DNase IIβ Distribution and Activity in the Mouse Lens

Abstract

purpose. To map the cellular and subcellular distribution of DNase IIβ activity in the mouse lens.

methods. DNase IIβ-specific activity was determined by assaying lens lysates prepared from wild-type or DNase IIβ-null mice. Regional nuclease activity was determined by microdissection of lens samples or a tissue-imprinting assay. Subcellular distribution was determined by density-gradient ultracentrifugation.

results. DNase IIβ transcripts increased 200-fold in abundance during fiber cell formation, and DNase IIβ activity accounted for approximately 50% of the acid nuclease activity in the cortical fiber cells. Examination of lenses from DNase IIβ-null mice confirmed that the enzyme was required for denucleation. In wild-type lenses, nuclei were TUNEL positive before denucleation, indicating that 3′-OH DNA ends had accumulated. However, DNase IIβ-mediated scission generates 3′-PO₄ ⁻ DNA ends only. This paradoxical finding was explained by the presence of phosphatases that converted the 3′-PO₄ ⁻ ends produced by DNase IIβ into 3′-OH ends. DNase IIβ activity was strongest early in differentiation, where it was associated with the lysosomal fraction. Later, an increasing proportion of DNase IIβ activity was found in the cytosol.

conclusions. DNase IIβ activity correlated with and was necessary for fiber denucleation and was most likely contained initially within fiber cell lysosomes before release into the cytoplasm.

The lens is a transparent cellular structure composed of two cell types: epithelial cells, which form a monolayer across the anterior lens surface, and fiber cells, which comprise the remainder of the tissue volume. Fiber cells are produced by differentiation of epithelial cells near the lens equator. Newly formed fibers are added to the surface of the preexisting cell mass. Consequently, the age of a fiber cell can be deduced from its radial position. The oldest fibers, those formed earliest in development, are located in the lens core. The most recently formed fibers are located near the surface. 

During lens fiber cell differentiation, nuclei and other cytoplasmic organelles are degraded. This process occurs over a period of a few hours in a restricted region of the lens cortex and results in the formation of a central organelle-free zone (OFZ). The dissolution of fiber nuclei is preceded by changes in shape and breakdown of the nuclear lamina. ^{1 2} In many species, condensed DNA accumulates at the nuclear envelope. ³ Disintegrating nuclei become TUNEL positive, ¹ indicating the accumulation of 3′-OH ends in the DNA, and, ultimately, nucleosome-sized fragments of DNA are released. ⁴ The degradation of nuclei and other organelles removes light-scattering structures from the visual axis. The abnormal persistence of organelles is observed in congenital cataracts in humans ⁵ and strains of mice with hereditary cataracts ^{6 7 8} or cataracts resulting from expression of certain transgenes. ^{9 10}  

Studies have sought to identify the enzyme(s) responsible for lens chromatin degradation, and several putative nucleases and nuclease activities have been detected. These belong to three categories: acidic cation-independent nucleases (e.g., l-DNase II ¹¹ and DNase IIβ ¹² ), Mg²⁺-dependent nucleases (e.g., DNA fragmentation factor ¹³ ), and Ca²⁺/Mg²⁺-dependent nucleases (e.g., DNase I²). However, a direct role in nuclear breakdown has been demonstrated only for DNase IIβ. ¹² Nishimoto et al., ¹² found that DNase IIβ-null mice develop nuclear cataracts. Opacification was associated with the persistence of nuclei in the innermost fibers. 

DNase IIβ (aka DLAD; EC. 3.1.22), molecular mass ∼41 kDa, shares 34% amino acid sequence identity with DNase IIα and, like that enzyme, does not require cofactors for activity, which is maximal at acidic pH. ¹⁴ Another characteristic common to both enzymes is that DNA scission generates 3′-PO₄ ⁻/5′-OH DNA termini. ¹⁴ The tissue distributions of DNase IIα and β, differ markedly. DNase IIα is expressed ubiquitously, ¹⁵ whereas DNase IIβ is restricted to the lens and the liver. ¹² To date, DNase IIβ is the only enzyme with a demonstrated role in fiber cell organelle breakdown. Our goal in the present study was therefore to gain a better understanding of the distribution and activity of DNase IIβ, in the hope that this might provide additional insights into the regulation of lens organelle degradation in general and denucleation in particular. 

Materials and Methods

Animals

DNase IIβ-null mice were kindly provided by Shigekazu Nagata of Osaka University, and C57-BL6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Animals were killed by CO₂ inhalation. The lens epithelium was stripped from the fiber cell body, and the fiber mass was separated into cortical and core fractions. Unless otherwise stated, 25- to 35-day-old animals were used in these studies. All procedures conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Studies Committee at Washington University. 

Histology

Eyes were fixed in paraformaldehyde (4%/PBS) overnight, washed, dehydrated and embedded in paraffin. Sections were cut in the sagittal plane. In some instances, unfixed cryosections were prepared by embedding lenses in optimal cutting temperature (OCT) medium (TissueTek; Miles, Elkhart, IN) and freezing them in 2-methylbutane over dry ice. Sections were cut on a cryostat at 50 μm. 

Labeling DNA Strand Breaks

Single-strand DNA breaks were identified using a modified TUNEL assay or an in-tube labeling reaction. TUNEL labeling was performed with a cell viability assay (In Situ Cell Death Detection; Roche, Basel, Switzerland) according to the manufacturer’s instructions. 

For in tube labeling, lens cortical lysates (10 μg) were mixed with substrate DNA (2.5 μg of phIX174 plasmid DNA) in 40 μL of activity buffer (50 mM MES-HCl [pH 5.9], 5 mM EDTA) and digested for 30 minutes at 37°C. Some reactions included phosphatase inhibitor (5 μL per tube of a 1:1000 dilution of product P5726; Sigma-Aldrich, St. Louis, MO). Nicked DNA was purified, precipitated and dissolved in water. DNA 3′-OH termini were labeled with biotinylated dNTPs using the Klenow fragment, according to the manufacturer’s instructions (Promega Corp., Madison, WI). Labeled products were separated on 1% agarose gels, transferred to nylon membranes and UV cross-linked. Membranes were blocked with nonfat milk (5%/PBS), and labeled DNA fragments were visualized with a horseradish peroxidase (HRP)-avidin conjugate (Sigma-Aldrich) and chemiluminescence detection (Supersignal; Pierce Biotechnology, Rockford, IL). 

Semiquantitative Nuclease Activity Assay

Acid nuclease activity was assayed by adding 2 μg of lens protein to 40 μL of activity buffer (50 mM MES-HCl [pH 5.9], 5 mM EDTA) containing 2.5 μg of λ DNA (23 kb) as substrate. Samples were incubated at 37°C, aliquots were withdrawn at intervals, and the reaction stopped with loading buffer (0.05% bromophenol blue, 40% sucrose, 0.1 M EDTA [pH 8.0] and 0.5% SDS). Cleavage products were separated on 0.6% agarose gels and stained with ethidium bromide. The rate of disappearance of full-length DNA was approximately linear over the 2-hour incubation period. The normalized rate of loss of full-length λ DNA (% · min⁻¹) was used as a semiquantitative measure of nuclease activity in the sample. 

Polymerase Chain Reaction

To quantify DNase IIα and DNase IIβ mRNA transcripts we used conventional (endpoint) and quantitative real-time PCR (QRT-PCR) techniques. RNA was purified (RNasy protect; Qiagen, Inc., Valencia, CA), and cDNA synthesis was performed (Retroscript kit; Ambion, TX). 

For end-point PCR, the following primers were synthesized: 5′-CGA CCA ACC TCC TAA ATC CA-3′ and 5′-GAG TGG TCC TCT GTG GCA CT-3′ (DNase IIα); 5′-TCC ACG AAT GAC ACA GCC TA-3′ and 5′-GGT CTT GGC GAG AAC TGA AG-3′ (DNase IIβ). The PCR conditions were 94°C for 5 minutes; 30 cycles of 94°C for 1 minute, 58°C for 1 minute, and 72°C for 1 minute, with a final extension at 72°C for 7 minutes. 

For QRT-PCR analysis, the following primers were synthesized: 5′-AAC CTA CAG GTC CAG TTC ACA-3′, 5′-ACA CCA TTT GGA GTG GTC CTC TGT-3′ for DNase IIα and 5′-TCT GCT GGT ATG GAA CAG AAC GCA-3′, 5′-CGA AAG TGA TGC AGA TGC CGG TTT-3′ for DNase IIβ. The expected PCR products were 150 and 131 bp, respectively. The QRT-PCR reaction was performed (iQ SYBR Green Supermix; Bio-Rad, Hercules, CA). Reactions were loaded onto 96-well thin walled plates and run on a real-time PCR detection system (iCycler iQ Multicolor; Bio-Rad) under the following conditions: 95°C for 5 minutes; 45 cycles of 95°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds; followed by 4 minutes at 72°C. PCR products were visualized by agarose gel electrophoresis to verify the amplification of single products. Each product was sequenced to confirm its identity and used to create dilution standards for QRT-PCR. 

In Situ Nuclease Activity Assay

To gauge the relative levels of nuclease activity in different regions of the lens we developed a “tissue imprinting” assay. An immobilized DNA substrate was prepared by UV-cross-linking high-molecular-weight DNA to a nylon membrane. A 50-μm-thick cryosection of lens tissue was laid on top of the substrate and the nylon membrane was placed on filter paper soaked in activity buffer. After a 5-hour incubation under humid conditions, the tissue was removed with proteinase K, and the DNA remaining on the membrane was stained with 1% toluidine blue. 

Cell Fractionation

Cortical tissue was homogenized in ice-cold buffer (50 mM TES [pH 7.4], 7 mM MgCl₂, 0.25 M sucrose and protease inhibitor mixture). Nuclei and cellular debris were eliminated by centrifugation (800g for 10 minutes). Postnuclear supernatant was centrifuged for 1 hour at 100,000g, to obtain cytoplasmic and membrane fractions. Alternatively, postnuclear supernatant was centrifuged for 15 minutes at 20,000g, to obtain the light-mitochondrial pellet. Lysosomes were purified from the light-mitochondrial fraction by ultracentrifugation in a discontinuous density gradient (Nycodenz; Axis-Shield, PoC AS, Oslo, Norway), as described. ¹⁶ Briefly, the gradient was diluted in homogenization buffer to prepare 19%, 24%, 26%, and 30% solutions that were loaded into the tubes from the bottom. The light-mitochondrial pellet, dissolved in homogenization buffer, was loaded onto the gradient and centrifuged overnight at 127,000g (4°C). After centrifugation, the contents of each tube were fractionated and assayed for each organelle marker. The activity of β-glucuronidase, a lysosomal marker, was measured using a fluorogenic assay. ¹⁷ The levels of calnexin and succinate-ubiquinol-oxidoreductase, markers for endoplasmic reticulum and mitochondria, respectively, were determined by Western blot and quantified by spot-densitometry (AlphaEase FC; Alpha Innotech Corp., San Leandro, CA). In all cases, the background was subtracted and values normalized to the highest values obtained. Analysis was performed on two independent samples, each composed of pooled material from 100 mouse lenses. Organelle assays were performed in duplicate. 

Western Blot

Proteins were separated by SDS-PAGE and transferred to nitrocellulose. Blots were incubated with calnexin (SPA-865; StressGen, Victoria, BC, Canada) or succinate-ubiquinol-oxidoreductase (A11142; Invitrogen-Molecular Probes, Eugene, OR) antibodies. The primary antibodies were detected using an HRP-conjugated secondary antibody (Pierce Biotechnology, Rockford, IL) and chemiluminescence (Pierce Biotechnology). 

Lysosome Distribution

Lenses were incubated for 30 minutes (37°C) in culture medium containing a fluorescent lysosomal probe (LysoTracker Green; 1:20,000; Invitrogen-Molecular Probes) and 0.05 μg/μL of the lipophilic membrane stain (FM464 Invitrogen-Molecular Probes) and then examined by confocal microscopy. 

Statistical Analysis

Differences between groups were assessed by unpaired two-tailed t-test. Differences were considered significant at P < 0.05. 

Results

The cation-independent nuclease DNase IIβ was recently implicated in the breakdown of fiber cell chromatin. ¹² This observation prompted us to examine the regulation of DNase IIβ mRNA during fiber differentiation (Fig. 1) . The expression of the related, ubiquitously expressed nuclease, DNase IIα, was also analyzed. In the lens epithelium, transcripts for DNase IIα and DNase IIβ were equally abundant. However, during fiber differentiation, DNase IIβ transcripts increased in abundance ∼200-fold, whereas DNase IIα transcripts decreased ∼40-fold. The ratio of DNase IIα to DNase IIβ, thus, shifted significantly during differentiation, from 1:1 in the epithelium to 1:8000 in the fiber cells. As a negative control, no PCR product was obtained after 30 cycles of amplification using cDNA from DNase IIβ-null lenses (data not shown). 

The lens expresses several nucleases and nuclease activities. ^{2 11 12 13} To determine the specific contribution of DNase IIβ, we developed a semiquantitative assay in which the rate of degradation of an exogenous DNA substrate was measured in lysates prepared from wild-type (Wt) or DNase IIβ-null lenses (Fig. 2A) . The difference in DNA degradation rate between lysates was taken as a measure of specific DNase IIβ activity. At acidic values (pH 5.5–6.0), the rate of DNA degradation in lysates from Wt lenses was approximately twice that of DNase IIβ-null lysates (Fig. 2B) . Thus, DNase IIβ constituted approximately 50% of the lens acid nuclease activity. At neutral pH values, the rate of DNA degradation was lower in both samples, and there was no significant difference between lysates prepared from Wt and DNase IIβ-null lenses. 

To map the distribution of DNase IIβ activity in the lens, we used two techniques (Fig. 3) . First, lenses from Wt and DNase IIβ-null animals were dissected into epithelial, cortical, and core fractions. The acid nuclease activity in each fraction was measured by semiquantitative nuclease activity assay. In both Wt and DNase IIβ-null lysates, acid nuclease activity was greatest in the epithelial sample and least in the core sample (Figs. 3A 3B 3C) . In the epithelium, the nuclease activities of Wt and DNase IIβ-null samples were statistically indistinguishable. In contrast, more than 50% of the acid nuclease activity of the cortical and core samples was due to DNase IIβ. To refine the tissue distribution of DNase IIβ activity further, we used a tissue-imprinting assay (Fig. 3D) . Acid nuclease activity was detected throughout the Wt samples, with the exception of a small region in the center of the lens (Fig. 3D) . When incubated under identical conditions, eye tissue from DNase IIβ-null animals contained significantly less acid nuclease activity, and little DNA degradation was noted in regions of the substrate underlying the lens. These data confirm the observation that DNase IIβ contributes significantly to the acid nuclease activity in Wt lenses, where it is present throughout the fiber cell mass with the exception of the oldest cells located in the center of the tissue. 

Regional dissections or tissue imprinting assays (Fig. 3)demonstrated that DNase IIβ activity was maximum in the cortex, the region of the adult lens containing cells undergoing organelle breakdown. However, organelle degradation is first observed in the lens during late embryonic development. ¹⁸ To determine whether the appearance of DNase IIβ activity correlates with the onset of organelle breakdown, lens lysates were prepared from Wt and DNase IIβ-null embryos and newborns, age E12 to P5 (Fig. 4) . Assays were performed under acidic conditions (pH 5.9). Early in development (embryonic day [E]12–E14), comparable acid nuclease activity was noted in both Wt and DNase IIβ-null lenses. However, by E18, the activity in the Wt lysates significantly exceeded that of the DNase IIβ-null lysates, suggesting that DNase IIβ was first activated in the lens in the interval E14 to E18. By E18, DNase IIβ constituted ∼50% of lens acid nuclease activity. A similar contribution was noted in the lenses of P5 (Fig. 4)and 1-month-old mice (Fig. 2) , in each of which DNase IIβ accounted for approximately half of the total acid nuclease activity. 

We examined the phenotype of DNase IIβ-null lenses by confocal microscopy. In Wt (Fig. 5A)and DNase IIβ-null (Fig. 5B)lenses, the nuclei of young fibers (i.e., those located near the lens surface) were oval-shaped with uncondensed chromatin (Fig. 5C) . Later, the nuclei became more spherical (Fig. 5D)and, shortly before the disappearance of nuclei in Wt lenses, particulate, propidium-stained material (presumably comprised of partially degraded DNA) was released from the nucleus into the perinuclear cytoplasm (Fig. 5E , arrow). Within the OFZ, small foci of amorphous propidium-stained material were often present (Fig. 5F) . In lenses from DNase IIβ-null animals, the denucleation process was inhibited. As a result, nuclei were present throughout the fiber cell mass, even in the lens core (Fig. 5B) , as noted previously. ¹² The discharge of propidium-stained material from fiber cell nuclei, a feature of nuclear degradation in Wt cells (Fig. 5E) , was not observed. Nuclei persisted in centrally located fiber cells in the DNase IIβ-null lens but marked and progressive changes in the structure of the retained nuclei were noted, including globularization of chromatin (Fig. 5G)and, in the innermost cells, frank nuclear disintegration (Fig. 5H) . 

Numerous studies in various animal models have demonstrated that fiber denucleation is preceded by DNA strand scission and accumulation of 3′-OH DNA termini. ^{1 3 4 19 20} The presence of nuclei in the central fiber cells of DNase IIβ-null lenses indicated that DNase IIβ was absolutely required for normal denucleation. Paradoxically, however, DNase IIβ digestion results in the production of 3′-PO₄ ⁻ DNA ends rather than 3′-OH. ¹⁴ To reconcile these apparently contradictory observations, the integrity of nuclear DNA was analyzed in differentiating Wt or DNase IIβ-null fiber cells by using a modified TUNEL assay (Fig. 6) . In this assay, 3′-OH ends were identified using fluorescein-labeled dUTPs with or without alkaline phosphatase pretreatment. The TUNEL assay confirmed the presence of 3′-OH DNA ends in Wt fiber cells bordering the OFZ (Fig. 6B) . In contrast, no 3′-OH labeling was observed in DNase IIβ-null lenses (Fig. 6D) . This observation indicated that DNase IIβ activity was required for the accumulation of 3′-OH ends in the disintegrating fiber nuclei. Wt lens sections pretreated with alkaline phosphatase before TUNEL labeling were not labeled more strongly than untreated cells (compare Fig. 6Bwith 6F), suggesting that the number of 3′-PO₄ ⁻ ends present was negligible compared with the number of 3′-OH ends. A plausible explanation for these observations was that during fiber denucleation, DNase IIβ activity generates DNA fragments with 3′-PO₄ ⁻ ends but that these are subsequently (and rapidly) converted to 3′-OH ends by the action of endogenous phosphatases. To test this hypothesis, we developed an in tube labeling assay that used a nuclease activity assay (Fig. 6I)in conjunction with a 3′-OH end labeling reaction (Fig. 6J) . Cortical lysates from Wt lenses were mixed under acidic conditions with an exogenous DNA substrate, resulting in degradation of the substrate (Fig. 6I) . The 3′-OH ends in the degraded DNA were then labeled using the Klenow fragment. The addition of a broad-spectrum phosphatase inhibitor to the lysate had no effect on the extent of DNA degradation (Fig. 6I)but reduced the number of labeled 3′-OH ends significantly (Figs. 6J 6K) . These data supported the notion that DNA degradation was accomplished through the actions of DNase IIβ that cut the DNA to produce 3′-PO₄ ⁻ ends, which were converted subsequently to 3′-OH ends by an endogenous phosphatase(s). 

The acidic pH optimum of DNase IIβ (Fig. 2B)suggests that DNase IIβ may reside in an acidic compartment within the lens fiber cells. Efforts to localize DNase IIβ at the subcellular level have been hindered by the lack of suitable antibodies. As an alternative approach, we used a cell fractionation protocol to analyze the subcellular distribution of DNase IIβ activity in lens fiber cells. Cytosolic and membrane fractions were prepared from cortical and core regions of Wt and DNase IIβ-null lenses. In cortical samples, ∼65% of the DNase IIβ activity present in the homogenate was removed on pelleting the membranes (Fig. 7) . Expressed per milligram of protein, the DNase IIβ activity associated with the membrane fraction was >20-fold greater than the cytosolic fraction. In contrast, DNase IIβ-specific activity was more evenly distributed between the membrane and cytoplasmic fractions in samples prepared from the lens core. 

Its acidic pH optimum suggests that, like DNase IIα, DNase IIβ may be a lysosomal enzyme. The distribution and abundance of lysosomes within the mouse lens has not been established previously; therefore, confocal microscopy was used to visualize lysosomes in intact, living lenses (Fig. 8) . Lysosomes were found to be most abundant in the lens epithelium. Lysosomes were also present at a lower density in the superficial fiber cells but were absent from the inner cortical and core fiber cells. Thus, the distribution of lysosomes in the lens paralleled that of the acid nuclease activity, which was highest in the epithelium and lowest in the lens core (Fig. 3) . 

To examine the subcellular distribution of DNase IIβ, a light mitochondrial fraction was obtained by density-gradient centrifugation of homogenates from Wt or DNase IIβ-null lenses. Two subpopulations of lysosomes (the light and dense fractions) were separated from the light mitochondrial fraction. ²¹ Soluble lysosomal proteins, such as β-glucuronidase and β-galactosidase, are present in both light and dense lysosomal fractions, whereas lysosome-associated membrane proteins (LAMPs) and acid phosphatase are present predominantly in the light lysosomal fraction. ²² The distributions of calnexin (an ER marker), succinate-ubiquinole oxidoreductase (a mitochondrial marker), and β-glucuronidase (a lysosomal marker) were determined (Fig. 9A 9B) . β-glucuronidase activity was highest in two peaks, consisting of fractions 3 to 6 (σ1.094–1.122 g/mL; light lysosomal fraction) and fraction 11 (σ1.171 g/mL; dense lysosomal fraction). Calnexin was found in fractions 1 to 5, partially overlapping in distribution with the light lysosomes. The mitochondrial marker succinate-ubiquinol-oxidoreductase was most abundant in fractions 4 to 7 (σ1.107–1.121 g/mL). DNase IIβ-specific activity was present in all fractions, although the activity varied approximately fourfold between the fraction with highest activity (F7) and the fraction with the least (F3). We discerned two broad activity peaks, corresponding to fractions 4 to 7 and 9 to 11, that paralleled the distribution of β-glucuronidase activity. DNase IIβ activity was also high in fraction 2 (Fig. 9C) , in which calnexin, the endoplasmic reticulum marker was located. These results are consistent with the notion that, as with DNase IIα, DNase IIβ is located primarily in lysosomes, although there may also be some association with elements of the endoplasmic reticulum. 

Discussion

Denucleation, first reported by Rabl, ²³ is a striking feature of lens fiber differentiation and the biochemical processes underlying this phenomenon have attracted considerable attention. To date, the only nuclease implicated directly in this process is DNase IIβ. ¹² In this study, we examined the temporal and spatial expression of DNase IIβ activity and its subcellular localization in the mouse lens. 

Early in embryonic development (<E14) acid nuclease activity was evident in fiber cells. This may be attributable to DNase IIα activity because, at the early embryonic stages, DNase IIβ-specific activity was undetectable. The appearance of DNase IIβ activity (at E14–E18) correlated well with the reported onset of organelle breakdown in the mouse lens, which commences at E17 to E18, when rounded nuclei with condensed chromatin are first observed. ¹⁸ Similarly, DNase IIβ mRNA is present initially in the lens at low levels and peaks at E17.5. ¹² Thus, increases in both DNase IIβ mRNA levels and DNase IIβ-specific nuclease activity presage denucleation in the embryonic lens. This pattern continues in the adult lens, where QRT-PCR analysis revealed a 200-fold increase in DNase IIβ transcript abundance during lens fiber cell differentiation. An increase in DNase IIβ mRNA abundance, albeit of smaller magnitude (twofold), has also been reported in cells bordering the OFZ. ²⁴ The increase in DNase IIβ mRNA is particularly noteworthy when compared with the downregulation of the related nuclease, DNase IIα, in the same cells over the same period. In the lens epithelium, the transcript levels for the two genes were approximately equal. In the fiber cells, however, the ratio of DNase IIα to DNase IIβ was 1:8000. In view of this large shift in ratio, it was somewhat surprising that DNase IIβ activity comprised only half of the acidic nuclease activity in the lens fiber cells and that the total acidic nuclease activity in the fibers was less than that of the lens epithelium. There was thus no clear relationship between the abundance of DNase IIβ mRNA and nuclease activity. 

The acidic pH optimum of DNase IIβ, its distribution within the lens, its association with the lysosome-enriched membrane fraction and its sequence similarity to DNase IIα (a lysosomal enzyme) together support the notion that, at least in the superficial lens fiber cells, DNase IIβ is primarily a lysosomal enzyme. The unavailability of specific antibodies against DNase IIβ has made it difficult to test this hypothesis directly. However, when expressed in HeLa S3 cells, chimeric DNase IIβ-GFP protein was distributed in a granular pattern consistent with a lysosomal location. ¹⁴ The mechanism by which a lysosomal nuclease might gain access to the fiber cell chromatin remains obscure. In other systems, entire organelles are sometimes degraded within specialized lysosomes through the process of autophagy. However, the programmed degradation of nuclei and other organelles appears to occur normally in Atg5^−/− mice, which are autophagy deficient. ²⁵ It is possible that DNase IIβ is released into the cytoplasm during organelle degradation. This notion is consistent with the observed redistribution of DNase IIβ activity from the membrane fraction to the cytoplasmic fraction in centrally located fiber cells. Although the pH optimum of DNase IIβ is <6.0, previous studies with recombinant enzyme have shown that residual enzymatic activity is present at pH as high as 6.8. ¹⁴ In the lens, a standing gradient of intracellular pH (pH_i) is generated as a result of its predominantly glycolytic metabolism. ²⁶ Consequently, pH_i in the inner fibers may be 0.5 pH units lower than at the lens surface. Thus, DNase IIβ may retain some activity at the cytoplasmic pH encountered in differentiating fiber cells. In the present study, however, we were unable to discern specific DNase IIβ activity at pH 6.5 or higher. This may reflect the fact that the measurements were derived from differences in activity between samples from wild-type and knockout lenses. This approach probably lacked the resolution to detect small differences in nuclease activity at pH far removed from the optimum. It remains possible, therefore, that a decrease in pH_i in cells located near the border of the OFZ may play a role in modulating the activity of DNase IIβ released into the fiber cell cytoplasm. 

In contrast to Wt lenses, nuclei were retained in the central fiber cells of lenses from DNase IIβ-null mice, confirming earlier reports. ¹² No TdT labeling (with or without alkaline phosphatase) was observed in DNase IIβ-null mice, indicating that either DNase IIβ is the only nuclease involved in fiber cell denucleation or that DNase IIβ activity is necessary for the activity of other, downstream nucleases. Although the rapid degradation of nuclei did not occur in the absence of DNase IIβ, structural alterations (including blebbing and disintegration) were noted in the persistent nuclei. These morphologic changes may reflect the activity of other nucleases in the central cells but could also be secondary to loss of the karyoskeleton and nuclear membrane. The latter appears to be degraded normally in the lenses of DNase IIβ-null mice. ¹²  

In searching for the nuclease(s) involved in fiber cell denucleation, much attention focused previously on the nature of DNA termini produced by digestion with various nucleases. Studies have consistently found that 3′-OH ends accumulate in the DNA of disintegrating fiber cell nuclei, leading to speculation that DNase I, an enzyme that generates 3′-OH ends, might be implicated. ^{2 27} It is intriguing, therefore, that such termini do not accumulate in the absence of DNase IIβ, an enzyme known to produce 3′-PO₄ ⁻ ends exclusively. Our data suggested that this discrepancy is most readily explained by the action of endogenous phosphatases, a number of which have been identified in the lens. ^{28 29} Of interest, dephosphorylation of DNA ends occurred under the acidic conditions of the nuclease assay (Fig. 6I) , suggesting the involvement of an acid phosphatase. The coordinated translocation of DNase IIα and acid phosphatase from lysosome-to-nucleus has been observed in human myeloid leukemia cells after γ-irradiation. ²⁹ Perhaps in the lens also the appearance of 3′-OH ends in the fiber cell DNA may be due to the coordinated activity of two enzymes (in this case, DNase IIβ and acid phosphatase) normally resident in the lysosome. 

A very recent paper, published after the present studies were concluded, supports the notion that DNase IIβ is primarily a lysosomal enzyme. Using an antibody raised to recombinant DNase IIβ, Nakahara et al., ³⁰ have been able to visualize the distribution of DNase IIβ within the lens and have proposed an interesting model in which lysosomes fuse with the fiber cell nucleus before chromatin degradation. 

 Supported by National Eye Institute Grants R01EY009852 and EY02687 (Core grant for vision research), and an unrestricted grant to the Department of Ophthalmology and Visual Sciences from Research to Prevent Blindness (RPB). SB is a William and Mary Greve RPB scholar.

 Submitted for publication June 26, 2007; revised July 24, 2007; accepted October 15, 2007.

 Disclosure: A. De Maria, None; S. Bassnett, None

 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.

 

Analysis of DNase IIα and DNase IIβ transcripts in epithelial and fiber cells. (A) End-point PCR. (B) QRT-PCR. Although equally abundant in the lens epithelium, DNase IIα and DNase IIβ transcripts were differentially regulated during fiber cell differentiation. There was a 40-fold decrease in the level of DNase IIα mRNA in the cortex compared with the epithelium. In contrast, the level of DNase IIβ mRNA increased ∼200-fold. Quantitative data represent the mean ± SD (n = 4). Ep, epithelium; Cx, cortex.

Figure 2.

 

DNase IIβ-specific activity in lens lysates. (A) Representative experiment showing the rate of degradation of a DNA substrate in the presence of lens lysates under acidic (pH 5.9) conditions. The DNA substrate was degraded more quickly by lens lysates from Wt animals than by lysates from DNase IIβ-null animals. (B) Quantitative analysis of lens nuclease activity as a function of pH. DNase IIβ activity (the difference in activity between samples obtained from Wt and DNase IIβ-null lenses) was maximal at pH 5.9 (where it represented ∼50% of the nuclease activity) and was undetectable at pH 6.5 and higher. Data represent the mean ± SD (n = 4).

Figure 3.

 

Localization of DNase IIβ activity in the lens. Nuclease activity assays were performed (at pH 5.9) on lysates from Wt (A) or DNase IIβ-null (B) lenses. When mixed with epithelial lysates from Wt or DNase IIβ-null lenses, the DNA substrate degraded rapidly. Acid nuclease activity was greater in the cortical fibers of Wt lenses than DNase IIβ-null lenses. Acid nuclease activity was present in Wt core lysates but was barely detectable in core lysates from DNase IIβ-null lenses. (C) Quantitative analysis of acid nuclease activity in various regions of Wt and DNase IIβ-null lenses. *Statistically significant differences between samples. Data represent the mean ± SD (n = 3). (D) Tissue-imprinting assay in which tissue sections were overlaid on an immobilized DNA substrate. The lighter areas correspond to regions of relatively high acid nuclease activity. Dashed circle: position of the lens. In Wt lenses, nuclease activity was highest in the lens cortex and was absent from a region in the center of the tissue (arrow). The nuclease activity in the vitreous humor (VH) of Wt samples is presumably due to DNase IIβ diffusing from the lens tissue during the incubation period because it was absent in sections of DNase IIβ-null tissue. Sections are representative of four independent experiments.

Figure 4.

 

Developmental expression of DNase IIβ-specific acid nuclease activity. Nuclease activity was measured in lens lysates prepared from Wt or DNase IIβ-null mice of varying ages, by using a semiquantitative nuclease assay at pH 5.9. At each age, the difference in nuclease activity between samples from the two genotypes reflects specific DNase IIβ activity. *Samples in which the nuclease activity differed significantly between genotypes. Note the emergence of DNase IIβ activity between E14 and E18. At E18 and P5, DNase IIβ constitutes ∼50% of the acid nuclease activity. Data represent the mean ± SD (n = 4).

Figure 5.

 

Fate of lens fiber cell nuclei in Wt and DNase IIβ-null mice. Sections of 3-week-old mouse lenses were stained with PI and imaged by confocal microscopy. (A) In Wt lenses, nuclei were evident in the overlying epithelium (Ep) and superficial fiber cells but were lost from cells bordering the OFZ. (B) In contrast, nuclei were present throughout the fiber cell mass in lenses from DNase IIβ-null mice. Progressive changes in chromatin organization were observed in nuclei located at various depths in the lens (corresponding to the positions indicated in A and B). (C) At the periphery, nuclei were elongated. (D) In the deep cortical region, the nuclei had a more spherical shape. (E) In Wt cells bordering the OFZ, propidium-stained material was released from the nucleus into the perinuclear cytoplasm (arrow). (F) Within the OFZ, PI-stained foci were visible in the Wt lens. (G) Nuclei present in the central region of DNase IIβ-null lenses were often globularized (G) or fragmented (H). Data are representative of sections examined from three lenses for each genotype.

Figure 6.

 

The nature of DNA termini produced in Wt or DNase IIβ-null mice. Sections were prepared from Wt (A, B, E, F) or DNase IIβ-null (C, D, G, H) mouse lenses. Some sections were stained with PI (red) to visualize the distribution of nuclei in the lens tissue. Other sections were TUNEL labeled (green), with (F, H) or without (B, D) alkaline phosphatase pretreatment. In Wt lenses nuclei were absent from the OFZ (A, E). During nuclear breakdown, the chromatin became TUNEL positive (B, arrow), signifying the accumulation of 3′-OH ends in the disintegrating nuclei. Pretreatment of Wt lens sections with alkaline phosphatase was not associated with an increase in TUNEL labeling (F), suggesting that 3′-PO₄ ⁻ ends, if present, are much less abundant than 3′-OH ends. In sections from DNase IIβ lenses nuclei were present throughout the lens (C) and TUNEL-labeled cells were not observed (D), even after alkaline phosphatase pretreatment (H). (I) Lysates prepared from Wt lenses were mixed with plasmid DNA. In control samples (C) the full length plasmid was present (arrowhead). Addition of Wt lens lysates (L) alone or with phosphatase inhibitor (L+I) led to degradation of the plasmid substrate. (J) In-tube labeling of degraded DNA with the Klenow enzyme suggests that inclusion of a phosphatase inhibitor reduces (by ∼twofold) the number of 3′-OH termini produced during DNA degradation. (K) Densitometric analysis of DNA 3′-OH termini labeled by the Klenow enzyme. Data represent mean ± SD (n = 3).

Figure 7.

 

Changes in DNase IIβ subcellular localization during fiber differentiation. The DNase IIβ-specific activity associated with the cortical membrane fraction was ∼20-fold greater than in the cytoplasmic fraction. In contrast, in core fibers, the activity was more evenly distributed between the membrane and cytosolic fractions. Data represent the mean ± SD (n = 4).

Figure 8.

 

Distribution of lysosomes in the mouse lens. Confocal x-z image of the anterior region of a living lens. (A) Lysosomes (green) were abundant in the lens epithelium (*) and present, at lower levels, in the underlying fiber cells. They were not detected in cells of the deep lens cortex. (B) The membranes of the fiber cells (red) are shown in cross-section. (C) Overlay of the green and red fluorescence signals.

Figure 9.

 

Subcellular localization of DNase IIβ in cortical fiber cells. Subcellular organelles were separated into 12 fractions (F1–F12) by density centrifugation. (A) Western blot of succinate-ubiquinol-oxidoreductase (a mitochondrial marker) and calnexin (ER marker) distribution in F1 to F12. (B) Normalized distribution of succinate-ubiquinol-oxidoreductase, calnexin, and β-glucuronidase (a lysosomal marker) activity. Two subpopulations of lysosomes were present (F3–F7 and F11). (C) Quantification of DNase IIβ-specific activity. DNase IIβ activity (the difference in acid nuclease activity between wild-type and DNase IIβ-null samples) was present throughout the gradient with apparent peaks in F2, F5 to F7, and F9 to F11. Data represent mean ± SD (n = 4).

The authors thank Stuart Kornfeld of Washington University for his advice on lysosome purification.