The debranching enzyme Dbr1 regulates lariat turnover and intron splicing

SUMMARY The majority of genic transcription is intronic. Introns are removed by splicing as branched lariat RNAs which require rapid recycling. The branch site is recognized during splicing catalysis and later debranched by Dbr1 in the rate-limiting step of lariat turnover. Through generation of the first viable DBR1 knockout cell line, we find the predominantly nuclear Dbr1 enzyme to encode the sole debranching activity in human cells. Dbr1 preferentially debranches substrates that contain canonical U2 binding motifs, suggesting that branchsites discovered through sequencing do not necessarily represent those favored by the spliceosome. We find that Dbr1 also exhibits specificity for particular 5’ splice site sequences. We identify Dbr1 interactors through co-immunoprecipitation mass spectroscopy. We present a mechanistic model for Dbr1 recruitment to the branchpoint through the intron-binding protein AQR. In addition to a 20-fold increase in lariats, Dbr1 depletion increases exon skipping. Using ADAR fusions to timestamp lariats, we demonstrate a defect in spliceosome recycling. In the absence of Dbr1, spliceosomal components remain associated with the lariat for a longer period of time. As splicing is co-transcriptional, slower recycling increases the likelihood that downstream exons will be available for exon skipping.


33
The removal of introns from pre-mRNA by splicing is an integral step in gene expression. Splicing is a 34 two-step process performed by the spliceosome, a large macromolecular machine that rivals the 35 ribosome in complexity. The first step involves the generation of a 2'-5' linkage between the 5' splice 36 site and a branchpoint nucleotide that typically resides near the 3' end of the intron. In the second step,

41
First identified as a modifier of retrotransposon activity in yeast, the lariat-debranching enzyme Dbr1 is 42 responsible for the linearization of intron lariats -the rate limiting step in intron turnover 3,4 . The 43 alignment of high-throughput RNA sequencing reads to the human genome has mapped core gene 44 expression elements like start, splice and polyadenylation sites to saturation in many tissues 5 . The 45 branchpoint is the only obligate cis-element of intron-containing genes that has not been mapped to a 46 similar level of completion.

47
The current set of annotated branchpoints has been discovered by lariat sequencing, a technique that 48 exploits the ability of reverse transcriptase to read through a branchpoint while copying a lariat into 49 cDNA 6 . Our lab created lariat-seq by adapting this method to high throughput sequencing data 7 . These

85
The predominantly nuclear Dbr1 supplies all debranching activity in HEK293T cell extract 86 Currently, annotated branchsites are inferred from lariat-seq which samples the steady state pool of 87 cellular lariats. The identity of lariats in this pool is influenced both by the sequence requirements of 88 branchsite selection in splicing and the sequence specificity of the debranching enzyme Dbr1. To 89 decouple these two processes, a commercial homologous recombination CRISPR kit was used to 90 knock out DBR1 in HEK293T cells. Multiple clones were isolated and screened for recombination 91 events (Supplemental Fig. 1A). To test if a disruption was sufficient to deplete Dbr1, whole cell extract 92 was prepared. Western blot analysis visualized moderate Dbr1 expression in C9 but greatly reduced 93 (C19) or absent (C22) expression in the DBR1-clones confirming the PCR screening (Fig. 1A).

94
Presumably, the faint band reactive to Dbr1 antibodies (C19 lane) represents an inactive missense 95 mutant that arose through non-homologous end-joining. We tested the extracts in a cell-free 96 debranching assay that utilized a quenched fluorophore in a synthetic branched RNA oligonucleotide 17 .

97
While C9 extract debranched at near wildtype levels, the debranching activity of C19 and C22 was 98 indistinguishable from the no-extract (i.e. null) control (Fig. 1B). The Dbr1 status of this cell line was 99 further characterized by immunohistochemistry. Dbr1 was localized to the nucleus in wild type 100 HEK293T cells but undetectable in the DBR1-negative cell line (C22, Fig. 1C). To measure the sub-101 cellular localization of lariats, single molecule FISH was designed to visualize the location of the Taok2 102 intron 13 in wild-type and Dbr1-depleted (C22) cells (Fig. 1D). As expected, the overall number of 103 lariats increased two-fold in the knockout clone but also shifted to the cytoplasm. Taken together, these 104 results suggest a primarily nuclear Dbr1 encodes the sole debranching activity in HEK293T cells.

105
Lariats that escape debranching in the nucleus accumulate in the cytoplasm and that the null C22   sizeable increases in overall intronic coverage for these lariat-associated introns (Fig. 2B)

139
introns from the same genes upon DBR1 knockout (Fig. 2E). Over 50,000 branchpoints in total were 140 identified from RNA-Seq of DBR1 knockout cells (Fig. 2F), and the lariats captured in a DBR1-knockout 141 background are more than two-fold enriched for 'A' branchpoints ( Fig. 2G). This increase in the 142 proportion of 'A' branchpoint lariats represents a molecular signature of lower DBR1 activity consistent 143 with earlier studies on the enzymatic specificity of Dbr1 13 . To characterize the relationship between 144 branchsite motif and splicing activity, we performed a massively parallel reporter assay that tested 145 every hexamer's ability to function as a branchsite in the context of APRT intron 2 (Supplemental Fig.   146 2). Annotating the branchpoints from both DBR1 knockout and wildtype contexts with the functional 147 activity scores learned from our splicing assay indicates that branchpoints mapped in the Dbr1-deficient 148 environment are more likely to be used than those found in the wildtype (Fig. 2H)

168
To better understand the mechanism of debranching, a mass spectrometry approach was used to 169 define the network of Dbr1 interacting partners. Briefly, FLAG-tagged Dbr1 was expressed in HEK293T 170 cells (Fig. 3A, left). Coomassie staining showed Dbr1 complexes were efficiently isolated through FLAG 171 antibody conjugated magnetic beads, and the components of these complexes were analyzed by mass 172 spectrometry (Fig. 3A, right). 120 proteins were identified as highly enriched in the eluate above control 173 (Supplemental Table 1). Analysis of these interactors using the Gene Ontology and Reactome 174 databases revealed an enrichment for many RNA processing pathways including rRNA and ncRNA 175 processing (Supplemental Table 2). Dbr1 interacts with members of the U5, U2, NTC, and NTR 176 spliceosomal sub-complexes as well as Cwf19L, the human homolog of yeast Drn1 (Supplemental 177 Table 1). Dbr1 has not been isolated with the spliceosome and is not part of existing structures.

193
Dbr1 also co-localize in vivo in nuclear speckles (Fig. 3D).   Our proteomic analysis revealed Dbr1 interacts with many splicing factors (Supplemental Table 1). To 251 explore the role of Dbr1 in splicing, total RNA was extracted from wildtype and Dbr1-deficient cells.

252
Isoform analyses were performed by rMATS to identify splicing events that differed between DBR1-253 knockout and wildtype states. The loss of Dbr1 affected the inclusion of cassette exons most strongly.  297 depletion (Fig. 6C, Supplemental Fig. 3A). Surprisingly, cloning small RNA from C22 and wildtype cells wildtype lariats (Fig. 2D), and minor spliceosomal introns with alternate 5' splice site motifs only 329 undergo about half the lariat level increase upon DBR1 knockout that major spliceosomal introns do 330 (Fig. 2E). This discrepancy suggests that minor intron lariats are more reliant on Dbr1-independent 331 pathways for degradation.

332
Dbr1 is unlikely to act alone as co-IP mass spectrometry reveals interactions with many other factors 333 involved in RNA processing (Supplemental Table 2

574
In vitro Dbr1 inhibition assay: 20 nM Dbr1 was combined with 0.2 uM AK88 and the concentrations of 575 G-G cap analog (NEB S1407), G-A cap analog (NEB S1406), and cGAMP (InvivoGen 576 nacga23).Product development was measured as described in the "Debranching assay and Western  conditions. For comparison, a control set of introns was constructed containing introns that lack these 588 RBP binding sites but are in the same genes as the RBP-bound set, and the lariat read recovery rate 589 was also calculated for this set.

645
QIAGEN CLC Genomics Workbench with the default setting. snoRNA reads were then normalized and 646 tested for differential expression using edgeR.