High urea and NaCl carbonylate proteins in renal cells in culture and in vivo, and high urea causes 8-oxoguanine lesions in their DNA

Abstract

Urea and NaCl are elevated in the renal inner medulla. We now find that a high concentration of urea or NaCl increases reactive oxygen species (ROS) in mouse renal inner medullary (mIMCD3) cells in culture. Previously, high NaCl, but not high urea, was found to cause DNA double-strand breaks. We now tested whether high urea or NaCl causes oxidative damage to DNA or cellular proteins. We find that high urea increases mIMCD3 cell DNA single-strand breaks and 8-oxoguanine lesions. High NaCl does not cause detectable 8-oxoguanine lesions. High urea or NaCl also greatly increases carbonylation of proteins in mIMCD3 cells. Carbonylation occurs within 5 min and with as little as 5 mM urea, a normal plasma level. It increases as urea is raised over the range in uremia. A high raffinose level increases ROS and carbonylation. High sorbitol and glycerol levels do not increase ROS or carbonylation. Carbonyl content is high in mouse renal inner medullas where interstitial NaCl and urea concentrations are normally high. There, numerous proteins are carbonylated, and carbonylation occurs in both collecting ducts and thin limbs. Conclusions: (i) Oxidative stress, associated with high urea, causes 8-oxoguanine DNA lesions in mIMCD3 cell DNA. (ii) High urea or NaCl carbonylates proteins in mIMCD3 cells and in renal inner medullary cells in vivo. (iii) In mIMCD3 cells a normal plasma concentration of urea causes carbonylation, and carbonylation increases over the uremic range of urea concentration, indicating that urea can contribute directly to the carbonylation found in uremia.

Cells in the renal inner medulla are normally exposed to variably high concentrations of urea and NaCl as a consequence of the urinary concentrating mechanism. Systemic urea concentration also increases in uremia. High urea denatures proteins, affects enzyme activity in vitro (1), and affects the function of cells in tissue culture. In mouse renal inner medullary (mIMCD3) cells, high urea activates several mitogenic signaling proteins (2). It increases transcription and activity of the immediate-early gene, Egr-1 (3), in an MEK-, ERK-, and Elk-1-dependent fashion (4). Urea also initiates other signaling events characteristic of a receptor tyrosine kinase-mediated pathway (5). In addition, high urea causes oxidative stress to mIMCD3 cells (6): (i) it increases mRNA and protein expression of growth arrest and DNA damage-inducible protein 153 (GADD153); (ii) the increases of GADD153 are completely prevented by pretreatment with the antioxidant N-acetylcysteine; and (iii) high urea decreases the level of reduced glutathione, a biological index of oxidative stress at the cellular level. Taken together, these findings suggest that high urea not only initiates signaling cascades but also causes oxidative stress, which in itself could act as a signal (7) in the renal inner medulla and in cells exposed to high urea because of renal failure.

Because oxidative stress damages cellular DNA and proteins, we hypothesized that high urea might also damage cellular DNA and proteins, either by a direct effect on the macromolecules or indirectly because of the oxidative stress that urea causes. Previously, when the neutral comet assay was used, high NaCl, but not urea, was found to cause DNA double-strand breaks (8). However, other forms of DNA damage, including single-strand DNA breaks and covalent modification of bases, are not detected by the neutral comet assay, but are detected by the alkaline comet assay (9) because at high pH, DNA unwinds revealing, in addition to double-strand breaks, single-strand breaks, including those that occur during excision repair of covalently modified bases. When the alkaline comet assay is used, numerous DNA breaks are found in renal inner medullary cells adapted to high NaCl and urea in tissue culture and in renal inner medullary cells in vivo (10). Oxidative stress induces formation of 8-hydroxyguanine in DNA. Human oxoguanine glycosylase (hOGG1; 7,8-dihydro-8-oxoguanine DNA glycosylase) catalyzes the excision repair of 8-oxoguanine lesions in cellular DNA (11). Furthermore, hOGG1 also excises 8-hydroxyguanine from DNA in vitro, causing single-strand breaks that provide a test for 8-hydroxyguanine lesions. Thus, we use alkaline comet assay and hOOG1 to determine whether urea causes DNA single-strand breaks and 8-oxoguanine lesions in mIMCD3 cells. Given the evidence that high urea causes oxidative stress, we also tested for oxidative damage to proteins by measuring protein carbonylation (12) in mIMCD3 cells exposed to high urea or NaCl, and in renal inner medullary cells during their normal exposure to high urea and NaCl in vivo. The term “carbonylation” means protein oxidation as determined by carbonyl derivatives.

Materials and Methods

Cell Culture. mIMCD3 (13) cells between passages 12 and 18 were grown at 37°C in 5% CO₂ on 100-mm Falcon plastic dishes in 1:1 Irvine Dulbecco's modified Eagle's medium (DMEM)/Ham's medium F-12, containing 2 mM l-glutamine and 10% FBS. The osmolality of the basal medium was 300 milliosmoles (mosmol)/kg (Advanced Osmometer, Advanced Instruments, Norwood, MA). Cells were osmotically stressed by replacing growth medium with hyperosmotic medium containing added urea, NaCl, or other solutes, as indicated. Urea solutions were specially prepared as follows. Immediately before an experiment, 20 ml of 10 M urea (Sigma) was mixed with 1.3 g of AG 501-X8 resin (Bio-Rad) at room temperature for 1 h to remove cyanate, then added to the medium to increase the osmolality, as indicated. As a control, an equal volume of ultrapure water was added to the basal medium. Cyanate forms spontaneously from urea. To evaluate possible effects of the cyanate, we used a conductivity meter (MultiFunction conductivity meter, Amber Science, Eugene, OR) to measure cyanate in urea solutions, comparing conductivity with that of potassium cyanate (Sigma) standards. Fresh urea solutions, treated with the ion-exchange resin, contain no cyanate.

Measurement of Reactive Oxygen Species (ROS). ROS were measured as described in ref. 14. In brief, dichlorodihydrofluorescein diacetate (Molecular Probes) was dissolved in DMSO and diluted with phenol red-free DMEM/F-12 containing 10% FBS to a final concentration of 20 μM in 0.1% DMSO. Cells were loaded with dichlorodihydrofluorescein diacetate for 1 h at 37°C in 5% CO₂ and then quickly rinsed three times with dye-free phenol red-free medium. Then a hyperosmotic medium containing added urea or NaCl or other solutes was added for 15 min at 37°C in 5% CO₂. Cells were imaged by laser confocal microscopy. To quantitate ROS levels, relative dichlorofluorescein fluorescence was measured in images of 60 randomly selected cells for each treatment.

Alkaline comet assay was performed by using a hOGG1 FLARE Assay Kit (Trevigen, Gaithersburg, MD), following the manufacturer's instructions. mIMCD3 cells were embedded in a thin layer of agarose (Comet LMAgarose, Trevigen) on a slide (FLARE Slides, Trevigen). Slides were incubated in cold lysis solution (Trevigen) at 4°C for 1 h to permeablize cell membranes then at 37°C for 1 h in a 1:500 dilution of the hOGG1 enzyme solution supplied in the kit. Slides were then incubated for 30 min at room temperature in the dark in an alkaline solution (6 g of NaOH and 1 ml of 500 mM EDTA in 499 ml of deionized water, pH > 13). Electrophoresis used the alkaline solution in a horizontal electrophoresis tray (Owl Scientific, Woburn, MA), applying 300 mA, 25 V for 30 min at 4°C. DNA was stained with SYBR Green (Trevigen) and analyzed with the epif luorescence microscope (Olympus, Melville, NY) of a laser scanning cytometer (CompuCyte, Cambridge, MA). Digital images were acquired with a Kodak Digital Science DC 120 Zoom digital camera. Comet moment in each cell was quantified for at least 50 random cells per sample by using scion image (Scion, Frederick, MD) software to outline each comet head and tail manually, then integrate the SYBR fluorescence within the outline. Comet moment is defined as ratio of DNA in the comet tail to total DNA (tail plus head).

Protein Isolation. Proteins were isolated as follows. For cell cultures we used ice-cold M-PER lysis buffer, containing a protease inhibitor mixture (Pierce), according to the manufacturer's instructions. For mouse kidney cortex or inner medulla we homogenized in ice-cold lysis buffer containing 50 mM Tris·HCl at pH 7.4, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100,1 mM Na₃VO₄, 1 mM NaF, and protease inhibitor mixture (Roche Diagnostics), and using a tissue homogenizer (Omni 1000 with a micro saw tooth generator). All samples were centrifuged at 20,897 × g for 5 min, and the supernatant was used for protein analysis. Total protein concentrations were measured (BCA kit; Pierce), and the samples were diluted to equal protein concentration with lysis buffer.

Protein carbonylation was measured as described in ref. 15 with an OxyBlot Protein Oxidation Detection kit (Chemicon) according to the manufacturer's instructions, as follows. Five microliters of protein was added to 5 μl of 12% SDS. Samples were derivatized by adding 10 μl of 2,4-dinitrophenylhydrazine solution for 15 min at room temperature, followed by addition of 7.5 μl of “neutralization solution” and 1.5 μl of 2-mercaptoethanol.

Dot-blot analysis used a Bio-Dot apparatus (Bio-Rad). Poly(vinylidene dif luoride) membranes (Millipore) were dipped in 100% methanol for 15 s, rinsed with water, then soaked in 20% MeOH/80% PBS (containing 200 mg/liter KCl, 200 mg/liter KH₂PO_4, 8,000 mg/liter NaCl, and 2,160 mg/liter Na₂HPO₄·7H₂O) for 5 min. Twenty-nine microliters of sample, containing 2 μg of protein, was loaded in each well and allowed to sit on the membrane for 2 min before applying vacuum for long enough to filter all of the fluid. The membranes were washed three times in PBS for 5 min each, then nonspecific protein binding sites were blocked with 1% BSA in PBS containing 0.05% Tween 20 (PBS-T) for 1 h at room temperature. Membranes were incubated for 1 h at room temperature with rabbit anti-2,4-dinitrophenyl antibody, diluted 1:150 with 1% BSA/PBS-T, washed three times with PBS-T, 10 min each, then incubated for 1 h at room temperature with the secondary antibody (goat anti-rabbit IgG couple to horseradish peroxidase), diluted 1:300 with 1% BSA/PBS-T, and washed. Dots were visualized with a luminol-based enhanced chemiluminescence substrate (LumiGLO, Kirkegaard & Perry Laboratories) and quantified by laser densitometry (Personal Densitometer SI, Molecular Dynamics). Membranes were then stained with amido black (0.1% amido black/45% methanol/10% glacial acetic acid/45% H₂O) to confirm equal protein loading.

To test whether urea can directly carbonylate cellular proteins in vitro, mIMCD3 cells grown at 300 mosmol/kg were homogenized in 300 mosmol/kg PBS, containing 25 mM bicarbonate, and gassed with 5% CO₂, using a pellet pestle (Kimble Glass, Vineland, NJ). Homogenates were left at 300 mosmol/kg or osmolality was increased to 600 mosmol/kg by adding urea, then incubated at 37°C for 15 min. After the resulting suspensions were centrifuged at 20,897 × g for 5 min, SDS was added to 6% final concentration and OxyBlot dot blots were performed, as above.

Western analysis used SDS/PAGE on 12% polyacrylamide gels. The proteins (5 μg each lane) were transferred to poly(vinylidene difluoride) membranes electrophoretically (Bio-Rad TransBlot SemiDry Transfer Cell).

To confirm that the signal measured carbonylation, carbonyl groups were reduced in some samples by incubation for 30 min at 37°C, pH 7.4, with 20 mM NaBH₄ (from a stock of 1 M NaBH₄ in 100 mM NaOH, diluted 1:50 with lysis buffer).

Immunohistological Detection of Carbonylated Proteins. Normal mouse kidney was fixed in 4% paraformaldehyde, paraffinembedded, cut, and mounted on silanized slides (American Histolabs, Gaithersburg, MD). Sections were deparaffinized with xylene and rehydrated in a graded series of ethanol concentrations. Endogenous peroxidase was quenched by placing the slides in 3% hydrogen peroxide in methanol for 10 min. Heat-induced epitope retrieval was performed by boiling the slides for 6 min in citrate buffer solution at pH 6.0 (Zymed, no. 00-5000). Slides were washed in PBS. To derivatize the carbonyl groups in the proteins to 2,4-dinitrophenylhydrazone, sections were incubated with 2,4-dinitrophenylhydrazine for 15 min. Slides were washed with PBS and stained with anti-2,4-dinitrophenyl antibody by using Histostain-Plus Kit (Zymed) according to the manufacturer's instructions. Briefly, sections were blocked with serum blocking solution then incubated successively with primary antibody, biotinylated secondary antibody, and streptavidin-peroxidase conjugate. Peroxidase was visualized by addition of 3,3′-Diaminobenzidine tetrahydrochloride substrate, which reacts with peroxidase to produce a brown deposit. Formaldehyde-based fixatives have been reported to produce nonspecific 2,4-dinitrophenylhydrazine adduction, which can be avoided by fixation with Methacarn (methanol/chloroform/acetic acid, 60:30:10) (16). Therefore, we also tested this fixative. The result (data not shown) was essentially the same as fixing with paraformaldehyde.

Statistical significance was tested by using the Student t test.

Results

High Urea, NaCl, or Raffinose Increases ROS, but Sorbitol or Glycerol Does Not. The level of ROS was determined by measuring fluorescence of dichlorodihydrofluorescein diacetate. Elevating osmolality to 600 mosmol/kg for 15 min by adding urea, NaCl, or raffinose increases ROS ≈2-fold, but addition of the same amount of sorbitol or glycerol has no significant effect (Fig. 1A).

Fig. 1.

ROS and DNA damage in mIMCD3 cells. (A) Effect of different solutes on ROS. The solutes were added for 15 min. Images are representative of three to five experiments. Mean values for relative fluorescence at 300 mosmol/kg, 243 ± 31, and at 600 mosmol/kg urea, 638 ± 112^*; NaCl, 577 ± 109^*; raffinose, 528 ± 72^*; sorbitol, 146 ± 23; and glycerol 246 ± 45. *, P < 0.05 vs. 300 mosmol/kg. (B) High urea causes 8-oxoguanine lesions in mIMCD3 cell DNA, but high NaCl does not in the comet assay. Raising osmolality from 300 to 600 mosmol/kg for 1 h by adding urea increases fragmentation of DNA, and incubation of the isolated nuclei with hOGG1 increases it even more. Adding NaCl also causes DNA fragmentation, but incubation with hOGG1 does not increase fragmentation.

High Urea Causes Single-Strand Breaks in DNA and 8-Oxoguanine Lesions. Previously, by using the neutral comet assay, it was observed that high NaCl causes double-strand breaks in DNA, but high urea does not (8). High urea causes oxidative stress (6), which can damage DNA. Because DNA becomes unwound at high pH, the alkaline comet assay can detect single-strand DNA breaks, including those resulting from oxidative lesions in DNA (11). In the present experiments, using the alkaline comet assay, we find that raising osmolality from 300 to 600 mosmol/kg for 1 h by adding NaCl causes DNA breaks (Fig. 1B). Comet moment (which quantifies damaged DNA) rises from 18% to 36.7% (Table 1). The DNA breaks caused by high NaCl presumably include the double-strand breaks detected with the neutral comet assay (8). Although high urea does not cause double-strand DNA breaks, as determined with the neutral comet assay (8), it does cause single-strand breaks, as determined with the alkaline comet assay (Fig. 1B). Raising osmolality in mIMCD3 cells from 300 to 600 mosmol/kg for 1 h by adding urea increases comet moment from 18% to 25.7% (Table 1). Single-strand DNA breaks occur during excision repair of altered DNA bases, for example those whose oxidation produces 8-oxoguanine lesions (11). Those lesions can be detected by incubation of isolated nuclei with the repair enzyme, hOGG1, which excises them in vitro, producing additional DNA breaks. Accordingly, incubation with hOGG1 increases DNA breaks in nuclei from cells exposed to high urea, but not cells exposed to high NaCl (Fig. 1B). Comet moment increases to 35.3% in nuclei from cells exposed to high urea but is unaffected in nuclei from cells exposed to high NaCl (Table 1). We conclude that urea causes single-strand DNA breaks, associated with oxidative lesions in DNA, but high NaCl does not.

Table 1. Comet moments in alkaline comet assay.

	Comet moment, %
Conditions	—hOGG1	+hOGG1
300 mosmol/kg	18.0 ± 1.7	18.7 ± 1.5
600 mosmol/kg, urea added	25.7 ± 1.8*	35.3 ± 1.8†
600 mosmol/kg, NaCl added	36.7 ± 2.3*	36.7 ± 2.4

^*, P < 0.05 vs. 300 mosmol/kg

^†, P < 0.05 vs. —hOGG1; n = 3

High Urea Carbonylates Proteins in Renal Cell Culture. Given the evidence that high urea causes oxidative stress, we tested for oxidative damage to proteins by measuring protein carbonylation in mIMCD3 cells exposed to high urea. Increasing osmolality from 300 to 600 mosmol/kg by adding urea for 1 h increases protein carbonylation (Fig. 2A), indicating that high urea does cause oxidative damage to proteins.

Fig. 2.

Effect of added solutes on carbonylation of mIMCD3 cell proteins in carbonyl dot-blot assay. (A) Raising osmolality from 300 to 600 mosmol/kg by adding 300 mM urea for 1 h increases carbonylation. The protein carbonylation is not an indirect effect of cyanate (CNO) in the urea solutions because adding 0.02 mM or 0.2 mM KCNO to mIMCD3 cells for 1 h does not produce carbonylation (*, P < 0.05, n = 3). (B) Time course of carbonylation induced by adding 300 mM urea (representative of three experiments). (C) Urea concentration dependence of carbonylation. The urea was added for 15 min (*, P < 0.05 vs. 300 mosmol/kg, n = 3). (D) Protein carbonylation produced by increasing osmolality by addition of different solutes for 15 min.

Cyanate, which forms spontaneously from urea, can modify proteins by carbamylating them (17). To assess the possibility that urea-induced carbonylation might be an indirect effect of cyanate-induced carbamylation, we used an ion-exchange resin to remove cyanate from the urea solutions that we used (see Materials and Methods). Nevertheless, cyanate is continuously produced from urea in solution. To estimate the amount of cyanate that might form from urea during experimental incubations, we measured cyanate in a 300 mM urea solution that had been freshly diluted from the 10 M stock solution, then incubated at 37°C. After 1 h of incubation, the concentration of cyanate is 0.02 mM. We also determined that the equilibrium concentration of cyanate in a 300 mM urea solution is 0.2 mM by testing a solution of urea that had been stored for 6 months at room temperature. In the absence of urea, those concentrations of cyanate do not cause protein carbonylation in mIMCD3 cells (Fig. 2 A). We conclude that high urea carbonylates proteins in mIMCD3 cells, and that this effect is not an indirect effect of protein carbamylation produced by cyanate in the urea solutions.

We next tested how rapidly urea carbonylates proteins in mIMCD3 cells and how much urea is required. Carbonylation is very rapid. Extensive protein carbonylation is already evident within 5 min after 300 mM urea is added (Fig. 2B). Further, very little urea is required. As little as 5 mM urea significantly carbonylates proteins in mIMCD3 cells within 15 min (Fig. 2C). The amount of carbonylation apparently peaks at 20 mM. However, oxidation may be even greater at higher urea concentration. Carbonylation at higher urea concentrations may decrease because the carbonyl groups are susceptible to further oxidation to carboxylic acids. For example, initial oxidation of glutamine synthetase converts an arginine residue to the carbonyl-bearing derivative, γ-glutamyl semialdehyde, but this derivative can be further oxidized to glutamic acid. Of note, 5 mM urea is a normal plasma concentration, and the range in uremia is 20–80 mM. Thus, the level of urea normally present in blood can cause protein carbonylation, and the higher urea in renal failure can account for the increased carbonylation observed in that condition (18).

High concentrations of urea directly affect enzyme activity and denature proteins (1). Therefore, it is conceivable that urea-induced carbonylation of cell proteins could be a direct chemical effect of urea. To test this hypothesis, we added 300 mM urea for 15 min directly to homogenates of cells that had been at 300 mosmol/kg. Direct addition of urea does not cause protein carbonylation (data not shown). Carbonylation by urea in cell-free homogenates presumably requires addition of components that have been diluted or modified by homogenization.

High NaCl or Raffinose Carbonylates mIMCD3 Cell Proteins, but High Sorbitol or Glycerol Does Not. To determine whether high levels of other solutes besides urea cause carbonylation, we tested the effects of adding 300 mosmol/kg NaCl, raffinose, sorbitol, or glycerol to mIMCD3 cells for 15 min. High NaCl causes as much carbonylation as urea (Fig. 2D). This result is physiologically significant for renal inner medullas, where concentrations of NaCl and urea are both high. Also, it is consistent with the observation (Fig. 1A) that high NaCl causes oxidative stress, as does urea. To determine whether the effect of high NaCl is due to a specific effect of Na or Cl, we tested raffinose. High raffinose increases protein carbonylation, as much as urea or NaCl (Fig. 2D), suggesting that hypertonicity might account for the effects of high NaCl, including oxidative stress. However, the lack of effect of high sorbitol (Fig. 2D) is against this hypothesis, because high sorbitol is also hypertonic. Nevertheless, the fact that high sorbitol does not increase carbonylation is consistent with its lack of effect on ROS (Fig. 1A). Finally, glycerol, which is a compatible solute that readily permeates cells, does not cause carbonylation (Fig. 2D) or increase ROS (Fig. 1A). The lack of effect of both sorbitol and glycerol makes it unlikely that hyperosmolality, per se, causes carbonylation.

High Urea Carbonylates Many Proteins in mIMCD3 Cells, but Not All Proteins Are Equally Carbonylated. We used Western blots to determine the generality of urea-induced protein carbonylation. The addition of 300 mM urea to mIMCD3 cells for 15 min carbonylates numerous proteins of various sizes (Fig. 3). However, not all proteins are equally carbonylated. Comparison of the abundance of particular size proteins (amido black staining) to the extent of their carbonylation reveals some nonabundant proteins that are heavily carbonylated and some abundant proteins that are not (Fig. 3).

Fig. 3.

Protein carbonylation in vivo and in mIMCD3 cells shown by using Western blots. Western blots used anti-2,4-dinitrophenyl antibody. Molecular masses of markers (kDa) are indicated on the left. Proteins in the blots were later stained nonspecifically with amido black to display their total abundance. (Left) Normal mouse renal cortex or inner medulla. (Right) mIMCD3 cells: control or after adding urea for 15 min.

Proteins Are Extensively Carbonylated in Normal Mouse Renal Inner Medulla in Vivo. Interstitial concentrations of both NaCl and urea are normally high in renal inner medullas. Because a high concentration of either NaCl or urea carbonylates mIMCD3 cell proteins (Fig. 2D), it was logical to suppose that protein carbonylation might occur in renal inner medullas in vivo. That is indeed the case. In normal mice, proteins from inner medullas, where NaCl and urea are high, are much more highly carbonylated than proteins from the renal cortex, where NaCl and urea are not elevated (Figs. 3 and 4).

Fig. 4.

Protein carbonylation in normal mouse kidney in vivo. (A) Immunohistological staining demonstrates widespread protein carbonylation in inner medulla and papilla, but little in cortex, except in distal nephron segments (arrows). (B) Carbonyl dot blots from cortex and inner medulla (IM); analytical control adding 20 mM NaBH₄, which reduces carbonyl groups in proteins. (C) Densitometric analysis of carbonyl dot blots from renal cortex and inner medulla (*, P < 0.05 vs. cortex, n = 3).

As an analytic control, we exposed the lysates to NaBH₄, which reduces carbonyl groups. NaBH₄ greatly decreases the signal intensity, consistent with bona fide detection of carbonylation (Fig. 4B).

We used immunocytochemistry to determine whether carbonylation in the inner medulla is limited to certain cell types. Carbonylation is widespread in the inner medulla, including the papilla (Fig. 4A). Essentially all cells, including collecting ducts and thin limbs, contain heavily carbonylated proteins. In contrast, in the cortex only some distal nephron segments, which form a minor part of the mass, are stained (Fig. 4A).

Carbonylation is greater in Western blots from mouse renal inner medulla than from cortex (Fig. 3), consistent with the dot blots (Figs. 4 B and C). Also, as with mIMCD3 cells exposed to high urea (Fig. 3), many proteins are carbonylated in the inner medulla, but not equally so. Proteins of 36 and 70 kDa are especially abundant and heavily carbonylated in the inner medulla (Fig. 3).

Discussion

DNA Damage. Although high NaCl was found to cause DNA double-strand breaks, high urea does not (8). In the present studies we find that in mIMCD3 cells acute elevation of urea, but not NaCl, causes 8-oxoguanine lesions. The 8-oxoguanine occurs in mammalian cellular DNA either as a byproduct of normal oxidative metabolism or as a result of exogenous sources of reactive oxidizing species, such as ionizing radiation, singletoxygen sensitizer dyes, and redox-active organic molecules (11). Thus, 8-oxoguanine is a form of oxidative damage to DNA, and its occurrence in the present studies is presumably caused by oxidative stress resulting from high urea. Formation of 8-oxoguanine is detrimental. It preferentially mispairs with adenosine during replication, giving rise to G·C to T·A transversion mutations, which can be tumorigenic (11). Oxoguanine glycolase implements excision repair of 8-oxoguanine lesions (11). Its level is high in the kidney (19).

DNA damage is high in normal mouse inner medullary cells in vivo (10). The damage apparently results from the high levels of urea and NaCl normally present in the inner medulla because it is rapidly repaired when urea and NaCl levels are lowered by the diuretic furosemide. At this point it is not clear how much of the damage is DNA double-strand breaks caused by the high NaCl and how much is 8-oxoguanine lesions caused by the high urea. At this point we do not know how the cells survive and function despite this damage. Possible ameliorating factors include high expression of heat proteins (20), which are antiapoptotic (21), and very limited cellular proliferation (22), which minimizes replication of any damaged DNA.

Protein Carbonylation. We find that high NaCl or urea, in addition to damaging DNA, also carbonylates proteins, and that protein carbonylation is extensive in the normal renal inner medulla in vivo, presumably a consequence of high interstitial urea and NaCl concentrations. Protein carbonyl content is a general biomarker of severe oxidative protein damage. High levels of protein carbonylation are associated with diseases including Alzheimer's disease, chronic lung disease, chronic renal failure, diabetes, and sepsis (23), and also with aging (24). For the most part, oxidatively modified proteins are not repaired and must be removed by proteolytic degradation, usually carried out by the proteosome, which selectively degrades oxidatively modified proteins. Proteins marked by carbonylation are degraded at a higher rate (24). Protein carbonylation has also been associated with important functional alterations in a variety of structural and enzymatic proteins, including inhibition of enzyme activity (25). At this point, however, we do not know what the functional consequences are of urea- and NaCl-induced oxidation of renal medullary proteins.

High NaCl and urea apparently do not carbonylate proteins directly, but carbonylate as a result of the oxidative stress that they cause. Thus, high urea does not carbonylate proteins in homogenates of cells (data not shown), presumably because it does not elevate reactive oxidizing species sufficiently under those conditions in vitro, or because it requires addition of components that have been diluted or modified by homogenization. Both high NaCl and high raffinose cause protein carbonylation (Fig. 2D). NaCl and raffinose have little in common except that a high concentration of each is hypertonic, making it likely that their effect is due to the hypertonicity, rather than a specific effect of either. On the other hand, high urea is not hypertonic because urea readily permeates cells. However, urea affects protein structure and activity (1), which could cause oxidative stress if the affected proteins were involved in oxidative reactions. Because high sorbitol and glycerol do not cause carbonylation (Fig. 2D), high osmolality is not of itself the cause the oxidative stress. The lack of effect of sorbitol (Fig. 2D) is puzzling. High sorbitol is hypertonic, like raffinose, but its effect evidently differs. The reason for this unexpected result is not clear. Although any sugar can function as a radical scavenger at very high concentrations, there is nothing unique about sorbitol, and the concentrations used in our experiments are well under those required for radical scavenging or an antioxidant action (26).

It is remarkable how little urea is necessary to carbonylate cellular proteins. Five millimolar urea, which is a normal plasma level, significantly increases carbonylation (Fig. 2C). Thus, the normal level of urea could contribute to the low levels of carbonylation that normally exist. Increasing urea above 5 mM greatly increases carbonylation (Fig. 2C). The carbonylation appears to peak at 20 mM urea, but this result is misleading. As oxidation proceeds, further oxidation forms products that the carbonyl assay does not detect, so that higher levels of oxidation are not recorded (27). The increased carbonylation of serum albumin found in chronic renal failure was attributed to oxidative stress (18). We suggest that high urea might be the cause of the oxidative stress. Further, the carbonylation of albumin might be occurring in the liver cells in which it is synthesized.

It has been suggested that urea might be a uremic toxin because cyanate, which forms spontaneously from urea, is elevated in chronic renal failure, resulting in increased protein carbamylation (17). The present studies add to the evidence that urea is a uremic toxin, given our observation that protein carbonylation is increased over the range of urea concentration that occurs in chronic renal failure (Fig. 2C). Both carbonylation and carbamylation can be toxic because they alter protein function (17, 25).

Numerous proteins are heavily carbonylated in normal mouse renal inner medulla (Fig. 3). Carbonylation of abundant 36- and 70-kDa proteins is particularly striking. The osmoprotective proteins aldose reductase (36 kDa) (28) and HSP70 (20) are both known to be particularly abundant in inner medullas. Further studies are required to confirm these identifications and to test what effect the carbonylation has on their function.

The high level of protein carbonylation in renal inner medullas in vivo suggests that renal inner medullary cells normally are under oxidative stress. Further studies are necessary to confirm these results directly. It is known, however, that oxygen tension is low in the inner medulla because of the medullary counter-current exchange system and that glycolysis is the predominant metabolic pathway in the inner medulla (29). Low oxygen tension and respiration may help to minimize the oxidative stress in this tissue.