Glyceraldehyde-3-phosphate dehydrogenase interacts with Rab2 and plays an essential role in endoplasmic reticulum to Golgi transport exclusive of its glycolytic activity.

Rab2 requires atypical protein kinase C iota/lambda (aPKC iota/lambda) to promote vesicle formation from vesicular tubular clusters (VTCs). The Rab2-generated vesicles are enriched in recycling proteins suggesting that the carriers are retrograde-directed and retrieve transport machinery back to the endoplasmic reticulum. These vesicles also contained the glycolytic enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH). We have previously established that GAPDH is required for membrane transport between the endoplasmic reticulum and the Golgi complex. Moreover, GAPDH is phosphorylated by aPKC iota/lambda and binds to the aPKC iota/lambda regulatory domain. In this study, we employed a combination of in vivo and in vitro assays and determined that GAPDH also interacts with Rab2. The site of GAPDH interaction was mapped to Rab2 residues 20-50. In addition to its glycolytic function, GAPDH has multiple intracellular roles. However, the function of GAPDH in the early secretory pathway is unknown. One possibility is that GAPDH ultimately provides energy in the form of ATP. To determine whether GAPDH catalytic activity was critical for transport in the early secretory pathway, a conservative substitution was made at Cys-149 located at the active site, and the mutant was biochemically characterized in a battery of assays. Although GAPDH (C149G) has no catalytic activity, Rab2 recruited the mutant protein to membranes in a quantitative binding assay. GAPDH (C149G) is phosphorylated by aPKC iota/lambda and binds directly to Rab2 when evaluated in an overlay binding assay. Importantly, VSV-G transport between the ER and Golgi complex is restored when an in vitro trafficking assay is performed with GAPDH-depleted cytosol and GAPDH (C149G). These data suggest that GAPDH imparts a unique function necessary for membrane trafficking from VTCs that does not require GAPDH glycolytic activity.

and function as transport intermediates delivering cargo from the endoplasmic reticulum (ER) to the Golgi complex (3). VTCs are the first site of segregation of the anterograde and retrograde paths and thereby perform a critical function in the early secretory pathway (4). The retrieval of membranes from VTCs back to the ER is essential to preserve organelle identity and allows reuse of the transport machinery required for ER export. Rab2 bound to a VTC subcompartment promotes recruitment of soluble components that facilitate membrane fission and cytoskeletal interaction resulting in the release of retrogradedirected vesicles (5)(6)(7). One of the components recruited by Rab2 is atypical protein kinase C / (aPKC/), which binds to the N terminus of Rab2 (8,9). aPKC/ kinase activity is essential for Rab2-mediated retrograde vesicle formation (8). The Rab2-generated vesicles also contain glyceraldehyde-3-phosphate dehydrogenase (GAPDH). In fact, GAPDH recruitment to VTCs is required for membrane trafficking between the ER and Golgi complex (10). Moreover, GAPDH is a substrate for aPKC/ and interacts directly with aPKC/ via the regulatory domain (7). GAPDH catalyzes the NAD-mediated oxidative phosphorylation of glyceraldehyde 3-phosphate (G3P) to 1,3diphosphoglycerate. However, it has become increasingly clear that this enzyme has multiple intracellular activities in addition to its role in glycolysis. These activities include promoting membrane-membrane fusion, modulation of the cytoskeleton, DNA repair, DNA replication, and tRNA export (11). It has been suggested that these diverse GAPDH functions may be dictated by the subcellular location of the enzyme that allows association with compartment specific proteins including posttranslational modifying enzymes resulting in the many GAPDH isoforms that have been detected (7,(12)(13)(14). These modifications could potentially impart a new activity to GAPDH and allow it to switch function.
Since Rab2 interacts with aPKC/ and aPKC/ associates with GAPDH, it seemed plausible that Rab2 might also directly bind GAPDH. Indeed, we found using a combination of in vivo and in vitro assays that Rab2 associates with GAPDH suggesting that Rab2, aPKC/, and GAPDH form a complex on VTCs. GAPDH binds to residues 20 -50 in Rab2. This Rab2 sequence includes the putative switch I domain that undergoes guanine nucleotide-dependent conformational changes recognized by Rab regulatory factors.
Although GAPDH interacts with Rab2 and aPKC/ on VTCs and is required for ER to Golgi transport, the precise function of GAPDH in the early secretory pathway is unknown. GAPDH associated with VTCs could ultimately provide energy in the form of ATP, a factor essential for membrane trafficking. To address the question whether GAPDH catalytic activity was critical for transport in the early secretory pathway, a conservative substitution was made at the invariant residue (Cys-149) located at the active site. As expected, GAPDH (C149G) was inactive when evaluated in a glycolytic activity assay. GAPDH (C149G) was further characterized biochemically to be assured that the mutation did not affect other physiological properties in addition to loss of enzyme activity. By employing a quantitative microsomal binding assay we found that Rab2 recruited the mutant protein to membranes and that the membraneassociated form was the tetrameric species. Moreover, GAPDH (C149G) binds directly to Rab2 and aPKC/. Importantly, purified recombinant GAPDH (C149G) rescued VSV-G transport between the ER and Golgi complex when an in vitro trafficking assay was supplemented with GAPDH-depleted cytosol. These combined results suggest that GAPDH provides a specific function essential for membrane trafficking from VTCs that does not require GAPDH glycolytic activity.

EXPERIMENTAL PROCEDURES
Materials-The human expression-tested clone GAPDH pDual was purchased from Stratagene (La Jolla, CA). Rab2 N-terminal peptides were synthesized at the University of Michigan Protein and Carbohydrate Structure Facility (Ann Arbor, MI). The Mammalian Matchmaker two-hybrid assay kit was purchased from Clontech Laboratories, Inc. (Palo Alto, CA).
Construction of Rab2 N-terminal Deletion Mutants, Rab2-Rab1 Chimeras, and GAPDH Catalytically Inactive Mutant-N-terminal deletions of human Rab2 were generated as previously described (9). Reciprocal exchanges between human Rab2 and rat Rab1b were made by a two-step PCR procedure involving two complementary overlapping oligonucleotides in combination with flanking 5Ј-and 3Ј-primers for either Rab2 or Rab1. To generate the Rab1-(1-52)/Rab2-(50 -211) chimera, the first reaction was performed with overlapping 5Ј-and 3Јfragments using Rab1 as the template. The 5Ј sense primer (5Ј-GCCA-CATATGAACCCCGAATATGACTAC-3Ј) and the 3Ј antisense primer (5Ј-TTTTATCTGTTTCCCATCAATCAGTTCAATGGTTCGAATCTT-3Ј) were used to amplify the first 156 base pairs of Rab1. The 5Ј-portion of Rab2-(50 -211) was generated using Rab2 as the template and the overlapping primer; 5Ј-AAGAATCGAACCATTGAACTGATTGAT-GGGAAACAGATAAAA-3Ј in tandem with the 3Ј antisense primer that included an EcoRI site; 5Ј-GGCAGAATTCGTCAACAGCAGC-CCCCAGC-3Ј. The two PCR products were combined to generate the full-length mutant in a second reaction using the Rab1 sense primer and Rab2 antisense primer. To generate the Rab2-(1-50)/Rab1-(52-201) chimera, the first reaction was performed with overlapping 5Ј-and 3Ј-fragments using Rab2 as the template. The 5Ј sense primer (5Ј-GG-CCATGGCGGTACGCCTATCTCTTCAAGTAC-3Ј) and the 3Ј antisense primer (5Ј-GATTTGATGGTTTTGCCATCCAGAATAGTTATCATTCG-AGCACCCTT -3Ј) were used to amplify the first 150 base pairs of Rab2. The 5Ј portion of Rab1 (52-201) was generated using Rab1 as the template and the overlapping primer; 5Ј-AAGGGTGCTCGAATGATA-ACTATTGTCGATGGCAAAACCATCAAACT-3Ј in tandem with the 3Ј antisense primer that included an EcoRI site; (5Ј-GGCAGAATTCGCT-AGCAGCAGCCACCACT-3Ј). The two PCR products were combined to generate the full-length mutant in a second reaction using the Rab2 sense primer and Rab1 antisense primer. The chimeras were subcloned into pGEX2T (Amersham Biosciences) and into the GAL4 DNA binding domain in the pM vector as described below. The GAPDH catalytically inactive mutant (C149G) was made by a two-step PCR procedure involving two complementary mutagenic oligonucleotides in combination with flanking 5Ј-and 3Ј-primers. In the first reaction, overlapping 5Јand 3Ј-fragments were produced using pDual GAPDH as the template. The 5Ј mutagenic oligonucleotide (5Ј-CAGCAATGCCTCCGGGACCAC-CAACTGCTTAGCACC-3Ј) and the 3Ј wild-type antisense oligonucleotide (5Ј-GGATCCTTACTCCTTGGAGGCCATGTGGGCCATGAGGTC-CAC-3Ј) were used to produce the 3Ј-portion of the molecule. The 5Јportion of the molecule was generated using either a 5Ј-oligonucleotide primer that included an NdeI or a BamHI site (5Ј-GGCCATATGGGG-AAGGTGAAGGTC-3Ј) in combination with the 3Ј-mutagenic oligonucleotide (5Ј-GGTGCTAAGCAGTTGGTGGTCCCGGAGGCATTGCTG-3Ј). The two PCR products were combined to generate the full-length mutant in a second reaction using the respective 5Ј-sense and 3Јantisense primers. GAPDH wild type and GAPDH (C149G) were subcloned into pET15B (Novagen, Madison, WI) and pGEX2T. The sequence of all constructs was verified by automated DNA sequence analysis.
Immunoprecipitation-HeLa cells (2 ϫ 10 6 ) were lysed in 50 mM Tris-buffered saline (TBS) (pH 8.0) and 1% Triton X-100 for 10 min on ice, and then clarified by centrifugation at 100,000 ϫ g for 60 min at 4°C. The soluble fraction was precleared with protein G plus/protein A-agarose (Novagen) and subjected to immunoprecipitation for 4 h at room temperature with an affinity-purified anti-Rab2 polyclonal antibody (10 g) or with preimmune serum and protein G plus/protein A-agarose. The immune complexes were collected by centrifugation at 5,000 rpm for 5 min, washed 3ϫ with TBS, 1% Triton X-100, and 100 mM NaCl, and then boiled in sample buffer. The immunoprecipitates were separated by SDS-PAGE and transferred to nitrocellulose. The blot was blocked in TBS that contained 5% BSA and 0.5% Tween 20, incubated with an anti-Rab2 polyclonal antibody and an anti-GAPDH monoclonal antibody (Chemicon International, Temecula, CA), washed, further incubated with a horseradish peroxidase (HRP)-conjugated anti-rabbit or an anti-mouse antibody, and then developed with enhanced chemiluminescence (ECL) (Amersham Biosciences).
Mammalian Two-hybrid Assay-Human Rab2, Rab2/Rab1, and Rab1/Rab2 cDNA was cloned in-frame to the EcoRI site of the GAL4 DNA binding domain in the pM vector and GAPDH cDNA cloned in-frame to the EcoRI site of the activation domain in pVP16 (Clontech Laboratories, Inc.). The two constructs (5 g each) were co-transfected with the reporter vector pG5CAT (5 g) that contains the chloramphenicol acetyltransferase (CAT) gene into HeLa cells (10 5 ) using a calcium phosphate transfection protocol (15). Control cells were transfected with a combination of vectors as described under "Results." The cells were collected 72 h post-transfection, lysed in sample buffer, separated by SDS-PAGE, and then transferred to nitrocellulose. The blot was blocked as described above, incubated with an anti-CAT polyclonal antibody (Invitrogen Life Technologies), washed, further incubated with an HRP-conjugated secondary antibody, and then developed with ECL.
GST Pull-down-BL21 (DE3) pLysS cells (Novagen) that contained the recombinant plasmids GST-GAPDH, GST-Rab2/Rab1, GST-Rab1/ Rab2, or GST-Rab1 were grown at 37°C to an OD 600 of 0.6 and then induced with 0.4 mM isopropyl-␤-D-thiogalactopyranoside for 3 h at 37°C. The liquid culture was centrifuged at 6,000 rpm for 30 min, and the pellet resuspended in cold TBS, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, sonicated, centrifuged at 12,000 rpm for 20 min, and then the supernatant applied to a glutathione-Sepharose 4B column (Amersham Biosciences). The column was washed with 10-bed volumes of TBS, and the fusion protein eluted with 5 mM reduced glutathione. GST-GAPDH, GST-Rab2-(1-50), GST-Rab2-(51-211), or GST (5 g) were preincubated with 20 l of glutathione Sepharose 4B for 1 h at room temperature, collected by centrifugation at 4,000 rpm for 5 min, and then washed 3ϫ with TBS and 1% Triton X-100 to remove any unbound protein. The beads were resuspended in 50 mM Tris (pH 7.5), 5 mM MgCl 2 , 100 mM NaCl, 10 M GTP␥S, or 10 M GDP (ICN Biochemicals Inc., Irvine, CA), and then 5 g of either Rab2 or His 6 -GAPDH added and incubated for an additional 2 h at room temperature. The beads were washed 4ϫ with 1 ml of 50 mM Tris (pH 7.5), 10 mM MgCl 2 , and 100 mM NaCl, and the bound proteins boiled in sample buffer, separated by SDS-PAGE, and then transferred to nitrocellulose. The blot was blocked as above and the membrane probed with an anti-GAPDH monoclonal antibody (Chemicon Intl.) or an affinity-purified anti-Rab2 polyclonal antibody, or a GST tag monoclonal antibody (Novagen), washed, incubated with an HRP-conjugated secondary antibody, and then developed with ECL.
ELISA-ELISA was performed as described previously (9, 16). Briefly, Nunc immunomodules were coated with peptides and purified Rab2 N-terminal-truncated recombinant proteins (1 g/100 l of 50 mM NaHCO 3 , pH 9.6) listed in Table I at 4°C overnight. The wells were washed in TBS, blocked in TBS and 5% fetal bovine serum for 1 h at 37°C, additionally washed in TBS, and then incubated with 5 g of purified recombinant His 6 -GAPDH for 3 h at 37°C. After each well was washed 3ϫ with TBS, a monoclonal antibody to GAPDH (Chemicon) in 50 mM Tris (pH 7.5), 5 mM MgCl 2 , 100 mM NaCl, 10 M GTP␥S was added for 2 h at 37°C, washed, and further incubated with an antimouse alkaline phosphatase-conjugated secondary reagent for 1 h at 37°C. The wells were again washed with TBS, developed with Sigma FAST p-nitrophenyl phosphate (Sigma-Aldrich) and then read at 405 nm on a microplate reader.
Assay for GAPDH Activity-Purified recombinant His 6 -GAPDH or His 6 -GAPDH (C149G) (1 g) were incubated in 10 mM sodium pyrophosphate (pH 8.5), 20 mM sodium phosphate, 0.25 mM ␤-NAD (ICN Biomedicals, Inc., Aurora, OH), and 3 M dithiothreitol at 25°C for 5 min to establish a baseline. The reaction was initiated by the addition of 0.4 M glyceraldehyde-3-phosphate (Sigma-Aldrich), and the catalytic reduction of NAD to NADH was monitored at the indicated times described under "Results" by an increase in absorbance at 340 nm. The GAPDH units correspond to the mass (m) of GAPDH converted per min per mg. The mass of NADH generated was calculated using ⑀ ϭ 0.622 Analysis of Transport in Vitro-NRK cells were infected for 4 h with the temperature-sensitive VSV strain ts045, and then biosynthetically radiolabeled with 100 Ci Expre 35 S 35 S (specific activity 1175 Ci/mmol, PerkinElmer Life Science Products) for 10 min at the restrictive temperature (39.5°C) to maintain the VSV-G mutant protein in the ER. The cells were then perforated by swelling and scraping and employed in the ER to cis/medial Golgi transport assay as described (17). Transport reactions were performed in a final volume of 40 l in a buffer which contains 25 mM Hepes-KOH, pH 7.2, 75 mM KOAc, 2.5 mM MgOAc, 5 mM EGTA, 1.8 mM CaCl 2 , 1 mM N-acetylglucosamine, an ATP regeneration system (1 mM ATP, 5 mM creatine phosphate, and 0.2 international units of rabbit muscle creatine phosphokinase), 5 l of semi-intact cells (ϳ5 ϫ 10 7 cells/ml, ϳ25-30 g of total protein) resuspended in 50 mM Hepes-KOH, 90 mM KOAc (pH 7.2), and 5 l of rat liver cytosol depleted of GAPDH by batch absorbed on Blue Sepharose (Amersham Biosciences) followed by immunoprecipitation for 4 h at 4°C with 10 g of anti-GAPDH monoclonal antibody (Chemicon) and protein G plus/protein A-agarose. The reactions were supplemented with the indicated concentration of purified recombinant His 6 -GAPDH or His 6 -GAPDH (C149G), incubated at 32°C for a total of 60 min, and then transferred to ice to terminate transport. Membranes were pelleted by centrifugation, solubilized in buffer, and digested with endoglycosidase H (endo H). The samples were analyzed by SDS-PAGE and the fraction of ts045 VSV-G protein processed to the endo H-resistant forms quantitated by a Storm PhosphorImager (Amersham Biosciences).
Quantitative Membrane-binding Assay-HeLa cells were washed three times with ice-cold phosphate-buffered saline. The cells were scraped off the dish with a rubber policeman into 10 mM Hepes pH 7.2 and 250 mM mannitol, and then broken with 15 passes of a 27-gauge syringe. The broken cells were pelleted at 500 ϫ g for 5 min at 4°C, and the supernatant removed and centrifuged at 20,000 ϫ g for 20 min at 4°C. The resultant pellet containing ER, pre-Golgi, and Golgi membranes was washed with 1 M KCl in 10 mM Hepes (pH 7.2) for 15 min on ice to remove peripherally associated proteins, re-centrifuged at 20,000 ϫ g for 20 min at 4°C, and then the membrane pellet resuspended in 10 mM Hepes (pH 7.2) and 250 mM mannitol and employed in the binding reaction, as described previously (5,8). Membranes (30 g of total protein) were added to a reaction mixture, which contained 27.5 mM Hepes (pH 7.2), 2.75 mM MgOAc, 65 mM KOAc, 5 mM EGTA, 1.8 mM CaCl 2 , 1 mM ATP, 5 mM creatine phosphate, and 0.2 units of rabbit muscle creatine kinase. Recombinant Rab2 and His 6 -GAPDH (C149G) were added at the concentrations indicated under "Results," and the reaction mix incubated on ice for 10 min. GAPDH-depleted rat liver cytosol (50 g) and 2.0 M GTP␥S were added, and the reactions shifted to 37°C and incubated for 12 min. The binding reaction was terminated by transferring the samples to ice, and then centrifuged at 20,000 ϫ g for 10 min at 4°C. The membrane pellet was separated by SDS-PAGE and transferred to nitrocellulose. The blot was blocked as above, incubated with a His tag (27E8) monoclonal antibody (Cell Signaling Technology, Inc., Beverly, MA), washed, further incubated with an HRPconjugated anti-mouse antibody, developed with ECL, and then quantitated by densitometry.

Rab2
Interacts Directly with GAPDH-Our previous studies showed that Rab2 recruits GAPDH to membranes in a dose-dependent manner (10). To explore the possibility that Rab2-GAPDH physically associate, we employed a combination of in vivo and in vitro approaches. First, the detergent lysate from HeLa cells was incubated with preimmune serum or with an affinity-purified anti-Rab2 polyclonal antibody. Rab2 and GAPDH co-immunoprecipitated whereas no interaction was detected with preimmune serum as assessed by Western blot analysis of the immune complex (Fig. 1A). This is the same reagent that co-precipitates Rab2 and aPKC/ (9).
To further verify Rab2-GAPDH interaction in vivo, we employed a mammalian two-hybrid assay that was used previously to demonstrate Rab2-aPKC/ in vivo association (9). This system allows detection of transient and weak proteinprotein interactions. For this assay, Rab2 cDNA was used as bait and cloned into pM to generate a GAL4 DNA binding domain-Rab2 fusion protein whereas GAPDH cDNA that serves as prey was cloned into pVP16 to generate a VP16 activation domain-GAPDH fusion protein. HeLa cells were cotransfected with these two constructs and with a reporter plasmid that contains the CAT gene. If the two fusion proteins interact in vivo, transcription of the CAT reporter gene is activated. Three days post-transfection, the cells were lysed in sample buffer and the lysate separated on SDS-PAGE and immunoblotted with an anti-CAT polyclonal antibody. Cells co-transfected with pM-Rab2 and pVP16-GAPDH expressed ϳ20-fold higher level of CAT protein compared with controls cells indicating that Rab2 and GAPDH interacted in vivo (Fig.  1B, lane d).
This potential in vivo interaction was further evaluated in a blot overlay assay that is routinely used to identify Rab-interacting proteins (18). Rab2 was separated on SDS-PAGE, transferred to nitrocellulose, and the membrane incubated in overlay buffer supplemented with purified recombinant His 6 -GAPDH. Any GAPDH bound to Rab2 was detected after incubation with a monoclonal antibody to GAPDH followed by a secondary HRPconjugated antibody and development with ECL. Consistent with the in vivo results, we detected His 6 -GAPDH binding to Rab2 (Fig. 1C). This in vitro result was further confirmed in a GST pull-down experiment in which GST-GAPDH was first immobilized on glutathione Sepharose 4B, and then incubated with purified recombinant Rab2 in a buffer containing GTP␥S or GDP. After extensive washing, agarose beads containing GST-GAPDH retained Rab2-GTP␥S whereas minimal interaction was detected with Rab2-GDP (Fig. 1D). These collective results indicate that activated Rab2 interacts directly with GAPDH.
To map the GAPDH binding site in Rab2, we aligned the sequence of Rab2 with Rab1 and swapped the corresponding residues (1-50), and then evaluated the chimeras in the mammalian two-hybrid assay described above. Like Rab2, Rab1 is essential for trafficking in the early secretory pathway (1). Cells co-transfected with pM-Rab2/Rab1and pVP16-GAPDH expressed ϳ15-fold higher level of CAT protein compared with the reverse construct and the control suggesting that GAPDH interacts with the first 50 residues in Rab2 ( Fig. 2A).
The two chimeras were then evaluated in GST pull-down experiments. The bacterial produced fusion proteins were prebound to glutathione Sepharose 4B, and then incubated with His 6 -GAPDH. Fig. 2B shows that similar to the in vivo assay, GAPDH specifically associated with Rab2/Rab1. We made use of reagents present in the laboratory that were first employed to identify the aPKC/ binding domain in Rab2 to eliminate sequence and to begin defining the minimal residues required for GAPDH association with Rab2 (9). Using an ELISA to evaluate protein interaction, we learned that GAPDH binds to Rab2 between amino acids 20 and 50 (Table I). It is noteworthy that this Rab2 segment contains the putative effector domain that mediates interactions with specific accessory proteins (19,20).
GAPDH Glycolytic Activity Is Not Required for ER to Golgi Transport-Although GAPDH is essential for ER to Golgi transport, we did not know whether GAPDH catalytic activity was required for Rab2-mediated events at the VTC. To address this issue, a conservative substitution was made at the invariant residue (Cys-149) located at the active site and the key residue that binds substrate. The mutated and GAPDH wildtype cDNA was subcloned into a His 6 -tagged expression vector, and the recombinant proteins purified by affinity chromatography on Ni 2ϩ -NTA-agarose beads. We first assessed the enzymatic activity of His 6 -GAPDH and His 6 -GAPDH (C149G) by measuring NADH production. As we anticipated, substitution of Cys-149 resulted in an inactive GAPDH mutant protein that possessed no dehydrogenase activity (Fig. 3A).
To learn whether His 6 -GAPDH (C149G) was recruited to FIG. 1. Rab2 binds directly to aPKC/. A, immunoprecipitation. HeLa cell (2 ϫ 10 6 ) lysate (Input) was incubated with preimmune serum or with an affinity-purified anti-Rab2 polyclonal antibody and protein G plus/protein A-agarose for 4 h at room temperature. The immune complexes were collected by centrifugation, washed, and analyzed as described under "Experimental Procedures." The blot was probed with an anti-GAPDH monoclonal antibody and an affinity-purified Rab2 polyclonal antibody, washed, incubated with HRP-conjugated secondary antibodies, and then developed with ECL. B, mammalian two-hybrid assay. The lysates from HeLa cells transiently transfected for 72 h with a, pG5CAT, pM, and pVP16 (basal control), or b, pM-Rab2, pVP-16 and pG5CAT (pM control), or c, pM, pVP16-GAPDH, and pG5CAT (pVP16 control), or d, pVP16-PKC/, pM-Rab2, and pG5CAT (experiment) were separated on SDS-PAGE and transferred to nitrocellulose. The blot was probed with an anti-CAT polyclonal antibody, washed, further incubated with an HRP-conjugated secondary antibody, and then developed with ECL. Interaction of the two fusion proteins will result in transcription of the CAT reporter gene. Cells co-transfected with pM-Rab2 and pVP16-GAPDH expressed ϳ20-fold higher level of CAT protein compared with controls cells indicating that Rab2 and GAPDH interacted in vivo (C) overlay binding assay. Purified recombinant Rab2 (5 g) was separated by SDS-PAGE and transferred to nitrocellulose. The affinity-purified Rab2 antibody recognizes Rab2 by Western blot whereas a GAPDH monoclonal antibody does not cross-react with Rab2. An equal load of Rab2 was separated by SDS-PAGE, transferred to nitrocellulose, first incubated in renaturation buffer as described under "Experimental Procedures" and then incubated in binding buffer as described under "Experimental Procedures" for 4 h at room temperature. After incubation, the blot was washed with TBS, probed with a GAPDH monoclonal antibody, washed, further incubated with an HRPconjugated secondary antibody, and then developed with ECL. Control was incubated in the absence of His 6 -GAPDH. D, GST pull-down. GST-GAPDH (5 g) or GST (5 g) were preincubated with 20 l of glutathione-Sepharose 4B for 1 h at room temperature, and then washed 3ϫ with TBS 1% Triton X-100 to remove any unbound protein. The beads were resuspended in 50 mM Tris (pH 7.5), 5 mM MgCl 2 , 100 mM NaCl, 10 M GTP␥S, or 10 M GDP, and then 5 g of purified recombinant Rab2 added and incubated for an additional 2 h at room temperature. The beads were washed 4ϫ with 1 ml of 50 mM Tris, pH 7.5, 10 mM MgCl 2 , and 100 mM NaCl, bound proteins boiled in sample buffer, separated by SDS-PAGE, and then transferred to nitrocellulose. The blot was probed with an affinity-purified Rab2 polyclonal antibody and an anti-GST tag monoclonal antibody, washed, incubated with an HRPconjugated secondary antibodies, and then developed with ECL. Shown are representative results. All experiments were performed a minimum of three times.  membranes in response to Rab2, we made use of a quantitative binding assay. For this assay, microsomes were prepared from HeLa cell homogenates and washed with 1 M KCL to remove peripherally associate proteins including GAPDH. These membranes were preincubated in buffer for 10 min on ice in the presence of increasing concentrations of recombinant Rab2 and 100 ng His 6 -GAPDH (C149G). The reaction was supplemented with GAPDH-depleted rat liver cytosol and GTP␥S, and then incubated for 12 min at 32°C to promote binding of GAPDH (C149G) and other soluble components. The membranes were collected by centrifugation at 20,000 ϫ g and then analyzed by SDS-PAGE and Western blot for the presence of His 6 -tagged GAPDH (C149G). Rab2 efficiently recruited the GAPDH mutant to membranes in a dose-dependent manner (Fig. 3B). Moreover, GAPDH (C149G) interacts with Rab2 when evaluated in the overlay binding assay (Fig. 3C). A similar result was obtained when the overlay assay was performed with aPKC/ and GAPDH (C149G) (Fig. 3C).

FIG. 3. Biochemical properties of GAPDH (C149G).
A, purified His 6 -GAPDH wt (open square) and His 6 -GAPDH (C149G) protein (closed circle) (1 g) were incubated in buffer A as described under "Experimental Procedures" for 5 min at room temperature. The reaction was initiated by the addition of 0.4 M G3P. The catalytic reduction of NAD to NADH was monitored at the indicated times by an increase in absorbance at 340 nm. Results shown are representative of three independent experiments. B, salt-washed HeLa microsomes prepared as described under "Experimental Procedures" were preincubated with Rab2 and His 6 -GAPDH (C149G) at the concentrations indicated under "Results" and the reaction mix incubated on ice for 10 min. GAPDH-depleted rat liver cytosol (50 g) and 2.0 M GTP␥S were added, and the reactions shifted to 37°C and incubated for 12 min. To terminate the reaction, the membranes were collected by centrifugation, and the membrane pellet was separated by SDS-PAGE and transferred to nitrocellulose. The blot was probed with a His tag monoclonal antibody and an affinity-purified anti-Rab2 polyclonal antibody, washed, further incubated with HRP-conjugated secondary antibodies, and then developed with ECL. The amount of recruited His 6 -GAPDH (C149G) and Rab2 was quantitated by densitometry. A representative Western blot is shown. The results are the mean Ϯ S.D. of three independent experiments performed in duplicate. C, purified recombinant Rab2 (5 g) or aPKC/ (5 g) were separated by SDS-PAGE and transferred to nitrocellulose. The blot was incubated in buffer described under "Experimental Procedures" to renature the protein and then incubated in 12.5 mM Hepes/KOH (pH 7.2), 1.5 mM MgOAc, 75 mM KOAc, 0.1% BSA, 10 M GTP␥S, 200 mM NaCl, and 10 g/ml purified His 6 -GAPDH (C149G) for 4 h at room temperature. After incubation, the blot was washed with TBS, probed with a GAPDH monoclonal antibody, washed, further incubated with an HRP-conjugated secondary antibody, and then developed with ECL. Control was incubated in the absence of His 6 -GAPDH (C149G). D, purified recombinant (a) His 6 -GAPDH and (b) His 6 -GAPDH (C149G) were separated on a 10% Tris nondenaturing gel and then transferred to nitrocellulose. The blot was probed with an anti-GAPDH monoclonal antibody, washed, incubated with an HRP-conjugated secondary reagent, and developed with ECL. Protein standards correspond to cross-linked albumin. E, purified recombinant His 6 -GAPDH or His 6 -GAPDH (C149G) were incubated in the absence (C149G) or presence of aPKC/ and 5 Ci of [␥ 32 P]ATP in kinase buffer for 20 min at 30°C, and then analyzed by SDS-PAGE and autoradiography.
The molecular composition of recruited GAPDH was determined by subjecting the membranes obtained after performing the binding reaction to nondenaturing electrophoresis. Both His 6 -GAPDH and His 6 -GAPDH (C149G) had a mobility of ϳ150 kDa, which closely approximates the known molecular weight of native GAPDH as well as the contribution by His 6 indicating that the recombinant GAPDH exist as tetramers on VTCs (Fig. 3D). The presence of the tetrameric species suggests that the oligomeric structure is required for GAPDH function on VTCs.
We previously reported that GAPDH is phosphorylated by aPKC/ and that phospho-GAPDH plays a role in microtubule dynamics in the early secretory pathway (7). To establish whether the active site mutation affected aPKC/-dependent GAPDH phosphorylation, an in vitro kinase assay was employed. A comparable level of aPKC/-dependent phosphorylation was observed when the assay was performed with His 6 -GAPDH or His 6 -GAPDH (C149G) demonstrating that the mutant GAPDH serves as substrate for the kinase (Fig. 3E).
To study the effect of GAPDH (C149G) on membrane trafficking in the early secretory pathway, we introduced the recombinant protein into an in vitro transport assay (17). For this assay, tissue culture cells are first infected with ts045 VSV-G, a virus that synthesizes a protein with a thermoreversible defect resulting in ER retention at 39.5°C. The plasma membrane of these cells is then perforated to release soluble content, but retain functional ER and Golgi stacks. When the semi-intact cells are incubated at the permissive temperature of 32°C, export of ts045 VSV-G from the ER is initiated and transport of VSV-G protein is measured by following the processing of the two N-linked oligosaccharides to endo H-resistant forms. Since this assay requires addition of rat liver cytosol to reconstitute intracellular transport, the cytosol was first depleted of GAPDH by a combination of affinity chromatography on Blue Sepharose, which binds NAD-requiring proteins, and by immunodepletion (Fig. 4A). When the transport assay was performed with the GAPDH-depleted cytosol, there was an ϳ76% reduction in the processing of VSV-G to endo H-resistant forms. Importantly, supplementing the GAPDH-depleted cytosol with increasing concentrations of His 6 -GAPDH or with His 6 -GAPDH (C149G) reversed the inhibition and restored transport to near 77% of the control level. These results are highly suggestive that GAPDH catalytic activity is not required for VSV-G transport from the VTC (Fig. 4B). Furthermore, these combined studies demonstrate that the biochemical properties of His 6 -GAPDH (C149G) are not compromised by the mutation. Therefore, the mutant has similar function on the VTC as GAPDH wild type. DISCUSSION Rab proteins associate with a variety of effectors to form multimolecular complexes (21,22). This fact combined with our previous finding that aPKC/ interacts with the Rab2 N terminus and that the aPKC/ regulatory domain binds GAPDH prompted us to consider the possibility that GAPDH may also associate with Rab2. We performed in vivo and in vitro assays to establish that Rab2-GAPDH directly interact. The GAPDH binding site resides within residues 20 -50 in Rab2 that includes the putative effector domain (residues [35][36][37][38][39][40][41][42]. This region is one of two regions found by structural comparison of the GTP-and GDP-bound forms of small G proteins to undergo guanine nucleotide-dependent conformational changes that are recognized by regulatory factors and downstream effectors that modulate activity (19,20). Rab1 and Rab2 share 45% identity within this segment. However, Rab1 does not recruit GAPDH to membranes (10). GAPDH interaction with this Rab2 sequence is compatible with the observation that activated Rab2 binds GAPDH. Based on results obtained from the gel overlay assay and GST pull-down studies that employed only purified Rab2 and GAPDH, the interaction does not require aPKC/. However, we cannot rule out the possibility that aPKC/ enhances Rab2-GAPDH binding affinity and stabilizes their interaction on VTCs, in vivo. This interpretation would be consistent with our finding that a peptide made to the aPKC/ pseudosubstrate domain interferes with Rab2 recruitment of GAPDH to membranes (7). Our results are highly suggestive that Rab2-aPKC/-GAPDH form a complex on VTCs. It is possible that GAPDH binding to aPKC/ and Rab2 on VTCs blocks the active site or induces a conformational change making the active site inaccessible to substrate. Indeed, the dehydrogenase activity of membrane-associated GAPDH is significantly inhibited in enzymatic activity when associated with NRK microsomes (data not shown). A similar result has been reported when GAPDH associates with isolated human erythrocyte membranes (23,24). Although aPKC/ and GAPDH interact with other intracellular proteins and participate in various biochemical and signaling pathways independent of their roles at the VTC, the co-association with Rab2 "compartmentalizes" a specific activity required for retrograde transport. In that regard, aPKC/ phosphorylates GAPDH on VTCs and phospho-GAPDH influences microtubule nucleation at the budding site defined by Rab2 and associated effectors (7).
GAPDH is a well-characterized key enzyme in glycolysis that is responsible for the oxidative phosphorylation of G3P by FIG. 4. GAPDH glycolytic activity is not required for ER to Golgi transport. A, rat liver cytosol was batch-absorbed on Blue Sepharose. The eluate was subjected to immunoprecipitation for 4 h at 4°C with 10 g of anti-GAPDH monoclonal antibody and protein G Plus/protein A-agarose. The immune complexes were collected by centrifugation at 5,000 rpm for 5 min. The supernatant was removed, and then an aliquot equal to the nontreated cytosol analyzed by SDS-PAGE and Western blot. The blot was probed with an anti-GAPDH monoclonal antibody, washed, further incubated with an HRP-conjugated secondary antibody, and then developed with ECL. B, NRK cells infected with ts045 VSV-G were metabolically labeled with [ 35 S]methionine at 39.5°C for 10 min. The cells were permeabilized, shifted to 32°C, and then chased in radiolabeled-free medium for 60 min in the presence of GAPDH-depleted cytosol supplemented with increasing concentrations of (closed square) His 6 -GAPDH or (open circle) His 6 -GAPDH (C149G), or in the presence of complete cytosol supplemented with (open triangle) His 6 -GAPDH or (closed circle) His 6 -GAPDH (C149G). ts045-VSV-G was immunoprecipitated from detergent-lysed cells, digested with endo H, and analyzed by SDS-PAGE and fluorography. The endo H-resistant forms of ts045 VSV-G were quantitated using a phosphorimager. NAD ϩ and inorganic phosphate. The structure of GAPDH has been elucidated and studies have identified two significant regions that include the NAD ϩ binding domain and the catalytic domain that binds substrate (25)(26)(27)(28). The active enzyme exists as a tetramer containing identical 37 kDa subunits. We have found that both His 6 -GAPDH and His 6 -GAPDH (C149G) are tetrameric when bound to VTCs, and therefore GAPDH has the potential to be enzymatically active. GAPDH catalytic mechanism involves the formation of a hemithioacetal between the substrate and the active site thiol located at position 149. Cys-149 is essential for activity and the target of numerous thiol agents that inactive the enzyme (25,29). The experiments within show that substitution of Cys-149 eliminates dehydrogenase activity. To our knowledge, this is the first study to characterize biochemically and functionally an active site GAPDH mutant. A battery of assays were performed with the mutant to be assured that the substitution had no influence on other biochemical properties inherent to the role of GAPDH in the early secretory pathway including; 1) Rab2-dependent recruitment to membranes, 2) interaction with Rab2 and aPKC/, 3) aPKC/-dependent phosphorylation, and 4) an essential role in membrane trafficking between the ER and Golgi complex. All assay results were identical when performed with His 6 -GAPDH or His 6 -GAPDH (C149G).
The involvement of GAPDH in multiple intracellular activities inclusive/exclusive of its role in gluconeogenesis is consistent with the observation that numerous proteins have two or more different functions. These multifunction proteins are referred to as "moonlighting proteins" and their activity can vary dependent upon subcellular location, oligomeric state, cell type, or ligand, substrate, and cofactor concentration (29). Interestingly, phosphoglucose isomerase is also a key enzyme in glycolysis but functions extracellularly as either a nerve growth factor, a cytokine, a motility factor or a differentiation and maturation mediator (30 -34). Since add-back of GAPDH (C149G) to the in vitro trafficking assay reversed the inhibition and promoted VSV-G trafficking from VTCs, GAPDH glycolytic activity is not required for Rab2-mediated events.
What function could GAPDH provide at the level of VTCs independent of a catalytic role? Our previous results suggest that phospho-GAPDH promotes MT assembly/bundling at anchor sites within the VTC subcompartment. In this scenario, the Rab2-generated retrograde-directed vesicles would use the MTs as tracks to direct movement back to the ER. It is also possible that vesicle-associated GAPDH promotes fusion of the retrograde-directed vesicle with the ER. Lopez Vinals et al. (35) found that rabbit muscle GAPDH was a potent fusogen of negatively charged liposomes whereas studies by Glaser and Gross (14) demonstrated that a GAPDH isoform catalyzes fusion between pancreatic islet secretory granules and the plasma membrane. Moreover, Peters et al. (36) reported GAPDH as part of the fusion complex involving V-type [H ϩ ]ATPase and Ca 2ϩ /calmodulin. These two activities are not mutually exclusive and GAPDH may have dual function at the VTC.