**1. Introduction**

Inactivating mutations in the gene encoding ATP-binding cassette (ABC) subfamily C member 6 (ABCC6) underlie the autosomal recessive disease pseudoxanthoma elasticum (PXE, OMIM #264800) [1–3], characterized by ectopic mineralization in the skin, eyes, and vascular system [4–6]. PXE is a slowly progressing connective tissue disorder that affects approximately 1 in 50,000 individuals worldwide [7]. There is currently no specific and effective therapy for PXE and the disease slowly progresses after initial diagnosis [8].

ABCC6 is predominantly expressed in the liver [9] where it mediates the release of ATP from hepatocytes into the bloodstream [10,11]. Outside the hepatocytes, yet still in the liver niche, the released ATP is converted into AMP and the mineralization inhibitor pyrophosphate (PPi), by ectonucleotide pyrophosphatase phosphodiesterase 1 (ENPP1) [12]. The absence of ABCC6-mediated ATP release in both PXE patients and Abcc6 null mice results in plasma PPi levels that are < 40% of those found in ABCC6-proficient individuals [11], providing a plausible biochemical explanation for their ectopic mineralization.

**Citation:** Szeri, F.; Corradi, V.; Niaziorimi, F.; Donnelly, S.; Conseil, G.; Cole, S.P.C.; Tieleman, D.P.; van de Wetering, K. Mutagenic Analysis of the Putative ABCC6 Substrate-Binding Cavity Using a New Homology Model. *Int. J. Mol. Sci.* **2021**, *22*, 6910. https://doi.org/10.3390/ijms22136910

Academic Editor: Satoshi Kametaka

Received: 8 May 2021 Accepted: 23 June 2021 Published: 27 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Moreover, plasma PPi levels decline during pregnancy, which might explain the increased risk of vascular calcification in multiparous individuals [13]. Recent data indicate that ATP efflux by the progressive ankylosis protein (ANK) is also a major determinant of plasma PPi levels [14]. Intriguingly, an ABC protein other than ABCC6 has been reported to also be involved in cellular ATP release, albeit indirectly, as two ABCG1 variants were found to control volume-regulated anion channel-dependent ATP release by regulating cholesterol levels in the plasma membrane [15]. Neither of these ABCG1 variants, however, have been implicated in the pathology of PXE.

Although low levels of circulating PPi explain why PXE patients suffer from ectopic mineralization, the molecular details of ABCC6-mediated ATP release remain elusive. Most ABC proteins of the C-branch function as ATP-dependent efflux transporters, though there are several exceptions. Thus, ABCC7 is the ATP-gated chloride channel cystic fibrosis transmembrane conductance regulator (CFTR) with inactivating mutations causing cystic fibrosis [16], and ABCC8 and ABCC9 are regulatory subunits of complex potassium channels [17].

Most of the currently available data indicate that ABCC6 is an ATP-dependent ATP efflux transporter: ATP efflux rates from ABCC6-transfected HEK293 cells are very similar to rates at which ABCC1, ABCC2, and ABCC3 transport morphine-3-glucuronide out of cells [18]. Moreover, our recent work indicates ABCC6 does not function as an ATP channel [19] and nor does it induce the exocytosis of ATP-loaded vesicles (our unpublished data). ABCC6 was initially implicated in the transport of glutathione conjugates in in vitro vesicular uptake assays [20,21] but these results proved difficult to reproduce in later studies [6].

In 2017, the structure of LTC4-bound bovine Abcc1 (bAbcc1) in the ATP-free state, with a bipartite transmembrane cavity open towards the cytosol (inward-facing) was reported using cryogenic electron microscopy (cryoEM) [22]. This report was later followed by the cryoEM structure of the ATP-bound, outward-facing state of bAbcc1, with the transmembrane cavity open to the opposite side of the membrane [23]. Given that (1) ABCC6 shares most sequence similarity with ABCC1 [24], (2) the genes encoding both proteins arose from a recent gene duplication [25], and (3) in vitro studies suggested both proteins might share LTC<sup>4</sup> as a substrate [4,21,26,27] though attempts to connect the transport of LTC<sup>4</sup> to the potential role of ABCC6 failed [6].

We used the ATP-free, LTC4-bound and ATP-bound, substrate-free, bAbcc1 cryoEM structures as templates to build inward- and outward-facing homology models of hABCC6 and rat Abcc6 (rAbcc6) as a means of identifying amino acids potentially forming the binding cavity for ATP. Amino acids in ABCC6 at the same positions as those in bAbcc1 comprising the proposed bipartite binding cavity of LTC<sup>4</sup> were subsequently mutated in rAbcc6 expression vectors and the mutant rAbcc6 proteins functionally characterized to determine if they play a role in ABCC6-dependent ATP release. Several of the introduced mutations did not markedly alter rAbcc6 activity and thus are not essential for ATP efflux. Strikingly, the generation of a rAbcc6 mutant in which all amino acids of the modeled binding cavity were changed into their ABCC1 counterparts, showed ATP efflux similar to the wild-type protein.
