In silico prediction of the interaction of legacy and novel per- and poly-fluoroalkyl substances (PFAS) with selected human transporters and of their possible accumulation in the human body

Per- and poly-fluorinated compounds constitute a wide group of fluorocarbon chemicals with widespread industrial applications, ranging from non-stick coating in cookware to water surfactants, from fire-fighting foams to water-repellent coatings on textiles. Presently, over 12,000 PFAS are known worldwide. In recent years, extensive research has focused on investigating the biological effects of these molecules on various organisms, including humans. Here, we conducted in silico simulations to examine the potential binding of a representative selection of PFAS to various human proteins known to be involved in chemical transportation and accumulation processes. Specifically, we targeted human serum albumin (HSA), transthyretin (TTR), thyroxine binding protein (TBG), fatty acid binding proteins (FABPs), organic anion transporters (OATs), aiming to assess the potential for bioaccumulation. Molecular docking simulations were employed for this purpose, supplemented by molecular dynamics (MD) simulations to account for protein flexibility, when necessary. Our findings indicate that so-called “legacy PFAS” such as PFOA or PFOS exhibit a higher propensity for interaction with the analysed human protein targets compared to newly formulated PFAS, characterised by higher branching and hydrophilicity, and possibly a higher accumulation in the human body. Supplementary Information The online version contains supplementary material available at 10.1007/s00204-024-03797-0.


Figures and Table List
. Analysed perfluorinated compounds database.Each acronym is derived from its common name except for the novel substances, which are usually called by their commercial names.IUPAC names and base structures are provided.The PDB file contained the biological assembly, but the crystallographic structure was resolved only for the dimer (rainbow-coloured).The grey half of the protein was obtained, by Zhang et al. (2016), using symmetry; c.Thyroxine Binding Globulin (TBG); d.Fatty Acids Binding Protein (FABP), Liver isoform; e. FABP, Intestinal isoform; f.FABP, heart isoform; g.FABP, adipocyte isoform; h.Peripheral myeling Protein 2 (PmP2) isoform; i.Organic Anion Transporters; the "inward" isoforms are reported on the left, the "outward" isoforms on the right.i1 and i2: OAT1; j1 and j2: OAT3; k1 and k2: OAT4; l1 and l2:        Least square fitting was performed on the pocket residues with the starting frame as reference structure, and the rmsd was then calculated (through GROMACS 4.6.1 package).The opaque line is the unweighted moving average, or rolling mean, calculated for 100 frames.

Table
Figure S1 (a-k) Visual representation of the proteins and related pockets.….…...

Figure S1 .
Figure S1.Visual representation of the studied set of proteins.The pockets reported are represented as red contours.a. Human Serum Albumin (HSA).Nine different pockets were considered for our docking studies, four of them for molecular dynamics simulations; b.Transthyretin (TTR) bound with PFOA (in both pockets).The PDB file contained the biological assembly, but the crystallographic structure was resolved only for the dimer (rainbow-coloured).The grey half of the protein was obtained, by Zhang et al. (2016), using symmetry; c.Thyroxine Binding Globulin (TBG); d.Fatty Acids Binding Protein (FABP), Liver isoform; e. FABP, Intestinal isoform; f.FABP, heart isoform; g.FABP, adipocyte isoform; h.Peripheral myeling Protein 2 (PmP2) isoform; i.Organic Anion Transporters; the "inward" isoforms are reported on the left, the "outward" isoforms on the right.i1 and i2: OAT1; j1 and j2: OAT3; k1 and k2: OAT4; l1 and l2: URAT1.

Figure S2 .
Figure S1.Visual representation of the studied set of proteins.The pockets reported are represented as red contours.a. Human Serum Albumin (HSA).Nine different pockets were considered for our docking studies, four of them for molecular dynamics simulations; b.Transthyretin (TTR) bound with PFOA (in both pockets).The PDB file contained the biological assembly, but the crystallographic structure was resolved only for the dimer (rainbow-coloured).The grey half of the protein was obtained, by Zhang et al. (2016), using symmetry; c.Thyroxine Binding Globulin (TBG); d.Fatty Acids Binding Protein (FABP), Liver isoform; e. FABP, Intestinal isoform; f.FABP, heart isoform; g.FABP, adipocyte isoform; h.Peripheral myeling Protein 2 (PmP2) isoform; i.Organic Anion Transporters; the "inward" isoforms are reported on the left, the "outward" isoforms on the right.i1 and i2: OAT1; j1 and j2: OAT3; k1 and k2: OAT4; l1 and l2: URAT1.

Figure S3 .
Figure S3.RMSD (nm) for the MD simulations of HSA complexed with PFOA and cC6O4 RS.The RMSD was calculated through GROMACS 4.6.1 package with the starting frame as reference structure.The opaque line is the unweighted moving average, or rolling mean, calculated for 100 frames.

Figure S5 .Figure S6 .
Figure S5.Hbond persistency calculated for the MD simulations of HSA complexed with PFOA (a) and cC6O4 (b).Each dot (node) represents a residue, in its one-letter code and number, or a ligand.Their color represents the number of link incidents on that node according to degree centrality.The number on each link represents the occupancy of each H-bond interaction expressed as a percentage of simulation time.Occupancies lower than 1% are not reported.Data have been calculated with the Bridge2 package.

Figure S7 .Figure S8 .
Figure S7.Superposition of the crystallographic pose (gray; PDB ID 5JID) and the docked pose (orange) of PFOA in TTR binding site after IFD.The protein is shown as cartoon, the ligands in capped sticks and the H-bonds as yellow dashed lines.

Figure S9 .Figure S11 .Figure S12 .
Figure S9.RMSD (nm) for the MD simulations of TTR complexed with PFOA (a) and cC6O4 RS (b).Calculations were performed with GROMACS 4.6.1 package on the protein backbone, using the starting frame as reference.The opaque line is the unweighted moving average, or rolling mean, calculated for 100 frames.The three lines correspond to the RMSD calculated for each MD replica.

Figure S13 .
Figure S13.Docking scores (kcal/mol) for PFAS series docked in TBG.Graphs are divided by compound classes, maxima and minima of each PFAS are reported in red dots, while intermediate poses are reported in grey.

Figure S14 .Figure S15 .Figure S16 .Figure S17 .Figure S18 .Figure S19 .Figure S20 .
Figure S14.Docking pose of PFOA and cC6O4 in TBG.The protein is shown as cartoon, the ligands in capped sticks and the H-bonds as yellow dashed lines.

Figure S22 .Figure S23 .
Figure S22.Visual representation of the docking poses of PFOA (a) and cC6O4 (b) in OAT1 outward.The highest affinity poses were selected for visualization.The protein is shows as cartoon, the ligands and crucial residues in capped sticks, H-bonds as yellow dashed lines.The exterior of the cell, in OAT1's case the lumen side of the kidney cell, is on the right of the figures, while the interior of the cell is on the left side.