Characterization of the Salmonella Typhimurium Proteome by Semi-automated Two Dimensional HPLC-Mass Spectrometry: Detection of Proteins Implicated in Multiple Antibiotic Resistance Nick G. Coldham* and Martin J. Woodward Department of Food and Environmental Safety, Veterinary Laboratories Agency, Addlestone, Surrey, KT15 3NB, United Kingdom Received December 22, 2003
The proteome of Salmonella enterica serovar Typhimurium was characterized by 2-dimensional HPLC mass spectrometry to provide a platform for subsequent proteomic investigations of low level multiple antibiotic resistance (MAR). Bacteria (2.15 ( 0.23 × 1010 cfu; mean ( s.d.) were harvested from liquid culture and proteins differentially fractionated, on the basis of solubility, into preparations representative of the cytosol, cell envelope and outer membrane proteins (OMPs). These preparations were digested by treatment with trypsin and peptides separated into fractions (n ) 20) by strong cation exchange chromatography (SCX). Tryptic peptides in each SCX fraction were further separated by reversedphase chromatography and detected by mass spectrometry. Peptides were assigned to proteins and consensus rank listings compiled using SEQUEST. A total of 816 ( 11 individual proteins were identified which included 371 ( 33, 565 ( 15 and 262 ( 5 from the cytosolic, cell envelope and OMP preparations, respectively. A significant correlation was observed (r2 ) 0.62 ( 0.10; P < 0.0001) between consensus rank position for duplicate cell preparations and an average of 74 ( 5% of proteins were common to both replicates. A total of 34 outer membrane proteins were detected, 20 of these from the OMP preparation. A range of proteins (n ) 20) previously associated with the mar locus in E. coli were also found including the key MAR effectors AcrA, TolC and OmpF. Keywords: Salmonella Typhimurium • proteome • OMPs • antibiotic resistance • 2-dimensional-HPLC-mass spectrometry
Introduction Food borne diseases caused by nontyphoidal Salmonella present a significant public health problem worldwide.1 The emergence of strains resistant to multiple antibiotics, such as the multi-drug resistant Salmonella Typhimurium DT104 which produced a recent epidemic in the United Kingdom, are cause for further concern. The mechanism for the emergence of resistant strains of clinical significance is unresolved but may involve intermediaries characterized by intrinsic low level resistance to multiple antibiotics.2 Such resistance to multiple antibiotics is often associated with differential expression of innate genes and characteristic of the MAR phenotype.3 The control of such gene expression is complex and involves a regulon within which the mar locus plays a pivotal role.4 Mutation within the regulatory elements of genes may occur and further increase the level of resistance to multiple antibiotics.5 Transcriptomic studies6 have indicated that in E. coli the marRAB locus mediates a global stress response involving a large network of genes. Constitutive expression of the activator marA has revealed differential transcription of over 60 genes; these included increased transcription of genes encoding the * To whom correspondence should be addressed. Telephone: -01932357827. Fax: -01932-357595. E-mail:
[email protected]. 10.1021/pr034129u CCC: $27.50
2004 American Chemical Society
efflux pump protein AcrA, the outer membrane channel TolC, and reduced transcription of the porin OmpF. Other genes of diverse function including those involved in energy metabolism, biosynthesis of amino and fatty acids and ribosomal subunits may also be responsive to the marRAB operon. In some cases, the responsive genes would not necessarily be directly associated with resistance to antibiotics. One test for the MAR phenotype is for resistance to cyclohexane,7 but those proteins which afford low level resistance to antibiotics may differ from those that provide protection from cyclohexane. Differential protein expression by MAR strains offers the opportunity for comparison with wild types and discovery of biomarkers for multiple antibiotic resistance. Proteome characterization is an important first step toward the detection and discovery of such biomarker proteins. Until recently, two-dimensional gel electrophoresis has been used for proteome analysis but, although this technology has high resolving power, there are also significant limitations.8 Recently, two-dimensional chromatography with on-line mass detection has been applied to the large scale automated/semiautomated characterization of prokaryote proteomes. This method has enabled detection of 1,504 and 1,147 proteins derived from the proteomes of Saccharomcyces cerevisiae9 and E. coli,10 respectively. Experimental formats differ8 but, often Journal of Proteome Research 2004, 3, 595-603
595
Published on Web 04/08/2004
research articles protein extracts are digested with a proteolytic enzyme, such as trypsin, and orthogonal strong cation exchange and reversed phase chromatographies employed for the separation of the highly complex tryptic peptide mixture with on-line introduction into the mass spectrometer. This approach has enabled the detection of proteins of predicted low abundance, high mass, extreme pI and hydrophobicity, which are usually intractable by traditional methods.11 Although the number of proteins detected in these studies represents a relatively small proportion of the predicted proteome, many genes may be induced and only expressed under certain conditions.12 A common feature of many studies11,13 is simplification of the proteome by fractionation of cellular constituents into organelles or division into extracts of differential solubility using detergents and chaotropes representative of the cytoplasmic (or cytosolic), cell envelope and outer membrane protein (OMPs) fractions.14 Detection of outer membrane proteins is particularly desirable for discovery of biomarkers for MAR since these include certain channel proteins, such as TolC, linked to efflux pump proteins, and porins which are intimately involved in molecular permeation and multiple antibiotic resistance.5,15 Specific extraction and solubilization procedures have been developed and successfully applied to the analysis of OMPs from E. coli.16 Other potentially important proteins associated with the MAR phenotype are found in the inner membrane and periplasm, also termed the cell envelope; these include AcrB an inner membrane drug anti-porter which is joined to AcrA a periplasm fusion linker between the former and an OMP channel protein, such as TolC17. The reduced solubility of cell envelope proteins has similarly been exploited to provide for their separation from cytosolic proteins to yield cell envelope18 and OMP preparations.14 The objective of the present investigation was to develop simple procedures for the characterization of the proteome of Salmonella Typhimurium. Criteria for effective proteome characterization depend on intended application and in the present context, this may include detection of proteins previously associated with MAR, those of the outer membrane which may include key protein effectors not associated with MAR and a diverse representation of other proteins.19 Assessment of assay sensitivity and reproducibility also provide essential validation objectives for assay development. Subsequent application of these methods will enable the selection of protein biomarkers for biological activity, such as those for multiple antibiotic resistance.
Methods Materials. Sequence grade trypsin was obtained from Promega (Southampton, UK), acetic acid from Fluka (Poole, UK), acetonitrile (HPLC grade) from Merck (Lutterworth, Leicester, UK) and all other chemicals were purchased from Sigma (Poole, UK). PicoFrit capillary HPLC columns (75 µm id; 15 µm orifice) were procured from New Objectives Inc. (Woburn, MA) and packed with Jupiter Proteo reverse phase material (3-4 µm particle) from Phenomenex (Macclesfield, UK); a BioBasic strong cation exchange HPLC column (2.1 × 100 mm; 5 µm particle) was from ThermoHypersil (Hemel Hempstead, UK). Microbial Culture. Starter cultures were prepared by inoculation of LB broth (without glucose; 20 mL; pH 7.2) with Salmonella enterica serovar Typhimurium (strain SL1344) and incubated overnight at 37 °C. Larger working cultures (n ) 4) for proteome extraction were initiated by inoculation of LB broth (100 mL) with an aliquot (2.5 mL) of starter culture and 596
Journal of Proteome Research • Vol. 3, No. 3, 2004
Coldham and Woodward
incubated at 37 °C in 350 mL conical flasks with orbital agitation at 150 rpm. Bacteria were harvested during the logarithmic growth phase when the optical density (λ 600 nm) reached 0.5 by cooling the cultures on ice for 20 min and collection of cells by centrifugation at 1000 × g for 10 min at 4 °C. The bacterial cells were washed by suspension in chilled phosphate buffered saline (PBS; 100 mM, pH 7.2) and collected by centrifugation (1000 × g; 10 min, 4 °C). The number of viable bacteria was estimated from the number of colonies found on LB (without glucose) agar plates after incubation of broth dilutions (n ) 3, 10-7 to 10-9) for 24 h at 37 °C. Extraction of Cytosolic Proteins. Bacterial cell pellets were suspended in PBS (10 mL) containing PMSF (100 µM) and lysed by 6 cycles of probe sonication (Vibra Cell, Sonics and Materials Inc., Banbury, CT, USA) at full power for 30 s on ice. Cell debris was removed from the sonicate by centrifugation at 300 × g and the supernatant retained. A low speed cytosol was produced from the sonicate supernatant by centrifugation at 32 000 × g for 30 min. The cytosol extract was desalted by dilution with ammonium bicarbonate (50 mM; pH 8.0) and concentrated to 0.5 mL by centrifugation in 5000 Da molecular weight cutoff (MWCO) Vivaspin centrifugal filters (VivaScience, Hanover, Germany). Cytosolic proteins were finally heat denatured by incubation at 100 °C for 5 min. Extraction of Cell Envelope Proteins. The efficacy of two buffers on the extraction of cell envelope proteins was investigated. Cell envelope proteins were prepared in duplicate by extraction of the cytosolic pellets with either buffer A (9M urea, 2% v/v Triton ×100 and 2% v/v Pharmalytes 3-10) or buffer B (5M urea, 2M thiourea, 2% v/v CHAPS and 0.5% v/v Pharmalytes) both 3 mL. Proteins were extracted by vortex mixing with buffer A or B, centrifuged at 32 000 × g and the supernatants containing the cell envelope proteins sampled to fresh tubes. The extracts were desalted by dilution with ammonium bicarbonate (50 mM pH 8.0) and concentrated to 0.5 mL by centrifugation in 5000 Da molecular weight cutoff (MWCO) filters. Heat denaturation of proteins in this fraction was not attempted due to the likelihood of carbamylation in the presence of high residual urea concentrations. Extraction of the Outer Membrane Proteins. Extracts of OMPs were prepared in duplicate from each insoluble cell envelope pellet by extraction with a single buffer. The pellets were first washed with PBS (10 mL) to remove excess urea (this may cause carbamylation during heat denaturation) and the insoluble OMPs collected by centrifugation at 32 000 × g. The pellets were suspended in 50 mM Tris (0.5 mL; pH 7.8) containing 10% glycerol, 0.2% SDS and 10 mM DTT and heat denatured by incubation at 100 °C for 5 min. Analysis of Extract Protein Concentration. The protein concentration in the microbial cell extracts was determined with the Bradford reagent. The extracts were diluted 10-fold and duplicate aliquots (25-100 µL) sampled to 2 mL of Bradford reagent. The concentration of protein in the cytosol, cell envelope, and OMP extracts was determined by interpolation from a calibration curve of bovine serum albumin prepared over the concentration range 0.025 to 1 mg/mL. Trypsin Digestion. Cytosolic, cell envelope (200 µg) and outer membrane (100 µg) protein extracts were sequentially reduced and carbamidomethylated by incubation with 10 mM DTT and 40 mM iodoacetamide respectively for 30 min at room temperature, diluted to 200 µL in ammonium bicarbonate (50 mM; pH 8.0) and incubated overnight with 5 µg of sequencing
research articles
Salmonella Typhimurium Proteome
Table 1. Summary Data for the Protein Concentration, Number of Proteins Detected and Number of Matched Proteins in Replicates of Different Cell Extracts of the S. Typhimurium Proteomea
fraction
N
protein concentration (mg; mean ( 1 s.d.)
total (buffer A extraction) cytosol cell envelope buffer A cell envelope buffer B OMP post buffer A OMP post buffer B
2 3 2 2 2 2
10.47 ( 0.09 7.00 ( 0.21 2.12 ( 0.03 5.85 ( 0.35 1.35 ( 0.03 1.13 ( 0.04
no. of proteins detected (mean + 1 s.d.)
no. of matched proteins
correlation coefficient between matched proteins (r2) and statistical significance
816 ( 11 371 ( 33 565 ( 15 251 ( 18 262 ( 5 165 ( 22
621 246 425 200 187 121
0.76; P < 0.0001 0.59; P < 0.0001 0.71; P < 0.0001 0.60; P < 0.0001 0.55; P < 0.0001 0.50; P < 0.0001
a Three cellular extracts representative of the cytosol, cell envelope and outer membrane proteins (OMPs) were prepared and replicates analyzed by 2-dimensional HPLC-MSn. The cell envelope was separately extracted with two different buffers (A & B) to also yield two different OMP preparations for analysis. Correlation coefficients and statistical significance are provided for comparison of the rank listing of individual proteins between duplicate extractions of the same cell preparation.
grade trypsin in Eppendorf tubes (0.5 mL). Digestion was terminated by the addition of 1 µL of formic acid. SCX Chromatography (1st Dimension). Tryptic digests were centrifuged (5000 × g, 1 min) to remove insoluble material and the supernatant sampled to 250 µL HPLC vial inserts. Tryptic peptides (100 µL of digest) were chromatographed using a Hewlett-Packard 1050 system on a BioBasic SCX HPLC column. Analytes were eluted at a flow rate of 0.3 mL/min with mobile phases comprising 75:25 2.5 mM ammonium acetate: acetonitrile pH 4.5 (A) and 75:25 250 mM ammonium acetate: acetonitrile pH 4.5 (B) and a binary gradient (t ) 0 min, A 100%; t ) 5 min, A 100%; t ) 18 min, 65% A; t ) 20 min, B 100%; t ) 22 min, A 100%; t ) 32 min, A 100%). The optical density of HPLC effluent was recorded at 280 nm and 20 fractions of 1 min duration were collected between 5 and 25 min. The SCX fractions were taken to dryness under reduced pressure at 50° C using a centrifugal concentrator (Gyrovap GT, Howe, UK). Reversed Phase Chromatography (2nd Dimension) and Mass Analysis. SCX fractions were dissolved in 0.1% v/v formic (20 µL) and analyzed by capillary-HPLC-MSn using a Famos autosampler (Dionex, Camberely, UK) a Surveyor HLPC pump equipped with an Accurate flow splitter (100 to 1) and LCQ ion-trap mass spectrometer (ThermoFinnigan, Hemel Hempstead, UK). SCX fractions (5 µL) were chromatographed on a PicoFrit column and peptides eluted at an estimated flow rate of 400 nL/min with a linear binary gradient of 0.1% v/v formic acid (A) and acetonitrile (B) (t ) 0 min, A 100%; t ) 40 min, 60% A; t ) 50 min, 40% A; t ) 52 min, 100% A, t ) 75 min, 100% A and next injection. The PicoFrit column (70 mm) was mounted in a steel filter (0.5 µm pore size) assembly to which the electrospray ionization potential (2.5 KV) was applied. The Picofrit tip was positioned slightly off axis within 3 mm of the LCQ mass spectrometer heated capillary inlet using a Protana nanospray source. Mass data for each SCX fraction was collected over the mass to charge (m/z) range 300-2000 using a ‘Big Three’ acquisition method with data dependent product ion scanning of 1st, 2nd, and 3rd most abundant ions above a threshold trigger of 3 × 105 counts per second. The mass isolation window and collision energy were set to 4 amu and 35%, respectively. The mass spectrometer was tuned to the doubly charged ion (m/z 820.5) derived from the synthetic peptide FNPGELLPEAAGPTQV (10 µg/mL) by static infusion at 200 nl/min through a 15 µm orifice PicoTip (New Objectives). Bioinformatic Data Analysis. The SEQUEST algorithm embedded within the Bioworks (ver 3.1) software package was used for the identification of proteins from tryptic peptide mass spectra. Proteins were identified by comparison of tryptic peptide product ion mass spectra against those generated
from the S. Typhimurium LT2.FASTA database derived from the National Center for Biotechnology Information (ftp:// ftp.ncbi.nih.gov/genomes/Bacteria/Salmonella_typhimurium _LT2/NC_003197.faa). The multi-consensus report function was used to assign tryptic peptides to individual proteins and compile rank listings of the proteomes. The default TurboSEQUEST search parameters included selection of trypsin with up to two missed tryptic cleavage sites, static and variable mass modifications of +57 and +16 for carbamidomethlyation of cysteine and oxidation of methionine residues, respectively. Identified proteins were ranked in ascending order according to consensus scores and false positive identifications minimized by filtration against 4 of the 5 following criteria XCorr > 2.0, DeltCn > 0.2, Sp > 400, rsp < 5, ions>30%.20,21 Further, where appropriate, protein identifications were checked manually to provide for a false positive rate of
