It could require massive amount of work.

To begin with, I would establish  the concentration range.

If your target binds your molecule with nM affinity you may use 

radio-labeled  molecule to locate your target during chromatography or

on a native gel. Alternatively you may try to immobilize your molecule 

on some sort of matrix and fish out the target  from the crude extract.

If you have any candidates inferred from the literature, knocking them out 

may be easier and less technically challenging.

Here are a few suggestions 

2D gels (have you tried radioactive or fluorescence labeling?)

stable isotope labeling (http://www.mcponline.org/content/1/5/376)

microarray (http://www.affymetrix.com/catalog/131498/AFFY/S.+aureus+Genome+Array#1_1)

coimmunoprecipitation followed by MS identification of target proteins

Seconding Co-IP followed by MS. It seems like a lot of work to get protein expression levels for absolutely everything unless you really have no idea what this treatment is doing.