|Home | About | Journals | Submit | Contact Us | Français|
We evaluated the effect of the T4 bacteriophage gene 32 protein (T4gp32) on in vitro transcription and reverse transcription. T4gp32 doubled the yield of in vitro transcripts obtained with T7 RNA polymerase and increased the yield of cDNA synthesis when used in combination with an RNaseH-deficient Moloney murine leukemia virus [Au: ok] reverse transcriptase. The positive effect could be correlated with the RNA chaperone activity of T4gp32. T4gp32 stimulated the synthesis of long cDNAs, particularly species longer than 7 kb. By comparison, thermal activation of reverse transcriptase with trehalose only boosted the production of shorter cDNAs. For the construction of an Arabidopsis thaliana cDNA library, where the average cDNA size is 1.2 kbp, both the presence of T4gp32 under standard reaction conditions as well as thermal activation resulted in similarly high percentages of full-length cDNA. However, the inclusion of T4gp32 in a standard reverse transcription reaction resulted in the highest cDNA yield. We conclude that the addition of T4gp32 in standard reverse transcription reactions can increase the quality and yield of full-length cDNA libraries.
The availability of full-length cDNA libraries greatly facilitates the rapid analysis of gene functions. In standard cDNA libraries, the percentage of full-length cDNAs has been estimated to be around 30%.1 Most of the partial cDNAs arise from the blockage of the reverse transcriptase (RT) at a hairpin or another structured region of the mRNA template.2 Recently, full-length cDNAs libraries have been constructed using trehalose to thermoactivate the RT, allowing the RT reaction to be performed between 50°C and 60°C.3;4 Point-mutated avian myeloblastosis virus (AMV) RT with no RNase H activity and a high temperature optimum is also commercially available now and has been successfully employed in reverse transcriptase polymerase chain reaction (RT-PCR) and 5′–RACE (rapid amplification of cDNA ends) with highly structured mRNA templates that could not be reverse transcribed under standard conditions.5
Unfortunately, although high temperature contributes to resolving secondary structures of RNA, it also augments the breakdown of RNA, which is thermolabile and susceptible to metal-catalyzed degradation. Since reverse transcriptases have an absolute requirement for divalent cations, significant RNA degradation can occur by incubation in RT buffer at a temperature higher than 55°C.6
In living cells, RT copes with hairpins and stably folded RNA templates with the help of RNA chaperones. All retroviruses described so far have an RNA chaperone, the nucleocapsid (NC) protein. The NC protein acts as an essential accessory protein at many steps of the retrovirus’ life cycle including the dimerization of viral RNA, the formation of primer tRNA:viral template hybrid, strand transfer steps after the initiation of reverse transcription, and the removal of hairpins in RNA templates during reverse transcription. NC proteins are small, highly basic proteins that bind viral RNA at high affinity, and other RNA at lower affinity. They act as RNA chaperones (reviewed in ref. 7). However, the strong sequence preference of retroviral NC proteins for their respective long-terminal repeats RNA8 restrains the usefulness of these chaperones for cDNA library construction.
In contrast, T4gp32 shows very little if any sequence preference. The requirement of T4gp32 for bacteriophage replication and its implication in many steps of the bacteriophage life cycle is reminiscent of the functions associated with the retroviral NC proteins.9 Since its isolation, T4gp32 has been extensively studied as a single-stranded (ss) DNA binding protein necessary for T4 bacteriophage replication (reviewed in ref. 10). During replication, T4gp32 molecules cooperatively bind ssDNA dissociated by T4 helicase. The relatively low affinity of the chaperone for DNA allows it to slide along the DNA strands, generating unstructured ssDNA while the following T4 polymerase advances over the DNA template.11
T4gp32 has been used as a molecular tool to enhance numerous enzymatic reactions including PCR,12,13 sequencing,14 reverse transcription in RT-PCR,15 and cDNA synthesis.16 The effect of T4gp32 directly on RT, however, has so far not been studied and has only been observed after PCR amplification.
We hypothesized that T4gp32 could improve reverse transcription by preventing the formation of hairpins and other structures commonly found in mRNA templates, thus increasing full-length cDNA synthesis. In this report, we evaluate the effect of T4gp32 on the reverse transcription of a population of five polyadenylated RNAs ranging in size from 1.3 to 9.5 kb as well as on mRNA isolated from Arabidopsis thaliana. We also compare the efficacy of full-length cDNAs synthesis when performed either in the presence of T4gp32, or at elevated temperatures, using both trehalose-thermostabilization and heat-stable reverse transcriptase.
Chemical reagents and oligos [SfiB-dT (ATTCTAGAGGCCGCCTCGGCCGTGCAG T(22)vn) and SfiC-C11 (CACACACACACACACAGGCCATTTAGGCCGTGCAGC(11))] were purchased from Sigma Aldrich and Sigma Genosys, respectively, (Oakville, ON, Canada). Restriction enzymes were from New England Biolabs (Missisauga, ON, Canada). Taq polymerase, terminal transferase, T4 DNA ligase and some restriction enzymes were from Fermentas (Burlington, ON, Canada). Bacterial strain DH5_ FT, reverse transcriptase Superscript II, Thermoscript, 0.24–9.5 kb polyadenylated RNA ladder, rNTPs, and dNTPs were purchased from Invitrogen (Burlington, ON, Canada). Proteinase K, Streptavidin-coated paramagnetic particles, High Pure spin columns, T4gp32, bovine serum albumin, glycogen, plasmid pHelix 1(+) were from Roche Diagnostics (Laval, QC, Canada). RNase ONE, Riboprobe in vitro transcription kit, m7GTP cap analog were from Promega (Madison, WI). MEGAscript T7 in vitro transcription kit was from Ambion (Austin, TX). ExTaq was purchased from TaKaRa (Fisher Scientific, Ottawa, ON, Canada). ImageQuant software, Phosphorscreen and 32P nucleotides were from Amersham Biosciences (Baie d’Urfé, QC, Canada). Plasmid pHelix1(+) was modified to give pHSX-Ci by inserting an adapter between the Eco RI and Kpn I site which contains two Sfi I sites, ggccgcctcggcc (Sfi I-B) and ggccatttaggcc (Sfi I-C) to accommodate directional cloning of cDNA inserts.
An increasing amount of T4gp32 (0.04 μM, 0.2 μM, 1 μM, 5 μM), or 166 μg/mL bovine serum albumin for the negative control reaction, was added to approximately 100 ng of heat-denatured, radiolabeled 1.89 kb RNA transcript in 15 μL reactions containing 10 mM Tris pH 8.0, 1 mM EDTA, 50 mM KCl, 7.4% glycerol, and incubated for 10 minutes at room temperature. 2.2 μL loading dye (40% sucrose, Xylene Cyanol, Bromophenol Blue) were added to each reaction before loading on a 6 cm long, 4% acrylamide:bisacrylamide gel that had been previously pre-run for 20 minutes at 100V and cooled to 4°C. Electrophoresis was for 3 hours at 100V. The RNA was visualized with a Phosphor Imager.
A radiolabeled 1.89 kb RNA transcript was synthesized from the pTRI-Xef positive control DNA provided with the MEGAscript T7 in vitro transcription kit and resulted in the 1.89 kb, polyadenylated RNA containing the Xenopus elongation factor 1α gene (GeneBank Acc. no. BC041196). Typically, in vitro transcription was performed on 0.5–1 μg linearized DNA template in a 20-μL reaction using the Riboprobe in vitro transcription system according to the manufacturer’s instructions. The 1.89 kb RNA was further purified on a denaturing poly-acrylamide gel as described in Sambrook and Russel.2 As a positive control for CAP trapper experiments, a 1.3-kb capped RNA was synthesized from a maize cDNA clone containing a poly A-tailed homolog of a putative protein phosphatase (GeneBank Acc. No. AY112251.1), using the T3 Riboprobe in vitro transcription system. The 20-μL reaction contained 1 μg of linearized DNA template, 0.5 mM of each rATP, rUTP and rCTP, 0.05 mM rGTP, 0.5 mM of Ribo m7G Cap Analog, 20 units of T7 RNA polymerase, buffer and dithiothreitol as recommended by the supplier. The reaction was incubated for 2 hours at 37°C. The ratio of 10:1 of cap analog to rGTP is expected to result in over 90% of transcripts that are capped.
In the first assay, the 0.24–9.5 kb polyadenylated RNA ladder was used as template for reverse transcription. Briefly, 200 ng template and 50 pmole SfiB-dT primer were heat-denatured for 2 min at 65°C and quickly cooled on ice. T4gp32 (final conc. 0.12, 0.6, and 3 μM), was added and the mix incubated for 5 min at room temperature before addition of the RT mix (final 1 X Superscript II buffer, 10 mM dithiothreitol, 1 mM dNTPs, 1 μL 32P(dCTP) (3000 Ci/mmole), 100 units Superscript II) and incubated for 15 min at 37°C. The RT was stopped by the addition of ethylenediaminetetraacetic acid to 10 mM, sodium dodecyl sulfate (0.2%) and Proteinase K (70 μg/mL) in a total volume of 200 μL. The reaction was incubated for 15 min at 45°C before phenol:chloroform:isoamyl (24:24:2) extraction and alcohol precipitation with 5 μg of glycogen. The radioactive pellet was resuspended in 10 μL water. A 3-μL aliquot was loaded on a 1% alkaline agarose gel which was run at 100 mA until the bromophenol blue reached two-thirds of the length of the gel, fixed in 10% trichloroacetic acid, dried and exposed to a Phosphorscreen. The volume of each band was quantified with the ImageQuant software after correcting the background using the local median.
In the second assay, two RT reactions were assembled as described above except that one contained 4.5 μM T4gp32 and the other 23% trehalose. The reaction containing trehalose was placed in a thermocycler programmed to cycle five times between 60°C (10 sec) and 50°C (5 min). The reaction containing T4gp32 was split in two: one half was incubated as described for the first assay, and the other half was incubated as described for the second assay. The completed reactions were processed as described above.
Six tubes, each containing 1 μg of A. thaliana mRNA, isolated as described elsewhere,17 and 180 pmoles of SfiB-dT primer in a total volume of 4 μL, were placed at 65°C for 5 min then immediately cooled on ice. Three micrograms of T4gp32 were added on ice to tubes 2, 4 and 6. Reactions 1 and 2 were reverse transcribed with 100 units of Superscript II (Invitrogen) in a 20-μL reaction containing 1X Superscript II buffer, 10 mM dithiothreitol, 0.5 mM dNTPs and 1 μL 32P(dCTP). Reactions 3 and 4 were reverse transcribed with the same mix as 1 and 2, plus 23% trehalose. Reactions 5 and 6 were reverse transcribed with 1 μL of Thermoscript in 20 μL reaction volume. Reactions 1 and 2 were incubated for 40 min at 45°C. Reactions 3 to 6 were assembled at 45°C, then immediately transferred to a thermocycler pre-heated to 60°C and programmed to cycle five times between 60°C 10 sec) and 50°C (5 min). The entire routine took 40 min. All the reactions were stopped at the same time and treated with Proteinase K as described above. The efficiency of the reverse transcription was evaluated by counting 0.05 volumes of the resuspended first strand cDNA in 5-mL scintillation cocktail.
The CAP trapper method was performed as described elsewhere.18 The captured RNA:cDNA duplexes were eluted by a 5-min incubation at room temperature in 200 μL 0.2 M NaOH then ethanol precipitated. Both the supernatant and eluate fractions were loaded on an alkaline 1% agarose gel as described above. The volume of each cDNA smear was quantified with the Image Quant software after correcting the background using the local object feature.
Two controls were performed in parallel with the RT samples. The positive control template was a capped, polyadenylated 1.3 kb in vitro transcript, reverse transcribed and CAP trapped in the same conditions as the RT reaction 1. As a negative control, we reverse transcribed with RT condition 1, and CAP trapped the 0.24–9.5 kb RNA ladder in which the RNA fragments are polyadenylated but uncapped. For both controls, we quantified the CAP trapped and the non-trapped fractions using the ImageQuant software as we did for samples 1 to 6.
RT condition 2 was scaled up in order to build an A. thaliana full-length cDNA library. Ten micrograms mRNA was used as starting material. RT was performed as described above in the presence of 1 nmole SfiB-dT primer and 20 μg T4gp32, in a total volume of 40 μL. CAP trapping was performed exactly as described above. The CAP trapped, full-length, enriched first strand cDNA was dG-tailed in presence of 0.1 mM dGTP, 1X TdT buffer, and 40 Units Terminal transferase in a total volume of 20 μL for 15 min at 37°C. After phenol:chloroform extraction and ethanol precipitation, the cDNA pellet was resuspended in 20 μL H2O and rendered double-stranded with 10 Units ExTaq thermostable polymerase in 1X ExTaq buffer, 1 μM SfiC-C11 primer and 0.125 mM dNTPs in a total volume of 100 μL. The reaction was incubated in a thermocycler (Biometra, Montreal, QC, Canada) for 2 min at 95°C, 30 min at 68°C, and 10 min at 72°C. After ethanol precipitation and resuspension in 40 μL H2O, cDNA was digested with 100 Units Sfi I in a 100-μL reaction for 3 h at 50°C. The cDNA was size fractionated on a CL-4B column equilibrated in STE buffer as described elsewhere.2 Fractions containing cDNA greater than 0.3 kbp were pooled, precipitated, and ligated into pHSX-Ci. The ligation mix was used to transform ultra-competent DH5α-FT cells. A library titer of 1.4 × 106 was obtained. A sample of 200 clones was sequenced from the 5′ end with a Li-Cor sequencer. Raw sequences were trimmed and submitted to a Mega BLAST (NCBI) analysis. Hits were analysed manually for the presence of the AUG start codon.
A radiolabeled 1.89 kb in vitro transcript was used as template. The 1.89 kb RNA alone migrated as a sharp band on a denaturing gel (Fig. 1A1A),), but as a smear on a native polyacrylamide gel (Fig. 1B1B,, lane 1). The presence of bovine serum albumin at 166 μg/mL (Fig. 1B1B,, lane 2) or T4gp32 at the lowest concentration tested (0.04 and 0.2 μM, lanes 3 and 4) did not alter the migration pattern of this RNA. However, at a T4gp32 concentration of 1 μM, and even more at 5 μM, the RNA smear was compressed into a sharper, slower migrating cluster of bands, presumably resulting from the relaxation of secondary RNA structures (Figure 1B1B,, lanes 5, 6).
The effect of T4gp32 on in vitro transcription (Figure 2A2A)) and on reverse transcription (Figure 2B2B)) was evaluated. No significant effect was observed at the lower concentrations of T4gp32 tested (0.12 and 0.6 μM). However, at a concentration of 3 μM, a positive effect was observed for both, the in vitro transcription (1.8–2-fold increase; n=2) of a 1.89 kb template and the reverse transcription (1.1–1.4-fold increase; n=4) of a 2.4 kb RNA template (Figure 2C2C)) Superscript II was used at 37°C instead of 42–45°C as recommended by the manufacturer, in order to exemplify the potential effect of the chaperone. We also repeated the same experiment with incubations at 42°C and 45°C, for 5 min to 1.5 hours, always with similar results (not shown).
Figure 33 shows the comparison of five different RT reactions of an RNA ladder, composed of 5 poly-adenylated RNAs, devoid of secondary structures based on computer prediction (Vienna RNA folding procedure according to Zuker’s RNA folding method) using GeneQuest 5 (DNAStar) with a set temperature of 37°C, and ranging from 1.4 kb to 9.5 kb. The radiolabeled cDNAs synthesized under each condition are shown in Figure 3A3A;; a histogram of the size distribution of the cDNA resulting from each condition is shown in Figure 3B3B.. In this experiment, the only two factors that resulted in an increased synthesis of cDNA compared with the standard condition (lane 1) were the presence of T4gp32 at 45°C (1.3-fold overall stimulation, lane 2) and the trehalose stabilisation at 50–60°C (2-fold overall stimulation, lane 5). The stimulating effect of T4gp32 is more pronounced on longer cDNA products: 1.1-fold for the 1.4 kb band, 1.3-fold for the 2.2–7.5 kb bands and 1.7-fold for the 9.5 kb band. The combination of trehalose/high incubation temperature favoured the synthesis of all cDNAs, except the longer 9.5 kb product: a 2-fold increase for the cDNA bands below 7.5 kb, compared with a 1.7-fold increase for the 7.5 kb and a 0.3-fold increase for the 9.5 kb product. Incubation at a high cycling temperature (50–60°C) without the addition of trehalose clearly led to a reduced cDNA yield (lane 3) even in the presence of T4gp32 (lane 4). Similar observations were obtained in three independent experiments (not shown).
We tested the effect of T4gp32 on the first strand synthesis of A. thaliana mRNA using three alternative protocols. First, we reverse transcribed with Superscript II for 40 min at 45°C, in the absence (RT1) versus the presence (RT2) of T4gp32. Figure 4A4A shows the unbound and incomplete cDNA (−) fractions, as compared with the bound and full-length cDNA (+) fractions, for each RT condition. These results are visualized in the histogram presented in Figure 4B4B,, which shows the RT efficiency for each RT condition and the percentage of first-strand cDNA that was CAP trapped. The presence of T4gp32 had very little effect on the overall RT efficiency (RT1: 19.6% without T4gp32; RT2: 21.2% with T4gp32), but improved drastically the amount of full-length cDNA that was CAP trapped (from 26.6% without to 44.3% with T4gp32).
In a second setup (RT3 and RT4), we tested the reverse transcriptase (Superscript II) thermostabilized with 23% trehalose. The RT reaction was carried out by cycling five times between 60°C (10 sec) and 50°C (5 min). This approach is comparable to what has been recommended elsewhere.3;4 The reaction efficiency was comparable to that obtained with RT1 and RT2: 21.2% without T4gp32, (RT3) and 19.2% with T4gp32, (RT4). Under these conditions, the presence of T4gp32 did not affect the population of CAP trapped cDNA (34.6% without and 35.1% with T4gp32).
In a third setup (RT5 and RT6), the reaction conditions were identical to that of RT3 and RT4, but a thermostable enzyme (Thermoscript) was used instead and trehalose was omitted from the reaction. The overall RT efficiency was lower (11.2% without, and 13.2% with T4gp32) than what was observed in RT1–4); however, the proportion of full-length cDNA was high, even in the absence of T4gp32 (41.7%, RT5), but did increase to 44.2% in the presence of T4gp32 (RT6).
The efficiency and specificity of the method was evaluated by analyzing the positive and negative controls. The CAP trapped positive control amounted to 50% of bound material (Fig. 4A4A,, lane 3). Most of the unbound cDNA was smaller, partial cDNA (Fig. 4A4A,, lane 2). Less than 5% of the negative control remained on the beads (not shown).
Both the RT efficiency (25%) and the proportion of CAP trapped cDNA (42%) achieved in the full-length cDNA library experiment were similar to that obtained for the small-scale reaction (Fig. 3A3A)) whereby the positive control was CAP trapped at 43% and less than 5% of the negative, uncapped control remained on the beads (not shown). The library titer was 1.4 × 106 cfu, with less than 1% background of vector colonies. The cDNA insert size averaged 1.5 kbp. The sequences of 200 expressed sequence tags (ESTs) from the RT2 library were subjected to a BLAST search using the NCBI A. thaliana database. Of these, 138 sequences matched with annotated complete coding sequences (cds), of which 93% had an AUG start codon. The average 5′ UTR length was 95 bp.
The CAP trapper selection of A. thaliana full-length cDNAs was repeated with Superscript III reverse transcriptase. A higher yield of cDNAs was achieved compared with Superscript II, but the percentage of CAP trapped full-length cDNA was similar (not shown). Other plant full-length cDNA libraries were also obtained with Superscript III, including one library from developing corn kernels and one library from developing wheat endosperm. In all cases, the presence of T4gp32 protein resulted in a CAP trapped selection of around 40% of the total cDNAs (not shown).
The addition of T4gp32 to in vitro transcription as well as to reverse transcription reactions led to an increase in product quantity compared with standard reactions. This stimulation was observed at a T4gp32 concentration of 3 to 4.5 μM, which is in the range of the minimal concentration required for cooperative binding in vitro.19 The effect of T4gp32 in RT reactions correlates well with its effect on RNA conformation, suggesting that T4gp32 exerts this mechanism through nucleic acid chaperone activity.
In the case of in vitro transcription, binding of T4gp32 to nascent RNA could lead to an increased template recycling. It has been shown recently, that E. coli ssDNA binding protein binds RNA newly synthesized by the bacteriophage N4 virion RNA polymerase (vRNAP), thus preventing it fromforming a stable RNA:DNA duplex with the DNA template. Such RNA-protein interaction is essential for template recycling because the vRNAP does not have any RNA binding domain, like T7 RNA polymerase does.20 Moreover, T7 in vitro transcription was shown to be enhanced by the binding of the E. coli cold shock protein A (CspA), an RNA chaperone, to the nascent transcript.21 A similar mechanism could take place with T4gp32.
In the case of reverse transcription, considerable evidence suggests that binding of T4gp32 to RNA increases full-length cDNA synthesis. Interaction of T4gp32 with four different components of the RT system is feasible:
The comparison of the effects of heat and T4gp32 binding on reverse transcription led to the realization that, if both treatments increased the overall cDNA synthesis, their effects were qualitatively not the same. When a mix of five RNA templates ranging in size from 1.4 to 9.5 kb was reverse transcribed under elevated temperatures, a strong over-representation of the low- to medium-sized cDNAs and a diminished proportion of the very long cDNA (9.5 kb) could be observed. Similar results were reported when both trehalose and sorbitol were used to thermoactivate the reverse transcriptase.4 We interpret this as the consequence of increased reaction efficiency at high temperature, but coupled with an increased thermal degradation of the RNA template. In contrast, T4gp32 binding tended to counterbalance the size bias observed in cDNA synthesis. As a result, the cDNA population synthesized in the presence of T4gp32 better reflects the actual size distribution of the mRNA population
In addition to the structured RT experiments with defined RNA templates, we also evaluated the usefulness of T4gp32 in full-length cDNA library production. We observed that the efficiency of reverse transcription of an mRNA population is somewhat dependent on the type of reverse transcriptase used, but is not influenced much by the presence of T4gp32 since the average size of an A. thaliana cDNA is only 1.2 kbp, and the largest cDNA isolated so far from that plant is 7.5 kbp.29 The amount of CAP trapped cDNA, however, was clearly larger in the presence of T4gp32 than in the standard reaction, reflecting perhaps a higher degree of secondary structures in the population than in the controlled experiment. In our experiments, the maximal proportion of cDNA that was CAP trapped was around 45%. This result was achieved in a standard reaction with T4gp32 at 45°C as well as in RT reactions at elevated temperatures using a thermostable enzyme with and without T4gp32. The effect of T4gp32 was pronounced at 45°C, but almost not noticeable at 50–60°C. A comparison of the two best conditions revealed that both reactions produced a comparative proportion of full-length cDNAs, but the standard reaction with T4gp32 gave more cDNA products.
Altogether, these data indicate that T4gp32 has a positive effect on in vitro transcription reactions as well as on the reverse transcription of mRNAs. This could represent a significant improvement for the production of full-length cDNA libraries, in particular mammalian cDNA libraries, where the average cDNA size is considerably larger than in plants.
This work was entirely funded by the Canadian Crops Genomics Initiative. We would like to thank Hélène Rocheleau and Titus Tao for helpful assistance. The helpful discussions of this work with Dr. Laurian Robert and Dr. David Wilkinson are gratefully acknowledged.