Abstracts

Speaker Abstracts

1. A Structuralist's View of Muscle Contraction K.C. HOLMES, Max-Planck Institute for Medical Research, Heidelberg, Germany

The structures of a truncated form (without the lever arm) of Dictyostelium discoideum (slime mold) myosin 2 motor domain have been elucidated by Rayment's lab with a variety of nucleotide analogues bound in the active site. Dominguez et al. have obtained similar results with smooth muscle myosin (for review see Geeves and Holmes. 1999. Annu. Rev. Biochem. 68:687–728). These studies reveal the occurrence of two conformations of myosin (OPEN and CLOSED active site) entailing a movement of 5–6 Å of the switch 2 region. This movement is coupled with a large movement of the converter domain (a 60° rotation). CLOSED may only be stable in the presence of the γ-phosphate. Truncated Dictyostelium myosin 2 with ADP.BeF₃ in the active site was found by Rayment et al. to be in the OPEN form. We have solved a truncated Dictyostelium myosin 2 with ADP.BeF₃ in the CLOSED form. A comparison of these very similar constructs in two different conformations allows an analysis of the conformational change CLOSED to OPEN. The converter domain rotates (with no translational component) about an axis passing close to residue E476 (Dictyostelium sequence) in the switch 2 (or relay) helix (an analysis of chicken smooth muscle compared with chicken skeletal yields very similar results). In the complex with actin, this rotation axis lies at right angles to the actin helix. On closing the active site, the switch 2 helix breaks at residue K488 and forms a kinked helix, with the distal end rotated by one hydrogen bond (100°). The converter domain is firmly attached to the distal end of the switch 2 helix and is thus rotated through ∼60°. During this transition, a number of side chains (e.g., F487 and W501), move from solvent-exposed to buried environments. The distal end of the lever arm would move ∼110 Å in an axial direction in response to the opening and closing of the active site. Myosin seems to transport actin by switching between these two states.

The polymorphism of myosin is in fact richer. Myosin binds to the actin filament in two distinct ways, termed weak and strong. The initial binding is weak: weak binding isomerizes to strong. This isomerization is connected with release of products of hydrolysis. The binding of ATP restores weak binding and the release of the cross bridge from the actin filament. Electron microscopic reconstructions at ∼18 Å resolution from our lab show that on strong binding to actin the deep cleft in the myosin cross bridge shuts, as suggested by Rayment et al., who noted that the hydrophobic surfaces of the cleft would favor closure. Studies from Chris Berger's lab using custom-made tryptophan mutants support this view. Thus, it appears that the cleft is held open in the myosin structure and normally shuts only on binding to actin. This shut conformation has yet to be revealed by crystallography.

We can speculate how combining these pairs of myosin states (CLEFT-OPEN, CLEFT-SHUT and OPEN, CLOSED) move a muscle. Our conjecture is that the shutting of the cleft on binding to actin triggers the opening of the active site by moving the switch1 loop. This somehow facilitates phosphate release (status CLEFT-SHUT, active site CLOSED, γ-phosphate gone). This in turn initiates the lever arm swing since the CLOSED state is unstable without a γ-phosphate. At the end of the power stroke, the combination of OPEN (active site) and CLEFT-SHUT (strong binding to actin) would appear to facilitate ADP release by opening the nucleotide binding pocket. Rebinding of ATP would cause CLEFT-OPEN (weak binding) and fast release from actin.

2. Moving Beyond Myosin STEVEN M. BLOCK, Departments of Biological Sciences and Applied Physics, Stanford University, Stanford, CA

The field of biomolecular motors is a curious world. Movement is fundamental to life, and yet there is much that remains mysterious about the molecular mechanisms responsible for generating motion from chemical energy. Myosin, the mother of all motor proteins, has been a subject of experimental study now for more than a century, and intensely so for the past few decades. Victory in our understanding has been proudly declared at several junctures. Has all this talk been premature? Are we just deluding ourselves? Perhaps more importantly, can any principles gleaned from the study of myosin be generalized to other sources of motion? Movement, after all, abounds in nature. There are plenty of classical mechanoenzymes to consider, such as the members of the kinesin and dynein superfamilies. And there is no shortage of other important proteins that generate motion, including the nucleic acid enzymes (polymerases, helicases, topoisomerases, nucleases, etc.), ATP synthases, flagellar motors, pumps and transporters, and so on. Can the study of any of these biomolecular systems inform our understanding of myosin, or vice versa? This talk will discuss our studies of nonmyosin motility, and utterly fail to do justice to any of these vital questions. (Supported by grants from NIH.)

3. Interaction of Mammalian Class I Myosins with Actin and Nucleotide MICHAEL A. GEEVES,¹ ABEL W. LIN,² CHRIS ARTHUR,² RONALD A. MILLIGAN,² JUSTIN E. MOLLOY,³ and LYNNE M. COLUCCIO,^{4 1} University of Kent at Canterbury; ² Scripps Research Institute, ³ University of York; ⁴ Boston Biomedical Research Institute (Sponsor: H. Lee Sweeney)

Class I myosins are actin-based molecular motors believed to be involved in motile events in the cell. Several class I myosins exist in higher cells. To assist in determining whether these isoforms have unique or similar roles, we have been investigating the properties of class I myosins from rat liver. Using steady-state and transient kinetic analyses, we have recently observed that rat liver 130-kD myosin I (also known as MI130, MYR 1, or MM1a) interacts with nucleotide and actin in much the same way qualitatively as other myosins; but, it is much slower. Also, acto.MI130 has a high affinity for ADP. These results led us to propose that MI130 is designed for efficient tension maintenance (Coluccio and Geeves. 1999. J. Biol. Chem. 274:21575–21580). Furthermore, we showed in single-molecule experiments that, unlike skeletal muscle myosin II, MI130 interacts with actin in two-steps (Veigel et al., 1999. Nature. 398:530–533). The second step may be a consequence of the high affinity of acto.MI130 for ADP.

Here, we examine MI110, a myr 2 (or MM1b) gene product, which is also widely expressed. The steady-state Mg²⁺-ATPase activity of MI110 is activated in Ca²⁺. Purified rat liver MI110 translocates actin filaments in vitro and, unlike MI130, the rate of translocation is greater at pCa 4 than at pCa 7. In transient kinetic studies, we have observed that the major phase of the ATP-induced dissociation of actin-MI110 (K₁k₊₂ = 0.0035 μM⁻¹s⁻¹) is due largely to the smaller value of the maximal observed rate, k₊₂=2 s⁻¹. This is 10 times slower than MI130 and places MI110 among the slowest myosins so far examined; 1/K₁ is comparable to that of other myosins. The affinity of ADP for A.MI110 is ∼5 μM and Ca²⁺ insensitive. The 3-D structure of actin filaments decorated with MI110 under conditions of rigor or in the presence of ADP shows an ADP-induced conformational change, a characteristic of actomyosin complexes having a high affinity for ADP. Results from optical trapping of this myosin I as it interacts with actin indicate that MI110 exhibits a two-step power stroke similar to that exhibited by MI130.

The study of these myosins has resulted in a new view of how motors work. Furthermore, they indicate that although these two myosins share some similar characteristics, they possess some unique mechanochemical properties which might permit them to carry out distinct functions. (Supported by NIH, Royal Society, and the Wellcome Trust.)

4. Three Dimensional Structures of Kinesin Family Motors Complexed to Microtubules KEIKO HIROSE,¹ YING WU,¹ TOSHIHIKO AKIBA,² MARIA ALONSO,³ ROBERT A. CROSS,³ XIAOHUA ZHANG,⁴ SHARYN ENDOW,⁴ and LINDA A. AMOS,^{5 1} Gene Discovery Research Center and ² National Institute of Advanced Interdisciplinary Research, National Institute of Advanced Industrial Science and Technology, Tsukuba, Ibaraki, Japan; ³ Marie Curie Research Institute, Oxted, Surrey, UK; ⁴ Duke University Medical Center, Durham, NC; ⁵ MRC Laboratory of Molecular Biology, Cambridge, UK

The kinesin family includes various proteins that differ in their motor properties, (e.g., the directions of movement, velocity, and processivity). Comparison of the 3-D structures of these proteins would tell us how a particular structural change is related to their motor functions. Using cryo-electron microscopy and 3-D image reconstruction, we have been investigating the structures of various kinesin family motor proteins complexed to microtubules in the presence and absence of nucleotides.

When microtubules were fully decorated with dimeric constructs of motors, we saw only one of the two heads directly attached to the microtubule. The largest difference between different nucleotide states and different motors were seen in the positions of the free heads relative to the bound head. In the presence of AMPPNP, all the plus end–directed motors so far studied (rat kinesin, Neurospora kinesin (Nkin), Eg5, and a chimera formed of ncd heads attached to Nkin necks) had the free heads tilted to the plus end, whereas the free head of ncd was pointing toward the minus end. Preliminary results suggest that a homodimer of Kar3, a minus end–directed motor, has the free head in a similar position to ncd. In the absence of nucleotides, two kinesins, which are highly processive, showed a large change in the position of the free heads, whereas the free heads of other, less processive motors moved relatively little. Changes in the directly attached heads are subtle for all the types of kinesins we have investigated. However, detailed studies of a monomeric construct of Kar3 show there are distinct changes in the structure of the attached head when different nucleotides are bound. (Supported by Human Frontier Science program.)

5. Kinesin Motors: Directionality of Movement and Structural Changes Related to ATPase Activity SHARYN A. ENDOW,¹ HIDEO HIGUCHI,² MIKYUNG YUN,³ XIAOHUA ZHANG,¹ CHEON-GIL PARK,³ and HEE-WON PARK,^{4 1} Duke University Medical Center, Durham, NC; ² Tohoku University, Sendai, Japan; ³ St. Jude Children's Research Hospital, Memphis, TN

The kinesin motors move unidirectionally along microtubules, using the energy from ATP hydrolysis to perform work. Motor movement involves conformational changes that bias or determine motor directionality, together with structural changes associated with ATP hydrolysis that enable translocation along the filament. The conformational changes that occur in the kinesin motors are largely undefined at present, but identifying these changes and how they are coupled to one another will be critical to understanding motor function.

We are investigating the conformational changes that underlie motor directionality by analyzing mutants using microtubule gliding assays and single-motor laser-trap assays. We showed recently that a point mutation in the neck of the Ncd motor protein causes the motor to move toward either the plus or minus end of the microtubule, rather than unidirectionally toward the minus end. Single-motor analysis showed that a displacement, or working stroke, of wild-type Ncd occurs upon binding of the motor to the microtubule and is biased toward the minus end, whereas that of the mutant motor occurs in either direction. The displacements occurred as single events, demonstrating that Ncd is a nonprocessive motor. The mutant analysis indicates that the directional bias of the working stroke is dependent on neck–motor core interactions. These findings explain the minus-end directionality of wild-type Ncd, implying that directionality is caused by a movement that positions the stalk-neck relative to the motor core when the motor binds to the microtubule.

We are also using mutants to detect conformational changes in the kinesin motors required for ATP hydrolysis and movement along microtubules. The crystal structures of the kinesins to date are all in the same state, with bound Mg·ADP, and, thus, have not revealed these changes, but it may be possible to trap the motors in normally transient conformations by mutating residues that stabilize the ADP state or destabilize other conformations. We have analyzed three point mutants that affect highly conserved or invariant residues of the motor domain. The three mutants block the microtubule-activated ATPase of the motor but do not block the basal ATPase, and bind more weakly or more tightly to microtubules than wild type, indicating that they represent different states of the motor. We have solved the crystal structures of the three mutants, together with a high resolution structure of wild-type Kar3. Two of the mutants show striking changes compared with wild type, including a disordered switch I loop and partially melted switch I helix. The structural changes define a signaling pathway within the motor for activation of its ATPase. The movements alter interactions of residues with the bound Mg²⁺, which could destabilize the Mg²⁺ and accelerate ADP release to activate the motor ATPase. (Supported by NIH, Human Frontiers Science Program, Japan Ministry of Education, Science, Sport and Culture, American Heart Association, and American Lebanese Syrian Associated Charities.)

6. Single Molecule Studies of Myosin I ALEX E. KNIGHT,¹ GREGORY I. MASHANOV,¹ CLAUDIA VEIGEL,¹ LYNNE M. COLUCCIO,² JOHN F. ECCLESTON,³ MICHELLE PECKHAM,⁴ and JUSTIN E. MOLLOY,^{1 1} Biology Department, University of York, York, UK; ² Boston Biomedical Research Institute, Boston, MA; ³ National Institute for Medical Research, London, UK; ⁴ School of Biomedical Sciences, Worsley Building, The University of Leeds, Leeds, UK

We have used single molecule techniques to study the mechanism of force production by myosin I. Myosin I's are ideal for investigating the myosin motor mechanism, because of their monomeric structure, slow biochemical kinetics, and long, two-phase power stroke.

Using optical tweezers, we have made mechanical recordings from individual myosin-I molecules. We found that myr2 (with 3 IQ motifs) has an overall working stroke of ∼4 nm compared with 11nm for myr1 (with 6 IQ's). To analyze these mechanical recordings, we developed methods based on Page's test, which we found gave better timing accuracy and robustness of event detection than other widely used methods (Knight et al. 2001. Progr. Biophys. Mol. Biol. In press).

Using TIRFM (Total Internal Reflection Fluorescence Microscopy), we observed the binding and release of single molecules of Cy3-EDA-ATP/ADP by myosin I molecules on a surface. We found that the ADP release rate of brush border myosin I (BBMI), under similar conditions to the mechanical experiments above, is 0.16 s⁻¹. We are combining this method with optical tweezers to perform correlated single molecule measurements to test if the second power stroke of myosin I correlates with ADP release.

By observing fluorescent actin filaments as they swivel on individual myosins, we found the torsional stiffness of BBMI to be 31.0 ± 4.7pN·nm·rad⁻¹ (corresponding to an r.m.s. rotation of ∼21°). From these numbers, it is likely that a myosin molecule would be able to bind to an actin filament even if it were presented orthogonally to the axis of the myosin molecule. This finding is important in the context of the living cell, in single molecule experiments and motility assays. (Supported by the Wellcome Trust, BBSRC, and the Royal Society.)

7. The Motor Domain, Not a Long Neck Domain, Determines the Large Step of Myosin V. T. YANAGIDA,^1,2 H. TANAKA,² K. HOMMA,³ A. H. IWANE,¹ E. KATAYAMA,⁴ R. IKEBE,³ J. SAITO,³ and M. IKEBE,^{3 1} Osaka University Graduate School of Medicine, Suita, Osaka 565–0871, Japan; ² Single Molecule Process Project ICORP, JST; ³ University of Massachusetts Medical School; ⁴ Tokyo University, Tokyo, Japan

Myosin V is an unconventional myosin that processively moves along an actin filament. Myosin V has two heads, each of which consists of a motor domain and an expanded neck domain (23 nm long) that contains six IQ motifs. Based on a lever arm model, myosin V is postulated to stride along the helical repeat (36 nm) of an actin filament by tilting the long neck domain (lever arm) of one head and leading the partner head to the neighbor helical pitch. Optical trapping nanometry has actually shown large and processive steps of −36 nm. Here, we measure the mechanical properties of a deletion mutant of myosin V with neck domain consisting of only one IQ motif, which is too short to allow two heads to span the helical pitch of an actin filament. The results show that the size and processivity of steps, and the sliding velocity are all similar to those of myosin V fully equipped with all six IQ motifs. Thus, the long extended neck domain is not essential for myosin V's large processive steps, suggesting that the motor domain is responsible for the unique operation of myosin V.

8. Kinesin Molecular Motors: Transport Pathways, Receptors, and Human Disease L.S.B. GOLDSTEIN, Howard Hughes Medical Institute, Department of Cellular and Molecular Medicine, University of California at San Diego School of Medicine, La Jolla, CA

Neurons are large, highly polarized cells that depend upon kinesin and dynein molecular motor proteins to generate and support their functions. Considerable work is beginning to hint at the possibility that these neuronal transport systems are “Achilles Heels” of the biology of these cells—particularly susceptible to damage or insult that can cause neurodegeneration. We recently discovered two proteins that are strong candidates for “receptor” molecules linking conventional kinesin, kinesin-I to axonal vesicular cargoes and that provide clues to the possible relationship of neuronal transport systems to neuronal damage and disease. The first candidate, identified by the sunday driver mutation (syd) in Drosophila links the kinesin light chain subunit of kinesin-I to an unknown class of axonal vesicles. Intriguingly, syd has been independently identified as a protein called JIP3 that serves a scaffold function for JNK/MAP kinase signaling modules. Thus, jip3/syd may serve as a link between neuronal transport and neuronal damage signaling pathways. The second candidate receptor protein is the amyloid precursor protein (APP), which plays a key role in the pathogenesis of Alzheimer's disease. We discovered a biochemical interaction between APP and the kinesin light chain subunit of kinesin-I, and have strong in vivo evidence that kinesin light chain is required for axonal transport of APP. We also identified a vesicular axonal transport pathway in the mouse peripheral nervous system that depends upon APP function and have identified the proteins whose transport depends upon APP. Finally, we have developed an overexpression model for APP in Drosophila that causes neurodegeneration and whose genetic requirements are reminiscent of bona fide Alzheimer's Disease. These data suggest that at least one important normal function of the APP protein is as a kinesin-I receptor on vesicles, and that disturbances of this function could play a role in the causes or progression of Alzheimer's disease.

9. Specification of Dynein Motor Function in Drosophila Development: A Mutational Analysis of Dynein Subunit Function KRISTIN BOYLAN, MINGANG LI, SARAH MISCHE, MADELINE SERR, ANDRE SILVANOVICH, ED WOJCIK, and THOMAS HAYS, Department of Genetics, Cell Biology and Development, University of Minnesota, Minneapolis, MN

An important set of questions about cytoplasmic dynein concerns how the motor functions in multiple cellular and developmental processes. Our present understanding proposes that a single isoform of the cytoplasmic dynein HC is targeted to multiple cellular functions through association with accessory subunits and cargo “adapters.” The analysis of dynein heavy chain mutations previously demonstrated dynein's essential function during Drosophila development. The mutant phenotypes reveal a role for dynein in spindle orientation and oocyte differentiation, as well as in centrosome behavior and spindle autonomy during the mitotic divisions of the syncytial embryo. To pursue the mechanistic basis for the observed defects, we are analyzing the site of dynein motor action using transgenic flies expressing epitope-tagged dynein and dynactin subunits. The observations provide support for dynein function at the fusome during oocyte specification. During syncytial mitotic divisions, dynactin associates with the nuclear envelope in prophase where it may serve to maintain centrosome attachment. A dynamic poleward streaming of dynactin from prometaphase kinetochores is also revealed and will be considered in light of checkpoint controls. Loss-of-function mutations in the heavy chain motor subunit should affect all dynein functions. In contrast, if the nonmotor, accessory subunits of the dynein complex confer functional specificity, then the phenotypes associated with mutations in the corresponding genes should be distinct. We have now recovered mutations in the intermediate, light intermediate, and 14-kD tctex-1 subunits of the cytoplasmic dynein motor complex. The analysis of mutant phenotypes suggests new requirements for cytoplasmic dynein in wing development and spermatogenesis and supports the hypothesis that specific subunits can modulate dynein function in specific tissues.

10. The Myosin VII FERM Domains Are Essential for Its Role in Cell–Substrate Adhesion. RICHARD I. TUXWORTH, GREGORY C. ADDICKS, and MARGARET A. TITUS, Department of Genetics, Cell Biology and Development, University of Minnesota, Minneapolis, MN

Myosin VII (M7) is an actin-based motor protein that plays an essential role in hearing in humans, mice, and zebrafish. The simple eukaryote Dictyostelium expresses a single M7 (DdM7). Null mutant analysis has revealed that DdM7 plays a critical role in cell–substrate adhesion required for phagocytosis and cell migration, as well as filopod extension, novel functions for a myosin (Tuxworth et al. 2001. Curr. Biol. 11:318–329). The available evidence suggests that DdM7 plays a role in organizing receptors into a high avidity complex capable of binding to and stabilizing initial contacts with substrata.

The tail region of M7 is postulated to interact either directly or indirectly with receptors. It contains a short region of predicted coiled coil, along with a tandem repeat of combined MyTH4 and FERM domains separated by an SH3 domain. FERM domains are found in numerous cytoskeletal proteins that contribute to cell–substrate adhesion, but the function of these domains has not been clearly identified. Complementation and domain overexpression analyses using GFP-tagged molecules were performed to determine the role of the DdM7 FERM domains. Our results demonstrate that deletion of either FERM domain results in a loss of DdM7 localization and function. However, the individual FERM domains do not localize to the leading edge and filopodial tips of wild-type cells. They also do not act as dominant negative inhibitors of phagocytosis and filopod extension. Therefore, each of the FERM domains is necessary but not sufficient for DdM7 localization and function. (Supported by grants from NIH, National Science Foundation, and the Human Frontiers in Science.)

11. Dynein Binds to β-Catenin and Tethers Microtubules at Adherens Junctions LEE A. LIGON, SHER KARKI, MARIKO TOKITO, and ERIKA L.F. HOLZ-BAUR, Department of Animal Biology, University of Pennsylvania, Philadelphia, PA

Interactions between microtubules and the actin-rich cortex are hypothesized to be critical for many mechanical and signaling events in the cell. The microtubule motor cytoplasmic dynein has been proposed to mediate many of these interactions. Here, we report that dynein is localized to adherens junctions in cultured epithelial cells. A specific subpopulation of dynein containing the light chain Tctex-1 is enriched at these sites and is recruited to the contacts early in their formation, along with the junctional proteins β-catenin and E-cadherin. Microtubules project toward these early contacts and we hypothesize that dynein captures and tethers microtubules at these sites. We probed for interactions between dynein and junctional proteins and found that dynein coimmunoprecipitated with β-catenin. Subsequent biochemical analysis showed that dynein binds directly to β-catenin, suggesting a possible mechanism for targeting dynein to adherens junctions. Overexpression of β-catenin dramatically disrupts both the distribution of dynein in the cell and the cellular microtubule array. These results identify a novel role for cytoplasmic dynein in capturing and tethering microtubules at adherens junction. These tethered microtubules may provide structural support for the cell, or they may establish a fast track for communication between the periphery and the cell center. (Supported by NIH GM48661 and an American Heart Association Established Investigator Award to E.L.F. Holzbaur, and an NIH postdoctoral fellowship award to L.A. Ligon)

12. The Role of Cytoplasmic Dynein in Brain Disease and Mitosis RICHARD B. VALLEE, CHIN-YIN TAI, NICOLE E. FAULKNER. and DENIS L. DUJARDIN, University of Massachusetts Medical School, Worcester, MA

Lissencephaly is a severe brain developmental disease characterized by grossly disrupted neuronal distribution within the cerebral cortex. Human type I lissencephaly is caused by haploinsufficiency at the LIS1 locus and is thought to result from deficient migration of neuronal precursors from the ventricular zone during early brain development. A role for LIS1 in the cytoplasmic dynein pathway was suggested by genetic analysis of LIS1-related genes in lower eukaryotes (Xiang et al. 1995. Mol. Biol. Cell. 6:297–310). We have found the mammalian LIS1 gene product to coimmunoprecipitate with cytoplasmic dynein and its accessory complex dynactin and to colocalize with dynein and dynactin to the cortex of dividing cells and mitotic kinetochores (Faulkner et al. 2000. Nat. Cell. Biol. 2:784–791; Vallee et al. 2001. Trends Cell Biol. 11:155–160). LIS1 overexpression, antisense, and antibody injection experiments produced severe mitotic delays, and revealed a role for LIS1 in mitotic spindle orientation and metaphase chromosome alignment. These results suggested that defects in LIS1 affect neuronal distribution indirectly through primary effects on cell division within the developing neuroepithelium. We now find that epitope- or GFP-tagged LIS1 are targeted to the kinetochores and cortex of mitotic MDCK cells, as well as to a new site, the plus ends of growing microtubules. Together with our earlier data, these results identify three important sites of LIS1 colocalization with dynein and dynactin, and support a role for LIS1 in mediating the interaction of microtubule ends with the cell cortex and mitotic kinetochores. Coexpression experiments reveal the COOH-terminal presumptive β-barrel domain of LIS1 to interact with the cytoplasmic dynein heavy chain and to be sufficient for kinetochore targeting. These results suggest the β-barrel domain, the locus of many known human LIS1 mutations, to serve as a binding platform for other polypeptides in the dynein pathway. (Supported by GM47434 and HD40182 to R. Vallee; Fairlawn Foundation to N. Faulkner; and Human Frontiers to D. Dujardin.)

13. Roles for Unconventional Myosins in Membrane Uptake Events CARMEN WARREN,¹ LILY ENG,¹ LORI ASCHENBRENNER,¹ KATIE KLINE,¹ WAYNE VOGL,² and TAMA HASSON,^{1 1} Division of Biology, Section of Cell and Developmental Biology, University of California at San Diego, La Jolla, CA; ² Department of Anatomy, University of British Columbia, Vancouver, Canada

Myosin-VI and myosin-VIIa are two unconventional myosins required for hearing and balance. Mutations in either myosin result in profound deafness and balance dysfunction in mammals. Recent studies on Class VII myosins in Dictyostelium have suggested a role for this myosin class in adhesion during phagocytosis (Tuxworth et al. 2001. Curr. Biol. 11:318–329). We find myosin-VIIa is associated with adhesion structures in rodent testis and retina, suggesting myosin-VIIa may serve a similar role in phagocytosis in mammals. We have identified a binding partner associated with myosin-VIIa within adhesion domains and we will present our recent work characterizing this binding partner. Myosin-VI also has been implicated in membrane uptake events, although no direct evidence for myosin-VI function in endocytosis has been presented. Myosin-VI moves toward the pointed-end of actin filaments (Wells et al. 1999. Nature. 401:505–508), which is an orientation that would allow myosin-VI to pull membrane components into the cell. Our localization studies place myosin-VI in the correct locale for a role in endocytosis in both tissues and cultured cell lines. Within these endocytic domains we find that myosin-VI is associated with at least two distinct plasma membrane adaptor proteins. Both proteins have already been implicated in receptor-mediated endocytosis and are associated with clathrin-coated pits. The identified adaptor proteins are known membrane receptor binding proteins suggesting a role for myosin-VI in the regulated endocytosis of membrane receptors. These data provide the first real link between myosin-VI and the endocytic process. (Supported by NIH grant EY12695 and a Basil O'Connor Starter Scholar Research Grant No. 5-FY99–757 from the March of Dimes Birth Defects Foundation.)

14. Characterization of Mice Lacking Brush Border Myosin-I MATTHEW J. TYSKA,¹ JIAN-DONG HUANG,² JOSEPH SKOWRON,¹ NANCY A. JENKINS,² NEAL G. COPELAND,² and MARK S. MOOSEKER,^{1 1} Department of Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT; ² Mouse Cancer Genetics Program, National Cancer Institute, Frederick, MD

Brush border myosin-I (BBMI) forms spirally arrayed links between the microvillar membrane and the underlying actin core of the intestinal epithelial BB. Proposed functions for this myosin include polarized targeting of newly synthesized apical membrane components and mechanoregulation of nutrient transport across the BB membrane. BBMI contains multiple calmodulin light chains and represents the major calmodulin-binding protein of the enterocyte; thus, roles for this myosin in regulating vitamin-D dependent uptake of Ca²⁺ have been proposed. Phenotypic characterization of mice lacking BBMI heavy chain fails to support these hypotheses: These mice are healthy, and have normal intestinal function as assessed by growth rate, stool output, and bone density. However, these mice do show deficiencies at the cellular level. EM analysis of duodenal mucosa reveals that the BB surface is largely normal; however, a subset of the enterocytes exhibit regions of disrupted BB surface consisting of cytoplasmic herniations and detachment of the apical membrane from the underlying microvillar cores. Biochemical and ultrastructural characterization of isolated BBs from BBMI null mice reveal that these BBs lack bridges and, as a result, the BB membrane attachment to the underlying actin cytoskeleton is destabilized. Moreover, as expected, these BBs are depleted in calmodulin. Interestingly, these BBs also contain reduced levels of other myosins associated with the BB including myosin-Ic and myosin-VI. Other major BB cytoskeletal proteins are present at normal levels. Given the presence of an intact BB, but apparent instability in microvillus structure, we postulate that BBMI may be involved in maintaining structural integrity of microvilli and/or the plasticity of the BB structure in response to dietary stress or cell injury.

15. Motor Domain–dependent Localization of Myo1b (myr-1) NANYUN TANG and E. MICHAEL OSTAP, Department of Physiology, University of Pennsylvania School of Medicine, Philadelphia, PA

Myosin-I is the single-headed, membrane-binding, member of the myosin superfamily that plays a role in membrane dynamics and transport. Its molecular functions and its mechanism of regulation are not known. In mammalian cells, myosin-I is excluded from specific microfilament populations, indicating its localization is tightly regulated. Identifying the mechanism of this localization, and the specific actin populations with which myosin-I interacts, is crucial to understanding the molecular functions of this motor. eGFP chimeras of myo1b were imaged in live and fixed NRK cells. Ratio-imaging microscopy shows that myo1b-eGFP concentrates within dynamic areas of the actin cytoskeleton, most notably in membrane ruffles. Myo1b-eGFP does not associate with stable actin bundles or stress fibers. Truncation mutants consisting of the motor or tail domains show a partially overlapping cytoplasmic localization with full-length myo1b, but do not concentrate in membrane ruffles. A chimera consisting of the light chain and tail domains of myo1b and the motor domain from nonmuscle myosin-IIb (Myh10) concentrates on actin filaments in ruffles as well as on stress fibers. In vitro motility assays show that the exclusion of myo1b from certain actin filament populations is due to regulation of the actomyosin interaction by tropomyosin. Therefore, we conclude that tropomyosin and spatially regulated actin polymerization play important roles in regulating the function and localization of myo1b. (Supported by grants from NIH and AHA.)

Poster Abstracts

16. Kinetic Mechanism and Regulation of Myosin VI ENRIQUE M. DE LA CRUZ, E. MICHAEL OSTAP, and H. LEE SWEENEY, Department of Physiology, University of Pennsylvania School of Medicine, Pennsylvania Muscle Institute, Philadelphia, PA

Myosin VI is the only pointed-end directed myosin identified and is likely regulated by heavy chain phosphorylation (HCP) at the actin-binding site in vivo. We undertook a detailed kinetic analysis of the actomyosin VI ATPase cycle to determine if there are unique adaptations to support reverse directionality and to determine the molecular basis of regulation by HCP. ADP release is the rate-limiting step in the cycle. ATP binds slowly and with low affinity. At physiological nucleotide concentrations, myosin VI is strongly bound to actin and populates the nucleotide-free (rigor) and ADP-bound states. Therefore, myosin VI is a high duty ratio motor that is adapted for maintaining tension and has the potential to be processive. A mutant mimicking HCP increases the P_i release rate that lowers the K _ATPase, but does not affect the rate of ADP release. These measurements are the first to directly measure the steps regulated by HCP for any myosin, provide biochemical evidence for an additional phosphate-bound intermediate in the cycle, and may account for activation by HCP in vivo. Further, we propose a model for regulation of myosin VI motility by HCP. (E.M. De La Cruz is a Burroughs Wellcome Fund Fellow of the Life Sciences Research Foundation. This work was supported by the National Institutes of Health grants GM57247 to E.M. Ostap and AR35661 to H.L. Sweeney).

17. Regulatory Proteins Tropomyosin (Tm) and Troponin (Tn) Enhance Isometric Tension in Bovine Myocardium HIDEAKI FUJITA,¹ DAISUKE SASAKI,² SHIN'ICHI ISHIWATA,¹ and MASATAKA KAWAI,^{1 1} Department of Anatomy and Cell Biology, University of Iowa, Iowa City, IA; ² Department of Physics, School of Science and Engineering, Waseda University, Tokyo, Japan

The role of regulatory proteins in the elementary steps of the cross-bridge cycle was investigated using bovine myocardium in which the thin filament was selectively removed by gelsolin, and actin filament was reconstituted without Tm or Tn as reported (Fujita et al. 1996. Biophys. J. 71:2307–2318). The thin filament was further reconstituted by adding Tm and Tn to the actin filament-reconstituted myocardium. Isometric tension reproduced to 107 ± 4% (N = 26) of original tension. The effects of MgATP and inorganic phosphate (Pi) on the rate constants of exponential processes were studied in control, actin filament–reconstituted, and thin filament–reconstituted myocardium at 200 mM ionic strength, pCa 4.66, pH 7.00, and at 25°C. MgATP association constant was 9.1 ± 1.2 mM⁻¹ (±SEM, n = 11), ∼20× that of rabbit psoas and similar to that of porcine myocardium. Pi association constant of bovine myocardium was 0.14 ± 0.04 mM⁻¹, which was similar to that of both rabbit psoas and porcine myocardium. Equilibrium constant of the cross-bridge detachment step was 2.6 ± 0.4 s⁻¹, and equilibrium constant of the cross-bridge attachment step was 0.76 ± 0.05 s⁻¹. In the actin filament–reconstituted myocardium without regulatory proteins, the association constant of MgATP decreased to 0.18× and the association constant of Pi increased to 2.8×. Equilibrium constant of cross-bridge detachment decreased to 0.18×, but equilibrium constant of cross-bridge attachment increased to 7×. These kinetic constants regained the original value by reconstitution of the thin filament. These results indicate that presence of regulatory proteins, Tm and Tn, enhances detachment of cross-bridges. These results further indicate that tension/cross-bridge in the presence of regulatory proteins is ∼2× of that in the absence of regulatory proteins. The fact that regulatory proteins modified the rate and association constants of elementary steps of the cross-bridge cycle implies that the regulatory proteins change the conformation of actin so as to make more stereospecific interaction with myosin possible.

18. Mass/length Measurement and Composition of Native Purified Myosin Filaments from Muscle CARLOS HIDALGO,^1,2 RAUL PADRON,² RODRIGO MEDINA,³ PAOLA TONINO,⁴ LORENZO ALAMO,² MARTHA SIMON,⁵ and ROGER CRAIG,^{1 1} Department of Cell Biology, University of Massachusetts Medical School, Worcester, MA; ² Departmento de Biología Estructural and ³ Centro de Física, Instituto Venezolano de Investigaciones Científicas, Caracas, Venezuela; ⁴ Centro de Microscopía Electrónica, Facultad de Ciencias, Universidad Central de Venezuela, Caracas, Venezuela; ⁵ Biology Department, Brookhaven National Laboratory, Upton, NY

Analysis of the structure and function of native thick filaments of muscle has been hampered in the past by the difficulty of obtaining a pure preparation. Although such preparations have been obtained, they have involved specialized centrifugation approaches in addition that the native helical filament structure is not preserved. Therefore, we have developed a simple method for purifying native myosin filaments from muscle filament suspensions based on severing the thin filaments into short segments using a Ca²⁺-insensitive fragment of gelsolin. The thick filaments can be readily purified by differential centrifugation, with actin contamination below 3.5%. Negative staining and cryoelectron microscopy (cryo-EM) demonstrate intact thick filaments, with helical cross-bridge order preserved, and essentially complete removal of thin filaments. By gel electrophoresis, the purified thick filaments from tarantula muscle show myosin heavy and light chains together with paramyosin and three nonmyosin components. The mass/length of these filaments has been measured in the heads and bare zones using quantitative scanning transmission electron microscopy. Using the mass/length data, the paramyosin/myosin heavy chain ratio and the cryo-EM 3-D map we are refining an atomic model of the thick filament (Offer et al. 2000. J. Mol. Biol. 298:239–260) to include the backbone. (Supported by NIH grants AR34711 and HL62468 [to R. Craig] and CONICIT S1–97001311 [to R. Padron]. The research of R. Padron was supported in part by an International Research Scholars grant from the Howard Hughes Medical Institute.)

19. Characterization of the Motor Function of Class X Myosin KAZUAKI HOMMA, JUNYA SAITO, REIKO IKEBE, and MITSUO IKEBE, Department of Physiology, University of Massachusetts Medical School, Worcester, MA

Myosin X is a member of the diverse myosin superfamily that is ubiquitously expressed in various mammalian tissues. Although its association with the actin cytoskeleton in cells has been shown, little is known about its biochemical and mechanoenzymatic properties at the molecular level.

We expressed bovine myosin X containing the entire head, neck, and coiled-coil domains in Sf-9 cells, and then purified it. The Mg-ATPase activity was significantly activated by actin and a maximum steady-state rate of 8 s⁻¹/head was obtained at 25°C. The velocity of actin translocation was 0.36 μm/s. Using dual fluorescent labeled actin filaments, the directionality of myosin X was determined to be one with the pointed end at the front of the movement. This means that myosin X moved toward the barbed end, as is known for the conventional myosins

The motility activity was completely abolished at pCa6 or higher Ca²⁺ concentration. Activity resumed upon decreasing the Ca²⁺ concentration to pCa7, but not upon addition of exogenous CaM. The Mg-ATPase activity, on the other hand, decreased at pCa5 or higher, and this was accompanied by a dissociation of CaM from the heavy chain. In fact, the addition of exogenous CaM removed the inhibition of the ATPase activity at high Ca²⁺.

Despite similarities to myosin V in Ca²⁺ regulation, the K_actin and K_ADP values estimated for myosin X from the actin-activated ATPase assay were much larger than those of myosin V. As is consistent with this data, myosin X readily dissociated from actin in the presence of ATP and little myosin X coprecipitated with actin in the presence of ATP. These results indicate that myosin X is a nonprocessive motor. (Supported by NIH grants GM55834 and AR41653.)

20. The Motor Domain, Not the Lever-arm/Converter Domain, Determines the Direction of Myosin Movement KAZUAKI HOMMA, MISAKO YOSHI-MURA, JUNYA SAITO, REIKO IKEBE, and MITSUO IKEBE, Department of Physiology, University of Massachusetts Medical School, Worcester, MA

Unlike other well-characterized myosins, myosin VI moves toward the minus end of actin filaments. Since myosin VI has a unique insertion of ∼50 amino acids between its converter domain and light chain binding domain, it has been postulated that this insertion acts as a “reverse” gear, causing the lever arm to swing in the opposite direction, and thus causing this myosin to move in the opposite direction.

To verify this hypothesis, we produced a chimeric myosin that has the myosin V motor domain followed by the myosin VI neck and coiled-coil domains, including the unique insertion of myosin VI. The directionality of the movement was determined by conventional in vitro motility assay using dual-labeled actin filaments that had bright caps at their pointed ends labeled with tetramethylrhodamine maleimide, followed by dim fluorescein-labeled tails.

Although the chimera had myosin VI's unique insertion, it still moved toward the more typical plus end. This result suggests that the unique insertion of myosin VI is not a critical factor in determining the movement directionality of myosin. (Supported by NIH grants GM55834 and AR41653.)

21. Mammalian Myosin VIIA Is a Plus-directed Fast Motor that Requires High ATP Concentration For Its Motor Activity AKIRA INOUE and MITSUO IKEBE, Department of Physiology, University of Massachusetts Medical School, Worcester, MA

Myosin VII is a member of diverse myosin superfamily. Although its function in various cell biological processes has been postulated, little is known for its biochemical and mechanoenzymatic functions at a molecular level.

We obtained a cDNA clone containing entire coding region of myosin VIIA from rat kidney. Rat myosin VIIA containing entire motor domain and five IQ domains corresponding to residues 1–948 was coexpressed with calmodulin in sf9 cells. Mg-ATPase activity of myosin VIIA was activated by actin with K_actin of 67 M. Myosin VIIA translocates F-actin filaments and sliding velocity was 1.1 μm/s. High concentration of Ca²⁺-activated Mg-ATPase activity but completely abolished the actin-translocating activity. One mole of calmodulin dissociates from myosin VIIA at high Ca²⁺, but the dissociation neither affected the ATPase activity nor the motility activity. The directionality of the movement was determined by using dual fluorescence labeled actin, and it was found that myosin VIIA is a plus-directed motor. Myosin VIIA requires extremely high ATP concentration for its ATPase activity, in vitro motility activity, and dissociation from actin. On the other hand, ADP markedly inhibited the ATPase activity, and the ATP-induced dissociation of myosin VIIA from actin. These results indicated that myosin VIIA is a plus-directed motor that requires high ATP concentration for its motor function. (Supported by NIH grants GM55834 and AR41653.)

22. Switching Myosin in Drosophila Indirect Flight Muscle RHEA J.C. LEVINE,¹ ERZSEBET POLYAK,² RODNEY A. BRUNDAGE,³ and CHARLES P. EMERSON, JR.,^{2 1} Department of Neurobiology and Anatomy, MCP Hahnemann University, Philadelphia, PA; ² Department of Cell and Developmental Biology, University of Pennsylvania, Philadelphia, PA; ³ Department of Biology, West Virginia University, Morgantown, WV

Several different isoforms of myosin II are expressed in the thorax muscles of wild-type Drosophila, including flight muscle myosin, which is found in the indirect flight muscles, and TDT myosin, which is localized to the jump muscles. These two muscle types differ in their morphology, physiological properties, and structure of their myosin (thick) filaments. TDT myofibrils are strap-like; the thick filaments are solid in cross-section, ∼2.4 μm long, surrounded by 9–12 thin filaments in a more-or-less circular array and have appreciable paramyosin content. These filaments appear to have a relaxed myosin organization similar to that of Limulus, tarantula, and scorpion muscles (Stewart et al. 1985. J. Cell Biol. 101:402; Levine et al. 1983. J. Cell Biol. 97:186). Indirect flight muscles (IFMs) have cylindrical myofibrils, in which the thick–thin filament array is “crystalline” in organization. Thick filaments are hollow in cross-section (except in the bare zone), ∼3.3 μm long, and surrounded by six thin filaments that lie between adjacent thicks. Both thick and thin filaments are arranged in a precise hexagonal array. The paramyosin content of IFM thick filaments is extremely low, and the relaxed array of myosin heads is disordered. Natural and engineered mutants of the IFM are used to study the properties of the muscle proteins instrumental in the ability to fly. One such mutant has alanines substituted for the constitutively phosphorylated serines at positions 66 and 67 of the myosin regulatory light chains. These mutants cannot fly (Tohong et al. 1995. Nature. 374:650). Here, we report the effect of expressing either the TDT myosin or the myosin regulatory light chain mutant myosin in the IFM on the structure of the myofibrils and thick filaments, examined by electron microscopy and optical diffraction.

Two copies of the TDT myosin gene were expressed in transgenic flies, using the actin promoter, in an IFM-null background. Two copies of the mutant myosin regulatory light chain gene were expressed in transgenic flies with IFM null for the wild-type light chain. The IFM of adult flies with either transgene appears very similar to wild-type IFM: the myofibrils are round, but not perfectly crystalline, in the case of the TDT transgenic fibers. Most thick filaments in these fibrils are hollow, but some are solid. Except at the fibril periphery, where 9–12 thin filaments frequently surround each thin filament and extra thin filaments are present, the thick and thin filaments are in a “normal” hexagonal array. Skinned fibers were incubated in the Ca-insensitive subunit of gelsolin, followed by elastase and, finally, briefly sonicated to release thick filaments for negative staining. Filaments with two structures were separated from the TDT transgenic fibers. The majority are hollow, 3.3 μm long, with a disordered surface array in relaxing solution, similar to wild-type IFM filaments, while some are solid, ∼2.3 μm long, with a relaxed ordered array of surface myosins, similar to wild-type TDT filaments. Fibrils from the light-chain mutants closely resemble the wild-type IFM fibrils. Separated filaments are long and hollow, but frequently split into two subunits along their lengths, which sometimes rejoin to form a single structure. The relaxed surface array of myosin is disordered on these filaments as it is in those from wild-type IFM. These differences reflect different functions. (Supported by NIH grant HL15835 to the Pennsylvania Muscle Institute.)

23. Cross-linking of SH1 (Cys717) to the NH₂-terminal Subdomain in Smooth Muscle Myosin S1: Evidence for Flexible SH1 Region YIN LUO, Muscle and Motility Group, Boston Biomedical Research Institute, Watertown, MA (Sponsor: Kathleen Morgan)

The reactive thiols in the COOH-terminal 20k subdomain of skeletal muscle myosin S1 (skS1), Cys707 (SH1) and Cys697 (SH2), are well known to be cross-linked by N,N′-1,4-phenylenedimaleimide (pPDM) and N,N′-1,2-phenylenedimaleimide (oPDM) (12–14-Å and 4–9-Å span, respectively) in the presence of ATP, resulting in the inactivation and the trapping of the nucleotide in the active site of S1 (Wells and Yount. 1982. Methods Enzymol. 85:93–116). According to the crystal structure of skeletal S1 (Rayment et al. 1993. Science. 261:50–58), this is only possible if one or both of the helices in which these two residues are located melt and become flexible. In this work, chicken gizzard smooth muscle myosin S1 (smS1) was treated with pPDM and oPDM. In both cases, the SH1 (Cys717) was found to cross-link readily to a cysteine in the NH₂-terminal 25k subdomain, presumably Cys94 which has been identified as the only reactive thiol in the NH₂-terminal subdomain of smS1 (Kojima et al. 1999. Biochemistry. 38:11670–11676). In the crystal structure of smS1 (Dominguez et al. 1998. Cell. 94:559–571) these two residues are 13 Å apart and their side chains face away from each other. The SH1 in skS1 (Ue. 1987. Biochemistry. 26:1889–1894) or in smS1 (Bonet et al. 1988. Biochem. Biophys. Res. Commun. 152:1–8) were previously found to cross-link by dibromobimane to Cys522 in the 50k subdomain. This residue is 27 Å and 32 Å from the SH1 and Cys94, respectively. These results together suggest that the SH1 region is flexible between the NH₂-terminal and the 50k subdomains in solution. The presence of MgATP significantly increased the rate of oPDM–cross-linking in smS1, indicating increased flexibility of the SH1 region. This is supported by the crystal structure of scallop S1 (Houdusse et al. 1999. Cell. 97:459–470). In contrast to the SH1-SH2 cross-linking in skS1, the SH1-Cys94 cross-linking in smS1 did not affect the basal and the actin-activated MgATPase activities, and did it not affect smS1's ability to move actin in in vitro motility assays. Therefore, unlike the case of skS1, the oPDM- or pPDM-treated smS1 cannot be regarded as the weak binding state analogue of S1.

24. Kinetic Analysis of Myosin II “Backdoor” Mutants C.A. MORRIS,¹ A.L. WELLS,¹ A. HOUDUSSE,² and H.L. SWEENEY,^{1 1} Department of Physiology, and Pennsylvania Muscle Institute, University of Pennsylvania, Philadelphia, PA; ² CNRS Institute Curie, Paris, France

Analysis of the crystal structure of chicken S1 and Dictyostelium myosin II motor domain (Rayment et al. 1993. Science. 261: 50–58; Fischer et al. 1995. Biophys. J. 68:19S–28S) led to the hypothesis that myosin binding to actin accelerates the conformational change allowing inorganic phosphate to exit via a “backdoor” while ADP remains bound at the active site (Yount et al. 1995. Biophys. J. 68:44S–49S). If this mechanism is responsible for P_i release, then it should be possible to either inhibit or accelerate the release of phosphate by site-directed mutagenesis of an amino acid residue that extends into “backdoor” channel. Dictyostelium myosin II motor domain and smooth muscle myosin II mutants with altered side chain sizes and charges at this position were expressed and purified using the baculovirus/Sf9 cell system. The enzymatic properties of the wild type and mutant constructs were examined using steady-state and transient kinetic methods. The results indicate specific elements of the ATPase mechanism are altered by the mutations, primarily through modulation of the step(s) associated with phosphate release.

25. Tilting of the Light Chain Region in Single Myosin Molecules Using Total Internal Reflection Fluorescence Polarization Microscopy MARGOT E. QUINLAN,¹ JOSEPH N. FORKEY,¹ JOHN E.T. CORRIE,² and YALE E. GOLDMAN,^{1 1} Pennsylvania Muscle Institute, University of Pennsylvania, Philadelphia, PA; ² National Institute for Medical Research, London, United Kingdom

To study conformational changes of myosin leading to force production and filament sliding, we measured orientation changes of individual regulatory light chains (RLCs) during active translocation of actin. Engineered chicken gizzard RLC was labeled with bis-iodoacetamidorhodamine at Cys¹⁰⁸ and Cys¹⁰⁰ (Corrie et al. 1999. Nature. 400:425–430) and exchanged for the endogenous RLC in rabbit skeletal muscle myosin. Labeled myosin was mixed ∼1:10⁵ with unlabeled myosin, bound to aedans-labeled actin stabilized by phalloidin, and attached to a triethoxychlorosilane coated fused silica slide. The aedans fluorescence was excited by a 355-nm evanescent wave. The rhodamine was excited at 514 nm by an evanescent wave switched every 10 ms between polarizations parallel and perpendicular to the optical (z-) axis of the microscope and between x- and y-incident directions. Single fluorophores that colocalized with actin were identified using an intensified CCD camera. The stage was then translated to project the fluorescence emission through a polarizing prism onto a pair of photon-counting avalanche photodiodes. The eight resulting intensities were combined to give six polarization ratios which are sufficient to determine the axial angle (relative to the z-axis), the azimuthal angle (around the z-axis) and mobility of the probe's dipole on the <10-ms time scale. During actomyosin interactions at 1 μM ATP, abrupt deflections of the single molecule fluorescence intensities were observed. Some intensities increased while others simultaneously decreased leaving total emission constant. Discrete changes of the polarization ratios at constant total emission indicate tilting of the RLC. The probe angles could be related to the actin axis from the aedans images. The amplitude of <10 ms wobble often changed abruptly by >30°. These results suggest that large changes in orientation and mobility of the light chain domain accompany energy transduction. (Supported by NIH, HHMI, AHA, MDA, and MRC.)

26. Kinetic and Spectroscopic Evidence for Three Actomyosin:ADP States in Smooth Muscle STEVEN S. ROSENFELD,¹ JUN XING,² MICHAEL WHITAKER,³ HERBERT C. CHEUNG,² FRED BROWN,⁴ AMBER WELLS,⁴ RON A. MILLIGAN,³ and H. LEE SWEENEY,^{4 1} Department of Neurology and ² Department of Biochemistry and Molecular Genetics, University of Alabama at Birmingham, Birmingham, AL; ³ Department of Cell Biology, Scripps Research Institute, La Jolla, CA; ⁴ Department of Physiology, University of Pennsylvania, Philadelphia, PA

Smooth muscle myosin II undergoes an additional movement of the regulatory domain with ADP release that is not seen with fast skeletal muscle myosin II. In this study, we have examined the interactions of smooth muscle myosin subfragment 1 with ADP to see if this additional movement corresponds to an identifiable state change. These studies indicate that for this myosin:ADP, both the catalytic site and the actin binding site can each assume one of two conformations. Relatively loose coupling between these two binding sites leads to three discrete actin-associated ADP states. After an initial, weakly bound state, binding of myosin:ADP to actin shifts the equilibrium toward a mixture of two states that each bind actin strongly, but differ in the conformation of their catalytic sites. By contrast, fast myosins, including Dictyostelium myosin II, have reciprocal coupling between the actin and ADP binding sites, so that either actin or nucleotide, but not both, can be tightly bound. This uncoupling, which generates a second strongly bound actomyosin:ADP state in smooth muscle, would prolong the fraction of the ATPase cycle time that this actomyosin spends in a force-generating conformation, and may be central to explaining the physiologic differences between this and other myosins.

27. Structure and Function of Myosin V and X JAMES R. SELLERS, FEI WANG, LINGFENG CHEN, HEIPENG MENG, and JOHN A. HAMMER, III, National Heart, Lung and Blood Institute, National Institutes of Health, Bethesda, MD (Sponsor: H.L. Sweeney)

We are studying the structure, function and regulation of myosins from two classes, V and X. Intact myosin V's MgATPase activity is highly activated by calcium, whereas the MgATPase of the recombinant HMM-like fragment is largely independent of the calcium concentration when assayed in the presence of excess calmodulin. Myosin X has three IQ motifs that are potential light chain binding domains. There is a short segment predicted to form a coiled coil that probably allows for dimerization to produce a two-headed molecule. We have engineered fragments corresponding to an HMM and an S1 of bovine myosin X for expression in Sf9 cells. Initially, we co-expressed these fragments with calmodulin. Both constructs yielded soluble myosin containing bound. The myosin X-HMM-like fragment binds actin in an ATP-dependent manner and has an actin-activated MgATPase with a V_max of 10 s⁻¹ and a K_m of 5 μM at 37°C and 40 mM KCl. The MgATPase activity is relatively insensitive to increasing ionic strength compared with conventional myosin II molecules. The myosin X HMM translocates actin filaments at a rate of 0.18 μm/s in the in vitro motility assay. Motility requires moderate densities of myosin bound to the surface and is enhanced by the presence of methylcellulose. Thus, myosin X does not appear to be a candidate for a processive motor. We have also coexpressed the myosin X heavy chain fragments with a calmodulin-like protein (termed CLP) that is found in several tissues with an abundance similar to that of calmodulin. CLP binds to the heavy chain and supports activity. The affinity of myosin X for CLP appears to be lower than that of calmodulin as the myosin X coexpressed in baculovirus with CLP alone is purified with some bound calmodulin.

28. The Role of Actin and the S2 Rod in Myosin Conformation and Motility PAUL R. SELVIN,¹ MING XIAO,¹ TANIA CHAKRABARTY,¹ AMBER L. WELLS,² LI-QIONG CHEN,² C. BALDACCHINO,² J. REIFEN-BERGER,¹ and H. LEE SWEENEY,^{2 1} Department of Physics and Biophysics Center, University of Illinois, Urbana, IL; ² Department of Physiology, University of Pennsylvania School of Medicine, Philadelphia, PA

We have used various forms of fluorescence resonance energy transfer (FRET) to address two critical questions regarding myosin motility: (1) How does actin affect the structure of myosin, and in particular, does myosin adopt unique structures found only in the presence of actin? (2) How tightly are the two heads in dimeric myosins structurally and functionally coupled together? This later question is closely related to the structural stability of the S2 rod, which holds the two heads together at the head–rod junction.

First, to understand the effect of actin on myosin conformation, we have created a “cys-light” smooth muscle myosin (S1) construct in which all reactive cysteines have been removed and unique cysteines reintroduced at desired positions for labeling with fluorophores. In particular, we have introduced one cysteine in the 25/50-kD loop of the catalytic domain, and a second one in the regulatory light chain. These are labeled with donor and acceptor FRET pairs, which undergo energy transfer in a distant-dependent manner. We show that the ADP-induced swing of smooth muscle myosin, previously found in the presence of actin (Whittaker et al. 1995. Nature. 378:748–751; Gollub et al. 1996. Nat. Struct. Biol. 3:796–802) requires the presence of actin. More generally, this result implies that myosin adopts unique conformations only in the presence of actin, and hence the myosin crystal structures, which are all in the absence of actin, likely represent only a subset of the total myosin conformations present in the catalytic cycle.

Second, we have examined the stability of the S2 rod. When a myosin dimer is bound via two heads in a rigor conformation to an actin filament, a potentially large strain is placed on the head–rod junction, tending to unwind the S2 rod. We have measured distances between the two regulatory light chains within a myosin dimer under such conditions. If S2 uncoils, the distances are expected to be ∼90 Å. If S2 remains intact, the distances are expected to be <50 or <60 Å. We find the latter to be the case, implying that S2 is likely to remain largely intact during the catalytic cycle. A similar result has recently been found for kinesin (Tomishige et al. 2000. J. Cell Biol. 151:1081–1092) and, hence, an intact S2 may be a general feature of motility in dimeric motors. In addition, our results imply that a distortion occurs elsewhere in the myosin heads to allow two-headed binding of myosin to actin. [Supported by NIH grants to P.R. Selvin and H.L. Sweeney]

29. Mechanism of Inhibition of Skeletal Muscle Actomyosin by N-benzyl-p-toluene sulfonamide M. ALEXANDER SHAW, E. MICHAEL OSTAP, and YALE E. GOLDMAN, Pennsylvania Muscle Institute, University of Pennsylvania, Philadelphia, PA

The compound N-benzyl-p-toluene sulfonamide (BTS) specifically inhibits contraction of fast skeletal muscle fibers (Dantzig et al. 2001. Biophys. J. 80:274a). To determine the mechanism of this inhibition, we studied the effects of BTS on kinetic parameters of the myosin subfragment-1 (S1) and actin-activated S1 (actoS1) ATPase cycles. BTS does not affect the rate of nucleotide binding or the rate of ATP cleavage as detected by changes in the intrinsic tryptophan fluorescence of S1, indicating that BTS is not a competitive inhibitor. The single turnover and steady-state rates of ATP hydrolysis by S1 are decreased about fivefold in the presence of saturating levels (100 μM) of BTS, and the inhibition constant (K _I) is 10 μM. ActoS1 ATPase rates are also inhibited about fivefold at saturating BTS concentration. BTS does not block dissociation of actoS1 by ATP, but it weakens the affinity of S1·ADP for actin. Our results suggest that inhibition of ATPase activity by BTS is due to the inhibition of P_i release. The decreased actin affinity of the “strong-binding” S1·ADP state may also contribute to suppression of force generation in muscle fibers. (Supported by NIH grants AR26846 and AR07584.)

30. Drosophila Alternative Exon 3 Encodes a Region that Modulates Actin Velocity and ATPase Rate DOUG-LAS M. SWANK,¹ AILEEN F. KNOWLES,² FLOYD M. SARSOZA,¹ WILLIAM A. KRONERT,¹ GEORGE E. MORRILL,¹ and SANFORD I. BERNSTEIN,^{1 1} Department of Biology, ² Department of Chemistry, and Molecular Biology Institute, San Diego State University, San Diego, CA

We are analyzing the function of the domain encoded by alternative exon 3 in the Drosophila myosin heavy chain (MHC). This corresponds to residues 69–116 of chicken skeletal MHC. There are two alternative choices for this region, encoded by exons 3a and 3b in the single-copy Drosophila Mhc gene. The exon 3–encoded region is one of only four alternative domains in the S-1 head. We isolated myosin from Drosophila expressing one of four transgenes in their indirect flight muscle (IFM). These myosins are: EMB, an embryonic isoform; IFI (positive control), the endogenous IFM isoform (these two isoforms differ in all four alternative domains); and two chimeras where the exon 3 region is exchanged between the two isoforms, EMB-3b and IFI-3a. Previously, we showed significant differences in ATPase rates and in vitro actin sliding velocities between the IFI and EMB forms. Preliminary data indicate that IFI-3a basal ATPase rates are ∼70% of IFI, whereas EMB-3b rates are ∼2.5-fold higher than EMB. The exon 3 region also affects actin sliding velocity in vitro. EMB-3b rates are about threefold higher than EMB and IFI-3a velocity appears to be slower than IFI. These changes in ATPase and actin sliding rates are not enough to convert the functional properties of the chimeric isoform to the isoform that provided the exon 3 domain. This is not surprising, as we previously showed that the exon 11 region also influences actin velocity and that the exon 7 region affects basal ATPase rates. EMB and EMB-3b transgenic flies are both flightless, but the EMB-3b IFM myofibril ultrastructure does not deteriorate as severely as EMB (IFM myofibrils crack and fray as the flies age). The IFI-3a flies do not appear flight impaired and IFI-3a IFM myofibril ultrastructure is identical to wild type. Thus, the changes in functional properties resulting from the 3b to 3a switch do not grossly affect muscle structure or muscle performance in the intact organism. (Supported by NIH GM32443 and R25 GM58906 and a AHA Western States postdoctoral fellowship to D.M. Swank)

31. The Structural Basis of Nucleotide Dependent Intrinsic Fluorescence Changes in Smooth Muscle Myosin MARILYN VAN DUFFELEN, LYNN R. CHRIN, CHRISTOPHER M. YENGO, and CHRISTOPHER L. BERGER, Department of Molecular Physiology and Biophysics, College of Medicine, University of Vermont, Burlington, VT

The intrinsic fluorescence of smooth muscle myosin is sensitive to both nucleotide binding and hydrolysis. We have examined this relationship by making smooth muscle myosin motor domain-essential light chain (MDE) mutants, which contain a single tryptophan residue at the seven positions found in wild type (29, 36, 441, 512, 548, 597, and 625). Previously, tryptophans 441, 546, 597, and 625 have been shown to be insensitive to nucleotide binding, whereas 512, located in the rigid relay loop, is sensitive to conformational changes during the MgATPase cycle (Yengo et al. 2000. J. Biol. Chem. 275:25481–25487). Further investigations of the transient kinetics of nucleotide binding and hydrolysis by stopped flow fluorescence demonstrated that the rate of fluorescence change in W512-MDE is greater than wild-type MDE. We further investigated the subsequent stages of hydrolysis, the rate of P_i and ADP release. Adjacent to the rigid relay loop is the converter region, thought to play a role in converting the energy of ATP hydrolysis into the mechanical power stroke. Smooth muscle myosin has two nonconserved tryptophans in this region, W29 and W36. We have observed no fluorescence change upon nucleotide binding in the W36 mutant, but a commensurate quenching of fluorescence in the W29 mutant, which might help explain the level of fluorescence enhancement upon nucleotide binding in W512-MDE relative to wild-type MDE.

32. Myosin V Dissociates From Actin in the Weak Binding States CHRISTOPHER M. YENGO, ENRIQUE M. DE LA CRUZ, DANIEL SAFER, E. MICHAEL OSTAP, and H. LEE SWEENEY, Department of Physiology, University of Pennsylvania School of Medicine, Philadelphia, PA

Myosin V is a molecular motor shown to move processively along actin filaments, unlike conventional myosins, by remaining bound to actin for a greater fraction of its ATPase cycle (duty ratio). This may be accomplished by maintaining a high affinity for actin during the “weakly bound” states and/or by populating the “strongly bound” states for a greater fraction of the cycle. To begin examining the affinities of the “weakly bound” states, we determined the affinity of the myosin V 1IQ monomer for actin in the presence of nonhydrolyzable ATP analogues, ATPγS and AMPPNP. Cosedimentation assays with myosin V in the presence of ATPγS and varying actin concentrations demonstrated that myosin V binds actin weakly (K _d = 10 μM) in the M.ATP state. However, in the presence of AMPPNP myosin V did not quench pyrene actin while its affinity for actin was quite high (K _d = 0.3 μM), indicating that this analogue may mimic a state not yet identified for myosin V.

Replacing either a conserved glycine (G440A) or glutamic acid (E442A) residue within the switch II region with an alanine, trapped MV in the M.ATP or M*.ATP state, respectively. The G440A and E442A mutants bound to actin with very different affinities (K _d = 0.2, and 2 μM, respectively), suggesting coordination of the γ-phosphate of ATP is extremely important for generating a “weak-binding” conformation and dissociating myosin V from actin.

Although the affinities of myosin V for actin in the “weak-binding states” are higher than conventional myosins, the data favor the dissociative pathway for ATP hydrolysis by myosin V as suggested previously (De La Cruz et al. 1999. Proc. Natl. Acad. Sci. USA. 96:13726; De La Cruz et al. 2000. Biophys. J. 79:1524). Extremely rapid rates of ATP hydrolysis and P_i-release as well as a high affinity for actin in the posthydrolysis states, allow us to model the mechanism of native dimer processivity.

33. Distribution of Myosin Heads Attached to Actin Filaments in the Weakly Bound States of A•M•ATP and A•M•ADP•P_i in Skinned Psoas Muscle Fibers L.C. YU, J. GU, and S. XU, National Institutes of Health, Bethesda, MD

The transition from the weakly bound to strongly bound states in the cross-bridge ATPase cycle is thought to be the structural basis for force generation (Eisenberg and Hill. 1978. Prog. Biophys. Mol. Biol. 33:55–82; Brenner et al. 1991. Proc. Natl. Acad. Sci. USA. 88:5739–5743). Therefore, it is critical that the structures of the weakly bound states are fully analyzed. However, there is little detailed structural information on the individual states either with unhydrolyzed ATP (the A•M•ATP state), or with ADP•P_i (the A•M•ADP•P_i state). In the present study, the two weakly attached states were studied individually by X-ray diffraction from skinned muscle fibers.

For studying the A•M•ATP state, skinned rabbit psoas fibers were first reacted with N-phenylmaleimide to inhibit ATP hydrolysis, so that the cross-bridges under relaxing conditions were distributed only between the M•ATP and A•M•ATP states. Displacing the equilibrium toward the A•M•ATP state was achieved by lowering the ionic strength. The results show that upon attachment, the intensities of both the myosin and the first actin layer lines increase, whereas the sixth actin layer line is not significantly affected.

A model that explains the X-ray diffraction patterns from the A•M•ATP state has the following features: (1) the myosin head binds to actin at a specific site but with a wide range of orientations, supporting the idea of nonstereospecific binding; (2) the end of the head close to the thick filament backbone deviates substantially from its helical lattice position; (3) the unattached head of the cross-bridge appears to be located close to the surface of the thick filament; and (4) the binding site on actin is near the NH₂ terminus of the actin subunit, a position significantly different from the putative rigor binding site (Rayment et al. 1993. Science. 261:58–65; Schroder et al. 1993. Nature. 364:171–174).

Recently, we have obtained preliminary results showing evidence that the conformation of the A•M•ADP•P_i state differs from that in the A•M•ATP state. Because of the low affinity between actin and myosin, the conformation of A•M•ADP•P_i complex has been elusive for detailed study. However, biochemical studies showed that the population in the A•M•ADP•P_i state could be increased substantially by raising temperature, lowering the ionic strength, and by adding polyethylene glycol (White et al. 1997. Biochemistry. 36:11828–11836; Highsmith et al. 1998. Biophys. J. 74:1465–1472). In the present study, X-ray diffraction patterns were obtained from relaxed skinned rabbit psoas muscle at 25°C, with ionic strength lowered from 200 to 50 mM, and/or 5% of PEG 1000. The interventions significantly increased the ratio I₁₁/I₁₀, signifying increasing cross-bridge attachment. Although the increase originates from a mixture of the weakly bound states, the diffraction patterns show characteristics different from those associated with increasing the A•M•ATP population alone. There is a pronounced increase in myosin layer line intensities, but no comparable increase in the intensity of the first actin layer line, which was observed in the A•M•ATP state. Another significant difference is an increase in the intensity of the sixth actin layer line, whereas little change was observed with the A•M•ATP population. The results suggest that the conformation of the actomyosin complex containing ATP is modified by hydrolysis.

34. On the Dynamics of Single Kinesin Mole-cules MICHAEL E. FISHER and ANATOLY B. KOLOMEISKY, Institute for Physical Science and Technology, University of Maryland, College Park, MD; Department of Chemistry, Rice University, Houston, TX (Sponsor: Jonathon Howard)

Recently, the velocity, dispersion and mean-run lengths of single kinesin molecules moving in vitro along microtubules have been measured accurately under a range of experimental conditions (for external loads F = 1 − 8 pN and [ATP] = 1 mM − 2 mM) by Block and co-workers (1999,2000). We show that these extensive data and other experiments under assisting loads (Coppin et al. 1997.) can be described in a unified manner by simple discrete-state stochastic models. An (N = 2)-state model with fixed load-distribution factors and rate constants (consistent with chemical kinetic experiments) fits force-velocity curves at different [ATP]. Our analysis indicates a “substep” of d ₀ > 2.0 nm for ATP binding, which is consistent with structural suggestions. To describe the randomness (a dimensionless measure of the dispersion), the addition of a waiting-time distribution characterized by a degree of mechanicity is required. Alternatively, an (N = 4)-state chemical kinetic model is needed. Our analysis shows that the load distribution patterns are similar for both descriptions. Based on fits to the experimental data, the mean dwell times in the intermediate chemical states are obtained as functions of load. With allowance for detachment rates, the mean-run lengths can also be described satisfactorily.

35. A Kinetic Study of the ATP-induced Reorientation of the Kinesin Neck Linker STEVEN S. ROSENFELD, PETER H. KING, and GERALDINE M. JEFFERSON, Department of Neurology, University of Alabama at Birmingham, Birmingham, AL

Recent models of the kinesin mechanochemical cycle provide some conflicting information on how the neck linker contributes to movement. Some spectroscopic approaches suggest a nucleotide-induced order-to-disorder transition in the neck linker. However, cryoelectron microscopic imaging suggests instead that nucleotide alters the orientation of the neck linker when docked on the microtubule surface. Furthermore, since these studies used transition state or nonhydrolyzable nucleotide analogues, it is not clear at what point in the ATPase cycle this reorientation of the neck linker occurs. We have addressed this issue by developing a strategy to examine the effect of nucleotide on the orientation of the neck linker, based on the technique of fluorescence resonance energy transfer. Transient kinetic studies using this approach support a model in which ATP binding leads to two sequential isomerizations, the second of which reorients the neck linker in relation to the microtubule surface.

36. Looking for Dimeric Kinesin's “Bridge” GEORGIOS SKINIOTIS,¹ YOUNG-HWA SONG², ECKHARD MANDELKOW,² and ANDREAS HOENGER,^{1 1} European Molecular Biology Laboratory, 69117 Heidelberg, Germany; ² Max-Planck-Unit for Structural Biology, DESY, 22602 Hamburg, Germany

Conventional kinesin's walk on microtubules involves a number of conformational changes, triggered by the alternate binding and hydrolysis of ATP between its two motor heads. A key component in this highly processive movement is the neck linker region, a stretch of around 15 amino acids, that connects each catalytic motor domain with the dimerization coiled coil. The conformation of each neck linker within a dimer is different at each substep of kinesin's walk, and allows the free motor to search and bind on the next binding position while the other remains attached on the microtubule lattice. Using cryo-EM and various image reconstruction methods, we have sought to visualize the directionality of this region and locate the connection in a dimer with both heads bound. To compensate for the small and, until now, EM-unresolvable volume of the neck-linker, we have engineered a number of cysteine-light rat kinesin chimeras for specific undecagold cluster labeling, and also introduced small and compact protein domains between the linker and the dimerization coiled coil. One of the latter constructs carries an SH3 insertion at this position (rk379-SH3). Electron microscopy images of unidirectionally shadowed rk379-SH3 bound on tubulin sheets in the AMP-PNP state indicate axial repeats of 16 nm, showing that this motor is able to bind both heads on adjacent tubulin dimers. The same type of experiments show cooperative binding along single protofilaments, as being observed with the untagged protein. Helical reconstructions of rk379-SH3 complexed with microtubules in different nucleotide states, and difference mapping with plain rk379 reveal the orientations of the neck linkers in the context of double headed binding. These results are supported by helical reconstructions of microtubules decorated with monomeric forms of untagged and SH3-tagged rat kinesin. Because of the limitations of helical averaging for 3-D reconstructions of irregular polymers such as conventional kinesin–microtubule complexes, we are also applying back-projection methods, and we are investigating the application of single particle averaging for resolving the possibly different conformations of the two bound heads.

37. Ncd–Microtubule Complexes Show Cooperative Effects at Various Nucleotide Conditions as Revealed by Cryoelectron Microscopy T. WENDT, K. GOLDIE, Y. SKINIOTIS, E. MANDELKOW, and A. HOENGER, European Molecular Biology Laboratory, Structures and Computational Biology Program, Heidelberg, Germany; Max-Planck-Unit for Structural Molecular Biology, Hamburg, Germany

We have reinvestigated the structure and microtubule binding patterns of double-headed Drosophila Ncd under various nucleotide conditions and different stoichiometric motor–tubulin ratios, using cryoelectron microscopy and helical reconstruction methods. In the presence of AMP-PNP or in the absence of nucleotide and at decoration conditions using 4–5 times as many motor constructs than available binding sites, our maps are identical with previous results. These maps were used to dock the atomic resolution Ncd-dimer structure into our EM-derived 3-D volumes to model the distinct conformational changes, which occur upon microtubule binding. In contrast to other reports, our experiments revealed a very weak binding affinity of Ncd to tubulin in the presence of excess ADP. Even in experiments working with high motor/tubulin ratios microtubules remained virtually undecorated. Using decoration conditions with low stoichiometric motor/tubulin ratios, however, we found some strikingly different binding patterns and some clear indications of cooperativity during decoration. Ncd molecules were observed to fully decorate single protofilaments along microtubules while others remained completely empty, indicating a cooperative decoration process in axial, but not lateral, direction. In addition, our Ncd motor construct seems to mediate bundling of microtubules by forming a zipper-like interaction with neighboring heads from another microtubule under the same conditions. We are currently trying to investigate their structure by a single particle approach. Furthermore, conditions have been found, which indicate that dimeric ncd constructs show similar “double-binding” decoration patterns as found for kinesins.

38. Removal of the α Heavy Chain from Chlamydomonas Outer Arm Dynein Reveals a Ca²⁺-controlled ATP-dependent Microtubule-binding Activity MIHO SAKATO and STEPHEN M. KING, Department of Biochemistry, University of Connecticut Health Center, Farmington, CT

In Chlamydomonas, the photoshock response involves an alteration in flagellar waveform and swimming direction that is signaled by an increase in intraflagellar Ca²⁺ from 10⁻⁶ to 10⁻⁴ M. This response is abnormal or missing in cells lacking outer dynein arms. The only component of the outer arm known to bind Ca²⁺ is the LC4 light chain that is associated with the γ heavy chain (HC). LC4 is a member of the EF-hand family of Ca²⁺-binding proteins and, like calmodulin, contains four helix-loop-helix motifs, although only two conform to the EF-hand consensus for Ca²⁺-binding loops. Scatchard analysis indicates that recombinant LC4 binds 1.2 mol Ca²⁺/mol protein with a K _Ca = ∼10⁻⁵ M. Dynein subparticles lacking the α HC, derived from the mutants oda11 (βγ HCs) and oda11 oda4-s7 (γ HC only), exhibit Ca²⁺-controlled ATP-dependent microtubule (MT) binding. This response was not observed in outer arm dynein from wild-type cells (containing the α, β, and γ HCs) or from the mutant oda4-s7 (αγ HCs; lacks the β HC motor domain). The mutations sup1–1 and sup1–2 delete small sections from the MT-binding stalk of the β HC and suppress flagellar paralysis caused by radial spoke and central-pair microtubule deficiencies. Outer arm dyneins from both strains show Ca²⁺-independent ATP-sensitive MT binding. In contrast, significant amounts of dynein obtained from the double mutant oda11 sup1–2 (contains intact γ HC and a β HC with a small deletion) did not bind MTs except at very high Ca²⁺ concentrations. Together, these data suggest that the MT-binding activity of the γ HC may be activated by high Ca²⁺ and that this response is overridden, at least in vitro, by the presence of the α HC. (Supported by NIH grant GM51293.)

39. The A,B,C of Nonmuscle Myosin II ROBERT S. ADELSTEIN, KAZUYO TAKEDA, MARY ANNE CONTI, ELIAHU GOLOMB, HIROKO KISHI, YVETTE PRESTON, ZU-XI YU, VICTOR J. FERRANS, and QIZE WEI, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, Maryland (Sponsor: H. Lee Sweeney)

We are studying the role of nonmuscle myosin II during mouse development and in cultured cells. (A) Ablation of nonmuscle myosin heavy chain II-A (NMHC II-A) by homologous recombination results in lethality before embryonic day 7.5 (E7.5), possibly due to a failure in visceral endoderm development, since these cells also appear to lack NMHC II-B. Interestingly, in humans, others have shown that point mutations of single amino acids cause a variety of defects including macrothrombocytopenia, nephritis, and deafness. HeLa cells transfected with NH₂-terminal truncated NMHC II-A show defects in morphology and focal adhesions suggesting its role in these processes (Wei and Adelstein. 2000. Mol. Biol. Cell. 11:3617–3627). (B) Ablation of NMHC II-B leads to abnormalities in heart and brain and lethality at E15–18. Cardiac abnormalities include myocyte hypertrophy in the absence of cardiomegaly by E12.5 and a marked increase in binucleated cells, suggesting a defect in cytokinesis. Brain abnormalities include hydrocephalus, due to disruption of the ventricular surface and disordered cell migration (Tullio et al. 2001. J. Comparative Neurol. 433:62–74). We have also generated hypomorphic mice that manifest a gene-dosage effect in that they develop less severe defects over a more prolonged period, reflecting decreasing levels of NMHC II-B (Üren et al. 2000. J. Clin. Invest. 105:663–671). (C) The human genome sequence revealed NMHC II-C at 19q13.3 (GenBank accession No. AC020906, AC019157, and AK023943). We used mouse and human cDNAs to generate three different NMHC II-C–-specific probes and detected message in a number of tissues including lung, kidney, heart, and colon. Use of peptide-specific antibodies confirmed its presence in a number of tissues.

40. The Role of Microfilaments in Zebrafish Epi-boly JACKIE C. CHENG, ANDREW L. MILLER, and SARAH E. WEBB, Calcium Aequorin Imaging Laboratory, Department of Biology, The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong

The basic body plan of zebrafish embryos emerges during the gastrula period when a series of extensive cell movements and rearrangements (epiboly, involution, convergence, and extension) lead to the formation of the three germ layers, the endoderm, mesoderm, and ectoderm, as well as the dorsoventral and anterioposterior body axes. Starting toward the end of the blastula period, epiboly consists of the thinning and spreading of both the blastoderm and the yolk syncytial layer (YSL) over the yolk cell toward the vegetal pole until, by the end of the gastrula period, the yolk is completed encompassed. It has been previously proposed that in zebrafish, epiboly is partly driven by a microtubule-dependent mechanism. Here we present new data that suggests that microfilaments (MF) might also play a crucial role in this process. At ∼65–70% epiboly, a distinct 10-μm-thick ring of F-actin appears in the yolk cytoplasmic layer (YCL) at the enveloping layer (EVL) margin. Treatment with cytochalasin B at various times during epiboly leads to the disruption of actin structures and results in a slowing of epiboly, a failure of blastopore closure, an elongation of the embryo along its animal pole–vegetal pole (AP-VP) axis, and eventual lysis of the embryo through the vegetal hemisphere. We suggest that the YCL band of actin occurs at a site of extensive endocytosis, which helps to drive epiboly via the removal of membrane. (Supported by HK Research Grants Council Grant HKUST6130/98M awarded to A.L. Miller.)

41. A Molecular and Genetic Analysis of the Forces Required for Morphogenesis D.P. KIEHART, J.W. BLOOR, J.M. WIEMANN, N.J. GERALD, V.S. WILLIAMS, J.D. FRANKE, and R.A. MONTAGUE, Developmental, Cell and Molecular Biology Group, Department of Biology, Duke University, Durham, NC; Department of Cell Biology, Duke University Medical School, Durham, NC

Our focus is on the cellular and molecular mechanisms of cell shape change and movement during morphogenesis, cell locomotion, and wound healing. We generated transgenic flies designed to express sequences that encode the actin binding region of Drosophila moesin fused to sequences that encode the green fluorescent protein, an autonomously folding fluorophore under the control of a number of different transcriptional regulators. These constructs function as cell shape and structure reporters. We analyze morphogenesis during dorsal closure in wild-type, mutant and laser-ablated Drosophila embryos. Our studies indicate the following: (1) that two tissues, the leading edge of the lateral epidermis and the amnioserosa, are required for dorsal closure; (2) nonmuscle myosin II function is required for dorsal closure in both tissues; and (3) that the small GTPase Rho mediates wound healing at a time that it is apparently not required for dorsal closure. We have also used these tools to investigate the behavior of the hemocyte/macrophage cell lineage. Previous immunofluorescent imaging of fixed specimens suggested that these cells were small, round and relatively “inert” cells, comparable to blood borne lymphocytes in vertebrate species. In contrast, imaging of living specimens demonstrate that these cells are highly dynamic, have extensive lamellipodia, and locomote in seemingly choreographed patterns. These movements depend on the same cues that are responsible for guiding axons in and through embryonic tissues. Together, our observations provide new insight into the molecular mechanisms of morphogenesis. (Supported by NIH grants GM33830 and GM61240, AFOSR, March of Dimes, Muscular Dystrophy Association and NIH training grant GM07184 to J.D. Franke.)

42. Xenopus Myosin-1b Is Specifically Upregulated during Meiotic Maturation and Is Required for Egg Activation ANNA M. SOKAC and WILLIAM M. BEMENT, Program in Cellular and Molecular Biology, University of Wisconsin, Madison, WI

Meiotic maturation is the process whereby meiosis-I–arrested oocytes are converted into fertilization-competent, meiosis-II–arrested eggs. The egg then undergoes “activation” in response to fertilization. In frogs, egg activation is characterized by two classic, developmentally essential events: cortical granule exocytosis, which provides the slow block to polyspermy, and cortical contraction, which abets union of the sperm and the egg pronuclei. Because many of the events of egg activation are dependent on the actin cytoskeleton, we used a PCR-based screen to characterize the expression pattern of unconventional myosins in oocytes and eggs. Although a total of 13 unconventional myosins were found to be expressed in oocytes, only myosin-1b expression was found to be regulated during oogenesis. That is, myosin-1b poly(A)⁺ mRNA levels increase two- to threefold during meiotic maturation, while at the same time other myosin poly(A)⁺ mRNA levels parallel those of housekeeping proteins and decrease by one- to twofold. The increase is paralleled by a twofold increase in protein levels, and is due to increased polyadenylation of the myosin-1b mRNA. Injection of in vitro–transcribed mRNA encoding eGFP-myosin-1b fusions revealed that myosin-1b localizes to a novel, cortical membrane compartment in oocytes and undergoes transient localization to cortical granules during egg activation. Further, dominant negative tail expression of myosin-1b results in a profound deficit during egg activation—the complete inability to undergo either cortical contraction or cortical granule exocytosis. This phenotype is quite specific, in that oocytes proceed through meiotic maturation with no apparent cell cycle defects and assemble normal meiotic spindles, and in the absence of activation, the eggs appear perfectly healthy. This work provides the first demonstration of a role for an unconventional myosin in vertebrate fertilization. (Supported by NIH GM52932.)

43. Characterization and Differential Expression of the Tctex1/Tctex2 Family of Dynein Light Chains LINDA M. DIBELLA and STEPHEN M. KING, Department of Biochemistry, University of Connecticut Health Center, Farmington, CT

The Tctex1 and Tctex2 proteins define two major subgroups within one class of light chain (LC) that is present in both flagellar and cytoplasmic dyneins. Both proteins were originally identified in mammalian sperm and are thought be involved in non-Mendelian chromosome transmission in mice. Tctex1 appears to interact directly with specific cellular cargoes such as rhodopsin, Fyn tyrosine kinase, Trk neurotrophin receptor, and the GTPase NEDD-3. In Chlamydomonas, Tctex2 is essential for assembly of the flagellar outer dynein arm. Members of both the Tctex1 and Tctex2 subgroups are expressed in a tissue-specific and developmentally regulated fashion in mammals. For example, Tctex1 levels decrease in the lung during development, whereas levels of the related rp3 protein are greatly increased in the adult brain. Similarly, Tctex2 expression levels decrease during development of the thymus, whereas a related isoform is upregulated in the adult versus fetal kidney. Immunofluorescence analysis of Vero cells using an antibody against Tctex2 revealed a punctate perinuclear staining supporting our earlier identification of Tctex2 in cytoplasmic dynein immunoprecipitates from kidney and spleen. We have recently obtained a highly specific antiserum to the Tctex2 isoform upregulated in adult kidney and find that this protein also exhibits a punctate, cytoplasmic staining pattern in Vero cells and in many cases is localized along microtubules. We are currently performing additional cellular and biochemical experiments to further characterize novel isoforms of Tctex2. (Supported by GM51293 from NIH and by Patrick and Catherine Weldon Donaghue Medical Research Foundation.)

44. Wheels within Wheels and Rings within Rings: Spatial Regulation of Actomyosin during Xenopus oocyte Wound Healing WILLIAM M. BEMENT, CRAIG A. MANDATO, and HELENE A. BENINK, Department of Zoology, University of Wisconsin, Madison, WI

Plasma membrane wounds in Xenopus oocytes rapidly assemble a contractile array of F-actin and myosin-2 which subsequently closes the wound. Time-lapse confocal microscopy revealed that this array assembles as a result of two experimentally separable processes: (1) formation of a localized “zone” of actin and myosin-2 assembly around the wound border and (2) myosin-dependent flow of stable F-actin to the wound. These processes give rise to a classic contractile ring composed of F-actin and myosin-2 found within the broader ring of de novo assembly. These two arrays, in turn, are contained within a radial array of microtubules. Taxol treatment splits both the radial array of microtubules and the zone of assembly into concentric rings around the wound. To determine how, exactly, microtubules exert spatial control of the zone of actin and myosin-2 assembly, a novel approach was developed to visualize active rho-class GTPases in situ. This approach revealed that rho activity is elevated in a ring-like zone on the interior of the actomyosin array around wounds, whereas rac/Cdc42 activity is localized in a ringlike zone on the exterior of the actomyosin array around wounds. Treatment of oocytes with taxol resulted in splitting of the ring of rac/Cdc42 activity. We conclude that wounding establishes concentric rings of small GTPase activity that regulate the assembly of F-actin and myosin-2 around wounds and, further, that these rings are in turn controlled by microtubules. In addition to wound healing, the results may be broadly relevant to other transiently assembled, actomyosin-based contractile structures such as cytokinetic rings. (Supported by NIH GM52932.)

45. Myosin Sequences from Evolutionarily Divergent Plants MAGDALENA BEZANILLA and RALPH S. QUATRANO, Department of Biology, Washington University, St. Louis, MO (Sponsor: Francisco Bezanilla)

The myosin superfamily consists of at least 17 distinct families. These families are defined by homology within the catalytic domain. Interestingly, myosins from plants appear to group into three families with representatives only from plants. However, less plant myosin sequences are available as compared with animals, fungi, and protozoa. The majority of plant myosins come from a single organism, Arabidopsis. Thus, the plant myosin representation must be broadened in order to determine if plants have evolved a completely different set of myosins as compared with the other kingdoms. We have taken a degenerate PCR approach to amplify genomic sequences from a wide variety of plants. The product amplifies a highly conserved region in the catalytic domain from the peptide GESGAGK to the peptide LEAFGNAK. These primers have been used by others to identify large number of myosin families in other organisms (Bement et al. 1994. Proc. Natl. Acad. Sci. 91:6549–6553). We have obtained PCR products from ferns, mosses, monocots, and dicots. A common feature for all the genomic sequences obtained for class XI to date is the presence of an intron between the exons containing GESGAGK and LEAFGNAK. In contrast, the class VIII myosins lack this intron. Class XI myosins appear to be found in most plants ranging from the single cell Chlamydomonas to higher plants such as maize and Arabidopsis. Interestingly mosses only appear to have class VIII myosins. Since the moss Physcomitrella patens is efficient at homologous recombination, we have cloned genomic fragments of the four moss myosin VIII's and are in the process of disrupting these loci. We hope to learn what function class VIII myosins have in moss. In parallel, we have screened a collection of insertional mutants in Arabidopsis for knockouts in class VIII and class XI myosins. We have candidate seeds, which we are in the process of screening to identify the mutant plant. Together these studies will focus on determining the role of these evolutionarily distinct myosins. (Supported by the Helen Hay Whitney Foundation.)

46. N⁶-Modified ADP Analogues Block Adaptation of Hair–Cell Transduction in Mice Transgenic for Y61G Myo1c (Myosin I) PETER G. GILLESPIE,¹ JEFFREY R. HOLT,² KEVAN S. SHOKAT,³ and JOHN A. MERCER,^{4 1} Oregon Hearing Research Center and Vollum Institute, Oregon Health Sciences University, Portland, OR; ² Department of Neurobiology, Harvard University, Boston, MA; ³ Department of Cellular and Molecular Pharmacology, University of California, San Francisco, CA; ⁴ McLaughlin Research Institute, Great Falls, MT

In hair cells, the sensory cells responsible for auditory and vestibular mechanotransduction, sustained deflections of their mechanosensitive hair bundles evoke receptor currents that adapt over fast (0.5–5 ms) and slow (20–200 ms) time scales. To test whether Myo1c (myosin I) is the force-producing molecule responsible for slow adaptation, we designed a mutation in Myo1c that maintains near-normal ATP hydrolytic and chemomechanical activity, yet allows selective inhibition to the mutant Myo1c. Using N⁶(2-methylbutyl) ADP as a selective inhibitor, Myo1c mutated at Tyr-61 to Gly (Y61G-Myo1c) is 60-fold more sensitive to inhibition of ATP-ase activity and 400-fold more sensitive to inhibition of in vitro motility than is wild-type Myo1c. We have expressed Y61G-Myo1c as a transgene in mice, using 6 kb of 5′ genomic sequence to allow control of expression by the natural Myo1c promoter. When we recorded from hair cells expressing Y61G-Myo1c while dialyzing with N⁶(2-methylbutyl) ADP, we found that currents and adaptation were indistinguishable from those of wild-type over the first few minutes of the recording. However, in over 30 Y61G-transgenic hair cells analyzed in detail, we observed significant alteration in the properties of adaptation 5–7 min after entering whole-cell mode. Over 40% of these cells completely lost adaptation to positive deflections; in those cells that retained adaptation, the time constant for adaptation increased fivefold. Adaptation to negative displacements disappeared in over 90% of the Y61G-positive hair cells that displayed such adaptation. By contrast, in the presence of N⁶(2-methylbutyl) ADP, adaptation to positive and negative deflections remained robust in wild-type hair cells and hair cells transgenic for wild-type Myo1c. In conclusion, acute inhibition of Y61G-Myo1c inhibits adaptation in mammalian hair cells, suggesting that wild-type Myo1c participates in adaptation. (Supported by NIH grants DC03279 and DC02368.)

47. The Role of Type I Myosins in Assembly of Actin Filaments at Cell Cortex TERRY LECHLER, GUDRUN JONSDOTTIR, and RONG LI, Department of Cell Biology, Harvard Medical School, Boston, MA

Type I myosins are a class of unconventional myosins conserved from yeast to human. There are two myosin-I's in budding yeast: Myo3p and Myo5p. Previous studies showed that these myosins localize to cortical actin patches and have redundant functions in the establishment of a polarized actin distribution during bud formation (Goodson et al. 1995. Cell Motil. Cytoskeleton. 30:73; Goodson et al. 1996. J. Cell Biol. 133:1277). To understand the role of the yeast myosin-I's in actin organization, we analyzed their interaction with actin nucleation factors. We and others found that these myosins bind to Bee1p, the yeast WASP-like protein, and Vrp1p, the yeast homologue of WASP-interacting protein (WIP; Evangelista et al. 2000. J. Cell Biol. 148:353; Lechler et al. 2000. J. Cell Biol. 148:363). The COOH termini of these myosins interact directly with the Arp2/3 complex, a highly conserved actin nucleator. Furthermore, a complex of myosin-I and Vrp1p is able to activate the nucleation activity of the Arp2/3 complex. Using an in vitro actin assembly assay in permeabilized yeast cells, we found that the motor activity of these myosins are also required for actin polymerization at cortical patches. We have also obtained preliminary evidence which suggests that the function of myosin-I in actin polymerization may be conserved in animal cells. We hypothesize that myosin-Is are important for maintaining active actin nucleation factors at the leading edge during lamellipodia protrusion. (Supported by NIH grant GM57063.)

48. Cardiac Myosin-binding Protein C (MyBP-C): Binding Affects Thick Filament Structure and Ventricular Function RHEA J.C. LEVINE,¹ IRINA KULIKOVSKAYA,² GEORGE McCLELLAN,² MATHIAS GAUTEL,³ and SAUL WINEGRAD,^{2 1} Department of Neurobiology and Anatomy, MCP Hahnemann University, Philadelphia, PA; ² Department of Physiology, University of Pennsylvania, Philadelphia, PA; ³ University of Dortmund, Dortmund, Germany

Cardiac MyBP-C, like that of skeletal muscles, exists in seven bands in each one-half A-band. It differs from the two skeletal isoforms in that it has an additional “module” at the NH₂ terminus (C0), three accessible phosphorylatable sites in the cardiac-specific link between modules C1 and C2 and additional residues in C5. The COOH-terminal module (C10) binds to the LMM region of myosin and one or more sites in the NH₂ terminus can bind to myosin S2; the latter binding is predominantly seen with unphosphorylated MyBP-C. We reported that the extent of MyBP-C phosphorylation in intact rat ventricular strips and papillary muscles is a function of the external calcium concentration, and that this affects thick filament structure, extent of interaction between thick and thin filaments and maximum force development (Fmax; McClellan et al. 2001. Biophys. J. In press; Levine et al. 2001. Biophys. J. In press). We also reported the effect of removal of MyBP-C on thick filament structure (disorder and fragmentation), Fmax (decrease), and myosin retention (lost, especially with lack of stimulation; McClellan et al. 2001. Biophys. J. 80:261).

We have examined the extent and effect of binding of expressed, unphosphorylated NH₂-terminal fragments of MyBP-C consisting of different modules (C0C1), (C0C2), (C1C2) and (C2C5) to skinned ventricular strips and isolated thick filaments in which the endogenous MyBP-C was either unphosphorylated or phosphorylated. The physiological effect of binding was determined on strips after extensive washout; the extent of binding of the fragments was determined on Western blots of SDS gels of such strips; and the structural effect was assessed by incubation of thick filaments separated from strips with both phosphorylated and unphosphorylated endogenous MyBP-C with the different fragments, respectively. Binding of C1C2 occurred in strips with phosphorylated endogenous MyBP-C, where it decreased Fmax. Fragments containing C0 were bound in strips with unphosphorylated protein and did not affect Fmax. Filament structure was affected by binding of C0C1 whether the endogenous protein was phosphorylated or not, while C1C2 had an effect only on filaments with phosphorylated MyBP-C (Kulikovskaya et al. 2001. Biophys. J. In press; Levine et al. 2001. Biophys. J. In press). Here we present the analysis of the structural changes seen in electron microscopy by optical diffraction of negatively stained preparations and also report filament structures seen after incubation with the constructs: C0C2, which includes the cardiac specific NH₂ terminus and the cardiac-specific linker; and C2C5, which does not include either cardiac-specific region of MyBP-C. These will be correlated with the extent, and the physiological effects, of binding of these fragments to muscle with unphosphorylated and phosphorylated endogenous MyBP-C. Thus far it appears as if phosphorylation of MyBP-C, thus, its release from binding sites in myosin S2, significantly increases Fmax and may be an important modulator of cardiac contractility. (Supported by an NIH grant to S. Winegrad.)

49. Myosin VI Plays a Role in Restricting Organelle Movement MICHAEL A. MANDELL, MICHELE C. KIEKE, and MARGARET A. TITUS, Department of Genetics, Cell Biology and Development, University of Minnesota, Minneapolis, MN

The class VI myosins (M6) have been implicated in particle and membrane movement in the mouse and fly. Deletion of a C. elegans M6 (SPE-15) results in defective asymmetric segregation of organelles and cytoskeletal elements required for formation of a motile sperm that occurs during spermatogenesis (Kelleher et al. 2000. Curr. Biol. 10:1489–1496), and also causes male sterility. The potential role of SPE-15 in organelle transport was addressed by real-time imaging of organelle movements during spermatid budding. Secondary spermatocytes (SS) undergoing the process of spermatid budding were obtained either from adult wild-type males or spe-15(ok136) males lacking M6. Budding was observed by differential interference contrast (DIC) microscopy over the course of 90 min. As the spermatids bud, organelles are moved into the growing spermatid and the remaining residual body (RB) appears transparent. However, in spe-15(ok136) SS, organelles appear to be segregated into the spermatids, but are then released back into the RB. Subsequently, the spermatids undergo abnormal contractions. The overall time required for spermatid budding also appears to be decreased in the spe-15(ok136) SS—the formation of a nascent spermatid occurs in about half of the time that it takes in the case of wild-type SS. These results suggest that M6 is not required for the translocation of organelles, but rather it restricts their localization to the budding spermatid. M6 may also provide an opposing force to prevent premature budding and spermatid deformation during the final stages of budding. Ongoing analysis of spermatogenesis, in combination with localization studies of M6, will allow us to define how this motor performs these roles. (Supported by grants from the American Heart Association the National Institutes of Health.)

50. Visualizing Single Myosin-I Molecules in Living Cells GREGORY I. MASHANOV,¹ DARYL TACON,² MICHELLE PECKHAM,² and JUSTIN E. MOLLOY,^{1 1} Biology Department, University of York, York, United Kingdom; ² School of Biomedical Science, University of Leeds, United Kingdom

The single-headed myosin-I's are thought to be involved in diverse functions such as membrane protrusion and tensioning, endocytosis and vesicle transport (Ruppert et al. 1995. J. Cell Sci. 108:3775). Here we have studied the cellular dynamics of the 130kD myosin-I, (rat myr1b). Myr1b has an NH₂-terminal motor domain, a regulatory domain with 5 IQ motifs and a COOH-terminal lipid binding domain. Its slow biochemical kinetics (Coluccio and Geeves. 1999. J. Biol. Chem. 274:21575) and ADP-induced shape change (Veigel et al. 1999. Nature. 398:530) and localization in cells suggest that it may be important in maintaining tension between membrane and cytoskeleton.

We used total internal reflection fluorescence microscopy (TIRF) to investigate the behavior of myr1b fused to GFP in cultured mouse myoblasts. Expression levels were low enough to visualize individual GFP-tagged molecules using TIRF. We used time-lapse recordings to identify molecules bound to the basal membrane of leading lamellae. We developed a computer algorithm to automatically detect single fluorophores using ‘DISH’ criteria: Diffraction-limited image size, Intensity similar to single fluorophore control values, Single step photobleaching, Halftime proportional to illumination intensity. The last criterion distinguishes between diffusive motion of diffraction-limited objects or aggregates and single-step photobleaching. Our two-pass algorithm detects stepwise intensity changes of spots of appropriate size and intensity by performing a running t test on a diffraction limited pixel area over adjacent 1-s time windows of image data. Then the average lifetime of detected spots were measured at different laser powers.

The density of myr1b-GFP on the basal membrane of the leading lamellae was sufficiently low to allow the observation of single fluorophores, 0.5–1 molecule/μm². We are now developing methods to track the position of individual myr1b-GFP molecules in the leading lamellae of migrating myoblasts. (Myr1b clone was a gift of Martin Bähler; work funded by BBSRC.)

51. Functional Dissection of MyoM: A Myosin with a Rac-GEF Tail FRANK REICHARDT, FENGWEI WANG, HANS-JÖRG WARNATZ, and THIERRY SOLDATI, Department of Molecular Cell Research, Max-Planck Institute for Medical Research, Heidelberg, Germany

We recently described a novel unconventional myosin from Dictyostelium discoideum, MyoM (Schwarz et al. 1999. Cell Biochem. Biophys. 30:413–435; Geissler et al. 2000. Traffic. 1:399–410). Primary sequence analysis of full-length MyoM 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 surface loop near the actin-binding site (loop2) is extended by ∼70 residues compared with skeletal myosin II, and contains clusters of asparagines and serines, which might influence the binding to actin. Preliminary data from blot overlays indicate that the two canonical IQ motifs in the neck do not bind calmodulin. The phosphorylation of the TEDS site is currently tested with antibodies raised against the phosphorylated TEDS-peptide. The GEF domain displayed selective activity on a subset of Rac-related GTP-ases, but not Rho or Cdc42. The viability of MyoM null cells is decreased by hyper- and hypoosmotic shock. In addition, in response to osmotic stress, cells overexpressing the GEF domain extended massive actin-driven protrusions at up to 2 μm/s. The GEF was enriched at the tip of growing protuberances, through its pleckstrin homology domain. Our evidence shows that the PH domain binds protein ligand(s) instead of phosphoinositide lipids. 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.

52. Rab27a Enables Myosin Va-dependent Melanosome Capture by Recruiting the Myosin to the Organelle XUFENG WU,¹ KANG RAO,¹ M. BLAIR BOWERS,¹ NEAL G. COPELAND,² NANCY A. JENKINS,² and JOHN A. HAMMER III,^{1 1} Laboratory of Cell Biology, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD; ² Mouse Cancer Genetics Program, National Cancer Institute-Frederick, Frederick, MD

The peripheral accumulation of melanosomes characteristic of wild-type mouse melanocytes is driven by a cooperative process involving long range, bidirectional, microtubule-dependent movements coupled to capture and local movement in the actin-rich periphery by myosin Va, the product of the dilute locus. Genetic evidence suggests that Rab27a, the product of the ashen locus, functions with myosin Va in this process. Here we show that ashen melanocytes, like dilute melanocytes, exhibit normal dendritic morphology and melanosome biogenesis, an abnormal accumulation of end-stage melanosomes in the cell center, and rapid, bidirectional, microtubule-dependent melanosome movements between the cell center and the periphery. This phenotype suggests that ashen melanocytes, like dilute melanocytes, are defective in peripheral melanosome capture. Consistent with this, introduction into ashen melanocytes of cDNAs encoding wild type and GTP-bound versions of Rab27a restores the peripheral accumulation of melanosomes in a microtubule-dependent manner. Conversely, introduction into wild-type melanocytes of the GDP-bound version of Rab27a generates an ashen/dilute phenotype. Rab27a colocalizes with end-stage melanosomes in wild-type cells, and is most concentrated in melanosome-rich dendritic tips, where it also colocalizes with myosin Va. Neither endogenous myosin Va nor an expressed, GFP-tagged, myosin Va tail domain fusion protein colocalize with melanosomes in ashen melanocytes, in contrast to that seen previously in wild type cells. These results argue that Rab27a serves to enable the myosinVa-dependent capture of melanosomes delivered to the periphery by bidirectional, microtubule-dependent transport, and that it does so by recruiting the myosin to the melanosome surface. We suggest that Rab27a in its GTP-bound and melanosome-associated form predominates in the periphery, and that it is this form that recruits the myosin, enabling capture. These results argue that Rab27a serves as a myosin Va “receptor,” and add to the growing evidence that Rab GTPases regulate vesicle motors as well as SNARE pairing. Finally, we discuss recent efforts to elucidate the nature of the physical connection between myosin Va and Rab27a.

53. The Isolation and Characterization of KIN5, a Novel Kinesin-II in Tetrahymena thermophila AASHIR AWAN, MITCHELL BERNSTEIN, AARON BELL, TOSHIKAZU HAMASAKI, and PETER SATIR, Albert Einstein College of Medicine, Bronx, NY

Kinesin-like proteins (klps) are motor proteins that drive microtubule directed movement in cellular processes such as mitosis, meiosis, and vesicular transport. We are studying the roles of klps in eukaryotic cilia. Tetrahymena klps were cloned by PCR, using degenerate primers against the highly conserved regions of the kinesin motor. Using Tetrahymena genomic DNA as a template, PCR yielded several products. Sequencing of the subcloned PCR products revealed the presence of six different Tetrahymena klps. Kin6, Kin7, and Kin9 did not fall into any subfamily of klps, while Kin8 showed identity to C-terminal calmodulin-binding klps. Two of the klps, Kin2 and Kin5, are related to a Chlamydomonas flagellar klp, Fla10, that is required for the anterograde intraflagellar transport (IFT) in flagellar biosynthesis and maintenance (Kozminski et al., 1993). Fla10 is a member of the KRP85/95 kinesin-II subfamily of klps that functions in biogenesis of sensory and motile cilia and in vesicular transport in higher eukaryotic cells. Of the two isolated Fla10-related klps, one klp (Kin5) was a novel Tetrahymena gene product and one klp had been identified previously in the laboratory of Dr. Jacek Gaertig (University of Georgia). Taken together, our results and those of Dr. Gaertig show the presence of at least nine Tetrahymena klps. Our objective is to determine the role of the Kin5 protein in Tetrahymena. The Kin5 motor domain shows 62.6% identity to the Fla10 motor domain. We have cloned the entire 4.1-kb genomic sequence of KIN5 and have generated a knockout mutant. The kin5 knockout mutant seems to exhibit a phenotype of a cytokinesis/mitosis defect inconsistent with kin1/kin2 knockout (Gaertig et al. 1999). This suggests the presence of distinct kinesin-II populations with different roles. These studies of Kin5 and other kinesin-II klps in Tetrahymena will help us understand the role of this subfamily. (Supported by NSF grant MCB 9808654.)

54. In Vitro Dissection of Microtubule-based Endocytic Vesicle Processing EUSTRATIOS BANANIS, JOHN W. MURRAY, RICHARD J. STOCKERT, PETER SATIR, and ALLAN W. WOLKOFF, Marion Bessin Liver Research Center and Department of Anatomy and Structural Biology, Albert Einstein College of Medicine, Bronx, NY (Sponsor: David C. Spray)

After asialoglycoprotein receptor (ASGPR)–mediated endocytosis of asialoorosomucoid (ASOR) by he-patocytes, ligand and receptor dissociate within early endocytic vesicles that subsequently undergo fission. Receptor-containing vesicles recycle to the cell surface, whereas ligand-containing vesicles traffic to lysosomes. To dissect the mechanism regulating this process, early endocytic vesicles were purified from rat liver after injection of Texas red ASOR. Vesicles were drawn into a 3-μl motility chamber in which Taxol-stabilized, rho-damine-labeled microtubules (MTs) were attached. Interactions between vesicles and MTs were recorded in real-time before and after 50 μM ATP addition. In some studies, endosomes were incubated with antibodies to ASGPR, dynein, kinesins (pan specific), KIFC2 (minus-end kinesin), Kinesin II (plus-end), Rab 4 or Rab 5 before or following MT binding.

Ligand-containing vesicles were associated with ASGPR, kinesins, and Rab 4, but not with dynein or Rab5. 20–25% of microtubule-associated endocytic vesicles moved upon ATP addition, and 15–20% of motile vesicles underwent fission and sorting of receptor from ligand. Motility and fission were eliminated by 1 mM AMP-PNP (kinesin inhibitor) and by conventional kinesin antibodies but were unaffected by 20 μM vanadate (dynein inhibitor) and by dynein antibodies. Specific Kin II and KIFC2 antibodies revealed colocalization in these vesicles. In some experiments, after vesicles were bound to microtubules, buffer alone or buffer containing 4 mM GDP, 4 mM GTP-γ-S, or 4 mM GDP-β-S, was flowed into the motility chamber, incubated for 5 min, washed, and 50 μM ATP was added. As compared with buffer alone, GDP and GDP-β-S treatment increased the number of motile vesicles by 40%, and the number of vesicles undergoing fission by 100% (P < 0.001). Experiments with GTP-γ-S did not differ from control (P > 0.50). Immunofluorescence microscopy revealed that over 76% (519/681) of ASOR-containing MT bound vesicles contained Rab 4 and their motility was abolished by inclusion of anti-Rab 4. Pretreatment with 4 mM GDP, a condition in which motility and fission are enhanced, resulted in a loss of Rab 4 immunostaining with only 8% (55/682) of vesicles now being positive.

We have successfully reconstituted early endocytic vesicle microtubule-based motility and fission in vitro. Kinesins but not dynein are required for this process. Furthermore, GDP upregulates motility and fission, and this correlates with release of Rab 4 from the vesicles. (Supported by NIH grants DK41918 and DK41296.)

55. Control of Organelle Size by Motors: Kinesin-II–driven Intraflagellar Transport Determines Flagellar Length by Balancing Distal Tip Microtubule Turn-over WALLACE F. MARSHALL and JOEL L. ROSENBAUM, Department of Molecular, Cellular, and Developmental Biology, Yale University, New Haven, CT

A central question in cell biology is how cells determine size. Flagellar length control is a convenient system for studying organelle size regulation. Kinesin-II–driven intraflagellar transport (IFT) is needed to maintain flagella at full-length after assembly, suggesting it may act to balance ongoing turnover of flagellar protein. To determine the cytological localization of this turnover, we developed an assay to visualize flagellar tubulin turnover in Chlamydomonas. We found flagellar outer doublet microtubules undergo continuous turnover at their distal ends, suggesting that IFT is required to bring flagellar subunits to the distal tip to balance the ongoing turnover. When IFT was blocked by a conditional mutation in the kinesin-II motor subunit fla10, the incorporation of new tubulin at the tip ceased but the outer doublets continued to turn over, leading to resorption of the flagellum. We investigated the connection between IFT, turnover, and flagellar length, using mutants affecting IFT and length regulation. Partial loss of fla10 function caused flagella to equilibrate at half their wild-type length. Measurements on a long flagellar mutant lf2 indicated that the long flagellar phenotype was accompanied by a reduction in the turnover rate at the tip. Combining our data, we find that the properties of IFT and turnover lead directly to a simple mechanism of length control that can account both for length mutants and for the ability of flagella in a single cell to equalize their lengths. (Supported by NIH grant 14642 and the Helen Hay Whitney Foundation.)

56. Axonal Transport of Mitochondria Requires Milton, a Novel Kinesin-associated Protein LAURA J. MEGEATH,¹ R. STEVEN STOWERS,² JOLA GORSKA,³ IAN A. MEINERTZHAGEN,³ and THOMAS L. SCHWARZ,^{1 1} Division of Neuroscience, Children's Hospital, and Department of Neurobiology, Harvard Medical School, Boston, MA; ² Department of Molecular and Cell Biology, University of California, Berkeley, CA; ³ Life Sciences Centre, Dalhousie University, Halifax, Nova Scotia, Canada

Molecular motors are hypothesized to use adaptor proteins for identifying and binding specific cargoes, determining destinations, and anchoring cargoes after transport. We have cloned a novel Drosophila protein, milton, that may serve as an adaptor protein between kinesin and mitochondria.

Milton was identified as a lethal genetic locus that, when made homozygous in the photoreceptors of the fly, prevented normal phototactic behavior despite normal responses of the photoreceptors to light. Milton expression is not restricted to photoreceptors; it is present throughout development though it is enriched in heads and in 0–2-h embryos. The locus encodes a novel protein of 1100 amino acids that has two homologues in the human genome. Kinesin heavy chain (KHC) was identified by mass spectroscopy of proteins co-purifying with milton. This association between milton and KHC was confirmed in coimmunoprecipitation experiments. Ultrastructural analysis of Drosophila photoreceptors lacking milton (mil−/−) revealed a defect in mitochondrial transport. mil−/− photoreceptors had no mitochondria in axons or terminals, though the cell bodies contain numerous mitochondria that are normal in appearance. mil−/− photoreceptors were ultrastructurally normal in other respects, including the presence of synapses and synaptic vesicles. Electroretinograms, however, indicate that synaptic transmission is severely compromised in mil−/− photoreceptors. Furthermore, milton protein cosediments with the mitochondrial marker cytochrome c on a sucrose gradient. We propose that milton is an adaptor protein for mitochondrial transport by kinesin and that without milton, nerve terminals malfunction due to a lack of mitochondria. (Supported by NIH grants MH48108 and NS41062 [to T.L. Schwarz], NS07473 [to L.J. Megeath], and EY03592 and MRC MT132Y2 [to I.A. Meinertzhagen].)

57. Intraflagellar Transport (IFT) Is Required for the Assembly and Maintenance of All Motile and Sensory (Primary) Cilia and Is Related to PKD, RP, and Situs Inversus JOEL L. ROSENBAUM,¹ GEORGE B. WITMAN,² GREG PAZOUR,² DOUGLAS COLE,³ JOSEPH BESHARSE,⁴ HONGMIN QIN,¹ and MAUREEN BARR,^{5 1} Department Molecular, Cellular and Developmental Biology, Yale University, New Haven, CT; ² Department of Cell Biology, University of Massachusetts School of Medicine, Worcester, MA; ³ Department of Biochemistry, University of Idaho, Moscow, ID; ⁴ Department of Cell Biology, Medical College of Wisconsin, Milwaukee, WI; ⁵ Department of Pharmacy, University of Wisconsin, Madison, WI

The bidirectional movement of multi-polypeptide particles underneath the flagellar membrane is called intraflagellar transport (IFT). It is required for the assembly and maintenance of all sensory and motile cilia and flagella, and is hypothesized to be the mechanism by which pre-assembled axonemal precursors are moved out to their tip assembly sites, and also for the movement of proteins into the ciliary membranes. The anterograde movement is powered by kinesin-II, and the retrograde movement by cytoplasmic dynein 1b. The IFT particles occur in chains of varying length called rafts, which are attached to the overlying ciliary membrane and the B-subfiber of the outer doublet microtubules.

IFT was first observed in the flagella of the green alga, Chlamydomonas where it was shown that the particles are composed of two complexes, A and B, containing up to 16 polypeptides. Since its initial observation, IFT has been documented in sea urchin cilia, the mechanosensory cilia of Drosophila and vertebrate primary cilia including the connecting cilia of the rods and cones, and the primary cilia on the cells of collecting tubules of the kidney. In C. elegans, IFT has been visualized directly in the sensory cilia by use of GFP-tagged constructs of kinesin and IFT genes.

Mutations affecting IFT in vertebrates have been linked to PKD (polycystic kidney disease), and Tg737 mice with autosomal recessive PKD, have a defect in IFT and lack normal primary cilia in the kidney. In C. elegans, homologues of the polycystins, the calcium dependent cation channel polypeptides which are defective in PKD, are localized on the primary sensory cilia. It is hypothesized from this that vertebrate PKD can result if the primary cilia are no longer present or abnormal so that the polycystins cannot be targeted to the ciliary membranes.

Defects in IFT in vertebrates have also been shown to cause lack of the cilia on nodal cells in the mouse (Henson's node in chick). If the nodal cilia are abnormal, or are unable to rotate, situs inversus results, i.e., asymmetry of organ placement on the right versus left side of the midline.

IFT is also important in the assembly and maintenance of rod and cone outer segments, both derivatives of primary cilia. The connecting cilia contain both the motors and the IFT particle polypeptides characteristic of IFT. Targeted knockout of the kinesin-II which supports IFT in the retina causes degeneration of rods and cones.

Since primary cilia are present on almost all vertebrate cells, it is quite possible that the membranes of these cilia are repositories for specific channels (e.g., polycystins) and receptors. For example, in the vertebrate brain, the somatostatin receptors are localized only on the primary cilia membranes in a variety of brain cells. Presumably, other sensory functions for the primary cilia in vertebrates will be determined.

58. Overexpression of Dynamitin in Motor Neurons Causes Progressive Degeneration and Muscle Atro-phy BERNADETTE H. LAMONTE, BETH A. HOLLOWAY, KAREN E. WALLACE, MARIKO TOKITO, and ERIKA L. F. HOLZBAUR, Department of Animal Biology, University of Pennsylvania School of Veterinary Medicine, Philadelphia, PA

Defects in axonal transport have been suggested as an underlying cause for the degeneration of motor neurons observed in diseases such as amyotrophic lateral sclerosis (ALS). SOD-1 and NF-H transgenic mouse models have demonstrated defects in axonal transport early in the development of disease. To test the hypothesis that defective axonal transport may be a critical component of the mechanism of disease progression, we sought to disrupt the function of the axonal motor dynein and its activator dynactin. To inhibit dynein/dynactin function in vivo, we generated transgenic mice that overexpress the dynamitin (p50) subunit of dynactin in mature motor neurons, as overexpression of this protein has been previously shown to inhibit dynein-mediated transport processes at the cellular level. Transgenic mice exhibited highest levels of dynamitin expression in motor neurons, resulting in neurofilament aggregates and axonal degeneration. These mice display a phenotype of late onset motor neuron degeneration, characterized by hind end weakness, trembling, abnormal gaits, and decreased stride length. Together, these observations support the hypothesis that disruptions in axonal transport mediated by dynein may result in motor neuron degeneration and may be involved in the pathogenesis of motor neuron diseases such as sporadic ALS. (Supported by NIH GM48661 and an Established Investigator Award from the American Heart Association to E.L.F. Holzbaur.)

59. Biochemical Characterization of Yeast Dynein/Dynactin Components WEI-LIH LEE, JESSICA R. OBERLE, RICHARD A. HEIL-CHAPDELAINE, and JOHN A. COOPER, Department of Cell Biology and Physiology, Washington University School of Medicine, Saint Louis, MO (Sponsor: Francisco Bezanilla)

Cytokinesis in budding yeast occurs at the narrow neck between mother and bud. To ensure equal partitioning of replicated chromosomes, the mitotic spindle must move from the mother cell into the neck. This occurs in two steps: First, the nucleus moves to a region adjacent to the neck; Second, the nucleus and mitotic spindle move into the neck. During the second step of spindle movement, cytoplasmic microtubules extend from the spindle pole body into the bud, associate laterally with and slide along the bud cortex, pulling the nucleus and the elongating spindle into the neck (Adames and Cooper. 2000. J. Cell Biol. 149:863–874). This microtubule sliding event depends on dynein/dynactin components (Dyn1p, Arp1p, Jnm1p, and Nip100p) and the cortical protein Num1p. Despite complete genome sequence information, identification of all yeast dynactin components has been difficult due in part to low sequence similarity with their vertebrate counterparts. Phylogenetic analyses of Jnm1p and Nip100p, for examples, do not conclusively classify them as yeast dynactin components. We have taken a biochemical approach to isolate yeast dynactin and its associated proteins, by affinity tagging the putative components as well as other gene products which are required for normal nuclear segregation. Preliminary studies have shown that myc-tagged Nip100p is functional and soluble when lysed in buffer containing 2% Triton X-100. Subsequent purification of Nip100-myc and its associated proteins will help identify the yeast dynactin components, the putative regulator of dynein during microtubule sliding events. (Supported by NIH grant GM47337 and the Cancer Research Fund of the Damon Runyon-Walter Winchell Foundation Fellowship, DRG-1671.)

60. PLAC-24: A Novel Protein that Binds to Dynein Intermediate Chain and Is Recruited to Developing Adherens Junctions in Epithelial Cells LEE A. LIGON, SHER KARKI, MARIKO TOKITO, and ERIKA L.F. HOLZBAUR, Department of Animal Biology, University of Pennsylvania, Philadelphia, PA

Cytoplasmic dynein provides motive force for a range of cellular functions including vesicle trafficking and spindle assembly, yet the mechanisms governing the intracellular targeting and regulation of the motor are not well-understood. To address these questions, we screened for polypeptides that bind to dynein intermediate chain. We identified a 24-kD protein (PLAC-24) that binds to cytoplasmic dynein but is not a component of the dynactin complex. Competition assays demonstrate that PLAC-24 and the p150^Glued subunit of dynactin have distinct binding sites in the dynein intermediate chain. Immunocytochemistry using PLAC-24-specific antibodies revealed a perinuclear distribution of the polypeptide in fibroblasts and isolated epithelial cells. However, as epithelial cells make contact with adjacent cells, PLAC-24 is recruited to the cortex at sites of contact, where the protein colocalizes with β-catenin and E-cadherin. Prominent PLAC-24 staining is observed coincident with the ends of stress fibers. Disruption of the cellular cytoskeleton with nocodazole or latrunculin indicates that the subcellular localization of PLAC-24 is dependent on intact actin filaments but is not dependent on microtubules. Based on these data and the recent observation that dynein is localized to sites of cell–cell contact in epithelial cells, we suggest that PLAC-24 may be a part of a multiprotein complex at adherens junctions that functions to tether microtubules to the actin-rich cellular cortex. (Supported by NIH GM48661 and AHA 0150418N to E.L.F. Holzbaur and an NIH postdoctoral fellowship to L.A. Ligon.)

61. EB1 Binds to the p150^Glued Subunit of Dynactin and Stabilizes Microtubules in Cultured Epithelial Cells LEE A. LIGON, SPENCER SHELLY, MARIKO TOKITO, and ERIKA L.F. HOLZBAUR, Department of Animal Biology, University of Pennsylvania, Philadelphia, PA

Several proteins, including APC, EB1, CLIP-170, and dynactin, have recently been shown to localize to the plus-ends of microtubules. However, neither the possible interactions among these proteins, nor the function of this plus-end localization is well-understood. Using affinity chromatography, we have shown that EB1 binds directly to the p150^Glued subunit of dynactin, but not to the alternatively spliced p135 subunit. It previously has been shown that the overexpression of p150^Glued bundles microtubules in cultured cells (Waterman-Storer et al. 1995. Proc. Natl. Acad. Sci. 92:1634–1638; Quintyne et al. 1999. J. Cell Biol. 147:1–14). Using immunocytochemistry, we show here that EB1 is recruited to these bundles, where it colocalizes with p150^Glued. Overexpression of EB1 also robustly bundles microtubules. These bundles are cold- and drug-stable, and immunocytochemistry shows that endogenous p150^Glued is recruited to these bundles. These data demonstrate a direct interaction between EB1 and the p150^Glued subunit of dynactin that may form the basis of a microtubule plus-end complex. Further, they suggest that EB1 may serve to stabilize microtubules in vivo. (Supported by NIH GM4866 and an NIH postdoctoral fellowship award to L.A. Ligon.)

62. km23: A TGFβ Signaling Intermediate that Interacts with Dynein QIAN TANG, CORY M. STAUB, and KATHLEEN M. MULDER, Department of Pharmacology, Pennsylvania State College of Medicine, Hershey, PA

TGFβ is a natural growth inhibitor for epithelial cells and components of TGFβ signaling pathways can function as tumor suppressors. However, the identification of specific cytoplasmic signaling components and pathways which mediate the diverse cellular responses of TGFβ has been difficult. The majority of those identified mediate transcriptional responses. Here we describe the identification of a novel component of the TGFβ receptor signaling pathway, termed km23. This 11-kD cytoplasmic protein interacts with the TGFβ receptors intracellularly and is phosphorylated by them. Stable expression of km23 results in a cellular phenotype that is similar to TGFβ treatment of epithelial cells. We demonstrate that km23 interacts with cytoplasmic dynein and that this interaction is rapidly induced by TGFβ in epithelial cells. In fact, km23 belongs to a superfamily of dynein-associated proteins that includes the Drosophila roadblock and bithoraxoid genes. Since cytoplasmic dynein is a molecular motor which facilitates the transport of cargo along polarized microtubules toward the minus ends, our results suggest that km23, and therefore TGFβ, may regulate specific dynein functions important for epithelial cell polarity. This is the first demonstration of a link between cytoplasmic dynein microtubule motors and a natural growth inhibitor. Further, we have identified amino acid alterations in km23 in human ovarian cancer cells within specific functional domains. (Supported by NIH grants CA51452, CA54816, CA68444, and CA90765 to K.M. Mulder.)

63. Independent Roles for Cytoplasmic Dynein IC and P150^Glued Phosphorylation in Dynein-dependent Membrane Transport PATRICIA S. VAUGHAN,¹ PE-DRO MIURA,¹ MATTHEW HENDERSON,¹ JOHN D. LESZYK,² and KEVIN T. VAUGHAN,^{1 1} Department of Biological Sciences, University of Notre Dame, Notre Dame, IN; ² Department of Biochemistry, University of Massachusetts Medical School, Worcester, MA (Sponsor: Dr. Criss Hartzell)

Previous studies have implicated the dynactin complex as a receptor or adaptor for cytoplasmic dynein on the surface of cargo. This interaction is mediated by the cytoplasmic dynein intermediate chains (ICs) and p150^Glued. To identify regulatory mechanisms that control the binding of cytoplasmic dynein to cargo, we mapped phosphorylation sites on the ICs and p150^Glued, and determined the impact of phosphorylation on protein function. A p150^Glued phosphorylation site was identified in the NH₂-terminal “head” domain through mapping of the epitope of a phospho-sensitive antibody. This phosphorylation site resides near the microtubule-binding domain of p150^Glued, and fits the consensus for cAMP-dependent PKA. Using GFP-fusions, we visualized site-directed mutants in living cells and determined that p150^Glued phosphorylation regulates the dynamic targeting of dynactin to microtubule plus-ends. To test IC phosphorylation as a regulatory mechanism, we analyzed cytoplasmic dynein by 2-D gel analysis and detected two IC variants, one of which was eliminated by phosphatase treatment. To measure the effect of this IC phosphorylation on the IC-p150^Glued interaction, we performed overlay assays and observed that p150^Glued bound dephosphorylated but not phosphorylated ICs. To map the site of IC phosphorylation, we subjected the purified cytoplasmic dynein ICs to mass spectrometry. A single phosphorylated tryptic fragment was identified corresponding to the p150^Glued-binding domain of IC-2C. Fragmentation and retention time analysis were used to map the phosphorylation site to serine 84. Site-directed mutants designed to mimic the dephosphorylated (S84A) or phosphorylated (S84D) IC disrupted both in vitro and in vivo IC phosphorylation. S84A mutants bound p150^Glued in vitro and overexpression perturbed transport of dynein-dependent membranes. The S84D mutant diminished p150^Glued binding in vitro and did not disrupt dynein-mediated transport when expressed in vivo. These findings represent the first evidence that IC phosphorylation plays a fundamental role in the binding of cytoplasmic dynein to dynactin and cargo. (Supported by AHA grant 3009N.)

64. Fab Fragment of Tubulin Monoclonal Antibody YL1/2 Can Be a Specific Motility Inhibitor YUUKO WADA, TYSON VALENTIN, TOSHIKAZU HAMA-SAKI, and PETER SATIR Department of Anatomy and Structural Biology, Albert Einstein College of Medicine, Bronx, NY

Dynein is a molecular motor which performs a variety of cellular functions, such as ciliary beating and membrane transport. Oka et al. (1997. Cell Struct. Funct. 22:44–463) has demonstrated that reactivated demembranated flagella motility was inhibited by an antitubulin monoclonal antibody YL1/2. We applied the antibody to the in vitro microtubule gliding assays with dyneins and a kinesin to determine whether the antibody directly inhibits motor–microtubule interactions. Intact YL1/2 antibody halted both inner and outer arm axonemal dynein-mediated microtubule gliding and slowed cytoplasmic dynein-mediated gliding velocity of microtubules. Another antitubulin antibody, B-5–1-2, had little effect on any motor-driven microtubule motility. Since intact IgG is large in comparison to a tubulin dimer, we made Fab fragments from YL1/2 antibody (YL1/2 Fab). The Fab fragment halted axonemal dynein-supported microtubule gliding, and that the motility could be easily restored with perfusion with ATP buffer without the Fab fragments. However, the YL1/2 Fab failed to inhibit cytoplasmic dynein- or kinesin-supported microtubule gliding. The inhibition is different from the known inhibitory effect of vanadate in the following points. First, vanadate inhibits both axonemal and cytoplasmic dynein at micromolar concentrations and requires much higher concentrations to inhibit kinesin activity, whereas YL1/2 Fab does not inhibit either cytoplasmic dynein or kinesin activity. Second, the inhibition of YL1/2 Fab is easily removed simply by perfusion of motility buffer without the Fab molecules, whereas that by vanadate sustains after a perfusion of the buffer; catecholamine, which oxidize vanadate, is required to remove the inhibitions. The epitope of the YL1/2 antibody may play an important role in motor-MT coupling in axonemes. The work will help us to develop a new generation of inhibitors for study of motor proteins in vitro and in vivo.

65. Control of Flagellar Dynein Activity by Phosphorylation: Localization of Kinases and Calmodulin in the Radial Spokes PINFEN YANG, ANNE ROUSH GAILLARD, and WINFIELD S. SALE, Department of Cell Biology Emory University School of Medicine, Atlanta, GA (Sponsor: H. Criss Hartzell)

Physiological and genetic evidence revealed that flagellar radial spokes regulate dynein activity through control of phosphorylation. We postulated that the radial spokes are molecular scaffolds that anchor protein kinases in position to control motility. Biochemical analysis of isolated Chlamydomonas flagellar axonemes and isolated radial spokes (obtained by a new procedure) revealed that: (1) calmodulin is located in the radial spokes; (2) radial spoke protein 2 (rsp2p) is a novel calmodulin-binding kinase; and (3) rsp3p is an A kinase anchoring protein (AKAP). These conclusions are based on diverse experimental evidence. Analysis of isolated spokes revealed calmodulin is located in the spoke stalk. In-gel kinase assays of axonemes and isolated radial spokes revealed that rsp2p displays kinase activity. RSP2 was cloned, and although the predicted amino acid sequence did not reveal domains typical for conventional kinases, purified, recombinant rsp2p undergoes autophosphorylation. The sequence also revealed two Ca²⁺-dependent calmodulin-binding motifs, and calmodulin binds rsp2p in a calcium-dependent manner. We postulate that calmodulin and rsp2p, located in the radial spoke stalk, play roles in calcium control of flagellar bending. Functional analysis has indicated PKA is anchored in the axoneme and controls dynein activity. In support of this, rsp3p binds labeled RII in blot overlays of isolated axonemes and recombinant rsp3p, indicating rsp3p is an AKAP. Analysis of truncated and mutant RSP3 clones revealed rsp3p contains a single RII binding site located at amino acids 161–178. We postulate that rsp3p, located at the base of the spoke stalk near the inner dynein arms, targets PKA in position to locally regulate dynein activity. We propose a new model in which the radial spokes act as mechanochemical transducers of calcium and/or mechanical signals (INPUT) for local control of kinases and dynein phosphorylation (OUTPUT). (Supported by NIH and the March of Dimes.)

Figure 1.

(Top) Longitudinal view of a superposition of seven typically bound myosin heads in the A•M•ATP state. (Bottom) Binding site for the A•M•ATP state compared with the binding site for the rigor state.