The functional of atetrameric chimaera are …Proc. Nati. Acad.Sci. USA Vol. 73, No. 10, pp. 3529-3533,October1976 Biochemistry Thefunctionalrepressorpartsofatetramericlacrepressor-f3

Proc. Nati. Acad. Sci. USAVol. 73, No. 10, pp. 3529-3533, October 1976Biochemistry

The functional repressor parts of a tetrameric lac repressor-f3-galactosidase chimaera are organized as dimers

(protein chimaera/electron microscopy/subunit crosslinking)

JURGEN KANIA AND DENNIS T. BROWNInstitut fur Genetik der Universitit zu K6ln, Weyertal 121, 5000 Cologne, Federal Republic of Germany

Communicated by Thomas F. Anderson, August 9,1976

ABSTRACT The chimaeric protein repressor-galactosidase,in which fully active Jac repressor is covalently linked to theactive enzyme #-galactosidase, was used as a system for probingthe quaternary structure of Jac repressor. Electron micrographsrevealed repressor-galactosidase to be a tetrameric aggregate.When Jac repressor, alone, was crosslinked with dimethyl su-berimidate, dimers, trimers, tetramers, and oligomers of theprotein subunit were produced, whereas crosslinking of thetetrameric repressor-galactosidase resulted in the productionof only dimers of the chimaera. Treatment of Jac repressor withiodine resulted in the formation of protein dimers; the sameresult was obtained with repressor galactosidase. After limitedproteolysis of Jac repressor, no crosslinking was obtained aftertreatment with dimethyl suberimidate, whereas iodine stillproduced a covalent linkage. These results are interpreted asevidence that the Jac repressor parts of the tetrameric repres-sor-galactosidase-chimaera are organized as dimers on the tet-rameric-ftgalactosidase core. Because this chimaera has beenpreviously shown to have normal repressor activity [B. Muller-Hill and J. Kania (1974) Nature, 249,561-63,l we conclude thatJac repressor still is biologically active as a dimeric aggre-gate.

lac repressor, a tetramer of identical subunits, specifically in-teracts with its DNA target, the lac operator, thus preventingthe expression of the structural genes of the lac operon. Geneticand biochemical analysis point to the NH2-terminus of lac re-pressor as the region which interacts specifically with DNA [seereview by Muller-Hill (1)]. The primary structure of the lacrepressor protein (2) and the lac operator (3) have been deter-mined. Little, however, is known about the tertiary and qua-ternary structure of wac repressor since crystals large enoughfor single crystal x-ray diffraction analysis have not yet beenobtained. Electron micrographs of the protein in solution exist(4, 5). The most detailed structural analysis was carried out bySteitz et al. (6), who proposed a model of quaternary structurefrom electron micrographs and powder x-ray diffractionanalysis of microcrystals. For this model, lac repressor subunitsare arranged in an exact or quasi D2 symmetry. As a conclusionto their investigations Steitz et al. proposed a model for re-pressor-operator interaction, where it was suggested that twooperator binding sites exist per tetramer, in the case of D2symmetry. This model, if correct, might imply that lac repressordimers are capable of recognizing the lac operator sequence,if somehow the repressor tetramer could be dissociated into theappropriate stable dimers. Repressor-galactosidase* is a fu-sion-protein, in which fully active lac repressor is covalentlyattached by its COOH-terminus to the NH2-terminus of active

Abbreviation: NaDodSO4, sodium dodecyl sulfate.* We will use the term "repressor-galactosidase" for the protein chi-maera purified from Escherichia coil strain iql (71-56-14), as de-scribed (7). The terms "repressor-part" and "galactosidase-part" areused for those amino acid sequences in repressor-galactosidase, whichare coded by the lac repressor gene and by the fl-galactosidase gene,respectively.

3529

fl-galactosidase (7), and is an appropriate system for the furtherexamination of models of the quaternary structure of lac re-pressor. Because the covalent linkage of lac repressor to fl-ga-lactosidase limits the theoretically possible arrangements of (acrepressor subunits, investigation on the geometry of this hybridmolecule enables us to test the model of Steitz et al. As a resultof the application of electron microscopy and crosslinkingagents on repressor-galactosidase, we propose that lac repressoris able to recognize lac operator as a dimer, fixed via itsCOOH-terminus to a soluble matrix, i.e., fl-galactosidase.

MATERIALS AND METHODSPurification of Repressor-Galactosidase. Repressor-ga-

lactosidase protein was purified as described previously (7), withthe modification that chromatography on DEAE-cellulose(Biorad) was performed after the phosphocellulose step.DEAE-cellulose was equilibrated with buffer D [0.01 M Trisat pH 7.27 (at 22°), 0.01 M MgAc, 0.01 M 2-mercaptoethanol,and 0.1 M NaCI]; the pure protein eluted at 0.21 M NaCl, whena linear gradient of 0.1 M NaCI to 0.5 M NaCl in buffer appliedto the column. According to sodium dodecyl sulfate (NaDod-SO4) gels, the protein was about 90% pure.

Electron Microscopy. A concentrated solution (15 mg/ml)of repressor-galactosidase was diluted 1000-fold into 1% NH4Acat pH 7.2 with or without 0.35% glutaraldehyde. After 5 minat room temperature, the protein was mounted on copper grids(400 mesh) having a thin carbon film. The carbon film had beenmade hydrophilic by glow discharging prior to use. Themounted specimens were washed with distilledtwater, andnegatively stained with a solution of 1.5%-uranyl-acetate indistilled water. Micrographs were made in a Siemens 101electron microscope, the magnification of which was calibratedwith a carbon grating replica, having 2160 lines/mm (ErnestFullam Co. no. 1002).

Crosslinking with Dimethyl Suberimidate and DimethylAdipinimidate. Diimidates were obtained from Pierce.Crosslinking was performed essentially as described by Daviesand Stark (8). Repressor-galactosidase was crosslinked at 3-4mg/ml of protein in the reaction mixture, lac repressor at anapproximate concentration of 6 mg/ml. The pH was 8.5 andthe diimidate concentration was 0.7 mg/ml. The concentrationsof protein and crosslinker were not critical within a factor oftwo. More concentrated repressor-galactosidase solutions onlyresulted in the appearance of numerous multimeres apparentlydue to intermolecular crosslinking. The reaction was stoppedafter 30 min at room temperature by making the solution- 1%of NaDodSO4 and 2-mercaptoethanol and incubating for 20min at 650. The presence of either 10 mM isopropyl-thio-fl-D-galactoside or 10 mM o-nitrophenyl-,6-rfucoside had noeffect on the yield of crosslinked material.

Crosslinking with 12/KI. lodination was performed byslightly modifying the conditions described by Fanning (9).

Dow

nloa

ded

by g

uest

on

May

22,

202

0

3530 Biochemistry: Kania and Brown

>z..-,.'9K>7x~~~~~~~-W--r.

FIG. 1. Electron micrographs of negatively stained repressor-galactosidase. (A) Low magnification field showing a number of tetramericcomplexes. A number of the tetramers have a prominent centrally located hole (1). In other tetramers the hole is less easily seen (2). (B) Sixsimilarly oriented tetramers superimposed and printed as a single image. The four subunits of the tetramer are easily seen. Magnification (A)X161,000 and (B) X737,000.

Proteins were dialyzed against TMS buffer [0.01 M Tris at pH7.5 (at 220), 0.01 M MgAc, 0.2 M KCl, and 0.1 mM EDTA]without 2-mercaptoethanol at 40. To 100 ,l of both lac repressor(8 mg/ml) and repressor-galactosidase (10 mg/ml), we added10 Al from a stock iodine solution (0.01 M 12, 0.04 M KI) madeup in distilled water at 40. After 1 min, the reaction was stoppedby adding 10 ,l of 2-mercaptoethanol. Doubling the reactiontime or the amount of iodine did not result in an increase incrosslinked material, neithr was the concentration of the proteincritical. Isopropyl-thio-f3-D-galactoside or o-nitrophenyl-f3-D-fucoside at 10-2 M had no effect. Some protein precipitatedduring the reaction, but could be dissolved by adding Na-DodSO4 to a final concentration of 1%. After incubation for 20min at 650, each gel received 50-100 ,ug of protein.NaDodSO4/Acrylamide Gel Electrophoresis. NaDodSO4

gels were made by standard procedures (10). In the case of re-pressor-galactosidase, 4% acrylamide gels were used and runfor 8.5 hr at 8 mA per tube. Crosslinked lac repressor was runon 5% acrylamide for 4.5 hr (if treated with proteases, then thetime was reduced to 3-4 hr) at 8 mA per tube.

RESULTSElectron Microscopy of Repressor-Galactosidase. We have

previously shown that repressor-galactosidase is a stable andhomogeneous oligomer which is similar in its sedimentationbehavior to fl-galactosidase itself (16S) (7). As the knowledgeof the number of subunits is essential for further investigationson the geometry of this protein, we attempted to establish the

state of aggregation of repressor-galactosidase subunits directlyby electron microscopy. By electron microscopy, the repres-sor-galactosidase molecule was found to be square in shape.Frequently, the complex could be clearly seen to be composedof four subunits (Fig. 1). The tetramers were of constant sizeand possessed a centrally located hole which was somewhatvariable in diameter. The complexes shown in Fig. 1 were ob-tained from glutaraldehyde treated preparations. Similar resultswere obtained with unfixed material. The tetrameric config-uration of the complex was enhanced by superimposing severalsimilarly oriented molecules and printing them photographi-cally as a single image (Fig. 1). For this purpose the tetramershaving the centrally located hole were chosen. The width of thetetramers along the edge was 217 A and measured on the di-agonal (corner to corner) 251 A. Each of the four subunitscomprising the tetramer were 82 A in cross section. We havealso examined the structure of purified j3-galactosidase byelectron microscopy (data not shown). We found fl-galacto-sidase was not as well preserved in preparations for electronmicroscopy as repressor-galactosidase. f3-Galactosidase wasfound to be tetrameric in structure as has been previously re-ported (11). The measured length of the edge of the (3-galac-tosidase tetramer was found to be slightly larger than that ofrepressor-galactosidase at 220 A. The larger size was probablydue to measuring structures in less well-preserved specimens.Lac repressor alone, although tetrameric in structure (12), isless than one half the size of repressor-galactosidase (4, 5, 13).Thus, the size and shape of the repressor-galactosidase complex,

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

May

22,

202

0

Proc. Nati. Acad. Sci. USA 73 (1976) 3531

a b c d e f

W lo

a b c dIV

a b ca b c d e f g

*N,

di- o *-

mono -1m SN If

di bdi -i A

fs

.I.mono -iP la 0 I mono -)P- I

d i

nono -o'AL

A B C D

FIG. 2. NaDodSO4/gel electrophoresis of crosslinked lac repressor and crosslinked repressor-galactosidase. The positions of monomers(mono), dimers (di), trimers (tri), and tetramers (tetra) of the protein subunit are indicated by arrows. (A) lac repressor crosslinked with dimethylsuberimidate and with I2/K1 (a) untreated; (b) treated with chymotrypsin before incubation with dimethyl suberimidate; (c) crosslinked withdimethyl suberimidate; (d) same as in (c) except that repressor-galactosidase was added as marker; (e) untreated lac repressor, bovine serumalbumin, and repressor-galactosidase added as marker proteins; (f) crosslinked with iodine. (B) Repressor-galactosidase crosslinked with dimethylsuberimidate (a) untreated; (b) crosslinked with dimethyl suberimidate; (c) same as in (b) except that myosin was added as marker; (d) untreatedrepressor-galactosidase and myosin. (C) Repressor-galactosidase crosslinked with I2/K1 (a) untreated; (b) crosslinked with iodine; (c) sameas in (b) except that myosin was added. (D) lac repressor crosslinked with 12/KI after limited proteolysis (a) untreated; (b) untreated core afterincubation with chymotrypsin; (c) same as in (b), treated with iodine; (d) untreated after incubation with trypsin and chymotrypsin; (e) sameas in (d), treated with iodine; (f) uncleaved lac repressor after treatment with iodine; (g) untreated lac repressor, bovine serum albumin, andrepressor-galactosidase added as marker proteins.

as revealed in the electron microscope, suggests that the ga-lactosidase part of the repressor-galactosidase has the same

structural organization as galactosidase alone and furtherimplies that the glactosidase moiety of the repressor-galacto-sidase complex forms the structural core of the tetramer (seeFig. 3). If the smaller repressor part of the repressor-galacto-sidase formed the structural core of the aggregate, then one

would have expected to see "dumbbell" shaped structuresconsisting of large dimers of the high-molecular-weight ga-lactosidase part connected by the smaller repressor portions ofthe complex. Such structures were never found.

Crosslinking with Diimidoesters. If lac repressor is treatedwith dimethyl suberimidate or dimethyl adipinimidate, severalbands are found on NaDodSO4 gels which can be interpretedas monomer, dimer, two trimers, one or more tetramers, andoligomers (Fig. 2A). Repressor-galactosidase, however, if treatedwith the same diimidoesters in a parallel experiment, can becrosslinked only to dimers (Fig. 2B). 3-Galactosidase alonecannot be crosslinked with these diimidoesters under the con-

ditions described (see also ref. 14). This suggests that only twolac repressor-parts, in the repressor-galactosidase aggregate,are in the configuration of operator-binding lac repressor whichallows them to be crosslinked. This result would be expectedif the repressor parts of the complex were peripherally locatedas dimers on a fl-galactosidase core which has the tetramericorganization described in the preceding section (see Fig. 3).

Crosslinking in the Presence of Iodine. Iodine proved tobe an excellent agent for crosslinking lac repressor, by applyingthe procedure described in Materials and Methods. After io-dination of lac repressor, two main bands can be resolved onNaDodSO4 gels at the positions of dimers and also some minor

bands at trimer and tetramer positions (Fig. 2A and D). If fl-galactosidase is treated with iodine in the same way, faint bandsat dimer, trimer, and tetramer positions were visible on Na-DodSO4 gels, only if the gel was heavily overloaded. However,treatment of repressor-galactosidase with iodine resulted in a

strong band at the dimer position on NaDodSO4 gels (Fig. 2C).Thus, iodine as well as diimidoesters crosslink repressor-ga-lactosidase only to dimers, whereas lac repressor is crosslinkedprimarily into dimers by iodine and to dimers, trimers, andtetramers by diimidoesters.

Crosslinking of lac Repressor after Limited Proteolysis.Platt et al. have shown that if lac repressor is incubated withtrypsin and chymotrypsin under native conditions, peptides are

released only from the termini of the molecule thus leaving anunnicked, aggregated core protein (15). We applied this tech-nique to distinguish between the crosslinking sites of dimethylsuberimidate and I2/KI. After treatment of lac repressor withchymotrypsin (3% wt/wt, 1 hr at 370), no crosslinking couldbe achieved by dimethyl suberimidate within the chymotrypticcore (Fig. 2A). This implies that the crosslinking sites (lysineresidues) are located at the termini of the polypeptide chain oflac repressor. Excessive treatment of lac repressor with chy-motrypsin and a mixture of trypsin and chymotrypsin (each 2%wt/wt, 30 min at 370, and then by incubation at 40 for 12 hr)did not result in an abolition of crosslinking by iodine (Fig. 2D).Gel (c) of Fig. 2 reveals that after digestion the iodinated sampleconsisted of a mixture of polypeptide chains of different lengths,due to the prolonged incubation with the protease. The upperband presumably consists of dimers which lack the NH2-ter-mini; the crosslinked material from the lower band should befree of both the NH2- and COOH-termini. Since trypsin si-

tetra -j

tri -Op

Biochemistry: Kania and Brown

Dow

nloa

ded

by g

uest

on

May

22,

202

0

3532 Biochemistry: Kania and Brown

multaneously acts at both termini of the lac repressor, a mixtureof trypsin and chymotrypsin leads to the double-band patternfound on gels (d) and (e). Precise molecular weight determi-nations of the iodinated material are difficult from gels (c) and(e) since iodinated, but not crosslinked polypeptides, seem tomigrate faster than iodinated, while the uncleaved, uncross-linked but iodinated subunit shows the same mobility asnoniodinated lac repressor [gel (f)]. From these results weconclude that iodine catalyzes the crosslinking of the proteaseresistant core fragments of lac repressor at or near theiraggregation sites.

DISCUSSIONProtein chemical investigations have revealed that no more thanfive amino acids are missing at the COOH-terminus of lac re-pressor in repressor-galactosidase, because the second to lasttryptic peptide can be identified after limited tryptic digestionof the protein (16). Because the subunit is calculated to be about155,000 daltons (7) (with RNA polymerase as a standard onNaDodSO4 gels), about 70 amino acids are missing from thef3-galactosidase portion. Because the increase in the molecularmass which results from the attachment of lac repressor doesnot result in a faster sedimentation relative to f3-galactosidase,we propose that the structure of repressor-galactosidase is el-lipsoidal rather than spherical.By electron microscopy, we were able to establish the active

repressor-galactosidase to be a tetramer with morphology andsize similar to 03-galactosidase alone. #-Galactosidase and lacrepressor are also both stable tetramers (12, 17). Langley et al.have shown that defective galactosidase protein from the de-letion mutant strain M15 of E. coli, which lacks residues 11-41of intact ,B-galactosidase, is a dimer under native conditions (18).This observation excludes C4 symmetry for the ,B-galactosidasetetramer, since in the case of such a symmetrical arrangement,dissociation by mutation would lead exclusively to monomers[see review by Matthews and Bernhard (19)]. This must also betrue for lac repressor, because substitutions of amino acids inthe presumed aggregation sites of lac repressor lead to disso-ciation of the repressor tetramer into a mixture of dimers and

V monomers, as found recently by Schmitz et al. (20). Our results,obtained by iodine-catalysed crosslinking of lac repressor,support this assumption. Actually, this is not surprising sinceall other tetrameric proteins of known structure are of D2symmetry (19). Therefore, in the repressor-galactosidase tet-ramer, lac repressor parts cannot be aggregated into a tetramerif f,-galactosidase parts are tetrameric with D2 symmetry andvice versa. Consequently, one of the two protein-parts mustretain its biological activity in the repressor-galactosidase tet-ramer in the state of a dimer. The crosslinking patterns of lacrepressor and repressor-galactosidase produced by diimidoestersand iodine in this study support the hypothesis that it is the lacrepressor portion of the repressor-galactosidase complex whichis organized into dimers.

As mentioned above, Schmitz et al. have found mutationsin the repressor gene which result in the dissociation of lac re-pressor into dimers and monomers which are no longer capableof operator binding (20). Such mutations map in two regions,one around tryptophan 209 and phenylalanine 215, the otherone around tyrosine 260 and tyrosine 269 in the repressor se-quence. It is at these positions that Schmitz et al. propose sub-unit-aggregation sites. It is attractive to propose that theseparticular aromatic amino acids are involved in the iodine-catalyzed crosslinking reaction. Crosslinking of repressor-ga-lactosidase catalyzed by iodine revealed that the aggregationsites necessary for operator-binding (as proposed by Schmitz

lac Repressor

Repressor-Ga Iactosidase

FIG. 3. Schematic drawing of the proposed model of repressor-galactosidase organization. The fl-galactosidase-parts are aggregatedas tetramers in a D2 symmetry as in the wild type molecule. Each ofthose two fl-galactosidase parts, which have point symmetry to thex-axis, carry the repressor-parts which, as shown by the crosslinkingdata, must be aggregated as dimers. The native lac repressor tetrameris also aggregated in a D2 symmetry; this is schematically indicatedin the above drawing. By comparing the corresponding dimers in thesetwo models, it becomes evident that the DNA binding site in the re-pressor tetramere is made by those two subunits, which have pointsymmetry to the x-axis. Consequently, the NH2-terminus of the lacrepressor polypeptide chain, which contains the DNA-binding site,must be located at the free narrow-end of each subunit in the abovemodel.

et al.) are present in repressor-galactosidase. This implies thatthe two repressor parts are properly aggregated for operatorbinding.Our conclusions, from this study, are presented in model form

in Fig. 3 and may be summarized as follows: the fl-galactosidaseparts in repressor-galactosidase are aggregated in a D2 sym-metry. Each of those f3-galactosidase parts, which are similarlyoriented, carry at their NH2-termini lac repressor-parts which,aggregated as dimers, are able to bind to lac operator DNA (7).Similar conclusions have been made based on other types ofevidence by Geisler and Weber (21) and by J. Miwa and J. R.Sadler (personal communication). This model supports theimplication of the model of Steitzet al. that lac repressor dimersare capable of recognizing the lac operator and is also consistentwith some aspects of the model suggested by Adler et al. (22),who proposed that the DNA binding site is situated on a pro-trusion at the NH2-terminus of the polypeptide chain. Themodel presented here predicts as yet undiscovered pointmutations in the i gene of lac repressor, in which the substitutionof an amino acid in an aggregation site different from thoseaggregation sites described by Schmitz et al. would lead to adissociation of the lac repressor tetramer into stable dimersexhibiting all biologicl activities of native lac repressor.

Since repressor-galactosidase does not aggregate through itsrepressor-parts into long chains, one might conclude theCOOH-terminus of lac repressor is involved in the aggregationof repressor dimers to tetramers. This is supported by the factthat Ll-lac repressor, in which the COOH-terminus of lac re-pressor is missing, is still able to repress at low levels in vow andforms exclusively dimers [Miller et al. (23)]. This third predictedaggregation site could be the hydrophobic cluster existingaround proline 307.

Proc. Natl. Acad. Sci. USA 73 (1976)

Dow

nloa

ded

by g

uest

on

May

22,

202

0

Proc. Natl. Acad. Sci. USA 73 (1976) 3533

We thank B. Muller-Hill for his stimulating interest and helpfuldiscussion throughout this work. Thanks are also due to T. G. Fanningand K. Beyreuther for advice and suggestions and for providing pu-rified lac repressor; and to K. Weber for providing myosin. This in-vestigation was supported by Deutsche Forschungsgemeinschaftthrough SFB 74 by grants to B.M.-H. and D.T.B.

1. Muller-Hill, B. (1975) Prog. Biophys. Mol. Biol. 30,227-252.2. Beyreuther, K., Adler, K., Geisler, N. & Klemm. A. (1973) Proc.

Nati. Acad. Sci. USA 70,3576-580.3. Gilbert, W. & Maxam, A. (1973) Proc. Nati. Acad. Sci. USA 70,

3581-584.4. Oshima, Y., Horiuchi, T. & Yanagida, M. (1975) J. Mol. Biol. 91,

515-519.5. Abermann, R., Bahl, C. P., Marians, K. J., Salpeter, M. M. & Wu,

R. (1976) J. Mol. Biol. 100, 109-114.6. Steitz, T. A., Richmond, T. J., Wise, D. & Engelman, D. (1974)

Proc. Natl. Acad. Sci. USA 71,593-597.7. Muller-Hill, B. & Kania, J. (1974) Nature 249,561-563.8. Davies, G. E. & Stark, G. R. (1970) Proc. Natl. Acad. Sci. USA

66,651-656.9. Fanning, T. G. (1975) Biochemistry 14,2512-2520.

10. Weber, K. & Osborn, M. (1969) J. Biol. Chem. 244, 4406-4412.

11. Zabin, I. (1963) Cold Spring Harbor Symp. Quant. Biol. 28,431-435.

12. Muller-Hill, B., Beyreuther, K. & Gilbert, W. (1971) in Methodsin Enzymology, eds. Grossman, L. & Moldave, K. (Academic

Press, New York), Vol. 21, pp. 483-487.13. Bourgeois, S. & Pfahl, M. (1976) Advances in Protein Chemistry

eds. Anfinsen, C. B., Edsall, J. T. & Richards, F. M. (AcademicPress, New York), Vol. 30, pp. 1-99.

14. Mfillner, H., Hucho, F. & Sund, H. (1975) Hoppe-Seylers Z.Physiol. Chem. 356, 256.

15. Platt, T., Files, J. G. & Weber, K. (1973) J. Blol. Chem. 248,110-121.

16. Kania, J., Ruth, C. & Mfiller-Hill, B. (1975) Hoppe-Seylers Z.Physlol. Chem. 356,243.

17. Craven, G., Steers, E. & Anfinsen, C. B. (1965) J. Blol. Chem. 240,2468-2477.

18. Langley, K. E., Villarejo, M. R., Fowler, A. V., Zamenhof, P. J.& Zabin, I. (1975) Proc. Natl. Acad. Scd. USA 72, 1254-1257.

19. Matthews, B. W. & Bernhard, S. A. (1973) "Structure and sym-metry of oligomeric enzymes," in Annual Review of Biophysicsand Bioengineering, eds. Mullins, L. J., Hagins, W. A. & Stryer,L. (Annual Reviews Inc., Palo Alto, Calif.), Vol. 2.

20. Schmitz, A., Schmeissner, U., Miller, J. H. & Lu, P. (1976) J. Biol.Chem. 251,3359-3366.

21. Geisler, N. & Weber, K. (1976) Proc. Nati. Acad. Sci. USA 73,3103-3106.

22. Adler, K., Beyreuther, K., Fanning; E., Geisler, N., Gronenborn,B., Klemm, A., Mfiller-Hill, B., Pfahl, M. & Schmitz, A. (1972)Nature 237,322-327.

23. Miller, J. H., Platt, T. & Weber, K. (1970) in The Lactose Operon,eds. Beckwith, J. R. & Zipser, D. (Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.), pp. 343-351.

Biochemistry: Kania, and Brown

Dow

nloa

ded

by g

uest

on

May

22,

202

0