Not so capacious, after all Chandler and Turelli’s response

Brucker and Bordenstein’s experiment elicited an immediate response by James Angus Chandler and Michaelle Turelli (2014): ‘Comment on “The hologenomic basis of speciation: Gut bacteria cause hybrid lethality in the genus Nasonia”’. In their response, Chandler and Turelli made basically two

points: First, that since Brucker and Bordenstein’s data did not include any

evidence about specific incompatibilities between concrete bacterial lineages and the host’s genome, it was not conclusive to prove the existence of hologenomic speciation. Second, that the evidence that Brucker and

Bordenstein provided to support phylosymbiosis—taken by the authors as a proxy for hologenomic speciation—was weak and inconclusive. The second criticism is highly technical and concerns the type of methodologies that are required to test phylosymbiosis, so I will not review it here. I will only concentrate on the first criticism, as it is the criticism with a broader philosophical reading.

Chandler and Turelli believe that Brucker and Bordenstein’s experiment lacks the crucial test that would convincingly prove coadaptation, namely: a phylogenetically informed cross-inoculation experiment of gut microbes among the different Nasonia species. As their experimental set up was displaced, Brucker and Bordenstein proved, at most, that the gut microbiome influences hybrid lethality, but not that it is the causative agent of lethality. An alternative explanation of their results could be the following: the genome of hybrids is so dysfunctional due to their high degree of genetic incompatibility that any encounter with any free-living bacterial species (no matter its history of interaction with the host) will make it susceptible to lethality (‘intrinsic hybrid disfunction’). But, if this interpretation of Brucker and Bordenstein’s results is correct, what is the opportunity for hologenomic selection? As Chandler and Turelli put this criticism:

‘Brucker and Bordenstein’s data demonstrate that bacteria can contribute to hybrid lethality, but not because of concordant phylogenetic divergence with their hosts. The data suggest that hybrids may be generally weakened and incapable of dealing with many free-living bacteria. There are many such examples in both animal and plants. Intrinsic hybrid dysfunction is fully

consistent with the (…) [notion] that host divergence leads to defective

hybrids, without invoking coadaptation between hosts and their microbiota as a driver of speciation. (…) [T]he hologenomic conjecture that

incompatibilities between lineage-specific, free-living, horizontally

transmitted microbes contribute to speciation remains testable speculation without experimental support’ (2014: 1011a)

This criticism is more important than it might seem, since it points to one of the

most serious flaws of HCE: where to put the direction of the ‘causal arrow’.35 _Let

me explain this with a little bit more detail. Brucker and Bordenstein have shown that germ-free reared hybrid wasps do not suffer hybrid lethality, whereas conventionally reared and bacteria inoculated hybrids die. However, they have not isolated the agent that causally explains the origin of the lethality or, in other words, where the incompatibility that generates the lethality in the first place comes from. Is there any specific bacterial species, or bacterial gene, whose interactions with the host’s genome are incompatible, thus producing lethality? Only if this is plausibly shown, can it be argued that the lethality is a consequence of a host-microbe incompatibility. But the evidence that Brucker and Bordenstein present is highly unspecific and, thus, inconsequential, about the mechanistic causes that produce lethality. Insofar as their data only shows a large effect of the microbiome, the direction of the causal arrow can always be reverted: what if the observed effect (lethality in conventionally reared wasps, recovery in germ-free reared wasps) is caused by the existence of a higher degree of susceptibility in hybrid genomes when compared to non-hybrid genomes? It would be precisely the existence of the susceptibility what would explain both the observed high rate of lethality among conventionally reared wasps and the lower rate among germ-free reared wasps. It turns out that for their hypothesis about Nasonia to be conclusive, Brucker and Bordenstein need to prove not only a large effect of the microbiome on lethality, but the specific causal role of an agent.

Brucker and Bordenstein published a response (2014) right after the paper by Chandler and Turelli was published, where they defended the interpretation of the data that they had presented in their original paper. Particularly, they disagree that their experiments are compatible with Chadler and Turelli’s ‘intrinsic hybrid dysfunction’, but do not prove host-microbiome coadaptation. In their view, this is an inaccurate interpretation of their data, since it presupposes—as Chandler and Turelli explicitly state—that the bacteria

that were inoculated to F2 male hybrids in Nasonia were random species—i.e.

not necessarily phylosymbiotic. However, this was not the case, since Brucker

35 _{For a very similar criticism to holobiont research and the attribution of causal powers to the}

and Bordenstein explicitly selected bacterial species that had been isolated or

taxonomically found in the wasps. Thus, Chandler and Turelli’s argument that

‘any bacteria’ can cause hybrid lethality is unjustified and, more importantly, the observation of that result would be completely irrelevant to discard adaptation. Why do Brucker and Bordenstein believe the latter to be so? Because, in their opinion, that ‘holds the ambiguous equivalence that any foreign mitochondria from different animals can cause mortality’ (2014: 1011b). Let me expand on the latter point a bit. The idea to grasp here is the following: to test the possibility that mitochondria are coadapted to their hosts in a way that breaks down in hybrids it is not necessary to discard the possibility that a randomly inoculated mitochondrion from any other species would have a similar effect on hybrids than the effect of a mitochondrion from the original species. It is enough to show the effect of an inoculation with a mitochondrion from the original species. And the same applies to testing the effects that certain genes might have on hybrids. In other words, Chandler and Turelli are demanding more evidence to accept a case of hologenomic speciation than the evidence they would demand for other cases (influence of a gene, influence of mitochondrion), and this seems to be unjustified from the perspective of Brucker and Borenstein. As they claim, what they are proposing is changing the ‘framework for studying the basis of any reproductive isolation mechanism’ (2014: 1011b). And either the new framework is accepted, or the evidence that will be demanded to prove its validity will always be biased from the perspective of the

previous framework.36

3.4.5. Pre/post-mating isolation and the hologenome concept of evolution. A