Peter Snyder, Ph.D., and Rabih E. Jabbour, Ph.D.
A white paper prepared for the June 25–26, 2013, workshop on Strategies for Cost-Effective and Flexible Biodetection Systems That Ensure Timely and Accurate Information for Public Health Officials, hosted by the Institute of Medicine’s Board on Health Sciences Policy and the National Research Council’s Board on Life Sciences. The authors are responsible for the content of this article, which does not necessarily represent the views of the Institute of Medicine or the National Research Council.
Mass spectrometry (MS) is being considered as a candidate for the Tier 1, Tier 2, and Tier 3 autonomous detection systems. Candidate status depends on how the figures of merit compare to the given desirable characteristics. Samples to be analyzed include a pathogenic and simulant spore former and a pathogen and simulant vegetative cell bacterium.
The mass spectrometer is an instrument designed to separate gas-phase ions according to their mass-to-charge ratio, m/z. The heart of the mass spectrometer is the analyzer. This element separates the gas-phase ions. The analyzer usually uses time-of-flight, quadrupole, ion trap, or a
combination of all three to move the ions from the region where they are created to a detector, where they produce a signal that is amplified. The m/z, and not only the mass, is of importance. A mass spectrometer can measure the mass of a molecule only after it converts the neutral molecule to a gas-phase ion. To do so, it imparts an electrical charge to molecules and converts the resultant flux of electrically charged ions into a proportional electrical current that a data system converts to digital information, displaying it as a mass spectrum.
Ions can be created in a number of ways:
• Laser ablation of a compound dissolved in a matrix on a planar surface by matrix-assisted, laser-desorption ionization (MALDI).
• Interaction with an energized particle or electron such as in electron ionization.
• Electrospray ionization (ESI) where the eluent from a liquid chromatography (LC) system receives a high voltage, resulting in an aerosol of ions.
ESI is useful in producing ions from neutral macromolecules, because it overcomes the propensity of these molecules to fragment when ionized. ESI is advantageous over other atmospheric pressure ionization processes (e.g., MALDI) because it may produce multiply charged ions, effectively extending the mass range of the analyzer to accommodate the kilodalton to megadalton orders of magnitude observed in proteins and their associated polypeptide fragments.
ESI is a so-called soft ionization technique since there is very little fragmentation of the parent compound. This can be advantageous in the sense that the molecular ion (or, more accurately, a pseudo-molecular ion) is almost always observed; however, very little structural information can be gained from the simple mass spectrum obtained. This disadvantage can be overcome by coupling ESI with tandem mass spectrometry (MS-MS).
The development of ESI for the analysis of biological macromolecules was rewarded with the Nobel Prize in chemistry to John Bennett Fenn in 2002.
The analyzer is operated under high vacuum, so that the ions can travel to the detector with a sufficient yield. MS-MS is the combination of two or more MS experiments. The aim is either to get structural information by fragmenting the ions isolated during the first experiment or to achieve better selectivity and sensitivity for quantitative analysis. MS-
MS is done either by coupling multiple analyzers (of the same kind or different kinds) or with an ion trap, by doing the various experiments within the trap.
The unit of measure has become the dalton, displacing other terms such as amu. One dalton is one-twelfth of the mass of a single atom of carbon-12 (12C).
Liquid chromatography (LC) technology gives analytical access to about 80 percent of the chemical universe that is unreachable by GC. In its simplest form, LC relies on the ability to predict and reproduce competing interactions between analytes in solution (the mobile or condensed phase) being passed over a bed of packed particles (the stationary phase). The development in recent years of columns packed with a variety of functional moieties and of solvent delivery systems able to deliver the mobile phase has enabled LC to become the analytical backbone for many industries. The abbreviation HPLC for high-pressure liquid chromatography was coined in 1970 by Csaba Horváth to indicate that high pressure is used to generate the flow required for LC in packed columns. The technique is now generally referred to as high-performance liquid chromatography, with the same abbreviation being used. HPLC is an important separation technique for the analysis of proteins and peptides because it can easily be coupled to a mass spectrometer. Moreover, the compatibility of solvents used in HPLC separations with ESI makes this hyphenated technique most commonly used in the final stage of proteomics analysis.
TIER 1: 2012 TO 2016
Currently there are two MS-based systems that may be viewed as having parameters that come close to satisfying the proposed figures of merit for the autonomous detection system using mass spectrometry. The relevant characteristics include actual system size, performance, and quantitative figures of merit, such as the probability of false-positives approaching one false-positive per year. Three significant differences between the two systems are size, the state of the sample collected, and liquid expendable.
Hamilton Sundstrand Mass Spectrometer
The modified Hamilton Sundstrand mass spectrometer (HSMS) (Basile et al., 1998; Griest et al., 2001; Hart et al., 2000) is fairly small in size, uses a liquid expendable, and collects bulk samples. The system was originally manufactured by Bruker-Franzen in the 1980–1990 time frame and then transferred to Oak Ridge National Laboratories in the late 1990s. It subsequently came into the hands of Hamilton Sundstrand. The system has matured, and the key feature is that the aerosol sample is collected in bulk in a few minutes onto a small filter paper. Microliter amounts of liquid are injected onto the spot. The liquid is the reagent tetramethylammonium hydroxide (TMAH). The spot is then heated (pyrolyzed) in a ballistic fashion to about 500–550°C within 14 seconds and maintained at that temperature for 4 minutes, and TMAH reacts with any fatty acids that are included in the contents of the spot. TMAH reacts and derivatizes the fatty acids, and fatty acid methyl esters (FAMEs) result. Fatty acids are typically found in all bacteria. There are certain fatty acids that are fairly unique to certain bacteria, but the better discriminator is the pattern of approximately 20 to 30 FAMEs that are produced to the exclusion of almost all other material in the spot. Residual char is heated to a higher temperature so as to develop a clean, fresh spot for the next cycle of aerosol collection. Meanwhile the vapors are transported into the ion trap system (MS-MS) for analysis. There is a resident database, and the system can produce an output with the likeliest bacteria present.
It is the 20 to 30 derivatized fatty acids that are the fundamental data used to describe, analyze, and identify the biological sample. It is not yet known if that is enough information to correctly identify all the agents of interest and to exclude all others at the desired probability of detection (Pd) and probability of false-positive (Pfp) levels. Very few Pd and Pfp analyses were performed by the HSMS team. There has been none published.
The HSMS has the following features:
• Vibration tested in a way that simulates the high-mobility multipurpose wheeled vehicle: HMMWV Model 1097 “Humvee.”
• Chemical ionization in addition to electron ionization. Chemical ionization significantly reduces background chemical noise (such as diesel and gasoline vapors).
• The vacuum system operates at about 35 W.
• The HSMS has four separate modules, and each module is further compartmentalized for easy troubleshooting and replacement.
• The entire HSMS is approximately 5.8 cubic feet and draws an average of 500 W power.
• Two-stage virtual impaction system to deliver bioaerosol to the heating area from 300 L/min to 1 L/min with 50–90 percent efficiency for aerosol aerodynamic diameters between 2 and 10 microns.
• Sample deposition, heating, and processing take place in a quartz tube, and the vapors are directed into the MS-MS system.
• The 330-L/min aerosol input flow rate with a 2-minute sampling time and 50 agent-containing particles per liter of air (ACPLA) equates to 33,000 particles collected, which produces S/N = 5. This is laboratory data.
• Outdoor Joint Field Trials-6 (JFT-6) at Defence Research Establishment Suffield (DRES) in Medicine Hat, Alberta, Canada, during August 2000 had Bacillus globigii (BG) spore and Erwinia herbicola aerosol releases at the 30 ACPLA level. S/N of the JFT-6 bacterial FAME spectra were between 7 and 10.
• No real quantitative figures of merit have been done with respect to false-positive (FP) and probability of false-positive (Pfp).
• The four main Category A pathogenic bacteria and numerous spore and vegetative simulants have been performed.
Bioaerosol Mass Spectrometer
The bioaerosol mass spectrometer (BAMS) produces low-end bacterial biochemical information such as basic dipicolinic acid spore information. The BAMS needs a significant reduction in size. However, the system has had a significant degree of success in real-time analysis situations.
Lawrence Livermore National Laboratory (LLNL) documented success in tackling the challenge of detection and identification of bacterial aerosols by time-of-flight mass spectrometry (TOF-MS). Much of the success is due to the multivariate data-analysis methods that delineate the simultaneously captured positive- and negative-ion mass spectra. The impetus for developing a bioaerosol-TOF-MS system originated in the modification of a system based on the analysis of ambient inorganic and
organic aerosols (Liu et al., 1997; Su et al., 2004) naturally found, generated, and released into the environment.
The bioaerosol-TOF-MS system developed at LLNL does not use chemical matrix, liquid, or solid consumables (Russell et al., 2005; Steele et al., 2003). Bioaerosol BG spores were drawn through a sizing region consisting of two lasers. A frequency-quadrupled, Q-switched Nd:YAG ablated, desorbed, and ionized each particle. Both positive and negative ions are scanned because two separate TOF-MS tubes emanate from the ion source at a separation of 180°. Typical mass spectra for a BG spore showed positive ions <150 Da and negative ions <200 Da. This was the first-ever recording of positive- and negative-ion mass spectra from the same biological particle.
The development of the BAMS (Fergenson et al., 2004) saw scrutiny on the concept of analyzing every particle in succession. This is a key point that lends itself to the autonomous detection system using mass spectrometry Pd and Pfp figures of merit. Spores of BG and Bacillus thuringiensis (BT) were used with no liquid reagent. The data analysis starts with positive and negative mass spectra of 350 total masses (elements). The experimental spectra are then compared to a database of organisms. If multiple standards match, then the closest spectrum match (positive and negative ions) was considered as the experimental analyte. Reproducibility depended strongly on laser wavelength and fluence. The wavelength of 266 nm was chosen because it was absorbed by dipicolinic acid (DPA) in the spores. In nature, DPA is only found in Gram-positive bacterial spores. Different bacterial growth media for both spores saw minimal mass spectral differences within each species.
Determination of individual particles was made in real time in two steps. First, a prescreening stage eliminated nonbacterial particles by analysis of the acquired spectra. The microbial-related spectra then went through mass-related criteria to refine and provide a database match to the experimental spectra. BG spores were identified 93.2 percent when compared to BT. Commonly found and commercially available white powders were separately mixed with BG and BT to test the data analysis algorithm for interference properties.
For specificity information, the spores were recognized 91 percent of the time with Gold Bond powder, 86 percent with growth media, 78 percent with Equal sweetener, 56 percent with fungal spores, and 46 percent of the time with Knox gelatin.
Single-Particle Aerosol Mass Spectrometer
The BAMS evolved into the single-particle aerosol mass spectrometer (SPAMS) system (Steele et al., 2008). The suite of biological substances originally investigated was expanded to include chemical, biological, radiological, nuclear, and explosive (CBRNE) materials as well as clandestine and illegal drug substances. The SPAMS system uses three continuous-wave laser beams to produce particle sizing properties. The SPAMS can track up to 10,000 particles per second.
The particle’s position and velocity are used to predict when it passes through subsequent regions of the instrument. After the sizing region, the particle is interrogated by a laser-induced fluorescence (LIF) region to determine the presence of ultraviolet (UV) fluorescence, which indicates a biological nature. If the particle produces UV fluorescence, it is then ionized by a 266-nm laser beam. Identification of a particle occurs by mass spectral pattern matching with a database.
The tested substances produce significantly different experimental mass spectra, and no false identification or false alarms have been observed with sequential challenges of the CBRNE materials. In addition to the analytes, there is a constant background of ambient outdoor particles such that the background particles dominate or are equivalent in temporal signal responses to the particular CBRNE challenge signals.
The SPAMS system was tested at the San Francisco, California, airport. The aerosol collector and particle inlet were cleaned once per week. The ambient atmosphere internal to the airport was sampled every minute, and the spectra were recorded. Approximately 1 million particles were tracked and recorded over a 7-week period. After the recording and storage of the aerosol data, it was analyzed in the laboratory. No real-time analyses or decisions were made in the field. In any 1-minute interrogation, no more than two particles were identified as BG or pentaerythritoltetranitrate (PETN) explosive, and this resulted in zero false alarms because the 2 particles/minute was below the alarm threshold. Thus specificity is excellent.
Here are some figures of merit and comments on SPAMS:
• Disadvantages: Very large size and high cost. However, operating costs are very low.
• For SPAMS, it is essentially a single ACPLA detector and classifier with limited identification.
• SPAMS allows particles in the 0.7- to 10-micron-diameter range to pass into the system.
• Six laser beams size and track the aerosol particles.
• A second series of lasers, if necessary, produces LIF.
• Particles enter into a bipolar TOFMS that can track 50 particles/s.
• The system is autonomous and can track 10,000 particles/s up front. Most difficult when trying to detect nothing when no threat agent present! SPAMS accomplished high sensitivity and low Pfp and is very fast. After release in the lab of certain simulants in the air, it took on average of 34 s for the SPAMS to respond.
A comprehensive set of receiver operating characteristics (ROCs) curves has been performed to detail the Pfp and detection of false-negatives (Pfn) figures of merit (Gard et al., DARPA proposal, 2007 [unpublished]). To the best of our knowledge, there is no published work showing any ROC curves or Pfp and Pd figures of merit. Their ROC curves have only been presented in a DARPA proposal for the BAND program (Gard et al., DARPA proposal, 2007 [unpublished]).
The system can currently detect and identify a limited suite of samples. The time to detection is 1 minute at 1 ACPLA, and the aerosol size range is 1–10 microns in diameter. To detect 1 ACPLA in 1 minute and sampling 100 L/min with sampling efficiency of 10 percent yields 10 agent-containing particles per minute. Alarm conditions are based on the measurement and assignment of more than 1 particle.
The probability of misclassifying a particle as an agent is vanishingly small because no false-positives were observed in 5,000 BG, BT, and background particle cases. The probability of a single BT particle being wrongly identified as BG is 10–3. Concerning the probability for a false alarm, the alarm conditions are based on the measurement and assignment of more than 1 particle. The probability of correctly classifying a particle as an agent can be inferred from the BG spore challenge, in which the system correctly identified the particles 93 percent of the time over BT and other background aerosols. BG remained unidentified in 7 percent of the cases. The BAMS system has a good chance at performing to and meeting most of the autonomous detection system using mass spectrometry technical requirements but with some false-negatives. This assessment can also be considered for organisms in addition to Bacillus subtilis. The main issues for the BAMS and SPAMS are the size and power requirements.
TIER 2: 2016–2020
Tier 2 and Tier 3 systems cannot begin to address the analytical figures of merit that are requested by DHS. Future systems just have not addressed any final or mature numbers because of the very fluid makeup and constitution of the basic research laboratory apparatus. For example, sensitivity and specificity change constantly depending on the hardware, methods, sample handling processes, and data analysis packages used. There is so much change as well as new modifications for the hardware and techniques.
Aerosol Collector Candidates
Possible candidates for aerosol collector are the Micromachined Virtual Impactor Collector and the Micromachined Radial Virtual Impactor Collector from MesoSystems Technology, Inc. However, essentially no sampling technique can ensure that the collected microbial specimen reflects the original state and can be directly used in bioanalysis (Heidelberg et al., 2000; Lee et al., 2004; Pasanen, 2001; Ren et al., 2001).
Pathogen agar plate culturing is generally the necessary step before analysis, mainly because the concentration of the collected pathogen is too low for direct bioanalysis by the methods mentioned above. Because of the size difference, the water used to rinse or wash the samplers is generally too much compared with the small amount of the pathogens collected in the samplers. Consequently, the concentration of the collected pathogens in aqueous media is too low for direct bioanalysis.
There are various possible solutions. Microfluidics, which handles liquid in the micrometer dimension, corresponding to nanoliters in volume, appears to concentrate pathogens in a relatively small amount of liquid (Baoa et al., 2008; Bhagat et al., 2011; Qi et al., 2010; You et al., 2011). Microfluidics is also economical, with much less reagent consumption, suitable for a large-scale deployment and field application (Holmes and Morgan, 2010; Li et al., 2006; Liu, 2010; Park et al., 2011). Cell capture by microfluidic chip has been reported (Jang et al., 2012; Lim et al., 2012; Loutherback et al., 2012; Reisewitz et al., 2010), but most reports focused on tissue cell capture and rarely refer to airborne bacteria cells captured directly from air.
A simple microfluidic device that is capable of fast and efficient airborne bacteria enrichment has recently been reported (Jing et al., 2013).
The initial concentration of E. coli bacteria suspension was 106 cell/mL, and an aerosol generator was used to generate the bioaerosol for 2 minutes. Under the vacuum created by a micropump, the bacteria aerosol is drawn into the channels of the microfluidic chip. At the same time, an LB culture dish is placed next to the microfluidic chip as a parallel control (sedimentation method). Bacteria may be captured through their adhering to the inside walls of the microchannels in the chip. The uncaptured bacteria will pass through the chip and enter the resuspension solution. After enrichment, 2 µL of buffer are loaded into the microchannel to wash the captured bacteria inside the microfluidic chip, and the solution is collected at the outlet for statistical analysis.
When the concentration of E. coli bacteria suspension was 105 cell/mL, there were 254 cells collected by the microfluidic chip, which is more than 4.53 times higher than the 66 cells collected by the agar-plate, direct-from-the-air sedimentation method. When the concentration of E. coli bacteria suspension was 104 cell/mL, there were still 130 bacterial particles collected by the microfluidic chip, which is 4 times higher than the 26 cells collected by the plate sedimentation method.
When the concentration of E. coli bacteria suspension decreased to 103 cells/mL, the microfluidic chip collected 56 bacteria cells, which is 55 times higher. The 130 E. coli bacteria captured by the microfluidic chip were enough for rapid detection methods, such as an ELISA-based test (100 bacteria are enough for ELISA and polymerase chain reaction [PCR]-based tests).
Moreover, the detection limit of the microfluidic device is much lower than that of the agar-plate sedimentation method. It can collect enough bacteria at a low aerosol concentration for a direct ELISA, loop-mediated isothermal amplification (LAMP) test, which is essential for rapid bacteria detection, especially compared with traditional bioaerosol collection techniques that need the downstream culturing or PCR amplification because of the relatively high capture limit.
Bacterial Proteome Analysis
The bacterial proteome represents the collection of functional and structural proteins that are present in the cell. The protein content of the cell represents the majority of the cell dry weight, which makes it an ideal cellular component to be utilized for bacterial characterization (Loferer-Krobacher et al., 1998).
Most of the Category A, B, and C biological threats from the Centers for Disease Control and Prevention (CDC) have their genomes fully sequenced and available for bioinformatics-based proteomics methods.
The predominant MS techniques used for bacterial identification and differentiation include ESI-MS-MS, MALDI-TOF-MS, and one- or two-dimensional sodium dodecylsulfatepolyacrylamide gel electrophoresis (1D or 2D SDS-PAGE).
MS techniques for bacterial identification and differentiation (Kollipara et al., 2011) rely on the comparison of the proteome information generated from either intact protein profiles (top-down) or the product ion mass spectra of digested peptide sequences (bottom-up) analyses (Fox et al., 2002; Pennington et al., 1997; Zhou et al., 2012). The different approaches include
• Top-down from intact proteins: Bacterial differentiation and identification are accomplished through the comparison of the MS data of intact proteins with an experimental mass spectral database (fingerprint spectrum) containing the mass spectral protein masses of the microorganisms (Demirev et al., 1999; Fenselau and Demirev 2001; Jabbour et al., 2005; Pineda et al., 2000).
• Bottom-up around 1.5 kDa: Bacterial differentiation using product ion mass spectral data of peptide sequences from the trypsin-digested proteins is accomplished through the use of search engines against publicly available sequence databases to infer identification (Williams et al., 2002).
• Middle-down: around 5 kDa (Zhou et al., 2012). This is the most mature method for the MS identification of bacterial proteins.
• Shotgun proteomics: Trypsin is used to digest and separate all peptides in an LC system for MS-MS analysis. This is the preferred method because it yields the most data.
• Peptide mass fingerprinting or peptide fingerprinting from proteins: Only for predominately expressed bacterial proteins in a MALDI-MS spectrum. Thus, either intact bacterial mass or protein extract can be trypsin-digested without any purification or separation of the proteins. This approach is usually used to target certain proteins that are overexpressed in a bacterium. Then one can compare a theoretical table of peptide masses with experimental masses; small acid-soluble proteins (SASPs) are a good example (Castanha et al., 2006; Demirev and Fenselau, 2008).
Bacterial Protein Processing
This is a critical step given that the protein portion must be isolated if MS is to be useful in the analysis of bacterial aerosols. In general, proteins isolated from lysed bacterial cells will contain constituents detrimental to their isolation, such as lipids, nucleic acids, and polysaccharides. Unfortunately, the presence of buffers, chaotropes, detergents, or cocktails of proteinase inhibitors, which are usually added to aid in protein extraction and to preserve the integrity of a proteome, may interfere with further processing and analysis of proteins. Therefore, they have to be removed from the sample before introducing the sample into a mass spectrometer (Wisniewski et al., 2009).
Jabbour et al. (2011) devised a “one-pot protein mixture purification avenue” that removes the extraneous background milieu low-molecular-weight impurities. The conventional in-solution digestion of the protein contents of bacteria is compared to a small disposable filter unit that is placed inside a centrifuge vial for the processing and digestion of bacterial proteins. Each processing stage allows the filtration of excess reactants and unwanted byproducts while retaining the proteins. Upon the addition of trypsin, the peptide mixture solution is passed through the filter while retaining the trypsin enzyme. This can be replaced with a solid-state trypsin system.
Micro-Total Analysis Systems: Raw Sample Refinement.
A micro-total analysis system (µTAS) can include bacterial background cleanup, lysis (Andersson and van den Berg, 2007), protein purification, and protein digestion where the created peptides are introduced into an HPLC column for separation. This can be thought of as a miniature “one-pot protein mixture purification avenue.” A µTAS is a microfluidic device for sample processing using minimal reagents and water buffer (Dittich et al., 2006). It is characterized by a microfabricated “lab on a chip” and pressurized to allow the flow of liquid through the system. Biochips have been manufactured to include micro-LC separation columns and ESI. They can be characterized as either separate (modular) sections or monolithic (one injection-molded piece) designs. The peptide digest can be injected into a micro-LC column.
Proteomics on a Chip
Significant efforts over the past decades have been focused on the development of “proteomics on a chip” in an attempt to incorporate the various components necessary for analytical operations onto a single cost-effective platform (Feng et al., 2009; Freire and Wheeler, 2006; Henion, 2009; Huikko et al., 2003; Lion et al., 2003; Liu et al., 2006; Ma et al., 2009; Sedgwick et al., 2008; Szita et al., 2010; Tian et al., 2011; Wang et al., 2000).
Microfluidic Proteomic Reactor Performance
A comparison of the performance of the microfluidic proteomic reactor to the conventional proteomic reactor was carried out by Ethier et al. (2006). A standard protein, BSA, was processed and analyzed by a nano-HPLC-MS-MS system. All of the peptide peaks that originated from BSA were labeled on a base peak chromatogram. This result demonstrates that (1) the protein sample is digested efficiently and that (2) the proteomic reactor on a polymeric chip does not contaminate the HPLC-ESI-MS-MS with residual chemical from the polymers.
Three general processing procedures are used to generate protein ions for subsequent characterization and identification of bacteria with MALDI-MS.
The simplest uses a mixture of the bacterial sample with a matrix deposited onto a metal MALDI target (Bright et al., 2002; Demirev et al., 2001; Gantt et al., 1999; Hettick et al., 2004; Lay, 2001; Lee et al., 2002; Wang et al., 1998). In this whole-cell analysis procedure, any deliberate steps to break open or fragment the exterior cell walls are avoided in part due to logistic and procedural concerns.
The second method consists of suspending whole cells in a solvent to solubilize or extract protein species from the bacterial sample (Amado et al., 1997; Williams et al., 2003). Different types and amounts of proteins are extracted depending on the polarity of the solvent and the presence of additives (Dickinson et al., 2004). A portion of the protein extract is mixed with organic liquid matrix and analyzed by laser desorption/ionization MS. The efficiency of protein extraction was investigated
with respect to different solvents (Domin et al., 1999; Madonna et al., 2000; Ruelle et al., 2004), pH, salt content, detergent (Li et al., 1997), and other additives.
The third method uses lysis techniques (Bright et al., 2002; Demirev et al., 2001; Halden et al., 2005; Krishnamurthy et al., 1996; Lay, 2001; Owen et al., 1999; Williams et al., 2003) to deliberately break open or fragment the bacterial cell, and this allows straightforward solvent extraction of cellular proteins. A protein extract is mixed with a matrix and the mixture is analyzed by MALDI-MS.
MALDI vs. LC-ESI
MALDI-MS is used extensively in clinical and microbiological laboratories primarily for differentiation and characterization objectives for bacteria (Dworzanski and Snyder, 2005). Only single pure bacterial samples from agar-plate colonies have been performed. With ESI, mixed bacterial samples and metaproteomics (protein analysis from a mixture of many bacteria) have been performed. Overall, relatively few protein markers desorb from bacteria in MALDI. Very little detail is resident in the MALDI mass spectral signals compared to LC-ESI-ion trap MS-MS. LC-ESI yields orders of magnitude greater mass signals (Yates, 2004; Yates et al., 2009) for sophisticated data analysis techniques.
However, the number of steps needed to process bacteria is greatly reduced for MALDI compared to LC-ESI. The water expendable is not necessary for MALDI, and it is required for LC-ESI. An organic unsaturated/aromatic ring substance is essential for MALDI, while water buffer solutions are required for LC-ESI. Even if a sample is of a complex nature, both MALDI and LC-ESI benefit from a set of sample purification stages. Both necessarily require some form of sample cleanup. MALDI procedures also have used a laser impinging directly on a bacterial colony taken from a growth dish.
MALDI data analysis is based on patterns of mass spectral peaks that are compared to a database of spectral replicates. Thus the reference database must be derived from experiments. Discrimination is usually based on 10 to 20 peaks as data input (Lau et al., 2012; Liu et al., 2007). LC-ESI bacterial analysis is based on a comparison between the experimentally derived peptide/protein analysis and the protein translations from genome data banks. Discrimination is usually based on many hundreds to thousands of peaks as data input.
In terms of “bang for the buck,” LC-ESI provides orders of magnitude more raw data than MALDI, and much more sophisticated data analysis and reduction packages exist for LC-ESI data than for MALDI data.
Peptide Analysis for Bacterial Identification
The peptide mixture from the “one pot” method by Jabbour et al. (2011) was analyzed by LC-MS-MS with an in-house BACid algorithm in order to compare the experimental unique peptides with a constructed proteome database of bacterial genus, species, and strain entries. The concentration of bacteria was also varied from 107 to 3.3 × 103 cfu/mL. The protein-processing method results in reliable identification of pure suspensions and mixtures at high and low bacterial concentrations. The peptide supernatant was concentrated and introduced into an LC system. The purpose was to provide for comprehensive processing of bacterial cell lysate proteins into a “one pot” design with no offline components, including the removal of reactants and byproducts after each step. Bacterial outer layers, membranes, and extraneous spore coat macromolecule material all need to be separated and removed from the soluble protein milieu so as not to interfere with the processing steps. Prior reagents, byproducts, and components were removed so as not to affect the protein for analyte integrity in subsequent steps. A 3-kDa molecular weight cutoff (MWCO) membrane was used without LC column or separation components; therefore, the proteins were retained during processing. Once peptides were generated on the MWCO membrane by trypsin digestion, they were passed through the membrane and loaded onto the analytical LC column for mass spectral analysis. The trypsin enzyme was retained by the membrane.
Further, different bacterial dilution protocols were performed for low-concentration studies. Dilute bacterial samples were analyzed in order to assess the peptide recovery and bacterial identification ability of the in-house bacterial classification and identification (BACid) algorithm (Deshpande et al., 2011). The “one pot” procedure and sample dilution method were investigated for identification and reproducibility concerns for a range of bacterial concentrations.
Bacterial Classification Analysis
Using the “one pot” method of Jabbour et al. (2010), analyses were performed with known and double-blind bacterial suspensions at different concentrations. Lower concentrations of bacteria in general provided a lower amount of peptide recovery. Lower bacterial concentrations of 3.3 × 103 and 3.3 × 104 organisms provided satisfactory identification capabilities. The E. coli K-12 strain provided a shortest single-linkage Euclidean distance in a dendrogram analysis to the database E. coli K-12 entry.
A double-blind sample was closest to S. aureus subspecies aureus MRSA252 at a relatively high concentration of 107 cfu/mL. At a lower concentration, S. aureus MRSA252 and S. aureus RF122 were determined to be the closest strains to the sample strain. The bacterium was subsequently revealed to be S. aureus American Type Culture Collection (ATCC) 12600. This particular strain is not contained in the database because its genome has not been sequenced. However, the analysis did include the experimental sample within the group of S. aureus strains.
This approach utilizes the knowledge of amino acid sequences of peptides derived from the proteolysis of proteins as a basis for reliable bacterial identification. To evaluate this approach, the tryptic digest peptides generated from double-blind biological samples containing either a single bacterium or mixture of bacteria were analyzed using LC-MS-MS. Bioinformatics tools that provide bacterial classification were used to evaluate the proteomics approach. Figure J-1 shows that bacteria in all of the double-blind samples were accurately identified with no false-positive assignment.
The approach also characterized double-blind bacterial samples when the experimental organism was not in the database due to its genome not having been sequenced. One experimental sample did not have its genome sequenced, and the peptide experimental record was added into the virtual bacterial proteome database. The MS proteomics approach proved capable of identifying and classifying organisms within a microbial mixture.
Several peptide searching algorithms (i.e., SEQUEST and MASCOT) have been developed to address peptide identification using proteomics databases that were generated from either fully or partially genome-sequenced organisms (Demirev and Fenselau, 2008; Ecker et al., 2005; Krishnamurthy et al., 2000). The above approach is based on SEQUEST and Agents of Biological Origin Identifier (ABOid). ABOid is based on bacterial proteins resident in an in-house proteome database translated
from an online database of sequenced microorganism genomes. The exploitation of this proteome database approach allowed for a faster search of the product ion spectra than that using genomic database searching. Also, it eliminated inconsistencies observed in publicly available protein databases caused by the utilization of nonstandardized gene-finding programs during the process of constructing the proteome database.
Blind Mixture Analysis
The BACid analysis of Sample 18 in Figure J-1 is shown in Figure J-2. BACid eliminated all the unwanted and degenerate peptides, and only the unique peptides that represented a 99 percent confidence level and above were retained for each organism. In this case, the number of
FIGURE J-1 Accurate identification of bacteria in double-blind samples: (a) experimental and (b) actual sample key. Sample 21 is a blank. Numbers in parentheses represent the number of proteins identified. Solid box, strain level ID; vertically hatched box, species-level ID; horizontal hatched box, genus-level ID.
unique peptides varied for the different bacterial candidates. E. faecalis had the highest number of unique peptides, followed by B. thuringiensis, and B. thailandensis had the least number of unique peptides. Also shown in Figure J-2 for Sample 18 are six bacterial candidates near the cutoff threshold within the Staphylococcus genus. Staphylococcus aureus ATCC 3359 strain present in the blind sample has not been sequenced and has not been reported in the public domains, and thus was not part of the constructed proteome database. However, BACid was capable of providing a nearest-neighbor match to the species level (aureus) and thus identified the bacterium correctly as S. aureus subsp. aureus. It is noteworthy to mention that this bacterial strain, which is not genomically sequenced, could only be identified to the species level.
FIGURE J-2 ABOid output for LC-MS-MS data processing of Sample 18. Ordinate provides actual number of SEQUEST generated and filtered unique peptides. Abscissa represents the bacteria found at least once in the 21 experimental samples.
A significant advantage of this approach is that if a particular strain has not been sequenced and yet the species is represented in the database, it is highly likely the unsequenced sample strain will be identified to that species level. Strain-level experimental identification is indicated by a single line (see Figure J-2) in the histogram (Enterococcus faecalis V538) or by a grouping of lines where one line clearly dominates (e.g. Burkholderia thailandensis E264 and Pseudomonas aeruginosa PAO1) with respect to the number of unique peptides. B. thuringiensis has two strains resident in the database, and the two provide similar sets of peptides. This occurs because the two strains do not display peptides that clearly distinguish themselves. This blind sample was correctly identified as a mixture of five bacteria: B. thuringiensis, S. aureus subsp. aureus, E. faecalis V583, B. thailandensis E264, and P. aeruginosa PA01, where S. aureus and B. thuringiensis were identified to the species level, and the other three were identified to the strain level.
Blind sample 17 was investigated for BACid characterization. The experimental set of peptides could provide results only to the Clostridium genus level because all nine Clostridia bacteria (species strains) resident in the database produced a histogram (data not shown) similar to that of Staphylococcus aureus in Figure J-2. The experimental peptides matched that portion of the virtual proteome common to all Clostridia. Therefore, the complete experimentally derived, tryptic peptide information record was stored as a separate bacterial line item as “Clostridium species 1” in the database of 881 bacteria. Another aliquot of the blind sample was processed with data reduction and searching in the new hybrid database. The highest match was with the Clostridium species 1 entry. After the results were submitted, the identity of Sample 17 was revealed to be Clostridium phytofermentans ISDg. This strain does not have its genome sequenced, yet BACid was able to match the virtual proteins that are similar with the Clostridium genus to the experimentally observed peptides. Thus, BACid was able to characterize Sample 17 as Clostridium without choosing one of the nine Clostridia strains resident in the database or other bacterial genera. BACid instead matched Clostridia species 1 to the experimental peptides, which indicated that there is sufficient information in the experimental peptides to differentiate Clostridium phytofermentans ISDg from the nine database Clostridia strains.
The results showed that the method was effective in identifying bacteria whether the sample was composed of one organism or a mixture or even if the sample was not resident in the database. No false-positives were observed for any of the blind samples that were analyzed, including the blank sample. There are some major advantages to the proteomic method over other molecular biology methods, such as the DNA-based methods, in that (a) no prior information about the sample is required for analysis, (b) no specific reagents are needed in the analysis process, (c) proteomics MS is capable of identifying an organism when a primer/probe set is not available, (d) proteomics MS requires less rigorous sample preparation than PCR, and (e) proteomics MS can provide a presumptive identification of a true unknown organism by mapping its phylogenetic relationship with other, known pathogens.
Sensitivity Performance of Mass Spectrometry–Based Proteomics
The MS proteomics method has shown promising results in specificity and sensitivity. While the latter parameter is highly dependent on the MS physical limit of detection, enhancing the biological sample processing is a crucial step to ease such dependency. Table J-1 shows bacteria and their sensitivity limits for the MS proteomics method.
TIER 3: BEYOND 2020
Future systems may combine various techniques, such as PCR with LC-ESI-MS-MS or MALDI-MS-MS or ESI-TOF-MS or antigen-antibody with LC-ESI-MS-MS or MALDI-MS-MS or ESI-TOF-MS.
The PLEX-ID is a product (Havlicek et al., 2013) that, although recently introduced (Jacob et al., 2012), has had a short-lived life in the commercial market. It and its predecessor, the Ibis T5000, were developed by Ibis Biosciences, Inc., which was acquired in 2009 by a subsidiary of the Abbott Diagnostics Group (Abbott Molecular Inc.). The product featured nucleic acid amplification (PCR) coupled with ESI-TOF-MS to carry out base-composition analysis (Ecker et al., 2008). The PLEX-ID instrument received the CE marking in March 2012 along with three assays for use on the system: PLEX-ID Viral IC Spectrum, PLEX-ID BAC Spectrum BC, and PLEX-ID Flu (Ibis, n.d.). In 2011 the company
TABLE J-1 Sensitivity Limits of Mass Spectrometry Proteomics Method
|Biological Agent Tested||Analytical Sensitivity||Reproducibility N=8||Specificity||Blind|
|% Detected||CV (95% confidence ID)||Positive ID||% detected|
|Yersinia pestis CO92||1.04 cfu||100||5%||✓||100|
|B. anthracis Ames, Sterne||6,000 cfu||100||5%||✓||100|
|Burkholderia mallei/pseudomallei||2,000 cfu||100||4.2%||✓||100|
|C. burnetii NMQ||8,000 cfu||90||5%||✓||100|
|F. tularensis type-A/type-B||1.1×104 cfu||100||5%||✓||100|
|E. coli O157/O104/O111 and O26||1,000 cfu||100||2%||✓||100|
|Dengue virus||1,200 cpu||100||12%||✓||88|
|Ricin toxin||75 pg||100||11%||✓||95|
|SEB toxin||25 pg||100||9%||✓||95|
also introduced the PLEX-ID Biothreat Assay. The system was reported to enable the identification and quantification of a broad set of pathogens, including bacteria (Sampath et al., 2012), all major groups of pathogenic fungi (Kaleta et al., 2011), protozoa, and the major families of viruses (MacInnes et al., 2011). In September 2012 Abbott discontinued the production of the PLEX-ID system; however, it did not exclude the possibility of developing a smaller, cheaper, and faster device based on similar principles in the future.
There are many research reports on viral typing performed by MALDI (Gijavanekar et al., 2012) or ESI (Jeng et al., 2012). There is a market gap in this field, although there are speculations about the renaissance of commercial techniques based on nucleic acid sequencing (Jeng et al., 2012) because the past Abbott PLEX-ID product was discontinued in September 2012.
Memory effect has not been found for the lab-made devices (Wu et al., 2004). After each digestion, the substrate and products left in the microchannels were cleaned out by pumping fresh water through the microchannels for a few minutes. Subsequently, a blank solution of 2-mM NH4HCO3 (pH 8.0) buffer was allowed to flow into the microchannels as for sample digestion, then collected and checked for any remaining samples in the channels using MALDI-TOF-MS detection procedures. No detectable peptide fragments were found after using the cleaning procedures described above, which indicates that the micro-reactors are less susceptible to memory effect. The lab-made micro-reactor devices can be used at least 50 times in one week without noticeable loss of activity with a proper storage at 4°C; the two ends of the microchannel were sealed to avoid drying-induced enzyme degradation (Shi et al., 1999). In fact, the ability to alter the digestion time by varying the flow rate could provide a powerful means to achieve the desired extent of digestion or to compensate for enzyme activity loss.
Finally, a new MS system currently being developed by Torion could be either Tier 2 or Tier 3. Torion has introduced a portable, ruggedized 32-pound, battery-operated GC-MS capable of providing two to three hours of operation on a single battery. The next-generation system is proposed to be on the order of 1,200 cubic inches, 15 pounds, and 40 watts of power.
Amado, F. M. L., P. Domingues, M. G. Santana-Marques, A. J. Ferrer-Correia, and K. B. Tomer. 1997. Discrimination effects and sensitivity variations in matrix-assisted laser desorption/ionization. Rapid Communications in Mass Spectrometry 11:1337–1352.
Andersson, H., and A. van den Berg. 2004. Microtechnologies and nanotechnologies for single-cell analysis. Current Opinion in Biotechnology 15:44–49.
Baoa, N., B. Jagadeesan, A. K. Bhunia, Y. Yao, and C. Lu. 2008. Quantification of bacterial cells based on autofluorescence on a microfluidic platform. Journal of Chromatography A 1181:153–158.
Basile, F., M. B. Beverly, C. Abbas-Hawkes, C. D. Mowry, K. J. Voorhees, and T. L. Hadfield. 1998. Direct mass spectrometric analysis of in situ thermally hydrolyzed and methylated lipids from whole bacterial cells. Analytical Chemistry 70:1555–1562.
Bhagat, A. A. S., H. W. Hou, L. D. Li, C. T. Lim, and J. Han. 2011. Pinched flow coupled shear-modulated inertial microfluidics for high-throughput rare blood cell separation. Lab on a Chip 11:1870-1878.
Bright, J. J., M. A. Claydon, M. Suofian, and D. B. Gordon. 2002. Rapid typing of bacteria using matrix assisted laser desorption ionization time-of-flight mass spectrometry and pattern recognition software. Journal of Microbiological Methods 48:127–138.
Castanha, E. R., A. Fox, and K. F. Fox. 2006. Rapid discrimination of Bacillus anthracis from outer members of the B. cereus group by mass and sequence of ”intact” small acid soluble proteins (SASPs) using mass spectrometry. Journal of Microbiological Methods 67:230–240.
Demirev, P. A., and C. Fenselau. 2008. Mass spectrometry in biodefense. Journal of Mass Spectrometry 43:1441–1457.
Demirev, P. A., Y.-P. Ho, V. Ryzhov, and C. Fenselau. 1999. Microorganism identification by mass spectrometry and protein database searches. Analytical Chemistry 71:2732–2738.
Demirev, P. A., J. S. Lin, F. J. Pineda, and C. Fenselau. 2001. Bioinformatics and mass spectrometry for microorganism identification: Proteome-wide post-translational modifications and database search algorithms for characterization of intact H. pylori. Analytical Chemistry 73:4566–4573.
Deshpande, S. V., R. E. Jabbour, P. A. Snyder, M. Stanford, C. H. Wick, and A. W. Zulich. 2011. ABOid: A software for automated identification and phyloproteomics classification of tandem mass spectrometry data. Journal of Chromatography and Separation Techniques S5:001.
Dickinson, D. N., M. T. La Duc, M. Satomi, J. D. Winefordner, D. H. Powell, and K. Venkateswaran. 2004. MALDI-TOFMS compared with other polyphasic taxonomy approaches for the identification and classification of Bacillus pumilus spores. Journal of Microbiological Methods 58:1–12.
Dittich, P. S., K. Tachikawa, and A. Manz. 2006. Micro total analysis systems: Latest advancements and trends. Analytical Chemistry 78:3887–3908.
Domin, M. A., K. J. Welham, and D. S. Ashton. 1999. The effect of solvent and matrix combinations on the analysis of bacteria by matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry. Rapid Communications in Mass Spectrometry 13:222–226.
Dworzanski, J. P., and A. P. Snyder. 2005. Classification and identification of bacteria using mass spectrometry-based proteomics. Expert Reviews of Proteomics 2:863–878.
Ecker, D. F., R. Sampath, L. B. Blyn, M. W. Eshoo, C. Ivy, J. A. Ecker, B. Libby, V. Samant, K. A. Sannes-Lowery, R. E. Melton, K. Russell, N. Freed, C. Barrozo, J. Wu, K. Rudnick, A. Desai, E. Moradi, D. J. Knize, D. W. Robbins, J. C. Hannis, P. M. Harrell, C. Massire, T. A. Hall, Y. Jiang, R. Ranken, J. J. Drader, N. White, J. A. McNeil, S. T. Crooke, and S. A. Hofstadler. 2005. Rapid identification and strain-typing of respiratory pathogens for epidemic surveillance. Proceedings of the National Academy of Science of the United States of America 102:8012–8017.
Ecker, D. J., R. Sampath, C. Massire, L. B. Blyn, T. A. Hall, M. W. Eshoo, and S. A. Hofstadler. 2008. Ibis T5000: A universal biosensor approach for microbiology. Nature Reviews Microbiology 6:553-558.
Ethier, F., W. Hou, H. S. Duewel, and D. Figeys. 2006. The proteomic reactor: A microfluidic device for processing minute amounts of protein prior to mass spectrometry analysis. Journal of Proteome Research 5:2754–2759.
Feng, X., W. Du, Q. Luo, and B.-F. Liu. 2009. Microfluidic chip: Next-Generation platform for systems biology. Analytica Chimica Acta 650:83–97.
Fenselau, C., and P. A. Demirev. 2001. Characterization of intact microorganisms by MALDI mass spectrometry. Mass Spectrometry Reviews 20:157–171.
Fergenson, D. P., M. E. Pitesky, H. J. Tobias, P. T. Steele, G. A. Czerwieniec, S. C. Russell, C. B. Lebrilla, J. M. Horn, K. R. Coffee, A. Srivastava, S. P.
Pillai, M. T. P. Shih, H. L. Hall, A. J. Ramponi, J. T. Chang, R. G. Langlois, P. L. Estacio, R. T. Hadley, M. Frank, and E. E. Gard. 2004. Reagentless detection and classification of individual bioaerosol particles in seconds. Analytical Chemistry 76:373–378.
Fox, A., M. Anderson, J. Dunn, B. Guenther, L. Parks, R. Pinnick, C. Reed, J. Rowe, R. Luftig, G. W. Long, R. Lontz, G. L. Marchin, A. T. McManus, P. Setlow, J. Siedow, A. J. Sievers, M. L. Tchikindas, and S. Tove. 2002. Report of the “Bioterrorism Workshop”—Duke University Thomas Center on 2–4 April 2002 organized by U.S. Army Research Office. Journal of Microbiological Methods 51:247–254.
Freire, S. L. S., and A. R. Wheeler. 2006. Proteome-on-a-chip: Mirage, or on the horizon? Lab on a Chip 6:1415–1423.
Gantt, S. L., N. B. Valentine, A. J. Saenz, M. T. Kingsley, and K. L. Wahl. 1999. Use of internal control for matrix-assisted laser desorption/ionization time-of-flight mass spectrometry analysis of bacteria. Journal of the American Society of Mass Spectrometry 10:1131–1137.
Gijavanekar, C., R. Drabek, M. Soni, G. W. Jackson, U. Strych, G. E. Fox, Y. Fofanov, and R. C. Willson. 2012. Detection and typing of viruses using broadly sensitive cocktail-PCR and mass spectrometric cataloging: Demonstration with dengue virus. Journal of Molecular Diagnostics 4:402-407.
Griest, W. H., M. B. Wise, K. J. Hart, S. A. Lammert, C. V. Thompson, and A. A. Vass. 2001. Biological agent detection and identification by the Block II chemical biological mass spectrometer. Field Analytic Chemistry and Technology 5:177–184.
Halden, R. U., D. R. Colquhoun, and E. S. Wisniewski. 2005. Identification and phenotypic characterization of Sphingomonas wittichii strain RW1 by peptide mass fingerprinting using matrix-assisted laser desorption ionizationtime of flight mass spectrometry. Applied and Environmental Microbiology 71(5):2442–2451.
Hart, K. J., M. B. Wise, W. H. Griest, and S. A. Lammert. 2000. Design, development, and performance of a fieldable chemical biological agent detector. Field Analytical Chemistry and Technology 4:93–110.
Havlicek, V., K. Lemr, and K. A. Schug. 2013. Current trends in microbial diagnostics based on mass spectrometry. Analytical Chemistry 85:790–797.
Heidelberg, J. F., J. A. Eisen, W. C. Nelson, R. A. Clayton, M. L. Gwinn, R. J. Dodson, D. H. Haft, E. K. Hickey, J. D. Peterson, L. Umayam, S. R. Gill, K. E. Nelson, T. D. Read, H. Tettelin, D. Richardson, M. D. Ermolaeva, J. Vamathevan, S. Bass, H. Y. Qin, I. Dragoi, P. Sellers, L. McDonald, T.
Utterback, R. D. Fleishmann, W. C. Nierman, O. White, S. L. Salzberg, H. O. Smith, R. R. Colwell, J. J. Mekalanos, J. C. Venter, and C. M. Fraser. 2000. DNA sequence of both chromosomes of the cholera pathogen Vibrio cholerae. Nature 406:477-483.
Henion, J. 2009. The reality of lab-on-a-chip technology for the mass spectrometry laboratory. LCGC North America 27:900–915.
Hettick, J. M., M. L. Kashon, J. P. Simpson, P. D. Siegel, G. H. Mazurek, and D. N. Weissman. 2004. Proteomic profiling of intact Mycobacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Analytical Chemistry 76:5769–5776.
Holmes, D., and H. Morgan. 2010. Single cell impedance cytometry for identification and counting of CD4 T-cells in human blood using impedance labels. Analytical Chemistry 82:1455-1461.
Huikko, K., R. Kostiainen, and T. Kotiaho. 2003. Introduction to microanalytical systems: Bioanalytical and pharmaceutical applications. European Jour-nal of Pharmaceutical Science 20:149–171.
Ibis. n.d. Welcome page. http://www.ibisbiosciences.com/index.html (accessed on October 27, 2012).
Jabbour, R., J. P. Dworzanski, S. V. Deshpande, A. P. Snyder, and C. H. Wick. 2005. Effect of gas phase fractionation of peptide ions on bacterial identification using mass spectrometry-based proteomics approach. Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, San Antonio, TX, TP31.
Jabbour, R. E., S. V. Deshpande, M. M. Wade, M. F. Stanford, C. H. Wick, A. W. Zulich, E. W. Skowronski, and A. P. Snyder. 2010. Double blind characterization of non-genome-sequenced bacteria by mass spectrometrybased proteomics. Applied and Environmental Microbiology 76:3637–3644.
Jabbour, R. E., S. V. Deshpande, M. F. Stanford, C. H. Wick, A. W. Zulich, and A. P. Snyder. 2011. A protein processing filter method for bacterial identification by mass spectrometry-based proteomics. Journal of Proteome Research 10:907–912.
Jacob, D., U. Sauer, R. Housley, C. Washington, K. Sannes-Lowery, D. J. Ecker, R. Sampath, and R. Grunow. 2012. Rapid and high-throughput detection of highly pathogenic bacteria by Ibis PLEX-ID technology. PLoS ONE 7(6):e39928.
Jang, K., Y. Tanaka, J. Wakabayashi, R. Ishii, K. Sato, K. Mawatari, M. Nilsson, and T. Kitamori. 2012. Selective cell capture and analysis using shallow antibody-coated microchannels. Biomicrofluidics 6(4):044117.
Jeng, K., C. Massire, T. R. Zembower, V. M. Deyde, L. V. Gubareva, Y. H. Hsieh, R. E. Rothman, R. Sampath, S. Penugonda, D. Metzgar, L. B. Blyn, J. Hardick, and C. A. Gaydos. 2012. Monitoring seasonal influenza A evolution: Rapid 2009 pandemic H1N1 surveillance with a reverse transcription-polymerase chain reaction/electro-spray ionization mass spectrometry assay. Journal of Clinical Virology 54:332-336.
Jing, W., W. Zhao, S. Liu, L. Li, C.-T. Tsai, X. Fan, W. Wu, J. Li, X. Yang, and G. Sui. 2013. Microfluidic device for efficient airborne bacteria capture and enrichment. Analytic Chemistry 85:5255–5262.
Kaleta, E. J., A. E. Clark, A. Cherkaoui, V. H. Wysocki, E. L. Ingram, J. Schrenzel, and D. M. Wolk. 2011. Use of PCR coupled with electrospray ionization mass spectrometry for rapid identification of bacterial and yeast bloodstream pathogens from blood culture bottles. Clinical Chemistry 57:1057-1067.
Kollipara, S., N. Agarwal, B. Varshney, and J. Paliwal. 2011. Technological advancements in mass spectrometry and its impact on proteomics. Analytical Letters 44:1498–1520.
Krishnamurthy, T., P. L. Ross, and U. Rajamani. 1996. Detection of pathogenic and non-pathogenic bacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Communications in Mass Spectrometry 10(8):883–888.
Krishnamurthy, T., U. Rajamani, P. L. Ross, R. Jabbour, H. Nair, J. Eng, J. Yates, M. T. D. Douglas, C. Stahl, and T. D. Lee. 2000. Mass spectral investigation of microorganisms. Toxin Reviews 19:95–117.
Lau, S. K. P., B. S. F. Tang, S. O. T. Curreem, T.-M. Chan, P. Martelli, C. W. S. Tse, A. K. L. Wu, K.-Y. P. Yuen, and C. Y. Woo. 2012. Matrix-assisted laser desorption ionization-time of flight mass spectrometry for rapid identification of Burkholderia pseudomallei: Importance of expanding databases with pathogens endemic to different localities. Journal of Clinical Microbiology 50:3142–3143.
Lay, J. O., Jr. 2001. MALDI–TOF mass spectrometry of bacteria. Mass Spectrometry Reviews 20:172–194.
Lee, K., D. Bae, and D. Lim. 2002. Evaluation of parameters in peptide mass fingerprinting for protein identification by MALDI-TOF mass spectrometry. Molecules and Cells 13:175–184.
Lee, K. S., K. H. Bartlett, M. Brauer, G. M. Stephens, W. A. Black, and K. Teschke. 2004. A field comparison of four samplers for enumerating fungal aerosols I. Sampling characteristics. Indoor Air 14:360-366.
Li, G., M. Waltham, N. L. Anderson, E. Unsworth, A. Treston, and J. N. Weinstein. 1997. Rapid mass spectrometric identification of protein from two-dimensional polyacrylamide gels after in gel proteolytic digestion. Electrophoresis 18:391–402.
Li, Y. B., and X. L. Su. 2006. Microfluidics-based optical biosensing method for detection of Escherichia coli O157:H7. Journal of Rapid Methods and Automation in Microbiology 14:96-109.
Lim, E., A. Tay, and A. G. Nicholson. 2012. Antibody independent microfluidic cell capture of circulating tumor cells for the diagnosis of cancer. Journal of Thoracic Oncology 7:E42-E43.
Lion, N., T. C. Rohner, L. Dayon, I. L. Arnaud, E. Damoc, N. Youhnovski, Z.-Y. Wu, C. Roussel, J. Josserand, H. Jensen, J. S. Rossier, M. Przybylski, and H. H. Girault. 2003. Microfluidic systems in proteomics. Electrophoresis 24:3533–3562.
Liu, C. 2010. Rapid fabrication of three-dimensional microfluidic chip using natural lotus leaf template. Microfluidics and Nanofluidics 9:923-931.
Liu, D.-Y., D. Rutherford, M. Kinsey, and K. A. Prather. 1997. Real time monitoring of pyrotechnically derived aerosol particles in the troposphere. Analytical Chemistry 69:1808–1814.
Liu, H., Z. Du, J. Wang, and R. Yang. 2007. Universal sample preparation method for characterization of bacteria by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Applied Environmental Microbiology 73:1899–1907.
Liu, Y., H. Lu, W. Zhong, P. Song, J. Kong, P. Yang, H. H. Girault, and B. Liu. 2006. Multilayer-assembled microchip for enzyme immobilization as reactor toward low-level protein identification. Analytical Chemistry 78:801–808.
Loferer-Krobacher, M., J. Klima, and R. Psenner, R. 1998. Determination of bacterial cell dry mass by transmission electron microscopy and densitometric image analysis. Applied Environmental Microbiology 64:688–694.
Loutherback, K. J. D’Silva, L. Y. Liu, A. Wu, R. H. Austin, and J. C. Sturm. 2012. Deterministic separation of cancer cells from blood at 10 mL/min. AIP Advances 2(4):e42107.
Ma, J., J. Liu, L. Sun, L. Gao, Z. Liang, L. Zhang, and Y. Zhang, Y. 2009. Online integration of multiple sample pretreatment steps involving denaturation, reduction, and digestion with microflow reversed-phase liquid chromatography- electrospray ionization tandem mass spectrometry for highthroughput proteome profiling. Analytical Chemistry 81:6534–6540.
MacInnes, H., Y. Zhou, K. Gouveia, J. Cromwell, K. Lowery, R. C. Layton, M. Zubelewicz, R. Sampath, S. Hofstadler, Y. S. Liu, Y. S. Cheng, and F. Koster. 2011. Transmission of aerosolized seasonal H1N1 influenza A to ferrets. PLoS ONE 6(9):e24448.
Madonna, A. J., F. Basile, I. Ferrer, M. A. Meetani, J. C. Rees, and K. J. Voorhees. 2000. On-probe sample pretreatment for the detection of proteins above 15 KDa from whole cell bacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Communications in Mass Spectrometry 14:2220–2229.
Owen, R. J., M. A. Claydon, J. Gibson, B. Burke, and A. Ferrus. 1999. Strain variation within Helicobacter pylori detected by mass spectrometry of cell wall surfaces. Gut 45(Suppl. 3):A28.
Park, S., Y. Zhang, S. Lin, T.-H. Wang, and S. Yang. 2011. Advances in microfluidic PCR for point-of-care infectious disease diagnostics. Biotechnology Advances 29:830-839.
Pasanen, A. L. 2001. A review: Fungal exposure assessment in indoor environments. Indoor Air 11:87-98.
Pennington, S. R., M. R. Wilkins, D. F. Hochstrasser, and M. J. Dunn. 1997. Proteome analysis: From protein characterization to biological function. Trends in Cell Biology 7:168–173.
Pineda, F. J., J. S. Lin, C. Fenselau, and P. A. Demirev. 2000. Testing the significance of microorganism identification by mass spectrometry and proteome database search. Analytical Chemistry 72:3739–3744.
Qi, A., L. Yeo, J. Friend, and J. Ho. 2010. The extraction of liquid, protein molecules and yeast cells from paper through surface acoustic wave atomization. Lab on a Chip 10:470-476.
Reisewitz, S., H. Schroeder, N. Tort, K. A. Edwards, A. J. Baeumner, and C. M. Niemeyer. 2010. Capture and culturing of living cells on microstructured DNA substrates. Small 6:2162-2168.
Ren, P., T. M. Jankun, K. Belanger, M. B. Bracken, and B. P. Leaderer. 2001. The relation between fungal propagules in indoor air and home characteristics. Allergy 56:419-424.
Ruelle, V., B. E. Moualij, W. Zorzi, P. Ledent, and E. De Pauw. 2004. Rapid identification of environmental bacterial strains by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Communications in Mass Spectrometry 18:2013–2019.
Russell, S. C., G. Czerwieniec, C. Lebrilla, P. Steele, V. Riot, K. Coffee, M. Frank, and E. E. Gard. 2005. Achieving high detection sensitivity (14 zmol) of biomolecular ions in bioaerosol mass spectrometry. Analytical Chemistry 77:4734-4741.
Sampath, R., N. Mulholland, L. B. Blyn, C. Massire, C. A. Whitehouse, N. Waybright, C. Harter, J. Bogan, M. S. Miranda, D. Smith, C. Baldwin, M. Wolcott, D. Norwood, R. Kreft, M. Frinder, R. Lovari, I. Yasuda, H. Matthews, D. Toleno, R. Housley, D. Duncan, F. Li, R. Warren, M. W. Eshoo, T. A. Hall, S. A. Hofstadler, and D. J. Ecker. 2012. Comprehensive biothreat cluster identification by PCR/electrospray-ionization mass spectrometry. PLoS ONE 7(6):e36528.
Sedgwick, H., F. Caron, P. B. Monaghan, W. Kolch, and J. M. Cooper. 2008. Lab-on-a-chip technologies for proteomic analysis from isolated cells. Journal of the Royal Society Interface 5:S123–S130.
Shi, H., W. B. Tsai, M. D. Garrison, S. Ferrari, and B. D. Ratner. 1999. Template-imprinted nanostructured surfaces for protein recognition. Nature 398:593–597.
Steele, P. T., H. J. Tobias, D. P. Fergenson, M. E. Pitesky, J. M. Horn, G. A. Czerwieniec, S. C. Russell, C. B. LeBrilla, E. E. Gard, and M. Frank. 2003. Laser power dependence of mass spectral signatures from individual bacterial spores in bioaerosol mass spectrometry. Analytical Chemistry 7:5480–5487.
Steele, P. T., G. R. Farquar, A. N. Martin, K. R. Coffee, V. J. Riot, S. I. Martin, D. P. Fergenson, E. E. Gard, and M. Frank. 2008. Autonomous, broadspectrum detection of hazardous aerosols in seconds. Analytical Chemistry 80:4583–4589.
Su, Y., M. F. Sipin, H. Furutani, and K. A. Prather. 2004. Development and characterization of an aerosol time-of-flight mass spectrometer with increased detection efficiency. Analytical Chemistry 76:712–719.
Szita, N., K. Polizzi, N. Jaccard, and F. Baganz. 2010. Microfluidic approaches for systems and synthetic biology. Current Opinions in Biotechnology 21:517–523.
Tian, R., X. D. Hoa, J.-P. Lambert, J. P. Pezacki, T. Veres, and D. Figeys. 2011. Development of a multiplexed microfluidic proteomic reactor and its applications for studying protein-protein interactions. Analytical Chemistry 83:4095–4102.
Wang, C., R. Oleschuck, F. Ouchen, P. Li, P. Thibault, and D. J. Harrison. 2000. Integration of immobilized trypsin bead beds for protein digestion within a microfluidic chip incorporating capillary electrophoresis separations and an electrospray mass spectrometry interface. Rapid Communications in Mass Spectrometry 14:1377–1383.
Wang, Z., L. Russon, L. Li, D. C. Roser, and S. R. Long. 1998. Investigation of spectral reproducibility in direct analysis of bacteria proteins by matrixassisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Communications in Mass Spectrometry 12:456–464.
Williams, T. L., P. Leopold, and S. Musser. 2002. Automated post processing of electrospray LC/MS data for profiling protein expression in bacteria. Analytical Chemistry 74:5807–5813.
Williams, T., D. Andrzejewski, J. O. Lay, Jr., and S. M. Musser. 2003. Experimental factors affecting the quality and reproducibility of MALDI TOF mass spectra obtained from whole bacteria cells. Journal of the American Society of Mass Spectrometry 1:342–351.
Wisniewski, J. R., A. Zougman, N. Nagaraj, and M. Mann. 2009. Universal sample preparation method for proteome analysis. Nature Methods 6:359–362.
Wu, H., J. Zhai, Y. Tian, H. Lu, X. Wang, W. Jin, B. Liu, P. Yang, Y. Xu, and H. Wang. 2004. Microfluidic enzymatic-reactors for peptide mapping: Strategy, characterization, and performance. Lab on a Chip 4:588–597.
Yates, J. R. III. 2004. Mass spectral analysis in proteomics. Annual Reviews of Biophysics and Biomolecular Structures 33:297–316.
Yates, J. R., C. I. Ruse, and A. Nakorchevsky. 2009. Proteomics by mass spectrometry: Approaches, advance, and applications. Annual Review of Biomedical Engineering 11:49–79.
You, D. J., P. L. Tran, H.-J. Kwon, D. Patel, and J.-Y. Yoon. 2011. Very quick reverse transcription polymerase chain reaction for detecting 2009 H1N1 influenza A using wire-guide droplet manipulations. Faraday Discussions 149:159-170.
Zhou, H., Z. Ning, A. E. Starr, M. Abu-Farha, and D. Figeys. 2012. Advancements in top-down proteomics. Analytical Chemistry 84:720–734.