ArticlePDF AvailableCritical Analysis of Particle Detection Artifacts in Synaptosome Flow CytometryMay 2019eNeuro 6(3):ENEURO.0009-19.2019DOI:10.1523/ENEURO.0009-19.2019Authors: Benjamin D HobsonColumbia University Peter A. SimsPeter A. SimsThis person is not on ResearchGate, or hasn t claimed this research yet. Download full-text PDFRead full-textDownload full-text PDFRead full-textDownload citation Copy link Link copied Read full-text Download citation Copy link Link copiedCitations (3)References (48)Figures (3)Abstract and FiguresFlow cytometry and fluorescence-activated sorting are powerful techniques that hold great promise for studying heterogeneous populations of submicron particles such as synaptosomes, but many technical challenges arise in these experiments. To date, most flow cytometry studies of synaptosomes have relied on particle detection using forward scatter (FSC) measurements and size estimation with polystyrene (PS) bead standards. However, these practices have serious limitations, and special care must be taken to overcome the poor sensitivity of conventional flow cytometers in the analysis of submicron particles. Technical artifacts can confound these experiments, especially the detection of multiple particles as a single event. Here, we compared analysis of P2 crude synaptosomal preparations from murine forebrain on multiple flow cytometers using both FSC-triggered and fluorescence-triggered detection. We implemented multicolor fluorescent dye-based assays to quantify coincident particle detection and aggregation, and we assessed the false colocalization of antigens in immunostaining analyses. Our results demonstrate that fluorescence triggering and proper dilution can control for coincident particle detection, but not particle aggregation. We confirmed previous studies showing that FSC-based size estimation with PS beads underestimates biological particle size, and we identified pervasive aggregation in the FSC range analyzed in most synaptosome flow cytometry studies. We found that analyzing P2 samples in sucrose/EDTA/tris (SET) buffer reduces aggregation compared to PBS, but does not completely eliminate the presence of aggregates, especially in immunostaining experiments. Our study highlights challenges and pitfalls in synaptosome flow cytometry and provides a methodological framework for future studies. Optical configuration of the LSRFortessa…  Optical configuration of the Influx…  List of antibodies…  Discover the world s research20+ million members135+ million publications700k+ research projectsJoin for freePublic Full-text 1 Accepted manuscripts are peer-reviewed but have not been through the copyediting, formatting, or proofreadingprocess.Copyright © 2019 Hobson and SimsThis is an open-access article distributed under the terms of the Creative Commons Attribution 4.0 International license, whichpermits unrestricted use, distribution and reproduction in any medium provided that the original work is properly attributed.This Accepted Manuscript has not been copyedited and formatted. The final version may differ fromthis version. A link to any extended data will be provided when the final version is posted online.Research Article: Methods/New Tools | Novel Tools and MethodsCritical Analysis of Particle Detection Artifacts in Synaptosome FlowCytometryBenjamin D. Hobson1,2,3 and Peter A. Sims1,2,41Department of Systems Biology, Columbia University Irving Medical Center, New York, NY, 100322Department of Biochemistry Molecular Biophysics, Columbia University Irving Medical Center, New York, NY,100323Medical Scientist Training Program, Columbia University Irving Medical Center, New York, NY, 100324Sulzberger Columbia Genome Center, Columbia University Irving Medical Center, New York, NY, 10032https://doi.org/10.1523/ENEURO.0009-19.2019Received: 7 January 2019Revised: 1 April 2019Accepted: 27 April 2019Published: 22 May 2019B.H. and P.A.S. designed research; B.H. performed research; B.H. analyzed data; B.H. and P.A.S. wrote thepaper.Funding: HHS | NIH | National Institute of General Medical Sciences (NIGMS)T32GM007367;Funding: HHS | NIH | National Institute of Neurological Disorders and Stroke (NINDS)R01NS103473;Funding: HHS | NIH | National Institute on Drug Abuse (NIDA)F30DA047775.Conflict of Interest: The authors report no conflict of interest.This work was supported by National Institute of General Medical Sciences Grant T32GM007367 (BH), NationalInstitute on Drug Abuse Grant 1F30DA047775-01 (BH), and National Institute of Neurological Disorders andStroke Grant R01NS103473 (PS). Research reported in this publication was performed in the CCTI FlowCytometry Core, supported in part by the Office of the Director, National Institutes of Health under awardsS10OD020056. The content is solely the responsibility of the authors and does not necessarily represent theofficial views of the National Institutes of Health. These studies used the resources of the Cancer Center FlowCore Facility funded in part through Center Grant P30CA013696.Correspondence should be addressed to: Benjamin D. Hobson at bh2573@columbia.edu or Peter A. Sims atpas2182@columbia.eduCite as: eNeuro 2019; 10.1523/ENEURO.0009-19.2019Alerts: Sign up at www.eneuro.org/alerts to receive customized email alerts when the fully formatted version ofthis article is published. 1 Title: Critical Analysis of Particle Detection Artifacts in Synaptosome Flow Cytometry 1 Abbreviated Title: Technical Methods in Synaptosome Flow Cytometry 2 Authors: Benjamin D. Hobson1,2,3 and Peter A. Sims1,2,4 3 Author Affiliations: 4 1Department of Systems Biology, Columbia University Irving Medical Center, New York, NY 5 10032 6 2Department of Biochemistry Molecular Biophysics, Columbia University Irving Medical 7 Center, New York, NY 10032 8 3Medical Scientist Training Program, Columbia University Irving Medical Center, New York, NY 9 10032 10 4Sulzberger Columbia Genome Center, Columbia University Irving Medical Center, New York, 11 NY 10032 12 13 Author Contributions: BH and PS Designed research; BH Performed research; BH Analyzed 14 data; BH and PS Wrote the paper. 15 16 Correspondence should be addressed to: Benjamin D. Hobson (bh2573@columbia.edu) or 17 Peter A. Sims (pas2182@columbia.edu), Department of Systems Biology, Columbia University 18 Medical Center, New York, NY 10032 19 20 Number of Figures: 8 Number of words for Abstract: 243 21 Number of Tables: 6 Number of words for Significance Statement: 83 22 Number of Multimedia: 0 Number of words for Introduction: 748 23 Number of words for Discussion: 2704 24 25 Acknowledgements: We thank Siu-Hong Ho, Lu Caisheng, and Wei Wang for excellent 26 technical support in the CCTI Flow Cytometry Core. 27 28 Conflict of Interest: The authors report no conflict of interest. 29 30 Funding sources: This work was supported by National Institute of General Medical Sciences 31 Grant T32GM007367 (BH), National Institute on Drug Abuse Grant 1F30DA047775-01 (BH), 32 and National Institute of Neurological Disorders and Stroke Grant R01NS103473 (PS). 33 Research reported in this publication was performed in the CCTI Flow Cytometry Core, 34 supported in part by the Office of the Director, National Institutes of Health under awards 35 S10OD020056. The content is solely the responsibility of the authors and does not necessarily 36 represent the official views of the National Institutes of Health. These studies used the 37 resources of the Cancer Center Flow Core Facility funded in part through Center 38 Grant P30CA013696. 39 2 Abstract (243 words) 40 41 Flow cytometry and fluorescence-activated sorting are powerful techniques that hold great 42 promise for studying heterogeneous populations of submicron particles such as synaptosomes, 43 but many technical challenges arise in these experiments. To date, most flow cytometry studies 44 of synaptosomes have relied on particle detection using forward scatter (FSC) measurements 45 and size estimation with polystyrene (PS) bead standards. However, these practices have 46 serious limitations, and special care must be taken to overcome the poor sensitivity of 47 conventional flow cytometers in the analysis of submicron particles. Technical artifacts can 48 confound these experiments, especially the detection of multiple particles as a single event. 49 Here, we compared analysis of P2 crude synaptosomal preparations from murine forebrain on 50 multiple flow cytometers using both FSC- and fluorescence-triggered detection. We 51 implemented multicolor fluorescent dye-based assays to quantify coincident particle detection 52 and aggregation, and we assessed the false colocalization of antigens in immunostaining 53 analyses. Our results demonstrate that fluorescence-triggering and proper dilution can control 54 for coincident particle detection, but not particle aggregation. We confirmed previous studies 55 showing that FSC-based size estimation with PS beads underestimates biological particle size, 56 and we identified pervasive aggregation in the FSC range analyzed in most synaptosome flow 57 cytometry studies. We found that analyzing P2 samples in sucrose/tris/EDTA (SET) buffer 58 reduces aggregation compared to phosphate-buffered saline (PBS), but does not completely 59 eliminate the presence of aggregates, especially in immunostaining experiments. Our study 60 highlights challenges and pitfalls in synaptosome flow cytometry and provides a methodological 61 framework for future studies. 62 63 64 Significance Statement (83 words) 65 66 Synaptosomes are an invaluable model for synaptic biology, but these synaptic particles are 67 traditionally analyzed in bulk preparations rather than at the level of single particles. Although 68 flow cytometry is a powerful technique for high throughput particle analysis, submicron particles 69 present unique challenges. Here, the authors investigate key elements of synaptosome flow 70 cytometry experiments, especially those related to artifacts that confound the analysis of single 71 synaptosomes. They identify aggregation as especially problematic and implement methods to 72 minimize its impact on flow cytometry analysis. 73 74 75 Introduction (748 words) 76 77 Synaptosomes are synaptic particles consisting of resealed presynaptic nerve terminals that 78 often remain bound to postsynaptic elements (Whittaker, 1993). These structures were originally 79 isolated by homogenization of brain tissue in isotonic sucrose (Gray and Whittaker, 1962), and 80 have been further purified using a variety of filtration and density gradient centrifugation 81 procedures (Dunkley et al., 1986; Hollingsworth et al., 1985). Synaptosomes retain functional 82 properties such as membrane potential and depolarization-induced neurotransmitter release, 83 making them key model systems for fundamental synaptic biology (Whittaker, 1993). However, 84 studies of synaptosomes in bulk are limited by the purity of the preparations, which contain a 85 mixture of synapse types (e.g. glutamatergic, GABAergic, etc.) as well as contaminating 86 3 neuronal and glial membranes. In principle, high throughput purification and analysis of single 87 synaptosomes is a powerful tool for addressing the incredible heterogeneity of the billions of 88 synapses in mammalian brains. 89 90 Flow cytometry employs a pressurized fluidic system to pass suspensions of cells or particles 91 through an optical flow cell, where scattered light and fluorescence measurements are collected 92 for each particle. Analysis of submicron particles on conventional flow cytometers faces a 93 variety of limitations and pitfalls due to the limited sensitivity of these instruments (Nolan, 2015). 94 Event detection in flow cytometers, known as triggering, occurs when a particle passing through 95 the flow cell causes the trigger parameter to rise above a manually set threshold. The trigger 96 parameter, usually FSC, must have a minimum threshold set as to avoid electronic and buffer 97 noise, but this limits detection of submicron particles. It is generally accepted that conventional 98 flow cytometers have a lower limit of FSC-triggered detection around 300-500nm PS beads (van 99 der Pol et al., 2013), but this does not correspond to the size of biological particles that can be 100 detected (Lannigan et al., 2016; Nolan, 2015). FSC intensity for submicron particles depends on 101 many factors besides particle size, especially refractive index, which is generally 1.59-1.61 for 102 PS beads, 1.40-1.46 for silica beads, and ranges from 1.33-1.40 for cell-derived particles 103 (Nolan, 2015). Biological particles therefore scatter light approximately 10-fold less efficiently 104 than PS beads (Chandler et al., 2011), and direct comparison of PS bead FSC intensity has 105 been shown to underestimate the size of biological particles (Chandler et al., 2011; van der Pol 106 et al., 2012; Simonsen, 2016). 107 108 In addition to difficulties with size estimation using light scatter, numerous reports have identified 109 particle detection artifacts in submicron flow cytometry. Particles below the trigger threshold can 110 be detected when analyzed at high concentrations, a phenomenon known as coincidence or 111 (Nolan and Stoner, 2013; van der Pol et al., 2012). While coincidence is 112 caused by the simultaneous presence of multiple, single particles in the path of the laser, others 113 have reported detection of single events comprised of multiple, aggregated submicron particles 114 (Erdbrügger et al., 2014). Although these artifacts have been recognized in the cell-derived 115 microparticle community (Lannigan et al., 2016; Nolan, 2015), most flow cytometry studies of 116 synaptosomes have not addressed triggering or detection artifacts (Fein et al., 2008; Gylys et 117 al., 2000; Postupna et al., 2014; Prieto et al., 2017). Despite claims that single synaptosomes 118 can be detected by FSC-triggering and identified apart from contaminating particles based 119 solely on FSC signal (Gylys and Bilousova, 2017; Gylys et al., 2004), this has been a point of 120 recent controversy (see comments on Prieto et al., 2017). In contrast to these studies, 121 Biesemann and Herzog found that sorting of FSC-triggered events bearing genetically encoded 122 vesicular glutamate transporter 1 (VGLUT1) fluorescence yielded samples contaminated with 123 GABAergic synaptosomes and myelin (Biesemann, 2010: Results 3.4, Fig. 14, pg. 86-87). 124 Instead, fluorescence-triggering and sorting events below the FSC trigger threshold yielded a 125 VGLUT1 synaptosomal sample of unprecedented purity (Biesemann et al., 2014). The flow 126 cytometry protocols published by these groups are clearly at odds in terms of the best practices 127 to ensure detection, analysis, and sorting of single synaptosomes (Gylys and Bilousova, 2017; 128 Luquet et al., 2017). 129 130 Here, we present a critical flow cytometry analysis of P2 crude synaptosome preparations on 131 two cytometers, the BD Influx and BD LSRFortessa. We investigated several experimental 132 aspects of synaptosome flow cytometry, including: 1) FSC- and fluorescence-triggered particle 133 4 detection, 2) FSC-based size estimation with PS and silica bead standards, 3) range of particle 134 concentration, 4) coincidence and aggregation as causes of false double-positive events, 5) 135 sucrose and PBS buffers for sample preparation and acquisition, and 6) false colocalization of 136 antigens in immunostained samples. Our study highlights technical challenges and identifies 137 methods to minimize their impact on experimental results. 138 139 140 Materials and Methods 141 142 Animals 143 144 Male C57BL/6J mice (6-10 weeks old) were used in all experiments. Mice were housed on a 12-145 hour light/dark cycle with food and water available ad libitum. All animal procedures were 146 performed in accordance with the [Author University] Institutional Animal Care and Use 147 Committee and followed NIH guidelines. 148 149 P1 Crude Nuclei and P2 Crude Synaptosome Preparations 150 151 Preparation of the P2 crude synaptosome fraction was performed using standard procedures 152 (Gray and Whittaker, 1962). Mice were sacrificed by cervical dislocation, after which forebrains 153 were rapidly dissected and placed in 10 volumes of ice cold buffer consisting of 0.32 M sucrose, 154 4 mM HEPES pH 7.4, and protease inhibitors (cOmpleteTM EDTA-free protease inhibitor, 155 Roche). Tissue was homogenized on ice in a glass-glass dounce homogenizer with 10 gentle 156 strokes of loose and tight clearance pestles. All subsequent purification steps were performed 157 on ice or at 4oC unless otherwise specified. The homogenate was centrifuged at 1,000 xg 158 (Eppendorf 5424R) for 10 min to remove nuclei and cellular debris, yielding a P1 pellet and an 159 S1 supernatant. 160 161 A crude nuclei preparation was prepared for flow cytometry according to established procedures 162 (Krishnaswami et al., 2016). The P1 pellet was resuspended in ice cold buffer containing 250 163 mM sucrose, 25 mM KCl, 5 mM MgCl2, 10 mM Tris pH 7.4, 1 μM DTT, 0.1% Triton X-100, and 1 164 μM Hoechst 33342 and homogenized again with 10 strokes of the tight pestle to facilitate 165 release of nuclei. The homogenate was rotated for 15 min at 4oC, filtered through a 40μm cell 166 strainer cap, and centrifuged at 500 xg (Eppendorf 5424R) for 5 min to yield a crude nuclear 167 pellet. 168 169 The S1 supernatant was further centrifuged at 10,000 xg (Eppendorf 5424R) for 20 min to 170 obtain the crude synaptosome pellet (P2). P2 pellets were cryopreserved by resuspension in 4 171 mM HEPES/0.32 M sucrose buffer + 5% DMSO and slowly frozen to -80o C using an 172 isopropanol freezing container. Frozen synaptosomes were used within 2 months. This protocol, 173 when combined with rapid thawing at 37oC on the day of the experiment, has been shown to 174 preserve synaptosome function and morphology (Daniel et al., 2012; Gleitz et al., 1993). After 175 thawing, all experiments were conducted with either PBS (137 mM NaCl, 2.7 mM KCl, 8 mM 176 Na2HPO4, and 2 mM KH2PO4) or SET buffer (320 mM sucrose, 5 mM Tris, 1 mM EDTA). 177 178 Flow Cytometry Instrumentation and Setup 179 180 5 All flow cytometry data acquisition was carried out using the instrument software FACSDiva (BD 181 Biosciences). All flow cytometry data analysis including gating, quantification, and generation of 182 density plots/histograms was performed using FCS Express 6 (De Novo Software). Number of 183 events, % of all events, and channel statistics (median, geometric mean, standard deviation, 184  185 186 All data acquisition on LSRFortessa (hereafter Fortessa , BD Biosciences) was conducted using 187 the lowest possible sample pressure settings. Optical configuration employed and fluorophores 188 detected in these channels are summarized in Table 1. 189 190 The powers of the lasers in the Fortessa are 405nm 100 mW, 488nm 50 mW, 561nm 100mW, 191 640 100mW. Detector voltages underwent minor fluctuations over the course of these studies in 192 order to maintain comparable sample fluorescence and FSC/SSC values for the fluorescent 193 microspheres, which were run at the beginning of each experiment. Approximate detector 194 voltages were as follows: FSC (300), SSC (265), Pacific Blue (407), FITC (588), PerCP-Cy5.5 195 (500), PE (511), and APC (537). 196 197  198 psi. The sample pressure differential was kept as low as possible. The frequency was set at 199 25.8 kHz, piezo amplitude between 5 and 10. Optical configuration employed and fluorophores 200 detected in these channels are summarized in Table 2. 201 202 The power of the lasers in the Influx are 405nm 100mW, 488nm 200mW, 561nm 120mW, 640 203 120mW. Detector voltages underwent minor fluctuations over the course of these studies in 204 order to maintain comparable sample fluorescence and FSC/SSC values for the bead 205 standards. Approximate detector voltages were as follows: FSC (22.70), SSC (25.77), BV421 206 (52.55), FITC (44.51), PerCP-Cy5.5 (41.68), PE (63.65), APC (54.78). 207 208 In the annual preventive and maintenance conducted by the BD engineer, a quality control (QC) 209 is run with the 6 peak rainbow beads. In the daily setup, the QC setting is retrieved and the daily 210 alignment is compared with the QC setting. The complete overlay of the 6 peaks with the QC 211 setting indicates the proper alignment of the laser. Drop delay is done with the drop delay beads 212 and the accudrop camera, a step which calculates the distance between the laser interrogation 213 point and the charging point. 214 215 Daily setup also includes running a mixture of fluorescent PS bead standards (Bangs Labs, Inc.) 216 and a mixture of non-fluorescent silica and fluorescent PS bead standards (Apogee Flow 217 Systems). To ensure reproducibility in FSC-positioning of samples across experiments, the 218 positions of bead populations on FSC vs. SSC and FSC vs. FITC (green fluorescence detector 219 for PS beads) plots are maintained within tight gates stored in the QC workspace. Only 220 occasionally did this require minor adjustments to detector voltages. 221 222 Flow Cytometry Gating 223 224 Gating of bead populations was performed using density plots of FSC vs. SSC and FSC vs. 225 FITC (green fluorescence detector for PS beads). Gating of all fluorescent channels was 226 performed using a polygon gate on a bivariate density plot of FSC vs. fluorophore. Boolean logic 227 gating was used to assess all intersections of positive vs. negative for each fluorophore (e.g. 228 double-negative, calcein AM+/calcein red-, calcein AM-/calcein red+, double-positive). Where 229 6 applicable, hierarchical gating was used to assess all intersections of positive vs. negative for 230 each fluorophore within each FSC gate. 231 232 FSC- and FM4-64-triggered detection 233 234 On both the Influx and Fortessa, the FSC voltages were set in order to place 500nm PS beads 235 just left of center on the FSC axis, near 102 on the Influx and near 103 on the Fortessa. This 236 enabled the beads to be detected just above the FSC noise threshold while also allowing 237 800nm PS and 1300nm Sil beads to be collected on scale. Critically, this also allowed for a 238 significant amount of floor space on the FSC axis below the FSC noise threshold, which 239 became accessible upon fluorescence-triggering with FM4-64. 240 241 For FSC-triggering on the Influx and Fortessa, the FSC trigger threshold was first lowered to the 242 minimum value on the instrument while running clean PBS to define the range of electronic, 243 optical, buffer, and all other sources of FSC noise. After collecting ~10,000 events, the FSC 244 voltage was raised to the minimum level just above the FSC noise threshold such that event 245 rates were 10 events/sec while running clean PBS. Using the configurations described above, 246 the FSC-trigger threshold was generally 250-300 on the Fortessa and generally 0.9-1.10 on the 247 Influx. 248 249 FM4-64 fluorescent triggering was implemented on the Influx and Fortessa using a procedure 250 similar to FSC-triggering. FM4-64 was obtained from ThermoFisher, reconstituted at 100 μg/μL, 251 and diluted into PBS or sucrose buffer at 1.5 μg/mL. This concentration was previously shown to 252 be effective for fluorescence-triggering of synaptosomes (Biesemann et al., 2014; Luquet et al., 253 2017). We found optimal detection of FM4-64 fluorescence in the PerCP-Cy5.5 channel and 254 used this channel for triggering on the Influx and Fortessa. As with FSC triggering, the 255 fluorescence trigger threshold was first lowered to a minimum while running clean buffer + FM4-256 64 to define the noise range. The fluorescent trigger threshold was raised to the minimum level 257 above the noise range such that event rates were 10 events/sec while running buffer + FM4-258 64. Using the configurations described above, the FM-trigger threshold was generally 120-150 259 on the Fortessa and generally 0.35-0.40 on the Influx. 260 261 Fluorescence-Activated Sorting 262 263 For sorting on the Influx, populations of interest were gated using the FACSDiva software. For 264 bead double-positive and double calcein-positive sorting, the event rate was 5,000 events/sec. 265 nts were sorted directly 266 into FACS tubes, which were then used to reanalyze the sorted samples. Reanalysis was 267 conducted 15 minutes after sorting using the exact same cytometer settings, and at least 268 1,500 events were re-analyzed. 269 270 Multicolor Labeling and Mixing Assays 271 272 Cryopreserved P2 samples were rapidly thawed at 37oC, washed in PBS, and centrifuged at 273 10,000 xg (Eppendorf 5424R) for 5 min at 4oC. Pellets were resuspended in PBS and equally 274 divided into separate tubes for single dye labeling. Synaptosomes were incubated at room 275 temperature in PBS supplemented with one of the following dyes: calcein AM for 15-30 minutes 276 at a final concentration of 1 μM, calcein red--30 277 7 minutes at a final concentration of 1.9 μM, violet amine-reactive dye for 30 minutes (1:1000 from 278  279 final concentration of 500 nM. Synaptosomes were returned to ice and subsequently washed 280 and re-centrifuged twice with ice-cold PBS to remove residual dye. All samples were maintained 281 on ice and protected from light after labeling. After the second wash, pellets were resuspended 282 in equal volumes of PBS + 1.5 μg/mL FM4-64 and mixed in equal volumes to generate mixtures 283 of single color-labeled synaptosomes. In most experiments, these mixtures were re-centrifuged 284 and resuspended prior to passing the suspension through a 40μm cell strainer cap prior to flow 285 cytometry. In time lapse experiments, individual samples were passed through a 40μm cell 286 strainer cap and mixed immediately prior to flow cytometry. For experiments employing different 287 buffer combinations, the above protocol was identical, with SET used in place of PBS as 288 -289 assay but with the mixing, 3rd wash, and flow cytometry conducted in SET instead of PBS, and 290 291 some variability in the calcein labeling efficiency across samples. Since only fluorescent 292 particles are eligible to report on coincidence or aggregation, we normalize each fluorescent 293 event type (i.e. single-positive/double-positive) to all fluorescent events within each sample. 294 295 Immunostaining 296 297 Crude nuclear pellets were resuspended in blocking buffer (PBS, pH 7.4 with 5% BSA) 298 supplemented with 1 μM Hoechst 33342 and blocked on ice for 15 min. Primary antibody 299 staining was conducted in blocking buffer at 4oC on a rotator for 30 min. Samples were washed 300 with blocking buffer, centrifuged at 500 xg (Eppendorf 5424R) for 5 min and resuspended in 301 blocking buffer with goat anti-mouse Alexa Fluor 488. Secondary staining was 4oC on a rotator 302 for 20 min. Samples were again washed with blocking buffer plus 1 μM Hoechst 33342 and 303 centrifuged at 500 xg (Eppendorf 5424R) for 5 min, following resuspension in PBS for flow 304 cytometry. 305 P2 crude synaptosomes were briefly fixed in 4% paraformaldehyde (PFA) in PBS, pH 7.4 for 15 306 minutes at room temperature. After washing in blocking buffer to quench residual PFA, 307 synaptosomes were centrifuged at 10,000 xg (Eppendorf 5424R) for 5 min at 4oC. The P2 pellet 308 was resuspended in blocking buffer + 0.1% Tween-20 at room temperature for 30 minutes prior 309 to addition of primary antibodies. Primary antibody staining was conducted at room temperature 310 for 30 minutes, followed by two washes in PBS with centrifugation and resuspension. Pellets 311 were resuspended in blocking buffer + 0.1% Tween-20 with secondary antibodies and stained 312 for 20 minutes at room temperature. Samples were washed twice with PBS, and final samples 313 were resuspended in PBS + 1.5 μg/mL FM4--314 the final wash and flow cytometry were conducted in SET buffer. 315 316 Absolute Particle Count Measurements 317 318 P2 samples from dilution series (Fig. 2-1C) or from repeated centrifugation experiments (Fig. 8 319 and Fig. 8-1) were diluted to a precise final volume (200-500μL) in PBS+FM or SET+FM. A 320 precise volume of resuspended CountBrightTM absolute counting beads was added to each 321 sample. Counting beads were run alone to confirm that they could be gated apart from FM4-64-322 triggered P2 events, and were gated based on SSC and blue fluorescence (BV421 channel). 323 8 324 using the number of FM4-64-triggered P2 events, count bead events, bead count of the lot 325 (beads/μL), and bead/sample volumes. For repeated centrifugation experiments (all using 326 indicated relative centrifugal force (xg) on Eppendorf 5424R), absolute particle counts were 327 calculated based on the total sample volume, and total particle counts in each fraction were 328 determined by back-calculation using the aliquot and sample volumes. 329 330 AlexaFluor488 MESF Calibration 331 332 QuantumTM AlexaFluor488 MESF calibration beads were run on the Influx at the same time and 333 under identical conditions (sample pressure, detector voltages, etc.) as the immunostained 334 samples. MESF calibration beads (7μm) could be identified based on SSC with only a slight 335 reduction (~10%) in the SSC voltage for this experiment. The five calibration beads (blank and 336 MESF beads 1-4) were run individually to confirm separation by the green fluorescence detector 337 (FITC channel). Beads were first gated based on SSC and then based on green fluorescence 338 histograms (see Fig. 7-1E) using full width at half-maximum gating as recommended by the 339 manufacturer. The median fluorescence intensity of the calibration beads was used to fit the 340 MESF calibration curve (see Fig. 7-1-E) using the QuickCal V.2.3 template from the 341 manufacturer (Bangs Laboratories, Inc.). The calibration curve was used to calculate the 342 median fluorescence intensity of indirect immunofluorescence (Ms anti VGAT + anti-mouse 343 AlexaFluor488) in MESF for subsets of gated events (see Fig. 7-1F). Numerical data are 344 available in Figure 7-3. 345 346 Beads, Chemicals, and Antibodies 347 348 Polystyrene bead standards (200, 500, and 800 nm diameter) with refractive index =1.59 and 349 Yellow-Green fluorescent label were obtained from Polysciences Inc. (Submicron Bead 350 Calibration Kit, Catalog no. BLI832, Bangs Laboratories Inc.). A bead standard mixture of 351 Yellow-Green fluorescent polystyrene beads with refractive index =1.59 (110 and 500 nm 352 diameter) and non-fluorescent silica beads with refractive index =1.43 (180, 240, 300, 585, 353 880, and 1300 nm diameter) was obtained from Apogee Flow Systems (Product #1493). 354 Unlabeled polystyrene beads (2.0 μm diameter) were obtained from Spherotech Inc. (Catalog 355 no. PP-20-10). CountBrightTM absolute counting beads were obtained from Invitrogen (Catalog 356 no. C36950). QuantumTM AlexaFluor488 MESF calibration bead kit was obtained from Bangs 357 Laboratories, Inc. (Catalog no. 488). 358 359 The following dyes were obtained from Molecular Probes/Invitrogen: Hoechst 33342 (Catalog 360 no. H3570), FM4-64 (Catalog no. T13320), calcein AM (Catalog no. C34852), calcein red-361 orange AM (Catalog no. C34851), LIVE/DEADTM Fixable Violet amine-reactive dye (Catalog no. 362 L34955), and MitoTracker Deep Red FM (Catalog no. M22426). Antibodies used in these 363 studies and their concentrations are summarized in Table 3. 364 365 Statistical Analyses 366 367 Gated data exported from FCS Express 6 were further processed in Excel and in R. Where 368 indicated, the robust standard deviation (rSD) was calculated using FCS Express 6 software. All 369 graphs were generated using the ggplot2 package in R, except those in Figure 8 and Figures 2-370 9 1C, 7-1C-F, and 8-1A-F which were generated in Excel. All statistical testing was conducted in 371 R. Where indicate-test was performed with the indicated contrasts (Fig. 5D, Fig. 372 6D-F, Fig. 6-1C). Where indicated, a two-way ANOVA was performed using FSC Gate and 373 Buffer as factors (Fig. 6H and Fig. 7B-C); degrees of freedom, F statistic, and p values for main 374 effects as well as interaction are listed in the figure captions. A summary of all statistical tests is 375 provided in the Statistical Table. 376 377 378 Results 379 380 FSC- and FM4-64-triggering of Crude Synaptosome Fraction P2 and Comparison to Bead 381 Standards 382 383 Conventional flow cytometers generally use forward scatter (FSC) measurements for detection 384 of cells, but the sensitivity of each instrument in FSC-triggered detection of submicron particles 385 varies widely depending on the refractive index of the particles, the angle of collection for 386 scattered light, detector sensitivity, and other factors (van der Pol et al., 2013). We used PS and 387 silica beads to determine the relative sensitivity of two cytometers: a BD LSR Fortessa analyzer 388 and a BD Influx sorter. We optimized FSC-threshold parameters on each instrument in order to 389 maintain a mixture of PS and silica bead standards ranging 585 -1300nm in the dynamic range 390 of detection, while simultaneously avoiding electronic and buffer noise. 391 392 As shown in Figure 1A, the lowest FSC signal resolvable above the background noise on the 393 Influx was the 585nm silica bead. Cross-referencing the FSC signal with the green fluorescence 394 signal allowed us to clearly identify PS beads apart from the non-fluorescent silica beads (Fig. 395 1A, right). As expected due to their higher refractive index, the FSC signals for PS beads were 396 substantially higher than for similarly sized silica beads. We also identified events that appeared 397 to be simultaneous detection of two 500nm PS beads, which we confirmed by comparing their 398 FSC and green fluorescence Figure 1-2, Stoner et 399 al., 2016). Although the FSC and fluorescent signals appeared additive in the 500 nm PS 400 doublet population (1.7-fold higher FSC, 1.8-fold higher green fluorescence), the FSC signal for 401 these events still fell below the 880nm silica beads. In contrast to the Influx, the Fortessa 402 displayed poor FSC resolution when running the silica/PS bead mixture (Fig. 1B), which 403 required separately running the individual PS beads (500nm and 800nm PS labeled with (B) in 404 Influx plots). The lowest FSC signal resolvable above the noise background on the Fortessa 405 406 only be separated by SSC or green fluorescence intensity (Fig. 1B). A complete summary of 407 FSC and SSC measurements for gated bead populations can be found in Figure 1-2. 408 409 Having established FSC-triggering limits and relative bead positions on the instruments, we next 410 sought to compare FSC- and fluorescence-411 samples. Despite contamination by a heterogeneous mixture of free mitochondria, myelin, and 412 membranous debris, P2 samples are enriched with synaptosomes and have been widely 413 employed in flow cytometry studies (Fein et al., 2008; Gylys et al., 2000; Postupna et al., 2014; 414 Prieto et al., 2017). As previously described by Biesemann et al. (2014), FM4-64 is a styryl dye 415 with favorable properties for fluorescence-triggered detection of synaptosomes. FM4-64 is 416 minimally fluorescent in aqueous media but becomes intensely fluorescent upon partitioning into 417 10 membranes (Vida and Emr, 1995), and its far-red emission is efficiently excited by a 488nm 418 laser, which is required for triggering on our cytometers. As shown in Figure 1C and 1D, FM-419 triggering enabled highly sensitive detection of P2 samples compared to FSC-triggering, with 420 more than half of particles in the P2 sample below the FSC noise threshold on both cytometers. 421 These results are consistent with the work of Biesemann (2010: Results 3.5, Fig. 15, pg. 87-90), 422 who found ~70% of particles in sucrose gradient synaptosome preparations were undetectable 423 by FSC-triggering on the BD FACS Aria. 424 425 Since previous studies have analyzed putative synaptosomal particles with FSC signals as high 426 as 4.5μm PS beads (Gylys et al., 2000), we sought to assess the size of biological particles in 427 such an FSC range. In the absence of a refractive index mismatch between PS beads and 428 biological particles, one would expect neuronal nuclei, which sediment in the P1 fraction, to 429 exhibit significantly higher FSC than 2μm PS beads due to their size. Although we had to lower 430 the FSC voltage to place 0.5, 0.8, and 2μm PS beads within the same dynamic range, 2.0 μm 431 PS beads were clearly resolved from 0.5 and 0.8 μm PS beads by FSC on both the Influx and 432 Fortessa (Fig. 1-1A). We identified neuronal nuclei based on Hoechst 33342 intensity and 433 NeuN immunofluorescence (Fig. 1-1B) and found that they were adjacent to and partially 434 overlapping with the FSC range defined by 2μm PS beads on the Influx (Fig. 1-1C). These 435 results confirm that FSC-based sizing relative to PS beads underestimates the size of biological 436 particles with lower refractive indices, and suggest that a majority of nuclei would be included in 437 FSC gates containing 4.5μm PS beads. 438 439 Serial dilutions define a range of linear detection for FM4-64-triggering 440 441 It has previously been demonstrated that high concentrations of particles below the trigger 442 threshold can be detected when they simultaneously occupy the focal point of illumination 443 (Nolan and Stoner, 2013; van der Pol et al., 2012). This phenomenon of coincident detection 444 445 FSC-triggered detection of submicron particles (Nolan and Stoner, 2013). In order to define a 446 linear range of particle detection and further compare FSC- vs. FM-triggering, we conducted 447 dilution series on both Influx and Fortessa. Compared to FM-triggering, FSC-triggered P2 448 samples did not maintain a constant FSC vs. FM4-64 profile on either the Influx or Fortessa 449 (Fig. 2A-B and Fig. 2-1). Specifically, the abundance of events near the FSC threshold with low 450 FM4-451 10 events/sec) became more apparent with dilution on the Fortessa. On the Influx, 452 quantification of the event rate across the dilution series revealed a sub-linear profile for both 453 FSC- and FM-triggering, which was fit with high accuracy using a quadratic model (Fig. 2C, left 454 panel, R2 = 0.99 for both series). FM-triggering was similar on the Fortessa, but FSC-triggering 455 was better fit with a linear model (Fig. 2D, left panel, R2 = 0.93 for FSC and R2 = 0.99 for FM 456 series). These results confirm that event rate saturates at high particle concentrations 457 (Kormelink et al., 2016) and that FSC-triggering underestimates the true event rate across the 458 entire dilution series. 459 460 Given that much of the P2 remains undetected by FSC-triggering, we also wondered whether 461 particle concentration might alter the FSC and FM fluorescence measurements. On the Influx, 462 the median FSC remained relatively constant for both FSC- and FM-triggering (Fig. 2C, middle 463 panel), while the median FM4-64 fluorescence varied dramatically for FSC-triggering but not for 464 11 FM-triggering (Fig. 2C, right panel). On the Fortessa, both median FSC and median FM4-64 465 fluorescence varied dramatically across the dilution series for FSC-triggering, but not FM-466 triggering (Fig. 2D, middle and right panels). The increase in median FM4-64 fluorescence with 467 dilution seen in FSC--468 above. Given the heterogeneous population of particles contained in P2 samples, we cannot 469 make strong conclusions about the mechanisms underlying these effects. Nonetheless, it is 470 clear that FM-triggering is superior to FSC-triggering in terms of the stability and accuracy of 471 event rate, FM4-64 fluorescence, and FSC of P2 particles across the dilution series on both 472 instruments. We conducted all further studies using only FM-triggering and maintaining 473 reasonably dilute samples so as to stay in the most linear range of the event rate dilution curves 474 (Fig. 2C and 2D, left panels, fewer than 10,000 events/sec on the Influx and fewer than ~3,000 475 events/sec on the Fortessa). Since sample flow rates and event rates may vary across 476 cytometers, we also measured the absolute particle concentration across a typical dilution 477 series on the Influx using volumetric count beads (Figure 2-1C). We found that the most linear 478 range of the event rate curve was for samples below ~1000 particles/μL. A previous study on 479 the Influx (Kormelink et al., 2016) showed that the trigger pulse baseline becomes continuously 480 elevated at high event rates where coincidence is prominent, so we used a digital oscilloscope 481 on the Influx to confirm that the trigger pulse remained at baseline under these conditions (data 482 not shown). 483 484 Development of a multicolor fluorescence assay for false colocalization 485 Although FM-triggering and dilution should help to reduce coincident particle detection, another 486 -487 aggregation of multiple particles (Erdbrügger et al., 2014). Multiple microscopy studies have 488 noted the presence of aggregated synaptosomes (Choi et al., 2009; Daniel et al., 2012; Gray 489 and Whittaker, 1962), but conventional flow cytometry does not provide a means to verify 490 whether events truly represent single particles. We sought to design a fluorescence-based flow 491 -Fig. 3). 492 The assay workflo493 (Krutzik and Nolan, 2006; Krutzik et al., 2011). A P2 sample is first split and single-labeled with 494 spectrally separated, non-transferable fluorescent dyes. After washing, single-labeled samples 495 are mixed prior to flow cytometry (Fig. 3A). Single particle events should therefore bear only 496 one fluorescent label, while coincidences and aggregates derived from separate single-labeled 497 samples will be positive for multiple fluorescent labels (Fig. 3B). Although the samples are 498 washed twice to remove any residual dye, it is critical to use dyes that cannot be transferred 499 between particles. Synaptosomes are well-labeled with Calcein AM dyes (Daniel et al., 2012; 500 Gylys et al., 2000; Prieto et al., 2017), lipophilic dyes that rapidly cross cell membranes and 501 become trapped following hydrolysis of the AM ester. In addition to calcein AM and Calcein 502 Red--reactive dye (covalent 503 reaction with proteins) and MitoTracker Deep Red FM (thiol-conjugation and retention in 504 mitochondria) to expand the number of dyes in some experiments. As shown in Figure 3-1A, 505 samples single-labeled with any of the four dyes do not exhibit any spectral overlap. 506 507 A representative set of density plots for a P2 sample single-labeled with calcein AM or calcein 508 red, as well as a mixture of the two, on the Influx is shown in Figure 3C. Note that because 509 even the background fluorescence is linearly related to FSC on the log-log plots, quadrant-510 12 based gating of double-positive events underestimates the number of positive events. To avoid 511 this problem, we used polygon gating of each fluorophore individually vs. FSC (Fig. 3C). 512 Boolean logic is used to determine the number of double-positive events, which can be 513 visualized by displaying only the positive-gated population of one fluorophore on a plot of FSC 514 vs. the other fluorophore (Fig. 3C, right). 515 516 Having verified the presence of calcein double-positive events in P2 mixtures (Fig. 3C), we 517 sought experimental confirmation that the dyes did not transfer between particles. Because only 518 the cleaved, polar dye molecules are retained, we reasoned that even if some membrane 519 disruption and dye leak occurred during mixing and centrifugation, the polar molecules would be 520 unable to label surrounding intact particles bearing the other calcein dye. We confirmed this by 521 sonicating a P2 sample truly double-labeled with both calcein dyes and adding an unlabeled P2 522 sample to the sonicated solution using the exact same experimental workflow. As shown in 523 Figure 3-1B, no calcein AM- or calcein red-positive events were detected. This demonstrates 524 that even after complete membrane disruption to liberate the entire dye content of one sample, 525 we do not observe measurable transfer to an unlabeled sample after washing and mixing. 526 527 Aggregation, but not coincidence, causes false double-positive events in P2 samples 528 529 With our multicolor fluorescence assay in hand, we turned towards distinguishing between 530 coincidence and aggregation. Similar to previous studies (Libregts et al., 2018), we used 531 mixtures of green- and red-fluorescent 500nm PS beads to model coincidence. We repeated the 532 dilution series experiments with mixtures of beads or mixtures of single calcein-labeled P2 533 samples on both Influx (Fig. 4A-B) and Fortessa (Fig. 4-1A-B). For bead mixtures, we observed 534 a clear population of double-positive events (red- and green-bead coincidences) on both 535 instruments. At high concentrations, the Fortessa performed particularly poorly with regard to 536 coincidence, with a wide smear of red-fluorescent bead multiplets and double-positive events 537 (Fig. 4-1A). On both instruments, we observed a linear reduction in the number of double-538 positive bead events with dilution, with virtually none detected at event rates 500 events/sec 539 (Fig. 4C-D). In contrast, the percentage of double-positive events from P2 samples remained 540 relatively constant across the entire dilution series. P2 double-positive events persisted even at 541 event rates 100 events/sec, which strongly suggests that coincidence is not the source of 542 these false double-positive events (Fig. 4C-D). 543 544 We speculated that the double-positive events might come from the physical association of two 545 separate fluorescent particles in the P2 samples. To further probe the nature of these events, 546 we sorted double-positive events from bead or P2 mixtures on the Influx and reanalyzed the 547 sorted material (Fig. 5A-B). Sorting of bead double-positive events failed to increase the 548 frequency of these events in the sorted sample, with reanalysis showing a ~4-fold decrease 549 (Fig. 5C). Consistent with the dilution experiments, this result suggests that the majority of 550 green- and red-fluorescent beads are not physically associated with each other during detection 551 of double-positive events. In contrast, sorting of P2 double-positive events increased the 552 frequency of these events ~3-fold in the sorted sample reanalysis (Fig. 5C). We note that a 3-553 fold increase in fluorescent particles during sort reanalysis is similar to that achieved by 554 (Biesemann et al., 2014) for fluorescent VGLUT1+ synaptosomes, suggesting that P2 double-555 positive events behave as stable, 556 high pressures encountered during sorting physically disrupt some aggregates, as suggested by 557 13 Biesemann (2010: Results 3.2, Fig. 12, pg. 79-81), which would further decrease the frequency 558 of double-positive events in reanalysis. 559 560 The divergent behavior of bead and P2 samples in dilution and sorting experiments strongly 561 suggests that the majority of P2 double-positive events are aggregates. Since we routinely 562 vortex samples at high speed for 5 seconds prior to flow cytometry, this suggests that the 563 double-positive aggregates are relatively stable. We hypothesized that these aggregates form 564 during centrifugation and are not fully disrupted during resuspension and filtration. To test this 565 hypothesis, we conducted the multicolor dye-labeling assay with and without centrifuging the 566 sample mixture and monitored the samples over time on the Influx. As shown in Figure 5D, 567 centrifugation-resuspension significantly increased the frequency of double-positive events 568 compared to mixing without centrifugation. Intriguingly, the frequency of double-positive events 569 increased over time for nearly every sample (Fig. 5D, left panel). 570 571 Taken together with the dilution and sorting experiments, these results suggest that particle 572 aggregation in P2 samples is an ongoing process that is accelerated by centrifugation. We 573 emphasize that our assay actually underestimates the frequency of true aggregates; green-574 green or red-red aggregates are not detected as double-positive events, but such aggregates 575 are surely formed during centrifugation and washing of single-labeled samples. The mixing of 576 single-labeled samples without centrifugation shown in Figure 5D ples) actually 577 underestimates the true proportion of aggregates, and we therefore conducted all other 578 experiments in this study with centrifugation of the mixtures. 579 580 Aggregation of cells can also affect conventional flow cytometry experiments, and we wondered 581 582 P2 samples. We plotted the trigger pulse width (in this case, FM4-64) vs. FSC on the Influx, or 583 FSC-H vs. FSC-A on the Fortessa, for events gated as single or double calcein-positive (Fig. 5-584 1585 gated apart from the linear singlet profile (Wersto et al., 2001). We found no such deviation from 586 linearity in single-positive gated events, which represent the majority of the sample. Although 587 they bear only one fluorophore, single-positive events are expected to contain at least as many 588 or even more aggregates than double-positive gated events (i.e. green-green, red-red, red-589 nonfluorescent, green-nonfluorescent, etc.). The lack of deviation from linearity for any events in 590 these pulse prof591 recent flow cytometry study using this gating procedure in analysis of FSC-triggered P2 samples 592 (Prieto et al., 2017). The lack of a 593 -594 positive events, which are predominantly aggregates and display a similar linear pulse profile to 595 single-positive events (Fig. 5-1). Thus, we found no evidence that pulse profile is an effective 596 means of removing false double-positive aggregates from downstream analysis of submicron 597 particles. 598 599 600 Double-positive event frequency increases with FSC and is reduced in nonionic buffer 601 602 Although bead standards do not provide accurate size estimation of biological particles, we 603 wondered whether the relative size information afforded by FSC intensity would correlate with 604 14 the presumably larger size of aggregates. We analyzed five discrete regions ranging from low to 605 high FSC on the Influx, based on beads and the FSC noise threshold (Fig. 6A-B). A 606 representative overlay of the FSC gates onto gated double-positive P2 events is shown in 607 Figure 6C. Strikingly, the top two FSC gates contain ~50% of all double-positive events 608 (percentages in red) but only ~19% of all calcein-positive events (percentages in black), while 609 the bottom two FSC gates contain only ~25% of all double-positive events amongst ~65% of all 610 calcein-positive events. Accordingly, the percentage of fluorescent events within each FSC gate 611 that were double-positive steadily increased with FSC (Fig. 6C, percentages in blue). For all 612 fluorescent events within each FSC gate, we quantified the percentage of single-, double-, and 613 triple-positive events on the Influx (Fig. 6D-F). Although the overall frequency of single-positives 614 was ~90%, this steadily decreased from ~97% in the FSC Noise region to ~70% in the region 615 above or equal to 1300nm silica beads (Fig. 6D). The overall frequency of double-positives was 616 ~10%, but steadily increased from ~2% in the FSC Noise region to ~27% in the region above or 617 equal to 1300nm silica beads (Fig. 6E). Similarly, the overall frequency of triple-positives was 618 ~2%, but increased from ~0% in the FSC Noise range to ~12% in the region above or equal to 619 1300nm silica beads (Fig. 6F). These results suggest that although aggregates are present in 620 all FSC regions, their presence is strongly correlated with increasing FSC. Again, we emphasize 621 that our fluorescence assay underestimates aggregation, even in samples where the measured 622 double-positive frequency exceeds 40% in the highest FSC regions (e.g. Fig 6C). 623 624 Several groups familiar with microscopic analysis of synaptosomal preparations have noted that 625 ionic buffers (e.g. PBS) cause synaptosomes to aggregate, while nonionic media (e.g. sucrose 626 buffer) favors an even distribution of single particles (Choi et al., 2009; Daniel et al., 2012). We 627 repeated our double calcein assay on the Influx using the sucrose/EDTA/Tris (her628 buffer recommended by Daniel et al. (2012). Because many assays are not possible in nonionic 629 media, we tested labeling in PBS and then switching to SET for flow cytometry (designated 630 -Fig. 6G). Compared to PBS, 631 both PBS-SET and SET samples displayed similar FSC vs. FM4-64 profiles (Fig. 6-1A), 632 indicating that SET did not impair FM-triggered particle detection or FSC measurement. FSC 633 and green-fluorescence measurements of PS beads in PBS and SET were also 634 indistinguishable (Fig. 6-1B). However, we noticed that the total number of calcein-positive 635 events was reduced to ~20% in when calcein-labeling was conducted in SET, compared to 636 ~40% for PBS or PBS-SET (Fig. 6-1C). We suspect that the nonionic media and divalent cation 637 sequestration by EDTA somewhat reduce the esterase activity of synaptosomes; nonetheless, 638 the labeling efficiency was sufficient to analyze a large number of labeled particles present in 639 these mixtures. As shown in Figure 6H, the percentage of fluorescent events within each FSC 640 gate that were double-positive was reduced in SET compared to PBS. A stepwise pattern was 641 observed, wherein the double-positive frequency was highest for PBS, lower for PBS-SET, and 642 lowest for SET. Two-way ANOVA revealed highly significant main effects of both FSC gate (Df = 643 5, F = 20.06, p=2.7e-15) and buffer (Df = 2, F = 20.14, p=1.8e-8) as well as a significant 644 interaction (Df = 10, F = 2.78, p=0.004). The significant interaction makes sense given that 645 buffer has little effect in the lower FSC regions, while the reduction of double-positive events is 646 greater in the higher FSC regions (Fig. 6H). These results provide further evidence that double-647 positive events are aggregates, and that such aggregates are less abundant in nonionic SET 648 buffer. 649 650 False colocalization of antigens in immunostained P2 samples 651 15 652 Given reports claiming high purity of synaptosomes in FSC ranges between 500  1500nm PS 653 beads (Gylys et al., 2004; Postupna et al., 2014; Prieto et al., 2017), we suspected that many of 654 the aggregates in our P2 samples contain synaptosomes. Although the violet amine reactive 655 and calcein dyes efficiently label many particles in P2 samples, they do not provide information 656 about the identity of these particles. We conducted three-color immunostaining of P2 samples 657 on the Influx, targeting presynaptic markers expressed by excitatory, inhibitory, and 658 monoaminergic neurons, respectively: VGLUT1, vesicular GABA transporter (VGAT), and 659 vesicular monoamine transporter 2 (VMAT2). Although we did not attempt immunostaining in 660 SET buffer, we did mimic the PBS-SET condition described above by resuspending stained and 661 washed samples in SET for flow cytometry. As shown in Figure 7A, immunopositive events for 662 all three transporters were observed across the entire FSC range and reflect their abundance in 663 the brain (VGLUT1 VGAT VMAT2). Looking at all immunopositive events across FSC gates 664 (Fig. 7-1A), VGAT+ and VMAT2+ events displayed an asymmetric distribution centered below 665 500nm PS beads, while VGLUT1+ events were distributed fairly uniformly just below 880nm 666 silica beads. Although the distribution of all immunopositive events across FSC gates was not 667 affected by buffer (Fig. 7-1B), the percentage of immunopositive events within each FSC gate 668 was slightly reduced in PBS-SET samples (Fig. 7B). Two-way ANOVA revealed a significant 669 main effect of buffer for VGAT (Df = 1, F = 4.70, p=0.034) and VMAT2 (Df = 1, F = 5.21, 670 p=0.026) but not VGLUT1 (Df = 1, F = 0.36, p=0.55). A highly significant main effect of FSC 671 gate was observed for all three markers (Df = 5, F = 21.51, p=2.9e-12 for VGAT; Df = 5, F = 672 68.9, p 2e-16 for VGLUT1; Df = 5, F = 12.54, p=2.4e-8 for VMAT2), where the percentage of 673 immunopositive events within each FSC gate increased with FSC (Fig. 7B). All buffer x FSC 674 gate interactions were non-significant. Consistent with recent studies (Prieto et al., 2017), the 675 percentage of VGLUT1+ events in the highest FSC gate approaches ~60%. 676 677 Although specific cases of neurotransmitter co-transmission and co-release have been 678 observed (reviewed in Hnasko and Edwards, 2012; Vaaga et al., 2014), the expression of 679 VGLUT1, VGAT, and VMAT2 should be mutually exclusive for the majority of presynaptic 680 terminals. For all immunopositive events within each FSC gate, we quantified the percentage of 681 single-, double-, and triple-positive events (Fig. 7C). Similar to double calcein-positive events, 682 we observed a steady decrease in single-positive and a steady increase in double- and triple-683 positive events with increasing FSC. Two-way ANOVA revealed a significant main effect of FSC 684 gate for each event type (Df = 5, F = 19.83, p=1.3e-11 for single-positive; Df = 5, F = 23.25, 685 p=6.6e-13 for double-positive; Df = 5, F = 8.37, p=4.7e-6 for triple-positive). PBS-SET samples 686 had higher single-positive event frequencies and lower double-positive event frequencies 687 compared to PBS (Fig. 7C). The main effect of buffer was significant for single-positive (Df = 1, 688 F = 6.74, p=0.012) and double-positive (Df = 1, F = 8.00, p=0.006), but not for triple-positive 689 events (Df = 1, F = 1.91, p=0.17). None of the buffer x FSC gate interactions were significant. 690 Thus, although PBS-SET samples did display reduced aggregation, the effect was modest, 691 especially for triple-positive events. Similar to the percentage of immunopositive events within 692 each FSC gate, the frequency of false co-localization increases with FSC. To further illustrate 693 this point, we calculated the percentage of immunopositive events for each antigen that were 694 single-, double-, or triple-positive within each FSC gate (Figure 7-2). In the FSC noise region 695 we found that ~80% of VGAT+, ~75% of VGLUT1+, and ~63% of VMAT2+ events were single-696 positive. Strikingly, in the highest FSC gate, the single-positive percentage was only ~12% for 697 VGAT+, ~51% for VGLUT1+, and ~9% for VMAT2. These results suggest that many of the 698 16 699 synaptosomes. 700 701 The quantification of double-positive events described above represents the sum of the three 702 possible subtypes (i.e. VGAT+/VGLUT1+, VGLUT1+/VMAT2+, and VMAT2+/VGAT+). We 703 wondered whether the frequency of these double-positive subtypes was related to the overall 704 abundance of their respective antigens (Fig. 7-1C). We used the percentage of all events 705 immunopositive for each antigen to construct a simple probability model. In this model, a 706 double-positive event represents the random sampling of two events with replacement. The 707 probability of a particular double-positive subtype is therefore obtained by multiplying the 708 Fig. 7-1D). Consistent with the expected frequencies 709 of double-positive subtypes, we found that VGLUT1+/VGAT+ events were by far the most 710 abundant (~75%), followed by VGLUT1+/VMAT2+ (~20%), and finally VMAT2+/VGAT+ (~5%, 711 Fig. 7-1D712 crude synaptosome preparation. Furthermore, they strongly suggest that many double-positive 713 events observed in these immunostaining experiments are aggregated particles rather than 714 single synaptosomes co-expressing multiple vesicular transporters. 715 716 Although scatter measurements of submicron particles may vary dramatically across 717 cytometers, fluorescence measurements can be readily calibrated. We determined the 718 fluorescence intensity of VGAT/anti-mouse AlexaFluor488 immunostaining in molecules of 719 equivalent soluble fluorophore (MESF) using calibrated bead standards (Fig. 7-1E). Because 720 particle fluorescence is directly related to FSC intensity regardless of the immunostaining 721 procedure, many immunopositive events at a given FSC intensity would exhibit fluorescence 722 intensity less than or equal to immunonegative events at a higher FSC intensity nearby. We 723 therefore analyzed the median fluorescence intensity of gated immunopositive and 724 immunonegative events within each FSC gate (Fig. 7-1F). Numerical MESF data are also 725 included in Figure 7-3, which should facilitate comparison across laboratories. 726 727 Particle Recovery Following Repeated Centrifugation 728 729 Given the increased frequency of aggregates after centrifugation (Figure 5D), in higher FSC 730 regions (Fig. 6D-F), and in PBS (Fig. 6H), we wondered whether buffer and centrifugation might 731 interact to bias sample composition towards aggregates following repeated centrifugations. 732 Indeed, Gray and Whittaker (1962) previously reported increased recovery of acetylcholine in 733 the low-speed P1 pellet when using saline instead of sucrose buffer, which was interpreted as 734 aggregation in saline increasing the effective size of synaptosomes (and thus sedimentation at 735 lower speed). To assess whether multiple centrifugation steps might bias particle recovery in 736 PBS vs. SET, we conducted two sequential 5-minute centrifugations of P2 samples in PBS or 737 SET buffer at varying speeds and used flow cytometry with absolute counting beads to 738 determine the particle concentration of the pellets and supernatants (Figure 8A). We found that 739 under carefully controlled conditions with the absolute counting beads, PBS reduced particle 740 abundance by ~3-fold prior to any centrifugation (Figure 8B). It is possible that certain particles 741 in the P2 sample are susceptible to destruction in PBS, but several features of the data suggest 742 that aggregation contributes to the reduction in particle counts. First, the median fluorescence 743 intensity of membrane staining by FM4-64 was higher in PBS (Figure 8B), which was due to 744 increased relative abundance of brighter, larger events with higher FSC (Figure 8C-D). In line 745 17 with the findings of Gray and Whittaker (1962)746 more efficiently pelleted in PBS at lower relative centrifugal force (2500-5000xg) compared to 747 SET buffer (Figure 8D). The relative abundance of these particles was dramatically increased in 748 the two pellets, P2-1 and P2-2, compared to the input sample and two supernatants (Figure 749 8E). The relative abundance of the large particles in PBS samples was highest in 2500xg pellets 750 and lowest in 10000xg pellets, while no clear relationship between relative abundance and 751 centrifugation speed was observed in SET buffer (Figure 8F752 is effect 753 was modest compared to the increase in PBS vs. SET (Figure 8F). 754 755 As expected, the overall efficiency of particle recovery in the pellet of each spin increased with 756 centrifugation speed, with the exception of the first spin in PBS (Figure 8-1A-C). Similar to the 757 input samples, it appears that a combination of aggregation and/or destruction contribute to the 758 loss of particle counts in the first PBS spin, although the recovery in the second spin was higher 759 than in SET buffer. Critically, the first 5-minute spin at 2500-5000xg in either buffer leaves over 760 half the P2 particles in the supernatant (Figure 8-1D). Collectively, these data suggest that P2 761 samples do undergo a certain degree of further fractionation during repeated centrifugation-762 based wash steps. Although some centrifugation bias is inevitable regardless of buffer and 763 centrifugation speed, our data suggest that lower centrifugation speeds and ionic buffers will 764 bias the overall sample composition towards larger particles that are more likely to be 765 aggregates. 766 767 768 Discussion (2704 words) 769 770 The goal of most flow cytometry experiments is to accurately detect and quantify the 771 fluorescence intensity of single particles or cells in suspension. Towards that goal, we employed 772 a variety of experimental assays that can be used to optimize synaptosome flow cytometry 773 experiments. We have summarized our approach in Table 4 and will mirror the workflow in our 774 discussion. We note that this workflow could be employed for any type of small particle, but our 775 discussion will focus on aspects specific to synaptosomes. 776 777 FSC triggering and bead-based size estimation 778 779 It is now widely accepted that FSC-based size estimation using PS beads underestimates 780 biological particle size (Lannigan et al., 2016; Nolan, 2015). This consequence of the refractive 781 index mismatch has been experimentally validated in numerous flow cytometry studies 782 (Chandler et al., 2011; van der Pol et al., 2012; Simonsen, 2016). The extent of underestimation 783 depends on a variety of factors including the scattering parameters used, refractive indices of 784 the sample and beads, and optical configuration of the cytometer. Estimates of the lipid vesicle 785 size range defined by PS beads have been conducted on various cytometers and are shown 786 in Table 5. 787 788 Synaptosomes are heterogeneous in size, but a generally accepted size range is 0.5  1.0 μm 789 (Dunkley et al., 1986, 1988; Fernández-Busnadiego et al., 2013; Gray and Whittaker, 1962; 790 Hollingsworth et al., 1985; Williams et al., 2009). Although certain preparations contain larger 791 (Robinson et al., 1989), most 792 studies of cortical synaptosomes consistently obtain mean diameters of ~500-600 nm (Table 6). 793 Based on the synaptosome size ranges in Table 6 and vesicle size ranges defined by PS beads 794 18 in Table 5, a majority of single synaptosomes would be expected to produce lower FSC signals 795 than 500nm PS beads and therefore would be undetectable by FSC-triggering on many 796 cytometers. Indeed, we could detect neither 585nm silica beads nor more than half of P2 797 samples by FSC-triggering on the Fortessa (Fig. 1C-D). Although the greater sensitivity 798 afforded by the Influx enabled detection of 585nm silica beads above the FSC noise threshold, 799 more than half of the P2 sample was still undetectable by FSC-triggering (Fig. 1C-D). Our 800 results are consistent with those of Biesemann et al. (2014), who sorted fluorescent VGLUT1 801 synaptosomes from below the FSC noise threshold using FM-triggering on a BD FACS Aria. 802 These electron microscopy, immunostaining, and proteomic studies convincingly demonstrated 803 that single synaptosomes can be purified by sorting events that fall below the FSC noise 804 threshold, which was well below 750nm PS beads on their cytometer (Biesemann, 2010: 805 Results 3.5, Fig. 15, pg. 87-90). 806 807 In contrast, another widely adopted protocol (Gylys and Bilousova, 2017) implements FSC-808 triggering and specifically analyzes only events in the FSC range defined by 0.5  1.5 μm PS 809 beads. Based on the high percentage of events that carry immunofluorescent signal for synaptic 810 markers, these events are assumed to represent single synaptosomes and material below 811 500nm PS beads is regarded as debris (Gylys et al., 2004)812 - 4.5 μm PS beads (Gylys et al., 2000). 813 We found that the FSC range above 2 μm PS beads includes particles as large as neuronal 814 nuclei (Fig. 1-1C). To our knowledge, the only single synaptosomes approaching this size are 815 derived from mossy fiber terminals of the hippocampus and cerebellum, and accordingly these 816 synaptosomes sediment along with nuclei in the P1 fraction (Taupin et al., 1994; Terrian et al., 817 1988). In our hands, the FSC region above 1.3μm silica beads (which would fall below both 1.5 818 and 4.5μm PS beads) represents 5% of all events in the P2 sample (Fig. 6C) and contains the 819 highest frequency of false double-positive events (Fig. 6E-H and Fig. 7C). Our results strongly 820 suggest that a large fraction of events in these regions are synaptosome-containing aggregates, 821 while single synaptosomes are often found below 500nm PS beads and the FSC noise 822 threshold. 823 824 Fluorescence-triggering and dilution to control for coincidence 825 826 Maintaining an appropriately low sample concentration is critical to prevent coincident particle 827 detection in submicron flow cytometry experiments. It has been suggested that coincidence is 828 especially prominent when operating in FSC-trigger mode (Nolan and Stoner, 2013), which 829 might explain the variations in FSC and FM4-64 fluorescence we observed with dilution in FSC-830 trigger mode on both cytometers (Fig. 2C-D). Importantly, we found that FSC-trigger mode 831 substantially underestimates the true event rate in P2 samples on both cytometers. In such a 832 situation, undetected particles in the sample volume associated with a detectable particle can 833 contribute to the fluorescence measurements and/or be sorted in the same drop (Libregts et al., 834 2018). Thus, with FSC-triggering at high event rates, one cannot determine whether sorted 835 material accurately reflects the detection and sorting of single particles. Although we found that 836 aggregation was the major source of double-positive events in our samples, we did not conduct 837 these assays using FSC-trigger mode or at high event rates. Serial dilutions and fluorescence-838 triggering should be conducted in all submicron flow cytometry experiments in order to avoid 839 coincidence. 840 841 Detection of coincidence and aggregation 842 843 To our knowledge, our study is the first to systematically address coincidence and aggregation 844 in flow cytometric analysis of synaptosomes (Fig. 3). We chose a strategy analogous to a 845 19 --sequencing 846 systems, where a mixture of mouse and human cells is analyzed in order to quantify cell-doublet 847 rates (Klein et al., 2015; Macosko et al., 2015). Our dye labeling strategy is also employed in 848 high throughput flow cytometry studies, where cells from different conditions are given non-849 transferable fluorescent barcodes prior to mixing, staining, and acquisition in a single batch 850 (Krutzik et al., 2011). Although the cellular fluorescent barcodes are decoded with high accuracy 851 in these protocols, even when multiple barcodes are encoded by discrete concentrations of a 852 single dye (Krutzik and Nolan, 2006), our results demonstrate that synaptosome preparations 853 do not behave similarly to cells in these experiments. Determining the extent of single-particle 854 detection in submicron flow cytometry experiments is a challenging task. Image cytometry, 855 which combines confocal microscopy with the fluidics of a cytometer, is powerful in this regard 856 since images of each event can be manually examined after acquisition. However, these 857 instruments require extensive optimization and are not widely available to all labs. The 858 fluorescent labeling and mixing assay we describe here can be implemented on most 859 conventional flow cytometers and should be broadly useful to the microparticle flow cytometry 860 community. 861 862 Distinguishing between coincidence and aggregation 863 864 -865 dilution and FM-triggering of our P2 mixtures (Fig. 4C-D), we began to suspect aggregation as 866 the source of these events. Aggregated particles would beh867 of dilution and coincidence, but nonetheless present serious problems for both cellular and 868 submicron flow cytometry analysis. The large scatter signals produced by cells result in 869 abnormal width/height vs. area pulse profiles, which can be used to remove cell-doublet events 870 in conventional flow cytometry experiments. To date, we are unaware of any study 871 demonstrating that this gating strategy works on submicron particles, and our results strongly 872 suggest it does not (Fig. 3-1C). We further demonstrated that double-positive events in our P2 873 mixtures can be enriched by sorting (Fig. 5A-C), while fluorescent PS bead double-positive 874 events are depleted by sorting. These results are highly consistent with our dilution series and 875 provide further evidence that the dominant source of double-positive events in P2 mixtures is 876 aggregation. Combined with our fluorescent labeling and mixing assay, comparison to 877 fluorescent PS bead mixtures in dilution series and sorting experiments provides a general 878 strategy to determine the extent of coincidence vs. aggregation in any sample type. 879 880 Synaptosome aggregation 881 882 Single events comprised of aggregated cell-derived microparticles have been definitively 883 identified using image cytometry (Erdbrügger et al., 2014). It is unclear how often and under 884 what conditions cell-derived microparticles aggregate, but one can speculate that their 885 endogenous function within physiological bodily fluids would make them somewhat resistant to 886 aggregation. In contrast to cell-derived microparticles, presynaptic nerve terminals do not exist 887 as soluble particles in vivo, and the aggregation of synaptosomes in ionic media (saline-based 888 solutions such as PBS) is a well-described phenomenon. Gray and Whittaker (1962) analyzed 889 the minimum concentration of various electrolytes required to induce aggregation of particles in 890 the P2 and sucrose-gradient synaptosome fractions. They describe how the particles behave as 891 though negatively charged, since divalent and trivalent cations are particularly effective in 892 causing aggregation. However, the minimum concentration for sodium chloride was 20 mM 893 (Gray and Whittaker, 1962), far below the 137 mM present in PBS. Others encountered such 894 aggregates using a f895 in saline (Hollingsworth et al., 1985). 896 20 897 Most synaptosome flow cytometry studies have employed ionic media for incubation, washing, 898 and sample analysis, whilst assuming that each event generated by the flow cytometer 899 represents a single synaptosome (Fein et al., 2008; Gylys et al., 2000, 2004; Postupna et al., 900 2014; Prieto et al., 2017). Studies of cell-derived microparticles have clearly demonstrated 901 situations where this assumption breaks down (Erdbrügger et al., 2014; Libregts et al., 2018), 902 and this issue was recently brought forth in synaptosome flow cytometry (see comments on 903 Prieto et al., 2017). The centrifugation and time-lapse variations of our fluorescent mixture 904 experiments provide further insight into particle aggregation in P2 samples (Fig. 5D). Most 905 strikingly, we found that centrifugation and resuspension of the single-labeled mixture 906 dramatically increases the frequency of double-positive events. This finding strongly suggests 907 that resuspension of the P2 pellet does not completely eliminate particles that aggregate during 908 centrifugation. We also observed a steady increase in the frequency of double-positive events 909 over time with or without centrifugation (Fig. 5D), suggesting that aggregation is an ongoing 910 process in P2 sample mixtures. These results also further argue against coincidence as a cause 911 of these double-positive events, as coincidence would not be expected to increase over time. 912 Although washing steps are critical to these experiments, our data suggest that repeated 913 centrifugation will bias the sample composition. The enhanced relative abundance of larger, 914 brighter particles following centrifugation in PBS (Fig. 8D-F), especially at low speeds, is 915 consistent with previous work (Gray and Whittaker, 1962) and provides further evidence that 916 many of these high FSC events are aggregated particles. Future studies should carefully 917 consider the effects of duration, speed, and number of centrifugations on downstream 918 experimental results. 919 920 Reducing synaptosome aggregation 921 922 Analogous to the microscopy results of other groups (Choi et al., 2009; Daniel et al., 2012), we 923 found that SET buffer significantly reduced the frequency of double-positive events in our P2 924 mixtures, especially in the higher FSC gates (Fig. 6H). However, calcein-labeling efficiency was 925 also lower in SET buffer (Fig. 6-1C) and SET would not be compatible with many physiological 926 assays (Choi et al., 2009). We found that conducting the labeling procedure in PBS and 927 switching to SET buffer for flow cytometry also reduced the double-positive event frequency, 928 although the effect was smaller than for SET alone (Fig. 6H). Our flow cytometry assay provides 929 a quantitative measure (albeit an underestimate) of synaptosomal aggregation in each buffer 930 condition and should be useful to monitor the impact of future protocol improvements. Even 931 without the fluorescent mixture assay, we observed that SET buffer reduces the relative 932 abundance of particles in the highest FSC regions most likely to contain aggregates (Fig. 8D-F). 933 Anecdotally, we noticed that the P2 pellet disperses relatively easily in SET buffer compared to 934 PBS. Based on the findings of Gray and Whittaker (1962), we suspect that removing cations 935 reduces aggregation by allowing the negatively charged neuronal membranes to repel each 936 other. 937 938 Pitfalls in Immunophenotyping of Synaptosomes 939 940 The heterogeneity of presynaptic nerve terminals, including specific cases of neurotransmitter 941 co-release (Hnasko and Edwards, 2012; Vaaga et al., 2014), makes high-throughput 942 immunophenotyping of synaptosomes by neurotransmitter content an appealing direction for 943 synaptosome flow cytometry. Indeed, Gajera et al., (2019) have recently demonstrated high-944 dimensional analysis of human synaptosomal preparations using mass cytometry. Our results 945 21 suggest that the presence of synaptosomal aggregates is a serious problem for such 946 experiments, especially in the top two FSC ranges in our study (higher than 880nm silica beads, 947 Fig. 7C). We found implausibly high co-labeling of VGAT, VGLUT1, and VMAT2 in our P2 948 samples (Figure 7-2), and the increasing frequency of double- and triple-positive events 949 observed across FSC gates corroborates the analogous finding in our fluorescent mixture assay 950 (Fig. 6D-F, Fig. 7C). These findings further support our conclusion that a high percentage of 951 events in this FSC range are synaptosome-containing aggregates. As expected based on 952 synapse abundance, VGAT+/VGLUT1+ double-positive events were observed most frequently 953 (Fig. 7-1C-D). Similarly, Biesemann previously demonstrated co-enrichment of VGAT and 954 VGLUT1 by sorting VGLUT1 fluorescent events between 0.75-1.5 μm beads (Biesemann, 2010: 955 Results 3.4, Fig. 14, pg. 86-87). Western blotting of the sorted material revealed enrichment of 956 VGAT and myelin proteolipid protein (PLP), proteins that should be depleted in a sort of pure 957 VGLUT1+ synaptosomes. After implementing FM-triggering and sorting only VGLUT1 958 fluorescent events below the FSC noise threshold, the authors demonstrated unprecedented 959 purity of sorted VGLUT1 synaptosomes with the expected depletion of VGAT and other glial 960 markers (Biesemann et al., 2014). Overall, our results are highly consistent with their conclusion 961 that the previously sorted events were aggregates containing at least one glutamatergic 962 synaptosome (Biesemann, 2010: Results 3.4, pg. 86). 963 964 As others have pointed out (Erdbrügger et al., 2014), immunophenotyping of submicron 965 particles can result in detection, but not necessarily accurate quantification. Unfortunately, we 966 found that SET buffer was less effective in reducing false double- and triple-positive 967 immunostained events (Fig. 7C) compared to those observed in the fluorescent mixture assay 968 (Fig. 6H). Since resuspension of the P2 pellet is unlikely to completely eliminate aggregates, we 969 suspect that some aggregated particles are cross-linked during formaldehyde fixation prior to 970 immunostaining. Accordingly, we found that single-positive event frequency was generally 971 ~90% below the FSC noise threshold in immunostained samples (Fig. 7C), while the single-972 positive frequency in this region was generally ~97% in the calcein mixtures (Fig. 6D). Although 973 this suggests that some of the double-positive events in the immunostained samples are single 974 synaptosomes with true colocalization (e.g. VGAT+/VGLUT1+ nerve terminals in cortical layer 975 V, Fattorini et al., 2009), distinguishing such events from aggregates would require extensive 976 experimental validation. Flow cytometry studies claiming to study single synaptosomes often 977 report a high-degree of co-labeling for markers that are expected to colocalize, such as 978 VGLUT1, PSD95, synaptophysin, SNAP25, etc. (Gylys et al., 2004; Prieto et al., 2017). Such 979 data does not provide evidence that the detected particles are single synaptosomes, since 980 synaptosome-containing aggregates would be expected to yield similar, if not even higher co-981 labeling for these markers. Instead, we propose that such studies should explicitly report the co-982 labeling of neurotransmitter-specific markers (i.e. VGLUT1, VGAT, etc.) that should be mutually 983 exclusive for most single synaptosomes. As suggested previously (Nolan and Stoner, 2013), 984 claims of co-localization should be supported by direct evidence that coincidence and 985 aggregation are not contributing to the measurement. 986 987 Challenges and Future Directions 988 989 Here we highlighted a number of technical challenges that hinder the detection and analysis of 990 single particles in synaptosome flow cytometry. In the absence of new purification methods, 991 synaptosome preparations will inevitably be contaminated with membranes, myelin, and 992 mitochondria. Currently, it seems that transgenic mice harboring fluorescent protein reporters 993 are required for sorting, since tractable surface markers are limited and most downstream 994 assays are not compatible with fixation and immunostaining of cytoplasmic proteins. A sorting 995 protocol for VGLUT1Venus synaptosomes has described by Herzog and colleagues (Biesemann 996 22 et al., 2014; Luquet et al., 2017; see also Poulopoulos et al., 2019 for growth cone sorting) and 997 should be readily adaptable for other fluorescent proteins and synapse types. However, if 998 immunostained synaptosome events are the analytical endpoint for an experiment, the methods 999 described in our study will minimize the effects of coincidence and aggregation on the analysis 1000 of single synaptosomes. Such procedures will need to be routinely performed on each 1001 cytometer, and confirmation of accuracy by microscopy seems prudent. We remain optimistic 1002 that future advances in instrumentation, reagents, and sample preparation will obviate the need 1003 for such extensive quality control. Despite current limitations, we expect synaptosome flow 1004 cytometry studies to continue enhancing our understanding of synaptic biology in health and 1005 disease. 1006 23 References 1007 1008 Arraud, N., Gounou, C., Turpin, D., and Brisson, A.R. (2016). Fluorescence triggering: A general strategy 1009 for enumerating and phenotyping extracellular vesicles by flow cytometry. Cytometry A 89, 184–195. 1010 Biesemann, C. (2010). Development of Fluorescence Activated Synaptosome Sorting (FASS) and analysis 1011 of VGLUT1 synapses from mouse brain. Georg August University Göttingen. 1012 Biesemann, C., Grønborg, M., Luquet, E., Wichert, S.P., Bernard, V., Bungers, S.R., Cooper, B., 1013 Varoqueaux, F., Li, L., Byrne, J.A., et al. (2014). Proteomic screening of glutamatergic mouse brain 1014 synaptosomes isolated by fluorescence activated sorting. EMBO J. 33, 157–170. 1015 Chandler, W.L., Yeung, W., and Tait, J.F. (2011). A new microparticle size calibration standard for use in 1016 measuring smaller microparticles using a new flow cytometer. J. Thromb. Haemost. JTH 9, 1216–1224. 1017 Choi, S.W., Gerencser, A.A., and Nicholls, D.G. (2009). Bioenergetic analysis of isolated cerebrocortical 1018 nerve terminals on a microgram scale: spare respiratory capacity and stochastic mitochondrial failure. J. 1019 Neurochem. 109, 1179–1191. 1020 Daniel, J.A., Malladi, C.S., Kettle, E., McCluskey, A., and Robinson, P.J. (2012). Analysis of synaptic vesicle 1021 endocytosis in synaptosomes by high-content screening. Nat. Protoc. 7, 1439–1455. 1022 Dunkley, P.R., Jarvie, P.E., Heath, J.W., Kidd, G.J., and Rostas, J.A. (1986). A rapid method for isolation of 1023 synaptosomes on Percoll gradients. Brain Res. 372, 115–129. 1024 Dunkley, P.R., Heath, J.W., Harrison, S.M., Jarvie, P.E., Glenfield, P.J., and Rostas, J.A. (1988). A rapid 1025 Percoll gradient procedure for isolation of synaptosomes directly from an S1 fraction: homogeneity and 1026 morphology of subcellular fractions. Brain Res. 441, 59–71. 1027 Erdbrügger, U., Rudy, C.K., Etter, M.E., Dryden, K.A., Yeager, M., Klibanov, A.L., and Lannigan, J. (2014). 1028 Imaging flow cytometry elucidates limitations of microparticle analysis by conventional flow cytometry. 1029 Cytom. Part J. Int. Soc. Anal. Cytol. 85, 756–770. 1030 Fattorini, G., Verderio, C., Melone, M., Giovedì, S., Benfenati, F., Matteoli, M., and Conti, F. (2009). 1031 VGLUT1 and VGAT are sorted to the same population of synaptic vesicles in subsets of cortical axon 1032 terminals. J. Neurochem. 110, 1538–1546. 1033 Fein, J.A., Sokolow, S., Miller, C.A., Vinters, H.V., Yang, F., Cole, G.M., and Gylys, K.H. (2008). Co-1034 localization of amyloid beta and tau pathology in Alzheimer’s disease synaptosomes. Am. J. Pathol. 172, 1035 1683–1692. 1036 Fernández-Busnadiego, R., Asano, S., Oprisoreanu, A.-M., Sakata, E., Doengi, M., Kochovski, Z., Zürner, 1037 M., Stein, V., Schoch, S., Baumeister, W., et al. (2013). Cryo–electron tomography reveals a critical role 1038 of RIM1α in synaptic vesicle tethering. J Cell Biol 201, 725–740. 1039 Gajera, C.R., Fernandez, R., Postupna, N., Montine, K.S., Fox, E.J., Tebaykin, D., Angelo, M., Bendall, S.C., 1040 Keene, C.D., and Montine, T.J. (2019). Mass synaptometry: High-dimensional multi parametric assay for 1041 single synapses. J. Neurosci. Methods 312, 73–83. 1042 24 Gleitz, J., Beile, A., Wilffert, B., and Tegtmeier, F. (1993). Cryopreservation of freshly isolated 1043 synaptosomes prepared from the cerebral cortex of rats. J. Neurosci. Methods 47, 191–197. 1044 Gray, E.G., and Whittaker, V.P. (1962). The isolation of nerve endings from brain. J. Anat. 96, 79-88.8. 1045 Gray, W.D., Mitchell, A.J., and Searles, C.D. (2015). An accurate, precise method for general labeling of 1046 extracellular vesicles. MethodsX 2, 360–367. 1047 Gylys, K.H., and Bilousova, T. (2017). Flow Cytometry Analysis and Quantitative Characterization of Tau 1048 in Synaptosomes from Alzheimer’s Disease Brains. Methods Mol. Biol. Clifton NJ 1523, 273–284. 1049 Gylys, K.H., Fein, J.A., and Cole, G.M. (2000). Quantitative characterization of crude synaptosomal 1050 fraction (P-2) components by flow cytometry. J. Neurosci. Res. 61, 186–192. 1051 Gylys, K.H., Fein, J.A., Yang, F., and Cole, G.M. (2004). Enrichment of presynaptic and postsynaptic 1052 markers by size-based gating analysis of synaptosome preparations from rat and human cortex. Cytom. 1053 Part J. Int. Soc. Anal. Cytol. 60, 90–96. 1054 Hnasko, T.S., and Edwards, R.H. (2012). Neurotransmitter corelease: mechanism and physiological role. 1055 Annu. Rev. Physiol. 74, 225–243. 1056 Hollingsworth, E.B., McNeal, E.T., Burton, J.L., Williams, R.J., Daly, J.W., and Creveling, C.R. (1985). 1057 Biochemical characterization of a filtered synaptoneurosome preparation from guinea pig cerebral 1058 cortex: cyclic adenosine 3’:5’-monophosphate-generating systems, receptors, and enzymes. J. Neurosci. 1059 Off. J. Soc. Neurosci. 5, 2240–2253. 1060 Klein, A.M., Mazutis, L., Akartuna, I., Tallapragada, N., Veres, A., Li, V., Peshkin, L., Weitz, D.A., and 1061 Kirschner, M.W. (2015). Droplet barcoding for single cell transcriptomics applied to embryonic stem 1062 cells. Cell 161, 1187–1201. 1063 Kormelink, T.G., Arkesteijn, G.J.A., Nauwelaers, F.A., Engh, G. van den, Hoen, E.N.M.N.-’t, and Wauben, 1064 M.H.M. (2016). Prerequisites for the analysis and sorting of extracellular vesicle subpopulations by high-1065 resolution flow cytometry. Cytometry A 89, 135–147. 1066 Krishnaswami, S.R., Grindberg, R.V., Novotny, M., Venepally, P., Lacar, B., Bhutani, K., Linker, S.B., Pham, 1067 S., Erwin, J.A., Miller, J.A., et al. (2016). Using single nuclei for RNA-seq to capture the transcriptome of 1068 postmortem neurons. Nat. Protoc. 11, 499–524. 1069 Krutzik, P.O., and Nolan, G.P. (2006). Fluorescent cell barcoding in flow cytometry allows high-1070 throughput drug screening and signaling profiling. Nat. Methods 3, 361–368. 1071 Krutzik, P.O., Clutter, M.R., Trejo, A., and Nolan, G.P. (2011). Fluorescent cell barcoding for multiplex 1072 flow cytometry. Curr. Protoc. Cytom. Chapter 6, Unit 6.31. 1073 Lannigan, J., Nolan, J.P., and Zucker, R. (2016). Measurement of extracellular vesicles and other 1074 submicron size particles by flow cytometry. Cytometry A 89, 109–110. 1075 Libregts, S.F.W.M., Arkesteijn, G.J.A., Németh, A., Nolte-’t Hoen, E.N.M., and Wauben, M.H.M. (2018). 1076 Flow cytometric analysis of extracellular vesicle subsets in plasma: impact of swarm by particles of non-1077 interest. J. Thromb. Haemost. JTH 16, 1423–1436. 1078 25 Luquet, E., Biesemann, C., Munier, A., and Herzog, E. (2017). Purification of Synaptosome Populations 1079 Using Fluorescence-Activated Synaptosome Sorting. Methods Mol. Biol. Clifton NJ 1538, 121–134. 1080 Macosko, E.Z., Basu, A., Satija, R., Nemesh, J., Shekhar, K., Goldman, M., Tirosh, I., Bialas, A.R., Kamitaki, 1081 N., Martersteck, E.M., et al. (2015). Highly Parallel Genome-wide Expression Profiling of Individual Cells 1082 Using Nanoliter Droplets. Cell 161, 1202–1214. 1083 Nolan, J.P. (2015). Flow Cytometry of Extracellular Vesicles: Potential, Pitfalls, and Prospects. Curr. 1084 Protoc. Cytom. 73, 13.14.1-13.14.16. 1085 Nolan, J.P., and Stoner, S.A. (2013). A trigger channel threshold artifact in nanoparticle analysis. Cytom. 1086 Part J. Int. Soc. Anal. Cytol. 83, 301–305. 1087 van der Pol, E., van Gemert, M.J.C., Sturk, A., Nieuwland, R., and van Leeuwen, T.G. (2012). Single vs. 1088 swarm detection of microparticles and exosomes by flow cytometry. J. Thromb. Haemost. JTH 10, 919–1089 930. 1090 van der Pol, E., Coumans, F., Varga, Z., Krumrey, M., and Nieuwland, R. (2013). Innovation in detection of 1091 microparticles and exosomes. J. Thromb. Haemost. JTH 11 Suppl 1, 36–45. 1092 Postupna, N.O., Keene, C.D., Latimer, C., Sherfield, E.E., Van Gelder, R.D., Ojemann, J.G., Montine, T.J., 1093 and Darvas, M. (2014). Flow cytometry analysis of synaptosomes from post-mortem human brain 1094 reveals changes specific to Lewy body and Alzheimer’s disease. Lab. Investig. J. Tech. Methods Pathol. 1095 94, 1161–1172. 1096 Poulopoulos, A., Murphy, A.J., Ozkan, A., Davis, P., Hatch, J., Kirchner, R., and Macklis, J.D. (2019). 1097 Subcellular transcriptomes and proteomes of developing axon projections in the cerebral cortex. Nature 1098 565, 356–360. 1099 Prieto, G.A., Trieu, B.H., Dang, C.T., Bilousova, T., Gylys, K.H., Berchtold, N.C., Lynch, G., and Cotman, 1100 C.W. (2017). Pharmacological Rescue of Long-Term Potentiation in Alzheimer Diseased Synapses. J. 1101 Neurosci. 37, 1197–1212. 1102 Robinson, P.J., Gehlert, D.R., Sanna, E., and Hanbauer, I. (1989). Two fractions enriched for striatal 1103 synaptosomes isolated by percoll gradient centrifugation: synaptosome morphology, dopamine and 1104 serotonin receptor distribution, and adenylate cyclase activity. Neurochem. Int. 15, 339–348. 1105 Simonsen, J.B. (2016). A liposome-based size calibration method for measuring microvesicles by flow 1106 cytometry. J. Thromb. Haemost. 14, 186–190. 1107 Stoner, S.A., Duggan, E., Condello, D., Guerrero, A., Turk, J.R., Narayanan, P.K., and Nolan, J.P. (2016). 1108 High sensitivity flow cytometry of membrane vesicles. Cytometry A 89, 196–206. 1109 Taupin, P., Zini, S., Cesselin, F., Ben-Ari, Y., and Roisin, M.-P. (1994). Subcellular Fractionation on Percoll 1110 Gradient of Mossy Fiber Synaptosomes: Morphological and Biochemical Characterization in Control and 1111 Degranulated Rat Hippocampus. J. Neurochem. 62, 1586–1595. 1112 Terrian, D.M., Johnston, D., Claiborne, B.J., Ansah-Yiadom, R., Strittmatter, W.J., and Rea, M.A. (1988). 1113 Glutamate and dynorphin release from a subcellular fraction enriched in hippocampal mossy fiber 1114 synaptosomes. Brain Res. Bull. 21, 343–351. 1115 26 Vaaga, C.E., Borisovska, M., and Westbrook, G.L. (2014). Dual-transmitter neurons: Functional 1116 implications of co-release and co-transmission. Curr. Opin. Neurobiol. 0, 25–32. 1117 Vida, T.A., and Emr, S.D. (1995). A new vital stain for visualizing vacuolar membrane dynamics and 1118 endocytosis in yeast. J. Cell Biol. 128, 779–792. 1119 Wersto, R.P., Chrest, F.J., Leary, J.F., Morris, C., Stetler-Stevenson, M., and Gabrielson, E. (2001). Doublet 1120 discrimination in DNA cell-cycle analysis. Cytometry 46, 296–306. 1121 Whittaker, V.P. (1993). Thirty years of synaptosome research. J. Neurocytol. 22, 735–742. 1122 Williams, C., Shai, R.M., Wu, Y., Hsu, Y.-H., Sitzer, T., Spann, B., McCleary, C., Mo, Y., and Miller, C.A. 1123 (2009). Transcriptome Analysis of Synaptoneurosomes Identifies Neuroplasticity Genes Overexpressed 1124 in Incipient Alzheimer’s Disease. PLOS ONE 4, e4936. 1125 1126 27 Figure Captions 1127 1128 Figure 1: Determining Detection Sensitivity Using Bead Standards and FSC- vs. FM-Triggering 1129 of P2 Fraction 1130 1131 (A) Mixture of non-fluorescent silica/fluorescent PS beads (Apogee; A) and fluorescent PS 1132 beads (Bangs Labs; B) detected in FSC-trigger mode on the Influx. Left: forward and side 1133 scatter, right: forward scatter and green fluorescence detector (FITC) allows fluorescent PS 1134 beads and bead doublets to be clearly distinguished from non-fluorescent silica beads. (B) 1135 Mixture of non-fluorescent silica/fluorescent PS beads (Apogee; A) detected in FSC-trigger 1136 mode on the Fortessa. Note that the Bangs Labs 500nm and 800nm fluorescent PS beads 1137 could not be run simultaneously with the Apogee mixture on the Fortessa due to poor FSC 1138 resolution. (C) FM4-64 stained P2 sample on the Influx. Left: FSC-trigger mode, right: FM-1139 (D) FM4-64 stained P2 sample 1140 on the Fortessa. Left: FSC-trigger mode, right: FM-1141 running clean PBS. See Figure 1-2 for median FSC and SSC values from gated bead 1142 populations. See also Figure 1-1. 1143 1144 1145 Figure 2: Dilution Series Define a Range of Linear Particle Detection in P2 Samples 1146 1147 (A) Representative density plots of a P2 dilution series detected on the Influx in FSC-trigger 1148 mode. FM4-64 Low gate was set arbitrarily to quantify disappearance of events in this region 1149 with dilution. (B) Representative density plots of same P2 dilution series as in A, but detected on 1150 the Influx in FM-trigger mode. (C) Measurements from full P2 dilution series on the Influx. Left: 1151 Event rate data fit with a quadratic model. Middle: Median FSC of all detected events. Right: 1152 Median FM4-64 fluorescence of all detected events. (D) Measurements from full P2 dilution 1153 series on the Fortessa. Left: FM-trigger series fit with a quadratic model and FSC-trigger series 1154 fit with a linear model. Middle: Median FSC of all detected events. Right: Median FM4-64 1155 fluorescence of all detected events. See also Figure 2-1. 1156 1157 1158 Figure 3: Experimental Assay for Measuring Coincidence and Aggregation 1159 1160 (A) Schematic representation of the P2 sample labeling workflow. A single sample is split into 1161 equal portions for single-color labeling, followed by two washes to remove residual dye. The 1162 samples are then mixed and washed a third time before flow cytometry. (B) Schematic 1163 representation of particle detection and artifacts in microparticle flow cytometry. Particles flow in 1164 the sample fluid, which is surrounded by sheath fluid. Note that the width of the sample stream 1165 is enlarged for illustrative purposes; we run our cytometers at the minimum sample pressure 1166 (and thus sample stream width) in order to ensure optimal hydrodynamic focusing. When the 1167 trigger parameter detector crosses a user-defined threshold, the time duration of this pulse is 1168 then integrated on all 1169 occur when multiple particles simultaneously occupy the focal point of laser illumination 1170 (coincidence) or are physically bound (aggregates). (C) Representative density plots of a P2 1171 sample from the double-calcein assay on the Influx. Left: Calcein AM-labeled P2 sample shows 1172 no cross-emission in the calcein red detector. Mid-left: Calcein Red-labeled P2 sample shows 1173 28 no cross-emission in the calcein AM detector. Mid-right: Mixture of single-labeled P2 samples is 1174 gated for calcein AM+ and calcein red+ particles, which are then cross-checked for the other 1175 fluorophore (diagonal arrows). Right: Calcein AM+ or calcein red+ gated particles are analyzed 1176 to quantify double calcein+ events. See also Figure 3-1. 1177 1178 1179 Figure 4: Dilution Series Reduces False Double-Positive Bead Events but Not Double-Calcein 1180 Positive P2 Events 1181 1182 (A) Representative density plots of a dilution series for a mixture of fluorescent 0.5 μm PS 1183 beads detected using FSC-triggering on the Influx. Individual PS beads and double-positive 1184 events are distinguished by their green (FITC detector) or red (PE detector) fluorescence. (B) 1185 Representative density plots of a dilution series for a mixture of single calcein-labeled P2 1186 samples detected using FM-triggering on the Influx. Only events gated positive for calcein red 1187 are displayed. (C) Measurements from full dilution series of bead mixture or single calcein-1188 labeled P2 mixtures (n=3) on the Influx. P2 data are displayed as mean +/- SEM. Left: Event 1189 rate. Right: Double-positive events expressed as a percentage of all fluorescent events. (D) 1190 Measurements from full dilution series of bead or single calcein-labeled P2 mixtures on the 1191 Fortessa. Left: Event Rate. Right: Double-positive events expressed as a percentage of all 1192 fluorescent events. See also Figure 4-1. 1193 1194 1195 Figure 5: Effects of Fluorescence-Activated Sorting, Centrifugation-Resuspension, and Time 1196 Lapse on Double-Positive P2 Events. 1197 1198 These experiments were conducted using FM-triggering on the Influx. (A) Representative 1199 density plots of pre-sorted samples. Left: Mixture of fluorescent 0.5 μm PS beads on the Influx. 1200 TMid left and right: Mixture of single calcein-1201 labeled P2 samples. Right: Calcein AM+ gated events from single calcein-labeled P2 samples 1202  (B) Representative 1203 density plots of sorted samples re-analyzed on the Influx. Left1204 Mid left and rightRight: 1205 le. (C) Summary of pre-sort vs. 1206 reanalysis of sorted double-positive bead and P2 (n=3) samples. P2 data are plotted as mean 1207 +/- SEM. (D) Comparison of time lapse data for centrifugation-1208 gle-labeled P2 samples. Left: Time lapse of double-positive 1209 events plotted expressed as a percentage of all fluorescent events. Samples include the 1210 following mixtures of single-labeled P2 samples: violet amine-reactive dye/MitoTracker Deep 1211 Red FM (n=2), calcein AM/calcein red (n=3), violet amine-reactive dye/calcein AM (n=1), violet 1212 amine-reactive dye/calcein AM/calcein red (n=2), all four dyes (n=2). Right: Comparison of 1213 averages for the individual trajectories shown on the left, plotted as mean +/- SEM. A 1214 t-1215 indicates p 0.01. See also Figure 5-1. 1216 1217 Figure 6: Distribution of Double-Positive P2 Events Across FSC Ranges and in Different 1218 Sample Buffers 1219 1220 29 These experiments were conducted using FM-triggering on the Influx. (A) Representative 1221 density plot of non-fluorescent silica and fluorescent PS beads (Apogee) showing the five FSC 1222 gated regions based on noise threshold and bead positions. (B) Same sample and FSC gates 1223 as in A but plotted to display green fluorescence of PS beads (FITC detector). (C) 1224 Representative density plot showing the five FSC gated regions on a mixture of single calcein-1225 labeled P2 samples. Only gated double-positive events are displayed. The percentage of all 1226 fluorescent (i.e. calcein-labeled) events that falls into each FSC gate is displayed in black (all 1227 five FSC gates sum to 100%). The percentage of all calcein double-positive events that falls into 1228 each FSC gate is displayed in red (all five FSC gates sum to 100%). The number of double-1229 positive events within each FSC gate, expressed as a percentage of all fluorescent events in 1230 that FSC gate, is displayed in blue. (D-F) Box and whiskers plots of single-positive, double-1231 positive, or triple-positive events, expressed as a percentage of all fluorescent events within the 1232 respective FSC gate. Central bar represents the median. Lower and upper edges correspond to 1233 25th and 75th percentiles. Lower and upper whiskers extend to the smallest or largest value no 1234 greater than 1.5 times the interquartile range away from the corresponding edge. Data for 1235 single- and double-positive events are derived from two-color mixtures of P2 samples single-1236 labeled with calcein AM/calcein red (n=4), violet amine-reactive dye/MitoTracker Deep Red FM 1237 (n=1) or calcein AM (n=1) or calcein red (n=1), as well as three-color mixtures single-labeled 1238 with violet amine-reactive/calcein AM/calcein red (n=3). Triple-positive event data are derived 1239 only from the three-color mixtures. -test was performed comparing each FSC gate 1240 to that directly above or below it. * indicates p 0.05, ** indicates p 0.01. (G) Diagram of 1241 1242 run on the fl-1243 labeling and the first two washes in PBS prior to mixing, a third wash, and flow cytometry in SET 1244 buffer. (H) Box and whiskers plot of double-positive events, expressed as a percentage of all 1245 fluorescent events within the respective FSC gate. Data are derived from two-color mixtures of 1246 P2 samples single-1247 --way ANOVA was performed, revealing 1248 significant main effects of FSC Gate (Df = 5, F = 20.06, p=2.7e-15) and Buffer (Df = 2, F = 1249 20.14, p=1.8e-8), as well as a significant interaction (Df = 10, F = 2.78, p=0.004). See also 1250 Figure 6-1. 1251 1252 1253 Figure 7: Fractional Abundance of Immunostained P2 Events in Different FSC Ranges and in 1254 Different Sample Buffers 1255 1256 These experiments were conducted using FM-triggering on the Influx. (A) Representative 1257 density plots of P2 samples immunostained for VGAT, VGLUT1, and VMAT2 and run in PBS. 1258 Top: Secondary-only controls in which primary antibodies were omitted. Bottom: immunostained 1259 with indicated primary antibodies. (B) Box and whiskers plot of immunostained events as 1260 indicated, expressed as a percentage of all events within the respective FSC gate. Central bar 1261 represents the median. Lower and upper edges correspond to 25th and 75th percentiles. Lower 1262 and upper whiskers extend to the smallest or largest value no greater than 1.5 times the 1263 interquartile range away from the corresponding edge. Data are derived from P2 samples triple-1264 stained and washed in PBS, and subsequently run in either PBS (n=6) or SET (n=6) as sample 1265 buffer. Two-way ANOVA was performed for each dataset. Left: For VGAT+ events, FSC Gate 1266 (Df = 5, F = 21.51, p=2.9e-12) and Buffer (Df = 1, F = 4.70, p=0.034) were both significant, with 1267 30 no significant interaction (Df = 5, F = 0.40, p=0.85). Middle: For VGLUT1+ events, only FSC 1268 Gate (Df = 5, F = 68.9, p 2e-16) was significant, while Buffer (Df = 1, F = 0.36, p=0.55) and the 1269 interaction (Df = 5, F = 0.19, p=0.97) were not. Right: For VMAT2+ events, FSC Gate (Df = 5, F 1270 = 12.54, p=2.4e-8) and Buffer (Df = 1, F = 5.21, p=0.026) were both significant, with no 1271 significant interaction (Df = 5, F = 0.27, p=0.93). (C) Box and whiskers plots of single-, double-, 1272 and triple-positive events expressed as a percentage of all fluorescent events within the 1273 respective FSC gate. Data are derived from the same immunostained P2 samples as in B. Two-1274 way ANOVA was performed for each dataset. Left: For single-positive events, FSC Gate (Df = 1275 5, F = 19.83, p=1.3e-11) and Buffer (Df = 1, F = 6.74, p=0.012) were both significant, with no 1276 significant interaction (Df = 5, F = 0.15, p=0.98). Middle: For double-positive events, FSC Gate 1277 (Df = 5, F = 23.25, p=6.6e-13) and Buffer (Df = 1, F = 8.00, p=0.006) were both significant, with 1278 no significant interaction (Df = 5, F = 0.40, p=0.85). Right: For triple-positive events, only FSC 1279 Gate (Df = 5, F = 8.37, p=4.7e-6) was significant, while Buffer (Df = 1, F = 1.91, p=0.17) and the 1280 interaction (Df = 5, F = 0.19, p=0.97) were not. See also Figure 7-1, Figure 7-2, and Figure 7-3. 1281 1282 Figure 8: Effects of Repeated Centrifugation in Different Sample Buffer on Particle Abundance 1283 and Composition of P2 Samples 1284 1285 These experiments were conducted using FM-triggering on the Influx. (A) Diagram of 1286 buffer/centrifugation scheme. A P2 sample resuspended in PBS or SET was aliquoted to 1287 measure input (P2-0), followed by two centrifugations for 5 min at the indicated relative 1288 centrifugal force (xg). An aliquot of the first pellet (P2-1) is reserved prior to the second spin. All 1289 samples are then analyzed by flow cytometry using absolute counting beads to measure particle 1290 number in each fraction. (B) Input samples (P2-0) in PBS or SET buffer (n=4 each). Left: Total 1291 particle counts. Right: Median fluorescence intensity of FM4-64 for all events. (C) 1292 Representative histogram of FM4-64 fluorescence of all events for two replicate input samples 1293 (P2-0) in PBS and SET buffer. (D) FSC vs. FM4-64 density plots for P2-1 (top) or S2-1 (bottom) 1294 samples in PBS or SET buffer centrifuged at 2500xg (left) or 5000xg (right1295 corresponds to the top 2 FSC gates from Fig. 6 and 7. (E) Gated ‘1296 as % of all events, plotted for PBS (left) or SET (right) in each of five fractions moving through 1297 the centrifugation protocol at indicated relative centrifugal force (xg). Left to Right in each panel: 1298 Input (P2-0), first pellet (P2-1) and supernatant (S2-1), second pellet (P2-2) and supernatant 1299 (S2-2). (F) Gated ‘-1 and P2-2 samples 1300 in PBS or SET at indicated relative centrifugal force (xg). See also Figure 8-1. 1301 1302 Figure 1-1: Comparison of FSC Signal from Neuronal Nuclei and 0.5, 0.8, and 2.0 μm PS 1303 beads 1304 1305 We note that in these experiments, the FSC detector voltage had to be lowered in order to allow 1306 the 2.0 μm PS to fall within the detectable range; the FSC axis is therefore substantially different 1307 from all other plots in this study. (A) Mixture of fluorescent PS beads (Bangs Labs; B, or 1308 Spherotech; S) detected in FSC-trigger mode on the Influx (left) and SSC-trigger mode on the 1309 Fortessa (right). We note that the lowered FSC voltage on the Fortessa resulted in a significant 1310 amount of FSC noise (B) Crude nuclei detected in FSC-trigger mode on the Influx. Left: 1311 Hoechst33342 fluorescent staining of DNA identifies nuclei apart from all other particles. Right: 1312 Gated nuclei are then assessed for immunofluorescence of NeuN to identify neuronal nuclei. (C) 1313 Overlay of 2.0 μm PS bead gate (left) and neuronal nuclei (right) using FSC-trigger on the Influx. 1314 31 1315 1316 Figure 2-1: Dilution Series Define a Range of Linear Particle Detection in P2 Samples 1317 (Fortessa) 1318 1319 (A) Representative density plots of a P2 dilution series detected on the Fortessa in FSC-trigger 1320 mode. FM4-64 Low gate was set arbitrarily to quantify disappearance of events in this region 1321 with dilution. Although FSC-trigger threshold was set to reduce noise to below 10 1322 events/second, at the low end of the dilution curve these events begin to comprise a significant 1323 fraction of collected data. (B) Representative density plots of same P2 dilution series as in A, 1324 but detected on the Fortessa in FM-trigger mode. (C) Typical dilution series with FM4-64-1325 triggering on the Influx, conducted with absolute counting beads. FM4-64 triggered event counts 1326 (excluding absolute counting beads) were used to determine particle concentration of each 1327 sample dilution. Event rate (left) and calculated particles/μL (right) are shown across the dilution 1328 series. 1329 1330 Figure 3-1: Lack of Spectral Overlap Between Dyes and Lack of Calcein Dye Transfer to 1331 Unstained Samples 1332 1333 (A) Matrix of representative density plots demonstrating lack of spectral overlap between any of 1334 the four dyes used in multicolor aggregation assays. Columns represent P2 samples single-1335 labeled with the indicated dye, while rows indicate the measured fluorescence in the detector for 1336 the indicated dye. All samples were detected by FM-triggering on the Influx. (B) A single P2 1337 sample was split into equal aliquots, one of which was left unstained, while the other was truly 1338 doubled-labeled with calcein-AM and calcein red. The double-labeled sample was then 1339 sonicated in the equivalent volume and concentration of PBS and subsequently mixed with the 1340 unstained sample as in the standard multicolor aggregation assay. A small aliquot of the double-1341 labeled sample (right) and the sonicated double-labeled/unstained mixture (left) were detected 1342 using FM-triggering on the Influx. As shown, sonication fully disrupted all double-labeled 1343 particles, and none of the released calcein dye was acquired by the unstained sample. 1344 1345 Figure 4-1: Dilution Series Reduces False Double-Positive Bead Events but Not Double-1346 Calcein Positive P2 Events (Fortessa) 1347 1348 (A) Representative density plots of a dilution series for a mixture of fluorescent 0.5 μm PS 1349 beads detected using FSC-triggering on the Fortessa. PS beads and double-positive events are 1350 distinguished by their green (FITC detector) or red (PE detector) fluorescence. In addition to 1351 clear overlaps between green and red fluorescent beads, we also observed what appeared to 1352 be doublet, triplet, and higher order multiplets of the Nile Red 0.5 μm PS beads. We confirmed 1353 this based on the 1354 left), these 1355 multiplets could also be detected together with the green fluorescent beads. (B) Representative 1356 density plots of a dilution series for a mixture of single calcein-labeled P2 samples detected 1357 using FM-triggering on the Fortessa. Only events gated positive for calcein AM are displayed. 1358 1359 Figure 5-1: Scatter-based Singlet Gating Strategies Employed in Conventional Flow Cytometry 1360 are Not Effective for P2 Samples 1361 32 1362 A mixture of single calcein-labeled P2 samples was analyzed by FM-triggering on the Influx or 1363 the Fortessa. Only events gated for either single calcein-positive events (top) or double calcein-1364 positive events (bottom) are displayed. Left: The Influx is an analog cytometer and does not 1365 measure Height and Width of pulses for the FSC detector; we instead compared the width of the 1366 trigger pulse (FM4-1367 the Fortessa). Right: Comparison of Area and Height of FSC pulses on the Fortessa. On both 1368 the Influx and Fortessa, both single- and double- 1369 1370 Figure 6-1: Effects of Different Sample Buffers on Calcein Labeling, FSC Intensity, and 1371 Fluorescence Measurements 1372 1373 These experiments were conducted on the Influx using FSC-triggering (for beads in B) or FM-1374 triggering (for P2 samples in A and C). (A) Representative density plots for mixtures of single 1375 calcein-labeled P2 samples. Columns indicate the buffer designation for each sample, while the 1376 rows indicated the gated population displayed. (B) FSC and green fluorescence (FITC detector) 1377 of fluorescent 0.5 and 0.8 μm PS beads (Bangs Labs; B) run in PBS (left) or SET (right) buffer. 1378 The median and standard deviation of FSC and green fluorescence signal intensity are shown 1379 below for both gated bead populations. (C) Box and whiskers plots of fluorescent events (all 1380 events positive for either calcein), expressed as a percentage of all events. Central bar 1381 represents the median. Lower and upper edges correspond to 25th and 75th percentiles. Lower 1382 and upper whiskers extend to the smallest or largest value no greater than 1.5 times the 1383 interquartile range away from the corresponding edge. Data are derived from the same single 1384 calcein--1385 -test was performed comparing each buffer pair. * indicates p .05. 1386 1387 Figure 7-1: Distribution of Immunostained P2 Events Across FSC Ranges and Comparison 1388 Between Single- and Double-Positive Event Frequencies 1389 1390 All data are derived from the same samples as presented in Figure 7, i.e. P2 samples 1391 immunostained in PBS for VGAT, VGLUT1, and VMAT2 and run on the Influx with FM-triggering 1392 in PBS or SET as sample buffer. (A) Representative histogram with overlay of FSC gates. Only 1393 events gated as positive for VGAT (left), VGLUT1 (middle), or VMAT2 (right) are displayed. (B) 1394 Box and whiskers plots of immunopositive events within each FSC gate, expressed as a 1395 percentage of all immunopositive events for the indicated antigen. Analogous to the histograms 1396 above, the sum across FSC gates is 100% for any one sample. Central bar represents the 1397 median. Lower and upper edges correspond to 25th and 75th percentiles. Lower and upper 1398 whiskers extend to the smallest or largest value no greater than 1.5 times the interquartile range 1399 away from the corresponding edge. (C) Relative abundance of all events immunopositive for a 1400 given antigen, expressed as a percentage of all events. Data for are plotted as mean +/- SEM 1401 for PBS or PBS-SET (n=6 for each). (D) Left: A simple model for collisions of single-positive 1402 events leading to double-positive events is given by sampling with replacement. The relative 1403 abundance of any given double-positive event is estimated by multiplying the relative 1404 abundances of the two single-positive events. Right: Comparison of the observed and expected 1405 frequency for each type of double-positive event, expressed as a percentage of all double-1406 positive events. Observed data represent the average for PBS or PBS-SET (n=6 for each). (E) 1407 AlexaFluor488 MESF calibration bead standards (left) used to construct a calibration curve 1408 33 1409 each bead (right). (F) A P2 sample stained for VGAT (with anti-mouse AlexaFluor488 1410 secondary) was gated for immunonegative and immunopositive events as shown in Fig. 7A. 1411 Plotted are the median fluorescence intensity and robust standard deviation for each FSC gate 1412 within the immunostaining gate. 1413 1414 Figure 8-1: Particle Recovery in P2 Samples Following Repeated Centrifugations 1415 1416 All data are derived from the samples presented in Figure 8, i.e. P2 samples subjected to two 1417 centrifugations for 5 min in PBS or SET buffer at varying relative centrifugal force (xg). Absolute 1418 counting beads were added to each sample to determine absolute particle counts. (A) Total 1419 particle counts in each fraction for P2-1, S2-1, P2-2, and S2-2 samples in PBS. (B) Same as A 1420 but for SET buffer. (C) Percent recovery in the pellet following each centrifugation at indicated 1421 rcf (xg). (D) Total particle counts for pellets and supernatants across varying centrifugation rcf 1422 (xg) and in PBS or SET buffer for Spin 1. (E) Same as D but for Spin 2. 1423 1424 1 Table 1: Optical configuration of the LSRFortessa LSRFortessa Optics Laser Band Pass Filter Dichroic Filter Detector name Fluorophores used in this study 405nm 450/50 Blank Pacific Blue Violet amine-reactive dye 488nm 488/10 Blank SSC 488nm 530/30 505LP FITC Alexa488, Calcein AM 488nm 710/50 685LP PerCP-Cy5.5 FM4-64 561nm 582/12 Blank PE Alexa555, Calcein red 633nm 670/30 Blank APC Alexa647, MitoTracker Deep Red FM 1 Table 2: Optical configuration of the Influx Influx Optics Laser Band Pass Filter Detector name Fluorophores used in this study 405nm 460/50 BV421 Hoechst 33342, Violet amine-reactive dye 488nm 488/10 SSC 488nm 530/40 FITC Alexa488, Calcein AM 488nm 692/40 PerCP-Cy5.5 FM4-64 561nm 589/29 PE Alexa555, Calcein red 638nm 670/30 APC Alexa647, MitoTracker Deep Red FM 1 Table 3: List of antibodies Primary Antibody List Origin Antigen Concentration Manufacturer Product No. RRID Mouse NeuN 1:5,000 Millipore MAB377 AB_2298772 Mouse VGAT 1:400 Synaptic Systems 131 011 AB_887872 Guinea Pig VGLUT1 1:500 Synaptic Systems 135 304 AB_887878 Rabbit VMAT2 1:500 Immunostar 20042 AB_10013884 Secondary Antibody List Origin Antigen Fluorophore Concentration Manufacturer Product No. RRID Goat Mouse IgG (H+L) Alexa Fluor 488 1:500 Invitrogen A-11001 AB_2534069 Goat Guinea Pig IgG (H+L) Alexa Fluor 555 1:500 Invitrogen A-21435 AB_2535856 Goat Rabbit IgG (H+L) Alexa Fluor 647 1:500 Invitrogen A-21244 AB_2535812 1 Table 4: Summary of Key Steps in Microparticle Flow Cytometry Experimental Workflow Purpose Relevant Figures Experimental Assays Key Notes Determine detection sensitivity of cytometer Fig. 1, 1-1 FSC vs. fluorescence triggering with lipid dye (eg. FM4-64 or other) PS and Silica bead standards A number of dyes are suitable for fluorescence-triggering (Arraud et al., 2016; Gray et al., 2015; Kormelink et al., 2016; Stoner et al., 2016). Multiple studies, including ours, have demonstrated the increased sensitivity, accuracy, and reliability of fluorescence-triggering (Arraud et al., 2016; Nolan and Stoner, 2013). FSC signal from bead standards (especially PS) does NOT accurately reflect the size of biological particles detected. Refractive index differences lead to size underestimation, the extent of which is highly cytometer-dependent (Table 5). Determine linear range of particle detection Fig. 2, 2-1 Dilution series Conducting a dilution series is critical to identify the range of linear detection on each cytometer. Detection of coincidence and aggregation Fig. 3, 3-1 Multicolor dye labeling A number of dyes are suitable for multicolor dye labeling assays, as described in cellular \"barcoding” assays (Krutzik and Nolan, 2006; Krutzik et al., 2011). Distinguishing between coincidence and aggregation Fig. 4, 4-1 Fig. 5 1) Dilution series with multicolor dye labeling 2) Sorting double-positive events from beads vs. samples 3) ‘Spin’ vs. ‘Mix’ multicolor dye labeling We found that mixtures of different color PS beads are a suitable model of coincidence. Double-positive sample events that do not decrease with dilution, and those that can be enriched by sorting, strongly suggest aggregates vs. coincidence. This can be further supported by testing whether the frequency of double-positive events is increased by centrifugation. Reduce aggregation Fig. 6, 6-1 Fig. 7, 7-1 Use of nonionic buffers As reported in previous microscopy studies (Choi et al., 2009; Daniel et al., 2012), we found that nonionic sample buffers tend to reduce aggregation of synaptosomes. This is not a complete solution to eliminating false double- and triple-positive events. Identify suitable FSC ranges Fig. 6, 6-1 Fig. 7, 7-1 Immunostaining and multicolor dye labeling FSC of beads provides relative references for avoiding regions with 2 high false double-positive rates Assess false colocalization of antigens Fig. 7, 7-1 Immunostaining of mutually exclusive antigens Highly abundant antigens that should be mostly exclusive are best for these experiments. 1 Table 5: Previous Comparisons of PS Bead vs. Lipid Vesicle Scattering Measurements Study Cytometer Model Parameter PS Bead Range Vesicle Range Chandler et al., 2011 Apogee A40 FSC 0.5 - 1.0 μm 1.4 – 2.7 μm van der Pol et al., 2012 Beckman Coulter FC 500 FSC 0.5 - 0.9 μm 1.00 – 1.75 μm van der Pol et al., 2012 Apogee A40 FSC 0.5 - 0.9 μm 1.25 – 2.0 μm van der Pol et al., 2012 Becton Dickinson FACSCalibur SSC 0.5 - 0.9 μm 2.3 – 4.6 μm Simonsen, 2016 Becton Dickinson LSRII SSC 110 nm 400 nm 1 Table 6: Previous Size Range Measurements of Mammalian Synaptosomes Study Method Species/Region Synaptosome Fraction Mean Diameter (± SEM) Dunkley et al., 1986 Electron microscopy Rat Cortex P2 Percoll Fraction 3 515 ± 17 nm Dunkley et al., 1986 Electron microscopy Rat Cortex P2 Percoll Fraction 4 568 ± 14 nm Dunkley et al., 1986 Electron microscopy Rat Cortex P2 Percoll Fraction 5 539 ± 12 nm Dunkley et al., 1988 Electron microscopy Rat Cortex S1 Percoll Fraction 1 320 ± 130 nm Dunkley et al., 1988 Electron microscopy Rat Cortex S1 Percoll Fraction 2 460 ± 150 nm Dunkley et al., 1988 Electron microscopy Rat Cortex S1 Percoll Fraction 3 550 ± 130 nm Dunkley et al., 1988 Electron microscopy Rat Cortex S1 Percoll Fraction 4 640 ± 120 nm Dunkley et al., 1988 Electron microscopy Rat Cortex S1 Percoll Fraction 5 630 ± 190 nm Robinson et al., 1989 Electron microscopy Rat Striatum S1 Percoll Fraction 3 591 ± 9 nm Robinson et al., 1989 Electron microscopy Rat Striatum S1 Percoll Fraction 4 905 ± 12 nm Hollingsworth et al., 1985 Electron microscopy Guinea Pig Cortex Filtration ‘synaptoneurosomes’ 560 ± 15 nm Hollingsworth et al., 1985 Light microscopy Guinea Pig Cortex Filtration ‘synaptoneurosomes’ ~600 nm Williams et al., 2009 Phase contrast microscopy Human Cortex Filtration ‘synaptoneurosomes’ 300 – 700 nm Fernández-Busnadiego et al., 2013 Cryo-electron tomography Mouse Cortex Percoll Fraction 3+4 500 – 1000 nm 1 Statistical Table Line Figure Comparison Data Structure Statistical test Power p value F/t value a 5D Mix 0min vs. Spin 0min Normal distribution Two-sided Welch’s t-test 0.9950 0.0072 t(3.49) = -5.64 b 5D Mix 40min vs. Spin 40min Normal distribution Two-sided Welch’s t-test 0.8313 0.0046 t(5.23) = -4.74 c 5D Mix 75min vs. Spin 75min Normal distribution Two-sided Welch’s t-test 0.8251 0.0103 t(7.86) = -3.35 d 6D FSC Noise vs. 500nm PS Normal distribution Two-sided Student’s t-test 0.6783 0.0260 t(11.29) = -2.56 e 6D 500nm PS vs. 880nm Sil Normal distribution Two-sided Welch’s t-test 0.8487 0.0085 t(11.62) = 3.16 f 6D 880nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.4452 0.0738 t(14.71) = -1.92 g 6D 1300nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.7761 0.0124 t(14.83) = 2.85 h 6E FSC Noise vs. 500nm PS Normal distribution Two-sided Welch’s t-test 0.6953 0.0242 t(10.95) = 2.61 i 6E 500nm PS vs. 880nm Sil Normal distribution Two-sided Welch’s t-test 0.8366 0.0092 t(11.73) = -3.11 j 6E 880nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.4428 0.0734 t(15.73) = 1.92 k 6E 1300nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.7423 0.0142 t(15.54) = -2.76 l 6F FSC Noise vs. 500nm PS Normal distribution Two-sided Welch’s t-test 0.0911 0.5286 t(2.00) = 0.76 m 6F 500nm PS vs. 880nm Sil Normal distribution Two-sided Welch’s t-test 0.9965 0.0229 t(2.00) = -6.5 n 6F 880nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.2397 0.2468 t(2.01) = 1.62 o 6F 1300nm Sil vs. 1300nm Sil Normal distribution Two-sided Welch’s t-test 0.7653 0.0230 t(3.99) = -3.59 p 6-1C PBS vs. PBS-SET Normal distribution Two-sided Welch’s t-test 0.0666 0.6775 t(14.11) = 0.425 q 6-1C PBS-SET Normal Two-sided 0.2500 0.1779 t(12.49) = 2 vs. SET distribution Welch’s t-test 1.43 r 6-1C PBS vs. SET Normal distribution Two-sided Welch’s t-test 0.5481 0.0417 t(17.44) = 2.20 s 6H Main effect; FSC Gate Normal distribution Two-way ANOVA 0.9793 1.83E-13 F(4,125) = 21.40 t 6H Main effect; Buffer Normal distribution Two-way ANOVA 0.5278 1.15E-06 F(2,125) = 15.29 u 6H Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.2030 0.0043 F(8,125) = 2.98 v 7B, VGAT Main effect; FSC Gate Normal distribution Two-way ANOVA 0.9753 2.84E-10 F(4,50) = 21.21 w 7B, VGAT Main effect; Buffer Normal distribution Two-way ANOVA 0.0815 0.0610 F(1,50) = 3.67 x 7B, VGAT Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.0526 0.8060 F(4,50) = 0.40 y 7B, VGLUT1 Main effect; FSC Gate Normal distribution Two-way ANOVA 0.9998 2E-16 F(4,50) = 68.31 z 7B, VGLUT1 Main effect; Buffer Normal distribution Two-way ANOVA 0.0500 0.724 F(1,50) = 0.13 a1 7B, VGLUT1 Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.0504 0.956 F(4,50) = 0.16 b1 7B, VMAT2 Main effect; FSC Gate Normal distribution Two-way ANOVA 0.8516 4.67E-07 F(4,50) = 12.35 c1 7B, VMAT2 Main effect; Buffer Normal distribution Two-way ANOVA 0.0781 0.069 F(1,50) = 3.46 d1 7B, VMAT2 Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.0513 0.887 F(4,50) = 0.28 e1 7C, Single+ Main effect; FSC Gate Normal distribution Two-way ANOVA 0.9756 2.71E-10 F(4,50) = 21.27 f1 7C, Single+ Main effect; Buffer Normal distribution Two-way ANOVA 0.1074 0.029 F(1,50) = 5.06 g1 7C, Single+ Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.0504 0.96 F(4,50) = 0.16 h1 7C, Double+ Main effect; FSC Gate Normal distribution Two-way ANOVA 0.9876 1.68E-11 F(4,50) = 25.33 i1 7C, Double+ Main effect; Buffer Normal distribution Two-way ANOVA 0.1335 0.0162 F(1,50) = 6.20 j1 7C, Double+ Interaction: FSC Gate x Buffer Normal distribution Two-way ANOVA 0.0530 0.7857 F(4,50) = 0.43 k1 7C, Triple+ Main effect; FSC Gate Normal distribution Two-way ANOVA 0.6865 1.69E-05 F(4,50) = 8.88 l1 7C, Triple+ Main effect; Buffer Normal distribution Two-way ANOVA 0.0544 0.255 F(1,50) = 1.33 m1 7C, Interaction: Normal Two-way 0.0507 0.938 F(4,50) =