Dimeric lac repressors exhibit phase-dependent co-operativity1

Dimeric lac repressors exhibit phase-dependent co-operativity1

Article No. mb992253 J. Mol. Biol. (1998) 284, 851±857 COMMUNICATION Dimeric Lac Repressors Exhibit Phasedependent Co-operativity Johannes MuÈller,...

246KB Sizes 0 Downloads 26 Views

Article No. mb992253

J. Mol. Biol. (1998) 284, 851±857

COMMUNICATION

Dimeric Lac Repressors Exhibit Phasedependent Co-operativity Johannes MuÈller, Andrew Barker, Stefan Oehler and Benno MuÈller-Hill* Institut fuÈr Genetik der UniversitaÈt zu KoÈln Weyertal 121, 50931 KoÈln Germany

Transcription of the lac operon in Escherichia coli is repressed by the binding of Lac repressor (LacR) to lac operator O1, a pseudo-palindromic sequence centred 11 bp downstream of the transcription start. Full repression of the wild-type promoter by wild-type, tetrameric LacR requires the presence of at least two operator sequences that must not only be in close proximity to O1, 401 bp and 92 bp for the auxiliary operators O2 and O3, respectively, but must also be present on the same side of the DNA helix. LacR mutants lacking the C-terminal heptad repeat and thus only capable of dimer formation still repress, but at a much reduced level. Their repression of the lac promoter is comparable to repression by tetrameric LacR when both auxiliary operators are destroyed. We have examined the residual repression, by dimeric LacR, of a series of constructs containing a CAP-independent promoter and two lac operators, O1 and Oid, separated by a series of spacers increasing in size by single base-pair increments. Surprisingly, repression of these constructs still exhibits phase dependence. The periodicity of maxima is similar to the helical repeat of DNA in vivo, as measured by phase-dependent repression with tetrameric LacR, although the magnitude of repression is much smaller than that obtained in previous experiments with tetrameric LacR. Two additional variants of dimeric LacR with altered C termini that were tested also show phase dependence. Control experiments show that the presence of O1 is required for repression in this system. In the absence of O1, occupancy of the auxiliary operator does not lead to repression. The magnitudes of repression maxima correlate best with the overall basic nature of the C terminus. Weak, unspeci®c contacts by this region with DNA seem suf®cient to explain the observed periodicity. It remains to be seen whether additional factors are also involved in this residual repression. # 1998 Academic Press

*Corresponding author

Keywords: protein-DNA recognition; DNA looping; DNA helical repeat; Lac repressor; Gal repressor

Interactions between proteins bound to physically distinct sites on DNA, with loop formation by the intervening DNA strand, is a recurring theme in the regulation of transcription. Well characterPresent addresses: J. MuÈller, Max-Ernst-Gymnasium, Rodderweg 66, 50321 BruÈhl, Germany; S. Oehler, Department of Genetics, University of Cambridge, Downing Street, Cambridge, CB2 3EH, UK. Abbreviation used: IPTG, isopropyl-b-Dthiogalactopyranoside. E-mail address of the corresponding author: [email protected] 0022±2836/98/490851±07 $30.00/0

ised examples in prokaryotes include tetramer formation by dimers of the bacteriophage l CI repressor bound to distinct operator sites (Grif®th et al., 1986; Hochschild & Ptashne, 1986), loop formation by dimers of the Escherichia coli transcriptional activator protein AraC (Dunn et al., 1984; Lee & Schleif, 1989), and loop formation mediated by the tetrameric, E. coli Lac repressor (LacR) bound to two distinct sites (Mossing & Record, 1986; KraÈmer et al., 1987, 1988). In all cases, this loop formation serves to increase the activity of the protein at a primary site, by increasing the local concentration of the protein at that # 1998 Academic Press

852

Phase-dependent Co-operativity with Dimeric Lac Repressors

site (Bellomy & Record, 1990; Oehler & MuÈllerHill, 1994; MuÈller-Hill, 1998). Loop formation also depends on operator phase: the two bound sites must lie on the same face of DNA for a favourable interaction to occur (Dunn et al., 1984; Grif®th et al., 1986; Hochschild & Ptashne, 1986; KraÈmer et al., 1987, 1988; Bellomy et al., 1988; Lee & Schleif, 1989; Law et al., 1993; MuÈller et al., 1996). LacR is one of 14 related transcription regulators in E. coli (Blattner et al., 1997) that belong to the LacI-GalR family of proteins (MuÈller-Hill, 1983; Weickert & Adhya, 1992). At least one other member of this family, GalR, also exhibits phase-dependent co-operativity of binding to two operators, at least in vitro (Choy et al., 1995). However, as almost all other members of the family and unlike LacR, GalR is only capable of forming dimers, even at protein concentrations much higher than those encountered under physiological conditions (Majumdar et al., 1987). The interaction between DNA-bound dimers is too weak to be detected experimentally, and the co-operativity observed is dependent on the presence of an additional protein, the histone-like protein HU, in the loop complex (Aki et al., 1996). LacR mutants only capable of forming dimers have been isolated (Lehming et al., 1988; Oehler et al., 1990, 1994; Alberti et al., 1991; Brenowitz et al., 1991; Chakerian & Matthews, 1991; Kolkhof et al., 1995). The effect of such mutations on repression of the wild-type lac promoter is the same as if both auxiliary operators, essential for co-operative binding of a LacR tetramer, are destroyed (Oehler et al., 1990). Previously, we examined phase and distancedependence of co-operativity for wild-type, tetrameric LacR with a series of chromosomally located reporter constructs, containing a CAP-independent promoter, in which the distance between the primary operator, O1, and an upstream auxiliary ideal operator (Oid) was varied by single base-pair increments (MuÈller et al., 1996). Here, we examine repression in the same system by LacR variants that are only capable of forming dimers. Tetramer formation in LacR is mediated by a C-terminal mini-leucine zipper (Figure 1(a)), that forms a fourhelical bundle (Lehming et al., 1988; Alberti et al., 1991, 1993; Friedman et al., 1995; Lewis et al., 1996). Alteration of this sequence by frame-shift mutation (Lehming et al., 1988; Oehler et al., 1990), by replacement of any single leucine residue in the zipper with alanine (Alberti et al., 1991), or by substitution with heterologous sequence (Kolkhof et al., 1995), or removal by non-sense mutation (Oehler et al., 1994), leads to mutants fully active in DNA-binding, but defective in co-operative repression when auxiliary operators are present. The amino acid sequences of the C termini of variants examined here are shown in Figure 1. We ®rst examined the effect on repression of the nonsense mutant LacR331Stop (Figure 1(b)). As is apparent in Figure 2, a degree of co-operativity between two operators is still detectable with this variant, despite its complete inability to form tetra-

Figure 1. Amino acid sequence of the extreme C termini of the LacR variants used in this study: (a) wildtype, tetrameric LacR; (b) LacR331Stop; (c) LacR331Gal; and (d) LacRBasic. Leucine and valine residues that form the mini-leucine zipper responsible for tetramer formation in wild-type LacR are shadowed. The KRK motif referred to in the text is shadowed in LacR331Stop. Note also that residue 330 is leucine, not serine, in wildtype LacR. The substitution L330S is neutral in tetrameric LacR and was introduced for convenience in DNA manipulation (Oehler et al., 1994). Residues are numbered according to their position in wild-type LacR. Wild-type, tetrameric LacR was expressed from pSO1010-P1 (Oehler et al., 1990). LacR331Stop was expressed from pSO331Stop (Oehler et al., 1994), while LacR331Gal and LacRBasic were expressed from pSO331Gal and pSOBasic, respectively (Kolkhof et al., 1995).

mers (see below). An almost tenfold enhancement of repression is observed at an inter-operator spacing of 59.5 bp over that obtained in the absence of an auxiliary operator. This repression disappears and reappears as the operator spacing increases, with a periodicity of 11 bp between maxima. Maxima of decreasing magnitude are also observed at inter-operator spacings of 70.5 bp, 81.5 bp, 92.5 bp, and 115.5 bp. Note that these last two distances correspond to the spacing between O1 and O3 in the lac operon (Gilbert et al., 1976), and between OE and OI in the gal operon (Irani et al., 1983; Fritz et al., 1983), respectively. An identical effect is observed with a second LacR variant in which residues downstream of position 330 are replaced with the C-terminal 13 amino acid residues of GalR (Figures 1(c) and 2). Although these maxima correspond to those observed previously with tetrameric LacR, there are important differences. First, maximal repression with tetrameric LacR, an approximately 50fold enhancement over that seen in the absence of any auxiliary operator, occurs at an inter-operator spacing of 70.5 bp. Although a peak is still observed at a spacing of 59.5 bp, it is greatly reduced in magnitude, being about half that observed at 70.5 bp (MuÈller et al., 1996). The reason

Phase-dependent Co-operativity with Dimeric Lac Repressors

853

Figure 2. Repression values for chromosomally located lacZ constructs with an Oid auxiliary operator at the indicated distance from O1 in its naturally occuring position in the presence of dimeric LacR. Dimeric LacR was expressed at 5  wild-type amounts (about 200 monomers/cell) from pSO331Stop (®lled circles, continuous line) or pSO331Gal (open triangles, broken line). The straight, broken line indicates the repression value obtained for O1 without an auxiliary operator in the presence of pSO331Stop; the same value is obtained in the presence of pSO331Gal (see also Table 1). Important features of the construct series are illustrated by a cartoon above the curve. O1 and the auxiliary Oid are indicated by open boxes. The lacUV5 promoter is indicated by a ®lled arrow. The start of the lacZ reporter gene, the site for insertion of spacer DNA and the inter-operator distance are indicated by text. The CAPbinding site and auxiliary operator O2 within lacZ are depicted as a crossed ellipse and box, respectively, to indicate that both are destroyed in all constructs. The series of lacZ constructs is fully described by MuÈller et al. (1996). Numbers above maxima indicate the spacing between the centres of symmetry of the operators in base-pairs: the 0.5 basepair spacing derives from the fact that Oid, unlike O1, lacks a central base-pair (Sadler et al., 1983; Simons et al., 1984). Strains bearing individual reporter constructs were transformed with the indicated plasmid and, after growth in minimal media as described (Oehler et al., 1990; MuÈller et al., 1996), speci®c b-galactosidase values were determined (Miller, 1972). Values were obtained from at least three colonies from a minimum of two independent transformations. Repression values are the mean speci®c activity in the presence of pSO1000A, in which lacI has been inactivated (Oehler et al., 1990), divided by the mean speci®c activity in the presence of the indicated LacR variant. Note that a repression value of 1 indicates no repression. All mean values have a standard deviation of 20% or less.

for this reduced maximum is presumably an increase in the energetic barrier to closing a DNA loop of the shorter length by tetrameric LacR (MuÈller et al., 1996). Second, the effect has disappeared by an inter-operator distance of about 150 bp (Figure 2). The effect of the local concentration of tetrameric LacR on co-operativity is well established (Oehler et al., 1994; Oehler & MuÈllerHill, 1994; MuÈller et al., 1996; MuÈller-Hill, 1998). At the repressor concentrations used in this study (5  i‡, which is equivalent to about 200 monomers per cell, or a dimer concentration of 10ÿ7 M) a calculation of local concentration as a function of inter-operator distance indicates that the effect should be appreciable for inter-operator distances shorter than 500 bp (MuÈller et al., 1996). Experimental values obtained for tetrameric LacR are in excellent agreement with similar calculations (Oehler et al., 1994; MuÈller et al., 1996). Thus, the discrepancy between observed and predicted

values for dimeric LacR indicates that the local concentration must be very high before an appreciable effect is observed, and therefore that the effect involved must be very weak in comparison to the interaction between tetrameric LacR and the auxiliary operator. Given that LacR binding to O1 represses transcription from the adjacent promoter by directly competing with RNA polymerase for binding (Schlax et al., 1995), a possible explanation for the periodic maxima of repression observed with dimeric LacR is that a second repressor dimer bound to the upstream Oid operator is sterically hindering RNA polymerase in its initial approach to the promoter. This would also explain the appearance of maximum repression values at a distance apparently unfavourable for stable loop formation, at least with tetrameric LacR. To test this idea, we generated a series of constructs in which the upstream, auxiliary operator was replaced with

854

Phase-dependent Co-operativity with Dimeric Lac Repressors

a b-centred trp operator (Staake et al., 1990; GuÈnes et al., 1996) with an inter-operator spacing of: 59.5, 65.5, 70.5, 75.5, 81.5, 85.5, 91.5, 95.5, or 105.5 bp, corresponding to consecutive maxima and minima values shown in Figure 2. Generation of the constructs, and integration of single copies into the chromosome of E. coli DC41-2 (relevant genotype, trpR‡) or DC41-3 (relevant genotype trpR; Staake et al., 1990) by virtue of l lysogens, was performed as described for the generation of the Oid ±O1 series (MuÈller et al., 1996). Repression values were then determined for these constructs in the presence of LacR331Stop and the presence or absence of TrpR. However, no phase-dependent difference in repression was observed (data not shown), indicating that the presence of a bound repressor upstream of the ÿ35 region of the lac promoter is not suf®cient to account for the observed enhancement of repression. Additional control experiments also demonstrate that occupancy of the upstream operator per se is not responsible for the enhanced repression observed. In the absence of an auxiliary operator, O1 alone leads to a roughly 100-fold repression with both wild-type, tetrameric LacR, and all dimeric LacR variants examined here (Table 1). If O1 is destroyed, then, as expected, repression is completely abolished. Introduction of an upstream, auxiliary Oid operator at an inter-operator spacing of 70.5 bp, the same spacing at which the maximal increase in repression is observed for tetrameric LacR (MuÈller et al., 1996), leads to a less than twofold increase in repression in the absence of O1 (Table 1). Similarly, if O1 (in the absence of an auxiliary operator) is replaced with O3, a weak operator variant that only has a 50% chance of being occupied under the experimental conditions used here (Oehler et al., 1994), then repression is drastically affected, falling to less than twofold for tetrameric, and all dimeric LacR variants examined (Table 1). As expected (Oehler et al., 1994; MuÈller et al., 1996), the addition of an auxiliary Oid operator at an inter-operator spacing of 70.5 bp to this

construct leads to signi®cant repression for tetrameric LacR. However, the enhancement observed for all dimeric LacR variants tested is marginal (Table 1). Another possible explanation for the periodic enhancement of repression is that there is some interaction between LacR dimers bound to the two operators, when they are in-phase. A similar type of interaction is presumed to be involved in the interaction between GalR dimers bound to inphase operators, although in this case the presumed interaction requires the presence of an additional protein, HU, for stabilisation of the DNA loop (Aki et al., 1996). It is apparent from the structures of the various forms of LacR (Friedman et al., 1995; Lewis et al, 1996) that constraints are placed on how loop formation can occur. First, the quaternary structure of tetrameric LacR is a Vform. The two dimers are joined at the C terminus, with the two N-terminal DNA-binding domains located at the ends of the two arms of the V. Second, there is a slight bend in the bound DNA away from each dimer in the crystal form. Thus, regardless of whether the observed quaternary structure of a V-form of a dimer of dimers is a ¯exible or rigid structure, the geometry of a loop formed between operators separated by, for example, six turns of the DNA helix absolutely requires adoption of the V-form by the bound, tetrameric repressor. This geometric constraint would be expected to have substantially disappeared by an inter-operator distance of about 150 bp (Shore & Baldwin, 1983), by which point we no longer observe co-operativity for dimeric LacR (Figure 2). Therefore, we examined the crystal structures of LacR in its various forms (Friedman et al., 1995; Lewis et al., 1996) for possible interactions between the two dimers. The only potential interaction between dimers (other than the C-terminal leucine heptad repeats) are between residues forming a triad of basic residues in one monomer of each dimer, the KRK motif from positions 325 to 327, and the backbone of residues 234 to 236 in the

Table 1. An upstream operator does not contribute to repression in the absence of a primary operator Operator varianta LacR w.t. tetrameric

base-pairs

Oid

70.5

Oid

70.5

a

Repression values for LacRb

O1 Oÿ 1 Oÿ 1 O3 O3

120 1.2 1.5 1.5 50

LacR331Stop dimeric 110 1.1 1.4 1.1 2.4

LacR331Gal dimeric 120 1.0 1.4 1.2 2.5

LacR331Basic dimeric 120 1.0 1.5 1.2 3.5

Operator variants have the form indicated by the cartoon. The operator at the right is in the position of O1, as compared to the reporter constructs shown in Figures 2 and 3, while the operator to the left, where shown, is upstream of the CAP-independent promoter. The Oÿ 1 sequence used is that given by Oehler et al. (1994). All other features of the reporter constructs used are identical to the constructs used in Figures 2 and 3. b Repression values are the quotient of the b-galactosidase speci®c activity obtained in the absence of any active repressor and the speci®c activity obtained in the presence of the indicated LacR variant, as detailed in the legend to Figure 2. Note that a repression value of 1 indicates no repression.

Phase-dependent Co-operativity with Dimeric Lac Repressors

opposing dimer. Speci®cally, Ne of R326 is separated from the backbone O of residue 235 by about Ê in the non-IPTG-bound form of LacR (Lewis 6.8 A Ê from et al., 1996), while Nz of K327 is about 6.9 A the backbone O of residue 235 in the IPTG-bound form of LacR (Friedman et al., 1995; Lewis et al., 1996). These distances are too great for a de®nite interaction; for comparison, the interacting atoms in the internal salt-bridge between R326 and D275 of the same monomer are separated by a distance Ê in all crystal forms (Friedman et al., 1995; 42.9 A Lewis et al., 1996; Suckow et al., 1996). However, they are the only candidates for any potential interactions. R326 does not tolerate any substitution tested, while some substitutions are tolerated at K327 (Li & Matthews, 1995; Suckow et al., 1996). We therefore studied the phase-dependence of repression mediated by two substitutions at this position: K327L (Kolkhof, 1994) and K327G (generated by PCR mutagenesis; data not shown). Although widely differing in the effect these two substitutions have on overall repressor activity, presumably due to their differential effect on protein stability (see below), both mutants still exhibit a detectable phase-dependence on repression (data not shown). This indicates that even if the hypothesised salt bridge can form between dimers that are not linked by the tetramerisation domain, it is

855

not the primary cause of the periodic enhancement of repression observed. Furthermore, a dimer bound to the auxiliary Oid operator at an interoperator spacing of 70.5 bp causes a maximal increase in local concentration at O1 to approximately 10ÿ5 M. At this concentration, gel ®ltration experiments with puri®ed LacR331Stop show no evidence for tetramer or higher-order aggregate formation (data not shown), indicating that tetramer formation by two dimers bound to DNA separated by this inter-operator distance is unlikely. Note that the shorter inter-operator distance of 59.5 bp only leads to a maximal local concentration of about 2  10ÿ5 M. Note also that co-operativity of transcription activation in yeast does not necessarily involve direct protein-protein interactions between transcription factors (Vashee et al., 1998). The presence of a cluster of basic residues on the surface of a DNA-binding protein, even a surface far removed from the DNA-binding domain, is intriguing, especially given that the basic nature of the motif is present in many members of the LacIGalR family. For example, the comparable residues in GalR are RRH (Weickert & Adhya, 1992). In addition to the experiments reported above, substitutions of K325 or K327 by leucine or alanine (Kolkhof, 1994), of R326 by lysine, alanine, glutamate, leucine or tryptophan, or progressive

Figure 3. Repression values for chromosomally located lacZ constructs with an Oid auxiliary operator at the indicated distance from O1 in its naturally occuring position in the presence of LacRBasic. LacR was expressed at 5  wild-type amounts from pSO331Basic, coding for a dimeric LacR variant that has a C-terminal basic tail. The broken line indicates the repression value obtained for O1 without an auxiliary operator in the presence of pSO331Basic (see also Table 1). The same series of reporter constructs was used as for Figure 2. Again, a cartoon above the curve illustrates important features of the constructs. Features shown are as described in the legend to Figure 2, and the constructs are fully described by MuÈller et al. (1996). Numbers above maxima indicate the spacing between the centres of symmetry of the operators in base-pairs. The lacZ constructs and experimental procedures were as described for Figure 2.

856

Phase-dependent Co-operativity with Dimeric Lac Repressors

removal of the entire motif by deletion (Li & Matthews, 1995), have a profound effect on DNAbinding activity, although this is primarily due to destabilising effects on protein structure (Li & Matthews, 1995). On the other hand, increasing the basic nature of the C terminus of LacR, by addition of a so-called basic tail, dramatically increases repression of various constructs in vivo (Kolkhof et al., 1995). We therefore examined the phase and distance-dependence of repression by LacRBasic (Figure 1(d)), and present the results in Figure 3. Again, repression of constructs containing two operators is markedly phase-dependent, with a maximal increase of repression of about 15-fold at an inter-operator spacing of 59.5 bp. As seen with LacR331Stop and unlike wild-type, tetrameric LacR, this repression enhancement is greater at an interoperator spacing of 59.5 bp than at 70.5 bp, and the effect has completely disappeared by an interoperator spacing of about 150 bp. Again, local concentration must be extremely high for an observable enhancement of repression. Loop formation by two dimers of LacRBasic speci®cally bound to the two operators, with the intervening loop stabilised by non-speci®c contacts of DNA to the basic tails of the two dimers, could explain this effect (Kolkhof et al., 1995). One dimer would have to be located in the loop, which still necessitates a tight turn of the DNA. However, if the loop were stabilised by protein-protein interactions, that are of necessity different to the interactions between dimers in tetrameric LacR, then this should reduce the energetic cost of loop formation. Alternatively, the formation of a DNA-protein-DNA sandwich complex, in which two in-phase operators are contacted by repressor dimers which then bind nonspeci®cally with their C termini to another portion of the DNA that is spatially close but linearly distant from the operator DNA, seems equally plausible. In either case, as the basic nature of the C terminus increases, from LacR331Stop and LacR331Gal, to LacRBasic, so too does the repression. However, the weakness of the interactions involved is underlined by its rapid dissipation with distance (Figures 2 and 3). This also explains why a marked increase in the basic nature of the C terminus of dimeric LacR only leads to an approximately twofold increase in repression values for co-operative interactions (Figures 2 and 3). Furthermore, removal of HU does not completely abolish co-operativity for GalR in vivo (Aki et al., 1996), suggesting that similar, weak contacts could also play a role here. In conclusion, a periodic enhancement of repression is observed when an upstream operator is present in-phase with dimeric variants of LacR, as is also the case for the related GalR. Additional factors may help stabilise loop formation mediated by non-speci®c (basic)protein-(acidic)DNA contacts, thus making them more amenable to identi®cation (Aki et al., 1996), although it is not yet clear whether any such factors play a role in LacRmediated repression.

Acknowledgements We are indebted to Petra Hammes, Jutta Kun and Ingrid MecklenbraÈuker for their outstanding technical support, and to Brigitte Kisters-Woike and Tom Steitz for valuable discussions. Supported by Deutsche Forschungsgemeinschaft through SFB 274.

References Aki, T., Choy, H. E. & Adhya, S. (1996). Histone-like protein HU as a speci®c transcriptional regulator: co-factor role in repression of gal transcription by Gal repressor. Genes Cells, 1, 179± 188. Alberti, S., Oehler, S., von Wilcken-Bergmann, B., KraÈmer, H. & MuÈller-Hill, B. (1991). Dimer-to-tetramer assembly of Lac repressor involves a leucine heptad repeat. New Biol. 3, 57 ± 62. Alberti, S., Oehler, S., von Wilcken-Bergmann, B. & MuÈller-Hill, B. (1993). Genetic analysis of the leucine heptad repeats of Lac represor: evidence for a four-helical bundle. EMBO J. 12, 3227± 3236. Bellomy, G. R. & Record, M. T., Jr (1990). Stable DNA loops in vivo and in vitro: roles in gene regulation at a distance and in biophysical characterisation of DNA. In Progress in Nucleic Acid Research and Molecular Biology (Cohn, W. & Moldave, K., eds), vol. 39, pp. 81 ± 128, Academic Press, Orlando, FL. Bellomy, G. R., Mossing, M. C. & Record, M. T., Jr (1988). Physical properties of DNA in vivo as probed by the length dependence of the lac operator looping process. Biochemistry, 27, 3900± 3906. Blattner, F. J., Plonker, G., III, Bloch, C. A., Perna, N. T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J. D., Rode, C. K., Mayhew, G. F., Gregor, J., Davis, N. W., Kirkpatrick, H. A., Goeden, M. A., Rose, D. J., Mau, R. & Shao, Y. (1997). The complete genome sequence of Escherichia coli K-12. Science, 277, 1453± 1462. Brenowitz, M., Mandal, N., Pickar, A., Jamison, E. & Adhya, S. (1991). DNA-binding properties of a Lac repressor mutant incapable of forming tetramers. J. Biol. Chem. 266, 1281±1288. Chakerian, A. E. & Matthews, K. S. (1991). Characterisation of mutations in oligomerisation domain of Lac repressor protein. J. Biol. Chem. 266, 22206± 22214. Choy, H. E., Park, S. W., Aki, T., Parrack, P., Fujita, N., Ishihama, A. & Adhya, S. (1995). Repression and activation of transcription by Gal and Lac repressors: involvement of alpha subunit of RNA polymerase. EMBO J. 14, 4523± 4529. Dunn, T. M., Hahn, S., Ogden, S. & Schleif, R. (1984). An operator at ÿ280 base-pairs that is required for repression of araBAD operon promoter: addition of DNA helical turns between the operator and promoter cyclically hinders repression. Proc. Natl Acad. Sci. USA, 81, 5017± 5020. Friedman, A. M., Fischmann, T. O. & Steitz, T. A. (1995). Crystal structure of lac repressor core tetramer and its implications for DNA looping. Science, 268, 1721± 1727. Fritz, H.-J., BicknaÈse, B., Gleumes, B., Heibach, C., Rosahl, S. & Ehring, R. (1983). Characterisation of two mutations in the Escherichia coli galE gene inactivating the second galactose operator and comparative studies of repressor binding. EMBO J. 2, 2129± 2135.

Phase-dependent Co-operativity with Dimeric Lac Repressors Gilbert, W., Majors, J. & Maxam, A. (1976). How proteins recognise DNA sequences. In Organisation and Expression of Chromosomes (Allfrey, V., Bautz, E. K. F., McCarthy, B. J., Schimke, R. T. & Tissieres, A., eds), pp. 167± 178, Dahlem Konferenzen, Berlin. Grif®th, J., Hochschild, A. & Ptashne, M. (1986). DNA loops induced by co-operative binding of l repressor. Nature, 322, 750± 752. GuÈnes, C Ë . , Staake, D., von Wilcken-Bergmann, B. & MuÈller-Hill, B. (1996). Cooperative binding of two Trp repressor dimers to a- or b-centred trp operators. Mol. Microbiol. 20, 375± 384. Hochschild, A. & Ptashne, M. (1986). Cooperative binding of l repressors to sites separated by integral turns of the DNA helix. Cell, 44, 681±687. Irani, M. H., Orosz, L. & Adhya, S. (1983). A control element within a structural gene: the gal operon of Escherichia coli. Cell, 32, 783± 788. Kolkhof, P. (1994). Protein Design bei DNA bindenden Proteinen. PhD thesis, University of Cologne, Cologne. Kolkhof, P., Oehler, S., Alex, R. & MuÈller-Hill, B. (1995). A basic tail increases repression by dimeric Lac repressor. J. Mol. Biol. 247, 396± 403. KraÈmer, H., NiemoÈller, M., Amouyal, M., Revet, B., von Wilcken-Bergmann, B. & MuÈller-Hill, B. (1987). Lac repressor forms loops with linear DNA carrying two suitably spaced lac operators. EMBO J. 6, 1481± 1491. KraÈmer, H., Amouyal, M., Nordheim, A. & MuÈller-Hill, B. (1988). DNA supercoiling changes the spacing requirement of two lac operators for DNA loop formation with lac repressor. EMBO J. 7, 547± 556. Law, S. M., Bellomy, G. R., Schlax, P. J. & Record, M. T., Jr (1993). In vivo thermodynamic analysis of repression with and without looping in lac constructs. J. Mol. Biol. 230, 161±173. Lee, D.-H. & Schleif, R. (1989). In vivo loops in araCBAD: size limits and helical repeat. Proc. Natl Acad. Sci. USA, 86, 476± 480. Lehming, N., Sartorius, J., Oehler, S., von WilckenBergmann, B. & MuÈller-Hill, B. (1988). Recognition helices of lac and l repressor are oriented in opposite directions and recognise similar DNA sequences. Proc. Natl Acad. Sci. USA, 85, 7947± 7951. Lewis, M., Chang, G., Horton, N. C., Kercher, M. A., Pace, H. C., Schumacher, M. A., Brennan, R. G. & Lu, P. (1996). Crystal structure of the lactose operon repressor and its complexes with DNA and inducer. Science, 271, 1247± 1254. Li, L. & Matthews, K. S. (1995). Characterization of mutants affecting the KRK sequence in the carboxyl-terminal domain of lac repressor. J. Biol. Chem. 270, 10640± 10649. Majumdar, A., Rudikoff, S. & Adhya, S. (1987). Puri®cation and properties of Gal repressor: pL-galR fusion in pKC31 plasmid vector. J. Biol. Chem. 262, 2326± 2331. Miller, J. H. (1972). Experiments in Molecular Genetics, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

857

Mossing, M. C. & Record, M. T., Jr (1986). Upstream operators enhance repression of the lac promoter. Science, 233, 889± 892. MuÈller, J., Oehler, S. & MuÈller-Hill, B. (1996). Repression of lac promoter as a function of distance, phase and quality of an auxiliary lac operator. J. Mol. Biol. 257, 21 ±29. MuÈller-Hill, B. (1983). Sequence homology between Lac and Gal repressors and three sugar-binding periplasmic proteins. Nature, 302, 163±164. MuÈller-Hill, B. (1998). The function of auxiliary operators. Mol. Microbiol. 29, 13 ± 18. Oehler, S. & MuÈller-Hill, B. (1994). Prokaryotic control of transcription: how and why does it differ from eukaryotic control? In Inducible Gene Expression and Cytoplasmic/Nuclear Signal Transduction (Baeurle, P., ed.), pp. 1 ± 24, Springer Verlag, New York. Oehler, S., Eismann, E. R., KraÈmer, H. & MuÈller-Hill, B. (1990). The three operators of the lac operon cooperate in repression. EMBO J. 9, 973± 979. Oehler, S., Amouyal, M., Kolkhof, P., von WilckenBergmann, B. & MuÈller-Hill, B. (1994). Quality and position of the three lac operators of E. coli de®ne ef®ciency of repression. EMBO J. 13, 3348± 3355. Sadler, J. R., Sasmor, H. & Betz, J. L. (1983). A perfectly symmetrical lac operator binds the lac repressor very tightly. Proc. Natl Acad. Sci. USA, 80, 6785± 6789. Schlax, P. J., Capp, M. W. & Record, M. T., Jr (1995). Inhibition of transcription initiation by lac repressor. J. Mol. Biol. 245, 331± 350. Shore, D. & Baldwin, R. L. (1983). Energetics of DNA twisting. I. Relation between twist and cyclization probability. J. Mol. Biol. 170, 957± 981. Simons, A., Tils, D., von Wilcken-Bergmann, B. & MuÈller-Hill, B. (1984). Possible ideal lac operator: Eschericia coli lac operator like sequences from eukaryotic genomes lack the central G-C base pair. Proc. Natl Acad. Sci. USA, 81, 1624± 1628. Staake, D., Walter, B., Kisters-Woike, D., von WilckenBergmann, B. & MuÈller-Hill, B. (1990). How Trp repressor binds to its operator. EMBO J. 9, 1963± 1967. Suckow, J., Markiewicz, P., Kleina, L. G., Miller, J., Kisters-Woike, B. & MuÈller-Hill, B. (1996). Genetic studies of the Lac repressor XV: 4000 single amino acid substitutions and analysis of the resulting phenotypes on the basis of the protein structure. J. Mol. Biol. 261, 509± 523. Vashee, S., Melcher, K., Ding, W. V., Johnston, S. A. & Kodadek, T. (1998). Evidence for two modes of cooperative DNA binding in vivo that do not involve direct protein-protein interactions. Curr. Biol. 8, 452± 458. Weickert, M. J. & Adhya, S. (1992). A family of bacterial regulators homologous to Gal and Lac repressors. J. Biol. Chem. 267, 15869± 15874.

Edited by M. Yaniv (Received 30 June 1998; received in revised form 7 August 1998; accepted 15 September 1998)