Structural basis for functional interactions in dimers of SLC26 transporters

Abstract

The SLC26 family of transporters maintains anion equilibria in all kingdoms of life. The family shares a 7 + 7 transmembrane segments inverted repeat architecture with the SLC4 and SLC23 families, but holds a regulatory STAS domain in addition. While the only experimental SLC26 structure is monomeric, SLC26 proteins form structural and functional dimers in the lipid membrane. Here we resolve the structure of an SLC26 dimer embedded in a lipid membrane and characterize its functional relevance by combining PELDOR/DEER distance measurements and biochemical studies with MD simulations and spin-label ensemble refinement. Our structural model reveals a unique interface different from the SLC4 and SLC23 families. The functionally relevant STAS domain is no prerequisite for dimerization. Characterization of heterodimers indicates that protomers in the dimer functionally interact. The combined structural and functional data define the framework for a mechanistic understanding of functional cooperativity in SLC26 dimers.

Fig. 1

Dimer interfaces in 7TMIR proteins. a Side view of the membrane domains of NBCe1 (PDB: 6CAA) and UraA (PDB: 5XLS). Core and gate domain are colored orange and gray, respectively, with residues within 4 Å of the opposing protomer in pink. b Top views of the dimeric arrangements of NBCe1 and UraA. For each dimer, the gate domain of one of the protomers follows a rainbow coloring scheme (blue-to-red for N-to-C direction). TMs central in the respective dimers are numbered. c Side view of the membrane domain of SLC26Dg (PDB: 5DA0). Residues mutated to cysteine for site-directed spin labeling are colored blue. The circled numbers indicate the respective TMs

Fig. 2

Interspin distances in the SLC26Dg dimer. a Primary PELDOR data of detergent-solubilized K353R1. b–d Left panels: background-corrected PELDOR time traces for membrane-reconstituted K353R1, V367R1, and L385R1 (black traces), overlaid with the fit from Tikhonov regularization (green), and forward-calculated PELDOR time traces from BioEn spin-label rotamer refinement of the MD simulation model (magenta, dashed; θ = 10). Right panels: distance distributions obtained by Tikhonov regularization (green), overlaid with the distance distributions resulting from BioEn analysis of the MD simulation model (magenta, dashed). Original PELDOR data in Supplementary Fig. 2. e C_α-atom root mean squared distance (RMSD) values of the core, gate, TM13, and 14 relative to the monomer crystal structure as a function of MD time (1 μs)

Fig. 3

Model of the SLC26Dg dimer interface. a Side view of the SLC26Dg membrane domain in the same orientation as Fig. 1a. Core and gate domain are colored orange and gray, respectively, with residues within 4 Å of the opposing protomer in pink. b Top views of the dimeric arrangement of SLC26Dg. The gate domain of one of the protomers follows a rainbow coloring scheme (blue-to-red for N-to-C direction)

Fig. 4

Oxidative cysteine cross-linking between TM14 of SLC26Dg. a In gel GFP fluorescence analysis of disrupted E. coli cells expressing single-cysteine variants of SLC26Dg fused to superfolder GFP. Following oxidative cross-linking, samples were analyzed by non-reducing SDS-PAGE. Cysteine-free SLC26Dg (cysless) and L144C (TM5) represent negative controls. Black and white arrows indicate dimeric and monomeric SLC26Dg. Source data are provided as a Source Data file. b Side view of the SLC26Dg dimer model. Core and gate domain are colored orange and gray, respectively. Positions in TM14 susceptible to cross-linking are colored in green, non-susceptible residues are colored pink. The gate domain of the right protomer is depicted in surface representation. TM13 of the left protomer is contoured. The circled numbers indicate the respective TMs

Fig. 5

Generation and functional characterization of SLC26Dg-IL. a Surface representation of MD-simulated SLC26Dg clipped through the funnel toward the putative substrate-binding site. Cytoplasmic water molecules in a ~ 10 Å slab at the clipping plane are shown. Ile-45 and Ala-142 indicate the relative position of the cysteine mutants in the core (orange) and gate (gray) domain, respectively. b SDS-PAGE analysis of purified and cross-linked SLC26Dg-IL monomers in the absence and presence of DTT. Single and double stars indicate not-cross-linked and cross-linked protein, respectively. c Functional characterization of membrane-reconstituted and cross-linked SLC26Dg-IL (dark blue), wildtype SLC26Dg (orange), and both proteins mixed in equal ratio’s (pink). Closed and open symbols indicate the absence and presence of a pre-incubation step with DTT. d Initial transport rates of membrane-reconstituted and cross-linked samples containing wildtype and SLC26Dg-IL mixed in different ratio’s. Dark blue, pink, and orange dashed curves indicate the anticipated curves assuming an activity of the heterodimers corresponding to 0, 50, and 100% of the wildtype homodimers. These models were calculated assuming stochastic dimer formation (e.g., mixing WT:IL protomers in a 50:50 ratio results in 25% WT–WT, 50% WT–IL, and 25% IL–IL dimers) and specific transport activities of 32.3 or 6.8 nmol fumarate per mg WT or IL homodimer per min, respectively, and heterodimer activities corresponding to 0, 50, or 100% of WT homodimer. Data points represent mean and standard deviations of three technical replicates. Source data are provided as a Source Data file