Chapter 1.3. Macromolecular crystallography and medicine

In the last hundred years, crystallography has contributed immensely to the expansion of our understanding of the atomic structure of matter as it extends into the three spatial dimensions in which we describe the world around us. At the beginning of this century, the first atomic arrangements in salts, minerals and low-molecular-weight organic and metallo-organic compounds were unravelled. Then, initially one by one, but presently as an avalanche, the molecules of life were revealed in full glory at the atomic level with often astonishing accuracy, beginning in the 1950s when fibre diffraction first helped to resolve the structure of DNA, later the structures of polysaccharides, fibrous proteins, muscle and filamentous viruses. Subsequently, single-crystal methods became predominant and structures solved in the 1960s included myoglobin, haemoglobin and lysozyme, all of which were heroic achievements by teams of scientists, often building their own X-ray instruments, pioneering computational methods, and improving protein purification and crystallization procedures. Quite soon thereafter, in 1978, the three-dimensional structures of the first viruses were determined at atomic resolution. Less than ten years later, the mechanisms and structures of membrane proteins started to be unravelled. Presently, somewhere between five and ten structures of proteins are solved each day, about 85% by crystallographic procedures and about 15% by NMR methods. It is quite possible that within a decade the Protein Data Bank (PDB; Bernstein et al., 1977) will receive a new coordinate set for a protein, RNA or DNA crystal structure every half hour. The resolution of protein crystal structures is improving dramatically and the size of the structures tackled is sometimes enormous: a virus with over a thousand subunits has been solved at atomic resolution (Grimes et al., 1995) and the structure of the ribosome is on its way (Ban et al., 1999; Cate et al., 1999; Clemons et al., 1999).

Macromolecular crystallography, discussed here in terms of its impact on medicine, is clearly making immense strides owing to a synergism of progress in many scientific disciplines including:

Numerous aspects of these developments are treated in great detail in this volume of International Tables.

Knowledge of accurate atomic structures of small molecules, such as vitamin B₁₂, steroids, folates and many others, has assisted medicinal chemists in their endeavours to modify many of these molecules for the combat of disease. The early protein crystallographers were well aware of the potential medical implication of the proteins they studied. Examples are the studies of the oxygen-carrying haemoglobin, the messenger insulin, the defending antibodies and the bacterial-cell-wall-lysing lysozyme. Yet, even by the mid-1980s, there were very few crystallographic projects which had the explicit goal of arriving at pharmaceutically active compounds (Hol, 1986). Since then, however, we have witnessed an incredible increase in the number of projects in this area with essentially every major pharmaceutical company having a protein crystallography unit, while in academia and research institutions the potential usefulness of a protein structure is often combined with the novelty of the system under investigation. In one case, the HIV protease, it might well be that, worldwide, the structure has been solved over one thousand times – in complex with hundreds of different inhibitory compounds (Vondrasek et al., 1997).

Impressive as these achievements are, this seems to be only the beginning of medicinal macromolecular crystallography. The completion of the human genome project will provide an irresistible impetus for `human structural genomics ': the determination, as rapidly and systematically as possible, of as many human protein structures as possible. The genome sequences of most major infectious agents will be completed five years hence, if not sooner. This is likely to be followed up by `selected pathogen structural genomics', which will provide a wealth of pathogen protein structures for the design of new pharmaceuticals and probably also for vaccines.

This overview, written in late 1999, aims to convey some feel of the current explosion of `crystallography in medicine'. Ten, perhaps even five, years ago it might have been feasible to make an almost comprehensive list of all protein structures of potentially direct medical relevance. Today, this is virtually impossible. Here we mention only selected examples in the text with apologies to the crystallographers whose projects should also have been mentioned, and to the NMR spectroscopists and electron microscopists whose work falls outside the scope of this review. Tables 1.3.3.1 and 1.3.4.1 to 1.3.4.5 provide more information, yet do not claim to cover comprehensively this exploding field. Also, not all of the structures listed were determined with medical applications in mind, though they might be exploited for drug design one day. These tables show at the same time tremendous achievements as well as great gaps in our structural knowledge of proteins from humans and human pathogens.

Presently, an immense number of genetic diseases have been characterized at the genetic level and archived in OMIM [On-line Mendelian Inheritance in Man. Center for Medical Genetics, Johns Hopkins University (Baltimore, MD) and the National Center for Biotechnology Information, National Library of Medicine (Bethesda, MD), 1999. URL: http://www.ncbi.nlm.nih.gov/omim/ ], with many more discoveries to occur in the next decades. Biomolecular crystallography has been very successful in explaining the cause of numerous genetic diseases at the atomic level. The stories of sickle cell anaemia, thalassemias and other deficiencies of haemoglobin set the stage (Dickerson & Geis, 1983), followed by numerous other examples (Table 1.3.3.1). Given the frequent occurrence of mutations in humans, it is likely that for virtually every structure of a human protein, a number of genetic diseases can be rationalized at the atomic level. Two investigations from the authors' laboratory may serve as examples:

Impressive as the insights obtained into the causes of diseases like these might be, there is almost a sense of tragedy associated with this detailed understanding of a serious, sometimes fatal, afflictions at the atomic and three-dimensional level: there is often so little one can do with this knowledge. There are at least two, very different, reasons for this. The first reason is that turning a malfunctioning protein or nucleic acid into one that functions properly is notoriously difficult. Treatment would generally require the oral use of small molecules that somehow counteract the effect of the mutation, i.e. the administration of the small molecule has to result in a functional complex of the drug with the mutant protein. This is in almost all cases far more difficult than finding compounds that block the activity of a protein or nucleic acid – which is the way in which most current drugs function. The second reason for the paucity of drugs for treating genetic diseases is very different in nature: the number of patients suffering from a particular mutation responsible for a genetic disease is very small in most cases. This means that market forces do not encourage funding the expensive steps of testing the toxicity and efficacy of potentially pharmaceutically active compounds. One of several exceptions is sickle cell anaemia, where significant efforts have been made to arrive at pharmaceutically active agents (Rolan et al., 1993). In this case the mutation Glu6βVal leads to deoxyhaemoglobin polymerization via the hydrophobic valine. In spite of several ingenious approaches based on the allosteric properties of haemoglobin (Wireko & Abraham, 1991), no successful compound seems to be on the horizon yet for the treatment of sickle cell anaemia.

More recently, the spectacular molecular mechanisms underlying genetic serpin deficiency diseases have been elucidated. A typical example is α1-antitrypsin deficiency, which leads to cirrhosis and emphysema. Normal α1-antitrypsin, a serine protease inhibitor, exposes a peptide loop as a substrate for the cognate proteinase in its active but metastable conformation. After cleavage of the loop, the protease becomes trapped as an acyl-enzyme with the serpin, and the cleaved serpin loop inserts itself as the central strand of one of the serpin β-sheets, accompanied by a dramatic change in protein stability. In certain mutant serpins, however, the exposed loop is conformationally more metastable and occasionally inserts itself into the β-sheet of a neighbouring serpin molecule, thereby forming serpin polymers with disastrous consequences for the patient (Carrell & Gooptu, 1998). In vitro, the polymerization of α1-antitrypsin can be reversed with synthetic homologues of the exposed peptide loop (Skinner et al., 1998). This approach might be useful for other `conformational diseases', which include Alzheimer's and other neurodegenerative disorders.

Another frequently occurring genetic disease is cystic fibrosis. Here we face a more complex situation than that in the case of sickle cell anaemia: a range of different mutations causes a malfunctioning of the same ion channel, which, consequently, leads to a range of severity of the disease (Collins, 1992). Protein crystallography is currently helpful in an indirect way in alleviating the problems of cystic fibrosis patients, not by studying the affected ion channel itself, but by revealing the structure of leukocyte elastase (Bode et al., 1986), an enzyme responsible for much of the cellular damage associated with cystic fibrosis (Birrer, 1995). On the basis of the elastase structure, inhibitors were developed to combat the effects of the impaired ion channel (Warner et al., 1994). Also, structures of key enzymes of Pseudomonas aeruginosa, a bacterium affecting many cystic fibrosis patients, form a basis for the design of therapeutics to treat infections by this pathogen. Yet, to the best of our knowledge, no compound has been developed so far that repairs the malfunctioning ion channel.

However, in some cases there might be more opportunities than assumed so far. Several mutations leading to genetic diseases result in a lack of stability of the affected protein. In instances when the mutant protein is still stable enough to fold, small molecules could conceivably be discovered that bind `anywhere' to a pocket of these proteins, thereby stabilizing the protein. The same small molecule could even be able to increase the stability of proteins with different mildly destabilizing mutations. Such an approach, though not trivial by any means, might be worth pursuing. Proof of principle of this concept has recently been provided for several unstable p53 mutants, where the same small molecule enhanced the stability of different mutants (Foster et al., 1999)

Of course, mutations that destroy cofactor binding or active sites, or destroy proper recognition of partner proteins, will be extremely difficult to correct by small molecules targeting the affected protein. In such instances, gene therapy is likely to be the way by which our and the next generation may be able to improve the lives of future generations.

The impact of detailed knowledge of protein and nucleic acid structures on the design of new drugs has already been significant, and promises to be of tremendous importance in the next decades. The first structure of a known major drug bound to a target protein was probably that of methotrexate bound to dihydrofolate reductase (DHFR) (Matthews et al., 1977). Even though the source of the enzyme was bacterial while methotrexate is used as a human anticancer agent, this protein–drug complex structure was nevertheless a hallmark achievement. It is generally accepted that the first protein-structure-inspired drug actually reaching the market was captopril, which is an antihypertensive compound blocking the action of angiotensin-converting enzyme, a metalloprotease. In this case, the structure of zinc-containing carboxypeptidase A was a guide to certain aspects of the chemical modification of lead compounds (Cushman & Ondetti, 1991). This success has been followed up by numerous projects specifically aimed at the design of new inhibitors, or activators, of carefully selected drug targets.

Structure-based drug design (SBDD) (Fig. 1.3.4.1) is the subject of several books and reviews that summarize projects and several success stories up until the mid-1990s (Kuntz, 1992; Perutz, 1992; Verlinde & Hol, 1994; Whittle & Blundell, 1994; Charifson, 1997; Veerapandian, 1997). Possibly the most dramatic impact made by SBDD has been on the treatment of AIDS, where the development of essentially all of the protease inhibitors on the market in 1999 has been guided by, or at least assisted by, the availability of numerous crystal structures of protease–inhibitor complexes.

Figure 1.3.4.1| top | pdf | The structure-based drug design cycle.

The need for a large number of structures is common in all drug design projects and is due to several factors. One is the tremendous challenge for theoretical predictions of the correct binding mode and affinity of inhibitors to proteins. The current force fields are approximate, the properties of water are treacherous, the flexibility of protein and ligands lead quickly to a combinatorial explosion, and the free-energy differences between various binding modes are small. All this leads to the need for several experimental structures in a structure-based drug design cycle (Fig. 1.3.4.1). In this cycle, numerous disciplines are interacting in multiple ways. Many institutions, small and large, are following in one way or another this paradigm to speed up the lead discovery, lead optimization and even the bioavailability improvement steps in the drug development process. Moreover, a very powerful synergism exists between combinatorial chemistry and structure-based drug design. Structure-guided combinatorial libraries can utilize knowledge of ligand target sites in the design of the library [see e.g. Ferrer et al. (1999), Eckert et al. (1999) and Minke et al. (1999)]. Once tight-binding ligands are found by combinatorial methods, crystal structures of library compound–target complexes provide detailed information for new highly specific libraries.

The fate of a drug candidate during clinical tests can hinge on a single methyl group – just as a point mutation can alter the benefit of a wild-type protein molecule into the nightmare of a life-long genetic disease. Hence, many promising inhibitors eventually fail to be of benefit to patients. Nevertheless, knowledge of a series of protein structures in complex with inhibitors is of immense value in the design and development of future pharmaceuticals. In the following sections some examples will be looked at.

1.3.4.1. Infectious diseases

| top | pdf |

1.3.4.1.1. Viral diseases

| top | pdf |

Some icosahedral pathogenic viruses have all their capsid proteins elucidated, while for the more complex viruses like influenza virus, hepatitis C virus (HCV) and HIV, numerous individual protein structures have been solved (Table 1.3.4.1). However, not all 14 native proteins of the HIV genome have yet surrendered to the crystallographic community, nor to the NMR spectroscopists or the high-resolution electron microscopists, our partners in experimental structural biology (Turner & Summers, 1999). Nevertheless, the structures of HIV protease, reverse transcriptase and fragments of HIV integrase and of HIV viral core and surface proteins are of tremendous value for developing novel anti-AIDS therapeutics [Arnold et al., 1996; Lin et al., 1998; Wlodawer & Vondrasek, 1998; see also references in Table 1.3.4.1(a)]. A similar situation occurs for hepatitis C virus. The protease structure of this virus has been solved recently (simultaneously by four groups!), as well as its helicase structure, providing platforms on the basis of which the design of novel drugs is actively pursued (Le et al., 1998).

Table 1.3.4.1| top | pdf | Important human pathogenic viruses and their proteins (a) RNA viruses (i) Single-stranded Family Example Protein structures solved Reference Arenaviridae Lassa fever virus None   Bunyaviridae Hantavirus None   Caliciviridae Hepatitis E virus, Norwalk virus None   Coronaviridae Corona virus None   Deltaviridae Hepatitis D virus Oligomerization domain of antigen [1] Filoviridae Ebola virus GP2 of membrane fusion glycoprotein [2] Flaviviridae Dengue NS3 protease [3] Hepatitis C NS3 protease [4], [5] RNA helicase [6] Yellow fever None   Tick-borne encephalitis virus Envelope glycoprotein [7] Orthomyxoviridae Influenza virus Neuraminidase [8] Haemagglutinin [9] Matrix protein M1 [10] Paramyxoviridae Measles, mumps, parainfluenza, respiratory syncytial virus None   Picornaviridae Hepatitis A virus 3C protease [11] Poliovirus Capsid [12] RNA-dependent polymerase [13] Rhinovirus Capsid [14] 3C protease [15] Echovirus Capsid [16] Retroviridae HIV Capsid protein [17] Matrix protein [18] Protease [19], [20], [21] Reverse transcriptase [22], [23], [47], [48], [49] Integrase [24] gp120 [25] NEF [26] gp41 [27] Rhabdovirus Rabies virus None   Togaviridae Rubella None   (ii) Double-stranded Family Example Protein structures solved Reference Reoviridae Rotavirus None   (b) DNA viruses (i) Single-stranded Family Example Protein structures solved Reference Parvoviridae B 19 virus None   (ii) Double-stranded Family Example Protein structures solved Reference Adenoviridae Adenovirus Protease [28] Capsid [29] Knob domain of fibre protein [30] Hepadnaviridae Hepatitis B Capsid [31] Herpesviridae Cytomegalovirus Protease [32], [33], [34] Epstein–Barr virus Domains of nuclear antigen 1 [35] BCRF1 [36] Herpes simplex Protease [37] Thymidine kinase [38] Uracyl-DNA glycosylase [39] Core of VP16 [40] Varicella zoster Protease [42] Papovaviridae Papillomavirus DNA-binding domain of E2 [43] Activation domain of E2 [44] Poxviridae Smallpox virus None   Vaccinia virus (related to smallpox but non-pathogenic) Methyltransferase VP39 [45] Domain of topoisomerase [46] References: [1] Zuccola et al. (1998); [2] Weissenhorn et al. (1998); [3] Murthy et al. (1999); [4] Love et al. (1996); [5] Yan et al. (1998); [6] Yao et al. (1997); [7] Rey et al. (1995); [8] Varghese et al. (1983); [9] Wilson et al. (1981); [10] Sha & Luo (1997); [11] Allaire et al. (1994); [12] Hogle et al. (1985); [13] Hansen et al. (1997); [14] Rossmann et al. (1985); [15] Matthews et al. (1994); [16] Filman et al. (1998); [17] Worthylake et al. (1999); [18] Hill et al. (1996); [19] Navia, Fitzgerald et al. (1989); [20] Wlodawer et al. (1989); [21] Erickson et al. (1990); [22] Rodgers et al. (1995); [23] Ding et al. (1995); [24] Dyda et al. (1994); [25] Kwong et al. (1998); [26] Lee et al. (1996); [27] Chan et al. (1997); [28] Ding et al. (1996); [29] Roberts et al. (1986); [30] Xia et al. (1994); [31] Wynne et al. (1999); [32] Tong et al. (1996); [33] Qiu et al. (1996); [34] Shieh et al. (1996); [35] Bochkarev et al. (1995); [36] Zdanov et al. (1997); [37] Hoog et al. (1997); [38] Wild et al. (1995); [39] Savva et al. (1995); [40] Liu et al. (1999); [42] Qiu et al. (1997); [43] Hegde & Androphy (1998); [44] Harris & Botchan (1999); [45] Hodel et al. (1996); [46] Sharma et al. (1994); [47] Kohlstaedt et al. (1992); [48] Jacobo-Molina et al. (1993); [49] Ren et al. (1995).

A quite spectacular example of how structural knowledge can lead to the synthesis of powerful inhibitors is provided by influenza virus neuraminidase. The structure of a neuraminidase–transition-state analogue complex suggested the addition of positively charged amino and guanidinium groups to the C4 position of the analogue, which resulted, in one step, in a gain of four orders of magnitude in binding affinity for the target enzyme (von Itzstein et al., 1993).

1.3.4.1.2. Bacterial diseases

| top | pdf |

A very large number of structures of important drug target proteins of bacterial origin have been solved crystallographically (Table 1.3.4.2). Currently, the most important single infectious bacterial agent is Mycobacterium tuberculosis, with three million deaths and eight million new cases annually (Murray & Salomon, 1998). The crystal structures of several M. tuberculosis potential and proven drug target proteins have been elucidated (Table 1.3.4.2). The complete M. tuberculosis genome has been sequenced recently (Cole et al., 1998), and this will undoubtedly have a tremendous impact on future drug development.

Table 1.3.4.2| top | pdf | Protein structures of important human pathogenic bacteria Organism Disease(s) Protein structures solved Reference Staphylococcus aureus Abscesses Alpha-haemolysin [1] Endocarditis Aureolysin [2] Gastroenteritis Beta-lactamase [3] Toxic shock syndrome Collagen adhesin [4] 7,8-Dihydroneopterin aldolase [5] Dihydropteroate synthetase [6] Enterotoxin A [7] Enterotoxin B [8] Enterotoxin C2 [9] Enterotoxin C3 [10] Exfoliative toxin A [11] Ile-tRNA-synthetase [12] Kanamycin nucleotidyltransferase [13] Leukocidin F [14] Nuclease [15] Staphopain [16] Staphylokinase [17] Toxic shock syndrome toxin-1 [18] Staphylococcus epidermidis Implant infections None   Enterococcus faecalis Urinary tract and biliary tract infections NADH peroxidase [19] (Streptococcus faecalis)   Histidine-containing phosphocarrier protein [20] Streptococcus mutans Endocarditis Glyceraldehyde-3-phosphate dehydrogenase [21] Streptococcus pneumoniae Pneumonia Penicillin-binding protein PBP2x [22] Meningitis, upper respiratory tract infections Dpnm DNA adenine methyltransferase [23] Streptococcus pyogenes Pharyngitis Inosine monophosphate dehydrogenase [24] Scarlet fever, toxic shock syndrome, immunologic disorders (acute glomerulonephritis and rheumatic fever) Pyrogenic exotoxin C [25] Bacillus anthracis Anthrax Anthrax protective antigen [26] Bacillus cereus Food poisoning Beta-amylase [27] Beta-lactamase II [28] Neutral protease [29] Oligo-1,6-glucosidase [30] Phospholipase C [31] Clostridium botulinum Botulism Neurotoxin type A [32] Clostridium difficile Pseudomembranous colitis None   Clostridium perfringens Gas gangrene Alpha toxin [33] Food poisoning Perfringolysin O [34] Clostridium tetani Tetanus Toxin C fragment [35] Corynebacterium diphtheriae Diphtheria Toxin [36] Toxin repressor [37] Listeria monocytogenes Meningitis, sepsis Phosphatidylinositol-specific phospholipase C [38] Actinomyces israelii Actinomycosis None   Nocardia asteroides Nocardiosis None   Neisseria gonorrhoeae Gonorrhea Type 4 pilin [39] Carbonic anhydrase [40] Neisseria meningitidis Meningitis Dihydrolipoamide dehydrogenase [41] Bordetella pertussis Whooping cough Toxin [42] Virulence factor P.69 [43] Brucella sp. Brucellosis None   Campylobacter jejuni Enterocolitis None   Enterobacter cloacae Urinary tract infection, pneumonia Beta-lactamase: class C [44] UDP-N-acetylglucosamine enolpyruvyltransferase [45] Escherichia coli        ETEC (enterotoxigenic) Traveller's diarrhoea Heat-labile enterotoxin [46] Heat-stable enterotoxin (is a peptide) [47]  EHEC (enterohaemorrhagic) HUS Verotoxin-1 [48]  EPEC (enteropathogenic) Diarrhoea None    EAEC (enteroaggregative) Diarrhoea None    EIEC (enteroinvasive) Diarrhoea None    UPEC (uropathogenic)   FimH adhesin [49] FimC chaperone [49] PapD [50]  NMEC (neonatal meningitis) Meningitis None   Franciscella tularensis Tularemia None   Haemophilus influenzae Meningitis, otitis media, pneumonia Diaminopimelate epimerase [51] 6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase [52] Ferric iron binding protein Mirp [53] Klebsiella pneumoniae Urinary tract infection, pneumonia, sepsis β-Lactamase SHV-1 [54] Legionella pneumophila Pneumonia None   Pasteurella multocida Wound infection None   Proteus mirabilis Pneumonia, urinary tract infection Catalase [55] Glutathione S-transferase [56] Proteus vulgaris Urinary tract infections Pvu II DNA-(cytosine N4) methyltransferase [57] Pvu II endonuclease [58] Tryptophanase [59] Salmonella typhi Typhoid fever None, but many for S. typhimurium linked with zoonotic disease   Salmonella enteridis Enterocolitis None   Serratia marcescens Pneumonia, urinary tract infection Serralysin [60] Aminoglycoside 3-N-acetyltransferase [61] Chitinase A [62] Chitobiase [63] Endonuclease [64] Hasa (haemophore) [65] Prolyl aminopeptidase [66] Shigella sp. Dysentery Chloramphenicol acetyltransferase [67] Shiga-like toxin I [68] Vibrio cholerae Cholera Cholera toxin [69], [70] DSBA oxidoreductase [71] Neuraminidase [104] Yersinia enterocolitica Enterocolitis Protein-Tyr phosphatase YOPH [72] Yersinia pestis Plague None   Pseudomonas aeruginosa Wound infection, urinary tract infection, pneumonia, sepsis Alkaline metalloprotease [73] Amidase operon [74] Azurin [75] Cytochrome 551 [76] Cytochrome c peroxidase [77] Exotoxin A [78] p-Hydroxybenzoate hydroxylase [79] Hexapeptide xenobiotic acetyltransferase [80] Mandelate racemase [81] Nitrite reductase [82] Ornithine transcarbamoylase [83] Porphobilinogen synthase [84] Pseudolysin [85] Burkholderia cepacia Wound infection, urinary tract infection, pneumonia, sepsis Biphenyl-cleaving extradiol dioxygenase [86] cis-Biphenyl-2,3-dihydrodiol-2,3-dehydrogenase [87] Dialkylglycine decarboxylase [88] Lipase [89] Phthalate dioxygenase reductase [90] Stenotrophomonas maltophilia (= Pseudomonas maltophilia) Sepsis None   Bacteroides fragilis Intra-abdominal infections Beta-lactamase type 2 [91] Mycobacterium leprae Leprosy Chaperonin-10 (GroES homologue) [92] RUVA protein [93] Mycobacterium tuberculosis Tuberculosis 3-Dehydroquinate dehydratase [94] Dihydrofolate reductase [95] Dihydropteroate synthase [96] Enoyl acyl-carrier-protein reductase (InhA) [97] Mechanosensitive ion channel [98] Quinolinate phosphoribosyltransferase [99] Superoxide dismutase (iron dependent) [100] Iron-dependent repressor   Mycobacterium bovis Tuberculosis Tetrahydrodipicolinate-N-succinyltransferase [102] Chlamydia psitacci Psittacosis None   Chlamydia pneumoniae Atypical pneumonia None   Chlamydia trahomatis Ocular, respiratory and genital infections None   Coxiella burnetii Q fever None   Rickettsia sp. Rocky Mountain spotted fever None   Borrelia burgdorferi Lyme disease Outer surface protein A [103] Leptospira interrogans Leptospirosis None   Treponema pallidum Syphilis None   Mycoplasma pneumoniae Atypical pneumonia None   References: [1] Song et al. (1996); [2] Banbula et al. (1998); [3] Herzberg & Moult (1987); [4] Symersky et al. (1997); [5] Hennig et al. (1998); [6] Hampele et al. (1997); [7] Sundstrom et al. (1996); [8] Papageorgiou et al. (1998); [9] Papageorgiou et al. (1995); [10] Fields et al. (1996); [11] Vath et al. (1997); [12] Silvian et al. (1999); [13] Pedersen et al. (1995); [14] Pedelacq et al. (1999); [15] Loll & Lattman (1989); [16] Hofmann et al. (1993); [17] Rabijns et al. (1997); [18] Prasad et al. (1993); [19] Yeh et al. (1996); [20] Jia et al. (1993); [21] Cobessi et al. (1999); [22] Pares et al. (1996); [23] Tran et al. (1998); [24] Zhang, Evans et al. (1999); [25] Roussel et al. (1997); [26] Petosa et al. (1997); [27] Mikami et al. (1999); [28] Carfi et al. (1995); [29] Pauptit et al. (1988); [30] Watanabe et al. (1997); [31] Hough et al. (1989); [32] Lacy et al. (1998); [33] Naylor et al. (1998); [34] Rossjohn, Feil, McKinstry et al. (1997); [35] Umland et al. (1997); [36] Choe et al. (1992); [37] Qiu et al. (1995); [38] Moser et al. (1997); [39] Parge et al. (1995); [40] Huang, Xue et al. (1998); [41] Li de la Sierra et al. (1997); [42] Stein et al. (1994); [43] Emsley et al. (1996); [44] Lobkovsky et al. (1993); [45] Schonbrunn et al. (1996); [46] Sixma et al. (1991); [47] Ozaki et al. (1991); [48] Stein et al. (1992); [49] Choudhury et al. (1999); [50] Sauer et al. (1999); [51] Cirilli et al. (1993); [52] Hennig et al. (1999); [53] Bruns et al. (1997); [54] Kuzin et al. (1999); [55] Gouet et al. (1995); [56] Rossjohn, Polekhina et al. (1998); [57] Gong et al. (1997); [58] Athanasiadis et al. (1994); [59] Isupov et al. (1998); [60] Baumann (1994); [61] Wolf et al. (1998); [62] Perrakis et al. (1994); [63] Tews et al. (1996); [64] Miller et al. (1994); [65] Arnoux et al. (1999); [66] Yoshimoto et al. (1999); [67] Murray et al. (1995); [68] Ling et al. (1998); [69] Merritt et al. (1994); [70] Zhang et al. (1995); [71] Hu et al. (1997); [72] Stuckey et al. (1994); [73] Miyatake et al. (1995); [74] Pearl et al. (1994); [75] Adman et al. (1978); [76] Almassy & Dickerson (1978); [77] Fulop et al. (1995); [78] Allured et al. (1986); [79] Gatti et al. (1994); [80] Beaman et al. (1998); [81] Kallarakal et al. (1995); [82] Nurizzo et al. (1997); [83] Villeret et al. (1995); [84] Frankenberg et al. (1999); [85] Thayer et al. (1991); [86] Han et al. (1995); [87] Hulsmeyer et al. (1998); [88] Toney et al. (1993); [89] Kim et al. (1997); [90] Correll et al. (1992); [91] Concha et al. (1996); [92] Mande et al. (1996); [93] Roe et al. (1998); [94] Gourley et al. (1999); [95] Li et al. (2000); [96] Baca et al. (2000); [97] Dessen et al. (1995); [98] Chang, Spencer et al. (1998); [99] Sharma et al. (1998); [100] Cooper et al. (1995); [102] Beaman et al. (1997); [103] Li et al. (1997); [104] Crennell et al. (1994).

The crystal structures of many bacterial dihydrofolate reductases, the target of several antimicrobials including trimethoprim, have also been reported. Recently, the atomic structure of dihydropteroate synthase (DHPS), the target of sulfa drugs, has been elucidated, almost 60 years after the first sulfa drugs were used to treat patients (Achari et al., 1997; Hampele et al., 1997).

A very special set of bacterial proteins are the toxins. Some of these have dramatic effects, with even a single molecule able to kill a host cell. These toxins outsmart and (mis)use many of the defence systems of the host, and their structures are often most unusual and fascinating, as recently reviewed by Lacy & Stevens (1998). The structures of the toxins are actively used for the design of prophylactics and therapeutic agents to treat bacterial diseases [see e.g. Merritt et al. (1997), Kitov et al. (2000) and Fan et al. (2000)]. It is remarkable that the properties of these devastating toxins are also used, or at least considered, for the treatment of disease, such as in the engineering of diphtheria toxin to create immunotoxins for the treatment of cancer and the deployment of cholera toxin mutants as adjuvants in mucosal vaccines. Knowledge of the three-dimensional structures of these toxins assists in the design of new therapeutically useful proteins.

1.3.4.1.3. Protozoan infections

| top | pdf |

A major cause of death and worldwide suffering is due to infections by several protozoa, including:

(a) Plasmodium falciparum and related species, causing various forms of malaria; (b) Trypanosoma cruzi, the causative agent of Chagas' disease in Latin America; (c) Trypanosoma brucei, causing sleeping sickness in Africa; (d) Some eleven different Leishmania species, responsible for several of the most horribly disfiguring diseases known to mankind.

Drug resistance, combined with other factors, has been the cause of a major disappointment for the early hopes of a `malaria eradication campaign'. Fortunately, new initiatives have been launched recently under the umbrella of the `Malaria roll back' program and the `Multilateral Initiative for Malaria' (MIM). We are facing a formidable challenge, however, since the parasite is very clever at evading the immune response of the human host. Drugs are the mainstay of current treatments and may well be so for a long time to come. Protein crystallographic studies of Plasmodium proteins are hampered by the unusual codon usage of the Plasmodium species, coupled with a tendency to contain insertions of numerous hydrophilic residues in the polypeptide chain (Gardner et al., 1998) which provide sometimes serious obstacles to obtaining large amounts of Plasmodium proteins for structural investigations.

However, the structures of an increasing number of potential drug targets from these protozoan parasites are being unravelled. These include the variable surface glycoproteins (VSGs) and glycolytic enzymes of Trypanosoma brucei, crucial malaria proteases, and the remarkable trypanothione reductase (Table 1.3.4.3). The structures of nucleotide phosphoribosyl transferases of a variety of protozoan parasites have also been elucidated. Moreover, the structure of DHFR from Pneumocystis carinii, the major opportunistic pathogen in AIDS patients in the United States, has been determined. Several of these structures are serving as starting points for the development of new drugs.

Table 1.3.4.3| top | pdf | Protein structures of important human pathogenic protozoa, fungi and helminths (a) Protozoa Organism Disease Protein structures solved Reference Acanthamoeba sp. Opportunistic meningoencephalitis, corneal ulcers Actophorin [1] Profilin [2] Cryptosporidium parvum Cryptosporidiosis None   Entamoeba histolytica Amoebic dysentery, liver abscesses None   Giardia lamblia Giardiasis None   Leishmania sp. Leishmaniasis Adenine phosphoribosyltransferase [3] Dihydrofolate reductase-thymidylate synthase [4] Glyceraldehyde-3-phosphate dehydrogenase [5] Leishmanolysin [6] Nucleoside hydrolase [7] Pyruvate kinase [8] Triosephosphate isomerase [9] Plasmodium sp. Malaria Fructose-1,6-bisphosphate aldolase [10] Lactate dehydrogenase [11] MSP1 [12] Plasmepsin II [13] Purine phosphoribosyltransferase [14] Triosephosphate isomerase [15] Pneumocystis carinii Pneumonia Dihydrofolate reductase [16] Toxoplasma gondii Toxoplasmosis HGXPRTase [17] UPRTase [18] Trichomonas vaginalis Trichomoniasis None   Trypanosoma brucei Sleeping sickness Fructose-1,6-bisphosphate aldolase [19] Glyceraldehyde-3-phosphate dehydrogenase [20] 6-Phosphogluconate dehydrogenase [21] Phosphoglycerate kinase [22] Triosephosphate isomerase [23] VSG [24] Trypanosoma cruzi Chagas' disease Cruzain (cruzipain) [25] Glyceraldehyde-3-phosphate dehydrogenase [26] Hypoxanthine phosphoribosyltransferase [27] Triosephosphate isomerase [28] Trypanothione reductase [29] Tyrosine aminotransferase [30] (b) Fungi Organism Disease Protein structures solved Reference Aspergillus fumigatus Aspergillosis Restrictocin [31] Blastomyces dermatidis Blastomycosis None   Candida albicans Candidiasis Dihydrofolate reductase [32] N-Myristoyl transferase [33] Phosphomannose isomerase [34] Secreted Asp protease [35] Coccidiodes immitis Coccidioidomycosis None   Cryptococcus neoformans Cryptococcosis None   Histoplasma capsulatum Histoplasmosis None   Mucor sp. Mucormycosis None   Paracoccidioides brasiliensis Paracoccidioidomycosis None   Rhizopus sp. Phycomycosis Lipase II [36] Rhizopuspepsin [37] RNase Rh [38] (c) Helminths Organism Disease Protein structures solved Reference Clonorchis sinensis Clonorchiasis None   Fasciola hepatica Fasciolasis Glutathione S-transferase [39] Fasciolopsis buski Fasciolopsiasis None   Paragominus westermani Paragonimiasis None   Schistosoma sp. Schistosomiasis Glutathione S-transferase [39], [40] Hexokinase [41] Diphyllobotrium latum Diphyllobothriasis None   Echinococcus granulosus Unilocular hydatid cyst disease None   Taenia saginata Taeniasis None   Taenia solium Taeniasis None   Ancylostoma duodenale Old World hookworm disease None   Anisakis Anisakiasis None   Ascaris lumbricoides Ascariasis Haemoglobin [42] Major sperm protein [43] Trypsin inhibitor [44] Enterobius vermicularis Pinworm infection None   Necator New World hookworm disease None   Strongyloides stercoralis Strongyloidiasis None   Trichinella spiralis Trichinosis None   Trichuris trichiura Whipworm infection None   Brugia malayi Filariasis Peptidylprolyl isomerase [45], [46] Dracunculus medinensis Guinea worm disease None   Loa loa Loiasis None   Onchocerca volvulus River blindness None   Toxocara canis Visceral larva migrans None   Wuchereria bancrofti Lymphatic filariasis (elephantiasis) None   References: [1] Leonard et al. (1997); [2] Liu et al. (1998); [3] Phillips et al. (1999); [4] Knighton et al. (1994); [5] Kim et al. (1995); [6] Schlagenhauf et al. (1998); [7] Shi, Schramm & Almo (1999); [8] Rigden et al. (1999); [9] Williams et al. (1999); [10] Kim et al. (1998); [11] Read et al. (1999); [12] Chitarra et al. (1999); [13] Silva et al. (1996); [14] Shi, Li et al. (1999); [15] Velanker et al. (1997); [16] Champness et al. (1994); [17] Schumacher et al. (1996); [18] Schumacher et al. (1998); [19] Chudzik et al. (2000); [20] Vellieux et al. (1993); [21] Phillips et al. (1998); [22] Bernstein et al. (1998); [23] Wierenga et al. (1987); [24] Freymann et al. (1990); [25] McGrath et al. (1995); [26] Souza et al. (1998); [27] Focia et al. (1998); [28] Maldonado et al. (1998); [29] Lantwin et al. (1994); [30] Blankenfeldt et al. (1999); [31] Yang & Moffat (1995); [32] Whitlow et al. (1997); [33] Weston et al. (1998); [34] Cleasby et al. (1996); [35] Cutfield et al. (1995); [36] Kohno et al. (1996); [37] Suguna et al. (1987); [38] Kurihara et al. (1992); [39] Rossjohn, Feil, Wilce et al. (1997); [40] McTigue et al. (1995); [41] Mulichak et al. (1998); [42] Yang et al. (1995); [43] Bullock et al. (1996); [44] Huang et al. (1994); [45] Mikol et al. (1998); [46] Taylor et al. (1998).

1.3.4.1.4. Fungi

| top | pdf |

In general, the human immune system is able to keep the growth of fungi under control, but in immuno-compromised patients (e.g. as a result of cancer chemotherapy, HIV infection, transplantation patients receiving immunosuppressive drugs, genetic disorders) such organisms are given opportunities they usually do not have. Candida albicans is an opportunistic fungal organism which causes serious complications in immuno-compromised patients. Several of its proteins have been structurally characterized (Table 1.3.4.3) and provide opportunities for the development of selectively active inhibitors.

1.3.4.1.5. Helminths

| top | pdf |

Worms affect the lives of billions of human beings, causing serious morbidity in many ways. Onchocerca volvolus is the cause of river blindness, which resulted in the virtual disappearance of entire villages in West Africa, until ivermectin appeared. This remarkable compound dramatically reduced the incidence of the disease, even though it does not kill the adult worms. Therefore, a biological clock is ticking, waiting until resistance occurs against this single compound available for treatment. Schistosoma species are responsible for a wide variety of liver diseases and are spreading continuously since irrigation schemes provide a perfect environment for the intermediate snail vector. Other medically important helminths are Wuchereria bancrofti and Brugia malayi. Only a few protein structures from these very important disease-causing organisms have been unravelled so far (Table 1.3.4.3).

1.3.4.3. Non-communicable diseases

| top | pdf |

Of this large and diverse category of human afflictions we have already touched upon genetic disorders in Section 1.3.3 above. Other major types of non-communicable diseases include cancer, aging disorders, diabetes, arthritis, and cardiovascular and neurological illnesses. The field of non-communicable diseases is immense. Describing in any detail the current projects in, and potential impact of, protein and nucleic acid crystallography on these diseases would need more space than this entire volume on macromolecular crystallography. Hence, only a few selected examples out of the hundreds which could be described can be discussed here. Table 1.3.4.5 lists many examples of human protein structures elucidated without any claim as to completeness – it is simply impossible to keep up with the fountain of structures being determined at present. Yet, such tables do provide, it is hoped, an overview of what has been achieved and what needs to be done.

Table 1.3.4.5| top | pdf | Important human protein structures in drug design Proteins from other species that might have been studied as substitutes for human ones were left out because of space limitations. We apologize to the researchers affected. Pharmacological category Protein Reference Synaptic and neuroeffector junctional function None   Central nervous system function None   Inflammation Fibroblast collagenase (MMP-1) (also important in cancer) [1], [2], [3], [4] Gelatinase [5], [6] Stromelysin-1 (MMP-3) (also important in cancer) [7], [8], [9], [10], [11] Matrilysin (MMP-7) (also important in cancer) [12] Neutrophil collagenase (MMP-8) (also important in cancer) [13], [14], [15], [16] Collagenase-3 (MMP-13) [17] Human neutrophil elastase (also important for cystic fibrosis) [18], [19], [20] Interleukin-1 beta converting enzyme (ICE) [21], [22] p38 MAP kinase [23], [24] Phospholipase A2 [25], [26], [27] Renal and cardiovascular function Renin [28] Gastrointestinal function None   Cancer 17-Beta-hydroxysteroid dehydrogenase [29], [30] BRCT domain (BRCA1 C-terminus) [31] Bcr-Abl kinase [32] Cathepsin B [33] Cathepsin D [34], [35] CDK2 [36] CDK6 [37] DHFR [38], [39] Acidic fibroblast growth factor (FGF) [40] FGF receptor tyrosine kinase domain [41] Glycinamide ribonucleotide formyl transferase [42] Interferon-beta [43] MMPs: see Inflammation   p53 [44], [45] p60 Src [46] Purine nucleoside phosphorylase [47] ras p21 [48], [49], [50], [51] Serine hydroxymethyltransferase [52] S-Adenosylmethionine decarboxylase [53] Thymidylate synthase [54] Topoisomerase I [55], [56] Tumour necrosis factor [57] Interleukin 1-alpha [58] Interleukin 1-beta [59] Interleukin 1-beta receptor [60], [61] Interleukin 8 [62] Immunomodulation Calcineurin [63] Cathepsin S [64] Cyclophilin [65], [66], [67] Immunophilin FKBP12 [68], [69], [70] Inosine monophosphate dehydrogenase [71] Interferon-gamma [72], [73] Lymphocyte-specific kinase Lck [74] PNP [47] Syk kinase [75] Tumour necrosis factor [57] ZAP Tyr kinase [76] Interleukin 2 [77] Interleukin 5 [78] Haematopoiesis Erythropoietin receptor [79], [80] Coagulation AT III [81], [82], [83], [84] Factor III [85], [86] Factor VII [87] Factor IX [88] Factor X [89] Factor XIII [90] Factor XIV [91] Fibrinogen: fragment [92], [93] Plasminogen activator inhibitor (PAI) [94], [95], [96] Thrombin [97], [98], [99] tPA [100] Urokinase-type plasminogen activator [101] von Willebrand factor [102], [103], [104] Hormones and hormone receptors Insulin [105] Insulin receptor [106], [107] Human growth hormone + receptor [108] Oestrogen receptor [109], [110] Progesterone receptor [111] Prolactin receptor [112] Ocular function Carbonic anhydrase [113] Genetic diseases See Table 1.3.3.1   Drug binding Human serum albumin [114], [115] Drug metabolism Glutathione S-transferase A-1 [116], [117] Glutathione S-transferase A4–4 [118] Glutathione S-transferase Mu-1 [119] Glutathione S-transferase Mu-2 [120] Neurodegeneration Aldose reductase [121] JNK3 [122] Osteoporosis Cathepsin K [123], [64] Src SH2 [126] Various Interferon-alpha 2b [124] Bcl-xL [125] References: [1] Borkakoti et al. (1994); [2] Lovejoy, Cleasby et al. (1994); [3] Lovejoy, Hassell et al. (1994); [4] Spurlino et al. (1994); [5] Libson et al. (1995); [6] Gohlke et al. (1996); [7] Becker et al. (1995); [8] Dhanaraj et al. (1996); [9] Esser et al. (1997); [10] Gomis-Ruth et al. (1997); [11] Finzel et al. (1998); [12] Browner et al. (1995); [13] Bode et al. (1994); [14] Reinemer et al. (1994); [15] Stams et al. (1994); [16] Betz et al. (1997); [17] Gomis-Ruth et al. (1996); [18] Bode et al. (1986); [19] Wei et al. (1988); [20] Navia, McKeever et al. (1989); [21] Walker et al. (1994); [22] Rano et al. (1997); [23] Wilson et al. (1996); [24] Tong et al. (1997); [25] Scott et al. (1991); [26] Wery et al. (1991); [27] Kitadokoro et al. (1998); [28] Sielecki et al. (1989); [29] Ghosh et al. (1995); [30] Breton et al. (1996); [31] Zhang et al. (1998); [32] Nam et al. (1996); [33] Musil et al. (1991); [34] Baldwin et al. (1993); [35] Metcalf & Fusek (1993); [36] De Bondt et al. (1993); [37] Russo et al. (1998); [38] Oefner et al. (1988); [39] Davies et al. (1990); [40] Blaber et al. (1996); [41] McTigue et al. (1999); [42] Varney et al. (1997); [43] Karpusas et al. (1997); [44] Cho et al. (1994); [45] Gorina & Pavletich (1996); [46] Xu et al. (1997); [47] Ealick et al. (1990); [48] DeVos et al. (1988); [49] Pai et al. (1989); [50] Krengel et al. (1990); [51] Scheffzek et al. (1997); [52] Renwick et al. (1998); [53] Ekstrom et al. (1999); [54] Schiffer et al. (1995); [55] Redinbo et al. (1998); [56] Stewart et al. (1998); [57] Banner et al. (1993); [58] Graves et al. (1990); [59] Priestle et al. (1988); [60] Schreuder et al. (1997); [61] Vigers et al. (1997); [62] Baldwin et al. (1991); [63] Kissinger et al. (1995); [64] McGrath et al. (1998); [65] Kallen et al. (1991); [66] Ke et al. (1991); [67] Pfuegl et al. (1993); [68] Van Duyne, Standaert, Karplus et al. (1991); [69] Van Duyne, Standaert, Schreiber & Clardy (1991); [70] Van Duyne et al. (1993); [71] Colby et al. (1999); [72] Ealick et al. (1991); [73] Walter et al. (1995); [74] Zhu et al. (1999); [75] Futterer et al. (1998); [76] Meng et al. (1999); [77] Brandhuber et al. (1987); [78] Milburn et al. (1993); [79] Livnah et al. (1996); [80] Livnah et al. (1998); [81] Carrell et al. (1994); [82] Schreuder et al. (1994); [83] Skinner et al. (1997); [84] Skinner et al. (1998); [85] Muller et al. (1994); [86] Muller et al. (1996); [87] Banner et al. (1996); [88] Rao et al. (1995); [89] Padmanabhan et al. (1993); [90] Yee et al. (1994); [91] Mather et al. (1996); [92] Pratt et al. (1997); [93] Spraggon et al. (1997); [94] Mottonen et al. (1992); [95] Aertgeerts et al. (1995); [96] Xue et al. (1998); [97] Bode et al. (1989); [98] Rydel et al. (1990); [99] Rydel et al. (1994); [100] Laba et al. (1996); [101] Spraggon et al. (1995); [102] Bienkowska et al. (1997); [103] Huizinga et al. (1997); [104] Emsley et al. (1998); [105] Ciszak & Smith (1994); [106] Hubbard et al. (1994); [107] Hubbard (1997); [108] DeVos et al. (1992); [109] Schwabe et al. (1993); [110] Brzozowski et al. (1997); [111] Williams & Sigler (1998); [112] Somers et al. (1994); [113] Kannan et al. (1975); [114] He & Carter (1992); [115] Curry et al. (1998); [116] Sinning et al. (1993); [117] Cameron et al. (1995); [118] Bruns et al. (1999); [119] Tskovsky et al. (1999); [120] Raghunathan et al. (1994); [121] Wilson et al. (1992); [122] Xie et al. (1998); [123] Thompson et al. (1997); [124] Radhakrishnan et al. (1996); [125] Muchmore et al. (1996); [126] Waksman et al. (1993).

1.3.4.3.1. Cancers

| top | pdf |

Over a hundred different cancers have been described and clearly the underlying defect, loss of control of cell division, can be the result of many different shortcomings in different cells. The research in this area proceeds at a feverish pace, yet the development, discovery and design of effective but safe anti-cancer agents are unbelievably difficult challenges. The modifications needed to turn a normal cell into a malignant one are very small, hence the chance of arriving at `true' anti-cancer drugs that exploit such small differences between normal and abnormal cells is exceedingly small. Nevertheless, such selective anti-cancer agents would leave normal cells essentially unaffected and are therefore the holy grail of cancer therapy. Few if any such compounds have been found so far, but cancer therapy is benefiting from a gradual increase in the number of useful compounds. Many have serious side effects, weaken the immune system and are barely tolerated by patients. However, they rescue large numbers of patients and hence it is of interest that many targets of these compounds, proteins and DNA molecules, have been structurally elucidated by crystallographic methods – often in complex with the cancer drug. The mode of action of many anti-cancer compounds is well understood, e.g. methotrexate targeting dihydrofolate reductase, and fluorouracil targeting thymidilate synthase. These are specific enzyme inhibitors acting along principles well known in other areas of medicine. Several anti-cancer drugs display unusual modes of action, such as:

(a) the DNA intercalators daunomycin (Wang et al., 1987) and adriamycin (Zhang et al., 1993); (b) cisplatin, which forms DNA adducts (Giulian et al., 1996); (c) taxol, which not only binds to tubulin but also to bcl-2, thereby blocking the machinery of cancer cells in two entirely different ways (Amos & Lowe, 1999); (d) camptothecin analogues, such as irinotecan and topotecan, which have the unusual property of prolonging the lifetime of a covalent topoisomerase–DNA complex, generating major road blocks on the DNA highway and causing DNA breakage and cell death; (e) certain compounds function as minor-groove binders, e.g. netropsin and distamycin (Kopka et al., 1985); (f) completely new drugs which were developed based on the structures of matrix metalloproteinases, purine nucleotide phosphorylase and glycinamide ribonucleotide formyltransferase and which are in clinical trials (Jackson, 1997).

Meanwhile, it is sad that crystallography has not yet made any contribution to the molecular understanding of multi-drug resistance in cancer. The resistance is caused by cellular pumps that efficiently pump out the drugs, often leading to failed chemotherapy (Borst, 1999). On the other hand, the structures of major oncogenic proteins such as p21 (DeVos et al., 1988; Pai et al., 1989; Krengel et al., 1990; Scheffzek et al., 1997) and p53 (Cho et al., 1994; Gorina & Pavletich, 1996) are of tremendous importance for future structure-based design of anti-neoplastic agents.

1.3.4.3.2. Diabetes

| top | pdf |

The hallmark characteristic of type I diabetes is a lack of insulin. A major therapeutic approach to this problem is insulin replacement therapy. Unfortunately, the insulin requirements of the body vary dramatically during the course of a day, with high concentrations needed at meal times and a basal level during the rest of the day. Only monomeric insulin is active at the insulin receptor level, but insulin has a natural tendency to form dimers and hexamers that dissociate upon dilution. Thanks to the three-dimensional insight obtained from dozens of insulin crystal structures, as wild-type (Hodgkin, 1971), mutants (Whittingham et al., 1998) and in complex with zinc ions and small molecules such as phenol (Derewenda et al., 1989), it has been possible to fine-tune the kinetics of insulin dissociation. The resulting availability of a variety of insulin preparations with rapid or prolonged action profiles has improved the quality of life of millions of people (Brange, 1997).

1.3.4.3.3. Blindness

| top | pdf |

The main causes of blindness worldwide are cataract, trachoma, glaucoma and onchocerciasis (Thylefors et al., 1995). Trachoma and onchocerciasis are parasitic diseases that destroy the architecture of the eye; they were discussed in Section 1.3.4.1. The other two are discussed here. During cataract development, the lens of the eye becomes non-transparent as a result of aggregation of crystallins, preventing image formation. Crystal structures of several mammalian beta- and gamma-crystallins are known, but no human ones yet. In glaucoma, the optic nerve is destroyed by high intra-ocular pressure. One way to lower the pressure is to inhibit carbonic anhydrase II, a pivotal enzyme in maintaining the intra-ocular pressure. On the basis of the carbonic anhydrase crystal structure, researchers at Merck Research Laboratories were able to guide the optimization of an S-thienothiopyran-2-sulfonamide lead into a marketed drug for glaucoma: dorzolamide (Baldwin et al., 1989).

1.3.4.3.4. Cardiovascular disorders

| top | pdf |

Thrombosis is a major cause of morbidity and mortality, especially in the industrial world. Hence, major effort is expended by pharmaceutical industries in the development of new classes of anti-coagulants with fewer side effects than available drugs, such as heparins and coumarins. Because blood coagulation is the result of an amplification cascade of enzymatic reactions, many potential targets are available. At present most of the effort is directed towards thrombin (Weber & Czarniecki, 1997) and factor Xa (Ripka, 1997), responsible for the penultimate step and the step immediately preceding it in the cascade, respectively. Thrombin is especially fascinating owing to the presence of at least three subsites: a primary specificity pocket with the catalytic serine-protease machinery, an exosite for recognizing extended fibrinogen and an additional pocket for binding heparin. This knowledge has led to the design of bivalent inhibitors which occupy two sites with ultra-high affinity and exquisite specificity. Several of these agents are in clinical trials (Pineo & Hull, 1999).

1.3.4.3.5. Neurological disorders

| top | pdf |

Even a quick glance at Table 1.3.4.5 shows that crystallography contributes to new therapeutics for numerous human afflictions and diseases. Yet there are major gaps in our understanding of protein functions, in particular of those involved in development and in neurological functions. These proteins are the target of many drugs obtained by classical pre-crystal-structure methods. These proven drug targets are very often membrane proteins involved in neuronal functions, and the diseases concerned are some of the most prevalent in mankind. A non-exhaustive list includes cerebrovascular disease (strokes), Parkinson's, epilepsy, schizophrenia, bipolar disease and depression.

Some of these diseases are heart-breaking afflictions, where parents have to accept the suicidal tendencies of their children, often with fatal outcomes; where partners have to endure the tremendous mood swings of their bipolar spouses and have to accept extreme excesses in behaviour; where a happy evening of life is turned into the gradual and sad demise of human intellect due to the progression of Alzheimer's, or to the loss of motor functions due to Parkinson's, or into the tragic stare of a victim of deep depression. Human nature, in all its shortcomings, has the tendency to try to help such tragic victims, but drugs for neurological disorders are rare, drug regimens are difficult to optimize and the commitment to follow a drug regimen – often for years, and often with major side effects – is a next to impossible task in many cases. New, better drugs are urgently needed and hence the structure determinations of the `molecules of the brain' are major scientific as well as medical challenges of the next decades. Such molecules will shed light on some of the deepest mysteries of humanity, including memory, cognition, desire, sleep etc. At the same time, such structures will provide opportunities for treating those suffering from neurodegenerative diseases due to age, genetic disposition, allergies, infections, traumas and combinations thereof. Such `CNS protein structures' are one of the major challenges of biomacromolecular crystallography in the 21st century.

Vaccines are probably the most effective way of preventing disease. An impressive number of vaccines have been developed and many more are under development (National Institute of Allergy and Infectious Diseases, 1998). Smallpox has been eradicated thanks to a vaccine, and polio is being targeted for eradication in a worldwide effort, again using vaccination strategies. To the best of our knowledge, crystal structures of viruses, viral capsids or viral proteins have not been used in developing the currently available vaccines. However, there are projects underway that may change this.

For instance, the crystal structure of rhinovirus has resulted in the development of compounds that have potential as antiviral agents, since they stabilize the viral capsid and block, or at least delay, the uncoating step in viral cell entry (Fox et al., 1986). These rhinovirus capsid-stabilizing compounds are, in a different project, being used to stabilize poliovirus particles against heat-induced denaturation in vaccines (Grant et al., 1994). This approach may be applicable to other cases, although it has not yet resulted in commercially available vaccine-plus-stabilizer cocktails. However, it is fascinating to see how a drug-design project may be able to assist vaccine development in a rather unexpected manner.

Three-dimensional structural information about viruses is also being used to aid in the development of vaccines. Knowledge of the architecture of and biological functions of coat proteins has been used to select loops at viral surfaces that can be replaced with antigenic loops from other pathogens for vaccine-engineering purposes (e.g. Burke et al., 1988; Kohara et al., 1988; Martin et al., 1988; Murray et al., 1998; Arnold et al., 1994; Resnick et al., 1995; Smith et al., 1998; Arnold & Arnold, 1999; Zhang, Geisler et al., 1999). The design of human rhinovirus (HRV) and poliovirus chimeras has been aided by knowing the atomic structure of the viruses (Hogle et al., 1985; Rossmann et al., 1985; Arnold & Rossmann, 1988; Arnold & Rossmann, 1990) and detailed features of the neutralizing immunogenic sites on the virion surfaces (Sherry & Rueckert, 1985; Sherry et al., 1986). In this way, one can imagine that in cases where the atomic structures of antigenic loops in `donor' immunogens are known as well as the structure of the `recipient' loop in the virus capsid protein, optimal loop transplantation might become possible. It is not yet known how to engineer precisely the desired three-dimensional structures and properties into macromolecules. However, libraries of macromolecules or viruses constructed using combinatorial mutagenesis can be searched to increase the likelihood of including structures with desired architecture and properties such as immunogenicity. With appropriate selection methods, the rare constructs with desired properties can be identified and `fished out'. Research of this type has yielded some potently immunogenic presentations of sequences transplanted on the surface of HRV (reviewed in Arnold & Arnold, 1999). For reasons not quite fully understood, presenting multiple copies of antigens to the immune system leads to an enhanced immune response (Malik & Perham, 1997). It is conceivable that, eventually, it might even be possible for conformational epitopes consisting of multiple `donor' loops to be grafted onto `recipient capsids' while maintaining the integrity of the original structure. Certainly, such feats are difficult to achieve with present-day protein-engineering skills, but recent successes in protein design offer hope that this will be feasible in the not too distant future (Gordon et al., 1999).

Immense efforts have been made by numerous crystallographers to unravel the structures of molecules involved in the unbelievably complex, powerful and fascinating immune system. Many of the human proteins studied are listed in Table 1.3.4.5 with, as specific highlights, the structures of immunoglobulins (Poljak et al., 1973), major histocompatibility complex (MHC) molecules (Bjorkman et al., 1987; Brown et al., 1993; Fremont et al., 1992; Bjorkman & Burmeister, 1994), T-cell receptors (TCR) and MHC:TCR complexes (Garboczi et al., 1996; Garcia et al., 1996), an array of cytokines and chemokines, and immune cell-specific kinases such as lck (Zhu et al., 1999). This knowledge is being converted into practical applications, for instance by humanising non-human antibodies with desirable properties (Reichmann et al., 1988) and by creating immunotoxins.

The interactions between chemokines and receptors, and the complicated signalling pathways within each immune cell, make it next to impossible to predict the effect of small compounds interfering with a specific protein–protein interaction in the immune system (Deller & Jones, 2000). However, great encouragement has been obtained from the discovery of the remarkable manner by which the immunosuppressor FK506 functions: this small molecule brings two proteins, FKB12 and calcineurin, together, thereby preventing T-cell activation by calcineurin. The structure of this remarkable ternary complex is known (Kissinger et al., 1995). Such discoveries of unusual modes of action of therapeutic compounds are the foundation for new concepts such as `chemical dimerizers' to activate signalling events in cells such as apoptosis (Clackson et al., 1998).

In spite of the gargantuan task ahead aimed at unravelling the cell-to-cell communication in immune action, it is unavoidable that the next decades will bring us unprecedented insight into the many carefully controlled processes of the immune system. In turn, it is expected that this will lead to new therapeutics for manipulating a truly wonderful defence system in order to assist vaccines, to decrease graft rejection processes in organ transplants and to control auto-immune diseases that are likely to be playing a major role in cruelly debilitating diseases such as rheumatoid arthritis and type I diabetes.

At the beginning of the 1990s, Max Perutz inspired many researchers with a passion for structure and a heart for the suffering of mankind with a fascinating book entitled Protein Structure – New Approaches to Disease and Therapy (Perutz, 1992). The explosion of medicinal macromolecular crystallography since then has been truly remarkable. What should we expect for the next decades?

In the realm of safe predictions we can expect the following:

What if we dream beyond the obvious? One day, medicinal crystallography may contribute to:

One day, crystallography will have revealed the structure of hundreds of thousands of proteins and nucleic acids from human and pathogen, and their complexes with each other and with natural and designed low-molecular-weight ligands. This will form an extraordinarily precious database of knowledge for furthering the health of humans. Hence, in the course of the 21st century, crystallography is likely to become a major driving force for improving health care and disease prevention, and will find a well deserved place in future books describing progress in medicine, sometimes called `The Greatest Benefit to Mankind' (Porter, 1999).