Detection of Staphylococcal Enterotoxin B via Biomolecular Interaction Analysis Mass Spectrometry

Detection of Staphylococcus enterotoxin B (SEB) by biomolecularinteraction analysis mass spectrometry (BIA/MS) is presentedin this work. The BIA/MS experiments were based on a surfaceplasmon resonance (SPR) MS immunoassay that detects affinity-capturedSEB both via SPR and by means of exact and direct mass measurementby matrix-assisted laser desorption/ionization time-of-flight(MALDI-TOF) mass spectrometry. Experiments were performed withstandard samples and food samples to assess the BIA/MS limitof detection for SEB and to set the experimental parametersfor proper quantitation. Single and double SPR referencing wasperformed to accurately estimate the amount of the bound toxin.Reproducible detection of 1 ng of SEB per ml, correspondingto affinity capture and MS analysis of $~$ 500 amol of SEB, wasreadily achieved from both the standard and mushroom samples.A certain amount of SEB degradation was indicated by the signalsin the mass spectra. The combination of MS with SPR-based methodsof detection creates a unique approach capable of quantifyingand qualitatively analyzing protein toxins from pathogenic organisms.

INTRODUCTION

Top

Introduction

Detection of biological warfare agents has become a matter ofgreat concern in the last several years. An environmental exposureto even a subtoxic dose of certain biological agents can leadto serious outcomes, due to the amplification of toxicity viain vivo replication in the human host. The classical approachto pathogen detection involves sampling and growth in suitablemedia so that higher concentrations of the microorganisms areobtained for subsequent biochemical evaluation. However, ina post-biological incident environment, direct-reading methodsand instruments are needed for rapid monitoring and detectionof microbial pathogens and their toxins at very low concentration.Furthermore, the detection should be unambiguous and be ableto distinguish between pathogenic and similar nonpathogenicmicroorganisms.

One approach to detection involves recognition of protein phenotypescharacteristic of the specific pathogenic microorganisms. Thisis commonly achieved by immunoassays that utilize antibodiesto specific protein antigens or toxins. The immunoassays aremost frequently performed in an enzyme-linked immunosorbentassay format, although others have also been developed, includingassays based on strip tests and surface plasmon resonance (SPR),piezoelectric, fluorescence, chemiluminometric, and electrochemicaldetection. The immunoassays either involve indirect detectionby optically active reporters that bind to the antibody-retrievedantigen (i.e., amplification detection approach) or use theelectrochemical and optical attributes of the material to whichthe antibodies are immobilized to detect the bound antigen directly(as in surface plasmon resonance). In either case, quantitationof the targeted antigen is readily achieved. However, none ofthe above-mentioned immunoassay approaches is capable of deliveringqualitative (structural) information on the targeted antigens,and they can still suffer from issues such as nonspecific binding,which can lead to false-positive results. Since each proteinpathogen or antigen has a distinct molecular mass and structure,there is a clear advantage in designing and using immunoassaysthat have the ability to delineate the molecular mass of theimmunoassayed antigens.

Over the past several years, we have developed several technologiesthat combine immunoassays (e.g., affinity capture) with matrix-assistedlaser desorption/ionization time-of-flight (MALDI-TOF) massspectrometry (MS) (2, 4, 8, 26) for determination of the molecularmass of biomolecules. In one such approach, termed biomolecularinteraction analysis MS (BIA/MS) (9, 14, 15, 18, 19), the affinity-interactionanalysis between the surface immobilized ligands (e.g., antibodies)and analytes (e.g., proteins) in solutions is monitored andquantified by SPR. SPR is a label-free quantification methodthat utilizes an interaction of light photons with free electrons(surface plasmons) on a gold surface to quantify the changesin concentration or amount of biomaterial on the surface (1,6, 10, 12, 20). The SPR detection itself is nondestructive;therefore, the affinity capture-retrieved analyte(s) can befurther analyzed (from the same surface where they were captured)via MALDI-TOF MS, yielding the molecular mass of the analyte(s).Signals from other proteins, specifically or nonspecificallyretained on the ligand surface, can be also be detected, indicatingpossible binding of analyte variants, protein complexes, ornonspecific binding.

In this work, we investigated the detection of Staphylococcusenterotoxin B (SEB) via BIA/MS. SEB is one of the several toxinson the National Institute for Allergy and Infectious DiseasesBiodefense Priority Pathogens List for which rapid and sensitivemethods of detection are needed. SEB is produced by Staphylococcusaureus, and belongs to a family of heat-stable enterotoxinsthat includes eight other enterotoxins with 50 to 85% sequencehomology (27). The amount of enterotoxin needed to cause intoxicationis very small ( $~$ 1 ng/ml), and hence sensitive and specific detectionis essential. In a previous study we showed the general applicabilityof the BIA/MS approach in detection of staphylococcal toxinsin food (17). Presented here are experiments performed withstandard samples and food samples to assess the limit of detectionof BIA/MS for SEB and to set the experimental parameters forproper quantitation.

MATERIALS AND METHODS

Top

Materials and Methods

Proteins and antibodies.
Affinity-purified rabbit anti-SEB antibody (1 mg/ml, in 10 mMphosphate buffer [pH 7.5] [Toxin Technology Inc., Sarasota,Fla.]) and SEB (1 mg/ml in water [Sigma, St. Louis, Mo]) wereobtained from Avraham Rasooly (Center for Food Safety and AppliedNutrition, U.S. Food and Drug Administration, College Park,Md.).

Samples.
Solutions with decreasing SEB concentrations were prepared byserial dilution of the SEB stock with HBS-EP buffer (0.01 MHEPES [pH 7.4], 0.15 M NaCl, 3 mM EDTA, 0.005% surfactant P20).The mushroom sample was prepared by homogenizing the whole content(solid and brine) of a 7-oz can of commercial mushrooms. Thehomogenate was centrifuged for 5 min at 13,000 rpm (10,000 xg), resulting in $~$ 65% (vol/vol) clarified supernatant (mushroomstock), which was mixed in a 1:1 ratio with an appropriate SEBstock solution (at 20 and 2 ng/ml) to yield SEB-mushroom sampleswith 10 and 1 ng of SEB per ml, respectively.

SPR analysis.
A Biacore X instrument (Biacore AB, Uppsala, Sweden) was usedfor the first dimension of BIA/MS (affinity retrieval and SPRquantification). CM5 research grade sensor chips (carboxymethyldextran-derivatizedsurface [Biacore AB]) were used in the experiments, with HBS-EPrunning buffer at a flow rate of 5µl/min. The antibodywas immobilized on sections (flow cells, FC) on the chip surfaceby a standard EDC-NHS [N-ethyl-N'-(dimethylaminopropyl)carbodiimide-N-hydroxysuccinimide]coupling method (7). Regeneration of the antibody surfaces wasachieved by short injections of 0.06 N HCl.

MALDI-TOF MS analysts.
Following removal from the biosensor, chips were washed withthree 200-µl aliquots of distilled water and preparedfor MS by application of a MALDI matrix (aqueous solution of ${alpha}$ -cyano-4-hydroxycinnamic acid [ACCA] in 33% [vol/vol] acetonitrile-0.4%[vol/vol] trifluoroacetic acid) with a matrix aerosol applicationdevice (16). The MS analysis was performed on a homemade MALDI-TOFmass spectrometer (18).

RESULTS AND DISCUSSION

Top

Results and Discussion

Following the immobilization of $~$ 90 fmol of anti-SEB in FC1 and $~$ 180 fmol of anti-TSST-1 (toxic shock syndrome toxin 1) in FC2(which served as a reference FC), several aliquots of SEB sampleswere injected over both FCs. Shown in Fig. 1a is a sensorgramwith four consecutive, 100-µl injections of 10 ng of SEBper ml in HBS-EP buffer, 1 ng of SEB per ml in HBS-EP buffer,HBS-EP buffer only, and 1 ng of SEB per ml in HBS-EP bufferagain. The sensorgrams shown were baseline corrected in realtime by subtraction of the signal from the reference (anti-TSST-1)FC2 (i.e., the FC1 - FC2 signal is shown in Fig. 1). Even withthis correction, a constant positive bulk refractive index change(positive baseline drift) is noticeable in the sensorgram. Figure1b shows the overlay of the four response curves, clearly indicatingthe positive SPR response resulting from the buffer injection.To correctly assess the responses from the SEB injections, thebuffer injection curve was subtracted from the SEB curves. Fromthese doubly corrected binding curves (which are shown in Fig.1c), the true SPR responses resulting from the injections ofthe 10- and 1-ng/ml SEB aliquots were determined to be 145 and15 response units (RU), respectively. The two 1-ng/ml SEB bindingcurves superimpose well, indicating the reproducibility of theexperiments. The 15-RU response is indicative of binding of $~$ 500 amol of SEB (1 RU = 1 pg of protein; MW_SEB = 28,366). Themass spectrum taken from the surface of FC1 following the lastinjection of the 1-ng/ml SEB aliquot is shown in Fig. 1d. Multiply(singly, doubly, and triply) charged SEB signals are observedin the spectrum, confirming the binding of SEB in the flow cell.The signals are somewhat broader, suggesting the existence ofmultiple SEB forms.

View larger version (31K):
[in this window]
[in a new window]
FIG. 1. (a) SPR sensorgrams (baseline corrected from an anti-TTST-1-derivatized reference FC) resulting from four consecutive 100-µl injections of 10 ng of SEB per ml in HBS-EP buffer, 1 ng of SEB per ml in HBS-EP, HBS-EP buffer, and 1 ng of SEB per ml in HBS-EP. (b) Overlay view of the four binding curves. (c) SEB binding curves corrected by subtraction of the buffer injection sensorgram. (d) MALDI-TOF mass spectrum taken from the surface of FC1 following the last injection of the 1-ng/ml SEB aliquot.

To demonstrate the same level of SEB detection from a food sample,100-µl aliquots of mushroom extract samples containing10 or 1 ng of SEB per ml were injected over the surface of ananti-SEB-derivatized FC on a new CM5 sensor chip (Fig. 2a).Also shown and overlaid in the same figure is an injection ofa mushroom sample (SEB-free) to determine the amount of non-specificbinding to the active/antibody surface. As done in the precedingexperiment, the mushroom binding curve was subtracted from thetwo SEB-mushroom sample curves, yielding the sensorgrams shownin Fig. 2b. The readings taken from these singly corrected bindingcurves (130 and 19 RU) indicate binding of 130 and 19 pg ofmaterial from the injections of the 10- and 1-ng/ml SEB-mushroomsample aliquots, respectively. These responses are similar tothose observed in Fig. 1c, suggesting the retrieval of subfemtomoleamounts of SEB from the 1-ng/ml sample injection. After regenerationof the antibody surface, another aliquot of the 1-ng/ml SEB-mushroomsample was injected over the anti-SEB-derivatized flow celland the chip was undocked from the biosensor and prepared forMS. The resulting mass spectrum is shown in Fig. 2c. Signalsfrom the singly, doubly, and triply charged ions of SEB dominatethe spectrum, with very few other background signals. The SEBsignals are noticeably sharper and stronger than those observedin Fig. 1d. The peak broadening in Fig. 1d can be attributedto SEB degradation during storage; the experiments shown inFig. 1 were performed after the SEB sample had been stored at4°C for over 2 years. The existence of degradation productswas verified by direct MALDI-TOF MS analysis of the SEB stocksample, which revealed the presence of several truncated SEBforms (containing residues 5 to 239, 5 to 238, 6 to 238, 7 to238, 8 to 238 [or 7 to 237], and 8 to 237 [or 7 to 236]; nativeSEB contains 239 residues). This SEB degradation most probablycaused the diminished signal intensity in the mass spectrumshown in Fig. 1d (i.e., instead of one strong signal, multipleweaker signals are observed). The detection of these SEB isoformsdemonstrates the ability of BIA/MS to delineate the existenceof potentially significant posttranslational modification thatcan go undetected if SPR-only analysis is performed.

View larger version (60K):
[in this window]
[in a new window]
FIG. 2. (a) Overlaid SPR sensorgrams resulting from injections of 10 and 1 ng of SEB-mushroom samples per ml and a mushroom sample (SEB free) over the surface of an anti-SEB-derivatized flow cell. (b) SEB-mushroom binding curves corrected via subtraction of the mushroom sample injection sensorgram. (c) MALDI-TOF mass spectrum taken from the surface of FC1 following an injection of 1 ng of SEB-mushroom sample per ml.

Overall, detection of 1 ng of SEB per ml in buffer and in themushroom sample was readily achieved in the BIA/MS experiments.SEB signals were observed in the MALDI-TOF mass spectra fromall three experiments in which SEB at 1 ng/ml was analyzed,either in buffer or in a food matrix. We have demonstrated thesame level of detection for another protein, ß₂-microglobulin(15). Further optimization (i.e., increased sample volume ordecreased flow rates) may lower the SEB limit of detection (LOD)below the 1-ng/ml level. Nevertheless, this limit of detectionis on par with (or slightly better than) those reported by authorsusing other SPR-based methods. Detection of $~$ 5 ng of SEB perml was achieved by newly developed miniature fiberoptic (25)and wavelength modulation-based (5) SPR sensors. Another prototypeminiature SPR sensor was capable of detecting $~$ 15 ng of SEB perml in buffer and urine (13). When SPR was used in the indirectmode of detection (i.e., with signal amplification via bindingof secondary antibodies to the affinity capture-retrieved antigens),the LOD was lowered to 0.5 ng/ml with one amplification step(5, 13) and 3 pg/ml (SEB in buffer) with two amplification steps(13). Other (non-SPR) approaches to SEB detection include traditionalELISAs (11, 24) (LOD = 0.05 to 1.0 ng/ml), time-resolved fluorometry(22) (LOD = 4 to 20 pg/ml [SEB in buffer]), and fluorescence-basedarray sensors (3, 23) (LOD = 4 ng/ml), and all of them use somekind of signal amplification. However, what separates the MS-basedimmunoassays and BIA/MS from the rest of the immunoassays isthe ability to detect proteins exactly by means of direct massmeasurement. With the use of polyclonal antibodies, multipleforms of the toxins can be captured and analyzed. In such experiments,cross-reactivities with proteins from similar nonpathogenicmicroorganisms (21) can readily be delineated form the signalsin the mass spectrum (but not from the SPR signal itself) andeliminated from the data analysis. Also implied in the analysisis the detection of mutant forms of the protein that might be(potentially) genetically engineered to escape detection bymonoclonal antibody-based (i.e., monospecific) assays. BIA/MSoffers a unique capability to phenotypically trace toxins frompathogenic organisms and can serve as a high-performance technologyfor detecting and quantifying biodefense agents present in potentialtarget media of terrorist attacks.

ACKNOWLEDGMENTS

We thank Avraham Rasooly from FDA/CFSAN for providing the proteinsand antibodies.

This publication was supported in part by grant 2 R44 CA82079-03A1from the National Institutes of Health.

The contents of this paper are solely the responsibility ofthe authors and do not necessarily represent the official viewsof the National Institutes of Health.

FOOTNOTES

* Corresponding author. Mailing address: Intrinsic Bioprobes Inc., 625 S. Smith Rd., Suite 22, Tempe, AZ 85281. Phone: (480) 804-1778. Fax: (480) 804-0778. E-mail: dnedelkov{at}intrinsicbio.com.

REFERENCES

Top