Reflection Coefficient For NaCl Research Articles

Nanofiltration of saline oil-water emulsions: Combined and individual effects of salt concentration polarization and fouling by oil

This study employed NF270 nanofiltration membranes and SDS-stabilized hexadecane to identify mechanisms of membrane fouling by highly saline oil-water emulsions. Concentration dependencies of NaCl reflection coefficient σ and NaCl permeability coefficient B were measured and used to determine the separate contributions of osmotic pressure and fouling to the overall flux decline. The NaCl permeability coefficient asymptotically converged to the same steady state value across a range of feed salinities and concentration polarization conditions. The measured near-hyperbolic dependence of the reflection coefficient on the transmembrane concentration differential (σΔπm≈ const) negated the effect of concentration polarization on permeate flux. Oil caused an abrupt decrease in permeate flux but only for more saline feeds (seawater level and higher), which was interpreted as a result of membrane surface sealing by coalesced oil. Headloss analysis showed that over the longer term, the additional hydraulic resistance due to a layer of oil droplets on the membrane surface became the dominant fouling mechanism.

Effects of Proteoliposome Composition and Draw Solution Types on Separation Performance of Aquaporin-Based Proteoliposomes: Implications for Seawater Desalination Using Aquaporin-Based Biomimetic Membranes

Aquaporins are a large family of water transport proteins in cell membranes. Their high water permeability and solute rejection make them potential building blocks for high-performance biomimetic membranes for desalination. In the current study, proteoliposomes were prepared using AquaporinZ from Escherichia coli cells, and their separation properties were characterized by stopped-flow measurements. The current study systematically investigated the effect of proteoliposome composition (lipid type, protein-to-lipid ratio (PLR), and the addition of cholesterol) on water permeability and NaCl retention. Among the various lipids investigated, 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC)-based proteoliposomes were found to have excellent osmotic water permeability and NaCl reflection coefficient values. Increasing the PLR of DOPC proteoliposomes up to 1:200 increased their osmotic water permeability. However, further increase in the PLR reduced the osmotic water permeability probably due to the occurrence of defects in the proteoliposomes, whereas the addition of cholesterol improved their osmotic water permeation likely due to defects sealing. The current study also investigated the effect of major dissolved ions in seawater (e.g., Mg(2+) and SO(4)(2-)) on the stability of proteoliposomes, and design criteria for aquaporin-based biomimetic membranes are proposed in the context of desalination.

Solvent drag measurement of transcellular and basolateral membrane NaCl reflection coefficient in kidney proximal tubule.

The NaCl reflection coefficient in proximal tubule has important implications for the mechanisms of near isosmotic volume reabsorption. A new fluorescence method was developed and applied to measure the transepithelial (sigma NaClTE) and basolateral membrane (sigma NaClcl) NaCl reflection coefficients in the isolated proximal straight tubule from rabbit kidney. For sigma NaClTE measurement, tubules were perfused with buffers containing 0 Cl, the Cl-sensitive fluorescent indicator 6-methoxy-N-[3-sulfopropyl] quinolinium and a Cl-insensitive indicator fluorescein sulfonate, and bathed in buffers of differing cryoscopic osmolalities containing NaCl. The transepithelial Cl gradient along the length of the tubule was measured in the steady state by a quantitative ratio imaging technique. A mathematical model based on the Kedem-Katchalsky equations was developed to calculate the axial profile of [Cl] from tubule geometry, lumen flow, water (Pf) and NaCl (PNaCl) permeabilities, and sigma NaClTE. A fit of experimental results to the model gave PNaCl = (2.25 +/- 0.2) x 10(-5) cm/s and sigma NaClTE = 0.98 +/- 0.03 at 23 degrees C. For measurement of sigma NaClbl, tubule cells were loaded with SPQ in the absence of Cl. NaCl solvent drag was measured from the time course of NaCl influx in response to rapid (less than 1 s) Cl addition to the bath solution. With bath-to-cell cryoscopic osmotic gradients of 0, -60, and +30 mosmol, initial Cl influx was 1.23, 1.10, and 1.25 mM/s; a fit to a mathematical model gave sigma NaClbl = 0.97 +/- 0.04. These results indicate absence of NaCl solvent drag in rabbit proximal tubule. The implications of these findings for water and NaCl movement in proximal tubule are evaluated.

Transepithelial water permeability in an in vitro model of renal cysts.

Renal cysts develop from microscopic tubules and may enlarge progressively several thousandfold. Sustained epithelial proliferation, intracavitary fluid accumulation, and extracellular matrix remodeling are central elements in a multistep process that leads to the formation and enlargement of cysts. MDCK cells suspended within medium-hydrated collagen gels grow to form spherical, monolayered, fluid-filled cysts that enlarge steadily. Vasopressin and other agents that increase intracellular levels of cAMP stimulate the rate of MDCK cyst growth and net fluid/solute secretion when added to defined medium in vitro. In this model, net fluid secretion is the only means by which fluid can accumulate within the cyst cavity. We used this cyst-forming line of epithelial cells to evaluate several membrane transport properties that are important in the coupled movements of solute and water in the process of secretory fluid transport. Individual cysts were microdissected from collagen gels, held by a micropipet in a thermostated chamber, and examined at a high magnification by video microscopy. Transepithelial water flow was initiated by rapidly exchanging the bath medium with hyperosmotic solutions. Net water flux, Jv, determined from the initial rate of decrease in cyst diameter, was proportionate to the transmembrane osmotic gradient of NaCl or raffinose; the reflection coefficient for NaCl was indistinguishable from 1.0. Osmotic water permeability (cm3/cm2/osm/min x 10(-6)) was 739 +/- 99 (N = 11) in medium augmented by an NaCl concentration of 100 mosmol/kg. Hydraulic conductivity (Pt), estimated in control cysts, was 6.8 +/- 0.9 microns/s, a value similar to that of medullary and cortical thick ascending limbs of Henle.(ABSTRACT TRUNCATED AT 250 WORDS)

Model study of placental water transfer and causes of fetal water disease in sheep

The purpose of the computer simulation was to use experimentally measured parameters of placental water transfer to compute conceptual water acquisition during the last third of ovine pregnancy and to evaluate the possible role of each of these parameters in the pathophysiology of polyhydramnios. Total conceptual water at birth was almost insensitive to the value of the placental filtration coefficient. It was more sensitive to fetal and maternal placental blood flows, the concentrations of actively transported or metabolically produced solutes in fetal plasma (bicarbonate, fructose, alpha-amino acids, urea, lactate), and the hydrostatic pressure difference across the placental barrier. It was quite sensitive to the NaCl reflection coefficient and permeability. We conclude that the actively transported or produced solutes in fetal plasma constitute a primary driving force. The opposing diffusion gradient of NaCl, and to a much lesser extent that of glucose, are essential to restrain the process. The causes of polyhydramnios in this species are, in order of probability, an increase in the placental diffusion permeability of NaCl, a decrease in the placental reflection coefficient for NaCl, or an increase in the concentration of an osmotically effective solute in fetal plasma, by active transport or metabolic production.

Difference between hypertonic NaCl and NaHCO3 as osmotic diuretics in dog kidneys.

To compare the osmotic inhibitory effects of NaCl and NaHCO3 on proximal tubular fluid reabsorption, plasma osmolality was raised by 40 mosmol kg-1 H2O by infusing hypertonic NaCl and NaHCO3 in volume-expanded dogs receiving ethacrynic acid. In five dogs studied at constant plasma pH 7.5, both NaCl and NaHCO3 reduced water reabsorption by 29 +/- 2%. However, NaCl infusion reduced bicarbonate reabsorption by 31 +/- 2%, whereas bicarbonate reabsorption remained unchanged during NaHCO3 infusion. In six dogs, bicarbonate reabsorption was kept constant during NaCl and NaHCO3 infusion by adjustments of plasma pH. At similar glomerular filtration rates (42.4 +/- 2.9 ml min-1), water reabsorption was 28.7 +/- 1.7 ml min-1 in the control period, 29.4 +/- 2.5 ml min-1 during hypertonic NaCl infusion and 20.6 +/- 1.2 ml min-1 during hypertonic NaHCO3 infusion. Therefore, NaCl did not reduce proximal tubular water reabsorption by a direct osmotic effect. By calculating the regression coefficient for the relationship between measured chloride reabsorption and maximal convective chloride flux, the effective reflection coefficient for NaCl averaged 0.11 +/- 0.01. The combination of a low reflection coefficient and high permeability may explain why hypertonic NaCl is not an osmotic diuretic.

NaCl reflection coefficients in proximal tubule apical and basolateral membrane vesicles. Measurement by induced osmosis and solvent drag

Two independent methods, induced osmosis and solvent drag, were used to determine the reflection coefficients for NaCl (sigma NaCl) in brush border and basolateral membrane vesicles isolated from rabbit proximal tubule. In the induced osmosis method, vesicles loaded with sucrose were subjected to varying inward NaCl gradients in a stopped-flow apparatus. sigma NaCl was determined from the osmolality of the NaCl solution required to cause no initial osmotic water flux as measured by light scattering (null point). By this method sigma NaCl was greater than 0.92 for both apical and basolateral membranes with best estimates of 1.0. sigma NaCl was determined by the solvent drag method using the Cl-sensitive fluorescent indicator, 6-methoxy-N-[3-sulfopropyl]quinolinium (SPQ), to detect the drag of Cl into vesicles by inward osmotic water movement caused by an outward osmotic gradient. sigma NaCl was determined by comparing experimental data with theoretical curves generated using the coupled flux equations of Kedem and Katchalsky. By this method we found that sigma NaCl was greater than 0.96 for apical and greater than 0.98 for basolateral membrane vesicles, with best estimates of 1.0 for both membranes. These results demonstrate that sigma NaCl for proximal tubule apical and basolateral membranes are near unity. Taken together with previous results, these data suggest that proximal tubule water channels are long narrow pores that exclude NaCl.

Transition of permeability properties along the descending limb of long-loop nephron.

The isolated segments of the hamster descending limb of the long-loop nephron (LDL) were perfused in vitro to demonstrate the axial heterogeneity with respect to permeability properties. When a NaCl gradient from the lumen to bath was present, the lumen-negative diffusion voltage (VD) was generated in the upper portion (LDLU). When the VD was measured stepwise along the axis of tubules, the magnitude of the VD decreased in the portion within 0.5 mm before the border between the outer and inner medulla in most cases, indicating that a gradual functional transition to the lower portion (LDLL) occurs along the descending limb. The lumen-to-bath flux coefficients (K1----b) for Na+ were 14.3 +/- 3.7 X 10(-7) and 2.4 +/- 0.8 X 10(-7) cm2/s in the LDLU and LDLL, respectively. The K1----b for urea were 1.7 +/- 0.6 X 10(-7) and 7.9 +/- 4.1 X 10(-7) cm2/s, respectively. The LDLL was highly permeable to water, with osmotic permeability coefficient being 1,693 +/- 517 X 10(-9) cm2.s-1.atm-1. The reflection coefficients for NaCl and urea were not different from unity. From these observations, we conclude that the functional transition occurs along the LDL from the segment with a high Na+ permeability to that with a low Na+ permeability.

Video measurement of basolateral NaCl reflection coefficient in proximal tubule.

Isolated, lumen-collapsed, proximal and distally occluded segments of rabbit S1 and S2 proximal tubule were equilibrated in isotonic NaCl or isosmotic raffinose medium and then exposed acutely to hypotonic or hypertonic raffinose or NaCl solution. The result was a water flux per millimeter tubule length, JVo, across the basolateral cell membranes and a consequent cell swelling or shrinkage that could be measured by a video technique in the initial 0.1 s or less after a change from steady state. The cell volume change was proportional to the applied osmolality difference, delta pi, and differed consistently with the solute employed. From the equation JVo/delta pi = sigma LpA, where sigma is the basolateral membrane reflection coefficient for the osmotic solute used and LpA is the membrane hydraulic conductivity per millimeter tubule length, and from the assumption that sigma raffinose = 1, sigma NaCl was obtained by dividing the JVo/delta pi values from the NaCl studies by those from the raffinose studies. For both S1 and S2 segments, sigma NaCl was found to be approximately 0.5. A similar value was obtained from the rate of cell shrinkage immediately after isosmolar exchange of raffinose for NaCl medium.

In vivo and in vitro studies of urinary concentrating ability in potassium-depleted rabbits.

The factors responsible for the urinary concentrating defect associated with the potassium-depleted (KD) state are uncertain. The present studies were designed to, first, determine whether a urinary concentrating defect exists in potassium-depleted rabbits and, second, to use the technique of in vitro perfusion to evaluate directly the antidiuretic hormone (ADH) responsiveness of cortical collecting tubules (CCT) in this setting. Feeding female New Zealand White rabbits a potassium-deficient diet for 2 wk caused a significant fall in plasma potassium levels in both the ad-libitum and controlled water intake groups (P less than 0.001). Muscle potassium content after 2 wk of potassium restriction fell from 45.6 +/- 0.9 to 29.0 +/- 1.2 meq/100 g fat-free dry solids (P less than 0.001). Renal papillary sodium content fell significantly from a control value of 234.6 +/- 8.0 to 182.46 +/- 10.0 meq/kg H2O after 2 wk of potassium restriction. Maximal urinary osmolality measured after 12 h of dehydration and 1.25 U pitressin IM was significantly decreased in rabbits after 2 wk of potassium restriction in both the ad-libitum and controlled water intake groups (P less than 0.001). The relationship between plasma potassium concentration and maximum urinary osmolality was significantly correlated in both the ad-libitum and controlled water intake groups, r = 0.73 and 0.68 (P less than 0.001), respectively. In addition, refeeding KD rabbits with normal chow for 1 wk resulted in normalization of both plasma potassium levels and urinary concentrating ability. CCT from control and KD rabbits were perfused in vitro at 25 degrees C. The hydraulic conductivity coefficient, Lp, was significantly reduced at all doses of ADH tested in tubules from KD rabbits when compared with control tubules. In addition, the maximal hydraulic conductivity in tubules from KD rabbits when tested with 200 microU/ml ADH at 37.5 degrees C was only 23% of control values (P less than 0.05). Furthermore, this reduced ADH responsiveness persisted when the bath potassium was elevated from 5 to 20 mM. The reflection coefficient for NaCl when compared with raffinose was 0.91 in tubules from KD animals. Thus, these data suggest that the ADH-resistant urinary concentrating defect associated with potassium depletion is due, at least in part, to a diminished responsiveness of the CCT to ADH. Therefore, further studies were designed to investigate the cellular steps involved in this abnormal response. There was no difference in the 8-para-chlorophenylthio cyclic AMP induced hydroosmotic response between CCT from KD and control rabbits. Since the cAMP-induced hydroosmotic response was similar between KD and control CCT, experiments were performed to evaluate the contribution of phosphodiesterase (PDIE) activity by using the potent PDIE inhibitor isobutylmethylxanthine (10(-4) and 10(-3)M) in the presence of ADH (200 U/ml). Although Lp was increased by PDIE inhibition in CCT from both control and KD animals, the overall hydroosmotic response in CCT from KD rabbits was still significantly reduced when compared with controls. The final experiments used forskolin to evaluate further the adenylate cyclase complex. The resulting hydroosmotic response in CCT from KD rabbits was almost identical to that obtained in controls. In conclusion, these data suggest that the decreased responsiveness of CCT from KD rabbits to ADH involves a step at or proximal to the stimulation of the catalytic subunit of adenylate cyclase, and that PDIE activity makes no contribution to this abnormal hydroosmotic response.

Estimation of whole-body capillary transport parameters from osmotic transient data.

A mathematical model has been developed of short-term, extrarenal, whole-body fluid volume regulation. The Kedem and Katchalsky equations are used to describe rapid movements of crystalloid and colloid solutes and water between five fluid compartments. Simulation results showing rapid cell volume changes following a hyperosmotic crystalloid infusion demonstrated the necessity of considering the effect of cellular water shifts in osmotic transient experiments. From measurements of plasma volume and osmolality in acutely nephrectomized dogs subjected to isosmotic and hyperosmotic NaCl infusions and with the model, six parameters related to capillary membrane transport of water and NaCl were estimated. The mean capillary filtration capacity from six experiments was estimated as 0.01 ml.min-1.mmHg-1.100 g-1 of dog. This increased about threefold due to the hyperosmolality. Mean values of capillary diffusion capacity and osmotic reflection coefficient for NaCl were 0.37 cm3.s-1.100 g-1 and 0.087, respectively. These results support the use of the osmotic transient approach and a mathematical model to study the role of microvascular transport in whole-body fluid volume regulation.

Gallbladder epithelial cell hydraulic water permeability and volume regulation.

The hydraulic water permeability (Lp) of the cell membranes of Necturus gallbladder epithelial cells was estimated from the rate of change of cell volume after a change in the osmolality of the bathing solution. Cell volume was calculated from computer reconstruction of light microscopic images of epithelial cells obtained by the "optical slice" technique. The tissue was mounted in a miniature Ussing chamber designed to achieve optimal optical properties, rapid bath exchange, and negligible unstirred layer thickness. The control solution contained only 80% of the normal NaCl concentration, the remainder of the osmolality was made up by mannitol, a condition that did not significantly decrease the fluid absorption rate in gallbladder sac preparations. The osmotic gradient ranged from 11.5 to 41 mosmol and was achieved by the addition or removal of mannitol from the perfusion solutions. The Lp of the apical membrane of the cell was 1.0 X 10(-3) cm/s . osmol (Posm = 0.055 cm/s) and that of the basolateral membrane was 2.2 X 10(-3) cm/s . osmol (Posm = 0.12 cm/s). These values were sufficiently high so that normal fluid absorption by Necturus gallbladder could be accomplished by a 2.4-mosmol solute gradient across the apical membrane and a 1.1-mosmol gradient across the basolateral membrane. After the initial cell shrinkage or swelling resulting from the anisotonic mucosal or serosal medium, cell volume returned rapidly toward the control value despite the fact that one bathing solution remained anisotonic. This volume regulatory response was not influenced by serosal ouabain or reduction of bath NaCl concentration to 10 mM. Complete removal of mucosal perfusate NaCl abolished volume regulation after cell shrinkage. Estimates were also made of the reflection coefficient for NaCl and urea at the apical cell membrane and of the velocity of water flow across the cytoplasm.

A model for transepithelial fluid transport.

A number of experimental observations which are not easily, or at all, accommodated within the framework of the natriocentric theory for transepithelial fluid transport are discussed. These are shown to be not merely accommodated but actually predicted by a simple model for fluid transport in which the major osmotic force leading to fluid transport derives from solutes other than NaCl. The model is then analyzed for feasibility in light of existing data relating to reflection coefficients for NaCl and hydraulic conductivity. Certain lines of evidence for specific active sodium transport are critically reexamined.

Pulmonary microvascular filtration and reflection coefficients in two- and four-week-old rabbits

Several researchers have determined that pulmonary capillaries in young mammals are significantly more permeable to water and lipid-insoluble solutes than are the pulmonary capillaries in the adult counterparts. Our purpose was to quantify the microvascular fluid flux parameters in 2- and 4-week-old rabbit lungs and to compare our results with adult and newborn data previously reported. Using an isolated, Ringer-perfused lung preparation, we measured the values for the microvascular filtration coefficient and the reflection coefficients for NaCl and albumin. We found a filtration coefficient value of 2.57 × 10 −4 ml/sec-cm H 2O-m 2 for the 2-week-old population. For the 4-week-old rabbit we evaluated the filtration coefficient to be 3.18 × 10 −4 ml/sec-cm H 2O-m 2. These values are normalized for the estimated exchange area, and neither is significantly different from the adult value. The reflection coefficients for NaCl and albumin were measured using an osmotic transient technique. Correcting the albumin data for flow across the endothelial cell membrane, we arrived at the following values for the pulmonary microvascular reflection coefficient of 2- and 4-week-old rabbits, respectively: σ cap alb = 0.17, and σ cap alb = 0.23. Based on these values we estimated the equivalent pore radii of large pores in the pulmonary microvascular wall to be approximately 175 Å in the 2-week-old, and 145 Å in the 4-week-old rabbits. These results indicate that in the rabbit K f/m 2 has reached the adult value by 2 weeks of age; however, the number of pores increases, and the pore radius decreases, continuously even beyond 4 weeks of age.

Permeability changes in Necturus proximal tubule during volume expansion.

The permeability of Necturus proximal tubule to hydrophilic nonelectrolytes of varying molecular size was studied under control conditions and during isotonic expansion of the animal's extracellular volume. Transepithelial permeability was measured in perfused tubular segments under conditions of zero net water flux. During volume expansion, tubular permeability to urea increased slightly, whereas mannitol decreased slightly and permeability to sucrose was significantly decreased. Volume expansion had a greater effect on osmotic flow parameters; the NaCl reflection coefficient decreased from 0.64 to 0.47 (summer animals) and from 0.41 to 0.27 (winter animals). Osmotic water flux and hydraulic conductivity increased but only in the lumen-to-capillary direction. Reflection coefficients of nonelectrolytes measured at the apical surface were reduced during volume expansion for probing molecules greater than 3 A in radius and were unchanged for smaller molecules, less than 3 A, suggesting two pore populations. We propose that an increase in tight-junction permeability can account for modification of osmotic flow parameters, whereas the whole thickness of the epithelium, particularly the intercellular space, plays the dominant role in regulation of diffusional permeability.

Water and ion permeability of a native collagen membrane

With the help of a ribonucleoprotein it is possible to precipitate collagen in a layer of fibers with a 700 Å period. As collagen is a constituent of many membrane systems in the body, it seemed interesting to investigate the permeability of ions and water through a native collagen membrane. The experiments were carried out with the help of an acryl glass apparatus, where an osmotic pressure, a hydrostatic pressure difference or both can be maintained between the two bulk phases separated by the membrane. The diffusion coefficients for NaCl and KCl were found to be comparable with those in other biological membranes ( D s = 9 · 10 −7 cm 2 · s −1 ) whereas there is difference of more than three orders of magnitude in the hydraulic permeability ( L p = 6 cm 4 · J −1 · s −1 ). Volume flow measurements caused by an osmotic gradient indicated that the reflection coefficient for NaCl and KCl is very small. In hydrostatic pressure experiments, the membrane shows a preferred direction for volume flows which seems to have something to do with the mode of preparation of the membrane.

Microvascular pressure measurements by cannulation: Independency of concentration gradients and deviations in micro-servonulling

Small, steady-state departures from zero reference in servonulling pressure measuring systems were determined under static conditions by immersing the tips of microcannulae into a normal saline solution at atmospheric pressure. These positive and negative deviations from atmospheric pressure appearing when the servo loop is closed, are independent of changes in concentration of NaCl in the test solution and consequently not related to diffusion phenomena. Such dependence is precluded by thermodynamic considerations that would require a nonzero reflection coefficient for NaCl as a property of the microcannular tips. Therefore, the small steady-state departures from zero can undoubtedly be traced to a hydraulic defect in the system.

Sodium chloride and water transport in the descending limb of Henle.

The unique membrane characteristics of the thin descending limb of Henle (DLH) play an integral part in the operation of the countercurrent system. We examined these properties in vitro by perfusing isolated thin descending limbs of rabbits. Active transport of NaCl was ruled out by failure to demonstrate either net transport or transmembrane potential difference when perfusing with isosmolal ultrafiltrate of the same rabbit serum as the bath. Transmembrane potential was zero, and net fluid transport was -0.07 +/-0.06 nl mm(-1) min(-1), which also is not significantly different from zero. Passive permeability coefficient for Na(P(Na)) was determined from the disappearance rate of (22)Na from isosmolal perfusion solution. P(Na) was surprisingly low, 1.61 +/-0.27 x 10(-5) cm sec(-1), a figure which is significantly less than P(Na) in the proximal convoluted tubule (PCT). Reflection coefficient for NaCl (sigmaNaCl) was measured by perfusing the tubule with Na-free raffinose solution in a bath of rabbit serum to which sufficient NaCl was added to obtain conditions of zero net fluid movement. The measured sigmaNaCl of 0.96 +/-0.01 is significantly greater than sigmaNaCl in the PCT. Water permeability to osmotic gradients (L(p)) was determined by perfusing with ultrafiltrate of rabbit serum in a bath made hyperosmotic by addition of either 100 mOsm raffinose or NaCl. L(p) with raffinose was 1.71 +/-0.15 x 10(-4) ml cm(-2) sec(-1) atm(-1) and with NaCl 1.62 +/-0.05 x 10(-4) ml cm(-2) sec(-1) atm(-1), indicating much greater water permeability than in the PCT. In each case the measured increase in osmolality of the collected fluid was primarily due to water efflux without significant influx of solute. The finding of low permeability to sodium and high permeability to water is consonant with the hypotheses that high interstitial concentration of Na in the medulla generates an effective osmotic pressure which results in concentration of the fluid as it courses through the DLH primarily by abstraction of water without significant net entry of NaCl.