Abstract
To examine role of hsrω nuclear lncRNAs in Ras signaling cascade, we down- or up-regulated these transcripts in eye discs of Drosophila expressing sev-GAL4 driven activated rasV12 transgene. The sev-GAL4 driven RasV12 transgene expression dependent late pupal lethality and extra R7 photoreceptors in ommatidia, were significantly enhanced when levels of hsrω lncRNA were down/up-regulated. This was associated with enhanced p-MAPK expression, reduced Yan levels, and greater association of RafRBDFLAG with Ras, indicating elevated Ras activation which was both cell autonomous and non-autonomous. RNAseq analysis revealed significant increase in expression of certain sno/sn/scaRNAs and some RNA processing genes in sev-GAL4>RasV12 which was further modulated when hsrωRNA levels were co-altered. Down-regulation of hsrωRNAs elevated positive modulators of Ras signaling while their up-regulation reduced expression of negative modulators of Ras signaling, and thus both conditions have similar outcome. Further enhancement of activity of hyperactive Ras following changes in hsrω lncRNA levels in cell autonomous as well as non-autonomous manner emphasizes roles of lncRNAs in cell signaling during development and disease conditions associated with hyperactive Ras pathway mutants.
Summary Our findings highlight roles of hsrω lncRNA in conditionally modulating the important Ras signaling pathway and provide evidence for cell non-autonomous Ras signaling in Drosophila eye discs.
Introduction
Evolution of multi-cellularity and the associated division of labour between different cell types has necessitated inter-cellular signaling pathways with complex network of regulatory circuits. The evolution of biological complexity has also been paralleled by substantial increase in the non-coding component in diverse genomes. There is increasing realization in recent years that the large varieties of short and long non-coding RNAs (lncRNA) have great roles in cell signaling and gene regulation (Mattick and Makunin, 2006; Geisler and Coller, 2013; Huang et al., 2013; Morris and Mattick, 2014; Jose, 2015; Lakhotia, 2016; Lakhotia, 2017; Peng et al., 2017). The lncRNAs like acal, mrhl, CRNDE, LncND, and URHC have been reported to regulate JNK, Wnt, insulin, Notch and ERK/MAPK signaling pathways, respectively (Arun et al., 2012; Ellis et al., 2014; Xu et al., 2014; Ríos-Barrera et al., 2015; Katsushima et al., 2016). Besides their roles in developmental regulation (Katsushima et al., 2016; Kotake et al., 2016; Lakhotia, 2017; Misawa et al., 2017; Zhang et al., 2017), diverse lncRNAs have been reported to interact with components of the various signaling pathways in cancer, activating or repressing their strength and thus affecting outcomes like cell proliferation or apoptosis (Liu et al., 2015; Wang et al., 2015). An earlier study from our laboratory (Ray and Lakhotia, 1998) showed that mutant alleles of ras (rasE62K and rasD38N) dominantly enhanced the lethality due to nullisomic condition of the hsrω gene, which produces multiple lncRNAs (Lakhotia, 2011; Lakhotia, 2016). The RAS/RAF/MAPK signaling pathway regulates many developmental pathways as it affects cell division, proliferation, growth as well as death, besides its major roles in many human cancers (Fernández-Medarde and Santos, 2011; Pylayeva-Gupta et al., 2011). Ectopic expression of activated Ras causes hyperplastic growth of the concerned organ in Drosophila also (Karim and Rubin, 1998; Prober and Edgar, 2000).
In the present study, we further examined interaction between the hsrω gene and the Ras signaling pathway. We found that down- or up-regulation of hsrω nuclear lncRNAs, through sev-GAL4 driven RNAi or its over-expression, respectively, exaggerates the phenotypes following ectopic expression of activated Ras producing UAS-RasV12 transgene in developing eye discs. The sev-GAL4 driven expression of RasV12 is known to cause ommatidial derangement and rough eyes due to increase in number of R7 photoreceptor cells (Karim et al., 1996). Intriguingly, our results show that reduced as well as enhanced levels of hsrω lncRNAs in sev-GAL4>RasV12 expressing eye discs enhances Ras signaling due to substantial increase in levels of activated Ras and phosphorylated MAPK (p-MAPK) in cell autonomous as well as non-autonomous manner. This resulted in further increase in number of the R7 photoreceptors, with down-regulation of hsrω lncRNAs being more effective than its up-regulation. We found presence of activated Ras and its downstream component (RafRBDFLAG, Ras binding domain of Raf protein tagged with FLAG) in cells that do not express the sev-GAL4>UAS-RasV12 and UAS-RafRBDFLAG transgenes, indicating the possibility that activated Ras complex itself can move from the source cell to the adjacent cells, leading to cell non-autonomous Ras signaling. With a view to understand how changes in levels of hsrω lncRNAs enhance Ras signaling activity, we examined changes in transcriptomes of eye discs expressing activated Ras alone or in background of altered levels of hsrω RNAs. Unexpectedly, levels of transcripts of none of the known members of Ras/Raf/MAPK signaling pathway were found to be significantly affected by changes in hsrω transcripts levels in activated Ras expression background. Interestingly, several sn/snoRNAs along with a scaRNAs and some members of the RNA processing machinery were differentially expressed in activated Ras background and were further affected when the hsrω RNA levels were down-or up-regulated. In addition, while down-regulation of hsrω activity resulted in up-regulation of a few positive modulators of Ras signaling pathway, up-regulation of these transcripts caused down-regulation of the negative regulators of Ras/Raf/MAPK pathway, and thus resulting in increase in Ras activity in either cases.
The present study thus shows that expression of activated Ras not only affects activity of other protein coding genes but also causes altered expression of many non coding RNAs involved in RNA processing. An over- or under-expression of the nuclear lncRNAs produced by the hsrω gene, which is known to affect dynamics of hnRNPs and other proteins associated with omega speckles (Prashant et al, 2000; Lakhotia et al, 2012; Singh and Lakhotia 2015, 2016), seems to modulate levels of diverse RNA processing components and of certain other Ras signaling modulators. Together, these enhance the signaling cascade in cell autonomous as well as non-autonomous manner. The cell non-autonomous Ras signaling seems to correlate with Ras activity, with increase in activity inducing more of cell non-autonomous signaling, possibly through direct transfer of activated Ras complex to neighbouring cells which themselves do not express activated Ras.
Results
Alterations in hsrω RNA levels aggravate pupal lethality and enhance the rough eye phenotype due to sev-GAL4 driven expression of activated Ras in eye discs
We used UAS-RasV12 transgene (Karim et al., 1996), which upon GAL4 driven expression produces mutant Ras that does not undergo RasGTP to RasGDP transformation and can be active even in absence of activation by the upstream receptor tyrosine kinase (RTK). The sev-GAL4 driven ectopic expression of UAS-RasV12 leads to additional R7 rhabdomeres and rough eye morphology (Karim et al., 1996). To alter levels of the >10kb nuclear lncRNAs produced by hsrω gene, the UAS-hsrωRNAi transgene was used to down-regulate the hsrω nuclear lncRNAs (Mallik and Lakhotia, 2009) while for up-regulating this gene, the GAL4 inducible EP3037 (Liao et al., 2000; Mallik and Lakhotia, 2009) allele was used. In some experiments hsrω66, which is a near null allele (Johnson et al., 2011) of hsrω gene was also used to down-regulate hsrω RNA transcripts.
At 24±1°C, only 11–12% of sev-GAL4 driven RasV12 expressing pupae eclosed as adult flies with rough eye phenotype together with de-pigmented patches and occasional black spots in eyes (Fig 1A). Most of those failing to emerge died as pharates (Fig 1B). When either hsrωRNAi or EP3037 was co-expressed in sev-GAL4 driven RasV12 background, none of the pupae eclosed, with a majority dying as early pupae (Fig. 1A, B), indicating that changes in hsrω transcripts enhanced the effects of ectopic expression of activated Ras. With a view to examine the adult eye phenotypes, the above three genotypes were grown at 18±1°C, since the strength of the GAL4 driven expression is reduced at lower temperatures (Brand et al., 1994; Mondal et al., 2007). When reared at 18±1°C, more than 80% flies eclosed in each case, with no early pupal lethality (Fig. 1C). In this case also, sev-GAL4 driven activated Ras expression caused roughening of eyes (Fig. 1I, J) compared to normal eye (Fig. 1D, E). Interestingly, those coexpressing hsrωRNAi or EP3037 and RasV12 showed greater roughening of eyes and ommatidial fusion than in sev-GAL4>RasV12 eyes (Fig. 1F, G, K, L).
The sev-GAL4 driven activated Ras expression in the hsrω66, reared at 24±1°C, also resulted in greater roughening of eyes with fusion of ommatidia (Fig. 1H, M).
When third instar larval eye discs of these genotypes, grown at 24±1°C, were immunostained with anti-Elav to mark the neuronal photoreceptor cells in eye discs, it was evident that both down- or up-regulation of hsrω transcripts in sev-GAL4 driven activated Ras expression background further enhanced the number of photoreceptor cells and consequently the ommatidial disarray (compare Fig 2A–B with Fig 2C–E). In agreement with the reported expression of sev-GAL4 driver (Ray and Lakhotia, 2015), the sev-GAL4 directed UAS-GFP transgene, present in all these genotypes, was expressed in a subset of photoreceptor cells, and in the two future cone cells. Accordingly, the average number of GFP+ve and Elav+ve cells (photoreceptor cells with the Sevenless expression) in each sev-GAL4>UAS-GFP ommatidium varied between 3–4, that of GFP+ve and Elav-ve (future cone cells) between 1–2 cells while the GFP-ve and Elav+ve rhabdomeres (photoreceptor cells without the Sevenless expression) varied between 3–4. Identification of these three classes of cells in third instar larval eye discs from sev-GAL4>UAS-RasV12 larvae showed a small but specific increase only in GFP+ve and Elav+ve cells (Fig 2K), i.e., sev-GAL4 expressing photoreceptor cells. Interestingly, the numbers of GFP+ve and Elav+ve cells in eye discs that had altered levels of hsrω RNAs in sev-GAL4>UAS-RasV12 background, showed a much greater increase (Fig. 2K), with the most pronounced increase being in sev-GAL4>UAS-RasV12 UAS-hsrωRNAi genotype (Fig 2C, K).
It may be noted that, in agreement with an earlier report (Mallik and Lakhotia, 2011), sev-GAL4 driven expression of UAS-hsrωRNAi or EP3037 in normal wild type Ras expression background did not cause any roughening of eyes (data not presented).
Thus alterations in levels of hsrω transcripts in activated Ras expression background enhanced the number of sev-GAL4 expressing neuronal cells but not of the non-neuronal sev-GAL4 expressing future cone cells.
The additional photoreceptors in eye discs with altered levels of hsrω transcripts in activated Ras expression background are R7 type
The Ras/Raf/MAPK signaling dependent differentiation of R7 photoreceptor, the last one of the 8 photoreceptors to differentiate, is initiated by binding of the Boss ligand to the Sevenless receptor tyrosine kinase (RTK) (Tomlinson and Struhl, 2001; Mavromatakis and Tomlinson, 2016). Of the multiple R7 precursor cells, only one cell in which the Boss ligand binds with and activates the RTK, which in turn activates the downstream Ras by converting the GDP-bound Ras to active GTP-bound Ras, differentiates into R7 in normal development. Activation of Ras initiates signaling cascade involving a series of phosphorylation reactions culminating in phosphorylation of MAPK and its nuclear translocation, which eventually triggers R7 differentiation (Karin and Hunter, 1995). Since the RasV12 does not need ligand binding for activation, sev-GAL4>UAS-RasV12 expression directly drives differentiation of two or more R7 photoreceptor cells per ommatidium. To confirm that the additional GFP+ve and Elav+ve photoreceptor cells seen in the experimental genotypes indeed belonged to the R7 lineage, eye discs from larvae of different genotypes were immunostained with Runt antibody. In wild type third instar larval eye discs, each developing ommatidium shows two Runt-positive photoreceptors, viz. R7 and R8 (Edwards and Meinertzhagen, 2009; Tomlinson et al., 2011). Since the R7 and R8 photoreceptors are present one above the other in each ommatidium, a given optical section shows either the R7 or R8, which could be further distinguished because the sev-GAL4>UAS-GFP is expressed only in R7. In wild type discs, R7 cells formed a well arranged pattern of rows of cells with one Runt and GFP-positive rhabdomere in each developing ommatidium (Fig 2F). Most of the ommatidia in eye discs expressing sev-GAL4 driven activated Ras showed two Runt as well as GFP-positive R7 photoreceptor cells (Fig 2G), resulting in derangement of the regular pattern. Down-regulation of hsrω RNA in the same background led to about four R7 photoreceptors in each ommatidium with severely disarrayed ommatidial pattern (Fig 2H, L). Up-regulation of hsrω RNA through EP3037 expression in activated Ras background also increased the number of R7 photoreceptors to about three per ommatidium (Fig 2I, L), which is less than that in sev-GAL4>UAS-RasV12 UAS-hsrωRNAi eye discs. Down-regulating hsrω RNA level using the hsrω66 allele too, resulted in derangement of photoreceptor array in developing eye discs (Fig 2E) due to increase in neuronal cells (Fig 2K), which were confirmed by Runt staining to be R7 photoreceptors (Fig 2J, L).
Altered hsrω RNA levels further enhance Ras signaling in eye discs ectopically expressing activated Ras in cell autonomous as well as non-autonomous manner
The above noted increase in number of R7 photoreceptors in each ommatidium in eye discs with altered levels of hsrω transcripts in activated Ras expression background suggests further increase in Ras signaling. In order to measure the level of Ras signaling, we examined distribution of p-MAPK as phosphorylation of MAPK and its nuclear translocation is a measure of active Ras signaling (Karin and Hunter, 1995). In addition, we also examined levels of Yan, a transcription factor which is negatively regulated by Ras signaling (Brunner et al., 1994; O’Neill et al., 1994).
In the normally developing control eye discs, only a few cells in each ommatidium showed nuclear p-MAPK localization (Fig 2M, 3A). Expression of activated Ras led to greater number of cells showing nuclear p-MAPK staining (Fig 3B, 2M) besides an overall increase in p-MAPK presence. When hsrω RNA levels were either down- (Fig 3C) or up-regulated (Fig 3D), the number of cells with nuclear p-MAPK showed a steep increase (Fig 2M, 3C–D) with concomitant increase in overall p-MAPK levels in the eye discs. Interestingly, not only the cells expressing sev-GAL4 driver, identified by the UAS-GFP expression, but GFP-negative, and thus non sev-GAL4 expressing cells (marked by arrows in Fig 3B–D) also showed higher p-MAPK levels. This suggested a non-autonomous Ras signaling.
In control sev-GAL4>UAS-GFP third instar larvae, the Yan transcription factor is widely expressed in eye discs posterior to the morphogenetic furrow (MF), with especially stronger presence in the MF itself (marked by arrows in Fig 3E–H). The Yan staining progressively declined from anterior to posterior region of eye discs, where it was sparsely present in differentiated photoreceptors. Following sev-GAL4 driven expression of activated Ras a small but perceptible decrease in Yan expression along the antero-posterior axis of the discs and at the MF was noted (Fig 3F). Down- or up-regulation of hsrω RNA levels in the sev-GAL4 driven activated Ras expression background further reduced the expression of Yan in the eye discs (Fig 3G–H), including in the MF cells. Since the sev-GAL4 driver has no expression at the MF (Ray and Lakhotia 2015), the distinct reduction in Yan staining all over the eye disc, including the MF, clearly indicates a cell non-autonomous Ras signaling, which was further enhanced when levels of hsrω non-coding transcripts were reduced or elevated.
With a view to ascertain if above changes that suggested elevated Ras signaling in the test genotypes were associated with enhanced Ras expression and/or with a higher proportion of Ras being in an active form, we co-immunostained developing eye discs of wandering third instar control larvae of different genotypes for Ras and RafRBDFLAG since the UAS-RafRDBFLAG construct (Freeman et al., 2010), which expresses the FLAG-tagged active Ras binding domain of Raf, binds only with active Ras. Fig. 3I–L presents confocal images of immunostainied eye discs while the Fig. 4 shows results of quantification of mean intensity levels of Ras, RafRBDFLAG and co-localization of Ras and RafRBDFLAG, to provide an estimate of active Ras levels in eye discs of different genotypes.
Following expression of sev-GAL4 driven UAS-RafRBDFLAG in developing eye discs of wandering third instar control larvae with normal developmental Ras expression, little colocalization of the RafRBDFLAG with the native Ras was detectable (Fig 3I). In contrast, FLAG tagged RafRBD was substantially co-localized with the Ras in eye discs expressing sev-GAL4 driven activated Ras, as expected with the presence of Ras in active form (Fig 3J). Down- or up-regulation of hsrω RNAs in this background clearly enhanced the number of cells that showed distinctly co-localized RafRBDFLAG and Ras (Fig 3K, L). Interestingly, as may be noted from Fig 3J-L, eye discs co-expressing sev-GAL4 driven UAS-RasV12 and UAS-hsrωRNAi or EP3037 had a greater number of GFP-ve cells, adjoining the GFP+ve cell, which also showed colocalized Ras and RafRBDFLAG. Since neither RafRBDFLAG nor activated Ras was expressed in the non sev-GAL4 expressing GFP-ve cells, their co-localization in such cells reflects movement of activated Ras complex from sev-GAL4 expressing cells to the neighbouring cells.
With a view to know if the increased colocalization seen in greater numbers of cells in eye discs expressing activated Ras without or with altered hsrω RNA levels reflected equal or differential elevation in levels of total Ras, activated Ras and RafRBDFLAG proteins, we used the Histo option of the Zeiss Meta 510 Zen software to quantify the Ras, FLAG, DAPI and GFP fluorescence signals in these genotypes. As seen in Fig. 4A, the total Ras is expectedly higher in sev-GAL4>UAS-RasV12 than in sev-GAL4>UAS-GFP eye discs. The more than 2 times further increase in Ras staining in discs co-expressing UAS-hsrωRNAi or EP3037 with activated Ras, clearly shows that the under- or over-expression of hsrω transcripts in activated Ras overexpression background further enhanced Ras levels. The increase in GFP staining in sev-GAL4>UAS-RasV12 UAS-hsrωRNAi correlates with the earlier noted greater increase in sev-GAL4 driven GFP expressing cells in these eye discs. The more or less comparable levels of DAPI fluorescence in the samples of eyes discs in different genotypes indicates that the increase in Ras or GFP activity in specific genotypes is not due to a general increase in number of cells in some genotypes. It is significant that the levels of RafRBDFLAG protein showed the expected increase in sev-GAL4>UAS-RasV12 than in sev-GAL4>UAS-GFP eye discs but co-expression of UAS-hsrωRNAi or EP3037 with activated Ras was not associated with further increase in the FLAG staining (Fig. 4A). In order to determine how much of the enhanced levels of Ras in sev-GAL4>UAS-RasV12 and more so in sev-GAL4>UAS-RasV12 UAS-hsrωRNAi and sev-GAL4>UAS-RasV12 EP3037 eye discs was in activated form, we examined colocalization of Ras and RafRBFFLAG fluorescence signals (Fig. 4B). In agreement with the ectopic expression of activated Ras in sev-GAL4>UAS-RasV12 eye discs, nearly 40% of RafRBDFLAG was associated with Ras compared to only about 5% of the FLAG signal being associated with Ras in sev-GAL4>UAS-GFP discs. Interestingly, co-expression of UAS-hsrωRNAi or EP3037 in sev-GAL4>UAS-RasV12 eye discs resulted in further increase in association of Ras and RafRBDFLAG proteins, indicating that a greater proportion of Ras in these cells is in activated form with which the RafRBDFLAG can bind.
In order to see if the above noted increase in Ras and activated Ras levels was associated with increased transcription of the resident Ras and/or UAS-RasV12 transgene, we examined the levels of Ras transcripts derived from these two sources using semi-quantitative RT-PCR with primers designed to differentiate between these two transcripts (see Supplementary Methods). The normal resident Ras gene transcripts remained more or less comparable in all the four genotypes (sev-GAL4>UAS-GFP, sev-GAL4>UAS-RasV12, sev-GAL4>UAS-RasV12 hsrωRNAi and sev-GAL4>UAS-RasV12 EP3037) and likewise, the transcripts derived from the RasV12 transgene remained similar in sev-GAL4>UAS-RasV12 and those co-expressing hsrωRNAi or EP3037 with UAS-RasV12 (see Supplementary Fig S1 A and B). This indicated that the elevated Ras activity in the latter two genotypes with altered hsrω RNA levels is unlikely to be due to increased transcription of UAS-RasV12 transgene or the resident Ras gene. As noted later, the RNA-seq data also did not show any significant increase in Ras transcripts in sev-GAL4>UAS-RasV12 eye discs co-expressing hsrωRNAi or EP3037 compared to those expressing only activated Ras.
Down- or up-regulation of hsrω transcripts commonly affects many RNA processing components in activated Ras background but differentially affects positive and negative modulators of Ras signaling
With a view to understand how the down- and up-regulation of levels of hsrω lncRNAs similarly enhance Ras signaling activity in cells ectopically expressing activated Ras, we undertook transcriptome analysis of total cellular RNA in eye discs of different genotypes. The eye discs from sev-GAL4>UAS-GFP late third instar larvae served as control. We also examined changes in transcriptome following sev-GAL4 driven down- or up-regulation of hsrω transcripts in normal Ras background. Although detailed results of this analysis will be presented elsewhere, in the present we have used this data whenever required to compare with changes that are relevant in the context of their expression in activated Ras expression background.
The sev-GAL4 driven expression of activated Ras in eye discs resulted in differential expression of many genes, with 374 genes being down-regulated and 138 up-regulated, when compared with those in sev-GAL4>UAS-GFP eye discs (List 1 in Fig 5A and B, Supplementary Table S1, sheet1). Besides the expected increase in levels of transcripts of genes involved in cell growth, proliferation and differentiation, transcripts of many genes involved in RNA biosynthesis, metabolism and processing were also found to be up-regulated when compared with sev-GAL4>UAS-GFP eye discs (Supplementary Table S1, sheet2).
As expected, the levels of Ras transcripts were significantly higher in sev-GAL4>UAS-RasV12 than in sev-GAL4>UAS-GFP. In agreement with the above noted RT-PCR results (Supplementary Fig. S1A, B), the RNA seq data also showed that levels of Ras transcripts did not show any further significant increase when hsrω RNA levels were down- or up-regulated in sev-GAL4 driven activated Ras expression background (see, Supplementary Table S1, sheet 2). The RNA seq data also revealed levels of transcripts of the genes that act directly downstream in the Ras signaling cascade were not further up-regulated in eye discs co-expressing sev-GAL4 driven UAS-RasV12 and hsrωRNAi or EP3037.
A search of the RNA seq data on the basis of GO terms showed that eye discs with sev-GAL4 driven activated Ras expression showed up-regulation of several genes involved in R7 cell differentiation (e.g., salm, ten-m, cadn, svp, dab, nej) and also of certain genes involved in photoreceptor development (e.g., rno, doa, pdz-gef, jeb, atx2 and csn4) (see Supplementary Table S1, sheet2). However, none of these genes showed any further change when hsrωRNAi or EP3037 was co-expressed with activated Ras, except for svp, whose transcripts were further up-regulated in eye discs co-expressing sev-GAL4 driven activated Ras and hsrωRNAi but not in those co-expressing activated Ras and EP3037 (see later). Results of a real-time qRT-PCR analysis for levels of transcripts of many of the genes, which are part of Ras signaling cascade, in different genotypes (see Supplementary Fig. S1 C) validated the RNA seq data as none of them were found to show differential expression in eye discs expressing activated Ras along with changes in hsrω RNA levels compared to those expressing activated Ras alone.
In order to understand the basis for the unexpected similar enhancing effect of down- or up-regulation of hsrω transcripts on the Ras signaling in eye discs that were expressing activated Ras under the sev-GAL4 driver, we looked for genes that were commonly down- or up-regulated following expression of hsrωRNAi or EP3037 in sev-GAL4 driven activated Ras expression background (encircled in red and white, respectively, in Fig. 4A and B). The group that was down-regulated by activated Ras expression and further down-regulated when hsrωRNAi or EP3037 was co-expressed included only one gene (encircled in red in Fig 5A), while the group that was up-regulated following activated Ras expression and further up-regulated when hsrω RNA levels were reduced or elevated included three genes (encircled in red in Fig. 5B). The single gene in the first group, CG13900, codes for a splicing factor 3b subunit involved in mRNA splicing via spliceosome (U2 snRNP), while the three genes up regulated on Ras activation and further up-regulated on co-alterations in hsrω RNA levels were CG15630, Hsp70Bb and CG14516. On the basis of their described roles in the Flybase, however, none of these appeared to be directly involved in modulating Ras pathway. The CG13900 gene encoding a splicing factor, however, may be involved in post-transcriptional modulation of activities of other genes, including those in the Ras pathway.
When we looked for genes that were commonly down or up-regulated following hsrωRNAi or EP3037 co-expression in sevGAL4>RasV12 expressing eye discs compared to those of sevGAL4>RasV12 but which showed no change in sevGAL4>RasV12 as compared to sevGAL4>UAS-GFP control eye discs, a total of 88 genes were found to be commonly down-regulated (encircled in white in Fig. 5A, and Fig. 6A–C) and 45 (encircled in white in Fig. 5B, and Fig. 6D) commonly up-regulated. One group among the 88 commonly down-regulated genes comprised of several snoRNAs, snRNA and scaRNA. As shown in Fig. 6A, the genes in the sno/scaRNAs group were significantly up-regulated by activated Ras expression but coexpression of either hsrωRNAi or EP3037 with activated Ras led to their significant down-regulation. It is important that, except one, none of these sno/scaRNAs are significantly affected when hsrωRNAi or EP3037 is expressed under the sev-GAL4 in normal Ras background (Fig. 6A). Interestingly, sev-GAL4 driven hsrωRNAi or EP3037 expression in normal Ras background affects some other sno/scaRNAs which are not affected by activated Ras expression (data not presented). Six genes, associated with the GO term Transcription factors, were significantly up-regulated following sev-GAL4 driven activated Ras expression but all of them were commonly down-regulated when hsrωRNAi or EP3037 were co-expressed (Fig. 6B). Among these, the Ssb-c31a is a protein that binds with Raf (Friedman et al., 2011) and is reported to be a negative regulator of transcription (Lacoste et al., 1995).
The other genes among these commonly down-regulated 88 genes (Fig. 6C) belonged to diverse GO terms without any apparent and known association with Ras signaling. One of these genes, the dlc90F encodes a dynein light chain protein, which is reported to bind to Raf (Friedman et al., 2011) and to be a positive regulator of FGF signaling (Zhu et al., 2005); however, not much is known about its role in Ras signaling.
The group of 45 genes that were significantly up-regulated by co-expression of hsrωRNAi or EP3037 in activated Ras background (Fig. 6D) included diverse genes, none of which seem to be directly involved in Ras signaling pathways. However, the kuz (kuzbanian) gene, encoding a metalloprotease, is expressed in developing ommatidia with roles in neuroblast fate determination, round-about (Robo) signaling and neuron formation (Sotillos et al., 1997; Udolph et al., 2009; Coleman et al., 2010). Therefore, this may be one potentially important gene which is up-regulated only when hsrω RNA levels are altered in activated Ras expressing eye discs in parallel with increased numbers of R7 photoreceptors.
Finally, we looked at genes that were differentially affected when hsrωRNAi or EP3037 were coexpressed with activated Ras. Again, these genes belonged to different pathways (see Supplementary Table S1, sheets 4–7) but one group that appeared significant was that of positive and negative modulators of photoreceptor differentiation (Fig. 7). The Gene ontology search results revealed that down-regulation of hsrω transcripts in activated Ras background enhanced levels of R7 cell fate determining genes, namely, phly, svp, rau and socs36E compared to only activated Ras expressing eye discs (Fig 7 and Supplementary Table S1, sheet4).
In case of over-expression of hsrω RNA in activated Ras background, none of the above four genes involved in R7 differentiation were up-regulated. However, several other negative regulators of Ras signal transduction pathway like bru, klu, mesr4, cdep, epac,, nfat, ptp-er, cg43102, rhogap1a, rhogef2, spri, exn etc were down regulated (Supplementary Table S1, sheet 7, and Fig 7) in sev-GAL4>RasV12 EP3037 eye discs. On the basis of their GO term classification (David Bioinformatics), genes like bru, cdep, sos, pdz-gef, cg43102, rhogap1a, rhogef2, spri, exn in this group are involved in Ras guanyl nucleotide exchange factor activity while the other genes like klu, mesr4, nfat, ptp-er affect small GTPase mediated signal transduction. Being negative-regulators, their down-regulation by co-expression of activated Ras and EP3037 would lead to further enhanced Ras activity. It is interesting to note that in normal developmental Ras activity background, the sev-GAL4 driven expression of UAS-hsrωRNAi or EP3037 did not exert comparable differential effects on expression of these positive and negative modulators of Ras signaling since as shown in the last two columns in Fig. 7, these genes were either not affected or were commonly down regulated in both conditions.
Among the other groups of genes that showed differential changes upon down or up-regulation of the hsrω transcripts in activated Ras expressing eye discs, was a group involved in ribosome biogenesis that was down regulated in activated Ras expressing discs with down-regulated hsrω transcripts. Expression of these genes was, however, not affected when EP3037 was coexpressed with activated Ras (Supplementary material S2 sheets 6–7).
Discussion
Present study was initiated to examine interactions between long non-coding hsrω RNAs and Ras signaling cascade. Although, some studies have implicated oncogenic Ras to influence expression of several long non coding transcripts (Rotblat et al 2011, Kotake 2016, Zhang et al 2017 Jiang et al., 2017; Jinesh et al., 2017), regulation of Ras signaling cascade by lncRNAs is not yet known. Earlier study from our lab showed that a ras mutant allele enhanced hsrω-null phenotype (Ray and Lakhotia, 1998). The present study shows that alterations in levels of hsrω transcripts exaggerate phenotypes that follow when activated Ras is ectopically expressed in developing eye discs of Drosophila larvae. The UAS-RasV12 transgene (Karim et al., 1996) has been widely used for examining consequence of ectopic expression of ligand-independent activated Ras. Its expression in developing eye discs under the sev-GAL4 driver is known to disrupt ommatidial arrays because of recruitment of additional cells to R7 photoreceptor path (Karim et al., 1996). Our results clearly show that reduced as well as over-expression of hsrω transcripts in activated Ras background significantly enhanced the number of R7 photoreceptor per ommatidium unit. As revealed by detection of p-MAPK, Yan and activated Ras associated RafRBDFLAG in eye discs, the increase in R7 photoreceptors was distinctly correlated with the enhanced Ras activity levels in sev-GAL4>hsrωRNAi UAS-RasV12 and sev-GAL4>EP3037 UAS-RasV12 genotypes.
Of the multiple transcripts produced by hsrω gene (http://flybase.org; Lakhotia, 2011; Lakhotia, 2017), the >10kb nuclear transcripts are primarily involved in biogenesis of omega speckles and thus affect the dynamic exchanges of hnRNPs and other omega speckle associated proteins between different nuclear compartments (Lakhotia et al., 2012; Singh and Lakhotia, 2015; Lakhotia, 2016; Singh and Lakhotia, 2016; Lakhotia, 2017). The GAL4 induced expression of hsrωRNAi transgene and of the EP3037 allele of hsrω are known to primarily lower and elevate, respectively, levels of the larger and nucleus-limited transcripts of hsrω gene (Mallik and Lakhotia, 2009; Mallik and Lakhotia, 2011). Either of these conditions have comparable consequences for dynamics of omega speckles and their associated proteins, which are involved in RNA processing (Lakhotia et al., 2012; Singh and Lakhotia, 2015). Such common consequences for dynamics of several important members of RNA processing machineries may be responsible for the rather unexpected commonality in enhancing the Ras signaling cascade when the hsrω nuclear transcript levels were down or up-regulated in the background of ectopic expression of activated Ras.
The down- or up-regulation of hsrω transcripts apparently did not further enhance the transcription of UAS-RasV12 transgene and the other downstream Ras-signaling pathway genes since their transcript levels did not show any significant elevation in sev-GAL4>RasV12 UAS-hsrωRNAi or sev-GAL4>RasV12EP3037 over those in sev-GAL4>RasV12 eye discs. Therefore, the greatly enhanced Ras signaling seems to be mediated through post-transcriptional events. It is notable that in spite of the similar levels of Ras transcripts in sev-GAL4>RasV12 UAS-hsrωRNAi, sev-GAL4>RasV12EP3037 and sev-GAL4>RasV12 eye discs, the levels of total as well as activated Ras, as revealed by immunostaining and RafRBDFLAG association (Fig. 3I–L and Fig 4), appeared to be more enhanced in the former two genotypes. This clearly suggests effects at post-transcriptional levels.
It is significant that down- or up-regulation of hsrω transcripts in wild type Ras background does not affect ommatidial differentiation since the eyes in sev-GAL4>UAS-hsrωRNAi or sev-GAL4>EP3037 flies are normal with no extra rhabdomere (Mallik and Lakhotia, 2011). Immunostaining for Ras and RafRBDFLAG in sev-GAL4>UAs-GFP or sevGAL4>UAS-hsrωRNAi or sev-GAL4>EP3037 eye discs also did not show any difference between them (data not shown). Apparently, ectopic over-expression of activated Ras makes the cells detectably sensitive to alterations in hsrω transcript levels. This may have implications for a role of the hsrω transcripts in modulating any aberrant Ras activity during normal development.
The significant increase in Ras activity seen in sev-GAL4>RasV12 UAS-hsrωRNAi or sev-GAL4> RasV12 EP3037 eye discs seems to be due to modulation of activities of other regulators of the Ras signaling cascade. Our data suggest that sev-GAL4>RasV12 UAS-hsrωRNAi eye discs have significantly elevated levels of transcripts of some of the genes (phly, svp, rau and socs36E) whose products have been identified as positive modulators of Ras activity in R7 differentiation. The phly encodes a nuclear receptor, which acts downstream of Ras and is responsible for R7 fate determination (Chang et al., 1995). The Svp is an orphan nuclear receptor responsible for transforming cone cells to R7 photoreceptor cells in conjunction with activated Ras (Begemann et al., 1995; Kramer et al., 1995). Rau is known to sustain RTK signaling downstream of Htl or Egfr via two Ras association (RA) domains, which have a binding preference for GTP-loaded Ras to maintain its activity via a positive feedback loop (Sieglitz et al., 2013). Svp expression is also known to increase Rau expression (Sieglitz et al., 2013). The Socs36E too can amplify Ras/MAPK signaling in precursors of R7 cells (Almudi et al., 2010). Thus, up-regulation of these genes in eye discs with decreased hsrω RNA levels in activated Ras expression background correlates with the observed greater increase in R7 photoreceptors than when activated Ras is expressed alone. While none of these positive modulators showed enhanced activity in sev-GAL4>RasV12UAS-EP3037 eye discs, transcripts of another set of genes (bru, klu, mesr4, cdep, epac, nfat, ptp-er, cg43102, rhogap1a, rhogef2, spri, exn etc), which can act as negative modulators of the Ras cascade, were actually down regulated. On the basis of their GO term classification (David Bioinformatics), some of these genes, viz., bru, cdep, sos, pdz-gef, cg43102, rhogap1a, rhogef2, spri, exn, are involved in Ras guanyl nucleotide exchange factor activity (Lee et al., 2002; Jékely et al., 2005; Yan and Perrimon, 2015)) while others like klu, mesr4, nfat, ptp-er are negative regulators of small GTPase mediated signal transduction (Huang and Rubin, 2000; Brachmann and Cagan, 2003; Ashton-Beaucage et al., 2014). In either cases, therefore, the net result would be up-regulation of Ras activity as actually seen in sev-GAL4>UAS-hsrωRNAiUAS-RasV12and sev-GAL4>UAS-EP3037 UAS-RasV12 eye discs. The action of positive regulators can be expected to be stronger than that of the negative modulators and this may be one of the factors for the greater enhancement in Ras signaling in sev-GAL4> RasV12UAS-hsrωRNAi than in sev-GAL4> RasV12UAS-EP3037 eye discs.
The >10kb long nuclear lncRNAs, whose levels are primarily altered by the GAL4 driven expression of hsrωRNAi transgene or the EP3037 allele (Mallik and Lakhotia, 2009, 2011), are essential for biogenesis of the omega speckles and consequently for the dynamic movement of the omega speckle associated hnRNPs, other RNA binding proteins and some of the nuclear matrix proteins between different sub-cellular compartments (Mallik and Lakhotia, 2011; Lakhotia, et al 2012; Singh and Lakhotia, 2015). Any disturbance in the intra-cellular dynamicity of these proteins would affect their functions, which in turn would impinge upon a large variety of gene activity at transcriptional and post-transcriptional levels because of the central roles of different hnRNPS in prost-transcriptional processing of diverse RNAs (Han et al., 2010; Piccolo et al., 2014). Present results suggest that besides acting through omega speckle associated RNA-binding proteins, the hsrωRNAs may also directly or indirectly affect different small ncRNAs since certain snoRNAs, snRNAs and scaRNAs, which are highly up regulated in activated Ras background but were found to be significantly down-regulated when the hsrω transcripts were either up- or down-regulated. Such alterations in levels of the small ncRNAs may also contribute to the enhanced Ras signaling in these genotypes. Although the most widely appreciated function of snoRNAs is their roles in maturation of rRNAs, through extensive 2’-O-methylation and pseudouridylation modifications, guided, during or after their transcription (Henras et al., 2015; Sloan et al., 2017), especially by C/D box and H/ACA box snoRNAs, respectively (Dragon et al., 2006; Dieci et al., 2009), they have other regulatory roles as well, including modifications of some snRNAs (Falaleeva and Stamm, 2013; Dupuis-Sandoval et al., 2015; McMahon et al., 2015).The Cajal body associated scaRNAs are essential for proper functioning and maturation of snRNAs, which in turn are critical for appropriate processing of mRNAs (Darzacq et al., 2002; Kiss et al., 2002; Richard et al., 2003; Kiss, 2004; Deryusheva and Gall, 2009). The cumulative consequences of alterations in these diverse small RNAs (sca, sn and snoRNAs) on different gene’s transcriptional and translational activities can be very extensive. It would be interesting to examine the possibility that some of these small ncRNAs are part of an auto-inhibitory loop in Ras signaling such that their down regulation results in further enhanced Ras activation. Loss of two snoRNAs, SNORD50A and SNORD50B, in human cells has been shown to be associated with increased levels of active K-Ras leading to hyperactivated Ras/ERK signaling (Siprashvili et al., 2016). Homologs of these two snoRNAs are not yet identified in flies.
Several ribosomal protein genes show reduced expression following activation of the Ras pathway (Friedman et al., 2011). It remains to be examined if the observed reduction in some of the ribosomal proteins gene transcripts in sev-GAL4> RasV12UAS-hsrωRNAi eye discs has any effect on the very high Ras activity in these eye discs. Likewise, further studies are needed to identify roles, if any, of altered expression of genes like CG13900, dlc90F and kuz in further elevating the Ras signaling in the experimental genotypes.
An interesting finding of the present study was the marked increase in cell non-autonomous Ras signaling. The sev-GAL4 driven activated Ras as well as RafRBDFLAG transgenes were expressed only in a specific subset of cells in the eye discs. Therefore, the presence of RafRBDFLAG bound Ras in adjacent non sev-GAL4 expressing cells is unexpected and suggests that the activated Ras-Raf complex can move out of the source cells to neighbouring cells. The great reduction in Yan expression even in MF region, where the sev-GAL4 driver is not at all expressed, when activated Ras expression is accompanied by altered hsrω transcript levels also provides strong evidence for existence of cell non-autonomous Ras signaling. Although several studies (Uhlirova et al., 2005; Yan et al., 2009; Parry and Sundaram, 2014; Takino et al., 2014; Enomoto et al., 2015) have indicated cell non autonomous Ras signaling, only one study reported transfer of GFP tagged H-Ras to T cells from the antigen-presenting cells (Goldstein et al., 2014). Our results also strongly indicate movement of the activated Ras complex across to neighbouring cells.
Our findings, besides highlighting roles of non coding part of the genome in modulating important signaling pathway like Ras, also unravel new insights into the working of Ras signaling cascade itself. The observed non-autonomous spread of Ras signaling and increase in Ras activity by hsrω lncRNAs are also significant in view of the fact that activated Ras/Raf mutations are implicated in diverse malignancies. Although some earlier studies indicated roles of certain lncRNAs in ERK/MAPK activation in cancer (Xu et al., 2014; Kotake et al., 2016; Zhang et al., 2017), their roles in enhancing Ras activity itself has not been known. Future studies on interactions between the diverse small nc and lncRNAs and signaling pathways like Ras are expected to unravel new dimensions of cellular networks that regulate and determine the basic biological processes of cell proliferation, differentiation and death on one hand, and roles in cancer on the other.
Materials and Methods
Fly stocks
All fly stocks and crosses were maintained on standard agar-maize powder-yeast and sugar food at 24±1°C. The w1118; sev-GAL4; + (no. 5793; Bailey 1999), w1118; UAS-GFP (no. 1521), w1118; UAS-RasV12 (no. 4847) stocks were obtained from the Bloomington Stock Centre (USA):. For targeted (Brand and Perrimon, 1993) down-regulation of the hsrω transcripts, UAS-hsrω-RNAi3 transgenic line was used, while its up regulation was achieved through the GAL4 inducible EP3037 over expressing hsrω allele (Mallik and Lakhotia, 2009). These two lines are referred to in the text as UAS-hsrωRNAi and EP3037, respectively. For hsrω-null condition, the w: hsrω66/hsrω66 stock (Johnson et al., 2011) was used. The UAS-RafRBDFLAG stock (Freeman et al., 2010) was provided by Dr S Sanyal (Emory University, USA). Using these stocks, appropriate crosses were made to finally obtain progenies of the following genotypes:
w1118; sev-GAL4 UAS-GFP/UAS-GFP; dco2 e/+
w1118; sev-GAL4 UAS-GFP/UAS-GFP; dco2 e/UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-GFP; UAS-hsrωRNAi/UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-GFP; EP3037/UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-GFP; hsrω66/hsrω66 UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-RafRBDFLAG; dco2 e/+
w1118; sev-GAL4 UAS-GFP/UAS-RafRBDFLAG; dco2 e/UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-RafRBDFLAG; UAS-hsrωRNAi/UAS-RasV12
w1118; sev-GAL4 UAS-GFP/UAS-RafRBDFLAG; EP3037/UAS-RasV12
The w1118, dco2 and e markers are not mentioned while writing genotypes in Results.
Lethality Assay
For lethality assay, freshly hatched 1st instar larvae of sev-GAL4>UAS-GFP, sev-GAL4>RasV12, sev-GAL4>UAS-RasV12UAS-hsrωRNAi and sev-GAL4>UAS-RasV12EP3037 were collected during one hour interval and gently transferred to food vials containing regular food and reared at 24±1°C or at 18±1°C. The total numbers of larvae that pupated and subsequently emerged as flies were counted for at least three replicates of each experimental condition and/or genotypes.
Photomicrography of adult eyes
For examining the external morphology of adult eyes, flies of the desired genotypes were etherized and their eyes photographed using a Sony Digital Camera (DSC-75) attached to a Zeiss Stemi SV6 stereobinocular microscope or using Nikon Digital Sight DS-Fi2 camera mounted on Nikon SMZ800N stereobinocular microscope.
Nail polish imprints
Flies to be examined for organization of ommatidial arrays were anaesthetized and decapitated with needle and the decapitated head was briefly dipped in a drop of transparent nail polish placed on a slide. It was allowed to dry at RT for 5–10 min. The dried layer of nail polish was carefully separated from the eye tissue with fine dissecting needles and carefully placed on another clean glass slide with the imprint side facing up, and finally flattened by gently placing a cover slip over it as described earlier (Arya and Lakhotia, 2006). The eye imprints were examined using 20X DIC optics.
Whole organ immunostaining
Eye discs from actively migrating late third instar larvae of desired genotypes were dissected out in PSS and immediately fixed in freshly prepared 3.7%paraformaldehyde in PBS for 20 min and processed for immunostaining as described earlier (Prasanth et al., 2000). The following primary antibodies were used: rat monoclonal anti-Elav (DSHB, 7E8A10, 1:100), rabbit monoclonal anti-Ras (27H5, Cell signaling, 1:50), mouse anti-FLAG M2 (Sigma-Aldrich, India, 1:50), rabbit p-MAPK (Phospho-p44/42 MAPK (Thr202, Tyr204), D13.14.4E, Cell signaling, 1:200), mouse anti-Yan (8B12H9, Developmental Studies Hybridoma Bank, Iowa, 1:100) and guinea pig anti-Runt, a gift by Dr. K. Vijaya Raghavan, India, (Kosman et al., 1998) at 1:200 dilution.
Appropriate secondary antibodies conjugated either with Cy3 (1:200, Sigma-Aldrich, India) or Alexa Fluor 633 (1:200; Molecular Probes, USA) or Alexa Fluor 546 (1:200; Molecular Probes, USA) were used to detect the given primary antibodies. Chromatin was counterstained with DAPI (4’, 6-diamidino-2-phenylindole dihydrochloride, 1μg/ml). Tissues were mounted in DABCO antifade mountant for confocal microscopy with Zeiss LSM Meta 510 using Plan-Apo 40X (1.3-NA) or 63X (1.4-NA) oil immersion objectives. When required, quantitative estimates of proteins in different regions of eye discs and their co-localization were obtained with the Histo option of Zeiss LSM Meta 510 software. All images were assembled using Adobe Photoshop 7.0 software.
Next Generation RNA sequencing
Total RNAs were isolated from 30 pairs of eye discs from sev-GAL4>UAS-GFP, sev-GAL4>UAS-hsrωRNAi, sev-GAL4>EP3037, sev-GAL4>UAS-RasV12, sev-GAL4>UAS-hsrωRNAi UAS-RasV12 and sev-GAL4>EP3037 UAS-RasV12 third instar larvae using Trizol (Invitrogen, USA) reagent as per manufacture’s protocol. 1μg of the isolated RNA was subjected to DNAase treatment using 2U of TurboTM DNase (Ambion, Applied Biosystem) enzyme for 30 min at 37°C. The reaction was stopped using 15mM EDTA followed by incubation at 65°C for 5–7 min and purification on RNAeasy column (Qiagen). The purified RNA samples were processed for preparations of cDNA libraries using the TruSeq Stranded Total RNA Ribo-Zero H/M/R (Illumina) and sequenced on HiSeq-2500 platform (Illumina) using 50bp pair-end reads, 12 samples per lane and each sample run across 2 lanes. This resulted in a sequencing depth of ~20 million reads in each case. Triplicate biological samples were sequenced in each case. The resulting sequencing FastQ files were mapped to the Drosophila genome (dm6) using Tophat with Bowtie. The aligned SAM/BAM file for each was processed for guided transcript assembly using Cufflink 2.1.1 and gene counts were determined. Mapped reads were assembled using Cufflinks. Transcripts from all samples were subjected to Cuffmerge to get final transcriptome assembly across samples. In order to ascertain differential expression of gene transcripts between different samples, Cuffdiff 2.1.1 was used (Trapnell et al., 2012). A P-value <0.05 was taken to indicate significantly differentially expressing genes between the compared genotypes. Genes differentially expressed between experimental and control genotypes were categorized into various GO terms using DAVID bioinformatics Resources 6.8 (Huang et al., 2009) https://david.ncifcrf.gov for gene ontology search. In order to find out distribution of differentially expressing genes into various groups, Venn diagrams and Heat maps were prepared using the Venny2.1 and ClustVis softwares, respectively (Metsalu and Vilo, 2015).
Competing interests
Authors declare no conflicting interests
Author contributions
MR and SCL planned experiments, analyzed results and wrote the manuscript. MR carried out the experimental work and collected data.
Funding
This work was supported by a CEIB-II grant (no. BT/PR6150/COE/34/20/2013) from the Department of Biotechnology, Govt. of India to SCL. MR is supported by the Indian Council of Medical Research, New Delhi, India through senior research fellowship.
Data availability
The NGS data are being deposited at GEO (http://www.ncbi.nlm.nih.gov/geo/).
Acknowledgements
We thank the Bloomington Drosophila Stock Ctr and Drs S.Sanyal (Emory University, USA) and Stephen W. Mckechnie (Australia) for providing fly stocks. We thank Developmental Studies Hybridoma Bank (DSHB, Iowa, USA) for anti-Elav and anti-Yan, and Dr. K. Vijay Raghavan (India)) for anti-Runt. We also thank the Centre of Advanced Studies in Department of Zoology, DBT-BHU Interdisciplinary School of Life Sciences and the Centre of Genetic Disorders (CGD) at BHU for various facilities. We thank Dr Amit Chaurasia at CGD for RNA-sequencing. We thank the Department of Science &Technology, Govt. of India (New Delhi) and Banaras Hindu University for the Confocal Microscopy facility.
Footnotes
ORCID ID: Mukulika Ray: 0000-0002-9064-818X
S. C. Lakhotia: 0000-0003-1842-8411