Recombinant SFD Isoforms Activate Vacuolar Proton Pumps*

The vacuolar proton pump of clathrin-coated vesicles is composed of two general sectors, a cytosolic, ATP hydrolytic domain (V1) and an intramembranous proton channel, V0. V1 is comprised of 8–9 subunits including polypeptides of 50 and 57 kDa, termed SFD (SubFifty-eight-kDa Doublet). Although SFD is essential to the activation of ATPase and proton pumping activities catalyzed by holoenzyme, its constituent polypeptides have not been separated to determine their respective roles in ATPase functions. Recent molecular characterization of these subunits revealed that they are isoforms that arise through an alternative splicing mechanism (Zhou, Z., Peng, S.-B., Crider, B.P., Slaughter, C., Xie, X.S., and Stone, D.K. (1998) J. Biol. Chem. 273, 5878–5884).  To determine the functional characteristics of the 57-kDa (SFDα)1 and 50-kDa (SFDβ) isoforms, we expressed these proteins in Escherichia coli. We determined that purified recombinant proteins, rSFDα and rSFDβ, when reassembled with SFD-depleted holoenzyme, are functionally interchangeable in restoration of ATPase and proton pumping activities. In addition, we determined that the V-pump of chromaffin granules has only the SFDα isoform in its native state and that rSFDα and rSFDβ are equally effective in restoring ATPase and proton pumping activities to SFD-depleted enzyme. Finally, we found that SFDα and SFDβ structurally interact not only with V1, but also withV0, indicating that these activator subunits may play both structural and functional roles in coupling ATP hydrolysis to proton flow.

The vacuolar proton pump of clathrin-coated vesicles is composed of two general sectors, a cytosolic, ATP hydrolytic domain (V 1 ) and an intramembranous proton channel, V 0 . V 1 is comprised of 8 -9 subunits including polypeptides of 50 and 57 kDa, termed SFD (Sub Fiftyeight-kDa Doublet). Although SFD is essential to the activation of ATPase and proton pumping activities catalyzed by holoenzyme, its constituent polypeptides have not been separated to determine their respective roles in ATPase functions. Recent molecular characterization of these subunits revealed that they are isoforms that arise through an alternative splicing mechanism (Zhou, Z., Peng To determine the functional characteristics of the 57-kDa (SFD␣) 1 and 50-kDa (SFD␤) isoforms, we expressed these proteins in Escherichia coli. We determined that purified recombinant proteins, rSFD␣ and rSFD␤, when reassembled with SFD-depleted holoenzyme, are functionally interchangeable in restoration of ATPase and proton pumping activities. In addition, we determined that the V-pump of chromaffin granules has only the SFD␣ isoform in its native state and that rSFD␣ and rSFD␤ are equally effective in restoring ATPase and proton pumping activities to SFD-depleted enzyme. Finally, we found that SFD␣ and SFD␤ structurally interact not only with V 1 , but also withV 0 , indicating that these activator subunits may play both structural and functional roles in coupling ATP hydrolysis to proton flow.
Vacuolar, or V-type proton pumps acidify a wide array of intracellular compartments and are essential to functions of constitutive endocytotic and regulated secretory pathways. These H ϩ pumps are also found in the plasma membrane of polarized cells such as osteoclasts and renal tubular epithelial cells, where they function to acidify discrete extracellular compartments. In fact, V-pumps have been localized to virtually all intracellular compartments, except the nucleus and mitochondria, and their functions are equally diverse, ranging from promoting receptor-ligand dissociation in clathrin-coated vesicles and endosomes, to energizing neurotransmitter and catecholamine storage in synaptic vesicles and adrenal chromaffin granules (1)(2)(3).
Key questions thus arise regarding the mechanisms by which these pumps are targeted to their cellular sites and how differential regulation of the enzymes is achieved in these disparate locales. Several regulatory elements of V-pumps have been described, including activator proteins (4,5) and chloride channels that operate in parallel with V-pump to dissipate the charge generated by these electrogenic pumps and thereby facilitate pH gradient formation (6,7).
Recently we provided biochemical (8) and molecular (9) evidence for the roles of a 50-and 57-kDa polypeptide doublet, termed SFD (Sub Fifty-eight-kDa Dimer, or Doublet), in the activation of the V-pump of clathrin-coated vesicles (CCV) of bovine brain. These proteins were discovered in the course of our attempts to achieve biochemical resolution of components of the V-pump of CCV. This enzyme, like all V-pumps, is comprised of two general sectors, a multisubunit, cytosolic ATP hydrolytic domain (V 1 ), and a multisubunit proton channel (V 0 ). Biochemical and genetic studies have revealed that the V 1 domain in eukaryotic organisms is composed of 7 core subunits (A-G), some of which are present in multiple copies (10). Although less well characterized, the V 0 domain of the V-pump of CCV contains between 3-6 different subunits, ranging from a 116-kDa polypeptide (subunit a) to a small proteolipid (subunit c) (11).
Attempts at defining the components of V 1 (subunits A-G) and their functions revealed that an additional factor(s) was required for pump function, namely the polypeptides of SFD. When selectively depleted of these proteins, the V-pump of CCV, though assembled as a V 1 V 0 complex, cannot support ATP hydrolysis or proton pumping. Purified SFD, when added to SFD-depleted pump, was shown to restore these functions (8).
From a molecular standpoint, we recently determined the 57-and 50-kDa polypeptides of SFD, termed SFD␣ and SFD␤, respectively, arise from a single gene by an alternative splicing mechanism and that the SFD␤ isoform has a smaller molecular mass because of an 18-amino acid deletion (9). Further characterization revealed that the SFD proteins have sequence homology to the VMA13 product of Saccharomyces cerevisiae. As is the case with SFD, loss of the VMA13 gene product yielded a yeast vacuolar proton pump that was assembled but inactive (12).
Of note, the activation properties of SFD had been previously ascribed by others (13,14) to AP50, a component of the AP2 complex which is responsible for the assembly of the clathrin coats of coated pits and vesicles (15)(16)(17). Our more recent work demonstrated that SFD, and in particular its 50-kDa component (SFD␤), is molecularly distinct from AP50 and that removal of AP50 from impure pump preparations had no effect on enzyme activity, whereas removal of SFD␣ and SFD␤ accounted for the deactivation we had previously observed (9).
The current studies add final proof to the molecular identity of the SFD isoforms. Purified, recombinant SFD␣ and SFD␤ are shown to restore ATPase and proton pumping activity to V-pump depleted of SFD. Moreover, we have identified a source of V-pump (chromaffin granules) that in its native form has only SFD␣. When depleted of SFD, the chromaffin granule pump loses ATPase and proton pumping activities, and both are restored by addition of either SFD␣ or SFD␤. Finally, we have determined that SFD binds to both isolated V 1 andV 0 , thus providing evidence that it may function in a structural role of coupling ATP hydrolysis to proton pumping.

EXPERIMENTAL PROCEDURES
Preparations and Materials-Isolation of clathrin-coated vesicles from bovine brains (18) and preparations of liposomes from purified lipids (19) were performed as reported. Purification of the protontranslocating ATPase of clathrin-coated vesicles was performed by sequential solubilization with C 12 E 9 , hydroxylapatite chromatography, (NH 4 ) 2 SO 4 fractionation, and glycerol gradient centrifugations (20); purified H ϩ -ATPase had a specific activity of 14 -16 mol of P i ⅐mg of protein Ϫ1 ⅐min Ϫ1 . V 0 was isolated from purified V-pump (21), and recombinant subunit B (22) was prepared as described. Purified proton pump was depleted of SFD by treatment with Zwittergent 3-16, followed by glycerol gradient centrifugation, as reported (8). Partial purification of the vacuolar proton pump of chromaffin granules was achieved by modification of the protocol used for preparation of the vacuolar proton pump of clathrin-coated vesicles of bovine brain. Briefly, 500 mg of bovine chromaffin granule membranes were incubated at 4°C for 1 h with 20 ml of solubilization buffer, consisting of 10 mM Tris-HCl (pH 7.5), 1% C 12 E 9 , 0.5 mM EDTA, and 5 mM dithiothreitol. The mixture was centrifuged at 150,000 ϫ g for 1 h, and the supernatant was removed and mixed with (NH 4 ) 2 SO 4 to achieve 45% final saturation. After a 30-min incubation at 4°C, the mixture was centrifuged at 150,000 ϫ g for 30 min, and the supernatant was discarded. The pellet was dissolved in 1 ml of 20 mM Tris-HCl (pH 7.5), 0.05% C 12 E 9 , 5 mM dithiothreitol, and 0.5 mM EDTA and was layered over a 13-ml, continuous 15-30% glycerol gradient prepared in the same buffer. The gradients were centrifuged at 38,000 rpm for 20 h at 4°C in a Beckman SW40 rotor. Proton pump was harvested from the bottom one-third of the gradient in 1-ml fractions.
Restriction enzymes, T4 DNA sequencing ligase, and a nick translation kit were obtained from Roche Molecular Biochemicals; DNA sequencing reagents and enzymes and GeneAmp PCR reagent kit were from Perkin-Elmer-Cetus; E. coli strains DH5␣ and BL21 (DE3) pLys S and expression vector pET16b were from Novagen; radioactive reagents were from Amersham Pharmacia Biotech; nitrocellulose membranes for plaque lift were from Millipore Corp.; and chemicals for SDS-PAGE were from Bio-Rad. Bovine adrenal chromaffin granule membranes were the generous gift of Dr. Joseph Albanesi (University of Texas Southwestern Medical Center at Dallas).
Expression and Purification of Recombinant SFD Proteins-The coding region for bovine SFD␣ and SFD␤ were amplified by PCR using the cloned cDNAs SFD-RT4 and SFD21 (9), respectively, as templates and two synthetic oligonucleotides (5Ј-GGATCCGATGACCAAGATGGATA-TTCG-3Ј and 5Ј-GGATCCGGAGATCCAAGGGAAGCCCT-3Ј) as primers, which were designed to contain BamHI restriction sites to enable cloning into the bacterial expression vector pET16b. The amplification reaction was performed in a thermal cycler using the following conditions: 1 min at 94°C and 5 min at 65°C, for a total of 30 cycles. The PCR products were size-fractionated and purified by agarose gel electrophoresis; resulting fragments were identical to the parent cDNA sequences as determined by direct DNA sequencing. The PCR products and pET16b vector were separately digested with BamHI, and the desired digestion fragments were gel-purified and ligated by bacteriophage T4 DNA ligase. Ligation products were used to transform E. coli DH5␣. Plasmids recovered from independent clones yielded two expression vectors, pET16b-SFD␣ and pET16b-SFD␤, that were screened for the correct orientation by restriction analysis. Both of these plasmids were designed to express fusion proteins containing 10 neighboring histidine residues and a Factor Xa1 recognition site at the NH 2 terminus. For expression of SFD␣ and SFD␤, E. coli strain BL21 (DE3) pLys S was transformed with expression vectors pET16b-SFD␣ or pET16b-SFD␤, respectively, and grown and induced with IPTG at 37°C, as described (23).
Northern Blot Analysis-An 856-bp PCR product, originally used to isolate the cDNAs encoding SFD␣ (9), was used to generate a 32 Plabeled RNA probe with a MAXI-Script Sp6/T7 kit from Ambion. Designated amounts of bovine poly(A) ϩ -enriched mRNA were subjected to agarose (1.2%) gel electrophoresis and were subsequently transferred to nylon filter. After baking at 80°C for 1 h, the filters were prehybridized with 4ϫ SSC, 5ϫ Denhardt's solution, 0.1 mg/ml sheared, singlestranded salmon DNA, and 0.1% SDS at 65°C for 3 h. Subsequently the filter was hybridized with RNA probe (0.3 ϫ 10 6 cpm/ml) for 12 h at 65°C. Filters were sequentially washed with 2ϫ SCC, 0.1% SDS at 65°C for 15 min, 0.5ϫ SSC, 0.1% SDS at 65°C for 15 min, and 0.1ϫ SSC, 0.1% SDS at 65°C for 30 min. Washed filters were exposed overnight using Amersham Pharmacia Biotech Hyperfilm.
RT-PCR-RT-PCR was performed using a GeneAmp RNA PCR kit from Perkin-Elmer. Bovine poly(A) ϩ RNA (0.1 g) was reverse-transcribed using poly(dT)16 as primer, and PCR amplification was subsequently performed using Primer I (5Ј-CTGAGGAGAAGCAGGAGATG-3Ј) and Primer II (5Ј-GCATCAGCTGCAGACACCCG-3Ј). These primers were designed to yield a PCR product that included the 54-bp insertion site that distinguishes SFD␣ from SFD␤ (9). With SFD␣ transcript as template, these primers are predicted to yield a 450-bp PCR product, and with SFD␤ transcript template, a 396-bp PCR product. PCR products were separated electrophoretically on 1% agarose gels and were visualized by ethidium bromide staining.
Purification of Recombinant SFD␣ and SFD␤-Bacterial cells were harvested by centrifugation at 3,840 ϫ g for 20 min. The pellets were resuspended in 100 ml of lysis buffer (50 mM Tris-HCl (pH 7.5), 150 mM NaCl), and the cells were broken by sonication. The lysate was centrifuged at 186,000 ϫ g for 1 h, and the supernatants, which contained most of the recombinant proteins, were collected and loaded on 1 ml of Ni 2ϩ -NTA columns. After loading, the columns were washed with 20 ml of lysis buffer and then eluted with lysis buffer containing 20 mM imidazole (pH 7.5). The eluents were diluted 1:10 with 5 mM Tris-HCl (pH 7.5) and loaded on a Mono-Q (5 mm ϫ 5 cm) column. The column was washed with 10 ml of binding buffer (20 mM Tris-HCl (pH 7.5), and 10% glycerol). Proteins were eluted with linear (0 -400 mM) NaCl gradients prepared in 15 ml of binding buffer. The fractions (1 ml) contained as much as 10 -20 mg of the recombinant proteins, with Ͼ 95% purity, as assessed by SDS-PAGE (24) and by the Bradford protein assay (25).
Assembly of SFD-depleted Proton Pump with Recombinant SFD␣ and SFD␤-For each assay, 0.5 pmol of SFD-depleted proton pump was mixed with designated amounts (0 -2.5 pmols) of recombinant SFD␣ and/or SFD␤, and the final volume was brought to 10 l with buffer consisting of 20 mM Tris-HCl (pH 7.5), 0.05% C 12 E 9 , 0.5 mM EDTA, and 2 mM dithiothreitol. After incubation at room temperature for 10 min, reassembled pump was assessed for ATPase and/or proton pumping activities.
Measurement of ATPase and Proton Pumping-Measurement of ATPase activity was assessed by the liberation of 32 P i from [␥-32 P]ATP. Reassembled pump (0.5 pmol/assay) was incubated with 5 l of phosphatidylserine (0.5 mg/ml) at room temperature for 5 min. Reactions were initiated with addition of 200 l of assay buffer, consisting of 50 mM Tris-MES (pH 7.0), 30 mM NaCl, 3 mM MgCl 2 , 0.15 mM Na 3 VO 4 , and 3 mM [␥-32 P]ATP (200 -400 cpm/nmol). After incubation at 37°C for 30 min, reactions were terminated by the addition of 1 ml of 1.25 N perchloric acid, and liberated 32 P i was extracted and counted as described (26).
For assessment of ATP-driven proton pumping, samples of reassembled proton pump (0.5 pmol/assay) were reconstituted into liposomes (300 g/assay) prepared from purified lipids by the cholate-dilution, freeze-thaw method, as described (19). Proteoliposomes were diluted into 1.6 ml of pumping assay buffer consisting of 150 mM KCl, 10 mM Tricine (pH 7.5), 2.5 mM MgCl 2 , 0.15 mM Na 3 VO 4 , and 7 M acridine orange. Proton pumping was assessed by ATP-dependent quenching of the absorbance of acridine orange and was measured in Amino DW2C dual wavelength spectrophotometer as ⌬A492-540. Reactions were initiated by addition of 1.3 mM Na-ATP and 1 M valinomycin and were terminated by addition of 1 M 1799.
Western Blot Analysis-The preparation of antibodies Q50 (specific to SFD␣) and Q48 (directed against a common epitope of SFD␣ and SFD␤) were described previously (9). Designated amounts of protein were subjected to SDS-PAGE and were electrophoretically transferred to nitrocellulose filters. Protein A-purified Q50 or Q48 was incubated with filters at a 1:3000 dilution, and immunoreactivity was assessed by ECL using donkey, anti-rabbit IgG at a 1:4000 dilution.

RESULTS
Our previous attempts to resolve the 50-and 57-kDa polypeptides of SFD by conventional biochemical methods failed (8). While this may have owed to the fact that SFD existed as a heterodimer, it was also possible that their copurification owed to the fact that the two proteins differed by only 18 amino acids (9). Moreover, results of molecular mass determination by gel filtration chromatography were equivocal, perhaps because of the presence of residual amounts of detergent used for release of the polypeptide pair from holoenzyme. Key issues regarding the 50-and 57-kDa components of SFD are thus whether both components are required for pump functions, and if not, do differences exist in the functional properties of the two proteins.
To explore these points, we expressed and purified recombinant forms of the two polypeptides, as described under "Experimental Procedures." Shown in panel A of Fig. 1 4 and 5).
Of note was the finding that both rSFD␣ and rSFD␤ were present as soluble proteins in E. coli lysates. This differed from our experience with production of recombinant forms of all other V 1 subunits, each of which required detergents for solubilization, and many of which required denaturation with 8 M urea to achieve dissolution of inclusion bodies that contained the recombinant proteins.
To address whether SFD␣ or SFD␤ had independent activities, or synergistic effects when used in combination, reconstitution experiments were performed using the recombinant SFD proteins and biochemically prepared proton pump that had been depleted of SFD. As shown in Fig. 2, proton pump depleted of SFD lacked significant Mg 2ϩ -activated ATPase activity. Reassembly of recombinant forms of SFD with SFDdepleted pump resulted in a saturable activation of MgATPase activity. No differences were detected with the use of rSFD␣ alone, rSFD␤ alone, or equimolar mixtures of rSFD␣ and rSFD␤. Maximum restoration of MgATPase activity was achieved at a molar ratio of rSFD to SFD-depleted proton pump of 2. The final specific activity of reassembled enzyme was about 1.5 mols of P i ⅐mg of protein Ϫ1 ⅐min Ϫ1 . 2 To further characterize the functional properties of rSFD␣ and rSFD␤, we performed reconstitution experiments in which the rSFD proteins were first reassembled with SFD-depleted pump, and the resulting holoenzyme was then reconstituted into liposomes and assayed for proton pumping activity. Shown in Fig. 3 are the proton pumping activities of these preparations, as assessed by ATP-generated acridine orange quenching. SFD-depleted pump lacked activity (trace 1), whereas rSFD␣ (trace 2), rSFD␤ (trace 3), and equimolar mixtures of rSFD␣ and rSFD␤ (trace 4) activated protein pumping in SFDdepleted pump. In this particular experiment, proton pumping, as assessed by the initial rate of ATP-generated acridine orange quenching, was about 1.3-fold higher when a mixture of rSFD␣ and rSFD␤ were used (trace 4), as compared with rates attained with a single isoform of SFD (traces 2 and 3). This, however, was not a consistent observation, and coupled with the results obtained in the experiments of Fig. 2, we find no compelling evidence at present that there exist functional differences between the two isoforms.
Tissue distributions of the SFD isoforms were assessed by Northern blot analysis, as shown in Fig. 4. The highest copy number of transcripts was found in brain ( transcripts were detected in heart (lane 2). Because of the minor difference in molecular mass in the transcripts for SFD␣ and SFD␤, we performed RT-PCR to determine whether tissue differences in the transcript level of SFD␣ and SFD␤ existed. As described in "Experimental Procedures," primers were utilized that would predictably yield products of 450 bp for SFD␣ and 396 bp for SFD␤. Although transcripts for SFD␣ and SFD␤ were found in all tissues, it appears that SFD␤ is in relative abundance in brain (lane 2), whereas relatively more SFD␣ RT-PCR product was demonstrated in all other tissues examined.
We next sought to determine whether proton pump preparations from sources other than brain contained both SFD␣ and SFD␤ isoforms. As shown in Fig. 5, proton pump prepared from bovine brain (lane 1) contained roughly equimolar amounts of SFD␣ and SFD␤, whereas V-pump purified from bovine chromaffin granules (lane 2) had only the SFD␣ isoform, as determined by immunoblot analysis (panels B and C). This finding allowed us to examine by another approach whether any functional differences could be attributed to the isoforms, and we specifically sought to determine whether SFD␤ could substitute for SFD␣ in the proton pump of chromaffin granules, which, in its native form, contains only SFD␣.
To test this possibility, we first depleted the purified chromaffin granule V-pump of SFD (Fig. 6) by treatment of holoenzyme with Zwittergent 3-16, followed by glycerol gradient centrifugations as described under "Experimental Procedures." In panel A, SDS-PAGE reveals that the holoenzyme was found to migrate to its usual position in the glycerol gradient, as evidenced by the characteristic Coomassie stained bands of 70-, 58-, and 33-kDa, that can be visualized in lanes 3 and 4. In contrast, an immunoblot performed with anti-SFD antibody (panel B) demonstrate that all immunoreactive protein was present near the top of the gradient (e.g. lanes 8 and 9), where released SFD is typically found (8,9).
Next, we utilized SFD-depleted proton pump of chromaffin granules (as shown in Fig. 6, lanes 3 and 4) to compare the effects of rSFD␣ and rSFD␤ in restoration of enzyme activity. Shown in Fig. 7 are the results of these experiments. rSFD␣ and rSFD␤ were found to be equipotent in restoration of MgATPase (panel A) and proton pumping (panel B) activities of the SFD-depleted V-pump of chromaffin granules.
Finally, we have begun to investigate the structural basis of the effects of SFD on pump function. As described previously (8,9), and as demonstrated in these studies, SFD activates ATPase and proton pump activities of the holoenzyme. In addition, SFD has been shown to activate ATPase activity of isolated, SFD-depleted V 1 (8) To determine the domain(s) through which SFD regulates pump function, we performed binding experiments using rSFD, SFD-depleted proton pump, SFD-depleted V 1 , and isolated V 0 . For these studies (Fig. 8), recombinant, histidine-tagged SFD was incubated with SFDdepleted holoenzyme, SFD-depleted V 1 , or isolatedV 0 ; the mixtures were passed over Ni 2ϩ -NTA columns; and bound proteins  were subsequently eluted with imidazole. After SDS-PAGE, the eluents were then tested by immunoblot analysis to determine whether there had been binding of the 116-kDa polypeptide (a V 0 constituent) and/or the 70-kDa subunit (a V 1 constituent). As shown in panel A, SFD-depleted holoenzyme binds to both rSFD␣ (lane 3) and rSFD␤ (lane 4). In addition, both recombinant forms of SFD bind isolated V 0 (panel B, lanes 3 and 4) as well as SFD-depleted V 1 (panel C, lanes 3 and 4). Important controls for this study included passage of the biochemically prepared SFD-depleted pump, and V 1 orV 0, over a Ni 2ϩ -NTA column without preincubation with rSFD isoforms (lanes 1 of panels A, B, and C). In addition, histidine-tagged, recombinant (22) subunit B (a component of V 1 ) was substituted for rSFD and was demonstrated to not bind either V 1 ,V 0 , or holoenzyme (lanes 2 of panels A, B, and C). This study adds an additional insight into SFD function with the demonstration that SFD isoforms bind not only to V 1 , but also to V 0 , thus raising the possibility that SFD may play a structural role in the coupling of V 1 toV 0 , and thereby, the functional coupling of ATP hydrolysis to proton flow.

DISCUSSION
Expression of functional forms of 50-and 57-kDa components of SFD has allowed us to address several key questions regarding the role of SFD in V-pump function. First, a lingering question from our previous studies was whether both forms of SFD were essential to V-pump function. A strict biochemical approach to this issue was thwarted by our inability to separate the two isoforms (8). With the molecular cloning of SFD␣ and SFD␤, it became apparent that their (near) identical sequences might explain the difficulty in separating these two species by conventional biochemical methods (9). Moreover, attempts to determine the overall molecular mass of the putative SFD␣/ SFD␤ complex by size exclusion chromatography were equivocal, perhaps because of the presence of detergent-protein micelles resulting from residual Zwittergent 3-16. Two lines of evidence now demonstrate that the ␣ and ␤ isoforms can function independently in activation of ATPase and proton pumping activities.
First, we have shown that recombinant forms of either SFD␣ or SFD␤ can restore the ATPase and proton pumping activities of SFD-depleted proton pump prepared from clathrin-coated vesicles of bovine brain. Further, titration of mixtures of rSFD␣ and rSFD␤ did not yield synergistic effects. Second, we demonstrate that V-pump of chromaffin granules contains only the SFD␣ isoform, indicating that a native form of the enzyme functions with only one isoform. In additional experiments, rSFD␣ and rSFD␤ had equivalent effects on SFD␣-depleted pump of chromaffin granule, further demonstrating the independent, and interchangeable functions of the two isoforms. Additional evidence indicates that the purified recombinant SFD isoforms exist in monomeric forms prior to reassembly with the pump, as determined by high performance liquid chromatography performed with size exclusion columns (data not shown). We therefore believe it is most likely that the SFD protein exist in a monomeric form if, indeed, they are present as isolated species in the cytosol.
Our finding that optimal stimulation of pump activities occurs at a molar ratio of SFD to holoenzyme of 2:1 suggests that there may be more than one SFD binding site per holoenzyme. We, however, view the issue of SFD copy number as unresolved at present, in part because recombinant SFD isoforms may not be fully active and thus spuriously increasing the apparent SFD/holoenzyme ratio. Also, as discussed below, it is possible that SFD may act as a dissociable regulatory element, and the optimal ratio required for the binding of SFD to holoenzyme may differ from the copy number of SFD needed for activation of the enzyme, once binding has occurred. The finding that reassembly of SFD-depleted V-pump of CCV with mixtures of rSFD␣ and rSFD␤ does not yield additive, or synergistic, stimulation of activities indicates that the SFD binding sites of a given V-pump molecule are promiscuous, at least under these conditions. In addition, reassembly of SFD-depleted V-pump of chromaffin granule with either rSFD␣ or rSFD␤ provides additional evidence that SFD isoforms can act interchangeably, even in a pump which, in its native form, has only one (SFD␣) isoform. We thus favor the view that more than one isoform can be present in a single proton pump.
These studies add final clarification to a controversy regarding the nature of the 50-kDa polypeptide present in the V-pump of clathrin-coated vesicles. Previously it was reported by others (13,14) that this polypeptide was AP-50, a component of the AP2 complex that is essential to the assembly of clathrin coats (15)(16)(17). Previous work from our laboratory demonstrated that the 50-kDa polypeptide was in fact SFD␤ (9), and with the results of the current work, we have functional proof that SFD isoforms are causally related to the functional effect ascribed by others to AP-50. As we continue to note, these observations do not exclude the possibility that AP-50 may interact with V-pumps of clathrin-coated vesicles to induce changes in enzyme behavior that are at present unknown (e.g. targeting). In composite, however, our data do demonstrate that SFD components, and not AP-50, are required for V-pump function, as defined by activation of ATPase and proton pumping activities in enzyme depleted of this key component.
Given that SFD␣ and SFD␤ have interchangeable functions in activation of ATPase and proton pumping activities, a key question remains as to the roles these isoforms might play in the overall biology of V-pumps. Although we have no direct evidence to address this issue, it is likely that an important clue to this question resides in our finding that SFD isoforms can bind to both the V 1 and V 0 sectors of V-pumps. This observation is of some surprise, as all evidence to date has indicated that SFD is a V 1 component, as demonstrated by its copurification with isolated V 1 . Our current experiments now indicate that SFD subunits may play a structural role in coordinating the activities of V 1 and V 0 and may thus function not only as activators of V-pumps but also as true coupling factors. In this respect, it is notable that V-pumps (27), like F-type proton pumps (28), appear to have two links between the ATP hydrolytic sector and the intramembranous channel. In addition to the ␥ subunit that plays a key role in energy transduction, the F-type pump of E. coli has a stator arm that links F 1 and F 0 together in such a manner that F 1 is held stationary relative to the membrane. For the F-type pump of E. coli, this stator arm is composed of the b subunits of F o , and ␦ subunit of F 1 , whereas in mitochondrial F 1 F 0 this linkage is composed of the b subunits of F 0 and OSCP of F 1 (28). Based upon other investigations in our laboratory, it appears that the NH 2 terminus of the 116-kDa component (subunit a) of V 0 is cytosolic in orientation and may interact with SFD. 3 We thus speculate that subunit a with SFD may represent the stator arm of V-pumps. As such, SFD could play a pivotal role in governing the structural and functional assembly of V-pumps.
V-pumps, unlike F-type proton pumps, have been shown to reversibly dissociate into V 1 and V 0 components as a mechanism of recruitment of V-pump function in instances requiring increased proton pumping capability (29,30). Recently, we have demonstrated that the tendency of V 1 to dissociate from V 0 correlates with isoform diversity of subunit a (11). Thus it is possible that the isoform diversity of the SFD subunits interplays with the isoforms of subunit a to provide a multiplicative diversity in the interactions of V 1 withV 0 . This in turn, may represent an important means by which proton pumps undergo differential regulation in their myriad cellular locations.
We also believe that the solubility of recombinant SFD iso-forms may be highly significant. Of the 10 subunits and subunit isoforms of V 1 that we have expressed in functional, recombinant forms, only rSFD␣ and rSFD␤ exist as soluble proteins in expression cell lysates. Although our current data speak only to the roles these proteins play in activation of enzyme activity, and possibly coupling functions, we believe that SFD␣ and SFD␤ may exist as soluble, regulatory factors of V-pumps within the cytosol. Current studies are underway to investigate this issue.