- Email: [email protected]
A proteomics workflow for quantitative and time-resolved analysis of adaptation reactions of internalized bacteria
Methods 61 (2013) 244–250
Contents lists available at SciVerse ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
A proteomics workflow for quantitative and time-resolved analysis of adaptation reactions of internalized bacteria Henrike Pförtner a, Juliane Wagner a, Kristin Surmann a, Petra Hildebrandt a,d, Sandra Ernst a, Jörg Bernhardt b, Claudia Schurmann a, Melanie Gutjahr a, Maren Depke a,d, Ulrike Jehmlich c, Vishnu Dhople a, Elke Hammer a, Leif Steil a, Uwe Völker a, Frank Schmidt a,d,⇑ a
Interfaculty Institute of Genetics and Functional Genomics, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15a, D-17487 Greifswald, Germany Institute for Microbiology, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15, D-17487 Greifswald, Germany ZIK-HIKE Junior Research Group, Ernst-Moritz-Arndt-University Greifswald, Fleischmannstraße 42-44, D-17489 Greifswald, Germany d ZIK-FunGene Junior Research Group "Applied Proteomics", Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15a, D-17487 Greifswald, Germany b c
a r t i c l e
i n f o
Article history: Available online 30 April 2013 Keywords: Host–pathogen interaction S. aureus In vivo proteomics Pulse-chase labeling Cell sorting
a b s t r a c t The development of a mass spectrometric workflow for the sensitive identification and quantitation of the kinetics of changes in metaproteomes, or in particular bacterial pathogens after internalization by host cells, is described. This procedure employs three essential stages: (i) SILAC pulse-chase labeling and infection assay; (ii) isolation of bacteria by GFP-assisted cell sorting; (iii) mass spectrometry-based proteome analysis. This approach displays greater sensitivity than techniques relying on conventional cell sorting and protein separation, due to an efficient combination of a filtration-based purification and an on-membrane digestion. We exemplary describe the use of the workflow for the identification and quantitation of the proteome of 106 cells of Staphylococcus aureus after internalization by S9 human bronchial epithelial cells. With minor modifications, the workflow described can be applied for the characterization of other host–pathogen pairs, permitting identification and quantitation of hundreds of bacterial proteins over a time range of several hours post infection. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Infectious diseases are still one of the major challenges for the health care systems. In 2002 about 26% of all cases of death worldwide, totaling nearly 15 million people, were due to infectious diseases [1]. The development of new biomarker targets and therapies requires a thorough understanding of the pathophysiology of infectious diseases, particularly the intimate interplay between the pathogen and the host during the progression of disease. The introduction and development of the collection of functional genomics technologies (genomics, transcriptomics, proteomics, and metabolomics) commonly known as OMICS-technologies have paved ⇑ Corresponding author. Address: ZIK-FunGene Junior Research Group ‘‘Applied Proteomics’’, Ernst-Moritz-Arndt-University Greifswald, Germany. Fax: +49 3834 86 795871. E-mail addresses: [email protected] (H. Pförtner), [email protected] (J. Wagner), [email protected] (K. Surmann), [email protected] (P. Hildebrandt), [email protected] (S. Ernst), [email protected] (J. Bernhardt), [email protected] (C. Schurmann), Melanie.Gutjahr@fli.bund.de (M. Gutjahr), [email protected] (M. Depke), [email protected] (U. Jehmlich), [email protected] (V. Dhople), [email protected] (E. Hammer), [email protected] (L. Steil), [email protected] (U. Völker), [email protected] (F. Schmidt). 1046-2023/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymeth.2013.04.009
the way for a new level of understanding, because they can provide a holistic view of the adaptive changes that take place during infection in both host and pathogen. While genomics and transcriptomics are routinely used to address infection-related questions at a system-wide level, limitations in sensitivity and specificity in the field of proteomics until very recently restricted the generation of complementing data sets at the protein level. Integration of these different OMICS-datasets will provide an integrated view from adaptation at the gene expression level through physiological and cellular adaptation. With the development of the newest generation of high-accuracy mass spectrometers (MS) such as LTQ-Orbitrap XL MS, the perspective changed because gel-free protein identification and quantitation of complex samples in minute amounts down to the low picogram range became possible [2]. Compared to classical setups dealing with pure cell cultures or isolated tissue, the characterization of pathogenic bacteria from host–pathogen settings are even more difficult, because only extremely small amounts of sample are accessible for the infecting pathogen. While one challenge is the sensitivity of current proteomics techniques, an even greater problem is the selection of adequate pre-separation techniques for the low numbers of bacteria available in such experimental setups. Different approaches have
H. Pförtner et al. / Methods 61 (2013) 244–250
been developed for isolation of bacteria from host cells such as sucrose gradient centrifugation [3], immunomagnetic separation [4], and FACS sorting via fluorescent proteins [5]. A common feature of these techniques is the requirement of large quantities of cells (>108 cells) for sample preparation and protein identification, making them unsuitable for experiments with limited biomass availability, such as characterization of internalized bacterial pathogens. Recently, we introduced a workflow that allows time-resolved quantitative monitoring of the adaptive response of only 106 cells of S. aureus HG001 [6] carrying plasmid pMV158GFP [7] after internalization by 4 105 S9 human bronchial epithelial cells, a nonprofessional phagocytic cell line. Recent studies have already combined flow cytometry and 2-DE to investigate 108–109 internalized pathogens requiring a sorting time of several days [8]. In our approach, we gained an increase in sensitivity by optimizing an established workflow (Fig. 1) that combined pulse-chase labeling, FACS sorting via GFP, on-membrane washing [9] and digestion [10], high sensitivity nLC-MS/MS, and comprehensive data analysis (Fig. 2). The prior labeling of bacteria with SILAC allowed quantification even of low intense signals in a correct manner and discrimination of such low-abundance bacterial peptides (heavy) from
245
host peptides (light), which is still crucial in high complex proteomics approaches. Using this approach, one can now investigate intracellular bacteria with initial starting material two to three orders of magnitude lower than what was usually required for a quantitative proteomics approach [11]. This workflow is not only suitable for the characterization of individual species but a modified version has also been applied to the study of a mixture of bacteria [10] and we can envision its application for the analysis of complex consortia in metaproteomics approaches. 2. Materials and methods 2.1. Reagents, chemicals and media Acetone (ice-cold, stored at 20 °C; Merck, cat No. 1.00014. 1011). Acetic acid (AA; Merck, cat. No. 1.00063.2511). Acetonitrile (ACN; Roth, cat. No. T195.2). Alexa FluorÒ 568 phalloidin (Invitrogen, cat. No. SKU#A12380). Amino acid mixture group A (Promocell; customer formulation), alanine, valine, leucine, isoleucine; 50 mM each, cat. No. C-97027).
Fig. 1. General workflow to identify and quantify proteins from internalized S. aureus HG001 cells in a time-resolved manner. (a) S9 human bronchial epithelial cells were grown to confluence in eMEM containing light arginine and lysine. In parallel S. aureus was cultivated to exponential growth phase in pMEM, containing the 13C6 heavy isotopes of arginine and lysine. (b) The S. aureus culture was diluted with eMEM to a MOI of 25 and then transferred to the host cells. (c) After 1 h of internalization remaining non-internalized bacteria were killed by lysostaphin treatment. After internalization, bacteria can only incorporate light counterparts of the heavy amino acids into their proteins. (d) Samples were taken hourly, host cells were disrupted by Triton X-100 and internalized S. aureus were released. (e) At each point in time, GFP-positive bacteria were separated from cell debris via FACS and sorted on a low protein binding filter device. (f) and subjected to proteolytic digest first by lysostaphin. (g) followed by trypsin. (h) Tryptic peptides were purified by ZipTip and (i) measured via nLC–LTQ-Orbitrap XL MS. (j) Identification and determination of peptide ratios were performed with the Elucidator software package. (k) Protein light vs. heavy ratios with a more than twofold change relative to the median over the time course were defined as up- or downregulated. Proteins with a smaller deviation of their ratio from the average ratio of all proteins were defined as non-regulated. (l) Regulated proteins were interpreted in the context of their role in cellular physiology.
246
H. Pförtner et al. / Methods 61 (2013) 244–250
Amino acid mixture group B (Promocell; customer formulation), aspartate, glutamate; 50 mM each, cat. No. C97028). Amino acid mixture group C (Promocell; customer formulation), serine, threonine, cysteine; 50 mM each, cat. No. C-97029). Amino acid mixture group D (Promocell; customer formulation), proline, phenylalanine, histidine; 50 mM each, cat. No. C-97030). Ammonium bicarbonate (Sigma, cat. No. A6141). 13C6 L-Arginine–HCl (EURISO-Top GmbH, cat. No. CLM2265-0.25). Ethanol, analytic grade (Merck, cat. No. 1.00983.1011). FACSFlow (BD, cat. No. 342003). Fetal Bovine Serum (FBS; Biochrom AG, cat. No. S0115). L-glutamine, 200 mM (PAA, cat. No. M11-004). Glycerol (Sigma–Aldrich, cat. No. G6279). HEPES buffer solution, 1 M (PAA, cat. No. S11-001). High precision cover slips, 18 mm diameter (Marienfeld, cat. No. 011758018). Hoechst 33258 (Sigma–Aldrich, cat. No. 861405). 13C6 L-Lysine–2HCl (EURISO-Top GmbH, cat. No. CLM2247-0.25). Lysostaphin (AmbicinR L; AMBI Products, cat. No. LSPN-50). MEM Non Essential Amino Acids, 100 (NEAA; PAA, cat. No. M11-003). Minimal essential media (MEM Earl, w/o L-glutamine; Promocell, cat. No. C-75210). Minimal essential media 2 (MEM Earl; Promocell; customer formulation w/o stable glutamine, w/o L-arginine, w/o L-lysine, w/o sodium bicarbonate, with 6.8 g/L sodium chloride), cat. No. C-96187). p-Phenylendiamin (Roth, cat. No. 4499.1). Phosphate buffered saline Mg, Ca (PBS Mg, Ca; Dulbecco’s PBS 10 without Ca & Mg; PAA, cat. No. H15-011). Phosphate buffered saline+Mg, Ca (PBS+Mg, Ca; Dulbecco’s PBS 1 with Ca & Mg; PAA, cat. No. H15-001). Select Agar (Invitrogen, cat. No. 30391-023). Sodium Bicarbonate (PAN-Biotech GmbH, cat. No. P0444100). Sodium hydroxide 2 mol/L (Roth, cat. No. T135.1). Trifluoroacetic acid (TFA; Applied Biosystems, cat. No. 400028). Tryptophan (Promocell; customer formulation); 25 mM, cat. No. C-97031). Tryptic Soy Broth (TSB; BD Bacto™, cat. No. 211825). Triton X-100 (ESA Laboratories Inc., cat. No. 80-0072). Trypsin (Promega, cat. No. V5111). Trypsin EDTA, 1 (PAA, cat. No. L11-004). Water HPLC-grade (JT Baker, cat. No. 4218).
Lysostaphin: 50 mg vial was dissolved in 10 ml sterile HPLCgrade water, aliquoted and stored at 20 °C until use. Mowiol mounting medium: 6 g glycerol and 2.4 g Mowiol were dissolved in 6 ml of distilled water by shaking at RT for at least 2 h. Afterwards, 12 ml of 0.2 M TRIS/HCl (pH 8.5) was added and shaken at 53 °C until everything was dissolved. 0.1% p-phenylendiamine was added as antifading reagent and shaken again for 3 h. Aliquots of the solution were stored at 20 °C. 1 PBS Mg, Ca for bacterial dilution series: 10 PBS Mg, Ca were diluted 1:10 in A. dest. autoclaved and stored at room temperature until use. Phalloidin staining solution: 5 ll Phalloidin-Alexa Fluor 586 was diluted in 1 ml PBS Mg, Ca (final concentration 1 U/ml). The dilution was prepared freshly for each staining. 0.1% Triton X-100: The 1% stock solution was diluted 1:10 in sterile HPLC-grade water. The dilution was prepared one day before the experiment and sterile filtrated to avoid contaminations. Trypsin: A vial containing 20 lg was filled with 200 ll ammonium bicarbonate (20 mM) to dissolve trypsin. Aliquots of this solution were stored at 20 °C until use. TSB agar plates: 30 g TSB powder and 15 g select agar were dissolved in 1 l A. dest., autoclaved and used to pour plates. Solutions for ZipTip peptide purification (10 samples): each solution was prepared freshly before use (see Table 1). 2.3. Methods 2.3.1. Preparation of eMEM and SILAC pMEM media For internalization assays eMEM and for labeling of bacteria SILAC pMEM were prepared. For eMEM 4 ml fetal bovine serum (FBS), 2 ml glutamine (200 mM) and 1 ml NEAA (100) were added to 100 ml MEM (1). The medium was mixed thoroughly and stored at 4 °C. Before application to the cells the medium was preheated to 37 °C. For SILAC pMEM, compounds listed in Table 2 were used. The medium was further adjusted to a pH of 7.4 with 1 M NaOH in HPLC-grade water and sterilized by filtration using a pore size of 0.22 lm. 2.3.2. Bacterial growth conditions A glycerol stock of S. aureus HG001 [6] pMV158GFP [7] was serially diluted (10 3–5 10 6) in SILAC pMEM with tetracycline Table 1 Solutions for ZipTip peptide purification (elution).
100% ACN 5% AA HPLC-grade water Total Volume
80% ACN in 1% AA
50% ACN in 1% AA
484.0 ll 121.0 ll – 605.0 ll
302.5 ll 121.0 ll 181.5 ll 605.0 ll
Table 2 SILAC pMEM composition for 80 ml.
2.2. Reagent setup 20 mM ammonium bicarbonate (pH 7.8): 20 mg ammonium bicarbonate was dissolved in 12.5 ml HPLC-grade water. For each experiment fresh ammonium bicarbonate buffer was prepared. LC Buffer A: 2% acetonitrile and 0.1% acetic acid in HPLC-grade water. The buffer was stored at room temperature. LC Buffer B: 100% acetonitrile with 0.1% acetic acid. This buffer was also stored at room temperature. Hoechst staining solution: 20 lg/ml was prepared as stock solution and aliquots were stored at 20 °C. The Hoechst 33258 stock solution was diluted 1:20 in PBS Mg, Ca (final concentration 1 lg/ ml) for staining of DNA. Each dilution was freshly prepared for staining. 5 mg/ml.
Stock
Volume
Final concentration
2 MEM 100 NEAA 200 mM L-glutamine 1 M HEPES AA group A (Ala, Val, Leu, Ile) 50 mM each AA AA group B (Asp, Glu) 50 mM each AA AA group C (Ser, Thr, Cys) 50 mM each AA AA group D (Pro, Phe, His) 50 mM each AA Trp 25 mM 13 C6 Arg–HCl (125 mg/ml = 576 mM) 13 C6 Lys–2HCl (125 mg/ml = 556 mM) HPLC-grade water
40.0 ml 0.8 ml 1.6 ml 0.8 ml 3.2 ml
1 MEM 1 NEAA 4 mM L-glutamine 10 mM HEPES 2 mM each AA
3.2 ml 3.2 ml 3.2 ml 6.4 ml 82.9 ll 57.1 ll 17.46 ml
2 mM each AA 2 mM each AA 2 mM each AA 2 mM Trp 0.597 mM Arg 0.397 mM Lys
H. Pförtner et al. / Methods 61 (2013) 244–250
(20 lg/ml) and grown at 37 °C and 220 rpm for about 12 h. The 20 ml SILAC pMEM main culture was inoculated with an exponentially growing over night culture to an OD600 nm of 0.05. The culture was grown at 37 °C and 150 rpm in a water bath to OD600 nm 0.4 (exponential growth phase) and used for infection of eukaryotic cells. 2.3.3. Host cell line and cultivation conditions The S9 epithelial cell line [12] was cultured in eMEM supplemented with 4% of fetal bovine serum (FBS) at 37 °C in 5% CO2. For the S. aureus infection assay, epithelial cells were seeded 3 days in advance in 24-well tissue culture plates and cultured to confluence (4 105 cells/well). 2.3.4. Infection assays S9 cells were infected with S. aureus at a multiplicity of infection (MOI) of 25. For this purpose, S9 cells of one representative well were detached with 200 ll trypsin–EDTA and the trypsin reaction was stopped by adding 300 ll of eMEM. After homogenization, the cell solution was mixed with Trypan blue dye and the number of viable eukaryotic cells was counted with a Countess (Invitrogen, Karlsruhe, Germany). Afterwards, cells were incubated with an infection master mix, composed of S. aureus in eMEM (MOI of 25) and 29 ll NaHCO3 (7.5%) per ml of bacterial culture added to the infection master mix. Infection master mix and the S9 cells were co-incubated at 37 °C and 5% CO2 for 1 h allowing internalization of Staphylococci. Meanwhile 100 ll of the infection master mix were diluted in PBS Mg, Ca and plated on TSB agar plates to determine bacterial titer of the master mix for later calculation of the precise MOI reached. After 1 h of co-incubation the infection master mix was removed and pre-warmed eMEM with lysostaphin was added (final concentration 10 lg/ml). Incubation of cell culture plates was then continued at 37 °C and 5% CO2. For FACS sorting at desired points in time (1.5–6.5 h post infection), cell culture medium was removed and the S9 cells were washed twice with PBS+Mg, Ca. To lyze the S9 cells, 150 ll of Triton X-100 (0.1%) per well was added and the plate was further incubated for about 7 min at 37 °C. Host cells lysates with internalized bacteria were homogenized by pipetting and pooled in a 15 ml tube. This procedure did not disrupt bacteria. Cell culture plate wells were rinsed with 50 ll FACSFlow per well and this solution was added to the 15 ml tube. Samples were transferred immediately on ice to a FACSAria (Becton Dickinson, CA, USA) for sorting. In parallel, 100 ll of the homogenized sample were serially diluted in PBS Mg, Ca and plated to determine the titer of internalized bacteria. 2.3.5. Isolation of bacteria from host cells by FACS on a 96-well plate filter membrane The GFP labeled bacteria were sorted applying an appropriate fluorescence channel and SSC-A gates onto a low protein binding filter membrane (0.22 lm pore size) of a 96-well microtiter plate (Millipore, Schwalbach, Germany) by applying vacuum (450– 550 mbar) to the microtiter plate to allow constant removal of the fluid. The filter was rinsed with 200 ll FACSFlow and the membrane was cut out and divided into small pieces using a scalpel. The filter pieces were placed into a reaction vial for subsequent onmembrane digestion. 2.3.6. On-membrane protein digestion for Shotgun Mass Mapping (SMM) 12.33 ll of 20 mM ammonium bicarbonate (ABC) buffer was added to the filter membrane containing sorted bacteria. Afterwards, 1 ll lysostaphin (0.05 lg/ll) was added to the filter, mixed thoroughly and incubated for 30 min in a water bath at 37 °C. Then, 6.67 ll trypsin (0.1 lg/ll) was added, mixed thoroughly and incubated several hours or overnight (16 h) in water bath at 37 °C. The
247
enzymatic reaction was stopped by adding TFA to a final concentration of 0.1% and incubating the samples for 5 min at 37 °C. After centrifugation for 10 min at 13,000g at room temperature the supernatant was transferred into a new microfuge tube (0.5 ml). Peptides were purified with minor modification by ZipTip as recommended in the manufacturer’s instructions as follow: peptides were bound from supernatant to ZipTip, washed, and eluted by rinsing 10 times with 5 ll 50% ACN-buffer and 5 ll 80% ACN-buffer in 1% acetic acid and dried in a Speed-vacuum centrifuge (Eppendorf, Hamburg, Germany). 2.3.7. nLC–LTQ-Orbitrap XL MS measurements of peptides Dried peptides were dissolved using 10 ll LC buffer A (2% ACN, 0.1% AA) and transferred to LC vials. The mass spectra were acquired online after peptide separation by one-dimensional nLC. For the example presented here the following conditions were used: peptides were enriched on a BioSphere C18 pre-column and separated using an Acclaim PepMap 100 C18 column on a Proxeon Easy nLC (Proxeon, Odense, Denmark). Separation was achieved with the formation of non linear gradient of 290 min containing LC Buffer A and LC Buffer B 0% for 15 min, 0–30% for 260 min, 30– 60% for 15 min, 60–100% for 1 min and equilibrated for 5 min. Peptides were eluted with a flow rate of 300 nl/min and analyzed using a LTQ-Orbitrap XL MS (Thermo Scientific, Bremen, Germany) with an interfaced Nano ESI source. The full scan was carried out using a FTMS analyzer with normal mass range of m/z 300–2000. Data were acquired in profile mode with a resolution of 60,000 and positive polarity. This was followed by Data Dependent Acquisition of the top five most intense precursor ions for MS/MS scans using CID activation. A minimum of 1000 counts were activated for 30 ms with an activation of q = 0.25, isolation width 2.00 Da and a normalized collision energy of 35%. The charge state screening and monoisotopic precursor selection was enabled with the rejection of +1, +4, and unassigned charge states. Peptides were excluded for the next 30 s once MS/MS scans had been acquired. 2.3.8. Peptide identification and quantitation with the Trans Proteomics Pipeline (TPP) [13] Tandem MS spectra were used to identify peptides with the classical standard search engine Sequest™ and a S. aureus FASTA database embedded in Sorcerer™. A precursor tolerance of 20 ppm, one missed cleavage and [13C6]-arginine and [13C6]-lysine as possible modifications were applied. All first-ranked hits were further consolidated by the standard algorithm Peptide Prophet [14]. The linked survey-scan number and the assigned peptide identification were used to calculate the ion and time-dependent intensities of the light and corresponding heavy peptides, which were further used for quantitation by the XPRESS tool from the TPP. 2.3.9. Data analysis with Rosetta Elucidator In order to perform a spectral alignment, which cannot be performed by the TTP software so far, the final analysis was done with the Rosetta Elucidator software (Rosetta Biosoftware by Ceiba Solution Inc., MA, USA). First, raw data files from nLC–LTQ-Orbitrap XL MS were loaded. An experimental definition for quantitative protein analysis was created where an instrument mass accuracy of 10 ppm and modified amino acids arginine and lysines with mass shift of 6.02 Da were applied. Further, a retention time minimum cutoff of 40.0 (min) and a maximum cutoff of 270.0 (min) as well as a m/z minimum cutoff of 300.0 and maximum of 1200.0 were set. A peak time width minimum (min) of 0.2 was applied and spectral alignment (alignment search distance 10.0 min) was used for binning. For quantitation, a maximum of two labelings per peptide were allowed and detection parameters were set as follow: PPM error 10.0, left RT location tolerance (minutes) 0.8,
248
H. Pförtner et al. / Methods 61 (2013) 244–250
right RT location tolerance (minutes) 0.8, error model to apply: pair ratio error model, limit fold change to (±) 1000, and normalization was skipped. After identification (similar parameter as shown in 2.3.8), an automatic Peptide/Protein Tellers annotation was performed and only Peptide Teller results greater than 0.8 were used and at least two peptides were necessary for further consideration.
2.3.10. Quantitative data analysis using alpha values and ratios For quantitative analysis, a tab-delimited matrix was created from the Rosetta Elucidator Excel report, containing protein names, intensities, not normalized ratios (rnormalized = rnot normalized/
rmedian_all) and alpha values [15] (see Fig. 2i) from intensities of each point in time per independent biological sample series. Afterwards, a weighted or non-weighted linear regression for the six points in time was performed, and ratios from the values at t6.5 h and t1.5 h using linear regression equations (note: these ratios are more robust than using original values) were calculated. Furthermore, fold change cutoff and quantile-information were assigned using an R-script. The data were further classified by kmeans clustering (Fig. 2j) using the Analyst Software from GeneData (Genedata AG, Basel, Switzerland) and following parameter settings: Distance: Positive Correlation Distance (1 r), Centroid Calculation: Mean, Valid Values: 50, Max Iterations: 50.
Fig. 2. Data analysis workflow for proteome data. (a) After spectral alignment and isotope grouping, the corresponding isotopic peptide pairs were defined based on a mass shift of 6.02 Da. Image a shows a peptide pair of the protein PrsA, which was found to be upregulated during the time-course. (b) From each peptide pair, the areas of the selected ion chromatograms were calculated and further used for ratio calculation. (c) All available peptide ratios of a protein were squeezed to one protein ratio and normalized by the median. Furthermore, the corresponding ratios of three independent experimental series (bioreplicates) were compared. (d) In order to make the analysis more robust, a linear regression of all proteins was performed and their slopes were calculated. (e) The binning of ratios from time point 6.5 h divided by 1.5 h of the proteins of the three bioreplicates were plotted as histograms or (f) sigmoid curves. To assign significant changes a quantile calculation and fold change cutoffs were applied. (g) The normalized total intensity of the L9 ribosomal protein is given, where the general trend shows no significant regulation. (h) The normalized single intensities of the light and heavy peptides are also given, indicating a decrease in the amount of the heavy peptides over time. Together with the stable summed intensity of the light peptides this observation indicates enhanced degradation of L9. (i) The alpha value for normalization of protein levels was calculated by first calculating the sum of the peptide intensities for each individual protein and the sum of the intensity of all proteins observed and then dividing at each time point the intensities of individual proteins by the sum of all protein intensities at the same time point multiplied by factor of 100. For the calculation of the alpha values for single proteins the sum of intensities of the light and the heavy peptides of the protein were calculated. (j) Based on the alpha value data a Kmeans (n = 9) clustering (GeneData Analyst Software package) was applied. The nine clusters (C1–9) and the number of fitting proteins are shown. (k) Two representative clusters for groups of regulated proteins are shown. (l) In order to map the quantitative data onto pathways/regulatory circuits a Voronoi-like treemap was used. The figure shows an image section of the TCA, glycolysis and ribosomal proteins. The light red colors indicate a downregulation of many ribosomal proteins and some proteins belonging to cations and iron carrying compounds. The light blue color of proteins involved in TCA and glycolysis indicates an increase of the corresponding alpha values during internalization.
H. Pförtner et al. / Methods 61 (2013) 244–250
249
Fig. 3. (a) Enzymatic digestion of 102–105 GFP-labeled S. aureus bacteria. Filtered bacteria were incubated with lysostaphin first followed by trypsin for 1, 2, 3, 4, and 5 h and over night (grey triangles). The number of unique identified peptides identified ranged from 200 for 102 bacteria to 5000 for 105 bacteria, with enzymatic digestion times as short as one hour. Increasing incubation times did not significantly influence the number of peptides identified. (b) Fluorescence image of GFP expressing bacteria internalized by S9 human bronchial epithelial cells. The nuclei of the S9 cells are shown in blue and the f-actin in red. (c) Dot plot for sorting of GFP expressing bacteria. After internalization and lysis of the host cells, bacteria and attached human host proteins were sorted.
2.3.11. Data visualization and analysis using Voronoi-like treemaps First, a construction of a hierarchical structure of gene functional categories (e.g. Riley Scheme based TIGR S. aureus functional classification) was performed. Second, an iterative hierarchy level by level construction of a Voronoi-like treemap layout [16,17] was applied that contains all S. aureus proteins at the lowest level (mosaic tiles on Fig. 2l). Furthermore, mapping of protein quantitation data to an appropriate color scale and of colors to Voronoi cells representing the appropriate proteins was performed. 2.3.12. Visualization of internalized bacteria by confocal microscopy The internalization was performed as described in Section 2.3.4. At desired points in time cell culture medium was removed and cells were washed twice with PBS+Mg, Ca and fixed with 2% formaldehyde in PBS+Mg, Ca for at least 20 min at RT in the dark. After fixation formaldehyde was replaced with PBS+Mg, Ca. Plates can be stored at 4 °C in the dark until the staining of the eukaryotic cells is performed. At first, unspecific binding sites were blocked with 10% FBS in PBS+Mg, Ca and cover slips were incubated for 20 min at RT in the dark. The blocking solution was aspirated and 500 ll 0.1% Triton X-100 were added for 3–5 min. Wells were washed with PBS+Mg, Ca twice, 400 ll Phalloidin staining solution were added to stain the f-actin and wells were again incubated for 20 min at 37 °C and 5% CO2. Afterwards, cover slips were rinsed with PBS+Mg, Ca. In addition, 400 ll Hoechst staining solution were added to stain the DNA of the nuclei and incubation proceeded for 10 min at RT in the dark. Cover slips were then washed three times with distilled water and mounted upside down with MOWIOL containing antifading reagent on microscopic slides. Fluorescence pictures were obtained using a Leica TCS-SP5 DMI6000i confocal laser scanning microscope equipped with an argon laser (488 nm) for the excitation of GFP labeled bacteria, a UV laser (405 nm) for the excitation of Hoechst labeled nuclei and a helium–neon laser (594 nm) for the excitation of Phalloidin labeled f-actin skeleton of the S9 cells. Emission of each fluorescent dye was recorded according to manufacturer recommendations. The original images were processed with ZIK-ImageJ (free image analysis software) to optimize brightness, contrast and sharpness of the final pictures (Fig. 3b). 3. Results and discussion In order to identify and quantify proteins from internalized S. aureus cells, S9 human epithelial cells were cultivated in eMEM containing light arginine and lysine and in parallel S. aureus HG001 cells were cultivated in pMEM containing heavy arginine and lysine for at least five to six generations. The S9 cells were infected with heavily labeled S. aureus for 1 h and after internalization remaining non-internalized bacteria were killed by
lysostaphin treatment (Fig. 1a–c). The effective killing by lysostaphin allowed the exclusive analysis of internalized bacteria. Since concomitant with the infection the heavy amino acids lysine and arginine were replaced by their normal, light counterparts, internalized bacteria were forced to incorporate light lysine and arginine derivatives instead of heavy amino acids into their newly synthesized proteins after the start of the infection experiment. This pulse-chase design allowed the calculation of ratios between heavy and light peptides and thus quantification of protein synthesis after internalization. Information on intensity of heavy and light peptides could also be analyzed independently, thus even facilitating the distinction of altered synthesis and stability. To enrich the bacteria, S9 cells were carefully disrupted hourly by adding a nonionic surfactant until 6.5 h post infection and from each point in time GFP containing bacteria were selectively enriched by FACS (Fig. 1d and e). With this procedure, between 104–106 cells could be isolated per sample (determined by FACS sorting). To reduce the final sample volume, cells were collected on a low protein binding membrane where the proteolytic digestion (first by lysostaphin and followed by trypsin) took place (Fig. 1f and g). Lysostaphin is produced by S. simulans and contains an N-terminal glycyl-glycine-endopeptidase M23 and a C-terminal cell wall binding domain. The enzyme hydrolyzes the pentaglycine bridges in the peptidoglycan of S. aureus, even if the cells are not growing [18,19]. A homologue to lysostaphin, ALE-1, was also identified in S. capitis EPK1 [20] and both enzymes show a lytic activity specific to S. aureus. This enzymatic digestion of the S. aureus cell wall was used to improve the efficiency of cell disruption and to maximize access of trypsin to cytoplasmic proteins. For other bacteria lysozyme would be an alternative since this muramidase is a glycoside hydrolases that degrades bacterial cell walls by catalyzing hydrolysis of 1,4-beta-linkages between N-acetylmuramic acid and N-acetyl-D-glucosamine residues in peptidoglycan and between N-acetyl-D-glucosamine residues in chitodextrins. Subsequent to purification with ZipTips, peptides were measured via nLC–MS (Fig. 1h and i) and identification of peptides as well as ratio calculations were performed with the Elucidator software (Fig. 1j). When optimizing the digestion procedure using pure in vitro cultured bacteria, we were able to identify approx. 120 proteins (about 200 unique peptides) from 102 bacteria up to approx. 720 proteins (about 5000 unique peptides) from 105 bacteria with digestion times as short as one hour (Fig. 3a). The subsequent transformation and normalization of time dependent data have been performed by the program R and pathway analyses were done with e.g. Voronoi-like treemaps (Fig. 1k and l). Using the workflow presented in Fig. 1 we were finally able to identify about 600 S. aureus proteins over a period of 6.5 h post infection. However, only 40–50% of the peptides identified in total could be assigned to S. aureus because 50–60% of peptides were of
250
H. Pförtner et al. / Methods 61 (2013) 244–250
human origin. The majority of these mainly cytoplasmic proteins were originally localized in the endoplasmic reticulum (ER) or involved in intracellular protein transport (data not shown). This co-identification of human proteins might either reflect contamination of the gate area with human proteins or result from binding of these human proteins to S. aureus cells (Fig. 3c). On the other hand, the observation of these proteins in the S. aureus fraction might also indicate that the internalization process is likely ERmediated. Accordingly, co-localization studies showed that the bacteria often co-localized with the surrounded ER of the nucleus (Fig. 3b). Besides protein identification the workflow also included protein quantitation and pathway-oriented interpretation of the data. Since SILAC was used in a pulse-chase setup, different types of quantitative data were available. First, classical quantitation data were calculated by using the normalized ratios (Fig. 2a–f) and a fold change cutoff of two. With this method only 65% of the identified proteins could be quantified due to the fact that the quality of the labeled pairs was not always good enough to pass the threshold values. However, data analysis clearly showed that proteins belonging to the ribosome and the purine biosynthesis operon were down-regulated, whereas proteins belonging e.g. to the TCA cycle were increased in intensity after internalization. In addition to the relative quantitation, data about the turnover rate could be determined by SILAC as shown for the L9 subunit of the 50S ribosome. This subunit showed a higher than average ratio by only considering the normalized data. However, separate quantification of heavy and light peptides showed a clear and fast reduction in the level of heavy peptides following internalization which was specific to L9 (Fig. 2g–h), whereas intensity of light peptides did not display a significant change. Such a pattern was not seen for any other ribosomal protein monitored. The combination of these two sets of information indicates that the high heavy to light ratio observed for L9 was not due to increased synthesis but rather enhanced degradation. Yet another option for quantification is the application of alpha values (Fig. 2i), where the relative contribution of the sum of the intensities of the heavy and light peptides of a protein to the total protein amount is calculated. The quantification of total protein levels does not permit distinction of synthesis or degradation but allows quantification of all proteins identified and thus provided a more comprehensive view on the time dependent regulations of S. aureus simply because quantitative information was available for more proteins. Such quantitative data can be effectively visualized e.g. by kmeans clustering or Voronoi-like treemap (Fig. 2j–l), which focus on co-regulation of groups of genes or pathway/regulatory circuit centered presentation of the data. 4. Final remarks Using the experimental setup described in this communication, the analysis of pathogens in a wide variety of host–pathogen settings should become amendable to a thorough investigation. This approach will provide different qualitative and quantitative information (relative ratio, turnover, ratio trend) about the physiological adaptation reactions of pathogens to their intracellular environment at a thus far unprecedented depth with rather small numbers of prokaryotic or eukaryotic pathogens being required for analysis. At this point, we would like to mention that besides the classical use of SILAC completely labeled heavy amino acids and other labels such as SULAQ [21] or SULAQ34 [22] or heavy labeled substrate such as 13C-benzene [23] allow similar analyses. At a minimum, main branches of metabolism will be assessable. If the workflow is combined with pre-fractionation of host cells,
one might even be able to assign the adaptation reactions to particular sub-cellular compartments, e.g. the phagolysosome. However, such analyses might yield even fewer cells and thus further optimizations of the workflow to allow analysis of 102–105 cells (e.g. Fig. 3a) would be required. The value of the analysis of particular sub-cellular fractions would be further increased if combined with sophisticated imaging technologies (Fig. 3b). An additional application of the technology could be the dedicated and exhaust analysis of host proteins attached to the pathogens because these would be enriched with the bacterial cell sorting and be accessible to dedicated analysis as well (Fig. 3c). 5. Author contributions HP: Wet–lab experiments, data analysis, and writing. JW: Wet–lab experiments, data analysis, and writing. KS: Wet–lab experiments, data analysis, and writing. PH: Wet–lab experiments, data analysis, and writing. SS: Wet–lab experiments, and data analysis. JB: Data analysis, writing, and artwork. CS: Data analysis, and writing. MG: Wet–lab experiments, and data analysis. MD: Wet–lab experiments, and data analysis. UJ: Wet–lab experiments, data analysis, and writing. VD: Wet–lab experiments, and writing. EH: Concepts and design of experiments, and writing. LS: Data analysis, and writing. UV: Concepts and design of experiments, and writing. FS: Concepts and design of experiments, data analysis, artwork, and writing.
Acknowledgements We thank Manuela Gesell-Salazar, Lars Brinkmann for data acquisition and analysis, Susanne Engelmann and Leonhard Menschner for providing the strain HG001 containing pMV158GFP. This work was supported by funding from Deutsche Forschungsgemeinschaft via the SFB-Transregio 34, the Research Training Group GRK840, the BMBF within the framework of ZIK-FunGene, and the Alfried Krupp von Bohlen und Halbach-Stiftung. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
J.G. Lee, J.T. Yates Jr., J. Am. Chem. Soc. 126 (40) (2004) 13071–13078. J.R. Yates, C.I. Ruse, A. Nakorchevsky, Annu. Rev. Biomed. Eng. 11 (2009) 49–79. J.D. Boyce et al., Proteomics 6 (3) (2006) 870–880. J. Lee, R.A. Deininger, Luminescence 19 (1) (2004) 31–36. F. Schmidt et al., Proteomics (2010). K. Pohl et al., J. Bacteriol. 191 (9) (2009) 2953–2963. C. Nieto, M. Espinosa, Plasmid 49 (3) (2003) 281–285. D. Paape et al., Mol. Cell. Proteomics 7 (9) (2008) 1688–1701. C. Wiacek, S. Muller, D. Benndorf, Proteomics 6 (22) (2006) 5983–5994. N. Jehmlich et al., Appl. Microbiol. Biotechnol. 88 (2) (2010) 575–584. F. Schmidt, U. Volker, Proteomics 11 (15) (2011) 3203–3211. T.R. Flotte et al., Proc. Natl. Acad. Sci. USA 90 (22) (1993) 10613–10617. E.W. Deutsch et al., Proteomics 10 (6) (2010) 1150–1159. A. Keller et al., Anal. Chem. 74 (20) (2002) 5383–5392. R. Schleif, J. Mol. Biol. 27 (1) (1967) 41–55. J. Bernhardt et al., Genome Res. 13 (2) (2003) 224–237. A. Otto et al., Nat. Commun. 1 (2010) 137. J.A. Wu et al., Antimicrob. Agents Chemother. 47 (11) (2003) 3407–3414. J.K. Kumar, Appl. Microbiol. Biotechnol. 80 (4) (2008) 555–561. M. Sugai et al., J. Bacteriol. 179 (4) (1997) 1193–1202. N. Jehmlich et al., Proteomics 12 (1) (2012) 37–42. F.A. Herbst et al., Mol. Cell. Proteomics (2013). in press, Available online: 19 April, 2013. [23] M. Taubert et al., Proteomics 11 (11) (2011) 2265–2274.
Contents lists available at SciVerse ScienceDirect
Methods journal homepage: www.elsevier.com/locate/ymeth
A proteomics workflow for quantitative and time-resolved analysis of adaptation reactions of internalized bacteria Henrike Pförtner a, Juliane Wagner a, Kristin Surmann a, Petra Hildebrandt a,d, Sandra Ernst a, Jörg Bernhardt b, Claudia Schurmann a, Melanie Gutjahr a, Maren Depke a,d, Ulrike Jehmlich c, Vishnu Dhople a, Elke Hammer a, Leif Steil a, Uwe Völker a, Frank Schmidt a,d,⇑ a
Interfaculty Institute of Genetics and Functional Genomics, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15a, D-17487 Greifswald, Germany Institute for Microbiology, Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15, D-17487 Greifswald, Germany ZIK-HIKE Junior Research Group, Ernst-Moritz-Arndt-University Greifswald, Fleischmannstraße 42-44, D-17489 Greifswald, Germany d ZIK-FunGene Junior Research Group "Applied Proteomics", Ernst-Moritz-Arndt-University Greifswald, Friedrich-Ludwig-Jahn-Strasse 15a, D-17487 Greifswald, Germany b c
a r t i c l e
i n f o
Article history: Available online 30 April 2013 Keywords: Host–pathogen interaction S. aureus In vivo proteomics Pulse-chase labeling Cell sorting
a b s t r a c t The development of a mass spectrometric workflow for the sensitive identification and quantitation of the kinetics of changes in metaproteomes, or in particular bacterial pathogens after internalization by host cells, is described. This procedure employs three essential stages: (i) SILAC pulse-chase labeling and infection assay; (ii) isolation of bacteria by GFP-assisted cell sorting; (iii) mass spectrometry-based proteome analysis. This approach displays greater sensitivity than techniques relying on conventional cell sorting and protein separation, due to an efficient combination of a filtration-based purification and an on-membrane digestion. We exemplary describe the use of the workflow for the identification and quantitation of the proteome of 106 cells of Staphylococcus aureus after internalization by S9 human bronchial epithelial cells. With minor modifications, the workflow described can be applied for the characterization of other host–pathogen pairs, permitting identification and quantitation of hundreds of bacterial proteins over a time range of several hours post infection. Ó 2013 Elsevier Inc. All rights reserved.
1. Introduction Infectious diseases are still one of the major challenges for the health care systems. In 2002 about 26% of all cases of death worldwide, totaling nearly 15 million people, were due to infectious diseases [1]. The development of new biomarker targets and therapies requires a thorough understanding of the pathophysiology of infectious diseases, particularly the intimate interplay between the pathogen and the host during the progression of disease. The introduction and development of the collection of functional genomics technologies (genomics, transcriptomics, proteomics, and metabolomics) commonly known as OMICS-technologies have paved ⇑ Corresponding author. Address: ZIK-FunGene Junior Research Group ‘‘Applied Proteomics’’, Ernst-Moritz-Arndt-University Greifswald, Germany. Fax: +49 3834 86 795871. E-mail addresses: [email protected] (H. Pförtner), [email protected] (J. Wagner), [email protected] (K. Surmann), [email protected] (P. Hildebrandt), [email protected] (S. Ernst), [email protected] (J. Bernhardt), [email protected] (C. Schurmann), Melanie.Gutjahr@fli.bund.de (M. Gutjahr), [email protected] (M. Depke), [email protected] (U. Jehmlich), [email protected] (V. Dhople), [email protected] (E. Hammer), [email protected] (L. Steil), [email protected] (U. Völker), [email protected] (F. Schmidt). 1046-2023/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymeth.2013.04.009
the way for a new level of understanding, because they can provide a holistic view of the adaptive changes that take place during infection in both host and pathogen. While genomics and transcriptomics are routinely used to address infection-related questions at a system-wide level, limitations in sensitivity and specificity in the field of proteomics until very recently restricted the generation of complementing data sets at the protein level. Integration of these different OMICS-datasets will provide an integrated view from adaptation at the gene expression level through physiological and cellular adaptation. With the development of the newest generation of high-accuracy mass spectrometers (MS) such as LTQ-Orbitrap XL MS, the perspective changed because gel-free protein identification and quantitation of complex samples in minute amounts down to the low picogram range became possible [2]. Compared to classical setups dealing with pure cell cultures or isolated tissue, the characterization of pathogenic bacteria from host–pathogen settings are even more difficult, because only extremely small amounts of sample are accessible for the infecting pathogen. While one challenge is the sensitivity of current proteomics techniques, an even greater problem is the selection of adequate pre-separation techniques for the low numbers of bacteria available in such experimental setups. Different approaches have
H. Pförtner et al. / Methods 61 (2013) 244–250
been developed for isolation of bacteria from host cells such as sucrose gradient centrifugation [3], immunomagnetic separation [4], and FACS sorting via fluorescent proteins [5]. A common feature of these techniques is the requirement of large quantities of cells (>108 cells) for sample preparation and protein identification, making them unsuitable for experiments with limited biomass availability, such as characterization of internalized bacterial pathogens. Recently, we introduced a workflow that allows time-resolved quantitative monitoring of the adaptive response of only 106 cells of S. aureus HG001 [6] carrying plasmid pMV158GFP [7] after internalization by 4 105 S9 human bronchial epithelial cells, a nonprofessional phagocytic cell line. Recent studies have already combined flow cytometry and 2-DE to investigate 108–109 internalized pathogens requiring a sorting time of several days [8]. In our approach, we gained an increase in sensitivity by optimizing an established workflow (Fig. 1) that combined pulse-chase labeling, FACS sorting via GFP, on-membrane washing [9] and digestion [10], high sensitivity nLC-MS/MS, and comprehensive data analysis (Fig. 2). The prior labeling of bacteria with SILAC allowed quantification even of low intense signals in a correct manner and discrimination of such low-abundance bacterial peptides (heavy) from
245
host peptides (light), which is still crucial in high complex proteomics approaches. Using this approach, one can now investigate intracellular bacteria with initial starting material two to three orders of magnitude lower than what was usually required for a quantitative proteomics approach [11]. This workflow is not only suitable for the characterization of individual species but a modified version has also been applied to the study of a mixture of bacteria [10] and we can envision its application for the analysis of complex consortia in metaproteomics approaches. 2. Materials and methods 2.1. Reagents, chemicals and media Acetone (ice-cold, stored at 20 °C; Merck, cat No. 1.00014. 1011). Acetic acid (AA; Merck, cat. No. 1.00063.2511). Acetonitrile (ACN; Roth, cat. No. T195.2). Alexa FluorÒ 568 phalloidin (Invitrogen, cat. No. SKU#A12380). Amino acid mixture group A (Promocell; customer formulation), alanine, valine, leucine, isoleucine; 50 mM each, cat. No. C-97027).
Fig. 1. General workflow to identify and quantify proteins from internalized S. aureus HG001 cells in a time-resolved manner. (a) S9 human bronchial epithelial cells were grown to confluence in eMEM containing light arginine and lysine. In parallel S. aureus was cultivated to exponential growth phase in pMEM, containing the 13C6 heavy isotopes of arginine and lysine. (b) The S. aureus culture was diluted with eMEM to a MOI of 25 and then transferred to the host cells. (c) After 1 h of internalization remaining non-internalized bacteria were killed by lysostaphin treatment. After internalization, bacteria can only incorporate light counterparts of the heavy amino acids into their proteins. (d) Samples were taken hourly, host cells were disrupted by Triton X-100 and internalized S. aureus were released. (e) At each point in time, GFP-positive bacteria were separated from cell debris via FACS and sorted on a low protein binding filter device. (f) and subjected to proteolytic digest first by lysostaphin. (g) followed by trypsin. (h) Tryptic peptides were purified by ZipTip and (i) measured via nLC–LTQ-Orbitrap XL MS. (j) Identification and determination of peptide ratios were performed with the Elucidator software package. (k) Protein light vs. heavy ratios with a more than twofold change relative to the median over the time course were defined as up- or downregulated. Proteins with a smaller deviation of their ratio from the average ratio of all proteins were defined as non-regulated. (l) Regulated proteins were interpreted in the context of their role in cellular physiology.
246
H. Pförtner et al. / Methods 61 (2013) 244–250
Amino acid mixture group B (Promocell; customer formulation), aspartate, glutamate; 50 mM each, cat. No. C97028). Amino acid mixture group C (Promocell; customer formulation), serine, threonine, cysteine; 50 mM each, cat. No. C-97029). Amino acid mixture group D (Promocell; customer formulation), proline, phenylalanine, histidine; 50 mM each, cat. No. C-97030). Ammonium bicarbonate (Sigma, cat. No. A6141). 13C6 L-Arginine–HCl (EURISO-Top GmbH, cat. No. CLM2265-0.25). Ethanol, analytic grade (Merck, cat. No. 1.00983.1011). FACSFlow (BD, cat. No. 342003). Fetal Bovine Serum (FBS; Biochrom AG, cat. No. S0115). L-glutamine, 200 mM (PAA, cat. No. M11-004). Glycerol (Sigma–Aldrich, cat. No. G6279). HEPES buffer solution, 1 M (PAA, cat. No. S11-001). High precision cover slips, 18 mm diameter (Marienfeld, cat. No. 011758018). Hoechst 33258 (Sigma–Aldrich, cat. No. 861405). 13C6 L-Lysine–2HCl (EURISO-Top GmbH, cat. No. CLM2247-0.25). Lysostaphin (AmbicinR L; AMBI Products, cat. No. LSPN-50). MEM Non Essential Amino Acids, 100 (NEAA; PAA, cat. No. M11-003). Minimal essential media (MEM Earl, w/o L-glutamine; Promocell, cat. No. C-75210). Minimal essential media 2 (MEM Earl; Promocell; customer formulation w/o stable glutamine, w/o L-arginine, w/o L-lysine, w/o sodium bicarbonate, with 6.8 g/L sodium chloride), cat. No. C-96187). p-Phenylendiamin (Roth, cat. No. 4499.1). Phosphate buffered saline Mg, Ca (PBS Mg, Ca; Dulbecco’s PBS 10 without Ca & Mg; PAA, cat. No. H15-011). Phosphate buffered saline+Mg, Ca (PBS+Mg, Ca; Dulbecco’s PBS 1 with Ca & Mg; PAA, cat. No. H15-001). Select Agar (Invitrogen, cat. No. 30391-023). Sodium Bicarbonate (PAN-Biotech GmbH, cat. No. P0444100). Sodium hydroxide 2 mol/L (Roth, cat. No. T135.1). Trifluoroacetic acid (TFA; Applied Biosystems, cat. No. 400028). Tryptophan (Promocell; customer formulation); 25 mM, cat. No. C-97031). Tryptic Soy Broth (TSB; BD Bacto™, cat. No. 211825). Triton X-100 (ESA Laboratories Inc., cat. No. 80-0072). Trypsin (Promega, cat. No. V5111). Trypsin EDTA, 1 (PAA, cat. No. L11-004). Water HPLC-grade (JT Baker, cat. No. 4218).
Lysostaphin: 50 mg vial was dissolved in 10 ml sterile HPLCgrade water, aliquoted and stored at 20 °C until use. Mowiol mounting medium: 6 g glycerol and 2.4 g Mowiol were dissolved in 6 ml of distilled water by shaking at RT for at least 2 h. Afterwards, 12 ml of 0.2 M TRIS/HCl (pH 8.5) was added and shaken at 53 °C until everything was dissolved. 0.1% p-phenylendiamine was added as antifading reagent and shaken again for 3 h. Aliquots of the solution were stored at 20 °C. 1 PBS Mg, Ca for bacterial dilution series: 10 PBS Mg, Ca were diluted 1:10 in A. dest. autoclaved and stored at room temperature until use. Phalloidin staining solution: 5 ll Phalloidin-Alexa Fluor 586 was diluted in 1 ml PBS Mg, Ca (final concentration 1 U/ml). The dilution was prepared freshly for each staining. 0.1% Triton X-100: The 1% stock solution was diluted 1:10 in sterile HPLC-grade water. The dilution was prepared one day before the experiment and sterile filtrated to avoid contaminations. Trypsin: A vial containing 20 lg was filled with 200 ll ammonium bicarbonate (20 mM) to dissolve trypsin. Aliquots of this solution were stored at 20 °C until use. TSB agar plates: 30 g TSB powder and 15 g select agar were dissolved in 1 l A. dest., autoclaved and used to pour plates. Solutions for ZipTip peptide purification (10 samples): each solution was prepared freshly before use (see Table 1). 2.3. Methods 2.3.1. Preparation of eMEM and SILAC pMEM media For internalization assays eMEM and for labeling of bacteria SILAC pMEM were prepared. For eMEM 4 ml fetal bovine serum (FBS), 2 ml glutamine (200 mM) and 1 ml NEAA (100) were added to 100 ml MEM (1). The medium was mixed thoroughly and stored at 4 °C. Before application to the cells the medium was preheated to 37 °C. For SILAC pMEM, compounds listed in Table 2 were used. The medium was further adjusted to a pH of 7.4 with 1 M NaOH in HPLC-grade water and sterilized by filtration using a pore size of 0.22 lm. 2.3.2. Bacterial growth conditions A glycerol stock of S. aureus HG001 [6] pMV158GFP [7] was serially diluted (10 3–5 10 6) in SILAC pMEM with tetracycline Table 1 Solutions for ZipTip peptide purification (elution).
100% ACN 5% AA HPLC-grade water Total Volume
80% ACN in 1% AA
50% ACN in 1% AA
484.0 ll 121.0 ll – 605.0 ll
302.5 ll 121.0 ll 181.5 ll 605.0 ll
Table 2 SILAC pMEM composition for 80 ml.
2.2. Reagent setup 20 mM ammonium bicarbonate (pH 7.8): 20 mg ammonium bicarbonate was dissolved in 12.5 ml HPLC-grade water. For each experiment fresh ammonium bicarbonate buffer was prepared. LC Buffer A: 2% acetonitrile and 0.1% acetic acid in HPLC-grade water. The buffer was stored at room temperature. LC Buffer B: 100% acetonitrile with 0.1% acetic acid. This buffer was also stored at room temperature. Hoechst staining solution: 20 lg/ml was prepared as stock solution and aliquots were stored at 20 °C. The Hoechst 33258 stock solution was diluted 1:20 in PBS Mg, Ca (final concentration 1 lg/ ml) for staining of DNA. Each dilution was freshly prepared for staining. 5 mg/ml.
Stock
Volume
Final concentration
2 MEM 100 NEAA 200 mM L-glutamine 1 M HEPES AA group A (Ala, Val, Leu, Ile) 50 mM each AA AA group B (Asp, Glu) 50 mM each AA AA group C (Ser, Thr, Cys) 50 mM each AA AA group D (Pro, Phe, His) 50 mM each AA Trp 25 mM 13 C6 Arg–HCl (125 mg/ml = 576 mM) 13 C6 Lys–2HCl (125 mg/ml = 556 mM) HPLC-grade water
40.0 ml 0.8 ml 1.6 ml 0.8 ml 3.2 ml
1 MEM 1 NEAA 4 mM L-glutamine 10 mM HEPES 2 mM each AA
3.2 ml 3.2 ml 3.2 ml 6.4 ml 82.9 ll 57.1 ll 17.46 ml
2 mM each AA 2 mM each AA 2 mM each AA 2 mM Trp 0.597 mM Arg 0.397 mM Lys
H. Pförtner et al. / Methods 61 (2013) 244–250
(20 lg/ml) and grown at 37 °C and 220 rpm for about 12 h. The 20 ml SILAC pMEM main culture was inoculated with an exponentially growing over night culture to an OD600 nm of 0.05. The culture was grown at 37 °C and 150 rpm in a water bath to OD600 nm 0.4 (exponential growth phase) and used for infection of eukaryotic cells. 2.3.3. Host cell line and cultivation conditions The S9 epithelial cell line [12] was cultured in eMEM supplemented with 4% of fetal bovine serum (FBS) at 37 °C in 5% CO2. For the S. aureus infection assay, epithelial cells were seeded 3 days in advance in 24-well tissue culture plates and cultured to confluence (4 105 cells/well). 2.3.4. Infection assays S9 cells were infected with S. aureus at a multiplicity of infection (MOI) of 25. For this purpose, S9 cells of one representative well were detached with 200 ll trypsin–EDTA and the trypsin reaction was stopped by adding 300 ll of eMEM. After homogenization, the cell solution was mixed with Trypan blue dye and the number of viable eukaryotic cells was counted with a Countess (Invitrogen, Karlsruhe, Germany). Afterwards, cells were incubated with an infection master mix, composed of S. aureus in eMEM (MOI of 25) and 29 ll NaHCO3 (7.5%) per ml of bacterial culture added to the infection master mix. Infection master mix and the S9 cells were co-incubated at 37 °C and 5% CO2 for 1 h allowing internalization of Staphylococci. Meanwhile 100 ll of the infection master mix were diluted in PBS Mg, Ca and plated on TSB agar plates to determine bacterial titer of the master mix for later calculation of the precise MOI reached. After 1 h of co-incubation the infection master mix was removed and pre-warmed eMEM with lysostaphin was added (final concentration 10 lg/ml). Incubation of cell culture plates was then continued at 37 °C and 5% CO2. For FACS sorting at desired points in time (1.5–6.5 h post infection), cell culture medium was removed and the S9 cells were washed twice with PBS+Mg, Ca. To lyze the S9 cells, 150 ll of Triton X-100 (0.1%) per well was added and the plate was further incubated for about 7 min at 37 °C. Host cells lysates with internalized bacteria were homogenized by pipetting and pooled in a 15 ml tube. This procedure did not disrupt bacteria. Cell culture plate wells were rinsed with 50 ll FACSFlow per well and this solution was added to the 15 ml tube. Samples were transferred immediately on ice to a FACSAria (Becton Dickinson, CA, USA) for sorting. In parallel, 100 ll of the homogenized sample were serially diluted in PBS Mg, Ca and plated to determine the titer of internalized bacteria. 2.3.5. Isolation of bacteria from host cells by FACS on a 96-well plate filter membrane The GFP labeled bacteria were sorted applying an appropriate fluorescence channel and SSC-A gates onto a low protein binding filter membrane (0.22 lm pore size) of a 96-well microtiter plate (Millipore, Schwalbach, Germany) by applying vacuum (450– 550 mbar) to the microtiter plate to allow constant removal of the fluid. The filter was rinsed with 200 ll FACSFlow and the membrane was cut out and divided into small pieces using a scalpel. The filter pieces were placed into a reaction vial for subsequent onmembrane digestion. 2.3.6. On-membrane protein digestion for Shotgun Mass Mapping (SMM) 12.33 ll of 20 mM ammonium bicarbonate (ABC) buffer was added to the filter membrane containing sorted bacteria. Afterwards, 1 ll lysostaphin (0.05 lg/ll) was added to the filter, mixed thoroughly and incubated for 30 min in a water bath at 37 °C. Then, 6.67 ll trypsin (0.1 lg/ll) was added, mixed thoroughly and incubated several hours or overnight (16 h) in water bath at 37 °C. The
247
enzymatic reaction was stopped by adding TFA to a final concentration of 0.1% and incubating the samples for 5 min at 37 °C. After centrifugation for 10 min at 13,000g at room temperature the supernatant was transferred into a new microfuge tube (0.5 ml). Peptides were purified with minor modification by ZipTip as recommended in the manufacturer’s instructions as follow: peptides were bound from supernatant to ZipTip, washed, and eluted by rinsing 10 times with 5 ll 50% ACN-buffer and 5 ll 80% ACN-buffer in 1% acetic acid and dried in a Speed-vacuum centrifuge (Eppendorf, Hamburg, Germany). 2.3.7. nLC–LTQ-Orbitrap XL MS measurements of peptides Dried peptides were dissolved using 10 ll LC buffer A (2% ACN, 0.1% AA) and transferred to LC vials. The mass spectra were acquired online after peptide separation by one-dimensional nLC. For the example presented here the following conditions were used: peptides were enriched on a BioSphere C18 pre-column and separated using an Acclaim PepMap 100 C18 column on a Proxeon Easy nLC (Proxeon, Odense, Denmark). Separation was achieved with the formation of non linear gradient of 290 min containing LC Buffer A and LC Buffer B 0% for 15 min, 0–30% for 260 min, 30– 60% for 15 min, 60–100% for 1 min and equilibrated for 5 min. Peptides were eluted with a flow rate of 300 nl/min and analyzed using a LTQ-Orbitrap XL MS (Thermo Scientific, Bremen, Germany) with an interfaced Nano ESI source. The full scan was carried out using a FTMS analyzer with normal mass range of m/z 300–2000. Data were acquired in profile mode with a resolution of 60,000 and positive polarity. This was followed by Data Dependent Acquisition of the top five most intense precursor ions for MS/MS scans using CID activation. A minimum of 1000 counts were activated for 30 ms with an activation of q = 0.25, isolation width 2.00 Da and a normalized collision energy of 35%. The charge state screening and monoisotopic precursor selection was enabled with the rejection of +1, +4, and unassigned charge states. Peptides were excluded for the next 30 s once MS/MS scans had been acquired. 2.3.8. Peptide identification and quantitation with the Trans Proteomics Pipeline (TPP) [13] Tandem MS spectra were used to identify peptides with the classical standard search engine Sequest™ and a S. aureus FASTA database embedded in Sorcerer™. A precursor tolerance of 20 ppm, one missed cleavage and [13C6]-arginine and [13C6]-lysine as possible modifications were applied. All first-ranked hits were further consolidated by the standard algorithm Peptide Prophet [14]. The linked survey-scan number and the assigned peptide identification were used to calculate the ion and time-dependent intensities of the light and corresponding heavy peptides, which were further used for quantitation by the XPRESS tool from the TPP. 2.3.9. Data analysis with Rosetta Elucidator In order to perform a spectral alignment, which cannot be performed by the TTP software so far, the final analysis was done with the Rosetta Elucidator software (Rosetta Biosoftware by Ceiba Solution Inc., MA, USA). First, raw data files from nLC–LTQ-Orbitrap XL MS were loaded. An experimental definition for quantitative protein analysis was created where an instrument mass accuracy of 10 ppm and modified amino acids arginine and lysines with mass shift of 6.02 Da were applied. Further, a retention time minimum cutoff of 40.0 (min) and a maximum cutoff of 270.0 (min) as well as a m/z minimum cutoff of 300.0 and maximum of 1200.0 were set. A peak time width minimum (min) of 0.2 was applied and spectral alignment (alignment search distance 10.0 min) was used for binning. For quantitation, a maximum of two labelings per peptide were allowed and detection parameters were set as follow: PPM error 10.0, left RT location tolerance (minutes) 0.8,
248
H. Pförtner et al. / Methods 61 (2013) 244–250
right RT location tolerance (minutes) 0.8, error model to apply: pair ratio error model, limit fold change to (±) 1000, and normalization was skipped. After identification (similar parameter as shown in 2.3.8), an automatic Peptide/Protein Tellers annotation was performed and only Peptide Teller results greater than 0.8 were used and at least two peptides were necessary for further consideration.
2.3.10. Quantitative data analysis using alpha values and ratios For quantitative analysis, a tab-delimited matrix was created from the Rosetta Elucidator Excel report, containing protein names, intensities, not normalized ratios (rnormalized = rnot normalized/
rmedian_all) and alpha values [15] (see Fig. 2i) from intensities of each point in time per independent biological sample series. Afterwards, a weighted or non-weighted linear regression for the six points in time was performed, and ratios from the values at t6.5 h and t1.5 h using linear regression equations (note: these ratios are more robust than using original values) were calculated. Furthermore, fold change cutoff and quantile-information were assigned using an R-script. The data were further classified by kmeans clustering (Fig. 2j) using the Analyst Software from GeneData (Genedata AG, Basel, Switzerland) and following parameter settings: Distance: Positive Correlation Distance (1 r), Centroid Calculation: Mean, Valid Values: 50, Max Iterations: 50.
Fig. 2. Data analysis workflow for proteome data. (a) After spectral alignment and isotope grouping, the corresponding isotopic peptide pairs were defined based on a mass shift of 6.02 Da. Image a shows a peptide pair of the protein PrsA, which was found to be upregulated during the time-course. (b) From each peptide pair, the areas of the selected ion chromatograms were calculated and further used for ratio calculation. (c) All available peptide ratios of a protein were squeezed to one protein ratio and normalized by the median. Furthermore, the corresponding ratios of three independent experimental series (bioreplicates) were compared. (d) In order to make the analysis more robust, a linear regression of all proteins was performed and their slopes were calculated. (e) The binning of ratios from time point 6.5 h divided by 1.5 h of the proteins of the three bioreplicates were plotted as histograms or (f) sigmoid curves. To assign significant changes a quantile calculation and fold change cutoffs were applied. (g) The normalized total intensity of the L9 ribosomal protein is given, where the general trend shows no significant regulation. (h) The normalized single intensities of the light and heavy peptides are also given, indicating a decrease in the amount of the heavy peptides over time. Together with the stable summed intensity of the light peptides this observation indicates enhanced degradation of L9. (i) The alpha value for normalization of protein levels was calculated by first calculating the sum of the peptide intensities for each individual protein and the sum of the intensity of all proteins observed and then dividing at each time point the intensities of individual proteins by the sum of all protein intensities at the same time point multiplied by factor of 100. For the calculation of the alpha values for single proteins the sum of intensities of the light and the heavy peptides of the protein were calculated. (j) Based on the alpha value data a Kmeans (n = 9) clustering (GeneData Analyst Software package) was applied. The nine clusters (C1–9) and the number of fitting proteins are shown. (k) Two representative clusters for groups of regulated proteins are shown. (l) In order to map the quantitative data onto pathways/regulatory circuits a Voronoi-like treemap was used. The figure shows an image section of the TCA, glycolysis and ribosomal proteins. The light red colors indicate a downregulation of many ribosomal proteins and some proteins belonging to cations and iron carrying compounds. The light blue color of proteins involved in TCA and glycolysis indicates an increase of the corresponding alpha values during internalization.
H. Pförtner et al. / Methods 61 (2013) 244–250
249
Fig. 3. (a) Enzymatic digestion of 102–105 GFP-labeled S. aureus bacteria. Filtered bacteria were incubated with lysostaphin first followed by trypsin for 1, 2, 3, 4, and 5 h and over night (grey triangles). The number of unique identified peptides identified ranged from 200 for 102 bacteria to 5000 for 105 bacteria, with enzymatic digestion times as short as one hour. Increasing incubation times did not significantly influence the number of peptides identified. (b) Fluorescence image of GFP expressing bacteria internalized by S9 human bronchial epithelial cells. The nuclei of the S9 cells are shown in blue and the f-actin in red. (c) Dot plot for sorting of GFP expressing bacteria. After internalization and lysis of the host cells, bacteria and attached human host proteins were sorted.
2.3.11. Data visualization and analysis using Voronoi-like treemaps First, a construction of a hierarchical structure of gene functional categories (e.g. Riley Scheme based TIGR S. aureus functional classification) was performed. Second, an iterative hierarchy level by level construction of a Voronoi-like treemap layout [16,17] was applied that contains all S. aureus proteins at the lowest level (mosaic tiles on Fig. 2l). Furthermore, mapping of protein quantitation data to an appropriate color scale and of colors to Voronoi cells representing the appropriate proteins was performed. 2.3.12. Visualization of internalized bacteria by confocal microscopy The internalization was performed as described in Section 2.3.4. At desired points in time cell culture medium was removed and cells were washed twice with PBS+Mg, Ca and fixed with 2% formaldehyde in PBS+Mg, Ca for at least 20 min at RT in the dark. After fixation formaldehyde was replaced with PBS+Mg, Ca. Plates can be stored at 4 °C in the dark until the staining of the eukaryotic cells is performed. At first, unspecific binding sites were blocked with 10% FBS in PBS+Mg, Ca and cover slips were incubated for 20 min at RT in the dark. The blocking solution was aspirated and 500 ll 0.1% Triton X-100 were added for 3–5 min. Wells were washed with PBS+Mg, Ca twice, 400 ll Phalloidin staining solution were added to stain the f-actin and wells were again incubated for 20 min at 37 °C and 5% CO2. Afterwards, cover slips were rinsed with PBS+Mg, Ca. In addition, 400 ll Hoechst staining solution were added to stain the DNA of the nuclei and incubation proceeded for 10 min at RT in the dark. Cover slips were then washed three times with distilled water and mounted upside down with MOWIOL containing antifading reagent on microscopic slides. Fluorescence pictures were obtained using a Leica TCS-SP5 DMI6000i confocal laser scanning microscope equipped with an argon laser (488 nm) for the excitation of GFP labeled bacteria, a UV laser (405 nm) for the excitation of Hoechst labeled nuclei and a helium–neon laser (594 nm) for the excitation of Phalloidin labeled f-actin skeleton of the S9 cells. Emission of each fluorescent dye was recorded according to manufacturer recommendations. The original images were processed with ZIK-ImageJ (free image analysis software) to optimize brightness, contrast and sharpness of the final pictures (Fig. 3b). 3. Results and discussion In order to identify and quantify proteins from internalized S. aureus cells, S9 human epithelial cells were cultivated in eMEM containing light arginine and lysine and in parallel S. aureus HG001 cells were cultivated in pMEM containing heavy arginine and lysine for at least five to six generations. The S9 cells were infected with heavily labeled S. aureus for 1 h and after internalization remaining non-internalized bacteria were killed by
lysostaphin treatment (Fig. 1a–c). The effective killing by lysostaphin allowed the exclusive analysis of internalized bacteria. Since concomitant with the infection the heavy amino acids lysine and arginine were replaced by their normal, light counterparts, internalized bacteria were forced to incorporate light lysine and arginine derivatives instead of heavy amino acids into their newly synthesized proteins after the start of the infection experiment. This pulse-chase design allowed the calculation of ratios between heavy and light peptides and thus quantification of protein synthesis after internalization. Information on intensity of heavy and light peptides could also be analyzed independently, thus even facilitating the distinction of altered synthesis and stability. To enrich the bacteria, S9 cells were carefully disrupted hourly by adding a nonionic surfactant until 6.5 h post infection and from each point in time GFP containing bacteria were selectively enriched by FACS (Fig. 1d and e). With this procedure, between 104–106 cells could be isolated per sample (determined by FACS sorting). To reduce the final sample volume, cells were collected on a low protein binding membrane where the proteolytic digestion (first by lysostaphin and followed by trypsin) took place (Fig. 1f and g). Lysostaphin is produced by S. simulans and contains an N-terminal glycyl-glycine-endopeptidase M23 and a C-terminal cell wall binding domain. The enzyme hydrolyzes the pentaglycine bridges in the peptidoglycan of S. aureus, even if the cells are not growing [18,19]. A homologue to lysostaphin, ALE-1, was also identified in S. capitis EPK1 [20] and both enzymes show a lytic activity specific to S. aureus. This enzymatic digestion of the S. aureus cell wall was used to improve the efficiency of cell disruption and to maximize access of trypsin to cytoplasmic proteins. For other bacteria lysozyme would be an alternative since this muramidase is a glycoside hydrolases that degrades bacterial cell walls by catalyzing hydrolysis of 1,4-beta-linkages between N-acetylmuramic acid and N-acetyl-D-glucosamine residues in peptidoglycan and between N-acetyl-D-glucosamine residues in chitodextrins. Subsequent to purification with ZipTips, peptides were measured via nLC–MS (Fig. 1h and i) and identification of peptides as well as ratio calculations were performed with the Elucidator software (Fig. 1j). When optimizing the digestion procedure using pure in vitro cultured bacteria, we were able to identify approx. 120 proteins (about 200 unique peptides) from 102 bacteria up to approx. 720 proteins (about 5000 unique peptides) from 105 bacteria with digestion times as short as one hour (Fig. 3a). The subsequent transformation and normalization of time dependent data have been performed by the program R and pathway analyses were done with e.g. Voronoi-like treemaps (Fig. 1k and l). Using the workflow presented in Fig. 1 we were finally able to identify about 600 S. aureus proteins over a period of 6.5 h post infection. However, only 40–50% of the peptides identified in total could be assigned to S. aureus because 50–60% of peptides were of
250
H. Pförtner et al. / Methods 61 (2013) 244–250
human origin. The majority of these mainly cytoplasmic proteins were originally localized in the endoplasmic reticulum (ER) or involved in intracellular protein transport (data not shown). This co-identification of human proteins might either reflect contamination of the gate area with human proteins or result from binding of these human proteins to S. aureus cells (Fig. 3c). On the other hand, the observation of these proteins in the S. aureus fraction might also indicate that the internalization process is likely ERmediated. Accordingly, co-localization studies showed that the bacteria often co-localized with the surrounded ER of the nucleus (Fig. 3b). Besides protein identification the workflow also included protein quantitation and pathway-oriented interpretation of the data. Since SILAC was used in a pulse-chase setup, different types of quantitative data were available. First, classical quantitation data were calculated by using the normalized ratios (Fig. 2a–f) and a fold change cutoff of two. With this method only 65% of the identified proteins could be quantified due to the fact that the quality of the labeled pairs was not always good enough to pass the threshold values. However, data analysis clearly showed that proteins belonging to the ribosome and the purine biosynthesis operon were down-regulated, whereas proteins belonging e.g. to the TCA cycle were increased in intensity after internalization. In addition to the relative quantitation, data about the turnover rate could be determined by SILAC as shown for the L9 subunit of the 50S ribosome. This subunit showed a higher than average ratio by only considering the normalized data. However, separate quantification of heavy and light peptides showed a clear and fast reduction in the level of heavy peptides following internalization which was specific to L9 (Fig. 2g–h), whereas intensity of light peptides did not display a significant change. Such a pattern was not seen for any other ribosomal protein monitored. The combination of these two sets of information indicates that the high heavy to light ratio observed for L9 was not due to increased synthesis but rather enhanced degradation. Yet another option for quantification is the application of alpha values (Fig. 2i), where the relative contribution of the sum of the intensities of the heavy and light peptides of a protein to the total protein amount is calculated. The quantification of total protein levels does not permit distinction of synthesis or degradation but allows quantification of all proteins identified and thus provided a more comprehensive view on the time dependent regulations of S. aureus simply because quantitative information was available for more proteins. Such quantitative data can be effectively visualized e.g. by kmeans clustering or Voronoi-like treemap (Fig. 2j–l), which focus on co-regulation of groups of genes or pathway/regulatory circuit centered presentation of the data. 4. Final remarks Using the experimental setup described in this communication, the analysis of pathogens in a wide variety of host–pathogen settings should become amendable to a thorough investigation. This approach will provide different qualitative and quantitative information (relative ratio, turnover, ratio trend) about the physiological adaptation reactions of pathogens to their intracellular environment at a thus far unprecedented depth with rather small numbers of prokaryotic or eukaryotic pathogens being required for analysis. At this point, we would like to mention that besides the classical use of SILAC completely labeled heavy amino acids and other labels such as SULAQ [21] or SULAQ34 [22] or heavy labeled substrate such as 13C-benzene [23] allow similar analyses. At a minimum, main branches of metabolism will be assessable. If the workflow is combined with pre-fractionation of host cells,
one might even be able to assign the adaptation reactions to particular sub-cellular compartments, e.g. the phagolysosome. However, such analyses might yield even fewer cells and thus further optimizations of the workflow to allow analysis of 102–105 cells (e.g. Fig. 3a) would be required. The value of the analysis of particular sub-cellular fractions would be further increased if combined with sophisticated imaging technologies (Fig. 3b). An additional application of the technology could be the dedicated and exhaust analysis of host proteins attached to the pathogens because these would be enriched with the bacterial cell sorting and be accessible to dedicated analysis as well (Fig. 3c). 5. Author contributions HP: Wet–lab experiments, data analysis, and writing. JW: Wet–lab experiments, data analysis, and writing. KS: Wet–lab experiments, data analysis, and writing. PH: Wet–lab experiments, data analysis, and writing. SS: Wet–lab experiments, and data analysis. JB: Data analysis, writing, and artwork. CS: Data analysis, and writing. MG: Wet–lab experiments, and data analysis. MD: Wet–lab experiments, and data analysis. UJ: Wet–lab experiments, data analysis, and writing. VD: Wet–lab experiments, and writing. EH: Concepts and design of experiments, and writing. LS: Data analysis, and writing. UV: Concepts and design of experiments, and writing. FS: Concepts and design of experiments, data analysis, artwork, and writing.
Acknowledgements We thank Manuela Gesell-Salazar, Lars Brinkmann for data acquisition and analysis, Susanne Engelmann and Leonhard Menschner for providing the strain HG001 containing pMV158GFP. This work was supported by funding from Deutsche Forschungsgemeinschaft via the SFB-Transregio 34, the Research Training Group GRK840, the BMBF within the framework of ZIK-FunGene, and the Alfried Krupp von Bohlen und Halbach-Stiftung. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22]
J.G. Lee, J.T. Yates Jr., J. Am. Chem. Soc. 126 (40) (2004) 13071–13078. J.R. Yates, C.I. Ruse, A. Nakorchevsky, Annu. Rev. Biomed. Eng. 11 (2009) 49–79. J.D. Boyce et al., Proteomics 6 (3) (2006) 870–880. J. Lee, R.A. Deininger, Luminescence 19 (1) (2004) 31–36. F. Schmidt et al., Proteomics (2010). K. Pohl et al., J. Bacteriol. 191 (9) (2009) 2953–2963. C. Nieto, M. Espinosa, Plasmid 49 (3) (2003) 281–285. D. Paape et al., Mol. Cell. Proteomics 7 (9) (2008) 1688–1701. C. Wiacek, S. Muller, D. Benndorf, Proteomics 6 (22) (2006) 5983–5994. N. Jehmlich et al., Appl. Microbiol. Biotechnol. 88 (2) (2010) 575–584. F. Schmidt, U. Volker, Proteomics 11 (15) (2011) 3203–3211. T.R. Flotte et al., Proc. Natl. Acad. Sci. USA 90 (22) (1993) 10613–10617. E.W. Deutsch et al., Proteomics 10 (6) (2010) 1150–1159. A. Keller et al., Anal. Chem. 74 (20) (2002) 5383–5392. R. Schleif, J. Mol. Biol. 27 (1) (1967) 41–55. J. Bernhardt et al., Genome Res. 13 (2) (2003) 224–237. A. Otto et al., Nat. Commun. 1 (2010) 137. J.A. Wu et al., Antimicrob. Agents Chemother. 47 (11) (2003) 3407–3414. J.K. Kumar, Appl. Microbiol. Biotechnol. 80 (4) (2008) 555–561. M. Sugai et al., J. Bacteriol. 179 (4) (1997) 1193–1202. N. Jehmlich et al., Proteomics 12 (1) (2012) 37–42. F.A. Herbst et al., Mol. Cell. Proteomics (2013). in press, Available online: 19 April, 2013. [23] M. Taubert et al., Proteomics 11 (11) (2011) 2265–2274.