Introduction

Complementary DNA (cDNA) is often used in gene cloning, as gene probes or in the creation of a cDNA library, which is valuable for studying protein structure and function. The conventional method to acquire the cDNA fragment of a target gene is based on reverse transcriptase PCR (RT-PCR) amplification. Briefly, total RNA is isolated and reverse-transcribed into cDNA using a reverse transcriptase, then the resulting cDNA is used as a template for subsequent PCR amplification with primers specific for the target gene [1]. However, preparation of RNA is usually not a pleasant procedure as it requires intensive labor and high time consumption. Also, RNA is susceptible to endo- and exo-nuclease mediated degradation, rendering the initial stages of extraction and the downstream storage of the purified material more challenging than for DNA [2]. Although some strategies have been developed to prevent RNA degradation and increase yields such as thorough homogenization of tissue and cell samples, proper precipitation of RNA and optimized RNA storage conditions [3, 4], cDNAs of certain target genes are still difficult to obtain. These genes include the genes of rare transcripts, the genes of which the RNA templates are not always readily available such as those expressed in human cardiac tissue, and the long genes of which the RNA is not easy to be either reverse-transcribed in full-length using an ordinary reverse transcriptase or/and PCR-amplified using an ordinary polymerase. To solve the above-mentioned problems, several modified protocols combining chemical synthesis and PCR amplification of the target gene have recently been described for synthesis and assembly of such cDNA sequences. Representatives of these methods are the PCR-based thermodynamically balanced inside-out (TBIO) method [5], two-step total gene synthesis combining dual asymmetrical PCR (DA-PCR) and overlap-extension PCR (OE-PCR) [6], and PCR-based two-step DNA synthesis (PTDS) [7]. However, since all these strategies are based upon chemical synthesis of a number of short nucleotide sequences approx. 50–60 nucleotides in length, major drawbacks are the high cost and high error rate of chemical synthesis. Moreover, because the assembly of long sequences is carried out by performing at least two steps of PCR amplifications, this further increases the error rate. Synthesis of full-length cDNA through artificial splicing of two or even more PCR-amplified exons using overlap-extension PCR (OE-PCR) seems to be a better method, thanks to its lower cost. Using this method, several sets of primers are synthesized, which correspond, respectively, to the exons comprising the full-length cDNA. The adjacent primers share several nucleotides complementary to each other. The exons are separately PCR-amplified, then a second-round overlap-extension PCR is carried out to splice the exons [8]. The advantage of this approach is the relative low cost because only a few short oligonucleotide primers are synthesized. However, this method is somewhat time-consuming because at least two rounds of PCRs are required to obtain the full-length target gene. Moreover, it cannot guarantee a high accuracy of the obtained product, due primarily to the required minimum of two rounds of PCR amplifications. Therefore, it is of great interest to develop a simple, less error prone, and more cost-effective method that guarantees the successful synthesis of the desired gene whose cDNA is difficult to obtain by traditional reverse transcription.

In the current study, we developed a novel method to synthesize cDNA based upon the use of class IIS restriction enzymes like Eco31I. In brief, the exons of a certain gene are separately PCR-amplified, each using a primer pair containing a recognition sequence of a certain class IIS restriction enzyme. All the fragments are restricted using the enzyme(s), resulting in cohesive ends of each exon being complementary to the one in its adjacent exon. Then the fragments can be assembled together in their naturally occurring order. This method is rapid, cost-effective, and accurate as compared to previously published protocols.

Materials and Methods

PCR Amplification

The method was exemplified by the synthesis of the coding sequence of the Hoxa7 gene, a member of the Hox family. The Hox family of the clustered homeobox genes plays a fundamental role in the morphogenesis of the vertebrate embryo [9]. During organogenesis, the proteins encoded by these genes act to trigger positional identity of embryonic cells. This determines the patterning and segment identity along the anterior–posterior axis of the skeleton and a variety of organ systems. The family members share the features of temporally and spatially restricted expression. It is likely that Hoxa7 plays a role during skin development as the expression of human Hoxa7 was weakly detected in the fetal dermis during the second trimester of fetal development, but could not be observed in the newborn or adult dermis [10]. To better understand the function and regulatory process of the Hoxa7 gene, cDNA is often required for expression of the protein or generation of RNA probes. Our protocol provides an alternative method to synthesize a specific cDNA bypassing the traditional RT-PCR procedure in which RNA has to be extracted form fetal human tissues [10].

The coding sequence of Hoxa7 is composed of only two exons, of which several starting and ending nucleotides of the translated region are outlined in Fig. 1. To synthesize cDNA of the Hoxa7 gene (translated region), primers for amplification of the two exons of the gene were designed (Table 1). In the primers, we introduced the restriction sequence of Eco31I enzyme, which is a class IIS restriction endonuclease. Like other class IIS enzymes, Eco31I cuts outside its recognition site leaving 4-base 5′-overhangs that are predetermined by a downstream sequence, i.e.:

  • 5′...G G T C T C N^...3′

  • 3′...C C A G AG N N N N N^...5′ (N denotes any base).

Fig. 1
figure 1

Schematic diagram of generation of Hoxa7 cDNA. Note: The underlined bases indicate the restriction sites introduced by the primers. The starting and ending bases constituting the translated region of the exons are shaded gray

Full size image
Table 1 The primers used for generation of Hoxa7 cDNA
Full size table

In the primers, the boxed bases in F1 and R2 are the recognition sequences of restriction enzyme HindIII and EcoRI, respectively, which are to be used to facilitate the cloning of the Hoxa7 cDNA. The shaded regions in the primers R1 and F2 indicate the restriction sequence of Eco31I. Noteworthy is also that the bold bases ‘TCAG’ in F2 originate from the last four bases of the sense strand in the first exon of Hoxa7, such that the two expected 5′-overhangs (bold regions in the two primers) following Eco31I restriction are mutually complementary. The underlined regions in all the primers are homologous to the coding sequence of Hoxa7.

Two parallel PCR reactions were carried out to amplify the two exons separately in a volume of 50 μl, which included 10 ng of the human DNA extracted from blood cells, 0.4 μM of each primer, 200 μM dNTPs, 2.5 U Pfu DNA polymerase, 10 mM KCl, 10 mM (NH4)2SO4, 20 mM Tris–HCl (pH 8.8), 2 mM MgSO4, 0.1% Triton X-100, and 0.1 mg/ml nuclease-free bovine serum albumin (Fermentas). The reactions were performed in a Perkin Elmer 9600 thermocycler programed for an initial denaturation step of 3 min at 95°C, followed by 30 cycles of denaturation at 95°C for 30 s, annealing at 52°C for 30 s, and extension at 72°C for 1 min. The two PCR-amplified products were subjected to double digestion with HindIII or EcoRI (Fermentas) and Eco31I at 37°C for 2 h. Then the digested fragments were purified from 2% agarose gel using the Bioflux gel extraction kit (Bioflux).

Cloning of Hoxa7

The plasmid pUC18 was digested using HindIII and EcoRI and afterwards gel-purified. The multi-component ligation was performed in a total volume of 10 μl containing 25 ng of each of the two enzyme-digested segments, 50 ng of the cleaved vector pUC18 and 3 U T4 DNA ligase (Promega) and incubated at 16°C for 16 h. Then the ligated product was transformed into E. coli JM109 and then the cells were plated onto an LB plate containing isopropyl β-D-thiogalactoside (IPTG) and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside (X-gal). White colonies were picked and plasmids were extracted and cleaved using HindIII and EcoRI to confirm the presence of the desired cDNA. Three of the recombinant clones were randomly selected and sequenced using M13 forward primer in Yingjun biotech company.

Results and Discussion

The principle of the Eco31I restriction-based approach is outlined in Fig. 1. The most essential feature of this method suggested in this article is as follows. The two exons coding for the Hoxa7 gene (translated region) were amplified in two separate PCRs, with specially designed primers: two outer anchor primers, i.e., primers helping for the cloning, and two inner splicing primers, i.e., primers providing for the joining of the separately amplified segments. Each of the splicing primers consisted of a stretch identical to an edge of the corresponding DNA segment, and a pendant 5′-end which included a properly oriented recognition site of Eco3lI endonuclease. The unique feature of Eco3lI enzyme is that it recognizes six-base non-palindromic segments of a definite sequence and cuts DNA unilaterally with respect to the recognition site to yield a 5′-protruding 4-base overhang. This is quite different from usual class II endonucleases which recognize and cut inside palindromic sequences and result in self-complementary ends. The overhang generated by Eco31I is to be used for joining together the two DNA segments. For this goal, the adjacent splicing primers are designed in such a way that the two protruding ends at each prospective junction of the two segments are mutually complementary. Moreover, since this resulting overhang does not contain any part of Eco31I sequence, the joining will not introduce the Eco31I sequence, unlike the ligation following usual class II endonuclease treatment. Thus the joining of the segments leads to the naturally occurring sequence of the gene without introducing any undesired sequence.

In this work, the prospective four protruding bases “TCAG” in the pendent part of primer F2 originate completely from exon 1, while the protruding bases “CTGA” in the pendent part of primer R1 correspond only to the end of exon 1 itself. However, in the design of the adjacent splicing primers we had, to some extent, more flexibility. One splicing primer may have none to all of the protruding bases originating from its neighboring exon. And accordingly, the adjacent splicing primer should have all to none to ensure that the sum equals the number of the protruding bases produced by the restriction enzyme. For example, in the case of Eco31I restriction which produces a 4-base overhang, if one splicing primer contains one base from its neighboring exon, the adjacent splicing primer should have three bases from its neighboring exon such that the sum (i.e., four) equals the number of the protruding bases produced by Eco31I.

Thus, the paralleled amplifications of the two exons add the Eco31I site to the edge of the respective segment at the prospective junction. Following treatment using Eco31I, the two segments could be spliced to form the complete coding sequence of the target gene, which could be easily cloned into a suitably treated vector.

In this work, we successfully employed the Eco31I restriction enzyme to synthesize cDNA of the Hoxa7 gene. The two exons of Hoxa7 were successfully amplified with the primer pairs, as shown in Fig. 2 (lanes 1 and 2). Multi-component ligation was carried out and the end product was transformed into JM109 at an efficiency of approximately 6 × 105 cfu/mg DNA. Over 80% of the transformants were white colonies on indicator plates containing IPTG and X-gal, and all ten white colonies selected for further examination were confirmed to be recombinant since the digestions of the plasmids they contained using HindIII and EcoRI endonucleases released a fragment of approx. 700 bp (Fig. 2, lane 3). This indicated that the two separately amplified fragments were successfully spliced and ligated into pUC18 plasmid in a directional manner. Three independent recombinant clones were selected, sequenced and blasted with Genbank database. The results revealed that all the three clones harbored the correct sequence of the target gene.

Fig. 2
figure 2

Synthesis of cDNA of Hoxa7 gene by Eco31I restriction-based strategy. Two exons of Hoxa7 were successfully amplified with the primer pairs, which are 379 bp and 314 bp, respectively (lanes 1 and 2). Constructs were digested with EcoRI and HindIII endonucleases, and the DNA digests were separated on 2% agarose gel. The expected 693-bp fragment was released following cleavage of the construct (lane 3). As control, the pUC18 vector linearized using EcoRI and HindIII was run in lane 4

Full size image

Although we take the assembly of Hoxa7 coding sequence from two PCR-amplified exons as an example, the strategy described in this article is likely to work quite effectively if more exons are required to be spliced. In such cases, the target gene can be PCR-amplified into several segments determined by the number of exons it spans. Mutually complementary cohesive ends can be created among them using the strategy described herein. Then the segments can be ligated and cloned into the appropriately digested vector at one time. However, it should be noted that the protruding bases produced at different junctions should be different from one another, which is easy to achieve by designing the primers in such a way that the exons can be spliced in the proper arrangement.

We have demonstrated the high efficiency of this method in synthesizing cDNAs of several silent genes. Sometimes we use this method to synthesize and assemble redundantly expressed genes. Although this kind of genes can be amplified by using traditional RT-PCR method, our strategy has shown its attractive advantage in terms of time span and effort consumption by omitting the laborious RNA preparation and reverse transcription procedure. Also, it might be more accurate since it omits the reverse transcription which may increase the probability of mismatched nucleotide incorporation because the reverse transcriptase does not possess 3′–5′ exonuclease-dependent proofreading activity. The same approach can also be used to fuse different PCR fragments from distinct genes to create a chimeric gene or to perform site-directed mutagenesis.

As compared with overlap-extension PCR, which usually utilizes at least two rounds of PCRs for the synthesis of one fused DNA, one obvious advantage presented by our method is that only one round of PCR is required, which in turn minimizes the error rate of PCR. Accordingly, in the present study, the three randomly selected clones did not present mismatched nucleotides as confirmed by sequencing.

One concern about our method is that it may be restricted in use, due to the occasional presence of Eco31I recognition sites in target exons. However, this drawback can be easily overcome by employing other class IIS enzymes (Table 2). Moreover, different enzymes can be employed for different segments only if the enzymes used for the adjacent exons produce protruding bases of the same number.

Table 2 Alternative restriction sites of enzymes useful for the synthesis of cDNA
Full size table

Taken together, the cDNA assembly strategy described in this article provides an alternative to synthesize and amplify cDNA. It has at least the following advantages. First, it is simple, rapid, and far less effort-consuming as compared with traditional reverse transcription PCR, making it exceptionally appropriate for the synthesis of silent genes or long genes. Second, it employs only one round of PCR amplification, thus sharply limiting the mismatch of PCR.