rpm-1 encodes an E3 ubiquitin ligase of 3766 amino acids, containing multiple domains that include an RCC1 like domain at the N-terimus, PHR domains in the middle, and a RING domain and a zinc finger box domain at the C-terminus (Schaefer et al., 2000; Zhen et al., 2000) . Loss of function of RPM-1 causes abnormal presynaptic development in several types of neurons, including the GABAergic motor neurons (Zhen et al., 2000) and touch receptor neurons (Schaefer et al., 2000) , as well as overextension of touch neuron axons (Grill et al., 2007) . During the cloning of rpm-1 (Schaefer et al., 2000; Zhen et al., 2000) , it was discovered that F07B7.12 , located at 35 kb away from rpm-1 ( Figure 1A ), is a near-perfect genomic duplication of 12 kb of rpm-1 from the promoter region to coding sequences for ~1900 amino acids, hence named as rpms-1 ( rpm-1 s hort). A transgene expressing the duplicated region of rpm-1 in an rpm-1 null mutant did not rescue synapse defects, showing that rpms-1 cannot compensate for the function of rpm-1 (Zhen et al., 2000) . To address the role of rpms-1 in rpm-1- mediated function directly, we used CRISPR editing to generate a double knockout of rpm-1 and rpms-1 .

As rpm-1 and rpms-1 have nearly identical sequences, sgRNAs designed for rpms-1 also bind rpm-1 and result in similar genome editing. We thus took advantage of rpm-1 ( ok364 ) animals, which have 2,221-bp deletion, removing part of large exon 6 and exon 7 and causing protein truncation due to frameshift. We designed two sgRNAs within the ok364 deletion region. Using the co-CRISPR strategy (Friedland et al., 2013) , we generated a new allele ju1285 , which is a 1,218-bp deletion in exon 6 of rpms-1 and causes frameshift followed by a premature stop codon. The rpm-1 ( ok364 ) rpms-1 ( ju1285 ) animals are homozygous viable, grossly indistinguishable from rpm-1 ( ok364 ). We examined touch receptor neuron morphology and GABAergic neuron synapses, and found no significant differences in axon overextension or synaptic puncta, with double mutants showing a slight reduction in touch neuron synapse branches ( Figure 1B, 1C ). DLK-1 is negatively regulated by RPM-1 via protein degradation (Nakata et al., 2005) . We measured the fluorescence intensity of endogenously expressed GFP:: DLK-1 ( ju1579 ) as previously described (Sun and Jin, 2023) , and did not detect a significant difference between the single and double mutant ( Figure 1D, 1E ). Moreover, after axon injury by laser axotomy, both rpm-1 and rpm-1 rpms-1 mutants showed axon regrowth to a similar level, comparable to that of control ( Figure 1F ). Together, these analyses show that rpms-1 does not contribute to rpm-1 mediated regulation of neuronal development.

1. CRISPR-Cas9 mediated genome editing

We used CRISPR sgRNA design tool http://crispr.mit.edu (Feng Zhang's lab). Mutagenesis primers were ordered (EtonBio) and used to insert sgRNAs into eft-3 p ::cas9-NLS-pU6 empty vectors by Phusion PCR. Plasmids were treated with DpnI to digest plasmids of bacteria origin (methylated). DH5α transformation followed. The following plasmids were obtained and sequenced: pCZ890 (sgRNA#1) and pKEN268 (sgRNA#2). CZ3007 rpm-1 ( ok364 ) worms were injected with 50 ng/μl of each plasmid DNA and 5 ng/μl pCFJ104 myo-3 p ::RFP co-injection marker. F1 positive worms were let lay enough eggs and genotyped with primers YJ10660 (5'-ACCGATATGACTGGATATGAAAATCGTC) and YJ11134 (5'-GGCCATTCGCTCCCATAAC) for rpms-1 ( ju1285 ).

2. Laser axotomy

We cut PLM axons in anesthetized L4 worms using a near-infrared Ti-Sapphire laser (KMLabs) as described (Wu et al., 2007) . We used muIs32 ( mec-7 p ::GFP) to visualize touch neuron morphology and axon regeneration. To anesthetize worms for surgery and imaging, we put animals in the agar pad and used 0.1~1% 1-phenoxy-2-propanol in M9 buffer. For confocal imaging, we used an LSM710 confocal microscope to take Z-stack images of live anesthetized worms. Representative images for axon regrowth are projections or single layers of Z-stack images. Regrowth lengths were measured from maximum transparency projections of one single z-stack using Zeiss AIM software.

3. Fluorescence microscopy and GFP intensity measurement

Confocal images of the nerve ring region were collected from L4 immobilized in 2 mM levamisole (Sigma) in M9 buffer using a Zeiss LSM710 confocal microscope. Projections of Z-stack images (1 μm/section) are shown in the figure. GFP intensity from the region of interest (ROI) was analyzed in ImageJ, and the graph was generated in GraphPad Prism.

4. Statistics

For comparisons involving multiple groups, we used one-way ANOVA in GraphPad Prism. When comparing two groups, we used an unpaired t test with two-tailed P value.

CZ3007 : juIs1 [ unc-25p:: SNB-1::GFP]; rpm-1 ( ok364 ) V

CZ9840 : muIs32 [ mec-7 p ::GFP] II; rpm-1 ( ok364 ) V

CZ10969 : muIs32 [ mec-7 p::GFP] II

CZ22189 : juIs1 [ unc-25p:: SNB-1::GFP]; rpm-1 ( ok364 ) rpms-1 ( ju1285 ) V

CZ22190 : rpm-1 ( ok364 ) rpms-1 ( ju1285 ) V

CZ22191 : muIs32 ; rpm-1 ( ok364 ) rpms-1 ( ju1285 ) V

CZ26773 : muIs32 [ mec-7 p ::GFP] II; rpm-1 ( ok364 ) V

CZ26774 : muIs32 [ mec-7 p ::GFP] II; rpm-1 ( ok364 ) rpms-1 ( ju1285 ) V

CZ25941 : GFP:: dlk-1 ( ju1579 ) I

CZ26775 : GFP:: dlk-1 ( ju1579 ) I; rpm-1 ( ok364 ) rpms-1 ( ju1285 ) V

CZ26787 : GFP:: dlk-1 ( ju1579 ) I; rpm-1 ( ok364 ) V