Isolation and Characterization of the Methionine Aminopeptidase from Porcine Liver Responsible for the Co-translational Processing of Proteins*

A methionine aminopeptidase that specifically re- moves methionine residues from peptides with amino-terminal sequences of Met-Ala-, Met-Val-, Met-Ser-, Met-Gly-, and Met-Pro- but not Met-Leu- or Met-Lys- has been isolated to homogeneity from porcine liver by a procedure involving five chromatographic steps. The enzyme, whose specificity matches that predicted for the entity responsible for the co-translational amino- terminal processing of nascent polypeptide chains, has a measured molecular mass of 70,000 Da by SDS- polyacrylamide electrophoresis and 67,000 Da by gel chromatography (under nondenaturing conditions), suggesting the native molecule is a monomer. It is activated by Co2+ and inhibited by 8-mercaptoethanol and EDTA. With octapeptide substrates related to the amino-terminal portion of the @-chain of human hemoglobin (with a histidine in position 3), the enzyme had a pH optimum of 6.0. With a synthetic peptide devoid of histidine, it showed no pH dependence from 6.0 to 8.0. This sensitivity may be due to the propensity of peptides with histidine in the third position to bind divalent cations such as Co2+. The measured K,,, and kcat values were affected by residues in the second position. The peptide corresponding to the natural sequence (Met-Val-His-) gave a kcat/&, value of 260 m”’ s-’;

quenced, have been found to possess an amino-terminal methionine residue (1). Secreted proteins and most proteins imported into the mitochondrion typically contain a signal sequence that is removed by a specific peptidase, exposing an internal residue that becomes the amino terminus of the mature protein, which accounts for the missing methionine in these cases. However, most other intracellular proteins are also devoid of amino-terminal methionine, generally having instead the penultimate residue (originally adjacent to the initiator methionine) at the amino terminus, often blocked by an acetyl group (2). Thus amino-terminal processing, in different ways, is a common feature in the formation of most mature proteins.
In eukaryotes, both the cleavage of the initiator methionine and the addition of Ne-acetyl groups apparently occur cotranslationally (2), suggesting that the enzymes catalyzing these reactions are likely to be ribosomally associated. The removal of the methionine by a methionine aminopeptidase (MAP)' probably precedes N"-acetylation. Acting in concert, they can generate four classes of amino termini: Met-X-, X-, AcMet-X-, and AcX-, where X is the penultimate residue (2), and Ac is acetyl. In previous studies (1,(3)(4)(5)(6)(7), it has been shown both in vivo and in vitro that the substrate specificity of MAP depends almost exclusively on the physical size of the penultimate residue. When this residue is glycine, alanine, serine, threonine, proline, cysteine, or valine, the initiator methionine will be removed; the remaining residues with larger side chains will cause methionine to be retained. This specificity apparently has been widely conserved in prokaryotes and eukaryotes, even though it does not occur cotranslationally in the former, suggesting that at least a part of the purpose of this processing event may be conserved throughout living organisms. One possible role is suggested by the experiments of Varshavsky and colleagues (8), who have shown that the half-life of certain proteins depends on the residue at the amino terminus. In yeast, the stabilizing residues were the ones exposed by MAP, whereas the destabilizing ones were equivalent to the subset that prevented methionine removal. When these residues were left exposed (by post-translational proteolytic processing of a fusion protein), they directed a rapid ubiquitin-dependent degradation. Thus, at least one function of ribosomally associated MAP may be to mask or unmask determinants that govern the stability of proteins.
MAPS with a specificity generally consistent with that predicted from in vivo and in vitro studies have been isolated from two bacteria (9, 10) and yeast (11). The bacterial enzymes (Escherichia coli and Salmonella) have been cloned (9,12) and have molecular masses of -30,000 Da. The purified yeast enzyme has also been recently cloned and has a calcu-The abbreviation used is: MAP, methionine aminopeptidase. lated molecular mass of 43,269 Da (13). All three enzymes apparently utilize Co2+ for their activity.
In this report, we describe the first purification and characterization of MAP from a higher eukaryote (porcine liver). Although also stimulated by Co2+, it is substantially larger than the bacterial and fungal enzymes. Furthermore, in contrast to the yeast enzyme (ll), its substrate specificity matches very well that expected for the enzyme responsible for the co-translational removal of the initiator methionine in this tissue.

Materials
Phenylisothiocyanate, triethylamine, and trifluoroacetic acid were purchased from Pierce Chemical Co. Leupeptin, pepstatin, benzamidine, and the p-nitroanilide substrates were obtained from Sigma.
Peptide substrates were synthesized by the t-butoxycarbonyl method on an Applied Biosystems model 430A peptide synthesizer (Foster City, CA). The reagents utilized were also from Applied Biosystems, except the t-butoxycarbonyl amino acids, which were obtained from Bachem (Torrance, CA). Peptides were cleaved from the resin by HF and were purified by high performance liquid chromatography on a C4 column (1 X 25 cm) (Vydac) (The Separations Group, Hesperia, CA) using a linear gradient of 10% CH3CN, 0.1% trifluoroacetic acid to 30% CHgCN, 0.07% trifluoroacetic acid. Peptides were analyzed for amino acid composition after acid hydrolysis, as described below, and for amino acid sequence using an Applied Biosystems model 477A protein sequenator. The peptide Met-Pro-His-Thr-Leu-Pro-Glu-Glu was a gift of Dr. K. Hoey (Johnson and Johnson, La Jolla, CA).
Porcine liver was obtained from Farmer John in Los Angeles, CA and either used directly or flash-frozen in liquid nitrogen and stored at -70 "C. Enzyme prepared from the tissue stored at -70 "C for more than 6 months was generally obtained in lower yield and contained a greater amount of the lower molecular weight material (see below).

Methods
Assay of MAP Activity MAP activity was assayed by the addition of 5-20 pl of enzyme to a solution of 50 mM sodium phosphate buffer, pH 7.2, 50 mM NaCI, 0.1 mM CoClz (final concentration), followed by preincubation for 2 min at 37 "C. The reaction was initiated by adding 30 nmol of peptide substrate to a final volume of 100 pl. After incubation for 20 min at 37 "C, 0.1 volume of 0.1 N HC1 was added to stop the reaction, and a 50-pl aliquot was removed and dried in a Speed Vac (Savant, Hicksville, NY). The reaction products were derivatized with phenylisothiocyanate by a modification of the Picotag (Waters, Milford, MA) method (14). Briefly, 20 pl of methanol:H,O:triethylamine (2:l:l) was added to the dried aliquot, and resuspension was achieved by vortex mixing. Ten ~1 of methano1:triethylamine:phenylisothiocyanate (9:4:1) was added, and the reaction was incubated for 5 min at room temperature and dried in a Speed Vac. The derivatized material was resuspended in 50 mM ammonium acetate buffer, pH 7.0 ,with 5% CH,CN and analyzed on an ODS Ultrasphere PTH column (Beckman, Fullerton, CA) by the Picotag chromatography method (14).
The effect of P-mercaptoethanol, EDTA, and metal ions on MAP activity was determined by the addition of the germane agent to the enzyme solutions for 15 min at 4 "C prior to the initiation of the assay (at 37 "C). Enzyme used in these experiments was not previously dialyzed against 0.1 mM CoClZ.
Isolation of Porcine Liver MAP The procedure was routinely carried out on 120 g of liver. The values given correspond to this scale. All steps were done at 4 "C.
Step I : Homogenization-The tissue (fresh or thawed) was cut into small pieces, rinsed in 20 mM Tris-HC1, pH 7.4, 25 mM KC1, 10 mM MgC12, 25 mM sucrose, and homogenized with 4 mg/ml of leupeptin and pepstatin (each) and 1 mM benzamidine, 3 g of liver/ml of buffer using a Polytron homogenizer. The homogenate was centrifuged at 8,000 X g for 15 min, and the supernantant was filtered through cheesecloth. Solid NaCl was added to a final concentration of 0.5 M, and the solution was stirred for 15 min, followed by centrifugation at
Step 2: S-Fast Flow Chromatography-The dialyzed homogenate was passed over a DE52 (Whatman, Clifton, NJ) filter ( 5 X 10 cm), which was connected directly to an S-Fast Flow (Pharmacia LKB Biotechnology Inc.) column (1.5 X 20 cm), both being equilibrated with buffer A without CoCIz. This system was loaded at a flow rate of 80 ml/h. The column was subsequently equilibrated with 2 column volumes of 20 mM Tris-HC1, pH 8.4, 50 mM NaCl, 10% glycerol, 0.02% azide (buffer B) and eluted using fast protein liquid chromatography at a flow rate of 2.5 ml/min with a 500-ml linear gradient from 50 mM to 1 M NaCl (in buffer B) (Fig. 1). Fractions were collected, and active fractions were pooled, concentrated to approximately 10 ml by ultrafiltration with a YM-10 membrane (Amicon) and dialyzed against 20 mM potassium phosphate buffer, pH 7.2, 50 mM NaC1, 0.1 mM CoCl2, 10% glycerol, 0.02% azide (buffer C).
Step 3: Hydroxylapatite Chromatography-The sample from the S-Fast Flow step was centrifuged at 10,000 X g for 10 min and loaded onto a hydroxylapatite column (Bio-Rad) (1.5 X 10 cm) equilibrated in buffer C without CoClZ. The column was washed with 2 volumes of the same buffer, and the activity was eluted with a 200-ml linear gradient of 20 mM to 0.5 M potassium phosphate buffer, pH 7.2, 10% glycerol (Fig. 2). The active fractions were pooled and concentrated by ultrafiltration as described in the previous step.
Step 5: Heparin CL-6B Chromatography-The pooled, active fractions from the $200 column were loaded directly onto a Heparin CL-6B (Pharmacia LKB Biotechnology Inc.) column (1 X 5 cm) at a flow rate of 15 ml/h. The column was equilibrated in buffer C without CoC1, and eluted with a 60-ml linear gradient from 50 mM NaCl to 1 M NaCl in buffer C without CoC1, (Fig. 4). The active fractions were pooled and dialyzed against buffer C at pH 6.0.
Step 6: Mono S Chromatography-The protein solution from the previous step was loaded onto a Mono S HR5/5 column (Pharmacia LKB Biotechnology Inc.) equilibrated in buffer C at pH 6.0 without CoClZ. The activity was eluted at 0.5 ml/min with a linear gradient of buffer C to 0.9 M potassium phosphate buffer, pH 6.0,10% glycerol ( Fig. 5). The active fractions were dialyzed against buffer A with 10% glycerol and 0.02% sodium azide and stored at 4 "C. The enzyme remained fully active for several months under these conditions.

Polyacrylamide Gel Electrophoresis
Polyacrylamide gel electrophoresis analyses were routinely performed by the method of Laemmli (15) and were stained with Coomassie Brilliant Blue.

Amino Acid Analyses
Amino acid compositions were determined on a Beckman System Gold Analyzer on samples prepared by vapor-phase acid hydrolysis. Protein (or peptide) was reacted for 24, 48, and 72 h in vacuo with HCl. Values for serine and threonine were calculated by extrapolation to zero time.

Determination of Substrate Specificity, pH Optimum, and Kinetic Parameters
For these assays, 70 pmol of enzyme (dialyzed against 0.1 M CoC1,) were incubated with 200 PM substrate (final concentration), except for the kinetic analyses that were carried out over a range of 50-300 p~. For the specificity and pH dependence studies, assays of 10 min at 37 "C were used; in the kinetic experiments, measurements were made at 1,5, and 10 min. All assays were performed in 50 mM sodium phosphate buffer, pH 6.0 (or the indicated adjusted value).

RESULTS
Isolation and Properties-Using the protocol described above, a homogeneous preparation of a porcine liver MAP has been isolated in 34% yield. As summarized in Table I, five chromatographic steps provide a 23.6-fold purification with a recovery of 200 pg (commencing with 120 g of fresh tissue). The real increase in -fold purification was undoubtedly higher; however, it was impossible to determine the amount of this enzyme in the whole extract accurately because of other aminopeptidases. Assuming 100% recovery from the S-Sepharose step, the -fold purification from the extract would have been about 44,000.
The yield was also at least in part affected by the amount of a second form of MAP that is partially separated in the Porcine Liver Methionine Aminopeptidase final purification step. (Partial separation also apparently occurred on hydroxylapatite, but both forms were pooled together at that point (see Fig. 2)). The presence of this material is illustrated by the SDS-polyacrylamide gel electrophoresis analysis (Fig. 6) of the final separation on the Mono S column (Fig. 5). The electrophoretic analysis of the three principal fractions (fractions [14][15][16] indicated the presence of two species of 70 and 66 kDa, respectively. These occur in homogeneous forms in the first and last fractions, respectively, and as a mixture in the middle one. Characterization of the two forms revealed that they had very similar amino acid compositions, indistinguishable peptide profiles, as determined by Cleveland gel mapping (16), and identical specificity profiles with the battery of hemoglobin-related peptide substrates that varied in the second position (see below) (data not shown). However, the higher molecular weight form had twice the specific activity of the smaller one. Importantly, the amount of the 66-kDa species was found to vary from preparation to preparation and was clearly related to the state of the starting material (fresh or frozen) and the presence of protease inhibitors, particularly in the early purification steps. These findings strongly support the view that the 66-kDa protein is derived from the 70-kDa species by a limited proteolytic event(s); however, it was difficult to obtain large amounts of the 66-kDa species reproducibly, since the conditions that favored its formation resulted in poor overall yields of both species. Thus, at present, it cannot be rigorously ruled out that this molecule does not arise as a separate gene product or as the result of an alternative splicing mechanism during transcription.
The native molecular weight of this porcine liver MAP was determined on a calibrated column of Sephadex S-200. As compared with a standard mixture of proteins of known molecular weight (aldolase, serum albumin, trypsinogen, and cytochrome c), the MAP, which eluted as a single symmetrical peak, was determined to have a molecular mass of 66 kDa. This value was in excellent agreement with that determined from the SDS-polyacrylamide gel electrophoresis (Fig. 6) and clearly indicates that this protein is monomeric in its native state.
The amino acid composition of the 70-kDa form of MAP is shown in Table 11. The protein is relatively rich in glycine, alanine, and leucine. The protein contains over 100 acid plus amide residues, which is more than twice the number of basic residues. However, the behavior of the protein on Mono S chromotography suggests that a majority of the residues in this class will be in the amidated form.
A number of aminopeptidases that have been described appear to be metalloproteases (or a have a strong metal ion requirement for their activity). Leucine aminopeptidase, clearly the best characterized enzyme of this type, contains an active site zinc atom (17). Two prokaryotic aminopeptidases and a similar enzyme from yeast, which also are highly selective for methionine residues, are greatly stimulated by cobalt ions (9)(10)(11). Accordingly, several divalent metal ions, as well as two compounds known to disrupt metal ion binding, were tested for their effect on porcine liver MAP activity. As shown in Table 111, magnesium chloride was without effect on the activity, whereas both manganese chloride and zinc chloride were inhibitory. In contrast, CoC12 stimulated the activity of the enzyme over 2-fold, relative to undialyzed controls. This behavior is similar to that observed for the prokaryotic and yeast MAPS (9,11,12) and suggests that the porcine liver enzyme also utilizes this metal ion. @-Memaptoethanol and EDTA were both inhibitory, presumably because of their interactions with the metal ion.

TABLE I1
Amino acid composition of porcine liver MAP Purified MAP (Fig. 5, fraction 14  Substrate Specificity-The substrate specificity of the porcine liver MAP was tested using a series of peptides whose sequences were based on the amino terminus of the @-chain of human hemoglobin. In this series, only the second residue was varied. As shown in Fig. 7, the enzyme was active against substrates in which alanine, glycine, serine, proline, and valine occupied this position. It showed no activity with substrates in which this position was occupied by leucine or lysine. This is indeed the specificity predicted for the enzyme responsible for the co-translational amino-terminal processing of proteins. As additional support for this conclusion, the peptide in which the histidine in the third position was substituted for proline was assayed. Although the corresponding peptide (with valine in the second position) showed 75% activity (relative to the peptide with alanine in the second position) the peptide with the third position substitution was completely inactive (data not shown). This sequence corresponds to that found in two mutant hemoglobins (hemoglobin Long Island and Marseille) (18, 19), which have previously been shown to retain methionine. This also suggests that the porcine and human enzymes have similar specificities.
Catalytic Properties-As shown in Fig. 8, the enzyme is not particularly sensitive to pH in the range 6.0-8.0, using a synthetic substrate with the sequence Met-Ala-Ser-(Gly)5-(Leu)3. However, when the hemoglobin substrates with either valine or alanine in the penultimate position were assayed,  the enzyme showed an optimal activity at pH 6, and this decreased rapidly, approaching base-line levels by pH 8. The shape and position of this curve is indicative of the titration of a histidine residue. Since the effect is not observed in the substrate lacking histidine, the most likely candidate for this residue is in the substrate itself. However, the inactive form represents the uncharged state of the imidazole ring (which would be comparable with the uncharged synthetic substrate that is fully active at pH 8). A possible explanation that is more consistent with other observations is that the neutralized peptide becomes an efficient chelator of the Co2+ and, thus, acts in a fashion similar to EDTA or /3-mercaptoethanol (see above). Histidine residues in the third position of peptides are known to be excellent chelators of divalent metal ions in the deprotonated state (20).
The kinetic parameters for the hydrolysis of three representative peptides are shown in Table IV. In this study, three of the hemoglobin substrates with alanine, valine, and proline in the second position were examined. K,, VmaX, and kcat values were calculated from Lineweaver-Burk analyses. The alanine substrate, which showed the highest V,,, and kcat values, surprisingly was less well bound as judged by the K,,, value relative to the valine and proline peptides. However, both of these latter peptides were turned over more slowly, as reflected by kCat/K,,, values of 260 and 130 relative to that of the alanine substrate, which has a kCat/Km of 1,523 mM" s-'. These values suggest that the larger side chains of the penultimate residue provide greater binding energy (up to a certain size), but they are not as well accommodated by the catalytic machinery of the enzyme. The substrates with larger side chains in this position, which are not hydrolyzed at all (such as the leucine and lysine peptides) may either not be bound (because of steric hindrance in the appropriate subsite) or they may be bound but not hydrolyzed. If the latter situation were true, these substrates should act as competitive inhibitors.
Interestingly, the Met-Pro and Met-Val substrates showed substrate inhibition at concentrations near their K , values.
The effect was more pronounced with the Met-Pro substrate than the Met-Val substrate. The Met-Ala substrate also displayed this property but at higher concentrations (greater than 1 mM). These observations offer an explanation as to why the Met-Val substrate appears to be more rapidly hydrolyzed in the single-point assays used in the substrate specificity experiments (Fig. 7). Substrate concentrations used in ., Met-Ala-Ser-(Gly),-(Leu)3. those experiments were near the K , values of these substrates (200 mM).

DISCUSSION
Eukaryotic cells contain a variety of aminopeptidases, enzymes capable of removing the amino-terminal residues from peptide and protein substrates with varying degrees of specificity and selectivity. These enzymes are involved in a number of physiological functions, including digestion and absorption of amino acids from food, processing of the precursors of bioactive peptides, and protein degradation and turnover. Many enzymes that have been identified do not, as yet, have defined functions. The aminopeptidase isolated and characterized in this study shows a high degree of selectivity for methionine residues, and this specificity is clearly governed by the nature of the penultimate residue. Thus, it removes methionine from peptides in which the adjacent residues have relatively small side chains and not from those with more bulky side chains. Furthermore, the enzyme is not processive, i.e. it does not remove the next amino acid regardless of its nature. The selective nature of the catalytic activity of this enzyme, particularly against the substrates containing the Met-Pro sequence, suggest that it is the enzyme involved in the co-translational processing of proteins during eukaryotic translation. Both from direct measurements (3)(4)(5)(6)(7)21) and from an analysis of sequences occurring in the protein data base (1,3), the specificity of this enzyme(s) is well documented. The fact that this process occurs co-translationally and on nascent chains of relatively short length (20-40 amino acids) (22) has led to the suggestion that the enzyme is associated specifically with ribosomes. A direct demonstration of that remains to be ascertained for this enzyme.
There are many enzymes known to be capable of excising methionine from the amino-terminal position of peptides and proteins. Leucine aminopeptidase hydrolyzes amino acids from the amino terminus of polypeptides, naphthyl amides, and amino acid amides. Although it has a broad specificity and will continuously degrade peptides in a sequential fashion, it does not hydrolyze substrates with arginine or lysine in the first or second position (17,23), nor will it cleave substrates with proline in the second position. A protein of similar size and substrate specificity has been purified from the microsomal preparations of porcine kidney, but it is less well characterized than the cytosolic form (24). A third enzyme, aminopeptidase A, with similar physical properties to the microsomal leucine aminopeptidase, has also been isolated from several sources (25). This enzyme appears to be specific for substrates with amino-terminal aspartic and glutamic acid. Two membrane-bound forms of aminopeptidase (aminopeptidase M and aminopeptidase N) have been cloned from rat kidney and human intestinal mucosa, respectively (26,27). The human enzyme is the major protein in the microvillar membrane of the small intestine, but it is also found in the plasma membrane of other cell types. Both enzymes are zinc enzymes. The function of these enzymes is almost certainly distinct from the soluble aminopeptidases and may be related to amino acid uptake and/or precursor processing of bioactive peptides. A rat liver MAP has been partially purified following detergent extraction from subcellular organelles (28). It is active in the removal of methionine from the amino terminus of hemoglobin nascent chains but also acts on Met-Lys bradykinin and methionine 2-naphthylamide and leucine naphthylamide. It does not hydrolyze arginine naphthylamide or lysine naphthylamide. These latter activities clearly suggest that it is not the MAP involved in co-translational processing of proteins.
Three other enzymes have been identified that show similar substrate specificity profiles to the porcine liver enzyme. Two of these have been isolated from prokaryotes and correspond to proteins of approximately 30 kDa (9,10,12). Both proteins are activated by Co2+ and possess similar amino acid sequences, as determined from cDNA analyses.
The role of these enzymes in bacteria is not entirely clear. Prokaryotic translation is initiated with N-formylmethionine, and removal of this initiator residue must be preceded by the action of a deformylase enzyme. The subsequent removal of methionine is consistent with the action of these enzymes, although it clearly occurs in a post-translational fashion. The enzyme isolated from yeast behaves as a 34,000-Da protein (on SDS gel electrophoresis) (11) but has a calculated mass of over 43,000 Da from the cDNA sequence. However, the isolated protein, as judged by amino-terminal analysis, apparently lacks 10 amino acids, suggesting a molecular weight of -42,000 (13). The poor agreement in these values probably stems from anomalous behavior of the protein in SDS gels. As shown in Table V, its specificity (relative to alanine substrates) is somewhat distinct from that of either the bacterial or porcine preparations. It shows only very low activity against peptides with valine and serine in the penultimate position. As demonstrated in an in situ experiment (5), the profile of yeast processing is indistinguishable from that of higher organisms, suggesting that this enzyme may not be directly involved in co-translational processing. Alternatively, as suggested by Chang et al. (11), a second enzyme may also function in yeast to account for the complete spectrum of processing observed in that organism.
The substantially larger size of the porcine liver enzyme (approximately double that of the bacterial and fungal enzymes) is presently unexplained. It is likely that this enzyme contains at least one catalytic domain that is similar to that seen in the bacterial and yeast enzymes because of its common dependence on the biologically rarer metal ion, Co2+. This domain may, in fact, be repeated in the porcine enzyme, thus providing two catalytic centers in a single polypeptide chain. Alternatively, the extra molecular weight material may be involved in structures required to associate the enzyme with the ribosome and/or other components of that organelle, e.g. the N"-acetyltransferase, which also functions co-translationally (2). Chang et al. (13) have identified putative "zinc finger" domains in the amino-terminal sequence of the yeast enzyme that may participate in such interactions. Regardless of the function(s) of this extra material, it clearly marks the porcine liver enzyme as distinct from either the prokaryote or lower eukaryotic proteins. The presence of an enzyme with the appropriate predicted specificity for amino-terminal processing in a higher eukaryote supports the view that the function of this enzyme, as manifested in its essentially ubiquitous distribution in living organisms, must be well conserved. One such role, defined in yeast by Varshavsky and colleagues (a), suggests that the presence or absence of amino-terminal methionine residues ultimately can direct the turnover of at least some proteins (the N-end Rule). Briefly stated, proteins containing penultimate residues that direct the retention of methionine were shown to be inherently unstable after the processing of a fusion protein that led to their postribosomal exposure. It follows that the retention of methionine is at least in part a protection against premature degradation and further provides subsequent avenues for directing turnover by the later removal of methionine by other aminopeptidases, presumably located elsewhere in the cell. Therefore, the suggestion that higher eukaryotes show a different profile of stabilizing residues is not easily understood in the context of the retention line Aminopeptidase 20673 of the specificity of a co-translationally acting MAP (29).
However, it must be pointed out that the importance of the removal of MAP to produce proteins with the appropriate penultimate residue in the amino-terminal position (with or without an N-acetyl group) is not fully appreciated as yet either. Thus the retention of methionine, which is in effect a default situation, may be less important than the actual removal of this residue. Experiments to ascertain the role of this enzyme directly in in situ protein turnover in various paradigms are ndw under investigation.