BUILDING a RISK ASSESSMENT MODEL FOR MANAGEMENT OF PERSISTENT ENDODONTIC LESIONS

Copyright by Kee Zhi Ling Michelle 2015 Abstract Olfactory studies in zebrafish have provided enormous insights into how olfaction occurs in an aquatic environment, triggering innate and stereotypical responses that allow survival in the vast habitat One of the odors detected by zebrafish is an alarm substance from the lesioned skin of conspecifics, known as Schreckstoff A feature of the alarm substance is increased potency caused by heating, and glycan content To further characterize Schreckstoff, I have found that hyaluronan (HA), a simple linear polysaccharide, which is broken to active signaling fragments by heat, evokes activity in the olfactory system of larval zebrafish Through the use of wide-field fluorescence microscopy to perform calcium imaging on transgenic zebrafish expressing the genetically encoded calcium indicator GCaMP, I demonstrated that HA is able to evoke calcium responses in the olfactory epithelium and bulb of zebrafish as young as days post fertilization This suggests HA can also function as an odor, in addition to its diverse size-dependent roles in cell signaling Additionally, I also describe and demonstrate the use of a microfluidic chip, as a “proof-ofconcept” to aid in characterization of the olfactory sensory neurons that are responsible for detecting HA iv Dedicated in loving memory of Yeye “It doesn’t matter how long you take, but learn from the experience and enjoy the process.” Thank you Yeye, I finally did it v Acknowledgement This thesis would not be possible without the following people, to which my heartfelt gratitude and appreciation cannot be expressed in mere written words: To Associate Professor Suresh Jesuthasan – I thank you sincerely for your continuous mentorship, patience, guidance and support since 2007, especially during the course of my degree and my two pregnancies Research aside, you will always be the source of my motivation to seek maternal rights even as a PhD student To my thesis advisory committee members, Assistant Professor Marc Fivaz, Assistant Professor Adam Claridge-Chang, Associate Professor William F Burkholder – My sincere thanks and gratitude for your invaluable knowledge, advice and critical suggestions throughout these years To my colleagues in SJ lab - Thank you all for sharing your wealth of knowledge and advice, in all aspects of science and life Dr Ajay Mathuru, thank you for being my “personal critic” Special thanks to my fellow graduate student - Joanne, Lin Qian, Charlie and Mahathi - for keeping me healthy and mentally alive with our daily “4 p.m chocolate and apple time” Thank you all members of SJ lab, for converting me to a caffeine addict, without which I would probably not appreciate the close relationship between coffee and science To the members of MSB lab, both past and present – Thank you all for sharing your invaluable knowledge and advice on qPCR, microfluidics and RNA sequencing vi Last but not the least, I would like to express my deepest gratitude and appreciation for my dear family members, especially to my parents for finally supporting me in neuroscience endeavors; my husband for never failing to listen to my thoughts and providing me with a loving refuge from science; my lovely sons, for keeping me grounded, sane and youthful in the lab when the going gets tough and inspiring me to understand cognitive development, processing and functions in young children vii Contents Title Page i Abstract Signature ii Copyright iii Abstract iv Dedication v Acknowledgements vi Table of Contents viii List of Tables ix List of Figures x Chapter Introduction 1.1 Difference between olfaction and gustation 1.2 Olfaction 1.2.1 How olfaction differs in terrestrial and aquatic animals 1.2.2 Olfaction begins in utero 1.2.3 Food, foraging and homing odors 1.2.4 Olfaction in social context 1.2.4.1 Recognizing individuals 1.2.4.2 Recognizing kin 1.2.4.2.1 Kin recognition and mate choice – role of MHC 10 1.2.4.2.2 Kin recognition and mate choice – role of major urinary proteins 11 1.2.4.3 Recognizing predators 12 1.3 Recognizing danger – smell of threatening situations 13 1.3.1 Source of Schreckstoff 14 1.3.2 Components of Schreckstoff 14 1.4 Hyaluronan 17 1.4.1 Introduction of hyaluronan 17 1.4.2 Role of HA depends on its polymer size 18 viii 1.4.3 Role of HA depends on its binding proteins 20 1.4.4 HA degradation 21 1.4.4.1 Enzymatic fragmentation of HA 21 1.4.4.2 Thermal fragmentation of HA 22 1.4.4.3 Degradation of HA by free radicals 23 1.5 Introduction to zebrafish olfactory system 24 1.5.1 Advantages of zebrafish as a model system 24 1.5.2 Anatomy of the zebrafish olfactory system 25 1.5.3 Types of OSNs 26 1.5.4 Olfactory bulb 30 1.5.4.1 Odor coding in the olfactory bulb 32 1.5.5 Projection from OB to higher brain centers 33 1.6 Aims for this thesis 36 Chapter Zebrafish detect hyaluronan as an odor 37 2.1 Abstract 37 2.2 Introduction 37 2.3 Material and Methods 40 2.3.1 Animals 40 2.3.2 Skin extract preparations 41 2.3.3 Calcium imaging experiments 41 2.3.4 Odors stimulation 42 2.3.5 Calcium image processing and analysis 43 2.3.6 Dissection of olfactory epithelia 44 2.3.7 Confocal imaging 44 2.4 Results 45 2.4.1 Experimental workflow 45 2.4.2 Et(sqKR15-3A) labels olfactory sensory neurons 47 2.4.3 OSNs respond distinctly when exposed to separate odors 50 ix 2.4.4 A component in the WGA column elution buffer activates calcium activity in OSNs 56 2.4.5 Different sizes of HA activates the olfactory bulb 57 2.4.6 Distinct glomeruli can be identified with Tg(NBT:GCaMP5) 60 2.4.7 Exposure to a range of HA sizes at similar concentrations displayed distinct responses over time 62 2.4.8 Nanomolar concentration range of HA-L inhibited medial anterior glomeruli of the olfactory bulb 65 2.4.9 Similar concentrations of HA-L and HA-M evoked activity at maG, mdG and vmG clusters 69 2.4.10 Exposure to HA-S evoked excitatory activity specifically in mdG and vmG clusters 71 2.4.11 HA evoked similar glomerular activities across sizes, which were distinct from glomerular activities after L-lysine stimulation 75 2.4.12 Different sizes of HA at similar concentrations activated OSNs 79 2.4.13 The olfactory bulb remains active to HA exposure after two trials 80 2.5 Discussion 83 Chapter A microfluidic device to sort olfactory sensory neurons based on dynamic response to different odors 95 3.1 Abstract 95 3.2 Introduction 96 3.3 Material and Methods 98 3.3.1 Microfluidic device fabrication 98 3.3.2 Isolation and dissociation of olfactory epithelial cells 98 3.3.3 Cell viability assay 99 3.3.4 Calcium indicator dyes 99 3.3.5 Ionophore and odorants stimulation 100 3.3.6 Image acquisition and processing 100 3.3.7 qPCR of single OSN 101 3.4 Results 102 3.4.1 Microfluidic device design 102 x Reusch TB, Häberli MA, Aeschlimann PB, Milinski M (2001) Female sticklebacks count alleles in a strategy of sexual selection explaining MHC polymorphism Nature 414:300–302 Reutter K, Pfeiffer W (1973) Fluorescence microscopical demonstration of the alarm substance in the alarm substance cells of the European minnow J Comp Physiol 82:411–418 Ringach DL (2009) Spontaneous and driven cortical activity: implications for computation Curr Opin Neurobiol 19:439–444 Rolls BJ, Rolls ET, Rowe EA, Sweeney K (1981) Sensory specific satiety in man☆ Physiol Behav 27:137–142 Rooney P, Wang M, Kumar P, Kumar S (1993) Angiogenic oligosaccharides of hyaluronan enhance the production of collagens by endothelial cells J Cell Sci 105 ( Pt 1:213–218 Sander JD, Cade L, Khayter C, Reyon D, Peterson RT, Joung JK, Yeh J-RJ (2011) Targeted gene disruption in somatic zebrafish cells using engineered TALENs Nat Biotechnol 29:697–698 Sato Y, Miyasaka N, Yoshihara Y (2005) Mutually exclusive glomerular innervation by two distinct types of olfactory sensory neurons revealed in transgenic zebrafish J Neurosci 25:4889–4897 Sato Y, Miyasaka N, Yoshihara Y (2007) Hierarchical regulation of odorant receptor gene choice and subsequent axonal projection of olfactory sensory neurons in zebrafish J Neurosci 27:1606–1615 Sattar A, Rooney P, Kumar S, Pye D, West DC, Scott I, Ledger P (1994) Application of Angiogenic Oligosaccharides of Hyaluronan Increases Blood Vessel Numbers in Rat Skin J Invest Dermatol 103:576–579 Savani RC, Cao G, Pooler PM, Zaman A, Zhou Z, DeLisser HM (2001) Differential involvement of the hyaluronan (HA) receptors CD44 and receptor for HAmediated motility in endothelial cell function and angiogenesis J Biol Chem 276:36770–36778 Schaal B, Marlier L, Soussignan R (1998) Olfactory function in the human fetus: evidence from selective neonatal responsiveness to the odor of amniotic fluid Behav Neurosci 112:1438–1449 Schaal B, Marlier L, Soussignan R (2000) Human foetuses learn odours from their pregnant mother’s diet Chem Senses:729–737 Schaal B, Orgeur P, Arnould C (1995) Olfactory Preferences in Newborn Lambs : Possible Influence of Prenatal Experience Behaviour 132:351–365 Scheibner KA, Lutz MA, Boodoo S, Fenton MJ, Powell JD, Horton MR (2006) Hyaluronan fragments act as an endogenous danger signal by engaging TLR2 J Immunol 177:1272–1281 130 Shephard KL (1994) Functions for fish mucus Rev Fish Biol Fish 4:401–429 Sherborne AL, Thom MD, Paterson S, Jury F, Ollier WER, Stockley P, Beynon RJ, Hurst JL (2007) The genetic basis of inbreeding avoidance in house mice Curr Biol 17:2061–2066 Shpak G, Zylbertal A, Yarom Y, Wagner S (2012) Calcium-activated sustained firing responses distinguish accessory from main olfactory bulb mitral cells J Neurosci 32:6251–6262 Slevin M, Kumar S, Gaffney J (2002) Angiogenic oligosaccharides of hyaluronan induce multiple signaling pathways affecting vascular endothelial cell mitogenic and wound healing responses J Biol Chem 277:41046–41059 Smith TE, Tomlinson AJ, Mlotkiewicz JA, Abbott DH (2001) Female marmoset monkeys (Callithrix jacchus) can be identified from the chemical composition of their scent marks Chem Senses 26:449–458 Smotherman WP (1982) Odor aversion learning by the rat fetus Physiol Behav 29:769–771 Smotherman WP, Robinson SR (1990) Olfactory bulb transection alters fetal behavior after chemosensory but not tactile stimulation Dev Brain Res 57:175– 180 Sorensen PW, Fine JM, Dvornikovs V, Jeffrey CS, Shao F, Wang J, Vrieze LA, Anderson KR, Hoye TR (2005) Mixture of new sulfated steroids functions as a migratory pheromone in the sea lamprey Nat Chem Biol 1:324–328 Stephenson JF, Partridge JC, Whitlock KE (2012) Food and conspecific chemical cues modify visual behavior of zebrafish, Danio rerio Zebrafish 9:68–73 Stern R (2004) Hyaluronan catabolism: a new metabolic pathway Eur J Cell Biol 83:317–325 Stopfer M, Jayaraman V, Laurent G (2003) Intensity versus identity coding in an olfactory system Neuron 39:991–1004 Stosiek C, Garaschuk O, Holthoff K, Konnerth A (2003) In vivo two-photon calcium imaging of neuronal networks Proc Natl Acad Sci U S A 100:7319–7324 Sugahara KN, Murai T, Nishinakamura H, Kawashima H, Saya H, Miyasaka M (2003) Hyaluronan oligosaccharides induce CD44 cleavage and promote cell migration in CD44-expressing tumor cells J Biol Chem 278:32259–32265 Sullivan C, Charette J, Catchen J, Lage CR, Giasson G, Postlethwait JH, Millard PJ, Kim CH (2009) The gene history of zebrafish tlr4a and tlr4b is predictive of their divergent functions J Immunol 183:5896–5908 Sullivan RM, Wilson DA (1991) Neural correlates of conditioned odor avoidance in infant rats Behav Neurosci 105:307–312 131 Sun L, Müller-Schwarze D (1999) Chemical signals in the beaver: one species, two secretions, many functions? In: Advances in Chemical Signals in Vertebrates (Johnston RE, Müller-Schwarze D, Sorensen PW, eds), pp 281–288 Boston, MA: Springer US Swedlow JR, Platani M (2002) Live cell imaging using wide-field microscopy and deconvolution cell engineering for in vivo fluorescence imaging Cell Struct Funct 341:335–341 Symonds MRE, Elgar MA (2008) The evolution of pheromone diversity Trends Ecol Evol 23:220–228 Tabor R, Yaksi E, Weislogel J-M, Friedrich RW (2004) Processing of odor mixtures in the zebrafish olfactory bulb J Neurosci 24:6611–6620 Tallini YN, Ohkura M, Choi B-R, Ji G, Imoto K, Doran R, Lee J, Plan P, Wilson J, Xin H-B, Sanbe A, Gulick J, Mathai J, Robbins J, Salama G, Nakai J, Kotlikoff MI (2006) Imaging cellular signals in the heart in vivo: Cardiac expression of the high-signal Ca2+ indicator GCaMP2 Proc Natl Acad Sci U S A 103:4753–4758 Tang-Martinez Z (2001) The mechanisms of kin discrimination and the evolution of kin recognition in vertebrates: a critical re-evaluation Behav Processes 53:21– 40 Tay TL, Ronneberger O, Ryu S, Nitschke R, Driever W (2011) Comprehensive catecholaminergic projectome analysis reveals single-neuron integration of zebrafish ascending and descending dopaminergic systems Nat Commun 2:171 Taylor KR, Trowbridge JM, Rudisill JA, Termeer CC, Simon JC, Gallo RL (2004) Hyaluronan fragments stimulate endothelial recognition of injury through TLR4 J Biol Chem 279:17079–17084 Teh C, Chudakov DM, Poon K-L, Mamedov IZ, Sek J-Y, Shidlovsky K, Lukyanov S, Korzh V (2010) Optogenetic in vivo cell manipulation in KillerRed-expressing zebrafish transgenics BMC Dev Biol 10:110 Teicher MH, Blass EM (1977) First suckling response of the newborn albino rat: the roles of olfaction and amniotic fluid Science 198:635–636 Termeer C, Benedix F, Sleeman J, Fieber C, Voith U, Ahrens T, Miyake K, Freudenberg M, Galanos C, Simon JC (2002) Oligosaccharides of Hyaluronan Activate Dendritic Cells via Toll-like Receptor Assessment 195 Tian L, Hires SA, Mao T, Huber D, Chiappe ME, Chalasani SH, Petreanu L, Akerboom J, McKinney SA, Schreiter ER, Bargmann CI, Jayaraman V, Svoboda K, Looger LL (2009) Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators Nat Methods 6:875–881 Todd JH, Atema J, Bardach JE (1967) Chemical Communication in Social Behavior of a Fish, the Yellow Bullhead (Ictalurus natalis) Science (80- ) 158:672–673 132 Todrank J, Busquet N, Baudoin C, Heth G (2005) Preferences of newborn mice for odours indicating closer genetic relatedness: is experience necessary? Proc Biol Sci 272:2083–2088 Toole BP (2009) Hyaluronan-CD44 Interactions in Cancer: Paradoxes and Possibilities Clin Cancer Res 15:7462–7468 Treisman M (1964) Noise and Weber’s law: The discrimination of brightness and other dimensions Psychol Rev 71:314–330 Tricas TC, Kajiura SM, Summers AP (2009) Response of the hammerhead shark olfactory epithelium to amino acid stimuli J Comp Physiol A Neuroethol Sens Neural Behav Physiol 195:947–954 Varendi H, Porter R, Winberg J (1996) Attractiveness of amniotic fluid odor: evidence of prenatal olfactory learning? Acta Paediatr 85:1223–1227 Verheggen FJ, Haubruge E, Mescher MC (2010) Alarm pheromones-chemical signaling in response to danger Vladimirov N, Mu Y, Kawashima T, Bennett D V, Yang C-T, Looger LL, Keller PJ, Freeman J, Ahrens MB (2014) Light-sheet functional imaging in fictively behaving zebrafish Nat Methods 11:1–2 Volpi N (2007) On-line HPLC/ESI-MS separation and characterization of hyaluronan oligosaccharides from 2-mers to 40-mers Anal Chem 79:6390–6397 Volpi N, Zhang Z, Linhardt RJ (2008) Mass spectrometry for the characterization of unsulfated chondroitin oligosaccharides from 2-mers to 16-mers Comparison with hyaluronic acid oligomers Rapid Commun Mass Spectrom 22:3526–3530 Von Frisch K (1938) Zur Psychologie des Fisch-Schwarmes Naturwissenschaften 26:601–606 Von Frisch K (1942) Über einen Schreckstoff der Fischhaut und seine biologische Bedeutung Z Vgl Physiol 29:46–145 Wang HW, Wysocki CJ, Gold GH (1993) Induction of olfactory receptor sensitivity in mice Science 260:998–1000 Wang M-J, Jeng K-CG, Kuo J-S, Chen H-L, Huang H-Y, Chen W-F, Lin S-Z (2004) cJun N-terminal kinase and, to a lesser extent, p38 mitogen-activated protein kinase regulate inducible nitric oxide synthase expression in hyaluronan fragments-stimulated BV-2 microglia J Neuroimmunol 146:50–62 Ward J., Macdonald D., Doncaster C (1997) Responses of foraging hedgehogs to badger odour Anim Behav 53:709–720 Wedekind C, Füri S (1997) Body odour preferences in men and women: they aim for specific MHC combinations or simply heterozygosity? Proc Biol Sci 264:1471–1479 Wright RH (1982) The sense of smell Boca Raton, Fla.: CRC Press 133 Yaksi E, von Saint Paul F, Niessing J, Bundschuh ST, Friedrich RW (2009) Transformation of odor representations in target areas of the olfactory bulb Nat Neurosci 12:474–482 Yu CR, Power J, Barnea G, O’Donnell S, Brown HE V, Osborne J, Axel R, Gogos JA, Donnell SO (2004) Spontaneous neural activity is required for the establishment and maintenance of the olfactory sensory map Neuron 42:553–566 Zhang Y, Hoon MA, Chandrashekar J, Mueller KL, Cook B, Wu D, Zuker CS, Ryba NJP (2003) Coding of sweet, bitter, and umami tastes: Different receptor cells sharing similar signaling pathways Cell 112:293–301 Zhao GQ, Zhang Y, Hoon MA, Chandrashekar J, Erlenbach I, Ryba NJP, Zuker CS (2003) The receptors for mammalian sweet and umami taste Cell 115:255–266 Zhou B, Weigel JA, Fauss L, Weigel PH (2000) Identification of the hyaluronan receptor for endocytosis (HARE) J Biol Chem 275:37733–37741 Zimmer-Faust RK, Case JF (1982) Odors influencing foraging behavior of the California spiny lobster, Panulirus interruptus, and other decapod crustacea Mar Behav Physiol 9:35–58 134 Appendix A 135 A Microfluidic Device to Sort Cells Based on Dynamic Response to a Stimulus Swee Jin Tan1.Ô, Michelle Z L Kee2., Ajay Sriram Mathuru3, William F Burkholder1, Suresh J Jesuthasan2,3,4* Microfluidics Systems Biology Lab, Institute of Molecular and Cell Biology, Singapore, Singapore, Neuroscience and Behavioral Disorders Program, Duke-National University of Singapore Graduate Medical School, Singapore, Singapore, Neural Circuitry and Behavior Lab, Institute of Molecular and Cell Biology, Singapore, Singapore, Department of Physiology, National University of Singapore, Singapore, Singapore Abstract Single cell techniques permit the analysis of cellular properties that are obscured by studying the average behavior of cell populations One way to determine how gene expression contributes to phenotypic differences among cells is to combine functional analysis with transcriptional profiling of single cells Here we describe a microfluidic device for monitoring the responses of single cells to a ligand and then collecting cells of interest for transcriptional profiling or other assays As a test, cells from the olfactory epithelium of zebrafish were screened by calcium imaging to identify sensory neurons that were responsive to the odorant L-lysine Single cells were subsequently recovered for transcriptional profiling by qRT-PCR Responsive cells all expressed TRPC2 but not OMP, consistent with known properties of amino-acid sensitive olfactory neurons The device can be adapted for other areas in biology where there is a need to sort and analyze cells based on their signaling responses Citation: Tan SJ, Kee MZL, Mathuru AS, Burkholder WF, Jesuthasan SJ (2013) A Microfluidic Device to Sort Cells Based on Dynamic Response to a Stimulus PLoS ONE 8(11): e78261 doi:10.1371/journal.pone.0078261 Editor: Johannes Reisert, Monell Chemical Senses Center, United States of America Received June 5, 2013; Accepted September 9, 2013; Published November 8, 2013 Copyright: ß 2013 Tan et al This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited Funding: This work was supported by the Joint Council Office (JCO) of the Agency for Science, Technology and Research in Singapore (http://www.a-star edu.sg/About-A-STAR/A-STAR-Joint-Council.aspx), VIP Grant 092-110-0080 (Stephen R Quake, W.F.B program leader), JCO Career Development Award (12302FG014 to S.J.T.), and the Biomedical Research Council, Singapore (http://www.a-star.edu.sg/About-A-STAR/Biomedical-Research-Council.aspx) (S.J.J.) The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript Competing Interests: The authors have declared that no competing interests exist * E-mail: sureshjj@imcb.a-star.edu.sg These authors contributed equally to this work Ô Current address: Clearbridge Accelerator Pte Ltd., Singapore, Singapore single cells [14] However, these techniques introduce a significant external field, which can interfere with measurements of cellular responses and affect cell viability Other sorting techniques that are considered gentle to cells include separation through deterministic lateral displacement [15,16] or alteration of laminar flow characteristics in micro flows [17,18] All of these techniques, however, not allow the introduction of an external stimulant for sorting cells based on ligand-induced responses Here, we describe a device that allows monitoring of signaling in living cells, followed by sorting for molecular characterization We test the device using cells from the olfactory system of the zebrafish Like other vertebrates, the zebrafish has a large repertoire of odorant receptors [19,20,21] Each olfactory sensory neuron expresses one or a few receptors [22], but each receptor can bind several ligands The animal is able to recognize odorants occupying a large chemical space because of the diversity of the receptors [23] One method that has been used to identify olfactory sensory neurons that respond to a particular ligand is to attach a group of neurons to a cover slip, flow the odorant over the neurons and monitor increase in intracellular calcium levels [24] A responding cell can then be picked using a micromanipulator and molecularly characterized The device described here performs a similar procedure, but on cells in suspension, avoiding difficulties that arise from temporarily attaching cells that have to Introduction All biological systems, from multi-species microbial consortia to the adaptive immune system, cancer stem cells and the brain, are characterized by diversity at the single cell level [1,2,3] This richness can result from normal developmental processes, e.g Rag1/Rag2 mediated recombination [4] and gene expression changes following signaling events, or can be induced by environmental stressors that cause chromosomal changes [5] as well as the inherent stochasticity of biochemical reactions [6] Variation at the single cell level may be deleterious as in the case of nervous system disorders [7], but can increase fitness and improve performance of biological systems, for example by broadening the range of signal detection To understand the molecular basis of diversity in signaling, it would be useful to have a method that couples molecular analysis with an assessment of signaling at the single cell level Several techniques for characterizing the genome, transcriptome and metabolome of single cells are currently available [8,9,10] Additionally, several microfluidic-based methods for isolating single cells have been devised These include magnetic separation [11], electro-osmotic-based sorting [12], piezoelectric actuation in a continuous flow [13], and a dynamic array of traps that utilizes dielectrophoretic fields to capture and selectively recover multiple PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response schematic of the experimental setup; detailed instructions for assembling the electronic controller are available at http://www stanford.edu/group/foundry/testing/own_controller.html) Valve operation was automated using scripts written in NI Labview 2009 (National Instruments Corp., Austin, TX, USA; an executable file is available in File S3) Before use, control and flow lines within the microdevice were filled with buffer utilizing dead-end filling and pressure driven flow [28,29] The buffer in the flow lines was then replaced with 1% Pluronic F-127 (a surfactant that ensures cells not adhere to the surface) and incubated for hr These steps for preparing the chip took approximately 75 To operate the device after coating with Pluronic F-127, the flow channels were flushed for at psi with Ringer’s buffer (116 mM NaCl, 2.9 mM KCl, 1.8 mM CaCl2, mM HEPES, pH 7.2) to remove any traces of the surfactant before cell loading Introduction of the stimulant to the cell capture chamber shown in Figure 1C was achieved by opening micro-valves 11, 12 and 17 (Figure 1B) Subsequently, micro-valve 11 was closed and microvalve was opened to remove excess stimulant in the system by flushing with buffer, directing the flow to the waste outlet The device was now ready for the introduction of cells Cell suspensions (50–100 ml) were loaded and directed to the cell trap by the onchip peristaltic pump The pump was operated by sequential actuation of the valves in the pattern 101, 100, 110, 010, 011, 001 where represents an activated valve and shows an open valve Although a wide range of operating frequencies are possible (up to 100 Hz [28]), a frequency of 1.67 Hz was found to allow manual tracking of objects within the cell-trap, with a flow rate was approximately 200 mm/s When a single cell was observed to enter the cell trap region, the peristaltic pump was stopped and the micro-valves surrounding the cell trap were closed A baseline fluorescence image of the trapped cell was then taken Micro-valve 11 was subsequently opened for 15 s to expose the trapped cell to stimulant During this phase, fluorescence images were taken at intervals of 1–3 s and the maxima fluorescence intensity was recorded Positive cells (registering an intensity change of greater than 2%) were directed to the positive collection well Otherwise cells were discarded or sent to the negative collection well Flow from the stimulation line into the cell trap was tested with a green food dye (Tartrazine E102, Brilliant Blue E133, Carmosine E122; Star Brand, Singapore) Intensity was measured using imageJ be subjected to a flow We show that olfactory sensory neurons in suspension are able to respond to an odorant, and that the device can accurately sort cells responsive to the stimulus The modular nature of the system allows additional functionality, such as cell lysis or sample processing (mRNA extraction and cDNA synthesis) to be implemented on-chip, with the potential to increase the efficiency at which cells can be characterized We propose that this device is adaptable to other biological systems where the sorting of cells based on their dynamic response to a stimulus is an important step in understanding complexity Materials and Methods Microdevice fabrication procedures The microdevice was created using soft lithography as summarized in Figure S1 in File S1 using standard methods that have been described previously [25,26,27] The photo mask was drawn in AutoCAD 2011 (Autodesk, Inc., San Rafael, CA, USA) and produced on glass with critical dimensions of mm (IMRE, Singapore) The AutoCAD file is available in the accompanying File S2, and additional information regarding the file format and general design considerations for PDMS-based large-scale integrated microfluidic devices can be found at http://www.stanford edu/group/foundry/services/mask_design_rules.html) The master control mould (for the control lines shown in green in Figure 1B) was fabricated by spin coating (2700 rpm, 30 s) SU-8 2025 (MicroChem Corp., Newton, MA, USA) on a 4-inch silicon wafer achieving a 23 mm depth The flow mould (for sample and reagent flow lines shown in blue in Figure 1B) was fabricated by spin coating (1700 rpm, 60 s) SPR 220-7.0 (Rohm and Hass, Midland, MI, USA) on another 4-inch silicon substrate yielding a 15 mm thickness The flow mould was further subjected to a heat reflow process by incubating at 190uC for 30 minutes followed by a slow ramp down to room temperature This changes the photoresist profile from a rectangular cross-section to a curved surface that ensures that the microfluidic valves are closed completely during actuation [28] Using these moulds, the control and flow layers of the device were cast out of polydimethylsiloxane (PDMS) (Sylgard 184, Dow Corning Corp., Midland, MI, USA) and manually assembled under a stereoscope (Nikon, Japan) essentially as described [28,29] Base to curer ratios of 5:1 and 20:1 were used for the control and flow layers respectively The larger ratio used for the flow layer produced a fairly flexible membrane that bowed downwards easily when a positive pressure was applied in the control fluidic lines, whereas the smaller ratio used for the control layer resulted in a harder material that increased the overall rigidity of the device The control and flow layers were bonded to each other by first partially curing each layer at 80uC (20 for the control layer and 25 for the flow layer) [30] and then bringing them together and heating at 80uC for one hour After this, control fluidic ports (holes) and cell recovery wells were punched with a gauge 23 needle and mm biopsy punch respectively The finished device was then bonded onto a cleaned glass substrate by incubating at 80uC overnight Preparation of samples Experiments were approved by the Institutional Animal Care and Use Committee of the Biological Resource Centre at Biopolis (#120730) Olfactory epithelium cells isolation and dissociation methods were modified from Corotto et al [31] Briefly, AB wild type or TRPC2: Venus [32] zebrafish (a gift of Y Yoshihara from the RIKEN Brain Science Institute) were euthanized in ice-water [33] Olfactory rosettes were isolated in calcium-free Ringer’s buffer (116 mM NaCl, 2.9 mM KCl and mM HEPES, pH 7.2) Rosettes were suspended in 100 ml calcium-free Ringer’s with mU of Papain (Worthington Biochemical Corporation, Lakewood, NJ, USA) and units of DNaseI, and incubated for 15 mins at 37uC with shaking at 300 rpm After replacing the solution with standard Ringer’s buffer, cells were dissociated into suspension by trituration to minimize cell clumping For cell viability studies, trypan blue (Life Technologies, USA) was added to the cell suspension in a 1:1 ratio with stock solution (0.4%) For on-chip studies, the dye was introduced via the stimulation inlet For the ionophore or odorant stimulation experiments, mM of a fluorescent calcium indicator, Fluo 4-AM or Rhod 2-AM (Life Technologies Corp., USA), was added to the cell suspension after Device preparation and operation For pneumatic control of the device, the fluidic control ports (Figure 1B) were connected to pneumatic solenoid valves (Pneumadyne Inc., USA) using 23 gauge needle connectors attached to tygon tubing The valves were linked to pressure sources that provided constant pressures of and 20 psi and were operated electronically using a custom-built USB-based electronic controller unit interfaced to a PC (see Figure S2 in File S1 for a PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response Figure A microfluidic device for sorting cells based on dynamic responses (A) Outline of the operations performed using the chip (B) Schematic of the device The main components are the peristaltic pump, cell trap and selection wells The numbered circles represent points for insertion of 23 gauge needle connectors Blue lines represent flow lines, which are located in the lower layer of the device (Figure S1 in File S1), while green lines are control lines, which are located in the upper layer of the device (Figure S2 in File S1) Cells are introduced via a flow line, and their movement is regulated by the control lines that operate the peristaltic pump and seal the cell-trap by actuating push-down valves The stimulus is then delivered via another flow line After this, cells are directed to one of two large wells for recovery Flow lines for Pluronic-F127 and Ringer’s buffer are used to prepare the chip Other flow lines are available for introduction of multiple stimuli to the cell trap (C) Different components in the system for capturing single cells and introducing stimulants (cell trap), controlling the movement of cells (peristaltic pump), and cell sorting (selector), as seen in a finished device Scale bar represents 100 mm (D) An operational device mounted on a microscope The tygon tubes are used to provide control and introduce reagents doi:10.1371/journal.pone.0078261.g001 extraction and incubated for 15 mins at room temperature following manufacturer’s recommendation This led to a uniform labelling of the cell To remove excess dye, cells were collected by centrifugation, washed and resuspended in 100 ml Ringer’s buffer mM A23187 (Life Technologies Corp., USA) was used to induce calcium influx qRT-PCR of single cells Single cell qRT-PCR was performed to ascertain the expression profiles of the genes TRPC2, OMP, EF1a, B2M, and b-Actin (primer sets used for qRT-PCR are listed in Table S1 in File S1) The protocol was adapted from Dalerba et al [35] Briefly, after sorting a cell and verifying under bright-field microscopy that a single cell was present in the recovery well, ml of CellsDirect (Life technologies, USA) premixed with SuperaseIn RNase-inhibitor at a ratio of 50:1 (v/v; U of SuperaseIn RNase-inhibitor per sample) was added to the well and the cell was transferred to a PCR tube on ice Cells were lysed at 75uC for 10 and firststrand cDNA synthesis was performed using the SuperScript III reverse transcriptase kit (Life Technologies Corp, USA) following the manufacturer’s instructions qRT-PCR was performed on a Stratagene MX2005P (Agilent Technologies Inc, Santa Clara, CA, USA) Each reaction contained 10% of the total cDNA synthesized from each cell, Maxima SYBR Green qPCR master mix (Fermentas Inc., USA) and 0.3 mM of the forward and reverse primers (Table S1 in File S1) Amplification was performed with the thermal cycling conditions: cycle of 95uC for 10 mins; a preamplification step consisting of 20 cycles of 95uC for 15 s, 54uC for 30 s and 72uC for 30 s, followed by 40 cycles using the same thermocycling parameters Melting curve analysis was performed at the completion of each run Fluorescence signals for two channels using SYBR Green (495 nm) and ROX (535 nm) were recovered The delta-Ct values [(b-Actin Ct)-(OMP, E1a, or B2M Ct)] were then plotted Image acquisition The microdevice was mounted on the Olympus inverted microscope (IX71, Olympus Corp., Japan) and a 206 or 406 objective was used to image cells in the cell trap Bright field and fluorescence images were captured using a CoolSNAP EZ camera (Photometrics, Tucson, AZ, USA) and filter sets U-MWB2 (Ex: 460–490 nm/Em: 520 nm; YFP, Fluo 4-AM) and U-MWG2 (Ex: 510–560 nm/Em: 590 nm; Rhod 2-AM) from Olympus Singapore Pte Ltd (Singapore) Excitation was provided by a mercury arc lamp (100 W) Measurement of fluorescence intensities during the experiment was done via ImageJ software [34] Background subtraction and segmentation were applied to every image frame to obtain the average intensity of the whole cell The use of average intensity (i.e total intensity/area) compensated for changes in total intensity that may occur due to slight variations in focal plane The ratio of the fluorescence intensity change (DF) to the baseline fluorescence value before stimulation (F0) was calculated for each frame (DF/F0) PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response Figure Operational characteristics for stimulation and cell recovery (A) Tracking of a single cell (indicated by the green box) traversing through the device The cell can be visualized within the flow line, in the cell trap and in the recovery well (B) Visualization of the stimulation procedure using a green dye The dye is initially in the stimulation line, and fills the entire cell trap chamber within 15 seconds (C) Relationship between volume dispensed during the cell recovery process and the operational pressure Scale bar represents 100 mm doi:10.1371/journal.pone.0078261.g002 the micro-valves are actuated, and a frequency of 1.67 Hz, was determined empirically to enable manual tracking of objects The flow direction of cells in the micro-channel can also be manipulated to correctly position cells in the trap [29] In contrast to pressure driven flow, the use of an on-chip peristaltic pump provides greater flexibility and accuracy and enables a limiting number of cells to be examined with minimal loss The selector unit on the chip is the component that allows the user to direct a cell to one of the two collection wells based on the phenotype observed after stimulation in the cell trap To facilitate cell recovery, retrieval is performed by pressure driven flow at psi into the designated collection well shown in Figure 1C This enhances the turnover rate for screening and sorting cells The recovered cell can be visualized at the collection well and removed by pipetting out of the well Results Microfluidic device design A schematic representation and overview of the microfluidic chip is depicted in Figure The device contains three main components to achieve single cell manipulation, monitoring, and cell sorting: a cell trap, an on-chip peristaltic pump, and a selector to sort cells to either a ‘‘positive’’ selection well or a ‘‘negative’’ selection well (Figure 1B, C) Fabricated from PDMS (Figure S1 in File S1), the chip is optically transparent and compatible with most existing inverted and upright microscopes This minimizes start-up costs while permitting real-time visualization of cells by fluorescence microscopy or other methods before and after exposure to one or more treatment conditions The cell trap is a region enclosed by four micro-valves where a single cell enters and is held in place by actuating (closing) the four valves The trapped cell is exposed to a predetermined stimulus by controlling the micro-valve attached to the stimulant flow line With this layout, we can precisely control the exposure time and synchronize stimulation with imaging to measure the instantaneous changes within the isolated cell via fluorescence microscopy The cell can then be released to allow the entry of a new cell This provides robust and systematic interrogation of single cells The on-chip peristaltic pump provides a gentle yet precise means to direct cells to the cell trap from the inlet where cells are loaded onto the chip by pipetting The peristaltic pump consists of three push-down micro-valves (200 mm6100 mm) [28,29] that are arranged in series and actuated sequentially The speed that cells traverse through the device correlates with the frequency at which Microdevice performance Different components in the microfluidic device work in tandem to accomplish precise control and active stimulation before sorting each incoming cell Control of the process is semi-automated through custom scripts in Labview, which eases device operation for cell interrogation and sorting Figure 2A shows the tracking of a cell throughout the entire process, confirming the ability of the device to capture and recover single cells We characterized how the relative concentration of a stimulant changed over time after injection from the stimulation flow line into the cell trap using a green food dye (Figure 2B) To ensure efficient stimulant delivery without displacing the isolated cell, a Table Sorting efficiency and cell recovery from different input conditions Positive collection well Negative collection well Mix ratio Fluo4 (+) Fluo4 (2) Fluo4 (+) Fluo4 (2) Cells sorted 50-50 47 0 53 100 30-70 20 79 100 10-90 0 93 100 doi:10.1371/journal.pone.0078261.t001 PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response Figure Ionophore-induced calcium influx in cells extracted from zebrafish olfactory epithelium (A) Maximum fluorescence change in individual cells after stimulation with A23187 (n = 45) or Ringer’s (n = 14) The inset depicts the data for the same cells on a log scale (*p = 0.0144, unpaired one-tailed Student’s t-test) (B) Change in fluorescence intensity relative to baseline fluorescence intensity (DF/F0) plotted over time for representative cells The black trace indicates the averaged response A23187 was injected into the cell trap at the indicated time (80 s), and remained in the trap Control cells stimulated with Ringer’s buffer did not exhibit a significant change in fluorescence intensity doi:10.1371/journal.pone.0078261.g003 controlling pressures Over the range of conditions tested, we derived the relationship as denoted in the equation: positive pressure was kept in the trap during stimulant injection The deformability of the PDMS polymer used to fabricate the device allows the cell trap to expand to accommodate the injected volume of stimulant As shown in Figure 2B, the entire celltrapping chamber with dimensions of 100 mm by 200 mm was filled with green color dye within a span of 15 s The stimulantconnecting micro-valve was then closed to allow the space to equilibrate By measuring intensity, the cell trap was found to contain approximately 40% of the concentration of the dye in the stimulant line During the steps used to recover cells from the trap, the dye was flushed completely from the trap and replaced by cell suspension buffer from the flow line leading into the trap To prevent cross-contamination during the next cell stimulation, a portion of the green dye from the stimulus line was flushed out to waste For determining the optimal cell input concentration, we utilized varying concentrations of 10 mm polystyrene beads (Life Technologies Corp., USA) to simulate cells and monitored the number of beads captured within the cell trap during each cycle of trapping and recovery The probability of capturing two or more beads in the trap was very low (,2%) for samples ranging from 2,000 beads per ml to 2,000,000 beads per ml but sharply increased at higher concentrations (Table S2 in File S1) The number of beads that could be trapped and recovered per minute increased with concentration (Table S2 in File S1) Thus, the optimal input concentration for the pumping condition used (1.67 Hz) was around million beads per ml, which allowed an average of 53 beads to pass through the cell trapping region per minute, corresponding to approximately cell per second into the trap In cases when two or more cells are trapped, the cells can be discarded to the waste outlet utilizing the selector (Figure 1B) Central to the operation of the device is the cell sorting and recovery capability To ensure a reliable retrieval, the volume of media dispensed by the system during cell ejection was measured as a function of input pressure As illustrated in Figure 2C, the volume disbursed by the system is a function of both the input pressure and the amount of time given for cell ejection The relationship at each operating condition over time is linear since the fluidic resistance in the flow line remains the same at different PLOS ONE | www.plosone.org time~volume=½ð0:0218 x pressureÞ{0:0081Š where time (seconds) denotes the duration required to achieve the intended volume dispensed (microliter) at the designated operating pressure (psi) The ability to recover cells in a relatively constant volume of buffer is very beneficial for downstream applications such as single cell RNA isolation for qPCR Cell viability and sorting efficiency The effect of the device on cell survival was examined using cells isolated from the zebrafish olfactory epithelium Cell viability was assayed by incubation with trypan blue , a dye that is taken up only by dying cells [36] Immediately after dissociation, cell viability was estimated at 87% (n = 77 cells; counted on a heamocytometer) Of a different batch of 60 cells that were loaded onto the device, trapped, and exposed to trypan blue administered via the stimulus inlet dead cells were recorded over a period of hrs (85% viable) A count of the remaining cell suspension after hrs yielded 84% viable cells Thus, the device had minimal impact on cell viability As a test of sorting efficiency, samples of unlabeled and fluorescently labeled olfactory epithelial cells that had been premixed in different ratios were sorted on the device For each sample, 100 cells were sorted, and cells collected at the recovery well were tabulated An accuracy of close to 100% was observed (Table 1) Ionophore and L-lysine stimulations Our aim in designing the device was to identify a subset of cells within a heterogeneous population that responded to a specific stimulus, using increase in intracellular calcium levels as a measure of the cellular response, and to recover both responsive and nonresponsive cells for molecular characterization As a proof-ofconcept experiment, we assayed calcium influx in cells isolated from the zebrafish olfactory epithelium in response to treatment November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response Figure Lysine stimulation of a mixed cell population from the olfactory epithelium of TRPC2:Venus transgenic zebrafish (A) Relative change of fluorescence intensity in cells expressing Venus, following stimulation with L-lysine Gray lines represent the response of individual cells while the black line plots the averaged response L-lysine was injected into the cell trap at the time indicated by the dashed line (B) Relative fluorescence intensity change in cells with no observable Venus fluorescence, in response to L-lysine (C) Maximum fluorescence intensity change measured for each cell after stimulation with the indicated ligand Venus expressing cells (labeled Venus (+)) show a response to L-lysine, but not to Ringer’s or GCDA Cells that did not express Venus (labeled Venus (2)) did not respond to L-lysine or to Ringer’s For clarity, the last two columns replot the data for Venus-expressing cells stimulated with L-lysine, grouping the data points based on whether the maximum fluorescence change for a cell was above or below the threshold of 2.7% (D) qRT-PCR analysis of cells in panel (C), showing relative abundance of TRPC2 and b-Actin mRNA All Venus expressing cells expressed relatively high levels of TRPC2 TRPC2 mRNA was detected in only two cells that did not express Venus ND: not detected The corresponding data for OMP, EF1a, and B2M are shown in Figures S3, S4, and S5 in File S1 [(C): **p = 0.0005; (D): **p,0.0001; unpaired one-tailed Student’s t-test)] doi:10.1371/journal.pone.0078261.g004 As a control, we stimulated a number of isolated cells with Ringer’s buffer No significant change in fluorescence intensity was observed in these cells (Figure 3A, B) These data establish the feasibility of using the device to screen cells for changes in intracellular calcium levels in response to a stimulus They also establish that even though cells are in a suspension, fluorescence changes due to ligand introduction can be distinguished from any change that may occur simply by the introduction of a reagent such as Ringer’s buffer We utilized the device to identify and molecularly characterize a subpopulation of olfactory sensory neurons (OSNs) that respond to a specific odor cue, L-lysine, from the heterogeneous population of cells in the zebrafish olfactory epithelium In general, amino acids are sensed by neurons that express the channel TRPC2 [32,39,40], and these are labeled in the TRPC2:Venus transgenic with the ionophore A23187, which promotes the rapid diffusion of calcium across biological membranes [37] Freshly isolated cells suspended in Ringer’s buffer, containing 1.8 mM CaCl2, were labeled with a fluorescent calcium indicator (Fluo 4-AM) and loaded into the device Single cells were captured in the trap and exposed to mM A23187 for 15 s Fluorescence images were taken every seconds beginning minute before the stimulation event and extending to minutes after start of stimulation From a total of 45 randomly selected cells, the maximal fluorescence increase in response to the ionophore was, on average, 1.8 fold (Figure 3A), consistent with other studies using A23187 and similar concentrations of calcium in the cell suspension buffer [38] Although responses were heterogeneous, in terms of latency and magnitude, most cells reached a 50% increase in fluorescence intensity within the first 15 s of exposure to A23187 (Figure 3B) PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response line [32] In contrast, bile acids are sensed by OSNs that express the olfactory marker protein (OMP) We loaded cells from the olfactory epithelium of TRPC2:Venus fish onto the device and monitored the change in intracellular calcium levels as individual cells were exposed to one of three conditions: 10 mM L-lysine, Ringer’s buffer alone, or 10 mM of the bile acid glycochenodeoxycholic acid (GCDA) Rhod 2-AM was used as a calcium indicator here to avoid overlap with the fluorescence of Venus Based on the distribution of changes in fluorescence intensity (DF/F0) observed when cells were stimulated with buffer alone, we set the threshold for scoring a positive response to L-lysine or GCDA as DF/ F0.2.7%, which is similar to the threshold value used in previous studies [39,41,42] A subset of the cells that expressed Venus (68%) were scored as responsive to L-lysine (Figure 4A, C) In contrast, no significant increase in fluorescence intensity in response to Llysine was observed among the cells that did not express Venus (Figure 4B, C) None of the Venus-expressing cells tested responded to GCDA (Figure 4C) After screening each cell by fluorescence microscopy and recovering the cell in one of the recovery wells on the device, we manually transferred the cell to a PCR tube for cell lysis We collected cells from three independent experiments for mRNA extraction and qRT-PCR For this pilot experiment, we quantified the relative abundance of five genes: TRPC2, OMP, and the housekeeping genes EF1a, B2M, and ß-Actin As expected, all of the cells that responded to L-lysine expressed detectable levels of TRPC2 mRNA (Figure 4D) TRPC2 mRNA was undetectable in all Venus (2) cells, with the exception of two cells, which had TRPC2 mRNA levels comparable to the Venus-expressing cells These cells may represent a subtype in which transcription from the endogenous TRPC2 is activated via cis-acting sequences that are missing from the promoter fragment used to drive expression of the TRPC2:Venus reporter gene Consistent with previous reports that expression of TRPC2 and OMP in OSNs is generally mutually exclusive [32], we observed a sizeable fraction of OMP-expressing cells among the cells that did not express Venus (Figure S3 in File S1) There were, however, six cells that expressed both TRPC2 and OMP mRNA, five of which also expressed TRPC2:Venus (as determined by fluorescence, Figure S3 in File S1) Thus, there may be a small subset of OSNs in the zebrafish olfactory epithelium that express both TRPC2 and OMP provides a platform where tedious experimental protocols can be automated to reduce associated human errors Furthermore, these devices incur minimal costs and can be replaced for each experiment, which reduces chances of cross-contamination In order to achieve a gentle yet efficient sorting process, we have incorporated different components made up of micro-valves The device is capable of reliable stimulus delivery and performing consistent cell recovery Dynamic changes in the cells after stimulation are actively monitored via time-lapse fluorescence microscopy This creates the possibility of working with small number of input cells and is beneficial for sensitive measurements in situations such as biopsies and rare cell samples [43] We have also ascertained that cell viability is unaffected during or after the entire process Screening cells based on their dynamic properties by this method was low-throughput (approximately 30 seconds for stimulating and sorting) to ensure the reliability of sorting single cells with minimum error However, it may be possible to increase throughput by parallelization of the cell trap to allow simultaneous interrogation of multiple cells, as well as automating ligand delivery, image analysis and cell sorting In this case, real-time image analysis can be used to determine when cells enter the trap, as well as to determine their response upon stimulus introduction, as this would be reflected by a change in fluorescence intensity Supporting Information File S1 This contains all the supplementary figures Figure S1, Summary of device fabrication using soft lithography Figure S2, Schematic of the pneumatic connections for operating the device Figure S3, Relative OMP and b-Actin mRNA abundance for the cells shown in Fig ND: Not detected Figure S4, Relative EF1a and bActin mRNA abundance for the cells shown in Fig ND: Not detected Figure S5, Relative B2M and b-Actin mRNA abundance for the cells shown in Fig ND: Not detected Table S1, Primer sequences used in qRT-PCR Table S2, Effects of input concentration on single cell trapping (DOCX) File S2 AutoCAD file used to create the photo mask (DWG) File S3 A exe file used to operate the device The GUI allows control of valves, and hence movement of cells and delivery of stimulus or buffer to cells (EXE) Discussion We have developed a microfluidic device to interrogate single cells that allows recording of dynamic responses to a ligand, followed by sorting and enrichment of a desired subpopulation This was demonstrated here using olfactory sensory neurons of the zebrafish, and monitoring their response to the odorant L-lysine Integrated microfluidic systems offer several advantages for single cell analysis The micro-scale size of such devices ensures precise fluid control due to laminar flow, requires significantly lower volumes of reagents compared with standard formats, and Acknowledgments We thank Caroline Kibat for maintenance of the transgenic lines Author Contributions Conceived and designed the experiments: SJJ SJT ASM WFB Performed the experiments: MZLK SJT ASM Analyzed the data: SJT MZLK ASM WFB Wrote the paper: SJJ SJT ASM WFB Designed the software: SJT References Peisajovich SG, Garbarino JE, Wei P, Lim WA (2010) Rapid diversification of cell signaling phenotypes by modular domain recombination Science 328: 368– 372 Elowitz MB, Levine AJ, Siggia ED, Swain PS (2002) Stochastic gene expression in a single cell Science 297: 1183–1186 Iourov IY, Vorsanova SG, Yurov YB (2012) Single cell genomics of the brain: focus on neuronal diversity and neuropsychiatric diseases Curr Genomics 13: 477–488 Tang F, Barbacioru C, Nordman E, Li B, Xu N, et al (2010) RNA-Seq analysis to capture the transcriptome landscape of a single cell Nat Protoc 5: 516–535 Brenner K, You L, Arnold FH (2008) Engineering microbial consortia: a new frontier in synthetic biology Trends in biotechnology 26: 483–489 Clevers H (2011) The cancer stem cell: premises, promises and challenges Nature medicine 17: 313–319 Galli SJ, Borregaard N, Wynn TA (2011) Phenotypic and functional plasticity of cells of innate immunity: macrophages, mast cells and neutrophils Nature immunology 12: 1035–1044 Oettinger MA, Schatz DG, Gorka C, Baltimore D (1990) RAG-1 and RAG-2, adjacent genes that synergistically activate V(D)J recombination Science 248: 1517–1523 PLOS ONE | www.plosone.org November 2013 | Volume | Issue 11 | e78261 Cell Sorting Based on Dynamic Cellular Response Heinemann M, Zenobi R (2011) Single cell metabolomics Curr Opin Biotechnol 22: 26–31 10 Zhang L, Cui X, Schmitt K, Hubert R, Navidi W, et al (1992) Whole genome amplification from a single cell: implications for genetic analysis Proc Natl Acad Sci U S A 89: 5847–5851 11 Adams JD, Kim U, Soh HT (2008) Multitarget magnetic activated cell sorter Proc Natl Acad Sci U S A 105: 18165–18170 12 Fu AY, Spence C, Scherer A, Arnold FH, Quake SR (1999) A microfabricated fluorescence-activated cell sorter Nat Biotechnol 17: 1109–1111 13 Cho SH, Chen CH, Tsai FS, Godin JM, Lo YH (2010) Human mammalian cell sorting using a highly integrated micro-fabricated fluorescence-activated cell sorter ([small mu ]FACS) Lab on a Chip 10: 1567–1573 14 Voldman J, Gray ML, Toner M, Schmidt MA (2002) A microfabrication-based dynamic array cytometer Anal Chem 74: 3984–3990 15 Davis JA, Inglis DW, Morton KJ, Lawrence DA, Huang LR, et al (2006) Deterministic hydrodynamics: taking blood apart Proc Natl Acad Sci U S A 103: 14779–14784 16 Huang LR, Cox EC, Austin RH, Sturm JC (2004) Continuous Particle Separation Through Deterministic Lateral Displacement Science 304: 987–990 17 Wolff A, Perch-Nielsen IR, Larsen UD, Friis P, Goranovic G, et al (2003) Integrating advanced functionality in a microfabricated high-throughput fluorescent-activated cell sorter Lab Chip 3: 22–27 18 Bang H, Chung C, Kim J, Kim S, Chung S, et al (2006) Microfabricated fluorescence-activated cell sorter through hydrodynamic flow manipulation Microsystem Technologies 12: 746–753 19 Alioto TS, Ngai J (2006) The repertoire of olfactory C family G protein-coupled receptors in zebrafish: candidate chemosensory receptors for amino acids BMC Genomics 7: 309 20 Saraiva LR, Korsching SI (2007) A novel olfactory receptor gene family in teleost fish Genome Res 17: 1448–1457 21 Weth F, Nadler W, Korsching S (1996) Nested expression domains for odorant receptors in zebrafish olfactory epithelium Proc Natl Acad Sci U S A 93: 13321–13326 22 Sato Y, Miyasaka N, Yoshihara Y (2007) Hierarchical regulation of odorant receptor gene choice and subsequent axonal projection of olfactory sensory neurons in zebrafish J Neurosci 27: 1606–1615 23 Firestein S (2001) How the olfactory system makes sense of scents Nature 413: 211–218 24 Touhara K, Sengoku S, Inaki K, Tsuboi A, Hirono J, et al (1999) Functional identification and reconstitution of an odorant receptor in single olfactory neurons Proc Natl Acad Sci U S A 96: 4040–4045 25 McDonald JC, Duffy DC, Anderson JR, Chiu DT, Wu H, et al (2000) Fabrication of microfluidic systems in poly(dimethylsiloxane) Electrophoresis 21: 27–40 26 McDonald JC, Whitesides GM (2002) Poly(dimethylsiloxane) as a Material for Fabricating Microfluidic Devices Accounts of Chemical Research 35: 491–499 PLOS ONE | www.plosone.org 27 Xia Y, Whitesides GM (1998) Soft lithography Annual Review of Materials Science 28: 153–184 28 Melin J, Quake SR (2007) Microfluidic large-scale integration: the evolution of design rules for biological automation Annual Review of Biophysics and Biomolecular Structure 36: 213–231 29 Unger MA, Chou HP, Thorsen T, Scherer A, Quake SR (2000) Monolithic microfabricated valves and pumps by multilayer soft lithography Science 288: 113–116 30 Eddings MA, Johnson MA, Gale BK (2008) Determining the optimal ă PDMSA`PDMS bonding technique for microfluidic devices Journal of Micromechanics and Microengineering 18: 067001 31 Corotto FS, Piper DR, Chen N, Michel WC (1996) Voltage- and Ca(2+)-gated currents in zebrafish olfactory receptor neurons J Exp Biol 199: 1115–1126 32 Sato Y, Miyasaka N, Yoshihara Y (2005) Mutually exclusive glomerular innervation by two distinct types of olfactory sensory neurons revealed in transgenic zebrafish J Neurosci 25: 4889–4897 33 Wilson JM, Bunte RM, Carty AJ (2009) Evaluation of rapid cooling and tricaine methanesulfonate (MS222) as methods of euthanasia in zebrafish (Danio rerio) J Am Assoc Lab Anim Sci 48: 785–789 34 Abramoff MD, Magalhaes PJ, Ram SJ (2004) Image Processing with ImageJ Biophotonics International 11: 36–42 35 Dalerba P, Kalisky T, Sahoo D, Rajendran PS, Rothenberg ME, et al (2011) Single-cell dissection of transcriptional heterogeneity in human colon tumors Nat Biotechnol 29: 1120–1127 36 Strober W (2001) Trypan blue exclusion test of cell viability Curr Protoc Immunol Appendix 3: Appendix 3B 37 Pressman BC (1976) Biological applications of ionophores Annu Rev Biochem 45: 501–530 38 Jiang Z, Yin X, Jiang Q (2011) Natural forms of vitamin E and 139carboxychromanol, a long-chain vitamin E metabolite, inhibit leukotriene generation from stimulated neutrophils by blocking calcium influx and suppressing 5-lipoxygenase activity, respectively J Immunol 186: 1173–1179 39 Friedrich RW, Korsching SI (1997) Combinatorial and chemotopic odorant coding in the zebrafish olfactory bulb visualized by optical imaging Neuron 18: 737–752 40 Koide T, Miyasaka N, Morimoto K, Asakawa K, Urasaki A, et al (2009) Olfactory neural circuitry for attraction to amino acids revealed by transposonmediated gene trap approach in zebrafish Proc Natl Acad Sci U S A 106: 9884– 9889 41 Albantakis L, Lohmann C (2009) A simple method for quantitative calcium imaging in unperturbed developing neurons J Neurosci Methods 184: 206–212 42 Ashworth R (2004) Approaches to measuring calcium in zebrafish: focus on neuronal development Cell Calcium 35: 393–402 43 Schubert C (2011) Single-cell analysis: The deepest differences Nature 480: 133–137 November 2013 | Volume | Issue 11 | e78261 Biography Kee Zhi Ling Michelle Born February 27, 1984 in Singapore Education: 2010- 2015 2003-2007 Ph.D., Integrated Biology and Medicine A*STAR Graduate Scholarship Duke-NUS Graduate Medical School, Singapore B.Sc (Hons) in Biological Sciences (Second Upper) Nanyang Technological University Bursary Award Nanyang Technological University, Singapore Publications: 1: Tan SJ*, Kee MZ*, Mathuru AS, Burkholder WF, Jesuthasan SJ A microfluidic device to sort cells based on dynamic response to a stimulus PLoS One 2013 Nov 8;8(11):e78261 * Equal contributions 2: Tsuda S, Kee MZ, Cunha C, Kim J, Yan P, Loew LM, Augustine GJ Probing the function of neuronal populations: combining micromirror-based optogenetic photostimulation with voltage-sensitive dye imaging Neurosci Res 2013 Jan;75(1):76-81 3: Kee MZ, Wuskell JP, Loew LM, Augustine GJ, Sekino Y Imaging activity of neuronal populations with new long-wavelength voltage-sensitive dyes Brain Cell Biol 2008 Dec;36(5-6):157-72 4: Hendricks M, Mathuru AS, Wang H, Silander O, Kee MZ, Jesuthasan S Disruption of Esrom and Ryk identifies the roof plate boundary as an intermediate target for commissure formation Mol Cell Neurosci 2008 Feb;37(2):271-83 ... combinatorial detection of a vast number of odorants (Friedrich and Korsching, 1997; Malnic et al., 1999; Araneda et al., 2000; Katada et al., 2005; Nakagawa et al., 2005; Hallem and Carlson,... N-acetylglucosamine (GlcNAc) Glucuronic acid (GlcA) Iduronic acid (IdoA) N-acetylgalactosamine (GalNAc) Galactose (Gal) B GlcA GlcNAc Dermatan sulfate Figure 1-1: Schematic representation of GAGs (A) GAGs... GAGs (A) GAGs are linear polysaccharide chains consisting of a disaccharide unit of a hexose sugar (GlcNAc or Heparian sulfate GalNAc) and uronic sugar molecule (GlcA, IdoA or) /or galactose (Modified