They use the same crystals, just different wavelengths these operate in.
The three crystal (Nd:YAG) wavelengths used in these process are 946nm, 1064nm, and 1318nm. these specific wavelengths must only exit the crystal, so you need to apply a coating (number one to force the crystal to produce only it, number two to prevent your "pump" wavelength from leaking.) These coatings aren't as efficient as eachother and some of these wavelengths aren't as strong as the other.
532nm comes from a very strong frequency doubled 1064nm wavelength, with a very easy to produce coating and isn't that hard to align.
473nm uses a rather weak wavelength and a different coating, 946nm frequency doubled (wavelength halfed). Since it's a weak wavelength you can't produce as efficiently, less power, and the coating isn't easy to come by. Alignment isn't too easy on these either.
589nm uses a special type of DPSS, which is called SFG (Sum Frequency Generation). Basically, it combines two wavelengths the crystal produces, and frequency doubles the average. (formula=1/((1/WL1)+(1/WL2))) This method is EXTREMELY inefficient, EXTREMELY hard to align, and the coating is EXTREMELY hard to produce.
Hope I shed better light on the matter!