Shared operator recognition specificity between Trp repressor and the repressors of bacteriophage 434

J. Mol. Biol. (1991) 217, 599-602

Shared Operator Recognition Specificity between Trp Repressor and the Repressors of Bacteriophage 434 Ronald L. Somervillet, Gregg Bogosian$ and Jill H. Zeilstra-Ryallss Department of Biochemistry Purdue University West Lafayette, IN 47907, U.S.A. (Received 1 May

1990; accepted 15 October 1990)

Trp repressor is the only DNA-binding regulatory protein having a helix-turn-helix motif that has been reported to engage its operator target by a mechanism termed indirect readout: the Trp repressor-DNA interface is replete with hydrogen bonds between amino acid residues and non-esterified oxygen. atoms of the sugar-phosphate backbone, and eontains numerous specifically positioned water molecules. In Escherichia coli mutants repressor of phage 434 led to an eightfold reduction in trp deleted for trpR, the immunity promoter utilization. The Cro434 repressor also inhibited transcription from t,he trp promoter. The 434 repressors, considered to interact directly with operator targets, carry recognition heliees positioned near the N terminus of each protein. The DNA-recognizing elements of Trp repressor lie toward the C terminus. thus appears to The trp operator possess significant plasticity in terms of its ability to assume conformational states that allow complex formation with more than one class of regulatory protein.

As a result of crystallographic analyses, two distinguishable modes of operator recognition by regulatory proteins of the helix-turn-helix category have recently emerged. The complex between Trp holorepressor and its operator target involves a form of interaction (indirect readout) where there is negligible direct contact between the repressor protein and the purine-pyrimidine bases of the operator. Instead, water molecules mediate the contacts between amino acid residues in the recognition helices and the edges of bases in the major groove of duplex DNA that are critical to regulatory specificity (Otwinowski et al., 1988; Marmorstein & Sigler, 1989). In contrast, protein-DNA interaction between the lambdoid phage repressors and their operator targets involves a network of direct hydrogen-bonding and van der Waals’ contacts between specific amino acid side-chains and functional groups of purine and pyrimidines that are exposed within the major groove (Aggarwal et al., 1988;

Jordan & Pabo, 1988; Wolberger et al., 1988; Harrison & Aggarwal, 1990). In each situation, substantial localized structural deformations of the operator were noted. Segments of DNA that are in complexes with repressors are bent, overwound or underwound, changed in the width of their major and minor grooves, and are altered in various parameters of base-pair orientation and stacking (Aggarwal et al., 1988; Otwinowski et al., 1988; Wolberger et al., 1988). An operator target recognized b,y two different lambdoid repressors adopts different conformations upon forming complexes with the different proteins (Wolberger et al., 1988). The crystallographic analyses and much of the previous work on protein-DNA interaction have been predicated on assumptions of topographical uniqueness that conferred exclusivity upon the interacting protein and DNA surfaces. Few experiments have addressed the possibility that illegitimate protein-DNA interactions might occur, with consequent non-physiological effects. Here, we present evidence for cross-recognition of targets between the trp operator and lambdoid phage repressors of the 434 class. These repressor-operator systems have hitherto been regarded as totally noncognate, a notion consistent with the above-cited crystallographic studies (Aggarwal et al., 1988;

t Author to whom all correspondence should be addressed. $ Present address: Animal Sciences Division, Monsanto Corporation, Chesterfield, MO 63198, U.S.A. $ Present address: Department of Cellular, Viral and Molecular Biology, University of Utah Medical Center, Salt Lake City, UT 84132, U.S.A. 599 0@22-2836/91/040599-04

$03.00/O

0

1991 Academic

Press Limited

R. L. SomerviEEe et al.

600

80 90 IGln Arg GLu Leu Lys Asn Glut Leu Gly Ala GLy ]lle Ala Thr Ite Thr Arg Gly Ser Asn Ser Leu Lys Ala] . . . l * . . . . 1 / ‘;I I I

TRP R

Asn Gln Alo Gtu Leu Ala Glnl Lys Vat Gly Thr Thr Gtn Gtn Ser lie Glu Gtn Leu GCu Asn

REP434

I CR0434

20

I

I

I

I

I

,b

I

I

3:

Lys Met Thr Gln Thr GCu Leu Ata Thr Lys Ata Gly 1Vat Lys Gln Gtn Ser Ite GLn Leu Ile]

Figure 1. Primary amino acid sequences of those parts of Trp repressor (TRP R), 434 repressor (REP434) and 434 Cro proteins (CR0434) that make up the helix-turn-helix elements important to operator recognition. The boxed-in segments correspond to the bihelieal recognition motifs as determined crystallographically for Trp repressor (Otwinowski et ccl., 1988), 434 Cro protein (Wolberger et al., 1988) and 434 immunity repressor (Aggarwal et al., 1988). The sequences have been aligned to maximize amino acid identities. Across the helix-turn-helix motif there are 5 identities among the 3 sequences. There are 10 identities between 434 Cro protein and 434 immunity repressor. The numerals above certain amino acid residues refer to the distance of that residue from the N terminus. The dots beneath Trp repressor sequenees refer to locations where mutational alterations in protein structure diminish the binding of this protein to its primary operator target (Kelley & Yanofsky, 1985; Bass et al., 1988; Hagewood & Somerville, unpublished results).

Otwinowski et aZ., 1988; Wolberger et al., 1988). Our observations raise the possibility that certain repressors having the helix-turn-helix motif may be capable of interacting with a variety of operator targets, perhaps in ways that involve indirect readout mechanisms. The amino acid sequence of the three bihelical operator-recognition elements of Trp repressor, 434 repressor and 434 Cro protein are shown in Figure 1. There are a number of amino acid residue identities between these sequences that offer a basis for However, the known aligning primary structures. boundaries of the a-helical operator recognition elements, as determined crystallographically (Aggarwal et al., 1988; Otwinowski et al., 1988; Wolberger et al., 1988) cannot be aligned precisely on the basis of amino acid identities. The first indication that the immunity repressor of phage 434 can partially substitute for Trp repressor came from a study of Escherichia coli lysogens carrying phages wherein the 3, immunity

Table 1 promoter utilization in trpR+ and trpR E. coli strains whose cytoplasm contains either the 434 immunity repressor or the 2 immunity repressor

region of a trp-lae operon fusion phage, LGB8 (Bogosian & Somerville, 1984), had been replaced by t,he immunity region of phage 434 (Table 1). Also studied was a trpP (0”) derivative, iGBSS1, that arose spontaneously in a stock of E. coli Pu’K5031 (/ZGB2). By DNA sequencing it was established that the oc mutation of IGBZSl was an A. T to G. C to G/C transition at co-ordinate - 15, relative to the startpoint of transcription of the trp promoter (see Fig. 2). This mutant, is thus ident,ical to 0’6, previously described by Bennett, & Yanofsky (1978). These workers found that, a trp promoter with t’his oc mutation was repressible by a factor of about 5, and that the nucleotide switch did not affect, promoter strength. Our results are congruent with these earlier findings, alt’hough in our hands this particular 0’ mutation also has a promoter-up effect amounting to about, 2.4-fold. The operator-constitutive nature of this A. T to G. C switch is supported by the work of Bass et al. (1987). In host cells that were genetically disabled for Trp repressor production, the presence of the 434 immunity repressor imposed an eightfold reduction in trp promoter utilization (Table 1, lines I and 2). A similar degree of reduction in promoter utilization

trp

p-Galactosidase Promoter trpp+ trpP(0’)

Immunity &age

level

repressor/

1, (/ZGB2) 434 (aGB2i434) I. (IGBZSl) 434 (IGB2Sli434)

trpR+

AWR)

130 130 2860 1320

6200 830 14,900 1760

Single lysogens of 2 isogenie A(Eac) E. coEi strains, NK5031 (trpR+) and SP-361 (trpR) were grown at 30°C in minimal casein hydrolysate glycerol medium containing 100 pg L-tryptophan/ml and assayed as described (Bogosian & Somerville, 1984). The /3-galactosidase values are the results of multiple assays and were reproducible + 10%. The 1“M derivatives of IGB2 and IGBSSl were constructed by standard phage crosses (Parkinson, 1968) using AhsV3“ obtained from S. Adhya.

TGAACTiOTTAACTAGPTCA

trpo

$++ACAA$A~TTGT+T+

0434

--AAC-AG-TA--T-GTT-A

Consensus

Figure 2. B comparison

of consensus target sequences capable of entering into complex formation with Trp repressor, 434 immunity repressor and 434 Cro protein. The sequences shown are taken from the work of Bennett & Yanofsky (1978), Bass et al. (1987) and Wharton et aE. (1984). Residues critical for repressor binding, as revealed by mutatsional analysis, are indicated in boldface. A “consensus consensus” sequence indicating residues in common between these 2 sets of targets is shown. The asterisk indicates the position of the oc mutation in IGB2Sl (see the text). The central axis of symmetry of the trp operator shown above may not coincide with the dyad axis of bound Trp holorepressor (Staacke et al., 1990).

Communications

601

Table 3

Table 2 promoter utilization in trpR E. coli strains harboring an expression system for the Cro434 protein

trp

Host strain

Plasmid

PGalactosidase

SP361 (IGB2) SP361 (1GB2)

pGL139

5520 2370

SP968 (1GB2) SP968 (1GB2)

pGL139

5970 2000

level

Two different A(lac) E. coli strains both also deleted for trpR, were grown and assayed as described in Table 1. The Zac P/o-cro434 plasmid pGL139, donated by M. Ptashne, has been described (Laner et al., 1981). SP968 is essentially identical to SP361, except for the presence of a zjj :: TnlO element at 99 min region of the host chromosome.

was imposed by 434 repressor on a trp oc mutant (Table 1, lines 3 and 4). This result suggests that the base-pair altered by this oc mutation is not critical for the recognition of the trp operator by the 434 immunity repressor. In trpR+ host st,rains, the 434 immunity repressor had negligible effects on trp promoter utilization (Table 1, column 3). Independent control experiments (Table 3) established that 2 repressor had negligible effects on trp promoter utilization. We next evaluated the effect on trp promoter into utilization of the Cro434 protein, by introducing a pair of trpR mutants a plasmid wherein the cro434 gene is expressed from the Zac promoter (Table 2). was able to reduce trp Although Cro434 protein promoter utilization by a factor of only 2 to 3, the effect was highly reproducible. The 434 immunity repressor and Cro434 proteins are known to engage structurally identical operator targets in different ways (Wolberger et al., 1988), so it is not unexpected that there should be different consequences when these repressors interact with the heterologous trp operator. Additional supporting evidence for overlapping specificity between the trp and 434 repressor-operator systems comes from an analysis of the tryptophan analog susceptibility of strains harboring multicopy plasmids encoding the 434 immunity of repressor of Cro434 (Table 3). Both categories plasmid imposed 5methyltryptophan sensitivity upon a A(trpR) tester strain, a phenotype indicative trp of reduced transcription from the primary promoter (Somerville, 1983). A comparison of the consensus operator targets of Trp repressor and the lambdoid phage 434 repressors (Fig. 2) suggests that a common consensus sequence, AACNAN,TNGTT, in duplex form, may provide an interaction surface common to these three DNA-binding proteins. The variability in operator structure revealed by crystallographic analysis (Wolberger et al., 1988) and the fact that non-contacted bases within the 434 operator affect the interaction of this target sequence with regulatory proteins (Koudelka et al., 1987) is fully compat-

5-Methyltryptophan sensitivity of E. coli strains relation to repressor content of cells Strain NK5031 SP-967 SP-967(pGYlOl) SP-967(pGL139) SP-967(pJZ200)

Relevant

genotype

trpR+ WPR) A(trpR), imm434 A(trpR), c~o‘-~ A(trpR), imm”

in

Zone of inhibition (mm) 43 0 35 28 0

The susceptibility to inhibition by DL-5-methyltryptophan of the strains examined was performed as described (Kuhn et al., 1972). Plates were incubated at 37°C overnight: 80 pg of inhibitor was added to each disk. The diameter of the zone of inhibition reflects the ability of RNA polymerase to initiate transcription at the trp promoter (Somerville, 1983). Plasmid pGYlO1 has been described (Levine et al., 1979). Plasmid pJZ200 is a derivative of pTR262 (Roberts et al., 1980) having a kanamycin resistance element inserted at the unique BamHI site.

ible with the degeneracy in specificity that we have observed, and with earlier studies showing that Trp repressor may interact with a number of secondary operator targets (Bogosian & Somerville, 1983). Reciprocal interaction between the repressoroperator systems of phage 434 and Rhizobium meliloti phage 16-3 has been reported (Dallmann et al., 1987). It would be of interest to test whether trp promoter utilization is altered in the presence of the 16-3 repressor. Our results can be rationalized in several different ways: First, it is possible that the crystallographic analyses only partially describe the possible modes of interaction in vivo between the three repressors and their operator targets. That the available Trp holorepressor-operator crystals (Otwinowski et al., 1988) might involve non-specific protein-DNA has been suggested (Brennan & interaction Matthews, 1989; Lehming et al., 1990). Second, there may be enough conformational flexibility inherent within the operator targets that we have studied to allow the presentation by such segments of DNA of surfaces able to interact with more than one class of helix-turn-helix regulatory protein. How regulatory exclusivity is successfully maintained in the face of such potentially pathological intrusions by certain DNA-binding proteins into the regulatory domains of other control proteins is unclear. The 434 immunity repressor, at low levels, exerts little effect on trp promoter utilization in trpR+ cells. There was virtually no change in the rate of trp promoter utilization in trpR+ cells, whether or not an oc mutation was present, but, decreases of seven- to eightfold were observed in trpR mutant cells (Table 1). Perhaps Trp repressor protein, in a conformational state that is regulatorily neutral, is able to specifically associate with promoter-operator target regions of duplex DNA in a way that allows limited access to the promoter by RNA polymerase, yet shields the promoter from unauthorized regulatory proteins.

R. L. Somerville

602

That the helix-turn-helix motif of Trp repressor may be structurally distinct from similar motifs within phage immunity repressors has been proposed (Bass et al., 1988; Bushman et aE., 1989). Recently, however, a large number of similarities between Trp repressor and the lambdoid repressors have been noted (Pabo et al., 1990). The latter interpretations are consistent with our observations. We thank Barbara Nicodemus for expert technical assistance. This study benefited greatly from donations of biological materials by Byron Hagewood, Sankar Adhya, Naomi Franklin, Raymond Devoret and Mark Ptashne. This work was supported by a grant (GM22131) from the National Institutes of Health. This is Journal Paper no. 12, 132 from the Purdue University Agricultural Experiment Station.

References Aggarwal, A. K., Rodgers, D. W., Drottar, RI., Ptashne, M. & Harrison, S. C. (1988). Science, 242, 899-907 Bass, S., Sugiono, P., Arvidson, D. N., Gunsalus, R. P. & Youderian, P. (1987). Genes Develop. 1, 565-572. Bass, S. Sorrells, V. & Youderian, P. (1988). Science, 242, 240-245. Bennett, G. N. & Yanofsky, C. (1978). J. Mol. Biol. 121, 179-192. Bogosian, G. & Somerville, R. L. (1983). Mol. Gen. Genet. 191, 51-58. Bogosian; G. & Somerville, R. L. (1984). Mol. Gen. Genet. 193, 110-118. Brennan, R. G. & Matthews, B. (1989). J-. Biol. Chem. 264, 1903-1906. Bushman, F. D., Chang, S. & Ptashne, M. (1989). Cell, 58, 1163-1171. Dallmann, G., Papp, P. & Orosz, L. (1987). Nature (London), 330, 398-401. Harrison, S. C. & Aggarwal, A. K. (1990). Annu. Rev. Biochem. 59, 933-969. Edited

et al.

Jordan, S. R. & Pabo, C. 0. (1988). Science, 242, 8933899. Kelley, R. L. & Yanofsky, C. (1985). Proe. Nat. Acad. Sei., U.S.A. 82, 483487. Koudelka, G. B., Harrison, S. C. & Ptashne, M. (1987). Nature (London), 326, 886-888. Kuhn, J. C., Pabst, M J. & Somerville, R. L. (1972). J. Bacterial. 112, 933101. Laner, G., Pastrana, R., Sherley, J. & Ptashne, M. (1981). J. Mol. Appl. Genet. 1, 139-147. Lehming, N., Sartorius, J., Kisters-Woike, B., von Wilcken-Bergmann, B. & Miiller-Hill, B. (1990). EMBO J. 9; 615-621. Levine, A., Bailone, A. & Devoret, R. (1979). j. Mol. Biol. 131, 655-661. Marmorstein, R. Q. & Sigler, P. B. (1989). In Nucleic Acids and Molecular Biology (Eckstein, F. & Lilley, D. M. J., eds)? vol. 3, pp. 56-78, Springer-Verlag, Berlin/Heidelberg. Otwinowski, Z., Schevitz, R. W., Zhang, R.-G., Lawson, 6. L., Joachimiak, A., Marmorstein, R. Q.; Luisi, B. F. & Sigler, P. B. (1988). Nature (London), 335, 321-329. Pabo, C. O., Aggarwai, A. K.. Jordan, S. R., Beamer, L. J., Obeysekare, U. R. & Harrison: S. C. (1990). Science, 247; 1210-1213. Parkinson, J. S. (1968). Genetics, 59, 311-325. Roberts, T. M., Swanberg, S. L., Poteete, A., Riedel; G. & Backman, K. (1980). Gene, 12, 1233127. Somerville, R. L. (1983). In Amino Acids: Biosynthesis and Genetic Regulation (Herrmann, K. M. $ Somerville, R. L.; eds), pp. 351-378, AddisonWesley, Reading, MA. Staacke, D., Walter, B., Kisters-Woike, B., von Wilcken-Bergmann, B. 8i Miiller-Hill, B. (1990). EMBO J. 9, 1963-1967. Wharton, R. P., Brown, E. L. & Ptashne, M. (1984). Cell, 38, 361-369. Wolberger, C., Dong, Y., Ptashne, M. 85 Harrison S. C. (1988). Nature (London), 335, 789-795.

by J. H. Miller