Three protein extraction methods for mass spectrometry proteomics of fresh frozen mouse tissue

Human and animal tissues are frequently fresh frozen for archival storage. These specimens can be used for proteomic analysis to analyze proteome composition. A critical step in these proteomic analyses is the extraction of proteins from fresh frozen samples, which can affect the sampling of the proteomes. Here, we compare three different methods for protein extraction from fresh frozen mouse heart tissue: 1) extraction using SDS buffer followed by FASP purication, 2) extraction using SDS buffer followed by STrap purication, and 3) extraction using a guanidine hydrochloride buffer followed by in-solution digestion. Based on bicinchoninic acid assay, all three methods display similar recovery of total protein content. However, proteomics-based analysis identied far fewer proteins from the extraction guanidine hydrochloride. SDS-FASP identies as many proteins as the SDS-STrap method with good coverage of proteins from different cellular compartments.


Introduction
Human and patient tissue banks routinely keep fresh frozen tissue samples which can be further examined and investigated. Different extraction buffers and different puri cation methods have been proposed to maximize protein recovery and identi cation from fresh frozen tissue. Here, we present three different methods for protein extraction from fresh frozen mouse tissue, evaluating both the protein recovery and the protein identi cation by liquid chromatography mass spectrometry (LC-MS) proteomics.
The three methods are: 1) SDS-FASP ( lter-aided sample preparation). Extraction with a 2% SDS buffer and puri cation by FASP. 1 2) SDS-STrap. Extraction with a 5% SDS buffer and puri cation by S-Trap. 3) GH-IS. Extraction with a 6M guanidine hydrochloride extraction buffer with in-solution digestion.
We chose to study mouse heart tissue because it is composed both of cells and connective tissue and can serve as a model for multiple tissues, in which connective proteins may impede protein extraction from cells. We compared a detergent based buffer (SDS, sodium dodecyl sulfate) and a chaotropic agent containing buffer (GH, guanidine hydrochloride) for protein extraction. Detergents lyse cells by forming micelles and disrupting membranes, thus releasing contents within cellular organelles. Chaotropic agents, such as guanidine hydrochloride, tend to lyse cells by disrupting hydrogen bonding of water molecules and ionic interactions. We also compared different puri cation methods, by S-Trap (Proti ), FASP and by in-solution digestion. Presence of detergents used in the extraction process and of polymers will greatly reduce the sensitivity and accuracy of mass spectrometry proteomics. Here, the comparison of two different spin column-based methods of puri cation against an in-solution digestion method will allow determination the effect of sample puri cation on mass spectrometry protein identi cation. In addition, strong cation exchange (SCX) pre-fractionation of peptides before LC-MS proteomics further enhances the number of identi ed proteins in both the SDS-FASP and SDS-STrap methods.

Procedure
Preparation of fresh frozen tissue.
1. Euthanize mice and dissect the necessary tissue. Flash freeze tissue in liquid nitrogen. Store fresh frozen tissue at -80°C. All mouse experiments were carried out in accordance with institutional animal protocols, as approved by the Memorial Sloan Kettering Cancer Center Institutional Animal Care and Use Committee.
2. Before protein extraction, weigh fresh frozen tissue on an analytical balance and note tissue wet weight.
Protein Extraction Method 1: SDS Extraction with Filter-Aided Sample Preparation (FASP, adapted 1, 2 ) 3. Add 100μl of extraction buffer 1 to the fresh frozen heart tissue in a 1.7ml Eppendorf tube. 4. Grind tissue using a motorized tissue grinder vigorously for 3 minutes. 5. Add an additional 100μl of extraction buffer 1 to the tissue and repeat grinding for 3 minutes.
6. Add remaining extraction buffer (100μl extraction buffer per 5mg of tissue wet weight).
7. Repeat tissue grinding until large tissue chunks cannot be observed.
8. Transfer tissue suspension to a Covaris milliTUBE, with each milliTUBE holding up to 1ml of tissue suspension.
9. Sonicate suspension in Covaris Adaptive focused ultrasonicator with the following program: Time 5min. 30s constant pulse with peak watt of 130W and a pause of 30s in between each pulse. 10. Centrifuge tissue suspension at 18,000g for 10min to clear large particulates. 11. Transfer supernatant to a clean Eppendorf tube.
12. Take an aliquot of tissue lysate for quanti cation using Pierce BCA protein assay kit, following manufacturer's instructions.
13. For every 10μl of tissue lysate, add 500μl of urea buffer to dilute protein lysate. For downstream mass spectrometry analysis, 100-200μg of protein lysate is needed.
14. Load 500μl of diluted tissue lysate onto a Vivacon 500 lter unit. For each lter unit, load up to 500μl each time and not more than 200μg of protein in total.
16. Repeat loading and centrifugation for 500μl volumes of tissue lysate onto same lter unit, ensuring not more than 200μg of protein in total is loaded onto the same lter unit. Use multiple lter units if amount of protein to be puri ed exceeds 200μg.
18. Add 100μl of 100mM DTT in urea buffer. Incubate at room temperature for 20min. Centrifuge at 14,000g for 15min. Discard ow-through.
19. Add 100μl of 100mM IAA in urea buffer. Incubate at room temperature for 20min in the dark. Centrifuge at 14,000g for 15min. Discard ow-through. 35. Mass spectra raw les were analyzed using MaxQuant v1.6.0.16. 3,4 The target search database was the mouse proteome from SwissProt, containing isoforms (version November 2018), 5 supplemented with sequences of common contaminant proteins from cRAP. 6 Mass spectral analysis parameters included 2 maximum allowed missed cleavages and minimum peptide length of 7. Fixed modi cation of cysteine carbamidomethylation was set and up to 3 variable modi cations of methionine oxidation and asparagine, glutamine deamidation was set. Set precursor mass tolerance to 10 ppm and 1% false discovery rate (FDR) for PSM and protein levels.
36. For further data analysis: Venn diagrams were made using Venny version 2.1. 7 Bar graphs were plotted using Origin 2018 (Microcal). Gene Ontology Enrichment analysis was performed using PANTHER classi cation system. 8 Protein Extraction Method 2: SDS Extraction with S-Trap (Alternative to Steps 3-25. Adapted 9 ) 37. Add 100μl of extraction buffer 2 to the fresh frozen heart tissue in a 1.7ml Eppendorf tube. 38. Grind tissue using a motorized tissue grinder vigorously for 3 minutes.
39. Add an additional 100μl of extraction buffer 2 to the tissue and repeat grinding for 3 minutes.  73. Sonicate suspension in Covaris Adaptive focused ultrasonicator with the following program: Time 5min. 30s constant pulse with peak watt of 130W and a pause of 30s in between each pulse. 74. Centrifuge tissue suspension at 18,000g for 10min to clear large particulates.

Transfer supernatant to a clean Eppendorf tube.
76. Take an aliquot of tissue lysate for quanti cation using Pierce BCA protein assay kit, following manufacturer's instructions.
77. Adjust protein concentration of tissue lysate to 1μg/μl based on BCA assay results. Take 100μl (100μg) of protein for further steps.
78. Add 11μl of 100mM DTT in extraction buffer 2 to the tissue lysate. Incubate at 56°C for 1hr.
79. Leave solution to cool to room temperature before adding 12μl of 550mM IAA in extraction buffer 2. Incubate at room temperature for 30min in the dark.
80. Next, add 13μl of 100mM DTT in extraction buffer 2 to the tissue lysate. Incubate at 56°C for 30min. 81. Leave solution to cool to room temperature. Add 50μl of 0.1μg/μl of trypsin (Trypsin to Protein ratio 1:20) and dilute to a nal volume of 1ml with 50mM ammonium bicarbonate. Incubate at 37°C for 14-18hr.
82. After trypsin incubation, add formic acid to nal 1% v/v concentration. Proceed with solid phase extraction using C18 macrospin columns (Nest Group) following manufacturer's instructions.
83. Concentrate puri ed peptides by vacuum centrifugation. Store peptide pellets at -20°C. Proceed to step 26 if performing peptide pre-fractionation or directly to step 32 for LC-MS analysis.

Troubleshooting
Time Taken

Anticipated Results
We present results for the three different protein extraction methods from mouse heart tissue for mass spectrometry proteomics analysis. Based on tissue wet weight and the amount of protein extracted as measured by bicinchoninic acid (BCA) assay, we observed similar recovery e ciencies between protein extraction using SDS-GH-containing buffers, with 9.8%, 8.1% and 9.3% recovery e ciencies for 5%SDS, 2%SDS and GH buffers, respectively (Table 1). This suggests that both detergent and chaotropic agents can extract proteins from heart tissue at a relatively high e ciency.
When the extracted proteins are subject to liquid chromatography mass spectrometry (LC-MS) analysis, we however see a marked decrease in protein identi cations from the GH-IS method, with 1091 unique proteins identi ed from single dimension separation (Fig 1A). The SDS-FASP and SDS-STrap methods identi ed 1858 and 1596 unique proteins (Fig 1A). A majority of proteins, 926 of 2067 unique proteins, were identi ed from all 3 methods, while SDS-FASP method identi ed 342 proteins not identi ed in other methods while SDS-STrap identi ed 148 proteins not identi ed in other methods.
Peptides extracted by SDS-FASP and SDS-STrap methods were subjected to o ine strong cation exchange fractionation followed by LC-MS analysis. The use of 2 nd dimension fractionation over 6 fractions increased the number of unique proteins identi ed to 3224 and 3026 for SDS-FASP and SDS-Strap methods, respectively (Fig 1C). We observed comparable representation of proteins from different cellular compartments (Fig 1D) between the SDS-FASP and SDS-Strap method including the cytoplasm (73-74%), nucleus (32%), mitochondrion (26%), ribosome (5%), endoplasmic reticulum (12%) and plasma membrane proteins (6%).
The comparisons between the three different methods suggest that SDS-containing extraction buffers have a recovery e ciency comparable to GH containing extraction buffers. The use of FASP or S-Trap puri cation has minimal effect on the number of identi ed proteins and the use of 2-dimensional separation greatly increases protein coverage. The comparison of the different protein extraction methods will help us better understand how to approach analysis of fresh frozen tissues for clinical proteomics.
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium 11 via the PRIDE 12 partner repository with the dataset identi er PXD020256.