*2.2. Functional Analysis of Single Amino Acid rAbcc6 Mutants*

Our aim was to determine whether the residues in ABCC6 corresponding to those thought to be important in interaction with the physiological ABCC1 substrate LTC4, play a role in ABCC6-mediated ATP efflux. We used rAbcc6 in these studies, because it has higher activity in HEK293 cells than hABCC6 [10]. ATP and other nucleoside triphosphates (NTPs), the putative physiological substrates of ABCC6, carry multiple negative charges. We hypothesized such negatively charged substrates may be "coordinated" by positively charged residues in the substrate binding cavity of rAbcc6. Therefore, positively charged amino acid moieties (i.e., lysine, arginine, and histidine) at these positions, were replaced with uncharged residues (i.e., glutamine and alanine). Non-charged amino acid residues, according to the canonical/conservative mutagenesis practices, were changed into cavitycreating alanine residues, aimed at retaining the overall structure of the protein. The single amino acid rAbcc6 mutants generated for our study are summarized and positioned in a topology model of rAbcc6 below (Figure 3A). ‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ μ ‐ ‐ ‐ ‐ ‐ ‐ **Figure 3.** (**A**): Topology of the rAbcc6 amino acids analogous to those that comprise the LTC<sup>4</sup> binding cavity in bAbcc1. The inactivating mutation in the NBD2, E1426Q, is also indicated (**B**): Expression of the rAbcc6 single amino acid mutants in HEK293 cells. Of the total cell protein, 5 µg was fractionated on a 7.5%-polyacrylamide gel and bands corresponding to wild-type and mutant rAbcc6 proteins and the housekeeping protein tubulin were detected by Western-blot analysis using the K14 anti-rat Abcc6 antibody and the anti-tubulin antibody, respectively. The slight differences in electrophoretic mobility of some of the mutants may be attributed to altered glycosylation or other post-translational modifications.

‐ ‐ Levels of the mutant rAbcc6 proteins in HEK293 cells varied but were within the same range as those of wild-type Abcc6 (Figure 3B). We then characterized the functionality of the rAbcc6 mutants by following PPi accumulation in the culture medium as an indirect measure of NTP release (Figure 4A,B), as well as by directly determining ATP efflux using a luciferin/luciferase-based assay (Figure 4B). In both assays the untransfected, parental, HEK293 cell line as well as the cell line expressing the catalytically inactive E1426Q mutant, did not release substantial amounts of ATP into the culture medium. In contrast, cells overproducing wild-type rAbcc6 released large amounts of ATP, resulting in robust PPi accumulation in the culture medium (Figure 4). These results demonstrate the suitability of these assays for measuring the consequences of the mutations introduced into rAbcc6. ‐ ‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ ‐ ‐ **Figure 4.** Activity of rAbcc6 mutants in HEK293 cells. (**A**): PPi accumulation in culture medium and (**B**): ATP efflux from cell lines overexpressing rAbcc6 in which amino acid residues corresponding to those forming the bAbcc1 LTC<sup>4</sup> binding cavity were mutated. Data are presented as means ± SD for (**A**). For (**B**), means of representative experiments, each with at least 4 replicates are shown. In (**B**) data are presented in two graphs to better see results of individual mutants. Wild type: wild-type rAbcc6, control: parental HEK293 cells. The dashed line in (**A**) indicates the average amount of PPi in medium of HEK293 cells overproducing wild-type rAbcc6, which was set at 100%. Values have been adjusted to take any differences in protein expression of the mutants relative to wild type rAbcc6 into account. The same color coding was used for each mutant in panels A and B. \*\*\*\* *p* < 0.001 (ANOVA and subsequent Dunnett's multiple comparison test). Changes were considered biologically relevant when reduced by >50% compared to wild-type rAbcc6.

‐ ‐ ‐ ‐ ‐ ‐ ‐ Many of the rAbcc6 single amino acid mutants allowed cellular ATP efflux similar to that seen for wild-type rAbcc6, as determined by both PPi accumulation in the medium (Figure 4A) and the direct ATP efflux assay (Figure 4B). ATP release was substantially reduced (>75%) when M369, L534, R1168, T1214, and R1220 were mutated (Figure 4A,B). Changing L316 and H424 residues into alanine moderately reduced (>50%) efflux activity of rAbcc6. The substitution of the other residues did not reduce, or less substantially reduced, ATP efflux. The two arginine residues critical for function (rAbcc6 R1168 and R1220) belong to TMD2. In the ATP-free, LTC4-bound, inward-facing conformation, R1220 (TM17) localizes near one of the entrances of the modelled substrate-binding cavity, lined by TM15 and TM17 (Figure 5A). M369 (TM7), L534 (TM10), and R1168 (TM16) approximately lie on the same plane as R1220, and their side chains are exposed to the main cavity, with L534 located on the opposite side (Figure 5A). Among the residues that abolish ATP efflux when mutated, T1214 (TM17) is the one located further up in the transmembrane cavity (Figure 5A). In this conformation, among the other residues considered in

this study, the side chains of L316 (TM6) and S319 (TM6) are those further away from the main cavity (Figure 5A). In the outward-facing, ATP-bound state, the cavity opens towards the opposite side of the membrane. In this state many of the residues are more buried and located towards the bottom of the outward-facing cavity (Figure 5B). As mentioned earlier, R1168 and R1220 are conserved among all the sequences considered for model building (see Supporting Information files TMD1-NBD1\_Alignment.pdf and TMD2- NBD2\_Alignment.pdf), as well as all ABCC family members and their orthologs. The analogous amino acids are indispensable for all the transport activities of hABCC1, suggesting a key function in overall protein structure [32,33], although we cannot completely rule out interaction with its substrates. M369 is present as a phenylalanine and a tryptophan in the ABCC1 and ABCC5 sequences, respectively, while rAbcc6 L534 is a threonine in ABCC1 and a valine in ABCC5 (see Supporting Information file TMD2-NBD1\_Alignment.pdf). rAbcc6 T1214 is also not conserved among the sequences here considered, and it is present primarily as a tyrosine in ABCC6 and as a leucine in ABCC5 sequences (see Supporting Information file TMD2-NBD2\_Alignment.pdf). ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐

‐ ‐ ‐ ‐ ‐ ‐ **Figure 5.** Mutation of the amino acids forming the modelled rAbcc6 substrate-binding cavity affect ATP efflux to different degrees. View of the (**A**) ATP-free inward-facing and (**B**) ATP-bound, outwardfacing models of rAbcc6. rAbcc6 residues that when mutated abolished and reduced ATP efflux are shown as red and orange spheres, respectively. Residues that once mutated did not affect ATP efflux are shown as green spheres. The protein is shown as a white transparent surface and the volume of the main cavity in both models is shown as a gray surface. The volume was calculated with the 3V webserver [34].
