Mechanistic analysis of enzyme catalysed reactions

The studies of enzyme catalysed reactions have revealed many of the factors responsible for the astonishing rate enhancements and selectivities observed. The availability of high resolution X-ray structures of enzymes and better even of enzyme-substrate or enzyme-inhibitor complexes allowed to track down the structural details responsible for the astonishing properties of enzymes as catalysts. It could be shown that some of the properties of an enzyme can be often imitated by much smaller structures. It is also a challenge for chemists and biochemists to understand and correctly describe the events happening at the active site of an enzyme. Enzymes catalysing a sequence of synthetic steps are interesting candidates for mechanistic studies. On top of the questions which can be asked for all enzyme catalysed processes, there are additional problems to be solved for such multistep reactions: what is the sequence of events at the active site; is the sequence at the active site identical or different from the "chemical" sequence; does the enzyme accelerate all steps or maybe the enzyme acts mainly on one or two determining steps of the sequence. The detailed analysis of complex processes is possible, because the intermediates of such a process or analogues of the proposed intermediates are in principle available via synthesis. Therefore the "chemical" properties of such intermediates can be studied independently and the reactivity can be compared in the absence and in the presence of the enzyme. Finally from the point of view of evolution it is interesting to know how nature dealt with complex synthetic problems. To optimise a multistep sequence is a problem of higher complexity and we will learn more about the tactics and strategies used by nature studying such processes. The biosynthesis of porphobilinogen from 5-aminolevulinic acid is a typical case for such a multistep process. Adding to the usual interest in multistep transformations is the fact that 5-aminolevulinic acid is chemically undergoing a dimeroidization which is different from the one catalysed by the enzyme.

Inhibition studies of prophobilinogen synthase from R. spheroides

As a high resolution X-ray structure of porphobilinogen synthase (PBGS) is not yet available, we have to rely on other, less precise techniques such as inhibition studies. The results of the model studies for the formation of pyrroles from the appropriately substituted aldol products can be used to interpret the postulated mechanism of the biosynthetic pathway. With the knowledge from the model studies a chemically satisfactory interpretation of the postulated mechanism can be proposed. Neither the model studies nor the chemical logic of the interpretation proves the correctness of the postulated mechanism. In order to contribute to the knowledge of the enzymatic transformation we decided to start a systematic inhibition study with substrate analogues, analogues of the product and finally analogues of possible intermediates. We decided to use the enzyme isolated from one source the photosynthetic bacterium R. spheroides first. A major contribution to the knowledge of porphobilinogen synthase stems from Shemin's group. Shemin could show that g-ketoacids were good inhibitors, whereas a-ketoacids revealed no inhibition in his kinetic studies. Most of the other inhibitors known, have not been studied in an effort to understand PBGS, but have been isolated from natural sources. So the most potent inhibitor of PBGS known, succinylacetone, has been identified during a study of hereditary tyrosinaemia, a severe inborn metabolic disorder.

PBGS is catalysing a complex sequence of events. The kinetics of this process does not reflect one simple fundamental reaction step. Under these circumstances it is clear that the inhibition constants Ki are not the dissociation constant of the inhibitor enzyme complex either. As long as structurally similar inhibitors are used the experimentally determined inhibition constants should still reflect the recognition between the enzyme and the inhibitor.

We synthesised and screened substrate analogues where all positions of the 5-aminolevulinic acid had been modified with the exception of the ketofunction at position 4 (see Figures 35-38).

Structure KmKiInhibitor type
690.26 mM 0.02--
700.37 mM 0.0125 mM 3competitive
710.26 mM 0.0118 uM 3competitive

Figure 35

Structure KmKiInhibitor type
720.35 mM 0.03--
730.25 mM 0.02--
740.40 mM 0.10.43 mM 0.13competitive

Figure 36

Structure KmKiInhibitor type
rac-750.30 mM 0.012.2 mM 0.4competitive
rac-760.38 mM 0.025.5 mM 0.9competitive
rac-770.31 mM 0.031.2 mM 0.4competitive
rac-780.38 mM 0.097.3 mM 0.6competitive
rac-790.38 mM 0.0328 mM 2competitive

Figure 37

Structure KmKiInhibitor type
800.36 mM 0.041.0 mM 0.1competitive
390.35 mM 0.02--
810.20 mM--
820.35 mM3.1 mMcompetitive
830.40 mM 0.040.25 mM 0.05competitive
840.36 mM 0.042.7 mM 0.4competitive
850.30 mM 0.0260 uM 15competitive
860.37 mM 0.0119 mM 2competitive
870.28 mM 0.0327 mM 4competitive

Figure 38

The most powerful inhibitor was obtained when the carboxylic acid was replaced by the isosteric nitro group 71 (see Figure 35). Changing the substitution at position 2 and 3 gave moderate to good inhibitors. Most of the inhibitors where the substituent at position 5 had been modified were good inhibitors.

The first intermediate postulated by Shemin seemed to us to be an attractive candidate for inhibition studies. The chemical reactivity of the aldol product postulated by Shemin is such that even without the presence of the enzyme pyrrole formation would occur86. Therefore we had to use analogues of the postulated intermediate which lack the amino group involved in the pyrrole formation. The other amino group in the a-position from the ketone had also to be left out, because a-amino ketones dimerise forming dihydropyrazines 284,313. Therefore analogues were synthesised which lack both amino groups (see Figure 39).

Structure KmKiInhibitor type
rac-880.28 mM 0.0425 mM 2competitive
rac-890.23 mM 0.0111 mM 1competitive

Figure 39

The intermediate postulated by Shemin 163 can exist as four stereoisomers. In order to check if the aldol product is a potential intermediate we needed analogues which are chemically and configurationally stable. To assure that both conditions are fulfilled we decided to study the products rac 88 and rac 89. The two diastereoisomeric pairs of enantiomers rac 88 and rac 89 were studied separately. The inhibition studies showed that one of the pairs of enantiomers rac 88 was a very weak inhibitor, whereas the second pair of enantiomers rac 88 possessed an moderate inhibition constant.

The "mixed pyrrole" revisited

One of the arguments used by Shemin when he postulated his mechanism was the enzyme catalysed formation of the so called "mixed pyrrole". Shemin could at the time not prove the structure ot the "mixed pyrrole". He supposed that the structure of the newly appearing spot on paper chromatography had to be the one he proposed, because of the close analogy with the normal process leading to porphobilinogen.

Our method allowed us to synthesise the proposed "mixed pyrrole". The structure could be secured by spectroscopic methods. Repeating Shemin's experiment we could show that indeed a new spot could be detected if 5-aminolevulinic acid and levulinic acid were incubated together with the enzyme PBGS (see Figure 40). However comparing the chromatographic behaviour of the newly formed product with the synthetic material clearly indicated that the two compound are not identical.

Figure 40 Paper chromatogram of the formation of the mixed pyrrole. Experiment :1 Incubation of PBGS with 5-aminolevulinic acid; Experiment 2: Incubation of PBGS with 5-aminolevulinic acid and levulinic acid; Experiment 3: Incubation of PBGS with 5-aminolevulinic acid and methyl levulinate; Reference: the mixed pyrrole 18 obtained by synthesis

This result is rather surprising, because all the indirect information available was in favour of the proposed structure.

Conclusions

The inhibition studies of substrate analogues allows to draw the following conclusions. The carboxylate function at position 1 is important for the recognition. Additional substituents at the position C2 and C3 are tolerated. The interpretation of these inhibition results is difficult because it is not obvious if the "A" site or the "P" site or even both sites are blocked by the inhibitors. Variations at the position 5 are tolerated. The amino or ammonium group at C5 is not an essential element of recognition. This is in surprising contrast to other enzyme catalysed reactions, where the presence of the charged ammonium group is crucial for the recognition process. Neutral and polar substituents can be used to replace the amino group. Switching to a negatively charged carboxylate 86 or to a sterically demanding group like the monophthalimide 87 resulted in inhibitors with a low affinity. The nitrile derivative 85 is an excellent inhibitor. The protons at the methylene group at C5 are very acidic. In D2O as well as in CD3OH the 1H NMR spectra showed complete deuterium hydrogen exchange. The excellent inhibition by the nitrile might arise from the formation of the enamine into the "wrong" position at the active site of the enzyme (see Figure 41).

Figure 41

The picture of the recognition site of the enzyme porphobilinogen synthase from Rhodopseudobacter speroides has to be adapted to our results (see Figure 42).

Figure 42 Postulated recognition site of porphobilinogen synthase

Besides the carbonyl group at position C4, which is essential for the recognition, there is a high specificity for the planar carboxylate group at C1. Some space is available for additional substituents at position 2 and 3. As long as the second substrate does not interact with the active site thereby blocking the additional space available, this interpretation of the results is reasonable. The recognition of the 5-amino group is not very strong.

The most interesting result is the inhibition studies with the analogues rac 88 and rac 89 of the potential intermediate. Inhibition studies with the racemate of one diastereoisomer rac 88 showed no or only a very weak inhibition (see Figures 42 and 43). The inhibition studies with the racemate of the other diastereoisomer rac 89 gave an inhibition constant of 11 mM.

Figure 43 Analogues of the postulated intermediate

The mechanism first postulated by Shemin and latter adjusted by Jordan can now be modified so as to include the newly accumulated knowledge (see Figure 44).

Figure 44 Postulated mechanism for the enzymatic synthesis of porphobilinogen

The relative configuration corresponds to the relative configuration of the better inhibitor. The absolute configuration at the two chiral centres of the aldol intermediate is chosen to be 3R,4S. With this assumption the 1,2-elimination becomes the stereoelectronically preferred trans elimination and the final 1,4-elimination is syn if the pro R hydrogen is eliminated.

Acknowledgment

It is a particular pleasure to express my thanks to my coworkers for their commitment and their enthusiasm. My thanks are going to Anne Meunier, Hugo Schütz, Rainer Lüönd, Hugo Bertschy, Josef Walker, Vlado Mastihuba, Maurus Marty, Dominique Aeby, André Chaperon, Anne-Laurence Schrumpf, Matthias Henz and Thomas Engeloch. I would like to thank Cécile Pasquier for numerous literature searches. Our research efforts would not have been possible without the collaboration and help of PD Dr H.-P. Köst, Botanisches Institut der Universität München, Professor Dr C. Kratky, Karl-Franzens-Universität Graz, Prof. A.I. Scott and Dr C.A. Roessner, Texas A&M University College Station, Dr E.K. Jaffe, Fox Chase Cancer Center Philadelphia, Prof. H. Stoeckli-Evans, Institut de chimie Neuchâtel, and quite especially Prof. P. Schürmann, Institut de botanique Neuchâtel. I would also like to express my gratitude to the Schweizerischer Nationalfonds zur Förderung der wissenschaftlichen Forschung, the Stipendienfonds der Basler Chemischen Industrie and CIBA-GEIGY AG for support of the work described.