October 1, 2009 (Vol. 29, No. 17)

Patricia F. Fitzpatrick Dimond Ph.D. Technical Editor of Clinical OMICs President of BioInsight Communications

Alternatives Designed to Overcome the Limitations of 2-D Electrophoresis Are Under Development

Quantitative proteomics currently encompasses a variety of techniques including differential labeling of separate protein mixtures; digestion of combined, labeled protein mixtures such as cell lysates with separation of resulting peptides via multidimensional LC; peptide analysis using tandem mass spectrometry; and database searching to identify and associate peptide sequences with specific proteins as well as determination of relative protein abundance from the mass spectral data.

The historical workhorse for protein analysis has been high-resolution two-dimensional electrophoresis (2DE) combined with mass spectrometry. In this method, complex proteins are separated electrophoretically and then stained. Specific proteins are chosen for mass-spectrometric identification based on quantitative comparison of the 2DE staining patterns of control proteins. There are, however, factors limiting 2DE/mass spectrometry applicability including inability to distinguish, analyze, and quantitate low-abundance proteins. Frequently, analytical methods have required the development of reagents such as antibodies to detect specific proteins.

At the “Symposium on the Practical Applications of Mass Spectrometry in the Biotechnology Industries,” which was held during CASSS’ “Mass Spec” meeting in Philadelphia last month, industry scientists described applications of mass spec to simplify and streamline protein characterization and quantitation for biotech industry applications. Presentations covered a range of topics, including an LC/MS-based assay developed to detect residual host-cell proteins in biologics and use of deconvolution to resolve isotopically unresolved multiply charged states of intact proteins.


CE-MS analysis of intact glucoprotein produced from a murine cell line. Numbers indicate the number of sialic acids for the glycoforms. (Northeastern University)

ELISA Alternative

AlphaVax has adapted LC/MS assays to quantitate the vaccines it is producing for a range of infectious diseases and cancer. The company manufactures its vaccines using an alphavirus-like replicon particle (VPR) containing an RNA that encodes a protein of interest. The VPR is used to infect a cell monolayer in culture.

Following viral particle entry into a cell, the RNA is released and translated, resulting in accumulation of the desired antigenic protein in the cell. The cells are then lysed and digested with trypsin in preparation for MS analysis. The MS portion of the assay provides a means of quantifying the resulting tryptic peptides without the need to use isotopically labeled standards.

According to Jeremy Johnson, Ph.D., senior scientist, MS-based assay development does not depend on antibody-based reagents and permits rapid assay development without specific biologically derived reagents. Quantitation of the desired protein is accomplished by measuring the total ion count of a specific tryptic peptide present in cell lysate and correlating it against a standard curve generated from the equivalent synthetic peptide standard.

Dr. Johnson explained that the MS-based assay is a platform technology that can be used to screen vaccine candidates without the need to develop a full ELISA, thereby saving considerable time and money during early-phase product development. “Quantification based on peptide total ion count requires no additional experimental steps,” he added. “The direct correlation of total ion count to peptide concentration enables peptide and, therefore, protein quantitation. This method can be developed quickly with excellent specificity and a large dynamic range.”

AlphaVax will continue to use this protein assay method in product and process development. The data collected from its use will be utilized to support ELISA development as products mature. The availability of this data for material made in early development should be useful in supporting product/process consistency throughout the development cycle, Dr. Johnson explained.

Host-Cell Protein Assays

Biopharmaceuticals produced in recombinant cell lines can contain residual host-cell proteins (HCP) characteristic of the cell lines in which they were expressed. Present at extremely low concentrations, typically µg/g of recombinant protein, or ppm concentration, these contaminants can potentially elicit immune responses in individuals to whom they are administered.

According to Waters scientists, current analytical methodologies may not completely identify and quantify individual host-cell proteins, as such assays typically only provide a total concentration of host-cell protein. Also, current methods may not respond equally to all host-cell proteins present in a sample. As each therapeutic protein utilizes a unique purification process, generic host-cell protein assays may miss key host-cell protein contaminants present in the sample.

At the meeting, Catalin Doneau, Ph.D., senior scientist, described the development of a two-step approach to host-cell protein discovery and quantitation. In the discovery phase of a biopharmaceutical, 2D-nanoUPLC/MS/MS with a QTOF mass spectrometer, with database searching of tryptic peptides from the digested biopharmaceutical protein, is used to identify host-cell protein impurities. The identified proteins can be quantified using label-free quantitation against a spiked-in protein standard.

Once identified, individual host-cell proteins can be quantitated on a tandem quadrupole system by multiple reaction monitoring (MRM) assays of signature peptides from each identified HCP. The methodology was successfully used to analyze five low-abundance proteins introduced into a stock solution of a humanized monoclonal antibody, according to Dr. Doneau.

Jeff Mazzeo, Ph.D., biopharmaceutical business director, explained that the challenges inherent with quantitating HCP contaminants in recombinant protein production systems include never “really being sure whether you have detected all of the contaminating cell proteins in the product.”

Typically, he said, companies may use polyclonal antibodies for ELISAs rather than a targeted monoclonal to detect a specific host-cell protein. Another approach is to use mixtures of monoclonal antibodies for specific contaminants, but the key step is discovering all the potential host-cell protein contaminants. The issue, he explained “is that you don’t know whether you are detecting all of the proteins present and you can miss things” Further, he noted, regulatory agencies are asking companies for very specific information, requiring not only determination of total HCPs but individual discrimination among them.

Waters uses tools originally developed for proteomics and applies them to HCP detection and quantitation as well, Dr. Mazzeo said. For protein identification, “we perform two-dimensional LC separation, which allows us to load a lot of protein so we have a better chance of finding contaminants that are present in small amounts.”

LC separation followed by tryptic digestion produces peptides that are sequenced, and then identified by searching against a database to find peptide matches with specific protein sequences. If a complete protein sequence database is unavailable, as he said is the case with CHO cells, “we use a combination of mouse and rat genomes to identify CHO proteins, relying on homology.”

Once a process is established and an expected range of contaminants identified, HCPs can be quantitated directly by MRM. “Once you know what HCPs are present, they can be quantified with greater precision and a lot faster using a tandem quadrupole mass spectrometer. In an MRM experiment, one is monitoring the transition of the parent peptide ion to a specific fragment ion, which provides tremendous specificity and sensitivity. This approach has been used in proteomic studies to quantify biomarkers.” Once you know what the host-cell proteins are, you can use this MRM approach, he reported. “We suggest using the ID on a really dirty sample, then setting up a targeted MRM method for the next phase.”

Deconvolution of Isotopically Unresolved Intact Proteins

For spectra acquired through electrospray ionization of LC-separated biomolecules, charge deconvolution to yield enhanced ion statistics by combining peak intensities for all detected charge states is generally required. Maximum entropy charge deconvolution can produce spectra characterized by higher resolution and correct isotope distribution, enabling accurate determination of molecular weights and quantitation.

At the symposium, Natalia Belyaeva, Ph.D., a scientist at Thermo Fisher Scientific, described an alternative approach for resolving multiply charged proteins using a method based on graph search of all optimal paths.

Dr. Belyaeva explained that existing algorithms may not allow processing of data because spectral line widths vary with m/z region, even for the same compound. As a result, she and her coworkers developed an algorithm to resolve multiply charged protein entities utilizing a graph search of all optimal paths. All charge states that can belong to the compound form an individual path on the graph consisting of all determined peaks. They then applied the algorithm to analyzing different modifications such as glycosylation of an IgG.

“This algorithm assigns a score based on each determined charge state chain based on m/z values, peak quality, and charge envelope shape,” Dr. Belyaeva explained. The score assigned to each potential charge-state chain facilitates choosing the “best hypothesis” for all possible charge-state chains, and can avoid assigning a charge state to wrong mass.

To increase data quality by decreasing random noise and improving the charge-state envelope, scientists performed optimal averaging of spectra before the deconvolution algorithm was applied. The scientists showed that a deconvolved spectrum, after appropriate averaging to “get enough charge states and improve the charge-state envelope” revealed multiple glycosylation sites, as well as several antibody variants.

“High-accuracy deconvolution provides not only correct protein molecular weight determination, but also accurate quantification, which is important for such areas as characterization of different protein modifications or drug conjugates and can be applied to data acquired using both orbitrap and ion traps,” Dr. Belyaeva said.

All of these novel applications are expected to help biotechnologists identify, characterize, and quantitate complex protein therapeutics produced in a variety of recombinant systems.

Protein Glycosylation Characterization

Tomas Rejtar, Ph.D., research assistant professor at Northeastern University’s Barnett Institute, along with Professor Barry Karger and colleagues, described the development of an analytical method for intact glycoproteins based on high-resolution capillary electrophoresis separation coupled to MS, which was developed in collaboration with Momenta Pharmaceuticals.

Using this method, the investigators analyzed and compared glycoproteins obtained from different sources and discovered significant differences in glycoform abundance and composition among the proteins.

To perform the analyses, the scientists used a capillary electrophoresis (CE)-MS instrument with a pressurized liquid junction-based interface coupled to an LTQ-FT instrument. Proteins used to evaluate the analytic performance of this system included the alpha-subunit of human chorionic gonadotropin with two glycosylation sites.

High-resolution capillary electrophoresis enabled separation of intact protein glycoforms differing in the number of sialic acids as well as those with varying neutral glycan content. Analysis of the same sample showed that relative abundance of individual glycoforms could be quantitated by CE-MS with a relative standard deviation of less than 10%. The method can also be combined with enzymatic deglycosylation of the intact proteins to assign glycans to specific sites for further glycoprotein characterization, Dr. Rejtar said.

This method has the potential for application in the rapid profiling of glycosylation in therapeutic proteins, according to Dr. Rejtar. Unlike traditional methods that rely on analysis of released glycans, direct analysis of intact glycoproteins by CE-MS allows fast monitoring of changes in glycosylation or other modification with limited sample preparation, he added.

Patricia F. Dimond, Ph.D. ([email protected]), is a principal at BioInsight Consulting.

Previous articlePHD Diagnostics Develops Gene-Based Test to Predict Smokers’ Lung Cancer Risk
Next articleFormation of Dopamine-Producing Neurons Requires Cholesterol Derivative