Expression of PEX11β Mediates Peroxisome Proliferation in the Absence of Extracellular Stimuli*

Mammalian cells typically contain hundreds of peroxisomes but can increase peroxisome abundance further in response to extracellular stimuli. We report here the identification and characterization of two novel human peroxisomal membrane proteins, PEX11α and PEX11β. Overexpression of the human PEX11β gene alone was sufficient to induce peroxisome proliferation, demonstrating that proliferation can occur in the absence of extracellular stimuli and may be mediated by a single gene. Time course studies indicated that PEX11β induces peroxisome proliferation through a multistep process involving peroxisome elongation and segregation of PEX11β from other peroxisomal membrane proteins, followed by peroxisome division. Overexpression of PEX11α also induced peroxisome proliferation but at a much lower frequency than PEX11β in our experimental system. The patterns ofPEX11α and PEX11β expression were examined in the rat, the animal in which peroxisome proliferation has been examined most extensively. Levels of PEX11β mRNA were similar in all tissues examined and were unaffected by peroxisome-proliferating agents. Conversely, PEX11α mRNA levels varied widely among different tissues, were highest in tissues that are sensitive to peroxisome-proliferating agents, and were induced more than 10-fold in response to the peroxisome proliferators clofibrate and di(2-ethylhexyl) phthalate. Taken together, these data implicate PEX11β in the constitutive control of peroxisome abundance and suggest that PEX11α may regulate peroxisome abundance in response to extracellular stimuli.

Peroxisomes are ubiquitous components of eukaryotic cells, absent only from mature erythrocytes and certain primitive unicellular eukaryotes. One of the more intriguing aspects of peroxisome biogenesis is how cells control the abundance of this organelle. Mammalian cells contain hundreds of peroxisomes under normal growth conditions, suggesting that there are constitutive mechanisms for raising peroxisome abundance above one per cell. In addition, peroxisome abundance may change in response to extracellular stimuli, indicating the existence of a signal transduction pathway that exerts additional control over peroxisome abundance. Inducers of peroxisome proliferation include both hypolipidemic drugs (e.g. clofibrate) and plasticizing agents (e.g. di(2-ethylhexyl) phthalate (DEHP) 1 ), which act through PPAR␣, the ␣ isoform of the peroxisome proliferator-activated receptor (1)(2)(3). PPAR␣ is a member of the nuclear hormone receptor superfamily and functions as a heterodimer with retinoid X receptor (RXR), another nuclear hormone receptor. The activated PPAR␣⅐RXR heterodimer binds peroxisome proliferator-responsive elements (PPREs) and mediates transcriptional activation of a large array of PPRE-containing genes in a drug-dependent manner (4). However, the pathway between altered gene expression and peroxisome proliferation remains to be elucidated.
Peroxisome proliferation has also been observed in lower eukaryotes. In the yeast Saccharomyces cerevisiae, fatty acid oxidation is an exclusively peroxisomal process. Exposure to fatty acids, particularly oleic acid, leads to an increase in peroxisome abundance from 1-2/cell to 10 -20/cell (5). This example of peroxisome proliferation is also associated with dramatic changes in gene expression and requires the transcription factors PIP2 (6) and OAF1 (7,8). Together, these two proteins bind oleate-response elements within transcriptional control regions of responsive genes and are required for both the transcriptional response to oleic acid and the proliferation of peroxisomes. Of the many genes known to be induced by oleic acid, PEX11 is the only one (other than PIP2 itself) that is required for the normal peroxisome proliferation response: pex11 mutants accumulate only 4 -5 very large peroxisomes when incubated in oleic acid (5,9). Furthermore, overexpression of PEX11 can enhance fatty acid-induced peroxisome proliferation. These data demonstrate a role for PEX11 in the regulation of peroxisome abundance. However, there is no evidence that overexpression of ScPEX11 alone, in the absence of extracellular stimuli, can mediate peroxisome proliferation. Here we report the identification and characterization of two mammalian PEX11 genes that can induce peroxisome abundance in the absence of extracellular stimuli.

EXPERIMENTAL PROCEDURES
Plasmids-Plasmids corresponding to apparent full-length cDNAs for human PEX11␣ (GenBank accession number R18258, EST cDNA clone number 30793) and human PEX11␤ (GenBank accession number AA227332, EST cDNA clone number 663688) were obtained from Genome Systems (St. Louis). The cDNA inserts in these plasmids were sequenced in their entirety on both strands and the sequences of human * This work was supported by Grant DK45787 from the National Institutes of Health (to S. J. G.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  PEX11␣ (accession number AF093668 and PEX11␤ (accession number AF093670) are available from GenBank. The PEX11␣ expression vector, pBER22, was created by excising the PEX11␣ cDNA from clone 30793 by digestion with HindIII (partial) and NotI and inserting it between the HindIII and NotI sites of the mammalian expression vector pcDNA3 (Invitrogen, San Diego), downstream of the cytomegalovirus promoter. The PEX11␤ expression vector was created by excising the PEX11␤ cDNA from clone 663688 by digestion with EcoRI and XhoI and inserting it between the EcoRI and XhoI sites of pcDNA3. The COOHterminally tagged versions of PEX11␣ and PEX11␤ were generated by polymerase chain reaction using gene-specific oligonucleotides designed to amplify the appropriate open reading frame with the addition of the sequence 5Ј-GGTACC-3Ј immediately preceding the ATG of the open reading frame and with the sequence 5Ј-GGATCC-3Ј in place of the stop codon. Each polymerase chain reaction product was cleaved with Asp718 and BamHI and cloned between the Asp718 and BamHI sites of pcDNA3-myc (10). To generate the NH 2 -terminally myc-tagged versions of PEX11␣ and PEX11␤ each open reading frame was amplified using appropriate gene-specific primers designed to replace the start codon with the sequence 5Ј-GGATCC-3Ј and add the sequence 5Ј-CTCGAG-3Ј downstream of the stop codon. Once synthesized, each product was cleaved with BamHI and XhoI and inserted between the BamHI and XhoI sites of pcDNA3-Nmyc (a modified version of pcDNA3 that carries the sequence 5Ј-ACCATGGCGGAGCAGAAGCTGATCTCCGAGGAGG-ACCTGCTG-3Ј (start codon underlined) between the HindIII and BamHI sites of pcDNA3). Mus musculus PEX11␣ (accession number W42147, EST cDNA clone number 329428) and PEX11␤ (accession number AA237383, EST cDNA clone number 678785) cDNAs were also obtained from Genome Systems. The sequences of the murine PEX11␣ and PEX11␤ cDNAs are available from GenBank (accession numbers AF093668 and AF093671, respectively).
Transfections, Immunofluorescence, and Antibodies-HepG2 cells and 5756T human fibroblast cells were cultured under standard conditions (11). Transfections were performed by electroporation (12). Indirect immunofluorescence was performed on cells grown on glass coverslips. 2 days after transfection, cells were fixed by incubation in 3% formaldehyde in Dulbecco's modified phosphate-buffered saline, pH 7.5 (DPBS; Life Technologies, Inc.), for 20 min and then washed twice in DPBS. Next, cell membranes were permeabilized by incubation in either 0.1% Triton X-100 and DPBS for 5 min (standard permeabilization conditions) or in 25 g/ml digitonin and DPBS for 5 min (differential permeabilization conditions). Cells were then washed twice in DPBS and incubated with primary antibodies diluted in DPBS containing 0.1% bovine serum albumin (1:1,000 for the anti-PMP70 antiserum, 1:1 for the anti-myc hybridoma tissue culture supernatant, and 1:300 for the anti-SKL antibodies). After 10 washes with DPBS, cells were incubated with fluorescently labeled secondary antibodies diluted in DPBS (fluorescein or Texas Red), washed an additional 10 times, mounted, and viewed using an Olympus fluorescence microscope. To ensure the integrity of the peroxisome membrane in all differential permeabilization experiments, equivalent cell samples were stained with anti-SKL antibodies, which detect multiple peroxisomal matrix proteins (13). Data from differential permeabilization experiments were used only if these antibodies failed to detect peroxisomes in cells permeabilized with digitonin but did detect peroxisomes in matched samples permeabilized with Triton X-100. Antiserum directed against PMP70 was a generous gift from Dr. Suresh Subramani (UCSD, La Jolla, CA). Anti-myc antibody was obtained from the tissue culture supernatant of the hybridoma cell line 1-9E10 (14). Sheep anti-catalase antibodies were obtained from The Binding Site (San Diego). Secondary antibodies specific for rabbit, sheep, or mouse antibodies were obtained from standard commercial sources.
Northern Blots-To assess the regulation of gene expression by peroxisome-proliferating agents in rat liver, rats were fed either standard chow or chow supplemented with DEHP (2% w/w in rat chow) for 7 days or clofibrate (0.5% w/w in rat chow) for 2 weeks (15). Total RNA was isolated from livers of control, DEHP-, and clofibrate-fed rats and analyzed by Northern blot. The rat multitissue Northern blot (2 g of poly(A) plus mRNA/lane) was obtained from CLONTECH (Palo Alto, CA). Probes were generated by random primed labeling of gene-specific cDNA fragments, and hybridizations were carried out using standard protocols (16). Detection was by exposure to XAR-5 film (Kodak) or a Fuji PhosphorImager.

RESULTS
Identification of PEX11␣ and PEX11␤-We used the BLAST algorithm to scan the data base of expressed sequence tags (dbEST) for human cDNAs capable of encoding proteins similar to ScPEX11. Several overlapping ESTs from a single human gene were identified, and the longest cDNA for this gene was obtained and sequenced. The product of this human PEX11 homolog was more similar to ScPEX11 than to any other S. cerevisiae protein. However, its closest relative was the product of a second human gene, also represented by several ESTs. We obtained and sequenced cDNAs for this second human PEX11 homolog as well. These two human genes were designated PEX11␣ and PEX11␤ based on the similarity of their products to yeast PEX11 (Fig. 1). The presence of two human PEX11 genes was unexpected because all previously characterized PEX genes are present in a single copy in both yeast and humans. This result led us to examine the S. cerevisiae genome for any proteins similar to ScPEX11, but none was detected by the BLAST algorithm. The presence of two PEX11 genes was not unique to humans: PEX11␣ and PEX11␤ genes were also identified in mouse (Fig. 2). Although PEX11␣ and PEX11␤ are the mammalian proteins with greatest similarity to yeast PEX11, there is only ϳ20% amino acid identity between the mammalian proteins and their yeast homolog. Furthermore, these human proteins differ from ScPEX11 in that they contain a strongly predicted membrane-spanning domain at their COOH termini (residues 220 -239 of PEX11␣ and 230 -255 of PEX11␤) and a second membrane-spanning domain located about 100 amino acids from their NH 2 termini (residues 94 -114 of PEX11␣ and 94 -113 of PEX11␤). The transmembrane segment prediction algorithm (TM predict; http://ulrec3.unil. ch/software/TMPRED_form.html) also suggested that PEX11␣ and PEX11␤ would be oriented with their termini extending into the cytoplasm. In contrast, ScPEX11 is not predicted to have a membrane-spanning segment and does not behave as an integral peroxisomal membrane protein (PMP) (5).
Overexpression of PEX11␤ Induces Peroxisome Proliferation-Although the sequence similarities between yeast PEX11 and human PEX11␣ and PEX11␤ were relatively mild, they did suggest that these proteins may play some role in the control of peroxisome abundance. We explored this possibility by testing whether overexpression of PEX11␣ or PEX11␤ alone could stimulate peroxisome proliferation. Human cells were transfected with plasmids designed to express either PEX11␣ or PEX11␤. 2 days after transfection, peroxisome abundance was assessed by indirect immunofluorescence using antibodies specific for PMP70, a ubiquitous PMP (17). Peroxisome abundance was increased in cells transfected with the PEX11␤ expression vector (Fig. 3). The proportion of cells displaying increased peroxisome abundance in transfected cell populations corresponded roughly to the transfection efficiency in each instance, approximately 30 -50% in different trials and cell lines. In contrast to these results, peroxisome proliferation was detected in less than 1% of the cells transfected with the PEX11␣ expression vector (data not shown).
Peroxisomes can be quite heterogeneous in composition. Therefore, we tested whether expression of PEX11␤ also induced the proliferation of catalase-containing peroxisomes. Cells were transfected as above but processed for double indirect immunofluorescence using anti-catalase and anti-PMP70 antibodies. If there is any heterogeneity in the composition of these peroxisomes it is not apparent in the distribution of these proteins: both PMP70 and catalase were detected in almost all peroxisomes in cells that had undergone PEX11␤-mediated peroxisome proliferation (Fig. 4).
Peroxisome proliferation is normally mediated by extracellular stimuli. To test whether the changes in peroxisome abundance observed above were limited to those cells that overexpressed PEX11␣ or PEX11␤, we repeated these transfection experiments with plasmids designed to express myc epitopetagged derivatives of PEX11␣ and PEX11␤. Expression of a COOH-terminally myc-tagged version of PEX11␤, PEX11␤myc, revealed that 1) PEX11␤myc induced peroxisome proliferation in almost all expressing cells (Fig. 5, A and B), 2) peroxisome proliferation was detected only in those cells that expressed PEX11␤myc, 3) the degree of peroxisome proliferation correlated roughly with the extent of PEX11␤myc expression, and 4) PEX11␤myc colocalized with PMP70, demonstrating that it is a peroxisomal protein. We also examined the activity and subcellular distribution of an NH 2 -terminally tagged form of PEX11␤, mycPEX11␤. This protein also induced peroxisome proliferation in a cell-limited fashion and colocalized with PMP70 (Fig. 5, C and D).
During the course of these experiments we made the rather curious observation that PEX11␤myc and mycPEX11␤ were detected only rarely (and weakly) in cells that had been fixed and then permeabilized with Triton X-100 (0.1% for 5 min, our standard permeabilization protocol) but were detected readily in 30% of the same transfected cell population when the cells were fixed and then permeabilized with digitonin (25 g/ml for 5 min), as apparent from Fig. 5. Normally, these two permeabilization techniques allow one to determine whether a protein is intraperoxisomal (if detected in cells permeabilized with Triton X-100 but not in cells permeabilized with digitonin) or a peroxisomal membrane protein with at least some epitopes exposed to the cytoplasm (if detected in cells permeabilized with either agent). The ability to detect PEX11␤myc and my-cPEX11␤ under digitonin permeabilization conditions suggests that PEX11␤ is a peroxisomal membrane protein with its termini exposed to the cytoplasm and provides experimental sup- Note that peroxisome abundance is increased in the cell on the left, which expresses PEX11␤myc, relative to the cell on the right, which does not express PEX11␤myc. HepG2 cells expressing the NH 2 -terminally tagged mycPEX11␤ protein were processed in an identical manner and also displayed increased peroxisome abundance and co-localization of PMP70 (panel C) with the myc-tagged PEX11␤ (panel D). Bar ϭ 25 m. port for the predicted topology of PEX11␤ (see above). However, we have no explanation for why we were unable to detect the expression of PEX11␤myc or mycPEX11␤ in cells that were permeabilized with Triton X-100.
We also examined the activity and distribution of tagged derivatives of PEX11␣. Human fibroblasts were transfected with plasmids designed to express NH 2 -and COOH-terminally tagged versions of PEX11␣ and were processed for indirect immunofluorescence 2 days later. Both PEX11␣ proteins were expressed efficiently in human cells and were detected in approximately 30% of the transfected cell population. However, peroxisome proliferation was detected in less than 5% of the cells that expressed these proteins (data not shown). This result contrasted sharply with those obtained for PEX11␤, a protein that induced peroxisome proliferation in nearly all cells in which it was expressed (see above). Nevertheless, those rare cells that responded to the myc-tagged forms of PEX11␣ displayed an increase in peroxisome abundance (Fig. 6, A-D) similar to that induced by PEX11␤. Both tagged forms of PEX11␣ also colocalized with PMP70 under differential permeabilization conditions, data that support the hypothesis that PEX11␣ extends its termini into the cytoplasm. It should also be noted that the tagged versions of PEX11␣ could be detected easily in cells permeabilized with Triton X-100, suggesting that the inability to detect the myc-tagged versions of PEX11␤ under these conditions reflected a specific property of PEX11␤.
PEX11␤ Drives Peroxisome Proliferation in a Multistep Process-To shed some light on the process of peroxisome proliferation that is mediated by PEX11␤, we examined peroxisomes at various times after transfection with the PEX11␤myc expres-sion vector. At the earliest time point, just 6 h after transfection, a dramatic change in peroxisome shape was apparent: the peroxisomes in cells expressing PEX11␤myc displayed an elongated, tubular morphology (Fig. 7, A and B). Quantitation of peroxisome morphology in cells transfected with PEX11␤myc or vector alone demonstrated that this shift in peroxisome morphology from vesicular structures to tubules was significant (Fig. 7C). Peroxisome tubules declined in abundance over the following 3 days and were replaced by more numerous, smaller vesicles, as shown earlier (Figs. 3 and 4). Interestingly, the peroxisome tubules that were detected at early stages of proliferation displayed a significant spatial heterogeneity in the distribution of PMPs, with PEX11␤myc and PMP70 segregating to discrete bands along the peroxisome (Fig. 8). Many of these peroxisomes contained alternating bands of PEX11␤myc and PMP70, with PMP70 often concentrated at peroxisome termini.
Peroxisome-proliferating Agents Induce PEX11␣ but Not PEX11␤-The above data demonstrated that overexpression of PEX11␤ alone can drive the proliferation of peroxisomes. The activity of this gene led us to speculate that changes in its expression might be responsible for drug-induced peroxisome proliferation in mammalian systems. This possibility was tested in rat liver, the best model system for studying druginduced peroxisome proliferation. RNA was extracted from livers of rats that had been fed control diets or diets supplemented with either of the peroxisome proliferators, clofibrate or DEHP. Message RNA abundance was inferred from Northern blot experiments using the murine PEX11␣ and PEX11␤ cDNAs as probes. The PEX11␣ transcript was induced more than 10-fold FIG. 6. PEX11␣ is a peroxisomal membrane protein with its NH 2 and COOH termini exposed to the cytosol. HepG2 cells were transfected with plasmids designed to express the COOH-terminally myc-tagged form of PEX11␣ (panels A and B) or an NH 2 -terminally tagged form of PEX11␣ (panels C and D). 2 days after transfection the cells were processed for indirect immunofluorescence under differential permeabilization using antibodies specific for PMP70 (panels A and C) and the myc epitope tag (panels B and D). Bar ϭ 25 m. by either DEHP or clofibrate, whereas the PEX11␤ mRNA was unaffected by these drugs (Fig. 9). These results indicate that peroxisome-proliferating agents may act via PEX11␣ but are unlikely to act through PEX11␤, at least at the transcriptional level. It should also be noted that we have yet to establish directly whether the products of the rat PEX11␣ and PEX11␤ genes are involved in peroxisome proliferation.
To characterize further the expression of the PEX11␣ and PEX11␤ genes, we next performed Northern analysis on mRNAs from multiple rat tissues. These experiments revealed that the abundance of PEX11␣ mRNA varied significantly among different tissues, with highest levels in kidney, significant expression in liver, lung, brain and testis, and very low levels in heart, spleen and skeletal muscle (Fig. 10A). In contrast, PEX11␤ was expressed at roughly equivalent levels in all of these tissues (Fig. 10B). Similar loading and transfer of mRNA samples were confirmed by examination of actin mRNA abundance (Fig. 10C). Given that the specific activities of the PEX11␣ and PEX11␤ probes were similar and that the above exposures were for identical lengths of time, we infer that PEX11␤ is expressed at higher levels than PEX11␣ in many tissues. There are approximately three times as many expressed sequence tags for PEX11␤ in the data base of expressed tags as there are for PEX11␣, a result that is consistent with this hypothesis.

DISCUSSION
Under normal conditions, peroxisome proliferation is induced by extracellular signals and associated with altered expression of a wide array of genes. As a result, it has been difficult to determine whether increases in peroxisome abundance are mediated by altered expression of just a single gene or a set of multiple genes. Previous studies have identified several genes, including PEX10 (18) and PEX11 (5,9), which can augment proliferation when overexpressed in the context of a peroxisome proliferation event. However, in no instance has overexpression of these genes been shown to induce peroxisome proliferation in the absence of appropriate extracellular stimuli. Our observation that overexpression of human PEX11␤ alone can efficiently induce peroxisome proliferation demonstrates that altered expression of a single gene can indeed mediate peroxisome proliferation, even in the absence of extracellular stimuli. In contrast to our results with human PEX11␤, peroxisome proliferation was detected in less than 5% of the cells that overexpressed PEX11␣. However, the difference in peroxisome proliferation-promoting activity which we observed may reflect a limitation of our experimental system rather than a significant biochemical difference between these proteins.
The fact that myc-tagged versions of PEX11␤ retained peroxisome proliferation-promoting activity allowed us to follow the effect of PEX11␤ expression on peroxisome morphology and abundance over time as well as to gain some insight into the subcellular distribution of PEX11␤ itself. Tagged versions of PEX11␤ behave as peroxisomal membrane proteins and can be cells were processed for double indirect immunofluorescence using antibodies specific for PMP70 (panel A) and PEX11␤myc under differential permeabilization conditions (panel B). Bar ϭ 2 m. Panel C, peroxisome morphology at various times after transfection with the PEX11␤myc expression vector (open squares) or with vector alone (solid circles). The percentage of peroxisomes with the elongated, tubular morphology is presented on the x axis, and the time after transfection is presented on the y axis. The slight increase in the abundance of peroxisome tubules observed in the control population was expected because earlier studies have established that trypsinization alone stimulates the formation of peroxisome tubules and mild peroxisome proliferation (19 -21).
FIG. 7. PEX11␤myc induces peroxisome elongation before proliferation. HepG2 cells were transfected with the PEX11␤myc expression vector and seeded onto glass coverslips. 6 h after transfection the detected in peroxisomes just 6 h after transfection. The initial consequence of PEX11␤ overexpression is the conversion of peroxisomes from spherical vesicles into elongated tubules. This process is also associated with the generation of subdomains in the peroxisome membrane, some of which are enriched for PEX11␤, whereas others are enriched for PMP70. The formation of peroxisome tubules was followed by the appearance of numerous, small, vesicular peroxisomes, a conversion that was almost complete by 72 h after transfection. These observations suggest a precursor-product relationship between the tubules that are formed early after transfection and the more numerous vesicles detected at later time points. It should be noted that this general model for peroxisome division has been proposed to explain the proliferation of peroxisomes which is induced by the passaging of HepG2 cells (19 -21). Additional time course experiments, ultrastructural studies, and real time analysis of peroxisome proliferation will be required to test this model rigorously as well as to determine whether the type of peroxisome division we observed occurs by peroxisome fission or vesicle budding.
Although it is important to elucidate the morphological steps in peroxisome division, a molecular analysis of PEX11␤ activity is also required. PEX11␤ does not share any sequence similarity or motif with any proteins other than PEX11 homologs from FIGS. 9. Effect of peroxisome-proliferating agents on PEX11␣ and PEX11␤ expression. 10 g of total rat liver RNA from animals fed a control chow diet (lanes 1 and 4), a diet supplemented with DEHP ( lanes 2 and 5), and a diet supplemented with clofibrate (lanes 3 and 6) was separated by denaturing agarose gel electrophoresis, transferred to nylon membranes, and probed with a radioactively labeled fragment of the PEX11␣ cDNA to detect the 2.4-kilobase PEX11␣ mRNA (upper panel, lanes 1-3) and the PEX11␤ cDNA to detect the 2- kilobase  PEX11␤ mRNA (upper panel, lanes 4 -6). The abundance of 28 S rRNA in each lane is shown separately in the lower panels. other species. Thus, there is no clear picture of how PEX11␤ may mediate peroxisome proliferation. Interestingly, rat PEX11␣, which shares the same topology in the peroxisome membrane as human PEX11␣ and PEX11␤, recruits ARF and COPI to peroxisome membranes (22). ARF and COPI recruitment appears to occur at the dilysine motif, KXKXX-COOH , which is present at the COOH terminus of PEX11␣. These and other observations have led Passreiter et al. (22) to propose that PEX11␣ may mediate a coat-dependent budding of peroxisomes from pre-existing peroxisomes. Human and mouse PEX11␣ share the same COOH-terminal tail as rat PEX11␣, suggesting that they may function in a similar manner. However, both human and mouse PEX11␤ lack the dilysine motif at their COOH terminus and instead terminate in the sequence RXKX-COOH . It is interesting to note that the dilysine motif is also absent from yeast PEX11.
The existence of two mammalian PEX11 genes alone indicated that these genes may have distinct roles. The data of this report and of Passreiter et al. (22) provide additional support for this hypothesis. PEX11␣ appears to interact with ARF/ COPI and stimulate vesicle budding from peroxisomes (22). It also displays tissue-specific and drug-induced changes in expression but exhibits relatively weak peroxisome proliferationpromoting activity in our experimental system. Conversely, PEX11␤ lacks the sequence motifs common to ARF/COPI-binding proteins. Furthermore, PEX11␤ is expressed at similar levels in all tissues examined, is unresponsive to peroxisome proliferator drugs, and exhibits robust peroxisome-proliferating activity in our experimental system. These observations are consistent with a model in which PEX11␤ controls constitutive peroxisome division and PEX11␣ regulates peroxisome abundance in response to dietary, hormonal, or other stimuli. However, additional experiments will be required to test this hypothesis directly.