Reprinted from the ProceEpines or THe NATIONAL ACADEMY OF SCIENCES Vol. 54, No. 4, pp. 1174-1181. October, 1965. COMPARISON OF 8-GALACTOSIDASES FROM NORMAL (i-otzt) AND OPERATOR CONSTITUTIVE (-o’et) STRAINS OF E. COLI By Epwarp Sresrs, Jr., Gary R. Craven, anp Curisrian B. ANFINSEN LABORATORY OF CHEMICAL BIOLOGY, NATIONAL INSTITUTE OF ARTHRITIS AND METABOLIC DISEASES, NATIONAL INSTITUTES OF HEALTH, BETHESDA, MARYLAND Communicated August 26, 1965 Three genetic regions have been identified within the lactose (Lac) locus of Lscherichia coli which specify the structure of the enzymes §-galactosidase, galacto- sidase permease, and thiogalactoside transacetylase.! The expression of the three loci has been shown to be regulated by the product of another genetic region (designated 2) which is presumed to function by coordinately repressing the synthesis of these enzymes. The postulation of a repressor substance specified by the 7 locus leads to the prediction of a site of action for such a repressor. This site has been termed the “operator.” The isolation of certain mutants which conform to the properties expected for such a controlling element supports the concept of a specific operator region. Oniginally, Jacob and Monod characterized two different classes of mutants which defined the operator region. The first type was isolated from homozygous inducible diploids (¢+/7+). These mutants had the properties of being constitutive even in the presence of a competent repressor (7+). Thus, such constitutive organisms (termed 0°) were dominant over the 7+ genotype. Such dominance is in contrast to the usual constitutive state (¢~) which is recessive in the 7+/2z- diploid. Strains possessing the second class of mutants (0°) used to define the operator region are unable to synthesize any of the three enzymes specified by the Lac region. How- ever, these mutants mapped within a region contiguous with the extreme end of the 6-galactosidase-determining structural gene (z). Recent studies by Beckwith have shown that the 0° cluster is not part of the operator locus and that these mutants should be defined as members of the class of polar mutations.2. The o° mutants re- main as the only group which define the operator. Jacob, Ulmann, and Monod? have recently utilized the techniques of Beckwith to isolate a new series of operator constitutive mutants believed to represent deletions of varying sizes in the operator region (see Materials and Methods). They have examined the crude extracts from certain of these mutants in an effort to determine whether or not the galactosidase activities have properties similar to those of the wild-type protein. In all cases, the galactosidases, examined in crude extracts, were shown to have similar pH and temperature sensitivities. This paper reports the results of a continuation of these studies involving enzymic, physical, and chemical comparisons of the 8-galactosidases isolated from an operator constitutive organism (i-o°zt+) with the “wild-type” enzyme from a regulator constitutive organism (t-ote2*), Materials and Methods.—Selection of operator constitutive strain: Two considerations were in- volved in the selection of the operator constitutive strain used here. First, it was important that the operator constitutive mutant used represent as large a deletion of the operator gene as pos- sible. Second, the deletion mutation should include one of the terminal sections of the operator region. With regard to the first consideration Jacob et al.3 concluded that all operator constitutive mutations represented deletions of varying sizes. Based on this conclusion it was further inferred 1174 Vou. 54,1965 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN 1175 that the degree of constitutivity is proportional to the LACTOSE OPERON - E. cole length of the deletion, that is, the greater the deletion of the operator function the greater the constitutivity.4 To Revressor Sotactosidase Tronsave‘ylase Operator Permease satisfy both criteria a strain was chosen in which both the operator function and the repressor function were com- pletely deleted (Fig. 1). The mutant selected, designated se 0¢7, has been shown to be a complete constitutive (i.e., ab- ~ ot , solutely no increased synthesis of ; B-galactosidase in the Fr « 1.—Schematic genetic map presence of inducer) even in the diploid 7+o+z~/Fi-og7z*. of the Lac operon showing the pre- The mutant was obtained through the generosity of Drs. sumed deletion in strain 0%;. F. Jacob and J. Monod. Purification: The 6-galactosidases employed in this study were isolated from the regulator constitutive strain of Escherichia coli Ki2, 3300, and the operator constitutive strain Ky, 3000-067. The growth conditions for each strain were identical and are described in detail elsewhere.* The procedure for the purification of each enzyme, also described in detail in a previous publica- tion, was briefly as follows. The lysate from approximately 1 kg of cells (wet weight) was frac- tionated with ammonium sulfate and the 0-40% fraction was passed through Sephadex G-200. The enzymically active fraction from G-200 was further purified by ion exchange chromatography using DEAE Sephadex A-50. The appropriate fraction from DEAE chromatography, homo- geneous by polyacrylamide gel electrophoresis, was pooled and either stored under ammonium sulfate at 4° or carboxymethylated and stored in the inactive form. Carbozymethylation: The protein, dissolved in an 8 M solution of recrystallized urea, was incu- bated at 38° with 6-mercaptoethanol (0.1 M) for 4 hr and subsequently alkylated for 20 min at room temperature with a 5-fold molar excess of sodium iodoacetate prepared from recrystallized iodoacetic acid.’ The pH during the alkylation reaction was maintained at 8.2 by the addition of 6 N NaOH. The reduced-carboxymethylated product was dialyzed against 0.2 Af NHsHCO,, pH 8.2, and thoroughly lyophilized to remove the volatile salt. Trypsin digestion: The reduced-carboxymethylated product was dissolved at a concentration of 5-10 mg per ml and dialyzed against 0.2 @ NH,HCOs, pH 8.2. Diisopropylflucrophosphate- treated trypsin, 2 X crystallized (Worthington Biochem. Corp.), was added in a protein to enzyme ratio of 50:1 and the digestion allowed to proceed for 4 hr at 38°. The hydrolysates were lyo- philized. Reaction with cyanogen bromide: The reduced-carboxymethylated protein was dissolved in 70% formic acid at a concentration of approximately 5 mg per ml. A 50-fold excess of cyanogen bromide (Eastman Organic Chemicals) was added, the flask tightly stoppered, and the reaction® allowed to proceed at room temperature overnight (approximately 18 hr). The resulting mixture was then frozen and lyophilized. Electrophoresis: Polyacrylamide gel electrophoresis was performed with 7.5% ‘standard gel’ according to the manual supplied by the Canal Industrial Corporation (Bethesda, Md.). Free electrophoresis was carried out utilizing a Perkin-Elmer model 38 electrophoresis apparatus. Equal quantities of native enzyme and enzyme from o¢7 cells were mixed to give a final protein concentration of 14 mg per ml. The mixtures were dialyzed against buffers’? at pH 6.0, 7.5, and 9.0. Electrophoresis was carried out at 0° for periods of time ranging from 4 to 6 hr. Ultracentrifugal studies: A Spinco model E analytical ultracentrifuge equipped with both the standard schlieren and Raleigh interference optical systems was employed for all runs. Velocity measurements were carried out as described by Schachman on both enzyme preparations and on 1:1 mixtures of the two preparations. Molecular weights were determined utilizing the high-speed equilibrium technique of Yphantis.® All runs employed 3-mm columns of solution at speeds in the range of 7447 rpm and 8227 rpm, utilizing the AnJ rotor and protein concentrations of 0.1-0.3 mg/ml. All plate measurements were carried out using a Bausch and Lomb comparator. Amino acid analyses: The lyophilized, carboxymethylated protein was suspended in 5.7 HCl in thick-walled hydrolysis tubes. The tubes were evacuated, sealed, and heated at 110° for 20 hr. Following hydrolysis, the samples were taken to dryness and chromatographed on a Beckman model 120B amino acid analyzer according to the technique of Spackman, Moore, and Stein.® End-group analysis: Both the native and carboxymethylated galactosidases prepared from + 66 1176 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN Proc. N. A. 8. the 3300 and og7 strains were examined for NH;-terminal end groups by the fluorodinitrobenzene (FDNB) procedure.” Reaction with FDNB was carried out in 0.1 M borate buffer containing 8 M urea at pH 9.0, 40°, with shaking for 4 hr. Urea was added to help ensure availability of susceptible groups to the reagent. Control experiments with oxidized ribonuclease, both in the presence and absence of 8 M urea, yielded 0.8 moles of NH-terminal bis-DNP-lysine. The dinitrophenylated proteins were precipitated, washed, and hydrolyzed as described previously by Biserte.!! Ether-soluble DNP-amino acids were separated and determined as described pre- viously.!? High-voltage electrophoresis!! of the aqueous phase after ether extraction of the hydrolysate failed to reveal the presence of water-soluble DNP-amino acids. Immunochemical comparison: Antisera were prepared as outlined by Balbinder and Preer!3 using both the og7 and 3300 enzymes as antigens. Double diffusion utilizing agar plates was carried out according to the technique of Ouchterlony as described by Campbell e¢ al.!4 Quantitation of immunological activity was accomplished by the double-diffusion technique of Preer.6 The equivalence concentration of each enzyme was compared utilizing different antisera. Additional comparisons were carried out utilizing the “‘profile’’ technique of Oudin!¢ which has been modified by Finger and Heller.” In this analysis, the two enzyme preparations are compared qualitatively and quantitatively by reacting them against several different antisera in Preer double-diffusion tubes. Ultraviolet spectra: Ultraviolet spectral measurements were performed using a Cary model 15 recording spectrophotometer. Calculations for tyrosine to tryptophan ratios were done on spectral data obtained on carboxymethylated 8-galactosidase dissolved in 0.1 N NaOH according to Beaven and Holiday.'® Renaturation: The kinetics of return of activity following inactivation in 8.0 M urea were followed after a 1:10 dilution of the urea-protein solution (1.0 mg/ml protein, 8 M urea, Tris 0.05 M, NaCl 0.1 M, mercaptoethanol 0.1 Mf, pH 7.5) into buffer lacking urea at room tempera- ture. Aliquots were taken at various intervals for assay. Results.— Behavior during purification: The course of the purification procedure, summarized in the Materials and Methods section, was identical in all respects for both enzymes. Purification of the two enzymes was conducted simultaneously, and the homogeneity of each product was established independently. The final prod- ucts appeared to be homogeneous by polyacrylamide gel electrophoresis and yielded single boundaries in the analytical ultracentrifuge and during free electrophoresis. Free electrophoresis was carried out on 1:1 mixtures of the two enzyme preparations over a pH range of 6 through 9 (e.g., Fig. 2). Polyacrylamide gel electrophoresis patterns for the crude extracts, as well as for the final purified products (Fig. 3), were identical for both enzymes. The specific activities (840,000 units per mg) of the 3300 and 07 final products were equal within the precision of the measurements. pH 74 PERCENT ACTIVITY 20 7 fo} 20 40 60 a0 100. 120 Fig. 2.—Electro- TIME- MINUTES phoretic pattern “ : Fig. 4.—Kinetics of inactivation obtained at pH 7.4 Fie. 3.—Polyacrylamide gel of the two purified galactosidases at of 1:1 mixture of 8- _ electrophoresis patterns of crude 55-57° and 60-62°. @, 8-galactosi- galactosidases from extracts and purified 6-galactosi- dase from strain 3300; O, 8-galactosi- strains 0j7and 3300. —dases of strains o§7 and 3300. dase from strain o§7, > DESCENDING 312 mins. Vor. 54, 1965 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN 1177 Furthermore, no significant differences in K,,, determined using the Lineweaver- Burk plot. were observed between the two enzyme preparations. Sensitivity to heatand urea: The kinetics of inactivation of the two enzyme prepa- rations were examined at three different temperatures. No significant difference was observed (Fig. 4) between the kinetics of inactivation of the two enzyme prepa- rations at 55-57° and 60-62°C. Renaturation from 8 M urea: Zipser has shown that 6-galactosidase can be renatured to a large extent from 8 M urea upon removal of the urea by simple dialysis.1° The kinetics of this renaturation process can be conveniently studied utilizing direct dilution of the 8 M urea solution into buffer at room temperature containing Mg++, Nat, and mercaptoethanol (buffer B®). These studies have shown that the renaturation from 8 M urea represents a refolding of the extended chain (mol wt = 135,000) and subsequent aggregation into the native enzyme, having a molecular weight of 540,000. It would appear, therefore, that a study of the kinetics of renaturation from 8 M urea might be a sensitive measure for compari- son of the complex forces involved in the determination of the tertiary and quater- nary structure of the enzyme. In support of this, it should be pointed out that the heat-labile enzyme produced by a reversion mutant of the polar mutation, o*, (0% rev-1) is completely unable to renature from 8 M urea under conditions which result in the successful renaturation of the galactosidase of opm eee strain 3300. ash : Figure 5 shows the kinetics of reactivation when the two enzymes are diluted from 8 M urea. This test reveals no difference between the two enzyme prep- arations. Molecular weights: The molecular weight of native 8-galactosidase reported by Sund and Weber” has been corroborated by Craven et al.5 Using the high- speed equilibrium technique of Yphantis,® the mole- cular weights of the og7 and 3300 preparations were found to be 550,000 and 530,000, respectively. These values are in marked contrast to that determined for Wo. 4. J the enzyme produced by the CRM mutant, 3310, Tue shares which was found to have a molecular weight of tion of Bente ign ee 235,000.21 This type of comparison is, of course, strains 0%; and 3300 following di- only of value in detecting gross differences in quater- lution from 8 M urea solutions. nary structure such as that produced in the case of the CRM mutant, 3310. Immunochemical comparison: Antisera were prepared against the purified 3300 and og¢7 preparations as described in the Materials and Methods section. In both the homologous and heterologous reaction systems, the two enzyme preparations showed complete identity utilizing the double-diffusion technique of Ouchterlony (Fig. 6). For quantitative tests, the double-diffusion technique of Preer was employed to determine the equivalence concentration of each antigen preparation against a standard antiserum prepared against @-galactosidase of strain 3300 (i-otzt). In such tests, precise quantitative results may be obtained for any antiserum (or antigen) which shows some degree of cross reactivity with an antigen (or anti- PERCENT ORIGINAL SPECIFIC ACTIVITY 1178 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN Proc. N. A.8. OPTICAL DENSITY 1 1 L 280 200 300 320 340 WAVELENGTH, me Fig. 6.—Ouehterlony double-diffusion patterns obtained Fic. 7.—Ultraviolet with @-galactosidases from strains 0%; and 3300 using as spectra of §-galactosidases antisera in the center wells, anti-of; (87) and anti-3300 from strains of; and 3300. (810). 087, ----; 3300, ——. TABLE 1 serum). Table 1 presents the results QuaNTITATIVE PReciPitIN ANALYSIS* OF with two different antisera, comparing B-GALACTOSIDASES OF Srrains 3300 AND 067 the two antigens, og and 3300. No Antigen ug/ml Serum Average . . . ‘ 1220 significant antigenic differences were 3300 (270 tz) 1350 1290 observed between these proteinsin their TED ability to form precipitates with appar- ofl i-o%z*) 1360 1350 ently homologous antibodies. 1400 A third immunological procedure conc eapeions were performed at saulvatence. Values (termed “profile analysis”) which can pared against enzyme from strain 3300). yield a critical comparison of proteins is the technique of Oudin, which has been usefully modified by Finger and Heller.” The latter modification utilizes the Preer double diffusion technique for comparative analysis. It is possible, using the profile technique, to compare the reactions of a given antigen with a large number of antisera. The position of the precipitin band in the Preer tube reflects the level of “recognition” of the antigen-antibody reaction. Each antigen will thus give a characteristic “profile” against a series of antisera to which any other antigen prep- aration can be compared. The purified preparations of 8-galactosidase from the 3300 and og¢7 strains were tested by this method against five antisera to the 3300 protein. These sera had been prepared independently using five different rabbits. The relative reactivities of the two antigens to the five antibody preparations were constant, within the precision of the measurements, against all five sera. It is apparent that, by immunological methods, no differences are evident between the enzymes studied here. Ultraviolet spectra: The absorption characteristics of the two enzymes, og7 and 3300, were examined by scanning the two preparations after adjustment to approxi- mately equal concentrations. Figure 7 is a direct tracing of such an analysis where consecutive scans were superimposed upon each other. An analysis of these curves shows no differences and indicates identical maximum extinctions at 282 my, and 280/260 mp ratios. Furthermore, there is no apparent difference between the ratios of tyrosine to tryptophan in the two proteins as shown by the spectra of the car- boxymethylated proteins in 0.1 N NaOH. Vou. 54,1965 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN 1179 Amino acid analysis and peptide mapping: The amino acid compositions of the two enzymes were determined after acid hydrolysis using a Beckman-Spinco auto- matic amino acid analyzer. The composition of the protein from the 3300 strain has been reported previously® and the analyses of the 6-galactosidase from og; yielded identical results within the error of the method (approx. 3%). Fic. 8.—-Peptide maps obtained from trypsin digests of the galactosidases isolated from the constitutive strain (3300) and an operator constitutive mutant strain (0§7). Phenol red (PR) was added as a marker in the chromatographic dimension. A. : Boe A more direct chemical comparison is afforded ‘ : by the analysis of the peptide patterns obtained on tryptic digests of the proteins.2?_ Figure 8 shows the patterns which were obtained using digests of the carboxymethylated enzymes from strains 3300 and ogy. The minor differences ob- served were also seen in duplicate maps of a single protein preparation. CNBr fragmentation: A second direct. com- parison of primary structures was possible by analysis of the products produced by the specific action of cyanogen bromide. Cyanogen bro- mide has been demonstrated to cleave, specific- ally, peptide linkages following methionine res- idues.6 This technique has been used exten- sively with the “standard” protein of strain 3300.77 Figure 9 presents the results of such let]. 1-44... cleavage of both the reduced alkylated ,- Fic. 9.—Polyacrylamide gel patterns : . : c of CNBr digests of carboxymethylated galactosidase chains of strains 3300 and o¢7. 8 -galactosidases of strains oj; and 3300. ‘Two different cross-linked polyacrylamide gels Pattern A, 5 -linked gel; patter oy: BT 10 ozo rose Hinks Beh Pamern were utilized to compare the CNBr fragments 1180 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN Proc. N. A. 8. of the two proteins. The patterns have shown no reproducible differences. Should a significant deletion of information exist in the o° strain, such a genetic property might be expected to result in a significant alteration (or disappearance) of at least one of the resulting polypeptide fragments. End-group analysis: A deletion of the genetic region specifying the NH2-terminal structure of 8-galactosidase would increase the probability of detection of a new NH,-terminal end group. Applications of the FDNB procedure resulted in the recovery, from both the og7 and 3300 strains, of approximately 0.6 moles of threo- nine per 135,000 mol wt. Discussion.—As part of their original generalized concept of the Lac operon, Jacob and Monod? postulated three functions for the operator locus. It was suggested that the operator site was responsible not only for the “recognition” of the repressor substance and the initiation of message translation, but also for the specification of the proximal portion of the @-galactosidase protein itself. This latter function was deduced from the properties of a series of mutations of the polar variety, termed 0° mutations, which mapped approximately in the same area as the operator constitu- tive mutations. Since several reversions of oo mutants produced temperature sensi- tive @-galactosidase, it was inferred that the operator region specified the terminal structure of the enzyme itself. However, the experiments by Beckwith? demon- strated that the o° mutations are not within the operator locus but represent polar mutations of the @-galactosidase structural cistron. Thus, there remained no direct evidence that the operator specifies any portion of the 6-galactosidase molecule. The present studies, involving chemical, physical, and immunological examination of the §-galactosidases produced by the wild-type and the operator constitutive mutant, 0¢7, are consistent with the complete identity of these proteins. Perhaps most significant are the chemical results which include identical peptide maps, polyacrylamide gel patterns of cyanogen bromide fragments, and end-group data. The latter determinations involve some uncertainty because of the difficulty of detecting such NH-terminal residues as proline, glycine, histidine, and the acetyl- amino acids in general. Nevertheless, the same terminal residues were observed for both the normal and operator constitutive materials, and no new groups were de- tected in the latter protein.24 Since the deletion in og7 is believed to extend from within the operator through the z locus (see Materials and Methods), the terminal region of the operator must be deleted. If this characterization of the og7 mutation is correct, the present end-group data suggests that the proximal portion of the Z gene has not been altered and that the operator is not part of this structural gene. Summary.—The properties of purified 6-galactosidases from a normal constitutive strain of FB. coli K12 (3300, t-o+z+) and from an operator constitutive mutant (0¢7, t—-o°zt) have been compared using physical, chemical, and biological methods. No detectable differences were found between the two protein preparations. It is concluded that the operator locus does not specify any part of the structure of the 8-galactosidase molecule. ! Jacob, F., and J. Monod, J. Mol. Biol., 3, 318 (1961). ? Beckwith, J. R., J. Mol. Biol., 8, 427 (1964). 3 Jacob, F., A. Ulmann, and J. Monod, Compt. Rend., 258, 3125 (1964). ‘ Jacob, F., and J. Monod, Biochem. Biophys. Res. Commun., 18, 693 (1965). ' Craven, G. R., E. Steers, Jr., and C. B. Anfinsen, J. Biol. Chem., 240, 2468 (1965). VoL. 54,1965 BIOCHEMISTRY: STEERS, CRAVEN, AND ANFINSEN 1181 6 Gross, E., and B. Witkop, J. Biol. Chem., 237, 1856 (1962). 7 Miller, G. L., and R. H. Golder, Arch. Biochem. Biophys., 29, 420 (1950). 8 Yphantis, D. A., Biochemistry, 3, 297 (1964). $Spackman, I. H., 8. Moore, and W. H. Stein, Anal. Chem., 30, 1190 (1958). 10 Sanger, F., Biochem. J., 39, 507 (1945). u Biserte, G., in Chromatographic Reviews, ed. M. Lederer (Amsterdam: Elsevier, 1960), vol. 2. ” Redfield, R., and C. B. Anfinsen, J. Biol. Chem., 221, 385 (1956). 13 Balbinder, E., and J. R. Preer, Jr., J. Gen. Microbiol., 21, 156 (1959). 14 Campbell, D. H., J. 8S. Garvey, N. E. Cremer, and D. H. Sussdorf, Methods in Immunology (New York: W. A. Benjamin, Inc., 1963). 6 Preer, J. R., Jr., J. Immunol., 77, 52 (1956). 6 Oudin, J., in Methods in Medical Research, ed. A. C. Corcoran (Chicago: Year Book Pub- lishers, 1952), vol. 5, p. 335. 1 Finger, I., and C. Heller, J. Afol. Biol., 6, 190 (1963). 8 Beaven, G. H., and E. R. Holiday, Advan. Protein Chen., 7, 319 (1952). 19 Zipser, D., J. Mol. Brol., 7, 113 (1963). » Sund, H., and K. Weber, Biochem. Z., 337, 24 (1963). 4 Steers, E., Jr., unpublished experiments. » Katz, A. M., W. J. Dreyer, and C. B. Anfinsen, J. Biol. Chem., 234, 2897 (1954). 23 Steers, E., Jr., G. R. Craven, C. B. Anfinsen, and J. L. Bethune, /. Biol. Chem., 240, 2478 (1965). *4 Tt is assumed in this discussion that the direction of reading in the lactose operon is such that the NH.-terminal end of the protein chain is specified by the end of the galactosidase structural gene proximal to the operator locus as has been shown to be the case with tryptophan synthetase.® * Yanofsky, C., B. C. Carlton, J. Guest, D. R. Helinski, and U. Henning, these PRocEEDINGs, 51, 266 (1964).