TIBTECH - FEBRUARY 1991 [Vot. 9]
39
to try to build enzymes that are capable of site-specificallycleaving proteins to derive what would be the protein-chemists' equivalent of restriction enzymes! Paul {University of Nebraska} reported that catalytic polyclonal antibodies capable of cleaving vasointestinalprotein {VIP} can be elicited naturally5. This observation, however, remains un~-elimination, cycloaddition and confirmed and, until a similarly retrodimerizationI. Those catalytic active monoclonal antibody isderived antibodies studied closely resemble and the genes that encode it are enzymes mechanistically in their cloned and expressed, willremain so. primary mode of action, since the binding event gives rise to an Protein engineering antibody-substrate complex, which The modification of antibodies by dissociates into products following protein engineering is just one appliMichaelis-Menten kinetics. cation of the knowledge now availFor some reactions {e.g.hydrolysis able on antibody structure. The threeof p-acetamidophenyl esters},a rate dimensional structure of several accelerationof about 10 6 {inthe range antibody-antigen complexes have of esterolytic enzymes}, has been now been solved by X-ray crystalobserved 3, allowing the use of stop- lography6. This is a relatively fast flow techniques to analyse the re- technique when using molecular action in some detail {Benkovic). replacement of highly homologous Rate accelerations of this magnitude structures to aid resolution. Wilson are the exception rather than the rule {Scripps Clinic} elaborated some for the present generation of catalytic possible conclusions from these antibodies and there are, not sur- studies which were supported by prisingly, several approaches being nuclear magnetic resonance (NMR) followed to try to introduce catalytic studies (Arata, University of Tokyo} groups into existing antibodies in and modelling {Thornton, University order to speed things up. Shokat College London, UK). The emerging {University of California, Berkeley}, consensus is that the hypervariable presented the concept of using a CDR (complementarity-determining charged transition state to select an region, see Glossary} loops of both antibody with a complementary heavy and light chains are quite well charge in the binding pocket, in order conserved structurally, and contain to generate antibodies capable of important canonical residues which acid-base catalysis. It is still unclear, determine the topography of the however, how to introduce groups chain. The structure of the loops can which, at the active site, will generate be predicted quite accurately, with hydroxyl groups from H20, to pro- notable exceptions {such as the CDR3 vide the nucleophile often used by of the heavy chain, which is long and existing enzymes. the sequence of which is subject to Lerner {Scripps Clinic}, described the extra variation brought in by his group's recent w o r k 4 where somatic mutation during VDJ joining, the zinc-binding domain from car- see Glossary). bonic anhydrase was incorporated On binding antigen {whether a at the bottom of the binding pocket small hapten or the epitope of a of an antibody, by substituting polypeptide}, the loops do not alter residues in the light chain with those conformation very much, most of the capable of co-ordinating a zinc atom. movement being restricted to CDR3 This represents a first step towards of the heavy chain. It is also quite generating antibodies that include a clear that antibody-combining sites catalytic metal ion, and provides a are not exactly analogous to enzyme good illustration of deterministic active sites, and can be fiat binding ways of altering antibodies using p l a n e s - interactions occurring over a protein engineering. The approach wide surface of both antigen and may be applicable generally for antibody - or a hydrophobic cavity designing cofactor-binding sites. In into which a hapten, for example, can this way, there is the intention fit.For differentantibodies, binding
Catalytic antibodiessomething for everyone Catalytic antibodies have commanded much interest, and recent reviews on the topic are already available1'2. The intention in this workshop article is to capture the flavour of a recent CIBA symposium on catalytic antibodies*, to highlight the importance of the interaction between the enzymologists, protein chemists, immunologists, chemists and molecular biologists in this field, and to point out some of the directions the research is taking and the potential applications of catalytic antibodies. Organic reactions For the enzymologist, catalytic antibodies offer the opportunity to investigate the mechanism of organic reactions not normally catalysed by enzymes, and to examine novel mechanisms for biological reactions. As Benkovic (University of Pennsylvania), pointed out, the science is predicated on our understanding of organic reaction mechanism and transition-state theory, as put forward originally by Pauling and Haldane. This is coupled with the technical ability of being able to derive high-affinity monoclonal antibodies to small hapten organic molecules linked to keyhole limpet haemocyanin by hybridoma technology. The first reports of catalytic antibodies in 1986 involved the use of transition-state analogues to elicit antibodies which catalysed the hydrolysis of esters and carbonetes by transition-state stabilization. This was subsequently extended to include a number of important reaction types such as lactonization,
*The CIBA Foundation hosted a meeting 'Catalytic Antibodies' on 1-3 October, ~ 0 in London, UK. This symposium be published as a CIBA Foundation
~ k , Catalytic Antibodies - CIBA Foun~tion Symposium No. 159.
01~1, ElsevierScience Publishers Ltd (UK) 0167- 9430/91/$2.00
40
TIBTECH- FEBRUARY 1991 [Vol. 9]
~Glossary
sites are provided by amino acids in all or a subset of CDRs and by framework residues, the heavy-chain CDRs tending to contribute most. A more dynamic view of antigenantibody interaction can be obtained by NMR study of ~3C- or 1SN-labelled heavy and light chains (Arata).
Combinatorial associations The ability to make combinatorial libraries in lambda phage 7 has, in theory, made it possible to express the antibody repertoire directly in E. coli (Huse, IXSYS, San Diego). This system provides an alternative to selecting monoclonal antibodies that bind hapten from hybridoma supernatants. Instead, phage libraries are made by co-cloning polymerase chain reaction (PCR)-amplified copies of immunoglobulin heavy- and lightchain mRNAs isolated from spleen cells. These are then screened with radioactive hapten and the cDNAs that encode the combining sites constituting the expressed Fab fragment are identified. Expression of Fabs in this way, however, is not always reliable and other ways of expressing fragments of heavy and light chains in phage (e.g. on the surface of a filamentous phage as recently described for peptide librariesS-l°), would seem to be in order to facilitate screening. There is also an interesting debate going on concerning what the starting point for generating combinatorial libraries should be. The complete antibody repertoire clearly will not be represented if mRNA from primed B cells is used as a source of mRNA, because a good deal of selection has already occurredlL The alternative to taking mRNA from unprimed animals is complicated by the inherent sequence heterogeneity which may not allow amplification of all relevant mRNAs by the PCR primers. Pluckthun (Martinsreid, Germany) discussed the difficulty of expressing antibodies from phage vectors. This is overcome once the genes are isolated, by utilizing the sophisticated plasmids available for E. coli expression 12. Various ways of making Fabs and stable Fvs (e.g. by introduction of intervariable domain disulphide bridges or as single chains), were described and some of these are being used to produce small, catalytically active antibody fragments.
Combinatorial associations - the association of heavy chains, of any type, with light chains, of any type, in a given population of immunoglobulin molecules. CDR - complementarity determining regions: hypervariable regions of both heavy and light chains responsible, primarily, for epitope handling. Fab fragment-the fragment of the immunoglobulin molecule IgG obtained by treating the molecule with the enzyme papain. Fab consists of an intact light chain and the Fd fragment of one heavy chain, held together by a disulphide bond. (Two Fab fragments are obtained per IgG molecule, and each fragment contains one antigen-binding site.) Fd fragment - obtained by reduction of Fab, comprising that portion of the heavy chain that is joined to an intact light chain in the Fab. Fv fragment - the N-terminal portion of the Fab fragment, comprising the variable portions of one heavy- and one light chain. VDJ - variable diversity joining: recombination of three genetic elements forming CDR3 of the heavy chain.
The availability of these rapid techniques for expressing antibodies makes the site-directed approach to designing catalytically active molecules de novo, or by modification of an existing antibody, relatively simple although the outcome is anything but predictable. For example, amino acid residues in the combining site can be changed for amino acids potentially able to participate in catalysis (e.g. His, Ser or Cys) 13. The alternative approach of random mutagenesis and selection is also being pursued to increase the activity of an existing antibody. Hilvert (Scripps Clinic) described elegant experiments towards this end where active mutated antibodies with enhanced chorismate mutase activity are being used to complement yeast Aro 7 mutants.
Potential applications The industrial applications of catalytic antibodies are still some way off, though clear directions are emerging. Several activities of potential interest were described by Schultz (University of California, Berkeley) including an antibody that catalyses the cis-trans isomerization of a diarylenone, a 'Diels-Alderase' antibody generated against a bicycloctane hapten 14, and antibodies that carry out selective metallations 15 and stereoselective hydroxylations. Given the current interest and importance of chiral synthesis in organic chemistry, there is no doubt that catalytic antibodies now have to be considered, for this purpose, along with classical biotransformations using microbial enzymes. The overriding problem which is being addressed is that although catalytic antibodies work by reducing the free
energy of the transition state, they do not (in contrast to most enzymes~ release product very well. This was clearly demonstrated by Martin (IGEN, Maryland) in studies of phenylacetate hydrolysis, but the problem is found almost universally for all catalytic antibodies. Judicious choice of transition-state analogues (Blackburn, University of Sheffield, UK) may help to select binding sites from which reaction products will leave at a reasonable rate. Nevertheless, the time and chemistry involved in synthesising these transition-statemimics may not be trivial. Judgements will have to be made about the ease of synthesis and the raising of the catalytic antibodies, versus finding a suitable microbial enzyme or an alternative chemical route to the desired product. As indicated at the meeting by Lerner and others, however, consideration of application is presently premature; there are still too many problems to solve and, more importantly, fundamental insights still to be gained. There is little doubt that the combination of several different strategies will lead to the generation of highly active catalytic antibodies. This will require the interaction of expert enzymologists, chemists, molecular biologists and immunologists. Just the kind of people in fact who were present at the meeting and who are leading the work. Geographically speaking, it might have been more appropriate to hold the symposium in California, USA, instead of London, UK, as several of the prime movers in this technology (many of whom were speakers at the symposium) either work at,
TIBTECH - F E B R U A R Y 1991 [Vol. 9]
or originate from, the Scripps Clinic in La Jolla, or the University of California campus at Berkeley.
Acknowledgements ! would like to thank Dr A. Baxter, Dr C. Stylli (Glaxo) and Dr A. Mountain (Celltech) for comments on this article.
References 1 Shokat, K. M. and Schultz, P. G. (1990) Annu. Bey. Immunol. 8, 335-363 2 Blackburn, G. M., Kang, A. S., Kingsbury, G.A. and Burton, D.R. (1989) Biochem. J. 262, 381-390 3 Tramontano, A., Ammann, A. A. and Lerner, R. A. (1988) J. Am. Chem. Soc. 110, 2282-2286
41
4 Roberts, V. A., Iverson, B. L., Iverson, S.A., Benkovic, S.J., Lerner, R.A., Getzoff, E. and Tainer, J.A. (1990) Proc. Natl Acad. Sci. USA 87, 66546658 5 Paul, S., Voile, D. J., Beach, C. M., Johnson, D.R., Powell, M.J. and Massey, R.J. (1989) Science 244, 1158-1162 6 Stanfield, R. L., Fieser, T. M., Lamer, R. A. and Wilson, I. A. (1990) Science 248, 712-719 7 Huse, W. D., Sastry, L., Iverson, S. A., Kang, A. S., Alting-Mees, M., Burton, D. R., Benkovic, S. J. and Lerner, R. A. (1989) Science 246, 1275-1281 8 Cwirla, S. E., Peters, E. A., Barrett, R.W. and Power, W.J. (1990) Proc. Natl Acad. Sci. USA 87, 6378-6382 9 Devlin, J. J., Panganiban, L. C. and Devlin, P.R. (1990) Science 249,
Reconstituted viral envelopes 'Trojan Horses" for drug delivery and gene therapy? Robert Blumenthal and Abraham Loyter Reconstituted viral envelopes (RVEs) are formed by solubilizing intact virus in detergent and reassembling the envelope on removal of detergent. RVEs can be formed in the presence of ilgents that become encapsulated and can then be utilized in vitro and in vivo for drug delivery, cell destruction, transfer of membrane components, and as vectors for genetic engineering. The problems with biotechnological applications of RVEs and possible strategies for overcoming them are discussed in this article. A major problem in delivery of drugs and other macromolecules into cells is crossing the permeability barriers imposed by plasma membranes. Enveloped viruses have evolved a 'Trojan Horse' strategy to circumvent those barriers. Such viruses inject their genetic material into the cytoR. Blumenthal is at the Section on Membrane Structure and Function, LMMB, NCI, National Institutes of Health, Bethesda, MD 20892, USA. A. Loyter is at the Department of Biological Chemistry, Hebrew University, Jerusalem, Israel.
plasm of the host cell after fusion between the viral and cellular membranes 1. Biotechnologists have exploited this characteristic to develop strategies for in vitro and in vivo use of the viral envelope as a 'Trojan Horse'. This involves disassembly of the virus by detergent, followed by reassembly of its envelope to package the materials to be delivered into cells 2.
Viral entry Attachment of the viral particles to the host cell surface is mediated by the viral glycoproteins that consti-
© 1991,Elsevier Science Publishers Ltd (UK) 0167 - 9430/91152.00
40~ 'I.06 10 Scott, J. K. and Smith, G. P. (1990) Science 249,385-390 11 Caton, A. J. and Koprowski, H. (1990) Proc. Natl Acad. Sci. USA 87, 64506454 12 Skerra, A. and Pluckthun, A. (1988) Science 240, 1038-1041 13 Pollack, S, I. and Schultz, P. G. (1989) J. Am. Chem. Soc. 111, 1929-1931 14 Braisted, A. C. and Schultz, P. G. (1990) J. Am. Chem. Soc. 112, 74307431 15 Cochran, A. and Schultz, P. G. (1990) Science 249, 781-783 T. J. R. HARRIS
Glaxo Group Research Ltd, Greenford Road, Greenford, M i d d l e s e x U B 6 0 H E , UK.
tute the spikes in the viral envelope. The viral-spike glycoproteins recognize cellular receptors for the virus present in the plasma membrane of the target cell1; the specificity of this recognition process depends on the molecular structure of those viral glycoproteins and their corresponding receptors. The absence of the viral receptors on a specific cell confers resistance to viral infection. Significant advances have been made in recent years in terms of the knowledge available on the sequences and structure of viralenvelope proteins, and genetic and chemical methods have been developed for site-specific sequence alterations 3. Little is known, however, about molecular events involved in viral-envelope protein-mediated membrane fusion 4. Assays for studying the kinetics of fusion of fluorescently-labeled virus with a variety of target membranes have been developed using spectrofluorometric 5 and video microscopic techniques 6 (see Fig. 1). These have enabled fusion of viral envelopes with cell membranes to be studied for a variety of different virus strains including those that enter cells by acid-activated fusion following endocytosis (e.g. influenza virus and vesicular stomatitis virus), and those that fuse directly with the plasma membrane at neutral pH (e.g. Sendai virus, human immunodeficiency virus, vaccinia virus, and Epstein-Barr virus). It is possible to monitor fluorescence changes continuously
39
to try to build enzymes that are capable of site-specificallycleaving proteins to derive what would be the protein-chemists' equivalent of restriction enzymes! Paul {University of Nebraska} reported that catalytic polyclonal antibodies capable of cleaving vasointestinalprotein {VIP} can be elicited naturally5. This observation, however, remains un~-elimination, cycloaddition and confirmed and, until a similarly retrodimerizationI. Those catalytic active monoclonal antibody isderived antibodies studied closely resemble and the genes that encode it are enzymes mechanistically in their cloned and expressed, willremain so. primary mode of action, since the binding event gives rise to an Protein engineering antibody-substrate complex, which The modification of antibodies by dissociates into products following protein engineering is just one appliMichaelis-Menten kinetics. cation of the knowledge now availFor some reactions {e.g.hydrolysis able on antibody structure. The threeof p-acetamidophenyl esters},a rate dimensional structure of several accelerationof about 10 6 {inthe range antibody-antigen complexes have of esterolytic enzymes}, has been now been solved by X-ray crystalobserved 3, allowing the use of stop- lography6. This is a relatively fast flow techniques to analyse the re- technique when using molecular action in some detail {Benkovic). replacement of highly homologous Rate accelerations of this magnitude structures to aid resolution. Wilson are the exception rather than the rule {Scripps Clinic} elaborated some for the present generation of catalytic possible conclusions from these antibodies and there are, not sur- studies which were supported by prisingly, several approaches being nuclear magnetic resonance (NMR) followed to try to introduce catalytic studies (Arata, University of Tokyo} groups into existing antibodies in and modelling {Thornton, University order to speed things up. Shokat College London, UK). The emerging {University of California, Berkeley}, consensus is that the hypervariable presented the concept of using a CDR (complementarity-determining charged transition state to select an region, see Glossary} loops of both antibody with a complementary heavy and light chains are quite well charge in the binding pocket, in order conserved structurally, and contain to generate antibodies capable of important canonical residues which acid-base catalysis. It is still unclear, determine the topography of the however, how to introduce groups chain. The structure of the loops can which, at the active site, will generate be predicted quite accurately, with hydroxyl groups from H20, to pro- notable exceptions {such as the CDR3 vide the nucleophile often used by of the heavy chain, which is long and existing enzymes. the sequence of which is subject to Lerner {Scripps Clinic}, described the extra variation brought in by his group's recent w o r k 4 where somatic mutation during VDJ joining, the zinc-binding domain from car- see Glossary). bonic anhydrase was incorporated On binding antigen {whether a at the bottom of the binding pocket small hapten or the epitope of a of an antibody, by substituting polypeptide}, the loops do not alter residues in the light chain with those conformation very much, most of the capable of co-ordinating a zinc atom. movement being restricted to CDR3 This represents a first step towards of the heavy chain. It is also quite generating antibodies that include a clear that antibody-combining sites catalytic metal ion, and provides a are not exactly analogous to enzyme good illustration of deterministic active sites, and can be fiat binding ways of altering antibodies using p l a n e s - interactions occurring over a protein engineering. The approach wide surface of both antigen and may be applicable generally for antibody - or a hydrophobic cavity designing cofactor-binding sites. In into which a hapten, for example, can this way, there is the intention fit.For differentantibodies, binding
Catalytic antibodiessomething for everyone Catalytic antibodies have commanded much interest, and recent reviews on the topic are already available1'2. The intention in this workshop article is to capture the flavour of a recent CIBA symposium on catalytic antibodies*, to highlight the importance of the interaction between the enzymologists, protein chemists, immunologists, chemists and molecular biologists in this field, and to point out some of the directions the research is taking and the potential applications of catalytic antibodies. Organic reactions For the enzymologist, catalytic antibodies offer the opportunity to investigate the mechanism of organic reactions not normally catalysed by enzymes, and to examine novel mechanisms for biological reactions. As Benkovic (University of Pennsylvania), pointed out, the science is predicated on our understanding of organic reaction mechanism and transition-state theory, as put forward originally by Pauling and Haldane. This is coupled with the technical ability of being able to derive high-affinity monoclonal antibodies to small hapten organic molecules linked to keyhole limpet haemocyanin by hybridoma technology. The first reports of catalytic antibodies in 1986 involved the use of transition-state analogues to elicit antibodies which catalysed the hydrolysis of esters and carbonetes by transition-state stabilization. This was subsequently extended to include a number of important reaction types such as lactonization,
*The CIBA Foundation hosted a meeting 'Catalytic Antibodies' on 1-3 October, ~ 0 in London, UK. This symposium be published as a CIBA Foundation
~ k , Catalytic Antibodies - CIBA Foun~tion Symposium No. 159.
01~1, ElsevierScience Publishers Ltd (UK) 0167- 9430/91/$2.00
40
TIBTECH- FEBRUARY 1991 [Vol. 9]
~Glossary
sites are provided by amino acids in all or a subset of CDRs and by framework residues, the heavy-chain CDRs tending to contribute most. A more dynamic view of antigenantibody interaction can be obtained by NMR study of ~3C- or 1SN-labelled heavy and light chains (Arata).
Combinatorial associations The ability to make combinatorial libraries in lambda phage 7 has, in theory, made it possible to express the antibody repertoire directly in E. coli (Huse, IXSYS, San Diego). This system provides an alternative to selecting monoclonal antibodies that bind hapten from hybridoma supernatants. Instead, phage libraries are made by co-cloning polymerase chain reaction (PCR)-amplified copies of immunoglobulin heavy- and lightchain mRNAs isolated from spleen cells. These are then screened with radioactive hapten and the cDNAs that encode the combining sites constituting the expressed Fab fragment are identified. Expression of Fabs in this way, however, is not always reliable and other ways of expressing fragments of heavy and light chains in phage (e.g. on the surface of a filamentous phage as recently described for peptide librariesS-l°), would seem to be in order to facilitate screening. There is also an interesting debate going on concerning what the starting point for generating combinatorial libraries should be. The complete antibody repertoire clearly will not be represented if mRNA from primed B cells is used as a source of mRNA, because a good deal of selection has already occurredlL The alternative to taking mRNA from unprimed animals is complicated by the inherent sequence heterogeneity which may not allow amplification of all relevant mRNAs by the PCR primers. Pluckthun (Martinsreid, Germany) discussed the difficulty of expressing antibodies from phage vectors. This is overcome once the genes are isolated, by utilizing the sophisticated plasmids available for E. coli expression 12. Various ways of making Fabs and stable Fvs (e.g. by introduction of intervariable domain disulphide bridges or as single chains), were described and some of these are being used to produce small, catalytically active antibody fragments.
Combinatorial associations - the association of heavy chains, of any type, with light chains, of any type, in a given population of immunoglobulin molecules. CDR - complementarity determining regions: hypervariable regions of both heavy and light chains responsible, primarily, for epitope handling. Fab fragment-the fragment of the immunoglobulin molecule IgG obtained by treating the molecule with the enzyme papain. Fab consists of an intact light chain and the Fd fragment of one heavy chain, held together by a disulphide bond. (Two Fab fragments are obtained per IgG molecule, and each fragment contains one antigen-binding site.) Fd fragment - obtained by reduction of Fab, comprising that portion of the heavy chain that is joined to an intact light chain in the Fab. Fv fragment - the N-terminal portion of the Fab fragment, comprising the variable portions of one heavy- and one light chain. VDJ - variable diversity joining: recombination of three genetic elements forming CDR3 of the heavy chain.
The availability of these rapid techniques for expressing antibodies makes the site-directed approach to designing catalytically active molecules de novo, or by modification of an existing antibody, relatively simple although the outcome is anything but predictable. For example, amino acid residues in the combining site can be changed for amino acids potentially able to participate in catalysis (e.g. His, Ser or Cys) 13. The alternative approach of random mutagenesis and selection is also being pursued to increase the activity of an existing antibody. Hilvert (Scripps Clinic) described elegant experiments towards this end where active mutated antibodies with enhanced chorismate mutase activity are being used to complement yeast Aro 7 mutants.
Potential applications The industrial applications of catalytic antibodies are still some way off, though clear directions are emerging. Several activities of potential interest were described by Schultz (University of California, Berkeley) including an antibody that catalyses the cis-trans isomerization of a diarylenone, a 'Diels-Alderase' antibody generated against a bicycloctane hapten 14, and antibodies that carry out selective metallations 15 and stereoselective hydroxylations. Given the current interest and importance of chiral synthesis in organic chemistry, there is no doubt that catalytic antibodies now have to be considered, for this purpose, along with classical biotransformations using microbial enzymes. The overriding problem which is being addressed is that although catalytic antibodies work by reducing the free
energy of the transition state, they do not (in contrast to most enzymes~ release product very well. This was clearly demonstrated by Martin (IGEN, Maryland) in studies of phenylacetate hydrolysis, but the problem is found almost universally for all catalytic antibodies. Judicious choice of transition-state analogues (Blackburn, University of Sheffield, UK) may help to select binding sites from which reaction products will leave at a reasonable rate. Nevertheless, the time and chemistry involved in synthesising these transition-statemimics may not be trivial. Judgements will have to be made about the ease of synthesis and the raising of the catalytic antibodies, versus finding a suitable microbial enzyme or an alternative chemical route to the desired product. As indicated at the meeting by Lerner and others, however, consideration of application is presently premature; there are still too many problems to solve and, more importantly, fundamental insights still to be gained. There is little doubt that the combination of several different strategies will lead to the generation of highly active catalytic antibodies. This will require the interaction of expert enzymologists, chemists, molecular biologists and immunologists. Just the kind of people in fact who were present at the meeting and who are leading the work. Geographically speaking, it might have been more appropriate to hold the symposium in California, USA, instead of London, UK, as several of the prime movers in this technology (many of whom were speakers at the symposium) either work at,
TIBTECH - F E B R U A R Y 1991 [Vol. 9]
or originate from, the Scripps Clinic in La Jolla, or the University of California campus at Berkeley.
Acknowledgements ! would like to thank Dr A. Baxter, Dr C. Stylli (Glaxo) and Dr A. Mountain (Celltech) for comments on this article.
References 1 Shokat, K. M. and Schultz, P. G. (1990) Annu. Bey. Immunol. 8, 335-363 2 Blackburn, G. M., Kang, A. S., Kingsbury, G.A. and Burton, D.R. (1989) Biochem. J. 262, 381-390 3 Tramontano, A., Ammann, A. A. and Lerner, R. A. (1988) J. Am. Chem. Soc. 110, 2282-2286
41
4 Roberts, V. A., Iverson, B. L., Iverson, S.A., Benkovic, S.J., Lerner, R.A., Getzoff, E. and Tainer, J.A. (1990) Proc. Natl Acad. Sci. USA 87, 66546658 5 Paul, S., Voile, D. J., Beach, C. M., Johnson, D.R., Powell, M.J. and Massey, R.J. (1989) Science 244, 1158-1162 6 Stanfield, R. L., Fieser, T. M., Lamer, R. A. and Wilson, I. A. (1990) Science 248, 712-719 7 Huse, W. D., Sastry, L., Iverson, S. A., Kang, A. S., Alting-Mees, M., Burton, D. R., Benkovic, S. J. and Lerner, R. A. (1989) Science 246, 1275-1281 8 Cwirla, S. E., Peters, E. A., Barrett, R.W. and Power, W.J. (1990) Proc. Natl Acad. Sci. USA 87, 6378-6382 9 Devlin, J. J., Panganiban, L. C. and Devlin, P.R. (1990) Science 249,
Reconstituted viral envelopes 'Trojan Horses" for drug delivery and gene therapy? Robert Blumenthal and Abraham Loyter Reconstituted viral envelopes (RVEs) are formed by solubilizing intact virus in detergent and reassembling the envelope on removal of detergent. RVEs can be formed in the presence of ilgents that become encapsulated and can then be utilized in vitro and in vivo for drug delivery, cell destruction, transfer of membrane components, and as vectors for genetic engineering. The problems with biotechnological applications of RVEs and possible strategies for overcoming them are discussed in this article. A major problem in delivery of drugs and other macromolecules into cells is crossing the permeability barriers imposed by plasma membranes. Enveloped viruses have evolved a 'Trojan Horse' strategy to circumvent those barriers. Such viruses inject their genetic material into the cytoR. Blumenthal is at the Section on Membrane Structure and Function, LMMB, NCI, National Institutes of Health, Bethesda, MD 20892, USA. A. Loyter is at the Department of Biological Chemistry, Hebrew University, Jerusalem, Israel.
plasm of the host cell after fusion between the viral and cellular membranes 1. Biotechnologists have exploited this characteristic to develop strategies for in vitro and in vivo use of the viral envelope as a 'Trojan Horse'. This involves disassembly of the virus by detergent, followed by reassembly of its envelope to package the materials to be delivered into cells 2.
Viral entry Attachment of the viral particles to the host cell surface is mediated by the viral glycoproteins that consti-
© 1991,Elsevier Science Publishers Ltd (UK) 0167 - 9430/91152.00
40~ 'I.06 10 Scott, J. K. and Smith, G. P. (1990) Science 249,385-390 11 Caton, A. J. and Koprowski, H. (1990) Proc. Natl Acad. Sci. USA 87, 64506454 12 Skerra, A. and Pluckthun, A. (1988) Science 240, 1038-1041 13 Pollack, S, I. and Schultz, P. G. (1989) J. Am. Chem. Soc. 111, 1929-1931 14 Braisted, A. C. and Schultz, P. G. (1990) J. Am. Chem. Soc. 112, 74307431 15 Cochran, A. and Schultz, P. G. (1990) Science 249, 781-783 T. J. R. HARRIS
Glaxo Group Research Ltd, Greenford Road, Greenford, M i d d l e s e x U B 6 0 H E , UK.
tute the spikes in the viral envelope. The viral-spike glycoproteins recognize cellular receptors for the virus present in the plasma membrane of the target cell1; the specificity of this recognition process depends on the molecular structure of those viral glycoproteins and their corresponding receptors. The absence of the viral receptors on a specific cell confers resistance to viral infection. Significant advances have been made in recent years in terms of the knowledge available on the sequences and structure of viralenvelope proteins, and genetic and chemical methods have been developed for site-specific sequence alterations 3. Little is known, however, about molecular events involved in viral-envelope protein-mediated membrane fusion 4. Assays for studying the kinetics of fusion of fluorescently-labeled virus with a variety of target membranes have been developed using spectrofluorometric 5 and video microscopic techniques 6 (see Fig. 1). These have enabled fusion of viral envelopes with cell membranes to be studied for a variety of different virus strains including those that enter cells by acid-activated fusion following endocytosis (e.g. influenza virus and vesicular stomatitis virus), and those that fuse directly with the plasma membrane at neutral pH (e.g. Sendai virus, human immunodeficiency virus, vaccinia virus, and Epstein-Barr virus). It is possible to monitor fluorescence changes continuously
