The Regulation of Transcription Initiation in Bacteria

Annual Reviews www.annualreviews.org/aronline

Annu. Rev. Genet. 1985.19:355-387. Downloaded from arjournals.annualreviews.org by University of Wisconsin - Madison on 03/29/07. For personal use only.

Ann.Rev. Genet. 1985. 19:35547 Copyright©1985by AnnualReviewsInc. All rights reserved

THE REGULATION OF TRANSCRIPTIONINITIATION IN BACTERIA William S. Reznikoff I, Deborah A. Siegele 2, Deborah W. Cowingz, 2and Carol A. Gross ~ andBacteriology 2, Collegeof AgriculturalandLife Departments of Biochemistry Sciences, Universityof Wisconsin,Madison,Wisconsin53706

CONTENTS INTRODUCTION ..................................................................................... PROMOTER STRUCTURE ......................................................................... THEREGULATION OF E~TM TRANSCRIPTION INITIATION............................ NegativeRegulationof TranscriptionInitiation ............................................. PositiveRegulationof TranscriptionInitiation .............................................. DNATopologyRegulationof TranscriptionInitiation ...................................... DNAModification and the Regulationof GeneExpression............................... MODIFICATION OF HOLOENZYMESTRUCTURE AND THE REGULATION OF GENE EXPRESSION ................................................................... E. coil ~r7° .......................................................................................... TheHeatShockResponse and~r~2 ............................................................. The E. coli ntrA (glnF) Protein: AnotherSigmaFactor................................... B. subtilis Sporulation ........................................................................... SigmaFactorsin B. subtilis .................................................................... Phage-Encoded SigmaFactors................................................................. Conservationof Structure AmongSigmaFactors........................................... Modulation of SigmaFactorUse.............................................................. Role of Sigmain PromoterRecognition......................................................

355 356 360 361 363 365 366 366 367 368 370 371 372 376 377 377 379

INTRODUCTION Thefirst step in geneexpressionis the transcription initiation event, a multistep process catalyzed by RNA polymeraseholoenzyme.Transcriptioninitiation is a veryprecise eventoccurringat specific sites withspecific orientationson the 355 0066-4197/85/1215-0355502.00



356

REZNIKOFFET AL

chromosome.This specificity results from the recognition by RNApolymerase holoenzymeof DNAsequences termed promoters. The sequence of the promoter determinesthe location and orientation of the 5’ end of the mRNA and is an important element in determiningthe frequencyof transcription initiation. Whatis the structure of a promoter?Onegoal of this article is to present our current answer to this question. Weshowthat although a simple modelcan be used to define manyEscherichia coli promoters, numerousexceptions to this modelare known.It is thus possible that various alternative promoterstructures exist, possibly reflecting alternative routes in the transcription initiation process for different promoters, or else we have in general oversimplified our understandingof their structure. Animportant meansby which gene expression can be controlled is through the regulation of transcription initiation. That is, the frequencyof transcription initiation as programmed by a promoter sequence can be modulatedby a variety of mechanisms. The most well-known mechanism by which this is accomplished is through the binding of someother sequence-specific protein nearby or within the promoter. Othermechanisms of regulating transcription initiation are also knownto exist. The secondgoal of this article is to describe these various regulatory strategies, concentrating on systems found in E. coli. One mechanismthat functions in E. coli, the generation of different specics of holoenzymethrough the use of different ~r subunits, has been extensively studied in Bacillus subtilis and is also described. The reader mayalso be interested in articles which describe howgene regulation in somesystems occurs through the control of the transcription termination process (3a, 37a).

PROMOTER STRUCTURE Theprimary approachto defining the location and structure of a promoteris the identification of mutations that alter its recognition by RNA polymerase.These mutations alter (either raise or lower) the level of gene expression in cisdominant fashion and do so by directly affecting the RNApolymerasepromoterinteraction. Genetic analysis of such mutations for bacterial systems provides a simple commonresult. Promoter mutations are located prior to (upstream of) the structural gene(s) whose expression is being programmed.Classical genetic techniques obviously do not provide the details required to define promoter structure; instead, one must turn to DNAsequence analysis coupled to the biochemical definition of the precise mRNA start point. The value of DNA sequence determination is obvious: it provides the ultimate in fine-structure genetic mapping.The mRNA start-point determination is also critical because it providesan orientation point for the promotersequence;this allows it and its mutant variants to be comparedto other knownpromoters.

Annual Reviews www.annualreviews.org/aronline TRANSCRIPTION INITIATION REGULATIONIN BACTERIA

proposed promoter

canonical sequence

-45 0


lacP mutations consistent with canonical model

-35

+1

+I0

cot

GAC A TG T AA(+)

other lacP mutations

bacteriophage

-I0

17 bD ~ t t tg TAtAaT tcTTGACOt

357

A(+)

AAAA

TTGCT

TATAATA

T(=) C(+)A(+)

CATA TTGA

TS promoter homologies Figure 1 Promoter structure for ~r TM holoenzyme. The top line indicates the proposed canonical promoter structure for one strand of DNA(57). + I indicates thc position corresponding to the first base of the mRNA,+ numbers indicate downstream sequences, - numbers indicate upstream sequences. The indicated lacP mutations all decrease lacP expression unless they are followed by a (+) (increase in expression) or (=) (no change in expression). The mutations are described References (113) and (120). The bacteriophage T5 promoter homologies were described Reference (8).

Figure 1 summarizes (57) the general result obtained from this kind genetic, sequence, and mRNA start-point analysis for E. coli promoters recognized by ~r7° holoenzyme (Eo"7° is the major E. coli RNApolymerase species; an alternative species is discussed below). Using the mRNA start point as guide, one can see that there are two sequences whose presence is highly correlated with promoter activity. One hexadinucleotide sequence (TATAAT in the antisense strand) is centered 10 bp before the start point. A second hexadinucleotide sequence (TTGACA)is centered approximately 35 bp upstream of the start point. Although the spacing between the -10 region sequence and the start point is somewhatvariable, there does appear to be a distinctly favored spacing of 17 bp between the -35 and the -10 sequences. The results that lead to the above generalizations are the following: 1. The above commonalities were found in a statistical analysis of more than 100 E. coli promoter sequences. 2. Most mutations that alter promoter activity change the particular promoter’s sequence in an expected fashion (see second line of Figure 1). For instance, mutations that enhance the similarity of the particular promoter to the proposed canonical sequence (either by changing the - 10 region, the -35 region, or the spacing between the -10 and -35 regions) enhance the promoter’s activity; additional examples have been recently reviewed (57),

Annual Reviews www.annualreviews.org/aronline 358

REZNIKOFFET AL


and the extensive mutational analysis of the phage P22 ant promoter in particular offers evidence that changes awayfrom the proposed common sequence decrease promoter activity (151a). 3. In general one can obtain an approximate ranking of promoter activity vis-?~-vis other promoters by comparingtheir sequences (101). Obviously other types of RNApolymerases (either completely different proteins as in the case of bacteriophage-encoded enzymes, or holoenzymes containing different species of tr subunits) will recognize other sequencesas promoters (examplesare discussed below). Oneexpects these sequences to fit 7° into decipherable general patterns. The surprising observation with the ~r holoenzymeis that there are several exceptions to the proposed generalized pattern. This, in turn, suggests a caution in developingrigid canonical guidelines describing promoter sequences for other RNApolymerases; the exceptions mayprovide us with interesting information about howtranscription initiation occurs and howit can be regulated. Someof these exceptional findings are described next. 1. Somepromoter mutations do not affect any of the canonical promoter sequence characteristics described above and in Figure 1. The lac promoter provides three such examples(see Figure 1, third line); Pqa (at -i6), $7 +6), and prl 11 (at + 10) (25, 91). Thereare various possible explanations the ways in which these mutations enhance promoter function. H. Bujard’s laboratory performed a comparative sequence analysis using only extremely strong promoters that programbacteriophage T5 gene expression (8). This analysis extended the canonical sites described above to include an AAAA sequence near -43, a TTGA sequence downstreambetween + 5 and + 9, and a general tendency toward A/T richness (see Figure 1, fourth line). Thus the analysis that generated the picture presented in Figure I mayhave missedsome sites that contribute to the transcription initiation process, possibly becauseit included too manypromoters, weakand strong. Alternatively, other base pairs in the promoter could be recognized by RNApolymerase but might become significant in the overall process only whencanonicalrecognition sites are not present. 2. Somepromoters are located within DNAsequences that contain other RNApolymerase binding sites, some of whose sequences overlap with the promoterin question. The lac promoteris a good exampleof such an arrangement, although it is not unique. As shownin Figure 2, there are two RNA polymerasebinding sites that overlap lacP. Onesite (P2) is displaced 22 upstream from lacP (82, 112, 121). The second site (Pl15), which becomes obvious only as a consequenceof a single bp change, is displaced 12-13 bp downstreamfrom lacP (90, 113). A possible consequence of this type arrangementis that someapparent promoter mutations mayact not by changing the promoter sequence per se but by activating an alternative sequence to

Annual Reviews www.annualreviews.org/aronline TRANSCRIPTION INITIATION REGULATION IN BACTERIA 359 PII5 I


I

I p:)

PI

-22

I

"el

I +13

I

CAPSITE OPERATOR Figure2 Overlapping RNA polymerase bindingsites in the lac controllingelements. In addition to the promoter responsible for programming lacPexpression (P1),thereexisttwoothersequences recognizedby RNA polymerase;P2, displaced22 bp upstreamfromP1, and Pl15, 13 bp downstream (82, 90, 112,113,121).Mutation Prl 15results in enhanced CAP-cAMP-independent lacexpression, it hasnoeffectonP1expressiou (A/T--~ T/A(=) change at + 1, seeFigure1), ratheractivatesP115expression.Many mutations in the -35regionof PI are also in the -10 regionof P2.Alsoindicatedare the targetsites for the CAP-cAMP complex andthe repressor. program transcripts (an example would be the lac mutation Prll5) or by generating a competitive RNApolymerase-DNA interaction. The presence of multiple RNA polymerasebinding sites close to a promoter also suggeststhe interesting possibility that the alternate sites (or portions of these sites) mayserve to enhancepromoteractivity. For instance, a series of clustered sites that can form loose but specific contacts with RNA polymerase [similar to the product of the first step in RNApolymerase-promoterinteraction; sometimescalled a "closed complex"(80)] might serve as an "antenna" to enhance the probability of RNApolymerasemolecules being in the local environmentof the promoterand thus facilitate its productiveinteraction with the promoter.This possibility has not been critically examined.Mutations in such an antenna could enhanceor decrease promoter activity. 3. Somepromoters appear to lack one of the canonical sequences (a - 10 region or a -35 region) without a commensurateeffect on promoter activity. Anextreme exampleis the PRE*promoter, a mutant of the kPREpromoter. It differs fromthe hPREpromoterby three single bp substitutions, whichresult in the generation of a "canonical" - 10 sequence. However,PRE*does not appear to have a "-35 region" at all. Moreover,PRE*DNA molecules that end at - 17 (totally lacking any DNA sequencescorrespondingto a -35 region) are able programtranscription in vitro. This property is unlike that of any other tested promoter DNA fragment. It is proposed that in this case a sequence between - 13 and - 17 compensatesfor the lack of a - 35 region (S. Keitty, Y. S. Ho,G. Sathe, and M. Rosenberg, personal communication). The lac P115 promoter may be another example. The level of P115 programmed transcription is significantly above that predicted from its sequence relationship to the canonical sequence(101, 153). The primary discrepancy that P 115 has no appropriately positioned sequenceapproximatinga canonical -35 sequence (101, 113, 153). As discussed below, Pl15 mayalso have compensating sequence. 4. The canonical promoter sequence modeldescribed above predicts that DNAsequenceinformation upstream of the -35 region should have little or no influence on promoteractivity. This can be tested by synthesizing or isolating



360

REZN1KOFFET AL

deletions that progressively approachthe promoterfrom the upstream region. This type of study with the lacUV5promoter gives the expected result. The deletions have no effect on promoteractivity until they actually encroachupon the -35 sequence (153). However,quite a different result has been found for other promoters. In P115, the approachingdeletions have no effect until they reach position -42, at which point a 3-fold decrease in promoter activity is observed. These deletions maydefine an upstream sequence that compensates for the lack of a canonical -35 region in P115 (153). A more dramatic result was reported for the promoters programmingtyrT (71), rrnB (R. Gorse and M. Nomura,personal communication), ilvlH (56), ilvG (C. W.Adams,M. Rosenberg, and G. W. Hatfield, personal communication)and K. pneumoniaenifLA (28)3 expression. In these cases deletions ending 15 to 145 bp upstream from the -35 region decrease promoter expression by up to 10-fold. It is not known whetherthese deletions exert their effect by removinga specific site or by just changingthe overall sequencecomposition. However,in another case, the hisR promoterin S. typhimurium,a 3 bp deletion at position -70 (relative to the start point) results in a 2-fold reductionin promoteractivity (6). Thusin this case specific upstream site is implicated. In both this case and the rrnB promoter situation, the relevant upstream sequencesresult in an unusual DNA conformation as judged by aberrant electrophoretic mobility of DNA fragments carrying this region (6; R. Gorse and M. Nomura,personal communication). Howcan such long-range upstream effects be explained? Weconsider three general possibilities. The upstream sequences maynot be part of the promoter per se but rather serve a regulatory role. For instance, they maycontain a bindingsite(s) for a positive regulatoryprotein(s) or a site(s) of action for gyrase such that it generates the proper localized topology for maximalpromoter activity. Alternativelythey couldrepresent a part of the promoter.In this case they might be an extended antenna for RNApolymerase interaction as described above. Finally these sequences could locally affect nearby DNAconformation through "telestability" (7a) effects generated by virtue of their sequenceor composition.It is not clear whetherthis last possibility fits within the definition of a promoteror not. 5. Downstreamsequences mayalso affect the activity of some promoters, but except for the point mutations and T5 promotersequenceanalyses described above, this possibility has not yet been explored.

THE REGULATIONOF Eft 7° TRANSCRIPTION INITIATION Oneof the mostimportantconclusionsfromstudies of the last quarter century is that the regulation of geneactivity is often the result of regulating the frequency 3Note:nifLAmaybe transcribedbyan alternate formof holoenzyme as describedbelow.

Annual Reviews www.annualreviews.org/aronline TRANSCRIPTION INITIATION REGULATION IN BACTERIA 361 of transcription initiation. In the following sections we describe different mechanisms through whichthis can be effected. Theseinclude (a) the action regulatoryproteinsthat inhibit (repress) transcriptioninitiation; (b) the action regulatory proteins that stimulate (activate) transcription initiation; (c) regulation of DNAtopology; and (d) the modification of DNA bases within the promoter sequence.


Negative Regulation

of Transcription

Initiation

Thefirst modelto be elucidated was that gene expressioncould be regulated by a repressor protein that, whenboundto the DNA,wouldblock transcription initiation. The DNA binding property of the repressor wouldin turn be modulated by the binding to the repressor of a relevant small moleculeeffector. The lactose (lac) operon, shownin Figure 3, was an important systemfor studying this type of regulation, and to a first approximationit suggests an easily understood mechanismthat is readily applicable to manyother negatively regulated systems. The three lac genes are expressedat a high level only when the cells are grownin mediumcontaining compounds resembling the substrate for the pathway, lactose. Extensive genetic, physiological, and biochemical experimentshave indicated that this regulation occurs as follows. ThelacI gene encodes a repressor protein that when boundto the lac operator inhibits transcription initiation. This repressor-operator complexis destabilized (and transcription initiation is allowed) by the binding of any one of a numberof 13-galactoside compounds (for instance allolactose, an isomer of lactose, or isopropyl-13-t~-thiogalactoside ) to the repressor (see 63 and reviewsin 5, 95). Whilemanyother repressors havetheir activity modulatedby a similar effector binding mechanism,somerepressors, notably the bacteriophage lambda cl repressor, are inactivated by a proteolytic attack (122a). Howdoes the repressor act to block transcription initiation? Sequence analysis indicates that the repressor bindingsite (the operator) overlapswiththe promoter such that binding of the repressor and RNApolymerase should be P

T

PO

IOC Z

,/~-galactosidase

Lactose Thiogatoctoside Permease Transocetylase

Figure 3 The lactose operon. The transcription of the lacZ, Y, and A genes results from transcription initiation by RNApolymerase (RNAP)at lacP. This event is regulated in a positive fashion by the catabolite gene activator protein (CAP)when it is complexed with cyclic AMPand a negative fashion by the lacl gene product, the repressor. The repressor in turn is inactivated by the inducer (I). Translation (indicated by the wavyarrows) of each of the product proteins results ribosome (rbs) binding to the mRNA translation initiation signal for each gene. This figure is the same as that presented in Reference (120).



362

REZNIKOFFET AL

mutually exclusive competitive events (see Figure 2). In vitro transcription initiation experiments by Majors (84) measuring the kinetics of RNA polymerasebinding to the lacUV5promotersuggest that this is precisely the mechanismfunctioning for lac. That is, RNApolymerasedoes not bind to the promoter until the repressor is removedfrom the operator. The location of operators in manyother negatively regulated systems suggests that a similar mechanismfunctions there as well. There are important exceptions to this simple arrangement, two of whichare exemplified by the galactose operon system (61, 86): 1. galOc mutations define two gal repressor binding sites, both of which are required for maximum efficacy of the gal repressor, and 2. neither of the operators are located in a position that wouldbe presumedto result in direct competitive binding of the gal repressor and RNA polymerase.Onegal repressor binding site is located upstreamcentered at position -60; the other is located downstreamcentered at position +50. The arabinosc BADoperonalso showsa dual binding site modeof repressor action. This includes sites centered at -280 and a site within the araBAD controlling elements, probably at aral immediatelyadjacent to the promoter (30). Surprisingly, the lac systemalso showsan element of the multiple repressor binding site pattern, and someof its properties suggest a role for additional sites. It has long beenknownthat there are three repressor bindingsites in lac; the operator, a secondarysite with a repressor binding affinity approximately ten-fold lower than lacO centered at position +413 in the lacZ gene, and a tertiary repressor binding site with a still lowerrepressor affinity centeredat position -82 (41, 122). Deletion of the tertiary binding site has no measurable effect on lac regulation. Comparableexperiments removing the secondary binding site have not been done, but cloning the lacP-Oregion awayfrom the secondarybinding site does give constructs that are regulated in a qualitatively correct manner. Recent in vitro lac repressor-operator affinity measurements suggest that the secondarylac repressor binding site mayact to stabilize the repressor-operator complexand that this stabilization becomesparticularly capparent in cases where the repressor-operator affinity is weakenedby O mutations. That is, the repressor-operator affinity is weakerfor Oc mutations whenthe secondarybinding site is missing than whenit is present (100). This suggeststhat secondaryrepressor binding sites in general mayact by stabilizing repressor-operator complexesand that these secondary binding sites do not directly participate in the inhibition of transcription initiation. Weare still left with the secondquestion, howdoes the gal repressor function since in this case the presumedprimary repressor binding site does not overlap with the promoter sequences. It is possible that whenboundto both binding


363

sites simultaneously, the repressor holds the DNAin a conformation that is unfavorable for productive binding of RNApolymerase.


Positive Regulation of Transcription Initiation Starting with the studies on the arabinose (ara) regulon by Englesberg and his coworkers (see 73 for review), it was discovered that several systems are positively regulated. That is, the frequency of transcription initiation is enhanced by an activator protein. The activation of lacP expression by the catabolite gene activator protein-adenosine 3’ 5’-cyclic monophosphate (CAP-cAMP)complex is the model system we use to illustrate this type regulation (see Figure 3). In addition to the lac operon-specific control exemplified by the repressor-operator system described above, the maximal level of lac expression (and that of several other operons encoding catabolic functions) is modulated by the presence of (or, to be more precise, transport of) other available carbon sources in the media. Three classes of mutations have been isolated that prevent maximal level of lac expression and that define molecular components participating in this type of regulation. 1. Mutations in the crp gene that fail to make a functional CAP.This protein is known to be an activator for the expression of lac and other systems controlled by catabolite repression (24, 33, & 110). 2. Mutations in the cya gene that fail to synthesize adenyl cyclase (127). The cya mutants do not make cAMP.In vitro experiments have shown that cAMPis required for activation of CAP(24, 33). 3. Cis-dominant mutations associated with each regulated system that reduce the maximal stimulatory response of CAP-cAMP (4, 26, 152). Some these cis-dominant mutations define the DNAtarget site for CAP.This site has also been defined by biochemical experiments (126, 130, 185; also see Figure 2). As discussed below, the biochemical definition of this site is critical. This definition includes CAP-cAMP-IacPDNAbinding experiments that tested the effect of CAP site mutations on this reaction, and experiments that used chemical probe and nuclease protection protocols to examine the precise DNA binding site for CAP(85, 126, 130). -.30 t -,20 t -.10 t ÷.1 t -.20ti -.lOti *,lti AACGGAACCTTTCCCGTTTTCCAGGATCTGATCTTCCATGTGACCTCCTAACATGGTAACGTTCATGATAACTTCT a~ --’~

[ t mRNA

ti mRNA t protein

Figure4 Tn5 ~ransposase and transpos~lon ~nh~b~torpromoters. The ISS0 inse~ionsequence

bracketlng Tn5contains twopromoters tha~ program ~hesynthesis ofthetwoprotelns lnvo]ved in ~ran~po~l~lon. Thepromoter for~he~ransposase (~)lsexpressed a~lowlevels unless ~heindlca~ed do~sltes areunmethylated (LC.-P. Yin, M.Krebs andW.S.Reznikoff, unpubl~shod results). Therelevant slgna]s forthetransposition ~nhlb~tor promoter arelnd~cated by



364

REZNIKOFFET AL

Three important observations arise from studies on the CAP-cAMP DNA target site. 1. CAP-cAMP DNA target sites in a numberof systems can be defined, and comparative sequence analyses suggest that these sites approximatethe sequence AA-TGTGA-~-CACA-T(23, 30a). 2. CAP-cAMP DNA target sites are found at different locations vis-h-vis the transcription start site in different systems(23). 3. In lac, other mutations have been isolated with the sameLac phenotype (Lac- because lac expression does not respond properly to CAP-cAMP) but are not in the biochemically defined CAP-cAMP target site. These are "spacer" mutations; insertions or deletions that change the distance betweenthe CAPcAMP target site and the promoterand changethe structure of the overlapping upstream RNApolymerase binding site P2 (89, 152), Unlike the situation with negative regulation, a mechanismfor positive regulation is not intuitively obvious. (Of course, our intuition about negative regulatory mechanisms may be misleading for some systems). Below we discuss three mechanismsby which activator proteins might act. 1. Anactivator protein might act by perturbing the conformationof the DNA extending into the promoter(25). This "telestability"-like effect could, for instance, lower the energy requirements for open complexformation. [The possible steps in the transcription initiation process havebeendiscussed elsewhere(80)]. There is no evidence specifically supporting this model. It compatiblewith the fact that CAP-cAMP target sites are located differently in different systems. In this case the spacer mutations might alter a sequence necessaryfor the propagationof the "telestability" effect. 2. The activator protein might act as a contact point for RNApolymerase and thereby facilitate its formation of a productive open complex(40). There exists compellinggenetic evidencethat this is the mechanism by whichthe h cI protein activates expression from promoterPRra(51). A missensemutation has beenisolated in the cI genethat results in the synthesis of a productwith normal DNA binding properties but that fails to stimulate transcription initiation from PRr~. A cI protein modelgenerated from X-ray crystallographic data suggests that the altered residue is appropriately positioned for forminga contact with RNApolymerase bound to PRr~ (60). The relationship of this modelto CAP-cAMP’s modeof action is less clear. Thefact that CAP-cAMP target sites are located in different positions vis-a-vis the promoter in different systems suggests that CAP-cAMP does not stimulate transcription by means of interaction between CAP-cAMPand RNA polymerase. However,this modelis not ruled out because of the following considerations. In some systems, such as gal, the CAP-cAMP target site appears to be appropriately positioned in order to facilitate the proposed protein-protein contact. In other systems, such as the araBADoperon, the



TRANSCRIPTION INITIATION REGULATION IN BACTERIA

365

regulatory sites in the DNAhave the following arrangement. The CAPtarget site is next to the target site of the araactivator protein (araC), whichis in turn next to the promoter. Such an arrangementmight facilitate a protein (CAPcAMP)-protein(araC)-protein (RNApolymerase) interaction. Left unresolved are situations such as lac in whichthe CAP-cAMP target site is apparently located too far from the promoterto interact directly with an RNA polymerase boundto lacP, and yet there is no knownintervening binding site for another regulatory protein. Oneway to resolve this discrepancy wouldbe to propose that CAP-cAMP interacts with RNApolymerase, which binds to an upstream overlappingRNApolymerasebinding site, P2, and facilitates the movement of RNA polymerasefrom P2 to the promoter(89; see Figure 2). In this light it interesting to recall the "spacer"mutationsin lac, whichreduce the efficacy of lac stimulation by CAP-cAMP. Those spacer mutations which have been tested reduce the affinity of RNApolymerasefor P2 (154). 3. The activator protein might act as a repressor of a protein-DNA binding reaction that wouldotherwise inhibit the RNApolymerase-promoterinteraction. This was initially proposed to partially explain howCAP-cAMP might activate lac (82, 87, 121). If one assumes that P2-boundRNApolymerase inhibits RNApolymerase interaction with lacP, then CAP-cAMP might be envisionedas blockingthis inhibition. Theproperties of the "spacer" mutations in lac suggestthat, for this specific case, this modelis not correct. Thespacer mutations reduce the RNApolymerase-P2affinity. If the competitive model were correct, one would predict that these mutations would enhance CAPcAMP-independentlac expression. Such an enhancement is not observed (154). Thereis however,another systemin whichsuch a cascadeof competitive binding events mayoccur. This involves the CAP-cAMP stimulation of hutUH in K. pneumoniae. The hutUHcontrolling elements contain two divergent promoters that overlap and appear to compete with each other for RNA polymerase binding. The CAP-cAMP target site overlaps with the upstreamfacing promoter. Binding of CAP-cAMP to this site appears to inhibit RNA polymerasebinding to the upstream-facingpromoterand stimulate its binding - to the downstream-facingpromoter (107). DNATopology Regulation

of Transcription

Initiation

The in vivo template for transcription is knownto be a negative superhelix. Since the presence and amountof superhelicity affect the energy required to denature any region in the molecule, and since RNApolymerase-promoter open complexformation involves localized DNAmelting (128), changing the superhelical character of the template wouldbe expectedto changeits promoter properties; therefore, this mightbe a mechanism used for regulating transcription initiation. In fact, as reviewedby Gellert (39), the superhelicalcharacter the template does affect its template properties (although sometimesin



366

REZNIKOFFET AL

surprising fashion; manypromotersare stimulated by negative superhelicity but someare inhibited). Effects of superhelicity on the process of transcription initiation have been shownin vitro in numerousstudies [for instance, see Botchan(6a) and Malanet al (87)] and, throughthe use of gyrase inhibitors, vivo by Sanzey(125). Thereis one case in whichthe regulation of transcription initiation in response to variations in superhelical density is believed to be physiologically important. The expression of the gyrA and gyrB genes encoding the subunits of gyrase (whosefunction it is to generate supercoils) is enhanced by inhibiting gyrase activity (93). This observation suggests that gene regulation mayplay an importa,nt role in the homeostatic control of chromosome supercoiling. DNA Modification and the Regulation of Gene Expression A frequently invoked model for gene regulation in eukaryotes involves the presence or absence of methyl groups on the DNA.Somebacterial systems clearly use this mechanism(131a). The effect of DNA methylation on promoter activity was first suggested by experimentsof Denise Roberts and NancyKleckner in studies of transposition of Tnl0 (personal communicationand 67). Host mutants were isolated that enhancedthe frequency of Tn 10 transposition. Someof these were found to be in the damgene, whichencodes the methylase that forms N6-methyladenine at the DNAsequence GATC.The enhanced Tnl0 transposition in these mutants was found to be correlated with the enhancedsynthesis of the transposase mRNA. The same basic mechanismappears to be functioning in Tn5 (J. C.-P. Yin and W. S. Reznikoff, unpublished results). As shownin Figure 4, the sequenceof the Tn5transposase promoter suggests a reason for these observations. In these transposonsthe transposasepromotercontains either one (Tn 10), or two (Tn5) damsites within the promoter. It is proposedthat methylationof these sites inhibits the recognition of these promoters by RNA polymerase. A similar presence of damsites has also been observed for some promoters located in the bacteriophageP1 genome(13 l a). If one assumes that hemimethylated DNAalso has altered protein-DNA recognition properties in these cases, this damsite modification mechanism provides an intriguing meansby whichgene expression could respond to (a) the passage of the replicating fork, (b) the repair of damagedDNA,and (c) transfer of DNA via conjugation. If no methylationof both strands is required, this systemwouldalso haveinteresting implications for its int~’oduction into heterologous hosts; not all bacteria have a darn system (3). MODIFICATION OF HOLOENZYME STRUCTURE THE REGULATION OF GENE EXPRESSION

AND

The first evidence for dissociable RNApolymerase subunits required for selective initiation of transcription camefromthe discoveryof ~r7° by Burgess



TRANSCRIPTION INITIATION REGULATION IN BACTERIA 367 and Travers (10, 11). E. coli RNApolymerase, purified to homogeneityby chromatography on a phosphocellulose column (9), is highly competent transcribe calf thymusDNA,but is unable to transcribe T4 DNA efficiently. A search for the factor that permitted such transcription resulted in the discovery of the ~r7° subunit (10, 11), and the recognition that RNA polymeraseexists two forms: holoenzyme(Eo "7°) capable of selective initiation at promoter regions, and corc RNApolymerase(E) capable of elongation and termination but not of selective initiation. This characteristic of RNA polymerase,that the ability to initiate transcription selectively is dependenton the presenceof the ~r subunit, suggests that the specificity of gene expressioncould be modulatedby substituting one species of ~r for another(10). Belowweshowthat this is indeed the case and that this is an important meansof regulating gene expression. The discovery of ~7o was madepossible by twothings: (a) the existence of template containing promoterswhosetranscription required that sigma factor, and (b) the existence of a procedurefor separating the sigmasubunit from core RNApolymerase. Since that time, other sigma factors, encoded either by bacterial or phagegenomes,havebeenpurified. In each case their identification dependedon the ability to separate the putative sigmasubunit from core RNA polymerase and the existence of a template capable of demonstrating the specific promoter recognition properties of the reconstituted holoenzyme. The sigmafactors are used to regulate genes with diverse biological roles. Thefirst alternative sigmafactors wereidentified in B. subtilis and its phages (29, 37, 138, 139) and were implicated in the temporal regulation of development (reviewed by Losick &Youngman,79). Morerecently, alternative sigma factors were also implicated in cellular responses to various environmental stimuli. Below, we briefly describe the biological role of each sigma. We concludeby considering the interactions of the sigmafactors with each other, the similarities amongsigmas, and the role of sigmain selective initiation.

E. coli

(~70

The tr 7° subunit of RNApolymeraseis knownto be essential for cell growth since mutations in rpoD(encoding~r7°) confer a ts growthphenotype(55, 75, 102, 109). ~r7° is required for mosttranscription; cells containingthe temperature-sensitive sigmamutation, rpoDSO0,cease expressing most cellular genes shortly after temperatureupshift (48). E. coli cells contain about 3000moleculesof ~7o, enoughto bind about one third of the intracellular core RNApolymerase (32, 62). On the basis comparisonsof in vivo and in vitro transcription rates, McClure(81) estimated that only about 1%of the RNApolymerase holocnzymeis present as free poly~nerase. The remaining holoenzymeis sequestered by nonspecific DNA binding and acts to buffer the in vivo concentration of free polymerase.This notion is consistent with the observation that, although gene expression is 7° limited by transcription initiation, increasing the absolute concentrationof ~r


REZNIKOFFET AL

increases neither the overall rate of transcription nor the synthesis of major cellular proteins (103).


~2 The Heat Shock Response

and cr

The heat shockresponsein E. coli is regulated by the product of the htpR(rpoH) gene (105, 151), which was recently shownto be a sigma factor, 32 (50). Whencells are shifted from low to high temperature, the synthesis of a small numberof proteins, the heat shock proteins, transiently increases. The heat shock response in bacteria was recently reviewed (49, 106). The criteria that demonstrated the rpoH(htpR)gene product to be a sigma factor were similar to those described above(50). The factor responsible for transcription of a heat shock gene promoter was purified from crude extracts and separated from core RNApolymerase by electrophoresis on SDSpolyacrylamide gels. Uponrenaturation and reconstitution with core RNA polymerase, the new form of holoenzyme, E~r32, was shown to initiate 7° transcription from heat shock gene promoters, but not from two strong Ecr promoters: PlacUV5and the RNAI promoter of ColE1type plasmids (49, 50). A comparisonof the aminoacid sequences of ~r32 and cr v° showsthat they are very similar, consistent with the identification of cr32 as a sigmafactor (72, 155). (The regions of homologyamongsigma factors are discussed below). Five heat shock gene promoters, which precede the heat shock genes rpoD, dnaK,groE and the C62.5gene, have been characterized (20, 137). Transcription initiates from these promotersboth at low temperatureand after a shift to high temperature, but the amountof transcription increases after the shift to high temperature. These promoters are recognized in vitro by holoenzyme containing ~r32, Ecr32, but not by Ecr7° (20; Table1). Thesepromotersshare consensus seque,nce having T-tC-CcCTTGAAin the -35 region and CCCCATtTa in the -10 region. "32 The mechanismregulating the increase in transcription initiation by Eo after heat shockis unknown and is currently understudy in several laboratories. Twoalternative classes of modelscan explain enhancedinitiation at heat shock gene promoters.In the first class, activators or repressors acting at the heat shock gene promoters and affecting the frequency of initiation by Ecr32 are altered by the inducing stimulus. There is no evidence for auxiliary factors; however,an exhaustive search has not yet been carried out. In the secondclass of models, either the amountor activity of ~r32 itself transiently increases relative to other sigmafactors in the cell. In this case, the amountof ~r32 must limit the transcription of eventhe strongest of the heat shock promoters.This has beenshownto be true. cr32 is normallypresent in very small amounts(F. C. Neidhardt, personal communication; A. Grossman, unpublished data). When ~r32 is overproducedfrom a foreign promoter, the rate of synthesis of the heat shock proteins increases without a temperature upshift (A. Grossman,manuscript in preparation).



369


REZNIKOFFET AL

0 -32

is very unstable in vivo, with a T~/2 of about 3’ (A. Grossman, personal communication). Thus, changesin the rate of synthesis of 0-32 can rapidly alter its intracellular level. Whilethese findings are consistent with the secondclass of models, they do not prove such a model.


The E. coli

ntrA(glnF)

Protein:

Another Sigma Factor

Whenenteric bacteria (for example, E. coli, Salmonella, and Klebsiella) are nitrogen limited, the synthesis of a numberof proteins (including glutamine synthetase and aminoacid transport componentsand degradative enzymes)is induced 10-100-fold. The increased production of these proteins enlarges the capacity of cells to produce glutamate, the major precursor of other nitrogen-containing compounds.Someorganisms(for example, Klebsiella) also inducenitrogenaseand other proteins required for nitrogen fixation (83, 94, 143). Geneticstudies established that normalexpressionof the nitrogen-regulated genes requires both the ntrA(glnF) and ntrC(glnG) gene products, which function in vivo as positive regulators. In the absenceof these gene products, expressionof nitrogen-regulated genes is very low and does not respond to the imposition of nitrogen limitation (70, 83, 94). Becausecells lacking NtrC expressed glnA (encoding glutamine synthetase) at a higher rate than those lacking NtrA, it had been argued that NtrC also functioned as a repressor. Analysisof the in vivo transcription rates of the glnA geneusing Mud-1fusions established that both NtrAand NtrCwork at the transcription level (68, 83, 123). Finally, analysis of the 5’ ends of in vivo RNA provided a basis for molecular understanding of the dual action of NtrC(27, 119). These studies showedthat glnA is transcribed from two promoters (27, 119). In addition, there is a promoterinternal to the glnA operon (69, 83, 118,144). Comparison of transcription from ntr ÷ and ntr cells (118, 144) established (a) that represses transcription from the weak, upstream promoter and the internal promoters but i’s required for activity of the strong downstreampromoter responsiveto nitrogen regulation; and (b) that NtrAis required for activity the downstreampromoter but does not affect transcription from the upstream promoter. The probable biochemicalbasis of NtrCand NtrAaction has been elucidated by in vitro studies utilizing glnA as a template. It appears that NtrCis a DNA binding activator protein analogous to CAP(1, 59, 118) and that NtrA probably a new sigma factor (59). Footprinting experiments of Kustu and colleagues (59, 149) showedthat NtrCbinds with differential affinity to five sites upstreamof the glnA gene. Oneof the tightest binding sites overlaps the weakupstreampromoterand mayaccount for the ability of NtrCto repress its transcription (59, 118). The other sites are upstreamof the strong regulated promoter. Binding at these sites mayactivate the promoter. Using both an S30-basedtranscription-translation systemand a transcription system, Kustu



TRANSCRIPTION INITIATION REGULATION IN BACTERIA 371 and collaborators showedthat partially purified NtrApromotestranscription from the downstream regulated promoter upon addition of core RNA polymeraseand NtrC.Since ¢r7° is not present in the fraction and is not required for this reaction, this fraction mustcontain a newsigmafactor (59). Consistent with this idea, the sequence upstreamof the 5’ end of the RNA(see Table 1) bears no resemblance to an Eo"7° promoter but shows extensive homologyto several other ntrA dependentpromoters, including eight involved in nitrogen fixation (27, 59, 119). It is apparentlythe ntrA product itself whichstimulates glnA transcription rather than another protein whosesynthesis dependson the ntr system: Additionof an ntrA÷ plasmidto an ntrA- $30 resulted in a 30-fold stimulation of glnA expression and since protein synthesis in $30 extracts is dependent on an exogenousDNAtemplate, stimulation was presumably due to ntrA product synthesized from the plasmid (J. Keener and S. Kustu, unpublished). Hownitrogen limitation results in enhancedexpression of nitrogen-regulated genes is not known.Transcription of ntrA does not increase under nitrogenlimiting conditions, whichsuggests that the amountof NtrAdoes not regulate the response (22). Onthe other hand, the amountof NtrC does increase when cells are limited for nitrogen (69, 83, 118). In addition, somemutations mappingin the ntrC generesult in high-level constitutive productionof nitrogen-regulated proteins (70). Takentogether, these observations suggest that increases in the amountor activity of NtrC could result in the increased transcription of nitrogen-regulatedgenes. If so, the sensor of nitrogen limitation maywell affect the NtrC gene product. B. subtilis

Sporulation

It had long been knownthat RNApolymeraseisolated from sporulating Bacillus subtilis cells had a different polypeptidecompositionthan RNA polymerase purified from vegetative cells. This led to the speculation that newsigmasor other transcriptional factors might be responsible for altered transcriptional patterns during sporulation (78). Wepresent a brief description of the sporulation process in B. subtilis. Wethen describe howeach sigmain B. subtilis was identified and the current ideas about the biological role for each sigma. The precise stimuli inducing spornlation are unknown;however,it is efficiently induced by nutrient deprivation. This developmentalprocess has been divided into several phases based upon morphologicallandmarks. One of the earliest steps is invaginationof the plasmamembrane that divides the sporulating cell into a mothercell and a forespore compartment (35). The fore’spore engulfed by the mother cell membrane,develops a cortex and tough protein coat, and is eventually released as a dormantspore whenthe mothercell lyses. Both the spore and the mothercell contain a transcriptionally active chromosome (124), and there is some suggestion that gene expression is com-


REZNIKOFFET AL

partmentalized (74, 104). Mutations are knownthat arrest cells at various stages of sporulation. Genesidentified in such a mannerare namedfor the arrest stage; for example,spolIA mutants arrest at the secondstage. Mutationsthat arrest the sporulation process before visible landmarks have occurred are termed spoOmutants. Althoughsomeof the spoOmutants have altered phenotypes during vegetative growth, none of the spoOgene products has been shown to be essential for vegetative growth. Annu. Rev. Genet. 1985.19:355-387. Downloaded from arjournals.annualreviews.org by University of Wisconsin - Madison on 03/29/07. For personal use only.

Sigma Factors in B. subtilis The mostabundantsigmafactor in B. subtilis during vegetative growthis "43, 0 encodedby the rpoDgene located at 225° on the B. subtilis map.The gene was recently cloned and sequencedby Gitt et al (45). The changein nomenclature from 0.55 to 0.43 reflects the corrected M.W.assignment based upon DNA sequence determination. Sequencecomparisonshave established that 0.43 and 0.70 are homologous(45; also, see below). It is presumedthat the role of E0.43 is to transcribe most of the genes expressed during the vegetative state, but there are no conditional lethal mutationsin the B. subtilis rpoDgeneto test this assumption.Theonly existing rpoDmutations are the crsA mutations (133a; R. Doi et al, personal communication). The presence of glucose usually inhibits sporulation. Cells with crsA mutations sporulate even whenglucose is present in the medium.The molecular mechanismof this effect is not understood. Theconsensussequencefor promotersrecognizedby Eo"43 iS identical to that derived for E. coli Eo"7° (96, 98; Figure 1 and Table 1). However,strong promotersin B. subtilis appear to require additional information. For example, the tac promoter,a very strong promoterfor E. coli E0.7° that has the consensus sequenceat both the - 10 and -35 regions of the promoterwith a spacing of 16 bp, is not recognizedin vitro by B. subtilis E0.43 (96, 98). Possibly, B. subtilis E0.43 has very rigid spacing requirements.To date, all Eo"43 promotersidentified havea spacing of 17 or 18 bp. Twoadditional features of strongB, subtilis promoters have been noted: an extremely AT-rich region upstream of the promoter and the sequence RTRTG at positions -14 to -18 (96, 98). The importanceof these features for recognition by Eo"43 has not been tested. However,the AT-rich region upstream of several developmentally regulated genes has been found to enhanceexpression without altering regulation (2; R. Losick, personal communication),suggesting that the ATstretch maybe general structural feature of strong B. subtilis promoters. 0.28 is a minorsigmafactor present in vegetatively growingB. subtilis. The discovery of 0.28 is a prototypic exampleof howsigma factors are defined biochemically.Chamberlinand his collaborators (64) foundthat B. subtilis but not E. coli RNApolymerase makes a unique transcript from T7 DNA,termed the J transcript. Usingthe ability to makethis transcript as an assay, they




373

purified the enzymaticform with unique promoter recognition properties (64, 148). It proved to be a novel form of holoenzymecontaining a 28-kd protein rather than the normal 43-kd sigma factor protein. The 28-kd protein was dissociated from E~28 and shownto be responsible for the novel promoter recognition properties whenreconstituted with purified core, whichthus rigorously established that the protein was a sigma factor. The gene encoding this protein has not been identified. The numberof strong Err28 promotershas been estimated at 20-30 (43). To dissect the biological role of ~r28, Gilman&Chamberlin(42) madeuse the fact that this form of holoenzymehas a stringent promoter preference and has virtually no activity on templates lacking cognatepromoters. Theyidentified two B. subtilis DNA sequences recognized as promoters by E~r28 in vitro (44) and showedthat these promoters were also utilized in vivo. Gilman Chamberlin(42) showedthat these promoters were transcribed at a very low rate during vegetative growthand regulated by the sporulation machinery.The transcripts are absent in four different spoOmutants, spoOA,B, E, and F, and are shut off immediatelywhenexponential growth ends (preceding shutoff of Err55 transcripts by onehour). Notethat if all other tr 28 transcripts are similarly affected by the spoOloci, tr 28 must not be essential for vegetative growth. To determine if the lack of the tr 28 transcripts following the end of exponential growthresults from the loss ofE~r28, Wiggset al (147) comparedthe amountof E~r28 28 in growingand post-exponential cells and the amountof E~r in spoOAmutantand wild-type cells by followingErr28 activity after a one-step polymerase purification. They found reduced amounts of E~r28 activity in post-exponentialcells. In the absenceof an immunological assay, it is difficult to knowwhetherthis reduced activity reflects lack of tr 28 in the cell or the inability of ~r28 to associate with core RNA polymerase.The latter possibility has beenshownin the case of ~r43, whichis present duringsporulation but is not found in polymeraseprepared from sporulating cells (139). In contrast, Wiggs et al (148) recovered equivalent amountsof E~r28 from wild-type and spoOA mutants. Theseresults suggest that the spoOAlocus (and possibly other spo loci as well) encodesa protein required for E~r28 activity in vivo. Analternative possibility, that one of the spo loci encodes yet another sigma factor that actually reads these promoters in vivo, has not been excluded. Briat et al (7) recently showedthat E~r~8 recognizedthe E. coli heat shock promoterrpoDPhs in vitro, whichindicates that E. coli Eo"32 and B. subtilis E~r28 mayhave overlappingpromoterspecificity. It is also possible that these two ~r factors have homologous roles. Preliminarystudies indicated that the two identified tr 28 promotersdid not heat shockin either E. coli (7) or B. subtilis (42). Gilmanet al (43) recently isolated a newset of strong 28 promoters fromB. subtilis. It will be interesting to see the responseof these promotersto a temperatureshift.



374

REZNIKOFFET AL

Holoenzyme containing 0 .37 is present during vegetative growth and early during sporulation. Starting with a template containing spoVG,a gene that is actively transcribed early during sporulation and whoseexpression is under control of several spoO genes, Haldenwang&Losick (53, 54) purified the activity responsible for transcribing spoVGfrom extracts of cells early in sporulation (T1). They showed that RNApolymerase purified from these extracts containeda novel 37-kdprotein. This protein was separated fromother RNApolymerase subunits by chromatographyon phosphocellulose following denaturation in urea. Whenit was renatured and reconstituted with core RNA polymerase,the reconstituted enzymewas able to initiate transcription fromthe spoVGpromoter; this established that this 37-kdprotein was a sigmafactor. In vitro E0.37 also transcribes ctc (54), encodinga protein of unknownfunction, and sprE, encoding subtilisin (150). In vivo, all three genes are transcribed moreactively during stationary phase than exponential growth (99, 108,146). Transcription of spoVGand sprE, but not ctc, is under control of several spoO genes (97, 108, 114). The promoter sequencesimportant for E0. 37 recognition have been characterized by Moranand coworkers. A consensus sequence for E0. 37 promoters was proposed based on homology between the ctc and spoVG promoter regions (99). Toconfirmthe importanceof this sequenceand further define its features, Moranand colleagues analyzedthe interaction of Eo"37 with the ctc promoterby footprinting, DMSprotection experiments, and deletion and bisulfite mutagenesis of the cloned promoter. They find that the holoenzymebinds immediatelyupstream of the 5’ end of the gene and protects a region of DNA that includes the consensus sequence (96). The G residues on the nontranscribed strand at positions - 14, - 15, and - 16 (within or adjoining the - 10 consensus sequence) and at position -36 (within the -35 consensus sequence) were protected from methylation whenEo"37 was boundto the promoter region (96). Alteration of these same GCbase pairs to ATbase pairs by bisulfite mutagenesis weakenedpromoter activity both in vitro and in vivo (117, 135, 136). The biological role of 0 -37 has not yet been determined. Althoughthe three genesidentified as 0.37 templates in vitro are not transcribed by E0. 43, each is also transcribed by one or morealternate forms of holoenzy~ne,sometimeswith identical start points (spoVGis also transcribed by E0.32; ctc by both E0.32 and E0.eg) (65,136,146).Thein vitro studies makeit unlikely that E0.43 transcribes these promoters in vivo, but they do not establish which alternate form(s) transcribe these promotersin vivo. Identification of the gene encoding0.37 and isolation of mutantswill help to elucidate the role of 0.37 in B. subtilis development. ~32 is a very-low-abundance sigmafactor initially identified as a contaminat37 ing activity in an E0. preparation (58, 65). While relatively impure 37 E0.



TRANSCRIPTION INITIATION REGULATION IN BACTERIA 375 transcribed the spoVGgene from two start points separated by 10 bp, a more purified preparation of E0. 37 used only the upstreamstart. This difference was traced to a minorbandof 32 kd, visible only uponsilver staining the gel. When the 32-kd protein was eluted from the gel, renatured, and reconstituted with core RNA polymerase, the resultant holoenzymespecifically transcribed the spoVGgene from the downstreamstart. The biological role of this form of holoenzymehas not been investigated further. Sporulatingcells contain a 29-kdsporulation-specific protein associated with RNApolymerase (76, 104). Haldenwanget al. (52) purified RNApolymerase containing the 29-kd protein and used it to transcribe cloned DNAcontaining several sporulationgenesto see if they could detect a noveltranscript. Boththe ctc and spoVGgenes described above were transcribed by this holoenzyme form. In addition, a unique transcript termedL was obtained. Theyseparated the 29-kd protein from the other RNApolymerasesubunits by SDSgel electrophoresis, renatured it, and reconstituted it with core RNA polymerase.The reconstituted enzyme showed the same transcription properties as RNA polymerasecontaining the 29-kdprotein; hencethis protein is a sigmafactor. Haldenwangand collaborators have monitored the accumulation of -29 0 throughoutthe B. subtilis life cycle by using a monoclonal antibodyspecific to .29 is not detectably p 0 .29 (140--142). resent in vegetative cells a nd is present in significant amountsin sporulating cells only between stages T2 and T4. Accumulationof 0-29 is under sporulation-specific control and requires the wild-type gene products encoded by the spoOA, B, E, F, and H loci. The monoclonal antibodyto 0 .29 detected an additional protein of 31 kd in B. subtilis extracts. Like 0"29, P31waspresent only in sporulating cells. Theaccumulation of P31 precededthat of 0-29 by one hourand the twoproteins had related peptide maps, which suggests a precursor-product relationship for the two proteins. It has nowbeenshownthat 0 .29 is encodedby the spoIIGgene (133, 140-142) and is almostcertainly synthesizedas a 31-kdprecursormolecule.Stragier et al (133) showed that the spolIG gene encodes a protein of about 27 kd, homologousto the 0.70 protein of E. coll. Based on the M.W.of the open reading frame, they suggested that the locus might encode 0-29. TwospoIIG mutantswereshownto lack both 0-29 and P31. Significantly, one of the mutants produced immunologically reactive fragments of 25 kd and 21 kd, which suggests that this spoIIGmutation causes premature translation termination. Whenthe spollG plasmidwas put into this mutantstrain, proteins of 31 kd and 29 kd were detected, as well as the endogenous25 kd and 21 kd proteins. When the spolIG plasmidwas put into E. coli, only the 31-kd protein was detected. These observations are most easily explained by assuming that the 31-kd protein is the geneproduct of the spollG locus, whichis processed to a 29-kd protein in B. subtilis. 0.29 is the first sigma factor with a knownrole in sporulation. It is a



376

REZNIKOFFET AL

stage-specific sporulation protein encoded by the spolIG locus and is synthesized in a precursor form. Mutationsin spollG result in arrest at the second stage of sporulation; the cell is divided into the forespore compartmentand mothercell by the septum, but a detached spore within the mothercell does not form. It has beenreported that E0-29 is foundonly in the forespore compartment -29 m (104). ay be processed concomitantly w ith t ransport f rom m other cell t o forespore. Alternatively, the processing machinerymaybe active only in the forespore compartment. Since P31 associates only weakly with RNA polymerase and does not confer the ability to transcribe E0-29 promoters,the cell lacks 0-29 until the precursoris cleaved(141). Theprocessingof P31 to 0-29 may be intimately connected with the regulation of sporulation. Onespo locus has been implicated in the processing event; spollC mutations accumulatedP31 but not 0-29, Since~rz9 is present for onlya short time duringsporulation, it is likely that other sigmafactors will be implicated in the sporulation cascade. A potential sporulation sigma factor is the spollAC gene product which is 22 kd and homologousto E. coli cr 7° (36). It is not knownwhich sigma factor is the product of the spollAC gene. Phage-Encoded

Sigma Factors

The first evidencethat geneexpressionis regulated throughthe use of alternate sigma factors camefrom workon bacteriophage. Both E. coli and B. subtilis phages use alternate sigma factors for temporal regulation of phage gene expression. B. subtilis phagesSP01and SP82and coliphage T4 all encode three classes of transcripts: early, middle, and late (reviewedin 38, 111, 115). In all cases early transcripts are madeby host RNApolymerase. Production of middle and late transcripts requires phage-specifiedproteins. Transcription of SP01middle genes requires the expression of phage gene 28; appearanceof late transcripts is dependenton gene products 33 and 34. The products of genes 28, 33, and 34 have been isolated and shownto function as sigmafactors (19, 29, 138). Whengp28overproducedin E. coli is reconstituted with either B. subtilis or E. coil core, the holoenzymeformedtranscribes two middlepromoters (19). Production of SP01late RNA in vitro using total SP01 DNA as a template, requires the addition of gp33and 34, as well as the delta protein, to core enzyme(138). The related B. subtilis phage SP82encodes proteins analogous in size and function to gp28, 33, and 34 of SP01. Theseproteins are found associated with RNA polymeraseafter infection by SP82(131) and are likely to also be sigma factors. Achbergerand Whiteley (0) showedthat the 28 kd protein from SP82 functions as a sigma factor and allows both B. subtilis and E. coli core RNA polymerase to recognize SP82 middle promoters. RNApolymerase purified 5-10 minutes after T4 infection contains five




377

phage-encodedproteins (115). Twoof these proteins are the products of genes 33 and 55. The product of gene 55 has been shownto be a sigma factor (66, 88). In vivo transcription from T4 late prrmoters requires both gp33and gp55, but in vitro, purified gp55 and core RNApolymerase isolated from uninfectedE. coli is sufficient for transcription fromthe late promoterP23(66). The function of gp33 has not yet been determined. ~b29is a lytic phage of B. subtilis that encodes two temporal classes of mRNA. Appearanceof late mRNA requires the product of d¢29 gene 4 (reviewedin 38). Melladoet al. (92) haveoverproducedgp4 E. coli. Add ition of the partially purified protein to B. subtilis core RNA polymeraseis required for synthesis of two transcripts from the late region of the ~b29genome.These transcripts are not made when 429 DNAis transcribed by either core polymerasealone or core supplementedwith extracts containing a mutantgp4, suggesting that gp4 is a sigmafactor. Conservation

of Structure

Among Sigma Factors

In the past several years, the sequenceof nine sigmafactors has beenreported: 0.70 (12) and 0.32 (72, 155) fromE. coli; 0.43 (45), .29 (133), and spollAC(36) fromB. subtilis; 0.gp55(46) fromE. coli phageT4; ~rgp28(18) and IT gp34 (17) from B. subtilis phageSP01;and the p4 protein from B. subtilis phage29 (34). The derived amino acid sequences have been compared using a variety of computerprogramsdesigned to identify evolutionary relationships between proteins. The two E. coli sigma factors were the first to be identified as being homologous. Landicket al (72) and Yuraet al (155) foundsubstantial similarity between 0 .32 and the C-terminus of 0.70. Overall, the two proteins share (including identical amino acids and conserved replacement) 44%of their aminoacid residues in this region. Oneregion of 14 aminoacids is completely conserved betweenthe two proteins. In addition, 0.3~ was found to have two reasonably good matchesto the helix-turn-helix motif characteristic of DNA binding proteins. Subsequently, 0.43 from B. subtilis was sequencedand compared to these twoproteins. In additionto similarity in the C-terminus,0 .43 also exhibited similarity to 0.70 in the N-terminus(45). .29 (133) and the spolIAC gene product(47) also exhibit substantial similarity to the C-terminusof 0.70. Morerecently, eight of the sigmaproteins have been comparedby Gribskov& Burgess(47). Their results, presenteddiagramaticallyin Figure 5, showthat all eight of the proteins are homologous.All the proteins except T4 gp55have at least one region strongly resemblingthe helix-turn-helix motif (regions 3 and in Figure 5). Modulation

of Sigma Factor

Use

Changesin sigmafactors have been implicated in temporal regulation during irreversible processes such as sporulation and phagedevelopmentand transient


REZNIKOFFET AL 2

5

4

5

4

Sicjma- 70 SItjma - 45 2

Sigma- :52

2


2

4 5

2

Spo]IAc 4

2

4 2

I

I

I

Sigma - 29

SPOI qp28 SPOI gp34 T4 gp55

I

~00 200 300 400 500 600 Figure 5 Homologies in sigmastructures. Theshadedsection of thc bars indicate the locations of the highly conservedregions amongthe sigmafactors. (This figure used with the permissionof M. Gribskov.)

0

processessuch as the heat shockresponseor the responseto nitrogen limitation. The question remains, however: howdoes the cell control which 0- is used? Several strategies can be imagined, someof which are more appropriate for programmingtransient changes and some for irreversible changes in gene expression. Transient changesin gene expression require that the activity or amountof the alternate sigma be regulated in response to environmental conditions. Amongpossible mechanismsare these: 1. The sigmalevel could be changedby altering its synthesis or degradation in responseto an appropriate signal. Changesin relative geneexpressionwould reflect changesin the ratio of the two sigmas. Preliminary evidence suggests that increased amountsof 0-32 mayaccountfor the increased expressionof heat shock proteins under some conditions (A. Grossman and D. Straus, unpublished observation). 2. The amountor activity of additional transcription factors workingat the level of the regulated promoters could change in response to environmental conditions. These changes wouldchange the effectiveness of an alternative sigma factor. Such a mechanismhas been suggested for activation of the nitrogen-limited genes (S. Kustu, personal communication) and for the activation of 0 -37 (54) and 0-28 (42). Sequential, irreversible changes in gene expression could be accomplished




379

by successive loss of sigma factors. Amongthe possible mechanismsare the following: 1. The preexisting ¢r factors could be displaced by newlysynthesized with higher affinity for core RNApolymerase. Suchsigma factors wouldhave to be synthesized in amounts at least equimolar with existing core RNA polymerase. This may be the mechanismby which 0°29 displaces previous sigma factors (R. Losick, personal communication).Chelmet al (14) compared the ability of phageSP0100gp28 and B. subtilis 0043 to competefor core RNA polymerase.Theyfound that 00gp28 can effectively competewith ~r43 for binding to core. They concluded, however, that the degree of competition found was not sufficient to account for the shutoff of early transcription, and that other factors must be involved in the switch from early transcription by Eo"43 tO "gp28. middletranscription by Eo 2. Theactivity of preexisting 00 factors couldbe controlled by sigma-specific inhibitors synthesized as part of the developmentalprocess. This mechanism is utilized by T4 to prevent 0070function. A 10-kd protein, whichis first synthesizedstarting 10 nfin after infection, antagonizes0070,probablyby bindingto it (132). 3. Alternate sigmafactors maybe unstable. Turningoff the synthesis of an unstable sigmafactor wouldcause its level to decrease and thereby increase the amountof core available to other sigmas. ~r29 is unstable (140-142) and its replacementby a subsequent 00 factor later in sporulation maydependon its instability. 4. Twosigma factors maybind to core simultaneously but only one would direct the templateactivity. This maybe true for the T4sigmafactor 00gv55.It is knownthat O"gp55 can bind to holoenzyme containingcr7° (66, 116), but it is not knownwhetherthis results in displacementof 007o. Role of Sigma in Promoter Recognition Althoughthe sigmasubunit is required for specific binding of RNA polymerase holoenzymeat promoter sequences, howsigmaconfers specificity is unknown. Losick, Pero, and coworkers(72a, 77, 111) suggested that each sigma confers specificity by makingcontacts with both regions of the promoter. This model was based on the observation in B. subtilis that the promoters recognized by E0043 differ from promotersrecognizedby polymeraseswith phageor alternate bacterial ~rs at both the -10 and -35 regions (72a; reviewed in 77). Pero Losickalso consideredthe possibility that each ~r interacts directly only at the -10 region and affects recognition at the -35 region indirectly by inducing a novel conformation in core RNApolymerase. They thought this second model less likely; it seemedunlikely that each of the large numberof sigmafactors potentially available wouldconfer a novel conformation. SequencecomparisonofE. coli 0070 and 003zand B. subtilis 0 -43 with a set of



380

REZNIKOFFET AL

DNAbinding proteins has identified two possible DNAbinding regions in the C-terminus of the sigma factors (45, 47, 72, 155). The existence of two DNA binding domainswouldbe consistent with the interaction of sigmas with both regions of the promoters. The fact that 0-70 crosslinks at both regions of the promoter (15, 16, 129) is also consistent with this model. However, the ¯ crosslinks in the -35 region haveonly beenobtained with partially depurinated DNA,so these results maynot be biologically relevant. The following observations, on the other hand, would be consistent with sigmafactors recognizing only the - 10 region of the promoter: 1. The promotersrecognizedby E. coli Eo32 and Eo7° differ dramatically in the -10 region but are similar in the -35 region (20). 2. B. subtilis 0 -29 has only one region resemblingthe DNA binding region of DNAbinding proteins (47, 133). 3. crgp55from E. coli phage T4 recognizes promoters that have a consensus sequence only in the -10 region of the promoter (31). 4. The promotersrecognized by the ntrA gene product lack a conserved - 35 region, having instead two conserved regions centered around -10 and -20 and separated by 3 base pairs (27, 59, 119). Becauseof these observations, Cowinget al (20) recently proposed a newversion of the class of models which -35 region contacts are madeby the core subunits of holoenzyme,not by ~r. In this model,each 0- confers specificity to holoenzymeby interacting directly with the - 10 region. In’addition, the different size and shapeof each 0factor could alter the precise region of holoenzyme contacting the - 35 region. This altered geometry of the holoenzyme-DNA complex could lead to differences in the spacing betweenthe conservedsequencesand in the sequencein the -35 region recognized by holoenzyme. Accordingto the model, the - 10 regions of consensus promoters should be sufficiently different to account for the discrimination by various forms of holoenzyme. The -35 regions, recognized by subunits commonto each holoenzyme, could be more similar than the -10 regions recognized by different sigmas, but such similarity is not required. Analysisof the available consensussequencesindicate that in general they conformto these expectations (Table 1). The consensus - 10 regions of promoters recognized by alternate forms of holoenzyme all lack the highly conserved TA---T sequence characteristic of E0- 70 and Ecr 43 promoters, whichmakesit unlikely that 70 E0could interact with these sequences. The sequencesin the -35 regions are more similar to the conserved TTGbases in the -35 regions of Eo"7° and "43 Eo promoters. It is clear that comparisonof consensus sequences recognized by alternate forms of holoenzymecannot settle the issue of which DNAsequences are recognized by sigma. Additional approaches, such as mutational analysis of

Annual Reviews www.annualreviews.org/aronline TRANSCRIPTION INITIATION various sigmas and crosslinking answer this question.

studies

REGULATION IN BACTERIA to the closed

381

complex, are needed to

ACKNOWLEDGMENTS We wish to acknowledge the present and past members of our laboratories who have contributed to our understanding of transcription initiation. The research in our laboratories was supported by NIH grant GM-19670 and a 3M Foundation grant to W.S.R. and NIH grants AI 19635 and GM 28575 to C.A.G. Annu. Rev. Genet. 1985.19:355-387. Downloaded from arjournals.annualreviews.org by University of Wisconsin - Madison on 03/29/07. For personal use only.

C.A.G. was also

a recipient

of an NIH research

career

development award (AI

00573). Literature Cited 0. Achberger,E. C., Whiteley, H. R. 1980. The interaction of Escherichia coli core RNApolymerase with specificitydetermining subunits derived from unmodified and SP82-modified Bacillus subtilis RNApolymerase. J. Biol. Chem. 255:11957-64 1. Ames, G. F.-L., Nikaido, K. 1985. Nitrogen regulation in Salmonella typhimurium.Identification of an ntrC protein-binding site and definition of a consensus binding sequence. EMBOJ. 4:539-47 2. Banner, C. D. B., Moran, C. P. Jr., Losick, R. 1983. Deletion analysis of a complex promoter for a developmentally regulated gene from Bacillus subtilis. J. Mol. Biol. 168:351-65 3. Barbeyron, T., Kean, K.. Forterre, P. 1984. DNAadenine methylation of GATCsequences appeared recently in the Escherichiacoli lineage. J. Bacteriol. 160:586-90 3a. Bauer, C. E., Carey, J., Kasper, L. M., Lynn, S. P., Waechter, D. A., Gardner, J. F. 1983. Attenuation in bacterial operons. In GeneFunction in Prokarvotes, ed. J. Beckwith, J. Davies, J. G~]llant, pp. 65-89. Cold Spring Harbor: Cold Spring Harbor Laboratory. 328 pp. 4. Beckwith,J., Grodzicker,T., Arditti. R. 1972. Evidencefor two sites in the lac promoter region. J. Mol. Biol. 69:15560 5. Beckwith, J. R., Zipser, D., eds. 1970. The Lactose Operon. Cold Spring Harbor, NY:Cold Spring Harbor Lab. 437 pp. 6. Bossi, L., Smith, D. M. 1984. Conformational change in the DNAassociated with an unusual promoter mutation in a tRNAoperon of Sahnonella. Cell 39:643-52 6a. Botchan, P. 1976. Electron-microscopic comparisonof transcription on lin-

ear and superhelical DNA.J. Mol. Biol. 105:161-76 7. Briar, J.-F., Gilman, M., Chamberlin, M.J. 1985. Bacillus subtilis cr2~ and Escherichia coli ~32 (htpR) are minor cr factors that display an overlapping promoter specificity. J. Biol. Chem. 260:2038-41 8. Bujard, H., Niemann, A., Brennig, K., Roisch, U., Dresel, A., et al. 1982. The interaction of E. coli RNApolymerase with promotersof high signal strength. In Promoters: Structure and Function, ed. R. L. Rodriguez, M. J. Chamberlin, pp. 121-40. NewYork: Praegcr. 524 pp. 9. Burgess, R. R. 1971. RNApolymerase. Ann. Rev. Biochem. 40:711--40 10. Burgess, R. R./Travers, A. A. 1970. Escherichia coli RNApolymerase: Purification, subunit structure, and factor requirements. Fed. Proc. 29:116469 I 1. Burgess, R. R., Travers, A. A.. Dunn,J. J.. Bautz, E. K. F. 1969. Factor stimulating transcription by RNApolymerase. Natnre 221:43 12. Burton, Z., Burgess, R. R., Lin, J., Moore, D., Holder, S., Gross, C. A. 1981. The nucleotidc sequence of the cloned rpoD gene for the RNA polymerase sigma subunit from E. coli KI2. Nucleic Acids Res. 9:2889-2903 13. Deleted in proof. 14. Chelm,B. K., Duffy, J. J., Geiduschek, E. P. 1982. Interaction of Bacillus subtilis RNApolymerase core with two specificity-determining subunits. Compctition between~r aud the SP01gent 28 protein. J. Biol. Chem. 257:6501-8 15. Chenchik. A., Bibilashvili, R., Mirzabckov, A. 1981. Topography of interaction of Escherichia coli RNA polymcrasc subunits with /acUV5promoter. FEBSLett. 128:46-50



382

REZNIKOFF ET AL

16. Chenchik,A., Bibilashvili, R. Sh., Mirzabekov, A. D., Shik, V. V. 1982. Contact betweensubunits of Escherichiacoli RNApolymerase and the nucleotides of the lacUV5 promoter. Molekul. Biol. 16:26-36 17. Costanzo, M., Brzustowicz, L., Hannett, N., Pero, J. 1984. Bacteriophage SP01genes 33 and 34. Location and primarystructure of genes encodingregulatory subunits of Bacillus subtilis RNA polymerase. J. Mol. Biol. 180:533-47 18. Costanzo,M., Pero, J. 1983. Structureof a Bacillus subtilis bacteriophage SP01 gene encoding a RNApolymeraseo" factor. Proc. Natl. Acad. Sci. USA80:123640 19. Costanzo, M., Pero, J. 1984. Overproductionand purification of a bacteriophage SP01-encoded RNApolymerase sigma factor. J. Biol. Chem.259:668185 20. Cowing, D. W., Bardwell, J. C. A., Craig, E. A., Woolford,C., Hendrix, R. W., Gross, C. A. 1985. Consensus sequence for Escherichia coli heat shock gene promoters. Proc. Natl. Acad. Sci. USA 82:2679-83 21. Christensen, A. C., Young,E. T. 1982. T4late transcripts are initiated near a conserved DNAsequence. Nature 299:369-71 22. de Bruijn, F. J., Ausubel, F. M. 1983. The cloning and characterization of the glnF (ntrA) gene of Klebsiella pneumoniae: Role of glnF (ntrA) in the regulation of nitrogen fixation (n~ and other nitrogen assimilation genes. Mol. Gen. Genet. 192:342-53 23. de Crombrugghe,B. D., Busby, S., Buc, H. 1984. Activation of transcription by the cyclic AMPreceptor protein. In Biological Regulation and Development, ed. K. Yamamoto, 6B:129-67. New York: Plenum 24. deCrombrugghe, B. D., Chen, B., Anderson, W., Nissley, P., Gottesman, M., et al. 1971. RNApolymerase and cyclic AMPreceptor protein, cyclic AMP,lac repressor and inducer are the essential elementsfor controlled lac transcription. Nature New Biol. 231:139-42 25. Dickson, R. C., Abelson, J., Barnes, W. M., Reznikoff, W. S. 1975. Genetic regulation: The lac control region. Science 187:27-35 26. Dickson, R. C., Abelson, J., Johnson, P., Reznikoff, W. S., Barnes, W. M. 1977. Nucleotide sequence changes producedby mutationsin the lac promoterof Escherichia coli. J. Mol. Biol. 111:6575

27. Dixon, R. 1984. Tandempromoters determine regulation of the Klebsiella pneumoniaeglutaminesynthetase (g lnA gene. Nucleic Acids Res. 12:7811-30 28. Drumond, M., Clements, J., Merrick, M., Dixon,R. 1983. Positive control and autogenousregulation of the n/fl~A promoter in Klebsiella pneumoniae.Nature 301:302-7 29. Duffy, J. J., Geiduschek, E. P. 1977. Purification of a positive regulatorysubunit from phage SP01-modified RNA polymerase. Nature 270:28-32 30. Dunn, T. M., Hahn, S., Ogden, S., Schleif, R. F. 1984. Anoperator at -280 base pairs that is required for repression of araBADoperon promoter: Addition of DNAhelical turns between the operator and promoter cyclically hinders repression. Proc. Natl. Acad. Sci. USA81: 5017-20 30a. Ebright, R. H. 1982. Sequence homologies in the DNAof six sites knownto bind to the catabolite geneactivator protein of Escherichia coli. In Molecular Structure and Biological Activity, ed. J. F. Griffin, W.L. Duax, pp. 91-100. NewYork: Elsevier 421 pp. 31. Elliott, T., Geiduschek,E. P. 1984. Defining a bacteriophageT4late promoter: Absenceof a "- 35" region. Cell 36:21119 32. Engbaek, F., Gross, C. A., Burgess, R. R. 1976. Quantitation of RNApolymerase subunits in Escherichiacoli during exponential growth and after bacteriophage T4 infection. Mol. Gen. Genet. 143:291-95 33. Eron, L., Block, R. t971. Mechanismof initiation and repression of in vitro transcription of the lac operon of Escherichia coli. Proc. Natl. Acad. Sci. USA 68:1828-32 34. Escarmis, C., Salas, M. 1982. Nucleotide sequenceof the early genes 3 and 4 of bacteriophage ~b29. Nucleic Acids. Res. 10:5785 35. Fitz-James, P., Young, E. 1969. Morphology of sporulation. In The Bacterial Spore, ed. G. Gould, A. Hurst, pp. 39-72. NewYork: Academic. 724 pp. 36. Fort, P., Piggot, P. J. 1984. Nucleotide sequence of sporulation locus spollA in Bacillus subtilis. J. Gen. Microbiol. 130:2147-53 37. Fox, T. D., Losick, R., Pero, J. 1976. Regulatory gene 28 of bacteriophage SP01codes for a phage-induced subunit of RNApolymerase. J. Mol. Biol. 101:427-33 37a. Friedman, D., Gottesman, M. 1983.



TRANSCRIPTION

INITIATION

Lytic modeof lambda development. In Lambda H, ed. R. W. Hendrix, J. W. Roberts, F. W. Stahl, R. A. Weisberg, pp. 21-51. Cold Spring Harbor: Cold Spring Harbor Laboratory. 694 pp. 38. Geiduschek,E. P., Ito, J. 1982. Regulatory mechanismsin the development of lytie bacteriophagesin Bacillus subtilis. In The MolecularBiology of the Bacilli, ed. D. A. Dubnau, 1:203-46. NewYork: Academic. 378 pp. 39. Gellert, M. 1981. DNAtopoisomerases. Ann. Rev. Biochem. 50:879-910 40. Gilbert, W. 1976. Starting and stopping sequences for the RNApolymerase. In RNA Polymerase, ed. R. Losick, M. Chamberlin, pp. 193-206. Cold Spring Harbor, NY: Cold Spring Harbor Lab. 899 pp. 41. Gilbert, W., Gralla, J., Majors, J., Maxam, A. 1975. Lactose operator sequences and the action of lac repressor. In Protein-Ligand Interactions, ed. I-I. Sund, G. Blauer, pp. 193-210. Berlin: Gruyter 42. Gilman, M. Z., Chamberlin, M. J. 1983. Developmentaland genetic regulation of Bacillus subtilis 28 genestranscribed by ~r -RNApolymerase. Cell 35:285-93 43. Gilman, M. Z., Glenn, J. S., Singer, V. L., Chamberlin,M. J. 1984. Isolation of DNAsequences from Bacillus subtilis that contain promoters for sigma-28RNApolymerase. Gene 32:11-20 44. Gilman, M. Z., Wiggs, J. L., Chamberlin, M. J. 1981. Nucleotide sequences of two Bacillus subtilis promoters used by Bacillus subtilis sigma-28 RNA polymerase. Nucleic Acids Res. 9:59916000 45. Gitt, M. A., Wang, L.-F., Doi, R. H. 1985. A strong sequence homology exists between the major RNApolymerase sigmafactors of Bacillus subtilis and Escherichia coli. J. Biol. Chem.260:717885 46. Gram, H., Ruger, W. 1985. Genes 55, etgt, 47 and 46 of bacteriophage T4: The genomicorganization as deduced by sequence analysis. EMBO J. 4:257-64 47. Gribskov, M., Burgess, R. R. 1985. Sigmafactors from Escherichia coli, Bacillus subtilis, phage SP01, and T4 are homologous proteins with conserved secondary structures. In press 48. Gross, C. A., Grossman, A. D., Liebke, H., Waiter, W., Burgess, R. R. 1984. Effects of the mutant sigma allele rpoD800on the synthesis of specific maeromolecular components of the Escherichia coli K12cell. J. Mol. Biol. 172:283-302 49. Grossman, A. D., Cowing, D., Erickson,

REGULATION IN BACTERIA

383

J., Baker, T., Zhou, Y. N., Gross, C. 1985. Analysis of the Escherichia coli heat shock response. Microbiology. In press 50. Grossman, A. D., Erickson, J. W., Gross, C. A. 1984. The htpR gene product of E. coli is a sigmafactor for heatshock promoters. Cell 38:383-90 51. Guarente, L., Nye, J. S., Hochschild, A., Ptashne, M. 1982. Mutant h phage repressorwitha specificdefect in its positive control function. Proc. Natl. Acad. Sci. USA 79:2236-39 52. Haldenwang, W. G., Lang, N., Losick, R. 1981. A sporulation-induccd sigmalike regulatory protein fromB. subtilis. Cell 23:615-24 53. Haldenwang,W. G., Losick, R. 1979. A modified RNApolymerase transcribes a cloned gene under sporulation control in Bacillus subtilis. Nature 282:256-60 54. Haldenwang,W. G., Losick, R. 1980. A novel RNApolymerase sigma factor fromBacillus subtilis. Proc. Natl. Acad. Sci. USA77:700(~4 55. Harris, J. D., Heilig, J. S., Martinez,I. 1., Calendar, R., Isaksson, L. A. 1978. Temperature-sensitive Escherichia coli mutant producing a temperature-sensitive ~r-subunit of DNA-dependentRNA polymerase. Proc. Natl. Acad. Sci. USA 75:6177-81 56. Haugh,G. W., Squires, C. H., DeFelice, M., Lago, C. T., Calvo, J. M. 1985. Unusual organization of the ilvlH promoter of Escherichiacoli. J. Bacteriol. 163:186-98 57. Hawley, D. K., McClure, W. R. 1983. Compilationand analysis of Escherichia coli promoter DNAsequences. Nucleic Acids Res. 11:2237 55 58. Hay, R. E., Moran, C. P. Jr. 1985. Effects of base substitutions in a complex promoterfrom Bacillus subtilis. J. Bacteriol. In press 59. Hirschman, J., Wong,P.-K., Sei, K., Keener, J., Kustu, S. 1985. Products of nitrogen regulatory genes ntrA and ntrC of enteric bacteria activate glnA transcription in vitro: Evidencethat the ntrA product is a sigma factor. Proc. Natl. Acad. Sci. USA:In press 60. Hochschild, A., Irwin, N., Ptashne, M. 1983. Repressor structure and the mechanismof positive control. Cell 32:319-25 61. Irani, M. H., Orosz, L., Adhya,S. 1983. A control element within a structural gene: Thegal operonof Escherichiacoli. Cell 32:783-88 62. Iwakura, Y., Ishihama, A. 1975. Biosynthesis of RNApolymerase in Escherichia coli II. Control of RNA



384

REZNIKOFF

ET AL

polymerasesynthesis during nutritional shift up and down. Mol. Gen. Genet. 142:67-84 63. Jacob, F., Monod,L 1961. Genetic regulatory mechanismsin the synthesis of proteins. J. Mol. Biol. 3:318-56 64. Jaehning, J. A., Wiggs,J. L., Chamberlin, M. J. 1979. Altered promoter selection by a novel form of Bacillus subtilis RNApolymerase. Proc. Natl. Acad. Sci. USA 76:5470-74 65. Johnson, W. C., Moran, C. P., Losick, R. 1983. Two RNApolymerase sigma factors from Bacillus subtilis discriminate between overlapping promoters for a developmentallyregulated gene. Nature 302:800-4 66. Kassavetis, G. A., Geiduschek, E. P. 1984. Defining a bacteriophage T4 late promoter: Bacteriophage T4 gene 55 protein suffices for directing late promoter recognition. Proc. Natl. Acad. Sci. USA 8l:510l-5 67. Kleckner, N., Morisato, D., Roberts, D., Bender, J. 1984. Mechanism and regulation of Tnl0 transposition. Cold Spring Harbor Symp. Quant. Biol. 49: 235-44 68. Krajewska-Grynkiewicz, K., Kustu, S. 1983. Regulation of transcription of glnA, the structural gene encodingglutaf, mine synthetase, in g/nA::Mudl (Ap lac) fusion strains of Salmonellatyphimurium. Mol. Gen. Genet. 192:187-97 69. Krajewska-Grynkiewicz, K., Kustu, S. 1984. Evidencethat nitrogen regulatory gene ntrC of Salmonella typhimuriumis transcribed from the glnA promoter as well as from a separate ntr promoter. Mol. Gen. Genet. 193:135-42 70. Kustu, S., Sei, K., Keener, J. 1985. Nitrogen regulation in enteric bacteria. 39th Symp. Soc. for General Microbiology, ed. I. R. Booth, C. F. Higgins. Cambridge Univ. Press. In press 71. Lamond, A. I., Travers, A. A. 1983. Requirementfor an upstream element for optimal transcription of a bacterial tRNA gene. Nature 305:248-51 72. Landick, R., Vaughn, V., Lau, E. T., VanBogelen, R. A., Erickson, J. W., Neidhardt, F. C. 1984. Nucleotide sequence of the heat shock regulatory gene of E. coli suggests its protein product may be a transcription factor. Cell 38:175-82 72a. Lee, G., Pero, J. 1981. Conserved nucleotide sequences in temporally controlled bacteriophagepromoters. J. Mol. Biol. 152:247-65 73. Lee, N. 1978. Molecular aspects of ara regulation. See Ref. 95, pp. 389-409

74. Lencastre, H. de, Piggot, P. J. 1979. Identification of different sites of expression for spo loci by transformation of Bacillus subtilis. J. Gen. Microbiol. 114:377-89 75. Liebke, H., Gross, C., Walter, W., Burgess, R. 1980. A new mutation rpoD800,affecting the sigma subunit of E. coli RNApolymerase,is allelic to two other sigma mutants. Mol. Gen. Genet. 177:277-82 76. Linn, T., Losick, R., Sonenshein, A. L. 1975. Rifampin resistance mutation of Bacillus subtilis altering the electrophoretic mobility of the beta subunit of ribonucleic acid polymerase. J. Bacteriol. 122:1387-90 77. Losick, R., Pero, J. 1981. Cascades of sigma factors. Cell 25:582-84 78. Losick, R., Sonenshein, A. L. 1969. Change in the template specificity of RNApolymerase during sporulation of Bacillus subtilis. Nature 224:35 37 79. Losick, R., Youngman,P. 1984. Endospore formation in Bacillus. In Microbial Development,ed. R. Losick, L. Shapiro, pp. 63-88. Cold Spring Harbor, NY: Cold Spring Harbor Lab. 303 pp. 80. McClure, W. R. 1980. Rate-limiting steps in RNA chain initiation. Proc. Natl. Acad. Sci. USA 77:5634-38 81. McClure, W. R. 1983. In Biochemistry of Metabolic Processes, ed. D. L. F. Lennon, F. W. Stratman, R. N. Zahlten, pp. 207-17. NewYork: Elsevier 82. McClure, W. R., Hawley, D. K., Malan, T. P. 1982. The mechanism of RNA polymerase activation on the hPRMand lacP+ promoters. See Ref. 8, pp. 11120 83. Magasanik, B. 1982. Genetic control of nitrogen assimilation in bacteria. Ann. Rev. Genet. 16:135-68 84. Majors, J. 1975. Initiation of in vitro mRNA synthesis from the wild-type lac promoter. Proc. Natl. Acad. Sci. USA 72:4394-98 85. Majors, J. 1975. Specific binding of CAP factor to lac promoter DNA.Nature 256:672-74 86. Majumdar, A., Adhya, S. 1984. Demonstration of tw~ooperator elementsin gal: In vitro repressor binding studies. Proc. Natl. Acad. Sci. USA 81:6100-4 87. Malan, T. P., Kolb, A., Buc, H., McClure, W. R. 1984. Mechanism of CRP-cAMPactivation of lac operon transcription initiation activation of the P1 promoter. J. Mol. Biol. 180:881-909 88. Malik, S., Dimitrov, M., Goldfarb, A. 1985. Initiation of transcription by bacteriophage T4 modified RNApolymerase



TRANSCRIPTION

INITIATION

independently of host sigma factor. J. Mol. Biol. In press 89. Mandccki, W., Caruthers, M. H. 1984. Mutants of the lac promoter with large insertions and deletions betweenthe CAP binding site and the -35 region. Gene 31:263-67 90. Maquat, L. E., Reznikoff, W. S. 1980. lac Promotermutation Prl 15 generates a newtranscription initiation point. J. Mol. Biol. 139:551-56 91. Maquat, L. E., Thornton, K., Reznikoff, W. 1980. lac Promoter mutations located downstreamfrom the transcription start site. J. Mol. Biol. 139:53749 92. Mellado, R. P., Carroscosa, J. L., Salas, M. 1985.Control of the late transcription of the B. subtilis phage ~b29 DNA.In Sequence Specificity in Transcription and Translation. UCLASymp. Mol. Cell. Biol., (NS), Vol. 30, ed. R. Calendar, Gold. NewYork: Liss, In press 93. Menzel, R., Gellert, M. 1983. Regulation of the genes for E. coli DNA gyrase: homeostatic control of DNAsupercoiling. Cell 34:105-13 94. Merrick, M. J. 1982. A new model for nitrogen control. Nature 297:362-63 95. Miller, J. H., Reznikoff, W. S., eds. 1978. The Operon. Cold Spring Harbor, NY: Cold Spring Harbor Lab. 449 pp. 96. Moran, C. P. Jr., Johnson, W. C., Losick, R. 1982. Close contacts between 0-37-RNA polymeraseand a Bacillus subtilis chromosomal promoter. J. Mol. Biol. 162:709-13 97. Moran,C. P. Jr., Lang, N., Banner, C. D. B., Haldenwang, W. G., Losick, R. 1981. Promoter for a developmentally regulated gene in Bacillus subtilis. Cell 25:783-91 98. Moran,C. P. Jr., Lang, N., LeGrice, S. F. J., Lee, G., Stephens,M., etal. 1982. Nucleotide sequences that signal the initiation of transcription and translation in Bacillus subtilis. Mol. Gen. Genet. 186:33946 99. Moran, C. P. Jr., Lang, N., Losick, R. 1981. Nucleotide sequence of a Bacillus subtilis promoter recognized by Bacillus subtilis RNApolymerasecontaining "37. 0 Nucleic Acids Res. 9:5979-90 100. Mossing,M. C., Record, M. T. Jr. 1985. Thermodynamic origins of specificity in the lac repressor operator interaction: Adaptability in the recognition of mutant operator sites. J. Mol.Biol. In press 101. Mulligan, M. E., Hawley, D. K., Entriken, R., McClure, W. R. 1984. Escherichia coli promoter sequences predict in vitro RNApolymeraseselectivity. Nucleic Acids Res. 12:789-800


385

102. Nakamura, Y. 1978. RNApolymerase mutant with altered sigma factor in Escherichia coll. Mol. Gen. Genet. 165:16 103. Nakamura, Y., Osawa, T., Yura, T. 1977. Chromosomal location of a structural gene for the RNApolymerase 0- factor in Escherichiacoli. Proc. Natl. Acad. Sci. USA 74:1831-35 104. Nakayama,T., Williamson, V., Burtis, K., Doi, R. 1978. Purification and propcrtics of two RNApolymerases from sporulatingcells of Bacillussubtilis. Eur. J. Biochem. 88:155-64 105. Neidhardt, F. C., VanBogelen, R. A. 1981. Positive regulatory gene for temperature-controlled proteins in Escherichia coli. Biochem.Biophys. Res. Commun. 100:894-900 106. Neidhardt, F. C., VanBogelen, R. A., Vaughn,V. 1984. The genetics and regulation of heat-shock proteins. Ann. Rev. Genet. 18:295-329 107. Nieuwkoop,A. J., Boylan, S. A., Bender, R. A. 1984. Regulation of hutUH operon expression by the catabolite gene activator protein-cyclic AMP complexin Klebsiella aerogenes. J. Bacteriol. 159:934-39 108. Ollington, J. F., Haldenwang, W. G., Huynh, T. V., Losick, R. 1981. Developmentally-regulatedtranscription in a cloned segmentof the Bacillus subtilis chromosome. J. Bacteriol. 147:43242 109. Osawa, T., Yura, T. 1981. Effects of rcduccd amount of RNApolymerase sigmafactor on gene expression and growth of Escherichia coli: Studies of the rpoD40 (amber) mutation. Mol. Gen. Genet. 184:166-73 110. Perlman, R. L., Pastan, I. 1969. Pleiotropic deficiency of carbohydrate utilization in an adenyl cyclase deficient mutant of Escherichia coli. Biochem. Biophys. Res. Commun.37:151-64 111. Pero, J., Lee, G., Moran, C. P. Jr., Lang, N., Losick, R. 1982. Promoters controlled by novel sigma factors in Bacillus subtilis. See Ref. 8, pp. 283-92 112. Peterson, M. L., Reznikoff, W. S. 1985. In vivo and in vitro properties of lac P2, an overlapping RNApolymerase binding site within the lactose promoter.J. Mol. Biol. In press 113. Peterson, M. L., Reznikoff, W. S. 1985. The lactose promoter mutation Prll5 activates an overlapping promoterwithin the lactose control region. J. Mol. Biol. In press 114. Piggot, P. J., Coote, J. G. 1976. Genetic aspects of bacterial endosporeformation. Bacteriol. Rev. 40:908-62



386

REZNIKOFF ET AL

115. Rabussay, D. 1983. Phage-evoked changes in RNApolymerase. In Bacteriophage T4, ed. C. K. Mathews,E. M. Kutter, G. Mosig, P. B. Berget, 167-73. Washington DC: Am. Soc. Microbiol. 116. Ratner, D. 1974. The interaction of bacterial and phage proteins with immobilized Escherichia coil RNApolymerase. J. Mol. Biol. 88:373-83 117. Ray, C., Hay, R. E., Carter, H. L., Moran, C. P. Jr. 1985. Mutations that affect utilization of a promoterin stationary phaseBacillus subtilis. J. Bacteriol. In press 118. Reitzer, L. J., Magasanik, B. 1983. Isolation of the nitrogen assimilation regulator NRI,the product of the glnGgene of Escherichia coll. Proc. Natl. Acad. Sci. USA 80:5554-58 119. Reitzer, L. J., Magasanik,B. 1985. Expression of glnA in Escherichia coil is regulated at tandem promoters. Proc. Natl. Acad. Sci. USA82:1979-83 120. Reznikoff, W. S. 1984. Gene expression in microbes: The lactose operon model system. In The Microbe 1984: Part H, Prokamotes and Eukaryotes, ed. D. P. Kelly, N. G. Carr, pp. 195-218. Cambridge: CambridgeUniv. Press. 349 pp. 121. Reznikoff, W. S., Maquat, L. E., Munson, L; M., Johnson, R. C., Mandecki, W. 1982. The lac promoter: Analysis of structural signals for transcription initiation, and identification of a new sequence-specific event. See Ref. 8, pp. 80-95 122. Reznikoff, W. S., Winter, R. B., Hurley, C. K. 1974. The location of the repressor binding sites in the lac operon. Proc. Natl. Acad. Sci. USA71:2314-18 122a. Roberts, J. W.., Devoret, R. 1983. Lysogenic induction. See Ref. 37a pp. 123-44 123. Rothstein, D. M., Pahel, G., Tyler, B., Magasanik, B. 1980. Regulation of expression from the glnA promoter of Escherichia coil in the absenceof glutamine synthetase. Proc. Natl. Acad. Sci. USA 77:7372-76 124. Ryter, A., Whitehouse, R. 1978. Uracil incorporation in the forespore and the mother cell during spore developmentin Bacillus subtilis. Arch. Microbiol. 118:27-34 125. Sanzey, B. 1979. Modulation of gene expression by drugs affecting deoxyribonucleic acid gyrase. J. Bacteriol. 138:40-47 126. Schmitz, A. 1981. Cyclic AMPreceptor protein interacts with lactose operator DNA.Nucleic Acids Res. 9:277-92 127. Schwartz, D., Beckwith, J. R. 1970.

Mutants missing a factor necessary for the expression of catabolite-sensitive operons in E. coll. See Ref. 5, pp. 417-22 128. Siebenlist, U., Simpson,R. B., Gilbert, W. 1980. E. coil RNApolymerase interacts homologouslywith two different promoters. Cell 20:269-81 129. Simpson, R. B. 1979. The molecular topography of RNApolymerase-promoter interaction. Cell 18:277-85 130. Simpson,R. B. 1980. Interaction of the cAMPreceptor protein with the lac promoter. Nucleic Acids Res. 8:759-66 131. Spiegelman, G. B., Whiteley, H. R. 1978. Bacteriophage SP82induced modifications of Bacillus subtilis RNA polymeraseresult in the recognition of additional RNAsynthesis initiation sites on phage DNA.Biochem. Biophys. Res. Commun.81:1058-65 131a. Sternberg, N. 1985. Adenine methylation and gene function in procaryotes. J. Bacteriol. In press 132. Stevens, A. 1976. A salt-promoted inhibitor of RNApolymeraseisolated from T4 phage-infected E. coll. See Ref. 40, pp. 617-28 133. Stragier, P., Bouvier, J., Bonamy,C., Szulmajster, J. 1984. A developmental gene product of Bacillus subtilis homologousto the sigma factor of Escherichia coll. Nature 312:376-78 133a. Sun, D., Takahashi, I. 1982. Genetic mappingof catabolite-resistant mutants of Bacillus subtilis. Can. J. Microbiol. 28:1242-51 134. Deleted in proof 135. Tatti, K. M., Moran, C. P. Jr. 1984. Promoter recognition by sigma-37 RNA polymerase from Bacillus subtilis. J. Mol. Biol. 175:285-97 136. Tatti, K. M., Moran, C. P. Jr. 1985. Utilization of one promoterby two forms of RNApolymerase from Bacillus subtills. Nature 314:190-92 137. Taylor, W. E., Straus, D. B., Grossman, A. D., Burton, Z. F., Gross, C. A., Burgess, R. 1984. Transcription from a heat-inducible promoter causes heat shock regulation of the sigmasubunit of E. coil RNApolymerase. Cell 38:37181 138. Tjian, R., Pero, J. 1976. Bacteriophage SP01regulatory proteins directing late gene transcription in vitro. Nature 262:753 57 139. Tjian, R., Stinchcomb, D., Losick, R. 1975. Antibodydirected against Bacillus subtilis ~r factor purified by sodium dodecylsulfate slab gel electrophoresis: Effect on transcription by RNA polymerasein crude extracts of vegeta-

Annual Reviews www.annualreviews.org/aronline TRANSCRIPTION

INITIATION


143. 144.

145.

146.

147.

387

148. Wiggs, J. L., Gilman, M. Z., Chamberlin, M. J. 1981. Heterogeneity of RNA polymerases in Bacillus subtilis: Evidence for an additional sigma factor in vegetative cells. Proc. Natl. Acad. Sci. USA 78:2762-66 149. Wong,P., Keener, J., Sei, K., Hirschman, J., Kustu, S. 1985. Binding sites for ntrC protein in the glnA promoterL., Haldenwang, W. G. 1985. P , a regulatory region. Abstr. Ann. Meet. Am. Soc. Microbiol. p. 110 0-29-like RNApolymerase binding protein of Bacillus subtilis. In MolecularBi150. Wong,S. L., Price, C. W., Goldfarb, D. ology of Microbial Differentiation/ S., Doi, R. H. 1984. The subtilisin E gene of Bacillus subtilis is transcribed Spores, Vol. 9, ed. J. A. Hoch, P. Setfroma 0-37 promoterin vivo. Proc. Natl. low. Washington, DC: Am.Soc. Microbiol. Acad. Sci. USA81:1184-88 Trempy, J. E., Morrison-Plummer, J., 151. Yamamori,T., Yura, T. 1982. Genetic Haldenwang, W. G. 1985. Synthesis of control of heat-shock protein synthesis and its bearing on growth and thermal 0-29, an RNApolymerasespecificity determinant, is a developmentallyregulated resistance in Escherichia coli K-12. Proc. Natl. Acad. Sci. USA79:860~4 event in Bacillus subtilis. J. Bacteriol. 161:340-46 151a. Youderian, P., Bouvier, S., Susskind, Tyler, B. 1978. Regulation of the M. M. 1982. Sequence determinaats of assimilation of nitrogen compounds. promoter activity. Cell 30:843-53 Ann. Rev. Biochem. 47:1127-62 152. Yu, X.-M., Reznikoff, W. S. 1984. DeUeno-Nishio, S., Backman, K. C., letion analysis of the CAP-cAMP binding Magasanik, B. 1983. Regulation at the site of the Escherichiacoli lactose proglnL-operator-promoter of the complex moter. Nucleic Acids Res. 12:5449-64 glnALGoperon of Escherichia coli. J. 153. Yu, X.-M., Reznikoff, W. S. 1985. DeBacteriol. 153:1247-51 letion analysis of RNApolymerase inWang, P.-Z., Doi, R. H. 1984. Overteraction sites in the Escherichiacoli laclapping promoters transcribed by Baciltose operon regulatory region. J. Mol. lus subtilis 0-~s and O"37 RNA polymerase Biol. Submitted holoenzymesduring growth and station154. Yu, X.-M., Reznikoff, W. S. 1985. Deary phases. J. Biol. Chem. 259:8619-25 letion analysis of the Escherichiacoli lacWong,S.-L., Doi, R. H. 1984. Utilizatose promoter P2. Nucleic Acids Res. tion of a Bacillus subtilis ~37promoterby 13:2457-68 Escherichia coli RNApolymerase in 155. Yura, T., Tobe, T., Ito, K., Osawa,T. vivo. J. Biol. Chem. 259:9762-67 1984. ~eat shock regulatory gene (htpR) Wiggs, J. L. 1981. An RNApolymerase of Escherichiacoli is required for growth fromBacillus subtilis bearinga novel sigat high temperaturebut is dispensable at masubunit. PhDthesis. Univ. Calif., low temperature. Proc. Natl. Acad. Sci. Berkeley USA 81:6803-07

tive and sporulating cells. J. Biol. Chem. 250:8824-28 140. Trempy, J. E., Bonamy,C., Szulmajster, J., Haldenwang,W. G. 1985. Bacillus subtilis sigmafactor 0-29is the product of the sporulation-essential gene spollG. Proc. Natl. Acad. Sci. USA82:4189-92 141. Trempy, In pressJ. E., LaBe|l, T. L., Ra,Y3,~ G.

142.




The Regulation of Transcription Initiation in Bacteria

The Regulation of Transcription Initiation in Bacteria

Suggest Documents

Regulation by transcription factors in bacteria - BioMedSearch

Regulation of the human general transcription initiation factor TFIIF by ...

Regulation by transcription factors in bacteria: beyond ... - Google Sites

Regulation of WUSCHEL Transcription in the Stem

Regulation of transcription termination in the nematode

Multiple steps in the regulation of transcription

Mechanistic differences in transcription initiation at ...

Post-transcription initiation function of the ubiquitous SAGA complex in ...

Generality of the Branched Pathway in Transcription Initiation by

Transcription Initiation Activity Sets Replication

transcription initiation: the effectof supercoiling and ... - BioMedSearch

A Selector of Transcription Initiation in the Protozoan ... - Europe PMC

Dissecting the stochastic transcription initiation ...

TRANSCRIPTION REGULATION IN DIFERENTIAL ...

In vitro transcription initiation of the spinach chloroplast 16S rRNA ...

Genomic Organization of Human Transcription Initiation ... - PLOS

Regulation of polypeptide chain initiation and activity of initiation factor ...

Orthologous Transcription Factors in Bacteria ... - Semantic Scholar

Developmental Regulation of Transcription of the ... - ScienceDirect

Regulation of genes in Streptomyces bacteria ...

Negative regulation of transcription in eukaryotes

Epigenetic Regulation of Transcription in ...

TrpY Regulation of trpB2 Transcription in Methanothermobacter ...

Initiation of transcription of the yeast mitochondrial gene coding for ...