The Problem of Amino Acid Complementarity and Antisense Peptides

Current Protein and Peptide Science, 2004, 5, 507-527

507

The Problem of Amino Acid Complementarity and Antisense Peptides Ignacy Z. Siemion*, Marek Cebrat and Alicja Kluczyk Faculty of Chemistry, University of Wroclaw, Wroclaw 50-383, Poland Abstract: The review presents three hypotheses concerning the amino acid complementarity: 1) the Mekler-Blalock antisense hypothesis; 2) the Root-Bernstein approach based on stereochemical complementarity of amino acids and antiamino acids coded by anticodons read in parallel with the coding DNA strand; 3) Siemion hypothesis resulting from the periodicity of the genetic code. The current state of knowledge as well as the results of the implementations of these hypotheses are compared. A special attention is given to Root-Bernstein and Siemion hypotheses, which differ in only few points of the complementarity prediction. We describe methods of investigation of peptide - antipeptide pairing, including circular dichroism, mass spectrometry, affinity chromatography and other techniques. The biological applications of complementarity principle are considered, such as search for bioeffector – bioreceptor interaction systems, the influence of peptide – antipeptide pairing on the activity of peptide hormones, and the application of antipeptides in immunochemistry. The possible role of amino acid – anti-amino acid interactions in the formation of the spatial structures of peptides, proteins and protein complexes is discussed. Such problems as the pairing preferences of protein – protein interfaces, the role of the pairing in the creation of disulfide bonds and the possible appearance of such interactions in β-structure are also examined. The main intention of the paper is to bring the complementarity problem to the attention of the scientific community, as a possible tool in proteomics, molecular design and molecular recognition.

Keywords: complementary peptides, antisense peptides, antipeptides, antiidiotypic antibodies, hydropathic complementarity INTRODUCTION The problem of amino acid complementarity is intensively discussed in the scientific literature since 1980ties. The existence of such complementarity could be very important for the protein folding and protein-protein interactions, including molecular recognition between bioeffectors and the correspondent bioreceptors, immunoreactions, etc. The material presented in our review is organized in five chapters. In the first chapter the three hypotheses connected to amino acid complementarity problem are presented. All three hypotheses are considered in the review as equally valuable. It is impossible at the moment to decide which of the three approaches prevails over the other two in the usefulness and phenomenological justification. In the second chapter the possible role of amino acid pairing in the protein folding and protein tertiary structure determination is discussed. The material connected to this problem, presented in the literature, is limited, but it seems necessary to bring this important question to the attention of the readers. *Address correspondence to this author at the Faculty of Chemistry, University of Wroclaw, F. Joliot-Curie 14, Wroclaw 50-383, Poland; Tel: +48-71-3757206; Fax: +48-71-3282348; E-mail: [email protected]. uni.wroc.pl

1389-2037/04 $45.00+.00

The third chapter presents the data on peptide-antipeptide complexation and the methods of monitoring the complexation processes. The last two chapters are devoted to the practical usefulness of such complexes, with the chapter five focused on the complementarity phenomena in immunochemistry and receptor studies. I. THREE APPROACHES TO THE AMINO ACID COMPLEMENTARITY PROBLEM At present, there are three approaches to the problem of amino acid complementarity: i. the most developed approach of Blalock, based on the hydropathic complementarity principle, ii. Root-Bernstein's approach, funded upon a definite stereochemical model, iii. Siemion's approach, resulting from the periodicity of the genetic code. The Blalock's ideas are closely connected to the antisense strategy known in the nucleic acid chemistry. They also resemble earlier suppositions given by Mekler who introduced the idea of the specific interaction of “sense” and “antisense peptides” [1]. He suggested that peptides coded by complementary strand of DNA, read in 5'-3' direction, could interact with peptides resulted from reading of the coding strand. According to Mekler, such a relation could be of importance in the interaction of antigen idiotypes with © 2004 Bentham Science Publishers Ltd.

508

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

antibodies, and similar phenomena. “If the presented postulates are right – wrote Mekler – their consequence should be the existence of a code, which determines the specificity of complementary interactions between amino acid residues in polypeptide chains and, possibly, also between free amino acids”. The Mekler's theory was further elaborated by Idlis in terms of the graph theory [2]. Similar ideas were presented in 1981 by Biró. He developed the computerized step-by-step comparative method of a search for intra- and intermolecular compartments containing the complementary amino acid pairs. Complementary coding was considered by Biró as a background of Informational Complementarity, i.e. the molecular background of highly specific intra- and intermolecular interactions of proteins [3]. Both anti-parallel (5’-3’) and parallel (3’-5’) directions of the reading of complementary DNA strands were considered by Biró, but the main attention was paid to the antiparallel reading frame. A new aspect was added to this concept in the form of “hydropathic complementarity principle” of Blalock and Smith [4]. They pointed out that amino acids coded by the coding strand of DNA are generally complemented by amino acids with opposite hydropathic scores, coded by the DNA complementary strand. According to the Blalock's hypothesis, a hydropathic complementarity is a base for protein-protein and peptide-peptide interactions. Amino acids interact as a pair if they have approximately equal but opposite hydropathic scores. The complementary peptide results from the reading of complementary DNA strand in anti-parallel 5'-3' direction in the same reading frame as the coded peptide. This complementarity idea was tested for the first time by Blalock et al. on corticotropin (ACTH) and γ-endorphin. They found that these peptides interact selectively with their synthetic counterpart antipeptides determined by RNA sequences complementary to the mRNA for ACTH and γendorphin, respectively. They also found that ACTH1-24 antisense peptide induces antibodies specific for the ACTH receptor [5]. However, it should be noted that these observations were very quickly questioned by Eberle et al. [6]. The hydropathic complementarity of sense versus antisense peptides is encoded mainly by the central bases of the correspondent coding triplets, and therefore is should be independent of the direction of the frame reading. Taking this into account, Blalock concluded that the peptide products of complementary nucleic acid strands should bind to each other regardless of the direction of the DNA reading. This conclusion was confirmed by the observation that two complementary peptides that differ in the direction of reading of proper nucleic acid strand, bind to 125I-ACTH with the same affinity [7]. This conclusion might have been influenced by an earlier work of Root-Bernstein, preceding that by four years (see below), in which the idea that the peptides complementary to a given sequence are coded by complementary DNA strand read in 3’-5’ direction, i.e. in the parallel order to the coding strand, was presented. The critical remarks, saying that the interactions between hydrophobic and hydrophilic peptides resulting from

Siemion et al.

hydropathic complementarity hypothesis are rather impossible due to physical reasons, were deflected by Blalock with the assumption that, when in the proper conformation, both peptides can contact each other by conventional hydrophobic-hydrophobic and hydrophilichydrophilic interactions [8]. Such a possibility was also suggested by the model of complementary peptides interaction of Markus et al. [9]. They proposed that in the interacting peptides hydrophilic fragments of both peptide chains are oriented towards the aqueous phase, while the closely packed hydrophobic ones form the interphase between the two chains. They also suggested that the optimal interaction between the two chains, with hydrophobic and hydrophilic residues alternating along the chain, occurs when the chains appear in the extended conformation. It was also shown that the hydropathic complementarity of a given peptide and its 5'-3' counterpart takes place when both peptides are oriented antiparallel one to the other. On the other hand, in the case of 3'5' antipeptide the parallel orientation of both peptides is necessary. The practical recommendations for the design of complementary peptides were given by Bost and Blalock in [10]. They result from theoretical work of Dill et al. stating that the location of hydrophilic (polar) and hydrophobic (nonpolar) residues along the polypeptide chain provides the major force determining the protein molecule folding [11, 12]. Kamtekar et al. qualified this as “the binary code for the protein design” [13]. The “binary code” determines the hydropathy of successive fragments of polypeptide chains of proteins. Basing on this “binary code”, Blalock et al. synthesized a 3'-5' analog of the growth hormone – releasing hormone (GHRH) 1-23 sequence. The peptide was found to be an antagonist of GHRH binding to the GHRH receptors. The authors of that paper concluded that “since the pattern of hydropathy determines gross shape, a complete inversion of the exact pattern as occurs on the noncoding or complementary strand should result in an inverted or complementary shape” [14]. This is an important point of the molecular recognition theory (MRT) developed by Blalock. The theory states that since complementary codons specify hydropathically opposed amino acids and because peptides with opposite hydropathic patterns interact, then complementary nucleic acid sequences encode interacting peptides or proteins. If this is true, then the contact points between two interacting peptides or proteins can be identified through regions of complementarity in their respective nucleotide sequences [8]. An important part of MRT is a thesis on the existence of binary code of polar and nonpolar amino acids within the coded sequence, i.e. of linear array of its hydropathy [15]. As we have noted above, the strategy of antisense peptides, expressed by Blalock, resembles the antisense strategy known in nucleic acid chemistry. The DNA/RNA antisense strategy has recently found important and useful applications (for the review see: [16]). However, the practical application of peptide antisense strategy would probably be of a lesser value than DNA/RNA antisense strategy, because the amino acid pair interactions are much weaker and less specific than nucleic base pairing.

The Problem of Amino Acid Complementarity and Antisense Peptides

It must be noted, however, that several reports have questioned the general applicability of molecular recognition theory and the principle of peptide-antipeptide interaction. We quoted above the work of Eberle et al. which contradicts the results of Blalock et al. on the interaction of ACTH peptide with its antisense complement [6, 17, 18]. However, Clarke and Blalock have pointed out that 3'-5' ACTH complementary peptide binds to the surface of mouse Y-1 adrenal cells and to polyclonal anti-ACTH antibody. It mimics the properties of ACTH by stimulating cAMP synthesis and steroidogenesis in adrenal cells. In 1987 Rasmussen and Hesch showed that antisense peptide directed against parathyroid hormone does not bind to the hormone molecule [19]. Unfortunately, the binding properties of only the 5'-3' complement (Blalock`s assignment) were examined by these authors. Beattie and Flint obtained no evidence of any interaction of insulin-like growth factor-1 (IGF-1) with its 5'-3' antipeptide addressed at 21-40 fragment of IGF-1 [20]. Stefani et al. synthesized three antipeptides anchored to the resin by amide linker and directed against the 43-57 region of acylphosphatase and the 46-60 region of two isoenzymes: phosphotyrosine protein phosphatases AcPA and AcPA 2, respectively. They found that none of the resins was able to bind the corresponding protein. The authors of this report noted, however, that the lack of interaction between the peptides and respective proteins could be due to the steric hindrance or poor solvation of the resin-anchored peptides [21]. The idea that the complementary amino acid pairs may result from parallel reading of complementary DNA strands (i.e. when coding strand is read in 5'-3' direction and complementary strand in 3'-5' direction) was formulated by Root-Bernstein in 1982. He postulated that the resulting two peptides (peptide and antipeptide) may interact in sequencespecific manner involving the formation of parallel beta ribbon which aligns the amino acid side chains. It has been shown by Root-Bernstein that from the set of all 210 possible amino acid pairs no more than 26 fulfill the stereochemical condition indicated above, and 14 out of this number are genetically encoded in the indicated manner. For the Gly-Pro interaction Root-Bernstein postulated that it may be realized in collagen triple helix conformation rather than in the parallel β-structure [22]. He proposed also that the amino acid pairing could have important implications for the understanding of the origins of the genetic code. The fact that amino acid pairing is genetically encoded contradicts the hypothesis that the genetic code evolved as a series of “frozen accidents”, i.e. that the relation between the codons and amino acids they encode was of accidental origin [23]. As a rule, complementary peptides antagonize each other's biological activity. However, in several cases the agonistic activity resulting from the complexation was also noted. This problem will be discussed in the following text. The amino acid pairing hypothesis was applied by RootBernstein to the prediction of the sequence and location of Gly-Pro-Arg-Pro peptide binding site in fibrinogen molecule. It was known that this tetrapeptide inhibits fibrinogen aggregation. Based on the amino acid pairing hypothesis Root-Bernstein and Westall speculated that the possible

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

509

binding site in this protein could be connected to Asp-SerGly-Glu-Gly-Asp sequence, existing in fibrinopeptide A. The experiments performed with synthetic peptides showed that the tetrapeptide binds specifically to fibrinopeptide A with the binding constant of ca. 104 per mol [24]. One of the questions discussed by Root-Bernstein was a possible role of antisense complementarity in the living systems. In the opinion of this author antisense peptides may have evolved as naturally occurring physiological antagonists of the proper peptide bioeffectors or as the modulators of their functions. In particular, Root-Bernstein and Westall showed that bovine pineal anti-reproductive tripeptide Thr-Lys-Ser and luteinizing hormone-releasing hormone (LHRH) are complementary according to RootBernstein's convention. The anti-reproductive tripeptide interacts with the active site Trp-Ser-Tyr of LHRH, modulating its activity [25]. In the recent years Root-Bernstein and Dillon developed a more general molecular complementarity theory, oriented towards the problem of origin and evolution of life. The amino acid complementarity remains a particular question in this broad theory [26]. The question, which of the two presented approaches for the search of complementary peptides (5'-3' approach, or 3'5' approach) is the right one, is the question of controversy as yet. There are, however, some observations, which suggest that 3'-5' approach leads to more interesting results in terms of biological activity of the respective complementary peptides. For example, of the two peptides resulting from the reading the DNA strand complementary to that encoding angiotensin II in parallel and anti-parallel direction, respectively, only the product of parallel reading is a true angiotensin II antipeptide, whereas the Blalock's peptide is rather angiotensin homologue, working as the receptor antagonist [27]. In this particular case, however, the sequence of the “Blalock peptide” is more than 80% similar to angiotensin II (Glu-Gly-Val-Tyr-Val-His-Pro-Val versus Asp-Arg-Val-Tyr-Ile-His-Pro-Phe). Thus, in this case the hydropathic complementarity principle is only partially realized for the “Blalock peptide”. An independent approach to the amino acid complementarity problem resulted from investigations on the genetic code periodicity. In 1992 Siemion and Stefanowicz showed that the chemical reactivity of amino acid Nhydroxysuccinimide esters changes periodically within the genetic code. In order to see it the codons must be arranged in a closed ring in which the consecutive codons are connected by a regular series of one-step mutations of a type: 2, 3 3 3, 1, 3 3 3, 1, 3 3 3, 1, 3 3 3, 2, 3 3 3 ... (the numbers here denote a codon position in which a mutation takes place) [28]. The amino acid order resulting from the circular arrangement of their codons appears to be very similar to the “amino acid similarity ring” obtained by Argyle from the statistical analysis of the most frequent replacements of amino acid residues during the evolution of proteins [29]. Even more pronounced similarity occurs between Siemion's arrangement and the “amino acid ring” constructed by Pieber and Toha on the basis of codon replacement probability matrix [30]. It was also found that

510

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

Siemion et al.

Fig. (1). The folded form of the one-step mutation ring of the genetic code.

the amino acid compositional frequencies in proteins [31], as well as the Chou-Fasman conformational parameters of amino acids [32] change regularly within the genetic code ring, arranged in the manner described above. The periodical changes of Pα conformational parameters (the probability coefficients for the given amino acid to adopt the α-helical conformation) could be even described by a simple trigonometric equation [33]. These investigations were reviewed by Siemion [34]. The results were also confirmed by the algebraic analysis performed by Bashford and Jarvis [35]. The Fig. (1) presents the folded form of the one-step mutation ring of the genetic code. It can be seen that the so called equivalent codons, belonging to A and U, as well as C and G families of codons, respectively, appear in the same order within the corresponding rows or columns of the ring. The codon families are defined by the central bases of triplets. In the equivalent codons the first two bases of triplets are complementary, whereas the third is exactly the same. It was hypothesized that the respective equivalent codons correspond to amino acid – anti-amino acid pairs. Because of the lack of complementarity in the third position of the codons, the peptides designed according to this theory cannot be described as “antisense peptides”. However, in this case, the use of the term “complementary peptides” seems justified. Thus, according to this approach, the complementarity of amino acids is a consequence of the periodicity of the genetic code. The modes of amino acid pairing resulting from the three approaches presented above are compared in Table 1. It can be seen that the 3'-5' reading of complementary DNA strand strongly reduces the impact of the degeneracy of the genetic code on the number of amino acid complements. There are only minor differences in the assignments of “anti-amino acids” according to Root-Bernstein and Siemion approaches. Thus, according to Root-Bernstein, Asn forms a pair with Leu, Lys with Phe, and Tyr with Ile and Met. In the Siemion's prediction Asn interacts with Phe, and Lys with Leu, whereas the Met codon corresponds to one of the three

“stop” codons. (This last result seems especially interesting, as the Met codon which initiates the mRNA translation corresponds here to the codon of its termination). In this context it is of interest that according to the results of Chou and Fasman [36] Leu indeed can interact spontaneously in aqueous solution with Lys and Glu. The stabilizing ∆G value was found to be -1.0 kcal/mol in the case of Leu-Lys system (Siemion’s pair) and -0.5 kcal/mol in the case of Leu-Glu system (common in Root-Bernstein’s and Siemion’s assignment). However, the interaction of Phe with Lys was also discussed in the literature. As Burley and Petsko have established, the positively charged ε-amino group of Lys could interact with Phe aromatic ring by π-face hydrogen bonding [37]. It was mentioned above that the reversion of the direction of the polynucleotide chain reading does not change the hydropathy of the coded peptide. In some special cases identical sequences may be deduced from all three approaches indicated above. E.g., in the case of proctolin, a peptide hormone isolated by Brown and Starratt from the body extract of the cockroach Periplaneta americana [38], the same sequence of antipeptide could be found among other possible predictions by using each of these approaches: proctolin: Arg-Tyr-Leu-Pro-Thr antipeptides: 5'-3' reading: Ala-Ile-Glu-Gly-Cys N-terminus (retro-sequence) 3'-5' reading: Ala-Ile-Glu-Gly-Cys C-terminus (the sequences of both antipeptides are, however, reversed)

As Biró has pointed out [3], the meaning of symmetrical coding triplets does not depend on the direction of codon reading. Therefore, a group of eight amino acid pairs: ProGly, Ala-Arg, Val-His, Leu-Glu, Ile-Tyr, Phe-Lys, Ser-Arg, and Thr-Cys, is exactly the same in both Root-Bernstein's and Blalock's assignment systems. The hypothesis on the amino acid complementarity as a result of the genetic code periodicity was presented for the first time in 1992 [39]. Siemion et al. evaluated the hypothesis using a set of transforming growth factor β (TGF-

The Problem of Amino Acid Complementarity and Antisense Peptides

Table 1.

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

511

Complementarity of Amino Acids. Comparison of Hypotheses of Blalock, Root-Bernstein and Siemion “anti-amino acid” according to three hypotheses Amino acid Blalock

Root-Bernstein

Siemion

Stop

Ser, Leu

Thr, Ile

Thr, Ile, Met

Ala

Arg, Cys, Gly, Ser

Arg

Arg

Arg

Ala, Pro, Ser, Thr

Ala, Ser

Ala, Ser

Asn

Val, Ile

Leu

Phe

Asp

Ile, Val

Leu

Leu

Cys

Ala, Thr

Thr

Thr

Gln

Leu

Val

Val

Glu

Leu, Phe

Leu

Leu

Gly

Ala, Pro, Ser, Thr

Pro

Pro

His

Met, Val

Val

Val

Ile

Asn, Asp, Tyr

Stop, Tyr

Stop, Tyr

Leu

Gln, Glu, Lys, Stop

Asn, Asp, Glu

Asp, Glu, Lys

Lys

Leu, Phe

Phe

Leu

Met

His

Tyr

Stop

Phe

Glu, Lys

Lys

Asn

Pro

Arg, Gly, Trp

Gly

Gly

Ser

Arg, Ala, Gly, Stop, Thr

Ser, Arg

Ser, Arg

Thr

Arg, Cys, Gly, Ser

Cys, Stop, Trp

Cys, Stop, Trp

Trp

Pro

Thr

Thr

Tyr

Ile, Val

Ile, Met

Ile

Val

Asn, Asp, His, Tyr

Gln, His

Gln, His

β) protein fragments and their respective counterparts. In the crucial experiment a mixture of Tyr-Ile-Gly-Lys-Thr-ProLys-Ile (I) and Tyr-Tyr-Ile-Gly-Lys-Thr-Pro-Lys-Ile-Glu (II) peptides (TGF-β sequences) and Ile-Tyr-Pro-Leu-Cys(Acm)Gly-Leu-Tyr (III), Ile-Ile-Tyr-Pro-Leu-Cys(Acm)-Gly-LeuTyr-Leu (IV) and Ile-Tyr-Thr-Leu-Cys(Acm)-Gly-Leu-Tyr (V) (complementary peptides) was examined by mass spectrometry using ESI-MS method. In the MS spectrum only the pairs of peptide-complementary peptide dimers were visible. Neither the mass peaks of dimers of TGFpeptides, nor the dimers of respective antipeptides, nor the peaks of any homodimers were present in the MS spectrum (Fig. (2)) [40, 41]. In the experiments performed with cyclolinopeptide A (CLA; c-(Leu-Ile-Ile-Leu-Val-Pro-Pro-Phe-Phe)) the authors compared complexation of CLA with two linear peptides: Lys-Tyr-Tyr-Lys-Gln-Gly-Gly-Asn-Asn (A; Siemion's peptide), and Asn-Tyr-Tyr-Asn-Gln-Gly-Gly-Lys-Lys (B; Root-Bernstein's peptide). In the MS spectrum only the mass peak of CLA-A, but no peak of CLA-B complex was visible. The separated peak, appearing at 704.8 m/z [CLA + A +

3H]+3 was split in collision-induced dissociation experiment into CLA and A, and the stability of such complex was comparable with that of the most stable complex of TGF-β series (I-V) (see Fig. (3)) [42]. It may mean that Siemion's assignment of complementary amino acids could be more successful than Root-Bernstein's indications in case of the controversy regarding Leu codons. However, this conclusion needs further confirmation. A similar experiment was also performed with the fragment of interleukin 1 receptor antagonist protein (IL1Ra) Val-Thr-Lys-Phe-Tyr-Phe and two of its complementary peptides: Gln-Trp-Leu-Asn-Ile-Asn (constructed according to Siemion's approach) and Gln-Trp-Leu-Lys-IleLys (Root-Bernstein's approach). No heterodimer formation with Val-Thr-Lys-Phe-Tyr-Phe was observed for neither of the two peptides by mass spectrometry [43]. It is too early, in our opinion, to decide which of the three approaches presented above should be distinguished from the others. Blalock’s approach is founded on the solid base of the natural orientation of coding and complementary DNA strands in the living cells. In this case the degeneracy of the

512

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

Siemion et al.

Fig. (2). The result of MS experiment with the mixture of TGF-β peptide fragments and respective antipeptides [40, 41].

Fig. (3). The comparison of the dissociation of the most stable pairs appearing in TGF-β and CLA mass spectra [42].

genetic code produces, however, the increased number of amino acid complements. Root-Bernstein’s approach guarantees – according to the literature data – better results in regard to the peptide – antipeptide complex formation. The way to determine the complementary pairs, used in this approach, seems to be somewhat artificial. Siemion’s approach offers a solid ground for the prevelence of the 3’-5’ direction of the reading of the complementary DNA strand. It differs, however, in a few points from the Root-Bernstein’s approach to the amino acid complements prediction. The comparison of the hydropathic indices [90] of amino acid residues of Root-Bernstein’s (R-B) and Siemion’s (S) pairs shows that the first approach results in a better correspondence for asparagine complements, and the second

one – for lysine complements. For Asn-Leu pair (R-B), the hydropathic indices equal -3.5 vs. 3.8, whereas for Asn-Phe (S) pair -3.5 vs. 2.8. In the case of Lys-Phe (R-B pair) we have the values -3.9 vs. 2.8, and in Lys-Leu (S pair) -3.9 vs. 3.8. Thus, such a comparison does not set the preference of any of these two approaches. II. AMINO ACID COMPLEMENTARITY AND THE PROBLEM OF SECONDARY STRUCTURE FORMATION, PROTEIN FOLDING, AND COMPLEXATION From the beginning the authors of the amino acid complementarity hypotheses postulated that such complementarity should be of importance for protein interactions, recognition of protein effectors by their receptors, antigenantibody complex formation, as well as in protein folding and similar phenomena.

The Problem of Amino Acid Complementarity and Antisense Peptides

Some literature data, especially those connected to the problem of antipeptide activities in biological systems, suggest that such expectations have some merit. However, the data on the possible role of complementary amino acids interaction in protein structure formation are scarce and inconclusive. Some authors question the role of such interactions in the discussed phenomena [48, 49]. In our review we decided to present the data which argument for the validity of such expectations, although we believe that the problem is far from clarification. As we have noted, the best conditions for complementary amino acid interactions appear when both considered residues exist in the conformation typical for β-structures of polypeptide chains. Thus, our attention is directed at the parallel β-ribbons and β-sheets present in the protein molecules as the elements of their tertiary structures. The amino acids occupying the (i) and (i+3) positions within the β-turns of different types also accept the βstructure conformation and are located at the distance enabling the interactions. Thus, the turns in the proteins are also worth attention as objects in which the amino acid complementary interactions could occur. The Amino Acid Pairing and the Secondary Structure of Proteins Mandel-Gutfreund et al. have pointed out that a few amino acid pairs could guide the folding trajectory of β-sheet proteins [44]. In 1999 Hutchinson et al. showed that some amino acid pairs of Mekler-Blalock type, like Gly-Ala, ValTyr, Ser-Thr, and Arg-Thr, demonstrate a preference for specific positions in the adjacent β-strands of protein [45]. The frequency of occurrence of the nearest neighbor residue pairs on adjacent anti-parallel and parallel β-strands was analyzed in 1980 by Lifson and Sander, but on a relatively small number of proteins (30 structures) [46]. Some years ago the analysis of side chain interactions and pair correlation within anti-parallel β-sheets was performed by Wouters and Curmi who investigated a set of 253 nonredundant protein structures [47]. It was found that a set of 14 amino acid pairs with the highest pair correlation is dominated by the ionic pairs of the charged amino acid residues. A high correlation was also observed for the CysCys pair. Accordingly, the disulfide bonds and ionic interactions are of the greatest importance for β-sheets stabilization. However, among other pairs with high correlation there appears the Ser-Ser pair, predicted by RootBernstein’s and Siemion’s approaches. We would also like to direct the attention to the fact that complementary amino acids show nearly the same preferences for anti-parallel βsheets, e.g. Gly (0.64) – Pro (0.48); Val (1.68) – Gln (0.88), His (0.97) (Root-Bernstein’s and Siemion’s approaches). In this last case Val could form two pairs, with Gln and His, therefore a sum of Gln and His propensities should be compared with the propensity for Val. A similar analysis was performed by Hutchinson et al. [45]. They showed that ArgSer and Ile-Tyr complementary pairs belong to the group statistically favored in anti-parallel β-sheets. Both these pairs belong to the hydrogen bonded pairs, in which backbone carbonyl groups and amide hydrogens are bound to each other. The same situation appears in the case of Ser-Ser pair,

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

513

indicated by Wouters and Curmi [47]. It is worth noting that the Ile-Tyr cross-strand pair belongs, according to the results of Smith and Regan [48], to the group of pairs with the highest β-sheet stabilizing interaction energy, whereas the interaction energy of Thr-Trp pair was found to be destabilizing. According to Root Bernstein’s approach the amino acid complementarity should be more characteristic for parallel than for the anti-parallel β-sheets, although there are no proper statistical data available for such structures in proteins. However, in a very interesting structure of UDP-Nacetylglucosamine 3-O-acetyltransferase [49], dominated by parallel β-strands that form three parallel β-sheets, no complementarity of amino acid residues localized in adjacent strands is observed. Some years ago Heringa and Argos analyzed the presence of dense clusters of amino acid side chains in proteins with known tertiary structure. The clusters consisted generally of three to four amino acid side chains. According to the authors the clusters are likely to be important in the protein folding and stability. The clusters are dominated by complementary charged residues. However, the amino acid complementarity principle seems also to be important for their formation. E.g. in the case of γ-crystallin, one of the proteins analyzed by Heringa and Argos [50], four such clusters are present. They are composed of the following residues: 1)

Tyr6, Phe11, Ser34, Arg36, Asp61;

2)

Tyr134, Tyr139, Ser166, Arg168;

3)

Phe98, Asn125, Glu150, Arg152;

4)

Tyr45,Tyr50, Ser77, Arg79

It is remarkable that in each of these clusters there appears a pair of complementary amino acids (according to Siemion's assignment). Amino Acid Pairing in Protein Folding In 1989 Draper found that regions of self complementarity appear within the molecule of human albumin mRNA, especially in the region encoding the sequence of protein loop (325-345 fragment), in which several complementary amino acid pairs could be formed [51]. Later Baranyi et al. developed software allowing searching for sense-antisense regions within proteins based on the Blalock’s assignment of complementary amino acids. They described such regions as “antisense homology boxes”. In the opinion of Baranyi et al. the antisense homology boxes could be of importance for such phenomena as the aggregation of proteins, protein folding and chaperoning, and the protein-receptor interactions. The majority of such boxes, found in a number of proteins, contained the reverse turns, which suggests that they are located on the surfaces of molecules. The peptide Cys-Ala-Leu-Ser-Val-Asp-Arg-TyrArg-Ala-Val-Ala-Ser-Trp that corresponds to one of the antisense homology boxes was found to be a specific inhibitor of smooth muscle relaxation evoked by endothelin (ET) [52]. As Baranyi et al. pointed out, the indicated peptide is not only antisense to several other antisense

514

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

homology boxes, but also antisense to itself. Continuing this work, Baranyi et al. have shown that also four other ETA receptor fragment peptides of this type can inhibit ET-1 activity [53]. The antisense homology boxes approach was also applied by the same authors to C5a anaphylatoxin and its receptor [54]. They consider this approach as a new method for identification of potentially active peptides, interesting from the practical point of view. The problem of a role of antisense homology boxes in the formation and maintenance of tertiary structure of protein molecules was discussed separately by Okada [55]. Baranyi's technique of the determination of the “protein complementary boxes” closely resembles the method of “informational complementarity” investigation, developed earlier by Biró [3]. The folding of polypeptide chains of proteins leads to the formation of numerous reverse turns and loops on the molecule surface. It is interesting whether the amino acid complementarity plays any role in the formation of turns and loops. The β-turns, which are very common in proteins, are composed of four successive amino acid residues. The classification of different β-turns was proposed in 1977 by Chou and Fasman [56]. The turn formation brings into proximity the residues occupying positions i and i+3 of the turns. The protein loops were described by Leszczynski and Rose [57]. The loops are composed of 6 to 16 amino acid residues and constitute the segments of continuous polypeptide chain that trace a “loop-shaped” path in the

Siemion et al.

space. They are situated on the surface of the protein molecule and possess highly compact structures. Roughly 18% of 206 loops compiled by Leszczynski and Rose contain the complementary amino acid residues. The low number of the loops, in which the complementary pairs could be found, makes the idea that they can play a significant role in the formation of the loops rather doubtful. However, an interesting example of such a loop is present in the molecule of a prothrombin activator secreted by Staphylococcus aureus. The N-terminal hexapeptide fragment, important for the activity of this protein, is terminated on both sides by Ile1 and Tyr6 residues; the side chains of these two amino acids are oriented one towards the other and thus can form a complementary pair [58]. The protein loops represent sometimes a combination of associated β- and γ-turns. A set of 475 such combinations, resulted from X-ray data for 248 protein crystal structures, was collected very recently by Guruprasad et al. The combinations span from nine to twenty six amino acid residues along the polypeptide chain [59]. However, the most frequent are the nine and ten residue fragments. It can be seen that within a group of 100 such peptide sequences representing combinations of turns, roughly 70% contain two to even eight amino acid residues able to form complementary pairs. Thus, a significant role of amino acid pairing in such combinations seems very possible. In fact, in some of these loops a clear interaction between complementary amino acid chains takes place. Two examples of such situation are shown in Fig. ( 4) and (5). Fig. (4) shows the spatial orientation of Leu and two Asp residues

Fig. (4). κ-Carrageenase fragment 124-132. The distance between C β atoms of Leu 127 and Asp124 is 5.0 Å, between C β atoms of Leu 127 and Asp132 is 3.9Å. Structure from PDB file 1dyp [60]

Fig. (5). Ferredoxin thioredoxin reductase, catalytic chain fragment 85-98. The distance between Cβ atoms of Leu91 and Asn 95 is 4.1 Å. Structure from PDB file 1dj7 [61].

The Problem of Amino Acid Complementarity and Antisense Peptides

within a loop formed by κ-carrageenase (1,3α-1,4β-Dgalactose-4-sulfate-3,6-anhydro-D-galactose-4-galactohydrolase) [60]. The Leu side chain here is inserted between two Asp residues realizing the bilateral interaction of RootBernstein – Siemion type. The distances between Cβ atoms of Leu and both Asp residues reach values of 3.9 and 5.0 Å, respectively, and denote a high possibility of interactions. It is of interest that the conformations of the discussed residues are close to that of β-structure. The correspondent dihedral angles φ and ψ are -86.9° and 95.4° for Leu 127, -109° and 98° for Asp124, and -63.5° and 139.5 for Asp132. Thus, the conformation of these residues corresponds quite well with that proposed for interacting complementary amino acids. Both Asp residues are oriented, as regards Leu residue, in antiparallel direction. In Fig. (5) the interacting Root-Bernstein’s Leu-Asn pair within a loop which exists in ferrodoxin thioredoxin reductase is demonstrated [61]. The distance between Cβ atoms of Leu and Asn residues is 4.1 Å, allowing a close contact between these residues. As in the previous example, the dihedral angle values reach -110.5° and 142 for Leu91, and -60° and 136° for Asn95 which corresponds to somewhat distorted β-structure for Leu, and poly-Pro II structure for Asn. The residues are oriented antiparallel. In both these cases the interacting residues belong to the same peptide chain, but the turns appearing within the structure situate them in the proper spatial orientation, enabling the interaction. A possible participation of amino acid complementary pairs in the long-range interactions in proteins presents another problem. In 1989 Perry et al. identified such an interaction in dihydrofolate reductase between the Leu28 (located at the active site of the enzyme) and Glu139 residues, whose α-carbons are 15 Å away [62]. The mutations in the indicated positions result in the sufficient change in activation free energies of the folding processes of the protein. The Leu-Glu pair belongs to the set of complementary amino acid pairs according to all three assignments. During the folding of polypeptide chains several disulfide bridges between cysteine residues are usually formed. It seems possible that the creation of Cys-Thr complementary pair between threonine adjacent to one of the cysteine residues and the second Cys residue could facilitate the mutual recognition of the proper parts of the polypeptide chain. If so, the vicinity of the disulfide bridge should be enriched in Thr residues. The inspection of many polypeptide structures, especially those rich in disulfide bridges seems to partially confirm such a supposition. A situation like this takes place e.g. within the molecule of bicyclic trypsin inhibitor from sunflower seeds: Ser-Lys-Thr-Cys-Arg-Gly Asp Pro-Pro-Ile-Cys-Phe-Pro (the sequence taken from [63, 64]).

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

515

Beside the threonine adjacent to one of the Cys residues, an interaction between Pro and Gly complementary pair is also possible in this structure. In a recent work, Catalano et al. compared the sequences of four Bowman-Birk type inhibitors, isolated from snail Medicago scutellata, soybean, tracy soybean, and pea seed, respectively [65]. This polypeptide family is interesting from the indicated point of view because of the presence of seven disulfide bridges in each of the peptide molecules. In the nearest vicinity of Cys residues in these peptides there occurs 10 Thr residues (11.7 % of all non-cysteinic amino acid residues adjacent to cysteines). However, the same number of Ala and Ser residues appears at the positions adjacent to Cys. The threonine residue at the P2 position of BowmanBirk proteinase inhibitors is especially resistant to mutational changes. In 47 various Bowman-Birk proteins the Thr residue at the P2 position (i.e. adjacent to Cys which occupies position P3) is changed into other residues (Ala, Arg, or Asn) in six cases only [66]. However, another situation takes place in conotoxin peptides family. In the molecules of these peptides there appear four disulfide bridges. In the recent paper by Favreau et al. [67] the sequences of 11 ω-conotoxins, isolated from Conus snail venoms were compared. The threonine adjacent to Cys occurs in these peptides only three times. These data are presented here only as the illustration of the thesis expressed above. It seems, however, that the statistical analysis of the indicated tendency may be worth attention. Complementary Pairs at the Protein-protein Interfaces Recently Glaser et al. analyzed the residue frequencies and pairing preferences at the protein-protein interfaces. The analysis was performed on a very large group of 621 interfaces of known high-resolution protein structures. It was generally found that the highest pairing preferences occur for the large hydrophobic residues, like Trp and Leu, and that the contacts between the hydrophobic residues on one side and polar on the other are rather not favored. The charged residues tend to complement their charge by the interaction with the residues having the opposite charge [68]. Table 2 presents the values of residue-residue contact preferences normalized by residue volumes as determined by Glaser et al. Here we demonstrate these data in the order resulting from one-step mutation genetic ring, proposed by Siemion [28]. It can be seen from these data that the hydrophobic and ionic interactions dominate the interface stabilizing forces. The contact preferences of the majority of complementary amino acid pairs range within medium preference values. The pairs Gly-Pro and Ser-Ser are scarcely represented. On the other hand, the preferences for Ile-Tyr and Thr-Trp complementary pairs are very high. The pair Thr-Trp is the most distinct of all pairs formed by threonine. The pair IleTyr is also very representative for those formed by isoleucine; only the pair Ile-Trp has a somewhat higher preference coefficient (6.24 vs. 5.61 for Ile-Tyr). The data in (Table 2) are given as residue-residue contact preference Gij(v) values, calculated according to the equation [68]: Gij(v) = A x log(Qij(v) / Wi x Wj)

516

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

Table 2.

Siemion et al.

Residue-residue Contact Preferences Gij(v), Volume Normalized (Data From [68]; Ordered According to the Sequence of Amino Acids in the One-step Mutation Ring of the Genetic Code). Data for Complementary Pairs in Bold

G

E

D

N

K

Q

H

Y

W

C

S

A

T

P

R

L

F

V

I

M

G

-4.40 -0.89 -0.08 -0.54

1.33

0.70

1.08

1.25

1.42

-0.25 -1.53 -1.77

0.21

-0.51

1.51

-0.37

0.14

-0.14

0.77

0.91

E

-0.89

1.65

0.08

2.68

5.32

1.95

2.30

4.54

1.20

2.51

2.60

1.71

2.88

3.17

5.75

3.12

2.87

3.22

3.20

3.88

D

-0.08

0.08

0.13

3.85

3.90

3.26

5.20

1.76

2.62

0.24

2.94

1.18

3.88

1.46

4.94

1.40

0.99

1.93

2.30

0.36

N

-0.54

2.68

3.85

2.92

3.17

3.45

2.38

3.66

3.54

-0.42

1.77

1.69

2.52

3.09

3.85

2.31

3.11

1.36

1.59

2.30

K

1.33

5.32

3.90

3.17

3.24

3.50

2.72

5.26

5.76

2.05

2.74

2.13

3.67

3.75

2.29

3.15

3.57

4.45

3.23

3.93

Q

0.70

1.95

3.26

3.50

3.50

2.83

4.00

2.05

1.37

1.33

2.00

1.72

1.82

3.50

4.50

3.46

4.25

3.22

3.60

4.18

H

1.08

2.30

5.20

2.38

2.72

4.00

5.37

6.05

6.46

4.12

0.80

2.59

2.71

2.89

4.90

4.88

3.47

3.21

3.38

4.65

Y

1.25

4.54

1.76

3.66

5.26

2.05

6.05

5.93

6.13

2.47

2.30

2.47

3.14

4.22

5.28

4.19

5.83

3.95

5.61

4.81

W

1.47

1.20

2.62

3.54

5.76

1.37

6.46

6.19

5.85

2.14

2.87

3.37

5.12

7.87

8.57

5.77

5.83

2.92

6.24

4.89

C

-0.25

2.51

0.24

-0.42

2.05

1.33

4.12

2.47

2.14

7.65

2.48

1.46

1.03

2.74

2.81

2.93

3.68

2.89

1.76

1.84

S

-1.53

2.60

2.94

1.77

2.74

2.00

0.80

2.30

2.87

2.48

-0.09

0.39

1.91

1.33

2.82

1.41

1.75

1.42

1.00

1.61

A

-1.77

1.71

1.13

1.69

2.13

1.72

2.59

2.47

3.37

1.46

0.39

-0.52

1.21

1.22

1.90

2.77

3.00

3.57

2.84

2.30

T

0.21

2.88

3.88

2.52

3.67

1.82

2.71

3.14

5.12

1.03

1.91

1.21

1.27

2.65

3.77

2.07

3.34

2.83

3.05

2.09

P

-0.51

3.17

1.46

3.09

3.75

3.50

2.89

4.22

7.87

2.74

1.33

1.22

2.65

0.60

3.99

2.50

4.25

2.90

3.27

3.38

R

1.59

5.75

4.94

3.85

2.29

4.50

4.90

5.28

8.57

2.81

2.82

1.90

3.77

3.99

2.87

4.99

4.49

4.18

3.80

3.62

L

-0.37

3.12

1.40

2.31

3.15

3.46

4.88

4.19

5.77

2.93

1.41

2.77

2.07

2.50

4.99

4.03

4.86

4.20

4.59

5.32

F

0.14

2.87

0.99

3.11

3.57

4.25

3.47

5.83

5.83

3.68

1.75

3.00

3.34

4.25

4.49

4.86

5.34

4.69

5.33

5.28

V

-0.41

3.22

1.93

1.93

4.45

3.22

3.21

3.95

2.92

2.89

1.42

2.57

2.83

2.90

4.18

4.20

4.69

3.74

4.91

4.37

I

0.77

3.20

2.30

1.59

3.23

3.60

3.38

5.61

6.24

1.76

1.00

2.84

3.05

3.27

3.80

4.59

5.33

4.91

3.89

5.25

M

0.91

3.88

0.36

2.30

3.93

4.18

4.65

4.81

4.89

1.84

1.61

2.30

2.09

3.38

3.62

5.32

5.32

4.37

5.25

6.02

where A is a constant arbitrarily set to 10, Qij(v) is the residue-residue contact number normalized by residue volumes, and Wi and Wj are the normalized frequencies of i and j residues within the analyzed set of proteins. The amino acid pair was considered to be in contact when the distance between their Cβ atoms (Cα for Gly) were smaller than a certain cutoff distance.

The correlation depicted in Fig. (6) is based on Siemion’s assignment of complementary amino acids. Using the RootBernstein’s pairs, a very similar result, with the correlation coefficient R2 = 0.8604 was obtained. However, applying Blalock’s assignment leads to significantly worse correlation (R2 = 0.6566) which is mainly due to the irregularity introduced by Trp-Pro pair specific for Blalock`s approach.

As it is shown in Fig. (6), there is a linear correlation between the contact preferences of complementary pairs and the summary residual volumes of the correspondent amino acid residues. The pairs with the highest summary volumes possess the highest preference coefficients. In the construction of relations the values of residual volumes in Å3 given by Goldsack and Chalifoux [69] were adopted by us. The correlation coefficient of this relation reaches a value of R2 = 0.8543, thus, the level of the confidence can be classified as a very satisfying.

In the control, the similar relations for the groups of pairs formed by a given amino acid with all other proteinaceous amino acids were examined. The relation of the contact preferences to the summary volumes of the corresponding pairs is preserved but merely as the general tendency. Only for the groups of pairs formed by Gly, Pro, Leu, Ile, Phe, and Met relatively good linear relations were observed; however, the correlation coefficients were in each case distinctly lower than in the case depicted in Fig. (6). The presented data suggest that the formation of complementary pairs may be involved in the protein

The Problem of Amino Acid Complementarity and Antisense Peptides

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

517

Fig. (6). The contact preferences Gij(v) (taken from Table 2) of complementary amino acid pairs (Siemion`s assignment) versus their summary residual volumes in Å3.

complexation process. However, the role of complementary pair formation seems to be of secondary significance in this process. It seems interesting to mention here the ideas of Chipens concerning the two-step model of protein complex formation. Chipens assumes that the interaction between peptide and protein begins with a specific recognition and primary complex formation. The complementary amino acids pairing plays a crucial role at this stage of complexation. During the second stage the primary complex rearranges and a new structure is determined by pairwise contacts of hydrophobic and polar residues [70]. However, there are no experimental data confirming this interesting hypothesis. The data presented above show that there exists some preference for Ser-Ser, Arg-Ser, and Ile-Tyr complementary pairs within adjacent β-strands in proteins. It is of interest that two of these pairs (Arg-Ser and Ile-Tyr) are found in all three approaches to the problem of amino acid complementarity. A possible role of such interactions in protein loops formed by several successive β- and γ-turns, as well as at the protein-protein interfaces is also worth attention. The Ile-Tyr and Thr-Tyr pairs appear preferentially at such interfaces. It is remarkable that the same Ile-Tyr pair is privileged in pairs present in β-strands of protein molecules. The presence of many “antisense homology boxes” found by Baranyi in proteins also arguments for the possible importance of amino acid complementarity in protein folding and protein-protein complexation. III. PEPTIDE – ANTIPEPTIDE COMPLEXES The pairing of peptides with the proper complementary counterparts belongs to one of the main problems of investigations on the amino acid complementarity field. As one of us (I.Z.S.) remembers, the idea to synthesize the complementary peptides (coded by complementary DNA

strands) was firstly presented by J.S. Morley from Imperial Chemical Industries Laboratory during the 8th European Peptide Symposium in Noordwijk (1966). Later the results concerning this subject were published by Jones from the same research group [71]. The work was devoted to the synthesis of antipeptides from C-terminal tetrapeptide sequence of gastrin (Trp-Met-Asp-Phe-NH2). The complementary peptides resulted from both 5'-3' and 3'-5' reading direction were synthesized. The peptides obtained did not show stimulation of gastric acid secretion and were also devoid of inhibitory action on secretion stimulated by pentagastrin. Most of the papers published and connected to the problem of amino acid complementarity deal with the peptide hormone chemistry. There are also numerous review articles on this subject. Those of Blalock and Bost [72], Beattie [73], Root-Bernstein and Holsworth [74], and Heal et al. [75] could be quoted here. From this list a critical review of Root-Bernstein and Holsworth is worth special attention. The Root-Bernstein's paper tabularizes the best characterized examples of peptide – complementary peptide pairs. Both 5'3' and 3'-5' reading approaches are discussed. In the tables dissociation constants of the complexes, as well as the methods of investigation used in particular reports are given. In the review of Heal et al. the main biological protein – peptide systems from which complementary peptides have been derived are collected in a special table. The data on the peptide – antipeptide complexes are, however, not as complete as in the review of Root-Bernstein and Holsworth. In our review we concentrated the attention on the general properties of peptide - complementary peptide complexes, without the intention to present the complete survey of the literature. The formation of peptide – antipeptide pair requires the proper spatial orientation of interacting components. As we have noted above, for the 3'-5' complements Root-Bernstein proposed the parallel β-structure as the conformation needed

518

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

for complexation. In the case of Gly-Pro pair, however, the poly-Pro II structure is the proper one according to RootBernstein postulates. Both these conformations are of the extended type and do not differ sufficiently in the values of Φ and Ψ conformational angles (approximately -120°, +120° for β-structure, and -60°, +120 ° for poly-Pro II). Markus et al. have also pointed out that the strong hydropathic complementarity can take place only when the peptides resulting from 3'-5' reading are aligned parallel to their natural counterparts. On the other hand, the peptides resulted from 5'-3' reading must be oriented anti-parallel [9]. Thus, the extended anti-parallel β-structure must be also taken into account as the possible conformation of interacting peptides. In this context it is worth noting that the extended conformation seems to be typical for the antigenic peptides presented by the major histocompatibility complex class II HLA-DR1 molecules. Crystallographic analysis proposed by Jardetzky et al. showed that the presented peptide assumes regular conformation that is similar to a polyproline type II helix (poly-Pro II) [76]. The transconformational processes in the PrPc protein molecules, i.e. in the transition of helical into the β-conformation constitute also the ground of prion disease [77, 78]. According to the hypothesis of Gazit, the aggregated forms of proteins, in which the β-structure predominates, represent the true global minima of free energy of protein molecules, whereas the “correctly folded” proteins in solution may represent their metastable states [79]. The acceptation of this hypothesis leads to the conclusion that the peptide – complementary peptide complexes adopt the conformation which corresponds to that of global free energy minimum of proteins. It can be also anticipated that short linear peptides, complementary to the binding sites of respective proteins, could inhibit their aggregation processes. Such an idea was recently examined in case of β-amyloid aggregation and neurotoxicity [80]. Amyloid peptide 1-40 (Aβ1-40) is a significant component of neurofibrillary tangles, whose deposits characterize the neurodegenerative disorder – Alzheimer disease. Heal et al. examined the influence of the series of antipeptides ordered against fibrilization of A β1-40. It was found that only the 3'-5' complementary peptide binds specifically to Aβ1-40 and inhibits Aβ fibrilization and neurotoxicity. In the opinion of Heal et al. this peptide could form a base of a new therapeutic approach against Alzheimer disease. Another example of such approach presents the work of Gartner et al. [81] on the inhibition of fibrinogen binding to platelets. It was found, among others, that Asp-Pro-Pro-ArgPhe-Val-Arg-Pro-Leu-Gln peptide, a 3'-5' complement to Leu-Gly-Gly-Ala-Lys-Gln-Ala-Gly-Asp-Val sequence of γchain of fibrinogen, is able to bind specifically fibrinogen and to inhibit platelet aggregation, the adhesion of platelets to immobilized fibrinogen, and the clot reaction [82]. A view that the conformation of interacting complementary peptides should be close to β-structure is confirmed by some evidence from CD spectroscopy. In the CD spectra, the β-structure is characterized by a low negative CD band at about 220 nm, and a strong positive

Siemion et al.

band below 200 nm [83]. If the peptides in complementary peptide mixtures really interact their CD spectra should differ from the sums of ellipticies of the individual components, and the difference CD spectra may indicate the shifts in conformational equilibrium of peptide mixtures. However, the investigation performed by Najem et al. [84] with ACTH(1-24) peptide and its 5'-3' complement showed that there is no difference between the CD spectrum of the equimolar peptide mixture and the sum of the CD spectra of the individual components. In this case the lack of interaction was also confirmed by NMR spectroscopy. At the same time CD spectroscopy was used by Blalock's group in investigating the calcium octapeptide mimetic [85]. This peptide presents the antisense peptide towards the Ca2+ binding site of proteins of troponin superfamily. It can replace the Ca2+ ions in their biological effects. The CD spectra of the peptide showed that it binds, like Ca2+, with EDTA. Such a binding induces a shift in conformational equilibrium of the peptide chain towards β-structure. Fassina used the CD method in the investigation of complementary peptide for big-endothelin 16-29 sequence (Blalock`s approach). Big-endothelin displays a vasoconstrictor activity and is even more potent than angiotensin II. The binding between the antipeptide and bigendothelin partial sequences was controlled by CD and NMR spectroscopy. The changes in the spectra of peptide – antipeptide mixture compared to the spectra of the sum of individual components confirmed their interaction [86, 87]. In the last time the sufficient contribution to the CD spectroscopy of complementary peptides was done by Siemion et al. [41]. The authors evaluated the amino acid complementarity with TGF-β2 related peptides. The tendency of these peptides to form complexes with the correspondent complementary peptides was independently examined by ESI-MS spectroscopy, i.e. the direct visualization of the proper mass peaks of hetrodimers in the mass spectra. The difference CD spectra of 1:1 peptide – antipeptide mixtures demonstrated the shift (in comparison to the sum of the spectra of individual components) in the conformational equilibrium towards the β-conformation. It was evidenced by the positive CD band present below 200 nm. The samples, for which no complexation was visible in MS, showed the negative, or even no changes in this region of difference CD spectra. It was also shown that the peptides for which the CD spectra evidence the preponderance of folded forms of the peptide chains (i.e. characterized by the presence of turns) in the conformational equilibrium, do not form the complexes with the proper antipeptides. That means that the extended conformation is really needed for the complex formation. It also follows from these studies that the most efficient binding in peptide – antipeptide systems appeared at neutral pH. For the first time mass spectroscopy was proved to be useful for the elucidation and visualization of peptide complementary peptide complexes by Loo et al. who used ESI-MS to test for non-covalent interactions between human angiotensin II and its synthetic complementary octapeptides [173]. This method also enables the determination of relative stability of complexes. For the last purpose the collision experiments should be performed, in which the separated

The Problem of Amino Acid Complementarity and Antisense Peptides

ions of definite complexes are treated with argon atoms under indicated potential. In this context, the report of Madhusudanan et al. should be also quoted [88]. The Metand Leu-enkephalins and their complementary counterparts were the objects of these studies. In the study, both 5'-3' and 3'-5' complements of enkephalins were used. The complementary peptides did not show the opioid activity, but rather had a dose-dependent synergistic effects on enkephalin action [89]. In the MS spectra of the peptide – antipeptide mixtures, both homodimer and heterodimer peaks were visible, however, the preference for heterodimer formation was established. There were no sufficient differences in the binding of 5'-3' and 3'-5' complements. The amino acid sequence of 3'-5' antisense peptides (Ile-Pro-ProLys-Tyr and Ile-Pro-Pro-Lys-Asp) suggests, in our opinion, that the peptide conformation within the complexes should be rather the poly-Pro II and not the β-type. The sequences of complementary peptides are usually deduced from the nucleotide sequence of the complementary DNA strand, or RNA strand which is complementary to a proper mRNA, read in anti-parallel (Blalock) and parallel (Root-Bernstein) direction. In the approach of Siemion, the codon equivalence resulted from the genetic code periodicity was used as a basis of sequence deduction. As we indicated above, the antisense peptide complementarity was generalized by Blalock in his molecular recognition theory, in which a principle of hydropathic complementarity of interacting peptides is considered as the source of interaction. Basing on this principle, Fassina et al. developed a computerized approach for the determination of antipeptide sequences from the primary sequence of the target peptide. During this procedure to each residue of the target peptide a set of hydropathically complementary amino acids is assigned with the aim to find the sequence with maximal hydropathic complementarity. The table of complementary amino acids is given in the paper of Fassina et al. [87]. In the opinion of Fassina et al. the computer generated sequences show even higher binding affinities for their target peptides than the DNA deduced complementary peptides. Fassina's data show that increasing the average hydropathic complementarity of interacting peptides can improve the peptide - complementary peptide affinity. This approach converts the initial methods of antipeptide sequence determination into a more general one. However, it must be noted that the results of the proper experiments are not always consistent with the expectations resulted from Fassina's approach. The Kyte-Doolittle hydropathy scale is usually adopted for determination of mutual hydropathic complementarity of investigated peptides [90]. The positive results of the works based on the Fassina's approach to the problem of determination of antipeptide sequences suggest that there exists quite broad area of changes in the amino acid composition and sequences within peptide – antipeptide complementary pairs. Such a supposition is supported by the literature data. Shai et al. showed that S-peptide derived from pancreatic ribonuclease A is bound equally well by antisense 20-mer and by its retroanalog. That means that the interaction of S-peptide with its complement does not depend on the direction of amide bonds in the complementary peptide. These authors found,

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

519

however, that the affinities of complementary peptides weaken with the decreasing length of the peptide chain [91]. The “scramble” mutants of S-peptide complementary peptide, in which the order of amino acid residues was drastically changed while the amino acid composition was retained, were also bound to the S-peptide. The hydropathically sequence-simplified analog of antisense peptide was also obtained, in which all strongly hydrophilic residues were replaced by Lys, the strongly hydrophobic by Leu, and the rest, except for Gly, were replaced by Ala. This analog showed the affinity only an order of magnitude lower than that of original antipeptide [92]. Fassina et al. synthesized antisense complementary peptide to 356-375 sequence of c-raf protein. Among the peptides composed of L-amino acid residues, an analog with all D-residues was obtained. For both these antipeptides the identical binding properties were obtained. Binding was also unaffected by detergents and 8 M urea, but depended on ionic strength and pH of the solution. Substitution of the original residues of antipeptide with the residues with similar hydropathy does not significantly affect the binding. The binding was evaluated by analytical high performance affinity chromatography; the dissociation constants of the complexes were determined by isocratic zonal elution [86]. These data suggest that the interaction in peptide – antipeptide pairs depends on hydropathic arrangement of the amino acid side chains, without respect to amino acid configuration and peptide chain direction. On the other hand, there are also reports on the high specificity of peptide – antipeptide interaction. In the paper of Brentani et al. the antipeptide to Arg-Gly-Asp-Ser fibronectin fragment with the sequence Trp-Thr-Val-ProThr-Ala (Blalock`s assignment) was described. It was coupled to AH-Sepharose column. Fibronectin bound to this column was eluted much more efficiently by Arg-Gly-AspSer tetrapeptide than by its Arg-Gly-Glu-Ser analog. Assuming that the peptides are oriented in the complex antiparallel, the difference in affinity depends here on the change of Val – Asp interacting pair for the Val – Glu pair [93]. Johnson and Torres found that the Thr-Met-Lys-Val-LeuThr-Gly-Ser-Pro antipeptide to arginine vasopressin (AVP) binds to AVP and blocks AVP-induced interferon-γ production. However, the exchange of Lys3 residue by Asp or Val, respectively, sufficiently influences the effects exerted by this peptide [94]. These findings show that in some cases a single “one-point mutation” in the sequence of complementary peptide leads to a marked change in the binding affinity. The stability of peptide – antipeptide complexes is not very high. The affinities of interacting complements are situated usually in the milli- to micromolar range, and only exceptionally they reach the nanomolar value. In the determination of KD values for the interacting pairs several methods were used, including radiolabelled measurements of binding to solid support anchored peptides, competitive radiolabelled receptor - ligand binding [94, 95, 96], analytical high performance affinity chromatography [91], NMR spectroscopy (for the review see [97]), pH titration [25], and spectropolarimetry [84].

520

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

The reports on the application of NMR spectroscopy for the determination of reciprocal affinities of peptides and their peptide complements were reviewed by Curto and Krishma [97]. The changes of chemical shifts and line broadening, nuclear Overhauser experiments (NOESY), TOCSY (total correlation spectroscopy), and ROESY (rotating Overhauser effect spectroscopy) techniques were exploited in these investigations. Experimental data obtained for ACTH, angiotensin II, big-endothelin, luteinizing hormone – releasing hormone (LHRH), calcium like peptide, γ-loop of neurotoxin proteins from scorpion venom, trypsinmodulating oostatic factor, and antipeptides for enkephalin, are collected in this review. Because of some controversy existing in the literature, the data for ACTH and its complements should be quoted here. The NMR studies performed by Najem [84] showed that in the case of the complexes of ACTH(1-24) polypeptide with its antisense components the K D constant is greater than 1 mM, whereas in a solid matrix binding assay the value of 0.3 nM was obtained [7]. In the application of NMR spectroscopy to the studies of peptide – antipeptide complexes the works of RootBernstein's research group played the pioneering role. His works on fibrinopeptide A binding to Gly-Pro-Arg-Pro tetrapeptide [24], the interactions of bovine pineal antireproductive tripeptide with LHRH [25], and on the complements to angiotensin II [98] should be quoted here. From the recent reports, the paper of Misra et al., concerning the enkephalin complexes, as well as that of Moulia et al. on 2D NMR studies of peptide complements to B and T cell epitopes of La/SSB autoantigen for immunoregulation in Sjögren's syndrome may be mentioned [89, 99]. Curto and Krishna in their review article [97] concluded that there is an apparent discrepancy between the dissociation constants measured for peptide – antipeptide pair in solution and in the solid phase binding assays. In the first case they are greater than 10-4 M, but reach the range of 10-6 – 10-9 M in the second case. In the analytical high performance affinity chromatography experiments the sense – antisense peptide affinity is quantified from the degree of retardation during chromatographic elution on the sense peptide affinity matrix with and without mobile antisense peptide. Shai et al. have applied this method in the case of ribonuclease S-peptide system [91]. Fassina et al. have used it for peptides hydropathically complementary to 356-375 fragment of c-raf protein [86] while Ghiso et al. to a cystatin C 55-59 fragment and its antisense peptide located at position 611-614 of the β-chain of human C4 protein (the fourth component of the complement) [100]. It was found that the specific interaction between the native cystatin C and C4 proteins is inhibited by synthetic antisense peptides. Taking into account the data concerning the affinities for peptide – antipeptide systems determined by the indicated methods, Holsworth et al. concluded that in fact there appears the binding of 3'-5' complements to the target peptides, whereas the binding of 5'-3' complements does not take place [98]. This conclusion, however, needs further verification.

Siemion et al.

Antisense peptides have been used as immobilized ligands for the chromatographic separation of native peptides and proteins. The reviews of reports concerning such biotechnological applications of molecular recognition theory were published by Chaiken [101] and Fassina [102]. The works on purification of big endothelin [87], interleukin 1β [103], and tumor necrosis factor [104], are reviewed in Fassina's paper. The multimeric form of antipeptide to big endothelin was constructed in these investigations, starting from an octadentate polylysine core. The affinity of such multimeric form of antipeptide to the multimeric form of 1632 fragment of endothelin was found to be enhanced by at least two orders of magnitude as compared to the monomeric peptide pair [105]. The data presented in this chapter strongly suggest that the formation of peptide-complementary peptide complexes is really a natural phenomenon. The evidence confirms the conclusion that the complexation abilities of peptides resulting from 3’-5’ reading exceed those resulting from 5’3’ reading. The investigations of Siemion’s research group suggest also that the peptide complements generated according to the genetic code periodicity principle may be more effective in complexation than those constructed by using the classical Root-Bernstein’s approach (see also chapter 1). This statement needs, however, a further verification. The investigations of circular dichroism spectra of the peptide-complementary peptide mixtures suggest that the conformation of the peptides in the complexes should be of the β-structure type. IV. SELECTED PROBLEMS CONCERNING THE BIOLOGICAL ACTIVITY OF COMPLEMENTARY PEPTIDES Chapters IV and V present the attempts to apply the amino acid complementarity principle to solve practical problems, mainly of the medical character. Chapter IV describes selected data concerning the biological activity of complementary peptides directed against peptide and protein cellular effectors. Chapter V summarizes the results of the investigations on complementary peptides used in receptor studies and immunochemistry. In our opinion, these last results are most interesting as far as the possibility of practical applications is concerned. They also present some very interesting new ideas of general significance. The biological activity of complementary peptides was extensively presented in the reviews cited above. Therefore, only some selected questions are discussed here. As a rule, the complementary peptides antagonize the activity of their counterparts, usually acting as the competitors for respective cell receptors. From the numerous papers on such competition, the works of Root-Bernstein and Westall [24], Holsworth et al. [98], and a review by Tropsha et al. [106] can be quoted. However, several cases of such complexation resulting in agonistic activity have also been reported, e.g. presented by Dillon et al. in the paper concerning the angiotensin II antipeptides [27] or by Kluczyk et al. for interleukin-1 receptor antagonist (IL-1Ra) fragment antipeptides [43].

The Problem of Amino Acid Complementarity and Antisense Peptides

Thus, there are no general rules concerning the antipeptide activity. It seems, however, that the agonistic properties of antipeptides could be predicted when a distinct homology appears between the antipeptide and its target structure. The examples of such situation are presented throughout the text of this review. Because of the controversies indicated in the literature, the problem of angiotensin antipeptides is worth attention. In 1988 Elton et al. [107] found that octapeptide specified by RNA complementary to the mRNA of angiotensin II (AII) inhibits the binding of AII to rat adrenal membrane receptors by the direct interaction with AII. However, these results could be reproduced neither by de Gasparo et al. [108] nor by Guillemette et al. [109]. In their response, Blalock et al. have drawn the attention to the differences between the experimental systems employed in these studies and those used in earlier experiments [110]. Further studies showed that the dissociation constant KD of AII - complementary peptide complex indicates a low mutual affinity of both components [111]. Root-Bernstein et al. found that two AII complementary peptides: Lys-Gly-Val-Tyr-Met-His-Ala-Leu (Root-Brenstein's complement) and Glu-Gly-Val-Tyr-ValHis-Pro-Val (Blalock's complement) bind to AII receptor with KD values of 5 µM and 70 nM, respectively. Thus, affinity of Root-Bernstein's complement towards the receptor is much higher than that of Blalock's complement. It was also found that in the protected form the Root-Bernstein's complement interacts directly with AII, while the Blalock's complement with AII receptor only [98]. The affinity exerted by both these complements towards AII receptors is probably due to a very pronounced homology between AII and both of its complements. The inhibitory activity of antisense peptide in respect to the contractive action of AII was indicated by Moore et al. These authors concluded that Blalock's peptide acts rather as the AII receptor antagonist and not as AII antipeptide [112]. It is worth noting that the antisense complement of AII is now commercially available (BACHEM AG). However, the experiments of Dillon, consisted in the antagonization of AII-induced contractions on rabbit aorta smooth muscle by AII complements, showed that Blalock's peptide has no clinical usefulness in inhibition AII contractions. The Root-Bernstein's peptide inhibits the contractions at high concentrations, but increases the force of muscle contraction at nanomolar concentrations, showing a dual augmentation/antagonist activity [27]. Another example of investigations directed towards the possible therapeutic use of complementary peptides is the report on the inhibitory effect of such substances on ulceration in alkali-injured rabbit cornea. A primary trigger for neutrophil invasion into the alkali-injured cornea is a tripeptide Ac-Pro-Gly-Pro. The action of complementary peptide Arg-Thr-Arg (Blalock`s peptide) leads to the reduction in neutrophils infiltration and cornea ulceration. The all-D Arg-Thr-Arg peptide showed even a somewhat higher activity than its all-L counterpart [113]. The results obtained for interleukin-1 – interleukin-1 receptor system are also worth noting. There are two isoforms of interleukin-1 (IL-1): IL-1α and IL-1β. They are a potent cytokines and play an important role in immunity,

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

521

hematopoiesis, and inflammation. In 1992 Fassina and Cassani found that a peptide complementary to 204-215 fragment of IL-1β (88-99 fragment of the mature protein) forms a pair with this fragment with a dissociation constant of ca. 10 µM. The complementary peptide column was used for the isolation of 204-215 fragment of IL-1β from a complex mixture of peptides, as well as for the purification of recombinant human IL-1β protein directly from crude E. coli lysates. It should be also mentioned that the sequence of the complementary peptide was designated by a special computer program searching for the sequences of the maximal hydropathic complementarity [103]. On the basis of the comparison of IL-1α, IL-1β, and IL1Ra (IL-1 receptor antagonist) structures Davids et al. proposed the 48-54 peptide β-bulge of IL-1β as a possible interacting site of these proteins with the receptor, and found that the antisense peptides to this protein fragment act as the “mini-receptor” inhibitors for IL-1α and IL-1β. It was found that retro-sequence of “mini-receptor” inhibitor Val-Ile-ThrPhe-Phe-Ser-Leu interacts also with IL-1α and IL-1β, but with much lower affinity than the proper heptapeptide [114]. Using the Val-Ile-Thr-Phe-Phe-Ser-Leu heptapeptide sequence as a model, Heal et al. searched for the peptide chain fragments with the similar hydropathy profiles within the molecule of interleukin-1 type 1 receptor (IL-1R1). The peptide most complementary to the β-bulge loop of IL-1 (“best fit peptide”), with a sequence Leu-Ile-Thr-Val-LeuAsn-Ile, was found to bind IL-1β and inhibit the response to IL-1 in the IL-1 induced protein synthesis. However, as shows the crystal structure of IL-1β – IL-1R1 complex, the sequence does not contact the IL-1β loop. The loop contacts a discontinuous array of amino acid residues, complementary in hydropathy to IL-1 [115]. A series of the papers concerning these questions was reviewed recently by Heal et al. [116]. Verdoliva et al. investigated selective binding of the peptide complementary to 15-27 fragment of IL-2 (Gly-PheArg-Lys-Tyr-Leu-His-Phe-Arg-Arg-His-Leu-Leu, computer designed hydropathically complementary peptide) to the native IL-2. Among the antisense peptides, its retro-analog (direction of the polypeptide chain reversed), inverso-analog (configuration on the chiral centers changed), and retroinverso analog (both changes introduced) were examined. It was found that all of the investigated peptides recognized IL2 equally well. The result suggests that the linear array of side chains, rather than the peptide backbone polarity, contributes to the recognition [117]. Kluczyk et al. synthesized the antipeptides to 143-148 sequence (Val-Thr-Lys-Phe-Tyr-Phe) of IL-1Ra and to a very similar sequence (Val-Thr-Arg-Phe-Tyr-Phe) that appears in the peptide chain of C10L protein of Poxvirus. Both Root-Bernstein and Siemion approaches were applied for the designation of antipeptide sequences. The CD spectra suggest that there may be an interaction between IL-1Ra peptide and its antipeptide constructed according to Siemion's approach, however, the complexation was not observed in MS experiments. It was also demonstrated that both the peptides and antipeptides were equally able to compete with IL-1 for its cellular receptor [43].

522

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

It was reported that the 84-97 fragment of gp160 protein of HIV-1 inhibits the infection of HIV-1 into M5-4 cells. This fragment, however, is complementary to the respective sequence located within gp41 protein of the virus. For this gp41 fragment the inhibition of infectivity was also noted [118]. As we have noted above, several reports had questioned the general applicability of molecular recognition theory and the principle of peptide – antipeptide interaction. We quoted above the work of Eberle et al. which contradicts the results of Blalock et al. on the interaction of ACTH peptide with the antisense complement [17, 18]. It must be also said that the data on the biological activity of complementary peptides should be approached with caution. There are many ways of action of definite biological effectors in the living systems. The sites of the interaction of “antipeptides” with protein molecule surfaces are, as a rule, not precisely determined. Therefore, the conclusions on the “antipeptides” activity should be considered now as a strong suggestion rather than as a conclusive statement. V. AMINO ACID COMPLEMENTARITY IN IMMUNOCHEMISTRY AND RECEPTOR STUDIES The majority of immunological applications of amino acid complementarity principle are due to the works of Blalock and his group. Earlier works related to this field were reviewed by Blalock et al. [119]. The general idea of these works is that the antibodies prepared against antipeptides (e.g. antipeptide sequences of peptide hormones) may selectively interact with peptide receptors (e.g. the receptors of peptide hormones) and/or peptide antibodies (i.e. the antibodies produced against the peptide hormones). These relations can be illustrated by the following scheme: Peptide

its receptor of peptide antibody

"antipeptide"

antipeptide antibody

Of course, the relation is based on the assumption that peptide receptor or antibody binding site is structurally related to the antipeptide, and that of antipeptide antibody to the proper peptide. The relation means that anti-idiotypic antibodies can specifically bind the idiotypic antibodies, thus, an idiotypic vs. anti-idiotypic relationship exists between the two immunoglobulins mentioned.

Siemion et al.

The observations endorsing such a hypothesis were collected and discussed in a short review of Boquet et al. [120]. This theory is strongly supported by the results of Bruck et al. [121]. These authors have shown that the antiidiotypic antibody directed against the antibody specific for the reovirus neutralizing epitope on reovirus type 3 hemagglutinin presents the internal image of the receptor binding epitope of reovirus. It is clearly evidenced by the comparison of the sequences of this epitope with the proper partial sequences of VH and VL peptide chains of antiidiotypic antibody (Fig. (7)). The relation shown on the scheme below found a support in the results of the investigations of Blalock's research group. It was shown that the antibody to the peptide which is complementary to corticotropin (ACTH antipeptide) recognized the adrenal cell ACTH receptors [5]. Based on this observation Bost et al. suggested that the antibodies against complementary peptides could be used for the easy purification and characterization of the corresponding receptors. This idea has found interesting implementations in the following years. The authors of the quoted paper also hypothesized that the ordered peptide-peptide interactions (e.g. of peptide effector – peptide receptor type) resulted from the earlier reading of both complementary DNA strands during the transcription process. This means that, according to this hypothesis, at the beginning of the history of the living organisms both DNA strands were read, what seems to be rather problematic. However, as we know, the fragments of complementary DNA strands (princoms, protein inverse complementary sequences) could be inserted into the coding strand by inversion and reinsertion mechanism. Like the duplication, it composed the important part of the mechanism for shaping genomes [122]. This may lead to the presence of both peptide and antipeptide sequences within the same protein molecules. It is believed that the selfbinding properties ascertained by antibodies specific for phosphorylcholine are probably related to the presence of complementary sequences in the antibody molecules [123]. Results similar to those described for ACTH were obtained by Carr et al. for γ-endorphin [124]. It was demonstrated that peptide complementary to γ-endorphin inhibits endorphin binding to δ-opiate receptors on neuroblastoma-glioma hybrid cells. The endorphin binding sites were also recognized by antibody to that complementary peptide. The antibody competed also with βendorphin and naloxone for the receptors on the surface of the cells. Using these antibodies immobilized on Sepharose,

Fig. (7). The comparison of the sequences of the reovirus epitope with the proper partial sequences of VH and VL peptide chains of antiidiotypic antibody The same amino acids are indicated by closed boxes, the similar by the broken boxes.

The Problem of Amino Acid Complementarity and Antisense Peptides

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

523

Carr et al. were able to isolate the receptor protein composed of four noncovalently associated subunits from sonicates of the cells.

hexapeptide sequence composed of tripeptide fragments of both VL and VH regions of the antibody as a possible mimetic of the 1-9 epitope of MBP.

The same research group produced also antibodies to the binding site of the receptor for lutenizing hormone – releasing hormone (LHRH) [125]. For this purpose a synthetic decapeptide encoded by RNA complementary to the mRNA for LHRH was used as the antigenic epitope (Blalock`s assignment). This suggests that “LHRH antipeptide” is similar to LHRH receptor binding site.

A very similar computer-assisted design of hydropathically complementary peptides was elaborated by Fassina and Melli [132]. The authors developed a procedure for the identification of interacting sites between the proteins and protein receptors. The peptide chain fragments characterized by the maximal level of hydropathic complementarity were indicated as the putative interaction sites between the two proteins. Of course, the amino acid sequences of protein ligands and their receptors must be known to apply this procedure. The method was used for the determination of interacting sites in interleukin 1β – interleukin 1β receptor system. The residues 88-99 in IL-1β and 151-162 in the receptor molecule were identified as the fragments of maximal level of hydropathic complementarity.

The most interesting possibility for the use of antipeptides for the production of anti-idiotypic antibodies consists, however, in their application in the autoimmune diseases. Autoimmune diseases are characterized by the appearance of antibodies directed against the molecules of host's own proteins. In such a case proper anti-idiotypic antibodies could, in principle, prevent the disease by neutralization of auto-antibodies. Blalock's research group has proved such a possibility with experimental autoimmune myasthenia gravis disease, evoked in rats [126]. Myasthenia gravis is caused, in part, by the production of autoantibodies against the 61-76 fragment of the α-chain of acetylcholine receptor. The immunization of rats with the peptide complementary to this fragment elicited the production of anti-idiotypic antibodies that block the recognition of native Torpedo acetylcholine receptor by its antibody. The complementary peptide was coupled to keyhole limpet hemocyanin for immunization. According to Araga and Blalock [127], this approach may provide a novel therapy for myasthenia gravis and, possibly, other B-cell mediated autoimmune disorders. Further progress of these investigations was directed towards the preparation of a peptide vaccine able to prevent experimental myasthenia gravis by specifically blocking Tcell help [128]. For this purpose Blalock et al. used the antipeptide to the 100-116 fragment of acetylcholine receptor (dominant Lewis rat T-cell receptor) conjugated to keyhole limpet hemocyanin or diphtheria toxin. The influence of an adjuvant (incomplete Freund adjuvant and aluminium hydroxide) on the immunization process was also examined. The therapeutic effect of the vaccines produced by this approach has been measured as the reversal of clinical disease progression and reduction of idiotypic antibody levels in rat model, as well as in naturally occurring canine myasthenia gravis [129]. The monoclonal antibody obtained in these studies induced significant protection and remission of clinical myasthenia gravis in Lewis rats and elicited a marked reduction in interferon-γ production in the acetylcholine receptor specific T-cell line [130]. Maier et al. developed a computer program for the identification of idiotypic – anti-idiotypic antibodies interactive determinants [131]. The program was based on the assumption that the peptide chain fragments of inverted hydropathy form potential interaction sites of the two proteins considered. The method was applied to the localization of amino acid residues involved in the antiidiotype of monoclonal antibody elicited by immunizing rats with a peptide complementary to the 1-9 fragment of human myelin basic protein (MBP). This experiment indicated a

The importance of hydropathic complementarity for the immunoreactivity of substance P-related peptides was indicated by Hanin et al. [133]. The peptides were modified in their C-terminal part so that the hydropathic profile has been conserved or changed. The main conclusion drawn from these studies was that most of the peptides with preserved hydropathic profiles but different from substance P (SP) in their primary sequences, showed significant cross reactivity with the monoclonal antibody directed against SP, although their affinities were reduced by 2 to 3 orders of magnitude. In 2003 Weathington and Blalock published a review on this subject oriented towards the problem of a rational design of peptide vaccines for autoimmune diseases [134]. The complementary peptides were also used in the investigation of experimental allergic encephalomyelitis (EAE) [135, 136, 137]. EAE is a T-cell mediated autoimmune disease consisting in demyelination of the central nervous system. It can be induced by the immunization of animals with myelin basic protein (MBP) and selected peptides, and serves as an animal model for the human disease - multiple sclerosis. The peptides complementary to 1-9 and 80-89 sequences of MBP were used in these experiments. The most interesting finding of this study was that the anti-idiotypic antibodies, stimulated by these peptides, may inhibit B-cell and T-cell activities related to the epitopes against which the idiotypic monoclonal antibodies are directed. In the recent years the humoral autoimmune response in patients with Sjögren's Syndrome was also treated in a similar way [138, 139]. This approach was also used to obtain a complementary peptide vaccine that prevents experimental allergic neuritis in Lewis rats [140]. A peptide complementary to the 60-70 sequence of the bovine P2 protein was used in these experiments. The induced antibodies inhibited the proliferation of the P2 specific T-cell line stimulated by 60-70 P2 peptide. An interesting finding resulted also from the work of de Souza et al. They used a heptapeptide Thr-Lys-Lys-Thr-LeuArg-Thr to isolate from a library of antibodies a molecule recognizing interstitial collagenase. The indicated heptapeptide was deduced as the complementary to the

524

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

collagenase-sensitive site in collagen and showed similarity in hydropathy to a sequence present in fibroblast collagenase (Ser-Gln-Asn-Pro-Val-Gln-Pro) and neutrophil collagenase (Ser-Ser-Asn-Pro-Ile-Gln-Pro). The isolated autoantibody partially inhibited the collagenolytic activity and enabled the purification of collagenase when immobilized on Sepharose column [141, 142]. It should be noted here that the inhibition of collagenase activity would be very useful in preventing tissue destruction and tumor cell invasion. A significant part of published works on anti-idiotypic antibodies presents the investigation of the peptide hormone receptors. The antibodies directed against the peptides complementary to the hormones can selectively interact with the receptors and enable their identification, characterization, and in some cases, isolation and purification. Of course, the binding sites of such anti-idiotypic antibodies do not reproduce exactly the sequence of hormone epitopes but rather mimic the biologically active conformation of the parent peptide hormone when it forms a complex with a receptor. The binding sites of these antibodies may represent the continuous peptide sequences, however, they could also model discontinuous fragments of the peptide chain, organized in the one interacting moiety. Earlier works concerning this question were reviewed in 1994 by McGuigan [143]. That paper presents the results of the investigation on the receptors of ACTH [5], LHRH [125, 144], binding domain of fibronectin molecule [93], insulin [145], arginine vasopressin [146, 147, 148], substance P [149], Met-enkephalin and γ-endorphin [150, 151, 152, 153]. A special short review by Johnson and Torres was also devoted to the use of complementary peptides in the exploration of neuropeptide receptors on lymphocytes [154]. The greatest attention in that review is given to the ACTH hormone and its receptor. However, the works on arginine vasopressin, substance P, and opioid receptors are also discussed. The work on arginine vasopressin is worth special attention. It was shown that the 3'-5' vasopressin complementary peptide blocks the arginine vasopressin helper signal for interferon-γ production by murine Tlymphocytes, whereas the 5'-3' antipeptide does not show such activity. Also the anti-idiotypic antibody produced against the 3'-5' antipeptide binds selectively to the spleen cells, i.e. the 3'-5' antipeptide could induce receptor-specific antibodies [147, 94]. Ruiz-Opazo et al. have used the DNA oligonucleotide probes coding the sequences complementary to arginine vasopressin and angiotensin II for the isolation from rat kidney a complementary DNA library of the cDNA clones corresponding to Arg-vasopressin and angiotensin II receptors. This is a very interesting extension of the molecular recognition theory into the field of applications of DNA antisense strategy. The DNA library was screened twice: first with angiotensin antisense oligonucleotide probe, followed by that of vasopressin. It was found that the binding sites for both these hormones are coded by the same DNA clone, thus that a dual angiotensin II/vasopressin receptor appears in the kidney tissue [155]. Budisavljevic et al. [156, 157] investigated angiotensin II-related idiotypic network inducing the production of polyclonal antibodies raised against monoclonal anti-AII

Siemion et al.

antibody. However, an interesting question whether these antibodies interact with AII-antipeptides was not explored in this work. The complementary peptide methodology was also used for the production of anti-idiotypic antibodies that can react with the receptors for the neuropeptide - substance P. The obtained antibodies were able to recognize specific receptors on the rat parotid cell membranes, as well as on the cells of rat cervical spinal cord, and to compete for these receptors with substance P [158]. A special attention was also devoted to fibronectin and its receptors. Fibronectin is a very prominent compound of the blood plasma. Its RGD sequence, located within the exposed peptide loop [159], interacts with such proteins as collagen, fibrin, and cell-surface receptors of the integrin superfamily, and plays a principal role in cellular adhesion phenomena. These subjects were intensively studied with the aim to develop new therapies for cardiovascular diseases, cancer, osteoporosis, and inflammation [160, 161]. In 1988 Brentani et al. used Trp-Thr-Val-Pro-Thr-Ala hexapeptide (a sequence deduced from polynucleotide sequence of rat DNA strand complementary to that coding Arg-Gly-Asp-Ser fragment) and Gly-Ala-Val-Ser-Thr-Ala hexapeptide, predicted similarly from the nucleotide sequence of human fibronectin, as the inhibitors of fibronectin binding to the receptors of MG63 human osteosarcoma cells. Anti-idiotypic antibody against the first peptide also recognized the Gly-Ala-Val-Ser-Thr-Ala sequence, and was able to bind to affinity-purified fibronectin receptors from human osteosarcoma cells [93]. The antibody against Trp-Thr-Val-Pro-Thr-Ala peptide also blocked ADP-mediated platelet aggregation, and was binding to affinity-purified platelet receptor complex glycoprotein (GP-IIIa). It is worth noting that whereas TrpThr-Val-Pro-Thr-Ala and Gly-Ala-Val-Ser-Thr-Ala peptides inhibited the interaction of fibronectin with the receptor, the analogs Gly-Ala-Gly-Ser-Thr-Ala and Gly-Ala-Arg-Ser-ThrAla were ineffective in this test [162]. These arguments for the distinct specificity of peptide – antipeptide interactions. The GP IIb/IIIa integrin is implicated as the antigen in the chronic idiopathic thrombocytopenic purpura (ITP), a frequent platelet disorder caused by the presence of antiplatelet autoantibodies. De Souza et al. demonstrated that GP IIb/IIIa positive ITP-sera can also react with Trp-Thr-ValPro-Thr-Ala peptide, therefore the peptide may be important as a possible diagnostic tool [163]. Fibronectin is also involved in the process of the release of allergic mediators from mast cells. The adhesion of bone marrow-derived mast cells to fibronectin can be induced by Mn+2 stimulation, which is antagonized by Ca+2. Using the antipeptides to the respective fragments of calcium binding EF-hands of calmodulin, Houtman et al. showed that the adhesion process can be inhibited by the peptides with inverted hydropathy [164]. The principle of hydropathic complementarity was also applied to the studies of the interactions of some bacterial proteins with their respective cellular receptors. Recently it was used by Jacchieri et al. for the determination of interacting sites of bacterial flagellin and Toll-like receptor 5

The Problem of Amino Acid Complementarity and Antisense Peptides

protein (TLR-5). It was found that the interacting sites correspond probably to 552-561 fragment of TLR-5 and 8897 fragment of flagellin connected together by hydropathic complementarity [165]. An interesting investigation was performed some years ago with the angiogenin complementary peptides. Angiogenin is a tumor-derived factor that induces neovascularization and promotes growth and metastasis of tumor cells [166]. It is a member of RNAse superfamily and binds specifically to the endothelial cell receptors of the actin type. The binding site of angiogenin consists of its 58-70 protein fragment. Both 5'-3' and 3'-5' antipeptides to that site inhibited the neovascularization promoted by angiogenin in chick chorioallantoic membrane assay and, possibly, may be used in cancer therapy [167]. It should be indicated that in this case there exists a close similarity between 5'-3' and 3'-5' antipeptides (see Table 3). A tetrapeptide fragment in the Cterminal part of the molecule is nearly the same in both antipeptides. Table 3.

Amino Acid Sequences of Human Angiogenin and its 5’-3’ and 3’-5’ Antipeptides. C-terminal Similar Residues Show in Bold

human angiogenin

GluAsnLysAsnGlyAsnProHisArgGluAsnLeuArg

5'-3' antipeptide

ValPheSerValArgValSerIleLeuValPhe

3'-5' antipeptide

LeuLeuPheLeuProLeuGlyValSerLeuLeuAspSer

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

525

confirmed this anticipation analyzing insulin, glucagon, and gastrin hormones and their respective receptors [172]. He also suggested that Dwyer's hypothesis of receptor evolution may be generalized by the assumption that the peptide complementarity served as an important factor in this process. It must be underlined here that several of these very interesting results were obtained by using Blalock’s scheme of amino acid pairing. It suggests that the Blalock’s approach to the amino acid complementarity assignments should not be prematurely disregarded. It follows from our presentation that the story of complementary peptides has been developed during the last 15 years in a broad area of investigations, possessing also some interesting practical applications. The peculiarity of these studies is that the attempts to find the practical applications for this approach are a long way ahead of theoretical studies. A lot of questions wait for their answers, the major problem being the bounds of applicability and legitimacy of the molecular recognition theory. The 5'-3' – 3'-5' reading controversy, and in regard to 3'-5' reading, the small controversy between Root-Bernstein's and Siemion's approaches also need confirmation. The thesis that the amino acid complementarity is a direct consequence of the genetic code periodicity also seems to be worth attention. REFERENCES

Another example of similar situation presents the investigation on bacterial toxin produced by Bacillus thuringensis, toxic to larvae of several insect species. A decapeptide fragment Ser-Ser-Thr-Leu-Tyr-Arg-Arg-ProPhe-Asn-Ile of the toxin molecule shows the hydropathic complementarity to Asn-Ile-Thr-Ile-His-Ile-Asp-Thr-AsnAsn decapeptide of Manduca sexta Bt-R1 receptor, suggesting that binding of the toxin to the receptor is determined by the hydropathic complementarity principle [168]. As Jacchieri et al. [165] have pointed out, up to the year 2003 there were 69 reports of the successful application of hydropathic complementarity principle to the investigation of protein complexation, and only three negative reports on this subject. This suggests that the Blalock's molecular recognition theory possesses a definite value. The presented data are also connected to the more general problem of the origin of receptor-ligand pairs in protein chemistry, and the evolution of receptor-ligand systems. According to the initial Blalock's hypothesis, the receptor sequences were encoded by antisense DNA strands, complementary to the strand coding the ligands [124]. However, this would imply a deficiency of stop codons in the complementary DNA strand, which actually does not happen [169]. According to the hypothesis of Dwyer, the protein receptor-ligand systems evolved from self-aggregating peptide complexes [170, 171]. If this hypothesis is true, the peptide receptors should contain the ligand-like peptide fragments within their binding sites. Root-Bernstein

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

Mekler, L.B. (1969) Biofizika (Russ.), 14, 581-584. Idlis, R.G. (1980) J. Wsiesojuz. Chim. Obszcz. Mendelejev's, 25, 431-434. Biró, J. (1981) Medical Hypotheses, 7, 981-993. Blalock, J.E. and Smith, E.M. (1984) Biochem. Biophys. Res. Commun., 121, 203-207. Bost, K.L., Smith, E.M. and Blalock, J.E. (1985) Proc. Natl. Acad. Sci. USA, 82, 1372-1375. Eberle, A.N., Drozda, R., Siegvist, W., Baumann, J.B. and Girard, J. ( 1989) in: Peptides, Proc. 20th Eur. Pept. Symp. Tubingen, pp. 673-675. Blalock, J.E. and Bost, K.L. (1986) Biochem. J., 234, 679-683. Blalock, J.E. (1990) TIBTECH, 8, 140-144. Markus, G., Tritsch, G.L. and Parthasarathy, R. (1989) Arch. Biochem. Biophys., 272, 433-439. Bost, K.L. and Blalock, J.E. (1989) Methods Enzymol., 168, 16-28. Lau, K.F. and Dill, K.A. (1990) Proc. Natl. Acad. Sci. USA, 87, 638-642. Yue, K. and Dill, K.A. (1992) Proc. Natl. Acad. Sci. USA, 89, 4163-4167. Kamtekar, S., Schiffer, J.M., Xiong, H., Babik, J.M. and Hecht, M.H. (1993) Science, 262, 1680-1685. Weigent, D.A., Clarke, B.L. and Blalock, J.E. (1994) Immunomethods, 5, 91-97. Blalock, J.E. (1995) Nat. Med., 1, 876-878. Kurreck, J. (2003) Europ. J. Biochem., 270, 1628-1644. Eberle, A.N., Drozda, R., Baumann, J.B. and Girard, J. (1989) Pept. Res., 2, 213-220. Eberle, A.N. and Huber, M. (1991) J. Recept. Res., 11, 13-43. Rasmussen, H.B. and Hesch, R.D. (1987) Biochem. Biophys. Res. Commun., 149, 930-938. Beattie, J. and Flint, D.J. (1992) Biochem. J., 283, 473-478. Stefani, M., Bucciantini, M., Taddei, N., Mincione, G., Manao, G. and Ramponi, G. (1994) Biotechnol. Appl. Biochem., 20, 241-249. Root-Bernstein, R.S. (1982) J. theor. Biol., 94, 885-894. Root-Bernstein, R.S. (1982) J. theor. Biol., 94, 895-904. Root-Bernstein, R.S. and Westall, F.C. (1984) Proc. Natl. Acad. Sci. USA, 81, 4339-4342.

526

Current Protein and Peptide Science, 2004, Vol. 5, No. 6

[25]

Root-Bernstein, R.S. and Westall, F.C. (1986) Brain Res. Bull., 17, 519-528. Root-Bernstein, R.S. and Dillon, P.F. (1997) J. theor. Biol., 188, 447-479. Dillon, P.F., Root-Bernstein, R.S. and Holsworth D.D. (1998) Hypertension, 31, 854-860. Siemion, I.Z. and Stefanowicz, P. (1992) BioSystems, 27, 77-84. Argyle, E. (1980) Origins Life, 10, 357-360. Pieber, M. and Toha J.C. (1983) Origins Life, 13, 139-143. Siemion, I.Z. (1994) BioSystems, 32, 163-170. Siemion, I.Z. (1994) BioSystems, 32, 25-35. Siemion, I.Z., Siemion, P.J. and Krajewski, K. (1995) BioSystems, 36, 231-238. Siemion, I.Z. (1995) Amino Acids, 8, 1-13. Bashford, J.D. and Jarvis, P.D. (2000) BioSystems, 37, 147-161. Chou, P.Y. and Fasman, G.D. (1974) Biochemistry, 13, 211-222. Burley, S.K. and Petsko, G.A. (1988) Adv. Prot. Chem., 39, 125189. Brown, B.B. and Starrett, A.N. (1975) J. Insect Physiol., 21, 1879. Siemion, I.Z. and Stefanowicz, P. (1992) Bull. Polish Acad. Sci., Chemistry, 40, 11-20. Stefanowicz, P., Zbozien, R. and Siemion, I.Z. (1998) Lett. Peptide Sci., 5, 329-331. Siemion, I.Z., Zbozien-Pacamaj, R. and Stefanowicz, P. (2001) J. Molec. Recogn., 14, 1-12. Zbozien-Pacamaj, R., Stefanowicz, P. and Siemion, I.Z. (2002) in Peptides 2002, Proc. 27th Eur. Pept. Symp., (Benedetti, E., Pedone, C., Eds.), Napoli, pp. 928-929. Kluczyk, A., Cebrat, M., Zbozien-Pacamaj, R., Lisowski, M., Stefanowicz, P., Wieczorek, Z. and Siemion, I.Z. (2004) Acta Biochim. Polon., 51, 57-66. Mandel-Gutfreund, Y., Zaremba, S.M. and Gregorat, L.M. (2001) J. Mol. Biol., 305, 1146-1159. Hutchinson, E.G., Sessions, R.B., Thornton, J.M. and Woolfson, D.N. (1998) Protein Sci., 7, 2287-2300. Lifson, S. and Sander, C. (1980) J. Mol. Biol., 139, 627-639. Wouters, M.A. and Curmi, P.M.G. (1995) Proteins: Structure, Function and Genetics, 22, 119-131. Smith, C.K. and Regan, L. (1995) Science, 270, 980-982. Raetz, C.R.H. and Roderick, S.L. (1995) Science, 270, 997-1000. Heringa, J. and Argos, P. (1991) J. Mol. Biol., 220, 151-171. Draper, K.G. (1989) Biochem. Biophys. Res. Commun., 163, 466470. Baranyi, L., Campbell, W., Ohshima, K., Fujimoto, S., Boros, M. and Okada, H. (1995) Nature Med., 1, 894-901. Baranyi, L., Campbell, W., Ohshima, K., Fujimoto, S., Boros, M. and Okada, H. (1998) Peptides, 19, 211-223. Baranyi, L., Campbell, W. and Okada, H. (1996) J. Immunol., 157, 4597-4601. Okada, H. (1998) Anticancer Res., 18, 3927-3930. Chou, P.Y. and Fasman, G.D. (1977) J. Mol. Biol., 115, 135-175. Leszczynski, J.F. and Rose, G.D. (1986) Science, 234, 849-855. Friedrich, R., Panizzi, P., Fuentes-Prior, P. and Richter, K. (2003) Nature, 425, 535-539. Guruprasad, K., Rao, M.J., Adindla, S. and Guruprasad, L. (2003) J. Peptide Res., 62, 167-174. Michel, G., Chantalat, L., Duee, E., Barbeyron, T., Henrissat, B., Kloareg, B. and Dideberg, O. (2001) Structure, 9, 513-525. Dai, S., Schwendtmayer, C., Schurmann, P., Ramaswamy, S. and Eklund, H. (2000) Science, 287, 655-658. Perry, K.M., Onuffer, J.J., Gittelman, M.S., Barmat, L. and Matthews, C.R. (1989) Biochemistry, 28, 7961-7968. Luckett, S., Garcia, R.S., Barker, J.J., Konarev, A.V., Shewry, P.R., Clarke, A.R. and Brady, R.L. (1999) J. Mol. Biol., 290, 525-533. Marx, U.C., Korsinczky, M.L., Schirra, H.J., Jones, A., Condie, B., Otvos, L.Jr. and Craik, D.J. (2003) J. Biol. Chem., 278, 2178221789. Catalano, M., Ragona, L., Molinari, H., Tava, A. and Zetta, L. (2003) Biochemistry, 42, 2836-2846. McBride, J.D., Brauer, A.B.E., Nievo, M. and Leatherbarrow, R.J. (1998) J. Mol. Biol., 282, 447-457. Favreau, P., Gilles, N., Lamthanh, H., Bournaud, R., Shimahara, T., Bouet, F., Laboute, P., Letourneux, Y., Menez, A., Molgo, J. and Le Gall, F. (2001) Biochemistry, 40, 14567-14575. Glaser, F., Steinberg, D.M., Vakser, I.A. and Ben-Tal, N. (2001) Proteins, 43, 89-102.

[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68]

Siemion et al. [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99]

[100] [101] [102] [103] [104] [105] [106] [107] [108]

Goldsack, D.E. and Chalifoux, R.C. (1973) J. theor. Biol., 39, 645651. Chipens, G.I. (1991) Ukr. Biokhim. Zh., 63, 13-20. Jones, D.J. (1972) J. Chem. Soc. Perkin I, 1407-1415. Blalock, J.E. and Bost, K.L. (1988) Recent. Prog. Horm. Res., 44, 199-222. Beattie, J. (1990) J. Endocrinol., 12, 179-181. Root-Bernstein, R.S. and Holsworth, D.D. (1998) J. Theor. Biol., 190, 107-119. Heal, J.R., Roberts, G.W., Raynes, J.G., Bhakoo, A. and Miller, A.D. (2002) Chembiochem., 3, 136-151. Jardetzky, T.S., Brown, J.H., Gorga, J.C., Stern, L.J., Urban, R.G., Strominger, J.L. and Wiley, D.C. (1996) Proc. Natl. Acad. Sci. USA, 93, 734-738. Prusiner, S.B. (2003) Prions, in Nobel Lectures in Physiology and Medicine, World Sci. Publ. Co., Singapore 2003, pp. 74-129. Rieck, R., Hornemann, S., Wider, G., Glockshuber, R. and Wunthrich, K. (1997) FEBS Lett., 413, 282-288. Gazit, E. (2002) Angew. Chem. Int. Ed., Engl., 41, 257-259. Heal, J.R., Roberts,G.W., Christie, G. and Miller, A.D. (2002) Chembiochem., 3, 86-92. Gartner, T.K., Taylor, D.B. and Derrick, J. (1994) Immunomethods, 5, 148-157. Gartner, T.K. and Taylor, D.B. (1989) J. Cell Biol., 17a, 109. Woody, R.W. (1985) The Peptides, Synthesis, Analysis, Biology, (Hruby V.J., Ed.), Acad. Press, Orlando (1985), pp. 15-114. Najem, E.S, Corigliano-Murphy, A. and Ferretti, J.A. (1989) FEBS Lett., 250, 405-410. Dillon, J., Woods, W.T., Guacello, V., LeBoeuf, R.D. and Blalock, J.E. (1991) Proc. Natl. Acad. Sci. USA, 88, 9726-9729. Fassina, G., Roller, P.P., Olson, A.D., Thorgeirsson, S.S. and Omichinski, J.G. (1989) J. Biol. Chem., 264, 11252-11257. Fassina, G., Consonni, R., Zetta, L. and Cassini, G. (1992) Int. J. Peptide Protein Res., 39, 540-548. Madhusudanan, K.P., Katti, S.B., Haq, W. and Misra, P.K. (2000) J. Mass Spectrom., 35, 237-241. Misra, P.K., Roy, R., Haq, W., Katti, S.B. and Mathur, K.B. (1996) Proc. Natl. Acad. Sci. (India), 66, 125. Kyte, J. and Doolittle, R.F. (1982) J. Mol. Biol., 157, 105-132. Shai, Y., Flashner, M. and Chaiken, I.M. (1987) Biochemistry, 26, 669-675. Shai, Y., Brunck, T.K. and Chaiken, I.M. (1989) Biochemistry, 28, 8804-8811. Brentani, R.R., Ribeiro, S.F., Potocnjak, P., Pasqualini, R., Lopes, J.D. and Nakaie, C.R. (1988) Proc. Natl. Acad. Sci. USA, 85, 364367. Johnson, H.M. and Torres, B.A. (1988) J. Immunol., 141, 24202423. Junak, M., Pavo, L. and Fahrenhol, F. (1993) J. Recept. Res., 13, 881-902. Kelly, J.M., Trinder, D., Phillips, P.A., Casley, D.J., Kemp, B., Mooser. V. and Johnston, C.I. (1990) Peptides, 11, 857-862. Curto, E.V. and Krishna, N.R. (1994) Immunomethods, 5, 98-106. Holsworth, D.D., Kiely, J.S., Root-Bernstein, R.S. and Overhiser, R.W. (1994) Peptide Res., 7, 185-193. Moulia, A., Hemmerlin, C., Phan Chan Du, A., Elhilali, Z., Tsikaris, V., Sakarellos-Daitsiotis, M., Sakarellos, C., Dotsika, E., Tzioufas, A.G., Moutsopoulos, H.M. and Manth-Thong, C. (2001) in Peptides 2000, Proc. 26th Eur. Pept. Symp., (Martinez, J., Fehrentz, J.A., Eds.), EDK, Paris, pp. 693-694. Ghiso, J., Saball, E., Leoni, J., Rontagno, A. and Frangione, B. (1990) Proc. Natl. Acad. Sci. USA, 87, 1288-1291. Chaiken, I. (1992) J. Chromatogr., 597, 29-36. Fassina, G. (1994) Immunomethods, 5, 121-129. Fassina, G., Cassini G. (1992) Biochem. J., 282, 773-779. Fassina, G., Consonni R. and Corti A. (1992) Arch. Biochem. Biophys., 296, 137-143. Fassina, G., Corti A. and Cassini G. (1992) Int. J. Peptide Protein Res., 39, 549-556. Tropsha, A., Kizer, J.S. and Chaiken, I.M. (1992) J. Mol. Recognit., 5, 43-54. Elton, T.S., Dion, L.D., Bost, K.L., Oparil, S. and Blalock, J.E. (1988) Proc. Natl. Acad. Sci. USA, 85, 2518-2522. de Gasparo, M., Whitebread, S., Einsle, K. and Heusser, C. (1989) Biochem. J., 261, 310-311.

The Problem of Amino Acid Complementarity and Antisense Peptides [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140]

Guillemette, G., Boulay, G., Gagnon, S., Bosse, R. and Escher, E. (1989) Biochem. J., 261, 309. Blalock, J.E., Elton, T.S. and Oparil, S. (1989) Biochem. J., 261, 311. Eaton, H.L., Austin, R.E., Fesik, S.W. and Martin, S.F. (1989) Proc. Natl. Acad. Sci. USA, 86, 9767-9769. Moore, G.J., Ganter, R.C. and Franklin, K.J. (1989) Biochem. Biophys. Res. Commun., 160, 1387-1391. Haddox, J.L., Pfister, R.R., Sommers, C.I., Blalock, J.E. and Villain, M. (2001) Invest. Ophthalmol. Vis. Sci., 42, 2769-2775. Davids, J.W., El-Bakri, A., Heal, J., Christie, G., Roberts, G.W. and Raynes, J.G. (1997) Angew. Chem. Int. Ed. Eng., 36, 962-967. Heal, J.R., Bino, S., Ray, K.P., Christie, G., Miller, A.D. and Raynes, J.G. (1999) Mol. Immunol., 36, 1141-1148. Heal, J.R., Bino, S., Roberts, G.W., Miller, A.D. and Raynes, J.G. (2002) Chembiochem, 3, 76-85. Verdoliva, A., Ruvo, M., Villain, M. and Fassina, G. (1999) in Peptides 1998, Proc. 25th Eur. Pept. Symp., (Bajusz, S., Hudecz F., Eds.), Budapest, pp. 806-807. Imai, M., Okada, N. and Okada, H. (2000) Microbiol. Immunol., 44, 205-212. Blalock, J.E., Whitaker, J.N., Benveniste, E.N. and Bost, K.L. (1989) Methods Enzymol., 178, 63-74. Boquet, D., Dery, O., Frobert, Y., Grassi, J. and Couraud, J.Y. (1995) Mol. Immunol., 32, 303-308. Bruck, C., Co, M.S., Slaoui, M., Gaulton, G.N., Smith, T., Fields, B.N., Mullins, J.I. and Green, M.I. (1986) Proc. Natl. Acad. Sci. USA, 83, 6578-6582. Goldstein, D.J., Fondrat, C., Muri, F., Nuel, G., Saragueta, P., Tocquet, A.S. and Prum, B. (2003) C. R. Biol., 326, 339-348. Kang, C.Y., Brunck, T.K., Kieber-Emmons, T., Blalock, J.E. and Kohler, H. (1988) Science, 240, 1034-1036. Carr, D.J., Bost, K.L. and Blalock, J.E. (1986) J. Neuroimmunol., 12, 329-337. Mulchahey, J.J., Neill, J.D., Dion, L.D., Bost, K.L. and Blalock, J.E. (1986) Proc. Natl. Acad. Sci. USA, 83, 9714-9718. Araga, S., LeBoeuf, R.D. and Blalock, J.E. (1993) Proc. Natl. Acad. Sci. USA, 90, 8747-8751. Araga, S. and Blalock, J.E. (1994) Immunomethods, 5, 130-135. Araga, S., Xu, L., Nakashima, K., Villain, M. and Blalock, J.E. (2000) FASEB J., 14, 185-196. McAnally, J.L., Xu, L., Villain, M. and Blalock, J.E. (2001) Exp. Biol. Med., 226, 307-311. Xu, L., Villain, M., Galin, F.S., Araga, S. and Blalock, J.E. (2001) Cell Immunol., 208, 107-114. Maier, C.C., Moseley, H.N., Zhou, S.R., Whitaker, J.N. and Blalock, J.E. (1994) Immunomethods, 5, 107-113. Fassina, G. and Melli, M. (1994) Immunomethods, 5, 114-120. Hanin, V., Dery, O., Boquet, D., Sagot, M.A., Creminon, C., Couraud, J.Y. and Grassi, J. (1997) Mol. Immunol., 34, 829-838. Weathington, N.M. and Blalock, J.E. (2003) Expert Rev. Vaccines, 2, 61-73. Zhou, S.R. and Whitaker, J.N. (1992) Clin. Immunol. Immunopathol., 63, 74-83. Zhou, S.R. and Whitaker, J.N. (1994) Immunomethods, 5, 136-147. Zhou, S.R., Han, Q., LaGanke, C.C. and Whitaker, J.N. (1994) Clin. Immunol. Immunopathol., 70, 251-259. Routsias, J.G., Touloupi, E., Dotsika, E., Moulia, A., Tsikaris, V., Sakarellos, C., Sakarellos-Daitsiotis, M., Moutsopoulos, H.M. and Tzioufas, A.G. (2002) Mol. Med., 8, 293-305. Routsias, J.G., Dotsika, E., Touloupi, E., Papamattheou, M., Sakarellos, C., Sakarellos-Daitsiotis, M., Moutsopoulos, H.M. and Tzioufas, A.G. (2003) J. Autoimmun., 21, 17-26. Araga, S., Kishimoto, M., Doi, S. and Nakashima, K. (1999) J. Immunol., 163, 476-482.

Current Protein and Peptide Science, 2004, Vol. 5, No. 6 [141] [142] [143] [144] [145] [146] [147] [148] [149] [150] [151] [152] [153] [154] [155] [156] [157] [158] [159] [160] [161] [162] [163] [164] [165] [166] [167] [168] [169] [170] [171] [172] [173]

527

de Souza, S.J. and Brentani, R. (1992) J. Biol. Chem., 267, 1376313767. de Souza, S.J., Madaio, M.P., Juliano Neto, L. and Brentani, R.R. (1994) Immunomethods, 5, 172-176. McGuigan, J.E. (1994) Immunomethods, 5, 158-166. Gorcs, T.J., Gottschall, P.E., Coy, D.H. and Arimura, A. (1986) Peptides, 7, 1137-1145. Knutson, V.P. (1988) J. Biol. Chem., 263, 14146-14151. Knigge, K.M., Piekut, D.T. and Berlove, D. (1988) Neurosci. Lett., 86, 269-271. Torres, B.A. and Johnson, H.M. (1990) J. Neuroimmunol., 27, 191199. Swords, B.H., Carr, D.J., Blalock, J.E. and Berecek, K.H. (1990) Neuroendocrinology, 51, 487-492. Pascual, D.W. and Bost, K.L. (1989) Pept. Res., 2, 207-212. Carr, D.J., Blalock, J.E. and Bost, K.L. (1989) Immunol. Lett., 20, 181-186. Carr, D.J., DeCosta, B.R., Kim, C.H., Jacobson, A.E., Bost, K.L., Rice, K.C. and Blalock, J.E. (1990) Neuroendocrinology, 51, 552560. McGuigan, J.E. and Campbell-Thompson, M. (1992) Gastroenterology, 103, 749-758. Campbell-Thompson, M. and McGuigan, J.E. (1993) Am. J. Med. Sci., 305, 365-373. Johnson, H.M. and Torres, B.A. (1994) Immunomethods, 5, 167171. Ruiz-Opazo, N., Akimoto, K. and Herrera, V.L. (1995) Nat. Med., 1, 1074-81. Budisavljevic, M., Geniteau-Legendre, M., Baudouin, B., Pontillon, F. and Verroust, P.J., (1988) J. Immunol., 140, 30593065. Budisavljevic, M., Ronco, P.M. and Verroust, P.J. (1990) J. Immunol., 145, 1440-1449. Bret-Dibat, J.L., Zouaoui, D., Dery, O., Zerari, F., Grassi, J., Maillet, S., Conrath, M. and Couraud, J.Y. (1994) J. Neurochem., 63, 333-343. Leahy, D.J., Aukhil, I. and Erikson, H.P. (1994) Cell, 84, 155-164. Craig, W.S., Cheng, S., Mullen, D.G., Blevitt, J. and Pierschbacher, M.D. (1995) Biopolymers, 37, 157-175. Ojima, I., Chakravarty, S. and Dong, Q. (1995) Bioorg. Med. Chem., 3, 337-360. Pasqualini, R., Chamone, D.F. and Brentani, R.R. (1989) J. Biol. Chem., 264, 14566-14570. de Souza, S.J., Sabbaga, J., D'Amico, E., Pasqualini, R. and Brentani, R. (1992) Immunology, 75, 17-22. Houtman, R., Ten Broeke, R., Blalock, J.E., Villain, M., Koster, A.S. and Nijkamp, F.P. (2001) J. Immunol., 166, 861-867. Jacchieri, S.G., Torquato, R. and Brentani, R.R. (2003) J. Bacteriology, 185, 4243-4247. Fett, J.W., Strydom, D.J., Lobb, R.R., Alderman, E.M., Bethune, J.L., Riordan, J.F. and Vallee, B.L. (1985) Biochemistry, 24, 54805486. Gho, Y.S. and Chae, C.B. (1997) J. Biol. Chem., 26, 24294-24299. Gomez, I., Miranda-Rios, J., Rudino-Pinera, E., Oltean, D.I., Gill, S.S., Bravo, A. and Soberon, M. (2002) J. Biol. Chem., 277, 3013730143. Goldstein, A. and Brutlag, D.L. (1989) Proc. Natl. Acad. Sci. USA, 86, 42-45. Dwyer, D.S. (1989) Life Sci., 45, 421-429. Dwyer, D.S. (1998) J. theor. Biol., 194, 11-27. Root-Bernstein, R.S. (2002) J. theor. Biol., 258, 71-84. Loo, J.A., Holsworth, D.D. and Root-Bernstein, R.S. (1994) Biol. Mass Spectrom., 23, 6-12.