Semi-nested PCR analysis of unknown tags on serial analysis of gene expression Wang-Jie Xu 1 , Qiao-Li Li 1 , Chen-Jiang Yao 1 , Zhao-Xia Wang 1 , Yang-Xing Zhao 1 and Zhong-Dong Qiao 1,2 1 College of Life Science and Technology, Shanghai Jiao Tong University, Shanghai, China 2 Shanghai Institute of Medical Genetics, Shanghai Jiao Tong University, Shanghai, China The serial analysis of gene expression (SAGE) tech- nique allows the construction of a comprehensive expression profile, in which each mRNA is defined by a specific 14-mer [1–4]. By analyzing a short sequence tag for a transcript, SAGE significantly decreases the overall scale of sequencing analysis and makes it possi- ble to analyze nearly all of the expressed transcripts from the genome, a capability matched by no other currently available method [5]. Application of the SAGE technique has provided valuable information in various biological systems [6,7]. Recently, millions of short cDNA sequences called SAGE tags have been collected from human tissues through the SAGE method [8,9]. It has been frequently observed that a large number of SAGE tags do not match the existing expressed sequences upon analysis of the SAGE data Keywords modified lock-docking oligo(dT); mRNA; RACE; serial analysis of gene expression (SAGE); two-step analysis of unknown SAGE tags (TSAT-PCR) Correspondence Z. Qiao, Shanghai Institute of Medical Genetics, Shanghai Jiao Tong University, Shanghai, China Fax: +86 21 54747330 Tel: +86 21 34204925 E-mail: zdqiao@sjtu.edu.cn (Received 3 August 2008, revised 3 September 2008, accepted 5 September 2008) doi:10.1111/j.1742-4658.2008.06671.x Serial analysis of gene expression (SAGE) is a powerful technique for studying gene expression at the genome level. However, short SAGE tags limit the further study of related data. In this study, in order to identify a gene, we developed a semi-nested PCR-based method called the two-step analysis of unknown SAGE tags (TSAT-PCR) to generate longer 3¢-end cDNA fragments from unknown SAGE tags. In the procedure, a modified lock-docking oligo(dT) with two degenerate nucleotide positions at the 3¢-end was used as a reverse primer to synthesize cDNAs. Afterwards, the full-length cDNAs were amplified by PCR based on 5¢-RACE and 3¢-RACE. The amplified cDNAs were then used for the subsequent two- step PCR of the TSAT-PCR process. The first-step PCR was carried out at an appropriately low annealing temperature; a SAGE tag-specific primer was used as the sense primer, and an 18 bp sequence (universal primer I) located at the 5¢-reverse primer end was used as the antisense primer. After 15–20 PCR cycles, the 3¢-end cDNA fragments containing the tag could be enriched, and the PCR products could be used as templates for the second- step PCR to obtain the specific products. The second-step PCR was per- formed with a SAGE tag-specific primer and a 22-bp sequence (universal primer II) upstream of universal primer I at the 5¢-reverse primer with a high annealing temperature. With our innovative TSAT-PCR method, we could easily obtain specific PCR products covering SAGE from those tran- scripts, especially low-abundance transcripts. It can be used as a method to identify genes expressed in different cell types. Abbreviations GLGI, generation of longer cDNA fragments from serial analysis of gene expression tags for gene identification; PLF, primary library forward primer; PLR, primary library reverse primer; RAST-PCR, rapid reverse transcription–PCR analysis of unknown serial analysis of gene expression tags; rSAGE, reverse serial analysis of gene expression; SAGE, serial analysis of gene expression; TSAT-PCR, two-step analysis of unknown SAGE tags; UP-I, universal primer I; UP-II, universal primer II. 5422 FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS [10,11]. It is possible, then, that the unmatched SAGE tags originating from potentially novel transcripts or novel genes are unidentified in the human genome. We have constructed a SAGE library on human spermatozoa in which we obtained more than 2500 unique tags. Of these, 54 were considered to be high- frequency tags, and no homology could be found in the GenBank database [12]. Therefore, those tags might represent unidentified genes. However, there was a major problem when the SAGE tag sequence was applied to the process of gene identification. Owing to the short length of SAGE tag sequences, it became dif- ficult to produce the 3¢-longer cDNA fragments and even whole cDNA sequences by PCR, which affected further studies on SAGE data. Moreover, the short tag has hindered the application of SAGE to the vast majority of eukaryotes, including expressed sequence tags and genome sequences without sufficient genomic resources [13]. In order to solve this problem, we have developed a technique called the two-step analysis of unknown SAGE tags (TSAT-PCR) to generate the 3¢-longer cDNA ends. The three key points of our method are as follows: first, it uses a modified lock-docking oligo(dT) primer, with two degenerate nucleotide positions at the 3¢-end, as a reverse primer to syn- thesize the first-strand cDNA; second, the primary cDNAs were enriched by PCR, and then served as templates for the subsequent TSAT-PCR experiment; and third, the semi-nested PCR principle was used as a reference in designing the two-step PCR method in order to obtain the 3¢-end cDNA tag-specific fragments. Currently, we have successfully used this procedure to test and analyze 11 of the 54 unmatched SAGE tags. Results and Discussion Enrichment of cDNA template Owing to RACE technology, we could now amplify full-length cDNAs to generate enough templates for the subsequent PCR, especially a few low-abundance cDNAs (Fig. 1A). In this study, the amplification of cDNAs was carried out as follows: first, owing to two degenerate nucleotide positions at the 3¢-end of the modified oligo(dT) primer in the RT-PCR pro- cess, these nucleotides position the primer at the start of the poly(A) + tail, thereby eliminating the 3¢-heter- ogeneity inherent in conventional oligo(dT) priming [14]. As the PrimeScript Reverse Transcriptase exhib- ited terminal transferase activity upon reaching the end of an RNA template, it added three to five resi- dues (predominantly dC) to the 3¢-end of the first- strand cDNA. The 5¢-cap oligonucleotide contained a terminal stretch of G residues that annealed to the dC-rich cDNA tail and served as an extended tem- plate for reverse transcription. In the subsequent PCR process, the reverse transcription product above was used as template. Primary library forward primer (PLF) and primary library reverse primer (PLR) paired with the 5¢-end and 3¢-end of all cDNAs, respectively, and after 25 cycles, the entire cDNAs were largely amplified for the next experiment. Figure 2 shows the amplified cDNAs. As can be seen, the length of the smear is distributed from about mRNA mRNA NBAAAAAAA-3′ NBAAAAAAA-3′ NBAAAAAAA Modified oligo (dT) NVTTTTTTT NBAAAAAAA NVTTTTTTT NBAAAAAAA NVTTTTTTT NBAAAAAAA NVTTTTTTT NBAAAAAAA NVTTTTTTT 16 16 16 16 16 16 16 16 5′ 5′ 5′-cap oligo NVTTTTTTT NVTTTTTTT NBAAAAAAA NVTTTTTTT GGG CCC GGG CCC GGG CCC GGG CCC GGG CCC Anneal first strand Primer to mRNA cDNA first strand synthesis Modified oligo (dT) Tag-specific primer UP-I The 2 nd PCR The 1 st PCR UP-II UP-II PLF PLR GGATCC GGATCC GGATCC cDNAs synthesis cDNA library AB Fig. 1. Detailed mechanism of the amplification of the whole cDNAs and the TSAT-PCR technique. (A) In this process, double-stranded cDNAs synthesized by modified lock-docking oligo(dT) and 5¢-cap oligonucleotides were used for PCR. During the PCR process, PLF and PLR were used as sense primer and antisense primer, respectively, to amplify the cDNAs. (B) The procedure involved two PCR reactions. The first PCR reaction was performed with a tag-specific primer containing a SAGE tag sequence and an 18 bp primer (UP-I) located at the 5¢-reverse primer end. The first PCR product was then used as the template for the second PCR reaction. The tag-specific primer and a 22-bp primer (UP-II) located near UP-I located at the 5¢-reverse primer were used as the sense primer and the antisense primer, respectively. W J. Xu et al. New method of 3¢-end amplification from SAGE tags FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS 5423 100 bp to over 2 kb, and is mostly focused on the 0.3– 1 kb range. The results demonstrate that high-abun- dance genes are not very variable in terms of length, as they mostly concentrate on a narrow span (0.3– 1 kb). Aside from the range, we can see that there are a few low-abundance genes that are either very long (50 kb) or short (50 bp). It seems that the smear of the genes did not become obvious because of their low abundance or short extension time in the PCR, or both. TSAT-PCR general strategy The amplified cDNAs served as primary templates for TSAT-PCR, as illustrated in Fig. 1B. The antisense primers [PLR, universal primer I (UP-I) and universal primer II (UP-II)] were all designed from the sequence of the modified oligo(dT) primer. The three primers shared some overlap with each other and their length was different considering the consistency of their equivalent sense primers (Fig. 3). Both UP-I and UP-II were used as nested primers in the TSAT-PCR reactions. The TSAT-PCR technique was developed from the principle of nested PCR, and the procedure included a two step-PCR reaction. For 15–20 cycles of the first PCR, an appropriately low annealing tempera- ture (about 55 °C) was used, a SAGE tag-specific primer and UP-I. As a result, the 3¢-end cDNA frag- ments containing the tag could be enriched while some nonspecific products were also generated simulta- neously, and then the PCR products could be used as templates for the second-step PCR to obtain the specific products. The second-step PCR was performed with a SAGE tag-specific primer and a nested primer (UP-II) at a high annealing temperature (‡ 60 °C). Afterwards, the specific products corresponding to tags could be amplified. Amplification of longer sequences from SAGE tags To test the TSAT-PCR procedure, we chose five tags corresponding to known genes, as well as 11 different- abundance tags corresponding to unknown genes, all identified in SAGE analysis of human spermatozoa (Table 1). Among the 16 tags, tag 4, A and E were used as representatives of low-frequency genes in order to help us determine whether or not the process worked on low-frequency tags. Upon application of the TSAT-PCR method, we obtained the PCR prod- ucts (Fig. 4) of all tags tested using the standard PCR condition (first PCR, 94 °C for 30 s, 55 °C for 30 s and 72 °C for 30 s for 15 cycles; second PCR, 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s for 25 cycles). The PCR products were electrophoresed through a 2.0% agarose gel, and cloned into a plasmid vector for sequencing analysis. As compared with the others, tag 1, 2, 3, 4 and A displayed very weak PCR bands in the agarose gel, especially the two low- frequency tags (Fig. 4). Aside from this, there were also two clear bands in the PCR product of no. 10. We further optimized the PCR annealing tempera- ture, as well as the cycle number, for each of the weak-band tags. Moreover, these bands were obviously clearer than the pervious ones (data not shown). We then verified whether or not each PCR product indeed represented a sequence downstream of the most 3¢ NlaIII site in the full-length cDNA by analyzing the M 12 3530 bp 1584 bp 947 bp 564 bp Fig. 2. PCR amplification of the full-length cDNAs. The cDNAs were amplified with PLF and PLR. M: kDNA ⁄ HindIII + EcoRI mar- ker. Lane 1: amplified full-length cDNAs. Lane 2: glyceraldehyde-3- phosphate dehydrogenase (GAPDH). GAPDH was used as control. Fig. 3. The sequences and relationships of the primers [modified oligo(dT), PLR, UP-I and UP-II] discussed in this article. New method of 3¢-end amplification from SAGE tags W J. Xu et al. 5424 FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS sequences of the products. If the tag sequence was presented at the predicted location, no NlaIII site would be present in the sequence of the obtained PCR product, whereas the PCR product would include the oligo(dT 16 ) sequence. All PCR products were cloned and sequenced successfully (Table S1). Through analy- sis of the sequencing result, we identified 16 of 17 PCR products (Figs 4 and 5) that met the standard men- tioned above. This indicates that the 16 PCR products represented a sequence downstream of the most 3¢ NlaIII restriction site. In contrast, the remaining PCR product was a large size band of the no. 10 product, in which sequences of UP-II and oligo(dT 16 ) were not found, although the tag-specific primer was found only in its sequence. This meant that the PCR product was amplified by PCR using only a single primer (the tag-specific primer no. 10). Sequencing could only determine the single primer-prone product. The sequencing results (Table 1) were analyzed using the blast program of the NCBI server (http://www. ncbi.nlm.nih.gov/BLAST/). Among the five fragments containing known tags (Table 1), four sequences corre- sponding to the tags A, B, C and E were matched to the 3¢-cDNA of genes predicted by Zhao based on the spermatozoa SAGE tags [12], whereas no. D was not matched to the gene (Hs. 436980). The reason for this was further investigated, and it was found that the no. D tag could not represent the gene (Hs. 436980), because seven NlaIII (CATG) sites were found between the site of the no. tag D tag and a poly(dA) among the cDNA of the gene (Hs. 436980). The blast results of another 11 sequences in the GenBank Table 1. Overview of all tags analyzed with the TSAT-PCR technique. The sequences from nos. 1 and 7 matched a single sequence. No. 11 matched multiclusters. The rest of the sequences did not match any clusters. Tag Tag sequence UniGene ID Abundance PCR product size (bp) Presence of NlaIII site Presence of oligo(dT) Blast results A ACTTACCTGC Hs. 431668 6 89 No Yes Consistency B GCGTGCCTGC Hs. 372658 302 211 No Yes Consistency C GCCCCTGCGC Hs. 435464 214 217 No Yes Consistency D GTGACCACGG Hs. 436980 126 189 No Yes Inconsistency E GTGGCACACG Hs. 34114 5 192 No Yes Consistency 1 AACGAGGAAT – 84 254 No Yes AK027322 2 GTAAGTGTAC – 44 97 No Yes Unmatched 3 AGAGGTGTAG – 30 232 No Yes Unmatched 4 TTGCCAACAC – 4 94 No Yes Unmatched 5 GAAGTCGGAA – 58 101 No Yes Unmatched 6 GCCGTTCTTA – 21 198 No Yes Unmatched 7 ATTAAGAGGG – 16 165 No Yes NR_003286 8 ATGCCTGTAG – 16 182 No Yes Mismatch 9 GCCTTGTTCA – 13 184 No Yes Unmatched 10 TTCTCAATGA – 10 274 No Yes Unmatched 317 No No – a 11 CCCATCGTCC – 9 123 No Yes BC010864 BC021246 BC013387 AY211920 BC092442 a Single-prime PCR product. 500 bp M 1 2 3 4 5 6 7 8 9 10 11 a b c d e 400 bp 300 bp 200 bp 100 bp Fig. 4. TSAT-PCR analysis of 16 tags. Lanes 1–11 were unknown SAGE tags corresponding to tags 1–11 in Table 1. Lanes a, b, c, d and e were known SAGE tags corresponding to tags A, B, C, D and E in Table 1. TSAT-PCR was performed as described in Results and Discussion. W J. Xu et al. New method of 3¢-end amplification from SAGE tags FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS 5425 database (refseq_rna: reference mRNA sequence and expressed sequence tags) revealed several cases (Table 1): match, multimatch, unmatch and mismatch. The corresponding accession numbers of matched and multimatched sequences are given in Table 1. No. 8 was defined as a mismatch, because the blast result showed that the site of the tag did not exactly match sequences in the GenBank database, due to nonspecific amplification. The genes corresponding to the matched sequences (corresponding to tags A and E) are Hs. 431668 (COX6B1, cytochrome c oxidase subunit Vib polypeptide 1) and Hs. 34114 [ATP1A2, ATPase, Na + ⁄ K + -transporting, a2(+) polypeptide], which are related to energy production for motility of the human spermatozoa. Hs. 372658, corresponding to no. B, is a gene coding for spermatogenesis-related protein 7, which could take part in spermatogenesis. The rest of the genes corresponding to tags C, 1 and 7 are Hs. 435464 (Homo sapiens neuritin 1-like), AK027322 (highly similar to signal recognition particle 68 kDa protein), and NR_003286 (Homo sapiens 18S ribo- somal RNA). Currently, as little is known of the func- tion of mRNAs in human spermatozoa, it was difficult to estimate whether the rest of the genes were related to the function of human spermatozoa, or just retained during spermatogenesis. For the unmatched sequences and multimatched sequences, the 5¢-RACE experiment should be carried out to obtain its full-length cDNA sequences and to determine whether the sequences represent new genes. During the course of our research on the SAGE data of the human spermatozoa, we became aware that other methods [rapid reverse transcription–PCR analy- sis of unknown SAGE tags (RAST-PCR) [15], genera- tion of longer cDNA fragments from SAGE tags for gene identification (GLGI) [16] and reverse SAGE (rSAGE) [17]] hardly generate the 3¢-fragment sequences of these unmatched tags. Although GLGI is more effective than RAST-PCR [17], the antisense pri- mer in GLGI is only composed of oligo(dT), so the rigorous PCR conditions, the Mg 2+ concentration, the number of PCR cycles and the annealing temperature would be optimized for each SAGE tag. In experi- ments, we often encountered nonspecific amplification or multiple fragments, and met difficulties in amplify- ing the product of low-frequency tags, due to the short antisense primer. The rSAGE method was derived from SAGE, and many steps and reagents are shared by these two protocols. However, step 4 (linker liga- tion) in the rSAGE protocol does not avoid self- ligation of the cDNA, and the self-ligation would lead to smearing in the following PCR amplification. In addition, the method requires more initial total RNA and poly(A) + than SAGE, because of the loss of RNA in each step. Thus, the demand for RNA restricts the application of this method during the low total RNA experiment, as each human spermatozoon is estimated to contain just 0.015 pg of total RNA [18], only 1 ⁄ 600 of the amount of somatic total RNA. To avoid this problem, we have used semi-nested PCR to improve the specific amplification, and devel- oped the method called TSAT-PCR. Using the condi- tions described in that article [17], we compared the two methods with six tags and obtained the results that we expected (Fig. 5). The bands obtained with TSAT-PCR are obviously clearer than those obtained within GLGI; moreover, the tags (4, A and E) with low abundance (< 6) were all obtained with TSAT-PCR. In comparison with other methods, ours is able to amplify our target PCR products from low-abundance transcripts. Also, the method needs a lower initial amount of mRNA than the with others. Furthermore, our method possesses the advantages of being simple, rapid, low in cost, and highly efficient. We have dem- onstrated that we could obtain a clear band of PCR products for each case, as well as enough full-length cDNAs as PCR templates for subsequent experiments through the novel PCR amplification method described above. Although the improved version of SAGE can gener- ate tags with lengths of 21 bases [19] and 26 bases [13], which theoretically can be uniquely assigned to a single 500 bp 400 bp 300 bp 200 bp 100 bp M E C B A 10 4 E C B A 10 4 TSAT-PCR GLGI Fig. 5. Comparison between GLGI and TSAT-PCR. A set of six SAGE tags was chosen for the analysis. Among the six tags, three tags (4, A and E) with low abundance (< 6) were examined. The same RNA from human spermatozoa and sense primers was used for both methods. The conditions used for GLGI followed the procedures described in [16]. New method of 3¢-end amplification from SAGE tags W J. Xu et al. 5426 FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS genomic position [20], there still exists a much earlier SAGE database constructed with the use of the conventional SAGE technique, which consists of shorter tags (14 bp). Converting short tags to 3¢-longer cDNA is a key step and a breakthrough for further studies on SAGE data. Our method would help SAGE to become a high-throughput technique that could be widely applied to gene expression. In summary, the study could be applied to further analyses of SAGE data gathered from humans and some eukaryotic species. Our approach has several important advantages, such he following: (a) it can obtain enough full-length cDNA templates for sub- sequent experiments, such as 5¢-RACE, 3¢-RACE and northern blotting, among others; (b) it can convert short SAGE tag sequences into 3¢-complementary DNAs; (c) it can obtain full-length DNA sequences containing specific tags from mRNA transcripts, espe- cially low-abundance mRNA transcripts, through the combined application of TSAT-PCR and 5¢-RACE; and (d) it can identify novel genes from SAGE data and confirm the existence of exons predicted by bio- informatic tools in genomic sequences. Experimental procedures Tag sequences In our SAGE library generated from human spermatozoa, each tag was homologously screened in the Unigene data- base (http://www.ncbi.nlm.nih.gov/SAGE/SAGEtag.cgi?tag) to identify its respective match. We chose 16 SAGE tags, including four tags corresponding to known genes, which served as a positive control for this experiment, and 11 dif- ferent-abundance tags from the 54 unmatched tags corre- sponding to unknown genes. RNA samples and cDNA synthesis Total RNA of purified spermatozoa was extracted using Trizol RNA isolation reagent (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s protocol (http://www. invitrogen.com/content/sfs/manuals/10296010.pdf). The quantity of extracted RNA was determined by UV absorp- tion. Meanwhile, cDNAs were generated with a modified RACE method through the PrimeScript Reverse Transcrip- tase (TaKaRa, Dalian, China), following the manufacturer’s instructions. Briefly, two kinds of primers were added in the RT-PCR reaction: one was the modified oligo(dT) primer (5¢-CCAGA CACTATGCTCATACGACGCAG-T 16 -VN-3¢; N =A,C,G,orT;V = A, G, or C), which was used as a reverse transcription primer to generate the first-strand cDNA; and the other was the 5¢-cap oligonucleotide primer (5¢-AAGCAGTGGTATCAACGCAGAGTACGCGGG-3¢), which annealed to the dC-rich cDNA tail and served as an extended template for reverse transcription. Thus, a set of full-length cDNAs can now serve as a primary library of spermatozoa cDNAs to be used for further studies. Amplification of primary library The full-length cDNAs in spermatozoa were amplified by PCR with the use of Takara Ex Taq Hot Start Version (TaKaRa), with the primary library sequences serving as the template. Briefly, PLF (5¢-AAGCAGTGGTATCAACGCA GAGT-3¢) was used as the sense primer, and was located at the 5¢-end of all cDNAs generated from the 5¢-cap oligonu- cleotide primer. Meanwhile, PLR, which used the sequence (5¢-CCAGACACTATGCTCATACGACG-3¢) in the 3¢-ends of all cDNAs incorporated from the reverse transcription primer, was used as the antisense primer in the PCR. The PCR program consisted of 25 cycles of 94 °C for 30 s, 66 °C for 30 s and 72 °C for 3 min. The final extension step con- sisted of 72 °C for 5 min. Ten microliters of the PCR product was checked by 1.2% agarose gel electrophoresis. TSAT-PCR The amplified primary library was diluted 10 3 -fold with sterile H 2 O for TSAT-PCR analyses. A 1-lL aliquot was directly used as a template for the first PCR amplification with the tag-specific primer (5¢-GGATCCXXXXXXXXXX, X represents each tag) and UP-I (5¢-CCAGACACTAT GCTCATA-3¢). The reaction was then carried out for 15 cycles with the following conditions: 94 °C for 30 s, 53– 55 °C for 30 s and 72 °C for 30 s extension with TaKaRa Ex Taq (TaKaRa), using a Bio-Rad Cycler (Bio-Rad, Her- cules, CA, USA). The resulting PCR product was diluted 10 3 -fold with sterile H 2 O, and a 1 lL aliquot was used as a template for the second nested PCR amplification with the tag-specific primer and UP-II (5¢-CACTATGCTCATAC GACGCAGT-3¢) with the following conditions: 25–30 cycles of 94 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s, using TaKaRa Ex Taq (TaKaRa). DNA cloning and sequencing The PCR products were cloned into pT19G-T vector (Gen- eray Biotech, Shanghai, China). Positive clones were screened by PCR with M13 reverse and M13 forward (220 bp) primers while located in the vector; sequencing reactions were performed by Sanny Bio-Tech (Shanghai, China). Acknowledgements This work was supported by Shanghai Leading Academic Discipline Project (B205). W J. Xu et al. New method of 3¢-end amplification from SAGE tags FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS 5427 References 1 Velculescu VE, Zhang L, Vogelstein B & Kinzler KW (1995) Serial analysis of gene expression. Science 270, 484–487. 2 Madden SL, Galella EA, Zhu J, Bertelsen AH & Beau- dry GA (1997) SAGE transcript profiles for p53-depen- dent growth regulation. Oncogene 15, 1079–1085. 3 Velculescu VE, Zhang L, Zhou W, Vogelstein J, Basrai MA, Bassett DE Jr, Hieter P, Vogelstein B & Kinzler KW (1997) Characterization of the yeast transcriptome. Cell 88, 243–251. 4 Zhang L, Zhou W, Velculescu VE, Kern SE, Hruban RH, Hamilton SR, Vogelstein B & Kinzler KW (1997) Gene expression profiles in normal and cancer cells. Science 276, 1268–1272. 5 Velculescu VE, Madden SL, Zhang L, Lash AE, Yu J, Rago C, Lal A, Wang CJ, Beaudry GA, Ciriello KM et al. (1999) Analysis of human transcriptomes. Nat Genet 23, 387–388. 6 Hashimoto S, Suzuki T, Dong HY, Nagai S, Yamazaki N & Matsushima K (1999) Serial analysis of gene expression in human monocyte-derived dendritic cells. Blood 94, 845–852. 7 Hibi K, Liu Q, Beaudry GA, Madden SL, Westra WH, Wehage SL, Yang SC, Heitmiller RF, Bertelsen AH, Sidransky D et al. (1998) Serial analysis of gene expression in non-small cell lung cancer. Cancer Res 58, 5690–5694. 8 Lal A, Lash AE, Altschul SF, Velculescu V, Zhang L, McLendon RE, Marra MA, Prange C, Morin PJ, Pol- yak K et al. (1999) A public database for gene expres- sion in human cancers. Cancer Res 59 , 5403–5407. 9 Boon K, Osorio EC, Greenhut SF, Schaefer CF, Shoe- maker J, Polyak K, Morin PJ, Buetow KH, Strausberg RL, De Souza SJ et al. (2003) An anatomy of normal and malignant gene expression. Proc Natl Acad Sci USA 99, 11287–11292. 10 Lee S, Zhou G, Clark T, Chen J, Rowley JD & Wang SM (2001) The pattern of gene expression in human CD15+ myeloid progenitor cells. Proc Natl Acad Sci USA, 98, 3340–3345. 11 Zhou G, Chen J, Lee S, Clark T, Rowley JD & Wang SM (2001) The pattern of gene expression in human CD34+ hematopoietic stem ⁄ progenitor cells. Proc Natl Acad Sci USA 98, 13966–13971. 12 Zhao YX, Li QL, Yao CJ, Wang ZX, Zhou Y, Wang YJ, Liu LM, Wang YF, Wang LY & Qiao ZD (2006) Characterization and quantification of mRNA tran- scripts in ejaculated spermatozoa of fertile men by serial analysis of gene expression. Hum Reprod 21, 1583–1590. 13 Matsumura H, Reuter M, Kru ¨ ger DH, Winter P, Kahl G & Terauchi R (2008) SuperSAGE. Methods Mol Biol 387, 55–70. 14 Borson ND, Sato WL & Drewes LR (1992) A lock- docking oligo(dT) primer for 5¢ and 3¢RACE PCR. PCR Methods Appl 2, 144–148. 15 van den Berg A, van der Leij J & Poppema S (1999) Serial analysis of gene expression: rapid RT-PCR analysis of unknown SAGE tags. Nucleic Acids Res 27 , e17. 16 Chen JJ, Rowley JD & Wang SM (2000) Generation of longer cDNA fragments from serial analysis of gene expression tags for gene identification. Proc Natl Acad Sci USA 97, 349–353. 17 Richards M, Tan SP, Chan WK & Bongso A (2006) Reverse serial analysis of gene expression (SAGE) char- acterization of orphan SAGE tags from human embry- onic stem cells identifies the presence of novel transcripts and antisense transcription of key pluri- potency genes. Stem Cells 24 , 1162–1173. 18 Miller D, Ostermeier GC & Krawetz SA (2005) The controversy, potential and roles of spermatozoal RNA. Trends Mol Med 11, 156–163. 19 Wahl MB, Heinzmann U & Imai K (2005) Long SAGE analysis significantly improves genome annotation: iden- tifications of novel genes and alternative transcripts in the mouse. Bioinformatics 21, 1389–1392. 20 Saha S, Sparks AB, Rago C, Akmaev V, Wang CJ, Vogelstein B, Kinzler KW & Velculescu VE (2002) Using the transcriptome to annotate the genome. Nat Biotechnol 20, 508–512. Supporting information The following supplementary material is available: Table S1. The amplified longer cDNA sequences. This supplementary material can be found in the online version of this article. Please note: Wiley-Blackwell is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. New method of 3¢-end amplification from SAGE tags W J. Xu et al. 5428 FEBS Journal 275 (2008) 5422–5428 ª 2008 The Authors Journal compilation ª 2008 FEBS . analysis of gene expression tags; rSAGE, reverse serial analysis of gene expression; SAGE, serial analysis of gene expression; TSAT -PCR, two-step analysis of unknown. to identify genes expressed in different cell types. Abbreviations GLGI, generation of longer cDNA fragments from serial analysis of gene expression tags for gene