Oncogene (1998) 16, 2403 ± 2406  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc SHORT REPORT A highly conserved processed PTEN pseudogene is located on chromosome band 9p21

Patricia LM Dahia1,2, Michael G FitzGerald2,3, Xue Zhang2,3, Debbie J Marsh1, Zimu Zheng1, Torsten Pietsch4, Andreas von Deimling4, Frank G Haluska2,3, Daniel A Haber2,3 and Charis Eng1,2,5

1Translational Research Laboratory, Department of Adult Oncology, Charles A. Dana Human Cancer Genetics Unit, and Program in Population Sciences, Dana-Farber Cancer Institute, SM822, Boston, Massachusetts 02115, USA; 2Department of Medicine, Harvard Medical School, Boston, Massachusetts 02115, USA; 3Massachusetts General Hospital Cancer Center, Boston and Charlestown, Massachusetts 02114, USA; 4Department of Neuropathology, University of Bonn Medical Center, D-53105 Bonn, Germany; 5Cancer Research Campaign Human Cancer Genetics Research Group, University of Cambridge, Cambridge CB2 2QQ, UK

PTEN/MMAC1/TEP1, encoding a dual-speci®city of human malignancies (Li and Sun, 1997; Li et al., phosphatase, is a tumor suppressor gene which has 1997; Steck et al., 1997). PTEN encodes a 403-amino recently been cloned and mapped to chromosome acid dual speci®city phosphatase, and most likely, 10q23.3. We have shown that germline mutations of functions as a tumor suppressor gene. Because of its PTEN are present in individuals with two hamartoma apparent role in a wide variety of tumors, as syndromes: Cowden Syndrome, associated with a preliminarily indicated by a high frequency of predisposition to breast and thyroid cancers, and mutations and/or deletions in tumor cell lines, PTEN Bannayan-Zonana syndrome. Somatic mutations of mutation analysis based on both genomic DNA and PTEN have been reported in a variety of human cancer cDNA templates in several human genetic disorders as cell lines, suggesting a potential role for this gene in the well as in primary cancers is in progress. pathogenesis of human malignancies. We report the We identi®ed the PTEN pseudogene while char- identi®cation of a highly conserved PTEN processed acterizing PTEN transcription levels in normal and pseudogene, cPTEN, which shares over 98% homology a€ected tissue from individuals with CS. Primers with the coding region of functional PTEN, and its spanning the 5' and 3' untranslated regions (UTRs) localisation to chromosome 9p21. The high sequence of PTEN (5'-CATCTCTCTCCTCCTTTTTCTTCA- homology of cPTEN with the PTEN transcript may 3'and R 5'-TTTTCATGGTGTTTTATCCCTCTT-3') potentially lead to misinterpretation when performing were used for PCR with normal control genomic mutation analyses based on cDNA templates. Caution DNA (Promega, Madison, WI) as template. Condi- should be exerted when using such screening approaches. tions were as described previously (Li et al., 1997). A 1287 bp fragment, similar in size to a product obtained Keywords: MMAC1; TEP1; 9p; Cowden; pseudogene with the same primers using cDNA template, was consistently ampli®ed with genomic DNA. Single- and double-stranded sequence analyses were performed as PTEN, also known as MMAC1/TEP1, has recently previously described (Dahia et al., 1997). Despite using been cloned and mapped to chromosome sub-band a genomic DNA template and primers which would 10q23.3, a region which had been previously shown to have spanned several introns, the PCR product size be linked to Cowden Syndrome (CS) (Li and Sun, obtained was that predicted for a cDNA template. 1997; Li et al., 1997; Myers et al., 1997; Nelen et al., Sequence analysis of this product revealed 98.6% 1996; Steck et al., 1997). CS is an autosomal dominant homology to functional PTEN cDNA, with only 19 disorder associated with hamartomas of the skin, nucleotides varying from the functional PTEN cDNA breast, thyroid, gastointestinal tract and other organs, sequence in a 1287 bp fragment spanning the whole as well as a predisposition to breast and thyroid coding region of the gene plus an additional 80 bp of 5' cancers (Eng, 1997; Longy and Lacombe, 1996). We and 3' UTRs (Figure 1). The lack of introns or intron- have recently shown that germline mutations of PTEN like regions in this new sequence leads us to the are present in individuals with CS (Liaw et al., 1997), conclusion that this is a processed pseudogene (Vanin, and also in patients with another hamartomatous 1985). Eighteen of the 19 nucleotide di€erences were disorder, Bannayan-Zonana syndrome (Marsh et al., found within the PTEN coding region. Most of the 1997), which shares some clinical features with CS, but di€erences were nucleotide transitions (79%), while without known predisposition to cancer. In addition, only four transversion substitutions were observed somatic mutations of PTEN have been reported in (Figure 1). This pattern of mutations found in cPTEN glioblastoma, breast, prostate and kidney cancer cell is consistent with that found in other pseudogenes, lines, suggesting it may play a role in the pathogenesis nucleotide substitutions C?T being the most com- monly observed (Gojobori et al., 1982). Although CpG dinucleotides are usually targets for mutation, only a minority (3/18) of the di€erences found in cPTEN Correspondence: C Eng Received 28 October 1997; revised 1 December 1997; accepted 1 were present in such sequence context. The most December 1997 noteworthy di€erence observed between the two PTEN pseudogene on 9p21 PLM Dahia et al 2404 sequences was at the start codon of functional PTEN scribed sequences in the genome is more randomly (ATG?AGG), which would preclude the use of the distributed. Hence, the processed pseudogene is usually original PTEN open reading frame (ORF). However, a not syntenic with the functional gene (Vanin, 1985). putative ORF with 369 predicted amino acids was The time of occurrence of the retrotransposition of a found in cPTEN, which retains high homology with processed pseudogene may be tentatively estimated functional PTEN, including full identity with the based on the di€erences found in comparison with the protein tyrosine phosphatase (PTPase) motif (Figure coding sequence of the functional gene. If the estimated 2). The di€erences observed between cPTEN and rate of *261079 per nucleotide per year is taken as a human functional PTEN were also present when the model ®gure for spontaneous mutation in primates human pseudogene sequence was compared to the (Vanin, 1985), and if we assume that all 18 coding canine and murine functional pten (Figure 2), which nucleotide di€erences were accumulated by the suggests that the substitutions occurred in nucleotides pseudogene while the functional gene remained un- that are highly conserved throughout these species. changed since their divergence, then this processed It has been suggested that a large fraction of pseudogene could have originated approximately 7 processed pseudogenes present in the mammalian million years ago. If a higher mutation rate, 4.761079 genome have been generated by retrotransposition per nucleotide per year (Li et al., 1987), estimated for (Maestre et al., 1995; Vanin, 1985). This would be mammals, is employed instead, the origin of the mediated by transposon-encoded reverse transcriptases, pseudogene would have been approximately 3 million followed by integration of the sequence into the years ago. genome. Unlike other mechanisms of non-processed A somatic-cell hybrid panel generated using mouse pseudogene formation, such as genomic duplication, and hamster backgrounds (Coriell Camden, NJ) was which results in the generation of pseudogenes in screened by PCR using PTEN primers (PTEN-AF 5'- nearby sequences, integration of these reverse-tran- AGTCCAGAGCCATTTCCATC-3' and PTEN-AR 5'- CATGGTGTTTTATCCCTCTTG-3'). Appropriate positive (human genomic DNA) and negative (mouse and hamster DNA) controls were included. cPTEN was found to map to chromosome 9. Further, a monochromosome 9 panel and a chromosome 9 radiation-hybrid panel (Research Genetics, Huntsville, AL) were PCR-screened to con®rm cOTEN chromo- some mapping and better de®ne its location. The linkage likelihood data was obtained by testing the results against a Radiation Hybrid Database (SHGC, Stanford Human Genome Center). cPTEN was

Figure 1 Comparison of nucleotide sequence between homologous regions of the PTEN cDNA and PTEN processed pseudogene (cPTEN). PTEN cDNA start and stop codons are underlined PTEN pseudogene on 9p21 PLM Dahia et al 2405 localized to a single interval (with a likelihood of a genomic DNA-based template, and a *200-bp band 41000 : 1) spanning the D9S1878 marker, which lies with cDNA templates. cPTEN primers failed to *20 cM proximal to the p16/CDKN2 tumor suppres- amplify the pseudogene product, while a co-ampli®ed sor gene on chromosome 9p21. Interestingly, this house-keeping transcript (GUSB) showed a product of region, similar to the 10q23 region, where functional the expected size (Figure 3). These results suggest that PTEN maps, has long been known to be frequently the pseudogene is not transcribed, at least in this deleted in human cancers (Diaz et al., 1988; Einhorn glioblastoma cell line. A normal thyroid cDNA, used and Heyman, 1993; Nobori et al., 1994). It is possible as a control, ampli®ed both PTEN and GUSB that the 9p21 region possesses intrinsic genetic transcripts, while a genomic sample yielded genomic instability, allowing for more frequent alterations of cPTEN and GUSB products. No product compatible genetic material to occur. In fact, a region of with genomic ampli®cation was detected using the duplication from chromosome 7 to chromosome 9 A172 template, which validated the ®nding of lack of containing several pseudogenes and genomic repeats, cPTEN transcription in the model used. This is in which appears to include cPTEN, has recently been accordance with what is usually found in other human noted (GenBank accession number AF029308). To determine whether cPTEN was transcribed, we designed a RT ± PCR assay using a glioblastoma cell line known to be homozygously deleted for functional PTEN, A172 (Li et al., 1997; Steck et al., 1997). A normal thyroid cDNA and a genomic template were used as controls. RNA was extracted using TriZol marker A172 thyroid genomic blank (Gibco/BRL, Paisley, UK), according to the manufac- turers' instructions. Two micrograms of total RNA were reverse transcribed into cDNA using random hexamers and Superscript RNase H7 Reverse Tran- scriptase (Gibco/BRL, Paisley, UK) according to the manufacturers' guidelines. PCR was performed as described above with primers spanning PTEN exons 1 to 5 (F5'-CATCTCTCTCCTCCTTTTTCTTCA-3' and R 5'-ATATCATTACACCAGTTCGTCCCT-3'), which would amplify both PTEN and cPTEN, along Figure 3 An ethidium-bromide stained agarose gel of duplex with primers for the house-keeping gene beta-glucur- RT ± PCR products in which PTEN primers and primers for the GUSB house-keeping gene were used at an 8 : 1 molar ratio. A onidase (GUSB), as published (Ivanchuk et al., 1997), normal thyroid cDNA sample ampli®ed both PTEN and GUSB at an 8 : 1 molar ratio. A product obtained with exonic transcripts, while A172, a PTEN deleted cell line, only ampli®ed PTEN primers and A172 cDNA, therefore, would the house-keeping gene band (*200 ± bp fragment). In the represent the pseudogene. Further, to control for absence of PTEN in this template, a product of equivalent size potential genomic DNA contamintation of the cDNA (450 ± bp) would represent cPTEN transcript. A genomic DNA template ampli®ed both cPTEN and a 400 ± bp fragment sample, cPTEN product was coampli®ed with house- corresponding to a genomic GUSB-related product. 100-bp keeping primers which generate a 400-bp product with ladder (Gibco, Gaithersburg, MD) was used as a size marker

Figure 2 Multiple sequence aligment (MSA program, PIMA, Baylor College of Medicine) comparing the protein sequences of human, dog and mouse pten with the cPTEN putative 369 residues ORF. The PTPase motif is underlined and the amino acids mismatched are shown in bold PTEN pseudogene on 9p21 PLM Dahia et al 2406 processed pseudogenes: due to the random nature of misleading results in mutation studies. Hence, the the transposition, processed pseudogenes are rarely in possibility of the presence of cPTEN sequence has to the appropriate context for transcription to occur be considered and appropriately excluded if a sequence (Vanin, 1985). variant is noted. Given our data, investigators wishing The identi®cation of a PTEN pseudogene is highly to use cDNA as template may now design primers in relevant for mutation detection, as PTEN is an the regions relatively unique between the two important gene which might be involved in a wide sequences. array of inherited conditions as well as in several human cancers. The processed pseudogene, with its high homology to the functional PTEN transcript, presents the potential for misinterpretation in studies Acknowledgements using cDNA as template for mutation detection in We thank Ricardo Aguiar for helpful discussions. This study was supported by the American Cancer Society human disease. Due to the lack of intronic sequence, (RPG-97-064-01VM), the Markey Charitable Trust, the any study involving the use of exonic primers needs to Barr Investigatorship and the Lawrence and Susan Marx be performed cautiously as many RNA preparations Investigatorship in Human Cancer Genetics (to CE); the are not without some genomic DNA contamination. Susan G Komen Breast Cancer Foundation (to PLMD and Also, the exclusive use of exonic primers in genomic- CE); and the Deutsche Forschungsgemeinschaft -SFB400 based template is potentially prone to causing (to TP). Genbank Accession No. AF040103.

References

Dahia PL, Marsh DJ, Zheng Z, Zedenius J, Komminoth P, Longy M and Lacombe D. (1996). Ann. GeÂnet., 39, 35 ± 42. Frisk T, Wallin G, Parsons R, Longy M, Larsson C and Maestre J, Tchenio T, Dhellin O and Heidmann T. (1995). Eng C. (1997). Cancer Res., 57, 4710 ± 4713. EMBO J., 14, 6333 ± 6338. Diaz MO, Zieman S, Le Beau MM, Pitha P, Smith SD, Marsh DJ, Dahia PL, Zheng Z, Liaw D, Parsons R, Gorlin Chilcote RR and Rowley JD. (1988). Proc. Natl. Acad. RJ and Eng C. (1997). Nat. Genet., 16, 333 ± 334. Sci. USA, 85, 5259 ± 5263. Myers MP, Stolarov J, Eng C, Li J, Wang SI, Wigler MH, Einhorn S and Heyman M. (1993). Leuk. Lymphoma, 11, Parsons R and Tonks NK. (1997). Proc. Natl. Acad. Sci. 191 ± 196. USA, 94, 9052 ± 9057. Eng C. (1997). J. Genet. Counsel., 6, 181 ± 191. Nelen MR, Padberg GW, Peeters EAJ, Lin AY, van den Gojobori T, Li W-H and Graur D. (1982). J. Mol. Evol., 18, Helm B, Frants RR, Coulon V, Goldstein AM, van Reen 36 ± 369. MMM, Easton DF, Eeles RA, Hodgson S, Mulvihill JJ, Ivanchuk SM, Eng C, Cavenee WK and Mulligan LM. Murday VA, Tucker MA, Mariman ECM, Starink TM, (1997). Oncogene, 14, 1811 ± 1818. Ponder BAJ, Ropers HH, Kremer H, Longy M and Eng C. Li D-M and Sun H. (1997). Cancer Res., 57, 2124 ± 2129. (1996). Nature Genet., 13, 114 ± 116. LiJ,YenC,LiawD,PodsypaninaK,BoseS,WangS,PucJ, Nobori T, Miura K, Wu DJ, Lois A, Takabayashi K and Miliaresis C, Rodgers L, McCombie R, Bigner SH, Carson DA. (1994). Nature, 368, 753 ± 756. Giovanella BC, Ittman M, Tycko B, Hibshoosh H, Steck PA, Pershouse MA, Jasser SA, Yung WKA, Lin H, Wigler MH and Parsons R. (1997). Science, 275, 1943 ± Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis 1947. T, Frye C, Hu R, Swedlund B, Teng DHF and Tavtigian Li W-H, Tanimura M and Sahrp PM. (1987). J. Mol. Evol., SV. (1997). Nature Genet., 15, 356 ± 362. 25, 333 ± 342. Vanin EF. (1985). Annu. Rev. Genet., 19, 253 ± 272. Liaw D, Marsh DJ, Li J, Dahia PLM, Wang SI, Zheng Z, Bose S, Call KM, Tsou HC, Peacocke M, Eng C and Parsons R. (1997). Nature Genet., 16, 64 ± 67.