Intermediates in the synthesis of TolC protein include an incomplete peptide stalled at a rare Arg codon

Eur. J. Biochem. 152, 151 - 155 (1 985) $3 FEBS 1985

Intermediates in the synthesis of TolC protein include an incomplete peptide stalled at a rare Arg codon Rajeev MISRA and Peter REEVES Department of Microbiology and Immunology, Faculty of Medicine, The University of Adelaide

(Receivcd May 28, 1985) - EJB 85 0566

TolC is a minor outer membrane protein of Escherichia coli K 12 and is initially synthesized as a precursor. A distinct intermediate polypeptide of M , about 46000 was consistently observed at the initial stages of biosynthesis. The further elongation of this peptide can be blocked by chloramphenicol. We have investigated the cause of the temporary accumulation of the 46000-M, intermediate and we postulate that the presence of a rare codon AGA (Arg) at codon 402 of the tolC mRNA halts translating ribosomes owing to a limiting amount of the tRNAArg (AGA) species in the cell. The translation of tolC mRNA can be increased by providing T4 tRNAArg(AGA), encoded on a plasmid.

TolC is an outer membrane protein of Escherichia coli K12 [l]. It has been established that proteins exported by E. coli K12 to the periplasmic space and outer membrane are synthesised as precursor forms containing an NH,-terminal extension, the signal sequence, of about 20 - 25 amino acid residues [2]. We have previously reported the DNA sequence of the tolC gene [3] and showed that the NH2 terminus of the mature protein is preceded by a typical signal sequence of 22 amino acid residues. The precursor form of the TolC protein was also detected in a pulse-chase experiment [3]. In this paper we report a detailed study of TolC protein synthesis in E. coli K 12 strains, which carry the tolC gene in multicopy plasmids.

MATERIALS AND METHODS

Strains and growth conditions

The strains used in this study are listed in Table 1. Cells were grown in double-strength Difco broth (Difco, 0003) with the addition of 5 g/l NaCl or in M 9 minimal salt medium [4] supplemented with growth factors. Isolation of whole cell envelopes was carried out as described by Morona and Reeves [5]. Plasmid DNA was purified either by a quick Triton lysate

Table 1 . Bucterial strains used in this study

Strains Relevant properties Source/ reference

AB 11 33

P602 AB1133 tolC203 P. Reeves

P2948 C600 (GM4113) source of cloned tRNAArp [14] P2964 P602 (pPR258) this study P2970 P2964 (GM4113) this study P3025 P2964 (pBR322) this study

F- thrl leu6 pro A 2 lac Y I supE44 galK2 his4 rpsL31 xy15 mtll argE3 thil ma14

A. L. Taylor

P2715 AB1133 (pPR42) P I

Correspondence to R. Misrd, Department of Microbiology, The

Abbreviation. SDS, sodium dodecyl sulfate. University of Sydney, Sydney, New South Wales, Australia 2007

method [6] or by the isopycnic CsCl density gradient method t71.

Plasmid constructions

A general scheme of plasmid construction is shown in Fig. 1. Plasmid pPR42 contains a PstI-EcoRI fragment of E. coli K 12 carrying the tolCgene in pBR322 [I]. Other plasmids, pPR257 and pPR258 were made by subcloning in each orien- tation the tolC piece from pPR178 into pACYC184. Equal amounts of TolC protein were produced by strain P602 (tolC-) carrying pPR257 or pPR258.

Acc! AVAI

Fig. 1 . Diugrammutic construction of plasmids used in this study

152

Experiments with labelled cells

a ) f 35SjMethionine labelling und pulse chase. Strains for pulse-chase studies were all based on AB 1133 and were grown at 37°C in 10 ml M9 minimal salts medium supplemented with L-arginine, leuci cine, L-histidine, L-proline, L-threonine (20 pg/ml each), thiamine (1 pg/ml) and glucose (0.5%). At Asoo = 0.5, cells were centrifuged and resuspended in one- tenth the volume of fresh M 9 medium and preincubated at 25'C for 5 min prior to addition of 100 pCi [35S]methionine (specific activity of 800 Ci/mmol). The radioactivity was chased with non-radioactive methionine (final concentration 20 mM) after the appropriate pulse period. Samples (100 pl) were taken at various times and quickly frozen at - 70°C. The cells were pelleted at 4°C in a microfuge for 5 min and resuspended in 100 p1 SDS buffer [3] and stored at - 20°C. Samples were heated for 3 min in boiling water for cell lysis prior to immunoprecipitation.

b ) Celljiuctionution qf luhelled cells. A 10-ml culture of P2964 was grown and centrifuged as in (a) and was resuspended in 1 ml fresh M 9 medium containing 0.2 pg unlabelled methionine. Cells were preincubated for 10 min prior to labelling with [3sS]methionine (100 pCi/ml) for 90 s. A 0.1-ml sample was removed to prepare a whole cell extract as in (a) and the remaining labelled cells were used to prepare cytoplasm, whole cell envelope, outer membrane and inner membrane fractions. These fractions were prepared essentially by the method of Ito et al. [8], which involved two sucrose step gradients. Each fraction thus obtained was used separately for immunoprecipitation.

c) Immunoprecipitation. 20 - 50-pl amounts of [35S]methionine-labelled cell lysates or fractions were used for immunoprecipitation [3], after diluting tenfold in Triton buffer (100 mM Tris/HCl pH 7.4, 1 mM EDTA, 0.02% NaN3, 1% Triton X-100).

d) SDS/polyucrylamide gel electrophoresis. Sodium dodecyl sulfate (SDS) gel electrophoresis of the [35S]methionine- labclled samples was carried out on 11% linear polyacrylamide gels as described previously [9]. Proteins were fixed in 7% acetic acid for 30 min, and the gels soaked in Amplify (Amersham) with agitation for 15 - 30 min, rinsed twice with water and dried in a Biolab gel drier. The gels were fluorographed at room temperature for 4 - 5 days using Fuji X-ray film. Fluorographs were scanned in order to quantify the radioactivity which produced the bands. The molecular masses of the bands were determined by comparison with molecular mass standards run on identical gels (or half of the same gel), stained with Coomassie brilliant blue G250 after fixing with 50% trichloroacetic acid.

RESULTS

Pulse-chase experiment using strain P2715

Strain P2715 contains a multicopy toEC plasmid (pPR42) and hence produces a large amount of TolC protein (Fig.2) [I]. P602 (talc-) does not produce any TolC protein and the parent strain, AB 1133, produces very low levels which were not detected in earlier work from our laboratory. A pulse- chase experiment carried out previously [3] showed that the TolC protein was first synthesised as a precursor with a signal sequence. We repeated that experiment using a shorter pulse time (10 s) and found that three major polypeptides, of apparent M , 46 000, 52 000 and 54 500, were immunoprecipitated using TolC antiserum (Fig.3). None of these proteins was

Fig. 2. Outer membrane profile ofAB1133, P402 and P2715 shorr,ing positions o j TolC cind other outer membrane proteins

Fig. 3. Iw?inmno~rec~itation oj'[3sS]methionine-lr~hc~lic~d ToiC trnd its p w ~ ~ i ~ r ~ o r polypeptides. Strain P2715 was pulse-labelled and chased as dcsci-ibcd i n Materials and Methods. Samplcs at indicated times were immunoprecipitated and subjected to polyacrylamide gel electrophoresis. The gel was dried and fluorographed. P indicates precursor and M mature TolC protein

precipitated when a tolC- strain (P602) was used in a control experiment (data not shown).

A polypeptide with an apparent M , of 46000 was the first to appear at 10 s and was still the major band at 25 s but was barely detectable by 40 s. A second polypeptide of M , 54500 (precursor) was immunoprecipitated at 25 s. This polypeptide was most abundant between 40 s and 120 s and disappeared during the chase while a concomitant increase in the amount of mature protein ( M , 52000) was observed.

153

Fig. 4. Fute of TolC biosynthetic intermediates in the presence or uhscxee of chlorumphenicol. Strain P2715 was pulse-labelled with [35S]methionine for 10 s and chased with unlabelled methionine either in the absence or presence of chloramphenicol (to final concentration of 250 pg/ml). Samples, taken at times indicated, were treated as for Fig.3. P, precursor; M, mature protein

Polypeptide of M, 46000 is a biosynthetic intermediate of'the mature TolC protein

The kinetics of appearance of the M,-46 000 polypeptide in the previous experiment suggested that it cannot be a degra- dation product but could be an incomplete TolC peptide, which requires further protein synthesis for completion. The following experiment was conducted to investigate this possibility.

Cells were pulse-labelled for 10 s with [35S]methionine and then chased with unlabelled methionine in the presence or absence of 250 pg/ml chloramphenicol (Fig.4). In the absence of chloramphenicol, the 46 000-M, polypeptide remained as a significant band up to 70 s. In the presence of chloramphenicol the 46000-M, polypeptide was present to a much greater extent at 25 s and, instead of disappearing after 70 s, was still present in the final sample at 200 s. Furthermore, 54500- M , completed precursor was fully chased into mature TolC protein as it was in the absence of chloramphenicol. These results strongly support the hypothesis that the 46000-M, polypeptide is an incomplete TolC protein and not a degrada- tion product.

Cause of the temporary accumulation of the 46000-M, polypeptide

It has been suggested previously that secondary structures in mRNA may cause a temporary pause in the translation process. Thus ompA mRNA is predicted to have about ten possible secondary structures [I 01 ; the several incomplete peptides reported when its biosynthesis was studied [ l l ] were attributed to such an effect. A similar explanation was suggested for the existence of nascent intermediates in the synthesis of the maltose-binding protein [12]. The tolC DNA sequence did not indicate any possibly secondary structure in

Fig. 5. Ej7ijc.t of cloned T4 tRNAArK plasmid on biosynthesis of Talc protein. The strains indicated were treated as for Fig. 3. P, precursor; M, inaturc protein

mRNA [I 31 which could cause a pause in translation to give an incomplete peptide of M , 46000. However, the presence of the rare codon AGA (arginine) at codon 402 of the tnlC mRNA coding region may cause a delay in translation owing to the limiting amount of relevant tRNAArg species in the cell [22, 231. This would result in the temporary accumulation of an incomplete peptide with a relative molecular mass of 46000. To test this hypothesis we have studied TolC biosynthesis in a strain carrying a cloned T4 tRNAArg gene. The tRNAArg was originally cloned in pBR322 [14] and in order to overcome incompatibility we subcloned the t o l e gene into pACYC 184 and constructed appropriate strains (Fig. 1 and Table 1).

The results of a pulse-chase experiment, using strains with or without the cloned tRNAArg plasmid, are presented in Fig. 5. The presence of cloned tRNAArg gene had two effects. Firstly, the precursor (M, 54500) appeared earlier (10 s) and was subsequently processed more quickly such that it was no longer detectable at 120 s. In the control strain the precursor appeared at 25 s, was still a major band at 120 s and detectable at 180 s or later. Secondly, although the amount of the 46 000-M, polypeptide detected at early sampling times was significantly reduced, nonetheless it remained long after the chase commenced. In contrast the strains lacking the cloned t RNAArg gene, including P 3025 carrying pBR322 as a control, exhibited 46000-M, peptide as a major component in the earlier stages of the pulse and chase, but then lost it completely.

Location ofthe 46000-M, polypeptide, precursor and mature TolC proteins

[35S]Methionine-labelled cells were fractionated as described in Materials and Methods. The TolC protein and related polypeptides in each fraction were immunoprecipitated and electrophoresed on a SDS/polyacrylamide gel (Fig. 6). The 46000-M, polypeptide was present in the whole cell ex-

154

Fig. 6. Fmcrionution u f [35S/methioriine-laheil~~d cells. Cells were labelled with 100 pCi [35S]methionine for 90 s. Radioactive isotope incorporation was stopped by adding non-radioactive methionine and sodium azide lo final concentrations of 2 0 m M and 0.02% rc- spectively and freezing the labelled cells at - 70"C. Labelled cells were fractionated into whole cell extract (track l ) , cell sonicate (track 2), whole cell envelopes (track 3). and inner membrane (track 4) and outer membrane (track 5) by two sucrose step gradients. These fractions were then immunoprecipitated separately with TolC antiserum (tracks 6-10 respectively). The gel was dried and fluorographed. P, precursor; M, mature protein

tract (track 6), cell sonicate (track 7), and whole cell envelope (track 8) fractions. Furthermore, i t was present in the inner membrane (track 9) but not in the outer membrane (track 10) fraction. These results show that the polypeptide of M , 46000 is associated with the cytoplasmic membrane. As expected the precursor protein ( M , 54500) was present mainly in the inner membrane fraction (track 9). The mature TolC protein precipitated mainly from the outer membrane fraction (track 10).

We conclude from these observations that the polypeptide of MI 46000 is bound to the inner membrane along with the conventional precursor ( M , 54500). The mature protein is located in the outer membrane.

DISCUSSION

Our results suggest that the synthesis of the mature TolC takes place in at least three distinct steps. A polypeptide of apparent M, 46000 appears within 10 s and is converted with a half-life of about 30 s into conventional precursor ( M , 54500). This precursor is first detected 25 s after labelling and is itself converted into mature protein with a half-life of about 60 s. The mature TolC protein (Mr 52000) is first detected 40 s after the beginning of labelling and the whole of the [3sS]methionine pulse is completely chased into this form by about 180 s.

Our experiments suggest that the 46000-MI polypeptide is an incomplete biosynthetic intermediate of the TolC protein. Several distinct incomplete polypeptides were reported previously for OmpA [l l ] protein; this could be explained by predicted stable secondary structures in ompA mRNA causing a blockage in peptide elongation and accumulation of biosynthetic intermediates. In the case of rolC mRNA there are no stable secondary structures in the main body of the gene. However, the presence of the rare codon AGA (codon 402 of the coding region) is the most likely reason for the temporary accumulation of the 46000-M, polypeptide.

Table 2. The usagr? ofrare codons in E. coli genes The usage of 6 rare codons was examincd. Figures are given for a total of 22 genes which are highly expressed. 16 weakly expressed genes were surveyed as were the genes for 5 regulatory proteins [I 5 - 201. ornp, major outer membrane proteins: genes included are ornpF, onipC, ompA, luniB and Ipp

Rare Genes Major tolC codons On?(, ~~

highly weakly regu- expressed expressed latory

AGA (Arg) AGG (Arg) CGG (Arg) CCC (Pro) GGA (Gly) CUA (Leu)

Total

Average frcqucncy per gene

3 9 3 1 4 6 0 8 14 2 8 13 5 I 13 1 5 I

14 41 56

0.636 2.56 11.2 0 10

The significance of rare codons in regulatory and minor proteins has been studied before [15-201, and recently [21] it has been shown that the insertion in tandem of four of the extremely rare codon (Arg AGG) into the E. coli cut gene significantly reduces the level of expression. For some E. coli proteins the synonymous codons are used in a non-random manner and the codons preferred are those recognised by the most abundant tRNA species in the cell; the concentration of each tRNA and the frequency of usage of the synonymous codons have been listed by Ikamura [22, 231. Codons AGA, AGG and CGG for arginine, CCC for proline, GGA for glycine, CUA for leucine were considered to be rare codons (Table2) and are also not present in the mRNA of major outer membrane proteins: lipoprotein [24], OmpA [25], OmpF [26], OmpC [27], and LamB [28]. However, these codons and rare codons for other amino acids are present in relatively high proportions in minor proteins and regulatory proteins [15-201, which are maintained in the cell in low concentrations. The rare codons in such proteins may cause the translating ribosome to pause as it proceeds along the mRNA, owing to the limiting amount of corresponding tRNA, thereby reducing expression of the gene. Absence of these rare codons from the mRN,4 of the major proteins would facilitate a high level of expression.

In this paper we have presented evidence for such a mecha- nism. TolC is a minor outer membrane protein and has ten amino acid residues encoded by the rare codons listed above. TolC is synthesised in large amounts in strains carrying the tolC gene in multicopy plasmids and we consistently observed a 46000-M, polypeptide at the earlier stages of biosynthesis in such circumstances. This would be expected if translating ribosomes pause at codon 402 (AGA). We have confirmed by our results that the 46000-M, polypeptide is a nascent biosynthetic intermediate of the mature TolC protein and that the rate of synthesis of mature protein can be increased bq providing extra tRNAArg (AGA, AGG). Under this hypothesis one would expect the translating ribosome to pause also at the other rare codons shown in Fig. 7. In our experimental conditions we did not consistently detect any other small peptides. It is possible that the other tRNA species involved are not in such limiting levels as that for tRNAArg (AGA,

155

cec ?<A

'C M M M 203 6 4 1 1 ,A

~ [ ' i 1 2 6 100

288

Fig. 7. Position ofrare codons and methionine residues ( M ) in the tolC gWW

AGG). However, we might at least expect the rare arginine codons at residues 260 and 262 to lead to accumulation of a small peptide of 28000 M,, as the same tRNA species is involved for codons ACA and AGG [29]. It should be noted (Fig. 7 ) that four of the seven methionine residues are encoded between codons 262 (AGG) and 402 (AGA). Therefore, one would expect the translating ribosomes to synthesise 46000- M , polypeptides, which are labelled to more than twice the specific activity of 28000-M, polypeptides. Furthermore, ribosomes which have passed codon 100 at the commence- ment of the pulse will not incorporate [35S]methionine into the 28 000-44, polypeptide. Together, these observations pre- dict that during a short pulse significantly less radioactive label will be incorporated into the 28000-M, polypeptide than into the 46000-M, polypeptide resulting in difficulty in detecting the smaller polypeptides. The fortunate circum- stance of four methionine codons preceding the AGA codon at position 402 may account for the relative ease of detecting it. It is also possible that other small polypeptides do not fold in a way which can be recognised by TolC antiserum raised against the mature TolC protein. We did observe several small polypeptides besides the 46000-M, polypeptide but only when immunoprecipitates were not washed thoroughly with buffer before loading on the gel (data not shown). This may reflect the weak antigenic reactivity of these additional small polypeptides with our TolC antiserum; we have not further investigated this possibility.

We also observed a second effect of the cloned tRNAArg in that a portion of the 46000-M, polypeptide persisted much longer than in the absence of the cloned tRNAArg gene. It appears that the additional tRNA may have one of two effects when a ribosome reaches codon 402: it may either enable rapid addition of arginine and avoidance of stalling at this site, or it may in some way terminate chain elongation. We have not yet investigated this latter aspect and assume it to be due either to the T4 tRNA being not fully functional or not fully charged. The cloned tRNA may for example not be properly processed to the mature form or may be at a high concentration and which saturates the charging process.

We thank Carry Pcnny for excellent technical assistance and Dr W. H. McClain for plasmid GM4113. This work was supported by a

grant from the Australian Rescarch Grants Committee to P. Recves and by an Adclaide University Scholarship to R. Misra.

REFERENCES 1. Morona, R., Manning, P. A. & Rceves, P. (1983) J . Bacieriol.

2. Inouye. M. & Halegona, S. (1980) CRC Crit. Rev. Biochem. 7,

3. Hackett, J., Misra, R. &Reeves, P. (1983) FEBS Lett. 156, 307-

4. Millcr, H . J. (1972) Experiments in mokeculnr generics, Cold Spring

5. Morona, R. & Reeves, P. (1982) J . Bactrriol. 150, 1016-1023. 6. Kahn, M., Kolter, R., Thomas, C., Figurski, D., Meyer, R.,

Renault. E. & Helinski, D. R. (1979) Methods Enzymol. 68, 268.

7. Davis, R. W., Botstein, D. & Roth, J . R. (1980) Advancedbacterial genetics. A mairualfirr genetic engineering, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

153,693-699.

339- 371.

310.

I larbor Laboratory, Cold Spring Harbor, New York.

8. Ito, K.. Sato, T. & Yura, T. (1977) Cell 11, 1016-1023. 9. Lugtenberg, B., Mcijers, J., Peters, R., Van der Hoek, P. & Van

10. Movva, Rao, Nakamura, K. & Inouye, M. (1980) J . Mol. Biol.

11. Crowlesmith, I. & Gamon, K. (1982) Eur. .I. Biochim. 124, 577-

12. Randall, L. L., Josefsson, L. G . & Hardy, S. J. S. (1980) Eur. J .

13. Hackelt, 1. & Rceves, P. (1983) Nucleic Acids Res. 11, 6487-

14. Maz7ara, G. P., Plunkctt, G . & McClain, W. H. (1981) Proc. Natl

15. Farabangli, P. J. (1978) Nature (Lond.) 274, 765-769. 16. Singleton, C. J., Roedcr, W. D., Bogosian, G.. Somervillc, R.

17. Grantham, R., Gautier, C., Gouy, M., Jacobzone, M. & Mercier,

18. Grosjean, H. & Fiers, W. (1982) Gene 18, 199-209. 19. Stoner, C. M. & Schleif, R. F. (1982) J . Ma/. Bid . 154, 649-652. 20. Konigsberg, W. & Godson, G. N. (1983) Proc. Nut1 Acad. Sci.

21. Robinson, M., Lilley, R., Little, S., Emtage, J . S., Yarranton, G., Stephens, P., Millican, A,, Eaton, M. & Humphrcys, G. (1984) Nucleic Acids Res. 12, 6663 - 6670.

22. Ikemura, T. (1981) J . Mol. B id . 146, 1-21. 23. Ikemura. T. (1981) J . Ma/. Biol. 151, 389-409. 24. Nakamura, K. , Pirtle, R. M., Pirtle, I . L., Takeishi, I. & Inoyc,

M. (1980) J . B id . Chem. 255, 210-226. 25. Bcck, E. & Bremer, E. (1980) Nucleic Acid.7 Res. 8, 301 1-3027. 26. Inokuchi, K., Mutoh, N., Matsuyama, S. & Mizushima, S. (1982)

27. Takeshi, M., Chou, M. & Inouye, M. (1983) J . Biol. Chem. 258,

28. Clement, J. M. & Hofnung, M. C. (1981) C d 2 7 , 507-514. 29. Cclis, T. F. R. & Maas, W. K. (1971) J . Mol. Biol. 62, 179-188.

Alphen, L. (1975) FEBS Lett. 58, 254-258.

143. 317-328.

583.

Biochem. 107, 375 - 379.

6495.

Acad. Sci. U S A 78, 889 - 892.

L. & Weith, H. L. (1980) Nucleic Acids Rrs. 8, 1551 - 1560.

R. (1983) Nucleic Acids Res. 9, 243-268.

USA 80,687-691.

Nucleic Acids Rrs. 10, 6957-6958.

6932 - 6940.

Documents

Intermediates in the synthesis of TolC protein include an incomplete peptide stalled at a rare Arg codon