A Novel Kinesin-Like Protein, KIF1Bβ3 Is Involved in the Movement of Lysosomes to the Cell Periphery in Non-Neuronal Cells


  • Masafumi Matsushita,

    1. Department of Biological Sciences, Graduate School of Science, Osaka University, Machikaneyama-cho 1–16, Toyonaka City, Osaka, Japan 560–0043
    Search for more papers by this author
  • Shingo Tanaka,

    1. Department of Biological Sciences, Graduate School of Science, Osaka University, Machikaneyama-cho 1–16, Toyonaka City, Osaka, Japan 560–0043
    Search for more papers by this author
  • Norihiro Nakamura,

    1. Department of Biological Sciences, Graduate School of Science, Osaka University, Machikaneyama-cho 1–16, Toyonaka City, Osaka, Japan 560–0043
    Search for more papers by this author
  • Hiroki Inoue,

    1. Department of Biological Sciences, Graduate School of Science, Osaka University, Machikaneyama-cho 1–16, Toyonaka City, Osaka, Japan 560–0043
    Search for more papers by this author
  • Hiroshi Kanazawa

    Corresponding author
    1. Department of Biological Sciences, Graduate School of Science, Osaka University, Machikaneyama-cho 1–16, Toyonaka City, Osaka, Japan 560–0043
    Search for more papers by this author

Hiroshi Kanazawa, kanazawa@bio.sci.osaka-u.ac.jp


The kinesin superfamily protein, KIF1Bβ, a splice variant of KIF1B, is involved in the transport of synaptic vesicles in neuronal cells, and is also expressed in various non-neuronal tissues. To elucidate the functions of KIF1Bβ in non-neuronal cells, we analyzed the intracellular localization of KIF1Bβ and characterized its isoform expression profile. In COS-7 cells, KIF1B colocalized with lysosomal markers and expression of a mutant form of KIF1Bβ, lacking the motor domain, impaired the intracellular distribution of lysosomes. A novel isoform of the kinesin-like protein, KIF1Bβ3, was identified in rat and simian kidney. It lacks the 5th exon of the KIF1Bβ-specific tail region. Overexpression of KIF1Bβ3 induced the translocation of lysosomes to the cell periphery. However, overexpression of KIF1Bβ3-Q98L, which harbors a pathogenic mutation associated with a familial neuropathy, Charcot-Marie-Tooth disease type 2 A, resulted in the abnormal perinuclear clustering of lysosomes. These results indicate that KIF1Bβ3 is involved in the translocation of lysosomes from perinuclear regions to the cell periphery.

Intracellular membrane vesicles and organelles including endosomes, mitochondria and lysosomes have their own distinct patterns of localization and motility. It is a well-accepted concept that the microtubule-dependent molecular motors, such as kinesin and dynein, regulate these motile processes. However, the molecular identification of motor proteins specifically involved in the transport of each membrane compartment has not been fully completed and the mechanisms of localization and motility of organelles remain unclear (1–5).

KIF1Bβ is a member of the kinesin superfamily proteins (KIFs), which are microtubule-dependent molecular motors that are involved in various intracellular organellar transport processes (6). Since kinesin-related proteins are categorized on the basis of their primary structures, KIF1Bβ is grouped in the KIF1 subfamily, which includes KIF1A, KIF1B, KIF1C and KIF1D (7). The KIF1 family proteins contain a highly conserved motor domain in their amino-terminal region, as well as a tail domain in their carboxyl-terminal region that is less well conserved and assumed to function as the cargo-binding region.

KIF1B is further divided into two major splicing isoforms, KIF1Bα and KIF1Bβ, produced by alternative splicing of the KIF1B gene, which contains at least 47 exons (8). KIF1Bα was the first reported KIF1B isoform, and is an 1150-amino acid protein with a predicted molecular weight of 130 kDa (9). KIF1Bβ consists of 1770 amino acids and has a predicted molecular weight of 200 kDa (10–12). The primary structure of 660 amino acids in the amino-terminal region is conserved between KIF1Bα and KIF1Bβ; however, their carboxyl-terminal sequences are completely different. Moreover, it has been reported that short amino acid sequences can be inserted within the conserved amino-terminal half of KIF1Bα and KIF1Bβ (11,13). Rat KIF1Bβ2 and mouse KIF1Bp204 have two such insertions of 6 and 40 amino acids in length, resulting in a 1816-amino acid protein sequence with a predicted molecular weight of 204 kDa (11,13).

KIF1Bα was originally reported to be the motor involved in anterograde transport of mitochondria in neuronal cells (9). Recent studies demonstrated that KIF1Bα bound its specific cargo via PSD-95 proteins and S-SCAM, and that it was involved in dendritic and axonal transport (14). KIF1Bβ associates with synaptic transport vesicles in neuronal cells (12,13). Heterozygous deletion of the mouse KIF1B gene causes a defect in transport of vesicles and these animals develop progressive muscle weakness (12). This phenotype is similar to the symptoms observed in the human familial neuropathy, Charcot-Marie-Tooth disease type 2A (CMT2A), and notably an amino acid substitution in the KIF1B gene has been found in CMT2A patients (12).

Although KIF1Bβ is not restricted to neuronal systems but is widely expressed in other tissues (10–13), the function of KIF1Bβ in non-neuronal cells is still unclear. To investigate this, we analyzed the intracellular localization of KIF1Bβ in non-neuronal cells by immunocytochemistry. In addition, we characterized a novel 190-kDa KIF1Bβ isoform (KIF1Bβ3) in non-neuronal cells that is involved in the translocation of lysosomes from perinuclear regions to the cell periphery.


Intracellular localization of KIF1B isoforms

KIF1B isoforms are expressed in various non-neuronal tissues and in simian kidney COS-7 cells as shown previously (12–14). We prepared the KIF1B-CBD antibody, which was raised against the conserved region of KIF1B isoforms, as described previously (13). Using this antibody, 190-kDa (β) and 130-kDa (α) protein isoforms were detected in various tissues including brain, lung and testis, as well as in COS-7 cells (Figure 1). To investigate the role of KIF1Bβ isoforms in non-neuronal cells, we first analyzed the intracellular distribution of KIF1B in COS-7 cells by indirect immunofluorescence microscopy with the KIF1B-CBD antibody. Punctate labeling was observed throughout the cell (Figure 2A,B) and this cytosolic staining was largely codistributed with that of lysosomal membrane glycoprotein-2 (LAMP-2) (15) (Figure 2A,B). The punctate labeling was dense in the perinuclear region and the signals within each labeled compartment were indistinguishable from each other. To further clarify the codistribution of KIF1B and LAMP-2, we examined their distribution in cells in which the microtubule cytoskeleton had been depolymerized with nocodazole. Although LAMP-2-labeled compartments were randomly scattered throughout the cytoplasm in nocodazole-treated cells, these compartments largely colocalized with KIF1B (Figure 2C,D). To further confirm the codistribution of KIF1B and the lysosomal compartment, we performed subcellular fractionation of COS-7 cells on sucrose density gradient centrifugation (Figure 3). The lysosomal enzyme acid phosphatase was highly enriched in fraction 7. KIF1Bβ was also significantly enriched in fraction 7. KIF1Bβ and acid phosphatase found in the top fraction (fraction 1) reflect the dissociation or release from the membrane compartments. Notably, KIF1Bβ was codistributed with the acid phosphatase, whereas KIF1Bα showed the significant peak at fraction 2. Other markers, EEA1 (early endosomes) and GM130 (Golgi apparatus), showed distinct distributions from the acid phosphatase. These data suggest that a significant portion of KIF1Bβ, rather than KIF1Bα, is localized in lysosomes.

Figure 1.

Figure 1.

Expression of KIF1B isoforms in tissues and the COS-7 cell line. Total protein from rat brain, lung, testis (12.5 μg) and COS-7 cells (5.0 μg) was resolved on SDS-PAGE, transferred to Polyvinylidene Fluoride (PVDF) membranes and probed with affinity-purified KIF1B-CBD antibody. The positions of α isoforms (130 kDa) and β isoforms (190–200 kDa) are shown.

Figure 2.

Figure 2.

Localization of KIF1B and LAMP-2 in COS-7 cells. The COS-7 cells were fixed, permeabilized and stained with rabbit KIF1B-CBD and mouse anti-LAMP-2 antibodies, followed by incubation with Alexa Fluor 488-conjugated anti-rabbit IgG and Alexa Fluor 546-conjugated anti-mouse IgG, respectively. Prior to fixation, the cells were treated with (C–D) or without (A–B) 0.1 mg/mL nocodazole in DMEM + 10% FCS for 1 h. Left column (yellow) shows overlays of LAMP-2 and KIF1B immunofluorescence. Middle column (green) and right column (red) show anti-KIF1B immunofluorescence and LAMP-2 immunofluorescence, respectively. White squares in B and D indicate the enlarged areas seen in A and C, respectively. Bars, 25 μm.

Figure 3.

Figure 3.

Subcellular fractionations of KIF1B.COS-7 cells were homogenized in homogenization buffer, and the postnuclear supernatant was centrifuged at 41 000 g for 30 min. The membrane pellet was homogenized and layered on a preformed 0.6–2.0 m (approximately 10–55%) sucrose gradient. After centrifugation at 110 000 g, fractions were collected. The density (dotted line) and acid phosphatase activity (gray bars) of each fraction is shown in the top panel. Immunoblot analyses of KIF1B, GM130 and EEA1 of each fraction are shown in the lower panels. The distributions of KIF1Bα (closed squares) and KIF1Bβ (open circles) were quantified and are shown in the top panel.

Overexpression of KIF1Bβ, which lacks the motor domain, affects the localization of late endocytotic organelles

In previous studies, it was shown that kinesin-related proteins lacking the motor domain were unable to transport cargoes and that the overexpression of dominant-negative forms of these proteins interfered with the normal distribution of their specific cargoes (16–18). If KIF1Bα or KIF1Bβ is involved in the translocation of endosomal/lysosomal compartments, the distribution of endosomal/lysosomal compartments would be specifically affected in cells overproducing the motorless KIF1B isoform. Thus, we expressed a mutated form of the KIF1B isoform, lacking its motor domain, and analyzed the intracellular localization of lysosomes. In Figure 4A, a schematic illustration of the expression constructs used in this study is provided. Expression vectors, encoding the carboxyl terminal halves of KIF1B isoforms fused with EGFP, were constructed and transiently expressed in COS-7 cells. Amino-terminal regions containing the 6- and 40-amino acid residue insertions in KIF1Bβ2 were not employed in these constructs. The integrity of expressed EGFP-fusion products was confirmed by immunoblotting with anti-GFP (Figure 4B). The EGFP-fusion proteins were observed throughout the cytosol (Figure 4C). Due to overexpression and the relatively weak interaction with cargo, excessive fusion proteins may be diffused in the cytosol. It should be noted that a similar diffused localization was reported when headless-kinesin II was overexpressed (18). The total intensity of LAMP-2 signal in each cell was not significantly affected in the overexpressing cells (Figure 4C, arrow b) compared to the control (Figure 4C, arrow a) cells. However, overexpression of the carboxyl-terminal half of KIF1Bβ (βCt) induced perinuclear accumulation of LAMP-2 (Figure 4C(b)), although its punctate staining throughout the cytoplasm was not affected significantly (Figures 4C(b),D). In contrast, expression of the carboxyl-terminal half of KIF1Bα (αCt) did not affect the distribution of LAMP-2 (Figure 4C,D). These results suggest that KIF1Bβ associates with late endosomal/lysosomal compartments and is involved in the distribution of such compartments. Further deletion of the last 220 amino acids from the carboxyl terminus (β-CtΔ1551) resulted in a motorless construct that failed to impair the normal distribution of LAMP-2 (Figure 4C,D). This result suggests that this 220-amino acid sequence in the carboxyl-terminal region of KIF1Bβ is important for lysosomal distribution.

Figure 4.

Figure 4.

Structure and expression of KIF1B deletions lacking motor domain.A)Structure of KIF1Bα, KIF1Bβ and KIF1B deletions. aa 595–1770 of KIF1Bβ (βCt), aa 595–1551 of KIF1Bβ (βCtΔ1551) and aa 660–1150 of KIF1Bα (αCt) were fused with EGFP in their amino terminal regions. B)Expression of KIF1B deletions. Total proteins from COS-7 cells expressing EGFP-fusion KIF1B deletions were resolved on SDS-PAGE, transferred to PVDF membranes and probed with anti-GFP. C)Localization of GFP-KIF1B deletion fusions and LAMP-2. COS-7 cells were transiently transfected with plasmids encoding GFP-fused deletions of KIF1B shown in A. At 24 h post-transfection, the localization of EGFP-fusion proteins and LAMP-2 was observed. GFP-fusion proteins were visualized by direct fluorescence (left column). Localization of LAMP-2 was visualized by indirect immunofluorescence using anti-LAMP-2 antibody and Alexa Fluor 546-conjugated secondary antibody (right column). Bars, 25 μm. D)Percentage of cells with perinuclear accumulation of LAMP-2 in cells expressing GFP-fusion proteins or EGFP alone as the control. Data shown are mean ± s.d. of four experiments.

Characterization of KIF1Bβ isoforms in non-neuronal cells

Since the KIF1Bβ isoform of non-neuronal cells and tissues (Figures 1 and 5B, arrow c) migrates more quickly in SDS-PAGE than that of brain (Figures 1 and 5B, arrows a, b) (12), we characterized this isoform in non-neuronal cells. Firstly, we tested whether in non-neuronal cells this isoform contains the short insertion found in KIF1Bβ2 (Figure 5A) by using the pep40 antibody which was raised against the 40-amino acid insertion sequence (Figure 5B). As described previously, KIF1Bβ2 was detected with anti-pep40 in rat brain extracts as a 200-kDa band (13). However, no signal of this size was detectable for COS-7 cell extracts. These results indicate that COS-7 cells express KIF1B isoforms that lack the 40-amino acid insertion.

Figure 5.

Figure 5.

Structures of KIF1Bβ isoforms and expression in COS-7 cells.(A)Schematic illustrations of KIF1B isoforms. α-tail and β-tail are indicated as hatched and crosshatched boxes, respectively. Motor domain and PH domain are indicated in gray and dark gray, respectively. The insertions found in KIF1Bβ2 are indicated as boxes with horizontal lines. The asterisk in KIF1Bβ2 represents the 40-amino acid insertion used as an immunogen for the pep40 antibody. The region amplified by RT-PCR is also indicated. (B)Immunoblot analysis with KIF1B-CBD and KIF1B-pep40 antibodies. Total proteins from COS-7 cells and rat brain were resolved by SDS-PAGE, transferred to PVDF membranes and probed with KIF1B-CBD or pep40 antibodies. (C)Comparison of isoforms in COS-7 with known KIF1Bβ isoforms. Total proteins from COS-7 cells with or without the transfection of KIF1Bβ isoform expression vectors were resolved by SDS-PAGE, transferred to PVDF membranes and probed with KIF1B-CBD antibody. Transfected cell samples were loaded to a 5% concentration relative to nontransfected cells.

To test whether the 190-kDa isoform expressed in COS-7 cells was identical to known KIF1Bβ isoforms, we expressed rat KIF1Bβ/β2 in COS-7 cells and compared the migrations of these known isoforms and the 190-kDa isoform on SDS-PAGE (Figure 5C). Expressed KIF1Bβ2 showed a similar migration pattern to the 200-kDa rat brain KIF1Bβ2 isoform (Figure 5C) (13). Expressed KIF1Bβ also migrated more slowly than the 190-kDa isoform in COS-7 cells (Figure 5C). These discordances of migration between KIF1Bβ/β2 and the 190-kDa isoform in COS-7 cells suggested that another KIF1Bβ isoform, corresponding to 190-kDa isoform, would be expressed in COS-7 cells.

To investigate the presence of this putative KIF1Bβ isoform in rat tissue, we performed reverse transcribed-polymerase chain reaction (RT-PCR) using KIF1Bβ-specific primers (Figure 6A). The RT-PCR product from rat brain was approximately 4.3 kbp, consistent with the expected length for rat KIF1Bβ2 (Figure 6A). However, the RT-PCR products from other tissues (Figure 6A), as well as from COS-7 cells (data not shown) were 4.0 kbp, which is smaller than the expected length for KIF1Bβ(4.2 kbp) or KIF1Bβ2 (4.3 kbp). To analyze these differences in transcript lengths, we sequenced the RT-PCR products from rat kidney and COS-7 cells. This indicated that the shorter product lacked 249 bp of the β-specific tail of KIF1Bβ (Figure 6B). The deleted regions (aa 846–928 of KIF1Bβ) were the same for the rat kidney and COS-7 cells (Figure 6C). To analyze whether this deletion occurred by alternative splicing, we surveyed the rat genomic sequence using the Rat Genome Browser program at UC Santa Cruz (19). While the entire tail region of KIF1Bα was encoded in a single exon (Figure 7A), the tail region of KIF1Bβ was divided into 27 exons. This genomic structure including arrangement of exon and introns of rat KIF1B gene was the same as that of human KIF1B gene (8). The deleted 249-bp sequence matched the 5th exon of the β-tail (Figure 7B). Thus it was concluded that the 4th exon was joined directly to the 6th exon in the 249 bp-deleted form of KIF1Bβ. This result suggests that the 249 bp-deleted KIF1Bβ is an alternatively spliced form of KIF1B, which lacks the 5th exon in the β-tail. The calculated molecular weight of KIF1Bβ with this 249-bp deletion is 190 kDa. In COS-7 cells, expressed 249 bp-deleted KIF1Bβ migrated to the same position as the 190-kDa protein (Figure 8). Taken together, we concluded that the 249 bp-deleted KIF1Bβ, which comprises 1687 amino acids, is expressed in non-neuronal cells, including COS-7 cells. Accordingly, we designated this molecule as KIF1Bβ3 to distinguish it from other KIF1B isoforms.

Figure 6.

Figure 6.

Identification of a new KIF1Bβ isoform in non-neuronal cells.(A)Differently sized RT-PCR products from rat tissues. RT-PCR was performed using the primers designed for the amplification of 4.3 kbp (KIF1Bβ2) or 4.2 kbp (KIF1Bβ) sequences in the carboxyl terminal (shown in B). (B)Schematic illustration of a novel identified molecule. The position of the 249-bp (83 aa) deletion is shown as a bar in the middle of the molecule (aa 846–928 of KIF1Bβ). The sequence data are available from GenBank/EMBL/DDBJ under accession number AB120429. (C)Alignment of nucleic acid sequences around the 249-bp deletion found in novel molecules identified from rat kidney and COS-7 cells.

Figure 7.

Figure 7.

Structure of KIF1B gene in rat genome.(A)Map of rat KIF1B gene on chromosome 5. The rat genomic sequence and map were obtained from The Human Genome Browser at UC Santa Cruz (http://www.genome.ucsc.edu/). Vertical lines represent exons. Only the coding regions are shown. The arrowed circle indicates the 5th exon of the β-tail, which is matched to the 249-bp deletion. (B)Rat genomic sequence of the 4th, 5th and 6th exon of β-tail. The exon sequences (bold and capitalized) were located between ag and gt sequences (bold and underlined).

Figure 8.

Figure 8.

Comparison between the newly identified isoform and isoforms in COS-7 cells.(A)Schematic illustration of a novel identified molecule KIF1Bβ3. The position of the 249-bp (83 aa) deletion is shown as a bar in the middle of the molecule (aa 846–928 of KIF1Bβ). (B) Comparison between KIF1Bβ isoforms and isoforms in COS-7 by migration on SDS-PAGE. Total proteins from COS-7 cells with or without the transfection of KIF1Bβ isoform expression vectors were resolved on SDS-PAGE, transferred to PVDF membrane and probed with KIF1B-CBD antibody. Transfected cell samples were loaded to a 5% concentration relative to nontransfected cells.

Overexpression of KIF1Bβ3 caused the translocation of lysosomes from the perinuclear region to the cell periphery

We overexpressed KIF1Bβ3 or the KIF1Bβ3-EGFP fusion protein in COS-7 cells and analyzed their effects on lysosomal distribution. In contrast to normal cells (Figure 2B), in cells overexpressing KIF1Bβ3, this protein was localized throughout the cell but was significantly accumulated in certain areas, notably in the cell periphery (Figure 9A,B). Although perinuclear staining of the lysosomal marker LAMP-2 was still observed, localization of LAMP-2 to the cell periphery was increased significantly by KIF1Bβ3 overexpression. Notably, LAMP-2 and KIF1Bβ3 were localized to the same short protrusions (indicated with arrows in Figure 9A,B), which were located in the edges of microtubule networks (Figure 9C). The overexpression of KIF1Bβ3 did not affect the organization of microtubules significantly (Figure 9C). Next, we analyzed the localization of fluorescent dextran, as an internalized endocytic tracer in KIF1Bβ3-overexpressing cells. Following internalization of fluorescent dextran for 24 h, excess dextran in the medium was washed out and the expression vectors for KIF1Bβ3 were transfected. In cells overexpressing KIF1Bβ3-EGFP, dextran-labeled lysosomal compartments were predominantly localized to the cell periphery (Figure 9D), whereas they exhibited perinuclear localization, similar to LAMP-2 (Figure 2B) in nonoverexpressed cells (data not shown). To examine translocation of other endocytotic compartments by KIF1Bβ3, we analyzed the localization of membrane markers for early endosomes (EEA1), trans-Golgi networks to late endosomes or plasma membranes (cation-independent mannose-6-phosphate receptor) in cells overexpressing KIF1Bβ3. None of these markers accumulated in the cell periphery or displayed an abnormal distribution in response to KIF1Bβ3 overexpression (Figure 9E,F).

Figure 9.

Figure 9.

Effects on localizations of endocytotic compartments by over-expression of KIF1Bβ3. COS-7 cells grown on coverslips were transiently transfected with an expression vector encoding KIF1Bβ3 (A)or KIF1Bβ3-EGFP (B–F). Effects on local-ization of LAMP-2 (A-B), tubulin (C), internalized fluorescent dextran (D), EEA1 (E)and cation-independent mannose-6-phosphate receptor (CI-MPR) (F), in cells overexpressing KIF1Bβ3 are shown. Localization of KIF1Bβ3 was visualized by direct fluorescence of EGFP (B–F, left) or indirect immunofluorescence using KIF1B-CBD antibody and Alexa Fluor 488-conjugated secondary antibody (A, left). Localization of LAMP-2 (A,B), tubulin (C), EEA1 (E) and CI-MPR (F) was visualized by indirect immunofluorescence using its primary antibody and Alexa Fluor 546-conjugated secondary antibody (middle). Internalized fluorescent dextran was visualized by direct fluorescence (C, middle). Right column is the overlay of left and middle columns. Bars, 25 μm.

These results suggest that KIF1Bβ3 is not involved in the localization of early endocytotic compartments, but is predominantly associated with the localization of lysosomes. Moreover, considering that KIF1B is oriented towards the plus-end of microtubules (9), these observations suggest that KIF1Bβ3 is involved in the translocation of late endosomal/lysosomal compartments to the cell periphery.

Overexpression of KIF1Bβ3-Q98L, bearing a mutation found in human Charcot-Marie-Tooth disease, induces accumulation of lysosomes in the perinuclear region

A point mutation, Q98 to L, of the KIF1B gene is associated with Charcot-Marie-Tooth disease type 2A (12). This amino acid substitution results in the functional loss of KIF1Bβ and the overexpressed mutant KIF1Bβ accumulates in the perinuclear region (12). To determine the effect of the Q98L mutation on lysosomal distribution, we expressed KIF1Bβ3-Q98L in COS-7 cells. At 24–32 h post-transfection, KIF1Bβ3-Q98L was accumulated in the perinuclear region (Figure 10). This localization was similar to the previous observation with KIF1Bβ-Q98L. In cells overexpressing KIF1Bβ3-Q98L, accumulated staining of the lysosomal marker LAMP-2 was prominent in the perinuclear region (Figure 10). However, the total intensity of LAMP-2 signal in each cell was not significantly affected. These observations suggest that the Q98L substitution induces a loss of KIF1Bβ3 motor activity and that the localization of lysosomes is impaired by the dominant negative effect of KIF1Bβ3-Q98L.

Figure 10.

Figure 10.

The effect of CMT2A mutation (Q98L) in KIF1Bβ3 on lysosomal distribution. COS-7 cells grown on coverslips were transiently transfected with an expression vector encoding KIF1Bβ3-Q98L-EGFP. KIF1Bβ3-Q98L-EGFP was visualized by direct fluorescence of EGFP (upper left). Localization of LAMP-2 was visualized by indirect immunofluorescence using anti-LAMP-2 and Alexa Fluor 546-conjugated secondary antibody (upper right). Merged image: KIF1Bβ3-Q98L-EGFP and LAMP-2 (bottom left), with differential interference contrast (DIC) image (bottom right). Bars, 25 μm.


In this study, we identified KIF1Bβ3, a new isoform of KIF1B, which is expressed in non-neuronal tissues and cells. Analysis of the transcript and genomic sequences revealed that KIF1Bβ3 is an alternatively spliced form of KIF1Bβ, which lacks a single exon in the β-specific tail. Subcellular fractionation showed the codistribution of KIF1Bβ3 and lysosomal compartment. Dominant negative experiments using headless-constructs revealed that α isoforms but not β isoforms were involved in the lysosomal distribution. Overexpression of KIF1Bβ3 in non-neuronal COS-7 cells induced a marked translocation of lysosomes to the cell periphery, especially to protrusions. Thus, we concluded that KIF1Bβ3 mediates lysosomal translocation in non-neuronal cells.

Generally, lysosomes are concentrated in the perinuclear region, but move dynamically between the perinuclear region and the cell periphery (20–22), which is consistent with our observations in COS-7 cells. A recent study showed that lysosomes fuse with the plasma membrane in fibroblasts (23). Lysosomes and lysosome-related organelles are also involved in the secretion of various biomolecules, such as histamines, and involved in the transfer of melanin from melanocytes to keratinocytes, as well as in antigen presentation in immune cells (24). KIF1Bβ3 might be involved in these lysosomal motile processes.

The kinesin-like motor proteins KIF2β and KIF5B are also involved in the outward transport of lysosomes (25–27). Overexpression of an ATPase-deficient rigor mutant KIF5B, or targeted disruption of KIF5B, impairs the dispersion of lysosomes (25,27). The functions of KIF1Bβ3 and KIF5B might partially overlap. However, lysosomes become abnormally large and are peripherally located at some distance from their usual perinuclear positions in cells overexpressing KIF2β (26). Thus, KIF1Bβ3 and KIF2β might contribute to lysosomal movement in different ways. Further analysis of KIF1Bβ3 function may therefore provide new insights into lysosomal movement.

Previous studies revealed that KIF1Bβ isoforms have two insertion sites in the amino-terminal regions (10,11). KIF1Bβ (1770 aa, 199 kDa) and KIF1Bβ2 (1816 aa, 204 kDa), both expressed in neuronal cells, associate with synaptic transport vesicles, although they are different in terms of the insertions (12,13). Identification of KIF1Bβ3 revealed another variable splicing site in the β-specific tail. Since KIF1Bβ3 does not have the 6- and 40-amino acid insertions in the amino-terminal region, and also lacks an 83-amino acid sequence from the 5th exon of the β-specific tail, it is a 1687-amino acid protein with a molecular weight of 190 kDa (Figures 6 and 8A). Transiently expressed KIF1Bβ3 showed the same migration pattern on SDS-PAGE as the endogenous 190-kDa isoform in COS-7 cells (Figure 8). The same-sized 190-kDa isoform was detected by immunoblotting in various cell lines − NRK (rat kidney), CHO (Chinese hamster ovary), HEp2 (human larynx), HeLa (human cervix) and NIH3T3 (mouse fibroblast). Therefore, KIF1Bβ3 is considered to be expressed widely in non-neuronal cells (data not shown). However, the 190-kDa isoform (Figure 5B (b)) was also detected in rat brain but migrated slightly more slowly than non-neuronal cells (Figures 1 and 5B). Although it is possible that this band is another isoform, the identification of this molecule was outside the scope of this study. The results shown in Figure 4 suggest that KIFBβ/β2 could potentially cause the translocation of lysosomes when exogenously expressed in non-neuronal cells. In fact, overexpressed KIFBβ/β2 also caused the translocation of lysosomes to the cell periphery (data not shown). While these results suggested that the 83-amino acid sequence from the 5th exon of the β-specific tail is not involved in lysosomal transport, at present the function of this region is not clear.

UNC104, a member of the KIF1/UNC104 subfamily, can transport phosphatidylinositol(4,5)bisphosphate (PtdIns (4,5)P2)-containing liposomes through its PH (pleckstrin homology) domain (28). KIF1Bβ/β2/β3 have a PH domain-like sequence in their carboxyl terminal region and a KIF1Bβ-tail lacking this PH domain is unable to interact with synaptic vesicles (12). Our result suggests that this region might also be important for lysosomal distribution (Figure 4). Accordingly, KIF1Bβ3 could interact with lysosome-associated phosphatidylinositol through its PH domain. However, KIF1A, which also has a PH domain in its carboxyl end, associates with lipin-α and cargo molecules via its middle (aa 657–1105) rather than carboxy region (29). This suggests that both the PH domain and lipin-α-binding regions are required for the cargo binding of KIF1A (29). KIF1Bβ3 might have similar mechanisms. Thus, the molecular mechanisms and basis of the interaction between KIF1Bβ3 and lysosomes requires further investigation.

Charcot-Marie-Tooth disease (CMT) is the most common inherited neuropathy. Recent studies demonstrated that mutations in Rab7, which is expressed ubiquitously and involved in the motility of late endosomal/lysosomal compartments, causes Charcot-Marie-Tooth disease type 2B (30,31). The Q98L mutation of KIF1B causes Charcot-Marie-Tooth disease type 2A (CMT2A) (12). Our results indicated that KIF1Bβ3-Q98L impaired normal lysosomal distribution (Figure 10), suggesting that this mutation affects lysosome translocation, thereby contributing to its pathogenicity. However, further analysis of KIF1Bβ3 and related proteins will be required to understand the molecular mechanisms of lysosomal dynamics. Ultimately, such studies could provide new insights into the pathophysiology and treatment of Charcot-Marie-Tooth disease.

Materials and Methods

Cell culture and transfection

Simian (Cercopithecus aethiops) kidney COS-7 cells were maintained with DMEM containing 10% fetal calf serum at 37 °C in 5% CO2. For transfection, cells were incubated in medium containing the FuGENE6 (Roche Diagnostics, Basel, Switzerland) DNA complex according to the manufacturer's instructions. After 24 h, cells were used for immunofluorescence or immunoblot analysis.


The monoclonal antibodies for tubulin (clone DM1B), EEA1 (clone 14) and cation independent mannose-6-phosphate receptor (MA1-066) were purchased from Sigma-Aldrich (St. Louis, MO), BD Pharmingen (San Diego, CA) and Affinity Bioreagents (Golden, CO), respectively. The monoclonal anti LAMP-2 antibody (Clone H4B4) was obtained from the Developmental Studies Hybridoma Bank of The University of Iowa. Alexa Fluor 488- and Alexa Fluor 546-conjugated secondary antibodies and rabbit polyclonal antibody for GFP were purchased from Molecular Probes (Eugene, OR). Affinity-purified anti-KIF1B-CBD raised against the CHP binding domain (KIF1B-CBD, aa 301–554) and KIF1B-pep40 antibody against the peptide corresponding to the 40-amino acid insertion (aa 396–435; DDYSGSGGKYLKDF) of KIF1Bβ2 were as described previously (13).

Subcellular fractionation and immunoblotting

COS-7 cells were homogenized with a tight fitting Dounce homogenizer in homogenization buffer (0.3 m sucrose, 1 mm EDTA, Complete protease inhibitors cocktail (Roche Diagnostics, Basel, Switzerland) 10 mm HEPES pH 7.5). A postnuclear supernatant was prepared by centrifugation at 500 g for 8 min. The resulting supernatant was centrifuged for 30 min at 41 000 g. The membrane pellet was then suspended in the homogenization buffer and layered on a preformed 0.6–2.0 m sucrose gradient containing the homogenization buffer. After centrifugation at 110 000 g for 3 h, the fractionated proteins were precipitated with 5% trichloroacetic acid and then dissolved in SDS-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. The samples were resolved by SDS-PAGE and transferred to PDVF membranes (32). Then the membranes were incubated with the indicated primary antibodies and the secondary horseradish peroxidase-conjugated antibodies against rabbit or mouse IgG (Jackson Immuno Research Laboratories, West Grove, PA). The signals were detected with ECL Western Blotting Detection Reagent and Hyperfilm-ECL (Amersham biosciences, Piscataway, NJ), as described previously (32). The quantification was performed by ImageJ software (National Institutes of Health), with a series of the films with nonsaturated exposures.

Measurement of acid phosphatase activity

Aliquots of fractions (40 μL) were incubated in 200 μL of reaction buffer (8 mm p-nitrophenyl phosphate, 90 mm sodium acetate pH 5.0) for 30 min. The reaction was stopped by addition of 0.6 mL of 0.25 m NaOH and absorbance of released p-nitrophenol at 405 nm was measured.

cDNA cloning and genomic analysis

RT-PCR was performed with using High Fidelity RNA PCR Kit (TaKaRa Biomedicals, Otsu, Japan) from total RNA, following the manufacturer's instructions. The primer sets (CGGAAATGGAAGTCT and GCTCGGGCATCTGCG, for COS-7) and (GTGACCCGACTGAAGGACCT and TACTaagcttTTAGTATTTTGGCTG, for rat tissues) were used. The obtained products were purified with Wizard SV Gel and PCR Clean-up System (Promega, Madison WI) and then directly sequenced or cloned into the pGEM-T vector (Promega). The rat genomic sequence (based on January 2003 assembly) was obtained and analyzed with The Human Genome Browser at UC Santa Cruz (http://www.genome.ucsc.edu/) (19).

Expression vectors of KIF1Bβs

For the construction of KIF1Bβ2 expression vector pcDNA-KIF1Bβ2, a HindIII site was introduced next to the stop codon by PCR and the entire coding region of KIF1Bβ2 was cloned into pcDNA3.1mycHisA(–) (Invitrogen, Carlsbad, CA) with NotI (located in the 4–11 bp upstream of the first ATG in cDNA) and HindIII. KIF1Bβ expression vector pcDNA-KIF1Bβ was generated by site-specific deletion of the insertion sequences using PCR with the overlapped primers (GCCTTGGCAGAGGTGAGTAAAAAGAAGAAG and CTTCTTCTTTTTACTCACCTCTGCCAAGGC for 18 bp deletion, GGCTGGGGGACATTATCGATGCATCCATGGGGTCCCTCAC and GTGAGGGACCCCATGGATGCATCGATAATGTCCCCCAGCC for 120 bp deletion) (33). For KIF1Bβ3 expression vector pcDNA-KIF1Bβ3, the 5th exon of β-tail was deleted by site-directed mutagenesis using pcDNA-KIF1Bβ as a template and the primers (CAAACTTGTGGGAAGGGCATTTGTGTACTT and AAGTACACAAATGCCCTTCCCACAAGTTTG). For EGFP-fusion KIF1Bβ3 expression vector KIF1Bβ3-EGFP, a HindIII site was introduced just prior to the stop codon and EGFP was introduced into this site. Thus the resulting vector expressed the entire KIF1Bβ3 with the linker Lys-Leu (AAGCTT, HindIII site) followed by EGFP. pcDNA-KIF1Bβ3-Q98L-EGFP vector was also generated by site-directed mutagenesis using pcDNA-KIF1Bβ3-EGFP as a template and the primers (TTTGCCTATGGGctgACCGGTGCTGGG and CCCAGCACCGGTcagCCCATAGGCAAA, cag(Q) was substituted to ctg(L) at the underlined position).

Deletion constructs of KIF1Bβ
For pEGFP-βCt and pEGFP-βCtΔC, the fragments digested with XhoI and HindIII (EGFP-βCt) or EcoRI (EGFP-βCtΔC) from pcDNA-KIF1Bβ were integrated into pEGFP-C1 (CLONTECH, Palo Alto, CA) at the same sites. For pEGFP-αCt, carboxyl terminal of rat KIF1Bα aa 660–1150) was amplified from a rat brain cDNA library by PCR with the primers (tagCTCGAGtaGCGGATTCTGATAGCGGGGA and TACTaagcttCTAGACTGTGGTTTCTCGA, the positions represent XhoI and HindIII sites, respectively) and integrated into pEGFP-C1 with XhoI and HindIII.

Immunofluorescence microscopy

For observation, cells seeded on coverslips were fixed with phosphate buffered saline containing 2% formaldehyde for 25 min and then residual formaldehyde was quenched using 0.1 m glycine in phosphate buffered saline. After incubation in phosphate buffered saline containing 0.4% saponin, 1% bovine serum albumin and 2% goat serum (34), samples were incubated with the primary antibodies for 1 h, followed by incubation with the secondary antibodies for 1 h. The samples were filled with 50% glycerol/phosphate buffered saline, and then observed under a BX-51 microscope (Olympus, Tokyo, Japan) equipped with a ×40 UPlanApo objective (numerical aperture 0 .85). Digital images were acquired and processed with an ORCA-ER1394 digital camera and AQUACOSMOS software (Hamamatsu Photonics, Hamamatsu, Japan).

Fluorescent-tracer internalization

COS-7 cells were incubated with tetramethylrhodamine-isothiocyanate dextran (T-1162, Sigma-Aldrich) at 0.5 mg/mL for the indicated times at 37 °C, and then transfected with KIF1Bβ3-EGFP expression vector. At 24 h post-transfection, cells were observed microscopically.

Gene manipulation and other procedures

Subcloning, ligation and sequencing were performed by following previously published procedures (35).


Cell culture reagents were obtained from Sigma-Aldrich and Invitrogen. Other chemicals were purchased from Sigma, unless otherwise specified.


This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan.