Here, we present a protocol to profile the interplay between host and pathogen during infection by mass spectrometry-based proteomics. This protocol uses label-free quantification to measure changes in protein abundance of both host (e.g., macrophages) and pathogen (e.g., Cryptococcus neoformans) in a single experiment.
The technological achievements of mass spectrometry (MS)-based quantitative proteomics opens many undiscovered avenues for analyzing an organism’s global proteome under varying conditions. This powerful strategy applied to the interactions of microbial pathogens with the desired host comprehensively characterizes both perspectives towards infection. Herein, the workflow describes label-free quantification (LFQ) of the infectome of Cryptococcus neoformans, a fungal facultative intracellular pathogen that is the causative agent of the deadly disease cryptococcosis, in the presence of immortalized macrophage cells. The protocol details the proper protein preparation techniques for both pathogen and mammalian cells within a single experiment, resulting in appropriate peptide submission for liquid-chromatography (LC)-MS/MS analysis. The high throughput generic nature of LFQ allows a wide dynamic range of protein identification and quantification, as well as transferability to any host-pathogen infection setting, maintaining extreme sensitivity. The method is optimized to catalogue extensive, unbiased protein abundance profiles of a pathogen within infection-mimicking conditions. Specifically, the method demonstrated here provides essential information on C. neoformans pathogenesis, such as protein production necessary for virulence and identifies critical host proteins responding to microbial invasion.
The prevalence of invasive fungal infections is vastly increasing and is correlated with unacceptably high mortality rates, most commonly reported in individuals with immunodeficient predispositions1. Cryptococcus neoformans is a notorious opportunistic fungal pathogen capable of intracellular survival within host macrophage cells. Inadequate antifungal intervention results in fungal dissemination and life-threatening manifestations of cryptococcal meningitis and meningoencephalitis2,3. The global increase in immunocompromised status has demanded a parallel increase in the use of antifungal agents, in which many fungal species, including C. neoformans, have increasingly evolved resistance towards4,5,6. Therefore, it is imperative to implement robust and efficient technologies to answer vital biological questions regarding host defence response and microbial pathogenesis.
The new age of technological advancement in mass spectrometry (MS), including the generation of powerful computational and bioinformatic pipelines, provides the foundation for an integrative vision for large-scale analysis of host-pathogen research7,8. Conventional pathogenesis-driven proteomic analysis commonly profiles the view of infection from either the host or pathogen perspective, including comprehensive methodologies such as protein correlation profiling, affinity chromatography combined with proteomics, and interactomics9. Investigations into the virulence of dangerous pathogens in a host system are of immense clinical importance; however, the application of a dual perspective analysis in a single experiment was formerly considered unattainable. For example, the pathogen’s perspective towards infection is often overwhelmed by highly abundant host proteins resulting in reduced sensitivity for the detection of low-abundant fungal proteins7. Furthermore, the high sample complexity invites many targets to investigate in a single experimental system and proves challenging to elucidate mechanisms of action for a specific pathogen protein.
Bottom-up proteomics is a popular MS technique that enables manageable sample preparation, in which peptides are generated by sequence-specific enzymatic digestion followed by liquid chromatography separation, identification, and quantification by MS10,11. Here, we present a method demonstrating a data-dependent acquisition strategy purposed to achieve an unbiased coverage of an infection-based proteome or ‘infectome’. Specifically, label-free quantification (LFQ) sheds the dependence on chemical or metabolic labels for robust and accurate identification of protein level changes across multiple proteomes, reducing sample handling and processing steps12,13. This universal application interrogates produced proteins at a given moment within a cell independent of any expected protein production; thus, novel insights may be discovered that are critical to infection.
The workflow described herein is optimized to explore protein level changes of C. neoformans during infection-mimicking conditions with host immune cells (Figure 1). Rather than relying on the isolation and separation of cell types, this approach extracts the host and pathogen proteome together, and utilizes bioinformatic separation using two organism-specific databases to distinguish species-specific protein production. This method offers advantages for an unlimited number of samples to be processed without the extra costly preparation steps necessary in isotope-based labelling studies or fractionation. Furthermore, this workflow supports optimized protein extraction protocols transferable to a wide range of fungal and bacterial pathogens capable of targeting and infecting host immune cells. Overall, this protocol outlines the steps to complete an unbiased protein extraction and sample processing for high-resolution MS, followed by data and statistical analysis, capable of providing a wealth of knowledge of fungal proteins significant for infection combined with comprehensive profiling of the host defense response.
An immortalized line of macrophages derived from BALB/c mice were used for the following protocol approved by the University of Guelph Animal Utilization Protocol 4193. Notably, other strains of mice or other sources of immortalized cells can be applied to the outlined protocol with sufficient testing to optimize the detailed parameters. The following protocol will navigate the steps beginning with a frozen vial of macrophage cells. Cells are stored in 10% FBS (fetal bovine serum), 1% L-glutamine and 5% Pen/Strep (Penicillin-Streptomycin) mixture to DMEM (Dulbecco's Modified Eagle Medium) and 20% DMSO (dimethyl sulfoxide).
1. Culturing of C. neoformans
2. Culturing of macrophage cells
NOTE: Ensure work environment is sterilized prior to cell culture work.
3. Infection of macrophage cells with C. neoformans
NOTE: Upon reaching 70-80% confluence, there will be approx. 1.2 x 106 macrophage cells per well. To achieve the desired multiplicity of infection (MOI) of 100:1, 1.2 x 108 fungal cells are required for each reaction. Cultures must be set accordingly in biological quadruplicate.
DISCLAIMER: A MOI of 100:1 has achieved desirable results in our research group and is meant as a suggestion to readers. A lower MOI may be required for more infectious C. neoformans strains or for less resilient macrophage cell lines. Verification of infection (section 3.5) can be used to determine the ideal MOI for particular C. neoformans – macrophage combinations.
4. Sample collection
5. Cellular proteome
NOTE: Sufficient lysis must be optimized for the cell type analyzed (i.e., the quantity of cycles and amplitudes depends on cell pellet size and the power percentage of probe sonicator model).
6. Mass spectrometry
7. Data analysis
NOTE: MS data can be processed with numerous bioinformatics pipelines. In this protocol, we describe processing using the publicly available MaxQuant and Perseus platforms but recommend individual users to evaluate bioinformatic tools appropriate for the analysis, preference, and usage.
The protocol outlined above enables identification and quantification of proteins derived from both the fungal pathogen, C. neoformans, and the host, macrophage cells, in a single experiment. Following co-culture, cells are collected and processed together and bioinformatically separated based on peptide profiles specific to each species. This is a powerful approach for defining the interplay of the host-pathogen relationship during infection. The number of proteins identified from the experiment depends on the starting material, sample preparation, gradient length, MS instrumentation, and bioinformatic workflow. Using the protocol described herein, we typically, identify approx. 8,000 proteins from the experiment with 1,500 C. neoformans proteins and 6,500 host proteins. Following processing of the datasets, we generate a Principal Component Analysis (PCA) plot to observe critical factors driving our analysis (Figure 2A). Here, we observe the largest component of separation among the data is infected vs. non-infected samples, as we would anticipate from the experimental design (component 1, 79.8%), and a second distinguishing feature of the samples is biological variability (component 2, 5.7%). Next, a Pearson correlation combined with hierarchical clustering by Euclidean distance groups the samples and enables quantification of the variability among the replicates (Figure 2B). In our analysis, we observed distinct clustering of infected vs. non-infected samples and replicate reproducibility ranging from 95-96%, representing good reproducibility among the replicates. Lastly, we perform a Student’s t-test corrected for multiple hypothesis testing using a Benjamini-Hochberg false discovery rate (FDR) (p-value ≤ 0.05; FDR = 0.01; s0 = 1) to identify proteins with significant differences in abundance during infection compared to non-infected controls (Figure 2C). Here, we identify 760 proteins with significant changes in abundance, including 117 host proteins with 86 showing a significant decrease and 31 showing a significant increase upon infection. Notably, we also observe significant increases in abundance of fungal proteins, as expected during infection. With these data, subsequent analyses, including network mapping, in silico characterization, and follow-up experiments are performed to validate the data and explore the molecular mechanisms underpinning the host response to virulence.
Figure 1: Mass spectrometry-based proteomics workflow for analysis of macrophages infected with C. neoformans. The workflow begins with collection of macrophages either infected with C. neoformans or non-infected controls. Proteins are extracted by mechanical and chemical disruption, followed by reduction and alkylation, acetone precipitation, and enzymatic digestion. Peptides are purified on C18 STAGE tips, separated by high-performance liquid chromatography, subjected to electrospray ionization, and measured on a high-resolution mass spectrometer. Data is processed, analyzed, and visualized in the publicly available bioinformatics platforms, MaxQuant (with Andromeda) and Perseus14,15,16. Experiments performed in biological quadruplicate. Please click here to view a larger version of this figure.
Figure 2: Representative data for C. neoformans infection of macrophage cells. (A) Principal component analysis demonstrates distinction between infected vs. non-infected macrophage (component 1, 79.8%), and clustering of biological replicates (component 2, 5.7%). (B) Heat map of Pearson correlation plotted by hierarchical clustering by Euclidean distance to show clustering of samples (infected vs. non-infected) and replicate reproducibility (>95%). (C) Volcano plot of identified proteins. Blue = fungal proteins with significant change in abundance; black = macrophage proteins with significant change in abundance. Student’s t-test (p-value ≤ 0.05), FDR = 0.01; s0 = 1. Please click here to view a larger version of this figure.
Critical steps in the protocol include preparation of macrophage cells and collection of co-culture samples for protein processing with minimal disruption to the cells. It is important to perform steps of washing, inoculating, and removing adherent macrophage cells gently and carefully to prevent unnecessary lysis of cells prior to collection. Establishing the correct MOI for the experiment is also critical as inoculating with an excessively high MOI can cause rapid macrophage cell death and difficulty in collecting and processing samples for MS. Conversely, low MOI numbers will lead to fewer phagocytosed fungal cells and limited detection of fungal proteins in the biological system. To overcome such limitations, we recommend performing test experiments with varying MOIs, supported by cell death assays (e.g., LDH quantification) to define the number of fungal cells that initiate a host response but do not kill the host cells prior to collection. For the experiments, we aim to identify infection-associated fungal proteins, requiring a high MOI (100:1) to adequately detect fungal proteins among highly abundant host proteins. We routinely perform macrophage cytotoxicity assays to assess MOI impact on host cell death prior to performing the entire experiment. Timing of incubation of C. neoformans cells with macrophages is also crucial as the fungal cells may possess large polysaccharides capsules and therefore, macrophage require more time for engulfment. We select to use a co-culture incubation time of 3 h for the outlined experiment, as we found good coverage of the fungal proteome at this time point and to provide a ‘snap-shot’ of host response; however, researchers may wish to explore earlier and later time points and observe how timing impacts fungal and host protein production. Alternatively, priming the macrophage for opsonization has been performed to assist the phagocytic process17,18.
For sample collection, if samples are not being processed immediately flash freezing in liquid nitrogen will help prevent unwanted degradation of proteins by proteases present in the samples. In addition, for MS-based proteomic analyses, the potential for contamination from dust or keratin (e.g., skin and hair cells) should be limited through the use of nitrile gloves, laboratory coats, and washing of all surfaces with 70% ethanol prior to commencing experiments. Moreover, the protocols outlined above are specific to the co-culture sample set described but can be modified for protein extraction workflow optimization, as needed19,20. Opportunities for modifying the workflow and optimizing for specific cell types include the selected mechanical and chemical disruption techniques, duration and temperature of enzymatic digestion, and separation of the samples. For instance, lysis of C. neoformans cells is typically performed by mechanical bead beating; however, we have observed increased proteome coverage following probe sonication and therefore, recommend it for mechanical disruption of the infected cells19,21,22. For example, fractionating peptide samples into aliquots by high-pH fractionation or size exclusion chromatography may reduce sample complexity and improve depth of coverage on the mass spectrometer9. Moreover, to achieve the depth of coverage to identify approx. 8,000 host and fungal proteins, a high-resolution mass spectrometry system is required (e.g., QExactive Exploris, Fusion Lumos, timsTOF Pro).
The use of LFQ for quantifying changes in protein levels during infection is a reliable and cost-effective approach for MS-based proteomics12. It enables quantification of proteins by relative abundance without the need for additional sample processing steps. In addition, the analysis is performed following completion of the experiment, lending itself to universal applications and a flexible study design. However, limitations of LFQ include increased instrumentation time as samples must be run sequentially and cannot be combined, comparability between samples can be challenging, and the need for imputation to replace missing values can be high23. Alternative approaches for quantifying protein abundance include metabolic (e.g., stable isotope labeling of amino acids in cell culture) and chemical (e.g., tandem mass tags) labeling techniques, which permit combining samples to reduce instrument time, provide reliable comparability among samples, and commonly lead to less missing values24,25. However, such approaches require additional sample handling and processing steps, increased wet lab experiment complexity and time, as well requiring a fixed experimental design. To choose a quantification method optimal for proposed experiments, users should consider study design and comparability of samples, wet lab complexity and data analysis, as well as instrumentation time availability and cost.
The novelty of the protocol presented is the ability to define changes in the proteome from both the host and pathogen perspectives in a single experiment. The depth of coverage of both proteomes enables new insight into how the pathogen initiates infection and how the host responds in defense. Notably, the approach focuses on infection of the entire cell; however, opportunities to define sub-proteomes or compartmentalized responses to infection exist through combination with spatial localization techniques (e.g., centrifugation, enrichment, labeling)26,27. Here, we detail the interaction between macrophages and the fungal pathogen, C. neoformans; however, the approach is universal, and can be applied to interactions between a diverse array of biological systems. For example, we recently used a similar workflow to uncover general and site-specific responses of neutrophils derived from a murine model of ocular keratitis28,29,30. Moreover, the infectome datasets generated from this Protocol can be integrated with in vitro proteome and secretome profiling of a pathogen to detect proteins with altered abundance in the presence of host cells. Such proteins, referred to as infection-associated proteins, provide a plethora of known and novel virulence factors for further characterization, including temporal regulation, localization, and direct protein-protein interactions with the host. Overall, the outlined MS-based proteomics workflow provides a new opportunity to investigate the intricate relationship between host and pathogen in a single experiment with universality and comprehension not commonly available.
The authors have nothing to disclose.
The authors thank Dr. Jonathan Krieger of Bioinformatics Solutions Inc. for operating the mass spectrometer for representative experiments, as well as members of the Geddes-McAlister group for their assistance with experimental set-up and manuscript feedback. The authors acknowledge funding support, in part, from the Banting Research Foundation – The Jarislowsky Fellowship Discovery Award, New Frontiers Research Fund – Exploration (NFRFE-2019-00425), and the Canadian Foundation for Innovation (JELF 38798) for J.G.M., as well as NSERC Canada Graduate Scholarship – Masters and Ontario Graduate Scholarship for B.B., and Queen Elizabeth II Graduate Scholarship in Science and Technology for A.S..
100 mM Tris-HCl, pH 8.5 | Fisher Scientific | BP152-1 | Maintain at 4°C |
60 x 15 mm Dish, Nunclon Delta | ThermoFisher Scientific | 174888 | |
6-well cell culture plate | ThermoFisher Scientific | 140675 | |
Acetonitrile, MS grade | Pierce | TS-51101 | |
Acetic Acid | Sigma Aldrich | 1099510001 | |
Acetone | Sigma Aldrich | 34850-1L | |
Ammonium bicarbonate (ABC) | ThermoFisher Scientific | A643-500 | Prepare a stock 50 mM ABC solution, stable at room temperature for up to one month. |
Bel-Art™ HiFlow Vacuum Aspirator Collection System | Fisher Scientific | 13-717-300 | Not essential, serological pipettes can be used to remove media. |
C18 resin | 3M Empore | 3M2215 | |
Cell Scrapers | VWR | 10062-906 | Not essential, other methods to release macrophage cells can be used. |
Centrifugal vaccuum concentrator | Eppendorf | 07-748-15 | |
Complete Filtration Unit | VWR | 10040-436 | |
Conical falcon tubes (15 mL) | Fisher Scientific | 05-539-12 | |
Countess II Automated Cell Counter | ThermoFisher Scientific | AMQAX1000 | Not essential, haemocytometer can be used as an alternative. |
CytoTox 96 Non-Radioactive Cytotoxicity Assay | Promega | G1780 | |
Dithiothreitol (DTT) | ThermoFisher Scientific | R0861 | Prepare bulk stock solution of 1 M DTT, flash frozen and stored at -20 °C until use. Discard after each use (do not freeze-thaw repeatedly). |
DMEM, high glucose, GlutaMAX Supplement | ThermoFisher Scientific | 10566016 | |
Fetal Bovine Serum (FBS) | ThermoFisher Scientific | 12483020 | Heat inactivate by incubating at 60°C for 30 minutes. Prepare 50 ml aliquots and flash freeze. Thaw prior to media preparation |
Haemocytometer | VWR | 15170-208 | |
HEPES | Sigma Aldrich | H3375 | Prepare 40 mM HEPES/8 M Urea in bulk stock solution, flash frozen, store at -20°C until use. Discard after each use (do not freeze-thaw repeatedly). |
High-performance liquid chromatography system | ThermoFisher Scientific | LC140 | Gradient length is based on sample complexity, recommended 120 min gradient for infectome samples. |
High-resolution mass spectrometer | ThermoFisher Scientific | 726042 | |
Iodoacetamide (IAA) | Sigma Aldrich | I6125 | Prepare 0.55 M bulk stock solution, flash frozen, store at -20°C until use. Discard after each use (do not freeze-thaw repeatedly). |
L-glutamine | ThermoFisher Scientific | 25030081 | Can be aliquot and frozen for storage. Thaw prior to media preparation. |
LoBind Microcentrifuge tubes | Eppendorf | 13-698-794 | |
MaxQuant | https://maxquant.org/ | MaxQuant is a public platform that offers tutorials, such as the MaxQuant Summer School, outlining the computational analysis steps of large MS data sets | |
Microcentrifuge | Eppendorf | 13864457 | |
Penicillin : Streptomycin 10k/10k | VWR | CA12001-692 | Can be aliquot and frozen for storage. Thaw prior to media preparation. |
Peptide separation columns | ThermoFisher Scientific | ES803 | |
Perseus Software | http://maxquant.net/perseus/ | ||
Phosphate Buffered Saline | VWR | CA12001-676 | Puchase not required. PBS can also be prepared but sterile filteration must be performed before use. |
Pierce BCA Protein Assay | ThermoFisher Scientific | 23225 | |
Pipette, Disposable Serological (10 mL) | Fisher Scientific | 13-678-11E | |
Pipette, Disposable Serological (25 mL) Basix | Fisher Scientific | 14955235 | |
Probe sonciator | ThermoFisher Scientific | 100-132-894 | |
Protease inhibitor cocktail tablet | Roche | 4693159001 | |
Sodium dodecyl sulfate | ThermoFisher Scientific | 28364 | 20% (w/v) |
Spectrophotometer (Nanodrop) | ThermoFisher Scientific | ND-2000 | |
STAGE tipping centrifuge | Sonation | STC-V2 | |
Thermal Shaker | VWR | NO89232-908 | |
Trifluoroacetic acid | ThermoFisher Scientific | 85183 | |
Trypsin/Lys-C protease mix, MS grade | Pierce | A40007 | Maintain at -20 °C. |
Ultrasonic bath | Bransonic | A89375-450 | Stored in cold room (4C) |
Urea | Sigma Aldrich | U1250-1KG | Prepare 40 mM HEPES/8 M Urea in bulk stock solution, flash frozen, store at -20 °C until use. Discard after each use (do not freeze-thaw repeatedly). |
Yeast-extract peptone dextrose broth | BD Difco | BM20 |