Vol. 127, No. 2 Printed in U.S.A.

JOURNAL OF BACTERIOLOGY, Aug. 1976, p. 837-847 Copyright © 1976 American Society for Microbiology

Deoxyribonucleic Acid-Binding Studies on the hut Repressor and Mutant Forms of the hut Repressor of Salmonella typhimurium DAVID C. HAGEN' AND BORIS MAGASANIK* Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139

Received for publication 24 May 1976

In Salmonella typhimurium the genes coding for the enzymes of histidine utilization (hut) are clustered in two adjacent operons, hutMIGC and hut(P,R,Q)UH. A single repressor, the product of the C gene, regulates both operons by binding at two operator sites, one near M and one in (P,R,Q). The deoxyribonucleic acid (DNA)-binding activity of the repressor was measured using DNAs containing separate operators. The repressor had greater activity when assayed using DNA containing the operator of the (P,R,Q)UH operon than when assayed using DNA containing the operator of the MIGC operon. The binding to either operator was absent in the presence of the inducer, urocanate. The DNA-binding activities were also determined for two super-repressors. The super-repressors had altered DNA-binding properties, although the self-regulated nature of the repressors complicated the analysis of the results. A purification procedure for the wild-type repressor is presented. The purified repressor was somewhat unstable, and additional experiments using it were not performed. In Salmonella typhimurium, the genes coding for the four enzymes of histidine utilization (hut) are clustered in two adjacent operons, hutMIGC and hut(P,R,Q)UH (1, 9, 15-17). A single repressor regulates both operons. The gene coding for the repressor, C, is itself a member of one of the operons (the MIGC operon), and the repressor thus regulates its own synthesis (5, 15). There are two distinct promoter-operator regions, one in the MIGC operon near M, and one in the (P,R,Q)UH operon in (P,R,Q). Transcription is rightward from these two regions. Urocanate, the first degradation product of histidine and the inducer of the hut operons, prevents the binding of the repressor at either operator. (The hut regulatory mechanism is illustrated in Fig. 1.) M and (P,R,Q) are cis-acting control regions of the two operons (17). Mutations in M result in an increased level of 4-imidazolone-5-propionate amidohydrolase (the product of the I gene), N-formimino-L-glutamate formiminohydrolase) (FGA hydrolase) (the product of the G gene), and repressor in cells grown in the presence of inducer (16, 17). No operator mutations have yet been found in the MIGC operon. (An operator undoubtedly does exist and, most probably, is located in the vicinity of M.) 1 Present address: Department ofBiochemistry, Stanford University, Stanford, Calif. 94305.

(P,R,Q) represents a promoter-operator complex (1, 9, 17). Mutations in P result in a greatly reduced level of histidase (the product of theH gene) and urocanase (the product of the U gene) in cells grown in the presence of inducer. Mutations in R render the synthesis of histidase and urocanase insensitive to catabolite repression. Mutations in Q render the synthesis of histidase and urocanase partially constitutive and insensitive to catabolite repression. There are two indications that the repressor has different affinities for the two operators, the affinity for the operator of the MIGC operon being less than the affinity for the operator of the (P,R,Q)UH operon. First, the enzymes of the MIGC operon are present in high levels in uninduced cells and increase only two- to threefold upon induction; the enzymes of the (P,R,Q)UH operon, on the other hand, are present in low levels in uninduced cells and increase approximately 100-fold upon induction (1, 15, 17). Second, imidazole propionate, a urocanate analogue and gratuitous inducer of the hut operons (1, 9), induces the enzymes of the MIGC operon to the maximum level, but induces the enzymes of the (P,R,Q)UH operon to only one-third of the maximum level (10). In a previous study (5), the hut repressor was assayed in extracts of S. typhimurium. The 837

838

HAGEN AND MAGASANIK M

J. BACTERIOL.

I

6

C

(P, R,Q)

U

H

U

H

Un induced:

(I)

M

(G) Repressor

I

G

C

(PR,Q)

Induced:

I

In hIDI I

G

IPA-iFGA

Repressor Uro

zIzzzI H

U

Glu + HCONH2 -Uro

4-

His

NH3

FiG. 1. Model of the mechanism regulating the synthesis of the enzymes of histidine utilization (hut). The coding for the enzymes of histidine utilization are clustered in two adjacent operons, hutMIGC and hut(P,R,Q)UH (1, 9, 15-17). In the uninduced state, the repressor, the product of the C gene, binds at two operator sites, one site near M in the MIGC operon and the other site near (P,R,Q) in the (P,R,Q)UH operon, and represses hut mRNA synthesis. The enzymes of the MIGC operon, 4-imidazolone-5-propionate amidohydrolase (represented as I) (EC 3.52.7), the product of the I gene, and FGA hydrolase (represented as G) (EC 3.5.3.8), the product of the G gene, are synthesized at relatively high basal levels. The enzymes of the (P,R,Q)UH operon, urocanase (represented as U) (EC 4.2.1.49), the product of the U gene, and L-histidine ammonia lyase (histidase, represented as H) (EC 4.3.1.3), on the other hand, are synthesized at very low basal levels. In the induced state, the repressor, although present at a high level, is fully inactivated by the inducer, utocanate (Uro). All four enzymes are synthesized at a high level, and histidine (His) is degraded via urocanate, imidazolone propionate (IPA), and formimino glutamate (FGA) to glutamate (Glu), ammonia (NH3), and formamide (HCONH3).

genes

basis for the assay was the retention by repressor of labeled deoxyribonucleic acid (DNA) bearing the hut operators on nitrocellulose filters. The DNA binding was shown to be specific for DNA bearing the hut region, the binding was abolished by the presence of urocanate, and the binding was absent if the extracts were of cells bearing a deletion of the entire hut region or of cells bearing an amber mutation in hutC, the gene coding for the repressor. The DNA used for the studies had been extracted from a Xphut transducing phage bearing the entire hut region and thus both operator sites. The DNA-binding experiments reported in the present communication employ DNA extracted from phages bearing single operator sites: phages 4S167 and 4S129 bear only the operator of the MIGC operon, whereas phage 4S473 bears only the operator of the (P,R,Q)UH operon. (The phages are described in more detail in Materials and Methods.) The affinity of the repressor for each operator is thus measured independently here, and the two operators are shown to have different repressor-binding properties.

In addition to wild-type and repressor-negative alleles of the hutC locus, two super-repressor alleles, hutC183 and hutC161, have recently been described (3). Enzyme assays of cells bearing the two super-repressor alleles suggested that the two super-repressors, although similar in several respects (most notably, their insensitivity to inducer), apparently differ from each other and from the wild-type repressor in their affinities for the hut operators. The C183 super-repressor has, apparently, acquired a stronger affinity for the operator of the MIGC operon, whereas the C161 super-repressor has, apparently, acquired a weaker affinity for both operators. In this publication, we describe the assay of an extract of cells bearing the C183 allele and of an extract of cells bearing the C161 allele for binding activity using both DNA of the MIGC operon and DNA of the (P,R,Q)UH operon. So far, DNA-binding assays of the hut repressor (and its mutant forms) have been performed on minimally purified cell extracts (5). The extracts represent all of the proteins of a cell lysate that: (i) are not sedimented by centrifu-

VOL. 127, 1976

hut REPRESSOR OF S. TYPHIMURIUM

839

TABLE 1. Bacterial strains gation at 100 x g, and (ii) bind to cellulose phosphate at low ionic strength and are eluted Genotype' Straina by 0.4 M KCI. Clearly, th,e repressor preparahut+, bio+) (P376) (F'gal+, NE414 hut*162 tions are quite heterogeneous and, as such, can NE446 hut*162 (F'gal+, hutM145, hutR9, bio+) (P376) be used for only semiquantitative determinaNE449 hutM145, hutR9, bio-25 (F'gal+, hutM145, hutR9, bio+) (P376) tions. Contaminating nucleases and nonspecific NE502 hut*162 (F'gal+, hutC161, bio+) (P376) DNA-binding activities prevent the accurate NE556 hut*162 (F'gal+, hutC7, bio+) (P376) measurement of the parameters of the represNE700 hut*162 (F'gal+, hutC183, hutR9, bio+) (P376) of utilization or the interactions sor-operator M65 gall-, gar-, str-, tonGS96 gal,-, galr, str-, ton- (A) the repressor in a cell-free system for the synthesis of hut messenger ribonucleic acid. In the aThe strains with the prefix "NE" are derivatives of S. ratter portion of this publication, a more exten- typhimurium 15-59. Strains M65 and GS96 are Escherichia sive purification of the repressor is reported. coli K-12 strains. M65 was obtained from E. Signer. bThe genetic nomenclature conforms with that of DeThe repressor, purified by these procedures, is et al. (2) except where serial isolation numbers are highly unstable, and further studies using puri- merec unknown or have not been assigned. The abbreviations for fied repressor have not been undertaken. the genetic loci are those of Sanderson (12) and Taylor and Trotter (19). hut*162 represents a deletion of hut-

MATERIALS AND METHODS Chemicals. The sources of the chemicals used in the research presented here have, in most cases, been listed previously (5). Cellulose phosphate, diethylaminoethyl (DEAE)-cellulose, and carboxymethyl cellulose were Whatman P11, DE52, and CM52, respectively, purchased from Reeve Angel. P11 was processed to remove fine particles by settling and decantation, washed with acid (0.5 M HCl) and base (0.5 M NaOH), and equilibrated with buffer before use. DE52 and CM52 were processed to remove fine particles and equilibrated with buffer before use. Sephadex G200 was purchased from Pharmacia and was swollen and otherwise handled according to instructions furnished by the manufac-

MIGC(P,R,Q)UH, bioA, chID, and uvrB. All ofthe S. typhimurium strains are lysogenic for phage P376, and strain GS96 is lysogenic for phage X.

turer.

erons.

Strains of bacteria and bacteriophage. The strains of bacteria referred to in this paper and their genotypes are listed in Table 1. The methods of cultivation and the procedures for strain construction have been previously reported (1, 9, 14, 15, 17). Bacteriophages 4S61, qbS129, 4S167, and 4S473 are Xphut transducing phages isolated by Smith (14) or by the techniques of Smith. 4S61 bears the entire hut region. 4S129 and 4S167 bear only portions of the hut region, MI and MIG, respectively, as a result of incomplete excision of hut material during their initial formation. 4S473 bears the hut region GC(P,R,Q)UH due to a deletion of the region MI from a phage bearing the entire hut region. The extents of the hut regions present were determined for each phage by a marker rescue technique as illustrated in Fig. 2. q4S167 and 4S473 bear common hut material as both rescue hutG markers (Fig. 2), and Hut+ recombinants are produced in a mixed infection. 4S129 and 4S473, on the other hand, probably have no common hut material as there are no hut markers that can be rescued by both (Fig. 2), and the two do not produce Hut+ recombinants in a mixed infection. (The mixed infections were performed on plates by spotting volumes of 10 ,ul of a mixture of the two phages each at concentrations of about 1010 phage per ml on a minimal medium plate containing histidine as the sole source of carbon and nitrogen and spread with about 108 cells of strain GS96, a Hut- strain lysogenic for X. Phages that recombine to produce a complete hut genome can

4S473, or AcI857Sam7, were adsorbed at a multiplicity of infection of 2 phage per cell for 15 min at 37°C. After adsorption of phage, the infected cells were transferred to 100 ml of LPC broth plus 0.4% (wt/vol) glucose and 5 mCi of carrier-free [32P]phosphate and grown for 3 h at 380C. After phage growth, the phage were purified and the DNA was extracted by reported procedures (5), except that the cells were lysed by freezing at -20°C and thawing at 37°C, rather than by treatment with chloroform at 370C. Experiments were performed with [32P]DNA that had been stored at 4°C for 2 to 6 weeks after extraction. Repressor preparations. Repressor preparations were of two types: "small-scale" and "large-scale." The small-scale preparation, made from a culture of 500-ml volume and employing a small (1 by 3 cm) cellulose phosphate column, has been previously described (5). The large-scale preparation was made from a 32-liter culture to allow more extensive purification. In a typical preparation, 2 16-liter volumes of succinate-ammonia minimal medium plus histidine were each inoculated with 200 ml of a stationary-phase culture of strain NE449 grown in the same medium. The cells were grown at 37°C with

transduce the host to the Hut+ phenotype and permit growth.) All of the Xphut transducing phages bear the X alleles c1857, a thermoinducible repressor allele, and Sam7, an amber-suppressible, lysis-defective allele. W857Sam7 was obtained from E. Signer. Media. LPC broth (5), succinate-ammonia minimal media (3, 14), and histidine minimal medium (3, 15) have been described; L-histidine has been added at 0.2% (wt/vol) to the succinate-ammonia minimal medium to obtain induction of the hut opBacteriophage [32P]DNA. Phage labeled with 32P were prepared by lytic infection of strain M65. Cells growing exponentially at 370C in LPC broth plus 0.2% (wt/vol) maltose were centrifuged and suspended at a density of 2 x 109 cells per ml in 1.0 ml of 0.02 M MgSO4. The phage, OS129, OS167,

840

HAGEN AND MAGASANIK

9a/1

J. BACTERIOL.

1M

hut G

II 1 105 1 168

I

16 172 101

A

J b2..

171 174

Lio..

3i

102 167

669

4S 61

C lI P 161

I

U

I

H

I

bio

1 25 A

cM N .

-i

-

i

R

-i-

#5129 #S167 4S473

FIG. 2. Analysis of Xphut transducing phages. The Xphut transducing phages O4S61, .bS129, 4S167, and were analyzed by their ability to 'rescue" the indicated hut alleles. The top line represents the hut genome with the numbers of various alleles written below. (More than one allele number at a given position indicates that the alleles have not yet been separated by mapping techniques.) A bacterial strain bearing any one of these alleles is Hut- in phenotype and cannot grow on a plate containing histidine as the sole source of carbon and nitrogen. The DNA of a phage bearing the corresponding wild-type allele can recombine with the host DNA to convert the cell to the Hut+ phenotype, and growth is observed. (See reference 14 for a more detailed description of this test.) The phage genomes are represented as follows: ( ) genetic material of A with A, J, b2, cII, N, and R as representative genes; (ON) the attachment site of the transducing phage; (C) bacterial genetic material, aligned with the map of the hut region above; and (a) genetic material missing in the phages 4S129, 4S167, and cfS473 in comparison to phage 4S61. The absence of genetic material is due either to incomplete excision of hut material during the original formation of the transducing phage (phages 4S129 and 4sS167) or to a citrate-selected deletion ofpreviously existing genetic material (phage 4S4 73). The attachment site of phage 4S473 is known to be absent by an attachment site assay (14). cII is known to be present in all the phages by their plaque morphologies (7). 4S473

forced aeration until the culture reached a density of approximately 4 x 109 cells per ml. Ice was then added to chill the culture, and all subsequent operations were performed either at 4°C or on ice. The cells were collected by centrifugation, washed with 2 liters of PM buffer (10 mM potassium phosphate, 10 mM 2-mercaptoethanol, pH 7.5), and suspended in 50 ml of PM buffer containing 5% (vol/vol) glycerol. The cells were lysed with 10 5-min sonic treatments with an MSE ultrasonic power unit. Cellular debris and ribosomes were removed by centrifugation for 20 min at 35,000 x g and then for 120 min at 100,000 x g. The final supernatant, approximately 150 ml, was taken as "crude extract" and contained approximately 60 mg of protein per ml. The crude extract was dialyzed against 2 2-liter volumes of PM buffer plus 5% glycerol for a total of 18 h. A precipitate that formed during dialysis was removed by centrifugation. The dialyzed crude extract was applied to a column (2.5 by 25 cm) of cellulose phosphate equilibrated with PM buffer plus 5% glycerol, and 10-ml fractions were collected. After all of the extract had entered the column, additional 10-ml fractions were eluted with 600 ml of PM buffer plus 5% glycerol, and then 8-ml fractions were eluted with 800 ml of a linear, 0.0 to 0.8 M gradient of KCI in PM buffer plus 5% glycerol. Samples were taken from selected fractions for determination of repressor activity, protein concentration (estimated by absorbance [A280] measured with a Zeiss spectrophotometer), and KCI concentration (estimated by conductivity measured with a Radiometer conductivity meter). The nine

fractions having the greatest repressor activity were pooled and dialyzed overnight against 2 liters of PM buffer plus 10% glycerol. The cellulose phosphate-purified repressor was next applied to a column (1.5 by 20 cm) of DEAEcellulose equilibrated with PM buffer plus 10% glycerol, and 10-ml fractions were collected. After the sample, additional 10-ml fractions were eluted with 200 ml of PM buffer plus 10% glycerol, and then 7-ml fractions were eluted with 600 ml of a linear, 0.0 to 0.3 M gradient of KCI in PM buffer plus 10% glycerol. The 13 fractions displaying the greatest repressor activity were pooled. Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis was occasionally performed on repressor preparations to estimate the degree of purity. The SDS gels were prepared and stained for protein by the procedures of Studier and Maizel (18). Molecular weights were estimated by the technique of Shapiro et al. (13), using bovine serum albumin (Miles Laboratories), yeast alcohol dehydrogenase (Worthington), and chicken egg white lysozyme (Worthington) as standards. Protein concentrations were determined by the method of Lowry et al. (8) using bovine serum albumin as a standard. Repressor assay. The hut repressor was assayed by the DNA-binding technique of Riggs et al. (11), modified slightly as reported earlier (5). The binding buffer, BBD, is identical in composition to BB developed by Riggs et al., except that 200 j.g of chicken blood DNA per ml was added and the pH is 7.6. The buffer FB is identical to FB of Riggs et al., except that the pH is 7.6. -

VOL. 127, 1976

RESULTS Separation of the hut operators. Previously reported DNA-binding assays of hut repressor (5) employed DNA extracted from a 4phut transducing phage bearing both hut operator regions. To study the binding of repressor to each operator independently, it was necessary to use DNA extracted from phages in which the two operators had been genetically separated. Phages 4S129 and 4S167, depicted in Fig. 2, lack the hut genetic material GC(PR,Q)UH and C(PR,Q)UH, respectively, and thus serve as sources of DNA carrying the operator of the MIGC operon. Phage 4S473, on the other hand, bears a deletion of the hut material MI and thus serves as a source of DNA carrying the operator of the (PR,Q)UH operon. The construction of 4S129, kS167, and 04473 and the determination of the extents of the respective deletions are described in Materials and Methods and in Fig. 2. Binding of repressor to separate operators. Strain NE449, a strain diploid for hutM145, a fivefold super-promoter mutation in the MIGC operon (Table 1), was selected as a source of repressor, considerably more gS473 DNA [containing the operator of the (P,R,Q)UH operon] was bound than 4S167 DNA (containing the operator of the MIGC operon). The binding of says. In the experiment depicted in Fig. 3, increasing amounts of the extract were assayed for DNA-binding activity using [32P]DNA extracted from Xphut phages k5473 or OS167 or from phage X. In each instance, the assays were performed in the absence and in the presence of the inducer, urocanate. With equal amounts of repressor, considerably more 4OS473 DNA (containing the operator of the (P,R,Q)UH operon) was bound than 4S167 DNA (containing the operator of the MIGC operon). The binding of each Xphut DNA was significantly greater than the binding of X DNA, and the binding was abolished by the presence of urocanate at 1 mM. (In the binding curve for OS167 DNA in the absence of urocanate, the low plateau value and the downward slope at high protein values are probably due to nuclease contamination in these extracts [see reference 5]). In an experiment not presented here, the concentration of urocanate reducing a subsaturating amount of repressor to half of its full activity was found to be approximately 0.1 mM, regardless of which Xphut DNA was employed in the assay. Binding of super-repressor to the hut operators. The isolation of two super-repressor mutants in the hut system has recently been described (3). Genetic and physiological analysis revealed that the two mutations are similar in

hut REPRESSOR OF S. TYPHIMURIUM

841

20 z 0

a')

o 0

10

20 30 PROTEIN (9Lg)

40

50

FIG. 3. Binding of repressor to separate operators. Increasing amounts of repressor, prepared from strain NE449 grown on succinate-ammonia minimal medium plus histidine, were added to 1.00 ml of BBD buffer. The reactions were started with the addition of 26 ng of 48S473 [32P]DNA (X), 22 ng of 4S167 [32P]DNA (U), or 20 ng of X [32P]DNA (A). The reactions were also performed in the presence of 1 mM urocanate and started, as before, with cS473 [32P]DNA (0), c167 [32P]DNA (C), or X [32P]DNA (A). [4S473 DNA bears only the operator of the hut(P,R,Q)UH operon, OS167 DNA bears only the operator of the hutMIGC operon, and A DNA bears neither hut operator.] The mixtures were incubated for 30 min at room temperature (22°C). Triplicate portions of 0.30 ml were filtered through nitrocellulose filters and washed with 0.30 ml of FB buffer. The radioactivity was determined by liquid scintillation counting. Each point represents the average of the three values obtained expressed as a percentage of the input cpm of the portion. Smooth curves were drawn Tby eye" to fit the data points. The input was 3,810 cpm for 0S473 DNA, 3,320 cpm for 4S167 DNA, and 3,290 cpm for XDNA, or 1,140 cpm, 996 cpm, and 987 cpm, respectively, per sample.

several respects: both map in the hutC region, both revert to the HutC- phenotype rather than to the HutC+ phenotype, both result in the uninducible synthesis of the enzymes of the hut operons, and both are dominant over the hutC+ or hutC- alleles in merodiploid strains. The two, however, differ from one another in the degree of repression of hut enzyme synthesis. The C183 super-repressor appears to have acquired a stronger affinity for the operator of the MIGC operon. The C161 super-repressor appears to have acquired a weaker affinity for both operators: a greatly lessened affinity for the operator of the MIGC operon and a slightly lessened affinity for the operator of the (P,R,Q)UH operon. Due to the self-regulatory property of the hut repressor, a super-repressor with increased affinity for the operator of the MIGC operon should result in the synthesis of

842

HAGEN AND MAGASANIK

J. BACTERIOL.

less repressor than that found in uninduced wild-type cells. Similarly, a super-repressor with decreased affinity for the operator of the MIGC operon should result in the synthesis of more repressor than that found in uninduced wild-type cells. In a comparison of the DNAbinding activity of super-repressor strains with wild type there are, therefore, two variables to be considered: the affinity of the repressor for the operator and the amount of repressor present in the extract. Figure 4 represents the results of an experiA

*

0

30

10

20

20

0

2

3

0

0

z 0

0

PROTEIN

4

0

6

(jig)

FIG. 4. Binding of super-repressors to the hut operators. Increasing amounts of repressor prepared from cells grown on succinate-ammonia minimal medium plus histidine were added to 1.00 ml of BBD buffer, and 150 ng of 4S473 [32P]DNA (A) or 4S129 [32P]DNA (B) was added to start the reaction. [4S473 DNA bears only the operator of the hut(P,R,Q)UH operon, and 4S129 DNA bears only the operator of the hutMIGC operon.] The genotypes represented are wild type (strain NE414, 0); superpromoter hutM145 (strain NE446, *); repressornegative hutC7 (strain NE556, *); super-repressor hutC161 (strain NE502, 0); and super-repressor hutC183 (strain NE700, A). The incubation time, filtering, counting, averaging, and curve-fitting were the same as in Fig. 3. The input was 14,800 cpm for 4&S473 DNA and 26,200 cpm for IS129 DNA. The bacterial strains are described in greater detail in Table 1.

ment in which the DNA-binding capacity of two strains bearing the super-repressor mutations was compared with that of wild type, a strain bearing the hutM145 super-promoter allele of the MIGC operon, and a strain bearing the hutC7 repressor-negative allele (an amber mutation). Each strain was grown in succinateammonia medium containing histidine. Extracts were made of each, and the DNA-binding capacity was measured using [32 P]DNA extracted from 4S473, carrying the operator of the (P,R,Q)UH operon (Fig. 4A), and from OS129, carrying the operator of the MIGC operon (Fig. 4B). The differing amounts of DNA-binding activity seen in Fig. 4 are a reflection of either quantitative or qualitative differences in repressor ofeach extract. When tested with either DNA, the extract of strain NE446, the strain bearing hutM145, shows a high level of repressor relative to that of the wild type, strain NE414. The extract of strain NE556, the strain bearing hutC7, shows very little DNA-binding activity. (The DNA binding that is seen with the extract of NE556 is equivalent to the background levels seen with extracts of wild type assayed in the presence of urocanate.) Neither extract of the super-repressor strains, NE502, bearing the C161 allele, and NE700, bearing the C183 allele, shows much DNA-binding activity. In the assay using DNA bearing the operator of the (P,R,Q)UH operon (Fig. 4A), the level of binding that is seen for both extracts, although quite low relative to that of the extract of wild-type cells, is significantly above that of the extract of the strain bearing hutC7, the background level. In the assay using DNA bearing the operator of the MIGC operon (Fig. 4B), the extract of NE700 displays a binding activity that is intermediate between that of the extract of wild-type cells and the background level; the extract of NE502, on the other hand, has no detectable activity above the background level. The binding of both super-repressors to DNA bearing the operator of the (P,R,Q)UH operon was unaffected by urocanate as high as 1 mM. At higher concentrations (up to 30 mM), the binding of both was partially abolished. (Data are not presented.) The effect of urocanate on the binding ofthe super-repressor to DNA bearing the operator of the MIGC operon was not determined. Purification of the wild-type repressor. In DNA-binding experiments reported previously (5) and in DNA-binding experiments reported above, the extracts have been only partially purified. A more complete purification was performed as described in Materials and Methods.

843

VOL. 127, 1976

hut REPRESSOR OF S. TYPHIMURIUM

The fractionation of repressor by successive cellulose phosphate and DEAE-cellulose column chromatography is illustrated in Fig. 5. The repressor was eluted from cellulose phosphate at a KCl concentration of about 0.3 M (Fig. 5A). The peak of repressor activity in the region of fraction 120 is seemingly broad and "flattopped." The shape of this peak is probably due to the saturation of the DNA by repressor in the fractions displaying the greatest activity. The large peak of DNA-binding activity found in the region of fraction 160 and the smaller peak in the region of fraction 145 are judged not to be repressor, as the activities are nonspecific in that they are found in the presence, as well as the absence, of urocanate. In the DEAE-cellulose chromatography the repressor was eluted at a KCl concentration of about 0.1 M (Fig. 5B). Again, probably for the same reason, the peak of repressor activity appears to be broad and "flat-topped." By this stage of the purification, the repressor has been freed from much of the contaminating protein, as can be seen by the great reduction of material absorbing at a wave-length of 280 nm. After DEAE-cellulose chromatography, columns of carboxymethyl cellulose, Sephadex G200, or cellulose phosphate (again) were used.

Each column gave increased purity, as judged by a diminished number of protein bands in SDS-gel electrophoresis, but the total amount of repressor activity recovered was drastically reduced. These repressor preparations, as well as the repressor preparation from DEAE-cellulose, lost their activity upon storage at 4°C with a half-life of about 3 days. Storage in the presence of glycerol (5 to 25%), sulfhydryl reagents (mercaptoethanol, 10 mM, or dithiothreitol, 0.2 mM), a heavy-metal chelating agent (ethylenedinitrotetraacetate, 0.1 mM), or a protease inhibitor (phenylmethylsulfonylfluoride, 0.24 mg/ml) did not prevent the rapid loss of activity. SDS-gel electrophoresis of the purest repressor preparations revealed one major protein band that may be the repressor. The major band comprised approximately 80% of the input protein and migrated to a position corresponding to a subunit molecular weight of 52,000; a few other less substantial bands were also visible. DISCUSSION In the first portion of this paper, we described the measurement of the DNA-binding activity of the wild-type repressor and two super-repressors of the histidine utilization system of S.

E

1.233

300 ~~

X200

i

0.8- 2

B

0

O

x''D xx---/do l

1~?00

0~~~~~~~~~~~~~~~~~~~~~~~~. B

0

200 150 100 FRACTION NUMBER FIG. 5. Purification of the wild-type repressor by (A) cellulose phosphate and (B) DEAE-cellulose column chromatography. The extract was prepared and the column chromatography was performed as described in Materials and Methods. The DNA-binding capacity of selected fractions was determined as follows: portions of 5 p1 (A) or 10 p1 (B) were added to 1.00 ml of BBD buffer or to 1.00 ml of BBD buffer plus 1.0 ,umol of urocanate, and the reactions were started with the addition of 190 ng of e4S473 [32P]DNA. Incubation time, filtering, counting, and averaging were the same as in Fig. 3. The input of 1S473 [32P]DNA was 3,950 cpm Symbols: (0) S473 [32P]DNA bound; (A) or3,200 cpm (B) (1,180 cpm or960 cpm, respectively, per portion)8. (0) c1.473 [32P]DNA bound in the presence of urocanate; ( ) A280; and (x) molarity of KCl. To avoid cluttering, individual data points for A.280 are not shown; the line for A.280 intersects all data points. A280 Of fractions 10 to 31 of the cellulose phosphate column (A) range from 5 to 150 and are beyond the scale of the graph. 50

844

HAGEN AND MAGASANIK

J. BACTERIOL.

typhimurium. The two operator regions of the curves for five extracts, each with a different hut operons were separated genetically, and concentration of repressor (A) and for five difthe DNA binding by the repressors was mea- ferent dissociation constants of the repressorsured using two different DNAs, one bearing operator complex (B). In the case of varying the operator of the (P,R,Q)UH operon and one repressor concentrations, the initial slopes of bearing the operator of the MIGC operon. the binding curves are directly proportional to DNA-binding curves were obtained by adding the repressor concentration of the extract. The increasing amounts of the partially purified re- initial slopes of the binding curves thus can be pressor preparation to fixed amounts of DNA used to compare the concentration of repressor bearing a given operator. Two variables influ- in preparations in which the dissociation conence the shape of such curves: (i) the concentra- stants of the repressors are known to be the tion of repressor in the extracts and (ii) the same. In the case of varying dissociation conaffinity of the repressor for the operator. The stants, the situation is more complex. For dissimplest model for the repressor binding to the sociation constants less than the initial operaoperator can be expressed as: R + 0 = RO, tor concentration (e.g., K = 0.1[0,0,] or K = where R is the free repressor, 0 is the free 0.01[Oto,]), the initial slopes of the binding operator, andRO is the repressor-operator com- curves are identical (and equal to 1). Under plex. At equilibrium, ([R] [O])/[RO] = K, where such circumstances, the repressor binds stoiK is the dissociation constant for the repressor- chiometrically to the operator at low repressor operator complex. The expression ([R] [O])/ concentrations. For dissociation constants [RO] = K can be more usefully expressed as greater than the initial operator concentration {([Rtot] - [RO]) ([°tot] - [RO])}/[RO] = K, (e.g., K = 10[O,o] or K = 10O[O,ot]), the initial where [Rtot] = [R] + [RO] and [Otot] = [0] + slopes of the binding curves are inversely pro[RO]. Theoretical DNA-binding curves can be portional to the binding constant. The initial derived by solving this latter expression for slopes of the binding curves thus can be used to compare the binding affinities of different [RO]: repressor preparations only when the binding constants are known to be significantly [RO] = [Rtot] + [Otot] + K greater than the initial operator concentration N/I([Rtot] + [Otot] + K)2 - (4[RtoJ11t0,01)} and when the concentration of the repressor in the preparations is known to be the same. The 2 fact that both the concentration of repressor in an extract and the affinity of the repressor for Figure 6 shows theoretical DNA-binding the operator can affect the shape of the DNA-

0

I.0

m

30 0 0.62 0~~~~~~~~~~~~~~~~~~~1

EXTRACT ADDED initial concentration of operator, [O,od, the concentration of curves. The FIG. 6. Theoretical DNA-binding the repressor-operator complex, [RO] (given in the ordinate), and the dissociation constant, K, are each expressed in the same arbitrary units of concentration. The units given in the abscissas represent the amount of extract added. In (A) five extracts are represented. One unit of the 100 x extract provides repressor at a concentration oflOO (in the above units of concentration); analogously, 1 unit of the 10 x, 1 x, 0.1 x, and 0.01 x extracts provides repressor at concentrations of 10, 1, 0.1, and 0.01, respectively. In (B) one extract is represented; 1 unit of this extract provides repressor at a concentration of 1. In (A), K = 1. In (B) five dissociation constants are illustrated: K = 0.01, 0.1, 1,10, and 100. In both panels [O,od = 1. The theoretical DNA-binding curves were derived from an equation that is given in the Discussion.

VOL. 127, 1976

hut REPRESSOR OF S. TYPHIMURIUM

binding curve needs to be considered in the analysis of the results presented in this paper. The results of Fig. 3 show that there are two sites, two operators, in the hut region to which the repressor binds. The binding of the two sites is not equivalent; a given amount of extract binds more DNA containing the operator of the (P,R,Q)UH operon than DNA containing the operator of the MIGC operon. This disparity suggests that the two operators have different binding properties. The simplest explanation of the results is that the repressor has a greater affinity for the operator of the (P,R,Q)UH operon than for the operator of the MIGC operon. An alternative explanation is that the repressor that binds to the operator of the (P,R,Q)UH operon is present at a greater concentration in the extract than the repressor that binds to the operator of the MIGC operon. Genetic arguments (1, 15, 16) strongly suggest that there is only one repressor gene, but two different forms of the same repressor, e.g., modified and unmodified or dimers and tetramers, could exist within a cell and within the extracts. That the repressor has less affinity for the operator of the MIGC operon (or that there is a lower concentration of the repressor that binds to this operator) is consistent with the observation that the MIGC operon is less tightly regulated than the (P,R,Q)UH operon (1, 9, 10, 15, 17). The initial slopes of the binding curves of Fig. 3 can be calculated to be 1.90 percentage units per jig of protein for the DNA bearing the operator of the (P,R,Q)UH operon and 0.42 percentage units per ,ig of protein for the DNA bearing the operator of the MIGC operon, or 1.82 and 0.34 percentage units per ,ug, respectively, after subtracting the background level of binding (calculations not presented). Assuming that the same repressor species binds to both operators, and considering the theoretical binding curves discussed above, the ratio of these latter values, 5.3, represents the minimum ratio of K for the operator of the MIGC operon to K for the operator of the (P,R,Q)UH operon. If operator concentrations present in the assay [26 ng/ml or 8.4 x 10-13 M for the operator of the (P,R,Q)UH operon and 22 ng/ ml or 7.1 x 10-'3 M for thetoperator of the MIGC operon] are significantly below the values of K, then the ratio 5.3, or rather (5.3 x 22 ng/ml)/(26 ng/ml) = 4.5, correcting for the unequal concentrations of the operators, is the actual ratio of the K values. {IThe value of K for the operator of the (P,R,Q)UH operon has been judged to be about 10-10 M by an unpublished experiment. In this experiment, a fixed amount of OS473 [32PIDNA was assayed in the

presence of increasing amounts of cold kS473 DNA and a fixed amount of repressor sufficient to half-saturate the [32P]DNA in the absence of competing cold DNA. The value of 10-10 M was roughly estimated from the shape of the resultant curve.} The analysis of the DNA binding by extracts of strains bearing super-repressor alleles is complicated by the self-regulatory nature of the hut repressor. In wild-type cells, the level of repressor is found to vary with the growth conditions (5). Because the repressor is synthesized as a member of the MIGC operon, the repressor level increases upon induction in proportion to the level of hut enzymes controlled by this operon (e.g., FGA hydrolase). A mutant with a super-repressor, which differs from the normal repressor solely by being insensitive to the presence of inducer, will have an uninducible hut operon. The enzymes, as well as the repressor itself, will be found at basal, uninduced levels regardless of the presence or absence of inducer in the growth medium. Super-repressors, in addition to having an altered response to inducer, can also have altered operator affinities (6); such appears to be the case with the hut super-repressors Cl 61 and C183. Previous studies based on enzyme activities suggested that the C161 repressor has lost most of its affinity for the operator of the MIGC operon but has retained some of its affinity for the operator of the (P,R,Q)UH operon (3). In the DNA-binding results presented in Fig. 4, all strains were grown in the presence of histidine. Under these conditions, strain NE502, bearing C161, synthesizes FGA hydrolase at about the same level as wild-type strain NE414. The amount of repressor synthesized by strain NE502 must, therefore, also be approximately equal to the amount synthesized in the wild-type cells. The DNA-binding activities of the C161 repressor, however, are greatly altered compared with those of the wild-type repressor. Binding of the C161 repressor to DNA bearing the operator of the (P,R,Q)UH operon (Fig. 4A) is greatly reduced although significantly above the background level of binding, represented by the binding of the extract of strain NE556, bearing the repressor-negative allele C7, an amber mutation. The binding of the C161 repressor to DNA bearing the operator of the MIGC operon (Fig. 4B) is undetectable over the background level. Thus the DNA-binding studies here are consistent with the conclusions based on enzyme assays. The C183 repressor, in contrast to the C161 repressor, appeared, by the previous studies,

845

846

HAGEN AND MAGASANIK

to have an increased affinity for the operator of the MIGC operon and an unchanged affinity for the operator of the (P,R,Q)UH operon (3). Strain NE700, bearing C183, produces FGA hydrolase at a greatly reduced level compared with that of wild-type cells when growth is in the presence of histidine, as was the case in the experiment described in Fig. 4. The amount of repressor that is synthesized can be assumed to be similarly reduced. The DNAbinding capacities of the extract of NE700 are seen to be less than that of the extract of wildtype cells due, presumably, to this reduced concentration of repressor. A comparison of the DNA-binding activity of strain NE700 for the two different DNAs supports the contention that the affinity of the C183 repressor for the operator of the MIGC operon has increased relative to its affinity for the operator of the (P,R,Q)UH operon. Whereas the DNA-binding activity is reduced in both instances relative to that of wild type, the reduction is much less in the assays using the DNA bearing the operator of the MIGC operon (Fig. 4B) than in the assays using the DNA bearing the operator of the (P,R,Q)UH operon (Fig. 4A). The increased affinity for the operator of the MIGC operon apparently partially counteracts the loss in activity due to the reduced amount of repressor. Again, the conclusions based on the enzyme assays are corroborated by the DNAbinding results. Table 2 shows a semiquantitative comparison of the DNA-binding activities of strains NE414, NE446, NE502, and NE700. These DNA-binding activities were calculated from the data of Fig. 4 by determining the initial slopes of the binding curves and subtracting

J. BACTERIOL.

the corresponding initial slope for the background binding activity. NE446, bearing the super-promoter allele M145, is seen to have an increased DNA-binding activity relative to the wild type, NE414. Significantly, the magnitude of this increase, six- to sevenfold, is the same for both DNAs, 4S473 DNA, bearing the operator of the (P,R,Q)UH operon, and 4S129 DNA, bearing the operator of the MIGC operon, and is approximately the same as the increase in the level of enzyme synthesis that results from this super-promoter allele (16, 17). The extracts of strains NE414 and NE446 both contain the wild-type species of repressor; only the concentration of repressor differs. It is not surprising, therefore, to see that the ratio of the binding activity with respect to kS473 DNA to the binding activity with respect to OS129 DNA is approximately the same for the two strains: 3.5 for NE414, and 4.1 for NE446. This ratio, as discussed earlier, is inversely proportional to the corresponding ratio of the dissociation constants of the two DNAs, assuming that the DNA concentrations in the assays is significantly below the values of the dissociation constants. [In the assays of Fig. 4, the DNA concentration used for both DNAs, 150 ng/ml or 4.8 x 10-12 M, is probably well below the values for the dissociation constants. The smaller of the two dissociation constants, that for the operator of the (P,R,Q)UH operon, has been estimated by the unpublished experiment previously mentioned to be on the order of 10-1o M.] The ratio of 4.5, determined in the preceding discussion for an extract of strain NE449, also bearing the wild-type repressor, is in reasonable agreement with the ratios for NE414 and NE446. In the case of the strains bearing the super-repressor alleles, the reduction in the TABLE 2. DNA-binding activities ofcells bearing the DNA-binding activities is readily apparent. wild-type repressor and super-repressor alleles Despite the low activities, alterations in the DNA-binding relative affinities for the two operators can be seen from the ratios. The C161 repressor of Relevant activities" Ratio Strain genotypea strain NE502 has an increased affinity for the qS473 DNA 4S129 DNA of the (P,R,Q)UH operon (4S473 operator 3.5 0.63 0.18 NE414 hutC+ DNA) relative to the operator of the MIGC 4.1 4.47 1.08 NE446 hutM145, operon (OS129 DNA): the ratio of the DNAhutC+ >7 NE502 hutCl61 0.07

Deoxyribonucleic acid-binding studies on the hut repressor and mutant forms of the hut repressor of Salmonella typhimurium.

Vol. 127, No. 2 Printed in U.S.A. JOURNAL OF BACTERIOLOGY, Aug. 1976, p. 837-847 Copyright © 1976 American Society for Microbiology Deoxyribonucleic...
2MB Sizes 0 Downloads 0 Views