Proc. Natl. Ac ad. Sc i. USAVol. 87, pp. 9966-9970, December 1990Neurobiology

Another mechanism for creating diversity in y-aminobutyrate typeA receptors: RNA splicing directs expression of two forms of y2subunit, one of which contains a protein kinase Cphosphorylation site

(alternative splicing/cDNA/polymerase chain reaction)

PAUL WHITING*, RUTH M. MCKERNAN, AND LESLIE L. IVERSEN

Neuroscience Research Centre, Merck Sharp and Dohme Research Laboratories, Eastwick Road, Harlow, Essex, CM20 2QR, U.K.

Contributed by Leslie L. Iversen, September 14, 1990

ABSTRACT Diversity of y-aminobutyrate type A(GABAA) receptors has recently been proposed to be achievedby assembly of receptor subtypes from a multitude of subunits(al-6, (11-3, yl-2, and 6) encoded by different genes. Here wereport a further mechanism for creating GABAA receptordiversity: alternative RNA splicing. Two forms of bovine y2subunit cDNA were isolated (y2S and y2L) that differed by thepresence or absence of a 24-base-pair (8-amino acid) insertionin the cytoplasmic domain between the third and fourthputative membrane-spanning regions. Polymerase chain reac-tion from RNA demonstrated that the two forms of y2 subunitare expressed in bovine, human, and rat brain. Sequencing ofgenomic DNA clones encoding the y2 subunit demonstratedthat the 24-base-pair insert is organized as a separate exon.Analysis of the sequence of the 8-amino acid insert revealed thatit contains a protein kinase C consensus phosphorylation site.Expression of the large cytoplasmic loop domains of y2S andy2L in Escherichia coli, followed by phosphorylation of therecombinant proteins by protein kinase C, demonstrated thaty2L, but not y2S, could be phosphorylated. Thus the two formsof y2 subunit differ by the presence or absence of a proteinkinase C phosphorylation site. This mechanism for creatingGABAA receptor diversity may allow differential regulation ofthe function of receptor subtypes.

y-Aminobutyric acid (GABA) is the major inhibitory aminoacid neurotransmitter in the vertebrate nervous system. Thisfast synaptic inhibition is mediated by the opening of achloride channel that is intrinsic to the GABAA receptormacromolecule. GABAA receptors represent the site of ac-tion of several clinically important drugs, such as the widelyprescribed benzodiazepine anxiolytics and barbiturates (1-2).

Heterogeneity of the GABAA receptor-associated benzo-diazepine binding sites (BZ1 and BZ2) was initially identifiedusing classic ligand-binding approaches (3). More recently,cloning of cDNAs encoding subunits of GABAA receptors(al-6, 81-3, yl-2, and 8) has demonstrated considerably morediversity than originally thought, and indicated that thesereceptors are members of a ligand-gated ion-channel genesuperfamily (4-13). Expression of various combinations ofthree subunits (a, /3, and y) individually or together allowsformation of GABA-gated chloride channels (4-15), but onlywhen the y2 subunit is present (bovine cDNAs) is benzodi-azepine binding conferred (7, 12, 15). The type of benzodi-azepine pharmacology appears to be defined by the type of asubunit: while al confers a BZ1-type pharmacology, a2, a3,

and a5 appear to define the BZ2 type (12, 15). a6 confersbinding of the "alcohol antagonist" Ro15-4513 (13).Here we report a further mechanism for creating diversity

within the GABAA receptor gene family. We show thatalternative RNA splicing results in the expression in brain oftwo forms of y2 subunit that differ by the presence or absenceofan 8-amino acid insert. This insert contains a protein kinaseC (PKC) phosphorylation site, suggesting a novel mechanismof regulating the function of receptor subtypes.t

MATERIALS AND METHODSIsolation and Sequencing of cDNA Clones. Poly(A)+ RNA

was isolated from bovine cortex by standard procedures, andfirst-strand cDNA synthesis was performed from 1 ,ug ofRNA, using avian myeloblastosis virus reverse transcriptase(Boehringer Mannheim) and hexanucleotide random primers(Boehringer Mannheim) (16). For polymerase chain reaction(PCR; ref. 17), the following oligonucleotide primers (synthe-sized using an Applied Biosystems 380B synthesizer) derivedfrom the published human y2 sequence were used, incorpo-rating HindIl sites before the initiating ATG and after the stopcodon: 5'-AAA-AAA-AAG-CTT-GCC-ATG-AGT-TCA-CCA-AAT-ATA-TGG-3'; 5'-TAG-GTC-GAA-GCT-TCT-CAC-AGG-TAG-AGG-TAG-GAG-ACC-C-3'. PCR was per-formed using 10% of the first-strand cDNA, 1 gg of eachprimer, and 200 AM deoxynucleoside triphosphates, with 2.5units of Thermus aquaticus (Taq) DNA polymerase and 10xbuffer from Cambio (Cambridge, England). Cycling was per-formed at 94°C (45 sec), 55°C (2 min), and 72°C (5 min) for 30cycles. A 31st cycle had the elongation step extended for 10min at 72°C. PCR products were extracted with phenol/chloroform, ethanol-precipitated, digested with HindIll, andthen resolved in a 0.8% agarose gel. The 1400-base-pair (bp)band was excised, the DNA was eluted and subcloned intopBluescript SK(-) (Stratagene), and y2 cDNA clones wereidentified by restriction mapping and partial sequencing. Bothstrands of two clones, y2S and y2L, were sequenced fromdenatured, double-stranded DNA by using Sequenase T7polymerase (United States Biochemical).

Isolation and Sequencing of Genomic DNA Clones. A humangenomic library (constructed with A DASH; obtained fromStratagene) was screened (5 x 105 phage) with a 32P-labeledprobe sequence (labeled by random priming; ref. 16) encom-passing the putative large cytoplasmic domain of y2L. Theprobe sequence was excised by digestion with Nde I andHindIII from the vector pRSET5a (18), into which it had been

Abbreviations: GABA, y-aminobutyric acid; PKC, protein kinase C;PCR, polymerase chain reaction.*To whom reprint requests should be addressed.tThe sequence reported in this paper has been deposited in theGenBank data base (accession no. M55563).

9966

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Proc. Natl. Acad. Sci. USA 87 (1990) 9967

engineered for bacterial expression (see below). Nylon filters(Hybond, Amersham) with the transferred phage were hy-bridized overnight at 42TC in 40% (vol/vol) formamide/5xDenhardt's solution ( x is 0.02% Ficoll/0.02% polyyinylpyr-rolidone/0.02% bovine serum albumin)/5 x SSPE (1 x is 0.18M NaCI/0.01 M sodium phosphate, pH 7.4/1 mM EDTA)/0.5% SDS containing denatured salmon sperm DNA at 15,tg/ml and were washed at 60TC (twice, 15 min, 5x SSPE;twice, 10 min, 0.3x SSPE). Positive clones were plaque-purified and larger amounts of DNA were prepared fromplate lysates (16). The DNA inserts were restriction-mappedand fragments were subcloned into pBluescript SK(-) bystandard DNA cloning techniques (16). Subclones were se-quenced using oligonucleotide primers (16-mers, synthesizedon an Applied Biosystems 380B synthesizer) located in andaround the putative large cytoplasmic domain sequence.

Analysis of the Expression of Subunit mRNA by PCR.Poly(A)+ RNA was isolated from brain tissue (human brainRNA was purchased from Clontech) and first-strand cDNAsynthesis was performed as described above. PCR usedoligonucleotide primers derived from the published sequenceof human y2: forward primer (starting at bp 1051, Fig. 1),5' -GTG-GAG-TAT-CAT-ATG-TTG-CAT-TAT-TTT-GTC-3'; reverse primer (starting at bp 1341, Fig. 1), 5'-GAT-CCA-AGG-TTA-GGA-CTA-CAT-lTT-GGC-3'. PCR wasperformed essentially as described above, with the exceptionthat 4 ,tCi of [a-32P]dATP (3000 Ci/mmol; 1 Ci = 37 GBq) wasincluded and cycling was performed at 94°C (45 sec), 60°C (2min), and 72°C (2 min). The 31st cycle had an extension timeof 4 min. The PCR products were extracted with phenol/chloroform and precipitated with ethanol. The products werethen resolved in a denaturing 6% polyacrylamide gel andvisualized by autoradiography. Size was determined by usinga sequencing "ladder" resolved in the same gel.

Expression of Putative Large Cytoplasmic Loop Domains inEscherichia coli. The DNA sequences of y2L and y2S en-compassing the putative large cytoplasmic loop domain wereamplified from 1 ,ug of plasmid DNA by PCR using oligonu-cleotide primers and conditions described immediatelyabove. PCR products were digested with Nde I and HindIII,purified in 0.8% agarose gels, and ligated into Nde I/HindIII-digested pRSET5a, a T7 polymerase bacterial expressionvector (18, 19). For expression of protein, expression vectorswere transformed into E. coli strain BL21(DE3) (Lys S).Bacteria were grown in LB medium containing ampicillin (20,ug/ml) to an OD6w of 0.6, and then expression of recombi-nant protein was induced by addition of 1 mM isopropyl/3-D-thiogalactoside. After 3 hr of induction, cells were har-vested by centrifugation (1000 x g, 10 min) and the pellet washomogenized (three times, 20 sec, setting 10, Ultra Turraxhomogenizer) in 10 mM Tris-HCl/5 mM EDTA/0.1 mMphenylmethylsulfonyl fluoride, pH 7.5. Insoluble material,containing the recombinant proteins, was recovered by cen-trifugation (20,000 x g, 15 min) and washed twice further byrepeated homogenization and centrifugation. Recombinanty2L and y2S proteins were further purified by preparativeSDS/15% PAGE and recovered by electroelution. Proteinswere analyzed by SDS/15% PAGE and staining withCoomassie blue, and the protein sequence was confirmed byN-terminal amino acid analysis.PKC Phosphorylation. For phosphorylation by PKC, re-

combinant proteins, either as crude, insoluble material or asprotein purified by preparative SDS/PAGE, and syntheticpeptide (Pro-Leu-Leu-Arg-Met-Phe-Ser-Phe-Lys-Ala, syn-thesized by Peninsula Laboratories) were incubated at 30°Cfor 4 min with 1 mM [y-32PJATP (30 Ci/mmol), PKC (con-taining a, /31, and y isoforms, gift of Peter Parker, LudwigInstitute, London) at 1 unit/ml, 0.7 mM CaC12, 0.125 mMMgCI2, and 50 mM Hepes (pH 7.5) in the presence or absenceof lipids (phosphatidylserine, 125 mg/ml; phorbol 12-

myristate 13-acetate, 1.25 /lg/ml). 32p incorporation wasmeasured by binding of protein to P81 ion-exchange paper(Whatman), washing three times for 10 min with 30% (vol/vol) acetic acid, and counting of Cerenkov radiation.

RESULTSIsolation of cDNAs Encoding Two Forms of Bovine y2

Subunit. In preliminary studies, oligonucleotide primers de-rived from the human y2 sequence (7) flanking the largeputative cytoplasmic loop domain of the y2 subunit (Leu317-Met4O2) were used to isolate cDNAs from bovine brainmRNA by PCR. PCR products were subcloned, and whenone of the subclones was sequenced the deduced amino acidsequence was found to be identical to the human y2 sequenceexcept for an insertion of 8 amino acids (Leu-Leu-Arg-Met-Phe-Ser-Phe-Lys) between Pro337 and Ala338 (data notshown). Full-length cDNAs encoding the bovine y2 subunitwere then isolated by using oligonucleotide primers derivedagain from the published human y2 sequence, but flanking thestart and stop codons (see Materials and Methods). Isolatedy2 cDNA clones were initially characterized by sequencingthe region containing the additional 8 amino acids.

Five out of 14 cDNA clones contained the additionalsequence. Two clones, one that had the insertion (y2L) andone that did not (y2S), were completely sequenced. Notincluding the 8-amino acid insertion, the two cDNAs wereidentical apart from base 687 (G -* A, Gly -* Ser in y2L) andbase 897 (C -+ T, Ser -- Phe in y2L) (Fig. 1). These two basechanges may represent polymorphism or, more likely, errorsincorporated in the PCR. Bovine y2S has 98.5% amino acidsequence identity with human y2 (7). It thus contains themotifs of the ligand-gated ion-channel gene family, includingfour putative transmembrane regions and a large putativecytoplasmic loop domain between the third and fourth trans-membrane regions, where the 8-amino acid insert is located.As discussed in more detail below, this insert contains a PKCconsensus phosphorylation site.The 8-Amino Acid Insertion Is Encoded by a Separate Exon.

To characterize further the basis of the 8-amino acid inser-tion, a human genomic library was screened with a cDNAprobe derived from the sequence encompassing Leu317-Met402 of y2L. A clone containing a 14-kilobase (kb) insertwas isolated and a 6.5-kb Xba I fragment, which hybridizedto an oligonucleotide probe just 5' to the 24-bp insert, and a4.8-kb EcoRI-Xba I fragment, which hybridized to an oligo-nucleotide probe just 3' to the 24-bp insert, were subcloned.By using sequencing primers located around and within thelarge putative cytoplasmic loop domain, the intron/exonorganization of this part of the gene was determined (Fig. 2).The 24-bp insertion, which was identical to the correspondingbovine sequence, was found to exist as a separate exon.The base sequence around the downstream intron/exon

boundaries corresponds well to consensus donor-acceptorsplice sequences (20). This correspondence was less obviousfor the upstream donor-acceptor splice sequences, suggest-ing that splicing at this site involves a different mechanism,which could perhaps be differentially regulated. No otherintrons were found in the genomic sequence corresponding toLeu317-Met402. These data strongly suggest that the twoforms of y2 subunit arise by alternative splicing of RNAderived from a single gene leading to the inclusion, oromission, of the 24-bp sequence in the mature mRNA.Both y2S and y2L mRNA Are Expressed in Nervous Tissue.

PCR using oligonucleotide primers flanking the large putativecytoplasmic loop of the y2 subunit was used to examine theexpression of y2S and y2L mRNA. Two products werevisualized, 314 bp (corresponding to y2L) and 290 bp (cor-responding to y2S), in bovine brain and also in rat and humanbrain (Fig. 3). Previous in situ hybridization studies using an

Neurobiology: Whiting et al.

9968 Neurobiology: Whiting et al. Proc. Natl. Acad. Sci. USA 87 (1990)

ATG AGT TCA-39 MET Ser Ser

TTT TCA SAG-21 Phe Ser n

CTC ACT AGA-3 Leu Thr Arg

16 TrpCTC

34 Leu

TTA52 Leu

ATG70 ME

AAA88 Lys

ATC106 Ilie

GTC TTGVa Leu

CTG GAALeu lu

ATT CACIle His

GAA TATGlu Tyr

TTT AACPhe Asn

TGG ATTTrp Ile

CCA AAT ATA TGGPro Asn Ile Trp

AAA ATG ACG CTGLys MET Thr Leu

AAGin

ACTThr

GASTY

ACAThr

AGCSer

CCAPro

AAA TCT GATLys Ser Asp

CCA AAA GTTPro Lys Val

IAT GAT AATTyr Asp Asn

GAC AGITATAsp ME yyr

ATT GAT ATAIle Asp Ile

§C ATT AAAhr IeeLys

GAC CT TTCAsp Thr Phe

ATA ACC ACC CCT AAT AGG124 Ile Thr Thr Pro Asn Arg

ACT TTA142 Thr Leu

ATG GAT160 MET Asp

GAA ATT178 Glu Ile

TGG AGA196 Trp Arg

AAG ACG214 Lys Thr

AGA ATG232 Arg MET

CTA TCT250 Leu Ser

AGA CTG ACA ATTArg Leu Thr Ile

STA CAT TCC TGCGy His Ser Cys

GTC TAT CAG TGGVal Tyr GIn Trp

CTT TAT CAA TTTLeu Tyr Gin Phe

ACT TCT GGG GATThr Ser Gly Asp

GTC TAC TTC ACCGly Tyr Phe Thr

TGG GTG TCT TTCTrp Val Ser Phe

AGC ACA GGA AGCSer Thr Gly Ser

IGG ATT CTG CTCTrp Ile Leu Leu

GAT GAC TAT GAAAsp Asp Tyr Glu

CCG SAG GGT GATro Giu Gly Asp

AAA CTT CGA CCTLys Leu Arg Pro

GTG AAT AGC ATTVal An Ser Ile

TT TTGC APhehe A a I

GTC CTC CGA CTGVal Leu Arg Leu

TTC AGA AAT TCCPhe Arg Asn Ser

ATQ CTGMET Leu

GAT GCTAsp Ala

CCC CTGPro Leu

AAG CGTLys Arg

TCA TTTSer Phe

TAT GTGTyr Val

ATC CAGIle Gln

TGG ATCTrp Ile

TTA GGA ATC ACC ACT GTC CTG268 Leu Gly Ile Thr Thr Val Leu

TCT CTC CCC AAG GTC TCC TAC286 Ser Leu Pro Lys Val Ser Tyr

TTC ATC TTT GTC TTC TCT GCT304 Phe Ile Phe Val Phe Ser Ala

TCA GTC TAC TCGSer Val Tyr Ser

CTG CTA TCG CTCLeu Leu Ser Leu

GAT TAT GCT TCTAsp Tyr Ala ser

GTC ACA GTC ATCVal Thr Val Ile

GAT ATA GGA GTGAsp Ile Gly Va1

CQTsCA GTT AATu y Pro Va Asn

ACA TGG TAT GACThr Trp Tyr Asp

AAT AGC AAC AIGAsn Ser Asn MET

AAA AAG GCA GACLye Lys Ala Asp

AGA ATT TGG AAT GATArg Ile Trp Asn Asp

SAG TGC SAG TTA CAGG1U Cys Gln Leu Iln

AA TTC TCC AGT TATGlu Phe Ser Ser Tyr

AGT TCT GTT GAA GTGSer Ser Val Glu Val

GTT GGA CTG AGA AATVal Gly Leu Arg Asn

GTC ATG ACT GTC TATVai MET Thr Vai Tyr

ACC TAC ATC CCC TGTThr Tyr Ile Pro Cys

AAT AAG GAT GCT GTTAen Lys Asp Ala Val

ACA ATG ACT ACCThr MET Thr Thr

GTi TACA G~iGVa h A a

TTG GTG GAGLeu Val Glu

ST CGAy Arg

TTG CACLeu His

GGC TATG y Tyr

GGT GACGly Asp

ACC ACTThr Thr

TCT GACSer Asp

ACA CTCThr Leu

CCA GCCPro Ala

CTC AGC ACC ATTLeu Ser Thr Ile

ATG GAT CTC TTT GTAMET Asp Leu Phe Val

TAT GGC ACT CTG CATTyr Gly Thr Leu His

5~4ACT CCT GAhr Pro Val

TAC pCG >CTyr ro y

162AAC AAA A CAAsn Lys hr

TTA AAC WLeu Asn Asn

eAA CCA ACys ro hr

32GlCT AITC 3AAA a Is Asn

AGA CGT FSArg Arg Leu

432GT 8 GG AAAVa y Lys

G5A CAC T 8A a His rp

GTG CTC AVa Leu Tyr

AAT pTTpTAAsn he ro

CCA CGT 8ARro Arg Ilu

ACA AGG TThr Arg Ser

7%gAA GTA GTlu Val Val

610CTG AGC AGALeu Ser Arg

ATT GTT CIle Val Val

AGA ACT 9T8TArg Thr Ser

972GCC CGG AAAla Arg Lys

TCT GT TGCSer Va Cys

TAT TTT8GTyr Phe Vai

AGC AAC CGG AAA CCA AGC AAG GAC322 Ser Asn Arg Lys Pro Ser Lys Asp

GAC ATC CGCAsp Ile Arg

GAG AGA GATGlu Arg Asp

TTT TTC TGCPhe Phe Cys

CAC ATC CGTHis Ile Arg

TTC TGC TTGPhe Cys Leu

CCA CGA TCAPro Arg Ser

GAA GAA TACGlu Glu Tyr

TGT TTT GAACys Phe Glu

ATT GCC AAGIle Ala Lye

TTC AAT CTGPhe Aen Leu

CTT CTT CGG ATG TTT TCC TTC AAGLeu Leu Arg Met Phe S r Phe LysI * ,H IA:1

AAA GAC AAA AAG AAG AAA AAC pCT GCCC CTLys Asp Lys Lys Lys Lys Asn ro A a ro

GCA ACC ATT CAA ATG AAC AAT GCC ACA CACAla Thr Ile Gln MET Asn Asn Ala Thr His

GGG TAT GAG TGT CTG GAT GGC AAG GAC GGly Tyr Glu Cys Leu Asp Gly Lys Asp Cys

GAT TGC CGA ACA GGG GCT TGG AGA CAClipAsp Cys Arg Thr ly Ala Trp Arg His

ATG GAC TCC TAT GCG CGG ATC TTC TTC CCCMET Asp Ser Tyr Ala Arg Ile Phe Phe Pro

GTT TAT TGG GTC TCC TAC CTC TAC CTG AVal Tyr rp Val Ser Tyr Leu yr Leu .

FIG. 1. Nucleotide and deduced amino acid sequence of bovine GABAA receptor y2 cDNA clone. The nucleotide sequence is numberedat top right of each line, and the amino acid sequence, using the numbering for the human y2 sequence (7), is numbered to the left of each line.Putative transmembrane regions are underlined. The 24-bp/8-amino acid insert of y2L is included as a separate line. The consensus PKCphosphorylation site is indicated (0).

oligonucleotide probe that would have detected both y2S andy2L demonstrated that the y2 subunit is relatively ubiquitousin its distribution in the brain (7, 8). Here we have confirmedexpression of y2 in the cortex, cerebellum, striatum, andhippocampus. Further, it appears that y2S and y2L may vary

considerably in their relative abundance in different brain

regions, suggesting that their expression is differentiallyregulated.The 8-Amino Acid Insert of y2L Contains a PKC Phosphor-

ylation Site. Analysis of the inserted 8-amino acid sequenceindicated that it contained a PKC consensus phosphorylationsite (Ser-Xaa-Lys) (21). Additionally, the presence of an

337 338..AsnPro LeuLeuArgMetPheSerPheLys AlaPro..

..AACCCTgtatgta.. ..cccaaagCTTCTTCGGATGTTTTCCTTCAAGgtataat.. ..gtcccagGCCCCT..

FIG. 2. Intron/exon arrangement of the human y2 gene around the 24-bp insertion sequence. The exon base sequence is shown in uppercase,and the intron base sequence in lowercase.

ATTIle

CAAGln

AGTSer

ATAlie

GCTAla

ACA340 Thr

CTT358 Leu

GCC376 Ala

AGG394 Arg

4 CT412 Thr

Proc. Natl. Acad. Sci. USA 87 (1990) 9969

.

0co

r

CZEI

CCmLo

x Oa) -

3.1

EE Co

CD) 0 0CO Q

oI._C) C) I

E

-Q-0

* -314 bp

lb --- -290 bp

FIG. 3. Isoforms ofGABAA receptor y2 subunit are expressed innervous tissue. Electrophoresis and autoradiography were used todetect 32P-labeled PCR products from RNA, synthesized usingoligonucleotide primers flanking the putative large cytoplasmic loopof the y2 subunit. The 314-bp product corresponds to y2L, and the290-bp product corresponds to y2S.

arginine three residues N-terminal to the serine would prob-ably serve to enhance the Vmax and Km of phosphorylation ofthe serine residue (21). The only other GABAA receptorsubunits reported to contain a PKC consensus phosphory-lation site are a5 and a6 (12, 13). The location of the 8-aminoacid insert in the large putative cytoplasmic loop domain,where phosphorylation sites are known to exist in othermembers of the ligand-gated ion-channel gene superfamily(22), suggested that this sequence may be a bona fide PKCphosphorylation site. To test this possibility, we investigatedwhether PKC could phosphorylate (i) the synthetic peptidePro-Leu-Leu-Arg-Met-Phe-Ser-Phe-Lys-Ala, (ii) the bacte-rially expressed fragment Leu317-Met42 of y2S, and (iii) thebacterially expressed fragment Leu317-Met402 of y2L. Fig. 4A

A

Y21. Y2S1...

46 --

30 -_

21.5 -5

14.3

16.5 -~- N

B _o0

x

._

c

0)

04-~0

0.

D

0.

0

E- CL

'-

E4)

0

shows SDS/PAGE and protein stain of the bacterially ex-pressed fragments. The large putative cytoplasmic loop pro-teins of y2L and y2S are the predominant polypeptides in theinsoluble material prepared from bacterial lysates. The pre-dicted molecular weights of 12,850 and 11,980 are in goodagreement with the apparent molecular weights of 15,500 and14,500 (y2L and y2S, respectively). The recombinant pro-teins were also further purified to homogeneity by prepara-tive SDS/PAGE and electroelution (data not shown). Fig. 4Bshows that PKC phosphorylates, in a phospholipid-dependent manner, Leu317-Met42 of 'y2L, but not Leu317-Met12 of y2S, both as "crude" insoluble material and asprotein purified by preparative electrophoresis. The syn-thetic peptide was also phosphorylated by PKC in a phos-pholipid-dependent manner. Thus the 8-amino acid insert ofy2L contains a PKC phosphorylation site.

DISCUSSIONPharmacological studies suggested the existence of threemain classes of GABAA receptor (2, 3, 23), BZ1, BZ2, and athird class of receptor that is insensitive to most "classical"benzodiazepines but has high affinity for the "alcohol antag-onist" Rol5-4513. cDNA cloning of GABAA receptor sub-units has demonstrated the existence of a gene family of evengreater diversity and heterogeneity (4-13). Here we demon-strate a further level of GABAA receptor diversity: alterna-tive forms of the y2 subunit created by RNA splicing.

Alternative RNA splicing as a mechanism for creatingdiversity has been reported for other receptor systems.Recent studies have demonstrated that two forms of D2dopamine receptor exist as a result of alternative RNAsplicing from a single gene (24-27). The physiological role ofthe two forms, which differ by a 29-amino acid insertion in thethird cytoplasmic loop, has yet to be determined. Similarly,Beeson et al. (28) have reported that RNA splicing results in

51-

4

3

2H

0Substrate

PS/PMA

- - Y2S Y2L Y2S Y2L Peptide

+- + +- +-_ +-_ +-_

*- crude - 4- purified -cytopbasmric loop domain cytoplasmic loop domain

FIG. 4. Phosphorylation by PKC of the putative large cytoplasmic loop domain of y2L but not y2S. (A) SDS/PAGE and protein stain ofthe insoluble material prepared from bacteria expressing the putative cytoplasmic loop domains of y2S and y2L. Arrows at left indicate theapparent molecular weights (Mr X 10-3) of marker proteins resolved in the same gel. Arrows to the right of the lanes indicate the protein bandsof y2L (upper arrow) and -y2S (lower arrow) cytoplasmic loop proteins. (B) Phosphorylation by PKC of the bacterially expressed putativecytoplasmic domains (both as a crude bacterial cell pellet and as protein purified by preparative SDS/PAGE) and the synthetic peptide(Pro-Leu-Leu-Arg-Met-Phe-Ser-Phe-Lys-Ala) encompassing the 8-amino acid insert of y2L. Incorporation of 32p is expressed as cpm/10 ,g ofprotein and is shown as the mean ± SD of triplicate incubations. In the absence of PKC, there was no incorporation of 32p (data not shown).PS, phosphatidylserine; PMA; phorbol 12-myristate 13-acetate.

Neurobiology: Whiting et al.

6r

T

- -Ir J-1 , rz,-Tl

T

9970 Neurobiology: Whiting et al.

expression of two forms of muscle nicotinic acetylcholinereceptor a subunit that differ by the presence or absence ofa 25-amino acid insertion at amino acid 58 of the polypeptide.Again, the function of this insertion remains unknown.Our data suggest that isoforms of GABAA receptors exist

that differ in the presence or absence of a PKC phosphor-ylation site on the y2 subunit. Further studies are needed todetermine which other subunits are associated with y2S/y2Lin vivo to form a native receptor macromolecule, whether thePKC site of y2L can be phosphorylated in the native recep-tor, and what effect this has upon receptor function/pharmacology. It is known that GABAA receptors can bephosphorylated by PKC and that activators of PKC signifi-cantly decrease GABAA receptor function (29-31). Addition-ally, there is a precedent from other receptor systems for thispossibility. For instance, nicotinic acetylcholine receptors ofTorpedo electric organ, which are structurally homologous toGABAA receptors (32), can be phosphorylated by PKC atsites located in the large putative cytoplasmic loop domain ofthe 8 subunit. This PKC phosphorylation site is conserved inthe subunit of muscle nicotinic acetylcholine receptor (33),and activation of PKC in cultured muscle cells results in adecreased sensitivity to acetylcholine and increased rate ofreceptor desensitization (34). Activation of receptor systems(e.g., some muscarinic, domaminergic, and serotininergicsystems) coupled through the phosphatidylinositol second-messenger system leads to the generation of 1,2-diacylglyc-erol, which activates PKC. Thus it is possible to speculatethat PKC-mediated phosphorylation of GABAA receptors,and subsequent modulation of their function, may allowinteraction and cross-talk between neurotransmitter path-ways. Interestingly, the different subspecies of PKC have adiscrete cellular distribution in the rat cerebral neocortex;80% of the ,81 type has been shown to be localized adjacentto the plasma membrane of GABAergic neurons (35). It willbe important to map the distribution of the y2L subunit.Additionally, since the y2 subunit is required for binding ofbenzodiazepine drugs to GABAA receptors (7), it may beimportant to determine whether PKC phosphorylation ofy2L-containing GABAA receptors (resulting perhaps inchanges in receptor function) has any relationship to themechanisms underlying the tolerance and dependence thatdevelop in response to long-term administration of benzodi-azepine anxiolytic drugs.

We thank Dr. Peter Parker (Ludwig Institute, London) for kind giftofPKC and for helpful discussions. We thank Dr. M. Sardana (MerckSharp and Dohme, West Point, PA) for protein sequencing.

1. Olsen, R. W. & Venter, J. C., eds. (1986) Benzodiazepine/GABA Receptors and Chloride Channels: Structural and Func-tional Properties (Liss, NY).

2. Lippa, A. S., Garrett, K. M., Tabakoff, B., Beer, B., Wenno-gle, L. P. & Meyerson, L. R. (1985) Brain Res. Bull. 14,189-195.

3. Squires, R. F., Benson, D. I., Braestrup, C., Coupet, J., Klep-ner, C. A., Myres, V. & Beer, B. (1979) Pharmacol. Biochem.Behav. 10, 825-830.

4. Schofield, P. R., Darlison, M. G., Fujita, N., Burt, D. R.,Stephenson, F. A., Rodriguez, H., Rhee, L. M., Ramachan-dran, J., Reale, V., Glencorse, T. A., Seeburg, P. H. & Bar-nard, E. A. (1987) Nature (London) 328, 221-227.

5. Levitan, E. S., Schofield, P. R., Burt, D. R., Rhee, L. M.,Wisden, W., Kohler, M., Fujita, N., Rodriguez, H. F.,

Stephenson, A., Darlison, M. G., Barnard, E. A. & Seeburg,P. H. (1988) Nature (London) 335, 76-79.

6. Ymer, S., Schofield, P. R., Draguhn, A., Werner, P., Kohler,M. & Seeburg, P. H. (1989) EMBO J. 8, 1665-1670.

7. Pritchett, D. B., Sontheimer, H., Shivers, B. D., Ymer, S.,Kettenmann, H., Schofield, P. R. & Seeburg, P. H. (1989)Nature (London) 338, 582-585.

8. Shivers, B. D., Killisch, I., Sprengel, R., Sontheimer, H.,Kohler, H., Schofield, P. R. & Seeburg, P. H. (1989) Neuron3, 327-337.

9. Ymer, S., Draguhn, A., Kohler, M., Schofield, P. R. & See-burg, P. H. (1989) FEBS Lett. 258, 119-122.

10. Krestchatisky, M., MacLennan, A. J., Chiang, M.-Y., Xu, W.,Jackson, M. B., Brecha, N., Sternini, C., Olsen, R. W. &Tobin, A. J. (1989) Neuron 3, 745-753.

11. Malherbe, P., Sigel, E., Baur, E., Persohn, E., Richards, J. G.& Mohler, J. (1990) FEBS Lett. 260, 261-265.

12. Pritchett, D. B. & Seeburg, P. H. (1990) J. Neurochem. 54,1802-1804.

13. Luddens, H., Pritchett, D. B., Lohler, M., Killisch, I., Kein-anen, K., Monyer, H., Sprengel, R. & Seeburg, P. H. (1990)Nature (London) 346, 648-651.

14. Pritchett, D. B., Sontheimer, H., Gorman, C. M., Ketten-mann, H. K., Seeburg, P. H. & Schofield, P. R. (1989) Science242, 1306-1308.

15. Pritchett, D. B., Luddens, H. & Seeburg, P. H. (1989) Science245, 1389-1391.

16. Berger, S. L. & Kimmel, A. R., eds. (1987) Methods Enzymol.152, 33-556.

17. Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi,R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science239, 437-489.

18. Diehl, R. E., Whiting, P., Potter, J., Gee, N., Ragan, C. 1.,Linemeyer, D., Schoepfer, R., Bennett, C. & Dixon, R. A. F.(1990) J. Biol. Chem. 265, 5946-5949.

19. Rosenberg, A. H., Lack, B. N., Chui, D. S., Lin, S. W.,Dunn, J. J. & Studier, F. W. (1987) Gene 56, 125-135.

20. Shapiro, M. B. & Senopathy, P. (1987) Nucleic Acids Res. 15,7155-7174.

21. Woodgett, J. R., Gould, K. L. & Hunter, T. (1986) Eur. J.Biochem. 161, 177-184.

22. Hemmings, H. C., Nairn, A. C., McGuiness, T. L., Huganir,R. L. & Greengard, P. (1989) FASEB J. 3, 1583-1592.

23. Harris, C. M. & Lal, H. (1988) Drug Dev. Res. 13, 187-203.24. Toso, R. D., Sommer, B., Ewert, M., Herb, A., Pritchett,

D. B., Bach, A., Shivers, B. D. & Seeburg, P. H. (1989)EMBOJ. 8, 4025-4034.

25. Giros, B., Sokoloff, P., Martres, M.-P., Riou, J.-F., Emorine,L. J. & Schwartz J.-C. (1989) Nature (London) 342, 923-925.

26. Monsma, F. J., Jr., McVittie, L. D., Gerfen, C. R., Mahan,L. C. & Sibley, D. R. (1989) Nature (London) 342, 926-928.

27. Chio, C. L., Hess, G. F., Graham, R. S. & Huff, R. M. (1990)Nature (London) 343, 266-269.

28. Beeson, D. A., Morris, A., Vincent, A. & Newsom-Davis, J.(1990) EMBO J. 9, 2101-2106.

29. Sigel, E. & Baur, R. (1988) Proc. Nati. Acad. Sci. USA 85,6192-61%.

30. Moran, 0. & Dascal, N. (1989) Mol. Brain Res. 5, 193-202.31. Browning, M. D., Bureau, M., Dudek, E. M. & Olsen, R. W.

(1990) Proc. Nati. Acad. Sci. USA 87, 1315-1318.32. Safran, A., Neumann, D. & Fuchs, S. (1987) J. Biol. Chem. 262,

10506-10511.33. Nef, P., Mauron, A., Stalder, R., Alliod, C. & Ballivet, M.

(1984) Proc. Nati. Acad. Sci. USA 81, 7975-7979.34. Eusebi, F., Molinaro, M. & Zani, B. M. (1985) J. Cell Biol. 100,

1339-1342.35. Tsujino, T., Kose, A., Saito, N. & Tanaka, C. (1990) J.

Neurosci. 10, 870-884.

Proc. Natl. Acad. Sci. USA 87 (1990)