1.5. Diacids as analogues of the intermediates

Whatever is the mechanism of PBGS at some point during the reaction an intermediate is created which essentially is a long chain dicarboxylate. The chain length of this diacid will be either seven or ten carbon atoms long, depending on the mechanism. Studying the inhibition potency of this type of compounds seemed to us an adequate means to obtain more information in the recognition sites on the enzyme surface and hopefully also about the sequence of events occuring at the active site of the enzyme. In view of distinguishing between the two major mechanistic proposals the diacids containing seven carbon atoms (Shemin et Jordan 2) [12,14] or the diacids containing ten carbon atoms are especially interesting (Jordan 1) [46] (Scheme 8). At first we tested pimelic (23) and sebacic acid (29), because they are commercially available.

Scheme 8: Mechanistic rationale for testing diacids containing seven or ten carbon atoms

The next candidates would be compounds where a nitrogen atom has been introduced into an adequate position. Also of interest are analogues where one or may be even two of the carboxylate groups have been exchanged against a nitro group. This sort of variation would permit to detect a differenciation of the recognition sites for the P- and A-side substrates. Varying the chain length should allow to deduce the importance of the exact distance between the two carboxylate ends of the molecules.

Part of the work of R. Lüönd was already directed towards studies of analogues of one of the postulated intermediates [47]. He synthesised the diacid containing seven carbon atoms which was at the same time adequately substituted to mimick more closely the intermediate postulated by Shemin (Scheme 9). The molecule synthesised and tested as an inhibitor was in his racemic form.

Scheme 9: Inhibitor tested as analogues of the intermediate postulated by Shemin

The two possible diastereoisomers for this compound were separated by column chromatography and tested separately. The amino groups had been left out, in order to avoid unwanted side reactions, like pyrrole formation or the formation of the pyrazine. These side reactions would severely hamper the use of these compounds as inhibitors. The methoxy group was a consequence of the synthetic procedure applied for the synthesis. An important factor in favour of the presence of the methoxy substituent is also the fact that the methoxy substitued compounds are more stable. The retroaldol reactivity is considerably lower and lactone formation which is a major problem with the corresponding alcohol can be avoided. Besides these advantages of the methoxy group compared to the alcohol function at this position, there are at least two disadvantages as far as the recognition at the active site is concerned. The polarity is clearly reduced by introducing the methoxy group in the position of the alcohol function. On top of that the methoxy group clearly introduces additional steric hinderance at a sensitive site of the enzyme. The inhibition tests of the two racemic diastereoisomers rac-25 et rac-26 using PBGS isolated from Rhodopseudomonas spheroïdes showed that both inhibitors have a relatively weak interaction with the enzyme rac-26 Ki =11'000 mM, and rac-25 Ki=25'000 mM. This deyievingly high values are counterbalanced by the fact that the enzyme obviously distinguishes between the two diastereoisomers. In order to obtain more significant results the compounds rac-25 and rac-26 have to be resolved, so that individual stereoisomers can be tested. This would be a sensitive test to find out if compounds of this structure are recognised as analogues of the substrate, which are not fitting well into the active site, because they are too highly substituted, or if they are interacting as analogues of the postulated intermediate.

1.6. Inhibition studies of PBGS using diacids reported by other groups

In all the inhibition studies reported so far analogues of postulated intermediates have been very scarcely studied so far. Major contributions are due to the group of P.M. Shoolingin-Jordan, who published recently a study using a series of diacids [23]. He studied in parallel the inhibition efficiency of his compounds in the test with the enzymes from three different sources: Pisum sativum (peas), Escherichia coli and Saccharomyces cerevisiae (yeast). These three enzymes are members of the two categories mentioned: the Mg2+ dependent enzymes are represented by PBGS fromPisum sativum the two other enzymes (fromEscherichia coli and from Saccharomyces cerevisiae) are Zn2+ dependent. Known inhibitors were tested in order to have standards for comparison with earlier results obtained with enzymes from other sources (Scheme 10). The inhibition experiment using succinic acid (16) gave only a very weak inhibition (Ki bigger 20'000 mM) the corresponding 2,2-difluoro compound 9 shows only a moderately better competitive inhibition of the enzyme (Ki = 7'000 mM). Methylated glutaric (10) and adipinic acid (11), (12) were tested as well. They also showed only a moderate inhibition efficiency. However Jordan and his collaborators could observe that increasing the chain length considerably by passing from a diacid with 5 carbon atoms to the diacid with 10 carbon atoms in the chain, as is the case for 2,2,7,7-tetramethyl-sebacic acid (13), changed the mode of inhibition from a competitive inhibitor to a non-competitive inhibitor. The interpretation of this results postulated in their paper is, that the longer diacids exercise their inhibition by a non specific binding to the hydrophobic regions of the enzyme. The competitive inhibition however is attributed to a simultaneous interactions at the site of recognition for the A- and the P-side substrate.

Scheme 10: Substrate tested by the group of Shoolingin-Jordan [23].

In the second publication of the same group another diacid has been tested [35]. The compound 14 (Scheme 11) mimicks the intermediate proposed by Shemin. A time dependent inhibition could be observed and the formation of a stable complex could be detected.

Scheme 11: The analogue 14 of the postulated intermediate, which was tested as inhibitor by the group of Shoolingin-Jordan [35].

The most recent publication of Leeper,[36] is the first one which mentions the synthesis of a compound mimicking the intermediate of Jordan's first mechanistic proposal (Scheme 12). This diacid mimicks perfectly, the excepted intermediate and only one amino group is missing. Testing this compound with PBGS form Baccillus subtilus did not show any appreciable inhibition.

Scheme 12: The analogue of the intermediate postulated by Jordan, which was tested as potential inhibitor [36].

1.7. Choice of diacides tested as inhibitors

Most of the diacids are derived from sebacic and pimelic acid. But compounds with four carbon atoms, which are more easily available, were tested too. In this category , 19 and 20 were tested too and it is interesting to see the influence of the steric effect of an additional aromatic ring. Between the pimelic and the sebacic analogues, the influence of the chain length between the two acid functions was investigated and compounds with eight carbons 27 and with nine carbons 28 were prepared. In the series of diacids containing ten carbons atoms, the influence of the introduction of one or two nitrogen at different places of the chain were investigated.