The multidrug transporter Pdr5 from Saccharomyces cerevisiae is an ergosterol-modulated efflux pump

Nentwig, L.-M.; Helmbrecht, J.; Kaplan, E.; Khosa, S.; Song, H.; Westhoff, P.; Robinson, C. V.; Smits, S.; Szewczak-Harris, A. M.; Exterkate, M.; Luisi, B.; Schmitt, L.

doi:10.64898/2026.10.29.762749

bioRxiv · 10.64898/2026.10.29.762749

The multidrug transporter Pdr5 from Saccharomyces cerevisiae is an ergosterol-modulated efflux pump

Nentwig, L.-M.·Helmbrecht, J.·Kaplan, E.·Khosa, S.·Song, H.·Westhoff, P.·Robinson, C. V.·Smits, S.·Szewczak-Harris, A. M.·Exterkate, M.·Luisi, B.·Schmitt, L.

View original source ↗Export metadata ↓

Abstract

The pleiotropic drug resistance ABC transporter Pdr5 and its homologs play a critical role in multidrug resistance of numerous fungi, including pathogens. Previous studies demonstrated that alterations in the composition of the plasma membrane in which Pdr5 is embedded, specifically the ergosterol content, affected transport activity. Here we report the cryo-EM map of Saccharomyces cerevisiae Pdr5 in complex with ATP/ADP at 2.2 [A] resolution, revealing densities on the surface of the transmembrane domains as well as inside the substrate-binding site, that were assigned as ergosterol, identified and quantified by GC-MS, and the sphingolipid M(IP)2C, identified via native MS. The ergosterol molecules are bound to the protein in well-defined grooves and unexpected locations. We found that the activity of Pdr5 follows a Gaussian-like distribution, increasing with the ergosterol-concentration to reach a maximum before declining at higher concentrations, but this modulation was not observed in the presence of cholesterol, the sterol analogue found in mammalian cells. Mutational analysis corroborates the role of a single ergosterol-binding site in the modulation of Pdr5 activity. Our structural analysis identifies a protein-lipid network that could serve as a center for lipid engagement linking the surrounding lipid environment to the substrate-binding region of Pdr5. These findings advance our understanding of Pdr5 function by demonstrating that the lipid environment is closely linked to Pdr5 structure and activity.

Explore related subjects

biochemistry Preprint

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Nentwig, L.-M., Helmbrecht, J., Kaplan, E., Khosa, S., Song, H., Westhoff, P., Robinson, C. V., Smits, S., Szewczak-Harris, A. M., Exterkate, M., Luisi, B., Schmitt, L.. 2026-10-30. The multidrug transporter Pdr5 from Saccharomyces cerevisiae is an ergosterol-modulated efflux pump. https://doi.org/10.64898/2026.10.29.762749

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related preprints

bioRxiv2026

Transient diquat inclusion and carboxylate-rim binding to carboxymethyl-α-cyclodextrin: a molecular dynamics study

Rotaxane-based supramolecular scavengers have shown promise for diquat detoxification, but the relative contributions of cavity inclusion and interactions at the macrocycle rim remain unclear. We used all-atom molecular dynamics simulations in explicit water to examine diquat binding to carboxymethyl--cyclodextrin (CM--CD) and the behaviour of CM--CD on a poly(ethylene glycol) axle. Five systems were simulated: host-guest mixtures containing a neutral host (pH 3 model) or a hexaanionic host (pH 10 model), an explicitly constructed inclusion complex, and two-macrocycle pseudopolyrotaxanes initiated from pre-threaded or random configurations. The total production sampling was approximately 4.5 s. A relaxed threaded geometry was obtained, with the diquat centroid 0.03 nm from the cavity axis and 9 of its 14 heavy atoms within the defined cavity cylinder. The guest left this geometry within the first 50 ps sampling interval and subsequently remained associated mainly with the carboxylate rim. During 1.2 s of unbiased sampling of the hexaanionic host-guest system, only 5 of 802 guest-frame pairs contained a guest atom within the cavity cylinder, and all of these events occurred near a portal. Rim contacts were more frequent: each guest formed an average of 3.8-4.4 host heavy-atom contacts per frame, with 12.8-14.0% of guest heavy atoms in contact with the host. These interactions were substantially reduced for the neutral host. Sodium ions remained predominantly solvent-separated from the carboxylates. In the pseudopolyrotaxane simulations, pre-threaded rings satisfied the threading criterion in 1.6-8.4% of frames, and no spontaneous threading was observed from random configurations during 1 s. Under the simulated conditions, CM--CD therefore bound diquat primarily through dynamic rim interactions rather than persistent cavity inclusion.

biochemistry↗

bioRxiv2026

Validated LC-MS/MS quantification of kynurenine pathway metabolites and pathway-balance ratios in brain tissue and microglial culture supernatant: application to neonatal stroke

Kynurenine pathway (KP) metabolite imbalance is implicated in perinatal brain injury, and KP metabolites are measured in human plasma and cerebrospinal fluid as markers of neuroinflammation. Methods validated in neonatal brain tissue or glial culture supernatant to the same standard are scarce, limiting comparisons across preclinical and clinical measurements. We optimised LC-MS/MS detection of tryptophan (TRYP), kynurenine (KYN), kynurenic acid (KYNA), 3-hydroxykynurenine (3-HK), 3-hydroxyanthranilic acid (3-HAA), quinolinic acid (QA), anthranilic acid and picolinic acid after protein precipitation and Enhanced Matrix Removal-Lipid clean-up, validated it against Clinical and Laboratory Standards Institute (CLSI) acceptance criteria in brain tissue homogenate and cell culture supernatant, and applied it to a rat model of neonatal stroke and immune-activated BV2 microglia, reporting KYN/TRYP, KYNA/KYN, 3-HK/3-HAA and QA/KYNA ratios. The lower limit of quantification was 6.53-34.17 nM; imprecision was 0.57-11.47%, accuracy was 84.9-108.3%, recovery was 83.2-112.3%, and matrix effect was 87.8-108.8%, with linear calibration (R2 > 0.99) over 6.5-2000 nM. KYN/TRYP and QA/KYNA ratios were higher in the stroke hemisphere than in control brain at 1, 7 and 14 days (p<0.001), and KYNA/KYN was lower; lipopolysaccharide- and cytokine-stimulated BV2 microglia showed the same change in all four ratios (p<0.01). The assay quantifies eight KP metabolites and four pathway-balance ratios in brain tissue and microglial supernatant within clinical acceptance limits, allowing direct comparison with KP ratios reported in human plasma and cerebrospinal fluid. Changes in these ratios after neonatal stroke identify KYN/TRYP and QA/KYNA as candidate targets for neonatal biofluid studies.

biochemistry↗

bioRxiv2026

One lysine, two acyl marks: 2-hydroxyisobutyrylation and acetylation of G6PD differ in outcome and regulation

Lysine 2-hydroxyisobutyrylation is a widespread post-translational modification enriched in enzymes of central carbon metabolism, yet its site-specific biochemical consequences remain poorly defined and usually inferred from amino acid substitutions or heterogeneously modified protein preparations. Using genetic code expansion technology, we generated homogeneous human glucose-6-phosphate dehydrogenase (G6PD) carrying an authentic 2-hydroxyisobutyryl group at nine individual lysine residues. 2-hydroxyisobutyrylation affected G6PD activity in a strongly site-dependent manner: modification within or adjacent to the catalytic pocket (K171, K205, K360), or at the tetramer interface (K288), abolished activity, whereas modification of K386, near the structural NADP+ binding site, reduced catalytic efficiency and destabilized the enzyme. HibK386 G6PD was less oligomeric, while inactive HibK288 G6PD formed stable oligomers, suggesting that in HibK288 G6PD, assembly and catalysis are separable. Molecular dynamics simulations indicated that 2-hydroxyisobutyrylation of K386 reorients the side chain toward the solvent and weakens hydrogen bonding and salt-bridge across the tetramer interface. K386 is also subject to acetylation, allowing the comparison between two acyl groups at one residue; acetylation destabilized G6PD without substantially impairing catalysis, whereas 2-hydroxyisobutyrylation did both, and the two marks were differentially recognized by their erasers. SIRT2 removed both and restored the activity of the purified enzyme, whereas SIRT1 readily deacetylated AcK386 but was a poor de-2-hydroxyisobutyrylase of this position. SIRT2 also de-2-hydroxyisobutyrylated HibK288 G6PD and PKM2 HibK62, but not PKM2 HibK115 or HibK305. Together, these findings identify site-specific 2-hydroxyisobutyrylation as a potent inhibitory modification of G6PD, demonstrate that chemically distinct acylations at the same site can exert different functional effects, and establish deacylase specificity as a function of both modification chemistry and residue context.

biochemistry↗