I'll be nitpicky for a moment: you probably don't want to ask 'what should a rationalist care about?', rationality is a set of tools, not values. Also, keep your identity small.
Most of us, if we found that some of our values rested on confusions, would say those values had never been our true values. This is true not just of confusions about means-ends relationships but also of other confusions. So it's unsafe to assume, the way people here often do, that if a disagreement seems on the surface to be about terminal values, it's not up for rational debate.
Also, what one cares about is a different question from what one is concerned about; to me, "concern" about some value implies a claim that the degree of caring, combined with the practical details of the situation, makes it worth sometimes choosing the value over values that one would otherwise have pursued.
It's an interesting question that I'm pondering.
Now, while I do question the intellectual honesty of this blog, I'll link to it anyways, since the evidence does seem interesting, at the very least: http://wattsupwiththat.com/2010/01/04/where-are-the-corpses/
http://wattsupwiththat.com/2011/05/19/species-extinction-hype-fundamentally-flawed/
It does seem that environmentalism can mimic some qualities of religion (I know, since I used to be an environmentalist myself). As such, it can cause many extremely intelligent people to reject evidence that goes against their worldview.
Furthermore, it's also possible that computational chemistry may soon be our primary agent for drug discovery, rather than discovering more biological compounds in certain ecosystems (that being said, drug discovery is entirely different from drug synthesis, and discovering a gene that codes for a particular protein and splicing it into an E Coli bacterium is going to be far easier than anything computational chemistry can do in the near future).
With that all being said, what now? I do believe that there is something of value that does get lost as habitat gets destroyed. But it's hard to quantify value in these cases. Certain animals, like crows, chimpanzees, orcas, and elephants, are cognitively advanced enough to have their own cultures. If one of their subcultures get destroyed (which can be done without a fullscale extinction), then is anything valuable that gets lost? (besides value for scientific research that has potential to be applicable elsewhere?) And is it more important to worry about these separate cultures, as compared to worrying about different subspecies of the same animal? Certainly, we're now beginning to discover novel social networks in dolphins and crows. But most of these animals are not at risk of extinction, and even the chimpanzees and bonobos will only get extinct in the wild (at the very worst). There are other less advanced animals that have a higher risk of permanent extinction.
What we're prone to systematically underestimating, of course, is the possible permanent loss of micro-organisms. And of novel biological structures (and networks) that may be contained within them.