Upregulation of COX4-2 via HIF-1α in Mitochondrial COX4-1 Deficiency

Abstract

Cytochrome-c-oxidase (COX) subunit 4 (COX4) plays important roles in the function, assembly and regulation of COX (mitochondrial respiratory complex 4), the terminal electron acceptor of the oxidative phosphorylation (OXPHOS) system. The principal COX4 isoform, COX4-1, is expressed in all tissues, whereas COX4-2 is mainly expressed in the lungs, or under hypoxia and other stress conditions. We have previously described a patient with a COX4-1 defect with a relatively mild presentation compared to other primary COX deficiencies, and hypothesized that this could be the result of a compensatory upregulation of COX4-2. To this end, COX4-1 was downregulated by shRNAs in human foreskin fibroblasts (HFF) and compared to the patient's cells. COX4-1, COX4-2 and HIF-1α were detected by immunocytochemistry. The mRNA transcripts of both COX4 isoforms and HIF-1 target genes were quantified by RT-qPCR. COX activity and OXPHOS function were measured by enzymatic and oxygen consumption assays, respectively. Pathways were analyzed by CEL-Seq2 and by RT-qPCR. We demonstrated elevated COX4-2 levels in the COX4-1-deficient cells, with a concomitant HIF-1α stabilization, nuclear localization and upregulation of the hypoxia and glycolysis pathways. We suggest that COX4-2 and HIF-1α are upregulated also in normoxia as a compensatory mechanism in COX4-1 deficiency.

Keywords: COX4-1; COX4-2; HIF-1α; cytochrome c oxidase; mitochondria.

Figure 1

Decreased COX4I1 expression, COX activity and mitochondrial respiration with elevated glycolysis trend in COX4-1-deficient cells. mRNA expression levels of COX4 isoforms were measured by RT-qPCR in the knockdown vs. control cell lines (COX4-1 knockdown was obtained using a stable expression of COX4I1-targeting or nonmammalian-targeting shRNAs, respectively) (A) and in patient’s and healthy control fibroblasts (B). The results reveal decreased levels of COX4I1 mRNA expression and reciprocally elevated levels of COX4I2 mRNA expression in both COX4-1-deficient cells. Enzymatic activity of COX determined by spectrophotometry; shows decreased activity (C,D) in both shCOX4I1 and patient’s cells, relative to corresponding controls. Oxygen consumption rate (OCR) and extracellular acidification rate; extracellular acidification rates (ECARs) were measured in shCOX4I1 (E,G,I) and patient cells (F,H,J) and in corresponding controls, with subsequent addition of oligomycin, FCCP and antimycin/rotenone. Basal, ATP-linked and maximal OCR were calculated (G,H), and OCR was plotted against ECAR to construct energy maps (I,J). All values were normalized to cell content measured by methylene blue (A620). OCR was reduced, relative to the corresponding controls. Both COX4-1-deficient cell lines showed a tendency towards more glycolysis than the corresponding controls. Values are presented as normalized mean ± SEM; * p < 0.05 and ** p < 0.01 compared to corresponding controls.

Figure 2

Decreased COX4-1 with a reciprocal elevation of COX4-2 in COX4-1-deficient cells. HFF-shCOX4I1, patient’s and corresponding control cells were incubated with MitoTracker red, fixed and stained separately with antibodies against COX4-1 and COX4-2. The results displayed in A and B demonstrate markedly decreased COX4-1 levels in both COX4-1-deficient cells (HFF-shCOX4I1 and patient’s cells), while COX4-2 staining was increased, relative to respective controls. Nuclei were visualized with Hoechst-33342 (A). The micrographs were quantified and are depicted as histograms of signal intensity per cell ± SEM of at least 100 cells * p < 0.05; ** p < 0.01 (B).

Figure 3

CEL-Seq2 analysis identified hypoxia as one of the top upregulated pathways in COX4-1-deficient cells. Total RNA was isolated from both COX4-1-deficient cells (HFF-shCOX4I1 and HFF) and from both types of control cells (HFF-CV and healthy control fibroblasts). A HiSeq assay was performed at the Technion Genome Center using the CEL-Seq2 method. Differential expression analysis was performed with the DESeq2 package. Significance threshold was set as FDR < 0.1. In order to identify biological functions that were expected to be influenced (either to increase/decrease) given the observed gene expression changes (between HFF-shCOX4I1 and HFF fibroblasts), pathway analysis was subsequently performed. Upregulation of hypoxia (left) and glycolysis (right) detected by gene set enrichment analysis (GSEA). NES: normalized enrichment signal; FDR: false discovery rate (A). In the volcano plot, each dot represents a gene (B). The x-axis indicates the log2 (fold change) of the expression of HFF-shCOX4I1 relative to healthy control fibroblasts, and the y-axis reflects –log10 of the FDR-adjusted p-value of this comparison. The colored dots pass the threshold for FDR. Selected HIF-1 target genes in the volcano plot (PDK1, GLUT1, HK1 and HK2) were validated by RT-qPCR. Values of RT-qPCR validation are presented as the log2 (fold change) in ± SD of three biological duplicates, * p < 0.05, ** p < 0.01 (C).

Figure 4

Increased levels of HIF-1α in COX4-1-deficient cells. Western blot analysis was performed on cell extracts from both COX4-1-deficient cells (patient and HFF-shCOX 1) and from both corresponding controls. The extracted cells were probed with anti-HIF-1α and anti-actin antibodies as loading control. As a positive control for HIF1-α, untreated (UT) human foreskin fibroblasts (HFF) were preincubated with cobalt chloride to simulate hypoxia (A). The histogram represents the results normalized to actin (B). The figure depicts one representative experiment out of three, showing an increased level of HIF-1α in COX4-1-deficient cells.

Figure 5

HIF-1α nuclear accumulation was present in COX4-2-positive cells. COX4-1-deficient (HFF-shCOX4I1 and patient) and control cells were stained with antibodies against HIF-1α (red) and COX4-2 (green). Nuclei were visualized by Hoechst-3334. The results demonstrate increased nuclear localization of HIF-1α in COX4-1-deficient cells relative to controls. The observed accumulation of HIF-1α was correlated with COX4-2-positive cells (upper panel). The micrographs were quantified and are depicted as histograms of HIF-1α (lower panel) relative signal intensity per nucleus ± SEM of at least 100 nuclei ** p < 0.01.

Figure 6

Chemical activation of HIF-1α increased COX4-2 expression in control cells, whereas in COX4-1-deficient cells, the levels remained unchanged. COX4-1-deficient (HFF-shCOX4I1 and patient) and control cells were preincubated either with or without cobalt for 6 h prior to performing co-staining of COX4-2 and HIF-1α (A). The stabilization of HIF-1α was demonstrated by its translocation to the nuclei of each treated cell. COX4-1-deficient cells did not show any difference in COX4-2 levels with and without cobalt treatment, while nuclear accumulation of HIF-1α was increased (A). The quantified results are depicted in the histogram (B). Values are normalized to the corresponding cobalt-treated controls. Mean ± SEM of at least 70 cells; * p < 0.05, ** p < 0.01, *** p < 0.001 compared to corresponding control; ns—not significant.