A/C box skins (wing, tailplane, fin) might be as much as 60% spanwise fibre in some places. This needs unidirectional material (unbiased all cloth layups can be at most 50% 0 degree). Such a skin might vary between 40% and 60% spanwise. It depends on the other forces in the box. +-45 degree plies need to be present to take box torsion. Typical percentages might be 30 to 50%. Chordwise fibres are needed as well, particularly near significant load inputs such as trailing edge control surface hinges and actuators. Skin stiffeners can be more 0 degree, as they don't need as many angle plies or 90 degree ones.
Spar webs need lots of 45 degree plies. Ribs need a good mix, with (usually) quite a few fibres in the vertical direction to resist Brazier crushing loads and tension due to over-fuel pressures and crash pressures.
Before the aerodynamic effects of gross structural coupling between incidence and torsion were understood some terribly dangerous monoplanes were designed. In WWI quite a few German pilots were killed by such effects.
Deliberate material-based bend-twist coupling was probably most famously used on the X29 with its forward swept wing.
While you can achieve such coupling with uneven numbers of + and - angle plies, it can also be done by using the distribution through the thickness and to keep the same number in each direction. (This is probably more practical for a more monolithic structure than a box.) Just put all the +angle plies nearer the centerline of the beam than the -ones (or vice versa, depending on what you're trying to do). Under bending, the +plies above the neutral axis try to shear one way and those below it shear the other (giving twist). The same goes for the -plies, giving twist in the opposite direction. However, the plies furthest away from the neutral axis are more effective, giving a nett twist in their direction.