Lens Crystallin Modifications and Cataract in Transgenic Mice Overexpressing Acylpeptide Hydrolase*

Background: Acylpeptide hydrolase (APH) in the lens may have an integral role in cataract formation. Results: Transgenic overexpression of APH results in crystallin cleavage, impaired lens development, and cataract. Conclusion: APH may be involved in the generation of peptides that have the potential to induce protein aggregation. Significance: The transgenic APH mouse model could help us understand the role of crystallin fragments in cataractogenesis. The accumulation of crystallin fragments in vivo and their subsequent interaction with crystallins are responsible, in part, for protein aggregation in cataracts. Transgenic mice overexpressing acylpeptide hydrolase (APH) specifically in the lens were prepared to test the role of protease in the generation and accumulation of peptides. Cataract development was seen at various postnatal days in the majority of mice expressing active APH (wt-APH). Cataract onset and severity of the cataracts correlated with the APH protein levels. Lens opacity occurred when APH protein levels were >2.6% of the total lens protein and the specific activity, assayed using Ac-Ala-p-nitroanilide substrate, was >1 unit. Transgenic mice carrying inactive APH (mt-APH) did not develop cataract. Cataract development also correlated with N-terminal cleavage of the APH to generate a 57-kDa protein, along with an increased accumulation of low molecular weight (LMW) peptides, similar to those found in aging human and cataract lenses. Nontransgenic mouse lens proteins incubated with purified wt-APH in vitro resulted in a >20% increase in LMW peptides. Crystallin modifications and cleavage were quite dramatic in transgenic mouse lenses with mature cataract. Affected lenses showed capsule rupture at the posterior pole, with expulsion of the lens nucleus and degenerating fiber cells. Our study suggests that the cleaved APH fragment might exert catalytic activity against crystallins, resulting in the accumulation of distinct LMW peptides that promote protein aggregation in lenses expressing wt-APH. The APH transgenic model we developed will enable in vivo testing of the roles of crystallin fragments in protein aggregation.

2). These modifications, individually or collectively, lead to increased crystallin aggregation and precipitation. Lens proteases are involved in morphogenesis and in maintaining the clarity of the lens by degrading oxidized and modified proteins and peptides (3)(4)(5)(6). Several proteolytic enzymes have been shown to play a role in lens aging and cataract formation (7)(8)(9)(10)(11)(12)(13). Human cataractous lenses have increased proteolysis and protein aggregation as compared with age-matched noncataractous lenses (14,15). During aging, the total proteolytic activity of the human lens diminishes, and the accumulation of crystallin fragments increases (16 -20). Certain low molecular weight crystallin fragments that accumulate in the aging human lens interact with intact protein and induce crystallin aggregation (21,22). Recently we showed in bovine and human lenses that ␣A66 -80 peptide and its truncated forms can be generated by the action of multiple lens proteases (23). We found that some crystallin peptides possess the capability to function as anti-chaperones, thereby diminishing the amount of ␣-crystallin available to chaperone unstable proteins (16,24). Certain crystallin mutations increase the susceptibility of crystallins to proteases, and the cleaved crystallin fragments influence protein aggregation (25).
A link between cataracts and increased proteolytic activity or crystallin fragmentation has emerged from many studies of transgenic mice (26 -28). Proteolysis of crystallins is known to occur in nonhuman lenses during aging (29 -32) and in animal models of cataract (33)(34)(35)(36). Extensive studies of calpain, including Lp82 (37)(38)(39), and of transgenic mice expressing the HIV protease (40) have confirmed that proteolysis of crystallins is one of the factors in cataract formation. However, unlike in rodent cataract, lensspecific calpain Lp82 has not been found to be involved in human cataractogenesis and, because HIV protease is foreign to the lens, other proteases are responsible for the proteolysis in human agerelated cataract lenses and diabetic lenses.
In a previous study, we found that in bovine lens, the highest concentrations of water-insoluble protein and crystallin fragments were in the nucleus, where the known protease activities were minimal but measurable (7). However, among the known lens proteases, acylpeptide hydrolase (APH) 2 activity was greater than the activity of other endoproteases or endopeptidases (7). APH, also known as acylamino acid-releasing enzyme (AARE; EC 3.4.19.1), is a tetrameric enzyme protein of 300 kDa and a member of the prolyl oligopeptidase family (41). APH catalyzes the hydrolysis of N-acetylated peptide to an acyl amino acid and to a peptide with a free N terminus shortened by one amino acid residue (42). However, the enzyme cannot remove acetylated NH 2 -terminal residues from structural proteins such as crystallins (43,44). Removal of N-terminal acyl residue from crystallins is needed for the action of other proteases. We have demonstrated in earlier studies proteolytic activity in a 55-kDa truncated fragment generated in vitro by trypsin treatment of the APH. The 55-kDa truncated enzyme is also present in bovine and human lenses, and its generation is mediated through proteolytic processing (45)(46)(47). Interestingly, the 55-kDa active fragment generated in vitro unblocked ␣A-crystallin and exhibited ␣A-, ␣B-, and ␤-crystallin hydrolyzing activity (46,47).
Native APH exhibits endopeptidase activity toward glycated and oxidatively damaged proteins in animal (48) and plant cells (49). Recently APH was shown to degrade ␤-amyloid peptide, the pathogenic fragment in Alzheimer disease (50). Because APH is known to exert activity on native and modified proteins, we hypothesized that expression of this enzyme in lens will generate crystallin fragments and that the accumulation of these fragments will result in cataract formation. To test our hypothesis, we generated a transgenic mouse model that overexpresses APH specifically in the lens, using a modified ␣A-crystallin promoter. To serve as control, we made transgenic mice having Ser 587 3 Ala mutation in the APH gene to express inactive enzyme in the lens (51). The existence of an animal model of APH-overexpressing mice will help in delineating the role of crystallin fragments in cataractogenesis in vivo.

EXPERIMENTAL PROCEDURES
Transgene Construct-We used porcine liver APH cDNA, cloned onto pET23a vector with a reading frame, consisting of 756 residues having the full-length APH subunit sandwiched between the N-terminal T7 tag, MASMTGGQQMGRGS, and the C-terminal His tag, AALEHHHHH (51). All plasmids were maintained and propagated in XL1-Blue super competent cells (Agilent Technologies, Santa Clara, CA). For the construction of transgene, APH cDNA along with the N-and C-terminal tags were amplified by high fidelity PCR master kit (Roche Applied Science), using the following set of primers: 5Ј-ATAGGGAG-ACCACAACGGTTT-3Ј and 5Ј-TCTAGACCCGTTTAGAG-GCCC-3Ј. The amplified cDNA was then cloned on to pCRII-TOPO vector (Invitrogen) to generate pCRII-TOPO-APH plasmid. To direct the expression of transgene under ␣A-crystallin promoter, we used ␦en-␣A-plasmid (52), which was previously modified by PCR to include an additional KpnI site at the 3Ј end using the following primer pairs: 5Ј-GCCTGTTGAATTAAT-TCGAGGTACCAATTCGCCCTATAGTGAG-3Ј (sense) and 5Ј-CTCACTATAGGGCGAATTGGTACCTCGAATTAAT-TCAACAGGC-3Ј (antisense). The APH cDNA from pCRII-TOPO-APH plasmid was released and inserted into the modified ␦en-␣A vector at the EcoRI site to obtain the ␦en-␣A-APH plasmid. The orientation of the gene was confirmed by restriction mapping and DNA sequencing. The ␦en-␣A-APH plasmid was digested with KpnI to release the active APH (wt-APH) mini-gene fragment, as shown in Fig. 1A. To generate the inactive APH (mt-APH), the Ser 587 residue in the APH segment of the reading frame was replaced with Ala using the QuikChange site-directed mutagenesis kit (Agilent Technologies). The wtand the mt-APH genes express active and inactive enzyme, respectively.
Generation of Transgenic Mice-The studies conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research, and the protocols were approved by the University of Missouri Animal Care and Use Committee. Transgenic mice were generated at the University of Missouri transgenic animal core facility by pronuclear injection of the purified mini-gene fragment into one-cell stage inbred FVB/N embryos. Potential transgenic mice were identified by PCR on tail DNA using primers homologous to the APH portion of the transgene. The identified founders were crossed with normal FVB mice to establish a hemizygous line. All experiments were performed using lenses from the hemizygous mice unless otherwise mentioned.
Hematoxylin and Eosin (H&E) Staining-Eyes from transgenic mice sacrificed at appropriate time points were fixed in 10% neutral-buffered formalin overnight, dehydrated with ethanol, and embedded in paraffin for sectioning. Tissues sections (5 m) were stained with H&E using the standard procedure.
Immunohistochemistry-To determine the APH expression pattern, deparaffinized sections were rehydrated, and antigen retrieval was done by placing the sections at subboiling temperatures in 0.1 M sodium citrate (pH 6.0) for 10 min. The sections were washed in PBS, and the endogenous peroxidases were quenched in 10% methanol, 3% H 2 O 2 in PBS followed by blocking in 5% BSA solution prepared in PBS. The sections were then treated with anti-APH antibody (Sigma; HPA029700) and probed with biotinylated anti-IgG (Vector Laboratories, Burlington, CA; BA-1000). The biotin was detected using a Vectastain Elite ABC kit (Vector Laboratories) followed by color development using diaminobenzidine as a substrate (Sigma-Aldrich).
Transmission Electron Microscopy-Eye globes were carefully removed from transgenic mice expressing active and inactive APH, fixed in 2% paraformaldehyde and 2% glutaraldehyde in 0.13 M sodium cacodylate and 0.13 mM CaCl 2 (pH 7.4), and kept overnight at room temperature. Slits were made in the globe for the fixative to penetrate the lens. The samples were washed with cacodylate buffer, postfixed in 1% osmium tetroxide in 0.1 M cacodylate buffer for 1 h, dehydrated through a graded ethanol series, and embedded in epoxy resin. Thin sagittal sections were cut through at least one-third of the lens and stained with uranyl acetate and lead citrate. The sections were examined and photographed with a transmission electron microscope (1200EX; JEOL, Tokyo, Japan).
Enzyme Activity-Lenses from the transgenic and nontransgenic mice, sacrificed at various time points, were homogenized in 500 l of Tris buffer (TB) pH 7.5 using a micro tissue homogenizer. The samples were centrifuged at 5,000 rpm, and the supernatant was used to measure APH activity after estimating the protein content using Bio-Rad protein assay reagent (Bio-Rad). APH activity was measured using N-acetyl-Ala-p-nitroanilide (Ac-Ala-pNA) (Bachem Americas, Inc., Torrance, CA) as the substrate. To 5-25 l of the extract taken in 1 ml of TB was added 5 l of 15 mg/ml Ac-Ala-pNA prepared in DMSO. The enzyme activity was monitored kinetically at 37°C for 5 min at 405 nm on a Shimadzu UV-visible spectrophotometer. The number of moles of substrate hydrolyzed was calculated from the extinction coefficient of p-nitroaniline, 8.27 mM Ϫ1 cm. The unit of APH activity is reported as moles of p-nitroanilide released min Ϫ1 mg Ϫ1 protein.
SDS-PAGE and Western Blotting-Lenses isolated at appropriate time points were homogenized in ice-cold TB containing 6 M urea and protease inhibitor mixture. The samples were centrifuged, and the supernatant was collected. To compare the protein profiles of various transgenic lines, 150 g of the protein was loaded onto a 4 -12% Bis-Tris precast gel with XT-MOPS buffer and run on a criterion system (Bio-Rad). Gel images were obtained using Bio-Rad Chemidoc XRSϩ, and the transgenic protein levels were estimated using Image Lab software (Bio-Rad).
For Western blotting, samples containing 1-25 g of protein were run on 4 -20% Mini-Proten TGX precast gel (Bio-Rad) and electrotransferred on to PVDF membrane using Bio-Rad Trans-Blot SD semi-dry electrophoretic transfer cell. The membranes were washed in TB, blocked in 5% milk, and incubated with the appropriate primary antibodies in TB with 0.1% Tween at 4°C overnight, followed by incubation with horseradish peroxidase-labeled secondary antibodies, and developed by SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific, Rockford, IL). The primary antibodies used in the study were obtained from the following sources: APH and ␤-actin (Sigma-Aldrich); ␣A-crystallin (a kind gift from Dr. Cenedella, A.T. Still University of Health Sciences, Kirksville, MO); ␣B-crystallin (Enzo Life Sciences Farmingdale, NY); ␤-crystallin (LifeSpan Biosciences, Seattle, WA); and ␥D-crystallin (Santa Cruz Biotechnology, Dallas, TX).
Sample Processing for Two-dimensional Differential Gel Electrophoresis (DIGE)-Three-month-old lenses (n ϭ 4) from a transgenic mouse expressing wt-APH and from a nontransgenic mouse were used. The proteins were isolated using a modified protocol originally described for isolation of proteins from plant tissues (53). Briefly, the lenses were ground up to a fine powder under liquid nitrogen. The protein was extracted from the powdered sample using equal amounts of buffered phenol and extraction media (0.1 M Tris-HCl, pH 8.8, 10 mM EDTA, 0.4% 2-mercaptoethanol, and 0.9 M sucrose), and the slurry was placed on a nutator to agitate for 30 min at 4°C. After agitation, the sample was centrifuged at 4000 ϫ g for 30 min at 4°C. The phenol layer was isolated, and the process was repeated two more times. Added to the pooled phenol layer were 5 volumes of 0.1 M ammonium acetate in methanol. Proteins were allowed to precipitate overnight in Ϫ80°C. The samples were then centrifuged at 4000 ϫ g for 30 min, and protein pellets were washed two times with 0.1 M ammonium acetate in methanol and 10 mM DTT, keeping the samples at Ϫ20°C for 20 min between washes. The pellets were washed finally with 80% acetone. The 80% acetone protein pellets were suspended in 100 l of sample buffer (7 M urea, 2 M thiourea, 4% (w/v) CHAPS, 40 mM Tris base, 1% (w/v) DTT, and a protease inhibitor mixture (EMD Millipore, Darmstadt, Germany)). The samples were quantified using EZQ protein quantification kit (Invitrogen). The nontransgenic mouse lens proteins (50 g) were labeled with 200 pmol of Cy3, and the same amount of proteins from wt-APH lens were labeled with Cy5 dye and kept on ice for 30 min. The reaction was quenched with 10 mM lysine. The labeled samples were pooled and mixed with 350 g of unlabeled protein from each sample.
Two-dimensional Difference Gel Electrophoresis and Gel Imaging-The pooled samples were loaded onto a 23-cm nonlinear rehydrated immobilized pH gradient (IPG) strip, pH 3-10 (GE Healthcare) and focused on a Bio-Rad PROTEAN isoelectric focusing apparatus. After first-dimensional isoelectric focusing, the strip was placed on a lab-casted 8 -20% gradient gel for the second dimension. Gel was run until the dye front reached the bottom of the gel. The gel was scanned using the FUJI FLA-5000 and then stained with colloidal Coomassie stain overnight for spot visualization. The images were analyzed by DeCyder differential analysis software (GE Healthcare). The differential in-gel analysis module compared the individual Cy3-labeled proteins to the Cy5-labeled proteins and selected fold changes that were only two or more in normalized spot volume. The Coomassie-stained gel was scanned using a flatbed scanner.
Protein Identification by Peptide Mass Fingerprinting-Coomassie Blue (G250)-stained two-dimensional gel plugs were destained and in-gel digested using bovine trypsin (G-Biosciences, St. Louis, MO), according to the protocol described and posted online by the University of Missouri proteomics core. The digests were analyzed by 4700 MALDI-TOF/TOF mass spectrometer (Applied Biosystems). The spectra were processed and batch-analyzed in the combined MS and MS/MS mode with Applied Biosystems GPS Explorer software. Database searches were performed with the Mascot search engine (Matrix Science) against the NCBInr mammalian protein database.
Analysis of Low Molecular Weight (LMW) Peptides-Peptides of Ͻ10 kDa were isolated and analyzed by MALDI-TOF-MS and nanospray QqTOF MS/MS analysis. Lenses (5 per group) were homogenized in Tris buffer (20 mM, pH 8.0) containing 5 mM Tris (2-carboxyethyl) phosphine hydrochloride, 6 M urea, and protease inhibitor mixture. The sample was centrifuged at 10,000 ϫ g for 20 min, and the supernatant was filtered using a 10-kDa filter. The filtrate was desalted using Prevail C18 cartridge, and the bound peptides were eluted using 70% acetonitrile containing 0.1% formic acid. The samples were initially analyzed in the presence of ␣-cyano-4-hydroxycinnamic acid matrix by MALDI-TOF-MS (Applied Biosystems, Voyager DE Pro), in the positive ion reflector delayed-extraction MS mode over the mass range 500 -6000 Da. For the identification of protein, the peptides of interest were further analyzed by nanospray QqTOF MS. Multiple charge ions corresponding to the [MϩH] ϩ of the peptides of interest were subjected to collision-induced dissociation MS (MS/MS) for fragmentation analysis.

In Vitro Hydrolysis of Lens Proteins by APH and Estimation of LMW Peptides-Five nontransgenic mouse lenses (30 -40
postnatal days old) were homogenized in 500 l of TB using a micro tissue homogenizer. The sample was centrifuged at 5,000 rpm, and the supernatant containing 10 mg of water-soluble proteins was used. APH used in the incubation with lens proteins was purified from 30-day postnatal transparent KKS2 lenses using the His60 nickel gravity column (Clontech). Bound wt-APH was eluted from the column using PBS containing 300 mM imidazole. The activity was monitored using Ac-Ala-pNA substrate. The enzyme-rich fractions were concentrated and dialyzed against TB and were found to be Ͼ95% pure by SDS-PAGE. The lens proteins (10 mg) were incubated with 30 g of purified wt-APH in 500 l of TB at 37°C for 24 h. At the end of the incubation, protease inhibitor mixture was added to stop proteolysis activity. A tube containing the same amounts of the incubation mixture and treated with protease inhibitor mixture, prior to the start of incubation, served as a positive control. Another tube containing lens extracts and without wt-APH served as a negative control. At the end of incubation, solid urea (6 M) and Tris (2-carboxyethyl) phosphine hydrochloride were added to the samples, vortexed, and filtered through a 10-kDa filter. The filtrate was analyzed for LMW peptides using MALDI-TOF-MS as described previously.

Generation of Transgenic Mice Expressing APH in the Lens-
Porcine APH cDNA containing an N-terminal T7 tag and a C-terminal His tag was inserted into modified ␦en-␣A vector at the EcoRI site between the rabbit ␤-globin intron and human growth hormone poly(A) signal (Fig. 1A). The ␦en-␣A promoter directs lens-specific transgene expression in the lens epithelial and fiber cells (52). The transgene was released from the vector using KpnI, purified, and microinjected into the pronuclei of FVB/N mouse embryos. Seven wt-APH (active) and four mt-APH (inactive) expressing lines were established. The wt-APH lines were designated as KKS1, KKS2, KKS3, KKS4, KKS5, KKS6, and KKS7, and the mt-APH lines were labeled as KKS1M, KKS2M, KKS3M, and KKS4M. The founder mice and the hemizygous offspring were bred with normal FVB mice and followed for 10 (F10) generations.
Lens Morphology and APH Activity-The transgenic mice exhibited a normal appearance, and their lenses appeared normal when they opened their eyelids. The mice were examined periodically for cataract development. The majority of wt-APH-expressing lines developed cataract by 30 -40 postnatal days, whereas the lines expressing mt-APH had normal transparent lenses (Fig. 1B). The APH activities of transgenic mouse lenses were estimated at various postnatal intervals, from days 1 to 21. Total APH activity increased with age. However, the specific activity of APH remained more or less constant throughout the assay period. For example, in transgenic line KKS2, total APH activity per lens was 0.25 units on day 1 and reached 4.1 units on day 21, but the specific activity remained at ϳ1.9 units during this period, suggesting that the ratio of APH protein to lens crystallins remained the same during the entire postnatal period. Fig. 2A shows the specific activity of APH at postnatal day 21 in lenses of different transgenic lines. The endogenous APH activity in the mouse lens is low and cannot be measured during the 5-min assay period. As expected, the mice expressing mt-APH did not have any measurable APH activity. Transgenic mice expressing wt-APH showed very high activities (Ͼ1,000-fold) when compared with mt-APH and nontransgenic controls. The specific activity varied among the wt-APH lines. KKS2, KKS4, and KKS7 had the highest APH activity (1.9 units); KKS3 and KKS6 had moderate activity; and KKS1 and KKS5 had the least activity. Fig. 2B lists the postnatal days required to develop cataract in F5 generation of transgenic mice overexpressing APH. Of the seven transgenic lines expressing active APH, five developed cataract in 3-9 weeks after birth. The sixth transgenic line (KKS1) exhibited cataract after 8 weeks in only 5% of the population, and the seventh line, which had the least amount of APH activity (KKS5), never developed cataract. The onset of cataract in the transgenic lines correlated with the total APH activity in the lens, with cataract developing at an earlier postnatal day as the total APH activity became greater. Fig. 3 shows the morphological features of the lenses under stereoscope at various time intervals. Cataract development in mice expressing wt-APH begins with a clear demarcation of the inner nuclear area, which continues with the ejection of primary fiber cells through the posterior suture into the vitreous. The region expelled into the vitreous manifests as an opaque protein mass resembling dense cataract. The major portion of the lens, including the epithelium and cortex, remains transparent, with some degree of cloudiness in the nuclear region. KKS1 and KKS5 exhibited nuclear demarcation at 3 months, but there was no extrusion of fiber cells or cataract development at any time. Lenses from transgenic lines expressing mt-APH were transparent and looked similar to nontransgenic lenses.
SDS-PAGE Analysis-APH protein levels in transparent 30-day-old transgenic lenses were determined by SDS-PAGE. Lenses homogenized in 6 M urea buffer and supernatants containing equal amounts of protein were run on a SDS-PAGE gel (Fig. 4). As expected, an 84-kDa band, corresponding to the recombinant APH protein, was seen in transgenic lenses. The APH protein levels varied among the transgenic lines and correlated with their Ac-Ala-pNA hydrolyzing activity. Except for the APH band, no other change was apparent in the protein profiles of 30-day-old transparent wild-type, mutant, and nontransgenic lenses. The amount of transgenic protein was estimated from the gel picture using Image Lab software (the values are included in parentheses on top of the lanes in Fig. 4). At 30 days, APH accounted for approximately 2.5% of the total protein in KKS2 and KKS4, 0.48% in KKS1, and 0.36% in KKS5 lenses. All mutant lines had similar levels of APH protein (0.6% of total protein). SDS-PAGE analysis of the cataract and the transparent regions of transgenic lenses expressing active APH (Fig. 4) showed that cataract development in these mice is not due to the precipitation of overexpressing APH, because the precipitate contained mainly crystallins, and the more intense APH band was seen in the protein fraction from the transparent region of the lens. Along with the crystallins, a prominent low molecular weight band was also seen at the 6.5-kDa region in the cloudy aggregate.
Western Blot Analysis of Transgenic Lenses-To look for changes in specific proteins, Western blot analysis was done on nontransgenic and transgenic (KKS4) lenses at different postnatal days using APH, ␣A-crystallin, ␣B-crystallin, ␤-crystallin, and ␥D-crystallin antibodies (Fig. 5A). Lenses with or without the cataract aggregate were homogenized in urea buffer, and the protein levels were normalized based on GAPDH levels. The transgenic lenses showed an intense band for APH. The APH band intensity decreased marginally with KKS4 after cataract development. Interestingly, at 30 days (just prior to cataract development) and beyond, an additional protein band was seen below the main APH band, indicating cleavage of the APH protein.
Protein bands with corresponding mobility were not detected in nontransgenic lenses, including for the endogenous APH (Fig. 5A), suggesting that the endogenous APH levels are extremely low. There was no apparent difference in the ␣Aand ␣B-crystallin bands among transgenic and nontransgenic lenses. The band intensity of ␤-crystallin was slightly less in transgenic lenses when compared with their nontransgenic counterparts. The difference in the ␥D-crystallin levels between transgenic and nontransgenic lenses was quite dramatic, especially in lenses having cataract. Transgenic lenses had lower levels of ␥D-crystallin when compared with age-matched nontransgenic controls. The ␥D levels decreased with the aging of the lens in both groups (Fig. 5A). Based on the normalized Western blot band intensity, the ␥D-crystallin levels of the non-FIGURE 2. Characteristics of transgenic mice. A, specific activity of APH in 21-day-old transgenic mouse lenses. The activity was measured kinetically for 5 min using 100 nmol of Ac-Ala-pNA substrate. The activities are average Ϯ S.E. from five lenses estimated individually. Activities of APH in lenses expressing active enzyme were significantly higher than those of mutant and nontransgenic controls (p Ͻ 0.0001). Mice expressing mt-APH did not have measurable activity, which was comparable with nontransgenic controls. B, time required for cataract development in various transgenic lines.
transgenic mouse lens peak at 30 days postnatally, after which it drops drastically. At 4 months postnatally, the ␥D-crystallin levels of nontransgenic lenses were Ͻ10% of the levels at 1 month postnatally.
We also compared the protein levels among different transgenic lines (Fig. 5B) at 30 days postnatally. Cleavage of APH was seen in KKS2 and KKS4, which expressed high levels of APH, but no such cleavage was seen in KKS5 and KKS4M, suggesting that APH cleavage could have a role in cataract development. ␣Aand ␣B-crystallin intensities were not significantly altered in these lenses. However, ␤and ␥D-crystallin levels appeared to be reduced in transgenic lenses, with a greater reduction in KKS2 and KKS4. The ␤-actin levels were also affected in these lenses (Fig. 5B) because such all protein normalizations were done using GAPDH.
Histological Analysis of Transgenic Lenses-Up until the formation of cataract, the lenses of mice expressing active APH were not distinguishable by H&E staining from those of mice expressing inactive APH nor from those of mice that did not develop cataract (Fig. 6). At an early stage of cataract develop-  ment, the anterior portion of the transgenic lens was less affected than the posterior region. Lens capsule was ruptured at the posterior pole of the lens (Fig. 6B), with primary fiber cells expelled into the vitreous. The lens epithelial cells were abnormal and vacuolated. At an advanced stage of cataract, the lenses of mice that developed cataract were asymmetrical and had large vacuoles (Fig. 6C). No apparent changes were found by H&E staining in lenses of transgenic mice expressing inactive APH (Fig. 6, D and E) and of transgenic mice expressing low levels of active APH (KKS5 and KKS1) (data not shown). The histological appearance of transgenic lenses from the mutant, KKS1 and KKS5, was comparable with that of nontransgenic lenses at all age-matched periods.
APH expression patterns in the lenses were assessed at various postnatal time periods using APH antibody (Fig. 6, G-L). Intense and uniform APH staining was seen throughout the lens at postnatal day 7 in transgenic lenses expressing wt-and mt-APH. The intensity of staining was proportional to the APH activity of the lens. The less intense staining in the nucleus of mutant lenses at 90 days postnatally (seen in Fig. 6K) is not due to the absence of APH protein but to the dense packing of primary fiber cells, impeding the penetration of antibody. Nontransgenic lens showed faint staining, indicating low levels of endogenous APH in mouse lens (Fig. 6L). Just prior to cataract development, at postnatal day 30, the lenses of KKS2 attained the shape of an inverted pear (data not shown), whereas the normal spherical shape was seen in lenses from lines that express active APH and do not develop cataract and in lenses from mutant lines. It is difficult to conclude from the immunohistochemistry data that APH protein levels change with time. At an advanced stage of cataract, APH staining was seen in fiber cells expelled into the vitreous, which manifests as cataract.
Transmission electron microscopy of transgenic lenses expressing high levels of wt-APH and developing cataract and of mt-APH lenses showed that the lenses expressing active protease had vacuoles between and within the fiber cells at 2 months postnatally, suggesting that APH overexpression disrupts the fiber cell integration (data not shown). The lens expressing mt-APH appeared normal.

MALDI-TOF-MS Analysis of LMW peptides in Transgenic
Lenses-To identify the native substrate(s) for APH in the lens and to evaluate the role of low molecular weight peptides in cataract development, we assessed the mass spectrometry profile of Ͻ10-kDa peptides from the 30-day-old postnatal transgenic lenses expressing wt and from mt-APH lenses. Only transparent wt-APH lenses were selected for analysis. An explicit difference was found in the peptide profiles of transgenic lenses expressing active and inactive APH (Fig. 7). Many of the peptides seen in mt-APH lenses were also present in wt-APH lenses at equal abundance. However, lenses expressing wt-APH had more Ͻ10-kDa peptides when compared with age-matched lenses expressing mt-APH, suggesting increased proteolysis in lenses expressing active APH. MS/MS analysis of a few distinct ions indicated that these peptides come from ␣Aor ␣B-crystallin. ␣-Crystallin fragments-viz. ␣A38 -56, ␣A43-57, ␣B43-57, ␣B43-56, and ␣B45-57-were present only in lenses expressing wt-APH (Fig. 7), indicating that APH activity could have a role in the generation of these peptides. Previously, we reported that ␣A43-56 is present in young, aged, and cataractous human lenses and that ␣B45-57 is present in young human lenses (16). In lenses expressing inactive APH, acetyl-␣A1-26 and acetyl-␣A1-24 were present in abundance when compared with the levels in lenses expressing active APH. Such a difference is expected, because APH is known to hydrolyze N-acetyl-peptides, which are then readily hydrolyzed by aminopeptidases.
Two-dimensional DIGE and Peptide Mass Fingerprinting-To compare the protein levels of nontransgenic and transgenic lenses, we performed a two-dimensional DIGE analysis of nontransgenic and transgenic lenses labeled with Cy3 and Cy5, respectively (Fig. 8). The relative fold differences in protein levels between nontransgenic and transgenic lenses were estimated from the scanned fluorescent images using Decyder differential analysis software. The difference in the spot volume was used to estimate the fold change. Protein spots that changed by 2-fold in volume (Fig. 8A) were selected for mass spectrometric analysis. Fig. 8 (B and C) shows an enlarged view of the gel highlighting the analysis for APH spots marked in  MARCH 28, 2014 • VOLUME 289 • NUMBER 13 pink. We analyzed 35 protein spots that showed difference in their levels. Table 1 gives the list of proteins identified from the spots using MS analysis and their fold difference in volume between nontransgenic and transgenic lenses. At least three independent spots for the full-length transgene protein (84 kDa) were seen (spot numbers 221, 226 and 230), suggesting that APH protein has undergone modification that affects its pI. We also found 3 spots (373, 375 and 381) corresponding to truncated forms of APH and having a molecular mass of ϳ57 kDa. The MASCOT MS/MS ion search of tryptic peptides from these spots matched the APH transgene sequence from residue 282 to the C-terminal end, indicating that the 57-kDa APH fragment is generated by the removal of the N-terminal end. The cytoskeletal protein vimentin, implicated earlier in cataract development (54) was increased over 2-fold in transgenic lens compared with nontransgenic lens. More than 50% of the protein spots identified were crystallins, which included ␣A-, ␣B-, ␤B1-, and ␤B2-crystallins. Of the six ␣A-crystallin spots, spots 1217 and 1227 were present only in nontransgenic lenses, and spots 1313, 1515, and 1516 were either reduced or absent in nontransgenic lenses. It is plausible that spots 1217 and 1227, having a molecular mass of ϳ19 kDa, represent modified forms of full-length ␣A-crystallin cleaved in the transgenic lenses, resulting in 6.5 kDa spots, 1515 and 1516. Similarly, in the transgenic lenses, the majority of longer length ␣B-crystallin polypeptides (spots 1108, 1059 and 1065) were decreased and the majority of shorter length ␣B polypeptides (spots 1121, 1147 and 1200) were increased, suggesting enhanced cleavage of ␣B-crystallin in lens overexpressing APH. Compared with nontransgenic lenses, the levels of ␤B2-crystallins were elevated and ␤B1-crystallin was decreased in transgenic lenses. In Vitro Hydrolysis of Mouse Lens Proteins by APH-To confirm the role of APH in the generation of LMW peptides, we incubated nontransgenic mouse lens proteins with wt-APH purified from KKS2 lenses using a nickel affinity column. At the end of the incubation, the LMW peptides present in the incubation mixture were analyzed by MALDI-TOF-MS. We found a Ͼ20% increase in the LMW peptides after incubation with wt-APH as compared with LMW peptide levels after incubation with lens proteins and wt-APH in the presence of protease inhibitors (positive control) or after incubation with lens proteins without APH (negative control) (Fig. 9). The LMW peptide profiles of negative and positive control samples (Fig. 9, A  and B) were largely similar, with a small increase in peptides in negative control sample, which could be due to the action of proteases present in a normal mouse lens. LMW peptides increased in samples incubated with wt-APH (Fig. 9C), and the peptide profile differed significantly from that of the control samples. Interestingly, one peptide (2197 Da) was 8 -10-fold more abundant in the active APH incubations. We have yet to identify this peptide. Except for the 1730-Da peptide, peptides generated by the action of APH in vitro (Fig. 9C) and the peptides isolated from 30-day-postnatal mouse lenses expressing wt-APH were different (Fig. 7). This is expected, because in vivo the proteases are compartmentalized, and under the conditions of in vitro incubation, APH and other proteases and peptidases act simultaneously. Moreover, in vivo the peptides accumulate gradually, over an extended period of time, whereas they accumulate rapidly with in vitro hydrolysis. Our data clearly suggest a role for APH in the generation of LMW peptides. . The two-dimensional DIGE analysis of 90-day-old nontransgenic and transgenic mice expressing active APH. A, two-dimensional gel of whole lens extract labeled with Cy3 (nontransgenic) and Cy5 (transgenic) dyes. The image was scanned and analyzed using Decyder software. The gel was stained with Coomassie Blue, and the protein spots that were picked for mass spectrometry are shown as dotted circles. See Table 1 for the identity of proteins present in each spots picked from A. B, spot 226 (area marked in pink) from two-dimensional DIGE gel was identified as APH by mass spectrometry. C, spot 375 (area marked in pink) from two-dimensional DIGE gel was identified as truncated APH. The areas marked in blue and red are other spots selected by the software to estimate the relative fold difference in protein levels between transgenic and nontransgenic lenses.

DISCUSSION
Cataractous human lenses exhibit increased proteolysis and protein aggregation as compared with age-matched, noncataractous lenses. However, the role of crystallin fragments in disrupting the ordered protein structure of the lens and inducing cataract formation is unclear. During aging, protein aggregates composed of significant amounts of modified and C-terminally truncated ␣Aand ␣B-crystallins and crystallin fragments are formed. We have previously shown that the interaction between in vivo generated crystallin fragments and crystallins leads to protein aggregation and insolubilization, resulting in a loss of transparency (16,21). To test the role of the human lens protease APH in the formation of LMW peptides that promote cataract development, we generated transgenic mice that overexpress APH specifically in the lenses. We found that mice expressing active transgene develop cataract and have peptides similar to those accumulating in human cataractous lenses, establishing a direct correlation between crystallin fragmentation and cataractogenesis.
At the outset, APH activity appears to be responsible for cataract development, because clouding of the lens is seen only in mice expressing active APH. However, the data do not clearly indicate such an association, because lens opacity did not develop in all lines expressing active APH. The specific activity of APH was constant in all age groups of lenses, from postnatal day 1 onward, but cataract developed only after 3 weeks, discounting the direct role for APH protein in cataract formation because of "stuffing of the fiber cells." The onset of cataract in the transgenic mice correlated with the total APH activity in the lens (Fig. 2), and it took a few weeks for cataract to develop. Thus, it may be argued that APH is likely targeting proteins critical for the later stages of lens development. However, at this time we are unable to narrow down cataract formation to a specific protein or cytokine or growth factor(s) targeted by APH.
Our observations of APH and cataract development are similar to those described in the previously reported studies of transgenic mice expressing HIV-1 protease in the lens under the control of ␣A-crystallin promoter (40). Mice with active HIV-1 protease developed cataract, and mice bearing inactive protease had normal lens. HIV-1 transgenic mice displayed degradation of some lens crystallins and cytoskeletal proteins prior to cataract formation on postnatal days 23-25, and the lens morphology looked somewhat similar to that of lenses with APH (55,56). HIV-1 protease has no direct involvement in cataract development, however; its expression activates endogenous cysteine proteases, which may have a role in opacification (55). We have not measured the activity of other endogenous proteases. If other proteases responsible for crystallin breakdown are activated, we should have also seen cataract

Proteins identified using Mascot search for the MALDI-TOF-MS/MS ions of the mice lens two-dimensional gel protein spot digests
development in KKS5 because this line has enough activity to activate other proteases. In another study involving lens cysteine protease, overexpression of inactive calpain-2 did not lead to lens opacification (37), and the relative amount of mutant protease to lens proteins in that study is not known.
There appears to be a direct correlation between transgene APH levels and cataract development. Clouding of the lens was seen when APH protein levels were Ͼ2.6% of the total lens protein. The transgenic line expressing active APH and exhibiting no cloudiness (KKS5) had APH levels that were Ͻ0.6% of the total lens protein at postnatal day 30, and the APH levels never went above 1% at any time. All the mutant lines had transgene APH protein levels below 1% of the total lens protein.
Our attempt to increase the transgene protein levels in mutants using homozygous mice did not increase the level to Ͼ1.5% (and those also did not have increased incidence of cataract). There was only 1.7-fold difference in the mt-APH and wt-APH levels, yet only the lenses that expressed wt-APH developed cataract. It is tempting to suggest that a developing mouse lens has a limited tolerance for transgenic protein, even if the protein expressed is native to the lens. When the transgenic protein amounts increase beyond the critical level, proper maturation of the lens might be affected, culminating in clouding. In another study, the transgenic lenses with 6-fold overexpression of glutathione peroxidase did not develop cataract (57). Petrash and co-workers (58) have reported that the transgenic lenses overexpressing wt-␣A-crystallin resulting in 4% of the total ␣A-crystallin were clear. It would be interesting to see whether the mutant lines develop cataract when their transgenic protein levels go over 2.6% of total lens protein. It has been suggested that transgenic overexpression would activate an unfolded protein response, which can have adverse effects on lens fiber cell differentiation (59,60). We believe that the expression of APH may not directly activate an unfolded protein response in our model, because APH is native to the mouse lens and is expressed as soluble protein. In addition, the KKS5 lenses as well as the mutant lenses exhibit no gross morphological changes, which would generally be expected to occur in lenses with an unfolded protein response.
We found in our model that the days to develop cataract and the extent of severity depend on transgene APH levels. Similar observations were made with transgenic mice overexpressing vimentin in the lens (54). Vimentin overexpressing mice exhibit posterior capsule rupture and disintegration of primary fiber cells, as do mice expressing active APH, but there are also differences between vimentin expression-induced cataracts and wt-APH expression-induced cataracts. The vimentin cataract lenses showed extensive impairment of the denucleation process and increased density of nuclei in the equatorial region (54), whereas the cross-section of cataract forming wt-APH expressing lens (KKS4 line) did not have similar conditions (Fig. 6B). Further, there was suppression of the lens fiber cell elongation in vimentin-induced cataract model, whereas this was not obvious in APH model. Other studies have also suggested a role for vimentin in cataract development (61,62). Loss of HSF4 function has been shown to reduce crystallin levels, which activates vimentin expression (59). Our two-dimensional DIGE data show an increased cleavage of crystallins and a 2.36-fold increase in vimentin protein in transgenic lenses (Table 1). It is not clear whether elevated vimentin levels are a consequence of lens degeneration or a cause of cataract development. Our Western blot analysis data show (Fig. 5B) that the level of the cytoskeletal protein ␤-actin is significantly decreased prior to cataract development in mice expressing active APH. Reduction of ␤-actin levels could have caused vacuole formation in the lens epithelial and fiber cells of mice expressing high levels of active APH. The decreased levels of the microfilament and increased levels of intermediate filament collectively might have caused lens disintegration and cataract formation.
We previously showed that APH is unable to act on intact crystallins and that cleavage of a 22-kDa N-terminal fragment from the APH generates a 55-kDa active enzyme capable of unblocking and hydrolyzing crystallins (45,47). The 55-kDa fragment is also present in human and bovine lenses, in addition to full-length APH (46). The two-dimensional DIGE and Western blot analysis for the APH protein in the current study show that a small fraction of the transgene is cleaved in the lenses that develop cataract (Fig. 5). Interestingly, the fragment has a mass of ϳ57 kDa, and the MASCOT MS/MS ion search of tryptic peptides from the spot showed the cleavage of the N-terminal fragment, releasing the catalytic fragment of the enzyme. We did not see cleavage of APH at day 15, despite high APH levels. Fragmentation was seen at day 30, just prior to cataract development (Fig. 5A). The appearance of the 57-kDa fragment correlated with cataract development. KKS5 and mutant lines did not show cleavage of APH. It is plausible that the 57-kDa APH fragment possesses hydrolytic activity against crystallins, similar to our previous observation made in vitro with a 55-kDa APH fragment (47). In support of this view, we saw an increased accumulation of crystallin fragments in mouse lens expressing active APH even before the development of cataract. We also saw an increase in the generation of peptides during in vitro incubation of lens proteins with wt-APH. The 57-kDa APH fragment was also present (Ͻ 1%) along with full-length APH in our in vitro experiments because we were not able to separate them during purification. Although it is certain that APH has a role in the generation of LMW peptides, we were unable to show whether or not full-length APH or truncated APH is involved in hydrolyzing the proteins. Additional studies are needed to explain why cleaved APH is seen only in the mice that develop cataract and whether the 57-kDa APH fragment has any role in the generation and accumulation of LMW peptides in the transgenic lenses.
Increased cleavage of crystallins was seen in our transgenic mice, with the breakdown of crystallin occurring even before the development of cataract. At postnatal day 30, only a small fraction of crystallins are fragmented, and it is difficult to envisage widespread crystallin breakdown from our SDS-PAGE data (Fig. 4) because of dense protein bands. A clear difference is seen in the amount of LMW peptides (Ͻ3 kDa) that accumulate in the lenses (Fig. 7). Transgenic lenses expressing wt-APH showed increased accumulation of LMW peptides compared with lenses expressing mt-APH. Analysis of the few abundant peaks by MS/MS suggests that these peptides mainly originate from ␣-crystallin. We showed in previous studies that aging and cataractous human lenses accumulate crystallin fragments that are Ͻ3 kDa in mass (16). Some of these peptides promote protein aggregation at low concentrations in crowded environments (21). Recently we reported the presence of enzymes in bovine and human lenses that have the potential to release LMW peptides that promote protein aggregation (23). The importance of LMW peptides in lens aging and cataract development has also been suggested by others (63,64). In the current study, several LMW peptides were common to both active APH and inactive APH transgenic lenses, but the lenses with wt-APH showed several peptides distinct to that group. Interestingly, the peptides that were unique to wt-APH lenses are also present in aged and cataract human lenses (16). We believe that LMW accumulating in lenses with wt-APH may have a role in the impaired fiber cell differentiation and cataract development observed in our transgenic model. To our knowledge, this is the first study showing increased accumulation of LMW peptides in transgenic mice that develop cataract.
Partial modification and degradation of crystallins occur in the mouse lens during development (29). MALDI imaging studies have also confirmed the modification of lens crystallins during aging in humans and other models (20). We observed increased modifications and truncation of crystallins during our two-dimensional DIGE analysis. However, the analysis was carried out at a later stage, when the lens had developed cataract. Therefore, we cannot conclude that the observed changes contribute to fiber cell breakdown or constitute a consequence of lens clouding. Most of the changes are seen with ␣and ␤-crystallin. We did not see any change with ␥D-crystallin during DIGE analysis, but we found a significant decrease in ␥D-crystallin in our Western blot analysis. This could be due to extremely low levels of ␥D-crystallin in the 3-month-old lenses we used for DIGE analysis. The increased cleavage of crystallins might have led to the accumulation of LMW peptides in mice with wt-APH.
In conclusion, several factors, individually or in combination, including fragmentation of APH, modification and cleavage of crystallins, accumulation of distinct LMW peptides, capsule rupture, fiber cell disintegration, changes in cytoskeletal proteins, and a decrease in ␤and ␥D-crystallin levels, might have led to cataract development in our transgenic model. However, the most significant factor we observed is the removal of N-terminal end from APH, resulting in a fragment having the potential to hydrolyze crystallins and led to the accumulation of distinct LMW crystallins, because these changes are the first to occur prior to cataract development. The transgenic mice we developed, although they may not duplicate all the age-related changes occurring in the human lens, could serve as a model to study the role of LMW peptides in lens crystallin aggregation and cataract development.