Abstract
Unbalanced copper (Cu2+) homeostasis is associated with neurological development defects and diseases. However, the molecular mechanisms remain elusive. Here, central neural system (CNS) myelin defects and down-regulated expression of Wnt/Notch signaling and their down-stream mediator hoxb5b were observed in Cu2+ stressed zebrafish larvae. Loss/knockdown-of-function of hoxb5b phenocopied the myelin and axon defects observed in Cu2+ stressed embryos. Meanwhile, activation of Wnt/Notch signaling and ectopic expression of hoxb5b could rescue copper-induced myelin defects, suggesting Wnt&Notch-hoxb5b axis mediated Cu2+ induced myelin and axon defects. Additionally, whole genome DNA methylation sequencing unveiled that a novel gene fam168b, similar to pou3f1/2, exhibited significant promoter hypermethylation and reduced expression in Cu2+ stressed embryos. The hypermethylated locus in fam168b promoter acted pivotally in its transcription, and loss/knockdown of fam168b/pou3f1 also induced myelin defects. Moreover, this study unveiled that fam168b/pou3f1 and hoxb5b axis acted in a seesaw manner during fish embryogenesis, and demonstrated that copper induced the down-regulated expression of the Wnt&Notch-hoxb5b axis dependent of the function of copper transporter cox17, coupled with the promoter methylation of genes fam168b/pou3f1 and their subsequent down-regulated expression dependent of the function of another transporter atp7b, making joint contributions to myelin defects in embryos. Those data will shed some light on the linkage of unbalanced copper homeostasis with specific gene promoter methylation and signaling transduction as well as the resultant neurological development defects and diseases.
Author summary In this study, we first unveiled that copper induced central neural system (CNS) myelin defects via down-regulating Wnt/Notch-hoxb5b signaling, and parallel with hypermethylating promoters of genes fam168b/pou3f2 and their subsequent down-regulated expression. Additionally, we unveiled that fam168b/pou3f1 and hoxb5b axis acted in a seesaw manner during fish embryogenesis. Genetically, we unveiled that copper was trafficked to mitochondrion via cox17 then led to the down-regulation of Wnt&Notch-hoxb5b axis, and was trafficked to trans-Golgi network via atp7b to induce the hypermethylation and the down-regulated expression of pou3f1/fam168b genes, making joint contributions to myelin defects in embryos.
Introduction
Many neurological diseases with behavioral changes and neurological disorders are associated with the unbalanced copper homeostasis in human, such as Alzheimer’s disease (AD), Wilson’s disease (WD) and Wallerian degeneration (WD)(1, 2). Excess copper has been reported to damage nervous system and lead to the behavioral abnormalities in fish (3, 4). However, the detailed molecular characteristics and the potential mechanisms underlying copper-induced neural defects, especially axonal and myelin defects, remain unclear.
Axons transmit electrical impulses to neuron’s targets, which is an essential process for the establishment of the nervous system. Axonal damage has been shown to cause neurological disorders, such as stroke, traumatic brain/spinal cord injuries, multiple sclerosis (MS) and Wallerian degeneration (WD) (5, 6). Myelin sheaths are essential for the rapid and efficient propagation of action potentials as well as for the support for the integrity of axons in the vertebrate nervous system. In the central nervous system (CNS), oligodendrocytes spirally wrap axons in multilamellar plasma membrane and eventually compact to form the myelin sheaths (7, 8). The compacted myelin sheaths increase the resistance of axons and reduce their capacitance by several orders of magnitude(8, 9). The failure of compacted myelin formation leads to delayed or interrupted signal conduction, contributing to motor, sensory, and cognitive behavioral deficits(10, 11). Even a subtle defect of CNS myelin can cause a persistent cortical network dysfunction and induce neuropsychiatric disorders in mouse(12–14). Myelin disorder in human has been reported to associate with a series of neurodegenerative diseases such as MS, Menkes diseases (MD), Parkinson’s diseases (PD) and Huntington’s diseases (HD) (14, 15).
DNA methylation is implicated in many copper-induced disorders, including AD, WD, MS and MD (16, 17). In WD patients, accumulated copper dysregulates methylation status (17, 18). The methylation level of PAD2 is reduced in copper-accumulated MS brain, leading to changes of mbp expression (19, 20). Copper has been reported to upregulate the expression of DNA methylation- and stress-related genes in zebrafish (21, 22). However, few reports are available about copper-induced locus-specific DNA methylation during embryogenesis, and few studies have linked this methylation with its induced myelin developmental damages in vertebrates.
The normal trafficking of the copper is essential for normal cellular functions. In vertebrate, copper is delivered to different organelle by independent copper chaperones, such as cox17, atp7b, and etc(23).. Cox17 helps copper being delivered to mitochondrial cytochrome c oxidase, and Trans-Golgi network (TGN) resident atp7b helps copper being pumped to the circulation or to extracellular (24, 25).
In this study, zebrafish in vivo system was used to investigate the cellular and molecular mechanisms of copper-induced CNS myelin defects. Copper was revealed to induce myelin defects via down-regulating the expression of Wnt&Notch-hoxb5b axis. Additionally, the gene fam168b, similar to pou3f1 and pou3f2, showed hypermethylation in the promoter regions and reduced expressions in copper stressed embryos, which might mediate in a parallel pathway to Wnt&Notch-hoxb5b in copper-induced myelin defects. Moreover, cox17-/-and atp7b-/- mutants were used to verify the correlation of copper trafficking deficiency in specific organelle and the occurrence of developmental myelin defects in copper stressed embryos in this study.
Results
Cu2+ induces embryonic CNS myelin and axon defects in zebrafish
Transmission electron microscopy (TEM) detection revealed compacted myelin sheaths in the spinal cord in the control larvae at 5 dpf (Fig 1A1-A2), but significantly thinner (increased g-ratio) and uncompacted myelin in the spinal cord in Cu2+ stressed larvae (Figs 1A3-A5). Additionally, WISH assays detected significantly reduced expression of mbp,plp1a, and olig2 (Figs 1B, S1A, and S1B) in the spinal cord in Cu2+ stressed embryos at 96 hpf. qRT-PCR results further conformed the reduced expression of mbp and mpz in the Cu2+ stressed embryos at 96 hpf (Fig 1C). The Cu2+ effects on myelin and axon development were further tested by analyzing the fluorescence in Cu2+ stressed transgenic Tg(mbp:EGFP) and Tg(olig2:dsRED) embryos, with a significant down-regulation observed in their respective fluorescence at 96 and 48 hpf (Fig 1D). Furthermore, the length of axons was remarkably reduced in Cu2+ stressed embryos (Fig 1E).
CNS myelin and axon formation in hoxb5b loss- and gain-of-function embryos
Hoxb5b exhibited significantly and specifically reduced expression in Cu2+ stressed embryos (22), and was reported to function importantly in axon guidance in mouse(26). Thus, in this study, hoxb5b was assumed to be a potential mediator in Cu2+ induced myelin and axon defects. To validate the hypothesis, an anti-sense morpholino (hoxb5b-MO) and a hoxb5b null mutant with a 4-bp deletion in the first exon (hoxb5b-/-) (Fig 2A) were applied to test the hoxb5b roles in Cu2+ induced myelin and axon defects. Embryos injected with hoxb5b-MO at 48 hpf exhibited brain hypoplasia, eye hypoplasia, trunk abnormalities, and reduced body size (Fig 2B2), which phenocopied the defects observed in Cu2+ stressed embryos. However, hoxb5b-/- mutant embryos exhibited almost normal-like phenotype at 48 hpf (Figs 2B3, B4).
Additionally, compared with the WT larvae, hoxb5b morphants or hoxb5b-/- mutants exhibited significantly decreased expression in the CNS myelin markers mbp and plp1a at 96 hpf (Figs 2C, 2D S2A1, and S2A2), identical to that in Cu2+ stressed larvae. The olig2 promoter driven fluoresce was observed to be significantly down-regulated in hoxb5b-MO injected olig2:dsRED transgenic embryos, compared with that in the control embryos at 48 hpf (Fig 2E). Furthermore, the length of each axon was significantly reduced in hoxb5b morphants (Fig 2E6), which also phenocopied the defects observed in Cu2+ stressed embryos.
The expression of CNS myelin markers mbp and plp1a (Figs 2F, 2G and S2A3) and the olig2 promoter driven fluoresce in CNS and the length of axon (Fig 2H) were partially rescued in Cu2+ stressed embryos via ectopic expression of hoxb5b.
Activation of Wnt or Notch signaling rescues myelin and axon defects in Cu2+ stressed embryos v ia recovering hoxb5b expression
The microarray data showed that the expressions of Wnt and Notch signaling genes were reduced in Cu2+ stressed embryos (Figs 3A, S2B1 and Table S9), and qRT-PCR assays confirmed the down-regulated expressions of Wnt signaling(27) and Notch signaling genes (Fig S2B2) in Cu2+ stressed embryos. It has been reported that both Wnt and Notch signaling specified the oligodendrocyte fate (1, 28–30), and hoxb5b, is downstream of these two signaling pathways(31, 32). In this study, both Wnt agonist BIO and NICD notch3 mRNA partially rescued hoxb5b expression in Cu2+ stressed embryos separately (Figs 3B, 3C). WISH and qRT-PCR analysis exhibited the expression of mbp and plp1a was recovered to nearly normal level in Cu2+ stressed embryos co-exposed with BIO (Figs 3D, 3E). Additionally, BIO significantly recovered the down-regulated fluorescence level and the length of the fluorescent axon to nearly normal level in Cu2+ stressed olig2:dsRED transgenic embryos (Fig 3F).
WISH and qRT-PCR assays showed the expression of mbp and plp1a was significantly rescued via ectopic expression of NICD notch3 mRNA in Cu2+ stressed larvae (Figs 3G, 3H, and S2B3), and the fluorescence for the expression of olig2 and the length of fluorescent axon was partially rescued in the Cu2+ stressed olig2:dsRED transgenic embryos with ectopic expression of NICD mRNA (Fig 3I).
DNA methylation and transcriptional activity of myelin genes in Cu2+ stressed embryos
Under stress conditions, epigenetic DNA methylation has been reported to function importantly in disease process and intergenerational inheritance (33, 34). Thus, the whole genome methylation level in Cu2+ stressed larvae was examined to unveil the potential epigenetic mechanisms underlying Cu2+ induced myelin and axon defects. It has been unveiled that Cu2+ induced the expression of 26 hyper-methylated and 31 hypo-methylated genes in Cu2+ stressed larvae(35). Among them, genes pou3f1, pou3f2 and fam168b, which were associated with myelin and axon, were hypermethylated in the promoter domain (Figs 4A, S3A1 and S3B). Based on the microarray data reported previously(22), this study unveiled the down-regulated expression of genes pou3f1, pou3f2, and fam168a, the homolog gene of fam168b in Cu2+ stressed embryos (Fig S3A2), which was further confirmed by qRT-PCR analysis (Fig 4B1). Specifically, pou3f1 showed obviously decreased expression in the brain of Cu2+ stressed larvae (Figs 4B2-B4 and S3A3). Additionally, the expression of fam168b was also significantly reduced in Cu2+ stressed larvae from 24 hpf to 96 hpf (Figs 4B1 and C), and its promoter exhibited significant hypermethylation at 96 hpf (Figs 4D and S3C).
Loci in the fam168b promoter from -1672 to -1414, -1414 to -1240, and -1240 to -927 were obviously hypermethylated in Cu2+ stressed larvae (Fig 4A). Thus, we further investigated the roles of the aforementioned hypermethylated loci in regulating gene transcription. Different truncated promoter driven GFP fluoresces were almost distributed throughout the neural ectoderm in the injected embryos (Fig 4E), indicating their transcriptional activities during embryogenesis. Cu2+ slightly down-regulated the GFP fluoresce driven by different truncated promoters in embryos (Figs 4E2, E4, E6, E8 and S3D). Additionally, compared to GFP fluoresce driven by fam168b promoter from -1672 in the injected embryos (Figs 4E1, E2 and S3D), the GFP fluoresce from -1414 was obviously reduced (Figs 4E3, E4 and S3D), with a further reduction in the GFP fluoresce from -1240 (Figs 4E5, E6 and S3D) and -927 (Figs 4E7, E8 and S3D) truncated promoter driven GFP plasmids. The luciferase activity assays revealed that the gradient truncation of the fam168b promoter led to a gradual decrease of the transcriptional activity in the sequence of -927 promoter mutant < -1240 mutant <-1414 mutant < -1672 (Figs 4G and S3D). The schema of truncated fam168b promoter constructs was shown in Fig 4F.
CNS myelin formation in fam168a/fam168b loss- and gain-of-function embryos
The function of fam168a and fam168b during embryogenesis was further tested by knockdown and knockout of fam168a and fam168b in embryos. The transcripts of fam168a and fam168b were distributed ubiquitously among the whole embryo at the early stages (Fig S4A and S4B). Their predominant expression in the brain was observed at 96 hpf (Figs 5A1 and A2), similar to the expression pattern of pou3f1 in the embryos at this stage (Fig 5A3). The WT embryos injected with fam168a or fam168b MO exhibited similar developmental defects, such as shortened body, microcephalia, and slight ventralization at 24 hpf (Figs 5B2 and 5B3) and 96 hpf (Figs 5B6 and B7), similar to the developmental defects observed in pou3f1 morphants (Fig 5B4 and B8). Meanwhile, fam168a-/- mutant with a 4-bp deletion (Fig S4C1) exhibited a normal-like phenotype (Fig 5B10) and fam168b-/- mutant with 1-bp deletion (Fig S4C2) displayed microcephalia and slight ventralization at 96 hpf (Fig 5B11).
Transcriptional profiles in fam168a and fam168b morphants were investigated by KEGG pathway (Figs 5C1, S4D1 and Table S10, S11) and cellular component GO analyses (Figs 5D1, S4E1 and Table S12, S13). They showed enrichment in the nervous system and synapse for the down-regulated DEGs, identical to transcriptional profiles in pou3f1 morphants (Figs S4D2, S4E2, Table S14 and S15). Additionally, 85 genes in the nervous system (Fig 5C2) and 8 genes in synapse (Fig 5D2) were down-regulated and overlapped in the three fam168a, fam168b, and pou3f1 morphants. Meanwhile, 104 genes in the nervous system (Fig S4D3) and 8 genes in synapse (Fig S4E3) were down-regulated and overlapped in both fam168a and fam168b morphants.
In this study, CNS myelin and axon development in fam168a/b loss/knockdown-of-function embryos were further tested in term of mbp and plp1aexpression. Mbp and plp1a exhibited obviously reduced expression in both fam168a/b morphants and fam168a homozygous mutant by qRT-PCR and WISH assays (Figs 5E, 5F, S4F and S4G), similar to its expression in pou3f1 morphants (Figs 5E and 5F) and in Cu2+ stressed embryos.
Additionally, fam168a, fam168b, and pou3f1 mRNA were injected separately into Cu2+ stressed embryos to test whether they could rescue the myelin formation. Mbp expression was partially recovered in the Cu2+ stressed embryos injected separately with fam168a, fam168b, and pou3f1 mRNA (Fig S4H).
Wnt&Notch-hoxb5b signaling and fam168a/fam168b/pou3f1 transcriptional factors in embryogenesis
The crosstalk between Wnt&Notch-hoxb5b and fam168a/fam168b/pou3f1 transcriptional factors underlying Cu2+-induced myelin and axon developmental defects was explored separately by analysis of the expression of hypermethylated genes pou3f1, fam168a, and fam168b in hoxb5b morphants and hoxb5b-/- mutants and vice versa. Pou3f1, fam168a, and fam168b showed significantly increased expression in both hoxb5b morphants and hoxb5b-/- mutants at 96 hpf (Figs 6A, 6B and S5A). So did hoxb5b in pou3f1, fam168a, and fam168b morphants (Figs 6C, 6D and S5B).
Furthermore, we detected the combined effects of down-regulation of the two signaling pathways on the embryonic development and mbp expression. Morphants injected with the combined MOs of hoxb5b, pou3f1, fam168a and fam168b exhibited similar phenotypic defects (Fig 6E) and obviously reduced expression of CNS myelin marker mbp (Figs 6F and S5C). Meanwhile, hoxb5b+/-fam168a+/- mutants exhibited normal-like phenotype at 96 hpf (Fig 6G), but an obviously reduced expression of CNS myelin marker mbp (Figs 6H and S5D).
CNS myelin and axon formation in copper stressed cox17-/-, atp7b-/-, and atp7a-/- mutants
The question of in which organelle Cu2+overload resulted in the changed expression of the down-stream signaling and the CNS myelin and axon defects was investigated by using cox17-/- (Fig 7A1) and atp7b-/- (Fig 7A2) null mutants. Cox17-/- and atp7b-/- null mutants exhibited normal-like phenotypes at 96 hpf (Figs 7B). However, RNA-seq analysis revealed that genes in the nervous system (Fig S6A1), synapse (Fig S6A2), and axon (Fig S6A3) exhibited reduced expression in cox17-/- mutants.
Furthermore, the expression of the CNS myelin markers mbp, plp1a and genes pou3f1, and fam168a&fam168b was tested in cox17-/- or atp7b-/- mutants with and without Cu2+ stimulation. When compared with the WT control, mbp and plp1a showed no expression change in cox17-/- mutants, and so did fam168a and fam168b in cox17-/- mutants, but pou3f1 exhibited a slightly reduced expression in cox17-/- mutants (Figs 7C and S6B). When compared with their expression in cox17-/- mutants without copper stimulation (Figs 7C3, C7, C9, C10 and Figs S6B3, B7, B11 B13 and B14), mbp, plp1a, pou3f1, fam168a, and fam168b exhibited reduced expression in Cu2+ stressed cox17-/- mutants at 96 hpf (Figs 7C4, C8, C9, C10 and Figs S6B4, B8, B12, B13 and B14), similar to their expression tendency in Cu2+ stressed WT embryos (Figs 7C1, C2, C5, C6, C9, C10 and Figs S6B1,B2,B5, B6, B9,B10, B13 and B14). The percentages of embryos with reduced expression of the aforementioned genes were significantly increased in either WT or cox17-/- mutants after Cu2+ stimulation (Fig S6C). Additionally, RT-PCR analysis also unveiled the significantly reduced expression of mbp, plp1a (Fig S6D1), pou3f1, fam168a, or fam168b (Fig S6D3) in either Cu2+ stressed WT or cox17-/- embryos, but no significant change of hoxb5b in Cu2+ stressed cox17-/- mutants (Fig S6D2).
Myelin specification was further tested in atp7b-/- embryos after copper stimulation. When compared with their expression in WT embryos, mpb, plp1a, pou3f1, hoxb5b, and fam168a exhibited significantly reduced expression in atp7b-/- embryos (Figs 7D, 7E, and Figs S6E-F), with the expression of mbp (Figs 7D4, D9, 7E1, and Fig S6F1), plp1a (Fig 7E1) and hoxb5b (Fig 7E2) being more significantly reduced in atp7b-/- embryos after copper stimulation. However, WISH and qRT-PCR assays revealed no significant expression change in pou3f1 and fam168a&fam168b in atp7b-/- embryos after copper stimulation (Figs 7D8-D10, 7E3 and S6E, S6F2).
ER stress antagonist PBA was used to further study the role of copper-induced ER stresses in copper- induced down-regulated expression of mbp, hoxb5b, and fam168a. No significant recovery was observed in the expression of the three genes in Cu2+ stressed embryos after PBA co-treatment (Fig S6G).
Discussion
Cu2+ has been reported to induce dysfunctional locomotor in zebrafish larvae(22), but the underlying mechanisms are still poorly understood. In this study, Cu2+ was revealed to induce uncompacted and thinner myelin in the spinal cord, which was consistent with the observations in epb41l2 mutants with dysfunctional locomotor behaviors(36).
It is reported that mbp, a widely used marker for myelin(37), expressed in both the CNS and PNS myelin. Olig2 expressed in oligodendrocyte and olig2 driven fluoresce specifically labels oligodendrocytes and axon in the olig2:DsRed transgenic zebrafish line. In this study, mbp and olig2 exhibited significantly reduced expression in the spinal cord, and the length of axon was significantly reduced in Cu2+ stressed embryos, indicating Cu2+ induced CNS myelin and axon defects in zebrafish. Additionally, the shortened axon might be the secondary damage of defective myelin formation in Cu2+ stressed larvae, as indicated by previous studies showing that myelin abnormalities might precede evidence of axonopathies(38, 39).
Cu2+ specifically induced the down-regulated expression of Wnt signaling and Notch signaling and their down-stream mediator hoxb5b, rather than other hox genes in zebrafish embryos(22). Inhibition of Wnt signaling has been shown to induce hypomyelination, whereas the activation of Wnt signaling significantly increased the transcription of myelin genes in mouse(29). Notch signaling has been revealed to regulate the differentiation of oligodendrocyte precursor cells, and influence oligodendrocyte maturation and myelin wrapping(30, 40), and hox5 has been unveiled to regulate axon extension in motor neurons(26). Consistently, knockdown/out of hoxb5b zebrafish phenocopied the defective CNS myelin and axon observed in Cu2+ stressed embryos, whereas ectopic hoxb5b expression rescued the defects of CNS myelin and axon, indicating Cu2+ partially inhibited CNS myelin and axon marker expression via suppression of hoxb5b. The normal-like morphology of hoxb5b-/- mutant might result from the genetic compensation reported recently(41). Additionally, this study unveiled that both Wnt agonist BIO and Notch signaling activator NICD not only recovered the reduced expression of hoxb5b, but also recovered the myelin and axon defects in Cu2+ stressed embryos, further demonstrating that down-regulated expression of Wnt&Notch-hoxb5b signaling mediated Cu2+-induced myelin and axon developmental defects.
DNA methylation has been suggested to involve in regulation of gene expression and associate with a series of copper induced demyelinating diseases such as WD and AD (16, 17). In this study, it was unveiled that pou3f1, pou3f2 and fam168b exhibited down-regulated expression but hypermethylation separately in their promoter in Cu2+ stressed embryos. Their promoter hypermethylation and reduced expression in Cu2+ stressed embryos suggested the potential correlation of gene transcription with its promoter methylation in Cu2+ stressed larvae.
In this study, it was shown that both fam168b and fam168a exhibited down-regulated expression in Cu2+ stressed embryos, with a highly similar expression pattern to that of pou3f1 during fish embryogenesis. Additionally, similar transcriptional profiles and gene expression patterns, such as enrichment of nervous system and synapse for the down-regulated DEGs as well as down-regulated expression of CNS myelin genes, were observed in both fam168a&fam168b loss/knockdown-of-function embryos and pou3f1 morphants. Pou3f1 and pou3f2 were critical transcription factors in the conversion of embryonic stem cells into neuron and glial cells(42), and the function of pou3f2 was largely overlapped with that of pou3f1 in driving the transition from promyelinating to myelinating cells(43). Fam168b, a novel neural gene identified recently, has been reported to control neuronal survival and differentiation as well as be specifically expressed in myelinated neuron in the CNS in human and mice(44, 45), to exhibit significantly down-regulated expression in AD brains (44), but has never been reported to be involved in myelin development. In this study, similar transcriptional profiles were observed in fam168a, fam168b, and pou3f1 morphants, and ectopic expression of fam168a, fam168b, or pou3f1 could rescue CNS myelin defects in Cu2+ stressed embryos. Taken together, fam168a and fam168b might be novel transcriptional factors similar to pou3f1 in oligodendrocyte differentiation and the subsequent myelin cell development.
In this study, truncated promoter driven GFP and luciferase assays unveiled that the different hypermethylated loci in fam168b promoter, such as locus from -1672 to -1414, -1414 to -1240, and – 1240 to – 927, were critical for its transcriptional regulation during embryogenesis and in cells. The deletion of the aforementioned loci in fam168b promoter could induce significant down-regulation of its transcriptional activity, suggesting the hypermethylated loci are required and pivotal for fam168b transcriptional activation. Collectively, we demonstrated for the first time that Cu2+ might induce hypermethylation in the fam168b promoter, which is correlated with its down-regulated expression in Cu2+ stressed embryos. However, the down-regulated expression of fam168b occurred at 24 hpf, followed by hypermethylation of its promoter at 96 hpf in copper stressed embryos, suggesting the chromatin structure of transcriptional complex with its binding chromosome DNA might be damaged before promoter hypermethylation. This is consistent with the point in recent reports showing that regional methylation could be a secondary consequence of changes in transcriptional complex and chromosome DNA structure(34, 46).
Additionally, up-regulated expression of epigenetic mediators pou3f1/fam168a/fam168b was observed in hoxb5b loss/knockdown embryos, but significantly increased expression of hoxb5b was observed in pou3f1, fam168a, or fam168b morphants, not only suggesting pou3f1/fam168a/fam168b and hoxb5b acted in embryogenesis in a seesaw manner, but also indicating that hypermethylated pou3f1 and fam168a&b were parallel factors to Wnt/Notch-hoxb5b signaling axis in mediating copper-induced myelin and axon defects.
The transfer of copper to mitochondria was assumed to be blocked in copper stressed cox17-/- mutant, and cox17-/- embryos fail to produce ROS after copper stimulation (47), but defects of CNS myelin and axon were still observed in this study, suggesting that copper-induced myelin and axon defects might not be essentially mediated by copper-induced ROS and by the function of cox17 alone. Moreover, in this study, endoplasmic reticulum (ER) stress alleviant PBA was found unable to recover the expression of mbp in Cu2+ stressed WT embryos, suggesting copper-induced ER stresses might not alone mediate copper-induced CNS myelin development defects.
The cox17-/- larvae exhibited significantly reduced expression in pou3f1, fam168a, or fam168b after copper stimulation, while Cu2+ stressed atp7b-/- larvae exhibited slightly down-regulation in the expression of pou3f1 but no expression change in fam168a and fam168b. However, hoxb5b exhibited significantly reduced expression in Cu2+ stressed atp7b-/- mutants but not in Cu2+ stressed cox17-/- mutants. This not only suggested that copper induced changes in the promoter chromatin structure and the down- regulated expression of the fam168a/fam168b/pou3f genes independent of the function of cox17 alone, but also implying that copper required the integral function of atp7b rather than cox17 to induce the promoter methylation and the resultant reduced expression of genes pou3f1/fam168a/fam168b, and required integral function of cox17 rather than atp7b for the down-regulated expression of Wnt&Notch- hoxb5b axis. This might help to explain why myelin defects still occurred in either Cu2+ stressed cox17-/- or atp7b-/- embryos. In this study, we demonstrated that the epigenetic methylation of pou3f1/fam168a/fam168b in Cu2+ stressed cox17-/- embryos or the down-regulated expression in the Wnt&Notch-hoxb5b axis in Cu2+ stressed atp7b-/- embryos separately mediated the down-regulated expression of myelin genes in the Cu2+ stressed mutants. It has been unveiled that copper could locate in cell nucleus and damage the chromatin structure directly (48, 49). Thus, this study provided the direct evidence for the first time that copper damages chromatin structure independent of ROS in DNA methylation during fish embryogenesis.
In summary, this study confirmed the structural and detailed molecular characters of CNS myelin and axon defects occurring in copper stressed embryos. It was shown that copper induced ROS and led to down-regulation of Wnt&Notch-hoxb5b axis, with copper directly inducing locus-specific methylation and the down-regulated expression of pou3f1/fam168a/fam168b genes to mediate myelin and axon defects in copper stressed embryos. The working model is illustrated in Fig 8 for an intuitive understanding of how copper induces CNS defects. The combined data from the current study added novel insights into the mechanisms underlying the unbalanced copper homeostasis in cells linking with neurological disorders.
Materials and methods
Ethics statement
All experiments involving fish in this study were performed in accordance with the recommendations in the Guide for the care and use of Laboratory Animals of the Ministry of Scienc e and Technology of China, which was approved by the Scientific Committee of Huazhong Agricultural University (permit number HZAUFI-2016-007).
Fish stocks
Wild-type zebrafish (Danio rerio) (AB) maintenance, breeding and staging were performed as described previously(50). Tg(olig2:dsRED) and Tg(mbp:EGFP) transgenic lines were obtained from China Zebrafish Resource Center (http://www.zfish.cn/), and the catalog numbers of the lines used were listed in Table S1.
Morpholinos and Cas9/gRNA
The CRISPR/Cas9 genome editing system was reported as an effective tool for gene editing in organisms (51, 52). In this study, the CRISPR/Cas9 system was used to construct homeobox B5b (hoxb5b), family with sequence similarity 168 member A (fam168a), family with sequence similarity 168 member B (fam168b), ATPase copper transporting alpha (atp7a), ATPase copper transporting beta (atp7b) and cytochrome c oxidase copper chaperone COX17 (cox17) mutants. The guide RNAs (gRNAs) were designed to target the first exon of aforementioned genes by ZiFiT Targeter Version 4.2 at the following URL (http://zifit.partners.org/ZiFiT/CSquare9Nuclease.aspx). Sequences of gRNAs are listed in Table S2. The morpholinos (MOs), including hoxb5b-MO, pou3f1-MO, fam168a-MO, and fam168b- MO, were purchased from Gene Tools, LLC (Philomath, Oregon, USA) and their sequences are listed in Table S3.
Drug exposure
Copper and 6-Bromoindirubin-3′-oxime (BIO) (Sigma-Aldrich, USA) were prepared as described previously (22, 27). Zebrafish embryos developed to sphere stage (4 hpf, hours post fertilization) or early were exposed to 3.9 μM copper at random. BIO was added at bud stage. Embryos were harvested at indicated stages. Each group was biologically repeated 3 times.
Transmission electron microscope (TEM) analysis
TEM was performed to test CNS myelin structure in the control and copper stressed embryos at 5 dpf (days post fertilization). A transmission electron microscope (Hitachi H-7650 TEM Japan) was used to acquire the images. G-ratio (axon diameter/myelinated fiber diameter) was calculated to assess myelin thickness. A lower g-ratio indicated a greater myelin thickness. The axon diameter and myelinated fiber diameter were measured using the image J software (NIH, Bethesda, Maryland).
Plasmid construction
The full-length hoxb5b, fam168a, fam168b, POU class 3 homeobox 1 (pou3f1), and the intracellular domain of notch receptor 3 (notch3) (NICD) were amplified using the primers shown in Table S4 and cloned into pCS2 vector for synthesizing mRNAs. 5’ unidirectional deleted mutants of fam168b promoter, including -1672, -1414, -1240, -927, -623, and -284, were amplified using the primers shown in Table S5 and cloned separately into pCS2-GFP vector and pGL3 vector. All constructs were verified by sequencing.
mRNA Synthesis and Injection
For mRNA preparation, capped mRNAs were synthesized using the mMessage mMachine kit (Ambion) according to the manufacturer’s instructions. The synthesized mRNAs were diluted into different concentrations and injected into one-cell stage embryos as reported previously(53).
Quantitative RT-PCR analysis
Zebrafish embryos were collected at indicated stages. Total RNA was isolated from 50 whole embryos/sample using Trizol reagent (Invitrogen). cDNA was synthesized using a M-MLV Reverse-Transcript Kit (Applied Biological Materials Inc, BC, Canada). qRT-PCR was performed as described previously(22, 53, 54). The sequences of the RT-PCR primers were listed in Table S6.
Whole-mount in situ hybridization
Probes for zebrafish myelin basic protein a (mbp), oligodendrocyte lineage transcription factor 2 (olig2), hoxb5b, pou3f1, fam168a, and fam168b were amplified from cDNA pools using primers shown in Table S7. Whole-mount in situ hybridization (WISH) was performed as described previously(50, 53, 55). WISH embryos were photographed with a Leica M205FA stereomicroscope. The signal area in each image was calculated by Image J software (NIH, Bethesda, Maryland). Embryos with changed expressions in the tested genes were identified and their percentage was calculated as reported in our previous works(22, 53, 56).
RNA-sequencing (RNA-Seq) and analysis
WT embryos, WT embryos injected with a different MO of hoxb5b, pou3f1, fam168a, or fam168b, cox17-/- and atp7a-/- mutants and Cu2+ stressed atp7a-/- mutants at 96 hpf were lysed by Trizol reagent (Ambion, Life Technologies) for RNA preparation. The RNAs were then reversely transcribed, and amplified cDNA were sequenced on a BGISEQ-500 platform (BGI, Wuhan, China). Quantile normalization and subsequent data processing were performed using the RSEM v1.2.8 software package. Pathway and GO (Gene ontology) analyses were carried out to determine the roles of the differentially expressed genes (DEGs). The Hierarchical Cluster Tree (dendrogram) was constructed to show the relationships of expression levels among different samples.
Confocal microscopy
Embryos were anesthetized with a low dose of tricaine and mounted on dishes with 1% low-melting agarose. Confocal images were captured by a Leica (Wetzlar, Germany) TCS SP8 confocal laser microscope. The fluorescence intensity of the positive cells in embryos was analyzed by software of Image J. Axon tracing and measurement was performed by using the Neuron J (Image J) software (National Institutes of Health, Bethesda, MD).
Bisulfite PCR validation
Whole genome bisulfite sequencing (WGBS) has been performed in the control and the Cu2+ stressed embryos at 96 hpf, and 57 differential methylated genes (DMGs) were unveiled (35). In this study, the regions for differentially methylated loci of the candidate genes such as fam168bbetween the control group and Cu2+ stressed group were used for bisulfite PCR to validate the results of whole- genome bisulfite sequencing. The target fragments were amplified using specific primers (Table S8) designed with Methyl Primer Express v1.0 (http://www.urogene.org/cgi-bin/methprimer/methprimer.cgi). The obtained PCR products were purified using Min Elute Gel Extraction kit (OMEGA) and cloned into the pMD19-T Vector (Takara). The positive clones were confirmed by PCR and 12 clones were sequenced for each subject.
Luciferase reporter assay
Different truncated mutant promoters of fam168b were used for luciferase assays in this study. The luciferase reporter assays were performed as described previously (53). The data were reported as the mean ± SD of three independent experiments in triplicate (53).
Statistical analyses
The sample size used for different experiments in each group was larger than 10 embryos (n>10), and 2-3 biological replicates were performed for each test. Percentage analysis of the results among different groups was performed using hypergeometric distribution in the R-console software(57). Statistical data of the signal area and fluorescence level in different samples were analyzed using t-test by GraphPad Prism 7.00 software. Each dot represents signal level in an individual embryo. Statistical data of axon length were processed by GraphPad Prism 7.00 software. Each dot represents the length of each axon. The qRT-PCR data were analyzed by one-way analysis of variance (ANOVA) and post hoc Tukey’s test in the Statistic Package for Social Science (SPSS) 19.0 software. Each dot represents one repeat. The statistical analysis for luciferase reporter assay results was performed using GraphPad Prism 7.00 software (unpaired t-test) (GraphPad Software Inc). Data were presented as mean ± SD, *P < 0.05, **P < 0.01, ***P < 0.001.
Supplementary Information
Supplementary materials include 6 supplementary figures and figure legends, and 15 supplementary tables.
Funding
This work was supported by National Key R&D Program of China (2018YFD0900101), by the project 2662018JC024 of the Fundamental Research Funds for the Central University (to J-X. L.), and by the project of key laboratory of Biodiversity and Conservation of Aquatic Organisms (to J-X. L.), and by National Natural Science Foundation of China (Program No. 31771402 to GL. L.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Author contributions
J.X.L conceived the project and wrote the manuscript; T.Z performed most of the experiments, analyzed data and wrote the manuscript. P.P.G and G.Z contributed to analyzed data and experiments. G.L.L, Y.P.F, H.F and J.F.G supervised the project and approved the final manuscript.
Conflict of Interest
The authors declare no competing interests.
Supporting Information Legends
Fig S1. Expression of central neural system (CNS) myelin genes in Cu2+ stressed embryos. Related to Fig 1.
(A)WISH analysis of expression for myelin oligodendrocyte marker olig2 at 48hpf (A1-A2) in the control or Cu2+ stressed embryos, and the quantification analysis of the WISH data in different samples (A3). (B) The percentage of embryos exhibited reduced expression of indicated CNS myelin genes in different samples. Each experiment was repeated three times, and a representative result is shown. Data are mean ± SD. A1-A2, lateral view, anterior to the left and dorsal to the up. The red arrow indicates olig2-expressing in oligodendrocyte. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant. Scale bars, 100 μm.
Fig S2 Cu2+ induces CNS myelin and axon defects by down-regulating Wnt & Notch - hoxb5b regulatory axis in zebrafish embryos. Related to Fig 2 and 3.
(A) The percentage of embryos exhibited reduced expression of indicated CNS myelin genes in WT embryos, hoxb5b Morphants, or hoxb5b-/- mutants at 96hpf(A1-A2). (A3) The percentage of embryos exhibited reduced expression of mbp in the control, Cu2+ stressed embryos, or Cu2+ stressed embryos with ectopic hoxb5b expression at 96hpf. (B) Clustering analysis of Notch signaling genes with reduced expression in Cu2+ stressed embryos at 24 hpf(B1). qRT-PCR analysis of Notch signaling genes notch1a, dlc and mib expression in the control or Cu2+ stressed embryos at 24gpf (B2). The percentage of embryos exhibiting reduced expression of mbp in the control, Cu2+ exposed, or Cu2+ and BIO co-exposed embryos with NICD ectopic expression at 96 hpf(B3). Each experiment was repeated three times, and a representative result is shown. Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant.
Fig S3 DNA methylation and transcriptional activity of myelin genes in Cu2+ stressed embryos. Related to Fig 4.
(A) Analysis of bisulfite sequencing data for fam168b, pou3f1 and pou3f2 methylation levels in the control or Cu2+ stressed embryos(A1). RNA-seq analysis of fam168a, pou3f1 and pou3f2b expression in the control or Cu2+ stressed embryos(A2).The percentage of embryos exhibited reduced expression of pou3f1 in the control or Cu2+ stressed embryos at 96hpf(A3).(B) Graphical representation of methylation patterns in the promoter domain of genes pou3f1 (B1) and pou3f2 (B2) separately in the control or Cu2+ stressed embryos. (C) Bisulfite PCR validation of fam168b in the control or Cu2+ stressed embryos. (D) Quantification of fluorescence level in 24hpf WT embryos injected separately with plasmid containing different truncated fam168b promoter driven GFP reporters. Each experiment was repeated three times, and a representative result is shown. Data are mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant.
Fig S4 Myelin and axon formation in fam168a/fam168b loss- and gain-of-function embryos Related to Fig 5.
(A) The expression pattern of zebrafish fam168a during embryogenesis. (B) The expression pattern of zebrafish fam168b during embryogenesis. (C) Schematic diagram showing the genomic structure and a genetic mutation of fam168a (C1) and fam168b (C2) used in this study. ATG denotes the translation start codon; TGA denotes the translation terminate codon; PAM denotes the protospacer adjacent motif; slash denotes intron; blue horizontal bar denotes exon; dotted lines denote deletion of fam168a and fam168b; Numbers denote the length of mutant base. (D) Enrichment of genes exhibited down-regulated expression in fam168a or pou3f1 morphants at 96hpf via KEGG pathway analysis (D1-D2) and venn diagrams representing the overlapping down-regulated nervous system genes in fam168a and fam168b morphants at 96 hpf (D3). (E) Gene ontology (GO) classification of the genes exhibited down-regulated expression in fam168a(E2) or pou3f1 morphants at 96hpf (E2) and venn diagrams representing the overlapping down-regulated synapse genes in fam168a and fam168b morphants at 96 hpf (E3). (F)WISH analysis of expression for CNS myelin marker plp1a at 96hpf in the WT, fam168a-/- or fam168b+/- embryos (F1-F3), and the quantification analysis of the WISH data in different samples (F4). (G) The percentage of embryos exhibiting reduced expression in different samples. (H) WISH analysis of CNS myelin marker mbp expression in the control, Cu2+ stressed embryos, or Cu2+ stressed embryos with ectopic different mRNA expression(H1-H5). (H6) Quantification analysis of the WISH data in different samples. (H7) Percentage of embryos exhibited reduced expression of mbp in different samples. Each experiment was repeated three times, and a representative result is shown. Data are mean ± SD. A1-A6, B1-B6, dorsal view, anterior to the top, A7-A8, B7-B8, F1-3, H lateral view, anterior to the left and dorsal to the up. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant. Scale bars, 100 μm.
Fig S5 percentage of embryos exhibited abnormal expression in different samples. Related to Fig 6.
Fig S6 RNA-seq,WISH and qRT-PCR analysis unveils differentially expressed nervous system genes in cox17 and atp7b mutant embryos. Related to Fig 7.
(A) Clustering analysis of endoplasmic reticulum genes exhibited down-regulated expression in atp7a-/- mutant embryos at 96 hpf(A1). Clustering analysis of the expression of endoplasmic reticulum stress genes in WT embryos, Cu2+ stressed WT embryos, atp7a-/- mutants, or Cu2+ stressed atp7a-/- mutants at 96hpf (A2). Clustering analysis of myelin genes in WT embryos, Cu2+ stressed WT embryos, atp7a-/- mutants, or Cu2+ stressed atp7a-/- mutants at 96hpf (A3). (B) WISH analysis of the myelin genes plp1a, fam168a and fam168b expression in WT embryos, Cu2+ stressed WT embryos, cox17-/- mutant embryos or Cu2+ stressed cox17-/- mutant at 96hpf (B1-B12). Quantification analysis of the WISH data in different samples (B13-B14). (C) Percentage of embryos exhibited reduced expression in different samples. (D) qRT-PCR analysis of CNS myelin markers mbp, plp1a(D1), transcriptional factor hoxb5b(D2), pou3f, fam168a and fam168b(D3) expression in WT embryos, Cu2+ stressed WT embryos, cox17-/- mutant embryos or Cu2+ stressed cox17-/- mutant at 96hpf. (E) WISH analysis of myelination transcriptional factors fam168a and fam168b in WT embryos, Cu2+ stressed WT embryos, atp7b-/- mutant embryos or Cu2+ stressed atp7b-/- mutant at 96hpf (E1-E8). Quantification analysis of the WISH data in different samples (E9). (F) Percentage of embryos exhibited reduced expression in different samples. (H) qRT-PCR analysis of mbp (H1), hoxb5b (H1) and fam168a (H3) in WT embryos, Cu2+ stressed, and Cu2+ and PBA co-exposed embryos at 96hpf. Each experiment was repeated three times, and a representative result is shown. Data are mean ± SD. B1-A16, F1-F8, lateral view, anterior to the left and dorsal to the up. The red arrow indicates mbp-expressing in the spinal cord. *P < 0.05, **P < 0.01, ***P < 0.001. NS, not significant. Scale bars, 100 μm.
Acknowledgments
We are grateful to Dr. Yibing Zhang (Institute of Hydrobiology, Chinese Academy of Science) for the gift of HEK293T cells. We thank Mr. Qinhan Xu, Bei Cui and Haojie Sun (Huazhong Agricultural University) for participating in the study.