Mitigating In-Column Artificial Modifications in High-Temperature LC–MS for Bottom–Up Proteomics and Quality Control of Protein Biopharmaceuticals

Elevating the column temperature is an effective strategy for improving the chromatographic separation of peptides. However, high temperatures induce artificial modifications that compromise the quality of the peptide analysis. Here, we present a novel high-temperature LC–MS method that retains the benefits of a high column temperature while significantly reducing peptide modification and degradation during reversed-phase liquid chromatography. Our approach leverages a short inline trap column maintained at a near-ambient temperature installed upstream of a separation column. The retentivity and dimensions of the trap column were optimized to shorten the residence time of peptides in the heated separation column without compromising the separation performance. This easy-to-implement approach increased peak capacity by 1.4-fold within a 110 min peptide mapping of trastuzumab and provided 10% more peptide identifications in exploratory LC–MS proteomic analyses compared with analyses conducted at 30 °C while maintaining the extent of modifications close to the background level. In the peptide mapping of biopharmaceuticals, where in-column modifications can falsely elevate the levels of some critical quality attributes, the method reduced temperature-related artifacts by 66% for N-terminal pyroGlu and 63% for oxidized Met compared to direct injection at 60 °C, thus improving reliability in quality control of protein drugs. Our findings represent a promising advancement in LC–MS methodology, providing researchers and industry professionals with a valuable tool for improving the chromatographic separation of peptides while significantly reducing the unwanted modifications.

: Predicted minimum lengths of the capillary connecting the trap and separation column S7 to avoid temperature mismatch Table S4: Recent publications involving peptide mapping of biopharmaceuticals S8 Stock solutions of seven iRT peptides 1 (LGGNEQVTR, GAGSSEPVTGLDAK, YILAGVENSK, TPVITGAPYEYR, ADVTPADFSEWSK, GTFIIDPGGVIR, LFLQFGAQGSPFLK) were mixed so that the concentration of each was around 0.125 µg/µL.The mixture was 4.4fold diluted for microflow analyses.The sample contained 20% acetonitrile and 0.001% PEG 20 000. 2 The aliquots of the mixture were stored at -20 °C.

Trypsin digestion of proteins of Francisella tularensis and Jurkat cells
The F. tularensis live vaccine strain (LVS) was obtained from Chamberlain medium culture with OD600 of 0.6 to 0.7.The pellet was lysed in 2% sodium deoxycholate at 70 °C for 5 min.The proteins were incubated with 250 UI of benzonase for 30 min at 37 °C in the presence of 100 mM Tris-HCl buffer, pH 7.5.
The Jurkat cells (ATCC TIB-152) were cultured in 150 cm 2 cultivation flasks (TPP) in an RPMI 1640 medium supplemented with 10% fetal bovine serum.The cells were washed with phosphate-buffered saline and lysed on ice with 2.5% sodium deoxycholate containing 125 UI/mL benzonase.The protein concentrations were determined using a bicinchoninic acid assay (Merck/Sigma Aldrich). 3e milligram of the proteins from both lysates was subjected to reduction of disulfides in 5 mM TCEP for 60 min at 37 °C.Free thiols were blocked in 15 mM iodoacetamide at room temperature within 30 min in the dark.Iodoacetamide was quenched by incubation in 20 mM L-cysteine.The mixtures were then diluted with 50 mM Tris-HCl buffer to reach a pH above 7.The proteins were digested using trypsin in a 1:50 enzyme-protein ratio at 37 °C overnight.The digests were acidified with trifluoroacetic acid (TFA) to quench the enzymatic reaction and induce precipitation of deoxycholic acid, which was subsequently extracted into ethyl acetate saturated with water. 4The remnants of ethyl acetate were evaporated in a vacuum centrifuge at 30 °C within 30 min.The peptides were desalted using a Pierce Peptide Desalting Spin Column (Thermo Fisher Scientific) according to the manufacturer's manual.The eluates were vacuum-dried.The dried peptides were dissolved in 0.1% aqueous TFA with 0.001% PEG 20 000 and stored at -80 °C till analysis. 2,5ypsin digestion of protein biopharmaceuticals Biopharmaceuticals were denatured in 5 M guanidinium chloride and reduced in 20 mM DTT for 30 min at room temperature.7][8] The denatured biopharmaceuticals were diluted with the Low-Artifact Digestion Buffer containing DTT so that its concentration was 2 mM.The proteins were digested by trypsin added in a 1:20 enzyme-protein ratio at 37 °C for 2 hours.The digests were acidified with TFA, supplemented with 0.001% PEG 20 000, and stored at -80 °C till analysis.The predicted length of a stainless-steel capillary to be necessarily placed in a column thermostat to preheat the 100% water mobile phase to   calculated using Equation 7 at various flow rates.A capillary passively heated in a forced-air thermostat with 0.1 mm inner and 1/16 in.outer diameter was considered.The following parameters were used: mobile phase density  = 1 g/mL; mean specific heat of water,  = 4.323 J/g °C and capillary resistance to heat transfer,  ℎ = 0.03 W/°C.The value of  ℎ is taken from Yan et al. 9 , which is relevant for a silicone oil bath.Since we used an air bath instead of a silicone oil one as a column thermostat, the calculation includes that a roughly 7-fold longer capillary is needed to approach the same temperature using an air bath. 10Note that the calculated length is the part of the capillary that is to be necessarily placed in a thermostat, while the entire capillary should be adequately larger to ensure the connection.  is lower than  ℎ by 1 °C because it approaches  ℎ asymptotically and the remaining 1 °C is the hardest to complete.It was reported that   can be up to 5 °C lower than the separation column temperature (or  ℎ ) without a noticeable drop in the separation performance. 9,11,12Likely, this results from the frictional heating that cancels the column cooling to a small extent. 13That is why we tolerated a little temperature mismatch, which enabled significant shortening of the connecting capillary.Peptides identified in 0.25-min bins during a 30-minute analysis of the F. tularensis digest injected using a system with a 2position/6-port valve situated between the autosampler and separation column.The valve directed the mobile phase flow either through a 2.1 × 5 mm BEH C4 trap column maintained at 35 °C, or through a bypass capillary.Peptides were loaded onto the trap column and isocratically eluted for 2 minutes with 0.5% component B into the separation column.The valve was then switched, and peptides retained in the separation column at 80 °C were separated using a 30-minute gradient (red).Peptides from the trap column were separated during the subsequent blank injection (yellow).For reference, peptides identified from a direct injection into the separation column with the same 2-minute isocratic hold, are also shown.The grey zone highlights peptides that were not effectively retained in either the trap or the separation column.These peptides did not benefit from the trap column but, due to their early elution, were not excessively exposed to the high separation column temperature.The percentage of peptides benefiting from the BEH C4 trap column installation was calculated by comparing the number of peptides unretained on the trap column to those identified via direct injection.Peptides eluted in the grey zone were not considered.The experiment was conducted three times; data from the second replicate are presented.

Figure S1 :
Figure S1: Isocratic separation of small organic analytes S11 Figure S2: Comparison of connecting capillaries S12 Figure S3: Portion of peptides effectively retained in the trap column upon injection S13 Figure S4: Peptides identified only as modified and sequence duplicates S14 Figure S5: Peak broadening after the trap column installation S15 Figure S6: Abundance of artifacts in analyses using the optimized temperature setting S16 Figure S7: Significance of in-column artificial modification for peptides of monoclonal antibody S17 Figure S8: Peak capacity and peak width distribution in model analyses of trastuzumab S18 Figure S9: Illustrative chromatograms with and without the trap column S19 Figure S10: Artificial modifications in four protein biopharmaceuticals S20 Supporting References S21

Figure S1 :
Figure S1: Isocratic separation of small organic analytes

Figure S3 :
Figure S3: Portion of peptides effectively retained in the trap column upon injection

Figure S5 :
Figure S5: Peak broadening due to the trap column installation

Figure S6 :
Figure S6: Abundance of artifacts in analyses using the optimized temperature setting

Figure S7 :
Figure S7: Significance of in-column artificial modification for peptides of monoclonal antibody

Figure S8 :
Figure S8: Peak capacity and peak width distribution in model analyses of trastuzumab

Figure S10 :
Figure S10: Artificial modifications in four protein biopharmaceuticals

Table S2 :
Settings of DDA experiments S5

Table S1 :
Ion source settings

b Artificial modifications and separation performance in peptide mapping of protein biopharmaceuticals
a) Maximum injection time of 75 ms was set for 240 min analyses to enhance method sensitivity.b) Exclusion time in seconds was calculated as gradient time in minutes divided by two.Only the most intense charge state of each peptide was subjected to fragmentation in all DDA experiments.

Table S3 :
Predicted minimum lengths of the capillary connecting the trap and separation column to avoid temperature mismatch

Table S4 :
Recent publications involving peptide mapping of biopharmaceuticals