Lafora disease E3 ubiquitin ligase malin is recruited to the processing bodies and regulates the microRNA-mediated gene silencing process via the decapping enzyme Dcp1a

Abstract

Intracellular transport, processing and stability of mRNA play critical roles in the functional physiology of the cell and defects in these processes are thought to underlie the pathogenesis in a number of neurodegenerative disorders. One of the cellular sites that regulate the mRNA half-life is the processing bodies, the dynamic cytoplasmic structures that represent the non-translating mRNA and the ribonucleoprotein complex that also control the decapping and translation of mRNA. In the present study we explored the possible role of malin E3 ubiquitin ligase in the mRNA decay pathway via the processing bodies. Defects in malin are associated with Lafora disease (LD)—a neurodegenerative disorder characterized by myoclonus seizures. We show here that malin is recruited to the processing bodies and that malin regulates the recruitment of mRNA decapping enzyme Dcp1a by promoting its degradation via the ubiquitin proteasome system. Depletion of malin results in elevated levels of Dcp1a and an altered microRNA-mediated gene silencing activity. Our study suggests that malin is one of the critical regulators of processing bodies and that defects in the mRNA processing might underlie some of the disease symptoms in LD.

Keywords: :

• neurodegenerative disorder
• mRNA dysregulation
• epilepsy
• post-translational modification
• RNA granules
• stress response
• neuronal processing bodies

Introduction

An emerging theme in the understanding of various neurodegenerative disorders suggests that malfunctions in the mRNA metabolic pathways could be one of the common underlying mechanisms that lead to the neurodegenerative changes.¹ For example, in disorders caused by the triplet repeat expansion, the transcripts bearing the expanded repeat form nuclear foci that recruit several splicing factors leading to abnormal splicing of transcripts.^2-6 Besides the mRNA, the protein products of some of these transcripts result in the formation of intracellular inclusions which recruit the transactive response binding protein of 43 kDa (TDP-43) and fused-in-sarcoma (FUS) proteins that normally bind to the RNA and are thought to regulate mRNA processing.^7-13 Indeed one of the pathological hallmarks of the TDP-43 and FUS in neurodegenerative conditions is their redistribution from the nucleus to the cytoplasm and studies have shown that this nuclear loss could result in the neurodegenerative changes.^14-22 Intriguingly, the wild-type version of the polyglutamine disease associated proteins, the ataxin-2 and the huntingtin protein, are involved in the formation and/or the function of processing bodies and stress granules.^23,24 Among these, the processing bodies (PBs) represent the ubiquitous form of cytoplasmic structures that house the RNA decay machinery.^25,26 PBs are thought to mediate the mRNA silencing by several distinct mechanisms, and the number and position of PBs could vary depending on the cell type and/or the amount of mRNAs degraded via the silencing pathways.^26-28 Emerging studies suggest that PBs might help in regulating several neuronal functions, including the learning and memory, thereby underscoring the role of RNA processing in the neurodegenerative disorders.²⁹

Lafora disease (LD) is a fatal form of progressive myoclonus epilepsy characterized by neurodegeneration and the presence of polyglucosan inclusions in the neurons and other affected tissues.^30-32 The symptoms of LD include ataxia, dementia and hallucination besides the epileptic seizures.^30-32 LD can be caused due to defects in the EPM2A gene that codes for the laforin protein phosphatase or the NHLRC1 gene that codes for the malin E3 ubiquitin ligase.³¹ Studies on cellular models have shown that laforin and malin interact with each other, and appear to regulate, as a functional complex, several cellular functions such as glycogen metabolism, cellular stress response, and the proteolytic processes.^33-41 Laforin is known to associate with the polyribosomes, therefore a possible role for the two proteins in mRNA metabolism was proposed but not tested.^42,43 Here we show that malin is recruited to the PBs, and is one of the critical regulators of the PB resident protein, Dcp1a.

Results

Malin is recruited to processing bodies (PBs)

We have shown earlier that laforin and malin show stress-specific subcellular localization; proteasomal blockade results in their recruitment to aggresomes,⁴¹ heat shock induces their translocation to the nucleus,³⁸ and the glucose starvation induced only laforin to translocate to the nucleus.⁴⁴ We extended our studies to other stressors, and found a significant increase in the frequency of cells exhibiting cytoplasmic granule-like foci for the transiently expressed malin when the cells were treated with zinc or puromycin (Fig. 1A and B). Such cytoplasmic foci were also seen in untreated cells but at a lower frequency as compared with cells treated with either zinc or puromycin (Fig. 1B). Such effects were seen in more than one cell line (HeLa, COS7, and Neuro2A) (Fig. 1A; Fig. S1) and were independent of the fusion tags used for detecting the overexpressed protein (Fig. 1C; Fig. S1). Since zinc and puromycin treatments are known to induce PBs^45,46 − the subcellular sites where decapping and degradation of mRNAs takes place^25,26—we wanted to check whether the malin-positive cytoplasmic foci are indeed PBs. For this we coexpressed malin with argonaute2 (Ago2), an established PB marker.⁴⁷ Interestingly, malin showed complete colocalization with the cytoplasmic foci of Ago2 (Fig. 1C). Similar observations were made when malin was transiently expressed with the mRNA-decapping enzyme 1A (Dcp1a) − another well-known PB marker (Fig. 1C). The cytoplasmic granules of transiently expressed malin also showed colocalization with the endogenous PB markers GW182 and Xrn1 (Fig. 1D).^48,25 None of the available antibodies detected the endogenous malin in immunofluorescence staining, therefore we were unable to establish whether the endogenous malin gets recruited to the PBs or not. Transiently expressed laforin however did not co-localize with endogenous PB markers (Fig. S2). Therefore laforin was not considered for the rest of the analyses.

Figure 1. Recruitment of malin to the PBs. (A) Representative images showing subcellular localization of transiently expressed GFP-tagged malin in HeLa cells. Note the punctate cytoplasmic granules of malin (identified with arrowheads). (B) Bar diagram showing the frequency of cells showing the punctate cytoplasmic granules (as shown in (A) for the GFP-tagged malin in HeLa cells upon treatment with puromycin or zinc or without any treatment (control), as indicated (***, p < 0.0005; t-test; n = 3). (C) Representative images showing the colocalization of Myc-tagged malin with the PB marker Ago2 (GFP-tagged) or GFP tagged malin with Dcp1a (RFP-tagged) in HeLa cells when transiently coexpressed. (D) Representative images showing the colocalization pattern of transiently expressed malin (Myc-tagged) with endogenous markers for PBs—the GW182 and Xrn1, as indicated. Scale bar, 10 μM (A, C-D).

Malin regulates the recruitment of mRNA-decapping enzyme 1A (Dcp1a) to processing bodies

Decapping is a critical step for the degradation of cytoplasmic mRNA, and this process is regulated by a multiprotein complex.^49,50 Among them Dcp2 is the primary decapping enzyme whose activity is modulated by a few co-factors including the Dcp1a.^49-51 Decapping is necessary for the degradation of mRNAs, both in normal mRNA turnover and in the nonsense- mediated mRNA decay.^26,51 Dcp1a, when expressed alone, resulted in the formation of PBs in nearly 100% of the transfected cells (Fig. 2A). However, when coexpressed with the wild-type malin there was a significant reduction in the number of cells that showed Dcp1a-positive PBs (Fig. 2A and B). However no such reduction was noted when Dcp1a was co-expressed with catalytically inactive malin mutants (Fig. 2A and B). The malin mutants used are known to cause LD phenotype in recessive condition, and are known to be inactive as ubiquitin ligase.³¹ It is intriguing to note that cells that strongly expressed wild-type malin showed loss of Dcp1a-postive foci (Fig. 2B, panel 1) while no such difference was noted when malin mutants were coexpressed with Dcp1a (Fig. 2B, panels 2 and 3). Similar observations were made when the endogenous Dcp1a was visualized using an antibody in the wild-type or the mutant malin expressing cells (Fig. S3). Interestingly, this effect of wild-type malin was restricted to Dcp1a as Ago2 did not show such a difference in localization pattern when co-expressed with the wild-type malin (Fig. 2A), suggesting that malin could probably regulate the recruitment of only Dcp1a to the PBs. To investigate this possibility we co-transfected cells with the constructs that code for Dcp1a, Ago2 and either the wild-type malin, or its catalytic inactive mutant (C26S), or an empty vector and the frequency of cells showing PBs positive for both Dcp1a and Ago2 was scored. As shown in Figure 3A, nearly 75% of the transfected cells showed PBs that were positive both for Dcp1a and Ago2 when only these two proteins were transiently co-expressed or when they were co-expressed with the mutant malin. However, coexpression of wild-type malin resulted in a dramatic reduction in number of cells with PBs that were positive both for Dcp1a and Ago2 (Fig. 3A), suggesting that malin probably negatively regulates the recruitment of Dcp1a to PBs. To further confirm this observation, we looked at the number of endogenous Dcp1a-positive PBs in cells in which malin was partially knocked down by the RNAi approach. As shown in Figure 3B, there was a significant increase in the number of PBs that were positive for the endogenous Dcp1a when malin was suppressed. Similar observations were made even for the overexpressed Dcp1a but not for overexpressed Ago2 (Fig. 3C and D), suggesting that malin indeed regulates the recruitment of Dcp1a to the PBs. The efficiency of shRNAi knockdown constructs have been established previously^40,44 and have been validated once again in the present study (Fig. S4A).

Figure 2. Malin regulates the recruitment of Dcp1a to PBs. (A) Bar diagram showing frequency of cells having Dcp1a or Ago2-positive PBs in HeLa cells cotransfected with the expression constructs as indicated (***, p < 0.0005; t-test; n = 3). (B) Representative images showing the colocalization pattern of wild-type malin or its mutant forms—C26S mutation affecting the RING domain, and deletion mutant F216_D233 del (identified as “NHL del”) affecting the carboxyl-terminal NHL domain—when coexpressed with the RFP-tagged Dcp1a in HeLa cells. Both mutants are known to cause LD in recessive condition, and are shown to lack E3 ubiquitin ligase activity.³¹ Note the absence of Dcp1a positive PBs in the cell that had higher levels of wild-type malin expression and the presence of Dcp1a-positive PBs in the cell that had lower expression of wild-type malin (identified by arrows). No such difference was noted when malin mutants were coexpressed.

Figure 3. Malin regulates the recruitment of Dcp1a to the PBs. (A) Bar diagram showing frequency of cells having PBs that are positive both for Dcp1a and Ago2 when they were coexpressed with the wild-type malin or its mutant C26S, as indicated (***, p < 0.0005; t-test; n = 3). (B) A bar diagram showing the number of endogenous Dcp1a-positive PBs observed per cell when the cells were transfected with the RNAi construct to knockdown malin, or with the non-silencing RNAi (NS-RNAi) construct, as indicated (***, p < 0.0005; t-test; n = 3). (C) Bar diagram showing number of Dcp1a-, or Ago2-positive PBs observed per cell when Dcp1a or Ago2 was transiently expressed with the knockdown construct for, malin or the non-silencing control RNAi (NS-RNAi) construct, as indicated (***, p < 0.0005; t-test; n = 3). (D) Representative images showing the Dcp1a- or Ago2-positive PBs in HeLa cells transfected with the malin and non-silencing RNAi (NS-RNAi) constructs as indicated. Scale bar, 10 μM.

Malin regulates the cellular levels of Dcp1a by promoting its degradation via the ubiquitin-proteasome system

Since malin is an E3 ubiquitin ligase, and since overexpression of malin excludes Dcp1a from the PBs, we speculated that malin might interact with and promote the degradation of Dcp1a. To test this possibility, we looked at the cellular level of endogenous Dcp1a in conditions when either the wild-type or catalytically inactive mutant malin was overexpressed. As shown in Figure 4A, transient expression of wild-type malin (lane 2), but not it’s catalytic inactive mutant (lane 3), was able to decrease the cellular level of endogenous Dcp1a. Further, treating wild-type malin expressing cells with the proteasomal blocker MG132 rescued the loss of endogenous Dcp1a (Fig. 4A; compare lane 2 with 4) suggesting that malin probably regulates Dcp1a by promoting its degradation via the proteasome. Similar observations were made with the overexpressed Dcp1a as well; wild-type malin (Fig. 4B) but not its mutant (Fig. 4C) was able to decrease the cellular level of overexpressed Dcp1a in a dose-dependent manner. We also found an increase in the cellular level of endogenous Dcp1a when malin was knocked down using the RNAi approach (Fig. 5A), suggesting that malin promotes the degradation of Dcp1a. This effect of malin was specific to Dcp1a as the protein levels of two other PB markers GW182, and Ago2 did not alter when malin was partially knocked-down (Fig. 5A). Having found that malin regulates the cellular level of Dcp1a, we checked whether Dcp1a levels are altered in the malin knockout mouse. Malin knockout mice develop most of the symptoms/pathology seen in LD.⁵² As shown in Figure 5B, the level of Dcp1a was significantly higher in the brain tissue of malin knockout mice as compared with the brain tissue of wild-type littermates. These results clearly suggest that loss of malin results in increased levels of Dcp1a both in the cellular and animal models.

Figure 4. Malin promotes the degradation of Dcp1a. (A) Immunoblot showing the cellular level of endogenous Dcp1a in HeLa cells transiently transfected with the indicated expression constructs and probed for indicated proteins. Note the reduced levels of endogenous Dcp1a in the presence of wild-type malin (lane 2) and its rescue when the cells were treated with the proteasomal blocker MG132 (lane 4). The bar diagram shown below depicts the fold difference in the levels of endogenous Dcp1a (as measured by densiometric analysis) upon expressing the wild-type or the mutant form (C26S) of malin and when the cells were treated with MG132 as identified. (***, p < 0.0005; t-test; n = 3). (B and C) Immunoblot analysis for the lysates of the cells transfected with the expression constructs for RFP-tagged Dcp1a, GFP (in lanes 2-to-4) and increasing proportion of constructs (in lanes 3 and 4; identified by the slanting triangle on the top) for the wild-type (WT) (B) or the mutant form (C26S) of malin (C), as indicated. GFP served as the control for transfection efficiency. The bar diagrams shown below depicts the fold difference in the levels of RFP-tagged Dcp1a (as measured by densiometric analysis) when increasing concentration of wild-type (B) or mutant form of malin was coexpressed (C). (**,p < 0.005; ***, p < 0.0005; t-test; n = 3).

Figure 5. Malin interacts with and ubiquitinates Dcp1a. (A) Immunoblot showing the cellular levels of endogenous Dcp1a, GW182, and the transiently expressed GFP-tagged Ago2 in cells transfected with the RNAi construct for the knockdown of malin or a non-silencing control RNAi construct (NS-RNAi), as indicated. Probing for tubulin served as the loading control. Note that except Dcp1a, no other PB marker showed change in level on malin knockdown. For Ago2, the relative level was evaluated using an expression construct coding for GFP-tagged Ago2 and probed using anti-GFP antibody as indicated. (B) Immunoblot showing the levels of endogenous Dcp1a in two pairs of brain tissues of the malin knockout mice (KO) as compared with their wild-type littermates (WT). (C) Malin physically interacts with endogenous Dcp1a. Myc-tagged malin or an empty vector (pcDNA) was transiently expressed in HeLa cells and the cell lysates were processed for immunoprecipitation using anti-Myc antibody. The immunoprecipitate (IP) or the whole cell lysate (WCL) was probed with the indicated antibodies. The cells were pretreated with MG132 to arrest the degradation of Dcp1a. (D) Malin ubiquitinates Dcp1a. HeLa cells transiently transfected with an expression construct coding for RFP-tagged Dcp1a and wild-type malin or its catalytically inactive mutant form (C26S), treated with the proteasomal blocker MG132 and the cell lysate was immunoprecipitated (IP) using anti-ubiquitin antibody to trap the ubiquitinated proteins. The IP fraction and the whole cell lysates (WCL) were immunoblotted with indicated antibodies. Arrow identifies the expected molecular weight at which the full length Dcp1a-RFP should migrate (105 kDa).

We next checked whether malin physically interacts with Dcp1a to promote its degradation. For this, cells transiently transfected with an expression construct coding for the Myc-tagged malin or an empty vector were treated with proteasome blocker MG132 to prevent Dcp1a degradation, and the cell lysates were processed for immunoprecipitation with anti-Myc antibody and immunoblotting with an antibody against Dcp1a. As shown in Figure 5C, malin was able to pull-down endogenous Dcp1a. We further analyzed whether malin ubiquitinates Dcp1a to promote its degradation. For this, we transiently coexpressed RFP-tagged Dcp1a with either the wild-type malin or its catalytically inactive mutant (C26S). The cells were treated with the proteasomal blocker MG132, the ubiquitinated proteins were pulled down using anti-ubiquitin antibody, and the pulled-down product and the whole cell lysates were immunoblotted to detect the transiently expressed RFP-tagged Dcp1a. As shown in Figure 5D, coexpression of wild-type malin resulted in a marked increase in the high molecular weight ubiquitinated forms for Dcp1a as compared with the set wherein the mutant malin was coexpressed, suggesting that the difference noted for Dcp1a in the ubiquitinated fraction is indeed due to the ubiquitin ligase activity of malin.

Malin regulates mRNA silencing via the microRNA pathway

PBs are multi-protein complexes, and depletion of some of their components is known to compromise the PB-mediated mRNA surveillance. For example, depletion of GW182, Xrn1 or Dcp2 are known to alter the number and size of PBs^45,53 and loss of GW182 or the Dcp1-Dcp2 complex affects miRNA-mediated silencing of mRNAs.⁵³ Since malin regulates the cellular level of Dcp1a and its recruitment to the PBs, we wanted to check whether loss of malin would alter the PB function. The PBs are known to be important for the function of the miRNA pathway.^47,53 To check for the possible role of malin on miRNA-dependent mRNA silencing, we first knocked down endogenous malin in HeLa cells by RNAi approach and after 24 h, the cells were transfected with a reporter construct coding for an mRNA that has two target sequence sites for the endogenous let-7 miRNA (referred to here as “Let7R-WT”) or with a reporter construct encoding a transcript with a mutant version of the target sequence for let-7 miRNA such that the latter cannot be silenced by the endogenous miRNA (referred to here as “Let7R-MT”). Both constructs had a coding sequence for the Renilla luciferase, hence the activity of this protein was measured to determine the miRNA-mediated mRNA silencing. To normalize the transfection efficiency, a construct that codes for the firefly luciferase was cotransfected. We have used a shRNA construct for the PB marker GW182 (Fig. S4B) and shRNA construct for GFP as the positive and negative controls, respectively. As shown in Figure 6A, the knockdown of GW182 significantly reduced the miRNA-mediated repression of Let7R-WT reporter as noted previously,^24,53 while the knockdown of GFP did not show any such effect. Partial knockdown of malin using two independent shRNA constructs resulted in a significant increase in miRNA mediated Let7R-WT reporter repression, suggesting that malin might impair the let-7-dependent mRNA silencing (Fig. 6A). There was no change in the activity for the mutant construct Let7R-MT for any of the knockdowns (GW182, malin, GFP) (Fig. 6A). We have also checked whether overexpression of malin would inhibit the let-7-mediated mRNA degradation. This we did by co-expressing the wild-type or mutant form of malin with the reporter constructs and measured the luciferase activity. As shown in Figure 6B, overexpression of wild-type malin resulted in a significant reduction in the miRNA- mediated silencing of the Let7R-WT reporter activity as compared with the cells that coexpressed the mutant malin or GFP. No significant change in the activity for the Let7R-MT mutant reporter was noted (Fig. 6B). To further establish that the observed difference in the let-7 mediated reporter activity is indeed due to the malin-dependent changes in the Dcp1a level, HeLa cells were transiently transfected with RNAi constructs for malin, Dcp1a, and the Let-7 reporter construct and the luciferase activity was measured. As shown in Figure 6C, the double knockdown significantly rescued the reporter repression seen when only malin was knocked-down. The efficiency of Dcp1a siRNA was established by immunoblotting (Fig. S4C). Taken together our results suggest that malin negatively regulates the let-7 mediated mRNA silencing by modulating the cellular level of Dcp1a.

Figure 6. Malin negatively regulates the let-7 miRNA-mediated gene silencing pathway. (A) Bar diagram showing the let-7 miRNA-mediated repression of a reporter construct. HeLa cells were transfected with shRNAi constructs for GW182, malin, or GFP and at 24 h post-transfection they were split into two sets and transfected again with a construct that code for renilla luciferase and having a target sequence for the endogenous let-7 miRNA (LetR-WT) or a mutated target site for the let-7 miRNA (Let7R-MT) and a construct that code for the firefly luciferase as indicated. The luciferase activity for each set was measured as ratio of firefly to renilla and was normalized with activity of control set (taken as 1 fold) and plotted as fold change in the miRNA-mediated repression of the reporter. Similar assay was done on cells that expressed the wild-type or the mutant form of malin in the place of the knockdown constructs(B), or in conditions of either malin knockdown, or Dcp1a knockdown or in cells were both malin and Dcp1a were knocked down (C) and the activity of the luciferase was measured. Each bar in A, B and C represents the mean of three independent experiments (***, p < 0.0005; t-test). (D) Immunoblot showing the level of Dicer in HeLa cells transfected with shRNAi for malin or a non-silencing shRNAi construct (NS-RNAi), as indicated. (E) Immunoblot showing the level of Dicer in the brain tissue lysates of malin knockout (KO) or their wild-type littermates (WT), as indicated. Two different pairs of animals were used for the immunoblot. The same blot shown in Figure 5B was used to probe with anti-Dicer antibody.

To further probe malin’s role in miRNA-mediated mRNA silencing we looked at the level of one of the known endogenous targets for let-7 miRNA. Dicer, an endoribonuclease of RNase III family,^54,55 is known to be regulated by let-7 miRNA.^56,57 Studies have shown that overexpression of let-7 reduces the dicer expression significantly both at the protein and mRNA levels and let-7 knockdown has the opposite effect.⁵⁶ Let-7 target sequences are present in the coding region of the Dicer transcript and therefore, let-7 is thought to have a direct effect on Dicer expression.^56,57 If knockdown of malin led to an increase in the let-7 mediated mRNA silencing then dicer levels should also go down. As expected, there was a reduction in the level of Dicer, both at protein (Fig. 6D) and mRNA levels (Fig. S4D), when malin was transiently knocked down in HeLa cells. A reduction in the level of Dicer protein was also observed in the brain tissue lysates of malin knockout mice (Fig. 6E). Taken together these results are strong and compelling enough to suggest that malin negatively regulates the miRNA mediated gene silencing by regulating the level of Dcp1a both at the cellular and organism levels.

Discussion

We show here that transiently expressed malin is recruited to the PBs and this observation has been confirmed by co-staining of malin with the overexpressed as well as endogenous markers for these dynamic structures. The possible role of malin in PBs was further established by using reporter assays and also by evaluating some of the endogenous target genes. We however could not establish the presence of endogenous malin in the PBs since none of the available antibodies against malin work well in the in situ detection.

The Dcp1a is an indispensable member of the mRNA-decapping complex that recruits mRNAs to the PBs for the translational repression and their eventual degradation via the miRNA pathway.^58,59 It is known that the individual components of PBs, including the decapping enzymes, shuttle dynamically within the cytoplasmic pool and this dynamics are regulated by some critical cofactors.^60-62 For example, phosphorylation regulates the localization of Ago2 and Dcp1a to the PBs.^63,64 In addition, Ago2 localization is also regulated by mLin41, an E3 ubiquitin ligase, by modulating the half-life of Ago2.⁶⁵ Our observations that malin’s cellular level shows an inverse correlation with the Dcp1a-positive PBs (number of cells with PBs or the number of PBs per cell) proposes malin to be one of the regulators of Dcp1a. Since the loss or the overexpression of malin did not alter the level and localization of Ago2 (and other makers tested) to PBs, it could be argued that malin regulates the targeting of specific components of PBs, like the Dcp1a, but not the formation of PBs itself. This notion was supported by our demonstration that malin physically interacts with and promotes the degradation of Dcp1a via the ubiquitin-proteasome pathway. Increased level of Dcp1a was also observed in the brain tissue lysates of malin deficient mice. Thus, malin appears to regulate the PB-specific functions of Dcp1a by modulating its turnover—a mechanism very similar to the one proposed for the mLin41 E3 ligase on the stability and recruitment of Ago2 to PBs; overexpression of mLin41 partially rescued the silencing of let-7 miRNA targets by antagonizing Ago2.⁶⁵ In this regard it is of interest to note that Trim32, a malin-like ubiquitin ligase,⁶⁶ interacts with the Argonaute 1 protein and regulates the miRNA-mediated inhibitory mechanisms.⁶⁷ Since malin and Trim32 are phylogenetically related, and since both share some substrates for ubiquitination, it appears that the RING family ubiquitin ligases have functional roles in the miRNA-mediated gene silencing. Consistent with this model we have shown here that loss of malin resulted in an enhanced repression of a reporter for the let-7 miRNA and decreased levels of Dicer, an endogenous target of let-7, possibly due to the increased level of Dcp1a. The contrasting levels of Dcp1a and Dicer were also observed in the brain tissues of malin knockout mice—an animal model for LD that replicated most of the pathological changes seen in human LD.⁵² In this regard it is intriguing to note that loss of Dicer in the brain results in neurodegeneration, the hyperphosphorylation of tau protein, and symptoms like seizures and ataxia,^68-71 and these defining neuropathological changes are also seen in LD.^30-32,72 Therefore it is tempting to speculate that dysregulation in mRNA metabolism may underlie some of the symptoms in LD, and these could be secondary to the increased levels of Dcp1a or the decreased levels of Dicer. It is equally likely that malin may also promote the degradation of an additional PB associated factor not tested in the present study, and loss of which—along with Dcp1a—could result in the deficient miRNA activity and the neurodegeneration in LD. Indeed defects in the dicer-micro RNA network have been implicated in etiology of several neurodegenerative disorders like Parkinson disease, Alzheimer disease and several polyglutamine disorders including Huntington’s disease.^68-70

The emerging knowledge on the role of PBs in the neuronal functions suggests that deficits in PBs may affect several functions of the neurons. For example, PBs have been identified in neuronal dendrites and near synapses and they have been implicated in the signal-dependent mRNA trafficking, processing and/or localized protein synthesis.^29,73 Given the critical roles of PBs in neuronal processes, and the suspected role of LD protein in the mRNA regulation,^42,44 the present set of findings call for further studies on malin in the mRNA processing, and its specific role in the LD neuropathology.

Materials and Methods

Cell culture, transfections and treatments

All experiments were performed in HeLa cells unless otherwise stated. The cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) with 25 mM of glucose and 10% (vol/vol) fetal bovine serum. Cells were transfected using Polyfect reagent (Qiagen) as recommended by the manufacturer. Duration and concentration of various treatments given were as follows unless mentioned otherwise: MG132 (12 h, 20 μM), ZnCl₂ (2 h, 100 µM) and puromycin (1 h, 100 μg/ml).

Expression constructs and chemicals

The mammalian expression constructs that code for the wild-type or mutant form of malin are previously described.^41,42 The construct that code for red fluorescent protein (RFP)-tagged Dcp1a was generously provided by Ken Fujimura (University of Tokyo, Japan). The shRNA knockdown constructs (RNAi) for malin, GFP and the non-silencing RNAi (NS-RNAi) were purchased from Open Biosystems USA, and have been validated in our previous studies^40,44 and are shown here as well (Fig. S4). The shRNAi construct for GW182 and the siRNA duplex for Dcp1a were obtained from Sigma-Aldrich (India) Pvt Ltd and their efficieny has been confirmed (Fig. S4). The miRNA reporter constructs Let7R-WT and Let7R-MT, and the one that code for green fluorescent protein (GFP)-tagged argonaute-2 (Ago2) were obtained from Addgene (Plasmid ID 11324, 11325 and 11590, respectively). All the chemicals were purchased from Sigma-Aldrich Pvt Ltd unless stated otherwise.

Immunostaining and antibodies

Cells, grown on a gelatin-coated sterile glass coverslip, were processed for immunofluorescence microscopy as described earlier.^40,41 Briefly, cells were fixed with 4% paraformaldehyde, permeabilized with TritonX-100 (0.05%) and subsequently incubated with primary and secondary antibodies. For staining the nuclei, cells were treated with 10 µM of 4',6-diamidino-2-phenylindole (DAPI) for 5 min. Immunofluorescence images were captured using an epifluorescence microscope (Nikon Eclipse 80i, Japan, 40× oil objective) and the images were assembled using Adobe Photoshop. The following antibodies were used in the present study: Anti-Myc and anti-GFP (Roche India); anti-Myc (Cell Signaling Technologies); anti-Xrn1, anti-Dcp1a, anti-Dicer and anti-tubulin (Sigma-Aldrich); anti-GW182 (generously gifted by Dr. Marvin J Fritzler); and anti-polyubiquitin (Enzo Life Sciences). Secondary antibodies were obtained from Jackson Immuno Research Inc.

Cell counting

Cell counting was performed in a blinded manner. Each experiment was done in triplicates and 300 cells were counted in each experimental set.

Immunoprecipitation

Immunoprecipitation was performed as detailed earlier.⁷² Briefly, cell lysates were pre-cleared by incubating with protein A agarose beads (Bangalore Genei, India) for 2 h at 4°C, followed by incubation with anti-Myc antibody (CST) overnight at 4°C. Finally the protein A agarose beads were added and incubated for 2-to-3 h at 4°C. The beads were then washed 3 times and the bound proteins were eluted by boiling the beads with SDS sample buffer.

Malin knockout mouse tissue lysates

Brain tissues of malin-deficient mice and their wild-type littermates,⁵² obtained in frozen condition from Dr. Berge Minassian (The Hospital for Sick Children, Canada), were homogenized and processed for immunoblotting as reported earlier.⁷²

Immunoblotting

For immunoblotting, protein samples were resolved on SDS-PAGE gel and transferred onto 0.22 µm nitrocellulose membrane (MDI). After blocking with 5% non-fat modified milk (G Biosciences), the membranes were processed for sequential incubations with primary antibody followed by incubation with HRP- tagged secondary antibody, as reported previously.⁴⁴ Immunoreactive proteins on the filter were visualized using a chemiluminscent substrate (SuperSignal West PICO, Pierce).

Ubiquitination assay

The cells were treated with 10 µM MG132 at 24 h post-transfection for 12 h, and the lysates were then processed for immunoprecipitation with anti-polyubiquitin antibody and for immunoblotting as described.

Micro-RNA reporter assay

The reporter assay as described by Savas et al.²⁴ was followed to measure the miRNA activity. Briefly, the cells were transfected with shRNAi construct to knockdown GW182, malin, GFP (control) or with siRNA for Dcp1a or with an expression construct for the wild-type or mutant form of malin or GFP. At 24 h post-transfection, the cells were split into two sets and transfected again with either the Let7R-WT construct having a target of endogenous let-7 miRNA and coding for renilla luciferase or with the Let7R-MT construct having a mutated target site for let-7 miRNA and coding for renilla luciferase. In each set a construct that code for the firefly luciferase was cotransfected to serve as transfection control. The luciferase activity for each set was measured as ratio of firefly to renilla and was normalized with activity of control set and plotted as fold change in the miRNA-mediated repression of the reporter.

Statistical analysis

Data were analyzed by two-tailed, unpaired Student’s t-test using GraphPad software. Differences were considered significant at p < 0.05 and the same was denoted by *, ** and *** when the p-value is less than 0.05, 0.005 and 0.0005, respectively.

Abbreviations:
LD	=	Lafora Disease
PBs	=	Processing bodies
Ago2	=	argonaute-2
SDS-PAGE	=	sodium dodecyl sulfate-polyacrylamide gel electrophoresis
GFP	=	green fluorescent protein
RFP	=	red fluorescent protein
TDP-43	=	transactive response binding protein of 43 kDa
FUS	=	fused-in-sarcoma
GW182	=	glycine-tryptophan protein of 182 kDa

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

We thank Dr. Ken Fujimura, University of Tokyo, Japan, for providing us the RFP-tagged Dcp1a expression construct, Dr. Marvin J. Fritzler, University of Calgary, Canada, for the gift of GW182 antibody, Dr. Berge Minassian, Hospital for Sick Children, Canada, for the tissue lysates of the malin knockout mice, and Dr Amitabha Bandyopadhyay, IIT Kanpur, for extending his laboratory facilities. We also thank the anonymous reviewers for their comments and suggestions on an earlier version of this article. This work was supported by a sponsored research grant from the Department of Atomic Energy, Government of India, to S.G. S.G. is a Ramanna Fellow and Gill-Joy Chair Professor at IIT Kanpur. S.S. and P.K.S. were supported by a senior research fellowship from the Council of Scientific and Industrial Research, Government of India.

Supplemental Material

Supplemental material may be found here:

http://www.landesbioscience.com/journals/rnabiology/article/22708/