(C) PLOS One [1]. This unaltered content originally appeared in journals.plosone.org.
Licensed under Creative Commons Attribution (CC BY) license.
url:https://journals.plos.org/plosone/s/licenses-and-copyright

------------



FOXR1 regulates stress response pathways and is necessary for proper brain development

['Andressa Mota', 'Department Of Biology', 'Boston University', 'Boston', 'Massachusetts', 'United States Of America', 'Hannah K. Waxman', 'Rui Hong', 'Bioinformatics Program', 'Gavin D. Lagani']

Date: 2022-01 

The forkhead box (Fox) family of transcription factors are highly conserved and play essential roles in a wide range of cellular and developmental processes. We report an individual with severe neurological symptoms including postnatal microcephaly, progressive brain atrophy and global developmental delay associated with a de novo missense variant (M280L) in the FOXR1 gene. At the protein level, M280L impaired FOXR1 expression and induced a nuclear aggregate phenotype due to protein misfolding and proteolysis. RNAseq and pathway analysis showed that FOXR1 acts as a transcriptional activator and repressor with central roles in heat shock response, chaperone cofactor-dependent protein refolding and cellular response to stress pathways. Indeed, FOXR1 expression is increased in response to cellular stress, a process in which it directly controls HSPA6, HSPA1A and DHRS2 transcripts. The M280L mutant compromises FOXR1’s ability to respond to stress, in part due to impaired regulation of downstream target genes that are involved in the stress response pathway. Quantitative PCR of mouse embryo tissues show Foxr1 expression in the embryonic brain. Using CRISPR/Cas9 gene editing, we found that deletion of mouse Foxr1 leads to a severe survival deficit while surviving newborn Foxr1 knockout mice have reduced body weight. Further examination of newborn Foxr1 knockout brains revealed a decrease in cortical thickness and enlarged ventricles compared to littermate wild-type mice, suggesting that loss of Foxr1 leads to atypical brain development. Combined, these results suggest FOXR1 plays a role in cellular stress response pathways and is necessary for normal brain development.

Exome sequencing of an individual with severe neurological symptoms including postnatal microcephaly, progressive brain atrophy, and global developmental delay implicated a de novo missense variant in the FOXR1 gene as potentially causative. FOXR1 is a member of the forkhead box (FOX) family of transcription factors with unknown function. Overexpression of FOXR1 in cultured cells show diffuse nuclear localization, while the FOXR1 mutant led to an accumulation of nuclear aggregates due to protein misfolding. As a transcription factor, FOXR1 was found to regulate a large number of genes including those involved in protein folding pathways, while the mutant showed impaired regulation of stress-responsive genes. Although FOXR1 is expressed at low levels in most tissues, we detected Foxr1 expression in mouse embryonic brain tissue. Using CRISPR gene editing, deletion of the Foxr1 gene in mice led to reduced survival at birth. Brain pathology of Foxr1 knockout mice revealed decreased cortical thickness and an enlargement of ventricles. Our data reveal that FOXR1 regulates genes involved in proper protein folding and lack of Foxr1 in mice is associated with reduced survival and brain pathology consistent with observations found in the human brain.

Funding: This work was supported in part by the Intramural Research Program of the National Human Genome Research Institute (HG000215 to W.A.G.) and by a National Institutes of Health grant (R21GM114629 to U.B. and A.H.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.

Here, we report a human neurodevelopmental disorder associated with a rare variant in FOXR1. We demonstrate that the de novo missense M280L variant decreases FOXR1 protein expression and exhibits nuclear puncta aggregates in HEK293T cells, suggesting that impaired FOXR1 function can be pathogenic. In addition, we show that the FOXR1 M280L mutant has a compromised ability to respond to stress, in part due to impaired regulation of downstream target genes that are involved in the stress response pathway. Further, our analysis revealed Foxr1 knockout mice exhibit a severe survival deficit. Surviving newborn Foxr1 knockout mice show cortical thinning and enlarged ventricles suggesting that the architecture of the mammalian brain is dependent on Foxr1.

FOXR1, also known as FOXN5 (forkhead box N5) or DLNB13, is a 292 amino acid protein that contains a fkh DNA-binding domain [ 26 ]. The human FOXR1 and rat Foxr1 gene consist of six exons with conserved exon-intron structure, indicating that FOXR1 is well-conserved between human and rat genomes [ 27 ]. The Genome-based tissue expression consortium indicate that FOXR1 is expressed in the human brain and reproductive organs [ 28 ]. The Human Brain Transcriptome shows that FOXR1 is expressed in all brain regions during embryonic and postnatal development and its expression level in the brain is maintained throughout life ( https://hbatlas.org ). Furthermore, in situ hybridization showed that mouse Foxr1 expression was present in all brain regions and enhanced within cellular nuclei, consistent with the human tissue expression profile based on the Allen Brain Atlas [ 29 ]. However, little is known about the function of FOXR1. Several studies have shown that mouse Foxr1 is involved in spermiogenesis [ 30 ]. In addition, several point mutations within human FOXR1 have been shown to be associated with a variety of carcinomas, although functional characterization of these oncogenic FOXR1 mutants has not been performed [ 31 – 33 ]. Recently, Foxr1 was found to be an essential maternal–effect gene in zebrafish that is required for proper cell division and survival [ 34 ].

FOXR1 is a member of the evolutionarily conserved forkhead box (Fox) family of transcription factors named after the ectopic head structures observed in mutants of the Drosophila gene forkhead (fkh) [ 4 – 6 ]. Mutations in the Drosophila fkh gene cause defects in head fold involution during embryogenesis, resulting in a characteristic spiked head appearance in adult flies. Since the discovery of fkh, hundreds of Fox genes have been identified in organisms ranging from yeasts to humans, making it one of the largest but least explored families of higher eukaryotic transcription factors (reviewed in [ 7 – 8 ]). All members of the Fox gene family of transcription factors are monomeric, helix-turn-helix proteins that harbor a core fkh DNA-binding domain comprised of three α-helices connected via a small β-sheet to a pair of loops resembling butterfly wings or a “winged-helix” [ 9 – 11 ]. Despite the high degree of conservation identity in the DNA-binding domain, Fox proteins bind different target sequences with great specificity. Fox proteins affect transcriptional regulation of large array of genes directing major developmental processes such as cell proliferation and cell fate specification [ 9 , 12 – 14 ]. Human genetic analyses show several FOX genes have important biological functions associated with brain development; these include FOXG1 (potential determinant of forebrain size; [ 15 – 17 ]) and FOXP2 (vocal learning; [ 18 – 20 ]). Further, mutations in FOXG1, FOXC2, FOXL2, FOXP1 and FOXP2 have profound effects on human brain development including microcephaly, intellectual impairments, and language disorders [ 21 – 25 ].

Neurodevelopmental disorders result from abnormal brain development and the inability to reach cognitive, emotional, and motor developmental milestones. Progress in genomics has advanced the prognosis of human neurodevelopmental disorders and provided insights into the molecular mechanisms of disease [ 1 – 3 ]. While some causal genes are highly penetrant, there are also many rare single-nucleotide changes that have deleterious effects on genes of unknown function. Through exome sequencing, the NIH Undiagnosed Diseases Program (NIH UDP), a clinical site of the NIH Undiagnosed Diseases Network (UDN), identified a variant (M280L) in a single allele of the FOXR1 gene (forkhead box R1; NM_181721.2) in an individual with severe neurological symptoms including postnatal microcephaly, progressive brain atrophy, and global developmental delay.

Results

Exome sequencing identified an individual with developmental delay carrying a de novo missense variant in FOXR1 The NIH UDP identified a proband with severe neurological symptoms including postnatal microcephaly, progressive brain atrophy, and severe muscle hypotonia from early infancy. Brain MRI showed progressive hypoplasia in the cerebral cortex, pons and cerebellum and ventricular enlargement from age 1 to 5 compared to age-matched normal MRI brain scans (Fig 1A and 1B). The proband also exhibits growth delay, decreased body weight, short stature, scoliosis, hip dysplasia, ankle clonus, and bell-shaped thorax (S1 Table). Ophthalmic abnormalities include optic atrophy, cortical visual impairment, and retinitis pigmentosa. Neuromuscular abnormalities include hyperactive deep tendon reflexes, joint hypermobility, severe muscle hypotonia, and poor head control. In addition, the proband has myopathic facies, preauricular pits, anteverted nares and low set ears. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. De novo FOXR1 missense variant in a proband with microcephaly and brain atrophy. (A) MRI scans of mid-sagittal (top) and horizontal (bottom) view of normal age-matched and the proband at 1 year old. (B) MRI scans of mid-sagittal (top) and horizontal (bottom) view of normal age-matched and the proband at 5 years old. Arrow on mid-sagittal images indicate hypoplasia of the pons in the proband. Also, arrow on horizontal view show dilation of ventricle in the proband compared to age-matched normal individual. (C) Pedigree of the family where the letter P in red (black square) indicates the proband. (D) Sanger sequence analysis confirming the de novo FOXR1 variant. Sequence chromatograms demonstrate the presence of the heterozygous variant in the proband, II-4 (indicated by the red arrow) and the reference allele in both parents and siblings (green arrows). Letters on top indicate amino acid residues (Q = glutamine, C = cysteine, M = methionine, L = leucine, S = serine, P = proline). https://doi.org/10.1371/journal.pgen.1009854.g001 Exome sequencing was performed on the proband and the siblings and parents who are all unaffected. Three likely pathogenic candidate genes, rapamycin and FKBP12 target (RAFT1), ATPase Na+/K+ transporting subunit alpha 3 (ATP1A3), and FOXR1 were identified. RAFT1 functions as a kinase that regulates cell growth, proliferation, motility, and survival [35–36]. The proband has a homozygous RAFT1 missense variant, but the EXAC database identified an unaffected individual with the same RAFT1 variant. The second candidate, ATP1A3, maintains plasma membrane sodium and potassium gradients [37]. Investigations discovered an individual with the same variant who displays a mild phenotype involving learning disability and episodes of dizziness. Variants in ATP1A3 were considered to have contributed to the final phenotype and were returned to the family as a partial diagnosis (OMIM disorders 182350 and / or 128235). The last candidate is a de novo missense variant in FOXR1, a gene of unknown function, and the variant was not identified in the siblings or parents (Fig 1C). The heterozygous de novo nonsynonymous variant results in a methionine-to-leucine substitution at position 280 (M280L) and was confirmed by Sanger sequencing (Fig 1D). M280 is found in the C-terminal segment of the FOXR1 protein, which is downstream of the DNA-binding domain. M280 is highly evolutionarily conserved, from mammals, birds, reptiles to frogs and zebrafish (S1 Fig). In addition, the M280L variant is predicted to be damaging and disease-causing based on scores of Combined Annotation Dependent Depletion (score of 29.9 where a score of 30 means that the variant is in the top 0.1% of deleterious variants in the human genome), PolyPhen-2 (score: 0.994/1.0), and Mutation Taster (score: 0.99/1.0). Although a preliminary diagnosis implicating the ATP1A3 variant for this patient has been made, a synergistic contribution from additional variants including the FOXR1 M280L variant cannot be ruled out.

The FOXR1 M280L mutant leads to a decrease in FOXR1 protein expression To examine whether the FOXR1 M280L mutant was properly expressed in vitro, we transiently transfected FOXR1 wild-type (WT) or the M280L mutant in HEK293T or COS7 cells and immunoblotted for FOXR1 or GFP-tagged FOXR1 protein. FOXR1 levels were significantly decreased in the M280L mutant (Fig 2A, 2B and 2C). Since FOXR1 is a transcription factor, we next tested whether the M280L mutant affects FOXR1 nuclear localization in HEK293T cells transfected with either untagged or GFP-tagged FOXR1 WT or M280L. Western blot analysis demonstrated that both FOXR1 WT and M280L protein are localized in both cytoplasmic and nuclear fractions with higher levels found in the nuclear fraction (Fig 2D). However, protein levels of the M280L mutant was reduced compared to FOXR1 WT in both cytoplasmic and nuclear fractions. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. The M280L mutant destabilizes FOXR1 protein. (A) Representative immunoblots and quantitative analysis of FOXR1 from HEK293T cells transfected with pCMV-SPORT6 human FOXR1 WT or M280L mutant. GAPDH served as a loading control. Graph represents FOXR1 over GAPDH normalized to WT. Unpaired t-test (n = 4 independent experiments, ** p = 0.0025). (B) Representative immunoblot and quantitative analysis of FOXR1 from HEK293T cells transfected with GFP, GFP-tagged human FOXR1 WT or M280L mutant. GAPDH served as a loading control. Graph represents FOXR1 over GAPDH normalized to WT. Unpaired t-test (n = 5 independent experiments, *** p < 0.0001). (C) Representative immunoblot and quantitative analysis of FOXR1 from COS7 cells transfected with GFP, GFP-tagged human FOXR1 WT or M280L mutant. GAPDH served as a loading control. Graph represents FOXR1 over GAPDH normalized to WT. Unpaired t-test (n = 4 independent experiments, ** p = 0.0013). (D) Representative immunoblots and quantitative analysis of cytoplasmic (c) and nuclear (n) fractions of FOXR1 from HEK293T cells transfected with pCMV-SPORT6 or GFP-tagged human FOXR1 WT or M280L. GAPDH and Histone H3 served as cytoplasmic and nuclear loading markers, respectively. Graph represents FOXR1 over GAPDH normalized to WT. Unpaired t-test (n = 5 independent experiments, *** p < 0.0001). The percentages of total cellular FOXR1 in the cytoplasmic and nuclear fractions were determined. (E) Quantitative PCR (qPCR) to quantify FOXR1 mRNA levels from HEK293T cells transfected with GFP, GFP-tagged human FOXR1 WT or M280L mutant. Graph represents relative FOXR1 mRNA expression normalized to GFP. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments). (F) Representative immunoblot and quantitative analysis of FOXR1 from HEK293T cells transfected with GFP-tagged human FOXR1 WT or M280L mutant. Protein stability was monitored by quantitative immunoblotting after blocking with proteasome inhibitor MG132. Graph represents FOXR1 over GAPDH normalized to untreated WT. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, * p = 0.0245, ** p = 0.0003). (G) Representative immunoblot and quantitative analysis of FOXR1 from HEK293T cells transfected with GFP-tagged human FOXR1 WT, M280L mutant or FOXR1 C-terminal truncation mutant lacking the last 12 amino acids (Δ280–292). Protein stability was monitored for FOXR1 Δ280–292 mutant by blocking proteasome degradation with MG132. GAPDH served as a loading control. Graph represents FOXR1 over GAPDH normalized to untreated WT. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, *** p < 0.0001). https://doi.org/10.1371/journal.pgen.1009854.g002 We next investigated whether the decrease in FOXR1 levels in the M280L mutant was due to transcription or protein stability changes. In HEK293T-transfected cells, we detected equal amounts of FOXR1 mRNA levels of FOXR1 WT and M280L, indicating that decreased M280L protein levels are not due to decreased transcription (Fig 2E). To measure protein stability, we blocked the proteasome pathway by treating transfected HEK293T cells with MG132, a cell-permeable proteasome inhibitor. Protein levels of both FOXR1 WT and M280L were approximately the same after proteasome inhibition. This suggests that the M280L variant destabilizes the FOXR1 protein, likely due to protein misfolding which, make it susceptible to proteolysis and degradation through the proteasome pathway (Fig 2F). Finally, we investigated whether the short C-terminal tail containing M280 is necessary for protein stabilization. We generated a FOXR1 C-terminal truncation mutant lacking the last 12 amino acids from M280 (Δ280–292). Indeed, HEK293T cells transfected with GFP-tagged Δ280–292 have decreased FOXR1 protein levels, which increased following MG132 treatment, suggesting that the FOXR1 C-terminal tail is critical for FOXR1 protein stability (Fig 2G).

FOXR1 M280L induces a nuclear aggregate phenotype To examine whether the M280L mutant alters the cellular localization of FOXR1, we transfected HEK293T cells with GFP, GFP-tagged FOXR1 WT, M280L, or Δ280–292 mutant. Immunostaining for GFP shows FOXR1 WT mainly in a diffuse pattern in the nucleus, co-localizing with DAPI, a nuclear marker (Fig 3A). In contrast, about 13% of cells transfected with the M280L mutant form discrete nuclear puncta (Fig 3B and 3C). We observed a similar phenotype in COS7 cells transfected with the M280L variant (S2 Fig). In nuclei containing >15 puncta, the average size of individual puncta was <2 μm2, whereas nuclei containing <5 puncta had aggregates of >4 μm2 (Fig 3D). These results suggest that the larger puncta may form by coalescing from small nuclear foci. In addition, cells transfected with the FOXR1 Δ280–292 mutant displayed a similar nuclear puncta pattern, suggesting that the C-terminal tail of FOXR1 is necessary for proper folding of the protein. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. The M280L variant induces nuclear puncta phenotype. (A) Fluorescence images of HEK293T cells transfected with GFP or GFP-tagged human FOXR1 WT, M280L or Δ280–292 mutants. DAPI (blue) served as a nuclear marker. Scale bar = 20 μm. (B) Fluorescence images of HEK293T cells transfected with GFP-tagged M280L mutant showing a range of nuclear puncta phenotypes. Scale bar = 5 μm. (C) Quantitative analysis of the percentage of cells showing FOXR1 puncta phenotype. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, ** p = 0.0048, *** p = 0.0002). (D) Correlation analysis of the average size of the aggregate to the number of puncta per nucleus. https://doi.org/10.1371/journal.pgen.1009854.g003

Identification of novel FOXR1-dependent transcripts by RNA sequencing analysis To identify target genes regulated by FOXR1 and to investigate the effect of FOXR1 M280L, we performed an unbiased transcriptomic screen by RNA sequencing (RNAseq) in HEK293T cells transiently transfected with GFP, GFP-tagged FOXR1 WT or M280L. Principal component analysis showed that the three groups clustered separately excluding experimental covariates and batch effects (S3A Fig). We plotted a heat map of the log (-2) fold change for all the differentially-expressed genes (DEGs) and delineated five coherent clusters (Fig 4A). Differential gene expression analysis between GFP and FOXR1 WT transfected cells identified 2644 DEGs of which 1315 (49.7%) were upregulated and 1329 (50.3%) were downregulated transcripts (Figs 4A and S3B). To determine the effect of FOXR1 M280L, we compared WT and M280L, and identified 735 DEGs of which 561 (76.3%) were upregulated and 174 (23.7%) were downregulated (S3B Fig). We paid special attention to those transcripts whose levels showed a 2-fold increase in FOXR1 WT and a decrease in M280L as delineated in cluster E (Fig 4B). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. RNAseq analysis of FOXR1 wild-type and M280L mutant. (A) Heatmap of hierarchical clustering indicates differentially-expressed genes (rows) between GFP, GFP-tagged FOXR1 WT and M280L (fold-change > 2, p < 0.05). Red indicates up-regulated genes and blue indicates down-regulated genes. (B) Heatmap of gene cluster ‘E’ indicates differentially-expressed genes (rows) that are upregulated in FOXR1 WT and down-regulated in M280L compared to WT. (C) Distribution of gene ontology (GO) terms annotated in biological processes of highly-regulated genes in FOXR1 WT and down-regulated in M280L. (D) Heatmap of gene cluster ‘E’ highlighting several chaperone proteins that were differentially expressed in FOXR1 WT and down-regulated in M280L. (E) Volcano plots of differentially expressed genes between FOXR1 WT versus GFP control, M280L versus GFP and FOXR1 WT versus M280L. Significantly up-regulated genes are in red while down-regulated genes are in blue. Non-significant genes are in gray. (F) Quantitative real-time PCR verifying the RNAseq analysis showing FOXR1 drives expression of HSPA6, HSPA1A and DHRS2 and are misregulated in the M280L mutant. Graph represents relative expression. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, * p < 0.05, ** p < 0.005, *** p < 0.0001). https://doi.org/10.1371/journal.pgen.1009854.g004 Gene ontology (GO) analysis for biological processes within cluster E shows genes involved in the heat shock response. This cluster contains genes that are functionally-related to negative regulation of inclusion body assembly, chaperone cofactor-dependent protein refolding, de novo protein folding, cellular response to stress, and regulation of HSF1-mediated heat shock response where these are enriched in FOXR1 WT and downregulated in M280L (Figs 4C and S4). Based on the volcano plots that summarize both the expression fold-change and the statistical significance, the upregulated genes in response to FOXR1 WT and downregulated in M280L include HSPA1A and HSPA6 (both members of the Hsp70 family of heat shock proteins, Hsps), and DHRS2 (Dehydrogenase/Reductase SDR Family Member 2, a mitochondrial reductase enzyme) (Fig 4D and 4E). These proteins play roles in protecting against oxidative stress. In addition, when we examined the volcano plot between M280L relative to GFP, we found overlapping transcripts between M280L and GFP and between WT and GFP (Fig 4E). In fact, this was confirmed by a high Pearson’s correlation (r = 0.96) examining the log 2 (fold change) between WT with GFP and M280L with GFP suggesting the M280L mutation functions as a hypomorphic loss of function mutation due to reduced levels of the FOXR1 protein (S3C Fig). Quantitative real-time-PCR (qRT-PCR) supported the RNAseq data for HSPA6, HSAPA1A and DHRS2 (Fig 4F), confirming upregulation of gene expression in FOXR1 WT but not in the M280L mutant. Other Hsps such as SACS, DNAJC21 and DNAJC6 were increased in both FOXR1 WT and M280L groups. Not all members of the Hsp70 family were misregulated in the M280L mutant; for example, the HSPA12A transcript was found to be upregulated in both FOXR1 WT and the M280L mutant. These results indicate that FOXR1 drives expression of specific Hsps and an important NADPH-dependent reductase enzyme that is likely related to cytoprotective pathways alleviating oxidative stress. To determine whether the DEGs contain consensus sequences for FOXR1 response elements [14], we examined the promoter regions of DEGs for each cluster except for cluster C which comprised of only a few genes. Each cluster contains a subset of DEGs carrying the FOXR1 consensus element which may be direct targets of FOXR1, supporting FOXR1 playing a role as both a transcriptional activator and repressor (S3D Fig).

FOXR1 controls gene expression of heat shock chaperones and an antioxidant NADPH-dependent reductase To determine whether HSPA6, HSPA1A and DHRS2 are directly regulated by FOXR1, we manually performed a de novo motif analysis of target promoters to identify consensus DNA-binding sites upstream of the ATG start site (Figs 5A and S5). We found strong consensus sequences for FOXR1 response elements [14] within the promoter regions of at least three of the top FOXR1-regulated genes, HSPA6, HSPA1A and DHRS2 (Fig 5B). To determine whether FOXR1 regulates the expression of these three genes through interaction with their promoter sequences, we utilized a dual luciferase system under the control of proximal upstream regions of human HSPA6 (-1119 to -113 bp), HSPA1A (-1053 to -210 bp) or DHRS2 (-3329 to -2313 bp) and co-transfected with either GFP control, FOXR1 WT or M280L mutant in HEK293T cells. We found that HSPA6, HSPA1A, and DHRS2 are activated by FOXR1 WT but not by M280L, indicating that these promoter regions contain FOXR1 responsive sequences and are targets of FOXR1 WT (Fig 5C). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Human DNA binding-site motifs bound by FOXR1. (A) FOXR1 response elements showing consensus primary and secondary sequences bound by FOXR1 (adapted from [14]). (B) Putative FOXR1 response elements are denoted in the promoters of three of the top-regulated FOXR1-targeted genes: HSPA6, HSPA1A and DHRS2. (C) Dual luciferase reporter assays where GFP control, FOXR1 WT or M280L were co-transfected into HEK293T cells with the corresponding HSPA6, HSPA1A and DHRS2 luciferase reporters. Data are plotted as luciferase activity normalized to GFP control. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, * p < 0.05, *** p < 0.0002). (D) Consensus primary sequences bound by HSF1. The putative HSF1 response elements are denoted in the promoter of FOXR1. Dual luciferase reporter assays where GFP control or GFP-HSF1 were co-transfected in HEK293T cells with corresponding FOXR1 WT or Mut luciferase reporter. FOXR1 mutant (Mut) consists of the HSF1 response elements in FOXR1 where the two TT residues in FOXR1 WT are mutated to GG (underlined). Data was plotted as luciferase activity normalized to GFP control. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, ** p = 0.0062). https://doi.org/10.1371/journal.pgen.1009854.g005 Expression of many Hsps is known to be regulated by the transcription factor heat shock factor 1 (HSF1), which has a high affinity for cis-acting DNA sequence elements, including the heat shock elements (HSEs) found in the promoters of HSF-responsive genes such as Hsp70 proteins [reviewed in 38]. There is also precedence that HSF1 target genes extend beyond molecular chaperones. For example, in C. elegans, the protective effects of reduced insulin signaling requires both HSF1 and the FOXO transcription factor, DAF-16, to prevent damage by protein misfolding and to promote longevity [39–41]. Based on the GO analysis for biological processes, transcripts that were upregulated in FOXR1-transfected cells were genes related to regulation of HSF1-mediated heat shock response (S4 Fig). We therefore, tested whether HSF1 may regulate FOXR1 since we identified a consensus sequence for HSF1 binding within the promoter region of FOXR1 (Fig 5D). Utilizing a dual luciferase system under the control of an upstream region of human FOXR1 (-633 to +1 bp), FOXR1 was found to be activated by GFP-HSF1 (Fig 5D). However, HSF1-mediated FOXR1 activation was not observed when the HSF response element in FOXR1 was mutated from TTCTAGAA to GGCTAGAA (Mut) in vitro, indicating that human FOXR1 is a target of HSF1, which may be regulated by cellular stress.

FOXR1 expression is increased in response to cellular stress Because FOXR1 regulates expression of HSPA6 and HSPA1A transcripts and they are also direct targets of HSF1, we hypothesized that FOXR1 expression might be directly regulated following stress-induced paradigms. We induced cellular stress using two different paradigms: serum deprivation (metabolic stress for 24 hours) and CO 2 -deprivation (oxidative stress for 24 hours). Cells transfected with FOXR1 WT exhibited a 2.5- and 3.3-fold increase in FOXR1 protein levels under serum- and CO 2 -deprivation, respectively, when compared to the non-stressed condition (Fig 6A). The increase in FOXR1 protein levels coincided with an increase in nuclear FOXR1 (Fig 6B). In contrast, FOXR1 M280L protein levels also exhibited a 3.3-fold increase under CO 2 -deprivation but not during serum-deprivation, indicating that the M280L mutant may be sensitive to different types of environmental stressors. In fact, the number of nuclear aggregates in cells transfected with the M280L mutant in response to CO 2 -deprivation was increased but not in response to serum-deprivation (S1 Video). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. FOXR1 expression is increased in response to cellular stress. (A) Representative immunoblots and quantitative analysis for FOXR1 from HEK293T cells transfected with GFP, GFP-tagged FOXR1 WT or M280L mutant in response to serum and CO 2 deprivation. GAPDH served as loading control. Graph represents FOXR1 over GAPDH normalized to untreated WT. One-way ANOVA Tukey’s multiple comparisons test (n = 4 independent experiments, ** p < 0.0051). (B) Fluorescence images of HEK293T cells transfected with GFP-tagged human FOXR1 WT or M280L in response to serum and CO 2 deprivation. Scale bar = 20 μm. (C) Fluorescence images of HEK293T cells transfected with GFP-tagged human FOXR1 WT or M280L and treated with PMA, a NADPH oxidase activator known to enhance reactive oxygen species (ROS). Cells were fixed after 24 hours of treatment and assessed for ROS generation using CellROX, a photostable ROS sensor. Scale bar = 20 μm. Quantitative analysis of the percentage of cells expressing FOXR1 puncta phenotype. Unpaired t-test (n = 4 independent experiments, *** p < 0.0001). (D) Lactate dehydrogenase (LDH) levels from conditioned media of HEK293T cells following PMA treatment. Positive control is a set of cells treated with the lysis buffer. Data are expressed based on the absorbance reading at 490 nm normalized to positive control. One-way ANOVA Tukey’s multiple comparisons test (n = 3 independent experiments, *** p < 0.0001). (E) Representative immunoblots and quantitative analysis of HEK293T cells following PMA treatment showing an increase in FOXR1 expression. Graph represents FOXR1 over GAPDH normalized to untreated WT. One-way ANOVA Tukey’s multiple comparisons (n = 5 independent experiments, *** p < 0.0001). (F) Quantitative analysis of HSPA6, HSPA1A and DHRS2 protein levels from HEK293T cells transfected with GFP, GFP-tagged human FOXR1 WT or M280L. Graph represents protein of interest over GAPDH normalized to GFP. One-way ANOVA Tukey’s multiple comparisons (n = 3–5 independent experiments, * p < 0.05, ** p < 0.005, *** p < 0.0005). (G) Quantitative analysis of HSPA6, HSPA1A and DHRS2 protein levels from HEK293T cells transfected with GFP, GFP-tagged human FOXR1 WT or M280L and treated with PMA. Graph represents protein of interest over GAPDH normalized to GFP. One-way ANOVA Tukey’s multiple comparisons (n = 2–3 independent experiments, * p < 0.05). https://doi.org/10.1371/journal.pgen.1009854.g006 To further explore the relationship between FOXR1 and oxidative stress, we treated FOXR1-transfected HEK293T cells with phorbol 12-myristate 13-acetate (PMA), a pharmacologic NADPH oxidase activator known to enhance reactive oxygen species (ROS) through a protein kinase C-mediated pathway [42]. We assessed ROS generation by fluorescence imaging using CellROX, a photostable ROS sensor. Consistent with other stress paradigms, PMA enhanced ROS generation in HEK293T cells transfected with FOXR1 WT and M280L (Fig 6C). PMA enhanced the diffuse FOXR1 fluorescence in the nucleus of HEK293T cells transfected with FOXR1 WT. The number of nuclear aggregates in cells transfected with the M280L mutant was increased by 3.9-fold compared to non-PMA treatment (Fig 6C and S2 Video), suggesting ROS-induced aggregation of mutant FOXR1 protein in response to stress. To determine whether ROS-induced aggregation of FOXR1 protein is cytotoxic, we measured the amount of lactate dehydrogenase (LDH) released into the medium. While the PMA induced some ROS toxicity, we found no LDH changes between cells transfected with GFP alone and GFP-tagged FOXR1 WT or between FOXR1 WT and M280L, indicating that the nuclear aggregates were not cytotoxic (Fig 6D). We found FOXR1 protein levels were increased 2.3- and 1.8-fold in cells transfected with FOXR1 WT and M280L mutant after PMA treatment, respectively (Fig 6E). Concomitantly, we found an increase in both HSPA6 and DHRS2 protein levels in cells transfected with FOXR1 WT (Fig 6F and 6G). HSPA6 levels were increased in response to PMA treatment in the M280L mutant. In contrast, we did not observe any changes in DHRS2 protein expression levels in cells transfected with M280L regardless of PMA treatment. In addition, while we observed a significant increase in HSPA1A mRNA levels in cells transfected with FOXR1 WT (Fig 4), we did not detect any changes in HSPA1A protein levels in cells transfected with FOXR1 WT or M280L. However, we did consistently see a decrease in HSPA1A protein levels in cells transfected with M280L compared to FOXR1 WT, but this difference disappeared when cells were treated with PMA.

FOXR1 nuclear puncta in M280L mutant are insoluble To determine whether the nuclear puncta that form in HEK293T cells transfected with the M280L mutant were aggresomes, which are known to serve as storage bins for misfolded or aggregated proteins [43], transfected HEK293T cells were treated with PMA and stained with the Proteostat dye. The dye detects misfolded and aggregated proteins in cells. We found bright punctate staining for proteostat-positive aggregates colocalized with the nuclear puncta in cells expressing the M280L mutant but not in FOXR1 WT (Fig 7A). These results were similar in transfected cells expressing M280L that were treated with the cell-permeable proteasome inhibitor MG132, further supporting that the M280L variant destabilizes FOXR1 protein and forms nuclear aggregates (Fig 7B). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. M280L nuclear aggregates are insoluble misfolded proteins. (A) Fluorescence images of HEK293T cells transfected with GFP-tagged human FOXR1 WT or M280L and treated with PMA. Cells were fixed after 24 hours of treatment and immunolabeled with Proteostat marker. White square box in the middle panels indicate images presented in the bottom panel at higher magnification. Top and middle panels, scale bar = 20 μm. Bottom panels, scale bar = 10 μm. (B) Fluorescence images of HEK293T cells transfected with GFP-tagged M280L and treated with MG132. Cells were fixed after 24 hours of treatment and immunolabeled with Proteostat marker. Scale bar = 10 μm. (C) Time-lapse imaging of HEK293T cells transfected with GFP-tagged M280L. Top panel represents images showing nuclear aggregates undergoing extensive movements and fusions. Bottom panel illustrates schematic drawings of the fusion events. Scale bar = 5 μm. (D) FOXR1 was sequentially extracted with Tris-HCl, Triton X-100, Sarkosyl and SDS. Quantification shows that the amount of FOXR1 in the sarkosyl fraction was not significant (n.s.) between WT and M280L. However, the SDS fraction was significantly higher in the M280L mutant when compared to the overall Tris-HCl total fraction. One-way ANOVA Tukey’s multiple comparisons (n = 2 independent experiments, *** p = 0.0003). https://doi.org/10.1371/journal.pgen.1009854.g007 Misfolded proteins often expose their hydrophobic domains, leading to aggregation [44–45]. In addition, most aggregated proteins tend to coalesce and form large deposits such as aggresomes or inclusion bodies [46–47]. Previous studies have shown that nuclear and cytoplasmic aggregates of poly-Q proteins such as ataxin-1 are dynamic and exchange their components whereas ataxin-3 are immobile [48–49]. In fact, time-lapse live cell imaging of HEK293T cells transfected with GFP-tagged M280L showed that the nuclear aggregates are quite dynamic and undergo extensive movements and fusions, with small aggregates moving toward each other and fusing to form larger aggregates (Fig 7C and S3 Video). Another criterion of misfolded proteins deposited within aggresomes is that they are largely detergent insoluble [46,50–53]. Thus, we examined the biochemical properties of M280L aggregates versus FOXR1 WT, testing protein lysates from HEK293T cells transfected with GFP, GFP-tagged FOXR1 WT or M280L for their solubility in different detergents. Protein extracts were sequentially extracted by Tris-HCl buffer, Tris-HCl buffer containing 1% Triton-X100, 1% Sarkosyl, and finally by 2% SDS. The amount of FOXR1 extracted in each fraction was assessed by immunoblotting for GFP-FOXR1. GFP-FOXR1 WT was detected in Tris-HCl soluble, Sarkosyl soluble, and SDS soluble fractions but was not present in the Triton X-100 fraction, suggesting that the majority of the FOXR1 WT protein was soluble and, not associated with membrane-bound proteins (Fig 7D). However, the majority of M280L was detected in the SDS fraction and not in the Sarkosyl fraction indicating a significant portion of the protein was insoluble and aggregating, which is consistent with the increased aggregation shown by the Proteostat immunolabeling.

[END]

[1] Url: https://journals.plos.org/plosgenetics/article?id=10.1371/journal.pgen.1009854

(C) Plos One. "Accelerating the publication of peer-reviewed science."
Licensed under Creative Commons Attribution (CC BY 4.0)
URL: https://creativecommons.org/licenses/by/4.0/


via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/