ALDH3A1 Plays a Functional Role in Maintenance of Corneal Epithelial Homeostasis

Aldehyde dehydrogenase 1A1 (ALDH1A1) and ALDH3A1 are corneal crystallins. They protect inner ocular tissues from ultraviolet radiation (UVR)-induced oxidative damage through catalytic and non-catalytic mechanisms. Additionally, ALDH3A1 has been postulated to play a regulatory role in the corneal epithelium based on several studies that report an inverse association between ALDH3A1 expression and corneal cell proliferation. The underlying molecular mechanisms and the physiological significance of such association remain poorly understood. In the current study, we established Tet-On human corneal epithelial cell (hTCEpi) lines, which express tetracycline-inducible wild-type (wt) or catalytically-inactive (mu) ALDH3A1. Utilizing this cellular model system, we confirmed that human ALDH3A1 decreases corneal cell proliferation; importantly, this effect appears to be partially mediated by its enzymatic activity. Mechanistically, wt-ALDH3A1, but not mu-ALDH3A1, promotes sequestering of tumor suppressor p53 in the nucleus. In the mouse cornea, however, augmented cell proliferation is noted only in Aldh1a1^-/-/3a1^-/- double knockout (DKO) mice, indicating in vivo the anti-proliferation effect of ALDH3A1 can be rescued by the presence of ALDH1A1. Interestingly, the hyper-proliferative epithelium of the DKO corneas display nearly complete loss of p53 expression, implying that p53 may be involved in ALDH3A1/1A1-mediated effect. In hTCEpi cells grown in high calcium concentration, mRNA levels of a panel of corneal differentiation markers were altered by ALDH3A1 expression and modulated by its enzyme activity. In conclusion, we show for the first time that: (i) ALDH3A1 decreases corneal epithelial proliferation through both non-enzymatic and enzymatic properties; (ii) ALDH1A1 contributes to the regulation of corneal cellular proliferation in vivo; and (iii) ALDH3A1 modulates corneal epithelial differentiation. Collectively, our studies indicate a functional role of ALDH3A1 in the maintenance of corneal epithelial homeostasis by simultaneously modulating proliferation and differentiation through both enzymatic and non-enzymatic mechanisms.     

Collagen XVIII and corneal reinnervation following keratectomy

The significance of collagen XVIII in the regulation of corneal reinnervation remains largely unknown. We used whole-mount immunoconfocal microscopy to localize collagen XVIII to the nerve basement membrane of wild-type (WT) mouse corneas. Transmission electron microscopy showed corneal nerve disorganization in collagen XVIII knockout mice (col18a1(-/-)). Antibody 2H3-specific neurofilament colocalized with collagens XVIII and IV and laminin-2 in WT mouse corneas, but did not colocalize with collagen IV and laminin-2 in col18a1(-/-) mouse corneas. Following keratectomy, col18a1(-/-) mice displayed decreased corneal neurite extension compared to WT mice. Our data indicate that collagen XVIII may play an important role in corneal reinnervation after wounding.     

Corneal and stomach expression of aldehyde dehydrogenases: from fish to mammals

We have studied the distribution of the ALDH3A1, ALDH1A1 and ALDH2 proteins in the cornea and stomach of several animal species, including mammals (C57BL/6J and SWR/J mice, rat and pig), birds (chicken and turkey), amphibians (frog) and fish (trout and zebrafish). High ALDH3A1 protein levels and catalytic activities were detected in C57BL/6J mouse, rat and pig. We found complete absence of the ALDH3A1 protein in SWR/J mice, which carry the Aldh3a1(c) allele characterized by four amino acid substitutions (G88R, I154N, H305R and I352V) and lack of enzymatic activity. This indicates that the SWR/J mouse strain is a natural gene knockout model for ALDH3A1. Traces of ALDH3A1 were detected in rabbit, whereas expression was absent from chicken, turkey, frog, trout, and zebrafish. Interestingly, significant levels of the cytosolic ALDH1A1 and mitochondrial ALDH2 proteins were detected by immunoblot analysis in all examined species that are deficient in ALDH3A1 expression. In contrast, no ALDH1A1 or ALDH2 protein was detected in the species expressing ALDH3A1. It can, therefore, be concluded that corneal expression of ALDH3A1 or ALDH1A1/ALDH2 occurs in a taxon-specific manner, supporting the protective role of these ALDHs in cornea against the UV-induced oxidative damage.     

Corneal and stomach expression of aldehyde dehydrogenases: from fish to mammals

We have studied the distribution of the ALDH3A1, ALDH1A1 and ALDH2 proteins in the cornea and stomach of several animal species, including mammals (C57BL/6J and SWR/J mice, rat and pig), birds (chicken and turkey), amphibians (frog) and fish (trout and zebrafish). High ALDH3A1 protein levels and catalytic activities were detected in C57BL/6J mouse, rat and pig. We found complete absence of the ALDH3A1 protein in SWR/J mice, which carry the Aldh3a1^c allele characterized by four amino acid substitutions (G88R, I154N, H305R and I352V) and lack of enzymatic activity. This indicates that the SWR/J mouse strain is a natural gene knockout model for ALDH3A1. Traces of ALDH3A1 were detected in rabbit, whereas expression was absent from chicken, turkey, frog, trout, and zebrafish. Interestingly, significant levels of the cytosolic ALDH1A1 and mitochondrial ALDH2 proteins were detected by immunoblot analysis in all examined species that are deficient in ALDH3A1 expression. In contrast, no ALDH1A1 or ALDH2 protein was detected in the species expressing ALDH3A1. It can, therefore, be concluded that corneal expression of ALDH3A1 or ALDH1A1/ALDH2 occurs in a taxon-specific manner, supporting the protective role of these ALDHs in cornea against the UV-induced oxidative damage.     

Molecular cloning and baculovirus expression of the rabbit corneal aldehyde dehydrogenase (ALDH1A1) cDNA

Most mammalian species express high concentrations of ALDH3A1 in corneal epithelium with the exception of the rabbit, which expresses high amounts of ALDH1A1 rather than ALDH3A1. Several hypotheses that involve catalytic and/or structural functions have been postulated regarding the role of these corneal ALDHs. The aim of the present study was to characterize the biochemical properties of the rabbit ALDH1A1. We have cloned and sequenced the rabbit ALDH1A1 cDNA, which is 2,073 bp in length (excluding the poly(A+) tail), and has 5' and 3' nontranslated regions of 46 and 536 bp, respectively. This ALDH1A1 cDNA encodes a protein of 496 amino acids (Mr = 54,340) that is: 86-91% identical to mammalian ALDH1A1 proteins, 83-85% identical to phenobarbital-inducible mouse and rat ALDH1A7 proteins, 84% identical to elephant shrew ALDH1A8 proteins (eta-crystallins), 69-73% identical to vertebrate ALDH1A2 and ALDH1A3 proteins, 65% identical to scallop ALDH1A9 protein (omega-crystallin), and 55-57% to cephalopod ALDH1C1 and ALDH1C2 (omega-crystallins). Recombinant rabbit ALDH1A1 protein was expressed using the baculovirus system and purified to homogeneity with affinity chromatography. We found that rabbit ALDH1A1 is catalytically active and efficiently oxidizes hexanal (Km = 3.5 microM), 4-hydroxynonenal (Km = 2.1 microM) and malondialdehyde (Km = 14.0 microM), which are among the major products of lipid peroxidation. Similar kinetic constants were observed with the human recombinant ALDH1A1 protein, which was expressed and purified using similar experimental conditions. These data suggest that ALDH1A1 may contribute to corneal cellular defense against oxidative damage by metabolizing toxic aldehydes produced during UV-induced lipid peroxidation.     

Galectins-3 and -7, but not galectin-1, play a role in re-epithelialization of wounds

Disorders of wound healing characterized by impaired or delayed re-epithelialization are a serious medical problem. These conditions affect many tissues, are painful, and are difficult to treat. In this study, using cornea as a model, we demonstrate for the first time the importance of carbohydrate-binding proteins galectins-3 and -7 in re-epithelialization of wounds. In two different models of corneal wound healing, re-epithelialization of wounds was significantly slower in galectin-3-deficient (gal3(-/-)) mice compared with wild-type (gal3(+/+)) mice. In contrast, there was no difference in corneal epithelial wound closure rates between galectin-1-deficient and wild-type mice. Quantitation of the bromodeoxyuridine-labeled cells in gal3(+/+) and gal3(-/-) corneas revealed that corneal epithelial cell proliferation rate is not perturbed in gal3(-/-) corneas. Exogenous galectin-3 accelerated re-epithelialization of wounds in gal3(+/+) mice but, surprisingly, not in the gal3(-/-) mice. Gene expression analysis using cDNA microarrays revealed that healing corneas of gal3(-/-) mice contain markedly reduced levels of galectin-7 compared with those of gal3(+/+) mice. More importantly, unlike galectin-3, galectin-7 accelerated re-epithelialization of wounds in both gal3(-/-) and gal3(+/+) mice. In corresponding experiments, recombinant galectin-1 did not stimulate the corneal epithelial wound closure rate. The extent of acceleration of re-epithelialization of wounds with both galectin-3 and galectin-7 was greater than that observed in most of the published studies using growth factors. These findings have broad implications for developing novel therapeutic strategies for treating nonhealing wounds.     

Impairment of thrombospondin-1 expression during epithelial wound healing in corneas of vitamin A-deficient mice

The purpose of this study is to investigate the expression of thrombospondin-1 (TSP-1), a multifunctional extracellular matrix protein, during re-epithelialization in wounded corneas of vitamin A-deficient mice. Epithelial defects were created in the corneas of normal and Vitamin A-deficient mice with a microgrinder. Wounded corneas were stained with fluorescein and photographed for evaluation of re-epithelialization. Histological examination and immunohistochemical analysis of TSP-1 expression were also performed on the specimens from wounded corneas. In vitamin A-deficient mice, re-epithelialization of the wounded corneal epithelium was significantly delayed compared with that in normal mice. TSP-1 was detectable neither in the unwounded corneal epithelium of normal mice nor in that of vitamin A-deficient mice. In normal mice, linear staining of TSP-1 was observed on the wounded corneal surface and stroma at 30 min and 8 h to 16 h, respectively, after abrasion, and this TSP-1 expression disappeared at 36 to 48 h, when re-epithelialization was completed. In contrast, no TSP-1 staining was observed in the wounded corneas of vitamin A-deficient mice, except for the endothelial cells, throughout the wound healing process. Histological examination revealed a progressive increase in polymorphonuclear neutrophil infiltration in the stroma of the corneas of vitamin A-deficient mice during the healing process. These findings suggest that vitamin A may modulate the expression of TSP-1 in the corneas to accelerate the re-epithelialization of wounded corneas.     

Multiple and additive functions of ALDH3A1 and ALDH1A1: cataract phenotype and ocular oxidative damage in Aldh3a1(-/-)/Aldh1a1(-/-) knock-out mice

ALDH3A1 (aldehyde dehydrogenase 3A1) is abundant in the mouse cornea but undetectable in the lens, and ALDH1A1 is present at lower (catalytic) levels in the cornea and lens. To test the hypothesis that ALDH3A1 and ALDH1A1 protect the anterior segment of the eye against environmentally induced oxidative damage, Aldh1a1(-/-)/Aldh3a1(-/-) double knock-out and Aldh1a1(-/-) and Aldh3a1(-/-) single knock-out mice were evaluated for biochemical changes and cataract formation (lens opacification). The Aldh1a1/Aldh3a1- and Aldh3a1-null mice develop cataracts in the anterior and posterior subcapsular regions as well as punctate opacities in the cortex by 1 month of age. The Aldh1a1-null mice also develop cataracts later in life (6-9 months of age). One- to three-month-old Aldh-null mice exposed to UVB exhibited accelerated anterior lens subcapsular opacification, which was more pronounced in Aldh3a1(-/-) and Aldh3a1(-/-)/Aldh1a1(-/-) mice compared with Aldh1a1(-/-) and wild type animals. Cataract formation was associated with decreased proteasomal activity, increased protein oxidation, increased GSH levels, and increased levels of 4-hydroxy-2-nonenal- and malondialdehyde-protein adducts. In conclusion, these findings support the hypothesis that corneal ALDH3A1 and lens ALDH1A1 protect the eye against cataract formation via nonenzymatic (light filtering) and enzymatic (detoxification) functions.     

Thrombospondin-1 accelerates wound healing of corneal epithelia

To investigate a role of thrombospondin-1 (TSP-1), a multifunctional extracellular matrix protein, in corneal epithelial wound healing, we analyzed the expression of TSP-1 in the normal and wounded mouse corneal epithelia and the effect of exogenous TSP-1 on the wound healing. In immunohistochemical analyses of unwounded corneas, TSP-1 was only detectable in endothelial cells. In contrast, TSP-1 appeared on the wounded corneal surface and on the corneal stroma, at 30 min and 8-16 h, respectively, after making an abrasion on the corneal epithelium. This expression of TSP-1 disappeared after 36-48 h, when re-epithelialization was completed. The TSP-1 mRNA level in the wounded corneas increased as much as three fold compared with that in the unwounded corneas. In organ culture, exogenous TSP-1 stimulated the re-epithelialization of corneal epithelial wounds whereas anti-TSP-1 antibody significantly inhibited the re-epithelialization. These findings suggest the possibility that epithelial defects in the corneas stimulate the expression of TSP-1 in the wound area, resulting in the accelerated re-epithelialization of the cornea.     

Induction of class 3 aldehyde dehydrogenase in the mouse hepatoma cell line Hepa-1 by various chemicals.

The mouse hepatoma cell line Hepa-1 was shown to express an aldehyde dehydrogenase (ALDH) isozyme which was inducible by TCDD and carcinogenic polycyclic aromatic hydrocarbons. The induced activity could be detected with benzaldehyde as substrate and NADP as cofactor (B/NADP ALDH). As compared with rat liver and hepatoma cell lines, the response was moderate (maximally 5-fold). There was an apparent correlation between this specific form of ALDH and aryl hydrocarbon hydroxylase (AHH) in the Hepa-1 wild-type cell line--in terms of inducibility by several chemicals. However, the magnitude of the response was clearly smaller for ALDH than for AHH. Southern blot analysis showed that a homologous gene (class 3 ALDH) was present in the rat and mouse genome. The gene was also expressed in Hepa-1 and there was a good correlation between the increase of class 3 ALDH-specific mRNA and B/NADP ALDH enzyme activity after exposure of the Hepa-1 cells to TCDD. It is concluded that class 3 ALDH is inducible by certain chemicals in the mouse hepatoma cell line, although the respective enzyme is not inducible in mouse liver in vivo.     

Identification of 3-deoxyglucosone dehydrogenase as aldehyde dehydrogenase 1A1 (retinaldehyde dehydrogenase 1)

One of the metabolic fates of 3-deoxyglucosone, a product of protein deglycation and a potent glycating agent, is to be oxidized to 2-keto-3-deoxygluconate, but the enzyme that catalyzes this reaction is presently unknown. Starting from human erythrocytes, which are known to convert 3-deoxyglucosone to 2-keto-3-deoxygluconate, we have purified to near homogeneity a NAD-dependent dehydrogenase that catalyzes this last reaction at neutral pH. Sequencing of a 55 kDa band co-eluting with the enzymatic activity in the last step indicated that it corresponded to aldehyde dehydrogenase 1A1 (ALDH1A1), an enzyme known to catalyze the oxidation of retinaldehyde to retinoic acid. Overexpression of human ALDH1A1 in HEK cells led to a more than 20-fold increase in 3-deoxyglucosone dehydrogenase activity. In mouse tissues 3-deoxyglucosone dehydrogenase activity was highest in liver, intermediate in lung and testis, and negligible or undetectable in other tissues, in agreement with the tissue distribution of ALDH1A1 mRNA. 3-deoxyglucosone dehydrogenase activity was undetectable in tissues from ALDH1A1(-/-) mice. ALDH1A1 appears therefore to be the major if not the only enzyme responsible for the oxidation of 3-deoxyglucosone to 2-keto-3-deoxygluconate. The urinary excretion of 2-keto-3-deoxygluconate amounted to 16.7 micromol/g creatinine in humans, indicating that 3-deoxyglucosone may be quantitatively a more important substrate than retinaldehyde for ALDH1A1.     

A role for chromosomal protein HMGN1 in corneal maturation

Abstract Corneal differentiation and maturation are associated with major changes in the expression levels of numerous genes, including those coding for the chromatin-binding high-mobility group (HMG) proteins. Here we report that HMGN1, a nucleosome-binding protein that alters the structure and activity of chromatin, affects the development of the corneal epithelium in mice. The corneal epithelium of Hmgn1 ^−/− mice is thin, has a reduced number of cells, is poorly stratified, is depleted of suprabasal wing cells, and its most superficial cell layer blisters. In mature Hmgn1 ^−/− mice, the basal cells retain the ovoid shape of immature cells, and rest directly on the basal membrane which is disorganized. Gene expression was modified in Hmgn1 ^−/− corneas: glutathione-S-transferase (GST)α 4and GST ω 1, epithelial layer-specific markers, were selectively reduced while E-cadherin and α-, β-, and γ-catenin, components of adherens junctions, were increased. Immunofluorescence analysis reveals a complete co-localization of HMGN1 and p63 in small clusters of basal corneal epithelial cells of wild-type mice, and an absence of p63 expressing cells in the central region of the Hmgn1 ^−/− cornea. We suggest that interaction of HMGN1 with chromatin modulates the fidelity of gene expression and affects corneal development and maturation.     

Alpha-enolase is restricted to basal cells of stratified squamous epithelium.

We have developed a monoclonal antibody against a 50-kDa protein that binds preferentially to basal cells in the limbus of rat, rabbit, and human corneas (J. D. Zieske, G. Bukusoglu, and M. A. Yankauckas, Invest. Ophthalmol. Visual Sci. 33, 143-152, 1992). Here we report on the purification and identification of the antigen. The 50-kDa antigen was purified from rabbit limbal and corneal epithelium using HPLC methodology including anion exchange (DEAE) followed by reverse-phase (C18) chromatography. The purified 50-kDa protein was then digested with endoproteinase Lys-C, and a reproducible profile comprising approximately 20 peptides was observed by reverse-phase HPLC of the digest. Sequence analysis of five peptides ranging in length from 4 to 20 residues revealed that the 50-kDa protein was alpha-enolase, a glycolytic enzyme. Overall, 57 amino acids were identified with a 95% sequence homology. Localization of alpha-enolase in rat epithelium by immunofluorescence microscopy demonstrated that simple epithelium contained low or undetectable levels of the enzyme. Stratified squamous epithelium, however, showed high levels of alpha-enolase, which was localized specifically to cells of the basal layer. Epidermal, corneal limbal, oral mucosal, vaginal, and laryngeal epithelium all showed cytoplasmic binding specific to the basal cells. These data indicate that the glycolytic enzyme alpha-enolase is preferentially localized in the basal cell layer of stratified squamous epithelium and suggest that glycolytic activity is concentrated in these cells. The localization pattern suggests that a major change in metabolism occurs as cells leave the mitotically active basal cell layer and migrate toward terminal differentiation in the suprabasal cell layers.     

Aldh3a1 protects human corneal epithelial cells from ultraviolet- and 4-hydroxy-2-nonenal-induced oxidative damage

Aldehyde dehydrogenase 3A1 (ALDH3A1) is one of the most abundant proteins found in corneal epithelial cells of mammalian species, with several postulated protective roles that include detoxification of peroxidic aldehydes, scavenging of free radicals, and direct absorption of ultraviolet (UV) radiation. In the present study, the protective role of ALDH3A1 against UV- and 4-hydroxy-2-nonenal- (4-HNE-) induced oxidative damage was studied. For this purpose, human ALDH3A1 was stably transfected in a human corneal epithelial cell line (HCE) lacking endogenous enzyme. Cells transfected with ALDH3A1 were more resistant to UV- and 4-HNE-induced cytotoxicity than mock-transfected cells. DNA fragmentation assays revealed that both treatments induced apoptosis in mock-transfected cells, but not in ALDH3A1-expressing cells. Apoptosis appeared to occur via caspase-3 activation and subsequent PARP cleavage. The Michaelis-Menten constant (K(m)) for 4-HNE was 54 microM in ALDH3A1-transfected cells; the addition of 100 microM 4-HNE increased NAD(P)H levels by 50% above that in mock-transfected cells. We also found that ALDH3A1 expression prevented 4-HNE-induced protein adduct formation. Taken together, these data suggest that ALDH3A1 is a regulatory element of the cellular defense system that protects corneal epithelium against UV- and 4-HNE-induced oxidative damage.     

Mechanisms involved in the protection of UV-induced protein inactivation by the corneal crystallin ALDH3A1

Various lines of evidence have shown that ALDH3A1 (aldehyde dehydrogenase 3A1) plays a critical and multifaceted role in protecting the cornea from UV-induced oxidative stress. ALDH3A1 is a corneal crystallin, which is defined as a protein recruited into the cornea for structural purposes without losing its primary function (i.e. metabolism). Although the primary role of ALDH3A1 in the metabolism of toxic aldehydes has been clearly demonstrated, including the detoxification of aldehydes produced during UV-induced lipid peroxidation, the structural role of ALDH3A1 in the cornea remains elusive. We therefore examined the potential contribution of ALDH3A1 in maintaining the optical integrity of the cornea by suppressing the aggregation and/or inactivation of other proteins through chaperone-like activity and other protective mechanisms. We found that ALDH3A1 underwent a structural transition near physiological temperatures to form a partially unfolded conformation that is suggestive of chaperone activity. Although this structural transition alone did not correlate with any protection, ALDH3A1 substantially reduced the inactivation of glucose-6-phosphate dehydrogenase by 4-hydroxy-2-nonenal and malondialdehyde when co-incubated with NADP(+), reinforcing the importance of the metabolic function of this corneal enzyme in the detoxification of toxic aldehydes. A large excess of ALDH3A1 also protected glucose-6-phosphate dehydrogenase from inactivation because of direct exposure to UVB light, which suggests that ALDH3A1 may shield other proteins from damaging UV rays. Collectively, these data demonstrate that ALDH3A1 can reduce protein inactivation and/or aggregation not only by detoxification of reactive aldehydes but also by directly absorbing UV energy. This study provides for the first time mechanistic evidence supporting the structural role of the corneal crystallin ALDH3A1 as a UV-absorbing constituent of the cornea.     

Structural and functional modifications of corneal crystallin ALDH3A1 by UVB light

As one of the most abundantly expressed proteins in the mammalian corneal epithelium, aldehyde dehydrogenase 3A1 (ALDH3A1) plays critical and multifaceted roles in protecting the cornea from oxidative stress. Recent studies have demonstrated that one protective mechanism of ALDH3A1 is the direct absorption of UV-energy, which reduces damage to other corneal proteins such as glucose-6-phosphate dehydrogenase through a competition mechanism. UV-exposure, however, leads to the inactivation of ALDH3A1 in such cases. In the current study, we demonstrate that UV-light caused soluble, non-native aggregation of ALDH3A1 due to both covalent and non-covalent interactions, and that the formation of the aggregates was responsible for the loss of ALDH3A1 enzymatic activity. Spectroscopic studies revealed that as a result of aggregation, the secondary and tertiary structure of ALDH3A1 were perturbed. LysC peptide mapping using MALDI-TOF mass spectrometry shows that UV-induced damage to ALDH3A1 also includes chemical modifications to Trp, Met, and Cys residues. Surprisingly, the conserved active site Cys of ALDH3A1 does not appear to be affected by UV-exposure; this residue remained intact after exposure to UV-light that rendered the enzyme completely inactive. Collectively, our data suggest that the UV-induced inactivation of ALDH3A1 is a result of non-native aggregation and associated structural changes rather than specific damage to the active site Cys.     

Integrin expression during epithelial migration and restratification in the tenascin-C-deficient mouse cornea.

In the unwounded cornea, tenascin-C localizes to a short stretch of the basement membrane zone at the corneoscleral junction or limbus. To determine whether the function of the limbus is affected by the absence of tenascin-C, mice possessing a deletion of tenascin-C and strain-matched wild-type mice are used in corneal debridement wounding experiments. The expression of integrins (alpha3, alpha9, and beta4) in the tenascin-C knockout corneas is evaluated by producing polyclonal cytoplasmic domain antipeptide sera and performing immunofluorescence microscopy. In addition, we evaluate the localization of several other proteins involved in wound healing, including fibronectin, laminin beta1, nidogen/entactin, and VCAM-1, in both the tenascin knockout and wild-type mice. There are no differences in healing rate, scarring, or neovascularization after corneal debridement wounds. alpha9 integrin is expressed at the limbal border of unwounded tenascin-C knockout animals and is upregulated during migration only after the larger wounds. At 8 weeks after larger wounds, the localization of alpha9 again becomes restricted to the limbal border. Results show that tenascin-C is not required for development or maintenance of the corneal limbus or for normal re-epithelialization of corneal epithelial cells after debridement wounding.