ACCELERATING DIGESTION OF PROTEOMES USING TRYPSIN ... · INTRODUCTION Mass spectrometry (MS)-based...
1
INTRODUCTION Mass spectrometry (MS)-based protein characterization is typically performed using the shotgun approach where proteins are subjected to enzymatic digestion or chemical hydrolysis to generate peptides. Reproducible, rapid and complete digestion of proteins in complex mixtures is crucial for in-depth protein identification and quantification in bottom-up proteomics. Conventional in-solution tryptic digestion is time-consuming and prone to artifacts, reducing dynamic range of MS detection in complex samples. Immobilized trypsin has several advantages, allowing higher protein to enzyme ratio, small sample volume, easy sample manipulation, limited autolysis products and shorter incubation times. Immobilized enzyme on magnetic particles offers a distinct advantage over soluble enzyme as they serve both as absorbers of microwave radiation that accelerates the digestion process and affinity probes that facilitate the separation of enzyme from products by using an external magnetic field. There are several different types of commercially available trypsin and numerous digestion procedures cited in literature. The objective of this study is to evaluate the relative digestion efficiency and fidelity of the different types of trypsin (modified, unmodified) and digestion strategies (in solution, solid phase and microwave-assisted). In addition to measuring protein and peptide yield, this study also analyzes the percentage of canonical sequences, missed cleavages and over-digestion products under different conditions over a time course with a goal of finding the optimal conditions for quantitative MS studies. The combination of the technologies employed here is particularly well suited to do this type of assessment. First, the depth of acquisition provided by the TripleTOF™ 5600 System makes it much more likely to observe multiple state variants of the desired canonical peptide, including partitioning to low stoichiometry species. Secondly, the combination of the use of sequence tags in the Sequence Temperature Value approach and use of feature probabilities to structure search space allows the Paragon™ Algorithm in ProteinPilot™ Software to look for very large numbers of digestion and modification variations without the usual loss in identification discrimination. The ProteinPilot Descriptive Statistics Template enables easy distillation of the massive amounts of identification data into informative plots and tables. METHODS Preparation of Trypsin Immobilized Magentic Particles (TIMPs): Unmodified porcine pancreatic trypsin (USBioAnalyzed/Affymetrix) was immobilized onto amine-functionalized superparamagnetic particles (BcMag amine- terminated magnetic beads,1 μm ; Bioclone Inc.) by cross-linking the primary amines of trypsin and the particles using bis-succinimidyl ester-activated PEG compounds. The amount of trypsin immobilized onto the magnetic particles was determined by measuring the protein amount in solution before and after the immobilization reaction using BCA assay. Comparsion of Different dDigestion Strategies: E.coli cell lysate was denatured using 1% N-octylglucoside (OGS), reduced with 50 mM TCEP and alkylated with 100mM iodoacetamide. In solution digestion was done using Proteomics grade trypsin (TPCK-treated, Sigma) at an enzyme: substrate ratio (E/S) of 1:20 at 37 ºC across 4 time points. Solid-phase digestion was done using 300 μg of TIMPS at 37 ºC and monitored across 5 time points. Microwave- assisted solid phase digestion was carried out using a domestic microwave with output power of 1200 W. Comparison of Different Types of Denaturants and Trypsin: 293T cell lysate was denatured using either 1% N-octylglucoside (OGS) or 0.1 %RapiGest followed by reduction and alkylation as above. In-solution digestion were done using either Promega Sequencing Grade Modified Trypsin (Trypsin 1) or Sigma Proteomics grade Trypsin (Trypsin 2) at an E/S of 1: 20 at 37 ºC. The digestion was monitored every 10 min up to 60 min followed by 90 min, 3h, 6h and 16h. The samples were prepared in triplicate on three separate days. Chromatography: The sample was analyzed using the Eksigent nanoLC-Ultra ® 2D System plus combined with the cHiPLC ® -Nanoflex system configured in trap elute mode (Eksigent, Dublin, USA). The cell lysate (1 ug total protein) was loaded at 2 μL/min onto a 200 μm x 0.5 mm precolumn and eluted at 300 mL/min onto a analytical 75 μm x 150 mm column. Both pre-column and analytical column were filled with ChromXP C18-CL 3μm 120 Å phase (Eksigent). An elution gradient of 10-35% acetonitrile (0.1% formic acid) with a 60 min gradient was used. Mass Spectrometry: Data was acquired on the TripleTOF™ 5600 System in IDA mode using an acquisition rate of 20 MS/MS per cycle x 50 msec minimum accumulation time. High sensitivity MS/MS mode was used so MS/MS spectra had resolution >15,000. Data Processing: All data was processed using ProteinPilot™ Software 4.0, using the integrated false discovery rate analysis. Results were then analyzed using the PSPEP Comparison Template and the ProteinPilot Descriptive Statistics Template with some additional canonical sequence analysis capabilities added in. ACCELERATING DIGESTION OF PROTEOMES USING TRYPSIN-IMMOBILIZED MAGNETIC NANOPARTICLES Christie Hunter 1 , Uma Kota 2 , Sean Seymour 1 , Mark Stolowitz 2 1 AB SCIEX, USA; 2 Canary Center at Stanford for Cancer Early Detection, Dept. of Radiology, Stanford School of Medicine, Palo Alto, CA, USA. CONCLUSIONS • Multiple digestion strategies and reagents were evaluated for their yield and fidelity. • Analysis of digestion time courses concurrent with the analysis of canonical peptides and their under- and over- digested counterparts is key to understanding and optimizing digestion for quantitative MS assays. • The combination of deep acquisition with the TripleTOF™ 5600 system and extensive identification provided by the analysis with the Paragon™ Algorithm enable this level of interrogation in a way that has not been possible before. • Use of Trypsin Immobilized Magnetic Particles (TIMPs) and the combination with microwave digestion shows a loss in peptide yield, possibly due to surface adsorption of some proteins/peptides to the magnetic particles or aggregation of proteins at high temperatures caused by microwave irradiation. Further improvements to the design and synthesis of TIMPs will be performed. • Use of lower grade trypsin that has not been TPCK treated shows a dramatic increase in over-digestion as expected. Similarly, longer digestion times can produce an increase in the sample preparation artifacts and therefore may be less desirable than previously believed. • Proteomics grade trypsin from Sigma showed a higher rate of missed cleavages across the time points suggesting that perhaps this enzyme had lower activity or required a higher protein to enzyme ratio (needs to be investigated). • N-Octylglucoside and RapiGest provided similar profiles in terms of proportion of digestion artifacts observed, with RapiGest providing a slightly higher yield of peptides (~10%). The current data do not enable a definite conclusion about the relative performance of the two denaturants. • Detailed characterization of sample preparation strategies are critical for MS quantitative studies, it must be reproducible and have high fidelity especially in targeted proteomics studies where absolute quantitative values are desirable. Human Cell Lysate TIMPs digestion at 37°C 5 time points LC/MS/MS Analysis 0.1% Rapigest Reduction/ Alkylation Dilute TIMPs + microwave- assisted 6 time points In-solution digestion 10 time points Sigma Trypsin In-solution digestion 10 time points Promega Trypsin Quench Reaction ProteinPilot™ Software Database Search ProteinPilot Descriptive Statistics Template Analysis 1% N-octylglucoside (OGS) Reduction/ Alkylation Dilute In-solution digestion 4 time points Sigma trypsin 1% N-octylglucoside Reduction/ Alkylation Dilute In-solution digestion 10 time points Sigma Trypsin In-solution digestion 10 time points Promega Trypsin E. coli cell lysate Figure 1. Overview of Experimental Design. Analysis of all samples was performed in the same way. Steps shown in grey boxes were held constant. Colored boxes indicate the parameters that were varied. Figure 2: Comparison of Yield from the Different Digestion Strategies. 20 μg of E.coli cell lysate was denatured, reduced and alkylated followed by digestion with soluble trypsin (E/S of 1:20) and 300 μg of TIMPs with or without microwave irradiation. Samples were randomized and analyzed in duplicate by LC/MS/MS. This plot shows the false discovery rates (FDR) at critical threshold values for both global and local error rate calculations at the protein, distinct peptides and spectral level indicating that in-solution digestion gives better yield compared to immobilized trypsin. The poor peptide yield from solid phase-based digestion methods could be due to non- specific sample absorption and/or protein aggregation on the TIMPs. Solution 6 h-Rep 1 Solution 6 h-Rep 2 TIMPs 6 h-Rep 1 TIMPs 6 h-Rep 2 TIMPs + Microwave 90 sec-Rep 1 TIMPs + Microwave 90 sec-Rep 2 Percent deviation versus Reference Figure 4: Digestion Map Comparing the Cleavage Frequency by Residue Pair for the Different Grades of Trypsin. Matrix view of the digestion frequency between each pair of residues for E.coli cell lysate digested in solution with TPCK-treated trypsin (A) or unmodified trypsin (B). Untreated trypsin has chymotrypsin activity that cleaves after Leu (L), Phe (F), Trp (W) and Tyr (Y) residues. The chymotrypsin activity in the untreated trypsin as seen in 4(B) explains the higher percentage of over-digestion observed in Figure 3B. A. Digestion frequencies for TPCK-treated trypsin B. Digestion frequencies for unmodified trypsin A. In solution digestion with TPCK-treated trypsin B. In solution digestion with unmodified trypsin Figure 3: Comparison of Digestion Efficiency between TPCK-treated and Unmodified Trypsin. In the case of TPCK-treated trypsin digestion (A), the percentage of canonical sequences shows a decrease at all time points, earlier time points were then used in subsequent studies. As digestion proceeds overnight, the percentage of canonical and under-digested peptides decreases with a concurrent increase in the percentage of over-digested products. A similar trend is seen for unmodified trypsin (B) in all three categories, however the percent over - digested peptides is much higher than the TPCK-treated trypsin at 1 h and continues to increases with longer digestion time. This drastic increase in over-digested products can be attributed to the chymotrypsin activity in the untreated trypsin as measured in Figure 4. RapiGest / Sigma Trypsin RapiGest / Promega Trypsin OGS / Sigma Trypsin OGS / Promega Trypsin Figure 5: Comparison of Different Digestion Reagents for Impact on Canonical Peptides. Modified trypsin from different vendors in combination with commonly used denaturants (OGS and RapiGest) was used to digest human cell lysate to understand their effect on the quality and efficiency of digestion. (Continued above) Figure 6: Characterization of Sequencing Grade Promega Trypsin. Analysis of the peptide yield (A) and canonical analysis done in Figure 6, indicate that the 3 hr digestion time point with either OGS or RapiGest produced a digest of highest fidelity. The breakdown of these peptide components as a function of precursor signal intensity is shown (B) and the fraction of peptides that are canonical (C) is plotted. As expected, there is a decrease in the fraction of peptides that are canonical as the less abundant proteins (measured as precursor signal) in the sample are analyzed. Figure 7: Analyzing Sample Preparation Artifacts Produced During Digestion Time Courses. In addition to the canonical peptide analysis, we looked at the increase in sample preparation modifications that could occur during digestion. Data shows a steady conversion of N-terminal Glu residues to pyro-Glu. Minimal changes are seen in the other sample preparation modifications analyzed. These trends looked similar between the four digestion conditions discussed in Figure 5, data shown here is from the Promega Trypsin / RapiGest digestion. RESULTS The relative efficiency of trypsin digestion under the many different conditions tested here was measured in terms of yield at the peptide and protein level, percentage canonical sequences, missed cleavages and over-digestion products as well as artifacts formed during the course of in-solution digestion. Comparison of the two denaturants relative to each other when the same enzyme was used shows very little difference between using RapiGest or OGS (RapiGest providing ~10% more peptides).However, comparing the two different types of trypsin shows that Promega trypsin produces fewer missed cleavages across all time points with little difference seen in the percentage of over-digestion components. A B C ACKNOWLEDGEMENTS This research is funded by the Canary Foundation. www.canaryfoundation.org We would like to thank Ken Lau for help with the sample preparation. TRADEMARKS/LICENSING For Research Use Only. Not for use in diagnostic procedures. The trademarks mentioned herein are the property of AB Sciex Pte. Ltd. or their respective owners. AB SCIEX is being used under license. © 2011 AB SCIEX.
Transcript of ACCELERATING DIGESTION OF PROTEOMES USING TRYPSIN ... · INTRODUCTION Mass spectrometry (MS)-based...
INTRODUCTIONMass spectrometry (MS)-based protein characterization is typically performed using the shotgun approach where proteins are subjected to enzymatic digestion or chemical hydrolysis to generate peptides. Reproducible, rapid and complete digestion of proteins in complex mixtures is crucial for in-depth protein identification and quantification in bottom-up proteomics. Conventional in-solution tryptic digestion is time-consuming and prone to artifacts, reducing dynamic range of MS detection in complex samples. Immobilized trypsin has several advantages, allowing higher protein to enzyme ratio, small sample volume, easy sample manipulation, limited autolysis products and shorter incubation times. Immobilized enzyme on magnetic particles offers a distinct advantage over soluble enzyme as they serve both as absorbers of microwave radiation that accelerates the digestion process and affinity probes that facilitate the separation of enzyme from products by using an external magnetic field.There are several different types of commercially available trypsin and numerous digestion procedures cited in literature. The objective of this study is to evaluate the relative digestion efficiency and fidelity of the different types of trypsin (modified, unmodified) and digestion strategies (in solution, solid phase and microwave-assisted). In addition to measuring protein and peptide yield, this study also analyzes the percentage of canonical sequences, missed cleavages and over-digestion products under different conditions over a time course with a goal of finding the optimal conditions for quantitative MS studies.The combination of the technologies employed here is particularly well suited to do this type of assessment. First, the depth of acquisition provided by the TripleTOF™ 5600 System makes it much more likely to observe multiple state variants of the desired canonical peptide, including partitioning to low stoichiometry species. Secondly, the combination of the use of sequence tags in the Sequence Temperature Value approach and use of feature probabilities to structure search space allows the Paragon™ Algorithm in ProteinPilot™ Software to look for very large numbers of digestion and modification variations without the usual loss in identification discrimination. The ProteinPilot Descriptive Statistics Template enables easy distillation of the massive amounts of identification data into informative plots and tables.
METHODSPreparation of Trypsin Immobilized Magentic Particles (TIMPs): Unmodified porcine pancreatic trypsin (USBioAnalyzed/Affymetrix) was immobilized ontoamine-functionalized superparamagnetic particles (BcMag amine- terminated magnetic beads,1 µm ; Bioclone Inc.) by cross-linking the primary amines of trypsin and the particles using bis-succinimidyl ester-activated PEG compounds. The amount of trypsin immobilized onto the magnetic particles was determined by measuring the protein amount in solution before and after the immobilization reaction using BCA assay.
Comparsion of Different dDigestion Strategies: E.coli cell lysate was denatured using 1% N-octylglucoside (OGS), reduced with 50 mM TCEP and alkylated with 100mM iodoacetamide. In solution digestion was done using Proteomics grade trypsin (TPCK-treated, Sigma) at an enzyme: substrate ratio (E/S) of 1:20 at 37 ºC across 4 time points. Solid-phase digestion was done using 300 μg of TIMPS at 37 ºC and monitored across 5 time points. Microwave- assisted solid phase digestion was carried out using a domestic microwave with output power of 1200 W.
Comparison of Different Types of Denaturants and Trypsin: 293T cell lysate was denatured using either 1% N-octylglucoside (OGS) or 0.1 %RapiGest followed by reduction and alkylation as above. In-solution digestion were done using either Promega Sequencing Grade Modified Trypsin (Trypsin 1) or Sigma Proteomics grade Trypsin (Trypsin 2) at an E/S of 1: 20 at 37 ºC. The digestion was monitored every 10 min up to 60 min followed by 90 min, 3h, 6h and 16h. The samples were prepared in triplicate on three separate days.
Chromatography: The sample was analyzed using the Eksigent nanoLC-Ultra® 2D System plus combined with the cHiPLC®-Nanoflex system configured in trap elute mode (Eksigent, Dublin, USA). The cell lysate (1 ug total protein) was loaded at 2 μL/min onto a 200 µm x 0.5 mm precolumn and eluted at 300 mL/min onto a analytical 75 µm x 150 mm column. Both pre-column and analytical column were filled with ChromXP C18-CL 3μm 120 Å phase (Eksigent). An elution gradient of 10-35% acetonitrile (0.1% formic acid) with a 60 min gradient was used.
Mass Spectrometry: Data was acquired on the TripleTOF™ 5600 System in IDA mode using an acquisition rate of 20 MS/MS per cycle x 50 msec minimum accumulation time. High sensitivity MS/MS mode was used so MS/MS spectra had resolution >15,000.
Data Processing: All data was processed using ProteinPilot™ Software 4.0, using the integrated false discovery rate analysis. Results were then analyzed using the PSPEP Comparison Template and the ProteinPilot Descriptive Statistics Template with some additional canonical sequence analysis capabilities added in.
ACCELERATING DIGESTION OF PROTEOMES USING TRYPSIN-IMMOBILIZED MAGNETIC NANOPARTICLES
Christie Hunter1, Uma Kota2, Sean Seymour1, Mark Stolowitz2
1AB SCIEX, USA; 2Canary Center at Stanford for Cancer Early Detection, Dept. of Radiology, Stanford School of Medicine, Palo Alto, CA, USA.
CONCLUSIONS• Multiple digestion strategies and reagents were evaluated for their
yield and fidelity.
• Analysis of digestion time courses concurrent with the analysis of canonical peptides and their under- and over- digested counterparts is key to understanding and optimizing digestion for quantitative MS assays.
• The combination of deep acquisition with the TripleTOF™ 5600 system and extensive identification provided by the analysis with the Paragon™ Algorithm enable this level of interrogation in a way that has not been possible before.
• Use of Trypsin Immobilized Magnetic Particles (TIMPs) and the combination with microwave digestion shows a loss in peptide yield, possibly due to surface adsorption of some proteins/peptides to the magnetic particles or aggregation of proteins at high temperatures caused by microwave irradiation. Further improvements to the design and synthesis of TIMPs will be performed.
• Use of lower grade trypsin that has not been TPCK treated shows a dramatic increase in over-digestion as expected. Similarly, longer digestion times can produce an increase in the sample preparation artifacts and therefore may be less desirable than previously believed.
• Proteomics grade trypsin from Sigma showed a higher rate of missed cleavages across the time points suggesting that perhaps this enzyme had lower activity or required a higher protein to enzyme ratio (needs to be investigated).
• N-Octylglucoside and RapiGest provided similar profiles in terms of proportion of digestion artifacts observed, with RapiGest providing a slightly higher yield of peptides (~10%). The current data do not enable a definite conclusion about the relative performance of the two denaturants.
• Detailed characterization of sample preparation strategies are critical for MS quantitative studies, it must be reproducible and have high fidelity especially in targeted proteomics studies where absolute quantitative values are desirable.
Human Cell Lysate
TIMPs digestion at
37°C5 time points
LC/MS/MS Analysis
0.1% Rapigest
Reduction/ Alkylation
Dilute
TIMPs + microwave-
assisted6 time points
In-solution digestion
10 time points Sigma Trypsin
In-solution digestion
10 time points PromegaTrypsin
Quench Reaction
ProteinPilot™ Software Database Search
ProteinPilot Descriptive Statistics Template Analysis
1% N-octylglucoside (OGS)
Reduction/ Alkylation
Dilute
In-solution digestion
4 time pointsSigma trypsin
1% N-octylglucoside
Reduction/ Alkylation
Dilute
In-solution digestion
10 time pointsSigma Trypsin
In-solution digestion
10 time points PromegaTrypsin
E. coli cell lysate
Figure 1. Overview of Experimental Design. Analysis of all samples was performed in the same way. Steps shown in grey boxes were held constant. Colored boxes indicate the parameters that were varied.
Figure 2: Comparison of Yield from the Different Digestion Strategies. 20 μg of E.coli cell lysate was denatured, reduced and alkylated followed by digestion with soluble trypsin (E/S of1:20) and 300 μg of TIMPs with or without microwave irradiation. Samples were randomized and analyzed in duplicate by LC/MS/MS. This plot shows the false discovery rates (FDR) at critical threshold values for both global and local error rate calculations at the protein, distinct peptides and spectral level indicating that in-solution digestion gives better yield compared to immobilized trypsin. The poor peptide yield from solid phase-based digestion methods could be due to non-specific sample absorption and/or protein aggregation on the TIMPs.
Solu
tion
6 h-
Rep
1
Solu
tion
6 h-
Rep
2
TIM
Ps 6
h-R
ep 1
TIM
Ps 6
h-R
ep 2
TIM
Ps +
Mic
row
ave
90 s
ec-R
ep 1
TIM
Ps +
Mic
row
ave
90 s
ec-R
ep 2Pe
rcen
t dev
iatio
n ve
rsus
Ref
eren
ce
Figure 4: Digestion Map Comparing the Cleavage Frequency by Residue Pair for the Different Grades of Trypsin. Matrix view of the digestion frequency between each pair of residues for E.coli cell lysate digested in solution with TPCK-treated trypsin (A) or unmodified trypsin (B). Untreated trypsin has chymotrypsin activity that cleaves after Leu (L), Phe (F), Trp(W) and Tyr (Y) residues. The chymotrypsin activity in the untreated trypsin as seen in 4(B)explains the higher percentage of over-digestion observed in Figure 3B.
A. Digestion frequencies for TPCK-treated trypsin
B. Digestion frequencies for unmodified trypsin
A. In solution digestion with TPCK-treated trypsin
B. In solution digestion with unmodified trypsin
Figure 3: Comparison of Digestion Efficiency between TPCK-treated and Unmodified Trypsin. In the case of TPCK-treated trypsin digestion (A), the percentage of canonical sequences shows a decrease at all time points, earlier time points were then used in subsequent studies. As digestion proceeds overnight, the percentage of canonical and under-digested peptides decreases with a concurrent increase in the percentage of over-digested products. A similar trend is seen for unmodified trypsin (B) in all three categories, however the percent over -digested peptides is much higher than the TPCK-treated trypsin at 1 h and continues to increases with longer digestion time. This drastic increase in over-digested products can be attributed to the chymotrypsin activity in the untreated trypsin as measured in Figure 4.
RapiGest / Sigma TrypsinRapiGest / Promega Trypsin
OGS / Sigma TrypsinOGS / Promega Trypsin
Figure 5: Comparison of Different Digestion Reagents for Impact on Canonical Peptides. Modified trypsin from different vendors in combination with commonly used denaturants (OGS and RapiGest) was used to digest human cell lysate to understand their effect on the quality and efficiency of digestion. (Continued above)
Figure 6: Characterization of Sequencing Grade Promega Trypsin. Analysis of the peptide yield (A) and canonical analysis done in Figure 6, indicate that the 3 hr digestion time point with either OGS or RapiGest produced a digest of highest fidelity. The breakdown of these peptide components as a function of precursor signal intensity is shown (B) and the fraction of peptides that are canonical (C) is plotted. As expected, there is a decrease in the fraction of peptides that are canonical as the less abundant proteins (measured as precursor signal) in the sample are analyzed.
Figure 7: Analyzing Sample Preparation Artifacts Produced During Digestion Time Courses. In addition to the canonical peptide analysis, we looked at the increase in sample preparation modifications that could occur during digestion. Data shows a steady conversion of N-terminal Glu residues to pyro-Glu. Minimal changes are seen in the other sample preparation modifications analyzed. These trends looked similar between the four digestion conditions discussed in Figure 5, data shown here is from the Promega Trypsin / RapiGestdigestion.
RESULTSThe relative efficiency of trypsin digestion under the many different conditions tested here was measured in terms of yield at the peptide and protein level, percentage canonical sequences, missed cleavages and over-digestion products as well as artifacts formed during the course of in-solution digestion.
Comparison of the two denaturants relative to each other when the same enzyme was used shows very little difference between using RapiGest or OGS (RapiGest providing ~10% more peptides).However, comparing the two different types of trypsin shows that Promega trypsin produces fewer missed cleavages across all time points with little difference seen in the percentage of over-digestion components.
A
B C
ACKNOWLEDGEMENTSThis research is funded by the Canary Foundation. www.canaryfoundation.orgWe would like to thank Ken Lau for help with the sample preparation.
TRADEMARKS/LICENSINGFor Research Use Only. Not for use in diagnostic procedures.The trademarks mentioned herein are the property of AB Sciex Pte. Ltd. or their respective owners. AB SCIEX is being used under license.© 2011 AB SCIEX.
