Enhanced Raman spectroscopy offers capabilities to detect molecules in the complex molecular environments and image chemical heterogeneity in a wide range of samples. It has been shown that plasmonic interactions between a TERS tip and a metal surface produce significant enhancements. In this report we show how SERS spectra from purified molecules can be used to selectively image proteins on surfaces and in cell membranes. The SERS response from the purified protein can be used to create a multivariate regression model that can be applied to nanoparticles that bind to protein receptors. Filtering the observed TERS spectra with the regression model can then selectively image the protein receptor. Experiments with mutant proteins suggest that key amino acids provide significant contributions to the observed TERS signal, which enables the differentiation of protein receptors. These results demonstrate the selectivity that can be obtained in TERS images through a controlled plasmonic interaction. This approach has further implications for identifying membrane receptors that bind specific molecules relevant to drug targeting and chemical signaling.