Download Heterodimerization of the Two Motor Subunits of the Heterotrimeric

J. Mol. Biol. (1995) 252, 157–162
COMMUNICATION
Heterodimerization of the Two Motor Subunits of the
Heterotrimeric Kinesin, KRP85/95
Dana J. Rashid, Karen P. Wedaman and Jonathan M. Scholey*
Section of Molecular and
Cellular Biology, Division of
Biological Sciences
University of California
Davis, CA 95616, USA
The heterotrimeric kinesin-related motor protein, KRP85/95 is assembled from
two kinesin-related polypeptides, SpKRP85 and SpKRP95, together with an
uncharacterized 115 kDa polypeptide. Here we report the deduced amino
acid sequence of SpKRP95, a close relative of SpKRP85. Both SpKRP85 and
SpKRP95 are predicted to have a tripartite domain organization consisting
of an N-terminal motor domain, a central stalk domain capable of coiled-coil
formation, and a second globular C-terminal domain. The sequences of the
central stalk domains predict that SpKRP85 and SpKRP95 should be capable
of forming heterodimeric coiled coils. Furthermore, SpKRP85-SpKRP95
complexes can be immunoprecipitated from a cell-free translation system,
providing direct evidence that SpKRP85 and SpKRP95 are capable of
heterodimerization.
7 1995 Academic Press Limited
*Corresponding author
Keywords: heterotrimeric kinesin; KRP85/95 ; motor protein
The heterotrimeric kinesin-related motor protein,
KRP85/95, is a member of the kinesin superfamily of
motor proteins (Bloom & Endow, 1994) that
hydrolyzes ATP and transports cargo towards the
plus ends of microtubules at approximately 0.4 mm
per second (Cole et al., 1992, 1993). Two of the three
subunits of this trimeric protein, SpKRP85
(Mr = 85 kDa) and SpKRP95 (Mr = 95 kDa), are
thought to function as kinesin-related motor
polypeptides, whereas a third 115 kDa subunit is
thought to be a non-motor accessory polypeptide,
possibly analogous to the kinesin light chains (Cole
et al., 1993). KRP85/95 has recently been immunolocalized to detergent-sensitive particles in the metaphase
half spindles and anaphase interzone of sea urchin
embryonic mitotic apparatuses, leading us to
hypothesize that this motor protein transports
vesicles that deliver new membrane to the developing
cleavage furrow of dividing cells (Henson et al.,
1995).
A novel feature of KRP85/95 is the apparent presence
of two distinct kinesin-related polypeptides in a
single holoenzyme (Cole et al., 1992, 1993), raising
questions concerning the molecular architecture of
KRP85/95. For example, we hypothesize that SpKRP85
and SpKRP95 could bind together within the
KRP85/95 holoenzyme, perhaps by forming heterodimeric coiled coils analogous to the homodAbbreviations used: SpKRP, Strongylocentrotus
purpuratus kinesin related polypeptide; mAb,
monoclonal antibody.
0022–2836/95/370157–06 $12.00/0
imeric coiled coil that binds the two identical motor
polypeptides of conventional kinesin (deCuevas
et al., 1992). However, it is also possible that SpKRP85
and SpKRP95 do not physically interact, but are held
together in the heterotrimeric complex by the
115 kDa subunit acting as a linker protein. Here we
report the cloning and sequencing of a cDNA
encoding full length SpKRP95 and we describe
sequence analysis and co-immunoprecipitation
experiments which support the hypothesis that
SpKRP85 and SpKRP95 are indeed capable of
heterodimerization.
A full length SpKRP85 cDNA and a partial cDNA
encoding approximately 40% of SpKRP95 were
previously described (Cole et al., 1993). The partial
SpKRP95 cDNA was used as a probe to obtain a full
length cDNA by plaque hybridization screening of
an unfertilized sea urchin egg cDNA library. The
amino acid sequence of SpKRP95, deduced from the
nucleotide sequence of the full length clone, predicts
a kinesin-related polypeptide of 742 amino acids
(Figure 1(a)) with a molecular weight of 84 kDa. The
SpKRP95 sequence displays 51.6% identity to that
of SpKRP85, but SpKRP95 is 43 residues longer
than SpKRP85 (699 amino acids long, predicted
Mr = 79 kDa) with the extra residues located at the C
terminus of SpKRP95. Like SpKRP85, SpKRP95 has
a tripartite structure consisting of a globular
N-terminal motor domain (residues 1 to 356,
predicted pI = 9.9), a central stalk domain with a
high probability of a-helical structure (residues 357
to 588, predicted pI = 5.04), and a second globular
7 1995 Academic Press Limited
(a)
(b)
Figure 1. (a) Deduced amino acid sequence of full length SpKRP95. A l ZAP Strongylocentrotus purpuratus unfertilized
egg cDNA library was screened with a partial SpKRP95 cDNA fragment (Cole et al., 1993). The resulting 3.1 kb clone was
sequenced and found to contain an open reading frame encoding 742 amino acids. The X in position 668 is resolved as R.
This peptide sequence, and the corresponding nucleotide sequence, is available from GenBank under accession number
U00996. Library screening and sequencing were performed by standard methods (Wedaman et al., 1993). (b) Dot matrix
comparisons between SpKRP95 and closely related kinesin-related polypeptides. The full length SpKRP95 amino acid
sequence (vertical axes) was compared to the full length sequences of S. purpuratus SpKRP85, murine Kif3A, Drosophila
Klp68D, Chlamydomonas KHP1, Caenorhabditis elegans Osm-3 and S. purpuratus kinesin heavy chain (SpKHC) (horizontal
axes). The plots were generated from the Wisconsin GCG dotplot program, using a window size of 8 and stringency 7.
159
Communication
Table 1. Percent identities between SpKRP95 and related motor proteins
SpKRP95
Full
Motor
SpKRP95
SpKRP85
Kif3A
KHP1
Klp68D
Osm3
SpKRP85
Full
Motor
Kif3A
Full
Motor
KHP1
Full
Motor
Klp68D
Full
Motor
Osm3
Full
Motor
SpKHC
Full
Motor
51.6
50.7
72.5
47.8
48.2
49.5
47.2
41.6
41.9
38.8
41.0
42.1
39.4
41.3
39.1
33.8
33.6
35.1
35.8
30.6
33.2
72.8
70.1
80.5
68.5
66.1
65.2
60.0
57.2
57.2
54.4
60.7
58.8
55.5
59.5
52.0
46.0
44.9
44.7
43.4
44.3
43.2
The full length or motor region sequences of S. purpuratus KRP95, S. purpuratus KRP85, mouse Kif3A, Chlamydomonas KHP1, Drosophila
Klp68D, C. elegans Osm-3, and S. purpuratus conventional kinesin heavy chain are compared.
region at the C terminus (residues 589 to 742,
predicted pI = 9.62). The pI values of the corresponding domains in SpKRP85 and SpKRP95 are almost
identical, but the motor domains of SpKRP95 and
SpKRP85 are considerably more basic than the
kinesin motor domain (pI = 8).
The heterotrimeric kinesin, KRP85/95, was first
discovered and purified from sea urchin eggs by
Cole et al. (1992), but it is now apparent that its two
motor subunits, SpKRP85 and SpKRP95 (Cole et al.,
1993; this report) are members of a kinesin subfamily
referred to as the kinesin II or Kif3 subfamily (Bloom
& Endow, 1994; Goldstein, 1993). Close relatives
of SpKRP85 and SpKRP95 in other organisms (Figure 1(b); Table 1) include murine Kif3A (Aizawa et al.,
1992; Kondo et al., 1994), Drosophila Klp68D and
Klp64D (Pesavento et al., 1994), the Chlamydomonas
Fla10 gene product, KHP1 (Walther et al., 1994), and
Caenorhabditis elegans Osm-3 (Shakir et al., 1993;
Tabish et al., 1995). The motor domains of SpKRP85
and SpKRP95 display approximately 60% or greater
sequence identity to the motor domains of these close
relatives, compared to approximately 45% sequence
identity with the motor domain of sea urchin kinesin
heavy chain (SpKHC; Table 1). In addition there is
significant sequence identity outside the motor
domains of SpKRP85, SpKRP95, Kif3A, Klp68D and
KHP1 but not Osm-3 or SpKHC (Figure 1(b)). Thus
SpKRP95 is more closely related to mouse Kif3A,
fruitfly Klp68D and Chlamydomonas KHP1 than it is
to conventional kinesin heavy chain from the same
organism. The remarkably high sequence conservation between SpKRP85 and Kif3A suggests that
these two proteins are true homologs, and the
recently described Kif3B, whose sequence is not yet
available, may be a homolog of SpKRP95 (Yamazaki
et al., 1994; see also Pesavento et al., 1994).
The amino acid sequences of SpKRP85 and
SpKRP95 were analyzed in order to evaluate their
potential for heterodimerization. The algorithm of
Lupas et al. (1991) revealed a high probability of
coiled-coil formation extending approximately between residues 340 and 590 of SpKRP85 and
SpKRP95, corresponding to the stalk domain
(Figure 2(a)). Visual inspection and helical wheel
analysis of the sequence itself reveal obvious heptad
repeats indicative of coiled-coil formation between
residues 354 to 588 of SpKRP95 and residues 357
to 594 of SpKRP85, ending in a conserved helixbreaking FIP-like sequence. This region could
heterodimerize to form a coiled coil of approximately
235 residues. Assuming a residue repeat of 0.15 nm,
this predicts a 35 nm long rod separating the
head from the tail in the heterotrimeric KRP85,95
holoenzyme, which is considerably shorter than the
50 to 70 nm stalk of kinesin.
The aforementioned helical wheel analysis revealed that SpKRP85/SpKRP95 heterodimers would
display a hydrophobicity pattern that is typical of
coiled coil proteins. This pattern is consistent with
the formation of one set of hydrophobic bonds
between the A positions of SpKRP85 and SpKRP95,
and a second set of hydrophobic bonds between the
D positions of the two peptides, that would form a
hydrophobic core throughout virtually the entire
length of the stalk of an SpKRP85/SpKRP95
heterodimer (Figure 2(b)). Interestingly, however, the
glycine and proline-rich regions corresponding to
the first major discontinuity predicted by the Lupas
et al. algorithm in both SpKRP85 and SpKRP95
(asterisks in Figure 2(a)) contain stretches of charged
residues that are oppositely charged in corresponding positions of SpKRP85 and SpKRP95 (Figure 2(c)).
These regions would be expected to give rise to
attractive electrostatic forces within SpKRP85/
SpKRP95 heterodimers and repulsive electrostatic
forces within SpKRP85/SpKRP85 or SpKRP95/
SpKRP95 homodimers. Therefore these residues
may play a role in stabilizing heterodimers relative to
homodimers. In addition, unfavourable electrostatic
interactions between residues of like charge in the A
and D positions themselves (Figure 2(b)) should
preferentially destabilize homodimers.
To test the hypothesis that SpKRP85 and SpKRP95
are indeed able to form heterodimers in vitro, we
used the cell-free translation system that has been
used extensively to examine the heterodimerization
of the cellular oncoproteins, fos and jun (Ransone
et al., 1989). Thus, cDNAs encoding SpKRP85 and
SpKRP95 were transcribed and translated in a rabbit
reticulocyte lysate, and an antibody that recognized
only the SpKRP95 polypeptide was tested for its
ability to immunoprecipitate SpKRP85/SpKRP95
complexes (Figure 3). This was done by expressing a
fusion protein consisting of full length SpKRP95
fused to the T7 gene 10 leader peptide recognizable
by the T7 tag monoclonal antibody. A partial
SpKRP85 construct, missing one-third of the motor
domain, was expressed alone, lacking the gene 10
leader peptide. As expected, the T7 tag mAb was able
to immunoprecipitate SpKRP95 (Figure 3 lane 4)
but not SpKRP85 (Figure 3 lane 3) from lysates
(a)
(b)
(c)
Figure 2. Sequence analysis of SpKRP85 and SpKRP95. (a) Probability of a-helical coiled coil formation as predicted
by the method of Lupas et al. (1991). The horizontal axes indicate residue number, and vertical axes represent the
probability of coiled coil formation. The plots show a striking similarity in the predicted domain organization of SpKRP85
(bottom) and SpKRP95 (top). In particular, both peptides are predicted to form a similar sized coiled-coil, consistent with
them forming a heterodimeric coiled-coil. The abundance of turns and proline residues, together with the predicted lack
of extensive a-helical structure in the N-terminal heads and C-terminal tails, indicate globular ends on either side of the
stalk domains. Asterisks indicate the glycine/proline-rich charged region that may favour heterodimerization over
homodimerization. (b) Hydrophobicity of A and D positions in the heptad repeats of the SpKRP85 and SpKRP95 stalk
regions. Helical wheels were generated using residues 420 to 594 of SpKRP85 and residues 410 to 588 of SpKRP95
(downstream from the first discontinuity (a)). The helical wheels were adjusted for skip residues; there were two skips
in the SpKRP85 sequence and one skip in SpKRP95. The A and D positions show a typical pattern of hydrophobicity found
in coiled coil proteins; not every residue is hydrophobic, but there is ample opportunity for hydrophobic bonding between
positions A to A and D to D within SpKRP85/SpKRP95 heterodimers. (c) Amino acid region of opposing charge. This
region, 33 amino acids in length in both SpKRP85 and SpKRP95, lies within the stalk domains but is not predicted to
form an a-helical coiled coil. This segment contains proline residues in both subunits, and may impede homodimerization
and/or appropriately align the peptides for heterodimerization by ionic interactions between oppositely charged
side-chains.
Communication
Figure 3. Heterodimerization of SpKRP85 and SpKRP95
in vitro. This autoradiograph depicts 35S-labelled proteins
that were expressed in and immunoprecipitated from a
rabbit reticulocyte lysate cell-free translation system.
Lane 1 shows expression of a truncated SpKRP85 construct missing one-third of its motor domain, and lane 2
shows the expression of T7 tag-SpKRP95 fusion protein in
the reticulocyte lysate. Lane 3, control showing that the T7
tag mAb does not immunoprecipitate any radiolabelled
polypeptides from lysates expressing only SpKRP85
lacking the T7 tag, whereas lane 4 shows that the T7 tag
mAb does immunoprecipitate the T7 tag-SpKRP95 fusion
protein from lysates containing T7 tag-SpKRP95 fusion
protein but no SpKRP85. Finally lane 5 shows that the T7
tag mAb immunoprecipitates both SpKRP85 and T7
tag-SpKRP95 from lysates containing both these polypeptides. As the SpKRP85 polypeptide was not fused to the T7
tag epitope, it could only have been immunoprecipitated
by the T7 tag mAb as a result of its binding to SpKRP95
which was fused to the T7 tag. A truncated SpKRP85
construct was used in these experiments because SpKRP95
fragments tended to mask full length SpKRP85 on the
autoradiograms. Methods: The coding sequences of the
SpKRP85 and SpKRP95 cDNAs were amplified by PCR
using pfu polymerase (Stratagene) from the pBluescript
library and subcloned into either pGEM-3Z (Promega) or
PRSET (Invitrogen) vectors. The SpKRP95 construct was
subcloned into the PRSET A vector for expression as a
fusion protein containing the T7 gene 10 leader peptide at
its N terminus, which is recognised by the T7 tag antibody.
The SpKRP85 construct was subcloned into pGEM-3Z for
expression of a truncated SpKRP85 missing one-third of its
motor domain, and lacking the T7 gene 10 leader peptide.
In vitro transcription of the constructs was performed with
the Promega Riboprobe kit following the manufacturer’s
protocols. The RNA transcripts were translated in rabbit
reticulocyte lysate containing [35S]L-methionine (Promega).
For some experiments (not shown) the Promega TNT rabbit
reticulocyte lysate expression system was used instead.
Immunoprecipitations were carried out as previously
161
containing only SpKRP95 or SpKRP85, respectively.
However when SpKRP85 and SpKRP95 were
expressed then incubated together in a cell-free
lysate, SpKRP85 was immunoprecipitated together
with SpKRP95 by the T7 tag antibody (Figure 3
lane 5), suggesting that SpKRP85 must have bound
to SpKRP95 in the lysate. This experiment provides
direct evidence that SpKRP85 and SpKRP95 are
capable of heterodimerization.
Two types of control experiment were used to
assess the specificity of heterodimerization in these
cell-free lysates (data not shown). First we tested
the ability of T7 tag-SpKRP85 and T7 tag-SpKRP95
fusion proteins to form complexes with radiolabelled
sea urchin egg kinesin heavy chain (SpKHC) that
could be immunoprecipitated with the T7 tag mAb.
Visual inspection of autoradiograms and scintillation
counting of the immunoprecipitates revealed only
background levels of radioactivity in the precipitates,
suggesting that SpKRP85 and SpKRP95 did not form
heterodimers with SpKHC. Therefore SpKRP85
and SpKRP95 do not heterodimerize promiscuously
with all kinesin-related polypeptides. Secondly,
we investigated the possibility that SpKRP85 and
SpKRP95 could homodimerize by testing the ability
of unlabelled T7 tag-SpKRP85 or T7 tag-SpKRP95
fusion proteins to bind to radiolabelled SpKRP85 or
SpKRP95, forming radioactive complexes that could
be immunoprecipitated by the T7 tag mAb. Visual
inspection of autoradiograms and scintillation counting of the immunoprecipitates revealed an increase in
radioactivity above background consistent with
homodimerization of SpKRP85 but not SpKRP95.
However, the extent of SpKRP85 homodimerization
was estimated to be at least fivefold lower than
heterodimerization. These experiments are consistent with the idea that a small fraction of the 300 kDa
KRP85/95 holoenzyme might consist of the 115 kDa
subunit complexed with SpKRP85 homodimers,
but the majority consists of the 115 kDa subunit
complexed with SpKRP85/SpKRP95 heterodimers.
Based on these and previous studies (Cole et al., 1992,
1993), it seems reasonable to propose that each
KRP85/95 molecule consists of one 115 kDa subunit,
one SpKRP85 subunit and one SpKRP95 subunit, but
additional immunoelectron microscopy studies are
being initiated to further test this hypothesis.
Are the close relatives of SpKRP85 and SpKRP95
from other organisms also components of heterotrimeric complexes? Holoenzymes assembled from
described, with minor modifications (Ransone et al., 1989).
Briefly, constructs were expressed separately, then
incubated together at 18°C. The lysates were then
incubated with the T7 tag monoclonal antibody (Novogen)
on ice, and the T7 tag antibody-conjugated peptides were
precipitated using Pansorbin cells (Calbiochem), washed
in RIPA buffer (50 mM NaCl, 25 mM Tris-HCl (pH 7.5),
0.5% (w/v) NP-40, and 0.1% (w/v) sodium deoxycholate),
then resuspended in SDS-PAGE sample buffer. The
samples were subsequently electrophoresed on 7.5%
(w/v) SDS-PAGE. The acrylamide gels were Coomassie
stained, dried, and autoradiographed.
162
the close relatives of SpKRP85 and SpKRP95
described in Table 1 have not yet been purified from
native tissue so their subunit composition is
unknown. However, in a recent independent study
described in abstract form, mouse Kif3A and Kif3B
were shown to heterodimerize in vitro, and they
could be co-immunoprecipitated with a 115 kDa
polypeptide from brain tissue (Yamazaki et al., 1994).
In addition, Drosophila Klp64D (whose sequence is
not yet published) and Klp68D are reported to have
sequences that predict similar sized coiled coils,
consistent with the idea that they too may form
heterodimeric coiled coils (Pesavento et al., 1994).
These results are consistent with the idea that the
close relatives of SpKRP85 and SpKRP95 are all
components of heterotrimeric complexes that may
represent a new heterotrimeric subfamily of kinesins
(Cole & Scholey, 1995).
In summary, KRP85/95 is a heterotrimeric motor
protein composed of two heterodimerized motor
subunits, SpKRP85 and SpKRP95, and one 115 kDa
subunit. Sequence analysis is consistent with the
hypothesis that heterodimerization between SpKRP85 and SpKRP95 could occur by a-helical coiled
coil formation within their stalk domains, with a 33
amino acid residue segment rich in glycine, proline
and charged residues serving to appropriately
align SpKRP85 and SpKRP95 and stabilize heterodimers relative to homodimers. Additional structural studies will be aimed at testing this hypothesis
and analyzing the contribution of the 115 kDa
subunit to the structure of the KRP85/95 holoenzyme.
In a broader context, the ability of distinct
kinesin-related polypeptides to hetero-oligomerize
may represent a general mechanism by which cells
generate diversity in their complement of kinesin
motors. For example, hetero-oligomeric assemblies
of kinesin-related polypeptides may have properties
that differ from the corresponding homo-oligomers,
and consequently the combinatorial assembly of
different kinesin-related polypeptides into a variety
of oligomeric states could broaden functional
diversity within the kinesin family. Currently, many
kinesin-related polypeptides have been sequenced
(Bloom & Endow, 1994; Goldstein, 1993) but only
conventional kinesin, the heterotrimeric kinesin
KRP85/95, and the homotetrameric kinesin KRP130
have been purified and characterized as native
multimeric holoenzymes (Cole & Scholey, 1995). It
will now be interesting to determine if any of the
kinesin-related polypeptides analyzed only at the
primary structural level also form hetero-oligomers
in their natural host cell.
Acknowledgements
We thank members of the Scholey laboratory, Scott Chinn
and Dr Sharyn Endow for helpful discussion. This work
Communication
was supported by NIH grants no. GM46376 and
no. GM50718, and no. BE-46 from the American Cancer
Society.
References
Aizawa, H., Sekine, Y., Tekamura, R., Zhang, Z., Nangaku,
M. & Hirokawa, N. (1992). Kinesin family in murine
central nervous system. J. Cell Biol. 119, 1287–1296.
Bloom, G. S. & Endow, S. (1994). Kinesins. Protein Profile,
1, 1059–1116.
Cole, D. G. & Scholey, J. M. (1995). Structural variations
among the kinesins. Trends Cell Biol. 5, 259–262.
Cole, D. G., Cande, W. Z., Baskin, R. J., Skoufias, D. A.,
Hogan, C. J. & Scholey, J. M. (1992). Isolation of a sea
urchin egg kinesin-related protein using peptide
antibodies. J. Cell Sci. 101, 291–301.
Cole, D. G., Chinn, S. W., Wedaman, K. P., Hall, K., Vuong,
T. & Scholey, J. M. (1993). Novel heterotrimeric
kinesin-related protein purified from sea urchin eggs.
Nature, 366, 268–270.
de Cuevas, M., Tao, T. & Goldstein, L. S. (1992). Evidence
that the stalk of Drosophila kinesin heavy chain is an
alpha-helical coiled coil. J. Cell Biol. 116, 957–965.
Goldstein, L. S. B. (1993). With apologies to Sheherazade:
tails of 1001 kinesin motors. Annu. Rev. Genet. 27,
319–351.
Henson, J. H., Cole, D. G., Terasaki, M., Rashid, D. &
Scholey, J. M. (1995). Immunolocalization of the
heterotrimeric kinesin related protein KRP85,95 in the
mitotic apparatus of sea urchin embryos. Dev. Biol. In
the press.
Kondo, S., Sato-Yoshitake, R., Noda, Y., Aizawa, H., Nakata,
T., Matsuura, Y. & Hirokawa, N. (1994). Kif3A is a new
microtubule-based motor in the nerve axon. J. Cell Biol.
125, 1095–1107.
Lupas, A., Van Dyke, M. & Stock, J. (1991). Predicting
coiled coils from protein sequences. Science, 252,
1162–1164.
Pesavento, P. A., Stewart, R. J. & Goldstein, L. S. (1994).
Characterization of the Klp68D kinesin-like protein in
Drosophila: possible roles in axonal transport. J. Cell
Biol. 127, 1041–1048.
Ransone, L., Visvader, J., Sassone-Corsi, P. & Verma, I, M.
(1989). Fos-jun interaction: mutational analysis of the
leucine zipper domain of both proteins. Genes Dev. 3,
770–781.
Shakir, M., Fukushige, T., Yasuda, H., Miwa, J. & Siddiqui,
S. S. (1993). C. elegans osm-3 gene mediating osmotic
avoidance behaviour encodes a kinesin-like protein.
Neuroreport, 4, 891–894.
Tabish, M., Siddiqui, Z. K., Nishikawa, K. & Siddiqui, S. S.
(1995). Exclusive expression of C. elegans Osm-3
kinesin gene in chemosensory neurons open to the
external environment. J. Mol. Biol. 247, 377–389.
Walther, Z., Vashishtha, M. & Hall, J. L. (1994). The
Chlamydomonas Fla10 gene encodes a novel kinesinhomologous protein. J. Cell Biol. 126, 175–188.
Wedaman, K. P., Knight, A. E., Kendrick-Jones, J. & Scholey,
J. M. (1993). Sequences of sea urchin kinesin light
chain isoforms. J. Mol. Biol. 231, 155–158.
Yamazaki, H., Nakata, T., Okada, Y. & Hirokawa, N. (1994).
Kif3B forms a heterodimer with Kif3A and works as
a new microtubule-based anterograde motor of
membrane organelle transport. Mol. Biol. Cell, 5, 32a.
Edited by J. Karn
(Received 17 May 1995; accepted 28 June 1995)