One of the factors keeping H5N1 from gaining H2H transmissibility is that humans are a dead-end host for the virus. Since its reservoir is birds, the selective pressures on the virus are for it to maintain its avian-optimized infectivity and transmissibility profile. The bigger fear with H5N1 is that it can infect a host such as swine where it would have the opportunity to swap genes with another influenza virus.
With MERS, we are somewhat handicapped in that we do not know what kinds of mutations would make it easily transmissible. At least, since its genome consists of a single RNA strand, recombination with a more transmissible coronavirus (like the one that causes common colds) is not very likely.
Humans are not a dead end host for H5N1 at all. This is why there was such concern with the Egyptian clade, which had lower lethality and was primarily a pediatric form.
In fact, humans would be an ideal host for H5N1. With somewhere between 7-8 billion humans, we are a primary target for any opportunistic pathogen. And with a 7 day to 2 week incubation period, large collections of people would be like a flock of chickens, seemingly healthy, and by the next day, all dead or dying.
The five mutation barrier to humans just needs to be broken in the right place, like an urban area. Because it is a novel pathogen, it would take far less of it to cause an infection, a fraction of the amount of ordinary influenza.
And it has a far greater number of animal vectors than it typical, including those with very different immune systems.