From a first approximation, the depletion zone of a reverse-biased diode is simply an insulating region. But this doesn't explain the collector junction of a transistor. We need to look at this insulation-effect in more detail.
In truth, a depletion zone does not block the motion of any charge carriers found there. Instead, there's (usually) no significant carrier population there in the first place. A depletion zone is an insulator like an empty vacuum: a voltage placed across a vacuum will produce zero current, showing that a vacuum is insulating ...yet any charges injected from outside would easily flow.
In a reversed diode, electrons from the n-doped side might invade the depletion zone. But they'd be forced back by the strong e-field in that zone. The same thing happens if holes from the p-doped side should invade the depletion zone: they're pushed back again.
But what if we dumped a bunch of electrons into the p-doped side of our diode? Sure, many would be swallowed up by the holes there. But some would pour into the depletion zone, where they'd be strongly forced across the junction and into the n-doped side. (The larger the reverse-bias voltage, the faster those charges would move.) So, dumping charges into the wrong side of a reversed diode will cause a large current.
And that's exactly what transistors do: in an NPN transistor, the Emitter region dumps large numbers of electrons into the p-doped Base. From the viewpoint of the CB junction, those electrons are on the wrong side of that diode. A few are swallowed by holes, but the majority wander through the Base region, and make it all the way over to the Collector's depletion zone. If they touch it, it grabs them and accelerates them with the full Vcb voltage-field, flinging them into the collector region. (Their large K.E. causes the Collector to heat up.)
So, heh, a BJT is much like a vacuum tube triode, where the Collector region is like a positively-charged metal plate, and the depletion zone of the Collector junction is like a vacuum with a large voltage placed across. And even worse[*], with NPN transistors, if we force the initially-positive Vbe to become more and more negative, it turns off the electron flow, just like a Grid electrode does.
[*] Worse for those who dislike the idea that BJT transistors are like FETs and vacuum-triodes: they behave as transconductance components, with output current controlled by a voltage signal.