Loss of the WDR23 proteostasis impacts mitochondrial homeostasis in the mouse brain

Abstract

Mitochondrial adaptation is important for stress resistance throughout life. Here we show that WDR23 loss results in an enrichment for genes regulated by nuclear respiratory factor 1 (NRF1), which coordinates mitochondrial biogenesis and respiratory functions, and an increased steady state level of several nuclear coded mitochondrial resident proteins in the brain. Wdr23KO also increases the endogenous levels of insulin degrading enzyme (IDE) and the relaxin-3 peptide (RLN3), both of which have established roles in mediating mitochondrial metabolic and oxidative stress responses. Taken together, these studies reveal an important role for WDR23 as a component of the mitochondrial homeostat in the murine brain.

INTRODUCTION

Mitochondria play critical roles in maintaining cellular homeostasis, generating ATP, buffering calcium, but sometimes at the cost of generating reactive oxygen species (ROS) [1]. Mitochondria are sophisticated organelles that require the precise coordination of protein homeostasis with resident proteins transcribed, translated, and imported from the nucleus, but also proteins coded within the mitochondrial genome. To maintain a healthy lifespan, organisms must prevent the accumulation of protein aggregation by targeting misfolded proteins for proteasomal degradation or removal by autophagy [2]. The inability to maintain homeostatic function of the mitochondrial proteome can lead to organelle disruption and even loss of cellular functions [3].

Given its nickname as the powerhouse of the cell, the electron transport chain (ETC) is required for efficient oxidative phosphorylation (OXPHOS), supporting ATP synthesis en masse [4]. However, the ETC and complex I that passes electrons from NADH to ubiquinone is a major generator of ROS. Mitochondria contain a protective arsenal of enzymes to minimize ROS, including the capacity to dismutase superoxide anions to hydrogen peroxide (H₂O₂) and catalases which can further catabolize H₂O₂ to water and molecular oxygen with the assistance of glutathione antioxidant systems [5]. In the absence of these protective measures, OXPHOS deficiency promotes mitochondrial dysfunction and cell damage that drives several age-related pathologies (e.g., dementia, sarcopenia, and insulin resistance) [6-9].

WDR23, as a part of the cellular proteostasis machinery, regulates cytoprotection in metabolic tissues, but the functions of WDR23 in the brain are not documented. As part of the CUL4-DDB1 E3 ubiquitin ligase, WDR23 is an adapter for substrates that are targeted for degradation. Several specific substrates have been identified, including SKN-1/NFE2l2 [10], GEN1 [11], SLBP [12], and p21 [13]. Although a mitochondrial resident protein has yet to be established as a direct target of WDR23, several of the established direct targets play critical roles in mitochondrial homeostasis.

Mitochondria contain thousands of proteins of which only 13 are encoded by the mitochondrial genome with most mitochondrial resident proteins being nuclear-encoded, translated in the cytoplasm, and imported to their correct subcompartment within the mitochondria by translocases of the outer mitochondrial membrane (TOM) and translocases of the mitochondrial inner membrane (TIM) [14]. As such, mitochondrial biogenesis and maintenance requires a coordinated effort across multiple cellular compartments for proper function. Nuclear respiratory factor 1/2 (NRF1/2) transcribes nuclear genes that encode ETC proteins, transcription factor A (TFAM), and other proteins critical for mitochondrial biogenesis [15].

In this work we study the effects of whole animal genetic ablation of Wdr23 in the murine brain, which importantly alters steady state levels of mitochondrial proteins involved in OXPHOS and stress adaptation. These alterations could play critical roles in the homeostatic regulation of aging and aging-associated degradation of mitochondrial functions. Our work provides a new molecular player and an important biomarker of mitochondria function that could be modulated to ensure healthy aging.

MATERIALS AND METHODS

Animals

All animal protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the University of Southern California and all the procedures were conducted in compliance with institutional guidelines and protocols.

Wdr23 knock-out (Wdr23KO) mice were generated by Wellcome Trust Sanger Institute [16, 17]. Wdr23KO animals were subsequently backcrossed nine times into our C57BL/6J (WT) strain from the Jackson laboratory [10, 11]. Heterozygous (Wdr23 +/−) dams and sires were then mated to generate Wdr23+/+ and Wdr23−/− animals that were maintained as WT and KO, respectively. Male mice (n=4-6/group) were kept in a 12:12 h light-dark cycle, constant temperature, and humidity room. All animals were allowed ad libitum access to water and food.

Rodent diets

Mice were fed ad lib with irradiated D12450K rodent chow (Research Diets), containing 16.1kJ/g of digestible energy (animal-based protein 3.22kJ/g, carbohydrate 11.27kJ/g, fat 1.61 kJ/g).

RNA extraction and RNA sequencing

The hippocampus tissue of 44 weeks old mice was collected and lysed in RNeasy Lipid Tissue mini kit reagent (QIAGEN). RNA was extracted according to the manufacturer’s protocol. Qubit RNA BR Assay Kit was used to determine RNA concentration. Isolated RNA was sent to Novogene for library preparation and deep sequencing in biological triplicate. The read counts were used for differential expression (DE) analysis by using the R package DEseq2 (R version 3.5.2). Differentiated expressed genes were analyzed using p value <0.05 and fold change >1.5 as cutoff.

Western blot analysis

The hippocampus tissue of 8-14 weeks old mice was collected for WB analysis. Whole cell lysates were prepared in M-PER buffer (1x Mammalian Protein Extraction Reagent (Thermo Scientific), 0.1% Halt Protease & Phosphatase inhibitor (Thermo Scientific) according to the manufacturer’s protocol. Total protein concentrations were quantified by Bradford assay (Sigma). An equal amount of protein (20 μg) was separate on 4%-12% bis-tris polyacrylamide gel (Invitrogen,) in MOPS running buffer (Invitrogen) and then transferred to nitrocellulose membranes (GE Healthcare Life science). After blocking for 1 h with 3% BSA in PBST (PBS, 0.1% Tween 20), the membranes were subjected to immunoblot analysis. Antibodies used include: IDE (Millipore sigma, 1:5000), HO-1 (Abcam, 1:2000), TOMM20 (Abcam, 1:50000), AIF (Abcam, 1:1000), PGC1β (Abcam, 1:5000), TIMM17A/TIM17 (Abcam, 1:1000), VDAC1/porin+VDAC2 (Abcam, 1:5000), GCLC (Cell signaling, 1:1000), RXFP3 (Novus Biologicals, 1:1000), GSTP1 (Novus Biologicals, 1:1000), RLN3 (Novus Biologicals, 1:1000), β-actin (Millipore Sigma, 1:10000) and HRP-conjugated secondary antibodies (Thermo Fisher, 1:10,000). Specific protein bands were visualized and evaluated using FluorChem HD2 (ProteinSimple).

Statistical Analysis

All experiments were performed at least in triplicate. Data are presented as mean ± SEM. Data handling and statistical processing were performed using GraphPad Prism 8.0. Comparisons between two groups were done using unpaired Student’s t-test. Differences were considered significant at the p ≤ 0.05 level. *p<.05, **p<.01, ***p<.001, ****p<.0001, compared to C57BL/6J (WT) control.

RESULTS

Wdr23KO animals increase the expression of NRF1-regulated transcripts in the hippocampus

The brain is under energetic constraints that must continuously balance basal cellular function, repair damage, and monitor molecular turnover throughout the lifespan. We first measured the transcriptional response to loss of Wdr23 by performing RNA-sequencing on dissected hippocampus from Wdr23KO mice and compared to samples from age-matched WT animals. 46 genes displayed a significant difference in expression (Table 1 and Figure S1A). Among the classes of genes most highly upregulated include proteins with roles in neuronal function (e.g., synuclein, synaptogamin, potassium voltage-gated channels), cellular signaling (e.g., G protein coupled receptors, nicotinic cholinergic receptors), and metabolism (e.g., phospholipase C, aldehyde dehydrogenase family 1, prostaglandin synthase).

Table 1.

RNAseq of hippocampus (include annotation of NRF binding sites [18-20].

log2 FoldChange	padj	gene_name	NRF1	Description
−1.7072027	0.03524594	6530403H02Rik		lncRNA gene
−1.3668744	6.53E-48	Dcaf11	X	DDB1 and CUL4 associated factor 11
−0.8368631	0.03814108	Nkd2	X	naked cuticle 2
−0.4551116	0.03814108	Dgkb		diacylglycerol kinase, beta
−0.4442013	0.03814108	Tomm34	X	translocase of outer mitochondrial membrane 34
−0.2877003	0.03797079	Cpe		carboxypeptidase E
0.51703545	0.03814108	Selenon	X	selenoprotein N
0.56451625	0.04018394	Relt	X	RELT tumor necrosis factor receptor
0.59936546	0.00684726	H2-D1		histocompatibility 2, D region locus 1
0.60724433	0.03814108	Ier5l		immediate early response 5-like
0.67891576	0.00596244	Necab3		N-terminal EF-hand calcium binding protein 3
0.73693371	0.03417076	Epb41		erythrocyte membrane protein band 4.1
0.7453428	0.04302323	Myoc	X	myocilin
0.74586031	0.04302323	Rgs16	X	regulator of G-protein signaling 16
0.76312796	0.04347063	Klhl33		kelch-like 33
0.77229776	0.04302323	Hr		demethylase and nuclear receptor corepressor
0.77549126	0.04302323	Gm42756		lncRNA gene
0.83948279	0.04449878	Gpr3		G-protein coupled receptor 3
0.86185001	0.04560792	Coro6		coronin 6
0.87967952	0.03187854	H2-T23		histocompatibility 2, T region locus 23
0.95498895	0.03417076	Kcnab3	X	potassium voltage-gated channel
0.98129538	0.03417076	Plcb4		phospholipase C, beta 4
1.05204779	0.00209247	Lrrc55		leucine rich repeat containing 55
1.05445468	0.01057753	mon		adenylate cyclase activating polypeptide 1
1.06106885	0.04120696	Syt9	X	synaptotagmin IX
1.16137096	0.04018394	Ptgds		prostaglandin D2 synthase (brain)
1.1752958	0.01666918	Thbs2		thrombospondin 2
1.18602288	0.00016837	Cpne9		copine family member IX
1.25753393	6.85E-05	Ogn		osteoglycin
1.42302222	0.00011283	Syt6		synaptotagmin VI
1.45521832	0.00596244	Aldh1a2	X	aldehyde dehydrogenase family 1, subfamily A2
1.47753245	0.02033239	Sema3f	X	secreted, (semaphorin) 3F
1.50140755	0.00771786	Synpo2		synaptopodin 2
1.51850694	0.01057753	Fmod		fibromodulin
1.53999291	0.00255224	Sncg		synuclein, gamma
1.54251685	0.03814108	Ret	X	ret proto-oncogene
1.74768711	0.0111357	Rnf17		ring finger protein 17
1.88080386	0.04018394	Chrna3	X	cholinergic receptor, nicotinic, alpha peptide 3
1.93834849	0.01057753	Fst		follistatin
1.96654318	0.03417076	Gpr139	X	G protein-coupled receptor 139
2.04025612	0.02289672	Chrnb3		cholinergic receptor, nicotinic, beta polypeptide 3
2.05172614	0.01057753	Irx1	X	Iroquois homeobox 1
2.44897092	0.00596244	Krt80		keratin 80
3.73482426	0.04302323	Gng14		G protein subunit gamma 14 (Gm5741)
5.51020051	0.01566976	Gstp2	X	glutathione S-transferase, pi 2
5.79164125	0.03417076	Rln3	X	relaxin 3

We also noted a significant reduction in the expression of translocase of the outer mitochondrial membrane 34 (Tomm34) protein, which plays a role in the import of nuclear-coded mitochondrial proteins across the mitochondrial outer membrane. Moreover, 37% of the dysregulated genes have binding sites in the promoter region of the gene for nuclear respiratory factor 1 (NRF1) [18-20]; a key mediator of genes involved in oxidative phosphorylation (OXPHOS) and mitochondrial biogenesis [21, 22]. Taken together, the transcriptional signature suggests a change in mitochondria in the brain of animals lacking Wdr23.

Loss of Wdr23 results in OXPHOS complex subunit imbalance in the hippocampus

The functions of the ETC require the precise coordination of protein import and assembly of nuclear and mitochondrial subunits of each complex, and imbalances in this system can lead to defects that drive the generation of reactive oxygen species (ROS) and cellular dysfunction. Based on the established role that NRF1 plays in regulating oxidative phosphorylation, we first looked for changes in the mitochondrial ETC (Figure 1A). We measured the abundance of the five ETC complexes by Western blot with antibodies that specifically recognize NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 8 (NDUFB8) in complex I, succinate dehydrogenase [ubiquinone] iron-sulfur subunit (SDHB) in complex II, cytochrome b-c1 complex subunit 2, mitochondrial (UQCRC2) in complex III, mitochondrially encoded cytochrome c oxidase I (MTCO1) in complex IV, and ATP synthase F1 subunit alpha (ATP5A) in complex V (Figure 1B-G). Rather than a change in all ETC complex components, we measured a significant increase in the subunits for complex I (NDUFB8, Figure 1C) and complex III (UQRC2, Figure 1E), but not complex II (SDHB, Figure 1D), complex IV (MTCO1, Figure 1F), or complex V (ATP5A, Figure 1G) in hippocampus of Wdr23KO animals. In contrast, only complex III (UQRC2, Figure S1) was significantly increased in cortex of Wdr23KO animals, albeit the level of variation for all OXPHOS complex subunits was higher than that observed in the hippocampus.

Figure 1. Mitochondrial OXPHOS complex I and III are increased in the hippocampus of Wdr23KO mice.

(A) Cartoon model of ETC complexes. (B) Western blot analysis of total protein from hippocampal homogenates probed for subunits of the mitochondrial electron transport chain complexes I-V and quantification of differences between WT and Wdr23KO samples, normalized against β-actin, for (C) NDUFB8 of complex I, (D) SDHB of complex II, (E) UQCRC2 of complex III, (F) MTCO1 of complex IV, and (G) ATP5A of complex V.

Wdr23KO displays increased expression of stress signaling proteins

We next examined the steady state level of several stress response pathways in homogenates of dissected hippocampi and cortex from WT and Wdr23KO animals (Figure 2A). Our previous studies of liver tissue from the Wdr23KO mouse revealed activation of NFE2l2/NRF2 pathway due to the negative regulation of this transcription factor by the Cul4-WDR23 ligase and the ubiquitin proteasome system [10, 11]. Unsurprisingly, we noted an increase of log2fold change of 5.5 in the transcript levels of the established NFE2l2/NRF2 target gene glutathione s-transferase 2 (Gstp2) (Table 1 and Figure S1A). We confirmed a significant increase in the steady state level of GSTP proteins with an antibody that recognizes all three orthologous GSTP isoforms in the hippocampus (Figure 2B) but not in the cortex (Figure S2) of Wdr23KO animals. We next measured the expression of mitochondrial apoptosis inducing factor (AIF) that helps mediate the biogenesis and maintenance of complex I [23]. We noted a modest, but significant increase in the steady state levels of AIF (Figure 2C) and the voltage-dependent anion channel (VDAC1) of the outer mitochondrial membrane (Figure 2D) in hippocampus, but not in the cortex of Wdr23KO animals (Figure S2). Although we measured a clear enrichment for NRF1 targets, we could not detect a difference in the abundance of the upstream regulator of NRF1 activation, PGC1α/β [24] in both the hippocampus (Figure 2E) and the cortex of Wdr23KO animals (Figure S2). Lastly, we measured the steady state level of the TOMM20, a member of the translocase of the mitochondrial outer membrane, which was significantly reduced (Figure 2F), similar to transcript levels of the TOMM complex subunit Tomm34 (Table 1) in hippocampus of Wdr23KO animals. Interestingly, TOMM20 protein levels were increase in the cortex of Wdr23KO animals (Figure S2). Taken together, these results demonstrate that in addition to changes in the abundance of distinct OXPHOS complex proteins (Figure 1), several mitochondrial resident proteins with roles in stress adaptation are altered in the hippocampus, but not equally responsive in the cortex of Wdr23KO mice.

Figure 2. Loss of Wdr23 proteostasis increases mitochondrial stress factors in hippocampus.

(A) Western blot analysis of total protein from hippocampal homogenates probed for proteins with roles in stress adaptation and homeostasis. Differences between WT and Wdr23KO samples were quantified and normalized against β-actin, for (B) GSTP1, (C) AIF, (D) VDAC1/2, (E) PGC1β, and (F) TOMM20. Comparisons analyzed by t-test; *p<0.05, **p<0.01.

In addition to the oxidative stress pathway genes that are activated in Wdr23KO mice [10, 11], our recent work revealed an increase in the expression of the insulin degrading enzyme (IDE) in the liver of animals lacking WDR23 proteostasis [25]; noting that IDE can localize to the mitochondria [26, 27]. Surprisingly, the most highly expressed gene in the hippocampus of Wdr23KO mice, relative to WT, was the relaxin-3 locus, which was upregulated log2fold change of 5.7 (Table. 1), which has established roles in cellular responses to oxidative stress [28]. As such, we measured the steady state abundance of these two proteins in our hippocampus homogenates from Wdr23KO and WT animals (Figure 3A). We noted a ~3.5-fold increase in the abundance of IDE (Figure 3B) and a >1.5-fold increase in RLN3 (Figure 3C). Notably, we could not detect a change in the steady state level of the Relaxin Family Peptide Receptor 3 (RXFP3, Figure 3D). In the cortex of Wdr23KO animals, we noted a similar increase in IDE but not RLN3 (Figure 3). Taken together, these data suggest that in addition to changes in the abundance of several essential mitochondrial resident proteins, signaling molecules and metabolic regulators are significantly altered in the hippocampus in response to loss of Wdr23.

Figure 3. Wdr23KO mice display increased IDE and RLN3 expression in the hippocampus.

(A) Western blot analysis of total protein from hippocampal homogenates probed for proteins with roles in metabolic signaling. Differences between WT and Wdr23KO samples were quantified and normalized against β-actin for (B) IDE, (C) RLN3, and (D) RXFP3. Comparisons analyzed by t-test; **p<0.01.

DISCUSSION

Loss of coordination between the subunits within each complex of the electron transport chain and across the complexes themselves can lead to reduced mitochondrial function (e.g., reduced ATP and increased ROS) across model systems [29-33]. To correct this imbalance, the mitochondria and nucleus utilize sophisticated signaling pathways to communicate, which if unsuccessful, can lead to additive effects of age-associated damage to the organelle and the cell. Our understanding of the cause and consequence of loss of proteostasis and aging is complicated by the fact that endogenous and exogenous stressors accumulate with age and can promote the aggregation of misfolded proteins when cellular chaperones are unable to correct the defect which drives age-related pathologies. Moreover, with age, the activity of these chaperones is reduced resulting in the further imbalance of proteostasis. Understanding the homeostatic mechanisms that protect mitochondrial health through the integration of nuclear and mitochondrial signaling pathways to support metabolic plasticity and stress responses in the nervous system is of great importance. We identify Wdr23, a member of the Cul4-DDB1 E3 ubiquitin ligase complex, as a new molecule with the capacity to alter the steady state abundance of mitochondrial resident proteins in the murine hippocampus. The identification of new biomarkers that maintain homeostasis is of great importance and provides new molecular approaches for genetic, molecular, and pharmacological manipulation of physiological systems.

In this study we restricted out analysis to male mice as estrogen has been demonstrated to activate cytoprotective transcription factors [34]. We identified an enrichment for transcripts that are likely regulated by the transcription factor nuclear respiratory factor 1 (NRF1), which has well-documental roles in regulating the expression of genes that influence mitochondrial function. In addition to NRF1, we identify three proteins that are upregulated in the absence of Wdr23 that each have previously established roles in regulating the complexes of the mitochondrial OXPHOS system. AIF is essential for oxidative phosphorylation and has been linked to the stabilization of complex I and III, both of which display increased abundance in Wdr23KO brain [35]. AIF itself can also translocate to the cytoplasm and eventually the nucleus where it can regulate cell death [23, 35], however the abundance of AIF was only modestly increased in Wdr23 hippocampi and perhaps not to the level necessary to drive apoptosis. This finding is correlated with the increased abundance of VDAC1/2 in hippocampal homogenates from Wdr23KO animals. The accumulation of VDAC1 can increase the mitochondrial permeability transition pore (mPTP) opening [36, 37] under mitochondrial stress conditions, but whether the increase in VDAC1 is a response or result of Wdr23 loss remains to be elucidated.

Second, IDE, which can localize to the mitochondria [26], and beyond its role in regulating the levels of insulin, can interact with multiple mitochondrial ribosomal proteins in addition to proteins that facilitate the synthesis and assembly of mitochondrial complex I and IV [27]. Both complex I and IV are increased in the Wdr23KO brain albeit the latter failed to reach significance, but the trend for increased MTCO1 of complex IV was observed. Loss of complex I integrity, and the resulting imbalance of metabolism and ROS production has been implicated in metabolic diseases including diabetes and neurological conditions such as mood disorders and Parkinson’s disease [6].

Third, the most differentially expressed transcript in the hippocampus of Wdr23KO mice, as compared to WT, was Rln3, which encodes an insulin-like peptide linked to age-related neurological conditions including memory dysfunction, appetite control, and depression [28]. In addition, relaxin peptides have been used pharmacologically to provide protection from ischemia, which is thought to occur through the activation of mitochondrial respiratory pathways [38]. Rodent and non-human primate models suggest relaxin-3 signals through the G-protein coupled receptor, RXFP3, which has a broad range of effects on neuroendocrine functions associated with stress responses [39]. However, unlike the increase measured in RLN3 signaling peptide, we did not observe a similar increase in RXFP3, suggesting the loss of Wdr23 genetically mimics the pharmacological use of RLN3 in vivo.

Taken together our study documents a change in the abundance of several mitochondrial resident proteins in the hippocampus of animals harboring a genetic mutation in Wdr23. Moreover, the loss of Wdr23 results in significant increases in AIF, IDE, and RLN3, all of which have established roles in regulating mitochondrial ETC complexes and OXPHOS activity. Intriguingly, the changes in the steady state abundance of mitochondria-related proteins in the Wdr23KO mouse brain were differentially altered between the hippocampus and the cortex, which suggests a specificity of the response. Future work to define the potential influence of sex and the age-related impact of loss of Wdr23 on these molecules will be of great interest.

HIGHLIGHTS.

• Loss of Wdr23 increases several nuclear-coded mitochondrial resident proteins.

• Promoters of transcripts dysregulated in the hippocampus of Wdr23KO mice are enriched for NRF1 regulatory sequences.

• Insulin degrading enzyme (IDE) expression, which can localize to the mitochondria, is increased in the brain tissues lacking WDR23.

• Wdr23KO animals have increased expression of relaxin-3 (RLN3) peptide, but not RLFP3 receptor.

ABBREVIATIONS

ETC

electron transport chain

IDE

Insulin degrading enzyme

NRF

nuclear respiratory factor

OXPHOS

Oxidative phosphorylation

ROS

Reactive Oxygen Species

TOM

translocase of the outer membrane