Direct evidence for the cytoplasmic location of the NH2- and COOH-terminal ends of the Neurospora crassa plasma membrane H+-ATPase.

Reconstituted proteoliposomes containing Neurospora plasma membrane H+-ATPase molecules oriented predominantly with their cytoplasmic portion facing outward have been used to determine the location of the NH2 and COOH termini of the H+-ATPase relative to the lipid bilayer. Treatment of the proteoliposomes with trypsin in the presence of the H+-ATPase ligands Mg2+, ATP, and vanadate produces approximately 97-, 95-, and 88-kDa truncated forms of the H+-ATPase similar to those already known to result from cleavage at Lys24, Lys36, and Arg73 at the NH2-terminal end of the molecule. These results establish that the NH2-terminal end of the H+-ATPase polypeptide chain is located on the cytoplasmic side of the membrane. Treatment of the same proteoliposome preparation with trypsin in the absence of ligands releases approximately 50 water-soluble peptides from the proteoliposomes. Separation of the released peptides by high performance liquid chromatography and spectral analysis of the purified peptides identified only a few peptides with the properties expected of a COOH-terminal, tryptic undecapeptide with the sequence SLEDFVVSLQR, and NH2-terminal amino acid sequence analysis identified this peptide among the possible candidates. Quantitative considerations indicate that this peptide must have come from H+-ATPase molecules oriented with their cytoplasmic portion facing outward, and could not have originated from a minor population of H+-ATPase molecules of reverse orientation. These results directly establish that the COOH-terminal end of the H+-ATPase is also located on the cytoplasmic side of the membrane. These findings are important for elucidating the topography of the membrane-bound H+-ATPase and are possibly relevant to the topography of other aspartyl-phosphoryl-enzyme intermediate ATPases as well.

These results directly establish that the COOH-terminal end of the H+-ATPase is also located on the cytoplasmic side of the membrane. These findings are important for elucidating the topography of the membrane-bound H+-ATPase and are possibly relevant to the topography of other aspartyl-phosphorylenzyme intermediate ATPases as well.
The long-term goal of this laboratory is to elucidate the molecular mechanism by which the plasma membrane H+-ATPase of Neurosporu crassa transduces the chemical energy of ATP hydrolysis into a transmembrane electrochemical proton gradient. Of critical importance in achieving this goal is an understanding of the molecular structure of the H+- ATPase and its topography relative to the membrane in which it is embedded. Previous investigations of the topography of the H+-ATPase have been for the most part limited to interpretations of hydropathy profiles generated from analyses of the gene sequence. However, recent studies in this laboratory have provided methodology for preparing reconstituted H'-ATPase proteoliposomes in which the enzyme molecules are oriented predominantly with their cytoplasmic side facing outward (l), and methodology that allows for direct analyses of the protein chemistry of the H+-ATPase (2). The combination of a well-defined, membrane-bound H'-ATPase preparation and methodology that allows for detailed analysis of the protein chemistry of the H+-ATPase provides the essential ingredients for rigorous physical and chemical investigations of the topography of the H+-ATPase.
In the study described here, we present an improved procedure for preparing the reconstituted H+-ATPase proteoliposomes in which the bulk of the protein-free liposomes and any unreconstituted H'-ATPase molecules are removed and the proteoliposomes are concentrated, greatly simplifying subsequent analyses. We then initiate our investigation of the H+-ATPase topography by establishing the location of the NH2-and COOH-terminal ends of the H'-ATPase relative to the lipid bilayer in these proteoliposomes.
The establishment of the location of the ends of the H'-ATPase polypeptide relative to the bilayer imposes simple but important constraints on subsequent topographic models of the enzyme, including whether there are an even or odd number of transmembrane segments in the H+-ATPase, and, as a model of the intramembranous portion of the enzyme is developed (3), the direction in which each of the membrane-spanning segments must traverse the membrane.

EXPERIMENTAL PROCEDURES
Purificution and Reconstitution of the H'-ATPast--The plasma membrane H+-ATPase of Neurospora crassa was purified as described previously (4,5). The H+-ATPase was reconstituted into asolectin liposomes using a modification of the methods of Scarborough and Addison (1) and Goormaghtigh et al. (6). Briefly, 2 g of asolectin, purified by the procedure of Kagawa and Racker (

532
This is an Open Access article under the CC BY license.
EDTA, pH adjusted to 6.8 with KOH, centrifuged at 10,000 X g for 20 min at 20 "C, and the supernatant fluid containing the liposomes decanted.
Five ml of the purified H'-ATPase preparation (-2.6 mg of protein) was added to the liposome suspension, and the resulting mixture was inverted several times, frozen in a dry ice/methanol bath, and thawed at 30 "C.
Purification and Concentration of the H+-ATPuse-bearing Liposomes-Five-ml aliquots of the proteoliposome preparation were OVerlaid onto eight 35ml linear gradients of 24-40% (w/v) glycerol in buffer A. The gradient tubes were centrifuged at 60,000 rpm in a Beckman 70-Ti rotor (262,000 X g at raverage) for 18 h at 4 "C to separate the protein-free liposomes from those containing the reconstituted H'-ATPase (6). After centrifugation, the gradients were fractionated into 2.1-ml aliquots and assayed for turbidity and activity in the presence and absence of nigericin (1). Fractions containing the peak activity from all of the gradients were pooled (total volume approximately 84 ml), diluted to 120 ml with buffer A, and 30-ml aliquots were underlaid with 10 ml of a solution containing 40% (w/ v) glycerol in buffer A. The step-gradients were then centrifuged at 60,000 rpm in a Beckman 70-Ti rotor for 2 h at 4 'C to concentrate the proteoliposomes.
The band of turbidity near the density interface in each tube was isolated, pooled (total volume typically 15 ml), and the proteoliposome suspension either used immediately or stored at -20 "C and used within 3 days. When stored at -20°C the proteoliposome suspension did not freeze, due to the high concentration of glycerol in the solution, and showed no loss of activity in the presence and absence of nigericin over the course of 1 week.  (1,2,4).

AND DISCUSSION
We have previously shown that approximately 90% of the H+-ATPase molecules in H+-ATPase proteoliposomes prepared by our original freeze-thaw procedure are functional and oriented with their cytoplasmic surface on the outside of the liposome (1). This is a potentially useful preparation with which to investigate certain aspects of the H+-ATPase topography. However, significant drawbacks of this reconstituted preparation for studies of the H+-ATPase topography are the presence of a large excess of H+-ATPase-free liposomes, which interferes with subsequent protein chemical analyses, the possible presence of unreconstituted H+-ATPase molecules, which could lead to topographical artifacts, and a rather low concentration of protein, which also makes subsequent analyses difficult. As we have shown in a recent communication (6), the density of H'-ATPase-containing liposomes is greater than that of H+-ATPase-free liposomes, owing to the greater density of the protein. It is therefore possible to separate the H+-ATPase-containing and H+-ATPase-free liposomes on a glycerol density gradient. Fig. 1 shows the results of glycerol gradient purification of the freeze-thaw reconstituted H+-ATPase proteoliposome preparation. The five fractions with the greatest H'-ATPase activity contain approximately onethird of the phospholipid, as measured by phospholipid phosphorus assay (data not shown), and two-thirds of the ATPase activity loaded onto the gradient, which constitutes a significant enrichment of the H'-ATPase-containing liposomes. The gradient-enrichment step offers the added advantage that any unreconstituted H'-ATPase molecules, which are hexamers of the lOO-kDa monomers as isolated (14), are sedimented to the bottom of the gradient and thus removed (data not shown). Fig. 1 also shows that the H'-ATPase activity in the enriched liposome pool is stimulated greater than 2-fold by the addition of nigericin in the presence of K+. As we have developed in greater detail elsewhere (l), this degree of stimulation by nigericin indicates that the majority of the functional, cytoplasmic-side-out H+-ATPase molecules in the reconstituted, gradient-enriched proteoliposomes translocate protons into a sealed space, which is an important consideration for any topography study. Finally, concentration of the enriched proteoliposome preparation by centrifugation on a glycerol step-gradient, as described under "Experimental Procedures," results in a prep- The H'-ATPase was reconstituted into sonicated asolectin liposomes and the preparation subjected to glycerol density gradient centrifugation as described under "Experimental Procedures." The gradients were fractionated and the fractions assayed for turbidity and H+-ATPase activity in the presence and absence of nigericin.
aration containing approximately 100 lg of protein/ml and 50 mg of phospholipid/ml (data not shown). The specific activity of the H'-ATPase in a typical preparation is about 10 pmol of P,/min/mg of protein in the absence of nigericin and about 22 Fmol of P,/min/mg of protein in the presence of nigericin, comparable to that of proteoliposome preparations before the gradient enrichment and concentration steps (1,6), indicating that the gradient enrichment and concentration steps do not significantly affect the functionality of the reconstituted H+-ATPase molecules or the liposomes themselves. These specific activities also indicate that the functionality and orientation of the H+-ATPase molecules in the final liposome preparation are similar to that previously established (1).
Additional support for these conclusions is presented in Fig. 2. As described in detail before (l), SDS-PAGE analysis of H'-ATPase-bearing proteoliposomes in the presence and absence of H+-ATPase ligands can be used to estimate the percent of H'-ATPase molecules in the preparation that are functional and oriented with their cytoplasmic side facing outward.
Membrane-embedded H+-ATPase molecules oriented with their cytoplasmic side out are readily degraded by trypsin with the production of relatively small peptides. Additionally, recent data (not shown) indicates that H+-ATPase molecules with the reverse orientation are extremely resistant to any degradation by trypsin. Fig. 2A shows a time course of tryptic digestion of the H'-ATPase proteoliposome preparation described in this report in the absence of H+-ATPase ligands. Densitometric analyses of these gel lanes indicated that approximately 85-90% of the H'-ATPase molecules in the proteoliposome preparation were reduced to small peptides upon exposure to trypsin, which means that 85-90% of the H'-ATPase molecules in the preparation are oriented with their cytoplasmic side facing out. Conversely, lo-15% of the H'-ATPase molecules in the proteoliposome preparation were resistant to tryptic digestion, presumably representing a cytoplasmic-side-in population of H'-ATPase molecules. Tryptic digestion of H'-ATPase molecules in the presence of the active site ligands Mg*+ plus vanadate or Mg*+ plus ATP plus vanadate results in truncation of the H'-ATPase molecule to -97-, 95-, and 88-kDa forms, with relatively slow degradation of these forms to smaller peptide fragments (15,16). And, as elaborated in detail earlier (l), this ligandprotection of the H'-ATPase against tryptic digestion is a measure of the functionality of the H'-ATPase molecules in the reconstituted proteoliposome preparation. the time course of tryptic digestion of the H'-ATPase proteoliposome preparation described in this report in the presence of Mg'+, ATP, and vanadate. Densitometric analyses of these gel lanes showed that approximately 80% of the mass of the H'-ATPase was protected as the 97-, 95-, and 88-kDa forms until the later stages of the tryptic digestion, indicating that about 80% of the H'-ATPase molecules in the preparation, virtually all of the cytoplasmic-side-out molecules, are responsive to the H'-ATPase ligands and are therefore functional. The combined results of all of these analyses thus demonstrate that 85-90% of the H'-ATPase molecules in the proteoliposome preparation used in these studies are oriented with their cytoplasmic surface facing outward and that the great majority of these molecules are functional.
The improved reconstituted H'-ATPase preparation is therefore well-defined with respect to the H'-ATPase orientation and functionality, enriched in the H+-ATPase-containing liposomes, essentially free of unreconstituted H'-ATPase molecules, and adequately concentrated with respect to H'-ATPase protein, and is thus a very useful system for studies of the H'-ATPase topography. The results shown in Fig. 2 also establish the topographic location of the amino-terminal end of the H+-ATPase molecule. A recent study by Mandala and Slayman (16) demonstrated that conversion of the lOO-kDa native polypeptide to the -97-, 95-, and 88-kDa forms by trypsin is due to cleavage at LYS*~, Lys"', and Arg7", respectively.
Thus, the generation of these truncated forms of the H'-ATPase is a measure of tryptic cleavage at the NH2 terminus of the molecule. Therefore, because the great majority of the H'-ATPase molecules in the proteoliposomes used in these studies were oriented with their cytoplasmic side out, the fact that most of the H+-ATPase molecules in the preparation are converted to -97-, 95-, and 88-kDa truncated forms by trypsin in the presence of Mg*+, ATP, and vanadate, demonstrates that the NH, terminus is located on the cytoplasmic side of the membrane.
Our strategy for determining the presence or absence of the COOH terminus of the H'-ATPase in the cytoplasmic portion of the enzyme molecule was to analyze the peptides released from the above-described H'-ATPase proteoliposomes after trypsin treatment in the hopes of identifying a fragment from the COOH terminus of the enzyme, Two factors greatly simplified the search for such a peptide among the numerous tryptic cleavage products released from the proteoliposomes. First, hydropathy analyses of the amino acid sequence of the H+-ATPase (2,17,18) and investigations of the fragments of the H'-ATPase remaining associated with the liposomes after trypsin treatment (3) indicate that there are approximately 40 residues on the COOH-terminal side of the last membranespanning segment. This COOH-terminal portion of the enzyme contains six potential tryptic cleavage sites, with the largest tryptic fragment expected to contain between 11 and 29 residues and no tyrosine or tryptophan (Fig. 3). Thus, if one or more peptides are released from the COOH terminus of the H+-ATPase by trypsin treatment of the proteoliposomes, the largest of these peptides should have a substantial absorbance at 214 nm and relatively little absorbance at 280 nm. The second factor that simplified the search is the analytical capability of currently available HPLC data acquisition and analysis software. With such software, it is possible to definitively examine the absorbance characteristics of each peak in an HPLC eluate and thus pinpoint peptides of a specific type for further purification and analysis. Fig. 4 shows the results of HPLC analysis of the peptides released from the H+-ATPase proteoliposomes by tryptic digestion.
In this experiment, the proteoliposomes were treated with trypsin and subjected to gel filtration chromatography to separate the released peptides from the liposomes. The eluate fractions containing the released peptides were pooled, and the peptides concentrated by lyophilization and subjected to HPLC analysis. Out of a total of approximately 50 peaks in the profile, four peaks, indicated by the arrows with the numbers, had relatively high A214 ,,,,, and distinctively high ratios of AZ,, nm to A 280 nm, and were thus chosen as most likely to be the COOH-terminal peptide of interest. NH*terminal sequence analysis of the fraction containing peak number 1, eluting at about 108 min, was determined first and yielded the sequence TVEEDHPIPEEVDQAYK, identifying the major peptide in this sample as residues 483-499 of the H'-ATPase by reference to the gene sequence (17,18 ATPase molecules in the preparation are resistant to trypsin, presumably indicating a reverse orientation of these molecules. In order to ensure that the SLED.. . peptide was released from the predominant population of cytoplasmicside-out H+-ATPase molecules, rather than from the minor population of H'-ATPase molecules with the other orientation, it was important to quantitate the recovery of this peptide relative to the amount of H'-ATPase used in the experiment. To do this, the hydrophilic peptides generated from limit tryptic hydrolysis of the purified H+-ATPase were used as a calibration standard for the HPLC analyses. Previous analyses of this peptide mixture indicated that, subsequent to HPLC analysis, approximately 80% of the mass of the peptides is recovered relative to the amount expected from the starting mass of purified H'-ATPase (2). Thus, the areas of the peaks in the 214 nm profile from HPLC analysis of the calibration standard reflect on average approximately 80% of the mass of each peptide expected from tryptic cleavage and