The extent of both files looks like you have reprojected the data to a projected CRS, but then assigned a longlat CRS in the end. In this way, units are metres, not degrees as it should be with longlat.
If you know which CRS you have reprojected to, you can assign that to the shapefile with
Vector -> Data Management Tools -> Define current projection, and then reproject the data to WGS84 with Rightclick
-> Save As... and choosing the CRS.
If you don't know, but can give the extent of the original shapefile in degrees, we could try to find out the projected CRS.
UTM 39N (EPSG:32639) applied to the first extent looks rather good:
red stars are "bad" UTM39N coordinates, green ones "good" UTM 36N coordinates.