Effects of second-codon mutations on expression of the insulin-like growth factor-II-encoding gene in Escherichia coli

&V?C, 9x ( I99 I) 217-223 217

Eisevier

GENE

03872

Effects of second-codon Escherichia coli (Recombinant

DNA;

mutations on expression of the insulin-like growth factor-II-encoding

cassette

mutagenesis;

transcription

and translation;

Amanda S. CantrellS, Stauley G. Burgett=, James A. Cwkb,

trypto~han

oxidative

gene in

cleavage)

Michele C. Smith b and Hansen M. Hsiung il

‘IDepartment qf Molecular 3iolog.v md ” Biochemistry Research, Lilly Research Laboratories, Eli Lilly mzd Compmy. Indiunapoiis. IN 46285 (U.S.A.) Received by R. Wu: 22 June 1990 Revised: 11 September 1990 Accepted: 12 September 1990

SUMMARY

Expression plasmids encoding random sequence mutant proteins of insulin-like growth factor II (IGFII) were constructed by cassette mutagenesis, to improve the efficiency of IGFII synthesis in Escherichia coli. A pool of oligodeoxyribonucleotide linkers containing random trinucleotide sequences were used to introduce second-codon substitutions into the gene encoding Met-Xaa-Trp-IGFII in expression vectors. E. co/i RV308 cells transformed with these vectors synthesized IGFII at levels varying from O-22y0 of total cell protein. This variable synthesis is a function of the random second-codon sequence and its corresponding amino acid, Xaa. Our data showed that mRNA stability, protein stability and translational efftciency all contributed to variable expression levels of Met-Xaa-Trp-~~Frr in E. co&. Furtherm~)re, an efficiently synthesized IGFII mu&ant protein, diet-His-Trp-IGFII, was converted to natural sequence IGFII by a simple oxidative cleavage reaction.

INTRODUCTION

Insulin-like growth factor II (IGFII) is a small, mitogenic 67-aa polypeptide closely related to insulin and IGFI (Blundell and Humbel, 1980). Although the biological functions of insulin and IGFI are largely known. the function of IGFII in animds is still unclear. IGFII expresCwrespondence10: Dr. H.M. Hsiung, Department Lilly Research (U.S.A.)

Laboratories,

Tel. (317)276-4609;

Abbreviations:

encoding

IGF;

phoiino]propanesulfonic otide(s);

oligo,

electrophoresis; streptomycin; aa.

Biology, IN 46285

Fax(317)276-1414.

aa. amino acid(s); bp, base pair(s); Cm, chloramphenicol;

DMF, dimethylformamide; ance liquid chromatography; (DNA)

of Molecular

Eli Lilly and Co., fndianapoiis,

DTT, dithiothreitol; HPLC. high-performIGF, insulin-like growth factor; IGF, gene

kb, kilobase

or 1000 bp; MOPS,

acid; NCS, N-chlorosuccinimide;

oii~odeoxyr~bonucleotide; Rif, rifampicin;

Tc. tetracycline;

SDS,

PAGE, sodium

tsp,transcription

3-[N-mor-

nt (Nj, nucle-

polyacrylam~de-~ei

dodecyl

sutfate;

start point(s);

Sm,

Xaa, any

sion is less dependent on the level of growth hormone than that of IGFI (Zapf et al., 1981). It is an anabolic agent in vivo although apparently less potent than IGFI (Schoenle et al., 1982; Shaar et al., 1989). IGFII is synthesized in fetal rat tissues, whereas the levels of IGFII expression are much lower in adult tissues (Zapf et al., 1981). These observations have led to the suggestion that IGFII is important in fetal growth and development. More recently, IGFII has been found in high concentrations in human bone (Frolik et al., 1988). Together with the anabolic activities of IGFII, these results suggest potential applications for IGFII in wound healing, as an adjunct for parenteral/enteral nutrition, in fracture healing, and in reversing other catabolic states. We were, therefore, interested in efficiently producing large quantities of IGFII, to assess these potential applications. Human IGF%f cDNA has been cloned (Bell et al., 1984) and the protein has been previously produced in E. co& as a &>LEl-Met-IGFII fusion protein (Furman et al., 1987).

Xbal-CTAGAGGGTATTACATATGNNNTGGGCTTAT-Taal

expression. Cassette mutagenesis (Wells et al., 1985) was therefore used to introduce a ‘leader’ sequence ‘ATGNNN-TGG’ to the 5’-terminal coding region of the IGFII gene with NNN trinucleotide capable of coding for any aa. Using this approach, we constructed many mutant plas-

Fig. 1. The IGFII

expression

plasmids.

The plasmid

pBR322 derived vector, was used as the starting mutant

plasmids.

Using enzymatic

ligation,

mutant

plasmids

were generated

by replacing

spanning plasmid flanked

by XbaI and ToyI cohesive

two complementary

sequence

a short

DNA

IGFII

fragment

oligo linkers.

ends, were generated

These

linkers,

by annealing

The plasmid

(Applied

Biosystems)

by incorporat-

RESULTS

pPRO-IGFII

i. repressor,

contained

a synthetic

was under the control

and an E. colilpp ribosome-binding

site sequence.

Met-Pro-

of a A:p,_ promoter

A ~I857 gene, encoding

was also present in the plasmid to regulate

of the /I p, promoter

To increase the efficiency of IGFII production, we attempted to express the ZGFZI gene as a nonfusion protein. Studies of foreign protein expression in E. coli (Hsiung and MacKellar, 1987; Schoner et al., 1984) have shown that the first codon after the start codon is critical for efficient gene % IGFII A

0

‘P

Arg(CGT) Arg(CGA)

Expression 2p

as a function

work and has been described

previously

(a) Expression of various IGFIZ mutants in Escherichia coli RV308 Mammalian genes encoding small-M, proteins such as IGFII are not usually expressed efficiently in E. coli. Many factors may be responsible for this phenomenon. Small foreign proteins tend to be degraded rapidly by an ATPdependent protease encoded by the E. coli lon gene (Chung and Goldberg, 1981; Goff et al., 1984; Gottesman et al., 198 1). In addition, translational efficiency of mRNA (Schoner et al., 1984; Iserentant and Fiers, 1980), turnover of mRNA (Donovan and Kushner, 1986; Nilsson et al.,

B

30

21.8 (pHS 246)

Leu(TTA) Met(ATG) Gln(CAA) m His(CAT) 2 Trp(TGG) E Val(GTA) 8 Leu(TTA) i Ser(AGC) Phe(TTT) Ser(TCC) His(CAC) Arg(CGG) Ala(GCA) Pro(CCG)

Fig. 2. IGFII expression

AND DISCUSSION

of each nt, dA. dT, dG or dC, at the random

IGFII gene whose expression

the activity

all

oligo pools. These two oligo pools were synthesized

ing equal molar amounts

a thermosensitive

a

to construct

gene in the pPRO-IGFII

with a pool of double-stranded

on a model 380B DNA synthesizer position.

the random

the 5’ end of the Met-Pro-IGFII

pPRO-IGFII.

plasmid

mids expressing various Met-Xaa-Trp-lGFI/ genes where Trp is the cleavage site for selective oxidative cleavage and Xaa is a random aa residue. To understand how IGFII gene expression was affected by the random sequence mutations, we investigated the effects of these mutations on transcription, translation, protein stability and mRNA turnover. We also performed in vitro run-off transcription experiments to compare the transcription rates of various mutant IGFII genes.

M

K

W

ATG AAG TGG

-lGFII

M E W ATG GAG TGG

--IGF”

M ATG

--IGF”

P w CCT TGG

M P W ATG CCA TGG

ofthe random

aa and its corresponding

-lGFII

IO pg Tc:‘ml. The cultures

(pHS 286)

codon. (A) E. coli K-12 RV308 was used as a host strain for all expression

and MacKellar, 1987). The E. coli cells harboring mutant overnight at 3_7’C in TY medium (Rao et al., 1987) containing

(Hsiung

a single transformation and were grown inoculated 1: 50 into fresh TY (2 ml) containing

(pHS284)

were maintained

IGFII plasmids were randomly selected from IO pg Tc/ml. The overnight culture was then

at 32°C in a shaker-incubator

until absorbance

(A,,,,) reached 0.4. The temperature of incubation was then raised to 41 “C to inactivate the cl857-encoded repressor, thereby inducing expression. IGFII protein accumulation reached the highest lcvcl3 h after induction. The cells were then harvested, pelleted and resuspended

at 550 nm IGFII gem in a protein

sample buffer (2”,, SDS/30”, glycerol;‘1 M 2-mercaptoethanol/h M urea/O.125 M Tris HCI pH 6.8) for O.l”‘, SDS-15 % PAGE analysis (Laemmli, lY70). The gels were stained with Coomassie blue (BioRad) and analyzed with a Shimadzu 910 gel scanner on line with a Hewlett-Packard 2100 computer that integrated the areas under the peaks. Mutant IGFII cxprcssion levels were determined by this densitometric scanning of stained gels and were represented as a”,, of total cellular protein. (B) Four additional IGFII mutants which did not yield a visible band of IGFII protein on a Coomassie blue stained MKW,

gel wcrc listed. The gene sequences for Met-Xaa-Trp-IGFII mutants MEW and MPW are single-letter abbreviations for three N-terminal

were determined by the chain termination method aa encoded by Met-Xaa-Trp-IGFII gems.

of Sanger et al.

( 1977).

21Y 1987), strength of promoters and the ribosome-binding sites (Reznikoff and McClure, 1986; Stormo, 1986), and copy

t1/2

(min)

numbers of the expression plasmids can all contribute to the low expression of small foreign proteins. Earlier study of foreign gene expression in E. cali has shown that expression levels can be increased by using A + T-rich codons in the 5’-coding region of the gene (Hsiung and Mackellar, 1986). The Met-Pro-ZGFZZ genes containing both E. coli-preferred codons and A + T-rich codons were synthesized (data not shown). However, these genes were not e~ciently expressed in the E. cob RV308 host strain, yet the genes were expressed efficiently in several E. coli protease (ion) deficient mutant strains (data not shown). Unfortunately, these protease deficient strains do not grow to a high cell density in large fermentors. To increase expression levels in RV308, we performed cassette mutagenesis using a pool of oligo linkers containing random trinucleotide sequences (Fig. 1 and the legend). Using these linkers, we generated Met-Xaa-Trp-ZGFZZ expression vectors and the transformed mutant RV308 hosts. Of the 60 mutants randomly selected from a single transfor~nation, only 16 expressed detectable levels of IGFII as measured by densitometry of stained gels with the detection limit approx. 1“4, of the total cellular protein (Fig. 2). The yield of mutant IGFII production from 16 producer clones varies from 3-22% of total cellular protein. For all 16 producers as we11 as four non-producers of IGFII, the identity of the random aa was determined by dideoxysequencing of the mutant plasmid DNAs (Fig. 2). The sequencing results also showed that we had isolated only two identical (Met-Arg-Trp-IGFZZ) plasmids out of the 21 plasmids sequenced. This low incidence of identical plasmids suggests that the 60 mutants randomly chosen represent a majority of the 64 possible permutations at the second codon of Met-Xaa-Trp-ZGFZZ genes.

10’

0’

5’ lo’

2’

0’

5’

2’

20’

pHS 278 MHW-IGFII (CAT) 19%

10’20’

5’ 10’20’ %mhr

Fig. 3. The post-translational induced

1 =lO

pHS 280 MAW-IGFTI (GCA) 8%

=4

pHS283 MHW-IGFII &AC) 10%

=lO

protein

E. cnii RV308 cells harboring

stability

of lGFl1

the mutant

IGFIf

mutants.

Fully

plasmids

were

labeled for 2 min with L-[“Slcysteine

(20 pCi,lmmol,

the addition

I mg/ml. Aliquots of culture were

of unlabeled

I.-cysteine to

NEN) followed by

taken at 0, 2, 5, 10, and 20 min after the addition of unlabeled L-cysteine PAGE under reducing conditions and analyzed by 0.1”: SDS-15% (Laemmli,

1970). The gels were then dried onto Whatman

and auLoradiographed. densitometric scanner

The amount oflabeled

scanning

The constructs

by Met-Xaa-Trp-IGFII

by

using a 910 Shimadzu

These values were used to calculate

IGFII proteins.

MRW, MHW and MAW are single-letter minal aa encoded

3MM paper

protein was determined

of the autoradiograms

on line with a computer.

half lives of mutant

TABLE

(b) Stabitities of IGFII mutant proteins The contribution of post-translational protein turnover to ZGFZZ expression was explored using pulse-chase analysis with L-[ “Slcysteine and unlabeled L-cysteine. By this analysis, the half-lives of Met-Xaa-Trp-IGFII mutants were determined (Fig. 3). In general, the half-lives of mutant IGFII proteins were longer (> 10 min) for a high producer (pHS246) and shorter (4 min) for a low producer (pHS280). However, for the cells harboring the plasmids pHS278 and pHS283, the half-lives were the same (10 min) but the production levels differed by twofold. After examining more mutant IGFII proteins produced in E. coli cells, we concluded that although there was some correlation between expression level and protein stability. protein turnover alone could not account for all the differences in ZGFZZ expression. Some specific examples arc given in Table I which listed

2’

the

are shown in Fig. 2A.

abbreviations

for three N-ter-

genes.

I

Met-Xaa-Trp-IGF1I

expression

Codon

*<,IGFI

vs. codon

I

preference

and tRNA

Relative

E. w/i codon

tRNA

prcfercnce”

levels

content .’ Arg

His

Pro

CGI? CGA

22 21

Major Major

Nonpreferred

CGG

9

Minor

Nonpreferred

CAU

1Y

Minor

Preferred

CAC

10

Minor

Preferred

CCG

3

Major

Preferred

CCU

NPh

Minor

Nonpreferred

CCA

NP

Major

Nonpreferred

.’ The tRNA content and codon and Kastelein (IY86). h NP, nonproducer

of mutant

preference IGFII.

Preferred

were cornplIed

by De Boer

220 three Met-Xaa-Trp-IGFII

mutant

proteins

with Xaa en-

coded by synonymous codons. These identical mutant proteins were produced in E. coli with varied efficiency. For example, Met-Arg-Trp-IGFII in which ‘Arg’ was encoded by CGU and CGA codons was produced at higher levels (22 and 21%) than the same protein in which ‘Arg’ was encoded

by CGG

(9%). The same was true for Met-His-

Trp-IGFII in which ‘His’ was encoded by two synonymous codons (CAU, 190;, vs. CAC, IO?;) and for Met-Pro-TrpIGFII in which ‘Pro’ was encoded by three synonymous codons (CCC, 3:/, vs. CCU, < I :/, or CCA, < I?‘,). These results suggested that factors other than the post-translational protein stability must affect IGFIl expression. Table I also compares the expression levels of mutants substituted with synonymous codons as a function of tRNA abundance and E. coli codon preference (De Boer and Kastelein, 1986; Ikemura, 1981). It was shown that tRNA abundance or codon preference alone did not correlate completely with the IGFfl expression levels. However, a preferred codon in combination with an abundance of tRNA [i.e., CCG(Pro) vs. CCU or CCA(Pro)] increased expression. Conversely, a rare codon for which there was a low level of tRNA resulted in lower IGFII expression levels [i.e., CGG(Arg)]. Looman et al. (1987) have also reported that second codon variation can affect IucZgene expression by as much as 15fold. When we compared their results with ours, we found that the codons that caused high-level LacZ production differed from the ones responsible for high-level IGFII production. For example, CGA used in their study results in only one-ninth the protein yield of CGT, yet we found no difference between these two synonymous codons in mutant IGFII production (CGA vs. CGT, Fig. 2). The discrepancy in results was probably due to the differences in the genetic backgrounds in these E. co/i expression studies. We concluded that the preferred codons for highlevel production of the protein had to be individually determined for each gene in different host/vector systems. With the convenient SDS-PAGE protein gel assay that we described in this report, random mutagenesis was probably the best method to discover optimal codon choices for IGFII protein production. (c) Messenger RNA stability Although mRNA secondary structures can affect translational efficiency (Iserentant and Fiers, 1980), we found it was difficult to correlate mutant IGFIZ gene expression with random nt changes by analyzing hypothetical mRNA secondary structures. Instead, we studied the turnover of mutant mRNAs using the transcriptional inhibitor Rif. Individual Met-Xaa-Trp-1GFII mRNA stabilities were determined by measuring the amounts of mRNA surviving 0, 2,4 and 6 min after the addition of 200 pg Rifiml to fully

induced cultures. One-ml cell culture aliquots were taken and frozen immediately on dry ice to quench cell metabolism and their RNAs isolated. Whole-cell E. coii RNA was isolated by the following technique. Aliquots (1 ml) of induced-cell cultures were pelleted, then resuspended in a 7504 solution of 0.2u/, SDS/l0 mM EDTAjlOO mM Tris . HCl pH 8.0/50 mM NaCl. These suspensions were heated to 95°C for 5 min. then allowed to cool to 37°C. The samples were then digested with proteinase K (BRL, 50 pg,/ml) for 1 h at 37 “C. Each reaction mixture was extracted three times with phenol~chloroform (1: 1, v/v) and once with chloroform before precipitation in 2 ~01s. of ethanol for 1.5 min at -20’ C. The dried pellets of nucleic acids were then digested at 37°C for 20 min with RNase-free DNase 1 (BRL, buffer containing 50 mM 20 pg/ml) in a reaction Tris . HCl/l mM EDTA/S mM MgSO, pH 7.5. These mixtures were extracted with phenol/chloroform and precipitated with 2 ~01s. of ethanol. The RNA pellet was resuspended in 20 ~1 of IO mM Tris . HCl/ 1 mM EDTA buffer pH 7.5. Final concentrations of RNA were determined from A 2h0 measure~lents. All mRNAs were characterized using Northern-blot analysis as described by Maniatis et al. (1982). IGFfI mRNAs were quantitated by densitometry ofthe autoradiographs and the half lives of mRNAs were calculated from the densitometry results. The corresponding autoradiograms (data not shown) showed that the stability of MetXaa-Trp-ZGFII mRNAs correlated well with their levels of

pHS 246 ARK-IGFII ‘2c2G,T’

pHS 247 ~~~-lGFII y&y

0’ 4’O8’

0’ 4’*8’

0’ 4’ 8’

0’ 4’ 8’

A

0’

4’

8’

6

Fig. 4. Effect of protecting Fully induced E. after temperature the addition

ribosomes

on mutant

IGFII

mRNA stability.

harboring the indicated mutant plasmids (I h shift to 41 “C) were harvested at 0, 4 and 8 min after

co/i cells

of IO0 pg Sm/ml, 200 pg Rif/ml (panel B) or at the same time

intervals after the addition of 100 JL~Cm/ml, 200 pg Rifiml (panel A). ‘rhc MXW-IGFII RNAs were isolated and Northern-blot analysis was run as previously described. MRW, MMW, MPW and MXW represented N-terminal aa encoded by the mutant lGFlI genes.

three

221 expression. The approximate half lives for the mutant mRNAs were greater than or equal to 4 min for high pro-

(d) In vitro transcription In vitro run-off transcription

ducers (pHS246 and pHS278) and less than 4 min for low producers (pHS280 and pHS283). The effect of protein translation and ribosome attachment on mRNA stability were explored indirectly by using a translational inhibitor (Cm or Sm) combined with a transcriptional inhibitor (Rif) in induced mutant E. coli cultures.

of transcriptional efficiency in maintaining steady-state mRNA levels. The data (Fig. 5) indicated that three MetXaa-Trp-ZGFII mutant templates transcribed at similar rates in vitro regardless of their different expression levels in vivo. This result also suggested that the rate of transcription was not affected by minor changes in the 5’-coding region of the mutant IGFII genes. The run-off transcript of

At the concentrations used, Cm should stall ribosomes on the mRNA, while Sm should dissociate ribosomes from the mRNA (Pestka, 1977). Whole cell RNA was isolated 0, 4, and 8 min after antibiotic addition and subjected to Northern-blot analysis (Fig. 4,A and B). The results show that mRNA from Cm-treated cells (Fig. 4A) is more stable than mRNA from Sm-treated cells (Fig. 4B). This suggests

the IGFII gene also showed doublet bands (Fig. 5) of IGFII transcripts, suggesting possible multiple tsp. (e) Conversion of Met-X-Trp-IGFII cleavage

pHS246 MRW-

‘888- -

603-

pHS271

pL11OC bGH

SP6

--

to IGFII by oxidative

One of the efficiently expressed mutants, Met-His-TrpIGFII, was subjected to a tryptophan oxidative cleavage reaction to remove the tripeptide extension and generate natural sequence IGFII. Tryptophan oxidative cleavage reactions have been applied to proteins to generate smaller fragments suitable for further sequence analysis (Schechter et al., 1976) and more recently to recombinant fusion proteins to liberate the mature protein of interest (Villa et al., 1988). Human IGFII contains no tryptophan residues and

that dissociation of ribosomes from mRNA (Sm-treated cells) leads to greater message instability than simply arresting mRNA on ribosomes (Cm-treated cells). Presumably the free mRNA species released during Sm treatment are more susceptible to nucleases than the mRNA still bound to ribosomes. This may explain why the efficiently translated mRNA, which would be protected by ribosomes, was more stable than the less efficiently translated mRNA.

4x174 Hae III

was used to explore the role

MPW-

pHS286 MPW-

(CCG) 3%

(CCA) NP

bGH-

422f43_ -

- 230-

234-

MXW-IGF

II

194-

iia- C

72-

Fig. 5. Transcriptional essentially polymerase. EDTA/O.I

efficiency

of IGFIl

mutants

with the procedure

of Gardner

The templates

were transcribed

mM DTT/4 mM Mg. acetate).

250 pg/ml. The RNA transcripts acid, dried and autoradiographed. included

were precipitated

on an 8”, polyacrylamide

samples.

pg) ofBamH1

experiments.

linearized

In vitro run-off

Met-Xaa-Trp-IGFII

in 50 ~1 transcription

transcriptions

plasmid

buffer (200 mM Tris

GTP, CTP and ATP were added to a final concentration

were carried

as template acetate

out

for E. coli RNA pH 7.9/l mM

of 1.5 mM each. Transcripts

were

(800 Ci/mmol, NEN). The reactions were incubated at 37°C for IO min followed by addition of heparin incubated for an additional 10 min, then stopped by the addition of Rif to the final concentration of in two volumes

gel containing

In vitro transcription

of ethanol

8 M urea in 1

x

after the addition

TBE buffer (Maniatis

of the bovine growth

as a positive control. Two nucleic acid size markers

transcription

by run-off transcription

with 1 unit of E. co/i RNA polymerase

The ribonucleotides

labeled by the addition of 50 PCi of a-[“P]UTP (10 pg/ml). The reaction mixtures were further electrophoresed

as measured

(1982). We used 1 pmol(0.3

($X 174 DNA/HaeIII

hormone

(bGH)

of 10 pg of tRNA

carrier.

The RNA samples

were then

et al., 1982). The gel was fixed in 5”” methanol/5”~0 gene in an identical

digests and SP6 transcripts)

expression

plasmid

acetic

(pL11OC) was

were run at the same time with the in vitro

222 is therefore an excellent candidate for using tryptophan oxidative cleavage chemistry to generate natural sequence IGFII. A tryptophan oxidative cleavage reaction was carried out on granules isolated (Schoner et al., 1985) from E. coli cells. Dried granules (500 mg) were dissolved in 50 ml of 50”; acetic acid/3.75 M urea with stirring at room temperature for 3 h. Three 2-ml aliquots of a lOO-mM solution of NCS (iildrich) in DMF were added 20 min apart for a total reaction time of 1 h. Methionine (9 ml of a 200 mM solution in SO?; acetic acidl3.75 M urea) was added to stop the reaction. The samples were dried in vacua, converted to the S-sulfonate and analyzed as described (Furman et al., 1987). The IGFII-S-sulfonate was isolated with a 357” yield and contained a low level of cysteic acid (1 Oo) as a side product. The modification of the published procedure used in this work involved the use of less harsh conditions with a comparable overall yield. IGFII prepared from the tryptophan oxidative reaction was refolded to its native conformation by a published procedure (Smith et al., 1989). I-IPLC analysis (data not shown) of folded IGFII prepared with the cleaved material showed that it was identical to an IGFII standard that has the native conformation and correct disulfide bonds (Smith et al., 1989).

ACKNOWLEDGEMENTS

The authors thank G.W. Becker, S. Kuhstoss, G. Chan, S. Kaplan and B.E. Schoner for their critical reading of the manuscript.

RbFERENCES Bell, G.I., Merryweather,

R.. Stcmpien,

L., Scott. J. and Rail. L.B.: Sequence of a cDN.4

encoding

human

preproinsulin-like

growth

factor

II.

.

M.M

clone

Nature

3 IO

families: pancreatic

hor-

(1984)775-777. Blundell,

T.L.

and Humbcl,

R.E.:

Hormone

mones and homologous growth factors. Nature Chung. C.H. and Goldburg, Exherichia

di

its role

in optimizing

I., (Eds,), RIA,

protensc, protcasc

R.A.: Biased codon usage’ an explanation

of translation.

~~axilllizin~

Gene

W.P. and Kushncr.

ribonucleasc Ekherichio

In: Reznikoif,

Expression.

W. and Gold.

Butteworth.

Stoneham.

S.R.: Pol~l~ucl~otide p~~~)sp~~~~r~l~i~c and

II are required for cell viability and mRN.4

twn of insulin-like Furman,

turnover in

coii K-12. Proc. Natl. .4cad. Sci. USA 83 (19x7)

Frollk, C.A., Ellis, L.F. and Williams. grov,th

Res. Commun.

I.C..

Epp.

J.,

Mendelsohn.

L.G..

Recombmant

human

D.C.:

factor

Hsiung,

HoskIn\,

B., Smtth,

insulin-hke

J..

D.P.

growth

and

factor

II

Lens,

(;.I...

Smith,

M.(‘.:

expressed

ln

5 ( 1987) lO37- IO5 1.

J.t.: Isolation. pausing, and tcrmmation

of transcrIption

m the

cok J. Bid ~‘hent.

threoninc operon regulatory region ofE.sciwr~chi~ 257 (1982)

bone. Biochcm

101 I-101X.

Il.hl..

Schoncr.

1X-114.

Isolation and characterlza-

II from human

151 (19X8)

Escherichicr cnli. Bio:Technology

Gofi

1,;~. Proc

1986, pp. 215-285.

Donovan,

Gardner,

78 I-7X7.

78 (1981) 4931-4935.

De Boer. H.A. and Kastelein,

of’

787 (lY80)

A.L.: The product ol’thc /otz (r,cl/>R) gcnc 1n

is the ATP-dependent

Natl. Acad. Sci. USA

Biophq,.

(f) Conclusions Our study showed that the expression levels of IGFII mutants varied as a function of the random aa and its corresponding codon NNN at the second position of mutant IGFII genes. However, we could not correlate this variability to any obvious physicoche~lical properties of the substituted aa such as hydrophobicity, hydrophilicity or aromaticity. We also observed that the random mutation of the Met-Xaa-Trp-IGFZZ gene in the second codon could affect IGFII production in multiple ways. (1) Some of the mutant Met-Xaa-Trp-IGFII proteins were produced at various levels due to different post-translational protein stabilities. (2) The mutant IGFZI genes had similar transcriptional rates in vitro, regardless of the drastic differences in expression levels in vivo. This result suggests that traIlscriptiona1 efficiency of mutant fGFlI genes probably is not responsible for different levels of expression. (3) The mutant Met-Xaa-Trp-IGE;II mRNAs have different half lives, suggesting that mRNA stability plays a role in expression. The mRNA turnover study showed that the mutant IGFZI mRNAs attached to ribosomcs and stalled translationally were more stable than those released from ribosomes, suggesting that IGFZI mRNA stability may depend on the transiational efficiency. (4) A tryptophan oxidative cleavage procedure was developed, resulting in the conversion of Met-His-TrpIGFII to native IGFII.

J.P., Scenchez-Pescador,

Priestley.

3896-3904.

S.A., Cassun, L.P. and Goldberg,

hrpR influcnccs rates ofprotein gene in Eschrrichirr

to/i.

Proc.

I. Gottesman,S.. Gottesman.M..

A.L.: ItIcat shock regulator> fcnu

degradation Natl.

and expression of the 11~

.4cad.

X1

Sci. USA

( 19X-1)

6647-665

degradation

Shaw, J.E. and Pearson, M.L..

in E. m/i: the IO/I mutation and bactcriophagc

and cell protein stability. Ccl1 23 Hsiung, H.M.

and MacKellar,

mow derivatives

Protein

lambda N

( IL)8 1) 115-233

Cf:.c‘.: Expression of bovmc growth hor-

in Escherir~l~irrco/i and the use of the dcrivatlvcs

to

produce natural sequence growth hormone bq catbepsin c‘ clca\agc. Methods lkcmura.

Enzymol.

153 ( 19X7) 3’10-30 I.

T.: Correlation

the occurrence Biol. 136 (19X1) Isercrrtant.

hetwccn the abundance

of the rcspectisr

cicncq of translation

initiation.

the head of bacteriophage Looman. A.C., Bodlaender,

J., Comstock,

and Van Kmppenbcrg,

Harbor,

NY.

of mRN:1

127 (IY70)

P.H.: Inllucnce

J. h

6X0-685.

L.J.. Eaton. D., Jhurani, P., De

I 1987)

L.1,. and Sambrook,

Manual.

and clti-

I-I 2.

initlatmn codon on the expreaswn

I _, l‘rltsch,

Laboratory

( IWO)

proteins durmg the assembly of

‘1‘4. Nature

gene in ,?‘rchwichici wii. EMBO Mania&,

structure

Gene Y

Lacrnmh, IJ.K.: Cleavage of structural

Boer, H.A.

and

I-21.

I>, and Fiers, ‘v\‘.: Sccondarl

lowing the AUG

of I:. <
codons m it> protein gcncx. 1 Mol.

Cold Spring Harbor

of the codon To-

of a modified

Iw%

248%2JY7.

J., hloiccular Laboratory,

Clvmng.

A

Cold Spring

19X?.

Nilsson. G., Belasco. J.G.. Cohen, S.N. and van Gabam, premature termination

of tr:tnslatinn on mRNA

.4.: Etliict 01

stahllity depend\ on

the site of ribosome

release.

Proc. Natl. Acad.

Sci. USA 84 (1987)

Pestka,

S.: Inhibitors

of protein

synthesis.

In: Weissbach,

S. (Eds.), Molecular Mechanisms of Protein Press, New York, 1977, pp. 468-553. Rao, R.N., Richardson,

M.A. and Kuhstoss,

for cloning and analysis of Streptornyces

H. and Pestka,

Biosynthesis.

S.: Cosmid

Academic

shuttle vectors

DNA. Methods

Enzymol.

153

In: Reznikoff,

W.

W.S. and McClure,

and Gold, Stoneham, Sanger,

W.R.: E. colipromoters.

L. (Eds.), Maximizing MA. 1986, pp. l-33.

F., Nicklen.

terminating

S. and Cot&on,

inhibitors.

Proc.

Gene

Expression.

Butterworth.

A.R.: DNA sequencing Natl.

Acad.

Sci.

with chain-

USA

74 (1977)

E., Zapf,

J., Humbel,

growth factor I stimulates 296

R.E. and Froesch,

growth

L.M.: Insulin-like

in hypophysectomized

rats, Nature

( 1982) 252-253.

Schoncr.

B.E., Hsiung,

H.M., Belagaje,

mone expression

translational

R.M., Mayne, N.G. and Schoner, efficiency

in bovine growth

in Escherichia co/i. Proc. Natl. Acad.

hor-

Sci. USA 81

f 1984) 5403-5407.

growth

human

growth

age of tryptophanyl

and activity

bonds

factor II on disullide 9314-9321. Stormo,

G-D.: Translation

(Eds.), Maximizing

Y.: Selective

15 (i976)

bond

pairing.

initiation.

chemical

cleavwith

5071-5075.

human

J.L.: Structure

insulin-like

J. Biol. Chem.

In: Reznikoti;

Gene Expression.

II

Endocr.

chlorination

T.C. and Occolowitz.

of recombinant

B.:

factor

rats.

by oxidative

Biochemistry

dependence

growth

in hypophysectomized

peptide

,V-chlorosuccinimide.

J.A. and Neubauer,

insulin-like

Res. 15 (1989) 403-411. Shechter, Y., Patchornik, A. and Burstein,

granules

hormone.

Villa, S., DC Fario,

growth

264 (1989)

W. and Gold,

Btntcrworth.

BE.: Isolation

and puri~cation

from E~c~er~e~jffcoli ceils overproducing BiojTechnoIogy

3 (1985)

151-154.

bovine

G., Donini, S., Tarchi,

in E.tchcn’chicr coli and characterization factor.

Wells, J.A., Vasser, cficient

Eur. J. Biochem. M. and Powers,

method for generation

Stoncham,

L.

MA,

G. and Canosi, of human

U.: Expression

growth

hormone

171 (1988) 137-141. D.B.: Cassette

of multiple mutations

mutagenesis:

an

at defined sites,

Gene 34 (1985) 315-323. Zapf, J., Walter,

R.G., Ellis, L.F. and Schoner,

of protein

stimulates

releasing

R.G.: Role of mRNA

Schoner,

(hIGF-II)

DNA-derived

1986, pp. 195-224.

5463-5461. Schocnle,

C.J., Tinsley, F.C., Smith, M.C., Clemens,

Smith. M.C., Cook, J.A., Furman,

(1987) 166-198. Reznikoff,

Shaar,

Recombinant

4890-4894.

H. and Froesch,

nation of insulinlike in patients glycemia.

growth

with growth

E.R.: Radio~mmLlnol(~gical

factors

disorders

J. Clin. Invest. 68 (1981)

I and II in normal and extrapancreatic 1321-1330.

determi-

subjects tumor

and hypo-