SEARCH

SEARCH BY CITATION

Keywords:

  • Leaderless mRNA;
  • Ribosomal protein S1;
  • Translation

Abstract

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References

Leaderless mRNAs beginning with a 5′-terminal start codon occur in all biological systems. In this work, we have studied the comparative translational efficiency of leaderless and leadered mRNAs as a function of temperature by in vitro translation competition assays with Escherichia coli extracts. At low temperature (25°C) leaderless mRNAs were found to be translated comparatively better than mRNAs containing an internal canonical ribosome binding site, whereas at high temperature (42°C) the translational efficiency of canonical mRNAs is by far superior to that of leaderless mRNA. The inverse correlation between temperature and translational efficiency characteristic for the two mRNA classes was attributed to structural features of the mRNA(s) and to the reduced stability of the translation initiation complex formed at a 5′-terminal start codon at elevated temperature.


1Introduction

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References

Leaderless mRNAs are present in all domains of life[1], and several studies demonstrated that leaderless mRNAs of different origin are translated in heterologous systems [2–6]. The observation that leaderless λc I mRNA did not form a binary complex with ribosomes and that its translation required the presence of translation initiation factor IF2 led to the proposal that a ribosome–initiator tRNA complex, analogous to that formed during translation initiation in higher eukaryotes, is required for translation initiation complex formation at the 5′-terminal start codon of leaderless mRNAs[5]. The absence of downstream cis elements that contribute to ribosomal recognition of leaderless mRNAs [5–9] lends further support to this hypothesis.

Since leaderless and leadered mRNAs co-exist in the same cells, a relevant issue concerns the possibility that factors and/or environmental conditions might differentially modulate the translational efficiency of the two classes of mRNAs. For instance, an increase in the IF2 concentration – a condition occurring in Escherichia coli during cold-shock[10]– was found to enhance leaderless λc I mRNA translation both in vivo and in vitro, and enabled this mRNA to compete with a canonical mRNA for ribosomes [5,11]. In contrast, translation initiation factor IF3 was shown to act as a down-regulator of leaderless mRNA translation, since an increase of the intracellular concentration of this factor was shown to be inversely correlated with the translational efficiency of a reporter gene translationally fused to a leaderless mRNA[12]. Furthermore, discrimination of 5′-terminal start codons by IF3 was shown to require the presence of ribosomal protein S1[13]. Moreover, protein S1 was shown to be dispensable for translation of leaderless mRNA, while it is essential for in vitro 30S-initiation complex formation on mRNAs comprising a structured 5′-untranslated region (UTR) [14,15]. In this study, using both in vitro translation competition assays and toeprinting, we have analyzed the effect of temperature changes on translation of leaderless and canonical mRNAs. Our results show that temperature variation has differential effects on the translation of the two classes of mRNA.

2Materials and methods

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References

2.1Plasmids and PCR templates used in this study

Plasmid pAXL7 harbors the full-length c I gene[5]. Plasmid pKS0235[16] served as a source for wild-type ompA mRNA. Plasmid pcI-1[5] served as a source for full-length c I mRNA with a 5′UTR containing the Shine–Dalgarno (SD) sequence of phage T7 gene 10. Plasmid pIR12[17] harbors the entire leaderless rro gene of Lactococcus lactis phage r1t. Each of the encoded mRNAs was under transcriptional control of a T7 promoter and was transcribed in vitro using T7 RNA polymerase.

2.2In vitro translation assays

The in vitro translation of c I and ompA mRNAs with E. coli S100 extracts was essentially performed as described[5]. Mix I contained 16.6 mM MgOAc, 80 mM NH4Cl, 30 mM Tris/HCl pH 7.7, 3.3 mM DTT, 1.6 mg ml−1E. coli tRNA, 0.2 mM citrovorum, 16.6 mM KCl, 0.33 mM amino acids (−lysine), 66.6 μM [14C]lysine, 3.3 mM ATP, 0.66 mM GTP, 16.6 mM phosphoenolpyruvate and 0.04 mg ml−1 pyruvate kinase. Mix II contained 2 mM Tris/HCl pH 7.7, 60 mM NH4Cl, 10 mM MgOAc, 0.33 μl S100 extract, 3.3 pmol μl−1 IF free 70S ribosomes and 0.66 pmol μl−1 IFs. The translation mixture contained in a final volume of 130 μl, 78 μl of Mix I, 26 μl of Mix II and 45 pmol of each mRNA. The translation reactions were started by the addition of mRNA and incubated at 25°C, 30°C, 37°C or 42°C, respectively. 12 μl samples were withdrawn after different incubation times. The reactions were stopped by addition of 2 volumes of protein sample buffer. The labelled proteins were separated on 15% SDS–polyacrylamide gels. The gels were dried under vacuum and exposed to a Molecular Dynamics PhosphoImager screen for quantitation. The relative signal intensities were corrected for the number of lysine residues in the different proteins.

2.3Preparation of ribosomes and primer extension inhibition analysis (toeprinting)

Initiation factor free 30S subunits were prepared as described[18]. They were judged as pure when 23S rRNA was absent from the preparations. The preparation of S1-depleted 30S subunits was performed by affinity chromatography using poly(U)-Sepharose 4B (Pharmacia) as described previously[19]. The absence of S1 on 30S(-S1) ribosomes was verified by Western-blot analysis with anti-S1 antibodies.

The ompAΔ117 mRNA was prepared from a PCR template[14], and the Ava II primer[13] was used for toeprinting. The toeprinting assays were carried out with 30S ribosomal subunits and initiator-tRNA, tRNAfMet, essentially as described by Hartz et al.[20]. The mRNA (0.04 pmol) was pre-incubated at either 25°C, 30°C, 37°C or 42°C for 5 min together with 4 pmol 30S subunits and 20 pmol tRNAfMet. Then, the MMLV reverse transcriptase reactions were performed at the corresponding temperatures. The MMLV reactions were terminated after 10 min. The relative toeprints (%) were calculated as described[21] after quantitation of the toeprint and extension signals with a PhosphoImager using the equation: toeprint signal/(toeprint signal+extension signal)×100. Since the extension signal serves as an internal standard, the equation provides an intrinsic correction for the temperature-dependent differential activity of the MMLV.

The toeprinting assays on λc I(SD) mRNA were performed with 30S and 30S(-S1) ribosomes and primer O8[13] as described above at 37°C.

3Results and discussion

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References

3.1At low temperatures leaderless mRNA can compete for ribosomes with leadered mRNA

Leaderless mRNAs are translated at rather low levels in E. coli[12,22]. At the optimal growth temperature for E. coli (37°C), and at physiological stoichiometric ratios between IF1, IF2, IF3 and ribosomes (≅ 0.15:0.15:0.15:1), translation of a canonical mRNA by far outcompetes that of a leaderless mRNA [5,12]. In these experiments we asked whether the temperature can affect the competition for ribosomes between mRNAs containing a 5′-terminal start codon and an mRNA containing a canonical ribosome binding site (rbs). The studies were performed in vitro with the rationale to avoid interfering effects resulting from variables difficult to control in vivo such as varying mRNA concentrations or different mRNA and/or product stability at different temperatures.

We prepared an E. coli in vitro system under conditions in which ribosomes are limiting (∼1:1 ratio of ribosomes and mRNA) for translation and programmed it with both leaderless λc I mRNA (Fig. 1A) and E. coli ompA mRNA (Fig. 1C) encoding the λ repressor and the outer membrane protein A, respectively. The latter mRNA contains a highly structured, 135-nucleotide-long 5′UTR comprising a canonical SD sequence, and its translation requires the presence of ribosomal protein S1 [14,15]. When equimolar amounts of the above mRNAs were translated with the system, the relative rates of their translation varied as a function of the incubation temperature: translation of c I mRNA was found to be 65% and 45% of that of ompA mRNA at 25°C (Fig. 2A) and at 30°C (Fig. 2B), respectively. In general these experiments indicated that the lower the incubation temperature, the lower the translational efficiency of ompA mRNA and correspondingly the higher the relative rate of leaderless mRNA translation. At 37°C, corresponding to the optimal growth temperature of E. coli, ompA mRNA was well expressed, whereas the leaderless c I mRNA was translated at a rate equal to approximately one fifth of that of ompA mRNA (Fig. 2C); when the temperature was raised to 42°C, CI repressor synthesis was further decreased relative to OmpA synthesis (Fig. 2D). These results reflect primarily the different capacity of the two mRNAs to compete with each other for the limiting amount of available ribosomes. In fact, as shown in Fig. 2E, in absolute terms translation of both mRNAs is increased with increasing temperature. However, the temperature-dependent increase in translation rate is different for the two mRNAs: for a 30 min translation time, the absolute translation rate of ompA mRNA increases by approximately 33-fold going from 25°C to 37°C while the corresponding increase for c I mRNA was only approximately 8.5-fold (Fig. 2E). Furthermore, while translation of ompA mRNA can further improve at higher temperature, the increase from 25°C to 42°C being approximately 60-fold (Fig. 2E), the c I mRNA translation rate becomes actually lower at 42°C compared to 37°C (Fig. 2E), a finding which could simply reflect a reduced stability of the 30S initiation complex formed at the 5′-terminal start codon at the elevated temperature.

image

Figure 1. Structures of the 5′UTR and the initial coding region of the mRNAs used in this study. A: λc I mRNA; B: c I(SD) mRNA; C: E. coli ompA mRNA; D: ompAΔ117 mRNA; E: phage r1t gene rro mRNA. The authentic start codons of the c I, c I(SD), rro, ompA mRNAs as well as the internal start codon AUG68 of c I mRNA[12] are underlined. On ompAΔ117, the underlined 5′-terminal AUG has been artificially created[14]. Authentic and putative SD sequences are indicated by bars. The secondary structures shown in ompA mRNA[27] and c I mRNA[12] have been experimentally verified. According to the computer predictions (RNA-FOLD), the artificial 5′UTR of c I(SD) is unstructured.

Download figure to PowerPoint

image

Figure 2. Competition of the leaderless λc I mRNA with the canonical ompA mRNA for ribosomes at different temperatures. A–D: Quantitative analysis of the in vitro translation rates with equimolar concentrations of leaderless λc I mRNA and ompA mRNA. The relative synthesis of the CI (filled circles) and OmpA (open circles) proteins is shown for different incubation times. CI synthesis is normalized to that of OmpA in each panel. The relative signal intensities have been corrected for the number of lysine residues in each protein. E: Translation rates of both c I and ompA mRNA at different temperatures. The in vitro translation reactions with both c I (black bars) and ompA mRNA (white bars) were carried out for 30 min at the different temperatures shown. The translation rates of both mRNAs were normalized to the respective signal intensity obtained at 25°C.

Download figure to PowerPoint

3.2Temperature-dependent differential stability of ternary complexes at a 5′-terminal start codon and at a canonical rbs

To examine the effect of temperature on the stability of 30S initiation complexes formed at a 5′-terminal AUG and an internal canonical rbs, toeprinting assays were performed between 25°C and 42°C with 30S ribosomes on ompAΔ117 mRNA containing a 5′-terminal AUG (AUG1) and a downstream AUG (AUGi) preceded by the canonical SD sequence of ompA mRNA (Fig. 1D). Since 5′-terminal or internal 30S initiation complex formation is exclusive on ompAΔ117 mRNA[13], and other factors can be disregarded, the minimal experimental system is suited to study the effect of different temperatures on 30S-ternary complex formation at either start codon. As shown in Fig. 3A, lane 2, and Fig. 3B, at 25°C, the 5′-terminal AUG was able to compete with AUGi for ribosomes. The toeprint signals obtained at higher temperatures mirrored the in vitro translation assays shown in Fig. 2 in that at higher temperatures the 30S subunits were shifted to AUGi at the expense of AUG1 (Fig. 3A, lanes 3–5, and Fig. 3B). Thus, the simplest explanation for the inverse translational efficiency of leaderless and leadered mRNAs at elevated temperature is the reduced stability of the translation initiation complex at a 5′-terminal AUG when compared to that at a canonical rbs which can be readily explained by the additional SD-aSD contacts between a canonical mRNA and 16S rRNA.

image

Figure 3. Temperature-dependent competition for 30S subunits of the 5′-terminal AUG with the canonical rbs on ompAΔ117 mRNA. A: Toeprint analysis. Lane 1, primer extension in the absence of 30S subunits and initiator-tRNA. Lanes 2–5, toeprint analysis with 30S subunits at different temperatures. The reaction mixtures were incubated at 25°C, 30°C, 37°C, or 42°C as indicated on top. The stoichiometries of ribosomes, tRNAfMet, and mRNA are given in Section 2. The toeprint signals for AUG1 and AUGi are indicated by a filled and an open circle, respectively. B: Relative toeprints obtained on ompAΔ117 mRNA for AUG1 and AUGi at the different incubation temperatures.

Download figure to PowerPoint

It is interesting to note that all leaderless mRNAs identified in E. coli stem from accessory genetic elements, and two out of the three[1] specify repressor proteins which are not required in high quantities. Nevertheless, both the TetR and CI repressor must be continuously synthesized over the entire temperature range supporting cell division to maintain a latent state of the Tn1721 transposon and of phage λ, respectively. Since the rate of translation is temperature dependent[23], at elevated temperatures the increased rate of protein synthesis may compensate for the reduced stability of translation initiation complexes formed on leaderless mRNAs (see Fig. 3) as well as for the increased competition by leadered mRNAs for ribosomes.

3.3A canonical mRNA containing an unstructured 5′UTR competes successfully with leaderless mRNA over a temperature range from 25°C to 42°C

Upon lowering the temperature below a threshold value (approximately 20°C) most of the translational activity of E. coli stops while a small set of (cold-shock) proteins is synthesized. It is generally believed that the translational block is caused by a defect in translation initiation, possibly in association with the acquisition by the mRNAs of unfavorable secondary structures [24,25] which are considered to be a key factor in determining the efficiency and the regulation of translation initiation in prokaryotes. Several lines of evidence suggest that translation of highly structured mRNAs requires ribosomal protein S1 most likely because of its RNA unfolding capacity[26]. The 5′UTR of ompA mRNA (Fig. 1C) contains large stem-loop structures[27], and S1 has been shown to be essential for ternary complex formation on ompA mRNA[14]. The results shown in Fig. 2 could indicate that at low temperatures the RNA unfolding capacity attributed to S1 is not sufficient to effectively resolve the secondary structures around the rbs of ompA mRNA. This could explain why at lower temperatures the leaderless λc I, which is devoid of 5′-secondary structures, is able to compete with leadered ompA mRNA for translation by a limiting amount of ribosomes (Fig. 2).

If this conjecture is correct, then translation of a canonical mRNA with an unstructured translation initiation region should outcompete leaderless mRNA translation with equal efficiency at both low and high temperatures. To test this hypothesis, the leaderless λc I mRNA was modified to contain a short 5′UTR comprising the strong translational initiation signals of phage T7 gene 10[28] (Fig. 1B). According to computer predictions (RNA-FOLD), the artificial 5′UTR of c I(SD) (Fig. 1B) is unstructured. Our previous studies have indicated that S1 is dispensable for translation initiation complex formation on mRNAs containing a 5′-proximal rbs when it is devoid of secondary structure(s)[14]. Therefore, we tested first whether or not in vitro formation of the 30S-initiation complex on this mRNA is affected by protein S1. The toeprint assay (Fig. 4A) revealed that both 30S(-S1) and 30S ribosomes formed ternary complexes with the same efficiency on c I(SD) mRNA indicating that structure(s) does not affect ribosomal recognition of this mRNA.

image

Figure 4. Competition of the leaderless rro mRNA with c I(SD) mRNA for ribosomes at different temperatures. A: Toeprinting analysis on c I(SD) mRNA (Fig. 1B) with 30S and 30S(-S1) subunits at 37°C. The concentrations of ribosomes, tRNAfMet and mRNA were as described in Section 2. Lane 1, primer extension in the absence of ribosomes and tRNAfMet. Lanes 2 and 3, reactions with equimolar concentrations of 30S ribosomes in the presence and absence of S1, respectively. The toeprint signals are marked by a filled circle. B–E: Quantitative analysis of the in vitro translation rates with equimolar concentrations of rro mRNA and c I(SD) mRNA. The relative synthesis of the Rro (filled circles) and CI (open circles) proteins is shown for different incubation times at different temperatures. Rro synthesis is normalized to that of CI in each panel. The relative signal intensities have been corrected for the number of lysine residues in each protein.

Download figure to PowerPoint

Next, we used c I(SD) mRNA in the in vitro translation assay as a competitor with leaderless rro mRNA (Fig. 1E) which is derived from the lactococcal phage r1t. The latter mRNA was chosen because this template encodes a product with different electrophoretic mobility from the product of c I(SD) mRNA thereby facilitating the task of quantifying the product of each mRNA. As shown in Fig. 4B–E, when equimolar amounts of rro mRNA and c I(SD) mRNA were co-translated, CI repressor synthesis by far exceeded Rro synthesis at all temperatures tested. This result lends further support to the premise that it is the dependency on mRNA structure(s) and/or ribosomal protein S1 for 30S initiation complex formation that determines the temperature-dependent differential translation of leaderless and leadered mRNAs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References

We thank Drs. R. Koebnik and A. Nauta for the gift of plasmids. A part of this work was performed by S.G. in the laboratory of C.O.G. who is supported by MURST-CNR 95/95, CNR (Progetto Strategico ST74). The work in U.B.'s laboratory was supported by Grant P-12065 MOB from the Austrian Science Foundation.

References

  1. Top of page
  2. Abstract
  3. 1Introduction
  4. 2Materials and methods
  5. 3Results and discussion
  6. Acknowledgements
  7. References
  • [1]
    Janssen, G.R. (1993) Eubacterial, archaebacterial, and eukaryotic genes that encode leaderless mRNA. In: Industrial Microorganisms Basic and Applied Molecular Genetics (Baltz, R.H., Hegeman, G.D. and Skatrud, P.L., Eds.), pp. 59–67. American Society for Microbiology, Washington, DC.
  • [2]
    Lopez, P., Martinez, S., Diaz, A., Epinosa, M. (1987) Streptococcus pneumoniae polA gene is expressed in Escherichia coli and can functionally substitute for the E. coli polA gene. J. Bacteriol. 169, 48694871.
  • [3]
    Wu, C.J., Janssen, G.R. (1996) Translation of vph mRNA in Streptomyces lividans and Escherichia coli after removal of the 5′-untranslated leader. Mol. Microbiol. 22, 339355.
  • [4]
    Wu, C.J., Janssen, G.R. (1997) Expression of a Streptomycete leaderless mRNA encoding chloramphenicol acetyltransferase in Escherichia coli. J. Bacteriol. 179, 68246830.
  • [5]
    Grill, S., Gualerzi, C.O., Londei, P., Bläsi, U. (2000) Selective stimulation of translation of leaderless mRNA by initiation factor 2: evolutionary implications for translation. EMBO J. 19, 41014110.
  • [6]
    Moll, I., Huber, M., Grill, S., Sairafi, P., Londei, P., Müller, F., Brimacombe, R., Bläsi, U. (2001) Evidence against an interaction between the mRNA downstream box and 16S rRNA in translation initiation. J. Bacteriol. 183, 34993505.
  • [7]
    Resch, A., Tedin, K., Gründling, A., Mündlein, A., Bläsi, U. (1996) Downstream/anti-downstream-box interactions are dispensable for translation initiation of leaderless mRNAs. EMBO J. 15, 47404748.
  • [8]
    O'Connor, M., Asai, T., Squires, C.L., Dahlberg, A.E. (1999) Enhancement of translation by the downstream box does not involve base pairing of mRNA with the penultimate stem sequence of 16S rRNA. Proc. Natl. Acad. Sci. USA 96, 89738978.
  • [9]
    La Teana, A., Brandi, A., O'Connor, M., Freddi, S., Pon, C.L. (2000) Translation during cold adaptation does not involve mRNA-rRNA base pairing through the downstream box. RNA 6, 13931402.
  • [10]
    Goldenberg, D., Azar, I., Oppenheim, A.B., Brandi, A., Pon, C.L., Gualerzi, C.O. (1997) Role of Escherichia coli cspA promoter sequences and adaptation of translational apparatus in the cold shock response. Mol. Gen. Genet. 256, 282290.
  • [11]
    Grill, S., Moll, I., Hasenöhrl, D., Gualerzi, C.O., Bläsi, U. (2001) Modulation of ribosomal recruitment to 5′-terminal start codons by translation initiation factors IF2 and IF3. FEBS Lett. 495, 167171.
  • [12]
    Tedin, K., Moll, I., Grill, S., Resch, A., Graschopf, A., Gualerzi, C.O., Bläsi, U. (1999) Translation initiation factor 3 antagonizes authentic start codon selection on leaderless mRNAs. Mol. Microbiol. 31, 6777.
  • [13]
    Moll, I., Resch, A., Bläsi, U. (1998) Discrimination of 5′-terminal start codons by translation initiation factor 3 is mediated by ribosomal protein S1. FEBS Lett. 436, 213217.
  • [14]
    Tedin, K., Resch, A., Bläsi, U. (1997) Requirements for ribosomal protein S1 for translation initiation of mRNAs with and without a 5′-leader sequence. Mol. Microbiol. 25, 189199.
  • [15]
    Moll, I., Grill, S., Gründling, A. and Bläsi, U. (2002) Effects of ribosomal proteins S1, S2 and the DeaD/CsdA DEAD-box helicase on translation of leaderless and canonical mRNAs in Escherichia coli. Mol. Microbiol., in press.
  • [16]
    Ried, G., Koebnik, R., Hindenach, I., Mutschler, B., Henning, U. (1994) Membrane topology and assembly of the outer membrane protein OmpA of Escherichia coli K12. Mol. Gen. Genet. 243, 127135.
  • [17]
    Nauta, A., van Sinderen, D., Karsens, H., DeSmit, E., Venema, G., Kok, J. (1996) Inducible gene expression mediated by a repressor-operator system isolated from Lactococcus lactis bacteriophage r1t. Mol. Microbiol. 19, 13311341.
  • [18]
    Spedding, G. (1990) In: Ribosomes and Protein Synthesis: A Practical Approach (Spedding, G., Ed.), pp. 1–29. Oxford University Press, New York.
  • [19]
    Suryanarayana, T., Subramanian, A.R. (1983) An essential function of ribosomal protein S1 in messenger ribonucleic acid translation. Biochemistry 22, 27152719.
  • [20]
    Hartz, D., McPheeters, D.S., Traut, R., Gold, L. (1988) Extension inhibition analysis of translation initiation complexes. Methods Enzymol. 164, 419425.
  • [21]
    Hartz, D., McPheeters, D.S., Gold, L. (1991) Influence of mRNA determinants on translation initiation in Escherichia coli. J. Mol. Biol. 218, 8397.
  • [22]
    Baumeister, R., Flache, P., Melefors, Ö., Gabain, A., Hillen, W. (1991) Lack of a 5′ non-coding region in Tn1721 encoded tetR mRNA is associated with a low efficiency of translation and a short half-life in Escherichia coli. Nucleic Acids Res. 19, 45954600.
  • [23]
    Farwell, A., Neidhardt, F. (1998) Effect of temperature on in vivo protein synthetic capacity in Escherichia coli. J. Bacteriol. 180, 47044710.
  • [24]
    Broeze, J., Solomon, C.J., Pope, D.H. (1978) Effects of low temperature on in vivo and in vitro protein synthesis in Escherichia coli and Pseudomonas fluorescence. J. Bacteriol. 134, 861874.
  • [25]
    de Smit, M.H., van Duin, J. (1990) Secondary structure of the ribosome binding site determines translational efficiency: a quantitative analysis. Proc. Natl. Acad. Sci. USA 87, 76687672.
  • [26]
    Thomas, J.O., Szer, W. (1982) RNA-helix-destabilizing proteins. Prog. Nucleic Acid Res. Mol. Biol. 27, 157187.
  • [27]
    Rosenbaum, V., Klahn, T., Lundberg, U., Holmgren, E., von Gabain, A., Riesner, D. (1993) Co-existing structures of an mRNA stability determinant. J. Mol. Biol. 229, 656670.
  • [28]
    Olins, P.O., Devine, C.S., Rangwala, S.H., Kavka, K.S. (1988) The T7 phage gene 10 leader RNA, a ribosome-binding site that dramatically enhances the expression of foreign genes in Escherichia coli. Gene 73, 227235.