In the rest frame of the wire, the positive ions are not moving, but the electrons are moving. You can think of the wire as an infinite rod-shaped region of a certain negative charge density (i.e. with a certain number of electrons per meter) superimposed on an infinite rod-shaped region of positive charge density. There are two important bits here:
1) Since the electrons are moving, the rod of electron density is length-contracted in the wire's rest frame. Length contraction of an infinite rod increases the apparent density of the rod.
2) The wire is assumed to be electrically neutral in its rest frame. This means that the positive charge density is equal to the length-contracted negative charge density.
Now we transition to a frame where the electrons are at rest (meaning that the positive ions are moving). Two things happen during this transition:
1) The electrons aren't moving in this frame, so the rod of negative charge density is no longer experiencing length contraction. This means that the apparent negative charge linear density decreases relative to the wire's rest frame.
2) The positive ions are moving in this frame, so the rod of positive charge density is experiencing length contraction. This means that the apparent positive charge linear density increases relative to the wire's rest frame.
Both of these effects work together to give the wire an apparent net positive charge. Ultimately, the thing that breaks the symmetry between frames is the assumption that the wire is electrically neutral in one of them.