NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Madame Curie Bioscience Database [Internet]. Austin (TX): Landes Bioscience; 2000-.

Cover of Madame Curie Bioscience Database

Madame Curie Bioscience Database [Internet].

Show details

Experiments in Evolution


Conceptual science interfaces with the experimental world as predictions emanate from the internal logic of a paradigm. Hypotheses concerning the origin and unfolding of life have their first brush with reality when they are held up against fossils. Molecular genealogy is the first prediction evolving from the Darwinian paradigm of “descent with variation” and fossil and protein lineage must confirm each other if the original idea is correct. Evolution by random mutations amounts to an intrusion of divinity into science and it is satisfying that the concept does not yield acceptable results. The next prediction of the old model pertains to the mechanism of the evolution-induced genomic changes and that is amenable to experimental testing. New technology has made it possible to progress to a deeper level of the genealogy, the rationale for mutation acceptance.

The accumulation of the primordial genome is a chance event by any hypothesis whereby one must grant the possibility that there may have been a tendency to form a certain sequence faster than another, but that problem is left waiting until one knows how primordial condensation occurred. In any case, the genomist claims that all variations observed today are due to prebiotic events1 in contrast to the old model.2 Proteins would change continuously in the Darwinian system with survival as the only selective force. Proteins in this type of study come, for obvious reasons, exclusively from survivors. The testable aspect of the hypothesis is the proposal that functionally important amino acids remain constant in a protein and that functionally unimportant ones are subject to mutational replacement.

For easy reference the primary sequences of homologous proteins (relaxins for example) from different species are shown side by side for counting the number of differences.3 Here significant assumptions enter the picture, namely that all molecules have once been alike and that differences in amino acids in defined positions in a protein are due to mutations.4 Amino acids that remained unchanged are important for biological functions because their exchange would have more or less severely compromised the owner of such mutants . The primary structures of a series of relaxins shown in Fig. 1 have been aligned at the cysteine (C) residues (cross-links) which are identical in all relaxins and insulins except for mouse relaxin.3

Figure 1. This figure displays the majority of known relaxin molecules.

Figure 1

This figure displays the majority of known relaxin molecules. The sequences are aligned at the cross-linking cystein residues which are constant except for the C-terminal cystein in the mouse relaxin A-chain. The length of the chains is not of functional (more...)

The letters stand for specific amino acids and the identities of letters in certain positions are scored when we talk about levels of identity between two or more proteins. One then observes which positions (say 12, 13, and 17 in the B chain) have identical amino acids in all relaxin molecules. These are called “preserved” in the old model and that goes along with the assumption that all relaxins come from one ancestor gene and that all uncritical residues have changed due to mutational events.5

Furthermore, the neo-Darwinian hypothesis says that the insulin gene had been duplicated long ago and that the “left-over” copy has mutated into a relaxin.6 Once the astronomical number of mutations had led to an active relaxin any further mutation that would hit the invariant positions would have killed or severely handicapped the owner of that hormone, and therefore all constant residues are important. My colleague Dr. Erika Büllesbach has developed an ingenious as well as practical way to synthesize relaxin and insulin for our NIH-funded research.3 These derivatives allowed us to experimentally test some of the postulated mechanisms in the Darwinian model of molecular evolution.

Human relaxin was synthesized with either L-alanine or the unnatural D-alanine, substituting for the constant glycines in position 12 in the B chain, and we found that the modified molecules were just as active as the native ones. Furthermore, the constant glycine in B24 could be replaced by alanine with only minor disturbance!3 Yet only once so far has alanine been observed in a natural relaxin (hamster relaxin). From these experiments it follows that the glycines in position B12 and B24 must be constant for reasons other than functionality (Fig. 1).

Next, glycine 14 in the A-chain loop (Fig. 1) was exchanged for isoleucine (a large molecule) and this change inactivated the hormone.7 Here then was a case where a constantly appearing residue was also required from a functional point of view. However, probing further, a relaxin with an alanine (slightly larger than glycine) in that position would function very well! Again, only once do we see another fairly small amino acid (serine) in tammar relaxin in that position but alanine, a quite common amino acid in proteins, has not yet been seen in position A 14.

The arginines in the B chain, however, told a different story. They could not be replaced by anything without destroying the hormonal activity of relaxin. The constant arginines in the B chain proved to be the receptor-interacting site and to remove them and replace them with anything else would be like grinding the serrated edge of a key; the hormone would no longer interact with its receptor.8 Here then is a case where the constant residues are required for functionality, and our experiments have therefore led to the clear-cut conclusion that functionally important residues are constant but that not all constant residues are functionally important! The clarion is muffled when one learns that humans apparently do not need relaxin so that the relaxin molecule should be under no “pressure” to retain its active structure. The argument that an arginine to X exchange in relaxin would kill (a human) is not tenable! What keeps nonfunctional residues constant and what keeps functional residues constant when the whole molecule is not important for survival? Biology seems to run on rails, the controls that keep us alive keep us from changing, and they do so for important and unimportant amino acids equally well. The experiments favor the Genomic Potential Hypothesis, which places the creation of variety into the prebiotic era and makes biology steady.

Were an inactivating mutation to hit insulin, the victim would die, but by no stretch of the imagination can one extrapolate this observation to relaxin. Furthermore, insulin has at least 35% replaceable residues but differs only very little among mammals (except for the guinea pig). In this case a large number of ‘replaceable residues’ have remained constant without “evolutionary pressure” to do so and that also points to stability as the ground cord of life.

Experimenting with whale and pig relaxin, we found the two molecules to be all but identical which was sensational against the background of high variability of relaxins among terrestrial mammals.9 Because the pig and whale fossils are already distinct in 50 million-year-old layers we looked for a reason why, in spite of numerous changeable positions, their relaxins have not changed during that period with respect to each other. In fact, the relaxins of so-called closely related animals are so dramatically different that the pig/whale similarity is truly astounding. This observation invited another set of experiments which, for convenience, was targeted to the N-terminal A chain end of pig relaxin10 (Fig. 1). After careful chemical dissection of the first seven residues that had remained completely constant between whale and pig, the molecule was inactive. We then synthesized the N-terminal pentapeptide of insulin and coupled it to the inactive truncated relaxin, and noted that the relaxin activity returned. In an even more drastic experiment the truncated relaxin was coupled to a penta-alanine peptide and that too was sufficient to restore the biological activity of relaxin. The functional requirement was merely that a chain be present that could form an a-helix. Between whale and pig the relaxin N-terminal ends of the A chain had remained constant for 50 million years, ostensibly without interbreeding, even though this region of the molecule could have mutated to many amino acids without destroying the function of relaxin.

Meanwhile we discovered porcine relaxin in tunicates (Chapter 13) and again one wonders what keeps these amino acids constant against the background of purportedly constant mutational activity which is credited with creating all the variety of life on earth?11 The gene for “porcine relaxin” was known since the very beginning of the age of animals; it did not need to be invented by mammals, and again, the minimal message is that fossil and molecule genealogy do not match.

Nature is an honest adversary. She never lies but often hedges, and like a good witness, she will never say more than the question warrants. In the case of mouse relaxin she was explicit. When one compares the mouse relaxin in Fig. 1 with rat relaxin one notices that the C-terminal cysteine in the mouse hormone A chain is displaced outward by one residue.12 This change makes mouse relaxin different from any other relaxin.

The impact of this finding on the Darwinian picture of the evolution of rodents is quite interesting. The story suggests that physical obstacles once upon a time caused the original primitive rodent ancestor population to split into hares, squirrels, the guinea pigs, and the muridae (mouse and rat). These events are thought to have occurred many million years ago, but the murine split into rat and mouse was a relatively recent event. Of course, one would notice that the segregation scheme is kind of compromised by the fact that the three taxa continued to co-exist in the same landscape, but we will ignore that to keep the story flowing. Guinea pig relaxin was synthesized in our laboratory and found it to be very active in that species.3 We synthesized rat relaxin as well, which proved to be one of the very active relaxins, and both of these rodent molecules have the regular disulfide bond structure. Since both rat and guinea pig are purportedly ancestral, a normal disulfide bond pattern is the older one and the mouse, after separating from the rat, must have converted a good relaxin gene to one with an extra residue in the A chain (according to neo-Darwinists). We have synthesized the regular mouse relaxin and noticed relatively low bioactivity in the mouse. Curiosity caused us to synthesize the mouse relaxin minus the tyrosine (Y) (Fig. 1) that had purportedly been inserted after the mouse/rat separation.12 Biological activity measurements in mouse tissue showed clearly that the synthetically “reverted” mouse relaxin is superior to the real mouse relaxin. According to the old paradigm the mouse gene had suffered an insertion mutation and the inferior relaxin had propagated against selection pressure through the whole mouse taxon to the exclusion of the better gene. Darwinian selection would, by definition, always drift toward the better molecule which is how evolution to complexity is envisioned.13 Here nature tells us, with rare frankness, that that idea is wrong! Genomists agree, the mouse never had a different gene, and that gene was sufficient to provide us with all the mice we need.

Neither proteins nor the encoding genome have sensors for needs or directionality. In contrast, life has developed an unmatched system of control and repairs which keeps the genome free of mismatches. That status quo-protecting system is not judgmental as concerns the quality of a gene. Conversely, the genome has no means of eliminating weakly active molecules or to improve them, but because of the natural redundancy of functions in all living systems loss of a particular function on account of mutational activity can, in some cases, be overcome. This has become particularly clear with the advent of gene knock-out technology which has enabled us to remove an “important” gene only to find that in some cases the animal survives happily via compensatory functions. A whole gene has been kept constant without a need for it in the organism.

All those basic residues in the relaxin A chains (R and K in Fig. 1), the high isoelectric point (high net positive charge), was all of that necessary? Human relaxin was synthesized with all four basic residues in the A chain, replaced by the neutral unnatural amino acid citrulline. The modification lowered the isoelectric point into the acidic range but had no effect on the relaxin activity! The relaxin receptor recognized this molecule as well as an unmodified relaxin, and the idea of evolutionary pressure causing all relaxins to retain global molecular features, such as the isoelectric point, seemed seriously unconvincing.

Almost all cells in an organism have insulin receptors as well as receptors for many other hormones, but relatively few have relaxin receptors. The process whereby the specific receptor distribution came about is far beyond our present understanding of biology. It has been suggested in the literature that the insulin gene long ago duplicated to give rise to a relaxin gene via a stream of mutations.6 A new function , however, is thought to be targeted, i.e., the receptor should be present and the hormone should mutate to match this target (or vice versa).14 That means that one of the members of the receptor/hormone pair mutates to a new function without a target.

Not really, they say, because at the same time and in the same cells, the insulin receptor gene duplicates and when only this duplicate picks up mutations to produce a relaxin receptor, we have our target. Meanwhile the old insulin receptor stays intact because all cells need carbohydrate metabolism, for example. Clearly, this “conjugated miracle” model of targeted development cries for divine intervention, and it is impossible to point out how many non-contiguous and anticipatory changes had to occur to bring these events to a proper conclusion. If one would want to back away from divinity and argue chance, such as in mutations, one would need to consider the basic reality of this process. For the 50 some amino acids of insulin it would take 20 × 1050 trials to explore all random possibilities on the way to a relaxin structure. Receptors contain more than 1000 amino acids so that the conversion of the insulin (or any other) receptor to a relaxin receptor would require about 20 × 101000 trials. This would mean about 4 × 101050 trials for the development of this receptor hormone pair. Probabilities such as 1 in 101,000 simply mean that such an event would not happen. The number of failures would fill the universe many times.

What kind of phenomenon could keep nonessential residues constant over millions of years if change is the stew that nourishes novelty?15 The Genomic Potential Hypothesis says that there is no stew, stability is the standard and changes are accidents. How is it possible that many proteins in different species look alike to some extent without functional needs to keep them so? The new view is that functions are recruited from untargeted pools of similar nucleic acid seqences that can appear in many organisms and that do not change during development. Some proteins taken from bacteria or protozoa are active in the mammalian cell and even a transcription factor from archaebacteria interacts properly with human DNA.16,17 Some proteins have a mosaic pattern of runs of 20–30 amino acids identical to homologous proteins in other species, and other stretches of amino acids that are matching several stretches of proteins with totally different functions.18,19 These observations, I think, are a telltale sign of a generic origin of life, a clonal affair based upon nucleic acid chemistry that coalesced in many places into biogenic droplets like fog condenses on the ground.

All observations such as gene duplications and redundancies are primordial chemical events that have been converted to mutations, gene duplications, and lateral gene transfers by the Darwinian/neo-Darwinian ideology.

These experiments had been performed in my laboratory but the literature is full of such examples. Why do others not see the same phenomenon and acknowledge the problem? Perhaps they were primed to look at their results from the platform of the old prejudices. Frank Plumpton Ramsey20 and his theory tells us in brief that any large enough, apparently random, collection of items will contain an orderly substructure, and that the complexity of possible substructures depends upon the number of members of a basic set.

Lately the age of the genetic code was researched21 using a statistical analysis of tRNA sequence relatedness. Data in the literature plus assumptions about the inverse of the mutation accumulation rate (sort of a biological Hubble constant22) were used to deduce a focal point where all tRNA sequences would become one and the same, the point where the genetic code was created. The time at which this happened clearly depended upon the meaning given to the sequence differences between the contemporary tRNAs, which in turn depended upon the paradigm that had produced the differences via mutations.

The uniformity of the genetic code in all living creatures examined so far has always been interpreted as evidence for a single-point origin.21 The paper offers the conclusion that the genetic code might be older than life itself, but not so old that one would need to presume an extraterrestrial origin. The conclusion supports the Genomic Potential Hypothesis (Chapter 6), but it does not follow from the paper unless the Darwinian basis is taken as self-evident.

The author of the Science paper admitted, in an answer to my argument, that an original distribution of tRNAs, just as they are seen today, would invalidate his conclusions, but he was sure that “God is not malicious”.

The sequence differences between the many tRNAs that these investigators examined are a fact. The evolutionary distances between the species, i.e., the phylogeny of eubacteria, archaebacteria, and eukaryotes, are based on the idea that they were derived from each other by mutations. This means that the tRNA distances, indicated by the sequence differences that are supposedly an indicator of branching, are in fact a restatement of a hypothesis. Thus, we have no independent means other than a paradigm to tie together the various tRNAs into a common ancestor which most likely never existed. This intrusion of reality made no impression upon the investigators who assert that: “kinship relations are revealed by alignment of sequences”, whereas only similarities are revealed by such a comparison. Kinship is a derived property that comes from a paradigm; there are no independent means whereby one can determine how these tRNAs came about, there are no fossil molecules.

These studies, like those mentioned in the previous chapter, have appeared in a very prestigious journal where many of these types of papers are published because editors are under palpable “paradigm pressure” and are often uncomfortable in matters of epistemology. What makes investigators accept a model so uncritically and, more importantly yet, what makes them force fit data into such a model? Here, I think, a biological uncertainty principle comes into play, which states that our mind cannot view a set of data without at once seeking for an underlying order. More often than not the order is achieved according to an internal (intuitive) set of parameters, and when mental coziness spreads, stars and molecules will move into the proper positions to satisfy our pictures. Ramsey's theory explains this phenomenon.20

The reader has witnessed a head-on collision of experimental evidence with a major postulate of the neo-Darwinian model. As a consequence, differences are no longer mutations but rather have reverted to just differences. Both, gene duplications and mutations, are mimicked by primordial variability that was stabilized by the constraints of biology.


Schwabe C. Evolution and chaos. Computers Math Applic. 1990;20:287.
Benton MJ. Diversification and extinction in the history of life. Science. 1995;268:52. [PubMed: 7701342]
Schwabe C, Buellesbach EE. RelaxinSpringer Verlag, Heidelberg, Germany1998.
Mayr E. Darwin's influence on modern thought. Scientific American. 2000:79.
Samols D, Thornton CG, Murtif VL. et al. Evolutionary conservation among biotin enzymes. J Biol Chem. 1988;263:6461–6464. [PubMed: 2896195]
Dayhoff MO. Atlas of protein sequence and structure. The national biomedical research foundation. 1972;vol 5
Bullesbach EE, Schwabe C. Structural contribution of the A-chain loop in relaxin. Int J Pept Protein Res. 1995;46:238–243. [PubMed: 8537177]
Bullesbach EE, Yang S, Schwabe C. The receptor-binding site of human relaxin II. A dual prong-binding mechanism. J Biol Chem. 1992;267:22957–22960. [PubMed: 1331071]
Schwabe C, Buellesbach EE, Hayne H. et al. Cetacean relaxin: Isolation and sequence of relaxins from Balaenoptera acutorostrata and Balaenoptera edeni. J Biol Chem. 1989;264:940. [PubMed: 2910872]
Buellesbach EE, Schwabe C. Relaxin structure: N-terminal A-chain helix. J Biol Chem. 1987;262:12496. [PubMed: 3624270]
Georges D, Schwabe C. Porcine relaxin, a 500 million-year-old hormone? The tunicate Ciona intestinalis has porcine relaxin. FASEB. 1999;13:1269–1275. [PubMed: 10385617]
Bullesbach EE, Schwabe C. Mouse relaxin: synthesis and biological activity of the first relaxin with an unusual crosslinking pattern. Biochem Biophys Res Commun. 1993;196:311–319. [PubMed: 8216305]
Mayr E. Evolution. Sc Am. 1978;239:46–55. [PubMed: 360394]
Shapiro JA. Adaptive mutation: who's really in the garden? Science. 1995;268:373. [PubMed: 7716540]
Ohno S. Dispensable genes. Trends in Genetics. 1985;1:160.
Garrett RA, Dalgaard J, Larsen N. et al. Archaeal rRNA operons. TIBS. 1991;16:22. [PubMed: 1905072]
Pennisi E. Static evolution. Science News. 1994;145:168.
Baron M, Norman DG, Campbell ID. Protein modules. TIBS. 1991;16:13. [PubMed: 2053133]
Yonekura H, Nata K, Watanabe T. et al. Mosaic evolution of prepropancreatic polypeptide. II. Structural conservation and divergence in pancreatic polypeptide gene. J of Biol Chem. 1988;263:2990–2997. [PubMed: 3343236]
Graham RL, Spencer JH. Ramsey's theory. Sci Am. 1990;263:112.
Eigen M, Lindemann BF, Tietze M. et al. How old is the genetic code? Statistical geometry of tRNA provides the answer. Science. 1989;244:673. [PubMed: 2497522]
Osterbrock DE, Gwinn JA, Brashear RS. Edwin Hubble and the expanding universe. Scientific American. 1993;84:&#0.
Copyright © 2000-2013, Landes Bioscience.
Bookshelf ID: NBK6358
PubReader format: click here to try


Related information

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...