I have done some work on vitamin content of fruit-juice-based beverages. For water-soluble vitamins, use a polar-embedded column with a phosphate buffer and gradient from 0 to 20% acetonitrile. A diode-array detector is very desirable. (
http://www1.dionex.com/en-us/acclaim_li ... 28432.html)
B3 and B6 are not too hard. B12 is a lost cause for HPLC for 3 reasons: its normal concentration is extremely low, it elutes in the same region as the plant pigments, and it is not stable in the presence of high concentrations of vitamin C. Vitamin E is a separate assay.
Depending on which fruits you are analyzing, you will need to alter the method to move the interferences away from the vitamins. Usually this involves small changes to pH. Exact pH control is important to repeatable results. Also you can use a different polar-embedded column; I have used both the PolarAdvantage and PolarAdvantage-II columns. For these reasons, you can't transfer the method for one fruit to another without revalidation.
Preparation for fortified beverages is usually just filtration. Sometimes, treatment with a Dionex OnGuard-P cartridge is helpful to remove pigments and phenolics.