There are some mistakes in karkas's #2. The square root is missing from the denominator of the equation's right-hand side. Also, karkas's argument shows that if a particle travels at c, it must have zero rest mass. But what particlemania asked for was a proof that if a particle has zero rest mass, it must travel at c. Finally, the statement that "it is a postulate that light travels at maximum universal speed c" is not necessarily incorrect, but it is somewhat misleading in the present context, because: (1) there are various axiomatic foundations for relativity, and some do not require constancy of c as an axiom (Rindler 1979); (2) regardless of the axiomatization, one can always prove that if there is some frame-independent velocity, then massless particles must travel at that velocity.
As an intuition-building warmup, imagine that in purely Newtonian physics, we had a particle with zero mass. Even the tiniest perturbation, with the most miniscule energy, would be sufficient to accelerate it to an infinite velocity. This is more or less how it works in relativity, except that the relevant limit is not v->infinity but v->c. For instance, neutrinos have almost no mass, and therefore essentially all neutrinos that we observe are moving at very close to c.
A more rigorous argument is that E^2-p^2=m^2 (in units where c=1). The case of zero mass gives |p|=E, and this is only possible if, in the limit m->0, we have m\gamma v=m\gamma, so v=1.
Rindler, Essential Relativity: Special, General, and Cosmological, 1979, p. 51