The following is not a legal analysis (IANAL), but a social/academic one. The problem, as I'm sure you recognize, is that passing on an electronic version makes it all too easy for it to be widely distributed by the recipient. Hopefully you can trust them to honor a request not to pass it farther. Hopefully, they will comply with such a request.
If you want to be absolutely sure that you are staying within all bounds, make a printed copy of the document and send that, along with a request to have it returned to you when your colleague is finished with it, however long that might be. In a working group that has always been done and is certainly covered under fair use for scholarly work. This is probably more than is really necessary, but it is conservative enough that no publisher should object. That is especially true if this paper is important to your collaboration.
But you almost certainly have a right to make a printed copy and you almost certainly have a right to share it. Hopefully your colleague won't make and distribute a can of the document, but that is on them in any case.
Sharing such a document with a colleague isn't "publishing", and while it might reduce the economic value of the publication to a minor (minuscule) extent, a publisher would only be likely to object to it if they were willing to spend money to "make an example" of someone. And likely it would only occur if the person sued had deep enough pockets or if the problem became extreme in some way.
And publishers need to be able to support scholarly work if the next work they hope to publish is to ever be created. The recognize this, actually, and benefit from it.