This "problem" is merely an artifact of your model. You seem to assume that the species you call B at present is the same as it was as the first branching point and that it must have lost gene 451 between the first and the second branching point.
However, the more likely scenario is the one where species B still has gene 451 after branching off from species A and only loses it after splitting off from species C.
This confusion could be avoided if you used different labels after each branching point, i.e.:
Species A branches into species B and C. Later, species C branches into species D and E.
Today we observe that species B and E both have gene 451 but not species D. This may be explained by the fact that species C lost the gene but somehow species E magically regained it or which is more likely, that species C still had gene 451, inherited it to both D and E but later D lost it. (Of course there is also the scenario where species A didn't have this gene and B and E 'evolved' it independently of each other, but this is even more unlikely)
For those creationoids who think there's a "gotcha" here, I suggest they think (and I mean "think") twice.
The apparent species skipping is just not so.
Your (BMCDA) post is quite good.
js1138 (?) also showed that the problem could well be associated with incorrect taxonomy, something that is always being updated and corrected. Science is not static.
The other possibility is that the gene is really not "missing", just cryptic or recessive or modifed (could even be post translation modification, or mutated simply so that it can revert to wildtype at a later stage.
All of these are much more probable and fit Occam's Razor much better that ID or some other mystical mush.