Targeted Genomic Deletion of the Lens-Specific Intermediate Filament Protein CP49

Abstract

purpose. To deduce the function of the lens-specific cytoskeletal structure, the beaded filament, by blocking expression of the fiber cell–specific beaded filament protein CP49.

methods. The first exon of the mouse CP49 gene was deleted by using targeted genomic deletion techniques. Gene deletion was assessed through Southern blot analysis and PCR. Translation and protein expression were characterized by Northern and Western blot analysis of both CP49 and its assembly partner filensin. The architecture of knockout lenses was compared with that of wild-type lenses at the histologic level by light microscopy. Lens clarity was assessed in situ by direct ophthalmic examination and slit lamp microscopy.

results. Transcription and translation of CP49 were successfully negated in knockout animals. Lenses homozygous for the CP49 deletion showed no obvious changes in lens architecture at the light microscope level. Filensin levels were sharply reduced, although filensin mRNA levels appeared unchanged. Direct examination of lenses showed no obvious loss of lens clarity, but slit lamp examination revealed the emergence of opacification in even the youngest animals. The opacification worsened with age.

conclusions. The absence of CP49 causes a subtle loss of optical clarity in the ocular lens, a loss that worsens with age. However, CP49 is not essential for the assumption or maintenance of overall fiber cell shape or long-range order of fiber cells. CP49 appears to regulate the protein levels of its assembly partner filensin, suggesting a mechanism for the regulation of beaded filament protein stoichiometry.

The intermediate filament (IF) is a cytoskeletal structure that is essentially ubiquitous among cells of vertebrate organisms. ¹ Although these filaments are structurally similar from cell to cell, the proteins that comprise them are drawn from a relatively large gene family consisting of more than 30 members in humans. The IF proteins range in molecular weight from approximately 45 to 150 kDa, and may share as little as 15% to 20% sequence identity. All cytoplasmic IF proteins, however, share a common domain structure consisting of a central rod domain flanked by an amino head domain and carboxyl tail domain. The central rod domain is well conserved in size and predicted secondary structure, whereas the head and tail domains show considerable variability in both size and sequence. 

Which IF protein(s) is used in the construction of an IF in a given cell is highly regulated and varies with cell type, stage of development, and stage of differentiation. Presumably, such variation permits adaptation of a given IF to cell-specific functions. Thus, it is not surprising that the lens fiber cell expresses two IF proteins that are unique to it: CP49 and filensin. However, both proteins were the most divergent IF proteins yet identified. ^{2 3 4 5 6 7} Several features which were otherwise highly or absolutely conserved among all other cytoplasmic IF proteins were lacking in one or the other of the two fiber cell specific IF proteins. Notably, these lens IF proteins were localized to a filamentous structure called the breaded filament (BF). This represented the first example of cytoplasmic IF proteins’ being localized to a structure other than a classic IF. ^{8 9 10 11} CP49 and filensin have thus been referred to as beaded filament (BF) proteins. 

The observation that the two BF proteins were restricted in expression to the lens fiber cell, combined with their extreme degree of divergence from the remainder of the IF family, led to the suggestion that BF proteins must play a critical role in the biology of the lens fiber cell, a role that is unique to that cell. This hypothesis was consistent with reports that implicated two separate point mutations in human CP49 as causative in two families with autosomal dominant congenital cataract. ^{12 13} However, few data have been generated that might suggest a function for the BF in the lens fiber cell. 

We sought to deduce the role of the BF by targeted genomic deletion of the CP49 gene in mouse lens. We hypothesized that the BFs may play a role in the structural specializations that are unique to the lens: (1) individual lens cells undergo extreme structural differentiation as they mature from epithelial cell to fiber cell, including a transient redistribution of organelles and (2) fiber cells are arranged in very precise columns, assuming a long-range ordering of cells into a tissue. As a corollary, we hypothesized that the elimination of CP49 expression by knockout technology would create a loss-of-function mutant that could suggest a role for the BF in the biology of the lens. We report herein the characterization of the CP49 null mouse. 

Materials and Methods

Construction of the CP49 Targeting Vector and Generation of Knockout Mice

A P1 clone containing the entire CP49 gene was isolated from a mouse 129/OLa library (Incyte Genomic, St. Louis, MO) by using a PCR reaction specific for exon 1. A pKO scrambler NTK vector (Stratagene, La Jolla, CA) was used to construct a targeting vector (see Fig. 1 ) that incorporated the following: (1) an MCI promoter thymidine kinase (TK)-TK polyA signal expression cassette of pKO in the same orientation as the CP49 sequence, (2) a 4.5-kb HindIII restriction fragment extending from the 5′ untranslated region that ended 80 bp upstream of the ATG translational start site, (3) the polyA signal-neomycin phosphotransferase-PGK promoter expression cassette in opposite transcriptional orientation to the CP49 gene, and (4) a 6.5-kb blunted EcoRV-EcoRI restriction fragment from intron B, which started at a point 500 bp downstream of exon 1. ¹⁴  

This plasmid was linearized with NotI, and 25 μg was electroporated into 4 × 10⁷ GK129 embryonic stem cells derived from 129/OLa mice. ¹⁵ Cells were selected by double selection with the neomycin analogue G418 and the nucleoside analogue 2′-deoxy-2′-fluoro-β-d-arabinofuranosyl-5-iodouracil (FIAU). ¹⁶ Approximately 570 resistant clones were screened by Southern blot for homologous recombination at the CP49 locus. Genomic DNA was digested with BamHI and hybridized with a 500-bp genomic fragment from 3′ flanking region of a 6.5-kb fragment. Eleven correctly targeted embryonic stem (ES) cell clones were obtained, and four were expanded and further tested by Southern blot, using the 3′ flanking probe and two internal probes specific to homology fragments. These clones were karyotyped. Two correctly targeted ES clones were microinjected into C57BL/6J blastocysts and transferred to the uteri of pseudopregnant CD1 female mice (University of California Davis Targeted Genomics Laboratory). Chimeric animals were backcrossed onto a C57BL/6J background to screen for germline transmission. Heterozygous mice were interbred to obtain homozygous mice that were CP49 knockouts. To genotype the resultant offspring, mouse genomic DNA was used as a template for PCR with primers specific for the wild-type allele and the targeted allele. Primers used to identify the CP49 wild-type allele were specific to exon 1: 5′ AAGAGGAGAGTGGCAGCGGACTTG; 3′ GAGCCCGCAGTTGTGTTTCCAGTT. These resulted in amplification of a 450-bp fragment of exon 1. To identify the disrupted allele, primers were used that amplified a 550 bp fragment of the neomycin gene cassette (Neo): 5′ GCCGCCAAGCTCTTCAGCAATATC; 3′ TGCCCTGAATGAACTGCAGGACGA. To identify the disrupted allele, we also used a primer that flanked the Neo site, and a primer that flanked the 4.5-kb HindIII fragment. These primers amplified a 500-bp fragment: 5′ CTGGCTGCATAAGGATTTTGAGGC; 3′ TGCCACTCCCACTGTCCTTTCCTA. 

Unless otherwise indicated, all subsequent characterizations were performed on the offspring of each of two separate clones. Similarly, comparisons of wild-type, heterozygous, and knockout animals were conducted on litter mates unless otherwise stated. All procedures conformed to the provisions of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

Northern Blot Analysis

Total RNA was isolated from 4- to 6-week-old litters from heterozygous–heterozygous breeding, using a single extraction with an acid guanidinium thiocyanate-phenol-chloroform mixture. ¹⁷ Total RNA was calculated by spectroscopy. Ten micrograms of total RNA was electrophoresed in formaldehyde-agarose gels and transferred to membrane (Immobilon-Ny+; Millipore, Bedford, MA), according to standard procedures. ¹⁴ After UV cross-linking, the blots were probed with ³²P-labeled cDNAs for CP49 (full-length cDNA probe), or CP115 (partial cDNA probe), washed, exposed, and developed according to standard procedures. Resultant blots were normalized against a γS-crystallin cDNA probe. 

SDS PAGE-Western Blot Analysis

Whole lenses were solubilized in standard SDS gel denaturing cocktail, and identical volumes were compared on 12.5% SDS polyacrylamide gels. For Western blot analysis, samples were transferred to membranes overnight (Immobilon P; Millipore), rinsed in Tris-buffered saline (TBS)–Tween, and blocked in TBS-Tween containing 5% normal goat serum and 2% powdered milk. Rabbit antiserum to recombinant bovine filensin, recombinant mouse CP49, and recombinant human vimentin were used to probe Western blots of wild-type, heterozygous, and knockout lenses. ^{4 5 18}  

Histology

Whole eyes were immersed in cold, phosphate-buffered, 4% paraformaldehyde overnight, and processed into glycolmethacrylate according to the manufacturer’s recommendations (Polysciences, Warrington, PA), but with extended infiltration times. One-micrometer sections were stained with toluidine blue. 

In Situ Examination of Lenses

All mice were examined using slit-lamp ophthalmoscope without anesthesia (FS-2 photograph slit lamp; Nikon, Tokyo, Japan). Mouse eyes were dilated with a drop of a 1:1 mixture of 1% tropicamide (Alcon, Fort Worth, TX) and 10% phenylephrine hydrochloride (Akorn, Abita Springs, LA). The angle of the slip lamp was approximately 40°, and the slit width was approximately 0.2 mm. Examinations were recorded on digital video (Optura Pi; Canon, Tokyo, Japan). Still images were captured (Premiere 6.0; Adobe, San Diego, CA) and processed with image-management software (Photoshop 6.0; Adobe). 

Results

Figure 1a presents a schematic of the targeted CP49 locus and the targeting vector. The short and long arms of the targeting vector were taken from the 5′ untranslated region and the first intron, respectively. In the targeting vector, these regions flanked an insert that included the neomycin resistance gene in opposite orientation. A unique BamHI site was included in the targeting vector to facilitate screening of ES cell clones. This BamHI site converted a ∼12 kb restriction fragment in the wild-type locus to ∼2.5- and ∼9.5-kb fragments in the null mouse. Positive-negative selection using thymidine kinase and neomycin resulted in many clones. Five hundred seventy-six resistant clones were selected and screened by Southern blot analysis (Fig. 1b) , taking advantage of the change in restriction fragment size due to the BamHI restriction site introduced by the targeting vector. 

Eleven clones were identified, and four were expanded for further characterization. Two ES clones were injected into C57BL/6J mice. Resultant chimeras were backcrossed, and resultant heterozygotes bred to obtain mice homozygous for the modified CP49 locus. Southern blot analysis (Fig. 1c) and PCR (Fig. 1d) were used to verify correct modification of the CP49 locus in the resultant offspring. 

Northern blot analysis was used to verify that the gene targeting effectively negated transcription (Fig. 2a) . γS-crystallin was used as a positive control and standard for RNA levels for both the CP49 and filensin Northern blot analysis. Total lens RNA from wild-type (+/+), heterozygous (+/−), and knockout (−/−) lenses were probed with full-length CP49 cDNA probes. Wild-type lenses yielded a positive signal of the appropriate size, as did heterozygous lenses. The somewhat reduced signal in the heterozygote derives, at least in part, from overall lower levels of RNA, determined by probing with γS-crystallin. However, the possibility that total message level is reduced in the heterozygote cannot be ruled out. No hybridization occurred in the lane containing RNA from the knockout animals, indicating the absence of transcripts containing CP49 coding sequence. 

Northern blot analysis (Fig. 2a) was also performed using filensin cDNA probes as a second internal standard to verify the integrity of RNA isolated from knockout lenses, but also to determine the impact of CP49 negation on transcription of filensin. Hybridization appeared comparable in both knockout and wild-type lenses, suggesting no significant impact on the steady state levels of filensin mRNA. Again, reduced levels of signal in the heterozygote resulted, at least in part, from lower levels of total RNA when normalized to γS-crystallin internal controls. 

Western blot analysis was performed to examine the impact of locus disruption on the accumulation of CP49 protein. Figure 2b , left, shows Coomassie blue-stained SDS PAGE profiles of wild-type (+/+), heterozygous (+/−), and knockout (−/−) lenses. The approximate positions of filensin (95 kDa) and CP49 (49 kDa) are indicated. The contribution of the CP49 to total lens protein is low enough that it is not evident in gels stained for total lens protein, whereas the lower-molecular-mass crystallins dominate the gel profile. Figure 2b , right, shows comparably loaded samples probed by Western blot analysis using anti-CP49 antiserum. The intense reactivity of the CP49 can be seen in the Western blots of wild-type and heterozygous lenses but no signal could be detected in the lenses of knockout animals, confirming that CP49 protein expression had been eliminated. If there was a reduction in total CP49 protein in the heterozygote, it was not dramatic. 

Figure 2b also confirms that the lower-molecular-mass bands that are commonly observed in Western blots with the CP49 antibody and that have been observed even with monoclonal antibodies, were derived from CP49. The sequence similarity between CP49 and other IF proteins leaves open the possibility that antiserum or monoclonal antibodies to CP49 may cross-react with other IF proteins, some of which have been reported in the lens. 

CP49 has been colocalized with filensin to the beaded filament of the lens, and together they are thought to comprise the backbone of this cytoskeletal structure. ^{8 9 10 11 19} Thus, we were interested in knowing the fate of filensin in the absence of an assembly partner. Coomassie blue–stained total lens extracts and filensin Western blots were prepared from several litters, totaling more than 72 animals. The results generated from one litter are shown in Figure 3a . In all cases, the total level of filensin immunoreactivity was comparable between wild-type and heterozygous lenses, but was sharply reduced in knockout lenses. To assess the approximate degree of filensin reduction in the knockout lenses, we prepared serial dilutions of wild-type whole lens extracts and compared them with the stock knockout lens extract (Fig. 3b) . Only at dilutions in the range of 1:64 to 1:128 did the signal in the wild-type lens approximate that of the knockout lens, suggesting that filensin levels are reduced in the knockout lenses on the order of 100-fold. In comparing the knockout and the more dilute wild-type lanes, it can be seen that the pattern of filensin degradation in the knockout animal is also different from that of the wild type. The bottom arrowhead highlights a breakdown product that was prominent in the wild type but absent from the knockout. 

The impact of CP49 deletion on the histology of the lens was assessed by light microscopy on sections of whole lenses from wild-type and knockout litter mates. Figure 4 presents images taken from 3-week (Fig. 4a) and 1-month (Fig. 4b) knockout lenses cut in orthogonal planes. These images show that CP49 deletion did not noticeably alter the long-range organization of lens fibers, the regularity of their hexagonal profiles, or the placement of the fiber cell nucleus. Although this histologic characterization does not rule out more subtle changes, it suggests that the remarkable organization of lens fiber cells is not critically dependent on CP49. 

Both direct ophthalmic examination and slit lamp microscopy were performed on lenses of homozygous null, heterozygous, and wild-type animals (Fig. 5) . Twenty-three CP49 null mice, ranging in age from 1 month to 10 months, were examined. Every CP49 knockout lens showed opacification, starting with a mild loss of clarity in the youngest animals (Fig. 5 , middle) that progressively worsened with age (Fig. 5 , left). The loss of clarity was subtle and not immediately obvious during direct ophthalmic examination or by transillumination. Lenses of wild-type animals were usually clear (Fig. 5 , right), although minor opacification was observed in some older animals. Lenses of heterozygotes were indistinguishable from wild-type lenses, which is consistent with the similarity in expression of CP49 in heterozygous and wild-type lenses. 

Discussion

We report the targeted deletion of the first exon of the mouse lens-specific intermediate filament protein CP49, which resulted in the complete absence of both CP49 mRNA and protein. Histologic examination of the knockout lenses showed no marked changes in the size of lenses, the architecture of the fiber cell, the long range order of fiber cells across the lens, or the positioning and ultimate denucleation of lens fiber cells with differentiation. Thus, CP49 does not appear to be critical to the assumption of lens fiber cell shape, long-range order among fiber cells, or nuclear positioning, at least as assessed by light microscopy. More subtle structural changes may emerge with electron microscopic examination (manuscript in progress). 

Elimination of CP49 resulted in sharply reduced levels of filensin, the assembly partner of CP49, but not in filensin mRNA. This suggests a mechanism by which the stoichiometry of BF protein levels may be achieved. Filensin, which is not soluble in physiologic solution, is stabilized by coassembly with CP49. In the absence of such stabilization, filensin is targeted for degradation, presumably by the ubiquitin-dependent pathway. This form of posttranslational regulation may be uniquely beneficial to lens, as accumulated insoluble protein is likely to be light scattering. A similar mechanism has been suggested for posttranslational regulation of cytokeratin pairs in the epidermis. ²⁰  

We present results of both Northern and Western blots of CP49 and filensin in wild-type, heterozygous, and knockout lenses. None of these data suggest a reduction in either CP49 or filensin levels in the heterozygous animals. However, the multistage processing required for both techniques makes detection of twofold differences a relatively difficult endeavor. Thus, although it is evident that there is no dramatic reduction of CP49 expression in the heterozygote, we cannot rule out a more subtle reduction. 

Slit lamp examination revealed light scattering in each of the knockout lenses examined, a scattering that worsened with age. The opacification was subtle, not evident by direct ophthalmic examination, nor even by low-angle slit lamp examination. Thus, the loss of CP49 is not catastrophic to the lens clarity but is important in achieving optimal clarity. The progressive worsening of light scattering with age that was seen in the knockout lenses suggests that CP49/beaded filaments may enhance the capacity of the lens to resist age-dependent opacification. Alternatively, the subtle light scattering may result from the presence of low levels of filensin, which is insoluble in physiologic solution. It is interesting to speculate that the effect of abnormal CP49 expression in human lenses could have the same subtle effect on lens transparency. The loss of more than one element of the lens cytoskeleton may be necessary for complete opacification. This hypothesis will be directly tested by targeted deletion of filensin. 

It is worth noting that point mutations in the human CP49 appear to produce a greater negative impact on optical clarity ^{12 13} than the deletion of the gene product in mice. Several factors may account for this difference: (1) At the time of examination the human lenses were far older than the mice examined in this study; (2) point mutations may result in failed assembly, a result that may produce light-scattering accumulation of improperly folded or assembled product. From an optical standpoint, this may be worse than complete absence; and (3) mouse lenses do not accommodate and thus may not be exposed to a stress that exacerbates the loss of function. 

Although in this study CP49 and beaded filaments provided a measurable improvement of optical quality in lenses, the mechanism by which this was achieved remains unknown. Several candidate functions that the intermediate filament literature may suggest that for the beaded filament, functions such as elongation, maintenance of fiber cell architecture, maintenance of long-range order of fiber cells, and successful denucleation of fiber cells, clearly do not require CP49 and/or the beaded filament. More subtle roles for CP49 and beaded filaments, roles in which such processes are enhanced, stabilized, or extended (i.e., in which a slight competitive advantage is conferred by their presence) are harder to identify, and will emerge only over time with larger numbers of animals. These studies are under construction. 

 Supported by National Eye Institute Grants RO1 EY08747 (PGF), P30 EY12576 (UCD), and RO1 NEI EY04542 (JIC), and the Office of Research, School of Medicine, University of California, Davis.

 Submitted for publication April 19, 2002; revised July 9, 2002; accepted July 29, 2002.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

(a) The CP49 locus, targeting vector, and predicted structure of the disrupted locus. Restriction sites are indicated as H (HindIII); B (BamHI); EI (EcoRI); EV (EcoRV). The location of the probe used for Southern blot analysis (b, c) is indicated as “3′ probe.” Open arrows: PCR primers used in genotyping. (b) Southern blot analysis of ES clones after BamHI digestion, and the 3′ probe indicated in (a). Homologous recombination of the construct introduces a BamHI site in the targeted allele, resulting in the conversion of a ∼12-kb restriction fragment in the wild type (+/+) to a ∼ 9-kb fragment in the disrupted allele (+/−). (c) Southern blot analysis of resultant offspring using the 3′ probe, contrasting wild-type (+/+), heterozygotes (+/−), and homozygous null (−/−). (d) PCR characterization of the targeted locus, establishing the presence of the neo gene and the absence of exon 1.

Figure 2.

 

(a) Northern blot of wild-type (+/+), heterozygous (+/−), and knockout (−/−) mice. CP49 mRNA was present in both wild-type and heterozygous mice, and was absent in knockout mice. γS-crystallin was used as a control to assess gel loading levels, and established that the heterozygote lane was somewhat underloaded, compared with the wild-type and knockout lanes. Northern blot analysis with probes for filensin established that filensin mRNA levels were not noticeably altered in the knockout lenses. (b) SDS PAGE and Western blot analysis of wild-type (+/+), heterozygous (±), and knockout (−/−) lenses. CP49 protein levels were comparable in the wild-type and heterozygous animals, but no CP49 could be detected in the knockout animals.

Figure 3.

 

(a) Whole lenses from each animal (375–383) of a litter resulting from the breeding of two heterozygotes, were solubilized in SDS-PAGE lysis buffer, resolved by SDS-PAGE, and stained for proteins with Coomassie blue (top). Genotypes were established by PCR. Identical samples were resolved, transferred, and probed by Western blot analysis with antiserum to recombinant filensin. Wild-type and heterozygous lenses showed indistinguishable levels of filensin, whereas knockout lenses showed sharply reduced levels. (b) Serial dilutions of sample 376 (a) were resolved and compared with a stock sample of 377 to estimate the level of filensin reduction that occurs in knockout lenses.

Figure 4.

 

One-micrometer methylmethacrylate sections of 3-week (a) and 1 month (b) knockout lenses. (a) Bow region sectioned in a plane approximately parallel to the long axis of the fiber cells. (b) A plane of section perpendicular to the long axis of fiber cells, from somewhat deeper in the lens.

Figure 5.

 

Slit lamp views of 2-month-old (middle) and (right) 10-month-old CP49 null mice are compared with wild-type (left) 2-month-old lenses. In the slit lamp view, the bright line on the left (anterior) is the cornea, and the aqueous chamber appears as a dark band directly to the right (posterior) of the cornea. The lens capsule and epithelium appear as the next bright band that is anterior to the transparent fiber cells in the wild-type lens. In CP49 null mice, layers of cells became increasingly opaque at 2 (middle) and 10 (left) months and the opaque layers surrounded the dense nuclear opacity. Wild-type lenses remained clear with increasing age. Heterozygous lenses (not shown) were indistinguishable from wild-type lenses. The bright spot in the center of each image, just posterior to the epithelium, is a reflection of the illumination of the slit lamp.

The authors thank Janet Lee for excellent assistance with the generation of the knockout mice.