Zilber-Rosenberg and Rosenberg acknowledge that HCE requires not only the existence of interactions between the host and the symbionts (first piece of evidence in support of HCE), but that these interactions: (1) are reliably transmitted transgenerationally; (2) affect the fitness of the holobiont. If there are host-symbiont interactions but they do not have any (or they only have one) of the two shortlisted properties, then the holobiont cannot be considered a unit of selection, but a conglomerate of independent units of selection interacting ecologically with each other (and thus, HCE would be plainly false). In their original paper, Zilber-Rosenberg and Rosenberg present evidence to support the claim that the symbionts that compose a holobiont are intergenerationally transmitted with sufficient fidelity to support the claim that holobionts are units of selection (Figure 3). It is important to note, although in passing, that Zilber- Rosenberg and Rosenberg argue that what needs to be intergenerationally transmitted, if the holobiont can be considered a unit of selection, are the different genomes that constitute the holobiont. In their words:
in chapter II, section 3, when I analyse some of the criticisms that have been raised against the hypothesis.
15 These observations, as well as the evidence that supports them, have slightly evolved with
‘The hologenome theory [concept] of evolution relies on ensuring the continuity of partnerships between holobiont generations. Accordingly, both host and symbiont genomes must be transmitted with accuracy from one generation to the next.’ (2008: 726, emphasis added)
Figure 3. Table presenting the original evidence of (1) rate of transmission of a host’s
microbiota and (2) phenotypic effects of the microbiota in the host’s phenotype, as introduced by Zilber-Rosenberg and Rosenberg. (From Zilber-Rosenberg and Rosenberg 2008: 727, Table 2).
This is important, since this is a very specific requirement about: first, the type of inheritance that is required for units of selection; second, the nature of the
holobiont, which is here conceived as a purely genomic conglomerate of interacting species.16
Zilber-Rosenberg and Rosenberg do not consider the transmission of the host genome problematic, since it relies on the well-studied rules of Mendelian inheritance. What is more problematic, though, is the transmission of the genomes of microbes that constitute the host’s microbiome. What are the
mechanisms that ensure that the same microbes will reappear
transgenerationally, to ensure the inheritance of the hologenome? Drawing upon previous work by McFall-Ngai (2002), Zilber-Rosenberg and Rosenberg distinguish direct from indirect modes of transmission of the microbes within the holobiont. The most direct case of transmission, they argue, can be found in the organelles (mitochondria, chloroplasts) of the eukaryotic cells, which are transgenerationally transmitted by cytoplasmic inheritance. A less direct, but not indirect mode of transmission occurs when the symbionts are transmitted with the reproductive cells of the host, as happens in the aphid-Buchnera symbiosis. Another still less direct mode of transmission occurs when the direct contact between the host and its offspring induces the passage of the microbes of the microbiota of the parents. This mode of transmission has been observed in humans, through the birth canal or through breast-feeding. The case of breast- feeding is similar to the cases of parent-offspring coprophagy, like the one observed in termites, where adult workers feed juveniles with feces. This mechanism is thought to be used to guarantee the acquisition of the hindgut microbiota by juveniles. A less direct mode of transmission occurs when the microbiome needs to be acquired horizontally, from the environment. Some cases of horizontal transmission, despite being almost indirect, can be very precise after all. One example is the Vibrio fischeri that bobtail squids acquire every generation to develop their light organ. Even if the transmission is not vertical, meaning that the squids need to acquire their bacteria from the environment every generation, the acquisition is very precise, and it seems that bobtail squids have developed barriers to prevent colonization from V. fischeri that do not emit light, suggesting that horizontal transmission does not
16 It is important to point this here because the notion of the holobiont that I will present later will
be different, and so will be the conception of the units of selection that I will develop, especially in chapter V, where I introduce my model of the units of selection.
necessarily prevent a faithful transmission of the microbiome that guarantees a faithful transgenerational reconstitution of the holobiont.
After they review all the possible mechanisms of hologenome transmission, Zilber-Rosenberg and Rosenberg end their section with the following highly illuminating quote:
‘The large varieties in modes of transmission have an interesting implication: individuals can acquire and transfer symbionts throughout their lives, and not just during their reproductive phase. This means that the parents, grandparents, nannies, siblings, spouses or any organism that is in close contact with an offspring can transfer symbionts and thereby influence the holobiont of the next generation.’ (2008: 728, emphasis added)
The highlighted part in their quote reflects an important assumption in Zilber- Rosenberg and Rosenberg’s first definition of HCE. According to the authors, the transgenerational transmission of the holobiont is a consequence of the transgenerational transmission of the host and microbial genomes (genomic view), and this transmission does not necessarily occur during the host’s reproductive phase, but it can occur during its whole lifespan, and be a consequence of the interaction with any member of the population (nannies, spouses, etc.), not necessarily parents. This assumption will be crucial for some of the criticisms to HCE that I will review in chapter II.
Part of the information included in Figure 3 referred to the fitness influences of the symbionts in the holobiont. I will not extend this here, since I will talk more extensively about it again in section 4 (see McFall-Ngai et al. 2013). For the moment, it is enough to say that the symbionts can have different phenotypic effects on the holobiont, including not only metabolic effects (respiration, digestion, etc.), but also behavioural effects (e.g. determining mate preference), immunological effects (e.g. protection against pathogens),
morphological effects (e.g. organ development), etc. Some of these phenotypic effects will influence the fitness of the holobiont, thus potentially generating fitter and less fit holobionts in competition with each other, as natural selection requires. Thus, according to Zilber-Rosenberg and Rosenberg, this evidence suggests that the fitness of the holobionts is comparable, which is a requirement for an entity to be a unit of selection.