Restriction of sulfur-containing amino acids alters claudin composition and improves tight junction barrier function

S. Skrovanek, M. C. Valenzano, J. M. Mullin


Restriction of sulfur-containing amino acids (SCAA) has been shown to elicit a similar increase in life span and decrease in age-related morbidity as caloric restriction. The singular importance of epithelial barrier function in both physiological homeostasis and prevention of inflammation raised the issue of examining the effect of SCAA restriction on epithelial tight junction structure and permeability. Using a well-described in vitro, epithelial model, the LLC-PK1 renal epithelial cell line, we studied the effects of SCAA restriction in culture medium. Reduction of methionine by 90%, cysteine by 50%, and total elimination of cystine resulted in dramatically lower intracellular pools of these amino acids and their metabolite, taurine, but the intracellular pools of the non-SCAA were all elevated. Cell growth and differentiation were maintained, and both confluent cell density and transepithelial short circuit current were unaffected. Certain tight junctional proteins, such as occludin and claudins-1 and -2 were not altered. However, claudins-3 and -7 were significantly decreased in abundance, whereas claudins-4 and -5 were markedly increased in abundance. The functional result of these structural changes was improved barrier function, as evidenced by increased transepithelial electrical resistance and decreased transepithelial (paracellular) diffusion of d-mannitol.

  • caloric restriction
  • diet
  • aging
  • permeability
  • methionine

higher animal life is (both phylogenically and ontologically) a complex derivative of the barrier structure and fluid compartmentation that is seen in a very simple form in the coelenterate or the blastocyst. For every multicellular organism, there is always a luminal compartment and an antiluminal compartment—a paradigm theme of higher animal life (Fig. 1). In the human body, with its manifold organs, there exists a wide array of luminal compartments (many of which communicate with the outside environment with its attendant high pathogen level) and a common antiluminal compartment in the vasculature and stroma. The compartments are separated by an epithelial or endothelial layer, which itself consists of two components in parallel, the cells themselves and the tight junctional seals that surround each cell circumferentially. This construction is readily obvious in the urinary bladder, the cornea, or the colon, but is equally true, though less obvious, for the blood-brain barrier separating the central nervous system from the bloodstream. Interestingly, the immune system typically resides in the protected antiluminal compartment, a barrier away from the relatively unregulated luminal space in communication with the environment. Pathogens have ready access to luminal compartments and luminal surfaces of many of these barriers, and it is only by crossing these barriers that contact between pathogen and immune system components is made possible, followed by the entire cascade, which we then term inflammation.

Fig. 1.

Diagrammatic cross section of an epithelium, indicating various routes of transport. Both paracellular and transcellular flux routes are indicated. Paracellular flux is passive and regulated by the selectively permeable tight junction. Transcellular flux occurs by both transporter-mediated means and simple diffusion.

A strong case is made in the published literature that this manifold of epithelial and endothelial barriers constituting an organism becomes compromised as the organism ages. Because of its enormous surface area and bacterial floral component in its luminal compartment, an age-related compromise of the gastrointestinal (GI) barrier is particularly noteworthy, as the GI tract represents the greatest source of bacterial antigens in the body. Age-related compromise of the GI barrier seemingly occurs specifically at the tight junction (TJ), a gasket-like proteinaceous seal, which circumferentially bands each and every epithelial and endothelial cell and prevents free solute diffusion along the paracellular space (69, 77). Evidence exists that a similar phenomenon is occurring in the lung (74), the blood-brain barrier (46), and the epididymis (35). Increased paracellular leakiness was also observed in the colon of 22-mo-old rats (relative to 6-mo-old rats), with diet playing a role in the effect (53). It is unclear whether the age-related change is due directly to the structural components of the barrier or to signaling pathways like PKC and RAF/MEK/ERK that are known to regulate those structural components (49, 53).

Regardless of the molecular mechanism for the leaks, their existence would mean that immunostimulatory agents like endotoxin, specific bacterial toxins, or bacteria per se can move from luminal to antiluminal compartments and activate immune cascades in the process. Such translocation is obvious in instances of acute trauma to a barrier but may be just as active, though more subtle, in a long-term state of chronic leak on a molecular level. Furthermore, once an inflammatory state ensues, it is very noteworthy that proinflammatory cytokines are capable of causing further transepithelial/transendothelial leak (27, 51, 68).

This present research was undertaken with the hypothesis that a connection exists between the compromise of epithelial and endothelial tissue barriers as a function of aging, and the low-level systemic inflammation that is increasingly associated with aging. Numerous studies have indicated that a low-grade inflammatory state is common in advancing age (8, 11, 22, 64). This finds representation in high circulating levels of pro (and anti) inflammatory cytokines (38, 64) and C-reactive protein (56), as well as altered white blood cell compositions (14, 73) in the elderly. It also is represented in age-related incidence of certain inflammatory disorders ranging from arthritis to macular degeneration to cardiovascular disease to neurodegenerative diseases (43, 66). A heightened proinflammatory state in aging would derive from one of two possibilities: 1) direct, age-related alterations in immune system cells, and/or 2) age-related alteration in the avenues by which these cells are activated. Without excluding or minimizing the likelihood of the first possibility, this study focuses on the second.

The increase of median and maximum life span (as much as 30–40%) that accrues from caloric restriction was shown in the 1990s to also occur when specific essential amino acids such as methionine were restricted in the diet of rodents (59, 84). Extension of life span seems to occur in concert with, and perhaps be related to, a delay in age-related morbidities (16, 41, 55, 79). The central hypothesis of this current research study is that this dietary-related forestalling of age-associated morbidities may be linked, in part, to a dietary enhancement of epithelial barrier function. Such enhancement should result in improved separation of immunostimulatory molecules like endotoxins from the LPS receptors of white blood cells or even stromal cells that trigger cytokine production and the inflammatory cascade process.

This is the first study of sulfur-containing amino acid restriction effects on epithelial barrier function. It uses a differentiated renal epithelial cell culture model, LLC-PK1, which is well known to form an efficient and well-characterized epithelial barrier in vitro, with similarity to the kidney proximal tubule, one of the best studied of all epithelial tissue barriers (54, 63).


Cell culture.

The porcine renal cell line, LLC-PK1, was a gift of Dr. Robert Hull (Eli Lilly) and was used from passages 188 to 200. LLC-PK1 is an epithelial cell line derived from the outer cortex of porcine kidney and was chosen for use in this study because it is a widely used differentiated, polar epithelial cell culture model with which our laboratory has extensive experience. Cultures were passaged weekly, upon reaching confluence, by trypsinizing with 0.025% wt/vol trypsin in 0.02% wt/vol EDTA (CellGro), and seeding into vented Falcon T75 tissue culture flasks with 25 ml of Eagle's minimal essential medium alpha-modification (α-MEM) without deoxyribonucleosides and ribonucleosides and with Earle's salts (HyClone). This medium was further supplemented with 2 mM l-glutamine and 10% defined FBS (HyClone). Cultures were incubated at 37°C in an atmosphere of 95% air-5% CO2.

For electrophysiology and tracer flux experiments, cells needed to be grown on permeable supports. Cells were seeded at 1 × 106 (confluent density) onto sterile Millipore Millicell polycarbonate (PCF) filter ring assemblies (pore size 0.4 μm and diameter 30 mm) in culture medium on a Monday. Three to four Millicell PCF units were placed in a 100-mm sterile Petri dish. Medium in the inserts became the “luminal” fluid compartment. Medium in the Petri dish became the “antiluminal” fluid compartment. Cells were refed with either control or sulfur-containing amino acid (SCAA)-restricted media on Tuesday, Friday, and the following Tuesday (the day of the experiment).

Amino acid-modified medium.

The diet for the Fischer 344 rats reported in Orentreich et al. (55) was the model used in cell culture media preparation for these experiments. Custom-made α-MEM without cysteine, cystine, or methionine was purchased commercially (HyClone), and appropriate levels for control and restricted media were added back to the media using tissue-culture grade amino acids (Fisher) dissolved in HPLC-grade water. Normally, cysteine is present at 0.569 mM in α-MEM, cystine at 0.198 mM, and methionine at 0.1 mM. It was found that the cells could not survive without at least 50% of this level of cysteine; however, given this, cystine could be completely eliminated without adverse effects. Under these conditions, methionine could be decreased to 10% of normal levels without affecting differentiation state and only marginally slowing cell growth.

LLC-PK1 cells did not survive when cysteine levels were decreased below 50% of normal but methionine was raised to 300% of normal. This indicates that conversion of methionine to cysteine is either impossible or extremely reduced in these epithelia. It should be noted that mammalian cells cannot convert cysteine or cystine to methionine, explaining why methionine is an essential amino acid; cysteine and cystine freely interconvert, as cystine is simply a cysteine dimer.

The final amino acid concentrations for the restricted medium were as follows: 1) cysteine: 0.285 mM, or 50% of its normal concentration (in α-MEM); 2) cystine: 0 mM, or 0% of its normal concentration; 3) methionine: 0.01 mM, or 10% of its normal concentration.

To ensure that the sulfur-containing amino acids were indeed restricted to the levels reported, FBS was dialyzed to remove free amino acids. While this does not eliminate the potential for cells to scavenge amino acids from larger proteins, it does bring the amount of amino acids in the serum down to the minimum level possible and make the culture medium the only source of free amino acids. Control medium was also prepared with dialyzed FBS. Cells were observed to grow to confluence and differentiate at the same rate and in the same manner as cells in nondialyzed FBS.

FBS in dialysis tubing (Spectrum SpectraPor, 3500 Dalton molecular weight cutoff) was dialyzed against three changes of PBS (pH 7.4 and osmolarity of 280 mOsm/l) at 4°C over 3 days. FBS was removed from the tubing and filter sterilized using 0.4-μm pore filters.

Transepithelial electrophysiology.

Before electrophysiological measurements, cells were refed in fresh culture medium and allowed to incubate at 37°C for 1–2 h. Voltage was measured at 37°C using 1 M NaCl bridges in a series with calomel electrodes and read off a multimeter (Fluke 8020B), as previously described (50). Transepithelial electrical resistance (TER) and short-circuit current (ISC) were measured by passing 10 μamp/cm2 current pulses of 1-s duration across the cell sheet using Ag/AgCl electrodes in series with 1 M NaCl bridges. The voltage deflection was measured on an autoranging multimeter (Keithley 197A). TER and ISC were then calculated using Ohm's law.

Radiotracer flux.

Immediately after electrophysiology readings, the 15 ml of (antiluminal) media in the Petri dishes containing the Millicell filter ring assemblies was aspirated and replaced with medium of the same type but now containing 0.1 mM, 0.1 μCi/ml 14C-d-mannitol (Dupont NEN) or 0.1 mM, 0.1 μCi/ml 14C-polyethyleneglycol (Amersham) and incubated at 37°C. Antiluminal medium samples (25 μl) were taken for liquid scintillation counting to calculate the activity and specific activity. Duplicate 25-μl samples were taken from the rings (luminal compartment) at 30, 60, 120, and 180 min, also for liquid scintillation counting, to calculate the flux rates (in cpm·min−1·cm−2 and pmol·min−1·cm−2) of the tracer molecule crossing the cell sheets.

Cell morphology studies.

For phase contrast microscopy, cells were seeded at 1 × 105 into Falcon T75 culture flasks with 25 ml α-MEM supplemented with 2 mM l-glutamine and 10% FBS and allowed to reach confluence (1 wk). Upon reaching confluence, flasks were washed once in 4°C, 1× PBS and then refed in one of the amino acid-restricted media. Cells were refed in their respective culture medium again 4 days later and once more 4 h before observing and photographing their morphology. This was done so as to mimic the conditions that the cells seeded into filter ring assemblies were subjected to for physiology studies. Cells were then observed and photographed using phase contrast microscopy (Nikon MS) at ×100.

For thick-section microscopy, cell sheets were seeded at confluent density on filters in the Millicell PCF assemblies. A week postseeding (with intervening refeedings every 4 days), cell sheets were washed three times in 4°C PBS before the filter was cut from each ring and placed in formalin for ∼2 h before undergoing dehydration with successively increasing concentrations of ethanol. The filter was then infiltrated with liquefied paraffin and encased in more paraffin for long-term storage. Sections (∼10 μm in thickness) were cut with a microtome and floated onto 44°C deionized water before being mounted on uncoated glass slides and allowed to air dry overnight. Sections were then deparaffinized and rehydrated using xylene and successively decreasing concentrations of ethanol before being stained with hematoxylin and eosin.


Cells were seeded at a density of 1 × 106 onto sterile Millicell-PCF filter ring assemblies as above on a Monday. Cells were refed in either control or amino acid-restricted conditions on Tuesday, Friday, and the following Tuesday. That same day, cell sheets were washed three times in 4°C, 1× PBS before the filter was cut from each ring and fixed in fresh 3.2% paraformaldehyde in PBS at room temperature for 15 min. Immediately after fixation, filters were washed 3× in PBS for 5 min each, then permeabilized in 0.1% Triton X100 for 3 min. Cell sheets were again washed in PBS three times for 5 min each, before blocking at room temperature in 10% goat serum in PBS for 45 min. Cell sheets were then exposed to the primary antibody, rabbit polyclonal ZO-1 (Zymed) diluted in 10% goat serum, and allowed to incubate in a humid chamber at room temperature for an hour and a half. After incubation with the primary antibody, cell sheets underwent three additional 5 min PBS washes before exposure to the fluorescent-labeled, secondary antibody, Affini-Pure goat anti-rabbit IgG (H+L) Cy3 (Jackson Laboratories) diluted 1:200 in PBS. Cell sheets were then incubated at room temperature in the dark for 1 h, followed by three final, brief PBS washes and one brief wash in double-distilled water to prevent crystal formation. Cell sheets were then mounted, cell side up, onto slides in VectaShield wet mount with DAPI (Electron Microscopy Sciences) and covered with glass coverslips. Slides were kept in the dark at 4°C until ready to be viewed at which point 10 frames were randomly selected to be photographed and counted.

Intracellular amino acid pool quantification.

Cells were seeded at 1×105 into Falcon T75 culture flasks, grown to confluence, and refed on the same schedule and in the same manner as described above. Two T75s were used per condition to ensure sufficient cells for quantification. One week postconfluence, flasks were washed three times each in 4°C PBS, then scraped into 2 ml of double-distilled water and placed on ice. The cell suspensions were sonicated on ice for 60 s, allowed to sit on ice for 30 s to permit heat dissipation, and sonicated again on ice for 60 s. Suspensions were then flash-frozen using a methanol-dry ice bath and shipped overnight on dry ice to Scientific Research Consortium in St. Paul, MN, where free, intracellular amino acid pools were determined by HPLC.

Claudin and occludin expression by Western immunoblot.

Cells were seeded at 1 × 105 into Falcon T75 culture flasks, grown to confluence, and refed in the same manner as cells used for morphology studies. One week postconfluence, flasks were washed three times each in 4°C PBS, then flash-frozen in a methanol-dry ice bath, and stored at −80°C until analysis. At that time, flasks were quick-thawed and 2 ml of 4°C Buffer A (20 mM Tris·HCl, pH 7.5, 0.25 M sucrose, 10 mM EGTA, 2 mM EDTA) with protease and phosphatase inhibitors (final concentrations: 0.8 μM aprotinin, 20 μM leupeptin, 50 μM bestatin, 1 mM 4-(2-Aminoethyl)benzenesulfonyl fluoride hydrochloride, 10 μM pepstatin) (Calbiochem) was added to each, cells were scraped into the buffer, the suspension mechanically disrupted, and then sonicated for 60 s on ice and transferred to an ultracentrifuge tube. Tubes were centrifuged in a chilled Beckman 50TI rotor at 39,000 rpm for 1 h at 4°C. Supernatants (“cytosolic fraction”) were discarded. To the remaining pellets, 400 μl of cold buffer A with 1% Triton X-100 and protease and phosphatase inhibitors was added, and pellets were mechanically broken up. Suspensions were then rocked for 90 min at 4°C and centrifuged again at 39,000 rpm in a chilled Beckman 50TI rotor for 1 h at 4°C. The supernatant from this final spin was the “membrane fraction.” Total protein was measured using the Bio-Rad DC protein assay kit.

Samples of these fractions were analyzed by PAGE using a Novex XCell SureLock MiniCell apparatus and a 4–20% gradient Novex Tris-glycine, precast, 10-well, 1.5-mm-thick gel (Invitrogen). Precision Plus Kaleidoscope Protein Standards (Bio-Rad) were also included in each gel. Gels were run at 125 V, constant voltage, for 1 h at room temperature.

Proteins were transferred from the gel to a polyvinylidene difluoride membrane using a Novex XCell MiniCell. Transfer was run at 30 V, constant voltage, for 2 h at room temperature. At the end of the transfer, to check for protein transfer efficiency, the membranes were stained with Ponceau S (Sigma) for 10 min, destained with double-distilled water, air-dried, and then photographed. The membrane was then rehydrated and washed three times for 10 min each with PBST (1× PBS with 0.3% Tween-20). The membranes were then blocked with 5% milk/PBST overnight at 4°C.

Blots were incubated with the specific primary antibody at a concentration of 0.3 to 1 μg/ml for 1 h at room temperature. The blots were then incubated with secondary antibody labeled with horseradish peroxidase along with Western Lighting chemiluminescence reagents (Perkin Elmer). For occludin, claudin-1, -3, and -7, the secondary used was goat anti-rabbit, diluted 1:8,000 in 5% milk/PBST; for claudin-2, -4, and -5, the secondary used was rabbit anti-mouse, diluted 1:6,000 in 5% milk/PBST. The blots were then placed against reflection autoradiography film (Kodak) and developed in a Kodak M35A X-OMAT processor.

Films were analyzed for protein expression level by measuring optical density units with a Personal Densitometer SI (Molecular Dynamics).

All primary antibodies against claudin-1, -2, -3, -4, -5, and -7 and against occludin were purchased from Zymed. Occludin, and claudin-1, -3, and -7 primary antibodies were rabbit polyclonal, while primary antibodies for claudin-2, -4, and -5 were mouse monoclonal.

DNA, RNA, and protein determination.

Total protein was determined using the Bradford method (6a). Absorbance at 595 nm was read for all samples using a single-beam spectrophotometer, and a standard curve was created using linear regression and the A595 values of the albumin standards. Total protein for each filter was calculated by linear regression.

Total RNA was determined using the modified orcinol procedure (2) and RNA standards from torula yeast (Sigma). Absorbance at 500 nm was read using a single-beam spectrophotometer. A standard curve was created using linear regression and the A500 values of the torula yeast RNA standards. Total DNA was determined by precipitating cellular DNA in 0.4 M perchloric acid, hydrolyzing in 1 M perchloric acid, and reacting with diphenylamine, along with calf thymus DNA standards (10). Absorbance at 600 nm was read for all samples using a single-beam spectrophotometer. Total DNA for each cell sheet was calculated using linear regression.


Effect of restriction of sulfur-containing amino acids on intracellular amino acid pools and morphology.

LLC-PK1 cells were able to survive with complete elimination of cystine from the culture medium, but cysteine could not be decreased below 50% of its normal level (see materials and methods). With cysteine at 50%, methionine could be decreased to 10% of normal levels without affecting differentiation state and only marginally slowing cell growth. The final SCAA-restricted media were as follows: 1) cysteine: 0.285 mM, or 50% of its normal concentration in α-MEM; 2) cystine: 0 mM; and 3) methionine: 0.01 mM, or 10% of its normal concentration.

When LLC-PK1 cells were cultured in this medium, intracellular amino acid pools of the sulfur-containing amino acids were sharply reduced (Fig. 2). This includes cysteine, cystine, methionine and taurine. The remaining intracellular amino acid pools (glycine, valine, threonine, alanine, serine, proline, leucine, isoleucine, phenylalanine, lysine, arginine, aspartate, and glutamate) were all elevated, by ∼30%–40% (data not shown).

Fig. 2.

Intracellular sulfur-containing amino acid levels in control vs. restricted conditions. Two confluent 75-cm2 tissue culture flasks per condition were placed under control (100% cysteine/100% cystine/100% methionine) (light gray bar) or sulfur-containing amino acid-restricted (50% cysteine/0% cystine/10% methionine) (dark gray bar) conditions for 1 wk before being harvested. Cells were sonicated and sent off for analysis of soluble amino acid pools (SRC, Minnesota). Duplicate 100-μl aliquots were also taken from the sonicates for both DNA and total protein quantification. Results shown are from one quantification but are representative of three quantifications performed, with cysteine, methionine, and taurine always decreasing in sulfur-containing amino acid-restricted conditions, and cystine either decreasing or remaining the same. Results are comparable when evaluated on the basis of milligrams of total protein.

When cells were seeded on Millicell PCF-permeable filters as described in materials and methods, healthy, polar, one-cell-layer-thick cell sheets were observed in cross sections stained with hematoxylin and eosin (data not shown). In phase contrast microscopy (×100) of confluent cell sheets in 75-cm2 culture flasks, the characteristic fluid-filled domes (hemicysts) were observed in control and SCAA-restricted media, further indication of bioenergetically healthy and polar cell sheets (Fig. 3). Domes were less numerous but larger in the SCAA-restricted media.

Fig. 3.

Phase contrast micrograph of 1 wk postconfluent cultures of control vs. sulfur-containing amino acid-restricted conditions. Cells were seeded into 75-cm2 tissue culture flasks and allowed to reach confluence (1 wk). Flasks were then refed in either control (100% cysteine/100% cystine/100% methionine) or sulfur-containing amino acid-restricted (50% cysteine/0% cystine/10% methionine) media on the day confluence was achieved, again 4 days later, and again at 1 wk postconfluence (2 wk postseeding). Results are typical and show the three-dimensional, fluid-filled domes characteristic of confluent LLC-PK1 cultures, with the domes being larger in the restricted condition (right). The control condition (left) resembles a typical 1 wk postconfluent LLC-PK1 cell sheet with numerous but very small domes; the restricted condition appears more like a typical freshly confluent LLC-PK1 cell sheet. Since doming is indicative (in LLC-PK1) of a healthy, polarized cell sheet that is actively and transepithelially transporting ions, these photographs suggest that the restricted media did not negatively affect the basic functions of LLC-PK1 cells as transporting epithelia. Scale bar = 250 μm.

Transepithelial electrophysiology and radiotracer flux studies.

After seeding cells into Millicells at confluent density and maintaining cell sheets in SCAA-restricted medium for a week, there was no effect observed on short-circuit current (Fig. 4, top). For LLC-PK1 cells, this indicates that transcellular Na+ transport (a function of apical Na+/glucose cotransport and basal-lateral Na+-K+-ATPase) was not significantly affected. However, a 30% increase in transepithelial electrical resistance was observed, which indicated decreased permeability of the TJ, and consequently the epithelium (Fig. 4, bottom).

Fig. 4.

Sulfur-containing amino acid-restricted cell sheets have comparable short-circuit currents (ISC) compared with controls but exhibit a significant increase in transepithelial electrical resistance (TER). Cells were seeded at confluent density onto permeable (0.4-μm pore size) Millipore-PCF filter ring assemblies and after 24 h kept on control (100% cysteine/100% cystine/100% methionine) (light gray bar) or sulfur-containing amino acid restricted (50% cysteine/0% cystine/10% methionine) (dark gray bar) conditions for 1 wk before potential difference (PD) and voltage readings were taken using electrodes in series with NaCl agar bridges and a modified Ussing chamber setup, respectively. ISC and TER were calculated using Ohm's law. Data shown are expressed as the means ± SE of n = 9 cell sheets. Significance was tested using Student's t-test, assuming equal variances.

This finding of enhanced barrier function was further tested using radiolabeled mannitol and polyethyleneglycol, both chosen for their inability to enter cells. Consequently, these molecules can only cross the epithelium by the paracellular route (48). The transepithelial diffusion rate of d-mannitol, a sugar alcohol, was found to be significantly decreased by over 60% in SCAA-restricted conditions (Fig. 5, top). Mannitol is a relatively small nonelectrolyte (MW 182), so a larger nonelectrolyte, polyethyleneglycol (PEG; MW 4000), was then studied. Paracellular, transepithelial permeability to polyethyleneglycol was not increased as a result of culture in SCAA-restricted medium (Fig. 5, bottom).

Fig. 5.

Sulfur-containing amino acid-restricted cell sheets exhibit a significant decrease in transepithelial 14C-mannitol flux (MW 180) but show no difference in transepithelial 14C-polyethylene glycol (PEG) flux (MW 4000) compared with controls. Cells were seeded at confluent density onto permeable (0.4-μm pore size) Millipore-PCF filter ring assemblies and after 24 h kept on control (100% cysteine/100% cystine/100% methionine) (light gray bar) or sulfur-containing amino acid restricted (50% cysteine/0% cystine/10% methionine) (dark gray bar) conditions for one wk. The radiotracer probe (either mannitol or PEG) was put in the dish media and allowed to incubate for 3 h, during which time samples were taken from the ring media at fixed intervals. Flux was calculated using scintillation counts. Data shown represent the means ± SE of n = 9 cell sheets. Significance was determined by Student's t-test, assuming equal variances.

The data presented in Fig. 5 indicate that (paracellular) flux of a smaller probe (mannitol) decreases, but flux of a larger probe (PEG) is unchanged. If one accepts a working model of the frontal “face” of the TJ as an impenetrable web punctuated here and there by pores of various sizes and of varying frequencies (e.g., number of pores per square nanometer of junctional “face”), the explanation of these data may be as follows. If there are a relatively small number (very low frequency) of large pores (e.g., allowing solutes greater than 5,000 Da to pass) through which a solute such as PEG can permeate, then the unchanged flux of PEG (as a function of methionine restriction) suggests that these pores do not change significantly in frequency or structure. However, if the smaller pores (e.g., those allowing only solutes less than 200 Da to pass), occur at a much higher frequency than the larger pores (e.g., 100 to 1), and their frequency decreases dramatically as a result of methionine restriction, the flux of the small probe, mannitol, will be markedly curtailed even as the flux of PEG is unaffected. Even though mannitol can more easily fit through the large pores—and even multiple mannitol molecules could go through simultaneously—the much higher frequency of the smaller pores (through which mannitol, but not PEG, can pass) makes the smaller pores a much greater factor in overall mannitol (small solute) flux. This explanation of course covers only (uncharged) nonelectrolytes, and the situation would be different for (charged) salts, amino acids, and other charged molecules due to the charged groupings of these claudin “pores.”

Effect of SCAA-restricted medium on cell density.

It is possible that an apparent decrease in transepithelial paracellular permeability can be due merely to decreased cell density rather than a permeability change in the TJs. This is so because decreased cell density (in an otherwise confluent cell sheet) results in a decreased density of paracellular spaces (TJs) per unit area, that is, less opportunity for paracellular flow (62). To assess this possibility, immunofluorescent labeling of the TJ was performed using an anti-ZO-1 antibody (see materials and methods) to observe the cell density (and junctional density) of cell sheets on filters used in the above permeability studies (Fig. 6). This allowed for easy determinations of cell density as reported in Table 1. Cell density in control medium and SCAA-restricted medium, however, was not significantly different. It follows that linear junctional density would likewise be similar under the two media conditions and, therefore, that differences in linear junctional density were not the mechanism for the observed decrease in permeability. This was then verified by two separate sets of studies.

Fig. 6.

Indirect immunofluorescence micrographs of control (left) vs. sulfur-restricted (right) cell sheets whose tight junctions have been labeled with Cy3-conjugated secondary antisera against the primary antisera for the junction-specific protein, ZO-1. Cells were seeded at confluent density onto permeable (0.4-μm pore size) Millipore-PCF filter ring assemblies and after 24 h kept on control (100% cysteine/100% cystine/100% methionine) (left) or sulfur-containing amino acid restricted (50% cysteine/0% cystine/10% methionine) (right) conditions for 1 wk. Filters were then excised from rings and labeled with the Cy3-conjugated secondary antisera against the primary antisera for the junction-specific protein, ZO-1. Scale bar = 20 μm.

View this table:
Table 1.

Cell density determined by immunofluorescence for ZO-1

Observation of unaltered cell density by the method above would suggest that the amount of cell material on a standard area of the filter assemblies would be the same. This was shown to be the case whether total DNA, total RNA, or total protein was used to assess the amount of cellular material (Table 2). This lack of difference highlights a change in the structure and permeability of the TJ as the cause of the altered transepithelial permeability and also underscores the subtle and seemingly rather directed changes being brought about in the LLC-PK1 cells by the restriction of SCAAs.

View this table:
Table 2.

Total DNA, RNA, and protein levels in control (100/100/100) versus restricted (50/0/10) conditions

A change in the structure of the TJ was tested by observing for changes in the relative abundance of various specific TJ proteins by PAGE and Western immunoblot (Fig. 7). The ubiquitous TJ protein, occludin, showed essentially no change in abundance in SCAA-restricted conditions. Occludin frequently remains unchanged in various stress conditions (18) There was likewise no significant change in abundance of certain TJ claudin proteins, namely, claudins-1 and -2, as verified by measurements of optical density of the TJ protein bands for several different passages of cells in control vs. SCAA-restricted media. Claudins-3 and -7, however, showed significant decreases in abundance, whereas claudins-4 and -5 increased significantly in abundance. The nature of the TJ barrier, in its component claudin proteins, had thus changed, and done so selectively, as a result of SCAA restriction.

Fig. 7.

Western immunoblots of integral tight junction proteins in sulfur-containing amino acid-restricted (left) vs. control (right) conditions. Cells were seeded into 75-cm2 tissue culture flasks and allowed to reach confluence (1 wk). Flasks were then refed in either control (100% cysteine/100% cystine/100% methionine) or sulfur-containing amino acid-restricted (50% cysteine/0% cystine/10% methionine) media on the day confluence was achieved, again 4 days later, and again at 1 wk postconfluence before being scraped and placed in Buffer A and treated as described in materials and methods. Results shown are from the membrane fraction (triton-soluble). All primary antisera were products of Zymed. A chemiluminescent-labeled secondary antibody was used in detection. The occludin band migrated between 65 and 70 kDa. All claudin bands migrated between 20 and 30 kDa. Immunoblots shown are representative of n = 4 runs using four different cell passages. The percentage increase/decrease represents the mean ± SE and was calculated by densitometry of the appropriate bands of the four separate PAGE and Western immunoblots of the membrane fraction of cells. No significant difference was found between sulfur-containing amino acid-restricted and control conditions for claudins -1, -2, and occludin, claudin-3 significantly decreased (34%), claudin-4 increased (126%), claudin-5 increased (87%), and claudin-7 decreased (33%), in the sulfur-containing amino acid-restricted cells. Significance was determined using Student's t-test, assuming equal variances.


In epithelial cell cultures cultured on impermeable substrata, the size and shape of domes have long been known to relate back to transepithelial transport across the cell sheet (1). In this study, we have observed that short-circuit current was unchanged as a result of SCAA restriction (Fig. 4), but a change in the number and size of domes also occurred (Fig. 3). Transepithelial transport is, however, only half of what is involved in dome formation, the other half being cell-cell and cell-substratum adhesion (44, 75, 76, 85). This present study on SCAA restriction is focused on permeability per se, and investigations concerning the effects of SCAA restriction on adhesion proteins such as E-cadherin or integrin were not performed, although SCAA-restriction-induced changes in these proteins and thereby in adhesion itself could certainly underlie the alterations in dome morphology that are seen here and should be a subject of future studies. As with so many other properties, the specific effects of SCAA restriction on cell-cell and cell-substratum adhesion have never been reported. In addition to effects of SCAA-restriction on adhesion properties of the cells, it is likewise possible that domes were also affected by the action of SCAA restriction on transepithelial transport. Studies of the effects of SCAA restriction on secretagogue “stimulation” of transepithelial transport (and dome morphology) would be useful here.

In addition to examining the effects of SCAA restriction on cell adhesion in future studies, the effects on cell polarization would also be promising investigations. It is well known that the TJ exerts a fence function around the polar epithelial membrane domains, as well as a gate function for the cell sheet as a whole (30). It is reasonable, therefore, that an effect of SCAA restriction on permeability or gate function, as shown in this manuscript, may also influence cell polarity (fence function) features. The LLC-PK1 cell line used in these present studies may be an especially good model in such fence function studies because of the well-described imperfect polarization of LLC-PK1 cells (21).

Restriction of sulfur-containing amino acids in the culture medium of the differentiated, polar renal epithelial cell line, LLC-PK1, was undertaken to approximate in cell culture the studies done with various animal models, wherein their diet was dramatically reduced in its methionine, cysteine, and cystine content. In rodent models, the purpose of this dietary change was to observe whether a life span extension could be achieved, similar to that observed with caloric restriction (45, 59, 80, 84). Our purpose in this present study was to observe whether SCAA restriction might alter—and in fact improve—the barrier function of a widely studied epithelial model. We were guided by the fact that epithelial barriers (and tissue fluid compartmentation) figure very prominently in the physiology of higher animal life, and compromise of epithelial barriers is a hallmark of a very wide variety of disease states, ranging from dust mite allergies to bacterial infections to cancer (67).

In this present study, as a result of SCAA restriction, barrier function was changed along the paracellular route (increased transepithelial electrical resistance and decreased transepithelial mannitol flux) but with surprisingly no change of transcellular sodium transport (short-circuit current). A structural change in the claudin composition of TJs paralleled this functional permeability change, with highly significant changes in the abundance of certain claudins (3, 4, 5, and 7) but no significant change in other TJ proteins (occludin and claudins- 1 and -2). Future studies on effects of SCAA restriction should not simply focus on the barrier proteins per se, but also on the linker proteins, such as ZO-1, ZO-2, ZO-3, which physically connect claudins and occludin to the cytoskeleton, as changes to these TJ-associated proteins are certainly known to affect TJ structure and permeability (23, 30, 32). Studies should also focus on the regulatory proteins (PKC, myosin light chain kinase, etc.) that are known to control the phosphorylation state and functionality of the overall tight junctional complex (40, 52, 57). Examination of freeze-fracture images of SCAA-restricted TJs could also be highly informative.

It is tempting to suggest a specific mechanism by which the decreases in claudins -3 and -7 and the increases in claudins-4 and -5 work in concert to produce the observed enhancement of barrier function in our SCAA-restricted cell culture model. Claudin-5, for example, is a major component of the TJs of the blood-brain barrier, and its transfection and overexpression in various cell culture models is known to correlate with increased barrier function as evidenced by a drop in paracellular permeability to the radiolabeled probes, inulin, and sucrose (7, 58). Additionally, induced expression of claudin-4 in Madin-Darby canine kidney II cells and LLC-PK1 cells significantly increases barrier function as measured by TER (78). Such results seem to correlate with our findings about claudins-4 and -5 in this study. However, the effect of given claudins on TJ permeability can vary widely depending on the model used and the conditions to which a model is subjected. Also, focusing on the changes in each claudin individually fails to address the heterotypic interactions between different claudins and the overall complexity of the TJ in general.

Investigations of the overall mechanism by which amino acid restriction alters cell phenotype have ranged from oxidants (6, 83) to the sirtuin genes (28, 29, 39) to proinflammatory cytokine levels and apoptotic activity (4, 60, 61, 70, 71), to the growth hormone/IGF axis (3, 65).

Among the growing number of disease states in which TJ changes are being observed, inflammation is prominent. Alterations in TJs have now been associated with collagenous colitis (9), psoriasis (82) multiple sclerosis/encephalomyelitis (5), Crohn's disease (24, 31), ulcerative colitis (26), and perhaps arthritis (17). Furthermore, in a different line of investigations, proinflammatory cytokines have been shown to alter epithelial barrier permeability by a combination of inducing apoptosis in individual cells and direct action on TJ structure and composition (25, 27, 36, 37, 51, 72, 81). It would be intriguing to learn in future studies what effect, if any, SCAA restriction might have on inflammatory diseases.

A recent, widely recognized theory of systemic aging suggests that the life-long burden of infectious disease and resultant increase of systemic inflammation is a major determinant of life span (13, 19, 20, 47). An ever increasing number of studies has correlated increases in mediators of systemic inflammation with age, as well as correlated age with inflammatory conditions per se (15, 34, 42, 66). Interestingly, caloric restriction is observed to not only extend life span but also alleviate the systemic surplus of proinflammatory mediators and reactive oxygen species that appear to link inflammation and aging (12, 33). Taken together, these findings indicate a potential coupling between diseases of aging and a systemic proinflammatory state in aging.

With the epithelial barriers of the respiratory, urinary, and gastrointestinal tracts being the greatest surface area boundary between the environment (with its pathogens and allergens) and the vasculature (with its immune system components), an aging-related, infectious disease-related, and inflammation-related compromise of these boundary barriers is noteworthy. Ability of dietary amino acid restriction to “repair” such barrier pathologies could prove clinically beneficial. In this study, we have attempted to show that a well-known dietary means of extending life span also happens to support epithelial barrier function at the level of the TJ. Future studies will aim to determine whether this holds true for epithelial cell types other than just renal, and whether it holds in animal tissue models, as well as cell culture models. A careful approach to the molecular mechanism behind this dietary regulation of TJs can then begin.


Support was provided by the Pennsylvania Department of Health, Sharpe-Strumia Research Foundation, National Institutes of Health (1R03-AG023246-01A1), and the Cancer Research and Prevention Foundation.


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


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 6a.
  8. 7.
  9. 8.
  10. 9.
  11. 10.
  12. 11.
  13. 12.
  14. 13.
  15. 14.
  16. 15.
  17. 16.
  18. 17.
  19. 18.
  20. 19.
  21. 20.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
  42. 41.
  43. 42.
  44. 43.
  45. 44.
  46. 45.
  47. 46.
  48. 47.
  49. 48.
  50. 49.
  51. 50.
  52. 51.
  53. 52.
  54. 53.
  55. 54.
  56. 55.
  57. 56.
  58. 57.
  59. 58.
  60. 59.
  61. 60.
  62. 61.
  63. 62.
  64. 63.
  65. 64.
  66. 65.
  67. 66.
  68. 67.
  69. 68.
  70. 69.
  71. 70.
  72. 71.
  73. 72.
  74. 73.
  75. 74.
  76. 75.
  77. 76.
  78. 77.
  79. 78.
  80. 79.
  81. 80.
  82. 81.
  83. 82.
  84. 83.
  85. 84.
  86. 85.
View Abstract