Latent TGF-β-binding protein-2 (LTBP-2) is an extracellular matrix protein associated with microfibrils. Homozygous mutations in LTBP2 have been found in humans with genetic eye diseases such as congenital glaucoma and microspherophakia, indicating a critical role of the protein in eye development, although the function of LTBP-2 in vivo has not been well understood. In this study, we explore the in vivo function of LTBP-2 by generating Ltbp2–/– mice. Ltbp2–/– mice survived to adulthood but developed lens luxation caused by compromised ciliary zonule formation without a typical phenotype related to glaucoma, suggesting that LTBP-2 deficiency primarily causes lens dislocation but not glaucoma. The suppression of LTBP2 expression in cultured human ciliary epithelial cells by siRNA disrupted the formation of the microfibril meshwork by the cells. Supplementation of recombinant LTBP-2 in culture medium not only rescued the microfibril meshwork formation in LTBP2-suppressed ciliary epithelial cells but also restored unfragmented and bundled ciliary zonules in Ltbp2–/– mouse eyes under organ culture. Although several reported human mutant LTBP-2 proteins retain normal domain structure and keep the fibrillin-1-binding site intact, none of these mutant proteins were secreted from their producing cells, suggesting secretion arrest occurred to the LTBP-2 mutants owing to conformational alteration. The findings of this study suggest that LTBP-2 is an essential component for the formation of microfibril bundles in ciliary zonules.
Primary congenital glaucoma (PGC) is an inherited disease commonly found in children as early as their first year of life, often resulting in blindness. Recently, homozygous mutations in LTBP2 were reported in patients with PCG who develop high intraocular pressure (IOP), megalocornea, buphthalmos and ectopia lentis (1–3). On the other hand, other groups reported on homozygous LTBP2 mutations in patients with an autosomal recessive ocular disease who presented with megalocornea, microspherophakia and lens dislocation with secondary glaucoma later in life or without glaucoma (4–6). These results clearly identify causative LTBP2 mutations that lead to genetic eye diseases including glaucoma, although it remains unclear whether glaucoma is the primary phenotype or a secondary result of other primary defect such as lens dislocation.
LTBP2 encodes a matrix protein, latent transforming growth factor beta (TGF-β)-binding protein-2 (LTBP-2), which is one of four LTBP family proteins (LTBP-1–4) (7,8). LTBPs are multidomain glycoproteins containing epidermal growth factor (EGF)-like domains and eight-cysteine domains, both of which are also found in other matrix proteins, namely fibrillin-1, -2 and -3. As the name indicates, LTBP-1, -3 and -4 bind to TGF-β molecules and function to keep the signaling molecule inactive in the extracellular matrix of tissues until further induction releases active TGF-β from the complex (7). Despite the homology with and a domain structure similar to that of other LTBPs, LTBP-2 does not bind to TGF-β molecules and seems to have a function distinct from direct involvement in TGF-β signaling (9). Several reports, including ours, indicated an association between LTBP-2 and microfibrils, which mainly comprise fibrillin-1 and -2, and suggested a role for LTBP-2 in the elastic fiber assembly process (10–13). Shipley et al. created Ltbp2-deficient mice and reported that the homozygous mutant mice died at an early embryonic stage, suggesting an essential role of LTBP-2 in early embryogenesis of mice (14). However, this gene-knockout study of mice was inconsistent with human studies because no birth defects and/or early developmental abnormalities other than eye diseases were reported in human LTBP2 mutations. Because of this discrepancy, it has been understood that LTBP-2 has different functions among species in vivo.
In this study, we independently created conditional Ltbp2-mutant mice with minimum alterations in the mouse genome and created systemic homozygous null mutant mice to confirm their phenotype as previously reported (14). Surprisingly, the homozygous null mutant mice survived to adulthood without any gross abnormalities. We analyzed the phenotypes of the mutant mice and concluded that Ltbp2 null mutant mice developed lens dislocations owing to the disorganized formation of ciliary zonules, which comprise bundles of microfibrils. We further analyzed the function of LTBP-2 and found that the protein plays an essential role in the assembly of large microfibril bundles in both cultured cells and organ-cultured mouse eyes. We also investigated the pathogenic effects of human LTBP2 mutations and found that secretion arrest occurs in all reported LTBP-2 mutants that conserve the fibrillin-1-binding domain.
Ltbp2-mutant mice survived to adulthood
To investigate the biological function of LTBP-2 in vivo, we created conditional knockout mice that could delete exon 1 of Ltbp2 in the presence of Cre-recombinase (Fig. 1A). The neomycin cassette (pgk-Neo) was removed from the floxed allele by FRT/Flp-mediated recombination. We tested the Cre-recombinase-dependent elimination of Ltbp2 expression in vivo by mating heterozygous floxed mice with transgenic mice expressing Cre-recombinase in the germline to produce systemic heterozygotes (Ltbp2+/–). The Ltbp2+/– mice were intercrossed to produce systemic Ltbp2 homozygous mutant mice (Ltbp2–/–). Using reverse transcription–polymerase chain reaction (RT–PCR), western blot and immunohistochemical analyses, we observed no detectable Ltbp2 transcripts or LTBP-2 proteins in Ltbp2–/– tissues (Fig. 1B–D), indicating that the expression of Ltbp2 was completely lost in the homozygous mutant mice. Contrary to a previous report (14), Ltbp2–/– mice survived to adulthood without showing gross developmental defects; furthermore, the homozygous mutant mice were fertile, indicating no reproductive abnormalities. By intercrossing Ltbp2+/– mice, we obtained homozygous mutant mice at the expected Mendelian ratio (Supplementary Material, Table S1), also indicating no embryonic lethality of the mutant mice. From these results, we concluded that lack of Ltbp2 expression caused no severe effects on development and reproduction of mice in vivo.
Lack of LTBP-2 had no obvious contribution to tissue elasticity
LTBP-2 has been found to associate with fibulin-5, a protein necessary for elastogenesis (15–17), and was suggested to regulate elastic fiber assembly in vitro (8,10,11). Down-regulation of LTBP2 by siRNA in human dermal fibroblasts allows fibrillin-1-independent deposition of fibulin-5 and elastin in the matrix (10). Therefore, we investigated whether there were any detectable abnormalities in the elastic fibers in Ltbp2-mutant mice. Histological examination did not reveal any morphological changes such as a sign of emphysema in Ltbp2–/– lung tissues (Supplementary Material, Fig. S1A). Electron microscopic observation also demonstrated normal elastic fiber formation in the extracellular matrices of the lungs and normal elastic lamellae in the aortic media of the mutant mice at different developmental stages (Supplementary Material, Fig. S1B). We also analyzed elastic features of the aortae by measuring arterial diameters under continuous increases in intra-aortic pressure and found no significant difference in aortic stiffness between the wild-type and Ltbp2–/– mice (Supplementary Material, Fig. S1C). On the basis of these results, we concluded that the lack of LTBP-2 does not lead to major abnormalities of elastic fibers in elastin-rich tissues in mice.
Null mutations in Ltbp2 does not result in congenital glaucoma in mice
Although LTBP2 mutations have been reported to cause congenital ocular diseases including glaucoma in humans (1–6,18), it remains unclear whether glaucoma is the primary phenotype or a secondary symptom owing to other congenital diseases such as ectopia lentis. These reports prompted us to investigate whether Ltbp2-mutant mice had any phenotypes related to glaucoma. Anterior chamber angle closure is one of the morphological abnormalities linked to high IOP, which leads to the optic nerve damage found in glaucoma. By histological examination, we found that the angles of Ltbp2–/– mouse eyes were not closed compared with those of wild-type eyes (Supplementary Material, Fig. S2A). Transmission electron microscopy showed no evident structural abnormalities in the trabecular meshwork and Schlemm's canal in Ltbp2–/– eyes (Supplementary Material, Fig. S2B). We measured the IOP of wild-type and Ltbp2–/– mice by placing a glass needle into the eyes under anesthesia in vivo and found no significant increase of IOP in Ltbp2–/– mouse eyes at both young and old ages (Supplementary Material, Fig. S2C). These results suggested that Ltbp2–/– mice developed none of the typical phenotypes of glaucoma that involved an increase in IOP. It is often found that human patients develop normal-tension glaucoma, wherein optic nerves are damaged even though the IOP is within normal range. Therefore, we tested whether there were any neuronal damages in Ltbp2–/– mouse eyes by measuring the density of the optic nerve and retinal ganglion cells; however, we found no difference on comparing with wild-type mouse eyes (Supplementary Material, Fig. S2D and E). Finally, we examined the visual function of the Ltbp2–/– mice by observing the retinotopic map in the visual cortex of the mice using optical imaging of intrinsic signals (Supplementary Material, Figs S3 and S4). The retinotopic maps in the wild-type and Ltbp2–/– mice appeared to cover a wide range of the visual field (–30°–20° in azimuth), and we found no noticeable difference between them (Supplementary Material, Fig. S3B). Quantitative analysis of the cortical representation of the visual field also showed no significant difference between the wild-type and Ltbp2–/– mice (Supplementary Material, Fig. S3C). Therefore, we concluded that Ltbp2–/– mice do not have a visual field deficit in the horizontal plane. We also analyzed the visual acuity of Ltbp2–/– mice and found no significant difference compared with the acuity of wild-type mice (Supplementary Material, Fig. S4). These results indicated that Ltbp2–/– mice have no deficit in detecting projected images on the retina and in transmitting the information to the visual cortex of the brain.
Ltbp2 mutations cause ectopia lentis because of ciliary zonule fragmentation
LTBP2 mutations have also been reported in humans with ectopia lentis (3,4,6,18). Therefore, we next analyzed phenotypes related to lens dislocation in Ltbp2–/– mice. Indeed, the mutant mice developed lens dislocation into the posterior chamber of the eyes (Fig. 2A and B). Ciliary zonules act as ligaments anchoring the lens to the ciliary process and are made of bundles of fibrillin microfibrils (3,19). Immunofluorescent staining, immunotransmission electron microscopy (immuno-TEM) and western blot analyses revealed the presence of LTBP-2 in ciliary bodies, including the ciliary epithelial cells and ciliary zonules, in wild-type mouse eyes (Fig. 2C, Supplementary Material, Figs S5 and S6A and C). Immuno-TEM of negatively stained ciliary zonules revealed that LTBP-2 is localized to the junction of microfibrils and connecting filaments, a structure previously reported in ciliary zonules (20) (Supplementary Material, Fig. S6D and E). The fibrous structure of ciliary zonules, which is visualized by immunohistochemical examination using anti-fibrillin-1 antibody, was present in wild-type mouse eyes, whereas the structure was disorganized in Ltbp2-mutant mouse eyes at early stages after birth (Supplementary Material, Fig. S6B) and largely missing in later stages (Fig. 2C), even though the expression levels of fibrillin-1 were similar in the eyes of both genotypes (Supplementary Material, Fig. S5A). Electron microscopy also detected microfibrils as thick bundled structures on the surface of the ciliary epithelial cells of wild-type mouse eyes, whereas no bundled structures of microfibrils were observed on the ciliary epithelial cells of Ltbp2–/– mouse eyes, leaving only fragmented remnants of microfibrils (Fig. 2D). These results indicated that LTBP-2 has an essential role in forming the thick bundled structure of microfibrils in ciliary zonules. mRNAs for Ltbp1 and 3 were expressed in the ciliary body, suggesting that these could not compensate for the absence of LTBP-2 (Supplementary Material, Fig. S5D).
LTBP-2 is required for microfibril assembly on ciliary epithelial cells
From these results, we expected that LTBP-2 functioned to construct the microfibrils of ciliary zonules. To test this hypothesis, we cultured human non-pigmented ciliary epithelial cells and analyzed microfibril formation in the presence or absence of LTBP-2. The primary cultured ciliary epithelial cells produced microfibrils that were visualized by the anti-fibrillin-1 antibody on the surface of the cells (Fig. 3A). We found that LTBP-2 co-localized to fibrillin-1, showing a fibrous pattern on the surface of the cells. When we suppressed the expression of LTBP2 mRNA by siRNA, we found that fibrillin-1-positive microfibrils were markedly decreased on the LTBP2 knockdown cells (Fig. 3A), although the siRNA suppressed the expression of LTBP-2 but not that of fibrillin-1 (Fig. 3B and C). We then supplemented the culture medium of ciliary epithelial cells transfected with LTBP2 siRNA with recombinant LTBP-2 protein and observed the restoration of microfibril formation on the cells; this excluded the possibility that the decrease in microfibrils by LTBP2 siRNA was an off-target effect (Fig. 3A). The added LTBP-2 protein co-localized with fibrillin-1 on the microfibril meshwork. The requirement of LTBP-2 in microfibril formation seems to be cell type-specific, because suppression of LTBP2 mRNA by siRNA in human dermal fibroblast did not affect the development of microfibril meshwork (Supplementary Material, Fig. S7). Through an organ culture experiment ex vivo, we further tested whether supplementation of recombinant LTBP-2 was effective in restoring ciliary zonules in Ltbp2-mutant mouse eyes. We cultured the anterior cup of eye from P5 animals in culture medium for 7 days with or without recombinant LTBP-2 supplementation and observed microfibril formation in the ciliary body by electron microscopy. Only fragmented microfibrils were detected on the ciliary epithelial cells of Ltbp2-mutant mouse eyes without LTBP-2 supplementation; however, the production of bundled unfragmented microfibrils was observed in mutant eyes that were cultured in the presence of LTBP-2 (Fig. 3D). These results suggested that LTBP-2 promotes the assembly of mature microfibrils in ciliary zonules.
Pathological LTBP2 mutations cause loss of secretion and lack of fibrillin-1 binding activity
It has been previously reported that LTBP-2 directly binds to fibrillin-1, one of the main components of microfibrils (11). Several groups have determined that missense, nonsense and frame-shift mutations in LTBP2 in humans are responsible for familial eye diseases such as glaucoma and microspherophakia (Supplementary Material, Table S2) (1–6). We expected that these mutations would cause a loss of function of LTBP2, resulting in malformation of microfibrils in the eyes; therefore, we tested the impact of mutations in LTBP-2 for secretion and binding to fibrillin-1. We first narrowed down the critical domain for fibrillin-1 binding on LTBP-2. Because intact LTBP-2 binds to the N-terminal domain of fibrillin-1 (11), we expressed the recombinant N-terminal domain of fibrillin-1 (rF23) (21) with a myc-tag in HEK293T cells and recovered the protein in cultured medium to use for LTBP-2 binding. For LTBP-2, we separated the protein into five domains and expressed each domain with a FLAG tag into the culture medium of HEK293T cells (Fig. 4A). We then mixed both the culture media containing myc-tagged fibrillin-1 fragment and FLAG-tagged LTBP-2 fragments and immunoprecipitated LTBP-2 fragments by anti-FLAG antibody beads to find out which LTBP-2 domain binds to fibrillin-1. We found that LTBP-2 binds to fibrillin-1 at the latter half of the calcium-binding EGF (cbEGF)-like repeat domain (LTBP-2-D; Fig. 4B). When we looked at the reported human LTBP2 mutations, we found that most of the LTBP-2 mutants did not have this fibrillin-1-binding domain because of premature termination (Supplementary Material, Table S2); however, we noticed three frame-shift mutants and one missense mutant containing the fibrillin-1-binding domain. To test whether these mutants had the ability to bind to fibrillin-1, we prepared expression vectors for these mutants by site-directed mutagenesis. As a control, we also prepared expression vectors for LTBP-2 variants that contained single amino acid replacements caused by single-nucleotide polymorphisms (SNPs) deposited at Genbank (Supplementary Material, Table S3). Prepared expression vectors were individually transfected into HEK293T cells, and the culture medium from these transfected cells was collected for immunoprecipitation with fibrillin-1 fragments. However, when we tested for the presence of recombinant proteins in the cultured medium, we found that none of the LTBP-2 mutants were efficiently secreted in the medium, although the proteins were detected in cell lysates (Fig. 5A). Normal and polymorphic LTBP-2 variants were all detected in both the cell lysate and cultured medium, indicating that the failure of secretion is specific to mutant proteins. Because all the LTBP-2 mutants were produced in the cells, we prepared cell lysates and analyzed the fibrillin-1 binding activity of the mutant proteins. Among the four LTBP-2 mutants, three mutants had no binding activity to fibrillin-1, whereas one missense mutant LTBP-2 (p.C1438Y) showed fibrillin-1 binding activity equivalent to that of normal and polymorphic LTBP-2 variants (Fig. 5B). These data suggest that pathogenic mutations cause structural alterations in the proteins, leading to secretion failure, inactivation of fibrillin-1 binding activity or both.
LTBP-2 is the only protein that does not bind to TGF-β in the LTBP family; therefore, it is suspected to have a biological function other than TGF-β signal transduction. Shipley et al. reported on the early embryonic lethality of systemic homozygous Ltbp2 knockout mice (14), suggesting an essential function of the protein at the embryonic stage in mice, although the reason of the lethality was not identified. On the other hand, it has been reported by several groups that human homozygous mutations of LTBP2 cause ocular diseases including primary or secondary glaucoma, megalocornea, microspherophakia and ectopia lentis but no developmental defects in organs other than eyes (1–6). These findings indicate a sharp contrast of phenotypes between mice and humans, because many of the human LTBP2 mutations were nonsense mutations at the N-terminal region and considered equivalent to null mutations. In the process of generating conditional Ltbp2 knockout mice, we generated systemic Ltbp2 homozygous mutant mice by deleting the first exon containing the starting ATG codon. Unexpectedly, the homozygous mutants showed no outcomes such as embryonic lethality, gross developmental abnormalities or impaired fertility. Because we deleted the pgk-Neo cassette (Fig. 1A), the artificial effect of the exogenous sequences was minimal in our study compared with that in the original Ltbp2-mutant mice analysis. We assume that the Ltbp2-mutant mice created in the previous report may have had an unexpected disruption of a critical gene(s) in addition to Ltbp2 or that the remaining artificial DNA sequences, including the pgk-Neo cassette, may have affected the expression of the neighboring genes that were required for embryonic development.
We extensively analyzed any detectable ocular abnormalities and visual function in the Ltbp2-mutant mice and found that the mutant mice exhibited severely fragmented ciliary zonules in the eye without any sign of glaucoma. Homozygous mutations in LTBP2 have been originally reported to cause glaucoma in humans (1,2); however, some of the following reports have also described LTBP2 mutations for other eye diseases such as megalocornea, microspherophakia and Weill–Marchesani syndrome (3–6,18). Currently, researchers disagree as to whether glaucoma is a primary symptom or a secondary effect of another abnormality caused by LTBP2 mutations. Interestingly, the reported symptoms for all patients with LTBP2 mutations include zonule weakness and lens dislocation, suggesting a malformation of ciliary zonules as the primary defect caused by LTBP2 mutations. Indeed, lens dislocation is often associated with secondary glaucoma, as observed in Marfan syndrome and Weill–Marchesani syndrome (22). Increasing number of evidences suggest association of microfibril deficiency with glaucoma (23). One possible explanation is that moving lens causes adhesion of iridocorneal angle by pushing the iris forward. Another possibility is that fragmented zonule microfibrils occlude the trabecular meshwork or Schlemm's canal, resulting in high IOP. We therefore assume that the LTBP2 mutations do not cause PCG; however, some patients with the mutations develop symptoms of glaucoma at an early stage of life as a secondary effect of lens dislocation and/or fragmentation of ciliary zonules. Screening of gene mutations in the United States, United Kingdom, India and Saudi families and sporadic cases of PGC did not find LTBP2 mutations (24–27), possibly because these PCG subjects were all primary and had no other ocular manifestations, including lens dislocation. Our mice data are in accordance with these clinical investigations.
We observed that Ltbp2-mutant mice developed lens luxation because of abnormal ciliary zonule formation (Fig. 2, Supplementary Material, Figs S6A and S7). Through electron microscopic observation, thick and continuous microfibrils were barely detected in the mutant mice, indicating that LTBP-2 is involved in the formation of mature ciliary zonules made of bundled microfibrils (Fig. 2D). Because LTBP-2 binds to fibrillin-1, which is the major component of microfibrils, LTBP-2 may have a role in the development of mature bundles of microfibrils so that ciliary zonules obtain mechanical strength. Microfibrils consist of fibrillins, particularly fibrillin-1 in case of ciliary zonule microfibrils, with the parallel head-to-tail alignment (28); this occurs without a major structural contribution of other extracellular matrix proteins. LTBP-2 may not be directly involved in the formation of the fibrillin polymer; instead, the protein may be associated with the fibrillin polymer to form larger bundles of fibers. By proteomics analysis, Cain et al. reported that extensively purified zonular microfibrils consist of fibrillin-1 as the largest component, with a small contribution from microfibril-associated glycoprotein-1 (MAGP-1) as the matrix proteins (29). Although they did not detect LTBP-2 by their proteomics analysis, we detected the presence of LTBP-2 in ciliary zonules in intact eye tissue (Fig. 2C, Supplementary Material, Figs S5 and S6), indicating LTBP-2 is present but may not covalently or strongly interact with fibrillin-1 in ciliary zonules. Davis et al. reported that ciliary zonule microfibrils were laterally connected by small filaments (20). Our electron microscopic observation detected a lateral association of LTBP-2 to a microfibril (Supplementary Material, Fig. S6E), suggesting that LTBP-2 may contribute to the formation of thick microfibrils by tying fibrillin microfibrils into bundles. Further investigation is necessary to determine whether LTBP-2 constitutes the small connecting filaments.
LTBP-2 binds to fibrillin-1 at the C-terminal half of the long cbEGF-like repeat domain (Fig. 4A and B) that corresponds to exon 23 to 28 in the LTBP2 genomic structure. Most of the LTBP2 mutations found in humans are located on earlier exons of LTBP2, resulting in premature termination of the gene product. However, some of the pathogenic LTBP2 mutations are localized to the C-terminal region in the primary structure of the protein; therefore, the mutant proteins should contain the fibrillin-1-binding domain. To uncover the pathogenic mechanism of LTBP2 mutations, we produced recombinant LTBP-2 mutant proteins in cultured HEK293T cells to perform immunoprecipitation with a recombinant fibrillin-1 fragment and found that none of the mutants were efficiently secreted from the producing cells (Fig. 5A). All the LTBP-2 variant proteins that contained nonpathogenic amino acid substitutions because of SNPs were as efficiently secreted as the original LTBP-2 protein; therefore, only pathogenic mutations affected the secretion of the matrix protein, which could be caused by conformational alteration. It is interesting that although the mutations are downstream of the fibrillin-1-binding domain, all but one missense LTBP-2 mutants had no binding to fibrillin-1 by immunoprecipitation (Fig. 5B). This finding suggested that these mutations may cause such conformational changes that result in not only secretion arrest but also loss of fibrillin-1 binding activity.
The addition of recombinant LTBP-2 in a culture medium of human ciliary epithelial cells transfected with LTBP2 siRNA restored the microfibril meshwork formation in the extracellular matrix of the cells (Fig. 3A). The recombinant protein also increased the production of stable microfibrils in organ-cultured eye tissues from both wild-type and Ltbp2-mutant mice (Fig. 3D), suggesting an essential function of LTBP-2 in the construction of ciliary zonules. Ltbp2-deficient mice, however, did not show an apparent abnormality in the matrix of other tissues such as the lung, aorta, and skin (Supplementary Material, Fig. S1), although Ltbp2 mRNA was abundantly expressed in these tissues in wild-type mice (14). We assume that stable microfibril formation in ciliary zonules critically depends on LTBP-2; however, other matrix proteins may compensate for the lack of LTBP-2 in the other tissues. There are 3 other LTBP family members both in mice and human. Among these, Ltbp1 and 3 were expressed in the ciliary body (Supplementary Material, Fig. S5D), suggesting that LTBP-1 and 3 are not sufficient to compensate for the lack of LTBP-2. Mutant mice lacking an isoform of LTBP-4 develop pulmonary emphysema as well as cardiomyopathy associated with disrupted elastic fiber formation (30). Further studies using double-mutant mice deficient in Ltbp2 and Ltbp4 are required to address the question whether LTBP-2 has a cryptic function in tissues other than ciliary zonules.
In summary, we demonstrated the essential function of LTBP-2 in the formation of ciliary zonules, which consist of microfibrils. Homozygous mutant mice lacking LTBP-2 exhibited lens dislocation because of the disruption of normal ciliary zonule formation, thus resembling some symptoms of human patients with LTBP2 homozygous mutations. Further studies including Ltbp2-mutant mice and gene mutations related to ectopia lentis in humans will lead to understanding of the microfibril organization as well as the pathology of glaucoma.
MATERIALS AND METHODS
All mice used in this study were maintained on normal lab diet. All procedures were conducted according to the Guideline for Animal experimentation at Kansai Medical University and Tottori University.
The human cadaveric eye was obtained from San Diego Eye Bank. A written consent was acquired from the donor regarding eye donation for research purpose. Tissue procurement was adherent with the Declaration of Helsinki. The globe was initially cut into three parts and homogenized in 8 M urea using a Polytron homogenizer (PT10–35, Kinematica AG). After centrifugation, the supernatants were subjected to western blotting.
Generation of Ltbp2 knockout mice
The coding sequence in the exon 1 of mouse Ltbp2 was flanked by two loxP sequences inserted in 5′ untranslated region of the exon 1 and intron 1. The pgk-Neo cassette flanked by FRT sites was removed by crossing with flippase-expressing mice (31). To generate Ltbp2 null allele, Ltbp2flox/+ mice were crossed with Ayu-1 Cre knock-in mice, which express Cre in multiple tissues including germ line (32). See Supplementary Material, Methods, for construction of the targeting vector.
Antibodies and immunodetection
Primary antibodies used were as follows: rabbit anti-LTBP-2 polyclonal (made by immunizing rabbits with recombinant mouse LTBP-2, 1:1000, Supplementary Material, Fig. S2C, Fig. 5A and B; 1:100, Fig. 1D, Supplementary Material, Fig. S6A; 1:50, Supplementary Material, Fig. S6C and E), chicken anti-LTBP-2 polyclonal (made by immunizing chickens with recombinant mouse LTBP-2, 1:100, Fig. 2C), mouse anti-LTBP-2 monoclonal 8F1 (made by immunizing mice with recombinant human LTBP-2, 1:100, Fig. 3A; 1:1000, Supplementary Material, Fig. S5C), rabbit anti-fibrillin-1 polyclonal PR217 (Elastin Products Company, 1:100, Fig. 3A), rabbit anti-mouse fibrillin-1 polyclonal (made by immunizing rabbits with recombinant mouse fibrillin-1 fragment, 1:100, Fig. 2C and Supplementary Material, Fig. S6B) and rabbit anti-γ-Synuclein polyclonal (Abcam, 1:100, Supplementary Material, Fig. S2E) antibodies. See Supplementary Material, Methods, for details.
HEK293T cells and human non-pigment ciliary epithelial cells (HNPCEC) purchased from Sciencell were maintained in DMEM (Invitrogen) supplemented with 2 mm glutamine, 100 units/100 mg/ml penicillin/streptomycin and 10% FBS at 37°C in 5% CO2. To develop fibrillin microfibrils, confluent HNPCECs were cultured on coverslips in DMEM/F12 (Invitrogen) supplemented with 2 mm glutamine, 100 units/100 mg/ml penicillin/streptomycin and 10% FBS with or without 15 µg/ml of recombinant LTBP-2 at 37°C in 5% CO2 for 7 days without changing the medium.
Human full-length LTBP-2 cDNA was kindly provided by J Keski-Oja (University of Helsinki). pEF6/ssFLAG, which contains a preprotrypsin signal sequence followed by a FLAG tag and 6×His tag before multiple cloning sites, was used to add an N-terminal FLAG tag (33). pEF6/FLAG or pEF6/Myc were used to add a C-terminal FLAG tag or a C-terminal Myc tag (16). FLAG-tagged human LTBP-2 fragments LTBP-2A, B, C, D and E were previously described (10). The following four mutant cDNAs and three polymorphism cDNAs of human LTBP-2 were generated by site-directed mutagenesis and subcloned into pEF6/ssFLAG: c.4700G>A (p.C1438Y), c.5242C>T (p.Q1619X), c.5763delC (p.Y1793fsX55), c.5833dupC (p.H1816PfsX28), c.5008G>A (p.E1541K), c.5351A>G (p.Y1655C), c.5737C>G (p.P1784A). All constructs were confirmed by sequencing (ABI Prism 3100). An N-terminal fragment of human fibrillin-1 rF23 cDNA was kindly provided by Lynn Sakai (Shriners Hospital for Children) and subcloned into pEF6/Myc and pEF6/ssFLAG.
Recombinant human and mouse LTBP-2 with a FLAG tag and a 6×His tag was purified using TALON affinity resin (Takara) from serum-free conditioned medium of HEK293T cells stably transfected with pEF6/ssFLAG-human LTBP-2 or pEF6/FLAG-mouse LTBP-2.
Transfection, in vitro binding assay and western blotting
HEK293T cells were transfected using LipofectAMINE PLUS (Invitrogen) or FuGENE HD Transfection Reagent (Promega). After transfection, they were cultured in serum-free DMEM/F12 (Invitrogen). The conditioned media were subjected to immunoprecipitation with anti-FLAG M2 affinity gel followed by western blotting as described previously (33).
Duplex RNA oligonucleotides (Stealth Select RNAi) and control oligonucleotides (Stealth RNAi™ siRNA negative control) were purchased from Invitrogen. Mixed siRNA duplexes (at final concentration 10 nm) were reverse-transfected into HNPCECs using Lipofectamine RNAi MAX (Invitrogen). The oligonucleotides sequences of LTBP2 siRNAs are provided in Supplementary Material, Table S6.
T.I., T.O., T.O.A. and T.N. designed the study; T.O., Y.F., H.Y., K.N., M.H., K.K., Y.H. and T.N. performed the experiments; T.I., T.O., Y.F. and T.N. collected and analyzed the data; K.T. and K.K. provided the reagents; T.I., T.O.A. and T.N. wrote the manuscript.
This work was supported by NEXT Program in Japan (LS120 to T.N.), Japan Society for the Promotion of Science (26670147 to T.N., 25462741 to T.O.A.), Japan Science and Technology Agency (08090019 to T.N.), NIH grant R01 (EY014620 to T.O.A), The Vehicle Racing Commemorative Foundation (T.N.) and Takeda Science Foundation (T.N.).
We thank Ms A. Kimura and Ms M. Masuda for technical assistance, and San Diego Eye Bank for providing human eye tissues.
Conflict of Interest statement. None declared.
- extracellular matrix proteins
- bodily secretions
- culture media
- ectopia lentis
- eye diseases, hereditary
- glaucoma, congenital
- lens, crystalline
- lens (device)
- organ culture techniques
- epithelial cells
- rna, small interfering
- mutant proteins
- suspensory ligament of lens