Search

The first relaxin sequence published was porcine relaxin and all subsequent sequences were necessarily compared with it. Rat relaxin was purified two years later,1 and within another two years the sequence was published by John et a1.2 The sequence was significantly different from that of porcine relaxin which was surprising in light of the similarity of rat and pig insulins. Variability proved to be the rule rather than the exception for relaxin structures isolated from different animals (Fig. 4.1). Rat relaxin would eventually play a significant role in the development of structure activity relations, in particular the receptor-binding site configuration.

The relaxin work lasted through several grant applications and it has become common knowledge by now that the route from one grant to another is perilous at best. Several strategic decisions are necessary in terms of a funding agency and review committees in order to improve the odds of success. For awhile we worked with the reproductive division of Child Health and Human Services and found it extremely difficult to impress this group with the importance of relaxin for their mission!? This was a conflict that led to a temporary break in our grant support. One of all the derivatives we had proposed to synthesize, in order to further explore the importance for bioactivity of the various surface regions of relaxin, involved the replacement of the absolutely constant glycine in the A chain loop, and the comment was that “replacement of a mere glycine is hardly an adventurous proposal.” The reader may be wondering for a moment about this remark and why it made it to the headline of this chapter. There is no general correlation between the size of an amino acid and its relative importance in biology. Osteomalasia is a fatal disease caused by the exchange of only one of 900 glycines in type I collagen molecules for any other amino acid.

Rat relaxin had been known for about 18 years and had been thoroughly characterized mainly through by Sherwood et al.1–4 The relaxin used in those studies had been isolated by the method of Walsh et al in which homogenized ovaries were extracted with a mixture of trifluoroacetic acid (15%), formic acid (5%), sodium chloride (1%) and hydrochloric acid (1M).5,6 The conditions seemed rather rough but an active hormone was obtained which was used for all subsequent studies until very recently. Rat relaxin (Fig.13.1), always considered to be a relaxin of low potency, made a natural connection to our structure function studies, in particular the beneficial effect of replacement of histidine B10 in insulaxin for a glutamic acid. In the rat this position was occupied by glycine and based on a central dogma in protein chemistry, this was an unfavorable situation. Glycine is considered a helix-breaking residue and its presence in the major B chain helix right at the receptor interaction site seemed a clear indication that glycine B14 was compromising the bioactivity of rat relaxin. The opportunity was irresistible to test one’s ability to predict how certain features in a molecule would influence its bioactivity and subsequently to improve on a natural product. First rat relaxin was needed to establish a baseline and consequently a few milligrams were synthesized by the new method.7

The new chemical synthesis of human relaxin finally provided the tool to deny or confirm earlier observations with CHD-modified relaxin, and if the basic idea was correct one should be able to obtain unequivocal answers as to which of the arginines is important if not both, or whether merely a positive charge is required. Another important question pertains to the distance between the core of relaxin and the guanidino group which could be answered by incorporation of homoarginine at the binding-site.

Preliminary work on relaxin, or rather on the relaxin perimeters, had built up such a “curiosity pressure” that the trials and tribu-lations of a total synthesis became the lesser problem. When a peptide chemist faces a new synthesis he will immediately divide the problem into two major categories, one the overall plan of development of a primary structure (fragment condensation versus synthesis by sequential addition of amino acids) and secondly the decision concerning semipermanent side chain protections that must be compatible with the overall plan. If, for example, repeated exposure to weak acid is part of the synthesis of the primary backbone, acid labile semipermanent protecting groups can only be used if they require a much stronger acid for cleavage (trifluoroacetic acid vs. hydrogen fluoride) and so on. We will describe the synthesis of relaxins, including the special problems, in detail so that a student may derive sufficient information from the general concept for the synthesis of other peptides under similar conditions. For that reason a few of the approaches that did not work will be useful. In chapter 9 the problem of proper documentation of syntheses will be discussed as well as the recording of this type of research for it to become part of the serious scientific literature.

Floating on top of the world, a small world that is, one is given to making heady predictions. From what had been learned about insulin and relaxin it seemed justified to predict that, by careful compromise between relaxin-and insulin-specific requirements, a true dual hormone could be designed which would recognize the relaxin as well as the insulin receptor.

It is time to interrupt the main scenario now for an extensive description of the procedures required to demonstrate the purity of synthetic human relaxin. The physiologists should particularly appreciate a rigorous approach to this portion of the research for if there are questions about the purity or authenticity of a hormone, enzyme, factor or whatever, unequivocal interpretation of their measurements in whole animals would be even more difficult. Relaxin obliges to provide an example of such an occurrence which casts a shadow over many years of painstaking measurements in whole animals. This chapter is set aside for important technical know-how that is required to ascertain authenticity and purity of the product of a synthesis. Students of this discipline might think of it as a protein chemists’ bootcamp, designed to make him exert a great deal of effort to destroy his illusions about a “pure” product.

When biochemists use abbreviations such as this title, the mean-ing is usually discoverable by contemplation (as opposed to the meaning of abbreviations in molecular biology). ‘Retro’ means the reverse and it is occasionally used exchangeably with ‘inverto’, and both refer to the direction into which the amino acid backbone is pointing. Peptides are always written with the amino group on the left side and the carboxyl group on the right side, and each single amino acid is incorporated into this peptide in the same direction. The retro- or inverto peptide has each amino acid remaining in the same spot relative to the others, but flipped, so that the carboxyl group now points to the left. Thus, the inverto of Leu→Ala→Gly→Ile would be Leu←Ala←Gly←Ile where the arrows point from the amino terminus to the carboxyl terminus. Using the same example and the conventional way of writing the peptide sequence from the N to the C terminus the retro-form of the original peptide Leu-Ala-Gly-Ile would be Ile-Gly-Ala-Leu. Figure 15.1 explains the concept. The reader will notice that ‘retro’ does not refer to a natural property of a peptide but rather to one sequence relative to another, i.e., there is no retro peptide without the reference peptide to which it is ‘retro’.

The fetuses of a sow would have to be as long as the width of a man’s fist before the ovaries would contain significant amounts of relaxin. About 100 kg of ovaries collected under contract in an abattoir had to be discarded because the collectors had ignored the order to do this rough measurement on the fetuses. A little inconvenience had turned them into endocrinologists which made them decide that the large yellow cysts that often occurred in pig ovaries were a better sign of pregnancy. This costly mistake leaves something to be said for specialization.

The second prominent member of the murine family, and equally 1 as successful as its larger family member, is the mouse. Against the background of contemporary evolutionary philosophy it is easy to see the mouse as a small rat or vice versa and on that basis alone one would expect the relaxins between the two species to be nearly identical. There is always an experiment to ruin a beautiful concept. In this case it was the publication of a mouse cDNA that encoded a relaxin with the most drastic difference yet observed, i.e., a different distribution of the disulfide bonding pattern which had become the hallmark of the relaxin/insulin family.1 While there was a fair degree of homology between other regions of the two molecules, the C-terminal cysteine had been moved out one residue relative to other relaxins. This extended the interchain disulfide loop by one extra tyrosine residue.

Spectacular successes, alternating with equally impressive failures, have kept up expectations and, even in the mind of skeptics, enough conditioned doubts for them to recall relaxin whenever connective tissue causes problems.

At least one of the authors is old enough to remember how puzzled protein chemists were to see a distinct insulin structure in lieu of any of a other possible cross-link isomers. At that time students would also think about such problems and one of them actually presented a hypothesis concerning this phenomenon plus experiments as a research proposal to his examination committee. This was several years before Steiner and his colleagues discovered a single-chain proinsulin.1 The student’s rationale was that a monomolecular reaction (single chain) is always faster than a bimolecular one and that the folding of the single chain could force the disulfide bonds always into the same parallel direction as they are found in insulin and relaxin (which was not known at that time). This proposal hit the examining committee like Jason’s stone hit the dragon man. They were fighting among each other until they realized about an hour later that the candidate was off the hook. Diffusion-controlled folding of a single chain molecule was many orders of magnitude faster (1011 per second) than the actual process of protein synthesis which occurs at a rate of approximately one or two residues per minute in mammalian cells and up to 4 residues per second in prokaryotes. In either case a growing chain can sample all possible energy conformations in the time it takes to add another residue and can select momentarily the lowest energy valley at any stage of chain growth. This means that the product may not necessarily be in the lowest cumulative energy valley at the end of the folding process which becomes apparent as the molecule rearranges itself slightly after connecting peptides are split from the chain. The physical chemist on the committee suggested that this idea be published in one of the highly visible journals but the thesis advisor, a great person in every respect, had grown up with the standards of the Lutherans in Lake Wobegone, i.e., “publishing ideas is immodest; stick to experiments.” In retrospect this is a peculiar attitude since science is primarily about ideas.

Relaxin as referred to in this chapter is B29 porcine relaxin which has an A chain starting with arginine and ending with cysteine, and a B chain starting with pyroglutamic acid and ending with arginine in position B29 (hence B29 relaxin). Due to the presence of pyroglutamic acid in position B1, the N-terminal region of the A chain lends itself to modifications or even to total exchange. The questions to be answered then were: 1) is the A chain N-terminal region important, and if so, 2) what is its role?

The relaxin receptor has been elusive for an unusually long time 1 and still is not terribly forthcoming with detailed information. While the interaction with its natural ligand is quite strong the high nonspecific binding, which is a property of relaxin, has made the usual radioligand-mediated detection methods quite difficult. In contrast, insulin has low nonspecific binding, but shows specific binding up to 80% of the total radioactivity bound. The relaxin system gives 50% specific binding at best and usually shows around 30%, and that only after major efforts in tracer design, preparation and purification.

Clearly the receptor-interaction site is the most interesting feature of a hormone, but how would one start to look for it? We knew that at the same time several other research groups were having this discussion, among them one of the most successful biotechnology companies, and that gave our efforts a certain sense of urgency.

Testicular Leydig cells contain the message for a protein of the insulin/relaxin-like family which, in the heat of discovery, was called the Leydig insulin-like peptide (LEY I-L)1 or INSL3, corresponding to the chromosomal location of the human gene.2 The authors felt that this was a Leydig cell specific message and predicted an insulin-like character for the gene product.1 Upon close scrutiny one will notice that the critical position in the A chain-loop, which decides between insulin or relaxin-like conformations, was occupied by glycine (relaxin configuration) instead of isoleucine (insulin configuration). Furthermore, the cDNA sequence of the B-chain also showed more relaxin-like characteristics, in particular the binding-site sequence RXXXR which occurred four residues further toward the C-terminal end of the molecule (Fig. 16.1). This meant a displacement of exactly one helix turn with the arginines still projecting essentially away from the core as they do in relaxin. All this was enough reason to synthesize the molecule according to the cDNA sequence and to study its physicochemical and biological properties. Within a month this new factor was synthesized and measurements of circular dichroism as a function of wavelength showed a conformation very similar to that of porcine relaxin.3 A tracer was produced by direct iodination of a synthetic precursor as given below.