It depends on each specific protein conformation, the number of disulphide bonds in the protein, the composition of the buffer it is dissolved in, the presence or absence of any detergents that might change the conformation of your protein, any reagents like urea or Guanidine-HCl, and the pH of the buffer. You can reduce the disulphide bonds with high concentration of the reducing agent (like 100 mM DTT) and then remove the excess of DTT by desalting spin column in the buffer containing freshly prepared 0.5 - 1.0 mM DTT at pH about 7.4.  If your buffer has physiological pH and no other influencing reagents as above, you can be sure that all your Cys residues will remain in reduced conditions for at least about 2 hours. 

If your goal is to reduce the disulfide bond in order to react the cysteine residues with something, you may wish to consider using TCEP instead of DTT. TCEP is compatible with maleimide chemistry, so you can reduce the disulfide and react the cysteines with a maleimide without removing the TCEP.

Thanks for the info Viktor Y Butnev and Adam B Shapiro. I was also considering TCEP but I absolutely need to remove it and learned that they sell TCEP immobilized beads. Assuming room temp and PBS buffer, I am hoping Cys  will stay reduced for at least an hour after removing the TCEP. 

Keep in mind the solubility of your reduced protein at acidic conditions and the pKa of the thiols of proteinyl cysteines.  Once the protein is reduced at ~pH8.5 then by lowering the pH to acidic conditions you protonate the sulfur atom so that electrons are not completely free for covalent bond formation and the oxidation rate drops tremendously; a way to stabilize the situation you're describing and then to process further. by re-adjusting the pH when needed.  Is your reduced protein soluble ~pH 4.5?

Thanks for the info @Grant Shimamoto. My protein is not stable at pH 4.5. I would like to keep it at pH 7.4 after removing the TCEP. Were you thinking along the lines of buffer exchanging out the PBS after TCEP treatment for a low pH buffer to keep it resistant to oxidation for a little longer?

And one downside of TCEP that it's not very stable in phosphate buffers, especially at neutral pH, which is exactly what I have here. I don't have to use this combo of buffer and pH but prefer to keep conditions close to physiological

The assumption was that you were still going to use the spin columns and soluble reducing agent.  So after the reduction with whichever reagent you select the idea is to equilibrate the spin column in vacuum/helium purged buffer that's in the acidic pH range, maybe pH ~5.5 to ~7 if you want to be closer to physiological (The protein is already denatured if the disulfides are broken and many commercial proteins are liquid-formulated at pH 5 for stability.)   The eluate can then be stored under vacuum (don't purge due to foaming denaturation/shear) until you're ready for the next reaction with pH re-adjustment.  I doubt you'll have any oxidation  This works for caspase 3 after the DTT is removed and is stable for days at 5C;  even more, it works for the cleavage reaction as well when undesirable traces of DTT can reduce the cleavable target protein that has labile disulfides.  Immediately after the spin column you might check your product with Ellman's reagent and then check again just before the next reaction if you want to get really analytical about disulfide re-formation.  Also check with RP-HPLC.  If checking with SDS-PAGE non-reducing be sure you include NEM or IAA in the sample loading buffer to prevent in situ re-oxidation.

Solid phase phosphines is convenient  but be sure there is no loss in SN2 reducing efficiency and no re-oxidation during the reaction time course (maybe 1 hr at 25C, might need chaotropes) at your protein concentration.   Be cautious of side reactions:  Journal of the American Society for Mass Spectrometry, Volume 21, Issue 5, May 2010, pg 837–844.   If you proceed with cysteine stabilization through pH post reduction, you probably would forego spin columns but instead use a buffer system of MES or Bis-Tris for your entire procedure because it would allow pH readjustment in the relevant range.  Phosphines have a broad working pH range due to utilization of water rather than thiols.  Another possible approach would be sulfitolysis and then use your spin column afterwards.  The sulfhydryl adduction is highly reversible once you're ready for the next reaction.  

Residual TCEP could reduce disulfide bonds, depending on the concentration.

It might be tricky to get the TCEP concentration just right. Another approach is to use immobilized TCEP, which can be removed after the reduction step:

https://www.thermofisher.com/order/catalog/product/77712

Still another way is to use dithiothreitol, then after the reduction is done, add N-ethylmaleimide in slight molar excess over thiols to consume the remaining thiol groups.

The idea of using N-ethylmaleimide to inactivate the DTT reductant was to avoid having to do the column separation.

how phosphate buffer promote oxidation of cysteine