Conformational restriction of peptidyl immunogens with covalent replacements for the hydrogen bond Arnold C. Satterthwait*, Thomas Arrhenius*, Robert A. Hagopian*, Fidel Zavala t, Victor Nussenzweig t and Richard A. Lerner* A new strategy for designing synthetic vaccines is presented. In this approach synthetic peptides are conformationally restricted by replacing putative hydrogen bonds with covalent mimics. The chemistry for substituting a hydrazone-ethane link (N-N=CH-CH2-CHz) for an ~ + 4) --* i hydrogen bond in a pentapeptide with m-helical potential is reported. Chemically shaping peptides to mimic the three-dimensional surfaces of proteins may enhance their immunogenicity. To test this strategy, a potential synthetic vaccine for malaria, Cys-(Asn-Pro-Asn-Ala)x--NHz, was conforma- tionally restricted by replacing putative hydrogen bonds between asparagine side chains with a covalent replacement, an ethylene bridge, to give first generation chemically shaped immunogens. Antibodies to one of the shaped malarial peptides show a strong reaction with living Piasmodium falciparum sporozoites, a form of malaria which infects hundreds of millions of people yearly. Keywords: Malaria; peptides; conformation; mimics; a-helix; reverse turn Introduction Antlpeptide Peptides Antibodies Protein Ab-Protein The effectiveness of synthetic peptide vaccines is dependent on the surprising if not improbable reaction of antipeptide antibodies with proteins. This reaction is surprising since short peptides under twenty amino acids in length are largely disordered structures in water ~-5. Yet antibodies to these disordered peptides, with binding pockets which presumably mirror this disorder, are in many instances capable of reacting with the ordered surfaces of proteins ° . The facts are that some antipeptide antibodies formed against surface amino acid sequences do not bind to proteins 7, many show some affinity for proteins and some bind quite well6'8. This puzzling result has been referred to as the disorder-order paradox 1°. The resolution of this paradox has implications for vaccine design. Currently, two explanations are under consideration. First, contrary to established dogma, short peptides may form small populations of ordered structures in water. In support of this are recent exper- iments which show that a few short peptides are capable of forming helices 1'11-14 and reverse turns 15'16. Small populations of ordered or partially ordered peptides could stimulate the formation of neutralizing antibodies (Figure 1). The effectiveness of short peptides as immu- nogens would as a first approximation be proportional to the fraction of peptide which mimics the ordered structure of the cognate sequence in the protein (Equa- tion 1, Figure 1). For most cases this mechanism pre- dicts that the immune response to peptides will fall short of the immune response to native proteins. A second rationale for the disorder-order paradox *Research Institute of Scripps Clinic, La Jolla, CA, USA. tNew York University Medical Center, New York, NY, USA. eq 1 Kob s = fKideal Figure 1 The immune response to partially ordered peptides ( ~ ), yields antipeptide antibodies of which a fraction binds to the ordered cognate sequence in the protein considers the possibility that surfaces of proteins may prove more flexible than anticipated t7'18. The binding of an antipeptide antibody to a flexible surface of a protein can be described in terms of two extremes of the probable binding pathway (Figure 2). At one extreme, the antipeptide antibody binds to the native three- dimensional structure of the cognate sequence (K1) in some manner in a weak reaction to yield an initial complex that undergoes a rearrangement of interactive surfaces (Ku,) to maximize the overall binding affinity (Kobs). Alternatively, at the other extreme the cognate sequence in the protein unfolds (Ku) to give a locally disordered structure which then combines tightly (K2) with the corresponding antipeptide antibody. Both pathways which lead to the same final complex are thermodynamic equivalents (Equation 2, Figure 2) of the probable pathway which may proceed by an inter- mediate route. However, the relationship between pro- 0264-410X/88/020099-05 $03.00 ~) 1988 Butterworlh& Co. (Publishers) Ltd. Vaccine, Vol. 6, April 1988 99