Streptavidin nonspecific in matrix

Using a CM4 chip I've amine coupled a protein and had success binding an antibody (Ab) to this protein. Regeneration seemed to be complete using 10mM glycine, pH 2.2. This antibody is biotin labeled so I combined the Ab with streptavidin (SA) and repeated the injection over the immobilized protein. Problem now is that the pH 2.2 no longer regenerates the surface and it appears from doing other controls that the SA is nonspecifically incorporating into the CMD matrix.

I was wondering if anybody else has seen this and what they're done to fix the problem.
Profplum

When you combine the streptavidin with the biotinylated antibody you can make a complex with four antibodies on one streptavidin. Each antibody having two interaction points you make a octavalent molecule! No wonder that regeneration is very hard. You have to try something else. For instance a lower pH, higher salt, or a combination.
This page can be of help: www.sprpages.nl/kinetics/regeneration.html

The non-specific interaction is more difficult. I have no experience with SA but you can try to add CM5 dextran to the injection or raise the NaCl concentration to 250 mM. This will help sometimes.
Arnoud