Loss of FYCO1 leads to cataract formation

Autophagy is a degradation process of cytoplasmic proteins and organelles trafficked to degradation vesicles known as autophagosomes. The conversion of LC3-I to LC3-II is an essential step of autophagosome formation, and FYCO1 is a LC3-binding protein that mediates autophagosome transport. The p62 protein also directly binds to LC3 and is degraded by autophagy. In the present study, we demonstrated that disrupting the FYCO1 gene in mice resulted in cataract formation. LC3 conversion decreased in eyes from FYCO1 knockout mice. Further, FYCO1 interacted with αA- and αB-crystallin, as demonstrated by yeast two-hybrid screening and immunoprecipitation analyses. In eyes from knockout mice, the soluble forms of αA- and αB-crystallin, the lens’s major protein components, decreased. In addition, p62 accumulated in eyes from FYCO1 knockout mice. Collectively, these findings suggested that FYCO1 recruited damaged α-crystallin into autophagosomes to protect lens cells from cataract formation.

We determined the presence or absence of cataract formation after we observed lens under a microscope objectively. Although it is easy to distinguish normal from mild and moderate symptoms, it is difficult to classify severity. For example, it is difficult to distinguish mild from moderate due to subjectivity in observation and analysis. In Fig. 2, we classified the lens based on cataract (+) or cataract (−).

Histological analysis of lens from FYCO1 knockout mouse. Histological analysis of lenses from
wild-type and KO mice was performed (Fig. 3).
Wild-type mice (left upper and lower panels) had almost normal lenses (left upper panel) except one lens with mild degeneration of lens fibers in the anterior pole of the lens (arrow, left lower panel). In KO mice (right panel), the lens capsule was ruptured in the posterior pole of the lens, and swollen lens fibers were prolapsed into the vitreous body with hypertrophied lens epithelium (right upper panel, arrow and arrowhead). In severe cases, the lens nucleus was also prolapsed to the vitreous body, and almost all lens fibers and epithelium were vacuolated and hypertrophied (right lower panel, arrow and arrowhead).
As shown in Fig. 3, we observed the phenotypical variability between the cataracts of the right and left eyes. The results suggest that there is a certain threshold for cataract formation. Crystallin aggregates and cataracts become more noticeable when the threshold is exceeded. Autophagy, p62 expression and α-crystallin solubility. Because FYCO1 regulates autophagy, the conversion of LC3-I to LC3-II was quantified in lenses of wild-type (+/+) and KO (−/−) mice. The conversion of LC3-I to LC3-II was analyzed by western blotting. As shown in Fig. 4A, LC3-I/II conversion was active in wild mice (LC3-II/LC3-I = 2.07). Contrastingly, conversion of LC3-I to LC3-II was decreased in lenses of FYCO1 KO (−/−) mouse (LC3-II/LC3-I = 0.64; Fig. 4A). The results were reproducible. This suggested that autophagy activity was decreased in the eyes of the FYCO1 KO mouse. α-crystallin, which is comprised of A and B subunits, is the major type of human lens protein. Because cataract is thought to be due to crystallin aggregation, αA-crystallin and αB-cystallin water solubility was analyzed  Figure 1. Analysis of FYCO1 tissue distribution and generation of FYCO1 knockout mice. (A) Extracts from various 4-week-old male C57BL/6J mouse tissues and MEFs (30 μg of protein) were subjected to western blot analysis with anti-FYCO1 antibody. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a loading control. This experiment was performed once using one mouse. (B) The mouse FYCO1 gene was disrupted by the insertion of a neomycin resistance gene cassette (Neo) in the first coding exon. Open and filled boxes represent coding and noncoding exons, respectively. The diphtheria toxin A gene cassette (DT-A) was inserted outside of the 3' homologous region for negative selection. Restriction enzyme sites and probes used for Southern blot analysis are indicated. (C) Lysates from 4-week-old WT and KO mouse tissues (heart and brain) were subjected to western blot analysis with anti-FYCO1 antibody. This experiment was performed once using a mouse. (D) Lysates from 4-week-old WT and KO mouse tissues (eyes) were subjected to western blot analysis with anti-FYCO1 antibody. This experiment was performed once using a mouse.  p62 accumulation in FYCO1 KO mice. As impairment of autophagy is accompanied by p62 accumulation, p62 was next analyzed using western blotting. p62 was increased in the water-soluble and SDS-soluble fractions of eyes from FYCO1 KO (−/−) mice, relative to the WT (+/+) mice. p62 expression was also examined using immunohistochemistry. Ring-shaped p62 aggregation was observed in lenses from KO mice, while p62 expression was not observed in lenses from WT mice (Fig. 4C). Additionally, p62 expression was robust outside of the OFZ in lenses from KO mice (Fig. 4C). These data indicated that ablation of FYCO1 caused lens p62 accumulation, which was suggestive of impaired autophagy.
Interaction between FYCO1 and αA/B-crystallin. The yeast two-hybrid system was used to assess the hypothesis that FYCO1 interacted directly with α-crystallin. As shown in Fig. 5A, the results clearly showed the interaction between FYCO1 and αA-crystallin and FYCO1 and αB-crystallin.
To further confirm the interaction between FYCO1 and αA/B-crystallin, COS-7 cells were transfected as indicated, and cell extracts were subjected to immunoprecipitation (IP) with anti-Myc antibody and western blot (WB) analysis with the indicated antibodies. The interaction between FYCO1 and αA-crystallin and αB-crystallin was clearly confirmed (Fig. 5B).
We speculated that in the lens, FYCO1 is essential for clearance of degenerated crystallin. Therefore, we determined if FYCO1 associated with αA-and/or αB-crystallin in WT mouse eyeballs. When WT mouse eyeball lysates were immunoprecipitated with FYCO1 antibody, αB-crystallin ( Fig. 5C-b) and αA-crystallin ( Fig. 5C-c) were detected compared with lysates immunoprecipitated using control antibody ( Fig. 5C-a). These data indicated that FYCO1 directly interacted with αA-crystallin and αB-crystallin in the lens.  Experiments were performed in duplicate with similar results, and a representative result is shown. This experiment was performed twice using a mouse per experiment. Both eyeballs from a 4-week-old mouse were used, and 50 µg protein was loaded in electrophoresis. (B) Analysis of αA-crystallin and αB-crystallin water solubility and water-insolubility (SDS-solubility and SDS-insolubility) in lenses from WT and KO mice. Protein levels of αA-crystallin and αB-crystallin were measured by western blot analysis. P62 levels were also assessed by western blot analysis. This experiment was performed twice using a mouse per experiment. Both lenses (right and left) were used and 10 µg protein was loaded in electrophoresis. (C) Immunohistochemical analysis of p62 expression. Lenses of WT and KO mice were stained with anti-p62 antibody. Mouse eyeballs were enucleated and immediately fixed using SUPER FIX (Kurabo). Paraffin sections were incubated with primary antibodies overnight at 4 °C, followed by secondary antibodies for 1 h. This experiment was performed twice using one mouse per experiment. www.nature.com/scientificreports/ FYCO1 was not required for organelle degradation. We determined if FYCO1 was essential for organelle degradation. We stained paraffin sections of lenses with DAPI (nuclei staining dye), anti-Tom20 (mitochondrial marker protein), and anti-KDEL (ER marker protein GRP78). In lenses from WT and KO mice, nuclei, mitochondria, and ER were present in the cortical region but not in the OFZ (Fig. 6). These data suggested that FYCO1 was not required for organelle degradation in the OFZ. Immunoprecipitates from WT mice eyeball lysate using FYCO1 antibody or control antibody were probed with α-A or α-B crystallin antibodies. Compared with immunoprecipitants using control antibody, those using FYCO1 antibody were detected with α-A and α-B crystallin. Blots were reprobed with anti-rabbit IgH antibody as a loading control. Asterisk indicates a nonspecific band. This experiment was performed twice using one mouse per experiment. www.nature.com/scientificreports/ Proposed mechanism of action. The proposed model of FYCO1 function in lens protein quality control is illustrated (Fig. 7). FYCO1 recruited damaged αA/αB crystallin into autophagosomes and facilitated its degradation in autolysosomes to prevent cataract formation.

Discussion
In the present study, we demonstrated that disruption of the FYCO1 gene in mice caused cataract formation. In eyes from FYCO1 KO mice, LC3 conversion decreased, and p62 accumulation increased, suggesting impaired autophagy. In addition, FYCO1 interacted directly with αA-and αB-crystallin, as demonstrated by yeast twohybrid screening and immunoprecipitation analyses. In eyes from FYCO1 KO mice, water-soluble forms of αAand αB-crystallin, the lens's major components, were decreased. These findings suggested that FYCO1 recruited damaged α-crystallin into autophagosomes to protect lens cells from cataract formation. Based on these results, we have proposed a new mechanism for cataract formation, as illustrated in Fig. 7. FYCO1 recruits damaged α-crystallin (αA-and αB-crystallin) into autophagosomes and facilitates its degradation in autolysosome to prevent cataract formation (Fig. 7). Cataract is the leading cause of blindness worldwide 22,23 , and is characterized by lens protein aggregation 24 . The human lens is comprised of three major types of proteins, α-, β-and γ-crystallins, which account for Figure 6. FYCO1 was not required for organelle degradation. The presence of nuclei, mitochondria, and ER in the cortical region and OFZ in lenses from 12-week-old WT and KO mice was evaluated using immunohistochemical staining with anti-KDEL antibody, anti-Tom20 antibody, and DAPI. Equatorial and cortical regions are shown. Scale bars, 40 μm. This experiment was performed twice using a mouse in each experiment.  34,35 . The transparency of the lens depends on maintaining the tertiary structures and solubility of lens crystallin proteins. Homogenization of human lens in water or lysis buffer yields two fractions: a water-soluble fraction and a water-insoluble fraction 36,37 . About 50% of lens proteins from aged human eyes accumulate in the water-insoluble fraction 38,39 , and water-insoluble proteins are known to increase with aging 36,37 . Yang et al. reported that lens epithelium soluble αA-crystallin and αB-crystallin are decreased in age-related and congenital cataracts 40 . These findings are consistent with those of the present study (Fig. 4B). Yang et al. did not assess the role of FYCO1. Mutations of more than 50 genes such as crystallin and HSF4, have been reported in congenital cataract 30 . Shiels and Hejtmancik suggested that when mutations in crystallins or other lens proteins are sufficient in and of themselves to cause rapid and direct protein aggregation, they usually result in congenital cataract formation 29 . However, these studies did not assess FYCO1. further discussed that loss of FYCO1 function inhibits autophagosome transport from the perinuclear area to the periphery leading to vesicle accumulation and loss of transparency. However, they did not evaluate the interaction between FYCO1 and crystallin.
In summary, we demonstrated that disruption of the FYCO1 gene in mice resulted in cataract formation. In the eyes of FYCO1 KO mice, LC3 conversion was decreased. In addition, FYCO1 interacted with αA-and αB-crystallin. In eyes from FYCO1 KO mice, the soluble forms of αA-and αB-crystallin, the major lens protein components, were decreased. In addition, p62 accumulated in eyes from FYCO1 KO mice. These findings suggested that FYCO1 recruited damaged α-crystallin into the autophagosome to protect lens cells from cataract.

Western blot analysis. Tissues and cells were homogenized in buffer A (20 mM Tris
The resulting supernatants were designated as SDS-soluble fractions. The pellets were resuspended in buffer A containing 1% SDS and 1.3 M 2-mercaptoethanol, boiled at 95 °C for 3 min, and centrifuged at 19,000 × g for 10 min. The resulting supernatants were designated as SDS-insoluble fraction. Generation of FYCO1-deficient embryonic stem cells. A FYCO1-targeting vector was constructed by inserting a neomycin resistance gene cassette into a site just downstream of the FYCO1 start codon, preserving 7.0 kb (5') and 0.9 kb (3') of the flanking homologous regions at the FYCO1 locus. The targeting vector was linearized with NotI and electroporated into B6N-S1Utr embryonic stem (ES) cells derived from C57BL/6J mice 42 . Embryonic stem cells were cultured on a feeder layer of mitotically inactivated mouse embryonic fibroblasts (MEF) using a co-culture method described previously 42 Fig. S1). In brief, 3 μg of DNA was digested with restriction enzymes (digestion with EcoRI or double-digestion with ApaI and BssHII), followed by electrophoretic separation on agarose gels. DNA was transferred to nylon membranes in 20 × SSC, and hybridized with DIG-labeled DNA probes. Nylon membranes were washed with 0.1 × SSC/0.1% SDS. DNA fragments were detected using alkaline phosphatase labelled anti-DIG antibodies (Roche) and CDP-Star chemiluminescent substrate (Roche). Ultimately, four homologous recombination ES clones were established.

Animals.
To generate FYCO1 deficient mice, targeted ES cells were aggregated with CD-1 embryos as described previously 42 . Chimeric males were crossed with C57BL/6J females to examine the germline transmission of the targeted allele. Heterozygous (FYCO1 ±) mice were then backcrossed to C57BL/6J mice for at least 5 generations. The resulting heterozygous mice were intercrossed to produce homozygous (FYCO1−/−) mutants. There were no phenotypical differences among the number of backcrosses. Genotyping of mice was performed by PCR using the following primers: neo primer, 5'-ATT TTG AAT GGA AGG ATT GGA GCT ACGG-3' , which is homologous to the neomycin resistance gene cassette; FYCO1 forward primer, 5'-TAA ACA GGA AGG TGA AAA ACT TGG AGG -3' , which lies just upstream of the 3' homologous region in the targeting vector; and FYCO1 reverse primer, 5'-GTG GTT GTG AGC TAA GAC TGG TGC TG-3' , which lies just downstream of the 3' homologous region in the targeting vector. The 951-bp and 1095-bp fragments represented the wild-type and targeted alleles, respectively. C57BL/6J and CD-1 (ICR) mice were purchased from Charles River Laboratories Japan (Atsugi, Japan). In addition to eyes, other major organs of KO mice were examined macroscopically and histologically using light microscopy (data not shown).

Approval of animal experiments.
All animal experiments were conducted according to the institutional and national animal experimental guidelines. All experimental protocols were reviewed and approved by the Tokyo University Animal Care, Ethics and Experimentation Committee (Approval number 19-32).
Plasmid construction. cDNAs for human FYCO1 and αB-crystallin were amplified by RT-PCR using human heart total RNA (Clontech), and subcloned into expression vectors. The identity of the PCR products was confirmed by sequencing analysis. Human αA-crystallin cDNA was kindly provided by Dr. N. Fujii (Kyoto University). FLAG-tagged human FYCO1, Myc-tagged human αA-crystallin, and Myc-tagged human αB-crystallin were subcloned into the pTARGET vector (Promega) for transient gene expression.
Yeast two-hybrid analysis. Yeast  www.nature.com/scientificreports/ AH109 was co-transformed with plasmids encoding various protein regions fused to the GAL4 DNA-binding domain (GAL4BD) or GAL4 activation domain (GAL4AD). Protein-protein interactions were analyzed by growth of the transformants on minimum medium plates lacking histidine and adenine. To identify novel FYCO1-binding proteins, human brain and heart cDNA libraries were screened using various regions of human FYCO1 as bait. We performed yeast two-hybrid screening to identify FYCO1-binding proteins. In addition to αA-crystallin and αB-crystallin, LC3 and α-crystallin were also identified.
Immunohistochemistry. Mouse eyes were dissected and fixed immediately using SUPER FIX (Kurabo).
Pathological analysis. Eyes were harvested from WT and FYCO1 KO mice, and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). After fixation for 24 h. The eyes were subjected to histological examination. The eyes were dehydrated in a sequential ethanol series (50-100%) by an automated processor and embedded in paraffin wax. Serial sections (5 µm) were cut and stained with hematoxylin and eosin.
This study was carried out in compliance with the ARRIVE guidelines (http:// www. nc3rs. org. uk/ page. asp? id= 1357).