Magnetic resonance and fluorescence studies on pyruvate hydrogenase complexes and their small molecular weight constituents

The articles presented in this thesis do not describe at first glance one well-defined subject. They are, however, in fact connected by one central theme: the study of large enzyme aggregates by molecular physical methods. Chosen was the pyruvate dehydrogenase complex (PDC) because of its physiological importance and because it is, perhaps therefore, one of the most studied multi-enzyme complexes. Furthermore the presence of a fluorescent group FAD and the easy replaceable Mg <sup><font size="-1">2+</font></SUP>ion make it an accessible complex. Detailed knowledge of its constituent molecules is also of importance to gather as much information as possible. Thus parallel with the label studies on the PDC, detailed studies on its small molecular constituents i.c. thiamine pyrophosphate and isoalloxazine were carried out. Originally only the smallest complex known yet, that isolated from A.vinelandii, was studied, later the methods developed were also applied to the complex from E. coli . Because of the multitude of possibilities and the difficulties of working with biological material, no complete study is presented here. The information in the articles of this study will serve in the future as starting-point for further studies.The aim is to estimate distances between groups of the component enzymes of the complexes and construct from these distances a model about the working (regulation) mechanism of these pyruvate dehydrogenase complexes. Detection and analysis of the conformational changes induced upon the binding of metabolites and nucleotides to PDC are for this purpose of major importance.The observations made in the third article prove that such differences exist and can be detected, thus stimulate further research in this respect. In this article it is proven that the fluorescence properties of the label ANM (N-(1-anilino-naphthyl-4) maleimide) are dependent on the way incorporation is accomplished. Clearly energy transfer between the label and FAD occurs, which indicates that estimates of distances can be made, when a more specific probe is used. From the difference in energy transfer however, it is already apparent that the distance of label attached to the lipoyl moiety, via reduction of the lipoyl S-S bridge with NADH is smaller relative to FAD, than in case it is attached via reductive alkylation of the lipoyl S-S bridge with pyruvate. This illustrates that indeed the proposed aim can be reached in principle. Thus pyruvate as a metabolite and NADH as a nucleotide induce different conformations. Extensive study of such conformers will give insight in the working mechanism of this kind of complex. A conformational change observed with spin label illustrates a similar effect. In this case removal of the phosphate, which has an activating effect (Bresters, T.W., De Kok, A. and Veeger, C. (1975) European Journal of Biochemistry 59, 347- 353) on the overall reaction, induces a structure in which spin-spin interactions occur. The origin of these spin-spin interactions are not completely clear from the theoretical point of view. We have ascribed them for the time being to two kind of interactions, one weak in which the hyperfine coupling a is much larger than the interaction term ( a /J>>1) and one in which the interaction term is strong, leading to a triplet-like spectrum. In order to explain the extremely narrow lines it is assumed that the two spins have a correlated movement and are projected on a common axis-system, which moves much more rapidly itself, than the pyruvate dehydrogenase complex tumbles in solution. In this way the relatively low doublet-splitting between the triplets, the smaller hyperfine splitting of the nitrogen and the narrow linewidth are due to the projection on the new axis-system. This explanation requires at least two different types of bound spin labels, which are in interaction with each other, the magnitude of the interaction being a function of the distance between these two. In this case a relatively slow movement of the lipoyl arm (slow exchange) connects the two environments in which the interactions take place. For this explanation no additional prove exists at this moment, but it is a reasonable explanation in view of the postulated mechanism of Reed (Reed, L.J. (1969) in Current Topics in cellular regulation 1, 233-251) although at least two transacetylase (LTA)-components should be placed between the pyruvate dehydrogenase (PDH) and the lipoamide dehydrogenase in both complexes. Such an explanation is in agreement with the observation that in the A. vinelandii complex two different types of LTA are present. If, however, the temperature dependence of the signals is analysed, it appears that the linewidth of the narrow type (I) is very slowly varying with viscosity (30% glycerol-H _{<font size="-1">2</font>} O, temperature range 40°C, at the maximum viscosity which is about 15 centipoise the linewidth is 1.7 G, not published) indicating that the strong interaction type stays in the narrowing region over a rather long range. In addition only the label incorporated via pyruvate shows in a phosphate buffer- glycerol mixture a relatively narrow species upon extrapolation to 1 centipoise (&lt; 0.3 ns, but anisotropic). There are, however, clearly at least two differently surrounded spin labels present, which are in phosphate buffer not strongly interacting, indicating a rather large conformational change upon removal of phosphate of the PDC from A. vinelandii.Similar studies are in progress concerned with the PDH-site. In this context the first article and the Mn <sup><font size="-1">2+</font></SUP>binding studies of the second article must be mentioned. The first article derives from specific spin-lattice and relaxation line broadening effects in NMR, the conformation of the Mn <sup><font size="-1">2+</font></SUP><strong>.</strong> TPP complex in solution.Because Mn <sup><font size="-1">2+</font></SUP>can replace Mg <sup><font size="-1">2+</font></SUP>both in the overall reaction of the PDC and in the partial PDH-reaction the conformation of this complex will be similar to the Mg <sup><font size="-1">2+</font></SUP><strong>.</strong> TPP complex. If bound to the PDC it is of structural (and thus kinetic) im portance to detect possible conformational changes of this Mn <sup><font size="-1">2+</font></SUP><strong>.</strong> TPP molecule. Because of the large influence paramagnetic Mn <sup><font size="-1">2+</font></SUP>exerts on the relaxation of pro tons it is possible to use a relatively high excess of TPP relative to Mn <sup><font size="-1">2+</font></SUP>and still have a measurable influence on the relaxation. Therefore, with the aid of Fourier-Transform NMR, it is possible to undertake such a study, even with these large complexes (with a concentration of 10 μM PDC, a concentration of about 1-10 mM TPP is needed, which can be measured indeed in a reasonable time). In fact both the association constant for the Mn <sup><font size="-1">2+</font></SUP><strong>.</strong> TPP complex derived in the first article as well as the binding constants of Mn <sup><font size="-1">2+</font></SUP>to the enzyme complex, given in the second article, are already preliminary studies needed for this purpose.Studies in this context have already been performed, as mentioned in the second article. In order to establish in which manner TPP is bound to the complex, water-relaxation studies in the presence of Mn <sup><font size="-1">2+</font></SUP>are for this purpose essential. In principle it is possible that TPP is directly attached to the complex and Mn <sup><font size="-1">2+</font></SUP>is bound to the TPP (enzyme-TPP-Mn = E-S- M complex) or that Mn <sup><font size="-1">2+</font></SUP>is then bound to another site (S-E-M complex), or that TPP is bridged to the complex by Mn <sup><font size="-1">2+</font></SUP>(E-M-S complex). Although it is highly probable that an EMS complex is the real situation, no direct information is available. A water (solvent) relaxation study is one of the few methods which can be used to prove, which ternary complex is present. Apart from the dissociation constants of the binary complexes (Mn.TPP and Mn.PDC,) also the effect they have each on the water relaxation must be determined. Therefore also titrations in spin-echo NMR were carried out (not published). This method determines directly the amplitude of the signals due to the solvent, after a series of pulses. Only one series of such titrations was carried out, because the instrument was at that moment only available at the University of Groningen. The data obtained indicate that the Mn <sup><font size="-1">2+</font></SUP>is bound quite deeply within the complex, because no large influence on the relaxation was observed.When TPP was added a slight decrease in relaxation was observed, indicative for an EMS type of complex. If in addition pyruvate was added a large decrease was observed, which indicates a large difference of binding of the thus formed quarternary complex. All these observations will be studied in more detail on the instrument, now available in Wageningen. An observation of interest which caused us to stop measuring at the University of Groningen has to be mentioned in this context. For this type of Mn <sup><font size="-1">2+</font></SUP>-binding studies it is necessary to work in a buffer, which does not bind Mn <sup><font size="-1">2+</font></SUP>, as phosphate does. Therefore the PDC was dialysed against Tris-HCl buffer (pH 7.0), in Wageningen frozen in liquid nitrogen and then transported to Groningen. The complex showed after this treatment only a residual activity of 1-2 μmol NADH min <sup><font size="-1">-1</font></SUP><strong>.</strong> mg- <sup><font size="-1">-1</font></SUP>instead of the usual 8-10 μmol NADH min <sup><font size="-1">-1</font></SUP><strong>.</strong> mg- <sup><font size="-1">-1</font></SUP>. If the plots, obtained from EPR-measurements, are compared with the one shown in the second article, it appears that in particular the first strong binding site differs. The data of EPR, T _{<font size="-1">1</font>} - and T _{<font size="-1">2</font>} -measurements could only be made graphically identical if instead of one strong binding site only half such a binding site for Mn <sup><font size="-1">2+</font></SUP>relative to FAD was present. This observation sug gests that this strong binding site is connected with a site which is essential for the overall activity and not with the PDH-components. Another point of interest is that the data obtained from EPR, T _{<font size="-1">1</font>} and T _{<font size="-1">2</font>} cover each other only when two types of independent Mn <sup><font size="-1">2+</font></SUP>-binding sites, a strong and a weaker one, are as sumed, thus excluding negative cooperativity of Mn <sup><font size="-1">2+</font></SUP>binding.One additional experiment must be mentioned. It was tried to obtain information about the distance between the PDH and the lipoamide moieties. For this purpose spin-labelled PDC was titrated with Mn <sup><font size="-1">2+</font></SUP>. It is expected according to J.S. Leigh Jr. ((1970) Journal of Chemical Physics 52, 2808-2612) that paramag netic quenching of the spin label occurs when the spins are not too far apart. Quenching started to be observable at a 10-fold excess of Mn <sup><font size="-1">2+</font></SUP>relative to PDC, which is far above the concentration needed to occupy the specific sites (cf. Fig. 4, second article). Furthermore Mg <sup><font size="-1">2+</font></SUP>shows, although at about a 60- fold ex cess relative to PDC, a similar influence on the spin label spectrum. From this it can be concluded that no specific spin quenching is observed and that the distance between Mn <sup><font size="-1">2+</font></SUP>and spin label is thus rather large. Presumably even an extra group between TPP and the lipoyl moiety is needed to explain these data. Experiments with fluorescent and spin label analogues of TPP are now in progress to study these effects in more detail. The fact that Mn <sup><font size="-1">2+</font></SUP>does not quench the FAD fluorescence, but a spin label attached to the lipoyl moiety does quench, in dicates that the distance between the lipoyl moiety of LTA and FAD is much smal ler than the distance between the PDH-active site and the lipoyl moiety.The fluorescence properties of the prosthetic FAD group of lipoamide dehydrogenase is also of interest. In the third article it is shown that its properties vary considerably for the different sources from which the enzyme is isolated. Particularly the observation of non-identical binding of the FAD groups in a dimeric enzyme with identical peptide chains is of interest. Knowledge about the flavin molecule is therefore also of importance. As shown in the fourth and fifth articles a large variety of isoalloxazine (the major part of FAD, determining the fluorescence properties) - analogues have been studied. Especially the cationic species are of interest because they can be visualized as stable analogues of intermediates normally occurring in reactions, catalysed by flavoproteins. The detailed information obtained from our NMR studies thus adds considerably to our understanding of the catalytic properties of the isoalloxazine. It is apparent from articles IV and V, that the theoretical description of the flavin molecule lacks still accuracy to explain the differences observed between flavin molecules as well as the different types of reactions catalysed by flavoproteins. It is clear from this thesis that molecular physical methods can be used on these large enzyme complexes and that valuable information can be obtained. With a lot of patience and carefulness much more will be learned