Hydrophobic Regions Adjacent to Transmembrane Domains 1 and 5 Are Important for the Targeting of the 70-kDa Peroxisomal Membrane Protein*

The 70-kDa peroxisomal membrane protein (PMP70) is a major component of peroxisomal membranes. Human PMP70 consists of 659 amino acid residues and has six putative transmembrane domains (TMDs). PMP70 is synthesized on cytoplasmic ribosomes and targeted posttranslationally to peroxisomes by an unidentified peroxisomal membrane protein targeting signal (mPTS). In this study, to examine the mPTS within PMP70 precisely, we expressed various COOH-terminally or NH2-terminally deleted constructs of PMP70 fused with green fluorescent protein (GFP) in Chinese hamster ovary cells and determined their intracellular localization by immunofluorescence. In the COOH-terminally truncated PMP70, PMP70(AA.1-144)-GFP, including TMD1 and TMD2 of PMP70, was still localized to peroxisomes. However, by further removal of TMD2, PMP70(AA.1-124)-GFP lost the targeting ability, and PMP70(TMD2)-GFP did not target to peroxisomes by itself. The substitution of TMD2 in PMP70(AA.1-144)-GFP for TMD4 or TMD6 did not affect the peroxisomal localization, suggesting that PMP70(AA.1-124) contains the mPTS and an additional TMD is required for the insertion into the peroxisomal membrane. In the NH2-terminal 124-amino acid region, PMP70 possesses hydrophobic segments in the region adjacent to TMD1. By the disruption of these hydrophobic motifs by the mutation of L21Q/L22Q/L23Q or I70N/L71Q, PMP70(AA.1-144)-GFP lost targeting efficiency. The NH2-terminally truncated PMP70, GFP-PMP70(AA.263-375), including TMD5 and TMD6, exhibited the peroxisomal localization. PMP70(AA.263-375) also possesses hydrophobic residues (Ile307/Leu308) in the region adjacent to TMD5, which were important for targeting. These results suggest that PMP70 possesses two distinct targeting signals, and hydrophobic regions adjacent to the first TMD of each region are important for targeting.

function for Pex19p that stabilizes newly synthesized PMPs in the targeting process. These discrepancies remain to be resolved, and the precise function of Pex19p is still a matter of debate.
The PTS of PMPs, termed mPTS, has yet to be identified. The mPTS was first defined in Candida boidinii PMP47 (34). Dyer et al. (35) reported that the targeting information of PMP47 resides on the matrix-oriented loop between transmembrane domain 4 (TMD4) and TMD5, which is enriched in positively charged amino acids. mPTSs have been determined for several PMPs, including Pex3p, PMP22, and PMP34 (36 -40). Although no consensus primary amino acid sequences or common structural properties have been delineated, nearly all PMP fragments that can target to peroxisomes contain a cluster of basic amino acids in conjugation with at least one TMD. Interestingly, several PMPs have been shown to contain multiple nonoverlapping mPTSs (38,40,41). However, it is still unclear whether these multiple mPTSs reflect the specific properties of PMP targeting.
To better understand the molecular mechanisms of peroxisome membrane synthesis and especially the import of human PMPs, we characterized the regions responsible for the targeting of PMP70. PMP70 is one of the major components of mammalian peroxisomal membranes and belongs to the ATP-binding cassette (ABC) protein superfamily (42,43). It consists of 659 amino acid residues, and the putative topology of PMP70 predicts six transmembrane segments with the NH 2 and COOH termini facing the cytoplasm (44). We describe that PMP70 possesses two distinct nonoverlapping targeting signals; one is in PMP70(AA. , and the other is in PMP70(AA.264 -375). Furthermore, we found that the hydrophobic segments not containing the positively charged cluster, residing just adjacent to the first TMD of both targeting ele-ments, were important for targeting but not for the binding of Pex19p.

EXPERIMENTAL PROCEDURES
Materials-pEGFP-N1 and pEGFP-C3 were purchased from Clontech (Palo Alto, CA). pQE30 and pEU3-NII were from Qiagen (Valencia, CA) and TOYOBO (Osaka, Japan), respectively. PRO-MIX TM L-[ 35 S] in vitro cell labeling mix (70% L-[ 35 S]methionine and 30% L-[ 35 S]cysteine, Ͼ37 TBq/mmol) was purchased from Amersham Biosciences. PROTEIOS TM , a wheat germ cell-free protein synthesis core kit, was obtained from TOYOBO (Osaka, Japan). Nucleotides, such as ATP, CTP, UTP, and GTP, for mRNA synthesis were obtained from Promega (Madison, WI). The protein G-agarose was from Sigma. Rabbit anti-Living Colors A. v. peptide antibody was obtained from Clontech. Mouse anti-His G antibody was from Invitrogen. Preparation of the antibody against the COOH-terminal 15 amino acids of rat PMP70 is described in Ref. 45. The anti-rat liver catalase antibodies were raised in a rabbit (46).
Construction of PMP70 Expression Plasmids for Subcellular Localization and Immunoprecipitation-pEU3-NII/PMP70(AA.  and expression constructs encoding PMP70(AA.1-659)-GFP and PMP70(AA.1-144)-GFP were prepared as described in Ref. 27. Different NH 2 -or COOH-terminal truncation mutants of PMP70 were generated by PCR using a full-length human PMP70 cDNA (47) as a template. The oligonucleotide primers used are listed in Table 1. PCR-generated fragments with XhoI or PstI and BamHI restriction sites were subcloned in frame into a pEGFP-N1 expression vector. To construct NH 2terminal GFP fusion proteins, NH 2 -or COOH-terminally truncated cDNA fragments of PMP70 were amplified from each PMP70-GFP expression vector encoding the corresponding truncated cDNA fragment of PMP70 using the forward primer TABLE 1 Oligonucleotide primer sequences used for the generation of PMP70 deletion mutants 5Ј-AAATGGGCGGTAGGCGTGT-3Ј, which annealed to a site upstream of the unique PstI site in pEGFP-N1, or 5Ј-CTC-GAGATGGCGGCCTTCAGCAAG-3Ј, which introduced an XhoI site at the amino terminus, and the reverse primer 5Ј-CGCTGAACTTGTGGCCGTTTA-3Ј, which annealed to a site downstream of the unique BamHI site in pEGFP-N1. PCRgenerated fragments with PstI or XhoI and BamHI restriction sites were subcloned in frame into a pEGFP-C3 expression vector. To construct PMP70(AA.1-124/TMD4)-GFP and PMP70(AA.1-124/TMD6)-GFP, in which TMD4 or TMD6 of PMP70 was inserted just downstream of PMP70(AA.1-124), the PMP70(AA.1-124) cDNA fragment was generated by PCR using the forward primer 5Ј-CTCGAGCCGCCATGGCGGC-CTT-3Ј and the reverse primer 5Ј-CTGCAGTCTCTTGAAA-TCTTTCCTGCTACG-3Ј. PCR-generated fragments with XhoI and PstI restriction sites were subcloned in frame into PMP70(TMD4)-GFP and PMP70(TMD6)-GFP expression vectors. The identity of all subclones was confirmed by semiautomated sequencing on an ABI 310 DNA sequencer (PerkinElmer Life Sciences).
Culturing Conditions and Transient Transfection-CHO-K1 cells were cultured in F12K medium (ICN, Aurora, OH) with 10% fetal bovine serum at 37°C and 5% CO 2 . 48 h before transfection, 5 ϫ 10 3 cells were seeded on a Lab-Tek TM Chamber Slide TM system with eight chambers on a glass slide (Nalge Nunc, Rochester, NY). All transfections were performed using an Effectene Transfection Reagent (Qiagen, Valencia, CA) according to the manufacturer's instructions. Two days after transfection, the cells were washed three times with phosphatebuffered saline and fixed for 10 min in 5% paraformaldehyde in phosphate-buffered saline for indirect immunofluorescence.
Indirect Immunofluorescence-Immunostaining was performed by essentially the same procedure as described in Ref. 7. The fixed cells were permeabilized in 0.1% (w/v) Triton X-100 in phosphate-buffered saline for 10 min, washed twice with phosphate-buffered saline, and incubated with the primary antibodies for 1 h at room temperature. The primary antibodies used in this study were a rabbit antibody against the COOHterminal 15 amino acids of rat PMP70 (1:200) and a rabbit antibody against rat catalase (1:200). Cy3-conjugated goat anti-rabbit IgG antibody (Amersham Biosciences) was used to label the first antibody. The cells were mounted in 90% glycerol in 100 mM Tris-HCl (pH 8.0), and the samples were examined by confocal microscopy (LSM510; Carl Zeiss, Jene, Germany). To analyze the efficiency of peroxisomal localization, samples were examined by TCS-SP5 software (Leica, Wetzlar, Germany). The per pixel scatter diagrams were generated using the built in software of the Leica TCS-SP5. Peason's correlation coefficient (PCC) and the peroxisome colocalization rate were employed to evaluate colocalization. PCC is one of the standard measures to assess the relationship between fluorescence intensities (48). Its value ranges between Ϫ1.0 and 1.0, where Ϫ1.0 represents no overlap and 1.0 represents complete colocalization. The peroxisome colocalization rate was expressed as a ratio of colocalization area showing certain red pixel intensity of peroxisomal marker and certain green pixel intensity of each GFP fusion protein against the area foreground. In a typical experiment, 20 randomly chosen areas containing some of the cells expressing GFP fusion protein were examined for each culture, and each experiment was repeated at least three times.
Purification of His-Pex19p-Purification of the NH 2 -terminal His 6 -tagged human Pex19p (His-Pex19p) was performed by essentially the same procedure as described in Ref. 27. M15 pREP4 Escherichia coli cells (Qiagen, Valencia, CA) harboring pQE30/PEX19 were grown at 37°C in LB medium containing 0.1 mg/ml ampicillin. At a cell density of 0.5 (A 600 ), protein expression was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside for 5 h at 37°C. The cells were harvested by centrifugation at 4,000 ϫ g for 20 min, resuspended in 35 ml of the lysis buffer (50 mM Tris-HCl, pH 7.5, 0.3 M NaCl, 5 mM imidazole, 0.1 mM phenylmethylsulfonyl fluoride), and disrupted 20 times for 20 s in an ice bath by an Astrason XL-2020 ultrasonic processor (Misonix Inc., Farmingdale, NY). The lysate was cen- GGTGGAACACCTACATAATTTCAATCAGTTTCGGTTTTCAATGGGC trifuged at 20,000 ϫ g for 30 min and the His-Pex19p in the supernatant was immediately applied to 10 ml of TALON Metal affinity resin (Clontech) equilibrated with the lysis buffer. After extensive washing, the His-Pex19p was eluted with the lysis buffer containing 250 mM imidazole. The eluted fractions containing His-Pex19p were dialyzed against 50 mM Tris-HCl, pH 8.0, 50 mM NaCl, and 10 mM dithiothreitol and stored at Ϫ80°C.
In Vitro Transcription and Translation-The plasmids encoding wild type and mutant PMP70s were transcribed in vitro using T7 RNA polymerase, and the synthesized mRNAs were isolated by a MicroSpin G-25 column (Amersham Biosciences). Using the purified mRNA, cell-free translation was performed according to the bilayer method using PROTEIOS TM , a wheat germ cellfree protein synthesis core kit, according to the manufacturer's procedure. In a typical experiment, the synthesized mRNAs were translated for 24 h at 26°C in a 300-l wheat germ cell-free protein synthesis system containing 50 Ci of [ 35 S]methionine in the presence of 100 g of His-Pex19p. After translation, the reaction mixture was centrifuged for 20 min at 17,000 ϫ g, and the supernatant was used for co-immunoprecipitation.
Co-immunoprecipitation-Translation products (50 l) were precleaned with an appropriate amount of protein G-agarose in 200 l of the binding buffer (20 mM Hepes-KOH, pH 7.5, 110 mM potassium acetate, 5 mM sodium acetate, 2 mM magnesium acetate, 1 mM EDTA, 0.2% Triton X-100, 10 mM dithiothreitol). After this step, the supernatant was removed and incubated with protein G-agarose beads saturated with anti-His G antibody. After incubation of the suspensions for 2 h at 4°C, the beads were collected by centrifugation and washed five times with 250 l of the binding buffer. Immunoprecipitated proteins were analyzed on a 7-15% SDS-polyacrylamide gradient gel. The gels were dried, and the radioactivity of the band corresponding to PMP70 was quantified by a Fuji BAS 5000 imaging analyzer (Fuji Film, Tokyo, Japan).
Other Methods-Protein was assayed as described previously (43). Western blot analysis was performed with primary antibodies and a second antibody, donkey anti-rabbit IgG antibody conjugated to horseradish peroxidase (Amersham Biosciences). Antigen-antibody complex was visualized with ECLϩPlus Western blotting detection reagent (Amersham Biosciences). ization by immunofluorescence. We have recently shown that PMP70(AA.1-659)-GFP and PMP70(AA.1-375)-GFP, which possess whole NH 2 -terminal transmembrane segments, were localized to peroxisomes and that PMP70(AA.1-144)-GFP was still targeted to peroxisomes, indicating that PMP70 possesses an mPTS in the NH 2 -terminal 144-amino acid region (27). It is suggested that the targeting characteristics are influenced by the position of GFP. Therefore, we expressed various COOH-terminal deletion constructs of PMP70 in fusion with the COOH terminus of GFP (Fig. 1A). As shown in Fig. 1B, GFP-PMP70(AA.1-375) exhibited a punctated immunofluorescent pattern that coincided with that of peroxisomes. GFP-PMP70(AA.1-324) was still localized to peroxisomes, and further COOH-terminal deletion constructs, including GFP-PMP70(AA.1-276), GFP-PMP70(AA.1-228), and GFP-PMP70(AA.1-144), were still directed to peroxisomes. These results also indicate that, even in the case of NH 2 -terminal GFP fusion, the mPTS of PMP70 could still exist in the NH 2 -terminal 144amino acid region.

Subcellular Localization of COOH-terminal Truncated
Role of TMD2 in the Targeting of PMP70-To identify the minimal region in the NH 2 -terminal region of PMP70 that is sufficient for its proper peroxisomal localization, we made small deletions in the NH 2 -terminal region and examined the subsequent intracellular localization ( Fig These values were almost the same as in the case of GFP alone (PCC of 0.28 Ϯ 0.03, and 10% of the fluorescence coincided with that of peroxisomes). Furthermore, subcellular fractionation showed that the distribution of PMP70(AA.1-124)-GFP was different from that of endogenous PMP70 and was almost the same as that of calnexin, a marker protein of endoplasmic reticulum (data not shown). PMP70(TMD2)-GFP did not target to peroxisomes by itself either. PMP70(TMD2)-GFP was mainly recovered in the nuclear fraction and some was in the cytosol fraction (data not shown). To elucidate the role of the second TMD in the peroxisomal targeting of PMP70, TMD2 in PMP70(AA.1-144)-GFP was swapped for TMD4 or TMD6 of PMP70, respectively. PMP70(TMD4)-GFP and PMP70(TMD6)-GFP did not target to peroxisomes by themselves. However, PMP70(AA.1-124/TMD4)-GFP and PMP70(AA.1-124/TMD6)-GFP displayed the peroxisomal localization. These results suggest that the important targeting information of PMP70 is contained within an NH 2 -terminal 124-amino acid region and that at least two TMDs are required for integration into the peroxisomal membrane.
Basic Amino Acid Clusters in the NH 2 -terminal Region of PMP70 Are Not Essential for the Peroxisomal Targeting of PMP70-In the NH 2 -terminal 124-amino acid region, there are three clusters of basic amino acids (the first cluster is at amino acid positions 28, 29, 30, 31, 38, and 39; the second is at positions 53, 54, 56, 61, 66, 72, and 77; the third is at positions 117, 119, 120, 123, and 124) (Fig. 3A). The positively charged amino acid cluster is suggested to be the peroxisomal targeting motif of other PMPs. To examine whether these basic clusters function as an mPTS of PMP70, we replaced these basic amino acids in the respective parts of PMP70(AA.1-144)-GFP with Ala. PMP70(AA.1-144)-GFP was localized to peroxisomes (Fig. 2B).  Fig. 4). These data suggest that these basic clusters are not essential for the targeting of PMP70, and the mPTS of PMP70 is located in another part of the molecule.
Hydrophobic Motifs Adjacent to the NH 2 -terminal Side of TMD1 Are Important for the Stability and the Peroxisomal Targeting of PMP70-Based on the hydropathy profiling, peroxisomal ABC proteins possess two hydrophobic segments adjacent to the NH 2 -terminal side of TMD1 (Fig. 3A). To examine whether these hydrophobic motifs are important for the targeting of PMP70, we disrupted these hydrophobic properties by the mutation of L21Q/L22Q/L23Q or I70N/L71Q and examined the subcellular localization. When these amino acid residues were substituted for alanines considered to retain the minimum hydrophobic property of these sequences in Interaction between Mutant PMP70 and Pex19p-We recently found that Pex19p, a protein required for peroxisomal membrane biogenesis, interacts with the NH 2 -terminal region of PMP70 to maintain it in soluble and proper conformation in the cytosol (27). Therefore, we investigated the interaction between Pex19p and mutant PMP70s in which the NH 2 -terminal hydrophobic motifs were disrupted. As shown in Fig. 5A, wild type and mutant PMP70s were translated in a wheat germ in vitro translation system in the presence or absence of purified His-Pex19p. Wild type PMP70 was solubilized by the addition of purified His-Pex19p, and about 60% of wild type PMP70 synthesized in the cell-free system was recovered in the supernatant fraction after centrifugation at 17,000 ϫ g for 20 min (Fig. 5, A and B). The solubilized PMP70 was co-immunoprecipitated with His-Pex19p (Fig. 5, C and D). described above. These data suggest that a set of the NH 2 -terminal hydrophobic residues, Leu 21 -Leu 22 -Leu 23 , is necessary for the interaction with Pex19p to maintain PMP70 in the soluble and proper conformation in the targeting process, and the second set of hydrophobic residues, the pair of Ile 70 -Leu 71 , is required in the targeting step after PMP70 has already associated with Pex19p. PMP70 Possesses a Second and Distinct mPTS in the Region of TMD5-TMD6-Furthermore, we examined the subcellular localization of NH 2 -terminal truncated PMP70 in fusion with the COOH terminus of GFP (Fig. 6). GFP-PMP70(AA.1-375) was localized to peroxisomes as described above. GFP-PMP70(AA.113-375), which lacks the NH 2 -terminal hydrophobic region required for the targeting of PMP70, still retained the ability to target to peroxisomes (PCC of 0.81 Ϯ 0.03). GFP-PMP70(AA.176 -375) was still partially localized to peroxisomes (PCC of 0.74 Ϯ 0.04). By the further removal of TMD3 from PMP70, GFP-PMP70(AA.224 -375) lost the targeting ability. Image quantitation implied the random localization of GFP-PMP70(AA.224 -375) (PCC of 0.51 Ϯ 0.01), and GFP-PMP70(AA.224 -375) did not exhibit any punctated pattern like peroxisomes. However, GFP-PMP70(AA.263-375), which comprises TMD5 and TMD6 of PMP70, restored the peroxisomal localization (PCC of 0.82 Ϯ 0.04). The localization was diminished by the removal of TMD5, and GFP-PMP70(AA.314 -375) was diffused in the cytosol and partially localized to endoplasmic reticulumlike structures. These data suggest that, in addition to the NH 2 -terminal hydrophobic motif, PMP70 possesses a second and distinct mPTS in the region of PMP70(AA.263-375).
We also examined the subcellular localization of NH 2 -terminal truncated PMP70 in fusion with the NH 2 terminus of GFP (   Fig. 4C). These data suggest that the hydrophobic motif located adjacent to the NH 2 -terminal side of TMD5 is also important for the peroxisomal targeting of PMP70.
To characterize the hydrophobic motif in the targeting process of PMP70 more precisely, we examined the interaction between PMP70(AA.1-659 I307N/L308Q) and Pex19p. As shown in Fig. 9, PMP70(AA.1-659 I307N/L308Q) translated in the wheat germ in vitro translation system was almost recovered as a soluble protein in the presence of purified His-Pex19p, and PMP70(AA.1-659 I307N/L308Q) interacted with Pex19p at almost the same efficiency as PMP70(AA.1-659). These data suggest that the hydrophobic region constituted by Ile 307 and Leu 308 could function as an mPTS rather than a region that is required to stabilize PMP70 in the targeting process through interaction with Pex19p.  luminal loop of PMP70 functions as sufficient information for peroxisomal targeting (data not shown).

DISCUSSION
Most of the PMPs are synthesized on free cytosolic polysomes and posttranslationally targeted to peroxisomes. However, the precise targeting process of PMPs is still unknown, and furthermore the common peroxisome targeting signal of the PMPs is not yet identified. In this study, we examined the characteristics of the mPTS of PMP70.
At first, it was found that a targeting element of PMP70 existed within the NH 2 -terminal 124 amino acids, and at least two TMDs were required for the integration into peroxisomal membrane by the following observations. 1) PMP70(AA.1-144)-GFP was localized to peroxisomes. 2) PMP70(AA.1-124)-GFP, which was deleted of the region of TMD2(AA.125-144), lost the ability to localize to peroxisomes, but substitutions of TMD2 in PMP70(AA.1-144)-GFP with TMD4 or TMD6 of PMP70 did not affect the peroxisomal localization (Fig.  2B).
3) The finding that TMD2, TMD4, or TMD6 did not locate to peroxisomes excludes the possibility that these regions possess sufficient information for peroxisomal localization (Fig. 2B). As for the mPTS in the NH 2 -terminal region of PMP70, Sacksteder et al. (22) reported that the COOH-terminally Myc-tagged PMP70(AA.1-124) was able to target to peroxisomes, and the efficiency of targeting was much decreased in Myc-tagged PMP70(AA.1-61). Biermanns and Gärtner (49) determined that a region of 20 amino acids (positions 61-80) contained important targeting information from their observations that GFP-PMP70(AA.61-180) targeted to peroxisomes and that GFP-PMP70(AA.81-160) did not display any peroxisomal localization. Thus, our data were partly consistent with these observations. On the other hand, we found an additional novel mPTS of PMP70 in the region of amino acids 263-375, including TMD5 and TMD6 (Fig.  6B). We have previously shown that the PMP70 produced by the in vitro translation system was inserted into rat liver peroxisomes, and the truncated PMP70 that was deleted of the NH 2 -terminal 20-kDa region adjacent to TMD3 was also associated with peroxisomes (45). These data imply the existence of another mPTS, which is outside of the NH 2 -terminal region. Indeed, in some PMPs, it is suggested that there are multiple targeting signals. As for the targeting of PMP47, Dyer et al. (35) first reported that the targeting information was contained in the basic cluster within matrix loop 2, which connects TMD4 and TMD5 of PMP47. In addition to this observation, Wang et al. (41) showed that TMD2 and an adjacent region of cytosolic loop 1 were also crucial for the targeting of PMP47. Jones et al. (40) reported that PMP34 contained multiple distinct targeting signals, and Brosius et al. (38) showed that PMP22 contained two distinct and nonoverlapping peroxisomal membrane targeting signals. The meaning of the presence of multiple mPTSs in these PMPs is not well understood yet, but Wang et al. (50) suggest that two mPTSs of PMP47 function differently among the peroxisome populations in Saccharomyces cerevisiae. These data give impetus to the possibility that PMP70 possesses another mPTS in a region distinct from the NH 2 -terminal region, and PMP70 adopts the mPTSs depending on the condition of the peroxisomes. Indeed, PMP70 has the property to be highly enriched in the rodent liver peroxisomal membrane under the condition in which peroxisomes are induced by the administration of hypolipidemic agents, such as clofibrate (51,52).
Our results also showed that the NH 2 -terminal basic clusters did not function as the mPTS of PMP70, but the hydrophobic motifs just adjacent to the first TMD were important for the targeting of PMP70, as suggested by the following observations. 1) Disruption of basic amino acid clusters that exist in the NH 2terminal region did not affect the peroxisomal targeting efficiency of PMP70(AA.1-144)-GFP (Fig. 3B). acid pair existing just adjacent to TMD5 was important for the targeting to peroxisomes; GFP-PMP70(AA.263-375 I307A/L308A) was directed to peroxisomes, but GFP-PMP70(AA.263-375 I307N/L308Q) was not (Fig. 8). In a previous paper (27), we proposed that the mPTS of PMP70 was located near the TMD2, including a positively charged cluster of basic amino acids, in a study using COOH-terminal truncated PMP70-GFP. However, the present study suggests that the region is essential for the insertion of PMP70 into the peroxisomal membranes, but a positively charged cluster is not involved in the process. Concerning the targeting element of PMPs, positively charged amino acid clusters in the matrix loop are suggested to be essential. Dyer et al. (35) first defined a basic cluster of amino acids of the sequence KIKKR existing on the second intraperoxisomal loop as a targeting motif of PMP47. Baerends et al. (36) also reported that RHKKK at the NH 2 terminus of Hansenula polymorpha Pex3p was important for sorting. In addition, a matrix-oriented positively charged amino acid cluster was seen in many PMP fragments that can target to peroxisomes, such as PMP22, Mpv17-like protein, Pex16p, and ascorbate peroxidase (39,41,(51)(52)(53)(54)(55)(56). Our data are inconsistent with these observations. However, Pause et al. (37) reported that YX 3 LX 3 PX 3 (K/Q/N), a conserved motif among PMP22 orthologues, comprised the core of the mPTS of PMP22, and the basic cluster in the first peroxisomal matrix loop was not essential for targeting. They also found that the hydrophobic sequence was necessary for the targeting and/or insertion of PMP22. Recently, Landgraf et al. (57) identified a 14-amino acid motif (F(F/L)X(R/Q/K)(L/F)(L/I/ K)XLLKIL(F/I/V)P/) as an mPTS of adrenoleukodystrophy protein, one of the peroxisomal ABC proteins, and found that the substitution or deletion of these hydrophobic residues significantly reduced the targeting efficiency. In particular, the deletion of three amino acids (Leu 78 -Leu 79 -Arg 80 ) lost peroxisomal targeting of adrenoleukodystrophy protein.
The region corresponds to Ile 70 -Leu 71 -Lys 72 in PMP70, which we found to be an mPTS of PMP70. Taking these observations into consideration, we suggest that in a group of PMPs, the mPTS is composed of hydrophobic regions but not basic amino acid clusters. Another point we addressed in this study is the function of Pex19p in the targeting process of PMP70. Pex19p is a farnesylated protein essential for the early steps of peroxisome biogenesis and most likely is involved in peroxisomal membrane synthesis. Pex19p is mainly located in the cytosol and is known to bind multiple PMPs. From these findings, Pex19p has been proposed to function either as a receptor for the mPTS of PMPs or as a chaperone that stabilizes PMPs in the cytosol. Sacksteder et al. (22) found that the targeting regions of multiple PMPs were also recognized by Pex19p. Jones et al. (28) reported that the attachment of a nuclear localization signal to Pex19p lead to accumulations of mPTS regions of PMP34, PEX11␤, PEX16, PMP22, and PMP70 in the nucleus. Furthermore, Rottensteiner et al. (29) deduced the amino acid sequence, X 3 (C/ F/I/L/T/V/W)X 2 (A/C/F/I/L/Q/V/W/Y)(C/I/L/V)X 2 (A/C/F/I/ L/V/W/Y)(I/L/Q/R/V)X 3 as the common Pex19p-binding motif in PMPs. They also reported that the Pex19p-binding site in conjugation with one or more adjacent TMDs of Pex13p possessed peroxisome targeting ability, and a mutation within the Pex19p-binding site that disrupted the Pex19p binding impaired the peroxisomal targeting of Pex13p (29). A similar Pex19p-binding site-dependent targeting was observed in Pex17p (31). These findings strongly support the function of Pex19p as a PMP import receptor, which recognizes the mPTS of peroxisomal membrane proteins and delivers them to peroxisomes. Recently, we found that an NH 2 -terminal 61-amino acid region and TMD5-TMD6 of PMP70 interacted with Pex19p in in vitro binding experiments (27). These regions partly overlap with the mPTSs which we found in this study. However, we also found that PMP70(AA.1-659 I70N/L71Q) and PMP70(AA.1-659 I307N/L308Q) lost the targeting activity, although these mutant proteins still interacted with Pex19p and were solubilized by Pex19p, suggesting that the targeting element and Pex19p-binding site of PMP70 are functionally separated. This finding seems to claim the function of Pex19p as a mPTS receptor. Snyder et al. (23) reported that the domains of Pex3p, Pex10p, Pex13p, and Pex22p, which interact with Pex19p, did not function as an mPTS. Fransen et al. (24) also found that the Pex19p-binding sites of Pex3p and Pex12p were separated from their mPTS regions. Further, they found that the mutant Pex13p did localize to peroxisomes but did not show binding affinity for Pex19p (58). Recently, Vizeacoumar et al. (33) showed that both Pex30p and Pex32p of S. cerevisiae interacted with Pex19p in regions that did not overlap with their mPTSs. These data suggest that Pex19p functions as a chaperone for PMPs in the targeting process rather than acting as an mPTS receptor. In this study, some of the deletion constructs of PMP70 still possessed peroxisomal targeting ability. These deletion constructs might have different conformation with respect to native PMP70. However, similar results were obtained for various PMPs (34 -41, 49, 54, 57). These data suggest that these PMP fragments possess the ability to form a proper conformation by themselves. In this process, Pex19p is supposed to play an important role. Consistent with the hypothesis, the fragments of various PMPs, including Pex16p, PMP22, and PMP34, which can target to peroxisomes, were shown to be able to interact with Pex19p (28). As for the PMP70, our recent study showed that various COOH-terminal or NH 2 -terminal deletion constructs still interacted with Pex19p, although the NH 2 -terminal 61-amino acid region and TMD5-TMD6 of PMP70 are required for efficient binding (27). These interactions might prevent entire conformational changes of deletion constructs and might keep them in a proper conformation for the peroxisomal targeting.
In summary, based on our results, we propose a hypothetical model for the targeting of PMP70 (Fig. 10). After being synthesized on free cytosolic ribosomes, PMP70 interacts with Pex19p through the NH 2 -terminal hydrophobic motif constituted by Leu 21 -Leu 22 -Leu 23 and the region of TMD5-TMD6. By the assistance of Pex19p, PMP70 is kept in soluble and proper conformation in the cytosol, and PMP70, which cannot associate with Pex19p, forms aggregates and/or is degraded by proteasomes. Then the PMP70-Pex19p complex is transported to peroxisomes by the mPTSs located in the NH 2 -terminal 124amino acid region and the region of PMP70(AA.263-375) (the hydrophobicities of Ile 70 -Leu 71 and Ile 307 -Leu 308 might be essential). Finally, PMP70 is inserted into peroxisomal membranes through the unidentified proteinaceous components on the peroxisomal membranes. In this process, at least two TMDs are required for correct insertion.