Lambda phage cro repressor

(1982) 162, 251-266

J. Mol. Biol.

Lambda Phage cro Repressor DNA

Sequence-dependent

FRANK

BOSCHELLI,

Interactions

Seen by Tyrosine

Fluorescence

KIM ARNDT, HARM- NICK?, QIJIU ZHANG~, PON~Y Lv§ Department of Chemistry University of Pennsylvania Philadelphia, Pen,n. 19104, U.S.A. ASD

YOSHIX)RI TAKEDA Department University Catonsville.

(Receivrd

of Chemistry of Maryland

Md 21228, U.S.A. 5 April

1982)

cro repressor of coliphage lambda plays a vital role in the regulation of the phage lifecycle. It exerts control of phage functions at the transcriptional level by its ability to recognize and specifically bind the operators controlling the expression of the early genes of the phage. We have examined the interaction of cro with double and singlestranded DNA and RNA, and a synthetic lambda pseudo-operator, by monitoring the quenching of the tyrosyl fluorescence of cro repressor by added nucleic acid. cro repressor was found to bind single-stranded DNA with affinities similar to that seen for non-specific double-stranded DNA. It also binds synthetic RNA to a significant degree. The lambda pseudo-operator, which differs in three base-pairs out of 17 from Oa3, the operator with the greatest affinity for O-O,binds cro repressor more tightly than does non-specific DNA. The stability of the cro repressor-pseudo-operator complex as a function of salt concentration shows that the specific complex is stabilized by other factors in addition to the ionic interactions seen with cro repressor-non-specific DNA complexes.

1. Introduction of gene expression in the temperate coliphage lambda involves a complex series of closely interrelated processes, in which transcriptional and translational control as well as host-related factors play a part (Herskowitz & Hagen, 1980; Echols, 1980). Special interest lies in the regulatory elements that’ participate in the lysis-lysogeny decision during the phage life-cycle: the phage encodes two repressor proteins that play essential but opposing roles. The product The regulation

t Present address: Biological

Laboratories,

Harvard

University,

Cambridge MA, U.S.A.

1 Permanent address: Academia Sinica, Beijing, People’s Republic of China. 5 Author to whom correspondence

should be addressed. 2.51

OO22-ZS36/82/340251-16 $03.00/00

0 1982 Academic Press Inc. (London) Ltd.

"5,"

F.

BOSC’HELLI

E7’

.1 /,.

ofthe cI gene, cl repressor, is required for the establishment and the maintenance of the lysogenic state, whereas the product of the cro gene, cro repressor. is necessary during the lytic cycle of phage development. Both cro and cl repressors compete for and bind t,wo operator regions cont’rolling the promoters P,, P,, and P, within the phage genome. These operator regions each consist of three binding sites for the repressor molecules (Ptashne et al., 1980,1982), and the respective repressors have different relative affinities for each of the sites within the operator regions (Johnson et aZ., 1981). The interaction at these binding sites by cro and cl repressors is a clear example of a developmental switch (Echols. 1980: Herskowitz & Hagen. 1980; Ptashne et al.. 1980). Thus, it is of interest to study in detail the proteinnucleic acid interactions in this system. These protein-nucleic acid interactions have been analysed on the DN4 by Johnson rt al. (1978.1979,1981) with the chemical probe experiments of Gilbert et al. (1976). We have looked at this interaction at the protein by characterizing the fluorescence of the t’yrosines of cro repressor with added nucleic acids. including both specific and non-specific double-stranded DNA as well as single-stranded DNA and the synthetic RNA. The results show that it is possible to examine the structure of the cro repressor-DNA complex as w:ell as the kinetics of its formation by direct physical chemical methods. A spectroscopic study of cro repressor is of value because both the primary (Hsiang rt al.. 1977: Itobert,s rt al., 1977) and the t,hreedimensional structures (Anderson et al., 1981) are known. The cro repressor is a dimer with t\vo identical subunits of 2V2, 7351. each having three tyrosyl, three phenylalanyl, one histidyl, and no trypt,ophanyl residues (Takeda et al.: 1977: paper (Boschelli, 1982) we have Hsiang et al.. 1977). In this and the accompanying focused on the fluorescence of the tyrosyl residues located at positions lo,26 and 5 I, and the quenching of this fluorescence by added nucleic acid. ,4 survey of wo repressor interactions with a variety of nucleic acids is presented here, while in the accompanying paper, the interaction of cro with double-stranded DNA is characterized quantitatively. We show that the fluorescence measurements can distinguish clearly between the specific and non-specific cro repressor-DNA complexes.

2. Materials cro repressor

was isolated

and Methods

from (ISHSti (Miller.

1972) cells freshly

pTR214 DNA, which contains the wo gene rloned distal 1979). Cells were grown to saturation in the prestnoe

medium.

The isolation

procedure

\vas rssentially

transformed

with

to the lac promoter (Koberts it al.. of 40 pg ampicillin/ml in minimal

that of Takeda rf al. (1977). w&h the

exception that the DNA-cellulose column chromatography step was omitted. cro repressorcontaining fractions from the phosphocellulose column were detected by radioimmunoassay. whereas fractions from t,he G75 Sephadrx (Pharmacia) were assayed for CTO by the fluorescence emission spectrum. The purity of the cv-o repressor could be monitored conveniently by the fluorescence emission spect,rum, since cro contains 3 tyrosyl and no tryptophanyl residues. For the fluorescence titrations. only fractions emitting with an emission maximum of 310 nm or less were pooled. These fractions were greater than 98’7” purr as judged by polyacrylamide gel electrophoresis. The pooled fractions were precipitated with 540 mg ammonium sulfate/ml followed by resuspension in and dialysis at 4°C: versus cro

LAMBDA

PHA(:E

CTO REPRESSOR

“53

storage buffer (0.1 iw-Tris. HCl (pH 7.3), 0.2 M-KCl, 0.1 mM-KDTA, 304/, (v/v) glycerol) and stored at - 20°C. The molar absorptivity of cro repressor was determined by comparing the absorbance ofcro repressor in storage buffer to the absorbance of cro repressor in buffered 6 iv-guanidine. H(‘1. Since guanidine.HCl does not affect the electronic properties of tyrosine (Cowgill, 197(i), an accurate determination of the molar absorbance of the protein can be made in guanidine HCI. Triplicate determinations showed that the A,,, value for cro repressor is 0.54 unit/mg per m I This is slightly higher than the value of 0.5 predicted from the number of tyrosyl residues in cro repressor. and is consistent with the observation that the absorpt,ion spectrum of cro repressor is red-shifted by 1 nm to 276 nm relative to free tyrosine. (b) DNA

for titrations

Calf thymus. salmon testes and Micrococcus Z&us DNA were obtained from Sigma. Synthetic nucleic acids were purchased from Collaborative Research. Poly[d(A-T)] was synthesized by the de nova mechanism using DNA polymerase I (Kornberg, 1980). lac operator fragment was isolated as described by Lillis etal. (1982). All DNA preparations were suspended in BB buffer (10 mw-Tris.HCl, 0.1 mM-EDTA, pH 7.3 at 4°C) with KC1 added to desired levels according to the titration performed. In the case of the native calf thymus DNA preparations, small DNA or nucleotides were removed by re-extraction through precipitation with ethanol and spooling on a glass rod to obtain only the high molecular weight DNA. D1\‘A stocks more than a week old were extracted with chloroform/isoamyl alcohol (24: I, v/v), followed by extraction with ether and precipitation with ethanol to spool out the extracted DNA. Bacteriophage fd DNA was prepared from intact phage by extraction 4 times with chloroform/isoamyl alcohol (24: 1, v/v) followed by extraction with ether and precipitation M-ith ethanol. The molar absorptivities of the double-stranded DNAs were taken to be 6.5 x IO3 Azeo units/m01 base: fd DNA, 7.4 x IO3 A,,, units/m01 base (Berkowitz XT Day. 1974). The synthetic nucleic acids were assumed to possess t,hemolar absorbance cited by the supplier. (c) Double-stranded

DSA

fragment

prqaration

From 1 to 2 g of DNA (calf thymus, salmon testes (Sigma), or pMB9 DNA (Rodriguez et al., 1976)) were suspended in 500 ml of 10 mmTris.HCl (pH 7), 0.1 mm-EDTA by vigorous t,hr sample was made 0.1 M in sodium borate shaking overnight. After resuspension. (?Ja,B,O,) and the pH adjusted t,o 9.0, and then CaCl, was added to 0.01 M. Approx. 3000 to 5000 units of Staphylococcal nuclease (Worthington) were added and the digestion of the et al.. DNA was monitored by the increase in the optical density at 260 nm (Cuatrecasas 1967). Temperatures ranging from ambient to 37°C were employed with only the rate of digestion being affected. Sfter a 25’+, increase in the optical density at 260 nm, the reactioti was terminated by the addition of solid Na,EDTA to 0.05 M. Orcasionally, addit,ional enzyme was required durmg the digestion to achieve the optical density increase desired. After the added EDTA was dissolved. the digested DNA was loaded on a 5 cm x 5 cm DEAE-Sephadex A50 (Pharmacia) column equilibrated with 0.2 iv-NaCl DB buffer (0.2 XINaCl, 10 mM-Tris.HCl (pH 7), 10 rnM EDTA). Under these conditions, calcium, mononucleotides and small oligonucleotides are eluted, while larger fragments are retained by the column. The column was washed with 0.2 M-NaCl DB buffer until the optical density of the eluent. was less than 0.05 at 260 nm, whereupon a 1 M-NaCl DB buffer step gradient was applied to the column to recover the DNA. The eluted DX‘A was precipitated with ethanol followed by extraction with chloroform/isoamyl alcohol. This extracted sample w-as precipitated with ethanol, resuspended in a minimal volume, and loaded on a Sephacryl S300 (Pharmacia) column (2.5 cm x 2.5 m) equilibrated with 0.04 M-NaCl DB buffer. The DNA in the fractions were assayed for size on an 8 cm native I;i”/,, (w/v) polyacrylamide gel (37 : 1 acrylamide to bisacrylamide ratio). The desired size range fractions were pooled and precipitated with ethanol. Figure 1 shows the migration pattern of the DiYA fragments compared with a 36 base-pair

254

F. BOSC’HELLI

ET

AL.

FIN:. 1. High-resolution native polvacrylamide gel of’siz~,~f’rac.tiorlat~tl DSA fi-agments. A 40 cm l.i”,, (w/v) acrylamide (37: 1, w/w ac~rylamitle/bisac~rylamitlr) gel with a 3 c’m .V, ac*rylamide stacking gel was employed. The gel was run at 1000 V wit.h 40 mw’l%a-borate (pH 8.3) reservoir buffer. Bromophenol blue and xylem cyanol dye markers were used, with the sylene ~yanol running just, behind the 36 baw pair lac operator rrst,rict,ion fragment (indicated by the arrow in lane a) used as a size standard.

Zac operator rrst,riction f’ragment. Discrete bands are observed in this native gel, indicating the double-stranded nature of these fragment,s with a size range of’25 to 40 base-pairs. The Figure shows a 40 cm gel, but we have observed hands. less well-resolved. on 8 cm gels. The condit,ions used for electrophoresis arc indicated in the Figure legend. Approximately 20 mg of fragments in the 25 to 50 base-pair size range \verc obtained from 1 g of DNX. (d) LarnDda pseudo-oprator The lambda pseudo-operator. a 17 has0.05 M-Na+ : Shen & Hearst, 1976), the actual contribution of the single-stranded DNA to the cro interaction may be masked. It is shown in the accompanying paper that cro repressor binds single-stranded deoxyoligonucleotides. (c) cro repwssor-RNA

hteraction

Figure 7 shows that cro repressor binds to RNA. The synthetic polymers poly(rG) and poly(rA) bind cro with markedly differing affinities, as well as possibly causing dissimilar quenching of the protein fluorescence. The poly(rA) titration does not saturate, and it is not possible to rule out the eventual attainment of the same maximal quenching seen for poly(rG). Since poly(rG) exists as a multistranded entity at neutral pH (Pochon & Michelson, 1965), whereas poly(rA) is singlestranded at neutral pH (Brahms et al., 1966), it is difficult to attribute the affinity differences to a particular property of the respective polynucleotides. The cro repressor-calf thymus DNA titration shown for comparison indicates that the binding of cro repressor to poly(rG) is much tighter than that seen for doublestranded DNA. Note that saturation with poly(rG) occurs near 60 bases per cro

FIG. 6. Comparison ofcro repressor binding at 4°C to singleand double-stranded DNA at differing ionic< strengths. The 0.05 M-KC1 BB buffer titration was performed with initial era repressor concentration P, = 1.17 @wdimer; the 0.16 M-KC1 BB buffer titration had P,-, = 1% pwdimer. (0) Calf thymus DNA: (0) fd DNA.

F.

BOS(‘HELL1

ET

.-II,

Ease-po1rs (xU5) 0 60-

I I

2 I

3 1

FIG. 7. Binding of cro repressor t,o synthetic RNA. The binding of wo repressor to poly(rC+), (cr.0 concentration, 1% PM) and poly(rA) (O-Orepressor concentration, 1.26 PM) in Wlti M-KC1 BB buffer at 4°C’. The broken line shows t,htl era repressorvalf thymus DSA titrat,ion for comparison.

repressor dimer, while 400 base-pairs (800 bases) of calf thymus DNA per dimer are needed for 99.8% saturation (Boschelli, 1982). (d) cro repressor-h

operator

interaction

The 17 base-pair lambda pseudo-operator synthesized by Kawashima et al. (1977), which differs from 0,3 at three positions (see Fig. 2) was utilized here for the purpose of observing the interaction of cro repressor with a specific site. The pseudooperator was isolated from an EcoRI digestion of plasmid containing a 58 base-pair dimer of this specific sequence flanked with Hind111 linkers and EcoRI linkers, as described in Materials and Methods. Figure 8 shows the spectrum of cro repressor quenched by added pseudo-operator DNA and the return of the original intensity by addition of KC1 to l-6 M. The spectrum shown here is similar to that seen for cro-calf thymus DNA (Fig. 4), and indicates that no unusual quenching behavior occurs upon the cro-pseudo-operator interaction. The titration of cro protein with pseudo-operator is shown in Figure 9. The binding of cro to pseudo-operator is significantly different from that seen for native calf thymus DNA under the same conditions. Since the pseudo-operator is relatively short (68 base-pairs) and comparable to the size of the cro-DNA binding site size (10 base-pairs of 20 nucleotides long at this ionic: strength (Boschelli, 1982)), it is necessary to control for the effect of the length of the fragment on the observed binding (McGhee & von Hippel, 1974; Cantor & Schimmel, 1980). In addition, the pseudo-operator contains EcoRl single-stranded ends that may affect the observed binding. To control for these factors, cro repressor was titrated with a 40 base-pair restriction fragment containing the lac operator flanked by single-stranded EcoRI ends (Sadler et al., 1980). The binding of cro repressor to the Eat operator appears to be weaker than that seen for both calf thymus DNA and the pseudo-operator. To

LBMBDA

PHAGE

cro

REPRESSOR

261

t

ZO-

Q

II 300

I 320

I

I 340

I

I 360

11 300

11 400

Wavelength (nm)

FIG. 8. Spectrum of era repressor and the CTOrepressor-pseudo-operator complex at 0.16 M and at 1.6 M-KC1 BB buffer. The quenched spectrum and the salt-released spectrum have been corrected for t,he inner filter effect and for dilution. (0) era repressor alone. @I6 wKCI; (0) cro represaor-pseudooperator complex, 0.16 M-KU; (A) crc-pseudo-operator in the presence of 1.6 M-KC1 (PO = 1.26 PMdimer) with DNA added to 42 base-pairs/dimer.

Bose-pairs 0 I

2 I

I I

I 0

IO

(x10-? 3 I

4 I

5 I

I

I

30 20 Base-palrs/dlmer

40

I

FIG. 9. Titration of cro with pseudo-operator (0) compared to native calf thymus DNA (unbroken curve), calf thumus DNA fragments (A), and a 40 base-pair Zaeoperator-containing restriction fragment (0). All titrations were performed at 4°C in 0.16 M-KC] BB buffer with an initial cro repressor concentration of I.26 PM

%W

F.

BOS(‘HELLI

E’f’

..(I,.

control for possible anomalous behavior of cro repressor toward lac, operator (O’Neill, 1976), i.e. tight binding with a lower maximal quenching, cro repressor was titrated with double-stranded random sequence DNA fragments 25 to 40 base-pairs long. The binding seen for these fragments is similar to that noted for CTO repressor-lac operator binding. Thus, the different behavior noted for cro repressor-Zac operator binding compared to native calf thymus DNA is due to the length effect, which reduces the effective DNA concentration (i.e. number of sites available for cro repressor binding). While some sort of specific interaction of’cro repressor with pseudo-operator is suggested by the data of Figure 9, the consideration that cro repressor fluorescence might be quenched to a differing extent by pseudo-operator (i.e. saturation is not approached), with cro possessing the same affinity for pseudo-operator as for calf thymus DNA, is not easily ruled out. Because of this, t]he experiment shown in Figure IO was performed to illustrate the specific nature of the cro repressor-pseudo-operator complex. Here, samples of cro repressor-calf thymus DNA and CTOrepressor-pseudo-operator DNA complexes were dissociated by titration with salt (Kowalczykowski et al., 1980). The protein concentration and DNA to protein ratio were approximately equal, as described in the Figure legend. The dramatically dissimilar behavior of the complexes is demonstrated clearly in this experiment. Whereas the cro repressor-non-specific DNA complex is completely dissociated by 0.4 to 0.5 M-KC], at least 1.6 M-KC] is

0

I-O

0.5

I.5

[KC11 (M)

FIG. 10. (‘omparison ofthe cm represaor~alfthymus DNA complex (0) and thewo repressor-pseudooperator complex (0) dissociation by added .salt. The scale on the ordinate represents an increase in the fluorescence of the sample from the initially quenched level. All of the data shown have been corrected for t,he inner filter effect of the DNA as well as for the dilution of the sampleby the added DNA and salt. The cro repressor-c:alf thymus DXA complex was formed in 0.06 M-KU BB buffer, whereas the CTO repressor-pseudo-operator was formed in 0.16 M-KU BB buffer. Both samples had the DNA concentration adjust.ed t.o 4p base-pairs/rlimrr. The initial protein concentration (after the addkion of DNA) before addition of KCI was 1.08 x 10m6 M dimer for the era repressor-pseudo-operator sample and 1.05 x 10M6 M dimer for the cro repressor-calf thymus DNA sample. Roth titrations (duplicates of each) were performed at 4°C. In each case. salt was added to the control cuvett. with CTOrepressor alone as a control for possible salt-dependent fluorescence variations.

LAMBDA

PHA(:E

CTO REPRESSOR

“63

required to completely dissociate the cro repressor-pseudo-operator complex in the sub-micromolar protein concentration range used here. The high ionic strength required suggests that more than electrostatic interactions are involved. These results show that we are looking at a specific complex and not an unusual quenching property of the pseudo-operator.

4. Discussion The results presented above show the usefulness of tyrosine fluorescence for probing nucleic acid interactions. Similar observations have been made for the nonspecific Dh’A-binding gene V product of phage fd (Pretorius et al., 1975) and the N-terminal DNA-binding fragment of Zac repressor (Jovin et al., 1977). The most significant finding is summarized in Figures 9 and 10, where we show that the cro repressor-DNA affinity observed by the quenching of the tyrosyl fluorescence of the cro repressor is dependent on the DNA sequence. Observations at the tyrosine residues will allow direct comparison with nuclear magnetic resonance data, since their signals would fall in a window of the nucleic acid ‘H resonances (K. Arndt, F. Boschelli, Y. Takeda & P. Lu, unpublished results). The titration of era with the lambda pseudo-operator (Fig. 21, a dimer of the left half of 0,3. which differs at three base-pairs from OR3 (Kawashima et al., 1977). illustrates an inherent problem in physical studies of specific protein-DNA interactions at the relatively high concentrations of protein and DNA used in spectroscopic studies. The non-specific DNA interactions of the repressor protein must be considered, even if this affinity is small, since the number of non-specific sites is very large (each base may be the start of a site), while here there are only two specific sites. Thus, the non-specific complex is most likely a major species over much of the binding isotherm. These lower affinity sites will predominate in the early stages of a titration, since they are present in much greater number relative to the higher affinity specific site (Gaugain et aE., 1978; Shafer, 1980). This is the observed behavior for the cro repressor-pseudo-operator interaction under the conditions employed in this study. Similar binding is seen in the early portion of the titration curve for each type of natural DNA employed, since predominantly nonspecific binding occurs in this region of high protein to DNA ratios. After the proportion of cro repressors bound to specific sites becomes appreciable, we see evidence of tighter specific binding in the greater quenching seen for pseudooperator. The quenching observed in the cro repressor-pseudo-operator titration of 70% is very near the total quenching observed for cro repressor-non-specific DNA binding at lower ionic strengths where saturation is approached. The calculated quenching maximum for the cro repressor-native calf thymus DNA titration at 0.16 M ionic strength (Boschelli, 1982) is near 65%. This calculated level is reached at DNA concentrations greater than 20 times the level of the pseudo-operator used here. This suggests that the quenching properties are similar for both specific and non-specific complexes, and that the binding of cro repressor to the pseudo-operator is tighter than that noted for cro repressor-non-specific DNA binding. Quantitation of the binding of cro repressor to pseudo-operator is difficult because of the IO

ai4

F.

BOSCHELLI

ET’

AL.

competing equilibria occurring under these conditions, i.e. the effect of the DNA fragment length on non-specific binding is hard to quantitate. The salt dissociation titration of the cro repressor-non-specific and the cro repressor-specific DNA complexes (Fig. 10) best illustrates the difference in the nature of the two complexes. At the initial ionic strengths employed (0.06 M-K+, 0.16 M-K+, respectively), greater than 95:& of the cro repressor is bound for the nonspecific case, while at least this amount is bound in the specific case, as will be shown below. Since the protein concentration and the DNA to protein ratios are virtually identical, direct comparison of the two complexes is valid. The immediate dissociation of the cro repressor-non-specific complex is consistent with the data presented in the accompanying paper (Boschelli, 1982), where the changed binding site size of cro repressor and a reduced DNA affinity in this ionic strength range causes an effective decrease in the binding of cro to non-specific DNA. cro repressor-pseudo-operator binding is insensitive to increasing ionic strength at this protein concentration (PM) range until a salt concentration of 0.6 M-KC1 is at,tained, in contrast to the non-specific complex, which is completely dissociated above 0.4 MKCl. This is the strongest evidence that cro repressor forms a specific complex with the pseudo-operator, which is also tightly bound by the Xi repressor (Kawashima et al., 1977). At the DNA to protein ratio used for Figure 10, the titration illustrated by Figure 9 appears to show that saturation was not reached. The fact that’ no dissociation was observed (Fig. 10) until the KC1 concentration increased above 0.6 M indicates that the cro repressor was completely bound in 0.16 ~-Kc1 BB buffer. If all of the cro protein were not bound (i.e. the K, value for the cro repressor-pseudo-operator complex formation was not large enough) addition of salt should have resulted in immediate dissociation. We cannot rule out the possibility that cro repressor-pseudo-operator complex formation is ionic strengthindependent. Then the salt dependence of the quenching that we see is a reflection of a change in cro repressor or pseudo-operator structure. For all of the natural double-stranded DNAs studied, no dependence of the magnitude of the quenching or the affinity on the base composition was noted. Plasmid DNA pBR325 (-50% C+C), JZ. Zuteus DNA (70 molY/b G+C), poly[d(A-T)], and calf thymus DNA (39 mol4, C + C), show identical binding and quenching behavior in their interaction with cro when monitored by fluorescence quenching. The lack of a base composition effect on the cro repressor-non-specific DNA interaction is unlike that seen for lnc repressor, where an increased affinity for poly[d(A-T)] versus calf thymus DNA was noted (Riggs et aZ., 1970; Revzin & von Hippel, 1977). cro seems to have a versatile capacity to bind many different forms of nucleic acids, including single-stranded fd DNA as well as synthetic RNA. This is also very different from lac repressor, which does not interact with denatured DNA (Riggs et al., 1970), although binding of that repressor to single-stranded poly(dA) and poly(A) has been reported (Maurizot B Charlier, 1977). The binding of cro repressor to synthetic RNA is of interest, especially in view of the biochemical observations which suggest that cro repressor might be a translation factor as well. During the lytic development of lambda phage, early messenger RNA synthesis continues at a reduced rate even after cro repressor starts and protein synthesis from this exerting its negative effect on transcription,

LAMBDA

I’HAGE

CTO REPRESSOR

265

message (e.g. for phage exonuclease) is completely shut-off. Furthermore, Takeda & Kuwano (1975) reported that cro repressor affects stability and translation of early messenger RNA through a possible interaction of cro repressor with RNA. In view of the observation made here, it will be interesting to test whether cro repressor recognizes any specific sequence in lambda early mRNA, and to test directly whether cro repressor acts as a translation factor. In any event, it is possible that cro repressor performs some other functions in vivo, which are uncharacterized, in addition to its transcriptional control function at the lambda phage operators. The work was supported by National Institutes of Health grants (P.L., research and career award: Y.T., research award; and K.A., University of Pennsylvania training grant). The fluorimeter was purchased with funds from the American Cancer Society. We thank Will McClure of Carnegie-Mellon University, Pittsburgh for the poly[d(A-T)], Tom Roberts of Harvard University, Cambridge for the pTR214 plasmid and Marvin Caruthers of the University of Colorado for a plasmid containing the h pseudo-operator.

REFERENCES Anderson, W. F., Ohlendorf, D. H., Takeda, Y. & Matthews, B. W. (1981). Naturr (London), 299, 754-758. Berkowitz, S. A. & Day, L. A. (1974). Biochemistry, 13, 48254831. Bolivar, F. (1978). Gene, 4, 121-136. Bolivar, F.. Rodriguez, R., Greene, I’., Betlach. M. C., Heyneker. H., Boyer, H. W.. Cross, J. & Falkow, S. (1977). Gene, 2, 95-113. Boschelli, F. (1982). J. Mol. Biol. 162, 267-282. Brahms, *J., Michelson, A. M. & Van Holde, K. E. (1966). J. Mol. Biol. 15, 467-488. Brun, F., Toulme, J.-J. & Helene, C. (1975). Biochemistry, 14, 558-563. Cantor. C. R. & Schimmel, P. R. (1980). Riophysical Chemistry part 111, W. H. Freeman and Co., San Francisco. (‘owgill. R. W. (1976). Biochemical Fluorescence, Concepts, C’.II (Chen & Edelhock, eds). pp. 441486, Marcel Dekker, New York. (‘uatrecasas, P., Fuchs, S. & Anfinsen. C. B. (1967). J. Biol. Chrm. 242, 1541-1547. Echols, H. (1980). In The Molecular Geneticsof Development (Loomis, W. & Leighton, T., eds), pp. I-16, Academic Press, New York. Gaugain, B., Barbet, J., Capelle, N., Rogue, B. P., LePecq, J-B. & Le Bret, M. (1978). Biochemistry, 17, 5078-5088. (Gilbert, W., Maxam, A. & Mirzabekov, A. D. (1976). In Control of Rihosomr Synthesis, Th,e Alfred Bensor Symposium IX (Kjelgaard & Maaloe, eds), pp. 139-148, Minksgaard, Copenhagen. Herskowit,z. 1. & Hagen, D. (1980). Annu. Rec. Genet. 14! 399445. Hsiang. M. W., Cole, R. D., Takeda, Y. & Echols, H (1977). Nature (London), 270,275277. ,Johnson, A.. Meyer, B. J. & Ptashne, M. (1978). Proc. Nat. Awd. Sci., U.S.A. 75, 1783%1787. ,Johnson, A., Meyer, B. J. & Ptashne, M. (1979). Proc. Nat. Acad. Sci., U.S.A. 76, 5061-5065. Johnson. A. D., Poteite, A. R., Lauer, G., Sauer, R. T., Ackers, Q. K. & Ptashne, M. (1981). ,Vaturr (London), 294, 217-223. Jovin. T., Geisler. PI’. & Weber, K. (1977). &atu.re (London). 269, 66s672. Kallai, 0. B., Rosenberg, J. M., Kopka, M. L., Takano, T., Dickerson, R. E., Kam, .J. & Riggs, A. (1980). Biochim. Biophys. Acta, 606, 113-124. Kawashima, E., Gadek, T. & Caruthers, M. H. (1977). Biochemistry, 16, 42094217. Kirby, E. P. (1971). Excited States of Proteins and Nucleic Acids, Plenum Press, New York, Kornberg, A. (1980). DNA Replication, W. H. Freeman and Co., San Francisco. Kowalczykowski, S. C., Lonberg, N., Newport, J. W. & von Hippel, P. H. (1980). J. ~!Iol. Biol. 145. 7%104.

266

F. BOSCHELLl

ET AL.

Lillis, M. T., Nick, H. N., Caruthers, M. H., Sadler, J. D. & Lu, P. (1982). Anal. Biochem. 129, 52-58. Longworth, J. W. (1971). Excited States of Protein,7 an,d Nucleic Acids (Steiner & Weinryb, eds), pp. 319484, Plenum Press, New York. Maurizot, J. C. & Charlier, M. (1977). FEBS Letters, 83, 107-l 10. McGhee, J. D. & von Hippel, P. H. (1974). J. Mol. Biol. 86, 469489. Miller, J. H. (1972). Experiments in Molecular Genetics, Cold Spring Harbor Laboratory, Cold Spring Harbor. O’Neill, M. C. (1976). Nature (London), 260, .5X-%4. Pesce, A. J., Rosen, C.-G. & Pasby, T. L. (1971). Fluorescen,ce Spectroscopy: An In,troduction for Biology and Medicine, Marcel Dekker, New York. Pochon, F. & Michelson, A. M. (1965). Proc. ,Vat. Acad. Sci., IT.S.A. 53, 1425-1430. Pretorius, H., Klein, M. & Day, L. (1975). J. Biol. Chem. 250, 9262-9269. Ptashne, M., Jeffrey, A., ,Johnson, A. D., Maurer, R., Meyer, B. J., Pabo, C. O., Roberts, T. M. & Sauer, R. T. (1980). Cell, 19, l-l 1. Ptashne, M., Johnson, A. D. & Pabo, (1. 0. (1982). Scientijk American, 247(5). 128140. Revzin, A. & von Hippel, P. H. (1977). Biochemistry, 16, 47694776. Riggs, A.D., Suzuki, H. & Bourgeois, S. (1970). J. Mol. Biol. 48, 67-83. Roberts, T. M., Shimatake, H., Brady, C’. & Rosenberg, M. (1977). Nature (London), 270, 274-275. Roberts, T. M., Kacich, R. & Ptashne, M. (1979). Pror. Nat. Acad. Sci., I’.S.A. 76,760-764. Rodriguez, R. L., Bolivar, F., Goodman, H. M., Boyer, H. W. & Betlach, M. C. (1976). In Molecular Mechanisms in the Control of Gen,e Expression (Nierlich, Rutter, & Fox, eds), pp. 47 1477, Academic Press, New York. Sadler, ,J. R., Tecklenberg, M. & Betz, 0. J. (1980). fkne, 8, 279-300. Shafer, R. H. (1980). Biopolymers, 19, 419430. Shen, C.-K. J. & Hearst, J. E. (1976). Proc. nrat. Acad. Sci.. r,J.S.A. 73, 2649-2653. Takeda, Y. (1975). Biochim. Biophys. Acta, 407, 73-82. Takeda, Y. & Kuwano, M. (1975). Biochim. Biophys. Acta, 40’7, 83-90. Takeda, Y., Folkmanis, A. &, Echols, H. (1977). J. Biol. Chem. 252, 61776183 Edited

by M.

Gellert