Acanthamoeba Myosin IC Colocalizes with Phosphatidylinositol 4,5-Bisphosphate at the Plasma Membrane Due to the High Concentration of Negative Charge*

The tail of Acanthamoeba myosin IC (AMIC) has a basic region (BR), which contains a putative pleckstrin homology (PH) domain, followed by two Gly/Pro/Ala (GPA)-rich regions separated by a Src homology 3 (SH3) domain. Cryoelectron microscopy had shown that the tail is folded back on itself at the junction of BR and GPA1, and nuclear magnetic resonance spectroscopy indicated that the SH3 domain may interact with the putative PH domain. The BR binds to acidic phospholipids, and the GPA region binds to F-actin. We now show that the folded tail does not affect the affinity of AMIC for acidic phospholipids. AMIC binds phosphatidylinositol 4,5-bisphosphate (PIP2) with high affinity (∼1 μm), but binding is not stereospecific. When normalized to net negative charge, AMIC binds with equal affinity to phosphatidylserine (PS) and PIP2. This and other data show that the putative PH domain of AMIC is not a typical PIP2-specific PH domain. We have identified a 13-residue sequence of basic-hydrophobic-basic amino acids within the putative PH domain that may be a major determinant of binding of AMIC to acidic phospholipids. Despite the lack of stereospecificity, AMIC binds 10 times more strongly to vesicles containing 5% PIP2 plus 25% PS than to vesicles containing only 25% PS, suggesting that AMIC may be targeted to PIP2-enriched regions of the plasma membrane. In agreement with this, AMIC colocalizes with PIP2 at dynamic, protrusive regions of the plasma membrane. We discuss the possibility that AMIC binding to PIP2 may initiate the formation of a multiprotein complex at the plasma membrane.

that a short-tailed mouse myosin I, MYO1C, binds PIP 2 with high specificity through a putative PH domain in its tail.
AMIC has been shown to bind to acidic phospholipids (19,20) and to cell membranes (21)(22)(23)(24) through the basic region of its tail (19). In this study, we initially asked if AMIC binds PIP 2 with higher affinity than it binds other acidic phospholipids, if binding to phospholipids is affected by interactions between the basic region and SH3 domain of AMIC, and if the putative PH domain in the basic region is important for the binding of AMIC to acidic phospholipids. In contrast to related studies, we used full-length as well as truncated myosin.
We found that the apparent interactions between the putative PH domain and the SH3 domain revealed by NMR do not affect the affinity of AMIC for acidic phospholipids and that AMIC does not bind preferentially to phospholipid vesicles containing PIP 2 relative to vesicles containing PS when normalized to net negative charge. Consistent with this conclusion, the mutation of an Arg to Ala that blocks the binding of mouse MYO1C to PIP 2 (18) had only minimal effect on the binding of AMIC to PIP 2 . However, we identified a 13-residue basic-hydrophobic-basic sequence in the putative PH domain that may be important for binding AMIC to acidic phospholipid vesicles. Importantly, we found that, despite its lack of stereospecificity for PIP 2 , AMIC concentrates at PIP 2 -enriched regions of the plasma membrane in situ (most likely because of the high concentration of negative charges) at sites of cell protrusion (e.g. pseudopods and phagocytic and pinocytic cups).

MATERIALS AND METHODS
Proteins-Full-length AMIC (FL-AMIC) (25) or T4 (26) heavy chains, with N-terminal FLAG, were expressed together with myosin IC light chain (27) in SF9 cells and purified using anti-FLAG-agarose affinity chromatography. The R779A mutant DNA of FL-AMIC was prepared by site-directed mutagenesis performed by Mutagenex Inc. (Piscataway, NJ) using FL-AMIC DNA as starting material. Purified proteins were dialyzed against myosin I storage buffer (20 mM Tris, pH 7.5, 100 mM KCl, 50% glycerol, 1 mM dithiothreitol) and stored at Ϫ20°C.
Binding Assay-Proteins (40 -100 nM) and vesicles were mixed together in 20 mM imidazole, pH 7.0, 100 mM NaCl, and 1 mM EGTA containing 0.5 mg/ml bovine serum albumin to block nonspecific binding. The mixture was centrifuged for 40 min at 200,000 ϫ g. Aliquots of the supernatants were removed, and the amount of unbound FL-AMIC or T4 was determined by measuring K-ATPase activity (28). Vesicle concentration and pelleted vesicles were monitored by fluorescence.
The K 50 values (concentration of phospholipid required to bind 50% of FL-AMIC or T4) were calculated from assays in which the concentration of the protein was constant and the concentration of phospholipid varied. The values compared in each figure and in the table were obtained with the same batch of phospholipids and with proteins that were purified in parallel and had the same K-ATPase activity. Although absolute values varied up to 2.5-fold when different batches of vesicles or proteins with lower K-ATPase activity were used, the ratios of the K 50 values of FL-AMIC and T4 were the same in all experiments. K 50 is equivalent to K d when, as in these experiments, the phospholipid concentration is greatly in excess of the protein ligand.
Synthetic Peptides and Peptide Competition Assay-Peptides were synthesized and purified by EZbiolab (Westfield, IN). The peptide competition assays were performed with an excess of lipid over protein (i.e. when 100% of protein (T4) was bound). The peptide was then added at increasing concentrations, and the amount of T4 released from the vesicles was measured by K-ATPase activity, as described above for the binding assay. IC 50 is the concentration of peptide required to release 50% of bound T4. The K 50 for peptide binding, which can be compared with the K 50 for T4, was calculated, assuming that peptide and T4 bind to the same sites (composed of one or more phospholipids) on the vesicles, by adaptation of the equation developed by Nikolovska-Coleska et al. (29) for the inhibition of ligandprotein binding, where I 50 is the concentration of free inhibitor at 50% inhibition of binding, L 50 is the concentration of free ligand at 50% inhibition, K d is the dissociation constant of the ligand-protein complex, and P 0 is the concentration of free protein at 0% inhibition. We substituted T4 for L, PS or PIP 2 for P, peptide for I, K 50 (T4) for K d , and K 50 (peptide) for K i and solved for K i using the on-line K i calculator (52).
For immunolocalization microscopy, amoebae were grown on 22-mm diameter, collagen-coated coverslips (BD Biocoat, Bedford, MA) placed in 6-well plates or collagen-coated 8-well cell culture slides (BD Biocoat). Amoebae from rotating cultures were diluted to 2-3 ϫ 10 6 cells/ml, and 2-3 ml of diluted cells were placed in each well and left at 30°C for 2 h or overnight. The cells were incubated in fresh medium for about 10 min and moved to room temperature, and the medium was replaced with fixing solution (1% glutaraldehyde plus 3% formaldehyde in phosphate-buffered saline) for 40 min. This was followed (as described earlier) (30) by three 5-min washes with blocking buffer (10% fetal bovine serum, 0.02% NaN 3 in phosphate-buffered saline), a 1-h incubation with the first antibody in blocking buffer with the addition of 0.2% saponin, three 5-min washes with blocking buffer, a 1-h incubation with the second antibody in blocking buffer with 0.2% saponin, three 5-min washes with blocking buffer, and one rinse with phosphate-buffered saline.
For phagocytosis experiments, yeasts grown in shaking cultures were centrifuged, resuspended in phosphate-buffered saline to final A 600 ϭ 0.6, diluted 10-fold in amoeba growth medium, and added to amoebae grown overnight on coverslips (2-3 ml/well) after removal of old medium. Cells were incubated for 40 min and fixed and stained as described above. For membrane staining, fixation, and antibody staining, procedures were as described above, except the fixation time was reduced to 30 min, and incubation with secondary antibodies was in the absence of saponin. Amoebae were incubated for 15 min with concanavalin A conjugated to Alexa Fluor 594 or CM-DiI (Molecular Probes), diluted 1:150, for 15 min, either before or after the amoebae were fixed and stained with antibodies with similar results. CM-DiI (1:200 dilution) was incubated with fixed amoebae for 20 min. Cells were viewed with a Zeiss LSM 510 confocal microscope with a ϫ63 plan-Achromat 1.4 lens, and slice thickness of 1 m.

RESULTS
We first studied the binding of AMIC to acidic phospholipids in vitro. FL-AMIC ( Fig. 1) and AMIC heavy chain truncated after the basic region (T4; Fig. 1), both with N-terminal FLAG tags, were co-expressed with AMIC light chain (25)(26)(27) in Sf9 insect cells and purified by affinity chromatography (26). T4 contains the entire basic region but not GPA1, GPA2, or the SH3 domain that, as indicated by NMR data (10), interacts with the basic region of FL-AMIC. The myosin constructs were mixed with large unilamellar phospholipid vesicles of different compositions and at different concentrations, and the fraction of bound myosin was calculated from the K-ATPase activity remaining in the supernatant when the vesicles were pelleted (see "Materials and Methods" for experimental details).
FL-AMIC and T4 Bind with the Same Affinity to Acidic Phospholipid Vesicles-The lipid concentration dependences of the binding of FL-AMIC and T4 to vesicles containing either 75% PS or 5% PIP 2 (with neutral PC accounting for the remainder of the phospholipid) are shown in Fig. 2, A and B. The K 50 values (total lipid) were 1.8 M for FL-AMIC and 2.2 M for T4 binding to 75% PS vesicles, and 42 M for binding of both FL-AMIC and T4 to 5% PIP 2 vesicles (Table 1). Although their binding affinities varied with the composition and concentration of the phospholipid vesicles, FL-AMIC and T4 always had similar affinities for vesicles of the same composition and concentration (Table 1). Therefore, at least under the conditions of these experiments, neither the folding of the FL-AMIC tail, shown by cryoelectron microscopy, nor the interaction between the SH3 domain and the basic region, indicated by NMR spectroscopy, affected the binding of AMIC to acidic phospholipid vesicles.
Comparison of Binding to PIP 2 and PS-At low content of acidic phospholipids, PIP 2 vesicles were much better binding partners of both FL-AMIC and T4 than were PS vesicles. For example, the K 50 (total lipid) for binding to T4 was 42 M for 5% PIP 2 vesicles compared with Ͼ4,000 M for 15% PS vesicles (Table 1). Increasing the PS content, however, greatly improved binding, and the K 50 value for binding of T4 to 25% PS vesicles was 90 M (Table 1).
Interestingly, doubling the PS content from 25 to 50% increased the affinity of T4 binding by more than 10-fold (  Table 1, total lipid), The domain structures of the heavy chains, the subdomains of the tails, and the amino acid sequence of the putative PH domain are shown. Residues in red are generally conserved among PH domains, and underlined residues are predicted to form ␤-strands (10); Arg-779, which was mutated to Ala in one experiment (see Fig. 4), is indicated by the red arrow, and the sequence (residues 802-814) corresponding to synthetic peptide BHB-1 (see Fig. 5) is in boldface type.
but further increase of the PS content to 75% had a much smaller effect (K 50 ϳ 2 M, Table 1). The addition of 5% PIP 2 to 25% PS vesicles had a similar effect as doubling the PS content ( Fig. 2C and Table 1). These results suggest that strong binding of AMIC to PS vesicles requires a minimum charge density and that a molecule of AMIC may bind to more than one PS head group. The dependence of binding on the content of acidic phospholipid and the equivalence of the affinities for PIP 2 and PS vesicles, when normalized for net negative charge, Ϫ1 for PS and Ϫ4 for PIP 2 (31), are well illustrated in the experiments shown in Fig. 3, in which the total phospholipid concentration was kept constant and the phospholipid composition was varied. The lower curve in Fig. 3A shows that vesicles containing 5-15% PS had essentially the same low affinity for FL-AMIC with an increase in binding when the PS content was increased to 75%.
The importance of charge density is supported by the affinity of FL-AMIC with vesicles containing 1-10% PIP 2 in addition to 25% PS (Fig. 3A, upper curve). Although the addition of PIP 2 to 25% PS vesicles caused a much greater increase in affinity than an equal molar percent increase in PS concentration (Fig. 3A, lower curve), this difference can be fully explained by the higher negative charge of PIP 2 . Corrected for the 4-fold difference in negative charge between PIP 2 and PS (31), the two curves are essentially identical (Fig. 3B). In agreement with this, the K 50 for binding of FL-AMIC and T4 to vesicles containing 5% PIP 2 plus 25% PS was the same as for vesicles containing 50% PS and more than 10 times higher than for vesicles containing only 25% PS ( Table 1). The last result implies that in vivo AMIC may preferentially bind to regions of membranes enriched in PIP 2 (see below).
When calculated on the basis of accessible acidic phospholipid, the K 50 for binding of 5% PIP 2 vesicles to both FL-AMIC and T4 was 1 M (Table 1, acidic lipid only), which is a typical affinity for proteins with a PIP 2 -specific PH domain (15)(16)(17). However, as discussed above, vesicles containing 50% PS and no The percentage of acidic phospholipid lipid was kept constant, and the concentration of total lipid was varied. Note that the scales of the x axes differ in different panels. Binding assays were performed as described under "Materials and Methods." FL-AMIC and T4 concentrations were 50 nM. In this and all figures, a percentage sign represents mol %. Each experiment in this and subsequent figures is representative of at least two independent experiments.  K 50 values for binding of FL-AMIC and T4 to phospholipid vesicles with different phospholipid compositions were determined as described under "Materials and Methods" and illustrated in Fig. 2. The total lipid concentration was varied, and the concentration of FL-AMIC and T4 was constant at 50 nM. K 50 is the concentration of lipid at which 50% of FL-AMIC or T4 was bound. Total lipid is the concentration of total lipid in the assay (acidic and neutral phospholipids), and accessible lipid is half of the total lipid, since only half of the phospholipid is in the outer layer of the vesicle and available for myosin binding. Accessible acidic lipid is the accessible acidic phospholipid: PS, PIP 2 , or PIP 3 . The K 50 values for T4 and FL-AMIC binding to vesicles of the same composition were obtained in the same assay with the same batch of vesicles and proteins that were purified in parallel and had similar K-ATPase activities. Each value is representative of at least two experiments. NA, not applicable because two different acidic lipids were present.
Although PH domains generally have low sequence homology, several highly conserved basic residues are required for high affinity binding of PH domain proteins to PIP 2 vesicles (34). For example, mutation of Arg 903 in mouse MYO1C to Ala decreased the affinity of MYO1C for PIP 2 by Ͼ Ͼ8-fold (18). We found that the homologous mutation of FL-AMIC, R779A, caused only a 2-4-fold reduction in the binding of T4 to 50% PS vesicles (Fig. 4) or to 5% PIP 2 plus 25% PS vesicles (not shown). Although this may be a greater decrease in affinity than would be expected from the decrease in net positive charge of the basic region (from ϩ15 to ϩ14), it is a much smaller decrease than expected for a PIP 2 -specific PH domain (34). Therefore, although Arg 779 may contribute to the binding specificity of AMIC, it is not a major determinant, and this region of the putative PH domain most likely does not play a major role in binding of FL-AMIC to acidic phospholipids.
The AMIC Basic Region Contains a Basic-Hydrophobic-Basic (BHB) Sequence That Blocks Binding of T4 to Acidic Phospholipid Vesicles-Some proteins, including MARCKS and several cytoskeletal and signaling proteins (35) (for reviews, see Refs. 36 -38), bind to acidic phospholipids through a region consisting of two basic sequences separated by a hydrophobic sequence. The putative PH domain of the tail of AMIC has such a sequence, 802 KKVKPFLYVLKRR 814 (Fig. 1), the first five residues of which are in the long loop between the third and fourth ␤-strands (10). We found that a synthetic peptide, BHB-1, with this sequence inhibited the binding of T4 to vesicles containing 50% PS, 5% PIP 2 plus 25% PS, or 6% PIP 2 (Fig. 5, A-C) with IC 50 values (concentration of peptide required for 50% inhibition of T4 binding) of 16, 15, and 40 M, respectively. The IC 50 values for binding of BHB-1 to 50% PS vesicles and 6% PIP 2 vesicles correspond to K 50 values for accessible acidic phospholipid of 0.3 and 5 M, respectively (see "Materials and Methods"), which are similar to the K 50 values for binding of FL-AMIC and T4 (Table 1). A peptide with Ala replacing the middle six hydrophobic residues of BHB-1, KKVKAAAAAAKRR (BHB-2), was very much less inhibitory of T4 binding (Fig. 5, A-C). Replacement of two different pairs of hydrophobic residues, KKVK-PAAYVLKRR (BHB-3) and KKVKPFLYAAKRR (BHB-4), had a smaller but still substantial effect on the peptide's ability to  inhibit T4 binding (Fig. 5D). These results suggest that residues 802-814 in the AMIC tail may be a major determinant of binding of AMIC to acidic phospholipid vesicles and that both hydrophobic and electrostatic interactions may be involved.

Inhibition of T4 Binding by BHB Regions in Tails of Other Amoeba
Class-I Myosins-AMIA, AMIB, and Dictyostelium myosins IA, IB, IC, and ID (DMIA-DMID) have BHB regions in their tails similar to the BHB region of AMIC (Fig. 6A). 4 By pairwise comparison, the BHB regions of DMID and AMIB are most similar to the BHB region of AMIC, and DMIB, DMIC, AMIA, and DMIA are less similar, in that order. Notably, the last three contain an acidic residue adjacent to the basic residues.
We tested the ability of synthetic peptides corresponding to the BHB regions of DMID, AMIB, and AMIA to inhibit binding of T4 to phospholipid vesicles containing either 50% PS or 6% PIP 2 (Fig. 6, B and C). The peptides corresponding to the BHB regions of DMID and AMIB were effective inhibitors of T4 binding to both PS and PIP 2 vesicles, the DMID peptide being substantially more inhibitory and the AMIB peptide slightly less inhibitory than the corresponding peptide from AMIC. In stark contrast, the peptide corresponding to the BHB region of AMIA had essentially no inhibitory effect on T4 binding. These results support the idea that the BHB region of AMIC (and of other, but not all, long-tail class-I myosins) has a significant role in binding to acidic phospholipid vesicles and that the binding affinity can be modulated by relatively small differences in the sequences of the BHB region.  thought to be high (ϳ25%) and uniform, whereas there is much less PIP 2 (ϳ1%). However, the amount and localization of PIP 2 can be regulated, thus allowing recruitment of PIP 2 -specific binding proteins to spatially and functionally distinct regions of plasma membranes (for reviews, see Refs. 11, 12, and 31). The characteristics of binding of AMIC to acidic phospholipid vesicles described above, especially the fact that FL-AMIC bound 10 times more strongly to vesicles containing 25% PS and 5% PIP 2 than to vesicles containing only 25% PS, suggested that AMIC might be targeted to PIP 2 -enriched regions of the Acanthamoeba plasma membrane, despite a lack of specificity for PIP 2 , because of the high negative charge density of such regions.

Colocalization of Endogenous AMIC and PIP 2 in Plasma Membranes of Acanthamoeba castellanii-The content and distribution of PS in plasma membranes of live cells is generally
We used amoebae grown on a collagen-coated surface and localized endogenous AMIC and endogenous PIP 2 by immunofluorescence confocal microscopy of fixed cells stained with AMIC-specific antibodies (22) and either of two commercial PIP 2 -specific antibodies (see "Materials and Methods") with similar results. Because most of the PIP 2 in Acanthamoeba (39), as in other cells (for a review, see Ref. 16), and almost all of AMIC (22,23) resides on the plasma membrane, we adjusted conditions to optimize immunostaining of the plasma membrane (21). When harsher staining conditions were used, AMIC was also seen associated with contractile vacuole membranes (not shown) (22), but plasma membrane staining was disturbed under these conditions, as previously observed (21,22).
Colocalization of AMIC and PIP 2 in pseudopods in several typical cells is shown in Fig. 7 (row A). Most of the regions enriched in PIP 2 and AMIC also contained F-actin, but F-actin frequently occurred in regions that did not contain either AMIC or PIP 2 . Occasionally, we also observed punctated PIP 2 staining, mostly on the ventral surface (Fig. 7B), which was only sporadically associated with AMIC staining. We did not char-acterize this punctated staining any further. In agreement with earlier observations (22,23), we also saw relatively uniform AMIC staining of plasma membranes in regions that did not stain for PIP 2 (Fig. 7, rows B and C); the staining of AMIC was weaker in these regions than in PIP 2 -enriched regions.
AMIC and PIP 2 colocalized in both broad (Fig. 7, row B) and small (Fig. 7, row C) pseudopods but not necessarily in all pseudopods when there were more than one pseudopod in a single cell. For example, the cell in Fig. 7 (row B) has three pseudopods, of which only two, and principally mostly one, stained positively for AMIC and PIP 2 . The presence, individually, of AMIC (40) and PIP 2 (39) in broad pseudopods were reported before, but here we show for the first time that AMIC and PIP 2 colocalize in the same pseudopods. Presumably, AMIC and PIP2 colocalized in the pseudopods that were most active at the time the cells were fixed. For example, the small pseudopod in Fig. 7 (row C) that stained positively for both AMIC and PIP 2 may have been a newly formed pseudopod, and, in contrast to broad pseudopods, it was not adherent to the surface (not shown). Fig. 8 (row A) shows the colocalization of AMIC and PIP 2 in a nascent (pinocytic?) cup and the absence of both AMIC and PIP 2 in filopodia. Fig. 8 (rows B and C) shows cells reaching toward (row B) and beginning to phagocytose (row C) yeast. In both cells, AMIC and PIP 2 are colocalized at the base of the phagocytic cup but absent from extensions from the cups and from filopodia. Similarly, AMIC and PIP 2 colocalized at the base of, but not in, filopodia extending from the broad, thin   A, B, C) and at the base of filopodia (D). Amoebae were stained for AMIC (red) and PIP 2 (green). Gray panels are differential interference contrast images. Bar, 10 m. pseudopod seen in Fig. 8 (row D). Both AMIC and PIP 2 were also absent from phagocytic vesicles after the yeasts were internalized (not shown). All of the images in Figs. 7 and 8 strongly suggest that AMIC and PIP 2 colocalize at the plasma membrane at regions of enhanced membrane activity.

Enrichment of AMIC and PIP 2 at the Plasma Membrane Is Not an Artifact Caused by Localized Membrane Accumulation-
To eliminate the possibility that the apparent colocalization of AMIC and PIP 2 was an optical artifact caused by overlapping of membranes that contained independently bound AMIC and PIP 2 , we stained amoebae with two membrane-targeted dyes, CM-Dil and concanavalin A, in addition to the AMIC and PIP 2 antibodies. The conditions were slightly modified to allow simultaneous staining with CM-Dil or concanavalin A, and AMIC and PIP 2 antibodies (see "Materials and Methods"). As illustrated by the examples in Fig. 9, there was no significant evidence for an increase in membrane concentration in the regions where AMIC and PIP 2 were colocalized.

DISCUSSION
This study began with a few straightforward questions. The tail of AMIC comprises a basic region, followed by two Gly/Pro/ Ala-rich regions separated by an SH3 domain (6 -8). Cryoelectron microscopy (9) and NMR spectroscopy and homology modeling (10) had shown that the tail was folded back between the BR and GPA1 with interactions between SH3 and a putative PH domain within the BR. Experiments with bacterially expressed peptides (19) had shown that the basic region (and only the basic region) of the tail bound to acidic phospholipid vesicles and with an affinity equivalent to the binding of fulllength AMIC. The questions we asked were as follows. Do interactions between the SH3 domain and BR in the folded tail affect the binding of AMIC to acidic phospholipids? Is binding a function of the putative PH domain? Is binding specific for PIP 2 ?
The data in this paper show that full-length AMIC binds to acidic phospholipids of various compositions and concentrations with the same affinity as tail-truncated AMIC, retaining only the basic region. These results are consistent with the evidence from experiments with expressed peptides that binding of AMIC to acidic phospholipids is driven solely by the basic region in the tail (19). Additionally, our data show that the folded tail of FL-AMIC and the interactions between the SH3 domain and the putative PH domain do not inhibit binding of FL-AMIC to acidic phospholipids, for which there are at least two possible explanations. Under our experimental conditions, the tail of AMIC could remain folded, and acidic phospholipids bind to residues that are not involved in the intratail interactions. Alternatively, the tail may unfold when AMIC binds to phospholipid vesicles, but the energy required for unfolding might be too small to affect the phospholipid-binding assay. However, even if the folded tail has no effect or minimal effect on the binding of AMIC to acidic phospholipid vesicles (and, by inference, to membranes), it is possible that the intratail interaction between the SH3 domain and the basic region regulates the interaction of the SH3 domain with its binding partners.
The small fraction of PH domain proteins that are known to bind specifically to PIP 2 contain a sequence, KX n (K/R)XR, in the loop between the first two ␤-strands (41). A similar, but not identical, sequence, 769 KSFWGSKVERR 779 , occurs in that position in the putative PH domain of AMIC (Fig. 1), and AMIC binds PIP 2 with an affinity similar to the affinities of proteins that contain a PIP 2 -specific PH domain (16). However, when normalized for net negative charge, AMIC had the same affinity for PS vesicles and PIP 2 vesicles, the mutation R779A had minimal effect on the binding to both PS vesicles and PIP 2 vesicles, and soluble inositol phosphate head groups did not inhibit binding of AMIC to acidic phospholipid vesicles. Thus, the putative PH domain of AMIC, like a majority of PH domains (15), does not exhibit the stereospecificity characteristic of PIP 2 -specific PH domains. Although the K 50 values (ϳ1 M) for binding of FL-AMIC and T4 to PS vesicles and PIP 2 vesicles are similar to the apparent K d values for high affinity binding of other proteins to phospholipid vesicles, the apparent K d values calculated by Doberstein and Pollard (19) for binding of AMIC tail peptides are very much lower (ϳ5 nM). This difference may be due to differences in assay methods or to differences in the binding of the proteins (FL-AMIC and T4) used in our assays and the bacterially expressed ␤-galactosidase fusion peptides, used by Doberstein and Pollard (19).
In addition to PH domains, at least two other motifs can be the basis for binding of proteins to membrane phospholipids: a cluster of basic amino acids, or two polybasic clusters separated by a hydrophobic region (35) (for reviews, see Refs. 36 -38). The AMIC putative PH domain contains the latter motif, 802 KKVK-PFLYVLKRR 814 , and a peptide with this sequence, BHB-1, specifically blocked the binding of AMIC to acidic phospholipid vesicles. Importantly, mutations of two or more of the hydrophobic residues to Ala greatly reduced the ability of BIB-1 to inhibit binding of AMIC to acidic phospholipids. Thus, this region is a reasonable candidate for the phospholipid/membrane-binding domain of AMIC.
The possible role of the BHB region in the binding of AMIC to acidic phospholipids is supported by the experiments with synthetic peptides corresponding to the BHB regions of DMID, AMIB, and AMIA. Peptides corresponding to the BHB regions of DMID and AMIB, which are most similar to the BHB region of AMIC, inhibited binding of T4 to acidic phospholipids as effectively as the AMIC peptide. In contrast, the peptide corre- sponding to the AMIA BHB region, which has less similarity to AMIC, did not inhibit T4 binding, possibly because of the Asp residue within its hydrophobic core.
Interestingly, the abilities of the BHB peptides derived from the Acanthamoeba class-I myosins to inhibit binding of AMIC to acidic phospholipids correlate with the associations of the myosins with the plasma membrane in vivo. AMIC and AMIB are mostly localized on the plasma membrane, and AMIA is mostly cytosolic (21)(22)(23). Also, a significant fraction of DMID is associated with the plasma membrane of Dictyostelium amoebae, but this association varies during cell development (42)(43)(44).
By analogy to the peptides, binding of AMIC may involve nonstereospecific, electrostatic interactions between the basic side chains of the BHB region, which reside in a flexible loop (10), and acidic phospholipid heads, followed by hydrophobic interactions in which the hydrophobic side chains are inserted into the phospholipid bilayer. The residues that are obvious candidates for such insertion are strongly hydrophobic nonpolar residues, Val, Leu, and Ile, but aromatic residues may also be important, since they are present in the three BHB peptides that inhibit AMIC binding to acidic phospholipids and not in the BHB peptide (AMIA) that does not.
However, since replacing only two hydrophobic residues in the AMIC peptide BHB-1 greatly reduced its ability to inhibit binding of AMIC to acidic phospholipid vesicles, it is difficult to predict the binding properties of the myosins based only on the sequence of this region. Even if this region is the principal basis for the localization of these myosins in vivo, other sites must also be involved, because, for example, AMIB and AMIC have overlapping, but different and dynamic, localizations in the amoebae (22)(23)(24). Similarly, although the tails of both mouse MYO1C and AMIC have a basic region containing a putative PH domain, the two myosins bind to acidic phospholipid vesicles by different mechanisms: MYO1C by stereospecific binding to PIP 2 through the putative PH domain (17,18) and AMIC by nonstereospecific interactions, possibly both electrostatic and hydrophobic, involving the BHB region (this paper).
Previously, Baines and Korn (21) had shown, by immunofluorescence and immunogold electron microscopy, the close association of AMIC with the plasma membrane and contractile vacuole membrane. Ostap et al. (24) showed the dynamic localization of AMIC in macropinocytic and phagocytic cups and pseudopods by immunofluorescence of living and fixed cells, and Kong and Pollard (40) observed that expressed enhanced green fluorescent protein-myosin IC transiently concentrated around macropinocytic cups and contractile vacuoles. We have confirmed the localization of AMIC to pseudopods and endocytic cups and, importantly, find that PIP 2 colocalizes with AMIC in these regions. Although, to our knowledge, this is the first report of PIP 2 localization with pinocytic and phagocytic cups in amoebae, PIP 2 has been shown to be present in phagocytic cups in other organisms (for a review, see Ref. 45).
But which comes first, AMIC or PIP 2 ? It would be consistent with our in vitro data for AMIC to bind to regions in the plasma membrane where PIP 2 had previously clustered. But it is also possible that the basic residues in the BHB region in the PH domain might cluster PIP 2 , as do similar regions in other proteins (46,47), after initial binding of AMIC to the plasma membranes.
The mechanism of association of the long-tailed amoeba myosins with the plasma membrane becomes even more complicated when one considers that the scaffolding protein Acan125/CARMIL (48,49), which binds to the SH3 domain of long-tailed class-I myosins, colocalizes with these myosins at dynamic, protrusive regions of Acanthamoeba and Dictyostelium similar to those where we observed colocalization of AMIC and PIP 2 (48,49). Acan125/CARMIL also binds G-actin, actin-capping protein, and the Arp2/3 complex, all of which have been shown to localize to regions of cell protrusion (49,50) similar to those at which we observe AMIC, F-actin, and PIP 2 . One, but not the only, possible sequence of events for colocalization of these multiple proteins in Acanthamoeba and Dictyostelium would begin with a local increase in concentration of PIP 2 , to which the long-tailed class-I myosins would bind. F-actin could then bind to the ATP-independent F-actin binding site in the myosin I tail, and Acan125/CARMIL could bind to the SH3 domain of myosin I. Acan125/CARMIL, with its binding sites for G-actin, Arp2/3, and capping protein, could initiate actin polymerization at these dynamic regions.
This speculative sequence of events is compatible with the observation of Novak and Titus (51) that the SH3 domain of myosin I is required for myosin I function in vivo but is not required for the proper membrane localization of myosin I in Dictyostelium myosin I null cells (i.e. association with Acan125/ CARMIL may be required for myosin I activity but is not required for binding myosin I to the plasma membrane). The apparently contradictory observation by Kong and Pollard (40) that proper membrane localization of transiently expressed AMIC in wild-type Acanthamoeba required the SH3 and head/ neck domains, in addition to the basic region of the tail, might be explained by the necessity, in those experiments, for the expressed AMIC to compete with endogenous myosin I, whose membrane association might be stabilized by multiple interactions subsequent to its initial binding to the membrane.