ZCCHC8, the nuclear exosome targeting component, is mutated in familial pulmonary fibrosis and is required for telomerase RNA maturation

Abstract

Short telomere syndromes manifest as familial idiopathic pulmonary fibrosis; they are the most common premature aging disorders. We used genome-wide linkage to identify heterozygous loss of function of ZCCHC8, a zinc-knuckle containing protein, as a cause of autosomal dominant pulmonary fibrosis. ZCCHC8 associated with TR and was required for telomerase function. In ZCCHC8 knockout cells and in mutation carriers, genomically extended telomerase RNA (TR) accumulated at the expense of mature TR, consistent with a role for ZCCHC8 in mediating TR 3' end targeting to the nuclear RNA exosome. We generated Zcchc8-null mice and found that heterozygotes, similar to human mutation carriers, had TR insufficiency but an otherwise preserved transcriptome. In contrast, Zcchc8^-/- mice developed progressive and fatal neurodevelopmental pathology with features of a ciliopathy. The Zcchc8^-/- brain transcriptome was highly dysregulated, showing accumulation and 3' end misprocessing of other low-abundance RNAs, including those encoding cilia components as well as the intronless replication-dependent histones. Our data identify a novel cause of human short telomere syndromes-familial pulmonary fibrosis and uncover nuclear exosome targeting as an essential 3' end maturation mechanism that vertebrate TR shares with replication-dependent histones.

Keywords: RNA processing; ciliopathy; lung disease; nuclear RNA exosome; telomerase RNA.

Figure 1.

Linkage analysis and whole-genome sequencing identify novel disease gene ZCCHC8 in familial pulmonary fibrosis with low telomerase RNA (TR). (A) Pedigree with pulmonary fibrosis proband (arrow) with affected relatives are indicated by the shaded symbols (key). The clinical history below each of the four shaded pedigree symbols refers to the age of onset of lung disease including idiopathic pulmonary fibrosis (IPF). (?) Asymptomatic individuals who had unknown affected status at the time of clinical assessment; (gray shading) unknown cause of death; (*) individuals with DNA who were included in the linkage analysis. (B) TR levels measured by quantitative real time PCR (qRT-PCR) in lymphoblastoid cell lines (LCLs). Arrow refers to proband (red) and pedigree identifiers refer to A. TR level from a DKC1 mutation carrier is a positive control. The data represent a mean of three experiments, each from independent RNA isolations. (C) Telogram shows age-adjusted lymphocyte telomere length by flow cytometry and fluorescence in situ hybridization (flowFISH) in the proband (arrow) and family (pedigree designations as in A). The validated telogram is based on 192 controls. (D) Phenotype assignments used in linkage (key) and genotype below each individual refers to ZCCHC8 SNP. Italicized genotypes refer to obligate carriers. (E) Log of the odds (LOD) ratio across autosomal chromosomes calculated from SNP data from 14 individuals, with arrow on chromosome 12 pointing to maximum LOD. (F) p.P186L conservation across eight vertebrate ZCCHC8 species with darker shading denoting more conserved residues. CCHC refers to Zinc-knuckle domain; PSP refers to proline-rich domain.

Figure 2.

ZCCHC8 loss of function is sufficient to cause low TR levels. (A) Immunoblot of ZCCHC8 in lymphoblastoid cell lines (LCLs) from healthy controls (C1 and C2) and unaffected relatives and mutation carriers labeled with pedigree identifiers from Figure 1A. Quantification from one blot and result replicated twice from independently harvested protein lysates. (B) Immunoblot of ZCCHC8, SKIV2L2, and RBM7 levels in proband's primary skin fibroblasts. (C) Immunoblot of transfected Myc-tagged (293FT cells) and endogenous ZCCHC8. (D) Mean ZCCHC8 mRNA levels ± SEM from LCLs in unaffected family members (n = 4) and ZCCHC8 p.P186L mutation carriers (n = 3). (E) Chromatogram showing that ZCCHC8 p.P186L mutation is expressed in LCL mRNA from proband (also verified in two other mutation carriers). (F) Immunoblot showing efficiency of shRNA knockdown of Luciferase (Luc), ZCCHC8, and NAF1 in HeLa cells. (G) Total TR levels measured by qRT-PCR (mean ± SEM from three independent knockdowns and RNA isolations). (H) Northern blot of TR after stable knockdown of ZCCHC8 and NAF1 (replicated twice with independent RNA isolations). (**) P < 0.01; (***) P < 0.001 (Student's t-test, two-sided).

Figure 3.

ZCCHC8 is required for its 3′ end maturation and telomerase function. (A) Compound heterozygous frameshift (fs) mutations introduced using CRISPR/Cas9 in HCT116 pseudodiploid cells. (B) Immunoblot for ZCCHC8 in HCT116-edited cells. (C) Scheme summarizing TR 3′ rapid amplification of cDNA ends sequencing (3′RACE-seq). TR 3′ ends were generally divided into mature (451 bp) and extended (>451 bp) where extensions are denoted by gray N's. (D) Summary of TR 3′RACE-seq fractions in isogenic ZCCHC8^+/+ and ZCCHC8^−/− cells. Color-coded key shows four categories of TR forms: mature (451 nt), adenylated (A)n, short genomically extended (g)n (<465 nt), and long genomically extended (>465 nt). Data are mean of three independent 3′RACE-seq analyses from three RNA isolations each from a different aliquot of a single clone. (E) qRT-PCR of extended TR forms beyond the 451 mature end (>20, >51, >784 nt). Data are mean of three independent RNA isolations similar to D. (F) TR Northern blot of edited ZCCHC8^+/+ and ZCCHC8^−/− cells. (G) Total TR levels by Northern bot (six blots from three RNA isolations). (H) Telomerase activity measured by telomere repeat amplification protocol (TRAP) assay in ZCCHC8^+/+, ZCCHC8^−/−, and NAF1^S329/S329 HCT116 cell extracts. Activity was quantified on serially diluted extracts (1, 1/5, 1/25, and 1/125) against a PCR-amplified internal control (IC). RNase-treated wild-type extract and no template PCR reaction are included as negative controls. (I) Mean TRAP activity of 1/5× diluted extracts (three independent TRAP assays, each from a different lysate). (J) Summary of 3′RACE-seq of TR forms from control and proband's primary skin fibroblasts with speciation as in D. (K) qRT-PCR values of extended TR forms in primary skin fibroblasts as in E, mean of three technical replicates). (L) Amplified TR from input and Myc-ZCCHC8 immunprecipitated fractions (293FT cells) using primers falling within the mature TR sequence. (M) qRT-PCR of extended TR (>51 nt extended beyond the 3′ mature TR end) after transfection of tagged ZCCHC8, DIS3, EXOSC10/RRP6, and PARN into HCT116 ZCCHC8^−/− cells (three to four independent transfections/experiment). Data are expressed as mean ± SEM (*) P < 0.05; (**) P < 0.01 (Student's t-test, two-sided).

Figure 4.

Zcchc8-null mice have TR insufficiency. (A–C) Immunoblot for ZCCHC8, SKIV2L2, and RBM7, respectively, on lysates from mouse ear fibroblasts. (D,E) Northern blot for mouse TR and quantification. For E, mean reflects mice Zcchc8^+/+ (n = 4, 2M/2F), Zcchc8^+/− (n = 4, 2M/2F), Zcchc8^−/− (n = 3M), mTR^+/− (n = 2, sex unknown) and mTR^−/− (n = 2, sex unknown). (F) TR 3′ extended levels (>20 bp) relative to Hprt as measured by qRT-PCR. Mouse numbers and M/F designations as in E. Data are expressed as mean ± SEM. (*) P < 0.05; (**) P < 0.01 (Student's t-test, two-sided).

Figure 5.

ZCCHC8 complete loss causes progressive and fatal neurodevelopmental phenotype. (A, top row) Images showing head profile of Zcchc8 wild-type, heterozygous, and homozygous null mice (41–46 d-old). The labels show the genotype with a male (left) and a female (right) for each genotype. Zcchc8^−/− mice have abnormal head profiles with domed crania, as outlined by the dashed line. (Middle row) CT head mid-sagittal images show Zcchc8^−/− mice have dome-shaped crania. (Bottom row). Volume-rendered (VR) CT images of mouse calvaria show widened cranial sutures in Zcchc8^−/− mice (seen in three of six imaged). None of Zcchc8^+/+ or Zcchc8^+/− mice (four mice imaged/genotype) had this feature. Each vertical image group is from the same mouse except the last column (two different females). (B) Representative H&E coronal sections from 8-wk-old heads (all male) show no differences in Zcchc8^+/− mice (11 examined) compared with Zcchc8^+/+ mice (10 examined). In contrast, Zcchc8^−/− mice had severe ventricular dilation (nine of 11 examined). (C) E12.5 brain sections show Zcchc8^−/− have microcephaly but intact brain structures with no ventriculomegaly in utero. (D) Image of E12.5 embryos from a single dam showing expected Mendelian ratios but Zcchc8^−/− embryos have small crania. (E) Cranial area of newborn (P0) pups measured on VR CT images and corrected to left femur length on the same images (Zcchc8^+/+ n = 7, 3M/4F; Zcchc8^+/− n = 10, 3M/7F; Zcchc8^−/− n = 4, 3M/1F). (F) Mean cranial area ± SEM relative to age for all three genotypes showing that Zcchc8^−/− mice develop macrocephy after birth. Newborn mice include those listed in E and older mice (Zcchc8^+/+, n = 4, 3M/1F; Zcchc8^+/−, n = 4, 3M/1F; Zcchc8^−/−, n = 5, 3M/2F). Dashed lines denote 95% confidence intervals. (**) P < 0.01 (Student's t-test, two-sided).

Figure 6.

The Zcchc8^−/− transcriptome shows up-regulation and misprocessing of low-abundance intronless RNAs other than TR. (A) Heat map with dendrogram of gene expression showing unsupervised analysis of 9788 high-quality genes from brain RNA-seq analysis. Colors denote mean-subtracted FPKM expression values on a log₂ scale (Zcchc8^+/+, n = 5; Zcchc8^+/−, n = 3; Zcchc8^−/−, n = 6 embryonic brains sequenced). Each column is labeled below by WT, HET, and KO followed by the embryo number (1, 2, 3, etc.), referring to respective Zcchc8 genotypes. The log₂ expression value was subtracted from the mean log₂ expression value of the entire cohort. The dendrogram showing relatedness of the samples is above, and relatedness of the gene transcripts is at the left. The differential change in RNA expression is shown as positive and negative change on color scale in the key above the top right corner. (B,C) Volcano plots depicting the log₂-fold changes (X-axis) versus −log₁₀ P-values calculated by two-tailed one-way ANOVA (Y-axis) for the Zcchc8^+/− and Zcchc8^−/− versus Zcchc8^+/+ comparisons, respectively. Each dot represents a single transcript. (D) Histogram of number of genes at each expression value denoted on the x-axis by the mean log₂FPKM values obtained from Zcchc8 wild-type embryos (n = 5). RNAs that have more than two SD higher levels in the Zcchc8^+/+ versus Zcchc8^−/− comparison are shown in red (n = 197) and fall on the low end of the histogram with TR and its mean FPKM in wild-type embryos shown. Down-regulated RNAs, defined as less than two SD (n = 43), are shown in blue appear uniformly distributed on the distribution. (E) Histogram of the most up-regulated (>2SD) transcripts in the Zcchc8^−/− versus Zcchc8^+/+ by exon number shows the largest subset is intronless RNAs (42 of 188 with known gene structure, 22%). The pie chart divides the intronless RNAs by functional category. (F) Annotation of 28 up-regulated intronless RNAs (TR, histones and cilia) shows a majority of the histones represented are replication-dependent histones (RDH) (23 of 24, 96%). The majority have an annotated transcript size in the range of TR between 400 and 560 (22 of 28). Columns referring to 5′ end and 3′ end refer to visualized additional reads beyond annotated gene boundaries with 5′ end reads referring to upstream reads that are not necessarily contiguous (manually identified in the Integrative Genome Viewer [IGV]). (G,H) Genome browser read coverage plots from IGV viewer showing extended 3′ ends as labeled above from two histone genes in each of Zcchc8^+/+, Zcchc8^+/−, Zcchc8^−/− transcriptomes. (I,J) Coverage plots for two coiled-coil domain containing cilia genes Ccdc89 and Ccdc182, respectively, by genotype. For Ccdc89, there is also an increase in discontinuous upstream of gene 5′end reads that resemble so-called promoter upstream transcripts (PROMPTs).