Identification of the Major Cytoplasmic Regions of the Neurospora crussa Plasma Membrane H+-ATPase Using Protein Chemical Techniques*

The transmembrane topography of the Neurospora crassa plasma membrane H+-ATPase has been inves- tigated using purified, reconstituted components and direct protein chemical techniques. Reconstituted pro- teoliposomes containing H+-ATPase molecules ori- ented predominantly with their cytoplasmic surface facing outward were treated with trypsin to liberate peptides present on the cytoplasmic surface of the H+- ATPase J.

liposomes in the preparations was determined from the phosporous content by the procedure of Bartlett (8).
Materials-The sources of most of the materials used have been described previously (1).

AND DISCUSSION
As demonstrated in a recent paper (l), 85-90% of the H+-ATPase molecules in proteoliposome preparations identical to those used in the present study are functional and oriented with their cytoplasmic portion facing outward. The remaining ATPase molecules in the preparation are extremely resistant to any detectable degradation by trypsin (1) and presumably are present in the liposomes with the opposite orientation. With this information taken into account, Fig. 1 outlines the concept of our approach to defining regions of the H'-ATPase molecule present on the cytoplasmic side of the membrane. ATPase molecules oriented with their cytoplasmic surface facing outward are cleaved with trypsin, resulting in the formation of liposomes containing the membrane-associated and exocytoplasmic portions of the molecule, together with a collection of peptides from the cytoplasmically located regions of the molecule that are released from the liposomes. The released peptides are then separated from the liposome-bound peptides by Sepharose CLGB column chromatography and are then separated from one another by reversed-phase HPLC. Fig. 2 shows the results obtained when the released peptide collection is fractionated by HPLC on a Cls column as described previously (1). More than 50 peaks with significant absorbance at 214 nm can be discerned in the sections of the chromatogram shown. Ideally, every peptide peak in such a chromatogram that can be identified by NH*-terminal amino acid sequence analysis as originating from the H+-ATPase molecule by reference to the gene sequence (9, 10) represents evidence for the location of that peptide on the cytoplasmic side of the membrane. However, minor uncertainties as to the physical state of all of the ATPase molecules in the preparation warrant caution in assigning a cytoplasmic location to peptides present in very low yields. That is, although it is reasonably certain that 85-90% of the ATPase molecules in the proteoliposome preparation are functional and present with their cytoplasmic surface exposed, and that the remaining lo-15% of the ATPase molecules in the preparation are not detectably degraded by trypsin (l), it is possible that a small amount of the ATPase molecules in the preparation is present in other physical states. Moreover, it is likely that certain of the peaks present in the HPLC profile of Fig. 2 are derived from proteinaceous or non-proteinaceous contaminants of the asolectin used in these experiments and are therefore of little interest. For these reasons, it was deemed necessary to confirm the ATPase origin of the peptides present in the HPLC analysis of Fig. 2  relative to the mass of the H'-ATPase present in the proteoliposomes at the beginning of the experiment, so that only major ATPase peptides would be chosen for NHz-terminal sequence analysis. This was done by comparing the area in the 214-nm profile of each peak with that of the corresponding peak in the profile of a standard tryptic digest of the purified H'-ATPase, as described in detail previously (1). Only peptides with counterparts in the standard tryptic digest and with recoveries greater than 50% were chosen for subsequent NH2terminal sequencing.
The peaks chosen on this basis are indicated by the asterisks in Fig. 2. The asterisks with numbers indicate the peaks that yielded unambiguous sequences. Table I summarizes  This information provides major constraints for models of the transmembrane topography of the ATPase and together with information obtained from other recent experiments allows the formulation of a reasonably detailed and accurate model for the disposition of the ATPase polypeptide chain in the membrane. Fig. 3 presents a model that takes into account all of the available pertinent information.
In the model, residues with topographic locations directly established are indicated by the larger, bold letters using the single-letter nomenclature for amino acid identities.
Beginning at the NH,terminal end of the molecule, we have recently demonstrated (1) that treatment of ATPase-bearing proteoliposomes identical to those used in the present investigation with trypsin in the presence of the H'-ATPase ligands, Mg++, ATP, and vanadate produces approximately 97-, 95-, and 88-kDa truncated forms of the H'-ATPase similar to those shown by Mandala and Slayman (11) to arise from cleavage of the ATPase molecule at Lysz4, LYS~~, and ArgT3, respectively, which provided evidence for the cytoplasmic location of these residues. The cytoplasmic location of these tryptic cleavage sites was also demonstrated recently by Mandala and Slayman (12). Because the cleft in which trypsin substrates bind during catalysis reaches several residues beyond the scissile bond in both directions along the cleaved polypeptide chain (13), at least 4 residues in each direction from the cleaved bond are presumed to be accessible in the polypeptide chain in order for trypsin to act. Therefore, in the model of Fig. 3, in addition to the residues identified by peptide purification and NHp-terminal sequence analysis, 4 residues in each direc-  LGLGGGGDMPGS 584-615 ' From Fig. 2. b Calculated from the area of the peak in the HPLC run of Fig. 2 the membrane with reasonable certainty, because these stretches are too short to form transmembrane helices that cross the membrane and return between residues established to be on the cytoplasmic side of the membrane. Residues with such inferred locations are indicated in the model by the smaller letters. The caveats with such deduced stretches are, of course, that they could conceivably enter into the plane of the membrane inside the structure of the H'-ATPase or cross the membrane bilayer as non-helical structures. For these reasons, it is useful to assign smaller letters to these stretches in order to acknowledge the fact that their location is likely, but not quite as certain as the stretches established directly by NH*-terminal amino acid sequence analysis of peptides released from the membrane. Continuing along the polypeptide chain, the identification of peptides 6 and 4 described in Fig. 2 and Table I establish residues 70-100 as located on the cytoplasmic side of the membrane, as indicated by the bold letters in the model. At this point, the polypeptide chain approaches its first entry into the lipid bilayer. From experiments similar in concept to that schematized in Fig. 1, but in which the peptides remaining associated with the liposomes were ana-lyzed by our recently developed methodology for analyzing the hydrophobic portions of the ATPase (6, 14), we have obtained evidence that the great majority of the membraneembedded domain of the ATPase is contained in three tryptic fragments with approximate molecular masses based upon their mobilities in SDS-polyacrylamide electrophoresis gels of 7, 7.5, and 21 kDa, beginning at residues 100, 272, and 660, respectively, and with 8-12 membrane crossings (15 and unpublished data, with M, values altered to take into account the information in Ref. 16). The model in Fig. 3 was constructed to take this information into account, but aminoacid letter assignments have not been made for these regions because the emphasis of this paper is on the cytoplasmic regions of the H+-ATPase molecule and because our investigation of the membrane-embedded region of the ATPase is not quite complete. The results of these latter studies will be elaborated upon in detail in a forthcoming paper. However, for the purposes of the present discussion, suffice it to say that it is reasonably certain that the ATPase polypeptide chain crosses the membrane and returns between residues 100 and 186, accounting for the -7-kDa hydrophobic peptide beginning at residue 100 that remains associated with the membrane after tryptic cleavage.
At this point, the polypeptide chain appears again on the cytoplasmic side. The identification of peptide 13 described in Fig. 2 and Table I establishes residues 186-219 as located on this side of the membrane as shown. Likewise, the identification of peptide 3 defines residues 238-256 as also present on the cytoplasmic side, and, by deduction, indicates the presence of residues 220-237 on the same side. The chain then approaches the next membrane-embedded region and proceeds with two to four membrane crossings, accounting for the -7.5-kDa hydrophobic peptide beginning at residue 272, found associated with the liposomes (see above).
From the information thus far available, the topographical location of the next -80 residues (-360-440) is unclear. From our studies of the chemical state of the 8 cysteine residues in the H'-ATPase molecule (17), the retention times in the HPLC system used in the present investigation of two tryptic cysteine-containing peptides comprising residues 363-379 and 388-414 are well established at about 160 and 179 min, respectively. Moreover, during the course of the cysteine work, these two peptides did not exhibit any problems with respect to losses, and were recovered quantitatively.
Nevertheless, repeated attempts to find these peptides in the released peptide collection have been uniformly negative. Because the first of these peptides contains the active site aspartate at position 378, it is likely that this part of the stretch in question is accessible from the cytoplasmic side of the membrane, as indicated in the model. However, the failure to find these peptides in the released peptide collection may indicate that this part of the ATPase molecule, including the active site aspartate, is embedded in the membrane, or is intimately associated with other parts of the molecule that are. It is pertinent in this regard that the Ca*+-binding sites of the closely related Ca'+-translocating ATPase of sarcoplasmic reticulum appear to be located close to one another and deep in the membrane-embedded region of the molecule (18-25, but also see 26), and also appear to be near (20, 23), probably within 4 A of (27), the terminal phosphoryl group of bound ATP next to the active site aspartate, suggesting that the active site aspartate in the Ca"-ATPase may also be embedded in the membrane.
Clearly, additional work will be required to elucidate the reasons for the absence of these peptides in the released peptide collection and ascertain whether or not they are present in the membrane-embedded region of the H+-ATPase.
It might be mentioned in this context, however, that if the active site aspartate is indeed close to or embedded in the membrane in the aspartyl-phosphoryl-enzyme intermediate-type ATPases, then direct energy coupling mechanisms for these enzymes such as those proposed in detail earlier (28, 29) become entirely reasonable possibilities.
The ATPase polypeptide chain then enters into the last membrane-embedded region and proceeds with four to six membrane crossings, accounting for the -21 kDa hydrophobic peptide beginning at residue 660 that remains associated with the liposomal membrane after tryptic cleavage (see above). And finally, the identification by NHz-terminal sequence analysis of an undecapeptide beginning at SeTgo in our previous investigation of the peptides released from the reconstituted proteoliposomes (1) directly established the presence of residues 897-915, and by deduction, residues 916-920 of the ATPase molecule, on the cytoplasmic side of the membrane as indicated in the model of Fig. 3.
In conclusion, in this paper we have demonstrated by direct protein chemical methods the presence of 14 peptide segments of the N. crassa plasma membrane H+-ATPase, and by inference, the presence of several additional segments of the molecule, on the cytoplasmic side of the membrane. This information, taken together with the results of previous investigations of the locations of the NH, and COOH-termini, has localized at least 417 of the 919 residues in the H'-ATPase molecule as cytoplasmically located. With this information, and additional preliminary results regarding the membraneembedded segments of the molecule, a first generation topographical model of the H'-ATPase containing a substantial amount of established detail has been constructed. The model should be a useful starting point for further experimentation on the topography and other structural aspects of the H+-ATPase molecule and could also be useful for comparative studies with the other transport ATPases in the aspartylphosphoryl-enzyme intermediate family.