1 Biotransformation of bisphenol-A bis(diphenyl phosphate): In vitro, in silico, and (non-) 1 target analysis for metabolites in rat and bird liver microsomal models 2 3 Sofia M. Herczegha,b, Shaogang Chua, Robert J. Letchera,b* 4 5 a Ecotoxicology and Wildlife Health Division, Environment and Climate Change Canada, 6 National Wildlife Research Centre, Carleton University, Ottawa, ON, K1A 0H3, Canada 7 b Department of Chemistry, Carleton University, Ottawa, ON, K1S 5B6, Canada 8 9 10 11 12 13 14 15 16 17 18 * Corresponding author: Robert J. Letcher; Phone: 1-613-998-6696; Fax: 1-613-998-0458; 19 E-mail: robert.letcher@ec.gc.ca. 20 21 2 Highlights 22 23  44% bisphenol-A bis(diphenyl phosphate) (BPADP) in vitro was depleted in rat assays 24  In rat assays 2.6% and 3.9% of BPADP was converted to DPHP and BPA, respectively 25  No significant BPADP depletion or BPA and DPHP formation in herring gull assays 26  In silico estimation of metabolites BPA and DPHP confirmed by rat and gull assays 27  Other BPADP oxidation metabolites and conjugate identified by non-target screening 28 29 3 Graphic Abstract 30 31 32 33 34 4 Abstract 35 Increased production and usage of organophosphate esters (OPEs) as flame retardants and 36 plasticizers has trended towards larger and ‘novel’ (oligomeric) OPEs, although there is a dearth 37 of understanding of the environmental fate, stability, toxicokinetics, biotransformation and 38 bioaccumulation of novel OPEs in exposed biota. The present study characterized in vitro 39 biotransformation of the novel OPE bisphenol-A bis(diphenyl phosphate) (BPADP) using 40 Wistar-Han rat and herring gull liver based microsomal assays. Hypothesized target metabolites 41 bisphenol-A (BPA) and diphenyl phosphate (DPHP) and other metabolites were investigated by 42 applying a lines of evidence approach. In silico modelling predicted both BPA and DPHP as rat 43 metabolites of BPADP, these metabolites were quantified via UHPLC-QQQ-MS/MS. Additional 44 non-target metabolites were determined by UHPLC-Q-Exactive-Orbitrap-HRMS/MS and 45 identified by Compound Discoverer software. Mean BPADP depletion of 44 ± 10% was 46 quantified with 3.9% and 2.6% conversion to BPA and DPHP, respectively, in the rat assay. 47 BPADP metabolism was much slower when compared to the well-studied OPE, triphenyl 48 phosphate (TPHP). BPADP depletion in gull liver assays was far slower relative to the rat. 49 Additional non-target metabolites identified included two Phase I, O-dealkylation products, five 50 Phase I oxidation products and one Phase II glutathione adduct, demonstrating agreement 51 between lines of in vitro and in silico evidence. Lines of evidence suggest that BPADP is 52 biologically persistent in exposed mammals or birds. These findings add to the understanding of 53 BPADP stability and biotransformation, and perhaps of other novel OPEs, which are factors 54 highly applicable to hazard assessments of exposure, persistence and bioaccumulation in biota. 55 Keywords: Organophosphate esters, Biotransformation and metabolism, In silico, In vitro, bird 56 and rat models, non-target analysis 57 5 1. Introduction 58 Flame retardant (FR) chemicals, used in a variety of consumer products as both 59 plasticizers and FRs, consist of a variety of compounds categorized by chemical properties and 60 structure (van der Veen and de Boer, 2012; NIEHS, 2018; Yao et al., 2021). Brominated FRs 61 (BFRs) include polybrominated diphenyl ethers (PBDE) technical formulations (pentaBDE, 62 octaBDE and decaBDE) comprised of varying congeners, which were produced and used 63 extensively before largely being replaced by organophosphate ester FRs (OPEs) (van der Veen 64 and de Boer, 2012; Blum et al., 2019). Following strong evidence of the persistent, 65 bioaccumulative and toxic (PBT) nature of PBDEs, pentaBDE and octaBDE were listed under 66 Annex A (elimination) of the Stockholm Convention on persistent organic pollutants (SC-POPs) 67 in 2009 and decaBDE was listed in 2017 (UNEP, 2019). 68 Bisphenol-A bis(diphenyl phosphate) (BPADP; CAS RN 5945-33-5) is an OPE proposed 69 as a replacement for decaBDE (Rossi and Heine, 2007). In 2006, the production volume of 70 BPADP in the United States was estimated to be within the range of 4500-23000 tons (van der 71 Krowech et al., 2016; Veen and de Boer, 2012). Similar to the majority of OPEs, BPADP is an 72 additive FR compound, posing a higher risk of entering the environment compared to reactive 73 and chemically-bonded FRs (Velencoso et al., 2018). Based on physico-chemical properties and 74 limited available literature concerning the compound, BPADP could be considered a ‘novel’, 75 oligomeric OPE along with other OPEs such as resorcinol bis(diphenylphosphate (RDP) and 76 tetrakis(2,6-dimethylphenyl)-m-phenylene biphosphate (RDX). As industry trends move towards 77 polymerization and longer-chain, larger replacement FRs such as BPADP, more study is needed 78 concerning the fate, transformation and effects of such compounds. 79 6 Considering physico-chemical properties and environmental behaviour of BPADP, an 80 experimentally derived logKOW value for BPADP was ‘≥6’ (ECHA, 2011) and logKOC values of 81 4.00 ± 0.473 and 4.76 ± 0.252 were reported for water – suspended particulate matter (SPM) 82 and water – sediment, respectively (Zhong et al., 2021). Such coefficients suggest lipophilicity 83 and high affinity to sediment, with bioaccumulation factors (BAF) of 4.0e3 L water/g lipid (4.0e6 84 L water/kg lipid) and 1.3e2 to 2.1e2 L water/g lipid (1.3e5 to 2.1e5 L water /kg lipid) for plankton 85 and fish, respectively, reported in a study of Taihu Lake, China (Zhao et al., 2019). 86 Bekele et al. (2021) concluded that only a few studies have reported on the metabolic 87 transformation of OPEs in vitro in cells or microsomes, and even fewer studies have shown 88 metabolism of OPEs in vivo. These studies are limited to six “legacy” OPEs including TPHP. 89 Cytochrome P450 (CYP450) enzymes have been found to mediate the metabolism of several 90 OPEs (e.g. TPHP, tris(2-chloroethyl) phosphate (TCEP), tris(1-chloro-2-propyl) phosphate 91 (TCIPP), tris(1,3-dichloro-2-propyl) phosphate (TDCIPP)), and require nicotinamide adenine 92 dinucleotide phosphate (NADPH) (van den Eede et al., 2013). However, a preliminary study of 93 BPADP human liver microsomal (HLM) and S9 fraction metabolism suggested NADPH-94 independent enzymes may play an important role in the metabolism of BPADP (Alves et al., 95 2018). An example may be paraoxonase (PON) enzymes, specifically paraoxonase 1 (PON1), 96 which has been found to exhibit high efficiency in catalyzing the metabolism of the particular 97 organophosphates diazoxon and chlorpyrifos oxon (Draganov and La Du, 2003). 98 The metabolic pathways of O-dealkylation, abiotic hydrolysis, glucuronidation and 99 sulfation appear to be relevant for BPADP (Alves et al., 2018). BPADP may degrade into BPA, a 100 plastic additive, estrogen-mimicking compound (Rubin, 2011) and aquatic toxicant (Wu and 101 Seebacher, 2020). In the present study, a lines of evidence approach informed on BPADP 102 7 biotransformation and environmental fate, combining in silico modelling, in vitro inter-species 103 biotransformation using target metabolite identification and non-target analysis for possible 104 additional metabolites. 105 2. Materials and methods 106 2.1 In silico metabolite profiling 107 Prior to in vitro microsomal assays, rat liver S9 and microsomal metabolites as well as rat 108 in vivo metabolites were simulated using the publicly available OECD Toolbox v4.4.1 (OECD, 109 2021) (Paris, France). Predicted metabolites were generated in ranked order of probability of 110 occurrence, displayed via unique Simplified Molecular Input Line Entry System (SMILES) code 111 (Table S-1). A test of the OECD Toolbox was performed using the model OPE, TPHP, for which 112 target metabolites from an in vitro rat assay have been identified (Chu and Letcher, 2019) to 113 verify the utility of the model for identification of both Phase I and II metabolism reactions 114 relevant to OPEs. BPADP was further screened through EPI SuiteTM v4.1.1 (US EPA, 2012) for 115 environmental fate predictions. Model training sets were reviewed with certain models excluded 116 given the high molecular weight and low water solubility of BPADP. 117 2.2 Chemicals and reagents 118 BPADP standard (98 %) was obtained from Toronto Research Chemicals (Toronto, ON, 119 CA) and a spiking solution was prepared by dissolving BPADP in dimethyl sulfoxide (DMSO). 120 Internal standards d15-TPHP (> 98 % purity), and d10-DPHP (95 % purity), were obtained from 121 Wellington Laboratories (Guelph, ON, Canada) and Toronto Research Chemicals (Toronto, ON, 122 Canada), respectively. BPA, DPHP and TPHP (each > 99+ % purity), as well as the internal 123 standards 13C12-BPA (> 98 % purity) and trans-1,4-cyclohexanediol bis(diphenyl phosphate) 124 (TCHBDP; > 97 % purity) were obtained from Sigma-Aldrich (St. Louis, MI, USA). Gentest 125 8 Male Wistar-Han rat liver microsomes (20 mg/mL protein), and NADPH regenerating system 126 solutions A (NADP+ and Glc-6-PO₄) and B (G6PDH) were obtained from Corning Inc. 127 (Corning, NY, USA). All glassware were cleaned at 450 oC overnight prior to each assay to 128 reduce possible OPE contamination. 129 Rat and gull liver microsomes and NADPH system solutions were stored at -80 oC, with 130 aliquots thawed on ice for the assays. Deactivated microsomes were prepared by heating of 131 microsomes to 100 °C for 5 min in a water bath and then stored in freezer (-80 °C) until further 132 use. Potassium phosphate buffer (0.5 M; pH=7.4) was purchased from Alfa Aesar (Ward Hill, 133 MA, USA). L-Glutathione reduced (≥ 98 % purity) was obtained from Sigma-Aldrich in 134 powdered form. A complete list of referenced compounds is provided in Table S-2. 135 2.3 In vitro liver microsomal assay optimization for BPADP 136 BPADP metabolism kinetics were investigated to determine the concentration of BPADP 137 that achieves enzyme substrate saturation (zero-order kinetics) and maximum biotransformation 138 rate ([BPADP] >> 2(KM)). A kinetics assay was designed based on previous findings that 139 enzyme system saturation is achieved at a final dosing concentration of 2 µM (Greaves et al., 140 2016; Strobel et al., 2018), therefore, initial depletion of BPADP and formation of DPHP were 141 quantified in a 10 minute incubation with aliquots taken at t0, t2, t5 and t10. Kinetics assays were 142 conducted at a range of five BPADP incubation concentrations between 0.514 µM and 2.57 µM. 143 The day prior to each assay, a BPADP spiking solution was prepared in potassium 144 phosphate buffer (50 mM, pH 7.4) and deionized water then sonicated in an amber bottle and 145 then sonicated in a sonication bath for 24 hours to increase BPADP bioavailable within the 146 aqueous assay given its low water solubility and high molecular mass. 147 9 Triplicate biotransformation assays were conducted at BPADP incubation concentrations 148 of 1780 ppb (or 2.57 µM), with each active, blank and control test tube (each at total volumes of 149 950 μL) shaken at 80 rpm and 37 oC for 2 minutes minimum. Addition of 50 μL of microsomal 150 suspension (either Wistar-Han rat or herring gull) to the respective incubation tube (total volume 151 1 mL) initiated the reaction. All biotransformation incubations were vortexed for 20 seconds to 152 allow for homogenization (TPHP positive control vortexed for 5 seconds due to known rapid 153 metabolism), and a 100 μL aliquot was taken within 30 seconds of reaction initiation. Aliquots 154 from the biotransformation assay were also taken at the 5, 10, 20, 40, 60, 90 and 120- min time 155 points, with a 2 min time point aliquot additionally taken for TPHP positive controls. To 156 terminate each reaction, aliquots were added to the appropriate time point test tube containing 157 300 μL acetonitrile (ACN) and 100 μL of 40 ppb combined internal standard for a total volume of 158 500 μL. All samples were filtered through pre-rinsed VWR Modified Nylon 0.2 μm, 500 μL 159 centrifugal filters, centrifuged at 10 000 rpm for 5 minutes and transferred to UHPLC vials for 160 analysis. 161 A modified in vitro rat liver microsomal (RLM) assay was conducted to generate samples 162 suitable for Target and non-target analysis (NTA) by UHPLC-Q-Exactive-HRMS/MS for 163 screening of additional BPADP metabolites. Reactions were terminated by addition of 4 mL 164 diethyl ether, centrifuged at 3500 rpm for five minutes, frozen for a minimum of one hour before 165 retention of ether phase, then volatilized with 100 µL UHPLC-grade methanol added to dried 166 sample. All samples were sonicated for 10 minutes and filtered through pre-rinsed PALL Life 167 Sciences Nanosep 0.2 µm wwPTFE centrifuge filters for five minutes at 3500 rpm then 168 transferred to vials for UHPLC-Q-Exactive-HRMS/MS analysis. Following the findings of Chu 169 and Letcher (2019), 0.15 g/mL GSH solution was prepared in deionized water immediately prior 170 10 to the experiment for investigation of potential BPADP adducts with GSH. Method blanks of the 171 modified assay for NTA were prepared at an equivalent total incubation volume and therefore 172 used as non-GSH controls. 173 2.4 UHPLC-QQQ-MS/MS analysis 174 Analysis of time-dependent in vitro biotransformation assay samples was conducted 175 utilizing Waters Acquity Ultra High Performance Liquid Chromatography coupled to a Waters 176 Xevo TQ-S Triple Quadrupole Mass Spectrometer (UHPLC-QQQ-MS/MS) operated with an 177 electrospray ionization (ESI) source. The method used in the present study was based on our 178 previous methods with some modifications (Chu and Letcher, 2015, 2018; Su et al., 2015b). A 179 Kinetex EVO C18 column (50 x 2.1 mm, 1.7 μm particle size) was used to separated analytes. 180 Column and sample temperature were 45 °C and 20 °C, respectively. The mobile phases were 181 water (A) and methanol (B) and both contained 2 mM ammonium acetate, and the injection 182 volume was 10 μL. The initial flow rate was set to 0.5 mL/min, mobile phase B starting at 5 % 183 and the gradient increasing for five minutes to 95 % mobile phase B, held for five minutes and 184 concluding at a total 15 minute run time. The MS/MS was operated in both ESI+ and ESI- 185 modes. During 1 to 4.2 min MS/MS operated in negative mode, after that it switched to positive 186 mode. The nebulizing gas was nitrogen and argon was used as the collision gas. The capillary 187 voltage was 2.5 kV, and source and desolvation temperatures were 150 ℃ and 500 ℃, 188 respectively. The desolvation gas flow rate was 1000 L/hr and the cone gas flow rate 150 L/hr. 189 MS/MS analysis was performed using the multiple reaction monitoring (MRM) mode and six 190 functions are set in the MS method. The MRM transitions, compound dependent operation 191 parameters, and retention times are listed in Table S-3. 192 11 After UHPLC-Q-Exactive HRMS/MS analysis of the assay fractions (see next section, 2.5) 193 and using an alternate UHPLC-QQQ-MS/MS analytical method, additional metabolite targeting 194 and re-analysis occurred for in vitro triplicate samples, method blanks, enzymatically deactivated 195 and NADPH deficient controls samples. In this alternate UHPLC-QQQ-MS/MS analytical 196 method the operation parameters were the same as described for the initial method except that 197 the MS/MS was operated in ESI(+) mode for monitoring transitions of the four additional 198 targeted metabolites, i.e. BPA-DPP, BPA-DPP-MPP, BPADP+O and BPADP+2O (Table S-3). 199 UHPLC-QQQ-MS/MS sample analysis and MassLynx v4.1 software (Waters, 2014) 200 quantified the initial target BPADP depletion and BPA and DPHP metabolite formation with 201 back calculation to represent assay incubation concentrations. An 8-point calibration curve was 202 quantified between every 16 samples to ensure linearity of response, with the calibration 203 standards prepared to have metabolite concentrations 1/5 of the BPADP concentration and a 204 constant combined internal standard concentration to account for low metabolite formation. For 205 the additional four target metabolites (Table S-4), quantification was not possible due to the lack 206 of suitable internal standards. Thus, the relative concentration change of these analytes was 207 determined by UHPLC-QQQ-MS/MS peak response over time. 208 2.5 UHPLC-Q-Exactive-HRMS/MS analysis 209 Target analysis and non-target analysis (NTA) for identification of additional metabolites 210 and GSH adducts was via a Vanquish UHPLC coupled with a quadrupole orbitrap mass 211 spectrometer (UHPLC-Q-Exactive-HRMS/MS; ThermoScientific, Waltham, MA, USA). Details 212 of all these analyses are fully described in the Supplemental Information. 213 Target and non-target metabolite screening and identification following UHPLC-Q-214 Exactive-HRMS/MS in vitro sample fraction analysis was conducted using the automatic search 215 12 capabilities of Compound Discoverer version 3.2 (ThermoFisher Scientific, Waltham, MA, 216 USA) and included automatic blank subtraction with the following transformations searched: 217 Phase I, dearylation, dehydration, demethylation, desaturation, hydration, oxidation, reduction, 218 ionization and loss of a diphenyl phosphate group. For GSH adduction search the transformation 219 search also included Phase II GSH conjugation 1(->C10H15N3O6S) and GSH conjugation 2 (-220 >C10H17N3O6S). Results were filtered to minimize interference and background levels, 221 specifically filtering for retention time (RT) greater than three minutes, mass shift less than 5 222 ppm and t120/t0 response ratio > 2 (Figure S-4). 223 Verified in FreeStyle version 1.6 (ThermoFisher Scientific, Waltham, MA, US.A), mass 224 spectra and chromatographs of each potential metabolite were manually evaluated and confirmed 225 if the m/z and RT matched at least two of three adduct ions present in the buffer ( [M + H] +, [M 226 + NH4] + and [M + Na] + ). The isotope mass distribution was also used to confirm identification. 227 Peaks were averaged, background subtracted and the Gaussian 3 smoothing operator applied. 228 2.6 Quality Assurance and Control 229 The design of assay sets included one Wistar-Han rat biotransformation assay that 230 consisted of three active replicates (containing NADPH), a method blank, TPHP positive control 231 and two negative controls (NADPH-deficient and enzyme deactivated). Due to a limited amount 232 of herring gull liver microsomes, two assays were conducted resulting in a total n=5 active 233 replicates, n=4 NADPH-deficient controls, n=2 method blanks and n=2 enzyme-deactivated 234 controls. Each active replicate contained 890 μL 2ppm (2.89 µM) BPADP standard buffer solution, 235 50 μL active Wistar-Han rat or herring gull microsomes, 50 μL NADPH A and 10 μL NADPH B. An 236 identical procedure was applied to method blanks, which contained 790 μL deionized water and 100 237 μL pH 7.4 buffer rather than BPADP. NADPH-deficient controls contained 60 μL deionized water 238 13 rather than NADPH A and B while 50 μL microsomes denatured at 100 oC for 5 minutes replaced 239 active microsomes in enzyme deactivated controls. TPHP positive controls consisted of 2 μL 60 ppm 240 (183.88 µM) TPHP, 100 μL potassium phosphate pH 7.4 buffer and 788 μL deionized water rather 241 than BPADP standard buffer solution. All test tubes were covered with aluminum foil throughout the 242 assay to avoid evaporation and BPA contamination from plastic caps. For the initial target BPADP, 243 BPA and DPHP metabolites, and ISs, Table S-4 details all UHPLC-QQQ-MS/MS instrument limits 244 of detection and quantification (ILOD/ILOQ). 245 For target and non-target analyses, sample sets for UHPLC-Q-Exactive-HRMS/MS 246 analysis consisted of a method blank, four replicates at the zero minute time point and four 247 replicates at the 120- min time point. The blank sample incubation contained 1748 µL of 0.1 M 248 Biotage potassium phosphate (pH 7.4) buffer while all replicates (t=0 and t=120) contained 1748 249 µL of 2 ppm sonicated BPADP solution. All samples contained 32 µL deionized water, 100 µL 250 Wistar-Han rat liver microsomes, 100 µL NADPH A and 20 µL NADPH B. 251 2.7 Data analysis 252 The difference in BPADP concentration between t0 and t2 during the in vitro RLM 253 kinetics assays was divided by two minutes to represent the initial rate of reaction (V0). This was 254 calculated for each dosing concentration, and further for additional time points (t5, t10). A Hanes-255 Woolf plot was used to investigate maximal rate of biotransformation (Vmax) and the Michaelis 256 rate constant (Km), with further use of both Lineweaver-Burk and Eadie-Hofstee plots for 257 confirmation. 258 For UHPLC-QQQ-MS/MS determined and quantified data of target BPADP, BPA and 259 DPHP, all statistical analyses for the were conducted using GraphPad Prism v9.3.1 (GraphPad 260 Software, San Diego, CA, USA). D’Agostino-Pearson normality tests assessed data distribution 261 14 compared to Gaussian, where sample size was too small for the D’Agostino-Pearson and 262 Anderson-Darling tests the Kolmogorov-Smirnov test of normality was used. Non-linear one-263 phase decay/association least squares fit functions were then used to plot time-dependent 264 BPADP depletion and formation of metabolites. The Brown-Forsythe test was applied to 265 determine the equality of means, with significance set to a p-value of < 0.05. Depletion of 266 BPADP and target metabolite formation at each time point was tested for level of significance 267 when compared to t0 using a one-way analysis of variance (ANOVA) and post-hoc Dunnett’s 268 multiple comparisons test. 269 3. Results and discussion 270 3.1 BPA and DPHP target in vitro metabolites of BPADP 271 Compared to TPHP, BPADP depletion was very slow within the 120 min of the 272 biotransformation in vitro in the RLM and HGLM assays, likely due to the higher molecular 273 mass and structural complexity. DPHP formation in all kinetics assays was insufficient for the 274 calculation of Vmax or Km values using Hanes-Wolf, Lineweaver-Burk and Eadie-Hofstee plot 275 methods, therefore dosing concentrations of 1780 ng/mL (ppb; or 2.57 µM) were used for all 276 further RLM and herring gull liver microsomal biotransformation assays. In the positive RLM 277 microsomal assays that contained NAPDH, TPHP was 94 ± 8.19% (n=3) depleted by the 20-278 minute time point (Figure 1) at a 1.53 µM t0 concentration. Optimized RLM assays at 2.57 µM, 279 containing NADPH, demonstrated a mean 44 ± 10 % (n=9) BPADP depletion and mean 20 ± 7.0 280 % (n=3) depletion in the absence of NADPH (Figure 2a). Both target metabolites BPA and 281 DPHP were quantified as in vitro metabolites of BPADP in RLM assays, though the former only 282 in NADPH-deficient control samples (Figure 2a). The percent conversion 283 ( [𝑡𝑜𝑡𝑎𝑙 𝑚𝑒𝑡𝑎𝑏𝑜𝑙𝑖𝑡𝑒 𝑓𝑜𝑟𝑚𝑒𝑑] [𝑡𝑜𝑡𝑎𝑙 𝐵𝑃𝐴𝐷𝑃 𝑑𝑒𝑝𝑙𝑒𝑡𝑒𝑑] ∗ 100 ) of BPADP to DPHP was 2.6 % in RLM assays containing 284 15 NADPH, with a 3.9 % conversion rate of BPADP to BPA in NADPH-deficient assays. Depletion 285 of BPADP in active replicates was significant (p < 0.05) at t90 and t120 relative to t0, with 286 significant formation of both target metabolites found to be statistically significant (p < 0.05) 287 relative to t0 (BPA: t90, t120; DPHP: t20, t40. t60, t90, t120). TPHP concentrations were quantified in 288 all active microsomal, enzyme deactivated, and NADPH-deficient samples, remaining low 289 throughout all assays and attributable to impurities in the BPADP standard. A 46 ± 12 % 290 conversion of TPHP to DPHP was quantified in positive controls. 291 In vitro metabolism of BPADP to DPHP appears to be linked to mediation by both CYP450 292 enzymes and NADPH-independent enzymes, potentially paraoxonase (PON) enzymes and/or 293 aryl esterases. In the active microsomal samples, quantification of a time-dependent increase in 294 BPA formation in vitro may have occurred but the presence of CYP450 enzymes could have 295 feasibly resulted in the Phase II conjugation of the hydroxyl groups of BPA, and thus depletion 296 of BPA. Other studies have demonstrated this previously in in vitro assays using rat liver S9 297 fractions (Yoshihara et al., 2004; Ousji et al., 2020). Preliminary evidence suggests paraoxonase 298 (PON) enzymes may play a major role in BPADP in vitro HLM metabolism, where six 299 hydroxylated and O-dealkylated NADPH-independent metabolites of BPADP were identified 300 (Alves et al., 2018). The findings of Alves et al. (2018) as well as the present study demonstrate 301 NADPH-deficient controls are crucial to an experimental design when investigating BPADP 302 metabolism. This contrasts with the current understanding that CYP-mediated metabolism is an 303 important pathway for many legacy OPEs (Bekele et al., 2021; van den Eede et al., 2013), 304 suggesting the involvement of different enzymes in the mediation of novel OPE metabolism. 305 Herring gull liver microsomal assays showed no significant BPADP depletion or target BPA 306 and DPHP metabolite formation, with similar depletion in both active replicates and enzyme 307 16 deactivated controls (Figure 2b), suggesting involvement of both enzyme-mediated and abiotic 308 hydrolysis. While BPA concentrations in all samples were indistinguishable from the 309 background, DPHP concentrations were quantifiable both in active and NADPH-deficient 310 samples. Higher background levels of TPHP were quantified in herring gull liver microsomal 311 assays (Figure S-1), however the mean 19 % (n=2) depletion of TPHP cannot fully account for 312 the formation of DPHP quantified. While rapid in vitro depletion of several legacy OPEs has 313 been reported in herring gulls (Greaves et al., 2016), the present results illustrate BPADP 314 depletion to be considerably slower than that of lower molecular mass and structural complexity. 315 3.2 Additional target and non-target metabolites in the in vitro assays 316 UHPLC-Q-Exactive-HRMS/MS analysis identified eight additional in vitro metabolites 317 of BPADP, which included bisphenol A-(diphenyl phosphate) (BPA-DPP), bisphenol A-318 (diphenyl phosphate)-(monophenyl phosphate) (PBA-DPP-MPP), three BPADP+O metabolite 319 isomers (X1, X2, X3), and two BPADP +2O metabolite isomers (X4, X5) (Table 1, Figure 3). 320 These results were confirmed by a Compound Discoverer (CD) search via retention times, mass 321 shifts, t120/t0 response ratios and full scan mass spectra (Table 1, Figure S-2). In full scan mass 322 mode (HRMS), identification of these metabolites were depended on their m/z and isotope 323 distribution. The results showed that the measured m/z and isotope distribution matched very 324 well with the theoretical value (Figure S-2). Metabolites of BPA and DPHP were not confirmed 325 by UHPLC-Q-Exactive-HRMS/MS) with ESI(-) analysis, likely due to low conversion rates and 326 lower sensitivity of the UHPLC-Q-Exactive-HRMS/MS analysis compared to UHPLC-QQQ-327 MS/MS analysis. 328 Four of the addition metabolites identified by UHPLC-Q-Exactive-HRMS/MS analysis 329 could be measured by UHPLC-QQQ-MS/MS with negligible response interferences, and where 330 17 each displayed clearly increasing peak area responses during the 0 to 20 min incubation time 331 (Figure 3). However, quantitation was not possible due to lack of available chemical standards. 332 The results described were by comparison of the m/z response level (representing the mass 333 chromatographic peak area) at all time points relative to control samples. In both rat and gull 334 assays, each oxidation metabolite demonstrated an increase in peak response over time which 335 required NADPH, suggesting CYP450-mediated metabolism. BPA-DPP-MPP displayed a time-336 dependent increase in peak area in the presence and absence of NADPH and in both model 337 species, while BPA–DPP was identified in rat only in the NADPH-deficient control but both in 338 the presence and absence of NADPH in gulls. Alves et al. (2018) identified BPA–DPP via both 339 enzymatic and non-enzymatic hydrolysis pathways and while abiotic hydrolysis of BPADP was 340 not investigated in the present study, the findings also suggested BPA–DPP may be formed by 341 both enzymatic (likely PON-mediated) and abiotic/chemical hydrolysis. While formation rates 342 could not presently be calculated for these compounds, the identification of additional target 343 BPADP metabolites increases the understanding of the array of degradation compounds that 344 make up the total BPADP biotransformation mass balance. Corresponding to the much slower 345 depletion rate of BPADP in gulls, the peak areas of each of the additional four target metabolites 346 were lower for gull compared to rat assays. 347 A glutathione (GSH) adduct of BPADP was confirmed via mass spectra (Figure S-3) 348 from UHPLC-Q-Exactive-HRMS/MS analysis and Compound Discoverer in the fraction from 349 the in vitro assay with RLM (Table S-5). The posited structure in Table 1 is accurate as to the 350 exact mass and molecular formula identified via UHPLC-Q-Exactive-HRMS/MS, however, the 351 specific adduct location is unconfirmed and the possible structure presented here is based on 352 recent study in our lab of TPHP-GSH adducts from an o-hydroquinone intermediate and reaction 353 18 with o-benzoquinone (Chu and Letcher, 2019). Glucuronidation and sulfation of the BPADP 354 metabolites were previously identified in HLM S9 fraction incubated with BPADP, chemical 355 formulas C33H33O11P and C27H25O8PS, respectively (Alves et al., 2018). The present results are 356 therefore the first report on the identification of a specific Phase II metabolite from BPADP GSH 357 conjugation namely C49H49N3O16P2S (BPADP + GSH + 2O - 2H). 358 3.3 In silico metabolite predictions and physico-chemical property estimations 359 Rat liver S9 and microsomal in silico simulated metabolites of BPADP (Table 1) included 360 the target metabolite DPHP, while BPA was simulated only as a rat in vivo metabolite. EPI 361 SuiteTM predicted physico-chemical properties varied widely from limited available experimental 362 values (Table 2). Predicted logKOW and logKOC of BPADP were 10.02 and 6.24, respectively, 363 both estimates being very high relative to well-studied compounds and suggesting 364 overestimation of physicochemical properties due to the high molecular mass and low water 365 solubility of BPADP. Greater than 4.5, the predicted logKOC value suggests very high adsorption 366 to and low mobility in sediment (US EPA, 2012) and possible deleterious effects to terrestrial 367 ecosystems. 368 The combination of predicted high lipophilicity (or superhydrophobicity) and soil adsorption 369 may indicate persistence and bioaccumulation of BPADP in sediment or terrestrial environments, 370 however this may also limit contaminant availability in aquatic environments (Gobas et al., 371 2003). In both phytoplankton and zooplankton the OPEs tricresyl phosphate (TMPP), 372 2‑ethylhexyl diphenyl phosphate (EHDPP), triphenyl phosphate (TPHP), tris(2-ethylhexyl) 373 phosphate (TEHP) and tris(1,3-dichloro-2-propyl) phosphate (TDCIPP) were quantified and 374 bioavailability between sediment and biota increased with hydrophobicity (Wang et al., 2019). 375 Bioavailability was also found to plateau at logKOW values >5.73 and was affected by 376 19 biotransformation rates of individual organisms at various trophic levels. High sediment 377 concentrations of OPEs can contribute to aquatic species ingesting OPEs on a continuous basis, 378 potentially leading to magnification up a trophic system depending on the properties of 379 individual compounds (Yao et al., 2021). Additional experimentally derived physicochemical 380 properties are needed, particularly a specific logKOW value, in order to better understand the 381 bioaccumulation and biomagnification potential of BPADP. Persistence and accumulation 382 warrants further investigation given that all predicted or measured logKOW values of BPADP 383 exceed the threshold of 5, denoting potential of bioaccumulation, and the estimated persistence 384 in sediment exceeds 365 days/12 months as per CEPA 1999 Persistence and Bioaccumulation 385 Regulations (Environment Canada, 2004). BPADP lipophilicity may be similar to or exceed that 386 of previously studied contaminants including polychlorinated biphenyl (PCB) and PBDE 387 congeners (Table 2), which have demonstrated high bioconcentration and bioaccumulation 388 potential. The environmental fate of BPADP possibly being similar would be highly relevant if 389 production and usage increases in the future. 390 One metabolite (Supplier Name ‘SCHEMBL12670502’) was corroborated by all lines of 391 evidence, i.e. predicted in silico and confirmed experimentally as an in vitro metabolite via 392 UHPLC-MS/MS and Compound Discoverer, and is the first experimental identification of this 393 BPADP metabolite. In silico profiling via the OECD ToolboxTM also predicted the metabolites 394 BPA-DPP-MPP, BPA and DPHP which were each confirmed in vitro. Despite being simulated 395 only in vivo, BPA was confirmed in vitro in the present study. Simulated metabolites did not 396 include any Phase I oxidation metabolites of Phase II conjugates, therefore predictions should be 397 interpreted conservatively- particularly for novel OPEs such as BPADP for which literature data 398 is scarce. Given the advantage of time saved, in silico modelling applications such as the OECD 399 20 ToolboxTM could be useful in identifying novel OPEs for which metabolites may be of concern, 400 and in identifying compounds of research priority. The present study being an in vitro 401 investigation of BPADP biotransformation, additional in vivo metabolites are feasible as are in 402 vitro S9 fraction and non-GSH Phase II in vitro conjugates. 403 3.4 Implications for the environmental fate of BPADP and risk assessment relevance 404 The slow metabolism, biotransformation and depletion of BPADP in the present study 405 suggests the potential for environmental persistence and bioaccumulation, coupled with the low 406 but quantifiable conversion to BPA, a well-established xenoestrogen (Badamasi et al., 2020) and 407 aquatic toxicant (Wu and Seebacher, 2020). In the absence of specific experimentally derived 408 physicochemical properties, bioaccumulation and biomagnification of BPADP and other novel 409 OPEs remains poorly understood. Limited available literature suggests BPADP partitions to soil 410 and potentially bioaccumulates (Zhao et al., 2019) illustrating the relevance of BPADP 411 bioavailability to hazard and risk assessment. Low conversion rates of BPADP to target 412 metabolites may result in chronic, low-dose exposure of BPA and DPHP to biota. 413 3. Conclusions 414 BPADP metabolism is presently shown to be slow in both the Wistar-Han rat 415 (mammalian) and herring gull (avian) models. Both metabolites of BPA and DPHP were 416 confirmed in the rat model, with eight additional metabolites identified and confirmed as in vitro 417 metabolites of BPADP (BPA-DPP, BPA-DPP-MPP, BPADP+O, BPADP+2O, 418 BPADP+GSH+2O-2H). In silico modelling proved to be a valuable prediction tool for BPADP 419 metabolite formation although these predictions should be interpreted with caution, and a lines of 420 evidence approach is preferable given the lack of available literature data for novel OPEs. 421 21 Scientific understanding as to the persistence, fate and biotransformation products of 422 novel OPEs is limited amidst projected increases in consumption, highlighting a need for 423 efficient study of these compounds. The present study suggests BPADP may be highly stable and 424 potentially bioaccumulative in environmental media and biota, however species-specific 425 differences in metabolism could lead to variation in metabolite formation across exposed species. 426 The structural complexity of novel OPEs allows for varied and complex degradation products, 427 which may pose additional persistence and/or toxicity concerns and may vary between species. 428 Addressing these knowledge gaps can elucidate whether exposure of novel OPEs to biota may 429 add to the burden of PBT contaminants in a similar fashion as the PBDEs that OPEs were 430 developed to replace. 431 432 Associated Content 433 Further information on UHPLC-MS/MS parameters (all multiple reaction monitoring transitions, 434 instrument limits of detection/quantification), molecular mass spectra and background triphenyl 435 phosphate (TPHP) can be found online. 436 437 Author Information 438 Corresponding Author 439 *Phone: +1 613-998-6696; fax: +1 613-998-0458; e-mail: Robert.Letcher@ec.gc.ca 440 441 Notes 442 The authors declare no competing financial interest. 443 444 22 Acknowledgements 445 Funding for this project was provided by the National Science and Engineering Research 446 Council (NSERC) CREATE-REACT Program (to S.M.H), by BizNGO of Clean Production 447 Action through Carleton University (to R.J.L), and the Environment & Climate Change Canada 448 Chemical Management Plan (CMP) (to R.J.L.). 449 450 References 451 Alves, A., Erratico, C., Lucattini, L., Cuykx, M., Ballesteros-Gómez, A., Leonards, P.E.G., 452 Voorspoels, S., Covaci, A., 2018. Mass spectrometric identification of in vitro-generated 453 metabolites of two emerging organophosphate flame retardants: V6 and BDP. Chemosphere 454 212, 1047–1057. doi: 10.1016/j.chemosphere.2018.08.142 455 Badamasi, I., Odong, R., Masembe, C., 2020. Threats posed by xenoestrogenic chemicals to the 456 aquatic ecosystem, fish reproduction and humans: a review. African Journal of Aquatic 457 Science, 45:3, 243-258, doi: 10.2989/16085914.2020.1746233 458 Bekele, T.G., Zhao, H., Yang, J., Chegan, R.G., Chen, J., Mekonen, S., Qadeer, A. 2021. A 459 review of environmental occurrence, analysis, bioaccumulation, and toxicity of 460 organophosphate esters. Environmental Science and Pollution Research (2021) 28, 49507–461 49528, doi: 10.1007/s11356-021-15861-8 462 Blum, A., Behl, M., Birnbaum, L.S., Diamond, M.L., Phillips, A., Singla, V., Sipes, N.S., 463 Stapleton, H.M., Venier, M., 2019. Organophosphate Ester Flame Retardants: Are They a 464 Regrettable Substitution for Polybrominated Diphenyl Ethers? Environmental Science & 465 Technology Letters 6(11), 638–649. doi: 10.1021/acs.estlett.9b00582 466 23 Center for Disease Control Agency for Toxic Substances and Disease Registry. Unknown date. 467 PCBs – Chemical and Physical Information. Washington, DC, USA. Available at: 468 https://www.atsdr.cdc.gov/toxprofiles/tp17-c4.pdf 469 Chu, S., Letcher, R.J., 2015. Determination of organophosphate flame retardants and plasticizers 470 in lipid-rich matrices using dispersive solid-phase extraction as a sample cleanup step and 471 ultra-high performance liquid chromatography with atmospheric pressure chemical 472 ionization mass spectrometry. Analytica Chimica Acta 885, 183–190. doi: 473 10.1016/j.aca.2015.05.024 474 Chu, S., Letcher, R.J., 2018. A mixed-mode chromatographic separation method for the analysis 475 of dialkyl phosphates. Journal of Chromatography A 1535, 63–71. doi: 476 10.1016/j.chroma.2017.12.069 477 Chu, S., Letcher, R.J., 2019. In vitro metabolic activation of triphenyl phosphate leading to the 478 formation of glutathione conjugates by rat liver microsomes. Chemosphere 237, 124474. 479 doi: 10.1016/j.chemosphere.2019.124474 480 Draganov, D.I., La Du, B.N., 2004. Pharmacogenetics of paraoxonases: A brief review. Naunyn-481 Schmiedeberg’s Archives of Pharmacology 369(1), 78–88. doi: 10.1007/s00210-003-0833-1 482 Environment Canada, 2004. Canadian Environmental Protection Act, 1999 Environmental 483 Screening Assessment Report on Polybrominated Diphenyl Ethers (PBDEs) – Draft for 484 Public Comments. Available at: https://www.ec.gc.ca/lcpe-485 cepa/documents/substances/pbde/pbde_draft-eng.pdf 486 Environment Canada, 2013. ARCHIVED- Environmental Screening Assessment Report on 487 Polybrominated Diphenyl Ethers (PBDEs) – Draft for Public Comments. Available at: 488 https://www.ec.gc.ca/lcpe-cepa/default.asp?lang=En&n=DF7DE982-1&offset=3&toc=hide 489 24 European Chemicals Agency (ECHA), 2011. (1-methylethylidene)di-4,1-phenylenetetraphenyl 490 diphosphate- Registration Dossier; Physical & Chemical Properties. Helsinki, Finland. 491 Available at: https://echa.europa.eu/registration-dossier/-/registered-492 dossier/6592/4/8/?documentUUID=dae1147a-fc9c-4818-905c-78b84a7265a1 493 European Commission Joint Research Centre, 2001. Diphenyl Ether, Pentabromo Derivative 494 (Pentabromodiphenyl Ether). Belgium. Available at: 495 https://echa.europa.eu/documents/10162/781ee1e9-6c90-467e-998b-8910ca2793e5 496 Gobas, F. A. P. C., Kelly, B. C., Arnot, J. A. Quantitative Structure Activity Relationships for 497 Predicting the Bioaccumulation of POPs in Terrestrial Food-Webs. QSAR & Combinatorial 498 Science 22, no. 3: 329–36. doi: 10.1002/qsar.200390022. 499 Greaves, A.K., Su, G., Letcher, R.J., 2016. Environmentally relevant organophosphate triesters 500 in herring gulls: In vitro biotransformation and kinetics and diester metabolite formation 501 using a hepatic microsomal assay. Toxicology and Applied Pharmacology 308, 59–65. doi: 502 10.1016/j.taap.2016.08.007 503 Gustafsson, K., Björk, M., Burreau, S., Gilek, M., 1999. Bioaccumulation kinetics of brominated 504 flame retardants (polybrominated diphenyl ethers) in blue mussels (Mytilus edulis). 505 Environmental Toxicology and Chemistry 18(6), 1218–1224. doi: 10.1002/etc.5620180621 506 Krowech, G., Hoover, S., Plummer, L., Sandy, M., Zeise, L., Solomon, G., 2016. Identifying 507 chemical groups for biomonitoring. Environmental Health Perspectives 124, A219-A226. 508 doi: 10.1289/EHP537 509 National Institute of Environmental Health Sciences, 2018. Flame Retardants. North Carolina, 510 USA. Available at: 511 https://www.niehs.nih.gov/health/topics/agents/flame_retardants/index.cfm. 512 25 Organisation for Economic Co-operation and Development, 2021. The OECD QSAR Toolbox. 513 Paris, France. Available at: https://www.oecd.org/chemicalsafety/risk-assessment/oecd-qsar-514 toolbox.htm 515 Ousji, O., Ohlund, L., Sleno, L., 2020. Comprehensive In Vitro Metabolism Study of Bisphenol 516 A Using Liquid Chromatography-High Resolution Tandem Mass Spectrometry. Chemical 517 Research in Toxicology 33, 6, 1468-1477. doi: 10.1021/acs.chemrestox.0c00042 518 Rossi, M., Heine, L., 2007. The Green Screen for Safer Chemicals: Evaluating Flame Retardant 519 for TV Enclosures. Somerville, MA, USA. Available at: 520 https://www.bizngo.org/static/ee_images/uploads/resources/EvaluatingFlameRetardants_Gre521 enScreenSaferChemicals_2007.pdf. 522 Rubin, B.S., 2011. Bisphenol A: An endocrine disruptor with widespread exposure and multiple 523 effects. The Journal of Steroid Biochemistry and Molecular Biology 127:27–34. doi: 524 10.1016/j.jsbmb.2011.05.002 525 Schlechtriem, C., Böhm, L., Bebon, R., Bruckert, H.-J., Düring, R.-A., 2017. Fish 526 bioconcentration studies with column-generated analyte concentrations of highly 527 hydrophobic organic chemicals. Environmental Toxicology and Chemistry 36(4), 906–916. 528 doi: 10.1002/etc.3635 529 Strobel, A., Willmore, W.G., Sonne, C., Dietz, R., Letcher, R.J., 2018. Organophosphate esters 530 in East Greenland polar bears and ringed seals: Adipose tissue concentrations and in vitro 531 depletion and metabolite formation. Chemosphere 196: 240–250. doi: 532 10.1016/j.chemosphere.2017.12.181 533 Su, G., Letcher, R.J., Yu, H., 2015. Determination of organophosphate diesters in urine samples 534 by a high-sensitivity method based on ultra high pressure liquid chromatography-triple 535 26 quadrupole-mass spectrometry. Journal of Chromatography A 1426, 154–160. doi: 536 10.1016/j.chroma.2015.11.018 537 United Nations Environment Programme, 2019. Stockholm Convention Polybromodiphenyl 538 ethers—Overview. Chatelaine, Switzerland. Available at: 539 http://chm.pops.int/Implementation/IndustrialPOPs/BDEs/Overview/tabid/5371/Default.asp540 x 541 United States Environmental Protection Agency, 2012. Estimation Programs Interface Suite™ 542 for Microsoft® Windows, v 4.11. Washington, DC, USA 543 van der Veen, I., de Boer, J., 2012. Phosphorus flame retardants: Properties, production, 544 environmental occurrence, toxicity and analysis. Chemosphere 88(10), 1119–1153. doi : 545 10.1016/j.chemosphere.2012.03.067 546 Van den Eede, N., Maho, W., Erratico, C., Neels, H., Covaci, A., 2013. First insights in the 547 metabolism of phosphate flame retardants and plasticizers using human liver fractions. 548 Toxicology Letters 223:9–15. doi: 10.1016/j.toxlet.2013.08.012 549 Velencoso, M.M., Battig, A., Markwart, J.C., Schartel, B., Wurm, F.R., 2018. Molecular 550 Firefighting—How Modern Phosphorus Chemistry Can Help Solve the Challenge of Flame 551 Retardancy. Angewandte Chemie International Edition 57(33), 10450–10467. doi: 552 10.1002/anie.201711735 553 Wang, X., Zhong, W., Xiao, B., Liu, Q., Yang, L., Covaci, A., Zhu, L., 2019. Bioavailability and 554 biomagnification of organophosphate esters in the food web of Taihu Lake, China: Impacts 555 of chemical properties and metabolism. Environment International 125, 25–32. doi: 556 10.1016/j.envint.2019.01.018 557 27 Wu, N., Seebacher, F., 2020. Effect of the plastic pollutant bisphenol-A on the biology of aquatic 558 organisms: A meta-analysis. Global Change Biology 26: 3821–3833. doi: 559 10.1111/gcb.15127 560 Yao, C., Yang, H., Li, Y., 2021. A review on organophosphate flame retardants in the 561 environment: Occurrence, accumulation, metabolism and toxicity. Science of the Total 562 Environment 795, 148837. doi: 10.1016/j.scitotenv.2021.148837 563 Yoshihara, S., Mizutare, T., Makishima, M., Suzuki, N., Fujimoto, N., Igarashi, K., Ohta, S., 564 2004. Potent Estrogenic Metabolites of Bisphenol A and Bisphenol B Formed by Rat Liver 565 S9 Fraction: Their Structures and Estrogenic Potency. Toxicological Sciences 78(1), 50–59. 566 doi: 10.1093/toxsci/kfh047 567 Zhao, H., Liu, L., Li, Y., Zhao, F., Zhang, S., Mu, D., Liu, J., An, L., Wan, Y., Hu, J., 2019. 568 Occurrence, Bioaccumulation, and Trophic Transfer of Oligomeric Organophosphorus 569 Flame Retardants in an Aquatic Environment. Environmental Science & Technology Letters 570 6(6), 323–328. doi: 10.1021/acs.estlett.9b00262 571 Zhong, W., Cui, Y., Li, R., Yang, R., Li, Y., Zhu, L., 2021. Distribution and sources of ordinary 572 monomeric and emerging oligomeric organophosphorus flame retardants in Haihe Basin, 573 China. Science of the Total Environment 785, 147274. doi: 10.1016/j.scitotenv.2021.147274 574 575 28 Tables 576 Table 1. Comparison of the multiple lines of evidence of the metabolites of BPADP from in 577 vitro biotransformation (targeted and NTA) and predicted in silico via OECD Toolbox v4.4.1 (in 578 vivo and in vivo rat simulation). 579 Metabolite Chemical Formula Abbreviation In silico In vitro target analysis In vitro NTA Posited Chemical Structure C27H25O5P SCHEMBL811 002; BPA- DPP   n/a C33H30O8P2 SCHEMBL126 70502; BPA- DPP-MPP    C12H11O4P diphenyl phosphate (DPHP)   x C6H6O phenol  n/a x C15H16O2 bisphenol A* (BPA)   x C39H34O9P2 BPADP + O (X1, X2, X3 isomers) x   29 C39H34O10P2 BPADP + 2O (X4, X5 isomers) x   C49H49N3O16P2S BPADP+GSH+ 2O-2H x   580 30 Table 2. EPI SuiteTM predicted physico-chemical properties of bisphenol A bis(diphenyl phosphate) (BPADP), resorcinol diphenyl 581 phosphate (RDP) and tetrakis(2,6-dimethylphenyl)-m-phenylene biphosphate (RDX). Comparison is made to available experimental 582 values and legacy persistent, bioaccumulative and toxic (PBT) pollutants, bolded values exceed Canadian Environmental Protection 583 Act (CEPA) guidelines. 584 BPADP RDP RDX Water Solubility (mg/L) WSKOWWINTM WATERNTTM Experimental 1.09e-7 1.88e-6 0.212* to 0.4* 1.11e-4 6.88e-3 0.015* 3.70e-9 6.87e-7 n/a Soil Adsorption Coefficient (logKOC) (KOCWINTM) Experimental 10.52 4.00 ± 0.473; 4.76 ± 0.252 8.32 10.03 Octanol-Water Coefficient (logKOW) (KOWWINTM) Experimental 10.02 ≥6 7.41 11.79 Octanol-Air Coefficient (logKOA) (KOAWINTM) 21.74 18.33 22.37 Air-Water Coefficient (logKAW) (HENRYWINTM) -11.72 -10.92 -10.58 Vapour Pressure (MPBPVPTM) 5.99e-10 to 5.61e-6 5.99e-10 to 5.61e-6 5.99e-10 to 5.61e-6 Estimated Fugacity t1/2 (hr) (LEV3EPITM) Air; Water; Soil; Sediment 10.94; 1440; 2880; 1.30e4 12.13; 900; 1800; 8100 5.48; 4320; 8640; 3.89e4 Comparison of BPADP to legacy PBT compounds BPADP PCB-153 tetraBDE (congeners present in commercial PeBDE, OBDE and DBDE) pentaBDE (2 congeners, specific nomenclature not disclosed) logKOW 8* 6.72, 8.35 6.81 6.64-6.97; 6.57 Molecular Weight (amu) 692.63 360.88 485.79 564.7 Bioconcentration Factor (BCF) (L/kg wet weight) 644.3* 1.8539e4 1.3e6 ; 6.67e4 2.74e4 ; 1.77e4 ; 1440 Bioaccumulation Factor (BAF) (L/kg wet weight) 1790* 2.1e5 n/a 1.4e6 *conservative estimate of BPADP logKOW between experimental value and in silico predictions, used as an input value to predict BCF and BAF via EPI SuiteTM 585 Environment Canada, 2004, 2013; Gustafsson et al., 1999; EU 2001; CDC ATSDR; Schlechtriem et al., 2016 ; Zhong et al., 2021 586 31 Figure Legends 587 Figure 1. Comparison of the in vitro depletion of BPADP and the triphenyl phosphate (TPHP) 588 positive control over a 60-minute incubation period in a rat liver microsomal assay. Each data 589 point is the average of triplicate samples (n =3). 590 591 592 593 32 Figure 2. BPADP depletion and formation of the target metabolites diphenyl phosphate (DPHP) 594 and bisphenol A (BPA) in assays of a 1780 ppb (2.57 µM) incubation concentration in both 595 Wistar-Han rat (WH Rat) (a) and Great Lakes herring gull (HERG) (b) liver microsomal in vitro 596 assays. Each herring gull active replicate data point is the mean of inter-day triplicate (n=3) and 597 duplicate (n=2) assays (total n=5 replicates). Error bars represent the standard deviation of each 598 mean and an asterisk represents a time point where the active microsomal concentration of 599 BPADP or DPHP or the NADPH-deficient concentration of BPA differs significantly (p < 0.05) 600 from that of zero minutes at the start of the assay. 601 602 603 604 605 * * 33 Figure 3. UHPLC-QQQ-MS/MS analysis response vs. incubation time for four additional 606 metabolites of BPADP: (a) BPADP + O (X2); (b) BPADP + 2O (X4); (c) BPA-DPP; (d) BPA–607 DPP-MPP in Wistar-Han rat (WH Rat) and Great Lakes herring gull (HERG) liver microsome 608 assays. All data points are the mean of three replicates (n=3), error bars represent standard 609 deviation. 610 611 612 613 614 34 Supplementary Information 615 616 Biotransformation of bisphenol-A bis(diphenyl phosphate): In vitro, in silico, and (non-) 617 target analysis for metabolites in rat and bird liver microsomal models 618 619 Sofia M. Herczegha,b, Shaogang Chua, Robert J. Letchera,b* 620 621 a Ecotoxicology and Wildlife Health Division, Environment and Climate Change Canada, 622 National Wildlife Research Centre, Carleton University, Ottawa, ON, K1A 0H3, Canada 623 b Department of Chemistry, Carleton University, Ottawa, ON, K1S 5B6, Canada 624 625 626 627 628 629 630 631 632 633 * Corresponding author: Robert J. Letcher; Phone: 1-613-998-6696; Fax: 1-613-998-0458; E-634 mail: robert.letcher@ec.gc.ca. 635 35 UHPLC-Q-Exactive-HRMS/MS analysis 636 Non-target analysis (NTA) for additional metabolites was via a Vanquish UHPLC 637 coupled with a quadrupole orbitrap mass spectrometer (UHPLC-Q-Exactive-HRMS/MS; 638 ThermoScientific, Waltham, MA, USA). The mobile phases were the same as that used in 639 UHPLC-QQQ-MS/MS. A Kinetex XB-C18 column (100 x 2.1 mm, 1.7 µm particle size) was 640 used for separation. The flow rate was 0.3 mL/min and the gradient elution ramping was 5 % B 641 to 95 % B to 10 minutes, held for 10 minutes and after that the mobile phase composition was 642 returned to initial conditions in 1 min and then equilibrated for 5 min. The UHPLC-Q-Exactive-643 HRMS/MS was operated in both full scan mode and data depended acquisition fragment analysis 644 (dd-MS2/dd-SIM, DDA) mode. The MS was performed in positive polarity (ESI (+)) and 645 negative polarity (ESI(-)) with 70 000 resolution, scanning over the range of 70 to 1050 m/z. Ion 646 source parameters consisted of: 2.5 K spray voltage, S-lens RF level of 55, 45 sheath gas flow 647 rate, 10 auxiliary gas flow rate and zero sweep gas flow rate, 350 °C desolvation capillary 648 temperature, 450 °C aux gas heater temperature. Lock masses were 445.12003 m/z (Si(CH3)2O))6 649 for positive mode and 255.23295 m/z (C16H32O2) for negative mode. In full scan mode, AGC 650 target was 106 and maximum IT is 200 ms. In dd-MS2/dd-SIM mode, the resolution is 17500; 651 AGC target was 105; maximum IT was 50 ms; lop count was 5; isolation window was 1.2 m/z; 652 NCE was 30. 653 For identification of GSH adducts via UHPLC-Q-Exactive-HRMS/MS, a Zorbax eclipse 654 Plus C18 Rapid Resolution HT column (2.1 x 100 mm 1.8 µm particle size ) was used. The 655 column temperature was 40 °C and mobile phase flow rate was 0.3 mL/min. Mobile phase were 656 water (A) and ACN (B) and both contained 0.1 % formic acid. The gradient started at 5 % B, 657 held for 30 seconds before ramping up to 95 % B in 10 minutes, then held for 5 minutes, 658 36 decreased to 5 % B at minute 16, and then held until the end of the run at 21 minutes. MS was 659 performed in t-SIM mode with positive polarity (ESI(+)) to increase sensitivity, three most 660 possible target BPADP GSH adducts ions were selected in inclusion list, which included 661 BPADP+GSH-2H, BPADP+O+GSH and BPADP+O2+GSH-2H. The MS resolution was 70 662 000 and scanning range was 150 to 1200 m/z with a 4.5 m/z isolation window and AGC target 663 was 5e4 and maximum IT was 200 ms. Ion source parameters consisted of: 1.00 kV spray 664 voltage, S-lens RF level of 40.0, 80 sheath gas flow rate, 20 auxiliary gas flow rate and zero 665 sweep gas (N2) flow rate, 300 °C desolvation capillary temperature, and 500 °C aux gas heater 666 temperature. 667 668 37 Table S-1. SMILES codes of all in silico OECD Toolbox v4.4.1 predicted metabolites of 669 BPADP (generated via in vivo rat simulation and rat liver in vitro S9/microsomal simulation) and 670 corresponding IUPAC Name, abbreviation and CAS RN. 671 Metabolite Identifier Label Identified Compound M1. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(O)cc1)c1ccc(O)cc1; 4-[2-(4-hydroxyphenyl)propan-2-yl]phenol / bisphenol-A; 80-05-7 M2. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(O)cc1)c1ccc(OP(=O)(Oc2ccccc2)Oc2ccccc2)cc1; [4-[2-(4-hydroxyphenyl)propan-2-yl]phenyl] phenyl hydrogen phosphate / SCHEMBL811002; n/a M3. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(O)cc1)c1ccc(OP(O)(=O)Oc2ccccc2)cc1; [4-[2-(4-hydroxyphenyl)propan-2-yl]phenyl] phenyl hydrogen phosphate / SCHEMBL2290182; n/a M4. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(OP(O)(=O)Oc2ccccc2)cc1)c1ccc (OP(=O)(Oc2ccccc2)Oc2ccccc2)cc1; [4-[2-[4-[Hydroxy(phenoxy)phosphoryl]oxyphenyl]propan-2-yl]phenyl] diphenyl phosphate / SCHEMBL12670502; n/a M5. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(OP(O)(=O)Oc2ccccc2)cc1)c1ccc (OP(O)(=O)Oc2ccccc2)cc1; [4-[2-[4-[Hydroxy(phenoxy)phosphoryl]oxyphenyl]propan-2-yl]phenyl] phenyl hydrogen phosphate / SCHEMBL3693240; n/a M6. SMILES code IUPAC Name/Abbreviation CAS RN CC(C)(c1ccc(OP(O)(O)=O)cc1)c1ccc(OP(=O)(Oc2ccccc2)Oc2ccccc2)cc1; n/a; n/a M7. SMILES code IUPAC Name/Abbreviation CAS RN OP(=O)(Oc1ccccc1)Oc1ccccc1; diphenyl hydrogen phosphate / diphenyl phosphate; 838-85-7 M8. SMILES code IUPAC Name/Abbreviation CAS RN OP(O)(=O)Oc1ccccc1; phenyl dihydrogen phosphate / monophenyl phosphate; 701-64-4 M9. SMILES code IUPAC Name/Abbreviation CAS RN Oc1ccc(O)cc1; benzene-1,4-diol / hydroquinone 123-31-9 M10. SMILES code IUPAC Name/Abbreviation CAS RN Oc1ccccc1; phenol; 108-95-2 M11. SMILES code IUPAC Name/Abbreviation CAS RN Oc1ccccc1O; 1,2-dihydroxybenzene / catechol; 120-80-9 672 673 38 Table S-2. Complete list of chemical reagents and compounds referenced, identified by CAS RN 674 and including molecular weight and purity. 675 Reagent Name Molecular Weight (amu) CAS # Abbreviation Supplier Purity 13C12-Bisphenol A 240.20 263261-65-0 13C12-BPA Sigma-Aldrich >98 % Bisphenol A 228.29 80-05-7 BPA Sigma-Aldrich 99+ % Bisphenol A bis(diphenyl phosphate) 692.64 5945-33-5 BPADP Toronto Research Chemicals 98 % d10-Diphenyl phosphate 260.25 1477494-97-5 d10-DPHP Toronto Research Chemicals 95 % d15-triphenyl phosphate 341.38 1173020-30-8 d15-TPHP Wellington Labs >98 % Diphenyl phosphate 250.19 838-85-7 DPHP Sigma-Aldrich 99 % Trans-1,4-cyclohexanediol- bis(diphenyl phosphate) 580.516 No CAS# T-CH-BDP Sigma-Aldrich >97 % Triphenyl phosphate 326.28 115-86-6 TPHP Sigma-Aldrich 99+ % L-Glutathione reduced 307.32 70-18-8 GSH Sigma-Aldrich ≥98 % Compound Name Molecular Weight (amu) CAS # Abbreviation Supplier Purity Phenol 94.11 108-95-2 n/a n/a n/a Tetrakis(2,6- dimethylphenyl)-m- phenylene biphosphate 686 139189-30-3 RDX; PBDMPP n/a n/a resorcinol bis (diphenylphosphate) 574.4 57583-54-7 RDP; PBDPP n/a n/a Tetrakis(2- chloroethyl)dichloroisopent yldiphosphate 583 38051-10-4 V6; BCMP- BCEP n/a n/a 676 677 678 679 39 Table S-3. UHPLC-MS/MS analysis for determination of bisphenol A bis(diphenyl phosphate) 680 (BPADP), targeted metabolites diphenyl phosphate (DPHP), triphenyl phosphate (TPHP), and 681 bisphenol A (BPA), corresponding internal standards and additional targeted metabolites: 682 functions, channels and compound-dependent operation parameters. 683 Function Compound for Quantitative Analysis Precursor ion (m/z) Product ion(m/z) Cone (v) Coll. (v) ESI RT(min) Internal Standard 1 TPHP 327.1 152.1 48 32 (+) 4.41 d15-TPHP 327.1 77 48 36 d15-TPHP 342.2 160 48 32 342.2 82 48 36 2 T-CH-BDP 581 251 42 18 (+) 4.95 581 331.1 42 6 3 BPADP 693.2 367.1 80 38 (+) 5.45 T-CH-BDP 693.2 327.1 80 34 4 BPADP * 693.2 327.1 2 38 (+) 5.45 T-CH-BDP 693.2 367.1 2 34 5 BPA 227 133.1 52 26 (-) 3.45 13C12-BPA 227 212.1 52 12 13C12-BPA 239 139.1 52 26 (-) 3.45 239 224.1 52 12 6 DPHP 249 93.2 65 25 (-) 2.3 d10-DPHP 249 155.1 65 20 d10-DPHP 259.1 98.2 65 25 (-) 2.26 259.1 159.1 65 20 Function Additional Compound Precursor ion (m/z) Product ion (m/z) Cone (v) Coll. (v) ESI RT(min) 1 BPA-DPP 461.20 461.20 135.00 327.10 80 80 40 40 (+) 4.68 2 BPA-DPP-MPP 617.1 617.1 327.1 367.1 80 80 40 40 (+) 4.27 3 BPADP+O 709.2 709.2 327.1 367.1 80 80 40 40 (+) 5.10 4 BPADP+2O 725.2 725.2 367.1 465.2 80 80 40 40 (+) 4.86 * This function is for monitoring high concentrations of BPADP. 684 The italicized channels are for confirmation. 685 686 687 688 689 690 691 40 Table S-4. UHPLC-MS/MS limit of detection (ILOD) and limit of quantification (ILOQ) values 692 for target compounds: bisphenol A bis(diphenyl phosphate) (BPADP), diphenyl phosphate 693 (DPHP), triphenyl phosphate (TPHP) and bisphenol A (BPA). 694 # Name ILOD (ng/mL) ILOQ (ng/mL) Linearity-H (ng/mL) Linearity-L (ng/mL) r 1 TPHP 0.001 0.004 0-10 >0.99 3 BPADP 0.0008 0.003 0-10 >0.99 4 BPADP* 0.05 0.15 0-500 >0.99 5 BPA 0.09 0.29 0-500 0-500 >0.99 6 DPHP 0.01 0.04 0-500 0-500 >0.99 695 Table S-5. Chemical formula, parent compound, transformations, composition change (from 696 parent compound), Δ mass shift, molecular weight, retention time (RT) and replicated group 697 areas of the identified GSH adduct obtained from Q-Exactive HRMS (ThermoFisher Freestyle 698 version 1.6 software). 699 Chemical Formula 1: C49H49N3O16P2S Parent Compound Bisphenol A bis(diphenyl phosphate) Transformations Oxidation, Oxidation, GSH Conjugation 1 Composition Change +C10 H15 N3 O8 S Δ mass shift (ppm) 0.33 Molecular Weight (amu) 1029.23122 RT (min) 7.241 Replicated Group Areas (t120) Replicate 1: 1.3e5; Replicate 2: 3.95e4; Replicate 3: 0.0e0; Replicate 4: 4.99e5 700 701 702 703 704 705 706 707 708 41 709 710 Figure S-1. (a) Mean TPHP and DPHP concentrations in each active replicate sample, NADPH-711 deficient control and enzyme deactivated control of the assays conducted at a 1780 ppb BPADP 712 incubation concentration as presented in Figure 2. Each active microsomal data point is the 713 mean of three triplicate samples (n=9), each control data point is the mean of triplicate samples 714 (n=3). (b) Mean (n=2) TPHP depletion versus DPHP formation in Great Lakes Herring Gull 715 positive control samples. 716 717 718 a b 42 719 720 721 722 723 724 Figure S-2. Full scan mass spectra of confirmed metabolites of bisphenol A bis(diphenyl 725 phosphate) (BPADP) identified via QE-Orbitrap and Compound Discoverer v3.2: (a) BPA-DPP; 726 (b) BPA-DPP-MPP; (c) BPADP+O (X2) ; (d) BPADP+2O (X4); (e) Extracted ion 727 chromatograms of BPA+O (m/z = 709.175, 726.202, 731.157); (f) Extracted ion chromatograms 728 of BPA+2O (m/z = 742.197, 747.152) 729 43 730 Figure S-3. Molecular mass spectra of the confirmed glutathione (GSH) adduct of BPADP: 731 BPADP+GSH+2O-2H. 732 733 44 734 735 736 Figure S-4. Compound Discoverer v3.2 workflow tree developed for bisphenol A bis(diphenyl 737 phosphate (BPADP) metabolite Non-Target Analysis (NTA). 738 739 740 1. Introduction 2. Materials and methods 2.1 In silico metabolite profiling 2.2 Chemicals and reagents 2.3 In vitro liver microsomal assay optimization for BPADP 2.4 UHPLC-QQQ-MS/MS analysis 2.5 UHPLC-Q-Exactive-HRMS/MS analysis 2.6 Quality Assurance and Control 2.7 Data analysis 3. Results and discussion 3.1 BPA and DPHP target in vitro metabolites of BPADP 3.2 Additional target and non-target metabolites in the in vitro assays 3.3 In silico metabolite predictions and physico-chemical property estimations 3.4 Implications for the environmental fate of BPADP and risk assessment relevance 3. Conclusions Acknowledgements References UHPLC-Q-Exactive-HRMS/MS analysis