S-Trap eliminates cell culture media polymeric surfactants for effective proteomic analysis of mammalian cell bioreactor supernatants

Proteomic analysis of bioreactor supernatants can inform on cellular metabolic status, viability, and productivity, as well as product quality, which can in turn help optimize bioreactor operation. Incubating mammalian cells in bioreactors requires the addition of polymeric surfactants such as Pluronic F68, which reduce the sheer stress caused by agitation. However, these surfactants are incompatible with mass spectrometry proteomics and must be eliminated during sample preparation. Here, we compared four different sample preparation methods to eliminate polymeric surfactants from filtered bioreactor supernatant samples: organic solvent precipitation; filter-assisted sample preparation (FASP); S-Trap; and single-pot, solid-phase, sample preparation (SP3). We found that SP3 and S-Trap substantially reduced or eliminated the polymer(s), but S-Trap provided the most robust clean-up and highest quality data. Additionally, we observed that SP3 sample preparation of our samples and in other published datasets was associated with partial alkylation of cysteines, which could impact the confidence and robustness of protein identification and quantification. Finally, we observed that several commercial mammalian cell culture media and media supplements also contained polymers with similar mass spectrometry profiles, and we suggest that proteomic analyses in these media will also benefit from the use of S-Trap sample preparation.


Introduction
The biopharmaceutical market continues to grow, with current and emerging highly posttranslationally modified proteins driving research and development to optimize production efficiency while controlling quality [1][2][3] . Bioreactor operation is critical for optimizing production [4][5][6][7] , with changes in bioreactor operational conditions leading to changes in product yield and quality [7][8][9][10] . In addition to the product, characterizing the co-secreted proteome is especially important, since the complexity and content of the secreted proteome can advise on the metabolic status and needs of the cells, as well as impacting the efficiency of downstream product purification and characterization [11][12][13] .
Liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) is a versatile and effective tool to perform qualitative and quantitative measurements of proteins in a sample [14][15] . Recent years have seen remarkable progress both in the hardware and software underpinning proteomics 15 . Mass spectrometry proteomics can now provide a detailed overview and quantification of the proteins and their post-translational modifications in a wide variety of samples [16][17][18][19][20][21][22] . However, even the highest-performing mass spectrometry instruments require effective and robust sample preparation to enrich analytes of interest and deplete contaminants. Many methods have been developed to aid in the preparation of samples for mass spectrometry proteomics, with the optimal selection depending on the specific sample content and experimental questions at hand [23][24][25][26][27][28][29][30][31][32][33][34][35][36][37][38][39][40] .
Culturing mammalian cells in bioreactors requires the use of surfactants such as Pluronic F68 to protect the cells from hydrodynamic damage caused by agitation [41][42] . However, these compounds can interfere with LC-MS/MS analysis and are best eliminated during sample preparation 43 . Here we compared four different methods for mass spectrometry proteomics sample preparation: organic solvent precipitation [44][45][46] , filter-assisted sample preparation (FASP) [33][34] , S-Trap 29,32,[36][37][47][48] , and single-pot, solid-phase, sample preparation (SP3) [27][28] . We compared the ability of these methods to eliminate Pluronic F68 and similar polymers, and the number of proteins and peptides identified with each method. We found that FASP using Amicon columns was not an adequate method to prepare samples that contained these type of polymers (including bioreactor samples and commercial laboratory CHO medium). We also found that organic solvent precipitation, S-Trap, and SP3 were able to reduce the amount of polymer in the samples, but that SP3 and S-Trap performed better than precipitation. We also found that samples prepared with SP3 showed partial alkylation, a suboptimal feature for proteomic experiments. Together, S-Trap provided the most consistent polymer removal for robust, reproducible, and comprehensive proteomic analysis.

Spent media from CHO cells incubated in flasks were obtained as follows. Chinese
Hamster Ovary (CHO)-S cells were grown in suspension in CD CHO medium supplemented with 200 mM Glutamax. Cells were seeded at a density of 0.2x10 5 cells/mL and cultured at 37 °C, 7.5% CO2, and 130 rpm shaking. Cells were cultured for 3-5 days until reaching a density of 1-2x10 6 cells/mL. Cells were pelleted at 700 rpm for 10 min and the supernatant was removed for analysis as spent medium.
The spent media was obtained from CHO cell fed batch culture. A CHO K1SV cell line stably expressing a modified version of human Coagulation factor IX (rFIX, accession number P00740 UniProtKB, with Q2G and P44V amino acid substitutions) integrated using the glutamine synthetase expression system 49 , and also expressing the protease PACE/Furin (Accession number P09958, UniProtKB) (provided by CSL, Marburg, Germany) was seeded at 0.3x10 6 cells/mL in 3 L of CD CHO medium supplemented with 30 mg/L reduced menadione sodium bisulphite.
Bioreactors were fed EfficientFeedB as a daily bolus starting on working day (WD) 3 until WD10, up to the equivalent of 40% (1.2 L) to a total working volume of 4.2 L in the bioreactor, following manufacturer's instructions. Samples of 15 mL were collected, centrifuged at 2,000 rcf for 10 min, filtered through a 0.2 µm polyethersulfone filter (Pall), dispensed in matrix tubes to a volume of 0.5 mL and stored at -80 °C.
Saccharomyces cerevisiae were obtained as follows. S. cerevisiae strain YBS10 50 was grown at 30 ˚C overnight with shaking at 200 rpm in YPD (1% yeast extract, 2% peptone, 2% glucose). 3 mL of saturated culture was centrifuged at 18,000 rcf for 3 min at room temperature.
The medium was discarded and the cells were resuspended in 200 µL of ice-cold 50 mM HEPES buffer pH 7.4 containing 1 x protease inhibitor cocktail (Roche) and 1 mM phenylmethylsulfonyl fluoride. Cells were lysed by glass bead beating for 20 min at 4 °C. The lysate was transferred to a new microfuge tube and was clarified by centrifugation at 18,000 rcf for 1 min at room temperature. The supernatant was transferred to a protein LoBind tube (Eppendorf) and frozen at -20 °C. The organic solvent precipitation protocol was performed as previously described 51 , except as noted. Briefly, 300 µL of sample was clarified by centrifugation at 18,000 rcf for 3 min at room temperature, and 25 µL was transferred to a protein LoBind tube containing 200 µL of denaturation buffer (6 M guanidine hydrochloride, 50 mM Tris HCl buffer pH 8, 10 mM dithiothreitol (DTT)). Samples were incubated at 30 °C for 30 min in a MS100 Thermoshaker at 1500 rpm, acrylamide was added to a final concentration of 25 mM, and samples were incubated at 30 °C for 1 h in a Thermoshaker at 1500 rpm. DTT was added to an additional final concentration of 5 mM to quench excess acrylamide, and samples were precipitated by addition of 4 volumes of 1:1 methanol:acetone and incubation at -20 °C for 16 h. Samples were centrifuged at 21,000 rcf for 10 min at room temperature, the supernatant was removed, samples were centrifuged again for at 21,000 rcf for 1 min, and the remaining solvent was removed. Samples were air dried for ~15 min at room temperature, resuspended in 50 mM ammonium bicarbonate containing 0.5 µg of trypsin (T6567, Sigma) and incubated at 37 °C for 16 h in a Thermoshaker at 1500 rpm. All samples were desalted by ziptipping with C18 ZipTips (ZTC18S960, Millipore). For comparison of the extent of alkylation using SP3 or precipitation, the denaturation, reduction, and alkylation steps were performed identically to the SP3 protocol as described below, with a heat denaturation incubation step of 95 °C for 10 min.

Mass spectrometry sample preparation
The FASP protocol was performed as previously described [33][34][35] . Briefly, samples were centrifuged, denatured, reduced, and alkylated as described above for the precipitation. Briefly, 300 µL of sample was clarified by centrifugation at 18,000 rcf for 3 min at room temperature, and 25 µL was transferred to a protein LoBind tube containing 200 µL of denaturation buffer (6 M guanidine hydrochloride, 50 mM Tris HCl buffer pH 8, 10 mM dithiothreitol (DTT)). Samples were incubated at 30 °C for 30 min in a MS100 Thermoshaker at 1500 rpm, acrylamide was added to a final concentration of 25 mM, and samples were incubated at 30 °C for 1 h in a Thermoshaker at 1500 rpm. DTT was added to an additional final concentration of 5 mM to quench excess acrylamide. Samples were then loaded onto a 10 or a 30 kDa cut-off Amicon column and centrifuged at 10,000 rcf for at least 10 min. The filter was washed twice with 500 µL of 50 mM ammonium bicarbonate, and 150 µL of ammonium bicarbonate solution containing 0.5 µg of trypsin was added to the column. The column was then incubated overnight at 37 °C in a rotatory shaker. The column was transferred to a new collection tube, and the liquid was eluted by centrifugation at 10,000 rcf for at least 10 min at room temperature. To recover the digested peptides that were still trapped in the filter, 50 µL of 50 mM ammonium bicarbonate was added to the column, followed by centrifugation at 10,000 rcf for at least 10 min at room temperature.
Samples were desalted with C18 ZipTips. without agitation. To recover the peptides, the column was first rehydrated with 80 µL of 50 mM ammonium bicarbonate, and after 15 min of incubation at room temperature the columns were centrifuged at 1,000 rcf for 1 min at room temperature. This was followed by subsequent elution with 80 µL of 0.1% formic acid followed by 80 µL of 50% acetonitrile and 0.1% formic acid. All elutions were pooled, samples were dried in a MiVac Sample concentrator (SP Scientific), resuspended in 30 µL of 0.1% formic acid, and desalted with C18 ZipTips.
The SP3 protocol was performed as previously described 24  Inhibitor Cocktail (4603159001, Roche)) was added. Samples were incubated at 60 °C for 30 min or 95 °C for 10 min and then cooled to room temperature before acrylamide was added to a final concentration of 25 mM. After incubation at 30 °C for 1 h in a multivortex, DTT was added to a final concentration of 5 mM, samples were incubated at room temperature for 5 min. SP3 beads were prepared by mixing equal volumes of hydrophobic and hydrophilic beads into one Eppendorf tube, the tube was placed in a magnetic rack (Invitrogen Dynal), beads were pelleted, the liquid was removed by pipetting, and the beads were washed 4 times with 500 µL of water. Beads were reconstituted in water to a concentration of 20 µg/µL. Beads were kept at 4 °C until use. Denatured and reduced/alkylated protein samples were transferred to regular Eppendorf tubes. 5 µL of bead mix was added and mixed by pipetting, and 100% ethanol was added to a final concentration of 50% v/v. The mix was incubated at room temperature in a thermomixer at 1000 rpm for 25 min.
Tubes were centrifuged briefly to collect the liquid and beads, and the tubes were placed in a magnetic rack and incubated for 2 min until the beads settled. The supernatant was removed and discarded and beads were washed 3 times with 200 µL of 80% ethanol. 50 µL of 50 mM ammonium bicarbonate with 0.5 µg of trypsin was added to each tube. Tubes were sonicated at low intensity at 17 °C for 15 s in a water bath to reconstitute the beads, and were incubated overnight at 37 °C in a Thermoshaker at 1000 rpm. The tubes were centrifuged at 20,000 rcf for 1 min to pellet the beads, and placed in a magnetic rack for 2 min. When the beads settled the supernatant was recovered into a protein LoBind tube and peptides were desalted with C18 ZipTips.
Details of the methods, samples, and replicates are provided in Supplementary information.

Results and discussion
Our goal was to develop a simple, efficient, and robust method to prepare samples from bioreactor supernatant for mass spectrometry proteomic analysis that eliminated polymeric contaminants and achieved high performance protein identification and characterization. We first tested the FASP method using 0.5 mL 10 kDa cut-off Amicon columns. FASP is a commonly used sample preparation method for bottom-up proteomics that provides consistent and reproducible results for a variety of samples [33][34][35] , including bioreactor supernatants 56-60 . We processed 20 µL of filtered supernatant from a fed batch culture in CD CHO medium supplemented with CHO CD EfficientFeed B and vitamin K. We identified 526 proteins and 2441 distinct peptides in the sample (1% Global FDR) (Fig. 1A). However, we also observed the presence of an abundant polymer  (Fig. 1B). We therefore concluded that there was a polymer in the bioreactor samples that the regular FASP protocol was not able to eliminate.  The polymer in the bioreactor sample was likely Pluronic F68, which is a common media additive. However, we were intrigued by our observation that a polymer with similar characteristics was also present in filtered spent CD CHO medium supplemented with anticlumping agent (ACA) from transient CHO cell transfections (Supplementary Figure S1). We therefore tested multiple fresh mammalian cell culture media or media supplements for the presence of a similar polymer (Fig. 1). Indeed, a polymeric contaminant with a similar chromatography and fragmentation profile was observed in fresh CHO CD EfficientFeed B and CD CHO, and less abundantly in DMEM (Fig. 1C, D, and E). CHO CD EfficientFeed A, ACA, EMEM, and Optimem showed little or no polymeric contaminant (Fig. 1C, D, Figure S3). We concluded that FASP using Amicon or Microcon columns was not suitable to prepare bioreactor supernatant samples for proteomic analysis.
To change strategy, we compared the FASP protocol to three different bottom-up proteomics sample preparation methods: organic solvent protein precipitation [44][45][46] , S-Trap columns 29,32,[36][37][47][48] , and SP3 with paramagnetic beads (Fig. 2 and Supplementary Table S1) [23][24][26][27][28][29][30][31] . The organic solvent used here was a 1:1 mixture of methanol and acetone [44][45] . All methods have been previously shown to eliminate detergents from diverse samples [24][25][46][47][48] , and organic solvent precipitation [61][62][63][64] and FASP 56-60 have been previously used to prepare bioreactor supernatant samples for LC-MS/MS proteomics. We found that all three methods could remove the polymer(s) from the samples and that SP3 and S-Trap were better than precipitation (Fig. 2B-E). SP3 was recently shown to eliminate polymeric contaminants such as Pluronic F68 from samples for mass spectrometry proteomics 25 , and we corroborated this result, although in our hands we observed inconsistent polymer removal ( Fig. 2C and Supplementary Fig. S4). On the other hand, S-Trap consistently achieved almost complete polymer removal from the samples over multiple trials (Fig. 2D). We also found that all three methods showed comparable total ion chromatograms (TICs), which had higher intensities than the FASP TICs ( Fig. 2A, orange) even though the same amount of starting material was used ( Fig. 2A). S-Trap also identified the same number or more proteins than the other three methods (Fig. 2F, G). Indeed, S-Trap has also been shown to provide better protein coverage and more robust bottom-up proteomic results than FASP for other sample types, such as mammalian whole cell lysates and milk fat globules 32,36 . The comparative protein and peptide identification results for all methods were consistent with prior publications, showing that S-Trap outperformed FASP and precipitation in terms of the number of proteins identified 23,29,36 (Fig. 2F). Further, S-Trap, SP3, and precipitation allowed identification of more peptides than FASP (Fig. 2G). Together, due to the more consistent polymer removal and improved proteomic results, we concluded that S-Trap was the method of choice to eliminate Pluronic F68 from bioreactor samples and to eliminate similar polymeric contaminants present in mammalian cell media and media supplements for mass spectrometry proteomic applications. SP3 is a recently developed sample preparation method for bottom-up proteomics [23][24][25][26][27][28][29][30][31] , and we verified that it was easy to perform and that it provided superior results to FASP (Fig. 2).
However, during the course of our analyses we uncovered a potential limitation of the SP3 technology: partial peptide alkylation (Fig. 3). A key step during sample preparation for bottomup proteomics is reduction and alkylation of cysteine residues. Cysteine alkylation ensures that disulfide bonds do not re-form after reduction, facilitating peptide identification and measurement.
Normally, cysteine alkylation is set as a fixed parameter during software identification searches under the assumption that alkylation has proceeded to completion and with the expectation that non-modified peptides will not be identified. However, most database searching software provides flexibility in these search parameters, allowing discovery of unexpected or missing modifications.
Using ProteinPilot (SCIEX) and Preview (Protein Metrics) we observed that while the mammalian cell culture samples processed using S-Trap, precipitation, and FASP showed complete alkylation, samples processed with the SP3 method consistently showed partial alkylation (Fig. 3, Supplementary Figure S5, and Supplementary Tables S1 and S2). In order to rule out an effect of polymeric surfactants or sample type on the partial-alkylation observed with SP3 we also tested yeast whole cell extracts from yeast grown in polymer-free YPD medium (Fig. 3). To control for the potential impact of differences in the sample preparation methods prior to loading the sample into the beads or columns, the same buffers and procedure to reduce and alkylate the yeast whole cell extract sample were used for the precipitation and SP3 methods ( Fig. 3 and Supplementary  Table S4). Analyses using Preview showed that duplicate yeast whole cell extract SP3 samples showed 61% and 90% of peptides alkylated compared to 100% for the same samples processed with S-Trap or precipitation. Notably, the abundance of some non-alkylated peptide variants in the SP3 samples was higher than the abundance of the equivalent alkylated peptides (Fig. 3A and D).  Table ST2) showed complete alkylation. The partial alkylation observed with SP3 sample preparation was not due to the type of sample, since we observed a similar partial alkylation in bioreactor supernatants, spent CD CHO medium from transfected CHO cells, and yeast whole cell extract. The partial alkylation was also not due to the presence of a polymer in the samples (yeast whole cell extract samples were prepared in the absence of any obvious polymeric compound), and was observed when the denaturation incubation condition was 60 °C for 30 min or 95 °C for 10 min (Fig. 3, Supplementary Tables S1 and S2). Finally, the partial alkylation was not due to differences in the denaturation, reduction, and alkylation steps between the protocols, because for analysis of yeast samples the precipitation and SP3 sample preparation steps were performed identically and simultaneously. To exclude the possibility that the partial alkylation we observed was a consequence of our sample preparation technique, we inspected published data that had been obtained using SP3 (PXD008698 24 , Supplementary Table S3). In agreement with our data, Preview analyses of this published data also showed partial alkylation in SP3 samples. For example, Preview found alkylation of only 53%, 76%, and 86% of 72, 63, and 59 peptides analyzed in three raw data files from this dataset. The partial alkylation observed with SP3 is intriguing, since the denaturation, reduction, and alkylation steps happen before the proteins are mixed with the beads (see the Experimental Section). Partial alkylation is problematic for the quality of protein identification, and especially for the robustness of peptide quantification. In addition, partial alkylation could be a symptom of other unanticipated (and at this point unclear) sample preparation shortcomings associated with SP3.

Conclusions
We compared four bottom-up proteomics sample preparation techniques to identify the optimal method for use with mammalian cell bioreactor culture spent media. We found that several commercial mammalian cell media and media supplements, including CD CHO medium and EfficientFeed B media supplement, contained a polymer with similar MS profiles to Pluronic F68.
We found that FASP was unable to remove polymeric surfactants from the samples, that precipitation could partially remove polymeric surfactants, and that SP3 could inconsistently remove polymeric surfactants. During the course of our analyses we observed that SP3 sample preparation was associated with partial alkylation of cysteines, and confirmed this result by analyzing previously published data from other laboratories. While the cause of partial alkylation is unclear and was not investigated in this work, it is a caveat of this technique that should be considered when choosing SP3 as a method for quantitative proteomics. We demonstrate that sample preparation using S-Trap gives consistent, robust, high quality mass spectrometry proteomic results, achieving effective removal of Pluronic F68 and other polymeric contaminants present in mammalian cell culture media.