The Tail Domain of Myosin M Catalyses Nucleotide Exchange on ...

Download PDF
6 downloads 4746 Views 585KB Size Report
Comment
motor domain and a guanine nucleotide exchange factor. (GEF) domain for ...... confocal files and Dr Bernd Helms for hosting and help during the nucleotide ...
Traffic 2000 1: 399–410 Munksgaard International Publishers

The Tail Domain of Myosin M Catalyses Nucleotide Exchange on Rac1 GTPases and Can Induce Actin-Driven Surface Protrusions

Heidrun Geissler, Ronald Ullmann and Thierry Soldati* Department of Molecular Cell Research, Max -Planck -Institute for Medical Research, Jahnstrasse 29, D -69120 Heidelberg, Germany * Corresponding author: T. Soldati, [email protected] -heidelberg.mpg.de

Members of the myosin superfamily play crucial roles in cellular processes including management of the cortical cytoskeleton, organelle transport and signal transduction. GTPases of the Rho family act as key control elements in the reorganization of the actin cytoskeleton in response to growth factors, and other functions such as membrane trafficking, transcriptional regulation, growth control and development. Here, we describe a novel unconventional myosin from Dictyostelium discoideum, MyoM. Primary sequence analysis revealed that it has the appearance of a natural chimera between a myosin motor domain and a guanine nucleotide exchange factor (GEF) domain for Rho GTPases. The functionality of both domains was established. Binding of the motor domain to F-actin was ATP-dependent and potentially regulated by phosphorylation. The GEF domain displayed selective activity on Rac1-related GTPases. Overexpression, rather than absence of MyoM, affected the cell morphology and viability. Particularly in response to hypo-osmotic stress, cells overexpressing the MyoM tail domain extended massive actin-driven protrusions. The GEF was enriched at the tip of growing protuberances, probably through its pleckstrin homology domain. MyoM is the first unconventional myosin containing an active Rac-GEF domain, suggesting a role at the interface of Rac-mediated signal transduction and remodeling of the actin cytoskeleton. Key words: Actin cortex, Dictyostelium discoideum, guanine nucleotide exchange factor, myosin, Rac GTPase Received 12 December 1999, revised and accepted for publication 31 January 2000

The actin cytoskeleton is essential for the maintenance of cell shape and locomotion, but also provides tracks for active intracellular transport. Many actin-based processes, such as motility, cytokinesis, phagocytosis, endocytosis, polarized secretion and exocytosis, organelle movement and mRNA transport (1) have been shown to involve myosins. These motor proteins form a superfamily currently encompassing 15 phylogenetic and structural classes (1–3). Despite adapta-

tion to diverse tasks in the cell, all myosins share a tripartite modular structure: the head or motor domain, site of the actin-activated ATPase activity; the neck acting as lever arm and binding site for light chains; and a class-specific tail reflecting functionality by its domain composition. Recent data indicate that myosins also play an active role in signal transduction (4,5). The tail domain of various myosins contains elements including Src homology 3 (SH3) and pleckstrin homology (PH) domains thought to mediate protein– protein or protein – lipid interactions within signaling cascades. Particularly, the tail of class IX myosins harbors a GTPase activating protein (GAP) domain specific for Rho (6), a member of the family of small GTPases that play key roles in growth factor-regulated cytoskeletal reorganization and cell cycle progression (7 – 9). Mammalian cells overexpressing myr5, a rat myosin IX, lose stress fibers and focal contacts and consequently round up (6). This phenotype correlates well with the effects induced by inactivation of Rho (6,10). The presence of a GAP domain raises the question whether and how this activity is functionally coupled to the motor activity. To be completed, beside GTP hydrolysis the GTPase cycle requires the exchange of nucleotide to proceed from the inactive, GDP-bound conformation to the active, GTP-bound form. This step is controlled by guanine nucleotide exchange factors (GEFs) forming a family of proteins related to Diffuse B-cell Lymphoma (Dbl), a human proto-oncogene (11). The GEFs share 250 residues long Dbl homology (DH) domain, which is the site of catalytic activity and transforming potential, followed by a PH domain. A natural chimera combining a myosin motor domain with a GEF domain has not yet been identified in higher eukaryotes. We recently reported the identification of two novel divergent unconventional myosins in Dictyostelium discoideum (12). As a professional phagocyte and a very active and motile amoeba, this unicellular eukaryote is a model organism well-suited to the investigation of actin-dependent processes. The present study describes the cloning of the complete myoM sequence and analysis of primary structure, revealing the first myosin with a DH domain in its tail. Both moieties of this natural chimeric protein were shown to be functional in vitro. Investigations of its role in vivo suggest that the tail domain of MyoM induces the extension of ramopodia, massive actin-driven plasma membrane protrusions, in response to external stimuli such as hypo-osmotic stress.

399

Geissler et al.

Results MyoM is the founding member of a novel myosin class MyoM was initially identified in a PCR screen for novel unconventional myosins in D. discoideum and found to represent an unknown myosin locus on chromosome VI (12). Phylogenetic analysis does not assign MyoM to any of the 15 myosin classes currently known. On average, pairwise sequence comparisons of motor domains indicated values around 30% identity with MyoM, for example D. discoideum mhcA (class II) and Drosophila melanogaster 95F (class VI). Similarity is highest to B. taurus myosin X (36%) and H. sapiens myosin IX (32%), whereas class I myosins appear less related to MyoM (27% identity with D. discoideum MyoB).

A distinguishing feature of the MyoM head domain are two insertions that in length and position, but not in sequence, are reminiscent of those found in myosins of class VI and IX, respectively (13) (Figure 1a). The shorter one, of  15 residues, occurs at the same site as the 21/22 residue insertion of myosin VI. The larger extension of  90 residues is located at the start of the flexible loop at the 50/20 kDa junction of the head domain, a region in contact with actin. In myosins IX, a 110 residues insertion is found precisely at this position. At a site in the head domain designated by the TEDS rule (14), embedded in a recognition sequence for Rho-dependent kinases of the PAK/Ste20 family, MyoM carries a serine residue (Figure 1a,b). Phosphorylation of this residue is known to be crucial for function in amoeboid Class I myosins (reviewed in 15).

Figure 1: Primary structure analysis of MyoM reveals a novel myosin isoform with a Dbl-like GEF domain. (a) Protein sequence. MyoM (1737 aa, accession number AF090533) contains two insertions in the head domain (light gray shading) and a phosphorylatable Ser residue (*) at the TEDS rule site (residues of the conserved DALAK motif and the myosin I heavy chain kinase recognition sequence are highlighted in light gray). IQ motifs are boxed. The coiled-coil domain is shaded in black (prediction of 1.0 for 107 aa with a window of 28 residues was made by the COIL algorithm (50)). The first half of the tail domain is distinguished by the preponderance of Ser, Thr and Pro residues (together  50%). The C-terminal Dbl-homology (DH) and pleckstrin homology (PH) domains are shaded in dark gray. (b) Domain structure. Color coding is the same as in (a). L1, L2: flexible loops at the 25/50 kDa and 50/20 kDa junction, respectively. P: TEDS rule phosphorylation site. (c) Schematic presentation of the constructs generated in this study. H8: octa-His tag. GST: glutathione S-transferase tag.

400

Traffic 2000: 1: 399 – 410

MyoM acts as a Rac1-GEF

Figure 2: Conserved residues in Dbl-homology domains cluster in two regions. DH domains of MyoM and the seven most closely related Dbl family proteins were aligned with the ClustalW program. Conservation of residues is indicated by shading (light: \ 70% conservation; dark: 100% conservation). Identity of DH domains (accession numbers) in pairwise comparisons with MyoM (AF090533) are as follows: Hs Tiam1 (Q13009), 35.3%; Hs Vav2 (P52735), 28.5%; Hs Y142 (Q14155), 29.1%; Sp SCD1 (P40995), 25.2%; Sc Cdc24 (P11433), 30.9%; Ce YL14 (Q11100), 32.2%; Hs Dbl (P10911), 29.9%.

The neck domain of MyoM contains two classical IQ motifs [IQxxxRGxxxR (16)]. This domain is followed by a stretch of about 150 residues with a high propensity to form a coiledcoil (score of 1.0 for 107 amino acids [aa], as predicted by the COIL algorithm, Figure 1). MyoM is therefore likely to occur as a dimer. Although no significant overall sequence homology was found for the region encompassing the 400 residues C-terminal to the coiled-coil domain, its aa composition is strikingly biased towards Pro, Ser and Thr (together 50%). It harbors two potential PEST sites and 17 putative protein kinase C and casein kinase II phosphorylation sites. Computer homology searches revealed the presence of a DH domain followed by a PH domain at the C-terminus of the protein. This tandem arrangement of the two domains is characteristic for proteins of the Dbl oncogene family (11). With 35% overall sequence identity, the MyoM DH domain appears related closest to the DH domain of Tiam1 (Figure 2), an oncoprotein promoting tumor invasiveness and metastasis (17). Conserved aa mainly cluster in two regions belonging to two a-helices at the domain surface which, according to structural data, form the substrate binding site (18,19).

3D modeling of the PH domain hints at a potential protein ligand Despite low sequence conservation, PH domains exhibit a characteristic 3-D structure consisting of a seven-stranded antiparallel b-sandwich delimited on one edge by a single C-terminal a-helix. This prompted us to examine by homology modeling whether the putative MyoM PH domain could adopt this fold. A PROSA value of −0.8 kT calculated for the

Traffic 2000: 1: 399– 410

resulting model indicated good overall reliability (20). The electrostatic polarization of most PH domains supports an association with phosphoinositide-containing membranes (20). The ligand-binding site of the BTK and PLCd sits in the loop between the b1 and b2 strand, and is surrounded by a positive potential. Several positively charged residues there form a pocket for accommodating an anionic phosphoinositide headgroup (Figure 3a,b). The MyoM PH domain is distinguished by a reversed polarity, the b1b2-loop being located in a lobe of negative potential. This particular distribution of electrostatic potential occurs in about half of the Dbl family proteins (20). In MyoM, the loop carries three additional acidic residues, while the only two positively charged residues face away from the ligand-binding site (assuming the model correctly represents side chain orientation) (Figure 3c). Taken together, these findings argue against phospholipid binding in the usual way and thus make interactions with proteinaceous binding partners appear more likely.

MyoM is present only during the first half of the developmental cycle Upon starvation, D. discoideum cells commit to a developmental program leading to the formation of a fruiting body within 24 h. Northern blot analysis of poly(A) +RNA revealed a 5.3-kb transcript, in accordance with the calculated sequence length of myoM, present at constant low abundance during the first half of the developmental cycle (Figure 4a). While this signal gradually decreases as the cells progress into slug stage, a second band, slightly lower than the first one, appears. It is not clear yet whether this shorter RNA represents a processing or degradation product or an

401

Geissler et al.

gave the first indication that the IQ domains are ‘functional’ and might each bind to a light chain, as MheadIQ eluted at a significantly higher apparent size ( 200 kDa).

Figure 3: The potential ligand-binding site of the MyoM PH domain is different from that of PH domains known to bind phosphoinositides. The PH domains of Bruton’s tyrosine kinase (BTK), Phospholipase Cd (PLCd), and MyoM were built by homology modeling on the basis of all eight solved PH domain structures. The sequences corresponding to the ligand-binding site of BTK and PLCd are presented on top of the panels and are shown as ribbon models in (a) and (b), respectively. The same region of the MyoM PH domain is shown in (c). For orientation, this potential ligand-binding site is indicated by an arrow on the rear-upper right of the MyoM PH fold (c, arrow in inlet). The residues making contact with phosphoinositide ligands in BTK (a) and PLCd (b) are color-coded according to charge (Arg, Lys: blue; Asp, Glu: red, others: green). For MyoM (c), all charged residues in the b1 and b2 strands and the loop in between are shown.

unrelated but cross-reacting transcript. Monitoring the developmental regulation at the protein level revealed a band at  195 kDa absent from myoM null cells (Figure 4b). Like the larger mRNA signal, the protein is detectable during the first half of the cycle only, suggesting a functional link between MyoM and the activities and motility associated with the amoeboid life style. From this basic characterization, it can be concluded that MyoM is a complex molecule, with the appearance of a chimera between a myosin motor domain and a Rho-GEF tail domain. To elucidate the biological function of MyoM, the endogenous gene was disrupted in wild-type cells. In addition, to assess the functionality of the motor and tail domains individually, we expressed each of the two moieties separately in D. discoideum.

The binding of MyoM to F -actin may be phosphorylation -dependent The motor domain of MyoM (called Mhead), with or without the potential light chain binding sites (IQ), and lacking the predicted coiled-coil region should be a (heavy chain) monomer and soluble. Indeed, fractionation of D. discoideum cells expressing Mhead (100 kDa) or MheadIQ (105 kDa, Figure 6d) indicated that both fragments were almost exclusively found in the 100000 × g supernatant (not shown). Gel filtration analysis (Figure 5a) subsequently confirmed that the head domain was likely to be monomeric (100 kDa) and

402

One of the basic properties expected from a myosin is ATP-dependent binding to F-actin. This axiomatic hypothesis was tested experimentally. Co-sedimentation assays with detergent insoluble rigor cytoskeleton were performed to study the behavior of Mhead. As expected, in the presence of ATP, neither Mhead nor myosin II efficiently bound to F-actin. Surprisingly, in the absence of ATP, Mhead also failed to interact strongly with F-actin, whereas myosin II was quantitatively recovered in the pellet fraction (Figure 5b). It is reasonable to assume that, because of the 90 residue insertion in the surface loop 2, the interaction of MyoM with actin might be regulated differently than in myosin I and II. In contrast to myosin II, MyoM carries a Ser residue at the position predicted by the TEDS-rule. Thus, the potential influence of phosphorylation on actin-binding was investigated. Inclusion of a broad spectrum phosphatase inhibitor, NaF, during the actin-cosedimentation resulted in 30 – 50% recovery of Mhead in the pellet fraction (Figure 5b). The effect was less pronounced when a cocktail of b-glycerophosphate and the more specific inhibitors okadaic acid and cypermethrin were used (not shown). Importantly, NaF had no effect on myosin II and did not induce sedimentation of Mhead in the presence of ATP. In order to test whether this interaction is nevertheless reversible, the rigor pellet was subsequently incubated in the presence or absence of ATP. Myosin II is not released from actin without ATP, whereas we observed that small quantities of Mhead were liberated. After incubation with ATP, both myosin II and Mhead were quantitatively released (Figure 5c). The data indicate that binding of Mhead

Figure 4: MyoM is expressed during the first half of the developmental cycle. (a) Northern Blot analysis of poly(A) + RNA (5 mg/lane; * 12 h: 15 mg) isolated from cultures seeded onto starvation plates at 0 h. A 624-bp head domain probe revealed two signals, the one at 5.3 kb gradually decreasing during the second half of the developmental cycle, accompanied by a parallel increase of the one at 5.1 kb. The very faint signal at 0 h is likely to be due to a technical problem, as the myoM mRNA was detected in similar amounts at 0 and 2 h by RT-PCR (not shown). (b) Analysis of MyoM protein expression in cells harvested at corresponding time points. Extract of myoM null cells (M −) served as negative control. The blot was probed with pAb 1849 directed against the DH/PH domains and indicated a similar expression profile for MyoM protein and the 5.3 kb transcript.

Traffic 2000: 1: 399 – 410

MyoM acts as a Rac1-GEF

to actin, even when induced by NaF, was ATP-dependent. NaF had no effect on myosin II and did not affect the minute extent of Mhead released in the absence of ATP. We conclude that, potentially owing to its particular architecture, phosphorylation of the TEDS-site may regulate the binding of MyoM to actin, a role different to that observed for myosin I (14) and myosin VI (21).

Overexpression rather than absence of MyoM affects viability and morphology Successful homologous recombination events with a head domain construct disrupting the endogenous myoM gene by insertion of a blasticidin resistance cassette (Figure 6a) were confirmed by Southern blot analysis (Figure 6b,c) and western blotting (Figure 6d,e). MyoM null mutants and the parental Ax2 wild-type strain showed similar growth and phagocytosis rates, as well as similar viability under hyperand hypo-osmotic stress conditions. Development was also found to progress with no obvious differences in morphology or kinetics (data not shown). Thus, data obtained with one representative clone are shown below. In many cases, ablation of a cytoskeleton protein in D. discoideum failed to produce an obvious phenotype, while overexpression of the whole protein, or parts thereof, led to severe effects. Moreover, it has been shown that the expres-

Figure 5: Characterization of actin- and light chain-binding to MyoM. (a) The fractions obtained after gel filtration of cytosolic extracts of cells expressing either Mhead or MheadIQ were probed with the pAb 1081 anti-MyoM antibodies. On the left (− ) indicates a total protein extract from wild-type cells and (+ ) an extract from cells expressing Mhead or Mhead +IQ, respectively. On top of the gels, molecular size markers used to calibrate the gel filtration column are indicated, whereas on the right, size markers for the SDS-PAGE are shown. (b) Actincosedimentation of Mhead (Mhead) and myosin II (MyoII) with the Triton insoluble cytoskeleton and (c) ATP release. Cells were lysed in 0.5% Triton in the presence ( +ATP) or absence (− ATP) of ATP, as well as in combination of the presence (+ NaF) or absence ( − NaF) of a phosphatase inhibitor, and centrifuged. Pellets (P) and supernatants (S) were analyzed by western blotting with anti-Myosin II and anti-MyoM antibodies. (c) Subsequently, the pellet from the precipitation experiment ([b], last lane, P− ATP+ NaF) was incubated in the presence or absence of 10 mM ATP. The release was also performed in the presence or absence of NaF.

Traffic 2000: 1: 399– 410

Figure 6: Disruption of the myoM gene by homologous recombination. (a) A vector for gene disruption was generated by replacing a 1.1-kb SpeI fragment with a 1.4-kb blasticidin (bsr ) resistance cassette in a construct encompassing the entire head domain coding sequence. Only those restriction sites used for vector construction or analysis of transformants (see [b] and [c]) are shown. Double arrows above the genomic locus represent myoM genomic DNA (m5g1, m3g1) and cDNA clones from phage library screening (ZC1, ZD3, ZD4) or RT-RACE amplification (8RT), respectively. (b) Confirmation of the myoM null genotype. Genomic DNA isolated from wild-type (wt) and transformants (M − ; one representative clone shown here) were digested with various enzymes (ClaI, BamHI, or BglII/XhoI) and subjected to Southern blot analysis. The different fragment lengths detected with a myoM head domain probe (position indicated in [a]) correlated with the sizes calculated for wild-type and myoM − mutant genotype. (c) Confirmation of single recombination event by Southern blot analysis of the same membrane with a bsr cassette probe (position indicated in [a]). (d) and (e) Recombinant proteins expressed from the constructs presented in Figure 1c. (d) Western blot of whole cell extracts from Ax2 (Wt), full-length MyoM overexpresser (M +), myoM null mutant (M − ), Mhead, and MheadIQ overexpresser strains, probed with pAb 1081 recognizing a fragment of the head domain. The arrow points to MheadIQ and the arrowhead to Mhead constructs, respectively. (e) Western blot of whole cell extracts from Ax2 (Wt), full-length MyoM overexpresser (M +), myoM null mutant (M−) and Mtail domain overexpresser (GEF +) strains probed with pAb 1849 recognizing the DH/PH domain. The arrow points to the MyoM and the arrowhead to the Mtail constructs.

403

Geissler et al.

sion of (mutated) Dbl-type GEF domains in mammalian cell lines affects cytoskeletal organization (11). Therefore, D. dis coideum cells were transformed with expression constructs containing the coding sequence for either the entire protein or the tail domain only. In both cases, stable clones were obtained in the myoM −, but not in the wild-type genetic background (Figure 6e). Both MyoM + and GEF+ clones showed longer doubling time (up to 16 instead of 8 h), decreased survival after cell freezing/thawing and a slightly delayed developmental cycle (up to 30 instead of 24 h). These observations suggested that elevated amounts of the tail domain were detrimental to cell viability especially in the presence of the MyoM motor domain. The morphology and dynamics of adherent cells growing in medium were monitored by time lapse video microscopy (Figure 7A). Wild-type cells showed the expected range of amoeboid movement, with the cyclic extension and retraction of pseudopodia (Figure 7Aa). Cells lacking MyoM are barely distinguishable from wild-type cells, with the exception of a higher number of cells extending filopodia and adopting a ‘spiky’ morphology (Figure 7Ab). In contrast, even though the GEF + cells were heterogeneous, many cells strikingly extended large, semi-circular lamellipodia reminiscent of migrating mammalian fibroblasts or fish keratocytes (Figure 7Ac). During the chemotactic aggregation phase induced by starvation, the cells tended not to form clearly defined and compact streams, but instead migrated individually towards the center of the aggregating cell mass (data not shown).

Cells overexpressing the MyoM tail domain are hyper -reactive to osmotic stress Early observations indicated that GEF + cells were responding acutely to mechanical stress such as occurring during culture passaging. After replating, many cells were exaggeratedly flattened, and some exhibited extremely extended cytoplasmic bridges resembling those described for myosin II null cells dividing by traction-mediated cytofission (22). In addition, washing of the cells in low osmolarity phosphate buffer resulted in a similar response. Therefore, we subjected the cells to well-defined osmotic stress conditions. The most remarkable effects were seen in cells immersed in water (Figure 7B), but cells also responded to 300 mM sorbitol with reversible morphological aberrations (not shown). Wild-type cells showed the typical reaction to water, in the first few minutes cells stop their amoeboid movements and round up (Figure 7Ba). After about 30 min, some activity is back and cells soon start their developmental program. MyoM − cells showed the same kinetic of changes, with a slightly more turgid morphology and less residual activity (Figure 7Bb). In sharp contrast, after a few seconds in water, GEF + cells reacted strongly and extended protrusions that sometimes exceeded the resting cell diameter in size (Figure 7Bc,d). The protrusions contained amorphous material, visible as a smooth darker gray in phase contrast, and were clearly devoid of organelles (Figure 7Bf). We propose to name these protrusions ‘ramopodia’, as they

404

Figure 7: Time lapse video microscopy of cells in medium or in water. Wild-type (WT), myoM − cells (M−) and Mtail overexpressers (GEF + ) were observed in growth medium ([A], medium) or after immersion in water for about 10 minutes ([B], water). Frames were taken every 3 s (arbitrarily starting at 0’’, and then at 3’’, 6’’ and 9’’). The arrowheads in (A) indicate: a wild-type cell (a); a myoM − cell (b); and two GEF+ cells (c). The arrow points to a myoM − cell extending excessive filopodia (b). The arrowheads in (B) indicate: two wild-type cells (a) with residual membrane deformation activity; and a myoM − cell (b) showing some blebbing. The GEF + cells show extensive and highly dynamic protuberances. In (c), the arrowheads point to a cell showing a ‘bifurcation’, and in (f) to a cell extending a huge protrusion, both growing at about 2 mm per s. The arrow in (d) indicates a cell with a lamellipodia-like extension that splits into smaller protrusions. Some cells even adopt a ‘cog-wheel’ morphology (e). Bars, 20 mm

seem to result from poking of the plasma membrane under the action of a ram. Some cells extended apparently large lamellipodia-like protrusions which often split in smaller domains (Figure 7Bd). The number of cells exhibiting these phenotypes steadily increased and after about 10 – 30 min, most cells showed aberrant morphologies. In extreme cases, cells adopted a relatively long lasting ‘cog-wheel’ appearance (Figure 7Be). It can be concluded that GEF overexpression did not lead to a strong constitutive phenotype, but an acute

Traffic 2000: 1: 399 – 410

MyoM acts as a Rac1-GEF

reaction was induced in response to a stress, in the present case a hypo-osmotic shock. The amplitude of the phenotypes correlated with the level of Mtail expression and, thus, data obtained from a clone showing a representative and homogeneous phenotype are presented here. In contrast, the MyoM+ cells did not display such morphological phenotypes. This is possibly due to either the higher expression level of Mtail compared with full-length MyoM (Figure 6e) or to the lack of a negative regulatory element in the Mtail construct. The latter is reminiscent of the oncogenic potential of Dbl family GEFs which can be revealed by deletion of adjacent repressor domains (11).

Ramopodia are a novel type of actin -driven protrusions The organization of the actin cytoskeleton, revealed by phalloidin staining (Figure 8A), perfectly reflected the respective morphologies observed by video microscopy in parallel conditions (Figure 7). In order to better appreciate the 3-D morphology of the GEF + cells in water, reconstruction by

deconvolution of confocal stacks was used (Figure 8B). The ramopodia were filled with actin (green) and had a collar of coronin, an actin-associated protein, at their base (red). The partial overlap in distribution showed up in yellow. The structure resembling a lamellipodium when viewed from the top (Figure 8Bc,d, arrow) was actually not extending on the substratum.

The GEF domain localizes at the tip of ramopodia An affinity-purified antibody (see also Figure 6e) directed against the DH/PH domain (Figure 9a) was used to localize the GEF in cells responding to the osmotic shock (Figure 8C). In cells with short ‘burgeons’, the mainly cytosolic GEF was slightly enriched under the plasma membrane of the nascent outgrowth, and became more and more abundant as the stubby protuberances grew. In cells with fully extended fingers, the GEF was localized at the tip of the ramopodia. In conclusion, upon osmotic shock, cells overexpressing the GEF produced a novel form of actin-filled protrusions. These ramopodia have an internal structure that can be subdivided in three regions: a cap of membrane associated MyoM GEF, an actin-filled core and a base enriched with coronin, an actin-binding protein.

Figure 8: Morphology of the cortical cytoskeleton of cells in medium and in water. Wild-type (WT), myoM − cells (M − ) and Mtail overexpressers (GEF + ) were fixed directly after growth in medium ([A], medium) or after immersion in water for about 10 min ([A, B and C], water). Cells were stained with only OregonGreen phalloidin (A), or with a combination of phalloidin (green) and anti-coronin antibody (red) (B), or with an antibody directed to the Mtail domain (C). After confocal microscopy and image processing (A) and (C) present whole cell extended focus projections, whereas (B) presents projections (b and d) as well as perpendicular ‘side views’ in the form of single sections or projections of a few sections cut at the places marked by small white arrowheads on the left and right of the central projections. In addition, shadow projections of three dimensional reconstructions are presented (a and c). For the observation in medium, the arrows in (A) indicate wild-type and myoM − cells extending classical bowl shaped structures called crowns, and two GEF+ cells with a morphology close to wild-type cells. The arrowheads point to wild-type and myoM − cells with filopodia and to GEF + cells harboring extensive lamellipodia. After short incubation in water, wild-type and myoM − cells have a rounded up, swollen morphology, whereas GEF + cells extend numerous long ramopodia filled with F-actin. The representations in (B) emphasize the shape of the projections and the organization of the cytoskeleton: F-actin (green) fills the tip of the projections and coronin, an actin-associated protein (red) is enriched at their base. Arrowheads indicate the position of the same projection in the four corresponding panels. The arrow points to a protrusion which resembles a lamellipodia when viewed from the top (c and d, central projection) but did not lay on the substratum (d, right side view). (C) presents a gallery of GEF + cells in water, stained with an anti-DH/PH antibody. Arrowheads point to protrusions or ‘burgeons’ with higher local concentration of Mtail. Bars, 5 mm.

Traffic 2000: 1: 399– 410

405

Geissler et al.

The MyoM GEF domain selectively acts on a subset of Rac -GTPases To confirm in vitro the predicted GEF activity of MyoM, we expressed the DH and DH/PH fragments as GST fusion proteins in Escherichia coli and purified them to homogeneity (Figure 9a). The recombinant proteins were tested in filterbinding assays for their ability to stimulate the exchange of bound GDP for GTPgS on Rho family GTPases. Control reactions confirmed that the purified fragments themselves do not bind guanine nucleotides (not shown). D. discoideum Rac1A and RacC showed similar intrinsic rates for nucleotide exchange, with 20% of GTP being exchanged on total active protein after 40 min at 20°C. In the presence of the MyoM DH domain, exchange activity of Rac1A was markedly accelerated (Figure 9b), whereas no effect was observed with RacC (Figure 9c). The stimulating effect was found to be independent of the PH domain under the conditions used (data not shown), as the DH/PH had similar activity. To further explore the specificity of MyoM GEF activity, a variety

of Rho family GTPases were tested as substrates. Significant GEF activity was exclusively observed with members of the subclass of Rac GTPases, and within this subclass the MyoM DH domain again was only active on a subset of proteins related to Rac1, irrespective of the species of origin (Figure 9d). Importantly, both Rac1A and Rac1B show the same developmental expression profile as MyoM (23). These results demonstrate a distinct substrate selectivity for MyoM.

Figure 9: The MyoM DH domain is a functional GEF selectively catalyzing guanine nucleotide exchange on Rac1-related small GTPases. (a) Coomassie-stained polyacrylamide gel of two constructs presented in Figure 1c, GST-DH (53 kDa) and GST-DH/PH (69 kDa) purified from E. coli. Association of GTPgS with D. discoideum Rac1A (b) and RacC (c) was monitored in the absence () or presence ( ) of bacterially expressed, purified GST-DH in a 1:2 molar ratio. The absence of nucleotide binding to GST-DH alone was confirmed in separate control reactions (not shown). (d) Substrate specificity of GST-DH. The same filter-binding assay was used to test a variety of Rho family GTPase substrates from D. discoideum (D.d.), H. sapiens (H.s.) and Saccharomyces cerevisiae (S.c.). Exchange activity was assessed in reaction mixtures containing 0.5–1.5 mM GTPase (depending on the specific activity of the individual protein preparation) and either control buffer to measure intrinsic exchange rates (open columns) or GST-DH in 5-fold molar excess (closed columns) after 30 min incubation (RacE: 10 min) at 20°C. Results are expressed as a percentage of the maximal binding obtained in the presence of 200 mM ammonium sulfate.

Overexpression of the MyoM tail domain had a higher impact on the cells than gene disruption. Whereas the null mutants behaved like wild-type cells in functional assays for phagocytosis and growth rates, development and osmoregulatory capacity, cells overexpressing the tail domain exhibited an altered morphology with unusually wide lamellipodia, even under normal growth conditions. As an acute response to hypo-osmotic stress, we observed a burst of massive actindriven protrusions emanating from the cells. These ramopodia are distinct from the actin-driven protrusions induced by RacC overexpresssion (24) and are indeed likely to be triggered by a different signaling pathway, as the MyoM GEF domain was not active on RacC in vitro. Confocal microscopy revealed that the overexpressed tail domain was concentrated at the tip of ramopodia. In all Dbl-type GEFs, and consistently in MyoM, the DH domain is followed by a PH domain which is thought to function in targeting the protein to specific cytoskeletal or membrane locations (25 – 29). It

406

Discussion D. discoideum MyoM is the first myosin to contain a Dbl-type GEF domain in its tail. Also, the motor domain itself carries features divergent from the myosin head consensus sequence. The most prominent ones are two insertions reminiscent in length and position of the insertions found in myosins of class VI and IX, respectively; nevertheless, overall sequence identity is highest with B. taurus myosin X. Moreover, the TEDS-site, like in class VI-myosins and ameboid myosins of class I, is occupied by a phosphorylatable (Serine) residue, suggesting that motor activity may be regulated by phosphorylation. Considering these peculiarities in MyoM structure, the functionality of both the head and DH domain was tested in vitro. Binding of the motor domain to F-actin was found to be dependent on ATP and on the presence of a phosphatase inhibitor, in sharp contrast with conventional myosin II. Thus MyoM may act not as a true motor but rather serve to anchor the tail in actin-rich regions of the cell, as has been suggested for Myosin IX (6,10). In addition, whereas the genome of D. discoideum does not seem to encode a myosin with a GAP domain, a MyoM homolog might be hidden in the human genome. Furthermore, as was shown for myosin VI (21), MyoM might be a substrate for the recently identified PAK65/Ste20 kinases known as effectors of Rho family GTPases. Taken together with the findings that the MyoM tail DH domain acts as a GEF selective for Rac1related GTPases in vitro, a combination of these two functional elements in the one MyoM protein offers an attractive potential for regulatory circuits (Figure 10).

Traffic 2000: 1: 399 – 410

MyoM acts as a Rac1-GEF mg/ml BlasticidinS (Calbiochem), the Mhead + and MheadIQ + with 5 mg/ml, and M + and Mtail + cells with 2.5 mg/ml G418 (Gibco BRL), respectively. The transformation procedure was based on published protocols by (37,38). Clones were selected on dense lawns of Klebsiella aerogenes (36).

Southern and northern blotting Preparation of genomic DNA and total RNA was performed as described previously (12). Poly(A) + RNA was isolated from total RNA using the polyATtract kit (Promega). Southern and northern blotting were performed according to standard procedures (39), hybridized with DIG-labeled probes (Boehringer Mannheim) and finally developed with CDP-Star as detection substrate.

Figure 10: A working model for MyoM function. Rho-type GTPases are key regulators relaying extracellular stimuli from surface receptors to the organization of the actin cytoskeleton and transcription activation. In D. discoideum, such receptors may be serpentine receptors such as the cAMP receptors or other undefined transmembrane signaling molecules. The PH domain of Dbl-family GEFs, including MyoM, may be recruited to membranes by binding to phosphoinositides or proteins such as the Gb subunit of trimeric G proteins. MyoM having both a GEF domain and a TEDS phosphorylation site for a PAK/Ste20-like kinase potentially sits in a feed-back loop. The situation depicted here is reminiscent of signaling complexes found in S. cere visiae. There, the adaptor molecule Bem1, itself an actin-binding protein, was shown to bind simultaneously Cdc24, a GEF for Cdc42, and its downstream effector Ste20 or Cla4, both PAKtype kinases. PAK, p21-activated kinase.

will be exciting to learn how the signal transduction cascade involving MyoM will feed into the machinery triggering the spatial and temporal control of actin filament dynamics, probably involving known players such as PAK kinases, cofilin, SCAR and the Arp2/3 complex (3031 and references therein). While the signaling cascades triggering protection of the cell against a hyper-osmotic environment are well studied (32 – 35), the discovery of MyoM has opened a door in the investigation of the comparatively poorly understood cellular processes responding to hypo-osmotic stress. Altogether, the data reveal that MyoM represents a new facette of the complex coupling between signal transduction and cytoskeletal remodeling (Figure 10).

Materials and Methods Dictyostelium discoideum cell culture and transformation Dictyostelium discoideum strain Ax2 was grown at 23°C in HL-5c medium (36). MyoM − cells were maintained with 5

Traffic 2000: 1: 399– 410

Cloning of myoM cDNA and genomic locus MyoM was identified by a PCR screening strategy (12). A 624-bp DIG-labeled DNA probe derived from this fragment was used to screen a lZAP cDNA library (kindly provided by Dr W. Loomis, UCSD). This procedure yielded three overlapping clones covering the entire MyoM head domain plus some 350 bp of 5% UTR (ZC1, ZD3 and ZD4; see Figure 6a). Genomic DNA fragments were isolated by cloning size-fractionated Cla I digests of D. dis coideum DNA. Two 2.7-kb clones carrying myoM fragments were identified by colony hybridization with the same head domain probe used for screening the phage library (clone pm5g1), or a second probe directed against a 400-bp stretch, close to the neck domain (clone pm3g1). The 3% end of myoM was amplified following a RACE-protocol: first strand cDNA was prepared from 0.5 mg poly(A) + RNA using the SuperScript II kit (Gibco BRL) with an anchored oligo(dT) primer. Subsequently, PCR reactions were carried out on this cDNA with the oligo(dT) primer and a specific primer MyoM6, encompassing an endogenous EcoRI site (underlined; 5% C AGA GAG TAT TTG GAA TTC TG 3%). The PCR product of 750 bp (8RT) was cloned into the pCRII-TOPO vector (Invitrogen). The D. discoideum genome project (40) has recently produced shotgun sequencing results confirming over 85% of the myoM sequence.

Sequence analysis Sequencing of genomic and cDNA clones was carried out either with the DNA Sequencing kit Version 2.0 (Amersham Life Science) or the dideoxy terminator cycle sequencing kit (Perkin Elmer). Oligonucleotide primers for PCR and sequencing were synthesized in the facilities of our institute. Sequences were analyzed for primary and secondary motifs and structures by softwares contained in the HUSAR package (DKFZ, Heidelberg), as well as on the ExPASy website. Database searches were performed with BLAST and FASTA. Sequence alignment was performed using the clustalW program.

Generation of myoM expression constructs and mutant strains A construct pBS-Mhead comprising the complete coding sequence for the MyoM head domain was assembled from the lZAP-cDNA clones pZD4 and pZC1 into pBS vector. The resulting pBS-Mhead construct was used as template for PCR with primers hybridizing before and after the IQ motifs, giving rise to the Mhead (residue 1-892) and MheadIQ (residue 1-937), respectively. The PCR fragments containing the introduced 5% BamHI and 3% XhoI restriction sites were cloned into the corresponding sites of the D. discoideum expression vector pDXA-HC (41).

407

Geissler et al. The pBS-Mhead construct was further modified for disruption of the myoM gene by homologous recombination: a central 1.1-kb SpeI fragment was replaced by a blasticidin resistance (bsr ) cassette released by XbaI digestion from the vector pBsR 503 (42). The interrupted coding sequence was then excised with EaeI and PstI and dephosphorylated. Transformation was performed with 20 mg of DNA. Resulting colonies were subsequently cloned by cultivation on bacterial lawns and screened by PCR. Genomic DNA was purified using DNAzol (Molecular Research Center). Positive candidates were first identified by PCR and then confirmed by Southern and western blotting. A construct encompassing the entire tail domain of MyoM, pBS-Mtail was obtained by inserting into pm3g1 a HincII/XhoI fragment from the pCR-TOPO vector containing the 3% end of myoM. A full-length construct for MyoM was obtained by joining the respective fragments from pBS-Mhead and pBS-Mtail at a common BsmI site. Expression of the M-full length, Mhead, MheadIQ and Mtail constructs was obtained by electroporation of AatII and PvuI doubly digested plasmids and selection with G418, resulting in the M +, Mhead+, MheadIQ+ and GEF+ strains, respectively.

Construction of His -tag and Glutathione -S -transferase (GST) fusion proteins To generate an antibody against the MyoM head domain, the coding sequence for a 38 kDa segment (residues 154-483) was modified by PCR to introduce a 5% NcoI and a 3% BglII restriction sites, a start methionine, an N-terminal octa-His tag and multiple stop codons. The amplified fragment was double-digested and ligated into the respective sites of pET8c. Glutathione S-transferase (GST) fusion products of the MyoM DH domain with and without pleckstrin homology domain (DH, 53 kDa and DH/PH, 69 kDa) were generated for in vitro measurements of GEF activity and antibody production. The pBSMtail construct was used as template for the following PCR reactions. The DH/PH domain (residues 1323-1737) was amplified by PCR, introducing 5% XmaI and SacI sites and 3% XhoI and XmaI sites. The PCR product was digested with XmaI and fused in-frame behind GST in the pGEX-3X bacterial expression vector (Pharmacia). A GST fusion protein omitting the PH domain (residues 1323-1595) was generated in a similar fashion but introduced a stop codon after the DH domain.

Expression and purification of recombinant proteins The pGEX and pET plasmid constructs described above were transformed into the E. coli strain BL-21 (DE3) (43). Overnight cultures were grown at 30°C in the presence of 2% glucose and 150 mg/ml ampicillin. A liter of LB-medium containing 100 mg/ml Ampicillin was inoculated 1:20 with these starter cultures and grown at 30°C to an OD600 of 0.6–0.9. Cultures transformed with the His-tagged head domain fragment construct were induced with 0.5 mM IPTG for 4 h at 30°C. Cells were harvested by centrifugation, resuspended in lysis buffer (6 M guanidinium hydrochloride, 0.1 M NaH2PO4 and 10 mM Tris-HCl, pH 8.0) and disrupted with a probe sonicator (Branson). After clarification, the extract was purified on a Ni-NTA column (QIAGEN) under denaturing conditions. Eluates were dialyzed overnight against 50 mM Hepes, pH 7.2, 0.15 M NaCl, 0.005% Tween 20, 10 mM DTT, supplemented with 200 mM O phenantroline, 100 mM phenylmethylsulfonyl fluoride and 5 mM benzamidine. Under these conditions, the protein quantitatively

408

precipitated, was collected by centrifugation, resuspended in sterile PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 15 mM K2HPO4, pH 7.4) to a concentration of 1 mg/ml and used to immunize rabbits. To obtain soluble material for immobilization on a resin, the concentration of chaotropic agent in the eluate was lowered by 2 × 2-h dialysis against the same buffer containing 4 M urea and coupled to Affi-Gel 10 (BioRad). Expression of GST fusion proteins was induced by addition of IPTG to a final concentration of 0.2 mM. Cells were harvested after overnight growth at 16°C, washed and lysed by sonication. Affinity chromatography was carried out on glutathione Sepharose 4B (Pharmacia). Bound protein was eluted in a buffer containing 100 mM Tris-HCl, pH 8.5, 50 mM KCl, 20 mM glutathione, 1 mM DTT and 10% glycerol. Further purification and buffer exchange was achieved by gel filtration on a HS 75 10/10 FPLC column (Pharmacia) in 1 ×PBS. Fractions containing the main protein peak were pooled, supplemented with 2 mM MgCl2, 1 mM DTT and 10% glycerol, aliquoted and frozen in liquid nitrogen.

Antibodies Polyclonal antibodies were raised against MyoM (Biogenes, Berlin) by immunizing rabbits with the recombinant His-tagged motor domain fragment (pAb 1081) and the GST-tagged DH/PH protein produced in E. coli (Ab 1849). Crude sera were further immunopurified as follows. First, the sera were cross-adsorbed to 1% (w/v) acetone powder from E. coli BL21 cells. The serum against GST-DH/PH was depleted from antibodies against the GST moiety by incubation with immobilized GST (Affi-Gel 10, BioRad). The flowthrough was further affinity-purified against the DH/PH or head domain protein, respectively, by incubation with the antigens coupled to Affi-Gel 10 (BioRad). Bound antibodies were eluted as described previously (Harlow and Lane, 1988). Finally, affinity-purified antibodies were diluted 1:10000 in 3% non-fat dry milk in TBST (150 mM NaCl, 50 mM Tris/Cl pH 7.4, 0.1% Tween-20) and cross-adsorbed overnight to extracts from 1.5 ×107 myoM − cells transferred to nitrocellulose membrane. This antibody solution was then used to probe western blots or for immunofluorescence. The anti-D. discoideum coronin (mAb 176-306-3) and anti-myosin II (mAb 56-396-5) were a gift from Dr G. Gerisch (MPI for Biochemistry, Martinsried). F-actin was visualized using 2U of OregonGreen-conjugated phalloidin (Molecular probes) per coverslip. Rabbit polyclonal and monoclonal antibodies anti-GST and anti-His-tag were purchased from Sigma and QIAGEN, respectively.

Microscopy Time-lapse video microscopy was performed on an Axiovert 100 equipped with a 40 × Achroplan objective (NA 0.6) and a CCD camera (CCD-IRIS, Sony). Images were captured at 3-s intervals by the Metamorph system, and converted to quicktime movies by the NIH Image software. Immunofluorescence was performed as described previously (44), using a methanol fixation protocol. Alternatively, in order to visualize F-actin with phalloidin, cells were fixed with picric acid and paraformaldehyde (45) without the ethanol post-fixation. Secondary antibodies were goat-anti-mouse or anti-rabbit IgGs coupled to Cy3 (BioTrend) or Alexa488 (Molecular probes). The antibodies were diluted in PBS/0.2% gelatine. Mounted samples were analyzed with an Axiophot2 microscope (Zeiss) or a Leica

Traffic 2000: 1: 399 – 410

MyoM acts as a Rac1-GEF confocal microscope DM/IRB using a 63 × Plan-Apo objective with NA 1.40. Confocal optical sections were recorded at 50 nm per vertical step with 4-fold accumulation. Image reconstruction and processing were carried out using the Huygens/Selima/ Imaris software package from Bitplane AG (Zurich, Switzerland).

Co -sedimentation with Triton -insoluble ‘rigor’ cytoskeletons Triton insoluble cytoskeletons were obtained according to Manstein and Hunt (46) with minor modifications described previously (47). In some sedimentation and release reactions, 50 mM NaF was added as a broad spectrum phosphatase inhibitor. The pellets and supernatants resulting from the co-sedimentation and the release were analyzed by western blotting with anti-MyoM and anti-myosin II antibodies. Protein extracts and imunoblotting procedures were carried out as described previously (48). Signals were visualized by chemiluminescence (ECL or ECLplus, Amersham) and directly recorded by a Luminescent Image Analyser LAS-1000 (FujiFilm) allowing for quantification of signal intensities within a broad linear range. Determination of GEF activity The guanine nucleotide exchange activity of the MyoM DH domain was monitored in an in vitro assay (49). The GTP-binding proteins (0.3 – 1.5 mM) were incubated with a 5-fold stoichiometric excess of isolated DH protein in the presence of 5 mM guanosine 5%-(g-(35S)thio)triphosphate (Sigma) at 20°C. All samples contained 25 mM Hepes pH 7.4, 100 mM NaCl, 2.5 mM MgCl2 and 1 mg/ml ovalbumin. At given time points, 50-ml aliquots from the reactions were diluted into 1 ml of ice-cold stop buffer (50 mM Tris-HCl, pH 7.4, 50 mM NaCl, 10 mM MgCl2) and subsequently filtered over BA-85 nitrocellulose filters (Schleicher & Schuell). Radioactivity bound to the filters was determined by liquid scintillation counting. Results are expressed as a percentage of maximal binding as assessed in samples containing 200 mM ammonium sulfate (49). Samples containing the recombinant GEF domain, but no GTPase, were set up as negative controls. In all assays, the GST fusion proteins were used without prior removal of the GST tag.

2.

3.

4. 5.

6.

7. 8.

9. 10.

11. 12.

13.

14.

15. 16. 17.

18.

Acknowledgments 19.

In addition to all the laboratory members, we would like to thank Jutta Schleich and Uschi Buhre for their excellent technical assistance, Dr Gu¨nther Giese for help with the processing of confocal files and Dr Bernd Helms for hosting and help during the nucleotide exchange assays. Special thanks go to Drs Menno Knetsch and Dietmar Manstein, Daria Illenberger and Peter Gierschik, Hans Faix and Arturo deLozanne, for providing expression vectors and purified Rho-GTPases. Invaluable expertise for the homology modeling of the PH domain was provided by Drs Nicklas Blomberg and Michael Nilges. The work was supported by the Max-Planck-Gesellschaft and grants from the DFG (SFB 352, SPP 1068 and GK Molekulare Zellbiologie).

20. 21.

22. 23.

24.

Note added in proof: MyoM was independently identified by Drs Adachi and Sutoh and coworkers (Accession no. AB 017910). 25.

References 26. 1.

Mermall V, Post PL, Mooseker MS. Unconventional myosins in cell-movement, membrane traffic, and signal-transduction. Science 1998;279: 527–533.

Traffic 2000: 1: 399– 410

Wang AH, Liang Y, Fridell RA, Probst FJ, Wilcox ER, Touchman JW, Morton CC, Morell RJ, Nobentrauth K, Camper SA, et al. Association of unconventional myosin myo15 mutations with human nonsyndromic deafness dfnb3. Science 1998;280: 1447 – 1451. Probst FJ, Fridell RA, Raphael Y, Saunders TL, Wang AH, Liang Y, Morell RJ, Touchman JW, Lyons RH, Nobentrauth K, et al. Correction of deafness in shaker-2 mice by an unconventional myosin in a BAC transgene. Science 1998;280: 1444 – 1447. Ba¨hler M. Myosins on the move to signal transduction. Curr Opin Cell Biol 1996;8: 18 – 22. Soldati T, Schwarz EC, Geissler H. Unconventional myosins at the crossroad of signal transduction and cytoskeleton remodeling. Protoplasma 1999;209: 23 – 37. Mu¨ller RT, Honnert U, Reinhard J, Ba¨hler M. The rat myosin myr 5 is a GTPase-activating protein for Rho in -vivo -essential role of arginine 1695. Mol Biol Cell 1997;8: 2039 – 2053. Symons M. Rho family GTPases: The cytoskeleton and beyond. Trends Biochem Sci 1996;21: 178 – 181. Tapon N, Hall A. Rho, Rac and Cdc42 GTPases regulate the organization of the actin cytoskeleton. Curr Opin Cell Biol 1997;9: 86 – 92. Hall A. Rho GTPases and the actin cytoskeleton. Science 1998;279: 509 – 514. Chieregatti E, Gartner A, Stoffler HE, Bahler M. Myr 7 is a novel myosin IX RhoGAP expressed in rat brain. J Cell Sci 1998;111: 3597 – 3608. Cerione RA, Zheng Y. The Dbl family of oncogenes. Curr Opin Cell Biol 1996;8: 216 – 222. Schwarz EC, Geissler H, Soldati T. A potentially exhaustive screening strategy reveals two novel divergent myosins in Dictyostelium. Cell Biochem Biophys 1999;30: 413 – 435. Cope MJT, Whisstock J, Rayment I, Kendrick-Jones J. Conservation within the myosin motor domain: Implications for structure and function. Structure 1996;4: 969 – 987. Bement WM, Mooseker MS. TEDS rule: A molecular rationale for differential regulation of myosin by phosphorylation of the heavy chain head. Cell Motil Cytoskel 1995;31: 87 – 92. Coluccio LM, Myosin I. Am J Physiol 1997;273: C347 – C359. Rhoads AR, Friedberg F. Sequence motifs for calmodulin recognition. FASEB J 1997;11: 331 – 340. Habets GGM, Collard JG. Identification of an invasion-inducing gene, Tiam-1, that encodes a protein with homology to GDP-GTP exchangers for rho-like proteins. Cell 1994;77: 537 – 549. Liu X, Wang H, Eberstadt M, Schnuchel A, et al. NMR structure and mutagenesis of the N-terminal Dbl homology domain of the nucleotide exchange factor Trio. Cell 1998;95: 269 – 277. Soisson SM, Nimnual AS, Uy M, Bar-Sagi D, Kuriyan J. Crystal structure of the Dbl and pleckstrin homology domains of the human Son of sevenless protein. Cell 1998;95: 259 – 268. Blomberg N, Nilges M. Functional diversity of PH domains: An exhaustive modeling study. Fold Des 1997;2: 343 – 355. Buss F, KendrickJones J, Lionne C, Knight AE, Cote GP, Luzio JP. The localization of myosin VI at the Golgi complex and leading edge of fibroblasts and its phosphorylation and recruitment into membrane ruffles of A431 cells after growth factor stimulation. J Cell Biol 1998;143: 1535 – 1545. Spudich JA. In pursuit of myosin function. Cell Reg 1989;1: 1 – 11. Bush J, Franek K, Cardelli J. Cloning and characterization of seven novel Dictyostelium discoideum Rac-related genes belonging to the Rho family of GTPases. Gene 1993;136: 61 – 68. Seastone DJ, Lee E, Bush J, Knecht D, Cardelli J. Overexpression of a novel Rho family GTPase, RacC, induces unusual actin-based structures and positively affects phagocytosis in Dictyostelium dis coideum. Mol Biol Cell 1998;9: 2891 – 2904. Yao LB, Janmey P, Frigeri LG, et al. Pleckstrin homology domains interact with filamentous actin. J Biol Chem 1999;274: 19752– 19761. Glaven JA, Whitehead I, Bagrodia S, Kay R, Cerione RA. The Dbl-related protein, Lfc, localizes to microtubules and mediates the activation of rac signaling pathways in cells. J Biol Chem 1999;274: 2279 – 2285.

409

Geissler et al. 27. Ren Y, Li R, Zheng Y, Busch H. Cloning and characterization of GEF-H1, a microtubule-associated guanine nucleotide exchange factor for rac and rho GTPases. J Biol Chem 1999;273: 34954 – 34960. 28. Zheng Y, Zangrilli D, Cerione RA, Eva A. The pleckstrin homology domain mediates transformation by oncogenic dbl through specific intracellular targeting. J Biol Chem 1996;271: 19017 – 19020. 29. Levine TP, Munro S. the pleckstrin homology domain of oxysterolbinding protein recognises a determinant specific to Golgi membranes. Curr Biol 1998;8: 729 – 739. 30. Machesky LM, Gould KL. The Arp2/3 complex: A multifunctional actin organizer. Curr Opin Cell Biol 1999;11: 117 – 121. 31. Machesky LM, Insall RH. Signaling to actin dynamics. J Cell Biol 1999;146: 267–272. 32. Insall RH. Osmoregulation: Cyclic GMP and the big squeeze. Curr Biol 1996;6: 516–518. 33. Kuwayama H, Ecke M, Gerisch G, van Haastert PJM. Protection against osmotic stress by cGMP-mediated myosin phosphorylation. Science 1996;271: 207 – 209. 34. Kuwayama H, van Haastert PJM. Chemotactic and osmotic signals share a cGMP transduction pathway in Dictyostelium discoideum. FEBS Lett 1998;424: 248 – 252. 35. Schuster SC, Noegel AA, Oehme F, Gerisch G, Simon MI. The hybrid histidine kinase DokA is part of the osmotic response system of Dictyostelium. EMBO J 1996;15: 3880 – 3889. 36. Sussman M. Cultivation and synchronous morphogenesis of Dic tyostelium under controlled experimental conditions. Meth Cell Biol 1987;28: 9–29. 37. Howard PK, Ahern KG, Firtel RA. Establishment of a transient expression system for Dictyostelium discoideum. Nucl Acids Res 1988;16: 2613–2623. 38. Kuspa A, Loomis WF. Tagging developmental genes in Dictyos telium by restriction enzyme-mediated integration of plasmid DNA. Proc Natl Acad Sci USA 1992;89: 8803 – 8807.

410

39. Sambrook J, Fritsch EF, Maniatis T. In Nolan C (Ed.), Molecular Cloning: A Laboratory Manual. New York: Cold Spring Harbor Laboratory Press, 1989. 40. Loomis WF. The Dictyostelium genome sequencing project. Protist 1998;149: 209 – 212. 41. Manstein DJ, Schuster HP, Morandini P, Hunt DM. Cloning vectors for the production of proteins in Dictyostelium discoideum. Gene 1995;162: 129 – 134. 42. Puta F, Zeng C. Blasticidin resistance cassette in symmetrical polylinkers for insertional inactivation of genes in Dictyostelium. Folia Biologica 1998;44: 185 – 188. 43. Luderer-Gmach M, Liebig HD, Sommergruber W, Voss T, Fessl F, Skern T, Kuechler E. Human rhinovirus 2A proteinase mutant and its second-site revertants. Biochem J 1996;318: 213 – 218. 44. Neuhaus E, Horstmann H, Almers W, Maniak M, Soldati T. Ethanefreezing/methanol-fixation of cell monolayers. A procedure for improved preservation of structure and antigenicity for light and electron microscopies. J Struct Biol 1998;121: 326 – 342. 45. Humbel BM, Biegelmann E. A preparation protocol for postembedding immunoelectron microscopy of Dictyostelium discoideum cells with monoclonal antibodies. Scanning Microsc 1992;6: 817 – 825. 46. Manstein DJ, Hunt DM. Overexpression of myosin motor domains in Dictyostelium : Screening of transformants and purification of the affinity tagged protein. J Muscle Res Cell Motil 1995;16: 325 – 332. 47. Schwarz EC, Neuhaus EM, Kistler C, Henkel AW, Soldati T. Dictyos telium myosin IK is involved in the maintenance of cortical tension and affects motility and phagocytosis. J Cell Sci 2000; in press. 48. Ulbricht B, Soldati T. Production of reagents and optimization of methods for studying calmodulin-binding proteins. Prot Expr Purif 1999;15: 24 – 33. 49. Zheng Y, Hart MJ, Cerione RA. Guanine nucleotide exchange catalyzed by the Dbl oncogene product. Meth Enzymol 1995;256: 77– 84. 50. Lupas A, Van Dyke M, Stock J. Predicting coiled coils from protein sequences. Science 1991;252: 1162 – 1164.

Traffic 2000: 1: 399 – 410