Anal. Chem. 2009, 81, 7357–7362
Proteomic Profiling of a High-Producing Chinese Hamster Ovary Cell Culture Tyler Carlage,† Marina Hincapie,‡ Li Zang,† Yelena Lyubarskaya,† Helena Madden,† Rohin Mhatre,† and William S. Hancock*,‡ Biogen Idec, 15 Cambridge Center, Cambridge, Massachusetts 02142, and Barnett Institute, Northeastern University, 360 Huntington Avenue, Boston, Massachusetts 02115 The productivity of mammalian cell culture expression systems is critically important to the production of biopharmaceuticals. In this study, a high-producing Chinese hamster ovary cell culture which was transfected with the apoptosis inhibitor BclXL gene was compared to a low-producing control that was not transfected. Shotgun proteomics was used to compare the high and low-producing fed-batch cell cultures at different growth time points. The goals of this study were twofold; it would be of value to find a biomarker that could predict cell lines with higher growth efficiency and to gain mechanistic insights into the effects of the introduction of a foreign gene that is known to have growth regulating properties in human cells. A total of 392 proteins were identified in this study, and 32 of these proteins were determined to be differentially expressed. In the highproducing cell culture, several proteins related to protein metabolism were upregulated, such as eukaryotic translation initiation factor 3 and ribosome 40S. In addition, several intermediate filament proteins such as vimentin and annexin, as well as histone H1.2 and H2A, were downregulated in the high producer. The expression of these proteins may be indicative of cellular productivity. A growth inhibitor, galectin-1, was downregulated in the high producer, which may be linked to the expression of Bcl-XL. The molecular chaperone BiP was upregulated significantly in the high producer and may indicate an unfolded protein response due to endoplasmic reticulum (ER) stress. Several proteins involved in regulation of the cell cycle such as RACK1 and GTPase Ran were found to be differentially expressed, which may be due to a differentially controlled cell cycle between low- and high-producing cell cultures. The biopharmaceutical industry has invested heavily in mammalian cell culture platforms to express recombinant glycosylated protein products. Common platforms include Chinese hamster ovary (CHO), baby hamster kidney (BHK), and mouse hybridoma (NS0). The productivity1 of these cell cultures directly affects the overall profitability of a given biomanufacturing process. For this reason, there has been a significant investment within the industry to engineer high-producing cell culture processes in order to increase product yield. Some of these approaches use targeted * To whom correspondence should be addressed. E-mail: wi.hancock@ neu.edu. † Biogen Idec. ‡ Northeastern University. 10.1021/ac900792z CCC: $40.75 2009 American Chemical Society Published on Web 08/10/2009
genetic engineering strategies to gain more control over specific pathways within the cell, while other techniques aim to manipulate cellular metabolism, so that cells can process metabolites more efficiently and increase growth rates. For example, upregulation of pyruvate carboxylase in human HEK cells resulted in greater productivity and lower metabolite consumption, including reduced ammonia and lactate production. Glutamine synthetase has been transfected into NS0 cells to create a glutamine-independent phenotype, thereby reducing the growth-inhibiting buildup of ammonia in the cells.2 Other approaches have targeted the cellular machinery responsible for processing and secretion of proteins in the endoplasmic reticulum (ER) and Golgi apparatus, to improve efficiency of protein metabolism within the cells. For instance, overexpression of protein disulfide isomerase (PDI), a key protein involved in disulfide bond formation in the secretory pathway, was shown to increase specific productivity of a CHO cell line for the production of an antibody therapeutic by 40%.3 However, in another example, transfection of hybridoma cells with PDI failed to increase productivity of an antibody.4 Moreover, overexpression of the molecular chaperone glucose-regulated protein 78 (BiP), a key chaperone involved in protein folding in the ER, also yielded mixed results.5,6 These results indicate that cellular response to overexpression of selected control proteins may be cell line dependent. Another common engineering method used to increase cellular productivity is modification of cells with apoptosis inhibitors, to disrupt programmed cell death, a process which can inhibit growth in mammalian cell cultures. Apoptosis can be initiated by external stimuli such as nutrient depletion or metabolite accumulation.7-9 The mitochondrial pathway of apoptosis involves the controlled release of pro-apoptotic proteins such as cytochrome C from the intermembrane space of the mitochondria into the cytoplasm. These proteins (1) Elias, C. B.; Carpentier, E.; Durocher, Y.; Bisson, L.; Wagner, R.; Kamen, A. Biotechnol. Prog. 2003, 19, 90–97. (2) Bell, S. L.; Bebbington, C.; Scott, M. F.; Wardell, J. N.; Spier, R. E.; Bushell, M. E.; Sanders, P. G. Enzyme Microb. Technol. 1995, 17, 98–106. (3) Borth, N.; Mattanovich, D.; Kunert, R.; Katinger, H. Biotechnol. Prog. 2005, 21, 106–111. (4) Kitchin, K.; Flickinger, M. C. Biotechnol. Prog. 1995, 11, 565–574. (5) Robinson, A. S.; Bockhaus, J. A.; Voegler, A. C.; Wittrup, K. D. J. Biol. Chem. 1996, 271, 10017–10022. (6) Shusta, E. V.; Raines, R. T.; Pluckthun, A.; Wittrup, K. D. Nat. Biotechnol. 1998, 16, 773–777. (7) Al-Rubeai, M.; Singh, R. P.; Goldman, M. H.; Emery, A. N. Biotechnol. Bioeng. 1995, 45, 463–472. (8) Arends, M. J.; Wyllie, A. H. Int. Rev. Exp. Pathol. 1991, 32, 223–254. (9) Earnshaw, W. C.; Martins, L. M.; Kaufmann, S. H. Annu. Rev. Biochem. 1999, 68, 383–424.
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
7357
form a complex which activates the caspase cascade initiating apoptosis events.10,11 Apoptosis can lower cell viability in mammalian cell cultures, thereby reducing the overall productivity of the upstream process. Therefore the understanding of how apoptosis is controlled is of high importance to the biopharmaceutical industry. Several genes have been identified that demonstrate antiapoptotic activity, such as Bcl-2 and Bcl-XL.12 These proteins inhibit apoptosis by binding to pro-apoptosis proteins in the mitochondrial membrane or cytoplasm, thereby disrupting the caspase activation necessary for apoptosis to occur. The application of these inhibitors in CHO and BHK cell cultures has been studied. In some cases, overexpression of Bcl-2 and Bcl-XL prolonged cell viability; however, these results varied between different cell lines and conditions.13 In another study, transfection of CHO with Bcl-XL was shown to increase overall productivity by 80% through increased cell growth and specific productivity.14 The benefits of expressing apoptosis inhibitors in mammalian cell culture are not well understood; however, the targeting of such proteins continues to be employed as one strategy for increasing productivity of therapeutic proteins. While various process strategies have been applied to enhance production of recombinant therapeutic proteins from cell cultures, the fundamental understanding of the effects of cell culture engineering on global protein expression is lacking. Recent developments in proteomic technologies have enabled the protein profiling of sequenced and nonsequenced organisms under different culture conditions. CHO cell cultures studied using proteomics approaches include the effect of sodium butyrate treatment,15 low culture temperature,16 lactate metabolism,17 and cell growth rates.18 Proteomics has also been used to characterize expression changes associated with overexpression of 6-phosphogluconolactonase.19 The study of differential expression of proteins in highly productive cells will likely increase our understanding of how cellular engineering impacts internal cell biology or it can also potentially lead to the discovery of new targets for cellular engineering. Furthermore, protein expression in these cell cultures may provide new insights with regards to how cellular productivity is linked to the proteome measurement; this information could lead to the identification of relevant pathways involved in the cellular production of biopharmaceutical products or to the discovery of novel proteins, which can be used as biomarkers predictive of cellular expression system productivity. The aim of this study was to compare the protein expression of a high and low-producing CHO cell culture. Chinese hamster ovary cells were transfected with Bcl-XL and grown using an optimized media profile in order to enhance productivity. These cells showed higher cell growth properties (high producer) (10) (11) (12) (13) (14) (15) (16) (17) (18)
Hengartner, M. O. Nature 2000, 407, 770–776. Strasser, A.; O’Connor, L.; Dixit, V. M. Annu. Rev. Biochem. 2000, 69, 217–245. Fussenegger, M.; Bailey, J. E. Biotechnol. Prog. 1998, 14, 807–833. Mastrangelo, A. J.; Hardwick, J. M.; Zou, S.; Betenbaugh, M. J. Biotechnol. Bioeng. 2000, 67, 555–564. Chiang, G. G.; Sisk, W. P. Biotechnol. Bioeng. 2005, 91, 779–792. Yee, J. C.; de Leon Gatti, M.; Philp, R. J.; Yap, M.; Hu, W. S. Biotechnol. Bioeng. 2008, 99, 1186–1204. Baik, J. Y.; Lee, M. S.; An, S. R.; Yoon, S. K.; Joo, E. J.; Kim, Y. H.; Park, H. W.; Lee, G. M. Biotechnol. Bioeng. 2006, 93, 361–371. Pascoe, D. E.; Arnott, D.; Papoutsakis, E. T.; Miller, W. M.; Andersen, D. C. Biotechnol. Bioeng. 2007, 98, 391–410. Nissom, P. M.; Sanny, A.; Kok, Y. J.; Hiang, Y. T.; Chuah, S. H.; Shing, T. K.; Lee, Y. Y.; Wong, K. T.; Hu, W. S.; Sim, M. Y.; Philp, R. Mol. Biotechnol. 2006, 34, 125–140.
7358
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
compared to the control without Bcl-XL (low producer). We performed a differential proteomic measurement of the CHO cell cultures at different growth time points. Systematic analysis of the proteome was performed by shotgun proteomics, an approach which identifies proteins using nanoLC-MS/MS analysis of the proteolytic digest of the whole proteome. Differential expression was determined by comparison of spectral counts for each identified protein. The reason that the shotgun approach was chosen over the more common twodimensional gel electrophoresis methods is the higher throughput of the former method and the ability to be able to identify a large number of proteins in a single LC-MS experiment.20 EXPERIMENTAL SECTION CHO Cell Lines. Both cell cultures studied were derived from the same Chinese hamster ovary DG44 host cell line by stable transfection of a plasmid encoding genes for DHFR and a humanized recombinant fusion protein in Biogen Idec (Cambridge, MA). Fusion protein production was further amplified by cell line selection in increasing concentrations of methotrexate. In addition, the high-producing cell line was transfected with a plasmid encoding Bcl-XL and G418. Stable clones were selected in the presence of neomycin and methotrexate. Cell Culture. The fed-batch control culture was grown in a 2 L sparged B. Braun bioreactor (Sartorius, Goettingen, Germany) for 13 days using a proprietary custom in-house serum-free medium supplemented with complex hydrolysate. The high-producing culture was grown in a 200 L custom-made stainless steel stirred tank bioreactor for 16 days using a modified version of the media used for the control. The nutrient profile and complex media components are different between the cell cultures, with the high producing culture having media optimized for higher cell growth. The pH was controlled at 7.15 using sodium carbonate for both cultures. Cell number and viability were measured by trypan blue staining and using a Cedex automated cell counter (Innovatis, Bielefeld, Germany). A volume equivalent to 3E7 cells was sampled from the bioreactor at varying time points (days 0, 5, 10, and 13 for the lowproducing culture and 1, 5, 10, and 16 for the high producing culture). Samples were centrifuged at 500g for 10 min. After removal of the supernatants, the pellets were reconstituted in 5 mL of PBS and centrifuged again at 500g for 10 min. The obtained pellets were stored at -70 °C until further analysis. Cell Lysis. Cell pellets were thawed at room temperature and reconstituted in a lysis buffer consisting of 50 mM Tris, pH 7.5 and 0.1% Rapigest (Waters, Milford, MA). Samples were then sonicated in a water bath for 3 cycles of 15 s each. Following sonication, samples were centrifuged at 5 000g for 10 min. Supernatants were transferred to clean tubes. The total protein concentration of each cell lysate was measured by bicinchoninic acid (BCA) (Pierce, Rockford, IL) according to the manufacturer’s instructions. Samples were stored at -70 °C prior to tryptic digestion. Trypsin Digestion. A total of 20 µL of each CHO lysate, containing approximately 100 µg of total protein, were mixed with 45 µL of 8 M guanidine, 50 mM Tris pH 7.5, and 1 µL of 500 mM DTT and incubated at 60 °C for 15 min for denaturation and reduction. The proteins were alkylated by adding 5 µL of 300 mM iodoacetic acid to each sample and incubated at room temperature in the dark for 1 h. After the remaining iodoacetic acid was quenched using 5 µL of 500 mM DTT, the samples were desalted
using Microspin SEC spin-columns (BioRad, Hercules, CA) that were pre-equilibrated in 50 mM ammonium bicarbonate, pH 8.0. The desalted samples were brought to a final volume of 200 µL in 50 mM ammonium bicarbonate. A volume of 10 µL of trypsin was added to each, and digestion was conducted at 37 °C for 18 h, and 5 µL of 1% TFA was added to each sample after the digestion. LC-MS Analysis. All samples were analyzed in triplicate using a Ultimate 3000 nanoLC (Dionex, Sunnyvale, CA) interfaced with an LTQ linear ion trap mass spectrometer (Thermo Fisher Scientific, Waltham, MA). Mobile phases used were 0.1% (v/v) formic acid in water (A) and acetonitrile (B). A volume of 1 µL of peptide digest from each sample was first injected onto a CapTrap precolumn (Michrom Bioresources, Auburn, CA). The trap column was washed with 100% A at a flow rate of 20 µL/min for 10 min before the valve switched to connect CapTrap with a 0.075 mm × 150 mm C18AQ column (Michrom Bioresources, Auburn, CA) for separation at a flow rate of 300 nL/min. The separation gradient ran from 2% B to 35% B over 110 min and then increased to 90% B in 20 min. It was then held at 90% B for 35 min and back to 2% B in 5 min. The column was re-equilibrated for 60 min before the next injection. The Dionex HPLC was controlled using Chromeleon v.6.80, and the LTQ was controlled using XCalibur 2.0.6 software (Thermo Fisher Scientific, Waltham, MA). The electrospray conditions were as follows: spray voltage 1.70 kV, capillary voltage 48 V, tube lens 70 V, capillary temperature 225 °C. MS data were acquired using the data-dependent mode with one MS scan in the centroid mode, in the m/z range of 300-2000 amu, followed by 9 MS/MS scans of the 9 most intense peaks. Dynamic exclusion was enabled for a duration of 30 s and repeat count of 1. A normalized collision energy of 35%, isolation width of 2.0 m/z, and activation Q of 0.25 was used for each MS/MS scan. Protein Identification. Peptide sequences and proteins were identified by searching all MS2 spectra against theoretical fragmentation spectra of a mouse protein database (Swiss-Prot, updated in September, 2007, 12 902 entries) using the Sequest algorithm incorporated into the Bioworks software, version 3.1, SR1.4 (Thermo Electron, San Jose, CA). Database search parameters included carbamidomethylation of cysteines, ±1.4 and ±1.0 Da tolerance for precursor and product ion masses, respectively. Only peptides resulting from tryptic cleavages with up to one missed cleavage were searched. The Sequest results were filtered by correlation score (Xcorr) values selected to obtain highly confident peptide and protein identifications: Xcorr 1.9, 2.2, and 3.75 for singly, doubly, and triply charged peptide ions, respectively, and all with dCn g 0.1. Protein identifications were then validated using ProteinProphet software, accepting identifications made with g95% confidence. Only proteins found with two or more unique peptides were considered valid identifications. Assessment of Relative Abundance of Peptides and Proteins. The spectral counting method was used for estimation of relative peptide and protein abundance.21 This method uses the number of scans generated by the mass spectrometer for every peptide identified from a specific protein as a semiquantitative measurement of protein abundance and has been shown previously to be useful for comparing abundance between different samples in LC-MS experiments.
being transfected with the Bcl-XL gene to inhibit apoptosis and enhance cellular productivity (14). The media profile for the highproducer was also optimized in order to improve metabolite availability and limit bottlenecks to cell growth. The viable cell density (VCD) of both cell cultures was monitored on a daily basis using trypan blue staining and cell counting (see Figure 1). The control culture was harvested after 13 days and reached a maximum density of 5.8 × 106 cells/mL on day 10. The cells maintained densities >5 × 106 cells/mL until day 13, when it decreased to 3.6 × 106 cells/mL. This change also corresponds to a small decrease in cell viability (data not shown). The high producer was harvested after 16 days, and it reached a maximum density of 15.4 × 106 cells/ mL on day 7 and maintained a similar cell density until day 16. The recombinant fusion protein titer was determined by measuring fusion protein concentration in the secreted media at different days by affinity chromatography. Product titer was also higher in the high producer compared to the control (data not shown). Extraction of Proteins from CHO Cells. A robust and reproducible cell lysis method is imperative for proteomics. Many cell lysis techniques use strong detergents to solubilize proteins. While having a high efficiency, these methods are typically not compatible with mass spectrometric analysis. Hence, for solubilization of proteins from CHO cells, a mass spectrometry compatible detergent, Rapigest, was used.22,23 Cell pellets were reconstituted using a Tris buffer containing 0.1% Rapigest and subjected to three cycles of sonication to disrupt cell membranes. We evaluated the efficiency of this extraction method and compared it to a commercially available Mammalian Protein Extraction Reagent (M-PER) kit (Pierce). Five replicates of the sample CHO cell culture sample were analyzed using either method, and the protein concentration was measured by the BCA method. The
RESULTS Cell Growth and Specific Productivity. The high producer and control CHO cells are both DG44 clones, with the high producer
(19) Wang, Y.; Wu, S. L.; Hancock, W. S.; Trala, R.; Kessler, M.; Taylor, A. H.; Patel, P. S.; Aon, J. C. Biotechnol. Prog. 2005, 21, 1401–1411. (20) Hancock, W. S.; Wu, S. L.; Shieh, P. Proteomics 2002, 2, 352–359. (21) Liu, H.; Sadygov, R. G.; Yates, J. R., 3rd Anal. Chem. 2004, 76, 4193–4201.
Figure 1. Cell growth over time: viable cell density was measured by trypan blue staining using a CEDEX cell counter on various days for both control and high producer cell cultures. The solid line indicates the high producer, while the dashed line indicates the control. Table 1. Concentration of Cell Lysates Prepared by Sonication and M-PERa
average standard deviation % CV
sonication
M-PER
4.1 mg/mL 0.2 mg/mL 3.9%
4.0 mg/mL 0.4 mg/mL 9.2%
a Cell pellets were lysed using Pierce Mammalian Protein Extraction Reagent (n ) 5) and a sonication method using Rapigest (n ) 5). The total protein concentration was measured for each lysate using the Pierce BCA kit.
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
7359
Table 2. Proteins Identified in CHO Samplesa
control high producer
total
nucleus
cytosol
ribosome
cytoskeleton
mitochondrion
endoplasmic reticulum
352 339
105 97
67 75
44 51
47 37
34 31
24 20
a Identified proteins were categorized according to cellular compartment using DAVID (http://david.abcc.ncifcrf.gov/). All proteins were identified by at least two unique peptides.
Table 3. Differentially Expressed Proteinsa fold change day 5 alanyl tRNA synthetase T-complex protein 1 subunit delta T-complex protein 1 subunit eta eukaryotic translation initiation factor 3 subunit 5 epsilon (EIF3) glucose-regulated protein 78 (BiP) 60S ribosomal protein L30 40S ribosomal protein S6 40S ribosomal protein S7
Protein Metabolism 5.0 1.9 2.1 2.4
day 10
day 16/13
2.0 2.9 1.5 1.4
1.3 1.0 -1.1 2.7
2.0 1.5 2.1 1.2
2.6 4.8 1.6 1.2
2.8 2.4 1.5 2.1
histone H1.2 histone H2A type 1-F nucleosome assembly protein 1-like 1 heterogeneous nuclear ribonucleoprotein A2/B1
Transcription -2.3 -1.8 1.2 -1.1
-3.0 -4.0 1.3 -1.2
-3.2 -4.6 2.4 -2.1
annexin-A2 adenylyl cyclase-associated protein 1 filamin-A myosin-9 myosin regulatory light chain 2-B vimentin
Cytoskeleton -1.4 2.1 -1.1 -1.3 -1.2 -2.4
-1.3 1.6 -1.1 -2.3 -2.4 -6.2
-2.2 1.0 -2.1 -2.2 -3.4 -1.1
1.3 1.2 1.6
-1.4 -2.3 -1.4
-2.4
-2.8
2.3 2.9
1.9 1.2
-1.6 -2.7
Miscellaneous -1.2 3.1 3.9 -1.2 2.5 -2.1 -1.1 2.5
-1.1 1.6 2.6 -2.3 1.5 -5.5 -2.0 -1.1
-2.0 2.0 2.3 -1.3 1.3 -1.2 -2.1 -2.4
receptor for activated C kinase (RACK1) calcyclin GTP-binding nuclear protein Ran galectin-1
Cell Cycle Regulation 2.2 2.1 2.0 Cell Growth -1.3 Glycolysis
R-enolase glyceraldehyde-3-phosphate dehydrogenase chloride intracellular channel protein 1 dihydrofolate reductase recombinant fusion protein osteoclast-stimulating factor 1 phosphoserine aminotransferase proteasome activator complex subunit 1 prostaglandin E synthase 3 thioredoxin
a Differentially expressed proteins were identified by calculating the ratio of spectral counts for each protein at each timepoint. Proteins showing a 2-fold change up or down and which had a CV e 0.5 were considered as differentially expressed.
results are presented in Table 1 and indicate that the two methods yield a similar amount of total protein from CHO cells, while the sonication method was more reproducible between the five replicates. On the basis of these results, the sonication method using Rapigest for protein solubilization was used for proteomics analysis of the CHO lysates. Classification of Identified CHO Proteins. Cell lysates were treated with trypsin to digest the proteins into peptide fragments. (22) Arnold, R. J.; Hrncirova, P.; Annaiah, K.; Novotny, M. V. J. Proteome Res. 2004, 3, 653–657. (23) Yu, Y. Q.; Gilar, M.; Lee, P. J.; Bouvier, E. S.; Gebler, J. C. Anal. Chem. 2003, 75, 6023–6028.
7360
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
The resulting fragments were analyzed by LC-MS/MS, and sequence information was generated by searching against the Sequest database of protein sequences for identification. In this study, 392 proteins were identified with conservative criteria for protein assignment as well as the measurement of at least 2 unique peptides per protein (see the Experimental Section). To determine the cellular origin of the identified proteins, the DAVID bioinformatics tool was used to categorize proteins based on cellular compartment (http://david.abcc.ncifcrf.gov/). As shown in Table 2, a similar number of total proteins were identified in both cell cultures. More nuclear and cytoskeletal proteins were identified in the control cell culture, while more cytosolic and ribosomal
Figure 2. Proteins upregulated in the high-producer: The relative abundance of proteins was determined by spectral counts for (A) recombinant fusion protein, (B) RACK1, (C) BiP, and (D) R-enolase. The solid line indicates the results for the high-producer, while the dashed line indicates control. Error bars correspond to 1 standard deviation (n ) 3).
proteins were identified in the high producer. These results correlate with differential expression patterns that we identified with our proteomic measurements and discussed in the following section. Differential Expression between Control and High Producer. Differentially expressed proteins were identified based on the ratio of spectral counts between the control and high producer for each identified protein. The ratio of spectral counts was used to calculate fold changes between the control and high producer at day 5, day 10, and the end points (day 13 for control, day 16 for high producer). In addition, the coefficient of variance (CV) of the spectral count values was measured with three replicates for both the control and high producer cell lines. Proteins that showed a fold change of greater than 2.0 or less than -2.0 and that had a CV of less than or equal to 0.5 were identified as differentially expressed. A list of selected proteins with the greatest level of differential expression is shown in Table 3. A total of 32 differentially expressed proteins were identified. The major functionalities of these proteins include protein metabolism, cytoskeletal structure, and cell cycle control. Both glucose-regulated protein 78 (BiP) and the recombinant fusion protein showed the highest level of upregulation in the high producer, with fold changes over 2.0 at all three time points (see Figure 2). Other proteins such as 40S ribosome, eukaryotic translation initiation factor 3 (EIF3), and alanyl-tRNA-synthetase were upregulated to a lesser degree across all 3 time points. Several proteins were consistently downregulated at all 3 time points, such as histone H1.2, vimentin, and galectin-1 (see Figure 3). Other proteins such as the receptor for activated C kinase (RACK1), R-enolase, and calcylcin showed upregulation and certain time points and downregulation at others.
The upregulation of proteins such as alanyl-tRNA synthetases, EIF3, and 40S ribosome indicate that protein metabolism is increased in the high producer. These proteins play crucial roles in the translation of proteins. In addition, the recombinant fusion protein was detected at 2-3-fold higher levels in the high producer compared to the control. This indicates that the intracellular concentration of this protein is significantly higher in the high producer, which supports the increased productivity observed in the high-producing cell culture. Previous studies of Bcl-XL transfected CHO cells have shown an increase in specific productivity over nontransfected cells,14 and our study which uses the insights generated by proteomic measurements suggests that the increased levels of product are related to a greater level of protein biosynthesis under the fermentation conditions used in this study. The molecular chaperone BiP was significantly upregulated in the high producer. This is relevant since BiP is a key chaperone involved in protein folding in the ER. When unfolded proteins accumulate to a certain level in the ER, chaperones such as BiP are upregulated to clear the proteins from the ER for degradation.24 The upregulation of BiP may indicate ER stress in the high producer due to high intracellular concentrations of the recombinant fusion protein, which can lead to an unfolded protein response in the cell. The expression profiles for BiP and the recombinant fusion protein are similar, which is consistent with the suggestion that higher intracellular concentrations of the protein over time result in a proportional response by the cell to express BiP. Galectins are a class of carbohydrate-binding proteins that modulate various activities within cells, such as differentiation, (24) Patil, C.; Walter, P. Curr. Opin. Cell. Biol. 2001, 13, 349–355.
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
7361
Figure 3. Proteins downregulated in the high producer: The relative abundance of proteins was determined by spectral counts for (A) annexinA2, (B) histone H2A, (C) galectin-1, and (D) vimentin. The solid line indicates the results for the high producer, while the dashed line indicates the control. The error bars correspond to 1 standard deviation (n ) 3).
cell growth, apoptosis, and tumor progression.25 Galectin-1 has specifically been shown to inhibit cell growth and activate apoptosis in T cells and the expression of galectin-1 may be regulated by Bcl-XL.26 Galectin-1 was downregulated in the high producer, indicating that it may be responsible for inhibiting cell growth in the control, and thus it is a candidate as a biomarker for successful cell engineering and product yield. Several proteins that are involved in cell cycle regulation were differentially expressed in this study, including GTPase Ran, which has a role in the regulation of mitosis,27 and RACK1, a kinase receptor. RACK1 has been shown to inhibit the tyrosine kinase Src, which can lead to G0/1 cell cycle arrest.28 Both of these proteins showed a similar expression profile over time. They were upregulated in the high producer at days 5 and 10 and were downregulated at the end points. The differential expression of these proteins may indicate that the cell cycle is controlled differently in the high producer compared to the control. We also observed the differential expression of several cytoskeletal proteins, such as vimentin, annexin, myosin, and filamin. These intermediate filament proteins have several functions, including maintaining cell shape, intracellular transport, and formation of mitotic spindles during cell division. The differential expression of these proteins may also be related to control of the cell cycle and related to cellular productivity. (25) Yang, R. Y.; Liu, F. T. Cell. Mol. Life Sci. 2003, 60, 267–276. (26) Brandt, B.; Buchse, T.; Abou-Eladab, E. F.; Tiedge, M.; Krause, E.; Jeschke, U.; Walzel, H. Histochem. Cell. Biol. 2008, 129, 599–609. (27) Rensen, W. M.; Mangiacasale, R.; Ciciarello, M.; Lavia, P. Front. Biosci. 2008, 13, 4097–4121. (28) Mamidipudi, V.; Zhang, J.; Lee, K. C.; Cartwright, C. A. Mol. Cell. Biol. 2004, 24, 6788–6798.
7362
Analytical Chemistry, Vol. 81, No. 17, September 1, 2009
The downregulation of histones in the high producer could also be evidence of differential cell cycle control. Several histones, which are responsible for condensation of DNA into chromatin structures, were downregulated in the high producer. Lower expression of histones results in greater accessibility of DNA for transcription. Similar results were observed by Nissom et al., who observed downregulation of histone 1.2 in a high-producing CHO culture.18 CONCLUSIONS The application of shotgun proteomics to mammalian cell cultures is a useful tool for understanding how cellular productivity and protein expression are related and identifying proteins associated with cellular productivity. In this study, proteomics was used to identify several potential protein markers that may be indicative of high productivity in CHO cells. The upregulation of protein metabolism, as indicated by proteins such as eukaryotic translation initiation factor 3 and ribosome 40S, as well as the downregulation of intermediate filaments such as vimentin and annexin, and histones H1.2 and H2A may be related to productivity, possibly by way of differential cell cycle control. Other proteins such as galectin-1 and BiP were also potentially linked to productivity. The relationship between the expression of these proteins and cellular productivity will be verified in future experiments. ACKNOWLEDGMENT The authors thank Rashmi Kshirsagar, Jason Wong, and Correne George for their assistance with the cell culture experiments described in this paper. Received for review April 13, 2009. Accepted July 16, 2009. AC900792Z
