Summation of the International Conference on Radiation and Health: radiobiology viewpoint.

The idea to hold an international conference on radiation and health arose as a natural response to the mounting evidence of serious health effects on populations resulting from the Chernobyl accident. The nature of the Chernobyl accident, where large amounts of radioactive nuclides spread in the reactor area and all over Europe, dictated the emphasis of the presented papers covering both direct radiation effects and effects of radioactive contaminants on large populations. In addition to classical radiobiologic studies of the effects of radiation on cells, organs, humans and their environment, basic studies of DNA damage and repair, the possibility to enhance and repair, and studies of radioprotectors, emphasis was directed at the effects of radiation on large populations and at the inaccuracy of dosimetry. The lack of accurate physical dosimetry resulted in the need to develop new methods of biological dosimetry. Health effects as a result of Chernobyl were expressed in many disorders, some not recognized before as radiation effects, and many related to psychosocial behavioral symptoms. Special emphasis was given to the plight of liquidators who were exposed while carrying out cleanup operations around the reactor and to the observed increase in thyroid cancer among area inhabitants. All of the observations and results were taken into consideration when discussing and proposing appropriate health policies based on risk analysis and the observed phenomena.

The accepted mechanism of action of thiamine in its enzymes was proposed by Breslow [3] and involves the addition of pyruvic acid to the C 2 atom of thiazolium, following its deprotonation and the ylide formation (see Scheme 2).
Intermediate (A) is called an "active aldehyde" and it presents an O − −S + electrostatic interaction, contributing to the internal neutralization of the charge of thiazolium. Such "active aldehyde" intermediates can be isolated [1].
Despite the general acceptance of Breslow mechanism, there remain the following questions unanswered in it: (1) the role of bivalent metal ions, both in vivo (Mg 2+ ) or in vitro (Mg 2+ , Co 2+ , Zn 2+ , Ni 2+ , Cd 2+ , etc.); (2) the role of various parts and the conformation taken by TPP during the various enzymic steps.
We have attempted to provide answers to these questions and proposed a detailed mechanism of action by studying the interactions of bivalent metals, with thiamine model compounds, in recent years [1,4]. We have subsequently tried to immobilize thiamine and derivatives into silica and look upon the catalytic properties of the composite materials produced.
In this paper, we are briefly reviewing all these efforts and conclusions drawn and proposing future applications.

INTERACTION OF THIAMINE WITH BIVALENT METAL IONS
To investigate the role of bivalent metals in the enzymatic action of thiamine, it is important to elucidate the way the metals are bound to thiamine derivatives, such as coordina-tion sites. Towards this goal, early research was directed in the preparation and elucidation of structures of bivalent metals with thiamine and derivatives. However, all these failed to give an answer to this question, due to the net positive charge on thiazolium ring and the easy protonation of N 1 atom of pyrimidine (pKa ∼ = 5) resulting in double charged ; X is Cl − , Br − ; Th is thiamine). All these contained no direct metal-ligand bonding. An example is the Hg 2+ complex [5] of the formula in Scheme 3.
The earlier reported Pt 2+ and Pd 2+ thiamine derivatives (Scheme 4), however, were for the first time proposed to have a zwitterionic structure, with the metals bound at the N 1 site of pyrimidine [6]. This demonstrated the importance of this position as potential coordination site of metal ions. It indicated also that the net positive charge on N 3 of thiazolium was the real reason for the difficulty to form complexes with direct metal-ligand bonds. The proposed structures were later confirmed by X-ray crystal structure determinations for these and other bivalent metals [1].
In the "active aldehyde" intermediates of thiamine catalysis (Scheme 2), the net positive charge on N 3 of thiazolium is partially internally neutralized, due to the S + −O − electrostatic interaction, resulting by a partial positive charge migration to the S 1 position of the ring. We therefore thought that if "active aldehyde" derivatives of thiamine were used, instead of thiamine itself, the problem of complex formation with direct metal-ligand bonds might be overcome. This was proven to be true, using the "active aldehydes" 2-(α-hydroxybenzyl)thiamine (HBT) and 2-(αhydroxycyclohexylmethyl)thiamine (HCMT) and the metals Zn 2+ , Cd 2+ , Hg 2+ , Co 2+ , and Ni 2+ [7,8]. Working at pH ∼ 5.3, all complexes obtained corresponded to the zwitterionic formulae MLCl 3 and presented an M-N 1 direct interaction [7,8]. Important was also the fact that all ligands either free or complexed were adopting the S conformation, and that the internal S + −O − interaction continued to be the same in the complexes, like the ligands.
These findings led to the following conclusions: (1) the easy formation of complexes with direct metalligand bonds of bivalent metals with thiamine "active aldehyde" derivatives, than thiamine itself, strongly indicates that the metal interaction in the enzymatic action should follow the formation of the "active aldehyde" intermediates.
(2) The S conformation may be important in the enzymatic cycle, since it is retained after the formation of the direct metal-ligand bonds.

INTERACTION OF THIAMINE MONO-AND PYROPHOSPHATE "ACTIVE ALDEHYDE" DERIVATIVES WITH BIVALENT METAL IONS
Since the phosphate esters of thiamine are the factors for its enzymatic action, a better conclusion on the role of metal ions cannot be drawn without a detailed study of the interaction of bivalent metals with thiamine phosphate esters. With this aim we continued our efforts starting with the interaction of "active aldehyde" derivatives of thiamine monophosphate (HBTMP) with bivalent metal ions, followed by "active aldehyde" derivatives of thiamine pyrophosphate (HBTPP and HETPP). Depending on pH, three types of complexes were obtained from the interaction of HBTMP with Zn 2+ , Cd 2+ , and Hg 2+ in aqueous solutions. There exist at pH ∼ 1 the double salts of formula [MCl 4 ] 2− [LH] 2 2+ (L is HBTMP; and M are the named metals), at pH ∼ 3.5 the M(LH)Cl 3 of a zwitterionic formula, with the metals bonded through the phosphate moiety, and at pH ∼ 6 complexes of formula MLCl 2 [9], with the metals simultaneously bonded through N 1 and pyrophosphate oxygens [9]. Hg 2+ produced the complex HgL 2 Cl 2 , at pH ∼ 6 with the metal bound at the N 1 site of the pyrimidine moiety only. The crystal structure of the free ligand and the HgL 2 Cl 2 complex showed an S conformation for the ligand [9]. 1 H NMR ROESY spectra confirmed the existence of the S conformation for the free ligand in solution, in all cases, demonstrated by a cross peak between the C 4 −CH 3 group of thiazolium and the C 6 −H of pyrimidine [10].
Using HETPP and HBTPP as ligands (L), we also prepared complexes of formulae {K 2 Cl 2 ], and Zn(LH) 2 -Cl 2 . Both sites of pyrimidine and pyrophosphate oxygens are once more shown to be potential coordination sites [10,11]. The X-ray crystal structure determination of the ligand has again shown the S conformation for it [10]. The same conformation was adopted by the ligand in all isolated complexes, as the 1 H NMR ROESY spectra revealed [10,11].
These studies led to the following conclusions: (1) both the N 1 site and the phosphate group of thiamine phosphates may be important metal binding sites; (2) the importance of the S conformation in the mechanism of the enzymatic action of thiamine is emphasized; (3) the nature of the metal is important in determining the coordination site. For example, in contrast to the lighter Zn 2+ , Cd 2+ metals, the heavier Hg 2+ reacts only with N 1 site of thiamine monophosphate, even at pH ∼ 6.

THE CONFORMATION THAT THIAMINE MAY UNDERTAKE DURING THE ENZYMATIC ACTION
Using the model peptide Asp· Asp·Asn·Lys·Ileu mimicking the protein environment, surrounding the pyrophosphate moiety of TPP, we studied the ternary system [M 2+ ]- [peptide]-[HETPP] of Zn 2+ , Cd 2+ , and Cu 2+ [12,13]. The pyrophosphate oxygen atoms and the N 1 site of pyrimidine were once more proven to be the metal coordination sites and thiamine once more was found in the S conformation. It should be noted here that the base approaching the C 2 −H atom of thiazole was earlier proposed to be the 4 -NH 2 group of pyrimidine, thereby forming the ylide, by attracting its proton and initiating the reaction [14]. Such a role of the 4 -NH 2 group requires a V conformation for thiamine, found only rarely in its free derivatives. However, this conformation was constantly found in the crystal structure of several thiamine examples, obviously imposed and stabilized by the enzymic environment [15,16]. Metals like Mg 2+ and Ca 2+ were linking thiamine pyrophosphate with the apoenzyme through the pyrophosphate group [15,16].
These results showed that both the V and S conformations are important and possibly succeed each other in the mechanism of action of thiamine enzymes [4,17], as shown in Scheme 5. The V conformation initiates the reaction by attracting a proton from C 2 and creating an ylide, and the S conformation is adopted after the formation of the "active aldehyde" intermediates, when a metal may also coordinate with N 1 , besides the pyrophosphate group. The S conformation favors an S + −O(2a) − electrostatic interaction, which facilitates the release of a proton and aldehyde and regenerates the ylide.

BIOMIMETIC CATALYSIS
The tethering of thiamine pyrophosphate and its derivatives on silica via the pyrophosphate group is expected to produce novel composite biomimetic materials, where silica would replace the protein environment of the natural enzyme. These materials would also be expected to catalyze thiamine reactions in vitro. In fact, TPP and its derivatives were immobilized on silica, [18,19] according to the reaction in Scheme 6.
In the same way, the "active aldehyde" intermediates HBTPP and HETPP were also immobilized on silica and their catalytic properties were evaluated and compared with the ones of TPP itself.
The results on benzoyl-formate decarboxylation are summarized in Table 1 and compared with the homogeneous systems. It is thus obvious that the immobilized thiamine pyrophosphate derivatives are far more efficient catalysts than the corresponding homogeneous systems, reducing the reaction time from 330 minutes of the homogeneous system to less than 5 minutes only, for 100% yield in both reactions (1), (3) and (2), (4). Detailed mechanisms of action have been proposed for both reactions as shown in Scheme 7.
These very effective and promising heterogeneous biocatalysts may be evaluated in future studies for the enantioselective production of optically active 2-hydroxy-ketones.