- Email: [email protected]
Wnt-16a, a Novel Wnt-16 Isoform, Which Shows Differential Expression in Adult Human Tissues
Biochemical and Biophysical Research Communications 278, 814 – 820 (2000) doi:10.1006/bbrc.2000.3852, available online at http://www.idealibrary.com on
Wnt-16a, a Novel Wnt-16 Isoform, Which Shows Differential Expression in Adult Human Tissues Mark W. Fear,* David P. Kelsell,* Nigel K. Spurr,† and Michael R. Barnes‡ ,1 *Centre for Cutaneous Research, St. Bartholomew’s and the Royal London Hospital School of Medicine and Dentistry, 2 Newark Street, London E1 2AT, United Kingdom; and †Department of Genetic Technologies and ‡Department of Bioinformatics, SmithKline Beecham Pharmaceuticals, New Frontiers Science Park (North), Third Avenue, Harlow, Essex CM19 5AW, United Kingdom
Received October 17, 2000
The WNT genes encode a large family of secreted glycoprotein signalling molecules important from the earliest stages of development through to the adult. We have identified a novel isoform of the recently described WNT family member, Wnt16, following analysis of chromosome 7q31 genomic sequence. We find differential organisation of Wnt16 with the generation of two mRNA isoforms, Wnt16a and Wnt16b. These isoforms differ in the composition of their 5ⴕ-UTR and first exons and show evidence of differential expression. In normal human tissues, Wnt16a is expressed at significant levels only in the pancreas, whereas Wnt16b is expressed more ubiquitously with highest levels in adult kidney, placenta, brain, heart, and spleen. Wnt16 is one of a growing number of WNT genes showing evidence of distinct isoforms. We present evidence to suggest that these isoforms may be regulated from alternative promoters and discuss the potential functional differentiation afforded by these WNT isoforms. This may reveal subtle new mechanisms of regulation of WNT expression and function. © 2000 Academic Press
Key Words: WNT; Frizzled; oncogene; E2A-PBX1; isoform; gene prediction; promoter analysis.
The WNT family of secreted glycoproteins are a group of signalling molecules that have been shown to control a diverse range of developmental processes including cell fate specification, cell proliferation, cell polarity and cell migration (1, 2). In common with many developmental pathways, WNT signalling extends a significant role into oncology—loss or inappropriate activation of WNT expression is associated with a wide range of tumour types (3). Implicit in this comThe sequence reported here for the Human Wnt-16a gene has been deposited in the GenBank database under Accession No. AF152584. 1 To whom correspondence should be addressed. Fax: ⫹44 1279 622929. E-mail: [email protected]. 0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
plex array of interactions is an equally complex regulatory system. Study of this system is confounded by the size of the gene families involved. In mammals, there are 19 known members of the WNT family, complemented by a large family of putative WNT receptor proteins, encoded by the Frizzled genes, of which 10 mammalian homologues have been described. Both the WNT and Frizzled protein families share a high degree of homology within each family, which has led to speculation and some experimental support (4) of the assumption that several WNT proteins may activate individual Frizzled receptors. To add to the huge complexity of these interactions, there is also potential for expression of multiple protein isoforms from each gene. This potential diversity of interactions, which probably involves some degree of redundancy, presumably provides the complexity required for the WNT signalling pathway to fulfill such diverse roles in the many different cell types and tissues throughout development and adult life. Using in silico bioinformatic gene prediction techniques we were able to identify and confirm in vivo, that the recently described Wnt16 gene (5) actually consists of two isoforms. These Wnt16 isoforms, which we have called, Wnt16a and Wnt16b, share three of four exons, differing only in the composition of their 5⬘-UTR and first exons. Following expression analysis of both human Wnt16 isoforms, we demonstrate different expression profiles of Wnt16a and Wnt16b in adult human tissues. Bioinformatic analysis suggests the existence of alternative promoters driving the expression of each isoform, one of which may play a role in oncogenesis. This may impact on the analysis of WNT gene expression and regulation, since it may no longer be sufficient to consider each gene responsible for one product, but rather several products which may have different expression patterns and effects in diverse cells and tissues.
814
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Schematic representation of the human Wnt-16 gene. Exons are indicated by the numbers 1– 4. Alternative first exons are indicated by 1a and 1b. Locations of the alternative putative promoter regions are indicated by P1 and P2. (B) Schematic representation of the spliced Wnt16 mRNA isoforms, Wnt16a and Wnt16b. PCR primers used for expression analysis of each isoform are indicated above each isoform. †, primers for Wnt16a analysis; ‡, primers for Wnt16b analysis; §, primers used by McWhirter et al. (1999) for analysis of Wnt-16 expression (sequences common to Wnt16a and Wnt16b).
MATERIALS AND METHODS Cloning of Wnt16a. Wnt16a was cloned from placental cDNA (clontech) using nested PCR with primers designed using genomic prediction to span the ORF of Wnt16a. To reduce sequence errors, Pfu turbo polymerase (stratagene) and low cycle PCR were carried out. Outer primer sequences were Wnt16outerF (5⬘) ACCCAGTGCTCTTTCCACTG and Wnt16outerR (3⬘) CCCCAAATCATCAAAAGGTG. Nested primer sequences were Wnt16nF(5⬘) AGCCTGCAAAAACCACAGAG and Wnt16nR (3⬘) GGGATTCCACTGCAAGAGTC. 3⬘ A overhangs were added to this nested PCR product and the product was then subcloned into PCRIITOPO vector (Invitrogen, K4600-01). Sequencing was carried out on both strands of three subclones using ABI 310 and BigDye terminator sequence kit (PE Applied Biosystems). Final sequence was assembled using Lasergene package. cDNA expression analysis. PCR analysis of normal human tissues was carried out using Multiple Tissue cDNA panels (Clontech #K1420-1, K1421-1) following standard Clontech protocols. Primers used for Wnt-16a were W16aF–CAGAAAGATGGAAAGGCACC and W16aR–ATCATGCAGTTCCATCTCTC (276 bp product). Primers used for Wnt-16b were W16bF–TGCTCGTGCTGTTCCCCTAC and W16bR–ATCATGCAGTTCCATCTCTC (226 bp product). PCR conditions were 3 min at 94°C, 30 (40 s 94°C, 30 s 62°C, 1 min 72°C), 5 min 72°C.
RESULTS Identification and in Silico Gene Structure Prediction of Wnt16 The Wnt16 gene was initially identified in unfinished human genomic sequence on a BAC from chromosome 7q31.31 (AC006364) by a TBLASTN search (6) using the human Wnt7a amino acid sequence as a search probe. Further refinement of homology was achieved using the GeneWise tool (7). This homology based approach identified three exons of the Wnt16 gene, that appeared to comprise all but the first exon of Wnt16, which consisted almost entirely of the putative signal peptide region and showed no homology to other
WNT proteins. To identify the first exon of Wnt16 a combination of gene prediction, splice site and promoter prediction algorithms were performed, including GENESCAN (8), Grail (9), NNPP and Splice—two neural network methods for promoter and splice site prediction (10). These methods identified two potential first exons for the Wnt16 gene, exon 1a and exon 1b, each with a putative promoter region (Fig. 1). GENESCAN predicted exon-1b (Wnt16b) with high confidence, separated by a 179 bp intron from exon 2. A combination of splice site and promoter prediction predicted Exon-1a (Wnt16a) separated from exon 2 by a 4082 bp intron. Predicted promoter regions were also analysed for known transcription factor binding sites, using TFSearch and the TRANSFAC database (11). Transcription factor binding site analysis of both predicted promoter regions revealed, a number of developmentally significant potential transcription factor binding sites (Fig. 2). In particular, three TCF binding sites were identified upstream of exon 1a, another was identified upstream of exon 1b. The predicted properties of the putative signal peptides encoded by exon 1a and 1b differ significantly. The Wnt16b isoform has a hydrophobic signal sequence with a good predicted cleavage site (12) between residues 29 –30 (AQG-NM). The signal sequence in the Wnt16a isoform is less hydrophobic, and using the same prediction method (12) it does not have a good predicted cleavage site, however this is within the parameters of other known WNT signal peptides, e.g., Wnt7b, which also contains a weakly hydrophobic signal sequence without a good predicted cleavage site. Alignment of both Wnt16 isoforms, with the most closely related WNT proteins shows some degree of sequence conservation across signal peptide regions (Fig. 3). Both Wnt16 isoforms, share highest homology
815
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Human Wnt-16 putative promoter regions and alternative first exons. Sequence of the human Wnt16a putative promoter region (A) and the human Wnt16b putative promoter region 2 (B). Exons are indicated in upper case, introns are in lowercase. Predicted RNA Polymerase II promoter sites are indicated with dashed underlining (based on NNPP prediction). Putative transcription factor binding sites identified using TFSearch and the TRANSFAC database (Heinemeyer et al., 1998), are represented in bold with underline, the name of the site is indicated underneath along with the percentage match to the binding site identity matrix in brackets.
with mouse Wnt7b (48% identity). A phylogeny of Wnt16 in relation to other known vertebrate Wnts is presented in Fig. 4. Based on this phylogeny, Wnt16 does not appear to have a close paralog within the known vertebrate WNT family, indeed it appears to be one of the most divergent members of the WNT family, yet it has retained all of the most strongly conserved signatures of the family, including the 24 strongly conserved cysteine residues. Isolation of Putative Wnt16 Isoforms and Sequence Analysis Considering the strong supporting evidence for the alternative Wnt16 exon 1 predictions, we designed
primers to confirm the existence of these potential isoforms based on alternative organisation with exon 1a or exon 1b. Using RT-PCR we were able to identify both isoforms in cDNA libraries (data not shown). We cloned Wnt16a from placental cDNA, sequence analysis revealed an ORF of 1086 base pairs, coding for a 361 amino acid protein. This ORF concurred closely with the predicted Wnt16a gene from the BAC sequence, except for a 4 bp ACCC insertion, 11 bp upstream of the ATG start codon. The sequence was deposited in the Genbank database, under Accession No. AF152584. Subsequently to this Mcwhirter et al. (5), cloned and characterised a Wnt16 gene, 100% homologous to Wnt16b (AF169963).
816
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
Expression Analysis of Wnt16 Isoforms in Adult Human Tissues
DISCUSSION
Expression analysis of Wnt16 in normal adult tissues was carried out by PCR analysis of commercially prepared cDNA libraries using isoform specific primers. To ensure signal was obtained from mRNA only, primers were designed to span introns of the genomic sequence (Fig. 1b). In normal human tissues, Wnt16a is expressed at significant levels only in the pancreas, whereas Wnt16b is expressed more ubiquitously with highest levels in adult kidney, placenta, brain, heart, and spleen (Fig. 5).
Using a combination of in silico gene prediction and promoter characterisation techniques we have identified and confirmed in vivo, a novel Wnt16 isoform formed from alternative first exon use. Wnt16a and Wnt16b, differ in the composition of their 5-UTR, first exons and predicted signal peptide properties. This highlights an issue in dealing with gene predictions from genomic sequence. Exon prediction tools alone (e.g., GRAIL or GenScan) predict only the Wnt16b isoform. However intense analysis of a 6 kb region upstream of Wnt16 exon-2, revealed other potential
817
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Amino acid alignment of human WNT-16a and b, and closer related members of the WNT family: mouse WNT-4, human WNT-7a, mouse WNT-7b, and mouse WNT-1. Asterisks indicate the 24 strongly conserved cysteine residues, which define all members of the WNT family.
promoters and WNT signalling associated transcription factor binding sites. When all this data was combined and reviewed with a biological rationale tailored for the specific target pathway, it was possible to detect potential gene structures which eluded conventional gene prediction. This is not the first description of a WNT gene with alternative isoforms, very recently Katoh et al. (13)
FIG. 4. Rooted phylogenetic tree illustrating the putative evolutionary relationship between Wnt16 and other known mammalian Wnts. Human WNT proteins are prefixed with an h, mouse WNT proteins are prefixed with an m.
described “alternative splicing” of first exons in the Wnt2B gene to form alternative isoforms with distinct expression profiles. The authors did not explore the mechanism underlying the generation of these isoforms. We propose that Wnt16 isoforms are not generated by alternative splicing but rather, they are the product of separate promoters and hence Wnt16a and Wnt16b might be considered to be separate transcriptional and regulatory units rather than splice variants from a single promoter. There are several precedents in the literature to support this, Van den Wijngaard et al. (14) described two promoters involved in the transcriptional regulation of the human BMP-4 gene, one upstream of exon 1, the second located in intron 1, upstream of exon 2, giving rise to different transcripts in a cell type and differentiation-dependent manner. In another example, Ohnishi et al. (15) isolated a type 2C protein phosphatase (PP2Cbeta) with different 5⬘ termini. Subsequent analysis of the 5⬘ flanking regions of exon 1 and 2 by reporter gene showed that these regions acted as distinct promoters. Analysis of PP2Cbeta transcripts by RT-PCR showed that exon-1 and exon-2 derived transcripts were differentially expressed, suggesting that alternative promoter usage regulated tissue-specific expression of PP2Cbeta (15). We suggest that in a similar manner to PP2Cbeta and BMP-4, Wnt16 isoforms may also be regulated by separate promoters, with different regulatory cues. Some indication of these potential cues emerge after analysis of consensus transcription factor binding sites
818
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Tissue expression pattern of Wnt-16b and Wnt-16a mRNA in normal human tissues. Commercially available CLONTECH cDNAs prepared from various normal human tissues were used as templates for PCR. (A) Pattern of Wnt16b mRNA: 30 cycles of PCR with primers W16bF and W16bR were used, product size is 225 bp. (B) Pattern of Wnt16a mRNA: 30 cycles of PCR with primers Wnt16aF and Wnt16aR were used, product size is 275 bp. (C) Housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) control: 22 cycles of PCR were used, predicted size is 452 bp. Lanes 2–16 contain adult human tissue cDNAs (tissue indicated above lane). 100 bp ladder (Roche Molecular Biochemicals) DNA size markers are shown in lane 18.
in the 5⬘ regions of Wnt16 exon 1a and exon 1b. The 5⬘ region preceding exon 1b contains two consensus binding sites for members of the T-cell factor (Tcf) family (Fig. 2). Upon interaction with the WNT signalling transducer molecule beta-catenin, Tcf factors become potent transactivators of many developmental processes (16). Also potentially significant, is an AP-1 site. Treatment with lithium, a GSK-3 inhibitor, leads to activation of an AP-1-luciferase reporter in Xenopus embryos (17). The 5⬘ region preceding exon 1a is rich in potential transcription factor binding sites with a known role in WNT signalling. There are four consensus binding sites for members of the Tcf family. There is also an E2F1 binding site, which functions to regulate apoptosis and to suppress cell proliferation and is also found in the WNT-responsive cyclin E and cyclin D1 promoters (18). Perhaps most significantly the Wnt16a promoter contains three consensus binding sites for the oncogenic homeodomain transcription factor, E2A-Pbx1. Mcwhirter et al. (1999) recently demonstrated upregulation of Wnt16 by E2A-Pbx1 in acute lymphoblastoid leukaemia (5). Interestingly, the Wnt16 gene in this publication corresponded to Wnt16b, they did not describe the Wnt16a isoform, hence Leukaemic cell expression analysis was performed using RT-PCR primers common to both Wnt16a and Wnt16b. Hence the results would reflect the expression of both Wnt16 isoforms (Fig. 1b). Considering the tandem E2A-Pbx1 response elements in the putative Wnt16a promoter, it is tempting to speculate that Wnt16a rather than Wnt16b may be specifically up-
regulated in lymphoblastoid leukaemia. This might suggest a wider role for Wnt16a and other WNT isoforms (e.g., Wnt2b isoforms) in oncogenesis in diverse tissues. Wnt16a and Wnt16b both display very distinct and specific expression patterns in panels of normal adult tissue. Of the 16 adult tissues screened for the alternative isoforms, only the pancreas displayed significant levels of Wnt16a. In contrast, Wnt16b was more widely expressed, showing highest levels in the kidney, placenta, brain, heart, and spleen. These differential expression patterns of the Wnt16 isoforms, further support the hypothesis that alternative promoters may be driving tissue specific expression of one or the other isoform of Wnt16. The generation of functionally distinct isoforms from a single WNT gene adds further complexity to the WNT signalling paradigm. It may be that alternative promoters allow the expression of the WNT gene to be controlled by a wider range of independent pathways, so the response can be initiated by different regulatory cues. Alternatively, because the actual gene product is altered, this may alter the properties of the WNT ligand itself, perhaps by altering secretion efficiency or by altering subcellular targeting (e.g., nuclear targeting (19)). This may indicate that each WNT isoform has a different effector pathway. Coupled with the recent evidence for alternative splicing of Wnt2B (13), this evidence for isoforms and alternative promoter usage in Wnt16 suggests the complexity of regulation of WNT
819
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
signalling may be harder to unravel than previously anticipated. ACKNOWLEDGMENTS Many thanks go to Kathy Ellington and Tania Testa for technical assistance and advice. This research was supported by an MRC Collaborative studentship.
REFERENCES 1. Wodarz, A., and Nusse, R. (1999) Annu. Rev. Cell. Dev. Biol. 14, 59 – 88. 2. Nusse, R., and Varmus, H. E. (1982) Cell 31, 99 –109. 3. Cavallo, R., Rubenstein, D., and Peifer, M. (1997) Curr. Opin. Genet. Dev. 7, 459 – 466. 4. Gazit, A., Yaniv, A., Bafico, A., Pramila, T., Igarashi, M., Kitajewski, J., and Aaronson, S. A. (1999) Oncogene 18, 5959 –5966. 5. McWhirter, J. R., Neuteboom, S. T., Wancewicz, E. V., Monia, B. P., Downing, J. R., and Murre, C. (1999) Proc. Natl. Acad. Sci. USA 96, 11464 –11469. 6. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389 –3402. 7. Birney, E., Thompson, J. D., and Gibson, T. J. (1996) Nucleic Acids Res. 24, 2730 –2739. 8. Burge, C. B., and Karlin, S. (1998) Curr. Opin. Struct. Biol. 8, 346 –354.
9. Lopez, R., Larsen, F., and Prydz, H. (1994) Genomics 24, 133–136. 10. Reese, M. G., Harris, N. L., and Eeckman, F. H. (1996) Biocomputing: Proceedings of the 1996 Pacific Symposium (Hunter, L., and Klein, T. E., Eds.), World Scientific Publishing Co., Singapore. 11. Heinemeyer, T., Wingender, E., Reuter, I., Hermjakob, H., Kel, A. E., Kel, O. V., Ignatieva, E. V., Ananko, E. A., Podkolodnaya, O. A., Kolpakov, F. A., Podkolodny, N. L., and Kolchanov, N. A. (1998) Nucleic Acids Res. 26, 364 –370. 12. Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijne, G. (1997) Int. J. Neural. Syst. 8, 581–599. 13. Katoh, M., Kirikoshi, H., Saitoh, T., Sagara, N., and Koike, J. (2000) Biochem. Biophys. Res. Commun. 275, 209 –216. 14. Van den Wijngaard, A., Pijpers, M. A., Joosten, P. H., Roelofs, J. M., Van zoelen, E. J., and Olijve, W. (1999) J. Bone. Min. Res. 14, 1432–1441. 15. Ohnishi, M., Chida, N., Kobayashi, T., Wang, H., Ikeda, S., Hanada, M., Yanagawa, Y., Katsura, K., Hiraga, A., and Tamura, S. (1999) Eur. J. Biochem. 263, 736 –745. 16. Roose, J., and Clevers, H. (1999) Biochim. Biophys. Acta. 1424, M23–M37. 17. Hedgepeth, C. M., Conrad, L. J., Zhang, J., Huang, H. C., Lee, V., and Klein, P. S. (1997) Dev. Biol. 185, 82–91. 18. Inoshita, S., Terada, Y., Nakashima, O., Kuwahara, M., Sasaki, S., and Marumo, F. (1999) Kidney Int. 56, 2085–2095. 19. Baird, J. W., Nibbs, R. J., Komai-Koma, M., Connolly, J. A., Ottersbach, K., Clark-Lewis, I., Liew, F. Y., and Graham, G. J. (1999) J. Biol. Chem. 274, 33496 –33503.
820
Wnt-16a, a Novel Wnt-16 Isoform, Which Shows Differential Expression in Adult Human Tissues Mark W. Fear,* David P. Kelsell,* Nigel K. Spurr,† and Michael R. Barnes‡ ,1 *Centre for Cutaneous Research, St. Bartholomew’s and the Royal London Hospital School of Medicine and Dentistry, 2 Newark Street, London E1 2AT, United Kingdom; and †Department of Genetic Technologies and ‡Department of Bioinformatics, SmithKline Beecham Pharmaceuticals, New Frontiers Science Park (North), Third Avenue, Harlow, Essex CM19 5AW, United Kingdom
Received October 17, 2000
The WNT genes encode a large family of secreted glycoprotein signalling molecules important from the earliest stages of development through to the adult. We have identified a novel isoform of the recently described WNT family member, Wnt16, following analysis of chromosome 7q31 genomic sequence. We find differential organisation of Wnt16 with the generation of two mRNA isoforms, Wnt16a and Wnt16b. These isoforms differ in the composition of their 5ⴕ-UTR and first exons and show evidence of differential expression. In normal human tissues, Wnt16a is expressed at significant levels only in the pancreas, whereas Wnt16b is expressed more ubiquitously with highest levels in adult kidney, placenta, brain, heart, and spleen. Wnt16 is one of a growing number of WNT genes showing evidence of distinct isoforms. We present evidence to suggest that these isoforms may be regulated from alternative promoters and discuss the potential functional differentiation afforded by these WNT isoforms. This may reveal subtle new mechanisms of regulation of WNT expression and function. © 2000 Academic Press
Key Words: WNT; Frizzled; oncogene; E2A-PBX1; isoform; gene prediction; promoter analysis.
The WNT family of secreted glycoproteins are a group of signalling molecules that have been shown to control a diverse range of developmental processes including cell fate specification, cell proliferation, cell polarity and cell migration (1, 2). In common with many developmental pathways, WNT signalling extends a significant role into oncology—loss or inappropriate activation of WNT expression is associated with a wide range of tumour types (3). Implicit in this comThe sequence reported here for the Human Wnt-16a gene has been deposited in the GenBank database under Accession No. AF152584. 1 To whom correspondence should be addressed. Fax: ⫹44 1279 622929. E-mail: [email protected]. 0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
plex array of interactions is an equally complex regulatory system. Study of this system is confounded by the size of the gene families involved. In mammals, there are 19 known members of the WNT family, complemented by a large family of putative WNT receptor proteins, encoded by the Frizzled genes, of which 10 mammalian homologues have been described. Both the WNT and Frizzled protein families share a high degree of homology within each family, which has led to speculation and some experimental support (4) of the assumption that several WNT proteins may activate individual Frizzled receptors. To add to the huge complexity of these interactions, there is also potential for expression of multiple protein isoforms from each gene. This potential diversity of interactions, which probably involves some degree of redundancy, presumably provides the complexity required for the WNT signalling pathway to fulfill such diverse roles in the many different cell types and tissues throughout development and adult life. Using in silico bioinformatic gene prediction techniques we were able to identify and confirm in vivo, that the recently described Wnt16 gene (5) actually consists of two isoforms. These Wnt16 isoforms, which we have called, Wnt16a and Wnt16b, share three of four exons, differing only in the composition of their 5⬘-UTR and first exons. Following expression analysis of both human Wnt16 isoforms, we demonstrate different expression profiles of Wnt16a and Wnt16b in adult human tissues. Bioinformatic analysis suggests the existence of alternative promoters driving the expression of each isoform, one of which may play a role in oncogenesis. This may impact on the analysis of WNT gene expression and regulation, since it may no longer be sufficient to consider each gene responsible for one product, but rather several products which may have different expression patterns and effects in diverse cells and tissues.
814
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. (A) Schematic representation of the human Wnt-16 gene. Exons are indicated by the numbers 1– 4. Alternative first exons are indicated by 1a and 1b. Locations of the alternative putative promoter regions are indicated by P1 and P2. (B) Schematic representation of the spliced Wnt16 mRNA isoforms, Wnt16a and Wnt16b. PCR primers used for expression analysis of each isoform are indicated above each isoform. †, primers for Wnt16a analysis; ‡, primers for Wnt16b analysis; §, primers used by McWhirter et al. (1999) for analysis of Wnt-16 expression (sequences common to Wnt16a and Wnt16b).
MATERIALS AND METHODS Cloning of Wnt16a. Wnt16a was cloned from placental cDNA (clontech) using nested PCR with primers designed using genomic prediction to span the ORF of Wnt16a. To reduce sequence errors, Pfu turbo polymerase (stratagene) and low cycle PCR were carried out. Outer primer sequences were Wnt16outerF (5⬘) ACCCAGTGCTCTTTCCACTG and Wnt16outerR (3⬘) CCCCAAATCATCAAAAGGTG. Nested primer sequences were Wnt16nF(5⬘) AGCCTGCAAAAACCACAGAG and Wnt16nR (3⬘) GGGATTCCACTGCAAGAGTC. 3⬘ A overhangs were added to this nested PCR product and the product was then subcloned into PCRIITOPO vector (Invitrogen, K4600-01). Sequencing was carried out on both strands of three subclones using ABI 310 and BigDye terminator sequence kit (PE Applied Biosystems). Final sequence was assembled using Lasergene package. cDNA expression analysis. PCR analysis of normal human tissues was carried out using Multiple Tissue cDNA panels (Clontech #K1420-1, K1421-1) following standard Clontech protocols. Primers used for Wnt-16a were W16aF–CAGAAAGATGGAAAGGCACC and W16aR–ATCATGCAGTTCCATCTCTC (276 bp product). Primers used for Wnt-16b were W16bF–TGCTCGTGCTGTTCCCCTAC and W16bR–ATCATGCAGTTCCATCTCTC (226 bp product). PCR conditions were 3 min at 94°C, 30 (40 s 94°C, 30 s 62°C, 1 min 72°C), 5 min 72°C.
RESULTS Identification and in Silico Gene Structure Prediction of Wnt16 The Wnt16 gene was initially identified in unfinished human genomic sequence on a BAC from chromosome 7q31.31 (AC006364) by a TBLASTN search (6) using the human Wnt7a amino acid sequence as a search probe. Further refinement of homology was achieved using the GeneWise tool (7). This homology based approach identified three exons of the Wnt16 gene, that appeared to comprise all but the first exon of Wnt16, which consisted almost entirely of the putative signal peptide region and showed no homology to other
WNT proteins. To identify the first exon of Wnt16 a combination of gene prediction, splice site and promoter prediction algorithms were performed, including GENESCAN (8), Grail (9), NNPP and Splice—two neural network methods for promoter and splice site prediction (10). These methods identified two potential first exons for the Wnt16 gene, exon 1a and exon 1b, each with a putative promoter region (Fig. 1). GENESCAN predicted exon-1b (Wnt16b) with high confidence, separated by a 179 bp intron from exon 2. A combination of splice site and promoter prediction predicted Exon-1a (Wnt16a) separated from exon 2 by a 4082 bp intron. Predicted promoter regions were also analysed for known transcription factor binding sites, using TFSearch and the TRANSFAC database (11). Transcription factor binding site analysis of both predicted promoter regions revealed, a number of developmentally significant potential transcription factor binding sites (Fig. 2). In particular, three TCF binding sites were identified upstream of exon 1a, another was identified upstream of exon 1b. The predicted properties of the putative signal peptides encoded by exon 1a and 1b differ significantly. The Wnt16b isoform has a hydrophobic signal sequence with a good predicted cleavage site (12) between residues 29 –30 (AQG-NM). The signal sequence in the Wnt16a isoform is less hydrophobic, and using the same prediction method (12) it does not have a good predicted cleavage site, however this is within the parameters of other known WNT signal peptides, e.g., Wnt7b, which also contains a weakly hydrophobic signal sequence without a good predicted cleavage site. Alignment of both Wnt16 isoforms, with the most closely related WNT proteins shows some degree of sequence conservation across signal peptide regions (Fig. 3). Both Wnt16 isoforms, share highest homology
815
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2. Human Wnt-16 putative promoter regions and alternative first exons. Sequence of the human Wnt16a putative promoter region (A) and the human Wnt16b putative promoter region 2 (B). Exons are indicated in upper case, introns are in lowercase. Predicted RNA Polymerase II promoter sites are indicated with dashed underlining (based on NNPP prediction). Putative transcription factor binding sites identified using TFSearch and the TRANSFAC database (Heinemeyer et al., 1998), are represented in bold with underline, the name of the site is indicated underneath along with the percentage match to the binding site identity matrix in brackets.
with mouse Wnt7b (48% identity). A phylogeny of Wnt16 in relation to other known vertebrate Wnts is presented in Fig. 4. Based on this phylogeny, Wnt16 does not appear to have a close paralog within the known vertebrate WNT family, indeed it appears to be one of the most divergent members of the WNT family, yet it has retained all of the most strongly conserved signatures of the family, including the 24 strongly conserved cysteine residues. Isolation of Putative Wnt16 Isoforms and Sequence Analysis Considering the strong supporting evidence for the alternative Wnt16 exon 1 predictions, we designed
primers to confirm the existence of these potential isoforms based on alternative organisation with exon 1a or exon 1b. Using RT-PCR we were able to identify both isoforms in cDNA libraries (data not shown). We cloned Wnt16a from placental cDNA, sequence analysis revealed an ORF of 1086 base pairs, coding for a 361 amino acid protein. This ORF concurred closely with the predicted Wnt16a gene from the BAC sequence, except for a 4 bp ACCC insertion, 11 bp upstream of the ATG start codon. The sequence was deposited in the Genbank database, under Accession No. AF152584. Subsequently to this Mcwhirter et al. (5), cloned and characterised a Wnt16 gene, 100% homologous to Wnt16b (AF169963).
816
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 2—Continued
Expression Analysis of Wnt16 Isoforms in Adult Human Tissues
DISCUSSION
Expression analysis of Wnt16 in normal adult tissues was carried out by PCR analysis of commercially prepared cDNA libraries using isoform specific primers. To ensure signal was obtained from mRNA only, primers were designed to span introns of the genomic sequence (Fig. 1b). In normal human tissues, Wnt16a is expressed at significant levels only in the pancreas, whereas Wnt16b is expressed more ubiquitously with highest levels in adult kidney, placenta, brain, heart, and spleen (Fig. 5).
Using a combination of in silico gene prediction and promoter characterisation techniques we have identified and confirmed in vivo, a novel Wnt16 isoform formed from alternative first exon use. Wnt16a and Wnt16b, differ in the composition of their 5-UTR, first exons and predicted signal peptide properties. This highlights an issue in dealing with gene predictions from genomic sequence. Exon prediction tools alone (e.g., GRAIL or GenScan) predict only the Wnt16b isoform. However intense analysis of a 6 kb region upstream of Wnt16 exon-2, revealed other potential
817
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. Amino acid alignment of human WNT-16a and b, and closer related members of the WNT family: mouse WNT-4, human WNT-7a, mouse WNT-7b, and mouse WNT-1. Asterisks indicate the 24 strongly conserved cysteine residues, which define all members of the WNT family.
promoters and WNT signalling associated transcription factor binding sites. When all this data was combined and reviewed with a biological rationale tailored for the specific target pathway, it was possible to detect potential gene structures which eluded conventional gene prediction. This is not the first description of a WNT gene with alternative isoforms, very recently Katoh et al. (13)
FIG. 4. Rooted phylogenetic tree illustrating the putative evolutionary relationship between Wnt16 and other known mammalian Wnts. Human WNT proteins are prefixed with an h, mouse WNT proteins are prefixed with an m.
described “alternative splicing” of first exons in the Wnt2B gene to form alternative isoforms with distinct expression profiles. The authors did not explore the mechanism underlying the generation of these isoforms. We propose that Wnt16 isoforms are not generated by alternative splicing but rather, they are the product of separate promoters and hence Wnt16a and Wnt16b might be considered to be separate transcriptional and regulatory units rather than splice variants from a single promoter. There are several precedents in the literature to support this, Van den Wijngaard et al. (14) described two promoters involved in the transcriptional regulation of the human BMP-4 gene, one upstream of exon 1, the second located in intron 1, upstream of exon 2, giving rise to different transcripts in a cell type and differentiation-dependent manner. In another example, Ohnishi et al. (15) isolated a type 2C protein phosphatase (PP2Cbeta) with different 5⬘ termini. Subsequent analysis of the 5⬘ flanking regions of exon 1 and 2 by reporter gene showed that these regions acted as distinct promoters. Analysis of PP2Cbeta transcripts by RT-PCR showed that exon-1 and exon-2 derived transcripts were differentially expressed, suggesting that alternative promoter usage regulated tissue-specific expression of PP2Cbeta (15). We suggest that in a similar manner to PP2Cbeta and BMP-4, Wnt16 isoforms may also be regulated by separate promoters, with different regulatory cues. Some indication of these potential cues emerge after analysis of consensus transcription factor binding sites
818
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 5. Tissue expression pattern of Wnt-16b and Wnt-16a mRNA in normal human tissues. Commercially available CLONTECH cDNAs prepared from various normal human tissues were used as templates for PCR. (A) Pattern of Wnt16b mRNA: 30 cycles of PCR with primers W16bF and W16bR were used, product size is 225 bp. (B) Pattern of Wnt16a mRNA: 30 cycles of PCR with primers Wnt16aF and Wnt16aR were used, product size is 275 bp. (C) Housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) control: 22 cycles of PCR were used, predicted size is 452 bp. Lanes 2–16 contain adult human tissue cDNAs (tissue indicated above lane). 100 bp ladder (Roche Molecular Biochemicals) DNA size markers are shown in lane 18.
in the 5⬘ regions of Wnt16 exon 1a and exon 1b. The 5⬘ region preceding exon 1b contains two consensus binding sites for members of the T-cell factor (Tcf) family (Fig. 2). Upon interaction with the WNT signalling transducer molecule beta-catenin, Tcf factors become potent transactivators of many developmental processes (16). Also potentially significant, is an AP-1 site. Treatment with lithium, a GSK-3 inhibitor, leads to activation of an AP-1-luciferase reporter in Xenopus embryos (17). The 5⬘ region preceding exon 1a is rich in potential transcription factor binding sites with a known role in WNT signalling. There are four consensus binding sites for members of the Tcf family. There is also an E2F1 binding site, which functions to regulate apoptosis and to suppress cell proliferation and is also found in the WNT-responsive cyclin E and cyclin D1 promoters (18). Perhaps most significantly the Wnt16a promoter contains three consensus binding sites for the oncogenic homeodomain transcription factor, E2A-Pbx1. Mcwhirter et al. (1999) recently demonstrated upregulation of Wnt16 by E2A-Pbx1 in acute lymphoblastoid leukaemia (5). Interestingly, the Wnt16 gene in this publication corresponded to Wnt16b, they did not describe the Wnt16a isoform, hence Leukaemic cell expression analysis was performed using RT-PCR primers common to both Wnt16a and Wnt16b. Hence the results would reflect the expression of both Wnt16 isoforms (Fig. 1b). Considering the tandem E2A-Pbx1 response elements in the putative Wnt16a promoter, it is tempting to speculate that Wnt16a rather than Wnt16b may be specifically up-
regulated in lymphoblastoid leukaemia. This might suggest a wider role for Wnt16a and other WNT isoforms (e.g., Wnt2b isoforms) in oncogenesis in diverse tissues. Wnt16a and Wnt16b both display very distinct and specific expression patterns in panels of normal adult tissue. Of the 16 adult tissues screened for the alternative isoforms, only the pancreas displayed significant levels of Wnt16a. In contrast, Wnt16b was more widely expressed, showing highest levels in the kidney, placenta, brain, heart, and spleen. These differential expression patterns of the Wnt16 isoforms, further support the hypothesis that alternative promoters may be driving tissue specific expression of one or the other isoform of Wnt16. The generation of functionally distinct isoforms from a single WNT gene adds further complexity to the WNT signalling paradigm. It may be that alternative promoters allow the expression of the WNT gene to be controlled by a wider range of independent pathways, so the response can be initiated by different regulatory cues. Alternatively, because the actual gene product is altered, this may alter the properties of the WNT ligand itself, perhaps by altering secretion efficiency or by altering subcellular targeting (e.g., nuclear targeting (19)). This may indicate that each WNT isoform has a different effector pathway. Coupled with the recent evidence for alternative splicing of Wnt2B (13), this evidence for isoforms and alternative promoter usage in Wnt16 suggests the complexity of regulation of WNT
819
Vol. 278, No. 3, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
signalling may be harder to unravel than previously anticipated. ACKNOWLEDGMENTS Many thanks go to Kathy Ellington and Tania Testa for technical assistance and advice. This research was supported by an MRC Collaborative studentship.
REFERENCES 1. Wodarz, A., and Nusse, R. (1999) Annu. Rev. Cell. Dev. Biol. 14, 59 – 88. 2. Nusse, R., and Varmus, H. E. (1982) Cell 31, 99 –109. 3. Cavallo, R., Rubenstein, D., and Peifer, M. (1997) Curr. Opin. Genet. Dev. 7, 459 – 466. 4. Gazit, A., Yaniv, A., Bafico, A., Pramila, T., Igarashi, M., Kitajewski, J., and Aaronson, S. A. (1999) Oncogene 18, 5959 –5966. 5. McWhirter, J. R., Neuteboom, S. T., Wancewicz, E. V., Monia, B. P., Downing, J. R., and Murre, C. (1999) Proc. Natl. Acad. Sci. USA 96, 11464 –11469. 6. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D. J. (1997) Nucleic Acids Res. 25, 3389 –3402. 7. Birney, E., Thompson, J. D., and Gibson, T. J. (1996) Nucleic Acids Res. 24, 2730 –2739. 8. Burge, C. B., and Karlin, S. (1998) Curr. Opin. Struct. Biol. 8, 346 –354.
9. Lopez, R., Larsen, F., and Prydz, H. (1994) Genomics 24, 133–136. 10. Reese, M. G., Harris, N. L., and Eeckman, F. H. (1996) Biocomputing: Proceedings of the 1996 Pacific Symposium (Hunter, L., and Klein, T. E., Eds.), World Scientific Publishing Co., Singapore. 11. Heinemeyer, T., Wingender, E., Reuter, I., Hermjakob, H., Kel, A. E., Kel, O. V., Ignatieva, E. V., Ananko, E. A., Podkolodnaya, O. A., Kolpakov, F. A., Podkolodny, N. L., and Kolchanov, N. A. (1998) Nucleic Acids Res. 26, 364 –370. 12. Nielsen, H., Engelbrecht, J., Brunak, S., and von Heijne, G. (1997) Int. J. Neural. Syst. 8, 581–599. 13. Katoh, M., Kirikoshi, H., Saitoh, T., Sagara, N., and Koike, J. (2000) Biochem. Biophys. Res. Commun. 275, 209 –216. 14. Van den Wijngaard, A., Pijpers, M. A., Joosten, P. H., Roelofs, J. M., Van zoelen, E. J., and Olijve, W. (1999) J. Bone. Min. Res. 14, 1432–1441. 15. Ohnishi, M., Chida, N., Kobayashi, T., Wang, H., Ikeda, S., Hanada, M., Yanagawa, Y., Katsura, K., Hiraga, A., and Tamura, S. (1999) Eur. J. Biochem. 263, 736 –745. 16. Roose, J., and Clevers, H. (1999) Biochim. Biophys. Acta. 1424, M23–M37. 17. Hedgepeth, C. M., Conrad, L. J., Zhang, J., Huang, H. C., Lee, V., and Klein, P. S. (1997) Dev. Biol. 185, 82–91. 18. Inoshita, S., Terada, Y., Nakashima, O., Kuwahara, M., Sasaki, S., and Marumo, F. (1999) Kidney Int. 56, 2085–2095. 19. Baird, J. W., Nibbs, R. J., Komai-Koma, M., Connolly, J. A., Ottersbach, K., Clark-Lewis, I., Liew, F. Y., and Graham, G. J. (1999) J. Biol. Chem. 274, 33496 –33503.
820