Creating optimized peptide libraries for AMP discovery via PepSAVI-MS

Abstract

Antimicrobial peptides (AMPs) are host defense peptides with a range of functions/activities/modes of action that are ubiquitously expressed across all forms of life. Continued discovery of novel AMPs presents exciting opportunities to address evolving resistance to existing treatments in multiple fields, including agricultural pathogens/pests as well as antimicrobial and chemotherapeutics for human health. However, typical discovery methods including bioassay-guided fractionation and genome mining generally lack the capacity for robust AMP discovery in non-model/unsequenced organisms. PepSAVI-MS (Statistically guided bioactive peptides prioritized via mass spectrometry) was developed as an AMP discovery approach that addresses some of the limitations associated with these standard methods. PepSAVI-MS is a multi-pronged pipeline that includes peptide library creation, bioactivity screening, LC-MS analysis, and statistical modeling for putative AMP identification. The original implementation of PepSAVI-MS outlined strategies for the fractionation of plant extracts with strong cation exchange chromatography (SCX). Herein, we elaborate on recent improvements to peptide library creation through the use of orthogonal fractionation methods, specifically crude SCX chromatography and reversed-phase liquid chromatography (RPLC). This optimization of the “peptide library creation” step has demonstrated improvements for processing and AMP identifications via PepSAVI-MS.

1. Introduction

Antimicrobial peptides (AMPs) are host defense peptides that are expressed by all taxa and possess a broad range of activity profiles (Mahlapuu et al., 2016; Wang et al., 2015). Endogenous AMPs are typically small (<10 kDa) and cationic with a hydrophobic region (Tam et al., 2015). Additionally, AMPs are generally cysteine-rich with multiple disulfide bonds for enhanced stability against proteases, temperature, and pH. Overall these characteristics aid in the broad bioactivity profile of AMPs to position their emerging use as therapeutic agents (Harvey, 2008; Mahlapuu et al., 2016; Tam et al., 2015). Compilations of AMPs from bacteria, archaea, protists, fungi, plants, and animals can be found in ever-growing publicly available databases including the Antimicrobial Peptide Database (APD, https://aps.unmc.edu/AP/) (Wang et al., 2016, Wang, 2021). The APD includes bioactive natural AMPs with known amino acid sequences less than 100 residues in length and contains 3273 AMPs from six kingdoms (Wang et al., 2016). Other databases include the Data Repository of Antimicrobial Peptides (DRAMP, http://dramp.cpu-bioinfor.org/) (Kang et al., 2019), the Collection of Anti-Microbial Peptides (CAMP_R3, http://www.camp3.bicnirrh.res.in/) (Waghu et al., 2016), and the Database Linking Antimicrobial Peptides (LAMP, http://biotechlab.fudan.edu.cn/database/lamp/) (Zhao et al., 2013).

Traditionally, AMP discovery is facilitated through bioassay-guided fractionation and/or genome mining approaches. Bioassay-guided fractionation requires iterative rounds of fractionation and bioactivity screening to isolate the active compound (Sharma & Gupta, 2015; Strömstedt et al., 2013). Significant material and activity losses can occur with each round of fractionation resulting in depletion of the bioactive components within the extract (Weller, 2012). Bioactivity screening is both concentration and pathogen-dependent, so some present AMPs may remain undetected via bioassay-guided fractionation and often results in the identification of previously characterized and highly abundant AMPs (Weller, 2012). Overall, this method is laborious and requires large amounts of material without the guarantee of identifying novel bioactive AMPs (Sharma & Gupta, 2015; Strömstedt et al., 2013).

AMP discovery via genome mining relies on DNA/RNA sequencing of the organism, knowledge regarding biosynthetic pathways, and computational resources to aid in the identification of cryptic biosynthetic gene clusters producing bioactive peptides (Skinnider et al., 2015, 2016). Though this approach has been readily applied in bacterial and fungal species with well-annotated and predictable bioactive natural product biosynthetic gene clusters, various challenges remain when implementing this process in higher-order and non-model organisms, including plants. For example, there is limited access to high-quality genomic and/or transcriptomic resources for non-model organisms which limits genome mining capabilities. Additionally, higher-order organisms generally lack the typical tight clustering of AMP-producing gene clusters that is common in bacteria (Russell & Truman, 2020). Additionally, genome mining approaches typically do not employ any means to detect bioactivity.

While both bioassay-guided fractionation and genome mining are powerful approaches for discovery, methods that address their limitations help facilitate novel AMP discovery, especially in non-model organisms. PepSAVI-MS (Statistically guided bioactive peptides prioritized via mass spectrometry) screens and identifies bioactive peptides from natural sources while addressing the aforementioned limitations (Kirkpatrick et al., 2017). This pipeline includes the creation of a peptide library via the extraction of peptides from source material followed by crude fractionation, bioactivity screening of these fractions, LC-MS-based peptidomics, and statistical analysis for higher-throughput discovery of putative bioactive AMPs. PepSAVI-MS has enabled the discovery of several novel bioactive AMPs with antibacterial, antifungal, and/or anticancer properties from various plants and microbial secretomes, and has revealed new bioactivities for known AMPs (Kirkpatrick et al., 2017; Kirkpatrick, Parsley, Bartges, Cooke, et al., 2018; Kirkpatrick, Parsley, Bartges, Wing, et al., 2018; Moyer et al., 2019; Parsley et al., 2018). The utilization of mass spectrometry is very beneficial for high-throughput AMP identification and quantification as it can determine peptidyl abundance which is then correlated with the bioactive regions found from the bioassay. Additionally, PepSAVI-MS encompasses an AMP discovery workflow that utilizes direct bioactivity measurements with minimal fractions and avoids reliance on genomic information which is not always available or well-annotated for non-model organisms.

Initially, PepSAVI-MS employed a single iteration of crude fractionation via strong cation exchange chromatography (SCX) for peptide library creation (Kirkpatrick et al., 2017). This minimizes material requirements, decreases the probability of peptide degradation via denaturation, and increases bioassay throughput in comparison to the bioassay-guided fractionation method. However, we have vastly improved downstream processing by instead implementing a two-step orthogonal SCX and reversed-phase liquid chromatography (RPLC) strategy for peptide library creation. SCX is used to remove the bulk of unretained small molecules (with the exception of N-containing small molecules) and create a bulk ‘superfraction’ of cationic peptides. This SCX ‘superfraction’ is then fractionated via RPLC generating the peptide library for bioactivity screening and mass spectrometry (Figure 1). RPLC has multi-faceted advantages including 1) reducing labor-intensive sample handling steps to sublimate salts from SCX buffers, 2) decreasing false positives in bioassays resulting from residual salts in SCX peptide fractions, and 3) providing higher resolution separation compatible with downstream isolation efforts for characterization. Specifically, this improved library creation allows for easier identification of lower abundance peptides and/or multiple contributors to bioactivity. The orthogonal fractionation utilizing SCX and RPLC can be adjusted to employ different chromatography that would best suit the experiments of interest. Though there is a small risk for bioactivity loss following exposure to organic solvent in RPLC, AMPs that denature during RPLC would be difficult to isolate and characterize in requisite downstream experiments. Altogether, this additional fractionation is advantageous for the aforementioned PepSAVI-MS pipeline. This optimized peptide library creation for implementation in the PepSAVI-MS pipeline can be combined with in silico AMP predictions and downstream MS analyses as described in Culver et. al (Culver & Hicks, 2021). Herein, we detail the protocols for optimal peptide library creation from plants to be implemented within PepSAVI-MS or for other similar bioactivity-guided approaches aimed at novel AMP discovery.

Figure 1.

PepSAVI-MS pipeline highlighting the optimization of peptide library creation using SCX for ‘superfraction’ collection and RPLC for fractionation.

2. Extraction of Plant Peptides

The first step of the PepSAVI-MS pipeline is to extract peptides from plant material. Throughout the extraction, it is important to prevent peptide denaturation/degradation from warm temperatures and protease activity. It is necessary to prepare the extract for fractionation with dialysis as a preliminary means to remove small molecules. This process yields a crude extract for further fractionation to create the peptide library (Section 3)

2.1. Equipment

• 2.1.1Porcelain mortar and pestle or spice grinder (Hamilton Beach Fresh Grind Electric Coffee Grinder, ASIN B005EPRFKO)
• 2.1.2Analytical balance (Mettler Toledo, New Classic MF, Model #MS104S, Mass range 0.1 mg – 120 g)
• 2.1.3Stirring hot plate (Corning, 6795-420D, Item # UX-84303-40)
• 2.1.4
  Centrifuges

  • 2.1.4.1
    Superspeed centrifuge; temperature range: - 20 °C – + 40 °C; max speed 22,000 rpm (Thermo Scientific Sorvall RC 6 Plus Centrifuge, Product Code 12121680)
  • 2.1.4.2
    Large benchtop centrifuge; temperature range: - 9 °C – + 40 °C; speed range: 200 – 14,000 rpm (Eppendorf Centrifuge 5810 R, Catalog # 022625101)
  • 2.1.4.3
    Microcentrifuge; max speed 15,060 rpm (Eppendorf Centrifuge 5425 R, Catalog # 2231000815)
• 2.1.5Centrifuge bottles (Thermo Scientific Nalgene Oak Ridge High-Speed Polycarbonate Centrifuge Tubes, Catalog # 3118-0085PK)
• 2.1.6Stericup filtration (Corning, 150 mL Bottle Top Vacuum Filter, 0.22 μm Pore, PES Membrane, Low Binding Protein, Product # 431160)
• 2.1.7Vacuum pump (BrandTech Scientific, Vacuubrand ME1C, MFG # 2071103)
• 2.1.8Centrifugal filter devices (Millipore, Amicon Ultra-15, 30 kDa nominal MWCO, Regenerated Cellulose Membrane, Catalog # UFC203024)
• 2.1.9Plastic test tubes (Globe Scientific, 13 x 100 mm Plastic Test Tubes, 8 mL, Polypropylene, Item # 110445)
• 2.1.10Vacuum centrifuge (Labconco Acid-Resistant CentriVap Centrifugal Concentrator Catalog # 7810016 with CentriVap Cold Trap −50°C Catalog #7811021)
• 2.1.11pH meter (Thermo Scientific, Orion Combination pH Electrode, Refillable Ag/AgCl, Catalog # 9156BNWP)
• 2.1.12Dialysis tubing (Spectrum Labs, SpectraPor Float-A-Lyzer G2 500-1000 MWCO 10 mL devices, Catalog # 888-97239)
• 2.1.13Extra-long transfer pipettes (Thermo Fisher Scientific, Samco, 9”, Catalog # 262PK)
• 2.1.14Safe-Lock Microcentrifuge Tubes, Polypropylene (Eppendorf, 2mL, Natural color, Catalog # 20901-540)

2.2. Chemicals

• 2.2.1Dry ice
• 2.2.2Liquid nitrogen
• 2.2.3Milli-Q water
• 2.2.4Acetic acid (Millipore Sigma; CAS # 64-19-7)
• 2.2.5Roche EDTA-free protease inhibitor cocktail (Millipore Sigma; SKU 11873580001)
• 2.2.6Pepstatin (Sigma-Aldrich; CAS # 26305-03-3)
• 2.2.7Polyvinylpolypyrrolidone (Millipore Sigma; CAS # 9003-39-8)
• 2.2.8Ammonium formate (Fisher Scientific; CAS # 540-69-2)
• 2.2.9Formic acid (Millipore Sigma; CAS # 64-18-6)
• 2.2.10Ethanol (Fisher Scientific; CAS # 64-17-5)

2.3. Buffers / Solutions to be Prepared

• 2.3.1
  0.5 mg/mL pepstatin stock

  • 2.3.1.1
    Resuspend pepstatin in 80% ethanol and 20% Milli-Q water with final concentration of 0.5 mg/mL
• 2.3.2
  Extraction buffer

  • 2.3.2.1
    3 mL of 10% acetic acid solution per gram of pulverized material (i.e., 300 mL for 100 g of tissue)
  • 2.3.2.2
    1 tablet/50 mL buffer Roche EDTA-free protease inhibitor cocktail
  • 2.3.2.3
    1 mg/L pepstatin
  • 2.3.2.4
    3% polyvinylpolypyrrolidone (g/mL)
    Tip: If using plant seeds, you should prepare the buffer at 30 mL/g of pulverized seeds (i.e., 150 mL for 5 g of seeds) due to the lower water content in seeds vs. plant tissue.
• 2.3.3
  Dialysis buffer

  • 2.3.3.1
    5 mM ammonium formate, pH 2.7 using formic acid (tested with pH meter), 1 L for every 10 mL of the concentrated extract (i.e., 3 L for 30 mL concentrated extract).
    Tip: Need to prepare two volumes of dialysis buffer for the two dialysis steps.

2.4. Protocol: Extraction

• 2.4.1Grind plant tissue (aerial or root) material with a pre-chilled mortar and pestle that are nestled in dry ice, adding liquid nitrogen as needed such that the plant material remains frozen throughout the grinding process.

  Tip: If using plant seeds instead, use a room temperature spice grinder to pulverize the seeds for extraction.

• 2.4.2Transfer pulverized plant material to a tared and chilled weigh boat to measure and record the mass of the pulverized plant material.

  Tip: The target total mass for pulverized aerial/root tissue and seeds is approximately 150 and 5 g, respectively, which could take iterative rounds of grinding to this pulverize total mass. This should be a good estimate of tissue or seed mass for a preliminary exploration of AMPs in plant tissue, but additional material is likely needed for additional isolation and characterization if intriguing AMPs are found from the first iteration with the PepSAVI-MS pipeline.

• 2.4.3Store pulverized plant material at −80 °C until next steps.
• 2.4.4Add plant material to extraction buffer (3 mL/g for aerial/root tissue or 30 mL/g for seeds) and stir vigorously via stir plate for 15 min at room temperature.

  Tip: Use an oversized glass beaker to accommodate the volume of plant material and extraction buffer.

• 2.4.5Transfer to a cold room at 4°C and continue stirring vigorously via stir plate for 4 h.
• 2.4.6Transfer the plant extract into centrifuge bottles and centrifuge for 45 min, 13,200 RCF, and 4°C to pellet insoluble material.

  Tip: If a high-speed centrifuge is not available, it is possible to have a longer centrifugation step at slower speeds. This will generally increase the time required for the Stericup filtration (Section 2.4.7).

• 2.4.7Carefully decant supernatant into a Stericup filtration unit and vacuum filter liquid to remove residual insoluble material. Discard the remaining pellet of insoluble material.

  Tip: Depending on the viscosity of plant material, this filtration step can range from 5 min to several days for all material to be filtered. It is typically on the time scale of minutes, however, keep extract cold in an ice bath for a multi-day filtration step.

• 2.4.8Precondition 30 kDa MWCO centrifuge filter devices with Milli-Q water per manufacturer’s protocol. Discard the Milli-Q water flow-through before adding the sample (Stericup filtrate).

  Tip: The size of the MWCO centrifuge filter devices can change depending on the size of the peptide of interest.

• 2.4.9Centrifuge for 1 h at 3,200 RCF at 4°C. Repeat until all input is filtered removing large proteins from the extract.
• 2.4.10Combine all MWCO sample filtrate and record the volume. Distribute the extract equally into polypropylene test tubes and concentrate in a vacuum centrifuge to a 10-fold decrease in volume (i.e., 30 mL remaining from 300 mL starting volume).

  Tip: Monitor the volume of liquid remaining every few hours and pool material to ensure that it does not concentrate to dryness as this may lead to precipitation of peptides.

• 2.4.11Prepare the dialysis tubing according to the manufacturer’s instructions by filling the dialysis tubing with 10% ethanol and slowly stirring on a stir plate for 10 min. Repeat by filling the tubing with Milli-Q water and slowly stirring on a stir plate for 20 min.
• 2.4.12Fill dialysis tubing with the concentrated extract.

  Tip: If any precipitation is visible in concentrated extract, combine in a conical tube. Try to solubilize any precipitate via vortexing and then centrifuge to pellet any insolubilized precipitation. Add supernatant to the dialysis tubing and discard the pellet.

• 2.4.13Place filled dialysis tubing in fresh dialysis buffer (i.e., 3 L beaker) and dialyze with gentle stirring via stir plate in a cold room at 4°C for 4-6 h. Transfer to fresh dialysis buffer and continue to dialyze overnight (~16 h).
• 2.4.14Remove dialysis tubing from the buffer and remove dialyzed samples from the tubing using long transfer pipettes. Distribute the dialyzed sample equally into polypropylene test tubes and concentrate in a vacuum centrifuge. Pool material into 2 mL microcentrifuge tubes as the volume decreases to produce a concentrated crude extract (i.e., concentrate the 300 mL extraction buffer to 4 mL crude plate extract) that can be considered the “1X” concentration.

  Tip: Monitor the volume of liquid remaining every few hours and pool material to ensure that it does not concentrate to dryness as this may lead to precipitation of peptides.

• 2.4.15Store “1X” aliquot at −80°C until further use.

3. Peptide Library Creation

Following the extraction of peptides from plant material into the “1X” crude extract aliquot, two fractionation steps are needed to yield the peptide library for bioactivity screening via bioassay (Section 4) and peptidomic analysis (Section 5). Herein, we discuss the use of strong cation exchange chromatography (SCX) and reversed-phase liquid chromatography (RPLC) as orthogonal fractionations for peptide library creation, but recall that this can be adjusted to employ different chromatography as needed. The first step for peptide library creation is SCX fractionation (Figure 2A), which yields a ‘superfraction’ of the late-eluting fractions after the initial removal of unretained small molecules. This ‘superfraction’ is then desalted via solid-phase extraction (SPE) in preparation for further fractionation via RPLC (Figure 2B) to yield the small-increment fractions which compose the peptide library. The key advantages for the additional fractionation via RPLC are for removing residual salts from the SCX eluate and having higher resolution isolation.

Figure 2:

Sample gradients and UV traces (a) for SCX injection with 10-min fractions collected. The first 10 minutes are unretained small molecules, but the rest (10-30) are combined to create SCX ‘superfraction’ as highlighted with the red box. (b) for RPLC injection from SCX ‘superfraction’ with 1-min fractions collected to create peptide library. UV detector 220 nm (black) and 280 (pink) and Mobile Phase B gradient (blue).

3.1. Equipment

• 3.1.1Large benchtop centrifuge; temperature range: - 9 °C – + 40 °C; speed range: 200 – 14,000 rpm (Eppendorf Centrifuge 5810 R, Catalog # 022625101)
• 3.1.2Analytical balance (Mettler Toledo, New Classic MF, Mass range 0.1 mg – 120 g, Model #MS104S)
• 3.1.3Stirring hot plate (Corning, 6795-420D, Item # UX-84303-40)
• 3.1.4pH meter (Thermo Scientific, Orion Combination pH Electrode, Refillable Ag/AgCl, Catalog # 9156BNWP)
• 3.1.5Ultrasonic Cleaner (VWR, Symphony, Operating Frequency: 35 kHz, Catalog Code 97043-944-EA)
• 3.1.6High-Performance Liquid Chromatograph (Shimadzu, LC-20AT, with UV-Vis Detector SPD-20A)
• 3.1.7Fraction collector (Bio-Rad, Model 2110 Fraction Collector, Catalog # 7318122)
• 3.1.8Plastic test tubes (Globe Scientific, 13 x 100 mm Plastic Test Tubes, 8 mL, Polypropylene, Item # 110445)
• 3.1.9PolySULFOETHYL A column (PolyLC, 100 mm x 4.6 mm, 3 μm particles, Item # 104SE0303)
• 3.1.10pH Test Strips (VWR Chemicals BDH, pH range 1.7-3.8, Catalog # BDH35315.607)
• 3.1.11Vacuum pump (BrandTech Scientific, Vacuubrand ME1C, MFG # 2071103)
• 3.1.12Vacuum manifold (Phenomenex, 24-port Vacuum SPE Manifold, Part # VM24)
• 3.1.13Sep-Pak C₁₈ Vac Cartridge (Waters, 500 mg Sorbent per Cartridge, 6cc, SKU WAT043395)
• 3.1.14C₁₈ column (Phenomenex Jupiter, 5 μm, 300 Å, LC Column 150 x 4.6 mm, Fully Porous Silica, Part # 00F-4053-E0)
• 3.1.15Safe-Lock Microcentrifuge Tubes, Polypropylene (Eppendorf, 2mL, Natural color, Catalog # 20901-540)
• 3.1.16LC-MS Total Recovery Vials (Waters, 12x32 mm glass screw neck vial, PTFE/silicone septa, Catalog # 186000385c)

3.2. Chemicals

• 3.2.1Ammonium formate (Fisher Scientific; CAS # 540-69-2)
• 3.2.2Formic acid (Millipore Sigma; CAS # 64-18-6)
• 3.2.3Acetonitrile (HPLC Grade; Fisher Scientific; CAS # 75-05-8)
• 3.2.4Water (HPLC Grade and LC-MS; Fisher Scientific; CAS # 7732-18-5)
• 3.2.5Trifluoroacetic acid (LC-MS Grade; Fisher Scientific; CAS # 76-05-1)

3.3. Buffers / Solutions to be Prepared

• 3.3.15% formic acid: 5% formic acid in LC-MS Grade water
• 3.3.2SCX Mobile Phase A: 5 mM ammonium formate, 20% acetonitrile, pH 2.7 using formic acid (tested with pH meter)
• 3.3.3SCX Mobile Phase B: 500 mM ammonium formate, 20% acetonitrile, pH 3.0 using formic acid (tested with pH meter)
• 3.3.4SPE Wash: 0.1% formic acid
• 3.3.5SPE Elution: 80% acetonitrile with 0.1% formic acid
• 3.3.6RPLC Mobile Phase A: 5% acetonitrile with 0.1% TFA
• 3.3.7RPLC Mobile Phase B: Acetonitrile with 0.1% TFA

  Tip: Sonicate all mobile phases for 30 min to de-gas before use.

3.4. Protocol: Generate ‘Superfraction’ via Strong Cation Exchange (SCX)

• 3.4.1Thaw “1X” aliquot of extract and centrifuge in a large benchtop centrifuge for 5 min at 3,220 RCFto pellet any protein precipitate.
• 3.4.2Perform SCX to remove neutral and negatively charged small molecules from the “1X” crude extract. Extract (450 μL) is injected on PolySULFOETHYL A column with a flow rate of 0.5 mL/min with UV detector set at 220 nm and 280 nm. The 30-min gradient has the following steps: 0-5 min (0% SCX Mobile Phase B), 5-17 min (linear ramp from 0-100% B), 17-27 min (100% B), 27-29 min (linear ramp from 100-0% B), and 29-30 min (0% B) (Figure 2A). Between injections, allow column re-equilibration. UV absorbance is monitored at 220 nm and 280 nm to monitor the elution of peptides containing aromatic residues (280 nm) and confirm the increasing concentration of ammonium formate throughout the gradient (220 nm).

  Tip: Peptides that do not contain aromatic residues will be masked by ammonium formate.

• 3.4.3Using a fraction collector, collect 10-min fractions in polypropylene test tubes. The first 10-min fraction is unretained small molecules and is discarded. The three fractions collected after 10 minutes are combined as a ‘superfraction’ of approximately 15 mL (Figure 2A).

  Tip: The timing of fractions collected is adaptable depending on what the UV traces look like since there may be a more convenient place to switch tubes.

• 3.4.4Dry the ‘superfraction’ in a vacuum centrifuge to remove all of the acetonitrile.
• 3.4.5Resuspend dried ‘superfraction’ in 5 mL of 0.1% formic acid.

3.5. Protocol: Desalting via Solid Phase Extraction (SPE)

• 3.5.1Acidify ‘superfraction’ by adding small aliquots of 5% formic acid until pH is less than 3 when measured with a pH test strip.

  Tip: The volume needed will depend on the acidity of the sample, but this can be upwards of 1 mL to acidify to pH 3.

• 3.5.2
  Desalt the extract with solid-phase extraction on C₁₈ Sep-Pak (500 mg) with vacuum manifold.

  • 3.5.2.1
    Resin is conditioned with 6 mL of SPE Elution.
  • 3.5.2.2
    Equilibrate with 6 mL of SPE Wash.
  • 3.5.2.3
    Load acidified sample onto the cartridges and collect the flowthrough.
  • 3.5.2.4
    Repeat 3.5.2.3 twice to ensure all material is bound to the resin.
  • 3.5.2.5
    Wash resin with 18 mL of SPE Wash.
  • 3.5.2.6
    Elute peptides with 6 mL of SPE Elution.
• 3.5.3Dry eluate in a vacuum centrifuge.

3.6. Protocol: Reversed-Phase Liquid Chromatography (RPLC) Library Creation

• 3.6.1Resuspend desalted SCX eluate in 500 μL of RPLC Mobile Phase A.
• 3.6.2Perform RPLC on the entire desalted SCX eluate. Add the 500 μL of resuspended SCX eluate to an LC-MS vial and inject it on a Phenomenex Jupiter C₁₈ column at a flow rate of 1.0 mL/min with UV detector monitoring 220 nm and 280 nm and column oven temperature at 40 °C. The 42-min gradient has the following steps: 0-4 min (0% RPLC Mobile Phase B), 4-29 min (linear ramp from 0-40% B), 29-36 min (linear ramp from 40-80% B), 36-40 min (80% B), and 40-42 min (linear ramp from 80-0% B) (Figure 2B). Between injections, allow column re-equilibration.
• 3.6.3Using a fraction collector, collect 1-min fractions across full injection time into 2 mL microcentrifuge tubes and dry in a vacuum centrifuge.
• 3.6.4Reconstitute dried fractions in LC-MS grade water for the formation of the peptide library. Fractions 11-40 will be needed for the bioassay screening (Section 4) and LC-MS analysis (Section 5) to obtain peptide abundance data for statistical modeling.

4. Bioassay

Following the creation of the peptide library, all fractions should be screened for bioactivity. A variety of pathogens can be cultivated in the lab and assayed in a 96 well plate format, including antibacterial (Kirkpatrick, Parsley, Bartges, Wing, et al., 2018; Moyer et al., 2019; Moyer, Allen, et al., 2021; Moyer, Purvis, et al., 2021; Parsley et al., 2019), antifungal (Kirkpatrick, Parsley, Bartges, Cooke, et al., 2018; Parsley et al., 2018), and anticancer (Parsley et al., 2018) activities. Herein, we detail the use of a 96-well plate bioassay against E. coli utilizing optical density to determine cell growth and resazurin to estimate cell death. The relative bioactivity in library fractions will identify regions containing putative AMPs and guide statistical modeling efforts (Section 6).

4.1. Equipment

• 4.1.1
  Incubators

  • 4.1.1.1
    Isotemp Incubator (Fischer Scientific, Model 650F, Catalog # 11-690-650F)
  • 4.1.1.2
    Incubatory Shaker (New Brunswick Scientific, Classic C25KC, MFG # M1248-0010)
• 4.1.2Biosafety Cabinet (The Baker Company, SterilGARD Class II Type A2, Model VBM-400)
• 4.1.3Petri dish (Fisher Scientific, 100 x 15 mm, Sterile, Polystyrene, Stackable Lid Catalog # FB0875712)
• 4.1.4Inoculating Combi Loop (Fisher Scientific, HIPS, White, Flexible 1 μL/10 μL IWP, Catalog #22-363-601)
• 4.1.5Culture Test Tubes (Fisher Scientific, Sterile, Translucent Polypropylene with blue snap cap, 17x100 mm, Catalog # 14-956-1J)
• 4.1.6Spectrophotometer Cuvette (Globe Scientific, Micro, 1.5 mL, Polystyrene, 2 clear sides, Item # 112137)
• 4.1.7Plate and cuvette reader (Cambridge Scientific, Molecular Devices SpectraMax M5 Multimode Microplate Reader, ID # 17000)
• 4.1.896 Well Plates (Greiner, Microplate, 96 well, Polypropylene, Flat Bottom, Natural, Item # 655261)
• 4.1.9Microplate Lids (Corning, General Assay Microplate Low Evaporation Lids, with corner notch, Catalog # 07-200-598)
• 4.1.1015 mL Falcon tubes (Corning, 15 mL High-Clarity Polypropylene Conical Tube, Produce # 352096)
• 4.1.11Pipetting Reservoirs (Argos Technologies, 25 mL, Sterile, Individually Wrapped, Polystyrene, Catalog # EW-04395-23)
• 4.1.12
  Multi Channel Pipettes

  • 4.1.12.1
    12-Channel 30-300 μL (Eppendorf Research Plus MFG # 3125000060)
  • 4.1.12.2
    12-Channel 0.5-10 μL (Eppendorf Research Plus MFG # 3125000028)
• 4.1.13Large benchtop centrifuge; temperature range: - 9 °C – + 40 °C; speed range: 200 – 14,000 rpm (Eppendorf Centrifuge 5810 R, Catalog # 022625101)
• 4.1.14Microplate Rack (Fisher Scientific, Eppendorf, Catalog # 14-278-200)
• 4.1.15Parafilm (Fisher Scientific, Bemis Parafilm M Laboratory Wrapping Film, Catalog # 13-374-16)

4.2. Chemicals

• 4.2.1Müeller Hinton Agar (Fisher Scientific, BD Difco, Dehydrated Culture Media, Catalog # DF0252-17-6)
• 4.2.2Müeller Hinton Broth (Fisher Scientific, BD Difco, Dehydrated Powder, Catalog # DF0757-17-6)
• 4.2.3Milli-Q water
• 4.2.3E. coli freezer stock (Strain ATCC 29522)
• 4.2.4Ampicillin, sodium salt (VWR, CAS # 69-52-3)
• 4.2.5Water (LC-MS Grade; Fisher Scientific; CAS # 7732-18-5)
• 4.2.6Resazurin sodium salt (Millipore Sigma, BioReagent, CAS # 6275-13-8)

4.3. Buffers / Solutions to be Prepared

• 4.3.1
  Müeller Hinton Agar Plates

  • 4.3.1.1
    Suspend 38 g Müeller Hinton Agar powder in 1 L Milli-Q water.
  • 4.3.1.2
    Heat to boiling with gentle stirring to dissolve powder completely.
  • 4.3.1.3
    Sterilize media in autoclave at 121 °C for 15 min.
  • 4.3.1.4
    Pour a thin layer (~20 mL) of media into Petri dishes such that the entire plate is covered. Allow cooling in the biosafety cabinet until solidified.
  • 4.3.1.5
    Store at 4°C.
• 4.3.2
  1x Müeller Hinton Broth

  • 4.3.2.1 S
    uspend 21 g Müeller Hinton Broth powder in 1 L Milli-Q water.
  • 4.3.2.2
    Warm with gentle stirring to dissolve powder completely.
  • 4.3.2.3
    Sterilize media in autoclave at 121 °C for 15 min.
  • 4.3.2.4
    Store at room temperature.
• 4.3.3
  2x Müeller Hinton Broth

  • 4.3.3.1
    Suspend 42 g Müeller Hinton Broth powder in 1 L Milli-Q water
  • 4.3.3.2
    Warm with gentle stirring to dissolve powder completely
  • 4.3.3.3
    Sterilize media in autoclave at 121 °C for 15 min
  • 4.3.3.4
    Store at room temperature.
• 4.3.4
  Amp₁₀₀ stock

  • 4.3.4.1
    Dissolve ampicillin sodium salt in LC-MS grade water to yield a final concentration of 100 mg/mL as the Amp₁₀₀ stock
  • 4.3.4.2
    Dilute Amp₁₀₀ stock 1:100 with water for a working concentration of 1mg/mL ampicillin
• 4.3.5
  50 mM resazurin

  • 4.3.5.1
    Dissolve resazurin sodium salt with water to yield a final concentration of 50 mM
  • 4.3.5.2
    Filter sterilize 50 mM resazurin with 0.22-micron filter

4.4. Protocol: Bioassay

Tip: It is important to use sterile technique for all aspects of the bioassay protocol.

• 4.4.1Streak Müeller Hinton agar plate with E. coli from a freezer glycerol stock. Use the quadrant streaking method to yield isolated single colonies. Incubate overnight at 37°C.
• 4.4.2Store the streaked plate at 4°C for up to two weeks.
• 4.4.3Transfer a single colony from the streaked plate into a culture tube with 5 mL of Müeller Hinton Broth (MHB) using an inoculating stick. Incubate overnight at 37°C with shaking (250 rpm).
• 4.4.4Perform 10x dilution of overnight culture in a cuvette (1 cm path length) and measure optical density (OD) at 600 nm. Back calculate to determine the undiluted OD₆₀₀.
• 4.4.5Subculture from the overnight culture into a new culture tube with 5 mL total and a final OD₆₀₀ of 0.25.

  Tip: Use M₁V₁=M₂V₂ to determine the volume of culture needed.

• 4.4.6Incubate subculture for 1 h at 37°C with shaking (250 rpm). At 1 h, the OD₆₀₀ should be approximately 0.5.
• 4.4.7During the incubation period, prepare sterile 96-well plates with the peptide library fractions, controls, and blanks. Add 10 μL of fraction, 1:100 dilution of Amp₁₀₀ stock (positive control), or water (negative control) to test wells. 50 μL of 1x MHB is used as a blank. Each plate should have 4 wells of positive control, negative control, and blanks.

  Tip: To obtain results in triplicate, 3 plates are needed.

• 4.4.8Following the 1-h incubation, confirm culture density is approximately 0.5 OD₆₀₀ in a cuvette (1 cm path length).
• 4.4.9
  Using this starter culture, 2x MHB, and 1x MHB, make a ‘master mix’ of OD₆₀₀ 0.125 culture. For three replicates, prepare 8 mL of culture as follows:

  • 4.4.9.1
    Calculate the volume of culture required to prepare 8 mL of OD ₆₀₀ 0.125 ‘master mix’ and add to a 15 mL Falcon tube.
    Tip: Use M₁V₁=M₂V₂ to determine the volume of culture needed.
  • 4.4.9.2
    Add 2 mL of 2x MHB to Falcon Tube.
  • 4.4.9.3
    Bring the ‘master mix’ to 8 mL with 1x MHB.
• 4.4.10Using the sterile pipetting reservoir and multichannel pipette, add 40 μL of ‘master mix’ to all wells with peptide fractions, positive control, and negative control, discarding tips after each transfer.

  Tip: The ‘master mix’ is not needed in the blank controls which confirm that the media is not contaminated and be used as a measure of background absorbance.

• 4.4.9Spin plates briefly (15 s) in benchtop centrifuge to remove bubbles and to ensure the wells are coated with the sample.
• 4.4.10Check and record the initial optical density of the 96-well plates.
• 4.4.11Incubate the plates at 37°C and with shaking (250 rpm) for 4 h to grow cells to mid-log phase (OD₆₀₀=0.5). Wrap plate in parafilm to prevent wells from drying out and avoid splashing.
• 4.4.12Check and record the optical density following the 4-h incubation. The OD₆₀₀ should be approximately 0.5.
• 4.4.13Using the sterile pipetting reservoir and multichannel pipette, add 1 μL of 50 mM resazurin to every well.

  Tip: Examine the level of resazurin in pipette tips before transfer to ensure equal amounts are added to each well.

• 4.4.14Spin plates briefly (15 s) in benchtop centrifuge to remove bubbles and to ensure the wells are coated with the sample.
• 4.4.15Incubate the plates at 37°C with shaking (250 rpm) for 1 h. Wrap plate in parafilm to prevent wells from drying out and avoid splashing.
• 4.4.16Using plate reader, read fluorescence with λ_ex at 544 nm and λ_em at 590 nm. Save readings as *.csv.
• 4.4.17
  Determine the percent activity of each fraction using the below equation.

  $$PercentActivity=100\times \left[1-\left(\frac{{Fraction}_{raw}-{PostiviteControl}_{avg}}{{NegativeControl}_{avg}-{PositiveControl}_{avg}}\right)\right]$$

  Tip: It is beneficial to use Excel to template these calculations.

• 4.4.18Determine the average percent activity of each fraction. Format this data in a new *.csv file with row 1 with the fraction numbers (11-40) and row 2 with average percent activity from each fraction. This format will be needed for the PepSAVI-MS modeling.

5. Quantitation via LC-MS

Following the creation of the peptide library, fractions 11-40 are analyzed using LC-MS analysis to measure the relative abundance of peptidyl species. The LC-MS data is compiled using Progenesis QI for Proteomics software to provide chromatographic peak alignment, deisotoped peak lists, and quantification of peptides in each fraction. This data will be compiled with the bioassay results (Section 4) for statistical modeling (Section 6). As listed in a review of computational approaches for proteomic analysis, other publicly available software packages can output peptide lists equivalent to Progenesis QI ( Gracia & Husi, 2019). A key part of Progenesis that is needed for the PepSAVI-MS statistical modeling is that the ions are quantified before identification. Therefore, if using another integrated pipeline, this same approach to quantification should be implemented.

5.1. Equipment

• 5.1.1
  LC-MS/MS platform

  • 5.1.1.1
    nanoAcquity UPLC (Waters, Milford, MA)
  • 5.1.1.2
    TripleTOF 5600 mass spectrometer (AB Sciex, Framingham, MA)
• 5.1.2Symmetry C₁₈ trap column (Waters, Acquity UPLC M-Class, 100 Å, 5 μm, 180 μm x 20 mm, 2G, V/M, SKU 186007496)
• 5.1.3HSS T3 C₁₈ analytical column (Waters, Acquity UPLC M-Class, 100 Å, 1.8 μm, 75 μm x 250 mm, SKU 186007474)
• 5.1.4pH Test Strips (VWR Chemicals BDH, pH range 1.7-3.8, Catalog # BDH35315.607)
• 5.1.5LC-MS Total Recovery Vials (Waters, 12x32 mm glass screw neck vial, PTFE/silicone septa, Catalog # 186000385c)
• 5.1.6Progenesis QI for Proteomics software (Nonlinear Dynamics, v.2.0)

5.2. Chemicals

• 5.2.1Formic acid (LC-MS Grade; Fischer Scientific; CAS # 64-18-6)
• 5.2.2Acetonitrile (LC-MS Grade; Fisher Scientific; CAS # 75-05-8)
• 5.2.3Water (LC-MS Grade; Fisher Scientific; CAS # 7732-18-5)

5.3. Buffers / Solutions to be Prepared

• 5.3.1LC-MS/MS Mobile Phase A: Water with 0.1% formic acid
• 5.3.2LC-MS/MS Mobile Phase B: Acetonitrile with 0.1% formic acid

  Tip: Sonicate all mobile phases to de-gas before use.

5.4. Protocol: LC-MS Analysis

• 5.4.1Determine the most concentrated fraction based on the intensity of the 220 nm UV trace from the RPLC chromatogram. Acidify this peptide fraction and perform a dilution series with LC-MS grade 95% water/5% acetonitrile/0.1% formic acid, starting with the lowest concentration (i.e., 1:100) and increasing the concentration as needed until appropriate loading level is achieved. Confirm peptide fraction is acidified to pH less than 3 when measured with a pH test strip before injection.
• 5.4.2Inject acidified dilution series (1-10 μL injections depending on sample loop size) onto the nano-LC-ESI-MS/MS platform monitoring signal intensity to identify an appropriate loading level (i.e., low e7 TIC for AB Sciex TripleTOF 5600). Prior to the analytical column, injections are made to the Symmetry C₁₈ trap column for 3 min at a flow rate of 5 μL/min with 99% LC-MS/MS Mobile Phase A and 1% LC-MS/MS Mobile Phase B. Samples are then separated on the HSS T3 C₁₈ analytical column using a 60-min gradient at a flow rate of 0.3 μL/min. The gradient has the following steps: 0-1 min (5% LC-MS/MS Mobile Phase B), 1-40 min (linear ramp from 5-50% B), 40-41 min (linear ramp from 50-85% B), 41-45 min (85% B), 45-47 min (linear ramp from 85-5% B), and 47-60 min (5% B). The mass spectrometer is operated in positive polarity and high sensitivity mode. MS survey scans have a mass range of 350 – 1600 kDa, a survey spectrum acquisition time of 250 ms, and information-dependent acquisition (IDA) of MS/MS data using an 8-second dynamic exclusion window and a cycle time of 2 seconds. Features above the intensity threshold of 150 counts/second, with a charge state of +2 to +5, were fragmented using rolling collision energy (±5%).

  Tip: MS/MS data is supplemental for PepSAVI-MS statistical modeling but is not essential.

• 5.4.3Dilute the fractions 11-40 of the peptide library the same as the optimized, representative fraction. Inject diluted samples following the LC-MS/MS parameters listed above and rerun the representative fraction.

  Tip: It is essential for downstream data processing steps that all fractions are analyzed using consistent dilution ratios and injection volumes.

  Tip: If there is limited access to LC-MS, but a clear bioactivity region is observed in the bioassay, it is possible to limit the LC-MS data to the fractions of the bioactive region with a few surrounding fractions instead of all 30 fractions.

5.5. Protocol: Progenesis

• 5.5.1Open new Progenesis experiment and import the LC-MS/MS data from all fractions as *.wiff format for AB Sciex.

  Tip: Other data files can be processed via Progenesis, including data from Thermo instruments. For Thermo data files, they will be imported in *.raw format.

• 5.5.2Click “Start automatic processing” when the data is imported into Progenesis. Within this step, Progenesis will select an alignment reference, align the runs, and perform peak picking with maximum sensitivity. During this step, do not set up an experimental design as this will be completed in another section. After the automatic processing is complete, review the information and click “Section Complete” to progress to the next step.
• 5.5.3Review the automatic peak alignment and choose a reference run that contains a region of activity from the bioassay screening.
• 5.5.4Click “Section Complete” through the “Filtering” stage.
• 5.5.5Complete the “Experiment Design Setup” by manually grouping all fractions to the same condition.
• 5.5.6Click “Section Complete” through the “Review Peak Picking” stage.
• 5.5.7Export data at the “Peptide Ion Statistics” stage and save as *.csv file. This file should be formatted to have columns for m/z, retention time, mass, charge, and intensity from each fraction (11-40).

6. Statistical Modeling and Analysis

Data from the bioassay results (Section 4) and LC-MS analysis (Section 5) are integrated utilizing the R programming language with the PepSAVIms CRAN package (https://cran.r-project.org/web/packages/PepSAVIms), to compile a list of the Top 20 ranked candidates that are most likely to contribute to the bioactive properties as seen from the bioassay results (Kirkpatrick et al., 2017). This data can be used to prioritize further investigation, including isolating or synthesizing the peptide of interest for determining minimum inhibitory concentration, probing biological activity, or other downstream characterization approaches (Balboa & Hicks, 2021).

6.1. Equipment

• 6.1.1RStudio (v 3.5)
• 6.1.2PepSAVIms CRAN package (https://cran.r-project.org/web/packages/PepSAVIms)

6.2. Protocol: PepSAVI-MS Statistical Analysis

• 6.2.1Open PepSAVI-MS statistics CRAN package, publicly available at https://cran.r-project.org/web/packages/PepSAVIms (Kirkpatrick et al., 2017).
• 6.2.2Load the *.csv files from Progenesis and formatted bioassay results into R.
• 6.2.3
  Use the binMS function to combine peptide ion dataset entries that are the same. The R script notation for this function is as follows:

  $$\begin{gathered} \text{binMS}(\text{mass}\_\text{spec},\text{mtoz},\text{charge},\text{mass},\text{time}\_\text{peak}\_\text{reten},\text{ms}\_\text{inten},\text{time}\_\text{range}, \\ \text{mass}\_\text{range},\text{charge}\_\text{range},\text{mtoz}\_\text{diff},\text{time}\_\text{diff}) \end{gathered}$$

  In the above notation, ‘mass_spec’ represents the mass spectrometry data matrix from the *.csv file. The next five parameters (‘mtoz’, ‘charge’, ‘mass’, ‘time_peak_reten’, and ‘ms_inten’) correspond to the columns from the data matrix (m/z, charge, mass, retention time, and intensity from each fraction). Filters for retention time, mass, and charge are employed with the ‘time_range’, ‘mass_range’, and ‘charge_range’ parameters, respectively. The last two parameters (‘mtoz_diff’ and ‘time_diff’) are used to specify the largest difference in m/z rations and retention time shifts that is allowed for binning rows together.

• 6.2.4Typical AMPs have masses, charge states, and retention times of 1,000-15,000 Da, +2-9, and 15-45 min, respectively. These values should be used in the above notation for binning.
• 6.2.5The summary(binMS_out) function will summarize what was consolidated and report the number of features remaining after each data reduction criteria.
• 6.2.6
  Use the filterMS function to refine the binned mass spectrometry data for the compounds that are most likely to contribute to the bioactive regions. The R script notation for this function is as follows:

  $$\text{filterMS}(\text{msObj},\text{region},\text{border},\text{bord}\_\text{ratio},\text{min}\_\text{inten},\text{max}\_\text{chg})$$

  In the above notation, the output *.csv from the binMS function is the ‘msObj’ parameter. The ‘region’ parameter is used to specify the filtering region based on the observed bioactivity results and all other fractions are considered in the ‘border’ parameter. The ‘bord_ratio’ parameter is used to specify the maximum ratio of a compound’s abundance outside of the bioactivity region compared to the observed abundance of that feature. This step is critical for filtering out compounds that are unlikely to be the cause of the observed bioactivity. The last two parameters, ‘min_inten’ and ‘max_chg’ are used to eliminate noise by filtering a minimum intensity and maximum charge value, respectively.

• 6.2.7The summary(filterMS_out) function will summarize what was consolidated and report the number of features remaining after each data reduction criteria.
• 6.2.8
  Use the rankEN function for penalized linear regression to determine which of the candidate compounds are best matched to the observed bioactivity profile. The R script notation for this function is as follows:

  $$\text{rankEN}(\text{msObj},\text{bioact},\text{region}\_\text{ms},\text{region}\_\text{bio},\text{lambda},\text{pos}\_\text{only},\text{ncomp})$$

  In the above notation, the output *.csv from the filterMS function is the ‘msObj’ parameter. The ‘bioact’ parameter is the *.csv formatted bioassay results. The next two parameters (‘region_ms’ and ‘region_bio’) are used to specify that the mass spectrometry data and bioassay results need to be compared. The ‘lambda’ parameter is used to set the penalty parameter used for estimation; it is typically 0.01 for our data sets. To ensure only compounds that have a positive correlation are considered for modeling, the ‘pos_only’ parameter is needed. To specify the maximum number of compounds that will be reported, use the ‘ncomp’ parameter.

• 6.2.9This yields a list of ranked candidate compounds and the Top 20 are reported in the script. Candidates with multiple charge states ranked in the Top 20 should be prioritized for further investigation.

  Tip: Possible further investigation includes peptide isolation or synthesis for the determination of minimum inhibitory concentration (MIC) studies. AMP identification can be confirmed via downstream analyses (Culver & Hicks, 2021).

7. Summary

AMPs compose a diverse class of compounds with a wide range of potential therapeutic applications. Unfortunately, common discovery workflows, including bioassay-guided fractionation and genome mining, can be hampered by low-throughput processes, compound rediscovery, lack of requisite genomic/transcriptomic databases, or predictive approaches lacking a direct bioactivity measurement. These challenges are often exacerbated when working with non-model organisms, including plants, which lack high-quality resources and knowledge of biosynthetic pathways. This chapter highlights the improvements to the PepSAVI-MS pipeline, an alternative AMP discovery method, which seeks to address the abovementioned challenges in AMP discovery. PepSAVI-MS optimization includes the adaption of sample preparation to utilize RPLC fractionation to reduce bioassay interference from SCX buffers and improve separation quality. This will allow for more confident identification of bioactivity, including the peptide features that are most likely contributing to this activity. Putative AMPs prioritized via the PepSAVI-MS pipeline are then subjected to further characterization as desired. Overall, the flexibility of the PepSAVI-MS pipeline is key for its widespread use. This flexibility has been noted with its adaptable use of orthogonal fractionation with various chromatography as needed. Additionally, though we detail the use of the PepSAVI-MS pipeline with Sciex TripleTOF 5600, other available MS platforms with sufficient resolving power can be utilized for comparable data. This optimized workflow of PepSAVI-MS can also be implemented in combination with different processes, including in silico AMP predictions and downstream mass spectrometry analysis (Culver & Hicks, 2021).