Multivariate analysis of single quadrupole LC-MS spectra for routine characterization and quantification of intact proteins.

Modern high-throughput proteomic platforms allow incomparable protein mixture resolution and identification. However, such sophisticated facilities are expensive and not always accessible for routine analysis of simple mixtures. In this paper, we propose a simple methodology, based on detection of intact, nondigested proteins by LC coupled to single quadrupole MS (sqLC-MS), followed by the analysis of the resulting spectra by multivariate analysis (MA). By doing so, even large molecular weight (MW) proteins, generating complex spectra, can be characterized to a level that allows isoform discrimination, while standard algorithms, such as MS spectrum deconvolution, cannot. To demonstrate the effectiveness of the proposed approach, we have analyzed the spectra of a set of purified, intact albumins from seven different organisms (bovine, human, rabbit, rat, sheep, mouse, and pig) as a model of microheterogenous proteins, using Projection to Latent Structure Discriminant Analysis (PLS-DA). Although these proteins are very similar (less than 1% difference in MW), sqLC-MS/MA allowed their classification, and the identification of unknown source samples. In addition, MA allowed precise protein quantification from the same data (calibration curve R2 = 0.9966). The ability to rapidly characterize and quantify proteins, together with simplicity and affordability, could make of combined sqLC-MS/MA a routine method for the characterization of simple mixture of known proteins.

Cell culture tracking by multivariate analysis of raw LCMS data.

Liquid chromatography mass spectrometry (LCMS) is a powerful technique that could serve to rapidly characterize cell culture protein expression profile and be used as a process monitoring and control tool. However, this application is often hampered by both the sample proteome and the LCMS signal complexities as well as the variability of this signal. To alleviate this problem, culture samples are usually extensively fractionated and pretreated before being analyzed by top-end instruments. Such an approach precludes LCMS usage for routine on-line or at-line application. In this work, by applying multivariate analysis (MA) directly on raw LCMS signals, we were able to extract relevant information from cell culture samples that were simply lyzed. By using the recombinant adenovirus production process as a model, we were able to follow the accumulation of the three major proteins produced, identified their accumulation dynamics, and draw useful conclusions from these results. The combination of LCMS and MA provides a simple, rapid, and precise means to monitor cell culture.