Cation Channel Activity of Apolipoprotein L1 is Modulated by Haplotype Background

Two variants in apolipoprotein L1 (ApoL1) are known to be responsible for the increased risk of progressive proteinuric kidney disease in people of recent African ancestry.^1,2 Wild-type ApoL1 protein (designated G0) is an amphipathic protein that can insert into lipid membranes at low pH, where it increases anion permeability³; on titration back to neutral, the permeability transitions to cation selectivity.^3–6 The two kidney disease–associated variants (designated G1 and G2) show increased cation selective channel activity.⁵

In addition to G1 and G2, there are other ApoL1 protein isoforms, not known to be involved in kidney disease, which are inherited as distinct haplotypes.⁷ Three of these haplotypes, designated KIK, EIK, and EMR, account for the majority of protein isoforms in the human population (Supplemental Table 1 and Supplemental Appendix 1). KIK is the most prevalent worldwide. EIK is the most common haplotype in people of African ancestry and is absent from other populations; G1 and G2 variants are only naturally found in the EIK background. EMR (the GenBank refseq, thought to be of Neanderthal origin) is most frequently encountered in European populations.

ApoL1 has direct cytotoxic effects that are altered by haplotype background.⁷ To test whether haplotype alters ion permease activity, we generated expression constructs encoding the G0, G1, and G2 variants in the KIK, EIK, and EMR haplotypes, and assayed resulting proteins for ion permease activity and membrane association.

ApoL1 potassium permeability assays are shown in Figure 1A. Output (mV) from the chloride-selective electrode is shown, starting with 10 μM KCl in buffered sucrose (pH 7.5). Protein and KCl-loaded liposomes (combined at pH 6.0) with no extravesicular KCl are added at the first arrow. Chloride ionophore is added at the second arrow, inducing lumen-positive potential driving KCl efflux, limited by the K permeability. Triton X-100 is added at the third arrow, releasing remaining intravesicular KCl. The maximal rate of fractional Cl release after addition of ionophore is taken as the potassium permeability. Permeability conferred by each protein is equal to the rate in presence of protein minus the rate in the absence of protein.

Figure 1.

Ion efflux assays. (A) Examples of raw data from K permeability assays. First arrow indicates addition of protein-vesicle mixture after passage through a desalting spin column. Second arrow indicates addition of chloride ionophore. Third arrow indicates addition of Triton X-100. Dotted tracing is of no-protein control, solid black tracing is of G1 variant in the EIK haplotype, solid gray tracing is of G1 variant in the EMR haplotype. The straight lines indicate the maximal rates of chloride release after addition of ionophore, reflecting K permeability. Permeability conferred by the protein is determined from the difference between the efflux rates in the presence and absence of added protein. (B) Example of raw data from Cl permeability assays. Tracings as in (A), except the second arrow indicates the addition of valinomycin rather than chloride ionophore. The tracings from EIK-G1 and EMR-G1 are nearly superimposed. (C) K permeability conferred by each variant in each haplotype reported as the fractional Cl release (%/s). Circles indicate individual data points. Horizontal line with error bar indicates the median and interquartile range. Solid brackets indicate P<0.05 for comparisons among haplotypes for a single variant. Dashed brackets indicate P<0.05 for comparisons among variants within a single haplotype. (D) Results of membrane association assay under conditions supporting K permease activity, reported as ng protein bound. Markers as in (B). See Supplemental Table 2 for medians, interquartile ranges, n, and P values for panels (C) and (D). (E) Cl permeability conferred by each variant in each haplotype reported as the fractional Cl release (% per sec) for each variant in each haplotype as marked. Markers as in (C). (F) Results of membrane association assay under conditions supporting Cl permease activity, reported as ng protein bound. Markers as in (D). See Supplemental Table 3 for medians, interquartile ranges, n, and P values for (E) and (F).

Potassium permeability conferred by each variant in each haplotype is shown in Figure 1B and Supplemental Table 2. For each variant, the permeability in EIK is 1.8- to 2.3-fold higher than the permeability in EMR, whereas the permeability in EIK is not significantly different from that in KIK.

Within each haplotype the potassium permeability conferred by G1 and G2 is nearly double that conferred by G0, very like our previous report using the KIK haplotype.

Stable ApoL1 membrane association under optimal conditions for the cation permease is shown in Figure 1C and Supplemental Table 2. The stable association of EMR is significantly less than EIK for variant G0 and G1, but not G2. The differences in stable membrane association among the haplotypes are much less than the differences in activity; within each variant, the ratio of stable membrane association of EIK to EMR haplotypes is about 1.2, whereas the ratio of K permeability of EIK to EMR is about 2.

Within each of the haplotypes, the amount of the G1 and G2 variants stably associated with membranes is approximately double that of G0, paralleling the differences in K channel activity and matching our previous data for the KIK haplotype.⁵

Chloride permeability assays are shown in Figure 1B, identical to the potassium permeability assay except that protein and liposomes are mixed at pH 5, efflux step is performed at pH 5, and the ionophore is valinomycin. Chloride permeability conferred by each isoform was determined (Figure 1D and Supplemental Table 3). There are no significant differences among the haplotypes. Among the variants, G1 confers greater Cl permeability than G0 in KIK and EMR, and than G2, in all three haplotypes. The relative differences are small, with the activity of G1 ranging between 1.1- to 1.4-fold greater than G0 or G2. Stable membrane association under conditions that support Cl permeability (Figure 1E and Supplemental Table 3) show no pattern of significant differences among the haplotypes or among the variants.

Thus, we find that naturally occurring haplotype-encoded isoforms of ApoL1 have a significant effect on the cation channel activity without an effect on the chloride permease activity. The cation channel activity is greatest in the EIK (African) haplotype and is least in the EMR (Neanderthal) haplotype. The relative differences in cation channel activity among the wild-type and disease-associated variants is independent of haplotype: in all three haplotypes, the disease-associated variants have approximately double the cation channel activity of the wild type.

The effect of haplotype on cation channel activity of the variants parallels the previously reported effects of haplotype on cell toxicity,⁷ supporting a model in which cation channel activity itself is driving cellular injury that leads to kidney disease. In contrast, lack of correlation of anion permease activity with cell toxicity argues against a central role for the anion permease in ApoL1-associated cellular toxicity or disease. The parallel deviations of activity and stable membrane association among variants suggest activity is altered through the effects on membrane association/insertion rather than the intrinsic activity of membrane-inserted protein. In contrast, the lack of correlation between activity and stable membrane binding among haplotypes indicate these changes in activity are due to differences in intrinsic activity of membrane-inserted protein.

Methods and further observations are presented in the Supplemental Methods and in Supplemental Appendix 2.

Disclosures

P. Buchanan reports having employment with Allsup. J. Edwards reports having employment with Bayer Crop Science; and reports receiving research funding from Relypsa Inc. (Vifor Pharma). All remaining authors have nothing to disclose.

Funding

This work was funded by National Institutes of Health National Institute of Diabetes and Digestive and Kidney Disease grant 1R01DK120651 (J. Edwards).

Author Contributions

R. Winkler, J. Bruno and J. Edwards conceptualized the study and were responsible for the data curation; P. Buchanan and J. Edwards were responsible for the formal analysis; J. Edwards was responsible for the funding acquisition; P. Buchanan was responsible for the statistical methodology and software; J. Bruno, P. Buchanan, J. Edwards, and R. Winkler were responsible for the investigation; J. Bruno, J. Edwards, and R. Winkler were responsible for the methodology; J. Edwards provided supervision; J. Edwards wrote the original draft; and J. Bruno, P. Buchanan, J. Edwards, and R. Winkler reviewed and edited the manuscript.

Data Sharing Statement

All data used in this study are available in this article.

Supplemental Material

This article contains the following supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2022020213/-/DCSupplemental.

Supplemental Appendix 1. Haplotype and variant sequences used in this study.

Supplemental Appendix 2. Theory, performance, and interpretation of vesicle-based selective ion permeability assays.

Supplemental Table 1. Amino acids at positions of variance among the most common ApoL1 haplotypes and their frequencies in global, African, East Asian, and European populations.

Supplemental Table 2. Potassium permeability and stable membrane association under conditions supporting cation permeability.

Supplemental Table 3. Chloride permeability and stable membrane association under conditions supporting Cl permeability.

Supplemental Methods