MSACL 2016 US Abstract

Development of a Mass Spectrometry Method for Quantifying Glycoprotein in Ebola Virus-Like-Particles

Michael Ward (Presenter)

Bio: Michael completed his undergraduate degree in Biology in 1990. He is currently a Staff Scientist in the laboratory of Dr. Sina Bavari in the Molecular and Translational Sciences Division of the US Army Research Institute of Infectious Disease (USAMRIID). Mr. Ward has 25 years of experience in protein biochemistry and mass spectrometry, and is one of the lead researchers for proteomic studies at USAMRIID where he is involved in the development of mass spectrometry approaches for therapeutic monitoring and discovery of biomarkers of infection. Mr. Ward is also responsible for all proteomic data analysis at USAMRIID as well as analytical software development for in-house research projects.

Authorship: Michael D. Ward1, Paul Demond2, Ernie Brueggemann1, Chris Mahone1, Tara Kenny1, John Nuss1, Trevor Glaros2, Lisa Cazares1,3 and Sina Bavari1
(1) U.S. Army Medical Research Institute of Infectious Diseases, Frederick, MD 21702, (2) Edgewood Chemical Biological Center, Gunpowder, MD 21010 (3) DoD BHPCSAI, Fort Detrick, MD 21702

Short Abstract

Virus-like particles (VLPs) are a promising vaccine platform composed of a subset of viral components that mimic the wild-type virus structure but lack genetic material. Ebola VLPs are produced by expressing recombinant Ebola viral glycoprotein (GP) and Ebola viral matrix protein VP40 in culture. Ebola VLPs protect non-human primates from lethality after Ebola challenge when administered as a vaccine. In order to better predict the success of each VLP lot, we developed a mass spectrometry method to quantify the amount of Ebola glycoprotein. Our results revealed a strong correlation between survival and the total amount of full length viral glycoprotein.

Long Abstract


Ebola is an extremely pathogenic virus that causes hemorrhagic fever and can result in mortality rates of up to 90%. The 2014 Ebola endemic in western Africa brought global attention to a disease that was once only an isolated-regional problem. More than a year later and with a death toll greater than 10,000 people, there is an urgent need for novel therapeutic strategies including treatment and prevention. Virus-like-particles (VLPs) represent a new type of prophylactic vaccine that has had success and is commercialized in products such as GlaxoSmithKline’s Engerix (hepatitis B virus), Cervarix (human papillomavirus), Merck’s Recombivax (hepatitis B) and Gardasil (human papillomavirus). Ebola VLPs (eVLPs) are produced by expressing recombinant Ebola viral glycoprotein (GP) and Ebola viral matrix protein VP40 in HEK293T cells. When these proteins are expressed they self-assemble and bud from ‘host’ cells resulting in VLPs that contain viral GP, VP40, and other randomly packaged host proteins[1]. eVLPs are therefore composed of a subset of viral components that mimic the wild-type virus structure but lack viral genetic material thus rendering them non-infectious. Ongoing Ebola vaccine VLP studies, performed by the US Army, have shown that eVLPs provide protection to a lethal dose of Ebola virus in mice and non-human primates when administered with an appropriate adjuvant. However, in our experience, the effectiveness of each small scale VLP preparation can be quite variable. Since the full-length Ebola glycoprotein is known to be immunogenic and expressed as multiple isoforms, we developed as mass spectrometry method to quantify the amount of GP in order to provide quality control and accurately predict the effectiveness of each eVLP vaccine preparation.


Five tryptic peptide sequences were initially identified as candidate quantitation targets primarily due to their ionization characteristics, lack of know post-translational modifications, and relative distance within the GP1 sequence. Isotopically labelled AQUA Ultimate™ peptides (Thermo Scientific) of each peptide sequence were synthesized and evaluated as quantitation standards. During initial method development we found two of these peptides (173- GTTFAEGVVAFLILPQAK[13C6, 15N2]-190), (634-TLPDQGDNDNWWTGWR[13C6, 15N4]-649) failed to show consistent linearity and one other (479- LGLITNTIAGVAGLITGGR[13C6, 15N4]-497) proved to be too hydrophobic for reliable quantitation using C18 column fractionation. The remaining 2 peptides (65-SVGLNLEGNGVATDVPSATK[13C6, 15N2]-84 and 303-SEELSFTAVSNR[13C6, 15N4]-314) provided highly reproducible linear standard curves and were therefore selected for use in the assay. The SVG peptide, given its proximity to the N-terminus, is common to all potential GP isoforms (Gp, sGp, ssGp) but the SEE peptide is unique to the full length Ebola GP only. Due to a variable but persistent occurrence of a missed cleavage at the C-terminus of the SVG peptide and at the N-terminus of the SEE peptide, the missed cleavage standards (65-SVGLNLEGNGVATDVPSATK/R[13C6, 15N4]-86) and (301-IR/SEELSFTAVSNR[13C6, 15N4]-314) were also included in the final method. No other regional missed cleavages were observed for either peptide.

Initial Orbitrap survey scans of the digested VLPs revealed a single deamidation on one of the two asparagine residues in the SVG peptide, which resulted in a mass shift of 0.984 daltons and that accounted for approximately 15-20% of the SVG/R contribution. Since the XIC response of the deamidated analyte peptide was identical to the deamidated AQUA standard, the contribution of this peptide species was added to the total SVG quantitation value. The final quantitation method included integrating the peak areas of 12 parent ions, including the two most abundant charge states (2+ and 3+) for each labeled and un-labeled peptide as well as the deamidated SVG and SVG/R species.

In order to validate this methodology, a highly purified monomeric full-length GP standard was prepared and quantitated using amino acid analysis. For evaluation by mass spectrometry, the VLP and rGp standard digests were resuspended in either 100µL or 120 µL 40% acetonitrile, 0.1% formic and a 5-point, 2-fold serial dilution performed. AQUA peptides were then spiked into each analyte dilution at a 1:1 (v:v) ratio resulting in a 100 fmol/µL AQUA standard concentration. Each dilution was run in triplicate followed by a duplicate run of the entire quantitation. The total number of fmols was calculated for each peptide at each dilution by comparing the average analyte XIC response to the average AQUA peptide standard response at 4 ppm. The values for the SVG and SVG/R peptides (Set 1) and the IRSEE and SEE peptides (Set 2) were summed to represent the total stoichiometric contribution for each set.


Two mRNA species are transcribed from the wild-type EBOV GP gene as a result of transcriptional editing. The EBOV polymerase transcribes the unedited GP gene which contains 7 adenosines at the editing site (with >80% frequency), and these transcripts program the expression of the predominant GP gene product, sGP (secretory glycoprotein) [2]. In contrast, the GP1/2 transmembrane protein is expressed only when an extra (eighth) adenosine is inserted into the nascent mRNA via stuttering of the EBOV polymerase over the editing site. The secreted form of GP (sGP) is identical to GP1/2 in its 295 N-terminal amino acids (aa) but differs in the 69 C-terminal aa [3]. Due to the fact that sGP is produced in greater abundance than GP1/2, and since the proteins share a common N-terminus, it is speculated that sGP functions as a decoy molecule for EBOV-specific neutralizing and non-neutralizing antibodies [4]. In addition, recent studies have shown that sGP actively subverts the host immune response to induce cross-reactivity with epitopes it shares with membrane-bound GP1,2 [3]. Given that the recombinant glycoprotein coding sequence used to transfect the HEK293T cells was modified to prevent the transcription of the smaller sGP and ssGP variants, we were surprised to see antibody reactive fragments at molecular weights much lower than the monomeric mass of the full length protein detected by western blot. It was hypothesized that these N-terminal fragments would not contribute to the development of effective antibodies in a vaccinated host and would confound the accurate quantitation of full length GP with our chosen peptide standards. This was confirmed during initial quantitation experiments where the difference between the SEE and SVG peptide set quantitation (delta-S) differed by as much as 30%. Quantitation of the monomeric recombinant GP standard, however, resulted in a delta-S of less than 7%. In addition, the MS quantitation results were within 6% of the value obtained by amino acid analysis. These data suggest that during the production and purification of the VLP’s, shorter, N-terminal fragments of GP are incorporated into the particles and remain bound through an unknown mechanism. This is also supported by the observation that the delta-S values are consistent within each lot. Using this approach, mass spec analysis of all VLP lots resulted in a range of concentrations (0.176-1.8 mg/mL) with a replicate CV% across all lots of less than 8%. These results were then compared to a western blot-based strategy used for post-production quantitation and found that the full length GP values were vastly overestimated and that a portion of the calculated total protein concentration of the non-purified rGP standard was actually due to the presence of these N-terminal fragments. When the results of the MS quantitation were compared to the survival data of a mouse cohort vaccinated with various lots of eVLP, there was a strong correlation between the full length GP content and survival of the Ebola virus challenge, however there was no correlation when using the western blot quantitation results. The presence of these truncated forms is the most likely explanation for the observed variation between the SEE and SVG peptide sets in the VLP’s.


The use of a high resolution and accurate mass Orbitrap-based method proved to be essential for accurately quantitating GP in Ebola VLP’s. This method allowed detailed characterization of the highly variable Ebola glycoprotein so that potential issues could be addressed and accounted for prior to development of a more streamlined quantitation scheme. A close comparison of the mass spec based GP quantitation data with the survival data compiled from the VLP vaccine study shows that it is the relative amount of full-length GP present in each VLP preparation that correlates most highly with survival and hence is the best predictor of vaccine efficacy. This mass spectrometry method could serve as the first line of quality control for new eVLP preparations. In addition, given the reproducibility of the SVG and SEE peptide pairs chosen for this work, these peptides many prove to be useful markers of infection should a sensitive QqQ method be developed.

References & Acknowledgements:

1. Licata JM, Johnson RF, Han Z, Harty RN. Contribution of ebola virus glycoprotein, nucleoprotein, and VP24 to budding of VP40 virus-like particles. J. Virol. 2004 Jul; 78(14): 7344-51

2. Sanchez, A., et al., The virion glycoproteins of Ebola viruses are encoded in two reading frames and are expressed through transcriptional editing. Proc Natl Acad Sci U S A, 1996. 93(8): p. 3602-7.

3. Mohan, G.S., et al., Antigenic subversion: a novel mechanism of host immune evasion by Ebola virus. PLoS Pathog, 2012. 8(12): p. e1003065.

4. Martins, K.A., T.K. Warren, and S. Bavari, Characterization of a putative filovirus vaccine: virus-like particles. Virol Sin, 2013. 28(2): p. 65-70.

Financial Disclosure

Board Memberno

IP Royalty: no

Planning to mention or discuss specific products or technology of the company(ies) listed above: