Purification of the enzyme and cloning of the gene

Although the absolute configurations of the products are opposite to that of antiinflammatory active compounds, and the substrate specificity is rather restricted as to the steric bulkiness around the reaction center, the enzyme system of A. bronchisepticus was proved to have a unique reactivity. Thus, detailed studies on the isolated enzyme were expected to elucidate some new interesting mechanism of the new type of decarboxylation. Thus, the enzyme was purified. (The enzyme is now registered as EC 4.1.1.76.) The molecular mass was about 24 kDa. The enzyme was named as arylmalonate decarboxylase (AMDase), as the rate of the decarboxylation of phenylmalonic acid was faster than that of the a-methyl derivative.9

To clarify the characteristics of AMDase, the effects of some additives were examined using phenylmalonic acid as the representative substrate.9 The addition of ATP and coenzyme A did not enhance the rate of the reaction, different from the case of malonyl-CoA decarboxylase and others; in those, ATP and substrate acid form a mixed anhydride, which in turn reacts with coenzyme A to form a thiol ester of the substrate. In the present case, as both ATP and CoA-SH had no effect, the mechanism of the reaction will be totally different from the ordinary one described above. It is well established that avidin is a potent inhibitor of the formation of the biotin-enzyme complex.10-13 In the case of AMDase, addition of avidin has no influence on the enzyme activity, indicating that AMDase is not a biotin enzyme.

Thus, AMDase requires no cofactors and this fact is entirely different from those of known analogous enzymes, such as acyl-CoA carboxylases,14 methylmalonyl-CoA decarboxylases10 and transcarboxylases.1415

A strong inhibitory effect on AMDase activity was found for sulfhydryl reagents (at 1 mM), such as HgCl2 (relative activity, 0%), HgCl (8%), AgNO3 (3%), iodoacetate (3%) and p-chloromercuribenzoate (PCMB) (0%). N-ethyl maleimide (at 10 mM) causes 72% inhibition of the decarboxylase activity. Accordingly, AMDase was revealed to be a thiol decarboxylase, i.e., at least one of the cysteine residues is present as a free SH form and plays an essential role in the active site of the enzyme. The activity of the enzyme was not affected upon incubation with the following reagents: several divalent metal ions, such as Ni2+, Co2+, Ba2+, Mg2+ and Ca2+, carbonyl reagents, such as NaN3, NH2OH, KCN, metal-chelating agents, such as EDTA, 8-quinolinol, bipyridil, 1,10-phenanthroline, and serine inhibitors, such as phenylmethanesulfonyl fluoride (at 10 mM). In conclusion, AMDase can be considered an unusual enzyme containing neither metal ions nor coenzymes, which ordinary decarboxylases and transcarboxylases have.

For more detailed studies on this unique enzyme, the gene of AMDase was cloned using the direct expression method. The gene was clarified to be consisting of 720 bp, indicating that the enzyme consists of 240 amino acids (Fig. 9).

A Pstl-Hindlll (1.2 kbp) fragment was subcloned in pUC19. The enzyme produced by the E. coli transformant was purified to homogeneity and shown to be identical to that of the original strain. Both enzymes had the same enzymological properties and N-terminal amino acid sequences.16

2.3. Reaction mechanism

2.3.1. Electronic effect

To study the reaction mechanism, the electronic effect of the substituents (p-MeO, p-Me, p-Cl, m-Cl and H) on the rate of the reaction of phenylmalonic acid was examined. The logarithm of kcat (X)/kcat (H) cleanly correlated in a linear fashion to Hammett a values (Fig. 10). The p-value was +1.9.9 The positive sign of the p-value indicates that the transition state has some negative charge. Thus, the most probable intermediate is the enolate form of the product as shown in the bottom part of Fig. 10.

2.3.2. Stereochemistry

As shown above, the electronic properties have a serious effect on the rate of the reaction. It means that the aromatic ring should occupy the same plane with that of the estimated intermediate enol moiety. Then, it is supposed that the conformation of the substrate is already restricted when it binds to the active site of the enzyme. The evidence that supports this estimation is the inactiveness of a-methyl-o-chlorophenyl and a-naphthylmalonic acids. This is a marked difference with the fact that a-methyl-p-Cl-phenyl and methyl-^-naphthylmalonic acids are

30 60

atg cag caa gca age act ccc acc ate ggc atg ate gtg cog ccc gco gcg ggt ctg gtg Met Gin Gin Ala Ser Thr Pro Thr lie Gly Met lie Val Pro Pro Ala Ala Gly Leu Val

90 120

cog gcg gat ggg gcg egg ctc tat ccc gat ctg ccc ttc att gcc age ggg ctg ggg ctg Pro Ala Asp Gly Ala Arg Leu Tyr Pro Asp Leu Pro Phe lis Ala Ser Gly Leu Gly Leu

150 180

ggc tcc gte acg cog gaa ggc tat gac gcc gtg ate gaa teg gtg gtg gae cat geg egc Gly Ser Val Thr Pro Glu Gly Tyr Asp Ala Val He Glu Ser Val Val Asp Kis Ala Arg

210 240

ego ctg caa aag cag ggc gcg gcg gtg gtt teg ctg atg ggc ace teg etc ago tte tac Arg Leu Gin Lys Gin Gly Ala Ala Val Val Ser Leu Met Gly Thr Ser Leu Ser Phe Tyr

egg ggc gcg gee tte aat gee gcg ttg aec gta gcg atg egc gaa gcc aeg gga ctg eea Arg Gly Ala Ala Phe Asn Ala Ala Leu Thr Val Ala Met Arg Glu Ala Thr Gly Leu Pro

tgc laeg ace atg age aeg geg gte etg aac gga ttg ege gee ctg ggc gtg ege cgc gtc Cys (Thr Thr Met Ser Thr Ala Val Leu Asn Gly Leu Arg Ala Leu Gly Val Arg Arg Val (101)1 390 420

gcg ttg gcg acg gcc tat ate gac gat gtg aac gag cgc ctg gcg gca ttc ctg gcc gaa Ala Leu Ala Thr Ala Tyr lie Asp Asp Val Asn Glu Arg Leu Ala Ala Phe Leu Ala Glu

450 480

gag ago ctg gtt ccc acc ggc Glu Ser Leu Val Pro Thr Gly egc gtg gat acg gcc acg ctg Arg Val Asp Thr Ala Thr Leu Val age gac ggc ate ctg ctg tct Ser Asp Gly He Leu Leu Ser tgc fegc age ctt gge ate aeg ggc gtg gag gco atg gog Oysferg Ser Leu Gly He Thr Gly Val Glu Ala Met Ala

gac ctg tgc jtg cgt gcc ttc gaa gcg geg ccc gat Asp Leu Cys /aI Arg Ala Phe Glu Ala Ala Pro Asp t 570 |(i7i) 600

gc ggc ttg etg acg ctg gac gee ata ccc gaa gtc ly Gly Leu Leu Thr Leu Asp Ala He Pro Glu Val 630 660

... __ ____ _ ____tg teg agt teg ccg gcg ggg ttc tgg gac gcc gtg

Glu Arg Arg Leu Gly Val Pro Val Val Ser Ser Ser Pro Ala Gly Phe Trp Asp Ala Val

690 720

egg ctt gcg ggg gga ggg gee aag gea agg ceo gga tac ggc egg otg tte gae gag too Arg Leu Ala Gly Gly Gly Ala Lys Ala Arg Pro Gly Tyr Gly Arg Leu Phe Asp Glu Ser tga

Figure 9: Nucleotide and deduced amino acid sequences of AMDase.

very reactive substrates. The different reactivities can be explained by supposing that the reaction proceeds smoothly when the conformation of the substrate is arranged in such a way that the o-substituent and a-methyl group take the syn-periplanar conformation. In the case of o-chlorophenyl-a-methylmalonic acid, it will be difficult for the substrate to take the syn-periplanar conformation because of the steric repulsion between chlorine and the methyl groups. o-Methylphenyl-a-methylmalonic acid was also revealed to be inactive and the reason is estimated to be the same. If this estimation is correct, fixing the two methyl groups in the syn-periplanar conformation will make the decarboxylation reaction proceed. The only way to fix the conformation of the substrate in an unfavorable one is to connect the two groups with a covalent bond. Thus, indane dicarboxylic acid is considered to be a good model of syn-periplanar conformation of o-methylphenyl-a-methylmalonic acid. As expected, cyclic substrate was smoothly decarboxylated to give the corresponding (R)-monobasic acid in high chemical and optical yields (Fig. 11).17

p- H.9 j

H ^ •//

p-CI

m-CI ..

p-HoCO —,,, 3 m^ P"CH3

Hammet constant (<r)

Hammet constant (<r)

Figure 10: Hammett plot of the kcat of the AMDase-catalyzed decarboxylation of substituted phenylmalonate.

Figure 10: Hammett plot of the kcat of the AMDase-catalyzed decarboxylation of substituted phenylmalonate.

It is noteworthy that the Km value of this substrate is smaller by one order compared to non-cyclic compounds. According to the discussions proposed above, this is considered to be due to its conformation already being fixed to the one that fits to the binding site of the enzyme. This estimation was demonstrated to be true by the examination of the effect of temperature on the kinetic parameters. Arrhenius plots of the rate constants of indane dicarboxylic acid and phenyl-malonic acid showed that the activation entropies of these substrates are -27.6 and -38.5 cal mol-1K-1, respectively.18 The smaller activation entropy for the cyclic compound demonstrates that the s;y«-periplanar conformation of the substrate resembles the one of the transition state.

The other interesting problem concerning the stereochemistry of the reaction is the mode of enantiotopos-differentiation. Does the enzyme distinguish two prochiral carboxyl groups? The clue to elucidation of this problem is to prepare both enantiomers of a-methyl-a-phenylmalonic acid which have 13C on either one of two carboxyl groups. Starting from 13C-phenylacetate, via optical resolution of an intermediate, both enantiomers of chiral 13C-containing a-methyl-a-phenylmalonic acid were prepared. The absolute configuration of the chiral

Figure 11: Effect of the conformation on the reactivity.

substrate was unambiguously determined by the optical rotation of the resolved intermediate.

The result of enzymatic decarboxylation was extremely clear.19 While (S)-compound resulted in 13C-containing product, (R)-compound gave the product with 13C no more than natural abundance. Apparently, the enzyme decarboxylated pro-(R) carboxyl group selectively and the reaction proceeds with net inversion of configuration. Thus, the presence of a planar intermediate can be reasonably postulated. Enantioface-differentiating protonation to the intermediate will give the optically active final product (Fig. 12).

2.3.3. Active site

DNA sequence indicated that AMDase contains four cysteine residues located at 101, 148, 171 and 188 from amino terminal (Fig. 9).16 At least one of these four is estimated to play an essential role in the decarboxylation. The most effective way to determine which Cys is responsible to enzyme activity will be site-directed mutagenesis. To determine which amino acid should be introduced in place of active Cys, its role was estimated as illustrated in Fig. 13. One possibility is that

(S)-13C-methy1pheny1malonate

(S)-13C-methy1pheny1malonate

(H)-13C-msthy1pheny1malonate

Figure 12: Enantiotopos-differentiating manner of decarboxylation.

(H)-13C-msthy1pheny1malonate

Figure 12: Enantiotopos-differentiating manner of decarboxylation.

Figure 13: Possible reaction mechanism.

it works as a nucleophile. Attack on the carbonyl carbon will result in the thiol ester, which will stabilize the enolate-type transition state in place of coenzyme A. Enantioface-differentiating protonation followed by hydrolysis will give (R)-a-arylpropionate. On the other hand, SH group can work as an acid. Partial protonation to a-carbon of the substrate will facilitate the C—C bond fission to generate an enolate-type intermediate and accelerate the reaction.

Then, substitution of the sulfur atom of Cys with an oxygen would greatly decrease the rate of reaction, because nucleophilicity, anion-stabilizing effect and proton-donating ability of OH group are far smaller than that of an SH group.

Table 2

Reactivities of wild-type and mutant enzymes

Wild 13.3 365 27.5

C101S 4.3 247 57.6

C148S 11.5 100 8.8

Still, a hydroxyl group is capable of more or less keeping the hydrogen bondings in which the cysteine residue might be incorporated. In addition, steric bulkiness of Cys and Ser is not so different. Accordingly, the conformation of the enzyme will not be seriously affected by the mutation and hence the activity of the enzyme was expected to remain partly, whichever Cys is replaced by Ser, different from the cases in which other totally different amino acid residue is introduced in place of Cys. Thus, four mutant genes in which either one of the four codons of Cys is replaced by that of Ser were prepared and expressed using E. coli. Four mutant enzymes were isolated, purified and incubated with phenylmalonate.20 Kinetic data for the mutant enzymes as well as the wild one are summarized in Table 2. Among four mutants, Cys188Ser showed a drastic decrease in the activity (kcat/Km), which was due to a decrease in kcat rather than affinity to the substrate (Km).

The CD spectrum of the C188S mutant is essentially the same as that of the wild-type enzyme, which reflects that the tertiary structure of this mutant changed little compared to that of the wild-type enzyme. Calculation of the content of secondary structure of the mutant enzyme based on J-600S Secondary Structure Estimation system (JASCO) also showed that there is no significant change in the secondary structure of the mutant. The fact that the kcat value of this mutant is extremely small despite little change in conformation clearly indicates that Cys188 is located in the active site.

Was this article helpful?

0 0

Post a comment