1 MASS SPECTROMETRIC STUDIES OF ENDOGENOUS THIO CYANATE IN EXHALED BREATH CONDENSA TES AND ORAL RINSES By FRANK A. KERO A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008
2 Frank A. Kero
3 To my loving family and my future wife, Stacey
4 ACKNOWLEDGMENTS I would like to thank Dr. Richard Yost for hi s support and guidance during the course of my graduate career. I am thankful for our act ive Yost group alumni, es pecially Dr. Randy Pedder and Dr. Jodie Johnson for valuable discussions on the fundamentals of mass spectrometry and data interpretation. I would also like to acknowledge UF alumnus Dr. Gregory Conner, associate professor of cell biology, anatomy and medicine at the University of Miami, for providing the initial motivation to st udy thiocyanate and exhaled breath condensates. I acknowledge Dr. Donn Dennis, Dr. Timothy Morey and Dr. Scott Wasdo at the department of anesthesiology (UF) for their financial support and valuable discus sions during my time with their group. This dissertation would not have been possible with out the administra tive contributions and guidance of Dr. Matthew M. Booth (also at the depart ment of anesthesiology). Matthew was a great mentor and friend who helped me navigate th e graduate experience. I also acknowledge Dr. Bruce Goldberger and Dr. Michelle Merves for valuable discussions regarding applications development, method validation and manuscript a dvice. I acknowledge my Florida family: future doctors David Richardson, Michael Napolitano a nd Jennifer Bryant for their friendship and encouragement. I also acknowledge Davids wife Samantha whose wonderful home cooked meals made me feel like I was ne ver far away from family. I am very grateful for the love and support of my mother and father, my sister Ta nya and for my Stacey, who left her beloved New York Yankees to follow me down to the swam ps of Gainesville. I acknowledge my current employers, Dr. Ben Blount and Dr. Liza Valentin-Blasini, who have become a welcome part of the final stages of my professional developm ent prior to my defense. Their words of encouragement, flexibility in work schedule, friendship and mentoring were very much appreciated in the final stages of this process.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................7LIST OF FIGURES .........................................................................................................................9ABSTRACT ...................................................................................................................... .............12 CHAP TER 1 INTRODUCTION .................................................................................................................. 14Exhaled Breath Condensates (E BC) Background and Significance ....................................... 14Project Motivation for Studying Thiocyanate in Airway Secretions: Cystic Fibrosis ........... 18Project Motivation for Measuring SCNin Human Oral Rinses: Background and Significance .................................................................................................................. .......21The Advantages of Saliva (Oral Rinses) Versus Other Biofluids .......................................... 22Hyphenated Analytical Method Lead to Improved Method Detection Limits ....................... 25Overview of the Dissertation ..................................................................................................272 MASS SPECTROMETRY AND TANDEM MASS SPECTROMETRY: OPTIMIZATION AND DEVELOPMENT ........................................................................... 30Overview ...................................................................................................................... ...........30Fundamentals of Mass Spectrometry: Quadrupole Mass Filters, Theoretical and Practical Consideration .......................................................................................................30Theory and Practice of Quadrupole Mass Filters ................................................................... 31Tandem Mass Spectrometry ...................................................................................................37Commercial Instrumentation .................................................................................................. 42Optimization of Mass Spectrometry Parameters .................................................................... 43Resolution .................................................................................................................... ....45Compound Selective Parameters before Q1 .................................................................... 47Optimization of Tandem Mass Spectrometry Overview ........................................................48Defining Tandem Mass Spectrome try Figures of Merit ..................................................49Justification of the Selection of the Appropriate SRM Transition .................................. 55Summary and Conclusions .....................................................................................................563 METHOD DEVELOPMENT AND OP TIMIZATION OF ELECTROS PRAY IONIZATION SOURCE PARAMETERS AND LIQUID CHROMATOGRAPHY ............58Overview ...................................................................................................................... ...........58Review of Atmospheric Pre ssure Ionization Sources ............................................................. 58Fundamentals of the Electrospray Ionization Source .............................................................60Commercial Instrumentation .................................................................................................. 70
6 Reversed Phase Liquid Chromatography ............................................................................... 72Determination of Dead Time: Nontraditional Calculation .....................................................75Evaluation of Analyte Suppression Due to Matrix Effects by Post-Column Infusion ...........76The Selection of Mobile Phase Flow Rate ............................................................................. 82HPLC Mobile Phase and Modifiers ........................................................................................ 84Evaluation of the Tailing Factor (T) ....................................................................................... 86Evaluation of EBC Samples ................................................................................................... 90Offline Sample Enrichment Using a Nitrogen Evaporator .....................................................95Summary and Conclusions .....................................................................................................974 QUANTITATIVE ANALYSIS: MEASUREMENT AND VALIDATION OF FREE THIOCYANATE CONCENTRATI ONS IN HUMAN ORAL RINSES (PROCESSED SALIVA) ....................................................................................................................... .......102Overview ...................................................................................................................... .........102The Institutional Review Board and the Prot ection of Human Subj ects: IRB History ........ 103Protocol Summary ................................................................................................................103Validation Summary .............................................................................................................104The Qualitative Identification of SCNin Oral Fluid ...........................................................104Evaluation of Analyte Recovery from the Dental Gauze and Salivette ............................... 108Stability .................................................................................................................................108Standard ................................................................................................................................109Sensitivity & Limitatio ns of Linearity ..................................................................................110Precision ..................................................................................................................... ..........110Human Subjects Data ...........................................................................................................115Replicate Measurements of a Calibration Control Standard ................................................ 118Evaluation of Carryover ....................................................................................................... 118Conclusions ...........................................................................................................................1215 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK ...................................... 123Conclusions ...........................................................................................................................123Suggestions for Future Work Part I: Signal-to-Noise: Increase Signal ................................ 125Conclusions Part II: Signal-t o-Noise: Decrease Noise ......................................................... 126APPENDIX A RECRUITMENT ADVERTISEMENT FO R HUMAN SUBJECTS STUDY .................... 128B IRB APPROVED PROTOCOL AND INFORMED CONSENT FORMS .......................... 129LIST OF REFERENCES .............................................................................................................136BIOGRAPHICAL SKETCH .......................................................................................................148
7 LIST OF TABLES Table page 1-1 Summary of current breath based diagnostic m ethods that have been approved by the FDA....................................................................................................................................161-2 The chemical composition of saliva responsib le for the protection of the human host, .... 231-3 A comparison of the intuitive order of a t ypical analytical experiment compared with the chronology of analytical development .........................................................................272-1 MS and MS/MS operational modes of a triple quadrupole mass spectrometer, showing the modes for the two mass filter s (Quadrupole 1 and Quadrupole 3) and the quadrupole collision cell (q2) ............................................................................................ 402-2 Intensities of precurs or and product ions of m/z 58 at 45 eV collision energy versus increasing chamber pressure (CAD). ................................................................................. 523-1 Reported optimal values for flow rate based on the internal diameter of an HPLC column......................................................................................................................... .......713-2 Recommended operational parameters for the positioning of the ESI capillary relative to the or ifice at various flow rates ......................................................................... 713-3 Summary of the MS/MS parameter optimization .............................................................. 743-4 Determination of the amount of sa liva collected in a 5 minute period .............................. 793-5 Summary of flow rate study ............................................................................................... 843-6 Evaluation of the effect of mobile pha se modifiers on the MS/MS peak areas for SCNby SRM m/z 5826. ................................................................................................863-7 Evaluation of the effect of variations in mobile phase composition on peak symmetry for SCN-. ............................................................................................................................ 893-8 Evaluation of the effect of injection on volume on peak symmetry. ................................. 904-1 The validation of the Salivette saliva sampling method by the determination of absolute recovery. ............................................................................................................1084-2 Peak areas for m/z 58 26 used in the determination of absolute recovery. ................... 1094-3 Stability study for three freeze-thaw cycles of SCNstandards at three concentrations to represent low, medium and high concentrations in clinical samples. 1104-4 Summary of the determination of the method detection limit (MDL) for SCN-. MDL = stdev x t; .................................................................................................................. ......111
8 4-5 Evaluation of the within r un drif t in elution time for SCNin five replicates of the same clinical specimen .................................................................................................... 1124-6 Evaluation of the precision of the autosampler using single variable ANOVA .............. 1134-7 ANOVA data for a 1 ppm SCNsolution standard .......................................................... 1134-8 ANOVA data for a 2 ppm SCNsolution standard .......................................................... 1144-9 ANOVA data for a 7 ppm SCNsolution standard .......................................................... 1144-10 ANOVA data for an 11.4 ppm SCNsolution standard ...................................................1154-11 The quantification of [SCN-] in human oral rinses in 18 human subjects by reversed phase LC-ESI-MS/MS. The lowest standard concentration was 50 g/mL. ................... 1164-12 Isotope dilution ESI-M S/MS calibration of SCNsolution standards ............................. 1174-13 The evaluation of a calibration control st andard to monitor instrument performance during the course of the experiment. ................................................................................ 1194-14 Comparison of the measured values for salivary [SCN-] with similar studies in literature. ..........................................................................................................................121
9 LIST OF FIGURES Figure page 1-1 The generation of microaerosols of surf ace liqu id from areas of high turbulence in the lung and trachea. ..........................................................................................................171-2 The number of EBC papers published versus time. ........................................................... 172-1 The quadrupole mass filter and its power supply .............................................................. 322-2 Mathieu stability diagram for an ion of m/z 219 in a quadrupole mass filter with 9.5mm diameter round rods and an RF frequency of 1.2 MHz, ............................................. 342-3 Mathieu stability diagram and resolution. .........................................................................362-4 Schematic of a triple quadrupole mass spectrometer (TQMS). ......................................... 382-5 Plot of the three-dimensional dataset generated by electron ionization/MS/MS of cyclohexane. .................................................................................................................. .....412-6 Overview of the Applied Biosystems Sciex API 4000 LC-MS/MS. ................................. 412-7 Illustration of the differe nt pressure regimes of the Applied Biosystems Sciex API 4000....................................................................................................................................432-8 The LINAC collision cell used in q2 of the API 4000: A) Conical rods B) T-shaped electrodes. ................................................................................................................... .......432-9 60 MCA full scans (Q1 only) for a 1 g/mL KSCN standard in 50/50 AcCN/H2O in negative ion mode under default system parameters. ........................................................ 462-10 Comparison of Q1 scans at three resolution settings to determine the appropriate setting to resolve m/z 58 prior to fragmentation ................................................................ 472-11 The isolation of m/z 58 in Q1 using the SIM operational mode.. ...................................... 492-12 Product ion scans m/z 58: The effect of collision cell pressure on the abundance of fragmentation from the precursor ions at m/z 58. .............................................................. 532-13 Proposed MS/MS CAD mechanism for SCN-................................................................... 542-14 Selectivity: Comparison of the LC-ES I-MS/MS selected reaction monitoring (SRM) transitions for chromatograms of a) m/z 5832 b) m/z 58 26. Experimental parameters are detailed in main body text. ........................................................................ 543-1 TurboIonSpray source. ..................................................................................................... ..70
10 3-2 Surface plot of the intensity of the dete c tor response obtained at various positions of the TurboIonSpray probe with re spect to the or ifice plate. ...............................................733-3 Post-column infusion experiment design. .......................................................................... 773-4 Post-column infusion chromatogram of m obile phase for the negative ion transition at m/z 5826 ..................................................................................................................... 793-5 Post-column infusion comparison of more d ilute clinical specimen of the same oral rinse specimen from a human volunteer for the transition of m/z 5826. 980 L mobile phase + 20 L oral rinses. ......................................................................................803-6 Post-column infusion comparison of a clinical specimen of an oral rinse from a human volunteer for the transition of m/z 5826. 900 L mobile phase + 100 L oral rinses. ..........................................................................................................................813-7 LC-MS/MS transition m/z 5826 at 200 L/min flow rate. ............................................. 833-8 LC-MS/MS transition m/z 5826 at 100 L/min flow rate. ............................................. 833-9 Evaluating the effect of mobile phase modifiers on the MS/MS peak area intensity for SRM m/z 5826. .........................................................................................................873-10 Defining the tailing factor (T) figure of merit for syst em suitability calculations. Image available at www.forumsci.co.il/HPLC/SST_abic.pdf ........................................... 873-11 Effect of peak symmetry on peak integr ation in terms of T. Image available at www.forumsci.co.il/HPL C/SST_abic.pdf ......................................................................... 883-12 Evaluation of the effect of variati ons in mobile phase composition on peak symmetry, as measured by tailing factor T. ....................................................................... 893-13 The RTube design to separate EBC and saliva. ................................................................. 913-14 Description of the path of expired breath after the saliva trap, at the entrance of the one-way valve. ...................................................................................................................913-15 LC-MS/MS chromatograms for an E BC samples collected from the Jaeger EcoScreen ..................................................................................................................... .....943-16 LC-MS/MS chromatogram for two E BC specimens collected using the RTube collection device and preconcentrated by a f actor of 2 using a nitrogen evaporator. S/N was calculated using Analyst software .......................................................................963-17 Triplicate injections of an EBC sample monitored by LC-MS/MS. Two transitions were monitored: m/z 58 26 (blue) and m/z 6026 (red). ...............................................983-18 Triplicate injections of a second EBC sample monitored by LC-MS/MS. Two transitions were monitored: m/z 58 26 (blue) and m/z 6026 (red). ............................. 99
11 3-19 Clinical sample monitored by SRM 5826 A) water B) standard C) spiked EBC D) EBC ..................................................................................................................................1014-1 Identification of SCNin human oral rins es by capture of the isotopic envelope. .......... 1064-2 Proof of concept: Validation for the identification of 32SCNin human oral fluid .......... 1074-3 Replicate injections of a 10 ng/mL SCNstandard LC-ESI-MS/MS transitions for m/z 58 26 for the verification of the met hod detection limit calculated for SCNby the EPA method. .............................................................................................................. 1114-4 Isotope dilution ESI-M S/MS calibration of SCNsolution standards .............................1164-5 Control chart for the determination of drift in measurement of clinical oral rinse specimens ..................................................................................................................... ....1174-6 Evaluation of carryover from clinical specimens of human oral rinses during chromatographic analysis................................................................................................. 120
12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MASS SPECTROMETRIC STUDIES OF ENDOGENOUS THIO CYANATE IN EXHALED BREATH CONDENSATES AND ORAL RINSES By Frank A. Kero December 2008 Chair: Richard A. Yost Major: Chemistry The detection of specific com pounds in breath has long been considered an important part of clinical diagnostics. Breath analysis was first reported by the ancient Greeks during the time of Hippocrates, when medi cal practitioners considered sp ecific aromas in breath as diagnostic markers of disease. As technology for breath analysis became more sophisticated, over 200 volatile compounds have been detected and their biochemical origins elucidated. Despite the persistent interest throughout history, only seven breath tests have met federal guidelines for implementation in the routine clinical laboratory and law enforcement screening procedures. It follows that developing analytical strategies to e xploit the breath matrix offers great potential in the future of clinical diagnostics. In the 1980s, Sidorenko reported the differences in total protein content when comparing saliva with condensed-phase breath samples. Th is was a particularly exciting study because it suggested that nonvolatile material s in breath, such as high molecu lar weight proteins, could be detected without invasive surgic al procedures or se mi-invasive sputum techniques to obtain samples of airway secretions. The detection of monosaccharides and inorganic ions in breath condensates soon followed and has b een reported at concentration levels on the order of pg/mL ng/mL. Of particular interest to this res earch is the detection of thiocyanate (SCN-), a novel
13 biomarker for cystic fibrosis, since prev ious reports of methods to monitor SCNin airway secretions have provided controversial results. The secretions of the airway were samp led as aerosols during normal tidal breathing using commercially available devi ces that cool expired breath to a condensed sample. Samples obtained by these methods are known as exhale d breath condensates (EBC). A novel LC-ESIMS/MS procedure was developed for the selective detection of SCN-. Method detection limits were determined to be ~10 ng/mL. This is th e first known report on the characterization of SCNunder reversed phase LC conditions. This is also the first known report on the utility of selected reaction monitoring to identify SCN(SCNCN-). Despite the improvement in selectivity, this method did not provide detection limits low enough to allow for SCNto be detected in EBC. Although this method could not be successfully a pplied to EBC, it was successfully applied to the measurement of free [SCN-] in oral rinses (processed sali va) from human volunteers. This method offered several advantages versus previo usly reported methods for this application, including improved detection limits and redu ced sample preparation. Given that [SCN-] in breath is below the limits of detection for this method and that SCNis easily detectable in saliva, it follows that SCNcould be considered a chemical ma rker for saliva contamination in EBC samples. A pilot study was performed to initiat e the evaluation of th is hypothesis. It is anticipated that methods to control the adul teration of breath samples, by saliva or other biofluids, will have significant im pact in the future of clinical diagnostics as testing methods move to noninvasive platforms.
14 CHAPTER 1 INTRODUCTION Exhaled Breath Condensates ( EBC) Background and Sig nificance The detection of specific com pounds in breath has long been considered an important part of clinical diagnostics. Breath analysis was first reported by the ancient Greeks during the time of Hippocrates1. Medical practitioners of this time period consider ed specific aromas in breath as diagnostic markers of disease. For ex ample, when the smell of fish (amines) was recognized, the patient was diagnosed with liver disease. A fruity smell (acetone) was associated with diabetes. The smell of urine (urea) was associated with kidney failure. As technology improved, interest in monitoring specific biochemi cal events in breath increased, and research groups around the world grew in number. In 1784 LaPlace was able to extend the work of Lavoisier through clinical studies of the breath of guinea pigs. He was able to demonstrate that the metabolism of food was detectable by m onitoring oxygen and carbon dioxide levels in expired breath of guinea pigs. He later repeated the experiment on himself, giving birth to the phrase to be a guinea pig1. As technology became more sophisticated, analytical instrumentation became more selective and sensitive. In the 1970s, the marriage of chroma tography and mass spectrometry came into vogue and over 200 volatile organic co mpounds were identified in human breath and reported2. This includes endogenous biomarkers of disease3 and chemical markers of exposure4-15. Given the abundance of compounds that are present in breath, it was very surprising to learn that there have only been seven breath tests that have been approved by the Food and Drug Administration (FDA) for routine patien t monitoring and law enforcement screening16. Table 1-1 gives a description of the breath tests currently in practice.
15 A reasonable question to ask, given the prolif ic nature of the scientific communitys literature on the subject of breath analysis and th e ability to selectively monitor new biomarkers, is why have breath-based analytic al methodologies failed to be impl emented into ro utine clinical analysis? Over time, significant limitations of br eath analysis have been identified but not addressed. These issues include the following: wide ranges of normal analyte concentrations that do not correspond with physiology, lack of standardization, and lack of anatomic specificity, which includes the potential for multiple biological events to produce the same small-molecule metabolite in the breath. Current initiatives in breath analysis are focused on addressing these issues by approaching the breath matrix in a new way16. This novel approach will be discussed in the following section. In the 1980s, Sidorenko reported on the studies of surface active properties/surfactant properties of lung fluid as well as differences in protein content when comparing saliva with condensed-phase breath samples, now more commonly referred to as exhaled breath condensates (EBC) 17. The samples were freeze-dried using sta ndard lyophilization techniques prior to a series of characterization experiments. This was a particularly exciting study, because it suggested that nonvolatile material s in breath, such as high molecu lar weight proteins, could be detected without invasive surgical procedures such as bronchoalveolar lavage (BAL) or seminvasive sputum techniques to obtain samples of airway secretions18. In general, BAL sampling methods are surgical procedures that involve the insertion of a fiberoptic tube19 in a sedated human or animal subject. 50-100 mL of sa line is injected through the fiberoptic tube in to the lung. The saline is recove red and analyzed for cellular co mponents, bacteria colonization as well as small molecules. These methods ar e typically high in cost and restraints to
16 accommodate patient comfort and safety preclude these techniques from studies that require longitudinal sampling. Table 1-1. Summary of current breath based diagnostic methods that have been approved by the FDA, Risby T, Solga S (2006) Current status of clinical breath analysis, Applied Physics B: Lasers and Optics, 85 (2-3), 421-426. Test Analyte Capnography Carbon dioxide Neonatal Jaundice Carbon monoxide Law enforcement screening Ethanol Asthma therapy Nitric oxide Gastrointestinal disaccharide deficiency Hydrogen Heart transplant rejectio n 9 methylated alkenes H. pylori infection Urea A mechanism to explain the presence of these nonvolatile, hi gh molecular weight molecules in breath was proposed20. The mechanism was based on the formation of microaerosols in areas of high turbulence in the lung and trachea. This is the case during normal breathing cycles (tidal breathing) where turbulent airflow causes the airway-lining fluids to be nebulized. A second possible source of nebulized mi croaerosols can be considered when closed respiratory bronchioles and alveol i open during normal ventilation16. Figure 1-1 shows an illustrated representation of this phenomenon. It fo llows that small molecu les of limited volatility like monosaccharides and inorganic ions that prov ide important metabolic fingerprints could also be contained in these microaerosols. The relatively small physical dimensions of these microaerosols and questionable dilution factor in breath (an artifact of the lack of standardization of EBC collection devices as well as variability between people) have led investigators to devel op novel and innovative analytical strategies for their characterization and have been the subject of several review papers21,22. Since the first report on the characterization of nonvolatile ma terials in EBC, the number of papers that have been reporte d in the current literature has slowly grown, as shown in Figure
17 1-2. This figure illustrates that this is a relatively new field of study with a great opportunity for contributors to have an impact in academic fundamentals and clinical applications. It is anticipated that fundamental studies, such as th e characterization of th e physical properties of microaerosols as described above, will support clini cal investigations in the development of EBC based noninvasive diagnostic applica tions. Of particular interest to this collaborati on is a breath based diagnostic method for cystic fibrosis23. Figure 1-1. The generation of microaerosols of surface liquid from areas of high turbulence in the lung and trachea. Reference Effros R, Dunning M, Biller J, Shaker R (2004) The promise and perils of exhaled breath c ondensates, American Journal of Lung and Cellular Molecular P hysiology, 287, L1073-L1080 Figure 1-2. The number of EBC papers published versus time. Each data point represents a restricted Google scholar literature search by year. Ai r -Li q uid Interface Turbulent Air Flow (tidal breathing)
18 Project Motivation for Studying Thiocyanat e in Airw ay Secretions: Cystic Fibrosis Cystic fibrosis (CF) is a progressive and degenerative ge netic disease that presents symptoms in the lungs and digestive system24,25. This disease is characterized by fibroid growths in the pancreas, which accounts for its original name, CF of the pancreas. As the disease progresses, the immune system becomes compro mised and the host is plagued by recurring lung infections. The ability to generate healthy and functioning mucus in the lungs is also compromised. This is particularly troubling sin ce solidified mucus (phlegm) in the respiratory tract causes severe discomfort. The patient soon becomes dependent on caregivers to perform airway clearance to prevent suffocation. The most common of these techniques is known as chest physical therapy (CPT) or conventional physiotherapy (also CPT)26. This airway clearance technique is performed by using cupped hands to clap on the back and ches t. Other symptoms of cystic fibrosis include recurring sinus inf ections, poor growth, diarrhea and infertility 25. The average life expectancy of a ch ild born with CF is approximately 37 years, although continuing improvements with clinical diagnostics and ge netic screening translate to longer life expectancies for a baby born today versus a baby born a few decades ago27. In 2004, The Boomer Esiason Foundation for CF reported that one in every 3,000 children in the United States is born with CF. An important contribution of the Conner res earch group (University of Miami, Miami, FL) to the characterization of CF is the hypothesis 28,29 that all of the components of the lactoperoxidase (LPO) system are important factors in the maintena nce and regulation of bacteria and microbial clearance in the host airway 29,30. This includes LPO expression and secretion, SCNtransport and H2O2 production. A recent study has reported 31 that variation in [SCN-] across the epithelium affects the ability to mainta in the LPO system. Failure to clear microbial
19 colonization results in chronic airway inflamma tion which is common in progressive and chronic diseases of the lung. Other reports from the Conner group suggests a strong link between [SCN-] in the secretions of the airway and the pathogenesis associ ated with cystic fibrosis 32,33. The evolutionary significance of SCNin regards to relevant enzymatic pathways for SCNand the pathogenesis associated with cystic fibrosis are described by the LPO mediated reactions presented below32,33: 1) H2O2 + LPO LPO-O + H2O 2) LPO-O + SCNOSCN+ LPO LPO, the catalyst for the oxidation of SCNto the biocidal OSCN-, is expressed by goblet cells present in the airway epitheli um. In fact, it is hypothesized28-33 that OSCNacts as a primary microbial scavenger in the airway, since it is a highly reactive species an d it is assumed to be energetically favorable for the process to pr ecede in this fashion. It follows that SCNtransport across the epithelium can have a delete rious effect on the generation of SCNand the result is patient symptoms with chronic airway in flammation due to the accumulation of hydrogen peroxide (H2O2). Recent evidence suggests that the cyst ic fibrosis transmembrane regulator (CFTR) protein may also play a role in SCNtransport28-33. Previous analytical stra tegies to monitor SCNin airway secretions have produced mixed results. The application of a bronchoalveo lar lavage (BAL) sampling method yielded no detection of SCN; however, the authors suggested38 that the large analyte dilution factor associated with this clinical sampling procedure may have masked the true [SCN-]. The Conner group responded with the development of an innovative and direct sampling method that allowed for the sampling of airway secretions by suc tioning secretions from intubated sheep. The recovered secretions yielded a [SCN-] = 160 mM If the re sults of this study are interpreted as evidence that SCNis detectable in the secretions of th e airway, it follows that the long term
20 goals of this research collaboration are: 1) de termine if the presence of thiocyanate can be detected in the aerosolized secretions of the airway, sampled as EBC, and 2) if SCNcan be detected, determine if ratios of [SCN-] / [I-] will be different in CF vs non-CF volunteers and therefore serve as a potential biomar ker for CF. The concentration of Iis significant since SCN-and Iare both regulated by the same ion pump in vivo 35 (i.e., the Na-I symporter). It is anticipated that [SCN-] is on the order of 20-50 nM (<10ng/mL) based on a 20,000 fold dilution factor when compared to serum concentrations. If SCNcould be demonstrated as a valid biomarker for CF, the scope of this projec t would expand, since a noninvasive screening platform (with a low rate of false positives) would be an advantage and may have a potential impact in the field of newborn screening. Currently, Scotland leads the world in the CF screening of newborns36 with an immunoassay of dried blood spots known as the immunoreactive trypsin (IRT) test27. Trypsinogen, which is expr essed in the pancreas and transported to the intestine, is then activated to form the enzyme trypsin. Measured levels of trypsin have proven to be a useful indicator of newborns with CF, since it has been observed that newborns with CF have thick mucus obstructions in pancreatic ducts that prevent trypsinogen from reaching the intestine and th erefore blood IRT levels will be el evated in newborns with CF. Positive results must be followed by other testin g methods to confirm the diagnosis, since there is a high rate of false positive results asso ciated with IRT measurements. Perhaps mass spectrometry could do better37. During the course of this investigation, it became apparent that the ease of salivary measurements for [SCN-] could not be ignored. The scope of this project expanded to include a small population pilot study of huma n subjects as proof-of-concept. Validation of single operator figures of merit (i.e. limitations of linearity, freezer stability etc.) will be included to support the
21 measurements of the quantitative study. This wo rk will be reported la ter in chapter 4. The relevance of these salivary measurements to the scientific community will be discussed in the following section. Project Motivation for Measuring SCNin Human Oral Rinses: Background and Significance Thiocyanate (SCN-) is an important endogenous ion that has been the subject of a number of studies in selected biofluids and was recently the subject of an in depth literature review by Valdes38. The quantitative de termination of SCNconcentrations in sel ected biological fluids (e.g., serum, urine, oral rinses, etc.) has been li nked to the health and function of two discrete enzymatic pathways. The first pathway involves lactoperoxidase (discussed earlier in this chapter), which regulates the inhi bition of microbial colonization in the host airway, as well as the oral cavity. The second pathway involve s rhodonase, which regulates cyanide (CN-) that has been internalized by diet39 or environmental exposure38 In this context, CNis converted to the less harmful SCN-. Several studies have been report ed on this pathway to monitor SCNas a biomarker of CNexposure from smoking. It has also been reported that SCNlevels increase with the digestion of certain vegetables th at contain glucosinolates (genus Brasicca)40, fruits, mustards41 and milk39. Recent studies suggest that SCNfrom these foods may play a role in the detoxification of carcinogens42. Although normal concentrations of SCNare present in the body as a consequence of diet and normal physiolo gical processes, when endogenous levels are elevated, symptoms such as arterial hypertyension, vertigo, nasal bleeding and unconsciousness can manifest43. Since SCNcompetes with Iin the biochemical events associated with the thyroid, studies have also been performed to determine signifi cance in the formation of goiter44. In summary, the scientific community has long held an interest in the measurement of SCNin complex biological fluids for a variety of app lications including biomonitoring for exposure of
22 environmental toxins (i.e. cyanide or compounds that degrade to cyanide) as well as general physiology and metabolism. Recent literature shows trends in monitoring salivary [SCN-] as a way to distinguish smokers from nonsmokers45,46,47. The Advantages of Saliva (Oral Rinses) Versus Other Biofluids Saliva is approxim ately 99% water 0 .3% enzymes and 0.3% mucin (which is responsible for the vi scous characteristic)48. A description of the com position of the saliva matrix is given in Table 1-249. Saliva contains a dynamic variety of low-to-high molecular weight compounds of various polarities. Th e relatively low protein content of this biofluid allows for analytical development strategies that are more similar to urine than serum or plasma. In practical laboratory applications, the reader can visualize a testing protocol where 10-20 L of an oral rinse is diluted into 1mL of mobile phase prior to measur ement, a so-called dilute-andshoot method. One would assume the amount of nonvolatile components for oral rinses entering the ionization source is minimal compared to seru m. The low protein content also suggests that analyte-protein binding is not a significant issue and therefore th e determination of free analyte concentrations in saliva is re presentative of the freely circul ation analyte in the host subject50. The noninvasive nature of sample procedures to obtain a saliva sample makes it an appealing matrix since materials are low in cost and re peated samples could be taken from a single volunteer for longitudinal studies with minimal patient discomfort. Methods that have been reported for the collection of saliva include mast ication with rubber bands, or stimulation with citric acid on the tongue50. This report employs the commercial methods of Sarstedt (available from Fisher Scientific) since the University of Florida reported previous success for the study of human subjects when this collection method was em ployed (unpublished study by the Goldberger research group, Rocky Pointe Labo ratories, University of Florida) .The procedure used for this
23 study will be described later in this section, but briefly, a piece of de ntal gauze is inserted in to a volunteers mouth for a designated period of time, the gauze is centrif uged and the recovered fluid is analyzed. The process is painless, rapid and portable, allowing for the screening of selected populations. Table 1-2. The chemical composition of saliva re sponsible for the protec tion of the human host, reference 49. Protective Functions Salivary Components Involved Lubrication Mucins, prolinerich glycoproteins, water Antimicrobial Amylase, complement, defensins, lysozyme, lactoferrin, lactoperoxidase, mucins, cystatin s, histatins, proline-rich glycoproteins, secretory IgA, secretory leukocyte protease inhibitor, st atherin, thrombospondin Growth factors Epidermal growth factor (E GF), transforming growth factor-alpha (TGF), tr ansforming growth factor-beta (TGF-), fibroblast growth factor (FGF), insulin-like growth factor (IGF-I & IGF -II), nerve growth factor (NGF) Mucosal integrity Mucins electrolytes, water Lavage/cleansing Water Buffering Bicarbonate, phos phate ions, proteins Remineralization Calcium, phosphate statherin, anionic proline-rich proteins The Statistical Advantages of Dilute and Shoot Methods The advantages of dilute and shoot analytical methods extend past the lower cost in time and labor for sample preparation. The advantages of reduced sample preparation have been described in terms of statis tical variance by Mitra and Brukh51. Consider the relationship: T 2 = S 2 + A 2 where: T 2 = total variance S 2 = variance due to sample preparation A 2 = variance due to analysis
24 The variance due to sample preparation can further be described by the following relationship: S 2 = h 2 + ex 2 + C 2 + Cl 2 where: S 2 = variance due to sample preparation h 2 = variance due to hom ogeneity of the sample ex 2 = variance due to extraction C 2 = variance due to concentration Cl 2 = variance due to sample cleanup As the variance due to sample preparation decreases ( S 2), the variance of the analytical process is essentially the variance of the anal ytical measurement devi ce. In terms of tandem quadrupole based mass spectrometric instrumenta tion, the inherent resolution elements are conducive to highly accurate a nd reproducible measurements; how ever, limitations have been considered at analyte-dependent mass ranges and concentration levels49. There have been several reports that ha ve determined the concentration of SCNin human oral rinses; however, the mo st relevant study to this collabo ration is the one most recently employed by the University of Florida to study [SCN-] in oral rinses collected from smokers and nonsmokers. The methods of Lundquist were select ed as the gold standard measurement based on its consumption of less toxic materials versus previously reported methods with no loss of accuracy. The details of Lundquists method were reported elsewhere45-47 but briefly: SCNis sequestered from a biological mi xture through the appli cation of a weak an ion exchange solidphase extraction resin (Amberlyst 21). A deri vatization procedure is then employed using the selective chemistry of th e Konig Reactions to generate an aromatic dye that absorbs at 607 nm. The Lundquist methods for the quantification of SCNhave proven an effective testing platform but are not without limitation. The sample preparat ion procedure takes approximately 3 to 4 h to
25 complete (as determined by this investigator). The limit of detection (LOD) associated with this method was reported as 1,000 ng/mL. Laborious sample preparation and poor detection limits are not uncommon for the analysis of SCN38. These two limitations are addressed by the LC-ESI-MS/MS method reported in this work (reported later in ch apter 4). It should also be note d that the Lundquist materials are no longer commercially available. While it was important to consider current methods in practice, simple modifications to these methods would not provide the necessary sensitivity needed for EBC studies. To summarize: the Lundquist methods have been reported as an improved strategy for the quantification of [SCN-] in biofluids; however there are two notable limitations: 1) the sample prepar ation (detailed above) is laborious, requiring hours of work to perform the analyte isolation and separation th rough the selective chemis try associated with derivatization procedures, and 2) detection limits are too high to be useful for the EBC platform. This seems to be a common limitation for methods to monitor SCN38. As stated previously in this work, a survey of the litera ture confirms that se veral previously reported methods to monitor SCNin biofluids have detection limits on the order of 1 g/mL; however, some ion chromatography (IC) methods have been reported with lo wer detection limits38,. A successful collaboration with two local research groups usi ng IC instrumentation could not be established and so this route could no t be further evaluated. Hyphenated Analytical Method Lead to Improved Method Detection Limits In response to the above lim itations, an LC-ESI-MS/MS method was developed. As a multi-stage tandem method, this platform should provi de lower limits of detection as rationalized in Bush53. With hyphenated analytical techniques, the chemical noise (background) levels decrease with increasing stages of analysis (number of hyphenated methods). For example, the methods of Lundquist detail offline sample pur ification by solid-phase extraction followed by
26 derivatization through selective chemistry and subse quent detection by a UV-vis spectrophotometer. Thus, the Lundquist methods pr ovide three stages of hyphenated analytical techniques (i.e. 3 noise gates) that result in detection limits of 1 g/mL45-47. The strategy detailed in this work is based on an LC-ESI-MS/MS plat form that will provide four stages of hyphenated analytical techniques. The LC provides a conde nsed-phase separation. The ESI (electrospray ionization, explained in detail in chapter 3) source separates positive ions from negative ions, the first and second stages of MS analysis are two additional stages of gas-phase separation. This strategy, a priori should provide lower limits of detec tion for the interrogation of clinical samples. The strategy for the development of a typi cal mass spectrometry procedure was reported by Gillespie54 and is described in the first column of Table 1-3. To review, a sample is purified for compatibility with the chromatographic system. The sample is then loaded onto a chromatographic column to allow the separati on and sequential ionizati on, of analytes with subsequent detection by mass sp ectrometry. In practice, this procedure is often developed in reverse, as reflected in the second column of Table 1-3. The parameters of the mass spectrometric detector are first optimized followed by the ionization parameters to enhance sensitivity. From there, chromatographic paramete rs are optimized (i.e., column stationary phase, mobile phase recipe, flow rate and injecti on volume). Since mass spectrometers allow for separation of co-eluting species by m/z the complete separation of the mixture on-column is not a necessity. Rather, the optimization of peak shap e is most important to allow for confidence in the chromatographic peak integration. It is an ticipated that full optim ization of all online parameters will often allow for the minimization or elimination of offline sample cleanup. Since the estimated EBC concentration of SCNwas near the anticipated detection limit20, samples
27 were to be preconcentrated using a nitrogen ev aporator. Sample preconcentration by solid-phase weak-anion extraction was not an option due to ES I compatibility issues with concentrated salt solutions traditionally used in the analyte elution step of these procedures. Table 1-3. A comparison of the intuitive order of a typical analytical experiment compared with the chronology of analytical developm ent reference Gillespie T (Ph.D. 1988) Dissertation: "Concepts for the Determin ation of Prostaglandins by Tandem Mass Spectrometry, University of Florida Order of analysis step # Typical mass spectrometry procedures Method development chronology for this investigation 1 Sample preparation Detection 2 Online separation Ionization 3 Ionization Online separation 4 Detection Sample preparation Overview of the Dissertation The scope of this dissertation is to evaluate the u tility of LC-ESI-MS/MS to a current topic of interest in the field of breath analysis, more specifically, the feasibility of tandem mass spectrometry to measure low levels of SCNin EBC. Measurements of SCNin saliva (oral rinses) were also considered. The measurement of SCNin saliva (oral rinses) is important for this EBC study since saliva is the most likely source of EBC sample adulteration. To further detail the course of this study, ch apter summaries are included here. Chapter 1 has provided background and history in the field of breath analysis and current trends towards the analysis of nonvolatile analytes (such as proteins, sugars and, inorganic salts). The long term goals of this project were iden tified, including a noninvasi ve screening method for the screening of cystic fibrosis by the application of breath-based diagnostics and the potential of SCNas a novel biomarker. A brief review of relevant analytical methods to monitor SCNthat has been provided with emphasis on two limitatio ns: 1) laborious sample preparation 2) poor
28 detection limits. It was emphasized that tande m mass spectrometry have often provided the necessary selectivity and detection limits to measure low levels of selected analytes with reduced sample preparation. Chapter 2 provides an in-depth review of the fundamentals associated with mass spectrometry relevant to this work. The fundamentals of quadrupole mass analyzers and quadrupole analyzers in tandem will be presente d. The optimization of mass spectrometry and tandem mass spectrometry parameters is detailed in accordance with relevant figures of merit. The selectivity and performance of the mass analyzers were investigated to determine if the methods provided the necessary reso lution elements to separate SCNfrom intrusive chemical noise. Chapter 3 reports on the optimization of the selected ionization source for this experiment. A literature review focused on th e ESI source is included. Evaluation of a novel reversed-phase liquid chromatography (LC) column will also be reported. This strategy retains the interferences on the LC column and allows the SCNto elute in the dead volume of the column. Since SCNelutes in the dead volume, further c onsiderations to the effect of matrix suppression were investigated w ith a series of post-column infusion experiments. The chapter will end with the evaluation of real clinical specimens of EBC. Chapter 4 focuses on the application of this novel LC-ESI-MS/MS method to other body fluids, specifically saliva. A validation protocol was developed to quantify [SCN-] in human oral rinses (processed saliva). A human subject study was performed in accordance with the guidelines of the university institution review bo ard (IRB) for the protection of human subjects. [SCN-] was successfully determined in oral rinses from a population of volunteers. Experiments to define single operator figures of merit for validation were developed and reported.
29 Chapter 5 provides an overall summary of th is research and provides suggestions for future work that includes the development of a novel portable instrument for improved sensitivity in breath analysis.
30 CHAPTER 2 MASS SPECTROMETRY AND TANDEM MASS SPECTROMETRY: OPTIMIZATION AND DEVE LOPMENT Overview Quadrupole m ass analyzers are dynamic mass filt ers that play a central role in the characterization of biological materials, yet th ese devices are not wit hout limitation. A balanced knowledge of theoretical and practical considerations for quadrupole mass analyzers and quadrupole mass analyzers in tandem is required to develop effective analyt ical methods that are highly selective and specific. For this reason, a review of the fundame ntals of quadrupole mass filters, including a description of the relevant operational scan modes, is presented prior to discussion of data. The translation of fundamental concepts to experiment is presented as an overview of the relevant commercial instrumentat ion employed in this investigation. Data to support the development process wi ll be presented. The complementary scan modes of the triple quadrupole mass spectrometer (TQMS), the MS/MS in strument selected for this study, proved useful in the determination that there was more than one ion at m/z 58, complicating the interpretation of MS/MS spectra and the subseq uent selection of the appropriate selective reaction monitoring (SRM) transi tions. Further investigation s uggests that the TQMS has the necessary resolution elements to separate SCNfrom the isobaric interfer ence(s). Justification for the choice of SRM transition is provided. Fundamentals of Mass Spectrometry: Quadrupole Mass Filters, Theoretical and Practica l Consideration Quadrupoles are used both alone and in tande m for the isolation ( by mass-to-charge) of ions. These devices may also be used (in RF-onl y mode) to contain ions during collision-induced dissociation (CID)56. Gas-phase ions are generated by a source external to the quadrupole mass analyzer, often at atmospheric pressure57. Biological mixtures are ty pically introduced into the
31 ion source in the liquid phase ei ther from an HPLC column (for LC-MS), by flow injection analysis (with no HPLC column), or by conti nuous direct infusion through a fused silica capillary. For heterogeneous samples, a sample cl eanup or separation technique such as HPLC or solid-phase extraction is typically incorporated into the method development scheme to allow for the sequential introduction of multiple analytes in to the ion source. Simultaneous ionization of multiple analytes, particularly at widely vary ing concentrations, may result in ion suppression and prove deleterious in quantitative investigations57,58,59. The most common ionization source coupled with quadrupole mass analyzers for the i nvestigation of polar an alytes and preformed ions in solution is electrospr ay ionization; the fundamentals57,60,61 of this source will be discussed in chapter 3. The purpose of this review is to provide the reader with a balanced treatment of practical and theoretical cons iderations for quadrupole mass analyzers, and to present the advantages of interfacing quadrupoles in tandem. Theory and Practice of Quadrupole Mass Filters A quadrupole m ass analyzer consists of four el ectrically isolated hyperbolic orcylindrical rods linked to RF (radio frequency) and DC (d irect current) voltages, as described by the schematic in Figure 2-161. The combination of RF and DC voltages creates a region of strong focusing and selectivity known as a hyperbolic (m ore commonly quadrupolar) field. In simplest terms, the ratio of RF/DC allows for the select ive transmission of ions of a narrow range of mass-to-charge from the total population of i ons introduced from the ionization source. The idealized quadrupolar field can be descri bed in terms of Cartesian coordinates ( x and y directions toward the rods, and z direction along the rods axis). Ions of a selected m/z follow a stable trajectory around the center of the field, and are transmitted in the z direction through the device62. Motion of these stable ions in the x and y directions are small in amplitude, as the lowest energy pathway lies toward the cen ter of the hyperbolic field. Ions of other m/z will have
32 unstable trajectories, with increasing displacement in the x and / or y directions away from the center of the hyperbolic field, and thus strike the quadrupole rods, wher e they neutralize upon contact. These ions will not be transmitte d through the quadrupole to the detector. The application of RF and DC voltages creates a regi on of stability for transmission of ions of a limited m/z range. In this regard, the quadrupole analyz er is operating as a mass filter. A mass spectrum is produced by ramping the RF and DC voltages at a constant ratio. Ions of increasing m/z will sequentially achieve stable trajectories a nd reach the detector in order of increasing m/z Figure 2-1. The quadrupole mass filter and its po wer supply, reference Kero F, Pedder R, Yost R (2005) Quadrupole Mass Analyzers: Theo retical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons
33 To understand the behavior of ions in the quadrupole, a brief introduction to the mathematics associated with ion motion is worthwhile. For a co mplete discussion, the reader is referred to the work of March and Hughes60. The motion of ions through a quad rupole is described by a secondorder linear differential equation called the Mathieu equation, whic h can be derived starting from the familiar equation relating fo rce to mass and acceleration, F = ma yielding the final parameterized form, with the following substitutions for the parameters a and q64: (d2u / d 2 ) + (au 2qu cos 2 ) u = 0 au = (8eU ) / (mro 2 2 ) qu = 4eVmr0 22 The u in the above equations represents position along the coordinate axes ( x or y ), is a parameter representing t/2, t is time, e is th e charge on an electron, U is the applied DC voltage, V is the applied zero-to-peak RF voltage, m is the mass of the ion, r0 is the effective radius between electrodes, and is the applied RF frequency in radians s 1. The parameters a and q are proportional to the DC voltage U and the RF voltage V respectively. The analytical solution to this second-order linear differential equation is: which reduces to a similar infinite sum of sine and cosine functions: The solutions to the Mathieu equation can be pres ented graphically, as shown in Figure 2-2 in a so-called stability diagram. Points in (U V ) space (DC, RF voltage sp ace) within the lines lead to stable trajectories for the m/z 219 ion; points outside the lines will lead to an unstable
34 trajectory. The dotted line is called the scan line, and shows the ramp of RF and DC voltage at a constant ratio (the slope of the line). If the line passes just below the apex of the stability region, as shown here, the m/z 219 ions will have a stable trajec tory at those voltages, and will be transmitted to the detector. Figure 2-2. Mathieu stability diagram for an ion of m/z 219 in a quadrupole mass filter with 9.5mm diameter round rods and an RF frequenc y of 1.2 MHz, reference Kero F, Pedder R, Yost R (2005) Quadrupole Mass An alyzers: Theoretical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons Figure 2-3 shows the stability di agram for three different ions, m/z 28, 69, and 219. Note that ions of lower m/z have stable trajectories at lower RF and DC voltage s (i.e., they are stable at points in ( U V) space closer to the origin). The solid scan line again shows the ramp of RF and DC voltage at a constant ratio61. The scan line passes near the apexes of the stability regions for
35 all three ions. As the RF and DC voltages ramp up in amplitude, ions of increasing m/z have stable trajectories and are transmitted, as show n in the solid mass spectrum below the stability diagram. If the scan line has a lower slope (the dotted line), it passes through a larger portion of each stability region, resulting in a mass sp ectrum with lower resolution, as shown as by the dotted spectrum. The selec tion of RF and DC voltages, V and U the RF frequency, and the inscribed radius r0 between the rods determines the performance of the quadrupole mass filter. Typical values are 5 kV RF(0-p), 2 kV DC, 2MHz RF frequency, and 1 cm radius, and yield a mass-to-charge range of ~ 2000. As can be noted from equation (4), the Mathieu q parameter is proportional to the RF voltage and inversely proportional to56,49 the square of the inscribed radius and RF frequency; thus, the mass range of the mass filter can be increased by increasing the RF and DC voltages, or by decrea sing the inscribed radius or RF frequency. The mass resolution is a function of the ratio of the RF and DC voltages (the slope of the scan line in Figures 2-2 and 2-3); increasing the slope to move the scan line clos er to the apex of the stability region (moving from the dotted line to the solid lin e, for instance) increases the mass resolution, as shown in the spectra in Figure 2-3(b). Unfortunately, the increase in mass resolution is accompanied by a loss of sensitivity (decrease in tr ansmission of ions through the mass filter). The ultimate resolution that can be achieved is determined not only by how close the scan line approaches the stability region apexes, but also by the precision and accuracy of the quadrupole power supplies and the quadrupole rod dimensions. An important operational mode of the quadrupole is achieved when U the DC potential, is equal to zero (a = 0 ) This corresponds to a scan line along the x -axis
36 in Figures 2-2 and 2-3. This RF-only mode results in transmission of ions of a very wide range of m/z values, and is thus often termed total ion mode62. Note that in this mode, there is a limit to the lowest m/z ion that can achieve a stable trajectory. This can be seen as the right-hand edges of the stability regions in Figure 2-3. This total ion mode is used for the collision cell in tandem quadrupole mass spectrometers, as discussed below; it is also often employed in mass spectrometer systems to help transmit ions form the ion source region to the quadrupole or other mass analyzer. Figure 2-3. Mathieu stability diagram and resolu tion. (a) Mathieu stability diagram for ions of m/z 28, 69, and 219. Two scan lines are show n, one dotted and a second solid. (b) Simulated mass spectra showi ng these ions as the RF a nd DC voltages are scanned along the two scan lines (dotted and solid) s hown, reference: Kero F, Pedder R, Yost R (2005) Quadrupole Mass Analyzers: Theo retical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons
37 Tandem Mass Spectrometry Most applications of m ass spectrometry in proteomics employ mass spectrometry in tandem with one or more other analytical stages66,67; one of those stages is often a separation stage, typically liquid chromatography or capilla ry electrophoresis, lead ing to LC-MS or CE-MS and another stage may be a second stage of mass spectrometry (the combination of two or more stages of mass spectrometric analysis in series is termed tandem mass spectrometry68,69 or MS/MS, or even MSn, where n can be 2). Indeed, a common approach is combining a stage of separation with tandem mass spectrometry that is LC-MS/MS and LC-MSn. These tandem (sometimes termed hyphenated) methods can of fer dramatically improved selectivity and information to solve tough analytical problems, as often encountered in a wide range of applications. Informing power is a figure of merit that provides a means of quantifying the amount of information available in such an anal ytical procedure. Fetterolf and Yost70 demonstrated the use of informing power to illustrate the advantag es of hyphenated methods such as LC-MS/MS over single-stage analytical methods. No te that MS/MS is particularly important in LC-MS (compared to combined gas chromatography/mass spectrometry GC-MS) for two reasons. First, LC typically provides poorer chromatographic reso lution than capillary GC, meaning that LC separations are often incomplete, with coeluting peaks that MS/MS can help resolve. Second, LC-MS employs ionization methods such as ESI and APCI that provide only molecular-type ions and little or no structur al information (GC/MS, in cont rast, typically employs electron ionization (EI), which provides si gnificant fragmentation and, ther efore, structural information with only a single stage of mass spectrometry). Thus, LC-MS/MS is invaluable in providing structural information for the id entification of components eluting from the LC column and for detecting those compounds with high selectiv ity. MS/MS is readily performed tandem-in-
38 space (i.e., with two mass analyzer s, typically quadrupole mass filte rs, in sequence). Note that MS/MS can also be performed tandem-in-time, with two or more stag es of mass analysis performed sequentially in time in a single ion tr ap such as a quadrupole ion trap. The triple quadrupole mass spectrometer (TQMS, see Figure 24) is the most common MS/MS instrument in use. Figure 2-4. Schematic of a triple quadr upole mass spectrometer (TQMS). Reference: Kero F, Pedder R, Yost R (2005) Quadrupole Mass Analyzers: Theoretical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons The operational principles of the triple quadr upole tandem mass spectrometer can be described in a similar fashion to the single quadrupole. I ons from the ion source enter the first quadrupole mass filter (Q1), which selects ions of a selected m/z and transmits them to the second quadrupole. The second quadrupole ( q2) functions as a collision cell in RF-only mode; CID is accomplished by adding a collision gas, typically nitrogen or argon, at a pressure of around 1 millitorr (110 microbar). The third quadrupole (Q3) another mass filter, provides a means of mass-analyzing the products of CID from the co llision cell. Note that the RF-only quadrupole collision cell in triple quadrupole instruments is often replaced with a higher-order multipole (a hexapole or an octapole, with 6 or 8 rods, respectively)67. The use of an RF-only higher-order multipole provides effective transmission of a wider mass range than an RF-only quadrupole. The stability diagram for a hexapol e or octapole has le ss well-defined stabili ty boundaries, so the
39 low-mass cut-off characteristic of an RF-only quadrupole is not an i ssue with higher-order multipole collision cells. A practical advantage that should be noted is that these multipoles require less demanding tuning procedures for th e optimization of volta ges and frequencies. A fundamental understanding of the scan modes associated with the TQMS is essential for understanding the MS and MS/MS capabilities of the instrument66,68,69,. These are summarized in Table 2-1. First, note that the triple quadrupo le tandem mass spectrometer can readily be used to perform a si ngle stage of mass spectrometry (M S) by setting two of the three quadrupoles into RF-only (total ion) mode. Either Q1 or Q3 can then be used as a mass filter, scanned to obtain a full spectru m, or set to pass a single fixed m/z (or a few fixed m/z values), a mode termed selected ion monitoring (SIM).When Q1 and Q3 are both used as mass analyzers, there are four possible MS/MS scan modes. In the daughter ion scan mode (also called product ion scan), Q3 is scanned to obtain a spectrum of all the daughter ions generated by CID in q2 of the parent or precursor ion mass-selected with Q1. In the parent ion scan mode (also called precursor ion scan), Q1 is scanned to obtain a spec trum of all of the parent ions that upon CID in q2 produce the daughter or product i on mass-selected with Q3. In a neutral loss scan mode, both Q1 and Q3 are scanned at the same rate, but with a fixed difference in mass to detect only those ions that undergo a specific neutral lo ss in q2 such as a loss of mass 18 (H2O). An important MS/MS scan mode for quantifying targeted compounds with maximum sensitivity is the selected reaction monitoring mode (SRM) in which one (o r a few) selected parent-ion-to-daughter ion transitions are monitored by setti ng Q1 and Q3 to pass specific m/z values. SRM is the MS/MS scan mode analogous to the MS selected ion monitoring scan mode (SIM). The relationship between the MS/MS scan modes can be explored by examining the plot of the three-dimensional MS/MS data intensity (on a logar ithmic scale) versus parent ion m/z versus daughter ion m/z
40 for a single compound, as shown in Figure 2-545. Such a three-dimensional data set could be obtained by a series of daughter i on scans of each parent ion form ed in the ion source, or by a series of parent ion scans of each daughter ion. It could even be obtained by a series of neutral loss scans for every possible neutral loss (i.e., by interrogating the surface in a series of lines parallel to the front lin e of the plot). Finally, a series of SRM experiments could measure the intensity of every possible parent iondaughter ion pair (i.e., ever y point on the surface). Note that all of these MS and MS /MS scan modes are readily achieved on a TQMS, and indeed, on most tandem-in-space MS/MS instruments. Note that this is not so on tandem-in-time MS/MS instruments, where the selection of a parent ion precedes (temporally, not spatially) the mass selection of the daughter ion. On tandem-in-time MS/MS instruments such as ion traps, parent scans and neutral loss scans are not available scan modes. Table 2-1. MS and MS/MS operational mode s of a triple quadrupole mass spectrometer, showing the modes for the two mass filter s (Quadrupole 1 and Quadrupole 3) and the quadrupole collision cell (q2), referen ce Kero F, Pedder R, Yost R (2005) Quadrupole Mass Analyzers: Theoretic al and Practical Considerations Encyclopedia of Genetics, Genomics, Prot eomics and Bioinformatics, John Wiley & Sons Scan mode Q1 q2 Q3 Parent (Precursor) Scanninga RF only Fixed Daughter (Product) Fixed RF only Scanning Neutral loss Scanning RF only Scanning b SRM Fixed RF only Fixed a Note that the MS scan modes can be performed by mass selection with either Q1 or Q3, with the other one in RF-only (total ion) mode. b Q3 scanned at a difference in mass corresponding to the mass of the selected neutral. ions that upon CID in q2 produc e the daughter or product ion mass-selected with Q3. To summarize: the quadrupole mass analyzer (and the RF-only quadrupole and multipole collision cells) enjoy a central ro le in the success of mass spectrometric methods in the clinical laboratory. Development of an understanding of the theoretical and practical aspects of quadrupoles can improve ones ability to utilize single and triple quadrupole mass spectrometers
41 effectively to solve important problems in clinic al chemistry. For further reading, see reference 72. Figure 2-5. Plot of the threedimensional dataset generated by electron ionization/MS/MS of cyclohexane. A similar plot of 1-propanol could be found in reference Kero F, Pedder R, Yost R (2005) Quadrupole Mass An alyzers: Theoretical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons Figure 2-6. Overview of the Applied Bios ystems Sciex API 4000 LC-MS/MS. A) Harvard Apparatus syringe pump B) mobile phase re servoirs (raised above the HPLC pumps) C) Perkin Elmer series 200 HPLC pumps connected by a mixing chamber D) Perkin Elmer series 200 autosampler with 100-vial tray E) the PursuitMS column F) vacuum pump (hidden by th e cabinet door). E A B C D F
42 Commercial Instrumentation The fundamental operating principles of th e triple quadrupole mass spectrometer were presented in the previous section. An overview of the specifications of the API-4000 Applied Biosystems Sciex triple quadrupole mass spectro meter is given in Figure 2-6 and will be discussed in this section. The data acq uisition software used was Analyst 3.0.. The quadrupole mass filters have a range of 5 3,000 m/z and a mass accuracy of 0.01%. The vacuum regions are maintained by three independent pumps. A rotary vane pump maintains the pressure at the interface betw een orifice and skimmer. Pressures in this region are usually less than 2 Torr. A turbomolecular pump maintains the Q0 region at a pressure of approximately 8 x10-3 Torr. The high-vacuum mass anal yzer region is held at 2 x10-5 Torr A diagram of the pressure regimes is given in Figure 2-7. The second quadrupole re gion (q2) is a LINAC71-73 collision cell (shown in figures 2-7 and 2-8). The proprietary name LINAC refe rs to a linear accelerator quadrupole cell characterized by a tapered design with a larger op ening in the front of the cell. LINAC cells are typically operated at higher pres sures than other collision cells74-77. This higher pressure creates a collisional focusing effect that allows for the fo cused transmission of a beam of ions from q2 to Q3. One of the disadvantages of earlier tapered de signs was the relatively long transit times that were the result of a spectrum of internal ion en ergies. This was the result of ions moving at different speeds from q2 to Q3. The addition of an axial field on the order of 1 V/m addressed these issues. The advantages of the addition of the axial field include improved transmission from Q1 to q2 and faster switching of RF/DC para meters to eliminate hysteresis (cross-talk) in methods that require multiple reaction monitoring (MRM). It was also determined that this design minimizes chemical reactions during the dwell time for ions in the collision cell74-77. The
43 commercial design of the LINAC cell is given in Figure 2-8. The conical rods (2-8A) are tapered. The axial field is create d by applying a DC potential to ea ch of the T-shaped electrodes (2-8B). Figure 2-7. Illustration of the di fferent pressure regimes of the Applied Biosystems Sciex API 4000. Figure 2-8. The LINAC collision cell used in q2 of the API 4000: A) Coni cal rods B) T-shaped electrodes.Reference: http://www.physics.umanitoba.ca/~ens/ASMS2000LobodaPoster.pdf Optimization of Mass Spectrometry Parameters The for mation of gas-phase negative ions by ESI can be considered some what of an art, since strategies to obtain good sensitivity are not as intuitive as the optim ization of experiments performed in the positive ionization mode77-80. This is illustrated in many areas of the development process including solvent selection. Considerations for the ideal ESI solvent for both positive and negative ESI have been reported by Cech and Enke in a recent review article78. In positive ionization mode, so me percentage (usually ~ 0.5 1.0 % by volume) of acid is typically added to a polar organi c / water solution. The organic co mponent of this mobile phase is typically either methanol82 or acetonitrile81. It is believed that the solvent molecule is first Ionization Source Interface Assembly Rotary vane pump 2 Torr q0 ion focusing Turbomolecular pump1 8 x 10-3 Torr Atmospheric Pressure Turbomolecular pump2 3 x 10-5 Torr Mass Analyzers (Q1-q2-Q3)
44 protonated by the acid, and in turn, the analyte is ionized by the protonated solvent. A key point here is that charge is displaced through the so lvent, and proper selection of the mobile phase helps to maintain a stable spray current. Since the addition of acids in positive ion mode aids the generation of gas-phase ions79, intuitively one would believe that the addition of bases in negative ion mode would produce a similar effect. In fact, this is not the case77. At low analyte concentrations, basic solutions tend to form a ve ry unstable spray due to the absence of charged species in solution, since basic compounds do not ionize to an appreciable extent through electrochemical reduction. Several strategies have been considered to address this issue. Halogenated compounds can ionize under the conditions of electrochemical reduction. Cole reported on the utility of chlo roform as a mobile phase modifier for the analysis of negative ions83. This method was demonstrated as an advantage for analytes that lack acidic sites and thus exhibit weak [M-H]signals. The method re lies upon attach ment of Clions present in electrosprayed solutions of ch lorinated solvents such as ch loroform. Attachment of these chloride ions to form [M+Cl]ions was observed for co mpounds characterized by pKa values as high as 27 (aniline). Fluorinated solven ts such as 2,2,2-trifluoroethanol and hexafluoroisopropanol have also been evaluated81. These modifiers function not only as a means of generating anions to displace ch arge, but also as electron scavenge rs in the source. In negative ion mode, issues of arcing and uncontrolled corona discharge have been determined to limit sensitivity84-86. Electron scavenging gases such as SF6 have also been evaluated, but are not routinely used due to the relativ ely high cost of these materials87. Optimization of the instrumental parameters of the TQMS for SCNwas performed by direct infusion using a Harvard Apparatu s syringe pump at a flow rate of 10 L/min. Figure 2-10 is the spectrum from a direct infusion experime nt under default parameters in negative ionization
45 mode. The standard was 1 g/mL of KSCN in 50/50 acetonitr ile/water. Note that a peak at m/z 58 was not detected but rather a poorly resolved mass spectrum was obtained. This spectrum was particularly interesting since low-mass interferences were not expected (sin ce solvent clusters are less common in negative ion mode compared to positive ion mode). A study to determine the effect of mobile phase composition on the intensity of the response for SCNwill be reported later in chapter 3 (rec all the optimization order: detecti on, ionization, online separation, offline separation presented in Table 1-3) To begin, the mobile phase re cipe was kept as simple as possible to obtain a point of reference for optim ization. An additional advantage of acetonitrile versus methanol is the lower surface tension th at eases desolvation processes. The effect of variation in surface tension on th e current produced by ESI was described in the equations of Kebarle82. Resolution Since there was no discerning of individual ma sses (Figure 2-9), resolution parameters were the first to be optimized. The optimization of the resolution parameters were considered completed when individual mass peaks could be di scerned. With the TQMS operating in Q1 full scan mode, the resolution of Q1 was evaluated at three settings low (Figure 2-10a), unit (Figure 2-10b) and high (Figure 2-10c). Wellresolved mass peaks were only discernable when Q1 was operated with the high setting. Th is suggests that the normal mass calibration procedure with propylene glycol (PPG) does not allow for adequate tuning of mass peak widths at these low masses. The peak width specificatio ns for the unit resolution setting are 0.7 +/0.1 amu. The peak width specifications for th e high resolution setting are 0.5 +/0.1 amu. Quadrupole mass analyzers are usua lly set through the instrument control software to afford a constant peak width of 0.8 to about 1 Da. When operated in the high resolution mode, the API
46 Figure 2-9. 60 MCA full scans (Q1 only) for a 1 g/mL KSCN standard in 50/50 AcCN/H2O in negative ion mode under default system parame ters. The direct infusion flow rate was 10 L/min. m/z m/z = 58
47 Figure 2-10. Comparison of Q1 scans at three re solution settings to determine the appropriate setting to resolve m/z 58 prior to fragmentation a) lo w resolution b) un it resolution c) high resolution. The m/z window was 54.5 61.5 for each experiment respectively. 4000 quadrupole Q1 a peak width of ~0.5 Da (Fi gure 2-10c). Furthermore, the mass calibration places the m/z 58 ion on the correct m/z value. Compound Selective Pa rameters before Q1 W ith the ion(s) at m/z 58 sufficiently resolved in Q1 high resolution mode, the next stage of the development process was the optim ization of compound dependent parameters. The TQMS was first optimized to detect the precursor ion at m/z 58 in the first quadrupole (Q1) at maximum intensity. The second qua drupole (q2) is operated as an RF-only quadrupole ion guide a) m/z = 58
48 that allows ions of all m/z to focus and be passed to the third quadrupole (Q3). Q3 was also operated as an RF-only ion guide allowing all ions mass selected in Q1 to reach the channel electron multiplier. The relevant parameter optimized first was the declustering potential (DP) in the q0 region of the instrument. The DP is the po tential difference applied to the orifice with the skimmer held at ground. The greater the magnitude of the voltage, the more efficiently ions clustered with solvent molecule s are dissociated. The working ra nge for DP was 0 to -100 V for the negative ion operational mode. Based on experiment, the DP selected was 34 V. The entrance potential (EP) is the di fference between the skimmer (gr ound) and the entrance to Q0. The working range is from 10 to +10V with a positive bias applied for positive ions and a negative bias applied for negative ions. Th e optimized EP was determined to be 10 V. After the optimization for these parameters was complete, the precursor ion at m/z 58 was considered optimized and the parameters that are relevant to the fragmentation of SCNwere then evaluated. Optimization of Tandem Mass Spectrometry Overview The ability to identify a single analy te in a complex mixture is a practical problem that most analytical chemists encounter68. The proverbial expression nee dle-in-a-haystack is often used to describe this problem. In MS, high sensitivityand thus high transmission is critical. Selectivity, reflected in mass re solution, is required for discrimination against other components of the mixture. MS/MS provides ev en greater selectivity than MS, as it provides two rather than a single stage of mass analysis. The increase in the am ount of chemical information afforded by tandem mass spectrometry techniques was reported by Fetterolf and Yost67. It was determined in this investigation that the appl ication of MS/MS was not only adva ntageous, but an essential tool in providing a method for the identification of SCNin a biological sample. Prior to the optimization of the relevant MS/MS parame ters, it was verified that the ions of m/z 58 could be isolated from the mixture in a low-mass ne gative ion mass spectral window. Figure 2-10c
49 demonstrated the detection of m/z 58 in a mixture of ions. Figure 2-11 demonstrates the isolation of m/z 58 prior to fragmentation and the second stage of mass analysis. This process allows for the assignment of fragment ions to be considered exclusively from m/z 58. This gas-phase separation is useful for sepa rating co-eluting species in chromatographic columns. The details of the experiment are as follows: 1 g/mL 50/50 AcCN/H2O delivered by a Harvard Apparatus syringe pump by direct infusion at 10 L/min; all mass spectrometery parameters were default except for q0 and Q1 (not e: lower case nomenclature for q indicates that the ion focusing quadrupole is typically not operated as a mass filter); the declustering potential (DP) at the entrance of the skimmer (shown in Fi gure 2-7) was optimized for signal intensity to -34 v (automatic feature of the Analyst software ); Q1 resolution was se t to high; data was acquired as 60 MCA scans (spectral addition); the m/z window was 10-100. Figure 2-11. The isolation of m/z 58 in Q1 using the SIM operational mode. Details of experiment are in the main body text. Defining Tandem Mass Spectro metry Figures of Merit MS/MS is perform ed here using three quadrupoles in tandem (two operated as mass filters and one operated as an RF only collision cell). In the first stage of MS, a selected ion of I n t e n s i t y
50 interest (precursor ion) is frag mented to produce characteristic fragments for the characterization of a specific analyte in a complex mixture. The fragmentation of the precu rsor ion is typically achieved by collision-induced dissociation (CID) or collisionally activated dissociation (CAD). Since ESI is considered a soft ionization sour ce (see chapter 3 for more detail), structural information obtained from ESI is typically li mited and therefore MS/MS is required for qualitative purposes. Gas-phase ions formed by ESI can be fragmented by controlled collision with an inert gas (nitrogen was used for this investigation). Upon co llision, some of the translational energy of the ion is converted in to internal ener gy, and increasing bond vibrations can lead to fragmentation. Below is a general description of CID for a positive ion where Mp + is the precursor ion and N is the collision gas. Ma + and Mb are fragmentation products of the precursor ion. A description for ne gative ions will be included la ter in this chapter (Figure 213).The symbol represents the activated precursor ion after the collision event89. Mp + + N Mp +* + N Mp +* Ma + + Mb The optimization of MS/MS parameters focused on the selection of proper settings for collision energy and pressure. Consider the equation E com = E LAB [m target/ (m ion + m target)] where E com = the energy of the ion in center-of-mass coordinate, E LAB = the kinetic energy of the ion in the lab frame (the collision energy), m target = mass of neutral collision gas, m ion = mass of the precursor ion. The maximum internal en ergy that can be depos ited into an ion in a collision is Ecom. Note the internal energy of the Mp + ion after collision is not a single value but a distribution of values, with a maximum governed by this equation. The internal energy of the Mp + is also a function of experime ntal conditions such as the ener gy of the analyte, molecule or
51 ion before ESI and the internal energy gained dur ing the ESI process, and changes in the internal energy during the transit from atmospheric pre ssure to vacuum. From the equation above, it follows that the collision energy required for e fficient collision-induced dissociation generally increases with the mass of the ion to be fragment ed. Precursor ions with too little internal energy produce few if any product ions, and those in low abundance. Neutraliz ation and scatter can compete with CID, resulting in a decrease in the transmission of the precursor ion. These competing processes are present influenced by both the levels of collision energy and collision cell pressure. When these instrument parameters are not optimized prope rly, CID will not occur efficiently. The optimum collision energy for CI D depends upon a number of factors, including the mass, charge, composition and internal energy of the ion to be fragmented and the mass of the collision gas. Collision energy and collision pressure were optimized using fragmentation efficiency as a figure of merit. Fragmentation efficiency (Ef)86,87 is the fraction of ions present following CID which are product ions, sometimes referred to as daughter ions (D), as opposed to remaining precursor ions, P, and can be calcu lated using the following equation. Ef = Di / (P+ Di) Table 2-2 Ef for versus variation in collision pressure at three different collision energies for m/z 58. Note that the pressure reco rded is the chamber pressure outside of the collision cell. The actual collision cell pressure is not measure d, but will be higher. The figure shows that the fragmentation efficiency for the ion(s) at m/z 58 is ~40% under optimal MS/MS parameters. Since SCNis such a low mass ion, it follows from th e equations presented here that low mass ions are anticipated to obtain higher internal energies upon collision with N2 collision gas. One would anticipate that the increa se in internal energy would result in a good abundance of
52 Table 2-2. Intensities of pr ecursor and product ions of m/z 58 at 45 eV collision energy versus increasing chamber pressure (CAD). MS/MS figures of merit were defined in the main body text on the previous page. Chamber pressure x 10-5 torr m/z 58 m/z 32 m/z 26 Ef (%) Ec(%) ECID(%) 0 8,923,233 3,362 3,968 0 1.0 0.00 1 170,630 5,165 3,751 5 1.1 0.05 2 138,747 12,596 3,288 10 1.1 0.11 3 138,887 38,214 2,313 23 1.3 0.29 4 130,713 54,616 2,226 30 1.4 0.43 5 102,350 65,942 1,790 40 1.7 0.66 6 87,873 29,597 1,609 26 1.4 0.36 7 105,363 49,051 2,331 33 1.5 0.49 8 56,013 15,946 1,337 24 1.3 0.31 9 48,632 5,183 1,549 12 1.1 0.14 10 44,108 1,148 1,109 5 1.1 0.05 fragment ions; however this was not the case si nce a low abundance fragmentation pattern was observed. Table 2-2 is interesting in that the trend of the calcula ted fragmentation efficiency is atypical, reflecting the atypical nature of this low mass target analyte as described by the corresponding equation. Recall that three physical processes have been reported to cause a decrease in the precursor ion intensity: 1) frag mentation 2) scatter 3) collisional cooling. The fragmentation efficiency of an ion is expected to increase as pressures increase. Figure 2-12 was included to support the findings of this study in defining this MS/MS figure of merit for SCN-. Note the narrowing of the precursor ion at m/z 58. This spectral symptom is indicative of over resolved peaks (too many RF cycles) which may be the result of the ions moving too slowly through the third quadrupole, perhaps an artifact of the Q3 offs et. This parameter is not typically considered a compound selective pa rameter and it is generally recommended that this parameter should not be changed, however this data may indi cate that future considerations should include studies varying this parameter w ith observations to the effect of the product ion spectrum. It is
53 also possible that the spectrum is an artifact of the transmissi on of the ion population from q2 to Q3. Figure 2-12. Product ion scans m/z 58: The effect of collision cell pressure on the abundance of fragmentation from the precursor ions at m/z 58. For all experiments, Q1 resolution is set to high, Q3 is set to unit. The CID behavior could furthe r be described by the overall CI D efficiency as defined by the following equation: %ECID = ( Di / Po) = Ef x Ec x 100%
54 Where ECID = overall CID efficiency; Di = sum of the intensities of the product ions and Ef was defined previously. Figure 2-14 presents an overall summary of the MS/MS mechanism for SCN-.The discussion of this mechanism was presen ted previously in this chapter (page 66) 86. MS/MS of an SCNSolution Standard: 2 Step Process 1) Activation (fast 10-15 to 10-14 seconds): SCN+ N2 SCN-* + N2 2) Unimolecular dissociation of excited precursor ion: SCN-* S + CNSCN-* S+ CN Figure 2-13: Proposed MS/MS CAD mechanism for SCNFigure 2-14. Selectivity: Comparison of th e LC-ESI-MS/MS selected reaction monitoring (SRM) transitions for chromatograms of a) m/z 5832 b) m/z 5826. Experimental parameters are detailed in main body text.
55 Justification of the Selection of th e Appropriate SRM Transition The MS/MS transition from precursor to product ion is generally accepted to be a highly informing method for qualitative analysis. To furt her substantiate this cl aim, previous studies have employed monitoring multiple transitions from one precursor ion. In cases where fragmentation is extensive, such as electron ionization (EI) methods, three transitions are commonly monitored (the three ion rule88). A figure of merit known as the branching factor (Rx) has been developed to describe the sp ecificity of a method for two transitions92. Since SCNis so low in mass, there are only two transitions to choose from (shown in Figure 2-13). From the experiment described below, it was determined that one of these transitions was not selective for SCN-. The study was performed under the mass spectrome tric parameters as follows: Curtain gas = 10 arbitrary units, source gas 1 = 40 arbitrary un its, source gas 2 = 50 arbitrary units, IonSpray voltage = -3.0 kV, interface heater = on, temperature = 550C chamber pressure = 5 x 10-5 torr, declustering potential V, entrance poten tial = -12V, collision energy = 45 eV, cell exit potential = -15 V, Q1 resolution = high, Q3 resolu tion = unit. Figure 2-14(a) demonstrates that a continuous signal for an injection of SCNstandard when the MS/MS transition is specified to monitor SCNScorresponding to m/z 5832 for this reversed-phase LC experiment. The continuous nature of the signal s uggests interference in the mobile phase is present that has the same transition. The appropriate transition that demonstrated selectivity for SCNwas the transition for SCNCNthat corresponded to m/z 5826. This was demonstrated in how Figure 2-14(b). The asymmetric LC peak shap e inspired additional development experiments that are reported later in the chapter. In summary, m/z 5826 produced one LC peak while the transition for m/z 5832 shows continuous intensity where the peak is not indicating signal contribution from the mobile phase. Th e interference at the transition of m/z 58 could possibly be
56 [CH3CNOH]or [CH2CNH2O]since either ion upon collision could potentially provide a neutral loss of 26 and pr ovide a product ion at m/z 32. Summary and Conclusions The funda mentals of mass spectrometry and ta ndem mass spectrometry were presented in terms of a graphical approach. This discussion tr anslates mathematical concepts that may be foreign to the novice reader (e.g. a and q space) to the more pedestrian units of voltage and frequency. This approach was used to explain the fundamentals of resolution and MS peak generation. The discussion closed with an overvie w of commercial instru mentation that includes a detailed explanation of the complementary scan modes of the TQMS. Of particular relevance to this work was the explanation of selected reaction monitoring (SRM). The translation of fundamental concepts to the laboratory was pr esented in through the method development process. Some issues with the conventional PPG calibration and tuning procedure may have been discovere d since there were issues with achieving unit resolution at the low mass region in negative ion mode. This issu e was circumvented by operating Q1 in high resolution mode. In order to in crease the method sensitivity, Q3 was operated at unit resolution. The reasoning to justify this in strument configuration is the majority of interferences are removed prior to the collision cell, either by the LC column or by Q1. It is therefore reasonable to operate Q3 at unit resolution to improve transmission. The instrument optimization continued detaili ng the compound selective parameters that were determined for this investigation. The MS/MS development process detailed the optimization of collision energy and chamber pressure with the appropriate figures of merit. It was determined that controlling the collision energy and pressure resulted in satisfactory fragmentation of the precursor ion; however, th e collection efficiency of the product ions was very poor and therefore limits the sensitivity of this method. Whether this is an inherent
57 limitation of the transmission from the LINAC collis ion cell to Q3 or from Q3 to the detector could not be determined at this time. Future studies to answer this question may consider evaluating the voltages applied to Q3 as an offset to improve transmission of the product ion beam to the detector. The development of an SRM method for SCNproved to be an interesting study. It was clearly demonstrated through the comparison of two different SR M transitions that MS/MS is not only an advantage for monitoring SCNbut a necessity. With the MS and MS/MS parameters optimized, the source and LC parameters will now be addressed in the next chapter.
58 CHAPTER 3 METHOD DEVELOPMENT AND OPTIMIZA TION OF ELECTROS PRAY IONIZATION SOURCE PARAMETERS AND LIQUID CHROMATOGRAPHY Overview The electrospray ionization source (ESI) was selected for the analysis of SCNbecause of well reported studies that ESI e fficiently generates gas-phase ions from ions that are preformed in solution. The source was operated in negati ve ionization mode. The fundamentals of the electrospray ionization (ESI) source will be pr esented in this chapter along with special consideration to the mechanisms of gas-phase ion formation. An overview of the commercial instrumentation used in this study will also be provided. A second goal of this chapter is to present the reader the deta ils of the development of the chromatographic parameters used in this e xperiment. The column that was selected for evaluation was the Varian PursuitMS It was an ticipated that the selectivity of the pi-pi interactions afforded by the columns diphenyl stationary phase would provide favorable conditions to slightly retain SCNunder reversed-phase conditi ons. This would allow for ESI compatibility and subsequent mass analysis It was determined, however, that SCNeluted in the dead volume of this column. Since SCNeluted with other ions, c onsiderations of matrix and ionization suppression will be presented. Review of Atmospheric Pre ssure I onization Sources Ion production at one atmosphere has an importa nt practical advantage, in that it allows for the easy exchange of ion sources without disturbing the vacuum of the mass analyzer region88. Refer to Figure 2-8, which illustrates th e different pressure regimes of the mass spectrometer used in this research as well as the ionization source options amenable for atmospheric pressure. There are several atmosphe ric pressure ionization techniques that are amenable to online chromatographic separation94. Two options that dominate current literature
59 are atmospheric pressure chemical ionizati on (APCI) and electrospr ay ionization (ESI). Atmospheric pressure photoionization95 has recently generated strong interest since the technique is characterized by similar advantages of APCI, with less background inte rferences due to higher specificity. The fundamentals of AP PI will not be reported here, si nce a survey of the literature indicated that this type of ionization method has had little impact on the field of EBC research, most likely due to the polarity of the target analytes.12,13,15,16. In APCI, a liquid-phase solution of mob ile phase and sample is pumped through a nebulizing region held at temperatures of 400 C or above96,97,98. The nebulized droplets are vaporized and the analytes of the injected mixt ure enter an atmospheric pressure region where the solvent molecules are ionized by a corona discharge. In positive ion mode, analytes are typically protonated by reaction w ith the ionized solvent. In negative ionization mode, [M-H]ions are generated by deprotonati on by negative solvent ions with higher proton affinity. APCI offers relatively high ionization efficiencies due to the large number of collisions at atmospheric conditions that facilitate proton transfer from solvent to analyte. A second reason for the efficiency of APCI is that it is less susceptibl e to the analyte suppression mechanisms that are associated with ESI (more on this in the next section of this chap ter). ESI and APCI are complementary techniques, as reflected in the ph ase in which charge transfer processes occur99. Ionization for ESI occurs in the liquid phase, wh ereas ionization for APCI occurs in the gas phase. Both APCI and ESI are suitable ioni zation sources for a quadrupole mass analyzer because the ions are generated and transmitted as a continuous beam of ions. The most common ionization source coupled with quadrupole mass analyzers for proteomics studies is electr ospray ionization (ESI), since it is ideal for compounds
60 such as peptides and proteins that can be ionized in solution. The fundamentals of ESI ionization mechanisms will be detailed later in this chapter as well as elsewhere100. A key advantage of ESI is the ability to impose multiple charges on an analyte upon ionization. Recall that ions are separated in mass analyzers as a function of their respective mass-to-charge ratios ( m/z )88. It follows that multiple charging reduces the mass-to -charge ratio of ions, and thereby extends the mass range of the instrument. Atmospheric pressure chemical ionization (APC I) is a highly efficient alternative to ESI for less polar compounds. Note that APCI produces only singly charged ions. Another recently developed ionization method is matrix -assisted laser desorption/ionization101 (MALDI). Since MALDI with a pulsed laser typica lly produces narrow pulses of ions it is a far better marriage with pulsed mass analyzers (such as time-of-flight mass analyzers) than with continuous mass analyzers such as quadrupoles. Furthermore, MA LDI yields almost exclusively singly charged ions, and thus benefits from the essentially unlimited mass range of the time-of-flight mass analyzer. Since the target compounds of inte rest in EBC are mostly polar compounds or preformed ions, such as SCN-, APCI was not considered a suit able ionization source for these studies. Electrospray ionization (E SI) was selected due to its well documented ability to generate gas-phase ions from ions that are preformed in solution100. The fundamental principles of the ESI process are described in deta il in the next section. Fundamentals of the Electrospray Ionization Source In the m id 1990s the fundamentals of electrosp ray ionization mass spectrometry were the subject of intense debates at the conferences of the American Society for Mass Spectrometry (ASMS). John Fenn, who won the Nobel prize for pioneering this technique, likened the participants in these debates to the blind me n in the classic poem, The Blind Men and the
61 Elephant, each man describing the elephant in a completely diffe rent manner depending on the body part he was touching103 (partially in the right but completely in the wrong). The debates that Fenn was describing focu sed on the two key processes involved in electrospray: 1) the production of charged droplets and 2) the pr oduction of gas-phase ions from these charged droplets. The mechanism for the formation of charged droplets is based on electrophoretic motion, partially in the condensed phase, partially in the gas phase. It was noted that this mechanism has not been accept ed by the entire mass spectrometry community100. The importance of understanding the fundamentals of gas-phase ionization was documented by Tang and Smith105. It was suggested that a complete unders tanding of the mechanisms by which ions enter the gas phase will lead to the developmen t of improved instrumentation. While the majority of the review presented in this chapter is fo cused on the production of charged droplets, it is important to include current pers pectives on how the analyte ions enter the gas-phase. At present, it is still unclear whether gas phase chemistry alters the mass spectrum produced by an analyte100. This chapter will present current perspectiv es on how analyte ions enter the gas phase by describing both the ion evaporation model and the charge residue model. In general terms, the electrospray ionizat ion process involves in jecting a solution of analyte and solvent into a capilla ry held at a constant voltage such that the fluid, under the influence of an electric field, will be pulled towards a large plan ar counter electrode in the form of a disperse plume of small droplets100. Electrochemical reactions at the capillary are responsible for the charging of the droplets and the maintenance of a continuous current106. The droplets function as a delivery vessel for the an alyte prior to it reac hing the gas phase. The general processes in the ESI source are usuall y described for the production of positive ions. Note that for negative ions, the bias of the power supply will be reversed.
62 The properties of the solvent used to deliver the analyte in to the ionization source have several effects on the mechanisms of ESI 81,102. It has been shown that hi ghly polar solvents lead to more multiply charged analyt e ions, an advantage for high molecular weight compounds such as peptides and proteins. By multiplying the ch arges on an analyte ion, the mass-to-charge ratio is decreased to within the working limitations of a practical quadrupole mass filter. Solvent properties will be discussed in more detail in relevant sections of this review. In general, low flow rates of high conductivity solvents generate the smallest droplets82,97, 100. It is advantageous to make the smallest droplets possible, because a high ion yield is favored by the formation of small droplets and therefore a high transmission of i ons is achieved. Enke further explained this phenomenon based on the charge partitioning model and the competition for surface charge110. The analyte is carried by th e selected solvent through a small diameter metal capillary held at a constant biased voltage. In the positiv e ionization mode, the voltage is held positive, and multiple protonation of the analyte is possible96. In the negative ionization mode, the voltage is held negative and removal of multiple protons or attachment of multiple anions is possible. A typical voltage for the capillary is between 3 kV and 6 kV. This voltage generates an electric field described by the equation97,100: E = 2Vc/ [Rln(4d/R)] where E = electric field; R = capillary outer radius in mm; Vc = the applied electric potential in kV; d = distance from the capillary tip to the co unter electrode in mm. Typical electric field strengths are on the order of 106 to 107 V/m. The electric field emitting from the capillary serves several purposes. If the mechanis m of charge separation is indeed electrophoretic, the strength of the electric field would directly influence the initial charge separation in the liquid phase. Descriptions of this mechanism begin with the el ectric field as it penetrates the surface of the
63 liquid, influencing the positive and negative ions present in solution. In positive ionization mode ESI-MS, negative ions are attracted towards the capillary, whereas positive ions are attracted towards the counter electrode100. Highly polar solvents stabilize this charge separation. The opposite effect occurs in the ne gative ionization mode, where posit ive ions are attracted towards the capillary and negative ions are attracted towards the counter electrode. Excess unipolar charges at the tip of the electr ode are electrically attracted to the counter electrode, causing the liquid emerging from the capillary to elongate in the direction of the counter electrode. The analyte is initially retained in the liquid phase because the impose d electric field is balanced by the surface tension of the solution100. Balancing these forces changes the elliptically shaped fluid into the dynamic Taylor cone102,103. The electric field can also be described including the geometry of the Taylor cone by the equation97: Eo = [(2 cos49 / oRc)] where Eo = electric field; = the surface tension of the liquid; cos49 = half-angle of the Taylor cone; o = permittivity of vacuum; Rc = radius of the capillary. Note that the minimum voltage required to form the Taylor cone increases with increasing surface tension of the liquid. After the cone draws tightly over a flight path of one milli meter, a fine jet of charged droplets begins to appear. The least stable point at the filame nt undergoes electrohyrodynami c disintegration when influenced by the electric field.106 The bias of the capillary (positive or negative) determines the bias of the droplets. Since there is a relative ly large endothermic loss due to evaporation and subsequent gas expansion into vacuum, the electric field promotes a warming of these droplets by forcing collisions which generate translational energy that is converted to internal energy. In this way, the electric field prevents the droplets from freezing97,106.
64 To summarize the operational principles of the ESI: there is a potential source maintained at a constant voltage on the order of 3-6 kV106; there is a counter elec trode held at or near ground; and in the space in between the two is an induced electric field. In this way, the ionization source can be thought of as a special kind of electrochemical cell. The cell is not typical for two reasons: 1) ion migration from wo rking electrode to counter electrode is in the gas phase and 2) typical electrochemical cell pot ential differences are on the order of millivolts or volts, whereas the potential acro ss the ESI configuration is on th e order of kV. Given that the ESI source is a special kind of electrochemi cal cell, one would expect some kind of reduction/oxidation chemistry at the metal-liquid in terface (capillary tip) typical of standard twoelectrode electrochemical cells114. Since electrochemical processes are occurring at the capillary tip, does the conversion of capillary material contribute to the ion transmission in electrospray? Blades et al. attempted to answer this question111. As previously noted in literature, only electrons can flow through the wires of an electro chemical cell; it follows that there must be some electrochemical oxidation process yielding these free electrons. If the ESI source is functioning as an electrochemical cell, these processes must be occurring. In electrochemical cells, these reactions occur at the metal-liquid interface on the or der of 10-300 angstroms from the electrode surface115. The metal-liquid interface in the ESI s ource is the capillary tip. To force an oxidation reaction at the metal-liquid interface Blades used a capillary material with a low reduction potential, zinc (Zn)111. The subsequent mass spectrum s howed the presence of Zn ions, thereby supporting that the re actions involving the capillary material occurred. The counter electrode in the ESI source is thought to function as charge balance by reduction reactions. Blades also demonstrated the possibility of the formation of low concentrations of iron ions from
65 stainless steel capillaries. Ions from the capillary material contribute to the total current produced from the ESI source. The ion current produced by ESI is a result of the continuous formation of unipolar droplets, redox reactions at th e capillary tip, and possible coro na discharge. The continuous production of charged droplets can be mainta ined because the electrophoretic separation effectively decreases the tendency for counter ions to remain attached to the multiply charged analyte in the subsequent gas-phase ions. The production of these char ged droplets generates a continuous current in the range of 0.1-1uA107. The magnitude of the current may also be dependent on the presence of ions in solution due to electrolytes dissolved in solvent. The current, when operating in cone-jet m ode (flow = typically on the order of L/min; concentration < 10-3 M) is described by the following equation100: I~ constant x ( Kf /Eo)1/2 where I = total droplet current, constant = 18 for water, methanol, acetonitrile, formamide (high dielectric constants), = surface tension, Kf = conductivity, = permittivity of solution, Eo = permittivity of vacuum. Note that the current incr eases with the concentration of total ionized electrolyte and also with increasing conductivity. Contributions to th e electrical current may also be due to corona discharges at the capillar y tip, so care should be taken in planning an experiment to minimize these interferences. There have been many attempts to descri be the electrohydrodynami c disintegration of liquids in to the unipolar charged dr oplets that generate this current79, 82,3792. Thus, it is informative to consider the physical and chemical properties of individual dr oplets. Recall that at the Taylor cone apex, the charge density on th e surface of the liquid phase is such that the
66 surface tension can no longer hold together the emer ging fluid. The droplet jets fission from the Taylor cone, forming a plume. In an attempt to understand the chemical changes at late stages of the droplet plume, Zhou et al.91 used laser-induced fluorescence spectromet ry. Experimental data were generated to compile a pH profile of this charged plume. The purpose of this study was to further the understanding of the chemistry involved in the electrospray process in order to improve efficiency of ion transmission. Profiles of pH changes were generated by measuring the emission of a pH-sensitive dye (SNARF). As the solvent ev aporates, the ions become more concentrated in each droplet. The pH profile of the plume downstream is dependent on the initial value for pH and the initial polarity of the solvent, and both decreases and increases in pH were observed at different capillary voltages. The la rge decreases in pH may be due to solvent evaporation, droplet cooling, droplet subdivision or heterogeneous subdivisions of charge within a single droplet. This study was performed in positive and negative mode and noted that polarity alone was not enough to explain changes in pH111. Investigations into the phys ical properties of individua l droplets have also been reported97,100-103 Not surprisingly, the two pr operties that are of partic ular interest are droplet charge and radius. These properties greatly in fluence the subsequent gas-phase ionization mechanisms that will later be described in th is chapter. The elongati on and deformation of a parent droplet to yield a series of droplets of solution with decreasing radius. Solutions with high conductivity (10-4 to 10-2 ohms-1 m) at low flow rates (a few L/min) produce the smallest droplets (a few m radius). Olumee et al.109 investigated the dynami cs of methanol-water droplets similar to droplets pr oduced during ESI. The experiment al data were generated using two-dimensional phase Doppler anemometry. The results of this study showed that, as the
67 conductivity of the solution increase s by 3 orders of magnitude, droplet size decreases from ~7 to ~2 m. The higher conductivity solutions showed smaller initial droplets at the capillary tip. It was also demonstrated that the increase of droplet size was po ssible. The gradual increase in droplet size was due to the coalescence of the sm aller droplets. It is wo rth noting that droplet radius has also been shown to increase propor tionally to an increase in solvent viscosity97. It has been shown that after th e initial break away from the apex of the Taylor cone, the droplet radius begins to decrease due to evap oration if high c onductivity solvents are used. As the solvent evaporates, the droplet radius decreas es. The charge on the droplet remains constant, so there is increasing electrostatic repulsion at the surface. Classical molecular dynamics explains that coulombic repulsion within individual droplets forces the charged species to maximize the distance between themselves, so it follows that the charged species would accumulate on the droplet surface110. When the force of this re pulsion becomes equal to the intrinsic surface tension, the Rayleigh limit is reached22,41: Q = 8 ( o R3) 1/2 where Q=excess charge; R=droplet radius; o=permittivity of vacuum; =surface tension. Lord Rayleigh (1842-1919) was interested in the behavior of charged drops and ice crystallines in clouds100. This equation was the result of his investigations into capillary waves on liquid jets and charged water drops. To further discuss the formation of gas-phase i ons from charged droplets, it is of interest to note that the driving force behind the formation of smaller and smaller droplets by droplet fission is an attempt at achieving stability. Dr oplet fission causes a re duction in coulombic energy by spreading the same charge over a relatively larger area, creating a temporary state of stability until further evaporation of the solvent brings the system to the Rayleigh stability limit
68 and the process occurs again. The solvent will evaporate from the progeny drop until the Rayleigh limit is reached and the process will repeat. Just how many times can this fissioning pro cess repeat? Gomez and Tang suggested this process can repeat for a total of 32 fission ev ents before the field desorption limit is reached102. Cech and Enke suggested that this proce ss can repeat 16 times, because after the 16th fission event, it is unlikely for the progeny droplet to produce ions. This also may be due to the possibility of ions ejecting pr ior to complete desolvation (see the discussion on the ion evaporation mechanism later in th is section). A second reason may be that charged droplets in the presence of a high electric field do not maintain a static shape100 and therefore uneven fissioning processes may be possible, generating offspring droplets with an uneven distribution of charge and mass. The effect of uneven fissi oning may occur because an analyte may have an aversion to the surface of the drop (or an attraction to the center of the drop). If this occurs, the analyte may be left behind in the fission proce ss. Taflin demonstrated that the accumulation of charges on the surface of the drop reflects the chemical composition of the progeny droplets113. If the analyte is not on the surface of the droplet, it could remain in the pa rent droplet after the fissioning process. As the series of coulombic fissioning events progresses, the system will eventually reduce to final small droplets th at have only a single analyte i on (the single analyte could have multiple charges). As the last solvent molecules evaporate, the residual charges settle on most stable part of the analyte. The analyte retains these charges and forms a free gas-phase ion. This is the charge residue model (CRM) proposed by Dole et al.114 The ion evaporation model (IEM) is slightly different. In this model97, it is theorized that the coulombic fission series progresses to a poin t where the droplet radius is very small (10-20
69 nm), at which point the imposed electric field can exert such force on the analyte ion as to eject the ion directly into the gas phase prior to the co mplete disintegration of the parent droplet. This means that the gas-phase ions can be emitted prior to the Rayleigh limit. The analyte will enter the gas phase partially solvated. Francisco De La Mora showed that large multiply charged species proceed via Doles charged residue model114. His work with globular proteins showed that ion formation from proteins with masses 3.3 kDa and larger proc eed by the CRM. This limit of 3.3 kda has been cited as the cutoff between the CRM and the IEM. To summarize, current perspectives on the fundamentals of electrospray ionization mass spectrometry offer insight into the two key questions: 1. How are charged droplets formed? 2. How are gas-phase analyte ions ge nerated from these droplets? Electrochemical processes at the metal-liquid interface form a continuous current of charged droplets, creating an environment similar to that of an electrochemical cell. The imposed electric field distorts the liquid th at emerges from the capillary tip, forming the dynamic Taylor cone. Droplets fission from the ap ex of the Taylor cone in to reduce the concentration of surface charge. After more of the solvent eva porates, the process repeats, forming smaller and smaller droplet s. The electrophoretic separati on of analyte and counter ions at the capillary tip may prove to be the justification for the unipolar nature of these droplets. The accumulation of charge has been shown to occur on the surface of the droplet, so uneven fissioning of mass and charge could occur if the nature of the analyte has an aversion to the surface of the liquid. Progeny droplets are the vehicl e to form gas-phase analyte ions by either the charge residue model, or the ion evapora tion model. Large molecules (>3,300 Da) appear to
70 follow charge residue model, whereas small molecules (< 3,300 Da) follow the ion evaporation model97,100. Commercial Instrumentation The Applied Biosystem s Sciex TurboIonspray source was used for th is investigation. The commercial design is essentially a pneu matic ESI source as described by Bruins119 that offers very efficient desolvation due to the high temperature of the ceramic heaters and therefore tolerance of high flow rates to accommodate a variety of HPLC columns. An overview of the source is given in Figure 3-1. Figure 3-1. TurboI onSpray source. The working parameters of the instrument can be described as source and gas parameters, compound selective parameters, resolution parameters and detector para meters. It is of interest to note that parameters that affect sensitivity are strongly dependent on flow rate. As flow rate increases, the most critical source parameters are those altering desolvation and these are the parameters that were chosen for this investigatio n. Table 3-1 illustrates ty pical flow rates chosen for columns of a specific internal diameter. Since the column obtained for the revers ed phase separation had a 2.1 mm internal diameter (length = 100 mm), a 200 L/min flow rate was chosen. Va riation of the flow rate was Horizontal Vertical Waste Neutrals to Waste Ceramic heaters max temp 600 C Spray Capillary Ionization source housing Orifice leading to the MS/MS
71 investigated; however, no increase in sensitivit y was observed. The ioniza tion parameters that were strongly affected by the selection of flow rate were curtain gas (CUR) pressure, the X,Y positioning of the spray capillary in relation to the orifice and temperature of the ceramic Vshaped heaters. The average of three LC-ESI-MS/ MS peak areas was measured and recorded at X, Y position. The data is plotted as a spatial intensity surface map in Figure 3-2. A total of 150 injections (20 L loop injections) were performed to ensure the optimum setting (X = 3 mm, Y = 2 mm). The manufactures recomme ndation for probe position at diffe rent flow rates is given in Table 3-2. Table 3-1. Reported optimal values for flow rate based on the in ternal diameter of an HPLC column. Internal Diameter (mm) Flow rate ( L/min) 4.60 1,000 2.10 200 1.00 50 0.30 4 0.15 1 Table 3-2. Recommended operational parameters for the positioning of the ESI capillary relative to the orifice at various flow rates. Information obtained from the API 4000 operators manual. Flow rate ( L /min) Vertical positioning (mm) Horizontal positioning (mm) 5-50 10-13 3-5 200-500 2-5 3-5 500-1,000 0-2 3-5 The maximum recommended operating temperat ure of the ceramic heaters was 600 C; thus a temperature of 550 C was selected for most efficient desolvation bu t extended life of the ceramic heaters. The manufacturers recommendati on for nebulizer gas pressure for this flow rate was 35 arbitrary units. This corresponds to a value between 4 and 8 L/min, although it is
72 unclear of the exact value. No increase in sensitivity was afforded by varying these parameters at this flow rate (variation in gas parame ters evaluated by direct infusion of 1 g/mL at 10 L/min). The IonSpray voltage was optimized for peak area to 3.0 kV (for reference 3 to 6kV is common, see chapter 1 for more detailed explanati on). The curtain gas enters after the spray and the orifice plate to prevent contamination of th e optics (refer to Figure 2-7). This parameter was set to 10 arbitrary units. To en sure these settings were optimum, a separate experiment was performed varying the capillary in the x-posi tion from 2.5 to 3.5 mm a nd the y-position 0.5 to 2.5 mm. No advantage in sensitivity was observed. Based on the results of these experiment s; a set of optimum mass spectrometric parameters were selected. These are summarized in Table 3-3. These parameters were used for subsequent experiments desc ribed in chapters 3 and 4. Reversed Phase Liquid Chromatography Reversed phase chromatography (RPC)124 includes any chromatographic method that uses a non-polar stationary phase. In the 1970s most liquid chromatography was done on nonmodified silica or alumina sta tionary phases with a hydrophilic surface chemistry and a stronger affinity for polar compoundshence it was cons idered "normal". The introduction of alkyl chains covalently bonded to the support surf ace reversed the elut ion order. Now polar compounds are eluted first, while non-polar compounds are retainedhence "reversed phase". Today, reversed phase column chromatogr aphy is the most common form of liquid chromatography. Some examples of common functionalities that are marketed as stationary phases for reversed phase high performance liquid chromat ography (HPLC) are C8 and C18 alkyl chains that are bonded on silica. It was anticipated that SCNwould have limited or no retention on
73 33 544 5 5 2 2.5 3 3.5 4 4.5 5 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 210 220 230 240 250 Figure 3-2. Surface plot of the intensity of the detector response obtained at various positions of the TurboIonSpray probe with re spect to the orifice plate. Horizontal position (mm) I n t e n s i t yV e r t i c a l p o s i t i o n
74 Table 3-3. Summary of the MS /MS parameter optimization Parameter Parameter Optimized setting CUR Curtain gas 10 arbitrary units GS1 Ion source gas 1 40 arbitrary units GS2 Ion source gas 2 50 arbitrary units IS IonSpray voltage 3.0 kV IHE Interface heater On TEM Temperature (ceramic heaters) 550C CAD Chamber pressure (for CAD or CID) 5 x 10-5 torr DP Declustering potential 34 V EP Entrance potential 12 V CE Collision energy 45 eV CXP Cell exit potential 15 V Q1 Res Q1 Resolution High Q3 Res Q3 Resolution Unit these columns, and would elute in the void (dead) volume with other electrolytic impurities. Ion chromatography122 (IC) has been reported as a successf ul solution to these retention issues. Commercially available ion s uppressor cartridges are attached between the column and the ionization source to allow for ESI compatibility; however, IC instrumentation was not available for these studies. To address this problem, more selective reversed phase stationary phases were evaluated. The diphenyl ligands of the PursuitM S LC column promised enhanced selectivity based on differences in an analytes double bond and aromatic group st ructure, creating a mechanism of retention utilizing electron interactions. The diphenyl stationary phase bound
75 to the silica scaffold. The diphe nyl stationary phase is similar in functionality to biphenyl stationary phases however the distinction in name indicates differences in proprietary manufacturing processes. The API 4000 LC-ESI-MS/MS was operated with the parameters described in Table 3-3. The elution time for SCNsolution standard was ~1 minute corresponding to <100 theoretical plates. This was determined by the following equation: N = (5.55 T2) / (W2) where N = number of theoretical plates; T = elut ion time ~ 1 min; W = peak width at half the maximum peak height ~ 0.3 min (fig 2-14). Typical plate counts (N) for LC applications are between 1,000 < N < 5,000125. Was SCN retained? The comparison of this elution time to the theoretical dead time can help answer this question. Determination of Dead Time: Nontraditional Calculation The dead time (in min) for a 4.6 mm ID column has been described as126 T0 = 0.01 x (L/F) where L = column length; F = flow rate; 0.01 is a constant presumably with units to cancel all labels save for time. It is generally accepted that to perform a method transfer between columns of different dimensions, the followi ng relationship must be considered in order to keep the linear velocity of the mobile phase constant: (Radius new2) / (Radius old2) = (new flow rate / old flow rate) Therefore (1.05 mm2) / (2.3mm2) = (0.208 mLmin-1/1 mLmin-1) Since this is a new method with no old flow rate pa rameter, the old flow ra te was selected as 1
76 mL/minute since this is a typical flow rate fo r 4.6mm ID columns. Since the experiment was run at 0.2 mL minute, the linear ve locity is nearly the same. So, for a 4.6 mm ID column, T0 = 0.01 x (100/1) = 1 min The corresponding theoretical dead time for a 2.1 mm ID column would be the same since the linear velocity is nearly th e same. So, for the 2.1 mm ID column of the same length, the calculated value for the T0 = 1 min indicating that it is not re tained on this column. Thus, one can conclude that the SRM traces in Figure 2-14 demonstrated that SCNeluted at ~1 min and therefore, it was determined that the interactions afforded by the PursuitMS column lacked the selectivity to effectively retain free SCN-. Therefore, this column will not separate SCNfrom other species that are not retain ed under reversed-phase condi tions (e.g. other ions). The only utility of this column would be to separate SCNfrom non-ionic species, which would be retained since the beginning of the LC runs ar e most susceptible to matrix effects (e.g. the potential for more than one unretained species to coelute), a post-column infusion evaluation of the effect of matrix on the analyte signal was performed. Evaluation of Analyte Suppression Due to Matrix Effects by Post-Column Infusion Strategies to control or eliminate the effect of matrix on the ionizat ion of efficiency of selected analytes include 1) column selectiv ity, 2) online diversion, and 3) offline sample cleanup (such as solid-phase extr action or liquid-liquid extraction) A useful method to map the matrix effects during a chromatographic run is a technique known as post-column infusions. This technique has been reported and applied by a number of investigators. Briefly, the selected analyte (SCNin this case) is delivered by an infusion pump (post-column) through a high pressure T (see Figure 3-3). Selected biofluids can then be injected sequentially with the autosampler (20 L fixed loop injection) onto the LC colu mn for evaluation. For each matrix, the
77 presence of components of the matrix (includi ng neutral molecules or competing electrolytes) that will suppress the analyte signa l are detected by deviation in the constant signal. For this experiment, the MS/MS parameters were descri bed in Table 3-3. The combined mobile phase flow rate from the two LC pumps (with mixing chamber) was 200 L/min. The 32SCNstandard (1 g/mL) was directly infused at a flow rate of 10 L/min. Thus, the rate of SCNintroduction was 10 ng/min. The mobile phase was 50/50 AcCN/H2O. Using this design, different matrices were evaluated. Figure 3-3: Post-column in fusion experiment design. Blank mobile phase (50/50 AcCN/H2O) was injected and no increase or decrease in intensity was detected as predicted (Figure 3-4) An oral rinse sample collected from a human volunteer using a Salivette was al so evaluated. Salivettes (Sarstedt, Newton NC) were selected
78 for sample collection in this p ilot study due to the products dem onstrated ability to generate ESI-compatible samples. An alternative technique for the collection of oral fluids was reported by having volunteers lean forward in a chair a nd decant saliva from their lower lip in to a collection vessel143. The potential for food particles fouli ng the ionization sour ce disqualified this method from consideration. The procedure to ob tain oral rinses from human volunteers was based on the commercially available Salivette me thods by Sarstedt. Volunteers were given two pieces of standard sized dental gauze and asked to place one in each side of their mouth for a collection period of five minutes. The recovered gauze pi eces were placed in to the Salivette centrifuge tubes. The Salivettes were centr ifuged at 3,000 rpm for 10 minutes. The recovered samples were diluted and tested immediately. To further describe the Salivette collection process, a study was performed to determine th e amount of saliva collected by each piece of dental gauze A 5 minute collecti on from a single volunteer was evalua ted for this study. The dry gauzed was weighed on a tared Mettler-Toledo anal ytical balance. After collection the gauze was weighed. The amount of saliva coll ected was determined by difference. The results are presented in Table 3-4. This data suggest s that ~ 4g of saliva (~4 mL) wa s collected during the collection process for two Salivettes. A 20 L aliquot was diluted to 1 mL w ith mobile phase and evaluated by post-column infusion (Figure 3-5). The data obta ined was very similar to the data obtained from the blank injection presented in Figure 34. To exaggerate the e ffect of matrix, a 100 L aliquot of the oral rinse sample was diluted to 1 mL with mobile phase and evaluated under the same conditions (Figure 3-6). The decrease in in tensity corresponding to the dead volume of the column indicates that there are endogenous coelu ting ESI competitors (ions or molecules) that exist in the oral rinse matrix. However, it wa s demonstrated that these interferences can be controlled by dilution and therefore a dilute-and -shoot strategy is satisfactory. Based on these
79 experiments, it is recommended that isotopically labeled internal standards (i.e coeluting standards that will mimic ionization suppression without compounding the issue) are recommended for quantitative measurement. Figure 3-4: Post-column infusion chromatogram of mobile phase for the negative ion transition at m/z 5826 Table 3-4. Determination of the amount of saliva collected in a 5 minute period Identification Mass (grams) Gauze (dry) 0.7730 One piece of gauze + saliva (10 minutes) 2.6856 Saliva by difference 1.9126 I n t e n s i t y
80 Figure 3-5. Post-column infusion comparison of more dilute clinical specim en of the same oral rinse specimen from a human volunteer for the transition of m/z 5826. 980 L mobile phase + 20 L oral rinses. I n t e n s i t y
81 Figure 3-6. Post-column infusi on comparison of a clinical speci men of an oral rinse from a human volunteer for the transition of m/z 5826. 900 L mobile phase + 100 L oral rinses. I n t e n s i t y Time
82 The Selection of Mobile Phase Flow Rate The selection of mobile phase flow rate is pa rt of the art of LC de velopment, since there is a balance between controlling analysis time chromatographic peak shape and system backpressure. The mass spectrometry parameters that were summarized in Table 3-3 were used in this experiment. The data acquisition software wa s Analyst 3.0. The Varian PursuitMS (diphenyl stationary phase, 2.1 x 100) column was used in this experiment. Samples were injected using a Perkin Elmer series 200 autosampler (20 L injection loop). The mobile phase was 50 % acetonitrile / 50% water. To ensure that the flow rate parameter was truly varied, before each flow level was evaluated, the system software was completely shutdown and restarted bring the LC pumps to standby. A 50 ng/mL solution standard of KSCN(aq) was pr epared to a concentration of in mobile phase. Figure 3-7 is the SRM 58 26 transition for this standard at the flow rate of 200 L/min (Varians recommended flow rate for this column). Since a lower flow rate could possibly offer the advantage of improved separati on from interferences that could be retained on this column, a 100 L/min flow rate was also evaluated (Figure 3-8). The data obtained at 200 L/min produced a smaller peak at half the maximum peak height (W) and therefore would provide better chromatographic resolution (Rs) (different from mass spectral resolution discussed in chapter 2, Figure 2-10). Cons ider this equation: Rs = [2(TB TA)] / (WA + WB) where A and B represent two analytes injected onto this chromatographic column and T is the corresponding retention times for each. Let analyte A be SCNand let analyte B be some interfering ion or molecule that has a stronger affinity for th e diphenyl stationary phase. It follows that as WA decreases, Rs increases. In addition to the an ticipated resolution advantage,
83 Figure 3-7. LC-MS/MS transition m/z 5826 at 200 L/min flow rate. Figure 3-8. LC-MS/MS transition m/z 5826 at 100 L/min flow rate. Time I n t e n s i t y Time I n t e n s i t y
84 the higher flow rate 200 L/min provided a faster analysis time. A summary of the figures of merit for comparison is given in Table 3-5. Base d upon the shorter elution time, narrower peak width and increase peak area for SCN-, the higher flow rate was chos en for further experiments. Table 3-5. Summary of flow rate study Flow rate ( L/min) Elution time, T (min) Peak width, W (min) Peak Height (counts) 100 1.84 0.6 28 200 1.04 0.3 32 HPLC Mobile Phase and Modifiers The additio n of mobile phase modifiers is a common practice in the development of HPLC methods with UVvis detectors th at do not have issues with an alyte suppression. The purpose of this is to modify the selectivity of the method to afford better se paration. The selec tivity factor is defined by the equation125: = kb / ka = [(Trb T0) / (Tra T0)] where = selectivity factor, kb = distribution constant for the mo re strongly retained species ka = distribution constant for the mo re weakly retained species, T0 = dead time, Tra and Trb are the respective retention times of the two co mpounds arbitrarily labeled as a and b. In ESI method development, the purpose of the addition of modifiers is slightly different. Since the role of mobile phase in the ESI process is to assist in maintaining the stable spray current by displacing charge, the selection of mob ile phase modifiers are is far more limited. For example, phosphate buffers that are typically us ed to control pH are c onsidered incompatible with the ESI source. Consider the equations of ESI current described by Francisco De La Mora.: iES = f ( r)( f r)1/2 where iES = electrospray ionization current, r = diecletric constant, = surface tension,
85 = electric conductiv ity of the liquid, f = flow rate. This equation, a priori, helps one predict the anticipated effect of two negative ion mode modifiers on the ESI curren t (considered here as proportional to the intensity of the detector response). Ammonium salts, such as ammonium acetate, are typically used in negative ion m ode experiments since th ere is no counter ion produced (NH4 + NH3). Acetate was not used here since the corresponding m/z = 59 is too close in m/z to SCN( m/z = 58). Ammonium hydroxide and formaldehyde were chosen as modifiers in this experiment. Ammonium hydroxide was chosen as a modifier for its solutionphase basicity, since it is believed that solutionphase chemistry strongly re flects the behavior of specific components in the gas phase100. The wrong way round explanation94 was suggested to explain the occurrence of [M H]ions in basic solutions and [M+H]+ for acidic solutions. Hydroxide is expected to deprotonate most an alytes; however, the addition of hydroxide ions will make the mobile phase more viscous, complicating the desolvation mechanism in the source. It is anticipated that the sensitivity wi ll decrease with the addi tion of hydroxide (decrease from equation above); however, a survey of th e literature suggests that there may be some advantage for specific compounds81. Dalton et al. reported the ability of weak acids (such as formaldehyde) to improve the sensitivity of weakly basic compounds in negative ion ESI129. The reasoning behind this is based on the work of Van Berkel107,130 that considers the ESI s ource as a special kind of electrochemical cell (see chapter one for more detail). In negative ion mode, it is believed that a reduction of the analyte occurs. 2 Analyte 2 [Analyte H]+ H2 Since the reduction reactions form H2 gas, Dalton suggests weak acids that donate hydrogen in the presence of the electrochemical processes occurring at the ESI capillary would improve the
86 sensitivity for the analysis. Upon evaluation, this study suggested that we ak acids improve the sensitivity of more ionic com pounds versus nonpolar compounds. SCNseemed an appropriate candidate to evaluate th e formaldehyde additive. An MS/MS experiment was performed with three replicate measur ement of a 200 ng/mL aqueous SCNstandard with variation in mobile phase The mass spectrometric parameters were described in Table 3-3. Th e injection volume was 20 L. The mobile phase was 50/50 acetonitrile and water. The effect of the addition of possible mobile phase modifiers to evaluate is summarized in Table 3-6 and Figure 3-9. The error bars represent one standard deviation of the mean above and below the average peak area. The data suggest that the selected modifiers that were evaluated were not appropriate for SCNand no modifiers were used for the remained of this project. Table 3-6. Evaluation of the effect of mobile phase modifiers on the MS/MS peak areas for SCNby SRM m/z 5826. Mobile phase Inj 1 peak area Inj 2 peak area Inj 3 peak area Average Standard deviation of the mean 50 / 50 AcCN / H2O 544 622 673 613 38 with 0.05% NH4OH 461 509 447 472 19 with 0.05% CH2O 264 64 163 164 59 Evaluation of the Tailing Factor (T) System suitability calculations have been reported to determine the effect of peak asymmetry on the calculated integration of peak area131. The tailing factor to desc ribe chromatographic peak symmetry is defined by the following relationship: T = (A 5%h + B 5% h) / (2 x A 5 % h ) where T = tailing factor; A is the peak width from th e front of the peak to the center of the peak; B is the peak width from the center of the peak to the back of the peak; 5%h is the peak height at 5% of the maximum intensity. The variables for this equation are define d in Figure 3-10. Figure
87 3-11 demonstrates how a slight di stortion in peak shape can result in considerable error in peak area determinations. Figure 3-9. Evaluating the effect of mobile phase modifiers on the MS/MS peak area intensity for SRM m/z 5826. Figure 3-10. Defining the tailing factor (T) figur e of merit for system suitability calculations. Image available at www.forumsci.co.il/HPLC/SST_abic.pdf
88 Figure 3-11. Effect of peak symmetry on peak integration in terms of T. Image available at www.forumsci.co.il/H P LC/SST_abic.pdf In summary, these figures illustrate the qua ntitative relationship between the obtained chromatogram and the confidence th at one can assign to the peak area obtained by the automatic integration features of commercially available data acquisition software. In this study, two parameters were determined to have an effect on T. These two parameters were mobile phase composition and injection volume mo re specifically, variation in the amount of analyte loaded on column. Figure 3-12 and Table 3-7 illustrates th e effect of mobile pha se composition on T. The optimum mobile phase composition was 45% AcCN : 55% H2O yielding a tailing factor of 1.7. A second parameter that can be useful to optimize the symmetry of chromatographic peaks is the injection volume. A working parame ter for injection volume can be estimated using the 15% method reported by Dolan153 where 15-20% of the peak volume is considered a good starting point for injection vol ume. Peak volume is determined by the following equation: Flow rate ( L/min) x peak width at half the maximum peak height (min) = peak volume ( L) thus, SCNwhich has a peak width at half the ma ximum peak height of 0.4 minutes at 200 L/min has a peak volume = 80 L. The working range for the injection volume at 15-20% of peak volume is 12-16 L. A 100 L calibrated loop was present in the Perkin Elmer series 200
89 autosampler and therefore was used for this study to evaluate the effect of variation in injection volume on the symmetry of the chromatographic p eak with points above and below the estimated working range. The results of this study are presented in Table 3-8. Figure 3-12. Evaluation of the effect of va riations in mobile phase composition on peak symmetry, as measured by tailing factor T. The optimal point is identified on the graph with a box. The injection volume was 20 L. Table 3-7. Evaluation of the effect of variations in mobile phase composition on peak symmetry for SCN-. % AcCN T 40 2.3 42 2.5 43 2.1 45 1.7 47 2.0 50 2.4 60 3.8 Mass spectrometry parameters were summarized in Table 3-3. The mobile phase recipe was described earlier in this chapter (45% AcCN / 55% H2O). The SCNstandard was 90 g/mL (a low-to-mid-range concentration for SCNin saliva). The results showed that the optimal
90 working range to minimize tailing for injection volume was 4-17 L. Since the lowest calibrated loop available was 20 L, it was selected for the remainde r of this work. The loss of peak symmetry afforded by the 20 L injection is balanced by the antic ipated gain in precision of the fixed calibrated loop. The evaluation of method precision will be pres ented in the chapter 4. Table 3-8. Evaluation of the effect of injection on volume on peak symmetry. Injection volume (mL) Expressed as g on column Tailing Factor (T) 0.001 0.09 not detected 0.006 0.54 1.5 0.012 1.08 1.5 0.014 1.26 1.1 0.017 1.53 1.3 0.020 1.80 1.7 0.025 2.25 1.7 Evaluation of EBC Samples The novel L C-ESI-MS/MS method described in chapters 2 & 3 was applied to real clinical specimens of EBC. Two commercially available breath sampling methods were evaluated to determine if SCNcould be detected in EBC by LC-ESI-MS/MS. The first method uses the RTube157, shown in several photos in w ww.rtube.com. The components of a commercially available EBC samp ling apparatus from Respiratory Research (Charlottesville, VA) consists of a polypropylene collection tube that is shaped with a 90 bend to facilitate the separation of saliva and EBC (figure 3-13). There are one-way valves before and after the 90 bend to separate expiratory from inspiratory breath. The second one-way valve opens in response to a volunteers breath, allowing entry in to the collection cham ber (Figure 3-14). This process also allows for expansion (therefore cooling
91 Figure 3-13. The RTube design to separate EBC and saliva. Figure 3-14. Description of the pa th of expired breath af ter the saliva trap, at the entrance of the one-way valve. The design of the one-way valve in to the polypr opylene tube allows for expansion to facilitate condensation.
92 of the breath) which facilitates condensation. The aluminum cylinder is usually kept in a laboratory freezer (~0-5C) and is immediately removed prior to testing. A thermal sleeve is placed over the chilled aluminum cylinder for patient comfort. The collection of EBC samples from healt hy (no chronic airway condition, nonsmokers) was conducted at the University of Miami Hu man Subject Laboratory in accordance with a validated IRB protocol (more on IRB regulations and ethical considerations for the study of human subjects in chapter 4). EBC samples from healthy volunteers were frozen and shipped on dry ice to the University of Florida and stored in the laboratory freezer. Volunteers were seated in a comfortable chair and given plastic nose clip s to wear (not pictured). The purpose of the nose clips was to standardize breathing patterns from person-to-person. The clips prevent the volunteer from nasal breathing. The application of nose clips in EBC sampling has been reported previously 158. The RTube is given to the volunteer and ex piratory breath is collected for a period of ten minutes. This is in accordance with curr ent recommendations for the standardization of EBC methods 159. Sample collection can be performed for a specific time or for constant volume sampling. Since controlled time sampling dominates the literature, EBC samples were collected in this manner for all experiments reported in this dissertation. Volunteers we re instructed to take breaks at any time during the testing procedure if there was discomfort such as headaches or dizziness. Upon completion of the sampling period, the RTube is disassembled and the pooled EBC sample is placed in labor atory freezer. Since enzyme concentrations in EBC are very low (if there at all), the temperatur e of the freezer is not considered a critical parameter to control 159. The second method of collection was th e Peltier-cooled Jaeger EcoScreen160. This offered several advantages. Breath collection performed in -house at UF eliminated the quality control concerns associated with the shipment of biologi cal specimens. The device also cools samples to
93 10C versus 0C with the RTubeThe advantage of the lower cooling temperature allows for a more efficient shorter collection time (sometimes just a few minutes was needed). Disadvantages versus of the EcoScreen versus the RTube include sacrificing the domiciliary aspect of sampling. Domiciliary sampling is a potential advantages for volunteers who are ill and cannot leave there home. Since the Peltier cooli ng unit weighs about 80 pounds, E BC sampling with this method could only be done at the Department of Anesthesiology at the Univer sity of Florida. Soyer et al. reported an analytical comparison of E BC collection by the RTube and EcoScreen161. It was previously determined that EBC collected with the EcoScreen allowed for greater sensitivity in the detection of lipids and prot eins; however, little was known a bout the collection efficiency. Over the period of my employme nt here at UF, volunteers who donated real EBC specimens did not wear nose clips. The sample was collected from a single volun teer, 10 min collection period, without nose clips (unavailable at time of sampling). Mass spectrometric parameters were described previously in Table 3-3. The injection volume was 20 L. The mobile phase was 45/55 AcCN/H2O. The flow rate was 200 L/min. Samples were run neat, in accordance with current practice at the Department of Anesthesiology, Sh ands Medical School, University of Florida. Data for EBC collected by the EcoS creen are shown in Figure 3-15. SCNwas not detected in this sample. It was noted that the background for m/z 58 26 in the EBC was approximately 20 fold higher versus the water sample. Since there was no peak at the elution time corresponding to SCNat approximately 1 min, it appears that sensitivity may still not satisfactory. In the next experiment, a method that details the application of a nitrogen evaporator to address this issue will be presented.
94 Figure 3-15. LC-MS/MS chromatograms for an EBC samples collected from the Jaeger EcoScreen: a) water b) EBC sample. SCNcould not be detected using the EcoScreen collection device. Two transitions at m/z 5826 (blue) and m/z 6026 (red) were monitored.
95 Offline Sample Enrichment Usin g a Nitrogen Evaporator The utility of a nitrogen evaporator for the preconcentration of EBC samples were evaluated to maximize the sensitiv ity of this analytical method. Th e nitrogen evaporator that was used is shown in Figure 3-16. Sa mple vials, each containing 300 L of an EBC specimen, are placed uncapped in to the HPLC vial carousel. Th e carousel is then lowered in to the heated water bath and nitrogen is applied to the samp le surface. The nitrogen delivery pressure is regulated by a knob under the flow meter. In general, the sample was blown down to dryness and reconstituted to a smaller volume to determ ine if the intensity of response (peak height) would vary linearly with the degree of preconcentration. During this process, volatile organic compounds (or water in the case of EBC) are se parated from the mixture and the analyte is reconstituted with mobile phase in an attemp t to provide a matrix-matched standard. The Organomation Associates Inc. (Berlin, Massachusetts) N-EVAP 112 nitrogen evaporator with an OAS heating system was used. The nitrogen deliv ery pressure was regulated to 10 psi. The heated water bath temperature was maintained between 50 and 55 C. Temperature and delivery pressure parameters were optimized to minimi ze the possibility of sample loss due to vigorous bubbling (bumping). In order to protect against random bumping events that cause sample to transfer from one vial to another, vials were arranged to leave a space on the carousel between each vial. Unfortunately, reconstitution with volumes less than 300 L (that is, preconcentration by a factor greater than 2) yielded less than expected response. This may reflect matrix effects and ion suppression (as noted for oral rinse sample s in Figure 3-5 to 3-7) or poor recovery of analyte from the vials at rec onstituted volumes less than 10 L. In any case, future samples were only concentrated by a factor of 2. Two EBC specimens obtained from the University of Miami were analyzed for a preliminary investigation. Samples were collect ed using an RTube for a 10 min collection
96 Figure 3-16. LC-MS/MS chromatogram for tw o EBC specimens collected using the RTube collection device and preconcentrated by a f actor of 2 using a nitrogen evaporator. S/N was calculated using Analyst soft ware: a) S/N ~ 2 b) S/N ~ 13. 0.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.126.96.36.199.04.24.44.64.8 Time, min 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 5.8 2.89 0.92 1.14 0.53 4.63 3.49 4.28 1.88 2.45 2.11 I n t e n s i t y 0.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.188.8.131.52.04.24.44.64.8 Time, min 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 1.06 1.60 2.06 2.42 3.86 2.63 4.91 0.34 4.53 4.37 3.53I n t e n s i t y b) a )
97 period. Volunteers wore noseclips. Samples were pr econcentrated by a factor of 2.Data for this experiment are shown in Figure 3-17. Mass spectr ometric parameters were summarized in Table 3-3. The mobile phase was 45/55 AcCN/H2O. The loop injection volume was 20 L. One sample produced a peak below S/N = 3 i.e. below detect able limits. The other sample produced a peak with a S/N = 13, a possible positive detection. It is generally accepted that an analyte at concentrations near the detection limit of analytical method will be detected ~30% of the time. Since these two samples were only analyzed by a single injection each, littl e can be interpreted from this data. For this reason, in the evaluation of the next shipment of EBC specimens under the same conditions, triplicate injections were performed. Figure 3-18 shows triplicate analysis of one EBC sample. A second sample was analyz ed in triplicate under the same conditions Figure 3-19 SCNcould not be identified in samples collected under similar conditions with confidence (i.e there was no peak co rresponding to the elution time for SCN-). Further evaluation was needed. A clinical sample collected with an RTube was spiked with a solution standard. Figure 3-19 shows a comparison of EBC and spiked EBC. This would help to compensate for the effect of matrix on analyte response. It was demo nstrated that a low level concentration of SCNis detectable in the EBC milieu. The response for this clinical specimen was very similar to the intensity response of the blank. Limited conclusions can be drawn from the characterization of one sample, but it seems that it is possible to detect SCNin EBC if it were present. If it were possible, an evaluation of a larger set of samples would have been useful to confirm this claim. Summary and Conclusions In chapter 2, the fundamentals of the sele cted mass analyzer, the TQMS, was detailed prior to evaluation and optimization. In this chapter, the selection of the ESI source was presented in similar fashion. Dr oplet dynamics and chemistry were first detailed through a
98 Figure 3-17. Triplicate injections of an EBC sample monitored by LC-MS/MS. Two transitions were monitored: m/z 58 26 (blue) and m/z 6026 (red). I n t e n s i t y Time I n t e n s i t y Time I n t e n s i t y Time
99 Figure 3-18. Triplicate injections of a s econd EBC sample monitored by LC-MS/MS. Two transitions were monitored: m/z 58 26 (blue) and m/z 6026 (red). I n t e n s i t y Time
100 literature review and selected topics were evaluated by experi ment.An overview was presented of the relevant commercial instrumentation was used in this experiment. During the development process, it was determined that mobile phase fl ow rate has a significant effect on the optimal parameter settings for the ionization source, in cluding gas flow, capill ary voltage and source temperature. It was interesting to learn that as flow rate increases, these parameters become less compound selective and more relevant to desolv ation of mobile phase. Additional experiments were performed to determine the optimal position of the ESI probe relative to the sampling orifice of the mass spectrometer. The optimization of LC parameters was also presented in this chapter. The internal diameter of the LC column has an effect on the se lected mobile phase flow rate (which in turn has an effect on the ESI source parameters). A discussion of peak symmetry was presented to illustrate the effect of peak distortion on the confidence of the integrated peak area. Variation of two key parameters was considered for the eval uation of chromatographic peak shape. These parameters, mobile phase composition and injection volume, were studied. When the method parameters were considered optimized, the evaluation of the method for analysis of EBC specimens was performed. In this study, it could not be determined if SCNis detectable in EBC. This is due to the limite d sensitivity of the method. It is also acknowledged that this pilot study did not include a large enough sample population where variation in individual physiology and diet may affect the obtained results. The feasibility of applying this method to study human oral rinses, when SCNis expected to be at much higher concentrations, is presented in the next chapter.
101 Figure 3-19. Clinical sample monitored by SRM 58 26 A) water B) standard C) spiked EBC D) EBC I n t e n s i t y Time I n t e n s i t y I n t e n s i t y I n t e n s i t y C D B A
102 CHAPTER 4 QUANTITATIVE ANALYSIS: MEASUREMENT AND VALIDATION OF FREE THIOCYANATE CONCENT RATIONS IN HUM AN ORAL RINSES (PROCESSED SALIVA) Overview This novel reversed phase LC-ESI-MS/MS m e thod has been applied to the selective detection and quantification of free SCNconcentrations in human oral rinses. This dissertation considers the operational definition fo r oral rinses as saliva samples that have been processed and filtered using the Salivette commercially availa ble sampling method. Single operator figures of merit that were considered for validation include d stability, analyte recove ry, and limitations of linearity. Studies to determin e the limitations of linearity for this methods include the determination of the upper limit of quantitation (ULOQ), lower limit of quantitation (LLOQ), method detection limit (MDL), precision and accur acy. The advantages of applying this direct measurement method versus previously reported methods include the elimination of laborious, complex sample cleanup and derivatization chemis try without a loss accuracy. A protocol for the study of human subjects was developed and a subsequent study was performed. The collection of oral rinse specimens was conducted under the guid elines for the protection of human subjects, including informed consent. Dietary cons umption (including smoking habits) has been previously reported as a factor in measured levels of SCN-. A comparison of the results of this study to similar work that has been previously re ported in literature is included for discussion. Suggestions for future work are also presented. To my knowledge, this is the first report of the measurement of [SCN-] by a reversed phase LC-MS/MS method. This approach has demonstrated improved detection limits (100x lo wer than Lundquists method), reduced sample preparation, requires less sample volume (10 L) and provides a plat form for high-throughput analysis.
103 The Institutional Review Board and the Pr otection of Human Subjects: IRB History Since 1945, various codes for the proper and responsible conduct of hum an experimentation in medical research have b een adopted by different organizations. These regulatory documents include the Nuremberg Co de of 1947, the Helsinki Declaration of 1964 (revised in 1975), the U.S. Department of H ealth, Education, and Welfare Guidelines of 1971 (codified into Federal Regulations in 1974), and the American Ps ychological Association Code of 1973132,133. Prior to the execution of experiments invol ving human subjects at the University of Florida, researchers are trai ned in a number of modules such as medical ethics, respect for persons, beneficence (decisions that are made in the best interest of the patient), and justice. Training is further focused on the practice of info rmed consent, assessment, and risk benefits. Ethical ways to recruit subjects are also consider ed as well as operational definitions of protected class populations. After completion of the appropriate training modul es, an IRB protocol for the collection of saliva samples from human vol unteers was developed and approved. Volunteers were recruited using flyers and advertisem ents in three buildings on the UF campus. Protocol Summary The approved protocol and inform ed cons ent are included in Appendices A & B for reference. In summary, volunteers were recruited by local advert isements and were invited to visit the Department of Anesthesiology at Sh ands Hospital (University of Florida) for the collection of oral rinse samples that would im mediately be tested after collection without freezing (see validation of freezer stability in this chapter). No restrictions on diet or activity were approved for this study. Since diet has a strong influence on the determination of [SCN-], volunteers were requested to arri ve at 9:00 a.m. Controlling th e sample collection time offered this investigation two advantages: 1) circ adian rhythms have an effect on measured concentrations of [SCN-]134; and it was anticipated that foods with high concen trations of SCN-
104 such as beer, mustards and green leafy vegetabl es would not be consumed during early morning hours. Variables like cigarette smoking were not controlled and we re anticipated to account for variability in the studied population44-47. Recall that the goal of this study was not to demonstrate the ability of [SCN-] to distinguish smokers from non-smok ers, since this has been reported by several groups using spectrophotometric methods44-47, but to demonstrate the feasibility of a reversed-phase LC-ESI-MS/MS approach for quanti tative measurement. Oral rinse samples were diluted 50x with mobile phase prior to analysis based on the results of the accuracy evaluation in the next section (dilution results in measurements in the middle of the calibration curve to ensure the most accurate measurement possible). Sample vi als were handled in such a way that provided no identification or personal link to any of the volunteers. Validation Summary Validation criteria for an alyte measurements were performed in accordance with recommendations described in both the Journal of Chromatography B135 and by the Environmental Protection Agency (EPA)136. Variables that were considered include limitations of linearity, method detection limit (MDL)137-142, freezer stability and analyte recovery. System suitability calculations describe d by the Food and Drug Administration131 (FDA) to quantify the effect of peak asymmetry on peak integration was demonstr ated in chapter 393,94. The Qualitative Identification of SCNin Oral Fluid The natural isotopic138 abundance of sulfur is 32S 96% and 34S 4%. This ratio was reflected in the isotopic distribution of 32SCNand 34SCNin standards and in clinical samples. A typical MS/MS strategy is to monitor the transi tion of the precursor i on to its most intense product ion (fragment). This is known as the q uant ion transition and is typically used for quantitation, since this will be the most sens itive transition. Sometimes in complex biological
105 mixtures, there is an issue with non-selective transitions such that other species may give similar MS/MS fragmentation patterns with the same precursor and product ion massesespecially if the LC retention time of the ion is relatively s hort (translating to a low number of theoretical plates). Therefore, it is common practice to monitor multiple transitions to generate confidence in the qualitative identification of an ion. The second precursor product i on transition monitored for a given ion is known as the qualifier. Quant and qualifier transitions will be revisited later in this chapter. For low-mass ions with little option for sele cting transitions, monitoring the two isotopic transitions for qualitative purposes is an importa nt analytical tool and can be considered analogous to the strate gy for larger ions145. The monitoring of the two isotopic transitions for qualitative purposes is presented in Figure 4-1. Transition 1 represents th e fragmentation of the precursor ion to the most selective product ion (32SCNCN-). This is the most abundant isotope. Transition 2 represents the heavy isot ope of sulfur for an analogous transition (34SCNCN-). Peak area for 32SCNCNm/z 58 26 = 12,483; Peak area for 34SCNCNm/z 60 26 = 572. This translates to an isotopic contribution of 32SCN= 95.5 and 34SCN= 4.5%, well within the experimental error of the experiment. Figure 4-2 shows LC-ESI-MS/MS transitions fo r oral rinses compared to water blanks and standards. These data demonstrated an increa se in the detector res ponse of the transition corresponding to SCNwhen the sample was spiked with SCN(aq) standard. The water blank also shows no background contribution from the aqueous standard. Based on the similar intensity of detector response, it can be es timated that the [SCN-] ~200 g/mL (2 g/mL x 100 dilution factor).
106 Figure 4-1: Identification of SCNin human oral rinses by captu re of the isotopic envelope. 0.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.184.108.40.206.04.24.44.64.85.0 Time, min 0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 0.97 Transition 1 32SCNCNTransition 2 34SCNCNI n t e n s i t y
107 Figure 4-2. Proof of concept: Valid ation for the identification of 32SCNin human oral fluid. A) mobile phase spiked with deionized water (negative control-blank) B) mobile phase spiked with 2 g/mL 32SCN(positive control) C) mobile phase spiked with oral fluid (clinical sample)
108 Evaluation of Analyte Recovery fr om the Dental Gauz e and Salivette Validation of the analyte recovery from Salivette was treated like a solid-phase extraction-recovery absolute recovery experiment135. Three concentrations of solution standards (KSCN dissolved in water) were deposited onto three respective pieces of gauze to represent low, medium and high concentrations found in clinical samples of hum an oral rinses. The Salivette would have been disqualified if re sults were < 50% recovery as recommended in the Journal of Chromatography B135; however, this was not the cas e. In fact, Salivettes did not contribute to any significa nt loss of signal for [SCN-] and were validated as an appropriate means of sample collection for this i nvestigation. Data from this validation experiment are given in Table 4-1. The raw data for evaluation is incl uded in Table 4-2. The LC-ESI-MS/MS parameters used for this experiment were previously de scribed in Table 3-3. Th e %absolute recovery reported reflects the average of three 20 L loop injections. The absolute recovery was determined for three concentrations using the following equation135: % Absolute recovery = [Response of anal yte spiked in to matrix (processed)] x 100% [R esponse of analyte of pure standard (unprocessed)] Table 4-1. The validation of the Salivette saliva sampling method by the determination of absolute recovery. Standards [SCN-] g/mL % Absolute recovery Blank 0 N/A Low 50 108 Medium 250 83 High 500 104 Stability A freezer stability study was designed and perform ed in accordance with current, standard recommendations135 The effect of three freeze-thaw cycles (12 hours each cycle) at three concentration levels reflecting low (50 g/mL), medium (250 g/mL) and high (500
109 g/mL) analyte concentrations in clinical speci mens was evaluated. Solu tions standards were stored in Salivette centrifuge tubes. Samples were stored at 80 C, thawed to ambient temperature and refrozen for three cycles. The internal standard was spiked in to an aliquot from the thawed standards after each of the freeze-thaw cycles and analyzed using the LC-ESIMS/MS parameters described in Tabl e 3-2. The peak area ratios for 32SCNversus the heavy isotope of 34SCNare reported in Table 4-3. Samples were considered stable if the measured 32SCN/ 34SCNratio was within a CV of 15% for 3 cycles. Freezer stability could not be verified at three c oncentration levels. Table 4-2. Peak areas for m/z 58 26 used in the determination of absolute recovery. Standard Concentration ( g/mL) Inj 1 peak area Inj 2 peak area Inj 3 peak area Average peak area Standard Deviation %RSD unprocessed 0 1 2 3 2 1 68 processed 0 7 0 0 3 4 173 unprocessed 50 424 400 422 415 13 3 processed 50 403 377 378 386 15 4 unprocessed 250 894 778 840 837 58 7 processed 250 978 963 1,001981 19 2 unprocessed 500 1,728 1,575 1,5341,612 102 6 processed 500 1,550 1,601 1,5091,553 46 3 The trend in decreasing concentr ations was not expected since typical reasons for analyte degradation, such as volatilization or photodegradation146, are not relevant to SCN-. It was also assumed that the low temperat ure -80C freezer sufficiently slowed uncontrolled oxidation reactions (SCNOSCN-). It was also assumed that the pl astic containers would not provide a favorable surface for analyte loss due to adsorption. Perhaps the most reasonable explanation is an uncontrolled precipitation reaction, however furt her evaluation is needed to determine the root cause of this freezer stability issue. A suggested experiment would be to measure SCN-, CNand OSCNduring the freeze-thaw cycles to determine a reasonable explanation for loss of analyte response147. Stability modifiers such as weak acids may also be considered; however, the
110 subsequent effect on the ioniza tion process should also be ev aluated. Since an effect was observed, the sample collection protocol spec ified immediate testing (without freezing) upon collection. Table 4-3. Stability study for three freeze-thaw cycles of SCNstandards at three concentrations to represent low, medium and high concentrat ions in clinical samples. The average of three replicate injections (n = 3) is reported [32SCN-] g/mL Start Avg Peak area ratio (n=3) Cycle 1 Avg Peak area ratio (n=3) Cycle 2 Avg Peak area ratio (n=3) Cycle 3 Average Stdev %CV 0 0.04 0.04 0.02 0.04 0.0 0.01 29 50 0.34 0.22 0.17 0.16 0.2 0.08 37 250 0.99 0.56 0.71 0.91 0.8 0.19 25 500 1.17 0.88 1.32 1.12 1.1 0.18 16 Sensitivity & Limitations of Linearity The m ethod detection limit (MDL) was determined with current EPA recommendations136. Seven replicate measurements of a 40 ng/mL solution standard of 32SCNwere used to determine the MDL for this method. The corresponding T-value was 3.707. The standard deviation obtained for the peak area was 0.054. It was determined that the MDL ~ 10 ng/mL. The method detection limit was verified by experiment and reported in Table 4-4 and summarized in Figure 4-3. The equation MDL = T x stdev was modified to include the 50 fold dilution factor for clinical specimens. The determined MDL = 10 ng/mL, a more conservative estimate than 200 pg/mL, the calculated value obtained without correct ion for dilution. The anticipated concentration of SCNin EBC specimens is anticipated to be lower than this MDL. It is noted that this method has adequa te sensitivity of for quantifying SCNin oral rinses. Precision Precision was considered in a few different ways during the course of developm ent and measurement. The within run precision of elu tion time by LC-ESI-MS/MS with the PursuitMS
111 Table 4-4. Summary of the determination of the method detection limit (MDL) for SCN-. MDL = stdev x t; stdev was determined by 7 injections; corresponding T value (7-1) injections = 3.707; MDL = 10 ng/mL (MDL = T x stdev x dilution factor; 3.707x 0.054 x 50 = 10). # replicate injections (n) 40 ng/mL Degrees of freedom (n-1) Corresponding T value Standard deviation ( ) MDL calculation 7 6 3.707 0.054 ~10 ng/mL Figure 4-3. Replicate injections of a 10 ng/mL SCNstandard LC-ESI-MS/MS transitions for m/z 58 26 for the verification of the met hod detection limit calculated for SCNby the EPA method. 0.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.220.127.116.11.04.24. 4 Time, min 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 4.9 0.96 4.09 0.20.40.60.81.01.21.41.61.82.02.22.42.62.83.03.18.104.22.168.04.24.44.64.85.0 Time, min 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 1.04 2.26 4.48 0.38I n t e n s i t y I n t e n s i t y
112 column was determined by performing five replicate injections of a solution standard spiked with an oral rinse specimen. Fi ve injections of 100x diluted saliva spiked with 2 g/mL 32SCN-and recording the time the eluted peak, what was (pre viously determined in this chapter to be the dead time). The fixed loop injection volume was 20 L. Data from this experiment are given in Table 4-5. For daily runs, a similar procedure is suggested with a solu tion standard since it typically takes 20 min at the beginning of each da y for the source temperature to stabilize. Since each injection is 5 min, this procedure would allow for the proper equilibration prior to experiment. Note that the body of work in this dissertation wa s not performed on one column. Also, consumables of the LC systems required service so deviation in elution time between experiments is easily explained. Consider this experiment design to monitor within run performance. Table 4-5. Evaluation of the within run drift in elution time for SCNin five replicates of the same clinical specimen Injection Time (minutes) 1 0.92 2 0.91 3 0.92 4 0.91 5 0.91 Average 0.9 % CV ~ 1 % In addition to the description of precision by replicate measurements at each respective concentration of the calibration curve, a second evaluation was performed using single-variable (or one-way) ANOVA148. A similar method has been described in the Journal of Chromatography B135. A 4 x 3 x 4 matrix was constructed using 4 independent concentrations, 3 replicate injections at each concentration le vel over 4 batches. The results of the ANOVA analysis are summarized in Table 4-6. When reviewing the ANOVA data, recall that the null hypothesis considers all population mean s as equal. The alternative hypo thesis is that at least one
113 mean is significantly different. The data in Tabl e 4-6 illustrates that p>0.05 at a variety of concentration levels. Table 4-6. Evaluation of the precision of th e autosampler using single variable ANOVA [SCN-] g/mL F Fcrit p 1.00 2.69 4.27 0.12 2.00 2.62 4.26 0.13 7.00 2.66 4.26 0.12 11.40 2.69 4.26 0.12 When the value for F > Fcrit, the null hypothesis is rejected indicating poor precision149. When p < 0.05, the null hypothesis is rejected and there is a significant difference between the groups. In the context of an autosampler, the precision of the method is poor if the means of replicate injections from the same vial are different150. This experiment was performed at di fferent concentration levels, a recurring theme that pervades the current literature on valida tion. The complete ANOVA data analysis obtained using Microsoft Excel is included in tables 4-7 to 410. This data supports the claim that measurements over the clinical range of interest are precise. Table 4-7. ANOVA data for a 1 ppm SCNsolution standard Standard 1.20 g/mL Batch 1 Batch 2 Batch 3 injection 1 0.47 0.44 0.45 injection 2 0.47 0.47 0.43 injection 3 0.45 0.47 0.45 injection 4 0.44 0.47 0.43 Anova: Single Factor SUMMARY Groups Count Sum Average Variance Column 1 4 1.8305620.4576410.00029 Column 2 4 1.8567450.4641860.000225 Column 3 4 1.7622960.4405740.000148 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 0.00118888 2 0.0005942.6905590.121346 4.256495 Within Groups 0.00198842 9 0.000221 Total 0.00317731 11
114 Table 4-8. ANOVA data for a 2 ppm SCNsolution standard Standard 2.30 g/mL Batch 1 Batch 2 Batch 3 injection 1 0.53 0.57 0.53 injection 2 0.54 0.56 0.52 injection 3 0.54 0.53 0.54 injection 4 0.58 0.57 0.54 Anova: Single Factor SUMMARY Groups Count Sum Average Variance Column 1 4 2.1950430.5487610.000361 Column 2 4 2.2301080.5575270.000334 Column 3 4 2.12747 0.5318688.27E-05 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 0.00136085 2 0.00068 2.62464 0.12648 4.256495 Within Groups 0.002333206 9 0.000259 Total 0.003694056 11 Table 4-9. ANOVA data for a 7 ppm SCNsolution standard Standard 7.00 g/mL Batch 1 Batch 2 Batch 3 Injection 1 0.94 0.93 0.93 Injection 2 0.90 0.89 0.92 Injection 3 0.93 0.91 0.98 Injection 4 0.92 0.92 0.96 Groups Count Sum Average Variance Column 1 4 3.6929758690.9232440.000297 Column 2 4 3.65102085 0.9127550.000275 Column 3 4 3.7862230220.9465560.000779 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 0.002395 2 0.0011972.6578020.123865 4.256495 Within Groups 0.004054 9 0.00045 Total 0.006449 11
115 Table 4-10. ANOVA data for an 11.4 ppm SCNsolution standard Standard 11.40 g/mL Batch 1 Batch 2 Batch 3 Injection 1 1.73 1.71 1.66 Injection 2 1.65 1.67 1.55 Injection 3 1.71 1.55 1.60 Injection 4 1.71 1.57 1.64 Anova: Single Factor SUMMARY Groups Count Sum Average Variance Column 1 4 6.8000211.7000050.001078 Column 2 4 6.4992921.6248230.0062 Column 3 4 6.4587991.6147 0.0024 ANOVA Source of Variation SS df MS F P-value F crit Between Groups 0.0173759 2 0.008688 2.69308 0.121154 4.256495 Within Groups 0.0290342 9 0.003226 Total 0.0464101 11 Human Subjects Data The determ ination of free [SCN-] in a selected population of 18 human volunteers is reported in Table 4-11. The mass spectrometric parameters were reported in Table 3-3. The calibration model for this measurement was (area 32SCN/area 34SCN-)= 0.7304([32SCN-]/[34SCN-]) + 0.3705. The isotope dilution calibration curve is detailed in Figure 4-4 and Table 4-12. Triplicate 20 L loop injections were preformed for each vial. Erro r bars reflect one standard deviation of the mean (n = 3) above and below the average pe ak area for 3 injections of each standard. A minimum of six concentration levels ove r one concentration decade is recommend135, so seven levels were prepared and teste d. The lowest concentration was 50 g/mL. Calibration data is presented in Figure 4-4 and Table 4-12. A control chart is included in Figure 4-5 to monitor the pe rformance of the instrument during the course of the run by char ting the intensity of the internal standard. If system events
116 Table 4-11. The quantification of [SCN-] in human oral rinses in 18 human subjects by reversed phase LC-ESI-MS/MS. The lowest standard concentration was 50 g/mL. Sample # Average g/mL (n=3) Within run %CV 1 361 4 2 255 8 3 479 4 4 Above highest standard 5 5 72 4 6 217 2 7 52 6 8 208 4 9 171 < 1 10 94 3 11 59 9 12 194 1 13 197 6 14 113 4 15 Below lowest standard 40 16 Below lowest standard 8 17 90 6 18 65 7 Figure 4-4. Isotope dilution ESI-MS/MS calibration of SCNsolution standards P e a k a r e a r a t i o Concentration ratio
117 Table 4-12. Isotope dilution ESI-MS/MS calibration of SCNsolution standards Vial [32SCN-] Concentration ratio Peak area ratio Stdev Stdev of mean (n = 3) %RSD 1 0.5 0.23 0.51 0.01 0.00 1.4 2 1.0 0.47 0.71 0.08 0.05 11.2 3 1.5 0.70 0.88 0.01 0.01 1.3 4 2.0 0.93 1.10 0.14 0.08 12.5 5 3.0 1.40 1.36 0.09 0.05 6.6 6 4.0 1.86 1.78 0.10 0.06 5.4 7 4.9 2.28 1.99 0.11 0.06 5.5 Figure 4-5. Control chart for the determination of drift in measurement of clinical oral rinse specimens. No significant drift (+/3 standard de viations) was detected.
118 occur that dramatically affect the sensitivity of the method, the shape of the data on the control chart may help when trying to determine the appropriate interpretation of data. For this measurement, the system performed as expect ed (i.e. random distribution of error in the measurement of the internal standard) and no loss of sensitivity was observed. Replicate Measurements of a Calibration Control Standard A calibration control standard wa s also prepared and evaluated. In an analytical run with replicate injections of standa rds and large num bers of sample s, the prepared biological specimens may sometimes be stored at room temp erature in the autosampler of the LC system for several hours. Therefore, it is important to monitor the perfo rmance of the analysis over time by replicate injections of the same standard vial at the beginning, middle, and end of the experiment. For larger population studies, further consideration of additional time points is suggested. There were 18 specimens of oral rins es tested in this study. The % CV of replicate injections of the same standard vial were pe rformed at the beginning (before sample 1), middle (before sample 10) and end (after sample 18) of the experiment was 2%. Therefore, no significant change in response due to sample st orage at room temperature was observed. Similar replicate injections of a biol ogical specimen, more specifically, a NIST certified reference material would indicate if storage of such samples at room temperature during extended analytical runs would lead to sample stability problems. Results are presented in Table 4-13. Evaluation of Carryover Carryover151 of analyte from injection-to-injection has been reported as a potential pitfall in the high-throughput screening of quantitative screening of biol ogical specimens with minimal sample preparation. Two types of carryover ha ve been identified in LC-MS analysis: 1) carryover from the autosampler needle 2) carryover from the column. To minimize the effects of carryover from the autosampler syringe, thre e syringe washes with Millipore water was
119 performed between each injection. In addition, a blank injection of Millipore water were also performed between each sample to minimi ze the effects of column carryover. Table 4-13. The evaluation of a calibration control standard to monitor instrument performance during the course of the experiment. Bracket standards [SCN-] g/mL Before sample 1 1.66 Before sample 10 1.69 After sample 18 1.64 std dev 3 average 1.7 %CV 2 Figure 4-6 is included to evaluate this cleaning protocol within run of clinical specimens. Figure 4-6-a shows the first injection of water. Figure 4-6-b shows an in jection of water after clinical specimen # 1. These sp ectra are representative of the entire run, so for brevity in presentation, only two spectra were included in this report Carryover would have been considered significant if the peak height of the water blank injected after the clinical specimen was greater than 20% of peak height for the lowest standard which was not the case. No carryover or memory effects (i.e artifacts of high concentration analytes bleeding signal from injection-to-injection) was observed in th e water spectra of this experiment. It is acknowledged that carryove r is a known cause of measurement bias and may be the reason for the atypically high sa mples present in the sample population studied. Since SCNis in the dead volume, it is anticipat ed that this is the case. Table 4-14 provides a comparison of the measured values for SCNin this study versus data compiled from similar salivary studies. The results obtained in this work are in good agreemen t with literature values. The three samples that were atypically high could be easily explained by the restrictio ns of our IRB protocol, not by carryover. Since we could not control diet, so me volunteers walked in consuming food and coffee.
120 Figure 4-6. Evaluation of carryover from clini cal specimens of human oral rinses during chromatographic analysis. A) An injection of water prior to the injection of a clinical specimen (peak height 17 at 1 min). B) An injection of water after the clinical specimen (peak height 13 at 1 min). a) b) I n t e n s i t y Time 5 0 Blue trace 32SCNCNRed trace 34SCNCN-
121 Table 4-14. Comparison of the measured values for salivary [SCN-] with similar studies in literature. Reference Sampling methods Measurement methods Salivary [SCN-] Range ( g/mL) Analytical Biochemistry,1996 ,240,7-12 Saliva decanted off bottom lip FT-IR at 2058 cm-1N = 25 50-200 Journal of health Science, 2000, 46(5), 343-350 Plastic tubes (no method details) Konig Reaction, Spectrophotometric measurement at 607 nm N = 40 0.8-148 Journal of Analytical Toxicology, 2006, 30, 511-515 Plastic tubes (no method details) Derivatization, GC-MS N = 10 17-60 Analytical Sciences, 2002, 18, 887-892 No description Ion selective electrodes N = 2 35 & 105 Unpublished UF study, Rocky Point Labs Salivettes Konig Reaction, Spectrophotometric measurement at 607 nm N = 20 56-235 This study Salivettes LC-ESI-MS/MS N =13 50-300 N = 3 >300 N = 2 < 50 Conclusions The novel L C-ESI-MS/MS method described in chapters two and three proved useful in measuring low levels of SCNin saliva (oral rinses). This claim is guarded by successful compliance with validation criteria recommended by the Journal of Chromatography B including recovery, post-preparat ive stability, linearity, precision and accuracy. Although freezer stability could not be demonstrat ed, the testing protocol was am ended for the immediate testing of samples collected from human subjects to ci rcumvent this limitation. The ease of detecting SCNin oral rinses was not without special consideration for the e ffect of matrix. For this reason,
122 isotopically labeled standards were used to mimi c any analyte suppression that may be imparted on the target analyte. It is ge nerally accepted that normalizing in tensity in units of peak area versus the peak area of the in ternal standard not only accounts for analyte suppression do to matrix but also allows for correction due to deviation in injection volume. Suggestions for future work include the valida tion of the utility of this method for serum to facilitate serum: sa liva concentration correlation. This method would then be useful for a variety the characterization of a variety of physiological process in vivo. Potential studies include the determination of SCNin response to diet or environmenta l factors to study th e effect of this analyte in the context of thyroi d inhibitors. It is recommended that measurements of the same sample be using two exclusive techniques be performed and analyzed using Bland Altman statistics.
123 CHAPTER 5 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK Conclusions Chapter 1 presented an o verview of breath analysis and discusses key issues that define future directions in this field. A review of historical achievements wa s included, with special emphasis on the practical limitations of the breath milieu for clinical diagnostics. Despite persistent interest throughout hi story, it was emphasized that on ly seven breath tests have met federal guidelines for implementation in the r outine clinical laboratory and law enforcement screening procedures. Current re search efforts in the scient ific community are focused on investigating the advantages of condensed-phase breath analysis. The most important advantage of condensed-phase breath analysis is the ab ility to monitor high molecular weight compounds, polar and nonvolatile compounds, and ions in breath due to the formation of aerosols in areas of high turbulence in the host airway during tida l breathing. These samples are very clean, with osmolatily measurements similar to deionized wa ter, allowing for the direct measurement of specific analytes with li ttle sample preparation. The analytical challenge is the determination of pg/mL to ng/mL levels of specific analytes in the presence of potential interferences at much higher concentrations. This chapter also detail ed why chromatographic separations in tandem with mass spectrometric detection were employed for this investigation, as well as theoretical and practical considerations for relevant instru mentation. Chapters 2 and 3 detailed the method development process in support of this work. In chapter 4, the novel reversed phase LC-ES I-MS/MS method presented in chapters 2&3 has been applied to the selective detection and quantification of [SCN-] in human oral rinses. A validation protocol was developed and applied to ensure confiden ce in the measured value for free concentrations of [SCN-]. The criteria for validation include analyte recovery, freezer
124 stability, post preparative stability (i.e., the e ffect of residence time in the autosampler), accuracy, precision and method detection limit. Th e advantages of this method include the elimination of laborious and complex sample cleanup (and derivatization chemistry) that was required for several previously reported methods81. A human subject protocol for the collection of oral fluid specimens was developed and a subsequent study was performed. The study was conducted under the guidelines for the protection of human subjects, including informed consent. The second part of Chapter 3 focused on wo rk inspired by the C onner research group, University of Miami Department of Cell Biology & Anatomy and Medicine. Recent publications from Conner et al. indicate that SCNmay be a potential biom arker for the pathogenesis associated with cystic fibrosis (CF). A brief review of the relevant literature was included. To support the efforts of the Conne r group, strategies to develop a noninvasive method to quantify [SCN-] in airway secretions were considered an d evaluated. Since e xhaled breath contains aerosolized secretions of the airway, EBC was evaluated as a matrix; however, [SCN-] could not be detected using current analytic al testing methods at the Univer sity of Florida (SPE-UV-vis). It was unclear if SCNwas not detected because of limited detection limits (1 ppm) or if SCNwas simply not present. A CE method was developed and briefly considered; however, difficulties arose with introducing the sample on column due to poor reproducibility and poor robustness. The CE platform was not suitable for population studies. A novel LC-ESI-MS/MS procedure was developed and yielded method de tection limits 100x lower than the reported values for the Lundquist method (described in chapter 1). The details of the instrument optimization were given in this chapter. This is the first known report of the separation of SCNunder reversed phase conditions. This is also the first report of using selected reaction
125 monitoring to detect SCN( m/z 5826). Despite the improvement in selectivity, SCNcould not (convincingly) be detected in EBC samples. Suggestions for Future Work Part I: Signal-to-Noise: Increase Signal Inadequate sensitivity and sele ctivity are lim itatio ns of current analytical methodologies associated with the implementation of EBC clin ical diagnostic methods, resulting in detection limits too high to examine the nonvolatile materials in breath. Two possible strategies to improve detection limits are: 1) increase signal (S ) 2) decrease noise (N ). One strategy to increase si gnal is the novel instrument m odification reported by Tang et al. (Pacific Northwest Laboratorie s) that details a microfabrica ted polycarbonate array with the ability to generate multiple Taylor cones (Taylor cones explained in Chapter 1) from the emitting spray capillary. Figure 5-1 shows an ex ample of the microfabricated array168. Of particular interest is the linear response in signal versus the number of sprays used. The maximum number of sprays that Tang observed to produce a linear response was nine. When considering detection limits are a function of S/N, the multiESI devi ce will improve detection limits by improving S. This phenomenon is due to the linear increase in sensitivity with increas ing Taylor cones as achieved by the array. In the next section, a portable instrument is proposed that combines the increased S with decreased N through the combin ation of the multiESI interface with FAIMS. Since the theoretical diluti on adjustment for an analyte in EBC is 20,000-fold less 20concentrated than in serum, the added sensitivit y for targeted analysis would be very useful. Sample preconcentration techniques, such as so lid phase extraction or nitrogen evaporation, may not be required if the suggested experiments in clude the evaluation of the multiESI emitter. The triple quadrupole mass spectrometer would be s uggested for these experiments given the many successful reports of quadrupole mass analyzers for targeted anal ysis in biological fluids.
126 A second possible route for future work is the application of the multiESI emitter in biomarker discovery. It is antici pated that this emitter, when pa ired with a quadrupole ion trap mass spectrometer (QITMS), will be a very sensitive platform for the characterization of a breath mixture. A third investigation may include fundamental spray emitter limitations. It is suspected that the Coulombic repulsions of multiple ion generators at close proximity would have a deleterious effect on ion signal. Modifications to the original design could possibly provide improved ion signal. Conclusions Part II: Signal-to-Noise: Decrease Noise The reduction of noise and increase in se lec tivity by the application of high-field asymmetric-waveform ion mobility spectrometr y (FAIMS) has been reported by several groups170-174. The principles of FAIMS has been reported in depth elsewhere; but to give a brief overview, the FAIMS cell is a post-ionization se paration device that operates at atmospheric pressure over a range of temperatures. FAIMS in strumentation has the ability to perform three tasks: 1) separate ions in the gas phase 2) focus ions 3) trap ions in three dimensions. Recent literature suggests that FAIMS dramatically improves the detection limits for several mass spectrometric methods through the reduction of in terferences afforded by increased selectivity. Future EBC studies centered on FAIMS may include the optimization of selective parameters to target specific ions in breat h. The development of a portable FAIMS device in breath analysis may also contribute to EBC res earch. The combination of the multispray emitter and the FAIMS may provide the platform with th e lowest possible detection limits over all. Evaluation of the instrument geometry may be useful. To conclude, current and future efforts in the field of breath analysis will be an important tool in the noninvasive diagnostics procedures th at are relevant to pa tient monitoring and the screening of drugs of abuse in forensic applica tions. Perhaps the real novelty is the discovery of
127 a new chemical marker for saliva contamination in EBC. Since SCNis so easily detected in saliva and is essentially not detected in most breath samples, it follows that SCNshould be evaluated as a chemical marker for saliva contam ination in EBC. Strategies to ensure sample integrity, such as this, should be considered in all stages of the development process.
128 APPENDIX A RECRUITMENT ADVERTISEMENT FOR HUMAN S UBJECTS STUDY Volunteers needed to participate in clinical research If anyone is interested in donating a saliva sample for a clinical research study, please contact us. We are looking for 50 volunteers. No restrictions on diet are required. Smokers and nonsmokers will be accepted. The donation process is noninvasive and takes approximately 15 minutes. Please contact FKLCMS@aol.com
129 APPENDIX B IRB APPROVED PROTOCOL AND INFORMED CONSE NT FORMS Protocol 1. Project Title: Development of a stable isotope dilution met hod to quantify endogenous thiocyanate in oral rinses by LC-MS/MS 2. Investigator: Frank Kero, Matthew Booth, Timothy Morey, Br uce Goldberger Richard Yost, Donn Dennis 3. Abstract: Thiocyanate (SCN-) has previously been reported as a biomarker for cyanide exposure due to smoking and environmental contaminants. Traditional methods to monitor SCNtypically require laborious sample preparation and suffer from poor detection limits. A novel LC-MS/MS method was developed to eliminate the need for sample cleanup/derivitization prior to analyte measurement. This application has demonstrat ed detection limits 100x lower than the SPE-UV method currently employed by Rocky Point laborato ries (UF). To validate this application, 50 volunteers will be recruited to donate oral rinse (saliva) samples. It is anticipated that this method will be able to differentiate between smokers and nonsmokers based on previously reported literature. It is also an ticipated that this method will be us ed as a generic platform for future medical and environmental applications. A point of novel science is the ability to retain an ion under reversed phase conditions which allows so lvent compatibility with the electrospray ionization source. 4. Background: The University of Florida has performed previ ous clinical studies that have focused on the quantification of [SCN-] as a biomarker to differentiate sm okers from nonsmokers. Smokers typically have a 3 fold higher concentrations of [SCN-]. 5. Specific Aims: The development and validation of a novel LC-MS-MS assay to quantify endogenous [SCN-] in the oral rinses of 50 volunteers.
130 6. Research Plan: 50 volunteers will report to the Department of Anesthesiology on a designated testing day. These volunteers will be give two pieces of dental ga uze. The volunteers will hold the gauze in their mouths for a period of 5 minutes. The volunteers w ill be free to leave after this collection period. For each volunteer, the gauze samples will be pl aced in to a Salivette centrifuged tube. These tubes will be centrifuged for 10 mi nutes at 3000 rpm. Three cycles of freezing and thawing will be performed to evaluate the effect of s hort term sample stability in the freezer. Saliva will be diluted 100x in to 60/40 Water/AcCN An internal standard of isotopically labeled 34SCNwill be spiked in to a concentration of 5g/L. The samples will be tested on the same day and quantified using a 6 point calibra tion curve. Results will be submitted as a manuscript to Clinical Chemistry. 7. Possible Discomforts and Risks: No additional risk will be added by the initiation of this study. 8. Possible Benefits: There are no possible benef its to the study subjects. 9.Conflict of Interest: There are no conflicts of interest.
131 IRB# 46-2007 Informed Consent to Pa rticipate in Research You are being asked to take part in a research study. This form provides you with information about the study and informs you of how your privacy will be protected. The Principal Investigator (the pers on in charge of this research) or a representative of the Princ ipal Investigator will also describe this study to you and answer all of your questions. Y our participation is entirely voluntary. Before you decide whether or not to take part, read the information below and ask questions about anything y ou do not understand. If you choose not to participate in this study you will not be penalized or lose any benefits to which you would otherwise be entitled. 1. Name of Participant ("Study Subject") ________________________________________________________________________ 2. Title of Research Study Development of a stable isotope dilution met hod to quantify endogenous thiocyanate in oral rinses by LC-MS/MS. 3. Principal Investigator and Telephone Number(s) Donn D Dennis, M.D. 846-1355 4. Source of Funding or Other Material Support University of Florida 5. What is the purpose of this research study? Thiocyanate is a chemical that is normally present in saliva. A method to monitor thiocyanate in saliva has been developed in our laboratory.
132 The purpose of this study is to evaluate the clinical value of this new method to measure thiocyanate in saliva (also known as oral rinses) from human subjects. 6. What will be done if you take part in this research study? You will place 2 pieces of sterile dental gauze in to your mouth for a period of five minutes. After five minutes, the gauze is removed and placed into a test tube. Ther e will be no identifiable information on this test tube and, once your sample is placed with other samples, we will be unable to determin e which sample is yours. At this time, the stud y is complete and you are free to go. 7. If you choose to participate in this study, how long will you be expected to participate in the research? This study will take no more than 15 minutes including 5 minutes for orientation and 5 minutes for specimen collection 8. How many people are expected to participate in this research? Up to 50 subjects, male and female, will take part in this study locally. 9. What are the possible discomforts and risks? There are no discomforts or risk s associated with this study. This study may include risks that are unknown at this time. If you wish to discuss the information above or any discomforts you may experience, you may ask questions now or call the Principal Investigator or contact person liste d on the front page of this form. 10a. What are the possible benefits to you? There is no personal benefit to you fo r participating in this study.
133 10b. What are the possible benefits to others? The possible benefit to others includes impr oved analytical methodologies that are more accurate and sensitive for the clini cal management of disease. 11. If you choose to take pa rt in this research study, will it cost you anything? There are no costs to you. 12. Will you receive compensation for taking part in this research study? You will not receive any compensation (money or services) for being in this study 13. What if you are injure d because of the study? If you experience an injury that is directly cau sed by this study, only professional consultative care that you receive at the University of Fl orida Health Science Ce nter will be provided without charge. However, hospital expenses will have to be paid by you or your insurance provider. No other compensation is offered. Please contact the Principal Investigator listed in Item 3 of this form if you experience an inju ry or have any questions about any discomforts that you experience while par ticipating in this study. 14. What other options or treatme nts are available if you do not want to be in this study? This is not a treatment study. The only other option is not to participate in this study. 15a. Can you withdraw from this research study? You are free to withdraw your consent and to stop participating in this research study at any time. If you do withdraw your consent, there will be no pena lty, and you will not lose any benefits you are entitled to. If you decide to withdraw your consent to participate in this research study for any reason, you should contact Donn D. Dennis, M. D., Ph.D. at (352) 846-1355. If you have any questions regarding your right s as a research subjec t, you may phone the Institutional Review Board (IRB) office at (352) 846-1494.
134 15b. If you withdraw, can information abou t you still be used and/or collected? If you withdraw immediately af ter providing the sample, the sample will be thrown away. However, should you wish to withdraw your sample at some later date, we will be unable to comply with your request as we will be unabl e to determine which sample is yours. 15c. Can the Principal Investigator withdraw you from this research study? You may be withdrawn from the study without your consent for the following reasons: Dr. Dennis or his designee is unavailable to collect the appropriate specimens. 16. How will your privacy and th e confidentiality of your research records be protected? Information collected about you w ill be stored in locked filing ca binets or in computers with security passwords. Only certain people have the legal right to review these research records, and they will protect the secrecy (confidentiality) of these records as much as the law allows. These people include the researchers for this stud y, certain University of Florida officials, the hospital or clinic (if any) involved in this res earch, and the Institutional Review Board (IRB; an IRB is a group of people who are responsible for looking after the rights and welfare of people taking part in research). Otherwise your research records will no t be released without your permission unless required by law or a court order. If the results of this research are published or presented at scientific me etings, your identity will not be disclosed. 17. How will the researcher( s) benefit from your being in this study? In general, presenting research results helps the ca reer of a scientist. Therefore, the Principal Investigator may benefit if the re sults of this study are presented at scientific meetings or in scientific journals.
135 18. Signatures As a representative of this study, I have ex plained to the participant the purpose, the procedures, the possible benefits, and the risks of this research study; the alternatives to being in the study; and how privacy will be protected: ________________________________________ ______ _______ ______________ Signature of Person Obtaining Co nsent Date You have been informed about this studys purpo se, procedures, possible benefits, and risks; the alternatives to being in the study; and how your privacy will be protected. You have received a copy of this Form. You have been gi ven the opportunity to as k questions before you sign, and you have been told that you can ask other questions at any time. You voluntarily agree to participate in this st udy. By signing this fo rm, you are not waiving any of your legal rights. ________________________________________ ______ _______ ______________ Signature of Person Consenting Date
136 LIST OF REFERENCES 1. Phillips, M (1992) Breath Tests in Medicine, Scie ntific American July, 74-79 2. Miekisch W, Schubert J, Noeldge-Schomburg G (2004) Diagnostic pote ntial of breath analysisfocus on volatile organic compounds Clinica Chimica Acta 347 (1-2), 25 3. Rahman I (2004) Reproducibility of oxidative stress biomarkers in breath condensate: are they reliable? European Respiratory Journal, 23,183 4. Mikuls T, ODell J, Ertl R, Bergman D, Renna rd SI (2006) Examining the exhaled levels of hydrogen peroxide in rheuma toid arthritis: a pilot study, Annals of Rheumatic Disease,65,1252-1253 5. Loukides S, Horvath I, Wodehouse T, Cole P, Ba rnes P (1998) Elevated levels of expired breath hydrogen peroxide in bronchiectasis, American Journal of Respiratory and Critical Care Medicine,158, 991-994 6. Wood L, Gibson P, Garg M (2003) Biomarkers of lipid peroxidation, airway inflammation and asthma, European Respiratory Journal 21,177-186 7. Gessner C, Kuhn H, Toepfer K, Hammerschmidt S, Schauer J, Wirtz H (2004) Detection of p53 gene mutations in exhaled breath condensate of non-small cell lung cancer patients, Lung Cancer, 2004, 43, 215-222 8. Cao W Duan Y (2006) Breath Analysis: Poten tial for Clinical Diagnosis and Exposure Assessment Clinical Chemistry, 52, 5, 800-811 9. Droz P, Guillemin M (1986) Occupational exposure monitoring using breath analysis, Journal of Occupational Medicine and Toxicology 28, 8, 593-602 10. Egeghy P, Nylander-French L, Gwin K, Hert z-Picciotto I, Rappaport, S (2002) Selfcollected Breath Sampling for Monitori ng Low-level Benzene Exposures among Automobile Mechanics Journal of Occupational Medicine and Toxicology, 28, 8, 593602 11. Burgaz S, Cok I, Ulusoy L, Tarhan U, Aygun N, Karakaya A (1997) Biological monitoring of n-hexane exposure in shoe workers, Biomarkers, 2, 1, 25 28 12. Carpagnano G, Kharitonov S, Foschino-Barbaro M, Resta O, Gramiccioni E, Barnes P (2003) Increased inflammatory markers in the exhaled breath conde nsate of cigarette smokers; European Respiratory Journal, 21, 589 13. Goldoni M, Catalani S, De Palma G, Manini P, Acampa O, Corradi M, Bergonzi R, Apostoli P, Mutti A (2004) Exhaled Breath Condensate as a Suitable Matrix to Assess Lung Dose and Effects in Workers Exposed to Cobalt and Tungsten, Environmental Health Perspectives 112, 13, 1293-1298
137 14. Dahnke H, Kleine D, Hering P, Murtz M (2001) Real-time monitori ng of ethane in human breath using mid-infrared cav ity leak-out spectroscopy, Applied Physics B: Lasers and Optics, 72, 8, 971-975 15. Balint B, Donnelly L, Hanazawa T, Kharitonov S, Barnes P (2001) Increased nitric oxide metabolites in exhaled breath condensat e after exposure to tobacco smoke, Thorax 56, 456-461 16. Risby T, Solga S (2006) Current status of clinical breath analysis, Applied Physics B: Lasers and Optics, 85 (2-3), 421-426 17. Sidorenko G, Zborovskii Levina D (1980) Su rface-active properties of the exhaled air condensate (a new method of studying lung function), Ter Arkh, 52, 65 18. Efrati O, Gonik U, Bielorai B, Modan-Mose s D, Neumann Y, Szeinberg A, Vardi A, Barak A, Paret G, Toren A, Fiberoptic bronchoscopy and bronchoalveolar lavage for the evaluation of pulmonary disease in childre n with primary immunodeficiency and cancer, Pediatric Blood & Cancer, 48, 3, 324-329 19. Larsen K, Tufvesson E, Malmstrom J, Morgel in M, Wildt M, Andersson A, Lindstrom A, Malmstrom A, Lofdahl C, Marko-Varga G, Bjermer L, Westergren-Thorsson G (2004) Presence of Activated Mobile Fibroblasts in Bronchoalveolar Lavage from Patients with Mild Asthma, American Journal of Respiratory and Critical Care Medicine, 170,10491056 20. Effros R, Hoagland K, Bosbous M, Castillo D, Foss B, Dunning M, Gare M, Lin W, Sun F (2002) Dilution of Respiratory Solu tes in Exhaled Condensates, American Journal of Respiratory and Critical Care Medicine ,165, 663-669 21. Effros R, Dunning M, Biller J, Shaker R ( 2004) The promise and perils of exhaled breath condensates, American Journal of Lung and Ce llular Molecu lar Physiology, 287, L1073L1080 22. Rosias P, Dompeling E, Hendriks H, Heijn ens J, Donckerwolcke R, Jobsis Q (2004) Exhaled breath condensate in children: Pearls and pitfalls, Pediatric Allergy and Immunology 15, 1, 4 23. ESI-MS/MS monitoring of trace le vels of thiocyanate and other low mass inorganic ions in exhaled breath (with F.A. Kero, L.K. Bajpai,, M.M. Booth, D. M. Dennis, G.E. Conner, R.A. Yost) The Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, June 5-10, 2005 San Antonio, Te xas, food and pharmaceuticals quantitation poster session 24. Quinton P (1999) Physiological Basis of Cys tic Fibrosis: A Historical Perspective, Physiological Reviews 79 ,1, S3-S22 25. Tait J, Gibson R, Marshall S, Cheng E, Sternan D, Cutting G(2001) Cystic Fibrosis, Gene Clinics
138 26. Arens R, Gozal D, Omlin K, Vega J, B oyd K, Keens T, Woo M (1994) Comparison of high frequency chest compression and conventi onal chest physiotherapy in hospitalized patients with cystic fibrosis, American Journal of Respiratory and Critical Care Medicine,150, 4, 1154-1157 27. Farrell P, Kosorok M, Rock M, Laxova A, Zeng L, Lai H, Hoffman G, Laessig R, Splaingard M (2001) Early Diagnosis of Cystic Fibrosis Through Neonatal Screening Prevents Severe Malnutrition and Improves Long-Term Growth, Pediatrics 107, 1, 1-13 28. Wijkstrom-Frei C, El-Chemaly S, Ali-Rachedi R, Gerson C, Cobas M, Forteza R, Salathe M, Conner G (2003) Lactoperoxida se and Human Airway Host Defense. American Journal of Respirat ory Celullar Molecular Biology 29, 206-212 29. Conner G, Salathe M, Forteza R (2002) Lactoperoxidase and H2O2 metabolism in the airway, American Journal of Respiratory Critical Care and Medicine 166, S57-61 30. Gerson C, Sabater J, Scuri M, Torbati A, Coffey R, Abraham J, Lauredo I, Forteza R, Wanner A, Salathe M, Abraham WM, Conner GE (2000) The Lactoperoxidase System Functions in Bacterial Cl earance of Airways, American Journal of Respiratory Celullar Molecular Biology, 22, 665-671 31. Conner G, Wijkstrom-Frei C, Randell S, Fernandez V, Salathe M(2007) The lactoperoxidase system links anion transpor t to host defense in cystic fibrosis. FEBS Lletters 581,271-278 32. El-Chemaly S, Salathe M, Baier S, C onner GE, Forteza R(2003) Hydrogen peroxidescavenging properties of normal human airway secretions. American Journal of Respiratory Critical Care and Medicine ,167,425-30 33. Conner G, Salathe M., Forteza R (2002) Lactoperoxidase and H2O2 metabolism in the airway. American Journal of Respiratory Critical Care and Medicine, 166,S57-61 34. Gerson C, Sabater J, Scuri M, Torbati A, Coffey R, Abraham J, Lauredo I, Forteza R, Wanner A, Salathe M, Abraham W, Conne r G (2000) The Lactoperoxidase System Functions in Bacterial Clearance of American Journal of Respiratory Celullar Molecular Biology,22,6,665-671 35. Van Sande J, Massart C, Beauwens R, Schout ens A, Costagliola S, Dumont J, Wolff J (2003) Anion Selectivity by the Sodium Iodide Symporter, Endocrinology, 144,1,247252 36. Bobadilla J,Macek M,Fine J,Farrell P (2002) Cy stic fibrosis: A worldwide analysis of CFTR mutations correlation with incidence data and application to screening, Human Mutation ,19,6,575 606 37. Wang Z,Milunsky J,Yamin M,Maher T,Oates R and Milunsky A (2002) Analysis by mass spectrometry of 100 cystic fibrosis gene mutations in 92 patients with congenital bilateral absence of the vas deferens, Human Reproduction 17,8, 2066-2072
139 38. Valdes M, Diaz-Garcia M (2004) Determina tion of thiocyanate w ithin physiological fluids and environmental samples: current practice and future trends, Critical Reviews in Analytical Chemistry, 34, 9-23 39. Laurberg P, Andersen S, Knudsen N, Ovesen L, Nohr S, Pedersen I (2002) Thiocyanate in Food and Iodine in Milk: From Domestic Animal Feeding to Improved Understanding of Cretinism, Thyroid 12, 10, 897-902 40. Heaney R,Fenwick G (1980)The analysis of glucosinolates in Bra ssica species using gas chromatography. Direct determination of the thiocyanate ion precursors, glucobrassicin and neoglucobrassicin, Journal of the Science of Food and Agriculture ,31,6, 593-599 41. Bialy Z,Oleszek W,Lewis J, Fenwick G (1990) Allelopathic potentia l of glucosinolates (mustard oil glycosides) and their degradation products against wheat, Plant and Soil ,129, 2,277-281 42. Wattenberg L (1992 )Inhibition of carcinoge nesis by minor dietary constituents Cancer Research 52,7 (Suppl),2085s-2091s 43. Glatz Z, Novakova S and Strebova H (2001) Anal ysis of thiocyanate in biological fluids by capillary zone electrophoresis, Journal of Chromatography A 916, 273-277 44. Lundquist P, Martensson J, Sorbo B, Oh man S (1979) Method for determining thiocyanate in serum and urine, Clinical Chemistry 25, 678-681 45. Lundquist P, Kagedal B, Nilsson L (1995) An improved method for determination of thiocyanate in plasma and urine, European Journal of Clinical Chemistry and Clinical Biochemistry 33, 343-349 46. Lundquist P, Rosling H, Sorbo B, Tibbling L (1983) Cyanide concentrations in blood after cigarette smoking, as determined by a sensitive fluorimetric method, Clinical Chemistry, 33, 1228-1230 47. Pawliszyn J (2002) Sampling and sample preparation for field and laboratory: Fundamentals and application, Critical Reviews in Oral Biology and Medicine 13,2,197212 48. Schramm W, Smith RH, Craig PA, Kidwell DA (1992) Drugs of abuse in saliva: a review, Journal Analytical Toxicology ,16, 1,1-9 49. Winefordner J, (2003). Sample preparation Techniques in Analytical Chemistry: A series of Monographs on Analytical Chemistry, John Wiley and Sons 50. Busch K, Spectroscopy (17) 10, October 2002 51. Kooley K (2003) Tandem mass spectrome try in the clinical laboratory, Clinical Biochemistry 36, 471-481
140 52. Gillespie T (Ph.D. 1988) Dissertation : "Concepts for the Determination of Prostaglandins by Tandem Mass Spectrometry, University of Florida 53. Yost R and Enke C (1979) Triple quadr upole mass spectrometry for direct mixture analysis and structure elucidation, Analytical Chemistry, 51, 1251AA 54. Cole R (1997) Electrospray Ionization Mass Spectrometry: Fundamentals, Instrumentation and Applications John Wiley & Sons 55. King R, Bonfiglio R, Fernandez-Metzler C, Miller-Stein C, Olah T (2000) Mechanistic investigation of ionization suppres sion in electrospray ionization, Journal of the American Society of Mass Spectrometry 11, 942. 56. Constantopoulos T, Jackson G and Enke C ( 1999) Effects of salt c oncentration on analyte response using electrospray ionization mass spectrometry, Journal of the American Society of Mass Spectrometry 10, 625. 57. Kebarle P (2000) A brief overview of the pres ent status of the mechanisms involved in electrospray mass spectrometry, Journal of Mass Spectrometry 35, 804. 58. Pedder R (2001) Practical Quadrupole Theory: Graphical Theory. Proceedings of the 49th ASMS Conference on Mass Spectrometry and Allied Topics, Chicago, IL. 59. Dawson P and Douglas D (1999) Use of quadrupoles in mass spectrometry, Encyclopedia of Spectroscopy and Spectrometry Academic Press 60. March R and Hughes R (1988) Quadrupole Storage Mass Spectrometry John Wiley & Sons 61. Dawson P (1976) Quadrupole Mass Spectrometry and its Applications, Elsevier 62. Pedder R (2002) Practical Quadrupole Theory : Peak Shapes at Various Ion Energies Proceedings of the 50th ASMS Conference on Mass Spectrometry and Allied Topics, Orlando, FL. 63. Busch K, Glish G and McLuckey S (1998) Mass Spectrometry/Ma ss Spectrometry: Techniques and Applications of Tandem Mass Spectrometry, VCH Publishers 64. Niessen W (1999) Applications of hyphenated techniques in mass spectrometry. Encyclopedia of Spectroscopy and Spectrometry Academic Press 65. McLafferty F (1983) Tandem Mass Spectrometry John Wiley & Sons 66. Fetterolf D, Yost R (1983) Tandem mass spectrometry (MS/MS) instrumentation, Mass Spectrometry Reviews 2, 1
141 67. Fetterolf D, Yost R (1984) Added resolution elements for greater informing power in tandem mass spectrometry, International Journal of Mass Spectrometry and Ion Processes, 62, 33 68. McCloskey J (1990) Methods in Enzymology : Volume 193, Mass Spectrometry Academic Press 69. Kero F, Pedder R, Yost R (2005) Quadrupol e Mass Analyzers: Theo retical and Practical Considerations Encyclopedia of Genetics, Genomics, Proteomics and Bioinformatics, John Wiley & Sons 70. Javahery G, Thomson B (1998) LINAC Journal of the Amer ican Society for Mass Spectrometry 8, 697 71. Dodonov A, Kozlovsky V, Loboda A, Raznikov V, Sulimenkov I, Tolmachev A, Kraft A, Wollnnik H, LINAC 2, Rapid Communications in Mass Spectrometry 1997, 11, 1649 72. Thomson B, Joliffe C, Javahery R Proceedings of the 44th ASMS Conference on Mass Spectrometry and Allied Topics: Portland OR, May 12-16, 1155 73. Mansoori B, Dyer E, Lock C, Bateman K: Boyd R: Thomson B (1998) Analytical performance of a high-pressure radio fr equency applied by using conical rods, Journal of the American Society for Mass Spectrometry 9, 775 74. Cole R, Kebarle P, Ho Y (1997) On the mechanisms of Electrospray Ionization Mass Spectrometry Wiley and Sons, New York 75. Schug K, McNair H (2002) Adduct formation in electrospray ioniza tion. Part 1: common pharmaceuticals, Journal of Separation Science, 25, 760-766 76. Schug K, McNair H (2003) Adduct forma tion in electrospray ionization mass spectrometry part II benzoic acid derivatives, Journal of Chromatography A, 985, 531539 77. Wu Z, Gao W, Phelps M, Wu D, Miller D, and Dalton J (2004) Favorable Effects of Weak Acids on Negative-Ion Electrospray Ionization Mass Spectrometry, Analytical Chemistry, 76, 839-847 78. Cech N, Enke C (2001) Practi cal implications of some recent studies in electrospray ionization fundamentals, Mass Spectrometry Reviews 20, 362. 79. Fenn J, Mann M, Meng C, Wong S, Whitehouse C (2005) Electrospray ionizationprinciples and practice, Mass Spectrometry Reviews 9, 1, 37 70 80. Cole R, Zhu J (1999) Chloride anion attachment in negative ion el ectrospray ionization mass spectrometry, Rapid Communications in Mass Spectrometry 13, 7, 607 611
142 81. Kebarle P (2000) A brief overview of the pres ent status of the mechanisms involved in electrospray mass spectrometry. Journal of Mass Spectrometry 35, 804 82. Kebarle P, Peschke M (2000) On the m echanisms by which the charged droplets produced by electrospray lead to gas phase ions, Analytica Chimica Acta, 406, 11-35 83. Kebarle P (1993) Dependence of Ion Intensity in Electrospray Mass Spectrometry on the Concentration of the Analytes in the Electrosprayed Solution Journal of Mass Spectrometry 35, 804-817 84. Ikonomou M, Blades A, Kebarle P (1991) Elec trospray mass spectrometry of methanol and water solutions suppression of electric discharge with SF6 gas, Journal of the American Society for Mass Spectrometry, 2, 6, 497-505 85. Watson J (1997) Introduction to Mass Sp ectrometry, Third Edition, Lippincott-Raven Publishers 86. Ospina M (1998) Internal Energy Effects in Tandem Mass Spectrometry, dissertation, University of Florida 87. Gillespie T, Yost R (1990) Differences in the low energy collision activated dissociation of the carboxylate anions from stru cturally similar prostaglandins E2, F2 D2 and DHKF2 Journal of the American Society for Mass Spectrometry ,1,389-396 88. Li L, Campbell D, Bennet P, Henion J (1996) Acceptance criteria for ultratrace HPLC tandem mass spectrometry, Analytical Chemistry, 68, 3397-404 89. Kushnir M, Rockwood A, Nelson G, Bingfa ng Y, Urry F (2005) Assessing analytical specificity in quantitative analys is using tandem mass spectrometry, Clinical Biochemistry 38, 319-327 90. Campbell S, Rodgers M, Marzlaff E, Beauch amp J (1995) Deuterium exchange reactions as a probe of biomolecule structure, Journal of the American Chemical Society ,117, 12840-12854 91. Zhou S and Cook K (2000) Protonation in el ectrospray mass spectrometry: wrong-wayround or right-way-round? Journal of the American Society for Mass Spectrometry 11, 961-966 92. Ardrey B (2003) Liquid Chromatography Mass Spectrometry: An introduction John Wiley and Sons 93. Hayen H, Karst U (2003) Strategies for th e liquid chromatographicmass spectrometric analysis of non-polar compounds Journal of Chromatography A 1000,1-2, 549-565
143 94. Chan E, Ho P (2000) High-performance liq uid chromatography/atmospheric pressure chemical ionization mass spectrometric met hod for the analysis of catecholamines and metanephrines in human urine, Rapid Communications in Mass Spectrometry ,14, 21, 1959 1964 95. Nakazawa H, Ino S, Kato K, Watanabe T, Ito Y and Oka H (1999) Simultaneous determination of residual tetracyclines in foods by high-performance liquid chromatography with atmospheric pre ssure chemical ionization tandem mass spectrometry, Journal of Chromatography B: Biomedical Sciences and Applications ,732, 1, 10, 55-64 96. Glish G, Vachet R (2003) The Basics of Mass Spectrometry in the 21st century Nature Reviews 2, 140-150 97. Cole R (1997) Electrospray Ionization Mass Spectrometry: Fundamentals, Instrumentation and Applications, John Wiley & Sons 98. Karas M, Bachmann D, Bahr U, Hillenkamp F (1987)Matrix-assisted ultraviolet laser desorption of non-volatile compounds, International Journal of Mass Spectrometry and Ion Processes, 78, 53-68 99. Zhou S, Hamburger M (2005) Effects of solven t composition on molecular ion response in electrospray mass spectrometry: Investigation of the ionization processes, Rapid Communications in Mass Spectrometry 9, 15, 1516 1521 100. Cole R. (2000) Some tenets pertaining to electrospray ionization mass spectrometry, Journal of Mass Spectrometry 35, 763-772 101. Blades A, Ikonomou M, Kebarle P (1991) M echanism of electrospray mass spectrometry. Electrospray as an electrolysis cell, Analytical Chemistry, 63, 2109-2114 102. Tang K, Smith R (1999) Theoretical prediction of charged droplet evaporation and fission in electrospray ionization, International Journal of Mass Spectrometry 185-187, 97-105 103. Van Berkel G (2000) Electrolytic deposition of metals on to the high-voltage contact in an electrospray emitter: implications for gas-phase ion formation, Journal of Mass Spectrometry 35, 773-783 104. De la Mora J, Van Berkel G, Enke C, Cole R, Martinez-Sanchez M, Fenn J (2000) Electrochemical processes in electrospray ionization mass spectrometry, Journal of Mass Spectrometry 939-952 105. Bard A, Faulkner L (2001) Electrochemical Methods: F undamentals and Applications Second Edition John Wiley and Sons
144 106. Quarmby S, Yost R (1999) Fundamental studies of ion injection and trapping of electrosprayed ions on a quadrupole ion trap, International Jo urnal of Mass Spectrometry 190-191, 81-102 107. Cech N, Enke C (2001) Pract ical implications of some recent studies in electrospray ionization mass spectrometry, Analytical Chemistry 73, 4632-4639 108. Kebarle P (1993) Dependence of Ion Intensity in Electrospray Mass Spectrometry on the Concentration of the Analytes in the Electrosprayed Solution, Journal of Mass Spectrometry 35, 804-817 109. Olumee Z, Callahan J, Vertes A (1998) Dropl et Dynamics Changes in Electrostatic Sprays of Methanol-Water Mixtures, Journal of Physical Chemistry A 102, 9154-9160 110. Znamenskiy V, Marginean I, Vertes A ( 2003) Solvated Ion Evaporation from Charged Water Nanodroplets, Journal of Physical Chemistry A 107, 7406-7412 111. Zhou S, Prebyl B, Cook K (2002) Profiling pH Changes in the Electrospray Plume, Analytical Chemistry, 74,19, 4885-4888 112. http://webserver.lemoyne.edu/ faculty/giunta/rayleigh.html 113. Taflin D, Ward C (1989) Electrified Droplet Fission and the Rayleigh Limit Langmuir 5, 376-384 114. De la Mora, J (2000) Electrospray ionization of large multiply charged species proceeds via Doles charged residue mechanism, Analytica Chimica Acta 406, 93-104 115. Busch K, Glish G and McLuckey S (1998) Mass Spectrometry/Ma ss Spectrometry: Techniques and Applications of Tandem Mass Spectrometry, VCH Publishers-dup? 116. Cech N and Enke CG (2001) Practical implications of some recent studies in electrospray ionization fundamentals. Mass Spectrometry Reviews 20, 362 117. Cole R (1997) Electrospray ionization mass spectrometry, Fundamentals, Instrumentation and Applications, John Wiley 118. Constantopoulos T, Jackson G and Enke C ( 1999) Effects of salt concentration on analyte response using electrospray ionization mass spectrometry, Journal of the American Society for Mass Spectrometry 10, 7, 625-634 119. Bruins A (1998) Mechanistic aspe cts of electrospray ionization, Journal of Chromatography A, 794,1-2, 345-357 120. Bruins A (1994) Atm ospheric-pressure-ioniz ation mass spectrometry. I: Instrumentation and ionization techniques, Trends in Analytical Chemistry ,13, 37-43
145 121. Braithwaite A, Smith F (1999) Chromatographic Methods, fifth edition, KluwerAcademic 122. Skoog, Holler, Nieman, (1997) Principles of Instrumental Analysis fifth edition Saunders Golden Sunburst 123. Dolan J (2005) Stay away from the cliffs, LCGC North America, 23, 1, 30-35 124. Sojo L, Lum G, Chee P (2003)Internal sta ndard signal suppression by co-eluting analyte in isotope dilution LC-ESI-MS, The Analyst,51-54 125. Enke C (1997) A predictive model for matr ix and analyte effects in electrospray ionization of singly ch arged ionic analytes, Analytical Chemistry 69, 4885-4893 126. Wu Z, Gao W, Phelps M, Wu D, Miller D, Dalton J (2004) Favor able Effects of Weak Acids on Negative-Ion Electrospra y Ionization Mass Spectrometry, Analytical Chemistry 76, 839-847 127. Van Berkel G, Quirke J, Tigani R, D illey A, Covey T (1998) Derivatization for Electrospray Ionization Mass Spectrometry. 3. Electrochemically Ionizable Derivatives, Analytical Chemistry, 70, 8, 1544 1554 128. The Belmont Report, Office of the Secretary, Ethical Principles and Guidelines for the Protection of Human Subjects of Research, The Nati onal Commission for the Protection of Human Subjects of Biomedical and Behavioral Research 129. Code of Federal Regulations, title 45 part 46, public welfare depart ment of health and human services protectio n of human subjects 130. Schultza C, Ahmeda M, Dawesb C, Mantscha H (1996) Thiocyanate Levels in Human Saliva: Quantitation by Fourier Transform Infrared Spectroscopy, Analytical Biochemistry 240, 1, 7-12 131. Causon R(1997) Validation of chromatogra phic methods in biomedical analysis: viewpoint and discussion, Journal of Chromatography B 689, 175-180 132. Williams K (2005) Laboratory manual for instrumental analysis 133. Clayton C, Hines J, Elkins P (1987) Det ection limits with specified assurance probabilities, Analytical Chemistry 59, 2506-2514 134. Currie L (1999) Nomenclature in evaluation of analytical methods including detection and quantification capabilities IUPAC Recommendations 1995, Analytica Chimica Acta, 391, 2, 105-126 135. Analytical Methods Co mmittee (Royal Soci ety of Chemistry, UK) Recommendations for the definition, estimation and use of detection limit, The Analyst, 112, 199-204 (1987)
146 136. Long G, Winefordner J (1983) Limit of De tection: A closer look at the IUPAC definition Analytical Chemistry, 712A-742A 137. Schneyer L (1956) Source of re sting mixed saliva in men, Journal of Applied Physiology 9, 79-81 138. Windholz M, Budavari S, Fertig M, Albers-Sch onberg G, Table of Molecular Weights, A Companion Volume to the Merck In dex, Merck and Co. Inc., pp. APP1-6 139. Valentn-Blasini L, Mauldin J, Maple D, Blount B (2005) Analysis of Perchlorate in Human Urine Using Ion Chro matography and Electrospray Tandem Mass Spectrometry, Analytical Chemistry, 77, 8, 2475 -2481 140. Dua S, Maclean M, Fitzgerald M, McAnoy A, Bowie J (2006) Is the hypothiocyanite anion (OSCN)the major product in the peroxidase cat alyzed oxidation of the thiocyanate anion (SCN)-? A joint experimental and theoretical study, Journal of Physical Chemistry A, 110, 14, 4930-6 141. Miller R, Brown B(1997)Beyond ANOVA: Basics of Applied Statistics, CRC press 142. Bross I (1971) Critical levels, statistical language and scientific inference, in Godambe VP and Sprott" Foundations of Statistical In ference, Toronto, Holt, Reinhart & Winston of Canada LTD 143. Hughes N,Wong E, Fan J, Bajaj N ( 2007) Determination of Carryover and Contamination for Mass SpectrometryBased Chromatographic Assays, AAPS Journal 9, 3, E353-E360 144. Garden J, Mitchell D, Mills W (1980) Nonc onstant variance regression techniques for calibration curve based analysis, Analytical Chemistry 52, 2310-2315 145. Dolan J (2005) Stay away from the cliffs, LCGC North America, 23, 1, 30-35 146. Bush K (2002) Chemical Noise in Mass Spectrometry part 1, Spectroscopy,17,10,32-37 147. Bush K (2002) Chemical Noise in Mass Spectrometry part 2, Spectroscopy,18,2,56-62 148. Bush K (2002) Chemical Noise in Mass Spectrometry part 3, Spectroscopy,18,5,52-55 149. Vass G, Huszar E, Barat E, Valyon M, Kiss D, Penzes I, Augustinovicz M (2002)Comparison of nasal and oral inha lation during exhaled breath condensate collection, American Journal of Respiratory and Critical Care Medicine, 167, 850-855 150. Soyer O, Dizdar E, Keskin O, Lilly C, Ka layci O (2006) Comparison of two methods for exhaled breath condensate collection, Allergy, 61, 8, 1016
147 151. Cap P, Chladek J, Pehal F, Maly M, Petr u V, Barnes P, Montuschi P (2004) Gas chromatography/mass spectrometry analysis of exhaled leukotrienes in asthmatic patients. Thorax 59,6,465-470 152. Papachristou G (2005) Inflammatory Markers of Disease Severity in Acute Pancreatitis, Clinics in Laboratory Medicine ,25,1,17 37 153. Boulos P, Whitfield P, Dave M, Faber R, Hobsley M (1980)Thiocyanate as a marker for gastric juice?, Gut ,18-22 154. Tang, K, Lin Y, Matson W, Kim T, Smith R (2001) Generation of mu ltiple electrosprays using microfabricated emitter arrays for improved mass spectrometric sensitivity, Analytical Chemistry, 73, 1658-1663 155. Purves R, Guevremont R, Day S, Pipich, C, Matyjaszczyk M (1998) Mass spectrometric characterization of a high-field asymmetr ic waveform ion mobility spectrometer, Review of Scientific Instruments 69,12, 4094-4106 156. Purves R, Guevremont R (1999) Electrospray ionization high-field asymmetric waveform ion mobility spectrometry-mass spectrometry, Analytical Chemistry 71, 2346-2357 157. Barnett. D, Guevremont R, Purves R (1999) De termination of parts-pe r-trillion levels of chlorate, bromate, and iodate by electrospra y ionization/high-field asymmetric waveform ion mobility spectrometry/mass spectrometry, Applied Spectroscopy, 53, 1367-1374 158. Barnett D, Purves R, Guevremont R ( 2000) Isotope separation using high-field asymmetric waveform ion mobility spectrometry, Nuclear Instruments and Methods in Physics Research A, 450,179-185 159. Futrell J (2000) Development of tandem mass spectrometry: one perspective, International Journal of Mass Spectrometry 200, 495-508
148 BIOGRAPHICAL SKETCH Frank Anthony Kero was born in Paterson, New Jersey in 1975. His fa mily later moved to East Brunswick, New Jersey where he was raised and attended high school. He graduated from Seton Hall University (SHU) with a BS in chemistry. His first research experience was as an undergraduate research assist ant under the supervisi on of Dr. John Sowa. His responsibilities included synthesizing organic precursors in suppor t of an axial chirality study. After graduation from SHU, Frank obtained employment in th e pharmaceutical industry and found focus and enjoyment in the field of chemical separations. Th is experience inspired a return to academics in pursuit of an advanced degree in analytical chemistry. Frank joined the Yost Research Group in 2003. He served as laboratory instructor for the ge neral chemistry (I & II) a nd analytical sections (quantitative analysis & instrument al analysis) as well as an unde rgraduate research mentor to four students. During his time at UF, Frank carried an additional appointment with the Department of Anesthesiology at Shands Hospita l. His responsibilities included analytical development in support of a number of depart mental projects includ ing a novel drug-delivery method for common anesthetics as well as preliminary evaluations for breath-based diagnostics and drugs-of-abuse screening procedures by LC-ESI-MS/MS and LC-APCI-MS/MS. He was very active with the regional chapter of the Am erican Chemical Society under the guidance of Dr. James Horvath and assisted with the responsibilities of the regi onal treasurer as needed. Prior to graduating from UF, Frank follo wed in the footsteps of several former Yosties as a research fellow at the Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia. His preliminary responsibilities at the CDC include the biomonitoring of selected populations for exposure to known thyroid i nhibitors (including SCN-) by IC-MS and IC-MS/MS.