Novel heterotrimeric kinesin-related protein purified from sea urchin eggs

Expression of kinesin heavy chain isoforms in retinal pigment epithelial cells

To examine the possible role of kinesin in pigment granule migration in the retinal pigment epithelium (RPE) of teleosts, we investigated the expression and distribution of kinesin heavy chain (KHC) in RPE. Blots of fish RPE lysates probed with two well-characterized antibodies to KHC (H2 and HD) displayed a prominent band at 120 kD. A third KHC antibody (SUK4) recognized a band at 118 kD. The 118 kD band was also occasionally present in blots probed with H2, suggesting the presence of two KHC isoforms in teleost RPE. Reverse transcriptase-polymerase chain reaction (RT-PCR) of mRNA from RPE using primers homologous to conserved regions of the KHC motor domain resulted in the identification of two putative KHC genes (FKIF1 and FKIF5) based on partial amino acid sequences. Previous studies had demonstrated a requirement for microtubules in pigment granule aggregation in RPE. In addition, the reported microtubule polarity orientation in RPE apical projections is consistent with a role for kinesin in pigment granule aggregation. Immunofluorescent localization of KHC in isolated RPE cells using H2 revealed a mottled distribution over the entire cell body, with no detectable selective association with pigment granules, even in cells fixed while aggregating pigment granules. Microinjected KHC antibodies had no effect on pigment granule aggregation or dispersion, although each of the three antibodies has been shown to block kinesin function in other systems. Thus we found no evidence for KHC function in RPE pigment granule aggregation. However, the two KHC isoforms may participate in other microtubule-dependent processes in RPE.

Identification and partial characterization of mitotic centromere-associated kinesin, a kinesin-related protein that associates with centromeres during mitosis

Using antipeptide antibodies to conserved regions of the kinesin motor domain, we cloned a kinesin-related protein that associates with the centromere region of mitotic chromosomes. We call the protein MCAK, for mitotic centromere-associated kinesin. MCAK appears concentrated on centromeres at prophase and persists until telophase, after which time the localization disperses. It is found throughout the centromere region and between the kinetochore plates of isolated mitotic CHO chromosomes, in contrast to two other kinetochore-associated microtubule motors: cytoplasmic dynein and CENP-E (Yen et al., 1992), which are closer to the outer surface of the kinetochore plates. Sequence analysis shows MCAK to be a kinesin-related protein with the motor domain located in the center of the protein. It is 60-70% similar to kif2, a kinesin-related protein originally cloned from mouse brain with a centrally located motor domain (Aizawa et al., 1992). MCAK protein is present in interphase and mitotic CHO cells and is transcribed as a single 3.4-kb message.

Active sites of ligands and their receptors are made of common peptides that are also found elsewhere

The simultaneous emergence in evolution of a ligand and its receptor might have entailed their active sites being drawn from the pool of common oligopeptides. This was tested on the principal components of cell-matrix interaction: the RGD (Arg-Gly-Asp) site of matrix proteins and the EKKD (Gly-Lys-Lys-Asp) site of integrin cell-surface receptor. In the 32 diverse proteins scrutinized, which totalled 14,806 residues, there were 104 Arg-Gly dipeptides. Most common of the tripeptides beginning with Arg-Gly were Arg-Gly-Leu, Arg-Gly-Gly, and Arg-Gly-Asp; each was found in ten copies. RGD tripeptide was one of the commonest; the fortuitous presence of an RGD site was noted in two enzymes, fibrinogen, a pituitary hormone precursor, and a viral structural protein. The 32 proteins also contained 121 Lys-Lys dipeptides. Of the tetrapeptides centered by Lys-Lys, the commonest was Lys-Lys-Lys-Lys, in four copies. Second most common were Gly-Lys-Lys-Lys, Val-Lys-Lys-Leu, and Glu-Lys-Lys-Asp; each occurred in three copies. The fortuitous presence of an EKKD site was noted in three proteins--an intracellular transport protein, a pituitary hormone precursor and a protein of the cerebrospinal fluid. In most instances, protein-protein interaction between the fortuitously present active sites appears to bring about deleterious consequences. Occasionally, however, the fortuitous active site appears to confer a new function to a protein bearing it.

Organelle transport and sorting in axons

Recent advances of the past year in the field of organelle transport have documented the existence of numerous kinesin-related proteins and the presence of multiple conventional kinesins within neurons. Biochemical and genetic mutant analyses indicate that different kinesin superfamily members transport different organelles. In addition to microtubule-based systems, actin filaments and myosin motors are associated with organelle transport in axons. The great diversity of motor proteins suggests that they may play a role in sorting, in addition to transport.

The Chlamydomonas FLA10 gene encodes a novel kinesin-homologous protein

Many genes on the uni linkage group of Chlamydomonas affect the basal body/flagellar apparatus. Among these are five FLA genes, whose mutations cause temperature-sensitive defects in flagellar assembly. We present the molecular analysis of a gene which maps to fla10 and functionally rescues the fla10 phenotype. Nucleotide sequencing revealed that the gene encodes a kinesin-homologous protein, KHP1. The 87-kD predicted KHP1 protein, like kinesin heavy chain, has an amino-terminal motor domain, a central alpha-helical stalk, and a basic, globular carboxy-terminal tail. Comparison to other kinesin superfamily members indicated striking similarity (64% identity in motor domains) to a mouse gene, KIF3, expressed primarily in cerebellum. In synchronized cultures, the KHP1 mRNA accumulated after cell division, as did flagellar dynein mRNAs. KHP1 mRNA levels also increased following deflagellation. Polyclonal antibodies detected KHP1 protein in Western blots of purified flagella and axonemes. The protein was partially released from axonemes with ATP treatment, but not with AMP-PNP. Western blot analysis of axonemes from various motility mutants suggested that KHP1 is not a component of radial spokes, dynein arms, or the central pair complex. The quantity of KHP1 protein in axonemes of the mutant fla10-1 was markedly reduced, although no reduction was observed in two other uni linkage group mutants, fla9 and fla11. Furthermore, fla10-1 was rescued by transformation with KHP1 genomic DNA. These results indicate that KHP1 is the gene product of FLA10 and suggest a novel role for this kinesin-related protein in flagellar assembly and maintenance.

Molecular phylogeny of the kinesin family of microtubule motor proteins

The rapidly expanding kinesin family of microtubule motor proteins includes proteins that are involved in diverse microtubule-based functions in the cell. Phylogenetic analysis of the motor regions of the kinesin proteins reveals at least five clearly defined groups that are likely to identify kinesins with different roles in basic cellular processes. Two of the groups are consistent with overall sequence similarity, while two groups contain proteins that are related in overall structure or function but show no significant sequence similarity outside the motor domain. One of these groups consists only of kinesin proteins with predicted C-terminal motor domains; another includes only kinesins required for mitotic spindle bipolarity. Drosophila Nod, presently an ungrouped protein, may represent a class of kinesins that, like the myosin I proteins, function as monomers. The analysis indicates that many types of kinesin proteins exist in eukaryotic organisms. At least two of the five groups identified in this analysis are expected to be present in most, or all, eukaryotes.

A new kinesin-like protein (Klp1) localized to a single microtubule of the Chlamydomonas flagellum

The kinesin superfamily of mechanochemical proteins has been implicated in a wide variety of cellular processes. We have begun studies of kinesins in the unicellular biflagellate alga, Chlamydomonas reinhardtii. A full-length cDNA, KLP1, has been cloned and sequenced, and found to encode a new member of the kinesin superfamily. An antibody was raised against the nonconserved tail region of the Klp1 protein, and it was used to probe for Klp1 in extracts of isolated flagella and in situ. Immunofluorescence of whole cells indicated that Klp1 was present in both the flagella and cell bodies. In wild-type flagella, Klp1 was found tightly to the axoneme; immunogold labeling of wild-type axonemal whole mounts showed that Klp1 was restricted to one of the two central pair microtubules at the core of the axoneme. Klp1 was absent from the flagella of mutants lacking the central pair microtubules, but was present in mutant flagella from pf16 cells, which contain an unstable C1 microtubule, indicating that Klp1 was bound to the C2 central pair microtubule. Localization of Klp1 to the C2 microtubule was confirmed by immunogold labeling of negatively stained and thin-sectioned axonemes. These findings suggest that Klp1 may play a role in rotation or twisting of the central pair microtubules.

KIF3A is a new microtubule-based anterograde motor in the nerve axon

Neurons are highly polarized cells composed of dendrites, cell bodies, and long axons. Because of the lack of protein synthesis machinery in axons, materials required in axons and synapses have to be transported down the axons after synthesis in the cell body. Fast anterograde transport conveys different kinds of membranous organelles such as mitochondria and precursors of synaptic vesicles and axonal membranes, while organelles such as endosomes and autophagic prelysosomal organelles are conveyed retrogradely. Although kinesin and dynein have been identified as good candidates for microtubule-based anterograde and retrograde transporters, respectively, the existence of other motors for performing these complex axonal transports seems quite likely. Here we characterized a new member of the kinesin super-family, KIF3A (50-nm rod with globular head and tail), and found that it is localized in neurons, associated with membrane organelle fractions, and accumulates with anterogradely moving membrane organelles after ligation of peripheral nerves. Furthermore, native KIF3A (a complex of 80/85 KIF3A heavy chain and a 95-kD polypeptide) revealed microtubule gliding activity and baculovirus-expressed KIF3A heavy chain demonstrated microtubule plus end-directed (anterograde) motility in vitro. These findings strongly suggest that KIF3A is a new motor protein for the anterograde fast axonal transport.

Molecular motors and cell motility in the brain

The advent of video computer-enhanced microscopy has provided a new vision of cell migrations, growth cones, and fast axonal transport in the nervous system. In images obtained in studies of fast transport in isolated axoplasm from the squid giant axon, a virtual torrent of membrane traffic could be seen moving in both directions. Similarly, examination of growth cones and cell migrations in vitro and in vivo revealed properties of cell motility that were previously unsuspected. Evidence has accumulated that many of these activities are driven by a variety of microtubule and microfilament based motors.

Isolation and analysis of microtubule motor proteins

Isolation of microtubule motor proteins is needed both for the discovery of new motors and for characterization of the products of motor-related genes. The sequences of motor-related genes cannot yet be used to predict the mechanochemical properties of the gene products. This was illustrated by the first kinesin-related gene product to be characterized. Protein expressed from the ncd gene moved toward the minus ends of microtubules (Walker et al., 1990; McDonald et al., 1990), while kinesin itself moves toward the plus ends. Until the relationship between mechanochemical function and amino acid sequence is more thoroughly understood, biochemical isolation and characterization of microtubule motor proteins will remain essential. Two approaches for getting useful quantities of microtubule motor proteins have been used: isolation from cytosol as described under Section II above and isolation from bacteria carrying cloned motor protein genes in expression vectors. Bacterial expression of functional microtubule motors has been successful to date in only a few cases (Yang et al., 1990; Walker et al., 1990, McDonald et al., 1990). Additional progress is expected with the expression of cloned genes from viral vectors in cultured eukaryotic cells, but broad success has not yet been reported. Biochemical isolation of motors from their natural cytosol has some distinct advantages. One can have confidence that a given motor will be folded properly and have normal post-translational modifications. In addition, if it exists in vivo as a heteromultimer, a microtubule motor isolated from its native cytosol will carry with it a normal complement of associated proteins. Studies of such associated proteins will be important in learning how motors accomplish their tasks in vivo. Drosophila cytosol should be a rich source of microtubule motors. Drosophila carry at least 11 and perhaps as many as 30 genes that are related to kinesin (Stewart et al., 1991; Endow and Hatsumi, 1991). The work of Tom Hays' lab indicates that Drosophila carry more than nine dynein related genes (Rasmussen et al., 1994). Relatively little effort to isolate the products of these genes from cytosol has been made. The only work that I am aware of has produced a kinesin-like microtubule motor (D.G. Cole, K.B. Sheehan, W.M. Saxton, and J.M. Scholey, in progress) that may be the Drosophila homolog of Xenopus eg5 (Sawin et al., 1992). This isolation was straightforward, and efforts to identify additional motors are almost assured of success.