The opening of a puzzle begins when it is first presented to folders. It ends (for you) whenever you are no longer using the initial configuration as a reference or jumping off point. Unlike in turn-based strategy games, you can always go back to the opening in FoldIt, and sometimes it may be a good idea, particularly if a given solve seems to be reaching the endgame too soon, and if your score is not yet where you'd like it to be.
One good approach to the opening may be as follows, although it should be kept in mind that FoldIt strategy has not yet developed to the point where any conclusions are certain.
All puzzles start out with a configuration that scores very low. The key to good opening strategy seems to be to try many different things and to explore a wide variety of different paths. Each one branches off from the starting point in a different direction. "Be bold" is one of the proverbs that seems most important here.
The first opening move is simply to global wiggle and watch how the entire structure changes. Do not worry about the score, although sometimes you can indeed score very well simply by wiggling the starting configuration. However, the true purpose of this wiggle is to get a rough idea of which elements "want" to be together, which ones do not, and where low-scoring trouble spots seem likely to arise.
After you've achieved a more or less stable structure using a single global wiggle, reset the puzzle. Keep in mind what you've noticed about it. Consider whether you can rebuild any of the elements. Rebuilding can often be more profitable in the initial configuration than in a more compact one. Consider where rubber bands may help. Adjust sidechains if they seemed to get in the way of the global wiggle's progress. You may even want to lock some parts of the structure, although lock is a difficult tool to use in the opening, since it temporarily raises the overall structure's mojo. During the opening, you want a structure that is highly variable and flexible, with lower mojo, so you can explore many different options.
Once you've applied some modifications to the initial configuration, wiggle it again. Again, watch the way it comes together. It won't be quite the same as the first time, and it may be better or worse. If it's better, try to continue in that direction using the tools at your disposal. If it's worse, figure out if you can move it in an opposite direction.
Repeat this process until you feel satisfied with your overall build. Then you're ready to proceed to the middlegame, keeping in mind that you can always go back to the opening if you aren't getting enough out of your middlegame setup.