Skip Navigation

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (98K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Pezo, V
Right arrow Articles by Wain-Hobson, S
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pezo, V
Right arrow Articles by Wain-Hobson, S
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© 1996 Oxford University Press 253-256

Footnote

Fate of direct and inverted repeats in the RNA hypermutagenesis reaction

Fate of direct and inverted repeats in the RNA hypermutagenesis reaction Valérie Pezo , Miguel Angel Martinez and Simon Wain-Hobson *

Unité de Rétrovirologie Moléculaire, Institut Pasteur, 28 rue du Dr Roux, 75724 Paris cedex 15, France

Received November 22, 1995 ; Accepted November 28, 1995

ABSTRACT

RNA hypermutagenesis results from cDNA synthesis in the presence of highly biased dNTP precursor concentrations and preferentially exploits human immunodeficiency virus type 1 (HIV-1) reverse transcriptase. Such reaction conditions slow down DNA synthesis, which might be conducive to strand transfer and deletion. This has been investigated. A 6 bp inverted repeat nested between 10 bp repeats was efficiently deleted at dCTP concentrations typically used. Inter- or intramolecular strand transfer between 10 bp repeated sequences separated by runs of templated G residues occurred, but at lower concentrations. If RNA hypermutagenesis of a sequence containing direct and inverted repeats is unavoidable, avian myeloblastosis virus (AMV) reverse transcriptase could be used, as strand transfer occurs with much diminished dCTP substrate dependence.

INTRODUCTION

Retroviral replication is an error prone process. Apart from the familiar point substitutions, frame-shift mutations and intramolecular deletions and insertions it is becoming increasingly apparent that complex mutations involving multiple strand transfer may also occur. The notorious examples involve proto-oncogene transduction, while others, less known but more frequent, may result in deletions with insertions and multiple insertions ( 1 - 8 ). In a single cycle reverse transcription assay deletions occurred frequently, but not exclusively, between short tracts of sequence identity ( 3 , 5 ). Inverted repeats were eliminated particularly efficiently ( 4 ).

To this catalogue of replication errors must be added G -> A hypermutation, which results from cDNA synthesis in the presence of highly biased intracellular dNTP pools ( 2 , 9 - 14 ). G -> A hypermutation may be reproduced in a simple in vitro reaction using RNA, reverse transcriptase and highly biased [dCTP]/[dTTP] ratios ( 15 , 16 ) and forms the basis of a powerful method for in vitro protein evolution ( 17 ). Among the plethora of hypermutants sequenced ( 15 , 16 ) a few deletions were noted, most of which occurred within small runs of 2-3 bp homologous sequences. In the present study the in vitro hypermutation reaction has been adapted to specifically study the influence of dNTP pool imbalances on nascent DNA strand transfer.

MATERIALS AND METHODS

The four RNA templates used are shown in Figure 1 . In all cases a 10 bp repeated sequence (DR1 and DR2) was separated by 1, 4 or 7 G residues or a GC inverted repeat. Single-stranded DNA oligonucleotides (72G1, 75G4, 78G7 and 81GC) were synthesized and rendered double-stranded by 20 rounds of PCR using the 19RT and 23R PCR primer pair and cloned into the pBluescript SK+ vector via Kpn I and Sac I restriction sites. Recombinants were verified by sequencing. RNA was made using T7 RNA polymerase and quantified as described ( 15 , 16 ). cDNA was synthesized using 0.5 pmol RNA with a 30-fold molar excess of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase (RT) (5 U/reaction, ~12-15 pmol; Boehringer), the 19RT primer and biased [dCTP]/[dTTP] ratios. Avian myeloblastosis virus (AMV) (Promega) and Moloney murine leukemia virus (MoMLV) (Gibco) RTs were also used at 5 U/reaction. After reverse transcription the cDNA was incubated with 1 [mu]g DNase-free RNase (Boehringer) and 0.5 U RNase H (Boehringer) and concentrated on a Centricon 30 (Amicon).


Figure 1 . Oligonucleotide sequences used in the in vitro hypermutagenesis assay. The sequences common to the four oligonucleotides are shown above while the differences are given underneath. Those used to produce double-stranded DNA and to amplify cDNA after reverse transcription, 19RT and 23R, are also given. 19RT was also used as the reverse transcription reaction primer. Kpn I and Sac I sites were used to clone into pBluescript SK+ while Xba I and Eco RI were used to clone the products into M13mp18. The template sequence was designed such that full-length and Xba I- and Eco RI-cleaved products from 72G1, 75G4, 78G7 and 81GC templated reactions would yield a blue phenotype on X-gal plates, while those resulting from deletion between the direct repeats would yield a white phenotype. The reading frame is indicated by dots under the first base of the codon.

In order to recover sufficient material for analysis (template ~0.5 pmol) the cDNA was PCR amplified under optimized conditions (2.5 mM Mg 2+ ) using the 19RT and 23R primer pair (Fig. 1 ) and Taq polymerase (Roche). The amplification profile was classic apart from three initial cycles with an annealing temperature of 37oC followed by 12 cycles at 55oC. This was done to facilitate annealing of the 23R primer to hypermutated cDNA. PCR products were precipitated and 5'-end-labelled with [[gamma]- 32 P]ATP and T4 polynucleotide kinase. Half of the reaction was electrophoresed through a 6% non-denaturing polyacrylamide gel, fixed in 10% acetic acid and dried. Band intensities were quantified using a Molecular Dynamics phosphorimager.

PCR products were purified from a 10% native polyacrylamide gel, digested with Eco RI and Xba I and cloned into M13mp18 RF DNA. The template sequences were designed such that cloned undeleted PCR-amplified products from all four different templated reactions would yield a blue phenotype on X-gal plates, while frame-shift mutations and other deletions, including deletion between the direct repeats, would yield a white phenotype. Sequencing reactions were carried out with the Sequenase 2.0 kit (USB).

RESULTS

The sequences were designed so as to allow reverse transcription to proceed to the end of DR1 unhindered by a dearth of dCTP. Depending upon the dCTP concentration the run of 1, 4 or 7 templated G residues would impose a pause in reverse transcription, probably allowing dissociation of the RT/template-primer complex. Fraying at the 3'-end of the primer would promote either intra- or interstrand transfer to DR2, offering the possibility of unimpeded reverse transcription up to the Eco RI site. Following PCR the proportion of deleted templates was quantitated by 5'-labelling of the extremities.

The raw data for the HIV-1 RT hypermutagenesis reactions using three RNA templates (75G4, 78G7 and 81GC) are shown in Figure 2 . In all cases the proportion of the deleted form increased with decreasing dCTP concentration. Band intensities were quantified by phosphorimager, the proportion of deleted products being given in Figure 3 A. Fifty percent of the template with the inverted repeat (81GC) was deleted at a dCTP substrate concentration of ~10 nM, with ~3 nM for 78G7 and <0.1 nM for 75G4. Reverse transcription of template 72G1 did not result in any deletion whatsoever. The nature of the deletions were verified by sequencing cloned PCR products. In all cases they occurred precisely between DR1 and DR2. That no deletions were found at high dCTP concentrations rules out any contribution from PCR (Figs 2 and 3 ).


Figure 2 . Raw phosphorimager data for kinased PCR products following gel electrophoresis. All reactions were carried out using HIV-1 RT. ( A ) Reactions using RNA template 75G4. Lane p indicates reaction products resulting from cDNA synthesis with physiological [dCTP], i.e. 10 [mu]M, while the dCTP concentration for reactions in lanes 1-5 was 10, 1 and 0.1 nM and 10 and 1 pM respectively. Lane m indicates the mock reaction products, i.e. without HIV-1 RT. After 15 cycles of PCR there was no signal, indicating that DNA contamination of the RNA template was negligible. ( B ) Reactions using RNA template 78G7. Lanes p and m are as above. The dCTP concentration for reactions in lanes 1-9 was 10, 8, 6, 5, 3, 1 and 0.1 nM and 10 and 1 pM respectively. ( C ) Reactions using RNA template 81GC. Lanes p and m are as above. The dCTP concentration for reactions in lanes 1-10 was 100, 50, 10, 8, 5, 3, 1 and 0.1 nM and 10 and 1pM respectively. [dATP] and [dGTP] were constant at physiological values (18), 40 and 20 [mu]M respectively, while [dTTP] was kept at 10* the physiological value, i.e. 440 [mu]M (18).


Figure 3 . Quantification of strand transfer between DR1 and DR2 by phosphorimager. ( A - C ) Reactions using HIV-1, AMV and MoMLV RTs respectively. Closed circles represent data for the 81GC template; closed squares, the 78G7 template; open squares the 75G4 template; open circles, the 72G1 template. The proportion of deleted product was calculated as the band intensity of the deleted product divided by the sum of the band intensities of the deleted and undeleted bands. The data for the HIV-1 reactions represent the mean of three reactions, while those for AMV and MoMLV are the means of two reactions. The error was +-10%.

Efficient deletion of the inverted repeat was particularly worrying given that the 10-30 nM dCTP range is typical of most G -> A hypermutagenesis reactions ( 15 , 16 ). As the sensitivity of different RTs to [dCTP]/[dTTP] changes is HIV-1 > AMV > MoMLV respectively ( 16 ), use of either of the latter two enzymes in the hypermutagenesis reaction might translate into less efficient strand transfer. Accordingly, reverse transcription of the same four templates was analysed under identical conditions but with the RTs from AMV (Fig. 3 B) and MoMLV (Fig. 3 C). For both RTs strand transfer and deletion between DR1 and DR2 occurred more frequently with the 81GC and 78G7 as opposed to the 75G4 and 72G1 templates. Perhaps the AMV enzyme could accommodate strand transfer a little better than the MoMLV RT, but given the experimental variation (+-10%) the differences are probably not significant. For both enzymes the proportion of deleted product was less than for the corresponding reactions with HIV-1 RT.

Approximately 8-10% of product remained undeleted following reverse transcription of the 81GC and 78G7 templates with HIV-1 RT, even at very low dCTP concentrations (Fig. 3 A). In fact, dCTP concentrations below 1 nM are only nominal, being supplemented by contaminating dCTP from the three other dNTPs, particularly dTTP at 440 [mu]M (data not shown). This explains why the proportion of deleted product never attained 100% (Fig. 3 A). In order to link the biased [dCTP]/[dTTP] ratio to both strand transfer and G -> A hypermutation some of the undeleted PCR material was cloned and sequenced. Only these full-length products encoded templated G residues and therefore could reveal the presence of G -> A transitions. Sequences derived from blue recombinants (cloned undeleted 78G7, 10 pM dCTP reaction products using HIV-1 RT; Fig. 3 A) are summarized in Figure 4 . Those in Figure 4 A represent 39 clones encoding single point mutations. The distribution over the seven sites was non-random ([chi] 2 = 13.8, 6 degrees of freedom, P < 0.05), with positions G1 and G3 particularly being substituted.


Figure 4 . Distribution of substitutions in undeleted cDNA products resulting from HIV-1 reverse transcription of the 78G7 template. In order to recover sufficient material cDNA from the 10 pM dCTP sample (Fig. 2B) was PCR amplified using 40 cycles and cloned into an M13mp18 vector. ( A ) Distribution of single base substitutions. ( B ) Distribution of multiple base substitutions. ( C ) Frequency (f) of substitution by A in the i + 1 position for G -> A transitions of the i th site. ( D ) Sequences of single -1 frame-shift mutations. The arrows indicates the sense of cDNA synthesis. Undeleted products from the 72G1 and 75G4 templated 1 pM reactions using HIV-1 RT were cloned and sequenced. Of the 72G1 derived clones seven of eight encoded G -> A substitutions, while of the seven 75G4 derived clones, six were mutated (one AGGG, three GGGA, one AGGA and one AAGA). PCR was not responsible for deletions between DR1 and DR2 (see Fig. 1 at high dCTP concentrations). Forty cycles of PCR were carried out on the mock reaction from the reverse transcription assay. The PCR products were cloned into M13mp18 and 20 recombinants sequenced. No base substitutions were found, indicating that all the substitutions noted above resulted from reverse transcription.

A further 18 recombinants derived from the same reaction encoded multiple G -> A substitutions (Fig. 4 B). Among these clones most of the G -> A transitions occurred at G1, G6 and G7, again giving rise to a non-random distribution ([chi] 2 = 41.0, 6 degrees of freedom, P < 0.001). Two clones encoded a U -> A substitution at position U8. This probably arose by a -1 dislocation of the primer strand with respect to the template between G5 and G7, misincorporation of a T opposite G5, G6 or G7, followed by relocation of the primer strand and subsequent elongation beyond the mismatched U[middot]T pair. The multiply substituted clones clearly showed signs of increasing dCTP depletion as both the frequency of G -> A substitutions, as well as those resulting from -1 nascent strand dislocation, increased towards the end of the run of seven G residues. This is best shown in Figure 4 C, which describes the probability of a substitution by A at i + 1 given a G -> A transition at position i .

A small number of white, rather than blue, plaques were found upon cloning the undeleted 78G7 DNA. Fourteen encoded a deletion of a single G within the run of seven G residues. As it is impossible to know where the deletion occurred and so a gap has been arbitrarily introduced at position G7 (Fig. 4 D). Twelve clones carried one or two G -> A substitutions in addition to deletion of a single G.

DISCUSSION

The data illustrate that both G -> A hypermutation and strand transfer may result from a common denominator, [dTTP]/[dCTP] biases. Presumably a pause in polymerization starting at the run of templated G residues resulting from the dearth of dCTP allows recycling of the RT and occasionally fraying of the template-nascent strand and either intra- or interstrand transfer between DR1 and DR2.

HIV-1 RT can apparently continue reverse transcription after strand transfer more easily than either the AMV or MoMLV RTs, while the differences between the latter two enzymes are minimal. However HIV-1 RT is the enzyme of choice for hypermutagenesis, as it demonstrates the greatest sensitivity to dNTP biases. Thus during G -> A hypermutagenesis reactions using this enzyme care must be exercised when using depleted dCTP concentrations, particularly below 10 nM. This is especially important if the target sequence harbours an inverted repeat. Above this concentration strand transfer, either intra- or intermolecular, between 10 bp direct repeats is tolerable. However, if a sequence proved particularly prone to internal deletions during HIV-1 RNA hypermutagenesis the AMV enzyme could be substituted, although the extent of G -> A hypermutation would be reduced by a factor of two to three ( 16 ). The MoMLV enzyme is to be avoided, for it shows much reduced dCTP-dependent G -> A hypermutation ( 16 ).

ACKNOWLEDGEMENTS

We would like to thank Henri Buc, Andreas Meyerhans, Jean-Pierre Vartanian and Lynn Ripley for helpful discussions. VP was supported by le Ministère de la Recherche et de l'Enseignement Supérieur and MAM by a post-doctoral fellowship from the European Union. This work was supported by grants from the Institut Pasteur and l'Agence Nationale pour la Recherche sur le SIDA (ANRS).

REFERENCES

1 Monk,R.J., Malik,F.G., Stokes,D. and Evans,L.H. (1992) J. Virol., 66, 3683-3689. MEDLINE Abstract

2 Pathak,V.K. and Temin,H.M. (1990) Proc. Natl. Acad. Sci. USA, 87, 6019-6023. MEDLINE Abstract

3 Pathak,V.K. and Temin,H.M. (1990) Proc. Natl. Acad. Sci. USA, 87, 6024-6028. MEDLINE Abstract

4 Pathak,V.K. and Temin,H.M. (1992) J. Virol., 66, 3093-3100. MEDLINE Abstract

5 Pusinelli,G.A. and Temin,H.M. (1991) J. Virol., 65, 4786-4797. MEDLINE Abstract

6 Varela-Echavarria,A., Garvey,N., Preston,B.D. and Dougherty,J.P. (1992) J. Biol. Chem., 267, 24681-24688.

7 Mansky,L.L. and Temin,H.M. (1994) J. Virol., 68, 494-499. MEDLINE Abstract

8 Mansky,L.M. and Temin,H.M. (1995) J. Virol., 69, 5087-5094. MEDLINE Abstract

9 Vartanian,J.P., Meyerhans,A., Åsjö,B. and Wain-Hobson,S. (1991) J. Virol., 65, 1779-1788. MEDLINE Abstract

10 Vartanian,J.P., Meyerhans,A., Sala,M. and Wain-Hobson,S. (1994) Proc. Natl. Acad. Sci. USA, 91, 3092-3096. MEDLINE Abstract

11 Johnson,P.R., Hamm,T.E., Goldstein,S., Kitov,S. and Hirsch,V.M. (1991) Virology, 185, 217-228. MEDLINE Abstract

12 Gao,F., Yue,L., White,A.T., Pappas,P.G., Barchue,J., Hanson,A.P., Greene,B.M., Sharp,P.M., Shaw,G.M. and Hahn,B.H. (1992) Nature (Lond.), 358, 495-499.

13 Perry,S.T., Flaherty,M.T., Kelley,M.J., Clabough,D.L., Tronick,S.R., Coggins,L., Whetter,L., Lengel,C.R. and Fuller,F. (1992) J. Virol., 66, 4085-4097. MEDLINE Abstract

14 Wain-Hobson,S., Sonigo,P., Guyader,M., Gazit,A. and Henry,M. (1995) Virology, 209, 297-303.

15 Martinez,M.A., Vartanian,J.P. and Wain-Hobson,S. (1994) Proc. Natl. Acad. Sci. USA, 91, 11787-11791. MEDLINE Abstract

16 Martinez,M., Sala,M., Vartanian,J.P. and Wain-Hobson,S. (1995) Nucleic Acids Res., 23, 2573-2578. MEDLINE Abstract

17 Martinez,M.A., Pezo,V., Marliere,P. and Wain-Hobson,S. (1996) EMBO J., in press.

18 Dahbo,Y. and Eriksson,S. (1985) Eur. J. Biochem., 150, 429-434. MEDLINE Abstract

Return

* To whom correspondence should be addressed
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 J. Virol.Home page
P. Blancou, N. Chenciner, R. H. T. Fang, V. Monceaux, M.-C. Cumont, D. Guetard, B. Hurtrel, and S. Wain-Hobson
Simian Immunodeficiency Virus Promoter Exchange Results in a Highly Attenuated Strain That Protects against Uncloned Challenge Virus
J. Virol., February 1, 2004; 78(3): 1080 - 1092.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Print PDF (98K) Freely available
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrowRequest Permissions
Right arrow Commercial Re-use Guidelines
for Open Access NAR Content
Google Scholar
Right arrow Articles by Pezo, V
Right arrow Articles by Wain-Hobson, S
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Pezo, V
Right arrow Articles by Wain-Hobson, S
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?