It depends on the length of you are column using. For maximum resolution, the sample volume should be 1-5% of the column bed volume. Take care that the viscosity of the sample should not be too high than the elution buffer (So dilute your sample if it is too viscous).
For example. If you are using column with 50 cm length and 1 cm diameter, then you can load 1-2 ml of sample after the column is packed.
The protein amount (as mass or moles) is not really a key variable for these kinds of columns. A broad range will work, as long as your detector can see it.
Resolution will be better a smaller amount of protein (assuming constant volume as others have discussed). You should try a few different amounts and see for yourself. There's no good substitute for experimentation here. It's too dependent on your situation.