Development and evolution

In the nineteenth century, ideas about development, heredity and evolution were inextricably mixed up, because it seemed natural to suppose that changes that first occurred in development could become hereditary, and so could contribute to evolution. This was not only Lamarck’s view but Darwin’s, expressed in his theory of pangenesis. Weismann liberated us from this confusion, by arguing that information could pass from germ line to soma, but not from soma to germ line. If he was right, geneticists and evolutionary biologists could treat development as a black box: transmission genetics and evolution could be understood without first having to understand development. Since Weismann, developmental biology has had only a rather marginal impact on evolutionary biology. One day, we have promised ourselves, we will open the box, but for the time being we can get along very nicely without doing so. Recent progress in developmental genetics, some of which has been reviewed in the last three chapters, oblige us to reopen the question. In fact, there are three related questions, not one. The first, which is most relevant to the theme of this book, is the ‘levels of selection’ question: why does not selection between the cells of an organism disrupt integration at the level of the organism? This is the topic of section 15.2. The second is the problem of the inheritance of acquired characters. This old problem has reappeared in a new guise. We now recognize the existence of cell heredity, mediated by different mechanisms from those concerned with transmitting information between generations. In section 15.3, we discuss whether cell heredity plays any role in evolutionary change. Finally, in sections 15.4 and 15.5, we ask whether recent molecular information sheds any light on another old problem—that of the extraordinary conservatism of morphological form, maintained despite dramatic changes of function. This conservatism has led anatomists to identify a small number of basic archetypes, or bauplans. There is little doubt that conservatism is real. Consider, for example, the fact that bones and cartilages, which in humans serve in swallowing, sound production and hearing, are derived from elements of the gill apparatus whereby our fish ancestors exchanged gases with seawater, and, before that, in all probability, from elements of a filter-feeding apparatus.

Development in simple organisms

Complex multicellular organisms, whose bodies consist of differentiated cells of many kinds, have evolved independently on three occasions—animals, higher plants and fungi. In addition, multicellular organisms with a lesser degree of cellular differentiation have evolved on a number of occasions. For example, the algae have given rise to ‘seaweeds’ several times. In this and the next three chapters, we discuss the origin and subsequent evolution of such organisms. Some 540 million years ago, at the beginning of the Cambrian, there appeared an array of multicellular marine animals, including the major phyla that exist today—coelenterates, platyhelminths, annelids, arthropods, molluscs, echinoderms and others. Chordates are also present in the Cambrian: they are not known from the earliest deposits, in which only hard parts are preserved, but are present in the slightly later Burgess Shale, in which soft-bodied forms are preserved. Forty years ago, this sudden appearance of metazoan fossils was not only a puzzle but something of an embarassment: the absence of any known fossils from earlier rocks was a weapon widely used by creationists. Today, the fossil evidence for prokaryotes goes back 3000 million years, and for protists some 1000 million years. The Cambrian explosion remains a puzzle, however, which has been only fitfully illuminated by the discovery of the enigmatic soft-bodied Ediacaran fauna, which had a worldwide distribution between 580 and 560 million years ago. There are still doubts about how these fossils should be interpreted (Conway Morris, 1993). The orthodox, and more plausible, view is that the fauna is dominated by coelenterates, with some specimens identified as echinoderms and annelids. An alternative interpretation (Seilacher, 1992) is that they belong to an extinct clade of multicellular eukaryotes, the ventobionts, probably lacking an alimentary canal, muscles and nervous system. Although such organisms may have existed, at least some of the Ediacaran fauna have been successfully compared to recent metazoans. If the interpretation of most of these fossils as coelenterates proves to be correct, it would fit in well with the morphological and molecular evidence. The molecular data suggest that coelenterates arose early, but probably not independently of other metazoans. Morphologically they are simple in being diploblastic (formed from two cell layers), in contrast to the triploblastic animals that predominate in the Cambrian.

The evolution of templates

In this chapter, we discuss the origin and early evolution of genetic replication. The argument is complex, so we start with a brief outline. Section 4.2 discusses the nature of replication. We draw a distinction between simple replicators, limited hereditary replicators and indefinite hereditary replicators. Continued evolution requires indefinite hereditary replicators: it seems that such replicators depend on some form of template reproduction. In section 4.3, we point out that there is an error threshold for the accuracy of replication: for a given total quantity of genetic information—for example, for a fixed number of bases—there is an upper limit on the error rate of replication. If the error rate rises above this limit, natural selection cannot maintain the information. This leads to what we have called Eigen's paradox. In the absence of specific enzymes, replication accuracy is low. Hence the total genome size must be small—almost certainly, less than 100 nucleotides. The genome is therefore too small to code for accurate replication machinery. There is a catch-22 situation: no enzymes without a large genome, and no large genome without enzymes. The next three sections discuss possible solutions to the paradox. Section 4.4 considers populations of replicating RNA molecules. We point out that the dynamics of replication are such as to lead to the stable coexistence of a diverse population, but we do not think that this constitutes a solution to the paradox. Section 4.5 discusses the hypercycle, a particular relationship between replicators that makes it possible for a greater total quantity of information to be maintained with a given accuracy of replication. We argue that the further evolution of hypercycles requires that they be enclosed within compartments, because otherwise they are sensitive to parasitic replicators. We also discuss, rather inconclusively, the possibility that, even in the absence of compartments, cooperation might evolve, by a processes analogous to kin selection, if the components of the hypercycle were confined to a surface. Finally, we discuss an alternative model, the stochastic corrector model. This also depends on the existence of compartments, but emphasizes the importance of stochastic effects arising if there are small numbers of each kind of molecule in a compartment. Essentially, small numbers serve to generate variation upon which selection can act.

What is life?

Imagine that, when the first spacemen step out of their craft onto the surface of one of the moons of Jupiter, they are confronted by an object the size of a horse, rolling towards them on wheels, and bearing on its back a concave disc pointing towards the Sun. They will at once conclude that the object is alive, or has been made by something alive. If all they find is a purple smear on the surface of the rocks, they will have to work harder to decide. This is the phenotypic approach to the definition of life: a thing is alive if it has parts, or ‘organs’, which perform functions. William Paley explained the machine-like nature of life by the existence of a creator: today, we would invoke natural selection. There are, however, dangers in assuming that any entity with the properties of a self-regulating machine is alive, or an artefact. In section 2.2, we tell the story of a self-regulating atomic reactor, the Oklo reactor, which is neither. This story can be taken in one of three ways. First, it shows the dangers of the phenotypic definition of life: not all complex entities are alive. Second, it illustrates how the accidents of history can give rise spontaneously to surprisingly complex machine-like entities. The relevance of this to the origin of life is obvious. In essence, the problem is the following. How could chemical and physical processes give rise, without natural selection, to entities capable of hereditary replication, which would therefore, from then on, evolve by natural selection? The Oklo reactor is an example of what can happen. Finally, section 2.2 can simply be skipped: the events were interesting, but do not resemble in detail those that led to the origin of life on Earth. There is an alternative to the phenotypic definition of life. It is to define as alive any entities that have the properties of multiplication, variation and heredity. The logic behind this definition, first proposed by Muller (1966), is that a population of entities with these properties will evolve by natural selection, and hence can be expected to acquire the complex adaptations for survival and reproduction that are characteristic of living things.

Gene regulation and cell heredity

Two cellular mechanisms are essential for development. The first, gene regulation, makes it possible to switch on different genes in different cells, in response either to conditions external to the cell or to the activity of other genes within the cell. The second, cell heredity, ensures that these states of gene activity, once induced, can be stably transmitted through cell division, without the need for the continued presence of an external inducer. In this chapter, we describe how gene regulation and cell heredity are achieved in metazoans, and point to some similar mechanisms that are already present in prokaryotes. The central problem of gene regulation was posed, in a social context, by the scholastic Master Eckhardt: ‘Quis custodiet ipsos custodes?’ [Who regulates the regulators?] Clearly, the proposition that every gene needs a separate regulator gene leads to an infinite regress. There are various ways of resolving the paradox, which include one regulator controls several other genes, including regulators; one gene, even a regulator, is controlled by several other genes; and some genes may be both regulatory and structural. Plenty of examples are known for each case. It is also necessary that some genes be regulated by signals from outside the cell. The essential mechanism of gene regulation was discovered by Jacob & Monod (1961; Fig. 13.1) in E. coli. A regulatory gene codes for a protein, which, by binding to a specific regulatory sequence of another gene, alters the activity of that gene (negatively in the case originally described by Jacob & Monod, but the effect can also be positive). The regulation can be modified by a specific inducing molecule that alters the effect of the regulatory protein by binding to it allosterically. It is interesting that these two properties of regulatory proteins—that they can recognize specific regulatory sequences, and that their effectiveness can be altered by binding allosterically to inducers—are already present in prokaryotes. The complexity of multicellular eukaryotic development requires that an average gene be controlled by many others. Whereas regulatory elements in bacteria are usually simple switches, eukaryotes tend to have ‘smart’ genes, controlled by a complex of several regulatory proteins (Davidson, 1990; Beardsley, 1991).

Symbiosis

The establishment of a permanent and obligate coexistence of genetic entities that were once capable of independent existence played an important part in the origin of the eukaryotes, and, if our earlier speculations are correct, in the origin of cells and chromosomes. In this chapter, we discuss other examples of symbiosis. The term is used to include all cases in which two or more different kinds of organism live in close association: thus it extends from parasitism to mutualism. Mutualism has been defined as a relationship from which both partners benefit. However, as will become clearer below, it is hard to measure, or even to define, ‘benefits’: in what sense is a mitochondrion today better off than its once free-living ancestors? The two questions that we shall ask are: • What are the selective force acting on the two partners in present-day symbioses? • Could such selective forces lead to the establishment of permanent and obligate coexistence? First, however, we review briefly some of the ecologically more important symbioses (for further examples, see Pirozynski & Hawksworth, 1988; Margulis & Fester, 1991). We mention only a fraction of the known mutualistic associations. Others, including cases of interaction between animals and prokaryotes, are discussed below. It is striking that symbiotic relationships have been important in the utilization by plants of nutrient-poor soils, the colonization of bare rock, life in deep-sea vents, the construction of coral reefs, and the utilization of plant material by several groups of insects. Sonea (1991; see also Sonea & Panisset, 1983) has pictured the world of bacteria as a single superorganism, whose individual component cells rely for their survival on ecological exchange of metabolites, and on genetic exchange via plasmids and phages. This picture has the virtue of emphasizing the important role played by plasmids and temperate phages in conferring on individual bacterial cells capacities needed in particular environments—for example, resistance to antibiotics, tolerance of heavy metals and new metabolic abilities. But the picture suffers from the drawback that is fatal to all holistic models of evolution, from the Gaia hypothesis downwards, of losing all sight of the units of selection, and hence of lacking any model of the dynamics of evolutionary change.

The origin of eukaryotes

The basic structures of a bacterial and a eukaryotic cell are shown in Fig. 8.1. The differences whose origins call for an explanation are as follows: • The bacterial cell has a rigid outer cell wall, usually made of the peptidoglycan, murein. In eukaryotes, the rigid cell wall is not universal, and cell shape is maintained primarily by an internal cytoskeleton of filaments and microtubules. • Eukaryotic cells have a complex system of internal membranes, including the nuclear envelope, endoplasmic reticulum and lysosomes. • Bacteria have a single circular chromosome, attached to the rigid outer cell wall. In eukaryotes, linear chromosomes are contained within a nuclear envelope, which separates transcription from translation: communication between nucleus and cytoplasm is via pores in the nuclear envelope. • Eukaryotes have a complex cytoskeleton. The actomyosin system powers cell division, phagocytosis, amoeboid motion and the overall contractility to resist osmotic swelling. Microtubules and the associated motor proteins (kinesin, dynein and dynamin) ensure the accurate segregation of chromosomes in mitosis, ciliary motility and the movement of transport vesicles. Intermediate filaments form the structural basis for the association of the endomembranes and nuclear-pore complexes with the chromatin to form the nuclear envelope, while other intermediate filaments help to anchor the nucleus in the cytoplasm. One crucial difference between prokaryotes and most eukaryotes has been omitted from Fig. 8.1: this is the presence of mitochondria, and, in plants and algae, of chloroplasts. The reason for the omission is that, on the scenario for eukaryote origins that seems to us most plausible, these intracellular organelles originated later in time than the structures shown in the figure. The differences between these cell types justifies the recognition of two empires of life (above the kingdom level): Bacteria and Eukaryota (Cavalier-Smith, 199la; Table 8.1). (It is interesting that this taxonomic rank was recognized by Linnaeus.) Within each of the empires, there are two major categories: Bacteria consist of the kingdoms Eubacteria and Archaebacteria, and Eukaryota are divided into the superkingdoms Archaezoa and Metakaryota. The justification for these divisions is as follows. The Archaebacteria, in contrast to the Eubacteria, never have murein cell walls, and their single cell membrane contains isoprenoidal ether rather than acyl ester lipids.

The origin of protocells

Cellularization has the following main aspects that we have to explain: • The need for active (self-generated) compartmentation when metabolism is liberated from the surface. • The origin of membranogenic molecules and membranes. • The origin and mechanism of spontaneous protocell fission. • The transportation problem. Simple membranes are not ‘leaky’ enough to permit important nutrients to pass through. • Were the first protocells autotrophs or heterotrophs? The evolution of the first autocatalytic metabolic cycle. • The iron-sulphur world and the RNA world: are they mutually exclusive or complementary? • The problem of the origin of the two membranes of negibacteria, the most ancient existing group of organisms. • The origin of chromosomes and DNA synthesis. We shall discuss these problems in turn. As we discussed before, the prebiotic pizza has the ability to localize metabolites and genes. This is advantageous for two reasons: Reactants are kept in each other’s proximity, which ensures that reaction rates will be high enough and that important compounds do not drift away. Genes will interact, directly (e.g. by influencing each other’s replication) or indirectly (by catalysing steps of metabolism), only with their neighbours: selection will thus be able to ensure cooperation among genes that would otherwise compete against each other. Life liberated itself from surfaces a long time ago. Somehow, passive localization must have been replaced by an active process of membrane generation, maintenance and fission. The basic structure of contemporary biomembranes is as follows. There is a molecular bilayer of lipids, to which proteins are attached in various ways. The bilayer is formed because the membrane constituents are so-called amphipathic molecules: they have a hydrophilic head and a hydrophobic tail. Since the binding interaction of water with itself is much stronger than that between water and hydrophobic compounds, the latter are expelled by water as much as possible; this results in tails coming together. A simple sheet of bilayer would still be not at the energy minimum because its edges would be exposed to water. An energetically favourable solution is the formation of a lipid vesicle.

Introduction

Living organisms are highly complex, and are composed of parts that function to ensure the survival and reproduction of the whole. This book is about how and why this complexity has increased in the course of evolution. The increase has been neither universal nor inevitable. Bacteria, for example, are probably no more complex today than their ancestors 2000 million years ago. The most that we can say is that some lineages have become more complex in the course of time. Complexity is hard to define or to measure, but there is surely some sense in which elephants and oak trees are more complex than bacteria, and bacteria than the first replicating molecules. Our thesis is that the increase has depended on a small number of major transitions in the way in which genetic information is transmitted between generations. Some of these transitions were unique: for example, the origin of the eukaryotes from the prokaryotes, of meiotic sex, and of the genetic code itself. Other transitions, such as the origin of multicellularity, and of animal societies, have occurred several times independently. There is no reason to regard the unique transitions as the inevitable result of some general law: one can imagine that life might have got stuck at the prokaryote or at the protist stage of evolution. There are obvious difficulties in discussing unique events that happened a long time ago. How can we ever know that our suggested explanations are correct? After all, historians cannot agree about the causes of the Second World War. We accept that certainty is impossible, but there are several reasons why we think the enterprise is worth while. First, we have one great advantage over historians: we have agreed theories both of chemistry and of the mechanism of evolutionary change. We can therefore insist that our explanations be plausible both chemically, and in terms of natural selection. This places a severe constraint on possible theories. Indeed, the difficulty often lies, not in choosing between rival theories, but in finding any theory that is chemically and selectively plausible. Further, theories are often testable by looking at existing organisms.

The origin of language

The past 30 years has witnessed a debate between the holders of two very different views about how humans are able to talk. The behaviourists, following B. F. Skinner, argue that we learn to talk in the same way that we learn any other skill. Children are rewarded when they speak correctly, and reproved when they make mistakes. We can talk, whereas chimpanzees cannot, because we are better at learning: there is nothing special about language. In contrast, Noam Chomsky and his followers have argued that humans have a peculiar competence for language, which is not merely an aspect of their general intelligence. We learn to utter, and to understand, an indefinitely large number of grammatical sentences, and to avoid an even larger number of ungrammatical ones, so we cannot possibly learn which sentences are grammatical by trial and error. Instead, we must learn the rules that generate grammatical sentences. These rules are of great subtlety, so that, although we acquire and apply them, we cannot formulate them explicitly. For example, consider the two following sentences: How do you know who he saw? (1) Who do you know how he saw? (2) How do you know who he saw? Who do you know how he saw? Every speaker of English knows at once that is grammatical, and is not. But what rule tells us this? No-one but a trained linguist would have any idea, any more than a non-biologist would know how the rate of beating of the heart is adjusted to meet changing demands. In section 17.3, we describe a hypothesis about the rule that tells us that is ungrammatical: it is a subtle rule, but as yet no-one has thought up a simpler one. It is hard to believe that we could so painlessly master such rules unless we were genetically predisposed to do so. More generally, it is still beyond the wit of linguists and computer scientists to write a language-translating programme, yet many 5-year-olds know two languages, do not mix them up, and can translate from one to the other. A second reason for thinking that we cannot learn to talk by trial and error lies in the poverty of the input on which a child must rely. After hearing a finite set of utterances, a child learns to generate an indefinitely large number of grammatical sentences. This implies that the child learns rules, and not merely a set of sentences.