Host specificity

The basic question for this week is: what determines parasite specificity? We can define specificity at a number of levels, from the breadth of a parasite's choice of spatial position within a host organ such as the gut, to differences in parasite performance in different host subpopulations (this connects to last week's RQ discussions), to the range of species that parasites are able to exploit.

The evolutionary reasoning and consequences behind parasite specificity are interesting and important. Parasite specificity for hosts has connections to the entirely general question in evolution of why there are species at all: why don't we just have one big life form that is good at everything? What is the balance between plasticity (the capability to express different phenotypes in different environments) and specialization?

Some preliminaries:

Encounter and compatibility filters

Combes is big on the idea that parasite success, and parasite specificity, are determined by these two processes or "filters". I don't think it explains everything, but it is a useful distinction.

Encounter filter: this filter determines whether the parasite can ever find the host. It is affected by the large-scale spatial distribution of the host and parasite (geography); the small-scale spatial distribution (habitat use); the behavior of H and P (e.g. circadian rhythms/chronobiology); the ecology of the host (e.g. for trophically transmitted parasites, whether the host ever eats the upstream host). Both parasites and hosts can "try" to open or close the encounter filter by modifying their behavior to chase or avoid each other.
Compatibility filter: determines whether the parasite succeeds in infecting the host and reproducing once it makes contact with the host. Biochemistry, immune system, etc..

Sampling issues

How do we really know what the host range of a parasite is? Often we try to draw conclusions about specificity from large databases assembled from the literature about the number of hosts that have been recorded for different kinds of parasites.

Sampling: if some kinds of species are systematically undersampled (e.g. species that live in inaccessible habitats, species that don't parasitize humans), then we could draw misleading conclusions.
Cryptic species: since parasites are often morphologically simple (often all we have to go on are mouthparts and genitalia, since everything else is unnecessary), parasites specific to two or more different hosts are often lumped into a single species. Equally, generalist parasites on different hosts could be misidentified as separate species (e.g. on the basis of plastic traits or morphological adaptations) even when there is gene flow between the subpopulations. Obviously, mistakes like this will bias our conclusions about specificity. Combes gives some examples.

Why specialization?

We can broadly divide explanations into two categories, adaptive and isolation/drift. Adaptive explanations are more closely connected with microevolutionary processes (the change of gene frequencies within populations), while isolation/drift arguments are more closely connected with macroevolutionary patterns (the patterns of species and higher-level taxa). Speciation is the link between micro- and macroevolution. Everything in evolution is either too big and slow, or too small and fast; studies of microevolutionary processes must be extrapolated to see if they match macroevolutionary patterns, while studies of macroev. patterns must be consistent with microev. processes.

Basic adaptive explanations for specialization:

There may be basic physiological tradeoffs that constrain the viability and fecundity of parasites on different hosts; you can do better by specializing on one than by generalizing on many (analogous to the frequency-independent lottery model for maintenance of sexual reproduction). Specialization may be increased by feedback, as parasites choose to restrict themselves to a specific host.
Tradeoffs may be frequency-dependent ("tangled bank") instead; specialists may be able to compete better with other parasites within a host
The advantage to specialization, and disadvantage to generality, may be in your ability to adapt to your host: the Red Queen returns

Isolation/drift arguments say that specificity is not driven by tradeoffs between hosts, but by the random loss (via mutation or drift) of the ability to utilize a host that is never seen by the parasite population.

Levels of specialization

Specialization can occur at the level of:

organs, or parts of organs, within a host;
individual hosts, in cases where there are multiple generations of parasites on a single host (Karban, "Fine-scale adaptation of herbivorous thrips to individual host plants": Nature, 340, 60-61), or (possibly) HIV within hosts
families or local populations: in measles epidemics in West Africa, there is some evidence that the first ("index") case in a family is milder than subsequent infections in the same family. This may be evidence that the virus is adapting to the shared portions of the family's genotype (although there are other explanations, such as the possibility that healthier children with better immune systems are more active outside the household and thus more likely to acquire infection).
species

We will mostly be discussing the latter two cases (local populations and species), although we will come back to #1 later when we discuss within-host competition.

Processes

Rather than looking at patterns at the broad scale of host and parasite phylogenies, and making inferences about what kinds of jumps and extinctions have happened in the past, we can look directly at the ecological or microevolutionary processes that lead to the different patterns (cospeciation, host jumps, extinction, etc.)

Examples:

Diplozoon gracile (monogenean) may be in the process of jumping back to Barbus barbus, in places where there is hybridization between B. barbus and B. meridionalis. The compatibility filter is open, as proved by experimental infection of B. barbus in the lab; the difference in infection rates is (presumably) not caused by genetic differences. Combes' explanation for the observed pattern that parasitism is correlated with meridionalis genes is that the encounter filter is related to behavior (habitat choice), which in turn is determined by genetics. (Comparisons of parasite load in hybrids can be tricky, since hybrids often have screwed-up immune systems, but this conclusion seems reasonable.)
Lepeophtheirus europaensis (copepod) on flounder and brill; the compatibility filter is open (cross-infection occurs in the lab) and the encounter filter is apparently open, since the two species live in similar habitats, but parasites are not found crossing over. This appears to be through "choice", since parasites have lower reproductive output on the "wrong" hosts. Is this speciation in action?
Howardula aeronymphus (nematode on Drosophila): by maintaining the parasite on one species in the lab, it loses the ability to infect other species (J. Jaenike, "Rapid evolution of host specificity in a parasitic nematode", Evolutionary Ecology 7, 103-108). This has important implications: it gives some concrete evidence for genetic divergence of isolated parasite populations. This raises the question of what is happening here: how, exactly, is the parasite losing its ability to infect the other species? Is this caused by the relaxation of tradeoffs, or just by random mutation in (say) the part of the genome that controls ability to infect the other species? Many generalist or opportunist pathogens have big genomes (Pseudomonas aeruginosa), presumably precisely for this reason. Is there actually a disadvantage to having a big genome? How would you tell?

Correlations

One way to step between micro- and macro-patterns is to look at broad-scale correlations between ecological and evolutionary characteristics of parasites/hosts and their specificity.

Ecological explanations

Some of the explanations are essentially ecological, based on observations and arguments about the compatibility and encounter filters:

orally transmitted or vector-borne parasites (esp. passively transmitted ones) are less specific than parasites with direct transfer [encounter filter]
parasites with complex life cycles are less specific? (need to allow flexibility/open encounter filter?)
parasites in smaller geographic areas are less specific (same as above, need to allow flexibility)
Host abundance: stable, abundant populations of hosts should promote specificity ("resource fragmentation" hypothesis, Janzen 1981 (available from JSTOR)

Evolutionary explanations

parasites from species-rich taxa will be less host-specific, since they have the opportunity to jump onto many "similar" hosts
older host-parasite associations more specific, since they have had more time to evolve specificity

Macroevolutionary explanations/explorations

Cospeciation and Fahrenholz' rule

Macroevolution refers to the processes driving extinction and speciation of different "higher-order taxa", typically at the species level or above.

Cospeciation is such a process: it corresponds to vicariant speciation in general evolution, as when continental drift separates two continents. If parasites stay with their hosts and speciate with them then the phylogeny of the hosts and of the parasites will match (Fahrenholz' rule). The possibilities are:

cospeciation: parasites remain in both host daughter species, and gene flow among parasites stops so that parasites speciate too (e.g. lice in pocket gophers)
"missing the boat": the parasite fails to transfer to one branch, or goes extinct at some point
species jump/host shift: some time later (possibly much later), the parasite transfers laterally to another taxon. Jumps to distantly related species may occur, in which case host and parasite phylogenies will no longer be congruent
Speciation of parasites within a host population, or "sympatric" speciation may occur, if parasites specialize/gene flow is interrupted in parasites that live on a single host species (e.g., specialization to different body parts in mites)
Parasite speciation may not occur at all, if parasite gene flow is maintained despite host separation

We can look at the patterns in host and parasite lineages at many different scales, from the large-scale patterns of parasite and host lineages to small-scale patterns of local adaptation (e.g., Curt Lively's snails and trematodes again). There is almost certainly enough gene flow among hosts and parasites in New Zealand lakes to keep the subpopulations from speciating, but there is little enough that we can still see evidence of "specialization" (better success of parasites on local hosts) in these populations.

It's worth being open-minded about gene flow; although species are fundamentally different from subpopulations because they have diverged evolutionarily, it can be very tough to draw hard and fast lines. For example, in Peter and Rosemary Grant's studies of Darwin's finches on the Galapagos, they have found that interspecific hybrids are fertile and that genes can thus move from one species to another. There are also many cases where populations below the level of species experience some degree of limited gene flow. The relative balance of gene flow and of parasite flow among populations will determine the overall degree of host specialization, parasite specialization, and host-parasite specificity.