Download pdf - Sequence and identification of the nucleotide binding site for the

Volume 15 Number 22 1987 Nucleic Acids Research

Sequence and identiflcation of the nucieotide binding site for the elongation factor Tu fromThermits thermophtius HB8

Lothar Seidler, Marcus Peter, Franz Meissner and Mathias Sprinzl

Laboratorium fur Biochemie and Institut fur makromolekulare Forschung, UniversitSt Bayreuth,Postfach 10 12 51, D-8580 Bayreuth, FRG

Received August 20, 1987; Revised and Accepted October 12, 1987

ABSTRACTTwo structural genes for the Thermus thermophilus elongation

factor Tu (tuf) were identified by cross-hybridization with thetufA gene from E. coli. The sequence of one of these tuf genes,localized on a 6.6 kb Bam HI fragment, was determined and con-firmed by partial protein sequencing of an authentic elongationfactor Tu from T.thermophilus HB8. Expression of this tuf gene inE. coli minicells provided a low amount of immuno-precipitablethermophilic EF-Tu. Affinity labeling of the T.thermophilus EF-Tuand sequence comparison with homologous proteins from other orga-nisms were used to identify the guanosine-nucleotide bindingdomain.

INTRODUCTION

Elongation factor Tu (EF-Tu) from E .coli mediates the binding

of aminoacyl-tRNA to programmed ribosomes (1) and plays a role in

maintaining the fidelity of translation (2) . Amino acid sequen-

ces were determined for the procaryotic elongation factors from

E. coli (3, 4), yeast mitochondrium (5) chloroplasts of Euqlena

qracilis (6) as well as for several eucaryotic elongation factor

I a species (7, 8, 9). EF-Tu possesses some sequence homologies

with eucaryotic elongation factor la and G-proteins mainly in the

guanosin-nucleotide binding domain (10).

EF-Tu from T.thermophilus HB8 is temperature insensitive up

to 65 °C (11) in contrast to EF-Tu from E.coli. It forms stable

aa-tRNA»EF-Tu>GTP ternary complexes and is well suited for physi-

cal and biochemical investigations. We used this thermophilic EF-

Tu to construct affinity columns for the isolation of aminoacyl-

tRNAs (12) and for the investigation of aminoacyl-tRNA • EF-Tu*GTP

interactions (13). In order to prepare sufficient quantities of

the thermophilic elongation factor and its mutants for physical

© IR L Press Limited, Oxford, England. 9 2 6 3

Downloaded from https://academic.oup.com/nar/article-abstract/15/22/9263/1268034by gueston 17 February 2018

Nucleic Acids Research

studies the expression of its genes in E.coli was attempted. In

this communication we present the sequence of one of the two

T.thermophilus EF-Tu genes and the identification, by affinity

labeling, of the nucleotide binding site in the protein.

MATERIALS AND METHODS

Enzymes, deoxynucleoside-5'-triphosphates and dideoxynucleo-

sides-5'-triphosphates were obtained from Pharmacia (Uppsala,

Sweden), BRL (Eggenstein, FRG) and Boehringer (Mannheim, FRG).

[3H]GDP (11.3 Ci/mmole), [U14C]GTP (500 mCi/mmole), 35S-labeled

methionine (1275 Ci/mmole) and 'Amplify' were purchased from

Amersham-Buchler (Braunschweig, FRG). [ P ]orthophosphate

and P-labeled dATP (800 Ci/ mmole) were obtained from Du

Pont/New England Nuclear (Bad Nauheim, FRG). [G32P]GDP (1000

Ci/mmole) was prepared as described elsewhere (14). CM-Sepharose

CL-6B and Sephadex LH-60 were from Pharmacia. NalO., NaBH3(CN)

and NaBH. were from Serva (Heidelberg, FRG). Acrylamide and N,N'-

methylenebisacrylamide were obtained from BRL. All other reagents

were of analytical grade and purchased from Merck (Darmstadt,

FRG).

The plasmid pPR 1 carrying the 1.0 kbp Nrul/Hpal fragment of

the E.coli tufA gene was obtained from Dr. L. Bosch (Leiden).

E.coli strains RR1 and Kill were U8ed for cloning experiments.

The E.coli strain R312A which was used for minicell expression

was from Dr. N. Schumann (Bayreuth).

For hybridization experiments 20 p.g of T•thermophilus chromo-

somal DNA were digested with several restriction enzymes using

33 mM Tri8-acetate pH 7.9, 66 mM potassium acetate, 10 mM magne-

sium acetate, 0.5 mM DTT, 0.1 mg/ml BSA as incubation buffer, in

a final volume of 60 jil. Cleavage was complete after 3 h at 37 C

in the presence of 20 units of restriction enzyme. The resulting

DNA fragments were separated by horizontal gel electrophoresis

(1.3 % agarose in 40 mM Tris-acetate pH 7.9, 25 mM sodium ace-

tate, 0.5 mM EDTA) run at 3 V/cm for 4 h. Transfer of separated32DNA to a nitrocellulose filter and hybridization to the P-

labeled E.coli tufA fragment of pPR 1 were carried out as pub-

lished elsewhere (15, 16).

9264



For cloning the T.thermophilus chromosomal DNA, it was cleaved

and separated as described above. 2 jig of isolated T.thermophilus

DNA (Bam HI or Bgl II restriction fragments) was incubated with

1 jig BamHI cleaved, dephosphorylated pBR 322 DNA at 15°C, over-

night. The ligation was performed in 50 mH Tris-HCl pH 8.0, 10 mH

HgCl2, 20 mH DTT, with 10 units T4 DNA ligase in a total volume

of 30 u.1 . Transformation was carried out according to the method

of Inone and Curtiss (17). Isolated plasmid DNA of positive

transformants was blotted onto nitrocellulose and hybridized32

with P-labeled E.coli tufA probe.

The subcloned 1.6 kbp Smal fragment was sequenced in both

directions using the Sanger H13/dideoxy method (18).

Minicell preparation, S-protein labeling in vivo, and gel

electrophoresis of crude cell lysates were done as described

elsewhere (19, 20).

EF-Tu'GDP was purified from T.Thermopnilus cells, strain HB8,

as described by Leberman et al. (21). Nucleotide-free EF-Tu was

prepared as follows: Purified EF-Tu-GDP ( 30 mg in a volume of

2 ml) was diluted ten fold with 10 mM NH H P0 pH 5.65, 5 M

urea, 10 mM Q-mercaptoethanol and 10 \i» PMSF (buffer A). The pH

was adjusted to 5.65 with acetic acid . The solution was then

applied to an equilibrated (buffer A) CM-Sepharose CL-6B column

(2 x 10 cm), and further washed with 50 ml of the same buffer.

GDP-free EF-Tu was obtained by elution with buffer B (buffer A +

200 mM KC1, pH adjusted to 7.5). Urea was then removed by dialy-

sis against labeling buffer (50 mM Na-borate, pH 7.5. 50 mM KC1

and 10 mM MgCl_). The protein had a GDP-binding activity of^ 3

21,000 U/mg. The [ H]GDP activity was measured by the nitrocellu-

lose membrane filter binding assay (11).

Affinity labeling with periodate oxidized GTP: 20 (iM of nu-

cleotide- free EF-Tu was incubated with 10 jiM [U14C]GTP for 10 min

at 37 °C in labeling buffer and dialyzed twice for two hours at

4 C against the same buffer. Oxidation was performed in situ

with 1 mM NalO. for one min at 37 C. The reduction then followed

by treatment with 20 mM NaBH,(CN) for one min at 37 °C. The

reaction was stopped by addition af 25 mM NaBH., the mixture was

desalted on a NAP-Column (Pharmacia), and lyophilized.

9265



3 2Photoaffinity labeling with [6 P]GDP: 20 jiM nucleotide-free

EF-Tu was incubated with 1 u.M [ 632P]GDP for 10 rain at 37 °C in

50 mM Tris/HCl pH 8.0, 100 mH NaCl, 5 mH MgCL. and 1 mH Q-

mercaptoethanol . 1 mH ATP was added and the reaction mixture was

irradiated on ice with a laser beam at 257 nm using an argon ion

laser (model 2000, Spectra Physics, Mountain View, USA) with a

KDP crystal at an intensity of 5 mW/cm for 5 min. The sample was

desalted as described above.

Peptide analysis: A 500 jig sample of labeled EF-Tu was dis-

solved in 500 (il of 70 % formic acid containing 12.5 mg CNBr.

After 24 hrs in the dark the mixture was diluted ten fold with

water and lyophilized. Labeled peptides were analyzed by a modi-

fication of the slab gel system described by Swank and Munkres

(22); 12.5 % acrylamide, 1.0 % N,N'-methylenebisacrylamide, 5 M

urea, 0.1 % SDS, 0.1 H Na-phosphate, pH 6.8 with an upper gel (3

cm high) containing 8 % acrylamide, 0.064 % N,N'-methylenebis-

auiylaniide in the same buffer. The gel was atained with Coomaasie

blue, treated for 30 min with "Amplify", and dried. Labeled

peptides were identified by autoradiography using Kodak XAR-5

film.

Sequencing of cyanogen bromide fragments was done as follows:

22 mg of EF-Tu were incubated with 220 mg CNBr in 22 ml of 0.1 N

HC1 for 48 hrs in the dark. After dilution with 200 ml of H O and

lyophilisation the peptides were purified on a Sephadex LH-60

column (2 x 100 cm) according to Gerber et al. (23). Further

purification was performed on reverse phase HPLC column (Vydac C.

300 8, Macherey und Nagel, Duren, FRG). Sequencing of the large

CNBr-fragments was carried out with a liquid phase sequencer.

RESULTS

Chromosomal DNA from T.thermophilus HB8 was isolated and di-

gested with Bam HI or Bgl II restriction endonucleases. The frag-

ments were separated by eletrophoresis and hybridized to an

E.coli tuf A probe. Bam HI fragments of 6.6 and 4.5 kbp could be

hybridized to E .coli tufA-DNA while Bgl II fragments of 9.0 kbp

and 6.6 kbp provided a positive signal (fig. 1 ) . This indicates

the presence of two EF-Tu genes in T.thermophilus, a situation

similar to E.coli (24). The two 6.6 kbp restriction fragments

9266



kb

- 4.0

Fig. 1. Hybridization of bacterialchromosomal DNA against a ^ P-labeledE.coli tufA probe. Lane 1: E.coliDNA, EcoRI-cleaved; lane 2:T.thermophilus DNA, BamHI-cleaved;lane 3: T.thermophilus DNA,Bglll-cleaved.

were cloned into pBR 322 using the Bam HI site. Attempts to clone

the 4.5 kbp and 9.0 kbp fragments containing the putative second

T.thermophilus tuf gene were not successful. In fig. 2 the re-

striction map of the cloned DNA fragments is shown. The smallest

region providing a positive hybridization signal with E.coli tuf

A is located on a 1.6 kbp Sma I restriction fragment. This DNA

was sequenced using the M13/Sanger sequencing system. The entire

sequence is presented in fig. 3. The T.thermophilus tufl gene

codes for a protein of molecular weight 44600 D. It is 70 %

homologous to the E. coli EF-Tu on the protein level.

k

n

f 1 , 1J T 1 1

<I

>T f r

pLS652

PLS601

8 kbp

Fig. 2. Restriction map of the two cloned T.thermophilus DNA-fragments and their orientation in the genome. There are about 12Smal sites in the 6.6 kbp BamHI fragment, though only 6 sites areshown in the map. PLS 601 was identified as tufl by hybridizationunder stringent conditions with 1.3 kb BamHI/HindllI restrictionfragment obtained from pLS 652 (indicated by batching). Somecommon restriction site of pLS 652 and pLS 601 were identified asindicated. Restriction endonuclease cleavage sites: BamHI (•)Bglll (o), Hindll (•), Hindlll (•) , Kpnl (T ) , Pvul (V), Smal (x).

9267



<CCC>CGC«UCGCCAOTUTCCCCCCirrrCCTSCCme«eCiUTSTTCCTACeCCACeUCCTeC.CTCCUUCCUCCCCCCACCCTCCTTCCTUIH 1LII l_CUCTACUC^CC^CCCCUSUGCTCCA«iCAACCTUTCAASCCTCUTACCSCCCCTrCCSCCT6CCCl 1111 IA6CCCC

75194

Ihiln-Dllgirw.cccccuiu.ii U_UCUSSACCCAC

ill LT< Sly «"» Pht Til 1 0 I»r L*l Fro I_l Til U l Til Clr Thr II. Sly Bit Til U p III

cc« uc «c us m en c« i n uc ca cu m uc en o« ACC ATT ccc uc CTC uc uc

• • I • 30 40 50I I I Lli Tkr Ikr I n m U i 111 Us Tbr Trr Til U« U i Alt els UD Pro Un Til Sin Til Lrt Up lyr Clr Up l i t Up LrtSSS U S ACS ICC CTS ICC CCC CCC TTA ICC TIT STC CCS CCC CCC U C U C CCS U T STA C1C CTT UC CAC TIC CM SAC I T T GAC U S

(0 70 10U i Fro Gls Clu Arg H i Arg Cly l i t Tbr III i in Thr H i Bit Til e l l TTT Cl« Thr H i Lrt Arg lit Tyr ttr l i t Vil Up CytCCC CCS U C IUC CCT CCC CCC CCC I T T ICC ITC U C ICC CCC C1C CTC C1C TIC U O ACC CCC AAS CCS UC TIT TCC CIC CTC CAT TCC

______ll_______! |L_LJJO 'lOO 110

Pro Clr lit H i U p Tyr lit Lr« Atn Htt lit Thr Cly H i All clt Hit U p Sly H i lit L M Til Til ttr Alt Alt U p Sly ProCCT SCC CAC CCC GAC TAC ATC U C U C ATC ATC KC OCT CCC CCC CAO ATC U C CCC CCC ATC CTT STS CTC TCC CCC CCC SAC SCS CCS

-1 CB 7 CB 7 11 CB 11 | L _ _ _ _ _• • • • • • 120 130 14?

Hit Pro Cln Thr Arg Sis Bit l i t Ua I n H i lrg sin Til c ly Til Pro Tyr III Til Til Pht l i t U i Lyt Ttl U p Hit Til U pATC CCC U C ACC CCC U C U C ATT TTS CTS OCC CCC CAC CTC SSS STS CCS TAC ATT CTC CTS TTC ATC U C U C CTC CAC ATO CTO SAC

Cl 9 |.C1 i-50 H O 170

U p Pro 61a U s U s U p U a Til Clt Htt Cla Til Arg U p U a U a All Cln Tyr Clt Pbt PTO Clr U p Clt Til PTO Til lit ArgCAC CCC U G TTC CTC CAC CTC CTC CAS ATC U C CTC CCC U C CTT TTS AAC U C TAC U S TTT CCT SSS UC U S CTT CCC STS ATT CSS

110 • • • • • • • • • • • • • • • • • •Clr <tr 111 U i Lto H i U I Cla Cln Hit Bit lrg Un Pro Lrt Ttr Arg Arg Clr Cla U n Clt Trp Til U p Lrt l i t Trp Sis U sSCC ACT CCT CTT TTC CCC CTT U S U S ATC UC UC UC CCC AAC ACS ACC CCT CCC UC U C U S TSS STS UC UC ATT TCS U S CTS

• • • 210 220 230U t U p H i lit U p Sis Tyr lit Pro Thr Pro Til Arg U p Til U p Lyt Pro Pht U i Hit Pro Til Cls U p Til Pht Thr lit TbrTTC CAC CCC ATT CAC SAS TAC ATT CCC ACS CCC CTS CSS U C STS SAC U S CCC TTC TTC ATC CCC CTG U C U C CTS TTT ACC ATC ACS

240 250 2t0Cly Arg Cly Thr Til H i Thr Sly Arg lit Sis lrg Sly Lrt Vil Lrt Til Cly U p Cla Til Sis lit Til Cly Ua H i Pro Cla ThrOCT CCT CCC ACS STS CCC ACC CST CCS ATT CAC CCC CCC U C GTS U S STT SCS U C SAS CTC U C ATT CTS SSC CTT CCT CCC SAC 1CS

JTO " • • • • • • • • • I t _T 290lrg Arg Thr Til Til Thr Sly Til Cla Htt 111 lrg Lyt Thr U a Cln cla Glr lit H i Sir U p U n Til Clr Til U a U a Arg 01rCSC ACC ICC STS STS ACC SCT CTS U S ATC U C CCC U C ACC TTG U S SAS SSC ATT CCT CCC U C U T CTS CCC CTC CTC CTC CSC GST

300 310 320Til Itr irg Sin Slo Til Sit Arg Sir Sin Til U s All Lyt Pro Cly S«r lit Thr Pro III Thr Lrt Pht Cls H i Sir Vil Tyr VilCTC ACC CCC CAC G1C CTC U C CCC CCC CAC CTC CTC CCC AAC CCT SCO ACC ATT ICC CCC U C ACC AAG TTT GAC CCC TCS STO TAT GTS

330 340 350Ltt Lrt lyt Cls Clt Cly Clr Arg lit Thr Clr Pht Pht Sir Clr TTT Arg Pro Sin Pht Tyr Pht Arg Thr Thr U p Til Thi '.ly TilTTS U C U C U S SAC S«T SSA CCS CAC ACS CCC TTT TTT TCS CCC TAC CCT CCC CAC TTT TAC TTT CCS ACS ACS SAC STS ACS SCC STS

T i l S i t Lts Pro Pro S i r T i l Cls Btt T i l Htt Pro C_r Asp Aim T t l Thr Phi Thr T i l Cls U s I I I Lrt Pro T i l Clr Lta cTo ClaSTS CAS TTC CCT CCC CCC CTC UC ATC CTC ATC CCT CCC CAC U C CTC ACS TTT ACS STS U C CTC ATC U S CCS CTC CCC CTS SAC CAC

Sir Ltt Art Pht All l i t Arf Sla Clr Clr Arg Thr Til Clr All Clr Til Til Tbr Lyt l i t Ua Cla »**CCT TTC CSS TTT CCC ATC CCT U S CCT CCC CCC ICC CTC CSC CCC CCC CTC CTC ACC U C ATC a s SAS T U SCTUCCTATCCCCUCATCCCC M l !

ATCUCCTCCC6CI.111 IIUCUCAAUCCCTCUCCCCTCCCrtUCAAUTCCTCSACSCSSCCCSCCCTTCCCCCCCCCACCTCTCCCCCCCUTCCCCCTACCCACCC [CCCI 1577

Fig. 3. Sequence of the tufl gene of T.thermophilus. The cyanogenbromide fragments CB1-CB12 were identified by protein sequencing(•) and deduced from the positions of methionine residues. CB1and CB4 indicated by thick lines were labeled by GTP . andphotolabeled by GDP, respectively. The putative Shine-Dalgarnosequence is indicated.

9268



Table 1. Nucleotide composition of the codons used in theEF-Tu genes of T.thermophilus tufl (TT) and E. coli tufA

(EC) in percent.

Nucl.

GATC

totalTT EC

41.818.718.620.9

27.424.122.625.9

1stTT

44.325.110.120.5

letterEC

42.225.19.622.9

2ndTT

17.830.531.221.0

letterEC

17.030.730.322.0

3rdTT

63.30.514.521.2

letterEC

22.816.527.932.8

The nucleotide composition in the codons of the thermophilic

tufl gene is summarized in table 1 and compared to the tufA gene

of E.coli. As expected, the G+C content in the thermophilic gene

is higher than that in E. coli. This difference is much more

pronounced in the third letter of the codons. Futhermore, in the

case of the T.thermophilua gene G is an especially prefered

nucleotide in this position. The structural gene is preceded by a

purine-rich, putative ribosomal binding sequence which continues

up to the ATG protein initiation site. The codon usage in T. ther-

mophilus is considerably different from that in E.coli (table 2).

Table 2. Comparison of the codon usage in the T. thermophilus tufl(TT) and E. coli tufA (EC) genes.

TTEC

TTEC

TTEC

TTEC

GlyGGG

241

ArgAGG

3

TrpTGG

21

CGG

19

*GGA

1

AGA

—

0

TGA

1

*

CGA

-

GGT

819

SerAGT

1

CysTGT

1

*

CGT

521

GGC

721

*AGC

2

*

TGC

12

*

CGC

2

GluGAG

377

LysAAG

205

o

TAG

GinCAG

00 C

O

*GAA

30

*AAA

18

0

TAA

1

*CAA

-

AspGAT

14

AsnAAT

2

TyrTAT

32

HisCAT

1

*GAC

2320

*AAC

97

ft

TAC

CO

C

O

*CAC

1210

ValGTG

454

MetATG

1210

LeuTTG

11

LeuCTG

1027

*GTA

110

HeATA

-

*TTA

-

*

CTA

—

*GTT

424

*ATT

133

PheTTT

101

*

CTT

51

*GTC

2

*

ATC

826

*

TTC

213

*

CTC

1

AlaGCG

198

ThrACG

29

SerTCG

3

ProCCG

1519

*GCA

5

*ACA

1

*TCA

-

*

CCA

1

*

GCT

313

*ACT

13

ft

TCT

7

*CCT

6

ft

GCC

51

*ACC

316

ft

TCC

13

*ccc

2

9269



I V H T C T I C B * D B G K T T L T A A

I V N V C T I C B f D B C K T T L T A A

P B V N I C T I C K V D B C K T T L T A A

EtRTEP D I M I C T I C I V D B C K T T L T A A

II P M V E Vtkj) Y CBI DKAPEERA R C

T i t CAAtAFDQ-

CANFLDYAA|l DKAPEERARC

-C NSK

' 6 0

99

96

9 *

I T I N T A K T E V E T A I R B Y S I V D C P C I A D T I I NHITCAAOHD CAILVVEAAD CPMPQTDEKlJ 110

ITIHtQlVET Q I P T « « T Q I V DCPCBADrfvjl NHITCAAOHD C A I L V V|ALJT|D GPHPOTREBID 119

I T I ^ T A I T E I ElAIRKTSBf DCPCBADYII NHITCAAQHD C A iQ f «|AJA|TJD CgMP9T I [^L 196

ITIMTABVEr E1 H m B T[A]I V DCPClADvfvlt MMITCAAgMD CAILVVBAAD CPMPQTKEBl| lit

TT

EC

SC

ES

L L A U I t V P T I V I F M I V D H

L L A H I C l k

LLAKOVCVI

VDDPELLDLV EHEV

i vrh.biK]CJDH VDDHELLELV EMEV

IVVFV4KVD|T HDDPEQ.ELV EH

mjEHg [7pp f l r LLEL I EJTIEII

TEFPCD E*PV

TEFPCDDI

I I C S A L L A L E

THI v ICSALIKALE

I_fl«CSAL(c ALE

PV IP SSALLS^E

1 to

1 T9

116

179

VDETI PTPE

vrjspr 1 PTPTITB

n *2 1 «

Fig• 4. Comparison of the tight domain of the EF-Tu sequencesfrom T.thermophiluB (TT), E. coll EF-TuA (EC), 5.cerevieiae mito-chondrium (SC) and Euqlena qracilis chloroplast (EG) . Regions ofhomology are framed. Amino acids of the putatiue GDP bindingdomain are underlined. The cys-81 which have all elongationfactors Tu in common is marked by an asterisk.

For example the lysine codon AAG is used twenty times in the

T.thermophilus EF-Tu gene but AAA is absent. On the other hand,

in the E.coli tufA gene the codon AAG is used only five times

whereas AAA appears eighteen times. The only exceptions to the

preference for G and C at position 3 of the codon exists for

phenylalanine and isoleucine. In these cases T is prefered over

C in the third position. This is in accordance to the observation

that in T.thermophilus codons with a C in the third position are

generally avoided (table 2).

The protein sequence of T.thermophilus elongation factor Tu is

highly homologous to the E.coli elongation factor. In fig. 4 the

sequences of elongation factor Tu from T . thermophilus, E . coli

(3, 4), Saccharomyces cerevisae mitochondrium (5) and Euqlena

qracilis chloroplasts (6) are compared. Identical regions were

found in the sequences of all four proteins. In general the four

proteins show a high degree of homology especially in the guano-

9270



sine nucleotide binding domain. Regions where the four sequences

are not homologous are confined mostly to loops. A remarkable

feature of the T.thermophilus elongation factor Tu are ten amino

acids in the region 181 - 190. This sequence is absent in E.coli

and is partially absent in yeast mitochondrial EF-Tu. In Euqle-

na qracilis chloroplast EF-Tu this sequence is present and con-

tains several basic amino acids as is the case with T.thermophi-

lus . By comparing the amino acid sequences there is no obvious

structural feature which could be attributed to the thermostabi-

lity of T.thermophilus EF-Tu. In contrast to an earlier work

which reported the presence of a disulfide bridge (11) we identi-

fied only one cysteine residue in the protein.

Expression of T . thermophilus EF-Tu was achieved by transferring

pLS 601 and pLS 652 carrying the 6.6 kb Bglll and the 6.6 kb

BamHI fragment, respectively, to an E.coli minicell producing

strain. In case of pLS 601 EF-Tu was identified as a radioactive

protein band comigrating with the T.thermophilus EF-Tu sample

(fig. 5, lane 3). This band showed a positive reaction with anti

EF-Tu GDP antibodies from rabbit (data not shown). Three addi-

tional proteins were expressed from pLS 601.

No expression occurs using pLS 652 suggesting that the promo-

tor region is situated at least 1.5 kbp upstream from the T.ther-

mophilus tufl gene.

In order to identify the nucleotide binding domain of EF-Tu

from T.thermophilus and especially to clarify the role of the

insertion of the additional 10 amino acids (residues 181-190) in

the protein, we performed affinity labeling experiments. Elonga-

tion factor Tu was therefore prepared in a nucleotide-free form

and charged with radioactive guanosine-5'-diphosphate or guanosi-

ne-5'-triphosphate. In one experiment EF-Tu*[ C]GTP was oxidized

by sodium periodate and subsequently reduced with sodium cyanobo-

rohydride. After cyanogen bromide cleavage of the labeled protein

the radioactive fragment was identified by autoradiography

(fig. 6). Assignment of cyanogen bromide fragments was performed

by their partial sequencing in a liquid phase sequenator. GTP

binds specifically to a region containing the 10.1 kD cyanogen

bromide fragment CB1 (fig. 3) . This part of the elongation

factor Tu originates from the N-terminus and contains a so called

9271



1 2 3 kDI

— - 80

50

- 30Fig• 5 . Autoradiography. of 35S-labeled

,- 20

crude cell lysatea of E. coli minicellacarrying different T.thermophilus DNAfragments cloned in pBR 322.Lane 1: minicells harbouring pBR 322;lane 2: minicella harbouring pLS 652;lane 3: minicells harbouring pLS 601.

glycine loop (residues 20-28) which was suggested to be a part

of the GDP binding domain in the E•coli protein (25). In another

affinity labeling experiment the nucleotide-free EF-Tu was char-32

ged with [B P]GDP and the complex irradiated with a laser. In

this case the 5.0 KD fragment CB4 (fig. 3) was predominantly

labeled by a guanosine nucleotide. This fragment contains a part

of T.thermophilus EF-Tu which corresponds to the helical domain F

of the E.coli protein (25). Eight of ten aminoacids comprising

the extra loop (residues 181-190) which is not found in E .coli

EF-Tu is localized at the N-terminus of this fragment.

The results of both affinity labeling experiments are consis-

tent with placing the extra loop in the vicinity of the GDP-

binding site in T.thermophilus.

DISCUSSION

The elongation factor Tu from T.thermophilus has a molecular

weight of 44,600 D, only slightly higher than the EF-Tu from

E. coli • Determination of the molecular weight by SDS-PAGE pro-

9272



A BkD 1 2 3 4 kD

10.1 -

I- 5.0

Fig. 6. Affinity labeling of the GDP/GTP-binding site of EF-Tu.E^jTu was labeled by [U C]GTP . (A) and photolabeled by[ P]GDP (B). Cyanogen bromide fragments were separated by SDS-urea-polyacrylamide gelectrophoresis (lane 1 and 3). The gel wasdried and analysed by autoradiography (lane 2 and 4). The 10.1 kDand the 5.0 kD cyanogen bromide fragments correspond to CB1 andCB4 in fig. 3. For details see 'materials and methods' .

vides a considerably higher value of 51,000 D. This apparent

difference is not as large as that found for the E•coli EF-

Tu (26), probably due to the more lipophilic character of the

thermophilic protein.

A typical feature of proteins from thermophilic bacteria is

their low cysteine content (27). Correspondingly, the T•thermo-

philus EF-Tu has only one cysteine residue compared to three in

the E. coli and B.stearothermophilus elongation factors Tu (28).

The cysteine present in the T.thermophilus EF-Tu corresponds to

residue 81 in the homologous E.coli protein. This cysteine is

conserved in all procaryotic elongation factors Tu and can be

specifically labeled with N-tosyl-L-phenylalanylchloromethane

(28). This modified EF-Tu interacts with guanosine nucleotides as

9273



does the native protein but binds the aminoacyl-tRNA poorly.

Therefore cys-81 is probably involved in the binding of aminoa-

cyl-tRNA to EF-Tu'GTP (28). This hypothesis is supported by the

fact that the sequence corresponding to residues 79 - 88 of the

T.thermophilus EF-Tu is identical for all known procaryotic fac-

tors Tu .

Typical changes are apparent comparing the amino acid compo-

sition of homologous proteins from thermophilic and mesophilic

bacteria. For instance in thermophilic proteins glutamic acid is

preferentially used in place of aspartic acid and arginine is

prefered over lysine (29). These changes result from an increased

usage of codons with a high G+C content in thermophilic bacteria.

In the case of EF-Tu from T.thermophilus there is a preference

for valine (38 in E. coli and 52 in T.thermophilus) over isoleu-

cine (29 in E.coli and 21 in T.thermophilus) . This is related to

the base composition of the codons for these amino acids. T.ther-

mophiluH uses the GTG valine codon very often whereas the ATA,

ATC and ATT codons for isoleucine are avoided. The evolutionary

pressure towards codons with a high G+C content is obvious for

almost all codons used in the T.thermophilus EF-Tu gene. A high

G+C content is most pronounced in the first and third codon

letter. This phenomenon was also observed for the moderate ther-

mophilic B.BtearothermophiluB genes (30, 31) - in a less pro-

nounced manner - and confirmed for the gene of isopropylmalate

dehydrogenase in T. thermophilus (32), where a very high G+C

content (89 %) in the third letter was found. A remarkable excep-

tion for the preference of codons with a high G+C content in

thermophilic genes is found in the codons for phenylalanine and

isoleucine. In the T • thermophilus tufl gene the phenylalanine

codon TTT is used ten times and the codon TTC only three times,

whereas an opposite ratio of 1 to 13 is found in the E.coli tufA

gene. Similarily the codon ATT is prefered over ATC for isoleu-

cine in the thermophilic protein in contrast to the mesophilic

variant. For unknown reasons in the thermophilic tuf gene the TTT

(ATT) codon has an advantage over the TTC (ATC) codon. However,

this is not the case in the T.thermophilus isopropylmalate dehy-

drogenase gene. Parker and Precup recently reported that leucine

is misincorporated with a higher frequency at UUC than UUU in

9274



E•coli (33), a finding which may reflect the stability of

codon : anticodon interactions involving UUN codons. As an alter-

native explanation the differences in codon usage for phenylala-

nine in E.coli and T. thermophilus tuf genes could be due to

different concentrations of tRNA isoacceptora in both bacteria.

However, in E . coli as well as in T.thermophilus only a tRNA

with an anticodon GAA was identified (K. Watanabe, M. Sprinzl,

unpublished). A possible explanation for the unusual phenylala-

nine and isoleucine codon usage Mould be the different distribu-

tion of the respective codons in the two T.thermophilus tuf

genes. In such a case, the unusual UUU and AUU codons could have

a regulatory function in expression of the different tuf genes.

The tuf2 gene of T.thermophilus has to be sequenced to test this

possibility .

The T.thermophilus EF-Tu gene has a long open reading frame in

the complementary DNA strand. Whether this is connected to the

fact that in highly expressed genes the codons complementary to

nonsense codons are extremely rare (34) or this arrangement is

important for a regulatory principle connected with an anti-sense

mRNA remains to be clarified.

Expression of T . thermophilus EF-Tu in E . coli is possible but

not efficient. There are several factors which could negatively

influence the expression. A toxic effect of the thermophilic EF-

Tu on the translation system of E.coli can be excluded since

other proteins encoded on the plasmid pLS 601 (e.g. Q-lactamase)

are normally expressed. Comparison of the translation products of

pBR 322 and pLS 601 indicates the additional expression of four

proteins. If the tuf 1 gene of T.thermophilus is a part of a

polycistronic mRNA, homologous to the E.coli str mRNA (35), these

translational products could correspond to the T. thermophilus

variants of the E.coli EF-G, EF-Tu and ribosomal proteins S7 and

S12. An identical gene organization exists for Bacillus stearo-

thermophilus (H. Kimura, personal communication). Since the two

putative S7 and S12 proteins (17 kD and 22 kD) are expressed to

an appreciable amount in the minicell system a misfunction of

their T.thermophilus promotor in E.coli is not likely. The long

ribosomal binding sequence in T.thermophilus as compared to E.co-

li (36) may be a reason for the low expression rate of T • thermo-

9275



philus EF-Tu in E. coli • Work is in progress to clarify these open

questions.

Acknowledgements

We thank Dr. B. Wittmann-Liebold and J. Brockmoller for se-

quencing the EF-Tu cyanogen bromide fragments, Dr. N. Schumann

for advice with the minicell expression system, R. Marmorstein

and Dr. H.G. Faulhammer for discussions and J. Beck for technical

assistance. This work mas supported by the Deutsche For-

schungsgemeinschaft SFB 213/D5 and Fonds der Chemiachen

Indus trie.

REFERENCES1. Miller, D.L. and Weissbach, H. (1977) in Molecular Mechanisms

of Protein Biosynthesis, Ueissbach, H. and Pestka, S. eds.,pp. 324-374, Academic Press, New York

2. Tapio, S. , Kurland, C.G. (1986) Mol. Gen. Genet. 202., 186-1893. Yokota, T., Sugisaki, H., Takanami, H. and Kaziro, Y. (1980)

Gene J^, 25-314. An, G. and Friesen, J.D. (1980) Gene 1_2, 33-395. Nagata, S., Tsunetsugu-Yokota, Y., Naito, A. and Kaziro, Y.

(1983) Proc. Natl. Acad. Sci. USA 80, 6192-61966. Montandon, P.-E. and Stutz , E. TT983) Nucl . Acids Res. 11 ,

5877-58927. Van Hemert, F.J., Amons, R., Pluijms, W.J.M., Van Ormondt, H.

and Holler, U. (1984) EMBO J. 2. 1109-11138. Nagata, S., Nagashima, K., Tsunetsugu-Yokota, Y., Fujimura,

K., Miyazaki, M. and Kaziro, Y. (1984) EMBO J. J5, 1825-18309. Brands, J.H.G.M., Haassen, J.A., Van Hemert, F.J., Amons, R.,

Mdller, U. (1986) Eur. J. Biochem. 15±, 167-17110. Moller, U. and Amons, R. (1985) FEBS-Lett. LB6, 1-711. Nakamura, S., Ohta, S., Arai, K., Arai, N., Oshima, T. and

Kaziro, Y. (1978) Eur. J. Biochem. 9J_, 533-54312. Derwenskus, K.-H., Fischer, W. and Sprinzl, M. (1984) Anal.

Biochem. 11±, 161-16713. Gebhardt-Singh, E. and Sprinzl, M. (1986) Nucl. Acids Res.

14, 7175-718814. Johnson, R.A. and Walseth, T.F. (1979) Adv. Cyclic Nucl.

Prot. Phos. Res. 10, 136-16715. Southern, E. (1975T J. Mol. Biol. 98, 152-17616. Rigby, P.U.J., Dieckmann, M., Rhodes, C. and Berg, P. (1977)

J. Mol. Biol. U 3 , 237-25117. Inone, M. and Curtiss, R. (1977) in Molecular Cloning of

Recombinant DNA, Scott, W.A. and Uerner, R. eds., pp. 248-261, Academic Press, New York

18. Messing, J. (1983) in Methods in Enzymology, Uu, R., Gross-man, L. and Moldave, K. eds., vol. 101, pp. 20-78. AcademicPress, New York

19. Meagher, R.B., Tait, R.C. and Boyher, H.W. (1977) Cell JJ),521-536

20. Leemmli, U.K. (1970) Nature 2^2, 680-685

9276



21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32 .

33.34 .35.

36.

, Antonsson, B., Giov/anelli, R., Guariguata, R.,and Wittinghofer, A. (1980) Anal. Biochem. 104,

R.T. and Munkres, K.D. (1971) Anal. Biochem. 22., 462-

Anderegg, R.J., Herlihy, W.C., Gray, C.P.,G.E.

Leberman, R.Schumann, R.29-36Swank ,477Gerber ,Biemann, K. and Khorana, H.G. (1979) Proc. Natl. Acad. Sci.USA 1±, 227-231Bosch, L., Kraal, B., Van der Meide, P.H., Duisterwinkel,F.J. and Van Noort , J.M. (1983) in Progress in Nucleic AcidsResearch and Molecular Biology, Cohn, E.W. and Moldav/e, K.eds., vol. 22., PP • 91-126. Academic Press, New YorkLa Cour, T.F.H., Nyborg, J., Thirup, S. and Clark, B.F.H.(1985) EMBO J. 4_, 2385-2388Blumenthal, T., Landers, T.A. and Weber, K. (1972) Proc.Natl. Acad. Sci. USA 69, 1313-1317Amelunxen, R.E. and Murdock, A.L (1978) in Microbial Life in

J. ed., pp.217-278. Academic

B. and Karas, K. (1986) Eur.

G.

Extreme Environments, Kushner,Press, New YorkJonak, J., Pokorna, K., HelounJ. Biochem. 154, 355-362Argos, P., Rossmann, H.G., Gran, U.H., Zuber, H., Frankand Tratschin, J.D. (1979) Biochemistry _18_, 5698-5703Winter, G., Koch, G.L.E., Hartley, B.S. and Barker, D.G.(1983) Eur. J. Biochem. 1_3_2, 383-387Ihara, H., Sasaki, T., Tsuboi, A., Yamagata, H., Tsukagoshi,N. and Udaka , S. (1985) J. Biochem. 9_£, 95-103Kagawa, Y., Nojima, H., Nukiwa, N., Ishizuka, M., Nakajima,T., Yasuhara, T., Tanaka, T. and Oshima, T. (1984) J. Biol.Chem. 2_5_9, 2956-2960Parker, J. and Precup, J. (1986) Hoi. Gen. Genet. 204, 70-74.Alff-Steinberger , C. (1984) Nucl. Acids Res. 12, 2235-2241Jaskunas, S.R., Fallon, A.M. and Nomura, M.~Tl977) J. Biol.Chem. 252, 7323-7336Stormo, G.D., Schneider, T.D. and Gold, L.H. (1982) Nucl.Acids Res. 10, 2971-2996

9277
