Brought to you by Portland Press Ltd.
Published on behalf of the International Federation for Cell Biology
Cancer Cell death Cell cycle Cytoskeleton Exo/endocytosis Differentiation Division Organelles Signalling Stem cells Trafficking
Cell Biology International (2006) 30, 74–77 (Printed in Great Britain)
On the osmotically unresponsive water compartment in cells
Gary D. Fullertona*, Kalpana M. Kanala and Ivan L. Cameronb
aDepartment of Radiology, The University of Texas Health Science Center at San Antonio, Floyd Curl Drive, San Antonio, TX, USA
bDepartment of Cellular and Structural Biology, The University of Texas Health Science Center at San Antonio, Floyd Curl Drive, San Antonio, TX, USA


Differences in colligative properties (freezing point, boiling point, vapor pressure and osmotic behavior) between water in living cells and pure bulk water were investigated by re-evaluating reports of the osmotic behavior of mammalian cells. In five different animal cells, osmotically unresponsive water (OUW) values ranged from 1.1 to 2.2g per g dry mass. Detailed analysis of human red blood cell (RBC) data indicates a major role for hemoglobin OUW-values, aggregation and packing in cell volume regulation that can be explained for the first time in relevant molecular terms.

Keywords: Osmosis, Interfacial water, Osmotically unresponsive water, Cell water.

*Corresponding author.

1 Introduction

Cells swell or shrink in response to the tonicity (osmolarity) of their environment. These observations have led many biologists to use cell volume or gravimetric water content measurements to study in vivo osmotic behavior. Because the plasma membrane is relatively permeable to water, water moves into or out of the cell down its concentration gradient. Thus, if an impermeable extracellular solute (osmolyte) is added outside the plasma membrane, the cell will behave as an osmometer that shrinks or swells in hypertonic or hypotonic osmolyte solutions. Although cells generally respond in this way, careful evaluation shows that no cell behaves like an ideal osmometer; that is to say, there is always a fraction of cellular water that is osmotically unresponsive under physiological conditions (Fig. 1). This led us to question whether osmotic cell behavior can be analyzed using the Fullerton molecular osmolyte model developed for BSA in the accompanying papers (Cameron et al., 2006; Fullerton et al., 2006).

Fig. 1

Generalized figure shows the equilibrium water content of cells plotted against the inverse of the osmotic pressure. Linear regression of the data derived solid line when extrapolated (dashed line) to the 1/π=0 intercept gives the amount of osmotically unresponsive (non-ideally behaved) water fraction in the cell. None of the cellular water is, strictly speaking, completely non-responsive or inactive as all of the water can be removed by drying in a vacuum oven at 100°C or by applying extremely high compressive forces.

2 Methods

The total water and OUW contents for five different cell types were obtained from the literature: pig lens fiber cells in situ (Cameron et al., 1988), fully grown Xenopus oocytes (Cameron et al., 1990), unfertilized sea urchin eggs (Merta et al., 1986), frog sartorius muscle (Ling and Negendank, 1970) and human erythrocytes (Heubrusch et al., 1985). Human erythrocyte data were also obtained by the methods described previously (Cameron et al., 1988). In addition, the OUW values for hemoglobin in bulk (extracellular) solution were obtained from Adair (1928) and Fullerton et al. (1993). The data were analyzed by the methods described in Fullerton et al. (2006).

3 Results and discussion

Table 1 lists comparative data on OUW content or I value in g water per g dry mass for the five cell types. The I values range from 1.1 to 2.2g per g or from 31 to 92% of the total water content of the cell type examined. These values overlap with the lowest measured I values for BSA (Fullerton et al., 2006), but are generally lower. This is partly due to a centrifugation artifact, described below.

Table 1.

Extent of osmotically unresponsive water (OUW) (g H2O/g dry mass) for five cell types compared to the extracellular value for hemoglobin

Cell typeWater contentOUW% OUWReference
Pig lens fiber cells in situ2.42.292Cameron et al. (1988)
Xenopus oocyte, full grown1.51.280Cameron and Fullerton (1990)
Sea urchin egg, unfertilized2.41.354Merta et al. (1986)
Frog sartorius muscle3.51.131Ling and Negendank (1970)
Human erythrocyte2.041.6a78Heubrusch et al. (1985) and present paper
Hemoglobin1.7Adair, 1928 and Fullerton et al., 1993
a Data from the isotonic range, see Fig. 2.

One report in the literature on the human red blood cell (RBC) is most amenable to detailed analysis of cell osmotic behavior (Heubrusch et al., 1985). First, most of the cell dry weight content of the human erythrocyte is attributable to a single protein, i.e. the cell impermeable protein is >90% hemoglobin. This simplification approximates the situation with the BSA experiments described in Fullerton et al. (2006). Heubrusch et al. (1985) also made an important improvement over other studies reported in the literature, including our own, in their method for measuring cell water content. They avoided packing or pelleting the cells by centrifugation, thus eliminating issues of extracellular space and, more importantly, the artifactual influence of the head of hydrostatic pressure caused by centrifugation. Hydrostatic pressure adds directly to the equilibrium osmotic pressure of the simulated plasma. This centrifugation artifact is both systematic and quite large where high g-forces are used. Centrifugation experiments, while useful, must be assessed with caution because decreases in the apparent cell water content caused by the g-force occur at specific applied osmotic pressures that can cause large decreases in apparent I values (Bogner et al., 1998).

Heubrusch et al. (1985) circumvented the centrifugation artifact by using Coulter counter measures of cell volume to allow direct calculation of cell water content. In addition, their osmotic pressure measurements cover a much wider range and thereby reveal interesting sources of non-linear osmotic behavior. Their results gave nearly identical RBC volumes for equivalent osmolalities applied with either NaCl or sucrose. Given the isotonic RBC water content of 1.99g water per g dry weight, and assuming that volume changes are due to solely to water, the data were reanalyzed (Fig. 2A), using the procedures introduced in Fullerton et al. (2006). For comparison we also plot the evaluation of in vitro osmotic pressure data for purified sheep hemoglobin (Adair, 1928; Fig. 2B).

Fig. 2

A) Mean water content in g water/g dry mass of human erythrocytes suspended in varying concentrations of NaCl expressed as reciprocal of the osmotic pressure 1/π(atm) as reanalyzed from literature data. The graph reveals three linear regions (hypotonic right, isotonic middle and hypertonic left). The intercept of each regions at the zero intercept 1/π(atm) yields an OUW fraction as listed in the text. The slope of each region was used to calculate the effective molecular weight of that region (see text for details). B) Molecular model analysis of in vitro osmotic pressure data on sheep hemoglobin has a single slope (molecular weight) and intercept I-value (OUW).

While the osmotic data on purified hemoglobin can be described by a single linear segment, the RBC response requires three linear segments that we have labeled hypertonic, isotonic and hypotonic regions. The multiple three-segment regression fit to the RBC data is excellent (r2=0.998, n=19). The Adair data on hemoglobin give an even better fit (r2=0.9998, n=18). It is encouraging to find that simple linear relationships describe cellular osmotic behavior over such a wide range of conditions. These equations should be useful with or without the underlying molecular basis. The effective molecular weights (Ae) and I values are listed on the figure. The slope of the hemoglobin best fit line gives the correct protein molecular weight and the intercept gives I=1.7g per g. The hemoglobin I value is similar to the BSA value when the protein is folded into the most compact particulate form (Fullerton et al., 2006). All three sections of the RBC plot, however, yield much higher slopes, which implies much lower molecular weights and we wondered whether this is reasonable.

3.1 Hypotonic region

Analysis of the data in the hypotonic region yields an OUW fraction of I=0.75g H2O per g dry mass and Ae=3800Da. Two molecular explanations must be considered. If the protein is denatured, we would expect to see AeAs as observed, but in no instance with BSA was the decrease so large (Fullerton et al., 2006). In addition, denaturation would cause a large increase in the I value, but in fact we measure a decrease relative to the hemoglobin I value=1.7g per g. This result makes sense only if the RBC contains a large number of small and intermediate sized molecules that are membrane impermeable, such that the mean Ae=>iniMi/ni includes these additional cosolutes. As small molecules have I values near zero, the sequestered water on the hemoglobin will be shared by all the molecules in the cell, which would explain the reduction from Ihemo=1.7–0.75g per g in the RBC. Finally, it is obvious from fundamental biology that functioning RBCs contain large numbers of small and intermediate sized molecules of salts, metabolites, sugars and many other solutes that serve as internal osmolytes.

3.2 Isotonic region

In the central isotonic region where the RBC resides in vivo, the shift in slope gives Ae=8600Da and I=1.6g per g, which approaches the value for native hemoglobin. At the transition point with the hypotonic region the RBC has only 2.23g water per g dry mass, approaching the hydration of molecular hemoglobin, Ihemo=1.7g per g. These facts suggest that small molecules associate with the protein as the cell water content approaches the OUW-value limit measured independently for hemoglobin. The isotonic region continues until the cell water content equals the I value for hemoglobin, where a final change in slope occurs.

3.3 Hypertonic region

Transition to this region occurs when less than 1.7g water per g solid is available in the RBC. Interpretation of the linear segment gives Ae=2100Da and I=1.0g per g, but these values are no longer meaningful molecular parameters as the solids are aggregated and the amount of cellular water is less than Ihemo=1.7g per g. The experiment is now similar to the hydration force experiments of Leikin et al. (1997, 1993) and more properly represents evaluation of the energy necessary to remove water from the surface of hemoglobin.

4 Conclusions

The molecular osmotic model therefore gives a plausible molecular description of the RBC osmometer responses to variable osmotic pressures in the extracellular environment; previously there has been no such plausible account. Although protein conformation changes play a role in cell volume regulation, it appears that the I values of the primary protein constituents dominate the cell water content. This value is, however, modulated by the concentration of smaller free-floating cosolutes such as salts and quantitatively important metabolites. It is already known that hemoglobin in the RBC is less mobile than hemoglobin at the same concentration in vitro (Bogner et al., 2005; Cameron et al., 1991). This strongly suggests that the hemoglobin is partially aggregated. Prouty et al. (1985) also demonstrated that human sickle cell hemoglobin is subject to massive polymerization, which results in changes in osmotic pressure, sickle cell shape and volume decrease. Both sets of observations are consistent with the molecular osmolyte model.

Polymerization of proteins (e.g. bovine serum albumin, lysozyme, actin and tubulin) displaces water of hydration or the water sheath on globular proteins (Fullerton et al., 1987). Thus, protein aggregation and disaggregation and tightness of polymer packing in vivo appear to be significant factors in regulating the extent of OUW in erythrocytes (see also Bogner et al., 2005 in this volume).

We conclude that the mean macromolecular I value of the cell constituents is the most important determinant of cell water content. Osmotically unresponsive water on macromolecules such as intracellular proteins provides the primary and energetically most favorable mechanism for maintaining equilibrium cell water content. The cell water consists of OUW plus a smaller secondary amount, which is attributable to water modulated by the passive osmotic response resulting from differences between the internal and external small molecule osmolytes. Large cellular inundations of water in disease states probably reflect changes in the OUW value that result from increased protein denaturation, as occurs in cell injury. Large increases in cell water content are predicted because of increased protein SAS and the solute/solvent interaction parameter I. Both predictions are consistent with observations made during cell breakdown and repair. In contrast, protein aggregation or ligand formation would cause decreased cellular water because of the decreasing I value, as occurs in sickle-cell anemia.


Adair GS. A theory of partial osmotic pressures and membrane equilibia, with special reference to the application of Dalton's law to hemoglobin solution in presence of salts. Proc R Soc London Series A 1928:120:573-603
Crossref   1st Citation   2nd   3rd  

Bogner P, Csutora, P, Cameron, IL, Wheatley, DN, Miseta, A. Augmented water binding and low cellular water content in erythrocytes of camel and camelids. Biophys J 1998:75:3085-91
Crossref   Medline   1st Citation  

Bogner P, Csutora, P, Cameron, IL, Wheatley, DN, Miseta, A. . Bogner P, Miseta A, Berent Z, Schwarcz A, Kotek G, Repa I. Osmotic and diffusive properties of intracellular water in camel erythrocytes: the effect of hemoglobin crowdedness. Cell Biol Int 2005 (in press).
1st Citation   2nd  

Cameron IL, Contreras, E, Fullerton, GD, Kellermayer, M, Ludany, A, Miseta, A. Extent and properties of nonbulk “bound” water in crystalline lens cells. J Cell Physiol 1988:137:125-32
Crossref   Medline   1st Citation   2nd   3rd  

Cameron IL, Cox, LA, Liu, XR, Fullerton, GD. Maintenance and mobility of hemoglobin and water within the human erythrocyte after detergent disruption of the plasma membrane. J Cell Physiol 1991:149:365-74
Crossref   Medline   1st Citation  

Cameron IL, Fullerton, GD. A model to explain the osmotic pressure behavior of hemoglobin and serum albumin. Biochem Cell Biol 1990:68:894-8
Medline   1st Citation  

Cameron IL, Merta, P, Fullerton, GD. Osmotic and motional properties of intracellular water as influenced by osmotic swelling and shrinkage of Xenopus oocytes. J Cell Physiol 1990:142:592-602
Crossref   Medline   1st Citation  

Cameron IL, Kanal, KM, Fullerton, GD. Role of protein conformation in pumping water into and out of a cell. J Cell Biol Int 2006:30:78-85
Crossref   1st Citation  

Fullerton GD, Finnie, MF, Hunter, KE, Ord, VA, Cameron, IL. The influence of macromolecular polymerization of spin-lattice relaxation of aqueous solutions. Magn Reson Imaging 1987:5:353-70
Crossref   Medline   1st Citation  

Fullerton GD, Zimmerman, RJ, Kanal, KM, Floyd, LJ, Cameron, IL. Method to improve the accuracy of membrane osmometry measures of protein molecular weight. J Biochem Biophys Methods 1993:26:299-307
Crossref   Medline   1st Citation   2nd  

Fullerton GD, Kanal, KM, Cameron, IL. Osmotically unresponsive water fraction on proteins: non-ideal osmotic pressure of bovine serum albumin as a function of pH and salt concentration. J Cell Biol Int 2006:30:86-92
Crossref   1st Citation   2nd   3rd   4th   5th   6th   7th  

Heubrusch P, Jung, CY, Green, FD. The osmotic response of human erythrocytes and the membrane cytoskeleton. J Cell Physiol 1985:122:266-72
Crossref   Medline   1st Citation   2nd   3rd   4th   5th  

Leikin S, Parsegian, VA, Yang, W, Walrafen, GE. Raman spectral evidence for hydration forces between collagen triple helices. Proc Natl Acad Sci USA 1997:94:11312-7
Crossref   Medline   1st Citation  

Leikin S, Parsegian, VA, Rau, DC, Rand, RP. Hydration forces. Annu Rev Phys Chem 1993:44:369-95
Crossref   Medline   1st Citation  

Ling GN, Negendank, W. The physical state of water in frog muscle. Physiol Chem Phys Med NMR 1970:2:15-23
1st Citation   2nd  

Merta PJ, Fullerton, GD, Cameron, IL. Characterization of water in unfertilized and fertilized sea urchin eggs. J Cell Physiol 1986:127:439-47
Crossref   Medline   1st Citation   2nd  

Prouty MS, Schechter, AN, Parsegian, VA. Chemical potential measurements of deoxyhemoglobin s polymerization, determination of the phase diagram of an assembling protein. J Mol Biol 1985:184:517-28
Crossref   Medline   1st Citation  

Received 25 July 2005; accepted 30 September 2005


ISSN Print: 1065-6995
ISSN Electronic: 1095-8355
Published by Portland Press Limited on behalf of the International Federation for Cell Biology (IFCB)