the operon revisited - pdfs.semanticscholar.org · the operon,functional polarized grouping of...

25
Copyright 1972. All rights reserved THE OPERON REVISITED WILLIAM S. REZNIKOFF Department of Biochemistry,Collegeof Agricultural and Life Sciences, University of Wisconsin, Madison, Wisconsin INTRODUCTION Simple observation of living organisms allows one to conclude that the entire genetic potential of any given cell is never expressed at any given moment.In bacteria this phenomenon is termed genetic adaptation and it occurs in two ap- parently contradictory ways. The synthesis of enzymes responsible for a catabolic process generally appears to be turned on (induced) by the substrate (1) while formation of anabolic enzymes appears to be turned off (repressed) by the end product of each given pathway (2). These observations have led to the question-- what mechanisms are responsible for genetic adaptation ? In 1961 Francois Jacob and Jacques Monod addressed themselves to this problem and, in an attempt to describe the mechanisms of genetic adaptation (and hopefully how gene regulation occurs in all organisms), they presented the "operon model" (3, 4). This model in fact described two phenomena: (a) structural genes (those coding for proteins) express themselves, and (b) how expression is regulated. Some parts of the operon model have been so consistent with experimental results that they have become articles of faith for most biologists. These will be described only briefly. Other aspects of the model have either had to be modified with time or have been found to apply to only a few systems. These eases will be discussed in greater detail. This review will be restricted almost entirely to Escherichia coli and Salmonella typhimurium. The reader is strongly advised to search out other recent reviews (5-17). THE EXPRESSION OF STRUCTURAL (~ENES Summary ofmodel.--Structural genes function by serving as templates for the synthesis of messenger RNA (mRNA). It is the mRNA that, in association with the protein synthesizing machinery (the ribosomes, translation factors, amino- acylatcd transfer RNAs, etc.), determines the primary structure of proteins. This mRNA is degraded soon after its synthesis. Structural genes coding for the proteins of a particular biochemical pathway are often clustered on the chro- mosome (e.g., the genes controlling lactose (lac) utilization shown in Fig. 1). Such a group of genes is expressed as a unit in a sequential polarized fashion. The synthesis of mRNA initiates at one end of the operon and progresses through the structural genes (lac z, lac y, and laca in the case of lac) to the end of the 133 3038 www.annualreviews.org/aronline Annual Reviews Annu. Rev. Genet. 1972.6:133-156. Downloaded from arjournals.annualreviews.org by University of Wisconsin - Madison on 03/29/07. For personal use only.

Upload: others

Post on 25-Sep-2019

12 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

Copyright 1972. All rights reserved

THE OPERON REVISITEDWILLIAM S. REZNIKOFF

Department of Biochemistry, College of Agricultural and Life Sciences,University of Wisconsin, Madison, Wisconsin

INTRODUCTION

Simple observation of living organisms allows one to conclude that the entiregenetic potential of any given cell is never expressed at any given moment. Inbacteria this phenomenon is termed genetic adaptation and it occurs in two ap-parently contradictory ways. The synthesis of enzymes responsible for a catabolicprocess generally appears to be turned on (induced) by the substrate (1) while formation of anabolic enzymes appears to be turned off (repressed) by the endproduct of each given pathway (2). These observations have led to the question--what mechanisms are responsible for genetic adaptation ?

In 1961 Francois Jacob and Jacques Monod addressed themselves to thisproblem and, in an attempt to describe the mechanisms of genetic adaptation(and hopefully how gene regulation occurs in all organisms), they presented the"operon model" (3, 4). This model in fact described two phenomena: (a) structural genes (those coding for proteins) express themselves, and (b) how expression is regulated.

Some parts of the operon model have been so consistent with experimentalresults that they have become articles of faith for most biologists. These will bedescribed only briefly. Other aspects of the model have either had to be modifiedwith time or have been found to apply to only a few systems. These eases will bediscussed in greater detail. This review will be restricted almost entirely toEscherichia coli and Salmonella typhimurium. The reader is strongly advised tosearch out other recent reviews (5-17).

THE EXPRESSION OF STRUCTURAL (~ENES

Summary ofmodel.--Structural genes function by serving as templates for thesynthesis of messenger RNA (mRNA). It is the mRNA that, in association withthe protein synthesizing machinery (the ribosomes, translation factors, amino-acylatcd transfer RNAs, etc.), determines the primary structure of proteins. ThismRNA is degraded soon after its synthesis. Structural genes coding for theproteins of a particular biochemical pathway are often clustered on the chro-mosome (e.g., the genes controlling lactose (lac) utilization shown in Fig. 1).Such a group of genes is expressed as a unit in a sequential polarized fashion.The synthesis of mRNA initiates at one end of the operon and progresses throughthe structural genes (lac z, lac y, and laca in the case of lac) to the end of the

133

3038

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 2: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

1.34

p i t:poI~I I~_| .I I

REZNIKOFF

y a

FIG. 1. The lac region of the E. coli chromosome. This is composed of two units oftranscription, the lae i gene, which codes for the lae repressor (REP), and the lac operonwhich includes three genes, lae z, lac y, and lae a, coding for/%galactosidase (B-Gz), galactoside permease (B-P) and thioglactoside transacetylase (TA) respectively. polymerase (RNP) binds to and initiates mRNA synthesis at the promoter, p. TheRNP-p interaction for the lac operon is dependent on the presence of the catabolitegene activator protein (CAP) and cyclic AMP (A). Transcription continues until RNP reaches the terminator (0. The REP prevents transcription of the lac operon bybinding to the operator, o. The REP-o interaction is prevented by the inducer (D.

operon. Such a unit of transcription was defined as an "operon" by Jacob &Monod (3, 4).

The messenger, unstable RNA.--By 1961 it was believed (a) that proteinswere not produced directly on the chromosome but rather on the ribosomes, (b)that RNA was an intermediate "messenger" template in protein synthesis, (c)that ribosomes did not provide the necessary specific messenger information,and (d) that there existed a metabolically unstable species of RNA, which couldbe found associated with the ribosomes, whose properties more accurately fittedthe expected properties of "messenger-RNA" (3, 4, 18-20). Subsequently it wasdiscovered that specific inhibitors of RNA synthesis would promptly block pro-tein synthesis (21-23) and that the unstable RNA contained species that couldform specific molecular hybrids with DNA known to be active in the ceils fromwhich the RNA was isolated (24, 25).

An important question is whether all mRNA species are equally unstableand, ff not, what factor(s) determine the degree of mRNA instability. SomemRNA species in E. coli appear to be quite stable. An almost trivial example isthe RNA molecule injected into E. coli by f2 and other RNA phages. Thismolecule is both the chromosome and the messenger of the phage (26). Anotherexample is the RNA coded for by the coliphage T7 (27). The relative stability T’/mRNA is, at least in part, an inherent property of the mRNA itself. Measure-ments of the stability of tryptophan (trp) mRNA and T7 mRNA in T7 infectedcells indicate that the two species of mRNA maintain their distinctly differenthalf lives (98). Whether T7 in addition produces a regulatory agent which"stabilizes" certain mRNA species (among which trp mRblA is not represented)is still an open question. The original studies by Summers suggested that bulk

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 3: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 13 5

E. coli mRNA is more stable after T7 infection (27) and, in addition, T4 earlymRNA seems to be more stable in T7 infected cells (R. Haselkorn, personalcommunication).

The operon,functional polarized grouping of structural genes.--Jacob & Monod(3, 4) were struck by an observation made by Demerec (29) that genes coding biochemically related functions are frequently linked together on both the E. coliand S. typhimurium chromosomes. They postulated that this linkage has a func-tional aspect--the linked genes are expressed sequentially as a group via thesynthesis of polygenic mRNA. The experimental evidence that supports thishypothesis includes: (a) The observation that the genes of the histidine (his)operon (and other ope~ons) are expressed coordinately (i.e., the repression/induc-tion ratios of all the enzymes coded for by the operon are identical) (30), (b) existence of polar mutations (mutations in one gene which decrease the expressionof neighboring genes; these effects are always manifested in a unidirectionalmanner, e.g. lac z--~lae a) (31L34), (c) The existence of regulatory mutationslocated at one end of the gene group in the operator region. These mutationsaffect the expression of all the linked genes (3, 4, 35), (d) The isolation of pleio-tropic transcription defective mutations in the promoter region located at thesame end of the operon as the operator (36), (e) The detection of polygenicmRNA corresponding to the lac operon (37, 38), the his operon (39), and thetrp operon (4,0), and (f) The observation that induction and repression quentially affect both the synthesis of the enzymes coded for by an operon (41)and the synthesis of the RNA messenger (40, 42-44).

In some cases the genes corresponding to a given pathway are not linked.Some people feel that this observation conflicts with the operon model. In fact,Jacob & Monod recognized and accounted for these cases by suggesting that insuch a system there may exist multiple operons each composed of one or morestructural genes and that all related operons could be controlled by one regula-tory gene in a similar but not necessarily coordinant fashion. The arginine (arg)system (45-47) was the example used by Jacob & Monod (3, 4). Maas & Clark(48) coined the term "regulon" to describe any system regulated by a singleregulatory gene regardless of how many operons are involved. Two other regulonsthat are composed of more than one operon are the arabinose (ara) system (132)and the trp system (164).

There are special conditions where the synthesis of the enzymes coded for bythe his operon occurs in a simultaneous rather than sequential fashion (49, 50).These conditions are those that produce high intracellular concentrations ofNl°-formyltetrahydrofolate. The mechanistic basis of these observations is un-known but they may represent an exception to the generally held hypothesis ofsequential polarized expression.

The promoter, transcription initiation ~ site.--The expression of an operon viathe polarized, sequential synthesis of polygenic mRNA implies a specific genetic

t The term "transcription initiation" refers to all the steps prerequisite to the forma-tion of the first mRNA phosphodlester bond.

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 4: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

13 6 REZNIKOFF

region associated with each operon where transcription initiation occurs. In theoriginal operon model this region was defined as identical to the repressor bind-ing site--the operator (3, 4). Subsequently, when mutations in the operator regionwere discovered not to have drastic effects on the rate of maximal lac expression,the transcription initiation site was postulated to be a distinct genetic entity--thepromoter (51, 52). Defim’tion of the promoter region awaited the isolation four mutations in this region by Scaife & Beckwith (36). These four promotermutations have these important characteristics: (a) They are pleiotropically lacnegative in a cis-dominant fashion (linked lac z, lacy and lae a expression are allcoordinately depressed) (36). (b) They are not suppressible by any known sense or polarity suppressors (36). (c) They are (at least qualitatively) + (theydo not manifest a defect in repressibility) (36, 53, 54). (d) In at least one case lac i-lac p deletion L1) a promoter mutation decreases in vitro lac transcriptionin a purified RNA polymerase-lac DNA system (55).

Genetic localization of these promoter mutations resulted from the isolationof lac deletions entering into or near the lac promoter region from both the"’lac/-side" and the "lac z-side." The generation of these deletions was madepossible by the isolation of two classes trek80 dlac prophages (56). In both casesthe lac region carried by the prophage is located dose to the tonB locus so thatselection for resistance to bacteriophage T1 can yield lac deletions (57). Most the tonB---lae p deletions were chosen from the total collection of tonB-’-lac-mutations by a very arduous procedure requiring the genetic screening of allpossible candidates (53, 54, 58). Using these mutations, it was possible to locatethe promoter between lac i and lac o as indicated in Figure 1 (53, 54).

Another approach to mapping the promoter, vis-a-vis the operator, is to askwhether it is possible to isolate lac o-lac z deletions that do not drastically altermaximal expression of the/~c operon. Such deletions have been found using atechnique first described by Jacob et al (51, 59). One of these deletions results in substantial inactivation of the operator while only reducing the maximal ex-pression of the lac operon 12%. This result indicates that the genetic materialmissing in the lac o-lac z deletion S145 is not necessary for the initiation of lactranscription. Furthermore, in vitro experiments, to be described later, suggestthat at least one of the steps involved in transcription initiation, the binding ofthe polymerase, can occur in the presence of the lae repressor (60).

A similar delineation of the p and o regions has been found for the trpoperon (61, 62) although in other systems such as his the distinction is not soclear (63). The operator and promoter are in the reverse order vis-a-vis the struc-tural genes in the ara operon (9).

In spite of these straightforward results, one might suspect that this descrip-tion of the promoter is too superficial. Transcription initiation is probably a seriesof sequential reactions (i.e., RNA polymerase binding, RNA polymerase move-ment, mRNA chain initiation) each of which may well require separate DNAsequence signals (64, 65). Furthermore, operator mutations do alter the maximallevel of expression of the lac operon (66) although it is not dear that these are

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 5: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 137

transcriptional as opposed to translational effects. Moreover, the promotermutants are not quantitatively as inducible as the parent wild type strain (54).

A further complication in our understanding of the lac p region arises fromthe fact that this promoter is recognized efficiently by RNA polymerase only inthe presence of the catabolite gene activator protein [CAP; in other articles it istermed CGA protein (67) and CRP (68)], and adenosine 3’-5’ cyclic mono-phosphate (cyclic AMP). The expression of the lac operon, and other catabolicsystems, is depressed when the ceils are grown in the presence of glucose orglucose-6-PO4 as opposed to glycerol. This phenomenon is called cataboliterepression. That the promoter is the target site for eatabolite repression is indi-cated by the work of Silverstone et al (69), Pedman et al (70) and Beckwith et (71). These groups have shown that lac promoter mutations result in a loss of/accatabolite repression sensitivity. Furthermore, high level expression of lac that isinsensitive to catabolite repression can result either from fusing the/ac genes to acatabolite insensitive promoter system (such as trp) (69) or by generating secondsite, closely linked, revertants of promoter point mutations (72). Since the lac Mac p deletion results in total insensitivity to catabolite repression and sincethe catabolite repression insensitive revertants of promoter mutations map out-side of L1 (towards lac z), the region of the promoter involved in catabolite repres-sion is covered by L1 (71).

The identification of cyclic AMP and CAP as the active components incatabolite repression came from studies showing that the intracellular level ofcyclic AMP is lowered when cells are subjected to catabolite repression (73), thatexogenous addition of cyclic AMP reverses catabolite repression (74, 75), andthat mutants unable to express any catabolite sensitive operon are defective eitherin adenyl cyclase or CAP (76-78). The hypothesis that CAP and cyclic AMP arenecessary for the initiation of mRNA synthesis at lacp has been supported by theobservation that CAP and cyclic AMP are absolutely required for correct invitro synthesis of lac mRNA in the purified RNA polymerase, lac DNA system(55, 60, 79). One would suspect that CAP would show a cyclic AMP dependentaffinity for lac p DNA, since it does not show an affinity for RNA polymerase(79) and since a mutation in CAP can phenotypically revert la c promotermutation (Plenge & Beckwith, personal communication). Unfortunately CAPshows a cyclic AMP stimulated affinity for any DNA (67).

Possible promoter mutations decreasing gene activity have been isolated inthe his operon (80), the leucine (leu) operon (81), and the trp operon (61) whilepromoter-like mutations increasing gene activity have been discovered associatedwith the lac i gene (82) and the gene coding for glucose 6-phosphate dehydro-genase (83).

The terminator, mRNA termination site.--While not mentioned in the originaloperon hypothesis (3, 4), the synthesis of specific polygenic operon mRNAclearly implies a specific mRNA termination site at the end of each operon.

The lac i gene has the same orientation as the lac operon (84, 85) (Fig. 1).

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 6: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

13 8 REZNIKOFF

therefore offers an opportunity to ask whether mRNA terminators exist, or, inother words, do RNA polymerase molecules that transcribe lac i stop beforereaching the lac operon ? The answer is YES--even where the transcription of laci has been dramatically elevated. However, transcription of the lac i gene doesread-through into lac when the lac #lac p deletion L1 is introduced, suggestingthat L1 removes the lac i gene terminator (84, 86). Alternatively one could arguethat normal termination of lae i---qae read-through is a secondary (polar) resultof translation termination. Evidence against this possibility comes from theobservation that the polarity suppressor SuA does not effect lac i-)lac read-through (87) and that, in order to fuse lae to another operon (i.e. trp), it seems tobe necessary that the fusion deletion extend beyond the end of lac i (88).

Transcription of the trp operon does not cease with the translation termina-tion signal of the last gene (trpA). Starting with the 480 dlac lysogen mentionedpreviously in which the transposed lac operon is located in the same orientationas the nearby trp operon

lac

(POED CBA ..... ~onB ..... i po~y~),

it has been possible to isolate trp-tonB--lac deletions that fuse the two operons.In two cases, the deletions do not cut into trp,4, indicating that the trp mRNAterminator is located some distance from the end of trp,4 (88). Recently Berger Hardman (personal communication) have obtained results that suggest the sameconclusion. They have isolated the tryptophan synthetase A protein made in astrain carrying a frameshift mutation near the end of the trl~,4 gene. The molecularweight of this frameshift protein is approximately 25% larger than that for normal

. tryptophan synthetase A. The sequence of this added segment is currently beingdetermined. We suspect that the region in which our trp,4 + trp-lac fusion dele-tions end and the region attached to the trpA frameshift protein is not an un-identified gene. Rose & Yanofsky have shown that there is not enough detectablemRNA produced beyond trp,4 to code for another "trp,4-size" protein (89).

In vitro evidence has suggested that a protein termed rho isolable from un-infected E. coli may play a role in the correct tcrmlnation of ~ mRNAs (90). direct in vivo evidence exists to suggest that rho (or a rho-like protein) acts terminate mRNA synthesis at the end of bacterial operons. However, Nissley etal (91) have found that the addition ofrho to an in vitro transcription system using~-gal DNA as a template blocks gal promoter stimulated transcription of k genes.The site of rho action may be the gal terminator or a k terminator closely linkedto the gal substitution.

REGULATION OF (~ENE EXPRESSION

Summary of model.--For each regulated operon there exists a specificmolecule termed the repressor, whose function is to control expression Of theoperon. The gene coding for the repressor is the regulator gene (an example is thelac i gene in Figure 1). The repressor acts by binding to the operator region (o)

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 7: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 139

thereby preventing transcription of the operon specific polygenic mRNA. Foreach repressor-operator system there exists an effector molecule (the inducer orthe corepressor) that interacts with the repressor affecting its affmity for theoperator region. The corepressor (an end product or related compound ofarmbolic systems) increases the repressor’s operator binding activity, while theinducer acts in a reverse manner. This model is pictorially represented for lacin Figure 1. Some systems are not regulated in this fashion.

Regulation occurs at the level of transcription.--The cornerstone of the operontheory is that the decision for or against gene expression is made at the levd oftranscription (3, 4). This hypothesis was primarily an economical one, sinceregulation at the translation level would result in valuable energy being squan-dered in making unstable, unused mRNA.

A result that seems to support this hypothesis is the observation that lacspecific mRNA (pulse labeled RNA hybrizable to lae DNA) is found only ininduced ceils (24, 25). However, since lac mRNA is also missing from induced cellsthat carry a translation defect in the lac z gene (a nonsense mutation) (25) we not be sure that the absence of messenger under repressed conditions is not anindirect result of inhibited translation.

Another way to discover the level at which regulation occurs is to determinethe target site for the repressor. Genetic studies have failed to distinguish whetherthe operator is on the DNA or on the mRNA. Gilbert & Miiller-Hill (92) andRiggs et al (93) found that the lac repressor binds specifically and tightly to lacoperator double-stranded DNA. Other repressors (gal, A and 434) also bind toDNAs carrying their specific operators (94-96). These data strongly support thenotion that operator DNA is the in vivo repressor target site, although a DNA-rnRNA hybrid or a hairpin mRNA double-stranded region hasn’t been ruled out.

Finally, recent in vitro experiments show that purified repressor can partiallyblock lac mRNA synthesis in a purified DNA-RNA polymerase system (55, 60).This suggests strongly that lac regulation occurs at the level of transcription.

Although we know that transcriptional control exists, several physiologicalexperiments suggest translational control on top of transcriptional control (97).One probable mechanism of translational control has been elucidated. T4 infec-tion inhibits the translation of RNA phage genomes (98). In vitro protein syn-thesis studies have shown that this defect in R17 or MS2 translation can belocalized in the initiation factors (99) and more specifically in the initiation factorF3 (100). While the experiments do not define how T4 infection alters F3 spec-ificity for mRNA ribosome binding sites, they identify a potential means ofreversible regulation of translation. One merely has to postulate that F3-mRNArecognition can be influenced by various effectors. It would be an even moreattractive hypothesis if the cell had mutiple species of F3 protein each with apreferred specificity for different mRNAs. Multiple species of F3 have been found(101-105).

The repressor, negatiDe controller of operon functioning.--The operon modelpredicted the existence of two genetic regulator~ elements. One element, the

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 8: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

140 REZNIKOFF

regulator gene, which codes for the repressor, is independent from the operonbeing regulated.

The important experimental characteristics of the lac repressor and its geneknown in 1961 include the following: (a) The repressor is distinct from the en-zymes whose synthesis it regulates. It has a distinctly different "substrate"specificity than do these enzymes (106), and mutations that alter its functioning(those resulting in unregulated high level expression of lac) reside at a site closeto but different from the structural genes of the lac operon (3,107, 108). (b) Theseso-called lac i- constitutive mutations result from a loss of function since theyare recessive and since some lac i- mutations are deletions (3, 109). (c) lac igene codes for a diffusible product since lac i + is dominant to (most) lac i-mutations even when the only functional lac structural genes are linked to thelac i- allele (3, 109). This information also confirms the conclusion drawn froma and b above. (d) The lac i gene does not code for an essential function sincestrains carrying lac i- deletions are viable (3). (e) The repressor is the only regula-tor (negative or positive) specific for lac since all known lac constitutive muta-tions reside either in lac i or in the repressor target site (o) (3) and all knownlac- mutations are located either in lac z or lacy or in lac i (so-called lac i"mutations that generate a repressor insensitive to the inducer) (3, 4, 110).

The initial belief that the lac repressor was RNA (3, 4) proved to be incorrect.The subsequent isolation of temperature sensitive mutations and suppressiblenonsense mutations in lac i suggested that the repressor was a protein (111,112)as did studies with protein synthesis inhibitors (113, 114). The protein nature the repressor was confirmed when it was isolated and purified (93, 115, 116).

The current biochemical knowledge about the lac repressor indicates that itundergoes three different recognition reactions, all of which are essential to itsfunctioning. Individual lac repressor protein molecules can recognize each otherand aggregate into tetramers (117, 118). Both the monomeric and the tetramericforms of the repessor bind isopropyl-fl-thiogalactoside (IPTG), an inducer of thelac operon (115, 117). The repressor tetramer binds specifically and tightly to thelac operator region (92, 93).

The ability of the repressor to bind to the operator (and thereby inhibit lactranscription) is determined by the other two reactions described above. Onlythe tetrameric form of the repressor can bind to the operator and there existsonly one binding site/tetramer (119). This suggests that repressors that fail aggregate will not repress. Presumably some i- mutations might lead to defectiveaggregation properties. Furthermore, mixed tetramers of operator bindingactive and inactive monomers should be inactive in operator binding. This isthought to be the explanation for the rare transdominant lac i-(t "-~) mutations(120, 121). IPTG causes a drastic reduction in the repressor’s affinity for theoperator (92, 93) presumably by altering the repressor’s conformation. Muta-tions that decrease the repressor’s affinity for the inducer shoiald have an un-inducible phenotype and, in fact, the lac i ~ mutations that manifest this phenotypedo have a lower affinity for IPTG (119).

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 9: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 141

Recently experiments by Eron & Block (55) and deCrombrugghe et al (79)have shown that the repressor can partially block in vitro synthesis of lac mRNAin a purified RNA polymerase-/ac DNA system and that this activity is reversedby IPTG.

To a first approximation, these results seem to apply to other systems. Inseveral cases apparent regulator gene mutants have been isolated. These usuallycarry recessive mutations that result in constitutive expression of the system.These regulator genes also supply a nonessential function to the cell. In manycases one suspects that the active repressor is a protein since suppressible non-sense mutations have been isolated and in at least four cases the repressor isthought to be an oligomer of identical subunits, since either intracistronic com-plementation or negative complementation has been detected [lon (122), metJ(123), the R-factor repressor (124), trpR(125)]. However, in on ly a very fewcases has the repressor been isolated and shown to act in vitro as predicted [thephage k repressor (95), the phage 434 repessor (96), and gal repressor (94)

In some other systems, the evidence is not yet consistent with the classicregulator gene concept. For instance, many man years of work have been in-vested in studying the regulation of the his operon and yet the regulator gene isstill to be identified.

The first striking fact about the his system is that there is a multiplicity of lociwhere mutations can produce constitutive his phenotypes. These include fivelocations (hisS, hisR, hisT, hisU, and hisW) in addition to the his operator region(126, 127). Mutations in three of these loci are known to be recessive and in allcases but one (hisR) the products of these genes are thought to be diffusibleproteins (11, 17). These data would suggest an embarrassment of riches (too manyrepressors) but in fact, closer examination suggests that all the loci can berelegated to roles in his regulation other than the production of the his repressorprotein. The real physiological effector (the corepressor in this case) is often "processed product" of the "effector" added to the cell suspension and in thiscase His-tRNA is thought to be the effector (11, 17, 128). All five of these hisregulatory loci appear to play roles in the production of or modification of His-tRNA; hiss codes for His-tRNA synthetase (129), hisR may code for tRNAais

itself (130) or a positive regulator of its production, hisT codes for a tRNAmodifying function that introduces two pseudouracils into the anticodon loopof tRNAms (131), and hisU and hisF¢ appear to play a role in the maturation oftRNArri~ and other tRNAs (11, 17, 128).

One can therefore conclude that the mere presence of constitutive mutationsin a particular locus, even recessive, nonsense, constitutive mutations that do notimpair cell growth, is insufficient evidence to identify the locus as the regulatorgene. (The reader should be warned that I shall make this mistake subsequently.)The product of the regulator gene must be demonstrated to have an affinity forthe real physiological effector and an affinity for the operator. Furthermore, itmust be shown to "do its thing" in vitro.

Before presenting the current favored hypotheses about the identity of the his

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 10: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

142 REZNIKOFF

regulator gene, it might be wise to define the concept of the regulator gene so thatthe reader can decide why the regulator gene may not have been found in this andcertain other cases.

(a) The regulator gene exists. A priori, one would predict that the cell mustprovide an operon-specific macromolecule (the regulator) through which an effec-tor can exert a regulatory influence on the mRNA or protein synthesizing ma-chinery. To require operon specific effectors to interact directly with the RNApolymerase complex (or protein synthesizing complex) thereby inttuencing itsrecognition of a series of operons, is to demand incredible flexibility and specific-ity of a generalized synthetic system. The regulator gene codes for this regulator.

(b) The regulator is a repressor--i.e., it acts in a negative fashion. Work the ara system (9, 132), the maltose (real) system (133), and the rhamnose (rha)system (134) has shown that this generalization is not valid. These systems are atleast partially controlled by specific regulatory proteins (so-called activators)which act in a positive fashion. Constitutive mutations in an activator shouldbe rare.

(c) The regulator acts through the cytoplasm. A cis-active protein is appar-ently coded for by the phage P2 A gene. This protein is necessary for P2 gene ex-pression or DNA synthesis, or both (135). The presence of a cis-active regulatoryprotein could be detected by the presence of cis-domiaant nonsense mutationseffecting derepression of the operon (or, for positive control, pleiotropic negativeshutdown of the operon).

(d) The regulator has no other function. In certain cases the regulator mayhave some other function necessary for cell survival. This would include the pos-sibility of the regulator being a necessary enzyme or being a regulator of anotheressential system.

(e) The regulator gene is an independent unit of transcription. The gene cod-ing for the regulator might be an operator proximal gene in an operon whose dis-tal genes perform necessary functions. Any polar mutation would prevent ex-pression of these necessary genes. One example of an apparent regulator genebeing part of a larger operon is found in the histidine utilization (hut) system(136).

(f) There is one unique regulator gene for each system. There may be morethan one regulator gene for a given system. Only very rare double mutationswould give rise to a constitutive phenotype. Recent experiments have indicatedthat the ~,-N gene operator, as defined by the ~2 mutation, is the target site forboth the cI and tofgene products (137).

To return to his, current hypotheses suggest that either the hiss gene (whichcodes for the His-tRNA synthetase) or hisG (which codes for the first enzyme inthe his pathway, phosphoribosyl-ATP synthetase), or both is the his regulatorgene (11, 17). The evidence for the role of hisS comes not from the constitutivehiss mutations (these can all be explained by their effect on the His-tRNA poolsize) but on some observations by Wyche, who isolated hiss mutations that pre-v~n~; full .derepression of the ttis operon, H.e ~tl~o fo~n~d th.at in h~terologou~

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 11: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 143

S. abony/S, typhimurium hisS+ diploids the basal level of his expression is raised(138). These results are compatible with the hisS protein having a direct positiverole in his regulation or might be explained by altered affinities of the His-tRNAsynthetases for tRNAm~. This latter possibility is being examined (B. N. Ames,personal communication). His-tRNA synthetase binds to DNA. Whether it isspecific for his operator DNA is also being studied (Ames).

The possibility of a regulatory role for the hisG enzyme comes from the fol-lowing observations. The enzyme’s feedback resistant property affects the regula-tion of his operon expression (139, 140), the enzyme displays an affinity for His-tRNA (141), and the enzyme binds to DNA in general and his-DNA preferentially(R. Goldberger, personal communication). Evidence that suggests that hisG isnot a regulatory gene (or at least not the unique his regulator gene) includes thefacts that no constitutive mutations have been isolated in hisG, and in the casestested so far, hisG nonsense and frameshift mutations have no effect on hisregulation (17).

The operator, the binding site for the repressor.--Jacob, Monod and coworkerspostulated that there must be a target site for the repressor. They termed thistarget site the operator. In 1960 mutations were isolated whose phenotype andlocation fit the predicted properties of operator mutations (35). These mutationshad a partially constitutive phenotype, they mapped between the promoter andthe start of the lac z gene, and their constitutive phenotype was cis-dominant (3,4, 35, 54). Subsequent in vitro experiments demonstrated that DNA containinglac o~ mutations has a lower affinity than wild type DNA for the lac repressorprotein (92, 93) and, therefore, that the operator region, as defined by these muta-tions, is the binding site for the repressor. Similar cis-dominant operator constitu-tive mutations have been isolated for many other systems.

The most detailed genetic study of an operator region was accomplished bySmith & Sadler, who isolated and characterized hundreds of independent pointmutations in the lac operator region (66, 142). Their findings indicate at least and probably 16 sites in the operator where a transition or a transversion mutationcan generate a detectable oc phentoype. Operator mutations found at a given siteall seem to generate the same phenotype. Although the total phenotype (i.e. themaximal level of expression and the induction ratio) is unique to the c mutationsof a given site, there appear to be pairs of sites in which the o° mutations haveidentical induction ratios. Moreover the tentative map order suggests that thesepaired sites are arranged in a bilaterally symmetrical fashion. It is surprising thatthe operator’s base sequence, as deduced from mutagenic data, does not reflectthis phenotypic symmetry (some matched sites do not have the same nucleotidepairs).

Another important genetic system for studying the lac o region results fromthe isolation of tonB- deletions ending in or near lac o. We have developed tech-niques that allow us specifically to isolate deletions ending in this region (86, 88).The parental strain contains: a ~b80 dlac prophage such that the trp and lac ope-

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 12: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

144 REZNIKOFF

rons are in the same orientation, a trpR- mutation resulting in constitutive exrpression of the trp operon, and a lac promoter point mutation reducing lac ex-pression. From this strain we isolated tonB- mutations that simultaneouslybecame lac+. All the deletions analyzed to date end between lac i and lac z. Wehope that these deletions will facilitate fine structure deletion mapping of lac o.Since some of these ddetions remove all or most of lac o but still recombine withall lac z mutations and have retained the ability to synthesize wild type/~-galacto-sidase, we believe that lac o and lac z are distinct genetic entities (58, 143). Thisresult confirms the previous finding of Steers et al that a lac oc mutation does notaffect the properties of/~-galactosidase (144).

In a positively controlled system there should be an operon linked target sitefor the activator protein. Starting with a strain carrying a deletion of the arapositive regulatory element (the araC gene), ethyl methane sulfonate and 2-amino-purine induced ara+ mutations in the putative target site for the araC activatorhave been isolated (145, 146). These mutations are called ~ (initiator c onstitu-tive) mutations. They are located near the arab gene, at one end of the operon,and result in expression of the ara operon in the absence of the activator. Similarmutations have been isolated by Hofnung & Schwartz in the real system (147).

However, the araIc mutations do not by themselves generate ara transcriptionwhen in the presence of an araC+ allele (145, 146, 148). This is thought to be be-cause the araC protein has both a repressor function (in the absence of the in-ducer, arabinose) and an activator function (in the presence of the inducer). araIo mutations obviate the necessity for the second function but still leave theoperon sensitive to the first, suggesting that the C-repressor has an additionaltarget site, the ara operator. Deletion analysis indicates that the ara operatorresides between aral and araC (araO-araC deletions make the aral~ stimulatedtranscription of ara insensitive to the presence of an araC+ allele). Such a doublecontrol system seems quite complicated. It required the evolution of two geneticsignals (araI and araO) which are recognized by the same protein in two ap-parently different ways. The ara system is currently being analyzed in vitro (149-151) to test this repressor-activator model.

The effector. A compound which controls the affinity of the regulator protein forthe operator (or initiator).--Essential to a description of gene regulation in bac-teria is the ducidation of how the metabolites that either induce or repress enzymesynthesis can effect their roles. The Jacob-Monod operon model predicts thatthese metabolites influence gene expression by interacting with the particularregulatory protein and altering its affinity for the particular genetic signal associ-ated with the operon. In general the in vitro work with the lac repressor has sub-stantiated this model.

However, the effector is frequently not identical to the metabolite added tothe cell suspension. For instance, in the case of the lac operon, lactose itself is notthe inducer (152). Rather the physiological inducer is allolactose, an isomer lactose generated from lactose by/~-galactosidase (153). Traditionally, people

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 13: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 145

studying the regulation of the lac operon have sidestepped this problem by usinggratuitous inducers, which are chemically synthesized, nonmetabolizable com-pounds such as isopropyl-~-D-thiogalactoside (IPTG).

Groups studying the regulation of biosynthetic operons have been very inter-ested in tracking down the true physiological effectors, Eidlic & Neidhardt (154)isolated a temperature sensitive mutant in E. coli that helped define one sucheffector. This mutant carried a defect in the Val-tRNA synthetase which, at non-permissive temperatures, not only stopped cell growth but also derepressed thesynthesis of the isoleucine-valine (ilv) regulon enzymes. Subsequent studies, withvaline analogs and with measurements of the level of charged tRNAva~ undervarious conditions, indicated that a likely candidate for an ilv regulon effector isVal-tRNA (155, 156). As indicated in my discussion of the his system, other (butnot all) amino acid system effectors are also charged tRNAs.

A particularly complex situation seems to exist in biosynthetic systemswhichlead to the synthesis of more than one end product. In several cases repressionappears to result from the presence of more than one effector. Experimentallytwo different kinds of multi-effector systems have been discovered: cumulativerepression, in which total repression seems to result from an additive repressionof each of the individual effectors [as in the regulation of carbamyl phosphatesynthetase synthesis by arglnine and a pyrimidine (157)], and multivalent repres-sion, in which all the effectors must be present for repression to occur (158).Despite the apparent differences between these two systems, we don’t knowwhether they are mechanistically different (H. E. Umbarger, personal communica-tion) so we will tentatively consider them as a group using the more commonterm multivalent repression.

Perhaps the best analyzed systems involving multivalent repression are thethreonine, isoleucine, valine, .and leucine metabolic pathways. Threonine is aprecursor for isoleucine biosynthesis; therefore, it is not surprising that the syn-thesis of the threonine biosynthetic enzymes is repressed by both isoleucine andthreonine (159). Most of the enzymes used in the biosynthesis of isoleucine fromthreonine are also used for valine biosynthesis and for the synthesis of a-ketoiso-valerate a precursor for leucine biosynthesis. Synthesis of these enzymes (or atleast those coded for by ilvA, ilvD, and ilvE) is subject to multivalent repressionby isoleucine, leucine, and valine. The steps unique to leucine biosynthesis aresubject only to leucine repression (158). In all these cases charged transfer RNAsmay be the actual effectors (155, 160-162).

Several questions emerge from these studies. Does multivalent repressionresult from all the effectors interacting with one repressor protein or are thereseveral different repressor proteins, each specific for a given effector ? In the lattercase all the repressors would have to act in concert in order to generate completerepression. If there is a single repressor for, say, ilvA, ilvD, and ilvEwith affinitiesfor multiple effectors, does the same repressor protein also regulate otheroperons that are sensitive to only one or two of the effectors ?

JEvid,enc¢ from another mu!tivalent repression system suggests that one r~pres-

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 14: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

146 REZNIKOFF

sor can be influenced by multiple effectors. The first step of the common aromaticamino acid biosynthetic pathway (PEPq-E4P---~DAHP) is catalyzed by three iso-enzymes. The synthesis of one of these isoenzymes (aroG-DAHP synthetase) isrepressed by phenylalanine and tryptophan (163, 164). The synthesis of a secondDAHP synthetase (that coded for by aroF) is repressed by tyrosine but high con-centrations of phenylalanine (and possibly tryptophan) can substitute for tyrosine(163, 165). This rather complex picture has been clarified considerably by the factthat mutations in one locus, tyrR, can result in derepression of aroG-DAHPsynthetase, and aroF-DAHP synthetase (165, 166). Other mutations in tyrR de-repress aroF-DAHP synthetase but not aroG-DAHP synthetase. If tyrR codes fora repressor protein, these results suggest that this repressor has affinities forphenylalanine, tyrosine, and tryptophan (or their derivatives), that these effeCtorsdifferentially affect the repressor’s affinity for different operator regions, and thatit is possible to alter by mutation one effector-repressor interaction withoutdrastically affecting the others.

There is one system thought to be controlled by more than one regulatoryprotein although it is not clear whether multiple effectors are involved. The galoperon is regulated by the gal repressor gene galR (167) and by Ion, the goneregulating capsular polysaccharide synthesis (122, 168). This regulation seems be independent and additive (168). It is similar to that seen for lac in trp-lae fusionstrains in which both trp and lac operators are present and where both repressorscan control lac expression in an additive fashion (86).

A general assumption in the study of E. coil gene regulation is that if an effec-tor (or effectors) influences the production of enzymes coded by genes in differentoperons, then one repressor must be responsible for directly influencing the pro-duction of all of them. This ignores the possible existence of "cascade regulation"where one of the operons codes for the production of an internal inducer (and anactivator ?) which turns on one or more of the other operons. It turns out that theilvC gene is probably controlled by an induction mechanism in whichthe productof the ilvB enzyme’s activity is the inducer (169).

Trm MOLECtrLAIi BASIS OF RECULATION

lntroduction.--The Jacob-Monod operon model is a broad conceptual frame-work containing few molecular details. Now that we can define the frameworkwith some accuracy (at least in the case of lac) it is probably worth consideringthe molecular basis for that framework. For instance: How does transcriptioninitiation occur ? How does the repressor act to block transcription ? What is thestructure (size and sequence) of the operator ? How do proteins such as the repres-sor and RNA polymerase recognize specific DNA sequences ? These questionsmerely scratch the molecular surface of the problem.

How does transcription initiation occur ?--Recent in vitro work, such as thatby Zillig et al (170) has suggested that transcription initiation is composed of series of reactions. The first of these is the binding of RNA polymerase to DNA.

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 15: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 147

One can ask whether RNA polymerase binding is the only sequence specific step(the other steps following automatically) or whether each step in transcriptioninitiation requires a specific DNA signal. In other words does the promoter con-tain one or many signals ?

We now have evidence that the X-N gene promoter is probably composed ofat least two spatially separate signals--a signal defined by the N promoter muta-tions t27 and sex (possibly the RNA polymerase binding site) and the "start"site where N-mRNA synthesis actually originates. Blattner et al (64, 65, Dahlberg& Blattner, personal communication) have accomplished a hybridization andsequence analysis of the first 160 nucleotides of in vitro synthesized N mRNA.This N mRNA sequence, which is identical to the in vivo sequence (171), startsoutside of the #nm434 region found in X-imm434 hybrids. From genetic studiesand electron micrographic heteroduplex mapping data we know that the t27 andsex mutations are probably located 200 or more nucleotides within the imm434region. This information indicates that for N-gene transcription to occur, RNApolymerase binds at one site, and then migrates a distance greater than its owndiameter to a second site where N-mRNA synthesis starts.

Indirect evidence suggests that the lac promoter mutations also alter a geneticsignal necessary for an early step in transcription initiation. Perhaps the signalaltered is the RNA polymerase binding site. In vitro studies by de Crombruggheet al (79) have suggested that cyclic AMP is required for the formation of a lacspecific, rifampicin resistant, RNA polymerase-DNA complex. The lac promotermutations reduce the cyclic AMP requirement for lac operon expression (69-71).Thus these mutations probably alter the signal necessary for the formation of theRNA polymerase-DNA complex.

How does the repressor block transcription ?--Does the repressor block bind-ing of RNA polymerase, initiation of RNA polymerase activity, or elongation ofthe messenger ? We (86) tried to generate an in vivo system for analyzing the modeof lac repressor action. Trp-lacfusion deletions wereisolated in which lac expressionwas the result of RNA transcription initiating at the trp promoter one or moregenes removed from the lac operator region. We then asked whether lac repressorcould block read-through transcription. If it did, one could conclude that therepressor was capable of(and by analogy, acted by) blocking progress of preboundRNA polymerase. Unfortunately the answer was ambiguous (partial repressionoccurred).

Chert et al (60) examined repressor action in vitro by adding lac repressoreither before or after adding RNA polymerase. Repressor added prior to RNApolymerase addition blocks lac mRNA synthesis 2- 4-fold. Addition of IPTG tothis mixture in the presence of rifampicin results in induction of lac mRNA syn-thesis suggesting that RNA polymerase binding is not prevented by the repres-sor. The opposite order of addition was also tried. Repressor can block mRNAsynthesis by prebound RNA polymerase and this repression is reversible by theaddition of IPTG.

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 16: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

14 8 REZNIKOFF

These results (and others relying on the in vitro lac system) are promising butthey should only be taken as tentative because of the nature of the in vitro system.The ontogeny of the in vitro lac system reflects a long search for conditions thataccurately reflect in vivo phenomena (i.e., sensitivity to repressor-inducer control,to CAP-cyclic AMP control, to promoter mutations). Though one suspects thatthe initial experimental difficulties were due to real problems, such as phagespecific promoter activities, it would have been more pleasing if the system hadworked correctly in a straightforward manner. Furthermore, the response of thein vitro system is only qualitatively equivalent to the in vivo results. For instancethe real repression ratio is 2,000 fold, not 2- 4-fold. This discrepancy is probablydue to real experimental difficulties but nonetheless should make us realize thatthe results might not reflect the true in vivo events.

Similar in vitro experiments have been done using a ~-DNA system (171a).These experiments show that N mRNA synthesis by prebound RNA polymerasecan be blocked by the ~,cI repressor. However, the reverse reaction sequence in-dicated that ~,cI repressor not only blocks mRNA synthesis, but also blocks RNApolymerase binding.

The previously mentioned studies on X-N transcription (64, 65, 171) also tellus something about the mode of action of the kcI and tof repressors. The Noperator, as defined by the o2 mutation, resides inside the imm434 substitutionclose to the sites occupied by the sex and t27 mutations. Since this is a consider-able distance from where the N mRiNA starts, the cI and tofgene products mustblock an early step in transcription initiation. In the lac and trp systems, however,genetic evidence suggests that the operators are transcribed [lac o-lac z and trpO-trp A deletions do not impair operon expression (59, 62).] Thus the lac andtrp repressors may prevent a late step in transcription initiation.

Structure of the lac operator.--One can argue a priori that the lac operatorregion must be equal to or larger than the minimum nonrepeating sequence lengthfor the E. eoli chromosome (or >12 nucleotide pairs). Qualifying statementswould include the fact that single nucleotide pair modifications of the operatormust also not be found elsewhere (142) (the necessary length increases) and the existence of operator-like repeats would not be just the result of random prob-ability but would be subject to negative selective pressure (the necessary lengthshrinks). The minimum operator length has been determined experimentally bySmith & Sadler, who discovered that there are at least 12 sites capable of generat-ing o° mutations (66, 142). Recently Sadler and coworkers (personal communica-tion) and Gilbert and coworkers (personal communication) have isolated opera-tor DNA using the fact that repressor bound to the operator will protect it fromdigestion by DNase. Sadler’s group estimates, from the fraction of input q~80dlac DNA that is protected by repressor from DNase attack, that the operatoris < 75 pairs long.

Mutagenic data indicate that both A-T and G-C pairs exist in the operatorwith A-T pairs predominating (66). This fits with some experiments by Riggs

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 17: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 149

et al (172) which show that the repressor interacts best with A-T rich DNA poly-mers, although this interaction is still several orders of magnitude less than thatmanifested by the repressor towards the real operator.

Structure and function of the repressor.--A strong effort is being made toelucidate the structure of the lac repressor using both genetic and chemical tech-niques. Thanks to a large collection of tonB--lac i- deletions isolated by Miller,it has been possible to map many lac i mutations (16, 173,174, 174a).

The lac i -d mutations are clustered close to the amino terminus of the lac igene. These mutations are known to result in the formation of repressor mole culesthat cannot interact with the operator (173, 174, Jackson, personal communica-tion). The effect of these lac i-d mutations may be due to a direct change in theoperator binding site, or they may alter the repressor’s conformational propertiesso that operator binding cannot occur, or they may destroy the operator bindingcapability by changing the repressor’s aggregation properties. The observationthat these mutations are transdominant suggests that their subunits may aggre-gate more slowly than normal repressor monomers (16). Sadler (personal com-munication) has recently isolated pseudorevertants of lac o~ mutations that mapin lac i. These mutations may help to define the operator binding site and it wouldbe interesting to see if they were located close to the lac i TM mutations.

Mutations known to decrease the affinity of the repressor for IPTG (the lac is

mutations) (119) map in five different locations (16, 173). Again we do not whether the effect is a direct or indirect result of the mutational change. In onecase (that located closest to the lac i-a region) the lac i s mutation also alters theaffinity for the operator (16, 119, 173, 174). The lac i t mutation [which generates arepressor showing a higher affinity for the inducer (115)] may be located in theIPTG binding site.

To achieve the long range goal of determining the structures of wild type andmutant forms of the lac repressor and perhaps of various repressor-inducer andrepressor-operator complexes, sequence, crystallographic, and other physio-chemical studies are being initiated. A sequence analysis of the first 50 aminoacids of the wild type repressor has been accomplished and sequences of repres-sors carrying various i -d mutations are being examined (174, 174a, Fanning,personal communication). Many of these mutations should alter this sequence(174). Crystals of the repressor have been obtained and are currently beinganalyzed (Steitz, personal communication).

Mechanism of protein-DNA recognition reactions.--See the article by Peteryon Hippel for a concise analysis of this topic (175). Many of the ideas presentedbelow are similar to his.

If one assumes that DNA is uniformly in the B structure, then one can im-mediately conclude that there exists no sequence specificity in the phosphategroups or the deoxyribose groups and that the normal hydrogen bonding sitesof the purine-pyrimidine pairs would also be unavailable to an outside reactant.

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 18: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

150 REZNIKOFF

BROAD GROOVE

.....I c--c c--c

\_, \ /

...,/...,ill..."

NARROW GROOVE

Flo. 2. Possible reactive groups in the broad and narrow grooves of DNA. Groupsabsolutely unique to a single base are circled. Those that perhaps are only unique to thepurines or pyrimidines are in squares.

Is there anything else ? Observation of a space-filled model of DNA suggests awealth of specific information available for protein-nucleic acid interactions inthe broad and narrow grooves. Most of these are possible hydrogen bondingsites but some hydrophobic groups are present too. Many of these sites have beendescribed by Wilkins (176). In Figure 2 1 have tried to indicate some potentiallyimportant groups. Those circled may be absolutely unique centers (defining single base) while those enclosed in squares may either provide absolute unique-ness or twofold uniqueness (distinguishing purines from pyrimidines). Kelly Smith found evidence for both types of specificity in the Haemophilis influenzarestriction enzyme recognition site (177). The data for lac o give evidence only forcompletely unique base pair recognitions (66, 142). Recently, Adler et al (174)

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 19: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 151

have used the sequence of the first 50 amino acids in the lac repressor to constructa molecular model of the repressor-operator interaction. In this model, a portionof the repressor’s N-terminus lies in the broad groove where it specificallyrecognizes some of the unique centers pictured in Figure 2.

Our assumption that DNA resides in a uniform B structure is probably notvalid. DNA seems to "breathe," that is, it undergoes localized denaturations(178), suggesting that the normal internal hydrogen bonding sites could be openedup for interactions. Since hydrogen bonds would be lost in the process, thisconfiguration probably would not be favored unless stabilized by a specific inter-action with a protein. Such an event might be considered a protein induced con-formational change in the DNA.

Furthermore, as postulated by Gierer (179), some stretches of DNA may composed of two identical antiparallel sequences capable of forming hairpinloops similar to those found in tRNA. Such loops might give uniqueness to thephosphate and deoxyribose groups. The possible twofold symmetry of theoperator (142) suggests such a structure. Since hairpin loops would be less stablethan the normal conformation because some hydrogen bonding sites would belost, they probably would have to be induced by protein too.

Finally, the B structure of DNA is probably an average structure. A shortstretch of DNA may have a slightly different conformation generated by thebase composition and sequence of the particular region. This possibility has beendeduced largely from physicochemical studies of synthetic DNA polymers com-posed of defined repeating sequences [e.g., see Wells et al (180)]. Sequencespecified changes in the helical structure could generate structure specific (i.e.,sequence specific) deoxyribose and PO~ groups for recognition purposes.

It is also possible that aromatic amino acid residues of a protein might inter-calate into a DNA molecule (181). A possible model for such an intercalationcomes from the crystallographic analysis of puromycin, which shows an inter-molecular stacking of adenine bases and the tyrosine aromatic rings (182). How-ever, it is not clear where the specificity for such a reaction would come from.

These are rather simple ideas about a very complex subject. Do we have anyevidence for any of these recognition mechanisms ? Right now we are at theintuitive guess stage. I personally believe that several mechanisms are active.For instance, at some specific stage in its interaction with DNA, RNA poly-merase must induce (or stabilize) localized denaturation. The repressor probablydoesn’t denature DNA (the protected piece is double stranded) (Sadler, personalcommunication). Furthermore, more than one type of reaction may be involvedin a particular instance. For instance, the lac repressor interacts withpoly(dA--dT), poly(dA-dT) equally well in the presence or absence of IPTG.The binding constant for this interaction is approximately two orders of mag-nitude lower than that of the repressor-IPTG complex for the operator (Riggs,personal communication). Perhaps the repressor may interact with any DNAthat has an overall structure similar to poly(dA--dT), poly(dA--dT) but the correct sequence, as determined by the reactive groups in one or both grooves,

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 20: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

15 2 REZNIKOFF

can result in very tight repressor-operator interaction. The inducer may inhibit aconformational change in the repressor necessary for this very tight binding.

My prejudice is that only base specific interactions in the broad and narrowgroove can give the necessary specificity for the tight binding. There are toomany sequence specific proteins (repressors, restricting enzymes, modifying en-zymes, etc.) for eonformafional variations to do it alone.

ACra~OW~.~DCES~,~TS

The author’s research is supported by a National Science Foundation Grant(GB-20462), a Wisconsin Alumni Research Foundation Grant, and a NationalInstitutes of Health Biomedical Sciences Grant. I am particularly indebted toJ. Beckwith, T. Davis, P. E. Hartman and C. A. Thomas, Jr. for stimulating myinterest in gene regulation. I am also very thankful to M. Sundaralingam,R. Wells, and F. Blattner for helping me develop my ideas about protein-nucleicinteractions. Others whose comments or preprints were very helpful include:B. N. Ames, H. Berger, S. Bourgeois, K. D. Brown, J. Cairo, G. Craven, J. Dahl-berg, J. E. Davies, W. Dove, G. H. Echols, E. Fanning, T. Fanning, R. Gold-berger, J. Hardman, R. Hazelkorn, D. Jackson, A. Jobe, B. Mtiller-Hill, I. Pastan,A. Riggs, J. Sadler, R. Sommerville, T. Steitz, W. Summers, W. Szybalski,R. Wixom, H. E. Umbarger, 3. Miller, and T. Platt.

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 21: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON REVISITED 153

LITEIIATUEE CITED

1. Monod, J., Cohn, M. 1952. Advan.Enzymol. 13:67-119

2. Vogel, H. J. 1957. The ChemicalBasis of Heredity, ed. W. D. Mc-Elroy, B. Glass, ~76-89. Baltimore,Md.: Johns Hopkins. 848 pp.

3. Jacob, F., Monod, J. 1961. J. Mol.Biol. 3:318-56

4. Jacob, F., Monod, J. 1961. ColdSpring Harbor Syrup. Quant. Biol.26:193-209

5. Beckwith, J. R. 1970. The LactoseOperon, ed. J. R. B eckwith, D.Zipser, 5-26. Cold Spring Harbor,N.Y.: Cold Spring Harbor Labora-tory. 437 pp.

6. Gilbert, W., Mfiller-Hill, B. See Ref.5, 93-110

7. Miller, J. H. See Ref. 5, 173-888. Magasanik, B. See Ref. 5, 189-2209. Englesberg, E. 1971. Metabolic Regu-

lation, Metabolic Pathways, ed.H. J. Vogel, 5:257-96. New York:Academic. 3rd ed. 5 vols. 576 pp.

10. Zubay, G., Chambers, D. A. See Ref.9, 297-349

11. Brenner, M., Ames, B. N. See Ref. 9,350-88

12. Margolin, P. See Ref. 9, 389-44613. Umbarger, H. E. See Ref. 9, 447-6314. Calvo, J. M.; Fink, G. R. 1971. Ann.

Rev. Bioehem. 40:943-6815. Umbarger, H. E. 1969. Attn. Rev.

Bioehem. 38:323-7016. Mtiller-Hill, B. 1971. Angew. Chem.

lnt. Edit. 10:160-7217. Goldberger, R. F., Kovaeh, J. S.

Current Topics in Cellular Regula-tion, ed. B. L. Horecker, E. R.Stadtman. New York: Academic.Inpress

18. Brenner, S., Jacob,.F., Meselson, M.1961. Nature 190:576-80

19. Volkin, E., Astrachan, L. See Ref. 2,686--94

20. Gros, F. et al. 1961. Nature 190:581-85

21. Leive, L.,Kolin, V. 1967. J. Mol. BioL24:247-59

22. Kaempfer, R. O. R., Magasanik, B.1967. J. Mol. Biol. 27:475-94.

23. Jacquet, M., Kepes, A. 1969.Biochem. Biophys. Res. Commun.36:84-92

24. Hayasbi, M., Spiegelman, S., Frank-ll’n, N. C., Luria, S. E. 1963. Proc.Nat. Acad. Sci, USA. 49:729-36

25. Attardi, G., Naono, S., Rouvi~re, J.,lacob~ F, Gros~ F. 1963, CoM

S2~ring Harbor Syrup. Quant. Biol.:363-7226. Nathans, D., Notani, G., Schwartz,

J. H., Zinder, N. D. 1962. Proc.Nat. Acad. Sci. USA 48:1424-31

27. Summers, W. C. 1970. J. Mol. BioL51:671-78

28. Marrs, B., Yanofsky, C. 1971. NatureNew Biol. 234:168-70

29. Demerec, M. 1956. Cold SpringHarbor Symp. Quant. Biol. 21:113-20

30. Ames, B. N., Garry, B. 1960. Proc.Nat. Acad. Sci. USA. 45:1453-61

31. Kalckar, H. M., Kurahashi, K.,Jordan, E. 1959. Proc. Nat. Acad.Sci. USA 45:1776-86

32. Ames, B. N., Garry, B., Herzenberg,L. A. 1960. J. Gen. Microbiol. 22:369-78

33. Franklin, N. C., Luria, S. E. 1961.Virology 15:299-311

34. Englesberg, E. ~961. J. Baeteriol. 81:996-1006

35. Jacob, F., Perrin, D., Sanchez, C.,Monod, J. 1960. C. R. Acad. Sci.1727-29

36. Scaife, J., Beckwith, J. R. 1966. Cold~pring Harbor Symp, Quant. Biol.1:403~8

37. Guttman, B. S., Novick, A. 1963.Cold Spring Harbor Syrup. Quant.Biol. 28:373-74

38. Kiho, Y., Rich, A. 1965. Proc. Nat.Acad. Sci. USA 54:1751-58

39. Martin, R. C. 1963. Cold SpringHarbor Symp. Quant. Biol. 28:357-61

40. Imamoto, F., Morikawa, N.,Sato, K.1965. J. Mol. Biol. 13:169-82

41. Alpers, D. H., Tomkins, G. M. 1965.Proc. Nat. Acad. Sci. USA 53:797-803

42. Kepes, A. 1967. Biochim. Biophys.Acm 138:107-23

43. Imamoto, F. 1968. Proc. Nat. Acad.Sci. USA 60:305-12

44. Baker, R. F., Yanofsky, C. 1968.Proc. Nat. Acad, Sci. USA 60:313-20

45. Vogel, H. J. 1961. Cold SpringHarbor Symp. Quant. Biol. 26:163-72

46. Gorini, L., Gundersen, W., Burger,M. 1961. Cold Spring HarborSyrup. Quant. Biol. 26:173-82

47. Maas, W. K. 1961. Cold SpringHarbor Syrup. Quant. Biol. 26;!83-9!

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 22: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

154

48. Maas, W. K., Clark, A. J. 1964. J.Mol. Biol. 8:365-70

49. Berberieh, M. A., Venetianer, P.,Goldberger, R. F. 1966. J. Biol.Chem. 241:4426-33

50. Berberich, M. A., Kovach, J. S.,Goldberger, R. F. 1967. Proc. Nat.Acad. ScL USA 57:1857-64

51, Jacob, F., Ullman, A., Monod, J.1964. C. R. Acad. ScL 258:3125-28

52. Epstein, W., Beckwith, J. R. 1968.Ann. Reo. Biochem. 37:411-36

53. lppen, K., Miller, J. H., Scaife, J.,Beckwith, J. 1968. Nature 217:825-27

54. Miller, J. H., Ippen, K., Seaife, J. G.,Beckwith, J. R. 1968. J. Mol. Biol.38:413-20

55. Eron, L., Block, R. 1971. Proc. Nat.Acad. ScL USA 68:1828-32

56. Signer, E. R., Beckwith, J. R. 1966.J. Mol. Biol. 22:33-51

57. Miller, J. H. et al. 1970. J. Bacteriol.104:1273-79

58. Reznikoff, W. S., Beckwith, J. R.1969. J. Mol. Biol. 43:215-18

59. Eron, L., Beckwith, J. R., Jacob, F.See Ref. 5, 353-58

60. Chert, B. et al. 1971. Nature New Biol.233:67-70

61. Callahan, R., Blume, A. J., Bal-binder, E. 1969. J. Mol. Biol.51:709--15

62. Balbinder, E. et al. 1970. Genetics66:31-53

63. Fankhauser, D. B. 1971. The pro-moter-operator region of the hisoperon in Salmonella typhimu-RIOM. Ph.D. thesis. The JohnsHopkins Univ., Baltimore, 95PP-

64. Blattner, F. R., Boettinger, J. K.,Dahlberg, J. E. 1971. Fed. Proc.30:1315 Abstr.

65. Blattner, F. R. et al. 1972. Fed. Proe.31:472 Abstr.

66. Smith, T. F., Sadler, J. R. 1971. J.Mol. Biol. 59:273-305

67. Riggs, A. D., Reiness, C. G., Zubay,G. 1971. Proe. Nat. Acad. Sci.USA 68:1222-25

68. Emmer, M., deCrombrugghe, B.,Pastan, I., Perlman, R. 1970. Proc.Nat. Aead. Sci. USA 66:480-87

69. Silverstone, A. E., Magasanik, B.,Reznikoff, W. S., Miller, J. H.,Beckwith, J. R. 1969. Nature 221:1012-14

70. Perlman, R. L., deCrombrugghe, B.,Pastan, I. 1969. Nature 223:810-12

71. Beckwith, J. R., Grodzicker, T.,

REZNIKOFF

Arditti, R. R. unpublished experi-ments

72. Silverstone, A. E., Arditti, R. R.,Magasallik, B. 1970. Proc. Nat.,4cad, Sci. USA 66:773-79

73. Makmam, R. S., Sutherland, E. W.1965. J. Biol. Chem. 240:1309-14

74. Perlman, R. L., Pastan, I. 1968.Biochem. Biophys. Res. Commun.30:656-64

75. Ullmann, A. Monod, J. 1968. Fed.Eur. Biol. Soc. Lett. 2:57-60

76. Schwartz, D., Beckwith, J. R. SeeRef. 5, 417-22

77. Perlman, R. L., Pastan, I. 1969.Biochem. Biophys. Res. Commun.37:151-57

78. Zubay, G., Schwartz, D., Beckwith,J. 1970. Proc. Nat..4cad. Sci. USA66:104-10

79. de Crombrugghe, B. et al. 1971. Na-ture New Biol. 231:139-42

80. Voll, M. J. 1967. J. Mol. Biol. 30:109-24

81. Margolin, P. 1963. Genetics 48:441-56

82. Miiller-Hill, B, Crapo, L., Gilbert,W. 1968. Proc. Nat. Acad. Sci.USA 59:1259-64

83. Fraenkel, D. G., Banergee, S. 1971.J. MoL Biol. 56:183-94

84. Miller, J. H., Beckwith, J., Mi~ller-Hill, B. 1968. Nature 220:1287-90

85. Kumar, S., Szybalski, W. 1969. J.MoL Biol. 40:145-51

86. Reznikoff, W. S., Miller, S. H., Scaife,J. G., Beckwith, J. R. 1969. J. MoLBiol. 43:201-13

87. Reznikoff, W. S. Unpublished ex-periments

88. Mitchell, D., Reznikoff, W. S., Beck-with, J. R., Michels, C. A., Un-published experiments

89. Rose, J. K., Yanofsky, C. 1971. J.Bacteriol. 108:615-18

90. Roberts, J. W. 1969. Nature 224:1168-74

91. Nissley, S. P., Anderson, W. B.,Gottesman, M. E., Perlman, R. L.,Pastan, I. 1971. J. Biol. Chem. 246:4671-78

92. Gilbert, W., Miiller-Hill, B. 1967.Proc. Nat. Acad. Sci. US.4. 58:241 5-21

93. Riggs, A. D., Newby, R. R., Bour-geois, S., Cohn, M. 1968. J. MoLBiol. 34:365-68

94. Parks, J. S. et al. 1971. Proc. Nat.Acad. Sci. USA 68:1891-95

95. Ptashne, M. 1967. Nature 214:232-3496. Pirrotta, V., Ptashne, M. 1969. Na-

ture 222:541-44

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 23: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

THE OPERON

97. Lavall~, R., DeHauwer, G. 1970. J.MoL Biol. 51:435-47

98. Hsu, W., Weiss, S. B. 1969. Proc.Nat. Acad. Sci. USA 64:345-51

99. Steitz, J. A., Duke, S. K., Rudlaad,P. S. 1970. Nature 226:824-27

I00. Pollack, Y., Groner, Y., Aviv, H.,Revel, M. 1970. Fed. Eur. Biol.Soc. Left. 9:218-21

101. Revel, M., Aviv, H., Groner, Y.,Pollack, Y. 1970. Fed. Eur. Biol.Scl. Lett. 9:231-17

102. Berissi, H., Groner, Y., Revel, M.1971, Nature New Biol. 234:44-47

103. Lee-Huang, S., Ochoa, S. 1971. Na-ture New Biol. 234:236-39

104. Grunberg-Manago, M., Rabinowitz,J. C., Dondon, J., Lelong, J. C.,Gros, F. 1971. Fed. Eur. Biol. Soc.Lett. 19:193-200

105. Vermeer, C., Talens, I, Bloemsma-Jonkman, F., Bosch, L., Davies,J. W. 1971. Fed. Eur. BioL Soc.Lett. 19:201-06

106. Monod, J., Cohen-Bazire, G., Cohn,M. 1951. Biochim. Biophys. Acts7:585-99

107. Lederberg, J. 1951. Genetics in the20th Century, ed. L. C. Durra, 263-89 New York: MacMillan. 634 pp.

108. Cohen-Bazire, G., Jolit, M. 1953.Ann. Inst. Pasteur. 84:940-45

109. Pardee, A. B., Jacob, F., Monod, J.1958. C.R. Acad. ScL 246:3125-27

110. Willson, C., Perrin, D., Cohn, M.,Jacob, F., Monod, J. 1964. J. Mol.Biol. 8:582-92

111. Sadler, J., Novick, A. 1965. J. Mol.BioL 12:305-27

112. Bourgeois, S., Cohn, M., Orgel,L. E. 1965. J. Mol. Biol. 14:300-02

113. Barbour, S. C., Pardee, A. B. 1966.J. Mol. Biol. 20:505--15

114. Horiuchi, T., Ohshima, Y. 1966. J.MoL Biol. 20:517-26

115. Gilbert, W., Mtiller-Hill, B. 19f6.Proc. Nat. Acad. Sci. USA 56:1891-98

116. Riggs, A. D., Bourgeois, S. 1968. J.Mol. Biol. 34:361-64

117. Riggs, A. D., Suzuki, J., Bourgeois,S. 1970. J. Mol. Biol. 48:67-83

118. Miiller-Hill, B., Beyreuther, K., Gil-bert, W. 1971. Methods Enzymol.21:483-87

119. Bourgeois, S., Jobe, A. See Ref. 5,325-41

120. Davies, J., Jacob, F. 1968. J. MoLBiol. 36:413-17

121. Miiller-Hill, B., Crapo, L., Gilbert,W. 1968. Proc. Nat. Acad. Sci.USA 59:1259-64

REVISITED 155

122. Markovitz, A. 1964. Proc. Nat. Acad.Sci. USA 51:239-46

123. Chater, K. F. 1970. J. Gen. Micro-biol. 63:95-109

124. Hoar, D. I. 1971. J. BacterioL 108:582-83

125. Corredor, J., Reznikoff, W. S. 1972.Abstr. Ann. Meet. Am. Soc. Micro-bioL 72:52

126. Roth, J. R., Ant6n, D. N., Hartman,P. E. 1966. J. Mol. Biol. 22:305-23

127. Ant6n, D. N. 1968. J. Mol. Biol. 33:533-46

128. Lewis, J. A., Ames, B. N. 1972, J.Mol. Biol. In press

129. Roth, J. R., Ames, B. N. 1966. J.Mol. Biol. 22:325-34

130. Silbert, D. F., Fink, G. R., Ames,B. N. 1966. J. Mol. Biol. 22:335-47

131. Singer, C. E., Smith, G. R., Cortesse,R. Ames, B. N. 1972. Nature Inpress

132. Englesberg, E., Irr, J., Power, J.,Lee, N. 1965. J. Bacteriol. 90:946-57

133. Schwartz, M. 1967. Ann. Inst. Pas-teur 112:673-700

134. Power, J. 1967. Genetics 55:557-68135. Lindahl, G. 1970. Virology 42:522-33136. Smith, G. R., Magasanik, B. 1971.

Proc. Nat. Acad. Sci. USA 68:1493-97

137. Sly, W. S., Rabideau, K., Kolber, A.1971. The Bacteriophage X, ed.A. D. Hershey, 575-88. ColdSpring Harbor, N.Y.: Cold SpringHarbor Laboratory. 792 pp,

138. Wyche, J. H. 1971. Histidyl-tRNAsynthetase mutants and regulationin the histidine operon of Salmo-nella typhimurium. Ph.D. thesis.Johns Hopkins Univ., Baltimore.87 pp.

139, Kovaeh, J. S., Berberich, M. A.,Venetianer, P., Goldberger, R. F.1969. J. Bacteriol. 97:1283-90

140. Kovach, J. S., Phang, J. M., Ference,M., Goldberger, R. F. 1969. Proc.Nat. Acad. ScL USA 63:481-88

141. Kovach, J. S. et al. 1970. J. Bacteriol.104:787-92

142. Sadler, J. R., Smith, T. F. 1971. J.Mol, Biol. 62:139-69

143. Bhorjee, J. S., Fowler, A. V., Zabin,I. 1969. J. Mol. Biol. 43:219-22

144. Steers, E., Craven, G. R., Anfinsen,C. B. 1965. Proc. Nat. ,4cad. Sci.USA 54:1174-81

145. Englesberg, E., Squires, C., Meronk,F. 1969. Proc. Nat..4cad. Sci. USA62:1100-07

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 24: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

156 REZNiKOFF

146. Gielow, L., Largen, M., Englesberg,E. 1971. Genetics 69:289-302

I47. Hofnung, M., Schwartz, M. 1971.Mol. Gen. Genet. 112:117-32

148. Englesberg, E. Sheppard, D., Squires,C., Meronk, F. 1969. J. Mol. Biol.43:281-98

149. Wilcox, G., Clemetson, K. J., Santi,D. V., Englesberg, E. 1971. Proc.Nat. Acad. Sci. USA 68:2145-48

150. Zubay, G., Gielow, L., Englesberg,E. 1971. Nature New Biol. 233:164-65

151. Greenblatt, J., Schleif, R. 1971. Na-ture New Biol. 233:166-70

152. Burstein, C., Cohn, M., Kepes, A.,Monod, J. 1965. Biochim. Biophys.Acta 95:634-39

153. Jobe, A., Bourgeois, S. 1972. J. MoLBiol. In press

154. Eidllc, L., Neldhardt, F. C. 1965.Proc. Nat. Acad. Sci. USA 53:539-43

155. Freundlich, M. 1967. Science 157:823-25

156. Williams, L. S., Freundlich, M. 1969.Biochim. Biophys. Acta 186:305-16

157. Pi6rard, A., Glandorff, N., Mergeay,M., Wiame, J. M. 1965. J. Mol.Biol. 14:23-36

158. Freundlich, M., Burns, R. O., Um-barger, H. E. 1962. Proc. Nat.Acad. Sci. USA 48:1804-08

159. Freundlich, M. 1963. Biochem. Bio-phys. Res. Commun. 10:277-82

160. Szentirmai, A., Szentirmai, M,, Um-barger, H. E. 1968. J. Bacteriol.95:1672-79

161. Dwyer, S. B., Umbarger, H. E. 1968.J. Bac teriol. 95:1680-84

162. Calvo, J. M., Fretmdlich, M., Um-barger, H. E. 1969. J. Bacteriol.97:1272-82

163. Brown, K. D., Doy, C. H. 1966. Bio-chim. Biophys. Acta 118:157-72

164. Brown, K. D. 1968. Genetics 60:31-48

1.65. Brown, K. D., Somerville, R. L.1971. J. Bacteriol. 108:386-99

I66. rm, s. w. K., Davidson, H., Pittard,J. 1971. J. Bacteriol. 108:400-09

167. Buttin, G. 1963. J. Mol. Biol. 7:183-205

168. Mackie, G., Wilson, D. B. 1971. Fed.Proc. 30:1262 Abstr.

169. Arfin, S. M., Ratzkin, B., Umbarger,H. E. 1969. Biochem. Biophys. Res.Commun. 37:902-08

170. Zillig, W. et al. 1970. Cold SpringHarbor Symp. Quant. Biol. 35:47-58

171. Lozeron, H. A., Funderburgh, M. L.,Dahlberg, J. E., Stark, B. P.,Szybalski, W. 1972. Abstr. Ann.Meet. Am. Soc. Microbiol. 72:237

171a. Wu, A. M., Ghosh, S., Echols, H.1972. J. Mol. Biol. In press

172. Riggs, A. D., L/n, S., Wells, R. D.1972. Proc. Nat. Acad. Soc. USA69:761-64

173. Pfahl, M. 1971. Kartierung and char-akterisierung yon lac-i-genmutantenin Escherichia coll. Ph.D. thesis.Univ. Koln, Germany. 67_pp.

174. Adler, K. et al. 1972. Nature In press.174a. Platt, T., Weber, K., Ganem, D.,

Miller, J. H. 1972Proc. Nat. Acad.Sci. USA In press

175. yon Hippel, P. H. 1969. J. CellPhysiol. 74: Suppl, 1 235-38

176. Wilkins, M. H. F. 1957. Biochem.Soc. Syrup. 14:13-25

177. Kelly, T, J., Smith, H. O. 1970. J.Mol. Biol. 51:393-409

178. Printz, M. P, yon Hippel, P. H.1965. Proc. Nat. Acad. Sci. USA53:363-70

179. Gierer, A. 1966. Nature 212:1480-81180. Wells, R. D., Larson, J. E., Grant,

R. C., Shortle, B. E., Cantor,D. R. 1970. J. Mol. Biol. 54:465-97

181. H61~ne, C. 1971. Nature New Biol.234:120-21

182. Sundaraligam, M., Arora, S. K.1969. Proc. Nat. Acad. Sci. USA64:1021-26

www.annualreviews.org/aronlineAnnual Reviews

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.

Page 25: The Operon Revisited - pdfs.semanticscholar.org · The operon,functional polarized grouping of structural genes.--Jacob & Monod (3, 4) were struck by an observation made by Demerec

Ann

u. R

ev. G

enet

. 197

2.6:

133-

156.

Dow

nloa

ded

from

arj

ourn

als.

annu

alre

view

s.or

gby

Uni

vers

ity o

f W

isco

nsin

- M

adis

on o

n 03

/29/

07. F

or p

erso

nal u

se o

nly.