|Home | About | Journals | Submit | Contact Us | Français|
Peroxisome biogenesis disorders (PBD) are a heterogeneous group of autosomal recessive neurodegenerative disorders that affect multiple organ systems. Approximately 80% of PBD patients are classified in the Zellweger syndrome spectrum (PBD-ZSS). Mutations in the PEX1, PEX6, PEX10, PEX12, or PEX26 genes are found in approximately 90% of PBD-ZSS patients. Here, we sequenced the coding regions and splice junctions of these five genes in 58 PBD-ZSS cases previously subjected to targeted sequencing of a limited number of PEX gene exons. In our cohort, 71 unique sequence variants were identified, including 18 novel mutations predicted to disrupt protein function and 2 novel silent variants. We identified 4 patients who had two deleterious mutations in one PEX gene and a third deleterious mutation in a second PEX gene. For two such patients, we conducted cell fusion complementation analyses to identify the defective gene responsible for aberrant peroxisome assembly. Overall, we provide empirical data to estimate the relative fraction of disease-causing alleles that occur in the coding and splice junction sequences of these five PEX genes and the frequency of cases where mutations occur in multiple PEX genes. This information is beneficial for efforts aimed at establishing rapid and sensitive clinical diagnostics for PBD-ZSS patients and interpreting the results from these genetic tests.
Peroxisomes are organelles present in almost all eukaryotic cells (Purdue and Lazarow, 2001; Schluter, et al., 2007; Wanders and Waterham, 2006). In humans, they are essential for the metabolism of branched chain and very long chain fatty acids, ether lipids, polyamines, amino acids, and glyoxylate ( Schluter, et al., 2007; Steinberg, et al., 2006; Wanders and Waterham, 2006). During some of these metabolic processes, they generate and subsequently inactivate reactive oxygen species (Schrader and Fahimi, 2006).
Based on genetic, bioinformatic, and proteomic analyses, it has been estimated that at least eighty-five proteins are associated with peroxisome structure and function in humans ( Schluter, et al., 2007). Peroxisome matrix proteins are synthesized on free ribosomes in the cytosol prior to import into the peroxisome. Peroxins, encoded by a family of sixteen human PEX genes (Steinberg, et al., 2004), are involved in peroxisome biogenesis with functions ranging from membrane synthesis and matrix protein import to organelle division.
Peroxisomal biogenesis disorders (PBD; MIM# 601539) are inherited in an autosomal recessive manner and are characterized by impaired peroxisome assembly (Wanders and Waterham, 2005; Weller, et al., 2003). Due to their heterogeneity, PBDs had been divided into four groups: Zellweger syndrome (ZS; MIM# 214100), neonatal adrenoleukodystrophy (NALD; MIM# 202370), infantile Refsum disease (IRD; MIM# 266510), and rhizomelic chondrodysplasia punctata (RCDP; MIM# 215100) Type 1. ZS, NALD, and IRD comprise the Zellweger syndrome spectrum (PBD-ZSS). Approximately 80% of PBD patients fall within the Zellweger syndrome spectrum, which has an overall incidence of 1:50,000 births in the United States and 1:500,000 births in Japan (Krause, et al., 2006; Shimozawa, et al., 2003; Steinberg, et al., 2004). PBD-ZSS patients with severe disorders generally die within the first year of life, while those with milder disorders can survive into young adulthood, but typically exhibit multiple organ dysfunction, sensorineural hearing loss, pigmentary retinal degeneration, and psychomotor impairments. This reflects the crucial roles peroxisomes play in neuronal migration, proliferation, differentiation and survival (Faust, et al., 2005; Krysko, et al., 2007; Powers and Moser, 1998).
It has been estimated that over 90% of PBD-ZSS patients have mutations in the PEX1 (MIM# 602136), PEX6 (MIM# 601498), PEX10 (MIM# 602859), PEX12 (MIM# 601758) or PEX26 genes (MIM# 608666) (Steinberg, et al., 2004). PEX1 and PEX6 mutations account for ~70% and ~10% of PBD-ZSS patients, respectively (Steinberg, et al., 2004). PEX1 and PEX6 belong to the AAA (ATPases Associated with diverse cellular Activities) ATPase family of proteins and form heterodimers associated with the cytoplasmic side of peroxisome membranes. These heterodimers are involved in the recycling of PEX5, the receptor for proteins with peroxisomal targeting signals (PTS) (Tamura, et al., 2006; Thoms and Erdmann, 2006). In contrast, PEX10, PEX12, and PEX26 mutations are less frequently observed in PBD-ZSS patients (Steinberg, et al., 2004). PEX26 is a peroxisomal membrane protein that assists in the localization of PEX1-PEX6 complexes to the peroxisomal membrane (Furuki, et al., 2006; Tamura, et al., 2006). PEX10 and PEX12 comprise part of a RING-finger complex of peroxisomal membrane proteins that function downstream of the initial association of cargo-laden PEX5 with its peroxisomal docking complex (Chang, et al., 1999; Okumoto, et al., 2000).
Here, we surveyed the mutational spectrum of the coding regions and splice junctions of the PEX1, PEX6, PEX10, PEX12, and PEX26 genes in a cohort of fifty-eight PBD-ZSS patients. Overall, we identified eighteen novel sequence variants predicted to disrupt protein function and two novel sequence variants predicted to be neutral. Based on these results, we estimate the relative fraction of point mutations in the coding and splice junction sequences of these five PEX genes in PBD-ZSS patients. This information is valuable for predicting the sensitivity and specificity of existing and future genetic tests for the diagnosis of PBD-ZSS.
Genomic DNAs extracted from cultured PBD-ZSS patient fibroblasts were obtained from the Kennedy Krieger Institute Peroxisomal Diseases Laboratory. Whole genome amplification of these samples was conducted using the Repli-G Mini Kit (Qiagen) and the manufacturer’s recommended protocols.
Primers flanking PEX gene coding exons of interest were designed using the Primer3 program (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) (Supp. Table S1). The GenBank reference numbers for each PEX gene were as follows: PEX1 (NM_000466.2), PEX6 (NM_000287.2), PEX10 (NM_002617.3), PEX12 (NM_000286.1), and PEX26 (AB089678.1). To simplify sequencing procedures, we appended T3 (5′-ATTAACCCTCACTAAAGGGA-3′) and T7 (5′-TAATACGACTCACTATAGGGA-3′) promoter sequences to the 5′-end of forward and reverse primers, respectively. PCR reactions were set up in 25 μL containing 60 ng of template DNA, 20 μM of each primer, 2.5 μL of 10× PCR buffer, 2.5 o μL of 25 mM MgCl2, 1.25 μL of 10 mM dNTPs and 2 units of AmpliTaq Gold polymerase (Applied Biosystems). Thirty-five cycles (94 °C for 30 seconds, 60–68 °C for 30 seconds, and 72 °C for 90 seconds) of PCR were performed followed by a final extension at 72 °C for 10 minutes. DNA-free controls were run in parallel to monitor for template contamination. All template-free control reactions and randomly selected reactions with templates were analyzed on agarose gels to verify the presence of appropriately sized products. When no product was observed, 2% or 4% of DMSO was added to replicate reactions to increase product yields.
All amplicons were treated with 10 units of exonuclease (Amersham) and 1 unit of shrimp alkaline phosphatase (Amersham) according to manufacturer’s recommended protocols in order to remove excess primer and dNTPs. Treated PCR products were sequenced using forward T3 (5′-ATTAACCCTCACTAAAGGGA-3′) and reverse T7 (5′-TAATACGACTCACTATAGGGA-3′) primers and the DNA sequencing kit with BigDye Terminator V3.0 from Applied Biosystems, according to manufacturer’s recommendations. However, the BigDye Terminator was diluted 1:4 with halfBD (Genetix Ltd).
Sequencing reactions were evaluated on the ABI3700 DNA Sequencer (Applied Biosystems, Inc.) at the DNA Core Facility of the University of Southern California Norris Comprehensive Cancer Center. Sequence variants were detected using Sequencher software V4.6 (Gene Codes Corp.). In cases where only a single deleterious PEX allele was detected, the sequencing chromatograms of the PEX gene with the single deleterious allele was inspected by eye to maximize the possibility of detecting the second allele. Nucleotide numbering reflects cDNA numbering with +1 corresponding to the A of the ATG translation initiation codon in the reference sequence, according to journal guidelines (www.hgvs.org/mutnomen). The initiation codon is codon 1.
In instances where insertions and deletions were detected in the coding regions or flanking splice junctions of PEX exons and the nature of the mutation could not be unambiguously ascertained by analyzing the sequence traces, the appropriate amplicons were subcloned using the TA TOPOII Cloning Kit (Invitrogen) and individual subclones were sequenced to confirm the identity of mutations. In all cases, the recommended mutation nomenclature guidelines provided on the Human Genome Variation Society (HGVS) website (http://www.hgvs.org/mutnomen) were followed.
PCR amplification using the PEX6 exon 10a primer set yielded two amplicons that could be readily resolved from one another via agarose gel electrophoresis in 9/60 (15%) samples. In order to avoid flanking intronic sequences not amenable to primer design, the forward and reverse exon 10a primers were designed to bind within exons 9 and exon 11 of PEX6, respectively. We sequenced the lower molecular weight band produced from the above mentioned nine samples and found that while PEX6 exon 9, 10, and 11 sequences were present, the intronic sequences flanking PEX6 exon 10 were fully deleted. Therefore, we hypothesized that this lower molecular weight amplicon was generated via priming off of the remnants of a PEX6 pseudogene located elsewhere in the genomes of these nine individuals. However, PCR analyses using additional primer sets designed to interrogate for the presence of a PEX6 pseudogene in these samples did not support this hypothesis (data not shown). Therefore, we designed a novel primer set (designated PEX6 exon 10b) designed to partially overlap the intronic regions flanking PEX6 exon 10 and found that it generated a single amplicon in the PCR analysis these same nine samples. These PCR products were sequenced and results of these analyses were included in our larger data set.
Cell fusion complementation analyses were conducted as previously described (Moser, et al., 1995). Primary fibroblasts from PBD704, PBD604, PEX1 null, PEX6 null, PEX12 null and PEX26 null patients were cultured in rich media (DMEM supplemented with vitamins, essential and nonessential amino acids, antibiotic and 10% fetal bovine serum) at 37 °C in 5% CO2. Upon reaching confluence, co-cultures of admixture of equal numbers of cells for two different lines for crosses were generated and incubated overnight at 40 °C. Afterwards, cells were washed with HBSS to remove trace culture media. Whole cell fusions were initiated by adding consecutively: (i) Solution 1 (4 mL 72% (w/v) polyethylene glycol 4000 (PEG) solution in 15% (w/v) DMSO and MEM) for 60 seconds followed by the addition of 4 mL 72% (w/v) PEG 4000 solution in HBSS. After 90 seconds, the total solution was diluted with 16 mL HBSS and incubated for two minutes prior to washing with HBSS. Afterwards, the cells were cultured in DMEM with 10% FBS and incubated overnight at 40 °C. The genotypes of mutated cell lines used in our complementation analyses were as follows: PEX1 null (c.2098insT/c.2916delA), PEX6 null (c.1314_1320del/c.2362−2A>G), PEX12 null (c.887_888delTC homozygote), and PEX26 null (c.37_38delAG/c.667+2T>C).
Fused or primary dermal fibroblasts were seeded onto glass coverslips and grown in six well plates for 72 hours at 40 °C. Cells were fixed for 20 min in 3% formaldehyde/1 × Dulbecco’s PBS solution (DPBS). After fixation, cells were washed 3 times in 1 × DPBS and permeabilized in 1% Triton X-100/1 × DPBS for 5 minutes. Subsequently, the cells were washed three times in 1 × DPBS, followed by a 30 minute blocking with 0.1% BSA in 1 × DPBS. Afterwards, the cells were incubated with primary antibodies for 60 minutes, washed seven times in 1 × DPBS, and then incubated with the appropriate secondary antibodies for 60 minutes. Rabbit polyclonal antiserum to rat PMP70 and sheep antiserum to human catalase were obtained from Sigma Aldrich and The Binding Site, respectively. Secondary antibodies were obtained from Jackson ImmunoResearch. Micrographs were captured on a Zeiss LSM510 confocal multi-photon microscope.
Previously, a cohort of 91 PBD-ZSS patients was subject to the PEX Gene Screen, an algorithm for the sequence analysis of select PEX1, PEX2, PEX5, PEX6, PEX10, PEX12, and PEX26 gene exons (Steinberg, et al., 2004). In these studies, two putative disease-causing alleles were found in fifty-five individuals and one putative disease-causing allele in seventeen individuals. No PEX gene mutations were found in the remaining nineteen individuals.
Here, we sequenced the complete coding regions and splice junctions of PEX1, PEX6, PEX10, PEX12 and PEX26 in fifty-eight of these ninety-one PBD-ZSS patients, irrespective of prior genotypes. Our cohort was comprised of twenty-four patients where two putative disease-causing alleles were known, seventeen patients where one putative disease-causing allele was known, and seventeen patients where no disease-causing alleles were previously identified. This cohort was selected for the purposes of: (i) determining the sensitivity of our sequencing analysis based on known sequence variants, (ii) identifying deleterious alleles missed by the PEX Gene Screen, and (iii) identifying common sequence variants in these five PEX genes.
All analyses were successful with the exception of PEX10 exon 3, which was removed from consideration due to the inability to consistently produce robust specific PCR products despite multiple primer designs. Overall, 71 unique sequence variations were found scattered throughout these five genes (Tables 1–2). We placed these variants into three specific categories: (i) deleterious mutations predicted to inactivate gene function via nonsense-mediated decay, premature translation termination, frameshift, splicing error, or alteration of a conserved amino acid (i. e. same in human, rhesus monkey, mouse, and rat orthologs), (ii) single nucleotide polymorphisms (SNPs), and (iii) rare neutral variants. We screened the scientific literature and dbpex (www.dbpex.org) for prior reports of each of the deleterious mutations identified. The information found in dbpex is publicly available and kindly provided by Drs. Steven Steinberg (Kennedy Krieger Institute/Johns Hopkins University School of Medicine) and Nancy Braverman (Johns Hopkins University School of Medicine/McGill University). Mutations cited as (www.dbpex.org) were identified by Drs. Steinberg and Braverman, but not reported in the scientific literature.
We categorize SNPs as being reported in the National Center for Biotechnology Information (NCBI) Single Nucleotide Polymorphism database (dbSNP: http://www.ncbi.nlm.nih.gov/projects/SNP/) in more than one individual or the scientific literature. Previously unreported silent changes were classified as rare neutral variants, even though we recognize that such variants can be deleterious by introducing cryptic splice junctions or interfering with exonic splicing enhancer (ESE) motifs (Boichard, et al., 2008). We did not find any examples of missense alleles of ambiguous functional significance that either (i) was not found in multiple individuals in dbSNP or (ii) involved a poorly conserved amino acid in the orthologous mouse and rat proteins.
Previously, 59 sequence variants (28 unique) were identified in our patient cohort through PEX Gene Screen analyses (Steinberg, et al., 2004) (Tables 1–3). Based on our blinded sequencing analyses, we uncovered all 59 (100%) of these known variants. This indicates that our sequencing analyses display excellent sensitivity and can reproducibly detect mutations in these PEX genes. Below, we discuss all 71 unique sequence variants uncovered in our current study (Table 1).
We identified twenty-nine patients with two deleterious PEX1 alleles and none with a single deleterious PEX1 allele (Table 1). The alleles c.2098insT and c.2528G>A (G843D) are strongly represented here, because these are known common alleles that provide the basis of the PEX Gene Screen to identify PEX1 deficient patients (Steinberg, et al, 2004). A total of twenty-six deleterious PEX1 alleles were uncovered. These included twelve novel deleterious mutations (7 frameshift: c.270delA p.Q91fs, c.643_647delACCAA p.T215fs, c.782_783delAA p.Q261fs, c.1714_1715delCA p.H572fs, c.1840delA p.E615fs, c.3693_3696delGTCA p.Q1231fs, c.3732_3736dupCATTA p.S1246fs; one nonsense: c.547C>T p.R183X; four splice junction: c.473−1G>A, c.1240+1G>T, c.1900+2T>C, c.3438+2T>C) and eight previously reported deleterious mutations (two nonsense: c.2614C>T p.R872X (Fukuda, et al., 1996), c.2992C>T p.R998X (FitzPatrick and Valle, 1996); three frameshift:: c.2097_2098insT p.1700fs p.S701fs (Collins and Gould, 1999; Maxwell, et al., 1999; Walter, et al., 2001), c.2537_2545del9insTCATGGT (Steinberg, et al., 2004), and c.2916delA p.G973fs (Maxwell, et al., 2002); one in-frame duplication: c.1952_1960dupCAGTGTGGA (Preuss, et al., 2002); one in-frame deletion: c.3022_3024delCCT p.P1008del (www.dbpex.org); one splice junction: c.2926+1G>A (Crane, et al., 2005; Steinberg, et al., 2004)) that result in a truncated PEX1 protein lacking the AAA1 and AAA2 functional domains.
In addition, we identified three novel deleterious missense mutations (c.1769T>G p.L590R, c.2843G>A p.R948Q, and c.3710C>A p.A1237E) affecting residues conserved in rhesus monkeys, mouse, and rat. However, the functional significance of the PEX1 p.R948Q allele in patient 604 is subject to interpretation since this individual has two deleterious PEX26 missense alleles, both p.R98W. Indeed, cell fusion complementation analyses with PEX1 and PEX26 null fibroblasts (Fig. 1A–E) indicate that cells derived from patient 604 have at least one functional PEX1 allele (Fig. 1D), but lack PEX26 function (Fig. 1E). Thus, we provide evidence that the p.R98W allele negatively impacts PEX26 function. However, we cannot rule out the possibility that PBD604 carries three inactivating PEX alleles (i.e. two PEX26 p.R98W alleles and PEX1 p.R948Q). R948 is localized in a region that is conserved in yeast, although this residue itself is only semi-conserved.
One previously identified missense mutation c.1777G>A p.G593R (Rosewich, et al., 2005; Walter, et al., 2001) associated with low PEX1 protein levels (Walter, et al., 2001) and two previously identified temperature-sensitive mutations, c.2528G>A p.G843D (Collins and Gould, 1999; Maxwell, et al., 2002; Maxwell, et al., 1999; Walter, et al., 2001) and c.2392C>G p.R798G (Crane, et al., 2005; FitzPatrick and Valle, 1996)) were detected. The c.2088A>G p.I696M allele present in two patients was reported to be either a rare variant or polymorphism in other studies (Crane, et al., 2005; Steinberg, et al., 2004) and is thus not included in Table 1.
PBD672 carries the deleterious p.G843D and p.R872X PEX1 alleles and the PEX6 p.R601Q allele. Although we do not know the phase of these PEX1 mutations, most likely they are on different chromosomes and responsible for the NALD/IRD patient diagnosis. As discussed below, we uncovered functional evidence that the PEX6 p.R601Q allele is deleterious in patient PBD704 (Fig. 1). Thus, we propose that PBD672 carries three inactivating PEX alleles. The significance of this complex genotype is uncertain since the clinical and biochemical phenotype of PBD672 is similar to PEX1 deficient patients who are compound heterozygotes for p.G843D and a nonsense or frameshift mutation.
We identified four patients with two deleterious PEX6 alleles, one patient with a single deleterious PEX6 allele, and one patient with three deleterious PEX6 alleles (Table 1). One was a novel frameshift (c.914delA p.D305fs) mutation that results in a truncated PEX6 protein lacking an intact AAA cassette domain. In addition, two previously reported splice junction mutations (c.882+1G>A (Steinberg, et al., 2004) and c.1962−1G>A (Steinberg, et al., 2004; Yahraus, et al., 1996)) were identified.
Two novel [c.1646C>T p.A549V and c.2546A>C p.N849T] and four previously reported deleterious missense mutations (c.821C>T p.P274L (Steinberg, et al., 2004), c.1802G>A p.R601Q (www.dbpex.org), c.2578C>T p.R860W (www.dbpex.org), and c.2579G>A p.R860Q (www.dbpex.org)) also were identified. All of the missense mutations affect amino acid residues conserved in rhesus macaque, mouse, rat, zebrafish and chicken orthologs. Nevertheless, one patient carried two protein-abolishing mutations (c.914delA p.D305fs and c.1962−1G>A) in addition to the p.A549V variant. Thus, the medical significance of the p.A549 allele is questionable, even though it is counted as a novel deleterious allele in Table 1 according to our nomenclature.
The functional significance of the c.2579G>A p.R860Q/c.1802G>A p.R601Q mutations in PBD704 were of particular interest given that this individual also carries the deleterious c.681−2A>C PEX12 allele. As such, we conducted cell fusion complementation analyses with PEX6 and PEX12 null fibroblasts (Fig. 1F–J) and demonstrated that PBD704 fibroblasts lack PEX6 function (Fig. 1I), but have at least one functional PEX12 allele (Fig. 1J). Given the nature of the PEX12 c.681−2A>C mutation, we propose that PBD704 carries three inactivating PEX alleles; it remains a tantalizing question as to whether this is of any biological or medical significance. However, we do not have any clinical information on PBD704 to make any genotype-phenotype assessment.
The previously reported PEX6 c.235G>C p.A79P missense allele (Steinberg, et al., 2004) was found in one patient (627) with two deleterious PEX1 alleles (both p.A1237E) and one patient (607) with two deleterious PEX12 alleles (both p.L296fs). In addition, it is conserved in rhesus macaques and mouse, but not rat. Thus, this is highly likely a neutral variant not causative of disease.
We identified two patients with two deleterious PEX10 alleles and none with a single deleterious PEX10 allele (Table 1). We uncovered four previously reported deleterious PEX10 mutations (c.4delG p.A2fs (Steinberg, et al., 2004), c.704_705insA p.L236fs (Warren, et al., 2000), c.730C>T p.R244X (Krause, et al., 2006) and c.835G>A p.E279X (Steinberg, et al., 2004)). All of those mutations result in a truncated PEX10 protein lacking the C-terminal RING finger domain (AA273-311) critical for its function (Krause, et al., 2006). No PEX10 missense alleles were found in this study.
We identified five patients with two deleterious PEX12 alleles and one patient with a single deleterious PEX12 allele (Table 1). Overall, we uncovered nine previously reported deleterious PEX12 mutations. Of those, four were frame-shift mutations: c.541_542insT p.Y181fs (Steinberg, et al., 2004), c.730_733dupGCCT p.L245fs (Chang and Gould, 1998), c.887delT p.L296fs (Steinberg, et al., 2004) and c.887_888delTC p.L296fs (Gootjes, et al., 2004b). Two in-frame deletions (c.531_533del p.Q178del and c.1047_1049del p.Q349del) were detected that affect residues conserved in rhesus, mouse and rat. The latter mutation lies in the C-terminal zinc-binding RING domain encompassing amino acids 260–359 (Chang, et al., 1999; Krause, et al., 2006; Okumoto and Fujiki, 1997). A previously reported splice site mutation (c.681−2A>C (Steinberg, et al., 2004)) was detected in PBD704, which we previously discussed as also carrying the PEX6 p.R601Q and p.R860Q alleles. PBD673 was homozygous for the c.959C>T p.S320F missense allele, which was previously reported to reduce PEX12 binding affinity to PEX5 and PEX10 (Chang, et al., 1999). Interesting, in another study, this same allele also was found in the homozygous state in a patient with a relatively mild PBD (Gootjes, et al., 2004a).
In PBD649, we detected the c.102A>T p.R34S PEX12 missense allele affecting an amino acid residue conserved in rhesus macaques, mouse, and rat, which was previously reported as a mild, temperature sensitive mutation in two probands (Zeharia, et al., 2007). We noted that this individual also carries two deleterious PEX1 alleles (pI700fs and p.H572fs), although we did not test if they reside on separate chromosomes. PBD649 had a severe early onset presentation consistent with Zellweger syndrome, thus being compatible with having two null alleles rather than a mild missense allele in conjunction with a null allele. Although PEX12 p.R34S may be associated with mild disease in isolation (hence its inclusion in Table 1), triallelic inheritance is unlikely to be an important contributor to the biochemical and clinical phenotype of PBD649.
We identified five patients with two deleterious PEX26 alleles and no patients with a single deleterious PEX26 allele (Table 1). Of the five unique deleterious PEX26 mutations detected, four (c.37_38delAG p.R13fs, c.192_216del p.S64fs, c.296G>A p.W99X, and c.574C>T p.R192X) were previously identified as protein-truncating mutations (Steinberg, et al., 2004). One (c.667+2T>C) represents a previously reported splice junction mutation (Steinberg, et al., 2004). We also identified two reported missense mutations (c.292C>T p.R98W (Matsumoto, et al., 2003) and c.353C>G p.P118R (Steinberg, et al., 2004)). Both could disrupt a PEX6 binding domain present in residues 29–163 of PEX26 protein (Weller, et al., 2003). The c.292C>T p.R98W PEX26 missense allele has been associated with a temperature-sensitive peroxisomal matrix protein import defect (Furuki, et al., 2006). Furthermore, homozygosity for PEX26 p.R98W has been reported in association with neonatal adrenoleukodystrophy (Matsumoto, et al., 2003; Weller, et al., 2005) and PBD604 has the same clinical diagnosis.
We detected a total of sixteen unique SNPs previously reported in dbSNP and/or the scientific literature and two unique neutral variants in the PEX1, PEX6, PEX10, PEX12, and PEX26 genes (Table 2). Twelve variants of this general type were found multiple times in our cohort. In contrast, six such variants (PEX1 c.3762T>C p.A1254A; PEX6 c.2770G>T p.A924S; PEX6 c.210G>A p.G70G; PEX6 c.271G>T p.P271A; PEX12 c.733T>A p.L245I and PEX12 c.867C>T p.D289D) were found only once in our cohort.
Here, we identified at least one deleterious PEX allele in eleven of the seventeen individuals where the PEX Gene Screen uncovered no PEX mutation. This led us to combine the results from our current analyses with that of the PEX Gene Screen. Based on these combined analyses, we demonstrate that the disease-causing gene could be assigned in 91.2% of the ninety-one PBD-ZSS cases (Table 3). In keeping with reports from multiple laboratories, the bulk of the cases could be explained by PEX1 mutations (58.2%). The remaining 33.0% of cases where at least one deleterious PEX allele was identified were distributed amongst PEX6, PEX10, PEX12, PEX26, PEX2, and PEX5 mutations (each gene ranging from 1.1 – 9.9% of cases explained).
There are several possible explanations for our inability to uncover all predicted PEX gene defects. Most likely, mutations may lie outside of the exon and splice junction sequenced, either elsewhere in these five PEX genes or, alternatively, in other loci. In addition, sequence variation could be missed due to the sensitivity of our sequencing analysis. As in any PCR-based method, the loss of primer binding sites could prevent the amplification of mutant alleles. In a similar manner, genomic deletions that eliminate large segments of any of these five genes would not be detected using our non-quantitative PCR-based methodology.
Overall, this study supports the design of the PEX Gene Screen. Although a limited number of coding exons in the PEX26 (two of five), PEX12 (two of three), and PEX10 (three of six) were sequenced in the PEX Gene Screen analysis of 91 patients (Steinberg, et al., 2004), we uncovered no deleterious mutations in the untested exons of PEX Gene Screen mutation-negative patients. Although it is likely that a select number of PBD patients have mutations localized solely to these exons untested by the PEX Gene Screen, our observations suggest that adding these exons to the PEX Gene Screen would not benefit test sensitivity significantly.
The identification of four patients (PBD649, PBD704, PBD604, and PBD672) with two predicted deleterious mutations in one PEX gene and a single deleterious mutation in a second PEX gene highlights potential difficulties in the interpretation of genetic tests. It is highly likely (but not certain) that the gene containing two deleterious alleles is responsible for disease. However, it is possible that a third allele on a second PEX gene further reduces the efficiency of peroxisome assembly and thus contributes to the clinical phenotype in some patients. Thus far, we have not identified a case consistent with the triallelic inheritance reported in other disorders, such as Bardet-Biedl syndrome (Badano, et al., 2006). As a corollary, it is not formally correct to assign the disease gene in the two patients (PBD650 and PBD720) for which one deleterious allele was found. Nevertheless, given the observed frequency of PBD-ZSS patients with mutations in multiple PEX genes, we anticipate that the genes in these two patients for which we found only one mutation by sequencing are responsible for disease.
The complex mutational spectrum in some PBD-ZSS patients makes this an important testing ground for newly developed resequencing technologies. These including oligonucleotide microarrays (Greiner, et al., 2006; Hacia, 1999; Hacia, et al., 1996; Hacia, et al., 2000; Karaman, et al., 2005) and next generation sequencing methods (Albert, et al., 2007; Dahl, et al., 2007; Hodges, et al., 2007; Porreca, et al., 2007). The wealth of functional biochemical and genetic assays for peroxisome activity in cell culture model systems as well as blood lipid profiling (Gootjes, et al., 2002; Steinberg, et al., 2008) will play a major role in deciphering the results of future large-scale resequencing analyses.
We thank Dr. Paul Watkins at the Kennedy Krieger Institute and the Johns Hopkins University School of Medicine, Dr. Nancy Braverman at the Johns Hopkins University School of Medicine and McGill University, and Drs. Patricia Dranchak and Krishna Ramaswamy at the University of Southern California for expert advice and discussion. This investigation was conducted in a facility constructed with support from Research Facilities Improvement Program Grant Number C06 (RR10600-01, CA62528-01, RR14514-01) from the National Center for Research Resources, National Institutes of Health.
Contract grant sponsor: National Institutes of Health; Contract grant number: R21 HD050277 and R01 GM072477.