Cloning and expression of the glutaredoxin (grx) gene of Escherichia coli

13

Gene, 43 (1986) 13-21 Elsevier GENE

1576

Cloning and expression of the giutaredoxin (grx) gene of ~sc~er~e~~uco& (Recombinant DNA; overproduction; Ml3 phage vector)

nucleotide sequence; oligodeoxyribonucleotide

probe; hybridization;

Jan-Olov HWg, Hedvig von Bahr-LindstriTm,Hans JiSrnvailand Arne Holmgren Department of Chemistry I, Karolinska Institutet. S-104 01 Stockholm, (Sweden) Tel. + 46-(8)-340560 (Received

November

(Revision

received

(Accepted

February

2&h, 1985) February

3rd, 1986)

4th, 1986)

SUMMARY

Two DNA segments, together comprising 1147 bp and containing the glutaredoxin (GRX) gene, grx, from Escherichiu coli K-12 were cloned and characterized in M 13mp9. The gene was identified by hybridization with

synthetic oligodeox~ibonucleotide probes corresponding to parts of the amino acid (aa) sequence of GRX. The sequence of 255 bp comprising the GRX structural gene gave a deduced aa sequence identical to the directly determined one. The coding region is preceded by two possible ribosome-binding sites and three possible promoters with - 10 and - 35 regions as judged by homology to consensus sequences. The presence of a stable stem-loop structure, dG = - 17.0 kcal, followed by six thymine bases indicates that the transcription of the grx gene is Rho-independently terminated. An over-representation of rare codons in the grx gene, as compared to the genes for thioredoxin (TRX) and highly expressed proteins, is suggested as one possible explanation for the large difference in the synthesis between TRX and GRX in wild-type E. coli cells. GRX production was amplified at least 100 fold in strain JM103[pEMBL9ECG] over that in wild-type E. coli cells, The protein purified from the overproducing strain was identical in aa composition and N-terminal aa sequence with the previously analyzed GRX protein.

INTRODUCTION

Glutaredoxin (GRX) from E. coli is a small dithiol protein (10 kDa) required for glutathione-dependent synthesis of deoxyribonucleotides catalyzed by the

Abbreviations: ethidium

glutaredoxin; tathione;

aa,

bromide; GSH,

amino GRX, reduced

nt, nucleotide(s);

polyacrylamide; [ 1, designates

acid(s);

bp,

glutaredoxin; gfutathione;

base

pair(s);

grx, gene GSSG,

oxidized

oligo, oligodeoxyribonucleotide;

RBS, ribosome-binding plasmid-carrier state.

EtdBr,

coding

for gluPA,

site; TRX, thioredoxin;

essential enzyme ribonucleotide reductase (Holmgren, 1976). GSH is regenerated from GSSG in the presence of NADPH and glutathione reductase (Holmgren, 1979). Originally, TRX, another small dithiol protein (12 kDa), was identified as hydrogen donor for ribonucleotide reductase in vitro (Laurent et al., 1964). Oxidized TRX is reduced by NADPH and TRX reductase, making NADPH the ultimate hydrogen donor for ~bonucleotide reductase in both systems. GRX was discovered in a mutant of E. coii lacking detectable TRX activity (Holmgren et al., 1978) but with a fully active NADPH-dependent

0378-1~~9~86/$03.50 0 1986 Elsevier Science Publishers B.V. (Biomedical D~vjsion)

14

ribonucleotide reduction system (Holmgren, 1976). The level of GRX in E. coli K-12 is regulated in a complex manner since mutants in TRX reductase (trxS) overproduce the protein about 5-fold and in contrast, derepression of ribonucleotide reductase (nrdA and nrdB) by thymine starvation leads to a decrease in the GRX content (Holmgren, 1983). The content of GRX in wild-type E. coliB cells has been estimated to be about loo-fold lower than that of TRX (100-200 vs. 10 000 copies per cell, respectively; Holmgren, 1981). However, the activity of ribonucleotide reductase in vitro is about IO-fold higher with GRX than with TRX (estimated turnover numbers are 110-150 per min for GRX and 13-15 per min for TRX; Holmgren, 1981). The relative contribution of TRX and GRX in ribonucleotide reduction by normal cells is not yet known. TRX is not essential for growth of E. coli cells as shown by analysis of mutants in the gene (trx_4) located at 84.7 min on the E. coli linkage map (see Holmgren, 1985). The role of GRX in E. coli can be analyzed using mutants in the grx gene. However, no mutants have yet been isolated, nor has the grx gene been mapped on the E. coii chromosome. Since no simple procedure exists for the isolation of grx - mutants, we have attempted to clone the grx gene. To identify the gene, we used the previously determined aa sequence of GRX (Ho&g et al., 1983) to construct an oligo probe. This study describes the cloning and ch~acte~zation of a 1.1-kb DNA segment containing the structural gene for GRX from E. coli K-12 including its potential promoter(s) and terminator, and the isolation and analysis of GRX from an overproducing strain.

MATERIALS

AND METHODS

tion enzymes were used under the conditions recommended by the manufacturer (Boehringer). DNA fragments were isolated from PA gels by diffusion and from agarose gels by electrophoresis onto an NA45 filter (Schleicher & Schlill) as described (Dretzen et al., 198 1). Additional purification was by ion-exchange chromatography on DE-52 (Whatman). Synthetic oligos used are listed in Table I. (b) Bacterial strains, vectors and growth conditions used for GRX production Vectors pUN121 (Nilsson et al., 1983), pBR322 and a 2.6-kb fragment of pBR325 (Wallace and Kushner, 1984) were transformed into E. coli K-12 strain HBlOl, and vectors pEMBL9 (Dente et al., 1983) and M13mp9 (Messing and Vieira, 1982) into strain JM103. Strains with plasmids containing the grx gene have the symbol ECG added to the vector name. HB 101 was grown in LB-medium and JM 103 in 2 x YT (Maniatis et al., 1982). (c) Purification and characterization of GRX Cells from early stationary phase cultures of JM 103[pEMBL9ECGl harvested by centrifugation, were lysed by sonication and lysozyme treatment (Maniatis et al., 1982). GRX was purified from the cell extract by a two-step procedure, ~munosorbent ~hromato~aphy followed by gel fdtration on Sephadex G-50 (Hoog et al., 1983). The GRX content was monitored by a GSH-disulphide transhydrogenase assay (Holmgren, 1979). N-terminal sequence analysis of purified GRX was performed by the manual d~ethyl~ino~obenzene isothiocy~ate method (von Bahr-Lindstrom et al., 1982). The total aa composition was determined on a Beckman 121M analyzer after acid hydrolysis.

(a) Preparation of DNA and fragments The E. coli K-12 strain ClO-17 (m&B, upp, udk, thyA& trxB) derived from strain KK1006 (Fuchs, 1977) was used as source of chromosomal DNA. Cells were grown overnight at 37°C in L-broth (Maniatis et al., 1982) with 1% glucose supplemented with 50 pg thyminelml before extraction of chromosomal DNA. Plasmid DNA was isolated using the alkaline lysis method (Maniatis et al., 1982). Restric-

RESULTS

AND DISCUSSION

(a) Identification and nt sequence of the grx gene DNA from the E. coli K-12 strain ClO-17 was fragmented with several restriction enzymes. The digests were separated on agarose gels, transferred to nitrocellulose filters and hybridized with a mixed

15

TABLE

(Fig. 2). The l.O-kb HaeIII

I

Synthetic

oligos used as hybridization

and 0.7-kb BglII-EcoRI

bands were cloned into M 13mp8 and M 13mp9 vec-

probes a

tors for further characterization. 82

78 Trp-Va

14-mer

I-Lys-Glu-Asn

3’-ACC

CAN

TT;

(JM103[M13mp9ECG3])

CT;

TT -5’

57

51 CGT

CCA

and

and positional

GGG

CAT

numbers

CTT-5’

refer to E. coli

K-12 GRX (Hoog et al., 1983). Oligos, kindly provided S. Josephson,

KabiGen

automatic

AB, Stockholm,

synthesizer

method (Josephson

by

the

[$‘P]ATP

(Amersham;

kinase (New England Nitrocellulose Amersham)

Biolabs)

(Maniatis

were performed the 21-mer

Cronex

phosphoamidite probes,

of phosphate

from

as described

(HBog et al., 1984).

or nylon

(Hybond

as recommended

N;

lift technique,

by the

or by the colony

et al., 1982). Hybridization et al., 1979). Hybridization

overnight Lightning)

over

recombinant

BglII-EcoRI

at a frequency

clone of 0.2%.

er, of the inserts in the hybridization-positive the grx gene. clones The

identified

them to contain

clones, parts of

JM103[M13mp9ECG3]

and

JM 103[ M13mp9ECG9] were subjected to sequence analysis. Altogether, the two clones comprise a sequence of 1147 nt (Fig. l), all determined on both strands. A coding region of 255 nt is in full agreement with the aa sequence 1983).

of E. coli GRX (Hoog

et al.,

to the filters by Southern lift

filters

at - 70” C using intensifying and Kodak

XAR-5

(b) Production of GRX

and washing

at 37°C for the mixed 14-mer, and at 45°C for

(Wallace

autoradiographed

in an

using T4 polynucleotide

& Schtill)

DNA was transferred

by the plaque

technique

(DuPont

by transfer

filters were used and treated

manufacturers. blotting,

solid-phase

3000 Ci/mmol),

(Schleicher

by Prof.

were synthesized

et al., 1984). For use as hybridization

the oligos were 5’ end-labelled

positive

clone

of 0.1%

Dideoxy sequence analysis, with the 21-mer as prim-

TTT

unambiguously a Amino acid sequences

plaques

a

(JM103[M13mp9ECG9])

Lys-Ala-Gly-Lys-Pro-Val-Glu 3’-TTC

at a frequency

(hybridization-positive plaques)

2 1-mer

Plaque hybridiza-

tion with the 21-mer showed a positive HaeIII

were

screens

film.

14-mer (Table I), synthesized to match a part of the aa sequence of E. coli GRX. Due to the high AT content and multiplicity of the 14-mer probe, no distinctly separated fragments of a size likely to contain the complete grx gene were found with this hybridization probe. However, the results indicated that the Sau3A digestion produced small fragments with sequences complementary to the probe. Accordingly, a mixture of all Sau3A fragments smaller than 400 bp was eluted from an agarose gel and cloned into the Barn HI site of M 13mp8. Plaque hybridization with the mixed 16mer revealed positive clones which were subjected to dideoxy sequence analysis. An insert of 118 bp corresponding to a part of the grx gene, position 139-256 in Fig. 1, was identified. The partial nt sequence determined was used to construct a unique 21-mer (Table I). Hybridization with the 21-mer to Southern blots of fragments from E. coli K-12 ClO-17 DNA revealed distinct bands of sufficient size to possibly contain the grx gene

containing the grx gene, of The insert, M 13mp9ECG3 was isolated and ligated into plasmid vectors, pBR322, pBR325, pUN121 and pEMBL9, and the resulting plasmids were transformed into E. coli strains HBlOl and JM103. Colonies were selected by hybridization to the GRXspecific 21-mer (Table I). Cultures of hybridizationpositive clones and of appropriate controls were grown to stationary phase, and the cells were collected by centrifugation. Cell extracts were assayed for GRX by double antibody radioimmunoassay and for total protein by UV-absorbance (Table II). The results show a background production of less than 160 ng GRX/mg protein, a lo-20-fold overproduction in JM103[M13mp9ECG3], HBlOl[pBR322ECG], HBlOl[pBR325ECG] and HBlOl[pUN121ECG] and a more than lOO-fold overproduction in JM103[pEMBL9ECG]. GRX was isolated from early stationary phase cultures of clone JM103[pEMBL9ECG] by immunosorbent chromatography and gel filtration (Hoog et al., 1983). The purified GRX was analyzed by aa analysis (Table III) and N-terminal sequence determination. The total GRX composition from the overproducing strain JM103[pEMBL9ECG] agrees within the experimental errors with that calculated from the aa sequence, and the N-terminal sequence

16

GGCCGAGAAAA TGGCTCAGGC TCCGATAAAT AGTGAGCC

TCCAGTATCA

CAGGTTAAAC ATA

CTAAAA

==%=

CTACTTTCAG

AGCGACAGGG

AGTCGCTTAC

CGACAGCAGC

GCCACCGCCA

GTAGCATCAT

-214

ACACGCCTTT

AGGCAATTTA

CCGATCGCGC

GCATACGCTT

CCCTCTGCAA

-124

CTTTG

Jf$G‘CGTJT

CGAATACATT

t Sau3A35 TTAGCGTGAT CATTACAGGC

AAI;BGT

-74 -h

ATAAATCTwi?@@%?$A

CAA

ACC

GTT

ATT

TTT

GGT

CGT

TCG

GGT

TGC

CCT

TAC

TGT

GTG

CGT

GCA

AAA

M

Q

T

V

I

F

G

R

S

G

C

P

Y

C

V

R

A

K

1

L

GCT

GAG

AAA

TTG

AGC

A

E

K

L

S

AAT N

GAA

CGC

GAT

GAT

TTT

E

R

D

D

F

CAG

TAT

CAG

GTA

GAT

Q

Y

Q

TAT Y

V

0

GTA

111

30

ATT

CGT

GCG

GAA

I

R

A

E

t Sau3A GGG ATC G

I

t ACT

AAA

GAA

T

K

E

Eigm GAT CTA D

L

CAA

CAA

AAG

GCA

GGT

AAA

CCC

Q

Q

K

A

G

K

P

GGC

GGC

TAT

ACC

GAT

CD ACC

54

10

20

GAA

I\U$AlGCAA

ATG

t Sau3A GAT CTG D

t AluI JA~~OG~TTTAGCA

GTG

168

11

50 CCG

CAG

ATT

TTT

GTC

f Sau3A GAT CAG

CAA

CAT

ATC

TTT

225

ETVPQXFV’DQQHIGGYTDF 60

70

GCT

GCA

TGG

GTG

AAA

GAA

AAT

CTG

GAC

GCC

A

A

W

V

K

E

N

L

D

A

80 TACTGATTTT

TTCTGTGCTG

$ Sau3A TGA TCGTCTGACA

AGCCCfCGCG *

TTGAGGGCTT 4

280

85 TGGTTTAAAC

AAACTACTGA

TAAATAAGAA

ACACAGTGCC

CCCAGCGCAC

356

GAATGAGCGC

GTCGGTGAAA

AAAACAGCCG

428

ACCAGAACAC

CGCGCTTAGT

AACCATGCCA

t AluI GCTCTTGCCA

CATAATGAGC

ATCGAACAGG

GTGCCGCCAG

CATTGCGCCA

AACAGAGGTT

TCAGGACTTC

TCTACGCTGT

498

t Ah.11 GAAAAGAAGC TGGAGACTGC

TCAGAAGAAT

GAAAAATAGC

AAGCCGATTT

CAGGAATGCC

CGGCAGCCCG

568

AAAAGCGCCT

TTCATGTCGT

CGCCAGAAAA

AGGCACACCA

CAATGAAGAG

GACAAAACAG

CAGATTGCCC

638

CCGCCCAACG

TTGTTTATGT

TTCACTCGTT

CCTCCTGACA

CTGCGTCTAT t

GTGGCGTTCA TGATACCTAA CTGCGCCCTG

CGAACACATT

TTTCGCCAGT

708

ATATAGCCGT

778

CTGAATGAAT

848

GTAAGATAAA

GCCGCTTCGC

ATTCCATGCT

AATATAGGCC

HaeIII AACGCAATTC

TGTGATTACA

CTAGTAAAAT

ATATTGTTAC

TTTACTATCG

TTTAGGTGCG

) EcoRI AATTC

863

Fig. 1. The sequence of 1147 nt comprising the E. co&K-12 grx gene and flanking regions (top), and of the 85 aa comprising E. coii K-12 GRX (bottom). The nt are numbered on the right margin in the 5’ to 3’ direction; nt + 1 is the A of the ATG codon; negative numbers refer to the _5’-flankingregion. The aa are numbered underneath the sequence. Proposed RBS’s are boxed and promoter regions ( - 10 and - 35) are underlined. Region of dyad symmetry (nt 262-294) is marked by convergent arrows, Restriction enzyme sites used for sequence analysis are indicated. DNA fragments were ligated into M13mp8 and M13mp9 vectors (Messing and Vieira, 1982). Sequence analysis was performed by the dideoxy chain termination method (Sanger et al., 1977) utilizing single-stranded Ml3 templates. Synthetic ohgos or an Ml3 specific universal primer (17-mer; Amersh~) were used as primers. The labelled nt was [ti’-S]dATP (Anlersham; 600 Ci/mmol), and reaction mixtures were separated by electrophoresis in a 0.2 mm thick 6% PA gel, using linear or wedge-shaped field-strength gradients (Olsson et al., 1984).

17

TABLE

II

Content

of GRX in crude extracts

of early stationary

phase cells

of E. co/i strains ng GRX a

Strain

mg protein 90

B

23150

HBlOl

60

JM103

160

ClO-17

550 130

HBlOl[pBR322]

9420 6560

1600

HBlOl[pBR322ECG]

70

HBlOl[pBR325]

2000

HBlOl[pBR325ECG]

140

HBlOl[pUNl21]

4380

1300

HBlOl[pUNl2lECG]

150

JMl03[Ml3mp9]

2000

JMl03[Ml3mp9ECG]

140

JMl03[pEMBL9]

2320 2020

a Cells were grown to stationary cultures washed EDTA.

were chilled

extract

by

was measured

560

by Holmgren

protein

(ng

content

and Bernlohr,

of fragments

from chromosomal

E. co/i K-

12 DNA digests containing

the grx gene. (A) EtBr staining

agarose-gel

of: (lane 1) a Hue111 digest of E. coli

electrophoresis

(lane 2) a EcoRI + BglII

(lane 3) a Hind111 + BglII digest shown on left margin. after

transfer

labelled

to nylon

digest

of E. coli DNA

of 1 DNA

(B) Autoradiograph

after and

with sizes in bp

of lanes 1 and 2 in A

filter and hybridization

with the “P-

21-mer probe (see Table I).

obtained, Met-Gln-Thr-Val-Be-, is identical to that determined for GRX from strain ClO-17 (Hoog et al., 1983). (c) Conclusions (1) GRX isolated from E. colz’ K-12 (strain ClO17) has methionine as N-terminus. In contrast, the

protein

1977). Strains

from

antibody

ECG added

of

the

radioimmunoassay

raised

was

of the pellet

content

and Luthman

protein/ml)

using the formula

gene have symbol

GRX

specific antibodies,

tially as described absorbance

Fig. 2. Identification

by a double

HCI pH 7.5, 1 mM

by sonication

The

The

and the ceil pellet was

in 50 mM Tris

were prepared

centrifugation.

using E. coli GRX

phase in lOO-ml cultures.

and centrifuged,

and resuspended Cell extracts

followed

DNA;

10200

JMl03[pEMBL9ECG]

in rabbits,

essen-

(1978). The total

determined

by

UV-

183 x A,,,

- 75.8 x A,,e

(Kalb

with plasmids

containing

the grx

to the vector

E. coli B has

name.

N-terminal

glutamic

acid/glutamine (Holmgren, 1979). Consequently, the protein from strain ClO-17 appears to start one residue earlier. The N-terminal methionine residue is present also in GRX isolated from the overproducing strain JM103[pEMBL9ECG], as proved by Nterminal sequence determination and supported by aa analysis. The nt sequence indicates that the N-terminal methionine is due to non-identical processing of N-formyl-methionine during protein maturation rather than to a difference in DNA structure between E. coli B and K-12. Empirical rules predicting when N-terminal methionine residues are removed/retained have been formulated (Flinta et al., 1986), and for the GRX sequence are compatible with both situations. Strain JM103[pEMBL9ECG] was chosen for GRX purification

18

TABLE

III

Amino acid composition producing calculated

strain

of E. coli GRX purified

JM103[pEMBL9ECG]

from the protein

structure

from the over-

compared determined

with

that

for GRX from

strain CIO-17 (Hii(ig et al., 1983) Amino

acid

pEMBL9ECG

ClO-17

Cys

a

2

Asx

9.3

10

Thr

4.1

4

Ser

2.9

2

Glx

13.1

14

Pro

3.4

3

Gly

6.6

6

Ala

7.2

7

Val

6.7

7

Met

1.0

1

Be

4.6

5

Leu

4.2

4

Tyr

3.5

4

Phe

3.7

4

Trp

a

1

6.2

6

LYS His

1.3

1

Arg

4.4

4

a Destroyed

during

acid hydrolysis.

as it overproduced GRX more than 100 fold, while the other strains tested showed a lo-20-fold overproduction (Table II). (2) The initiation codon, AUG, at position 1 is preceded by a possible translation start, at position - 15, in frame with the coding region. The resulting peptide, Met-Arg-Arg-Glu-Be, is too short (5 residues) and too hydrophilic to fulfill the requirements for a classical leader peptide (Kreil, 198 1). Neither is it preceded by a typical RBS, GGAGGT (Shine and Dalgamo, 1974) and is therefore not likely to be expressed. No indication of a longer form has been found during the isolation of GRX from E. coli strains ClO-17, B or JM103[pEMBL9ECG]. Both TRX and GRX are found in the periphery of E. coli cells (Nygren et al., 1981). The presence of a potential leader peptide has recently been observed for TRX at the gene level (Wallace and Kushner, 1984; Lim et al., 1985). The tentative TRX leader peptide is longer (18 aa residues), than that discussed for GRX, and also contains charged and hydrophilic residues. It is notable that both gene structures coding for these hypothetical leader peptides contain

more than 50% rare codons (Konigsberg and Godson, 1983). (3) The postulated GRX initiation codon at position 1 (Fig. 1) is preceded by two possible and partly overlapping RBS’s. The one that starts at position - 11 shows 4 out of 6 identities to the consensus sequence, GGAGGT (Shine and Dalgarno, 1974). The same is true for the one that starts at position - 14. The distances between the ATG and the putative RBS’s are within the accepted limits for both sites (Gold et al., 1981). (4) The 5’-lIzInking region also contains three possible promoters, all with optimally located - 10 and - 35 regions (Fig. 1; Hawley and McClure, 1983). The identities with the consensus sequence TATAAT, for the - 10 region are 4/6 for all the possible - 10 regions at positions - 83, - 97, and - 113, respectively. The corresponding - 35 region shows 516, 416, and 316 identities to the consensus sequence TTGACA. Further analysis at the transcription level is required to show if one of the possible promoter sites preceding the grx gene is preferentially used or if GRX-mRNAs of different lengths are produced. (5) The coding region of thegrx gene is terminated by a UGA stop codon at position 256. In addition, two extra UGA stop codons in frame with the coding region are present at positions 280 and 292. The 3’-flanking region of 605 nt contains a likely transcription termination signal. A sequence of dyad symmetry rich in G + C is centered around position 278 (Fig. 1) and is possibly forming a strong hairpin loop, dG = - 17.0 kcal (Tinoco et al., 1973), immediately followed by a row of six thymine bases. These structures could together act as a Rho-independent transcription termination signal (Rosenberg and Court, 1979). (6) The grx gene contains an overrepresentation of rare codons (Table IV; Konigsberg and Godson, 1983) and does not follow the codon usage typical for highly expressed E. coli genes (Table IV; Gouy and Gautier, 1982). The average frequency of rare codons in non-regulatory E. coli protein genes is 4 y0 (Konigsberg and Godson, 1983) as compared to 10.6% in the grx gene. Such a high frequency is typical for regulatory protein genes expressed in low amounts (Konigsberg and Godson, 1983). The trxA gene (Hoog et al., 1984) has a normal rare codon frequency of 2.8% and shows a codon usage close

0 0 1 2

L L L L

1 I 1 M

V V V V

CTT CTC CTA CTG

ATT ATC ATA ATG

GTT GTC GTA GTG

1

1 2

1

1

3 6 0 1

0 3 0 9

3 0 1

_~

TRX

-

,,

44 7 31 17

21 79 0 100

4 3 0 86

23 77 3 4

%

Codon usage

GCT A GCC A GCA A GCGA

1 3 0 0

ACT T ACCT ACAT ACGT

3 1

1

2 4 5

1

2

1

3 2 0

1 0 0 4

1 1 0

1

0 0 0

1

TRX

0 0 1

0

P P P P

CCT CCC CCA CCG

T-CT s TCC S TCAS TCG S

GRX

26 18

6

50

48 41 7 3

6 3 18 14

44 31 2 0

%

Codon usage

Y Y -

GATD GAC:D GAA E GAGE

7 I 5 f

2

0 5 I

AACN AAAK AAGK

I 0 4 4

3 1 0 0

MN

CAT H CAC H C_AAQ CAGQ

TAT TAC TAA TAG

GRX

5 6 3 2

5

0 4 x0 0

31 69 7.5 25

95 71 29

32 6X _14 86

17 83 -

%

_

Codon usage

0 1 I 2

2 0 1 0

TRX

GGT GGC GGA GGG

AGT AGC AGA AGG

CGT CGC CGA CGG

G G G G

S S R R

R R R R

TGT C TGC C TGA TGGW

3 2 0 1

0 I 0 0

3 1 0 0

1 1 1 1

GRX

.

5 3 0 1

1 1 0 0

1 0 0 0

0 2 0 2

TRX

58 40 1 1

6 17 0 0

70 28 0 0

100

20 80

%

Codon usage

a The GRX and TRX columns denote the number of times each codon is used in the E. co/igrx (Fig. 1) and rrxA (W66g et al., 1984) genes, respectively. Percent codon usage denotes ratio of codons to total codons of a specific residue for a highly expressed protein calculated from data of Gouy and Gamier (1982) Codons rare in E. cob non-regulatory proteins (Kon~gsberg and Godson, 1983) are underfined.

1 1 2 3

3 2 0 1

L

4 0 0

TTT F TTC F TTA L TTGL

GRX

Codon usage a.

TABLE IV

20

to the one for highly (Table II). This dserence the grx and trxA genes is regulation of their relative

expressed E. coli genes in codon usage between probably one part of the amounts.

Holmgren,

A.: The glutaredoxin

nius, S., Holmgren, of Glutathione:

Biochemical

Clinical Aspects. Holmgren,

lular localization

Toxicological

Annu.

Rev. Biochem.

and

Biochemistry

A., Ohlsson,

54 (1985)

M.: Tissue distribution

of bovine thioredoxin

ioimmunoassay.

Skillful technical assistance by Susie Bjorkholm, Ella Cederlund and Barbro Mattsson is gratefully acknowledged. We are grateful to Stat&n Josephson, KabiGen AB for the oligonucleotide synthesis. This work was supported by grants from the Swedish Cancer Society (projects 961 and 1806), the Swedish Medical Research Council (13X-3529, 3532 and 7148) and the Magn. Bergvall Foundation.

A., Orre-

B. (Eds.) Functions

Physiological,

A.: Thioredoxin.

Holmgren,

In Larsson,

Raven Press, New York, 1983, pp. 199-203.

237-271. Holmgren, A. and Luthman,

ACKNOWLEDGEMENTS

system.

A. and Mannervik,

and subcel-

determined

by rad-

17, (1978) 4071-4077.

I. and Grankvist,

M.-L.: Thioredoxin

from Escherichia coli. Radioimmunological

and enzymatic

terminations

defective in phage

in wild type cells and mutants

T7 DNA replication. Hodg, J.-O., Jornvall,

J. Biol. Chem. 253 (1978) 430-436.

H., Holmgren,

son, M.: The primary in. Eur. J. Biochem. Ho@,

de-

A., Carlquist,

M. and Pers-

of Escherichia coli glutaredox-

structure

136 (1983) 223-232.

J.-O., von Bahr-Lindstrom,

H., Josephson,

S., Wallace,

H. and Holmgren,

A.: Nucleo-

B.J., Kushner,

S.R., Jdrnvall,

tide sequence

of the thioredoxin

gene from Escherichia co/i.

Biosci. Rep. 4 (1984) 917-923. Josephson,

S., Lagerholm,

E. and Palm, G.: Automatic

of oligodeoxynucleotides

and mixed

using the phosphoamidite

method.

synthesis

oligodeoxynucleotides

Acta Chem. Stand.

B38

(1984) 539-545. Kalb, V.F. and Bernlohr,

REFERENCES

for von Bahr-Lindstrdm, ducts

H., Hempel,

as an aid in residue

J. and Jdmvall,

identification

quence analysis with dimethylaminoazobenzene ate. J. Prot. Chem. Dente,

L., Cesareni,

of single

peptide

se-

isothiocyan-

plasmids.

R.: pEMBL:

Nucl.

Acids

a new family

Res.

11 (1983)

G., Bellard, M., Sassone-Corsi,

reliable

method

agarose

and

for the recovery

acrylamide

P. and Chambon, of DNA

gels. Anal.

determinants

for cytosolic

Eur. J. Biochem. Fuchs,

J.: Isolation

thioredoxin

fragments

Biochem.

and Stormo,

of proteins

gene

processing.

initiation

S., Singer, B.S.

in prokaryotes.

An-

C.: Codon usage in bacteria: Nucl.

Acids

Escherichia coli promoter

W.R.: Compilation DNA sequences.

Res.

A.: Hydrogen

nucleoside-diphosphate tathione.

10

(1982)

and analysis

of

Nucl. Acids. Res.

donor system for Escherichia coli riboreductase

cleotides.

dependent

upon

glu-

Sci. USA 73 (1976) 2275-2279.

A.: Glutathione-dependent

synthesis

of deoxyribonu-

A.: Regulation

of ribonucleotide

Cold

Spring

J. Molecular Harbor

163

Cloning.

Laboratory,

NY 1982.

Messing, J. and Vieira, J.: A new pair of Ml3 vectors for selecting either

DNA

strand

B., Uhltn,

of double-digest

M., Josephson,

son, L.: An improved

reductase.

Curr.

positive

by oligonucleotide

restriction

fragments.

S., Gatenbeck, selection mediated

S. and Philip-

plasmid

vector con-

mutagenesis.

Nucl.

Acids Res. 11 (1983) 8019-8030. Nygren,

H., Rozell, B., Holmgren,

thioredoxin

microscopic

A. and Hansson,

localization

H.A.: Im-

of glutaredoxin

and

in Escherichia coli cells. FEBS Lett. 133 (1981)

145-150. Olsson,

A., Moks,

spaced

banding

field-strength

J. Biol. Chem. 254 (1979) 3664-3671.

Top. Cell. Reg. 19 (1981) 47-76.

synthes-

and characterization

donor from Escherichia coli B. J.

E.F. and Sambrook,

Manual.,

munoelectron

Proc. Natl. Acad.

P.: Enzymatic

D. and Fuchs, J.A.: Cloning and nucleotide

T., Fritsch,

A Laboratory

Nilsson,

correlation

11(1983) 2237-2255. Holmgren,

Rev.

of the trxA gene of E. coli K-12. J. Bacterial.

structed

D.K. and McClure,

Annu.

Gene 19 (1982) 269-276.

35 (1981) 365-403.

expressivity.

membranes.

IV. Isolation

the hydrogen

Cold Spring Harbor,

129 (1977) 967-972.

T., Shinedling,

G.: Translational

in

across

Biol. Chem. 239 (1964) 3436-3444.

Maniatis, deficient

Sci. USA 80 (1983) 687-691.

T.C., Moore, E.C. and Reichard,

(1985) 311-316.

J. Bacterial.

G.N.:

50 (1981) 317-348.

of thioredoxin,

7055-7074.

Holmgren,

Biochem.

N-terminal

D., Schneider,

Gouy, M. and Gautier,

Holmgren,

Kreil, G.: Transfer

from

154 (1986) 193-196.

nu. Rev. Microbial.

Hawley,

genes ofEsche-

sequence

protein

assay

82 (1977)

for use of rare

is ofdeoxyribonucleotides,

112 (1981)

Biochem.

Evidence

H. and von Heijne, G.: Sequence

of an Escherichia coli mutant

reductase.

Gold, L., Pribnow,

with

W. and Godson,

Lim, C.-J., Geraghty, B., Jornvall,

Anal.

codons in the dnaG gene and other regulatory

P.: A

295-298. Flinta, C., Persson,

extracts.

362-371. Konigsberg,

Laurent,

1645-1655. Dretzen,

R.W.: A new spectrophotometric

in cell

richia coli. Proc. Natl. Acad.

1 (1982) 257-262.

G. and Cortese,

stranded

H.: By-pro-

during

protein

83-90. Rosenberg,

T., Uhlen, pattern

gradient.

M. and Gaal,

in DNA sequencing J. Biochem.

A.B.: Uniformly gels by use of

Biophys. Meth. IO (1984)

M. and Court, D.: Regulatory

sequences

involved

in

21 the promotion and termination of RNA transcription. Annu. Rev. Genet. 13 (1979) 319-353. Sanger, F., Nicklen, S. and Coulson, A.R.: DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74 (1977) 5463-5467. Shine, J. and Dalgarno, L.: The 3’-terminal sequence of Escherichiu coli 16s ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites. Proc. Natl. Acad. Sci. USA 71 (1974) 1342-1346. Tinoco Jr., I., Borer, P.N., Dengler. B. and Levine, M.D.: Improved estimation of secondary structure in ribonucleic acids. Nature New Biol. 246, (1973) 40-41.

Wallace, B.J. and Kushner, S.R.: Genetic and physical analysis of the thioredoxin (trti) gene of Escherichiu coli K-12. Gene 32 (1984) 399-408. Wallace R.B., Schaffer, J., Murphy, R.F., Bonner, J., Hirose, T. and Itakura, K.: Hybridization of synthetic oligodeoxyribonucleotides to (PX 174 DNA: the effect of single base pair mismatch. Nucl. Acids Res. 6 (1979) 3543-3557. Communicated by S.R. Kushner.