PROFILES OF VOLATILE COMPOUNDS AS MICROBIAL MARKERS IN APPLICATIONS OF BIOSECURITY AND BIOENERGY By Kristen Leigh Reese A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree o f 2020 ABSTRACT PROFILES OF VOLATILE COMPOUNDS AS MICROBIAL MARKERS IN APPLICATIONS OF BIOSECURITY AND BIOENERGY By Kristen Leigh Reese All organisms emit volatile organic compounds (VOCs) and profil ing of volatiles (vol atilomics) is finding diverse applications. Some VOCs are consistently present throughout the lifecycle s of organism s , while other VOCs are biomarkers, quantifiable indicat or s of change s in physiological state or reflective of environmental stresses . Th is dissertation describes research into vola tile biomarkers of different microorganisms in the context of biosecurity and bioenergy. U ntargeted analys es of microbial biomarkers were accomplished using solid phase microextraction (SPME) coupled to gas chromato graphy - mass spectrometry (GC - MS) . In the context of biosecurity, pathogenic bacteria can be used as the basis for a bio - terrorism attack. There is a need for deeper understanding of the chemical signatures of organisms, in particular when they infect indi viduals, and a need for methods for detecting these pathogens in the context of infections of humans . Current research has performed metabolite profi ling of VOCs emitted in culture by surrogates for potential bacterial bioterrorism agents, Bacillus anthrac is Sterne and Francisella tularensis novicida in conjunction with measurements of VOCs released by their fully virulent counterparts, F. tularensis SCHU S4 and B. anthracis Ames , both on the CDC category A bioterrorism and disease agent list. Methyl ketone s, alcohols, esters, carboxylic acids, and nitrogen - and sulfur - containing compounds were attribute d to the bacteria. The two genera showed distinct VOC profiles whereas the taxa within each genus showed subtler differences in VOC profiles. Growth phase in fluenced absolute and relative VOC abundance s , indicating the potential for markers to discriminate growth phases. This in vitro determination of VOC profiles laid gro u ndwork for non - invasive prob ing of bacterial metabolism. Towards bioenergy efforts, micr oalgae present a renewable alternative to producing biofuels. However, biofuels are more costly per gallon compared to non - renewable fossil fuels due to production and harvesting costs. Therefore, research driving increases in biomass production are of int erest, specifically (1) better early - warning tools to anticipate and/or diagnose the presence of pr edators and (2) understand ing algae - bacteria interactions, as they are challenging to manage and may help or harm algal productivity . Research towards part ( 1) aimed to better define the physiological state of algae ponds . A biofuel - relevant alga, Microchloropsis salina , was infected with a predator, the rotifer Brachionus plicatilis . SPME - GC - MS aided discovery of seven putative culture crash biomarkers, inclu ding carotenoid degradation products trans - - - cyclocit ral, over several timepoints during active crashing of algal ponds that were not observed in healthy controls. T hese biomarkers offer potential as diagnostic tools to signal the need for cra sh mitigation strategies, as signals were detected before observed losses in algal cell density. Research towards part (2) aimed to detect and identify VOC biomarkers related to the micro - scale interactions of a model system of alga P. tricornutum and bact erium Marinobacter spp. 3 - 2 . The presence of Marinobacter spp. 3 - 2 , either in the form of live bacterial cells or sterile exudates , caused modest inhibition in growth rates of P. tricornutum . Substantial differences in VOC biomarker profiles were observed between 1) co - cultures of both organisms, 2) P. tricornutum exposed to Marinobacter spp. 3 - 2 exudates, and 3) Marinobacter spp. 3 - 2 exposed to P. tricornutum exudates, all relative to the VOC biomarker profiles of corresponding monocultures. Increasing the knowled ge base of algae - bacterial interactions will enable a deeper understanding of the basic science of microorganism signaling. iv ACKNOWLEDGEMENTS As this dissertation was the culmination of an unusual path, there are many people deserving of my deepe s t thanks for help along the way. First, I would like to thank to my committee. To my Michigan State University Ph.D. advisor Dr. A. Daniel Jones , y ou have led by example with high standards and strong work ethic, admirable qualities which , through em ulati o n, have allowed me to truly flourish. Your patience, understanding, and care have allowed me to finally reach the end of my graduate studies, and I will forever be grateful. To Dr. Matthias Frank, my LLNL mentor, thank you for your encouragement, sup port, and insight. You have helped me to envision the bigger picture of my research and provide context. Your invaluable mentorship over the past four years has truly been appreciated. Finally, to Dr. Dana Spence and Dr. Gary Blanchard , thank you both for servi n g on my committee and for guiding me through teaching assistant positions and academic courses. I would like to thank collaborators at Lawrence Livermore National Laboratory, including Dr. Amy Rasley, Dr. Julia Avila, Dr. Matt Cole man, Dr. Xavier Mayali, D r. Rhona Stuart, Dr. Vanessa Brisson, and Kristina Rolison. It has been a great privilege to work alongside these talented people and learn from them. I am extremely grateful to collaborators at Sandia National Laboratories , in par ticular Dr. Todd Lane, P a mela Lane, and Dr. Carolyn Fisher . Thank you for showing me the wonderful world of algae . I want to extend my appreciation to other members of the MSU Community that have helped me to not only persist but thrive over the entire course of my graduate caree r . To Dr. Ruth In particular, to v members of research group sharing your knowledge and experti se and helping me to hone my resea r ch skills. Thank you to Chu, a dear friend who has supported me throughout -- from listening to my presentations for the 1000 th time or willing to gr k now how I could have managed this journey without you. To Michelle McDaniel, Kristin Shannon, Tim Shannon, Tanner McDaniel, Corey Jones, Travis Bethel, and Pengchao Hao, thank you for reminding me to always embrace a life both inside and outside of the la b , to daily laughs on the Chemistry Chat, and that good friends persist even while far apart. To Yu - Ling Lien, always a voice of reason and wit with a smile, you are gone too soon but never far from my thoughts. Thank you to my family -- you are my reason to keep going. Mom, Dad, and Kelly, y my pursuit of h igher education understand my science - lingo. Y ou are m y biggest fan s , inspiring me even when I felt like nothing was going right. Finally, I will acknowledge several groups and funding sources that made my research possible. I thank the LLNL Fo rensic Science Center (FSC) and, in particular, Audrey Williams and Deon Anex, for use of l a boratory equipment and space and Roald Leif for assisting with instrumental setup and maintenance. Partial f unding for this work was provided by the LLNL Laboratory - Directed Research and Development (LDRD) project s 17 - LW - 021 and 19 - FS - 035 as well as by Sa n dia National Laborator ies. T his work was performed , in part s , under the auspices of the U.S. Department of Energy by Lawrenc e Livermore National Laboratory under Contract DE - AC52 07NA27344 and at Sandia National Laboratories for the U.S. Department of Ene r National Nuclear Security Administration under contract DE - NA0003525. vi PREFACE The Lawrence Livermore National Laboratory, Office of Scientific and Technical Information, Information Management (IM) number associated with this document is LLNL - TH - 81 1155. This document was prepared as an account of work sponsored by an agency of the U nited States government. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or i mplied, or assumes any legal liability or responsibility for the accuracy, completenes s, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to a ny specific commercial product, process, or service by trade name, trademark, manufact urer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Se curity, LLC. The views and opinions of authors expressed herein do not necessarily sta te or reflect those of the United States government or Lawrence Livermore National Security, LLC, and shall not be used for advertising or product endorsement purposes. vii TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ .......................... x LIST OF FIGURES ................................ ................................ ................................ ...................... xii CHAPTER 1: INTRODUCTION ................................ ................................ ................................ ... 1 1.1 VO LATILE MARKERS AS DIAGNOSTIC TOOLS TO ASSESS HEALTHY VS. UNHEALTHY STATES OF A VARIETY OF ORGANISMS ................................ ............... 1 1.2 OVERVIEW OF DISSERTATION C HAPTERS ................................ .............................. 2 1.2.1 VOC data acquisition and data processing Chapter 2 ................................ ............. 2 1.2.2 Health and Biodefense: Biomarkers for detection of pathogenic bacteria Chapter 3 ................................ ................................ ................................ ................................ ............. 3 1.2.3 Bioenergy: Evaluation of algal wounding signals in the presence or absence of predators Chapte r 4 ................................ ................................ ................................ .......... 6 1.2.4 Bioenergy: Introduction to algal - bacterial interactions Chapter 6 .......................... 8 REFERENC ES ................................ ................................ ................................ ............................. 11 CHAPTER 2: DEVELOPMENT OF AN UNTARGETED SPME - GC - MS WORKFLOW TO IDENTIFY VOLATILE BIOMARKERS EMITTED FROM LIQUID CULTURES ................. 15 2.1 INTRODUCTION ................................ ................................ ................................ ............ 15 2.2 STRATEGIES FOR DATA ACQUISITION OF VOLATILE METABOLITES ........... 15 2.2.1 Overview of common VOC samplers ................................ ................................ ...... 15 2.2.2 Sample collection and concentration via solid - phase microextraction .................... 17 2.2.3 Separation and detection of VOCs ................................ ................................ ........... 20 2.3 APPROACHES FOR DATA PROCESSING OF VOLATILE METABOLITES ........... 21 2.3.1 Untargeted metabolomics vs. targeted metabolite profiling ................................ .... 21 2.3.2 Development of data processing criteria to identify biomar ker signatures ............. 22 2.4 Example of VOCs contributed from commercial sampling flasks. ................................ .. 24 2.5 SUMMARY ................................ ................................ ................................ ...................... 25 REFERENCES ................................ ................................ ................................ ............................. 27 CHAPTER 3: METABOLIC PROFILING OF VOCs EMITTED BY THE PATHOGENS Francisella tularensis AND Bacillus anthracis IN LIQUID CULTURE ................................ .... 30 FOREWORD ................................ ................................ ................................ .......................... 30 3.1 INTRODUCTION ................................ ................................ ................................ ............ 30 3.2 METHODS ................................ ................................ ................................ ....................... 34 3.2.1 Strains and Growth Media ................................ ................................ ....................... 34 3.2.2 Preparation of Bacterial Headspace ................................ ................................ ......... 35 3.2.3 Sampling V OCs from Bacterial Headspace (RG2 strains) ................................ ...... 35 3.2.4 Sampling VOCs from Bacterial Headspace (RG3 strains) and Transfer of SPME Samples to BSL - 2 Facility ................................ ................................ ................................ 36 3.2.5 Determination of Bacterial Concentrations ................................ ............................. 37 3.2.6 Data Acquisition Parameters ................................ ................................ .................... 37 viii 3.2.7 Data Processing ................................ ................................ ................................ ........ 38 3.3 RESULTS ................................ ................................ ................................ ......................... 40 3.3.1 Results from RG2 species ................................ ................................ ........................ 42 3.3.2 Results from RG3 spe cies ................................ ................................ ........................ 54 3.4 DISCUSSION ................................ ................................ ................................ ................... 60 3.4.1 Ketones ................................ ................................ ................................ .................... 61 3.4.2 Alcohols ................................ ................................ ................................ ................... 62 3.4.3 Sulfur - containing compounds ................................ ................................ .................. 63 3.4.4 Nitrogen - containing compounds ................................ ................................ .............. 63 3.4.5 Esters and carboxylic acid compounds ................................ ................................ .... 64 3 .4.6 Evidence of dynamic metabolic processes ................................ .............................. 64 3.4.7 Comparison of VOC markers for RG3 vs. RG2 strains ................................ ........... 65 3.5 CONCLUSIONS ................................ ................................ ................................ ............... 67 APPENDICES ................................ ................................ ................................ .............................. 70 APPENDIX A: Figures ................................ ................................ ................................ ........... 71 APPENDIX B: Tables ................................ ................................ ................................ ............ 73 APPENDIX C: Protocol ................................ ................................ ................................ ......... 81 REFERENCES ................................ ................................ ................................ ............................. 90 CHAPTER 4. METABOLIC PROFILING OF VOCs IN THE HEADSPACE OF ALGAL CULTURES AS EARLY BIOMARKERS OF ALGAL POND CRASHES ............................... 95 FOREWORD ................................ ................................ ................................ .......................... 95 4.1 INTRODUCTION ................................ ................................ ................................ ............ 95 4.2 METHODS ................................ ................................ ................................ ....................... 99 4.2.1 Axenic algae cultures ................................ ................................ ............................... 99 4.2.2 Xenic marine rotifers ................................ ................................ ............................. 101 4.2.3 Preparation of cultures ................................ ................................ ........................... 101 4.2.4 Monitoring algae growt h and rotifer cultures ................................ ........................ 101 4.2.5 SPME headspace sampling and GC - MS data acquisition ................................ ..... 102 4.2.6 GC - MS data processing ................................ ................................ ......................... 103 4.3 RESULTS ................................ ................................ ................................ ....................... 104 4.3.1 Cell counts of infected and control cultures ................................ .......................... 104 4.3.2 Headspace VOC results ................................ ................................ ......................... 105 4.3.3 Abundance of VOCs ................................ ................................ .............................. 109 4.4 DISCUSSION ................................ ................................ ................................ ................. 111 4.5 CONCLUSIONS ................................ ................................ ................................ ............. 115 APPENDICES ................................ ................................ ................................ ............................ 117 APPENDIX A: Tables ................................ ................................ ................................ .......... 118 APPENDIX B: Fi gures ................................ ................................ ................................ ......... 123 REFERENCES ................................ ................................ ................................ ........................... 133 Chapter 5. EVALUATION OF DETECTION OF ALG AL - BACTERIAL INTERACTIONS BY TRACKING VOLATILE BIOMARKERS ................................ ................................ ................ 138 FOREWORD ................................ ................................ ................................ ........................ 138 5.1 INTRODUCTION ................................ ................................ ................................ .......... 138 5.2 METHODS ................................ ................................ ................................ ..................... 141 ix 5.2.1 Sample preparation ................................ ................................ ................................ 141 5.2.2 Estimation of algal cell d ensities ................................ ................................ ........... 144 5.2.3 VOC sampling ................................ ................................ ................................ ....... 145 5.2.4 VOC data acquisition, processing, and biomarker identification .......................... 145 5.3 RESULTS AND DISCUSSION ................................ ................................ ..................... 147 5.3.1 Cell counts of microorganisms in cultures ................................ ............................. 147 5.3.2 Headspace VOC results for Culture s of P. tricornutum and Marinobacter spp. 3 - 2 ................................ ................................ ................................ ................................ ......... 148 5.3.3 Headspace VOCs for monocultures of one species exposed to exudates from the other ................................ ................................ ................................ ................................ 155 5.4 CONCLUSION ................................ ................................ ................................ ............... 16 0 APPENDIX ................................ ................................ ................................ ................................ . 162 REFERENCES ................................ ................................ ................................ ........................... 177 Chapter 6: C ONCLUSIONS AND BROADER IMPACTS ................................ ....................... 181 x LIST OF TABLES Table 2.1 Recommended SPME polymer coatings for different analyte types, recommended by the commerc ial su ppl ier Supelco/Sigma - Aldrich (See SPME Applications Guide 925F) ........... 19 Table 3. 1 Number of VOCs from F. tularensis and B. anthracis taxa detected by GC - MS and remaining after application of fil terin g c riteria ................................ ................................ ............. 42 Table 3.2 Annotations of F. tularensis novicida - specific VOC markers through compound class, putative NIST ID, m/z , and retention index matching ................................ ................................ .. 44 Table 3.3 Annotations of B. anthracis Sterne - specific VOC markers through compound class, putative NIST ID, m/z , and retention index matching ................................ ................................ .. 44 Table 3.4 Average r ela tive abundances (Log - 10 Scale) of F. tularensis novicida - associated VOCs at all measured timepoints, separated by growth phase ................................ ..................... 48 Table 3.5 Average relative abundances (Log - 10 Scale) of B. a n thr acis Sterne - associated VOCs at all measured timepoints, separated by growth phase ................................ ................................ 50 Table 3.6 Annotations of F. tularensis SCHU S4 - specific VOC markers and average re lative abundances (n=3) at 6 and 24 hours (Hr) post inoculation ................................ ........................... 56 Table 3.7 Annotations of B. anthracis Ames - specific VOC markers an d average relative abundances (n=3) at 6 and 24 hours (Hr) post inoculation ................................ ........................... 58 Table A.3.1 Relative abundances of F. tularensis novic ida - associated VOCs for each replicate at all measured timepoints, separated by growth phase ................................ ................................ .... 73 Table A.3.2 Relative abundances of B. anthracis Sterne - associated VOCs for each replicate at all measured timepoints, separated by growth phase ................................ ................................ ......... 76 Table A.3.3 Bacterial concentration s o f F. tularensis SCHU S4 in modified Mueller - Hinton media and B. anthracis Ames in Brain - Heart Infusion media measured at the VOC sampling time points 6 h and 24 h post - inoculation of cultures. The numbers represent the mean of CFU/mL determined from 2 pla t e counts for each of the 3 culture replicates (total number of plate counts for each time point = 6), the e rrors represent standard deviation. ............................ 80 Table A.3.P.1 Protocol Document Revision His tor y ................................ ................................ .... 82 Table 4.1 VOCs robustly and repeatedly detected from Algae and Al gae + Rotifer experiments ................................ ................................ ................................ ................................ ..................... 107 Table A.4.1 Significant difference determination between mean levels of algal cell densities across replicates of Algae ( M. sali na ), Algae + Rotifer ( M. salina a nd B. p licatilis ) and Media ................................ ........ 118 xi Table A. 4.2 List of VOCs in Individual Experiments ................................ ................................ 119 Table 5.1 Description of experimental sample types, abbreviations, and number of sampl e replicates ................................ ................................ ................................ ................................ ..... 141 T able A.5.1 Annotations of VOC biomarkers measured from P. tricornutum sample s ( Algae , n=2) detected at several timepoints spanning 240 hours of sample growth ............................... 163 Table A.5.2 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 samples ( Bacteria , n=2) detect ed at sever al timepoints spanning 240 hours of sample growth ............. 165 T able A.5.3 Annotations of VOC bio markers measured from P. tricornutum and Marinobacter spp. 3 - 2 samples ( Co - cultures , n=3) detec ted at sever al timepoints spanning 240 hours of sample growth ................................ ................................ ................................ ............................. 166 Table A.5.4 Annotations of VOC biomarkers measured from P. tricornutum exudates ( AlgEx , n=2) detected at several timepoints span nin g 240 hour s of sample growth ............................... 168 Table A.5.5 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 exudates ( BacEx , n=2) detected at several timepoin ts spanning 240 hours of sampl e g rowth ................. 170 Table A.5.6 Annotations of VOC biomarkers measured from P. tricornutum + Marinobacter spp. 3 - 2 exudates ( Alg+BacEx , n=3) detected at several timepoints spanning 240 hours of sample gro wth ................................ ................................ ................................ ............................. 172 Table A.5.7 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 + P. tricornutum exudates ( Bac+AlgEx , n=3) detected at several timepoints spanning 240 hours of sample gr owt h ................................ ................................ ................................ ............................. 175 xii LIST OF FIGURES Figure 2.1 Schematic representing the partitioning of volatiles onto a SPME fiber (represented by black rectangle) suspended above a l iquid media, translatable to the experiments to be covered in Chapters 3, 4, and 5. The volatility of compounds is a spectrum, and some compounds with low volatility may still be detected depending on experimental conditions (e.g., sampling time, temper at u re). ................................ ................................ ................................ ................................ . 18 Figure 2.2 Schematic of a GC - MS system, where a sample collected on a SPME fiber is injected onto a capillary column. Flow o f helium carrier gas drives transport through the column, w her e analytes sep arate depending on interactions between the stationary phase and mobile phase (accelerated by increasing oven temperatures). Compounds exiting the column are ionized, fragmented, a nd separated in the mass analyzer (quadrupole, nominal m/z rep orted). Detect ed ions are summed by repetitive scanning through the mass spectrum, creating a total ion current chromatogram. Abundances of ions at each m/z value are stored, and mass spectra ca n be constructed from ion abundances for each m/z value. ................................ ................................ .. 21 Figure 2.3 Examples of the endogenous VOCs exhibited by polymer - comprise d sample flasks Polystyrene (PS), Polycarbonate (PC), and Polyethylene terephthalate glycol - modified (PEG) through compar iso n of the VOC to tal ion current chromatograms. Special annotation of peaks in PEG are given for 1) decamethyl - cyclopentasiloxane, 2) dodecamethyl - cyclohexasiloxane, and 3) tetradecamethyl - cycloheptasiloxane, high - outgassing VOCs contaminating the backgro und VOC profiles w hen no cultures are present. ................................ ................................ ................. 26 Figure 3.1 Examples of the chemical complexity exhibited by F. tularensis novicida cultures through comparison of the VOC total ion chromatogram s a t (a) 24 hours p ost inoculation, (b) corresponding Mueller - Hinton media control, and (c) enlarged overlay of both 1a and 1b, where stars indicate the bacteria - specific VOC emissions. ................................ ................................ ..... 41 Figur e 3 .2 Example workfl ow of criteria utilized to filter list of detected VOCs to bacteria - specific biomarkers produced during growth, s hown for F. tularensis novicida . ........................ 43 Figure 3.3 Growth curves of (a) F. tularensis novicida in modified Mueller - Hinton media over a 5 2 - hour time period and (b) B. anthracis Sterne in Brain - Heart Infusion media over a 24 - hour time period. Data points and error bars represent the means and standard deviations of CFU/mL de termined from thre e culture replicates. Red lines represent visually determined trends in bacterial growth across each growth phase. ................................ ................................ .................. 46 Figure 3.4 Mean combined peak areas (integrated detecto r c ounts) for compound s within individual classes at each timepoint post bacterial inoculation of cultures for (a) F. tularensis novicida and (b) B. anthracis Sterne. ................................ ................................ ............................ 51 Figure 3.5 PCA scores pl ots for VOC marker prof iles of (a) F. tularensis novicida and (b) B. anthracis Sterne generated using the peak areas of pathogen biomarkers across all timepoints. Each plot point represents one sample. Distinct chemical profiles were observed amongst label ed xiii growth phases. Corre sponding loadings plots to explain placement of samples are located in Appendix Figure A.3.1. ................................ ................................ ................................ ................ 53 Figure A.3.1 PCA loadings plots for VOC profiles of (a) F. tularensis novi cid a and (b) B. anthraci s Sterne generated using the relative abundances of pathogen VOCs measured using GC - MS across all timepoints (For PCA score plots see Figure 3.5). Points represent individual VOC markers (colored by compound class) explaining placem ent of samples on scores plot. ............. 71 Figure A.3.P.1 SPME fiber sampling of bacterial culture headspace ................................ .......... 86 Figure A.3.P.2 Submerging SP ME fiber exposed to headsp ace above liquid bacterial culture for viability testing. ................................ ................................ ................................ ............................. 89 Figure 4.1 (a) Adapted from McBride et al, showing 350 L open algal production ponds with a healthy algal p ond on the left compared wi th a crashed algal pond on the right. (b) Br achionus plicatilis (average length 160 µm), marine rotifer, in a field of algae, Microchloropsis salina . .. 97 Figure 4.2 Sch ema tic of experimental setup for growth M. salina (Algae, A) in the presence of B. plicatilis (Rotifer, R) for 5 days. Mass flow controllers (MFCs) mixed 1% CO2 with VOC - free air to sparge 15 L cultures at 150 cc min - 1. In total, three replicate experiment s ( Experiment 1, Experiment 2, Experiment 3) were performed using this setup. SPME fibers (Ex periment 1 used 1 fiber; Experiment 2,3 used 2 fibers) were used to sample the headspace of media blank (MB), Algae only (A), and Algae + Rotifer (A+R) carboys fo r 3 0 - 60 min each at various timepoints over 2 - 4 days. ................................ ................................ ................................ .............................. 100 Figure 4.3 Algae concentration as determined by fluorescence measurements collected for three experiments. Similar coloring and patterns re present biological replica tes of each condition: media blanks (MB), Algae (A), and Algae + Rotifer (A+R) cultures. Small fluorescence signals were observed in MB controls, most of which are not discernable on this scale. Error bars represent standard de via tion derived from duplicat e m easurements for each sample. Significance levels for conditions that exhibited p<0.05 are in Appendix Table A.4.1. Blue asterisks (*) indicate the time points for headspace VOC sampling by SPME fibers. ................................ ... 105 Figure 4.4 Example GC - MS chromatograms for observed VOCs sampled from Algae (A) and Algae + Rotifer (A+R) cultures between 16 and 24 min, (a) Total ion chromatogram with indicated VO Cs (Annotations See Table 4.1) , ( b - d) extracted ion chromato grams monitoring increase in compound 6 over time ( m/z 177, RT 23.46 min, RI 1495). ................................ ...... 106 Figure 4.5 Peak areas of extracted compound chromatograms for trans - - ionone an d - cyclocitral across Exper iments 1, 2, and 3, separated by biological replicates. Error bars represent standard deviation derived from duplicate measurements for each sample. The exposure time for SPME fibers was 30 minutes in Experiment 1 and 60 minu tes in Experiments 2 and 3. ................................ ................................ ................................ ................................ ........ 110 Figure 4.6 VOCs identified from the headspace of Algae + Rotifer cultures formed from the - carotene. Only those oxidation cl eav ages relevant to this study a re pictured, but all double bonds across the - carotene backbo ne are cleaved. .............. 112 xiv Figure A.4.1 Experimental mass spectra for all VOCs listed in Table 4.1, annotat ed either by NIST ID match or base peak and retention index pair. ................................ ............................... 123 Figure A.4.2 Peak areas of extracted compound chromatograms for Compounds 1, 2, 3, 5, an d 7 across Experiments 1, 2, and 3; Com pounds 4 and 6 are displayed in Figure 4.5 ...................... 130 Figure 5.1 Experimental setup during passive V OC sampling of algal and bacterial samples - using SPME fibers (n=1 fiber per 250 mL flask); avera ge exposure time ~ 2.5 hours); replicate cultures indicated by similar - colored labels. ................................ ................................ .............. 146 Figure 5.2 Growth curves of algae in samples containing P. tricornutum. Algae , Co - Culture , and Alg+BacE x , as determined by measurements of relative chlorophyll fluorescence units (RFUs) at increasing t imepoints post - inoculation of cultures. Averages RFUs (averaged over replicates) are plotted with error bars showing +/ - standard deviations. Colors of the er r or bars correspond to sample type in the legend. Counts of RFUs in ESAW, Bacteria, and BacEx samp les were simultaneously acquired but did not exceed 20 RFUs, hence considered to be negligible signal and omitted from this figure. VOCs were collected via SP M E sampling at 72, 120, 168, and 240 hours post - inoculation as indicated by arrows. ................................ ............. 148 Figure 5.3 Example total ion current chromatog rams for observed VOCs sampled from the Co - cultures and from ES AW at 240 hours post inoculation, with peaks indicated by * being potential VOC markers of algae - bacterial interactions. ................................ .............................. 150 Figure 5.4 Venn diagram of the overlapping VOC biomarkers amongst Co - cultures , Algae, and Bacteria as inclusive of markers detected across all measured timepoints. Ea ch group is derived from n=2 replicates. ................................ ................................ ................................ .................... 152 Figure 5.5 Venn diag ram of the overlapping VOC bi oma rkers among Algae and AlgEx across all measured timepoints. Each group is derived from n=2 replica tes. Greatest qualitative overlap occurred during sampling at the 72 hour timepoint. ................................ ................................ ... 156 Figu re 5.6 Venn diagram of the overlapping VOC biomarkers for organisms exposed to exudates from the respe ctive other species: (a) P. tricornutum exposed to Marinobacter exudates, (b) Marinobacter exposed to P. tricornutum exudates. Panel (c) evaluates the ove rla p of Co - cultures , with Alg+BacEx and Bac+AlgEx . All circles are summation of VOC biomarkers across all measured timepoints. The number of replicates for each sampl e type can be referenced in Table 5.1. ................................ ................................ ................................ ................................ 158 1 CHAPTER 1: IN TRO DUCTION 1.1 V OLATILE MARKERS AS DIAGNOSTIC TOOLS TO ASSESS HEALTHY VS. UNHEALTHY STATES OF A VARIE TY OF ORGANISMS The generation of volatile organic compounds (VOCs) from natural or man - made processes is ubiquitous throughout the world. VOCs have been de fined as low molecular mass carbon - containing compounds that have low boiling points and measurable vapor pressures at Standard Temperature and Pressure (20 °C, 1 atm) [1] . Human perception of VOCs is through the sensation of smell, whether fragrant or foul, arising from VOCs interacting with recept ors in the n asal cavity. Applications of VOCs are widespread in society, from industry pursuits - cosm etics, pharmaceuticals and medicine, environmental science, and food science to military pursuits - energy and national security. The study of volatiles in this ple thora has been recently referred to as a new branch of the - omics world, volatilomics, whe re the VOC profile of an organism [2 - 4] . R esearch within volatilomics has rap idl y expanded within the last decade [5] , indica tin g this field provides complementar y information to - abolomics and exposomics. The volatilome s of human, plant, and animal models are composed of VOCs that are consistently present throughout the lifecycle of the organism, while VOCs that vary in abundance are biomarkers, quantifiable indicator of a change in health state, such as an infection or disease or as a result of environmental exposure. The study of biomarkers in human and animal models has been acc ele rated by improvements in the analy tical approaches (sample collection and pre - concentration) and de velopment of novel instrumentation (chromatography and mass spectrometry), enabling the detection of qualitative and/or quantitative levels of VOCs in the he adspace of different sample matric es. 2 Areas of national security where the study of the volatilome has gained momentum include biosecurity and bioenergy. The measures designed to lessen the transmission or intentional exposure of infectious diseases to th e general public are characterized as biosecurity measures or biological countermeasures, while the restoration of a stable state to people, animals, or wildlife exposed to biological threats is characterized as remediation. Research presented in chapte r 3 investigates biosecurity efforts into detection of potential volatile biomarkers for diagnosing th e presence of pathogenic bacteria, specifically the species Francisella tularensis and Bacillus anthracis . For the area of bioenergy, increasing the produ cti vity of commercial algae ponds off ers potential to improve the economics of algal biofuels. In this light, research described in chapters 4 and 5 focuses on detection of volatile biomarkers indicative of biotic stress including the action of predators o r s ymbiosis/competition with bacteria in culture, respectively. 1.2 OVERVIEW OF DISSERTATION CHAPTERS 1.2.1 VOC data acquisition and data processing Chapter 2 Chapter 2 will present an overview of the data acquisition and data processing methods used in this dissertation. Solid - phase mic roextraction (SPME) is a popular approach to the collection and p re - concentration of VOCs in the gaseous headspace of samples and was used throughout multiple investigations . Within the context of this work, SPME offer e d n on - invasive, non - destructive monit oring of the metabolite profiles for microorganisms relevant to b iosecurity and bioenergy applications. Analytes adsorbed/absorbed onto the SPME fibers were separated and detected through a gas chromatography - mass spect rom etry (GC - MS) analysis. An untarget ed metabolomics approach was utilized in this work to broadly ann otate the metabolites 3 detected across growth states. VOCs were identified as candidate microbial biomarkers through development of filtering criteria desi gne d to eliminate background VOCs whi le allowing for biological variability observed in biological sys tems. 1.2.2 Health and Biodefense: Biomarkers for detection of pathogenic bacteria Chapter 3 Over a century of research has been dedicated to studying the relationship s between specific microorganisms and diseases caused by their interaction with humans and animals [6] . Some of the research has produced great benefits for human health and medicine, such as the development of vaccines and therapeutic drugs, which have increase d t he human lifes pan. Responsible and safe handling of infectious pathogens, facilitated by greater pr otective measures to minimize risks, occurs in thousands of labs daily, with few incidents of release into the general public. However, there is a subset of pathogens that pose severe threat s to human, animal, and plant health due to increased virulence, a nd some of them have been used deliberately to cause disease and instigate fear against military personnel and/or the general public. The U.S. Centers for Di sease Control and Prevention (CDC) lists pathogens that pose a high risk to national security as Ca tegory A agents because of high person - to - person transmission rates, high mortality rates, a high likelihood of public disruption, and requires special pr epa rations to ens ure public health [7] . Examples of Category A agents and the res ulting disease s include Bacillus anthracis (Anthrax), Clostridium botulinum (Botulism), Yersinia pe stis (Plague), Variola major (Smallpox), and Francisella tularensis (Tularemia). The National Institute of Allergy and Infection Diseases (NIAID) [8] y c hallenges to t great importance to develop (1) a deeper under standing of the organisms, in particular their 4 actions when actively infecting individuals, and (2) methods capable of detecting these pathogens in the en vir onment, in par ticular in the context of human/pathogen infections. One of the most direct routes f or human exposure for pathogens is through aerosolized microbes entering human airways via the mouth or nose during inhalation [9] . The pathogen can colonize and proliferate on human cells in the br onchial airways. Detection of pathogens at this early stage, before the individual displays relevant clinical symptoms, would facilitate early diagnosis a nd the work of healthcare providers by providing more time to treat an infection. Exhaled breath analy sis has been explored as a diagnostic tool due to large chemical complexity and information content [4] . Br eath is composed of thousand s o f VOCs and non - volatile trace compounds (e.g., proteins) embedded in aerosol droplets. Through the circulatory system, oxygen from the air is transferred via the lungs into blood, which is delivered to muscles and organs, and carbon dioxide is simultane ous ly removed. Besides oxygen and carbon dioxide exchanging at the blood - air interface , other compound s, both VOCs and non - volatiles, are incorporated into this system. The V OCs in this system are directly influenced by the imme diate environment of the sub jec t as well as the natural human metabolic processes. Thus, air exhaled from the lungs can be conside red a headspace of circulating blood, where compounds are constantly partitioning. Changes in breath volatiles in vivo have be en previously researched in rel ation to respiratory diseases, such as chronic obstructive pulmonary disease [10] and asthma [10] , or in relation to non - respiratory di seases, such as diabetes [11] and breast cancer [12] . The use of exhaled br eath holds additional advantages in the field of personal diagnostics, as the sampling is non - invasive and samplers have been developed that can be performed by subjects in a home environment. 5 Study of V OCs in breath research has also begun expansion beyon d in vivo profiling of biomarkers associated with a specific disease to design in vitro models that could approximate a host body, and thus study the cellular origins of VOCs. Select prior studies have e sta blished that metabolites from proliferating bacter ia detected in the headspace of in vitro cultures can be detected in the exhaled breath of infected animal models or humans, including Mycobacterium tuberculosis [13] , Heliobacter pyl ori [14] , and Pseudomonas aerugin osa [15] . However, replication of the complex cellular makeup of the human airways and respiration system in a cell cultu re is challenging. The generation of VOCs can occur f rom the mammalian cells and pathogen cells, which are likewise dependent on physiological state, and the host immune system may generate a specific response to the infection. Therefore, initial studies i nto the growth of pathogens can employ monocultures, allowing observations of species - specific and gro wth stage dependent VOC emissions on different nutrient - rich medias, all of which impact the volatile profile. Thereafter, increasing the complexity of in vi tro models, and eventually transitioning into in v ivo studies in animal/human models, would assist in establishing the potential origins of exhaled breath VOCs. Chapter 3 describes efforts to conduct comprehensive (untargeted) metabolite profiling of vo lat ile organic compounds (VOCs) emitted in culture by bacterial taxa Francisella tularensis (F. tulare nsis) novicida and Bacillus anthracis (B. anthracis) Sterne, surrogates for potential bacterial bioterrorism agents, as well as selective measurements of VOC s from their fully virulent counterparts, F. tular ensis SCHU S4 and B. anthracis Ames that are both on the CDC category A bioterrorism and disease agent list. F. tularensis and B. anthracis were grown in liquid broth for time periods that covered logari thm ic growth, stationary, and death phases. VOCs emitted from these cultures were collected from the h eadspace above the cultures as well 6 as from control samples without bacteria using solid ph ase microextraction (SPME) and were analyzed using gas chromato gra phy - mass spectrometry (GC - MS). Criteria were developed for distinguishing VOCs originating from bac teria versus background VOCs (originating from growth media or sampling devices). Analyses of collected volatiles revealed methyl ketones, alcohols, ester s, carboxylic acids, and nitrogen - and sulfur - containing compounds that were attributed to the bacteri a. The two bacterial genera showed distinct VOC profiles whereas the taxa within each genus showed more subtle differences in VOC profiles, illustrating t he potential for VOC profiles to distinguish biosecurity - relevant pathogens at the genus and species - l evel. Growth phase influenced VOC abundance, indicating the potential for VOC profiles to d iscriminate growth phases. This in vitro determination of VOC p rof iles lays the groundwork for non - invasive probes of bacterial metabolism. Such VOC profiles offer p rospects for detection of taxa - specific VOC biomarkers from various potential biowarfare ag ents, with applications in bio - detection and in targeted breath - ba sed diagnostics for detection of biological agents proliferating in the lungs of infected patients after a suspected biological attack. 1.2.3 Bioenergy: Evaluation of algal wounding signa ls in the presence or absence of predators Chapter 4 There is hi gh interest in the monitoring of VOC emissions from algal cultures due to the importance of algae a s renewable energy resource and as a potential commercial feed or food source [16,17] . Microalgae are important a uto trophs in ecological communities, serving as a sources of nutrition for higher organisms [18] . Microalgae are a div ers e collection of photosynthetic, microscopic organisms, with an estimated several thousand species r anging in size from 1 - 100 + microns [18] . Select algae species produce and accumulate an abundance of 7 lipids and carboh ydrates, which can be converted into biodiesel and ethanol, respectively, each of which can be used as alternative renewable energy source to compete with th e production of non - renewable petroleum [18,19] . Additiona lly, alga l proteins have a high nutritional value and can be ut ilized for human consumption. Biofuel production using algae as a precursor has several pot ent ial benefits. Algae fix atmospheric carbon dioxide and sunlight to create biomass for growth, produ cing approximately 50 - 66% of atmospheric oxygen as a by - product , and could be used to scrub industrial carbon dioxide emissions from the air [20] . High - production cul tivation of algae is facilitated by rapid growth rates . Algae can be grown in freshwater, saltwater , and terrestrial areas (including some not suited for agricult ural use) [19] . Algae can be cultured in the laboratory using controlled conditions, but relatively low volumes. Larger volumes of algae in com mercial practice involve culture in open ponds or photobioreactors, where the environmental and exposure variables are more challenging to control, simila r t o a natural marine habitat. However, generation and maintenance of low - cost volumes of biomass prod uction required for industrial pursuits is threatened by biological contaminations. In particular, algae ponds used in industrial bioma ss production are s usc eptible to pathogen or grazer infestation, resulting in pond crashes with high economic costs. Cur rent methods to monitor and mitigate unhealthy ponds are hindered by a lack of early indicators that precede culture crash [20] . It is known that algae produce VOC s a s secondary metabolites during growth, and levels are affected by the algal species, growth phase, and environmental conditions (light, pH, water, nutrients, dissolved gasses, etc.) [20] . The research presented in chapter 4 descri bes efforts towards identifying ear ly changes in the volatile biomarker metabolite profiles emitted from healthy and unhealthy culture s of Microchloropsis salina exposed to a known predator, the marine rotifer Brachionus plicatilis . An untargeted 8 analysis of VOCs, followed by preliminary id entification, was accomplished using SPME - GC - MS. The addition of B. plicatilis to healthy cultures of M. salina resulted in decreased algal cell numbers, relative to uninfected controls, and generated trans - - - cyclocitral, which were attribu ted to carotenoid degradation. The abundances of the carotenoid - derived VOCs increased with rotifer co nsumption of algae and were detected prior to when conventional monitoring methods indicated distress. The results indicat e that specific VOCs released by in fected algae cultures may be early indicators for impending pond crashes, providing a useful tool t o monitor algal biomass production and pond crash prevention. 1.2.4 Bioenergy: Introduction to algal - bacterial interacti ons Chapter 6 Expanding on th e d iscussion in Section 1.2.3, there are several routes to lowering the cost of biofuel production fro m microalgae to be competitive relative to petroleum. One avenue involves producing higher densities of alga l cultures in order to increase the biomass th at can be con verted into biofuel. In order to increase the density of algal communities, a potential r oute involves exploiting interactions between algae and microorganisms, such as bacteria, that naturally inhabit the same space. Bacteria and microalgae h ave coexisted over millions of years of evolution. Exchange of nutrients, signaling molecules, and oth er dissolved materials between organisms happens in the area of closest contact, which has been termed the phycosphere (analogous to the rhizosphere aroun d r oots) [21] . Within the phycosphere, the exchange of metabo lites, both volatile and non - volatile, can further affect the growth cycles of each organism. The detection, identification, and quantification of these m ole cules in complex communities would enable prediction and/or manipulation of these interactions for the pur poses of basic science and commercial applications. 9 Both symbiotic and inhibitory interactions have been observed between select combinations of a lga e and bacteria, where the complexity of the systems is limited in order to more accurately determin e the o rigins of identified metabolites. In combinations where bacteria promote algal growth, identification of exuded metabolites could identify targets tha t could be added to commercial cultures, and thus increase biomass production [22] . Several studies have concluded that algae can utilize vitamins and recycled minerals exuded by bacteria, while bacteria ca n utilize exuded ammonia from algae as a nitrogen source [2 3 - 25] . In combinations where bacteria inhibit algal growth, competition occurs between the organisms for nutrients and dissolved organic materials [2 6] . The effects can be either specific to a species or inhibit a broad range of algae. Studies have i ndicated that inhibitory or non - growth promoting interactions play an important role in organizing the structure of marine communities and ensuring surviv al [26,27] . Identification of inhibitory compounds may allow control of harmful phenomenon in algal systems, such as algae blooms, that produce toxic effects on marine wildlife and can contaminate sources of drinkin g w ater [28] . Chapter 5 presents efforts to detect, identify, and understand how algae and bacteria interact at the microscopic level using model alg ae - bacteria co - cultures. To this end, volatile metabolites detected in the headspace of co - cultures of model algae Phaeodactylum tr icornutum (P. tricornutum) and model bacteria Marinobacter spp. 3 - 2 were investigated. P. tricornutum is an alga used for bio fue l production because of its high lipid content (estimated ~30% of dry weight [29] ), and Marinobacter is a diverse genus of Gram - neg ative, aerobic bacteria found in most oceans [30] . This work demonstrated the feasibility of an untargeted SPME - GC - MS profiling of VOC marke r d etection and researched how VOC profiles changed between exponential and sta tionary growth states. Additionally, biomarkers from cultures of algae or 10 bacteria supplemented with exuded materials from the opposing organism were profiled to check if the sa me biomarkers from c o - cultures could be replicated when the presence of one spe cies is true, this would indicate soluble compounds produced by one species may trigger production of the VOCs by the other species. The presence of Marinobacte r s pp. 3 - 2, either as live bacteri a or as exudate containing bacterial m etabolites, inhibited the grow th of P. tricornutum to a small degree. Compared to monocultures of algae or bacteria, some differences were observed in biomarker profiles for 1) co - cult ure s of both organisms, 2) P. tricornutum exposed to Marinobacter spp. 3 - 2 exudates, and 3) Marinobact er spp. 3 - 2 exposed to P. tricornutum exudates. Increasing the knowledge base of algae - bacterial interactions at the phycosphere and alterations in microo rga nism physiology will enable better prediction and/or manipulation of these interactions for commerc ial purposes as well as a deeper understanding of the basic science of microorganism signaling. 11 REFERENCES 12 REFERENCES 1 U.S.E.P.A. Technical Overview of Volatile Organic Compounds , (2017). 2 Amann, A. et al. The human volatilome: volatile organic compounds (VOCs) in ex hal ed breath, skin emanations, urine, feces and saliva. J. Breath Res. 8 , 034001, doi:10.1088/1752 - 71 55/8/3/034001 (2014). 3 Rowan, D. D. Volatile metabolites. Metabolites. 1 , 41 - 63, doi:10.3390 /metabo1010041 (2011). 4 Breathborne Biomarkers and the Human Vol atilome . 2nd edn, (ed. Beauchamp, J., Davis, C, Pleil, J. D.) (Elsevier, 2020). 5 Broza, Y. Y., M ochalski, P., Ruzsanyi, V., Amann, A. & Haick, H. Hybrid volatolomics and disease detection. Angew. Chem. 54 , 11036 - 11048, doi:10.1002/anie.201500153 (2015 ). 6 Gest, H. The discovery of microorganisms by Robert Hooke and Antoni van Leeuwenhoek, Fellows of The Royal Society. Notes and Records of the Royal Society of London 58 , 187 - 201, doi:doi:10.1 098/rsnr.2004.0055 (2004). 7 U.S. Centers for Disease Control and Prevention. CDC. Bioterrorism Agents/Diseases , (2018). 8 U.S. Department of Heath and Human Services, U.S. National Institues of Heal th. NIAID Strategic Plan for Biodefense Research. (Bethesda, M D, 2007). 9 Tellier, R., Li, Y., Cowling, B. J. & Tang, J. W. Recognition of aerosol transmission of infectious agents: a commentary. BMC Infectious Diseases 19 , 101, doi:10.1186/s12879 - 019 - 3707 - y (2019). 10 van de Kant, K. D. G., van der Sande, L. J. T. M ., Jöbsis, Q., van Schayck, O. C. P. & Dompeling, E. Clinical use of exhaled volatile organic compoun ds in pulmonary diseases: a systematic review. Respir Res 13 , 117, doi:10.1186/1465 - 9921 - 13 - 1 17 (2012). 11 Minh, T. D. C., Blake, D. R. & Galassetti, P. R. Th e clinical potential of exhaled breath analysis for diabetes mellitus. Diabetes Res Clin Pract 97 , 195 - 205, doi:10.1016/j.diabres.2012.02.006 (2012). 12 Lavra, L. et al. Investigation of VOCs associated with different characteristics of breast cancer ce lls . Scientific reports 5 , 13246, doi:10.1038/srep13246 (2015). 13 13 Phillips, M. et al. Volatile bioma rkers of pulmonary tuberculosis in the breath. Tuberculosis 87 , 44 - 52, doi:10.1016/j.tube.2006.03.004 (2007). 14 Lechner, M. et al. H. pylori infection inc rea ses levels of exhaled nitrate. Helicobacter 10 , 385 - 390 (2005). 15 Shestivska, V. et al. Quantific ation of methyl thiocyanate in the headspace of Pseudomonas aeruginosa cultures and in the breath of cystic fibrosis patients by selected ion flow tube mas s s pectrometry. Rapid Communications in Mass Spectrometry 25 , 2459 - 2467 (2011). 16 Odjadjare, E. C., Mutanda, T. & Olaniran, A. O. Potential biotechnological application of microalgae: a critical review. Crit Rev Biotechnol 37 , 37 - 52, doi:10.3109/07388551. 201 5.1108956 (2017). 17 Hamed, I. The Evolution and Versatility of Microalgal Biotechnology: A Review . Comprehensive Reviews in Food Science and Food Safety 15 , 1104 - 1123, doi:10.1111/1541 - 4337.12227 (2016). 18 Chisti, Y. Biodiesel from microalgae. Biotech nol ogy Advances 25 , 294 - 306, doi:https://doi.org/10.1016/j.biotechadv.2007.02.001 (2007). 19 Chauton, M. S., Olsen, Y. & Vadstein, O. Biomass production from the microalga Phaeodactylum tricornutum: Nutrient stress and chemical composition in exponential f ed - batch cultures. Biomass and Bioenergy 58 , 87 - 94, doi:10.1016/j.biombioe.2013.10.004 (2013). 20 Ach yuthan, K. E., Harper, J. C., Manginell, R. P. & Moorman, M. W. Volatil e m etabolites e mission by i n v ivo m icroalgae An o verlooked o pportunity? Metabolites 7 , 39 (2017). 21 Seymour, J. R., Amin, S. A., Raina, J. B. & Stocker, R. Zooming in on the phycospher e: the ecological interface for phytoplankton - bacteria relationships. N ature microbiology 2 , 17065, doi:10.1038/nmicrobiol.2017.65 (2017). 22 Amin, S. A., Par ker, M. S. & Armbrust, E. V. Interactions between diatoms and bacteria. Microbiol Mol Biol Rev 76 , 667 - 684, doi:10.1128/MMBR.00007 - 12 (2012). 23 Le Chevanton, M. et al. Screening and selection of growth - promoting bacteria for Dunaliella cultures. Algal Res earch 2 , 212 - 222, doi:10.1016/j.algal.2013.05.003 (2013). 24 Kim, B. - H., Ramanan, R., Cho, D. - H., Oh, H. - M. & Kim, H. - S. Role of Rhizobium, a plant growth promoting bact erium, in enhancing algal biomass through mutualistic interaction. Biomass and Bioen erg y 69 , 95 - 105, doi:10.1016/j.biombioe.2014.07.015 (2014). 14 25 Wang, H., Hill, R. T., Zheng, T., Hu, X. & Wang, B. Effects of bacterial communities on biofuel - producing microalgae: stimulation, inhibition and harvesting. Critical Reviews in Biotechnology 3 6 , 341 - 352, doi:10.3109/07388551.2014.961402 (2016). 26 M ayali , X. & A zam , F. Algicidal b acteria in t he s ea and their i mpact on a lgal b looms1. Journal of Eukaryotic Microbiology 51 , 139 - 144, doi:10.1111/j.1550 - 7408.2004.tb00538.x (2004). 27 Paul, C., Mausz , M . A. & Pohnert, G. A co - culturing/metabolomics approach to i nvestigate chemically mediated interac tions of planktonic organisms reveals influence of bacteria on diatom metabolism. Metabolomics 9 , 349 - 359, doi:10.1007/s11306 - 012 - 0453 - 1 (2013). 28 Zerrifi , S . E. A., El Khalloufi, F., Oudra, B. & Vasconcelos, V. Seawe ed b ioactive c ompounds against p athoge ns and m icroalgae: Potential u ses on p harmacology and h armful a lgae b loom c ontrol. Mar Drugs 16 , 55, doi:10.3390/md16020055 (2018). 29 Yang, Y. H. et al. F att y a cid and l ipid c lass c omposition of the m icroalga p haeodac tylum tricornutum. Journal of oleo sci ence 66 , 363 - 368, doi:10.5650/jos.ess16205 (2017). 30 Gauthier, M. J. et al. Marinobacter hydrocarbonoclasticus gen. nov., sp. nov., a n ew, e xtremely h alot ole rant, h ydrocarbon - d egrading m arine b acterium. International Journal of Systematic and Evolutionary Microbiology 42 , 568 - 576, doi:10.1099/00207713 - 42 - 4 - 568 (1992). 15 C HAPTER 2: DEVELOPMENT OF AN UNTARGETED SPME - GC - MS WORKFLOW TO IDENTIFY VOLATILE BI OMA RKERS EMITTED FROM LIQUID CULTURES 2.1 INTRODUCTION The research described in this dissertation e xplores ways in which detection of biomarkers unique to the volatilomes of a micro organism can assist in biodefense and bioenergy applications. The objecti ves of the investigations described in subsequent chapters are divergent, but the types of data gener ated and the methods for separating, detecting, and identifying biomarkers from a complex matrix are similar. Therefore, this chapter provides a brief over vie w of the pertinent data acquisition and data processing strategies. 2.2 STRATEGIES FOR DATA ACQUI SITION OF VOLATILE METABOL ITES 2.2.1 Overview of common VOC samplers Several methods and/or devices have been developed for collection of VOCs from differ ent environments, each with advantages and disadvantages. Solid phase microextraction (SPME) has grow n to become one of the most popular approaches owing to its convenience and simplicity . SPME uses a sorbent material coated on the end of a fiber to passiv ely extract volatile analytes from the vapor phase. SPME fibers can be introduced into heated gas ch romatography (GC) injectors and the elevated t emperatures release trapped VOC analytes into the GC column. A competing approach, known as t hermal desorptio n ( TD) sampling , involves active sampling of a gaseous headspace by drawing vapors over a n ad sorbent mix packed inside a glass or stainless steel t ube. Adsorbed volatiles are removed by flowing carrier gas through the heated tube and into the GC column for fu rther analysis. A growing alternative to the use of selective, sorbent materials involves the use of broad - range sensor arrays in electronic nos es. Electronic noses have 16 been developed as mimics to the functionality of a human nose for rapid, real - time app roaches to VOC detection. This approach couples sensor arrays and pattern recognition software to create a characteristic profile of sensor resp onses for a given odor, and has been used to a great extent in food [1] and medical/pharmaceutical applications [2] . However, as the sensors have limited response selectivity, different VOCs can trigger very similar responses. Additionally, the individual VOCs comprising a novel pattern are not identified in the electronic nose d e tection process, precluding the detection and identification of novel biomarkers. Finally, when analytical instruments such as GC and GC - mass spectrometry (MS) systems cannot be located where headspace gas is generated (such as an environmental collection or collection from human/animal subjects), air samp les can be stored in plastic bags for transportation to a laboratory, sometimes made applicable for direct injection analyses in highly sensitive analytical detectors (see recent example with detection us i ng proton transfer reaction (PTR) mass spectrometry [3] ). A variety of factors affect which sampling technique should be chosen for a given experiment, such as selectivity, sample recovery rates, the potential for contaminants in the , storage conditions (e.g. , time and temperature), and intervals between collection and analysis in a lab (ranging from direct analysis to storage for extended periods of time). Recovery rates between different VOC devices were shown to vary: yields of >95 % recovery have been shown TD tubes in select conditions where compounds were well - suited to the given sorbent material [4,5] , while recoveries for SPMEs [6] and storage bags were lower [5] . Compared to SPMEs and TD tubes, bags also have a higher potential for cross - contamination during storage and the polymer material may introduce outgassing contaminants [7,8] . A higher sample capacity is achievable for TD tubes ( µ g to low mg) [9] compared to 17 levels) due to a larger surface area of adsorbent material [9,10] . Conversely, TD tubes have a disadvantage of req uiring the generation of a pressure gradient for active sampling (usually requiring active pumping) while SPME fibers involve passive sampling with minimal disruption of sample composition . This is important to consider for broader implications of VOC samp ling. While the research for this dissertation was conducted by sampling from controlled environment , f uture applications could be conducted in the field, thus requiring evaluation of the portabili ty and power requirements o f external devices to the sample rs. Ultimately, SPME samplers were chosen for the purposes of this work and will be further discussed. 2.2.2 S ample collection and concentration via s olid - phase m icroextraction The group of Janusz Pawliszyn developed SPME as a method to rapidly, cheaply, and easily achieve preconcentration of target analytes [11,12] , and this approach has been applied to applications in environmental work [13] , foods/fragrances [14] , medicine [15] , and forensics [16] . Thin polymer films are coated onto a support, typically a fused silica or stainless steel fiber that is a few hundreds of micrometer thick. Volatile or semi - volatile analytes can either absorb into the interior of the fiber coating, dependi ng on film thickness, or adsorb on the exterior of the polymer film. The fiber can be exposed to analytes through immersion in a liquid sample or through exposure to the headspace of a solid, liquid, or gaseous sample. SPME methods integrate a single step for sampling, extraction, and concentration prior to sample introduction into the heated injection port of a GC system. Conditions for VOC collection for all projects in this work involve a SPME fiber suspended in the headspace above actively outgassing s amples, similar to the illustration in Figure 2.1. In order for analytes to be 18 detected via SPME, two processes must occur. First, volatile analytes must partition between the liquid media and the gaseous headspace above the sample. Second, analytes must p artition between the gaseous phase above the sample and the SPME fiber coating. Once equilibrium has been established between the headspace and fiber coating, the amount of analyte deposited on the fiber is described by Equation 2. 1, which depends on the v olume of fiber coatings, volume of Figure 2.1 Schematic representing the partitioning of volatiles onto a SPME fiber (represented by black rectangle) suspended above a liquid media, translatable to the experiments to be covered in Chapters 3, 4, and 5. T he volati lity of compounds is a spectrum, and some compounds with low volatility may still be detected depending on experimental conditions (e.g. , sampling time, temperature) . (1) Equation 2. 1: n is the amount of analyte extracted into the SPME fiber, K fs is the distribution constant of analyte partitioning between the fiber coating and sample (independent of the location of the fiber in the system), V f is the fiber coat ing volume, V s is the liqui d sample volume, C o is the initial analyte concentration in the sample , K hs is the distribution constant of analyte partitioning between headspace and fiber coating, and V h is the volume of the gas in headspace. 19 sample matrix (i .e. gaseous headspace, liquid media ), the initial concentration of analyte in the sample, and equilibrium constants for each analyte between the sample/fiber or the headspace/fiber. The amount of analyte adsorbed on the fiber is proportiona l to its concent ration in the headspace, which in turn is influenced by its concentration in the liquid. Several different stationary phase coatings are commercially available, including polydimethylsiloxane (PDMS), divinylbenzene (DVB), carboxen (CAR), c arbowax - polyethy lene glycol (PEG), polyacrylate (PA), and C18 silica particles. Table 2.1 gives recommendations of which fiber types are advantageous to detect different analytes. Additionally , a mixture of sorbent materials can be coated on to the same fi ber, thus broade ning the capture performance to more substances. - Aldrich/Supelco), where the SPME coating is located at the end of an adjustable length needle and can be retracted Table 2.1 Recommended S PME polymer coatings for different analyte types, recommended by the commercial supplier Supelco/Sigma - Aldrich (See SPME Applications Guide 925F) Fiber Types Analytes Volatiles (MW < 275) 30 µm PDMS Non - polar semi - volatiles (MW 80 - 500) Non - polar high molecular weight compounds (MW 125 - 600) Gases and low molecular weight compounds (MW > 275) 65 µm PDMS/DVB Volatiles, amine s and nitro - aromatic compounds (MW 50 - 300) 85 µm PA Polar semi - volatiles (MW 80 - 300) 60 µm PEG Alcohols and polar compounds (MW 40 - 275) 50 µm / 30 µm DVB/CAR on PDMS Trace compound analysis, C3 - C20 (MW 40 - 275) 20 into an external casing. Only three polym er coatings were commercially available for the portable field samplers, limiting the extent to which the analysis could be tailored. The fiber PDMS/DVB coating, typic ally used for the broad analysis of VOCs and amines (nitro - aromatic compounds are rare in nature ). The portable samplers are suitable for use in a field - setting, are lightweight and easily transported back to a laboratory, and provide minimal sample loss for up to 1 month when properly stored (4°C - 20°C). 2.2.3 Separation and d etection of VOCs GC - MS is a n easily adaptable and commonly used method ology for the identification and measurement of VOCs [17] . Figure 2.2 shows the most basic components of a GC - MS system. Volatiles introduced in the GC heated inlet are carried via a mobile gaseous phase (e.g., helium gas) through interactions with a stationary phase based upon their volatility and affinity for each phase. Compounds eluting from the end of a chromatography column exit the column at different times after i are detected in th e mass spectrometer. Detection of analytes at each retention time, where 1) ionization of compounds occurs via bombarding compounds with a 70 eV beam of electrons, 2) energy imparted during ionization c auses fragmentation , with masses of fragment ions ofte n characteristic of functional groups, 3) molecular and fragment ions are separated by their mass - to - charge ( m/z ) ratio s in the mass analyzer. Extensive libraries of electron ionization ( EI ) mass spectra are available to facilitate identification of analyt es separated and detected in complex samples by GC - MS, a common example being the National Institute for Standards and T echnology (NIST) database that is updated every few years . 21 Figure 2.2 Schematic of a GC - MS system, where a sample collected on a SPME fiber is injected on to a capillary column. Flow of helium carrier gas drives transport through the column, where analytes separate depending on interactions between the stationary phase and mobile phase (accelerated by increasing oven temperatures). Compou nds exiting the column are ionized, fragmented, and separated in the mass analyzer (qu adrupole, nominal m/z reported). Detected ions are summed by repetitive scanning through the mass spectrum , creating a total ion current chromatogram. Abundances of ions at each m/z value are stored, and mass spectra can be constructed from ion abundances for each m/z value. 2.3 APPROACHES FOR DATA PROCESSING OF VOLATILE METABOLITES 2.3.1 Untargeted metabolomics vs. targeted metabolite profilin g Metabolomics is the quali tative and/ or quantitative study of small molecules, called metabolites, produced by a biological system (organism, tissue, or cells). Changes in the metabolites result from altered biochemical activities and vary with the growth state of a biological syst em, thus re presenting the metabolic phenotype. Strategies for defining metabolic phenotypes are divided into two approaches, untargeted (or discovery) metabolomics and targeted metabolite profiling. Untargeted metabolomics aims to generate a broad, compreh ensive anal ysis of both identified and unidentified analytes in samples without prior knowledge of metabolite identities. This approach attempts to detect as many compounds as possible (within the confines of the chosen sampling technique and detection sys tem), prese nting an avenue for novel biomarker discovery. Metabolomics commonly employ s data reduction techniques to narrow an extensive list to a more manageable, smaller set of signals, where the smaller set 22 reflects metabolites relevant to changes in th e biologica l system. These signals then are subjected to metabolite annotation using pre - existing in silico libraries or by experimental analysis of analytical standards on the same instrument. Targeted metabolite profiling involves measurements of pre - de termined gr oups of chemically annotated metabolites. When target compounds are known, the use of internal standards, whether similar in structure or isotopically labeled analogues, can be used to rigorously quantify changes of target compound level s across different sample types and has a higher sensitivity compared to untargeted approaches. Additionally, determining relationships in biochemical relationships between targeted metabolites, especially when changing between different physiological states, is e nhanced due to having prior knowledge of their metabolic pathways. In the hope of guiding pathway discoveries, the initial stages of development of research projects for biosecurity and bioenergy utilized an untargeted metabolomics approach for identificat ion of biomarkers within the datasets generated using SPME - GC - MS. 2.3.2 Development of data processing criteria to identify biomarker signatures A multi - step data processing workflow facilitated conversion of GC - MS datasets into a list of vola tile biomar kers for different matrices in this research. First, the information from Chromatographic deconvolution is the process of separating compounds in a chro matogram ba sed on differences in elution time of various detected ions and can be used to separate compounds that elute at similar elution times leading to chromatographic peak overlap in the total chromatogram . All mass spectral ion s that share the same c hromatograp hic peak shape and reach 23 compound. Deconvoluted compounds were reported by Agilent MassHunter Qualitative software for each sample. Retention time alignment of compoun ds across samples was accomplished using Mass Profiler Professional for each experiment , which corrects for minor retention time drift . Project - specific parameters utilized for peak detection are further described in each chapter. Detected VOCs were then annotated based upon comparisons to the NIST 2014 (NIST14) mass spectral database. subsequently identified by the name of the match with the highest score. The confidence in NIST14 annotation was f urthered by calculation of chromatographic Retention Index (RI) for each analyte on a non - isothermal, temperature - programmed gradient, where the RI of each analyte is calculated based upon its retention compared to a series of n - alkane standards analyzed u nder the sa me temperature program. Calculated experimental RIs differing more than 5% total deviation from the theoretical RI value in the reference library resulted in rejection of a NIST ID. Compounds that did not exceed the mass spectral match or retent ion index t hreshold were annotated using the base peak m/z and retention index (e.g. , m/z 177 at RI 1495). Identification of VOCs as candidate compounds for potential microbial biomarkers was achieved through development of two filtering criteria, with the purpose to eliminate background VOCs contributed from the experimental setup, growth media, and other comp ounds that exhibited inadequate consistency of detection. The first criterion required detection of a potential biomarker in at least two replicate samples at a given timepoint. This criterion was developed to assess levels of reproducibility while also al lowing for biological variability that leads to occasional non - detects that is often encountered in experiments involving live biological systems. Some of the d etected compounds were present at low concentrations and some biological variability could have easily pushed a compound below the detection threshold in one 24 of the replicates. This criterion avoids the analysis missing some potentially interesting markers by applying a too stringent criterion . For different experimental conditions, the number of sample replicates occasionally varied depending on the experimental setup; i.e. , an experiment with n=3 replicates would require detection in 66% of samples (2 out of 3 replicates) while an experiment with n=2 replicates required detection in 100% of sampl es (2 out of 2). Details on the sample types and sample replicates is included more within Chapters 3 - 5. The second filtering criterion concerned the presence or a bsence of a marker in a biological culture relative to the media blank controls appropriate f or each organism. A compound was removed from consideration as a potential candidate if its relative abundance in the biological replicates was less than ten times the average relative abundance in the contro l, where the threshold was set high due to obser ved changes in chromatographic peak areas across experimental timepoints regularly changing 1 - 2 orders of magnitude. 2. 4 Example of VOCs contributed from commerci al sampling flasks. The complexity of untargeted data analysis can be complicated when both pertinent and non - pertinent VOCs are present. One contributor of non - pertinent VOCs seen in this work that highlights that complexity includes outgassing from the experimental setup. Of note, Erlenmeyer flasks comprised of either glass or polymers were uti lized in different projects. Polymer - based Erlenmeyer flasks are advantageous for growing pathogens due to their lighter weight and resistance to breakage. Their u se minimizes potential safety hazards including exposure of researchers to pathogenic culture s. However, polymer - based materials contain higher levels of outgassing compounds that can be detected by VOC samplers, thus increasing endogenous background level s of VOCs. 25 Different polymer vessels were investigated for the purpose of growing pathogen cultures, comprised of polystyrene (PS), polyethylene terephthalate glycol - modified (PEG), or polycarbonate (PC). Figure 2.3 documents the VOC complexity observed for each flask when incubated at 37°C with 75 - 100 mL of liquid media present. Total ion curre nts were least for PS, and greatest for PEG, with PC in between. In particular, PEG released three VOCs of significantly higher abundance than anything observed in PS and PC. These intense signals matched the mass spectra of decamethyl - cyclopentas i loxane, dodecamethyl - cyclohexasiloxane, and tetradecamethyl - cycloheptasiloxane at 95%, 98% and 89% accuracy, respectively and all experimental RI values fell within 5% of theoretical values. The siloxane compounds are thought to originate during manufacturing, as siloxanes are used in proprietary additive mixtures to keep polymers malleable for manufacturing purposes as well as machinery lubricants [18] . As these vessels were adapted to VOC sampling and not the intended purp ose of these vessels, future studies should evaluate release of VOCs from flasks at the beginning of experiments. The PC flasks were utilized for the work in Chapter 3 due to product availability. 2. 5 SUMMARY Rationales for choosing SPME - GC - MS as the com bined VOC collection, data acquisition, and data processing methodology were outlined in this chapter. Minor modifications in the procedures, if any, will be addressed in each methods section of subsequent chapters. The developed workflow can be broadly em ployed to general VOC metabolomics datasets to enable detection and annotation of putative biomarkers and is not limited to the applications described in this work. 26 Figure 2.3 Examples of the endogenous VOCs exhibited by polymer - comprised sample flasks Polystyrene (PS), Polycarbonate (PC), and Polyethylene terephthalate glycol - modified (PEG) through comparison of the VOC total ion current chromatogram s. Special annotation of peaks in PEG are given for 1) decamethyl - cyclopentas i loxane, 2) dodecamethyl - cy clohexasiloxane, and 3) tetradecamethyl - cycloheptasiloxan e, high - outgassing VOCs contaminating the background VOC profiles when no cultures are present. 27 REFERENCES 28 REFERENCES 1 Sanaeifar, A., ZakiDizaji, H., J a fari, A. & Guardia, M. d. l. Early detection of contamination and defect in foodstuffs by electronic nose: A review. TrAC Trends in Analytical Chemistry 97 , 257 - 271, doi: 10.1016/j.trac.2017.09.014 (2017). 2 Wasilewski, T., W. Critical review of electronic nose and tongue instruments prospects in pharmaceutical analysis. Analytica Chimica Acta 1077 , 14 - 29, doi: 10.1016/j.aca.2019.05.024 (2019). 3 Papurello, D. Direct injection mass spectrometry technique for the odorant losse s at International Journal of Mass Spectrometry 436 , 137 - 146, doi: 10.1016/j.ijms.2018.12.008 (2019). 4 Kim, Y. - H. & Kim, K. - H. Extent of s ample l oss on the s ampling d evice and the r esulting e xperimental b iases w he n c ollecting v olatile f atty a cids (VFAs) in a ir u sing s orbent t ubes. Analytical Chemistry 85 , 7818 - 7825, doi:10.1021/ac401385m (2013). 5 Koziel, J. A. et al. Evaluation of s ample r ecovery of m alodorous l ivesto ck g ases from a ir s ampling b ags, s olid - p hase m i c roextraction f ibers, t enax TA s orbent t ubes, and s ampling c anisters. Journal of the Air & Waste Management Association 55 , 1147 - 1157, doi:10.1080/10473289.2005.10464711 (2005). 6 Merkle, S., Kleeberg, K. K. & Fritsche, J. Recent d evelopments and a pplicati o ns of s olid p hase m icroextraction (SPME) in f ood and e nvironmental a nalysis A Review. Chromatography 2 , 293 - 381 (2015). 7 Harreveld, A. P. v. Odor c oncentration d ecay and s tabilit y in g as s ampling b ags. Journal of the Air & Waste Management Association 53 , 51 - 60, doi:10.1080/10473289.2003.10466121 (2003). 8 Gilchrist, F. J. et al. An investigation of suitable bag materials for the collection and storage of breath samples containing hydrogen cyanide. Journal of Breath Research 6 , 036004, doi:10.1088/1752 - 71 5 5/6/3/036004 (2012). 9 Moein, M. M., Said, R., Bassyouni, F. & Abdel - Rehim, M. Solid p hase m icroextraction and r elated t echniques for d rugs in b iological s amples. Journal of Analy tical Methods in Chemistry 2014 , 921350, doi:10.1155/2014/921350 (2014). 10 P awliszyn, J. Theory of Solid - Phase Microextraction. Journal of Chromatographic Science 38 , 270 - 278, doi:10.1093/chromsci/38.7.270 (2000). 11 Belardi, R. P. & Pawliszyn, J. B. The a pplication of c hemically m odified f used s ilica f ibers in the e xtraction of o rganics from w ater m atrix s amples and their r apid t ransfer to c apillary c olumns. Water Quality Research Journal 24 , 179 - 191, doi:10.2166/wqrj .1989.010 (1989). 29 12 Arthur, C. L. & Pawliszyn, J. Solid phase microextraction with thermal desorption using fused silica optical fibers. Analytical Chemistry 62 , 2145 - 2148, doi:10.1021/ac00218a019 (1990). 13 Ouyang, G. & Pawliszyn, J. SPME in environmenta l analysis. Analytical and bioanalytical chemistry 386 , 1059 - 1073, doi:10.1007/s00216 - 006 - 0460 - z (2006). 14 Xu, C. - H. et al. Applications of solid - phase microextraction in food analysis. TrAC Trends in Analytical Chemistry 80 , 12 - 29, doi: 10.1016/j.trac.2016.02.022 (2016). 15 Looby, N. T. et al. Solid phase microextraction coupled to mass spectrometry via a microfluidi c open interface for rapid therapeutic drug monitoring. Analyst 144 , 3721 - 3728, doi:10.1039/C9AN00041K (2019). 16 Gherghel, S. et al. Development of a HS - SPME/GC MS method for the analysis of volatile organic compounds from fabrics for forensic reconstruct i on applications. Forensic Science International 290 , 207 - 218, doi: 10.1016/j.forsciint.2018.07.015 (2018). 17 compounds in effluent samples by gas chromatography A revie w . Journal of Chromatography A 1592 , 143 - 160, doi: 10.1016/j.chroma. 2019.01.045 (2019). 18 Landon, S. J. et al. Room temperature - cured siloxane sealant compositions of reduced gas permeability. USA patent US7674857B2 (2005). 30 C HAPTER 3: METABOLIC PROFILING OF VOCs EMITTED BY THE PATHOGENS Francisella tularensis AND Bacillus anthracis IN LIQUID CULTUR E FOREWORD The material presented in this chapter has been adapted from work first published in 20 20 in the journal Scientific Reports [1] . Contributions from others to conduct the experimen ts described in this chapter are as follows : A. Rasley and J. R. Avila were integral in providing stock cu ltures of RG2 pathogens, experimental development, and culturing RG3 pathogens and conducting volat ile organic compound sampling in Biosafety Level 3 (BSL3) facilities at LLNL. 3.1 INTRODUCTION The study and detection of volatile organic compounds (VOC s) originating from or interacting with organisms ranging from bacteria to humans have numerous app lications in biology, environmental sciences, medic ine, food industry, and national security. The measurement of VOCs and non - volatiles in exhaled breath is becoming an important rapid and non - invasive diagnostic tool to assess human physiology and health as well as a diagnostic tool for infections and sys temic disease [2 - 5] . VOC markers in exhaled breath are also bein g explored to assess human chemical pharmacokinetics and environmental exposures to drugs, toxic material s, chemical or biological agents and other il licit materials. In this context, in vitro systems are often used to explore human exposome, microbiome an d disease pathogenesis biomarkers [6,7] . One under - explored area of human volatile analysis is the analysis of exhaled - relevant bact erial pathogens. However, human breath VOC profiles may be specific to an individua l, 31 es, and immunological responses. B acteria emit VOCs as major metabolic products during their growth cycles, many of w hich have important functions as signaling molecules to either neig hboring bacteria or higher organisms [8] . Bacterial VOC profiles comprise complex mixtures containing diverse structural and chemical complexity. Monitoring of volatile emissions of bacteria has been facilitated by use of SPME - GC - MS , and extensive literature on bacteria related to food safety and hospital - acquired infe ctions demonstrates how volatile chemical signatures can differentiate bacteria from various genera, species, and subspecies. Prior work into sampli ng volatiles of active bacterial growth has largely focused upon pathogens related to clinical settings, wit h common examples originating from the genera Pseudomonas [9] , Staphylococcus [9] , Klebsiella [10] , and Mycobacterium [11] . For example, Rees, et al. [10] reported aliphatic 2 - keto nes as the most abundant VOCs produced by Klebsiella pneumoniae, a common Gram - negative human pathogen, wi th less - abundant compound classes including esters, benzene derivatives, heterocycles, and nitrogen - containing compounds. Chen et al. [12] focused on the time - dependent VOC emissions of several foodborne pathogens, identifying long chain methyl ketones (2 - heptanone, 2 - nonanone, 2 - undecanone) and alcohols (1 - octanol, 1 - decanol, 1 - 32 dodecanol) as markers of three Gr am - negative species ( E. col i, S. flexneri, and S. enteritidis ), while 3 - hydroxy - 2 - butanone was identified as a marker of two Gram - positive bacteria ( S. aureus and L. monocytogenes ). At present, there is less knowledge about volatiles released from bacteria l threat agents, which woul d be useful for detecting the presence of and distinguishing such threat agents . Horsmon and Crouse [13] used thermal desorption tubes coupled to gas chromatography - mass spectrometry (GC - MS) to describe VOC profiles emitted by cultures of Yersinia pestis (Y. pestis) , the causative agent of plague, and severa l strains from the genus Bacillus . They showed that VOC profiles and relative abundances of individual compounds distinguished bacterial genera as wel l as species within the same genus. However, their study did not provide a comprehensive analysis of detec ted VOCs and only qualitative differences determine d by inspection of the profiles were used to distinguish species or genera. Lonsdale et al. [14] used colorimetric sensor arra ys (CSAs) to differentiate volatiles in the headspace of Y. pestis and B. anthracis cultures , and r eported high specificity, accuracy, and sensitivity to very low bacterial concentrations. However, individ ual biomarkers leading to the colorimetric changes were not identified, and signal response could have been influenced by the growth media utilized. Two bacteria of concern to biosecurity and subjects of the work presented here are the aerobic, facultativ e intracellular pathogen Francisella tularensis (F. tularensis) and the obligate, endospore - forming pathogen Bacillus anthracis ( B. anthracis ). Both a re classified by the Center for Disease Control (CDC) as Tier 1 Select Agents on the CDC category A Bioter rorism Agents list [15] . F. tularensis , the causative agent of the disease tular emia, has been isolated from more than 200 separate organisms, and several subspecies are known human pathogens. The bacterium is highly infe ctious an d easily aerosolized, requiring as few as ten bacteria to cause infections 33 [ 16] . B. anthracis , the causative agent of the disease anthrax, forms resilient spores that survive chemic al treatments , heat, lack of nutrient s , and radiation, and has previously been developed into a bioweapon [17 ] . Det ection of volatile biomarkers specific to the presence and growth of F. tularensis or B. anthracis through headspace sampling would be an important step towards developing a non - invasive metabolomics tool for rapid diagnosis of their presence in t he lungs of subjects exposed to a biological attack. Prior studies aimed at identification and/or differentiation of metabolites from F. tularensis or B. anthracis have largely focused on measuring profiles of pre - selected molecular targets in whole cell extracts. In particular, fatty acids have been profiled using GC - MS following esterification. The Voorhees group di stinguished strains of F. tularensis, B. anthracis, Brucella spp. abortus, melitensis, and neotomae, and Yersinia pestis through analysis of fatty aci d methyl ester (FAME) profiles using pyrolysis mass spectrometry in combination with an in - situ thermal tr ansesterification [18,19] . Fatty acids of carbon chains ranging from 12:0 to 24:1 were identified, and princ ipal components analysis (PCA) of the fatty acid profile s discriminated bacterial species. Li et al. disti nguished Francisella tularensis subspecies novicida , Escherichia coli, and Bacillus subtilis by derivatizing fatty acids to form trimethylsilyl esters [20] . However, these studies required whole bacteria and sample preparation that was destructive to the bacteria, precluding analysis of metabolite changes over time in an unperturbed cult ure. For descriptions of risk group (RG) classifications of 34 infectious microorganisms and recommended biosa fety level (BSL) for their handling see the U.S . Department of Health and Human Services guide on Biosafet y in Microbiological and Biomedical Laboratories [21] or the World Health Organization Laboratory Biosafety Manual [22] .) solid phase microextraction in vitro determinatio n of VOC profiles lays the groundwork for non - invasive investigation of bacterial metabolism of such organisms and represent the first steps to wards p otential VOC - based detection of an infection by such agents. 3.2 METHODS 3.2.1 Strains and Growth Media F. tularensis subspecies novicida (strain U112; RG2) and subspecies tularensis (strain SCHU S4; RG3) were obtained from the CDC and Brigham Young Uni versity, respectively. B. anthracis (strain Sterne; RG2) and Ames (strain Ames; RG3) were obtained from a collaborator at Dugway Proving Grounds. The agar plates and liquid growth media for bacterial growth were prepared separately for each species. Differ ent media were chosen to achieve optimal bacterial growth. F. tularensis was grown using a modified Muelle r - Hinton (MH) growth media [23] (Becton Dickinson (BD) Difco , Franklin Lakes, NJ) supplemented with 0.1% glucose, 0.025% ferric pyrophospha te (Sigma - Aldrich, St. Louis, MO), and 0.02% IsoVitaleX (BD Difco); B. anthracis was grown with Brain - Heart Infusion (BHI) growth media (Becton Dickinson (BD) Difco, Franklin La kes, NJ). RG2 strains and RG3 strains were grown, prepared and sampled in a bio safety level 2 (BSL - 2) laboratory and biosafety level 3 (BSL - 3) laborat ory, respectively. 35 3.2.2 Preparation of Bacterial Headspace Bacterial colonies were selected after overnig ht incubation on agar plates and transferred to 10 mL of liquid modified MH med ia or BHI media, respectively. Bacteria were cultured in media under ae robic conditions with overnight incubation at 37 °C and 170 rpm shaking. For each species and experiment, three 100 - µL aliquots were inoculated into three separate 20 mL portions of fre sh liquid media (1:200 dilutions) and incubated in three 250 - mL disposa ble polycarbonate Erlenmeyer flasks with vented caps at 37 °C and 170 rpm shaking. The VOC profiles from t he headspaces of each of the triplicate bacterial cultures (replicates) and the number of viable bacteria were sampled and assessed at multiple timepo ints. In addition to the three replicates of each pathogen species, an uninoculated liquid media flask wa s simultaneously prepared and VOCs sampled from it as a negative (media - only) c ontrol. 3.2.3 Sampling VOCs from Bacterial Headspace (RG2 strains) Th e VOC profiles of bacterial headspaces and media controls were sampled at different time intervals dependi ng on bacterial growth rates and experimental setups using a protocol developed here that some of the authors also applied for headspace analysis of a lgal cultures in other work. [ 24] Ft novicida cultures were sampled at the following timepoints: 0, 2, 4, 8, 12, 16, 20, 24, 28, 32, 4 8, and 52 hours post - inoculation. Ba Sterne cultures were sampled at the following timepoints: 0, 4, 8, 12, 20, and 24 hours post - inoculation. At the time of sample collection, Erlenmeyer flasks were removed from the incubator - shaker and transferred to a b iosafety cabinet. Headspace VOCs wer e immediately collected for 30 minutes on a field - portable 2 cm solid - phase microextraction (SPME) fiber with a 65 µm polydimethylsiloxane/divinylbenzene (PDMS/DVB) coating (Supelco, Bellefonte, PA) with no 36 agitation of the flask. At each timepoint, one un exposed SPME fiber (fiber remaining retracted behind the septum in the SPME housing) was placed within the biosafe ty cabinet where the account for potential background vo latiles leaking onto retracted fibers over time during storage or transportation to the GC - MS analysis laboratory. concurrently with fibers exposed to cultures. After collection, all S PME fibers were stored in refrigerat ors at 2 - 4 °C until analysis. Data acquisition on the GC - MS occurred within 3 weeks of collection. 3.2.4 Sampling VOCs from Bacterial Headspace (RG3 strains) and Transfer of SPME Samples to BSL - 2 Facility The VOC profi les of RG3 Ft SCHUS4 and Ba Ames as well as corresponding media controls were sampled at the following timepoints: 0, 6, and 24 hours post - inoculation . Timepoints were chosen to capture the exponential and stationary growth phase in each species. At the ti me of sample collection, Erlenmeyer flasks were removed from the incubator - shaker and transferred to a biosafety cabinet within the BSL - 3 facility. Fl asks were allowed to sit in the BSC for 30 minutes prior to sampling in order to allow any aerosols to set tle. Headspace VOCs were collected f or 30 minutes on SPME fibers with no agitation of the flask. After collection, SPME fiber devices were decontamina ted by bleach wiping the entirety of their external housing for 1 min apiece, and residual bleach was remo ved via wiping. The process of bleac h wiping to prevent accidental transfer of pathogens out of the BSL - 3 facilities was tested and validated. The ove rall protocol was approved by the Institutional Biosafety Committee (IBC) at LLNL (see Protocol in Appendi x ). SPME fibers were transferred fro m the BSL - 3 to BSL - 2 facilities and 37 stored in refrigerators at 2 - 4 °C until analysis, as previously described. In analyzing the samples collected in the BSL - 3 we did not find any indication that bleach wiping may have al tered the compounds detected e.g. , b y introducing chlorinated compounds. 3.2.5 Determination of Bacterial Concentrations The growth phase (logarith mic, stationary, decline) of each organism was estimated by monitoring the concentration of viable bacteri a over the course of the experiment for all biological replicates. Aliquots (1 mL) of all bacterial cultures were collected immediately following VOC sampling at each of the timepoints, and the Erlenmeyer flasks were subsequently placed back into the incub ator - shaker. The aliquots were serially diluted between 10 - 2 to 10 - 7 depending on expected growth phase. A preliminary experiment was performed by pla ting in duplicate 10 - fold dilutions to determine the appropriate serial dilution for each growth phase. Th e dilution factor was sel ected to achieve a target concentration of 30 - 300 cells per plate for counting. Dilutions were plated in duplicate (100 - iquots) on agar plates to determine the number of colony - forming units (CFU). Bacterial concentrations are reported as CFU counts p er mL of liquid culture. 3.2.6 Data Acquisition Parameters The data acquisition followed a procedure similar to the one pre viously described for algal VOCs and is briefly summarized here. [24] VOC analyses were perfo rmed on an Agilent 5975T GC - MSD (Agilent T echnologies, Santa Clara, CA) using an Agilent HP - 5ms column (30 m x 250 µm x 0.25 µm) coupled to a single quadrupole mass analyzer with helium carrie r gas at 38 a constant flow rate of 1.2 mL/min. Volatiles absorbed by the SPME fiber were desorbed in the hea - seconds using splitless injection. The column temperature was programmed to start at 40 °C for 6 min, then heated at 8 °C/ min from 40 to 280 °C and held for 4 min (total run time = 40 mi n). Ions were generated using electron ion ization (EI) (70 eV) and acquired at 4 scans/s over m/z 35 - 450. Data acquisition was performed using ChemStation (version E.02.02). A commercial GC - MS reference standard (S - 22329; AccuStandard, New Haven, CT) was u sed to evaluate day - to - day performance of the GC - MS system and to calculate retention indices. 3.2. 7 Data Processing Af ter data acquisition, data processing procedures and criteria were appl ied to detect and identify taxa - specific biomarkers similar to t he work previously described for algal VOC s [24] . All ChemStation data files (consisting of data from biological replicates, media controls, and travel fibers) were translated using MassHunter GC/MS Translator B.07.05 fo r compatibility (version B.07.00 SP2) and Mass Profiler Pr ofessional (MPP) 12.6.1 software. These programs enabled sophisticated organization of individual MS files into complex datasets for chemometric analy ses. Chromatographic deconvolution and visualization were perfo rmed using MassHunter Qualitative using a Retention Time window size factor of 90.0, signal - to - noise ratio threshold of 2.00, and absolute ion height filter of 1000 counts. An arbitrary small value of 1 was assigned across all samples to the signal value f or compounds that were not detected. Detec ted peaks were transferred into MPP and inter - aligned using a retention time tolerance of 0.15 minutes, mass spectral match factor of 0.6 (of maximum 1.0), and a delta m/z tolerance of 0.2 Da. Annotation of 39 the ali gned compounds was performed by searching spectra against the NIST14 mass spectral name of the match wi th the highest score. Identifications with literature retention indices deviating more than 5% from the ex perimental retention indices were rejected. Compounds that did not exceed the mass spectral match or retention index threshold were annotated using th e base peak m/z and retention index (e.g. , m/z 121_RI 1 The reported abundance values in th is work are relative abundances of compounds, obtained by integrating the signal in their chromatographic peaks. Relative abundances are compared betw een different measurements (timepoints, species). Absolute quantification of VOCs in the headspace above b acterial cultures is challenging with our method. For example, our culture vessels were not fully enclosed due to the use of vented caps designed to f acilitate gas exchange and avoid pressure buildups, and some loss of VOCs may have occurred. The retention of analytes is also affected by sorbent material, sampling time, and potential saturation, whereas the desorption is affected by extraction time and temperature. Some relative quantitation could be achieved using internal standards, whether pre - loaded or spiked into cultures , but also has a number of practical issues. Therefore, for the purposes of our work, absolute quantification was not attempted. Two filtering criteria were used to identify relatively robust and reproducible VOCs as the most likely ca ndidate compounds fo r potential taxa - specific biomarkers. The first criterion required detection of a potential biomarker in at least two of three cul ture replicates at a given was chosen as a compromise to req uire some level of reproducibility while also allowing for some biological variability that is often encountered in experiments in volving live biological systems. Some of the detected 40 compounds were present at fairly low concentrations and some biological variability could have easily pushed a compound below the detection threshold in one of the replicates. The was chosen in order to avoid missing some potentially interesting markers by applying too stringent a cri terion. The second c riterion concerned the presence or absence of a marker in a biological culture relative to the media blank controls appropriate fo r each organism. A compound was removed from consideration as a potential candidate if its relative abunda nce in the biologica l replicates was less than ten times the relative abundance in the control. The VOCs identified as putative taxa - specific biomarke rs were compared with regard to both individual markers and groups of markers encompassing a compound clas s. First, the presen ce or absence of these markers in each growth phase (logarithmic, stationary, and decline) was determined. Second, the calculated peak areas of markers, also referred to as relative abundances here, were compared amongst biological repl icates to assess con sistency of detection. Finally, principal component analysis (PCA) was used as a dimension - reduction strategy to visualize covaria nce in the dataset. Only markers remaining after the filtering criteria were applied were utilized. Using the MPP software, pr ior to PCA analysis, markers were individually mean - centered and variance - scaled. PCA was performed on the transformed dataset, an d the results are presented as a scores plot of the first two principal components (PCs) and a loadings pl ot to elucidate the contribution of each marker to PC positioning. PCA was not performed on the RG3 taxa due to the limited number of acquired samples . 3.3 RESULTS The complexity of chromatographic peaks detected through GC - MS analysis of each sample is illustrated by an ob served profile of sampled 24 hours post - inoculation, 41 representing the early stationary phase (Figure 3. 1a). However, m any peaks originated from the Mueller - Hinton media and the SPME sampling device (Figure 3. 1b) and were con sidered background. Peaks representative of the bacterial signature were of lower relative abundance, highlighted on a smaller chromatographic scale i n Figure 3. 1c. This complexity was similar for Ft SCHUS4 and both B. anthracis taxa (not pictu red). Detec tion of thousands of volatile compounds in the various cultures and timepoints for each of the taxa (Table 3. 1) required specified data - filtering crit eria (see Methods) to remove background compounds and those compounds not reproducibly detecte d . For examp le, more than 2000 VOCs were detected across all samples. Eliminating VOCs that did Figure 3. 1 Examples of the chemical complexity exhi bited by F. tularensis novicida cultures through comparison of the VOC total ion chromatograms at (a) 24 h ours post inoculation, (b) corresponding Mueller - Hinton media control, and (c) enlarged overlay of both 1a and 1b, where stars indicate the bacteria - s pecific VOC emissions. 42 Table 3. 1 Number of VOCs from F. tularensis and B. anthracis taxa detected by GC - M S and remaining after application of filtering criteria Species Total VOCs Detected VOCs Fail Criterion 1 VOCs Pass Criterion 1 VOCs Fail Criterion 2 VOCs Pass Criterion 2 (putative biomarkers) F. tularensis novicida 2360 2239 121 103 18 F. tularensis SC HU S4 999 754 245 207 38 B. anthracis Sterne 1031 912 119 89 30 B. anthracis Ames 1022 745 277 221 56 not appear in at least two of the triplicate measurements (Criterion 1) narrowed the dataset to 121 VOCs, a reduction of approximately 95 % (Figure 3. 2). Further elimination of VOCs with relative abundances less than 10x the average relative abundance in t he negative controls (Criterion 2) narrowed the dataset to 18 putative volatile biomarkers that were confidently attributed to . The same c riteria were applied to the data from other bacterial species studied here, resulting in 38 putative VOC b iomarkers for Ft SCHUS4, 30 biomarkers in Ba Sterne, and 56 biomarkers in Ba Ames (Table 3. 1). 3. 3 .1 Results from RG2 species Candidate bacterial VOC biomarkers from all timepoints were annotated through examination of both mass spectral library matching scores using the NIST14 database and experim ental retention indices. Since all metabolite annotations in this report are based on comparisons to liter ature spectra and retention index values, they should be considered as satisfying confidence leve l 2 of the Metabolomics Standards Initiative recommen dations for 43 Figure 3. 2 Example workflow of criteria utilized to filter list of detected VOCs to bacteri a - specific biomarkers produced during growth, shown for F. tularensis novicida . identification of compounds [25] . For (Table 3. 2), 15 of the 18 biomarkers passed the set threshold of 70% match, while t matches below that threshold. For Ba Sterne (Table 3. 3), 18 of the 30 biomarkers passed the set thresh - species diversity in emitted VOC bio markers was observed. The profile contains odd - chain, aliphatic methyl ketones, alcohols , nitrogen - containing, and sulfur - containing vola tiles. The Ba Sterne volatile profile is comprised of branched methyl ketones, followed by esters, carboxy lic acids, alcohols, and sulfur - containing volatiles. Evaluation of potential markers requires asses sment of the growth phase at each sampled timepoi nt post - inoculation of the culture flask. The logarithmic, stationary, and decline ph ases were identified based upon CFU measurements taken alongside SPME - VOC sampling. The data 44 Table 3. 2 Annotations of F. tularensis novicida - specific VOC markers through compound class, putative NIST ID, m/z , and retention index matching Class Compound MS Base Peak ( m/z ) RI (Lit) RI (Exp) NIST 14 Match Factor Alcohol 1 - Butanol, 2 - methyl - 70 739 719 74 Alcohol 2 - Nonanol 45 1101 1113 80 Alcohol Phenylethyl Alcohol 91 1 116 1127 70 Alcohol 1 - Nonanol 56 1173 1183 83 Alcohol 2 - Undecanol 45 1308 1306 75 Methyl ketone 2 - Hepta none 43 891 890 71 Methyl ketone 2 - Nonanone 58 1092 1104 72 Methyl ketone 2 - Undecanone 58 1294 1298 73 Methyl ketone 2 - Tridecanone 58 1497 1492 93 Methyl ketone 2 - Pentadecanone 58 1698 1690 87 Methyl ketone 58.0@26.267332 58 1902 1892 83 Nitrogen - con taining Pyrazine, 2,5 - dimethyl - 108 917 911 80 Nitrogen - containing 2 - Methyl - 3 - isopropylpyrazine 121 1056 1064 83 Sulfur containing Dimethyl trisulfi de 126 970 974 96 Sulfur containing 1 - Propanol, 3 - (methylthio) - 106 981 993 70 Unknown m/z 121 _ RI 1002 121 1002 Unknown m/z 108 _ RI 1049 108 1049 Unknown m/z 133 _ RI 1110 133 1110 Table 3. 3 Annotations of B. anthracis Sterne - specific V OC markers through compound class, putative NIST ID, m/z , and retention index matching Class Compound MS B ase Peak ( m/z ) RI (Lit) RI (Exp) NIST 14 Match Factor Alcohols 4 - Heptanol 55 872 893 86 Carboxylic Acid Propanoic acid, 2 - methyl - 43 772 744 70 Car boxylic Acid Butanoic acid, 2 - methyl - 74 861 865 90 Carboxylic Acid Butanoic acid, 3 - methyl - 60 863 853 8 6 Ester Propanoic acid, 2 - methyl - , butyl ester 89 898 961 97 45 Table 3. 3 Ester Butanoic acid, butyl ester 71 995 1006 97 Ester Butyl 2 - m ethylbutanoate 103 1043 1053 89 Ester Butanoic acid, 3 - methyl - , butyl ester 85 1047 1058 83 Methyl Keton e Methyl Isobutyl Ketone 43 735 719 73 Methyl Ketone 2 - Hexanone, 5 - methyl - 43 862 848 80 Methyl Ketone 2 - Heptanone 43 891 889 94 Methyl Ketone 2 - He ptanone, 6 - methyl - 43 956 962 97 Methyl Ketone 2 - Heptanone, 5 - methyl - 43 971 973 73 Methyl Ketone 5 - Hept en - 2 - one, 6 - methyl - 108 986 977 70 Methyl Ketone 2 - Heptanone, 4,6 - dimethyl - 58 1045 1067 76 Sulfur containing compound Butanethioic acid, S - methyl e ster 43 874 834 78 Sulfur containing compound Thiopivalic acid 85 959 945 71 Unknown m/z 80 _ RI 715 80 715 Unknown m/z 57 _ RI 769 57 769 Unknown m/z 43 _ RI 791 43 791 Unknown m/z 43 _ RI 873 43 873 Unknown m/z 45 _ RI 901 45 901 Unkn own m/z 57 _ RI 912 57 912 Unknown m/z 43 _ RI 956 43 956 (methyl ketone) a m/z 58 _ RI 962 58 96 2 Unknown m/z 43 _ RI 997 43 997 Unknown m/z 90 _ RI 1005 90 1005 (methyl ketone) a m/z 58 _ RI 1104 58 1104 Unknown m/z 83 _ RI 1145 83 1145 (methyl ketone) a m/z 58 _ RI 1554 58 1554 a : GC/MS fragmentation similar to observed methy l ketones RI (Lit): Retention Index reported from NIST14 RI (Exp): Retention Index calculated from experiment 46 for both RG2 species, Ft novicida and Ba Sterne, are presented in Figure 3. 3. Logarithmic or l growth, was observed to last for 20 hours and 8 hours, respectively. The bacterial counts rose approximately 3 orders of magnitude for both species, pe aking at 1 - 2*10 9 CFU/mL for Ft novicida and 5*10 8 CFU/mL for Ba Sterne. For Ft novicida (Figure 3. 3 (a) ) the observed growth during that phase appeared rather variable. Ft cultures are known to be difficult to grow. Sampling more replicates may improve statistical confidence in future experiments. Stationary phase, occurring when the bacteria exhibit no Figure 3 .3 Growth curves of (a) F. tularensis novicida in modified M ueller - Hinton media over a 52 - hour time period and (b) B. anthracis Sterne in Brai n - Heart Infusion media over a 24 - hour time period. Data points and error bars represent the means and stan dard deviations of CFU/mL determined from three culture replicates. Red lines represent visually determined trends in bacterial growth across each gro wth phase. 47 additional growth due to a depleted nutrient source, was observed in both species. Ba Sterne m easurements were completed at 24 hours post - inoculation while still in stationary phase. For Ft novicida , further growth phase changes were observed t hrough a dec line in viable bacterial growth to ~1*10 6 CFU/mL at 32 h and no observable growth at the 48 h and 52 h post inoculation. The limit of detection for concentrations of viable bacteria was less than 1000 CFU/mL. Regardless, the CFU/mL counts were fairly consi stent across triplicate measurements in both taxa, allowing assessments of growth phase. The o bserved profiles of the VOC biomarkers varied with growth phase. The averaged relative abundances of biomarkers from Ft novicida and Ba Sterne are lis ted across all measured timepoints and grouped by compound class (Tables 3. 4, 3. 5; Appendix Tables A.3. 1, A.3. 2). The mean combined areas of each compound class are also stacked as a function of total area for each of the two RG2 species (Fig ure 3. 4). Of t he tentatively identified markers, chemical diversity was observed in the presence of ketones, aldehydes, alcohols, esters, carboxylic acids, nitrogen - or sulfur - containing markers, and alkanes. Although more biomarkers were detected for Ba Sterne (30) ve rsus Ft novicida (18), the combined peak areas (total signal) of markers from Ft novicida at its peak grow th (32 hours post - inoculation, stationary phase) were approximately 5x the total combined peak area of Ba Sterne VOCs at its peak growth (8 hours, log arithmic phase), attributed to the ~10x higher concentration of bacteria (compare Figure 3. 4 and Figure 3 . 3). There was only a moderate correlation between combined marker peak areas and the bacteria concentration at any single timepoint. Pe ak areas and b acterial counts rose during the logarithmic phase for both species, but cumulative peak areas were stagnan t or dropped during stationary phase despite bacterial concentration remaining steady. 48 Table 3. 4 Average relative abundances (Log - 10 Sca le) of F. tula rensis novicida - associated VOCs at all measured timepoints, separated by growth phase Time Post Inocula tion (Hours) 0 2 4 8 12 16 20 24 28 32 48 52 Growth Phase Log Phase Stationary Phase Decline Phase Class Compound Abundance (Log 10 Values) Alcohols 1 - Butanol, 2 - methyl - 0.00 0.00 0.00 0.00 5.87 b 5.84 b 6.77 6.55 a 6.77 6.76 5.70 5. 44 b Alcohols 2 - Nonanol 0.00 0.00 0.00 0.00 6.05 6.61 6.08 5.89 0.00 0.00 0.00 0.00 Alcohols Phenylethyl Alcohol 0.00 0.00 5.66 a 6.07 6.29 6.36 6.5 4 6.75 6.74 6.81 6.67 6.67 Alcohols 1 - Nonanol 0.00 0.00 5.80 6.35 6.70 7.26 7.46 6.98 6.43 6.24 5.90 5.85 Alcohols 2 - Undecanol 0.00 0.00 0.00 0.00 6.14 6.35 6.27 6.13 5.96 0.00 0.00 0.00 Methyl Ketones 2 - Heptanone 0.00 0.00 0.00 6.02 6.03 a 6.48 6.74 6. 84 6.82 6.88 6.46 6.33 Methyl Ketones 2 - Nonanone 0.00 0.00 0.00 6.44 6.60 6.84 7.11 7.31 7.41 7.48 6.96 6 .85 Methyl Ketones 2 - Undecanone 4.70 b,c 5.27 b,c 5.86 6.13 5.98 6.03 6.72 7.16 7.29 7.23 6.25 6.15 Methyl Ketones 2 - Tridecanone 0.00 0.00 0.00 0.00 0.00 0.00 6.14 6.91 7.01 6.89 6.09 6.02 Methyl Ketones 2 - Pentadecanone 0.00 0.00 0.00 0.00 0.00 0.00 4.9 6 b 6.23 6.50 6.51 6.03 6.04 Methyl Ketones 2 - Heptadecanone 0.00 0.00 0.00 4.71 b 0.00 0.00 0.00 5.27 5.80 5.88 5.63 5.69 Nitrogen - Containing Compou nds Pyrazine, 2,5 - dimethyl - 5.69 c 6.37 c 6.41 c 6.38 c 6.78 c 7.18 c 7.36 7.60 7.60 7.76 7.81 7.81 49 Table 3.4 Nitrogen - Containing Compounds 2 - Methyl - 3 - isopropylpyrazine 0.00 0.00 5.47 b 0.00 6.13 6.24 6.47 6.75 6.87 7.05 7.06 7.05 Sulfur - Cont aining Compounds Dimethyl trisulfide 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.90 6.84 Sulfur - C ontaining Compounds 1 - Propanol, 3 - (methylthio) - 0.00 0.00 0.00 0.00 5.72 a 5.76 a 5.98 6.05 6.00 6.07 6.08 6.13 Unknown m/z 121 _ RI 1002 0.00 6.14 5 .71 b 6.10 5.90 5.79 5.86 6.15 6.18 6.33 6.16 6.09 Unknown m/z 108 _ RI 1049 0.00 5.80 5.79 6.37 6.51 a 6 .50 a 6.54 a 6.62 6.60 6.69 6.56 6.58 Unknown m/z 133 _ RI 1110 0.00 0.00 0.00 0.00 0.00 0.00 5.07 b 6.07 6.01 6.08 5.94 5.87 Notes: a : VOC detec ted in 2/3 of triplicate measurements b : VOC detected in 1/3 of triplicate measurements c : VOC detected at levels less than 10x abundance in media blank 50 Table 3. 5 Average relative abundances (Log - 1 0 Scale) of B. anthracis Sterne - associated VOCs at all measured timepoints, separated by growth phase Time Post Inoculation (Hours) 4 8 12 20 24 Growth Pha se Log Phase Stationary Phase Class Compound Abundance (Log 10 Values) Alcohols 4 - Heptanol 6.15 0.00 0.00 0.00 0.00 Carboxylic Acid Propanoic ac id, 2 - methyl - 0.00 5.38 a 5.53 a 0.00 0.00 Carboxylic Acid Butanoic acid, 2 - methyl - 0.00 5.93 5.31 a 0.00 0.00 Carboxylic Acid Butanoic acid, 3 - methyl - 0.00 6.16 5.68 0.00 0.00 Ester Propanoic acid, 2 - methyl - , butyl ester 6.50 6.65 0.00 0.00 0.00 Ester Butanoic acid, butyl ester 6.70 6.22 0.00 0.00 0.00 Ester Butyl 2 - methylbutanoate 5.82 6.11 0.00 0.00 0. 00 Ester Butanoic acid, 3 - methyl - , butyl ester 5.44 6.23 4.81 a 0.00 0.00 Methyl Ketone Methyl Isobutyl Ketone 0.00 0.00 5.71 6.22 a 6.14 Methyl Ke tone 2 - Hexanone, 5 - methyl - 0.00 5.43 a 6.05 6.52 6.76 Methyl Ketone 2 - Heptanone 5.12 a 6.58 6.65 6.69 6.8 9 Methyl Ketone 2 - Heptanone, 6 - methyl - 0.00 0.00 6.62 6.92 7.02 Methyl Ketone 2 - Heptanone, 5 - methyl - 0.00 0.00 5.54 a 6.03 6.25 Methyl Ketone 5 - Hep ten - 2 - one, 6 - methyl - 0.00 0.00 0.00 4.34 5.49 Methyl Ketone 2 - Heptanone, 4,6 - dimethyl - 0.00 0.00 4.80 b 5 .26 a 5.64 Sulfur containing compound Butanethioic acid, S - methyl ester 0.00 5.35 5.57 5.89 0.00 Sulfur containing compound Thiopivalic acid 4.60 b 5.46 5.47 5.86 4.34 b Unknown m/z 80 _ RI 715 5.08 b,c 5.80 5.68 a 5.82 5.20 b Unknown m/z 57 _ RI 769 0 .00 0.00 4.54 b 5.19 0.00 Unknown m/z 43 _ RI 791 6.93 7.23 6.83 6.15 5.64 a Unknown m/z 43 _ RI 873 0.00 0.00 0.00 4.34 b 4.86 Unknown m/z 45 _ RI 901 0.00 0.00 5.16 a 5.30 0.00 Unknown m/z 57 _ RI 912 5.64 5.82 5.06 b 0.00 0.00 Unknown m/z 43 _ RI 9 56 5.18 0.00 0.00 0.00 0.00 (methyl ketone) a m/z 58 _ RI 962 0.00 5.82 0.00 0.00 0.00 Unknown m/z 43 _ RI 997 0.00 4.88 b 0.00 5.03 a 5.18 a Unknow n m/z 90 _ RI 1005 0.00 0.00 5.22 5.00 a 0.00 (methyl ketone) a m/z 58 _ RI 1104 5.62 5.79 5.67 5.13 b 5.7 9 Unknown m/z 83 _ RI 1145 0.00 5.11 4.57 a 0.00 0.00 (methyl ketone) a m/z 58 _ RI 1554 0.00 4.54 b 5.00 0.00 0.00 51 Notes: a : VOC detected in 2/3 of triplicate measurements b : VOC detected in 1/3 of triplicate measurements c : VOC d etected at levels less than 10x abundance in media blank d : GC/MS fragmentation similar to observed methyl ketones Figure 3. 4 Mean combined peak a reas (integrated detector counts) for compounds within individual classes at each timepoint post bacteria l inoculation of cultures for (a) F. tularensis novicida and (b) B. anthracis Sterne. The biomarker peak areas for Ft novicida steadily increased thro ughout the logarithmic and stationary phases before decreasing during the decline phase. Alcohols steadily rose in relative abundance throughout the log phase and were dominant in the early log and log phases. While some alcohols pe rsisted throughout the e ntire study, several were fully depleted at the 52 longest timepoints measured (see 2 - nonanol and 2 - undecanol ). Linear, odd - chain methyl ketones (or 2 - ketones) were present throughout all growth phases, with ketones consisting of more than 13 carbons (longer than 2 - tridecanone) being present only in the stationary phase and beyond. The contribution of methyl ket ones peaked in stationary phase growth, and their decrease in the decline phase lowered total VOC relative abundances. Nitroge n - containing markers wer e present throughout the analysis of F. tularensis species due to the presence of 2,5 - dimethylpyrazine, a marker that was a component of the growth media. However, the signal emitted from the bacterial cultures first exceeded 10x th e signal in the media co ntrol at the 20 - hour timepoint, prompting its inclusion as a potential F. tularensis marker. Combined with the signal from 2 - methyl - 3 - isopropylpyrazine, nitrogen - containing markers comprised almost 70% of the chemical profile for th e decline phase. Finally , Ft novicida noticeably displayed a large signal of dimethyltrisulfide as an abundant marker in the decli ne phase, comprising 7 - 8.5 % of the total VOC signal , formed as by - products of cysteine and methionine amino acid decompositio n after the death of bac terial cells. Biomarker areas for Ba Sterne also changed dependent on growth phase, though fewer timepoin ts were measured compared to Ft novicida . Esters, carboxylic acids, and alcohols comprised a significant portion of the logarithmic phase VOC marker p rofiles. Esters were based on butanoic or propanoic acids, with methyl groups at the 2 or 3 - carbon positio ns. Two carboxylic acids were also based on butanoic and propanoic acids, both methylated at the 2 - carbon position. Alcohols were only present in the early log phase. The non - detection of these markers during the stationary phases (with the exception of 2 - methyl - propanoic acid in early stationar y phase) suggests use as precursors for further synthesis. Relative abundances of methyl ketones significantly increased during stationary phase. Moreover, while Ft novicida was 53 dominated by straight - chain aliphatics , the methyl ketones in Ba Sterne contai ned methyl and aromatic substituents. Levels of VOC biomarkers for Ft novicida and Ba Sterne were subjected to principal component analysis (PCA) to visualize VOC profiles observed at different growth phases. The s cores plots in Figure 3. 5 ( a ) and 3. 5 ( b ) show clustering of all three culture replicates of the respective strains. The loadings plots, depicting the r elative importance of individual markers towards sample positioning on PCs 1 and 2, are described in great er detail in Appendix Figures A.3. 1a and A.3. 1b. PCA groupings similarly exhibited distinct groupings of timepoints into Figure 3. 5 PCA scores plot s for VOC marker profiles of (a) F. tularensis novicida and (b) B. anthracis Sterne generated using the pe ak areas of pathogen biomarkers across all timepoints. Each plot point represents one sample. Distinct chemical profiles were observed amongst labeled growth phases. Corresponding loadings plots to explain placement of samples are located in Appendix Figur e A.3. 1. 54 clusters as determined in Figure 3. 4 ( a ) and 3. 4 ( b ) for each species. The PCA scores plot provides additional verification of similarity of VO Cs from culture replicates - profiles from each timepoint (same color) were positioned more closely to eac h other than to replicates of an adjacent timepoint, demonstrating fairly reproducible VOC marker profiles. 3. 3 .2 Results from RG3 species Determinat ion of bacterial concentrations and VOC sampling of the RG3 pathogens grown in our BSL - 3 laboratory were p erformed at select time points, as shown in Appendix Table A.3. 3 for both Ft SCHUS4 and Ba Ames. Ba Ames exhibited growth throughout 24 hour s, with ba cterial concentrations rising to 2.5*10 7 CFU/mL at the 24 h time point. Meanwhile Ft SCHUS4 concentration remained stagnant around 6.7*10 5 CFU/mL throughout 24 hours of culture, hypothesized to remain in a lag phase after inoculation. The decont amination protocols developed for our work in the BSL - 3 laboratory on both Ft SCHUS4 and Ba Ames included wiping the SPME fiber exterior casing using bleach (see Appendix Information). There was a potential for VOCs adsorbed on the internal fibers to be in advertentl y oxidized. However, comparison of the VOC profiles from B. anthracis taxa obtained using the BSL - 2 protoc ol without bleach wiping and the BSL - 3 protocol that included the bleach wiping revealed a range of similar markers and/or compound classes, with no e vidence of oxidized by - products for the profiles obtained using the B SL - 3 protocol. The VOC marker profil e of Ba Ames displayed similarities to its RG2 counterpart Ba Sterne and is detailed in Table 3. 7, where 18 of the 56 putative markers were identified . The VOC marker profile at the 6 - hour timepoint of Ba Ames resembled the logarithmic VOC marker 55 profile o f Ba Sterne. Esters were the most abundant identified markers, consisting of propanoic and butanoic acid esters. Five esters were shared between both B. anthracis taxa. Subsequent compound classes included methyl ketone s and carboxylic acids. Conversely, t he VOC marker profile at the 24 - hour timepoint of Ba Ames more closely resembled the stationary VOC marker profile of Ba Sterne. Methyl ketones were t he dominant markers, while all esters have been depleted. Four methyl ketones were shared between B. anthr acis taxa. Principal component analysis (PCA) of the level of VOC biomarkers was also applied to Ba Ames to visualize VOC profiles at different growth phases. Similar to Ba Sterne, it appears that different chemical pro files can be associated with differen t growth phases of Ba Ames (not shown here), however, more data points would be needed to draw stronger conclusions. Conversely, the profile of VOCs from Ft SCHUS4 (Table 3. 6) had fewer similarities with Ft novicida . The majority of putative markers for F t SCHUS4 were classified as unknowns, with only 5 markers passing the conservative identification criteria. While none of the observed compounds passi ng the filtering criteri a were shared between either species, the 6 and 24 - hour timepoints both contained alcohols such as 4 - methyl - 3 - heptanol and 1 - dodecanol. Alcohols were also the dominant class of the logarithmic phase for Ft novicida . The lack of comp ound class similarities for determined markers could result from genetic differences between Ft SCHUS4 and Ft novicida . However, in agreement with the bacterial concentration data shown in Appendix Table A.3. 3, the CFU counts suggests tha t Ft SCHUS4 remain ed in a lag phase or a very early logarithmic phase throughout the first 24 h after inoculation. Additiona l measurements of the growth phases of Ft SCHUS4 over longer time periods are needed for a more comprehensive comparison of Ft SCHUS 4 VOC markers agai nst those of Ft novicida . 56 Table 3.6 Annotations of F. tularensis SCHU S4 - specific VOC markers and averag e relative abundances (n=3) at 6 and 24 hours (Hr) post inoculation Class Compound MS Base Peak ( m/z ) RI (Lit) RI (Exp) NIST14 Match Factor Abundance (Log10 Values) 6 Hr 24 Hr Alcohol 3 - Heptanol, 4 - methyl - 43 915 831 75 5.49 b 5.85 A lcohol 1 - Dodecanol 55 1473 1469 82 0.00 5.77 a Alcohol Phenol, 2,5 - bis(1,1 - dimethylethyl) - 191 1514 1508 76 4.59 b 4.87 a Aldehyde Furfural 96 833 8 18 73 4.77 b 4.94 a Nitrogen - containing compound Pyrazine, methyl - 94 801 802 76 0.00 5.43 a Unknown m/z 41 _ RI 672 41 672 0.00 5.65 a Unknown m/z 133 _ RI 913 133 913 0.00 7.42 Unknown m/z 57 _ RI 941 57 941 5.30 0.00 Unknown m/z 43 _ RI 986 43 986 4.81 a 0.00 Unknown m/z 105 _ RI 990 105 990 4.81 b 5.23 Unknown m/z 207 _ RI 998 207 99 8 0.00 5.46 a Unknown m/z 121 _ RI 1010 121 1010 5.30 0.00 Unknown m/z 69 _ RI 1034 69 1034 5.25 a 4.89 b Unknown m/z 43 _ RI 1063 43 1063 4.79 b 4.93 a Unknown m/z 71 _ RI 1185 71 1185 0.00 5.02 a Unknown m/z 57 _ RI 1199 57 1199 0.00 4.96 a Unknown m/z 57 _ RI 1303 57 1303 5.00 a 4.74 b Unknown m/z 119 _ RI 1365 119 1365 0.00 4.91 a Unknown m/z 57 _ RI 1370 57 1370 4. 90 0.00 Unknown m/z 57 _ RI 1398 57 1398 5.58 a 5.55 a Unknown m/z 69 _ RI 1527 69 1527 5.36 0.00 Unknown m/z 40 _ RI 1536 40 1536 4.74 a 0.00 Unknown m/z 163 _ RI 1667 163 1667 5.02 a 0.00 Unknown m/z 71 _ RI 1810 71 1810 4.23 b 4.61 a Unknown m/z 40 _ RI 1972 40 1972 4.50 a 0.00 Unknown m/z 73 _ RI 2363 73 2363 5.17 a 5.56 Unknown m/z 73 _ RI 2520 73 2520 5.00 a 4.56 b Unknown m/z 73 _ RI 2521 73 2521 4.80 b 5.31 a Unknown m/z 73 _ RI 2679 73 2679 4.58 b 5.09 a Unk nown m/z 96 _ RI 2777 96 2777 4.34 a 0.00 Unknown m/z 208 _ RI 3117 208 3117 5.52 a 0.00 Unknown m /z 207 _ RI 3154 207 3154 4.85 b 5.63 a 57 Table 3.6 Unknown m/z 97 _ RI 3306 97 3306 4.94 a 5.11 Unknown m/z 207 _ RI 3 356 207 3356 5.16 a 0.00 Unknown m/z 97 _ RI 3393 97 3393 0.00 4.88 a Unknown m/z 97 _ RI 3425 97 3425 0.00 4.88 a Unknown m/z 97 _ RI 3548 97 3548 4.63 b 4.95 a Unknown m/z 208 _ RI 3572 208 3572 0.00 4.84 a Notes: a : VOC de tected in 2/3 of triplicate measurements b : VOC detected in 1/3 of triplicate measurements RI (Lit): Reten tion Index reported from NIST14 RI (Exp): Retention Index calculated from experiment 58 Table 3.7 Annotations of B. anthracis Ames - specific VOC markers a nd average relative abundances (n=3) at 6 and 24 hours (Hr) post inoculation Class Compou nd MS Base Peak ( m/z ) RI (Lit) RI (Exp) NIST14 Match Factor Abundance (Log10 Values) 6 H r 24 H r Alcohol 3 - Octanol, 3,6 - dimethyl - 73 1043 1110 83 5 .96 b 6.03 a Carboxylic Acid Butanoic acid, 2 - methyl - 74 861 844 89 5.73 0.00 Ester Acetic acid, butyl e ster 43 812 794 96 7.29 0.00 Ester Propanoic acid, 2 - methyl - , butyl ester 89 898 964 97 7.03 0.00 Ester Butanoic acid, butyl ester 71 995 1009 80 5. 95 0.00 Ester Butyl 2 - methylbutanoate 103 1043 1055 97 6.48 0.00 Ester Butanoic acid, 3 - methyl - , butyl e ster 85 1047 1060 95 6.40 0.00 Ester 2 - Butenoic acid, 2 - methyl - , 2 - methylpropyl ester, (E) - 101 1112 1146 74 5.36 0.00 Ketone 3 - Octanone 43 986 999 71 5.28 5.31 Methyl Ketone Acetoin 45 713 691 85 6.86 0.00 Methyl Ketone Methyl Isobutyl Ketone 43 735 7 12 89 0.00 6.39 Methyl Ketone 2 - Pentanone, 3 - methyl - 43 752 722 77 5.29 a 5.84 Methyl Ketone 2 - Hexanone, 5 - methyl - 43 862 851 94 0.00 6.64 Methyl K etone 2 - Heptanone 43 891 892 96 6.28 6.67 Methyl Ketone 2 - Heptanone, 6 - methyl - 43 956 965 93 0.00 7.16 M ethyl Ketone 2 - Nonanone 58 1092 1069 86 0.00 5.97 Methyl Ketone Benzyl methyl ketone 91 1110 1141 85 0.00 5.67 Nitrogen - Containing Compound Pyrazine , tetramethyl - 136 1089 1099 77 4.85 b 5.80 Unknown m/z 42 _ RI 748 42 748 0.00 5.29 Unknown m/z 40 _ RI 874 40 874 5.01 a 0.00 Unknown m/z 42 _ RI 912 42 912 7.18 b 7.67 Unknown m/z 57 _ RI 915 57 915 5.68 a 0.00 Unknown m/z 43 _ RI 938 4 3 938 5.15 a 0.00 Unknown m/z 93 _ RI 939 93 939 4.63 b 5.04 a Unknown m/z 43 _ RI 976 43 976 4.95 a 6.26 59 Table 3.7 Unknown m/z 57 _ RI 1012 57 1012 5.89 a 0.00 Unknown m/z 69 _ RI 1034 69 1034 5.41 a 0.00 Unknown m/z 43 _ RI 1054 43 1054 0.00 5.23 Unknown m/z 55 _ RI 1057 55 1057 0.00 5.36 Unknown m/z 43 _ RI 1063 43 1063 0.00 5.00 a Unknown m/z 57 _ RI 1071 57 1071 4.60 b 4.99 a Unknown m/z 130 _ RI 1126 130 1126 0.00 5.40 a Unknown m/z 149 _ RI 1169 14 9 1169 5.41 a 0.00 Unknown m/z 57 _ RI 1180 57 1180 5.21 5.31 Unknown m/z 91 _ RI 1189 91 1189 0.00 5.29 Unknown m/z 55 _ RI 1201 55 1201 5.26 0.00 Unknown m/z 108 _ RI 1205 108 1205 4.42 b 4.83 a Unknown m/z 339 _ RI 1218 339 1218 4 .32 b 4.65 a Unknown m/z 71 _ RI 1285 71 1285 4.56 b 4.72 a Unknown m/z 119 _ RI 1365 119 1365 0.0 0 4.85 a Unknown m/z 401 _ RI 1564 401 1564 0.00 4.96 a Unknown m/z 405 _ RI 1685 405 1685 0.00 4.66 a Unknown m/z 40 _ RI 1747 40 1747 0 .00 4.36 a Unknown m/z 40 _ RI 2271 40 2271 4.46 a 0.00 Unknown m/z 73 _ RI 2521 73 2521 0.00 5.09 a Unknown m/z 281 _ RI 2674 281 2674 0.00 5.07 a Unknown m/z 281 _ RI 2983 281 2983 5.39 a 4.66 b Unknown m/z 281 _ RI 3008 281 3008 5.54 a 0.00 Unknown m/z 208 _ RI 3069 208 3069 5.86 a 5.43 b Unknown m/z 209 _ RI 3135 209 3135 5.31 a 0.00 Unknown m/z 207 _ RI 3154 207 3154 5.04 b 5.72 Unknown m/z 281 _ RI 3173 281 3173 5.71 a 0.00 60 Table 3.7 Unknown m/z 281 _ RI 3190 281 3190 5.64 a 0.00 Unknown m/z 208 _ RI 3439 208 3494 0.00 5.04 a Unknown m/z 97 _ RI 3521 97 3521 0.00 4.87 a Unknown m/z 97 _ RI 3533 97 3533 0.00 4.91 a Notes: a : VOC detected in 2/3 of triplicate measurements b : VOC detected in 1/3 of triplicate measurements RI (Lit): Retention Index reported from NIST14 RI (Exp): R etention Index calculated from experiment 3.4 DISCUSSION The methodology and results described here provide initial groundwork for detection and iden tification of volatile biomarkers f rom bacterial pathogens including fully virulent RG3 strains. The appli cation of this non - invasive methodology for VOC profiling applied to actively growing F. tularensis and B. anthracis bacterial cultures revealed dynam ic profiles, influenced by both the bacterial growth phase and bacterial concentration. At any given timep oint, isolation of the bacterial biomarkers was complicated by background volatiles, and data processing was applied uniformly across all sample types to identify bacterial biomarkers. As the VOC profiles observed here are discussed , one should keep in min d that measured VOC profiles were influenced by the sampling and detection methods used. For example, the type of sorbent material could have introduc e d a sampling bias (sampling effici ency is dependent on partition behavior of each compound) and the sampl ing time and mass spectral analysis method could have influence d the sensitivity with which compounds can be detected. Generally, detection limits for the basic type of SPME - GC - quadrupo le MS used here for untargeted analysis 61 (scan, not select ion mode) are on the order 1 ng of a compound injected into a column. It is conceivable that only the most prevalent VOCs were detected. A larger number of relevan t may be found if more efficient sa mpling techniques (e.g. , thermal desorption tubes) and more sensitive m ass spectrometry protocols such as selected ion monitoring are used. It should be noted here that absolute VOC quantification was not attempted here. For various practical reasons, absolute quantification of VOCs in entire cultures is challenging . The idea l standards, stable isotope - labeled internal standards , are susceptible to metabolic degradation. Also, cultures used here were not fully enclosed (fl asks with vented caps) because gas exchange (oxygen) was required to sustain growth and needed to be vente d to avoid buildup of pressure. However, relative abundances of compounds were compared among cultures by integrating chromatographic peaks across spe cies and timepoints. The cumulative VOC profiles of Ft novicida , Ft SCHUS4, Ba Sterne, and Ba Ames determ ined here included representatives of different compound classes such as methyl ketones, alcohols, nitrogen - containing compounds, sulfur - containing co mpounds, carboxylic ac ids, esters, and various unidentified biomarkers. Exhaustive identification of every pathway that produces these volatiles is beyond the scope of this discussion, but several likely routes of biosynthesis are enumerated below. 3.4.1 Ketones Ketones were abundant markers, present in all pathogens except Ft SCHUS4, and largely as methyl ke tones. The methyl ketones are likely formed by modifying products of the fatty acid - oxidation of fatty acids [26] . Odd - chain methyl ketones 62 can be formed through the decarboxyl ation of even - - keto acids. Conversely, even - carbon methyl ketones arise from odd - carbon fatty acids and occur with lower frequency [27] . Interestingly, methyl ketones with straight - chain a lkane branches were abundant in Ft novicida , while primarily branched and aromatic methyl ketones were prevalent in Ba Sterne and Ames. This difference may st em from B. anthracis being Gram - positive, whereas F. tularensis is Gram - negative. Synthesis of fat ty acids in Gram - positive and Gram - negative bacteria is controlled by enzymes with different preferred substrates. For example, comparison of the - ket oacyl - acyl carrier protein synthase III from Gram - negative E. coli and Gram - positive bacteria S. a ureus demonstrated a larger binding pocket in the Gram - positive bacteria, thus having a higher likelihood for synthesis of branched - chain alkyl substrates [28] . A smaller binding pocket for the Gram - positive bacteria would limit the use of branched - chain alkyl substrates. 3.4.2 Alcohols While alcohols were present for both F. tularensis and B. anth racis , their number and relative abundances were greater in F. tularensis . Alcohols may be synthesized from the - oxidation of fatty acids, for example after enzymatic reduction of carboxylic acids [27,29] . The fatty acid chains observed for the al cohols class exhi bited diversity, including straight - chain, branched chain, and aromatic substituents. 1 - nonanol was likely formed by reduction of fatty acid s . The 2 - alkanols (2 - nonanol and 2 - undecanol) are postulated to be derived from corresponding methyl ketones as redu ced intermediates, since the corresponding methyl ketones were also detected at all timepoints where the 2 - alkanols were detected, usually at a higher relative abundance. 2 - nonanol was not observed after 24 hours post - 63 in oculation, indicating it was utilize d by F. tularensis as a precursor for stationary phase metabolism. The aromatic alcohol phenylethyl alcohol is a widely occurring VOC produced by several bacterial species. Volatile alcohols have been shown to play a rol e in growth inhibition of several ba cteria and fungi [30 ,31] . 3.4.3 Sulfur - containing compounds Dimethyltrisulfide was an abundant VOC uniquely present in Ft novicida during the decline phase, when no viable bacteria were detected. Sulfur - containi ng VOCs are attributed to breakdown of the amino acids cysteine and methionine [27] . Dimethyltrisulfide has previously been observed as product of human decomposition caused by bacteria. 3.4.4 N itrogen - containing compounds The detection of nitrogen - containing pyrazine markers produced by bacteria is complicated by endogenous pyrazine VOCs present in the growth media [27] . The sterilizati on of growth media through aut oclaving heats amino acids and reducing sugars, producing pyrazines via the Maillard Reaction [32] . The Mueller - Hinton growth media controls consistently produced 2,5 - dimethylpyrazine, and a similar relative abundance was observed in the Ft novicida cultures through the first 16 h of growth. At 20 h of growth and beyond, the relative abundanc e of 2,5 - dimethylpyrazine rose more than 10x the relative abundance of the controls, suggesting the growing bacteria have active involvement in biosynthesis of pyrazines. An additional pyrazine, 2 - methyl - 3 - isopropylpyrazine, was also observed. Isopropyl su bstituents to pyrazine compoun ds are not common constituents in bacterial volatiles [27] . Therefore, we 64 hypothesize both pyrazines originate from Ft novicida under the chosen growth conditions. Pyr azines have also been observed as volatile byproducts of bacterial metabolism, for example, in the genera Streptomyces [33] and Bacillus [34] . Only one nitrogen - containing compound, tetramethylpyrazine, was observed in Ba Ame s during the last observed timepoint, estimated to be in the logarithmic phase, but was not as abundant in as in Ft novicida . This is compounded by the different growth media utilized, which emphasizes the need for careful evaluation when comparing biomark ers across different growth condition s and species. 3.4.5 Esters and carboxylic acid compounds Esters and carboxylic acids were detected exclusively in both Ba Sterne and Ba Ames, but not in the F. tularensis strains. The identified B. anthracis markers contained either propanoic or butanoic acid as the backbone for methylated esters or the side - chain for carboxylic acids. The formation of esters and carboxylic acids can be derived from shared metabolic pathways occurring during normal bacterial growth, such as oxidation of fatty acids or amino acid metabolism. As the ester and carboxylic acid compounds were only observed during the logarithmic growth stage, this demonstrates a shift in B. anthracis metabolism once bacteria reach the stationary ph ase. 3 .4.6 Evidence of dynamic metabolic processes For all four bacterial taxa studied here, their VOC marker profiles varied as function of time after inoculation/culture start and, as observed for Ft novicida , Ba Sterne and Ba Ames, varied distinctly across t heir growth phases. For Ft novicida and Ba Sterne, this was also shown 65 through application of PCA, which produced distinct groupings for the VOC markers of different growth phases. For example, in the Ft novicida cultures, the methyl ketones, once produced , were present throughout the remainder of the experiment. However, select alcohols (2 - nonanol and 2 - undecanol) were not detected after a mid - stationary (28 - hour) timepoint. This suggests alcohols were depleted in the liquid culture, potentially as precurs ors in ongoing bacterial metabolism. Once the basic metabolism of isolated pathogens is determined, changes in the marker profiles when additional variables are added (e.g. , different substrates) can help drive inferences on metabolic activity of c omplex s ystems. 3.4.7 Comparison of VOC markers for RG3 vs. RG2 strains In comparing the putative VOC biomarkers identified for Ba Ames (RG3) to those for Ba Sterne (RG2) some similarities were found , but also distinct differences. In contrast, the profile of VO Cs from Ft SCHU S4 (RG3) had fewer similarities with Ft novicida (RG2) . The relative similarities between Ba Ames and Ba Sterne may stem from the close genetic relationship of these two strains ( Ba Sterne is missing one of the two plasmids that Ba Ames has but is otherwise genetically very similar to Ba Ames [35] ). In co ntrast, Ft SCHU S4 and Ft novicida are genetically more distinc t [36] . If further confirmed in future studies, this may have implications related RG3 pathogens. In the b iodefense co mmunity Ba Sterne is generally considered a good simulant for Ba Ames. For Ft RG2 simulants other than Ft novicida may be considered. Future VOC sampling should be performed on additional subspecies of F. tularensis (e.g., spp. holarctica ) and B. anthracis (e.g., spp. Vollum or H9401) to investigate whether these profiles are unique to a subspecies, species, or bacterial pathogens in general. Several markers 66 identified in this study have been previously reported as emissions of other bacterial types. For ex ample, Chen et al. [12] reported 2 - heptanone, 2 - nonanone, and 2 - undecanone in E. coli but did not detect hig her carbon methyl ketones. Rees et al. [10,11] reported both even and o dd - chain methyl ketones, including 2 - hexanone, 2 - heptanone, 2 - nonanone, and 2 - decanone products from Klebsiella pneumoniae , where the presence of even - chain methyl ketones suggests a different or complementary synthesis pathway for volatile production. One of the long - term goals of this project seeks to use VOCs as breath - based diagnostic markers towards the detection of biowarfare agents in patients after a suspected biological attack. During a hypothetical pathogen infection in humans, the VOCs in breath may be derived from 1) the inva ding pathogen, 2) the human breath volatilome, or 3) interactions between the human host and pathogen. This study represents a first step in non - invasive methodology and data analysis optimization, extensively profiling two a ttenuated pathogen or RG2 speci es and screening their RG3 virulent counterparts in optimized growth media. The number of compounds identified in the human breath volatilome continues to grow through targeted and untargeted studies. A searchable database of breath - specific compounds in t curated by the U.S. Environmental Protection Agency (EPA) and is continuously updated [37,38] . A survey of this list against th e F. tularensis RG2 and RG3 profiles found here revealed 2 - heptanone and 2 - nonanone have been detected in human breath, while a comparison against the B. anthracis RG2 and RG3 profiles revealed 2 - methylpropanoic acid, 2 - hepta none, 6 - methyl - 2 - heptanone, and 5 - methyl - 5 - hepten - 2 - one that have been reported in human breath. Also, it is important to note that the volatiles in this database may not be commonly shared among all people, as human breath has been shown to be influenced iome, external exposures, and immunological responses. The effects of shared volatiles between 67 pathogens and a human host must be evaluated in further studies that better simulate an in vivo infection, as well as identifying markers unique to that interact ion. - specific volatilome, and host - microbe interactions are required for evaluation of VOCs as diagnostic tools for human health. Towards a pathogen - specific vo latilome, further efforts inclu de expanding both the number of bacterial species and evaluating the effects of chosen growth media on VOCs produced. Additionally, as animal model studies have established a low bacterial count can establish infections (e.g. , 10 bacterial counts for F. tu larensis in primate models), optimization of signal detection will also be investigated, as the conditions employed here used relatively high bacterial counts. Finally, future work should make the transition from in vitro stu dies into experiments more clos ely aligned with in vivo studies, such as initiating bacterial infection of human lung cell cultures and analyzing the resultant profiles for discovery of overlapping volatile compounds that may serve as diagnostic markers of human exposures to biosecurity - relevant pathogens. 3.5 CONCLUSIONS This study adapted a SPME - GC - MS methodology for noninvasive profiling of VOCs emitted from actively growing pathogens, specifically potential biowarfare bacterial agents and their surro gates, in both BSL - 2 and BSL - 3 settings. The devised methodology detect ed volatile biomarkers that were reflective of both the presence and physiological growth phase of pathogens. The data processing employed distinguished signals from the pathogens again st a complex chemical background, in this case aided by the use of powe rful software (MassHunter, MPP) for compound annotation and visualization of GC - MS data. Although the devised 68 methodology based on SPME - GC - quadrupole MS does not represent the pinnacle of sensitivity, a number of relatively robust and reproducible putative volatile biomarkers could be detected. Further confirmation of these markers should be pursued in more repeat experiments across a wider range of growth conditions. More efficient VOC collection methods and more sensitive mass spectral analysis techniques may also uncover additional markers in the future. Detection and identification of metabolites specific to taxa or species provides the first steps to understanding their formation vi a various metabolic pathways and the genetic basis for these pathways. It should be acknowledge d that the work presented here constitutes only initial scoping experiments. While this work demonstrates the applicability of this method and found a number of interesting volatile biomarkers, this work needs to be expanded to dete rmine the influence of various experimental factors on markers. F uture research should include determining the dependence of pathogen - produced volatiles on environmental conditions (e.g . , chosen growth media) and use of different VOC collection methods (e. g. , thermal desorption tubes) to achieve lower detection limits. Elucidation of comprehensive bacterial profiles is expected to provide clues about bacterial metabolism in controlled en vironments, which can further inform research into metabolic processes when pathogens are in other settings (i.e , . a host). Ultimately, such biomarkers may yield useful information about metabolism in bacterial taxa and may facilitate new applications in b iodetection. Distinct volatile profiles have potential to be used for t he detection of pathogens in the context of biosecurity - relevant exposures of humans during a biological attack. Results from this work have implications in the larger volatilomics com munity, both within the field of pathogens study and beyond. While vola tile compounds from B. anthracis have been previously studied, this is the first study, to our knowledge, to profile volatile 69 emissions of F. tularensis . Future databases can incorporat e biomarker signatures from various pathogen species for means of relev ant comparisons. 70 APPEND I CES 71 APPENDI X A : Figures Figure A.3. 1 PCA loadings plots for VOC profiles of (a) F. tularensis novicida and (b) B. an thracis Sterne generated using the relative abundances of pathogen VOCs measured using GC - MS across all timepoints (For PCA score plots see Figure 3. 5). Points represent individual VOC markers (colored by compound class) explaining placement of sample s on scores plot. The loadings plot s shown in Appendix Figure A.3.1 a and A.3.1 b depicts the relative importance of each volatile towards the positioning of individual samples in the scores plot s (Figures 3.5a and 3.5b, main manuscript) . For F. tularensis spp. n ovicida, the first two principal components (PCs) described 63 % and 14 ± % of the variance in data, respectively. Positioning on PC1 is dependent on the abundances of 2,5 - dimethylpyrazine and 2 - undecanone. Both compounds are present at all timepoint s and generally increase with higher timepoints, creating the left - right distribution seen in the scores plot. Positive loading on PC2 indicates a higher percentage of alcohol - containing compounds, indicative of the early log phase and early stationary pha se. A progressive shift towards a negative PC2 loading indicates an increasing percentage of methyl ketones, which rise to prominence in the late stationary phase, culminating in the appearance of dimethyltrisulfide. Notably, the volatiles that contributed the m ost to growth 72 phase separation, those of PC 2, were not the most abundant volatile classes, demonstrating the power of this analysis when characterizing trace compounds in a complex chemical profile. For B. anthracis Sterne, the first two PCs describ ed 32 % and 28 % of the variance in data, respectively. Positioning on PC1 in the scores plots exhibits strong dependence on the abundances of 4 - heptanol (negative loading) and the sulfur - containing compounds (S - methyl ester butanethioic acid and thiopival ic ac id) and ca rboxylic acids (methylated butanoic acids and 2 - ethyl - propanoic acid). A positive positioning on PC2 separates the logarithmic phase, containing a high percentage of esters, from the negatively positioned stationary phase, containing a high perce ntage of m ethyl ketones. 73 APP ENDIX B: Tables Table A.3.1 Relative abundances of F. tularensis novicida - associated VOCs for each replicate at all measured timepoints, separated by growth phase Time Post Inoculation (Hours) 0 2 4 8 12 16 20 24 28 32 48 52 Growth Phase Log Phase Stationary Phase Decline Phase Compounds Compound Class Replicate Abundance (Log 10 Values) 1 - Butanol, 2 - methyl - Alcohols 1 0.00 0.00 0.00 0.00 0.00 0.00 6.72 6.65 6.82 6.83 0.00 0.00 2 0.00 0.00 0.00 0. 00 6.35 6.31 6.66 0.00 6.33 6.51 6.18 5.92 3 0.00 0.00 0.00 0.00 0.00 0.00 6.89 6.79 6.96 6.85 0.00 0.00 2 - Nonanol Alcohols 1 0.00 0.00 0.00 0.00 6.02 6.61 6.07 5.88 0.00 0.00 0.00 0.00 2 0.00 0.00 0.00 0.00 6.06 6.60 6.10 5.87 0.00 0.00 0.00 0.00 3 0.00 0.00 0.00 0.00 6.07 6.63 6.09 5.90 0.00 0.00 0.00 0.00 Phenylethyl Alcohol Alcohols 1 0.00 0.00 5.88 6.12 6.31 6.34 6.52 6.79 6.81 6.80 6.72 6.64 2 0.00 0.00 5.79 6.03 6.27 6.35 6.47 6.70 6.68 6.83 6.66 6.72 3 0.00 0.00 0.00 6.04 6.28 6.40 6.61 6.7 5 6.73 6.79 6.64 6.64 1 - Nonanol Alcohols 1 0.00 0.00 5.88 6.44 6.72 7.25 7.46 7.02 6.48 6.34 5.94 5.83 2 0.00 0.00 5.75 6.26 6.72 7.28 7.54 7.11 6.52 6.26 5.96 5.95 3 0.00 0.00 5.76 6.33 6.67 7.24 7.35 6.69 6.26 6.09 5.76 5.76 2 - Undecanol A lcohols 1 0.00 0.00 0.00 0.00 6.13 6.34 6.26 6.16 5.97 0.00 0.00 0.00 2 0.00 0.00 0.00 0.00 6.15 6.33 6.25 6.17 5.90 0.00 0.00 0.00 3 0.00 0.00 0.00 0.00 6.15 6.37 6.30 6.03 6.00 0.00 0.00 0.00 2 - Heptanone Methyl Ketones 1 0.00 0.00 0.00 6.05 0.00 6 .52 6.71 6.82 6.84 6.87 6.51 6.21 2 0.00 0.00 0.00 5.74 6.18 6.46 6.73 6.82 6.68 6.88 6.46 6.38 3 0.00 0.00 0.00 6.16 6.24 6.47 6.79 6.87 6.90 6.90 6.41 6.38 2 - Nonanone Methyl Ketones 1 0.00 0.00 0.00 6.42 6.53 6.85 7.04 7.26 7.37 7.48 7.01 6.76 2 0.00 0 .00 0.00 6.37 6.68 6.81 7.12 7.25 7.32 7.44 6.97 6.87 3 0.00 0.00 0.00 6.52 6.58 6.85 7.17 7.41 7.52 7.51 6.89 6.91 74 2 - Undecanone Methyl Ketones 1 5.17 0.00 5.90 6.15 5.96 6.04 6.64 7.09 7.20 7.26 6.28 6.06 2 0.00 0.00 5.87 6. 12 6.05 6.05 6.60 7.07 7.23 7.17 6.25 6.19 3 0.00 5.75 5.80 6.12 5.92 5.99 6.87 7.30 7.41 7.26 6.21 6.20 2 - Tridecanone Methyl Ketones 1 0.00 0.00 0.00 0.00 0.00 0.00 5.96 6.82 6.94 6.97 6.09 5.92 2 0.00 0.00 0.00 0.00 0.00 0.00 5.99 6.84 7.0 0 6.80 6 .06 6.07 3 0.00 0.00 0.00 0.00 0.00 0.00 6.34 7.03 7.09 6.88 6.13 6.06 2 - Pentadecanone Methyl Ketones 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.20 6.43 6.55 6.03 5.92 2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.12 6.43 6.42 5.96 6.09 3 0.00 0.00 0.00 0.0 0 0.00 0.00 5.43 6.34 6.62 6.55 6.10 6.08 2 - Heptadecanone Methyl Ketones 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.42 5.90 5.65 5.68 5.65 2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.69 5.53 5.57 5.78 3 0.00 0.00 0.00 5.19 0.00 0.00 0.00 5.47 5. 78 6.17 5.62 5.64 Pyrazine, 2,5 - dimethyl - Nitrogen - Containing Compounds 1 0.00 6.35 6.45 6.46 6.72 7.10 7.36 7.57 7.56 7.78 7.71 7.76 2 5.72 6.39 6.36 6.30 6.86 7.24 7.44 7.71 7.44 7.66 7.82 7.78 3 5.97 6.38 6.40 6.36 6.77 7.19 7.27 7.48 7.74 7.81 7 .89 7.88 2 - Methyl - 3 - isopropylpyrazine Nitrogen - Containing Compounds 1 0.00 0.00 0.00 0.00 6.11 6.21 6.45 6.74 6.90 7.04 7.08 7.00 2 0.00 0.00 5.94 0.00 6.15 6.23 6.45 6.71 6.77 7.04 7.03 7.08 3 0.00 0.00 0.00 0.00 6.12 6.27 6.51 6.79 6.92 7.08 7.07 7.07 Di methyl trisulfide Sulfur - Containing Compounds 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 7.00 6.55 2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.91 6.86 3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6.74 7.01 1 - Propano l, 3 - (me thylthio) - Sulfur - Containing Compounds 1 0.00 0.00 0.00 0.00 5.87 0.00 5.88 6.04 5.93 6.11 6.09 6.05 2 0.00 0.00 0.00 0.00 5.93 5.86 6.01 6.01 5.93 6.07 6.07 6.19 3 0.00 0.00 0.00 0.00 0.00 6.00 6.03 6.09 6.11 6.03 6.08 6.13 m/z 121 _ RI 100 2 Unknow n 1 0.00 6.19 6.19 6.29 6.07 5.98 5.99 6.15 6.23 6.30 6.22 5.99 2 0.00 6.13 0.00 5.96 0.00 0.00 0.00 6.15 6.06 6.34 6.19 6.16 3 0.00 6.10 0.00 5.96 6.08 5.95 6.08 6.16 6.24 6.35 6.08 6.11 75 m/z 108 _ RI 104 9 Unknown 1 0.00 5.9 1 5.67 6.39 6.53 6.48 6.54 6.67 6.62 6.70 6.62 6.55 2 0.00 5.54 5.83 6.38 6.49 6.51 6.55 6.62 6.56 6.70 6.59 6.63 3 0.00 5.87 5.85 6.34 6.52 6.51 6.53 6.57 6.61 6.67 6.44 6.57 m/z 133 _ RI 1110 Unknown 1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 6 .22 6.09 6.03 5.92 5.83 2 0.00 0.00 0.00 0.00 0.00 0.00 5.55 6.02 5.92 6.05 5.98 5.91 3 0.00 0.00 0.00 0.00 0.00 0.00 0.00 5.90 6.00 6.14 5.91 5.87 76 Table A.3.2 Relative abundances of B. anthracis Sterne - associated VOCs for each replicate at all mea sured ti mepoints, separated by growth phase T ime Post Inoculation (Hours) 4 8 12 20 24 Growth Phase Log Phase Stationary Phase Compound Class Compound Replicate Abundance (Log 10 Values) Alcohols 4 - Heptanol 1 6.10 0.00 0.00 0.00 0.00 2 6.21 0 .00 0.00 0.00 0.00 3 6.12 0.00 0.00 0.00 0.00 Carboxylic Acid Propanoic acid, 2 - methyl - 1 0.00 5.49 0.00 0.00 0.00 2 0.00 5.62 5.93 0.00 0.00 3 0.00 0.00 5.20 0.00 0.00 Carboxylic Acid Butanoic acid, 2 - methyl - 1 0.00 5.82 0.00 0.00 0.00 2 0.0 0 5.78 5 .62 0.00 0.00 3 0.00 6.11 5.29 0.00 0.00 Carboxylic Acid Butanoic acid, 3 - methyl - 1 0.00 6.09 5.47 0.00 0.00 2 0.00 6.08 5.87 0.00 0.00 3 0.00 6.27 5.58 0.00 0.00 Ester Propanoic acid, 2 - methyl - , butyl ester 1 6.50 6.48 0.00 0.00 0.00 2 6.53 6.59 0.00 0.00 0.00 3 6.46 6.81 0.00 0.00 0.00 Ester Butanoic acid, butyl ester 1 6.70 6.12 0.00 0.00 0.00 2 6.74 6.15 0.00 0.00 0.00 3 6.66 6.35 0.00 0.00 0.00 Ester Butyl 2 - methylbutanoate 1 5.83 5.95 0.00 0.00 0.00 2 5.86 6.05 0.00 0.00 0. 00 3 5.75 6.26 0.00 0.00 0.00 77 Table A.3.2 Ester Butanoic acid, 3 - methyl - , butyl ester 1 5.42 6.10 4.94 0.00 0.00 2 5.48 6.17 0.00 0.00 0.00 3 5.40 6.37 5.03 0.00 0.00 Methyl Ketone Methyl Isobutyl Ketone 1 0.00 0.0 0 0.00 6.30 6.10 2 0.00 0.00 5.83 6.47 6.23 3 0.00 0.00 5.93 0.00 6.07 Methyl Ketone 2 - Hexanone, 5 - methyl - 1 0.00 0.00 6.04 6.55 6.63 2 0.00 5.40 5.90 6.63 6.96 3 0.00 5.75 6.18 6.30 6.59 Methyl Ketone 2 - Heptanone 1 0.00 6.50 6.71 6.6 8 6.68 2 5.30 6.60 6.54 6.82 7.12 3 5.30 6.64 6.68 6.52 6.72 Methyl Ketone 2 - Heptanone, 6 - methyl - 1 0.00 0.00 6.65 6.88 6.79 2 0.00 0.00 6.56 7.03 7.28 3 0.00 0.00 6.63 6.83 6.81 Methyl Ketone 2 - Heptanone, 5 - methyl - 1 0.00 0.00 5.66 5.99 6.04 2 0. 00 0.00 0.00 6.12 6.51 3 0.00 0.00 5.76 5.96 5.98 Methyl Ketone 5 - Hepten - 2 - one, 6 - methyl - 1 0.00 0.00 0.00 0.00 5.39 2 0.00 0.00 0.00 0.00 5.72 3 0.00 0.00 0.00 4.81 5.21 Methyl Ketone 2 - Heptanone, 4,6 - dimethyl - 1 0.00 0.00 5.27 0.00 5.4 6 2 0 .00 0.00 0.00 5.59 5.88 3 0.00 0.00 0.00 5.21 5.44 Sulfur containing compound Butanethioic acid, S - methyl ester 1 0.00 0.00 5.56 5.97 0.00 2 0.00 5.48 5.38 5.88 0.00 3 0.00 5.57 5.70 5.82 0.00 78 Table A.3.2 Sulfur containing c ompound Thiopivalic acid 1 5.08 5.41 5.49 5.99 0.00 2 0.00 5.50 5.37 5.71 0.00 3 0.00 5.48 5.54 5.84 4.82 Unknown m/z 80 _ RI 715 1 0.00 5.82 0.00 5.76 0.00 2 5.55 5.80 5.70 5.88 5.68 3 0.00 5.78 5.98 5.79 0.00 Unknown m/z 57 _ RI 769 1 0.00 0.00 5.0 2 5.05 0.00 2 0.00 0.00 0.00 5.28 0.00 3 0.00 0.00 0.00 5.21 0.00 Unknown m/z 43 _ RI 791 1 6.95 7.17 6.89 6.29 5.92 2 6.97 7.10 6.67 6.09 5.69 3 6.84 7.37 6.91 6.03 0.00 Unknown m/z 43 _ RI 873 1 0.00 0.00 0.00 0.00 4.92 2 0.00 0. 00 0.00 4.82 4.89 3 0.00 0.00 0.00 0.00 4.74 Unknown m/z 45 _ RI 901 1 0.00 0.00 0.00 5.41 0.00 2 0.00 0.00 5.21 5.07 0.00 3 0.00 0.00 5.43 5.35 0.00 Unknown m/z 57 _ RI 912 1 5.63 5.67 0.00 0.00 0.00 2 5.70 5.81 0.00 0.00 0.00 3 5.58 5.94 5.54 0. 00 0.00 Unknown m/z 43 _ RI 956 1 5.11 0.00 0.00 0.00 0.00 2 5.20 0.00 0.00 0.00 0.00 3 5.23 0.00 0.00 0.00 0.00 Unknown m/z 58 _ RI 962 1 0.00 5.82 0.00 0.00 0.00 2 0.00 5.81 0.00 0.00 0.00 3 0.00 5.84 0.00 0.00 0.00 79 Table A.3.2 (co Unknown m/z 43 _ RI 997 1 0.00 0.00 0.00 5.18 0.00 2 0.00 0.00 0.00 5.23 5.39 3 0.00 5.35 0.00 0.00 5.32 Unknown m/z 90 _ RI 1005 1 0.00 0.00 5.11 5.16 0.00 2 0.00 0.00 5.22 0.00 0.00 3 0.00 0.00 5.31 5.18 0.00 Unknown m/z 58 _ RI 1104 1 5.69 5.67 5.73 0.00 5.76 2 5.65 5.87 5.60 5.61 5.94 3 5.49 5.81 5.66 0.00 5.60 Unknown m/z 83 _ RI 1145 1 0.00 5.05 0.00 0.00 0.00 2 0.00 5.10 4.51 0.00 0.00 3 0.00 5.16 4.90 0.00 0.00 Unknown m/z 58 _ RI 1554 1 0.00 0.00 4.92 0.00 0. 00 2 0.00 5.02 5.10 0.00 0.00 3 0.00 0.00 4.97 0.00 0.00 80 Table A.3.3 Bacterial concentrations of F. tularensis SCHU S4 in modified Mueller - Hinton media and B. anthracis Ames in Brain - Heart Infusion media measured at the VOC sampling time points 6 h and 24 h post - inoculation of cultures. The numbers represent the mean of CFU/mL determined from 2 plate counts for e ach of the 3 culture replicates (total number of plate counts for each time point = 6), the errors represent standard deviation. 6 h 24 h Ft SCHU S4 (6.7 ± 1.6 ) * 10 5 (6.1 ± 2.9) * 10 5 Ba Ames (6.7 ± 1.4) * 10 6 (2.5 ± 0.4) * 10 7 81 APPENDIX C: Protocol Solid - Phase Microextraction (SPME) Sampling for Volatile Organic Compounds (VOCs) of Liquid Bacterial Cultures Lawrence Liver more Nation al Laboratory Protocol: Rasley - 2018 - 001 Facility/Program: SCA_BSL - 2_BSL - 3 Global Security Effective Date: 09/26/2018 Level of Use: Continuous Use LLNL - MI - 791181 This document was prepared as an account of work sponsored by an agency of the Unit ed States g overnment. Neither the United States government nor Lawrence Livermore National Security, LLC, nor any of their employees makes any warranty, expressed or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulne ss of any information, apparatus, product, or process disclosed, or represents that its us e would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacture r, or other wise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States government or Lawrence Livermore National Security, LLC. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States government or Lawrence Livermore National Security, LLC, and sh all not be used for advertising or product endorsement purposes. 82 Table A.3.P.1 Protocol Document Revision History Revision # Date Author Summary of Ch anges Init ial 07 - 03 - 2018 Amy Rasley This SOP was developed to establish the procedure for SPME fiber head - space sampling of liquid bacterial cultures without contamination of the SPME fibers with viable bacteria. Rev 1 9/17/2018 Amy Rasley Revised to inc lude non - co ntamination test results when SPME fibers are exposed to fully virulent Bacillus anth racis . Rev 2 9/26/2018 Amy Rasley Revised to include non - contamination test results when SPME fibers are exposed to fully virulent Francisella tularensis . 1. Purpose Th e purpose of this procedure is to demonstrate that solid - phase microextraction (SPME) fibers are not contaminated with viable agent when exposed to the head space of liquid bacterial cultures during sampling for volatile organic compounds (VOCs) prior to r emoval of SPME fiber sampling devices to a lower physical containment level [e.g., lo wer biosafety level (BSL), non - BSL laboratory] for GS/MS analysis. 2. Scope This procedure applies to and has been validated for head space sampling for VOC s using SPM E fiber VOC collection devices with the following agents: Bacillus anthracis Sterne (BSL - 2) Francisella tularensis subsp. novicida (BSL - 2) Bacillus anthracis Ames (BSL - 3) Francisella tularensis SCHU S4 (BSL - 3) 83 Note : This protocol has been d eveloped an d validated using attenuated, select agent - exempt strains in a non - registere d BSL - 2 laboratory for the purpose of demonstrating that SPME fibers used for head space sampling of VOCs from liquid bacterial cultures are not contaminated with viable agent duri ng the sampling process. Prior to removing SPME fibers from any registered BSL - 2 or BSL - 3 BSAT laboratory, this procedure shall be validated on the fully virulent agents and the SOP updated and approved by the Biosafety Officer (BSO) and Respo nsible Offi cial (RO). As of the current revision, this protocol has been validated f or use on the fully virulent Bacillus anthracis and Francisella tularensis . Validation results indicated that there is no potential for SPME fiber contamination during the samplin g process. The current SOP has been updated and approved by the Biosafety O fficer (BSO) and Responsible Official (RO). 3. Responsibility Research Personnel Research personnel who will be performing head space sampling using SPME fiber collect ion devices with any of the above listed agents for which this procedure has been validated must be familiar with and adhere to the exact procedures outlined in this SOP. Any deviation from the written procedure will require re - review and re - validation of the proced ure. 84 All research personnel who will be performing this procedure are required to read, understand and review this procedure with the appropriate PI/RI who developed the procedure and must sign the Training Record in Appendix A prior to perfo rming this procedure. Principal Investigator (PI)/Responsible Individual (RI) The PI/RI of the protocol is responsible for submitting the written procedure to the RO an d BSO for review and approval prior to initiating validation of the procedure. In add ition, the PI/RI is responsible for submitting validation data to the RO for review and final approval of the procedure prior to initiating experimental work utilizing th is procedure. The PI/RI is also responsible for training and reviewing this procedur e with all personnel who will be performing the validated inactivation procedure. 4. Procedure A. Preparation of Bacterial Cultures for SPME fiber head - space sampling of VOCs i. Using a sterile, disposable loop, transfer a loopful of frozen glycerol stock o f the bacte rium onto an appropriate agar medium. ii. Streak for isolation and incubate (static incubator) at 37°C for 24 - 28 hours until viable growth of isolated co lonies is observed on the plate. iii. Inoculate 10 ml of broth media with 1 - 5 colonies of bacteria an d incubate with shaking (170 rpm) overnight at 37°C. iv. Subculture bacteria into 25 - 50 mL of broth media based on desired optical density (OD) in disposable 250 mL Erlenmeyer flasks with vented caps. 1. Control Flasks: 85 Empty Flask (no media) Media Only Flask (n o bacteria) v. Incubate flasks with shaking (170 rpm) at 37°C for various time points (varies by experiment). Maximum bacteria concentrations estimated to not exc eed 10^9 colony - forming units (CFUs)/mL. B. Head Space Sampling using SPME fibers i. At each specific time point , remove the culture flasks from the incubator and transfer to the biological safety cabinet (BSC). 1. Allow the culture flasks to sit for a minimum of 30 minutes prior to head space sampling to allow for any potential aerosols generated during th e incubatio n phase to settle. ii. After 30 minutes, insert the SPME fiber (Supelco Portable Field Sampler, Product #57359 - U) through the vent in the cap into the he ad space above the liquid bacterial culture (see Protocol Figure A.3.P.1). Sampling times vary by experime nt. Distance between the fiber tip and liquid culture is approximately 3 inches, minimizing the risk of bacteria transfer. 86 Figure A.3.P.1 SPME f iber sampling of bacterial culture headspace iii. Once head space sampling is complete, retract the SPME fiber back into the sampling device. 1. Surface decontaminate the exterior of the sampling device prior to removing the device from the BSC. 2. Place the device in a secondary container for transport out of the laboratory. Surface decontaminate the ex terior of t he secondary container prior to r emoval from the laboratory. Note: Refer to Protocol Appendix A.3.P.B for procedures and data used to validate that this head space sampling procedure does not result in contamination of the SPME fiber with viable agent. 5. Change Control Revisions to this protocol will be made as necessary. This procedure will be reviewed, and revised as necessary, by the PI/RI, LLNL Responsible Official (RO) and BSO, at least annually or after any change in PI/RI, after any ch ange in the validated proced ure and after any failure of 87 the procedure. The PI/RI is responsible for communicating revisions to the appropriate research staff and training the staff on the changes to the procedure. 6. Responsible Individual The PI/RI who authore d the procedure i s responsible for change control. Protocol Appendix A.3.P.A. Training Record All research personnel who will be performing this procedure are required to read, understand and review this procedure with the appropriate PI/RI w ho develope d the procedure a nd must sign this Training Record prior to performing this procedure and again after any revision to this procedure. The PI/RI must submit an electronic copy of the completed training record(s) to the RO who will maintain fil e copies in the Laboratory S elect Agent Program Folder on UCM. I have read, understood and agree to fully comply with and adhere to the exact procedure, as written in SOP # Rasley - 2018 - 001 , Revision 2 , Solid - Phase Microextraction (SPME) Sampling for Volatile O rganic Compounds ( VOCs) of Liquid Bacterial Cultures . Name (print) Signature Date 88 Protocol Appendix A.3.P.B. Method Validation To demonstrate that SPME fibers are not contaminated with viable agent during headspace sampling o f liquid ba cterial cultures, viability testing was performed on the SPME fibers post - headspace sampling. Viability testing was performed one time using 3 bio logical replicates for each organism. A. Viability Testing Method 1. SPME fibers exposed to sampling h eadspace ab ove liquid bacterial cultures were submerged and agitated in 1 mL of an appropriate growth media for 30 seconds (see Protocol Figure A.3.P.2) and t hen removed. Brain Heart Infusion (BHI) B. anthracis Sterne Modified Mueller Hinton F. tularen sis subsp. novicida 2. Media exposed to SPME fibers was incubated (static) for 48 hours at 37°C ±2°C. 3. Controls: i. Broth only (no bacteria) negative control ii. 100 µL of inoculum from Replicate #1 in 1 mL media positive control 4. After 48 hours, media was observ ed for grow th (turbidity) and results were recorded. 5. 100% of the media exposed to the SPME fiber was then tra nsferred to the appropriate agar plates (100 µL x 10 plates). 6. Agar plates were incubated (static) for a minimum of 48 hours at 37°C ±2°C. 7. After 48 hours, plat es were observed for growth. 89 Figure A.3.P.2 Submerging SPME fiber exposed to headspace above liquid bacterial culture for viability testing. B. R esults: No growth was observed in either the liquid culture exposed to the SPME fibers used to sa mple the he adspace of the tested bacterial cultures for VOCs or when 100% of the exposed media was plated ont o agar plates. The negative control (media only, no bacteria) also showed no growth, whereas the positive control (inoculum from replicate #1) sho wed expecte d growth indicating that the media uses was capable of supporting the growth of the tested bacteri a. These results demonstrate that there is no potential for contamination of the SPME fibers when exposed to the head sampling space of a bacteria l liquid cu lture when following this protocol, as written. 90 REFERENCES 91 REFERENCES 1 Reese, K. L., Avila, J. R., Rasley, A., Jones, A. D. & Frank, M. Metabolic profiling of volatile organic compounds (VOCs) emitted by the pathogens Francisella tularensis and Bacillus anthracis in liquid culture. Sci. Rep. , doi:10.1038/s 41598 - 020 - 66136 - 0 (2020). 2 Amann, A. et al. The human volatilome: volatile organic compounds (VOCs) in exhaled breath, skin emanations, urine, feces and sal iva. J. Br eath Res. 8 , 034001, doi:10.1088/1752 - 7155/8/3/034001 (2014). 3 Lawal, O., Ahmed, W. M., Nijsen, T. M. E., Goodacre, R. & Fowler, S. J. Exhaled breath analysis: a review of 'breath - taking' methods for off - line analysis. Metabolomics 13 , 110, doi: 10.1007/s1 1306 - 017 - 1241 - 8 (2017). 4 Bruderer, T. et al. On - line analysis of exhaled breath focus re view. Chem. Rev. 119 , 10803 - 10828, doi:10.1021/acs.chemrev.9b00005 (2019). 5 Das, S. & Pal, M. Review Non - invasive monitoring of human health by exhaled brea th analysi s: A comprehensive review. J. Electrochem. Soc. 167 , 037562, doi:10.1149/1945 - 7111/ab67a6 (2020). 6 Pleil, J. D. Cellular respiration: replicating in vivo systems biology for in vitro exploration of human exposome, microbiome, and disease pathoge nesis biom arkers. J. Breath Res. 10 , 010201, doi:10.1088/1752 - 7155/10/1/010201 (2016). 7 Winters, B . R. et al. Review: Endogenously produced volatiles for in vitro toxicity testing using cell lines. Appl. In Vitro Toxicol. 4 , 129 - 138, doi:10.1089/aivt.2017 .0038 (201 8). 8 Audrain, B., Farag, M. A., Ryu, C. M. & Ghigo, J. M. Role of bacterial volatile compounds in bacterial biology. FEMS Microbiol. Rev. 39 , 222 - 233, doi:10.1093/femsre/fuu013 (2015). 9 Filipiak, W. et al. Molecular analysis of volatile metabol ites relea sed sp ecifically by staphylococcus aureus and pseudomonas aeruginosa. BMC Microbiol. 12 , 113, doi:10.1186/1471 - 2180 - 12 - 113 (2012). 10 Rees, C. A., Smolinska, A. & Hill, J. E. Volatile metabolic diversity of Klebsiella pneumoniae in nutrient - reple te conditi ons. M etabolomics 13 (2017). 11 Rees, C. A., Burklund, A., Stefanuto, P. - H., Schwartzman, J. D. & Hill, J. E. Comprehensive volatile metabolic fingerprinting of bacterial and fungal pathogen groups. J. Breath Res. 12 , 026001 (2018). 12 Chen, J., Tang, J., Shi, H., Tang, C. & Zhang, R. Ch aracteristics of volatile organic compounds produced from five pathogenic bacteria by headspace - solid phase micro - 92 extraction/gas chromatography - mass spectrometry. J. Basic Microbiol. 57 , 228 - 237, doi:10.1002/jobm.2 01600505 ( 2017). 13 Horsmon, J. R. & Crous e, K. Characterization of volatile organic compound profiles of bacterial threat agents. US Army Edgewood Chemical Biological Center (2008). 14 Lonsdale, C. L. et al. The use of colorimetric sensor arrays to discri minate bet ween pathogenic bacteria. PLOS O NE 8 , e62726, doi:10.1371/journal.pone.0062726 (2013). 15 U.S. Centers for Disease Control and Prevention. Select Agents and Toxins List , (2018). 16 McL endon, M. K., Apicella, M. A. & Allen, L. - A. H. Francisella tularensis: taxonomy, genetics, and Immunopathogenesis of a potential agent of biowarfare. Annu. Rev. Microbiol. 60 , 167 - 185, doi:10.1146/annurev.micro.60.080805.142126 (2006). 17 Goel, A. K. Anth rax: A dis ease of biowarfare and public he alth importance. World J. Clin. Cases. 3 , 20 - 33, doi:10.12998/wjcc.v3.i1.20 (2015). 18 Hendricker, A. D., Abbas - Hawks, C., Basile, F., Voorhees, K. J. & Hadfield, T. L. Rapid chemotaxonomy of pathogenic bacteria us ing in sit u thermal hydrolysis and methyla tion as a sample preparation step coupled with a field portable membrane - inlet quadrupole ion trap mass spectrometer. Int. J. Mass Spectrom. 190 - 191 , 331 - 342, doi:10.1016/S1387 - 3806(98)14266 - 5 (1999). 19 Basile, F. , Beverly, M. B., Voorhees, K. J. & Hadfie ld, T. L. Pathogenic bacteria: their detection and differentiation by rapid lipid profiling with pyrolysis mass spectrometry. Trends Anal. Chem. 17 , 95 - 109, doi:10.1016/S0165 - 9936(97)00103 - 9 (1998). 20 Li, Y. et al . Differen tiation of bacteria using fatty acid profiles from gas chromatography tandem mass spectrometry. J. Sci. Food Agric. 90 , 1380 - 1383, doi:doi:10.1002/jsfa.3931 (2010). 21 Biosafety in Microbiological and Biomedical Laboratories . 5th edn, (U.S. Depa rtment of Health and Human Services, Publi c Health Service, Centers for Disease Control and Prevention, National Institutes of Health, HHS Publication No. (CDC) 21 - 1112 Revised December 2009; https://www.cdc.gov/labs/BMBL.html). 22 WHO. Laboratory Biosafet y Manual . 3rd edn, (World Health Organiza tion, 2004; https://www.who.int/csr/resources/publications/biosafety/Biosafety7.pdf ). 23 El - Etr, S. H. et al. Francisella tularensis Type A strains cause the rapid encystment of Acanthamoeba castellanii and surviv e in amoeb al cysts for three weeks postinf ection. Appl. Environ. Microbiol. 75 , 7488 - 7500, doi:10.1128/AEM.01829 - 09 (2009). 24 Reese, K. L. et al. Chemical profiling of volatile organic compounds in the headspace of algal cultures as early biomarkers of al gal pond c rashes. Sci. Rep. 9 , 13866, doi: 10.1038/s41598 - 019 - 50125 - z (2019). 93 25 Sumner, L. W. et al. Proposed minimum reporting standards for chemical analysis Chemical Analysis Working Group (CAWG) Metabolomics Standards Initiative (MSI). Metabolomics. 3 , 211 - 221, doi:10.1007/s11306 - 007 - 0082 - 2 (2007). 26 Forney, F. W. & Markovetz, A. J. The biology of m ethyl ketones. J. Lipid Res. 12 , 383 - 395 (1971). 27 Schulz, S. & Dickschat, J. S. Bacterial volatiles: the smell of small organisms. Nat. Prod. Rep. 24 , 814 - 842, doi: 10.1039/b507392h (2007). 28 Qiu, X. et al. Crystal structure and substrate specificity of - ketoacyl - acyl carrier protein synthase III (FabH) from Staphylococcus aureus. Protein Sci. 14 , 2087 - 2094, doi:10.1110/ps.051501605 (2005). 29 Schmidt, R., Cordo vez, V., de Boer, W., Raaijmakers, J. & Garbeva, P. Volatile affairs in microbial interact ions. ISME J. 9 , 2329 - 2335, doi:10.1038/ismej.2015.42 (2015). 30 Elgaali, H. et al. Comparison of long - chain alcohols and other volatile compounds emitted f rom food - b orne and related Gram positive and Gram negative bacteria. J. Basic Microbiol. 42 , 373 - 380 , doi:10.1002/1521 - 4028(200212)42:6<373::AID - JOBM373>3.0.CO;2 - 4 (2002). 31 Togashi, N. et al. Antibacterial activity of long - chain fatty alcohols against St aphylococc us aureus. Molecules. 12 , 139 (2007). 32 Hwang, H. - I., Hartman, T. G., Rosen, R. T., Lech, J. & Ho, C. - T. Formation of pyrazines from the Maillard reaction of glucose and lysine - .alpha. - amine - 15N. J. Agric. Food Chem. 42 , 1000 - 1004, doi:10.1021/j f00040a031 (1994). 33 Brana, A. F. et al. Activation and silencing of secondary metabolites in Strep tomyces albus and Streptomyces lividans after transformation with cosmids containing the thienamycin gene cluster from Streptomyces cattleya. Arch. Microbio l. 196 , 34 5 - 355, doi:10.1007/s00203 - 014 - 0977 - z (2014). 34 Larroche, C., Besson, I. & Gros, J. - B. Hig h pyrazine production by Bacillus subtilis in solid substrate fermentation on ground soybeans. Process Biochem. 34 , 667 - 674, doi:10.1016/S0032 - 9592(98)00141 - 1 (1999). 35 Turnbull, P. C. Definitive identification of Bacillus anthracis -- a review. J. Appl. Mi crobiol. 87 , 237 - 240, doi:10.1046/j.1365 - 2672.1999.00876.x (1999). 36 Rohmer, L. et al. Comparison of Francisella tularensis genomes reveals evolutionary ev ents assoc iated with the emergence of human pathogenic strains. Genome Biol. 8 , R102, doi:10.1186/gb - 2007 - 8 - 6 - r102 (2007). 37 Pleil, J. D. & Williams, A. Centralized resource for chemicals from the human volatilome in an interactive open - sourced database. J. Breath Res. 13 , 040201, doi:10.1088/1752 - 7163/ab2fa2 (2019). 94 38 U .S. Environmental Protection Age ncy. Volatilome: Human breath. < https://comptox.epa.gov/dashboard/chemical_lists /VOLATILOM E > (2019). 95 CHAPTER 4. METABOLIC PROFILING OF VOCs IN THE HEADSPACE OF ALGAL CULTURES AS EARLY BIOMARKERS OF ALGAL POND CRASHES FOREW O RD The material presented in this chapter has been adapted from work first published in 2019 in the jour nal Scient ific Reports [1] . Contri butions from others to conduct the experiments described in this chapter are as follows : C. L. Fisher, P. D. Lane, and J. D. Jaryenneh were integral in set ting up and maintaining the experimental cultures and collect ing algal density data . 4.1 INTRODUCTIO N As the energy needs of the world increase, dependence on non - renewable sources of energy remains a concern. Increased production of corn starch or sugarcane - based ethanol has resulted in increased atmospheric carbon dioxide levels, diversion of arable land from food production, and increased consumer cost for sugar and corn [2] . For these reasons, microalgae production systems are considered a promising avenue for biofuel production. Microalgal strains are capable of growth in a range of environme nts (e.g. , freshwater, marine, hypersaline, highly acidic) including high - nutrient municipal wastewater systems [3] , allowing for simultaneous - potable (brackish or marine) water sources us i ng non - ar able land, combined with their high capacity for fixation of atmospheric carbon dioxide and high lipid - to - biomass ratios are significant advantages toward its use as a biofuel feedstock. The development of optimized systems for sustainable and de p endable b iofuel production through algal pond systems are necessary as global energy strategies continue to evolve (for review, see Kayitar 2017 [4] ). 96 A major challenge faced in algal production is often unpredictable losses of entire crops due to parasitism, grazing, weather, or many o t her facto rs. Closed photobioreactors are less likely to be susceptible to contamination with deleterious species (e.g. , viruses, fungi, protozoans, detrimental microbes), but involve higher capital costs [5] and, once contaminated, can be challenging to disinfect. Open algal ponds are less expensive to set up but are more likely to succ u mb to cra shes (Figure 4. 1A) caused by grazing or parasitism [6,7] . Notably, a single adult marine rotifer, Brachionus plicatilis (Figure 4. 1B), can consume 200 microalgal cells per minute and double in population w ithin 1 - 2 days [8] . It has been estimated that pond crashes account for 30 % loss of annualized algal production [5] and represent a significant economic barrier to biofuel production [9] . However, algae are curr ently more costly to produce per unit due to insufficient rates of biomass production and harvesting costs [10] . Basic research that drives increases in algal cell culture density, and thus total lipid accumulation, is an initial step to address these issues. To produce high er yields of algae crops, research is needed to (1) have better early - warning tools to anticipate and/or diagnose the presence of predators and (2) understand algae - bacteria interactions, as they frequently happen in nature. The study of potential VOC biom arkers for (1) is covered in this chapter and for (2) in chapter 5. Current pond crash mitigation strategies, both prophylactic and interdictive, are largely focused on chemical treatments, such as hypochlorite [11 ] , copper [12,13] , quinine sulfate [14] , rotenone [15] , additives that lower pH (to less than 3.0) [16] , and biocides, such as tossendanin [17] . The use of chemical countermeasures as a prophylactic strategy to prevent pond crashes is prohibitively expensive for most algal industry business models. Chemical additives can degrade 97 Figure 4. 1 ( a ) Adapted from McBride et al, showing 350 L open algal production ponds with a healthy algal pond on the left compared with a crashed algal pond on the right . ( b ) Brachionus plicatilis (average length 160 µm), marine rotifer, in a field of algae , Microchloropsis salina . from sun exposure and may need to be repeatedly added to cultures to maintain protection. In addition, frequent and repeated chemic al applica tion can be environmentally detrimental through the development of resistant pest species or through unacceptable off - target effects (e.g., pisicidal effects of rotenone). However, when applied early and in a targeted fashion after the detection of a delet erious species, chemical additives can be highly effective at saving algal cultures [for review of crop protection strategies, see Fisher et al. 2019 [18] ]. In order to attain the production levels of 25 g m - 2 d - 1 ash free dry weight (or 2,500 gallons of biofuel per acre per year) [19] deemed necessary for economic algal biofuel production, cost - effective pond monitoring strategies are necessary to reduce culture loss and increase annualized production. Currently, algal production facilities utilize light microscopy to identify contaminants, pathogens, and competing algal strains that could lead to the demise of the desired algal strain [20] . However, microscopy is slow, labor - intensive, and requires advanced operator training for differentiating various microbiota. Alternative methods involving automated and semi - automated technologies, such as FlowCAM i maging flow - cytometry [21,22] , 98 polymerase chain reaction, and hybridization - based ass ays [23,24] are under development to increase sensitivity and expedite analysis for daily algal culture monitoring. All of th e strategies described above require sampling of liquid from the algal culture, but an alternative appr oach involves sampling volatile molecular indicators from the air above algal cultures. Volatile organic compounds (VOCs) are carbon - containin g molecule s with high vapor pressures at ambient temperatures [25] that often occur in rich mixtures . Within the field of chemical ecology, VOCs have been identified as secondary metabolites and include, but are not limited to, pheromones, semiochemicals, o dorants, and phytohormone s [26,27] . Algal VOC production has been associated with intra - and inter - species communication, alle lopathy, s emiochemical production, and predator deterrence [for review, see Zuo 2019 [28] ]. A well characterized example of an algal volatile involves conversion of nonvolatile dimethylsulfoniopropionate (DMSP) to volatile dimethyl sulfide ( DMS). In i ntact Emiliania huxleyi cells, conversion of DMSP to DMS by the enzyme DMSP lyase is minimal , theorized to be segregated within the intact cell . However, during algal grazing, such as by the dinoflagellate Oxyrrhis marina , E. huxleyi is disrupted , releasin g DMSP. Once released , conversion of DMSP to DMS [29, 30] is catalyzed by DMSP lyases including those from bacteria . DMS acts not only as a deterre nt against herbivory by Oxyrrhis marina [31] , but additionally as an attractant for other species such as birds and fish [ 32] . The aims of this study were to 1) develop a methodology to detect VOCs from healthy algal cultures ( Microchloropsis salina) as well as algal c ultures in the presence of a grazer ( M. salina cultures with marine rotifer Brachionus plicatilis) and 2) evaluate whether specific VOCs could serve as early indicators of an imminent culture crash. A setup based upon solid - phase microextraction (SPME) fib ers couple d with gas chromatography - mass spectr ometry (GC - MS) 99 allowed for non - invasive monitoring of volatile emissions. Compounds present during the active grazing period of rotifers on algal cultures, but not produced in healthy controls, were deemed pot ential bio markers of high stress conditions. We propose that these biomarker compounds are potential diagnostic tools for chemical monitoring systems in algae cultivation s ystems to enabling the early detection of culture stress for improved algal crop pro duction . 4.2 METHODS The experimental setup (F igure 4. 2) facilitated headspace volatile monitoring of M. salina with and without the algal grazer and rotifer , B. plicatilis . In total, three replicate experiments were performed using this setup, labeled Ex periment 1 , Experiment 2, and Experiment 3. A t duplicates of every culture type. A lgal cell concentrations and VOC headspace samples were collected at various time points for M. salin a alone (a bbreviated Algae or A) , M. salina and B. plicatilis (abbreviated Algae + Rotifer or A+R ), and ESAW media blanks (abbreviated Media Blank or MB ) . 4.2.1 Axenic algae culture s Microchloropsis salina (CCMP 1776) was obtained as an axenic stock cult ure (as de termined by the supplier) from the National Center for Marine Algae and Microbiota (NCMA at Bigelow Laboratory, ME, USA). M. salina cultures were grown as previously described in Fisher et al, 2019 [33] . For volatilomics experiments, Enriche d Seawater , Artificial Water ( ESAW ) media was modified to contain 7.5 mM NaNO 3 and 0.5 mM Na 3 PO 4 . C ultures were 100 Figure 4.2 Schematic of experimental setup for growth M. salina (Algae, A) in the presence of B. plicatilis (Rotifer, R) for 5 d ays . Mass flo w controll ers (MFCs) mixed 1% CO2 with VOC - free air to sparge 15 L cultures at 150 cc min - 1. In total, three replic ate experiments (Experiment 1, Experiment 2, Experiment 3) were performed using this setup. SPME fibers (Experiment 1 used 1 fiber; Experimen t 2,3 used 2 fibers) were used to sample the headspace of media blank (MB), Algae only (A), and Algae + Rotifer (A +R) carboys for 30 - 60 min each at various timepoints over 2 - 4 d ays . grown in 15L of media in 20 - L polycarbonate carboys at room temperature (RT) of ~ 22°C with 24 - h light intensity of ~200 0 µmol m - 2 s - 1 for 5 d ays (d) Carbon dioxide gas, research purity 99.999 % (Matheson Tri - Gas, NJ, USA), and research grade air (70:30 N 2 /O 2 ), VOC Free (Matheson Tri - Gas, NJ, USA) , were supplied to all carboys via two ma ss flow controllers (one for CO 2 and one for air). The two mass fl ow controllers (Alicat, AZ, USA) were set to deliver 1% CO 2 (9.00 cc min - 1 ) and 99 % air (891 cc min - 1 ), for a total mass flow of 900 cc min - 1 split 101 equally across six carboys (150 cc min - 1 s parging rate for each sample through an air stone bubbler ). 4.2.2 Xenic marine rotifers Lots of 10 - 15 x 10 6 live, xenic, marine rotifers, Brachionus plicatilis , were obtained from Reed Mariculture, CA, USA 1 - 2 days before each inoculation and w ere shippe d overnight on ice. Upon arr ival, B. plicatilis were kept at 4 °C until concentrated and inoculated into algal culture s for each experiment. On the day of inoculation, B. plicatilis were allowed to warm for 1 - 3 h to room temperature ( 22 °C) , gen tly concen trated using a 53 µm screen filter (Florida Aqua Farms, FL, USA) down to 100 mL of culture , and rinsed twice with 200 mL of ESAW media. Rotifers were enumerated by direct counting using a Rafter counter. 4.2.3 Preparation of cultures Experimenta l cultures were grown in in 20 - L polyc arbonate carboys (ThermoFisher Scientific, MA USA) containing 15 L nutrient enriched ESAW medium . At the start of an experiment, 4 of the 6 carboys were inoculated with M. salina culture to a concentration of 4 - 5 x 1 0 6 cells m L - 1 in 15 L. After 48 hours (h) of M. salina growth and acclimation to culturing conditions in the four carboys , 1.32 x 10 6 live rotifers (final concentration of 88 rotifers mL - 1 ) were added to two of the four algal cultures . 4.2.4 Monitoring a lgae growt h and rotifer cultures Algal culture density was determined daily by chlorophyll fluorescence (430 nm excitation, 685 nm emission) using a Tecan i - control infinite 200 PRO , version 1.11.1.0 (Tecan 102 Group, Switzerland) . Monitoring growth of algal cult ures w as derived by calculation via a standard curve correlating chlorophyll fluorescence with algal density. Duplicate fluorescence measurements for each sample were averaged for each timepoint and then normalized to the final concentration measuremen ts f or the M. salina control in the absence of rotifers. Health and viability of rotifers within algal cultures was monitored daily via light microscopy. Significant differences between means of healthy or infected algal cultures were compared using two - wa y AN OVAs w 4.2.5 SPME headspace sampling and GC - MS data acquisition VOCs were sampled from the headspaces of each culture and media control vessel using portable field sampler SPME fibers, with 65 µm polydimethylsiloxane/divinyl - benze ne ( PDMS/D VB) coatings (Supelco, Bellefonte, PA). One fiber was used per carboy in Experiment 1, totaling 6 measurements per timepoint, while two fibers were used per carboy in Experiments 2 and 3, totaling 12 replicate measurements for each time point. Fo r th is wor k, we required a the time course of the experiments. The bi - phasic coating (one of three commercially available field - portable options) was chosen for sampli ng a wide range of compounds, including polar analytes, semi - volatiles, and larger weight volatiles. SPME samples were obtained within 1 - 2 h of the fluorescence measurements that were used to determine algae concentrations. SPME exposure times we re s horter for Experiment 1 (30 min) compared to Experiments 2 and 3 (60 min). SPME fibers were stored in sealed containers at 2 - 4 °C after sampling. Unexposed SPME potential extraneous volatiles arising fro m st orage conditions. Samples were analyzed by GC - MS within 2 weeks of collection. 103 An untargeted GC - MS approach was used to analyze the collected VOCs with an Agilent 5975T GC - MSD (Agilent Technologies, Santa Clara, CA) using an Agilent HP - 5ms column (30 m x 250 µm x 0.25 µm) coupled to a single quadrupole mass analyzer with helium carrier gas at a constant flow rate of 1.2 mL/min. VOCs absorbed in the SPME fiber were desorbed in the heated GC inlet (280 °C) for 15 seconds using splitless injection. The co lumn tempe rature was programmed, starting at 40 °C for 3 min, ramped at 5 °C/minute from 40 to 150 °C, ramped at 15 °C/min from 150 to 280 °C and then held for 2 min. The total run time was 35.67 min. Ions were generated using electron ionization (70 eV) a nd s pectra were acquired at 4 scans/s over m/z 35 - 450. Data acquisition was performed under control of ChemStation software (Agilent Technologies, version E.02.02). A commercial reference of 18 standard compounds (S - 22329; AccuStandard, New Haven, CT) was used to ev aluate day - to - day performance and to calculate retention indices. 4.2.6 GC - MS data processing After GC - MS data acquisition, data processing procedures and criteria were applied to detect and identify individual biomarkers in each condition. All Che mStati on data files (consisting of biological duplicates, media controls, and unexposed fibers) were translated for Chromatographic deconvolution and visualization were perfo rmed using MassHunter Qualitative (version B.07.00 SP2) using a Retention Time window size factor of 90.0, signal - to - noise ratio detection (threshold of detect ion 5 x 10 3 counts per peak). An arbitrary small value of 1 was assigned to the signal value for compounds that were not detected. 104 Detected peaks were transferred into Mass Profiler Professional (MPP) 12.6.1 software and aligned across all samples in the d ata set using a retention time tolerance of 0.15 minutes, mass spectral match factor of 0.6 (of maximum 1.0), and a delta m/z tolerance of 0.2 Da. Putative identification of the aligned compounds was performed by searching spectra against the Nationa l Inst itute of Standards and Technology (NIST) mass spectral database, NIST14. Compounds with did not exceed the mass spectral match threshold were annotated using th e base peak m/z and retention index (e.g. , m/z Two criteria were used to identify volatile biomarkers unique to the Algae or Algae + Rotifer conditions: (1) detection of the biomarker in at least three of the four replicates at each sample d timepoint and (2) a) the biomarker was present in the Algae or Algae + Rotifer condition and not detected in the media blank or travel blank conditions; OR b) the biomarker was present in the Algae or Algae + Rotifer conditions at an abundance greater th an 10x the abundance in the media blank or travel blank. The pe ak areas of potential biomarkers passing the filter criteria were compared across the three performed experiments, with regards to both individual biomarkers and groups of biomarkers belonging to the same compound class. The presence or absence of these bi omarkers in each experiment was determined, and the calculated peak areas were compared to algal density measurements. 4.3 RESULTS 4.3.1 Cell counts of infected and control cultures Measure d algae concentrations as a function of time for all three experiments is presented in Figure 4.3. At 48 h after inoculation, algal concentrations across all cultures were similar, 105 Figure 4.3 Algae concentration as determined by fluorescence meas urement s collected for three experiments. Similar coloring and patterns represent biological replicates of each condition: media blanks (MB), Algae (A), and Algae + Rotifer (A+R) cultures. Small fluorescence signals were observed in MB controls, most of wh ich are not discernable on this scale. Error bars represent standard deviation derived from duplicate measurements for each sample. Significance levels for conditions that exhibited p<0.05 are in Appendix Table A.4.1. Blue asterisks (*) indicate the time p oints f or headspace VOC sampling by SPME fibers. approaching the mid to late stages of logarithmic growth. At this time, B. plicatilis were added to two of the four M. salina cultures, resulting in time - dependent decreases in algal density relative to the axenic cultures (Figure 4. 3). Despite consistent growth conditions, 96 h after the initial cultures were started and 48 h after rotifers were added, the Algae + Rotifer cultu res displayed different extents of algal biomass loss attributed to rotifer grazing (see Figure 4. 3). This variation in rates of biomass loss may arise from differences in rotifer lots. 4.3.2 Headspace VOC results Qualitative and quantitative differences were observed in the VOC profiles of Algae + Rotifer cultures compared to the Algae cul tures. Example total ion chromatograms for Algae and Algae + Rotifer cultures taken approximately 24 h after addition of rotifers (Experiment #3) are 106 Figure 4.4 Example GC - MS chromatograms for observed VOCs sampled from Algae (A) and Algae + Rotifer (A+ R) cultures between 16 and 24 min, ( a ) Total ion chromatogram with indicated VOCs (Annotations See Table 4.1), ( b - d ) extracted ion chromatograms monitoring increase in compound 6 over time ( m/z 177, RT 23.46 min, RI 1495). shown in Figure 4. 4 a . Several VO Cs that differentiate the two culture conditions are indicated ( Figure 4. 4 a , annotations in Table 4. 1) and are potential early indicators of algal grazing or death. Extracted ion chromatograms were utilized to improve visualization of individual VOCs , as s hown in Figures 4. 4 b - d . These demonstrate the increase in an VOC displaying a base peak m/z 177 and retention index 1495, observed over the time course of the experiment in Algae + Rotifer cultures. Although the Algae chromatogram for m/z 177 also displays a small peak at the same retention time, this VOC was not detected using the given filteri ng criteria for data processing. 107 Table 4. 1 VOCs robustly and repeatedly detected from Algae and Algae + Rotifer experiments Compound # Mass Tentative Compound Class* NIST14 ID NIST % Match Experimental Retention Index Theoretical Retention Index Experiment # 1 2 3 VOCs detected in A+R cultures 1 82 Carotenoid 2,2,6 - trimethylcyclohexanone 79 1021 1036 X X 2 107 1181 X X X 3 121 Phenol 1191 X X 4 137 Caro tenoid 2,6,6 - trimethyl - 1 - cyclohexene - 1 - carboxaldehyde 81 1209 1220 X X X 5 121 Carotenoid 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 2 - butanone 76 1419 1433 X X X 6 177 Carotenoid trans - - ionone 94 1495 1486 X X X 7 57 Alkane 1691 X X X VOC s detected in A+R and A cultures 8 71 1039 X X 9 96 Methyl Ester 3 - Nonenoic acid, methyl ester 74 1134 1191 X X X 10 341 1139 X X 11 71 1293 X X 12 138 Terpene/Carotenoid 1338 X X 13 73 Fatty acid ( Hexade canoic acid ) 19 83 X X 14 192 2197 X X * Tentative compound class for unknown compounds is based on fragmentation in averaged mass spectra determined via chromatograp hic deconvolution and alignment. 108 The number of compounds detected from deconvolution of chromatographic peaks varied with each sample. The analysis of a single sample typically detected 100 - 200 chemical compounds, many of which were attributed to background ( present also in control measu rements) or not found reproducibly . Application of chromatographic peak alignment across the data from all samples and at every timepoint generated a list of more than 1800 compounds, consisting of both algal VOCs and extraneous signals from the experiment al setup. Application of the filtering criteria based upon algal abundance and detection frequency across experimental replica tes identified the most robust compounds as potential VOC biomarkers from either Algae or Algae + Rotifer cultures, removed irrep roducible compounds, and narrowed the extensive list to ~ 50 compounds in any single experiment . Table 4.1 shows only biomarker s that were observed across multiple experiments. For a detailed list of the volatile biomarkers detected in each individual exp eriment, refer to Appendix Table A.4.2. Analysis of VOC data from Experiments 1 - 3 revealed several VOCs that were reproducibly observed in 1) Algae + Rotifer cultures and 2) both Algae and Algae + Rotifer cultures, as shown in Table 4. 1, despite the differ ent rates in algal biomass loss. For example, Compound 6 monitored in Figure 4. 4B - D was identified with a 94% confidence score as trans - - ionone using the NIST14 library. Confidence in this identification increases when considering the calculated experimen tal retention index (RI) of 1495 was within 5% of the literature theoretical value (1486, NIST 14 database). Within the Algae + Rotifer cultures, all of the discriminating VOCs were structurally - related ketones or aldehydes: (a) Compound 6: trans - - ionone [IUPAC name: ( E ) - 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 3 - buten - 2 - one], (b) Compound - cyclocitral [IUPAC name: 2,6,6 - trimeth yl - 1 - cyclohexene - 1 - carboxaldehyde], (c) Compound 1: 2,2,6 - trimethyl - cyclohexanone, and (d) Compound 5: 4 - (2,6,6 - trimethyl - 1 - 109 cycl ohexene - 1 - yl) - 2 - butanone. Within the Algae and Algae + Rotifer cultures, 3 - nonenoic acid methyl ester has an adequate confidenc e for identification (74% spectral match, < 5% RI deviation from theoretical value). For those VOCs that could not be identified using the relatively conservative filtering criteria , the observed spectra and experimental retention indices provide suggestio ns for their identification. The suggested compound classes for unknown compounds are provided in Table 4. 1, and their experimental mass spectra are included in Appendix Figure A.4. 1. For example, the mass spectrum of Compound 12 has similar peaks and peak ratios to those of the identified carotenoids, suggesting a terpenoid structure with a molecular weight of 208 Da. The mass spectrum of Compound 13 contains ions characteristic of hexadecenoic acid including m/z 43, 60, 73, 129, 213, and 256 (M + ), and its experimental RI is within 1% of the literature RI of hexadecanoic acid (1968). 4.3.3 Abundance of VOCs In addition to the qualitative analysis, the relationship between rotifer d uration of grazing and the abundance of Algae + Rotifer distinguishing VOCs was examined . Levels of Compounds 4 and 6 were compared for individual Algae + R otifer cultures (Carboys 5 and 6) across individual experiments (Figure 4. 5). While there was no det ected signal from either compound before the addition of rotifers (48 h ), a ll detected signals after rotifer addition exceeded 2.0 x 10 5 counts, 2 - 3 orders of magnitude above detection threshold. Another - cyclocitral, appeare d in Experiment 2 after 24 h of rotifer feeding, with the signal increasing to more than 6.0 x 10 5 counts after 48 h of rotifer feeding. Similar comparisons for Compounds 1 - 7 are included in Appendix Figure A.4.2. Of note, 110 Figure 4.5 Peak areas of extracted compound chromatograms for trans - - ionone and - cyclocitral across Exp eriments 1, 2, and 3, separated by biological replicates. Err or bars represent standard deviation derived from duplicate measurements for each sample. The exposure time for SPM E fibers was 30 minutes in Experiment 1 and 60 minutes in Experiments 2 and 3. comparison of Figure 4. 5 to Figure 4. 3 reveals several instan ces where these VOCs were detected in Algae + Rotifer cultures before biomass loss was apparent as compared to alga e controls. For example, the second biological replicate of Algae + Rotifers in E xperiment 2 did not differ in algal density from the health y controls at the 72 h and 96 h timepoints. However, the signals for Compounds 4 and 6 were already large (6.0 x 10 5 and 1.0 x 10 6 counts, respectively). 111 4.4 DISCUSSION This SPME - GC - MS analysi s has identified seven discriminating VOCs in M. salina cultu res in the presence of actively - grazing B. plicatilis (Algae + Rotifer). The absence of these volatiles in the time - matched Algae control cultures (Figure 4. 3) suggests these chemicals are specif ic signals of active algal grazing , algal physiological stres s, or algal death. Many of these chemicals were detec ted within 24 h after rotifer addition and before algal cell densities changed substantially relative to controls (Figure 4. 2, 4. 4). Specifica lly, Compounds 4, 5, 6, and 7 were identified as early and ro bust grazing signals observed in M. salina cultures c ontaining rotifers. Several identified biomarkers Compounds 1, 4, 5, 6 and 7 were detected only during rotifer grazing and contained struc tural similarities, hinting at a shared metabolic pathway. Ma ny of these compounds (Table 4. 1) are known products of carotenoid oxidation [34,35] . Carotenoid - derived substances have been previously observed in algae volatile research, largely associated with investigations of flavor or smell components in food production. Carotenoids have important physiological functions as a component the light - harvesting complexes that transfer light energy to chlorophyll in photo systems [36] . Oxidative cleavage of the carotenoid backbo ne can occur through enzymatic (carotenoid cleavage dioxygenases) or non - enzymatic (light, oxygen, temperature) mechanisms [36] . Potential sites of - caro tene are shown in Figure 4. 6. In this work, the carotenoid - de rived VOCs could be generated - carotene released upon lysis of M. salina cells during the digestive process of B. plicatilis . This would be in agreement with studies of Ara bidopsis plants exposed to reactive oxygen species resulting - - cyclocitral [37] . The results from vascular plants suggest that carotenoid degradation products may be more ge neral indicators of stressed or wounded algae cultures, not s olely limited to the interaction of algae with rotifers. 112 Figure 4.6 VOCs identified from the headspace of Algae + Rotifer cultures formed from the - carot ene. Only those oxidation cleavages relevant to this study ar e pictured, but all double bonds across the - carotene backbone are cleaved. Although there are no reports of such analyses of M. salina , a small number of studies have examined algae from the g enus Nannochloropsis , of which M. salina is a close relative [38] . Van Durme et al [39] investigated the volatile composition of five microalgae species ( Botryococcus b raunii , Rhodomonas , Tetraselmis sp., Nannochloropsis oculata and Chlorella vulgaris ) by heating samples (40 °C) to enhance volatile signatures under heat stress conditions. - - - cyclocitral, were identified in all species tested. In terestingly, N. oculata contained a large abundance of ethanol , 2 - hydroxy - 2 - butanone, - cyclocitral and ionones were detected. Hosoglu [40] likewise characterized the volatilomes of several microalgae species using SPME - GC - MS and GC - olfactometry for both chemical profiles and olfactory properties to benefit incorporation into food products and to minimize unpleasant smells. The species C. 113 vul garis, C. protothecoides, and T. chuii reportedly contain distinguishing amounts of the carotenoid degradation prod - - ionone and 6 - methyl - 5 - hepten - 2 - one, while expressing a woody smell. While VOCs have been observed in analyses of chemical comp ositions of algae in destructive manners ( e.g., heating, sonication, solvent extraction etc.), there are fewer repo rts of volatiles emitted from live, actively - growing cultures. A variety of live algae - derived volatiles (terpenoids, aldehydes, halogenated compounds, etc.) have been shown to influence the odor quality of water [28] . Zhou et al [41] investigated changes in the volatilome of intact algae over different growth phases (logarithmic, stationary, and decline phase) for six microalgae ( Thalassiosira weissflogii , Nitzschia closterium , Chaetoceros calcitrans , Platymonas helgolandica , Nannochloropsis spp . (NMBluh014 - 1), and Dicrateria inornate) . The Nannochloropsis volatilome was largely dominated by alkanes and alkenes and 8 - heptadecene, but no carotenoid by - products were rep orted. Several functions for active ly - released VOCs, including carotenoids, have been postulated, such as tolerance of light and oxidative stressors, signaling the presence of predators [36] , and transfer of information throughout algal colonies [28] . - cyclocitral has pr eviously been reported as a volatile emitted by the bloom - forming cyanobacterium Microcystis as a defense mechanism against grazing by Daphnia magna [42] . Similarly, this work focused on the volatiles generated from active grazing of algae, which will generate more rapid algal death compared to natural growth cycles. Here, we confirmed the importa - cyclocitral as an indication of algal cell damage due to grazing. Fu ture work will determine if VOCs produced by M. salina in the presence of B. plicatilis ha ve similar role s in algal - defense as observed previously. 114 Using a non - invasive, non - dest ructive sampling and analysis technique, we have demonstrated that VOCs from the headspace of algae cultures can distinguish between algae cultures with grazing rotifers present and uninfected algal cultures, and may serve as general indicators of algal ce ll stress or death. It is worth noting that the list of biomarkers reported in this analysis may be considered conservative owing to the stringency of our filtering criteria. Additional biomarkers may be observed with more experimental repeats, more effici ent VOC sampl ing and higher - sensitivity analyses. In order to discover and validate additional diagnostic bio markers of grazer infection or other incipient crashes, more extensive stud ies of emitted volatiles from algae species are required. For example, l ow levels of grazer - associated VOCs in healthy algal cultures may result from background rates of algal death. Such background signals are likely modulated by physiological state of the culture (e.g. , exponential growth or stationary phase) or by nutrient limitation. T hus, an improved understanding of the threshold biomarker concentrations that indicate the need for interdictive treatment is paramount. Additionally, our SPME - GC - MS methodology is expected to have broader applications. Complex systems - level dynamics betw een algae, commensal bacteria, and various grazers will require more sophisticated sampling procedures alongside volatilomics data to include biological data sets, such as transcriptomics, metagenomics, and metabolomics. These systems - level an alyses and bi oinformatics analysis would be more likely to elucidate biological interactions or implications for the chemicals observed in the volatilome. Non - invasive and non - destructive VOC sampling is an attractive, analytical way to better understand a nd predict th e health of microbial cultures. 115 4.5 CONCLUSIONS The work presented in this chapter has aimed to increase the breadth and depth of reported algal and rotifer - specific VOCs, providing a tool to better define the physiological state of algae p onds and faci litate greater algal biomass production. A SPME - GC - MS methodology for non - invasive and non - destructive sampling of M. salina infected by B. plicatilis aided our discovery of seven putative culture crash biomarkers, including trans - - ionone and - cyclocitra l, over several timepoints during active crashing of algal ponds. These biomarkers were not detected in cultures displaying natural background levels of cell death, suggesting that these signals are produced by high stress conditions, such as rotifer grazi ng. Finally, these biomarkers offer potential as diagnostic tools to signal the need for crash mitigation strategies, as several signals were detectable before cell death was evident from changes in cell density. Both VOC baselines and signatu res from mult iple healthy and infected cultures will be compiled in a data base. Early use of this technique would then include surveilling for the emergence of targeted VOC biomarkers of algal distress or injury above healthy baseline thresholds that are indicative of imminent culture failure. The future of this work could see VOC based monitoring in open ponds . In such influenced by external sources (e.g. , VOCs fr om the enviro nment, wind effects, particulates, etc.), creating a variable background that would require correction for the levels of biomarker compounds. The creation of s (perhaps using a large fun nel), during sample collection, would serve to limit outside background The SPME fibers used in this experiment are field deployable and can easily be adapted to an algal pond production system. Although SPME - GC - MS has proven powerful for 116 untargeted d iscovery of algal volatile chemical signatures from healthy or grazed cultures, the cost of state - of - the art laboratory - based GC - MS systems and analyses efforts is prohibitive for using this method fo r continuous monitoring of industrial scale, open algal ponds. Knowledge gained and biomarkers annotated from our untargeted discovery efforts may guide development of targeted, lower - cost, field - deployable detectors capable of monitoring for changes in di agnostic chemical signatures and detecting volatile sign als of infection in real - time to facilitate the timely deployment interdictive strategies to prevent pond crashes. Miniaturized GC - MS systems [43,44] for field deployable detector systems is one such technology curr ently under development and optimization [45] for this typ e of application. 117 APPENDI CES 118 A PPENDIX A: Tables Table A.4.1 Significant difference determination between mean levels of algal cell densities a cross replicates of Algae ( M. salina ), Algae + Rotifer ( M. salina and B. plicatilis ) and Med ia Blank ( MB, ESAW) c alculated by ANOVA w Experiment 1 Experiment 2 Experiment 3 24 H 48 H 72 H 96 H 120 H 24 H 48 H 72 H 96 H 120 H 24 H 48 H 72 H 96 H 120 H MB 1 vs MB 2 - >0.9999 0.9942 >0.9999 >0.9999 >0.9999 >0. 9999 >0.9999 >0.9999 >0.9999 >0.9999 >0.9999 >0.9999 >0.9999 >0.9999 MB 1 vs Algae 1 - 0.0024 <0.0001 <0.0001 <0.0001 0.0202 <0.0001 <0.0001 <0.0001 <0.0001 0.4867 0.0696 <0.0001 <0.0001 <0.0001 MB 1 vs Algae 2 - 0.0011 <0.0001 <0.0001 <0.0001 0. 0075 <0.0001 <0.0001 <0.0001 <0.0001 0.3291 0.0136 <0.0001 <0.0001 <0.0001 MB 1 vs Algae + Rotifer 1 - 0.0004 0.2087 0.2671 0.2335 0.0088 <0.0001 <0.0001 0.0079 0.0051 0.2848 0.0115 0.0001 <0.0001 <0.0001 MB 1 vs Algae + Rotifer 2 - 0.002 0.6522 0. 9602 0.9978 0.0081 0.0001 <0.0001 <0.0001 <0.0001 0.2356 0.0086 0.0002 <0.0001 <0.0001 MB 2 vs Algae 1 - 0.0024 <0.0001 <0.0001 <0.0001 0.0202 <0.0001 <0.0001 <0.0001 <0.0001 0.4861 0.069 <0.0001 <0.0001 <0.0001 MB 2 vs Algae 2 - 0.0011 <0.0001 < 0.0001 <0.0001 0.00 75 <0.0001 <0.0001 <0.0001 <0.0001 0.3286 0.0135 <0.0001 <0.0001 <0.0001 MB 2 vs Algae + Rotifer 1 - 0.0004 0.0719 0.2662 0.2335 0.0088 <0.0001 <0.0001 0.008 0.0051 0.2844 0.0114 0.0001 <0.0001 <0.0001 MB 2 vs Algae + Rotifer 2 - 0.002 0.3351 0.959 8 0.9978 0.0081 0.0001 <0.0001 <0.0001 <0.0001 0.2352 0.0085 0.0002 <0.0001 <0.0001 Algae 1 vs Algae 2 - 0.9996 0.9827 >0.9999 0.9898 0.9987 >0.9999 0.2066 0.9996 0.6884 0.9997 0.9814 0.8709 0.6052 0.9915 Algae 1 vs Algae + R otifer 1 - 0.9818 <0. 0001 <0.0001 <0.0001 0.9994 0.9999 0.9711 0.0148 <0.0001 0.999 0.9722 0.9966 0.7581 0.1778 Algae 1 vs Algae + Rotifer 2 - >0.9999 <0.0001 <0.0001 <0.0001 0.9991 0.9868 0.7331 0.8134 <0.0001 0.9964 0.9503 0.9866 0.0302 0.064 A lgae 2 vs Algae + Rot ifer 1 - 0.9987 <0.0001 <0.0001 <0.0001 >0.9999 0.9992 0.0422 0.0069 <0.0001 >0.9999 >0.9999 0.6082 0.0622 0.0533 Algae 2 vs Algae + Rotifer 2 - 0.9999 <0.0001 <0.0001 <0.0001 >0.9999 0.9723 0.9268 0.9322 <0.0001 >0.9999 >0.99 99 0.5053 0.0005 0.01 63 Algae + Rotifer 1 vs Algae + Rotifer 2 - 0.9887 0.9615 0.7415 0.455 >0.9999 0.9984 0.2879 0.0006 0.0016 >0.9999 >0.9999 >0.9999 0.4138 0.9958 Timepoints are reported relative to the addition of algae to the growth media. Rotif ers were added to eac h condition after the 48 hour timepoint. - 119 Table A.4.2 List of VOCs in Individual Experiments Exper iment 1 Compound Number Mass NIST ID NIST % Match Experimental Retention Index Theoretical Retention Index Algae + Rotife r Cultures 1 208 1 074 2 152 1214 3 83 2 - Butanone, 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 70 1453 1433 4 177 trans - .beta. - Ionone 87 1506 1486 5 83 8 - Heptadecene 75 1691 1719 Algae + Rotifer AND Algae Cultures 1 57 113 4 2 148 1415 Experiment 2 Compound Number Mass NIST ID NIST % Match Experimental Retention Index Theoretical Retention Index Algae + Rotifer Cultures 1 59 987 2 82 1021 3 71 1101 4 107 1181 5 121 Phenol , 2,3,5 - trimethyl - 7 1 1190 1235 120 Table A.4.2 (cont'd) 6 137 1 - Cyclohexene - 1 - carboxaldehyde, 2,6,6 - trimethyl - 89 1209 1220 7 121 2 - Butanone, 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 85 1443 1433 8 177 trans - .beta. - Ionone 93 1495 1486 9 57 1692 10 143 1881 11 135 2036 12 71 2115 Algae + Rotifer AND Algae Cultures 1 55 Hexanoic acid, 2 - ethyl - , methyl ester 68 1031 1043 2 96 3 - Nonenoic acid, methyl ester 73 1134 1191 3 341 1138 4 71 1264 5 138 1338 6 73 1774 7 154 1854 8 73 1983 9 73 2079 10 192 2183 11 73 2210 Algae Cultures 1 71 757 2 56 1003 3 94 1207 4 91 1530 5 109 1619 6 119 2,4 - Diphenyl - 4 - methyl - 1 - pentene 83 1803 1 846 7 70 1955 8 70 2252 121 Table A.4.2 (cont'd) Experiment 3 Compound Number Mass NIST ID NIST % Match Experimental Retention Index Theoretical Retention Index Algae + Rotifer Cultures 1 118 971 2 82 Cyclohex anone, 2,2,6 - trimeth yl - 79 1021 1036 3 55 1054 4 107 1181 5 121 1191 6 137 1 - Cyclohexene - 1 - carboxaldehyde, 2,6,6 - trimethyl - 81 1209 1220 7 122 1379 8 121 2 - Butanone, 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 76 1419 1433 9 327 1429 10 401 1458 11 177 trans - .beta. - Ionone 94 1495 1486 12 179 1522 13 158 1675 14 57 1691 15 172 1744 16 73 1744 17 150 1776 18 73 1907 19 251 2406 20 73 2422 21 149 2496 122 Table A.4.2 (cont'd) Algae + Rotifer AND Algae Cultures 1 57 978 2 71 1039 3 96 3 - Nonenoic acid, methyl ester 74 1134 1191 4 341 1139 5 71 1293 6 138 1338 7 71 1370 8 57 1507 9 73 1983 10 192 2112 11 192 2197 12 73 2345 123 A PPENDIX B: Figures 1a) Compound 1: Cyclohexanone, 2, 2, 6 - trimethyl - 1b) Compound 2: m/z 107 , RI 1181 Figure A.4. 1 Experimental mass spectra for all VOCs listed in Table 4. 1, annot ated either by NIST ID match or base peak and retention index pair. 124 1c) Compound 3: m/z 121, RI 1191 - cyclocitral 125 1e) Compound 5: 2 - Butanone, 4 - (2,6,6 - trimethyl - 1 - cyclohexen - 1 - yl) - 1f) Compound 6: trans - - ionone 126 1g) Compound 7: m/z 57, RI 1691 1h) Compound 8: m/z 71, RI 1039 127 1i) Compound 9: 3 - nonenoic acid, methyl ester 1j) Compound 10: m/z 341, RI 1139 128 1k) Compound 11: m/z 71, RI 1293 1l) Compound 12: m/z 138, RI 1338 129 1m) Compound 13: m/z 73, RI 1983 : Hexadecanoic acid 1n) Compound 14: m/z 192, RI 2197 130 Figure A.4. 2 Peak areas of extracted compound chromatograms for Compounds 1, 2, 3, 5, and 7 across Experiments 1, 2, and 3; Compounds 4 and 6 are displayed in Figure 4. 5 131 Figure A.4. 2 132 Figure A.4. 2 133 REFERENCES 134 REFE RENCES 1 Reese, K. L. et al. Chemical p rofiling of v olatile o rganic c ompounds in the h eadspace of al gal c ultures as e arly b iomarkers of a lg al p ond c rashes. Scientific Reports 9 , 13866, doi:10.1038/s41598 - 019 - 50125 - z (2019). 2 Hannon, M., Gimpel, J., Tran, M., Rasala, B. & Mayfield, S. Biofuels from algae: challenges and potential. Biofuels 1 , 763 - 784 (2010). 3 Leite, G. B., Abdelaziz, A. E. M . & Halle nbeck, P. C. Algal biofuels: Challenges and opportunities. Bioresource technology 145 , 134 - 141, doi:10.1016/j.biortech.2013.02.007 (2013). 4 Katiyar, R., Kumar, A. & Gurjar, B. R. in Biofuels: Technology, Challenges and Prospects (eds Avi nash Ku mar Agarw al, Rashmi Avinash Agarwal, Tarun Gupta, & Bhola Ram Gurjar) 157 - 175 (Springer Singapore, 2017). 5 Richardson, J. W. et al. A financial assessment of two alternative cultivation systems and their contributions to algae biofuel economic via bility. Algal Re search 4 , 96 - 104, doi:doi.org/10.1016/j.algal.2013.12.003 (2014). 6 Day, J. G., Gong, Y. & Hu, Q. Microzooplanktonic grazers A potentially devastating threat to the commercial success of microalgal mass culture. Algal Research 27 , 356 - 365 , doi:1 0.1016/j.algal.2017.08.024 (2017). 7 Carney, L. T. & Lane, T. W. Parasites in algae mass culture. Front Microbiol 5 , 278, doi:10.3389/fmicb.2014.00278 (2014). 8 Hirayama, K. & Ogawa, S. Fundamental s tudies on p hysiology of r otifer for its m ass c ultu re - I: F ilter f eeding of r otifer. NIPPON SUISAN GAKKAISHI 38 , 1207 - 1214, doi:10.2331/suisan.38.1207 (1972). 9 McBride, R. C. et al. Contamination m anagement in l ow c ost o pen a lgae p onds for b iofuels p roduction. Industrial Biotechnology 10 , 221 - 227, doi:10.1 08 9/ind .2013.0036 (2014). 10 Zabed, H. M. et al. Biogas from microalgae: Technologies, challenges and opportunities. Renewable and Sustainable Energy Reviews 117 , doi:10.1016/j.rser.2019.109503 (2020). 11 Park, S. et al. The s elective u se of h ypochlorite t o p reve nt p ond c rashes for a lgae - b iofuel p roduction. Water environment research : a research publication of the Water Environment Federation 88 , 70 - 78, doi:10.2175/106143015x14362865227670 (2016). 12 Pradeep, V. et al. Use of c opper to s electively i nhibit Brachio nus calyciflorus ( p redator) g rowth in C hlorella kessleri ( p rey) m ass c ultures for a lgae b iodiesel p roduction. 135 International journal of molecular sciences 16 , 20674 - 20684, doi:10.3390/ijms160920674 (2015). 13 Fott, B. Phlyctidium s cenedesmi spec. nov a, a ne w chytrid destroying mass cultures of algae. Zeitschrift für allgemeine Mikrobiologie 7 , 97 - 102, doi:10.1002/jobm.19670070203 (1967). 14 Xu, C. et al. The u se of the s chizonticidal a gent qu inine s ulfate to p revent p ond c rashes for a lgal - b iofuel p rod uction. International journal of molecular sciences 16 , 27450 - 27456, doi:10.3390/ijms161126035 (2015). 15 Van Ginkel, S. W. et al. Taking advantage of rotifer sensitivity to rotenone to prevent pond crashes for algal - biofuel production. Algal Research 10 , 100 - 103 , doi:10.1016/j.algal.2015.03.013 (2015). 16 Liu, Z. & Lu, G. The sterilizing studies of flagellate and cilitate in marine unicellular algae liquid. Zhanjiang Aquacult. Coll. 6 , 36 - 41 (1990). 17 Wang, H., Zhang, W., Chen, L., Wang, J. & Liu, T. The contami nation and control of biological pollutants in mass cultivation of microalgae. Bioresource technology 128 , 745 - 750, doi:10.1016/j.biortech.2012.10.158 (2013). 18 Fisher, C. L. et al. Bacterial co mmunities protect the alga Microchloropsis salina from grazin g by the rotifer Brachionus plicatilis. doi:10.1016/j.algal.2019.101500 (2019). 19 U.S. Department of Energy. Bioenergy Technologies office Multi - Year Program Plan: May, 2013. Washington, D.C., 2 013 . 20 Borowitzka M. A. in Algal Culturing Techniques (ed Andersen R. A.) 205 - 218 (Elsevier Academic Press, 2005). 21 Day, J. G., Thomas, N. J., Achilles - Day, U. E. & Leakey, R. J. Early detection of protozoan grazers in algal biofuel cultures. Bioreso urce technology 114 , 715 - 719, doi:10.1016/j.biortech.2 012.03. 015 (2012). 22 Wang, Y., Castillo - Keller, M., Eustance, E. & Sommerfeld, M. Early detection and quantification of zooplankton grazers in algal cultures by FlowCAM. Algal Research 21 , 98 - 102, doi: 10.1016/j.algal.2016.11.012 (2017). 23 Carney, L., McB ride, R ., Smith, V. & Lane, T. Molecular Diagnostic Solutions in Algal Cultivation Systems. 183 - 204, doi:10.1201/b19464 - 9 (2016). 24 Carney, L. T. et al. Pond Crash Forensics: Presumptive identification of pond crash agents by next generation sequencing in replic ate raceway mass cultures of Nannochloropsis salina. Algal Research 17 , 341 - 347, doi:org/10.1016/j.algal.2016.05.011 (2016). 136 25 Achyuthan, K. E., Harper, J. C., Manginell, R. P. & Moorman, M. W. Volatile m etabolites e mission by i n v ivo m icroalgae - An o verlo oked o pportunity? Metabolites 7 , 39, doi:10.3390/metabo7030039 (2017). 26 Rowan, D. D. Volatile metabolites. Metabolites. 1 , 41 - 63, doi:10.3390/metabo1010041 (2011). 27 Leach, J. E., Triplett, L. R., Argueso, C. T. & Trivedi, P. Communication in the p hytob iome. Cell 169 , 587 - 596, doi:10.1016/j.cell.2017.04.025 (2017). 28 Zuo, Z. Why Algae Release Volatile o rganic c ompounds - The e mis sion and r oles. Front Microbiol 10 , 491, doi:10.3389/fmicb.2019.00491 (2019). 29 Wolfe, G. & Steinke, M. Grazing - Activate d Produ ction of Dimethyl Sulfide (DMS) by two clones of Emiliania huxleyi . Vol. 41 (1996). 30 Wolfe, G. V., Sherr, E. B. & Sherr, B. F. Release and consumption of DMSP from Emiliania huxleyi during grazing by Oxyrrhis marina . Mar. Ecol. Prog. Ser. 111 , 111 - 119 (1 994). 31 Wolfe, G. V., Steinke, M. & Kirst, G. O. Grazing - activated chemical defence in a unicellular marine alga. Nature 387 , 894 - 897, doi:10.1038/43168 (1997). 32 Hay, M. E. Marine chemical ecology: chemical signals and cues structure marine popul ations, communities, and ecosystems. Annual review of marine science 1 , 193 - 212, doi:10.1146/a nnurev.marine.010908.163708 (2009). 33 Fisher, C., Lane TW. Operational, p rophylactic, and i nterdictive t echnologies in a lgal c rop p rotection. Grand Challenges in Algae Biotechnology (2019). 34 C. M. d e Jesus Benevides ., M.C. da Cunha Veloso, P.A. de Paula Pereira, J.B. de Andrade. - carotene oxidation by ozone in an organic model system and the identification of the resulting products. Food Che mistry 126 , 927 - 934, doi: 10.1016/j.foodchem.2010.11.082 (2011). 35 Christaki, E., Bonos, E., G iannenas, I. & Florou - Paneri, P. Functional properties of carotenoids originating from algae. Journal of the science of food and agriculture 93 , 5 - 11, doi:10.1002 /jsfa.5 902 (2013). 36 Havaux, M. Carotenoid oxidation products as stress signals in plants. Th e Plant journal : for cell and molecular biology 79 , 597 - 606, doi:10.1111/tpj.12386 (2014). 37 Ramel, F. et al. Carotenoid oxidation products are stress signals t hat med iate gene responses to singlet oxygen in plants. Proceedings of the National Academy of Sciences of the United States of America 109 , 5535 - 5540, doi:10.1073/pnas.1115982109 (2012). 38 Fawley, M. W., Jameson, I. & Fawley, K. P. The phylogeny of the g enus Na nnochloropsis (Monodopsidaceae, Eustigmatophyceae), with descriptions of N. 137 australis s p. nov. and Microchloropsis gen. nov. Phycologia 54 , 545 - 552, doi:10.2216/15 - 60.1 (2015). 39 Van Durme, J., Goiris, K., De Winne, A., De Cooman, L. & Muylaert, K. Evalua tion of the volatile composition and sensory properties of five species of microalgae. Journal of agricultural and food chemistry 61 , 10881 - 10890, doi:10.1021/jf403112k (2013). 40 Isleten Hosoglu, M. Aroma characterization of five microalgae species using solid - phase microextraction and gas chromatography mass spectrometry/olfactometry. Food Chemistry 240 , 1210 - 1218, doi: 10.1016/j.foodchem.2017.08.052 (2018). 41 Zhou, L. et al. Chan ge of volatile components in six microalgae with different growth pha ses. Jo urnal of the science of food and agriculture 97 , 761 - 769, doi:10.1002/jsfa.7794 (2017). 42 Juttner, F., Watson, S. B., von Elert, E. & Koster, O. beta - cyclocitral, a grazer defence signal unique to the cyanobacterium Microcystis. Journal of chemical ecolog y 36 , 1387 - 1397, doi:10.1007/s10886 - 010 - 9877 - 0 (2010). 43 Whiting, J. J. et al. A high - speed, high - performance, microfabricated comprehensive two - dimensional gas chromatograph. Lab on a chip 19 , 1633 - 1643, doi:10.1039/c9lc00027e (2019). 44 Lewis, P. R. et al. Recent advancements in the gas - phase MicroChemLab. IEEE Sensors Journal 6 , 784 - 795, doi:10.1109/JSEN.2006.874495 (2006). 45 Snyder, D. T., Pulliam, C. J., Ouyang, Z. & Cooks, R . G. Miniature and f ieldable m ass s pectrometers: Recent Advances. Ana lytical chemistry 88 , 2 - 29, doi:10.1021/acs.analchem.5b03070 (2016). 138 Chapter 5. EVALUATION OF DETECTION OF ALGAL - BACTERIAL INTERACTIONS BY TRACKING VOLATILE BIOMARKERS FOREWORD Contributions from others to conduct the experiments described in th is chap ter are as follows: X. Mayali and R. K. Stuart provided stock cultures of Phaeodactylum tricornutum and Marinobacter spp. 3 - 2. K. Rolison assisted wit h experimental set up, maintenance of experimental cultures, and discussions of data analysis. 5.1 INTROD UCTION Algae have evolved in close relationships with other microbes, and interactions with other microbes influence their physiology and growth in na tural systems [1] and likely in engineered system s as we ll. Our ability to assess and predict algal physiology and activity in situ is limited due to a lack of understanding about dynamic interactions with other microbes that naturally inhabit the same space. In contrast and complementary to the work of Chapter 4, where we investigated effect of predator gazing on algae on emitted alg al volatile organic compound (VOC) markers, the research presented here in Chapter 5 aimed to lay further groundwork for understanding how microorganisms interact with each o ther at the molecular level using a model algal - bacterial co - culture and measuring volatile markers indicative of healthy (or stressed) algae in co - culture with bacteria. Organic matter is exuded from all microorganisms, including b acteria and algae , thr oughout their lifecycles. Contained within these complex mixtures are volatile, se mi - volatile, and non - volatile compounds. Such substances contribute to the biochemistry of aquatic systems y to fi nd food sources, 139 providing chemical defense from predators, or serving as a n intra - species indicator of danger. However, the presence and roles of these compounds have not been fully understood, nor their effects on other microorganisms. Algae and bacteri a have coexisted for millions of years. Exchange of nutrients, signaling mo lecules, and other dissolved materials between organisms [2,3] happens in the area of closest contact, the microscale chemical envir onment surrounding algal cells, which has been termed the phycosphere (analogous to the rhizosphere around roots) [4] . Within the phycosphere, the exchange of volatile and non - volatile exuded metabolites can furthe r affec t growth cycles of each organism. Changes in the metabolic processes that affect levels of volatile compound accumulation might be detected in the heads pace above actively growing algae ponds. The detection, identification, and quantification of the se mole cules in complex communities may enable diagnostics and/or manipulation of these interactions for a number of applications, including bioenergy, algal b ioproducts and agriculture, and carbon capture . Interactions of algae and bacteria, as well as t heir ex uded organic matter, have been investigated, where the complexity of the systems was limited in order to determine the origins of identified metabolites . Several studies have concluded that algae can utilize vitamins and minerals exuded by bacteria, while bacteria can utilize ammonia exuded from algae [5 - 7] . Some exchanges are bene ficial, where algae and bacterial growth is promoted through exchange of nutrients. Identification of exuded metabolites could gui de addi tion of target microbes to commercial cultures to increase biomass production [1] . In contrast, competition between the co - cultured organisms for nutrients and dissolved organic material may inhibit algal g rowth [8] . The effects can be either specific to a species or inhibit a broad range of algae. Studies have indicat ed that inhibitory or non - growth promoting interactions play an important role in organizing the 140 structure of marine communities a nd ensu ring survival [8,9] . Identification of inhibitory compounds may allow control of harmful phenomena in algal systems, such as algae blooms, that produce toxic effects on marine wildlife and can contaminate so urces o f drinking water [10] . The research presented in this chapter explored the feasibility of detecting and identifying algae - bacteria interactio ns at t he molecular level by tracking volatile metabolites in the complex gaseous headspace of the model alga Phaeodactylum tricornutum (P. tricornutum) and model bacterium Marinobacter spp. 3 - 2 at multiple timepoints over different algal growth states (ex ponenti al and stationary) . P. tricornutum is a diatom with a sequenced genome used for biofuel production because of its high lipid content (estimated ~30% of dry weight [11] including abundant polyunsaturated fatty acids), and Marinobacter is a diverse genus of Gram - negative, aerobic bacteria found in most o ceans [12] . The focus of this work was to detect and identify volatiles characteristic for each species as function of their growth states, and potentially, volatiles characteristic for the interaction of both species. To acc omplish this, unta rgeted VOC biomarker profiles were obtained using solid phase microextraction (SPME) coupled with gas chromatography - mass spectrometry (GC - MS) on co - cultures, simulated co - cultures (i.e. algae exposed to exudates of bacteria, and v ice ver sa), and controls. We hypothesized that monitoring VOC profiles of liquid cultures of algae and bacteria would reveal changes in the metabolism, specifically 1) co - culturing P. tricornutum and Marinobacter spp. 3 - 2 would produce a more complex VOC b iomarke r profile compared to monocultures of each organism, and 2) cultures of P. tricornutum exposed to exuded materials from Marinobacter spp. 3 - 2 cultures (and vice versa) would produce a more complex VOC biomarker profile (possibly more similar to that of the co - cultures) comp ared to monocultures due to the organisms potentially sensing volatile and non - volatile infochemicals emitted by their counterpart. 141 5.2 METHODS 5.2.1 Sample preparation In order to evaluate the feasibility of collecting and detec ting VO Cs emitted from the headspace of algae and bacteria we set up single organism cultures of alga Phaeodactylum tricornutum ( P. tricornutum ) and bacterium Marinobacter subspecies 3 - 2 as well a s co - cultures containing both organisms. The following nomen clature (Table 5.1) will be used throughout this chapter to designate mono - and co - cultures from P. tricornutum and Marinobacter spp. 3 - 2 microorganisms and samples comprised of exudates from mono cultures of these microorganisms . All samples prepared for VOC col lection were prepared using Enriched Seawater/Artificial Water (ESAW) media. All samples used for VOC sampling were prepared in 250 - mL glass Erlenmeyer flasks with stainless steel closures (Bellco Glass, Inc.) capable of allowing gas exchange and mo dified to accommodate insertion of solid - phase microextraction (SPME) fibers. The target working volume of each sample was 100 m L. Table 5.1 Description of experimental sample types, abbreviation s, and number of sample replicates Sample Type Nomenclature Number of Replicates P. tricornutum Algae 2 Marinobacter spp. 3 - 2 Bacteria 2 P. tricornutum Exudates AlgEx 2 Marinobacter spp. 3 - 2 Exudates BacEx 2 P. tricornutum + Marinobacter spp. 3 - 2 Co - Cultures 3 P. tricornutum + Marinobacter spp. 3 - 2 Exudates Alg+Bac Ex 3 Marinobacter spp. 3 - 2 + P. tricornutum Exudates Bac+AlgEx 3 Enhanced Seawater/Artificial Water (ESAW) Growth Media ESAW 1 142 ESAW media : ESAW media was prepared according to the National Center for Marine Algae and Microbiota (NCMA) guideline s and m odified to contain elevated nutrient levels of 1.4 mM NaNO 3 3 PO 4 . The ESAW media control sample was created by adding only 100 mL of liquid media to one of the flasks. Algae : P. tricornutum (strain CCMP2561) was purchased from the National Center for Marine Algae (NCMA, West Boothbay Harbor, Maine, USA ) and grown in ESAW media axenically (referred to in this work as monocultures), me aning no additional organisms were present at the start of the experiment. Stocks were incubated at 20°C under artificial sunlight on a 16:8 light: dark cycle. Stock culture s were regularly checked for contamination by collecting culture biomass on a filte r, applying a blue - - diamidino - 2 - phenylindole), which targets adenine - thymine rich regions of double - stranded DNA, a nd visually checking for the presence of bacteria using fluorescence microscopy. Aliquots of stock cultures we re tra nsferred to sterile media 1 - 2x per month to maintain cultures, then again immediately prior to the experiment. At the start of the experiment, stock cultures were measured (fluorescence readings at 680 nm) to determine appropriate di lutions to obtain a con sistent algal cell density (target of 100 - 150 relative counts). Algae replicates were created by inoculating sterile ESAW media with aliquots of the stock solution immediately prior to the start of the experiment. Bacteria : Marinobac ter spp. 3 - 2 (Genbank a ccession number ASM375135v1), a bacterium that was isolated and genetically characterized at LLNL, was grown axenically in Zobell media. Stocks were incubated at 20°C under artificial sunlight on a 16:8 light: dark cycle. Aliquots of stock cultures were tr ansferred to sterile Zobell media 1 - 2x per month to maintain cultures, then again immediately prior to the experiment. Replicate monocultures of Bacteria 143 were created by diluting 5 mL of culture (estimated concentration 10 7 cells/mL) in 95 mL of ESAW. Co - C ultures : Stock co - cultures of P. tricornutum and Marinobacter spp. 3 - 2 were maintained and provided by LLNL researchers. Stocks were incubated and maintained under the same conditions as Algae stock. At the start of the experiment, stock cultures were meas ured ( excitation wavelength of 440 nm, fluorescence readings at 680 nm) to determine appropriate dilutions to obtain a similar algal cell density to Algae . (Distortions in the fluorescence rea dings due to the presence of bacteria are expected to be negligi ble.) Co - cultures replicates were created by inoculating sterile ESAW media with aliquots of the stock culture immediately prior to the start of the experiment. AlgEx / BacEx : To generate AlgE x and BacEx , separate axenic algae and bacteria cultures, each w ith total volumes of 600 mL, were grown in 1 L baffled glass Erlenmeyer flasks with vented caps for air exchange. Algae cultures were grown until early stationary phase (approximately 144 hour s post - inoculation, average fluorescence maximum 3000 counts) wi th incubation on a rotating platform at 20°C. Bacteria cultures were consistently grown for the same amount of time and displayed fluorescence counts comparable to an ESAW control. Subsequent environment removed algal or bacterial biomass from the filtrate containing dissolved organic matter (hereafter referred to as AlgEx and BacEx ). Samples were prepared up to one month in advan ce of the experiment, sealed and stored in 1L Pyrex bottles at 2 0°C. For the control cultures, 100 mL of each exudate were added undiluted directly into the Erlenmeyer flasks. Bac+AlgEx / Alg+BacEx : Samples containing cultures of one species mixed with exudates form the other species were created using the exudates Alg Ex and BacEx as a base 144 media. Levels of nutrients (1.4 mM NaNO 3 3 PO 4 ) were replenished to support the growth of the new cultures. Subsequently, the same volumes of stock solution used to create the Algae and Bacteria stocks were added to exudat e samples of the respective other species . To illustrate, one r eplicate sample of Bac+AlgEx was created by adding 95 mL of AlgEx and 5 mL of t he same stock Marinobacter spp. 3 - 2 culture used to generate replicates of Bacteria . 5.2.2 Estimation of algal ce ll densities Indirect measurements of live algal biomass were performed using fluorescence measurements of chlorophyll content on all sample types using a commercial plate reader (Cytation 5 Cell Imaging Multi - Mode Reader, BioTek Instruments, Inc./Agilen t Technologies) . Algal growth of all samples was monitored usin excitation and emission wavelengths 440 nm/680 nm approximately every 24 hours after sample inoculation. Additionally, at 72, 120, 168, and 24 0 hours post - in oculation, a 1 mL aliquot of each sample was pla ced in separate Eppendorf tubes and fixed with formaldehyde to prevent further biological growth these archival samples were saved for future measurements of bacterial concentrations by micro scopy. A second planned method of measuring culture growth inv olved monitoring bacterial concentrations of bacteria in monocultures or in co - cultures using fluorescence microscopy and DAPI staining. Unpublished work at LLNL by researchers (X. Mayali, R.K. Stuart) on cultures of P. tricornutum and Marinobacter spp. 3 - 2 using DAPI staining and fluorescence microscopy to determine an estimated bacteria concentration of 1 - 6*10 6 cells/mL over 8 days of growth in co - cultures. During each timepoint of VOC samplin g (72, 1 20, 168, and 240 hours post - inoculation), aliquots of a ll sample types were taken in anticipation of these measurements. 145 However, I was unable to acquire the fluorescence microscopy data owing to limited laboratory access resulting from COVID - 19. T he bacte rial concentration of all samples is anticipated to be measured at a later time. For the purposes of this dissertation, it is assumed the bacterial concentration was comparable to the 1 - 6*10 6 cells/mL levels in Bacteria and Co - cultures samples meas ured pre viously in similar mono - and and co - cultures . 5.2.3 VO C sampling VOCs were collected from the headspaces of samples using PDMS - DVB SPME fibers, similar to the methods employed in Chapter 4 (Section 4.2.5). One fiber was used per flask. As demonst rated in Figure 5.1, within a sterile environment, SPME fibers were inserte d into samples and exposed to the headspace for an average of 2.5 hours. VOCs were sampled (approximately) after 72, 120, 168, and 240 hours post - inoculation. 5.2.4 VOC data acqui sition, processing, and biomarker identification The VOC analy sis and puta tive biomarker identification followed procedures similar to those described in Chapter 3 and 4 and in published works [13] . Briefly, an untargeted GC - MS approach was performed using a Agilen t 5975T GC - single quadrupole - MSD system using 70 eV electron io nization and an Agilent HP - 5ms column (the same system used for work described in chapters 3 and 4). Data acquisition was performed using ChemStation software (version E.02.02). 146 Figure 5.1 E xperimental setup during passive VOC sampling of algal and bact erial sample s - using SPME fibers (n=1 fiber per 250 mL flask); average exposure time ~ 2.5 hours); replicate cultures indicated by similar - colored labels. Data processing was performed using MassHunter Qualitative and Mass Profiler Professional. Two filt ering criter ia similar to the ones used in Chapters 3 and 4 were applied to detect putative biomarkers from the large VOC dataset, but required modification accounting for smaller replicate num bers. For the first criterion, samples of n=3 replicates requir ed a detecte d VOC to be present in two of three replicates for at least one timepoint to be further considered a biomarker candidate. For samples of n=2 replicates, a detected VOC was required to appear in both replicates for at least one timepoint to be f urther consi dered a biomarker. If a compound meeting either of the previous two requirements was also present in one replicate at another timepoint for the same sample type, its presence was no ted (see Results). The second criterion for considering a VOC a s a biomarke r was unchanged compared to the work in chapters 3 and 4: the compound was not detected in the ESAW media blank or travel blank conditions; OR b) the compound was present in the sam ple type at an abundance greater than 10x the abundance in the ESAW media b lank or travel blank. 147 Additional data processing was performed to identify the taxa - specific algal and bacterial biomarkers across all samples and timepoints. Retention indices wer e calculated with reference to a commercial reference standard, and identifications were performed by searching mass spectra against the NIST14 mass spectral database. Biomarker identification required a mass spectral library match greater than 70% and an experimental retention index within 3% match of the reference c ompou nd. Biomarkers unable to be identified were presented as the m/z value of the base peak at a given retention index ( e.g., m/z 71 at RI 1271). Compounds identified with values of base peak m/z 207, 262, and 281 were removed from biomarker consideration , as these are common contaminant ions from polysiloxanes originating from either the chromatography column or the SPME stationary phase. 5.3 RESULTS AND DISCUSSION 5.3.1 Cell counts of microo rganisms in cultures Figure 5.2 presents algal growth (as mea sured by chlorophyll fluorescence) as a function of time after inoculation for the sample types with live algae. A higher biological variation in growth was observed amongst replicates of Alg+B acEx compared to the re plicates of Algae and Co - cultures , as in dicated by the error bars in this figure. For all sample types, logarithmic growth is maintained throughout approximately 120 hours before beginning to level off, indicating the beginning of th e stationary phase. The average algal growth in Co - cultures (n= 3 replicates) and Alg+BacEx (n=3) samples was consistently lower than that of Algae cultures (n=2). As algal growth is inhibitedin both Co - cultures and Alg+BacEx cultures, the reduction is poss ibly due to the actions of a growth inhibitor secreted by the b acteria rather than competition between live algae and bacteria for nutrients in the growth media. 148 5.3.2 Headspace VOC results for Cultures of P. tricornutum an d Marinobacter spp. 3 - 2 This wo rk demonstrated that our methodology 1) detected differences in the metabolite VOC biomarker profiles from the growth of P. tricornutum and Marinobacter subspecies 3 - 2, either as mono - cultures or co - cultures, using unique VO C biomarkers and 2) the biomarke r profile varies within sample types across subsequent timepoin ts, indicating temporal metabolite changes. Data processing and analysis resulted in 150 - 250 compounds detected per sample. Peak alignment of all samples and tim epoints generated a list of more than 3200 compounds. Application of filtering criteria based o n detection frequency and compound abundance (described in the methods) removed over 90% of this list, thus identifying the most robust Figure 5.2 Growth curve s of algae in samples containing P. tricornutum. Algae , Co - Culture , and Alg +Bac Ex , as determine d by measurements of relative chlorophyll fluorescence units (RFUs) at increasing timepoints post - inoculation of cultures. Averages RFUs (averaged over replicates) are plotted with error bars s howing +/ - standard deviations. Colors of the err or bars corres pond to sample type in the legend. Counts of RFUs in ESAW, Bacteria, and BacEx samples were simultaneously acquired but did not exceed 20 RFUs, hence considered to be negligible signal and omit ted from this figure. VOCs were collected via SPM E sampling at 72, 120, 168, and 240 hours post - inoculation as indicated by arrows. 149 compounds as putative VOC biomarkers. In this section, the metabolite profiles for Algae , Bacteria , and Co - cultures samples will be described, followed (in section 5.3.3) by the results f rom the study of metabolite changes in a primary organism when exposed to exudates from a secondary organism ( Alg+BEx , Bac+AEx ). Algae: A total of 19 putative VOC biomarkers were detected i n Algae cultures and are presented in Appendix Ta ble A.5.1. The following compound classes were represented: amines, aliphatics (cyclic olefins, alkanes), carotenoids, and methyl ketones. Three biomarkers were given specific identifications based on NIST14 mass spectral matching, specifically a) 6 - [( Z ) - 1 - Butenyl] - 1,4 - cycloheptadiene (common name ectocarpene), b) 6, 10, 14 - trimethyl - 2 - pentadecanone (common name phytone or hexahydrofarnesyl acetone), and c) N - ethyl - N - methyl - 2 - propen - 1 - amine. Both 6 - [( Z ) - 1 - But enyl] - 1,4 - cycloheptadiene and 6, 10, 14 - trimet hyl - 2 - pentadecano ne were present at all timepoints measured. Additionally, 4 non - identified VOC biomarkers could be given tentative annotations to compound classes, such as alkanes, through use of compound clas s characteristic MS fragmentation patterns, ba se peak ( m/z ) val ues, and retention index values. One biomarker of note, annotated as m/z 177 at RI 1495, exhibited mass spectral fragments at m/z 43 and 135, similar to those of the carotenoid degradation prod - ionone that was observed as a signal of algal wounding in Chapter 4. Therefore, this was annotated as a carotenoid degradation product. Bacteria: A total of 11 putative VOC biomarkers were detected in the Bacteria cultures (Appendix Table A.5.2). O ne biomarker with m/z 68 at RI 1015 was given a specific identification, the monoterpene D - limonene. An additional 2 markers, m/z 57 at RI 10 and m/z 57 at RI 1816, were assigned to the chemic al functional groups of alkanes. Interestingly, biomarkers with base peak m/z ratios greater than 100 were obs erved more frequently at the 72 and 120 150 sampling points, while the 168 and 240 hour sampling points both contained biomarkers with base peak m/z r atios less than 100. Co - cultures : An example of the chemical c omplexity of P. tricornutum + Marinobacter spp. 3 - 2 VOCs versus the ESAW growth media at 240 hours post - inoculation is shown for the total ion current chromatogram in Figure 5 .3 , where the put ative VOC biomarkers for Co - cultures are indicated. A total of 3 1 putative VOC biomarkers were detected across Co - cultures (Appendix Table A.5.3). The following compound classes were represented: aliphatics (cyclic olefins, alkanes), carboxylic acids, caro tenoids, methyl ketones, and terpene s. Six biomarkers were given specific identifications based on NIST14 mass spectral matching, specifically a) D - limonene, b) 6 - [( Z ) - 1 - butenyl] - 1,4 - cycloheptadiene, c) dianhydromannitol, d) 6, 10, 14 - trimethyl - 2 - pentadeca none, e) Z - 11 - hexadecenoic acid, and f) n - hexadecanoic acid. For the monounsaturated fatty acid Z - 11 - hexadecenoic acid, it is not possible to distinguish the double bond position using EI fragmentation. Therefore, subsequent reference to it will be Fig ure 5. 3 Example total ion current chromatograms for observed VOC s sampled from the Co - cultures and from ESAW at 240 hours post inoculation, with peaks indicated by * being potential VOC markers of algae - bacterial interactions. 151 RI 196 b - - ionone, were identified as the substances annotated at m/z 137 at RI 1208 and m/z 177 at RI 1 495 based on comparison to the work in Chapter 4. Analy sis of VOC data from Algae , Bacteria , and Co - cultures revealed groups of biomarkers that were indicative of each sample type as well as select compounds shared across sample types. Figure 5. 4 details the overlap among Co - cultures , Algae , and Bacteria . VOC biomarkers found in Algae also observed in Co - cultures included a) 6 - [( Z ) - 1 - butenyl] - 1,4 - cycloheptadiene, b) 6, 10, 14 - trimethyl - 2 - - ionone, d) an unidentified alkane at m/z 71, RI 1271, e) unidentified marker m/z 108 at RI 1021, and f) unidentified marker at m/z 73 at RI 2572. Markers found in Bacteria also observed in algal - bacterial Co - cultures included a) D - limonene, b) m/z 42 at RI 815, c) m/z 95 at RI 1151, and d) m/z 73 at RI 1912. First, for this work we hypothesized that co - culturing P. tricornutum and Marinobacter spp. 3 - 2 would produce more chemic al complexity in the headspace VOCs, manifesting in a larger total number of produced biomarkers and/or production of additional, different biomarkers, in the Co - cultures profile compared to monocultures. Twenty - one of the VOC markers from Co - cultures were novel signatures not observed in the monocultures, and thus suggested that the metabolism of Algae and Bacteria was altered because of their interaction, thus causing additional markers to be produced. The origins of seve ral markers were traced back to VO C emissions of the monocultures, six from the Algae cultures and eleven from the Bacteria cultures, indicating the production of these markers was not altered. 152 Figure 5. 4 Venn diagram of the overlapping VOC biomarkers amongst Co - cultures , Algae, and Ba cteria as inclusive of markers detected across all measured timepoints. Each group is derived from n=2 replicates. Detected volatile biomarkers in Co - cultures, Algae, and Bacteria evolved over time within monocultures and co - cultures as observed when sam pling at 72, 120, 168, and 240 hours post - inoculation. For example, ectocarpene was observed to be present and increase at all timepoints for Algae and Co - cultures . Previous studies have shown high levels of ectocarpene and other C11 unsaturated olefins re sult from metabolic cleavage of eicosanoid precursors in diatoms, of which P. tricornutum is a member [9,14,15] . Ectocarpene has been observed in P. tricornutum and in species of brown algae, where the biochemical function wa s related to pheromone production and sexual attraction [14] . These results are in agreement, as a higher cell count at each subsequent timepoint produced greater levels of ectocarpene. 153 Two biomarkers indicative of the later timepoints, where growth phase approaches stationary phase, were two 16 - carbon fatty acids: hexadecenoic acid (16:1) at RI 1963 and n - hexadecanoic acid (16:0; also known as palmitic acid). Both biomarkers were observed o nly at the 240 hour sampling of Co - cultures . An observation or increase in the levels of saturated or monosaturated lipids for the stationary phase has been documented in previous studies with axenic P. tricornutum cultures [15 - 18] , and this research further reports its presence when grown in the presence of additional microorganisms. Fatty acids such as this are components of triacylglycerols, which serve as a storage source for carbon and energy. When algae have d epleted media nutrients during the stationary phase, internal triacylglycerols can be metabolized to produce energy for further growth, thus releasing fatty acids such as palmitic acid. Interestingly, some biomarkers only observed at the 240 hour sampling time of Algae and Co - cultures were annotated as carotenoid degradation products. These signals were previously observed in the research presented in Chapter 4 sp ecifically as early - warning VOC biomarkers for algal pond crashes caused by active predation o f the algae Microchloropsis salina [13] . Conv - ionone was present in Algae - - ionone were both observed in Co - cultures. The growth data presented in Figure 5.2 suggests Marinobacter exudates inhibit the growth of P. tricornutum , potentially indicating a toxic effect tha t could increase algal wounding and the release of carotenoid products. Additionally, different algae lineages have been shown to produce varying levels of endogenous carotenoids [19] - ionone only in the 240 hour timepoint of P. tricorn utum cultures suggests that rising levels of algal stress or cell death during the start of stationary phase releases endogenous levels of carotenoids at a level detectable by SPME - GC - M S. 154 A potential infochemical, the methyl ketone 6,10,14 - trimethyl - 2 - pen tadecanone, common name hexahydrofarnesyl acetone, was also present and increase in abundance at all sampled timepoints for Algae and Co - cultures . This ketone is a fragrant compound obs erved in the green algae genera Spirogyra [20] and Cladophora [21] , and brown alga Padina pavonia [22] . It has also been observed in higher plants, in particular an extract of the vining plant Vitis setosa , and demonstrated to contain antimicrobial properties [23] . Although exact identification was not possible, also alkane - like biomarkers were annotated in this work (ba sed on characteristic fragmentations at m/z 43, 57, and 71). Different annotated alkanes were present in each sample type and represented at every sampled timepoint. Studies report P. tricornutum accumulates a rang e of hydrocarbons similar to those found p etroleum - based fuels, such as octane, undecane, nonadecane, heneicosane, heptadecane, nonadecane, and eicosane, although the temperature at which the algae was grown played a role in whether shorter or longer chain hydrocarbons were produced [24] . The hydrocarbon content of P. tricornutum and other algae species has been extensively studied for determining which species would be opti mal to use in biofuel cultivation. The presence of a wide range of VOCs have been reported for P. tricornutum that were not observed in this work. In particular, isoprene [25] , aldehydes [26] , organohalogens comprised of chlorine and brom ine [15] , and sulfur - containing compounds [27] have been reported as originating from marine sources. I f present in P. tricornutum , this method did not possess either the selectivity or sensitivity to observe these compounds. Multiple methodologies a nd detection schemes should be considered when performing untargeted biomarker profiling of algae and associa ted microorganisms. 155 5.3.3 Headspace VOCs for monocultures of one species exposed to exudates from the other Biomarkers from cultures of one species supplemented with exudates of the respective other ( Alg+BacEx and Bac+AlgEx ) were profiled to check if a ny of the additional biomarkers from Co - cultures could be replicated when the presence of one species is simulat ed by adding exudates. If this is true, this would indicate soluble compounds produced by one species triggers production of the VOCs by the oth er species. First, the biomarker profiles for the exudates ( AlgEx and BacEx ) are described, followed by building up the analysis of one of the organism s exposed to exudates from the respective other organism ( Alg+BacEx , Bac+AlgEx ). AlgEx: A total of 19 p utative VOC biomarkers were detected in the AlgEx s amples (Appendix Table A.5.4), covering the compound classes of alkanes and carboxylic acids. Two carboxylic acids, a) hexadecenoic acid (16:1) at RI 1963 and b) n - hexadecanoic acid were assigned identific ations, while an additional three were assigned to classes based on a chemical functional group. Figure 5. 5 illustrates the overlap between compounds detected in Algae and AlgEx , showing a 21% overlap of each profile based upon the following shared but uni dentified biomarkers annotated as: a) m/z 71 at RI 1271, b) m/z 401 at RI 1424, c) m/z 135 at RI 1473, and d) m/z 73 at RI 2572. BacEx: A total of 24 putative VOC biomarkers were detected in the BacEx Appendix Table A.5.5). While no compounds could be a ssigned a confident chemical identification, two markers were annotated as alkanes: m/z 57 at RI 1521 and m /z 71 at RI 1759. The volatile biomarker profile of BacEx contained the largest number of markers at 72 hours post inoculation, the first sampling ti mepoint, and numbers of markers decreased for each timepoint thereafter. There was one compound shared betw een Bacteria and BacEx , m/z 73 at RI 1912. 156 Figure 5. 5 Venn diagram of the overlapping VOC biomarkers among Algae and AlgEx across all measured tim epoints. Each group is derived from n=2 replicates. Greatest qualitative overlap occurred during sampling at the 72 hour timepoint. Alg+BacEx : A total of 49 VOC biomarkers were detected in Alg+BacEx cultures (Table A.5.6), a larg er number than that prod uced in the Algae or BacEx alone. Chemically identified markers included the cyclic olefin ectocarpene, the methyl ketone 2 - pentadecanone, 6,10,14 - trimethyl - , and the carboxylic acid n - hexadecanoic acid, all of which had been previously observed in co - cult ures. 2,5 - di - tert - butyl - 1,4 - benzoquinone was identified at the 72 hour sampling timepoint, identified by mass spectral fra gments m/z 220, 205, 192, 177, 163 and a RI of 1475. An additional nine biomarkers could be annotated as alkanes ( m/z fragments at 55 and 71) or carotenoids ( m/z fragments 177 and 137). While 18 and 24 biomarkers were previously observed in monocultures of Algae and BacEx , respectively, unique production of 31 VOC biomarkers were observed in Alg+BacEx . Overlaps in biomarkers present in t he sample types combined to create Alg+B acEx are presented in a Venn diagram in Figure 5.6 a . Bac+AlgEx : In Bac+AlgEx (Table A.5.7), a total of 29 biomarkers remained after data filtering, a larger number of markers than the sum of numbers of markers found for Bacteria and A lg Ex . None of these 29 biomarkers could be given confident chemical identifications, alth ough 157 several markers displayed fragmentation patterns consistent with alkanes. Unique production of 22 biomarkers were observed in Bac+AlgEx w hile 7 markers were previously observed in either monocultures of Bacteria and AlgEx , (Figure 5.6 b ). Finally, a co mparison of the Co - cultures against that of Alg+BacEx and Bac+AlgEx (Figure 5.6 c ) revealed that a majority of the shared substances were produced by P. tricornutum and/or the presence of exudates of Marinobacter. Our second hypothesis that exuded volatile metabolites from one organism can be sensed by and affect the growth of the other organism, demonstrated here for 1) P. tricornutum exposed to Marin obacter spp. 3 - 2 exudates and for 2) Marinobacter spp. 3 - 2 expose d to P. tricornutum exudates, was supported by results from this study (Table 5.2). A greater complexity in VOC biomarker profiles was observed in Alg+BacEx compared to monocultures of Algae . However, the biomarker count in Bac+AlgEx was comparable to the biomarker count for monocultures of Bacteria . Alg+BacEx also contained more of the markers observed in Co - cultures . The VOC marker profiles are hypothesized to be more similar because the alg ae cell (average size 26 - 28 [28] ) is much larger in volume compared to the bacterial cell (average size 2 - length x 0.3 - mission profile of Co - cultures and Alg+BacEx are similar because the live algae biomarkers dominate th ose of Marinobacter . There was a low overlap of biomarker profiles between the following sample types: a) Algae vs. AlgEx and b) Bacteria vs. BacEx . Four biomarkers were confidently shared between Algae vs AlgEx samples while one shared between Bacteria and BacEx samples, with highest similarity during the 72 hour timepoint. Initially, the biomarker profiles from e xudates were hypothesized to have a large overlap with the biomarker profiles of live organisms. As described in the methods, stock cultures were filtered after 144 hours (early stationary phase) and the 158 Figure 5. 6 Venn diagram of the overlapping VOC bi omarkers for organisms exposed to exudates from the respective other species : ( a ) P. tricornutum exposed to Marinobacter exudates, ( b ) Marinobacter exposed to P. tricornutum exudates. Panel ( c ) evaluates the overlap of Co - cultures , with Alg+BacEx and Bac+A lgEx . All circles are summation of VOC biomarkers across all measured timepoints. The number of replicates for each sample type can be referenced in Table 5.1. exudates were stored. With furt her consideration on the methodology, the vacuum filtration pro cess may have expedited depletion of total VOC levels in the headspace of the filtrate while separating the biomass. Therefore, less abundant and more volatile VOCs may have been inadvertently removed. It is unknown whether the method of filtration and sto rage introduced VOC contaminants, but should be addressed in future experiments by performing the same 159 procedures with a media blank control. Subsequent to the filtration process, storage of ex udates for an extended period of time may have allowed addition al chemistry to take place, e.g.,. secreted non - volatile enzymes in the exudates could have broken down expected biomarkers within the media, even with storage at 20°C. Finally, once the experi ments were started, the exudates were not sampled for VOC emiss ions until 72 hours post - inoculation. Therefore, it is possible that biomarkers of interest were depleted in the media ( e.g., evaporation, increased O 2 or light exposure) prior to this time. Th e absence of living microorganisms from the exudates did not al low VOC markers to be replenished. Experimental evidence for additional chemistry taking place over the SPME sampling periods was seen in the biomarker profiles for AlgEx (Table A.5.4), where t he largest number of compounds were observed at the 240 hour sa mpling timepoint. Regardless, the exuded materials from algae and bacteria are comprised of both volatile and non - volatile compounds, including proteins and lipids. Therefore, even though some volatile biomarkers were depleted or degraded prior to sampling , compounds necessary to affect microorganism metabolism in Alg+BacEx and Bac+AlgEx were hypothesized to be present. Therefore, a concurrent measurement of non - volatile compounds via Liquid Chr omatography (LC) MS would be an appropriate further step to m ore deeply profile exuded materials. Future experiments should prepare exudate samples from the corresponding microorganism as close as possible to the start of the experiment, preferably the s ame day. While the VOC profiles for different growth states of P. tricornutum and related alga have been previously reported, to my knowledge, no study has concurrently studied volatile emission of P. tricornutum interacting with a bacterial partner. Scar ce research has been conducted on the VOC profiles of the genus Marinobacter , with a recent study by Lawson et al. (2020) [29] profiling VOCs from M. adhaerens HP15 interacting with the exudates from the 160 algae Symbiodiniaceae found to be the only published materials. Using thermal desorption, their work identified VOC biomarkers from M. adhaerens to include alcohols, aromatic hydrocarbons, esters, ethers, halogenated hydrocarbons, ketones, and sulfur - containing compounds. Whil e no VOC biomarkers were common between their study and this work, their work observed a greater complexity of bacterial biomarkers in the presence of algal exudates. Similarity to the work presented here underscores the importance of expanding research in to phycosphere - related interactions of algae - bacteria and the effects of exchanged metabolites in ecological systems and/or commercial systems ( e.g., biofuels industry). 5. 4 CONCLUSION This work presented in this chapter aimed to detect and identify VOC biomarkers related to the micro - scale interactions of a model system of P. tricornutum and Marinobacter spp. 3 - 2 and associated cultu res. The presence of Marinobacter spp. 3 - 2, either the primary bacterium or its metabolites in form of exudates from bacte rial cultures , caused unexpected, modest inhibition in the growth rates of P. tricornutum . Substantial differences in VOC biomarker pr ofiles were observed in 1) co - cultures of both organisms, 2) P. tricornutum exposed to Marinobacter spp. 3 - 2 exudates, and 3) Marinobacter spp. 3 - 2 exposed to P. tricornutum exudates, all relative to the VOC biomarker profiles of corresponding monocultures . Increasing the knowledge base of algae - bacterial interactions at the phycosphere and alterations in microorganism physio logy will enable better prediction and/or manipulation of these interactions for commercial purposes as well as a deeper understanding of the basic science of microorganism signaling. Further experiments are necessary to determine if individual biomarkers or compound classes are consistently indicative of monocultures of algae, bacteria, or algae - bacterial 161 interactions. Additionally, a changing complexity was observed in the qualitative profiles of each sample type over time and warrants further investigat ion. This information can in turn provide researchers further details on the transfer of biomarkers between organisms in the phycosphe re. Increasing the number of and combinations of algal and bacterial species, and a larger number of experimental replicat es, will strengthen and solidify the results presented here. Additionally, a large portion of detected biomarkers could not be chemica lly identified. Therefore, utilization of more sensitive and/or selective sampling methodologies and volatile detection me thods will undoubtedly expand upon the VOC biomarker profiles described in this work. Additionally, future work in this area could exp lore complementary techniques to SPME - GC - MS, such as the stable isotopic labeling of metabolic products to trace the origi ns of emitted VOCs. Finally, the approach and technologies developed in this field would also be applicable and transferable to other microbial communities under study, such as cyanobacteria, rhizosphere communities and biofilms. 162 APPENDIX 163 Appendix Tables T able A.5.1 Annotations of VOC biomarkers measured from P. tricornutum samples ( Algae , n=2) detected at se veral timepoints spanning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 44 Amine 2 - Propen - 1 - amine, N - ethyl - N - methyl - 72 773 X 2 108 10 21 X 3 57 **Alkane 1085 X a X a X 4 57 **Alkane 1134 X 5 91 Cyclic Olefin 6 - [( Z ) - 1 - Butenyl] - 1,4 - cycloheptadiene 75 1137 X a X X a X 6 339 1199 X a X 7 71 **Alkane 1231 X a X 8 125 1239 X 9 71 **Alkane 1271 X 10 415 1285 X X a 11 57 **Alkane 1316 X a X a X 12 401 1423 X a X 13 135 1473 X 14 177 **Carotenoid - ionone) 1495 X 164 T able A.5.1 15 58 Methyl Ketone 2 - Pentadecanone, 6,10,14 - trimethyl - 80 1863 X a X X X 16 149 1 893 X 17 149 2495 X a X 18 73 2572 X a X X X a 19 73 2626 X X a : Detection in 1 of 2 replicates **: Compound class assigned using characteristic fragmentation pattern 165 Table A.5. 2 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 samples ( Bacteria , n=2) detected at several timepoints spanning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 42 815 X 2 68 Terpene D - Limonene 74 1015 X 3 119 1065 X 4 57 **Alkane 1085 X a X 5 117 1133 X 6 95 1151 X a X 7 119 1159 X X a 8 401 1423 X a X 9 135 1473 X 10 57 **Alkane 1816 X 11 73 1912 X X X X a : Detection in 1 of 2 replicates **: Compound class assigned using characteristic fragmentation pattern 166 T able A.5. 3 Annotations of VOC biomarkers measured from P. tricornutum and Marinobacter spp. 3 - 2 samples ( Co - cultures , n =3) detected at several timepoints spanning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 42 815 X b 2 133 909 X b 3 68 Alipha tic D - Limonene 74 1015 X b 4 108 1021 X b 5 99 1028 X b 6 91 Cyclic Olefin 6 - [( Z ) - 1 - Butenyl] - 1,4 - cycloheptadiene 75 1137 X X X X 7 95 1151 X a X b 8 119 1158 X b 9 133 1163 X b 10 86 Diol Dianhydromannito l 73 1195 X b 11 137 **Carotenoid - cyclocitral) 1208 X a X 12 71 **Alkane 1271 X b 13 57 **Alkane 1372 X b X b 14 55 1473 X b 15 177 **Carotenoid - ionone) 1495 X b 16 135 1644 X 167 T able A.5.3 17 73 1770 X b 18 58 Methyl Ketone 2 - Pentadecanone, 6,10,14 - trimethyl - 80 1863 X b X a X X 19 100 1900 X b 20 73 1912 X a X b 21 71 **Alkane 1962 X b 22 55 Carboxylic acid Hexadecenoic acid (16:1) 80 1963 X b 23 73 Carboxyl ic acid n - Hexadecanoic acid 80 1982 X b 24 73 2454 X b 25 73 2572 X X a 26 73 2572 X a X b 27 208 2678 X a X b 28 97 2814 X b 29 248 3026 X a X b 30 97 3057 X b 31 248 3304 X b X a : Detect ion in 1 of 3 replicates **: Compound class assigned using characteristic fragmentation pattern X b : Detection in 2 of 3 replicates 168 Table A.5. 4 Annotations of VOC biomarkers measured from P. tricornutum exudates ( AlgEx , n=2) detected at several timepoint s spanning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 79 744 X 2 41 814 X 3 45 958 X 4 95 1151 X X a X 5 285 1220 X 6 68 1244 X 7 71 **Alkane 1271 X a X 8 57 **Alkane 1290 X 9 92 1340 X X a 10 154 1372 X 11 401 1424 X 12 135 1473 X 13 57 **Alkane 1692 X a X a X 14 57 **Alkane 1811 X a X a X 15 73 1912 X a X X 16 55 Carboxylic Acid Hexadecenoic acid (16:1) 80 1963 X 17 73 Carboxylic Acid n - Hexadecanoic acid 80 1982 X 169 Table A.5.4 18 73 2572 X a X X a 19 248 2901 X a X a : Detection in 1 of 3 replicates **: Compound class assigned using characteristic fragmentation pattern X b : Detection in 2 of 3 replicates 170 Table A.5. 5 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 exudate s ( BacEx , n=2) detected at several timepoints span ning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 103 976 X 2 121 981 X X X 3 109 998 X X X X 4 123 1078 X X X a 5 267 1101 X X a 6 127 1126 X 7 68 1244 X X X X 8 401 1424 X 9 55 1479 X 10 57 **Alkane 1521 X 11 216 1617 X 12 253 1627 X a X X a X 13 405 1690 X 14 71 **Alkane 1759 X 15 1 94 1889 X X a 16 73 1912 X a X 17 58 1924 X 171 Table A.5.5 18 71 1945 X 19 73 2142 X 20 73 2572 X X X a X a 21 73 2572 22 208 2697 X 23 248 3095 X 24 97 315 3 X X a : Detection in 1 of 2 replicates **: Compound class assigned using characteristic fragmentation pattern 172 Table A.5.6 Annotations of VOC biomarkers measured from P. tricornutum + Marinobacter spp. 3 - 2 exudates ( Alg +BacEx , n =3) detected at s everal timepoints spanning 240 hours of sample growth Compound # Base Peak m/z Tentative Compound Class NIST14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 45 768 X b 2 69 773 X b 3 42 815 X a X b 4 71 **Alkane 816 X b 5 42 945 X b 6 105 963 X a X a X b 7 121 981 X X a X b X a 8 57 988 X b X a 9 109 998 X X X b X b 10 108 1021 X b X 11 71 1029 X b 12 123 1078 X X b X a X a 13 57 **Alkane 1085 X b 14 57 **Alkane 1086 X a X b 15 91 Cyclic Olefin 6 - [( Z ) - 1 - Butenyl] - 1,4 - cycloheptadiene 75 1137 X b X b X b X b 16 95 1151 X b X b 173 Table A.5.6 17 105 1158 X b 18 119 1159 X a X b 19 57 **Alkane 1185 X b 20 137 **Ca rotenoid - cyclocitral) 1208 X b 21 68 1244 X X X X 22 122 1253 X b X b X X 23 71 **Alkane 1271 X b X b 24 57 **Alkane 1290 X a X b 25 133 1328 X b X a 26 109 1418 X b 27 401 1423 X a X b X a 28 135 1473 X b 29 55 1473 X a X b 30 205 Quinone 2,5 - di - tert - Butyl - 1,4 - benzoquinone 86 1475 X b 31 91 1500 X b 32 216 1617 X 33 405 1690 X b X b 34 73 1771 X b 35 57 **Alkane 1816 X b 174 Table A.5.6 36 4 21 1837 X b 37 58 Methyl Ketone 2 - Pentadecanone, 6,10,14 - trimethyl - 80 1863 X a X a X a X b 38 73 1912 X a X b 39 58 1924 X b X a X a 40 71 1945 X b 41 71 **Alkane 1948 X a X b 42 73 Carboxylic Acid n - Hexadecanoic acid 80 1982 X a X b 43 73 2142 X 44 135 2472 X b 45 73 2572 X b X X a 46 73 2572 X a X X b 47 97 2859 X a X b 48 97 3106 X b 49 97 3159 X b X a X a : Detection in 1 of 3 replicates **: Compound class assigned using cha racteristic fragmentation pattern X b : Detection in 2 of 3 replicates 175 Table A.5.7 Annotations of VOC biomarkers measured from Marinobacter spp. 3 - 2 + P. tricornutum exudates ( Bac+AlgEx , n=3) detected at several timepoints spanning 240 hours of sample grow th Compound # Base Peak m/z Tentative Compound Class NIST 14 ID NIST % Match Experimental Retention Index Day 3 Day 5 Day 7 Day 10 1 42 815 X b X a 2 91 947 X b 3 103 976 X b 4 121 981 X b X X a X a 5 267 998 X X b 6 59 1003 X b X a 7 57 **Alkane 1021 X b 8 99 1028 X b 9 57 **Alkane 1085 X b X a 10 61 1118 X b 11 95 1131 X b 12 68 1244 X b X X b 13 327 1318 X b 14 92 1340 X b 15 57 **Alkane 1348 X b 16 154 1372 X 17 170 1400 X b 176 Table A.5.7 18 401 1424 X b 19 343 1452 X b X a 20 91 1476 X b 21 405 1634 X b 22 236 1676 X b 23 57 **Alkane 1692 X b X a 24 71 **Alkane 175 9 X b 25 58 1924 X b X a 26 73 2142 X X a X a 27 97 2610 X b 29 208 2689 X b X a : Detection in 1 of 3 replicates **: Compound class assigned using characteristic fragmentation pattern X b : Detection in 2 of 3 replicat es 177 REFERENCES 178 REFERENCES 1 Amin, S. A., Parker, M. S. & Armbrust, E. V. Interactions between diatoms and bacteria. Microbiol Mol Biol Rev 76 , 667 - 684, doi:10.1128/MMBR.00007 - 12 (2012). 2 Amin, S. A. et al. Interaction a nd signalling between a cosmopolitan phytoplankton and associate d bacteria. Nature 522 , 98 - 101, doi:10.1038/nature14488 (2015). 3 Seyedsayamdost, M. R., Wang, R., Kolter, R. & Clardy, J. Hybrid biosynthesis of roseobacticides from algal and bacterial prec u rsor molecules. Journal of the American Chemical Society 136 , 15 150 - 15153, doi:10.1021/ja508782y (2014). 4 Seymour, J. R., Amin, S. A., Raina, J. B. & Stocker, R. Zooming in on the phycosphere: the ecological interface for phytoplankton - bacteria relations h ips. Nature microbiology 2 , 17065, doi:10.1038/nmicrobiol.2017.6 5 (2017). 5 Le Chevanton, M. et al. Screening and selection of growth - promoting bacteria for Dunaliella cultures. Algal Research 2 , 212 - 222, doi: 10.1016/j.algal.2013.05.003 (2013). 6 Kim, B. - H ., Ramanan, R., Cho, D. - H., Oh, H. - M. & Kim, H. - S. Role of Rhizobium, a plant growth promoting bacterium, in enhancing algal biomass through mutualistic interaction. Biomass and Bioenergy 69 , 95 - 105, doi: 10.1016/j.biombioe.2014.07.015 (2014). 7 Wa ng, H., H ill, R. T., Zheng, T., Hu, X. & Wang, B. Effects of bacterial communities on biofuel - producing microalgae: stimulation, inhibition and harvesting. Critical Reviews in Biotechnology 36 , 341 - 352, doi:10.3109/07388551.2014.961402 (2016). 8 M ayali , X. & Azam , F. Algicidal b acteria in the s ea and their i mpact on a lgal b looms1. Journal of Eukaryotic Microbiology 51 , 139 - 144, doi:10.1111/j.1550 - 7408.2004.tb00538.x (2004). 9 Paul, C., Mausz, M. A. & Pohn ert, G. A co - culturing/metabolomics approach to investigate c h emically mediated interactions of planktonic organisms reveals influence of bacteria on diatom metabolism. Metabolomics 9 , 349 - 359, doi:10.1007/s11306 - 012 - 0453 - 1 (2013). 10 Zerrifi, S. E. A., El Khalloufi, F., Oudra, B. & Vasconcelos, V. Seaweed Bioactive Compounds against Pathogens and Microalgae: Potential Uses on Pharmacology and Harmful Algae Bloom Control. Mar Drugs 16 , 55, doi:10.3390/md16020055 (2018). 11 Yang, Y. H. et al. Fatty a cid and l ipid c lass c omposition of the m icroalga Phaeodactylum tricor n utum. Journal of oleo science 66 , 363 - 368, doi:10.5650/jos.ess16205 (2017). 12 Gauthier, M. J. et al. Marinobacter hydrocarbonoclasticus gen. nov., sp. nov., a n ew, e xtremely h alotolerant, h ydro carbon - d egrading m arine b acterium. International Journal 179 of S y stematic and Evolutionary Microbiology 42 , 568 - 576, doi: 10.1099/00207713 - 42 - 4 - 568 (1992). 13 Reese, K. L. et al. Chemical p rofiling of v olatile o rganic c ompounds in the h eadspace of a lgal c ultures as e arly b iomarkers of a lgal p ond c rashes. Scientific Repo r ts 9 , 13866, doi:10.1038/s41598 - 019 - 50125 - z (2019). 14 Pohnert, G. & Boland, W. The oxylipin chemistry of attraction and defense in brown algae and diatoms. Natural product reports 19 , 108 - 122, doi:10.1039/a806888g (2002). 15 Prestegard, S. K. et al. Spec i fic m etabolites in a Phaeodactylum tricornutum s train i solated from w estern n orwegian f jord w ater. Mar Drugs 14 , 9 (2016). 16 Lupette, J. et al. The architecture of lipid droplets in the di atom Phaeodactylum tricornutum. Algal Research 38 , 101415, doi: 10. 1 016/j.algal.2019.101415 (2019). 17 Breuer, G., Lamers, P. P., Martens, D. E., Draaisma, R. B. & Wijffels, R. H. The impact of nitrogen starvation on the dynamics of triacylglycerol accumula tion in nine microalgae strains. Bioresource technology 124 , 217 - 2 2 6, doi:10.1016/j.biortech.2012.08.003 (2012). 18 Liang, Y., Beardall, J. & Heraud, P. Effects of nitrogen source and UV radiation on the growth, chlorophyll fluorescence and fatty acid composition of Phaeodactylum tricornutum and Chaetoceros muelleri (Bac i llariophyceae). Journal of photochemistry and pho tobiology. B, Biology 82 , 161 - 172, doi:10.1016/j.jphotobiol.2005.11.002 (2006). 19 Cezare - Gomes, E. A. et al. Potential of m icroalgae c arotenoids for i ndustrial a pplication. Applied Biochemistry and Biotech n ology 188 , 602 - 634, doi:10.1007/s12010 - 018 - 02945 - 4 (2019). 20 Abdel - Aal, E. I., Haroon, A. M. & Mofeed, J. Successive solvent extraction and GC MS analysis for the evaluation of the phytochemical constituents of the filamentous green alga Spirogyra longat a . The Egyptian Journal of Aquatic Research 41 , 23 3 - 246, doi: 10.1016/j.ejar.2015.06.001 (2015). 21 Kamenarska, Z. et al. Chemical composition and biological activity of the brackish - water green alga Cladophora rivularis (L.) Hoek. 47 , 215, doi: 10.1515/BOT . 2004.022 (2004). 22 Kamenarska, Z. et al. Chemica l composition of the brown alga Padina pavonia (L.) Gaill. from the Adriatic sea. Botanica Marina 45 , 339 - 345, doi:10.1515/BOT.2002.034 (2002). 23 Gobalakrishnan, R. Antimicrobial potential and bioactive co n stituents from aerial parts of Vitis setosa Wall. Journal of Medicinal Plants Research 8 , 454 - 460, doi:10.5897/JMPR2013.5356 (2014). 24 hydrocarbon production in the diatom Phaeo d actylum tricornutum due to growth 180 temperature. Jo urnal of Applied Phycology 27 , 1463 - 1472, doi:10.1007/s10811 - 014 - 0464 - 6 (2015). 25 Srikanta Dani, K. G. et al. Relationship between isoprene emission and photosynthesis in diatoms, and its implications for g lobal marine isoprene estimates. Marine Chemistry 189 , 17 - 24, doi: 10.1016/j.marchem.2016.12.005 (2017). 26 Wolfram, S. et al. A Metabolic p robe - e nabled s trategy r eveals u ptake and p rotein t argets of p olyunsaturated a ldehydes in the d iatom Phaeodactylum tr i cornutum. PLoS One 10 , e0140927, doi:10.1371/journal.pone.0140927 (2015). 27 Spielmeyer, A. & Pohnert, G. Influence of temperature and elevated carbon dioxide on the production of dimethylsulfoniopropionate and glycine betaine by marine phytoplankton. Mar i ne environmental research 73 , 62 - 69, doi:10.1016/j.marenvres.2011.11.002 (2012). 28 Renzi, M., Roselli, L., Giovani, A., Focardi, S. E. & Basset, A. Early warning tools for ecotoxicity assessment based on Phaeodactylum tricornutum. Ecotoxicology (London, E ngland) 23 , 1055 - 1072, doi:10.100 7/s10646 - 014 - 1249 - z (2014). 29 Lawson, C. A., Seymour, J. R., Possell, M., Suggett, D. J. & Raina, J. - B. The v olatilomes of Symbiodiniaceae - a ssociated b acteria Are i nfluenced by c hemicals d erived f rom t heir a lgal p artner. F rontiers in Marine Science 7 , doi :10.3389/fmars.2020.00106 (2020). 181 Chapter 6: C ONCLUSIONS AND BROADER IMPACTS From biosecurity to biodefense, SPME - GC - MS has proven useful for profiling volatile biomarkers emitted from actively growing microbial culture s . The presence and physiological growth states for monocultures of Bacillus anthracis and Francisella tularensis in optimal growth media were reflected in changes in the corresponding headspace composition . VOCs diagnostic of the health states of biofuel - r elevant algae were detected in the contexts of stress/predator grazing and for interactions with neighboring bact eria, both with the end goal of improving the yields of biomass from algae cultures. However, there are additional aspects to consider before t he results of this work could be implemented. While the data acquisition and data processing employed throughout this dissertation has proven capable of VOC marker collection, separation, detection, and annotation, a number of biomarkers could not be iden t ified compared to known entries in the NIST14 mass spectral database. An inability to identify these biomarkers m ay be a result of poor spectral qualit y for either the experimental data or the reference library , though it is also likely that many VOCs are novel compounds yet to be identified or added to spectral databases . Standardization of sample collection and dat a processing practices within the respective fields would facilitate inter - laboratory comparisons, thus increasing success rates in de - replica t ing whether biomarkers observed in this research have been previously observed. In a similar vein, exploring the use of a high - resolution mass spectrometer could improve confidence in annotations of unknown . Determination of exact ma s s for molecular and/or fragment ions would allow estimation of chemical formula, even if fragmentation patterns d o not match up with empirical databases or spectra predicted for unknowns in silico . Moreover, co - eluting 182 compounds with the same nominal mass but different chemical formulas (isobars) could be distinguished. Generalization of VOCs as diagnostic biomarker profiles for a given taxa requires investigation of a wider range of model microbes. In the research of biological warfare agents, define d sur r ogate organisms are used in place of a target pathogen due to safety requirements as well as strict regulations and laws. As seen in Chapter 3, VOC biomarker profiles were more similar for fully virulent Risk Group 3 (RG3) Bacillus anthracis Ames and its s urrogate Risk Group 2 (RG2) Bacillus anthracis Sterne compared to RG3 Francisella tularensis SCHU S4 and its surrogate RG2 Francisella tularensis novicida. Future VOC profiling should focus on the predictive strength of profiles generated for RG2 sur rogat e s , in particular those RG2 species that closely reproduce the phenotypes of RG3 virulent pathogens . The use of surrogates has implications in the protocols and personal protective equipment ( PPE ) used to safely handle different microorganisms. If sui table surrogates are not available, the protocols used in this work for the safe transfer of materials from BSL - 3 to BSL - 2 laboratories, which preserved key metabolites while avoiding destructive decontamination procedures, would be useful to BSL - 3 practit ioner s . Additional studies into the effects of different cellular environments on pathogen VOC profiles, in particular for growth conditions mimicking human - derived materials ( e.g. , cell lines) , should be investigated . For algal research , better understanding o f the relationship between volatile emissions, growth rates, and lipid production in different biofuel - relevant algae species would allow commercial producers to tailor biomass output. Different types of stress on algal ponds should be inves tigated to dete r mine how various triggers activate intrinsic metabolic pathways. Finally, research into changes in algal metabolism for different algal genotypes, both with and without the presence of bacteria, should be conducted, as some combinations may 183 produce protec t ive measures ( e.g., comb at the actions of a predator, etc.). Analysis from a larger number of alga l taxa should explore whether diagnostic marker s are specific to a given species or more generalizable to a genus. Experimental results can be validated usin g analytical standards to confirm the identi ties of biomarkers detected via mass spectrometry for inclusion in targeted methodologies. Future work should explore metabolic profiling using conditions more analogous to real - world applications. Experiments in laboratory settings are the most easily controlled, where conditions can be well - defined and any extraneous influences can be minimized. Translat ing that knowledge to microorganisms in their natural habitat or in large - scale, commercial production, where c onditions are ill - defined and can be externally contaminated, can provide researchers with selective and specific tools to target unhealthy ponds . E xperiments are necessary to determine if individual VOC markers or their compound classes are consistently i ndicative of the pathogens, algae, or bacteria studied across this work. Subsequently, knowledge gained and biomarkers annotated from untargeted metabolomics efforts may guide development of targeted metabolite monitoring for changes in diagnostic chemica l signatures and detecting volatile signals of metabolic changes in real - time. Miniaturized GC - MS systems for field deployable detector systems pr ovide one such technology currently under development. Applications utilizing such systems would require initi a l validation of accuracy and reproducibility of qualitative and quantitative analyses, for both analytical standards and for complex mixtures. The results presented in this dissertation present further research avenues and have implications for the wide r research communities with interests in microbial detection and volatilomics. Although the methodology based on SPME - GC - quadrupole MS in scanning mode does not represent the pinnacle of sensitivity, a number of relatively robust and reproducible 184 putative v olatile biomarkers could be detected in each study. Non - invasive monitoring of VOC biomarkers from the headspace of liquid cultures allows direct sampling with minimal perturbation of the culture. While the methods here applied to the headspace of liquid c ultures, this method could also apply for the headsp ace of other microbial communities, such as biofilms, cyanobacteria, and rhizosphere communities.