Ladies and gentes: Maternally inherited DNA and ancient honeyguide host races

Avian interspecific brood parasitism occurs when a female bird uses a different host species to incubate and rear her offspring. Studies of this parasitic behavior have provided some of nature’s most compelling examples of coevolutionary interactions (1). Avian brood parasites exhibit wide variation in host specificity, ranging from one or a few hosts to hundreds, and there is a concomitant degree of coevolution of traits between parasite and host. For example, within many species of cuckoos, and some species of cowbirds, parasite eggs visually and morphologically match those of their hosts. If there is variation in egg characteristics among different host species, and selection promotes egg matching to avoid host recognition and rejection of parasitic eggs, this can lead to the evolution of host-specific races of parasites (called gentes).

The evolution of gentes within species was considered problematic until it was determined that, unlike mammals, female birds have heterogametic sex chromosomes (i.e., females have 1 Z chromosome and 1 W chromosome, whereas males have 2 Z chromosomes). Thus, the genes that determine egg morphology and host nest selection behavior might reside on the female-specific W chromosome and not be available to recombine with genetic material from males (2). However, not until the recent advent of molecular tools has there been evidence for the role of female control of host egg matching or for the time frame of host-race evolution. Thus far, research has revealed somewhat conflicting evidence for direct female control of egg and nestling traits, and it has suggested very recent (<100 kya) host-race divergences in cuckoos and cowbirds (3–5). Now, however, using an mtDNA analog for the W chromosome, a study in PNAS by Spottiswoode et al. (6) on the brood parasitic greater honeyguide describes an unexpected ancient divergence among gentes and greatly expands our understanding of how such host races can evolve.

Among the 100 species of obligate brood parasites (about 1% of the roughly 10,000 described avian species) are the more familiar cuckoos and cowbirds, but less common examples include African viduid finches, a South American duck, and the relatively unstudied honeyguides (1). The honeyguides (Indicatoridae) consist of 17 species in four genera, with most species occurring in Africa and 2 species that leak out into southern Asia. Although drab and unremarkable in plumage, honeyguides are among the most interesting of birds because of two unusual features of their life history and associated behaviors (7). First, as their name implies, some honeyguide species are known to recruit and guide a select group of mammals, including humans (Homo sapiens), and perhaps ratels (Mellivora capensis), to beehives. The mammal will break open the hive, enabling it to retrieve the honey, leaving the honeyguide with a meal of wax and bee larvae (but apparently not the honey) from the honeycomb. Honeyguides are one of a few types of birds that practice cerophagy (wax eating). They have been suggested to use specialized enzymes and/or bacterial gut symbionts to assist in wax digestion (7, 8), but this has not been confirmed with recent experimental work (9).

Second, all honeyguides are obligate brood parasites, laying their eggs in the nests of a range of bird species, mostly those nesting in cavities. Female honeyguides will usually puncture a host egg before parasitizing the host nest with their own single white egg. They have another unusual adaptation for ensuring success of their young: Unlike many cuckoo nestlings, which will eject a host egg or nestling from the nest, honeyguide nestlings have a very sharp bill hook that they use to lacerate and kill host nestlings in the nest (7, 10). The “murderous” hook is present in very young chicks and is lost during development of the bill after about 2 wk of age.

Spottiswoode et al. (6) document that individual greater honeyguides that parasitize different host species show significant differences in egg size or shape that tend to match the size and shape of eggs of their hosts. However, unlike many other brood parasites, greater honeyguides do not need to match host egg color because they parasitize species that nest in low-light environments (tree cavities or terrestrial burrows) and color would be difficult to discern. In addition, the authors find two highly differentiated lineages of mtDNA amplified from within a single population of greater honeyguides in Zambia, and each lineage always corresponds to hosts with different nest types: One lineage is always found in nests made in tree cavities (e.g., those of hoopoes, different types of woodhoopoes and a single kind of kingfisher), and the other lineage is always found in nests in terrestrial burrows (mostly bee-eaters and another type of kingfisher; Fig. 1). They also show the same two lineages in other regions from across much of their range (e.g., Ghana). Because mtDNA in birds is inherited through the female line only, this reflects a mother-to-daughter-to-granddaughter inheritance pattern, with no input of DNA from males. When the authors apply an approximate and conservative calibration of mtDNA sequence evolution to the divergence between the two lineages, they find that the divergence time between the two mtDNA lineages has been at least 3 million years.

Fig. 1.

Shown is a schematic of the pattern of mtDNA sequence divergence in greater honeyguides that parasitize two types of hosts: those that nest in cavities in trees and those that nest in burrows in the ground [as described by Spottiswood et al. (6)]. MtDNA divergence indicates a surprisingly long time period since separation of female lineages, whereas similarity in nuclear genes suggests mating at random with respect to host nest type. Thus, selection of different host nest types and similarity in egg morphology are likely to be under female genetic control. Pictured are the greater honeyguide (Middle, Indicator indicator), little bee-eater (Top, Merops pusillus), and African hoopoe (Bottom, Upupa africana) (Top and Bottom, photographs courtesy of Mark Anderson; Middle, photograph courtesy of Warwick Tarboton.)

What makes the findings of Spottiswoode et al. (6) all the more exciting is that they find virtually no divergence at all in DNA sequences from genes present in the nucleus (i.e., genes that are descended from both the mother and father). This indicates that the divergent host-specific female lineages are exchanging nuclear genes seemingly at random with males;

Differentiated host races have been maintained in the greater honeyguide for several million years across most of its African range.

thus, the genetic control of the host specificity and morphological differences must be inherited only through the maternal line. Although they did not examine sequence from the W chromosomes of females, the mtDNA and W chromosomes are inherited essentially as a unit from mother to daughter; thus, variation in one is linked to variation in the other. In addition, the mtDNA genome has only a small number of genes, none of which are likely related to egg traits or laying behavior. Thus, the implication is that autosomal genes are not playing a guiding role in determining the female traits associated with brood parasitism in greater honeyguides. This has not necessarily been the story in cuckoos (3, 5) and cowbirds (4, 11), in which egg characteristics do not always assort with mtDNA type or autosomal gene sequences in gentes can differ along with mtDNA sequences. Next-generation sequencing methods should help immensely in the development of W-chromosome markers and may provide a means for direct determination of the genes responsible for these traits.

Also of significance is the conclusion that differentiated host races have been maintained in the greater honeyguide for several million years across most of its African range, without any evidence of reproductive isolation among the forms. There have been cases reported where vastly divergent mtDNA lineages have been maintained within populations of a single species with no evidence of nuclear DNA divergence [e.g., in ravens in western North America (12, 13)]; however, in these cases, there is often evidence that the pattern derived from secondary contact and subsequent breakdown in reproductive isolation among forms differentiated in allopatry. In the case presented by Spottiswoode et al. (6), there is very good evidence that this is not the case and that the divergent host races they find in the greater honeyguides arose in sympatry. Thus, these deeply divergent gentes probably evolved through the action of females and their daughters selecting and adapting to particular host nest types across millions of generations while mating at random with respect to the host origin of their male partners.