Yes, it applies to the modulation as well as the carrier frequency. Suppose for argument's sake that the peaks in the modulation occur every 1 million cycles of the carrier. This relation will not change upon reflection, so any shift in carrier means a shift in modulation as well.
To derive the value of the shift (skip to end if you just want the final result), first take the reference frame of the "moving" object. Here the incident and reflected signals both have the same frequency, since reflection from a stationary object will not shift the frequency.
However, in the measurement frame where the object is moving, the incident and reflected beams will have equal magnitude, but opposite sign, frequency shifts. Suppose the object is moving toward you. Then from the object frame's viewpoint, you are now moving in the same direction of the incident beam, and in the opposite direction as the reflected beam. Hence a red-shift for incident, and blue-shift for reflected beam frequencies.
From the observer's viewpoint, there is a doubling of the shift compared to the situation where the moving object simply emits a beam at a given frequency:
[tex]
\frac{\Delta f}{f}=2\frac{v}{c}[/tex]
where we are ignoring terms of order [tex](v/c)^2[/tex]
For an object traveling 100 mph, I calculate a 0.3 ppm (parts-per-million) frequency shift. For a 1 kHz modulation, that would be only a 0.0003 Hz shift. You probably want to use significantly higher than audio frequencies.