Proc. Natl. Acad. Sci. USAVol. 90, pp. 3058-3062, April 1993Biochemistry

Molecular cloning and expression of cDNA for mammaliantranslation initiation factor 5

(protein synthesis/GTP hydrolysis/GTPase superfamily proteins)

KALLOL DAS, JORGE CHEVESICH, AND UMADAS MAITRADepartment of Developmental Biology and Cancer, Albert Einstein College of Medicine, Bronx, NY 10461

Communicated by Salome G. Waelsch, January 5, 1993

ABSTRACT Eukaryotic translation initiation factor 5(eIF-5) catalyzes the hydrolysis of GTP bound to the 40Sribosomal initiation complex (40S AUG Met-tRNAf-eIF-2-GTP) with the subsequent joining of a 60S ribosomal subunitresulting in the formation of a functional 80S initiation com-plex. A rat cDNA that encodes eIF-5 has been isolated andexpressed in Escherichia coli to yield a catalytically active eIF-5protein. The 3.55-kb cDNA encodes a protein of 429 aminoacids (calculated Mr 48,926) with properties that are similar toeIF-5 isolated from rabbit reticulocyte lysates. The deducedamino acid sequence of eIF-5 contains sequence motifs char-acteristic of proteins of the GTPase superfamily.

The mammalian translation initiation factor 5 (eIF-5), inconjunction with GTP and other initiation factors, plays anessential role in the initiation of protein synthesis (for re-views, see refs. 1-4). Following scanning of mRNA by the40S ribosomal subunit and formation of the 40S initiationcomplex (40S-mRNA'Met-tRNAf eIF-2-GTP), eIF-5 inter-acts with the 40S initiation complex to promote quantitativehydrolysis ofribosome-bound GTP (5-8). Hydrolysis ofGTPis essential for the release ofeIF-2 and guanine nucleotide (asan eIF-2-GDP complex) from the 40S ribosomal subunit andthe subsequentjoining ofthe 60S ribosomal subunit to the 40Scomplex to form a functional 80S initiation complex(80S mRNA Met-tRNAf) that is competent for peptidyl trans-fer (5-8).The molecular weight of eIF-5 has been controversial.

Earlier reports from several laboratories suggested that eIF-5purified from rabbit reticulocyte lysates is a protein ofmolecular mass ranging from 125 kDa (9) to 150-168 kDa(10-12). In contrast, later reports from this laboratory on thepurification and immunochemical characterization of eIF-5showed it is a monomer that migrated on SDS gels as a proteinof 58 kDa (13-15). However, eIF-5 continues to be listed inrecent review articles as a protein of 125 kDa (4) or 150 kDa(3).

In this communication, we report the cloning, sequencing,and expression of a rat cDNA encoding functional eIF-5.*The results demonstrate that mammalian eIF-5 has a relativemolecular mass of 48,926 that migrates on SDS gels as aprotein of about 58 kDa.

MATERIALS AND METHODSMaterials. The preparation of 35S-labeled rabbit liver ini-

tiator Met-tRNAf, ribosomal subunits from Artemia salinaeggs, purified eIF-2, and elF-5 from rabbit reticulocytelysates has been described (8, 15). Purified eIF-2 containeda-, 3-, and -subunits, whereas homogeneous eIF-5 is amonomeric protein of 58 kDa (15).

Protein Sequence Analysis. Purified eIF-5 (1 nmol) wasdigested with 200 mg ofCNBr in 70%o formic acid as describedby Charbonneau (16) and the resulting polypeptides wereresolved by Tricine/SDS/PAGE as described by Schaggerand von Jagow (17). The separated peptides were electro-phoretically transferred to a Problott membrane, stained withCoomassie blue, and then destained. The stained bandscorresponding to different CNBr peptides were excised, andthe three well-resolved bands were subjected to automatedEdman degradation on an Applied Biosystems 477A proteinsequencer to determine their amino acid sequence.Immunochemical Techniques. The preparation of polyclo-

nal antisera in rabbits against purified rabbit reticulocyteeIF-5 and affinity purification of anti-eIF-5 antibodies havebeen described (15). Affinity-purified anti-eIF-5 antibodieswere specific for denatured eIF-5 antigen (15) and wereroutinely used for immunochemical detection and for immu-noscreening of the Agtll library. A goat anti-rabbit IgGcoupled to alkaline phosphatase was used as the secondaryantibody to detect the binding of the primary antibody to theimmunoblots (15). For the preparation of antibodies directedagainst native eIF-5, a laying hen was immunized with 30 pgof purified rabbit reticulocyte eIF-5 and given a boosterinjection of 15 ,g of eIF-5 as described by Gassmann et al.(18). Polyclonal antibodies (IgY) were isolated from egg yolksof hens and purified by polyethylene glycol precipitation asdescribed (18). Immunoprecipitation of eIF-5 was carried outby incubating labeled eIF-5 with chicken egg yolk anti-eIF-5antibodies followed by the addition ofrabbit anti-chicken IgYas the secondary antibodies. The antigen-antibody complexformed was then isolated by adding protein A-Sepharose.cDNA Cloning and Sequencing. A rabbit liver cDNA library

in phage Agtll (Clontech) was screened according to themethod of Snyder et al. (19) using monospecific rabbitanti-eIF-5 antibodies (15) as probes. Putative positiveplaques were isolated and rescreened at lower plaque den-sities until plaque purified. A single immunoreactive recom-binant phage containing a 1.135-kb cDNA insert was ob-tained. A 565-bp Pvu II fragment of this cDNA insert waslabeled with 32P by random-priming using [a-32P]dCTP(Boehringer Mannheim kit) and used as a hybridization probeto screen a size-fractionated rat insulinoma plasmid cDNAlibrary (20), as described by Sambrook et al. (21). Positiveclones were plaque-purified, and the BamHI inserts weresubcloned into pGEM3Zf+ (Promega) for further analysis. A3.55-kb DNA fragnent from the largest clone was subjectedto nested deletions in both directions, using the exonucleaseIII/S1 nuclease method (Erase-a-base, Promega). Plasmidclones with overlapping deletions were sequenced by thedideoxy chain-termination method (22) using a United States

Abbreviations: elF, eukaryotic (translation) initiation factor; Met-t-RNAf, initiator methionyl tRNA; IPTG, isopropyl ,-D-thiogalacto-pyranoside; ORF, open reading frame.*The sequence reported in this paper has been deposited in theGenBank data base (accession no. L11651).

3058

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.

Biochemistry: Das et al.

Biochemical sequencing kit. Regions not covered by thedeletion clones were sequenced using specific oligonucleo-tide primers based on newly acquired sequences. Sequenceanalysis was carried out using the University of WisconsinGenetics Computer Group program (23).

In Vitro Transcription and Translation. The HindIII frag-ment of the 3.55-kb eIF-5-cDNA was subcloned into

A -226 CTCTGCTTTTCTCTTTCAGAGCTCTTGCGCAGCCATTGGTACCTGTATTCGGGCA-163 TACAAGCAAGAAGCTTACAGCCTCAGTGGCGAAAAATTTTTCATGTCAGAGACCI-100 TGCAGTCGTTTATGTCATCCCTTCTTCTCCAGACAGAAGATACCAAAAAGTTGCI-37 TCTGTTCATCTTATTGATAAAGTCACi 3 CCAAAATGTCTGTCAACGTCA

M S V N V N26 TGTCAGACCACTTCTATCGCTACAAGATGCCCCGTTTGATTGCTAAGGTTGAGG(

S D Q F Y R Y K M P R L I A K V E G89 ATGGAATCAAGACAGTTATAGTCAACATGGTTGACGTTGCAAAGGCGCTTAATC(

G I K T V I V N M V D V A K A L N R152 CGTATCCCACCAAATATTTTGGTTGTGAGCTGGGAGCACAGACCCAGTTTGATGI

Y P T K Y F G C E L G A Q T Q F D V215 ACCGTTACATTGTCAATGGATCTCATGAGGCGAATAAGCTGCAAGACATGTTGGi

R Y I V N G S H E A N K L 0 D M L D278 TTAAAAAATTTGTTCTCTGTCCTGAGTGTGAGAATCCTGAAACAGATCTGCATGI

K K F V L C P E C E N P E T D L N V341 AGAAGCAAACAATAGGTAATTCTTGTAAAGCCTGTCGGTACCGAGGCATGCTTGC

K Q T I G N S C K A C G Y R G M L D404 ATAAACTCTGTACATTCATTCTCAAAAACCCACCTGAGAATAGTGACATTGGTAZ

K L C T F I L K N P P E N S D I G T467 AAGA -AGAATA G .GGCAGGACGGAATCTCTGTAT(

E K E K K N R K G K D K E N G S V S530 AGACACCACCACCTCCACCACCAAATGAAATTAGTCCTCCACATGCTGTGGAAGJ

T P P P P P P N E I S P P H A V E E593 ATGATGATTGGGGGGAGGATACAACTGAGGAAGCTCAAAGGCGCAGAATGGATG)

D D W G E D T T E E A Q R R R M D E656 ACCATGCAAAAGGTCTGACACTTAGCGATGATTTGGAAAGAACTGTAGAAGAGCZ

H A K G L T L S D D L E R T V E E R719 TCCTGTTTGATTTTGTTAAGAAAAAGAAAGCGAGGGCATTATTGATTCATCTGI

L F D F V K K K K E E G I I D S S D782 TTGTGGCTGAGGCAGAAAGACTGGATGTAAAAGCCATGGGCCCTCTCGTTTTGAC

V A E A E R L D V K A M G P L V L T845 TCTTTGATGAGAAGATAAGAGAGCAAATCAAGAAATACAGGCGCCATTTTTTAAC

F D E K I R E Q I K K Y R R H F L R908 ATAACAACAAAAAGGCCCAGCGGTACCTTCTTCATGGTTTGGAATGTCTGGTAGC

N N K K A Q R Y L L H G L E C V V A971 AAGCTCAGTTGATCTCCAAGATTCCACATATCTTGAAGGAGATGTATGATGCAG;

A Q L I S K I P H I L K E M Y D A D1034 AGGAAGAGGTCATTATCAGCTGGTCAGAAAAGGCCTCTAAGAAATATCTCTCAAi

E E V I I S W S E K A S K K Y V S K1097 CCAAAGAGATTCGTGTCAAAGCAGAGCCATTTATTAAATGGTTGAAGGAAGCGGI

K E I R V K A E P F I K W L K E A E1160 CTTCTGGTGCTGAGGAAGAAGACGAAGATGAAAATATTGAGGTGGTATATTCGk

S G G E E E D ED E N I E V V Y S K1223 GTGTACCAAAAGTTGAAACTGTGAAGTCTGACAACAAGGATGACGACATTGATAI

V P K V E T V K S D N K D D D I D I1286 TTTAAAAGGATGGATGCAACTTAGCTTAACAGTGTAATGCTGCAAATTTTTCTC(

1349 CCAGAAGTGCAACATGTATGTGCAAGAGCTAAAGTGGCTTAACATCATGCTACA(1412 AAAAAGCTATTACTGTGAGTGGTCTATAATTAAGCCCAATGAGACATCTAGGGA(1475 TATCAGTGAGCAGTTGTAGTTTGCTTATTTATAGCATGTTTCTTTCCGAAAAAC1538 ACACATTTGGATCACATTTATACAGTTATAAAAa&A5MGATTTGATTTTGGTCi1601 ACTTTGGGCTATGAATGGCTTATGCTGAAGTAATTG-TACTTTTAGGATGTTA(1664 ATAACTTAGACTTCTTAAGTTTGGTAGATTGTTAGGTACTGAAGACTTGAAGAA1727 TTATAATGACCTTACTCAGCCATTAAGAAATGAAGTATTTTGAAAGTTGTGTCT(1790 TCAGATTGGCAACTGACAATTCTTGTCATTCTAAGGAAATTTGATGATTTAATG;1853 CATCCTCATGAGAAGTAAAAATGACCTGTGTGTCCTATGGTTTAAGAGCAAATTI1916 GGAGTTGTGGTTTTTCAGTTTGTGTACACTCACCCCAAATTGTAGTCTATTGAGI1979 TTGCACGTTGGATAAGCCACGGAAATGACAAATAAGTATTTGTGTGTATTTAGC2042 TGTACTGAGAGAAAAGCTTTGAGGTGTGATTAAATCGTAAACTCTGATTCTATTI2105 CAGGAAAAAC-GTCCACTTAATCTAAAACAGCATAAGTTTTCAACTTTTACCCTTI2168 TTTCAAGATGTTTAGACATACTGTATCTTGTGTTTGATGTGTTCCCCCTCCCTAJ2231 TTTATTCTTTAATGCCTTTTAATTTGGATATAATAGCTTGTACTTTAGATTTTGC2294 TTGCCAa&AACGTGTTACTGTTTTTCAAGCTTGATCCCCTTCCCTGATTGTCTI2357 AGAAACTTXTCATACTTCTGAGTCAGAGCCTGTAtTTTGGTTAAGACTTGGC2420 TACTTCACATTGAATATAGCTGGATACCTGAGAAGTCTGGTGATGGCACTGGGTC2483 TAGCTAAGGCCTGACCAGCCCATTCAGAGCCTTGCACTTCAGACACAAAACGTGAC2546 CCACTTGCTGGTGTAAACTCTATCTGCGGTCCTGACTATATTTCAATACTTGTCI2609 AAAAAACATAGCACATTTTTCTTTCTACAAAAGTACATTCTGGAGTTAAGAACCC2672 GATTTGTGTGTGGCGTGCTAGCTCATACATTATTTGGATCTTATTCTTTGTGTCI2735 ACAGATTATAAGACTTTGATTAATGTAAAAACTATGCGTTAAAATCATACCAAAC2798 AAATTAAAACCTTGATGGGAGGCTGGGCGTGGAACAGGAGCCATATACCTGGAAM2861 GGGAATGTGCTATGTCACACCAAAGAAGTGGGACTTGGAAAGTCACTTGTCTCC12924 GACTCTTTGTTGCATGGGCAGCCCATCCATATGTCATTACTTTTTGAGATTCTCJ2987 AGCACATTTCGGCCCTCAGGTTGGCAAGATTTTGTCTTAGAGCTGTTGCTTTAAAC3050 TCAGGTCTTAGACACTTAGGAAGGTCTTGGGCTTCTGTTCATTCTGGTGCCAAAC3113 TTCAAATTTCACACAATCTGCGTTTTTATTCATGGAGGTTAACCTGGTAAGAGTA3176 TGGCTCTATTGAGGTGTCTTAAAAACTTTCCTGTTTCAAACAGCTACATTACTTC3239 AATGTTAT TTAAATTTCCCCCTCCTTTCATATTAAAAAAAAAAAAAAAAAA3302

B

Hindlll

aa \-. b N

Pvull Hpal

Sphl Apal

Proc. Natl. Acad. Sci. USA 90 (1993) 3059

pGEM7Zf+ vector (Promega) and the recombinant plasmidwas isolated and purified by CsCl/ethidium bromide centrif-ugation. The linearized plasmid was used as a template for invitro transcription by T7 RNA polymerase in the presence ofcap analog to produce capped transcripts. The RNA wasextracted with phenol/chloroform and ethanol precipitated.Approximately 2.5 ,ug of the RNA was translated in vitro by

RACATAGCA -164GAGAACTCT -101AATCAAAGA -38ACCGCAGCG 25

R S VGCAAAGGAA 88K G N

GGCCTCCAA 151P P T

TTAAGAATG 214K N D

ATGGATTCA 277G F I

TCAATCCAA 340N P K

ACACACATC 403T_ H H

CAGGAAAGA 466G K K

CCACCAGTG 529T S E

AAGAGGAAG 592E E D

AAATCAGTG 655I S D

GTGTTAACA 718V N I

ATAAAGACA 781K D I

CAGAAGTTC 844E V L

GATTTTGTC 907F C H

CAATGCATC 970M H Q

ACCTTTTAG 1033L E

MkA9XXCT'Tr 109 6E L A

AGGAGGAAT 1159E E S

AGACTGCCA 1222T A S

TTGATGCCA 1285D A I

CATTATCAG 1348

141114 7 415 3 7160016631 726178918521915197 8204 1210421672230229323562419248225 4 52608267 127 3 42797286029 2 32986304931123175323833 01

Partial Rabbit elF-5 clone

Hindil Hindll

IHpal Pvulil

500 bp

Hpa!

Complete Rat elF-5 clone

FIG. 1. Characterization of eIF-5 cDNA.(A) Nucleotide sequence and predictedamino acid sequence (single-letter aminoacid code) of rat cDNA encoding eIF-5. TheATG assigned as the translation start codon(numbered + 1) was found in frame with all ofthe partial amino acid sequences of CNBrpeptides of eIF-5 (underlined) and is pre-ceded by two consecutive in-frame transla-tion stop codons that are boxed. There aretwo other ATGs in the 5' untranslated re-gion, all of which terminate after a shortstretch of reading frame. The translationstop codon (nucleotide + 1288) is shown withan asterisk and the potential poly(A) signalsequences are doubly underlined. The arrow(nucleotide +2328) indicates the positionwhere a poly(A) tail was found in anothercomplete eIF-5-cDNA clone. The partialrabbit liver clone covers the nucleotide se-quence from +487 to + 1622 of the rat clone.In this overlapping region, the rat and therabbit clones are 90%o identical in nucleotidesequence and 96% identical in amino acidsequence. (For example, in peptide se-quence ii of rabbit eIF-5, valine at position 9is replaced by a glycine in rat eIF-5.) (B)Restriction map of eIF-5 clones from rabbitliver and rat pancreatic cDNA libraries.Shaded boxes indicate the coding region; aand b represent 565-bp Pvu II and 515-bp PvuII-Hpa I restriction fragments, respectively;c represents a 25-bp oligonucleotide (posi-tion +28 to +53) of the partial rabbit liverclone.

Proc. Natl. Acad. Sci. USA 90 (1993)

using rabbit reticulocyte lysate (Promega) for 1 hr at 30°C inthe presence of 20 pCi of [35S]methionine (1 Ci = 37 GBq).

Expression of elF-5 in Escherichia coli. Two primers weresynthesized bearing the N-terminal and C-terminal ends ofeIF-5-coding sequences that were flanked by Nde I andEcoRI sites:

N terminus 5'-dCCGGGATCCATATGTCTGTCAACGTCAACC-3'C terminus 5'-dCGCGAATTCCTGTTAAGCTAAGTTGCATCC-3'

These primers were used in PCRs using Pyrococcus DNApolymerase (Stratagene) to generate a product containing thecoding sequence of eIF-5 flanked by Nde I and EcoRI sites.The PCR product was digested with Nde I and EcoRI andcloned into Nde I and EcoRI sites of PET-5a plasmid (24)(Novagen). E. coli BL21 (DE3) (Novagen) containing recom-binant or the parental plasmid was grown in YT medium (21)containing ampicillin (50 ytg/ml) at 37°C. Cells were inducedwith 1 mM isopropyl 3-D-thiogalactopyranoside (IPTG)when A6 reached 0.6. After 3 hr, the cells were collected bycentrifugation and suspended in one-fifth of the culturevolume in a buffer containing 25 mM NaCl, 20 mM Tris HCl(pH 8.0), 5 mM EDTA, 10 mM 2-mercaptoethanol, and 0.5mM phenylmethylsulfonyl fluoride. Cells were broken byfreezing and thawing followed by sonication for 1 min on ice,and centrifuged at 15,000 x g for 10 min; the supernatant wasfrozen in small aliquots in dry ice/ethanol bath and stored at-70°C until used.

RESULTSCloning and Sequence Analysis ofeIF-5 cDNA. Partial amino

acid sequences of rabbit reticulocyte eIF-5 were determinedfrom three CNBr peptides isolated as described in Materialsand Methods. These sequences, in the single-letter amino acidcode, are (i) LDTHHKLCTFILKNPPENSD; (ii) DEISD-HAKVLTLSDDLERTVEERVNILFDF; and (iii) XXA-DLLEEEVIIX, where residue X indicates that no positiveidentification could be made. Simultaneously, monospecific

1 2

kb

- 9.49- 7.46

rabbit anti-eIF-5 antibodies (15) were used as a probe toimmunoscreen a Agtll phage cDNA expression library ofrabbit liver (Clontech). Of 2 x 106 recombinant phagesscreened, we isolated 1 immunoreactive clone with a 1.135-kbinsert. Sequencing ofthe subcloned fragment revealed that thecDNA insert contained peptide sequences ii and iii derivedfrom purified eIF-5 in one reading frame. A 515-bp restrictionfragment (Fig. 1B, fragment b) was then used to performNorthern analysis of total poly(A)+ RNA isolated from ratliver and HeLa cells. Three distinct transcripts of 3.5 kb, 2.8kb, and 2.2 kb were detected with rat liver, whereas HeLa celscontained a 3.5-kb and a 2.2-kb transcript (Fig. 2). A 565-bpPvu II fragment of the rabbit liver clone (Fig. 1B, fragment a)was then used to screen a rat insulinomacDNA plasmid library(20). Seven independent positive clones were isolated that alsohybridized with 5' 32P end-labeled 25-mer oligonucleotidederived from the 5'-most end of the partial rabbit liver clone(Fig. 1B, fragment c). Two clones containing the longestcDNA inserts (3.55 kb and 2.8 kb) were further analyzed. The3.55-kb insert was sequenced completely on both strands (Fig.1A). The longest open reading frame (ORF) encodes a proteinof 429 amino acids (Mr 48,926). This value is close to theapparent molecular weight of purified eIF-5 (Mr = 58,000)isolated from calf liver (13) and rabbit reticulocyte lysates (14).Sequences encoding all three CNBr-derived eIF-5 peptideswere found in this ORF. Furthermore, the ORF is preceded byseveral in-frame translation stop codons, indicating that this isan independent reading frame and not part of a larger poly-peptide.

Analysis of the 2.8-kb cDNA insert present in the otherpositive clone indicated that it had an identical 5' untrans-lated and coding sequence as the 3.55-kb insert but differedfrom the latter at the 3' untranslated region starting atnucleotide position +2328 from which the poly(A) tail was

A B a1 2 3 4 5

kDa

100. 6-71.2

43. 5--i.

28. 6

18.3-

1 2 3

*:..::*.

*- sss HOu. WY

.. .:.. , s.#ws . ? " ? ee :... l # ? .. <*? |.. B a | . .W.* ' *.. . . .. : ..

ssrxpts

1 2 3 4kDa

- 100.6- 71.2

- 43.5

-28.6

-18.3

3.5 -

2.8 -

2.2 - W 2.37

- 1.35

FIG. 2. Northern blot analysis of eIF-5 mRNA. Poly(A)+ RNAwas isolated from HeLa cells and rat liver using the acid guanidiniumthiocyanate/phenol/chloroform method (25) followed by oligo(dT)-cellulose column chromatography (21). Each RNA sample (20 ,ug)was incubated with glyoxal at 65°C for 1 hr, resolved on a 1.0%oagarose gel containing 2 M formaldehyde, blotted to a nylon mem-brane (Schleicher & Schuell), and baked (21). Following prehybrid-ization as described (21), the blot was hybridized in the same solutioncontaining 10% dextran sulfate at 42°C for 20 hr with a 515-bp PvuII-Hpa I fragment of the rabbit liver clone (Fig. 1B, fragment b),labeled with [a-32P]dCTP (6000 Ci/mmol) by random-priming. Theblot was washed three times in 0.1 x SSC/0.1% SDS (1 x SSC = 0.15M NaCl/15 mM sodium citrate, pH 7.0) at 68°C for 1 hr and exposedto an x-ray film for 3 days in the presence of an intensifying screenat -70°C. Positions of the RNA size markers (BRL) are indicated inkb.

FIG. 3. Expression of the eIF-5 cDNA clone. (A) SDS/PAGE(15% gel) of [35S]methionine-labeled eIF-5 produced in a reticulocytelysate. Lanes 1 and 2, in vitro translation in the absence ofRNA (lane1) or cloned eIF-5-derived mRNA (lane 2). 35S-labeled translationproducts formed in reticulocyte lysates were immunoprecipitatedwith protein A-Sepharose and rabbit anti-chicken IgY in the presenceof preimmune chicken IgY antibodies (lane 3) and in the presence ofchicken anti-eIF-5 antibodies (lane 4). Immunocomplexes formedwere washed as described by Harlow and Lane (28), boiled in SDSgel loading buffer (63 mM Tris-HCl, pH 6.8/2% SDS/140 mM2-mercaptoethanol/10o glycerol/0.04% bromophenol blue), andthen analyzed by gel electrophoresis. Purified rabbit reticulocyteeIF-5, labeled by incubation with [y-32P]ATP and purified caseinkinase II (15), was run in a parallel lane of the same gel as a marker(lane 5). Following electrophoresis, the dried gel was analyzed byautoradiography. (B) Expression of eIF-5 cDNA in E. coli. Cell-freeextracts from E. coli BL21 (DE3) transformed with either parentalplasmid pET-5a (lane 1) or recombinant plasmid pET-5a (lanes 2 and3) were subjected to SDS/PAGE (10% polyacrylamide) followedeither by Coomassie blue staining (panel a) or by immunoblotanalysis using monospecific rabbit anti-eIF-5 antibodies (panel b).Lanes 1, extracts of cells containing parental plasmid pET-5a; lanes2, extracts of uninduced cells containing recombinant pET-5a; lanes3, extracts of IPTG-induced cells containing recombinant pET-5a.Purified mammalian eIF-5 was electrophoresed at the same time andprobed with anti-eIF-5 antibodies (lane 4 of panel b).

3060 Biochemistry: Das et al.

Proc. Natl. Acad. Sci. USA 90 (1993) 3061

0

x0n._

'I

u

U--i

09.-

x

6 EULu)

0

12 1 6 20 24 28 32 36 40 44Fraction number

FIG. 4. Coelution of in vitro-expressed 35S-labeled translationproduct with eIF-5 isolated from rabbit reticulocyte lysates. [35S]Me-thionine-labeled translation products produced in reticulocyte lysatetranslation system (40 1,u containing about 1 x 106 cpm of proteinproducts as described in the legend to Fig. 3) were adjusted to 0.5 MKCI and centrifuged in a TLA-100.2 fixed-angle rotor for 1 hr at75,000 rpm, and the supernatant was mixed with 23 mg of 0.5 MKCl-wash proteins of rabbit reticulocyte polysomal pellets, obtainedas described (14). The total eIF-5 activity present in this mixture waspartially purified by successive step-wise elution from DEAE (10-mlbed volume) and phosphocellulose (2-ml bed volume) columns asdescribed (isolation of "partially purified eIF-2/eIF-5 preparation"(14). The phosphocellulose eluate containing all the eIF-5 activitywas then subjected to gradient elution (20 mM Tris HCl, pH 7.5/100mM KCl to 20 mM Tris HCl, pH 7.5/500 mM KCI) from a 1-ml bedvolume of a fast protein liquid chromatography-Mono Q column(Pharmacia). (All buffer solutions also contained 1 mM dithiothreitol,0.2 mM EDTA, and 10%o glycerol.) Aliquots from each gradientfraction were assayed for 35S radioactivity and eIF-5 activity. Thethin line represents absorbance at 280 nm.

added (Fig. 1A, arrow). It should be noted that the 3.55-kbcDNA contained multiple potential poly(A) sites (Fig. 1A).These results therefore suggest that some of the eIF-5 tran-scripts-e.g., 2.8-kb RNA-observed in RNA blot analysismay arise by alternative polyadenylylation at the 3' end ofeIF-5 mRNAs. It is also interesting to note that the 5'noncoding region of the 3.55-kb cDNA contains an excellent3' splice site, TCCCTTCTTCTCCAG, preceding the pre-dicted initiation ATG codon at + 1. Thus the possibility alsoexits that some of the shorter eIF-5 transcripts-e.g., 2.2-kbRNA-are derived by efficient splicing of eIF-5 mRNAs atthis site. Many vertebrate cDNA sequences that have up-stream ATG codons contain an unspliced 5' intron (26).The predicted amino acid sequence of eIF-5 is not closely

related to any other known proteins in the GenBank andEMBL data bases. However, a portion of the antisensesequence of eIF-5 was found to be identical to a 1017-bp ratbrain cDNA clone (27) ofunknown function. This cDNA wasisolated on the basis of its ability to hybridize with a 82-nt-long brain-specific RNA, termed "identifier sequence" (27).The significance of this homology is not apparent.

In Vitro Translation ofeIF-5 cDNA. To demonstrate that thecloned cDNA encodes eIF-5, the properties of the expressedproteins were tested (Fig. 3). A polypeptide of apparent Mr= 58,000, which showed the same mobility as purifiedmammalian eIF-5, was generated by in vitro translation ofcDNA-derived transcripts in a rabbit reticulocyte translationsystem (Fig. 3A, lane 2). Furthermore, the 35S-labeled trans-lation product was specifically immunoprecipitated withchicken anti-eIF-5 antibodies but not with chicken preim-mune antibodies (Fig. 3A, compare lanes 3 and 4). Additionalevidence that the 58-kDa protein was eIF-5 came from theobservation that when 35S-labeled translation products weremixed with a crude eIF-5 preparation from rabbit reticulocytelysates and subjected to purification of eIF-5 activity throughsuccessive DEAE and phosphocellulose chromatographic

.1bU

Cu

1.0

0.

0.

10 20Bottom TopFraction Number

FIG. 5. Formation of the 80S initiation complex by bacteriallyexpressed recombinant eIF-5. eIF-5 activity was measured by itsability to promote the joining of 60S ribosomal subunits to the 40Sinitiation complex containing bound [35S]Met-tRNAf to generate an80S initiation complex as described (13, 14). Reaction mixtures (45,ul each), containing 20 mM Tris HCl (pH 7.5), 100 mM KCI, 5 mM2-mercaptoethanol, 20 ,ug of bovine serum albumin, 0.3 mM GTP, 9pmol of [35S]Met-tRNAf (20,000 cpm/pmol), and 1 pg of purifiedeIF-2, were incubated at 37°C for 3 min to form the [35S]Met-tRNAreIF-2-GTP complex. Each reaction mixture was then adjustedto 5 mM MgCl2 and then supplemented with 0.5 A260 unit of 40Sribosomal subunits and 0.1 A260 unit of AUG; the mixtures wereincubated for 4 min at 37°C to form the 40S initiation complex(40S AUG [35S]Met-tRNAfeIF-2-GTP). Reaction mixtures werethen chilled in an ice bath and 0.8 A260 unit of60S ribosomal subunitsand 50 ng of crude E. coli cell extract (source of eIF-5) were addedas indicated. (a) No addition. (b) 60S subunits. (c) 60S ribosomalsubunits plus extract of E. coli BL21 (DE3) cells containing parentalplasmid pET-5a. (d) 60S ribosomal subunits plus extract of E. coliBL21 (DE3) cells containing recombinant plasmid pET-5a that wereinduced with IPTG. Following incubation at 37°C for 5 min, forma-tion of the 80S initiation complex containing bound [35S]Met-tRNAfwas measured by centrifugation in a 5-25% sucrose gradient cen-trifugation as described (13, 14).

steps followed by gradient elution from a fast protein liquidchromatography-Mono Q column, the 35S radioactivity co-eluted with eIF-S activity from the Mono Q column (Fig. 4).

Expression of eIF-5 in E. coli. To demonstrate the functionof the 58-kDa protein, the cDNA was placed under thecontrol of the phage T7 RNA polymerase promoter. Trans-formation of the resulting construct into E. coli BL21 (DE3)resulted, after induction with IPTG, in the production of apolypeptide of 58 kDa that migrated with the same mobilityas purified mammalian eIF-5 (Fig. 3B, panel a; compare lanes1 and 3). The 58-kDa polypeptide reacted strongly withanti-eIF-5 antibodies (Fig. 3B, panel b, lane 3). In the absenceof IPTG induction, only a low level ofthe 58-kDa polypeptidewas synthesized (Fig. 3B, panels a and b, lanes 2). Thesynthesis of this 58-kDa polypeptide was specific for eIF-5cDNA sequences, since an E. coli strain containing a plasmidlacking the cDNA failed to produce any immunoreactive58-kDa polypeptide (Fig. 3B, panels a and b, lanes 1). Mostsignificantly, the bacterially expressed recombinant eIF-5was highly active in promoting the in vitro formation ofan 80Sinitiation complex from a preformed 40S initiation complex(Fig. 5, compare d with a and b). Bacterial extracts derived

Biochemistry: Das et al.

Proc. Natl. Acad. Sci. USA 90 (1993)

Table 1. Sequence homology of mammalian eIF-5 with proteins of the GTPase superfamilyMotif G-1 G-2 G-3 G4

Consensus GXXGXGK(S/T) D-(X)n-T DXXG (N/T)(K/Q)XDMammalian

eIF-5 27GKGNGIKT34 soD-(X)n-T107 19DDWG2o2 419NKDD422

Comparison of the putative G1-G4 GTP-binding domains in the eIF-5 amino acid sequence with theconserved sequence motifs in GTPase superfamily (30). The single-letter amino acid code is used. Thenumbers preceding the sequence motifs in eIF-5 represent the position of the first amino acid in eachbinding motif. Ref. 30 contains a compilation of proteins belonging to the GTPase superfamily.

from cells lacking eIF-5 cDNA did not catalyze the formationof an 80S initiation complex (Fig. Sc). These results demon-strate that the isolated clone encodes a functional eIF-5 andthat the relative molecular mass of mammalian eIF-5 is48,926, which migrates on SDS gels with an apparent Mr of58,000. Data presented in Fig. S also allowed us to calculatethat the specific activity ofrecombinant eIF-5 in crude E. coliextracts was about 40,000 units/mg of protein. (One unit ofeIF-5 was defined as the amount of protein that promoted theformation of 1 pmol of 80S initiation complex.) Based on acomparable AUG-directed 80S initiation complex formationassay, this value is considerably higher than that reported forrabbit reticulocyte eIF-5 in earlier studies (9-12).

DISCUSSIONSeveral lines of evidence presented in this paper indicate thatwe have cloned the rat cDNA encoding mammalian transla-tion initiation factor, eIF-5. The coding region of the cDNAhas been expressed in E. coli to yield a catalytically activeeIF-5 protein (calculated Mr 48,926) that migrated on SDSgels with the same mobility as purified eIF-5 isolated fromrabbit reticulocyte lysates (apparent Mr 58,000). These re-sults confirm our original observations (13-15) that mamma-lian eIF-5 is a protein of 58 kDa and did not arise byproteolysis ofa 125- to 168-kDa protein during isolation of theinitiation factor from cell extracts. These experiments furthersuggest that earlier reports from other laboratories (5-8) thateIF-5 is a protein of 125-168 kDa probably reflected thepresence of the highly catalytic 58-kDa eIF-5 protein as acontaminant among other proteins unrelated to eIF-5. Thepossibility, however, exists that eIF-5 may exist as isoformsof different molecular masses (125-168 kDa and 58 kDa) ashas been observed with wheat germ eIF-4F (29).

Recently, the Saccharomyces cerevisiae gene that encodeseIF-5 has also been isolated and expressed in E. coli to yielda catalytically active eIF-5 protein (31). This single copy genecontains an intron-free ORF that encodes a protein of cal-culated Mr 45,346. Comparison of the derived amino acidsequence of the yeast and the mammalian eIF-S proteinsshows 60o homology and 39% identity.

Analysis ofthe derived amino acid sequence ofmammalianeIF-5 indicated several interesting features. eIF-5 is an acidicprotein of calculated pI 5.2. The amino acid sequence con-tains multiple potential phosphorylation sites for casein ki-nase II, protein kinase C, and cAMP-dependent proteinkinase A. It remains to be seen whether eIF-5, like manyother translation initiation factors, is phosphorylated in vivoand whether phosphorylation of eIF-5 is involved in regula-tion of its activity. eIF-5 is known to be phosphorylated invitro by casein kinase II (15).

Direct comparison of the eIF-5 amino acid sequence withthose of the GTPase superfamily (30) revealed regions withsequence similarities (Table 1). Members of the GTPasesuperfamily (30) include prokaryotic translation initiationfactor IF-2, elongation factors EF-Tu, EF-G, and EF-la; thea-subunit of the G proteins, Ha-RAS, yeast CDC 42 and SEC4 proteins, and a host of other GTPase proteins. Theseproteins have characteristic sequence motifs in four distinctdomains, designated G-1-G-4, as shown in Table 1. The motif

for a phosphate binding loop, GXXGXGK(S/T), is present ineIF-5 as GKGNGIKT. Insertion of an extra amino acid intothe consensus loop sequence suggests that eIF-5, by itself, isnot functional as a GTPase. However, such sequence simi-larities between mammalian eIF-5 and proteins ofthe GTPasesuperfamily correlate well with the known GTP hydrolysisfunction of eIF-5 in the initiation of protein synthesis (5-8).We are particularly indebted to Dr. Jack Dixon of the University

of Michigan for providing us with rat insulinoma cDNA library. Thisstudy was supported by Grant GM15399 from the National Institutesof Health and by Cancer Core Program Grant P30CA13330 from theNational Cancer Institute. Deoxyoligonucleotides were synthesizedby a core facility of the Albert Einstein College of Medicine and themicrosequencing of peptides was carried out in the laboratory ofMacromolecular Analysis of this institution.

1. Maitra, U., Stringer, E. A. & Chaudhuri, A. (1982)Annu. Rev. Biochem.51, 869-900.

2. Kozak, M. (1983) Microbiol. Rev. 47, 1-45.3. Hershey, J. W. B. (1991) Annu. Rev. Biochem. 60, 717-755.4. Merrick, W. C. (1992) Microbiol. Rev. 56, 291-315.5. Trachsel, H. & Staehelin, T. (1978) Proc. Natl. Acad. Sci. USA 75,

204-208.6. Peterson, D. T., Safer, B. & Merrick, W. C. (1979) J. Biol. Chem. 254,

7730-7735.7. Raychaudhuri, P., Chaudhuri, A. & Maitra, U. (1985) J. Biol. Chem. 260,

2140-2145.8. Chakrabarti, A. & Maitra, U. (1991) J. Biol. Chem. 266, 14039-14045.9. Merrick, W. C., Kemper, W. M. & Anderson, W. F. (1975) J. Biol.

Chem. 250, 5556-5562.10. Schreier, M. H., Emi, B. & Staehelin, T. (1977) J. Mol. Biol. 116,

727-753.11. Benne, R., Brown-Leudi, M. L. & Hershey, J. W. B. (1978) J. Biol.

Chem. 253, 3070-3077.12. Meyer, L. J., Brown-Leudi, M. L., Corbett, S., Tolan, D. R. & Hershey,

J. W. B. (1981) J. Biol. Chem. 256, 351-356.13. Raychaudhuri, P., Chaudhuri, A. & Maitra, U. (1985)J. Biol. Chem. 260,

2132-2139.14. Raychaudhuri, P., Chevesich, J., Ghosh, S. & Maitra, U. (1987) J. Biol.

Chem. 262, 14222-14227.15. Ghosh, S., Chevesich, J. & Maitra, U. (1989) J. Biol. Chem. 264,

5134-5140.16. Charbonneau, H. (1989) in A Practical Guide to Protein and Peptide

Purificationfor Microsequencing, ed. Matsudaira, P. T. (Academic, SanDiego), pp. 15-30.

17. Schigger, H. & von Jagow, G. (1987) Anal. Biochem. 166, 368-379.18. Gassmann, M., Thommes, P., Weiser, T. & Hfbscher, U. (1990) FASEB

J. 4, 2528-2532.19. Snyder, M., Elledge, S., Sweetser, D., Young, R. A. & Davis, R. W.

(1987) Methods Enzymol. 154, 107-128.20. Rodriguez, C., Brayton, K. A., Brownstein, M. & Dixon, J. E. (1989) J.

Biol. Chem. 264, 5988-5995.21. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning:A

Laboratory Manual (Cold Spring Harbor Lab., Plainview, NY).22. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci.

USA 74, 5463-5467.23. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids Res. 12,

387-395.24. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Dubendorff, J. W.

(1990) Methods Enzymol. 185, 60-89.25. Chomczynski, P. & Sacchi, N. (1987) Anal. Biochem. 162, 156-159.26. Kozak, M. (1991) J. Cell Biol. 115, 887-903.27. Milner, R. J., Bloom, F. E., Lai, C., Lemer, R. A. & Sutcliffe, J. G.

(1984) Proc. Natl. Acad. Sci. USA 81, 713-717.28. Harlow, E. & Lane, D. (1988) Antibodies. A Laboratory Manual (Cold

Spring Harbor Lab., Plainview, NY).29. Browning, K. S., Lax, S. R. & Ravel, J. M. (1987) J. Biol. Chem. 262,

11228-11232.30. Bourne, H. R., Sanders, D. A. & McCormick, F. (1991) Nature (London)

349, 117-127.31. Chakravarti, D. & Maitra, U. (1993) J. Biol. Chem., in press.

3062 Biochemistry: Das et al.