Variants in CAPZA2, a member of an F-actin capping complex, cause intellectual disability and developmental delay

Abstract

The actin cytoskeleton is regulated by many proteins including capping proteins that stabilize actin filaments (F-actin) by inhibiting actin polymerization and depolymerization. Here, we report two pediatric probands who carry damaging heterozygous de novo mutations in CAPZA2 (HGNC: 1490) and exhibit neurological symptoms with shared phenotypes including global motor development delay, speech delay, intellectual disability, hypotonia and a history of seizures. CAPZA2 encodes a subunit of an F-actin-capping protein complex (CapZ). CapZ is an obligate heterodimer consisting of α and β heterodimer conserved from yeast to human. Vertebrate genomes contain three α subunits encoded by three different genes and CAPZA2 encodes the α2 subunit. The single orthologue of CAPZA genes in Drosophila is cpa. Loss of cpa leads to lethality in early development and expression of the human reference; CAPZA2 rescues this lethality. However, the two CAPZA2 variants identified in the probands rescue this lethality at lower efficiency than the reference. Moreover, expression of the CAPZA2 variants affects bristle morphogenesis, a process that requires extensive actin polymerization and bundling during development. Taken together, our findings suggest that variants in CAPZA2 lead to a non-syndromic neurodevelopmental disorder in children.

Introduction

Intellectual disability (ID) is an early onset cognitive neurodevelopmental disorder including both intellectual and adaptive functioning impairment. Global developmental disorders (GDDs) are typically associated with children less than 5 years of age and are defined by a significant delay in two or more areas of development, and it is thought to predict a future diagnosis of ID (1,2). The etiology of GDD/ID is heterogeneous and is one of the most common condition observed in pediatric clinics as they affect 1–3% of children under the age of 5 (3,4). De novo mutations have been shown to contribute to over 40% of the GDD/ID in different studies, including severe non-syndromic sporadic ID patients and a large cohort of undiagnosed developmental disorders (5,6). To date, more than 800 genes are known to be involved in the pathogenesis of ID (7). Clinically, many IDs are syndromic and associated with dysmorphisms. Only ~14% of ID causative genes lead to non-syndromic ID (8).

Here, we report the identification of de novo missense variants in CAPZA2 (HGNC: 1490) that are associated with GDD/ID, hypotonia and rare seizures. CAPZA2 encodes a subunit of an F-actin-capping complex CpaZ. Dynamic actin filaments are regulated by numerous actin regulatory proteins. Newly formed filaments elongate rapidly, but elongation can be arrested by capping at the fast growing end or barbed end. Capping allows the cell to maintain fine temporal and spatial control over its F-actin distribution (9). Numerous capping proteins prevent elongation and depolarization by blocking the addition as well as loss of actin monomers. One of the ubiquitous capping complexes is named CapZ. It consists of an α and β heterodimer which is conserved from yeast to human. Both subunits have very similar secondary and tertiary structures yet lack any amino acid sequence similarity (10). In vertebrates, the α subunit is encoded by three CAPZA genes, CAPZA1, CAPZA2 and CAPZA3, which encode α1, α2 and α3 subunits, respectively (11). In contrast, the β subunits, β1 and β2, are produced by alternative splicing from a single CAPZB gene in vertebrates. Most tissues have both α1 and α2 isoforms, but the ratio varies widely, while α3 is testis specific (11). Both α and β subunits contain an actin-binding domain of ~30 amino acids at the carboxyl terminal end called the tentacle domain. In vitro studies in cell-free systems have shown that the α-tentacle plays a major role in actin affinity, and numerous highly conserved basic amino acids on the α-tentacle affect the binding to acidic residues on the F-actin ends. However, mutations of single basic amino acids disrupt capping mildly (12,13).

In vitro cellular studies have reported that knockdown of either the α or β subunit causes the loss of lamellipodia and explosive formation of filopodia in melanoma cells (14). Both α and β subunits of CapZ are expressed in the rat hippocampus, and knockdown of either subunit in primary hippocampal neuronal cultures results in a marked decline in spine density accompanied by increased filopodia-like protrusions (15).

A recent study reported a de novo, dominant, balanced chromosomal translocation that disrupts the CAPZB gene in a patient with GDD, micrognathia, cleft palate and hypotonia (16). Similarly, in zebrafish, homozygous loss of capzb causes craniofacial and muscular defects, whereas heterozygous animals did not display obvious phenotypes (16). However, homozygous CAPZA2 knockout mice are lethal prior to weaning and exhibit defects in multiple organ systems including the nervous system (17). To date, no patient with mutations in any of the CAPZA genes has been reported.

Flies carry a single CAPZA orthologue named cpa (capping protein alpha) that is evolutionarily highly conserved with a DIOPT score of 14/15 (18,19). cpa is an essential gene and its loss causes homozygous first instar lethality that is rescued by ubiquitous expression of wild-type cpa (20). Cpa regulates F-actin capping together with Cpb (Capping protein beta), the orthologue of the human CAPZB (DIOPT score 15/15). Null alleles of either cpa or cpb are lethal, and mutant cell clones cause a severe F-actin accumulation in larval imaginal disc cells (20–22). cpa and cpb play roles in multiple developmental processes including epithelial integrity of wing discs (20), eye development (21) and oogenesis (23).

Here, we report two pediatric probands who carry de novo mutations in CAPZA2 and exhibit neurological symptoms including global developmental delay, ID, hypotonia and a history of seizures. Expression of the human reference CAPZA2 in flies rescues the lethality associated with the loss of fly cpa. However, the CAPZA2 variants identified in the probands rescue at significantly lower efficiency than the reference cDNA. Moreover, expression of the CAPZA2 variants affects bristle morphogenesis, a process that requires extensive actin polymerization during development. Our data suggest that these de novo variants are mild loss of function mutations but that they can act as dominant negative variants in actin polymerization in bristles.

Results

Clinical profiles of probands

We identified two patients with variants in CAPZA2. The first proband, which is a 2.5-year-old Chinese female, conceived via in-vitro fertilization. She has two healthy brothers and a healthy fraternal twin sister. A few weeks after birth, she exhibited hypotonia and had feeding difficulty. She developed seizures at 7 months of age and infantile spasms occurred at 10 months of age. Electroencephalography (EEG) performed at 10 months revealed hypsarrythmia, but her brain MRI was normal. She was treated with adrenocorticotropic hormone for 14 days, but the spasms were controlled after a single treatment. Three subsequent EEGs showed no abnormalities. She presented with GDD, crawled at 18 months and was able to sit at the age of 2 years. She now is delayed in her speech and can only utter a few syllables. She has a head circumference of 48.2 cm (50th percentile), a height of 91 cm (50th percentile) and a weight of 10 kg (−2 SD) and has no dysmorphic features. By sharing our findings through GeneMatcher (24), we found the second proband, who is a 9-year-old girl of European descent with three healthy brothers. Hypotonia was first noted after birth, and she had difficulty suckling but was breast fed during the first 7 months. She had an atypical febrile seizure that was controlled without medication at the age of one, but no abnormalities were found on EEG, and she is controlled without medication. She presented with delay in gross and fine motor development, sitting and crawling at 18 months and walking at 3 years with an abnormal gait. She had very limited speech and has been diagnosed with autism at the age of 5. Currently, she has severe ID and speech delay, and her development is comparable to a 2-year-old. She also has difficulty with new situations and is easily distressed. Her MRI, at age 3, revealed late myelination and a non-specific white matter abnormality located in the right insula; at age 5, she exhibited a mild abnormal myelination in the frontal area, as well as parietal and occipital mildly dilated peripheral liquor spaces. She does not have facial dysmorphism. Her feeding is normal and her current height is 1.40 m (+0.49SD) and weight is 23 Kg (−1.74SD). More detailed clinical descriptions for each proband are in Table 1 and Supplementary Note.

Table 1.

Clinical features of individuals with heterozygous de novo mutations in CAPZA2

	Proband 1	Proband 2
Origin	Chinese	European
Variant	p.Arg259Ala	p.Lys256Glu
Gender	Female	Female
Age	2.5 years	9 years
Growth
Short stature	−	−
Microcephaly	−	−
Dysmorphic features	−	−
Development
Speech delay	+	+
Motor delay (milestones)	+	+
ID	NA	+
Neurological
Autism	NA	+
Hypotonia	+	+
Seizure history	+	+
MRI abnormality	−	+; mild abnormal myelination in frontal area, mild perivascular spaces dilation in parietal and occipital area
Others
Neonatal feeding difficulty	+	+
Additional findings	−	Hypopigmentation on right lower leg, mild hyperpigmentation in left inguinal area and both upper legs; toe walking

+, Feature present; −, feature absent; NA, feature not available for current age.

The two families are non-consanguineous. Copy-number variant (CNV) assays did not reveal any pathogenic candidate in the affected individuals. Whole-exome sequencing (WES) of proband and parents followed by Sanger sequencing revealed heterozygous de novo variants in CAPZA2 (GenBank: NM_00613) c.776G>T, p.Arg259Leu and c.766A>G, p.Lys256Glu in probands 1 and 2, respectively (Supplementary Material, Fig. S1). These correspond to some of the positively charged amino acids in or near the actin binding domain.

An additional case with de novo, heterozygous variant of CAPZA2 c.776G>T, p.Arg259Leu was found in the Deciphering Developmental Disorders database (6). However, we were unable to contact the investigators of this study.

The de novo variants of CAPZA2 are predicted to be deleterious

The CAPZA2 gene is intolerant of loss-of-function (LoF) with a pLI score of 1 and an o/e (observed/expected) ratio of 0, and no observed LoF variants were found in normal population databases, including ExAC (25) and gnomAD (26). There are no observed deletions in a DGV database of reference individuals (27). CAPZA2 is somewhat constrained to missense variation with a z score of 1.84 based on gnomAD with o/e = 0.58 (25). The affected residues in the probands are absent from the ExAC and gnomAD databases and are predicted to be deleterious by CADD (28), PolyPhen (29), PROVEAN (30), M-CAP (31) and Mutation Taster (32) (Table 2) (18). The variant residues from the patients are conserved basic amino acids: p.Arg259 is located at the beginning of the tentacle domain of CAPZA2, whereas the p.Lys256 is very near (Fig. 1A). Note that in vitro biochemical studies of chicken Capza1 have indicated that both p.Arg259 and p.Lys256 are important for F-actin binding, and mutating either residue mildly decreases actin capping rates (12,13). In summary, the above data suggest that these variants affect protein function and may be deleterious.

Table 2.

De Novo CAPZA2 variants

Proband	Variant	CADD	PolyPhen	PROVEAN	M-CAP	Mutation taster
1	7:116557836:G>T; c.776G>T; p.R259G	32	Probably damaging; 1	Deleterious; −6.476	Possibly pathogenic; 0.048	Disease causing; 1a
2	7:116557826:A>G; c.766A>G; p.K256E	39	Probably damaging; 1	Deleterious; −3.952	Possibly pathogenic; 0.368	Disease causing; 1a

^aA value close to 1 indicates a high security of the prediction.

Abbreviations: CADD, Combined Annotation Dependent Depletion; PolyPhen, Polymorphism Phenotyping; PROVEAN, Protein Variation Effect Analyzer; M-CAP, Mendelian Clinically Applicable Pathogenicity.

Figure 1.

Cpa, the ortholog of CAPZA1 and CAPZA2, is essential in flies. (A) Alignment of CAPZA1, CAPZA2 and Cpa. The tentacle domain is highlighted in magenta, the CAPZA2 variants identified in the probands are marked with a red box and the residues are conserved in human CAPZA1 and fly Cpa. (B) Schematic depicting mutant alleles of cpa and genomic span of genomic rescue construct inserted on chromosome 3 (VK31) of the fly. The lethality of transheterozygous cpa^107E/cpa^69E mutants is rescued by the 93 Kb genomic P[acman] construct.

Human CAPZA2 can substitute for the loss of fly cpa

The CAPZA2 gene is evolutionarily conserved in eukaryotes, and the fly homolog of CAPZA1, CAPZA2 and CAPZA3 is a single gene corresponding to cpa. Cpa shows a high sequence similarity to CAPZA1 and CAPZA2 (DIOPT score of 14 out of 15 for both) and a lower DIOPT score to CPAZA3 (7/15). Cpa and CAPZA2 share 77% similarity and 61% identity over the entire length of the proteins and contain a well-conserved tentacle domain at the carboxyl terminal end (Fig. 1A). Two available null alleles cpa^107E and cpa^69E contain non-sense mutations at amino acids 162 and 180, respectively, prior to the tentacle domain and are lethal at the first instar stage (Fig. 1B) (20). Heterozygotes of both alleles are viable and have no obvious phenotypes. The lethality of cpa^107E/cpa^69E trans-heterozygous flies but not either of the homozygotes can be fully rescued by the introduction of an 93 kb P[acman; CH321-25F16] (33) genomic BAC rescue construct encompassing the cpa gene (Fig. 1B). This indicates that cpa^107E and cpa^69E carry second site lethal mutations. Hence, we used the cpa^107E/cpa^69E transheterozygote mutants as cpa null mutants in subsequent experiments.

To study the CAPZA2 function, we generated transgenic flies with UAS-CAPZA2 reference cDNA. The reference and variant CAPZA2 proteins were expressed at comparable levels in flies (Supplementary Material, Fig. S2A). The lethality of cpa null mutants is fully rescued by ubiquitous expression of reference UAS-CAPZA2 under control of Tub-GAL4 at 25°C, showing functional conservation of human CAPZA2 and fly cpa. To examine whether the CAPZA2 variants are loss of function alleles, we ubiquitously expressed the CAPZA2 cDNAs in the cpa null mutant background. The lethality of cpa null mutants is rescued by ubiquitous expression of the two variant cDNAs under the control of Tub-GAL4 at 25°C (Table 3), indicating that the p.Arg259Leu and p.Lys256Glu variants are not severe loss of function alleles. Next, we lowered cDNA expression level by using the weaker Da-GAL4 driver and by lowering the temperature to 22°C to assess if the variants are partial loss of function (34). Using these conditions, the reference CAPZA2 is able to rescue the lethality of the cpa mutants at 23% of the expected frequency, whereas the p.Arg259Leu and p.Lys256Glu constructs rescue viability at 12.2 and 6% of the expected frequency, respectively (Table 3). The statistically significantly reduced rescue ability of p.Arg259Leu and p.Lys256Glu suggest that they only affect the function of CAPZA2 mildly in vivo.

Table 3.

The CAPZA2 variants are mild loss of function that rescued the lethality of cpa null mutants at lower efficiency

Number of adults		Observed	Expected	Viability
cpa/cpa; Tub-GAL4>UAS-CAPZA2 (25°C)	Reference	101	89.5	112.8%
	p.Arg259Leu	95	74	128.4%ns
	p.Lys256Glu	78	95.5	81.7%ns
cpa/cpa; Da-GAL4>UAS-CAPZA2 (22°C)	Reference	32	139	23.02%
	p.Arg259Leu	9	155.5	5.79%***
	p.Lys256Glu	17	139.5	12.19%*

Viability: observed/expected × 100%. Superscript: ns, P > 0.05; ^*P < 0.05; ^***P < 0.001 by chi-square test between each variants to corresponding reference.

CAPZA2 p.Lys256Glu variant overexpression leads to bristle defects in Drosophila

To determine if CAPZA2 variants can induce a gain of function or dominant negative effect, we ubiquitously overexpressed the CAPZA2 reference and variants in wild-type background (p.Arg259Leu and p.Lys256Glu). We varied the CAPZA2 proteins levels by culturing flies at 18, 22, 25 and 29°C given that the UAS-cDNA expression levels increase with increasing temperature. Note that the difference in expression between 18 and 25°C can be very significant (34). Ubiquitous expression of reference or variant CAPZA2 cDNAs by Tub-GAL4 in wild-type flies results in viable and fertile flies at all tested temperatures suggesting that CAPZA2 overexpression is not toxic under these conditions. However, we observed that the p.Lys256Glu variant induces a subtle phenotype in bristles. Sensory bristles of the peripheral nervous system correspond to one of the four cells of mechanosensory organs (35). The deflection of bristles provides proprioceptive information akin to hair cells in semicircular canals or Merkel cells in the skin (36). The bristle cells carry very long extensions and are supported by many longitudinal actin bundles, which are composed of a series of short actin modules attached end-to-end (Supplementary Material, Fig. S2B and C) (37). Bristle morphogenesis requires massive amounts of actin polymerization during development. Hence, they are extremely sensitive to actin disruptions and mutations in numerous actin regulatory proteins, including Profilin (38), ADF/cofilin (39), ADF/cofilin phosphatase (40) and Twinfilin (41). Interestingly, a partial loss of the partner of Cpa, Cpb, also leads to bristle defects including bending and branching caused by F-actin accumulation and an abnormal actin cytoskeleton (42). However, a function for Cpa in bristle morphogenesis has not previously been described.

Overexpression of p.Lys256Glu causes a mild morphological defect characterized by bending or branching of a subset of bristle cells, the macrochaetae and the largest bristles on the notum (Fig. 2A and D). In contrast, overexpression of the reference or p.Arg259Leu variant did not cause a bristle phenotype at all temperatures tested. We quantified the number of macrochaeta with defects of the 14 macrochaeta on the notum of each fly and observed only an average of 1.8 defective macrochaetae when p.Lys256Glu is expressed compared with none in flies with reference or p.Arg259Leu CAPZA2 overexpression. This suggests a subtle dominant gain or dominant negative function for the p.Lys256Glu variant during bristle morphogenesis.

Figure 2.

CAPZA2 variants act as partial loss of function as well as mild dominant negative mutations in fly bristle morphogenesis. (A–C) Optical images of fly bristles on the notum of the indicated genotypes. All flies were cultured at 25°C. (D) Quantifications of defective macrochaetae (out of a total of 14) on the notum of flies. The defects include bending and branching. The number of flies with different genotypes was analyzed which are in the corresponding columns; error bar: SEM. ns, P > 0.05; ****P < 0.0001 by one-way ANOVA with Dunnett’s multiple comparison test between each variants to corresponding reference. (E) Confocal projection images of longitudinal F-actin bundles labeled by fluorescent Phalloidin in pupal bristles of the indicated genotypes (bristle tip is up), followed by single cross sections of the F-actin bundles.

CPAZA2 mutations behave as partial loss of function and dominant negative mutations

If the CAPZA2 variant(s) act as weak dominant negative allele(s), reducing the level of wild cpa may exacerbate the bristle phenotype as described for antimorphic alleles. In contrast, if the mutations correspond to dominant gain of function alleles or hypermorphic alleles, reducing the cpa level may not affect the phenotype (43). We therefore examined the bristle phenotype caused by CAPZA2 overexpression in cpa/+ heterozygous flies. Note that cpa/+ heterozygous flies have no bristle phenotype. As a control, we overexpressed the reference CAPZA2: these flies did not exhibit any phenotype. In contrast, p.Lys256Glu expression showed bristle phenotype and exhibits 3.6 defective macrochaetae (Fig. 2B and D) consistent with a dominant negative model. Moreover, p.Arg259Leu expression also caused a mild bristle phenotype with 1.5 defective macrochaeta, suggesting a milder dominant negative effect of p.Arg259Leu variant.

Next, we examined the phenotypes of the overexpression of the CAPZA2 reference and variant cDNAs in cpa null mutants. Again, null cpa animals die as first instar larvae. However, cpa null mutant flies expressing the reference CAPZA2 enclosed as adults and showed a subtle bristle phenotypes with 1 macrochaetae being affected with a mild bending or split bristle phenotype (Fig. 2C and D). The actin cytoskeleton labeled by fluorescent Phalloidin was well organized in parallel arrays and attached to the cell membranes in developing pupal bristles of these flies (Fig. 2E). Hence, expression of the reference cDNA rescues lethality and does not disrupt bristle morphogenesis. Interestingly, expression of both CAPZA2 variants in cpa/cpa mutants also rescues lethality, again showing that they are functional proteins. However, they cause severe bristle defects, including bending, branching, shortening, missing bristles as well as multiple defects of single bristle (Fig. 2C and D). Importantly, the F-actin organization was severely disrupted in bristle for both variants (Fig. 2E). These results suggest that the variant proteins are functional with respect to rescuing lethality but have a dominant negative effect on bristle morphogenesis.

Discussion

The CapZ complex caps actin as a αβ heterodimer. Our data show that the expression of the human CAPZA2 rescues the lethality of cpa null mutants (Table 3), indicating that CAPZA2 potentially form a heterodimer with fly Cpb and function as a proper capping protein. In vitro biochemical data have provided evidence that the CapZ heterodimer containing the chicken CAPZA α1-subunit with mutations in conserved p.Arg259Ala or p.Lys256Glu leads to a reduced actin capping rate (12,13), but they do not seem to affect the binding with the β subunit to form the heterodimer (44). Our in vivo data show that the CAPZA2 p.Arg259Leu and p.Lys256Glu variants are not as efficient in rescuing lethality and bristle development of cpa null mutants when compared with the reference CAPZA2, consistent with the in vitro observations for CAPZA1, suggesting a capping issue caused by the two variants (Fig. 3). Moreover, in the wild-type or the cpa/+ heterozygous background, the reference CAPZA2 expression does not cause a phenotype, whereas the variants cause a mild dominant negative effect (Fig. 2). These data suggest that in addition to the variants causing a partial LoF of their biochemical function, the variants also act to inhibit fly Cpa in bristle morphogenesis. This dominant negative effect might be caused by a competition of the exogenously expressed CAPZA2 variants with the endogenous fly Cpa for heterodimer formation with Cpb causing a capping issue (Fig. 3).

Figure 3.

Model of CAPZA2 function. (Left) Capping protein composed of fly endogenous Cpa and Cpb caps the F-actin; (Middle) subtle defective capping of the complex composed of reference CAPZA2 and fly Cpb; (Right) capping complex composed of variant CAPZA2 and fly Cpb exhibits defective capping than that composed of reference CAPZA2 and Cpb.

Our data indicate that cpa in Drosophila is required for bristle morphogenesis, which is known to be regulated by cpb, the CapZ β subunit coding gene in the fly (42). In addition to bristle morphogenesis, other phenotypes caused by loss of Cpa or Cpb are very similar, including early lethality, defects in eye development (21), epithelial integrity of wing discs (20) and oogenesis (23). These phenotypes are probably due to F-actin capping defects as lack of either subunit disrupts the CapZ complex (20–23). However, in vertebrate, there are three α subunits encoding genes, and both CAPZA1 and CAPZA2 are expressed ubiquitously (11), suggesting that there may be functional overlap between the two genes. Hence, it is likely that their loss is associated with less severe defects (45). In contrast, the human genome only has a single β subunit encoding gene CAPZB, and the patient with a dominant de novo truncation in the CAPZB gene showed GDD, hypotonia and craniofacial defects (16). However, the proband is too young to establish ID. In contrast, no craniofacial malformations were observed in the two CAPZA2 probands in this study. Hence, CAPZA1 may partially compensate for some of the function of CAPZA2. Alternatively, the two variants described here may cause milder phenotypes solely because they are mild loss of function variants (Fig. 3). In this context, it is interesting to note that there are 20 children with deletions that contain the CAPZA2 gene in the Decipher database of chromosomal imbalance (46). Most of these children exhibit speech delay and many of have ID, apraxia and autism, similar to the two patients described here. However, some have physical or/and facial malformation (Supplementary Material, Table S1). Hence, the defects in these patients may be caused, at least in part, by the loss of one copy of CAPZA2. Our patients with CAPZA2 missense variants display phenotypes that are similar or milder than those associated with individuals who carry relatively small deletions that affect 20–50 genes (deletion size: 5.8–4.2 Mb; Supplementary Material, Table S1). Based on our data, we argue that some phenotypes may be caused by the loss of function, whereas others may be due to dominant negative mechanisms. The discovery of additional individuals with different variants may help unravel these issues.

Material and Methods

Diagnosis and human genetics

For family 1, genomic DNA of the proband and her parents were extracted, hybridized and enriched. The genomic DNAs were sequenced using NovaSeq 6000 platform (Illumina, San Diego, USA) with 150 bp pair-end reads. Raw image files were processed using CASAVA (v1.82) for base calling. Adaptor sequences of the raw data were trimmed at the tail of reads using Cutadapt (v1.15) (47) and then aligned to the human reference genome (UCSC hg19) with BWA (v0.7.15) (48). Duplicated reads were marked by Picard (v2.4.1; http://broadinstitute.github.io/picard/). Qualimap (v2.2.1) (49) was used to calculate base quality metrics, genome mapping rate and the coverage of targeted regions. Base quality score recalibration, indel realignment and variants (SNVs & InDels) calling were performed following the best practice protocol of the Genome Analysis Toolkit (GATK, v3.8) (50). Variant filtering was performed using a finely tuned in house script (Clabee Genomics). Pass-filter variants were annotated using the Pubvar variant annotation engine (www.pubvar.com) and VEP (51) (release 88). For genetic analysis, we separately identified variants that fit the dominant and recessive inheritance models. Variants that met anyone of the following criteria were excluded from genetic analysis: maximum population frequency was larger than 0.01, genotype confidence was low or predicted as benign by five algorithms including SIFT (52), MetaSVM (53), M-CAP (31), PolyPhen-2 (29) and MutationTaster (32). The pathogenic evidence of candidate disease-causing variants was scored by InterVar (1.0.8) (54) according to ACMG guidelines (55). A CNV assay used low-coverage whole-genome sequencing. All the above analysis was performed on Seqmax (www.seqmax.com).

For family 2, DNA isolated from whole blood from patient and parents was enriched for exomes using SureSelect V5 (Agilent Technologies) and sequenced on an Illumina HiSeq 2500 (Illumina). The CNV analysis was based on a single-nucleotide polymorphism array (Agilent 180 K oligo-array). The trio-WES data were analyzed by variant assessment by filtering for coding and splice site variants in line with recessive or de novo inheritance according to diagnostic standards of the UMCU (University Medical Center Utrecht) using Alissa Interpret Software (Agilent Technologies).

The identified variants have been submitted to Clinvar, accession number: SCV001190587, SCV001190600.

Drosophila strains

The following stocks were obtained from the Bloomington Drosophila Stock Center (BDSC): y¹ w*; P{tubP-GAL4}LL7/TM3, Sb¹ Ser¹ (RRID: BDSC_5138) and w*; P{w[+mW.hs] = GAL4-da.G32}UH1 (RRID: BDSC_5460). BAC line w¹¹¹⁸; Dp(2;3)GV-CH321-25F16, PBac{GV-CH321-25F16}VK00031 (Clone ID: CH321-25F16) (33) is from GenetiVision. The cpa mutant stocks were gifts from F. Janody: yw; FRT42D, cpa^69E/CyO, y + (FBID: FBal0193693) and yw; FRT42D, cpa^107E/CyO, y + (FBID: FBal0193694) (20).

Generation of CAPZA2 Drosophila transgenes

Transgenic stocks were generated as previously described (56). Briefly, pDONR221-CAPZA2 cDNA (GenBank: NM_00613; clone: IOH7253) was used for this study. Variants were generated using site-directed mutagenesis with the following primers: CAPZA2-p.Arg259Leu, FW 5′-CAAAGCCTTACTTCGACAGTTGCCAGTTACACGC-3′; RV 5′-CAACTGTCGAAGTAAGGCTTTGAAAGTAGTGTCCG-3′ and CAPZA2-p.Lys256Glu, FW 5′-GGACACTACTTTCGAAGCCTTACGTCGACAGTTG-3′; RV 5′-CGACGTAAGGCTTCGAAAGTAGTGTCCGACATTG-3′ followed by Sanger sequencing. The reference and variant pDONR221-cDNAs were cloned into the pGW-attB-HA vector (57) using Gateway cloning via LR clonase II (Thermo Fisher Scientific). These expression constructs were inserted into the VK33 (PBac{y + -attP-3B}VK00033) docking site by φ-C31-mediated transgenesis (58).

Viability

To quantify the viability of cpa null flies that express the UAS-CAPZA2-cDNA, we crossed cpa^69E/CyO, Act5C-GFP; UAS-CAPZA2/TM6B, Hu, Tb (reference or variant cDNAs) with cpa^107E/CyO, Act5C-GFP; enhancer-Gal4/TM6B, Hu, Tb (Tub-Gal4 or Da-Gal4) and cultured the flies at 22 or 25°C. The numbers of offspring with different genotypes without TM6B, Hu and Tb marker were quantified. The offspring with CyO marker is cpa heterozygous mutants, whereas flies without CyO marker are cpa homozygous null mutants. The expected number of Non-CyO-flies was counted as half of the observed number of CyO-flies. The viability of CAPZA2 rescued cpa null mutants was presented as percentage of observed non-CyO divided by the expected non-CyO fly numbers. Statistical significance between each genotype and the controls was determined with chi-square tests.

Immunochemistry

For immunoblots, proteins were extracted by cold RIPA buffer (Sigma-Aldrich) with proteinase inhibitors (genDEPOT) from whole third instar larvae then subjected to SDS-PAGE and immunoblotting. Mouse anti-CAPZA1/2 (DSHB 5B12.3, 1:100) and mouse anti-α-tubulin (Millipore-Sigma T6074, 1:20 000) were used in the assay.

For immunostaining of actin bundles in bristles, the pupae of 40–43 h APF (after puparium formation) were dissected and fixed in 4% paraformaldehyde for 1 h, washed in 0.2% Triton X-100 in PBS and incubated with Alexa 488 phalloidin (1:200) for 2 h.

Image collection and statistical analysis

Confocal images of F-actin were collected with a Leica confocal microscope SP8 using a 40 × 1.30 NA oil objective and LAS X software. Images of bristles on the notum and scutellum were collected by a Leica MZ16 stereomicroscope and Image-Pro Plus software (Media Cybernetics). All images of notum bristles were captured using identical settings among different genotypes. Images were processed by Fiji imageJ (59), and brightness, contrast and color were adjusted by Photoshop CC 2019 (Adobe).

For the quantification of the bristles, only the 14 scutellar and thoracic macrochaetae were analyzed in this study. The numbers of macrochaetae with major defects including bending, splitting and branching were quantified. Statistical significance between each genotype and the controls was determined by one-way ANOVA followed by Dunnett’s post hoc test using GraphPad Prism 6 (La Jolla California USA; www.graphpad.com). Data of average bristle numbers were expressed as mean ± standard error of the mean (SEM).

The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.

Funding

National Institutes of Health Office (R24OD022005 to H.J.B.); National Institute of General Medical Sciences (R01GM067858 to H.J.B.); Hunan Provincial Major Science and Technology (2019SK1010 to H.W.); and National Natural Science Foundation of China (81801136 to X.M.).