Gepze, 42 (1986) 283-292

Elsevier

283

GENE 1568

Cloning and characterization of several dominant-negative and tight-binding mutants of Znc repressor

(Recombin~t DNA; protein-DNA interactions)

Joan L. Betz

Department of Biochemistry, Biophysics and Genetics, University of Colorado Medical School, Denver, CO 80262 (U.S.A.) Tel. (303) 394-8261

(Received September lSth, 1985)

(Revision received January 14th, 1986)

(Accepted January 17th, 1986)

SUMMARY

Using both general recombination and molecular cloning techniques, 13 I-, Iwd and itb missense mutations in the ZacI gene were transferred from F’ la0 episomes to ColEl derivative plasmids. Two deletion derivatives of the lacl genes encoding the wild-type (wt) and the tight-binding (Itb) B3 and BS repressors were also constructed. The mutant repressors were examined for polypeptide size and stability, and for binding to the inducer isopropyl-~-D-thiog~actoside (IPTG). Several of the I-d repressors were shown to be partially degraded in vivo, in conflation of earlier results based on [ 14C]IPTG binding [Miwa and Sadler, J. Mol. Biol. 117 (1977) 843-8681. The sizes of polypeptides produced by lacl deletion derivatives were consistent with expectations based on the extent of deletion and the location of termination sites within the plasmid sequence. The first 400 bp of several mutant lacl genes were sequenced. Our wt lacl gene differs from another wt lacl sequence (Farabaugh, 1978), containing a single bp change that results in an Ala to Thr substitution at amino acid (aa) 109. We identified bp substitutions and the resultant aa changes for two Itb and two Isd genes; the positions correlated with prior genetic mapping data. Three of these new changes were in the N-terminal domain (headpiece) of repressor, with one change in the core domain at aa 99.

Abbreviations: aa, amino acid(s); Ap, ampicillin; bp. base INTRODUCTION

pair(s); CAMA, Casamino acids; d, deletion; f, la&; la@, up- promoter mutation for la&; lad phenotypes: I -, defective

repressor; I -d, dominant I- ; I’, reverse repressor in which

repression increases in presence of inducer; Itb, tight-binding

repressor; IPTG, isopropyl-/?-D-thiogalactoside; kb, 1000 bp;

lacZ’, N-terminal region of lacZ gene; MNNG, N-methyl-

N’-nitro-N-nitrosoguanidine; PA, polyacrylamide; R, resistance;

SDS, sodium dodecyl sulfate; Tc, tetracycline; wt, wild

type; XGal, 5-bromo-4-ch~oro-3-indoIy~-~-~-galactoside; Z/B

value(s), see Table I; [ ] designates plasmid-carrier state.

An important focus of molecular biology con- tinues to be deciphering protein structure - function relationships. The lac repressor protein remains a prototypic example of a site-specific DNA regulatory protein. Considerable evidence, both genetic and physical, has now accumulated for the presence of operator sequence-specific contact regions or res- idues in the N-terminal 60 aa of repressor. Mutants

0378-I 119/X6/%03.50 0 1986 Elsevw Science Publishers B.V. (Biomedical Division)

284

of both the protein and of its recognition site (the

operator) have been invaluable in elucidating

functional aspects of the luc repressor protein and in

defining its functional domains (reviewed in Miller,

1978; 1984, and Beyreuther, 1978).

Only three aa substitutions are currently known

which result in Itb repressors: X86 phenotype result-

ing from a Ser to Leu change at aa 61 (Files and

Weber, 1978); 112 a Pro to Tyr change at aa 3

(Schmitz et al., 1978) and the suppression of the A10

nonsense mutation affecting aa 61 by insertion of

Tyr (Miller et al., 1979). Six well-characterized Itb

repressors have been genetically mapped but not

sequenced (Pfahl, 198 1; Pfahl and Hendricks, 1984).

We had previously isolated ten Itb repressor mutants

(Betz and Sadler, 1976) which had increased non-

specific affinity for DNA with little or no increase in

specificity for operator. The Itb mutant B5 mapped

in the region of the gene encoding the first 57 aa and

the others in approx. aa 58-120, but aa substitutions

were not determined.

34 substitutions have been identified at 25 po-

sitions in the first 62 aa which result in the transdom-

inant I - (I Id) phenotype (Muller-Hill et al., 1975;

Weber et al., 1975; Miller, 1984; Mott et al., 1984).

In one study, 11 of 13 mutants tested which had

decreased nonspecific binding (to phosphocellulose

or DNA cellulose) were shown to result in labile

repressors, in which the N-terminus was degraded in

a mutant-specific manner (Schlotmann and

Beyreuther, 1979). This laboratory had isolated a

large number of Idd repressor mutants (Miwa and

Sadler, 1977) that were extensively characterized

and genetically mapped, but their aa alterations were

never determined, in part due to the instability of

some of the resultant proteins. These I d mutants

were characterized into four general and 39 specific

phenotypic groups based on IPTG binding, operator

binding, and degree of transdominance. The majori-

ty of these lucIpd mutations mapped in the first

57 aa, with some of them resulting in repressors sub-

ject to degradation in vivo.

Rapid DNA sequencing techniques, which have

made determination of changes in DNA sequence

easier than assessing protein primary structure, led

us to reexamine our missense I Id and Itb repressors.

Reported here is the cloning of mutant lad alleles

from F’luc episomes to plasmid pHIQ6 by means of

genetic recombination and their further molecular

subcloning onto small lad plasmids. Sequencing of

four of the mutants uncovered several new changes

in the N-terminal region. The mutant repressors have

also been characterized with regard to their affinities

for IPTG, and with respect to polypeptide size and

stability, both in fan strains (Gottesman and Zipser,

1978) and using the maxicell technique (Sancar

et al., 1979).

MATERIALS AND METHODS

(a) Hosts and plasmids

All strains are derivatives of Escherichia coli K- 12.

HBlOl (la&’ O’Z’ Y- recA; Bolivar and Back-

man, 1979), DH9, a recessive lad derivative (Hare

and Sadler, 1978) and D1210, a IucZq derivative

(Sadler et al., 1980), were recipients for transfor-

mation. Strain P91 (F- d(lac-proB) rpsL Thi- ) is a

derivative of J. Miller’s strain P90 which had been

cured of its prophage. BNN103 is the Alon strain of

Young and Davis (1983) obtained from A. Siddiqui.

Plasmid pHIQ6 is pMB9 carrying a lo-kb EcoRI

lacZq fragment of 1h80dlucZq (Hare and Sadler,

1978). Repressor alleles were obtained from

F’ ZacZ‘iproA + ,B + episomes for Itb alleles (Betz and

Sadler, 1976), I- and Ied alleles (Miwa and Sadler,

1977) and from an F’ZucZq episome for the X86 allele

(Chamness and Willson, 1970). Plasmid pJE13 is a

derivative of pBR322 with the EcoRI site filled in

and an 8-bp synthetic EcoRI linker inserted in AvaI-

cut, filled plasmid (J.L.B. and H.M.S., in prepa-

ration).

(b) Procedures

Methods for preparation of plasmid DNA, gel

electrophoresis, digestion with restriction endonu-

cleases, ligation, transformation of E. coli and XGal

plates have been described (Sadler et al., 1978;

1980). For primary transformations using DNA

from F’luc[pHIQ6] strains, plasmid DNA was pre-

pared as described by Holmes and Quigley (198 l),

with reprecipitation from 0.3 M sodium acetate fol-

lowing the isopropanol step. Direct DNA sequenc-

ing followed the protocols of Maxam and Gilbert

(1980) and the T > G reaction of Simoncsits and

285

Torok (1982), using restriction fragments end-label- ed with [ y-3”P]ATP (ICN or NEN).

RESULTS AND DISCUSSION

(a) Cloning of Iacl genes containing missense mu- tations

General genetic recombination was used to trans- fer the Ztb and IVd alleles from F’luc episomes to ColEl-derived plasmids. F’taclq proA + ,B + epi- somes carrying mutant alleles were introduced into a d(Zuc-prd) Ret + strain carrying plasmid pHIQ6 (wt repressor). After cultures were grown for a num- ber of generations, plasmid DNA was prepared and used to transform DH9. Colonies containing rare recombin~t plasmids with Imd or Itb alleles were distin~ished from those containing plasmid pHIQ6 by XGal media.

For the recessive I - and transdominant I - d rec- ombinants, roughly 0.1-0.2 y0 of transformants were dark blue (Lac-constitutive) and hence possible pHIQ-I _ recombinants. Plasmid DNAs prepared from transformants were used to retransform DH9 to ensure that only mutant lad plasmids were exam- ined. Several blue colonies from each of these secon- dary transformations were purified and the trans- dominant laci phenotypes verified by P-galactosid- ase assays of transformants of DH9 (la&- 0 + 2’ ), HBlOl (Zacl”O’Z’) and Dl210 (la0 O’+Z+) (Table I and not shown).

The X86 repressor (Chamness and Willson, 1970) and our tight-binding repressors (Betz and Sadler, 1976) are much less inducible by IPTG than is wt repressor. 3 x 10e5 M IPTG gave partial induction of DH9[pHIQ6] (blue colonies on XGal-Tc agar) while the recombinant DH9[pHIQ-Itb] clones remained white (uninduced). White colonies were observed at frequencies of about 0.1 y0 ; these white colonies were used to prepare DNA and retransform DH9. White retransformants on XGal-IPTG-Tc plates (the DH9[pHIQ-Itb] strains) were individual- ly purified and assayed for j?-galactosidase specific activities at a variety of IPTG concentrations to verify the mutant phenotypes (Table I). In two cases, for mutants B12 and B29, a class of transformants was found which appeared to carry only the fb B3

mutation; a recombinant with the B12 phenotype was not obtained. The Itb repressor mutants B5, B 12, B29 and B32 had been derived from mut~t B3 by further mutagenesis (Betz and Sadler, 1976).

(b) Analysis of recombinant plasmids

Plasmid DNAs prepared from DH9[pHIQ-Itb] or DH~[PHIQ-I-~] recombinant strains were the same size as the wt pHIQ6 plasmid. The DNAs were cleaved by EcoRI to give two fragments of similar size to those obtained from pHIQ6 (Hare and Sad- ler, 1978). Retransformants of DH9 carrying the mut~t repressor plasmids were assayed for @-galac- tosidase-specific activity (Table I). The recessive I - repressor E29 was fully constitutive, in that IPTG had no effect, although this repressor did not permit maximal induction of DH9 to the same specific ac- tivity shown by DH9[pMB9]. The basal and induc- ed Z/B values of the I - d and Ifb strains correlated well with those of the parental F’lac strains. Some of the DH9[pHIQ-Itb] cultures grew poorly in the absence of IPTG, notably X86, and B29 and B32 which also have the I’ phenotype of increased repression in the presence of IPTG. This poor growth probably reflects interference with expression of cellular genes by repressor binding at non- operator sites.

Isolation of purified fragments for direct DNA sequencing from these large pHIQ plasmids was inconvenient; therefore smaller fragments contain- ing the lad gene were subcloned into pJEl3, a deri- vative of pBR322. For some I d and Ifb mutants, a 1724-bp H&c11 partial digestion fragment encoding Zacl,P,O and part of ZacZ was ligated to EcoRI lin- kers and into the EcoRI site of pJE13 (the pIQ series plasmids; Fig. 1). For other mutants, a 635bp HincII-ApaI fragment of lad (containing the Zacl promoter and sequences encoding aa I-188) was ligated into a 4.5kb fragment of plasmid pIQ lacking this portion of Iucf (pIQ . series plasmids; Fig. 1). These small repressor plasmids were more stable, were excellent sources for repressor production, and were used for purification of fragments for DNA sequencing. In addition to the wt genes (both 19 and I + ), ten Zacl genes have been subcloned (Table I).

When extracts of strains carrying iucImd plasmids were tested for [ “C]IPTG binding at 8 x IO-’ M, the results agreed with previous findings for these

Subc

lone

d

in p

JE13

’

bp

subs

titut

ion8

aa

su

bstit

utio

n

TA

BL

E

I

Clo

ned

lacl

al

lele

s

Rec

ombi

nant

plas

mid

”

Phen

otyp

ei’

Mut

atio

n Z

/B

valu

e’

Bas

al

Indu

ced

Rep

ress

or

degr

aded

“

IPT

G

bind

ing”

pHIQ

6 w

t 0.

0000

5 0.

092

no

yes

PIQ

pH

IQE

ll6

pHIQ

E18

2

pHIQ

E14

1

pHIQ

E55

pHIQ

El5

1

pHIQ

E29

IId

IId

IId

I-d

IId

I-

El1

6

El8

2

El4

1

E55

El5

1

E29

0.08

0 0.

381

no

0.03

2 0.

460

yes

0.25

7 0.

426

no

0.04

9 0.

479

yes

0.10

7 0.

218

yes

0.10

0 0.

107

yes

yes

yes

yes

yes

yes

no

pIQ

Ell6

pIQ

E18

2

pIQ

E

l41

pIQ

. E

55

pIQ

.El5

1

353

G/d

A

/T

95

G/C

+

A/T

174

C/G

+

T/A

109

Ala

+

Thr

23

Val

--t

Met

49

Pro

+ L

euh

pHIQ

B3

I’b

B3

0.00

00

1 0.

017

no

yes

pIQ

B3

pHIQ

B5

1’s

B5

0.00

00

1 0.

0003

0 no

ye

s pI

QB

5

pHIQ

B21

pHIQ

B29

pHIQ

B32

pHIQ

B40

pHIQ

X86

Pb

I’b

1’s

I’b

Pb

B21

B29

B32

B40

X86

0.00

006

0.00

042

0.00

27

0.00

13

0.00

020

0.00

19

0.00

001

0.00

002

nd

0.00

073

no

no

no

no

no

yes

yes

yes

yes

yes

pIQ

.B21

pIQ

. B

29

pIQ

B

32

323

G/C

+

A/T

183

T/A

+

C/G

nd 39

T/A

+

C/G

210

G/C

-T/A

99

Val

+

Ile

52 V

al -

-t A

la’

nd’ 4

Val

+ A

la’

61

Ser+

Leu

pMB

9 -

1.17

1.

29

- -

no

* M

utat

ions

w

ere

reco

mbi

ned

from

F’

lacl

“ ep

isom

es

onto

pl

asm

id

pHIQ

6,

a IO

-kb

Eco

RI-

ende

d Iu

cZ f

ragm

ent

in p

MB

9 (H

are

and

Sadl

er,

1978

).

Ref

eren

ces

to m

utan

ts

are:

II

d se

ries

(Miw

a an

d Sa

dler

, 19

77);

It

b se

ries

(B

etz

and

Sadl

er,

1976

);

X86

(C

ham

ness

an

d W

illso

n,

1972

).

b Ia

cI p

heno

type

s ar

e I

~, r

eces

sive

de

fect

ive

repr

esso

r;

I _

d, t

rans

dom

inan

t de

fect

ive

repr

esso

r;

Pb,

tight

-bin

ding

re

pres

sor;

I’

, re

vers

e re

pres

sor

(rep

ress

ion

is i

ncre

ased

in

the

pr

esen

ce

of I

PTG

).

c Z

/B

valu

e is

the

sp

ecif

ic

activ

ity

of b

-gal

acto

sida

se

in s

trai

n D

H9,

a

Iact

- de

riva

tive

of H

BlO

l (H

are

and

Sadl

er,

1978

).

Cul

ture

s fo

r B

-gal

acto

sida

se

assa

ys

wer

e gr

own

in t

rypt

one

brot

h as

det

aile

d by

B

etz

and

Sadl

er

(197

6)

with

th

e ad

ditio

n of

10

ng

Tc/

ml;

indu

ced

cultu

res

wer

e gr

own

with

1

mM

IP

TG

. A

ssay

s w

ere

perf

orm

ed

and

spec

ific

ac

tiviti

es

calc

ulat

ed

as

desc

ribe

d (B

etz

and

Sadl

er,

1976

);

the

B29

an

d X

86

Pb s

trai

ns

whi

ch

also

ha

ve

an

I’ p

heno

type

gr

ew

poor

ly

in

abse

nce

of

IPT

G;

nd,

not

dete

rmin

ed.

d D

egra

datio

n w

as

asse

ssed

by

an

alys

is

of p

lasm

id-e

ncod

ed

prot

eins

on

SD

S-PA

ge

ls,

as

show

n in

Fi

g.

2. F

or

DH

9[pH

IQE

29],

only

a

trac

e of

rep

ress

or

was

de

tect

ed,

c [ 1

4C]I

PTG

bi

ndin

g w

as

asse

ssed

by

am

mon

ium

su

lfat

e co

prec

ipita

tion

usin

g 8

x IO

-’ M

[i4

C]I

PTG

an

d ce

ll-fr

ee

extr

acts

(B

etz

and

Sadl

er,

1976

). T

ypic

ally

, lo

-ml

cultu

res

wer

e gr

own

in d

oubl

e-st

reng

th

Pena

ssay

br

oth

(Dif

co)

cont

aini

ng

10 n

g T

c/m

l an

d re

susp

ende

d in

1

ml

for

lysi

s;

trip

licat

e 20

0~~1

aliq

uots

w

ere

used

fo

r ea

ch

assa

y.

r T

he

subc

loni

ng

of l

ucl

gene

s in

to

pJE

13,

a de

riva

tive

of p

BR

322,

is

des

crib

ed

in t

he

lege

nd

to

Fig.

1.

g N

ucle

otid

es

are

num

bere

d fr

om

star

t of

luc

k m

RN

A,

usin

g as

w

t se

quen

ce

that

of

Fa

raba

ugh

(197

8).

nd,

not

dete

rmin

ed.

h A

lso

cont

ains

th

e pI

Q

subs

titut

ion.

’ T

he

B5,

82

9 an

d B

32

lacl

ge

nes

also

co

ntai

n th

e B

3 m

utat

ion,

T

he

data

fo

r B

32

wer

e ad

ded

in t

he

proo

fs

and

are

not

refe

rred

to

in

the

te

xt.

281

Fig. 1. Construction of lucl plasmids. The pIQ series plasmids were constructed by ligation of a 1.7-kb HincII fragment (partial digest)

of the respective pHIQ plasmid to EcoRI linker segments, and then to EcaRI-cleaved plasmid pJE13. Plasmids of the pIQ series were

constructed by insertion of a 635-bp HincII-Apa1 fragment containing the lucl promoter and first 585 bp oflacl into WI-cut, Sl-treated

and ApaI-cleaved pfasmid pIQ, as shown for pIQ B32. The PvuII-deletion derivatives of the pIQ . series plasmids were constructed

by partial digestion with Pvufl and religation. The circled letters a-d represent the fragments produced by PvuII-digestion (Table II,

footnote a). The thin line represents the vector DNA; the thick line indicates the lac sequences, and the dotted segment represents the

lac sequences of the B32 mutant. Abbreviations: Av, Aval; A, ApaI; E, EcoRI; H, HincII; P, PvuII; S, SalI.

repressors relative to wt repressor (Miwa and Sad- ler, 1977). The El16 and El41 repressors bound IPTG to the same degree as wt repressor, whereas the E 182 repressor bound roughly 50 y0 as much and the E29 repressor less than 1% (not shown). The I - d repressors listed in Table I had previously been shown to have Kdiss values for IPTG very close to that of wt repressor (Miwa and Sadler, 1977), and thus the decreased binding by the El82 extract re- flected a decreased concentration of active repressor subunits.

The Itb B3 and B5 repressors also have affinities for IPTG similar to that of wt repressor (Betz and Sadler, 1976) and were expressed from the ZacIq promoter. Extracts of strains carrying the B3 and B5 mutations on F’luc episomes or pHIQ plasmids bound 55-80x as much [ ‘*C]IPTG as strains car-

rying analogous wt lad genes. Yet extracts of strain DH9 containing the pIQB3 or pIQB5 plasmids bound only about 25% as much IPTG as did ex- tracts of DH9[ pIQ], We have no explanation for this reduced expression of tight-binding repressor sub- units compared to wt from the pIQ plasmids.

(c) Nucleotide substitutions in Cud genes

The first 400 bp of the Zuci genes encoding our wt., Itb B3 and B5 repressors and two IId repressors have been sequenced, with substitutions listed in Table I. An apparently neutral change of Ala to Thr at aa 109 was located in our wt strain, as compared to another wt strain (Farabaugh, 1978). This change, which removes a Fnu4HI restriction site, was also present in the El82 ZucZ gene, but not in the other

288

mutant lad genes. The mutants were isolated on

F’lacPproA + ,B + episomes, whereas the la0 gene

was cloned from a kdlac strain of independent origin,

and so a neutral change of this type was not unex-

pected given the numerous ‘wt’ Zac strains in exis-

tence. One study of pseudo-wt repressors noted a

variation in operator allinity of 2-4 fold among the

three repressors (Jobe et al., 1972); no sequence ana-

lyses have been reported for these. During the genetic

recombination from F’lac to plasmid pHIQ6, some

of the recombinants would retain their own base

sequence for position 109, and others would obtain

the pHIQ6 allele.

For the Itb B3 repressor, the only change found

was a G/C to A/T transition at position 323, which

would result in a Val to Ile substitution at aa 99. The

B5 repressor, derived from B3 by further mutagenes-

is, had, in addition to the mutation at position 323,

a T/A to C/G transition at position 183. This substi-

tution changes aa 52 from Val to Ala in the so-called

hinge region of the protein (Beyreuther, 1978) and is

the first substitution at this site. This location was

consistent with previous deletion mapping which

placed the B5 mutation within the first 57 aa of the

protein (Betz and Sadler, 1976). No mutations have

been identified for the B29 and B32 repressors within

the portions of the first 400 bp that have been se-

quenced; the B3 mutation was identified in these

genes, as expected from their derivations.

Two Zpd mutations have been localized thus far.

The El 16 mutation, a strongly transdominant muta-

tion and the most N-terminal of that collection, re-

sulted from a G/C to A/T change at position 95; this

corresponds to a Val to Met substitution at aa 23 in

the N-terminus of repressor. Mutation E182, a

strongly transdominant mutation resulting in a labile

protein, resulted in a Pro to Leu substitution at aa 49

in the repressor. The El 16 mutation adds a second

substitution at aa 23, and the El82 mutation dupli-

cates a substitution at aa 49 (Miller, 1984), although

the exact bp change may be different. The bp changes

present in our mutants were all G/C to A/T or C/G

to T/A, as expected for MNNG mutagenesis (Miller,

1978).

(d) Deletion derivatives of ZucZ genes

From PvuII partial digestions of plasmids pIQ,

pIQB3 and pIQB5, followed by religation, we

obtained plasmids which had lost one or both PvuII

fragments from the 3’- end of the gene, in addition

to a fragment of pBR322 (Fig. 1; Table II). Repres-

sors specified by the plasmids with the smaller dele-

tion should have lost 12 aa from their C-termini; they

repressed from 10 y0 to 50 y0 as well as the respective

native repressors. The polypeptides specified by the

larger deletions should have lost 43 C-terminal aa;

they gave no repression. Cell-free extracts containing

the smaller deletion derivative repressors (pIQ- 1

series) bound 20-85% of the [ 14C]IPTG as did

extracts containing the corresponding intact repres-

sors (Table II). As anticipated, the Kdiss values for

IPTG did not differ significantly between the wt

repressors and the smaller deletion derivatives :

values ranged from 1.5-3 x 10.’ M for the six differ-

ent extracts tested. The polypeptides specified by the

larger deletion derivatives bound no IPTG (Tab-

le II), although as discussed below (section e), trun-

cated polypeptides were detected on polyacrylamide

gels. Their lack of IPTG binding probably reflects

disruption of the IPTG-binding site (Miller, 1978) or

partial degradation of the polypeptide.

(e) Stability and size of repressor polypeptides

Even though repressor is synthesized from the

laclq up-promoter and the IacIgene is on a multicopy

plasmid, no protein corresponding to repressor

could be distinguished on a Coomassie blue-stained

gel. With the maxicell technique for radiolabeling

plasmid-encoded proteins (Sancar et al., 1979) the

repressor proteins were easily visualized. The repres-

sors synthesized by plasmids carrying wt, lacZtb and

certain lacZed genes were stable and of wt size

(Fig. 2). In contrast, repressors specified by El82

and certain other IacI- d genes were present as a

mixture of wt size and slightly smaller fragments

(Table I and Fig. 2). Based on the mobilities of the

protein standards, the El82 repressor was estimated

to have lost fewer than 15 aa, although its missense

substitution is at aa 49 (see Table I). Only a trace of

wt-sized polypeptide was detected from cultures car-

rying the Zacl- plasmid pHIQE29 (not shown).

The repressor specified by plasmid pIQBS-1 was

expected to lack 6 aa (12 missing from fad with a

gain of 6 from translation of pBR322 sequences;

Table II), and, as expected, migrated with a mo-

bility essentially that of wt repressor (Fig. 2). The

289

TABLE II

Deletion derivatives of IucI plasmids

Plasmid Mutation PVUII

in lad fragments”

Repressor

aa

deleted

IPTG

bindingb

(% of

parental

plasmid)

Z/B value’ Quaternary

structured

Basal Induced

PIQ

PIQ-l

PIQ-2

wild type a,b,c,d

ad

a

- 100 0.00007 0.0964 t

I2 85 0.00014 0.139 d,t

43 0 1.15 0.98 nd

pIQB3 pIQB3-1

pIQB3-2

B3 a,b,c,d

a&d

a

-

I2

43

100 0.00005 0.0202 t

52 0.00024 0.270 d,t

0 1.05 0.95 nd

pIQB5 pIQB5-1

pIQB5-2

B5 a,b,c,d

a,d

a

- 100 0.00002 0.00034 t

12 22 0.00027 0.00236 d,t

43 0 1.02 1.03 nd

il Deletions of lacl were created by PvuII partial digestion and religation, as diagrammed in Fig. 1; an extra six aa should be added from

the translation of pBR322 sequences following the PvuII site at bp 2067.

b IPTG binding was assessed as described in Table I (footnote e). Results are given as a percent binding relative to each original plasmid;

the wt (pIQ) binding averaged 6100 cpm/mg protein in the crude extract. The pIQB3 and pIQB5 cell-free extracts bound only one fourth

the IPTG as did that of pIQ. The Kdlss values for IPTG ranged from 1.5 to 3 x IO-’ M for the original and small deletion derivative

repressors, with no significant differences between the six extracts.

’ Z/B value is the specific activity of fi-galactosidase in DH9; induced cultures were grown in the presence of 1 mM IPTG. See Table I,

footnote c.

d Quaternary structure of repressors was assessed using 5-20% (v/v) gly cerol gradients with repressor activity assayed by [14C]IPTG

binding: t, tetramer; d, dimer; nd, not determined.

67-

25-

Fig. 2. Plasmid-encoded proteins labeled by the maxicell tech-

nique (Sancar et al., 1979). A 3-ml culture ofstrain DH9 carrying

the indicated plasmid was irradiated to 250 J/m’, grown over-

night in the presence of 100 pg cycloserine/ml and then I ml

labeled for 2 h at 37°C with 2 pCi [14C]aa hydrolysate (ICN).

After lysis at 90°C in 50 nl of lysis buffer, 25 pl of sample was

PIQB5-2 repressor migrated as expected for a

truncated polypeptide which should be missing 31 aa

total (43 missing from lucl plus 6 translated from

pBR322). The repressors specified by the pIQ and

pIQB3 deletion derivatives behaved analogously to

the pIQB5 derivatives when analyzed on SDS-PA

gels (not shown).

The quaternary structure of the deletion derivative

repressors was analyzed using centrifugation on gly-

cerol gradients (Fig. 3 and not shown). The Q-l,

loaded per lane on a 12.5?, PA gel (30: I acrylamide: bisacrylam-

ide) containing 0.19, SDS (Laemmli, 1970). The numbers on the

left represent polypeptide sizes of standards (in kDa); mobilities

of unlabeled purified repressor and core repressor are noted by

R and CR, respectively. Autoradiography was for 2 days

following en3Hance (NEN) treatment of the dried gel.

/$lactamase was the prominent 28-kDa band; the 41-kDa TcR

protein ran anomalously fast as a weak 33-kDa band. The

additional bands between core repressor and p-lactamase are of

uncertain origin; their intensities varied in different experiments

and they were found irrespective of the plasmid used.

290

fraction

Fig. 3. Sedimentation of wild-type and truncated repressors.

0.5 ml of cell-free extract was layered onto an 11.5 ml 5-20%

glycerol gradient as described by Sadler and Tecklenburg (1976).

Repressor was detected using [14C]IPTG binding (see Table I,

footnote e). Fractions were 0.5 ml and are numbered from the

top; the arrowhead marks the position of the peak of horse

hemoglobin marker which migrates approximately with a dimer

of repressor (Miller et al., 1970). (A) The pIQB5 (0) and pIQBS-

1 (0) repressors; (B) the pIQ (0) and pIQ-1 (0) repressors.

B3-1 and B5-1 repressors lacking a few C-terminal

aa migrated more slowly than wt repressor, suggest-

ing a predominance of dimers rather than tetramers.

Because the tetramer is needed for operator binding,

a reduced concentration of tetramers could account

for most of the observed diminished repression (Tab-

le II). There was also a decreased concentration of

repressor subunits for the B3 and B5 truncated

repressors (Table II, IPTG binding column), which

contributed to the decreased repression. This finding

agrees with previous results with the Ll repressor

(Miller et al., 1970). The Ll repressor lacks nine

C-terminal aa, was present predominantly as a dimer

and was also subject to degradation.

To further probe the question of lability of I d

repressors, F’lacI dO + Z+proA + ,B + episomes

were transferred into a Alon strain and /&galactosid-

ase specific activities compared with those of the

Lon + strain P91. Deg mutants, which affect the

same locus as UV-sensitive lon mutants, had been

shown to degrade labile mutant P-galactosidases

more slowly than wt strains (Gottesman and Zipser,

1978), and might similarly affect labile repressors.

There was, however, no indication of any increased

stability of repressor in the Alon strain (i.e., decreas-

ed fi-galactosidase specific activity in the Alon strain)

even for those I - d repressors subject to degradation

(not shown). This result suggests that for these labile

repressors, proteases other than lon must be respon-

sible for the degradation.

Some of the I ~” and Pb mutants of lac repressor

may identify actual contact residues, although alter-

native explanations exist. The 51-aa to 59-aa head-

piece fragments of repressor bind strongly to phos-

pho- or DNA-cellulose, presumably via the seven

lysine and arginine aa in this domain, whereas the

tetrameric core does not (Weber and Geisler, 1975 ;

de Haseth et al., 1977). Eleven of 13 I-d repressors

with substitutions between aa 5 and 53 and decreas-

ed nonspecific DNA affinity were degraded in vivo:

the major classes of fragments lacked 20-30 aa or

40-60 aa from the N-terminus (Schlotmann and

Beyreuther, 1979). The strong phenotypic effects of

these mutants undoubtedly arise, in part, from the

loss of the electrostatic component of DNA binding.

For the I _ d mutants reported here, three of the five

tested were partially degraded in vivo, namely those

previously classified as degraded based on the total

amount of [ 14C]IPTG binding material present in

crude extracts (Miwa and Sadler, 1977). Only a few

of the strong I d repressors tested were capable of

nonspecific DNA binding: two substitutions at un-

determined positions (Schlotmann and Beyreuther,

1979), and substitutions at aa 16 and 19 (Weber

et al., 1975), to which can be added the El 16 change

at aa 23. The El 16 repressor has been successfully

purified using phosphocellulose chromatography

(K.S. Matthews, personal communication).

Residues 52-59 have also been postulated to par-

ticipate in the sequence-specific portion of repressor

binding (Beyreuther, 1978; von Wilcken-Bergmann

and Muller-Hill, 1982). The B5 repressor had an

alteration at aa 52. It will be important to separate

291

the B3 and B5 repressor mutations to analyze separ-

ately the effect of the Val 52 substitution on the

specificity and tightness of repressor’s action. Work

is in progress on the cloning of appropriate portions

of the B5 lucl gene, as well as the cloning and se-

quencing of additional mutants of our collection. A

separate communication details an examination of

the operator and non-operator aflinities of the tight-

binding repressors (J.L.B., in preparation).

Although identification of specific repressor con-

tacts has been hampered by difficulties in crystalliz-

ing repressor, either alone or in co-crystal complexes

with the operator, a three dimensional structure of

fuc repressor has recently been determined from nu-

clear magnetic resonance data (Kaptein et al., 1985).

The relative orientation of the cc-helices is similar to

that of other DNA-binding proteins whose crystal

structures are known (Matthews et al., 1982; Weber

et al., 1982). The analysis ofmutant repressors, espe-

cially those with alterations in the DNA binding

domain (the headpiece) will continue to aid in defin-

ing the molecular basis of the specific interaction

between the repressor and operator.

ACKNOWLEDGEMENTS

This manuscript is dedicated to the late John R.

Sadler, in memory of his friendship, guidance and

enthusiasm through the years. I would like to thank

Marianne Tecklenburg, Linde Fall, Chris Grant and

Rachel Ostroff for technical assistance, and thank

Alan Koop, Kathy Matthews and Henri Sasmor for

critically reading the manuscript. This research was

supported by National Institutes of Health grant

GM30709, and by grant BRSG-05357 awarded by

the Biomedical Research Grant Program, Division

of Research Resources, N.I.H.

NOTE ADDED IN PROOF

The B32 mutation has been identified as a

T/A -+ C/G substitution at position 39, resulting in a

Val + Ala change at aa 4.

REFERENCES

Betz, J.L. and Sadler, J.R.: Tight-binding repressors of the lac-

tose operon. J. Mol. Biol. 105 (1976) 293-319.

Beyreuther, K.: Chemical structure and functional organization

of lac repressor from Escherichia coli. In Miller, J.H. and

Reznikoff, W.S. (Eds.), The Operon. Cold Spring Harbor

Laboratory, Cold Spring Harbor, NY, 1978, pp. 123-154.

Bolivar, F. and Backman, K.: Plasmids of Escherichiu coli as

cloning vectors. Methods Enzymol. 68 (1979) 245-267.

Chamness, G.C. and Willson, C.D.: An unusual luc repressor

mutant. J. Mol. Biol. 53 (1970) 561-565.

de Haseth, P.L., Lohman, T.M. and Record Jr., M.T.: Non-

specific interaction ofluc repressor with DNA: an association

reaction driven by counterion release. Biochemistry 16 (1977)

4783-4790.

Farabaugh, P.J.: Sequence of the lacl gene. Nature 274 (1978)

765-769.

Files, J.G. and Weber, K.: Direct identification of the amino acid

changes in two mutant luc repressors. J. Mol. Biol. 123 (1978)

454-456.

Gottesman, S. and Zipser, D.: Deg phenotype oflkherichia coli

Ion mutants. J. Bacterial. 133 (1978) 844-851.

Hare, D.L. and Sadler, J.R.: Cloning ofthe l&gene into a ColEl

plasmid. Gene 3 (1978) 269-278.

Holmes, D.S. and Quigley, M.: A rapid boiling method for the

preparation of bacterial plasmids. Analyt. Biochem. 114

(1981) 193-197.

Jobe, A., Riggs, A.D. and Bourgeois, S.: luc repressor-operator

interaction, V. Characterization of super- and pseudo-wild-

type repressors. J. Mol. Biol. 64 (1972) 181-199.

Kaptein, R., Zuiderweg, E.R.P., Scheek, R.M., Boelens, R. and

Van Gunsteren, W.F.: A protein structure from nuclear mag-

netic resonance data: luc repressor headpiece. J. Mol. Biol.

182 (1985) 179-182.

Laemmli, U.K.: Cleavage of structural proteins during assembly

of the head of bacteriophage T4. Nature 227 (1970) 680-685.

Matthews, B.W., Ohlendorf, D.H., Anderson, W.F. and Takeda,

Y.: Structure of the DNA-binding region of luc repressor

inferred from its homology with cro repressor. Proc. Natl.

Acad. Sci. USA 79 (1982) 1428-1432.

Maxam, A.M. and Gilbert, W.: Sequencing end-labeled DNA

with base-specific chemical cleavages. Methods Enzymol. 65

(1980) 499-560.

Miller, J.H.: The lad gene: its role in luc operon control and its

use as a genetic system. In Miller, J.H. and Reznikoff, W.S.

(Eds.), The Operon. Cold Spring Harbor Laboratory, Cold

Spring Harbor, NY, 1978, pp. 31-88.

Miller, J.H.: Genetic studies ofthe luc repressor, XII. Amino acid

replacements in the DNA binding domain of the Escherichia

coli lac repressor. J. Mol. Biol. 180 (1984) 205-212.

Miller, J.H., Coulondre, C., Hofer, M., Schmeissner, U., Som-

mer, H., Schmitz, A. and Lu, P.: Genetic studies of the lac

repressor, IX. Generation of altered proteins by the sup-

pression of nonsense mutations. J. Mol. Biol. 131 (1979)

191-222.

Miller, J.H., Platt, T. and Weber, K.: Strains with the promoter

deletion Ll synthesize an altered luc repressor. In Beckwith,

292

J.R. and Zipser, D. (Eds.), The Lactose Operon. Cold Spring

Harbor Laboratory, Cold Spring Harbor, NY, 1970, pp.

343-35 1.

Miwa, J. and Sadler, J.R.: Characterization of i-“ repressor

mutations of the lactose operon. J. Mol. Biol. 117 (1977)

843-868.

Mott, J.E., Van Arsdell, J. and Platt, T.: Targeted mutagenesis

in vitro: lac repressor mutations generated using AMV re-

verse transcriptase and dBrUTP. Nucl. Acids Res. 12 (1984)

4139-4152.

Mtiller-Hill, B., Fanning, T., Geisler, N., Gho, D., Kania, J.,

Kathmann, P., Meissner, H., Schlotmann, M., Schmitz, A.,

Triesch, I. and Beyreuther, K.: The active sites of lac repres-

sor. In Sund, H. and Blauer, G. (Eds.), Protein-Ligand Inter-

actions. Walter de Gruyter, Berlin, 1975, pp. 21 l-224.

Pfahl, M.: Mapping of I gene mutations which lead to repressors

with increased affinity for luc operator. J. Mol. Biol. 147

(1981) 175-178.

Pfahl, M. and Hendricks, M.: Interaction of tight binding repres-

sors with lac operators: an analysis by DNA-footprinting. J.

Mol. Biol. 172 (1984) 405-416.

Sadler, J.R., Betz, J.L., Tecklenburg, M., Goeddel, D.V.,

Yansura, D.G. and Caruthers, M.H.: Cloning of chemically

synthesized lactose operators, II. EcoRI-linkered operators.

Gene 3 (1978) 211-232.

Sadler, J.R. and Tecklenburg, M.: Recovery of operator DNA

binding activity from denatured lactose repressor. Biochem-

istry 15 (1976) 4353-4356.

Sadler, J.R.,Tecklenburg, M. and Betz, J.L.: Plasmids containing

many tandem copies of a synthetic lactose operator. Gene 8

(1980) 279-300.

Sancar, A., Hack, A.M. and Rupp, W.D.: A simple method for

identification of plasmid-coded proteins. J. Bacterial. 137

(1979) 692-693.

Schlotmann, M. and Beyreuther, K.: Degradation of the DNA-

binding domain ofwild-type and ied lac repressors in Escheri-

chia coli. Eur. J. Biochem. 95 (1979) 39-49.

Schmitz, A., Coulondre, C. and Miller, J.H.: Genetic studies of

the lac repressor, V. Repressors which bind operator more

tightly generated by suppression and reversion of nonsense

mutations. J. Mol. Biol. 123 (1978) 431-456.

Simoncsits, A. and Torok, I.: A photoinduced cleavage of DNA

useful for determining T residues. Nucl. Acids Res. 10 ( 1982)

7959-7964.

von Wilcken-Bergmann, B. and Muller-Hill, B.: Sequence of

galR gene indicates a common evolutionary origin of lac and

gal repressor in Escherichia cob. Proc. Natl. Acad. Sci. USA

79 (1982) 2427-2431.

Weber, K., Files, J.G., Platt, T., Ganem, D. and Miller, J.H.: Lac

repressor. In Sund, H. and Blauer, G. (Eds.), Protein-Ligand

interactions, Walter de Gruyter, Berlin, 1975, pp. 228-235.

Weber, K. and Geisler, N.: lac repressor fragments produced

in vivo and in vitro: an approach to the understanding of the

interaction of repressor and DNA. In Miller, J.H. and Rezni-

koff, W.S. (Eds.), The Operon. Cold Spring Harbor Labora-

tory, Cold Spring Harbor, NY, 1978, pp. 155-175.

Weber, LT., McKay, D.B. and Steitz, T.A.: Two helix DNA

binding motifof CAP found in lac repressor and gal repressor.

Nucl. Acids Res. 10 (1982) 5085-5102.

Young, R.A. and Davis, R.W.: Efficient isolation of gents by

using antibody probes. Proc. Natl. Acad. Sci. USA 80 (1983)

1194-1198.

Communicated by D. Bastia.