INF1 is a novel microtubule-associated formin - PubMed (original) (raw)
INF1 is a novel microtubule-associated formin
Kevin G Young et al. Mol Biol Cell. 2008 Dec.
Abstract
Formin proteins, characterized by the presence of conserved formin homology (FH) domains, play important roles in cytoskeletal regulation via their abilities to nucleate actin filament formation and to interact with multiple other proteins involved in cytoskeletal regulation. The C-terminal FH2 domain of formins is key for actin filament interactions and has been implicated in playing a role in interactions with microtubules. Inverted formin 1 (INF1) is unusual among the formin family in having the conserved FH1 and FH2 domains in its N-terminal half, with its C-terminal half being composed of a unique polypeptide sequence. In this study, we have examined a potential role for INF1 in regulating microtubule structure. INF1 associates discretely with microtubules, and this association is dependent on a novel C-terminal microtubule-binding domain. INF1 expressed in fibroblast cells induced actin stress fiber formation, coalignment of microtubules with actin filaments, and the formation of bundled, acetylated microtubules. Endogenous INF1 showed an association with acetylated microtubules, and knockdown of INF1 resulted in decreased levels of acetylated microtubules. Our data suggests a role for INF1 in microtubule modification and potentially in coordinating microtubule and F-actin structure.
Figures
Figure 1.
INF proteins in animals. (A) An unrooted phylogenetic tree based on a comparison of the FH2 domains of INF1, INF2, and other animal formins. Sequences used for this analysis (with GenBank accession numbers in brackets) came from human (H. sapiens) INF1 (NP_203751.2) and INF2 (NP_071934.3); monkey (M. mulatta) INF1 (XP_001085372.1); mouse (M. musculus) INF1 (NP_001028473.1), INF2 (ABI20145.1), mDia1 (O08808.1), mDia3 (AAH86779.1), DAAM1 (AAH76585.1), delphilin (NP_579933.1), and FHOD1 (NP_808367.1); rat (R. norvegicus) INF1 (NP_001099907.1); platypus (Ornithorhynchus anatinus) INF1 (XP_001511976.1); toad (X. tropicalis or X. laevis) INF1 (CR942785.2), INF2 (NP_001072591.1), and FHOD1 (AAH84291.1); zebrafish (D. rerio) INF1 (predicted from BAC clone CH211-62K15), DAAM1 (XP_707353.2), and delphilin (XP_689509.2); pufferfish (Tetraodon nigroviridis) INF2 (CAG10691.1); sea squirt (Ciona intestinalis) INF1 (translated from AK173884.1); fruit fly (D. melanogaster or D. pseudoobscura) INFX (XP_001353341.1), diaphanous (NP_476981.1), DAAM (AAF45601.2), cappuccino (NP_722951.1), and FHOD (NP_729410.1); sea urchin (S. purpuratus) INFX (XP_793426.2 and XP_785094.2); wasp (Nasonia vitripennis) INFX (XP_001600053.1); honey bee (A. mellifera) INFX (translated from XR_015075.1); flour beetle (Tribolium castaneum). INFX (XP_970252.1); mosquito (A. aegypti) INFX (XP_001660600.1); and nematode worm (C. elegans and C. briggsae) inft-1 (NP_497334.1 and XP_001666551.1, respectively). The novel INFX group contains FH2 domains most similar to both INF1 and INF2, as determined by both BLAST and ClustalW2 analysis. Branch lengths correspond to evolutionary distances. (B) Domain structures of representative animal formins. Delphilin, DAAM1, and diaphanous, as with most known formins, contain C-terminal FH1/FH2 regions. These formins also contain regulatory domains in the N-terminal half. Shown are the PDZ domain of delphilin, and the GTPase-binding domain (GBD) and diaphanous-related formin domain (DRF; diaphanous inhibitory domain region) of DAAM1 and diaphanous. INF2 and INFX proteins share the DRF domain in the N-terminal part of the protein. Each of the INF proteins has a C-terminal region downstream from the FH2 domain that is more extensive than in other formins. In INF1, and in the nematode worm inft-1 sequences, the region upstream from the FH1 domains is considerably shorter relative to other formins, thus placing the FH1/FH2 region in the N-terminal half. There are no obvious regulatory regions upstream of the INF1 FH1 region in the N-terminus. At the C-terminal end of INF1, we have mapped a novel microtubule-binding domain (MTBD), in this study. (C) A ClustalW2 alignment of part of the novel C-terminal MTBD of INF1 showing two well conserved regions, which we have labeled as MTB1 and MTB2. Asterisks denote identical amino acid residues in each sequence, two dots denote conserved amino acids, and one dot denotes semiconserved amino acids.
Figure 2.
INF1 induces stress fiber formation and SRF activation in NIH 3T3 cells. (A–D) Full-length myc-INF1FL fusion protein induced the formation of actin filaments (B) in serum-starved 3T3 cells after 1 d of expression. (C) Microtubule staining in the transfected cells coaligned with the INF1 fusion protein. Coaligning INF1 and microtubules are indicated by arrows in the insets, which are enlarged from the boxed area of the transfected cell shown. Actin filaments also coaligned to a more limited extent (indicated by the arrow in the inset in B). The merged image is shown in D. (E) Schematics of the myc-INF1FL fusion protein, myc-INF1 truncation mutants, and myc-mDia1 full-length (FL) and FH2 domain fusion proteins used to examine stress fiber formation, along with quantifications of starved 3T3 cells expressing these proteins that displayed increased actin filament formation. Cells were scored as having an increase in thick filaments if the filaments displayed markedly brighter phalloidin labeling compared with untransfected cells (as in B) and as having an increase in thin filaments if there were more actin filaments compared with the untransfected cells, but with a similar intensity of phalloidin labeling. (F) Full-length (myc-INF1FL) and N-terminal (myc-Nterm and myc-FH2) INF1 fusion proteins induced SRF activation after 1 d of expression in serum-starved 3T3 cells. Plasmid, at 0.1, 0.3, and 1 μg, was transfected for each INF1 plasmid. Empty vector (EV) served as a negative control. Expression of a C-terminal INF1 fusion (myc-Cterm) did not induce SRF activation. Results are from three independent experiments, with SE of the mean shown. Scale bar, 20 μm in A and also the larger images in B–D.
Figure 3.
Endogenous INF1 protein is primarily associated with microtubules. (A) Full-length INF1FL-YFP expressed in Cos-1 cells was specifically detected using an affinity-purified INF1 pAb by immunofluorescence. Staining was not observed in untransfected cells. (B) Detection of endogenous INF1 by immunoblotting NIH 3T3 cell lysates. A band of ∼125 kDa was consistently the predominant band detected. Detection of this band was decreased by an average of 51% (three separate trials) using 1 nM INF1 siRNA duplex when compared with cells transfected with 1 nM control scrambled siRNA, in lysates collected 7 d after transfection. Mock-transfected cells displayed levels of INF1 similar to that of the control transfected cells. Lamin B is shown as a loading control. Lysates were collected with lysis buffer containing a high salt concentration and protease inhibitors, as described in Materials and Methods, though similar results were obtained with a standard SDS buffer in this case (data not shown). (C) Immunoblotting with INF1 pAb using protein lysates from Cos-1 cells expressing myc-INF1FL or GFP-mINF1, two neuronal cell lines (N2A and F11), NIH 3T3 fibroblast cells, a cardiomyoblast cell line (H9C2), and HeLa cells. Endogenous protein running at a molecular weight similar to myc-INF1 (∼160 kDa; marked with an arrow) was detected in undifferentiated and differentiated N2A cells (N2A and N2A diff, respectively), undifferentiated F11 cells, NIH 3T3 cells, and H9C2 cells, but not in HeLa or Cos-1 cells (note no band of this size in the GFP-mINF1 lane, with this fusion protein running higher because of the additional size of the GFP). In the N2A, F11, NIH 3T3, and H9C2 lysates, the lower molecular weight band of ∼125 kDa was detected much more prominently, and was overexposed in order to detect the higher molecular weight band. (D) INF1 localized in a filamentous pattern in NIH 3T3 cells. This pattern coaligned with a subset of microtubules (E). (F) F-actin staining was not consistently colocalized with the INF1 staining. (G) Merged image of D–G. The inset in G shows INF1 merged with microtubules alone from this cell, at a higher magnification. Scale bar, (F) 20 μm.
Figure 4.
INF1 localization is microtubule-dependent in NIH 3T3 cells. (A–D) Control cells treated with a 1:500 DMSO dilution, stained for INF1 (A), microtubules (B), and F-actin (C). INF1 localized in a filamentous pattern along microtubules in these cells. The merged image is shown in D, with enlargement from the boxed area in the inset. (E–H) Cells treated with the microtubule depolymerizing drug nocodazole (500 nM, 15 min) had dispersed microtubules (F) and dispersed INF1 staining (E). F-actin staining is shown in G, and the merged image in H, with enlargement from the boxed area in the inset. (I–L) Cells treated with the actin depolymerizing drug latrunculin A (1 μm, 15 min) had dispersed actin filaments (K) but INF1 (I) remained localized along the microtubules (J). The merged image is shown in L, with enlargement from the boxed area in the inset. Scale bar, 20 μm and applies to all images.
Figure 5.
Expression of INF1 in mouse tissues. (A) Immunoblot of mouse tissues showing expression of INF1 protein (arrow) in brain, heart, and lung. (B) By immunofluorescence analysis, INF1 was readily detectable in ventricular muscle of the heart. (C) At an identical exposure as in B, an immunofluorescence signal was barely detectable in kidney, corresponding to immunoblotting results shown in A. (D) INF1 (green) staining in a 5-d-old (P5) mouse cerebellum coronal section. The section was counterstained for F-actin (red) and chromatin (blue). INF1 staining in Purkinje neuron cell bodies is indicated by the arrowheads. (E) Sagittal brain section showing INF1 (green) staining with F-actin (red) labeling in the cerebellum of a 2-mo-old (adult) mouse. Arrows point to the white matter layer, and arrowheads to the Purkinje neuron cell bodies. (F–H) Higher magnification confocal images showing INF1 (F) with tuj1 (βIII-tubulin; G) in axons. The merged image is shown in E. Although INF1 was consistently associated with tuj1 staining, they did not necessarily colocalize. Scale bar, (D and E) 20 μm, (F–H) 10 μm.
Figure 6.
The C-terminus of INF1 is required for INF1 microtubule association. (A–C) Full-length INF1FL-YFP fusion protein expressed in Cos-1 cells. The fusion protein (A) primarily coaligned with microtubules (B) in 94% of the cells (n = 148 cells, two trials). The merged image is shown in C. Insets show an enlargement of the boxed region. (D–F) The truncated INF1ΔC1-YFP protein (D) was coaligned with microtubules in 38% of the cells expressing this fusion (n = 138 cells, two trials; E), though it was more commonly diffusely localized or was localized to puncta. The INF1ΔC2-YFP protein (G) was almost completely diffusely localized or localized to puncta and showed only a very limited association with microtubules (H). Microtubule coalignment was observed in just 2% of the INF1ΔC2-YFP–transfected cells (n = 169 cells, two trials). The merged image is shown in I. Insets, an enlarged image of the boxed region. (J–L) The C-terminal GFP-INF1C protein (J) localized with bundled microtubules (K) in the cytoplasm. This protein also localized to nuclei. The merged image is shown in L. Scale bar, 20 μm in L, and applies to all images.
Figure 7.
The INF1 C-terminus interacts with and bundles microtubules directly. Cos-1 cells transfected with GFP-INF1C showed the fusion protein (A) associated with bundled microtubules (B). Microtubule coalignment was observed in 62% of the transfected cells (n = 140 cells), with 33% of the cells displaying thicker, bundled microtubules. When the MTB1 region was removed (GFP-ΔMTB1), the fusion protein was primarily diffusely localized (C). Limited microtubule coalignment was observed in 15% of the GFP-ΔMTB1 transfected cells (n = 104 cells). Microtubules are shown in D. When the MTB2 region was removed (GFP-ΔMTB2), the fusion protein was again primarily diffusely localized (E). Limited microtubule coalignment was observed in 26% of the GFP-ΔMTB2 transfected cells (n = 98 cells). Microtubules are shown in F. (G) Purified INF1C protein was incubated with microtubules and centrifuged at 100,000 × g. The INF1C protein was found mostly in the pellet (P) fraction when incubated with microtubules, but remained in the supernatant (S) when incubated alone. GST incubated with or without microtubules remained in the supernatant. (H) Purified INF1C protein incubated with microtubules, but not GST (I), induced microtubule bundling, as observed by immunofluorescence microscopy. Scale bar, 20 μm in F and applies to all images of the cells.
Figure 8.
The INF1 C-terminus confers resistance against nocodazole-induced microtubule depolymerization. (A and B) Cos-1 cells expressing INF1FL-YFP were treated with 1 or 10 μM nocodazole for 1 h and then fixed and labeled with α-tubulin antibody. Cells expressing the full-length INF1 fusion still retained intact microtubules, though they were very limited at the 10 μM nocodazole concentration. Microtubules were scored as being extensive if there was more filamentous labeling than diffuse labeling within a cell and limited if the labeling was more diffuse than filamentous. (C and D) In cells expressing the C-terminal truncated INF1ΔC2-YFP fusion protein, very few cells had any filamentous tubulin labeling at either nocodazole concentration. In cells expressing the C-terminal fusion protein, GFP-INF1C, many cells were observed to have filamentous labeling, with this being primarily extensive at the lower nocodazole concentration and more limited at the higher concentration. (G) Quantification of the microtubule phenotypes from cells expressing INF1FL-YFP (n = 164 cells for 1 μM and 215 cells for 10 μM nocodazole), INF1ΔC2-YFP (n = 200 cells for 1 μM and 171 cells for 10 μM nocodazole), GFP-INF1C (n = 360 cells for 1 μM and 305 cells for 10 μM nocodazole), and GFP (n = 292 cells for 1 μM and 273 cells for 10 μM nocodazole) alone, observed from three separate trials. Error bars, SE of the mean. Scale bar, 20 μm in F and applies to all images.
Figure 9.
INF1 induces acetylated microtubule formation. Expression of INF1FL-YFP (B) in NIH 3T3 cells for 1 d induced an increase in labeling for acetylated microtubules (A) in 44% of the cells. Cells were counted as having increased labeling only when they showed a clearly stronger signal relative to all untransfected cells in the same field. (C) Acetylated microtubule formation was also induced in 28% of the cells expressing the INF1ΔC2-YFP fusion protein (D). Highly bundled acetylated microtubules (E) formed with the expression of the GFP-INF1C fusion protein (F) in 69% of the transfected cells. Expression of a Flag-FH2 protein (H) did not result in increased acetylated microtubule staining (G). Fusion protein expressing cells are marked with an asterisk in A, C, E, and G. The numbers of cells expressing each of these fusion proteins that displayed increased acetylated microtubule staining relative to the surrounding untransfected cells is quantified in I, along with data from cells transfected with GFP alone. Data are from four separate trials for the INF1FL-YFP (n = 189 cells total), INF1ΔC2-YFP (n = 172 cells total) and GFP-INF1C (n = 207 cells total) groups, and two trials for the Flag-FH2 (n = 166 cells total) or GFP (n = 176 cells total) groups, with SE of the mean shown. (J) The knockdown of INF1 using siRNA in NIH 3T3 cells resulted in a consistent reduction of acetylated tubulin levels. This did not involve a reduction in α-tubulin levels. Lamin B is also shown as a loading control. Scale bar, (A–H) 20 μm; (K and L) 20 μm.
References
- Chhabra E. S., Higgs H. N. INF2 Is a WASP homology 2 motif-containing formin that severs actin filaments and accelerates both polymerization and depolymerization. J. Biol. Chem. 2006;36:26754–26767. - PubMed
- Cooper G. T. Cytoskeletal networks and the regulation of cardiac contractility: microtubules, hypertrophy, and cardiac dysfunction. Am. J. Physiol. Heart Circ. Physiol. 2006;291:H1003–H1014. - PubMed
- Copeland J. W., Copeland S. J., Treisman R. Homo-oligomerization is essential for F-actin assembly by the formin family FH2 domain. J. Biol. Chem. 2004;48:50250–50256. - PubMed
Publication types
MeSH terms
Substances
LinkOut - more resources
Full Text Sources
Molecular Biology Databases
Miscellaneous