Guillaumestahl.free.fr

Nucleic Acids Research, 1995, Vol. 23, No. 9 1557–1560
Versatile vectors to study recoding: conservation of
rules between yeast and mammalian cells
Guillaume Stahl*, Laure Bidou1, Jean-Pierre Rousset1 and Michel Cassan
Institut de Génétique et Microbiologie, URA CNRS 1354, Bâtiment 400, Université Paris-Sud, 91405 OrsayCedex, France and 1UFR 927, Université Pierre et Marie Curie, 4 place Jussieu, 75252 Paris Cedex 05, France Received December 16, 1994; Revised and Accepted March 16, 1995 ABSTRACT
Accurate quantification of recoding efficiencies is fundamental for dissecting cis elements involved in these mechanisms. We have In many viruses and transposons, expression of some
previously constructed expression vectors to quantify recoding genes requires alternative reading of the genetic code,
efficiency in mammalian cells in culture (pRSVL series) (3). Using also called recoding. Such events depend on specific
these vectors, we demonstrated that a palindromic sequence mRNA sequences and can lead to read through of an
downstream of the frameshift site in HIV-1 acts as a stimulator of in-frame stop codon or to +1 or –1 frameshifting. Here,
frameshifting in different cell types (4). We were also able to show we addressed the issue of conservation of recoding
that HIV-1 infection of T lymphoid cells does not significantly rules between the yeast Saccharomyces cerevisiae
increase the frameshifting efficiency (4). However, identification and mammalian cells by establishing a versatile vector
of trans-acting factors is limited in mammalian cells because of the that can be used to study recoding in both species. We
lack of genetic approach. We thus decided to explore the yeast first assessed this vector by analysing the site of +1
Saccharomyces cerevisiae as an alternative eukaryotic model.
frameshift of the Ty1 transposon. Two sequences from
Here we present a versatile expression cloning vector based on higher organisms were then tested in both yeast and
a simple and highly sensitive assay that can be used to analyse cis- mammalian cells: the gag–pol junction of human
and trans-acting factors involved in recoding efficiency on immunodeficiency virus type 1 (HIV-1) (a site of –1
heterologous targets, either in yeast or in mammalian cells.
frameshift), and the stop codon region of the replicase
The following specific requirements are met with this system: (i) cistron from the tobacco mosaic virus (a site of UAG
expression in yeast and mammalian cells, (ii) an internal reference read through). We show that both sequences direct a
for translation initiation efficiency, and (iii) oriented cloning sites high level of recoding in yeast. Furthermore, different
to insert recoding target sequences.
mutations of the target sequences have similar effects
Using two heterologous targets: the –1 frameshift site of the on recoding in yeast and in mouse cells. Most notably,
gag–pro/pol junction of HIV-1 and the UAG read through of the a strong decrease of frameshifting was observed in the
TMV replicase cistron, we show that both are sites of highly efficient absence of the HIV-1 stem–loop stimulatory signal.
recoding in yeast. Furthermore, mutations known to decrease Taken together, these data suggest that mechanisms
recoding in the natural host cells have the same effect in yeast. These of some recoding events are conserved between lower
results establish that yeast is a suitable host to study recoding target and higher eukaryotes, thus allowing the use of
sequences from higher eukaryotes in vivo and suggest that S.cerevisiae as a model system to study recoding on
mechanisms involved in at least some –1 frameshifting and read target sequences from higher organisms.
through events are conserved between lower and higher eukaryotes.
INTRODUCTION
MATERIALS AND METHODS
The mechanism of translation is one of the cell features which is Recoding target sequences
best conserved in evolution. Triplets of nucleotides are sequentiallydecoded in amino acids which are added to the nascent peptidyl After NheI and BclI digest which destroys the lacZ ORF in pAC74 chain until a stop codon is encountered. However, there are (see Fig. 1), pairs of complementary oligonucleotides containing exceptions to this general rule; the best documented cases being the target sequence were inserted (white/blue selection). The frameshifting (i.e. slippage of the ribosome either backward or oligonucleotides are flanked by a NheI site and a BclI compatible forward), and read through (i.e. natural suppression of a stop codon cohesive end without the first T, in order to destroy the TGA stop in-frame between two coding regions). These modifications of the codon (see sequence in Fig. 1). The lacZ fragment (from the CmR classical reading of triplets depend on specific mRNA sequences pSG74 Z plasmid) is then re-inserted in the new plasmid at the and/or structures (1) and have been termed ‘recoding’ by NheI site (blue/white selection). Cm resistance and Amp sensitivity of pSG74Z allow a parental selection for the pAC derivatives: only * To whom correspondence should be addressed 468
Nucleic Acids Research, 1995, Vol. 23, No. 9 Figure 1. Description of the cloning vectors. pSG74 Z derives from pSG74, which was obtained by insertion of a PvuII fragment from pCYM1L, containing luc under
control of SV40 early promoter, in pMY237 HpaI site. The pCYM1L is a pCYM01 derivative (5), and pMY237 is a pUC derivative containing the CAT gene, 5′ (1
kb) and 3′ (150 bp) parts of ARG4 PstI–PstI fragment from S.cerevisiae (M. Cassan, unpublished results). A NheI–NheI fragment containing the E.coli lacZ ORF with
neither initiation nor stop codons was constructed from pUR278 (6) by PCR amplification using the oligonucleotides: W1 5′-GGAAACAGCTAGCACCATGAT-
TACG-3′ and C2 5′-CTAGAGTCGAGCTAGCGGGATCCCC-3′. The fragment was then inserted at the beginning of the luc coding sequence in the NheI site of
pSG74, in-frame with the luc AUG, giving rise to pSG74 Z. The largest PstI fragment of the pSG Z family constructs was introduced in the unique PstI site of pFL36,
giving rise to the pAC family. pFL36 is a shuttle vector, containing a replicative origin in E.coli and a β-lactamase gene (ampicillin resistance); it is also replicative
(ARS/CEN) and selectable (LEU2 ) in S.cerevisiae (12).
blue clones on ampicillin have the right structure. Each region [150 RESULTS AND DISCUSSION
nucleotides (nt)] surrounding the recoding site has been sequenced.
All recoding target sequences studied in this report are shown in General strategy
A hybrid lacZ–luc gene under control of the SV40 early promoter, was cloned on a shuttle expression vector (pFL36) replicative in Caesium chloride purified plasmids were used to transform E.coli and S.cerevisiae (12), giving rise to the pAC series (see Fig.
CM5α Escherichia coli strain (5,6).
1). As the SV40 promoter is functional both in yeast and higher Haploid S.cerevisiae strain Fy1679-18Bα (7) was transformed organisms (5), the pAC vectors can be used directly to transform using the lithium acetate method (8). At least three independent yeast cells and to transfect mammalian cells in culture, allowing transformants were grown in 2 ml rich media (1% yeast extract, direct and precise comparison of recoding efficiencies. Target 2% bactopeptone, 2% glucose) to early stationary phase. In these sequences were inserted at the junction of the lacZ–luc hybrid conditions, there is <4% plasmid loss per generation (data not gene in such a way that each ribosome initiating translation will shown). Culture aliquots were then diluted 1:20 in 4 ml of rich give rise to β-galactosidase activity but an active luciferase will medium, grown to reach OD600 = 1.5 and centrifuged. Crude be synthesized only if a recoding event takes place during cellular extracts were obtained by vortexing cell pellets for 30 min translation of the hybrid mRNA (see Fig. 1). The β-galactosidase at 4_C in 150 µl of luciferase assay buffer (9), including an equal is used as an internal reference to estimate the overall expression volume of glass beads. The crude extracts were cleared by level of each construct thus controlling for vector stability, centrifugation at maximum speed for 5 min at 4_C in a Eppendorf transfection efficiency, transcriptional and translational rates.
centrifuge. There were no significant clonal variation between Since introduction of an additional sequence at the amino terminal end of the luciferase protein may alter its specific The same constructs were transfected in NIH3T3 cells, using activity, we included control constructs in which no recoding is the calcium phosphate precipitation method (10). Crude extracts required to obtain a β-galactosidase–luciferase fusion protein.
were obtained as described before (3).
Each fusion protein synthesized from the control constructs willbe exactly the same as that obtained after recoding on the test Enzyme assays
constructs. To quantify recoding, we estimated the ratio betweenthe luciferase activity of the various test constructs to that of the Luciferase (9) and β-galactosidase activities (11) were measured corresponding control constructs, after standardization for ex- from the same crude extract as described elsewhere (4).
pression efficiency using β-galactosidase activity.
Nucleic Acids Research, 1995, Vol. 23, No. 9 469
Figure 2. Synopsis of recoding systems and sequence of the recoding sites. Name of the test vectors are in plain letters; those of control vectors are in italics. Sequence
of oligonucleotides are listed in references (3) and (4).
Ty1 +1 frameshifting
positioned downstream from the action site. The stimulatory signalis a stem–loop and possibly acts through slowing down the In order to assess whether the lacZ–luc hybrid gene is functional ribosome. By contrast, Wilson and co-workers (16) did not detect in yeast, we studied the slippery site of Ty1 transposon in its any effect of the palindromic sequence in a study where the HIV-1 natural host (13,14). This frameshift event is mediated by a 7 nt gag–pol frameshift has been analysed in yeast cells. This might be sequence CUU AGG C where the rare AGG arginine codon acts due to difficulties to accurately measure frameshifting in the as a stimulatory signal (14). When tRNAARG is in limiting detection system used or alternatively, to fundamental differences concentration, the tRNALEU slips one base forward, from CUU between yeast and mammalian cells regarding the mechanism of to UUA, thus achieving a +1 frameshift. In the control construct, frameshifting. We have investigated the effect of the potential a C was omitted in the first position of the slippery sequence CTT secondary structure on frameshifting efficiency, both in yeast and AGG C (14; and see Fig. 2). Figure 3a shows that, in yeast, the mammalian cells. We used pAC vectors to compare the frameshift- +1 frameshifting efficiency of the Ty1 slippery site carried by the ing efficiency obtained with the HIV-1 minimal sequence (pAC89) pACTy is 32% ± 2.5. Previous reports, using a lacZ reporter and that obtained in presence of the natural HIV-1 downstream system, have shown that the Ty1 +1 frameshifting efficiency stimulatory signal (pAC1789). We also included a mutated varies between 25 and 40% in yeast (13,14). Our results are sequence (pAC1815), where most of the base pairing of the stem therefore in good agreement with those already described, which has been destroyed and which has been shown to result in a 2-fold indicated that the lacZ–luc reporter hybrid gene is suitable for decrease in frameshifting efficiency in mammalian cells (4). Each quantification of recoding events in yeast.
control construct was designed by addition of an A after the TTA By contrast, only a very low recoding efficiency (7 × 10–4) was codon from the slippery sequence (17; and see Fig. 2).
observed with this construct in mammalian cells (Fig. 3a). This The results obtained in NIH3T3 mouse fibroblasts are shown in very low level may be due to minor but specific differences Figure 3b. The frameshifting efficiency is 0.6% ± 0.4 with the between the translational machineries of yeast and mammals. For minimal sequence (pAC89) and is increased five times (3% ± 1.4) example, the involved tRNAARG may be more abundant in when the stimulatory signal is present (pAC1789). The pAC1815 mammalian cells, thus inhibiting frameshift initiation. This construct with the disrupted palindrome gave an intermediate interpretation is consistent with the observation that the corre- frameshifting efficiency (1.3% ± 0.1). These recoding efficiencies sponding AGG codon is much more frequent in mouse than in are very similar to (i) those previously obtained using the pRSVL series vectors, with no lacZ internal control (4), (ii) other constructsin various mammalian cells (3,4,18), and (iii) another reporter HIV-1 frameshifting
system in avian or simian cells (19).
In heterologous yeast cells, a 3.3-fold increase of frameshifting This frameshift takes place at a 7 nt slippery site (action site): T efficiency was obtained in presence of the palindromic sequence: TTT TTA. In mammalian cells, frameshifting efficiency is 2% ± 0.2 for the pAC89 minimal construct, 6.6% ± 1.5 for the increased by a secondary structure (stimulatory signal) precisely pAC1789 construct containing the wild-type palindrome. The 470
Nucleic Acids Research, 1995, Vol. 23, No. 9 cells, ∼2-fold for the HIV-1 constructs and 10-fold for the TMVconstructs. Although the reason for such quantitative variationsremains unclear, this high level would be useful for isolatingdown mutants affected in the recoding process (specific ‘antisup-pressor’), which might lead to the characterisation of newtranslational factors.
Altogether, the results reported here show that some recoding rules are conserved between yeast and mammalian cells; theystrongly suggest conservation of the mechanisms involved. Theyalso clearly indicate that S.cerevisiae is a convenient and reliablehost to analyse this type of recoding events taking place on targetsequences active in higher eukaryotes.
Figure 3. Recoding efficiency in yeast and mouse cells. (a) Ty1 +1 frameshifting
(7 nt minimal sequence). (b) gag–pol –1 frameshifting: HIV-1 slippery sequence
ACKNOWLEDGEMENTS
with a wild-type (1789), mutated (1815) or without (89) palindromic downstream
sequence. (c) TMV replicase read through: wild-type ((TMV) or mutant ((TMG)
We are grateful to Marguerite Picard for constant support sequence. Results are expressed as the ratio of luciferase activity obtained withthe test vectors compared to their respective control, after β-galactosidase throughout the course of this work, and to Mounira Amor who normalisation. Each value represents the mean of at least five experiments in welcomed us in her laboratory for transfection experiments.
yeast and three in NIH3T3. Since identical results were obtained with pRSVL and During the course of this work, LB was supported by the pAC derivative vectors for the HIV-1 constructs, only pRSVL derivative were ‘Fondation Singer-Polignac’. Part of this work was supported by used for TMV and TMG measurements in NIH3T3. In NIH3T3, the constructwith the smallest efficiency (pACTy) gave a luciferase activity between 3 × 102 the ‘Association pour la Recherche sur le Cancer’ (ARC no.
and 6 × 103 relative light units (RLUs) (mock transfected cells gave a luciferase 3065) and by the ‘Agence Nationale de Recherches sur le Sida’ activity of 102 RLUs). For the control constructs, the luciferase activity was between 5 × 103 and 4 × 105 RLUs. In yeast, values obtained with pAC89 werebetween 8 × 102 and 1.2 × 103 RLUs while those of the control constructs werebetween 2.5 × 103 RLUs and 2 × 105 RLUs.
REFERENCES
1 Atkins, J.F., Weiss, R.B. and Gesteland, R.F. (1990) Cell, 62, 413–423.
frameshift level is intermediate for the pAC1815 construct 2 Gesteland, R.F., Weiss, R.B. and Atkins, J.F. (1992) Science, 257,
containing the disrupted palindrome (2.8% ± 0.5) (Fig. 3b).
Therefore, the stem–loop acts also on –1 frameshifting in yeast.
3 Cassan, M., Berteaux, V., Angrand, P.O. and Rousset, J.P. (1990) Res. These results suggest that the translational machineries of yeast Virol., 141, 597–610.
4 Cassan, M., Delaunay, N., Vaquero, C. and Rousset, J.P. (1994) J. Virol., and mammalian cells use similar mechanisms in recognizing the 68, 1501–1508.
frameshift site and the stimulatory signal.
5 Camonis, J.H., Cassan, M. and Rousset, J.P. (1990) Gene, 86, 263–268.
6 Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989) Molecular cloning: A
TMV read through
7 Dujon, B. (1994) Nature, 369, 371–377.
Finally, we examined whether yeast cells could achieve pro- 8 Gietz, D., St. Jean, A., Woods, R.A. and Schiestl, R.H. (1992) Nucleic Acids Res., 20, 1425.
grammed translational read through, using the region containing 9 Nguyen, V.T., Morange, M. and Bensaude, O. (1988) Ann. Biochem., 171,
the UAG leaky stop codon from TMV. We used a construct (pACTMV) containing the wild-type minimal region UAG CAA 10 Graham, F.L. and Van der Eb, A.J. (1973) Virology, 52, 456–467.
UUA (20) and a modified construct where the two CAA codons 11 Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring surrounding the UAG stop codon are replaced by two CAG 12 Bonneaud, N., Ozier-Kalogeropoulos, O., Li, G., Labouesse, M., codons (pACTMG; see Fig. 2). These positions are important for Minvielle-Sebastia, L. and Lacroute, F. (1991) Yeast, 7, 609–615.
a high level read through in plant cells (20) as well as in mouse 13 Clare, J.J., Belcourt, M. and Farabaugh, P.J. (1988) Proc. Natl. Acad. Sci. cells (21; Rousset and Cassan, unpublished). For the control USA, 85, 6816–6820.
vector, the UAG codon was replaced by a sense CAG codon (22).
14 Belcourt, M.F. and Farabaugh, P.J. (1990) Cell, 62, 339–352.
15 Wada, K., Wada, Y., Ishibashi, F., Gojobori, T. and Ikemura, T. (1992)
As shown in Figure 3c, the efficiency of read through at the TMV Nucleic Acids Res., 20, 2111–2118.
site is very high in yeast cells (21% ± 4.8) and is reduced 7-fold 16 Wilson, W., Braddock, M., Adams, S.E., Rathjen, P.D., Kingsman, S.M.
by mutations at the two critical sites (pACTMG: 3.1% ± 0.3). In and Kingsman, A.J. (1988) Cell, 1159–1169.
mouse NIH3T3 cells, the read through efficiency was 2% ± 0.7 17 Jacks, T., Power, M.D., Masiarz, F.R., Luciw, P.A., Barr, P.J. and Varmus, for pACTMV and 0.2% ± 0.08 for pACTMG. Therefore, the H.E. (1988) Nature, 331, 280–283.
18 Reil, H., Kollmus, H., Weidle, U.H. and Hauser, H. (1993) J. Virol, 67,
mutation also influences the recoding rate to the same extent in yeast and mouse cells. These results are also in agreement with 19 Parkin, N.T., Chamorro, M. and Varmus, H.E. (1992) J. Virol., 66,
those obtained in plant protoplasts (20).
20 Skuzeski, J.M., Nichols, L.M., Gesteland, R.F. and Atkins, J.F. (1991) J. Mol. Biol., 218, 365–373.
Conclusion
21 Berteaux, V., Rousset, J.P. and Cassan, M. (1991) Biochimie, 73,
The results we obtained in yeast compared to higher eukaryotic 22 Kuchino, Y., Beier, H., Akita, N. and Nishimura, S. (1987) Proc. Natl. cells indicate an overall increased recoding efficiency in yeast Acad. Sci. USA, 84, 2668–2672.

Source: http://guillaumestahl.free.fr/papers/StahlCassan95.pdf