The chromokinesin Kid (KIF22) forms a homodimer, moves processively along microtubules, and transports double-stranded DNA

  1. Shinsuke Niwa  Is a corresponding author
  2. Natsuki Furusaki
  3. Tomoki Kita
  4. Yuki Suzuki
  5. Kyoko Chiba  Is a corresponding author
  1. Frontier Research Institute for Interdisciplinary Sciences (FRIS), Tohoku University, Japan
  2. Graduate School of Life Sciences, Tohoku University, Japan
  3. Mie University, Graduate School of Engineering, Japan
8 figures, 1 table and 4 additional files

Figures

Figure 1 with 3 supplements
Kid is a processive motor.

(A) Schematic illustration of the domain organization in Xenopus Kid tagged with a fluorescent protein mScarlet (XKidFL) and human Kid tagged with mNeonGreen (hKidFL). The calculated molecular weights of the fusion proteins are indicated on the right. (B) Representative SDS-PAGE analysis of purified XKidFL and hKidFL fusion proteins. The proteins are visualized using a Stain-Free gel. The molecular weight standards are indicated on the left side of the SDS-PAGE images. (C, D) Representative kymographs showing the motility of XKidFL at 20 pM (C) and hKidFL at 20 pM (D) both in the presence of 2 mM ATP. Scale bars: horizontal, 10 µm; vertical, 60 seconds. (E) Dot plots showing the velocity of XKidFL and hKidFL. Each dot shows a single datum point. Green bars represent mean ± S.D. n=51 and 52, respectively. (F) Dot plots showing the run length of XKidFL and hKidFL. Each dot shows a single datum point. Green bars represent mean ± S.D. n=51 and 52, respectively.

Figure 1—figure supplement 1
Original SDS-PAGE data for Figures 1B, 3B and 5C.

Purified proteins were separated by electrophoresis and visualized by Stain-Free gel (Bio-Rad). Numbers indicate molecular weight of the size marker.

Figure 1—video 1
The motility of XKid protein.

The motility of XKid was analyzed through time-lapse imaging recorded at 0.5 frames per second (fps) and played back at 25 fps. The dimensions of the displayed frame are 30 µm by 7 µm.

Figure 1—video 2
The motility of hKid protein.

The motility of hKid was analyzed through time-lapse imaging recorded at 0.5 fps and played back at 25 fps. The dimensions of the displayed frame are 30 µm by 7 µm.

Figure 2 with 1 supplement
Kid forms a weak dimer.

(A) Size-exclusion chromatography profiles of hKidFL (black) and UNC-104(1-653)-sfGFP (cyan). Below the chromatography, an SDS-PAGE image shows the elution fractions. Asterisks indicate fractions used for mass photometry and single molecule assays. The molecular weight standards are indicated on the left side of the SDS-PAGE images. (B) Size-exclusion chromatography of XKidFL (black) and UNC-104(1-653) (cyan). The SDS-PAGE of the elution fractions is shown beneath the profiles. Asterisks indicate fractions used for mass photometry and single molecule assays. The number shown at the left side indicates molecular weight standard. (C) Mass photometry analysis of human Kid at 10 nM. Histograms show particle counts, and lines indicate Gaussian fits. The mean ± SD and percentage of each peak are shown. (D) Mass photometry analysis of human Kid at 10 nM. Histograms show particle counts, and lines indicate Gaussian fits. The mean ± SD and percentage of each peak are shown. Note that the majority of hKid and XKid are dimers in the size-exclusion chromatography, but they are mostly dissociated to monomers in mass photometry.

Figure 2—source data 1

Original file for SDS-PAGE data for Figure 2A.

https://cdn.elifesciences.org/articles/102828/elife-102828-fig2-data1-v1.zip
Figure 2—source data 2

Original file for SDS-PAGE data for Figure 2B.

https://cdn.elifesciences.org/articles/102828/elife-102828-fig2-data2-v1.zip
Figure 2—source data 3

Original file for SDS-PAGE data for Figure 2A indicating the relevant bands and elution volume of each fraction.

https://cdn.elifesciences.org/articles/102828/elife-102828-fig2-data3-v1.zip
Figure 2—source data 4

Original file for SDS-PAGE data for Figure 2B indicating the relevant bands and elution volume of each fraction.

https://cdn.elifesciences.org/articles/102828/elife-102828-fig2-data4-v1.zip
Figure 2—figure supplement 1
Original SDS-PAGE data for Figure 2A and B.

Purified proteins were separated by electrophoresis and visualized by Stain-Free gel (Bio-Rad). Numbers indicate molecular weight of the size marker.

Conserved coiled-coil domain is required for the processive motion.

(A) Schematic representation illustrating the domain organization of XKid(1-496) and XKid(1-437). The calculated molecular weights of the fusion proteins are indicated on the right. (B) Representative SDS-PAGE analysis of purified XKid(1-496) and XKid(1-437) proteins. The proteins are visualized using a Stain-Free gel. The molecular weight standards are indicated on the right side of the SDS-PAGE images. (C, D) Mass photometry analysis of XKid(1-496)-mSca and XKid(1-437)-mSca. The expected molecular masses are 86 and 79 kDa, respectively. Histograms show particle counts, and lines indicate Gaussian fits. The mean ± SD and percentage of total counts for each peak are shown. (E, F) Representative kymographs showing the motility of 10 pM XKid(1-496) (E) and XKid(1-437) (F) in the presence of 2 mM ATP. Note that no directional movement was detected in XKid(1-437). Scale bars: horizontal 10 µm; vertical 10 seconds. (G, H) Mean-square displacement (MSD) analysis of XKid(1-496) and XKid(1-437) trajectories. (G) Representative MSD curves fitted to the power-law relationship MSD = AΔtα, where α is the anomalous diffusion exponent. XKid(1-496) showed superlinear MSD scaling with α=1.68, consistent with persistent or directionally biased motion, whereas XKid(1-437) showed sublinear MSD scaling with α=0.77. (H) Distribution of α values obtained from individual trajectory fits. Each dot represents one trajectory; bars indicate mean ± SD. n=10 trajectories per construct. (I, J) Schematic drawing of XKidCC-mScarlet (I) and a representative result of size-exclusion chromatography (J). XKidCC-mScarlet (magenta) and mScarlet (cyan) are shown.

Human Kid-mStayGold and Kid-mScarlet3 co-migrate on microtubules.

Purified hKid-mStayGold and hKid-mScarlet3 were mixed at final concentrations of 20 pM each and analyzed by single-molecule motility assays using TIRF microscopy. Representative kymographs show hKid-mStayGold, hKid-mScarlet3, and merged signals. The schematic drawing illustrates examples of co-migrating particles on microtubules, defined as overlapping mStayGold and mScarlet3 signals moving together along the same microtubule. Co-migration was defined as overlapping mStayGold and mScarlet3 signals moving together along the same microtubule over the same time interval. Scale bars: horizontal, 10 µm; vertical, 100 seconds.

Figure 5 with 3 supplements
Untypical neck linker of Kid can support processive movement of KIF1A.

(A) Schematic representation illustrating the domain organization of KIF1A(1-393)LZ, XKid(1-496), KIF1A(1-350), and a chimera protein KIF1AMD-XKidSt. Note that KIF1A(1-393)LZ and XKid(1-496) are processive motors, and KIF1A(1-350) is a nonprocessive motor. Cyan, motor domain of KIF1A; orange, motor domain of Kid; magenta, neck linker. (B) Amino acid sequences of the neck linker region. KIF1A, XKid, hKid, KIF5C, and KIF1AMD-XKidSt are shown. Cyan, motor domain of KIF1A and KIF5C; orange, motor domain of Kid; magenta, neck linker; green, neck coiled-coil domain. (C) Representative SDS-PAGE analysis of purified KIF1AMD-XKidSt fusion protein. The protein is visualized using a Stain-Free gel. The molecular weight standards are indicated on the right side of the SDS-PAGE image. (D) Mass photometry analysis of KIF1AMD-XKidSt at 10 nM. The expected molecular mass is 103 kDa. Histogram shows particle counts, and lines indicate Gaussian fits. The mean ± SD and percentage of each peak are shown. (E, F) Representative kymographs showing the motility of KIF1A(1-393)LZ and KIF1AMD-XKidSt in the presence of 2 mM ATP. Note that KIF1AMD-XKidSt exhibits diffusion-like fluctuations while they are moving. This phenomenon is not observed in KIF1A(1-393)LZ. Scale bars: horizontal 10 µm; vertical 10 seconds. (G) Dot plots showing the velocity of KIF1A(1-393)LZ and KIF1AMD-XKidSt. Each dot shows a single datum point. Green bars represent mean ± S.D. ****p<0.0001, Unpaired t-test. n=273 and 434 particles for KIF1AMD-XKidSt and KIF1A(1-393)LZ, respectively. (H) Dot plots showing the run length of KIF1A(1-393)LZ and KIF1AMD-XKidSt. Each dot shows a single datum point. Green bars represent median value and interquartile range. ****p<0.0001, Mann–Whitney test. n=273 and 434 particles for KIF1AMD-XKidSt and KIF1A(1-393)LZ, respectively.

Figure 5—figure supplement 1
AlphaFold2 prediction of neck linker domain.

AlphaFold2 was used to predict structures of sequences spanning the neck linker (magenta) to the coiled-coil region (green). Compared with KIF5C (A), both human Kid (hKid, B) and Xenopus Kid (XKid, C) contain extended regions predicted to correspond to the neck linker. KIF5C is shown in an autoinhibited conformation. Whether the predicted conformations of hKid and XKid represent autoinhibited or active states was not assessed, as the analysis was intended to identify candidate neck-linker and coiled-coil regions. Scale bars, 10 nm.

Figure 5—video 1
The motility of KIF1AMD-XKidSt chimeric protein.

The motility of KIF1AMD-XKidSt was analyzed through time-lapse imaging recorded at 10 fps and played back at 25 fps. The dimensions of the displayed frame are 30 µm by 7 µm.

Figure 5—video 2
The motility of KIF1A protein.

The motility of KIF1A(1-393)LZ was analyzed through time-lapse imaging recorded at 10 fps and played back at 25 fps. The dimensions of the displayed frame are 30 µm by 7 µm.

Figure 6 with 2 supplements
DNA movement driven by Xkid.

(AD) sfGFP-tagged XKid at 1 nM was mixed with 20 nM Cy3-labeled double-stranded or single-stranded DNA and observed by TIRF microscopy. (AC) Representative kymographs showing the movement of full-length XKid with double-stranded 100 bp DNA (A), full-length XKid with single-stranded 100-base DNA (B), and XKid(1-496) with double-stranded 100 bp DNA (C). Scale bars: horizontal, 10 µm; vertical, 100 seconds. (D) Frequency of DNA movement along microtubules, normalized by microtubule length and observation time. Each dot represents an individual measurement from a different microtubule. Bars indicate mean ± SD. n=29 microtubules per condition. (EI) Single-molecule analysis of XKid–sfGFP on Cy3-labeled DNA. Purified XKid–sfGFP was mixed with Cy3-labeled DNA at final concentrations of 1 nM XKid–sfGFP and 20 nM DNA and observed by TIRF microscopy. (EG) Representative kymographs showing the movement of full-length XKid–sfGFP on 100 bp DNA (E), 1000 bp DNA (13 ng/µl; F), and 2000 bp DNA (26 ng/µl; G). Scale bars: horizontal, 10 µm; vertical, 100 seconds. (H) Velocity of DNA movement along microtubules. Each dot represents one DNA molecule. Bars indicate mean ± SD. n=48, 53, and 53 DNA molecules for 100, 1000, and 2000 bp DNA, respectively. (I) Run length of DNA movement along microtubules. Each dot represents one DNA molecule. Bars indicate mean ± SD. n=48, 53, and 53 DNA molecules for 100, 1000, and 2000 bp DNA, respectively.

Figure 6—figure supplement 1
hKid drives the movement of double-strand DNA.

Representative kymographs showing the movement of full-length hKid fused with mNeonGreen and double-strand DNA labeled with Cy3. Scale bars: horizontal, 5 µm; vertical, 5 minutes.

Figure 6—video 1
The motility of dsDNA driven by hKid protein.

The motility of hKid-mNeonGreen (green) and Cy3-labeled double-strand DNA (red) was observed through time-lapse imaging recorded at 0.5 fps and played back at 10 fps. The dimensions of the displayed frame are 25 µm by 8 µm.

AlphaFold3 prediction and functional validation.

(A) AlphaFold3-predicted structure of the XKid DNA-binding domain modeled as a monomer with double-stranded DNA (dsDNA). The prediction showed a low interface confidence score between XKid and dsDNA (ipTM = 0.12, pTM = 0.64). (B–D) AlphaFold3-predicted structure of the XKid DNA-binding domain modeled as a dimer with dsDNA. The two XKid molecules are shown in green and gray, and the two DNA strands are shown in red and blue. This dimeric model showed higher confidence for the XKid–DNA complex (ipTM = 0.79, pTM = 0.84). (B) Side view of the predicted XKid dimer–dsDNA complex. (C) Magnified view of the predicted XKid–DNA interface. K605 and K606, highlighted in yellow, are positioned near dsDNA. A cyan dashed line indicates the predicted hydrogen bond between K605 and DNA. (D) View of the same model shown in (B) after a 90° rotation. (E, F) Representative kymographs showing the movement of dsDNA along microtubules. Wild-type XKid supported movement of 1000 bp dsDNA (E), whereas the K605A mutant abolished detectable dsDNA movement (F). Scale bars: horizontal, 10 µm; vertical, 100 s.

Model.

Model for Kid-dependent DNA and chromosome transport. Kid forms processive dimers that can directly bind and transport double-stranded DNA along microtubules. On mitotic chromosomes, multiple Kid dimers may bind along chromosome arms and cooperate to generate polar ejection forces. This model is not drawn to scale and does not fully represent the structural complexity of condensed chromatin.

Tables

Table 1
Motile properties of constructs used in this study.

Motor protein constructs underwent purification through affinity chromatography followed by size-exclusion chromatography, as detailed in the ‘Materials and methods’ section. The reported velocities and run lengths represent mean values ± standard deviation (SD). Notably, XKid(1-437) failed to demonstrate consistent processive motion across three separate protein preparations, indicating a lack of detectable activity (ND: not detected).

ProcessiveOligomeric state in SECVelocity (µm/second)Run length (µm)
Full-length hKidYesDimer0.11±0.041.8±0.7
Full-length XKidYesDimer0.11±0.041.3±0.5
XKid(1-496)YesDimer0.12±0.033.1±2.4
XKid(1-437)NoMonomerNDND
KIF1A(1-393)LZYesDimer1.4±0.411±11
KIF1AMD-XKidStYesDimer0.67±0.186.7±6.0

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Shinsuke Niwa
  2. Natsuki Furusaki
  3. Tomoki Kita
  4. Yuki Suzuki
  5. Kyoko Chiba
(2026)
The chromokinesin Kid (KIF22) forms a homodimer, moves processively along microtubules, and transports double-stranded DNA
eLife 13:RP102828.
https://doi.org/10.7554/eLife.102828.3