The Muscleblind family comprises a group of evolutionarily conserved RNA-binding proteins that play key roles in various aspects of RNA metabolism, most notably in the regulation of alternative splicing. A defining feature of these proteins is the presence of tandem zinc finger domains, each composed of three cysteine residues and one histidine residue (Fernandez-Costa et al., 2011). Caenorhabditis elegans has a single ortholog of Muscleblind protein, MBL-1 , which shows prominent expression in the nervous system. Loss-of-function studies have shown that mbl-1 is essential for the synaptic formation at the neuromuscular junctions (Spilker et al., 2012), dendritic morphogenesis in PVD sensory neurons (Xie et al., 2023), microtubule stability and axonal growth in touch receptor neurons (TRNs), and alternative splicing of terminal selectors like mec-3 (Lee et al., 2024). In addition to its classic role as a regulator of RNA splicing, recent studies have shown that MBL-1 also modulates mRNA stability through direct binding to target transcripts (Puri et al., 2023; Verbeeren et al., 2023). Since certain MBL-1 isoforms possess a pair of nuclear localization signals (NLS) while other isoforms do not, the protein may function both as a splicing regulator in the nucleus and an mRNA stabilizer in the cytoplasm. To disentangle the two functions, we engineered three alleles that produced MBL-1 proteins exclusively with or without the NLS and found that its presence in the cytoplasm is likely necessary to promote axonal growth in the TRNs.

Previous studies identified a bipartite nuclear localization signal (NLS) in mammalian Muscleblind MBNL proteins, consisting of two repeats of lysine-arginine residues (KR motifs) that regulate their subcellular localization (Kino et al., 2015). This bipartite KR motif is evolutionarily conserved in C. elegans MBL-1 , with one located near the end of exon 8 and the second at the beginning of exon 9 (Verbeeren et al., 2023). Since exon 8 can be selectively included in some but not all isoforms, the gene can code for MBL-1 isoforms with one or two KR motifs ( Fig. 1A ). To perturb the NLS in MBL-1 , Verbeeren et al. previously generated the mbl-1 ( syb4318 ) and mbl-1 ( syb4345 ) alleles. The syb4318 allele featured a deletion of exon 7 and 8 and part of the flanking introns, resulting in the expression of isoforms with only one KR motif, whereas the syb4345 allele involves a deletion of exon 7 and its flanking introns, leading to the expression of MBL-1 isoform with two KR motifs if the connected exons 6 and 8 are included in the mRNA. However, if the exon 6&8 is skipped, which would happen in the alternative splicing of mbl-1 mRNA, the syb4345 would still produces proteins with only one KR motif. Moreover, whether the single KR motif could still contribute to nuclear localization is unclear.

To address the above issues, we created three additional mbl-1 alleles through CRISPR/Cas9-mediated gene editing. First, the unk249 allele deleted exons 7 and 8 along with their flanking intronic sequences (thereby directly joining exons 6 and 9) and produced MBL-1 proteins with only one KR motif. Second, the unk344 allele was built on top of unk249 by further deleting the remaining KR motif, thus generating MBL-1 proteins with no KR motifs. Third, the unk276 allele, in which exon 7, the introns flanking exon 7, and the intron between exons 8 and 9 were deleted, resulting in the fusion of exon 6, 8, and 9 and the production of MBL-1 proteins with two KR motifs only ( Fig. 1A ). We conducted RT-PCR to examine and sequence the transcripts of mbl-1 in animals carrying the three alleles and confirmed that the gene editing indeed changed the sequence of the mbl-1 transcripts as expected ( Fig. 1A ).

To understand the functional significance of these MBL-1 isoforms in promoting axonal growth, we crossed the above mbl-1 alleles into the mec-7 ( u278 ) mutants, which served as a sensitized background to test the effects on neurite growth. mec-7 codes for a TRN-specific β-tubulin, and the u278 (C303Y) is a gain-of-function mutation that led to the growth of a very long, ectopic posteriorly directed neurite in the ALM neurons (termed as ALM-PN) (Zheng et al., 2017). This ALM-PN does not exist or is very short in the wild-type animals. We previously found that the loss of mbl-1 completely suppressed the growth of ALM-PN in the mec-7 ( u278 ) mutants, suggesting that MBL-1 promotes neurite growth (Lee et al., 2024). Similar to the mbl-1 null allele, the unk276 allele (which only produces MBL-1 proteins with two KR motifs) also suppressed the ALM-PN growth ( Fig. 1B ), suggesting that the cytoplasmic presence of MBL-1 is likely required for its activity in promoting neurite growth. Both unk249 and unk344 alleles (which produces MBL-1 proteins with one or no KR motif) failed to suppress ALM-PN growth, suggesting that the KR motifs and nuclear localization may not be required for MBL-1 's function in inducing neurite growth in the mec-7 ( u278 ) background.

To confirm the subcellular localization of the MBL-1 proteins produced by the above three alleles, we cloned the mbl-1 gene from the mutants, fused them with GFP-coding sequences, and expressed the fusion proteins under the TRN-specific mec-17 promoter. These reporters were introduced into the wild-type, mec-7 ( u278 ) , and mec-7 ( u278 ) mbl-1 (-) animals. As expected, the proteins with two KR motifs were restricted to the nucleus, whereas the proteins with one KR motif showed diffusive expression throughout the TRN cell body ( Fig. 1C ). To our surprise, MBL-1 proteins with no KR motifs still showed a diffusive localization pattern in the cells and was not excluded from the nucleus. This result hinted that MBL-1 may be able to enter the nucleus in a mechanism that is independent of the two KR motifs. The shorter isoforms (which skipped exon 7 and 8) likely have both cytoplasmic and nuclear localizations.

To confirm that the nucleus-localized MBL-1 produced by the unk276 allele is capable of splicing target genes, we analyzed four known MBL-1-regulated splicing events (Lee et al., 2024). The unk276 animals could promote the normal splicing of pqn-52c , pqn-72b , lgc-22a , and kin-4h isoforms, suggesting that the long MBL-1 isoform is functional in controlling mRNA splicing. The mbl-1 ( tm1563 ) deletion mutants served as a negative control ( Fig. 1D ). The MBL-1 proteins with one or no KR motifs produced by unk249 and unk344 alleles, respectively, could not promote the normal splicing of pqn-52c and pqn-72b but was able to promote the splicing of lgc-22a and kin-4h to the same extent as MBL-1 produced by unk276 . Thus, the short MBL-1 isoform may still possess some ability to regular nuclear splicing, which is consistent with their nuclear localization.

MBL-1 is known to interact and stabilize the mRNAs of microtubule-related genes (such as mec-17 /tubulin acetyltransferase, mec-7 /β-tubulin, and mec-12 /α-tubulin) (Puri et al., 2023), and we suspect that this mRNA-stabilizing role in the cytoplasm is essential for MBL-1 's function in promoting microtubule stability and neurite growth. However, since we were not able to generate a version of MBL-1 that is exclusively cytoplasmic, it remains unclear whether its nuclear localization is also required for its function in neurite extension. Our previous work found that MBL-1 promotes the splicing of mec-3 , which activates the expression of mec-17 , mec-7 , and mec-12 . It is possible that MBL-1 's canonical function as a splicing regulator in the nucleus also contribute to microtubule stabilization and neuronal morphogenesis.

To generate the three mbl-1 mutant alleles, we used CRISPR/Cas9-mediated genome editing to introduce double-strained breaks at two targeted sites (exons 6 and 9) in the endogenous mbl-1 locus. Specifically, pairs of single guide RNAs (sgRNAs) were synthesized using the EnGen sgRNA Synthesis Kit (NEB, E3322V). A total of 1 μg of each sgRNA pair, combined with 20 pmol of recombinant Cas9 protein (EnGen S. pyogenes Cas9 NLS, NEB, M0646T), was microinjected into the gonads of young adult C. elegans . For precise editing of selected exons, we followed an established protocol that uses single-stranded DNA oligonucleotides (0.1 μg/μl) as homologous repair templates (Dokshin et al., 2018). To prevent re-cleavage by Cas9, synonymous mutations were incorporated into the repair templates at the protospacer-adjacent motif (PAM) sites.

To create TRN-specific fluorescent reporter constructs for the three mbl-1 mutant alleles, we amplified the corresponding mutant genomic sequences and inserted them into a vector downstream of a 1.9 kb mec-17 promoter and in-frame with GFP using Gibson Assembly (ClonExpress II One Step Cloning Kit, Vazyme Biotech, Nanjing, China). These plasmid constructs were then microinjected into the gonads of young adult C. elegans to generate transgenic lines carrying extrachromosomal arrays.

To conduct RT-PCR, total RNA was extracted from L4 animals using TRIzol reagent (Thermo Fisher). cDNA libraries were prepared through reverse transcription of the total RNA using SuperScript II Reverse Transcriptase with oligo(dT)s (Thermo Fisher). Four candidates were selected for semi-quantitative RT-PCR using isoform-specific primers based on previous studies (Lee et al., 2024). ama-1 was used as an internal control for RT-PCR.

Fluorescence imaging was performed on a Leica DMi8 inverted microscope equipped with a Leica K5 monochrome camera. Images were acquired and analyzed using Leica Application Suite X software (version 3.7.2.22383). Measurements of ALM-PN length were obtained from day-1 adult animals cultivated at 20 °C, with at least 20 individuals scored per genotype.