Prion radicals, a marriage between the big and the small
疯牛病毒蛋白自由基－大与小的结合

Chi Ming Yang
(中国科学院化学所 北京 100080)

Abstract When all the experimental evidence from prion research are reconciled, my interpretation of the available data argues that persuasive evidence suggests that putative forms of persistent sequence-specific prion radicals in transmissible spongiform encephalopathies (TSEs) can be responsible for the infectious agent. A mechanism corresponding to the self-replication of scrapie protein mediated by prion radicals is proposed. My analysis argues that prions replicate via a prion radical-mediated chain process and the generation of the prion radicals is associated with reactive oxidative species. All of the unusual nature of prion diseases in mammals can be explained by invoking the prion protein radicals.

Key words Persistent sequence-specific prion radicals, Transmissible spongiform encephalopathy (TSE), Scrapie self-replication, Protein oxidation, Creutzfeldt-Jakob Disease (CJD)

Prion radicals, a marriage between the big and the small

Chi Ming Yang
(San Diego Institute for Life Science & Health and Dept of Chem & Biochem, UCSD,
P. O. Box 12035, La Jolla, CA 92039, USA and Institute of Chemistry, Chinese Academy of Sciences, Beijing 100080)

Introduction
Prion diseases are fatal neurodegenerative disorders in mammalian species. This group of diseases includes scrapie in sheep, Bovine Spongiform Encephalopathy (BSE) in cattle, and Creutzfeldt-Jakob Disease (CJD) in humans ^[1]. Since the pioneering work by Prusiner and co-workers, who characterized the scrapie protein at a molecular level and built a bridge between research on scrapie and modern molecular biology of prions ^[2,3,4], past many years have seen tremendous advancement in prion research ^[1-6]. To date many important features of molecular genetics associated with prion diseases have been elegantly elucidated, including that the prion protein gene Prnp is a requirement for prion disease, elimination of this gene will abolish disease infection ^[4], a piece of strongly supporting evidence for the protein-only hypothesis. However, mainly because of the prion strain diversity ^[7] and lack of the ultimate prove of protein-only theory ^[5,6], it is suspected that something must have been missing ^[5,6,8]. Facing a scenario that bovine spongiform encephalopathy (BSE) may have passed to human ^[9], we urgently need a few new concepts to resolve the prion enigma in order to facilitate drug development.

On one hand, there is no doubt for the importance of prion protein folding problem in prion diseases ^[10]. On the other hand, facing the perplexing prion disease, an immediate question, which is central to mammalian species including humans, can be asked: In addition to those macromolecules involved in the diseases, is there any role for the smallest but the most vital species, oxygen, for mammals on this planet in the pathogenic process? The benign cellular prion protein (PrP^C) which is crucial to the disease is expressed on the cell-surface of neurons and massive experiments have shown that the transmissible agent of scrapie (PrP^Sc) is found most abundantly in brain homogenates of diseased individuals ^[11]. It is known that brain tissue has a remarkably high respiration rate, for instance, the human brain constitutes only about 2% of the adult body mass, but about 20% of its resting oxygen consumption of the whole body, independent of the state of mental activity. Unfortunately, throughout the prion research history, oxygen has very rarely been mentioned ^[12,13]. Prion infectivity is also called “slow virus” infection, during that from 5 up to 20 year incubation period, is there any involvement of oxygen in the pathogenesis? Moreover, it was reported from scrapie research that scrapie infectivity would not be lost even when the scrapie agent was allowed to stay in air for up to two years ^[14]. Now the scrapie protein has been excellently identified and sequenced ^[15], and several informative reviews on prions have recently sparked renewed interest in understanding the intriguing feature of the disease ^[1,5,8,16]. For chemists, it may be skeptical whether a protein molecule containing more than 10 His, more than 10 Tyr, and multiple Trp residues in its sequence, is so stable towards oxygen.

Rational to the prion replicating mechanism
Oxidative damage to proteins
Oxidative damage to proteins has been implicated in the pathogenesis of many neurodegenerative disorders such as Alzheimer’s disease ^[17,18]. The oxidative damage is usually believed to be mediated via rapid attack on proteins by reactive oxidative species such as hydroxyl radicals, which are generated by either Fenton type reactions or other metal ion-mediated reaction ^[19]. All the mammals use haemoglobin to transport oxygen from the lung to the tissue ^[20], which provides one possible pathway to the generation of hydroxyl radicals. In addition, recent biochemical and biophysical study revealed the tight and co-operative binding of the benign protein with copper (Ⅱ) ion ^[21,22], implying its good ability to generate reactive oxygen species, via a number of chemical processes. Oxidative damage to proteins produces protein radicals as primary products, which can be further reactive towards other protein molecules.

Tyrosyl radicals
Tyrosine radical has been believed to play important roles in biological systems such as in ribonucleotide reductase ^[23]. Considering the bond dissociation energies ^[24] of all the covalent bonds formed between a hydrogen atom and other atoms, oxygen, nitrogen, and carbon atoms of different hybridisation in a protein molecule such as prion protein without any free cysteinyl thiol groups, we can summarise their relative thermodynamic bond dissociation energy scales in the following table (Table 1)

As we know that the lower the bond dissociation energy scale is for a covalent bond, the easier the formation of the corresponding radical upon hydrogen atom abstraction by hydroxyl radicals; and the easier the formation of a radical, the more stable the radical ^[24]. Consequently, the phenoxyl radical on tyrosine (or called tyrosyl radical) is the most stable primary protein radical which can be generated from this prion protein molecule. Irvin and co-workers recently reported that radical transfer from one protein molecule to another protein molecule is mediated by protein oxidative damage ^[25]. Besides, recent studies also showed that Trp residue can form peroxyl radical which can then react rapidly with a wide range of proteins to give long-lived ( persistent) secondary radicals on the Tyr residues of the targeted proteins ^[26].

Table 1. Bond dissociation energies (BDE) of covalent bond H-X

Examine the sequence of the human normal prion protein PrP^C, in addition to many His and Trp residues at the N-terminus, more than ten Tyr residues at the C-terminal domain can be found in this molecule. The Tyr residues at the C-terminal domain of human prion protein sequence are illustrated here:

Y¹²⁸ --Y¹⁴⁵ -- Y¹⁴⁹Y¹⁵⁰ -- Y¹⁵⁷-- Y¹⁶²Y¹⁶³--Y¹⁶⁹ - Y²¹⁸ - Y²²⁵Y²²⁶

It can be expected that the protein molecule is very prone to oxidative attack at Tyr residues by reactive oxidative species such as the notorious hydroxyl radicals.

Initiation of prion radicals
Upon hydrogen atom abstraction from the protein molecule by a hydroxyl radical, presumably at the tyrosine residues, a protein free radical can be formed. According to the nature of radical species, it can subsequently damage other protein molecules, accompanied by cross-linking and production of secondary (new) protein free radicals. Recall this prion radical is on a >208-residue glycoprotein molecule with a rigid tertiary structure at the C-terminal core, it is not inconceivable that it may be very difficult to quench it, and its damaging property can be sequence-specific. Based on the unusual properties of the prion scrapie infectivity, which copurify with the scrapie protein PrP^{Sc [1]}, its unusual stability including resistance to UV irradiation which often kill virus containing nucleic acids, its unusual resistance to harsh chemical treatment ^[2], I hereby postulate that putative forms of sequence-specific prion radicals are responsible for the infectivity in the prion particles. In other words, the infectious, transmissible agent is the free radical form of scrapie prion protein. The scrapie radical is physically unable to be separated from the neutral form of prion scrapie and it is sequence-specific in both binding and damaging other molecules. The difference between the PrP^Sc and PrP^Sc radical is apparently one electron or one hydrogen atom. Shown here are the formation of simple forms of prion radicals without detailed discussion of their structures ^[27].

Prion radical-mediated scrapie replication
A free radical chain propagation is characterized by three sequential steps, i.e., radical initiation, radical propagation and radical termination ^[28]. Since there is no evidence for the involvement of any classical virion or viroids in the development of prion diseases after many year competitive research using powerful molecular biology technique, facing mounting evidence supporting that only a protein is involved, I therefore point out that prion disease replicating mechanism displays strikingly high similarity to a free radical-induced polymerization of a molecule through an autocatalytic chain propagation ^[28,29], with the help of partial sequence specificity in protein binding.

Here, using “N-C” to represent the cellular form of prion protein PrP^C, where N represents the N-terminal domain, which contains poly octarepeats, and C is C-terminal domain. Using“N-C ⁱ” to represent protein molecule i, with i = 1 and 2 for species 1 and species 2, respectively. Using “N-D” to represent the scrapie form of the prion protein. Then prion disease propagation can be described to proceed via the following free radical-chain propagation mechanism:

Protein radical initiation
The initial formation of prion protein radical, presumably on Tyr residues at C-terminus, via hydrogen atom abstraction by hydroxyl radicals, in step (b), occurs in the beginning of sporadic prion disease. The feasibility for the formation of this protein radical is governed by both the PrP protein sequence encoded by the Prnp gene and oxidative stress in the host.

Disease propagation
The protein radical then undergoes an irreversible radical-induced structural transformation at C-terminal domain to form scrapie protein, which is thermodynamically more stable, step (c), whereby intramolecular Tyr-Tyr and/or Tyr-His cross-linking may or may not be involved,

The resulting scrapie protein “N-D^1· ” is still a radical, therefore it can react with a second benign protein molecule in step (d), and carries the chain.

Step (d) is an inter-molecular sequence-specific radical-transfer process, which is dependent of sequence compatibility between the two protein molecules，“N-D¹”and “N-C²”, and it is also the rate-limiting step in the whole pathogenic process as soon as the first scrapie prion radical is generated. Therefore this stage is also the infectious stage if protein molecules N-C¹ and N-C² are of different sequences from two different species, with the protein radical, “N-D¹·”^, being the infectious agent. The resulting scrapie protein， “N-D¹”，becomes non-infectious scrapie after step (d), since it has transferred the radical to the second molecule，“N-C²” and the newly formed radical, “N-D²·”, becomes the infectious particle.

Besides being able to transfer the radical to a second molecule, the radical form of scrapie prion，“N-D^1· ”， may also undergo intermolecular radical addition-reaction to form radical-mediated cross-linking adducts with a second protein molecule, (see Figure 1) presumably on the aromatic rings of Tyr and/or His residues at the C-terminal domain in another prion protein molecule. The subsequently formed dimerized product is again in a radical form in scrapie deposits, shown in step (c)’,

Since the resulting scrapie protein adduct from step (c)’is again a radical, and it can react with still another benign protein molecule, step (d)’,

which is followed by scrapie formation of the second species in step (e). Thus, in the scrapie plaque, only those protein deposits which are in the radical form, are infectious in initiating further damaging processes to other benign protein molecules; while most deposits which are in non-radical forms, are not infectious. Theoretically, the Gibbs free energy change △G associated with each step in the chain mechanism is < 0, except step (d) or (d)’, which is the rate-determining step, with its △G value being dependent of sequence compatibility between the two protein molecules of different sequences from species 1 and species 2, indicating the species barrier and possibility of crossing the barrier. Since there are several pathways to generating scrapie prions, as shown in the above reactions, and the resulting prion radicals can further damage each other depending on the time of incubation and the location of the radicals within the protein molecules, recall that multiple Tyr and His residues are available in the molecule. All may explain the PrP^Sc strain diversity.

Therefore, the diseased-form protein PrP^Sc is the structurally transformed protein, here represented by，“N-D” and polymer，“(N-D¹)( N-C¹)” , formed via a free radical-chain process, with the free radicals， “N-D·”， and polymer radicals，“(N-D)(N-C)· ”, which can be of several protein structural conformations, being the infectious particles occluded in the solid deposits. The radical species retains partial sequence-specificity in attacking other molecules. We can also understand from here, as long as a scrapie protein radical, “N-D¹·” , is generated in the initial stage, the radical carries a very long chain, through sequential free radical transfer reactions to an unlimited number of other labile benign prion protein molecules.

Moreover, it is this radical species, “N-D·”, or polymer radicals，“(N-D)(N-C)·”, not only responsible for the persistently transmissible infectivity, but also being unable to be separated from the neutral form of scrapie，“ N-D” or polymer，“(N-D)(N-C)”, by virtue of the fact they have the closely identical molecular weight (±1 of molecular mass unit) as well as the same shape.

Termination
Theoretically, there is a termination step at the end of this free radical-chain propagation, which can be described as follows:

For very small radicals such as methyl radical CH₃ ·at a high concentration in solution phase, the rate of this coupling reaction is as fast as diffusion-controlled, with a bimolecular reaction constant of k ～10¹⁰ s ^-1 M ^-1. Very unfortunately, given the low concentration of the prion radicals, its solid state limiting its diffusion ability, and its highly steric hindrance with sequence specific properties of a 208-residues scrapie protein, (the radical is not really free) these three factors render the chance for two prion radicals come close to be within a distance of one to two Aº (10^-10 meter), in order to be able to form a covalent bond, is almost zero, leading to a scenario

Figure 1. Simple forms of possible structures of protein radical species in the scrapie prion

that radical-chain propagation going on until all the benign prion protein is consumed and structurally converted. The remaining radicals will still stay active, as it is shown in Table 1 that tyrosyl radical is one of the most stable radicals in biological fluids, no good hydrogen donor is available to quench it. This is demonstrated by its unusual resistance to long-time harsh chemical treatment. The radical nature of prion can also account for its inactivity towards the immune systems of mammals, since the infectious part of the agent is not a particular form of a protein but very tiny in size for it is an unpaired electron.

This interpretation using prion radicals explains species barrier, strain diversity, mutation effect of the diseases and the possibility of crossing species-barrier ^[1], because different species have both different PrP sequence including polymorphism and different glycoform of the cellular proteins, which affect the binding before a prion radical damages the second molecule. Protein radicals can also be powerful enough to damage other proteins regardless of poor binding between the two protein molecules of different species, thereby leading to the possibility of crossing the species barrier. The persistent and partially damaged prion radicals render itself both protease-resistant in nature and infectious pathogen-like properties.

The above suggested mechanism not only explains the prion disease fully at a molecular level, but also demonstrates that the prion protein is both necessary and sufficient (for it is infectious) to the disease for mammals which require oxygen. Meanwhile, the radical-mediated mechanism indicates that the susceptibility to scrapie is a function of PrP^C level in the host. Additionally, the radical mechanism is consistent with the unimolecular mechanistic model as well as the unknown protein X, proposed by Prusiner, based on their extensive molecular genetics study ^[30]. Moreover, the sequence-specific prion radical of highly damaging nature allow it to be capable of function like a “virino” species, leading to the widely observed strains diversity in the prion disease.

The poly octarepeats and the hydrophobic core
Prion protein (PrP) gene Prnp is found in all the mammals so far examined, but its normal physiological function is still unclear. The benign protein PrP^C is attached to brain cell surface membranes by its glycosylphosphatidylinositol anchor, and both the PrP^C and PrP^Sc were found to be encoded within a single exon of the chromosomal gene as proteins of 254 amino acids ^[31]. Both the five copies of the poly octapeptide repeats, (PQGGGGWGQ)(PHGGGWGQ)₄, at N-terminus and two hydrophobic segments, (A¹¹³GAAAAGAVVGGLGG Y) and (L²³⁴FSSPPVILLISFLIFLIVG) [human numbering]，at C-terminal domain are highly conserved among most mammal species such as human, cow, mouse and hamster ^[32].

Studies revealed that the N-terminal poly octarepeats in PrP^Sc is not required for either transmission or propagation of infectivity ^[30], implying that the infectious radical is located at the C-terminal domain, see Figure 2. It is reported from molecular genetic study that mutated goats having only three octapeptide repeats are non-pathogenic after challenge with scrapie ^[33]. In contrast, normal goats with five copies of octarepeats are very susceptible to scrapie. While familial (genetic) human CJD patients have from 5 up to 8 extra octarepeats as [PHGGGWGQ] in PrP, compared with five octarepeats at the N-terminus encoded in human prion protein gene for normal humans ^[34]. Nevertheless, transgenic mice ablating 3.5 of the five copies of octarepeats in PrP gene was still shown to be pathogenic towards PrP^Sc inoculation ^[4]. All these simply indicates that polyoctarepeats at the N-terminus of benign PrP may play a role in the aetiology of the diseases, or may be evolved to fight back against oxidative damage and the disease. Further investigation is required to clearify the role of the octarepeats in PrP protein.

Although it is unknown why the two hydrophobic segments, (A¹¹³GAAAAGAVVGGLGG Y) and (L²³⁴FSSPPVILLISFLIFLIVG), in PrP is so conserved in mammals, it can be speculated from radical-reaction mechanistic point of view that this hydrophobic cores may unfortunately provide a unique hydrophobic environment through aggregation in facilitating free radical-mediated cross-linking as soon as protein radicals are formed in the surrounding.

Figure 2. Prion protein components

Protein folding, the big and the small
It is required to be emphasised that protein folding undoubtedly play an important role in the pathogenesis, in particular in the generation of the protein radicals and its sequence specificity which determines both species barrier and possibility of crossing the barrier in the disease. Currently, while the three dimensional structure of the benign prion protein PrP^C has been revealed by NMR spectroscopy ^[35], but the non-crystalline, nonhomogeneous scrapie prion protein PrP^Sc obstructs its structure from being solved at a high-resolution level either by NMR or X-ray crystallography ^[30], which significantly blocks the effort towards a clear mechanistic elucidation through protein folding investigation.

By invoking the sequence specific prion radical which is generated from interaction of prion protein with reactive oxidative species, however, many of the controversial issues associated with the prion enigma can be resolved. Given biodiversity in nature evolution whereby proteins are employed to play a myriad role in biological processes, it is entirely conceivable for a chemically mutated protein through posttranslational processes involving reactive oxidative species, to protein radicals which are long-lived, transmissible, and sequence-specific damaging species.

Conclusion
Understanding the molecular mechanism has been the major focus of intensive research on prion diseases. After many year of investigation on prions, many lines of evidence mostly from molecular genetic study are strongly supporting the protein-only hypothesis. It is required to reconcile with the protein-only hypothesis together with a possibility that if it is likely for a modified protein behave as a new type of “virino” which modulates the structural transformation between PrP^C and scrapie PrP^Sc. My analysis argue persuasively, based on a wealth of knowledge and a body of evidence from prion research, that the feature of prion diseases can be very well described by a radical chain-propagation mechanism, exhibiting partial sequence specificity in transforming the benign prion protein to the scarpie protein. The relative ease in generating these persistent, damaging and highly sequence-specific protein radical can be used to interpret all the three forms of prion diseases, sporadic, transmissible, and familial pattern, which are manifested by the three pathways to acquire the damaging radicals, in the pathogenesis of prion diseases.

Research on protein radicals is still in its infancy ^[25,26], sequence specific protein radical is a new concept and it has never been reported before. However, sequence specific organic free radicals have been implicated in the function mechanism of antibiotics anti-cancer agents such as Calicheamicin and Bleomycin, which are among the naturally available chemical nucleases cleaving nucleic acids via a free radical mechanism ^[36].

The major arguments for prion protein radicals are summarised in the following:

1. Brain cells consume very high amount of oxygen, implying that reactive oxidative species are available;
2. Elimination of the PrP gene makes experimental mice totally resistant to scrapie infection, but cerebellar cells lacking PrP^C are more sensitive to oxidative stress and undergo cell death more readily than wild-type cells ^[13], indicating that oxidative damage to PrP is able to play an important role;
3. The PrP protein molecule is very prone to attack by reactive oxidative species;
4. Protein tyrosyl radicals can be very stable and persistent in the solid aggregate and transmissible;
5. Prion radicals will not be able to activate the immune systems in mammals since the infectivity is in a radical form without a defined shape;
6. A protein radical can be very damaging sequence-specifically to labile molecules of antioxidant features such as PrP;
7. Prion radicals are unable to be separated from neutral scrapie prions ^[1,4,6,8];
8. Prion radical-mediated chain reaction can be autocatalytic ^[1,4,6,8];
9. A protein radical can be quenched by antioxidant phenol and stabilized by electron-rich molecules ^[1,4,6,8].

While the detailed mechanism for the generation of these types of protein radicals and their structures remain to be determined with the help of free radical chemistry and protein folding study, nevertheless, it is clear this one-electron difference between prion scrapie protein and its free radical form has generated much debate for many decades. Future studies based on this hypothesis may allow us to determine the most effective approach to stop prion diseases, and drug design may be directed towards a combination of antioxidants and sequence-specific protein radical scavengers.

Acknowledgement
The author gratefully acknowledges the support of K. C. Wong Education foundation, Hong Kong. The author would also specially thank Dr. Stanley B. Prusiner in UCSF and his colleagues in UCSD for quickly became interested in the idea of protein oxidation and prion disease during the period of this menuscript preparation in early April 1999.

REFERENCES
[1] Prusiner S B, Proc. Natl. Acad. Sci. U.S.A., 1998, 95:13363
[2] Prusiner S B, Science, 1982, 216: 1364.
[3] Prusiner S B, Scott M R, DeArmond S J and Cohen F E, Cell, 1998, 93:337
[4] Weissmann C, FEBS Lett., 1996, 389: 3
[5] Aguzzi A and Weissmann C, Nature, 1997, 389:795
[6] Caughey B and Chesebro B, Trends Cell Biol., 1997, 7:56
[7] Farquhar C F, Somerville R A and Bruce M E, Nature, 1998, 391:345
[8] Chesebro B, Science, 1998, 279:42
[9] Hill A F et al., Nature, 1997, 389:448; Raymond G J et al., Nature, 1997, 388:285
[10] Harrison P M, Bamborough P, Daggett V, Prusiner S B and Cohen F E, Curr. Opin. Struct. Biol., 1997, 7:53; Jackson G S et al., Science, 1999, 283:1935
[11] Baldwin M A, Cohen F E and Prusiner S B, J. Biol. Chem., 1995, 270:19197
[12] Brown D R, Schmidt B and Kretzschmar H A, Nature, 1997, 380:345
[13] Brown D R, Schaeffer W J, Schmidt B and Kretzschmar H A, Experimental Neurology 1997, 146:104
[14] Pattison I H, Vet. Rec., 1988,123:661
[15] Stahl N et al., Biochemistry, 1993, 32:1991
[16] Edenhofer F, Weiss S, Winnacker E-L and Famulok M, Angew. Chem. Int. Ed. Engl., 1997, 36:1674
[17] Markesbery W R, Free Radical Biol. Med., 1997, 23:134
[18] Hensley K N et al., J. Neurochem., 1995, 65:2146
[19] Stadtman E R, Annu. Rev. Biochem., 1993, 62:797
[20] Wittenberg J B and Wittenberg B A, Annu. Rev. Biophys. Biophys. Chem., 1990, 19:217
[21] Viles J H, et al., Proc. Natl. Acad. Sci. U.S.A. 1999, 96:2042
[22] Stockel J, Safar J, Wallace A C, Cohen F E and Prusiner S B, Biochemistry, 1998, 37:7185; Brown D R, et al., Nature, 1997, 390:684
[23] Stubbe J, Proc. Natl. Acad. Sci. U.S.A., 1998, 95:2723
[24] Lowry T H & Richardson K S, Mechanism and Theory in Organic Chemistry (Harpe & Row, ed. 3, 1985), pp. 161-5.
[25] Irvin J A, Ostdal H and Davis M J, Archives Biochem & Biophysics, 1999, 362:94
[26] Ostdal H, Andersen H J and Davis M J, Archives Biochem & Biophysics, 1999, 362:105
[27] Although scrapie prions may contain covalent cross-linkages between Tyr and Tyr or between Tyr and His residues in the molecular deposits, these diaryl ether linkages (Figure 1) are undetectable by Edman sequencing because of a concentrated trifluoroacetic acid treatment in the Edman degradation procedure which could cleave the diaryl ether cross-linkage through acid catalysed hydrolysis. (See Voet D and Voet J G, Biochemistry, John Wiley & Son, Inc., ed. 2, 1995, pp. 108-110.)
[28] Walling C, Free Radicals in Solution (Wiley, New York, 1957).
[29] Huyser E S, Free Radical Chain Reactions (Wiley, New York, 1970).
[30] Cohen F E and Prusiner S B, Annu. Rev. Biochem., 1998, 67:793
[31] Caughey B, Br. Med. Bull. 1993, 49:860
[32] Marcotte E M and Eisenberg D, Biochemistry, 1999, 38:667
[33] Goldmann W, Chong A, Foster J, Hope J and Hunter N, J. General Virology, 1998, 79:3173
[34] Goldfarb L G et al., Proc. Natl. Acad. Sci. U.S.A. 1991, 88:10926
[35] Riek R, Hornemann S, Wider G, Glockshuber R and Wuthrich K, FEBS Lett., 1997, 413:282; Glockshuber R et al., TIBS, 1997, 22:241
[36] Nicolaou K C et al, Angew. Chem. Int. Ed. Engl., 1993, 32:1377; Stubbe J, Acc. Chem. Res., 1996, 29:322
[37] Dedicated to Prof. Jiang Xi-Kui on the occasion of his 75. In particular we thank Prof. Hu Guo-Fu in Harvard Medical School for proof reading of the manuscript and Dr. Taulane M. Goodmann for technical assistance in molecular modelling.

Chi Ming Yang, Ph.D., research associate professor.
email:   Chiming_Yang@usa.net or ccyang@gpu.srv.ualberta.ca
Received July 20,1999, revised August 15, 1999.