Vol. 149, No. 3JOURNAL OF BACTERIOLOGY, Mar. 1982, p. 1071-10810021-9193/82/031071-11$02.00/0

Physical and Genetic Localization of ilv Regulatory Sites in X

ilv BacteriophagesJOHN E. GRAY,1 D. CLARK BENNETT,2 H. E. UMBARGER,2 AND DAVID H. CALHOUN'*

Department of Microbiology, Mount Sinai School of Medicine, New York, New York 10029,1 and PurdueUniversity Biochemistry Program and Department ofBiological Sciences, Purdue University, West Lafayette,

Indiana 479072

Received 8 October 1981/Accepted 28 October 1981

A set of nine A dilv phages were used to transduce bacterial recipientscontaining point mutations or deletions in the ilv genes located at 84 min on theEscherichia coli K-12 chromosome. This genetic analysis indicated that twophages carry the entire ilvGEDAC cluster; others carry the complete ilvC geneand, in addition, bacterial DNA that extends to a termination point between ilvAand ilvC, within ilvD, within ilvE, or within ilvG. DNA extracted from the dilvphages was digested with EcoRI, HindlIl, KpnI, PstI, Sall, and SmaI. Therestriction maps revealed that these phages were generated after insertion at fourdistinct insertion sites downstream (clockwise) of iivC. The physical relationshipsbetween the various phages were further examined by electron microscopicheteroduplex analysis. The physical maps of the phages thus generated werestraightforward and in complete accord with the genetic data. No evidence forgenetic rearrangements of ilv DNA in the phage was obtained, thus validatingconclusions based on the use of these phages in previous and ongoing researchprojects. Bacterial cells with deletions of the ilv genes were made lysogenic with Adilv phage to examine the regulation of ilv genes present in the phage. The resultsconfirm previous studies showing that one site for control by repression andderepression is upstream (counterclockwise) of ilvG. It was shown, in addition,that the activities of dihydroxy acid dehydrase and threonine deaminase wereincreased when the prototrophic lysogens were grown with 20 mM leucine. Sincethis increase was exhibited even when the ilvG-linked control region was notcarried by the A dilv phage, additional control sites must be located within theilvEDA region of the ilvGEDA transcription unit.

In the past, specialized transducing phageshave been valuable in the analysis of biochemi-cal genetics in bacteria. The use of bacterio-phage X derivatives was originally restricted to afew Escherichia coli K-12 genes, but more re-cently methods have been devised that permitthe detection of phage A derivatives formed invivo that carry a wide range of bacterial genes(26, 27). A set of X dilv phages have beenisolated carrying overlapping segments of the ilvgenes at 84 min. These phages were previouslyused in genetic and biochemical analyses todetermine the order of ilv structural and regula-tory genes on the E. coli K-12 genome (2, 23).However, the validity of these conclusions isopen to question, since rearrangements of genesin specialized transducing phages have beenreported (21). Also, since only a limited numberof bacterial recipients with mutations in thesame gene have been tested, it has not beenpossible in many instances to differentiate be-tween the presence of all or part of an ilv gene orcontrol element in the X dilv phage.

In the studies reported here, we obtain de-tailed restriction maps, examine critical DNAheteroduplex structures, and test genetically forthe endpoint termini within ilv genes, particular-ly for the first structural gene, ilvG, and itscontiguous control site, for the X dilv phage.These studies clearly establish the physicalstructure of these phages, demonstrate that norearrangements have occurred in ilv DNA (7, 20,21) present in the phage, and reveal their modeof formation by aberrant excision. This knowl-edge of the precise context of ilv structural genespresent in the phage prompted an investigationof the presence of genetic control sites. Weconfirmed previous studies (4, 5, 16, 17, 19, 31)indicating that at least one control site precedesthe first structural gene, ilvG. We find clearevidence for regulation in at least one internalsite as well, and use the phage to identify itslocation.While this manuscript was in preparation, the

DNA sequence of the ilvG gene was completed,and the nature of the ilvO site was elucidated

1071

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1072 GRAY ET AL.

(16). Although i/lO has many properties of anoperator gene (mutations are cis-dominant andelevate expression of the linked ilvEDA structur-al genes), and this originally led to its designa-tion (discussed in reference 31), it has now beenestablished that ilvO is a 10-base-pair site withini/vG (the first structural gene of the i/.GEDAtranscription unit). Mutations in i/vO (actuallyilvG mutations) are 1-base-pair deletions or 2-base-pair insertions (16, 16a) that restore thetranslational reading frame of the entire i/lGgene. The wild-type E. coli K-12 carries a natu-rally occurring frameshift site within the i/vGgene (16). The product of the mutant, but not thewild-type, i/vG gene is a catalytically active cx-acetohydroxy acid synthase II isozyme that isresistant to inhibition of catalytic activity byisoleucine (13, 31). Thus, the mutant, but not thewild type, is Valr (resistant to growth inhibitionby valine). We feel that it is preferable, inagreement with a previous suggestion (16), todiscontinue the use of the misleading ilvO mne-monic, although any terminology is uncomfort-able for such a circumstance, since ordinarilythe wild type is genetically and phenotypicallyproficient and the mutant is deficient. Neverthe-less, the wild type is designated i/vG' (pheno-type, IlvG-- Vals; formerly i/vOt), and the mu-tant is designated ilvG (phenotype, IlvG' Val';formerly il.vO).

It should be emphasized that DNA sequenceanalysis is an ancillary technique that does notreplace conventional genetic analysis of pheno-typic determinants. The correlation of il/ pheno-types with specific physical DNA segments isthe topic of this paper. The experiments report-ed here provide the essential documentation forconclusions as to the effect of observed DNAsequence changes (16, 16a) on the correspondingilv phenotypes.

MATERIALS AND METHODSBacteria and bacteriophages. The bacterial and

phage strains used are listed in Table l. The methodsused for the isolation of the X dill phages and theirinitial characterization have been described (2, 23).

Transduction. Techniques used for specializedtransduction with the A dilt phages have been de-scribed (2, 23). In all cases, lysogenization by thetransducing phages was prevented by selecting recom-binants growing at 42°C. Since unusual cross-feedingpatterns between pairs of isoleucine and valine auxo-trophic mutants have been observed (23), the resultsreported have been confirmed by cloning and retestingof several transductants from each cross, includingboth valine-susceptible and valine-resistant types, onnonselective media (presence of leucine. isoleucine,and valine).Enzyme assays. The colorimetric assays for threo-

nine deaminase and dihydroxy acid dehydrase were aspreviously described (14, 29). Since the (-1857 allele ispresent in the helper and defective phage, the cells

were grown at 30°C to prevent prophage induction.For this reason, and because of the location andtranscriptional orientation of the il/ genes in thephage, read-through from phage promoters to i/l genesis blocked. Where indicated, the dehydrase was as-sayed with ,3-dihydroxy-3-methylvalerate instead of(x.4-dihydroxyisovalerate. Transaminase B was as-sayed as described by Duggan and Wechsler (10).Whether multivalent repression was exhibited in pro-totrophic, valine-susceptible strains was determinedby comparing the enzyme levels in cells grown in thepresence and absence of 0).6 mM leucine, 0.6 mMisoleucine, and 1 .2 mM valine. The repressing mediumfor the leucine auxotrophic, valine-susceptible lyso-gens was prepared by supplementing the minimalmedium of Davis and Mingioli (9) with 0.4 mM leucine,0.4 mM isoleucine, and 0.8 mM valine. For limitingisoleucine or limiting leucine, the concentration ofisoleucine or leucine, respectively, was reduced to0.02 mM. In addition, enzyme activities in the proto-trophic strains were also examined in cells grown inminimal medium supplemented with 20 mM leucine.

Preparation of phage and extraction of DNA. Themethods for growth of heat-inducible, lysis-defectivephages and the extraction of DNA have been de-scribed (32). A diiv58, A di/v37, A dilv26, and A dilv22were separated from the helper phage by isopycniccentrifugation in cesium chloride. A di/v62 and A dillZ73failed to separate from the helper during centrifugationand were used without separation.

Restriction endonuclease analysis. The DNA pre-pared from the phages was digested with BamHl,EcoRl, HitdIll, KpnI, PstI, Sall, and Stnal, obtainedfrom New England BioLabs. The cleaved fragmentswere separated by electrophoresis in horizontal agar-ose gels as described previously (19).

Heteroduplex analysis. The procedures for formingheteroduplexes and analyzing them by electron micro-scopic examination have been described (19).

RESULTS

Genetic analysis of the X dilv phages. Thepresence of i/v genes in the K dil/ phages haspreviously been examined in preliminary geneticexperiments by using only a few representativerecipient strains defective in specific structuraland regulatory genes (2, 23). The presence of acomplete and functional copy of the i/vA, ilvD,and ilvE genes was previously determined bymeasuring increased enzyme levels after thermalinduction of appropriate K dilv lysogens (2). Theanalysis in Tables 2 and 3 includes strains withadditional mutant alleles that permit a geneticdetermination of the termination point of bacte-rial DNA present in the K dilv phage within i/i'Dfor A dilv22, within i/vE for K di/v,62 and A dilv73.and within i/vG for A dilv26, A dil/v37, and Adil/43. Phage K dilv29 failed to generate proto-trophic transductants with any i/vA auxotrophstested, including one bearing the i1vA601 lesion(data not shown), which is the most i/vC proxi-mal of the known i/vA mutations. Thus, i/v DNAcarried by A dilv29 probably extends only to apoint between i/vA and il/C. Phages K di/v58 and

J. BACTERIOL.

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURE OF A dilv PHAGES 1073

TABLE 1. Bacterial and bacteriophage strains used

Strain Genotype Source or reference

F- A(ilvDAC)115 thi mtl malA rpsL his trpC tsx lacZ X-

F- galT12 X-

F+ ilvC462 rbs-224

F- rbs-221 leu455 galT12 X-F- ilv2049 leu455 gaIT12 X-F- A(ilvDACJ15) metE201 leu455 gaIT12 X-F- ilvG468 ilvA2058 rbs-215F- ilv2076 rbs-221 leu455 galT12 X-F- ilvG2095 ilvE2104 leu455 rbs-221 galT12 A-F- ilvG2096 ilvE2105 leu455 rbs-221 galT12 X-F- ilvG2113 ilvG2096 rbs-221 leu455 galT12 A-F- ilvG2111 ilvG2095 rbs-221 leu455 galTJ2 A-F- ilvG2095 ilvG2111 leu455 rbs-221 galT12 A-F- ilvG2096 ilvG2116 leu455 rbs-221 galT12 X-ilvC462F- A(ilvDAC)115 thi mtl malA rpsL his trpC tsx lacZ X

dilvGEDAC58F- A(ilvDAC)115 thi mtl mal rpsL his trpC tsx lacZ X

dilv'GEDAC26F- A(ilvDAC)115 thi mtl mal rpsL his trpC tsx lacZ X

dilv'EDAC75F- A(ilvDAC)115 thi mtl mal rbsL his trpC tsx lacZ X

dilv'GEDAC37

A(T-attP) [ilv'DAYC-rho] c1857 Sam7A(T-attP) [ilv'GEDAYC-rho'] c1857 Sam7A(FI-attP) [ilv'GEDAYC-rho] c1857 Sam7A(Z-attP) [rrfC-ilvGEDA YC] c1857 Sam7A(A-attP) [ilvDAYC-rho] c1857 Sam7A (A-attP) [ilvDAYC-rho] c1857 Sam7

B. Bachmann, E. coli Ge-netic Stock Center,Yale University Schoolof Medicine

Pledger and Umbarger(20)

Spontaneous Rbs- deriva-tive of CU1010 isolatedby E. L. Kline

Smith et al. (28)Watson et al. (33)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Smith et al. (28)Lysogenization of AB3590

Lysogenization of AB3590

Lysogenization of AB3590

Lysogenization of AB3590

Baez et al. (2)Baez et al. (2)Baez et al. (2)Baez et al. (2)Baez et al. (2)Baez et al. (2)

a The ilvG468, -2095 and -2096 alleles are activating ilvG mutations that confer the IlvG+ and Valr phenotype(previously ilvO); the iIvG2111, -2112 and -2113 alleles inactivate the catalytic activity of the activated ilvG gene

product. See text for discussion.

X dilv42 transduce all ilv auxotrophs tested andpresumably carry the entire ilv cluster (Fig. 1).The rationale for determining the presence of

ilvE, ilvD, ilvA, and ilvC structural genes fromthe experiment described in Table 2 is straight-forward. Since selection was made at 42°C, theprototrophs selected in each case were the resultof recombination between host- and phage-car-ried genes. Thus, the formation of prototrophictransductants at frequencies in excess of rever-

sion frequencies is evidence for the presence ofwild-type DNA in the A dilv phage correspond-ing to the site of the altered DNA in the recipientgenomes. The rationale for determining the pres-ence of ilvG in the A dilv phage depends upon theValr or Vals phenotype as a selected (Val) or

unselected (Valr or Vals) marker. The Valr phe-notype is dependent on ilvG activator mutations(e.g., ilvG264, ilvG671, ilvG468, ilvG2095, and

ilvG2096; previously called ilvO mutations) thatlead to the synthesis of a catalytically active a-

acetohydroxy acid synthase 11 (16). Additionalmutations in ilvG (e.g., iIvG2111, ilvG2113,ilvG605, and ilvG2112) abolish a-acetohydroxyacid synthase II activity (present in strains withan activated ilvG), presumably due to second-site mutations in essential coding regions ofilvG. The Valr transductants of strains CU853and CU856 (Table 2) resulted, therefore, fromrepair of the inactivating ilvG2113 and ilvG2111lesions, but retention of the respective activatingmutations ilvG2096 and ilvG2095. Note that theilvG2111 allele was mapped both by the directselection procedure (Table 2, strain CU856) andwithout selection (Table 3, strain CU859); ineach case, phages A dilv58 and A dilv42 were

found to carry ilvG2111+, but X dilv37, dilv43,and X dilv26 did not. Whereas the Valr pheno-

BacterialaAB3590

CU4

CU449

CU504CU505CU519CU595CU655CU692CU693CU853CU856CU859CU860Cu1o0oMSR168

MSR169

MSR170

MSR114

PhageA dilv22A dilv26A dilv37A dilv58X dilv62A dilv73

VOL. 149, 1982

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

00o o o

o~~~~~~~~~~~~~_0~~~~~~~~~~~

xx=

o x o,10

00, _ 0)+

oc- 0 Q

Z x .00i OC)0)

> O C

00o o o o ._0 ,

XX x x X > 0

0))

H~~~~~H~~~~ - Q0.

ZX Z X X X 0 0

oN , (N x0 000 0 0 0 0 0 0 n=>

X X X X X X X X Q D 4

O0r~ 4CC ( 0 > 0 O i t ') < _0..

OO'OO OO- O C C.)

+ + + + t:v

tm°+ -, + +o +

C)

00 00>0 ~ -

00 \0 £ > so00o0o 000p zt.D > D > D > > D > > D > D > U rc0

u o 0 o uc 0U 4

1074 GRAY ET AL. J. BACTERIOL.

f)

.h

i

"I

Xxv0

gmC .

XXZ

x:

0 °

X< xON

XX0L

, >Kz

XX'<

0421040

th

ct-.

0

C.tc:

02

-

ct

':1CO

CO

a..

C..

0

Z.

D0)

0'

-0

.°

0

0

0

0)

L-I

H: xX XON

00 rCC)

.X ZiO

+ 00rO10

c-

C11 0.O~c o

0~:-c

ON1 -

(NC.,*)N~Nj- aCCU~W

0

C/

0..

CO

._

L-

m;

C,

4

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURE OF k dilv PHAGES 1075

Iv rho

G (O) G E D A I|Y CI -I I

2111 2113 605 264 2112 2061 02104 16 454 60167 2 105 L468 2 09j A20762095 2 0

2096 12 Ail 5

1 X29~

A22

X73

X43

A37

X58j

FIG. 1. Genetic map of ilv DNA carried by A dilvphages. The genes in the ilv cluster are arranged as

they occur in the E. coli chromosome, with the clock-wise direction on the chromosome map as usuallyrepresented reading from left to right. The gene ar-

rangement is not drawn to scale. The bars belowrepresent the genetic material carried by the indicatedphages. The allele numbers enclosed in brackets (e.g.,264, 2112, 2061, etc.) were not ordered in this analysis.(0) indicates the location of the frameshift site in theilvG gene.

type can be selected directly, the Vals pheno-type must be identified as an unselected marker.The identification of ilvG2095+ and ilvG2096+derivatives of strains CU692 and CU693, respec-

tively (Table 3), was based on the presence ofthe unselected Vals phenotype among ilvE+ pro-totrophic transductants. The Valr recombinationwould thus have resulted from a crossover eventbetween the ilvE auxotrophic marker and therespective ilvG lesion.The order inferred from these and other cross-

es with the A dilv phages for these ilv loci isG2111-G21J3-G605-(G2112, G264, G671, G468,G2095, G2096, E2061, E2105, E2109, E2110,E12)-ED2104. The mutant alleles within paren-theses have not been ordered with respect toeach other by the K dilv mapping procedure,although four of the (G671, G468, G2095, andG2096) have been ordered by DNA sequenceanalysis (16a). Evidence from three-factor cross-es presented earlier (28) indicated the orderiIvG2096-i1vG2H12-i1vE; this genetic order is con-sistent with the results presented here, since allphages tested carry either both or neither of theilvG2096+ and ilvG2112+ alleles. The previousinterpretation (28) of these three-factor crosseswith either that ilvG2096 was unusual in itslocation or that the ilvG-activating and -inacti-vating markers were interspersed. The results ofthe crosses described here support the latterpossibility and would indicate that ilvG2112 (in-activating) is downstream (i.e., on the ilvE side)of the ilvG activation site, since it was the onlyone of the three ilvG inactivating markers exam-ined that A dilv43 and k dilv26 could repair.

TABLE 3. Detection of ilvG+ in X dilv phage as unselected markers in ilvE+ transductants"

Unselected Phage donor and ilv genes present in phageBacterial recipient marker X dilv42 X dilv58 X dilv37 X dilv43 X dilv26 X dilv73' A dilv62'detectedb (GEDAC) (GEDAC) ('GEDAC) ('GEDAC) ('GEDAC) ('EDAC) ('EDAC)

CU692 ilvG2095+ 41/78 45/96 8/96 8/63 6/96 0/96 0/66[iUvG2095, (53) (47) (8) (13) (6) (0) (0)A(ilvED)2104]

CU693 ilvG2096+ 45/145 44/95 24/141 19/141 8/137(ilvG2096 (31) (46) (17) (13) (6)ilvE2105)

CU859 iIvG2111+ 85/98 221/235 196/196 100/100 190/190(ilvG2095 (87) (94) (100) (100) (100)ilvE2109ilvG21J1)

CU860 ilvG2112+ 92/98 126/146 132/146 85/100 115/146(ilvG2096 (94) (86) (90) (85) (79)ilvE2110ilvG2112)a Prototrophic transductants (Table 2) were tested for their Vals phenotypes. The ratio given for each

represents (number of Vals transductants)/(number of transductants tested). For each cross, representativetransductants (Valr and Valr) were streaked for single-colony isolation on nonselective media (leucine,isoleucine, and valine); the phenotype (Val' or Val') was confirmed by transferring several well-isolated coloniesto plates with and without valine. Numbers in parentheses are percentages.

b The presence of i1vG+ alleles in the phage was detected by the formation of Val' transductants for CU692 andCU693 and by the presence of Valr transductants for CU859 and CU860. Alleles ilvG2095 and -2096 confer Valr,and alleles iIvG2111 and -2112 reverse Valr phenotypes. See text for discussion.

c No transductants are formed by phage A dilv73 and A dilv62 with strains CU693, CU859, and CU860 asrecipients (Table 2).

I I

VOL. 149, 1982

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1076 GRAY FTAL.J.BTEO.

Thus, the data in Table 2, obtained by thisdeletion mapping technique, and the data ofSmith et al. (27), obtained by three-factor cross-es with phage P1, are mutually reinforcing andprovide convincing genetic data that the activat-ing mutations in ilvG (previously designated i1vOmutations) lie within the coding part of the ilv,Gstructural gene. This conclusion has subsequent-ly been confirmed by direct DNA sequenceanalysis (16, 16a). From these genetic data (Ta-bles 2 and 3), confirmed by the physical analysis(see below), it can be determined that the K dilv,phages carry the following combinations; GE-DAC (K dil42, -58), 'GEDAC (K dilv37, -43, -26)'EDAC (K dilv62, -73), 'DAC (K dilv22), and C (Kdilv29).

Evidence interpreted to indicate the presenceof several ilv promoters (P1, P2, and P3) hasbeen reported (order, 1G-2E3DA) from the analy-sis of strains with polar mutations or insertionsin ilvGEDA (3-5, 29). A consensus promoterDNA sequence is located between ilvG and ji1vE(16, 18). Since these K dilv, prophages expressthe i/v genes only from ilv promoters (phage Krepressor inhibits transcription from phage pro-moters), the physical location and potential reg-ulation from these promoters in the K dilv phagecan be directly assessed in Ailv hosts (see be-low). In this regard, it should be noted thatsimilar analyses from i/v, segments cloned intomany plasmids (e.g., pBR322) are often ex-pressed by read-through from plasmid promot-ers. Also, interpretation of results from insertion

elements (3-5, 29) can be complicated by (i)promoters present in the insertion elements, (ii)undetected transposition to other nearby sites,and Oiii) ambiguous localization of these ele-ments within structural or regulatory sites, incontrast to the physically defined termini in theX dilv, phage.Physical mapping of the X dilv phages. The K

dilv phages were examined by restriction en-zyme cleavage, using EcoRI, Hindlll, Sail,SmnaI, Pstl, and Kpnl, and by electron micro-scopic analysis of heteroduplexes to give theresults depicted in Fig. 2. The restriction en-zyme analysis was simplified, since some re-striction enzyme data have been published forthe ilvGEDA YC genes present in X h80 dilv andthe Clark and Carbon plasmids, including about2.5 kilobases upstream or on the counterclock-wise si'de (as the chromosomal map is usuallyrepresented) of ilvG and less than 0.10 kilobasesdownstream of ilvC (4, 9, 13, 16, 25). In addition,a phage (K dilv) that carries rrnC (or rrnX) (21)and il/v has previously been studied, but with aprimary focus on the rRNA and tRNA genes.The DNA downstream of ilvC has not previous-ly been studied by restriction enzyme analysis.The rho-1uS allele is carried by phages K dilv,22,K dilv62 , and Xdilv73 (J. E. Gray, S. K. Guterman,and D. H. Calhoun, unpublished data). Therifampin hypersusceptibility of a chromosomalrho-115 mutation is suppressed by a pBR322derivative that carries the 6.6-kilobase Hindilllfragment downstream of i/sC (Gray and Cal-

PHYSICAL MAPS OF MdN/ PHAGES8 6 4 2 0 -2 4 -6 -8 -10 -12 -14 -16 -18 -20 -22 -24 -26 -28 -30 -32V ---T ~ I' m -1 T T- V -T ---F T - -

IrrrK fq~C Ii/vri/CG"OIEI P A 11Y 11CI

fq7C rr-fC '/irC a )id//v58211I t-2 ~0 . 11111I

Xd/A5'371.3 io.2 u ti 2

)id/,A26

Xd//v2257 - II II . 17 t vyE2

VI II 21IXd//v62

Iv. 11 411 77v 4 774

Xd//v73.1.57.. 11 th

A')P70 2 4 6 8 10 12 14 16 18 20

---5W£IF .1

Xis e.v red- re; -cl,22 24 26 28 30 32 34 36 38 40 42 44 46 48KiLOBASES

FIG. 2. Physical map of the DNA in the region of the i/v, gene cluster carried on the E. (o/i chromosome andon several K di/v, phages. The arbitrary kilobase coordinates for the chromosome are based on a zero point at theapproximate beginning of the i/v,C gene. The K kilobase coordinates are based on the K vegetative map. The 0"Osite indicates the presence of a naturally occurring frameshift site in the wild-type i/vG gene.Symbols: 1, EcoRl; 73., Hindlll; t, Kpnl;, V. Pst; ., Sa/l; I Smnal.

I1- ,I -, .I

J. BACTERIOL.

two,

im11

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURE OF A dilv PHAGES 1077

houn, unpublished data). Since none of thephages carries cya, these results support theorder ilv-rho-cya reported by Das et al. (8),rather than ilv-cya-rho (1).The structures of the X dilv phages as revealed

by restriction enzyme analysis were supple-mented and refined by examination of severalheteroduplexes between the complementarystrands of the parental phages and X dilv58, Xdilv37, A dilv26, A dilv62, and A dilv73.

In addition, heteroduplexes between A dilv58and both K dilv37 and A dilv26, between A dilv37and A dilv26, and between K dilv62 and K dilv73were examined. These combinations were suffi-cient to verify that they resulted from excision ofilv DNA by prophages inserted at three differentsecondary K attachment sites and that each hadunique left-arm K ilv DNAjunctions. The precisejunctions of K dilv37 and A dilv26 were impor-tant, since the genetic data of Table 2 indicatedthat K dilv37 carried more ilv DNA than did Kdilv26. This conclusion was verified by the het-eroduplex analysis between the two (Fig. 3a),which revealed a bubble structure in which thelarger bubble arm was formed by the extra left-arm K DNA carried by A dilv26 and the muchsmaller bubble arm was formed by the extra ilvDNA (about 300 bases) carried by A dilv37. Onthe other hand, neither the genetic tests nor the

heteroduplexes permitted an unequivocal com-parison of the ilv termini in K dilv62 and K dilv73,since only a single-stranded loop was formed(Fig. 3b). This loop was made up primarily ofleft-arm K DNA carried by K dilv62 but missingfrom A dilv73. It might also have included someilv DNA not present in A dilv73. It can thus beconcluded only that, if A dilv73 contains any ilvDNA not present in K dilv62, it is a length belowthe limits of resolution of the electron micro-scopic technique used.A heteroduplex between A dilv22 and the

parental phage was not examined, since therestriction enzyme analysis allowed a clear dem-onstration that it was derived from a prophageinserted at a fourth secondary A attachment site.Furthermore, the cleavage fragments were suchthat the junction between the left arm of A andthe ilv DNA could be as precisely defined asthey could have been by heteroduplex mapping.The genetic analysis (Tables 2 and 3) and the

physical characterization (Fig. 2) permit a corre-lation of genes and fragments of genes withrestriction enzyme sites. This correlation con-firms and extends the analyses based on cloningfragments from A h80 dilv (7, 20).

Expression of ilv genes in the X dilv phages. TheK dilv phages carry the ilvGEDA genes in over-lapping segments, including GEDA, 'GEDA,

FIG. 3. (a) Heteroduplex prepared between DNA strands of K dilv26 and k dilv37. The long single-strandedarm of the bubble structure is derived from left-arm k DNA carried by k dilv26; the short single-stranded arm isderived from chromosomal DNA in the vicinity of the ilvG gene. (b) Heteroduplex prepared between DNAstrands of A dilv62 and K dilv73. The single-stranded DNA loop is derived from the left-arm K DNA carried by Kdilv62 (but not by A dilv73). It could also be derived from any DNA in the vicinity of the ilvE gene carried by Kdilv62 but not by K dilv73. No ilv DNA in X dilv73 can be detected that is not also carried by K dilv62.

VOL. 149, 1982

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1078 GRAY ET AL.

'EDA, 'DA, and A. Lysogens derived fromstrain AB3590[A(i1vDAC)l15] were prepared totest for the presence of the ilv-specific regulationof ilvD and ilvA expression. The chromosomalilvGEDA gene products are subject to repressionin the presence of excess leucine, isoleucine,and valine. The synthesis of the ilvD gene prod-uct, dihydroxy acid dehydrase, and the ilvAgene product, threonine deaminase, were re-pressible (Table 4; compare growth with andwithout the branched-chain amino acid supple-ment) in lysogens containing X dilv58 but not inlysogens containing X dilh37, X dilv26, or Xdilv73. Thus, the site essential for multivalentrepression control is present upstream of the ilvDNA carried by phages X dilv26 and X dilv37;this site is presumably the attenuator identifiedby direct DNA sequence analysis (17, 22).There have been several reports showing that

prototrophic strains of E. coli growing in mini-mal medium supplemented with leucine exhibitan increased expression of the ilv genes due toan isoleucine limitation (13, 15, 23, 31). In pre-liminary experiments with strain CU4, it wasobserved that the maximal increase in ilv geneexpression was obtained when the medium wassupplemented with 20 mM leucine. This concen-tration also increased the generation time 1.5-fold (Table 4). The very high level of threoninedeaminase activity exceeded the typical dere-pressed levels seen in regulatory mutants or inamino acid-limited auxotrophs by approximately10-fold; it was approximately the same as thatobserved when an auxotrophic ilvG468 (activa-

tor) mutant was limited for valine or isoleucine(28).

It is striking that the leucine-induced elevationin activity of the dehydrase was less than that ofthreonine deaminase. In this respect, the leucineeffect was reminiscent of the downstream ampli-fication observed when isoleucine is limited (29).It was of interest to determine whether thiseffect of il/v gene expression was dependent onthe normal i/v-specific regulatory locus carriedby strain CU4 and by A dilv58 but missing fromthe other X dilv derivatives listed in Table 4. Asthe table shows, the effect of 20 mM leucine wasexhibited not only by the ilv genes carried on thephage with the normal ilvGEDA control regionbut also by two of the three that had not retainedthat site. Only in A dilv73, in which the ilv DNAextended from a point downstream of ilvC onlyinto, but not through, i/vE, did the addition of 20mM leucine fail to elevate the level of threoninedeaminase activity but, instead, was inhibitoryto growth. It should be noted, however, in thisstrain that in minimal medium the expression ofthe ilvD and ilvA genes was very low, andgrowth of the strain was extremely slow. Thequestion remains of whether the failure of thestrain to grow was due to loss of some elementbetween ilvG and ilvE that responds to excessleucine or to the inability of the low level ofthreonine deaminase to overcome the inhibitoryeffect of leucine on the enzyme. It is clear,however, that the response to 20 mM leucine isnot dependent upon the presence of the ilv-GEDA promoter-attenuator region.

TABLE 4. Expression of threonine deaminase (ilvA) and dihydroxy acid dehydrase (ilivD) specified by thegenes carried by several X dilv, prophages

Sp act

Straina [A(ilvDAC115]

CU4+(ilv+ A-)

MSR170(X dilv73) ['EDAC]C

MSR169(A dilv26) ['GEDAC]

MSR114(XA dilv,37) ['GEDAC]

MSR168(A dilv58) [GEDAC]

Growth medium

RepressingMinimal20 mM leucineRepressingMinimal20 mM leucineRepressingMinimal20 mM leucineRepressingMinimal20 mM leucineRepressingMinimal20 mM leucine

Generationtime (min) Threonine Dihydroxy acid

deaminase dehydraseh

90 123 9993 195 162132 2,120 530102 11 5780 14 6NGd NG NG72 84 3678 67 37138 260 9884 46 1490 45 20156 221 7890 61 1884 126 99120 700 296

a Strains MSR170, MSR169, MSR168, and MSR114 contain the X(ilvDAC)115 deletion and are lysogens ofstrain AB3590 containing the indicated prophage. Strain CU4 is an illv nonlysogenic control strain.

b The substrate used was a,,-dihydroxyisovalerate.' ilv gene carried on the phage genome.d NG, No growth.

J. BACTERIOL.

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURE OF K dilv PHAGES 1079

Derepression of the ilv genes in the Xdilv phagesby limiting isoleucine and by limiting leucine. Thecapacity of the ilv genes carried on the A dilvphages to respond to a limiting amino acid signalwas examined in strains carrying the leu455marker. In that way it was possible to examinethe effects of limiting leucine and, since thestrains were valine susceptible, of limiting iso-leucine. The phage recipients carried deletionsso that the enzyme activities measured werethose specified by the HivE, -D, and -A genes ofthe phages (or the ilvD and -A genes of K dilv73).Table 5 shows the activities observed in cellsgrown under the various conditions.

Strain CU505 lysogenized with A dilv58 exhib-ited the same quantitative pattern of derepres-sion as did the control strain, CU504, whichcarried the ilv gene cluster on the chromosome.The apparent derepression of threonine deami-nase was much greater than that of transaminaseB when isoleucine was limiting. This noncoor-dinate behavior of the ilvGEDA operon has beenseen before and is thought to be unrelated to theilv-specific transcriptional control exhibitedover the operon (28). The same effect on theactivity of threonine deaminase was observed inthe strains lysogenized with X dilv37 and Adilv26, neither of which exhibited derepressionof threonine deaminase with leucine limiting orof transaminase B with either amino acid limit-ing. These phages lack the attenuator-promoterregion of the ilv gene cluster. That an effect oflimiting isoleucine is also exhibited by the ilvAgene carried on X dilv73 indicates that the effectis mediated by the ilvA gene itself or by DNAthat lies well after the weak promoter between

ilvG and ilvE and which is probably functional inK dilv26 and K dilv37.

DISCUSSIONThe results reported here establish the genetic

and physical structure of a set of K dilv phagesthat are being used to probe the nature of thecontrolling elements regulating isoleucine andvaline biosynthesis. The restriction enzymeanalysis of the bacterial DNA in the K dilvphages and the examination of several heterodu-plexes has permitted a correlation of genes withDNA segments that is in accord with previouslypublished data (7, 20, 21). The K dilv phagesarose by insertion at four distinct sites down-stream or to the right of ilvC (as the chromosomeis represented in Fig. 2) and to the right of thefusion between chromosomal DNA and the leftarm +80 DNA of A h80 dilv. The existence ofsuch a nonrandom cluster of secondary pro-phage insertion sites was also observed by Shi-mada et al. (26). Since the DNA of bacterialorigin is inserted immediately to the left (asdrawn in Fig. 2) of the phage att site, it is clearthat they arose by the K gal type of aberrantexcision (26, 27). These phages have also madeit possible to extend the physical analysis toDNA upstream of ilvG into the rrnC operon anddownstream of ilvC beyond rho.

Examination of ilvA and ilvD gene expressionin K dilv phages containing all or part of the ilvGand ilvE genes has confirmed that the site essen-tial for multivalent repression by the branched-chain amino acids is upstream of ilvG. This sitepresumably includes, at least, ilvG proximalpromoter and leader sequence with an attenua-

TABLE 5. Derepression of the ilv genes carried by several k dilv prophagesSp act

Strain and relevant genotype Phage and ilv DNA Medium one acid Transaminasecarried ~~deaminase acdBdehydrasea

CU504 None Repressing 46 9 35bleu455 ilv+ Limiting isoleucine 247 32 77

Limiting leucine 97 32 88CU505 A dilv58, GEDAYC Repressing 101 12 44A(ilvGEDA YC)2049 leu455 Limiting isoleucine 429 35 73

Limiting leucine 174 38 92CU505 A dilv37, 'GEDA YC Repressing 105 8 44

.(ilvGEA YC)2049 leu455 Limiting isoleucine 155 12 49Limiting leucine 116 14 58

CU505 A dilv26, 'GEDA YC Repressing 29 10 31A(ilvGEA YQ2049 leu455 Limiting isoleucine 51 15 40

Limiting leucine 39 14 41CU519 A dilv73, 'EDA YC Repressing 32 5 43

A(iIvDAC)115 leu455 Limiting isoleucine 52 6 73Limiting leucine 25 8 89

a oa,P-Dihydroxy-,B-methylvalerate was the substrate.b Values of enzyme activities derived from a chromosomal gene are underlined.

VOL. 149, 1982

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

1080 GRAY ET AL.

tion site that has been identified by nucleotidesequence analysis and by in vitro transcriptionexperiments (17, 22). The control at this site isdemonstrated by (i) repression by growth ofprototrophic lysogens in the presence of thethree branched-chain amino acids (Table 4) or(ii) derepression by growth of valine-susceptibleleucine-auxotrophic lysogens in the presence oflimiting isoleucine or limiting leucine (Table 5).Two additional regulatory features of i/v gene

expression have been exhibited by these phagesthat cannot yet be fully explained. One is theelevated activity of threonine deaminase that isobserved in cells grown on limiting isoleucineeven in the absence of the cis-acting ilv-specificattenuation control that occurs upstream of i/vG.The increased activity did not occur when thesame lysogen as grown on limiting leucine andthus was not simply a consequence of the i/v-specific multivalent regulation. In the lysogencontaining A dilv58, which contains the i/v-spe-cific regulatory region, the unique effect of limit-ing isoleucine on threonine deaminase activity issuperimposed on the derepression that occurswhen either leucine or isoleucine is limiting andwhich is exhibited by all three of the enzymesexamined. This isoleucine-specific effect hasbeen reported before (29) and could be dueeither to a specific site of attenuation precedingthe ilvA gene or to paucity of isoleucine residuesin threonine deaminase relative to the occur-rence of isoleucine in E. coli bulk protein. Thesepossibilities are currently being tested by exam-ining regulation from specifically constructedplasmids with lacZ substituted for i/v structuralgenes (Gray and Calhoun, unpublished data).The second unexplained regulatory feature is

also independent of the regulatory region up-stream of il/v,G. This effect was observed whenthe prototrophic strains were grown in the pres-ence of excess leucine (Table 4). The effect maybe related to the effect of limiting isoleucinedescribed above and is similar to the down-stream amplification described by Smith et al.(29) in that the increase of threonine deaminaseactivity is greater than that of the dehydrase.One possible explanation of the effect of thesehigh levels of leucine is that the isoleucine poolhas been reduced because of the previouslyreported inhibition of threonine deaminase byleucine (6, 32). If so, the apparent derepressionwould thus be the result of an indirect limitationof isoleucine. Compatible with this possibility isthe fact that the high concentration of leucinereduced the growth rate of all the prototrophicstrains examined, and the growth inhibition wascompletely reversed by isoleucine (2, 6, 25;Calhoun, unpublished data). Furthermore, thegrowth of the prototrophic lysogen carrying Xdil/v73, which grew only slowly in minimal medi-

um and exhibited a low level of ilvD and ilvAexpression, was completely inhibited by thatamount of leucine. However, until the phenome-non is examined further, its basis must remainunexplained.

Finally, the combination of genetic and physi-cal characterization reported here for X dilvphages provides the essential foundation for theinterpretation of our nucleic acid sequence de-terminations (16, 16a), indicating specific basechanges associated with i/vG-activating muta-tions (previously il/O). Without such a specificand unambiguous foundation, we could not haverealistically attempted the DNA sequence deter-mination of ilvG-activating mutations, and theirinterpretation would be highly speculative. Weview DNA sequencing as an adjunct to conven-tional genetic and biochemical analyses. DNAsequence immeasurably sharpens our interpreta-tion of phenotype. At present, however, DNAsequence information alone does not adequatelypredict, and obviously cannot confirm, pheno-type.

ACKNOWLEDGMENTS

This work was supported by Public Health Service grantsGM12522 and GM23182 from the National Institutes ofHealth. D.C.B. is the recipient of a Public Health Servicetraineeship supported by a cell and molecular biology traininggrant (GM07211), awarded to Purdue University from theNational Institutes of Health. D.H.C. is the recipient of anIrma T. Hirschl career scientist award.

ADDENDUM IN PROOF

After this manuscript was submitted for publication,the paper by M. Uzan, R. Favre, E. Gallay, and L.Caro (Mol. Gen. Genet. 182:462-470, 1982, '1981")was published. Their structural analyses of the DNApresent in these phage are in substantial agreementwith ours except for phage X dil/73. Their preparationof this phage, but not ours, has undergone a generearrangement relative to the other X dil/ phages.The intercistronic region between il/vG and il/cE

present on a KpnI to Hindlll segment contains anactive and regulatable promoter as judged by its activi-ty in plasmid pMC81 (Gray and Calhoun, Mid-AtlanticExtrachromosomal Elements Conference, 1981).

LITERATURE CITED

1. Bachmann, B. J., and K. B. Low. 1980. Linkage map ofE.sclherichlia coli K-12. edition 6. Microbiol. Rev. 44:1-56.

2. Baez, M., D. W. Patin, and D. H. Calhoun. 1979. Deletionmapping of the i1vGOEDAC genes of Escherichai coli K-12. Mol. Gen. Genet. 169:289-297.

3. Berg, C. M., and K. J. Shaw. 1981. Organization andreg,ulation of the ilvGEDA operon in Saltnonella typhitnmur-ilum LTI. J. Bacteriol. 145:984-989.

4. Berg, C. M., K. J. Shaw, J. Vender, and NI. Borucka-Mankiewicz. 1979. Physiological characterization of polarTn-5 indUced isoleucine-valine auxotrophs in Esc/herichzia(o/i K-t": cvidence tor an internal promoter in theilif)fGEDA4 operon. Genetics 93:309-319.

5. Blazey, D. I., and R. 0. Burns. 1979. Genetic organization

J BACTFERIOIL.

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

STRUCTURE OF A dilv PHAGES 1081

of the Salmonella typhimurium ilv gene cluster. Mol. Gen.Genet. 177:1-11.

6. Calhoun, D. H. 1976. Threonine deaminase from Esche-richia coli: feedback hypersensitive enzyme from a genet-ic regulatory mutant. J. Bacteriol. 26:56-63.

7. Childs, G. J., H. Ohtsubo, E. Ohtsubo, F. Sonnenberg, andM. Freundlich. 1977. Restriction endonuclease mapping ofthe Escherichia coli K-12 chromosome in the vicinity ofthe ilv genes. J. Mol. Biol. 117:175-193.

8. Das, A., D. Court, and S. Adhya. 1976. Isolation andcharacterization of conditional lethal mutants of Esche-richia coli defective in transcription termination factorrho. Proc. Natl. Acad. Sci. U.S.A. 73:1959-1963.

9. Davis, B. D., and E. S. Mingioli. 1950. Mutants of Esche-richia coli requiring methionine or vitamin B12. J. Bacte-riol. 60:17-28.

10. Duggan, D. E., and J. A. Wechsler. 1973. An assay fortransaminase B enzyme in Escherichia coli K-12. Anal.Biochem. 51:67-79.

11. Gallant, J. A. 1979. Stringent control in Escherichia coli.Annu. Rev. Genet. 13:393-415.

12. Gayda, D. J., T. D. Leathers, J. D. Noti, F. J. Smith, J. M.Smith, C. S. Subrahmanyam, and H. E. Umbarger. 1980.Location of the multivalent control site for the ilvEDAoperon of Escherichia coli. J. Bacteriol. 142:556-567.

13. Guardiola, J., and M. laccarino. 1971. Escherichia coli K-12 mutants altered in the transport of branched-chainamino acids. J. Bacteriol. 108:1034-1044.

14. Hahn, J. E., and D. H. Calhoun. 1978. Suppressors of agenetic regulatory mutation affecting isoleucine-valinebiosynthesis in Escherichia coli K-12. J. Bacteriol.136:117-124.

15. laccarino, M., J. Guardiola, M. DeFelice, and R. Favre.1978. Regulation of isoleucine and valine biosynthesis.Curr. Top. Cell Regul. 14:29-73.

16. Lawther, R. P., D. H. Calhoun, C. W. Adams, C. A.Hauser, J. Gray, and G. W. Hatfield. 1981. Molecularbasis of valine resistance in Escherichia coli K-12. Proc.Natl. Acad. Sci. U.S.A. 78:922-925.

16a.Lawther, R. P., D. H. Calhoun, J. Gray, C. W. Adams,C. A. Hauser, and G. W. Hatfield. 1982. DNA sequencefine-structure analysis of ilvG (IlvG+), mutations of Esch-erichia coli K-12. J. Bacteriol. 149:294-298.

17. Lawther, R. P., and G. W. Hatfield. 1980. Multivalenttranslational control of transcription termination at atten-uator of ilvGE operon of Escherichia coli K-12. Proc.Natl. Acad. Sci. U.S.A. 77:1862-1866.

18. Lawther, R. P., B. Nichols, G. Zurawski, and G. W.Hatfield. 1979. The nucleotide sequence preceding andincluding the beginning of the ilvE gene of the ilvGEDAoperon of Escherichia coli K-12. Nucleic Acids Res.7:2289-2301.

19. Leathers, T. D., J. Noti, and H. E. Umbarger. 1979.Physical characterization of ilv-lac fusions. J. Bacteriol.

140:251-260.20. McCorkle, G. M., T. D. Leathers, and H. E. Umbarger.

1978. Physical organization of the ilvEDAC genes ofEscherichia coli strain K-12. Proc. Natl. Acad. Sci.U.S.A. 75:89-93.

21. Morgan, E. A., and M. Nomura. 1979. Deletion analysis ofthe expression of RNA genes and associated tRNA genescarried by a transducing bacteriophage. J. Bacteriol.137:507-516.

22. Nargang, F. E., C. S. Subrahmanyam, and H. E. Um-barger. 1980. Nucleotide sequence of ilvGEDA attenuatorregion of Escherichia coli. Proc. Natl. Acad. Sci. U.S.A.77:1823-1827.

23. Patin, P. W., and D. H. Calhoun. 1979. Mapping of ilvOloci of Escherichia coli K-12 with bacteriophage A dilv. J.Bacteriol. 137:1234-1242.

24. Pledger, W. J., and H. E. Umbarger. 1973. Isoleucine andvaline metabolism in Escherichia coli. XXI. Mutationsaffecting derepression and valine resistance. J. Bacteriol.114:183-194.

25. Rogerson, A. C., and M. Freundlich. 1970. Control ofisoleucine, valine, and leucine biosynthesis. VIII. Mecha-nism of growth inhibition by leucine in relaxed andstringent strains of Escherichia coli K-12. Biochim.Biophys. Acta 208:87-98.

26. Shimada, K., R. A. Weisburg, and M. E. Gottesman. 1973.Prophage lambda at unusual chromosomal locations. I.Location of the secondary attachment sites and the prop-erties of the lysogens. J. Mol. Biol. 63:483-503.

27. Shimada, K., R. A. Weisberg, and M. E. Gottesman. 1973.Prophage lambda at unusual chromosomal locations. II.Mutations induced by bacteriophage lambda in Escherich-ia coli K-12. J. Mol. Biol. 80:297-314.

28. Smith, J. M., F. J. Smith, and H. E. Umbarger. 1979.Mutations affecting the formation of acetohydroxy acidsynthase II in Escherichia coli K-12. Mol. Gen. Genet.169:299-314.

29. Smith, J. M., D. E. Smolin, and H. E. Umbarger. 1976.Polarity and the regulation of the ilv gene cluster inEscherichia coli K-12. Mol. Gen. Genet. 148:111-124.

30. Subrahmanyam, C. S., G. M. McCorkle, and H. E.Umbarger. 1980. Physical location of the ilvO determinantin Escherichia coli K-12 deoxyribonucleic acid. J. Bacte-riol. 142:547-555.

31. Umbarger, H. E. 1978. Amino acid biosynthesis and itsregulation. Annu. Rev. Biochem. 47:533-606.

32. Vonder Haar, R. W., and H. E. Umbarger. 1974. Isoleu-cine and valine metabolism in Escherichia coli K-12:detection and measurement of ilv-specific messenger ribo-nucleic acid. J. Bacteriol. 120:687-696.

33. Watson, M. D., J. Wild, and H. E. Umbarger. 1979.Positive control of ilvC expression in Escherichia coli K-12; identification and mapping of regulatory gene i1vY. J.Bacteriol. 139:1014-1020.

VOL. 149, 1982

on July 2, 2020 by guesthttp://jb.asm

.org/D

ownloaded from