We isolated peroxisome biogenesis mutants ZP128 and ZP150 from rat PEX2-transformed Chinese hamster ovary (CHO) cells, by the 9-(1′-pyrene)nonanol/ ultraviolet method. The mutants lacked morphologically recognizable peroxisomes and showed a typical peroxisome assembly-defective phenotype such as a high sensitivity to 12-(1′-pyrene)dodecanoic acid/UV treatment. By means of PEX cDNA transfection and cell fusion, ZP128 and ZP150 were found to belong to a recently identified complementation group H. Expression ofhuman PEX13cDNA restored peroxisome assembly in ZP128 and ZP150. CHO cell PEX13 was isolated; its deduced sequence comprises 405 amino acids with 93% identity to human Pex13p. Mutation in PEX13 of mutant ZP150 was determined by RT-PCR: G to A transition resulted in one amino acid substitution, Ser319Asn, in one allele andtruncationofa42aminoacidsequencefrom Asp265 to Lys306 in another allele. Therefore, ZP128 and ZP150 are CHO cell lines with a phenotype of impaired PEX13.
Peroxisomes are formed by division of pre-existing peroxisomes after post-translational import of newly synthesized proteins (1). Recent evidence obtained by studies using yeast mutants suggests the involvement of the endoplasmic reticulum (ER) in peroxisomal membrane biogenesis (2). Cis-acting peroxisomal targeting signals (PTSs) have been identified: a C-terminal SKL tripeptide motif (PTS1) and anN-terminal cleavable presequence (PTS2) of several proteins such as 3-ketoacyl-CoA thiolase (3). Genetic analysis of peroxisome-deficient mutants of yeast and mammalian cells has led to the identification of a number of protein factors, peroxins, essential for peroxisome biogenesis (3,4). The functional significance of human peroxisomes is highlighted by fatal genetic diseases linked to peroxisomal malfunction and failure of peroxisome biogenesis (3).
We previously isolated and characterized eight, mutually distinct Chinese hamster ovary (CHO) cell mutants, such as ZP65 (5) and ZP119 (6), and recently have isolated ZP126 (7). More than 15 genes are likely to be involved in mammalian peroxisome assembly, as deduced from complementation group (CG) analyses of CHO cell mutants and fibroblasts from patients with various peroxisome biogenesis disorders (PBD) (3,6,7). We isolated several PEX cDNAs, including PEX1, PEX2 (formerly PAF-1), PEX12 and PEX19, by genetic phenotype complementation assay of such CHO cell mutants, and further demonstrated that these PEX genes are responsible for the primary defects in PBD patients (8–12). Thus, peroxisome biogenesis-defective CHO cell mutants are an excellent mammalian somatic cell system for investigation of peroxisome assembly at the molecular and cellular levels, as well as for delineation of the genetic basis of PBD (3).
In this study, we isolated two novel peroxisome-deficient CHO cell mutants, ZP128 and ZP150, both belonging to CG-H (13). Transfection of PEX13 cDNA restored peroxisome biogenesis in ZP128, indicating that ZP128 and ZP150 are CHO cell lines with a phenotype of impaired PEX13. Mutation analysis of PEX13 in ZP150 is also reported.
Isolation and Complementation Group Analysis of CHO Cell Mutants
By using the 9-(1′-pyrene)nonanol/ultraviolet (P9OH/UV) selection method, we isolated 10 CHO cell mutant clones showing catalase in a diffuse staining pattern in the cytosol, as seen in ZP128 and ZP150 (Fig. 1A, a and e), indicative of a defect in peroxisome biogenesis (6,14) (Table 1). One mutant, ZP130, contained catalase-positive but morphologically aberrant, tube-like structures (data not shown), as in several earlier mutants (6,15). Mutant ZP128 apparently in a novel CG (see below) was stained with antibodies to PTS1 and 3-ketoacyl-CoA thiolase, a PTS2 protein (16). PTS1 proteins were visualized in punctate structures, presumably peroxisomes, in nearly half of the cells (Fig. 1A, b), while thiolase was present in the cytosol in most cells (Fig. 1A, c). Thus, import of PTS1 proteins appears to be mildly affected and PTS2 transport is more severely impaired in ZP128. Temperature-sensitive restoration of the impaired peroxisome biogenesis (17), revealed in fibroblasts from a CG-H patient with neonatal adrenoleukodystrophy (13), was not observed in ZP128 and ZP150 (data not shown). Peroxisomal remnants were seen in ZP128 (Fig. 1A, d), as in earlier CHO cell mutants (3,18) and fibroblasts fromPBD patients (19,20). ZP150 in the same CG as ZP128 (see below) also showed such a morphological phenotype.
To determine the CGs of 10 CHO mutants, each cell clone was transfected with PEX2, PEX5, PEX6 and PEX12 cDNAs (Table 1). ZP149, but not others including ZP128, was restored in catalase import by PEX12 (Fig. 1A, f and g), thus indicating that ZP149 belonged to CG-III (11). Three mutants were found to be in a PEX2-deficient CG (9), implying that introduced PEX2 cDNA was probably disintegrated during the initial cultivation of TKa (6); one mutant was in PEX5-defective CG (21). None of the mutants was in a PEX6-defective CG (22).
Accordingly, it is apparent that CHO mutants ZP128, ZP129, ZP145, ZP146 and ZP150 belong to CG(s) different from these four CGs. Next, these five unclassified cell mutants were examined by cell fusion. ZP145 and ZP146 were revealed to be in CG-E (CG-I in the USA). ZP129 was in the same CG as recently isolated ZP124 (7). ZP128 and ZP150 were both complemented by cell fusion with all cell mutants examined (Fig. 1A, i and j) (Table 1), except for a cell hybrid between ZP128 and ZP150 (Fig. 1A, h). Taken together, we conclude thatZP128 andZP150 are in the same CG but distinct from any of the previously identified CGs of CHO mutants. Moreover, fused cells of ZP128 and ZP150 each with the wild-type CHO-K1 showed catalase-positive particles (Fig. 1A, k and l), indicating that lesion(s) of allele(s) in the mutants are recessive, as in earlier cell mutants (3). Furthermore, no evidence of reestablishment of peroxisomes was obtained by cell fusion of ZP128 with fibroblasts from a CG-H patient (13) (Fig. 1B, a). In contrast, numerous peroxisomes were evident in the respective hybrid cells of ZP128 with CG-G (23) (Fig. 1B, b) and CG-VT (data not shown) for which no complementing genes are yet elucidated. Accordingly, these results strongly suggested that ZP128 and ZP150 are classified in CG-H.
PEX13 Restored Peroxisome Biogenesis in ZP128
Genetic functional complementation assay (8) of ZP128 using a human liver cDNA library led us to isolate first a positive clone containing an open reading frame (ORF) apparently encoding a 39 amino acid shorter Pex13p (24) (Fig. 2A, a), and then a full-length human (Hs) PEX13 (13,25) (Fig. 2A, b) at a later stage of the present work. The HsPEX13 complemented peroxisomal import of catalase inbothZP128 (see below) and ZP150 (data not shown). Several phenotypic abnormalities due to the peroxisome deficiency were found in ZP128 (see below), as in earlier CHO mutants (5,6,14,18). A stable HsPEX13 transformant of ZP128, named 128P13, showed peroxisomal localization of catalase, PTS1 proteins and PTS2 3-ketoacyl-CoA thiolase (data not shown). Restoration of catalase import was also verified by the digitonin titration method (5). About 70% of the catalase activity was latent at 100 mg/ml of digitonin in 128P13, as in CHO-K1, but ZP128 showed no latency (Fig. 2B, a). HsPEX13 expression also complemented peroxisome assembly in ZP150, but not in eight other CHO mutants and in five CGs of PBD patient fibroblasts (Table 2). Moreover, expression of Chinese hamster PEX13 (see below) restored peroxisome biogenesis in ZP150 (Fig. 2A, c). These data indicate that Pex13p is the peroxisome biogenesis factor for ZP128 and ZP150 of CG-H.
Abnormal biogenesis of peroxisomal enzymes was also complemented in 128P13 cells. All three polypeptide components of acyl-CoA oxidase (AOx) were evident in both 128P13 and CHO-K1 cells, as assessed by immunoblot analysis (Fig. 2B, b, lanes 1 and 3), whereas only the A component was seen in ZP128 (lane 2). PTS2-thiolase precursor was processed properly to the 41 kDa matured form in peroxisomes in 128P13, as in CHO-K1, but the 44 kDa precursor was stably present in ZP128 (Fig. 2B, b, lanes 4–6). As another typical phenotype, peroxisome-deficient mutants are killed specifically by the 12-(1′-pyrene)dodecanoic acid (P12)/UV selection, as there is no synthesis of plasmalogens (14,26), whereas the P9OH/UV selection kills wild-type cells which incorporate this fatty alcohol analog into plasmalogen molecules (14,27). 128P13 cells showed P12/UV-resistant and P9OH/UV-sensitive phenotypes, as was the case for CHO-K1 cells (Table 3). Conversely, ZP128 was resistant to the P9OH/UV and sensitive to the P12/UV treatment.
Taken together, these results indicate that PEX13 can fully complement the abnormalities in ZP128.
Dysfunction of Pex13p in CHO Mutants
Isolation of Chinese hamster PEX13. We isolated Chinese hamster (Cl) PEX13 by screening a CHO-K1 cDNA library with HsPEX13 as a probe. ClPEX13 was functionally active in complementing ZP128 (data not shown) and ZP150 (Fig. 2A, c) and encoded the 405 amino acid Pex13p (Fig. 3A). ClPex13p was longer by two amino acids than HsPex13p, with 93% amino acid identity, containing two hydrophobic, putative membrane-spanning regions (Fig. 3A, solid underlines) and a Src homology 3 (SH3) domain (28) (dashed underline), as found in yeast and human Pex13p (24,25,29,30). Two proline-rich regions (28) were located in the N-terminal part (Fig. 3, solid overlines), which were not evident in P.pastoris Pex13p (24).
PEX13 in mutants. On northern blotting, PEX13 mRNA was detected as an ∼1.9 kb band in CHO-K1 and ZP150, but not in ZP128, implying alteration at a transcriptional level or rather rapid degradation of PEX13 mRNA in ZP128 (data not shown). To investigate dysfunction of Pex13p in ZP150, we isolated PEX13 from ZP150 by RT-PCR, where two different sizes of PCR product were obtained (Fig. 3B, a). Subsequent sequencing of the respective cDNA clones indicated a compound heterozygous mutation: in one allele, a mutation of nucleotide G to A at position 956 (the A of the initiating ATG being 1), in a codon for Ser319, resulting in a codon for Asn319 (Fig. 3B, b; see Fig. 3A, solid arrowhead), named ClPEX13S319N; in another allele, truncation of 126 nucleotides from residues 793–918 or 792–917, named ClPEX13e/s, resulted in a 42 amino acid deletion from Asp265 to Lys306, presumably because of a defective splicing (Fig. 3B, c; see Fig. 3A, box). Thus, dysfunction of the Pex13p caused by a missense point mutation and/or truncation is most likely the primary defect in ZP150. Little RT-PCR product was obtained from ZP128-derived mRNA (Fig. 3B, a), consistent with the finding on northern blot analysis.
Complementation of protein import by ClPEX13. Expression of ClPEX13 complemented peroxisomal protein import of catalase in ZP128 cells (Fig. 3C, a), as was the case by HsPEX13 (Fig. 2A). To assess the impaired function of Pex13p, ZP150-derived ClPEX13S319N in pcDNA3.1/Zeo/TK vector, intentionally at a rather lower level of expression (see below), and ClPEX13e/s in pUcD2Hyg were transfected separately into ZP128. Catalase was present in the cytosol in most of the transfected cells (Fig. 3C, b), while a small number of particles were partly found, presumably due to a higher level of ClPEX13S319N expression as compared with that of ZP150 per se. This finding was more evident as assessed by using ClPEX13S319N in pUcD2Hyg with a more potent SRα promoter (data not shown). A similar result was likewise obtained when ZP150 was back-transfected with ClPEX13S319N (data not shown). ClPEX13e/s-transfected ZP128 cells showed catalase in the cytosol (Fig. 3C, c), as in ZP128 (see Fig. 1A, a), suggesting that ClPEX13e/s protein was biologically inactive. Taking these results together, we conclude that dysfunction of Pex13p caused by a S319N missense mutation and a 42 amino acid truncation is the primary defect in impaired peroxisome biogenesis in ZP150.
Characterization of PEX13 Protein
Localization of Pex13p in peroxisomes was confirmed by several lines of evidence. Upon sucrose density gradient centrifugation of the rat liver light mitochondrial fraction, Pex13p was co-sedi-mented with peroxisomal marker proteins, including catalase, thiolase and a peroxisomal integral membrane protein, Pex12p (11) (data not shown). Rat peroxisomal Pex13p was resistant to the extraction with sodium carbonate (31) and was recovered in the detergent phase by Triton X-114 treatment (32), thereby indicating that Pex13p is an integral membrane protein (data not shown). N-terminally flag-tagged human Pex13p was co-localized with catalase when expressed in CHO-K1 cells and detected by immunofluorescent microscopy (Fig. 4a and b). After permeabilization with 25 µg/ml of digitonin, under which conditions plasma membranes are permeabilized selectively and intra-peroxisomal proteins are inaccessible to exogenous antibodies (11), flag-Pex13p, but not catalase, was discernible in a punctate staining pattern using anti-flag antibody (Fig. 4c and d). Therefore, the N-terminal part of Pex13p is exposed to the cytosol. A cytosolic orientation of the C-terminal part (24) was confirmed with antibody to the C-terminal 19 amino acid peptide of human Pex13p (data not shown). Moreover, endogenous Pex13p was also detectable and superimposable with the location of catalase in CHO-K1 (data not shown).
After several cycles of the conventional mutant screening procedure, we isolated peroxisome-deficient CHO cell mutants ZP128 and ZP150 belonging to the PEX13-defective CG. ZP128 and ZP150 showed common properties such as absence from morphologically recognizable peroxisomes, no latency of catalase and high sensitivity to P12/UV treatment, despite normal synthesis of peroxisomal proteins, thereby representing typical somatic mammalian cell mutants defective in peroxisome biogenesis. Intriguingly, however, PTS1 proteins were partly imported to peroxisomal structures in ZP128 and ZP150. Fibroblasts from a CG-H patient also showed a similar phenotype (13).
It is noteworthy that the 364 amino acidPex13p identified by Gould et al. (24) was exactly the same ORF that we isolated initially by functional complementation of ZP128, indicating that the translation product of PEX13 starting at the second ATG for Met40 was biologically active. Such truncated Pex13pisunlikelytobepresentin vivo, because the size of HsPex13p synthesized in vitro by coupled transcription-translation was indistinguishable from that of a protein detected by immunoblot of purified peroxisomes (data not shown).
Pex13p may function as a component of the matrix protein import machinery, e.g. PTS1 receptor (Pex5p) docking protein (24). In ZP150, Pex13p is mildly affected by a missense mutation at S319N, while a 42 amino acid deletion eliminates the biological activity of Pex13p, as assessed by back-transfection of each mutant ClPEX13 to ZP128. Loss of Pex13p in yeast eliminated the import of PTS1 and PTS2 proteins (24,29,30). These findings imply that the severity of the phenotype caused by the affected Pex13p in the mutants depends on the type of mutations. Interestingly, apparent import of PTS1 proteins, possibly PTS2 proteins as well, is noted in some ZP128 and ZP150 cells. Such a different phenotype within a single mutant cell clone may reflect temporally regulated binding of Pex13p to its ligands, including the Pex5p-PTS1 protein complex, at various stages of peroxisome biogenesis or cell division, etc. Moreover, failure in conversion of AOx-A to B and C components (Fig. 2B), even in PTS1-positive particles visible in ZP128 and ZP150, may reflect the dysfunction or impaired import of a putative processing protease. In ZP128 and ZP150, import of catalase is completely impaired. Catalase may be transported by a pathway involving Pex13p but presumably distinct from the one, if any, used for PTS1 import, as deduced from the finding that human catalase possesses a PTS1-like but distinct signal (33). Both alleles of the PEX13 gene appear to be expressed in ZP150, at least at the mRNA level, although the type of gene expression in CHO cells has been postulated to be hemizygotic. The level of PEX13 mRNA derived from each allele as well as of Pex13p with S319N and/or a truncation from Asp265 to Lys306 remain to be determined. It is of interest to note that a similar type of expression of both alleles was also noted in the PEX14 gene in CHO cell mutants (34).
Given the fact that Pex13p is inactivated by mutations, one at S319N in the SH3 domain and deletion of 42 amino acids including nearly half of the SH3 region, the SH3 domain is probably important for the biological activity of Pex13p, presumably involved in the interaction with the proline-rich region ofproteins (28), including that of Pex13p itself. The SH3 domain inPex13p was shown to interact with Pex5p (24,29,30). It is of interest to note, however, that recently identified human Pex13p appears to interact little with Pex5p in an overlay ligand-binding assay (25). It is also possible that the N-terminal part, shown in the present study to be exposed to the cytosol, plays an important role in such interaction. To address such issues, the structure and function relationship of Pex13p in mammalian cells should be delineated fully at the molecular level.
Materials and Methods
Isolation of Peroxisome-Deficient CHO Cell Mutants and CG Analysis
TKa cells, wild-type CHO-K1 cells transformed with rat PEX2 (9,35), were mutagenized and used for isolation of mutant cells defective in peroxisome biogenesis by the P9OH/UV method (27), as described (14,35). Peroxisomes in CHO cells and human fibroblasts were visualized by indirect immunofluorescence light microscopy, as described (14,35), using rabbit antibodies to rat liver catalase (5), human catalase (14), PTS1 peptide (21), rat thiolase (5) and 70 kDa peroxisomal integral membrane protein (5,6). Antigen-antibody complex was detected using fluorescein isothiocyanate-labeled sheep anti-rabbit immunoglobulin G antibody (Cappel). CG analysis was done by cDNA transfection (6) and cell fusion (14), as described.
Stable Transfection of PEXI3 and Expression of Flag-Tagged Pex13p
An expression plasmid pUcD2Hyg·HsPEX13 was used to establish an HsPEX13-stable transformant 128P13, by selection with 200 µg/ml of hygromycin B (Sigma). An epitope was flag-tagged to the N-terminus of Pex13p by a PCR-based technique. HsPEX13 (nucleotide residues 4–506) was amplified using a forward primer HsPEX13.FLAG.F (5′-CGGTCGACATGGATTATAAAGATGATGATGATAAAG CGTCCCAGCCGCCACCTCCC-3′) and a reverse primer R6 (5′-GAAAAGTGATTTGCTACATCCAAT-3′). The PCR products were ligated into the SalI-NsiI site of pBlueScript SK(−)·HsPEX13, from which the SalI and SacI digest was then ligated into the SalI-KpnI site of pUcD2Hyg vector using SacI linker. Flag-Pex13p was detected using mouse monoclonal antibody to flag (Sigma) and Texas Red-labeled sheep anti-mouse IgG second antibody (Amersham Pharmacia Biotech).
Screening of a Chinese Hamster cDNA Library and Mutation Analysis
A 32P-labeled, 0.57 kb PstI-SphI fragmentof HsPEX13 was used as a probe to screen 2.0 × 105 independent clones of a CHO-K1 cDNA library (21). One of two positive clones, ClPEX13, was digested with SalI and NotI and ligated into the SalI-NotI site of pCMVSPORT and a lower level expression vector pcDNA3.1/Zeo/TK that had been constructed by using pcDNA3.1/Zeo (Invitrogen) and thymidine kinase promoter derived from the pTK-Hyg vector (Clontech). Total RNA was obtained from ZP128 and ZP150 cells, as described (11). Reverse transcription was performed using total RNA (2 mg), oligo(dT) primer (0.5 µg) and Superscript II reverse transcriptase (Gibco BRL), in 50 µl. To amplify the full length of the PEX13 ORF, PCR was carried out with a pair of ClPEX13-specific PCR primers, sense GCGGCCGCATGGCGTCTCAGCCGCC and antisense GTCGACTCAAAGATCTTGTTTCTCTCCATTTTTC, including the initiation and termination codons (underlined), and template DNA (the reverse transcription product), as described (8,11). The AflII-BglII fragment from the ZP150-derived mutant allele, ClPEX13S319N cDNA, was placed into the AflII-BglII site of pCMVSPORT ClPEX13.The KpnI-NotI fragment of ZP150-derived ClPEX13 in pCMVSPORT was ligated into the KpnI-NotI site of pcDNA3.1/Zeo/TK. The NotI-SalI fragment of a ZP150-derived, truncated form of PEX13, ClPEX13e/s, was ligated into the site of the pUcD2Hyg vector.
The catalase latency assay using digitonin was performed as described (5). Western blot analysis was done with the respective rabbit antibodies and a second antibody, donkey anti-rabbit IgG antibody conjugated to horseradish peroxidase (Amersham Pharmacia Biotech), using ECL western blotting detection reagent (Amersham Pharmacia Biotech). P12/UV and P9OH/UV resistance was determined at 2 mMfor1.5min and 6 mM for 2 min, respectively (11).
Chinese hamster ovary
peroxisome biogenesis disorders
- PTS1 and PTS2
peroxisome targeting signal types 1 and 2
Src homology 3
We thank K. Tateishi for participating in the initial stage of this work, A. Tamura for technical assistance, K. Okumoto, M. Honsho, T. Tsukamoto and T. Osumi for comments, and R. Tanaka for secretarial work. This work was supported in part by a CREST grant (to Y.F.) from the Japan Science and Technology Corporation; Grants-in-Aid for Scientific Research (Y.F.) from the Ministry of Education, Science, Sports and Culture of Japan.