1D SDS-PAGE with Coomassie blue and silver stain to detect total residual proteins in API samples
Unlike ELISA that relies on specific antibody-antigen binding to quantify protein amount, SDS-PAGE resolves all proteins according to their molecular weight, and its detection sensitivity depends on the staining method used, with Coomassie blue staining has a detection sensitivity of ~10-100 ng/band and silver staining has a detection sensitivity of ~0.1-1 ng/band (Weiss W., 2009). To estimate the amount of residual protein levels in the API samples, Prep. Lab batch API at different concentrations, along with a bovine serum albumin (BSA) standard obtained from micro BCA protein assay kit (Pierce Biotechnology, Rockford, IL) at different concentrations, were loaded and separated on 4-12% Bis-Tris gel as described in Materials and methods. Duplicate gels were either stained with Coomassie blue (Imperial stain) or silver stain. As shown in Figure 5, Coomassie blue stained gel (Figure 5a) can barely detect BSA at ~ 0.4 µg/mL load (~4.8 ng) and silver stained gel (Figure 5b) is able to detect BSA at as low as 0.05 µg/mL (~0.6 ng). In 1 mg/mL API, no protein band is detectable by Coomassie blue staining and a few bands are barely detectable with silver staining. When increasing the concentration of API, more proteins become detectable in both Coomassie blue and silver-stained gels. Although different proteins can have different response factors when stained with Coomassie blue (Congdon, Muth, & Splittgerber, 1993; ”The European Agency for the Evaluation of Medicinal Products Human Medicines Evaluation Unit
”, 1997; Tal, Silberstein, & Nusser, 1985), the amount of proteins detected in the API samples on the gel can be estimated using the intensity of protein bands observed and BSA as a reference standard (data analysis performed using the ImageQuant IQTL 8.0 software, GE healthcare (Piscataway, NJ). Since there is barely any protein band detectable in API at 1 and 2 mg/mL load, and BSA band is detectable at 0.4 µg/mL, assuming all proteins have similar detection sensitivity as BSA, then no individual protein in the API is above 200 ng/mg (0.4 µg/mL/2 mg/mL). The bands observed in 4, 10, and 30 mg/mL API were also used to estimate the total protein amount in those bands according to the calibration curve generated by BSA standard using IQTL8.0 software. A second gel was run and Coomassie blue stained to assess the variability of this estimate (gel image not shown). The ng/mg difference calculated between the two gels are included in Table 2. Across three different API concentrations, the average protein amount is estimated to be 764 ng/mg, ranging from 621 to 884 ng/mg. Of note, since Coomassie blue non-specifically binds all proteins through ionic interactions between dye sulfonic acid groups and positive protein amine groups as well as through Van der Waals attractions, in this method, all proteins over the detection limit should be visible on the gel. The increased appearance of protein bands along with the increase of loading concentration indicates that many protein species is at an abundance level lower than the detection limit when API is loaded at low concentrations, while the dominant proteins were shown as 6 major bands in Figure 5 and 6. Although using SDS-PAGE gel with Coomassie blue staining for protein quantitation can have high variations due to gel loading volume variability, the differences in response factors of staining between BSA and other proteins, the time difference on staining and destaining steps etc., the results from gel analysis indicates the presence of proteins at levels much higher than that quantified by Cygnus kit, which is also indicated in our previous publication using ImageJ software coupled with densitometry analysis on 1D SDS-PAGE gel(Smith et al., 2022). To further identify those proteins and investigate whether they can be detected by the Cygnus kit antibody, we excised those bands and did in-gel digestion coupled with nanoLC-MS/MS protein identification.