Activity of guinea pig liver transglutaminase toward ester analogs of amide substrates.

Abstract Calcium-activated guinea pig liver transglutaminase catalyzes the hydrolysis and aminolysis of γ-glutamyl methyl and ethyl ester analogs of a specific amide substrate, benzyloxycarbonyl(Z)-l-glutaminylglycine. Kinetic comparisons of amine incorporation into the γ-n-methylamide, and the γ-methyl and ethyl esters of Z-l-glutamylglycine indicate that the same aycl-enzyme intermediate is formed from each of these substrates. The differences in catalytic efficiency toward these substrates are due to differences in rates of formation of this intermediate, and may be a consequence of different orientations of the leaving groups (—OR and —NHR) at the catalytic site of the enzyme. An aliphatic ester, methyl acetate, and its amide analog, N-methylacetamide, are about equally effective, although poor, substrates for transglutaminase. This, together with the fact that the acyl-enzyme intermediate is formed from the γ-methyl ester of Z-l-glutamylglycine and from its γ-N-methylamide at approximately the same rate, is strong evidence that C—O and C—N bonds are about equally susceptible to cleavage by transglutaminase. The ability of an alcohol to function as an acyl acceptor in a transglutaminase-catalyzed amide transfer reaction and the reversibility of the ester to amide conversion has been demonstrated with the use of a fluorescent-labeled alcohol derivative, N-(5 - dimethylaminonaphthalenesulfonyl) - 5 - amino - 1 - pentanol.


SUMMARY
Calcium-activated guinea pig liver transglutaminase catalyzes the hydrolysis and aminolysis of y-glutamyl methyl and ethyl ester analogs of a specific amide substrate, benzyloxycarbonyl(Z)-L-glutaminylglycine.
Kinetic comparisons of amine incorporation into the y-n-methylamide, and the -y-methyl and ethyl esters of Z-L-glutamylglycine indicate that the same aycl-enzyme intermediate is formed from each of these substrates.
The differences in catalytic efficiency toward these substrates are due to differences in rates of formation of this intermediate, and may be a consequence of different orientations of the leaving groups (-OR and -NHR) at the catalytic site of the enzyme. An aliphatic ester, methyl acetate, and its amide analog, N-methylacetamide, are about equally effective, although poor, substrates for transglutaminase.
This, together with the fact that the acyl-enzyme intermediate is formed from the ymethyl ester of Z-L-glutamylglycine and from its r-Nmethylamide at approximately the same rate, is strong evidence that C-O and C-N bonds are about equally susceptible to cleavage by transglutaminase.
The ability of an alcohol to function as an acyl acceptor in a transglutaminase-catalyzed amide transfer reaction and the reversibility of the ester to amide conversion has been demonstrated with the use of a fluorescent-labeled alcohol derivative, N-(5 -dimethylaminonaphthalenesulfonyl) -5 -amino -lpentanol.
The kinetic data of hydrolysis and amine transfer (5, 6) c011form to .\Icchanism I, in \vhich ai, ncyl-c~izyme, F', is partitioned between water and another acceptor nuclcophilr, * This communication is the ninth paper in the series "hlechanism of Action of (;uinea Pig Liver Transglutanlirl:~sc." The preceding paper is llcf. 7. E F II, i.e. a primary amine.
In this mechanism, il is amide or active ester substrate, I' is the protonatcd leaving group, R is the product of hydrolysis, and & is the product of transfer to amine.
;\ctive site mapping (2) and "reporter group" studies (7) supply evidence that the glutamine side chain binding site of transglutaminase is a hydrophobic crevice in I\-hich is contained the active site --SII group and which assumes dimensions of appro~irnately 5 X X 5 h prior to, or during, acylatioii of the enzyme by substrate.
Straight chain and y-branched chain aliphatic amides are accommotlatctl within this active site pocket and, thus, are substrates by virtue of their simulatioii of the side chain portions of a pcptide-bound glutamine residue. Aliphatic amides that contain branches on their o( or /3 carbon atoms bind, but prevent proper formation of this active site pocket, and so do not act as substrates. Support for this concept comes from the finding that peptitle derivatives of cr-riictliylglutamilic act as substrates for the enzyme, whereas those of the isomers of fl-and y-methylglutamine do not (8). The specificity of transglutaniinase toward active esters of aliphatic carbosylic acids is different.
The results reported hcrc show that transglutarninase catalyzes the hydrolysis and aminolysis of esters that are not active esters. Further, they demonstrate that the specificity of the enzyme to\rartl these esters is a function of their similarity to the glutamine rrsidue, rather than of a relationship to active esters, and they indicate that tlicsc ester bonds arc as susceptible as amidc bonds to cleavage by the enzyme.  (14).

RESULTS
That Z-r-glutamyl(y-N-methylamido)glycine acts as a substrate for transglutaminase in the [i4C]mct,hylamine incorporation reaction (Table I) was as expected.
This is an example of an isotope exchange reaction, several of which have already been sl~om~~ to be catalyzed by the enzyme (5, 6). We did not anticipate that Z-L-glutamyl(y-ethyl ester)glycine and Z-Lglutamyl(y-methyl ester)glycine, the ester analog of the r-Nmethylamidc, would function as substrates. In fact,, it was concluded from an early survey of transglutaminasc-catalyzed hydroxamate formation that the y-cthyl ester acted as an inhibitor, but not as a substrate, for the enzyme (15). This incorrect conclusion is understandable in light of the poor substrate properties of this ester compared to Z-L-glutaminylglycinc (Table I) and the insensitive assay used.
When the initial velocities of methylaminc incorporation into the y-methyl and y-ethyl esters of Z-L-glutamylglyciIle were measured at several concentrations of [i4C]mcthylamine and the data were plotted in the usual fashion (for esamplc, see Ref. 6), intersecting patterns were obtained that were consistent with the steady state equat,ion (Equation 2) for formation of &, radioactive product, in Mechanism I. The kinetic constants, estimated from these data, together with those obtained for the glutamine substrate and for the y-N-mcthylamide substrate, are given in Table I. Also given in Table I are the K, values and maximum velocities for hydrosamatc formation with these substrates obtained at a high level of hydro~ylamine.
The question arises, is the activity of transglutaminase toward the two esters listed in Table I manifest as a consequence of  their structural  similarity  to glutamine  (and r-N-substituted  glutamine) substrate? There is substantial evidence that straight chain aliphatic amides, such as acctamide, function as substrates for transglutaminase because of their structural relationship to the side chain portion of peptide-bound glutamine residues, i.e. that these aliphatic amides bind at the glutamine side chain binding site in the enzyme (2). Since the y-N-methyl derivative of Z-L-glutaminylglycine and its ester analog are substrates (Table I), one would anticipate that N-methyl acetamide and its ester analog, methyl acetate, would also act as substrates.
Indeed, this proved to be the case. &rues 4 and 5 of Fig. 1 show the slow, but similar, rates of monodansylcadaverine incorporation into the two compounds. The rates of fiuorescent amine incorporation into acetamide (Curve .2) and into Z-Lglutaminylglycinc (Curve 1) under the same experimental conditions arc shown in this figure for comparative purposes.
It was not possible, because of their very limited water solubility, to test the substrate properties of branched chain aliphatic esters, such as methyl isobutyrate and methyl isovalerate. These esters arc analogs of the N-methyl derivatives of those branched chain amides that do not act as substrates, i.e. Q-methylprol~ionan~idc and /!-methylbutyramidc, respectively (2). Therefore, Z-r-aspartyl(/%cthyl cster)glycine was tested. This compound is the ester analog of the N-methyl derivative of Z-r-asparaginylglycinc.
The asparagine peptide derivative is not a substrate for transglutaminasc (5, 16). Z-r-Aspartyl(@ ethyl ester)glycinc did not act as a substrate (Curve 6, Fig. 1). No incorporation of fluorescent amine was found even after 24 hours of incubation under conditions of Fig. 1.
Formamide is not a substrate for transglutaminase (2). It has been suggested that this is the case because this amide has no hydrophobic attraction to the glutamine side chain binding site of the enzyme. We tested methyl formatc, the ester analog FIG. 1. Incorporation of monodansylcadaverine (0) and 5dansylamino-1-pentanol (0) as catalyzed by transglut,aminase. The reactions were carried out at 37" in 0.1 M Tris-acetate buffer, pH 6, containing 0.2 rnM fluorescent amine or fluorescent alcohol; 50 mM CaCl,, 1 mM EDTA, and 1 mg of enzyme per ml. The levels of amides and esters were: Curve 1, 60 mM Z-L-glutaminylglycine; Curve 2,500 mM acetamide; Curve S, 60 mM Z-L-glutaminylglycine; Curve 4, 500 rnM methyl acetate; Curve 5, 500 mM N-methyl acetamide; Curve 6, 500 mM methyl formate or GO m&f Z-L-aspartyl(Pethyl ester)glycine.
At the times shown, 1-J aliquots of reaction mixtures were applied to polyamide sheets ("Methods"), quickly dried in a stream of air, and chromatographed.
The percentage of incorporation designates the percentage of the total amine or alcohol incorporated. of the N-methyl derivative of formamide, and found that, indeed, this ester is not a substrate (Curve 6, Fig. 1).
The ability of an alcohol to act as an acceptor nucleophile in a transglutaminase-catalyzed reaction is demonstrated clearly in Curve S of Fig. 1. Here 5-dansylamino-1-pentanol, the alcohol analog of monodansylcadaverine, is incorporated in place of the -NH2 at the carboxamide group of Z-L-glutaminylglycine. The slow rate of incorporation compared to that of monodansylcadaverine is evident by comparison of Curves 1 and 8, Fig. 1. DISCUSSION We have observed that certain esters of aliphatic alcohols are substrates for transglutaminase. Substantial evidence is presented here that the substrate property of these esters derives from their structural relationship to amide substrates. Only those esters that correspond in their acyl portion with amide substrates are acted upon by the enzyme.
The kinetic data for amine incorporation with the y-methyl and y-ethyl esters of Z-L-glutamylglycine conform to Mechanism I and are, thus, consistent with the acyl-enzyme theory of transglutaminase action (4-6).
The acyl-enzyme mechanism predicts formation of a common covalent intermediate from the enzyme and the acyl portion of substrates that contain the same acyl group. The substrates listed in Table I contain the same acyl group, Z-L-glutamylglycine.
That the Kzbb values for these substrates are the same, within experimental error, is strong indication that, indeed, the same acyl-enzyme intermediate is formed during catalysis of each substrate.
This kinetic constant, Kzbb (Equation 3), is composed of rate constants that relate to the (3) reactions of acyl-enzyme intermediate with water and added nucleophile, methylamine in this case. It is totally independent of the rate of formation of acyl-enzyme.
Pronounced differences in Kibb values have been found for substrates that contain different acyl groups (5, 8), as well as for substrates that contain the same acyl group, but where different nucleophiles were added (5). The fact that the substrates of Table I show the same Kibb  values with methylamine is further evidence in support of the acyl-enzyme mechanism.
Accepting the evidence for a common acyl-enzyme intermediate formed from each substrate of Table I, one may conclude that, with t'he y-N-methylamide and the y-methyl and y-ethyl esters ka, the rate of enzyme acylation is limiting for both hydrolysis and transfer to amine, i.e. k3 < k7 and kg. This follows from the facts that (a) V, = vab (Equations 4 and 5, respecv = k&w% a kz + ks tively) for each substrate; (b) V, and vab = -kaE when k3 < k7 and kg; and (c) these maximum velocities are significantly smaller than the V, value for Z-L-glutaminylglycine. With this glutamine substrate acylation is not the rate-controlling step for hydrolysis, i.e. k3 > kg. This has been pointed out earlier (5) and derives from the fact that the velocity for amine incorporation is significantly faster than that for hydrolysis, i.e. v,b > V,. It is not known whether k3 is limiting for transfer with the glutamine substrate.
It is, however, obvious that the rate of acylation, ka, with this substrate is much faster than with the other substrates in Table I. The maximum velocities for hydroxylamine incorporation recorded in Table I are larger than those for methylamine incorporation.
It has been suggested that the normal enzymatic mechanism may be perturbed by the high concentration of hydrosylamine necessary in the assays (5). Inconsistencies with the acyl-enzyme theory observed with other enzymes in esperiments with hydroxylamine have been assumed to be a result of mechanism perturbations caused by this strong nucleophile (17). Comparison of the V& and V,,, values in Table I suggests that the rates of enzyme acylation with the N-methylamide and esters are also rate-limiting for hydroxamate formation. The relative differences between the maximum velocities for transfer to methylamine and hydroxylamine are similar and the V,,, value for the glutamine substrate is much larger than those for the methylamide and esters. The fact that Z-Lglutamyl(y-N-methylamido)glycine and its methyl ester analog display almost identical maximum velocities for both methylamine and hydrosylamine incorporation is strong evidence for approximately equal rates of enzyme acylation by these two compounds which differ only in the scissile bond, i.e. a C-N or a C-O bond. Thus, it may be concluded that these bonds are about equally susceptible to cleavage by transglutaminase. This is supported by the similar rates of monodansylcadaverine incorporation with N-methylacetamide and its ester analog, methyl acetate (Fig. 1). It is likely that enzyme acylation is also limiting with these acetyl derivatives, as appears to be the case with acetamide (2).
The chemical structure of the side chain portion, R, in the leaving group (-OR or --NH@ of transglntaminase substrates has a notable influence upon the rate of enzyme acylation. This has been observed previously with amide substrates (5, 18) and is obvious here in a comparison of the turnover values for Z-Lglutaminylglycine and its N-methyl derivative.
In this case substitution of a methyl for a hydrogen in the R position of the leaving group results in a large decrease in the rate of enzyme acylation.
Comparison of the maximum velocities for the two ester substrates of Table I shows that a small change in the size of the R group, -CH3 to -C&I-IS, results in a significant change in the rate of enzyme acylation.
These rhanges may be a consequence of different arrangements of the leaving groups in the active site of the enzyme. Dislocation of the amide or ester bond enough to change, but not to curtail, acylation could result.
When lia is rate-limiting, i.e. ka < ks and &, as in the case of the y-iV-methylamide and the esters of Table I, and OIK assumes lcz > ka, K,t (Equation 6) reduces to I<i, (Equation 7), the enzyme-substrate dissociation constant. The differences in Kat values for these three substrates are not great. In the hydroxylamine incorporation reaction enzyme acglation may bc rate-limiting with Z-L-glutaminylglycine, as n-cl1 as with the other substrates.
If so, the recorded I<, values are comparable measures of enzyme-substrate affinity under the conditions of this assay. That these values are essentially the same for all of the substrates is in agreement with earlier indications that hydrophobiL interactions alone in the glutamine side chain binding region of the active site of the enzyme do not play the major role in the over-all binding of substrates (8). There is evidence that the amino acids (or other residues) surrounding the glutamine moiety make a significant contribution to this binding (8).
The p-nitrophenyl (active) ester hydrolysis and transfer rcactions catalyzed by transglutaminasc are consistent with Mechanism I (5). The isolation of a stable trimethylacetyl-enzyme formed during the hydrolysis of p-nitrophenyl trimethylacetate (4) is strong support for an acyl-enzyme mechanism with the active esters. However, differential losses in the catalytic activities of transglutaminase toward amide and active ester substrates have been observed as a result of chemical modifications of the enzyme protein (U-20).
These modifications that affect esterase and amidase activities to different degrees are evidence for some differences in the mechanisms for active esters and amides. The present finding that esters of aliphatic alcohols function as transglutaminase substrates only by virtue of their structural similarity to amide substrates is in agreement. p-Nitrophenyl rsters of o(-and &branched acids, i.e. p-nitrophenyl isobutyrate (5) and p-nitrophcnyl isovalerate,' are substrates fo1 transglutaminase, whereas amides of these branched chain acids are not (2). Formamide (2) and methyl formate (Curve 6, Fig.  1) do not act as transglutaminase substrates. However, the enzyme does catalyze the hydrolysis of p-nitrophenyl formate.' That an aliphatic alcohol derivative, 5-daiisvlamino-l-peritanol, functions as an acceptor for the acyl group in a transglutaminase-catalyzed amide reaction (Fig. 1) is consistent with the action of the enzyme 011 esters of aliphatic alcohols and clearly demonstrates the reversibility of the reaction. The slow rate of incorporation of this alcohol derivative comparcd to monodansylcadaverinc, its amine analog (Curves 1 and S, Fig. I), probably results from its waker nuclcophilicity.
Ac!xnowledgmenl-The expert technical assistance of Xliss Norma K. Whetzel is gratefully acknowledged.