As long as your core devteam can handle the dev & maintenance (future evolutions of html5 ?) of this exporter without external help (contributions) than there is no reason why not close its source.
Just be sure to fully document the SDK of the exporter for plugins & third party developpers.
Also businesswise and as you described your will about the future of the devteam and construct it is a smart move.
The one element that would "force" incomes for Construct. As you said too, if scirra could be built into a "true" company, focusing on construct and eventualy tools of the same sort (no precise idea atm, maybe irrelevant, just opening for the future in a "who knows ?" tone ^^) the soft would be maintained on a more regular basis. Having this one element proprietary might help in building scirra.
Like many others users who stated it in different topics, I would be ready to pay good money for construct too. The tool has been providing great reward to me so far and I wouldn't be shocked to have to pay for it. The license use seems caring of the community and fair for scirra too.
I don't know exactly the impact of having it open source so far (I've heard that you haven't had THAT many contibutions). As I first stated, if you are sure with your actual core team to bring construct 2 to a level of stability and functionality that allows it to be profitable enough (before having to hire for more dev ppl, if/when needed on scirra's side, and for companies to rely on construct 2 for their commercial released products on the other side) then no reason to oppose that way of doing.