Blocking the 5′ splice site of exon 4 by a morpholino oligomer triggers APOL1 protein isoform switch

www.nature.com/scientificreports OPEN Received: 7 December 2017 Accepted: 21 May 2018 Published: xx xx xxxx Blocking the 5′ splice site of exon 4 by a morpholino oligomer triggers APOL1 protein isoform switch Amber M. Cheatham1, Shamara E. Davis1, Atanu K. Khatua1 & Waldemar Popik1,2 APOL1 risk alleles G1 or G2 are associated with a kidney disease phenotype exclusively in people of recent African ancestry. Here we show that exon 4 encoding a part of the APOL1 signal peptide is constitutively spliced in major APOL1 transcripts expressed in kidney glomerular and tubular cells. We demonstrate that constitutive splicing of exon 4 results from a suboptimal hnRNP A1 binding motif found in exon 4. Accordingly, a robust binding of hnRNP A1 protein to a consensus hnRNP A1 cis-acting element in exon 4 results in almost complete exclusion of exon 4 from the APOL1 minigene transcripts. Blocking the 5′ splice site at the exon 4/intron boundary with a specific antisense morpholino oligonucleotide excludes exon 4 from the splicing pattern of endogenous APOL1 transcripts. These transcripts are fully functional and produce APOL1 protein isoform that is not normally detectable in podocytes. Together with our previous data showing no cytotoxicity of overexpressed APOL1 isoform lacking exon 4, we propose that morpholino-induced APOL1 isoform switch may provide a new tool to identify in vivo molecular mechanism(s) by which risk alleles promote or mediate the kidney disease phenotype. Increased risk of nondiabetic kidney disease in African Americans has been associated with apolipoprotein L1 (APOL1) gene coding variants G1 and G2, found exclusively in people of recent African ancestry 1,2. APOL1 variants G1 and G2 confer protection against Trypanosoma brucei rhodesiense and possibly other parasites, resulting in high frequencies in Sub Saharan populations3. The absence of the APOL1 gene in most primates except humans, gorillas, and baboons has been a major obstacle for testing causality between risk alleles G1 or G2 and development of kidney disease. Recently, however, a transgenic mouse model with a podocyte-specific inducible expression of APOL1 demonstrated that G1 or G2 alleles, but not the wild-type G0 allele, produced albuminuria and glomerulosclerosis, hallmarks of kidney disease4. This study confirmed that development of the kidney pathogenic phenotype depends on the expression levels of risk alleles, supporting previous findings that individuals with two risk alleles are at increased risk for kidney disease, as compared to carriers of only one1,2,5. The mechanism by which APOL1 risk alleles mediate cell injury seems to be complex and several mechanisms have been proposed including apoptosis, necrosis, pyroptosis, or autophagic cell death4,6–11. Cytotoxicity of APOL1 risk alleles was also shown to result from depletion of cellular potassium and activation of stress-activated protein kinases12, interference with endosomal vesicle trafficking13, or dysfunctional mitochondria14,15. Nevertheless, overexpression of a wild-type G0 variant of APOL1, which has not been associated with an increased risk of the kidney disease, displayed significant cytotoxicity not considerably different from that of G1 or G2 variants6. This suggests that APOL1 variants may be inherently toxic but only variant G0 is effectively attenuated in vivo. Alternatively, toxicity of risk variants can be potentiated by interaction with additional factors or second hits16. One such factor, a soluble urokinase plasminogen activator receptor (suPAR) associated with chronic kidney disease (CKD)17, was recently shown to synergize preferentially with G1 or G2 variants for activation of αvβ3 integrin signaling pathway, triggering podocyte injury18. Preventing formation of this cell membrane multiprotein complex by lowering expression of extracellular suPAR or inhibiting secretion of intracellular APOL1 could block αvβ3-mediated signaling and reduce the toxicity of APOL1 kidney risk variants. As reported by several databases including NCBI and Ensembl, the APOL1 gene is composed of seven exons that could potentially produce several alternatively spliced transcripts by selective inclusion of exons 2 and 418. 1 Meharry Medical College, Center for AIDS Health Disparities Research, Department of Microbiology and Immunology, 1005 D. B. Todd Blvd, Nashville, TN, 37028, USA. 2Department of Internal Medicine, 1005 D. B. Todd Blvd, Nashville, TN, 37028, USA. Correspondence and requests for materials should be addressed to W.P. (email: ) SCIentIfIC REPortS | (2018) 8:8739 | DOI:10.1038/s41598-018-27104-x 1 www.nature.com/scientificreports/ As an exon 4-encoded sequence is part of the APOL1 signal peptide, deletion of exon 4 may likely affect APOL1 secretion and alter its intracellular localization and function. Consequently, although the G1 and G2 mutations are in terminal exon 7, deletion of exon 4 may potentially affect biological activity of APOL1 risk variants. In this context, we have previously demonstrated that cytotoxicity of the APOL1 variant G0 was mitigated by the deletion of the exon 4-encoded sequence18. A process of alternative splicing is regulated by the interaction between cis-acting RNA elements and trans-acting splicing factors. Cis-acting elements include exonic splicing enhancers (ESEs) and intronic splicing enhancers (ISE) that recruit positive trans-acting factors, such as serine/arginine-rich (SR) proteins, which generally promote exon inclusion19–21. Exonic splicing silencers (ESSs) and intronic splicing silencers (ISS) recruit negative acting factors, such as heterogeneous nuclear ribonucleoproteins (hnRNPs)19. The cooperative or antagonistic interactions between splicing factors affect assembly of the spliceosome, a multi-subunit ribonucleoprotein complex that together with a large number of auxiliary proteins removes introns from pre-mRNA transcripts22,23. Using a splicing reporter minigene, we identified several potential ESEs and ESSs in exon 4 that, due to their suboptimal binding motifs, did not considerably affect exon 4 splicing, suggesting that exon 4 is constitutively spliced. This conclusion was further supported by RNA pull-down assays showing only a minimal binding of SRSF1 and hnRNP A1 to their cognate binding sites in exon 4. We also demonstrate that blocking the 5′ splice sites (5′ss) of exon 4 by an antisense morpholino oligonucleotide results in exclusion of exon 4 from endogenous APOL1 transcripts. In line with our previous observation indicating that exon 4 confers toxicity on APOL1 proteins24, we propose that morpholino-induced expression of alternative APOL1 protein isoform lacking exon 4 may provide a new tool to investigate molecular mechanism(s) by which risk alleles promote or mediate the kidney disease phenotype. Results Splicing pattern of the APOL1 transcripts and expression of APOL1 proteins in human kidney cells. The APOL1 pre-mRNA is composed of seven exons that can be (...truncated)