In string theory the D-branes give those directions that the strings are allowed to move along. The string excitations give the fields that we detect. Is it correct to think of a particle propagating in 4d spacetime as a string propagating in the same 4d spacetime (in other words if we could zoom in sufficiently far a photon for example would reveal itself to be a string). If so, this implies that a realistic model should always have D-branes that extend in these 4 spacetime directions. Otherwise how would the strings, and their associated excitations, propagate in the 'normal' spacetime we see at low energies?
Yes, there must be branes filling the ordinary large spacetime dimensions in order for gauge fields to propagate there. Even the Type I string theory containing gauge fields (open strings) in 10 dimensions can be understood as having space-filling D9 branes on which the gauge fields live (as well as an orientifold O9 plane).
The exception is heterotic string theory, which has gauge fields that don't have an obvious natural interpretation in terms of branes. (Via dualities, you can always find some kind of brane interpretation, although the end-of-the-world branes in heterotic M-theory are different from ordinary D-branes.)