Fig 2: Functional prioritization of localized mRNAs and identification of Kif1c as a model for further study. (A) Outline of phenotypic screen. Coding sequences of candidate genes were targeted individually with CRISPR/Cas9, selected with puromycin, and then analyzed for phenotypes at two different time points. (B) Schematic of phenotypic analyses. (Left) Cells were measured along their longest axis (red lines). (Right) Speed and persistence were measured from cell tracks (blue arrows) after calculating distance (total path length) and displacement (straight line between first and last time point). (C–E) Only the top treatment (guide and time point) is shown for each candidate gene. Genes are ordered based on their median measurement; horizontal dashed lines mark the median of nontarget control treatments. Nontarget guides, Dynll2, Eif3f, and Kif1c are denoted in black or color. All other candidates are in gray. Cells were tracked migrating on uncoated plastic dishes for 5 h. N = >15 cells per assay for candidate genes. N = >150 cells per assay for nontarget guides. (C) Box plots of cell length measured from still frames. (D) Box plots of migration speed measurements. In C and D, the box shows the 25th to 75th percentiles, and the line in the middle is the median. Whiskers are drawn down to the smallest value and up to the largest value. (E) Migratory persistence over time. (F,G) Mean and SEM from qRT-PCR on cells with endogenous 3′ UTR deletions. qRT-PCRs on fractionated cells were normalized to the diffuse mRNA Arpc3. qRT-PCR on bulk mRNA was normalized to Ywhaz. N = 2–4 biological replicates per condition. All statistical tests are unpaired t-test compared with +/+. (***) P < 0.001, (**) P < 0.01, (*) P < 0.05.

Fig 3: Identification and prioritization of mRNAs localized to protrusions in melanoma cells. (A) YUMM1.7 cells are a morphologically diverse and migratory cell type. (B) YUMM1.7 cells were fractionated into protrusions and cell bodies using microporous membranes. (C) RNA-seq results for fractionated YUMM1.7 cells. Control mRNAs such as Ppia, Ywhaz, RhoA, and Arpc3 are uniformly distributed between protrusions and cell bodies (red dots), while nucleolar RNAs are depleted from protrusions (black). (D) Protrusion enrichment values for the top 137 localized mRNAs in YUMM1.7 cells. In C and D, Dynll2, Eif3f, Kif1c, and other candidate genes are denoted in color. (E) Thirty-five mRNAs that localize in melanoma cells are also localized in breast cancer cells (blue), embryonic fibroblasts (orange), or all three cell types (black). (F) Candidate mRNAs and their protrusion enrichment patterns in three nonneuronal cell lines. The table is arranged based on the candidates’ enrichments across three, two, or one cell type(s).

Fig 4: Kif1c mRNA localization is required for directed cell migration but dispensable for mRNA trafficking. (A) Schematic of the Kif1c genomic locus. The red arrowhead and stop sign in the inset denote cut site and location of premature termination codon in LOF alleles, respectively. LOF1 and LOF2 harbor 2- and 4-bp deletions in exon 11, respectively. (B) qRT-PCR on cells with endogenous deletions; mean and SEM are shown. N = 3–9 biological replicates. (C) Representative Western blot of total protein lysates. The experiment was repeated three times. (D,E) Cells were plated sparsely on fibronectin-coated plastic dishes for 18–24 h and then tracked every 3 min for at least 5 h. (D) Plot of migration persistence over time. Fifteen or more cells per genotype were used per experiment. The experiment was repeated three or more times per genotype. Error bars show SEM. (E) Individual cell tracks from a single replicate for the given genotypes. (F) qRT-PCR measurement of mRNA enrichment in protrusions in the indicated cell lines; mean and SEM are shown. All qRT-PCR data were normalized to Ppia and Ywhaz. N = 3–4 biological replicates per condition. All statistical tests are ordinary one-way ANOVA compared with parental. (***) P < 0.001, (**) P < 0.01.

Fig 5: Identification of endogenous KIF1C binding partners and their regulation by Kif1c mRNA localization. (A) Mass spectrometry of endogenous KIF1C IPs; the size of the dot represents the abundance of the interactor. Only hits more than threefold enriched over Kif1cLOF were included. Functional categories were determined using PANTHER and manual curation; many genes fall into more than one category. N = 3 biological replicates per condition. (B) Analysis of the expected and observed numbers of cell body- and protrusion-enriched proteins among the KIF1CΔGA-preferred interactors (interactors with log2FC > 0). (*) P < 0.05 (χ2). (C) Fold change (FC; Kif1cΔGA/Kif1c+/+), abundance rank, and a brief description of candidate proteins selected for validation. (D) Western blot analysis of co-IP assays with endogenous KIF1C from untagged cells (left) or mCherry-tagged KIF1C cells (right) using α-KIF1C or α-mCherry antibodies, respectively. Note that endogenous KIF1C is expressed at low levels and is undetectable in input samples under these conditions. (E) Proposed model: In Kif1c+/+ cells, Kif1c mRNA (purple) localizes to protrusions. KIF1C protein undergoes limited, specific protein–protein interactions and transports cargoes related to mRNA trafficking (e.g., APC) and cell migration (e.g., integrin). In Kif1cΔGA cells, Kif1c mRNA is diffuse in the cell body, and KIF1C protein engages with additional, aberrant interactors. KIF1C is still able to transport cargoes for mRNA trafficking, but cargoes for cell migration are affected through an unknown mechanism (faded cell migration cargo), which may include the cargo being lost, diminished, or unable to recycle properly. In Kif1cLOF cells, Kif1c mRNA and KIF1C protein are absent, and all cargos are affected.