Membrane type-1 matrix metalloprotease and stromelysin-3 cleave more efficiently synthetic substrates containing unusual amino acids in their P1' positions.

The influence of the substrate P1' position on the specificity of two zinc matrix metalloproteases, membrane type-1 matrix metalloprotease (MT1-MMP) and stromelysin-3 (ST3), was evaluated by synthesizing a series of fluorogenic substrates of general formula dansyl-Pro-Leu-Ala-Xaa-Trp-Ala-Arg-NH2, where Xaa in the P1' position represents unusual amino acids containing either long arylalkyl or alkyl side chains. Our data demonstrate that both MT1-MMP and ST3 cleave substrates containing in their P1' position unusual amino acids with extremely long side chains more efficiently than the corresponding substrates with natural phenylalanine or leucine amino acids. In this series of substrates, the replacement of leucine by S-para-methoxybenzyl cysteine increased the kcat/Km ratio by a factor of 37 for MT1-MMP and 9 for ST3. The substrate with a S-para-methoxybenzyl cysteine residue in the P1' position displayed a kcat/Km value of 1.59 10(6) M-1 s-1 and 1.67 10(4) M-1 s-1, when assayed with MT1-MMP and ST3, respectively. This substrate is thus one of the most rapidly hydrolyzed substrates so far reported for matrixins, and is the first synthetic peptide efficiently cleaved by ST3. These unexpected results for these two matrixins suggest that extracellular proteins may be cleaved by matrixins at sites containing amino acids with unusual long side chains, like those generated in vivo by some post-translational modifications.

Matrix metalloproteases (MMPs), 1 also known as matrixins, form a group of structurally related zinc endopeptidases collec-tively able to degrade all components of the extracellular matrix (1). MMPs are believed to be mediators of both normal and pathological tissue remodeling processes, and their increased expression has been observed in a variety of human disorders (2)(3)(4). In particular, membrane type-1 (MT1)-MMP, a progelatinase A activator (5), and stromelysin-3 (ST3), a matrixin with unusual functional properties (6), are expressed in most human carcinomas (7)(8)(9). MT1-MMP and ST3 both belong to a subgroup of MMPs which are believed to be intracellularly activated by furin or furin-like convertases, thereby suggesting that MT1-MMP and ST3 are present in tissues in an active form, in contrast to other MMPs which are secreted as inactive zymogens (10 -12). As part of a program aimed at developing therapeutic inhibitors for these two MMPs, the specificity of these enzymes in cleaving synthetic substrates has been investigated. Current data on the specificity of MT1-MMP toward the degradation of synthetic substrates are extremely limited (13), while no synthetic substrate has yet been reported to be cleaved by ST3.
In contrast to MT1-MMP and ST3, several studies have been devoted to the delineation of the specificity of other matrixin family members by developing both synthetic substrates (14 -18) and inhibitors for these enzymes (19,20). Such approaches have been greatly facilitated by the resolution of the crystal structures of several MMP catalytic domains (21)(22)(23)(24)(25)(26)(27), allowing structure-based design strategies to be used (28). According to these structural studies, the S 1 Ј subsite of these enzymes appears as a cavity of variable size, depending on the nature of the amino acid residue located near the bottom of this cavity. In stromelysin-1 (ST1), which contains leucine in this particular position, the S 1 Ј pocket is a deep cavity, forming a channel that extends through the whole body of the enzyme catalytic domain (29,30). X-ray structure analysis of a complex between ST1 and a carboxylalkyl inhibitor harboring a homophenylalanine in the P 1 Ј position has consistently shown that the homophenylalanine side chain only fills half of the S 1 Ј pocket of ST1 (29). A comparable situation is likely to occur in most other MMPs, including MT1-MMP, collagenase-2 (COL2), stromelysin-1 (ST1), stromelysin-2 (ST2), gelatinase A, and gelatinase B, due to the presence in their S 1 Ј pocket of a leucine at the same position. The high potency of several inhibitors, substituted in their P 1 Ј position by side chains longer than homophenylalanine, toward this subgroup of matrixins is consistent with this proposal (28,(31)(32)(33)(34)(35). In contrast to this subgroup of matrixins, collagenase-1 (COL1), matrilysin, and ST3 possess in their S 1 Ј * This work was supported by funds from the Commissariat à l'Energie Atomique, the Institut National de la Santé et de la Recherche Scientifique, the Center National de la Recherche Scientifique, the Center Hospitalier Universitaire Régional, the Bristol-Myers Squibb Pharmaceutical Research Institute, the Association pour la Recherche sur le Cancer, the Ligue Nationale Française contre le Cancer, the Comité du Haut-Rhin, the Fondation de France, the BIOMED 2 (contract no. BMH4CT96-0017) and BIOTECH 2 (contract no. ERBBIO4CT96-0464) Program, and a grant to P. Chambon from the Fondation Jeantet. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
While the particular shape of the S 1 Ј pocket in matrixins has been extensively exploited for the design of matrixin inhibitors, only one study so far examined the influence of this S 1 Ј cavity on the cleavage of synthetic substrates with unusual amino acids in their P 1 Ј position (37). It is worth remembering that most of the matrixin inhibitors developed to date and harboring long side chains in their P 1 Ј position are not transition-state analogues. Therefore, the ability of matrixins to cleave substrates containing in their P 1 Ј position amino acids with unusual side chains cannot be predicted from these inhibitor studies. To address this issue more systematically in the case of MT1-MMP and ST3, synthetic heptapeptides containing in their P 1 Ј position amino acids with arylalkyl or alkyl side chains of varying size were synthesized. The amino acid sequence covering the P 3 to P 3 Ј positions of these substrates was selected according to the canonical -Pro-Leu-Gly(Ala)-Leu-Trp-Ala-sequence previously established for matrixins (38,39). In these substrates, the cleavage site has been demonstrated to occur between the Gly(Ala)-Leu residues. Accordingly, several fluorogenic substrates were prepared by substituting the Nterminal side of this canonic sequence with a dansyl group (dns), as a quencher, the tryptophan in the P 2 Ј position of this sequence being retained as a fluorophore. An arginine residue was added at the C-terminal extremity of this sequence to improve peptide solubility. In the present report, these fluorogenic peptides were evaluated as substrates for MT1-MMP and ST3. Based on this study, a fluorogenic compound, containing dinitrophenyl-coumarin as a quencher-fluorophore pair, was also synthesized and examined for comparison.
Chemistry-Unusual Fmoc amino acids not commercially available (namely, 2-aminoheptanoic acid and 2-amino-5-phenyl-pentanoic acid) were synthesized by catalytic phase transfer alkylation of ethyl diphenylmethyleneglycinate with the appropriate alkyl bromides followed by hydrolysis, as described by O'Donnell et al. (40) and O'Donnell and Eckrich (41). 2-Aminoheptanoic, 2-aminooctanoic (Aldrich) and 2-amino-5-phenyl-pentanoic acids were converted into their Fmoc derivatives by the action of Fmoc-Cl in water/Na 2 CO 3 /dioxane solution according to the standard literature procedure (42). The Fmoc amino acids were obtained as white, crystalline compounds after their purification by flash chromatography on silica gel using ethyl acetate/hexane/acetic acid eluents. Their structure and the purity were confirmed by NMR and mass spectroscopy analysis. These unusual Fmoc amino acids were used for peptide synthesis as mixtures of enantiomers. For these peptides, the diastereoisomers were separated by HPLC. N 2 -Fmoc-N 3 -2,4dinitrophenyl-L-2,3-diaminopropionic acid (Fmoc-DpaOH) was synthesized starting from Fmoc-L-AsnOH, as described by Knight et al. (14).
Substrate Synthesis and Characterization-Solid phase synthesis of the substrates was performed in a model 357 Advanced Chemtech multiple peptide synthesizer on a Rink amide resin. Typically, three equivalents of an Fmoc amino acid, three equivalents of HBTU, and five equivalents of diisopropylethylamine in N-methylpyrrolidone were added to the resin, and the coupling reaction was allowed to proceed for 30 min. The Fmoc N-protection group was removed with a 30% solution of piperidine in N-methylpyrrolidone. N-terminal acylation of the peptides was achieved either with excess dansyl chloride (20 equivalents) in the presence of diisopropylethylamine or with triple coupling of McaOH, under the conditions described above. Cleavage of the peptides from the resin, together with the cleavage of the side chain protection groups, were performed by the action of trifluoroacetic acid containing 5% triisopropylsilane.
All peptides were purified by preparative HPLC column (Vydac, 218TP1022) performed on a Gilson system equipped with a variable wavelength detector. Gradient elutions were performed using solutions A (10% acetonitrile in 0.1% trifluoroacetic acid in water) and B (90% acetonitrile in 0.1% trifluoroacetic acid in water). All peptides were recovered by lyophilization. Peptide purities were checked by amino acid analyses, analytical HPLC (Vydac, 218TP104 column) and mass spectroscopy.
Enzymes-cDNAs corresponding to the catalytic domains of mouse ST3 (Phe-102 to Ser-276) and human MT1-MMP (Tyr-111 to Arg-298) were introduced into the expression vector pET-3b, expressed in Escherichia coli BL21 (DF3) cells after isopropyl-1-thio-␤-D-galactopyranoside induction and purified essentially as described in Noël et al. (43). Briefly, both MMP catalytic domains were solubilized from bacterial inclusion bodies with 8 M urea in the presence of 100 mM dithiothreitol and purified on a Q-Sepharose anion-exchange column (Pharmacia Biotech Inc.). Purified catalytic domains were then slowly refolded at a protein concentration of 50 g/ml by dialysis to dilute out the urea. The refolding step was followed by size exclusion chromatography, using a gel filtration column (Superdex-gf 200, Pharmacia) to eliminate the aggregates, and to retain the active monomeric protein alone.
Enzyme Kinetics-Substrate specificity assays were performed in 50 mM Tris/HCl buffer, pH 7.5, 10 mM CaCl 2 , in the absence (MT1-MMP) or presence of 0.2 M NaCl (ST3), at 25°C. Substrate concentrations were determined spectrophotometrically using ⑀ 340 nm ϭ 4300 M Ϫ1 cm Ϫ1 for dansyl peptides (44) and ⑀ 328 nm ϭ 12900 M Ϫ1 cm Ϫ1 for coumarin peptides (14). Substrates were prepared as 1 mM stock solutions in dimethyl sulfoxide. Enzyme concentrations were determined from optical density, using the method of Gill and von Hippel (45) to calculate the extinction coefficient of these two matrixins. In the case of MT1-MMP, values of k cat /K m were determined from first-order full-time course reaction curves obtained at [S] Ͻ Ͻ K m (S ϭ 0.2 M), at 10 nM final enzyme concentration. These progress curves were monitored by following the increase in fluorescence at 340 nm ( ex ϭ 280 nm), induced by the cleavage of the dns substrates, in a Biologic PMS 200 spectrophotometer. Due to the lower efficiency of ST3 in cleaving this series of substrates, the observation of full-time course reactions for ST3 has required the use of higher enzyme concentrations, leading to a high fluorescence background. Thus, the kinetic parameters for ST3 were based on HPLC (Thermo Separation Products system) allowing the separation of the unreacted substrate from the cleavage products and its quantification. Immediately after the initiation of the reaction, aliquots were withdrawn from this reaction solution by the autosampler (Spectra System AS300), at predetermined time intervals, and injected onto the column. Substrate and products (S ϭ 0.2 M, ST3 concentration from 50 to 200 nM) were separated on a C18 column (Vydac, 218TP104), eluted with a linear acetonitrile gradient in 0.1% trifluoroacetic acid. Products were detected using an FL300 Spectra System fluorescence detector ( ex , 280 nm; em , 340 nm) and were identified by mass spectroscopy analysis. Data analysis was performed with a Thermo Separation Products datajet integrator.
The two approaches used for MT1-MMP and ST3 yielded product progress curves, from which values for the specificity constant k cat /K m were determined by fitting these curves with the integrated Michaelis-Menten Equation 1, by nonlinear regression (46) where k ϭ (k cat /K m )⅐E t ϭ [total enzyme]. For each k cat /K m determination, three independent experiments were taken into consideration. Individual kinetic parameters (k cat and K m ) for MT1-MMP and ST3 were obtained from analysis of fluorescence curves, under steady-state rate conditions, over 0.2 to 5 K m substrate concentration ranges. The experiments were carried out in a 5 ϫ 5 ϫ 45-mm fluorescence cell, using excitation at 280 or 300 nm (slit width, 5 nm) and emission at 360 nm (slit width, 5 nm) for the dns derivatives, and excitation at 328 nm (slit width, 5 nm) and emission at 400 nm (slit width, 5 nm) for the Mca derivatives. Cleavage of the following substrates: dns-Pro-Leu-Ala-Leu-Trp-Ala-Arg-NH 2 , dns-Pro-Leu-Ala-Cys(OMeBzl)-Trp-Ala-Arg-NH 2 , and Mca-Pro-Leu-Ala-Cys(OMeBzl)-Trp-Ala-Arg-Dpa-NH 2 , was accompanied by a 12-, 13-, and 16-fold increase in fluorescence, respectively. Initial rates measurements for twelve different substrate concentrations were performed. K m and k cat values were determined by fitting these data to the equation of Michaelis-Menten by nonlinear regression analysis.
Depending upon the concentration of substrate used in a typical assay, it is possible that fluorescence measurements may be affected by an inner filter effect. To determine the extent of this effect in the case of the dns-derivatives, the fluorescence intensity of a 10 M Leu-Trp-Ala-Arg-NH 2 solution was measured in the presence of several dns-Pro-Leu-Ala-Leu-Trp-Ala-Arg-NH 2 substrate concentrations (from 1 M to 50 M). From these experiments, a correction factor for each substrate concentration was deduced and used to correct the fluorescence observed at high substrate concentrations for the inner filter effect. The same approach was used for the Mca derivatives, using a 1 M Mca-Pro-Leu-Ala solution as the fluorescence standard.

RESULTS
The different side chains in the P 1 Ј position of our series of fluorogenic substrates are described in Fig. 1. Arylalkyl side chains longer than that of phenylalanine were selected to assess the effect of the side chain's length on MT1-MMP1 and ST3 cleaving activities. In addition, substrates with arylalkyl side chains containing an oxygen or sulfur heteroatom were also examined. The effects of alkyl side chains longer than leucine were also studied.
Products resulting from the cleavage of this series of substrates by MT1-MMP or ST3 were characterized both by HPLC and mass spectroscopy. For each substrate, we observed a single cleavage site between the expected Ala-Xaa peptide bond, with no inhibition of the reaction by the released products (Fig. 2, and data not shown). Values of the specificity constant k cat /K m that were determined from the first-order progress curves for the hydrolysis of substrates by MT1-MMP are reported in Table I. The most rapidly cleaved substrate by this enzyme was that harboring the longest side chain in the P 1 Ј position (Cys(OMeBzl)). In the arylalkyl series, the lengthening of the side chain by one (hPhe) and two (pPhe) methylene(s) resulted in a marked increase in catalytic efficiency, as compared with the substrate containing phenylalanine in the P 1 Ј position. The comparison of substrates with a side chain of equal length (pPhe, Ser(Bzl), Cys(Bzl)) reveals the significant role played by a sulfur atom in the ␥ position of the side chain, which was associated with a much faster cleavage rate of the substrate by MT1-MMP. In the alkyl series, the rate of hydrolysis was also found to depend on the length of the side chain in the P 1 Ј position. Comparison between the arylalkyl and alkyl series of substrates reveals that, in addition to the side chain length, the presence of the phenyl aromatic group is a determinant factor for the optimization of the rate of hydrolysis, since the substrates in the alkyl series were always cleaved at a slower rate than their counterparts in the arylalkyl series ( Table I). The substitution Leu 3 Cys(OMeBzl) in these substrates led to a 37-fold increase in the k cat /K m ratio (Table I). In the case of ST3, the preferred side chain in the P 1 Ј position of the substrate was also the longest arylalkyl one (Cys(OMeBzl)) ( Table I). However, the substrates containing leucine, methionine or n-pentyl in the P 1 Ј position were cleaved more rapidly than that with an n-hexyl side chain. As compared with MT1-MMP, the former three substrates were rather well cleaved by ST3, while the n-hexyl compound was a poor substrate.
The K m , k cat and k cat /K m values determined for two substrates (Xaa ϭ Leu and Cys(OMeBzl)) are reported in Table II. For these matrixins, the substitution Leu 3 Cys(OMeBzl) was found to increase both the substrate affinity and the k cat value. However, while these two substrates displayed similar K m for MT1-MMP and ST3, the k cat values determined for ST3 on these substrates were about two orders of magnitude lower than those measured for MT1-MMP. The free energy difference (⌬⌬G ‡ ) associated with the substitution Leu 3 Cys(OMeBzl) was evaluated for each enzyme from the ratio of the k cat /K m values determined for these two substrates. For MT1-MMP, this substitution corresponds to a free energy change of 2 kcal/mol, while the same modification causes a free energy change of 1.55 kcal/mol for ST3.
The substitution of alanine by phenylalanine in the P 1 position, as well as the replacement of the Pro-Leu sequence by alanine residues in the P 3 and P 2 positions, led to substrates which were poorly hydrolyzed by MT1-MMP (Table I). In contrast, these two substrates were rather well cleaved by ST3. Thus, based on these preliminary data, the specificity requirement of the non-primed subsites of ST3 appears quite different from that observed for MT1-MMP.
The finding that the best substrate in this series was that containing a Cys(OMeBzl) residue in the P 1 Ј position led us to develop a fluorogenic substrate, Mca-Pro-Leu-Ala-Cys(OMeBzl)Trp-Ala-Arg-Dpa-NH 2 , characterized by the presence of the Mca-Dpa fluorophore-quencher pair. Such a fluorophorequencher pair has been demonstrated to provide more sensitive fluorescent assays than those based on the dns-Trp pair. Using the Mca-Pro-Leu-Ala-Cys(OMeBzl)Trp-Ala-Arg-Dpa-NH 2 substrate, the enzymatic activity of a 5 nM ST3 solution can be determined with high accuracy, in less than 20 min. On this time scale, only a very small fraction of the commercially available Mca-Dpa substrate for matrixins (Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH 2 ) was cleaved with the same amount of ST3,

DISCUSSION
In this study, we demonstrate that the substitution of the P 1 Ј position in synthetic substrates by unusual long arylalkyl side chains leads to peptides which are cleaved by MT1-MMP and ST3 much more rapidly than the corresponding substrates containing natural phenylalanine or leucine amino acids. The dns substrate with a Cys(OMeBzl) residue in the P 1 Ј position exhibited a k cat /K m value of 1.59 10 6 M Ϫ1 s Ϫ1 when tested with MT1-MMP, at 25°C. This substrate is therefore one of the most rapidly cleaved synthetic substrates so far reported for matrixins. For example, a k cat /K m value of 1.57 10 5 M Ϫ1 s Ϫ1 has been determined for the cleavage of the Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH 2 substrate by MT1-MMP, at the same pH, but at 37°C (13). Furthermore, the first synthetic substrates cleaved by ST3 were identified in this series. The cleavage of the Mca-Pro-Leu-Ala-Cys(OMeBzl)-Trp-Ala-Arg-Dpa-NH 2 fluoro-genic peptide provides a simple test for routine assay of ST3, useful for ST3 inhibitor screens. The influence of unusual amino acid substitutions in the P 1 Ј position of synthetic substrates was previously evaluated for two matrixins, COL1 and GEL B (37). A small increase (4-fold) in the k cat /K m value was observed for substrates containing ␥-S substituted cysteine in the P 1 Ј position, as compared with the parent compound with leucine in the P 1 Ј position. Based on the present work, this modest effect may be due to the fact that only ␥-S substituents of reduced size, with a methyl and an ethyl group, were investigated in this study.
The substitution Leu 3 Cys(OMeBzl) in the P 1 Ј position of the substrate improved the K m value of the substrate for MT1-MMP, a finding in agreement with x-ray data (25,29,30) and inhibitors studies (31)(32)(33)(34)(35). This suggests that in all matrixins containing a leucine residue in their S 1 Ј pocket, this pocket should be a deep cavity able to accommodate extremely long amino acid side chains. In contrast, the increase in substrate affinity resulting from the Leu 3 Cys(OMeBzl) substitution for ST3 is unexpected. ST3 is characterized by the presence of glutamine in its S 1 Ј subsite, a property which was predicted to confer a shallow S 1 Ј pocket to ST3 (36), paralleling the situation observed in the x-ray structures of COL1 and matrilysin (24,26,27). A possible interpretation of our results is that the glutamine side chain in the ST3 S 1 Ј pocket can move, allowing this matrixin to bind and cleave substrates harboring long side chains in the P 1 Ј position. Such a scenario has been proposed during inhibitor binding for the arginine side chain of COL1, since this matrixin was unexpectedly found to be potently inhibited by synthetic compounds with long side chains at the P 1 Ј position (34). A shift of the glutamine side chain in ST3, allowing substrate binding, may explain why the free energy change associated with the Leu 3 Cys(OMeBzl) substitution is lower for ST3 than MT1-MMP. In the case of ST3, the interactions engaged between the P 1 Ј position of the substrate and the S 1 Ј pocket would have to compensate for the energy associated with the displacement of the glutamine side chain.
Interestingly, the modification Leu 3 Cys(OMeBzl) not only improves the K m , but also the k cat value of the substrate for both MT1-MMP and ST3 enzymes. Therefore, for these two matrixins, the binding energy associated with the interaction of substrates containing an unusual amino acid side chain in the P 1 Ј position with the enzyme catalytic site is used to stabilize the energy of the substrate-enzyme complex both in the ground and transition-states. This last effect, as mentioned in the introduction, could not be predicted from previous studies aimed at evaluating matrixin-inhibitor interactions, since most inhibitors so far developed were not analogues of the substrate in the transition-state (20). For a long time, the functional significance of a deep S 1 Ј subsite in matrixins able to accommodate long P 1 Ј groups of inhibitors was a subject of debate. In this respect, our data clearly indicate that the filling of this S 1 Ј subsite by an appropriate group in the P 1 Ј position yields substrates with high k cat /K m values.
The k cat values reported in this study for two substrates are two orders of magnitude lower for ST3 than those observed for MT1-MMP. These low turnover numbers are in agreement with the very weak proteolytic activities reported for ST3 against most usual matrixin substrates, with the exception of ␣1 proteinase inhibitor (43,47,48). Arguments pointing to ST3 as a particular member of the matrixin family have been presented before and include, in addition to unusual proteolytic properties, both unusual proform processing (10,49) and regulation of gene expression (50,51). However, our observations do not rule out the possibility that ST3 may rapidly hydrolyze in vivo some, as yet unexamined, substrates. Our preliminary data on the specificity of ST3 for the nonprimed residues of synthetic substrates, in comparison to MT1-MMP, support the possibility that ST3 could cleave substrates which are not presently considered as matrixin substrates. Many reports have documented the preference of matrixins for small residues in the P 1 position of the substrate, as well as the key role of proline in their P 3 position (15, 16), a feature apparently not conserved in the case of ST3 (Table I). It is therefore important to define more exhaustively the specificity requirements of ST3 for the non-primed positions of the substrate. As illustrated in the present study, where only one position of the substrate was exhaustively investigated, such optimization could dramatically increase the k cat /K m value, as this involves a gain of only a few kilocalories/mol in the enzyme-substrate binding energy (Table II). An important issue is to determine whether it is possible to identify specific peptide sequences that are hydrolyzed by ST3 with an efficiency similar to that reported for the other matrixins. Our data showing that ST3 efficiently cleaves substrates harboring unusual long side chains in their sequence suggest that this matrixin may hydrolyze, in vivo, proteins also containing unusual amino acids. In this respect, extracellular proteins undergoing particular post-translational modifications may represent potential targets for ST3. Although a large number of protein substrates have already been described for MT1-MMP, as well as for the other matrixins containing a leucine residue in their S 1 Ј subsite, our observations raise the possibility that these matrixins may also cleave, Dns-Pro-Leu-Ala-Cys(OMeBzl)-Trp-Ala-Arg-NH 2 159 Ϯ 0.44 1.67 Ϯ 0.04 Dns-Pro-Leu-Ala-Cys(Bzl)-Trp-Ala-Arg-NH 2 65 Ϯ 0.32 1.12 Ϯ 0.09 Dns-Pro-Leu-Ala-Ser(Bzl)-Trp-Ala-Arg-NH 2 25 Ϯ 0.05 0.70 Ϯ 0.01 Dns-Pro-Leu-Ala-pPhe-Trp-Ala-Arg-NH 2 24 Ϯ 0.07 0.18 Ϯ 0.009 Dns-Pro-Leu-Ala-hPhe-Trp-Ala-Arg-NH 2 9.5 Ϯ 0.06 0.27 Ϯ 0.03 Dns-Pro-Leu-Ala-nHex-Trp-Ala-Arg-NH 2 6.3 Ϯ 0.04 0.001 Ϯ 0.004 Dns-Pro-Leu-Ala-Leu-Trp-Ala-Arg-NH 2 4.2 Ϯ 0.03 0.18 Ϯ 0.0015 Dns-Pro-Leu-Ala-Met-Trp-Ala-Arg-NH 2 3.6 Ϯ 0.03 0.52 Ϯ 0.006 Dns-Pro-Leu-Ala-nPent-Trp-Ala-Arg-NH 2 3.1 Ϯ 0.02 0.30 Ϯ 0.005 Dns-Pro-Leu-Ala-Nle-Trp-Ala-Arg-NH 2 1.8 Ϯ 0.01 0.13 Ϯ 0.005 Dns-Pro-Leu-Ala-Phe-Trp-Ala-Arg-NH 2 1.1 Ϯ 0.01 0.02 Ϯ 0.006 Dns-Pro-Leu-Phe-Cys(Bzl)-Trp-Ala-Arg-NH 2 Ͻ1 0.52 Ϯ 0.03 Dns-Ala-Ala-Ala-Cys(Bzl)-Trp-Ala-Arg-NH 2 Ͻ1 0.90 Ϯ 0.02 in vivo, proteins at sites characterized by the presence of amino acids with unusual long side chains.