On the Origins of New Forms of Life

7.10: Hybrid Infertility

EUGENE M. MCCARTHY, PHD GENETICS

Site Search

(Continued from the previous page)

Hybrid Infertility. Various theories have been offered to explain why hybrids are commonly less viable and/or fertile than their parents. One explanation, which makes sense with regard to the production of inviability, as opposed to infertility, is that hybridization combines at random two distinct genetic programs, which may interact in an inharmonious way. Here the causative factors are clear enough — production of organisms by a trial-and-error approach of this sort would be expected often to lead to bad results. The combination in a single organism of two genomes that are separately functional will frequently (but not always) result in adverse interactions because each of the genes in one genome must be compatible with all the genes in the other. Stabilization theory concurs with neo-Darwinian theory in explaining hybrid inviability in this way. Therefore, if one is attempting to judge the two theories on the basis of their explanatory powers, one cannot discriminate on this basis. However, as we shall see, the phenomenon of hybrid infertility is a different matter. In this latter case, stabilization theory provides a much more satisfactory explanation than does neo-Darwinian theory.

A frequently encountered explanation of the pervasiveness of hybrid infertility suggests that natural selection acts directly to increase reproductive isolation. In this scenario, two distinct somasets are derived from a single ancestral somaset by divergence in geographic isolation. This process, known as "reinforcement", supposedly occurs when the two somasets come back into contact ("secondary contact"). At this stage, natural selection is said to favor those individuals that mate with their own kind. Thus, it is said, reproductive isolation eventually becomes complete, as hybrids become less viable and fertile. Reinforcement, however, is controversial (Butlin 2005). Even those who accept the idea of reinforcement would admit many biologists do not accept it. Selection for infertile, inviable offspring is, in fact, a contradiction in terms — natural selection is a process favoring traits that help, not hinder, the production of offspring. Indeed, no explanation, accounting in terms of natural selection, for the general finding that hybrids are typically of reduced fertility is accepted by all biologists.

But many do accept an explanation not based on selection, the Dobzhansky-Muller (D-M) model (Bateson 1909; Dobzhansky 1937a; Muller 1942). This explanation, however, is seriously flawed. An explicit presentation of the D-M model would require the introduction of a number of technical concepts. But non-geneticists can simply ignore such technicalities and consider the D-M model as a simple black-box process. For those who wish it, Coyne and Orr (1997) give a concise description of this model, quoted in the footnote below. The process there described produces two genetically distinct populations. Coyne and Orr say it is "entirely possible," if individuals from those two populations interbred, that the resulting hybrids "could be sterile" (this is the sort of hypothetical language typically used in connection with the D-M model).

However, there is really no reason for us to expect such hybrids to be sterile. Since the D-M model is supposed to explain a general phenomenon (hybrid sterility), it should identify a general causative mechanism. An analogy will clarify: If we wished to take a boat out for a cruise, it would be "entirely possible" that it would sink. But in the absence of known causative mechanisms making such an event likely or inevitable (e.g., enemy gunboats on patrol, gale warnings, an incompetent captain), we would have no reason to expect it to occur. If shipwreck is not the expected outcome under such circumstances in even one case, then, certainly, we would have no reason to expect shipwreck to a regular outcome of going for a boat ride.

The same reasoning applies to the D-M model. True. It describes a process that could produce sterile hybrids. But its proponents fail to explain why we should expect it to do so. If it were somehow true that two distinct types gained an advantage by producing sterile offspring together, then we would expect natural selection to favor the production of sterile . We would therefore expect sterility generally to be a trait of hybrids. However, the D-M model eschews selection and it specifies no other causative mechanism. If, in the case of any particular cross, hybrid sterility is shipwreck, where is the gunboat that will make such cases likely? For, given that most crosses produce hybrids of reduced fertility, there must be some factor that makes it likely for crosses to have this characteristic. Under stabilization theory the gunboat is the widespread occurrence of karyotypic differences distinguishing even closely related forms (see Section 3). Under such circumstances hybrids are expected to be structural heterozygotes, which are typically less fertile than their homozygous parents. Structural heterozygosity is known to disrupt meiosis and interfere with the production of gametes.¹ Here the mechanism causing hybrid infertility is clear. Hybrids commonly have impaired fertility because they are often structural heterozygotes.

But in the case of the D-M model the causative mechanism (i.e., the mechanism making the production of such crosses not merely possible, but likely) is not at all clear. Proponents of the model claim a sufficient degree of genetic difference will prevent two individuals from producing fertile offspring together. But this is merely an assumption, not an explanation. It is not immediately clear why infertility should be produced by the combination of two genetically distinct individuals. After all, if any two individuals are not a pair of clones, they will differ genetically. So virtually all individuals that mate to produce fertile offspring do differ genetically. Since this is the case, why should we suppose additional randomly accumulated differences in genes would result in a pair producing sterile offspring? Neo-Darwinians fail to explain how these additional genetic differences would cause the production of sterile offspring with disrupted meiosis and few viable gametes. They also fail to explain why we should expect genetically distinct populations to produce infertile hybrids. Really, this seems to be little more than a vague, old idea that even predates any precise science of genetics. For example, according to Darwin (1872), the German botanist Max Ernst Wichura (1817-1866) maintained the "view of the sterility of hybrids being caused by two constitutions being compounded into one." This sounds well enough. But what exactly does it mean?

Proponents of the D-M model seem to confuse cause with correlation. True, a hybrid between parents highly distinct at the genetic level is — all other things being equal — more likely to be sterile than one whose parents are genetically alike. But this is only correlation. Karyotypic differences, too, are usually greater when parents are more distinct at the genetic level (Ferguson-Smith et al. 1998; Ruvinsky and Graves 2005: 352). Under such circumstances, one needs to decide which of the two phenomena is causes hybrid infertility. Is it genetic difference? Or is it karyotypic difference? Night correlates with day because it always follows day. But day does not cause night. The rotation of the earth causes both day and night. In the case of karyotypic differences, the causative mechanism is clear, well known, and well documented: recombination of rearranged homologs during meiosis in structural heterozygotes disrupts the process that produces gametes (Grant 1985; Griffiths et al. 1999; White 1973a: 215–232). It therefore reduces the fertility of structurally heterozygous hybrids.

On the other hand, while there may be adverse effects of purely genetic differences on fertility, there seem to be few, if any, well-documented cases. Studies purporting to demonstrate this phenomenon rarely control for the possibility that karyotypic differences are causing the sterility attributed to genetic differences. No one knows how genetic differences between the parents of hybrids might produce a disruption of meiosis in the hybrids themselves. Certainly, the idea that genetic difference in parents leads, in itself, to a disruption of meiosis in their hybrids has not been shown empirically to be a general phenomenon. Rather, the existence of such a general phenomenon is a mere inference drawn by neo-Darwinians on the basis of theory. How this happens, they cannot say. So there is no reason to suppose parental genetic differences should regularly cause sterility in hybrids. The D-M model therefore provides no real explanation of the fact that the typical hybrid is of reduced fertility.

Nineteenth century naturalists used the term "physiological species" to refer to forms that were unable to interbreed due to physiological incompatibility. Darwin's inability to account for such forms was a point of concern even to his supporters. When Darwin first proposed his theory, Huxley adopted it only "subject to the production of proof that physiological species may be produced by selective breeding" Huxley (1898: 150; see also: Huxley 1901: 257). Three decades later he was still of the same opinion: In a letter dated May 17, 1891, Huxley writes,

Even today, this proof has not been forthcoming.

Darwin himself concluded "after mature reflection" (Darwin 1872: 248) that natural selection could not account for the evolution of the "physiological species." Thus, in the sixth edition of the Origin (1872: 262) he states that "the sterility [characteristic] of first crosses and of their hybrid progeny has not been acquired through natural selection." Nor did he believe selection could cause forms to develop an inability to produce F₁ hybrids. Thus, in a letter to Alfred Wallace dated April 6, 1868, he writes:

In response, Wallace wrote (ibid, p. 297) "I will say no more, but leave the problem as insoluble, only fearing that it will become a formidable weapon in the hands of the enemies of Natural Selection." And yet Darwin did, in fact, think the generality of the phenomenon indicated "the cause, whatever it may be, is the same or nearly the same in all cases" (Huxley 1901: vol. II, 14).² In this regard, even Darwin posited an unknown force. He also refers to this mystery factor in The Descent of Man (1871: vol. I, 222–223, footnote):

Thus, among those who think in terms of the neo-Darwinian paradigm, there has been longstanding controversy over the causative factors underlying the general phenomenon of hybrid sterility. Neo-Darwinian explanations of this phenomenon are convoluted, logically flawed, and disputed. To those biologists weary of the intricacies of this debate, stabilization theory offers a clear, brief explanation of the general phenomenon of hybrid infertility: Populations treated as distinct species often belong to distinct chromosets. When individuals with different karyotypes mate, the resulting hybrids are structurally heterozygous. As we have seen, structural heterozygosity commonly has an adverse effect on fertility.^† It is for this reason that populations tend to break up spatially into karyotypically pure chromosets. Being like other members of the population pays a large reproductive dividend (any immigrant or mutant individual with an aberrant karyotype will not find a mate of its own kind and so will produce structurally heterozygous progeny of low fertility).

Karyotypic differences, then, can be used to explain why hybrid infertility is such a widely observed phenomenon — so long as it is assumed new forms commonly come into being via stabilization processes involving chromosomal mutations. The evidence presented thus far in this book strongly suggests that the origin of new forms by this means is indeed typical.

Origin of Karyotypic differences. And why do such karyotypic differences between populations arise in the first place? It is because the karyotype of each population comes into being independently of the karyotypes of other populations. Within a given sexual chromoset there is selection for a single uniform karyotype because variation of karyotypes results in the production of structural heterozygotes of low fertility. However, within a second chromoset there can be selection for uniformity with respect to some other karyotype. Thus, the karyotype of one population can evolve independently of the karyotype of another. As a result, natural selection for uniformity within each group stabilizes two different karyotypes. So there is nothing to coordinate the structures of the two resulting karyotypes, and they will likely differ in some, or many, respects.

Hybrids between such populations, then, are typically structural heterozygotes. Therefore, their fertility is expected to be reduced. Their sterility is not selected for in any direct way; rather it is incidental to the selection for the fertility associated with karyotypic uniformity within populations. Since populations treated as distinct species often have distinct karyotypes, matings between such populations often produce relatively infertile hybrids. It is unknown how many hybridizing taxa differ in this way. But there are certainly more than enough cases for sterility of this type to be perceived as a general phenomenon. Thus, the basis of the general phenomenon of hybrid sterility is clear under stabilization theory, but nebulous under orthodox theory. It can be stated concisely: Many, though by no means all, sexual somasets treated as separate species are distinct chromosets, so hybrids between such somasets are often structural heterozygotes of reduced fertility. Their sterility is an incidental result of selection for fertility (karyotypic uniformity) within populations.

Note: Coyne and Orr (1997) give the following, typical account of the D-M model: "If postzygotic isolation is based on incompatibilities between two or more genes, hybrid sterility and inviability can evolve unimpeded by natural selection. If, for example, the ancestral species had genotype aabb, a new mutation at one locus (allele A) could be fixed by selection or drift in one isolated population because the Aabb and AAbb genotypes are perfectly fit. Similarly, a new allele (B) at the other locus could be fixed in a different population since aaBb and aaBB genotypes are also fit. But it is entirely possible that when the AAbb and aaBB populations come into contact, the resulting AaBb hybrids could be sterile or inviable. The A and B alleles have never been 'tested' together within a [single individual], and so may not function properly in hybrids."

1. Darlington (1932, 1937); Delneri et al. (2003); Ford and Clegg (1969); Forejt (1996); Grant (1985); Griffeths et al. (1999); Gustavsson (1990); ; King (1980); King (1993: 164–168); Lewis (1966); McCarthy (2006); McClintock (1945); McCarthy et al. (1995); Piálek et al. (2001); Searle (1993); Stebbins (1950: 218–226, 1971); White (1973: 219, 1978).

2. Elsewhere Darwin (1872: 248) writes “Take the case of any two species which, when crossed, produced few and sterile offspring; now, what is there which could favour the survival of those individuals which happened to be endowed in a slightly higher degree with mutual infertility, and which thus approached by one small step towards absolute sterility? Yet an advance of this kind, if the theory of natural selection be brought to bear, must have incessantly occurred with many species, for a multitude are mutually quite barren.”