Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Nov 1;28(22):3029-3042.
doi: 10.1091/mbc.E17-03-0156. Epub 2017 Sep 6.

General and Specific Promotion of Flagellar Assembly by a Flagellar Nucleoside Diphosphate Kinase

Affiliations
Free PMC article

General and Specific Promotion of Flagellar Assembly by a Flagellar Nucleoside Diphosphate Kinase

Xiaoyan Zhu et al. Mol Biol Cell. .
Free PMC article

Abstract

Nucleoside diphosphate kinases (NDKs) play a central role in diverse cellular processes using the canonical NDK activity or alternative mechanisms that remain poorly defined. Our study of dimeric NDK5 in a flagellar motility control complex, the radial spoke (RS), has revealed new modalities. The flagella in Chlamydomonas ndk5 mutant were paralyzed, albeit only deficient in three RS subunits. RS morphology appeared severely changed in averaged cryo-electron tomograms, suggesting that NDK5 is crucial for the intact spokehead formation as well as RS structural stability. Intriguingly, ndk5's flagella were also short, resembling those of an allelic spoke-less mutant. All ndk5's phenotypes were rescued by expressions of NDK5 or a mutated NDK5 lacking the canonical kinase activity. Importantly, the mutated NDK5 that appeared fully functional in ndk5 cells elicited a dominant-negative effect in wild-type cells, causing paralyzed short flagella with hypophosphorylated, less abundant, but intact RSs, and accumulated hypophosphorylated NDK5 in the cell body. We propose that NDK5 dimer is an RS structural subunit with an additional mechanism that uses cross-talk between the two NDK monomers to accelerate phosphorylation-related assembly of RSs and entire flagella.

Figures

FIGURE 1:
FIGURE 1:
NDK5H121A expressed from a minigene caused paralyzed short flagella in WT cells. (A) Recombinant His-NDK51-201 exhibited NDK activity. Coomassie blue–stained SDS–PAGE showed Ni-NTA purification of His-tagged recombinant proteins (left). Pre, extract from bacteria expressing both MBP-His-NDK51-201 and the TVMV protease that could cleave off MBP; Post, the flow-through; Elu, the eluate. Identical amounts of proteins were analyzed in triplicate for NDK activity using the luciferase-based bioluminescence assay (right). RLU, relative light unit. (B) Representative transgenic strains from WT cells transformed with the minigene expressing full-length His-tagged NDK5 (left) or NDK5H121A (right). NDK5H121A transformants had short paralyzed flagella. The cells were fixed before light microscopy for length measurement. (C) Average flagellar length (Fla. Length) of transgenic strains expressing NDK5 or NDK5H121A. All strains from the NDK5 group, represented by randomly selected S1, S2, and S3 strains, contained only swimmers (S) like the parental strain. Four strains from the NDK5H121A groups contained mostly paralyzed cells but with a fraction of swimmers (P/S). Flagella of three of the four P/S strains had shorter flagella than those of the control strain. Asterisk indicates statistically significant differences (Student’s t test, p < 0.01, n = 50). (D) An anti-His Western blot revealed His-tagged NDK5 polypeptides in the flagellar samples from randomly selected S strains and representative P/S strains (left, top). Different protein loads indicated the signals were not oversaturated. Both His-tagged and untagged-NDK5 proteins were revealed by NDK5 polyclonal antibody (right, top). Total proteins were revealed by Ponceau red (bottom). Dynein, dynein heavy chains. (E) Plot Lanes analysis in the ImageJ software showed the ratio of His-tagged NDK5 vs. total NDK5 in the 20 μl flagellar samples in the Western blots in D.
FIGURE 2:
FIGURE 2:
Characterization of an ndk5 insertional mutant. (A) Schematic illustrating NDK5 protein and gene organization. NDK5 has an NDK domain followed by a Dpy-30 domain and an extended C-terminal tail containing three calmodulin-binding IQ motifs. The ndk5 mutant has an insertion of a 2.6-kb PMM-resistance cassette in the second intron. Black box, exon; white box, intron; black line, flanking sequence; arrowhead, the H121 codon; arrows, PCR primers. (B) PCR diagnosis of the insertion mutagenesis using templates prepared from ndk5 and the parental strain (Con.). A 0.8-kb fragment at the 3′ junction was amplified with the P1S and P1As primer pair from the mutant but not from Con. (left). A nearly 2-kb fragment at the 5′ region, including the second intron, was amplified with the P2S and P2As primer pair from the control but not from ndk5 (middle). An ∼0.2-kb fragment from the neighboring SPL4 gene, ∼5-kb upstream to the NDK5 gene, was amplified with the P3S and P3As primer pair from both templates (right).
FIGURE 3:
FIGURE 3:
The ndk5 mutant is defective in RSs and flagellar assembly. (A) Western blot analysis of axonemes revealed the absence of RSP1, NDK5, and HSP40 in the paralyzed ndk5 flagella. A fraction of the spoke scaffold protein RSP3 migrated slightly faster than that in WT cells (Con.). The other spoke proteins in the head (RSP4/6 and RSP9/10) and neck (RSP2 and RSP5) appeared normal. The axonemal proteins p28, IC1, and CPC1—subunits of inner dynein arms, outer dynein arms, and the central pair, respectively—were used as controls. The spoke-less pf14 was a negative control. (B) The antigen of mAb 3G3 is NDK5, and not RSP2, which is present in ndk5 axonemes. (C) NDK5 is a phosphoprotein. NDK5 in axonemes receiving the sham CIAP treatment migrated as double bands. In CIAP-treated axonemes, all NDK5 migrate with the lower band (arrowhead). CIAP treatment did not change the migration of the spoke scaffold protein RSP3. (D) Microscopy (left) and histogram (right) revealed short flagella (Fla.) in ndk5 cells cultured in minimal media, in contrast to the full-length flagella of the parental strain (Con.). The length phenotype was more pronounced in late log phase (compare day 4 and day 7, n = 50). (E) Average flagellar length corresponding to D. Asterisks indicate statistically significant differences (p < 0.01, n = 50). (F) ndk5 cells are deficient in flagellar regeneration. Flagella were excised from log-phase cells by pH shock and allowed to regenerate. Aliquots of cells were fixed periodically and imaged. Compared with the control, ndk5 cells regenerated flagella more slowly, and the final lengths were shorter (p < 0.01, n = 20).
FIGURE 4:
FIGURE 4:
Structure of radial spokes from WT (A, B) (EMD-2131; Bui et al., 2012), ndk5 mutant (C, D), and pf24 mutant (E, F; Pigino et al., 2011) Chlamydomonas. Density maps were obtained by averaging subtomograms involving the 96-nm periodic unit from cryo-electron tomography and shown as transverse (A, C, E) and longitudinal (B, D, F) sections. Red arrows indicate corresponding positions of the RS neck. The density at the position of the spokehead components, RSP1, 4, 6, 9, and 10, are missing in averaged cryo-electron tomograms from the ndk5 strain (C, D) similar to the pf24 headless mutant (E, F). Scale bar: 24 nm.
FIGURE 5:
FIGURE 5:
Heterogeneity of radial spokes in ndk5 axonemes. (A) Density map from the total 96 nm–based average shown as longitudinal sections. (B) A corresponding 3D variance map from aligned and normalized particles. Scale bar: 20 nm. The regions with high variance have darker pixel densities. An absence of variance and electron density in the spokehead and spokeneck indicates that the majority of spokes miss the bifurcated neck and head parts. High variance in the spoke stalks infers either orientation or compositional heterogeneity in that region. An absence of variance in individual MTDs serves as a positive control of applied analysis procedures. The representative density map subaveraged from subvolumes reveals typical straight, rigid conformation of spoke stalk with respect to the doublet in MTD 1 (C), in contrast to the subaverage from MTD 2 (D). Red dashed lines are drawn following the spoke-stalk density for the visual representation of the spoke tilt relative to the doublet. The subaverage from MTD 3 (E) shows complete absence of RS assembly. The numbering of the doublets is arbitrary and does not correspond to accepted numbering in the WT axoneme. Scale bar: 24 nm.
FIGURE 6:
FIGURE 6:
Genomic DNA expressing NDK5 or NDK5H121A rescued both the paralysis and length phenotypes of ndk5. (A) Flagellar-length (Fla. length) distribution showed that the flagella of ndk5 transgenic strains expressing NDK5 or NDK5H121A were longer than that of ndk5 but similar to that of the ndk5 parental strain. Cells cultured in minimal media were fixed at two indicated time points, and the lengths were measured from micrographs; n = 50. (B) Average flagellar length of corresponding data from Figure 4A. Asterisk indicates statistically significant differences (p < 0.01). (C) Western blot analysis of axonemes from representative transformants. RSP1, HSP40, and NDK5 polypeptides were restored. Note that NDK5 and NDK5H121A expressed by the BAC-derived transgene (triangle) migrated identically but faster than endogenous NDK5 (dot). (D) H121A mutation abolished NDK activity. The NDK activity of axonemes from ndk5 transformants expressing NDK5 was restored to the activity level of WT axonemes (Con.), whereas the activity of axonemes from NDK5H121A transformants, spoke-less mutant pf14, and ndk5 strains (either F0 or F2 from backcross) was ∼50% lower. The relative NDK activity was the RLU value normalized to the control. Quadruple aliquots from each sample were measured. Asterisk indicates statistically significant differences (Tukey HSD test, p < 0.01). (E) Live-cell imaging of transgenic ndk5 and pf14, respectively, expressing fluorescent NDK5 (top) and RSP3 (middle). ndk5 (bottom) was a negative control. NG-tagged NDK5 and RSP3 appeared identical in distribution and intensity throughout flagella.
FIGURE 7:
FIGURE 7:
Diminished RSs result in short flagella. (A) Flagellar-length (Fla. length) distribution of three pf14 strains cultured in minimal media. The defect in the RSP3 gene of cc613 and cc2496 that were derived from one common isolate was identical. Only flagella of cc613 were particularly short, and this was exacerbated at late log phase (day 5). The length phenotype and paralysis were rescued by a RSP3 transgene. (B) Average flagellar lengths of pf14 cells. Asterisks indicate statistically significant differences (p < 0.01, n = 50).
FIGURE 8:
FIGURE 8:
NDK5H121A expression from a genomic transgene in WT cells impaired flagellar motility, length, RS assembly, and RS phosphorylation. (A) The percentages of clones with swimmers (S), paralyzed cells (P), or a mixture (P/S) expressing NDK5 (first bar) or NDK5H121A (second bar) from one experiment. Average flagellar length (B) and flagellar-length (Fla. length) distribution (C) of representative strains cultured in minimal media showed that flagellar lengths of P strains (P1) were between the control and ndk5, whereas P/S strains (P/S1) had flagella of normal length. Asterisks indicate statistically significant differences (p < 0.01, n = 50). (D) Western blot analysis of axonemes from transformants with different motility levels. RSPs in the flagella of P/S strains (P/S1 and P/S2) appeared normal but reduced in P strains (P1 and P2) (top). In addition, NDK5 and the scaffold protein RSP3 were hypophosphorylated (arrow), especially NDK5H121A (triangle). IC140, a subunit of the inner dynein arm, was the loading control. *Con., a WT strain transformed with pNDK5. Dashed line, the arbitrary line to distinguish exogenous proteins expressed from the transgenes. (E) The ratio of rapidly migrating NDK5 from the transgenes relative to total NDK5 in D. Ratio of exogenous protein in Con. strain represents the background. Exo., Exogenous.
FIGURE 9:
FIGURE 9:
Accumulated hypophosphorylated NDK5 polypeptides in the cell body of the DN strain. (A) A representative Western blot of cell body extracts probed with anti-NDK51-201 serum. NDK5 in the DN strain P1 (lane 7) was more abundant than that in the control (lane 5) but similar to that in pf14 (x in lane 8), a control for reduced RS abundance in flagella. Negative control was ndk5 (lane 6). NDK5 bands in axonemes, by themselves or added to cell body extracts (lanes 1–4 and 9), served as markers for phosphorylated NDK5 and hypophosphorylated NDK5 that are endogenous (dot), compared with that from the transgene (triangle) or in pf14 (x). The axoneme sample loaded in lane 3 was one-sixth of the sample loaded in lane 1. (B) A Western blot probed with affinity-purified (Aff. Puri.) anti-NDK58-586 polyclonal antibody independently confirmed the hypophosphorylated state of endogenous NDK5 (dot) in the control. The band was undetectable in ndk5. The protein load was shown by Ponceau S stain (bottom).

Similar articles

See all similar articles

Cited by 1 article

  • The roles of a flagellar HSP40 ensuring rhythmic beating.
    Zhu X, Poghosyan E, Rezabkova L, Mehall B, Sakakibara H, Hirono M, Kamiya R, Ishikawa T, Yang P. Zhu X, et al. Mol Biol Cell. 2019 Jan 15;30(2):228-241. doi: 10.1091/mbc.E18-01-0047. Epub 2018 Nov 14. Mol Biol Cell. 2019. PMID: 30427757 Free PMC article.

References

    1. Attwood PV, Wieland T. Nucleoside diphosphate kinase as protein histidine kinase. Naunyn Schmiedebergs Arch Pharmacol. 2015;388:153–160. - PubMed
    1. Berg P, Joklik WK. Transphosphorylation between nucleoside polyphosphates. Nature. 1953;172:1008–1009. - PubMed
    1. Berthold P, Schmitt R, Mages W. An engineered Streptomyces hygroscopicus aph 7" gene mediates dominant resistance against hygromycin B in Chlamydomonas reinhardtii. Protist. 2002;153:401–412. - PubMed
    1. Biggs J, Tripoulas N, Hersperger E, Dearolf C, Shearn A. Analysis of the lethal interaction between the prune and Killer of prune mutations of Drosophila. Genes Dev. 1988;2:1333–1343. - PubMed
    1. Boissan M, Montagnac G, Shen Q, Griparic L, Guitton J, Romao M, Sauvonnet N, Lagache T, Lascu I, Raposo G, et al. Membrane trafficking. Nucleoside diphosphate kinases fuel dynamin superfamily proteins with GTP for membrane remodeling. Science. 2014;344:1510–1515. - PMC - PubMed

MeSH terms

Substances

LinkOut - more resources

Feedback