Yang et al . mRNA stability stops neuroblasts dividing May 11 , 2017 Regulating prospero mRNA stability determines when neural stem cells stop dividing

During Drosophila and vertebrate brain development, termination of neural stem cell (neuroblast) proliferation and their differentiation require the conserved transcription factor Prospero/Prox1. It is not known how the level of Prospero is regulated to produce an expression peak in pupal neuroblasts, which terminates proliferation. Here, we use single molecule fluorescent in situ hybridisation to show that larval neurons and terminal pupal neuroblasts selectively transcribe a long prospero isoform containing a 15 kb 3′ UTR stabilised by binding to the conserved RNA-binding protein Syncrip/hnRNPQ. The long prospero isoform and Syncrip are both required to stop neuroblasts dividing. Our results demonstrate an unexpected function for mRNA stability in limiting neuroblast proliferation required for normal brain development. Given that Prox1 regulates vertebrate neuroblasts and other stem cells, our findings suggest widespread roles for regulated mRNA stability in stem cell biology.


INTRODUCTION
The central nervous system (CNS) consists of a huge diversity and number of neurons and glia that originate from a limited population of neural stem cells, also known as neuroblasts (NBs) (Kelava and Lancaster, 2016).To produce a normal CNS, NBs must divide the correct number of times and undergo a precisely regulated programme of differentiation.Many factors and mechanisms regulating these processes have been studied in great detail, and in many cases are conserved between vertebrates and Drosophila (Homem and Knoblich, 2012).However, a key question in both vertebrates and Drosophila CNS development is how NB numbers are regulated by balancing proliferation and differentiation; this question has not yet been conclusively answered.
In Drosophila embryos and larvae, NBs divide asymmetrically, multiple times, while maintaining a single large cell that retains its stem cell properties.The asymmetric NB divisions 'bud off' smaller cells, called ganglion mother cells (GMCs) (Knoblich, 2008) that typically divide only once to produce a pair of differentiated cells (Homem and Knoblich, 2012).There are two classes of NB lineages; type I, in which GMCs originate directly form the NB divisions, and type II, in which NBs divide to give intermediate neuronal progenitors (INPs) that are stem cell-like in character, and divide further to generate multiple GMCs (Homem and Knoblich, 2012).Type I proliferation and differentiation are determined by the conserved homeodomain containing transcriptional regulator, Prospero (Pros) (Bayraktar et al., 2010).Pros activates the expression of genes required for the differentiation of type I NB progeny and suppresses the expression of genes required to promote stem-like properties (Lai and Doe 2014;Doe et al., 1991;Vaessin et al., 1991;Matsuzaki et al., 1992;Choksi et al., 2006;Bello et al., 2006;Betschinger et al., 2006;Lee et al., 2006).In contrast, type II NBs do not express Pros and their progeny remain stem cell-like as they have no immediate need for balancing proliferation and differentiation (Bayraktar et al., 2010).Prox1, the highly conserved vertebrate homologue of Pros, has related roles in the regulation of stem cells in diverse tissues (Kato et al., 2015;Dyer et al., 2002;Foskolou et al., 2013;Stergiopoulous et al., 2015;Elsir et al., 2012), but less is known about the details of Prox1 expression and function in comparison to Pros.
The best understood function of Pros is in promoting the differentiation of the progeny of NB into GMCs and then further into neurons.pros mRNA and Pros protein are segregated asymmetrically during NB division (Hirata et al., 1995;Spana and Doe, 1995;Knoblich et al., 1995), a process most extensively studied in embryos, but also operating in larvae (Kitajima et al., 2010).Pros is first expressed in NBs and is localised in a cortical basal crescent through binding to Miranda (Mir) protein, which prevents the import of Pros into the NB nucleus and ensures its inheritance by the GMC.Once Mir is degraded, Pros enters the nucleus of the GMC (Ikeshima-Kataoka et al., 1997).In this way, Pros sub-cellular localisation allows the stem-celllike properties of NBs to be maintained, while ensuring their GMC progeny do not adopt stemcell-like characteristics, which would lead to overproliferation.However, the asymmetric segregation of Pros is not sufficient to account for its increased expression first in larval neurons (Carney et al., 2013) and later in pupal NBs to stop proliferation (Maurange et al., 2008).
The proliferation of all NBs, other than those leading to Mushroom body neurons, is terminated by a pulse of Pros expression and its entry into pupal NB nuclei.Elevated Pros drives the last NB division that leads to a pair of neurons (Maurange et al., 2008).This final function of Pros is crucial for normal brain development, as it prevents overproliferation of the CNS (Kohwi and Doe, 2013;Maurange et al., 2008).However, the mechanism that causes cell-type-specific upregulation of Pros protein is poorly understood, despite its central importance in control NB proliferation.
Here, we elucidate the mechanism of up-regulation of Pros expression in larval neurons and terminal pupal NBs by adapting smFISH to work well in whole-mount brains.This method allows us to quantitate the levels of primary transcripts and cytoplasmic mRNA in individual cells at high temporal and spatial resolutions.We found that pros is transcribed at similar levels in NBs, GMCs and neurons, but its total RNA and protein level is significantly higher in new born larval neurons and pupal NBs that are in their final round of division.The burst of pros mRNA consists of an unusually long isoform of pros mRNA (pros long ) that contains a 15 kb 3' UTR.pros long is not transcribed in embryos nor in larval NBs, but is switched on in post-mitotic larval neurons and terminally dividing pupal NBs.We show that pros long is stabilised by binding to the conserved RNA-Binding Protein (RBP), Syncrip(Syp)/hnRNPQ (Liu et al., 2015;McDermott et al., 2014;Kuchler et al., 2014).We show that pros long and Syp are both required to stop proliferation of terminal pupal NBs, by enabling a burst of Pros expression.Our observations highlight a novel mechanism of terminating stem cell division though transcription of a long isoform followed by regulated mRNA stability involving binding of a trans-acting factor to a long 3' UTR.

Up-regulation of Pros protein in neurons is achieved through post-transcriptional regulation
In order to examine cell-type-specific regulation of Pros expression, we first performed a careful analysis of antibody staining of Pros in the type I NB lineage.Pros is known to be localised in GMC nuclei, where it promotes GMC differentiation into neurons (Choksi et al., 2006).We found that in addition to its expression in GMCs, Pros is up-regulated in new born neurons in comparison to NBs and GMCs (Figure 1A,B).To test whether the increase of Pros in immature larval neurons is achieved by an increase in primary transcription, we adapted smFISH methods (Buxbaum et al., 2015) to work efficiently in whole-mount Drosophila larval brains.smFISH can be used to quantitate reliably the intensity of nuclear nascent transcript foci as a measure of primary transcription levels (Zenklusen et al., 2008) in comparison to the levels of total pros mRNA and protein in the NBs and their progeny.NBs, GMCs and neurons were distinguished unambiguously using expression of an asense-Gal4>mcD8::GFP reporter that selectively marks type I NBs and GMCs, in combination with cell-type-specific markers as previously described (Figure S1, see methods for details).These stains identify the cells in the NB lineage definitively: NB express Asense (Ase) and Deadpan (Dpn), GMCs express Dpn only and neurons express Elav.We found that Pros protein expression correlates well with total pros mRNA levels, showing low expression in NBs and four-fold higher expression in neurons (Figure 1B,C,E,F).Unexpectedly, the level of pros primary transcription is similar in all cell types (Figure peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; 1D,G), including in NBs, where both pros mRNA and Pros protein levels are significantly lower.
Given that cytoplasmic pros mRNA levels were elevated in neurons, while the level of primary transcription was equal in all cells in the NB lineage, we conclude that the up-regulation of Pros protein in differentiating neurons depends on post-transcriptional mechanisms.
pros mRNA in larval brains contains an exceptionally long 3' UTR Post-transcriptional regulation is often linked to expression of distinct alternative splicing or 3' UTR isoforms.Neurons have been shown to express transcript isoforms with unusually long 3'UTRs (Hilgers et al., 2012;Hilgers et al., 2011;Oktaba et al., 2015).In FlyBase, pros is annotated as having six distinct isoforms with three different length 3' UTRs: ~1 kb, ~3 kb and ~15 kb (Figure 2A).In addition, previous publications showed that embryos only express a single ~6 kb pros RNA isoform corresponding to a 200 nt 3' UTR (Doe et al., 1991;Vaessin et al., 1991;Matsuzaki et al., 1992).To test which isoforms are expressed in larval brains, we carried out Northern blots able to detect very long transcripts.Our Northern blot analysis using a probe that is common to all the annotated isoforms of pros transcripts (probe ALL), also detected a ~6 kb embryonic isoform (Figure 2A-B).The Northern blots also show that in larval brain extracts, there are multiple sized pros transcripts including approximately 6 kb, 9 kb and 21 kb in length (Figure 2B), corresponding to 3' UTR lengths of 200 nt, 3 kb and 15 kb, respectively.No band was found corresponding to the annotated 8 kb pros isoform containing a 1 kb 3' UTR.We also detected the 21 kb pros band in larval brains using a probe targeting a region that is unique to its predicted 15 kb 3' UTR.
The pros genomic sequence contains four possible canonical polyadenylation signals (PASs) downstream of the stop codons for all isoforms, corresponding to the four 3' UTR lengths discussed above.We tested the usage of these four PASs by sequencing the 3' end of RT-PCR clones generated using a primer that is specific to the upstream region of the predicted PAS and a generic oligo (dT)15 primer (Davis and Ish-Horowicz, 1991).The sequencing results are consistent with the Northern blot data, indicating that polyadenylation occurs at positions leading to 3' UTRs 200 nt, 3 kb and 15 kb in length, corresponding to pros isoforms 6 kb, 9 kb and 21 kb in length, respectively (Figure S2).Considering all our results together, we conclude that embryos express only pros very-short , containing a 3' UTR 200 nt in length, whereas larval brains contain two additional isoforms, pros short and pros long , containing 3' UTRs 3 kb and 15 kb in length, respectively.
To investigate the patterns of expression of the three pros isoforms in larvae, we performed smFISH experiments using probes specific to the exonic 3 kb and 15 kb 3' UTR region of the pros transcript.pros long , containing the 15 kb 3' UTR, is absent from NBs and GMCs and is only peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May. 12, 2017; expressed in neurons (Figure 2C. Figure S3).We compared pros long and total pros primary transcription using smFISH with intron probes specific for the long isoform and against an intron that is common to all isoforms.We only detected transcription foci that are specific for the pros long isoform in neurons and not in NBs and GMCs (Figure 2D).We conclude that the pros long isoform is specifically transcribed in larval neurons.
To determine which pros isoform, pros very-short or pros short is expressed in larval NBs and GMCs where the pros long isoform is absent, we compared the signal intensity generated by the 3 kb 3' UTR probe, specific to pros short and the exon probe common to both (Fig S3).If pros very-short is the predominant isoform, a higher signal is expected from the exon probe common to both isoforms.If pros short is predominant, a similar signal level is expected from both probes.Our results show that similar signal was detected in NBs and GMCs from the pros exon and 3 kb 3' UTR.We conclude that pros short with the 3 kb 3' UTR is most likely the predominant isoform in NBs and GMCs.

The pros long isoform is required for the up-regulation of Pros protein in larval neurons and is stabilised by its 15kb 3' UTR
To test whether pros long mRNA expression could account for the up-regulation of Pros protein in larval neurons, we first established whether the peaks of pros mRNA and protein occur in the same cells.We found a strong positive correlation (R 2 =0.821) between expression levels of pros long and Pros Protein (Figure 3A-B).To test directly whether the pros long isoform is required for neuron-specific up-regulation of Pros expression, we created a mutation in the endogenous pros gene that specifically abolishes pros long without affecting the expression of the other isoforms.The mutant introduces a stop codon in the unique N terminal protein coding region of the long isoform (pros long-STOP ) that causes the transcripts to be degraded by nonsense mediated mRNA decay (NMD) (Figure 3C and Materials and Methods).smFISH experiments confirm that pros long mRNA is greatly reduced from the cytoplasm of pros long-STOP mutants despite being transcribed normally (Figure 3D-G), suggesting that the mutant transcripts are made but quickly degraded.Importantly, Pros protein levels are substantially reduced in pros long- STOP mutants and not up-regulated in neurons (Figure 3H-I).We conclude that pros long is required for the cell-type-specific up-regulation of Pros protein in larval neurons.
To test whether the presence or absence of the extended 3' UTR of pros influences the abundance of pros mRNA, we constructed Green fluorescent protein (GFP) transgenes with either the 3 kb or 15 kb pros 3' UTR under actin-gal-4 / UAS control (Figure 4A).Our results show that the presence of the 15 kb 3' UTR increases the stability of the transcript in the cytoplasm.We quantitated primary transcription and total mRNA of the two transgenes using peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May. 12, 2017; smFISH.The results showed that the presence of the long 15 kb 3' UTR did not affect transcription rate of the construct, but increased the level of cytoplasmic mRNA 5.75-fold (p<0.001) and of the resulting GFP protein 3.65-fold (p<0.001)relative to the short UTR reporter (Figure 4B-D).We conclude that the 15 kb 3' UTR of pros mRNA is sufficient to greatly increase transcript stability, although this does not exclude additional affects on translation efficiency.
Together, these results show that larval neurons up-regulate Pros protein expression by specifically transcribing the pros long isoform, which is stabilised post-transcriptionally by the presence of its 15 kb 3' UTR.
The RNA-binding protein (RBP) Syp is expressed in the NB lineage, binds to pros mRNA and stabilises the pros long isoform Post-transcriptional regulation in general and mRNA stability in particular depend on the action of specific RBPs.We previously showed that pros mRNA associates with Syp, a highly conserved RBP, in larval extracts (McDermott et al., 2014).Mammalian SYNCRIP/hnRNPQ, is an important regulator of neural development (Liu et al., 2015;Lelieveld et al. ,2016;Stoiber et al., 2015), and has a number of post-transcriptional roles including regulating mRNA stability through binding at the 3' end of transcripts (Kuchler et al., 2014;Kim et al., 2011).For Syp to be a suitable candidate for regulating pros mRNA stability, it would have to be expressed and associated with pros mRNA in larval brains.We tested both these conditions and found that they are satisfied.Staining with an antibody against Syp, shows that the protein is highly enriched in the larval NBs and their progeny (Figure 5A).Immunoprecipitations (IPs) of pros mRNA using anti-Syp antibody shows that 58.3±4.5% of pros IPs with Syncrip in the larval brain, compared with negligible IP of the negative control rp49 (Figure 5B,C).Moreover, IPs with IgG controls only pulled down a small fraction of both pros (1.10±0.83%)and rp49 (0.97±0.54%)RNA.These results demonstrate that Syp binds strongly and specifically to pros RNA.
Considering the presence of Syp in the NB lineage and its binding to pros mRNA, we conclude that Syp is a good candidate for an upstream regulator of pros mRNA stability.
To test whether Syp influences the levels of pros mRNA, we performed smFISH using pros exon and intron probes in syp mutant and wild-type brains.The results show that pros nascent transcripts levels are similar in both cases (Figure 5D,G) but total pros mRNA is significantly reduced in syp mutant compared to wild type (Figure 5E,G).Complementary biochemical methods confirmed these results using pros RT-qPCR on nuclear and cytoplasmic fractions of syp mutants and wild type brains (Figure S4 A-C and Methods).We also used quantitative smFISH to test whether the absence of cytoplasmic pros RNA in syp mutants is due to a defect Page ! of !7 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May. 12, 2017; in nuclear mRNA export.pros mRNA is not retained in the nucleus in the absence of Syp (Figure S4 D-F).
We also compared Pros protein levels in wild type and syp mutant brains, which showed that the neuron specific Pros up-regulation is abolished in syp mutant (Figure 5F,G).We conclude that Syp mediates the up-regulation of Pros protein in neurons primarily by stabilising cytoplasmic pros mRNA, not by affecting the primary transcription or nuclear export of pros transcripts.
To test directly whether loss of cytoplasmic pros mRNA in syp mutants is due to the instability of the long isoform, we carried out transcriptomics analysis of syp mutant versus wild type brains.
The results show that the abundance of pros long is specifically reduced in syp mutants, whereas the shorter pros forms are unaffected (Figure S5 A-B and Table S5).Northern blot analysis confirmed these findings, showing that the pros long isoform is reduced significantly in syp mutants whereas the abundance of the shorter pros isoforms remain unchanged (Figure 5H).
We conclude that syp is required to specifically stabilise the long isoform of pros.
To test whether Syp is required for the activation of pros long transcription in neurons, we used an smFISH probe that is specific to the intron region of the pros long isoform.We found that the intensities of pros long primary transcription foci are the same in syp mutants and in wild type larvae (Figure S5 C-E).In contrast, cytoplasmic pros long mRNA as detected by smFISH probes specific to the long 3' UTR smFISH probe, is reduced significantly in syp mutants versus wild type.Taken together, our results show that transcriptional regulation of pros alone is not sufficient to produce the up-regulation peak of Pros expression observed in larval neurons.
Therefore, we suggest that Syp regulates pros long stability post-transcriptionally.
We have demonstrated that both Syp and pros long are required to maintain high Pros protein level in larval neurons.We tested whether Syp and pros long are required to drive neuronal differentiation and to prevent neurons from reverting back to NBs, as is the case for some previously characterised mutants, such as midlife crisis (Carney et al., 2013).We quantitated the number of NBs and their rate of proliferation in larvae of pros long-STOP and syp mutants and found no significant difference in NB numbers or mitotic index, compared to wild type.We used the NB specific anti-Deadpan (Dpn) antibody, and anti-Phospho-Histone3 (PH3) antibody to identify and mark mitotic NBs (Figure 6A,C-D).We conclude that pros long , is not required to maintain larval neurons in a differentiated state, nor to control the rate of larval NB proliferation.
Page ! of ! 8 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

pros long is essential for pupal NBs to exit division
A burst of Pros expression is known to have an important role in terminating NB divisions in pupae.To test whether Syp or pros long also have a role in terminating pupal NB division, we quantitated NBs in syp and pros long-STOP mutant pupal brains compared with wild type.We found that in syp and pros long-STOP mutant NBs continue to divide, when NBs in wild type brains have already terminated.As previously published, we found that 48 hours after pupal formation (APF), in wild type pupal brains there are only 4 mushroom body NBs remaining in each hemisphere, while all other NBs have undergone terminal symmetric division to produce two neurons (Siegrist et al., 2010).In contrast, in both syp and pros long-STOP pupal brains, we found a large number of NBs remain and there is no significant change in mitotic index from the wild type larva, suggesting that these long-lived NBs continue to divide at the same rate (Figure 6B,C-D).In addition, we observed a significant number of PH3 + Dpn -cells adjacent to these long-living NBs in syp and pros long-STOP mutants (yellow arrows, Figure 6B).These cells are likely to be GMCs produced by the remaining NBs through regular asymmetric division.We conclude Syp and pros long are required for the termination of NB division in pupae.
Previous work showed that termination of pupal NBs division requires a transient up-regulation of Pros protein and that expression of transgenic Pros can terminate NBs at any stage (Maurange et al., 2008).To test whether a transient up-regulation of Pros in pupal NBs is achieved by the activation of pros long isoform transcription and subsequent transcript stabilisation by Syp, we performed smFISH with intron probes that are specific to the pros long isoform in conjunction with Dpn and Pros protein staining.In wild type pupal brains, we found that such transient nuclear Pros up-regulation in terminal pupal NBs correlates with the activation of pros long transcription (Figure 7A,E).In contrast, in pupal NBs that lack pros long mRNA, Pros protein level remains low (Figure 7B,E).In syp mutant pupal NBs, pros long mRNA is transcribed, but Pros protein fails to up-regulate, apparently due to failure of pros long mRNA to be stabilised in the absence of Syp (Figure 7C-D,E).Taking all our results together, we conclude that in pupae, NBs activate transcription of pros long and the transcript is stabilised by Syp binding to the 15 kb 3' UTR.Therefore, pros long expression and stabilisation lead to the transient upregulation of Pros protein that drives the final symmetric division of pupal NBs.
Page ! of ! 9 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

DISCUSSION
Many key regulators of NB proliferation and differentiation, in particular transcription factors (TFs), have now been identified and characterised (Kambadur et al., 1998;Isshiki et al., 2001;Novotny et al., 2002;Pearson and Doe, 2003;Grosskortenhaus et al., 2006;Li et al., 2013;Narbonne-Reveau et al., 2016;Kohwi and Doe, 2013).A general consensus has emerged in the field for both Drosophila and mammals, namely that a temporal sequence of TFs is predominantly responsible for controlling the gene expression landscape that governs NB proliferation and differentiation.Our use of quantitative smFISH methods opens up the possibility of addressing conclusively the roles of primary transcription and post-transcriptional regulation at an individual cell level within the intact brain.We have highlighted how a key aspect of the expression and function of the TF Pros is regulated by the stability of its long mRNA isoform (pros long ).These findings suggest that multiple, hitherto unappreciated layers of post-transcriptional regulation are likely to be important for regulating other aspects of stem cell proliferation and differentiation.Indeed, the generality of our findings is supported by the fact that many genes that function in the brain have complex gene structures such as multiple isoforms and long 3' UTRs, hallmarks of post-transcriptional mechanisms (Stoiber et al., 2015;Hilgers et al., 2011;Tekotte et al., 2002;Berger et al., 2012).

Post-transcriptional and transcriptional regulation work hand-in-hand to achieve cell specific gene expression
Our experiments show that pros long is selectively transcribed in larval neurons and in terminally dividing pupal NBs.It is currently unclear what induces the activation of this tissue specific pulse of pros long transcription, but we show that once transcribed, the presence of pros long is regulated through an exceptionally long 15 kb 3' UTR, which is stabilised by Syp.The use of long 3' UTRs has been widely described in other biological contexts (Yeh and Yong, 2016;Hilgers, 2015).
Moreover, gene specific 3' UTR extension has been reported during Drosophila embryonic neurogenesis (Hilgers el al., 2011;Smibert et al., 2012;Oktaba et al., 2015).In the case of pros, we have demonstrated a functional role for its long 3' UTR in stabilising pros long , required for NB termination.We hypothesise that the functional role of 3' UTR extension will also be important for other prominent examples of Drosophila neural genes with annotated extended 3' UTR isoforms.These include the regulator of the timing of neuronal development, Chinmo (Liu et al., 2015), and the mRNA-binding proteins with important roles in post-transcriptional regulation and NB biology, Brat (Bello et al., 2006;Betschinger et al., 2006;Lee et al., 2006) and Imp (Liu et al., 2015), all of which have complex isoform structures.Therefore, our data in the context of previous studies suggest a general principle of post-transcriptional regulation of the expression Page ! of ! 10 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; of many key neuronal transcripts, through cell-type specific presence of isoforms with distinct 3' UTRs. Previous work has shown that the loss of Pros in larval neurons is not sufficient to lead to their de-differentiation back to NBs (Carney et al., 2013).In agreement with these observations, we found that pros long-STOP mutants also do not cause neurons to de-differentiate into NBs.We interpret these observations as indicating that pros long has no essential role in the differentiation of neurons.Instead, pros long mRNA stabilisation is specifically required in pupal NBs to terminate their divisions by boosting the expression of Pros at a particular place, time and cell type in development.Therefore, pros long mRNA stability is regulated hand-in-hand with the choice of which pros isoform is transcribed, to enable the uncoupling of low level Pros expression required for neuronal differentiation from a specific burst of Pros expression needed to terminate NB division.In addition, the dependence of pros long , but not the shorter pros isoforms, on Syp for its stabilisation, allows further separation of control of pros long levels from those of the other isoforms.

Regulating expression levels by differential mRNA isoform stability and long 3' UTRs is likely to be a conserved mechanism
Although regulated mRNA stability has not been well characterised mechanistically in neural stem cells, there are many examples of unstable mRNAs in a variety of biological and medical contexts (Ulitsky et al., 2012;Miura et al., 2014).For example, regulated mRNA stability is very common during the early embryonic development of Drosophila (Chen et al., 2014;Laver et al., 2015).Moreover, in single cell models, including yeast, Drosophila S2 cells and mammalian tissue culture cells, mRNA stability is known to play many varied and important roles (Chávez et al., 2016;Sin and Nollen, 2015;Forrest and Khalil, 2017;Moszyńska et al., 2017;Maeda and Akira, 2017) and often involves isoforms with distinct 3' UTRs (Catalan et al., 2016;Hopkins et al., 2016;Mitra et al., 2015).More generally, post-transcriptional mechanisms are thought to be essential for many neuronal functions, ranging from synaptic plasticity (Capitano et al., 2017; Lee et al., 2 0 1 5 ) to neurodegeneration (Cookson, 2017;Tang, 2016;Kim et al., 2 0 1 1 ; Nussbacher et al., 2 0 1 5 ) and neurological diseases (Wang et al., 2 0 1 6 ).Therefore, we anticipate that regulated mRNA stability and the use of isoforms with distinct 3' UTRs will also play major and wide-spread roles in both vertebrate and invertebrate stem cell biology, especially in the nervous system.
Drosophila and mammals share many molecular mechanisms driving nervous system development.We propose that the post-transcriptional mechanisms we have uncovered are also conserved in mammals.Certainly, Prox1, the mammalian TF orthologue of Pros, is a tumour suppressor that regulates stem cell differentiation in the brain as well as many other organ systems (Stergiopoulos et al., 2015;Elsir et al., 2012).Like pros, prox1 also has several isoforms including one containing a long 3' UTR.Furthermore, a burst of Prox1 expression is required to drive the differentiation of immature granular neurons in the adult hippocampus (Hsieh, 2012).It is also likely that post-transcriptional regulation by RBPs helps determine the expression and translatability of Prox1.Future experiments will have to discover whether Prox1 expression, like pros, is regulated through stabilisation of its 3' UTR in order to terminate neural stem cell proliferation.
Our newly established quantitative application of smFISH to the Drosophila brain enables a new level of characterisation of post-transcriptional regulation mechanisms, also in other systems.
Indeed, there is increasing evidence that RNA-binding proteins play critical roles in Drosophila and mammalian brain development (Stoiber et al. 2015;Berger et al., 2012;Neumuller et al., 2011;Bryant et al., 2016), and that their dysfunction is associated with neurological disease (Nussbacher et al., 2015).We expect that applying smFISH in mammalian models will transform studies of the regulation of gene expression in complex tissues.
Page ! of ! 12 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

CONTACT FOR REAGENT AND RESOURCE SHARING
For further information and access to any reagents and fly strains used in this manuscript, please contact co-corresponding author Ilan Davis (ilan.davis@bioch.ox.ac.uk).
Ligations of multiple inserts were accomplished using the In-Fusion cloning kit (Clonetech).Total reaction mixture varied between 10 to 20 µl.Reaction mixture was incubated for 15 min at 50 ˚C and was immediately used for transformation into Stellar TM competent cells (Clonetech).
Plasmid DNA was extracted using the QIAprep Spin Miniprep Kit (Qiagen) following the manufacturer's instructions.For generating BAC insertion transgenic flies, embryo injection was done by GenetiVision (Houston, Texas, USA).For each genotype, between 400-600 embryos were injected and screened.
pros long-STOP mutant generation using CRISPR sgRNA construct design and validation was performed by Dr. Andrew Bassett -Genome Engineering Oxford (GEO).sgRNA (CCTCACAAGTCAACGGCGACGGC) targeting coding region that is unique to the pros Long isoform was injected into 400 vasa-cas9 embryos (Bloomington Stock: BL55821) as previously described (Bassett and Liu, 2014).Surviving embryos were crossed to a double-balancer stock TM3/TM6 and progeny were screened using high-resolution melt analysis (HRMA).Candidate mutant PCR products from the HRMA analysis were sequenced to confirm the introduction and location of the STOP codon.

Antibodies
The following primary antibodies were used: guinea pig anti-Syp (raised in the Davis lab as

RNA extraction
RNA extraction was performed using the RNAspin RNA Isolation kit (GE Healthcare).Modified manufacture instructions for RNA purification form cultured cells were followed with modifications to accommodate Drosophila samples.3rd instar larval brains were homogenised in 30 µl of RA1 buffer including 2 µl 40 U/µl RNAse inhibitor (RNAsin® Plus RNase Inhibitor (Promega)).Additional RA1 buffer was added to lysate after homogenisation to make up a total volume of 350 µl.An equal volume of 70% ethanol was added to the lysate and pipetted up down several times to allow good mixing.The lysate/ethanol solution was loaded onto the RNAspin Mini column and centrifuged for 30 s at 8,000 x g.Following centrifugation, the flow peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; through was discarded and the spin column was transferred to a fresh 1.5 ml collecting eppendorf tube.350 µl of membrane desalting buffer (MDB) was added to the column followed by 1 min centrifugation at 11,000 x g.In column DNA digestion was done by directly adding a 100 µl of DNase digestion solution (10 µl reconstituted DNAse I; 90 µl DNAse reaction buffer) to the RNAspin column and allowed to incubate at RT for 15 min.The column was subsequently washed with 200 µl RA2, 600 µl RA3 and 250 µl of RA3 buffer, each wash followed by centrifugation at 11, 000 x g for 1 min.After final wash, 40 µl of nuclease free H2O was applied to the column to elute the bound RNA and elute was collected by centrifugation at 11,000 x g for 1 min.To increase yield, elute was re-applied to the column and eluted a second time.
Concentration of RNA was measured on either Nano-drop or RNA Fluorometer (Qubit).
Extracted RNA could then be flash frozen with liquid nitrogen and stored at -80 ˚C or directly processed for cDNA library synthesis or Northern blot.
Following SDS-PAGE, proteins were transferred to nitrocellulose membrane using the XCell II™ Blot Module (Invitrogen) following manufacturer's protocol.Protein bands were visualised with the quantitative infrared imaging system (LI-COR Odyssey, LI-COR Biosciences; Lincoln, NE).
Intensity of protein bands was quantified using the LI-COR software.
RNA was transferred to pre-wet Nylon membrane (Hybond-N, Amersham).Blotted RNA was cross linked to the membrane with either a UV crosslinker or by baking at 80 ˚C for >1 hr.
The membrane was hybridised in hybridisation buffer (0.4 M Na2HPO4, 6% SDS and 1 mM EDTA) for 1 hr at 57 ˚C with RNA probes prepared using the Prime-a-Gene kit (Promega) (For list of primers used to generate RNA probes, see Table S1), washed twice with wash buffer 1 (40 mM Na2HPO4, 5% SDS and 1 mM EDTA) and twice with with wash buffer 2 (40 mM peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; protease inhibitor and 2 µl RNAse inhibitor (RNAsin® Plus RNase Inhibitor (Promega)) and topped up to 100 µl in IP buffer.For each reaction, 20 µl of lysate was taken as a 50% input sample which was taken directly to RNA extraction.40 µl of lysate was incubated with 25 µl of each antibody cross-linked beads over night at 4 ˚C on rotator wheel.For each reaction, 100 µl of lysate was incubated with 20 µl of 50% bead slurry over night at 4 ˚C on rotator wheel.Next day, supernatant was transferred to fresh tubes and beads were washed 5 times for 5 min each with 200 µl cold IP buffer at 4 ˚C.After final wash, beads were resuspended in 40 µl extraction buffer (50 mM Tris-HCl pH 8.0, 10 mM EDTA and 1.3% SDS, 1:100 RNAsin) and incubated at 65 ˚C, 1000 rpm for 30 min on thermomixer.The elution step was repeated and the supernatant were pooled.RNA was then extracted from the IP eluates and the input sample and used for cDNA library synthesis.

RNA Sequencing
Three biological replicates (n=3), each replicate consists of a pool of 100 larval brains.
Following RNA extraction, mRNA was enriched using NEB Next® Poly(A) mRNA Magnetic Isolation Module (NEB).Briefly, extracted RNA sample was mixed with Oligo d(T)25 beads and heated to 65 ˚C for 5 min followed by incubation at 4 ˚C for 1 hr to allow binding.Following incubation, the beads were washed 5 times for 5 min each at 4 ˚C and RNA was eluted by heating the beads at 80 ˚C for 5 min.Poly(A) enriched RNA was then used for library production using the Ion Total RNA-Seq Kit v2 for Whole Transcriptome Libraries (Life Technologies).
Following quality control steps, adaptors were hybridised to the RNA fragments and RT reaction was performed followed by cDNA amplification with Ion Xpress RNA Barcode primers.Prior to sequencing, quality of cDNA libraries were assessed using Agilent High Sensitivity DNA Kit with the Agilent 2100 Bioanalyser.Libraries were pooled to a total concentration of 100 pM, with three samples multiplexed per chip.Sequencing was performed on an Ion Proton Sequencer, using the Ion PI IC 200 Kit (Life Technologies).Ion PI chips were prepared following manufacturer's instructions and loaded using the Ion Chef System.

Poly(A)-site mapping
Poly(A)-site mapping was carried out using the method described in (Davis and Ish-Horowicz, 1991).In short, RNA was extracted from wild type larval brains and cDNA library was generated using a poly(T) primer containing a 5'G-C rich region and a NotI site.PCR reactions were subsequently carried out using the same poly(T) primer and a forward gene specific primer.The cDNA was purified using a standard PCR program except that instead of high fidelity Phusion DNA polymerase, 0.125 U Taq DNA polymerase (NEB) was used in a 25 µl volume (for peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; complete list of primers see Table S2).To optimise PCR, magnesium titration between 1 to 3 mM in 0.5 mM increments was tested.To increase the specificity of the PCR, nested PCR was performed using the product from the first PCR reaction.Product from nested PCR reaction was resolved on a 2.5% agarose gel and bands were extracted using the QIAquick Gel Extraction Kit (Qiagen) following manufacture's instructions.Extracted DNA product was cloned into a TOPO ® vector using a TA cloning kit (Invitrogen).Single colonies was selected and cultured in LB medium containing the appropriate antibiotics.Plasmid DNA was extracted using QIAprep Spin Miniprep kit (Qiagen) and sequenced by Sanger Sequencing (Source BioScience).

Cytoplasmic and Nuclear Fractionation of Larval Brains
30 dissected larval brains were dissociated to single cells by first incubating in 1 mg/ml liberase (mixture of collegenase I and II; Roche) for 10 min at 37 ˚C followed by pipetting up and down 100 times.Following dissociation, cells were collected by centrifugation at 8,000 rpm for 10 min at 4 ˚C and supernatant were discarded.Dissociated cells were resuspended in 20 µl of hypotonic Buffer A (10 mM HEPES, 1.5 mM MgCl2, 10 mM KCl, 0.5 mM DTT and 0.5 µl RNAse inhibitor, buffer prepared in ddH2O) and incubated on ice for 5 min.Cells were then homogenised to release cytoplasmic content and nuclei were pelleted by centrifugation at 1,000 rpm for 5 min at 4 ˚C.Supernatant was removed to fresh tubes (Cytoplasmic fraction).The nuclei pellet was resuspended in 20 µl S1 sucrose buffer (0.25 mM sucrose, 10 mM MgCl2 prepared in dH2O containing protease inhibitor and RNAse inhibitor), layered over a 40 µl sucrose cushion (0.88 mM sucrose, 0.5 mM MgCl2 prepared in dH2O including protease and RNAse inhibitor and centrifuged at 2,800 rpm for 10 min at 4 ˚C.The supernatant was discarded and the clean nuclei pellet re-suspended in Buffer A. The fractionated cytoplasmic and nuclei fraction were used directly for subsequent protein analysis or RNA extraction.

Immunofluorescence
Third instar larval brains were dissected in Schneider's medium and fixed in 4% formaldehyde (FA) solution (4% FA in 0.3% PBTX) for 20 min at room temperature.Following incubation with primary and secondary antibody, samples were mounted in VECTASHIELD anti-fade mounting medium (Vector Laboratories).

Identification of larval neuroblast, ganglion mother cells and neurons
Neuroblasts (NBs) were identified based on their unique shape and location in the larval brain.

RNA single molecule fluorescent in situ hybridisation (smFISH) for larval brains
Dissected 3rd instar larval brains were fixed in 4% paraformaldehyde (PFA) solution for 30 min at room temperature followed by 3 rinses in 0.3 % PBTX (PBS and 0.3% triton).Samples were washed 2 times for 20 minutes each in 0.3% PBTX at RT and incubated in pre-hybrdisation buffer (10% deionised formamide in prepared in 2x SSC) for 5 min at 37 ˚C.Hybridisation was performed by incubating samples overnight at 37 ˚C in the dark with gentle shaking in hybridisation buffer (10% deionised formamide, 5% dextran sulphate (sigma), 2x SSC) containing 250 nM gene specific fluorescently labeled Stellaris® DNA probe set (BioSearch Technologies).Following hybridisation, samples were rinsed 3 times in pre-hybridisation buffer and washed for a further 3 times, 30 min each time in pre-hybridisation buffer at 37 ˚C.DAPI was included during the second 30 min wash.A final two washes were performed using 2x SSC for 10 min each at room temperature and samples were mounted in VECTASHIELD anti-fade mounting medium (Vector Laboratories) and immediately imaged using point scanning confocal microscope.Samples were protected from light for all steps including hybridisation.

smFISH with immunofluorescence for larval brains
The protocol is identical to the smFISH protocol except for the following modifications.Before hybridisation, samples were blocked in blocking buffer (1% BSA in 0.3% PBTX) for 1 h at room temperature.Antibody at the appropriate dilution was included with the Stellaris® DNA probes during the over night hybridisation step.Counter-stain with secondary antibody was performed on the second day following hybridisation.After final wash, samples were mounted in VECTASHIELD anti-fade mounting medium (Vector Laboratories) and immediately imaged.

Image acquisition
Fixed imaging of larval brain was performed using an inverted Olympus FV1000 Laser Scanning Microsope with Becker and Hickel FLIM system and with an inverted Olympus FV1200 Laser Scanning Microscope with high sensitivity galium arsenide phosphide (GaAsP) detectors (Olympus).Images were acquired using x20 0.

Quantification and statistical analysis
Except for analysis of RNA sequencing data (see section "Bioinformatic analysis of RNA sequencing data"), all statistical analysis was performed using Microsoft Excel.For all bar graphs unless otherwise stated, data represents means ± standard error of the means (SEMs).
Two tailed Student's t-test was used to compare different genotypes or cell types.Significance was classified as p<0.05.

Data analysis for Immunoprecipitation
For each of three biological replicates, a dilution series of the input sample was produced (10%, 2%, 0.5%.0.01% of input).qPCR for each set of primers was performed on this series and the Cq values were plotted against log10 dilution to find the formula of the line.The Cq value of each pulldown sample (anti-Syp and anti-IgG) was inputted into this formula to calculate the % of input pulled down.For each set of primers, a two tailed Student's t-test was used to compare the % input pulldown in the test (anti-Syp) and control (anti-IgG) samples.

Bioinformatic analysis of RNA sequencing data
Base calling, read trimming and sample de-multiplexing was done using the standard Ion Torrent Suite.The reads were aligned to the Drosophila genome (ENSEMBL assembly BDGP5, downloaded 8 Jan 2015) using the TopHat aligner (v2.0.13) (Kim et al. 2013).To quantitate gene expression, uniquely aligned reads were assigned to the Drosophila exome (BDGP5) using htseq-count.Differential was assessed using negative binomial generalised linear models implemented in R/Bioconductor package DESeq2 (Love et al. 2014).

Image analysis for Immunofluorescence and smRNA FISH
The intensities of nascent transcript foci and cytoplasmic mRNA were analysed using method described in Zenklusen et al., 2008;Trcek et al., 2011 to FIJI and average intensity value for each pre-defined ROI was calculated using the "Measure" algorithm in FIJI.
Phenotype analysis for wild type, syncrip and pros long-STOP l a r v a l a n d p u p a l mutant Flies were allowed to lay for 2 h at 25 ˚C on apple juice plate.Hatched larvae were subsequently transferred to bottles containing standard fly food and larvae/pupae were removed at the appropriate time point for dissection.Two developmental stages were investigated: 72 h (3 rd instar) after larval hatching (ALH) and 48 h after pupal formation (APF).Brains were dissected and fixed in 4% FA and subsequently stained with the neuroblast specific marker Deadpan (Dpn), and the mitotic marker phospho-histone 3 (PH3).Images of the ventral brain were acquired using Olympus point-scanning confocal microscope at 2 µm steps.Image analysis was completed using Imaris Image Analysis software (Imaris, Bitplane).For counting the total number of NBs, and cells that are both positive for Dpn and PH3, the "Point" tool in Fiji program was used.For each genotypes, three brains (n=3) was analysed.

DATA AND SOFTWARE AVAILABILITY
We are currently acquiring a Gene Expression Omnibus (GEO) accession number for the presented RNA sequencing data and will provide the number prior to publication.
Page ! of !39 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

Figure 3 :Figure 4 :
Figure 3: pros long isoform is required for the up-regulation of Pros protein in larval neurons (A) Co-staining of Pros immunofluorescence with pros smFISH with UTRlong probe specific to pros long isoform, show Pros protein expression level is increased where there is expression of the more stable p r o s transcript isoform with extended 15 kb 3' UTR. ( B ) P r o s p r o t e i n e x p r e s s i o n l e v e l i s positively correlated with the expression pros long (C) Schematic of pros long-STOP mutant.A point mutation resulting in the formation of a stop c o d o n i s i n t r o d u c e d d o w n s t r e a m o f t h e transcription start site (TSS) of the pros long isoform but upstream of the pros short isoform.This will induce degradation of the pros long isoform transcripts via the NMD pathway.(D-G) RNA FISH confirming the successful elimination of the pros long transcripts.Only transcription foci but not cytoplasmic mRNA, were detected using UTR-long probe that is specific to the pros long isoform in the pros long-STOP mutant.Most cytoplasmic pros long-STOP transcripts are degraded by the NMD pathway following transcription.(H-I) Pros protein level is significantly reduced in pros long-STOP mutant and the neuron specific Pros up-regulation is abolished.

Figure 5 :
Figure 5: Syncrip selectively stabilises the pros long isoforms in neurons (A) Syncrip (Syp) protein is enriched in NB and progeny in 3 rd instar larvae.Scale bars: 20 µm and 10 µm (B) Western blot confirming Syp can be immunoprecipitated (IP) selectively and efficiently.⍺-Tubulin (Tub) was used as negative control.(C) pros RNA is enriched by Syp IP but not by IgG control beads.Negative control, ribosomal rp49 (n=7).(D) Syp does not affect primary transcription levels of pros, (E) but cytoplasmic pros mRNA levels are significantly reduced in syp mutants.(F) Up-regulation of Pros protein in neurons is abolished in the absence of Syp.(G) Quantitation of change in pros nascent transcripts, total mRNA and protein levels in syp mutant.(H) Northern blot showing the selective loss of pros long isoforms in syp mutants.The intensity of the 6.5 kb and 9 kb bands are similar between wild type and syp mutants (blue asterisks) whereas the 21 kb band is significantly reduced in syp mutants (red asterisk).

Figure 6 :
Figure 6: pros long is not required to maintain larval neurons in differentiated states but is essential for pupal NBs to exit division (A-B) NBs and dividing cells in the central brain region (grey dotted line) in larval (A) (72 hours after larval hatching) and pupal (B) (48 hours after pupal formation) brains, are labeled for Deadpan (green) and phospho-histone 3 (red), respectively.Inserts show clusters of dividing (yellow dotted outline) and non-dividing (white outline) NBs.Note Deadpan becomes faint and cytoplasmic in dividing NBs.PH3 + Dpn -cells were also detected in close proximity to NBs, these cells are likely to be dividing GMCs produced from NB asymmetric division (yellow arrows).Scale bar: 50 µm.(C-D) Total NB number (C) is significantly increased in syp and pros long-STOP pupae, while mitotic index (D) is unchanged from the wild type larva, as NBs failed to terminate proliferation.

Figure 7 :
Figure 7: Transient up-regulation of Pros in terminal pupal NBs is achieved through Syp-dependent stabilisation of pros long (A-B) Up-regulation of Pros protein in pupal NBs correlates with the activation of pros long transcription (yellow arrow) in wild type pupal brain.(C-D) pros long transcription (intron B probe, see Figure 2A) is not affected in syp mutant but nuclear Pros fails to up-regulate.(E) Quantification of Pros protein level in pupal NBs before and after pros long transcription activation; n=5 for Pros immunofluorescent quantification.(F) Graphical summary of pros long stabilisation by Syp triggering pupal NB exit from proliferation.

Figures
Figures S1: related to Figure 1

Figure
Figure S2: related to Figure 2

Figure
Figure S3: related to Figure 2

Figure
Figure S5: related to Figure 5 Figure S6: related to Figure 5

Figure S4 :
Figure S4: Biochemical confirmation of selective loss of cytoplasmic pros RNA in syp mutant Page ! of !11 47peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
production, the Bloomington Drosophila Stock Centre, P[acman] Resources.L.Y. was funded from the Clarendon Trust and Goodger Fund., T.J.S. was funded by Wellcome Trust Four-Year PhD Studentship (105363/Z/14/Z), F.R. was funded by a Marie Curie Postdoctoral Fellowship.C.P.Y. and T.L. were supported by Howard Hughes Medical Institute.D.I.H. was funded by University College London.I.D. was funded from a Wellcome Senior Research Fellowship (081858), which also supported A.I.J., MICRON Oxford (http://micronoxford.com, supported by a Wellcome Strategic Awards to ID (091911/B/10/Z and 107457/Z/15/Z) providing access to equipment and advice on advanced imaging techniques.
The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi: bioRxiv preprint first posted online May.12, 2017; pros Long +'ve NB pros Long -'ve NB peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
Page ! of ! 31 47 peer Page ! of !36 47peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.
. Briefly, for nascent transcript analysis, 3D image intron smFISH data set was first uploaded onto Imaris Image Analysis software (Imaris, -reviewed) is the author/funder.All rights reserved.No reuse allowed without permission. of interest.The coordinates of the mask was saved and imported as region of interest (ROI) image data was first reduced to a 2-dimensional image by maximum Z projection in FIJI (ImageJ 1.45r; M).Using membrane-tethered protein mCD8 staining which labels the nuclear and cytoplasmic membrane, a hand-drawn mask was created for the cytoplasmic region of the Page ! of !38 47 peercell

Table S1 -3 Included as separate file:Table S4 (
as Word file): Sequence used to generate Stellaris® DNA probe set

Table S5 (
as Excel file): RNA Sequencing results from wild type and syncrip mutant larval brains Page ! of !40 47 peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.

Table S1 : Primers for generating Northern blot probes
May. 12, 2017;47peer-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi:bioRxivpreprintfirst posted online May.12, 2017;The above primers are used to generate radioactive Northern blot probes.ORF -Open Reading Frame of pros transcript; UTR Long are used to generate probes that are specific to the 15 kb pros 3' UTR (probe LONG); UTR Short are used to generate probe that is against region that is common to all pros 3' UTRs (probe ALL).-reviewed) is the author/funder.All rights reserved.No reuse allowed without permission.The copyright holder for this preprint (which was not .http://dx.doi.org/10.1101/135848doi:bioRxiv preprint first posted onlineMay.12, 2017; peer

Table S2 : Primers for Poly(A) site mapping
Oligo d(T) was used in pair with primers that is specific to pros 3' UTR.Four predicted poly(A) sites were tested occurring 212 base-pair (bp) (pros 200F), 1355 bp (pros 1 kb F), 3265 bp (pros 3 kb F) and 15,530 bp (pros 15 kb F) after stop codon.

Table S3 : Primers used for RT-qPCR for Syncrip and IgG immuno-precipitation experiments pros
and housekeeping gene rp49 was used to assess the efficiency and specificity of Syncrip binding to pros.