GeIIP. 118 (1992) 137-141 0 1992 Elacvicr Scicncc

GENE

Publishers

B.V. All rights reserved.

137

0378-l 119:92:$05.00

06559

GAL4 fusion vectors for expression in yeast or mammalian

cells

(Transcriptional activators; ADHl promoter)

site: SV40 early promoter;

Brendan

Ivan Sadowski”, ‘I Depcrrmer~t

chimeric

proteins;

Alderle~~ Pork.

Rcccivcd

bq’ D.T. Dcnhardt:

/I-galactosidase;

Bell”, Peter Broadb

of‘ Biochemi.rtr~~. Univervit~~ of British

Mereside.

CAT;

Macclesfield,

Cheshire.

23 Januaq

Columhicr.

SK10

4TG,

1992; Accepted:

cloning

and Melvyn Hollisb

Vancouver. UK.

yeast; multiple

BC.

Vii T 123.

Crmrrdtr; md

h Depcrrtmmt

of Biotechtrolog~~. ICI Phnrnltrc~eutic~~~ls,

Tel. 144)625514305

16 February

1992; Reccivcd

at publishers:

13 April 1992

SUMMARY

We describe two sets of vectors, one for yeast (pY 1, pY2 and pY3) and one for mammalian

cells (pM1, pM2, and pM3),

that simplify the production of fusion proteins containing the DNA-binding domain of GAL4. This protein fragment, consisting of GAL4 amino acid (aa) residues l-147, binds to a specific 17-bp nucleotide sequence, but is incapable of activating transcription unless fused to a protein that can contribute an activating function. Genetic strategies exploiting this property of GAL4 (aa l-147) have been developed to characterize transcription factor functional domains, proteinprotein interactions, and site-specific proteolysis. The vectors we describe allow fusion to the C terminus of GAL4 (aa l- 147) in any reading frame, and thus facilitate these experimental strategies.

1NTRODI)CTION

Transcriptional activators are sequence-specific DNAbinding proteins which, when bound to enhancer or upstream activating sequences (UAS) on DNA, stimulate transcription of nearby genes through direct contact with

Corresp~~~~dewe to: Dr. 1. Sadowski,

Scicnccs

Mall, Vancouver,

U.B.C.

Biochemistry,

BC. V6T 123, Canada.

2146 Health

Tel. (604)822-4524;

Fax (604)X22-5227. Abbreviations:

aa. amino acid(s); ADHl,

gene encoding

alcohol dehydro-

genase I; Ap. ampicillin; ARS, autonomously replicating sequence: /3Gal. E. coli /&galactosidase: /3Gal, E. coli P-galactosidase; bp, base pair(s): C.AT, chloramphenicol ccntromere; (DNA)

GAL4.

encoding

acetyltransferase; transcriptional

GAL4:

cat. gene encoding

activator

HSV. Herpes

CAT; CE,v.

of GAL genes; GAL4, gene

simplex

virus; kb, kilobase

or

1000 bp; MCS. multiple cloning site(s); nt, nucleotide(s); ohgo, oligodeoxqribonucleotide; PolIk, Klenow (large) fragment of E. coli DNA polymerase I; pADHI, promoter ofADHl gene; K. resistance/resistant; RIPA, radioimmunoprecipitation assay: SDS. sodium dodecyl sulfate; SV40, simian virus 40; tADHI,

terminator

of ADHI

gene: TRPl,

isomerasc:

qucnce(s):

::. novel joint (fusion or insertion).

VP. viral protein;

C/,4S, upstream

gene encoding

phosphoribosylanthranilate

activating

se-

proteins of the general transcriptional initiation complex (Ptashne and Gann, 1990). These transcriptional regulatory proteins generally possess several separable functional domains that can invariably tolerate fusion to heterologous proteins. The GAL4 protein of Saccharonzyces cerevisiae is one of the most thoroughly characterized transcriptional activators (Johnston, 1987). Since the N-terminal 147 aa residues of GAL4 are sufficient to mediate specific and strong binding to DNA, but are incapable of efficient transcriptional activation (Ma and Ptashne, 1987) this protein fragment has frequently been used to confer specific DNA binding in experiments examining transcriptional activation function of heterologous proteins. This approach is facilitated by the finding that higher eukaryotes lack endogenous proteins that enhance transcription from the consensus GAL4-binding site. Fusions between GAL4(aa 1- 147) and activating domains from a variety of transcriptional regulatory proteins can activate transcription in yeast, insects, plants, and mammalian cells (Ptashne and Gann, 1990). Furthermore, in several instances, fusion of signal- and ligand-regulated transcription factors to GAL4 (aa 1- 147) produced appropriately regulated chimeras

138 (Webster et al., 1988: Song et al., 1991). These results demonstratc the extensive potential of this approach for char-

The vectors described in this report lend significant versatility to thcsc types of strategies by greatly simplifying production of fusions with the DNA-binding domain of GAL4.

actcrizing function and regulation of eukaryotic transcription factors. Scvcral additional cxpcrimcntal strategies have recently been developed which exploit GAL4s separable DNA binding and transcriptional activation functions. Fields and couorkers have described a unique ‘two-hybrid‘ approach,

bXPERIMENTAL

using GAL4 fusions in yeast. to identify specific proteinprotein interactions (Fields and Song, 1989; Chien et al.,

(a) Construction

14i’bSer

pM21pY2

EcoRl Smal GAA TTC CCG Glu Phe Pro

BamHl GGG ATC Gly Ile

TCG

ECORI CGG AAT Arg Asn

Smal BamHl TCC CGG GGA TCC Ser Arg Gly Ser

TCG

EcQRI GGA ATT Gly Ile

Smal CCC Pro

147bSer

pM31pY3

CCG Pro

147)&r

fusion

vectors

pY (6.7 kb)

pM (3.5 kb)

TCG

of the GAL4

The pM vectors are derived from a previously described GAL4 fusion vector pSG424 (Sadowski and Ptashne. 1989), which is itself a derivative of pECE (Ellis ct al., 1986). Since we found that pSG424 often gives poor ciclds of plasmid DNA, and there is a PHI site within the ApK gene of this plasmid. WC transferred the SV40 carlq promoter,!ori region . the coding sequcncc for GAL4 XI 1-147, and SV40 polyadenylation region from pSGJ24 to pSP72. All of the pSP72 MCS restriction sites Ilanking the insert were eliminated by digestion and end filling \vith PolIK. The Hilldill site immediately upstream from the GAL4 coding sequence was converted to an ;VheI site b>

1991); a variant of this technique has also been developed for use in mammalian cells (Vasavada et al., 1991). GAL4s DNA-binding and activating domains can activate transcription as a complex, when fused separately to proteins that form specific interactions. A reverse strategy, using GAL4 functional domains, has been dcvclopcd to charactcrizc specific proteolysis. A protcasc clcavagc site. placed between GAL4s DNA-binding and activating domains. inhibits transcriptional activation by the fusion: mutations in the protcolysis site can bc identified genetically by GALIhZ rcportcr gcnc cxprcssion (Dasmahupatra ct al.. 1992).

pMl/pYl

AND DISCUSSION

BamHl GGG GAT Gly Asp

SalI MlUl CGT CGA CGC Arg Arg Arg

PS!l GTC TGC Val Cys

Hindlll AGA AGC Arg Ser

Xbal TTC TAG Phe ---

Sal1 GTC Val

Mlul GAC GCG Asp Ala

Pstl TCT GCA Ser Ala

Hindlll GAA GCT Glu Ala

Xbal TCT Ser

Mlul ACG Thr

Pstl CTG Leu

Hindlll Xba AAG CTT CTA Lys Leu Leu

Sall CCG TCG Pro Ser

CGT Arg

CAG Gln

ATA

AGT

AA

AGA Arg

TAA ---

GTA

A

GAT Asp

AAG Lys

TAA ---

Fba.. I Ph\\ical map ofthc GAL1 fuslon;cxpreasion \cctors. The phl xectors are pSP72-derived plasmids with the SV30 earl> promotcr’w region (SVJII _; car-l) j,~n) directing rhc express~m of GAL4 (aa I- 147). Transcripts xc tcrminatcd within the SV40 earl> polqadenjlation region (SVSO p&A). The pk !ca\t shuttlc vectors ha\c GA1.1 (aa i-117) cxprcssed from the ADHI promoter (/ADHI): transcripts arc terminated hq the ADHI tcrmmator (fADHI I. The pY vectorsrcplicatc 3s single cop> plnsmids in yeast (ARS-C‘E,‘V) and have TRPI sclcction. Both sets ofvcctors have MCS following GAL4 (a;~ l-I-17). the scqucnccs encoded

of which are sho\ln

at the bottom.

hq the MCS arc indicated.

Each ofthc vectors ha\e

slop codons in all three rending frames follo\Qlng the MCS. Prcdrctcd :I:, rwrduc\

139 digestion, end-filling with PolIk, and religation. The MCS following GAL4 was modified by exchanging the &II-XbuI portion of the MCS with that of pMTL23 (Chambers et al., 1988); the resulting construct was called pM1. EcoRI sites were inserted into the two other reading frames following GAL4 codon 147 by digesting wt GAL4 DNA (Laughon and Gesteland, 1984) with CIaI and ligating to selfannealing adapters with the sequences: 5’-CGCGGAATTCCG and 5’-CGGGAATTCC. From these constructs, XhoI-EcoRI fragments, bearing GAL4 codons 74147, were subcloned into pM1 to form pM2 and pM3, respectively (Fig. 1). The yeast pY vectors are derived from pMH76, a GAL4 yeast expression vector (Sadowski et al., 1991). Plasmid pMH76 is pUC19-based, with a TRPI selectable marker and ARS-CEN replicon. TRPl was derived from an 830nt EcoRI-PstI fragment (Tschumper and Carbon, 1980), cloned into the Ah1 site at nt 629 of pUC19 using BglII linkers. The ARS-CEN fragment was inserted as a C/u1 fragment at nt 747 of pUC19, and consists of 400 nt of CENVI (Cottarel et al., 1989) and an uncharacterized 410nt ARS The ADHl promoter, GAL4 coding sequence and ADHI terminator were inserted between the EcoRI and HilzdIII sites of the pUC19 polylinker as a BanlHI fragment (Ma and Ptashne, 1987; Bennetzen and Hall, 1982) made blunt by treatment with PolIk. Plasmid pMH76 also has the HirzdIII site at the 5’ end of the GAL4 coding sequence changed to an NheI site (see above). A BamHI linker was inserted into the remaining HilldIII site at the 3’ end of the GAL4 coding sequence in pMH76. Then pY 1, pY2, and pY3 were constructed by inserting XhoI-BclI fragments containing GAL4 codons 74-147, polylinker, and the triple translational stop codons from pM1, pM2. and pM3, respectively, between the X/z01 and BurlzHI sites of the modified pMH76. (b) Transcriptional activation by GAL4-VP16 fusions produced from the pM and pY vectors For each of the vectors, we constructed in-frame fusions with the activating domain of HSV VP16. BglII-HifldIII fragments, bearing the coding sequence for aa 413-490 of VP16 and approx. 200 3’ untranslated nt, were subcloned from pCRF2, pCRF1, and pCRF3 (Triezenberg et al., 1988) between the BarnHI and Hind111 sites of pMl/pY 1, pM2/pY2, and pM3/pY3, respectively. We confirmed that the fusions were in the correct reading frame by doublestranded DNA sequencing using an oligo primer spanning GAL4 codons 133 to 138 (5’-TCGGAAGAGAGTAGTAAC) (Laughon and Gesteland, 1984); the junction sequences are shown in Fig. 2a. We found that all of the GAL4-VP 16 fusion proteins, and the proteins produced by the corresponding parent vectors, were easily detectable by immunoprecipitation from [ “Slmethionine-labeled COS- 1

GALCVP16 Junction Sequences

(a)

pM1VP16/pYiVP16

Smal EC&l TCG CCG GAA TTC CCG GGG ATC TGC Ser Pro Glu Phe Pro Gly I le Cys

GAL4 aa 147 Smal

ECORI

pM2VP16/pYNP16

GCC

Ala VP16 aa 413

TCG CGG AAT TCC CGG GGA TCl Ser Arg Asn Ser Arg Gly Ser GAL4 aa 147

GCG GCC Ala Ala VP16 aa

Smal ATT CCC GGG GAT CTG GCC I le Pro Gly Asp Leu Ala VP16 aa

EcoRl

pM3VP161pY3VP16

TCG

GGA

Ser Gly GAL4 aa 147

(4

M1

M2

M3

M2 VP16

“::6

M3 VP16

29 kDa 4

24 kDa -

20 kDa -

GAL4-VP16

+GAL4(1-147)

14 kDaFig. 2. GAL4

derivative proteins produced by the fusion/cwprcsaion [cc-

tora. (a) Nucleotide

and deduced aa sequencea of the GALA-VP10

junctions.

(b) Immunoprccipitates

rivativcs.

COS-1

cells transfected

of [“Slmcthlonine-labeled with the indicated

plasmids were labeled for 2 h with [“Slmethioninc. RIP.4 buffer (Tris.HCI 0.5”,, Na,deoxycholatc/ munoprecipitated

pH 8.0;’ 100 mM NaCI; O.l”,, SDS),

GAL4

GAL/

dc-

expresswn

Following

I mM EDTA,

fusion

GAL3

lqsis in

I “,, NP40/

d erivative proteins v.crc in-

with rabbit polyclonal antisera as described previouslq

(Gill et al.. 1990: Sadowski a 0.1 ‘I,, SDS-7.5”,

ct al., 1991). The proteins were resolved on

polyacrylamide

phi. The positions of GAL4

gel, and visualized by autoradiogra-

(aa 1-147)

and GALJ-VP16

fusion dcrlv-

ati\cs are indicated.

cells (Fig. 2b). Note that the immunoprecipitate of M2VP 16 contains a degradation product with mobility similar to thaL of GAL4(aa 1-147). Since we cannot detect such fragments of the other two fusions, and the N-terminal degraded product of M2VP16 is slightly smaller in apparent M, than the corresponding M2 GAL4( 1- 147), WC suspect that the junction sequence in that particular fusion contains a protease sensitive site (Fig. 2a). All of the GAL4-VP 16 fusion proteins activate transcription efficiently in both yeast and mammalian cells (Table 1. and Fig. 3). In mammalian cells WCused a cut reporter gent with a minimal promoter consisting offive consensus GAL4 DNA-binding sites immediately upstream from the adcnovirus Elb TATA box (Fig. 3, bottom). Fig. 3 shows the results of a cotransfection of the parental pM vectors and the GAL4-VP16 expression plasmids with this reporter in COS-1 cells. Although basal level expression of pG5EC is quite high in these cells (M 1. M2 and M3), we find that all of the GAL4-VP16 fusions activate transcription approx-

when Each least Table

produced from the pY vectors in of the GAL4-VP16 fusions activated 1200 fold over their corresponding I for /?Gal reporter enzyme encoded

yeast (Table I). transcription at pY parent (see by GALI- IuZ).

(c) Conclusions

See Figs. I and2. ” Yeaststrain YTO:: 17 I (Hm~mclfxb rcportu

ct al.. 1990). bearing a (;.4L/-ltrc%

gcnc integrated

nt L’RA3. wcrc tranaformed with the indicated and gro\sn to ,4(,,,,j,,,11= 0.6 in sclcctivc minimal media contammg

plaullid

2”,, zlvccrol 2” 0 lactic acid. flGal activity I! at,;

as doscribed

by Himmelfarb

M’

M2

was determined

from crude

Ml M2 M3 M3 vP16vP16vP16

HTATAH

:\CKNOWLEDGEMENTS

The experiments described in this report were supported by funds from the M.R.C. and N.C.I. of Canada. We thank Dyanne Niedbala for technical assistance. I.S. is a Research Scientist of the National Cancer Institute of Canada.

cat

5 X GAL4 binding sites Fig. 3. Transcriptional the phi \ectora derivatives

actlcation

were cotransfected

4X h post-transfcction

prwousl) consisting

dcriwtivcs

cxprcssing

with the pG5EC

COS-1 cells. using DEAE-dcxtran \csted

by GAL3

in COS cells. Plnsmlds

teins produced by the pM vectors and their VP16 fusion derivatives can be detected by immunoprccipitation. (2) The restriction sites in the MCS arc unique in all of the vectors, with the exception of XbaI, for which there is an additional site within TRPI of the pY vectors (Fig. 1). Each of the MCS are followed by stop codons in all three reading frames. (3) The VP16-activating domain is one of the strongest acidic activating domains characterized (Sadowski et al., 1988), and is often used in fusions with other DNA binding domains (Vasavada et al., 1991). The GAL4-VP16 fusion constructs described hcrc (Fig. la) have been useful for constructing such chimeras (unpublished). (4) The GAL4 DNA-binding domain can also function when positioned at the C terminus, or in the center. of fusion proteins; a plasmid for construction of such fusions has been described (Raycroft and Lozano. 1992).

et al. (1990).

pG5EC reporter construct I I I I I

(1) The vectors we describe simplify production of fusion proteins containing the DNA-binding domain of GAL4 by allowing fusion in any reading frame. Each of the vectors produce GAL4 (aa l-147), which in itself does not activate transcription, but, when DNA encoding the activating domain of VP16 is inserted into the MCS in frame with GAL4. very strong transcriptional activators arc produced. Pro-

produced

the indicated

reporter

construct

from GAL4 into

REFERENCES

(Gill et al.. 1990). The cells uere har-

and assayed

(Gill ct al.. 1990). pG5tC of five conscnbus ‘17.mcr

for CAT activltb a\ dcscrlbed

(bottom) has a minm~al pt-omotcr GAL4 binding sitcs upstream from

the a&no\ irus E I b T-\T \ box. The cur gcnc transcripts arc terminated within a fragncnt containing the SV30 carly pol>adcn>lation resjon (not shon n).

imatcly 600-fold (MlVP16, M2VPl6, and M3VP16). Similar levels of GAL4-VP 16 activation, but much lower basal levels, arc obscrvcd in CHO, HeLa, and Rat-2 cells (not shown). GAL4-VP16 fusions were also found to efficiently activate a GA Ll -hZ reporter gene (Ma and Ptashne, 1987)

Benneuen, J.L. and Hall, B.D.: The primary structure of the Strc.chtrn,fi?~,~e.s c~rrcrisirre gene for alcohol dehydrogenasc I. J. Biol. Chem. 257 (1982) 3018-3025. Chambers, S.P.. Prior, S.E.. Barsto\\. nit

cloning vectors, 1. Improved

the use of sonicated 130-119.

D.A. and hlmton. pUC polylinker

DNA for nuclcotide

Chien, C.-T.. Bartel. P.L.. Stern&lxx,

N.P.. The phlT1. regions to fxiiitatc

scqucncing.

Gcnc 68

C198X)

R. and Fields, S.: The t\\o-hybrid

aqstcm: n m&hod to identify and clone genes for protems that intcract with a protein of intcrcat. Proc. Natl. Acad. Sci. L’S,\ XX(1991) 9578-9582. Cottarcl. G., Shcro. J.H., Hieter. P. and Hegemann, J.H.: pair CENh DNA fragment is suficient for complete

,1 125.basemciotlc and

141 mitotic ccntromcre

functions

in Socchrrrwqws

cerrrisicre. Mol. Cell.

Biol. 9 (1989) 3342-3349. Dasmahapatra,

B., Dwyer. S., Ma, J., Sadowski,

J.: A novel genetic system to study a proteolytic

Proc. Natl. Acad.

I. and enzyme.

Sci. USA (1992) in press.

insulin-stimulated

kinase

activity

and

uptake

of

‘-deoxyglucosc. Cell 45 (1986) 721-732. Fields. S. and Song. 0.: A novel genetic system to detect protein-protein interactions.

Nature

Gill, G., Sadowski.

340 (1989) 245-246.

I. and Ptashne,

tivity of a transcriptional Proc. Natl. Acad. Himmclfarb. activators. Johnston.

that increase

the ac-

and mammalian

cells.

Sci. USA 87 (1990) 2127-2131.

H.J.. Pcnrlbcrg.

a geast mutation

in yeast

J.. Last. D.H. and Ptashne.

that potentiates

M.: GALI IP:

the effect of weak GAL4-derived

Cell 63 (1990) 1299-1309.

M.: A model fungal gcnc regulatory

mechanism:

the GAL genes

Rev. 5 I ( 1987) 458-476.

Laughon. :I. and Gcstcland. R.F.: Primary structure of the Sacch~lror~~,~ws c,rw\,;wre G/IL4 gcnc. Mol. Cell. Biol. 3 (1984) 260-267. scriptional

M.: Dclction

activating

scgmcnts.

analysis

of GAL4 dcfinca two tran-

Cell 4X (1987) 847-853.

L. and Lorano.

GAL4 DNA-bindmg

G.: A convenient domain.

Gene

S. and Ptashnc.

transcriptional

Sadowski.

1.. Nicdbala.

phorglatcd

activator.

GALJ( l-117)

M.: GAL4-VP16

Nature

Acad.

D.. Wood. K. and Ptashnc,

as a conscqucnce

i$ an

335 (1988) 563-

of transcriptional

M.: GAL4

1s phoa-

acti\ ation. Proc. Natl.

Sci. USA X8 (1991) IOjlO-lO514.

Song, O., Dolan. J.\V.. Yunn. Y.O. and Ficlda. S.: Pheromone-dcpcndcnt phosphorylation of the beast STE I7 protein corrclatcs with transcripTricrcnbcrg.

Genes

Dcvclop.

S.J.. Kingsbury.

5 (1991) 741-750.

R.C. and McKnight.

section of VPl6. the trans.activator ate earl! gcnc cxprcssion. Tschumpcr. G. and Carbon, containing

a chromosomal

cloning vector containing

118 (1992) 143-144.

Vasavada.

H.A., Ganguly.

S.hl.:

4 contingent

of hcrpcs

S.L.: Functional

dia-

simplex virus mlmedi-

Genes Develop. 2 (1988) 7 1X-729. J.: Sequcncc of a least DN,\ fragment and the 7RPI

replicator

protein

interactions

S.. Germino. replication

gcnc. Gene

IO

the

F.J.. Iyang. Z.X. and U’claaman.

assal

for the detection

of protcin-

in animal cell\. Proc. Natl. .Acad. Sci. L’SA XX

(1991) 10686-10690. Wcbatcr,

N.. Jin, J.R.. Green, S., Hollis, M. and Chambon.

UAS,, is a transcriptional

Ptaahnc. hl. and Gann. .A.A.F.: .Activators and targets. Nature 346 (1990) 329-331. Raycroft.

potent

for cxprcssing

Acids Rcs. I7 (1989) 7539.

(1980) 157-166.

of Suc,(,htrro,)~~,(,c.rcrw~~isitrr. Microblol.

Ma, J. and Ptashne.

M.: .A vector cells. Nucleic

I.. Ma. J., Trierenberg,

unusually

tional acti\at1on.

M.: Mutations

activator

Sadowski,

in mammalian

564.

Ellis. L., Clauser. E., Morgan, D.O.. Edery. M., Roth. R.A. and Rutter. W.J.: Replaccmcnt of insulin receptor tyrosine residues I 162 and 1163 compromises

1. and Ptashne,

fusions

B.. DiDomenico,

Schwartz.

Sadowski.

cncc of the GAL4

enhancer

trans-activator.

in human

P.: The qcast

HcLa cells in the prcs-

Cell 52 (19X8) 16% 178.