Proc. Natl. Acad. Sci. USA Vol. 77, No. 6, pp. 3346-3350, June 1980 Biochemistry

The L-arabinose : Binding sites of the regulatory proteins and a mechanism of positive and negative regulation (DNase protection/cooperativity/araC protein/cyclic AMP receptor protein/RNA polymerase) SHARON OGDEN, DENNIS HAGGERTY*, CAROL M. STONER, DAVID KOLODRUBETZ, AND ROBERT SCHLEIF Department of Biochemistry, Brandeis University, Waltham, Massachusetts 02254 Communicated by Charles Yanofsky, March 28, 1980

ABSTRACT The locations of DNA binding by the proteins quence determination methodology of Maxam and Gilbert (14) involved with positive and negative regulation of with the DNase I protection method of Galas and Schmitz (15) initiation of the L-arabinose operon in Escherichia coli have to determine the sites of the araC been determined by the DNase I protection method. Two cyclic DNA-binding protein and AMP receptor protein sites were found, at positions -78 to -107 CRP. These studies, data from a physiological experiment, and and -121 to -146, an araCprotein-arabinose binding site was consideration of published data suggest the following mecha- found at position -40 to -78, and an araC protein-fucose nism for regulation of the operon: under induction conditions, binding site was found at position -106 to -144. These loca- CRP is able to bind to a DNA site to be called CRPBAD. Its tions, combined with in vivo data on induction of the two div- presence on the DNA stabilizes the binding of araC protein in ergently oriented arabinose promoters, suggest the following its inducing conformation adjacent to CRP on a DNA regulatory mechanism: induction of the araBAD operon occurs sequence, when cyclic AMP receptor protein, araC protein, and RNA araL. The presence of araC protein on the DNA then stimulates polymerase are all present and able to bind to DNA. Negative the binding of RNA polymerase to the DNA. This induction regulation is accomplished by the repressing form of araC scenario can be blocked by the presence of araC protein in a protein binding to a site in the regulatory region such that it repressing conformation bound to a DNA site upstream from, simultaneously blocks access of cyclic AMP receptor protein but partly obscuring, the CRPBAD site. At the same time and to two sites on the DNA, one site of which serves each of the two from the same location, araC protein also partially represses promoters. Thus, from a single operator site, the negative reg- access of CRP to a second ulator represses the two outwardly oriented ara promoters. This Pc by blocking site, CRPC, located regulatory mechanism explains the known positive and negative on the opposite side of araC from CRPBAD. regulatory properties of the ara promoters. MATERIALS AND METHODS Studies on the L-arabinose operon of Escherichia coli have es- DNA fragments for sequence determination and protection tablished the following important facts on regulation of the were obtained from plasmid pMB9ara440 (8) by EcoRI en- divergently oriented ara promoters PC and PBAD (see Fig. 1). donuclease digestion and polyethylene glycol precipitation (16). The activity of both promoters is stimulated by the cyclic AMP CRP was purified as described (12), and araC protein was pu- (cAMP) receptor protein (CRP) in the presence of cyclic AMP rified as described (11), with the substitution of a hydroxy- (1-4). The PBAD is positively regulated by araC apatite column for the DEAE-Sephadex column. It was loaded protein in the presence of arabinose-i.e., the protein is re- in 0.014 M potassium phosphate buffer at pH 6.9, 10% (vol/vol) quired for activity of the promoter (1, 2, 5). Under noninducing glycerol, 0.01 M L-arabinose, 1 mM EDTA, and 1 mM di- conditions, the araC protein instead acts negatively to repress thiothreitol and eluted in the same buffer with the phosphate both the PC and PBAD promoters (1-3, 6). At least part of the raised to 0.13 M. DNA site necessary for repression of PBAD lies upstream from DNA fragments were 32P end-labeled by alkaline phospha- all of the sites necessary for induction of PBAD, as shown by the tase treatment and T4 polynucleotide kinase as described by existence of deletions entering the ara regulatory region from Maxam and Gilbert (14) and in protocols provided by these the Pc side that abolish repression of PBAD without affecting authors. Size standards for the DNAse-protected DNA frag- induction of PBAD (6, 7). ments were G>A sequencing fragments in which the methyl- The requisite components are now available for in vitro ation was performed either in the buffer described in the figure studies of the mechanism of regulation of the ara operon. The legends or in the DNase I digestion buffer. regulatory region DNA has been isolated, and its nucleotide DNase I protection was performed as described (15), with sequence has been determined (8, 9) and found to contain el- the exceptions noted here and in the figure legends. DNase I ements similar to the RNA polymerase-binding sites seen in digestion buffer is 50 mM NaCI/20 mM Tris-HCI, pH 8.0/3 other E. coli promoters at about 10 and 35 bases before the start mM MgCl2/0.1 mM dithioerythritol/0.2 mM cAMP. DNA was sites of transcription (8). The sequence also contains several incubated at 37°C for 10 min in DNase I digestion buffer, then stretches that resemble the CRP-binding site in the gal operon proteins were added and the incubation was continued 10 min (10). Also available are the proteins involved in the regulation to permit the proteins to bind. The temperature was shifted to of the ara operon: araC protein (11), CRP (12), and RNA 20°C and DNase I was added to a final concentration of 0.13 polymerase (13). ,ug/ml. After the addition of stop buffer, the DNA was prepared In the work reported here, we have utilized the DNA se- and separated by electrophoresis. The strain used in induction measurements of Pc was con- The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "ad- Abbreviations: cAMP, cyclic AMP; CRP, cAMP receptor protein. vertisement" in accordance with 18 U. S. C. §1734 solely to indicate * Present address: Industrial Products Group, Miles Laboratories, Inc., this fact. P.O. Box 932, Elkhart, IN 40515 3346 Downloaded by guest on September 28, 2021 Biochemistry: Ogden et al. Proc. Natl. Acad. Sci. USA 77 (1980) 3347 Regulatory region araC araB araA araD Pc PBAD

a~~~~ a i I -1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I I I I I aI'

1* -160 -140 -120 -100 GACCAAAGCCATGACAAAAACGCGTAACAAAAGTG7'CTATAA7'CACGGCAGAAAAG7'CCACATTGATTATT7'GCACGGCG C7GGTTTCGGTAC7'G'1717-1 TTGTTTTCACAGATATTAG7GCCG7CTTT7'CAGGTGTAACTAATAAACG7'GCCGC

-80 -60 -40 -20 1 TCACACTTTGCTATGCCATAGCATTTrATCCATAAGATTAGCGGATCCTACCTGACGCTTTTTATCGCAACTC7cTACTGTTTCTCCATACCCGTTTTT AGTGTGAAACGATACGGTATCGTAAAAATAGGTATTCTAATCGCCTAGGATGGACTGCGAAAAA TAGCGTTGAGAGATGACAAAGAGGTATrGGCAAAAA FIG. 1. Structure of the L-arabinose operon, with origins and directions of transcription of the , nucleotide sequence, and position notation.

structed by mating the araB episome, F'74 (17), into a leu- -170. The presence of cAMP was irrelevant for C protein- cine-requiring derivative of the pc-f3-galactosidase fusion strain binding in the absence of CRP, but necessary for CRP binding. FC-17 (3). The resulting strain, F'C + B-A +D + /araC-lacZ At less than saturating concentrations of araC protein, the B +A + D-, grows on L-arabinose. In the measurements shown, presence of CRP-cAMP enhanced the apparent binding of cells were grown overnight in M1O medium (18) containing araC protein in the presence of arabinose (data not shown). thiamine at 0.001%, glycerol at 0.2%, and Difco casamino acids Induction of the Pc Promoter. The binding sites determined at 1%. The cells were diluted 1:1000 and grown to 1 X 108 cells above suggested the mechanism of positive and negative reg- per ml, at which time the experiment began. At the indicated ulation of PBAD to be discussed in the next section. Additionally, times, 0.35-ml samples were taken into 0.5 ml of assay buffer the partial overlapping of araO with both CRPBAD and CRPC at 00C containing 200 ,gg of chloramphenicol per ml, and the suggested that the repression complex of araC protein with 0-galactosidase was measured as described (18). DNA might repress both the PBAD and PC promoters; therefore, induction by arabinose might both induce PBAD and derepress RESULTS Pc. Several years ago Casadaban addressed this very question Location of the CRPC and CRPBAD Binding Sites. Binding after noting that both ara promoters were in the same region of a protein to a specific DNA sequence can block or enhance (3). He constructed a pc-f3-galactosidase fusion strain to enable cleavage of the phosphate backbone by DNase I (15). The lo- convenient assay of Pc activity, but his measurements failed cation of the sites with altered cleavage frequencies is then to reveal any arabinose-mediated activation of Pc. On the basis determined by electrophoresis of the DNase I digestion prod- of the DNase I protection findings, we have reexamined this ucts: cleavage positions are determined by parallel electro- question. phoresis of the same DNA that has been prepared in accordance Our initial experiments confirmed Casadaban's findings that with the DNA sequence determination methods developed by the steady-state expression of Pc in an araC +B +A +D + strain Maxam and Gilbert. In such experiments with DNase I, we have is nearly the same in the presence or absence of arabinose. We located two CRP-binding regions: one between 20 and 50 base therefore explored the response of Pc immediately after the pairs from pc, position -121 to -146, and one from 78 to 107 addition of arabinose. Fig. 3 shows that Pc is indeed derepressed base pairs from PBAD, position -78 to -107 (Fig. 2 A and B). by the addition of arabinose. The initial activity of Pc, as In one of the experiments shown (Fig. 2A), almost complete demonstrated by the slope of the curve for f3-galactosidase level, protection of CRPBAD was achieved, and protection and en- is about 5-fold higher than the activity just preceding the ad- hancement of cleavage sites in CRPC were only barely detect- dition of arabinose. At about 15 min after arabinose addition, able. In the other experiment shown (Fig. 2B), the original data by which time the ,B-galactosidase level has almost doubled, the showed more complete protection of CRPBAD than CRPC, al- rate of Pc expression falls back to about its preinduction value. though the reproduced data show the differences less well. Subsequently (data not shown), Pc retains this preinduction From these observations we conclude that the in vitro affinity value so that the level of f3-galactosidase per cell is diluted by of CRP protein is higher for CRPBAD than it is for CRPC. cell growth back to its preinduction level. Location of the araC Protein Induction and Repression The expression of Pc appears to be strongly catabolite sen- Binding Sites. By similar DNase I experiments, the araC pro- sitive. As shown in Fig. 3, adding to the medium 30 min tein-binding sites have been located (Fig. 2 C and D). araC, in before adding arabinose drastically reduces the magnitude of the presence of arabinose, identifies a binding region from the Pc derepression. Additionally, as shown, a high concen- positions -40 to -78, and a region less well protected from tration of cAMP in the medium reverses the catabolite repres- -106 to -144 (Fig. 2C). When the same DNase I experiments sion caused by glucose. are performed with araC protein in the presence of the anti- D-fucose, the region -106 to -144 is protected, -and DISCUSSION to a lesser extent the region -40 to -78 (Fig. 2D). Thus the In Results we have presented data that locate the DNA-binding sequence from -40 to -78 is identified as the araI site, and the sites of the proteins involved in the joint regulation of the region from -106 to -144 is identified as the araO site in ac- arabinose operon promoters PC and PBAD. Determined were cordance with the nomenclature of Englesberg (6). the two binding sites of CRP protein, CRPC and CRPBAD, and No evidence of other araC protein-binding sites, either the binding sites araI and araO of araC protein in its inducing arabinose- or fucose-dependent, was seen in the region +45 to and repressing states. Additionally, DNase I protection exper- Downloaded by guest on September 28, 2021 3348 Biochemistry: Ogden et al. Proc. Nati. Acad. Sci. USA 77 (1980)

U1) U) 0 c (a) Z u) 4- 0 o O L. L. L O U OL Or + c c 0 + -1 0> a) U) z oD Z 0 11 0 o L- 00CL 0 CL < C 000X >C Xz~ 0 0 0 Z 0 -150'

-80- -140- ma -70 - -130 -60 _ -150 -120 - 0 -50_ -140 Q0 - -130 -140 -40 _ -110 -120 ~~~^ ~~ .~~-j30 -100 - zA* -lF -30* ___

- -120I -100 am =ft -90s

aw- _od -. ~~~~~~~~~-I _ "Nl -EA -80- -80 X CD

__ _ f-% .... _6% f __Ol _ _- - 100 -70 _t- , 46_okf JN A A B C D FIG. 2. Locations of CRPBAD, CRPc, araI, and araO by DNase I. Indicated are the regions containing bases with protected or enhanced cleavages. Arrows indicate the bases with cleavage enhanced by the presence of the protein. (A) Protection of CRPBAD from DNase I digestion. Buffer was 10 mM Tris-HCl, pH 7.9/50 mM KCl/lO mM MgCl2/2.5 mM CaCl2/0.1 mM dithiothreitol/30% (vol/vol) glycerol/0.2 mM cAMP. DNA was labeled at~~~~_the BamHI site at position -43. Lane 1 is G>A sequencing size control. For lanes 2-5 DNA was digested with DNase I at 0.13 jg/ml. Lane 2 is DNase I digestion with no other protein added. Lanes 3-5 are DNase I digestion with CRP present at concentrations of 6.5,32.5, and 65 gg/ml. (B) Protection of CRPC from DNase I digestion. DNA was labeled at the BamHI site at position -43. Lanes 1-3 contained CRP, a different preparation than that used in part A, at 12 Ag/ml; DNA was digested with DNase I at 0.10, 0.12, and 0.2 ag/ml. Lane 4 contained no other added protein and its DNA was digested with DNase I at 0.13 ,g/ml. (C) Protection of araI from DNase I digestion. DNA was labeled at the EcoRI end at position +48. The G>A sizing lane is not shown because it was substantially fainter and required longer exposure. DNA was digested with DNase I at 0.2 ,g/ml. Lane 1 is DNase I with no other added protein and lane 2 contained 2.5 ,l of araC protein and 100 mM L-arabinose. (D) Protection ofaraO from DNase I digestion. DNA was labeled at the BamHI site at position -43. Lane 1 is the G>A sizing standard. DNA was digested with DNase I at 0.1 jug/ml. Lane 2 contained no other added protein and lane 3 contained 2.5 jAl of araC protein and 100 mM D-fucose.

iments performed with araC protein plus CRP indicate that CRP to the CRPC sequence stimulates the activity of RNA the presence of the CRP-cAMP complex enhances the binding polymerase on Pc. Similarly, the binding of CRP to the CRPBAD of araC protein. Similarly, electron microscopy previously in- DNA sequence leads to greater occupancy by araC protein in dicated that RNA polymerase detectably binds to PBAD only its inducing state to a site adjacent to the CRP on the araI se- when CRP and araC proteins are present (4). The microscopy quence. In turn, the presence of araC protein increases occu- also showed that CRP-cAMP stimulated the binding of RNA pancy by RNA polymerase of the promoter DNA and hence polymerase to the Pc promoter about 3fold. This information, stimulates transcription from PBAD- plus the observation contained in this paper that addition of This mechanism directly explains the facts enumerated in arabinose induces both PC and PBAD, suggests the mechanism the Introduction on regulation of PBAD and Pc: (i) positive and sketched in Fig. 4: araC protein in the repressing state binds negative control by araC of PBAD; (ii) the CRP requirement for to the araO sequence. In this poition araC blocks access of CRP induction of PBAD; (iii) negative control on PBAD is exerted to the CRPc and CRPBAD sites located on either side. Thus PBAD upstream from positive control; (iv) araC protein represses Pc; remains fully off, but because RNA polymerase retains signif- and (v) the synthesis of araC protein can be stimulated by CRP icant "unaided" activity on the Pc promoter (3), a residual level (3,4). of araC protein synthesis remains. Upon the addition of arab- The mechanism leads to the prediction that under certain inose, araC protein leaves the araO site and permits the entry conditions the addition of arabinose to cells should derepress of CRP molecules to both CRP-binding sites. Such binding of the synthesis of araC protein. Furthermore, this derepression Downloaded by guest on September 28, 2021 Biochemistry: Ogden et al. Proc. Natl. Acad. Sci. USA 77 (1980) 3349 binding site adjacent to CRPBAD or convert the araC-binding site, I, to another CRP-binding site. From the relevant se- quences (8, 10) such results appear possible with one or two base changes. Mutations affecting araC protein have also been ob- served that permit expression of the arabinose operon in the absence of the normal inducer, arabinose (26, 27). Paradoxi- cally, these mutations, Cc, are not dominant to the wild-type araC protein. That is, uninduced cells containing both proteins possess basal levels of arabinose , whereas uninduced 0 cc mutants possess levels characteristic of wild-type fully in- o duced cells. The dominance of the repressing form of araC protein has long been hard to rationalize; however, it can now .00 0 be seen to be a natural consequence of the mechanism: occu- 0 pancy of the repression site blocks induction. A number of questions relevant to the proposed mechanism remain to be answered. Delineation of the RNA polymerase- £prt 0 00 binding sites on PC and PBAD by DNase protection has not yet been completed. We have not physically demonstrated that 00 ~ ~ araC protein bound to the araO site directly blocks the entry c 2 *

RNAP RNAP Steady-state Cind induction CRP

.1 BAD -~~~~~~~~~~~~~ E

-200 -160 -1 20 -80 -40 1 Base pairs FIG. 4. Proposed mechanism of ara operon regulation. Represented is the binding status of the regulatory proteins before, immediately after, and several minutes after induction by the addition ofarabinose. The positions of the binding regions are to scale, but not the sizes of the proteins. Also depicted are the activities of the leftward pC and rightward PBAD promoters in the three conditions. RNAP, RNA polymerase; Crep and Cind, araC protein in repressing and inducing conformations. a requirement for the complete set of proteins necessitates that 12. Eilen, E., Pampeno, C. & Krakow, J. (1978) Biochemistry 17, the binding of the proteins to DNA be cooperative. Such a 2469-2473. regulatory mechanism is versatile and easily accommodates 13. Burgess, R. & Jendrisak, J. (1975) Biochemistry 14, 4634- situations in which large numbers of must all be 4638. present in a to 14. Maxam, A. & Gilbert, W. (1977) Proc. Natl. Acad. Sci. USA 74, order for gene be induced. 560-564. 15. Galas, D. & Schmitz, A. (1978) Nucleic Acids Res. 5, 3157- We gratefully acknowledge Allen Maxam for his advice on the se- 3170. quence determination procedures and Ann Favreau for technical as- 16. Lis, J. & Schleif, R. (1975) Nucleic Acids Res. 2, 383-389. sistance. This work was supported by U.S. Public Health Service Re- 17. Schleif, R. (1972) Proc. Natl. Acad. Sci. USA 69,3479-3484. search Grant GM-18277 and Training Grant GM-212 from the Na- 18. Haggerty, D., Oeschger, M. & Schleif, R. (1978) J. Bacteriol. 135, tional Institutes of Health. 775-781. 19. Katz, L. & Englesberg, E. (1971) J. Bacteriol. 107,34-52. 1. Greenblatt, J. & Schleif, R. (1971) Nature (London) New Biol. 20. Lis, J. & Schleif, R. (1973) J. Mol. Biol. 79, 149-162. 233, 166-170. 21. Prisvant, M. & Lazdunski, C. (1975) Biochemistry 14, 1821- 2. Wilcox, G., Neuris, P., Bass, R. & Englesberg, E. (1974) J. Biol. 1825. Chem. 249,2946-2952. 22. Heffernan, L., Bass, R. & Englesberg, E. (1976) J. Bacteriol. 126, 3. Casadaban, M. (1976) J. Mol. Biol. 104,557-566. 1119-1131. 4. Hirsh, J. & Schleif, R. (1977) Cell 11, 545-550. 23. Heffernan, L. & Wilcox, G. (1976) J. Bacterlol. 126, 1132- 5. Sheppard, D. & Englesberg, E. (1967) J. Mol. Biol. 25, 443- 1135. 454. 24. Englesberg, E., Sheppard, D., Squires, C. & Meronk, F. (1969) 6. Englesberg, E., Squires, C. & Meronk, F. (1969) Proc. Nati. Acad. J. Mol. Biol. 43, 281-298. Sci. USA 62, 1100-1107. 25. Gielow, L., Largen, M. & Englesberg, E. (1971) Genetics 69, 7. Schleif, R. & Lis, J. (1975) J. Mol. Blol. 95, 417-431. 289-302. 8. Smith, B. & Schleif, R. (1978) J. Blol. Chem. a53, 6931-6933. 26. Englesberg, E., Irr, J. & Lee, N. (1965) J. Bacteriol. 90, 946- 9. Greenfield, L., Boone, T. & Wilcox, G. (1978) Proc. Nati. Acad. 956. Sci. USA 75,4724-4728. 27. Nathanson, N. & Schleif, R. (1975) J. Mol. Biol. 96, 185-199. 10. Taniguchi, T., O'Neill, M. & de Crombrugghe, B. (1979) Proc. 28. Majors, J. (1977) Dissertation (Harvard Univ., Cambridge, Nati. Acad. Sci. USA 76,5090-5094. MA). 11. Steffen, D. & Schleif, R. (1977) Mol. Gen. Genet. 157, 333- 29. Hirsh, J. & Schleif, R. (1976) Proc. Natl. Acad. Sci. USA 73, 339. 1518-1522. Downloaded by guest on September 28, 2021