Trypanosoma brucei Pex13.2 Is an Accessory Peroxin That Functions in the Import of Peroxisome Targeting Sequence Type 2 Proteins and Localizes to Subdomains of the Glycosome.

Trypanosoma brucei causes human African trypanosomiasis and a wasting disease called Nagana in livestock. Current treatments are expensive, toxic, and difficult to administer. Because of this, the search for new drug targets is essential. T. brucei has glycosomes that are essential to parasite survival; however, our ability to target them in drug development is hindered by our lack of understanding about how these organelles are formed and maintained. This work forwards our understanding of how the parasite-specific protein Pex13.2 functions in glycosome protein import and lays the foundation for future studies focused on blocking Pex13.2 function, which would be lethal to bloodstream-form parasites that reside in the mammalian bloodstream.

T rypanosoma brucei is a protozoan parasite that causes African trypanosomiasis (HAT) in humans and a wasting disease called Nagana in cattle. The parasite alternates between a mammalian host where it spends much of its time in the bloodstream as a bloodstream-form (BF) parasite and the tsetse fly as a procyclic-form (PF) parasite. Glycosomes are essential, parasite-specific membrane-bound organelles whose composition and function change during development (1) and in response to the environment (2). Small molecules that interfere with the protein interactions that regulate glycosome function are lethal to parasites (3); thus, these organelles and the processes that regulate them are attractive drug targets.
Glycosomes are related to peroxisomes and share many components of the machinery that facilitate protein import into these organelles. Peroxisome and glycosome biogenesis are regulated by proteins called peroxins (Pexs) that govern organelle formation, proliferation, and degradation as well as protein import into the organelle (4)(5)(6). Peroxisome and glycosome import involve binding of soluble receptor proteins, either Pex5 or Pex7, to a targeting sequence in the cargo protein (7,8). Pex5 binds to a C-terminal tripeptide with the consensus sequence SKL called a peroxisome targeting sequence 1 (PTS1), while Pex7 binds to a less conserved N-terminal sequence termed PTS2 (4,9,10). The PTS1 and PTS2 receptor-cargo complexes both dock at the peroxisome membrane through interactions with the glycosome membrane proteins Pex13 and Pex14, which make up the import channel (11). After import, the receptors are recycled via a ubiquitination process involving Pex2, -10, and -12 (12).
Kinetoplastids are unique in that they have two Pex13s, which have been designated Pex13.1 and Pex13.2 (13,14). These proteins share low sequence identity with each other or with Pex13s from higher eukaryotes. In previous studies, Pex13.1 localized to glycosomes. Silencing of the protein in BF and PF parasites yielded parasites with glycosome protein import defects and lowered growth rates (14). Later, iterative database searches resulted in the identification of Pex13.2 (13). Silencing of this second Pex13 via RNA interference (RNAi) in BF parasites resulted in mislocalization of Pex14 and aldolase and a defect in the growth rate. Prior to our work, Pex13.2 RNAi cell lines could not be established in PF parasites (13). Because glycosomes harbor a majority of the proteins involved in glycolysis, which is essential in BF, it is difficult to study Pex13.2 function in that life stage, as many of the expected phenotypes are likely lethal.
Here, we have partially resolved the topology of Pex13.1 and Pex13.2, identified several import protein complexes in PF parasites, and characterized Pex13.2-deficient PF cell lines. Results herein indicate that Pex13.2 interacts with known proteins of the protein import channel that form several high-molecular-weight membrane complexes and is essential for the efficient import of PTS2 sequences.

RESULTS
Pex13.2 is an integral glycosomal membrane protein with its N terminus exposed to the cytosol. To define Pex13.2 localization in PF parasites, we used immunofluorescence assays of cells expressing Pex13.2 fused to a myc epitope tag (myc.Pex13.2). Antibodies against the glycosome protein aldolase and myc both labeled punctate structures consistent with glycosome staining (Fig. 1A). To complement these microscopy studies, we took a biochemical approach to resolve Pex13.2 localization. We isolated organelles via density gradients and analyzed fractions by Western blotting with antibodies made against recombinant Pex13.2, the endoplasmic reticulum (ER) protein BiP, and aldolase. The ER is less dense than glycosomes and equilibrates higher on the gradient. As expected for a glycosome protein, Pex13.2 was detected in fractions 14 to 20 also containing aldolase (Fig. 1B), while BiP was detected in fractions 20 to 32.
Pex13.2 is predicted to have at least 2 transmembrane (TM) domains (13,15). To confirm that it is an integral membrane protein, we sequentially extracted membrane fractions with low salt, high salt, and sodium carbonate (Fig. 1C). The soluble glycosome matrix protein aldolase was detected in the low-salt supernatants. In contrast, Pex13.2 was detected only in the pellet following sodium carbonate extraction, consistent with being an integral membrane protein.
Most Pex13s in other organisms have both the N and C termini exposed to the cytoplasm (16)(17)(18), and previous work indicated that the C-terminal SH3 domain of Pex13.1 also faced the cytosol (14). The topology of Pex13.2 is unknown, and there is no information on whether the N terminus of Pex13.1 faces the cytoplasm or glycosome matrix. We used protease protection assays to assess the orientation of the N termini of both Pex13s. In these experiments, we used cells expressing N-terminally tagged Pex13s (myc.Pex13.2 and HA.Pex13.1) (Fig. 1D). Several attempts at tagging the C termini of either Pex13 were unsuccessful. Treatment of cells with proteinase K (PK) in the absence of Triton X-100 resulted in the loss of myc and hemagglutinin (HA) signal, indicating that the N termini of both proteins were accessible to PK and exposed to the cytosol (Fig. 1D). Pex13.1 antibodies were generated against the SH3 domain of the protein, which was shown to reside on the cytosolic face of the glycosome membrane (14). In our protease protection assays, the SH3 domain of HA.Pex13.1 was degraded by PK in the absence of Triton X-100, indicating that the topology of the Pex13.1 is not altered by addition of the HA epitope tag. As a control, we used aldolase, a matrix protein with a protease-resistant core (19). Full-length aldolase was detected in PK treatments without detergent, indicating that glycosome integrity was not compromised during treatment. After treatment with PK and detergent, we observed  a smaller proteolytic product as seen in previous studies documenting the proteaseresistant nature of aldolase (19). These results indicate that the N termini of Pex13.1 and Pex13.2 and the C terminus of Pex13.1 are on the cytosolic side of the glycosome membrane.
We next analyzed Pex13.1-, Pex13.2-, and Pex14-containing protein complexes in PF 2913 cells using 2-dimensional gel electrophoresis. Lysates were resolved in the first dimension by blue native gel electrophoresis and in the second dimension by SDS-PAGE. Western blotting revealed the presence of three complexes (Fig. 2B), where the largest complex (L) contained Pex13.1 and Pex13.2, the intermediate complex (I) contained Pex13.1, Pex13.2, and Pex14, and the smaller complex (S) contained primarily Pex14.
Silencing of Pex13.2 did not slow PF growth but affected glycosome size, number, and density. To gain insight into the specific function Pex13.2 plays in glycosome biogenesis, we generated PF Pex13.2-deficient parasites using tetracyclineinducible RNA interference (33). Upon addition of doxycycline (Dox), Pex13.2 expression was reduced 96.1% Ϯ 1.4% (Fig. 3A). Cells were counted daily after induction, and we observed no growth defects in Pex13.2-deficient cells. We next used transmission electron microscopy (TEM) to determine if glycosome number or size was altered upon depletion of Pex13.2 (Fig. 3B). Cells grown in the absence or presence of Dox had 13.5 glycosomes/100 nm 2 and 7.2 glycosomes/100 nm 2 , respectively. Cells grown without Dox had glycosomes with an average area of 0.20 m 2 , while those grown with Dox had an average area of 0.21 m 2 . Under standard culturing conditions in SDM79 medium (5 mM glucose), the disruption of normal glycosome function is often lethal, making phenotypic characterization difficult. However, when cells are grown in media lacking glucose, these lethal phenotypes can be rescued (20). We measured glycosome number and size in low-glucose medium (SDM79; 5 M glucose) with the expectation that under these conditions, we could score phenotypes that would be lethal in cells grown in high glucose. In low-glucose medium, the increase in glycosome size upon These results indicate that glycosome number and size were slightly but significantly affected (P values Ͻ 0.05) by Pex13.2 depletion, with induced cells having fewer but larger glycosomes. We next used density gradients to determine the density of glycosomes from uninduced and induced parasites (Fig. 3C). In uninduced cells, the glycosome proteins aldolase and Pex14 were detected in fractions 10 to 17, while the ER protein BiP was RNAi-induced cells grown in either SDM79 or the low-glucose medium SDM79. Glycosomes were measured for size (middle) as well as number of organelles (right). (C) Density gradient centrifugation was performed on both uninduced and induced cells; fractions were taken from the top of the gradients, and protein (2.5 g) was precipitated and analyzed by SDS-PAGE and Western blotting. Densitometry was used to quantify the protein signal in each fraction relative to the total protein signal (right). enriched in fractions 1 to 10. In induced cells, the majority of BiP was still detected in fractions 1 to 10, while the glycosome proteins aldolase and Pex14 were detected in higher fractions (4 to 13) of lower density. These results indicate that the glycosomes of Pex13.2-deficient parasites are lighter than those in cells harboring wild-type levels of Pex13.2.
Silencing Pex13.2 in PF parasites reduced import efficiency of PTS2 sequences. In higher eukaryotes, Pex13 functions in peroxisomal protein import, and silencing Pex13.2 in BF parasites compromises glycosome protein import (13). We used immunofluorescence assays (IFAs) to determine if Pex13.2 silencing altered glycosome protein import in PF parasites. In uninduced cells, the glycosome proteins aldolase, hexokinase (HK), fructose 1,6-bisphosphatase (FBPase), and phosphofructokinase (PFK) all localized to punctate structures characteristic of glycosomes (Fig. 4A). After 4 days of induction, IFA revealed an increase in the cytosolic staining of aldolase. In contrast, no change in staining of HK, FBPase, and PFK was observed.
To quantify the import efficiencies of each protein and identify mislocalization that is not detected by IFA, we used Western analysis. Lysates from uninduced and induced cells were centrifuged to obtain a membrane-rich fraction, and Western analysis and densitometry were used to calculate the percentage of each protein associated with membrane (glycosomes) and soluble (cytosolic) fractions (Fig. 4C)  cells. In contrast to that for aldolase, the percentages of FBPase, PFK, and HK localization associated with glycosome-enriched fractions did not change when Pex13.2 was silenced. While both HK and aldolase have type 2 peroxisomal targeting sequences (PTS2s), we were surprised that only aldolase localization was affected by Pex13.2 silencing. To determine if the differences in HK and aldolase localization in response to Pex13.2 depletion was due to the particular PTS2 sequence of each protein, we followed the localization of a reporter protein, enhanced yellow fluorescent protein (eYFP), fused to either the PTS2 of aldolase (AldoPTS2eYFP) or the PTS2 of HK (HKPTS2eYFP). Fluorescence microscopy of living cells revealed that both AldoPTS2eYFP and HKPTS2eYFP were mislocalized to the cytoplasm in Pex13.2-deficient cells (Fig. 4B), indicating that both PTS2s from HK and aldolase were affected by Pex13.2 depletion. Import experiments shown here were performed with cells grown in minimal glucose. When grown in high glucose, the mislocalization of aldolase and HKPTS2eYFP were observed but to a lesser degree (data not shown).
Pex13.2 partially colocalizes with Pex13.1 and is localized to discrete foci in glycosomes. IFAs, sucrose gradients, and membrane association experiments indicated that Pex13.2 is a glycosome membrane protein. Recent work in other systems has shown that components of the peroxisome import machinery localize to distinct regions of the peroxisome periphery (21). We used superresolution microscopy to determine if Pex13.2 exhibited similar localization. Antibodies against the matrix protein aldolase labeled punctate structures approximately 280 nm in diameter. In contrast, antibodies that recognize myc.Pex13.2 labeled smaller regions of the aldolasepositive structures, approximately 58 nm in diameter. We used Manders overlap coefficients (MOCs) to quantify the extent to which myc.Pex13.2 overlapped aldolase (Fig. 5E). MOC values range from 0 (no overlap) to 1 (complete overlap) (22,23). The MOC for aldolase and myc.Pex13.2 was 0.62 Ϯ 0.16, suggesting incomplete overlap. To better define the relationship between myc.Pex13.2 and aldolase staining, we calculated Manders M1 and M2 values, which reveal the extent to which pixels in one channel overlap the other. The M1 and M2 values were 0.49 Ϯ 0.20 and 0.91 Ϯ 0.10, respectively. These numbers indicate that 49% of the signal in channel 1 (aldolase) overlaps that in channel 2 (myc) and that 91% of the signal in channel 2 (myc) overlaps that in channel 1 (aldolase), and images reveal that myc.Pex13.2 is limited to a restricted portion of the aldolase-positive structures. Of the myc.Pex13.2-positive glycosomes, 37% had one focus, 35% had two foci, and 24% had 3 or more. Of aldolase-positive glycosomes, approximately 4% had no visible myc.Pex13.2 foci (Fig. 5G).
Coimmunoprecipitations indicated that HA.Pex13.1 and myc.Pex13.2 physically interact. Additionally, 2-dimensional gel electrophoresis suggested that these proteins form multiple, high-molecular-weight complexes. We next used superresolution microscopy to determine the extent to which these proteins overlap in cells (Fig. 4C and  D). Image analysis revealed that the proteins exhibit some colocalization, with an MOC of 0.77 Ϯ 10 (Fig. 4E) and M1 and M2 values of 0.49 Ϯ 0.13 and 0.95 Ϯ 0.03, respectively (Fig. 4F), indicating that 49% of Pex13.2 signal overlapped that of Pex13.1 while 95% of Pex13.1 overlapped Pex13.2. IFA revealed the presence of structures harboring HA.Pex13.1 but lacking detectable levels of myc.Pex13.2 in addition to Pex13.2 structures lacking detectable levels of HA.Pex13.1. These results indicate that the two proteins sometimes localize together but exhibit distinct localization.

DISCUSSION
Kinetoplastids have unique peroxisome-like organelles called glycosomes that are essential to parasite viability. Like peroxisomes, glycosomes have highly conserved protein machinery that facilitates the import of proteins into the organelle from the cytoplasm where they are synthesized. During the import process, the cargo protein is bound by either Pex5 or Pex7, which bind to a peroxisome targeting sequence 1 (PTS1) or PTS2, respectively. The receptor-cargo complex then docks at the membrane through interactions with Pex13 and Pex14. While eukaryotes have a single Pex13, kinetoplastid parasites are unique in that they have two Pex13s, Pex13.1 and Pex13.2.
In previous work (13), Pex13.1 localized to glycosomes with the C-terminal SH3 domain on the cytoplasmic side of the organelle. The silencing of the protein resulted in glycosome protein import deficiencies, growth defects, and mislocalization of Pex14, and yeast two-hybrid studies suggested that Pex13.1 interacted with Pex14 and Pex5 (13). Pex13.2 was identified later, and silencing it in BF parasites resulted in glycosome protein mislocalization and cell death (14). In those studies, PF RNA interference cell lines could not be obtained. In two of three yeast two-hybrid screens, Pex13.2 interacted with Pex13.1. Interactions between Pex14 and Pex13.2 were not detected, although the authors indicated they could not rule out technical reasons for this lack of interaction. While Pex13.1 and Pex13.2 are essential and nonredundant, it is unknown why kinetoplastids have two of these proteins when one is sufficient for most organisms. Because mislocalization of glycosome proteins is usually lethal, we hypothesize that Pex13.2 is necessary for the efficient import of proteins whose rapid sequestration in the glycosome is essential. Interference of glycosome biogenesis and protein import  kills parasites (24,25), and resolving the function of these novel Pex13s will likely reveal processes that can be exploited for pragmatic gain.
To understand the function of Pex13.2, we characterized the phenotype of Pex13.2deficient cells, partially resolved the topology of Pex13.2, demonstrated that it binds both Pex13.1 and Pex14, and identified three import complexes. The model kinetoplastid, Trypanosoma brucei, alternates between the tsetse fly and a mammalian host. Previous work in mammalian bloodstream-form (BF) parasites revealed that silencing of Pex13.2 disrupts glycosome protein import (14). Here, we have characterized Pex13.2deficient procyclic-form (PF) parasites that reside in the fly. This stage of the parasite is less sensitive to the disruption of glycosome function than BF, and we reasoned that we would observe phenotypes in this stage that are lethal in BF. Additionally, PF parasites are more tolerant of the expression of epitope-tagged glycosome proteins, which was required for these studies.
We found that Pex13.2 silencing resulted in mislocation of aldolase containing a PTS2, while the import of several PTS1 proteins was not affected. We were surprised to find that the localization of another PTS2 protein, HK, was not altered upon depletion of Pex13.2. There are two scenarios that may explain this observation. First, it may be that the two PTS2s differ in their relative strengths, with hexokinase PTS2 being a stronger targeting sequence that is less sensitive to Pex13.2 depletions than aldolase. There is precedence for such differences in PTS efficiencies. For example, replacement of the native low-efficiency PTS of Hansenula polymorpha catalase with a stronger high-efficiency PTS results in faster import of the protein from the cytoplasm. This increased import efficiency results in a decrease in overall catalase activity, as inactive aggregates are formed in the peroxisomes (26). It is hypothesized that the weaker native PTS allows more time for the protein to fold properly prior to import and is essential to maintaining high enzymatic activity. The finding that both HK PTS2 and aldolase PTS2 fused to the reporter eYFP exhibited similar mislocalization upon depletion of Pex13.2 suggests that the different behaviors of the two proteins is not a function of PTS2 efficiencies. We hypothesize that the lack of HK mislocalization in Pex13.2-deficient cells is the consequence of a cryptic PTS that targets HK to glycosomes when PTS2 targeting is disrupted. The presence of two independent PTSs has been demonstrated for the protein catalase A in Saccharomyces cerevisiae (27). Work is in progress to identify these cryptic targeting sequences.
Transmission electron microscopy revealed that Pex13.2-deficient cells had fewer but larger glycosomes, indicating that the protein is important for glycosome formation and/or proliferation as well as protein import. Glycosome morphology was not investigated in previous Pex13.2 studies (13,18). The finding that the N-terminal YG-rich region of Pex13.2 is exposed to the cytoplasm is significant, as it is this portion of the protein that interacts with Pex7 during the docking process. It is reasonable to propose that silencing of Pex13.2 reduces the number of Pex7 docking sites, which leads to compromised import of PTS2 proteins. Our findings are consistent with previous studies in which Pex13.2 silencing in BF parasites resulted in protein import defects. We are pursuing in vitro binding assays to determine if recombinant Pex7 binds the YG-rich region of Pex13.2. However, our inability to obtain large amounts of truncated Pex13.2 has hindered these efforts.
We demonstrated that Pex13.2 interacts with Pex13.1 and Pex14, known members of the docking complexes, further supporting its designation as a member of the glycosome protein import machinery. The observation that the association of Pex13.1 and Pex14 with the glycosome membrane is not disrupted in Pex13.2-deficient parasites suggests it is not essential to the formation of the primary Pex13.1/Pex14 import complex and is the basis for our designating it an accessory protein.
While Pex13.2 was essential in BF parasites (13), we did not observe a growth defect in Pex13.2-deficient PF parasites. There are two possibilities for these different outcomes. First, it is possible that there is a relaxed requirement for Pex13.2 in PF parasites. Because BF parasites are more sensitive to glycosome disruption, this scenario is not unreasonable. Second, the differences in growth phenotype may be a consequence of incomplete Pex13.2 silencing, where the remaining Pex13.2 is sufficient to support growth. While possible, this is unlikely as the protein levels were reduced ϳ96% upon induction.
There is limited information on the topology of Pex13.1 or Pex13.2. Previous work showed that the C terminus of Pex13.1 was on the cytoplasmic side of the glycosome (14). Our findings build a more complete picture of the import complex, revealing that the N termini of Pex13.1 and Pex13.2 are also on the cytoplasmic side of the glycosome. Unfortunately, multiple attempts to tag the C termini of Pex13.1 and Pex13.2 have been unsuccessful. While we cannot rule out a technical basis for this result, it may be that these regions are essential to critical aspects of Pex13 localization and/or function. The finding that Pex13.1, Pex13.2, and Pex14 interact in parasites validates and expands results obtained in yeast two-hybrid studies (13,14) where Pex13.1 interacted with Pex14 and Pex5 and Pex13.2 interacted with Pex13.1 and Pex19. In these studies, we demonstrated that those interactions occur in parasites and that Pex13.2 interacted with Pex14, a combination that could not be tested in previous work due to technical issues.
In other eukaryotes, several high-molecular-weight peroxisome import complexes have been characterized (28). Previous to our work, the number, size, and composition of the docking complexes in any kinetoplastid were unknown. Here, we resolved three complexes. One very-high-molecular-weight complex contained Pex13.1 and Pex13.2 but lacked detectable levels of Pex14. Similar large Pex13 complexes of unknown function have been identified in yeast and mammalian cells (28). To our knowledge, complexes containing Pex14 but lacking Pex13 have not been reported in eukaryotes. At this point, it is unknown whether any or all of these complexes support glycosome protein import.
Glycosome density was reduced in Pex13.2-deficient cells. This has been observed before in peroxisome mutants (29) and may be a function of reduced protein import or the result of having reduced amounts of Pex13.2 in the glycosome membrane. Mass spectrometry is necessary to discriminate between the two possibilities and determine if the import of other proteins is altered in Pex13.2-deficient cells.
We found that Pex13.2 localized to discrete regions of the glycosome periphery. Such asymmetric distribution of peroxins has been reported in other systems (21). Studies are ongoing to determine if other components of the import machinery exhibit similar suborganelle localization. This Pex13.2 localization pattern along with the reduction of glycosome number and the increase in size suggest that Pex13.2 may be involved in glycosome division. The ability to resolve individual organelles via superresolution microscopy is a critical advance that will allow us to identify residues required for this unique localization and assess the biological significance of this localization. Attempts to demonstrate this localization via immunoelectron microscopy has been unsuccessful, because antibodies that recognize these proteins do not work under fixation conditions required for this technology.
Here, we have identified three glycosome import complexes and demonstrated that Pex13.2 is a bona fide member of the docking complex in T. brucei and necessary for efficient import of PTS2 proteins. Current work is focused on resolving the composition and function of each complex. Because disruption of the protein interactions that facilitate glycosome protein import is toxic to T. brucei, a detailed understanding of composition and function of complexes that facilitate these processes is crucial to our ability to target them for therapeutic purposes.

MATERIALS AND METHODS
Cell culture and transfection of T. brucei. Procyclic-form (PF) 2913 and bloodstream-form (BF) 9013 expressing T7 polymerase and tetracycline (tet) repressor (30) were maintained in SDM79 (or the minimal glucose medium SDM79 containing 5 M glucose) (31) and HMI-9, respectively. Expression vectors for epitope-tagged proteins were generated by cloning the open reading frame of Pex13.2 into the pXS2 (PCF) vector or pXS6 (BSF) vectors possessing either a blasticidin resistance or puromycin resistance gene (32). For transfection, 20 g plasmid DNA was linearized (pXS2, pXS6: MluI; pZJM: NotI) and electroporated in 4-mm cuvettes (Bio-Rad GenePulser Xcell; exponential, 1.5 kV, 25 F). Twenty-four hours after electroporation, culture medium was supplemented with the appropriate drug for selection: 15 g/ml G418, 50 g/ml hygromycin, 2.5 g/ml phleomycin, 1 g/ml puromycin, or 10 g/ml blasticidin. RNA interference (RNAi) cell lines were generated by cloning nucleotides 41 to 441 of TbPex13.2 into the inducible pZJM vector possessing dual opposing T7 promoters and a phleomycin resistance marker (33). Usually, we grow RNAi cell lines in tet-free medium to reduce leaky expression from RNAi plasmids. However, this was not required for these cell lines, as we did not observe leaky expression.
Pex13.2 antibody production. Polyclonal guinea pig antiserum was generated against truncated recombinant Pex13.2 (Thermo Scientific). Amino acids 2 to 150 of Pex13.2 fused to an N-terminal His 6 tag were expressed using the pQE30 expression system (Qiagen) and purified using a nickel-nitrilotriacetic acid (Ni-NTA) column under denaturing conditions using 8 M urea as described (34) for use as antigen.
Growth curves. Cells possessing the tetracycline inducible pZJM:Pex13.2 vector were seeded at 10 5 cells/ml in SDM79 (PF) and induced with 1 g/ml doxycycline. Cells were allowed to grow to a density of 5 ϫ 10 6 cells/ml prior to passing back to 1 ϫ 10 5 . Culture density was monitored by flow cytometry at 24-h intervals using an Accuri C6 flow cytometer (BD Biosciences).
Protease protection assays. Protease protection assays were carried out using a modified protocol previously described (16). Cells (10 6 ) were harvested at 800 ϫ g for 10 min, washed once in PBS (150 mM NaCl, 1 mM KH 2 PO 4 , 5.6 mM Na 2 HPO 4 , pH 7.4), once in STE buffer (250 mM sucrose, 25 mM Tris-HCl [pH 7.4], 1 mM EDTA), and resuspended in 98 l ice-cold STEN buffer (STE buffer supplemented with 150 mM NaCl) and 0.1 mM phenylmethylsulfonyl fluoride (PMSF). Cells were permeabilized with 2 l 1 mg/ml digitonin (final concentration of 0.02 mg/ml), vortexed for 5 s, and incubated at room temperature for 4 min. Following permeabilization, cells were centrifuged at 20,000 ϫ g for 2 min and resuspended in 85 l STEN buffer. Pellets were treated with either 10 l water or Triton X-100 (1% [vol/vol] final) and either 5 l water or 2 mg/ml proteinase K. Reaction mixtures were incubated on ice for 30 min and stopped by addition of 10% (wt/vol) trichloroacetic acid (TCA). Precipitated proteins were centrifuged at 17,000 ϫ g for 10 min and washed once with acetone before being resuspended in cracking buffer (CB; 10% glycerol, 2% SDS, 2% ␤-mercaptoethanol, 100 mM Tris [pH 6.8], 0.1% bromophenol blue) and boiled at 100°C. Proteins were then analyzed by SDS-PAGE and Western blotting.
Membrane association assays. Membrane association assays were carried out as previously described (16). For extraction of membrane proteins, 10 7 cells were centrifuged at 800 ϫ g for 10 min and resuspended in 300 l of ice-cold low-salt buffer for 15 min (5 mM Tris-HCl [pH 7.8], 1 mM EDTA, 0.1 mM PMSF, 4 g/ml leupeptin). Cells were then passed through a pipette tip 10ϫ and centrifuged at 20,000 ϫ g for 30 min at 4°C. The insoluble pellet was resuspended in 300 l high-salt buffer (25 mM Tris-HCl [pH 7.8], 0.5 M KCl, 1 mM EDTA, 0.1 mM PMSF, 4 g/ml leupeptin) and incubated on ice for 15 min. After incubation, samples were centrifuged again at 20,000 ϫ g for 30 min at 4°C. The insoluble pellet was resuspended in 300 l 0.1 M Na 2 CO 3 and incubated for 30 min on ice. Samples were then centrifuged at 120,000 ϫ g for 1 h at 4°C with a 500-l cushion of 0.1 M Na 2 CO 3 , 0.25 M sucrose in a Beckman TLA100.3 rotor. Supernatant protein was precipitated by 10% (wt/vol) TCA and washed once with acetone before being resuspended in CB. Samples were then separated by SDS-PAGE and analyzed by Western blot.
Live-cell microscopy. Cells expressing either AldoPTS2eYFP or HKPTS2eYFP were washed once with PBS, mounted on a slide, and visualized using a Zeiss Axiovert 200M inverted fluorescence microscope with a 100ϫ lens objective (numerical aperture [NA] 1.3). Images were analyzed using AxioVision software version 4.8.2.