Larissa Karnaoukhova, Pei-Zhong Tang, Deming Wang, Lansha Peng, Barry Neiditch

click the image to enlarge Table 1. Libraries constructed from column-fractionated (A) and gel-fractionated (B) cDNA.

Introduction

A cDNA library is a pool of cDNA clones that represent the profile of mRNA expression in particular tissues or cells. cDNA libraries are used not only as a source of cDNA clones for gene discovery and protein expression, but can also be applied through normalization and subtraction techniques to study differential gene expression.^(1-4) In addition, gene expression analysis of micro-dissected and single cells creates a need for cDNA library construction from limited starting material.⁽⁵⁾ These applications demand the development of new cDNA library construction methods that afford high quality cDNA library standards including 5-10 million total colony forming units (cfu), a high percentage of clones with legitimate cDNA inserts (100% recombinants) and full-length cDNA clones representing the complete sequence of a gene (>70%). The recently developed CloneMiner (Invitrogen Corp., Carlsbad, CA) method of cloning uncut cDNA by recombination achieves just that.^(6,7) It is based on Gateway (Invitrogen Corp.) recombination cloning, which has been shown to be 2.5-5 times more efficient than DNA ligation-based cloning.^(8-10) Here we show the application of the CloneMiner method for library construction from a small amount of starting material and with a better representation of full-length cDNA. The CloneMiner libraries are Gateway-compatible and can be easily transferred into any Gateway destination vector, creating an expression-ready library for functional analysis. The CloneMiner method was also used to construct nano-quantity, normalized and full-length enriched cDNA libraries.

click the image to enlarge Table 2. Full-length enriched libraries.

Materials and methods

Tissues, cells and mRNA

Tissue (Pel-Freez, Arkansas), HeLa and 293 cells (Cellexbio) mRNA was extracted using FastTrack MAG mRNA isolation kits (Invitrogen).

Library construction

Libraries were constructed using the reagents and protocol provided with the CloneMiner kit. First strand synthesis reactions were performed using a biotin-attB2-oligo dT and SuperScript II reverse transcriptase (RT). After second strand synthesis and end-polishing reactions, the phosphorylated double-stranded attB1 adapter was ligated to the 5’ end of the cDNA. Adapted cDNA was size fractionated on a cDNA size fractionation column or by agarose gel electrophoresis as specified in the CloneMiner manual. For gel fractionation, cDNA of a certain size was excised from the gel, extracted with gel phenol, and precipitated. The fractionated cDNA was used in a BP reaction with pDONR 222 vector in the presence of BP Clonase. Following inactivation and ethanol precipitation the recombination reaction was used to electroporate ElectroMAX DH10B T1 Phage-Resistant competent cells. After electroporation the bacteria were incubated and plated at different amounts on L-broth kanamycin agar plates for library titer and total colony forming units (cfu) determination. For each library, 24 randomly selected clones were analyzed by BsrG I restriction digest analysis to determine the average insert size (AIS), % recombinants, and the insert size range according to the protocols provided in the CloneMiner manual.

LR library transfer

Library transfer reactions were done following the protocol in the CloneMiner manual. Briefly, the entire entry clone library was inoculated in LB media containing kanamycin and incubated for 6 h at 37 C. Plasmid DNA was prepared using the S.N.A.P Midiprep Kit (Invitrogen) and recombined with various destination vectors using the LR reactions in the presence of LR Clonase. The LR reactions were incubated at 25 C for 16 h and electroporated into ElectroMAX DH10B T1 Phage-Resistant competent cells. The resulting expression libraries were analyzed to determine the library titer, total cfu, AIS, and % recombinants according to the protocols provided in the CloneMiner manual.

click the image to enlarge Table 3. cDNA library transfer into multiple destination vectors to create expression libraries using LR recombination reaction.

Nano-quantity library construction

Nano-quantity library construction starts with as little as 5 to 10 ng of mRNA, or total RNA (1 μg) cell (50,000) or tissue (1 mg). The quality of mRNA is critical for the success of nano-quantity library construction. The magnetic bead-based nano-quantity level of “FastTrack MAG” mRNA Isolation Kit (Invitrogen) was used to isolate mRNA from total RNA, or directly from cell or tissues. Libraries were constructed using a modified CloneMiner system with the scaled down reagents and reduced reaction volumes. Briefly, 3 pmol of Biotin-attB2-d(T)19, 80 unit of SuperScript III RT and 5 nmol of dNTP were used in a total 10 μl reaction volume for first strand cDNA synthesis. For second strand cDNA synthesis, 4 units of E. coli DNA Ligase, 10 units of E. coli DNA Polymerase I and 1 unit of E. coli RNase H were used in a 75 μl volume reaction. One μg of attB1 adaptor was used in a 10 μl volume ligation reaction. Adapted cDNA was size-fractionated using the cDNA fractionation column and fractions 5-8 were pooled and ethanol precipitated. All recovered cDNA was then used in a BP reaction with 50 ng of pDONR 222 vector.

Full-length cDNA enriched library construction

The full-length-enriched cDNA library construction method is a combined procedure of Invitrogen’s propriety full-length mRNA selection method and the CloneMiner method. Following first strand cDNA synthesis with the attB2 oligo dT and RNase I (Invitrogen) digestion of not fully transcribed mRNA, the full-length mRNA:cDNA hybrids are recovered using an antibody specific to the mRNA 5’ CAP structure. Following second strand synthesis the attB1 adapter is ligated to the 5’ ends of full-length, enriched, double-stranded cDNAs. The adapted cDNA is recombined with the pDONR 222 vector to create the pENTR library of full-length enriched cDNA clones.

Library normalization

Invitrogen’s propriety library normalization procedure was performed using mouse testis full-length-enriched CloneMiner library constructed using modified pDONR 222 and attB2 oligo dT primer. The pDONR 222 vector was modified to include the f1 site while the attB2 oligo dT primer was modified to include the SP6 sequence. Double stranded cDNA plasmid library was linearized using Gene 2 and Exonuclease III (Invitrogen) digestion. A single stranded library serves as a target in the normalization procedure. The driver, a biotinylated RNA transcript, was made using SP6 RNA polymerase following the plasmid library digestion with Not I (Invitrogen). Single stranded circular targets and biotinylated RNA drivers were then hybridized at the optimal conditions and the DNA/RNA hybrids were removed with streptavidin and phenol extraction. The hybridization conditions are set to allow the most abundant targets to hybridize to the drivers and to be eliminated. The unbound single stranded plasmids were repaired using oligo dT and DNA polymerase I to generate a normalized cDNA library.

The level of normalization was determined as a fold reduction of abundant cDNA in normalized library. About 5,000 and 20,000 colonies (from the original and normalized libraries, respectively) were plated, lifted and hybridized to a labeled probe specific for a mouse elongation factor 1 alpha (Eef1a1) gene. The probe is a PCR product labeled with [a-³²P]-dCTP by DNA polymerase I-mediated nick repair. The percent of positive colonies was calculated for both libraries and the fold reduction was determined as a ratio of the percent positive colonies in the original library to that in the normalized library.

click the image to enlarge Table 4. Nano-quantity cDNA libraries.

Results and discussion

Using as little as 30-40 ng of high quality cDNA is generally sufficient to produce CloneMiner libraries with more than 5 million primary clones, 100% recombinants and an AIS of 1.8-2.3 kb (Table 1A). Gel fractionation can increase the AIS of a library as cDNA of a certain size can be excised from the gel. Thus, selecting the cDNA of 3.5 kb increased the AIS of a library to 3.1 kb with cDNA inserts up to 7.1 kb in size while maintaining library standards such as 4.1 million primary clones and 96% recombinants level (Table 1B).

     Sequence analysis of clones from the 5’CAP antibody selected CloneMiner libraries showed that 72-86% of them contain full-length cDNA (Table 2). One advantage of the recombination cloning for the full-length protocol is that a lot less starting material is required. The CloneMiner full-length-enriched protocol worked well with 2-6 μg of mRNA compared to the 15 μg of mRNA required for the full-length enriched library constructed using restriction enzyme digestion and DNA ligation cloning.

The CloneMiner method creates cDNA libraries as Gateway? entry clones. The cDNA in the CloneMiner library is flanked by attL recombination sites. These sites can be used to transfer a clone or the entire library into any Gateway expression vector using the LR reaction. The LR library transfer has no significant effect on the library AIS or the insert size range creating unbiased fully represented expression library (Table 3).

The CloneMiner method was used to construct cDNA libraries from as low as 5 ng of mRNA, 5吆⁴ cells and 1 mg of tissue (Table 4). In all cases the total number of cfu was at least 0.36吆⁶ with an AIS of 1.3 kb and above and 91% and above recombinants (Table 4). Since the yield of total RNA per 1 mg of tissue is 0.79 μg and the yield of mRNA per 1 μg of total RNA is 30 ng, the efficiency of library construction is 1.6吆⁶ cfu per 1 μg of starting total RNA or 5吆⁴ cfu per 1 ng of mRNA. Previously described methods of cDNA library construction from limited starting material use some form of RNA amplification such as PCR or in vitro transcription to amplify cDNA or cRNA.⁽⁵⁾ Any amplification method introduces size bias resulting in shorter transcripts being over-represented in the cDNA library. In contrast, this method is free of RNA amplification affording the construction of size and representation unbiased cDNA libraries.

The normalization procedure was applied for the mouse testis full-length CloneMiner library. A 36-fold reduction of the highly abundant elongation factor 1 alpha (Eef1a1) cDNA clones was observed after the normalization (Table 5). This is comparable to the historic data for nine standard normalized libraries where the fold reduction averaged at 93 with a 16-352 range depending on the starting material, the type of the library and the gene specific probe selection (data not shown).

click the image to enlarge Table 5. Normalized full-length cDNA library.

Conclusion

It was demonstrated that the CloneMiner method of cDNA library construction can be successfully used in various library applications including full-length enrichment, normalization and nano-quantity starting mRNA. The combination of highly efficient cDNA synthesis by RNaseH- SuperScript RT and recombination cloning of uncut cDNA converts starting mRNA into plasmid clones of cDNA at a rate of 5吆4 cfu per 1 ng of mRNA with >70% full-length clones. Since the cDNA is not cut with any restriction enzymes, full-length cDNA clones never before seen in traditional libraries can be recovered. The normalized libraries can be used to study differential gene expression and the Gateway format of the libraries can be used to create expression-ready libraries in the expression system of choice for functional analysis.

References

1. Mammalian Gene Collection (MGC) Program team. Generation and initial analysis of more than 15,000 full-length human and mouse cDNA sequences. PNAS 99, 26:16899-16903 (2002).

2. Carninci, P., Shibata, Y., Hayatsu, N., Sugahara, Y., Shibata, K., Itoh, M., Konno, H., Okazaki, Y. and Muramatsu, M. Normalization and subtraction of cap-trapper-selected cDNAs to prepare full-length cDNA libraries for rapid discovery of new genes. Genome Res. Oct 10 (10):1617-1630 (2002).

3. Otsuka, M., Arai, M., Mori, M., Kato, M., Kato, N., Yokosuka, O., Ochiai, T., Takiguchi, M., Omata, M. and Seki, N. Comparing gene expression profiles in human liver, gastric, and pancreatic tissues using full-length-enriched cDNA libraries. Hepato Res. Sep 27 (1):76-82 (2003).

4. Suzuki, Y., Yoshitomo-Nakagawa, K., Maruyama, K., Suyama A. and Sugano, S. Construction and characterization of a full length-enriched and a 5’-end-enriched cDNA library. Gene Oct 24 200, (1-2):149-156 (1997).

5. Ohara, R., Kikuno, R., Kitamura, H. and Ohara, O. cDNA library construction from a small amount of RNA: adaptor-ligation approach for two-round cRNA amplification using T7 and SP6 RNA polymerases. BioTechniques 38 (3):451-458 (2005).

6. CloneMiner cDNA Library Construction Kit. Expressions (Invitrogen) 10.3: 8-9.

7. Karnaoukhova, L., Taylor, M. and Neiditch, B. Construction of cDNA libraries by recombination using the CloneMiner cDNA library construction kit. Focus (Invitrogen) 25.2; 20-24 (July 2003).

8. Hartley, J.L., Temple, G.F. and Brash MA. DNA cloning using in vitro site-specific recombination. Genome Res. 10:1788-1795 (2000).

9. Ohara, O. and Temple, G. Directional cDNA library construction assisted by the in vitro recombination reaction. Nucleic Acid Res. 29:e22-e36 (2001).

10. Ohara, O., Nagase, T., Mitsui, G., Kohga, H., Kikuno, R., Hiraoka, S., Takahashi, Y., Kitajima, S., Saga, Y. and Koseki, H. Characterization of size-fractionated cDNA libraries generated by the in vitro recombination-assisted method. DNA Res. 9,47-57 (2002). About the authors

The authors are (need short author bio). More information about constructing cDNA libraries is available from: