View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Brief Communication 1131

Motor domain-dependent localization of myo1b (myr-1) Nanyun Tang and E. Michael Ostap

Myosin-I is the single-headed, membrane binding Results and discussion member of the superfamily that plays a role Myo1b-eGFP in membrane dynamics and transport [1–6]. Its Myo1b-eGFP (Figure 1a) has a nearly identical localiza- molecular functions and its mechanism of tion to endogenous myo1b in fixed normal rat kidney regulation are not known. In mammalian cells, epithelial cells (NRK; data not shown) and is found in myosin-I is excluded from specific the cell margins and in cytoplasmic punctae (Figure 1b,e), populations, indicating that its localization is tightly which are likely endosomes [6]. Myo1b-eGFP does not regulated. Identifying the mechanism of this concentrate on the major F--containing structures localization, and the specific actin populations with in the cell (Figure 1b–d) or on -containing which myosin-I interacts, is crucial to (Figure 1e–g). understanding the molecular functions of this motor. eGFP chimeras of myo1b [7] were imaged in live In live cells, myo1b-eGFP transiently concentrates in ar- and fixed NRK cells. Ratio-imaging microscopy eas undergoing rapid rearrangements of the actin cytoskel- shows that myo1b-eGFP concentrates within eton (Figure 2a; see Movies 1 and 2 in the Supplementary dynamic areas of the actin , most material available with this article online). Ratio images notably in membrane ruffles. Myo1b-eGFP does show that myo1b-eGFP concentrates in lamellipodia (Fig- not associate with stable actin bundles or stress ure 2a, asterisk) at a concentration that is 2- to 3-fold fibers. Truncation mutants consisting of the motor greater than in the central region of the cell. Myo1b- or tail domains show a partially overlapping eGFP is not concentrated at the leading edge of extending cytoplasmic localization with full-length myo1b, lamellipodia. but do not concentrate in membrane ruffles. A chimera consisting of the light chain and tail Myo1b-eGFP is most prominently concentrated in ruf- domains of myo1b and the motor domain from fling cell membranes (Figure 2a), i.e., lamellipodia that nonmuscle myosin-IIb (nmMIIb) concentrates on have lifted from the surface and undergone retrograde actin filaments in ruffles as well as to stress fibers. motion. Ratio images show that myo1b-eGFP is concen- In vitro motility assays show that the exclusion of trated Ͼ5-fold in ruffles over the central region of the myo1b from certain actin filament populations is due cell. The rates of retrograde movement of myo1b-eGFP- .and 40 nm/s 30ف to the regulation of the actomyosin interaction by labeled membrane ruffles range between tropomyosin. Therefore, we conclude that Upon retraction of the ruffle into the cell, myo1b-eGFP tropomyosin and spatially regulated actin localization rapidly dissipates. A membrane ruffle has polymerization play important roles in regulating the been highlighted in green to illustrate retraction and delo- function and localization of myo1b. calization of myo1b-eGFP (Figure 2a, arrow).

Address: Department of Physiology and The Pennsylvania Muscle We also find myo1b-eGFP in the perinuclear area of live Institute, University of Pennsylvania School of Medicine, B400 cells; however, distinct structures are difficult to resolve Richards, Philadelphia, Pennsylvania 19104, USA. because of background fluorescence and the motion of myo1b-eGFP within the cell during the relatively long Correspondence: E. Michael Ostap exposure time (0.5–1 s; see Supplementary material). We E-mail: [email protected] do not see a significant amount of eGFP fluorescence in the nuclei of live cells. Received: 23 March 2001 Revised: 17 May 2001 Myo1b truncation mutants Accepted: 30 May 2001 Myo1b-motor-eGFP (Figure 1a) localization (Figure 3a) is different from that of myo1b-eGFP (Figure 1b,e). Myo1b- Published: 24 July 2001 motor-eGFP is more diffusely localized in both live (Fig- ure 2b) and fixed cells (Figure 3a) than myo1b-eGFP. Current Biology 2001, 11:1131–1135 This diffuse localization is seen at all expression levels that are detectable over the background. Although myo1b-motor-eGFP appears to concentrate in membrane 0960-9822/01/$ – see front matter ruffles (Figure 2b, green panel), ratio images show that  2001 Elsevier Science Ltd. All rights reserved. this apparent concentration is due to variations in cell thickness (Figure 2b). Therefore, the tail domain is re- 1132 Current Biology Vol 11 No 14

Figure 1 Figure 2

(a) eGFP expression chimeras. The numbers indicate the amino acid position, and -G-G-G- indicates a three-glycine linker. Fluorescence The motor domain and tail domain of myo1b are required for localization micrographs of a fixed NRK cell expressing (b) myo1b-eGFP labeled to dynamic membrane ruffles. Fluorescence micrographs of live and with (c) rhodamine- phalloidin. (d) A superimposed image showing ruffling NRK cells expressing (a) myo1b-eGFP, (b) myo1b-motor- relative distributions of (green) myo1b-eGFP and (red) F-actin. (e) eGFP, (c) myo1b-IQtail-eGFP, and (d) nmMIIbM1b-eGFP. Each Myo1b-eGFP is not concentrated on (f) tropomyosin-containing panel shows fluorescence micrographs of (left) eGFP chimera , structures. Tropomyosin was visualized by indirect (right) red fluorescent protein (DsRed) used as a cell thickness immunofluorescence using an anti-tropomyosin (36/39 kDa) marker, and (bottom) time-lapse ratio images. The gray-scale monoclonal antibody. The different distributions are clearly seen in the calibration bars are shown in the first time-lapse image for each (g) superimposed image and in the insets that show magnification example and correspond to (a) 0–160 and (b–d) 0–100 intensity of the cell margin. Myo1b-eGFP is shown in green, and tropomyosin units. The time is shown in min:sec. The scale bars represent 5 ␮m. is shown in red. The scale bars represent 10 ␮m. See the text for a description of the asterisk and arrow in (a). See Supplementary material for animations and for details regarding image acquisition and ratio image normalization. quired for concentrating myo1b in lamellipodia. Myo1b- Nonmuscle myosin-IIb chimeras motor-eGFP does not concentrate on stress fibers or any NmMIIb-motor-eGFP (Figure 1a) is found on the stress other structures labeled strongly with rhodamine-phalloi- fibers of NRK cells (Figure 4a,b). Therefore, the intrinsic din (Figure 3a,b). actin binding properties of nonmuscle myosin-IIb allow the motor to interact with a subset of actin filaments not available to myo1b, and the ␣-helical coiled-coil domain Myo1b-IQtail-eGFP (Figure 1a) also diffusely localizes of myosin-II is not required for stress fiber localization. throughout the cytoplasm in live cells and does not con- We do not see nmMIIb-motor-eGFP concentrated in ruf- centrate in ruffles (Figure 2c), indicating that full-length fles in live cells (data not shown). myo1b is required for localization to the membrane ruffle. In fixed and permeabilized cells, we find that myo1b- To determine if the myo1b tail domain can specify the IQtail-eGFP and endogenous myo1b colocalize on cyto- localization of the nonmuscle myosin-IIb catalytic do- plasmic punctate structures (Figure 3e,f; inset), indicating main, we constructed a chimera consisting of the nonmus- that the tail contains information for myo1b targeting. cle myosin-IIb motor domain and the myo1b IQ and tail Myo1b-IQtail-eGFP and rhodamine-phalloidin do not domains (nmMIIbM1b-eGFP; Figure 1a). Like nmMIIb- colocalize (Figure 3c,d). motor-eGFP, nmMIIbM1b-eGFP concentrates on stress Brief Communication 1133

Figure 3 Figure 4

Fluorescence micrographs of fixed NRK cells transfected with nonmuscle myosin-IIb motor domain chimeras. (a) nmMIIb-motor- eGFP and (c) nmMIIbM1b-eGFP colocalize with (b,d) rhodamine- Fluorescence micrographs of fixed NRK cells transfected with myo1b- labeled actin in actin cables and stress fibers. (e) nmMIIbM1b-eGFP eGFP truncation mutants. (a) Myo1b-motor-eGFP is diffusely does not colocalize with (f) endogenous myo1b. The insets are distributed throughout the cytoplasm and does not colocalize with (b) magnifications to show localization details. The scale bars represent rhodamine-phalloidin-stained stress fibers or actin cables. (c) 10 ␮M. Myo1b-IQtail-eGFP is distributed throughout the cytoplasm and does not colocalize with (d) rhodamine-labeled actin cables or stress fibers. (e) Myo1b-IQtail-eGFP colocalizes with (f) endogenous myo1b on punctate structures within the cytoplasm. Endogenous myo1b trating the protein to the dynamic actin filaments at the was visualized with a myo1b polyclonal antibody that binds the motor cell periphery. In fixed cells, nmMIIbM1b-eGFP appears domain. The circled regions in the insets highlight regions of myo1b- to be more concentrated at the cell periphery than endoge- IQtail-eGFP and endogenous myo1b colocalization. The scale bars nous myo1b (Figure 4e,f). With the exception of stress ␮ represent 10 m. fiber labeling, the cytoplasmic distribution of nmMIIbM1b- eGFP is more diffuse than myo1b-eGFP. We do not find nmMIIbM1b-eGFP colocalized with endogenous myo1b fibers (Figure 4c,d), indicating that the motor domain on distinct cytoplasmic structures (Figure 4e,f; inset) as plays a defining role in the localization of the protein. seen with myo1b-IQtail-eGFP (Figure 3e,f; inset). Addi- The localization of nmMIIb-motor-eGFP ranges from a tionally, there is no clear alignment of endogenous myo1b very strong stress fiber localization (Figure 4c) to a more with stress fibers in cells expressing nmMIIbM1b-eGFP, diffuse cytoplasmic localization in which the stress fibers as would be expected if nmMIIbM1b-eGFP caused mis- are less apparent (Figure 4e). This variation in stress fiber localization of the cytoplasmic compartments that bind labeling by nmMIIbM1b-eGFP correlates with the size myo1b. However, since nmMIIbM1b-eGFP appears to of actin cables and stress fibers in the cell, as assessed by interact with multiple microfilament populations (i.e., not rhodamine-phalloidin labeling (data not shown). just the stress fibers), such an alignment would be difficult to detect. Therefore, although the myo1b tail domain We detect nmMIIbM1b-eGFP in lamellipodia (Figure helps concentrate the myosin-IIb motor domain to the 2d), confirming that the tail domain plays a role in concen- ruffles, the myosin-IIb motor domain causes a general 1134 Current Biology Vol 11 No 14 mislocalization of the tail domain. This mislocalization is Myo1b-eGFP is not found on most F-actin-containing most likely due to the ability of the nmMIIb motor domain structures (Figure 1b–d). The data clearly show that it is to bind microfilament populations not normally accessible not the binding of the myosin-I tail domain to membranes to myo1b. or other receptors that keeps the off of these structures, since myo1b-motor-eGFP (Figure 3a,b) does In vitro motility assays not concentrate on stress fibers, while nmMIIb-motor- We used the sliding filament assay to examine the effect eGFP and nmMIIbM1b-eGFP do bind to these structures of nonmuscle tropomyosin isoform, TM2 [8], on the abil- (Figure 4). Rather, the exclusion of myo1b from these ity of myo1b-eGFP to translocate actin filaments. Myo1b- stable actin structures is most likely due to the direct eGFP was expressed in 293T cells, partially purified by regulation of the actomyosin interaction [16]. Stable actin cation-exchange chromatography, and bound to anti-GFP structures are known to contain tropomyosin, while the antibodies that were adsorbed to nitrocellulose-coated highly dynamic microfilament compartments do not [17, motility chambers. In the absence of TM2, myo1b-eGFP 18]. It has been demonstrated that the effect of tropomyo- supports actin gliding motility at a rate of 31 Ϯ 8 nm/s at sin on the actomyosin interaction is dependent on the 37Њ (n ϭ 44). This rate is similar to that observed for myosin isoform [19]. For example, tropomyosin does not tissue-purified myo1b [9]. The presence of 1.5 ␮M TM2 inhibit the rate at which smooth muscle myosin-II and completely inhibited actin filament motility. Few fila- nonmuscle myosin-II propel actin filaments in the in vitro ments were bound to the coverslip in the presence of motility assay [20, 21], while nonmuscle tropomyosin com- TM2, indicating that myo1b-eGFP is not able to interact pletely inhibits brush border myosin-I [22]. We show that with regulated actin. myo1b-eGFP is concentrated in regions of high actin dy- namics, but is not concentrated on adjacent tropomyosin- containing filaments (Figure 1e–g, inset), and in vitro Mechanisms of myo1b localization motility assays show that myo1b-eGFP does not interact The actin cytoskeleton is very dynamic in lamellipodia, with actin and nonmuscle tropomyosin. Therefore, we where membrane extension is driven by actin polymeriza- propose that tropomyosin is a primary determinant of the tion [10]. We find myo1b-eGFP associated with lamelli- microfilament population on which myo1b can interact. podia, with the highest concentration in ruffling mem- branes (Figure 2a). The increased fluorescence intensity in ruffling lamellipodia is likely due to the concentration We offer the following simple model: First, the myo1b of the actin cytoskeleton via myosin-based contraction. tail domain restricts the diffusion of the motor by binding The rapid loss of myosin-I localization following retraction to acidic phospholipid domains [23] and possibly to un- likely reflects the disassembly of the actin cytoskeleton. known myosin-I receptors on intracellular membranes. However, the regulation of membrane binding by myosin-I Second, myo1b is not able to catalytically interact with via the tail domain is also possible [11]. While the precise the stable, regulated microfilament population. Third, role of myosin-I in retracting membranes is not known, upon the spatially regulated polymerization of actin [10], the localization of myo1b-eGFP and the similar rates of membrane-bound myosin-I rapidly binds to the newly membrane retraction (30–40 nm/s) and myo1b motility formed tropomyosin-free filaments and performs its un- (31 nm/s) are consistent with myosin-I playing a role in the known force-generating role, possibly retracting the actin retraction of the newly polymerized actin in the extended cytoskeleton [2] or maintaining cortical tension [15]. lamellipodia [2]. Fourth, disassembly of the actin cytoskeleton results in the delocalization of myosin-I. The inability of myo1b-motor-eGFP, myo1b-IQtail- eGFP, and nmMIIb-motor-eGFP to concentrate in lamel- Although this model is an over simplification (we do not lipodia indicates that both the motor and tail domains are consider how myo1b is anchored at the membrane or the essential for proper cellular localization [12–14]. Addition- poorly understood role of myo1b calcium regulation), its ally, our finding that myo1b-IQtail-eGFP does not con- application to describe the role of myo1b on the plasma centrate in the ruffles suggests that there are not specific membrane is straightforward, e.g., plasma membrane- myo1b receptors responsible for targeting the motor to bound myo1b rapidly concentrates in lamellipodia as actin lamellipodia. However, we can not rule out the presence is polymerized and contracts the cytoskeleton, resulting of myo1b receptors on the intracellular membranes (Fig- in membrane retraction. The proposed role of myosin-I ure 3e,f). Because myo1b is a low-duty ratio motor (i.e., in the localization of actin-nucleation machinery at the it spends a small fraction of its ATPase cycle strongly plasma membrane [24, 25] fits nicely with our finding bound to actin [15]), we propose that the tail domain that myosin-I interacts with the dynamic actin filament is needed to spatially restrict myo1b to the membrane, populations. A similar role for myo1b on intracellular allowing for the rapid recruitment of myo1b to regions of membranes is also expected, since proteins known to be membrane-based actin polymerization while preventing essential for the elongation of actin filaments may also myo1b from diffusing away from the actin. play a role in intracellular membrane trafficking [26]. Brief Communication 1135

Therefore, we predict that endosome-bound myo1b [6, populations of microfilaments in neurites and growth cones. Mol Cell Neuroscience 1997, 8:439-454. 27] interacts with newly polymerized actin filaments, re- 18. Gunning P, Hardeman E, Jeffrey P, Weinberger R: Creating sulting in membrane budding, transport, or changes in intracellular structural domains: spatial segregation of actin tropomyosin isoforms in neurons. Bioessays 1998, 20:892-900. endosome morphology. 19. Collins K, Matsudaira P: Differential regulation of vertebrate I and II. J Cell Sci 1991, 14:11-16. Supplementary material 20. Pato MD, Sellers JD, Preston YA, Harvey EV, Adelstein RS: Supplementary material including Materials and methods and movies of Baculovirus expression of chicken nonmuscle heavy the live cells shown in Figure 2 is available at http://images.cellpress.com/ meromyosin II-B. Characterization of alternatively spliced isoforms. J Biol Chem 1996, 271:2689-2695. supmat/supmatin.htm. 21. Umemoto S, Bengur AR, Sellers JR: Effect of multiple phosphorylations of smooth muscle and cytoplasmic myosins Acknowledgements on movement in an in vitro motility assay. J Biol Chem 1989, We thank J. Poole for technical assistance; H.L. Sweeney, E. De La Cruz, 264:1431-1436. S. Zigmond, and J. Sanger for helpful discussions; and S. Hitchcock-Degre- 22. Fanning AS, Wolenski JS, Mooseker MS, Izant JG: Differential gori for nonmuscle tropomyosin. We also thank A. Wells and M. Quinlan for regulation of skeletal muscle myosin-II and brush border advice on motility assays. This work was supported by grants to E.M.O. myosin-I enzymology and mechanochemistry by bacterially from the National Institutes of Health (GM57247) and the American Heart produced tropomyosin isoforms. Cell Motil Cytoskeleton Association (9730177N). 1994, 29:29-45. 23. Doberstein SK, Pollard TD: Localization and specificity of the phospholipid and actin binding sites on the tail of References Acanthamoeba myosin IC. J Cell Biol 1992, 117:1241-1249. 1. Sokac AM, Bement WM: Regulation and expression of 24. Evangelista M, Klebl BM, Tong AHY, Webb BA, Leeuw T, Leberer metazoan unconventional myosins. Int Rev Cyt 2000, E, et al.: A role of myosin I in actin assembly through 200:197-304. interactions with Vrp1p, Bee1p, and the Arp2/3 complex. J 2. Novak KD, Peterson MD, Reedy MC, Titus MA: Dictyostelium Cell Biol 2000, 148:353-362. myosin I double mutants exhibit conditional defects in 25. Lechler T, Shevchenko A, Shevchenko A, Li R: Direct involvement pinocytosis. J Cell Biol 1995, 131:1205-1221. of yeast type I myosins in Cdc42-dependent actin 3. Geli MI, Riezman H: Role of type I myosins in receptor-mediated polymerization. J Cell Biol 2000, 148:363-373. endocytosis in yeast. Science 1996, 272:533-535. 26. Rozelle AL, Machesky LM, Yamamoto M, Driessens MH, Insall RH, 4. Sto¨ ffler HE, Ruppert C, Reinhard J, Ba¨hler M: A novel mammalian Roth MG, et al.: Phosphatidylinositol 4,5-bisphosphate myosin I from rat with an SH3 domain localizes to Con induces actin-based movement of raft-enriched vesicles A-inducible, F-actin-rich structures at cell-cell contacts. through WASP-Arp2/3. Curr Biol 2000, 10:311-320. J Cell Biol 1995, 129:819-830. 27. Balish MF, Moeller EF, Coluccio LM: Overlapping distribution of 5. Jung G, Wu X, Hammer JA III: Dictyostelium mutants lacking the 130- and 110-kDa myosin I isoforms on rat liver multiple classic myosin I isoforms reveal combinations of membranes. Arch Biochem Biophys 1999, 370:285-293. shared and distinct functions. J Cell Biol 1996, 133:305-323. 6. Raposo G, Cordonnier MN, Tenza D, Menichi B, Durrbach A, Louvard D, et al.: Association of myosin I alpha with endosomes and lysosomes in mammalian cells. Mol Biol Cell 1999, 10:1477- 1494. 7. Ruppert C, Kroschewski R, Ba¨hler M: Identification, characterization and cloning of myr1, a mammalian myosin-I. J Cell Biol 1993, 120:1393-1403. 8. Hammell RL, Hitchcock-DeGregori SE: Mapping the functional domains within the carboxyl terminus of ␣-tropomyosin encoded by the alternative spliced ninth exon. J Biol Chem 1996, 271:4236-4242. 9. Williams R, Coluccio LM: Novel 130 kDa rat liver myosin-1 will translocate actin filaments. Cell Motil Cytoskeleton 1994, 27:41-48. 10. Pollard TD, Blanchoin L, Mullins RD: Molecular mechanisms controlling actin filament dynamics in nonmuscle cells. Annu Rev Biomol Struct 2000, 29:545-576. 11. Swanljung-Collins H, Collins JH: Phosphorylation of brush border myosin I by protein kinase C is regulated by Ca2؉-stimulated binding of myosin I to phosphatidylserine concerted with calmodulin dissociation. J Biol Chem 1992, 267:3445-3454. 12. Sto¨ ffler HE, Honnert U, Bauer CA, Hofer D, Schwarz H, Muller RT, et al.: Targeting of the myosin I myr3 to intracellular adherens type junctions induced by dominant actin Cdc42 in HeLa cells. J Cell Sci 1998, 111:2779-2788. 13. Ruppert C, Godel J, Muller RT, Kroschewski R, Reinhard J, Ba¨hler M: Localization of the rat myosin I molecules myr1 and myr2 and in vivo targeting of their tail domains. J Cell Sci 1995, 108:3775-3786. 14. Novak KD, Titus MA: The myosin-I SH3 domain and TEDS rule phosphorylation site are required for in vivo function. Mol Biol Cell 1998, 9:75-88. 15. Coluccio LM, Geeves MA: Transient kinetic analysis of the 130- kDa myosin I (MYR-1 product) from rat liver. A myosin I designed for maintenance of tension? J Biol Chem 1999, 274:21575-21580. 16. Pelham RJ Jr, Lin JJC, Wang YL: A high molecular mass non- muscle tropomyosin stimulates retrograde organelle transport. J Cell Sci 1996, 109:981-989. 17. Schevzov G, Gunning P, Jeffrey PL, Temm-Grove C, Helfman DM, Lin JJC, et al.: Tropomyosin localization reveals distinct