Splicing factors act as genetic modulators of TDP-43 production in a new autoregulatory TDP-43 Drosophila model

Abstract

TDP-43 is a critical RNA-binding factor associated with RNA metabolism. In the physiological state, maintaining normal TDP-43 protein levels is critical for proper physiological functions of the cells. As such, TDP-43 expression is tightly regulated through an autoregulatory negative feedback loop. TDP-43 is a major disease-causing protein in Amyotrophic Lateral Sclerosis (ALS) and Frontotemporal Lobar Degeneration (FTLD). Several studies argue for a pathogenic role of elevated TDP-43 levels in these disorders. Modulating the cycle of TDP-43 production might therefore provide a new therapeutic strategy. In this study, we developed a new transgenic Drosophila model mimicking the TDP-43 autoregulatory feedback loop in order to identify genetic modulators of TDP-43 protein steady-state levels in vivo. First, we showed that our TDP-43_TDPBR Drosophila model recapitulates key features of the TDP-43 autoregulatory processes previously described in mammalian and cellular models, namely alternative splicing events, differential usage of polyadenylation sites, nuclear retention of the transcript and a decrease in steady-state mRNA levels. Using this new Drosophila model, we identified several splicing factors, including SF2, Rbp1 and Sf3b1, as genetic modulators of TDP-43 production. Interestingly, our data indicate that these three RNA-binding proteins regulate TDP-43 protein production, at least in part, by controlling mRNA steady-state levels.

Introduction

Encoded by the TARDBP gene, TAR DNA-binding protein-43 (TDP-43) is an evolutionarily conserved member of the hnRNP (heterogeneous nuclear ribonucleoprotein) family of proteins (1,2). Like other members of this family, TDP-43 has been linked to numerous aspects of the mRNA life cycle, including transcription, pre-mRNA splicing, mRNA transport, mRNA stability and mRNA translation (3). TDP-43 also regulates non-coding RNAs (miRNAs, lncRNAs, etc.). TDP-43 protein predominantly resides in the nucleus, but is capable of nucleocytoplasmic shuttling (4,5).

Numerous studies showed that perturbation of TDP-43 levels by either increasing or decreasing TDP-43 in animal and cellular models results in severe consequences (6–10). For example, loss of TDP-43 is detrimental for early mouse embryogenesis. In zebrafish and invertebrates, this loss also causes developmental abnormalities leading to neurological dysfunction, behavioral alterations and in most cases premature death. On the other hand, studies have provided evidence that TDP-43 overexpression causes neurodegeneration in various model systems, including cell culture, C. elegans, Drosophila, zebrafish and rodents. In Drosophila and C. elegans, human TDP-43-mediated neurodegeneration was associated with impaired locomotive activity, paralysis and reduced life span, whereas in mice it presented as a very rapidly progressing paralysis. In addition, overexpression of wild-type TDP-43 in primary neurons led to cell death. Thus, maintaining normal TDP-43 protein levels is critical for proper physiological functions of the cells.

TDP-43 expression is under very tight control in order to maintain homeostatic levels. TDP-43 protein can modulate its own protein levels through a negative feedback loop triggered by binding to its own RNA in the 3′ UTR region (11–15). This autoregulatory process depends on a sequence called TDP-43 binding region (TDPBR) that contains a certain number of low-affinity binding sites for TDP-43. The model proposed holds that in steady-state conditions, most TDP-43 production within cells comes from the transcript that uses the polyadenylation site pA1. When TDP-43 concentration rises, increased binding of TDP-43 proteins on the TDPBR region promotes the splice of a normally silent intron that contains the TDPBR region and the pA1 sequences, forcing the system to use suboptimal polyadenylation sites. mRNAs using these alternative polyadenylation sites were shown to be partially retained in the nucleus and rapidly degraded.

The recognition of TDP-43 involvement in neurodegenerative diseases started in 2006, when it was reported that TDP-43 represents the major constituent of pathologic inclusions in the affected neurons of patients with amyotrophic lateral sclerosis (ALS) and frontotemporal lobar dementia (FTLD) (16,17). Besides FTLD-TDP and ALS, some degree of neuronal TDP-43 pathology was subsequently found in a range of additional neurodegenerative diseases (18), including Alzheimer’s disease (≤60% of the patients). TDP-43 proteins not only represent a hallmark pathological feature of ALS/FTLD, but missense mutations in the TARDBP gene are sufficient in causing ALS/FTLD (19), demonstrating that TDP-43 can directly trigger neurodegeneration. However, at present, the molecular basis underlying TDP-43 toxicity and the pathobiology of TDP-43 mutations remain to be elucidated.

Several studies have observed an increase in TDP-43 mRNA and protein levels in various tissues of patients suffering from FTLD or ALS (20–27). Additionally, TDP-43 mutant proteins show various degrees of prolonged half-life and enhanced stability (28,29), which could lead to an elevated steady-state level of TDP-43 proteins (30,31). Altogether, the observations that the level of TDP-43 is elevated in tissue samples from patients, the fact that TDP-43 expression is tightly controlled by autoregulatory mechanisms, and the strong evidence that overexpression of TDP-43 protein level is detrimental to central nervous system cells and can cause cell degeneration, argue for a pathogenic role of elevated TDP-43 levels. Breaking the cycle of TDP-43 production might therefore provide a new therapeutic strategy.

In this study, we used a new Drosophila model of TDP-43 proteinopathies to identify genetic modulators of TDP-43 protein steady-state levels in vivo. To our knowledge, all TDP-43 Drosophila models developed so far were based on the expression of human TDP-43 without its 3′ UTR region. The expression level of TDP-43 protein in these animal models was therefore independent of the autoregulatory process. We generated a new Drosophila transgenic model that expresses human TDP-43 complementary DNA under the control of the TDPBR region. We first demonstrated that the TDPBR region possesses a strong negative regulatory activity in flies, and that this regulation is specifically dependent on TDP-43 protein expression. Furthermore, we showed that our TDP-43_TDPBR Drosophila model recapitulates key features of the self-regulatory process of TDP-43 protein steady-state levels described previously in mammalian and cellular models, namely alternative splicing events, differential usage of polyadenylation sites, nuclear retention of the transcript and a decrease in steady-state mRNA levels. Using this new Drosophila model, we identified the splicing factors SF2, Rbp1 and Sf3b1 as genetic modulators of TDP-43 production. Interestingly, our data indicate that these three RNA-binding proteins regulate TDP-43 protein production, at least in part, by controlling steady-state mRNA levels, using distinct molecular mechanisms.

Results

The TDPBR region possesses strong negative regulatory activity in Drosophila

First, we investigated potential regulatory activities of the TDPBR region in flies by testing its ability to influence the expression of a heterologous transcript. We used the Gal4/UAS bipartite expression system that relies on transcriptional activation by the yeast protein Gal4 (32). We cloned the TDPBR sequence downstream of the GFP (Green Fluorescent Protein) reporter complementary DNA (Supplementary Material, Fig. S1A and B), and generated new UAS Drosophila transgenic lines carrying this UAS-GFP_TDPBR reporter transgene. The GFP expression was directed to retinal cells using the Glass Multiple Reporter (GMR)-Gal4 line that drives expression in all cells of the developing and adult eyes, including the photoreceptor neurons as well as accessory pigment cells. As expected, Western blot analysis of protein extracts from newly eclosed adult heads revealed a single band with an apparent molecular mass of ∼30 kDa that corresponds to the predicted size of the GFP sequence (Fig. 1A, control).

Figure 1

TDP-43-dependent negative regulatory activity of the TDPBR region. (A) Western blot analyses of proteins extracted from transgenic flies co-expressing the UAS-GFP_TDPBR reporter construct and TDP-43, β-galactosidase (LacZ) or FUS proteins in retinal cells, under the control of the GMR-Gal4 driver. Control flies: GMR-Gal4 > UAS-GFP_TDPBR. Blots were first probed with an anti-GFP antibody. Then, expression of TDP-43, FUS and LacZ proteins was qualified using specific antibodies. Expression of TDP-43 proteins significantly decreases GFP accumulation level (n = 5, P = 0.0004), whereas FUS (n = 5, P = 0.6805) and LacZ (n = 5, P = 0.9706) proteins do not. (B) Western blot analyses of transgenic flies co-expressing the UAS-GFP_TDPBR or UAS-GFP reporter construct with or without TDP-43 protein under the control of the GMR-Gal4 driver. Blots were probed with an anti-GFP or anti-TDP-43 antibody. TDP-43 expression down-regulates the GFP reporter including the TDPBR region (n = 4, P = 0.0007) but had no repressive effect on the UAS-GFP control transgene. A slight but significant increase in GFP production is detected in the absence of the TDPBR region (n = 4, P = 0.0398). (A, B) Representative blots are shown. Total protein was used as loading control with Stain-free (SF) technology and the normalized expression of the GFP protein is reported in the graphs [mean ± standard error of the mean (SEM)]. Controls were arbitrarily set at 100 arbitrary units. GFP protein levels were compared between both genotypes by using Student’s t-test.

Because the regulatory activities of the TDPBR region have been found to depend on TDP-43 binding, we next tested the outcome of TDP-43 expression on GFP accumulation. For this purpose, we developed new UAS Drosophila transgenic lines expressing an untagged wild-type form of human TDP-43 protein. As previously described (33), we observed that the expression of TDP-43 in the eye of fly caused a dose-dependent disruption of retinal structure (Supplementary Material, Fig. S2)—weak expression induced no discernable phenotype, whereas strong expression was associated with severe morphologic degeneration. Unless otherwise stated, to minimize the influence of TDP-43 mediated-cellular toxicity on the protein’s steady-state levels, we selected the UAS-TDP-43#2 transgenic line that expresses a modest level of TDP-43 protein and that did not display a retinal phenotype. When co-expressed with TDP-43, GFP levels significantly decreased (P = 0.0004) (Fig. 1A). The expression of GFP protein was reduced by approximately 60%. Normalization of the amount of GFP proteins was achieved using the Stain-free technology, a method based on total protein quantification that is more reliable, more robust and more sensitive to minute changes than using a single housekeeping protein as a loading control (34). In contrast, LacZ (P = 0.9706) or FUS (P = 0.6803) expression did not significantly modify the amount of GFP (Fig. 1A), arguing that this regulation is specifically dependent on TDP-43 protein expression. Correct expression of TDP-43, LacZ and FUS proteins was confirmed by using specific antibodies (Fig. 1A). Importantly, overexpression of TDP-43 had no repressive effect on the UAS-GFP control transgene (no TDPBR region, Supplementary Material, Fig. S1A) (Fig. 1B), highlighting that the TDPBR region is necessary to mediate downregulation of gene expression by TDP-43 overexpression. We even detected a slight but significant increase in GFP production in the absence of the TDPBR region (P = 0.0398), the reason remaining to be established. Because many experiments have underlined a high degree of functional similarity and interchangeability between the human TDP-43 and its fly ortholog called TBPH (TAR DNA-binding protein-43 homolog) (35), we also examined the regulatory effect of TBPH proteins on the UAS-GFP_TDPBR reporter construct expression. Two previously described UAS-TBPH transgenic lines were tested (36,37). As shown in Supplementary Material, Figure S3A and B, similarly to human TDP-43 proteins, TBPH over-expression resulted in a significant reduction in GFP protein production.

Our findings that the TDPBR region confers negative regulatory activity on the heterologous transgene, prompted us to investigate whether TDP-43 is also subject to regulation by the TDPBR region in vivo. We cloned the TDPBR sequence downstream of the complementary DNA encoding the untagged wild-type form of human TDP-43 (TDP-43_TDPBR) (Supplementary Material, Fig. S1A), and then generated UAS-TDP-43_TDPBR Drosophila transgenic lines. To allow comparison between UAS-TDP-43 transgenic lines described above and those that do not contain the TDPBR region, both transgenes were inserted at the same genomic site using the PhiC31 site-specific integration system. Using the GMR-Gal4 driver line, we targeted TDP-43 expression to retinal cells. The solubility of TDP-43 proteins was addressed by carrying out sequential extraction using RIPA (soluble) and urea (insoluble) buffers. Western blot analysis revealed in the RIPA-soluble fraction a single band with an apparent molecular mass of ∼43 kDa that corresponds to the predicted size of the 414 amino acid TDP-43 sequence (Fig. 2A). We found that the presence of the TDPBR region causes a drastic reduction in steady-state protein levels. After longer exposure times, only a very faint signal was detected in the RIPA-soluble fraction. We estimated signal intensity to be <3%, compared to that observed in UAS-TDP-43 transgenic lines. No increase in TDP-43 immunoreactivity was detected in the urea fraction, demonstrating that TDP-43 proteins were not trapped in insoluble structures in our Drosophila model. As expected, the expression level of the TDP-43_TDPBR transgene was insufficient to induce a retinal phenotype (data not shown). Interestingly, we found that reduction of TBPH expression in the context of the TDPBR region by specific RNAi induced only a slight increase in TDP-43 production (Supplementary Material, Fig. S3C and D), indicating that TBPH proteins are not key factors in the negative regulatory activity of the TDPBR region in vivo.

Figure 2

The TDPBR region activates alternative splicing, differential usage of polyadenylation sites, and a decrease in TDP-43 steady-state mRNA levels. (A) Western blot analyses of proteins extracted from transgenic flies that expressed the UAS-TDP-43 or the UAS-TDP-43_TDPBR reporter construct under the control of the GMR-Gal4 driver. Control flies: GMR > +. Blots were probed with an anti-TDP-43 antibody. Representative blots corresponding to the RIPA (soluble) and the Urea (insoluble) fractions are shown (n > 4). Short (S.E) and long (L.E) exposures are presented. Total protein was used as the loading control with Stain-free (SF) technology. The presence of the TDPBR region causes a drastic reduction in TDP-43 steady-state protein levels. (B) Determination of TDP-43 mRNA level by qRT-PCR or RT-QMPSF experiments. RNA was extracted from GMR > + (control), GMR > TDP-43_TDPBR or GMR > TDP-43 Drosophila heads. The graph represents the data obtained with the TDP-43^F1/R1 amplicon (Supplementary Material, Fig. S4) and the reference gene Cyp1. TDP-43 expression levels were compared between both genotypes by using Student’s t-test. The presence of the TDPBR region results in a significant reduction in TDP-43 steady-state mRNA levels (n = 4, p^qRT-PCR = 0.0006, p^RT-QMPSF = 0.0049). (C) The PCR result for 3′ RACE experiments performed using RNA described in (B). Molecular marker: 1 kb DNA ladder. Representative gel image is shown (n > 10). As expected, no PCR product is detected in control flies or when RNA samples were not retro-transcribed (-RT). In contrast to TDP-43 transgenic flies that expressed a single TDP-43 mRNA product, TDP-43_TDPBR transgenic flies displayed a complex pattern of 3′ RACE PCR products. (D) Schematic representation of TDP-43±TDPBR mRNA variants detected in flies. Note that TDP-43 cDNA was subcloned in the pUAST transgenesis vector that includes the SV40 small-t intron and polyadenylation site (pA) (SV40 term). (E) Quantification of the TDP-43 mRNA levels in nuclear and cytoplasmic fractions by RT-QMPSF experiments. RNA was extracted from GMR > TDP-43_TDPBR or GMR > TDP-43 Drosophila heads. The graph represents the data obtained with the TDP-43^F1/R1 amplicon and the reference gene Cyp1. TDP-43 expression levels were compared by using Student’s t-test (n = 5, P = 0.0314).

All together, these results show that the TDPBR region possesses a strong negative regulatory activity in Drosophila, and that this regulation is specifically dependent on TDP-43 protein expression.

The TDPBR region activates alternative splicing, use of different polyadenylation signals, nuclear retention and favours mRNA degradation in Drosophila

The proposed mechanisms for TDP-43 autoregulation link alternative splicing, use of different polyadenylation signals, nuclear retention and degradation to regulate the abundance of mRNA transcripts (11–15). To evaluate the relevance of our Drosophila models, we first quantified TDP-43 steady-state mRNA levels in our TDP-43 ±TDPBR Drosophila models, using two independent methods: (1) the qRT-PCR (quantitative reverse transcription-polymerase chain reaction) method, the gold standard method for the detection of RNA levels and (2) a simple multiplex RT-PCR assay (RT-QMPSF, reverse transcription-quantitative multiplex PCR of short fluorescent fragments) that has been previously developed in our group (38) (Supplementary Material, Fig. S4). This assay is based on simultaneous PCR amplification of short fluorescent fragments and allows for comparative quantitative analysis of mRNA. Total RNA was extracted from heads of GMR > +, GMR > TDP-43 or GMR > TDP-43_TDPBR transgenic adult flies. As shown in Figure 2B, both methods provided similar results. The TDP-43_TDPBR flies displayed a statistically significant decrease in TDP-43 mRNA levels (qRT-PCR: ∼65%, P = 0.0006; RT-QMPSF: ∼57%, P = 0.0049) as compared to TDP-43 flies, demonstrating that the presence of TDPBR region lowered TDP-43 steady-state mRNA levels. Importantly, for RT-QMPSF experiments, we obtained similar data using two different TDP-43 amplicons (F1/R1 and F2/R2) and two reference genes (RpL13A and Cyp1) (Supplementary Material, Fig. S4). Because the RT-QMPSF is a rapid, inexpensive and easily quantified assay for mRNA levels, this method was therefore used for the remainder of the study.

We next examined whether TDP-43 autoregulation in flies occurs via alternative splicing and differential usage of polyadenylation sites, performing 3′ Rapid Amplification of cDNA Ends (RACE) experiments. PCR products were amplified using human TDP-43-specific primer and an oligo-dT adapter primer, and then cloned and sequenced. The 3′ RACE experiments revealed that the UAS-TDP-43 transgenic flies expressed a single TDP-43 mRNA product (Fig. 2C and D). In contrast, flies expressing the UAS-TDP-43_TDPBR transgene displayed a complex pattern of 3′ RACE PCR amplification (Fig. 2C). Sequence analysis confirmed the presence of at least 12 alternative transcripts of different sizes resulting from alternative splicing events and/or differential usage of polyadenylation sites (Fig. 2D).

To determine the subcellular localization of the various TDP-43 mRNA isoforms, and therefore their potential nuclear retention, we performed cell fractionation on GMR > TDP-43 or GMR > TDP-43_TDPBR transgenic flies. Total RNAs were then extracted from nuclear and cytoplasmic fractions (Supplementary Material, Fig. S5). Quantification of TDP-43 steady-state mRNA levels was achieved by RT-QMPSF experiments. If all TDP-43 mRNA isoforms were similarly distributed between the nuclear and cytoplasmic fractions, we would expected similar TDP-43/TDP-43_TDPBR ratios of TDP-43 mRNA levels in both compartments. However, we detected a significantly higher ratio in the nuclear compartment (P = 0.0314) (Fig. 2E), indicating that the various TDP-43 mRNA isoforms produced in the context of the TDPBR region are subject to nuclear retention events in Drosophila. Altogether, these data showed that our TDP-43_TDPBR Drosophila model recapitulates key features of the self-regulatory process of the steady-state levels of TDP-43 proteins described previously in mammalian and cellular models, namely alternative splicing events, differential usage of polyadenylation sites, nuclear retention of the transcripts, and a decrease in steady-state mRNA levels.

The SF2, Rbp1 and Sf3b1Drosophila genes act as genetic modulators of TDP-43 production

In order to reveal genetic components potentially linked to TDP-43 production, we crossed selected stocks with GMR > TDP-43_TDPBR flies and assessed the F1 progeny for TDP-43 production by Western blot. Since it has been shown that the TDP-43 autoregulatory feedback loop involves alternative splicing, the use of different polyadenylation signals, and nuclear retention in mammalian and cellular models, we focused on seven genes known for their multifaceted functions in these events (39,40). All these genes act as genetic modulators of TDP-43 production in vivo (Table 1). Interestingly, six of them encode members of the family of serine/arginine (SR)-rich proteins. For the continuation of the study, we decided to focus on the three that presented the strongest effects on TDP-43 production in vivo, namely SF2 (Splicing factor 2), Rbp1 (RNA-binding protein 1) and Sf3b1 (Splicing factor 3b subunit 1). As shown in Figure 3A, downregulation of Sf3b1 using RNA interference (RNAi) or expression of GFP-tagged Rbp1 (GFP::Rbp1) proteins resulted in a drastic increase of TDP-43 protein steady-state level. Expression of SF2 (GFP::SF2) also induced TDP-43 accumulation, and importantly, the opposite effect was observed by its knock-down. Indeed, expression of SF2 RNAi using two different UAS transgenic lines consistently caused a complete loss of the TDP-43 signal. The upper band, detected when GFP::SF2 is overexpressed, probably corresponds to a dimeric TDP-43 species. We also tested the effect of the overexpression of an untagged form of human SF2/SRSF1 (41) on TDP-43 production and observed a strong increase of TDP-43 steady-state level (Supplementary Material, Fig. S6). Note that RNAi-mediated downregulation of Rbp1 did not modify TDP-43 accumulation level (Fig. 3A). This absence of effect could be due to the mutual negative feedback between Rbp1 and Rbp1-like (Rbp1L), the two Drosophila orthologs of mammalian SRp20: knock-down expression of Rbp1 causes a compensating increase in Rbp1L mRNA and protein (42,43). Expression of GFP::SF2 and GFP::Rbp1 was confirmed by Western blot (Supplementary Material, Fig. S7A), whereas quantitative real-time PCR (qRT-PCR) was used to validate the RNAi-mediated decrease of Rbp1 (Supplementary Material, Fig. S7B), SF2 (Supplementary Material, Fig. S7C) and Sf3b1 (Supplementary Material, Fig. S7D). Note that modulation of SF2, Rbp1 and Sf3b1 expression did not change LacZ steady-state level (Fig. 3B and C), indicating a specific effect of these factors regarding TDP-43 expression.

Table 1

Modulation of splicing factors expression affects TDP-43 accumulation levels

Human homolog	Drosophila gene	Overexpression	Loss-of-function allele	Effect on TDP-43 accumulation	Reference
SF3B1	Sf3b1/CG2807		UAS-IR Sf3b1	Increase	VDRC#25162a
–	RSF/CG5655	UAS-RSF1		Increase	Labourier et al. (62)
SRSF4/SRSF6	B52/CG10851	UAS-B52		Increase	Kraus and Lis (61)
			UAS-IR B52	Decrease	VDRC#38862
SRSF7	xl6/CG10203	UAS-GFP::9G8		Increase	Gabut et al. (59)
SRSF2	SC35/CG5442	UAS-GFP::SC35		Increase	Gabut et al. (59)
SRSF3	Rbp1/CG17136	UAS-GFP::Rbp1		Increase	Unpublished
			UAS-IR Rbp1	No effect	VDRC#21083
SRSF1	SF2/CG6987	UAS-hSF2		Increase	Allemand et al. (41)
		UAS-GFP::SF2		Increase	Gabut et al. (59)
			UAS-IR SF2	Decrease	VDRC#27775
			UAS-IR SF2	Decrease	VDRC#27776

a

VRDC: Vienna Drosophila Resource Center stock number.

Figure 3

Modulation of Drosophila Sf3b1, SF2 and Rbp1 expression alters TDP-43 steady-state protein levels. Western blot analyses of proteins extracted from transgenic flies that expressed the UAS-TDP-43 _TDPBR (A), the UAS-TDP-43 (D) or UAS-LacZ (B, C) reporter construct under the control of the GMR-Gal4 driver. Control flies: GMR  > +. The steady-state level of the different reporter proteins were analyzed in different genetic contexts. Note that flies were raised at 18 °C to limit putative toxicity associated with the modulation of Sf3b1, SF2 and Rbp1 expression. Blots were probed with an anti-TDP-43 antibody and representative blots are presented. Total protein was used as the loading control by Stain-free (SF) technology. Modulation of SF2, Rbp1 and Sf3b1 expression altered TDP-43 steady-state protein levels (A, D), but did not change LacZ steady-state level (B, C), indicating a specific effect of these factors regarding TDP-43 expression. TDP-43 expression levels were compared by using Student’s t-test (B, C) (n = 2) Control versus GFP::Rbp1 (P = 0.5731), GFP::SF2 (P = 0.5688), RNAi Sf3b1 (P = 0.2659). (D) Control versus RNAi Sf3b1 (P = 0.0423), GFP::Rbp1 (P = 0.8329), RNAi Rbp1 (P = 0.1128), GFP::SF2 (P = 0.2079), RNAi SF2²⁷⁷⁷⁵ (P = 0.0663), RNAi SF2²⁷⁷⁷⁶ (P = 0.0379).

SF2, Rbp1 and Sf3b1-mediated regulation of TDP-43 protein production depends on the presence of the TDPBR region

These three Drosophila RNA-binding proteins and their human orthologs are multifaceted regulators of gene expression. They notably influence mRNA splicing, export, stability and translation (44,45). To address the molecular mechanisms underlying these genetic interactions, we first determined whether the TDPBR region contributes to the SF2, Rbp1 and Sf3b1-mediated regulation of TDP-43 protein production, using UAS-TDP-43 transgenic lines (no TDPBR region) (Fig. 3D). Similarly to what we observed with the UAS-TDP-43_TDPBR construct (Fig. 3A), down-regulation of Sf3b1 significantly increased TDP-43 protein steady-state level (P = 0.0423). Expression of GFP::SF2 protein also consistently augmented TDP-43 protein steady-state level, whereas down-regulation of SF2 reduced TDP-43 production. However, except for the SF2 RNAi²⁷⁷⁷⁶ line (P = 0.0379), these effects did not reach statistical significance (GFP::SF2: P = 0.2079, SF2 RNAi²⁷⁷⁷⁵: P = 0.0663). In addition, it should be stressed that the levels of increase in TDP-43 production were much less important than those observed using the UAS-TDP-43_TDPBR transgene (Fig. 3A and D), suggesting that SF2 and Sf3b1-mediated regulation of TDP-43 protein production predominantly depends on the presence of the TDPBR region. On the other hand, modulation of Rbp1 expression did not significantly change TDP-43 steady-state level (P = 0.8329), suggesting that Rbp1-mediated control of TDP-43 protein production strictly requires the presence of the TDPBR region. Together, these results showed that the regulation of TDP-43 production by SF2, Rbp1 and Sf3b1 depends on the presence of the TDPBR region in our experimental model, with additional mechanisms of regulation for SF2 and Sf3b1.

Modulation of TDP-43 steady-state mRNA levels by SF2, Rbp1 or Sf3b1 factors

We next evaluated whether changes in TDP-43 steady-state mRNA levels could account for the observed modulation at the protein level. RT-QMPSF experiments showed that modulation of SF2, Rbp1 or Sf3b1 expression resulted in a significant increase of TDP-43_TDPBR mRNA level compared to the control (Fig. 4A). Furthermore, consistent with the data observed at the protein level (Fig. 3D), only modulation of SF2 or Sf3b1 expression consistently gave rise to an increase of TDP-43 mRNA level compared to the control (Fig. 4B). The overexpression of Rbp1 did not. All together, these results showed that SF2, Rbp1 or Sf3b1 factors modulate TDP-43 steady-state mRNA levels.

Figure 4

Modulation of Drosophila Sf3b1, SF2 and Rbp1 expression alters TDP-43 steady-state mRNA levels. Determination of TDP-43±TDPBR mRNA levels by RT-QMPSF experiments. The levels of expression of the UAS-TDP-43_TDPBR (A) or the UAS-TDP-43 (B) reporter constructs were analyzed in different genetic contexts (UAS-GFP::SF2, UAS-GFP::Rbp1 and RNAi Sf3b1). Driver line: GMR-Gal4. The graph represents the data obtained with the TDP-43^F1/R1 amplicon and the reference gene Cyp1. TDP-43 expression levels were compared between both genotypes by using Student’s t-test. (A) (n > 5) Control TDP-43_TDPBR versus GFP::SF2 (P = 0.0149), GFP::Rbp1 (P = 0.0432), RNAi Sf3b1 (P = 0.0063). (B) (n > 3) Control TDP-43 versus GFP::SF2 (P = 0.0311), GFP::Rbp1 (P = 0.0398), RNAi Sf3b1 (P = 0.0544).

Effect of the modulation of TDP-43 protein production by SF2, Rbp1 or Sf3b1 factors on TDP-43 phenotypic severity

Lastly, we assayed whether SF2, Rbp1 and Sf3b1-mediated regulation on TDP-43 protein production correlates with TDP-43 phenotypic severity. We first tested the outcome of SF2 expression knock-down on TDP-43-mediated neurodegeneration (Fig. 5A). We used the UAS-TDP-43#1 transgene that induced evident defects in Drosophila eyes. These TDP-43-induced defects were significantly reduced in eyes co-expressing TDP-43 with SF2 RNAi. This reduction of TDP-43-mediated neurodegeneration correlated with a decrease in TDP-43 protein accumulation level (Fig. 5B). Note that compared to control eyes, we did not observe any significant structural defects in eyes expressing SF2 RNAi. We have also tested the consequence of the expression of GFP::SF2 proteins or the downregulation of Sf3b1 expression on TDP-43-mediated neurodegeneration (Supplementary Material, Fig. S8). Remember that these two genetic contexts drive an increase on TDP-43 production (Fig. 3D). We used the UAS-TDP-43#2 transgene that induced no discernable phenotype by itself, and looked for the appearance of structural defects in eyes. Unfortunately, expression of GFP::SF2 or down-regulation of Sf3b1 alone produces a strong rough eye phenotype, making the results complex to interpret. However, we observed that TDP-43 co-expression significantly enhanced these defects. When Rbp1::GFP was expressed, no obvious phenotype was noticed. Consistent with our data showing that modulation of Rbp1 expression did not significantly change TDP-43 steady-state level in the context of the UAS-TDP-43 transgene (Fig. 3D). TDP-43 co-expression did not induce morphological changes in Drosophila eyes. All together, these results suggest that regulation of TDP-43 protein production by SR proteins could correlate with TDP-43 phenotypic severity.

Figure 5

Modulation of TDP-43 protein production correlates with TDP-43 phenotypic severity. (A) Light micrographs of new-born Drosophila adult eyes. Compared to control flies (GMR > +), SF2 RNAi expression triggered no structural defects (GMR > RNAi SF2²⁷⁷⁷⁵ or GMR > RNAi SF2²⁷⁷⁷⁶). In contrast, TDP-43 expression induced evident defects (GMR > TDP-43), which were suppressed by expressing SF2 RNAi (GMR > TDP-43/RNAi SF2²⁷⁷⁷⁵ or GMR > TDP-43/RNAi SF2²⁷⁷⁷⁶). (B) Western blot analysis of protein extracts from heads of the flies described in (A). Blots were probed with an anti-TDP-43 antibody. Representative blot is shown. Total protein was used as loading control with Stain-free (SF) technology and the normalized expression of the TDP-43 protein is reported in the graphs [mean ± standard error of the mean (SEM)]. Controls were arbitrarily set at 100 arbitrary units. TDP-43 protein levels were compared between both genotypes by using Student’s t-test. (n = 4) Control TDP-43 versus TDP-43/RNAi SF2²⁷⁷⁷⁵ (P = 0.0003), TDP-43/RNAi SF2²⁷⁷⁷⁶ (P = 0.0006).

Discussion

To identify genetic modulators of TDP-43 production in vivo, we first developed and characterized new Drosophila transgenic models that express human TDP-43 cDNA under the control of the TDPBR sensor region. In this study, we showed that the TDPBR region by itself, independently of additional sequences present in the 3′ UTR of TDP-43, displays a very strong negative autoregulatory activity in Drosophila, and that this regulation is specifically dependent on TDP-43 protein expression (Figs 1 and 2). Furthermore, our TDP-43_TDPBR Drosophila model recapitulates key features of the self-regulatory process of the steady-state levels of TDP-43 proteins described previously in mammalian and cellular models, namely alternative splicing events, differential usage of polyadenylation sites, nuclear retention of the transcripts, and a decrease in steady-state mRNA levels. Together, these data highlight the relevance of our experimental models (Fig. 2).

Performing quantitative analysis, we found that the presence of the TDPBR region lowered TDP-43 steady-state mRNA levels by ∼60%, but decreased TDP-43 level by about 97%. Furthermore, the study of TDP-43 mRNA subcellular localization revealed that these mRNAs are only partially retained in the nucleus, and that a significant amount of TDP-43 mRNA is present in the cytoplasm. Together, our data showed that the decrease in steady-state levels and the nuclear retention of TDP-43 mRNA do not fully account for the down-regulation seen at the protein level, suggesting that additional mechanisms are involved in TDP-43 autoregulation in our experimental models. Interestingly, several studies have suggested that one of the functions of TDP-43 in neurons is to regulate translation. It has been reported that TDP-43 localizes on ribosomes in primary neurons and in SH-SY5Y human neuroblastoma cells, and that an increase in cytoplasmic TDP-43 represses global protein synthesis (46). In Drosophila, TDP-43 associates with futsch mRNA in vivo and regulates its translation (47). More recently, it has been shown that TDP-43 acts as an adaptor protein to recruit the FMRP-CYFIP1 inhibitory complex to mRNAs, thereby repressing the initiation of translation (48). Thus, our study points out that TDP-43 binding on the TDPBR region could have an additional effect on TDP-43 production namely, translational repression.

Using our TDP-43_TDPBR Drosophila model, we have identified six genes (Rsf1, B52, x16, SC35, Rbp1, SF2) encoding members of the family of serine/arginine (SR)-rich proteins and the gene Sf3b1 as genetic modulators of TDP-43 production in vivo. In this study, we focused on SF2, Rbp1 and Sf3b1 because they have the strongest effect on TDP-43 production. The SF2, Rbp1 and Sf3b1 genes are the Drosophila orthologs of human SRSF1 (Serine/arginine-rich splicing factor 1), SRSF3 (Serine/arginine-rich splicing factor 3) and SF3B1 (Splicing factor 3B subunit 1) genes, respectively. The SR proteins SRSF1 and SRSF3 play multiple roles in eukaryotic gene expression pathways. If initially described as splicing factors, it is now clear that these factors are also implicated in RNA transcription, mRNA transport, stability and translation (45). Regarding the gene SF3B1, previous studies in Drosophila identified Sf3b1 as a spliceosomal component (49) and established a role of this factor in the splicing and subsequent export of transcript (44). In mammals, SF3B1 has been shown to drive spliceosome assembly, branch site selection and splicing (50,51).

Uncovering the mechanism of action of these factors on TDP-43 production is complicated by their multifaceted functions. We have only just begun to address these issues in this study. However, our data suggest the three RNA-binding proteins use distinct mechanisms to control TDP-43 production in our experimental model. Indeed, whereas Rbp1-mediated regulation on TDP-43 production is strictly dependent on the presence of the TDPBR region, the control on TDP-43 production by SF2 and Sf3b1 is only partially dependent on the sensor region (Fig. 3). The contribution of the TDPBR region is predominant, but additional mechanisms of regulation must be solicited in SF2 and Sf3b1 activity. Furthermore, if modulation of SF2, Rbp1 and Sf3b1 expression leads to a similar increase in TDP-43 protein production, the TDP-43 mRNA levels observed are more variable. SF2 expression and Sf3b1 down-regulation lead to a higher increase in TDP-43 steady-state mRNA levels compared to Rbp1 expression (Fig. 4), suggesting additional mechanisms of regulation for Rbp1.

Interestingly, while the SRSF1 and SRSF3 factors share numerous similar functions in RNA metabolism, several studies highlighted features that distinguish one from the other. First, HITS-CLIP analyses demonstrated that their consensus binding sites and their CLIP tag distribution across transcripts are distinct (40,52,53). On the other hand, if both SRSF1 and SRSF3 promote the export of mRNAs through their direct binding to the canonical mRNA export factor NXF1 (nuclear RNA export factor 1), a very recent study demonstrated that SRSF3 interacts most robustly with NXF1 and had the highest number of candidate mRNA export targets of all of the SR proteins (54). The authors also showed that SRSF3 preferentially recruits NXF1 to last exons and/or 3′ UTR. Lastly, Müller-McNicoll et al. (54) found that SRSF3 depletion induced 3′ UTR shortening and suggested that SRSF3 couples nuclear mRNA export to proper splicing and 3′ end maturation. Therefore, Rbp1 expression in flies could promote TDP-43_TDPBR mRNA nuclear export by recruiting NXF1 to the TDPBR sensor region. Enhancement of TDP-43_TDPBR mRNA transport towards the cytoplasm would consequently allow for an increase in protein production and a reduction in mRNA degradation. Of course, if the Rbp1-mediated regulation of TDP-43 production could at least in part be due to an increase in TDP-43_TDPBR mRNA nuclear export, Rbp1 expression could also influence other aspects of mRNA metabolism (alternative splicing, the use of alternative polyadenylation signals…).

As mentioned above, we found that SF2, Rbp1 and Sf3b1 factors regulate TDP-43 protein production by controlling steady-state mRNA levels. Again, future studies will be required to establish the mechanism through which these factors enhance TDP-43 production. However, we can note that the overexpression of both SF2 and SRSF1 proteins modulates TDP-43 production in flies (Fig. 3; Supplementary Material, Fig. S6). Interestingly, in contrast to its human ortholog SRSF1, the Drosophila SF2 proteins are concentrated in the nucleus and do not shuttle between the nucleus and cytoplasm (41), suggesting that the mechanism of action of SF2 on TDP-43 production involves only nuclear functions of the protein (pre-mRNA splicing, polyadenylation site selection…) in our experimental model. It will also be important to determine if the effects of SF2, Rbp1 and Sf3b1 factors on TDP-43 production are direct (depend on their binding to TDP-43 mRNA) or not (alteration of factors involved in the production of TDP-43). It should be stressed that SRSF3 crosslink sites were found in TDP-43 mRNA (52). Another aspect that will be interesting to uncover is the contribution of the TDP-43 protein itself in these processes of regulation. Indeed, proteomic studies performed on TDP-43 in several cell lines have detected a huge number of potential TDP-43 interacting partners, including the splicing factors SRSF1, SRSF3, SRSF7 and SF3B1 (55–57).

To conclude, we developed a new Drosophila model that recapitulates numerous key features of the TDP-43 auto-regulatory feedback loop. Using this new experimental model, we have identified splicing factors as modulators of TDP-43 production in vivo. The functions of these modifiers are fully consistent with the proposed mechanisms for TDP-43 autoregulation in mammals, highlighting the relevance of our model in identifying new genetic components potentially linked to TDP-43 production. It will now be important to unravel these factors’ mechanisms of action on TDP-43 production. Nevertheless, regardless of underlying mechanisms, the ability of SF2/SRSF1, Rbp1/SRSF3 and SF3B1 proteins to regulate TDP-43 production has demonstrated their role and relevance in studies on TDP-43 proteinopathies. Indeed, as TDP-43 functions in a feedback loop regulating its own expression, it is possible that a modest downregulation of single or multiple SR protein regulators would be sufficient to restore appropriate levels of TDP-43 proteins, without fully disrupting other necessary functions of the splicing factors. In this way, subtle modulation of combinations of individual SR proteins could be therapeutic in TDP-43 proteinopathies.

Materials and Methods

An ethics statement is not required for this work.

DNA constructs and development of transgenic flies

Using the pUAST-TDP-43-Flag plasmid described in Miguel et al., (58), PCR-amplification of human TDP-43 cDNA was achieved with the primers 5′-CGCAGGGCCGGACGGGCCCAAAATGTCTGAATATATTCGGGTAACCG-3′ and 5′-CGCAGGGCCCCAGTGGCCCTACATTCCCCAGCCAGAAGACTTAGAATCC-3′. The TDP-43 PCR products were next subcloned into the pUASTattB SfiI vector as a Sfi fragment and then sequenced. Amplification of the TDPBR region was achieved using a total of 2.5 ml of blood collected from a healthy donor directly into PAXgene Blood RNA Tubes (Qiagen, Valencia, CA, USA). Total RNA was extracted from the whole-blood sample using the PAXgene Blood RNA kit (Qiagen) and stored at −80 °C. RNA concentration was evaluated using the Nanovue (GE Healthcare, Chalfont St. Giles, UK). Three hundred nanograms of total RNA was reverse-transcribed into cDNA using the Verso cDNA kit with a blend of random hexamers and anchored oligo-dT (3:1) (Thermo Fisher Scientific Inc., Waltham, MA, USA). PCR-amplification of the TDPBR region was achieved using the primers 5′-GGGGCGGCCGCTCACAGGCCGCGTCTTTGACGGTGGG-3′ and 5′-CCGCTCG-AGCGGAAAACAAAGACACATATTATTTAAATCAG-3′. The PCR product was then subcloned into wild-type and mutant pUASTattB-TDP-43 vectors digested with NotI/XhoI, and sequenced. The GFP cDNA was PCR-amplified from pEGFP N3 (Clontech Laboratories Inc. Mountain View, CA, USA) using the primers 5′-CGCAGGGCCGGACGGGCCCAAAATGGTGAGCAAGGGCG-3′ and 5′-CGCAGGGCCCCAGTGGCCTTACTTGTACAGCTCGTCC-3′, and then subcloned into the pUASTattB SfiI vector as a Sfi fragment. The TDPBR fragment, PCR-amplified from the DNA constructs described above using the primers 5′-CGGAGATCTTCACAGGCCGCGTCTTTGACGGTGGG-3′ and 5′-GGGGCGGCCGCTCACAGGCCGCGTCTTT-GACGGTGGG-3′, was subsequently subcloned as a NotI/BglII fragment. Transgenic strains were generated by BestGene Inc. (Chino Hills, CA, USA), according to standard methods, using the y¹M ZH-2A w*; MZH-51C (cytological region 51C on the second chromosome, strain identifier at BestGene: 24482) as the recipient strain for TDP-43 constructs and the y¹M ZH-2A w*; M ZH-86Fb (cytological region 86F8 on the third chromosome, strain identifier: 24749) as the recipient strain for eGFP constructs. The UAS-GFP::Rbp1 line was established as described in (59).

Fly genetics

Drosophila were maintained on a 12:12 light/dark cycle on standard cornmeal-yeast agar medium at 25 °C unless otherwise stated. The following transgenic Drosophila strains were used in this study: UAS-FUS (60), UAS-GFP::SF2 (59). The GMR-Gal4 and UAS-LacZ lines were obtained from Bloomington Stock center. The Sf3b1 (ID 8837), Rbp1 (ID9289) and SF2 (ID38377 and ID38379) RNAi transgenic lines were obtained from the Vienna Drosophila Resource center (VDRC). Detailed fly genotypes are listed in Supplementary Material.

RNA isolation and reverse transcription

Total RNA extraction was performed on 30 newly eclosed adult fly heads using the Nucleospin RNA II kit (Macherey-Nagel GmbH and Co. KG, Düren, Germany) according to the manufacturers’ instructions. The amount of RNA per µl was measured on a NanoDrop Lite (Thermo Fisher Scientific Inc.). 1 µg of total RNA was reverse-transcribed into cDNA, using the RETROscript Reverse Transcription kit (Thermo Fisher Scientific Inc.) with Oligo(dT) primers according to the manufacturers’ instructions for 2-Step RT-PCR without heat denaturation.

Real-time quantitative RT–PCR

PCR reactions were performed in a final volume of 20 μl, using the SsoFast Evagreen Supermix (Bio-Rad Laboratories, Hercules, CA, USA) with primers at a final concentration of 250 nm (TDP-43: forward 5′-ATGGGTGGTGGGATGAACTTT-3′ and reverse 5′-CGATGGGCCTGACTGGTTCT-3′; Cyp1: forward 5′-TCGGCAGCGGCATTTCAGAT-3′ and reverse 5′-TGCACGCTGACGAAGCTAGG-3′). PCR amplifications were performed on a CFX96 Real-Time System thermal cycler (Bio-Rad Laboratories) using the following cycling steps: enzyme activation at 98 °C for 2 min; denaturation and annealing/extension, respectively, at 98 °C for 10 s and 60 °C for 15 s (40 cycles). The comparative −ΔΔCt method was then used to determine quantitative values for gene expression levels in each sample using Cyp1 as the housekeeping gene.

RT-QMPSF and data calculation

To specifically analyze the expression of TDP-43, targeted RT-QMPSF was developed as previously described (38). Briefly, short cDNA sequences (100–160 bp) were simultaneously PCR-amplified in a single tube, using two pairs of primers spanning 5′ (TDP-43 F1/R1) or 3′ (TDP-43 F2/R2) end of TDP-43, and two pairs of primers spanning two reference genes (RpL13A and Cyp1). Sense primers were 6-FAM-labeled. Primer sequences: TDP-43 F1: 5′-TGGGGAAATCTGGTGTATGTT-3′, TDP-43 R1: 5′-AATCGGATGTTTTCTGGACTG-3′, TDP-43 F2: 5′-ATGGGTGGTGGGATGAACTTT-3′, TDP-43 R2: 5′-CGATGGGCCTGACTGGTTCT-3′, Cyp1 F1: 5′-GCTCCCAGTTCTTTATTTG-3′, Cyp1 R1: 5′-AACCAGAGTTAGCCACAAT-3′, RpL13A F1: 5′-AGCTGAACCTCTCGGGACAC-3′, RpL13A R1: 5′-TGCCTCGGACTGCCTTGTAG-3′. All primers were used in a single PCR reaction volume of 25 μl. Multiplex fluorescent PCR assays were carried out using 2 mm MgCl2, 1 unit of Diamond Taq polymerase (Eurogentec), 8 mm of Triethylamine Acetate, 160 µm of dNTP, 0.3 µm of primers of TDP-43 F1/R1, 0.1 µm of primers of TDP-43 F2/R2, 0.06 µm of primers of RpL13A, 0.1 µm of primers of Cyp1 and 1 µl of cDNA. After an initial cycle of denaturation at 95 °C for 5 min, 22 cycles were performed consisting of denaturation at 95 °C for 20 s, annealing at 53 °C for 40 s, and extension at 72 °C for 20 s and final extension at 72 °C, in a DNA engine (Biorad T100) Peltier Thermal Cycler (Bio-Rad Laboratories, Hercules, CA). The number of cycles of amplification was determined by testing a range of cycle numbers in order to remain in the linear phase of the PCR. Fluorescent amplicons were separated on an ABI prism 3100 genetic analyser (Applied Biosystems, Foster City, California), and the resulting fluorescent profiles were analysed using GeneScan 3.7 software (Applied Biosystems). All QMPSF analyses were performed at least in triplicate. For comparative analyses, the average peak heights obtained for TDP-43 amplicons were compared to the mean peak height obtained for the control amplicon for each genotype. The ratio obtained is set at 100 for the control genotype (GMR > TDP-43). TDP-43 expression levels were compared between controls and each of the other genotype by using a Student’s t-test.

3′ RACE and TA cloning

The 3′ ends of the TDP-43 transcripts were determined using the 3′ rapid amplification of cDNA ends (RACE) system. RNA (100 ng) was annealed with a mix (0.5 μm of each) of anchored AdP Oligo(dT)₁₇ primers (5′-GGCCACGCGTCGACTAGTACTTTTTTTTTTTTTTTTTG-3′, 5′-GGCCACGCGTCGACTAGTACTTTTTTTTTTTTTTTTTA-3′ and 5′-GGCCACGCGTCGACTAGTACTTTTTTTTTTTTTTTTTC-3′) and RT–PCR with Superscript RT II in accordance with instruction (Invitrogen). The PCR was performed with 1 μl of the first-strand reaction, 0.5 μm of AdAmpP primer (5′-GGCCACGCGTCGACTAGAAC-3′), 0.5 μm of TDP-43 F5 primer (5′-GGATTCTAAGTCTTCTGGC-3′), using the FIREPol^® DNA polymerase (Solis Biodyne, Tartu, Estonia), as recommended by the manufacturer. A touchdown method was used with a DNA Engine (PTC-200) Peltier Thermal Cycler (Bio-Rad). Cycling times were: 3 min at 95 °C, followed by 40 cycles of denaturation at 95 °C for 10 s, annealing beginning at 65 °C and ending at 55 °C for 20 s, and extension at 72 °C for 1 min, with a final extension at 72 °C for 10 min. The annealing temperature was lowered 1 °C every cycle until it reached 55 °C; this annealing temperature was kept until the end of the cycling process. The resulting PCR products were subcloned using the TA Cloning^® Kit Dual Promoter (pCR^®II) (Invitrogen) and then sequenced.

Protein extraction

Thirty adult fly heads were dissected from newly eclosed adults and homogenized in 150 μl Radio Immunoprecipitation Assay (RIPA) buffer (25 mm Tris–HCl pH 7.6, 150 mm NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS) (Pierce Biotechnology, Rockford, IL, USA), supplemented with a cocktail of protease inhibitors (Sigma-Aldrich) and phosphatase inhibitors (Thermo Fisher Scientific Inc.) using the TissueLyser LT (Qiagen) through high-speed shaking (50 Hz) of samples in 2 ml microcentrifuge tubes with two 5 mm stainless steel beads for 2 min. Samples were then spun down to collect all lysated sample and transferred to clean, beadless tubes. Lysates were then centrifuged at 11 300g for 20 min at 4 °C. The supernatant, which contains the RIPA-soluble fraction of proteins, was reserved in a separate tube while the pellet was washed once in 50 µl of RIPA and centrifuged under the same conditions and the resulting supernatant was pooled with the first one. The remaining pellet was then homogenized in 200 µl of urea buffer (urea 9 m, Tris–HCl 50 mm pH 8, CHAPS 1%, and a cocktail of protease and phosphatase inhibitors) and centrifuged at 11 300g for 30 min. The supernatant was collected as the urea fraction. Protein concentrations of the soluble fraction were measured using the DC Protein Assay Kit (Bio-Rad Laboratories). Soluble and insoluble proteins were loaded for SDS-PAGE migration in a proportion of 1:1.

RNA and protein subcellular fractionation

Sixty newly eclosed adult fly heads were ground to powder using the TissueLyser LT (Qiagen) through three 1-min cycles of high-speed shaking (50 Hz) in 1.5 ml microcentrifuge tubes with two 2.5 mm stainless steel beads. Samples were then gently homogenized in 240 µl of fractionation buffer (Hepes 10 mm, NaCl 10 mm, MgCl2 3 mm, NP-40 0.5%, RNAse inhibitor 100 µ/ml (Promega, Fitchburg, WI, USA) on ice and centrifuged at 100g for 30 s to spin down debris. Lysates were then centrifuged at 2300g for 5 min at 4 °C to separate nuclei from cytoplasm. Nuclei (pellet) were washed four times in 500 µl of fractionation buffer and stored overnight at −80 °C. 20 µl of Sodium acetate 3m pH 5.2 and 600 µl of Ethanol 100% were added to cytoplasmic fractions (Supernatant). Samples were vortexed vigorously and then stored at −80 °C overnight. Cytoplasmic proteins and nucleic acids were then pelleted at 14 000g for 15 min at 4 °C and washed once with 500µl of Ethanol 70%. Proteins and RNA derived from nuclear and cytoplasmic fractions were then extracted using the Nucleospin RNA/protein kit (Macherey Nagel) using the manufacturer’s recommendations.

Immunoblot analysis

Proteins were resolved by TGX Stain-Free 12% gels (Bio-Rad Laboratories), and then transferred onto nitrocellulose membrane (Bio-Rad nitrocellulose Turbo transfer packs) for 7 min (25 V and 2.5 A) using the Trans-Blot Turbo system (Bio-Rad Laboratories). Membranes were then blocked (5% non-fat milk, Tween 0.05%) and incubated with antibodies. Gel loading was normalized by Stain-Free detection of total proteins (activation of the gel before transfer and visualization on the blotting membrane before ECL incubation) using a Geldoc™ EZ imager (Bio-Rad Laboratories), as recommended by the manufacturer. The Stain-Free signal obtained in each lane was quantified (ImageLab™ software, Bio-Rad Laboratories). The following primary antibodies were used: rabbit polyclonal anti-TDP-43 (1:5000; Proteintech, Chicago, IL, USA), mouse monoclonal anti-GFP antibody (clones 7.1 and 13.1) (1:1000; Roche, Mannheim, Deutschland), LacZ (1/10 000; Promega, Charbonnières-les-Bains, France), FUS (1/5000; Bethyl Laboratories, Inc. Montgomery, TX, USA). Membranes were incubated with secondary peroxidase-labelled anti-mouse or anti-rabbit antibodies (1:10 000) from Jackson Immunoresearch Laboratories (WestGrove, PA, USA), and signals were detected with chemiluminescence reagents (ECL Clarity, Bio-Rad Laboratories). Signals were acquired with a GBOX (Syngene, Cambridge, UK), monitored by the Gene Snap software (Syngene). The signal intensity in each lane was quantified using the Genetools software (Syngene), and normalized with the Stain-Free signal quantified in the corresponding lane.

Statistical analysis

All n reported are biological replicates. All statistical analyses were performed using a two-tailed Student’s t-test with Welch’s correction (GraphPad, San Diego, CA, USA). Data on graphs are expressed as mean values, error bars representing standard error of the mean (SEM). For significance symbols, one asterisk means P < 0.05, two asterisks mean P < 0.01 and three asterisks mean P < 0.001.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

We thank Françoise Charbonnier for technical assistance in RT-QMPSF experiments. We thank the Bloomington Drosophila stock center and the Vienna Drosophila Resource center for providing fly stocks.

Conflict of Interest statement. None declared.

Funding

This work was co-supported by a grant from the ‘Association pour la recherche sur la Sclérose Latérale Amyotrophique et autres Maladies du Motoneurone (ARSLA)’ to ML, the European Union and the Région Normandy. Europe gets involved in Normandy through the European Regional Development Fund (ERDF). MP is a PhD fellow of the French Ministry of Higher Education and Research.

Author notes

†

These authors contributed equally to this work.