key: cord-0262193-y1fj1jx3
authors: Haussmann, Irmgard U.; Wu, Yanying; Nallasivan, Mohanakarthik P.; Archer, Nathan; Bodi, Zsuzsanna; Hebenstreit, Daniel; Waddell, Scott; Fray, Rupert; Soller, Matthias
title: CMTr cap-adjacent 2’-O-ribose mRNA methyltransferases are required for reward learning and mRNA localization to synapses
date: 2021-06-24
journal: bioRxiv
DOI: 10.1101/2021.06.24.449724
sha: 8aa1880c4d7655b8a982e623b1bbb22d40c34daf
doc_id: 262193
cord_uid: y1fj1jx3

Cap-adjacent nucleotides of animal, protist and viral mRNAs can be dynamically O-methylated at the 2’ position of the ribose (cOMe). The functions of cOMe in animals, however, remain unknown. Here we show that the two cap methyltransferases (CMTr1 and CMTr2) of Drosophila can methylate the ribose of the first nucleotide in mRNA. Double-mutant flies lack cOMe but are viable. Consistent with prominent neuronal expression, they have a reward learning defect that can be rescued by conditional expression in mushroom body neurons before training. Among CMTr targets are cell adhesion and signaling molecules relevant for learning and cOMe is required for local translation of mRNAs at synapses. Hence, our study reveals a mechanism to co-transcriptionally prime mRNAs by cOMe for localized protein synthesis at synapses.

Methylation of cap-adjacent or internal nucleotides in messenger RNA (mRNA) is a major 47" post-transcriptional mechanism to regulate gene expression. Methylation of mRNA is 48" particular prominent in the brain, but the molecular function of methylated nucleotides and 49" their biological roles are poorly understood 1-5 .

Methylation of cap-adjacent nucleotides is an abundant modification of animal, protist and 51" viral mRNAs, that varies in different tissues and transcripts [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] . Dynamic O-methylation at the 52" 2` position of the ribose (cOMe) of cap-adjacent nucleotides is introduced co-transcriptionally 53" by two dedicated cap methyltransferases (CMTr1 and CMTr2) after capping at the beginning 54"

of an mRNA to a characteristic 5`-5` linked N7-methylated guanosine [18] [19] [20] .

The main function of the cap is to protect mRNAs from degradation and to recruit translation 56"

initiation factors, but also to promote splicing and 3` end processing 21 . The cap is initially 57" bound in the nucleus by the cap binding complex (CBC), consisting of CBP20 and CBP80.

Upon export from the nucleus, CBC is replaced by eIF4E, which is predominantly cytoplasmic 59"

and rate-limiting for translation initiation 22, 23 . N7-methylation of the cap guanosine is critical 60"

for both CBC and eIF4E binding. The importance of cap-adjacent nucleotide methylation in 61" animal gene expression, however, remains elusive, but is known to be essential in 62" trypanosomes and viruses including SARS-CoV-2 for propagation 15 . 63"

CMTrs act redundantly 66"

To elucidate the biological function of cap-adjacent 2`-O-ribose methylation (cOMe) in 67" animals we made null mutants of the CMTr1 (CG6379) and CMTr2 (adrift) genes in 68"

Drosophila. We generated small intragenic deletions in each gene by imprecise excision of a 69" P-element transposon to make CMTr1 13A and CMTr2 M32 mutant flies (Fig. 1a-c) . Both of these 70" aversive learning performance of CMTr1 13A ; CMTr2 M32 also suggest that cOMe deficiency 121" somehow specifically impairs reward learning.

Olfactory learning and memory in Drosophila is coded within the neuronal network of the 123" mushroom bodies (MBs) 30 . Valence learning can be coded as changes in the efficacy of 124" synaptic junctions between odor-activated Kenyon Cells (KCs, the intrinsic cells of the MB) 125"

and specific mushroom body output neurons. We therefore tested whether the reward learning 126" defect of CMTr1 13A ; CMTr2 M32 mutant flies could be rescued by restoring cOMe expression to 127"

KCs. Expressing a UAS-CMTr2 transgene in the KCs using MB247-GAL4 rescued the learning 128" deficits of CMTr 13A ; CMTr M32 double mutant flies (Fig. 2d) .

Next, we investigated whether the reward learning phenotype of CMTr 13A ; CMTr M32 double 130" mutant flies arose from a developmental origin, or from loss of an acute function in the adult 131"

stage. The gross morphology of the adult MBs appears to be normal in CMTr1 13A ; CMTr2 M32 132" mutants as judged from expressing a UAS-EGFP transgene with MB247-GAL4, or with the 133"

KC-subtype restricted drivers NP7175-GAL4 (!" core KCs), 0770-GAL4 (!" surface KCs) or 134" 1471-GAL4 (# KCs, Supplementary Fig. 3a) . Interestingly, restoration of CMTr2 expression 135"

to these more restricted KC subsets did not rescue the learning defect of CMTr 13A ; CMTr M32 136" double mutant flies (Supplementary Fig. 3b) .

We next tested whether the reward learning defect of CMTr 13A ; CMTr M32 double mutant flies 138" could be rescued by inducing CMTr2 expression just before training in adult flies. Since 139"

MB247-GAL4 was able to restore learning, we employed a MB247-driven Gene-Switch (GS) 140"

to conditionally induce CMTr2 expression by feeding flies with RU486. Only CMTr 13A ;

CMTr M32 flies that also harboured the MB247-GS and UAS-CMTr2 transgenes exhibited 142" restoration of memory performance when fed with RU486 (Fig. 2e) . Together these 143" experiments suggest that cOMe in the MB KCs plays a key role in olfactory reward learning. 144"

To investigate the impact of cOME on gene expression, we performed RNA sequencing on 147" cOMe deficient and control flies. Differential gene expression analysis revealed 197 and 701 148" genes that were significantly down-and up-regulated in CMTr 13A ; CMTr M32 double mutant flies 149" as compared to wild type controls (adjusted p-value<0.05, at least twofold change, Fig. 3a , 150" Data S1). GO term analysis revealed significant up-regulation of genes involved in 151" metabolism, receptor signalling and cell adhesion (Data S2). To obtain a high confidence list 152" of significantly differentially regulated genes, we took genes threefold differentially regulated 153"

(80 and 244 genes down-and up-regulated in double mutant flies compared to controls) and 154"

analysed them according to gene function by annotated protein domains. This analysis 155" confirms prominent effects on gene networks involved in metabolism, cellular signaling and 156" structural cell components, including a number of cell adhesion molecules and is qualitatively 157" different from loss of m 6 A or by regulation of synapse numbers by the transcription factor erect 158" wing ( Fig. 3b, Data S2 ) [31] [32] [33] .

Notably, immune genes were not significantly up-regulated in the double mutant flies (Data 160" S1) and CMTr1 knock-out mice 34 . Since only a proportion of all mRNAs have cOMe in both 161"

Drosophila and mice ( Fig. 1e and f) 9 , the primary role of cOMe is not self/non-self 162" discrimination, at least in Drosophila. The relevance of cOMe to prevent detection of non-self 163"

RNA by the evolutionary younger vertebrate immune system is linked to the interferon 164" response, which is absent in flies, and they also do not possess unmethylated cap RNA sensors 165"

Rig-I and IFITs 15 .

A potential role of cOMe could be to stabilize mRNA transcripts. However, we find a 3.5 fold 167" increase in up-regulated transcripts compared to down-regulated transcripts in the absence of 168"

cOMe, which does not support a general role of cOMe in protecting mRNAs from degradation 169"

in Drosophila. To further test, whether cOMe protects mRNAs from degradation, we generated 170" fully capped RNA oligonucleotides with or without methylation using the vaccinia capping 171" enzymes and noted that vaccinia CMTr can 2`-O-methylate the ribose of the first three 172" nucleotides ( fig. S4 ). When we incubated these RNA oligonucleotides that were uncapped, 173" capped and capped with cOMe in nuclear and cytoplasmic Drosophila S2 cell extracts, cOMe 174" did not affect RNA stability, while the lack of a cap resulted in increased degradation, which 175"

is consistent with observations in mammalian systems 35 (Fig. 3c) . 176"

CMTr2 has a dedicated set of target genes 178"

We next investigated how many genes produce mRNAs that contain cOMe. Since the levels of 179" cOMe are low ( Fig. 1e and f) , we reasoned that cOMe is either co-transcriptionally added to 180" mRNAs of only a few specific genes or, of only a fraction of all mRNAs. To distinguish 181" between these two possibilities, we stained polytene chromosomes from larval salivary glands.

CMTr1 prominently co-localized with RNA Pol II ( Fig. 4a-e) , suggesting that cOMe is 183" introduced co-transcriptionally and is wide-spread. In contrast, CMTr2 only prominently 184" localized to a subset of transcribed genes suggesting that CMTr2 has a preferred set of target 185" genes ( Fig. 4f-j) .

We subsequently used CLIP (crosslinking and immunoprecipitation) to identify targets for 187" CMTr1 and CMTr2. For these experiments we used a CMTr double knock-out line which 188"

contained genomic rescue constructs for CMTr1 and CMTr2 that are tagged with an HA or 189" FLAG epitope, respectively. From these experiments we obtained 36 and 701 protein coding 190" genes for CMTr1 and CMTr2, respectively, that were twofold or more enriched above input 191" (Data S3). Finding so few enriched genes for CMTr1 when CMTr1 co-stained with RNA Pol 192"

II on polytene chromosomes indicates that CMTr1 globally associates with most genes. In 193"

contrast, the larger number of CMTr2 enriched genes suggests that it introduces cOMe to a 194" more specific set of target transcripts.

To obtain a high confidence catalogue of CMTr2 CLIP targets, we took genes that were at least 196" threefold enriched (117 genes) and analysed them according to gene function. Consistent with 197" previous analysis of differentially expressed genes ( Fig. 3a and b , and Data S1 and S2), this 198" analysis revealed prominent effects on gene networks involved in cellular signalling including 199" a number of genes encoding ion channels or their regulators and synaptic vesicle release in 200" addition to many cell adhesion molecules (Fig. 4k, Data S3 ). Only few CMTr CLIP targets 201" are differentially expressed in CMTr mutants further supporting that cOMe does not affect 202" mRNA stability (Data S3). 203"

Since cOMe can enhance translation in trypanosomes 36 , we tested whether cOMe is required 206"

for local translation at synapses by puromycin incorporation. Indeed, in the absence of CMTrs 207" protein synthesis is significantly reduced at synapses of third instar neuromuscular junctions 208" (Fig. 5a) . 209"

Although known for over 40 years, the role of cOMe in animals has been enigmatic due to the 212" lack of knockout models 15 . Here we show that loss of cOMe has little obvious phenotypic 213"

consequences leading to development of healthy and fertile flies. In accordance with prominent 214" expression of mRNA methyltransferases in the brain, however, we find that cOMe is essential 215" for reward learning 3,4 . 216"

Short-term reward memory measured immediately after training is considered to be insensitive 219"

to blockers of protein synthesis 29,37 . It therefore seems somewhat enigmatic that cOMe would 220" play an acute role in the reward learning process. Moreover, cOMe occurs in the nucleus before 221" the mRNAs undergo a lengthy journey to the synapse. Our experiments demonstrate a role for 222" cOMe in adult KCs but the two days required to induce CMTr2 expression does not have the 223" required temporal resolution to distinguish between roles before and during learning itself. We 224" therefore currently favor a model for cOMe in establishing/maintaining the appropriate 225"

repertoire of locally-translated synaptic factors in adult KC synapses, that are necessary to 226" support reward learning, rather than directly in learning-induced synaptic change. Consistent 227"

with prior reports of neuronal localization of mRNAs encoding cytoskeletal proteins, 228" neurotrophins, membrane receptors and regulatory kinases important for synaptic activity and 229" plasticity 38 , we find that CMTr targets include many cell adhesion and signalling molecules.

To mention in particular as CMTr2 target is the volado-encoded α-integrin that was shown to 231" be defective in short-term memory performance 39 . Work in several organisms has also 232" demonstrated roles for neuronal cell adhesion molecules (NCAMs) in acute forms of plasticity 233"

and includes Drosophila mutants in the N-CAM homolog fasII 40, 41 . Although both of these 234"

Drosophila studies revealed defects in short-term aversive memory, other locally-translated 235" adhesion molecules could also be specifically required to support short-term and more 236" persistent reward memory.

It is well known that many mRNAs are transported and stored in various cellular locations 238"

including dendrites and synapses 38,42 . In dendrites, translation of mRNAs occurs in 239" polysomes, while in synapses the main form of translation is from monosomes 43 . Our 240" discovery of a function for mRNA cOMe in learning and local translation of transcripts at 241" synapses ( Fig. 5b) has important implications in understanding the role of these modifications 242"

in affecting gene expression in synaptic plasticity. 243"

Materials and Methods 245" 246"

The deletion allele y w CMTr1 13A (excision 13A) and y w; CMTr2 M32 (excision M32) were 248" obtained from imprecise excision of transposon P{EPgy2}CG6379 [EY08403] over Df(X)BSC869 249"

and P{EP}aft [G6146] were sequenced for validation. y w CMTr1 13A and y w; CMTr2 M32 excision lines were viable 255" when first generated. To normalize genetic backgrounds, excision lines were outcrossed to the 256"

Df lines for five generations. A control y w line was generated by crossing Df(X)BSC869 to 257"

P{EP}aft [G6146] and Df(2R)BSC347 to P{EPgy2}CG6379 [EY08403] for five generations and then 258"

combined. To determine survival of mutants, freshly hatched larvae were individually picked 259"

and grown in groups of 30 and surviving adults counted. 260"

To clone CMTr1 and CMTr2 cDNAs, total RNA was extracted with Tri-reagent (SIGMA) 263" from larval brains and reverse transcribed with Superscript II as described 44 and KpnI into the pUC 3GLA vector 45 containing an attB site for phiC31 mediated integration.

The GFP+-marked pUC 3GLA UAS CMTr2:FLAG construct was inserted into attP40 at 25C 285" by phiC31 transgenesis.

Genomic rescue constructs were made by recombineering from BAC clones. For gCMTr1, the 287" ends were amplified with Q5 polymerase (NEB) using primers dMtr end1F1 288"

(GGCACTAGTgcgcatgaattaagtgctaaaatgtg) and dMtr end1R1 289"

(ATCCCGGCTTATGTGTGTCCAACATG), and dMtr end2F2 290"

(ATCCCAAACCGAACCACATTAAAGG) and dMtr end2R2 291"

(CCGTGGTACCGGTGTTATGCTCGGACAGTGGTAATCGAATG) from BAC DNA 292" prepared as described 46 and cloned into pUC 3GLA using SpeI and KpnI. The 10.5 kb genomic 293" fragment was then retrieved using the ends vector linearized with EcoRV from BacR21I10 as 294" described 45 . The C-terminal HA TEV myc tag was then incorporated by PCR into a 495 bp 295"

AvrII and SbfI fragment and cloned with these sites. The GFP+-marked pUC 3GLA 296" gCMTr1:HATEVmyc construct was inserted into attP VK0022 at 57F by phiC31 transgenesis.

For CMTr2, the ends were amplified with Q5 polymerase (NEB) using primers aft end1 GCTTCAAAGATGTCATTTAATCCTCCAG) into a modified pUAST using BamHI and 308"

SpeI in a four-way ligation. The 6.7 kb genomic fragment was then retrieved using the ends 309"

vector linearized with StuI from BacR20E20 as described 45 . The w+-marked CASPR gCMTr2: 310"

TEVFLAG construct was inserted into attP40 at 25C by phiC31 transgenesis.

Essential parts of all DNA constructs were sequence verified. 312"

For negative geotaxis experiments, groups of 20 flies kept in two inverted fly vials (19 cm) 315"

were tapped to the bottom. A movie was then made to record the moving flies and a frame 316" about 5 sec after the flies started running upwards and before the first fly reached the top was 317" taken to measure the distance the flies have run upwards.

For learning and memory experiments, two to five day old flies of both sexes were used for 319" behavioral experiments in a T-maze. Odors used were 4-methylcyclohexanol (MCH) and 3-320" octanol (OCT).

For appetitive learning and 24 hour memory testing, flies were starved for 21-23 h prior to 322"

training and training was done as described 29 . Briefly, a group of about 120 flies were exposed 323"

first to the unconditioned odor (CS−) for two minutes followed by 30 seconds of air, and then 324"

to the conditioned odor (CS+) in the presence of dry sucrose for two minutes. For appetitive 325" learning or immediate memory, flies were tested immediately after training for their choice 326" between the two odors. For 24 hour memory, flies were transferred into a standard cornmeal 327" food vial after training and after one hour, they were transferred into food-deprivation vials 328" until testing on the next day.

Odor and sugar acuity tests were performed as described in 47 with some modifications. For 330" odor acuity tests, starved flies were directly placed into the T-maze to test for odor avoidance 331"

(OCT or MCH) against the smell of plain mineral oil. For sugar acuity test, a filter paper with 332" size 18x8cm was placed into a glass milk bottle (250ml). Half of the filter paper (~9x8cm) was 333" soaked with saturated sucrose and dried before use. For the test, starved flies were placed into 334" the bottle and the number of flies on both parts of the filter paper were counted separately two 335" minutes later. The performance index was calculated as [Nsugar/Ntotal] × 100, where Ntotal = Nsugar 336" + Nplain.

For conditional expression GSG GAL4 was used 48 , that is activated by feeding flies with the 338" progestin, mifepristone (RU486). Accordingly, flies were kept on RU486 (200 µM (SIGMA), 339" 5% ethanol) or control (5% ethanol) standard fly food for two days at 18°C before starvation 340" and training. 341"

Behavioral data was analyzed using GraphPad Prism 6. Two-tailed t tests were used for 344" comparing two groups, and one-way ANOVA followed by a Tukey`s post-hoc test was used 345"

for comparing multiple groups. 346"

Total RNA was extracted with Trizol (Invitrogen) and PolyA mRNA from two rounds of oligo 349"

dT selection was prepared according to the manufacturer (Promega). Alternatively, polyA 350" mRNA from one round of oligo dT selection was followed by ribosomal RNA depletion using 351" biotinylated oligos as described 49 . For each sample, 50 ng of mRNA was decapped using either 352" tobacco acid pyrophosphate (250 U; Epicenter) or RppH (NEB) in buffer provided by the 353" supplier and then dephosphorylated by Antarctic phosphatase (NEB). The 5`-end of 354" dephosphorylated mRNAs were then labeled using 10 units of T4 PNK (NEB) and 0.5 µl [#-355" 32 P] ATP (6000 Ci/mmol, 25 µM; Perkin-Elmer). The labeled RNA was precipitated, and 356" resuspended in 10 µl of 50 mM sodium acetate buffer (pH 5.5) and digested with P1 nuclease 357"

(SIGMA) for 1 h at 37º C. Two microliters of each sample was loaded on cellulose F TLC 358" plates (20x20 cm; Merck) and run in a solvent system of isobutyric acid:0.5 M NH4OH (5:3, 359"

v/v), as first dimension, and isopropanol: HCl:water (70:15:15, v/v/v) , as the second dimension.

TLCs were repeated from biological replicates. The identity of the nucleotide spots was 361" determined as described 9,50 . For the quantification of spot intensities on TLCs, a storage 362"

phosphor screen (K-Screen; Kodak) and Molecular Imager FX in combination with 363"

QuantityOne software (BioRad) were used.

For the analysis of CAGEseq data, nucleotides in the N1 position of mRNA following the m 7 G 365" artifact were counted in a loop using grep in bash on all fastq files available from SRP131270 366" nuclei were the resuspended in 50% of the volume in buffer C (20 mM HEPES, pH 7.6, 420 385" mM NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 0.5 mM DTT, 1 mM PMSF (stock: 0.2M in 386" isopropanol), 1 µg/ml leupeptin, 25 % v/v glycerol (Ultrapure, Gibco) using a pipette, a stirrer 387" added, the volume slowly increased by another 50% of the nuclei volume with buffer C and 388" then the nuclei were extracted for 30 min. The extract was then spun 30 min at 10 000 g at 4º 389" C and the supernatant taken off without the the white slur on top. This extract was then dialyzed 390"

in buffer E (20 mM HEPES, pH 7.6, 100 mM KCl, 0.2 mM EDTA, 0.5 mM DTT, 1 mM PMSF 391"

(stock: 0.2M in isopropanol), 1 µg/ml leupeptin, 20 % v/v glycerol (Ultrapure, Gibco) for 2 392"

hours. After dialysis, the supernatant was spun at 10 000 g for 10 min and aliquots frozen in 393" liquid nitrogen and extracts stored at -80º C. 394"

Generation of a cap labelled probe, UV-crosslinking, RNA stability assay, 396" immunoprecipitation and denaturing gel electrophoresis 397"

As probe for UV-crosslinking, RNA stability and binding experiments the trypanosome splice 398" leader oligo (trypSL, AACUAACGCUAUUAUUAGAAC) 53 was used. 6 pmole trypSL (1.25 399"

µl from a 50µM stock was kinased with 2 µl 32 PgammaATP (25 µM, 6000Ci/mmol, 150 400" mCi/ml, Perkin Elmer) with 10 U PNK in 10 µl with 20 U RNasin (Roche). After 1 h, the probe 401" was extracted by phenol/CHCl3 and precipitated. The second phosphate was then added with 402"

Myokinase (Sigma M3003, Myokinase was dialyzed into 100 mM NaCl, 50 mM TrisHCl pH 403" 7.5, 1 mM MgCl2, 1 mM DTT), 100 U in 20 µl, in a total volume of 40 µl to 2.4 pmole trypSL 404"

in the presence of 1 mM ATP and 20 U RNasin (Roche) in vaccinia capping buffer. After 2 h, 405" the RNA was extracted by phenol/CHCl3 and precipitated. Capping was then done in 20 µl 406" with vaccinia capping enzymes (NEB) according to the manufactures instructions and after 90 407" min 2 µl terminator nuclease buffer A and 0.7 U Terminator nuclease (Epicenter) were added.

After 30 min, the RNA was extracted by phenol/CHCl3 and precipitated. The RNA was then 409" analysed on 20% polyacrylamide gels, dried and exposed to a phosphoimager screen.

RNAse I digestion to analyse 2`-O-ribose methylation was done in the presence of 10 U T4 411" PNK (NEB) in 50 mM Tris-acetate (pH 6.5), 50 mM NaCl, 10 mM MgCl2 and 2 mM DTT to 412" remove 2`,3`-cyclic phosphate intermediates 54 .

UV-crosslinking was done as described 52 . Briefly, 32 P labeled capped trypanosome splice 414" leader oligo with or without cOMe was incubated in a total volume of 10 µL, in 40% (v/v) 415" nuclear or cytoplasmic extract, 1 mM ATP, 5 mM creatine phosphate, 2 mM MgAcetate, 20 416" mM KGlutamate, 1 mM, DTT, 20 U RNasin (Roche), and 5 µg/mL tRNA at room temperature 417"

for 25 min and UV cross-linked on ice at 254nm for 20 min in a Stratalinker (Stratagene), 418" followed by digestion with RNase A/T1 mix (Ambion) at room temperature for 15 min.

Samples were then taken up in SDS-protein gel buffer and run on 8% gels, the gels dried and 420"

exposed to a phosphoimager screen.

For RNA stability experiments, 32 P labeled uncapped and capped trypanosome splice leader 422" oligo with or without cOMe was incubated in a total volume of 10 µl, in 40% (v/v) nuclear or 423" cytoplasmic extract, 1 mM ATP, 5 mM creatine phosphate, 2 mM MgAcetate, 20 mM 424"

KGlutamate, 1 mM, DTT, 20 U RNasin (Roche), and 5 µg/mL tRNA on ice for 45 min. Input 425"

was take before the addition of nuclear extract. The RNA was extracted by phenol/CHCl3 and 426"

precipitated. Samples were then separated on 8% polyacrylamide gels, dried and exposed to a 427" phosphoimager screen.

For immunoprecipitations of CBP80, 32 P labeled capped Trypanosome splice leader oligo with 429" or without cOMe was incubated in a final volume of 120 µl in IP-Buffer (150 mM NaCl, 50 430" mM Tris HCL, pH 7.5, 1% NP-40, 5% glycerol) together with nuclear extract (40%, v/v), rabbit 431"

anti-CBP80 (4 µl, gift from D. Kopytova), 20 µl protein A/G beads (SantaCruz) in the presence 432"

of Complete Protein Inhibitor (Roche) and 40 U RNase inhibitors (Roche) for 2h at 4º C. After 433"

washing the beads, RNA was extracted by phenol/CHCl3 and precipitated. Samples were then 434" separated on 8% polyacrylamide gels, dried and exposed to a phosphoimager screen. 435"

In situ antibody stainings were done as described previously 31 using rat anti-HA (MAb 3F10, (1:250; Molecular Probes or Invitrogen, A11034). DAPI (4',6-diamidino-2-phenylindole) was 442" used at 1 µg/ml. For imaging, tissues were mounted in Vectashield (Vector Labs) for confocal 443" microscopy using a Leica TCS SP5/SP2. Images were processed using Fiji.

To analyse synapses at neuromuscular junctions (NMJ) third instar wandering larvae were 445" dissected in PBS and fixed with Bouin's solution (Sigma-Aldrich, HT10132) for 5 minutes 446"

using. The samples were washed three times in PBT (PBS with 0.1% Triton TM X-100 (Sigma, 447" T8787) and 0.2% BSA) for 15 minutes. Primary antibody were rat anti-HA (MAb 3F10 1:20, 448" Roche) or rabbit anti-FLAG (M2, 1:250, SIGMA), Mouse anti-NC82 (1:50, DSHB), rabbit 449"

anti-CBP80 (1:100, Gift from D. Kopytova) 55 and DAPI (4=,6=-diamidino-2-phenylindole,1 450" µg/ml) was carried out overnight at 4º C followed by secondary antibodies (conjugated 451" with Alexa Fluor 488 or Alexa Fluor 647 (1:250; Molecular Probes, Invitrogen) at RT for 4-5 452"

hours. NMJs were mounted in Vectashield (Vector Labs), scanned with Lecia TCS SP8 and 453" processed using FIJI. For quantification of synapse stainings the mean intensity of the 454"

boutons was calculated using the Nikon NIS-Elements Basic Research (BR) imagining 455" software, and the data was analysed using GraphPad Prism. 456"

Polytene chromosome preparations and stainings 458"

CMTr1 and CMTr2 were expressed in salivary glands with elav C155 -GAL4 from a UAS 459" transgenes tagged with HA or FLAG, respectively, as described 33 . Briefly, larvae were grown 460" at 18º C under non-crowded conditions. Salivary glands were dissected in PBS containing 4% 461"

formaldehyde and 1% TritonX100, and fixed for 5 min, and then for another 2 min in 50% 462" acetic acid containing 4% formaldehyde, before placing them in lactoacetic acid (lactic 463" acid:water:acetic acid, 1:2:3). Chromosomes were then spread under a siliconized cover slip 464" and the cover slip removed after freezing. Chromosome were blocked in PBT containing 0.2%

BSA and 5% goat serum and sequentially incubated with primary antibodies (mouse anti-PolII 466" H5 IgM, 1:1000, Abcam, and rat anti-HA MAb 3F10, 1:50, Roche, or rabbit anti-FLAG, 467" 1:1000, SIGMA) followed by incubation with Alexa488-and/or Alexa647-coupled secondary 468" antibodies (Molecular Probes) including DAPI (1 µg/ml, Sigma). 469"

Illumina sequencing and analysis of differential gene expression 471"

For sequencing, QuantSeq 3` FWD libraries were generated from y w control and y w CMTr 13A ;

CMTr M32 flies. The QuantSeq 3` FWD kit was used according to the manufacturer's 473"

instructions with the following modifications: RNA was not denatured, and 6 U of Heparinase 474" I (NEB) was added to the first strand cDNA synthesis mix. Pooled indexed libraries were 475" sequenced on an Illumina NextSeq 500 to yield between 10 and 30 million single-end 50bp 476" reads per sample.

After demultiplexing with Illumina bcl2fastq v1.8.4, sequence reads were aligned to the 478"

Drosophila genome (dmel r6.02) using STAR 2.6. Reads for each gene were counted using 479"

HTSeq-count and differential gene expression determined with DESeq2 and the Benjamini-480"

Hochberg for multiple testing to raw P-values with p<0.05 considered significant. 481"

For CLIP, RNA was prepared essentially as described from 14-18h old embryos of y w 484" CMTr 13A ; gCMTr1:TEVHA CMTr M32 gCMTr2:TEVFLAG 56 . Embryos were first 485" dechorionated in 50% bleach, washed and then fixed in heptane containing 5% formaldehyde 486"

(10 ml heptane, 1.75 ml 37% formaldehyde, and 1.3 ml PBS equilibrated for 30 min) for 10 487" min with vigorous shaking. Embryo extracts were then prepared in RIPA buffer (150 mM 488"

NaCl, 50 mM Tris-HCL, pH 7.5, 1% NP-40, 0.5% Na-deoxycholate, 0.05% SDS) in a 1-ml 489"

Dounce homogenizer. After 20-40 strokes with the tight pestil, 1 vol of immuno-precipitation 490"

(IP) buffer was added (150 mM NaCl, 50 mM Tris-HCL, pH 7.5, 0.05% NP-40). The extract 491" was then cleared by centrifugation for 15 sec. IPs were done with monoclonal anti-HA 492"

antibodies coupled beads (Sigma) or anti-FLAG antibodies and protein A/G beads (SantaCruz) 493"

in IP buffer containing 7 mM CaCl2, 40 U of RNase inhibitor (Roche), 2 U of TurboDNase 494"

(Ambion), and 15% of extract for 2 hr at room temperature. After washing and TEV Proteinase 495"

(Promega) digestion for 1 h on ice, the supernatant was taken off, Proteinase K digested (0.5 496" mg/ml in 150 mM NaCl, 100 mM Tris-HCl, pH 7.5, 10 mM EDTA, 0.25% SDS) for 30 min 497" at 37º C, and RNA was isolated by phenol/chloroform extraction and ethanol precipitation in 498" the presence of glycogen.

The RNA was then reverse-transcribed with Superscript II (Invitrogen) according to the 500" manufacturer's instructions using a random nonamer tagged partial p7 sequence 501"

(CACGACGCTCTTCCGATCTNNNNNNNNN) and the first strand synthesis product was 502" purified using AMPure XP beads (Beckman) following the manufacturer's instructions with 503"

1.8 volumes. To generate double stranded cDNA and sequencing-ready libraries, Lexogen's 504" quant-seq 3'FWD kit was used, proceeding from the RNA removal and second strand synthesis 505"

steps. The input library was generated the same way from RNA before IP. Size selection of 506" libraries were carried out with PAGE prior to sequencing with the NextSeq 500. Differential 507" gene expression analysis performed as above then provided a simple route to detecting enriched 508" transcripts following immuno-precipitation. 509" 510"

GO enrichment analysis was performed with Pantherdb. Gene expression data were obtained 512" from flybase. Visualization of RNA-seq data were carried out with the R packages 513"

EnhancedVolcano Version 1.4.0 and ggplot2 in the R studio environment 57,58 .

Hypergeometric p-values for the significance of overlapping genes between CMTr2 and FMRF 515" CLIP targets were calculated using the 197 successes in a sample size of 701 CMTR2 clip 516"

targets, compared to the 2432 successes of FMRP targets in cholinergic and GABAergic 517" neurons 59 in the whole population of 17421 coding genes that can be returned following 518" alignment (p=1.34 -23 ). 519" 520"

All data are available in the main text or the supplementary material and gene expression data 522"

have been deposited at GEO under the accession numbers GSE116212 and GSE138868. 523"

References 525" 

Methylated nucleotides block 5' terminus of HeLa cell messenger 539

Viral and cellular mRNA capping: past and prospects

A novel synthesis and detection method for cap-associated adenosine modifications in 543" mouse mRNA

Reversible methylation of m6Am in the 5' cap controls mRNA stability

Cap-specific terminal N (6)-methylation of RNA by an RNA polymerase II-associated 547" methyltransferase

PCIF1 Catalyzes m6Am mRNA Methylation to Regulate Gene Expression

Identification of the m(6)Am Methyltransferase PCIF1 Reveals the Location and 551" Functions of m(6)Am in the Transcriptome

Landscape and Regulation of m(6)A and m(6)Am Methylome across Human and Mouse 553

mRNA cap regulation in mammalian cell function and fate

RNA 2'-O-Methylation (Nm) Modification in Human Diseases

The Mammalian Cap-Specific m(6)Am RNA Methyltransferase PCIF1 Regulates 559" Transcript Levels in Mouse Tissues

Post-transcriptional modifications of mRNA. Purification and 561" characterization of cap I and cap II RNA (nucleoside-2'-)-methyltransferases from HeLa cells

Characterization of hMTr1, a human Cap1 564" 2'-O-ribose methyltransferase

2'-O-ribose methylation of cap2 in human: function and evolution in a horizontally 566" mobile family

Cap and cap-binding proteins in the control 568" of gene expression

Biophysical studies of eIF4E cap-binding protein: recognition of mRNA 5' cap 570" structure and synthetic fragments of eIF4G and 4E-BP1 proteins

CBP80-promoted mRNP rearrangements during the 572" pioneer round of translation, nonsense-mediated mRNA decay, and thereafter. Cold Spring Harb Symp 573

Transcription start site analysis reveals widespread divergent transcription in D. 575" melanogaster and core promoter-encoded enhancer activities

Using FlyAtlas to identify better Drosophila melanogaster 577" models of human disease

Diversity and dynamics of the Drosophila transcriptome

adrift, a novel bnl-induced Drosophila gene, required for tracheal pathfinding into the 580

Nucleotide modifications in messenger RNA and 582" their role in development and disease

Rapid consolidation to a radish and protein synthesis-dependent long-term 584" memory after single-session appetitive olfactory conditioning in Drosophila

Olfactory learning skews mushroom body output pathways to steer behavioral 587" choice in Drosophila

Erect wing regulates synaptic growth in Drosophila by 589" integration of multiple signaling pathways

Differential activity of EWG transcription factor isoforms identifies a 591" subset of differentially regulated genes important for synaptic growth regulation

m6A potentiates Sxl alternative pre-mRNA splicing for robust Drosophila sex 594" determination

CMTR1-Catalyzed 2'-O-Ribose Methylation Controls 596" Neuronal Development by Regulating Camk2alpha Expression Independent of RIG-I Signaling

Both the 7-methyl and the 2'-O-methyl groups in the cap of 599" mRNA strongly influence its ability to act as primer for influenza virus RNA transcription

Hypermethylated cap 4 maximizes 602" Trypanosoma brucei translation

Parametric and genetic analysis of Drosophila 604" appetitive long-term memory and sugar motivation

mRNA transport & local translation in neurons

Integrin-mediated short-term 608" memory in Drosophila

Drosophila fasciclinII is required for the formation of odor memories and for normal 610" sensitivity to alcohol

The role of cell adhesion molecules in synaptic plasticity and memory

Local translation in neurons: visualization and function

Monosomes actively translate synaptic mRNAs in neuronal processes

Differential and inefficient 618" splicing of a broadly expressed Drosophila erect wing transcript results in tissue-specific enrichment of 619" the vital EWG protein isoform

Plasmid-based gap-repair recombineered transgenes reveal a central role for 621" introns in mutually exclusive alternative splicing in Down Syndrome Cell Adhesion Molecule exon 4. 622

Concentration and localization of co-expressed 624" ELAV/Hu proteins control specificity of mRNA processing

Dopamine and octopamine differentiate between aversive and appetitive olfactory 626" memories in Drosophila

P[Switch], a system for spatial and temporal control of 628" gene expression in Drosophila melanogaster

Nascent-seq indicates widespread cotranscriptional pre-mRNA splicing in 630" Drosophila

Mobilities of modified ribonucleotides on two-dimensional cellulose thin-layer 632" chromatography

Accurate transcription initiation by RNA polymerase II 634" in a soluble extract from isolated mammalian nuclei

ELAV inhibits 3'-end processing to promote neural splicing of ewg pre-mRNA. 636

Mass spectrometry of 638" mRNA cap 4 from trypanosomatids reveals two novel nucleosides

Mechanism of RNA 2',3'-cyclic phosphate end healing by T4 polynucleotide 640" kinase-phosphatase

The DUBm subunit Sgf11 is required for mRNA export and interacts with Cbp80 in 642

ELAV-mediated 3'-end processing of ewg transcripts is 644" evolutionarily conserved despite sequence degeneration of the ELAV-binding site

Elegant Graphics for Data Analysis Using the Grammar of Graphics

Lasko and 655" the Developmental Studies Hybridoma Bank for antibodies, the University of Cambridge 656" Department of Genetics Fly Facility for injections, BacPAc for DNA clones, E. Zaharieva for 657" help with stainings, D. Balacco and F. Stappers for help with art work, and B. Muller and R. 658" Michell for comments on the manuscript. MS is funded by the BBSRC (BB/R002932/1) and 659" the Leverhulme Trust, RGF from BBSRC (BB/R001715/1), DH from WISB, a 660

Genomic 681" rescue fragments tagged either with hemaglutinin (HA, a) or FLAG (b) epitopes are indicated 682" at the bottom. (c) Validation of CMTr1 13A and CMTr M32 single and double mutants by genomic 683" PCR. (d) Survival of flies to adulthood after hatching from the eggshell (n=3-4). (e) Climbing 684" activity assessed by negative geotaxis assays, n=40, p≤0.001. (f) Schematic diagram of a 2D 685" thin layer chromatography (TLC) depicting standard and 2`-O-ribose methylated nucleotides. 686" (g-k) TLCs showing modifications of the first cap-adjacent nucleotides of S2 cells (e), adult 687" control (f) and CMTr1 13A and CMTr2 M32 single (g, h) and double (i) mutant females. (l) 688" Quantification of the mRNA first nucleotide from TLC (n=5) and CAGEseq data (n=8) from 689" adult Drosophila and S2 cells, respectively. 690" 691" Figure 2. mRNA cap 2`-O-ribose methylation is required for reward learning in 692" Drosophila. 693" (a and b) Appetitive memory immediately (a) and 24 hour (b) after training of control and 694" CMTr1 13A and CMTr M32 single and double mutant flies shown as mean±SE. n=8 for A and n=6 695" for B, p≤0.006. (c) Rescue of the learning defect in CMTr1 13A ; CMTr2 M32 double mutant flies 696" by genomic fragments shown as mean±SE. n=6, p=0.002. (d and e) Rescue of the learning 697" defect in CMTr1 13A ; CMTr2 M32 double mutant flies by constitutive (d) or conditional (e) 698" expression of CMTr2 in mushroom bodies from UAS shown as mean±SE. n=6, p≤0.0001. 699" 700" Figure 3. Impact of mRNA cap-adjacent 2`-O-ribose methylation on gene expression and 701" RNA stability. 702" (a) Volcano plot depicting differentially expressed genes in CMTr1 13A ; CMTr2 M32 double 703" mutant flies compared to control flies. (b) Functional classification of up-(bottom) and down-704" regulated (top) genes in CMTr1 13A ; CMTr2 M32 double mutant flies compared to control flies. 705" (c) Incubation of monophosphorylated RNA (pRNA) and capped RNA with or without 2`-O-706" ribose methylation of cap-adjacent nucleotides in nuclear (nu) and cytoplasmic (cy) extracts 707" from S2 cells. The graph to the right depicts the percent undegraded RNA left after 45 min as 708" mean±SE from three repeats. 709" 710" Figure 4. CMTr2 localizes to distinct sites of transcription and has a dedicated set of 711" targets. 712" (a-l) Polytene chromosomes from salivary glands expressing CMTr1::HA (a-e) or 713" CMTr2::FLAG (f-j) stained with anti-Pol II (magenta, d), anti-HA (green, c) and DNA (DAPI, 714" blue, b), or merged (white

The authors declare no competing interests. 675" 676"