Mutations in the TSC2 gene: analysis of the complete coding sequence using the protein truncation test (PTT)
Mutations in the TSC2 gene: analysis of the complete coding sequence using the protein truncation test (PTT)Inge van Bakel1, Tiina Sepp1, Susannah Ward1, John R.W. Yates1,2 and Andrew J. Green1,2,3,*
Departments of 1Pathology and 3Medical Genetics, University of Cambridge, Cambridge, UK and 2Department of Medical Genetics, P.O. Box 134, Hills Road, Addenbrooke's NHS Trust, Cambridge CB2 2QQ,UK
Received February 6, 1997;Revised and Accepted July 3, 1997
Mutations in the TSC2 gene on chromosome 16p13.3 are responsible for ~50% of familial tuberous sclerosis (TSC). The gene has 41 small exons spanning 45 kb of genomic DNA and encoding a 5.5 kb mRNA. Large germline deletions of TSC2 occur in <5% of cases, and a number of small intragenic mutations have been described. We analysed mRNA from 18 unrelated cases of TSC for TSC2 mutations using the protein truncation test (PTT). Three cases were predicted to be TSC2 mutations on the basis of linkage analysis or because a hamartoma from the patient showed loss of heterozygosity for 16p13.3 markers. Three overlapping PCR products, covering the complete coding sequence of mRNA, were generated from lymphoblastoid cell lines, translated into 35S-methionine labelled protein, and analysed by SDS-PAGE. PCR products showing PTT shifts were directly sequenced, and mutations confirmed by restriction enzyme digestion where possible. Six PTT shifts were identified. Five of these were caused by mutations predicted to produce a truncated protein: (i) a sporadic case showed a 32 bp deletion in exon 11, and a mutant mRNA without exon 11 was produced; the normal exon 10 was also spliced out; (ii) a sporadic case had a 1 bp deletion in exon 12 (1634delT); (iii) a TSC2-linked mother and daughter pair had a G -> T transversion in exon 23 (G2715T) introducing a cryptic splice site causing a 29 bp truncation of mRNA from exon 23; (iv) a sporadic case showed a 2 bp deletion in exon 36; (v) a sporadic case showed a 1 bp insertion disrupting the donor splice site of exon 37 (5007+2insA), resulting in the use of an upstream exonic cryptic splice site to cause a 29 bp truncation of mRNA from exon 37. In one case, the PTT shift was explained by in-frame splicing out of exon 10, in the presence of a normal exon 10 genomic sequence. Alternative splicing of exon 10 of the TSC2 gene may be a normal variant. Three 3rd base substitution polymorphisms were also detected during direct sequencing of PCR products. Confirmed mutations were identified in 28% of the families studied and on the assumption that half of the sporadic cases should have TSC2 mutations, a crude estimate of the detection rate would be 60%. This compares favourably with other screening methods used for TSC2, notably SSCP, and since PTT involves much less work it may be the method of choice.
Tuberous sclerosis (TSC) is an autosomal dominant condition characterised by tumour-like malformations (hamartomas) of the skin, brain, heart, kidney and other organs. The disease prevalence is estimated at 1 in 10 000, with two thirds of cases being sporadic and representing new dominant mutations (1 ).
Two loci for TSC have been identified by linkage studies: TSC1 located on chromosome 9q34 and TSC2 on 16p13.3 (2 ). Each accounts for ~50% of familial TSC cases. The TSC2 gene has been characterised, and intragenic deletions of 1-5 kb have been identified in this gene in a small number of cases of TSC, as well as a number of smaller mutations (3 -7 ). The gene product has been named tuberin, and is encoded by an ubiquitously expressed 5.5 kb mRNA. The gene spans ~45 kb of genomic DNA, and is made up of 41 exons. The exons vary in size from 70 to 213 bp, with one larger exon of 488 bp (8 ). A region of 58 amino acids towards the C-terminal of the predicted TSC2 protein shows similarity to the GTPase-activating protein (GAP) GAP3 (3 ). The TSC2 gene is immediately adjacent to the polycystic kidney disease PKD1 gene on chromosome 16p13.3, and contiguous gene deletions have been described in a small number of patients with TSC and polycystic kidney disease (9 ). The TSC1 gene remains uncloned, and lies in a region of 2.7 Mb between the anonymous markers D9S149 and D9S114 (10 ). The presence of large germline deletions of the TSC2 gene in cases of TSC, and the finding of loss of heterozygosity for markers in the region of the TSC2 gene in TSC hamartomas, would indicate that the TSC2 gene acts as a tumour suppressor (11 ). Eukaryotically expressed TSC2 protein has rap1-GAP activity, supporting the tumour suppressor hypothesis (12 ).
(i) skipping of exon 10 & 11 (ii) skipping of exon 11
(i) shortened by 94 aa (ii) shortened by 46 aa
3
15
1 bp
1634delT
frameshift
truncated by 1242 aa
7
23
G2715T
cryptic splice site created exon 23 truncated by 29 bp
truncated by 881 aa
6
36
2 bp
del4618AG
frameshift
truncated by 252 aa
1
37
29 bp
5007+2insA
normal splice site disrupted cryptic splice site in normal sequence used exon 37 truncated by 29 bp
truncated by 159 aa
Only a small proportion of cases of TSC show large deletions of the TSC2 gene detectable by Southern blot analysis. We therefore proceeded to look for small intragenic mutations in the gene. The large number of small exons makes the detection of genomic TSC2 mutations difficult. As the gene is likely to act as a tumour suppressor, we hypothesised that mutations giving rise to an inactive protein of altered size would be a common finding.
The protein truncation test (PTT) is designed to detect sequence changes that cause premature protein termination (13 ). A DNA template is generated by PCR, where the forward PCR primer incorporates a T7 polymerase binding site and a new ATG initiation codon. The DNA is converted into protein by a coupled transcription/translation system, and the resultant protein is analysed by SDS-PAGE. We therefore chose the protein truncation test to look for mutations in TSC2 mRNA from lymphoblastoid cell lines from cases of TSC.
Eighteen unrelated TSC patients were studied by PTT and six showed altered size protein products (see Table 1 ).
Patient 4, a sporadic case, showed two altered size protein products from fragment 1, as well as the normal sized protein. The RT-PCR fragments were sequenced, and showed an in frame deletion of nucleotides 1120-1257, representing the 46 amino acids encoded by exon 11. In addition, there was an in frame deletion of nucleotides 976-1257 representing the 94 amino acids encoded by exons 10 and 11. This case therefore showed three transcripts, one normal, one with exon 11 spliced out, and one with exons 10 and 11 spliced out. This was confirmed by PCR across exons 10 and 11, which showed two smaller fragments of the expected sizes. Exon specific restriction enzyme analysis confirmed the PCR results. The genomic sequence of exons 10 and 11 and flanking intronic regions was obtained from this case. There was a 32 bp deletion of bases 1160-1192 in the middle of exon 11. The sequence of exon 10, and flanking intron regions was normal. The mutant exon 11 containing an out-of-frame deletion is therefore being spliced out, to maintain expression of the TSC2 protein.
Patient 3, a sporadic case, showed an altered size protein produced from fragment 1 (see Fig. 1 A). Sequencing of TSC2 mRNA showed a 1 bp deletion of base 1634 in exon 15, confirmed on genomic sequencing of exon 15. The resulting frameshift truncated the predicted TSC2 protein by 1263 amino acids, and added a novel sequence of 21 amino acids, ending in a new stop codon. This patient presented with infantile spasms, and has continuing seizures with behavioural disturbance. He has hypopigmented skin macules, a forehead fibrous plaque, a fibroma on his back and calcified subependymal nodules on CT brain scan.
Large deletion mutations in the TSC2 gene comprise a very small percentage of mutations in cases of TSC. Five large deletions and five intragenic deletions were described in the original report of the cloning of the TSC2 gene (3 ). Six patients with TSC and polycystic renal disease have been shown to have large contiguous deletions including both the TSC2 and the adjacent PKD1 gene (9 ). However, the large majority of TSC cases does not show rearrangements of the TSC2 gene on Southern blot analysis. Alternative methods have therefore been used to detect small TSC2 mutations. This work is hampered by the fact that the TSC2 gene has 41 exons, between 70 and 210 bp in size, plus one larger exon of 488 bp.
Several papers have described small TSC2 mutations in individual cases, detected by genomic PCR-SSCP of individual exons. A de novo 1 bp deletion, 5110delA in exon 40, in association with a de novo 4 bp deletion in intron 38 was described in a child with sporadic TSC (4 ). This deletion was predicted to give rise to an elongated protein, with a more distal stop codon. In an Afro-American family linked to TSC2, a 4590/4591 delC mutation in exon 35 was shown to segregate with the disease (5 ). A 52A-T sequence change, causing the amino acid change K12X, was detected in a case of TSC, after screening exon 1 of the TSC2 gene in 116 unrelated patients (6 ). The mutation was predicted to give rise to a severely truncated protein.
A systematic study of the TSC2 gene in 30 patients with TSC has been performed, using PCR-SSCP from TSC2 mRNA expressed in lymphoblasts (7 ). The TSC2 gene was amplified in 34 overlapping fragments of 150-255 bp in size, and analysed for SSCP by MDE gel electrophoresis under one set of conditions. Fourteen cases showed SSCP mobility shifts, and were sequenced. Three showed sequence changes that were predicted to give rise to a truncated protein. One was a 29 bp tandem duplication in exon 34, the second was a 2 bp deletion in exon 10, and the third was a nonsense mutation in exon 14. One sequence change was an in frame deletion of Phe 1486 in exon 33. Five cases showed base substitutions giving rise to an altered amino acid. None of these sequence changes was found in 60 normal chromosomes. In one case the mutation was not found in the normal parents of the affected subject. The authors make the point that the missense mutations may be the cause of TSC in these patients, so that not all inactivating mutations are necessarily truncating. The five remaining SSCP shifts were silent nucleotide changes, and presumed to be TSC2 polymorphisms.
In our systematic study of the TSC2 gene, analysing the TSC2 mRNA by PTT rather than by PCR-SSCP, we found five mutations which produced truncated TSC2 protein in 18 possible TSC mutations, a detection rate of 28%. Using SSCP, Wilson et al. found three truncating mutations in 30 possible TSC mutations, a detection rate of 10% (7 ). We also incidentally found three silent coding sequence changes, which are polymorphisms, since they were found in a proportion of cases sequenced. Given the varied expression of TSC within families, and the small number of mutations found, we were unable to correlate the clinical phenotypes and genotypes in our cases.
The analysis of genomic sequence after detection of mRNA sequence abnormalities demonstrated the value of joint genomic and mRNA sequence analysis. Three different pathogenic abnormalities of mRNA splicing were found. One germline mutation created a new cryptic donor splice site in exon 23, truncating the mRNA from exon 23 by 29 bp. As the mutation was a silent third base substitution, such a genomic sequence change may not have been considered a mutation if detected by genomic DNA analysis alone. Analysis of mRNA by PTT and sequencing enabled us to confirm that this sequence change was pathogenic.
Another germline mutation caused a truncated TSC2 protein by the opposite mechanism. The mutation disrupted a normal splice site in exon 37, thereby using a cryptic splice site in the preceding normal exonic sequence, and the mRNA from exon 37 was truncated by 29 bp.
In one case we identified an alternative splicing event involving exon 10, which explained the altered size proteins seen on PTT. As the genomic sequence of exon 10 was normal in this case, it is not clear whether this was pathogenic or a rare splice variant. No evidence of alternative splicing of exon 10 was found in 15 normal controls. There is a precedent for alternative splicing in the TSC2 gene, which is known to occur for exon 25 (14 ).
One case showed two distinct abnormal protein truncation products. These represented one transcript with skipping of exons 10 and 11, and a second transcript with skipping of exon 11 alone. Sequencing of genomic DNA showed a germline 32 bp deletion in the centre of exon 11, with a normal sequence of exon 10 and normal flanking sequences to both exons. The truncating mutation in exon 11 was skipped to give TSC2 mRNA with an in frame deletion of exon 11. Such mutations have been described before in the dystrophin gene, where a frameshift mutation is rescued by exon skipping (15 ). Exon skipping in the dystrophin gene can recruit adjacent exons. It is not clear whether the additional exon 10 skipping in this case represents such recruitment, or is an alternative splicing variant, given that possible alternative splicing of exon 10 is seen in another case of TSC.
Eukaryotically expressed TSC2 protein has functional rap1-GAP activity, supporting the hypothesis that the TSC2 gene acts as a tumour suppressor (12 ). The first three truncating mutations we detected give rise to a protein that is terminated prior to the putative rap1-GAP region of the TSC2 protein, and we would predict that the protein would not have such activity. The third mutation in exon 37 causes the deletion of the last codon of the rap1-GAP region, and could therefore interfere with the rap1-GAP activity. The germline 32 bp deletion in exon 11 causes exon 11 to be skipped, and would suggest that the amino acids encoded by exon 11 are important for normal function of the TSC2 protein.
There are several reasons why more mutations were not found. Firstly, a proportion of the sporadic cases will have a mutation in the TSC1 gene on 9q34, and we would not expect to detect a TSC2 mutation. As ~50% of familial cases of TSC are linked to each locus, we would only expect to find a TSC2 mutation in about half of the 15 sporadic cases we analysed. On this basis a crude estimate of the detection rate for TSC2 mutations would be 60%. Secondly, PTT would not pick up missense mutations that did not alter the size of the protein. Thirdly, some TSC2 mutations may make the mutant mRNA unstable, so that it would not be detected in lymphoblasts. If the mutant mRNA were absent, then the mutation would in effect be inactivating. Fourthly, due to the nature of the PTT assay, frame shift mutations close to the 3' end of PCR products 1 and 2 would not create a new stop codon within the limits of the PCR product, and go undetected. This limitation could be overcome by analysing PCR products covering 3' regions of fragments 1 and 2 and additional 3' sequence. These problems do not arise in PCR fragment 3 where the PCR product covers 86 nucleotides 3' to the actual stop codon. The detection of larger mutant proteins is possible in these cases.
PTT has the considerable advantage over SSCP of being able to analyse large fragments of DNA, up to 2 kb, rather than the 150-250 bp size range of SSCP. The TSC2 mRNA can thus be generated in three PCR reactions, rather than the large number of small reactions required for SSCP. Also, the likelihood of finding mutations which are pathogenic is higher with PTT, which will not detect the silent base substitutions picked up by SSCP. The size of an altered PTT product will also give an indication of where in the DNA template to search for a mutation.
In summary, we have used PTT to identify five novel mutations in the TSC2 gene. One frameshift mutation was associated with apparent rescue by exon skipping, an unusual phenomenon not previously described in this gene. The detection rate using PTT compared favourably with other mutation screening methods, especially SSCP, and involves much less work. PTT may well be the best choice for the first stage of a mutation screening protocol for routine diagnostic use.
We studied 20 cases of TSC, representing 18 different TSC mutations. All cases were diagnosed according to the Gomez criteria (1 ). Four patients (representing two different mutations) were linked to chromosome 16 by haplotyping and one patient was linked to chromosome 16 by loss of heterozygosity studies. The other 15 patients were sporadic cases. None of the cases showed an altered hybridisation pattern by Southern blotting with the TSC2 cDNA probes.
Lymphocytes were transformed by EBV using standard techniques. DNA was extracted from lymphocytes by standard techniques. After culture to confluence only a fraction of the cells (~5 * 106 cells) was used for total RNA extraction. The rest of the cells were frozen at -70oC as dry pellets for future use.
Total RNA was extracted from lymphoblastoid cell lines using TRIzolTM (Life Technologies), according to the manufacturer's specifications. Cells (5 * 106) were enough to perform one cDNA synthesis using the SuperScriptTM Preamplification System (Life Technologies). cDNA was stored at 4oC. All cDNA was amplified using the provided oligo(dT) primers. Total RNA from non-TSC controls was extracted from whole blood samples. Lymphocytes were isolated using Ficoll-Paquer (Pharmacia) according to standard protocols. These were used for RNA extraction using TRIzolr LS (Pharmacia) according to the manufacturer's specifications. cDNA preparation was performed as described above. In general 10 ml of blood provided enough cells for four separate cDNA preparations.
Three PCR products were generated, designed to overlap by at least 80 nucleotides, spanning the entire coding region of the TSC2 gene. The PCR primers and annealing temperatures are shown in Table 3 . All forward primers were modified by the addition of the following sequence containing a T7 polymerase binding site and an ATG initiation codon: GGATCCTAATACGACTCACTATAGGAACAGACCACC-ATG- immediately 5' to the normal sequence. PCRs were carried out in a total reaction volume of 25 [mu]l. Initial denaturation was at 95oC for 2 min. All PCR cycles had a denaturation step at 95oC for 1 min, an annealing step for 1 min and an extension step at 72oC for 2 min. Cycles were repeated 35 times. Equal amounts of Taq Polymerase (Life Technologies) and Taq Extender (Stratagene) were used for each PCR (2 U). Twice the amount of cDNA template was used for fragments 1 and 2 compared with fragment 3 (2 [mu]l and 1 [mu]l respectively of undiluted cDNA). Sixty ng of each primer was needed. PCR reactions for fragment 2 and 3 were carried out in Taq extender buffer [Stratagene, 20 mM Tris-HCL (pH 8.8), 10 mM KCl, 10 mM (NH4)2SO4, 2 mM MgSO4, 0.1% Triton X-100 and 0.1 mg/ml nuclease-free bovine serum albumin] while those for fragment 1 were carried out in the following buffer: 10 mM Tris-HCl (pH 8), 50 mM KCl, 2.5 mM MgCl2 and 0.01% gelatin. The PCR products could be used directly for the PTT assay, without the need for purification.
The kit used (TnTTM T7-coupled reticulocyte lysate system) was provided by Promega. Reaction volumes were quartered as compared to the manufacturer's protocol. The amount of 35S-methionine used was brought down to one tenth the original amount and 50-100 ng of RT-PCR product was used. The reaction was incubated at 30oC for 2 h and translation products were separated on a 12% SDS-PAGE separating gel. The samples were run at 15 mA through the 4% stacking gel and at 30 mA through the 12% separating gel until the dye front had reached the bottom (gel size: 10 * 10 cm). The gel was fixed in 45% methanol, 7% acetic acid for 20 min, soaked in AmplifyTM (Amersham) for 15 min and dried at 80oC for 1 h under vacuum. The gel was developed by autoradiography using Kodak BiomaxTM MR film, overnight and for 7 days at -70oC. Prestained protein markers were supplied by Bio-Rad for sizing of the protein products.
All samples were sequenced using an ABI PrismTM 377 DNA Sequencer (Perkin Elmer). Dye terminator cycle sequencing kits were provided by Perkin Elmer. PCR products were amplified using a forward primer identical to the one used in the PTT-PCR without the presence of the T7 polymerase binding site and the start codon. Purification of templates was done using WizardTM Minicolumns (Promega). Reaction volumes were halved as compared to ABI protocols and annealing temperatures of the sequencing primers was always adjusted to either 58oC or 60oC, depending on the primer used. Between 60-80 ng of template was needed per reaction. Generally between 350-450 bases of readable sequence were generated. Several mutations were sequenced in two different cDNA preparations and all were confirmed on the opposite strand in at least one PCR. Numbering of the TSC2 sequence was according to the published genomic structure, including exons 25 and 31 (8 ). Automated sequencing of genomic DNA containing the relevant exons was used to determine the origin of the abnormal mRNA sequence.
Where possible, restriction digestion analysis was performed to confirm the mutations found by sequencing. The 29 bp deletion in patients 1 and 2 disrupts a BspEI site. BspEI digestion in normal individuals results in two fragments of 872 and 302 bp, whereas the PCR product from deleted patients fails to digest, confirming the mutation.
Patient 4 has a 29 bp deletion in exon 37, which contains an internal HindIII site and creates a new NciI site. A PCR product spanning exon 37 was generated, and digested with HindIII and NciI. Failure to digest with HindIII indicates a loss of the sequence containing the HindIII site. Normal individuals would give digestion products of 298 and 157 bp. The extent of the deletion could be confirmed by NciI digestion. Normal individuals do not have an NciI site in the PCR product. However, this particular patient, having lost 29 bp, contained the mutant 284 and 142 bp fragments generated after digestion of the new NciI site, as well as the normal 455 bp allele. Unfortunately the 2 bp deletion in patient 2, and the 1 bp deletion in patient 3 did not disrupt an enzyme site and could therefore not be confirmed.
Exons 10 and 11 were amplified from cDNA using PCR primers from the flanking exons 9 and 12. Specific restriction enzyme sites are present in exon 10 (TaqI) and exon 11 (HinCII). PCR of cDNA missing either exon 10 or 11 gave a smaller than expected product. Digestion of the smaller PCR product with either TaqI or HinCII was used to determine whether exon 10 or exon 11 was absent. In addition to the cases studied, 15 control RNA samples were analysed for alternative exon 10 splicing.
We thank the families, clinicians and the Tuberous Sclerosis Association (UK) for their help in this work. We thank N. Froggatt for her advice on PTT, and D. Walsh for help with cell culture. I. van Bakel is supported by a grant from the Westminster Medical School Research Trust. T. Sepp is supported by a grant from the Tuberous Sclerosis Association (UK).
1 Gomez, M.R. (1991) Phenotypes of the tuberous sclerosis complex with a revision of diagnostic criteria. Ann. NY Acad. Sci., 615, 1-7.MEDLINE Abstract
2 Povey, S., Burley, M.W., Attwood, J., et al. (1994) Two loci for tuberous sclerosis: one on 9q34 and one on 16p13. Ann. Hum. Genet., 58, 107-127.MEDLINE Abstract
3 The European Chromosome 16 Tuberous Sclerosis Consortium. (1993) Identification and characterisation of the tuberous sclerosis gene on chromosome 16. Cell, 75, 1-11.
4 Kumar, A., Wolpert, C., Kandt, R.S., et al. (1995) A de novo frame-shift mutation in the tuberin gene. Hum. Mol. Genet., 4, 1471-1472.MEDLINE Abstract
5 Kumar, A., Kandt, R.S., Wolpert, C., Roses, A.D., Pericak-Vance, M.A., Gilbert, J.R. (1995) Mutation analysis of the TSC2 gene in an African-American family. Hum. Mol. Genet., 4, 2295-2298.MEDLINE Abstract
6 Vrtel, R., Verhoef, S., Bouman, K., et al. (1996) Identification of a nonsense mutation at the 5'-end of the TSC2 gene in a family with a presumptive diagnosis of tuberous sclerosis complex. J. Med. Genet., 33, 47-51.MEDLINE Abstract
7 Wilson, P.J., Ramesh, V., Kristiansen, A., et al. (1996) Novel mutations detected in the TSC2 gene from both sporadic and familial TSC patients. Hum. Mol. Genet., 5, 249-256.MEDLINE Abstract
8 Maheshwar, M.M., Sandford, R., Nellist, M., et al. (1996) Comparative analysis and genomic structure of the tuberous sclerosis 2 (TSC2) gene in human and pufferfish. Hum. Mol. Genet., 5, 131-137.MEDLINE Abstract
9 Brook-Carter, P.T., Peral, B., Ward, C.J., et al. (1994) Deletion of the TSC2 and PKD1 genes associated with severe infantile polycystic kidney disease-a contiguous gene syndrome. Nature Genet., 8, 328-332.MEDLINE Abstract
10 Povey, S., Armour, J., Farndon, P., et al. (1994) Report and abstracts of the Third International Workshop on Chromosome 9. Cambridge, United Kingdom, 9-11 April, 1994. Ann. Hum. Genet., 58, 177-250.MEDLINE Abstract
11 Green, A.J., Smith, M., Yates, J.R.W. (1994) Loss of heterozygosity on chromosome 16p13.3 in hamartomas from tuberous sclerosis patients. Nature Genet., 6, 193-196.MEDLINE Abstract
12 Wienecke, R., Konig, A., DeClue, J.E. (1995) Identification of tuberin, the tuberous sclerosis-2 product. Tuberin possesses specific Rap1GAP activity. J. Biol. Chem., 270, 16409-16414.MEDLINE Abstract
13 van der Luit, R., Meera Khan, P., Vasen, H., et al. (1994) Rapid detection of translation-terminating mutations at the adenomatous polyposis coli (APC) gene by direct protein truncation test. Genomics, 20, 1-4.
14 Xu, L., Sterner, C., Maheshwar, M.M., et al. (1995) Alternative splicing of the tuberous sclerosis 2 (TSC2) gene in human and mouse tissues. Genomics, 27, 475-480.MEDLINE Abstract
15 Sherratt, T., Vulliamy, T., Dubowitz, V., Sewry, C., Strong, P. (1993) Exon skipping and translation in patients with frameshift deletions in the dystrophin gene. Am. J. Hum. Genet., 53, 1007-1015.MEDLINE Abstract
*To whom correspondence should be addressed at: Department of Medical Genetics, Addenbrooke's NHS Trust, P.O. Box 134, Hills Road, Cambridge CB2 2QQ, UK. Tel: +44 1223 216446; Fax: +44 1223 217054; Email: agreen@hgmp.mrc.ac.uk
-->
This page is maintained by OUP admin. Last updated Wed Aug 13 15:52:16 BST 1997. Part of the OUP Journals World Wide Web service.
Copyright
Oxford University Press, 1996
