Chemical state of the cysteine residues in the Neurospora crassa plasma membrane H(+)-ATPase.

The plasma membrane H(+)-ATPase of Neurospora crassa was treated with 5,5'-dithiobis(2-nitrobenzoate) to determine its cysteine content and with 2-nitro-5-thiosulfobenzoate to determine its cystine content. Six and seven mol of thiols/mol of H(+)-ATPase were detected in the 5,5'-dithiobis(2-nitrobenzoate) and 2-nitro-5-thiosulfobenzoate reactions, respectively, indicating that 6 of the 8 cysteine residues in the molecule are present as free cysteines and that 2 are present in disulfide linkage. The results of quantitative carboxymethylation experiments using [14C]iodoacetate under nonreducing and reducing conditions fully support this conclusion. Preparations of the ATPase 14C carboxymethylated under the above conditions were treated with trypsin, and the tryptic digests were resolved into hydrophilic and hydrophobic peptide fractions by our recently published procedure (Rao, U.S., Hennessey, J.P., Jr., and Scarborough, G.A. (1988) Anal. Biochem. 173, 251-264). Five of the six labeled free cysteine peptides partitioned into the hydrophilic peptide fraction and were purified and established to contain Cys376, Cys409, Cys472, Cys532, and Cys545. The labeled free cysteine residue in the hydrophobic peptide fraction was identified as either Cys840 or Cys869 by virtue of its presence in a large approximately 21-kDa hydrophobic peptide established previously to begin at Ser660. This in turn identified either Cys840 or Cys869 as one of the disulfide bridge cysteines. The other disulfide bridge cysteine was identified as Cys148 by purification and NH2-terminal sequencing of an additional peptide labeled in the reduced enzyme. The disulfide bridge is therefore between Cys148 and either Cys840 or Cys869. Because Cys148 is present in a putative membrane-embedded sector near the NH2 terminus of the ATPase molecule and Cys840 and Cys869 are present in a similar sector near the COOH terminus, it is possible that the disulfide bridge plays an important structural role in holding the two major membrane-embedded sectors of the molecule, distant in the linear sequence, together.

The plasma membrane H+-ATPase of Nezwosporu crassa was treated with 5,5'-dithiobis(2-nitrobenzoate) to determine its cysteine content and with 2nitro-5-thiosulfobenzoate to determine its cystine content. Six and seven mol of thiols/mol of H+-ATPase were detected in the 5,5'-dithiobis(2-nitrobenzoate) and 2 a large family of transport enzymes that includes the plasma membrane H+-ATPases of fungi, plants, and animals, the plasma membrane K+-ATPase of Escherichia co& the plasma membrane H+/K+-ATPase of gastric mucosa, the plasma membrane Na+/K+-ATPase of animal cells, and the Ca'+-ATPases of animal cell plasma membranes, sarcoplasmic reticulum, lysosomes, and Golgi (1). As one approach to this end, we are attempting to elucidate the topography of the H+-ATPase (2, 3) and to identify active site residues (4-6) using protein chemical techniques, As for all known integral membrane transport proteins, this has been a daunting task because of the presence in these molecules of a large hydrophobic domain that is almost totally refractory to conventional techniques of protein chemistry (7-10). However, in an important first step to surmounting this problem for the Neurosporu plasma membrane H+-ATPase, we have recently reported highly effective methodology for fragmenting this molecule and purifying, with high recoveries, virtually all of the hydrophilic and hydrophobic peptides produced (11). A key feature of the procedure developed is an initial separation of the tryptic hydrolysate into hydrophilic and hydrophobic peptide fractions, presumably representing the cytoplasmic and membrane-embedded domains of the molecule (2), which are then resolved by entirely different techniques. In this paper, this new methodology has been utilized to facilitate the elucidation of the chemical state of the 8 cysteine residues known to exist (12,13) in the H+-ATPase molecule. The results clearly indicate the presence of 5 free cysteine residues in the hydrophilic domain of the ATPase and 1 free cysteine and 1 disulfide bridge in the hydrophobic domain. HCl, or 0.8% (w/v) SDS also indicate the presence of six free sulfhydryl groups in the molecule.
Since there are 8 cysteine residues in the H+-ATPase molecule as deduced from the gene sequence (12, 13), these findings suggest that the remaining 2 cysteines in the molecule may be present in disulfide linkage and thereby not reactive with DTNB. This possibility was investigated by reacting the ATPase with NTSB under conditions described by Thannhauser et al. (18,19) for quantitating the disulfide bridge content of proteins. Fig. 1 also illustrates the results of the NTSB reaction with the H+-ATPase.
About 6 mol of NTB/ mol of ATPase is released in the first 5 min followed by a slower rise to a final value of about 7.1. This experiment has been carried out 10 times using several different ATPase preparations, with essentially the same results. As explained by Thannhauser et ul., the additional 1 mol of NTB released in the NTSB reaction compared with the DTNB reaction is the expected result of the sulfitolysis of a single disuhide bond in the enzyme. The abave data thus account for all 8 of the cysteine residues in the ATPase molecule.
The presence of fatty acylated cysteine residues has been demonstrated for several proteins (26-30), and it was therefore considered possible that the 2 cysteine residues in the H+-ATPase molecule which are unreactive with DTNB might be present as fatty acyl derivatives. To test this possibility, the behavior of two acyl thioesters, acetyl-CoA and palmitoyl-CoA, in the NTSB reaction was investigated. The results (not shown) showed clearly that acyl thioesters liberate NTB only slowly under the-conditions of the NTSB assay, much more slowly than any of the seven NTBs liberated by the ATPase (Fig. 1), presumably reflecting a slow rate of alkaline hydrolysis of the acyl thioesters followed by facile reaction of the liberated CoA with NTSB. In addition, when allowed to proceed to completion, the stoichiometry of such reactions was 1 mol of NTB produced/mol of acyl-CoA, not l/2 as would be required to explain the observed extra 1 mol of NTB liberated/m01 of ATPase in the NTSB reaction. Thus, it is extremely unlikely that the cysteine residues in the ATPase molecule which are unreactive with DTNB are present as fatty acyl thioesters.
Carboxymethylation of the H+-ATPase-To corroborate the results of the DTNB and NTSB assays, the stoichiometry of carboxymethylation of the ATPase with the well established thiol reagent iodo[Z-Y]acetic acid was investigated under nonreducing and reducing conditions. Table I summarizes the results obtained. ATPase treated with iodo[2-W]acetic acid in a 3-fold M excess over the 8 cysteine residues in the H+-ATPase molecule in the presence of 5.2 M guanidine HCl as described under "Experimental Procedures" incorporates almost exactly 6 mol of "'C-carboxymethyl group/m01 of enzyme. In these experiments, excess reagent was avoided in order to minimize the unwanted side reactions that take place with methionine, histidine, lysine, and tyrosine residues (20). However, the use of a 23-fold M excess of ['4C]iodoacetate did not result in any increased labeling of the enzyme (not shown). Moreover, the specific radioactivity of the iodo[2-14C]acetic acid used for these experiments varied anywhere from 1,000 to 21,000 cpm/nmol without any difference in the stoichiometry of labeling of the ATPase. These results are thus in complete agreement with the results of the DTNB assays of the free cysteine content of the H+-ATPase.
To account for the 2 remaining cysteines in the ATPase molecule, the ATPase was first carboxymethylated with nonradioactive iodoacetate under nonreducing conditions as described under "Experimental Procedures" to derivatize the 6 free cysteines and was then reduced with a high concentration of DTT under denaturing conditions, dialyzed to lower the DTT concentration to 2 mM, and then treated with a slight excess of iodo[2-'4C]acetic acid in the presence of 5.2 M guanidine HCl. The results of this experiment are also pre-sented in Table I. It can be seen that blocking of the free cysteines with nonradioactive iodoacetate followed by reduction and carboxymethylation with iodo[2-14C]acetic acid results in the radiolabeling of 2 additional cysteine residues, arising from the reduction of the disulfide bridge in the H+-ATPase molecule. To ensure that the labeling of the 2 additional cysteine residues in this experiment was indeed due to the reduction of a disulfide bridge and not due to any of the other manipulations required by this experiment, the same experiment was repeated without DTT and with a 400-fold M excess of iodo[2-'4C]acetic acid. The results are also shown in Table I. In this case, almost no cysteine labeling occurred, demonstrating that reduction by DTT is required for labeling of the additional 2 cysteine residues. Finally, the ATPase was fully carboxymethylated under reducing conditions.
The enzyme was iirst reduced with a high concentration of DTT in the presence of 5.2 M guanidine HCl for 3 h and then dialyzed as described under "Experimental Procedures." The enzyme was then carboxymethylated in the presence of guanidine HCl with iodo[2-'4C]acetic acid at a concentration slightly in excess of the thiol content present in the reaction mixture. The results of this experiment (Table I) showed that all 8 cysteine residues present in the enzyme are modified with iodoacetate under reducing conditions. The combined results of the carboxymethylation experiments thus fully corroborate the results of the DTNB and NTSB assays and indicate the presence of 6 free cysteines and a single disulfide bridge in the ATPase molecule. Importantly, they also pave the way for identifying both the free and disulfide bridge cysteine residues.
The enzyme preparation used for all of the above studies was isolated in a 30% (w/v) glycerol solution that did not contain DTT (see "Experimental Procedures"). It is thus possible that isolation of the ATPase in the absence of a reducing agent could allow the formation of an artifactual disulfide bond in the enzyme during purification.
To check this possibility, the ATPase was purified as described under "Experimental Procedures" but in the presence of 1 rnM DTT. The enzyme was then carboxymethylated with a slight excess of nonradioactive iodoacetic acid as above to block the free cysteines, dialyzed, and then assayed for free thiols and disulfide cysteines using the DTNB and NTSB assays described above. No reaction occurred with DTNB, indicating that all of the free cysteines had been derivatized with iodoacetate. However, in the NTSB reaction, 1.19 mol of NTB was released/m01 of ATPase, indicating that the disulfide bridge is present even when the ATPase is isolated in the presence of DTT. These results strongly argue that the disulfide bridge in the H+-ATPase is a natural and possibly important structural feature.
As isolated, the purified ATPase used for these experiments is in the form of hexamers (31), and it is possible that the disulfide bridge could be present between monomers in the hexamers. This possibility was ruled out by the absence of any higher molecular mass species above the lOO,OOO-Da ATPase band in SDS-polyacrylamide gel electrophoresis analyses of the enzyme run under nonreducing versus reducing conditions (not shown). The behavior of a ricin standard that contains a known intersubunit disulfide bridge (32) in these gels indicated that this procedure would have detected an intermonomer disulfide bridge if it were present. Tryptic Digestion of the Carboxymethylated H+-ATPase and Separation of the Products into Hydrophilic and Hydrophobic Peptide Fractions-we have reported recently a highly effective salt extraction procedure for quantitatively fractionating tryptic digests of the H+-ATPase into hydrophilic and hydrophobic peptide fractions (11). As a preliminary step toward identifying the free and disulfide bridge cysteines, the H+-ATPase was carboxymethylated by several of the procedures described above, digested with trypsin, and the distribution of the labeled peptides in the hydrophilic and hydrophobic peptide fractions was determined. Table I shows the results obtained.
When ATPase carboxymethylated to label the 6 free cysteine residues is used in such an experiment, 5 of the labeled free cysteines are found in the hydrophilic peptide fraction, and 1 is found in the hydrophobic peptide fraction. These pept,ides will be referred to below as the tive hydrophilic i4C-carboxymethylated free cysteine peptides and the hydrophobic 'C-carboxymethylated free cysteine peptide, respectively.
When ATPase specifically labeled at the disulfide bridge cysteines is used, 1 labeled cysteine is found in the hydrophilic peptide fraction, and 1 is found in the hydrophobic peptide fraction. The hydrophilic cysteine-containing peptide labeled in this way will be referred to below as the hydrophilic 14Ccarboxymethylated disulfide bridge cysteine peptide. When ATPase labeled at all 8 residues is used, 6 labeled cysteines are found in the hydrophilic peptide fraction, and 2 labeled cysteines are found in the hydrophobic peptide fraction. The combined results are thus highly quantitative and internally consistent and indicate that 5 of the free cysteine residues and 1 of the disulfide bridge cysteine residues in the H+-ATPase molecule are present in hydrophilic tryptic peptides and that 1 free and 1 disulfide bridge cysteine residue are present in hydrophobic tryptic peptides. Identification of the Hydrophilic %-Carboxymethylated Free Cysteine-containing Peptides- Fig.  2 shows the results obtained when the hydrophilic tryptic peptide mixture derived from ATPase "% carboxymethylated to label the free cysteine residues is resolved by HPLC as described under "Experimental Procedures." Two of the five "'C-carboxymethylated cysteine peptides, designated peptide I and II, respectively, are well resolved with retention times around 15 and 65 min.* The other labeled peptides elute as a partially resolved group between 155 and 180 min. Further purification of peptides I, II, III, and V by two or three additional HPLC runs as described under "Experimental Procedures" results in relatively pure peptide preparations as judged by the constant Az14 to AZ80 ratios in the purified peptide peaks (not shown). Each of these peaks could therefore be identified on the basis ' Peptide I occasionally elutes around 40-50 min. The reason for this variability is unknown. The H'+-ATPase was r4C carboxymethylated under nonreducing conditions, cleaved with trypsin, and the tryptic peptides generated from this procedure were separated into hydrophilic and hydrophobic peptide fractions as described under "Experimental Procedures." The hydrophilic peptides obtained from 1.2 mg of ATPase were then separated by HPLC as described under "Experimental Procedures." A, absorbance of the eluate at 214 nm. B, radioactivity distribution profile. of its amino acid composition.
Identification of these peptides on the basis of their amino acid compositions was straightforward since they each contain cysteine, and their amino acid sequences and NH* and COOH termini are predictable from the gene sequence (12, 13) and expected tryptic cleavage sites (11). Table II shows the amino acid compositions of peptides I, II, III, and V and their identities based on this information. Peptide I contained the 6 predominant residues shown and was therefore unambiguously identified as a hexapeptide containing Cy?', with the sequence Thr-Val-Cys-Glu-Ala-Lys beginning at Thr543 and ending at LASTS'. Peptide II was similarly identified as a pentapeptide containing CysY2 beginning at Ile470 and ending at LY&~ with the sequence Ile-Thr-Cys-Val-Lys.
Peptides III and V could also be identified in this way as CysT6-and Cys40g -containing peptides comprising residues 363-379 and 388-414, with the sequences Leu-Ser-Ala-Ile-Glu-Ser-Leu-Ala-Gly-Val-Glu-Ile-Leu-Cys-Ser-Asp-Lys and Leu-Ser-Leu-His-Asp-Pro-Tyr-Thr-Val-Ala-Gly-Val-Asp-Pro-Glu-Asp-Leu-Met-Leu-Thr-Ala-Cys-Leu-Ala-Ala-Ser-Arg, respectively. A second HPLC run of the pooled fractions from 162.5 to 175 min designated as peptide IV in the initial HPLC separation of the hydrophilic peptides (Fig. 2) indicated that there are several radioactive peptides in this region. Because the total amount of radioactivity in this region was equivalent to about 1 mol of "C-carboxymethylated cysteine/mol of the ATPase, it was suspected that this region contained incomplete tryptic cleavage products. The pooled peptide fractions from a similar preparation were therefore dried, redissolved, redigested with trypsin, and separated by HPLC as described under "Experimental Procedures." After redigestion, one major peak representing 65% of the radioactivity applied, and  two minor radioactive peaks were obtained. Although the amino acid composition of the major peptide (Table II) suggested that it comprises residues 520-537 with the sequence Gly-Glu-Gly-Ser-Trp-Glu-Ile-Leu-Gly-Ile-Met-Pro-Cys-Met-Asp-Pro-Pro-Arg, the identification was not as certain as for peptides I, II, III, and V. This peptide was therefore subjected to NHZ-terminal amino acid sequencing and yielded the sequence Gly -Glu -Gly -Ser -Trp -Glu -Ile -Leu -Gly -Ala -Met-Pro-X-Met-Asp-Pro-Pro, confirming the identity of this peptide. The five hydrophilic "C-carboxymethylated free cysteine peptides designated I-V in Fig. 2 thus comprise residues 543-548, 470-474, 363-379, 520-537, and 388-414, respectively. These identifications are further corroborated by the fact that the same peptides (labeled with 5((2-iodoacetamido)ethyl)l-aminonaphthalenesulfonic acid) were recently identified by Pardo and Slayman (33) and shown to elute in the same order as that shown in Fig. 2 under similar HPLC conditions. Identification of the Hydrophobic 14C-Carboxymethylated Free Cysteine Peptide- Fig.   3 shows the results obtained when the H"-ATPase hydrophobic peptide mixture containing the free i4C-carboxymethylated cysteine (Table I) is resolved by Sephadex LH-60 column chromatography by our recently described procedure (11). About half of the radioactivity emerges near the void volume in the fractions we have established previously (11,34) to contain only an -2l-kDa hydrophobic H+-ATPase peptide beginning at Ser6'jo and ending near Lys"' and a small amount of a concanavalin A contaminant (11) that contains no cysteine residues (35). This indicates that the free cysteine residue in the hydrophobic part of the ATPase molecule is either Cys%' or Cys@' because there are no other cysteine residues in the -2l-kDa hydrophobic peptide as deduced from the gene sequence (12,13). Since our original analysis (ll), we have established by NHZ-terminal sequencing (not shown) that a previously unidentified, minor -14-kDa hydrophobic ATPase peptide eluting just to the right of the -2l-kDa peptide in the Sephadex LH-60 profile also begins at Ser660 and because of its size, presumably ends at Ar$13. SDS-polyacrylamide gel electrophoresis analysis (34) of the column fractions collected in the experiment of Fig. 3 (not shown) showed the presence of roughly equal quantities of the -2l-and -14-kDa bands, indicating an increased efficiency of tryptic digestion of the -2l-kDa peptide compared with our earlier studies (ll), presumably as a result of denaturation and carboxymethylation of the ATPase. Conversion of the -2l-kDa hydrophobic peptide to the -14-kDa peptide would release one or more smaller peptides containing both Cysa4' and Cysa6', which explains the presence of radioactivity in the smaller peptides eluting after the -Zl-kDa peptide in the experiment of Fig. 3.
The establishment of either Cysa4' or Cys@j' as the 6th free cysteine residue in the ATPase molecule also establishes one or the other of these residues as the hydrophobic disulfide bridge cysteine (Table I) because the other 5 free cysteines have already been accounted for.
Identification of the Hydrophilic 14C-Carboxymethylated Disulfide Bridge Cysteine Peptide-Identification of the hydrophilic 14C-carboxymethylated disulfide bridge cysteine peptide defined as described above was by no means straightforward because of its extremely peculiar properties. Attempts to purify the peptide by HPLC using Cl8 and C4 reversed phase columns, DEAE-ion exchange column chromatography, Sephadex LH-60 column chromatography in chloroform/methanol/trifluoroacetic acid (ll), gel filtration on Sephadex G-50 in anionic and cationic detergents, two-dimensional peptide mapping on thin layer cellulose sheets, SDS-polyacrylamide gel electrophoresis, and electroblotting onto polyvinylidene difluoride membranes following SDS-polyacrylamide gel electrophoresis were all unsuccessful for a variety of reasons. However, molecular exclusion chromatography on a long column of Sephacryl S-100 HR equilibrated with 1 M guanidine HCl as described under "Experimental Procedures" eventually proved to be a satisfactory means of purifying this peptide. Fig. 4 shows the results of a typical column run. The eluted radioactivity profile is characterized by a small peak at around 310 ml with a high Azzo followed by a major radioactive peak representing more than half of the radioactivity applied at about 330 ml with a much smaller Azzo.
The major labeled peptide in this experiment was identified by NHZ-terminal amino acid sequencing. Although the repetitive yields for this peptide dropped rapidly, again presumably reflecting its peculiar properties, the sequence obtained clearly establishes that this peptide begins at Val'* with the sequence Val-Val-Pro-Glu-Asp-Met-Leu-Gin-Thr..
. . Because of its apparent size and because it must contain a cysteine residue,