S. M., and Ptashne, M. (1985). Specific DNA bindmg of GAL4, a positive regulatory protein of yeast Cell 40, 767-774. Godowski, f? J.. Picard, D., and Yamamoto, K. (1988) Signal transduction and transcriptional regulation by glucocorticoid receptor-LexA fusion proteins. Science 241, 812-816. Guarente, L., and Ptashne, M. (1981). Fusion to the cytochrome c gene of Saccharomyces Acad. Sci. USA 78, 2199-2203. of Escherichia co/i /acZ cerevisiae. Proc. Natl. Hall, B., and Johnson, A. D. (1987). Homeo domain of the yeast a2 IS a sequence specific DNA-binding domarn but IS not sufficient for repression. Science 237; 1007-1012. Harshman, K. D., Scott Moye-Rowley, W., and Parker, C. S. (1988). Transcription activation by the SV40 AP-1 recognition element in yeast is mediated by a factor similar to AP-1 that is distinct from GCN4. Cell 53, 321-330. Herr, W., Sturm, R. A., Clerc, R. G., Corcoran, L. M., Baltimore, D., Sharp, P A., Ingraham, H. A., Rosenfield, M. G., Finney, M., Ruvkun. G., and Horvitz, H. R. (1988). The POU domain: a large conserved region in the mammalian pit-l, act-I, act-2, and Caenorhabditis elegans uric-86 gene products. Genes Dev. 2, 1513-1516. K., and Brent, R. (1988). DNA-bound in yeast. Cell 52 179-184. segpro- Fos proteins Lehming, N., Satorius, J.. Niemiller, G. G., Wilcken-Bergmann, B. V., and Muller-Hill, B. (1987). The interaction of the recognition helix of lac repressor with lac operator. EMBO J. 6, 3145-3153. Lewis, M., Jeffrey, A., Wang, J., Ladner, R., Ptashne, M., and Pabo, C. (1983). Structure of the operator-binding domain of bacteriophage i repressor: implications for DNA recognition and gene regulation. Cold Spring Harbor Symp. Quant. Biol. 47; 435-440. Lillie. J. W., and Green, adenovirus Ela protein. M. R. (1989). Transcription Nature 338, 39-44. activation by the Ma, J., and Ptashne, M. (1987a). Deletion analysis of GAL4 defines transcriptional activating segments. Cell 48, 847-853. Ma, J., and Ptashne, M. (1987b). A new class of yeast transcription vators. Cell 51, 113-119. two acti- Maniatis, T., Fritsch, E. F., and Sambrook, J. (1982). Molecular Cloning: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). McGinnis, W., Levine, M. S., Hafen, E., Kuroiwa, A., and Gehring, W. J. (1984). A conserved DNA sequence in homeotic genes of Drosophils antennapedia and bithorax complexes. Nature 308, 428-433. Mihara, H., and Kaiser, E. T (1988). A chemically synthesized napedia homeo domain binds to a specific DNA sequence. 242, 925-927. Miller, J. (1972). Experiments bor, New York: Cold Spring Muller, M. M., Ruppert, in Molecular Genetics Harbor Laboratory). S., Schaffner, AntenScience (Cold Spring W., and Matthias, Har- I? (1988a). A r+r@;pected DNA Contact cloned octamer lymphoid-specific transcription promoters by Bicoid factor in non-B Recognition Helix stimulates transcription from cells. Nature 336, 644-551. Wharton, R. P, and Ptashne, M. (1985). Changing the binding specificity of a repressor by redesigning an a-helix. Nature 316, 601-605. Miiller, M., Affolter, M., Leupin, W., Otting, G., Wuthrich, K.. and Gehring, W. J. (1988b). Isolation and sequence-specific DNA binding of the Antennapedia homeodomain. EMBO J. 7, 4299- 4304. Wharton, R. P.. and Ptashne, M. (1987). A new-specificity mutant of 434 repressor that defines an amino acid-base pair contact. Nature 326, 888-891. Nusslein-Volhard, C., Frohnhofer, l-t. G., and Lehmann, termination of anteroposterior polarity in Drosophila. 1675-1681. R. (1987). DeScience 238, Wharton, R. P, Brown, E. L., and Ptashne, M. (1984). Substituting an a-helix switches the sequence-specific DNA interactions of a repressor. Cell 38, 361-369. Ohlendorf, D. H., Anderson, W. F., Takeda, Y., Fischer, R. G., Takeda, Y.. and Matthews, B. W. (1982) The molecular basis of DNA-protein recognition inferred from the structure of cro repressor. Nature 298, 718-723. Williams. T., Admon, A., Luscher, B., and Tijan, R. (1988). Cloning and expression of AP-2, a cell-type-specific transcription factor that activates inducible enhancer elements. Genes Dev, 2, 1557-1569. Otting, G., Qian, Y.-Q., Muller, M., Affolter, M., Gehring, W., and Wuthrich, K. (1968). Secondary stucture determination for the Antennapedia homeodomain by nuclear magnetic resonance and evidence for a helix-turn-helix motif. EMBO J. 7, 4305-4309. Pabo, C., and Sauer, R. T. (1984). Protein-DNA Biochem. 53, 293-321. Pabo, C., Krobatin, W., Jeffery, A., and N-terminal arms of 5 repressor wrap around 298, 441-443. recognrtion. Annu. Rev. Sauer, R. T. (1982). The the operator DNA. Nature Ptashne, M. (1986). A Genetic Switch. Gene Control Alto, CA, and Cambridge, MA: Blackwell Scientific and Phage 5 (Palo and Cell Press). Regulski, M., McGinnis, N., Chadwick, R., and McGinnis, W. (1987). Developmental and molecular analysis of Deformed; a homeotic gene controlling Drosophila head development. EMBO J. 6, 767-777. Samson, M.-L., Jackson-Grusby, L., and Brent, R. (1989). Gene activation and DNA binding by Drosophila Ubx and abd-A proteins. Cell 57, 1045-1052. Scheidereit, C., Cromlish, J. A., Gerster, T., Kawakami, K., Balmaceda, C.-G., Currie, A., and Roeder, R. G. (1988). A human lymphoid-specific transcription factor that activates immunoglobulin genes is a homoeobox protein. Nature 336, 551-557. Schleif. R. (1968). DNA binding by proteins. Science 241, 1182-1187. Scott, M. P, Tamkun, J. W., and Hartzell, G. W., Ill (1989). ture and function of the homeodomain. BBA Rev. Cancer, The strucin press. Sherman, F., Fink, G. R., and Lawrence, C. W., (1978). Methods in Yeast Genetics (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). Seeman, N. C., Rosenberg, J. M., and Rich, A. (1978). Sequencespecific recognition of double helical nucleic acids by proteins. Proc. Natl. Acad. Sci. USA 73, 804-806. Sorger, P. K., and Pelham, H. R. B. (1988). Yeast heat shock factor an essential DNA binding protein that exhibits temperature-dependent phosphorylation. Cell 54, 855-864. is Sorger, P K., Lewis, M. J., and Pelham, H. R. B. (1967). Heat shock factor is regulated differentially in yeast and HeLa cells. Nature 329, 81-84. Sturm, R. A., and Herr, W. (1988). The POU domain binding structure. Nature 336, 601-604. is a bipartite DNA- Sturm, R., Das, G., and Herr, W. (1986). The ubiquitous octamer binding protein Ott-1 contains a pou domain with a homeo subdomain. Genes Dev. 2, 1582-1599. Tautz, D. (1988). Regulation of Drosophila segmentation back by two maternal morphogenetic centers. Nature gene hunch332, 281-264. Tautz, D., Lehmann, R., Schnurch, H., Schuh, R., Seifert, A. K., Jones, K., and Jackie, H. (1987). Finger protein of novel structure encoded by hunchback, a second member of the gap class of Drosophila segmentation genes. Nature 327; 383-389. Thali, M., Miiller, M. M., Delorenzi, M., Matthias, P., and Bienz, M. (1968). Drosophila homoeotic genes encode transcriptional activators similar to mammalian CTF-2. Nature 336, 598-600. Vieira, J., and Messing, J. (1987). Production mid DNA. Meth. Enzymol. 753, 3-11. West, R. W., Yocum, R. R., and Ptashne, cerevisae GAL&GAL10 divergent promoter tion of the upstream activating sequence 2467-2478. of single-stranded plas- M. (1984). Saccharomyces region: location and funcUASo. Mol. Cell. Biol. 4, Yocum, R. R., Hanley, S.. West, R., Jr., and Ptashne, M. (1984). Use of /acZfusions to delimit regulatory elementsof the inducible divergent GAL7-GAL70 promoter in Saccharomyces cerevisiae. Mol. Cell. Biol. 4, 1985-1998.