Catalytic activity of the Bin3/MePCE methyltransferase domain is dispensable for 7SK snRNP function in Drosophila melanogaster

Fly stocks and husbandry

Fly stocks from the Bloomington Drosophila Stock Center (BDSC) used in this study are listed in Supplementary Table 1. Fly stocks we constructed for this study are listed in Supplementary Table 2. The genotypes of maternal and paternal flies, and the genotypes of their progeny used for experiments, and the figures in which those experiments appear, are listed in Supplementary Table 3. Fly were reared at 25°C and 60–80% humidity on a 12 hour light/dark cycle, on either Nutri-Fly MF food (Genesee Scientific) with ∼0.064 M propionic acid (Sigma Aldrich, P5561; or Apex Bioresearch Products, 20–271), or food made based on the Nutri-Fly MF formula: 25 g/l Inactive Dry Yeast (Genesee Scientific), 89.5 g/l Dry Molasses (Genesee Scientific), 57 g/l Fly Stuff Yellow Cornmeal (Genesee Scientific), 5.84 g/l Nutri-Fly Drosophila Agar, Gelidium (Genesee Scientific), and ∼0.064 M propionic acid (Apex Bioresearch Products).

Bin3 mutants

We previously described 2 excision alleles that remove some (bin3^4–7) or all (bin3^2–7) of the bin3 coding exons (Parks et al. 2004; Singh et al. 2011). Both alleles were made in the same background, and when placed in trans to produce bin3^4–7/bin3^2–7 mutants, could incur background effects, likely due to homozygosity for recessive second-site mutations. To produce bin3 mutants that avoid the possibility of these confounding effects, we crossed flies heterozygous for bin3^2–7 with flies heterozygous for a deficiency (hereon referred to as Df) uncovering bin3 (Fig. 1a), which is in a different genetic background than bin3^2–7 (Ryder et al. 2004).

Generation of transgenic bin3⁺ rescue flies

We previously showed that Bin3 is required for embryonic viability by demonstrating that a bin3 cDNA under control of the maternal Hsp83 promoter rescued the embryonic lethality of eggs laid by bin3 mutant females (Singh et al. 2011). To examine the role of Bin3 more broadly, we needed a rescue approach that would faithfully express bin3 under control of its own regulatory sequences in all bin3-expressing tissues, not just in the ovary. To do this, we used BAC clone CH322-154G8 (BACPAC Genomics) that comprises a 22.065 kb fragment of chromosome 2R, containing the bin3 P1 promoter (Singh et al. 2011), and 3 of the 4 predicted mRNA isoforms of bin3 (bin3-RA, bin3-RC, and bin3-RD). This BAC was inserted into the attP landing site on chromosome 3R in BDSC stock #9744 (y¹ w¹¹¹⁸; PBac{y⁺-attP-9A}VK00027) by BestGene, Inc.

Generation of the pUAS-DSCP-XbaI-attB vector via NEBuilder HiFi DNA assembly

High-copy plasmids pNP (Qiao et al. 2018; Wang et al. 2019) and pUASP-attB (Drosophila Genomics Resource Center Stock 1358; https://dgrc.bio.indiana.edu//stock/1358; RRID:DGRC_1358; Takeo et al. 2012) were purified from bacterial cells grown to confluence at 37°C in LB + 100 µg/ml ampicillin using the Monarch Plasmid Miniprep Kit T1010L from New England BioLabs, Inc. (NEB). To create pUAS-DSCP-XbaI-attB (the vector into which bin3 and 3xHA sequences would be cloned, see below), the 10XUAS-DSCP promoter was amplified from plasmid pNP with Q5 High-Fidelity DNA Polymerase (NEB) using primers UP0394 and UP0395, and inserted into pUASP-attB that had been digested with XhoI (NEB) and PstI (NEB), via NEBuilder HiFi DNA Assembly (NEB). NEBuilder HiFi DNA Assembly destroyed the XhoI site, and partially eliminated 1 of the 2 GAGA enhancers in pUASp-attB.

Ovarian RNA purification and reverse transcription for NEBuilder HiFi DNA assembly

One to 3-day-old w¹¹¹⁸ (BDSC #5905) females were incubated with males in yeasted vials overnight. Five pairs of ovaries were dissected from 2 to 4-day-old females in ice-cold 1X PBS, and flash-frozen in liquid nitrogen. Total RNA was extracted using the Monarch Total RNA Miniprep Kit (NEB), according to the manufacturer's instructions for RNA purification from “Tissues”. Total RNA (1 µg) was reverse transcribed using ProtoScript II First Strand cDNA Synthesis Kit (NEB, E6560S) with primer UP0118, which anneals to the very end of bin3 coding sequence (CDS). See Supplementary Table 4 for primer sequence.

Purification of BAC clone CH322-154G8 for NEBuilder HiFi DNA assembly

EPI300 bacterial cells (Lucigen) harboring the BAC clone CH322-154G8 (BACPAC Genomics) were grown to confluence at 37°C in LB + 12.5 µg/ml chloramphenicol, and plasmid replication was induced by diluting 1:10 in LB + 12.5 µl/ml chloramphenicol and 1X CopyControl Induction Solution (Lucigen). The BAC was purified using the Monarch Plasmid Miniprep Kit (NEB).

Generation of UAS-DSCP-3xHA-bin3 transgenic flies via NEBuilder HiFi DNA assembly

To generate plasmid pUAS-DSCP-3xHA-bin3-attB, which expresses Bin3 C-terminally tagged with 3xHA, under UAS control: Fragment #1, comprising the bin3-RA 5′-UTR to the end of the bin3 CDS, was amplified from bin3 cDNA using Q5 High-Fidelity DNA Polymerase, and primers UP0423 and UP0414. The 5′ end of Fragment #1 is homologous to sequences immediately upstream of the XbaI site in pUAS-DSCP-XbaI-attB; the 3′ end of Fragment #1 is homologous to 5′ end of a synthetic dsDNA fragment comprising Gly-Gly-Gly-Ser-3xHA (G₄S-3xHA; Integrated DNA Technologies) codon-optimized for Drosophila with the IDT Codon Optimization Tool (Integrated DNA Technologies). Fragment #2, comprising the entire bin3-RA 3′-UTR, was amplified from the BAC clone CH322-154G8 using Q5 High-Fidelity DNA Polymerase, and primers UP0415 and UP0424. The 5′ end of Fragment #2 is homologous to the 3′ end of the G₄S-3xHA fragment; the 3′ end of Fragment #2 is homologous to sequences immediately downstream of the XbaI site in pUAS-DSCP-XbaI-attB. Fragment #1, G₄S-3xHA, and Fragment #2 were assembled in this order into the XbaI site of pUAS-DSCP-XbaI-attB via NEBuilder HiFi DNA Assembly. The XbaI site was destroyed.

To generate plasmid pUAS-DSCP-3xHA-bin3^Y795A-attB, which expresses catalytically dead Bin3^Y795A, C-terminally tagged with 3xHA, under UAS control: Fragment #3, comprising the bin3-RA 5′-UTR to 11 nt downstream of the Y795 codon, was amplified from bin3 cDNA using Q5 High-Fidelity DNA Polymerase, and primer UP0423 and the mutagenic primer UP0419, which introduces a Y795A point mutation. The 5′ end of Fragment #3 is homologous to sequences immediately upstream of the XbaI site in pUAS-DSCP-XbaI-attB; the 3′ end of Fragment #3 is homologous to the 5′ end of Fragment #4. Fragment #4, comprising 11 nt upstream of the Y795 codon to the end of the bin3-RA CDS was amplified from cDNA using Q5 High-Fidelity DNA Polymerase, and mutagenic primer UP0420, which introduces a Y795A point mutation, and primer UP0414. The 5′ end of Fragment #4 is homologous to the 3′ end of Fragment #3; the 3′ end of Fragment #4 is homologous to 5′ end of the G₄S-3xHA fragment. Fragment #3, Fragment #4, G₄S-3xHA, and Fragment #2 (see above) were assembled in this order into the XbaI site of pUAS-DSCP-XbaI-attB via NEBuilder HiFi DNA Assembly. The XbaI site was destroyed.

To generate plasmid pUAS-DSCP-3xHA-bin3^ΔMSM-attB, which expresses Bin3^ΔMSM, C-terminally tagged with 3xHA, under UAS control: Fragment #5, comprising the bin3-RA 5′-UTR to 12 nt downstream of the L326 codon, was amplified from bin3 cDNA using Q5 High-Fidelity DNA Polymerase, and primer UP0423 and the mutagenic primer UP0417, which eliminates codons 311–326. The 5′ end of Fragment #5 is homologous to sequences immediately upstream of the XbaI site in pUAS-DSCP-XbaI-attB; the 3′ end of Fragment #5 is homologous to the 5′ end of Fragment #6. Fragment #6, comprising 12 nt upstream of the F311 codon to the end of the bin3-RA CDS was amplified from cDNA using Q5 High-Fidelity DNA Polymerase and mutagenic primer UP0418, which eliminates codons 311–326, and primer UP0414. The 5′ end of Fragment #6 is homologous to the 3′ end of Fragment #5; the 3′ end of Fragment #6 is homologous to 5′ end of the G₄S-3xHA fragment. Fragment #5, Fragment #6, G₄S-3xHA, and Fragment #2 (see above) were assembled in this order into the XbaI site of pUAS-DSCP-XbaI-attB via NEBuilder HiFi DNA Assembly. The XbaI site was destroyed.

See Supplementary Table 4 for primer sequences and the sequence of the G₄S-3xHA synthetic dsDNA fragment used to make these 3 constructs. Sequencing of each construct revealed that all UTRs corresponded to the bin3-RA mRNA isoform of bin3. Interestingly, each construct had an in-frame insertion of an Ala residue between L573 and V574, when mapped to the reference sequence of bin3 (NCBI Reference Sequence NM_165468.2). These 3 constructs were recombined into the attP landing site on chromosome 3L in BDSC stock #8622 (y¹ w^67c23; P{y^+t7.7=CaryP}attP2) by BestGene, Inc.

Expression of UAS-DSCP-3xHA-bin3 transgenes using the Trojan GAL4 approach

We used the UAS/GAL4 system (Brand and Perrimon 1993) and the Trojan GAL4 approach (Diao and White 2012; Diao et al. 2015; Lee et al. 2018) to induce the expression of UAS-DSCP-3xHA-bin3 (hereon referred to as UAS-bin3) transgenes. Intronic Trojan GAL4 insertions produce a null allele of a gene while also expressing GAL4 under the regulatory sequences of that gene. A Trojan GAL4 insertion into 1 of the bin3 introns (bin3^{MI08045-TG4.0}, hereon referred to as bin3^TG4.0) in trans to a deficiency (Df) uncovering bin3 produces bin3-null (bin3^TG4.0/Df) flies that express GAL4 in the spatiotemporal pattern of endogenous bin3 (Fig. 4d). When transgenes comprising UAS-bin3 cDNAs (Fig. 4g–i) are incorporated into bin3^TG4.0/Df flies, GAL4 induces expression of Bin3 proteins in the pattern of endogenous bin3 in an otherwise bin3-null background.

Total RNA purification from ovaries and qRT-PCR

Five 1 to 3-day-old females were incubated with males in yeasted vials overnight. Five pairs of ovaries were dissected from 2 to 4-day-old females in ice-cold 1X PBS, and flash-frozen in liquid nitrogen. Total RNA was extracted using the Monarch Total RNA Miniprep Kit (NEB), according to the manufacturer's instructions for RNA purification from “Tissues”. qRT-PCR was performed on 10 ng total RNA using the Luna Universal One-Step RT-qPCR Kit (NEB, E3005L). bin3, 7SK, U6, and rp49 mRNAs were reverse transcribed and amplified with primers UP0493 + UP0118, UP0146 + UP0147, OW1123 + OW1124, and UP0113 + UP0114, respectively (see Supplementary Table 4 for primer sequences). Reactions were performed using a CFX Opus 384 Real-Time PCR System (Bio-Rad).

Protein extraction and western analysis

Ten 1 to 3-day-old females were incubated with males in yeasted vials overnight. Ten pairs of ovaries were dissected from 2 to 4-day-old females in ice-cold 1X PBS, and flash-frozen in liquid nitrogen. Protein was exacted as previously described (Prudêncio and Guilgur 2015). Extracts were clarified through 0.22 µM Spin-X columns (Costar), and mixed 1:1 with 2X Sample Loading Buffer (Bio-Rad) + 5% β-mercaptoethanol, and boiled for 5 minutes. Extracts were spun down, and 15 µl was loaded onto 4–20% TGX gels (Bio-Rad), and ran at 100 V for 10 minutes, then 300 V for 20 minutes. Protein was transferred to PVDF-LF membranes (Bio-Rad) using a Trans-Blot Turbo Transfer System (Bio-Rad) using the High MW preset (1.3 A, 25 V, 10 minutes). Membranes were blocked in EveryBlot Blocking Buffer (Bio-Rad) for 5 minutes, and then incubated with primary antibodies diluted in EveryBlot Blocking Buffer for 1 hour at room temperature. Membranes were washed with TBST (1X Tris-buffered saline, 0.2% Tween-20) for 5 minutes 3 times, and then incubated with secondary antibodies diluted in EveryBlot Blocking Buffer for 1 hour at room temperature. Membranes were with TBST for 5 minutes 3 times, dehydrated with 100% methanol, and imaged using a ChemiDoc MP Imaging System with Image Lab Touch software (Bio-Rad). Primary antibodies used were 1 µg/ml rabbit anti-HA (abcam, ab9110), and 1 µg/ml mouse anti-beta Actin (abcam, ab170325), and 5 µg/ml IgG purified from sheep anti-CycT antiserum (Nguyen et al. 2012); using Pierce Protein G Magnetic Beads according to the manufacturer's instructions. Secondary antibodies were from Jackson ImmunoResearch, Inc., and used at 1 µg/ml: Alexa Fluor 488 AffiniPure Donkey Anti-Rabbit IgG (711-545-152), Alexa Fluor 488 AffiniPure Donkey Anti-Sheep IgG (713-545-147), Alexa Fluor 594 AffiniPure Donkey Anti-Mouse IgG (715-585-150), and Alexa Fluor 647 AffiniPure Donkey Anti-Mouse IgG (715-605-150).

Egg-laying assays

Ten 1 to 3-day-old virgin females were incubated with wild-type males en masse in yeasted vials overnight. Ten 2 to 4-day-old females were split individually into fresh vials with 2 Oregon-R males, and incubated for 24 hours. After 24 hours, eggs were counted to determine the number of eggs laid/female/day. Experiments were performed in biological triplicate.

Immunofluorescence staining of ovaries

Immunofluorescence staining of ovaries was performed as described (Maimon and Gilboa 2011). Ten 1 to 4-day-old females were incubated with males in yeasted vials overnight, and then 10 pairs of ovaries were dissected from 2 to 5-day-old females in ice-cold 1X PBS. The following steps were performed at room temperature (RT) unless otherwise indicated. Ovaries were fixed in 1 ml of Image-iT Fixative Solution (Thermo Scientific, R37814) + 0.3% Triton X-100 for 20 minutes. Ovaries were washed and permeabilized with 1 ml of 1% PT Buffer (1X PBS, 1% Triton X-100) for 5, 10, and 45 minutes. Ovaries were blocked with 1 ml PBT (1X PBS, 1% BSA, 0.3% Triton X-100) for 1 hour, and then incubated in 250 µl PBT and primary antibodies overnight at 4°C. Primary antibodies used were 1 µg/ml rabbit anti-HA (abcam ab9110) and 5 µg/ml mouse anti-Vasa (Developmental Studies Hybridoma Bank, Vasa 46F11). Ovaries were washed with 1 ml 0.3% PT Buffer (1X PBS, 0.3% Triton X-100) 3 times for 30 minutes each, blocked with 1 ml PBT-NDS (1X PBS, 1% BSA, 0.3% Triton X-100, 5% normal donkey serum) for 1 hour, and incubated in 250 µl PBT-NDS containing 10 µg/ml Hoechst 33342 (Thermo Scientific, H3570) and 2 µg/ml secondary antibodies for 2 hours in the dark. Secondary antibodies were from Jackson ImmunoResearch, Inc.: Alexa Fluor 488 AffiniPure Donkey Anti-Rabbit IgG (711-545-152) and Alexa Fluor 647 AffiniPure Donkey Anti-Mouse IgM (715-605-020). Ovaries were washed with 1 ml 0.3% PT Buffer in the dark 3 times for 30 minutes each, then mounted in 50 µl VECTASHIELD PLUS Antifade Mounting Medium (Vector Labs, Inc., H-1900-10).

Single-molecule fluorescence in situ hybridization (smFISH) in ovaries

Single-molecule fluorescence in situ hybridization (smFISH) on adult ovaries was performed essentially as described (Raj and Tyagi 2010; Abbaszadeh and Gavis 2016). Ten 1 to 4-day-old females were incubated with males in yeasted vials overnight, and then 10 pairs of ovaries were dissected from 2 to 5-day-old females in ice-cold 1X PBS. The following steps were performed at RT unless otherwise indicated. Ovaries were fixed in 1 ml of Image-iT Fixative Solution (Thermo Scientific, R37814) for 30 minutes, then washed with PBST (1X PBS, 0.1% Tween-20) 3 times for 5 minutes each. Ovaries were permeabilized by incubating in 7:3 PBST:methanol for 5 minutes, 1:1 PBST:methanol for 5 minutes, 3:7 PBST:methanol for 5 minutes, 100% methanol for 10 minutes, 3:7 PBST:methanol for 5 minutes, 1:1 PBST:methanol for 5 minutes, and 7:3 PBST:methanol for 5 minutes. Ovaries were then washed with PBST 4 times for 5 minutes each, followed by pre-hybridized in Wash Buffer A (20% Stellaris RNA FISH Wash Buffer A (Biosearch Technologies Cat# SMF-WA1-60) and 10% deionized formamide) for 5 minutes. Ovaries were incubated in Stellaris RNA FISH Hybridization Buffer (Biosearch Technologies Cat# SMF-HB1-10) with 10% deionized formamide and 125 µM Quasar 640-labeled oligonucleotide probes, overnight at 37°C in the dark. Oligonucleotide probes against 7SK were designed using the LGC Biosearch Technologies’ Stellaris RNA FISH Probe Designer (https://www.biosearchtech.com/support/tools/design-software/stellaris-probe-designer). Ovaries were washed in the dark with Wash Buffer A for 30 minutes at 37°C, and then with Wash Buffer A with 10 µg/ml Hoechst 33342 (Thermo Scientific, H3570) for 30 minutes at 37°C, and finally with Stellaris RNA FISH Wash Buffer B (Biosearch Technologies Cat# SMF-WB1-20) for 5 minutes at RT. Ovaries were mounted in 50 µl VECTASHIELD PLUS Antifade Mounting Medium (Vector Labs, Inc., H-1900-10).

Ovariole counting assays

Batches of 1 to 3-day-old females were incubated with sibling males in yeast vials overnight. Batches of 2 to 4-day-old females were anesthetized, transferred to 1.7 ml tubes, flash-frozen in liquid nitrogen, and stored at −80°C. Frozen females were thawed on ice or at RT and dissected in a 1:1 mixture of methanol and 1X PBS. 50% methanol dehydrates and extracts lipids from the gonadal sheath surrounding the ovary and from the ovarioles, turning the clear portions of the ovarioles opaque, and making ovarioles easy to separate and count. The mean number of ovarioles per ovary as well as the absolute value of the difference in ovariole number between each ovary per female (i.e. ovariole asymmetry) were calculated (Lobell et al. 2017). Experiments were performed in biological duplicate.

Climbing assays

Ten 2 to 5-day-old females were transferred to 15 ml tubes, and allowed to acclimate for 3 minutes. Flies were tapped to the bottom of the tube, and climbing was recorded for 12 seconds. Flies were then allowed to reacclimatize for another 3 minutes before performing the subsequent trial. Ten trials were performed per genotype to determine the percent of flies climbing 8 cm in 12 seconds per trial. All genotypes for a particular experiment were tested at the same time. Experiments were performed in biological triplicate.

Modeling of Bin3 protein sequence onto MePCE

Bin3 was modeled onto 289 amino acids of the MePCE active site (MePCE_MT) in complex with S-adenosyl-ʟ-homocysteine (SAH) and stem loop 1 of 7SK (PDB 6dcb; Yang et al. 2019) using hhpred (Söding 2005; Hildebrand et al. 2009; Meier and Söding 2015; Zimmermann et al. 2018; Gabler et al. 2020) and MODELLER (Šali et al. 1995; Zimmermann et al. 2018; Gabler et al. 2020). Alignment of the Bin3 model to 6DCB, RMSD calculation, and modeling of the Y421A and Y795A mutations was performed using PyMol (version 2.5.4).

Bin3 protein expression and purification

Residues P778-D1114 of Bin3 correspond to a region of the protein comprising the methyltransferase domain that aligns to the resolved portions of MePCE methyltransferase domain (PDB 6dcb; Yang et al. 2019). This 337-amino acid region of Bin3 was codon-optimized for expression in Escherichia coli (E. coli), and a wild-type version (Bin3^wt) or a Y795A mutant version (Bin3^Y795A) was cloned into the NdeI and BlpI sites of pET-28a (in frame with an N-terminal 6xHis tag) by GenScript. These plasmids were transformed into E. coli BL21-Gold(DE3) cells (Agilent Technologies) for protein expression. Bacterial cultures were grown in minimal media (1X M9 salts, 0.5% glucose, 8 mg/l thiamin, 2 mM MgSO₄, 0.1 mM FeCl₃, 1X trace elements) at 37°C to an OD₆₀₀ of 0.6, and the temperature was then lowered to 18°C for 1 hour before induction by the addition of 0.5 mM IPTG for 20 hours. Cultures were centrifuged to collect the cell pellet, which was resuspended in lysis buffer (20 mM HEPES, pH 8.0, 15 mM Imidazole, 1 M NaCl, 1 mM TCEP, 5% glycerol, 1 mM PMSF) supplemented with lysozyme and protease inhibitor cocktail tablet (Thermo Scientific A32965). The cell resuspensions were sonicated in an ice-water bath, centrifuged, and filtered to remove cell debris. 6xHis-Bin3 proteins were purified with a Ni-Sepharose affinity column (HisTrap HP; Cytiva 17524801), followed by size exclusion chromatography (SEC; HiLoad 26/600 Superdex 75; Cytiva 28989334) in SEC buffer (20 mM HEPES, pH 7.5, 250 mM KCl, 1 mM TCEP). Expression and purity of both proteins were confirmed by SDS-PAGE (Supplementary Fig. 2). The concentration of protein peak fractions were measured, and the protein was stabilized by addition of a 2-fold molar excess of cofactor S-(5′-adenosyl)-ʟ-methionine ρ-toluenesulfonate salt (SAM; Sigma A2408, ≥80% purity), or S-methyl-¹³C SAM (99 atom % ¹³C; Sigma Aldrich/ISOTEC 798231) for NMR studies, or a mixture of 4.6:1 SAM to S-(5′-adenosyl)-ʟ-[methyl-³H]methionine ([me³H]SAM) (stock solution of 55.9 µM (17.9 Ci/mmol) in hydrochloric acid; PerkinElmer Life Sciences) for methyltransferase assays. We hereafter refer to this mixture as “tritiated SAM”.

Design and in vitro transcription of Drosophila melanogaster (Dm) 7SK RNA

We designed a Dm7SK RNA substrate comprising a minimal stem 1 (S1) followed by a 6-nt single-stranded overhang, based on the predicted fold for Dm7SK RNA (Supplementary Fig. 3a). Of note, this minimal Drosophila 7SK substrate is similar to the minimal human 7SK substrate that we have used previously in methyltransferase assays with MePCE (Yang et al. 2019). S1 RNA sample was prepared by in vitro transcription as described previously (Eichhorn et al. 2016). S1 was transcribed from a chemically synthesized DNA template (Integrated DNA Technologies) using the T7 RNA Polymerase P266L mutant (Guillerez et al. 2005). The transcription mixture contained 40 mM Tris, pH 8.0, 1 mM spermidine, 0.01% Triton X-100, 2.5 mM DTT, 25 mM MgCl₂, 4 mM each of rATP, rUTP, rGTP, and rCTP, 0.5 µM DNA template, ∼1 µM T7 RNAP P266L, and 10% DMSO, and was incubated at 37°C for 5 hours. Transcribed RNA was purified using 12% denaturing polyacrylamide gel electrophoresis (Urea-PAGE) and extracted from the gel pieces using a crush and soak method in 1X TBE Buffer (90 mM Tris borate, 2 mM EDTA, pH 8.3). The extracted RNA was filtered and further purified with a weak anion exchange column (HiTrap DEAE FF; Cytiva 17515401) and eluted with high salt buffer (10 mM sodium phosphate, pH 7.6, 1 mM EDTA, 1.5 M KCl). Eluted RNA was buffer exchanged using a 3 kDa Amicon device into sterilized nanopure water, diluted to <100 µM, heated at 95°C for 5 minutes, and snap cooled on ice for 1 hour. Folded RNA was then concentrated with a 3 kDa Amicon device and stored at −20°C. The intended hairpin fold of S1 RNA was confirmed by NMR using ¹H-¹H nuclear Overhauser effect spectroscopy (NOESY) (Supplementary Fig. 3b).

Methyltransferase assays

In vitro RNA methyltransferase assays were performed as previously described (Yang et al. 2019). Briefly, reaction mixtures comprised reaction buffer (50 mM HEPES, pH 7.5, 150 mM KCl, and 1 mM TCEP) and contained 1 µM Bin3^WT or Bin3^Y795A and 20 µM tritiated SAM. A total of 25 µl of each reaction was preincubated in thin-walled PCR tubes at 37°C for 5, and 5 µl of 60 µM S1 RNA substrate (10 µM final) was added to initiate the reactions. Reactions were quenched by heating to 98°C for 2 minutes, and stored at −20°C. A total of 25 µl of the 30 µl reaction mixtures were spotted onto Amersham Hybond-N+ membrane (Cytiva RPN303B), washed, and counted as previously described (Yang et al. 2019).

NMR analysis of S1 RNA

Uncapped S1 RNA was buffer exchanged into NMR buffer (20 mM sodium phosphate, pH 6.06, 50 mM KCl, 5% D₂O) and concentrated to 0.25 mM. Capped S1 RNA for ¹³C NMR analysis was generated by incubating freshly purified Bin3^WT, S-methyl-¹³C SAM (see above), and RNA samples at a molar ratio of 1:2:2 in SEC buffer at 37°C overnight. RNA was purified from the reaction mixture by phenol-chloroform extraction, followed by ethanol precipitation, and then air dried and dissolved in NMR buffer. All NMR spectra were acquired at 17°C with a Bruker Avance III HD 600 MHz spectrometer equipped with QCI HCNP cryoprobe. NMR data were collected using TopSpin (Bruker), processed with NMRPipe (Delaglio et al. 1995) and analyzed using NMRFAM-Sparky (Lee et al. 2015).

Electrophoretic mobility shift assay (EMSA)

Stock solutions of S1 RNA, Bin3^WT, and Bin3^Y795A were prepared at 20 µM, 6 µM, and 6.8 µM, respectively. For electrophoretic mobility shift assay (EMSA), increasing amounts of protein were added to 20 pmol RNA with 1 µl of 30% glycerol loading dye. Mixed samples were loaded on a 5.7% non-denaturing polyacrylamide gel (37.5:1 crosslinking ratio) and run at 75 V for 48 minutes at room temperature in 0.5× TBE running buffer. Gel was stained with toluidine blue and destained in water prior to imaging.

Bin3 disorder prediction

Disorder in Bin3 was predicted by IUPred3, using the IUPred3 short disorder analysis type (Mészáros et al. 2018; Erdős and Dosztányi 2020; Erdős et al. 2021).

RNA immunoprecipitation

Twenty-five 1 to 3-day-old females with or without transgenes expressing bin3-3xHA cDNAs were incubated with males in yeasted vials overnight. Twenty-five pairs of ovaries were dissected from 2 to 4-day-old females in ice-cold, 1X PBS, and fixed in 1 ml 1.8% formaldehyde (a mixture of 55% 1X PBS and 45% Image-iT Fixative Solution, Thermo Scientific, R37814) for 15 minutes at RT on a nutating rotor. Fixation was terminated by addition of 163 µl 2.5 M glycine to a final concentration of ∼350 mM, and incubation for 5 minutes at RT on a nutating rotor. Ovaries were washed briefly with 1X PBS 3 times, and then flash-frozen in liquid nitrogen. Ovaries were homogenized by hand with a blue pestle on ice in 50 µl ice-cold RIPA Lysis and Extraction Buffer (Thermo Scientific, 89900) supplemented with 1 mM DTT, 3X Halt Protease and Phosphatase Inhibitor Cocktail (Thermo Scientific, 78440), and 1 U/µl SUPERase•In RNase Inhibitor (Invitrogen, AM2696). A total of 200 µl of additional buffer was added to the homogenate, and samples were incubated on ice for 15 minutes to allow for complete nuclear lysis. Homogenate was pelleted by centrifugation at top speed for 1 minute. A total of 2.5 µl of homogenate (1/100th homogenate volume) was removed and kept on ice in a total of 300 µl 1X Monarch DNA/RNA Protection Reagent (NEB, T2011L) as the input sample. The remaining homogenate was used for immunoprecipitation. A total of 25 µl (1/10th homogenate volume) of Pierce Anti-HA Magnetic Beads (Thermo Scientific, 88836) were washed with 175 µl 1X TBST (Pierce 20X TBS Tween 20 Buffer, Thermo Scientific, 28360), and then with 1 ml 1X TBST. Homogenate was added to the washed beads, and immunoprecipitation was performed for 4 hours at 4°C with end-over-end rotation. Beads were washed briefly with 1 ml 1X TBST 3 times. Bin3-3xHA complexes were purified by 2 sequential competitive elutions with 75 µl of Pierce HA Peptide (Thermo Scientific, 26184) diluted to 2 mg/ml in Nuclease-Free Water (NEB, B1500L), at 37°C for 15 minutes each. The eluates were pooled, and the 150 µl total eluate was added to 150 µl 2X Monarch DNA/RNA Protection Reagent. All samples (inputs and eluates) were incubated at 65–70°C for 30 minutes to reverse formaldehyde crosslinks. Total RNA was extracted using the Monarch Total RNA Miniprep Kit (NEB), according to the manufacturer's instructions for RNA purification from “Tissues”. Total RNA was eluted in 100 µl Nuclease-Free Water. A total of 2.9 µl of each sample was used in 10 µl reactions to reverse transcribe and amplify 7SK and U6 snRNAs with primers UP0146 + UP0147 and OW1123 + OW1124, respectively (see Supplementary Table 4 for primer sequences). Reactions were performed using the Luna Universal One-Step RT-qPCR Kit (NEB, E3005L) and a CFX Opus 384 Real-Time PCR System (Bio-Rad).

Bin3 is required for female fecundity

Reproductive success is influenced by fertility (offspring viability) and fecundity (the rate of offspring production). Previously, we showed that Bin3 plays a role in fertility, promoting embryonic viability by a mechanism of caudal translation repression (Singh et al. 2011). During that and other studies, we noticed that bin3 mutants lay fewer eggs than wild-type flies. Therefore, we asked if Bin3 also plays a role in fecundity by promoting a normal egg-laying rate. To do this, we compared the egg-laying rates of flies that were either wild type for bin3 (wt), heterozygous for bin3 (bin3^2–7/+), or deleted entirely for bin3 (bin3^2–7 in trans to a deficiency that removes the bin3 locus; bin3^2–7/Df). The genomic regions deleted by bin3^2–7 and Df are depicted in Fig. 1a. As expected, bin3 mRNA levels are reduced by half in bin3^2–7/+ heterozygotes and are undetectable in bin3^2–7/Df mutants (Fig. 1b). Interestingly, we found that bin3^2–7/Df females laid eggs at a significantly reduced rate compared to wt (P = 0.0001) and bin3^2–7/+ (P < 0.0001) females (Fig. 2a). The egg-laying rate defect was specific to the loss of bin3, as a single copy of a genomic fragment containing the bin3 ORF (bin3⁺, Fig. 1a) restored bin3 mRNA accumulation by half (Fig. 1b) and rescued the egg-laying defect of bin3^2–7/Df females (P < 0.0001, Fig. 2a). Rescue was comparable to that of bin3^2–7/+ controls. These results demonstrate that Bin3 is required for egg-laying.

Bin3 promotes female fecundity by repressing P-TEFb

We found that 7SK RNA levels were strongly reduced in bin3^2–7/Df females compared to wt (P < 0.0001, Fig. 1c), as we have previously shown (Singh et al. 2011). The canonical function of Bin3/MePCE is to stabilize 7SK RNA (Jeronimo et al. 2007; Singh et al. 2011). 7SK RNA in turn acts as a scaffold for proteins that form a snRNP that sequesters and represses P-TEFb (Krueger et al. 2008; Barboric et al. 2009; Xue et al. 2010; Muniz et al. 2013; Brogie and Price 2017; Yang et al. 2022). Therefore, destabilization of 7SK in bin3^2–7/Df females should lead to deregulation (activation) of P-TEFb, which could be responsible for the egg-laying defect. If this is true, then we expect that reducing the activity of P-TEFb in bin3^2–7/Df females would rescue the egg-laying defect. To genetically reduce the activity of P-TEFb, we introduced a mutant allele of CycT (CycT^jllB2) into bin3^2–7/Df flies, and confirmed that this mutation reduces the protein levels of the largest of the 2 CycT isoforms (Supplementary Fig. 1a). Indeed, we found that reducing the amount of CycT in bin3^2–7/Df females rescued the egg-laying defect (P < 0.0001, Fig. 2a), while having no effect on 7SK RNA levels (Fig. 1c). These results suggest that Bin3 stabilizes 7SK RNA to form the 7SK snRNP, which then sequesters and represses P-TEFb to promote a normal egg-laying rate. Interestingly, CycT^j11B2/+ heterozygous females exhibited a slightly elevated rate of egg-laying compared to wt females (P = 0.0367, Supplementary Fig. 1b), suggesting that P-TEFb activity intrinsically suppresses egg-laying.

Bin3 and 7SK are expressed ubiquitously in the ovary

To gain insight into the role of Bin3 in egg-laying, we determined where Bin3 (and its substrate 7SK) is expressed and localized in ovaries, which are composed of 16–20 individual ovarioles that can be thought of as assembly lines of germline development (Fig. 2b). To do so, we used the UAS/GAL4 (Brand and Perrimon 1993) and Trojan GAL4 (see Materials and Methods; Diao and White 2012; Diao et al. 2015; Lee et al. 2018), where a Trojan GAL4 insertion into the bin3 ORF was used to drive the expression of a 3xHA-epitope-tagged bin3 transgene in the spatiotemporal pattern of endogenous bin3 (bin3^TG4.0  > UAS-3xHA-bin3). Bin3-3xHA was detected by immunofluorescence staining of ovaries for the 3xHA tag. We found that Bin3 is expressed in and localized to the nuclei of all ovarian cells (Fig. 2, c and c′). Bin3 was found to be expressed in the germline cells of the germarium (located at the anterior of each ovariole), and in the 15 nurse cells that supply the oocyte with mRNAs and proteins. Bin3 was also found to be expressed in the somatic cells surrounding the germline cells and the oocyte, which communicate with germline cells to promote egg chamber and oocyte maturation. Control flies lacking the UAS-3xHA-bin3 transgene (bin3^TG4.0/+) yielded no detectable HA staining (Fig. 2, d and d′).

To identify where in the ovary 7SK RNA is expressed and localized, we performed smFISH (Raj and Tyagi 2010; Abbaszadeh and Gavis 2016) on ovaries from wt and bin3^2–7/Df females. We found that in wt ovarioles, 7SK follows an identical expression pattern to that of Bin3 (compare Fig. 2, e and e′ with Fig. 2, c and c′). As expected, 7SK staining was undetectable in bin3^2–7/Df ovarioles (Fig. 2, f and f′), which do not accumulate 7SK RNA (Fig. 1c). Taken together, Bin3 and 7SK are both expressed in the germline and somatic cells of the ovarioles, and localize to nuclei—commensurate with their role in transcription elongation regulation.

Bin3 is not required to establish the proper number of ovarioles

The total number of ovarioles per ovary directly correlates with the number of eggs insects can lay (Wayne et al. 1997; Wayne and Mackay 1998), and is established by proper proliferation and migration of the somatic terminal filament cells (TFCs; Sahut-Barnola et al. 1995, 1996). Given that we found bin3^2–7/Df mutant females had an egg-laying defect, and that Bin3 is expressed in the somatic cells of the ovarioles, we hypothesized that Bin3 promotes egg-laying, possibly by functioning in the TFCs to establish ovarioles. To test this hypothesis, we compared ovariole numbers in bin3^2–7/Df females with those of wt and bin3^2–7/+ controls. To do this, we dissected ovaries and determined (1) the average number of ovarioles per ovary, and (2) the difference in the number of ovarioles within a pair of ovaries (i.e. “ovariole asymmetry”; Lobell et al. 2017). Unexpectedly, we did not find a significant difference in either the mean number of ovarioles per ovary (Fig. 2g) or in ovariole asymmetry (Fig. 2h) in bin3^2–7/Df females compared to either wt or bin3^2–7/+ control females. Together, these results suggest that Bin3 is not required to establish ovarioles, and therefore promotes egg-laying by some other mechanism.

Expression of Bin3 in the germline is not sufficient to promote fecundity

Bin3 is also expressed in germline cells (Fig. 2, c and c′), therefore, we hypothesized that Bin3 might instead function in the germline to promote a normal egg-laying rate. To test this hypothesis, we used the nanos-GAL4::VP16 driver (Van Doren et al. 1998) to induce expression of the UAS-3xHA-bin3 transgene specifically in the germline cells of bin3^2–7/Df females, and asked whether this improved the egg-laying rate (Fig. 2g). Germline expression of bin3 did not rescue the egg-laying rate defect of bin3^2–7/Df; nanos-GAL4::VP16/UAS-bin3 females compared to bin3^2–7/Df; nanos-GAL4::VP16/+ mutant females (P = 0.9942, Fig. 2i) suggesting that Bin3 does not function in germline cells to promote a normal egg-laying rate.

Since Bin3 is expressed in the follicle cells (Fig. 2, c and c′), it remained possible that Bin3 might instead function in the follicle cells to promote a normal egg-laying rate. We attempted to rescue the egg-laying defect of bin3^2–7/Df females by expressing the UAS-3xHA-bin3 transgene specifically in all follicle cells (Tran and Berg 2003; Pai et al. 2006) using the GR1 GAL4 driver. Unfortunately, we only recovered rare bin3^2–7/Df; GR1/UAS-3xHA-bin3 escapers that were even sicker than bin3^2–7/Df; GR1/+ mutants. Therefore, we were unable to determine whether Bin3 functions in the follicle cells to promote fecundity.

Bin3 has a conserved role in neuromuscular function

A nonsense mutation in the MePCE gene of a patient with impaired neuromuscular ability was shown to reduce MePCE protein levels, causing P-TEFb hyper-activation and aberrant RNAPII transcription elongation (Schneeberger et al. 2019). If a conserved function of Bin3/MePCE is to repress P-TEFb to promote normal neuromuscular development, then it would be expected that bin3 mutant flies would also have neuromuscular defects, and that these defects should be rescued by reducing CycT levels (i.e. lowering P-TEFb activity).

Two ways in which neuromuscular function can be assessed in flies is by examining climbing ability (Manjila and Hasan 2018) and resting wing posture (Baehrecke 1997). Normally, flies will climb to the top of an enclosure after being knocked to the bottom, and will fold their wings one-over-the-other when resting. We found that a significant percentage of bin3^2–7/Df flies had a climbing defect compared to bin3^2–7/+ control flies (P = 0.0003, Fig. 3a). Additionally, whereas control bin3^2–7/+ flies properly folded their wings (Fig. 3b), bin3^2–7/Df mutant flies had a held-out wings phenotype (Fig. 3c). Both defects were specific to loss of bin3, as the bin3⁺ genomic fragment rescued the climbing (P = 0.0004, Fig. 3a) and resting wing posture defects of bin3^2–7/Df flies (compare Fig. 3d with Fig. 3c). Importantly, both climbing (P = 0.0008, Fig. 3a) and wing posture (compare Fig. 3e with Fig. 3c) defects of bin3^2–7/Df flies were also rescued by reducing CycT levels. Control CycT^jllB2/+ heterozygotes had no defects in climbing (Supplementary Fig. 1c) or resting wing posture (compare Supplementary Fig. 1e with Supplementary Fig. 1d). These results suggest that Bin3 and MePCE share a conserved role in neuromuscular function from Drosophila to humans, most likely by repressing P-TEFb and preventing aberrant transcription elongation. However, as we will demonstrate later, a conserved motif found in Bin3 and MePCE likely promotes specific neuromuscular functions through an additional pathway.

Neuromuscular disorders arise from developmental defects in motoneurons and/or the muscles that they control. We attempted to rescue the defects in climbing and wing posture (and defects in egg-laying, which is also in part a neuromuscular function; Deady and Sun 2015; Irizarry and Stathopoulos 2015; Liao and Nässel 2020; Garrett et al. 2023) of bin3 mutants by tissue-specific expression of bin3 in either the nervous system using elav-GAL4 (Hudson et al. 2008; Pfeiffer et al. 2008; Jenett et al. 2012; Massey et al. 2019; Nandakumar et al. 2020), or in the muscles (including ovarian muscles) using Mef2-GAL4 (Hudson et al. 2008; Kairamkonda and Nongthomba 2014). Unfortunately, we did not recover any bin3^2–7/Df; elav-GAL4/UAS-bin3 flies, and recovered only a single bin3^2–7/Df; Mef2-GAL4/UAS-bin3 fly. We speculate that genetic background effects are confounding our attempts at tissue-specific rescue in the current background. As an alternate approach, we attempted to identify the cells and tissues that require Bin3 for neuromuscular function by performing an RNAi screen using over 90 GAL4 drivers active in specific neuron and muscle cell types that drive the expression of 2 different short hairpin RNAs (shRNAs) targeting bin3 CDS (Ni et al. 2011). However, we did not identify any conditions under which bin3 RNAi knockdown recapitulated the neuromuscular phenotypes of bin3 mutant flies. The failure of our screen might be due to technical reasons such as weak GAL4 drivers and/or insufficient levels of RNAi knockdown, or to Bin3 being required in more than 1 tissue simultaneously for neuromuscular activity. Indeed, knockdown of bin3 induced by the Trojan-GAL4 bin3 allele (which also functions as a null allele, see Fig. 5) in the spatiotemporal expression pattern of bin3 recapitulated only the held-out wings phenotype of bin3 mutant flies (data not shown), suggesting that Bin3 might be required in multiple cell types simultaneously, at least for normal resting wing posture.

Bin3 catalytic activity is dispensable for 7SK stability and snRNP function in vivo

The importance of MePCE catalytic residues in 7SK binding and capping have been studied mostly in vitro (Xue et al. 2010; Shelton et al. 2018; Yang et al. 2019). To determine the importance of Bin3 catalytic activity in vivo, we sought to generate flies expressing a catalytically inactive Bin3 protein. We first analyzed the crystal structure of the methyltransferase domain of human MePCE (PDB 6dcb, Yang et al. 2019). As shown in Fig. 4a, Tyrosine 421 (Y421) is juxtaposed between the 5′ gamma phosphate of 7SK and S-adenosyl-ʟ-homocysteine (SAH; used in place of the S-adenosyl-ʟ-methionine (SAM) methyl donor for the purposes of crystallization). Y421 hydrogen bonds to the 7SK gamma phosphate. Mutation of this tyrosine to alanine (Y421A) completely abolished MePCE catalytic activity in vitro (Yang et al. 2019). We modeled Bin3 onto the crystal structure (Fig. 4a) and found that Y795 in Bin3 (red) aligned well with Y421 in MePCE (black). We created models (Fig. 4b) representing the Y421A mutation in MePCE (black) and the concomitant Y795A mutation in Bin3 (red), which revealed a complete loss of contact of this Tyr residue with 7SK and SAH (i.e. SAM).

Given the in vitro importance of the Y421 residue in MePCE for catalytic activity, and that Y795 in Bin3 aligns with Y421, we wanted to determine whether the Y795A mutation in Bin3 was catalytically dead like Y421A in MePCE. To do this, we performed in vitro methyltransferase assays on a minimal Drosophila 7SK substrate (S1) using purified methyltransferase domain of Bin3 without (Bin3^wt) or with (Bin3^Y795A) the Y795A mutation (see Materials and methods). We measured the activities of both enzymes at 3 different timepoints, and found that Bin3^WT completes a fast, single-turnover at 2 minutes, and continues to slowly methylate 7SK in multiple-turnovers (at 5 and 30 minutes, Fig. 4c). To verify that S1 RNA was capped by Bin3^WT, we obtained ¹³C HMQC NMR spectrum of S1 RNA that had been incubated with Bin3^WT and [methyl-¹³C]SAM. We observed 1 resonance corresponding to a single methyl group covalently attached to the 5′ end of the S1 RNA (Supplementary Fig. 3c). In contrast to Bin3^wt, Bin3^Y795A has nearly no methyltransferase activity at all 3 timepoints tested (Fig. 4c), demonstrating that Y795 is required for Bin3 catalytic activity, as Y421 is required for MePCE catalytic activity.

Using the Bin3^Y795A mutant, we next sought to determine the importance of Bin3 catalytic activity in vivo. We again used the Trojan GAL4 approach. Importantly, not only do Trojan GAL4 insertions express GAL4 in the pattern of the gene they are inserted into, but they simultaneously generate a null allele of the inserted gene. Indeed, bin3 Trojan GAL4 insertion flies (bin3^TG4.0/Df) (Fig. 5a) had nearly undetectable levels of both bin3 mRNA (P < 0.0001, Fig. 5b) and 7SK RNA (P < 0.0001 Fig. 5c) compared to wt, confirming that bin3^TG4.0 is a null allele of bin3. We generated flies expressing either Bin3^wt or Bin3^Y795A and found that GAL4 expressed from bin3^TG4.0 induced both UAS-bin3^wt and UAS-bin3^Y795A (Fig. 6a) transgenes similarly at the mRNA (P = 0.2188, Fig. 6b) and protein (P = 0.9836, Fig. 6c) levels. Thus, these flies could be used to study the importance of Bin3 catalytic activity in vivo, and we tested whether Bin3^Y795A was able to (1) bind and stabilize 7SK, and (2) rescue bin3 mutant phenotypes.

Bin3 catalytic activity is expected to be required for stabilization of 7SK in vivo via methyl capping; therefore, in flies expressing catalytically inactive Bin3^Y795A, 7SK should be destabilized. Remarkably, there was no significant difference in 7SK levels between Bin3^wt and Bin3^Y795A flies (bin3^TG4.0/Df; UAS-bin3^wt/+ vs bin3^TG4.0/Df; UAS-bin3^Y795A/+; P = 0.9996; Fig. 7a), suggesting that Bin3 catalytic activity is not required for the stability of 7SK in vivo. How then might Bin3 stabilize 7SK in the absence of catalytic activity? As suggested by prior studies (Xue et al. 2010), Bin3 might form a stable complex with 7SK independent of its catalytic function, and this may protect it from degradation. Consistent with this idea, we found that Bin3^wt and Bin3^Y795A bind to similar amounts of S1 RNA in vitro as analyzed by EMSA (Fig. 4d), and that Bin3^Y795A immunoprecipitated as much 7SK RNA as did Bin3^wt in vivo (P = 0.5350, Fig. 7b). Note that the concentration of SAM is higher than that of SAH in vivo (Kashio et al. 2016), and thus is likely present in the active sites of both Bin3^wt and Bin3^Y795A, leading to the formation of either Bin3^wt•SAH•me7SK complexes (after completion of the methyltransferase reaction) and Bin3^Y795A•SAM•7SK complexes (that are unable to complete the methyltransferase reaction). To best recapitulate in vivo conditions, SAM was added to the in vitro EMSA assays as well (see Materials and methods and Discussion). Overall, these results indicate that the Y795A mutation does not diminish the binding of Bin3 to 7SK in vitro or in vivo, and that binding alone is sufficient for 7SK stability in vivo.

Given that Bin3^Y795A could bind and stabilize 7SK, we expected that Bin3^Y795A would rescue the fecundity, climbing, and wing posture defects of bin3^2–7/Df mutant flies. First, we showed that bin3^TG4.0/Df flies recapitulate the fecundity (compare Fig. 8a to Fig. 2a), climbing (compare Fig. 8b with Fig. 3a), and wing posture (compare Fig. 8d with Fig. 3c) defects of bin3^2–7/Df flies. As expected, Bin3^Y795A expression rescued the fecundity (P = 0.0008, Fig. 8a), climbing (P < 0.0001, Fig. 8b), and wing posture (compare Fig. 8f–d) defects of bin3^TG4.0/Df flies, similar to Bin3^wt [P > 0.9999 for both fecundity (Fig. 8a) and climbing (Fig. 8b)]. Therefore, despite the importance of catalytic activity to 7SK capping in vitro, Bin3 catalytic activity (and thereby, the methylphosphate cap) does not seem to be required in vivo for 7SK stability, nor is it likely to affect 7SK snRNP assembly and function.

A metazoan-specific motif in Bin3 is required to promote a specific neuromuscular function

Much attention has been given to the methyltransferase domain of Bin3 orthologs; however, metazoans have acquired over evolutionary time additional protein sequences outside of the methyltransferase domain (Cosgrove et al. 2012). To determine whether there is conserved homology between Bin3 orthologs in these regions, we aligned several Bin3 orthologs that have been previously characterized, including Drosophila Bin3 (Zhu and Hanes 2000; Singh et al. 2011; Cosgrove et al. 2012), Homo sapiens MePCE (Xue et al. 2010; Muniz et al. 2013; Shelton et al. 2018; Schneeberger et al. 2019; Yang et al. 2019; Lee et al. 2020; Zhang et al. 2021), Danio rerio Mepcea (Barboric et al. 2009), and Schizosaccharomyces pombe Bmc1 (Páez-Moscoso et al. 2022; Porat et al. 2022). We discovered a 16-amino acid motif that is almost entirely conserved in metazoans (flies, zebrafish, and humans), but is absent in the fission yeast S. pombe (Fig. 6d). Therefore, we named this motif the metazoan-specific motif (MSM). The N-terminal portion of this motif is predicted to be more structured (IUPred3 score 0.2733 to 0.3764), while the C-terminal portion is predicted to be less structured (IUPred3 score 0.4062 to 0.5887). Given the strong conservation of the MSM, we hypothesized that the MSM must contribute to some important function of Bin3 and its metazoan orthologs. To test this hypothesis, we used the Trojan GAL4 system to generate flies expressing either Bin3^wt or Bin3 deleted of the entire 16-amino acid MSM (Bin3^ΔMSM) and asked whether Bin3^ΔMSM could (1) bind and stabilize 7SK, and (2) rescue bin3 mutant phenotypes.

We found that GAL4 expressed from bin3^TG4.0 induced both the UAS-bin3^wt and UAS-bin3^ΔMSM (Fig. 6a) transgenes to express similar levels of mRNA (P = 0.3840, Fig. 6b) and protein (P = 0.7994, Fig. 6c). Given that the MSM is outside of the Bin3 methyltransferase domain, we did not expect that the Bin3 MSM would be required for 7SK stability or binding. Accordingly, there was no significant difference in 7SK levels (P = 0.3846, Fig. 7a) between Bin3^ΔMSM flies (bin3^TG4.0/Df; UAS-bin3^ΔMSM/+) and Bin3^wt flies (bin3^TG4.0/Df; UAS-bin3^wt/+). Also, Bin3^ΔMSM immunoprecipitated as much 7SK (P = 0.9938, Fig. 7b) as did Bin3^wt.

Given that 7SK was bound and stabilized by Bin3^ΔMSM, we expected that Bin3^ΔMSM would rescue the fecundity, climbing, and wing posture defects of bin3^TG4.0/Df flies. Indeed, Bin3^ΔMSM expression rescued the fecundity (P = 0.0061, Fig. 8a) and climbing (P < 0.0001, Fig. 8b) defects of bin3^TG4.0/Df flies, similar to Bin3^wt [P = 0.6681 (Fig. 8a); P = 0.0685, (Fig. 8b)]. Unexpectedly, Bin3^ΔMSM did not rescue the held-out wings phenotype of bin3^TG4.0/Df flies (compare Fig. 8g with Fig. 8d). These results suggest that the MSM is not required for Bin3 function in the contexts of egg-laying and climbing, but that the Bin3 MSM is required for normal resting wing posture.

Since Bin3^ΔMSM retains the ability to bind and stabilize 7SK (Fig. 7), we hypothesize that the function of the Bin3 MSM in normal resting wing posture is independent of 7SK and therefore unlikely to involve the canonical function of Bin3 in repressing P-TEFb activity. This predicts that reducing CycT dosage in bin3^TG4.0/Df; UAS-bin3^ΔMSM/+ flies would not rescue the held-out wings phenotype. Unexpectedly, reducing CycT dosage in Bin3^ΔMSM flies (bin3^TG4.0/Df; UAS-bin3^ΔMSM/CycT^j11B2) did rescue the held-out wings defect (compare Fig. 8h with Fig. 8g). This suggests the possibility that the Bin3 MSM and P-TEFb work through parallel and/or redundant pathways to promote normal wing posture.

Bin3 binds to the U6 spliceosomal RNA, but is not required for its stability

MePCE in humans (Jeronimo et al. 2007; Muniz et al. 2013) and Bmc1 in S. pombe (Páez-Moscoso et al. 2022; Porat et al. 2022) bind to the U6 spliceosomal snRNA, but are not required for U6 stability. We asked whether Bin3 in Drosophila is also dispensable for U6 stability, and whether Bin3 also binds to U6.

Like MePCE and Bmc1, we found that Bin3 is not required for U6 stability, as U6 levels were not significantly affected in bin3^2–7/Df flies (P = 0.9314, Fig. 9a). U6 levels were reduced by approximately half in bin3^TG4.0/Df flies (P = 0.0471, Fig. 9b); however, these data were just below the threshold for significance (α = 0.05). Additionally, U6 levels were not different between flies expressing Bin3^Y795A (P = 0.8794) or Bin3^ΔMSM (P = 0.7529) and Bin3^wt (Fig. 9c). Finally, we found that Bin3^wt (P = 0.0113), Bin3^Y795A (P = 0.0033), and Bin3^ΔMSM (P = 0.0336) all bind to U6 snRNA, compared to un-tagged control (Fig. 9d). We also found that there is no significant difference in the amount of U6 bound by Bin3^Y795A (P = 0.7671) or Bin3^ΔMSM (P = 0.8486) compared to Bin3^wt (Fig. 9d). These results show for the first time that Drosophila Bin3 binds to U6, and that neither catalytic activity nor the MSM are required for binding. It is not surprising that Bin3^ΔMSM binds to U6, as Bmc1 lacks this region (Cosgrove et al. 2012), yet it binds to U6 (Páez-Moscoso et al. 2022; Porat et al. 2022). We conclude that Bmc1, Bin3, and MePCE have a conserved function in binding to the U6 snRNA but are not required for its stability.

Bin3 represses P-TEFb to promote normal fly physiology

The canonical function of MePCE is to repress gene expression by binding to and stabilizing 7SK snRNA, which acts as a scaffold for proteins that sequester P-TEFb (Peterlin et al. 2011; Cosgrove et al. 2012). This inhibits P-TEFb, which normally phosphorylates RNAPII to promote transcription elongation. Consequently, loss of MePCE function destabilizes 7SK, dissociates the repressive snRNP, releasing P-TEFb (Fujinaga et al. 2023). Hyper-activated P-TEFb induces aberrant gene expression by ectopic phosphorylation of RNAPII and other factors (He et al. 2008; Schneeberger et al. 2019). MePCE also has a noncanonical function where it activates gene expression by binding to the tail of histone H4 and recruiting free P-TEFb to chromatin. This noncanonical function was found be required for the induction of tumorigenic genes in a breast cancer cell line (Shelton et al. 2018). Here, we show that Bin3 contributes to normal fly physiology through a mechanism most consistent with its canonical function in repression of P-TEFb. We found that the egg-laying, climbing, and wing posture defects of bin3 mutants were rescued by genetically reducing the activity of P-TEFb with a mutation in CycT (which is needed for P-TEFb kinase activity). Thus, in fly development, Bin3 appears to repress P-TEFb (and gene expression) rather than to activate P-TEFb (Figs. 10a and 11a). However, with respect to normal resting wing posture, it appears that a motif in Bin3 (MSM) plays an additional role in regulating this process (see below).

Bin3 is required for reproductive success

Previously, we found that Bin3 plays a role in fertility by promoting embryonic viability (Singh et al. 2011). Here, we show that Bin3 is required for fecundity by supporting a normal egg-laying rate (Fig. 10a). Thus, Bin3 has a role in 2 different aspects of reproductive success: fertility (offspring viability) and fecundity (the rate of offspring production).

Although Bin3 and its substrate 7SK are expressed in all germline and somatic cells of the ovary, germline expression of Bin3 was dispensable for a normal egg-laying rate. A germline function for Bin3 might occur later in development, consistent with our previous finding that Bin3 appears to regulate the translation of grk in late-stage oocytes (Singh et al. 2011), which is important for embryonic patterning.

A previous report demonstrated that both wild-caught flies harboring indels in bin3, and transgenic flies harboring transposon insertions into bin3, have reduced numbers of ovarioles (Lobell et al. 2017). However, we found that our bin3^2–7/Df mutant females did not exhibit reduced ovariole numbers compared to controls. We suggest that the difference between our findings and those of Lobell et al. (2017) might be attributable to differences in genetic backgrounds. Nevertheless, the wild-type number of ovarioles established in our bin3 mutants is not sufficient to promote a normal egg-laying rate.

Egg-laying is in part a neuromuscular process, as muscle contractions of the ovary and uterus promote successful oviposition (i.e. egg-laying; Deady and Sun 2015; Irizarry and Stathopoulos 2015; Liao and Nässel 2020; Garrett et al. 2023). It is possible, therefore, that defects in neuromuscular function that we observe in bin3 mutants (see below) might also explain why bin3 mutant females lay fewer eggs, e.g. due to a reduction in the frequency or strength of these contractions. We attempted to express Bin3 specifically in the nervous system or in muscles (including ovarian muscles) of bin3 mutants to rescue bin3 mutant phenotypes (including the egg-laying defect). However, we only acquired rare, sick escapers that were not suitable for examination. Determining whether Bin3 functions in the nervous system or muscle to promote fecundity will be a focus for future experiments.

Conserved role of Bin3 and MePCE in metazoan neuromuscular development

A patient with a neurodevelopmental disorder characterized by mobility and musculature deficiencies was shown to carry a heterozygous nonsense mutation in MePCE (Schneeberger et al. 2019). Similarly, we found that bin3 mutant flies exhibited difficulties in climbing and an inability to properly fold their wings, phenotypes that are characteristic of neuromuscular defects (Shukla et al. 2013; Kairamkonda and Nongthomba 2014; Sanhueza et al. 2014; Manjila and Hasan 2018). Many mutations in Drosophila can affect wing posture (see below), including those in the highly conserved JAK/STAT pathway (Harrison et al. 1998; Callus and Mathey-Prevot 2002; Ayala-Camargo et al. 2013; Hatini et al. 2013; Johnstone et al. 2013). Therefore, the conserved roles of Bin3 and MePCE in metazoan neurodevelopment may occur through the JAK/STAT pathway. While the exact mechanisms remain to be determined, the phenotypes we observe in bin3 mutants (egg-laying, climbing, wing posture) can be attributable to neuromuscular defects, suggesting a strong conservation in the developmental function of Bin3 and MePCE in Drosophila and human development.

Bin3/MePCE methylates the gamma phosphate on the 5′ end of the 7SK snRNA, while another class of RNA methyltransferases methylates the N⁶ position of internal adenosines (N⁶-methyladenosine, m⁶A) on “RRACH” motifs (Wang et al. 2018) in a variety of RNAs (Lence et al. 2018). Interestingly, both human 7SK (Warda et al. 2017; Leger et al. 2021; Perez-Pepe et al. 2023) and mouse 7SK (Xu et al. 2022) contain the m⁶A modification. In HeLa cells, methylation of 7SK by the METLL3 m⁶A “writer” protein in response to EGF signaling allows hnRNPs to associate with 7SK; this causes HEXIM to dissociate from the 7SK snRNP and release P-TEFb, which activates transcriptional elongation (Perez-Pepe et al. 2023). In mice, methylation of 7SK at adenine 281 by METLL3 is required for the binding of m⁶A “reader” proteins to establish proper 7SK secondary structure (Xu et al. 2022). Drosophila 7SK contains a single RRACH motif at position 397-401; however, it is not known whether the adenine within this motif is methylated, or if m⁶A at that position would serve functions similar to those in HeLa cells or mice. Intriguingly, flies mutant for m⁶A writers (and m⁶A readers) exhibit the same neuromuscular defects in both climbing and resting wing posture (Lence et al. 2016; Kan et al. 2017) as flies mutant for bin3 (this study). Therefore, it is possible that Drosophila 7SK might also contain the m⁶A modification (at A399), and that m⁶A is important for 7SK stability and function in neurodevelopmental processes in Drosophila.

Although Bin3 and the m⁶A pathway might function synergistically to promote 7SK stability and function, they seem to act antagonistically on transcription elongation: Bin3 is required to prevent RNAPII pause-release, whereas the m⁶A pathway has been shown in Drosophila to promote RNAPII pause-release (Akhtar et al. 2021). Clearly, there is a complex balance between 7SK stability (via Bin3 binding and possible m⁶A methylation) as well as 7SK functions that sequester P-TEFb (Bin3-mediated function) or release P-TEFb (m⁶A modification function) that is governed by the activities of diverse classes of RNA methyltransferases, which in Drosophila contribute to neuromuscular function. Moreover, it appears that m⁶A modifications in 7SK might serve species-specific roles in 7SK biology.

Bin3 catalytic activity is dispensable for 7SK stability and snRNP function in vivo

MePCE modifies the 5′ end of 7SK RNA by catalyzing the addition of a monomethyl cap to the gamma phosphate (Gupta et al. 1990; Shumyatsky et al. 1990; Jeronimo et al. 2007; Xue et al. 2010; Yang et al. 2019). It has long been thought that the monomethyl cap is essential for 7SK stability. For example, in vitro transcribed capped 7SK was more stable than uncapped 7SK when injected into Xenopus oocytes (Shumyatsky et al. 1993). However, whether 7SK requires a monomethyl cap for stability in vivo has never been adequately addressed. We and others have shown that Bin3 and its orthologs are essential for 7SK stability, as deletion of bin3 (Singh et al. 2011), siRNA-mediated knockdown of MePCE in human cells (Jeronimo et al. 2007; Xue et al. 2010), and morpholino-mediated disruption of mepcea splicing in zebrafish embryos (Barboric et al. 2009) reduces 7SK levels. Importantly, these approaches do not explicitly test the in vivo requirement for methyltransferase activity, which requires the use of catalytic mutants.

Catalytic mutants of MePCE have been characterized in vitro. A triple mutation in the active site (V447A, L448A, and D449A) abolished methyltransferase activity, likely by preventing the interaction between MePCE and the methyl donor SAM. However, this mutant is also defective for binding to 7SK and therefore it is not a catalytic mutant per se (Xue et al. 2010). A double mutant in the active site (G451A and G455A) also abolished MePCE methyltransferase activity, again, presumably by preventing the interaction between MePCE and SAM; however, this mutant does bind to 7SK (Shelton et al. 2018). Finally, a single amino acid mutation in the active site (Y421A) reduced methyltransferase activity to 0.6% that of the wt enzyme. This is likely due to loss of hydrogen bonds between Y421 and both an adjacent residue (R425) and the 7SK gamma phosphate, which is the recipient of the methyl group from SAM (Yang et al. 2019). Importantly, Y421A was not expected to disrupt the overall structure of the active site, or binding to either the methyl donor cofactor (SAM) or the RNA substrate, but rather to disrupt the chemistry of methyl transfer to the gamma phosphate (Yang et al. 2019). Therefore, given the precise and specific nature of this active site mutation, our preferred choice for in vivo study of the importance of methyl capping was a Y795A substitution in Bin3 (orthologous to Y421A in MePCE).

Using in vitro methyltransferase assays, we showed that like the Y421A mutation in MePCE, the Y795A mutation in Bin3 completely abrogates catalytic activity. Unexpectedly, we found that flies expressing the catalytically inactive Bin3^Y795A mutant did not exhibit any of the physiological defects exhibited by bin3 deletion mutants, suggesting that catalytic activity is not required for Bin3 function in vivo. This result could be explained, in part, by our finding that Bin3^Y795A still bound and stabilized 7SK similarly to Bin3^wt (both in vivo and in vitro), despite its lack of methyltransferase activity in vitro (Fig. 10b). Thus, flies expressing the Bin3^Y795A allele retain the ability to assemble the 7SK snRNP to repress P-TEFb and promote normal development. Our findings suggest that in vivo capping of endogenous 7SK is not required for its stability per se, so long as Bin3 is able to bind to 7SK. Therefore, RNA binding contributes more to 7SK stability than does methyl capping, challenging the assumption that the gamma monomethyl cap is essential for 7SK stability.

That the Bin3^Y795A catalytic mutant protein was able to remain bound to 7SK is not entirely surprising. Y421 in MePCE (analogous to Y795 in Bin3) is located within a disordered region that becomes ordered upon binding to 7SK, forming an alpha helix (Yang et al. 2022); this disorder-to-order transition is required for MePCE methyltransferase activity. When LARP7 binds to 7SK and then associates with MePCE, the alpha helix containing Y421 adopts an extended conformation (Yang et al. 2022); this conformational transition induced by LARP7 binding inhibits MePCE methyltransferase activity (Xue et al. 2010; Brogie and Price 2017; Yang et al. 2022). Therefore, MePCE bound by LARP7 effectively phenocopies MePCE with a Y421A mutation, which is catalytically inactive (Yang et al. 2019). Thus, it appears that Y421 in MePCE and Y795 in Bin3 are not essential for 7SK snRNP assembly, but instead are essential only for methyl transfer. Additionally, the structure of the active sites of both Bin3^wt and Bin3^Y795A are not expected to be very different. SAM (i.e. the methyl donor) and unmethylated 7SK in the active site of Bin3^Y795A are expected to occupy a similar space as SAH and methylated 7SK in the active site of Bin3^wt, as the methyl group is positioned in the active site (of MePCE) in a near-transition state both prior to and after methyl transfer (Yang et al. 2019).

Bin3 and its orthologs are unusual enzymes, in that they remain associated with the products of the reaction they catalyze (Xue et al. 2010; Yang et al. 2019, 2022). MePCE has a higher affinity both for capped 7SK and SAH (the products) than uncapped 7SK and SAM (the reactants) (Yang et al. 2019). Therefore, the gamma monomethyl cap might be required not for 7SK stability per se, but to enhance the binding of MePCE to the 5′ end of 7SK to keep the enzyme bound after catalyzing the methyltransferase reaction. Constitutive binding of MePCE to 7SK would serve the dual purposes of physically protecting 7SK from pyrophosphatases (for which the gamma monomethyl cap is a substrate; Shumyatsky et al. 1990), and aiding in the recruitment of LARP7, which binds weakly to 7SK in the absence of MePCE (Xue et al. 2010), but enhances the affinity of MePCE for 7SK after binding (Muniz et al. 2013; Yang et al. 2022). However, our finding that the Bin3^Y795A catalytic mutant bound as much 7SK as Bin3^wt despite the lack of a gamma monomethyl cap is not consistent with such a model. Instead, our results support a model in which 7SK RNA sequence and secondary structure—and not the gamma monomethyl cap—is required for Bin3 and its orthologs to bind to 7SK. Indeed, MePCE bound equally well to transfected 7SK harboring either the normal gamma monomethyl cap or a tri-methyl guanosine (TMG) cap (Muniz et al. 2013); and Bmc1 (the S. pombe ortholog of Bin3) binds to both gamma monomethyl-capped U6 RNA and TMG-capped telomerase RNA (Páez-Moscoso et al. 2022; Porat et al. 2022). These findings suggest that a gamma monomethyl cap structure is not specifically required for constitutive or high-affinity binding by MePCE and its orthologs.

7SK sequence and secondary structure features stabilize the initial binding of MePCE (i.e. prior to the association of LARP7, which enhances the affinity of MePCE for 7SK). When MePCE initially binds to 7SK in the “nascent” conformation (adopted during 7SK synthesis), there is extensive hydrogen bonding between the methyltransferase domain and the 5′ hairpin of 7SK (Yang et al. 2019). Binding of MePCE to 7SK requires the first 2 (Singh et al. 1990) to 3 (Muniz et al. 2013) base pairs of the 5′ hairpin, and a portion of the single-stranded DNA 3′ to the hairpin (Shumyatsky et al. 1994; Muniz et al. 2013; Yang et al. 2019). Subsequently, LARP7 binds to the complex of MePCE and 7SK, and the conformation of 7SK changes from a “nascent” conformation to a “closed” conformation (adopted after formation of the core snRNP of MePCE, 7SK, and LARP7). In this conformation, MePCE forms protein–protein interactions with LARP7, and thereby has higher affinity for the “closed” conformation of 7SK than for the “nascent” conformation of 7SK, which lacks LARP7 (Yang et al. 2022). Therefore, both 7SK secondary structure and association with LARP7 are likely the major contributors to the constitutive binding of MePCE to 7SK.

One possible circumstance under which the gamma monomethyl cap on 7SK might be essential for stability is when the 7SK snRNP is actively repressing P-TEFb. For HEXIM to bind to 7SK and sequester P-TEFb, 7SK must undergo a conformational change from the “closed” conformation (i.e. the “core” snRNP conformation that does not bind P-TEFb) to the “linear” conformation (i.e. the “repressive” snRNP conformation that does bind P-TEFb). In this conformation, the 5′ end of 7SK is located outside of the MePCE active site (Yang et al. 2022), and therefore, in the context of the repressive 7SK snRNP, the 5′ end of 7SK might be solvent-exposed, which would necessitate a gamma monomethyl cap to protect 7SK from exoribonucleolytic degradation. However, these findings are based on cryo-EM structures using a recombinant MePCE comprising only the methyltransferase domain. Our finding that 7SK is not only bound but is also stabilized by Bin3^Y795A catalytic mutant protein (just like Bin3^wt protein) leaves open the possibility that there are regions outside of the methyltransferase domain in full-length Bin3/MePCE that protects the 5′ end of 7SK, even when the 5′ end is located outside of the active site. Moreover, the finding that the 5′ end of 7SK might not be bound by MePCE when in the “linear” conformation further supports a model in which the gamma monomethyl cap is not required for constitutive binding of MePCE to 7SK.

If the monomethyl cap is not required for either 7SK stability or for constitutive binding of Bin3, then what is the purpose of the methyltransferase domain in Bin3 and its orthologs? One possibility is that the methyltransferase domain is a vestige of an ancestral ortholog of Bin3 that did not remain constitutively bound to its RNA substrates, and therefore, the monomethyl cap would have been essential to protect these RNAs from 5′–3′ exoribonucleolytic degradation. The acquisition of additional protein binding constituents during evolution of a primordial 7SK snRNP complex might have necessitated that Bin3 remain associated with 7SK (after catalyzing the methyltransferase reaction) to facilitate the stable interaction of these additional components (like Larp7). This process would have rendered the monomethyl capping reaction dispensable for 7SK stability, as we observe. A second possibility, supported by the strong conservation of the methyltransferase domain over evolutionary time, argues that catalytic activity has an important function, perhaps on substrates other than 7SK, that has yet to be identified.

A metazoan-specific motif in Bin3 facilitates a 7SK-independent, tissue-specific neuromuscular function

S. pombe Bmc1, the most ancient ortholog of Bin3 studied to date, consists only of the conserved methyltransferase domain (Cosgrove et al. 2012; Páez-Moscoso et al. 2022; Porat et al. 2022). Over the course of evolution, Bin3 orthologs gained additional protein sequence flanking the methyltransferase domain (Cosgrove et al. 2012). We found a 16-amino acid motif N-terminal to the methyltransferase domain that is highly conserved in metazoans, and is mostly unstructured. As expected, given its location outside of the methyltransferase domain, this metazoan-specific motif (MSM) was not required for Bin3 to bind or stabilize 7SK in vivo, and was dispensable for fecundity and climbing ability. However, unexpectedly, flies expressing Bin3 deleted of the MSM (Bin3^ΔMSM) were unable to properly fold their wings, phenocopying bin3-null mutants (Fig. 10c). Genetically reducing the activity of P-TEFb in Bin3^ΔMSM flies rescued the held-out wings defect, a result that would not be expected if the Bin3 MSM functions independently of 7SK (and thereby, independently of P-TEFb). These findings suggest a model in which the Bin3 MSM plays a tissue-specific role in gene regulation, independent of 7SK, perhaps mediated by 1 or more MSM-Interacting Proteins (MIPs, Fig. 11, b and c). In this model (Fig. 11b), the Bin3 MSM mediates an interaction with 1 or more MIPs, that directly repress free P-TEFb (i.e. excess P-TEFb that is not sequestered into a 7SK snRNP) to prevent activation of genes that cause the held-out wing phenotype. Alternatively, a Bin3 MSM-MIPs complex could repress these genes through an independent, parallel pathway (Fig. 11c). The MIPs might be specifically expressed in the tissues involved in wing posture (e.g. the direct and indirect flight muscles); or, they might be ubiquitously expressed, but the interaction with the Bin3 MSM would only be necessary in tissues contributing to normal resting wing posture.

A 7SK-independent function of Bin3 is not without precedent. It was previously shown that fibroblasts from a patient with a nonsense mutation in MePCE (MePCE haploinsufficient cells) and cells deleted of LARP7 (LARP7 KO cells) both had reduced levels of 7SK, as expected, but that genes up-regulated in MePCE haploinsufficient cells were not affected in LARP7 KO cells. Moreover, exogenous MePCE rescued the up-regulation phenotype of MePCE haploinsufficient cells, despite only leading to a small increase in 7SK levels. Based on these findings, Schneeberger et al. (2019) hypothesized that MePCE might have a 7SK-independent role in regulating transcription, potentially by interacting with as-yet unknown proteins. Our finding that Bin3^ΔMSM binds and stabilizes 7SK but is not sufficient to rescue all bin3 mutant phenotypes supports the idea of a 7SK-independent function. We suggest that these as-yet unknown proteins interact with Bin3/MePCE through the MSM.