The Rho-GEF PIX-1 directs assembly or stability of lateral attachment structures between muscle cells

Figures

• 

  Fig. 1. Identification of pix-1 as a gene required for the assembly of IACs at muscle cell boundaries. a Confocal images of several body wall muscle cells immunostained with antibodies to PAT-6 (α-parvin) from wild type and the strain VC20366 identified by screening 574 MMP strains. Arrowheads point to the boundaries between muscle cells. b Schematic representation of domains in C. elegans PIX-1a, and the location and nature of 7 pix-1 mutants and evaluation of their phenotypes. The asterisk denotes the pix-1 mutant allele found in the original strain VC20366. c Muscle-specific expression of a wild-type cDNA for PIX-1 tagged with HA rescues the phenotype of pix-1(gk299374). The transgene is sfEx61[myo-3p::HA-PIX-1; sur-5::nls::GFP], in which sur-5::nls::GFP is the transformation marker showing GFP in nuclei. Note that PAT-6 has been restored to the muscle cell boundaries (indicated by arrowheads), and that HA-tagged PIX-1 localizes to muscle cell boundaries, dense bodies, and M-lines. d Comparison of wild type vs. pix-1(gk299374) immunostained with antibodies to the indicated IAC proteins and imaged by confocal microscopy. Arrowheads denote muscle cell boundaries. Note that all four…

• 

  Fig. 2. Live imaging of cortical F-actin at muscle cell boundaries. SIM images of portions of two adjacent body wall muscle cells from a nematode strain in which a muscle myosin MHC A was tagged with GFP by CRISPR and LifeAct-mCherry was expressed in muscle cells from a transgene. GFP-MYO-3 (MHC A) labels the middle of sarcomeric A-bands, and LifeAct-mCherry labels I-bands, except for F-actin at the boundary between two adjacent muscle cells (indicated by yellow arrows). Note how the signal from the F-actin at the boundary diminishes as the focal plane changes from close to the outer muscle cell membrane to deeper into the myofilament lattice whereas the F-actin signal from I-bands does not change. Also note that in the pix-1 nonsense mutant, gk299374, there are two bands of cortical F-actin at the boundary. Each image is a representative image obtained from imaging at three different animals of each strain. Scale bar, 5 μm.

• 

  Fig. 3. PIX-1 is most similar to human β-PIX and acts as a GEF for CED-10 (Rac) in muscle. a Schematic representation of predicted domains in C. elegans PIX-1a, human α-PIX and human β-PIX. The bottom triangles show a comparison of percentage identities for SH3, RhoGEF, and coiled-coil regions. b Validation of activation assay. Lysates were prepared from nematodes expressing HA-CED-10 in body wall muscle using a muscle-specific promoter, an excess of GDP or GTPγS, was added, and then beads coupled to GST-PAK-PBD were used to pull down activated CED-10 (i.e. bound to GTP or GTP + GTPγS). Samples were separated on a gel, blotted and incubated with antibodies to HA. The asterisk indicates the position of GST-PAK-PBD on the blot. Arrow indicates the position of HA-CED-10•GTP/GTPγS on the western blot. c Deficiency of PIX-1 results in a reduced level of activated CED-10. Lysates were prepared from two strains, each expressing HA-CED-10 in body wall muscle: wild type, and pix-1(gk299374), a nonsense mutant that results in no detectable PIX-1. These lysates were incubated with beads coupled to GST-PAK-PBD, and used to pull down activated CED-10. Both the Ponceau S stained blot, and the result…

• 

  Fig. 4. Both loss of function and overexpression of pix-1 results in reduced locomotion. a Swimming and b crawling assays show that loss of function mutations in pix-1 result in reduced locomotion compared to wild type. c Swimming and d crawling assays show that both the integrated rescued strain, pix-1(gk299374); sfIs20, and wild type expressing the rescuing transgene from an extrachromosomal array, sfEx61, have reduced locomotion. In contrast, CRISPR/Cas9 repair of the nonsense mutation in pix-1(gk299374), called pix-1(syb2137 gk299374), results in normalization of locomotion. In the graphs, each open circle represents the result from an independently selected animal. The exact n values vary, and these data can be found in the Source Data Files. Student’s two-sided t test was used to test for significance. Error bars: standard deviations; ****p ≤ 0.0001; n.s.: no significant difference.

• 

  Fig. 5. Overexpression of pix-1 results in disruption and CRISPR repair of pix-1(gk299374) results in normalization of muscle cell boundaries. a, b Western blot analysis showing overexpression of PIX-1 in the strain [pix-1(gk299374); sfIs20] in which the muscle cell boundary defect observed in the nonsense mutant pix-1(gk299374) has been rescued by an integrated array of wild-type pix-1 expressed from a muscle-specific promoter (see Fig. 1c). a A representative western blot is shown reacted with antibodies to PIX-1, and to the loading control histone H3. b Graphical summary of four independent western blot reactions for wild type vs. pix-1(gk299374); sfIs20. Means and standard errors of the means are shown. The two strains show statistically different levels of PIX-1 using a two-sided students t-test with p < 0.0001 (indicated by ****). Images of the entire western blots are provided in the Source Data Files. c This same array when expressed in a wild-type background [wild-type; sfEx61] disrupts the muscle cell boundary. d pix-1(syb2137gk299374) is pix-1(gk299374) after CRISPR/Cas9 was used to repair the nonsense mutation. Confocal images of anti-PAT-6 are shown and reveal that the…

• 

  Fig. 6. Antibodies to PIX-1 detect PIX-1a on a western blot and localize to muscle cell boundaries, M-lines, and dense bodies. a Schematic representation of domains in the predicted isoforms PIX-1a and PIX-1b and regions used as immunogens to generate antibodies. b Western blot detection of a protein of expected size for PIX-1a from wild type and from pix-1(gk893650) [P190S] but not from deletion or nonsense pix-1 mutants. Anti-paramyosin was used as a gel loading control. An image of the entire blot reacted with anti-paramyosin is available in the Source Data Files. c Antibodies to PIX-1 localize to muscle cell boundaries and with less intensity to M-lines and dense bodies. Each strain was co-stained with antibodies to PAT-6 and PIX-1, and imaged by confocal. Note the lack of PIX-1 immunostaining in the nonsense and deletion pix-1 mutants, but strong and disorganized staining in pix-1(gk893650) [P190S]. Arrowheads point to a muscle cell boundary. Each image is a representative image obtained from at least 2 fixation and immunostaining experiments, and imaging at least three different animals of each strain. Scale bar, 10 μm.

• 

  Fig. 7. Mutations in genes encoding known proteins of a PIX-1 pathway result in muscle cell boundary disruption. a Putative PIX-1 biochemical pathway based on what is known of PIX proteins in mammals and other cell types in C. elegans. Also indicated is a summary of the results shown in b. b Confocal images of body wall muscle from the indicated mutants immunostained with anti-PAT-6. Note reduced or disorganized PAT-6 localization at muscle cell boundaries (indicated by arrowheads) in pak-1, git-1, ced-10, and pak-2 mutants, but normal PAT-6 localization in mig-2, rac-2, and max-2 mutants. A higher magnification view (bottom row, right) of pak-2(ok332) is shown because it has a more subtle defect than the other mutants. Results on second alleles for rac-2 and ced-10 are shown in Supplementary Fig 3. Each image is a representative image obtained from at least 2 fixation and immunostaining experiments, and imaging at least three different animals of each strain. Scale bar, 10 μm.

• 

  Fig. 8. PIX-1 levels are reduced in git-1 and pak-1 mutants. Equal quantities of total SDS-soluble proteins from wild type, git-1(ok1848), ced-10(n3246), and pak-1(ok448), were resolved on a gel, blotted to the membrane and reacted with anti-PIX-1, and as a loading control with anti-histone H3. After normalization to the amount of histone H3, the levels of PIX-1 in the mutants were compared to the level of PIX-1 in wild type. Representative immunoblot results are shown below the graph. N = 4 independent western blot reactions from each strain; means and standard error of the means are shown; git-1 and pak-1 mutants show statistically different levels from wild-type based on a two-sided student’s t-test at p ≤ 0.0001 (indicated by ****). Images of the entire western blots are provided in the Source Data Files.

• 

  Fig. 9. P190 is conserved in PIX RhoGEF domains and required for PIX-1 function at muscle cell boundaries. a PFAM alignment of RhoGEF domain sequences of PIX proteins from 10 species showing that P190 is absolutely conserved. In contrast, R212 is not conserved, which might explain why pix-1(gk713465) [R212Q] has no obvious phenotype. b Confocal microscopy of wild type and pix-1(gk893650) [P190S] mutant co-stained with antibodies to ATN-1 (α-actinin) and UNC-95 (top two rows), and co-stained with UNC-89 (obscurin) and PAT-6 (α-parvin) (bottom two rows). Note that in pix-1(gk893650), at the muscle cell boundary, the two halves of the zipper are separated, whereas in wild type they are together. c Confocal imaging of pix-1(gk893650) [P190S] stained with antibodies to UNC-52 (perlecan), PAT-3(β-integrin) or UNC-112 (kindlin). Note that these IAC proteins also show reduced or disrupted localization at muscle cell boundaries. Arrowheads bracket a muscle cell boundary. Each image is a representative image obtained from at least 2 fixation and immunostaining experiments, and imaging at least three different animals of each strain. Scale bar, 10 μm.

• 

  Fig. 10. P190S may alter RhoGEF structure and interaction with Rac, and in muscle there is a reduction in activated Rac. a Comparison of helical conformations at the start (gray) and end (cyan) of PIX-1 and PIX-1-Rac1 complex MD simulations. Wild type PIX-1 maintains the predicted kinked helical conformation near P190. (This proline is on helix F in the β-PIX structure.) Mutant PIX-1 reveals a loss of the kink in the helix by the end of the simulation. In simulations of PIX-1 complexed with GTP-bound Rac1, the kinked conformation is maintained in the mutant complex. The same result is observed with GDP-bound Rac1 (not shown). b Contact surface areas were calculated for wild type and mutant PIX-1-Rac1 complexes. In GTP-bound Rac1 complexes, the mutation leads to an increase in contact surface area. In GDP-bound Rac1 complexes, the mutation leads to a decrease in the contact surface area. c The predicted PIX-1-Rac1 complex reveals an interaction on the interface near the P190 position. This interaction between A186 of PIX-1 with L70 of Rac1 (Switch II region) is quantified as an average distance between the residues and percentage of time the residue distance is <4.5 Å over the simulation.…