DHPLC Analysis of TSC1 and TSC2 Genes | Genetics | Biology

In this article we will discuss about the mutation detection in the Tuberous Sclerosis Genes TSC1 and TSC2 by DHPLC.

Overview of Tuberous Sclerosis (TSC):

TSC is a relatively variable clinical syndrome, in which several pathologic manifestations are highly specific for the disease. The cardinal feature is the cortical tuber, which is a distinctive form of brain hamartia. Hamartias and hamartomas usually involve several different organ systems in TSC patients, in addition to the brain.

Involvement of the skin, heart, and kidney are particu­larly frequent, occurring at some point during the lifetime of most patients. A hamartia is a group of dysplastic, disorganized cells within an organ, but occur­ring without unusual growth. A hamartoma is the occurrence of cells in a sim­ilar aberrant pattern, but with some growth tendency or potential.

TSC occurs as an autosomal dominant disorder. Large families with the con­dition are rare due to reduced reproductive potential among those affected, related to the occurrence of seizures, mental retardation, and significant devel­opmental delay. However, improvements in clinical care have led to increasing numbers of small families.

About two-thirds of TSC cases are sporadic, due to the occurrence of new mutations. TSC occurs in up to 1 in 6,000 live births without ethnic clustering. Linkage of TSC to chromosome 9q34 was achieved in 1987, and the corresponding TSC1 gene was identified in 1997.

Linkage of TSC to chromosome 16p13 was discovered in 1992, and the corresponding TSC2 gene identified in 1993. Among families large enough to permit link­age analysis, approximately half show linkage to 9q34 and half to 16p13. There is no evidence for a third locus.

Gene Structure of TSC1 and TSC2:

The TSC1 gene consists of 23 exons, of which 21 encode hamartin, with an 8.6-kb mRNA including a 4.5-kb 3′ untranslated region. There is variable splic­ing of the 2nd, untranslated exon. The gene occupies a genomic extent of 55 kb on 9q34. The TSC2 gene consists of 42 exons, of which 41 encode tuberin, with a 5.4-kb mRNA and relatively short 5′ and 3′ untranslated regions.

There is alternative splicing of exon 25, the first 3 bp of exon 26, and exon 31. Tuberin derived from mRNA isoforms missing exons 25 and 31 is relatively common, so that the functional importance of regions encoded by those exons appears minimal.

Mutational Analysis in TSC1 and TSC2- Strategic Considerations:

Several years ago when the mutational spectrum in the TSC1 and TSC2 genes was unknown, the best mutational analysis strategy was entirely uncer­tain. Well over 600 mutations in the TSC1 and TSC2 genes have now been identified (www.expmed(dot)bwh(dot)harvard(dot)edu/ts/review/), and they comprise a diverse set of inactivating mutations (see below).

It is clear that mutations can occur in nearly any of the exons of each of these genes, and that although there are ‘warmspots’ for mutation, there are no major hotspots for mutation that might have permitted a more focused mutation identification strategy.

It is also now clear that large deletions (>1kb) and genomic alterations account for at least 10% of all mutations occurring in these genes, particularly TSC2, so that any mutation identification strategy must include a method for identification of these mutations.

Numerous methods have been developed to scan amplified DNA fragments for genetic variation/mutation. In our laboratory, we have experience with five methods heteroduplex analysis by conformation sensitive gel electrophoresis (CSGE) on polyacrylamide gels, single-stranded conformation gel analysis (SSCP), denaturing gradient gel electrophoresis (DGGE), denaturing high pres­sure liquid chromatography (DHPLC), and re-sequencing.

The first four of these methods identify amplicons that are likely to contain sequence variants, but sequencing must be performed for confirmation and identification of the pre­cise variant.

The major focus of this article is to review our experience with DHPLC, but we briefly discuss our experience with these alternative methods. CSGE is an attractive method in which heteroduplexes are formed after PCR, and are iden­tified by polyacrylamide gel electrophoresis performed under partially denatur­ing conditions to provide the maximum discrimination between heteroduplexes and homoduplexes.

This method works extremely well for detection of insertion and deletion variants. Its sensitivity in detection of single nucleotide poly­morphism (SNP) or mutation variants is debated. In our hands, it was much less than 100% sensitive (see below).

It has been adapted for use of fluorescently labeled primers and analysis on ABI polyacrylamide gels and capillar­ies, and continues in use in some laboratories. SSCP analysis is one of the oldest methods available for variation analysis.

Although it continues to have adherents, and there are publications describing a sensitivity of over 95% for SNP detection when performed under a series of different conditions, our experience using it under more conventional conditions is that it is much less than 90% sensitive (see below). DGGE is another time-honored technique, which has been adapted to a multiplex format.

This method is extremely sen­sitive when performed properly, achieving SNP detection rates well over 90%. However, implementation of the method in multiplex format is difficult and time-consuming. Routine operation of the method is also complicated by par­tial assay failure. In addition, the equipment needed for operation of this method is non-standard and moderately expensive.

Resequencing analysis is increasingly popular as a method for identification of sequence variation. In high throughput genome-style laboratories with good bioinformatics support, this method is relatively efficient. In our limited expe­rience, the method is both expensive and time-consuming, in that the analysis of the sequence traces takes much time even for a skilled investigator or tech­nician.

Expense is considerably higher than any other method in which a muta­tion screening method is used followed by directed sequencing. However, as sequencing costs continue to fall, this approach becomes increasingly attrac­tive.

The infrastructure required in terms of sequencing capacity and bioinfor­matics is however significant. In addition, it is clear that resequencing is not 100% effective in detection of variation at the heterozygote level, and will miss many cases of mosaicism (see below).

Implementation of DHPLC Screening for Identification of Small Mutations in TSC1 and TSC2:

DHPLC analysis of heteroduplexes to detect mutations and sequence varia­tion has been used in our laboratory since 1999. The method is attractive because of the semi-automated nature of the analysis, so that 96 amplicons can be screened (14 to 18 hours per 96 well plate) easily during a 24-hour period in the standard analysis mode.

When we initially obtained our first machine, a WAVE DNA Fragment Analysis System (Transgenomic, Inc., Omaha, NE), there was relatively little guidance available for implementation. We designed primers so that exon amplicons would have a flat melting curve (www.insertion(dot)stanford(dot)edu/ meitdothtml), adding GC clamps (4 – 10b) to primers to make the melting curve have a variation of no more than 4°C over the length of the amplicon.

Amplicons were chosen to amplify all exons of TSC1 and TSC2, including at Least 20b of 5′ flanking and 10b of 3′ flanking sequence. For each amplicon, the run conditions on the DHPLC machine were determined empirically as follows. We used the Hansen and Oefner program (www.insertion(dot)stanford(dot)edu/meltdothtml) to determine a starting temperature for the DHPLC analysis.

Temperatures with­in a range of 2 degrees up or down from this temperature were then investi­gated, and the optimal temperature was chosen based upon three criteria: first, the DNA elution peak should be shifted by at least one minute to the left from the elution peak when run at 55°C; second, the highest temperature possible was used at which a homozygous sample eluted as a sharp peak; and third, the temperature should give the best resolution of heterozygous and homozygous samples, when samples with known mutations were available.

Although there have been some pitfalls in the use of the DHPLC machines, in terms of proper care and consistent performance, this approach has become and remained the standard approach for detection of small mutations in TSC1 and TSC2 in our laboratory. The primers and methods have been modified, tak­ing account of improved genomic sequence data for these genes.

Periodically, we have also reexamined the method sets for each of the amplicons. For example, some time ago we performed an analysis of the run temperatures derived from the Hansen and Oefner program, those derived from a proprietary program offered by Transgenomic, and our working methods.

We found that our oven temperatures were always within one or two degrees of the Hansen/Oefner pro­gram, but were generally 2-4°C hotter than those suggested by the Transgenomic software. We also tried out varying our temperatures for DHPLC runs of amplicons from multiple DNA samples for which mutations in TSC1 or TSC2 had not been found.

This exercise proved fruitless, and therefore we have continued to use our empirically derived temperatures. In addition, a blinded comparison showed that our DHPLC run conditions were capable of detecting nearly all sequence variation detected by other mutation screening procedures (see below).

The elution profile of each amplicon from the DHPLC machine is unique. Many are single sharp peaks but they can also be broad or have a shoulder. The development of software to permit discrimination of variants among the elution profiles would be beneficial.

Consequently, the human eye is at present the best detector of shifts that indicate the presence of variation. Printouts of elution profiles stacked one on top of the next facilitates visual discrimination of vari­ant elution profiles.

Comparison of DHPLC with CSGE and SSCP:

We performed a blinded comparison among three methods of mutation detection: CSGE, SSCP, and DHPLC. CSGE and DHPLC were applied in the analy­sis of the first 20 exons of TSC2 in 84 patients.

All sequence variants detected by CSGE were identified by DHPLC, consisting of 7 distinct insertion or deletion variants and 13 distinct single base substitutions. DHPLC analysis identified an additional 14 distinct SNPs, so that CSGE had detected only 15 of 28 (54%) SNPs overall.

In a second comparative trial we analyzed 15 DNA samples in which 18 distinct sequence variants had been detected by SSCP. In the blind­ed DHPLC analysis, all 18 variants were detected, and an additional 7 distinct sequence variants were identified. Thus, overall 11 of 16 (69%) SNPs were iden­tified by SSCP in these samples, while all were identified by DHPLC.

In a third comparative trial, sequence variants in the TSC1 gene that had been identified by a variety of techniques (CSGE, SSCP, and DGGE) were analyzed by DHPLC. Twenty-seven of 28 (96%) sequence variants were identified by DHPLC.

The one that was missed was present as a mosaic mutation, and had been detected by CSGE. Four or more additional distinct sequence variants were identified by DHPLC in these samples, and all of these had been missed by the earlier techniques.

Current Results of Mutational Analysis in TSC1 and TSC2:

Employing DHPLC as the primary mutation screening approach, we have studied 660 TSC patients. This set includes 224 patients which we have previ­ously reported. The remainder are a larger new group, not previously report­ed, for which the analysis is not yet complete.

Specifically, not all assays have been performed on each patient, some DNA samples are depleted, and careful review of clinical data has yet to be performed to remove patients who fail to meet TSC diagnostic criteria. Overall we have identified mutations in TSC1 or TSC2 in 469 (71%) patients (Figure 4-1, Table 4-1). This overall set of TSC patients and mutations is much larger than any other previously reported.

The large deletion and duplication mutations in this set have been identi­fied through the use of quantitative PCR (QPCR) and long range PCR (LRPCR) assays. We have developed these methods for use in the analysis of TSC1 and TSC2. The QPCR assay is performed by simultaneous amplification of 2-3 amplicons within TSC2 and a control amplicon, all with FAM-labeled primers and of different size, with only 18 cycles of PCR, followed by ABI 3100 capillary electrophoresis.

The method is sensitive, quantitative, and quite reproducible in our hands in the detection of reduced (indicating allele loss) or increased (indicating duplication of an allele) representation of an exon of TSC2 in the starting material. The LRPCR assay involves the generation of 19 amplicons providing overlapping coverage of the entire genomic extent of the TSC2 gene. The products vary in size from 1.7 – 11.6 kb and are analyzed by gel elec­trophoresis.

Analysis of Mutation Distribution and Frequency in TSC1 and TSC2:

We observe that there is a very broad spectrum of mutations in both TSC1 and TSC2, both in terms of type and distribution of mutations among the exons of these genes (Figure 4-1).

However, the distribution of mutations is not com­pletely uniform. Some exons of each gene are not involved by mutation in this set of patients, including the alternatively spliced exons 25 and 31 of TSC2, but also TSC1 exons 3, 6, 16, 22, and 23; and TSC2 exon 15. TSC2 mutations are much more common than TSC1 mutations (4.2:1 ratio), and this ratio is higher than that predicted by their relative genomic extents and coding regions. This observation may reflect a higher intrinsic mutation rate in TSC2.

Missense, in-frame deletion, and large deletion mutations are very rare in TSC1, although we have identified several possible and one proven TSC1 missense mutations (370T>C 50L>P; 403T>C 61L>P (confirmed by parental analysis); 499T>G 93L>R; 790G>C 190R>P; 958G>C 246R>T).

Two TSC1 deletion mutations have also been identified, including one characterized at the molecular level as a deletion of 1279 bp, extending from intron 16 to exon 18. Interestingly, the frequency of small deletion, missense, nonsense, splice, and big deletion mutations are all approximately the same in TSC2 (range 8.2% – 12.4%).

The distribution of mutations within TSC1 and TSC2 is clearly uneven (Figure 4-1). There are clear warmspots for mutation including exons 8 – 10, 15, and 17 – 19 of TSC1, and exons 13, 16, 26, 29, and 33 – 40 of TSC2. Some mutations are due to transitions of CG sequences to TG sequences creating stop or missense codons: in exons 8, 15, 17, and 18 of TSC1; and in exons 14, 16, 20, 33, and 38 of TSC2.

Most deletions occurring in these genes occur at short repeated sequences, as seen elsewhere in the genome, and similarly nearly all insertions represent single or multiple base duplication events. The in-frame deletion mutation TSC2 E40:5238-5255del18 is the most common mutation identified, accounting for about 2% of all mutations seen (Figure 4-1). There are several other mutations in TSC2 that are seen at about 1% frequency.

Mosaicism in TSC:

Mosaicism is well known in TSO as well as other tumor suppressor gene syndromes in which there is a high rate of sporadic cases due to new mutations. Detection of mosaic mutations is challenging, and it is very likely they would often be missed when sequencing alone is used as a strategy for mutation detection.

Although DHPLC cannot be expected to have 100% sensitivity for detection of mosaic mutations, it is likely to be more sensitive than sequence analysis. Using DHPLC, we often identify elution shifts for exon amplicons, for which subsequent sequence analysis fails to indicate a sequence variation. In these cases, we may alter the amplicon or sequencing primers to attempt to identify the cause of the elution shift, and are at times successful.

If these efforts fail, then we clone the amplicon products (Figure 4-2). Fifty clones are chosen and screened with vector primers to ensure that they all con­tain the correct sized insert. At least 25 clones and a DNA sample known not to contain a sequence variant (control) are then amplified using the amplicon primers.

The clone amplicons are then individually mixed with the control prod­uct, heteroduplexed, and analyzed on the DHPLC machine. Clones which give an elution shift in this analysis that is similar to (but often more prominent than) that seen with the original patient DNA sample are then sequenced. The muta­tion is typically easily identified in this case, as there is a single allele present in the clone.

To verify that the sequence variant is real and not the product of PCR or cloning artifact, we require that at least three independent clones show the same sequence variation. We also review the sequence traces of the patient for that amplicon looking for minor evidence of the sequence variant.

When performed in this way, in the last 120 patients studied we identified mosaic mutations in 4 (3%) patients. By counting the number of clones containing the mutation, we determined that the level of mosaicism in these patients was 7 – 26 percent (Table 4-2).

Conclusion:

DHPLC is a highly sensitive method for detection of mutations and other sequence variants, following PCR amplification. Our comparative trials have demonstrated that it has a sensitivity rate of 99% for variation detection, in comparison to much lower rates for CSGE and SSCP. The single case that was missed in these comparisons was a mosaic mutation.

Our studies have also shown that DHPLC is capable of detection of mosaic mutations in a substantial though small fraction (3%) of the TSC patients we have screened for mutations, and the method has detected mosaicism as low as 7%, clearly outperforming re-sequencing analysis.

DHPLC does require moderate capital investment and training and develop­ment of expertise in the operation of HPLC machinery. Nonetheless, once mas­tered it is an easily workable approach to the detection of mutation in human disease genes. The automated operation characteristics of the Transgenomic, Inc. machine are attractive, enabling moderate throughput.

Advances in soft­ware to permit automated screening of elution profiles would be valuable. Even without that advance, the greatest asset of the system is the speed with which elution profiles can be visually screened to permit identification of those con­taining sequence variation, and permitting a focused sequencing approach to mutation identification.