The magic thing about nitrocellulose is that almost all proteins can bind to it easily and irreversibly unless a high concentration of surfactant is applied. If your protein does not bind well, you can add 1-2% methanol to your protein buffer.
I have once compared the binding of 125I-labelled Hsc70 to various membranes. Whilst NC bound at most 60% and efficiency dropped with protein amount, PVDF bound 98% independent of protein amount (in the range tested, of course). Given that the protein was separated from free iodide only by a desalting column and hence may have contained some residual, this means that PVDF binds protein quantitatively, NC doesn't.