Search

C. elegans UNC-112 (kindlin) is required for muscle sarcomere assembly (Rogalski et al. 2000; Meves et al. 2009), and is one component of a conserved four-protein complex that associates with the cytoplasmic tail of integrin at the base of integrin adhesion complexes in muscle (Mackinnon et al. 2002; Lin et al. 2003; Norman et al. 2007; Qadota et al. 2014). UNC-112 binds directly with the cytoplasmic tail of PAT-3 (beta-integrin)(Qadota et al. 2012). UNC-112 binds to PAT-4 (ILK)(Mackinnon et al. 2002), and PAT-4 binds to both PAT-6 (alpha-parvin)(Lin et al. 2003), and UNC-97 (PINCH)(Mackinnon et al. 2002; Norman et al. 2007). A complex consisting of UNC-112, PAT-4, PAT-6 and UNC-97 has been demonstrated by co-immunoprecipitation (Qadota et al. 2014). UNC-112 is comprised of 720 amino acid residues and contains FERM and PH domains. The N-terminal half of UNC-112 (1-396 aa) can bind to the C-terminal half of UNC-112 (397-720 aa), and this interaction is inhibited by the association of PAT-4 (ILK) to the N-terminal half of UNC-112 (Qadota et al. 2012). In support of this model, we identified a D382V mutation that results in lack of binding to PAT-4 (Qadota et al. 2012). However, this residue is not conserved in human Kindlins (Figure 1D; D in UNC-112 but S in Kindlins), and mutation of this S to V in Kindlin-2 did not inhibit ILK binding (Huet-Calderwood et al. 2014). Here, we report identification of a novel UNC-112 mutation of a conserved residue that cannot bind to PAT-4. We found that E302G UNC-112 cannot bind to PAT-4, but still can bind to PAT-3 (beta-integrin) (Figure 1A). When we expressed HA tagged UNC-112 with E302G in C. elegans muscle (Figure 1B), HA tagged E302G UNC-112 cannot localize to the integrin adhesion complexes (dense bodies and M-lines) (Figure 1C). The E302 residue of UNC-112 is conserved in human Kindlin-1 and Kindlin-2, but not in Kindlin-3 (Figure 1D). Based on the only available crystal structure for a kindlin, that being for human kindlin-3 (Sun et al. 2020; Bu et al. 2020), we generated a homology model of UNC-112 and highlighted the locations of E302 and D382 (Figure 1E). Both E302 and D382 are located on surface loops of the structure. Mutating E302 to G, or mutating D382 to V resulted in no clashes with neighboring residues, based on the rotamer mutagenesis and energy minimization tools in Chimera (Pettersen et al., 2004). In fact, the lack of a sidechain of G, or the smaller sidechain of V, is likely to make these loops even more flexible. These mutations are predicted to not alter the overall structure of UNC-112 but could possibly affect the surface binding of UNC-112 to PAT-4. It should be noted, that in addition to our findings, a L located 6 residues C-terminal of D382, within the same FERM_M domain, and conserved in UNC-112 and all human kindlins (Figure 1D), when mutated to A, also greatly reduces binding to ILK (Huet-Calderwood et al. 2014). However, the results by Huet-Calderwood et al. were obtained by GST pulldown from lysates of tissue culture cells that overexpress both ILK and alpha-parvin. Our results, using the yeast two hybrid method and localization in worm muscle cells in which only UNC-112 was overexpressed, provide more evidence for an effect on direct binding between UNC-112 (kindlin) and PAT-4 (ILK). Our finding of a conserved residue in UNC-112 in the FERM_N domain that reduces binding to PAT-4, opens the door to testing the comparable residues in human Kindlin-1 and Kindlin-2 for their importance to binding to ILK. We do not know what the phenotype is of a nematode that is homozygous for the UNC-112 E302G mutation. Since UNC-112 function requires PAT-4 (Mackinnon et al. 2002), and all known mutations in pat-4 are Pat embryonic lethal, worms homozygous for UNC-112 E302G, might be Pat embryonic lethal. Several mutations in unc-112 are known to result in either the Pat embryonic phenotype or the adult viable Unc phenotype in which adults move slowly and have a disorganized myofilament lattice (Rogalski et al. 2000). The three known Pat alleles are unc-112(st562) and unc-112(st581) that are both nonsense mutations, and unc-112(gk1) that is a deletion. The Unc allele, unc-112(r367), is a missense mutation, T85I, and is also temperature sensitive. The molecular properties of UNC-112 T85I have not been explored, and thus, we do not know if the phenotype is due to reduced binding to PAT-4 or PAT-3, or whether it results in a unstable protein. Finally, there is one unusual allele created by CRISPR/Cas9, unc-112(kq715), L715E, which shows a defect in the migration of the distal tip cell (Park et al. 2020), but whether there was an effect on the myofilament lattice of body wall muscle, was not reported.

Additional file 1: Tables S1. List of the mutants, their measured muscle strength, normalized muscle strength and body diameter. Figure S1. The effect of body diameter on (A) the measured muscle strength and (B) the normalized muscle strength which is measured muscle strength divided by the cube of body diameter. Figure S2. 2D chemotaxis indices of WT and Unc mutants. None of the Unc mutants were significantly different from WT in the standard 2D chemotaxis on agar plates. Error bars are standard error of the mean. The number of replicates and average sample sizes (N) were, WT: 4, N=36, unc-54(s95): 4, N= 27, unc-22(e66): 2, N= 22, unc-54(s74): 4, N= 26, unc-22(sf21): 5, N=31. Statistical analysis was performed using a student’s t-test. Not significant: P-value>0.05. Figure S3. Phalloidin, MHCA, UNC-89, UNC-95, ATN-1 staining performed on sarcomere structural sites of wild type and muscle mutants. Supplementary note S1.

Using immunostaining with antibodies to myosin heavy chain A (MHC A), we show that there is a gradual decline in motility beginning at day 4 and the number of A-bands (a measure of thick filament assembly) beginning at day 8 adults. By day 12 and especially day 16, there is also disorganization of A-bands. This disorganization appears similar to that of unc-45 ts mutants grown at the restrictive temperature. We have found that in C. elegans a decline in hsp-90 mRNA (day 2) directly precedes a decline in HSP-90 protein (day3), which directly precedes a decline in UNC-45 protein (day4), which then in turn precedes a decline in Myosin B (day8), the main client of UNC-45. unc-45, unc-54, and myo-3 mRNA decline immediately after reaching maturity (day 1) and remain stable, though low – which is expected for proteins assembled into the sarcomere. We also observe a decrease in UNC-45 protein, but not mRNA, in an hsp-90 loss of function mutant, suggesting a role for HSP-90 in UNC-45 regulation and/or protein stabilization. We also observe early onset of sarcopenia when UNC-45 is lost during adulthood and an increase in an unknown UNC-45 post translational modification with age.

C. elegans UNC-112 (kindlin) is required for muscle sarcomere assembly (Rogalski et al. 2000), and is part of a conserved four protein complex that associates with the cytoplasmic tail of integrin at the base of muscle integrin adhesion complexes (Mackinnon et al. 2002; Lin et al. 2003; Norman et al. 2007; Qadota et al. 2014). UNC-112 consists of 720 amino acids and contains FERM and PH domains. UNC-112 N-terminal half (1-396 aa) can bind to UNC-112 C-terminal half (397-720 aa), and this interaction is inhibited by the association of PAT-4 (integrin linked kinase, ILK) to the UNC-112 N-terminal half (Qadota et al. 2012). The UNC-112 N-terminal half with either T346A or E349K mutations cannot bind to the UNC-112 C-terminal half, but can still bind to PAT-4 (ILK) (Qadota et al. 2012). To elucidate the molecular mechanism by which the UNC-112 N- and C-terminal halves interact genetically, we identified suppressor mutations in the UNC-112 C-terminal half that restore the ability of the UNC-112 C-terminal half to bind to the UNC-112 N-terminal half with T346A or E349K. We introduced random mutations into the UNC-112 C-terminal half (by error-prone PCR), and assayed for binding to the UNC-112 N-terminal half with T346A or E349K. From screening using the yeast two hybrid system and DNA sequencing of suppressor clones, we identified 5 single amino acid changes; R633G, S644C, N659D, I662T, and R663Q. S644C and R663Q were identified from screening with the UNC-112 N-terminal half containing T346A, and R633G, N659D, and I662T were identified from screening with the UNC-112 N-terminal half containing E349K. The wild type and all 5 suppressor clones failed to activate growth with an empty prey vector. Our results are summarized in Figure 1A. The UNC-112 C-terminal half with S644C or R663Q can bind to UNC-112 N-terminal half with T346A as expected from screening (see Methods), but also can bind to the UNC-112 N-terminal half with E349K. The UNC-112 C-terminal half with R633G, N659D, or I662T can bind to the UNC-112 N-terminal half with E349K as expected and but can also bind to UNC-112 N-terminal half with T346A. Interestingly, these 5 single amino acid changes are located within a 30 amino acid long segment (633 to 663; yellow bar in Figure 1). We speculate that these 30 amino acid residues of UNC-112 C-terminal half form a “Pocket” corresponding to the T346-E349 region of UNC-112 N-terminal half as a “Key”. We have shown previously, that only full-length UNC-112 can bind to cytoplasmic tail of PAT-3 (beta-integrin) by yeast two hybrid and biochemical methods (Qadota et al. 2012). Full length UNC-112 singly harboring the 5 suppressor mutations were tested for binding to the cytoplasmic tail of PAT-3. We found that 4 of 5 suppressors (R633G, S644C, I662T, and R663Q) could still bind to PAT-3, but full length UNC-112 with N659D could not bind (Figure 1B), suggesting that N659D causes a severe conformational change.