Kif18a is specifically required for mitotic progression during germ line development

Abstract

Genome integrity in the developing germ line is strictly required for fecundity. In proliferating somatic cells and in germ cells, there are mitotic checkpoint mechanisms that ensure accurate chromosome segregation and euploidy. There is growing evidence of mitotic cell cycle components that are uniquely required in the germ line to ensure genome integrity. We previously showed that the primary phenotype of germ cell deficient 2 (gcd2) mutant mice is infertility due to germ cell depletion during embryogenesis. Here we show that the underlying mutation is a mis-sense mutation, R308K, in the motor domain of the kinesin-8 family member, KIF18A, a protein that is expressed in a variety of proliferative tissues and is a key regulator of chromosome alignment during mitosis. Despite the conservative nature of the mutation, we show that its functional consequences are equivalent to KIF18A deficiency in HeLa cells. We also show that somatic cells progress through mitosis, despite having chromosome alignment defects, while germ cells with similar chromosome alignment defects undergo mitotic arrest and apoptosis. Our data provide evidence for differential requirements for chromosome alignment in germ and somatic cells and show that Kif18a is one of a growing number of genes that are specifically required for cell cycle progression in proliferating germ cells.

Keywords: Cell cycle; Gametogenesis; Germ cells; Kinesin; Laboratory mouse; Mitosis; Mitotic spindle.

Figure 1. The gcd2 mutation affects the highly conserved R308 residue within the KIF18A motor domain

Gcd2 is a G to A transition in exon 7 of Kif18a. This is a missense mutation that changes an AGA codon to an AAA, leading to the single amino acid change, R308K. Arginine 308 is a highly conserved amino acid in the motor domain of the protein.

Figure 2. Mutations in R308 disrupt KIF18A’s localization and chromosome alignment function

HeLa cells were treated with siRNA oligonucleotides targeting endogenous KIF18A. Cells were rescued with GFP labeled wild type KIF18A or with the mutant GFP labeled KIF18A-R308K or KIF18A-R308A (A). Unlike KIF18A^WT-GFP, KIF18A^R308K-GFP and KIF18A^R308A-GFP (green) failed to concentrate near kinetochores (red) and instead were distributed along the spindle between the centrosomes (γ-tubulin, blue) in KIF18A-depleted HeLa cells (B). To measure centromere alignment, the distribution of anti-centromere antigen fluorescence (ACA FL) along the normalized pole-to-pole axis was measured and the full width at half maximum (FWHM) was calculated for each condition. Representative images of control and Kif18A siRNA treated cells stained for γ-tubulin (red) and ACA (green) are shown (C). A plot of average ACA FL distribution (green circles) within a control cell is well fit by a single Gaussian function (dashed line) (D). Average FWHM measurements from HeLa cells treated with control or KIF18A siRNAs and expressing GFP or the indicated Kif18A constructs are plotted. Cells expressing KIF18A-R308 mutations displayed a similar increase in kinetochore distribution (p > 0.05), which were both significantly different from Kif18A-wt expressing cells (p < 0.001 ). The number of cells analyzed (N) is reported and error bars indicate SEM (E).

Figure 3. Primary Kif18a^gcd2 embryonic fibroblasts do not display a cell cycle delay

MEFs derived from mutant embryos were slow growing (A) and had reduced viability by trypan blue exclusion (B). There were no significant differences in histone H3 ser10 phosphorylation (C), in apoptosis related DNA fragmentation (TUNEL) (D) or cell cycle staging by DNA content between mutant and wild type MEFs. All experiments were performed with three independently derived primary cell lines per genotype and technical replicates as described in the Methods.

Figure 4. Kif18a^gcd2 mutant and wild type MEFs progress through mitosis with similar timing

Mutant and control MEFs were imaged over a 16h period by differential interference contrast (A). Time from nuclear envelope breakdown to anaphase was recorded at 2-minute intervals and no significant difference was found between mutant and wild type MEFs (p = 0.56), n = 93 cells for Kif18a^+/+ and 58 cells for Kif18a^gcd2/gcd2 from 2 cell lines per genotype (B).

Figure 5. Kif18a^gcd2 mutant MEFs exhibit chromosome alignment defects

Mitotic profiling of Kif18a^gcd2 mutant MEFs immunolabeled for tubulin (green) and CREST (red, centromeres) (A) revealed a high ratio of preanaphase mitotic cells with unaligned versus aligned chromosomes in mutant MEFs compared to wild type controls, n=3 cell lines per genotype, mean +/− s.d. is displayed (B).

Figure 6. Kif18a is ubiquitously expressed in the fetal gonad

In situ hybridization of Kif18a and the germ cell specific Pou5f1 (Oct3/4) in wild type E12.5 ovaries and testes shows that Kif18a expression is not restricted to the germ line (A). RT-PCR shows that Kif18a is expressed in the fetal gonad, as well as the adult testes and ovary. Kif18a expression positively correlates with the presence of proliferating germ cells (high in the wild type testes and reduced in the germ cell deficient Kit^W/W-v testes and ovaries). Lane 1, Kif18a^gcd2/Kif18a^gcd2 E11.5-12.5 fetal gonad; Lane 2, +/+ E11.5-12.5 fetal gonad; Lane 3, Kit+/+ E11.5-12.5 fetal gonad; Lane 4, Kit^W/W-v E11.5-E12.5 fetal gonad; Lane 5, C57BL/6J adult ovary; Lane 6, Kif18a^gcd2/Kif18a^gcd2 adult ovary; Lane 7, Kit^W/W-v adult ovary; Lane 8, C57BL/6J adult testes; Lane 9, Kif18a^gcd2/Kif18a^gcd2 adult testes; Lane 10, Kit^W/W-v adult testes; Lane 11, C57BL/6J adult liver; Lane 12, C57BL/6J adult brain; Lane 13, DNA, no RNA, no reverse transcriptase; Lane 14, no reverse transcriptase; Lane 15, PCR water control, 100 bp ladder (B).

Figure 7. Kif18a^gcd2 mutant germ cells arrest in mitosis with chromosome alignment defects

Cell cycle analysis by DNA content revealed a significantly higher percentage of Kif18a^gcd2 mutant germ cells in G2 compared to wild type (Student’s t-test, p=0.006) and a correspondingly decreased percentage of cells in G1 (Student’s t-test, p=0.002) (A). Phosphorylated histone H3 (B), and TUNEL labeling (C), both indicative of G2/M checkpoint activation were significantly increased in mutant germ cells (Student’s t-test (pH3, germ cells), p=0.005; Student’s t-test (pH3, somatic cells), p=0.02; Student’s t-test (TUNEL, germ cells), p=0.05). In all cases, error bars represent standard deviation from the mean value from at least 3 age-matched, sibling biological replicates per genotype from multiple litters. For mutant genotypes, more biological replicates were required to obtain sufficient germ cells numbers for flow cytometry. Mitotically dividing spermatogonial cells (MVH positive) from pre-pubertal, mutant testes had poor spindle organization and showed defects in mitotic chromosome alignment (D).