Autosomal dominant polycystic kidney disease (ADPKD) is a common genetic disease of the kidney (frequency 1/1000) accounting for approximately 5% of end-stage renal disease (ESRD) in Western countries. The disease is progressive with cyst development and enlargement resulting in bilateral polycystic kidneys and ESRD, typically in late middle age. ADPKD is genetically heterogeneous with two genes, PKD1 (chromosome region 16p13.3) and PKD2 (4q21) identified and characterized, plus evidence of rare unlinked families1,2,3,4,5. PKD1 accounts for approximately 85% of cases and is associated with a more severe disease course (average age at onset of ESRD of 53 compared to 69 years for PKD26). The PKD1 and PKD2 genes encode related proteins, polycystin-1 and -2, respectively, which probably form part of a receptor/channel complex involved in regulating intracellular Ca2+ levels7.
Presymptomatic diagnosis of ADPKD is possible by various imaging methods (abdominal ultrasound, computer tomography and magnetic resonance) and is relatively reliable in adult patients, especially in the more severe PKD1. However, there is a role for genetic diagnosis in this disorder, especially in patients with equivocal imaging results, those with a negative family history and in cases where a definite diagnosis is required in younger individuals, such as for living related kidney donors. Genetic diagnosis by linkage analysis is possible but is of limited utility due to the requirement for a relatively large family with multiple diagnosed members, to determine the gene involved.
Mutation analysis would be a more direct method of genetic diagnosis and allow genotype/phenotype correlations to be studied. However, mutation screening has proven uniquely difficult in ADPKD. The PKD1 transcript is large, encoded by 46 exons, and embedded in a complex duplicated genomic area1,2. The 5' region of PKD1, from upstream of exon 1 to exon 33, is reiterated at least five times further proximally on the same chromosome. As this reiteration is relatively recent in evolutionary time, only approximately 2% sequence divergence exists between the PKD1-like pseudogenes (HG loci) and PKD1, complicating specific amplification of the disease gene8. In addition to these structural problems, the genetic heterogeneity and marked allelic heterogeneity at PKD1 and PKD2, with most mutations unique to a single family9,10, have further complicated direct mutation based genetic diagnosis. Although, several screens of the entire approximately 3kb coding region of PKD2 have been described10,11,12, only limited analyses of the duplicated part of PKD1 have been published. These studies have used methods of long-polymerase chain reaction (PCR), with primers anchored in single copy DNA, and PKD1 specific-PCR, exploiting the rare differences between the PKD1 and HG loci, to analyze the duplicated region9,13,14,15,16,17,18,19. To date, only one screen of the entire PKD1 gene has been described, showing a mutation detection rate of approximately 50%9. However, the methods used in that study, including the use of mRNA derived template and the protein truncation test, are such that the strategy is not readily adapted for routine diagnostic needs.
Recently, a new semi-automated method has been introduced for mutation analysis in human genes, denaturing high-performance liquid chromatography (DHPLC)20,21. This system is based on the differential adsorption of homo- and hetero-duplexes to a hydrophobic matrix in a chromatographic column. Amplicons with a mismatch have decreased interaction with the matrix and are thus eluted earlier than the normal amplicon, resulting in an alteration of the elution profile. This method has a high level of sensitivity and specifically, is semi-automated and does not require analysis by electrophoresis. Due to the advantages of this technique it has now become the method of choice for screening large and multi-exon genes involved in human genetic disease22.
In the current study, we describe a mutation screen of the entire PKD1 and PKD2 genes by DHPLC using genomic DNA from a cohort of 45 ADPKD patients. An overall detection rate of approximately 70% was achieved. We propose this approach as an efficient and effective way to perform molecular diagnostics in ADPKD.
METHODS
Selection of the cohort
Forty-five unrelated ADPKD probands were selected for screening. Informed consent was obtained from each participant and the study was approved by the appropriate IRB and/or Ethics Committee. This cohort consisted of typically sized families that had not previously been subjected to linkage analysis. All probands in the study met established ultrasound criteria for ADPKD diagnosis23. A blood sample for DNA isolation was drawn from each proband and from available family members interested in participating.
DNA isolation and field inversion gel electrophoresis (FIGE)
Genomic DNA was extracted by standard phenol-chloroform procedures. An aliquot of DNA (5 
g) was digested with EcoRI and separated by FIGE using the FIGE Mapper Electrophoresis System (Bio-Rad, Herts, UK). Gels were Southern blotted and hybridized with the PKD1 IVS1 probe CW10, as previously described1.
PCR amplification and generation of DHPLC amplicons
The duplicated region of PKD1 was amplified as five PCR fragments that were either anchored in single copy DNA or mismatched with HG sequence (details are in the Results section and Table 1). Long-PCR fragments (>3kb) were amplified as previously outlined9 using the rTth DNA polymerase (PE Applied Biosystems) in the supplied DMSO containing buffer. A combination of Amplitaq (PE Applied Biosystems, Foster City, CA, USA) and Taq Extender (Stratagene, Heidelberg, Germany) was used for the GC-rich exons 1 of PKD1 and PKD29. Each amplicon for DHPLC analysis was amplified as a fragment of 150 to 450 bp using Amplitaq and the supplied (non-DMSO) buffer for PKD2 and a dimethyl sulfoxide (DMSO) containing buffer for PKD1 (detailed in Tables 2 and 3 and9,24). The primers were positioned approximately 20 to 30 bp from the intron-exon boundary to allow the detection of splicing defects but minimize intronic polymorphism. Due to the presence of a frequent intronic deletion (IVS9+28del7, found in 8 of 46 samples), a second reverse primer (9 int), positioned only 2 bp from the intron/exon boundary, was used to screen samples carrying the deletion polymorphism. The availability of a known mutation of this splice site (IVS9 +1G
T) was used to show the utility of the closer primer.
Heteroduplexes of amplicons were generated by heating the PCR products at 95°C for three minutes, cooling 0.1°C per second to 65°C, incubating at 65°C for 30 minutes, cooling 0.1°C per second to 37°C, with a final incubation of 10 minutes.
DHPLC and gradient generation
Denaturing high-pressure liquid chromatography was performed using the Wave system (Transgenomic Inc., Omaha, NE, USA). Crude PCR product, sufficient to generate an elution peak of 3 to 5mV (200 to 500 ng, in Buffer C), was injected into a preheated, fully equilibrated chromatographic column (C18 reverse-phase, non-porous polystyrene-divinyl benzene matrix; DNASep Column; Trangenomic Inc.). A linear gradient of 5% triethylammonium acetate (TEAA) (Buffer A) and 25% acetonitrile + 5% TEAA (Buffer B) with a flow rate of 0.9 mL/min was generated for each amplicon. Each DHPLC run included a DNA loading step (5% drop for loading of Buffer B), a linear separation gradient (2% Buffer B slope per min, 4.5 min), a wash step (75% acetonitrile; Buffer D, 0.5 to 1 min) and an equilibration step (0.9 to 1.2 min). When DMSO buffer was used in the PCR reaction, longer wash and equilibration steps were required. Total run times were therefore 9.4 minutes and 8.6 minutes for the PKD1 and PKD2 amplicons, respectively.
Fragment melting profile analysis and optimization of DHPLC conditions
The melting profile of each fragment was analyzed using the Wavemaker version 4.0.32 (Transgenomic Inc.) software. The process by which the melting point is determined is based on the Fixman and Friere implementation of Poland's algorythmn25, which calculates the probability that a base is in the helical duplex form or the non-helical, single-stranded form.
DHPLC conditions for each amplicon were optimized by use of the available positive controls (details are in Tables 2 and 3). Each fragment was eluted at the predicted melting temperature, and plus and minus 1 and 2°C. If multiple domains were predicted by the Wavemaker software, then a wider range of temperatures was tested and the final analysis performed under two sets of conditions (see Tables 2 and 3 for the final conditions for each fragment). A set of eight random samples, plus the positive control, were used for each optimization process and conditions were adjusted so that the DNA eluted from the column at, or greater than, four minutes. Resolution was optimized so that the elution profile of the positive control differed most from the normal sample. When no positive control was available, conditions were considered to be optimal at the temperature and corresponding gradient immediately before a significant decrease in the retention time of the amplicon was observed and/or an excessive broadening of the peak, indicating excess denaturation. This temperature was typically located in the range of 50 to 75% helical fraction for the amplicons analyzed.
DNA sequencing
Samples showing an aberrant elution profile were re-amplified and subjected to direct sequencing using the Big-Dye Terminator kit (PE Applied Biosystems). The PCR product was purified by polyethylene glycol 8000 (PEG) and ethanol precipitation and resuspended in water. An aliquot of the product (30 to 180 ng) was combined with each primer (3.2 pmol) and DMS0 (1 L) for sequencing reactions, analyzed on an ABI377 sequencing machine and assembled using Sequencher 3.1.1 software.
Validation of mutations
To analyze the segregation in families of characterized mutations the DHPLC elution profiles of all family members (always including two affected cases from different generations) were compared. Due to the reproducibility of the elution profile, the mutant genotype was assigned to samples resembling the mutant profile. When more than one sequence change was present in the same fragment, validation was performed by direct sequencing of all samples. Missense mutations were screened on 146 normal chromosomes by DHPLC.
RESULTS
DHPLC analysis of the ADPKD genes
A method was developed to screen the entire PKD1 and PKD2 genes from genomic DNA by DHPLC. The duplication of the PKD1 genomic region (exons 1 to 33) required that first-round long-PCR reactions were necessary to specifically amplify these exons. This was achieved for the regions bordering single copy DNA by employing a primer anchored in single copy DNA; the method used to amplify exon 1 and exons 22 to 34 has previously been described (Table 1 and Figure 1)9. For the rest of the duplicated area, specific amplification of PKD1 was obtained by employing primers that matched differences between PKD1 and HG sequences. Two previously described primer pairs exploiting a number of mismatches in exon 15 were used to amplify exons 13 to 2117. To analyze exons 2 to 12, one primer was positioned in the region of IVS1 that is deleted in most copies of the HG8 and the second in exon 13, which is not present in the IVS1-containing HG locus (manuscript in preparation). The specificity of these products was determined by amplification of somatic cell hybrids containing either just PKD1 (radiation hybrid Hy145.19) or only the HG loci (P-MWH2A)1.
Figure 1.
Map of the PKD1 gene showing the fragments employed for specific amplification of the duplicated part of the gene (top) and the locations of detected mutations in the indicated pedigrees (bottom; details are in Table 4). The mutation types are: I/F, in-frame deletion; M, missense; N, nonsense; F/S, frame-shifting deletion/insertion; and S, splicing.
Full figure and legend (26K)All exons were amplified as fragments of 150 to 450 bp, a size range found to be ideal for DHPLC analysis Tables 2 and 3. For the duplicated region, the PKD1-specific first-round product was used as the template, while for the remainder of PKD1 and for PKD2, DHPLC amplicons were amplified directly from genomic DNA. Large exons were split into more than one fragment (exon 15 was analyzed as 13 overlapping fragments) and intronic primers located 25 to 30 bp from the intron/exon boundaries. The 15 PKD2 exons were amplified using previously described primer pairs and conditions Table 324. In total PKD1 was amplified as 64 fragments and PKD2 as 17, making a total of 81 to analyze both genes.
The temperatures and initial Buffer B concentration of the elution gradient for DHPLC analysis were determined for each of the amplicons using the Wavemaker software and empirically (Methods section; Tables 2 and 3). Analysis temperatures for PKD1 were generally higher than for PKD2, reflecting the higher GC content. If the melting profile of the fragment was reasonably homogenous one analysis temperature was used. However, for 14 PKD1 and 8 PKD2 fragments, two different melting domains were detected and, in these cases, the amplicon was analyzed under two different sets of conditions Tables 2 and 3. An important determinant for establishing ideal DHPLC conditions was the availability of positive controls. Previously characterized mutations or polymorphisms were available as positive controls for 58/64 PKD1 and 15/17 PKD2 amplicons Tables 2 and 3. These controls included a variety of changes, but the most subtle change (a substitution) was selected if available. As well as helping to determine ideal analysis conditions, the positive controls also allowed the quality of analysis to be monitored by inclusion with each run of the fragment. The analysis of positive controls also permitted the sensitivity of the method to be assessed. All of our previously described base-pair mutations and polymorphisms9 were detected with the amplicons and analysis conditions described in our current study, indicating that DHPLC is a very sensitive method for detecting base-pair changes.
Mutation screening by DHPLC in a cohort of ADPKD patients
To determine the achievable detection rate, a cohort of 45 uncharacterized ADPKD patients was analyzed. Initial screening of PKD1 for large-scale rearrangements by FIGE did not identify any mutations. The entire coding regions of both the PKD1 and PKD2 genes (or until a clear mutation was identified) were screened by DHPLC. A total of 266 aberrant profiles were detected and characterized, illustrating the variable nature of the PKD1 gene. In the 32 patients in whom both genes were fully screened, an average of 5.4 DNA changes (range 0 to 14) were found per patient. These changes were characterized by sequencing, and segregation with the disease was analyzed when samples from other family members were available. Figures 2 and 3a show examples of aberrant DHPLC profiles from analyses of the PKD1 and PKD2 genes and examples of sequencing are shown in Figure 4. The way in which the reproducibility of the DHPLC profile was used to trace the mutation within a family is illustrated in Figure 3b.
Figure 2.
Examples of aberrant denaturing high-pressure liquid chromatography (DHPLC) profiles obtained from analysis of PKD1 exon fragments, as indicated.
Full figure and legend (51K)Figure 3.
(A). Examples of aberrant DHPLC obtained from analysis of PKD2 exon fragments, as indicated. (B; top) Segregation analysis of the mutation Q2158X; found in the proband (2116) and affected father (2115) but not normal control and (bottom) S2164X; present in the proband (1565) and her affected father (1590), but not unaffected brother (1652) or control.
Full figure and legend (38K)Figure 4.
Examples of sequence data showing: (A) nonsense mutation Q266X; 1007CT; (B) missense change Y2092C; 6486AG and (C) insertion, 12521insA.
Of the total changes, 29 were definite mutations, 26 from PKD1 and 3 from PKD2, consisting of 13 nonsense mutations, 12 truncating deletions and/or insertions, a large and a small in-frame deletion and 2 splicing defects Table 4. One of these mutations, 5225delAG (a previously described mutation;15) was found in two families, while two nonsense mutations, Q1838X and S2164X were present in the same patient and found to be inherited on the same chromosome from the affected father. In this case it is not clear if both mutations occurred simultaneously or if a second mutation occurred on an already mutant allele. In total, definite mutations were identified in 29 of 45 pedigrees (64%). Samples were available to show segregation with the disease in 16 of these pedigrees.
A second group of probable mutations were defined, consisting of non-conservative missense changes in which segregation with the disease was demonstrated. These changes were not found on 146 normal, unrelated chromosomes or previously described as a polymorphism. Two mutations fitted this category, Y2092C, a change in the 16th PKD repeat and L3851P a previously described26 substitution in the 3rd extracellular loop in the transmembrane region. Leucine 3851 is a highly similar valine in polycystin-1 orthologs in mouse and Fugu fish, while tyrosine 2092 is conserved in the C-
strand of many PKD repeats27 and also found at this position in the mouse, although it is a cysteine in Fugu28. Two further non-conservative missense changes: R1340W (PKD repeat 7 and conserved as a positively charged histidine in mouse and Fugu) and E1811K (PKD repeat 12, conserved as a negatively charged glutamic acid or aspartic acid in mouse and Fugu) were characterized. Neither of these changes was found in the cohort of normal individuals or has previously been described, however, due to the lack of available samples, segregation could not be tested. Finally, a conservative change, I3167F, which has not been found on normal chromosomes, but where segregation could not be demonstrated, was identified. This residue is an isoleucine in mouse and Fugu and lies within the possible lipid binding PLAT domain29. Although it is difficult to prove if these substitutions are mutations, it is worth noting that the entire coding region of PKD1 and PKD2 was screened and these were the only likely pathogenic changes detected. Inclusion of the segregating missense changes increases the overall detection rate to 69%, while inclusion of all the possible missense substitutions shows a mutation detection rate of 76%.
In addition to the mutations, and reflecting the frequency of DHPLC profile changes, a high level of polymorphism was detected. Seventy-one different polymorphic changes to PKD1 and two PKD2 polymorphisms were detected Table 5. Thirty-four of these have been described in previous studies, leaving 53% as newly characterized changes. The changes were defined as polymorphisms because they were intronic changes beyond the splice consensus sites (N = 14), in the 3' UTR (N = 1), silent exonic changes (not altering an amino acid) (N = 35) or amino acid substitutions that did not segregate with the disease in this or previous studies (N = 21). These polymorphisms are clearly a complication to mutation analysis in PKD1. However, because of the reproducible nature of the DHPLC profile (demonstrated by the segregation analysis; Figure 3b), frequent polymorphisms, once identified by sequencing, were recognized in subsequent samples by DHPLC "signature-based" genotyping. A total of 54 (
20%) DHPLC false positives were detected in the screen of 45 patients, in which no sequence change was identified. This, however, probably overestimates the final false positive rate that decreased as optimal analysis of each fragment was achieved during the study.
DISCUSSION
This study describes the first screen of the entire PKD1 and PKD2 genes for mutation. A largely genetically uncharacterized cohort of ADPKD patients was analyzed, similar to that which may be seen in a routine diagnostic setting. The method employed, DHPLC, has proven suitable for rapid detection of base pair changes with an overall detection rate of approximately 70%.
DHPLC is becoming the method of choice for analyzing large multi-exon genes, such as those causing cystic fibrosis and tuberous sclerosis30,31,32, because of the semi-automation of the analysis, the high level of sensitivity (100% of the known mutations in this study) and the cost of analysis, calculated at $1 per sample run31. While direct sequencing may be considered the best means to analyze small genes (with few exons)33, larger more complex genes benefit from a screening approach, such as DHPLC, to flag abnormal exons, which thus decreases the number of exons to be sequenced, and hence the total cost of the analysis. This consideration is of particular importance when selecting a method for routine molecular diagnostics where high throughput and cost considerations are important.
The results of this screen have illustrated the sensitivity of DHPLC with all previous changes detected and the majority of new changes being simple substitutions. The position of the change within the amplicon did not influence the detection rate, as in some other conformation sensitive systems (such as SSCA and gel-based heteroduplex analysis34), with substitutions detected immediately adjacent to the primer (for example, see exon 9 in the Methods section). This allowed the use of primers just 20 to 30 bp from the intron/exon boundaries, minimizing the detection of intronic polymorphisms. Initial analysis of the single copy exons of PKD1 with previously described primers located 40 to 60 bp from the splice sites35 showed an unacceptably high level of variant profiles due to intronic changes. Our screen of the ADPKD genes also has allowed the optimal size of amplicons to be analyzed. Previous studies have suggested optimal sizes for DHPLC analysis of 150 to 700 bp22, however, we found a significant decrease in detection rate (using positive controls) and an increased rate of false positives in fragments greater than 450 bp. Consequently, although larger fragments would have been beneficial to minimize the number of analyses for large exons (like exon 15), amplicons of no larger than 450 bp were employed.
Conditions for DHPLC were determined using the Wavemaker 4.0.32 software, but for some regions were also calculated employing a Stanford University Web based site, Stanford Melt program (http://insertion.stanford.edu/melt.html). Although the predictions were generally similar, for some amplicons the WM4.0.32 program predicted significantly higher temperatures and in general we had greatest success with these calculations. For a few fragments, neither prediction was accurate and in these cases the availability of positive controls was invaluable for establishing optimal analysis conditions.
A particular problem that was previously evident in mutation screens of PKD1, but emphasized in this study of the entire gene for base-pair changes, is the level of polymorphism9,16,36. In cases where the entire gene was analyzed, an average of 5.4 changes were detected, indicating that the ratio of polymorphic changes to mutation is very high at 4:1 to 5:1. The problem of common polymorphism was overcome to some extent by using the reproducible shape of the DHPLC profile to identify the change without sequencing ("signature-based" genotyping). Furthermore, because of the change specific shape of the profile, we did not find that common polymorphisms masked other mutations. When a mutation and polymorphism were detected in the same fragment a novel profile was generated.
The difference in the rate of polymorphism between PKD1 and PKD2 was striking in this study, with only two polymorphisms detected in PKD2. Although it is possible that there are special reasons for the level of changes at PKD1 (such as the polypyrimidine tract in IVS2137 or conversion by HG changes14) the GC rich nature of PKD1 is probably the major explanation. The GC richness results in a high level of CpG dinucleotides that are known to be hot spots for mutation because the methylated cytosine is susceptible to spontaneous deamination to thymine38. Although this change only accounted for 4 of 20 mutation-associated substitutions in PKD1, 54% of substitution polymorphisms occurred at CpG dinucleotides. An influence of the unusual IVS21 sequence was not clear, as changes were found throughout the gene and no mutations and only 8 of 70 polymorphisms matched known HG sequences. The large size of the PKD1 transcript also may partly explain the higher level of polymorphism (a target four times greater than PKD2).
The high level of polymorphic change at PKD1 makes it difficult to determine if a missense substitution is a mutation. Tests of whether the change is conservative and if the residue is conserved in polycystin-1 orthologs seems to be of only limited value as some of the polymorphisms also appear to be non-conservative and some are at conserved sites9. Segregation with the disease is an important test, as lack of co-inheritance rules out the change as a mutation, as long as the diagnosis in other family members is clear. Therefore, even within a routine molecular diagnostic setting, the availability of at least one sample from another affected family member would be beneficial. Of course, segregation with the disease is not proof that the missense change is a mutation, but hopefully as more molecular screening of PKD1 is reported, it will be possible to assemble lists of segregating mutations and non-segregating polymorphisms. The difficulties associated with missense changes are, however, likely to remain a complication of analysis at this locus until a functional test for the protein is available.
The uncertainty over the status of many missense changes means that the final mutation detection rate in this study cannot be reported with certainty. The lowest rate, just considering the definite mutations, is 64% and higher than previous studies of PKD1 alone, partly reflecting the approximately 10% of mutations to the PKD2 gene detected in this typical ADPKD population. If all the possible missense changes also are considered as mutations, the detection rate rises to 76% and is comparable to comprehensive screens of other large genes associated with disorders with genetic and allelic heterogeneity, such as tuberous sclerosis or Alport syndrome32,39. The undetected changes presumably consist of more subtle splicing defects (caused by changes away from the donor and acceptor splicing sites and by silent exonic changes), promoter mutations and the likelihood of a low level of changes undetected by DHPLC. In addition, larger rearrangements of PKD2 were not sought and the possibility that some of these cases may be due to as yet unmapped ADPKD genes should be considered.
As in previous studies, mutations were spread throughout the PKD1 and PKD2 genes and, except for one case, were unique to a single family within the study9. However, with the accumulating data on mutations, it is interesting to see that five changes matched previously described mutations Table 4. Even if all the putative substitutions are included as mutations the majority of changes (82%) were clear truncating or substantial in-frame changes, likely to inactivate the gene. This finding is of importance in the diagnostic setting indicating that, despite the problem differentiating missense mutations and neutral polymorphisms, clear results of mutation detection could be reported in 64% of cases. We had previously estimated from sequencing small regions of PKD1 that missense mutations would represent a higher level of mutations9, but that has not been borne out by this study. It remains possible, however, that missense mutations in highly conserved regions of the protein will be important sites of mutation. The nature of the detected changes is consistent with cyst development being associated with polycystin loss, through a two-hit mechanism40, or haploinsufficiency. Despite the evidence that mutations may inactivate, it will be interesting to see if any significant correlations with phenotype are associated with the type and/or position of mutations. Such studies should now be possible with this improved and simplified mutation screening approach.
We have illustrated the possibilities for routine molecular diagnosis in ADPKD by direct mutation detection using DHPLC. Genetic diagnosis is likely to become increasingly important as potential therapies become available, to determine the status of at-risk individuals at an early stage before significant renal damage has occurred.
References
| 1. | Consortium EPKD. The polycystic kidney disease 1 gene encodes a 14 kb transcript and lies within a duplicated region on chromosome 16. Cell 1994; 77: 881–894. | PubMed | ISI | |
| 2. | Hughes J, Ward CJ & Peral B et al. The polycystic kidney disease 1 (PKD1) gene encodes a novel protein with multiple cell recognition domains. Nat Genet 1995; 10: 151–160. | Article | PubMed | ISI | ChemPort | |
| 3. | International Polycystic Kidney Disease Consortium. Polycystic kidney disease: The complete structure of the PKD1 gene and its protein. Cell 1995; 81: 289–298. | PubMed | ISI | |
| 4. | Mochizuki T, Wu G & Hayashi T et al. PKD2, a gene for polycystic kidney disease that encodes an integral membrane protein. Science 1996; 272: 1339–1342. | PubMed | ISI | ChemPort | |
| 5. | Daoust MC, Reynolds DM & Bichet DG et al. Evidence for a third genetic locus for autosomal dominant polycystic kidney disease. Genomics 1995; 25: 733–736. | Article | PubMed | ISI | ChemPort | |
| 6. | Hateboer N, van Dijk MA & Bogdanova N et al. Comparison of phenotypes of polycystic kidney disease types 1 and 2. Lancet 1999; 353: 103–107. | Article | PubMed | ISI | ChemPort | |
| 7. | Somlo S & Ehrlich B. Human disease: Calcium signaling in polycystic kidney disease. Curr Biol 2001; 11: R356–R360. | Article | PubMed | ISI | ChemPort | |
| 8. | Loftus BJ, Kim U-J & Sneddon VP et al. Genome duplications and other features in 12 Mbp of DNA sequence from human chromosome 16p and 16q. Genomics 1999; 60: 295–308 10.1006/geno.1999.5927. | Article | PubMed | ISI | ChemPort | |
| 9. | Rossetti S, Strmecki L & Gamble V et al. Mutation analysis of the entire PKD1 gene: Genetic and diagnostic implications. Am J Hum Genet 2001; 68: 46–63. | Article | PubMed | ISI | ChemPort | |
| 10. | Veldhuisen B, Saris JJ & de Haij S et al. A spectrum of mutations in the second gene for autosomal dominant polycystic kidney disease (PKD2). Am J Hum Genet 1997; 61: 547–555. | PubMed | ISI | ChemPort | |
| 11. | Pei Y, He N & Kasenda M et al. A spectrum of mutations in the polycystic kidney disease 2 (PKD2) gene from eight Canadian kindreds. J Am Soc Nephrol 1998; 9: 1853–1860. | PubMed | ISI | ChemPort | |
| 12. | Torra R, Viribay M & Tellería D et al. Seven novel mutations of the PKD2 gene in families with autosomal dominant polycystic kidney disease. Kidney Int 1999; 56: 28–33. | Article | PubMed | ISI | ChemPort | |
| 13. | Peral B, Gamble V & Strong C et al. Identification of mutations in the duplicated region of the polycystic kidney disease 1 (PKD1) gene by a novel approach. Am J Hum Genet 1997; 60: 1399–1410. | PubMed | ISI | ChemPort | |
| 14. | Watnick TJ, Gandolph MA & Weber H et al. Gene conversion is a likely cause of mutation in PKD1. Hum Mol Genet 1998; 7: 1239–1243. | Article | PubMed | ISI | ChemPort | |
| 15. | Watnick T, Phakdeekitcharoen B & Johnson A et al. Mutation detection of PKD1 identifies a novel mutation common to three families with aneurysms and/or very-early-onset disease. Am J Hum Genet 1999; 65: 1561–1571. | Article | PubMed | ISI | ChemPort | |
| 16. | Thomas R, McConnell R & Whittacker J et al. Identification of mutations in the repeated part of the autosomal dominant polycystic kidney disease type 1 gene, PKD1, by long-range PCR. Am J Hum Genet 1999; 65: 39–49. | Article | PubMed | ISI | ChemPort | |
| 17. | Phakdeekitcharoen B, Watnick TJ & Ahn C et al. Thirteen novel mutations of the replicated region of PKD1 in an Asian population. Kidney Int 2000; 58: 1400–1412 10.1046/j.1523-1755.2000.00302.x. | Article | PubMed | ISI | ChemPort | |
| 18. | Phakdeekitcharoen B, Watnick T & Germino GG. Mutation analysis of the entire replicated portion of PKD1 using genomic DNA samples. J Am Soc Nephrol 2001; 12: 955–963. | PubMed | ISI | ChemPort | |
| 19. | Perrichot R, Mercier B & Quere I et al. Novel mutations in the duplicated region of PKD1 gene. Eur J Hum Genet 2000; 8: 353–359. | Article | PubMed | ISI | ChemPort | |
| 20. | OEFNER PJ & UNDERHILL PA. DNA mutation detection using denaturing high-performance liquid chromatography (DHPLC). Curr Prot Hum Genet 1998; 7: 10.11–10.12. |
| 21. | Jones AC, Austin J & Hansen N et al. Optimal temperature selection for mutation detection by denaturing HPLC and comparison to single-stranded conformation polymorphism and heteroduplex analysis. Clin Chem 1999; 45: 1133–1140. | PubMed | ISI | ChemPort | |
| 22. | Xiao W & Oefner PJ. Denaturing high-performance liquid chromatography: A review. Hum Mutat 2001; 17: 439–474 10.1002/humu.1130. | Article | PubMed | ISI | ChemPort | |
| 23. | Ravine D, Gibson RN & Walker RG et al. Evaluation of ultrasonographic diagnostic criteria for autosomal dominant polycystic kidney disease 1. Lancet 1994; 343: 824–827. | Article | PubMed | ISI | ChemPort | |
| 24. | Hayashi T, Mochizuki T & Reynolds DM et al. Characterization of the exon structure of the polycystic kidney disease 2 gene (PKD2). Genomics 1997; 44: 131–136. | Article | PubMed | ISI | ChemPort | |
| 25. | Fixman M & Freire JJ. Theory of DNA melting curves. Biopolymers 1977; 16: 2693–2704. | Article | PubMed | ISI | ChemPort | |
| 26. | Aguiari G, Savelli S & Garbo M et al. Novel splicing and missense mutations in autosomal dominant polycystic kidney disease 1 (PKD1) gene: Expression of mutated genes. Hum Mutat 2000; 16: 444–446. | Article | PubMed | ChemPort | |
| 27. | Bycroft M, Bateman A & Clarke J et al. The structure of a PKD domain from polycystin-1: Implication for polycystic kidney disease. EMBO J 1999; 18: 297–305. | Article | PubMed | ISI | ChemPort | |
| 28. | Sandford R, Sgotto B & Aparacio S et al. Comparative analysis of the polycystic kidney disease 1 (PKD1) gene reveals an integral membrane glycoprotein with multiple evolutionary conserved domains. Hum Mol Genet 1997; 6: 1483–1489. | Article | PubMed | ISI | ChemPort | |
| 29. | Bateman A & Sandford R. The PLAT domain: A new piece in the PKD1 puzzle. Curr Biol 1999; 9: R588–R590. | Article | PubMed | ISI | ChemPort | |
| 30. | Le Marechal C, Audrezet MP & Quere I et al. Complete and rapid scanning of the cystic fibrosis transmembrane conductance regulator (CFTR) gene by denaturing high-performance liquid chromatography (D-HPLC): Major implications for genetic counselling. Hum Genet 2001; 108: 290–298. | PubMed | ISI | ChemPort | |
| 31. | Choy YS, Dabora SL & Hall F et al. Superiority of denaturing high performance liquid chromatography over single-stranded conformation and conformation-sensitive gel electrophoresis for mutation detection in TSC2. Ann Hum Genet 1999; 63: 383–391. | Article | PubMed | ISI | |
| 32. | Jones AC, Sampson JR & Hoogendoorn B et al. Application and evaluation of denaturing HPLC for molecular genetic analysis in tuberous sclerosis. Hum Genet 2000; 106: 663–668. | Article | PubMed | ISI | ChemPort | |
| 33. | Klein B, Weirich G & Brauch H. DHPLC-based germline mutation screening in the analysis of the VHL tumor suppressor gene: Usefulness and limitations. Hum Genet 2001; 108: 376–384. | Article | PubMed | ISI | ChemPort | |
| 34. | Sheffield VC, Beck JS & Kwitek AE et al. The sensitivity of single-strand conformation polymorphism analysis for the detection of single base substitutions. Genomics 1993; 16: 325–332 10.1006/geno.1993.1193. | Article | PubMed | ISI | ChemPort | |
| 35. | Perrichot RA, Mercier B & Simon PM et al. DGGE screening of PKD1 gene reveals novel mutations in a large cohort of 146 unrelated patients. Hum Genet 1999; 105: 231–239 10.1007/s004390051094. | Article | PubMed | ISI | ChemPort | |
| 36. | Peral B, San Millán JL & Ong ACM et al. Screening the 3' region of the polycystic kidney disease 1 (PKD1) gene reveals six novel mutations. Am J Human Genet 1996; 58: 86–96. | ISI | ChemPort | |
| 37. | Van Raay TJ, Burn TC & Connors TD et al. A 2.5 kb polypyrimidine tract in PKD1 gene contains at least 23 H-DNA-forming sequences. Microb Comp Genomics 1996; 1: 317–327. | PubMed | ChemPort | |
| 38. | Cooper DN, Krawczak M & Antonorakis SE. The nature and mechanisms of human gene mutation,. inMetabolic and Molecular Bases of Inherited Disease 1995; (7th ed), edited by Scriver C, Beaudet AL, Sly WS, Valle D New York, McGraw-Hill pp 259–291. |
| 39. | Martin P, Heiskari N & Zhou J et al. High mutation detection rate in the COL4A5 collagen gene in suspected Alport syndrome using PCR and direct DNA sequencing. J Am Soc Nephrol 1998; 9: 2291–2301. | PubMed | ISI | ChemPort | |
| 40. | Qian F, Watnick TJ & Onuchic LF et al. The molecular basis of focal cyst formation in human autosomal dominant polycystic kidney disease type 1. Cell 1996; 87: 979–987. | Article | PubMed | ISI | ChemPort | |
| 41. | Rossetti S, Bresin E & Restagno G et al. Autosomal dominant polycystic kidney disease (ADPKD) in an Italian family carrying a novel nonsense mutation and two missense changes in exon 44 and 45 of the PKD1 gene. Am J Med Genet 1996; 65: 155–159. | Article | PubMed | ISI | ChemPort | |
| 42. | Badenas C, Torra R & San Millán JL et al. Mutational analysis within the 3' region of the PKD1 gene. Kidney Int 1999; 55: 1225–1233. | Article | PubMed | ISI | ChemPort | |
Acknowledgments
This work was supported by the National Institutes of Health (RO1 DK58816), the PKD Foundation (S.R. is a PKDF fellow) and the Mayo Foundation. The work was presented at the American Society of Nephrology meeting, 2001, and is published in abstract form. We thank the patients and their families for taking part in this study; Dr. Peters for supplying some control samples for PKD2, and Drs. F. Couch and W. Liu for early assistance with DHPLC and access to their WAVE equipment.
