## ABSTRACT

The exchange of isotopic water, ^{2}H_{2}O and H_{2}^{18}O, has been studied in amphibian eggs. The experiments were carried out with ovarian and body cavity eggs of *Rana temporaria* and unfertilized eggs of *Ambystoma mexicanum*. The cytoplasmic diffusion coefficient for H_{2}^{18}O was found to be 4·6 × 10− ^{4} cm^{2}/s, somewhat higher than that for ^{2}H_{2}O, 3·4 × 10^{−6} cm^{2}/s. The total change in reduced weight, Δ*RW*;, during the isotope experiments was compared with the total amount of water in the egg cell, *m*. The ratio *&RW/m* was significantly higher than would be expected from calculations using ordinary water density values. The results are discussed in terms of different phases of structured cell water.

## INTRODUCTION

For many years it has been suspected that the water inside the living cell differs from bulk water with respect to its structure and various physical properties. The great interest in this problem is evidenced by a number of recent articles discussing the physiological role and importance of intracellular water (Walter & Hope, 1971; Hazlewood, 1973; Cooke & Kuntz, 1974; Wiggins, 1975; Cope, 1976; Kolata, 1976; Ling, 1977). In spite of all this attention, nobody has been able to provide a direct demonstration of the properties of the intracellular water in cells under normal physiological conditions. All that has been observed are a number of ‘anomalies’ in the properties of the living cells which can be accounted for on the assumption that part of the cell water is structured (‘bound’) in some way. The study of this problem is hampered by the fact that the living cell usually is so small that few analytical methods are applicable. This may be one reason why so many data on cell water concern amphibian eggs, these being among the largest cells known.

One anomaly in amphibian eggs is that they do not behave as perfect osmometers: not all of their water participates in the adjustment to changes in the osmolarity of the external medium. Thus, when a frog egg is placed in a isotonic solution only about 65% of the water content is osmotically active (Sigler & Janacek, 1969, 1971). It has also been found that the diffusion coefficient for water in cytoplasm does not vary smoothly as in bulk water, but exhibits a break at 16 ° C (Hansson Mild Lovtrup, 1974a). At this temperature, unexpected changes likewise occur in the internal hydrostatic pressure of eggs incubated in hypotonic solutions (Hansson Mild, Løvtrup & Bergfors, 1974) and in the apparent osmotic water permeability coefficient of frog ovarian eggs (Hansson Mild, Carlson & Løvtrup, 1974). Measurements of the relaxation times with NMR have shown a reduction relative to free water, which again suggests an increased structural order in the cell water in the frog egg (Hansson Mild, James & Gillen, 1972).

Several forms of structured water are known, at least thirteen polymorphic forms of ice exist, and all of them have densities different from that of ordinary water, one being lighter and the remaining ones heavier (Kamb, 1972). It might therefore be anticipated that if cell water is structured, then its density should differ from unity. In the present paper we present a method which, indirectly, may give information about the density of the cell water in the amphibian egg in the normal physiological state. By means of this method we have concluded that cell water, like most types of structured water, has a density higher than that of bulk water.

## MATERIAL AND METHODS

### Biological material

The experiments were carried out with ovarian and body cavity eggs of *Rana temporaria*, and ovarian and unfertilized eggs of *Ambystoma mexicanum*. The frogs were purchased from commercial dealers and kept at 5 °C until used. Ovulation was induced as described by Rugh (1952). The axolotls were raised in the laboratory, where they are kept at 16 ° C. Spawning was induced by injecting the female with 150 i.u. of human gonadotropin.

The ovarian eggs of both species were obtained by withdrawing surgically a small portion of an ovarium and removing the follicle membrane mechanically with forceps. In the unfertilized eggs of *Ambystoma* the jelly was removed with forceps. All eggs were kept in Ringer solution. In all cases healthy looking eggs were chosen for experimentation. Prior to each experiment the radius of the egg was measured with an optical screw micrometer, an operation having a standard deviation of 1 %. All experiments were carried out at 18 ° C.

### Isotope exchange method

The isotopic water exchange method has previously been used to determine the diffusion coefficient of water in the cytoplasm and the water permeability coefficient of the plasma membrane of the amphibian egg (Hansson Mild & Lovtrup, 1974a, *b)*. The theory of this method has been published earlier (Løvtrup, 1963; Bergfors, Hansson Mild & Løvtrup, 1970; Hansson Mild, 1972; Hansson Mild & Løvtrup,1974 *a*)

The exchange of isotopic water is followed by determination of the changes in the reduced weight (*RW*) of an egg placed in isotonic Ringer solution containing either 20% ^{2}H_{2}O or 20% H_{2}^{18}O, by means of the automatic diver balance (Bergfors *et al*. 1970). The exchange curve is registered in two ways, on a X-Y recorder, whet” the RW is plotted as a function of time, and on a paper tape with data points taken at 10 s intervals. The experimental curves are then compared with the theoretical expression for the exchange process (Hansson Mild, 1972) by a computer curve fitting procedure (see Hansson Mild & Lovtrup, 1974*a*).

The exchange process, started by loading the diver with an egg, causes a mechanical disturbance of the system and reliable readings are therefore obtained only after approximately 30 s. This means that the initial value of the curve, *RW*_{0}, the reduced weight at *t* = o, and hence also the total change in *RW* during the isotope exchange process, must be calculated by the computer. We have estimated the error thus introduced by weighing the same eggs on two different balances, one floating in full strength Ringer solution made up with ordinary water, another in a medium consisting in part of isotopic water. On the first balance the value of *RW*_{0}is obtained directly, and on the other one, as mentioned, indirectly. A series of experiments of this kind has shown that the error involved in the indirect method is less than 1 %.

After completion of the isotope exchange, the egg is removed from the diver and dried in an oven at 105 ° C to constant weight (2 h), and the dry weight determined on a microbalance.

### Calculations

*RW*, of an object is its weight when submerged in a liquid medium, i.e. the wet weight minus the buoyancy. It is thus given by: where

*V*is the volume of the object (in our case an amphibian egg),

*ρ*its density and

*M*its wet weight, while

*ρ*

_{m}is the density of the medium. Rearranging the terms of equation (1) we obtain: The density of the medium is known from pycnometer determinations and

*RW*from diver balance experiments. With the volume obtained from determinations of the radius, we may thus calculate the wet weight. The mass of water in the egg,

*m*, is then obtained as the difference between the wet weight and the dry weight,

*dw*. An analysis of the errors involved in the outlined procedure gives a total error of about 4% in

*m*.

The total change in reduced weight, Δ *RW*, during the isotope exchange process, is obtained from the experimental curve by the computer curve-fitting procedure mentioned earlier. The error involved in this determination of Δ *RW* is estimated to be about 3%. An illustration of the steps involved in our method is given in Fig. 1, the reproduction of an experimental curve for a body cavity egg placed in Ringer containing 20% H_{2}^{18}O.

*RW*, can be expressed as where the index

*i*refer to the isotope-containing medium, and 0 to the ordinary Ringer solution,

*V*

_{i}—

*V*

_{0}thus being the change in volume. In equation 3, the index

*0*may either mean zero, i.e. the value of the parameter at time

*t*= 0, or the letter 0, i.e. the value of the parameter in ordinary Ringer solution, either interpretation being equivalent in the following text. Furthermore, we have where the last term refers to the volume and the density of the cell water. Assuming the dry weight, as well as its associated volume, to be constant during the isotope exchange process, these parameters may be eliminated from equation (3), giving If v

_{i}=

*v*

_{0}

*+*Δ

*v*, we arrive at The volume change, Δ

*v*, may have two sources, being associated either with the isotope-exchange experiment proper or with an osmotic imbalance between egg and medium.

*P*is the osmotic water permeability coefficient, Δ

*C*is the concentration difference across the membrane, and the partial molar volume of water. Assuming a 5 % difference in the tonicity between the ordinary Ringer and the isotope-containing solution gives .

*P*is of the order of 2× 10

^{−4}cm/8 for a body cavity egg (Hansson Mild

*et al*. 1974). For small change in radius equation (7) may be replaced by If for Δ

*t*we insert the duration of an isotope exchange experiment (20 min) we get Δ

*r*∼ 5 × 10

^{−5}cm. The corresponding volume change Ar is about 5 × 10

^{−8}cm

^{3}(0·2%).

The volume change due to the isotope exchange proper can be calculated approximatively by assuming that the isotope-containing water has the same density in the cytoplasm as in solution, and furthermore that the exchange takes place on a molemole basis. This leads to v_{4} = 1·0007 *v*_{0} for the ^{2}H_{2}O solution.

If we assume that both of the above mentioned volume changes act in the same direction, we thus have at most Δ*v* ∼ 0·3% or about 7× 10^{−8} cm^{3}. The density difference in equation (6), , is of the order 0·02 g/cm^{3} and thus the contribution of a volume change is of the order 0·1 *μ*g, as compared, with a total Δ *RW* of 40 *μg*. We therefore feel entitled to neglect the influence of volume changes on Δ *RW*.

^{−2}for the

^{2}H

_{2}O solution and 2·47 × 10

^{−2}for the H

_{2}

^{18}O solution. These values are hereafter referred to as ‘bulk values’.

## RESULTS

The experiments with the isotope H_{2}^{18}O in *Rana* oocytes gave an average cytoplasmic water diffusion coefficient of 4·6 × 10^{−8} cm^{2}/s in the ovarian eggs. The water permeability coefficient was too high to be measured with our method. In the body cavity eggs the membrane permeability is 2·7 × 10^{−8} cm/s. These results should be compared with the results previously obtained by us on these eggs, using heavy water as a tracer, where we found a diffusion coefficient of 3·4× 10^{−6} cm^{2}/s in the ovarian eggs and a permeability coefficient of 1·4×10^{−4}cm/s in the body cavity eggs (Hansson Mild & Lovtrup, 1974a). Thus, the isotope H_{2}^{18}O gives Rightly higher values for the measured parameters, as might be expected since the self-diffusion coefficient for H_{2}^{18}O is known to be higher than for ^{2}H_{2}O.

In Table 1 are listed some representative experimental values of the water content and the change in reduced weight (Δ*RW*). Figs. 2 and 3 show plots of *μ* Δ *RW* versus *m* for the isotope exchange with ^{2}H_{2}O and H_{2}^{18}O, respectively. The average values for the ratio Δ *RW/m* were not significantly different between the two species and the different egg types employed in this study. The mean value for Δ *RW/m* was therefore calculated for all experiments with the isotope ^{2}H_{a}0 and was found to be 2·6 × 10^{−2} with a standard deviation of 0·2 × 10^{−2} (n = 36). As shown above, the bulk value is 2·17 × 10^{−2}, and the difference between this and the experimental value is highly significant (P < o·001). The mean value for the H_{2}^{18}0 experiment was 2·8 × 10^{−2}, standard deviation 0·2 × 10^{−2} (n = 16). Also for this isotope there is a significant deviation from the bulk value 2·47 ·10^{−2}*(p <* 0·001).

In *Ambystoma* the water content varies from about 48% in ovarian eggs to 60% in unfertilized eggs. This provides an opportunity to establish the relationship between egg density and water content, as shown in Fig. 4. The line is the leastsquare regression line (correlation coefficient = 0·99). Assume that the egg density |s given by

*X*is the water content (

*m/M*) . The regression analysis of the data from Fig. 4 gives the equation and thus by comparing equations (11) and (12) we get

*ρ*

_{g}= 1·208 and

*ρ*

_{o}=1·015g/cm

^{3}.

## DISCUSSION

*RW*and

*m*. A comparison shows that the experimental values of the ratio Δ

*RW/m*are significantly larger than those obtained from equation (9) when the ‘bulk values’ are used. The slope

*k*of the lines in Figs. 2 and 3 is Inserting the experimental values for

*k*into this equation, and taking into account that we are dealing with 20% isotope solutions, we obtain the following relations between the densities of pure isotope water and ordinary water isotope in the cell cytoplasm: and Our data pertaining to the density of cell water are represented by equations (13) – (15). Unfortunately, these are given as ratios and differences, a circumstance that evidently limits the interpretation. In interpreting equation (13) we first assume that Δ p is the same as for bulk solutions; inserting the experimental values of

*k*we thus obtain

*ρ*

_{0}

*=*0·85 and 0·89 for the

^{2}H

_{2}O and H

_{2}

^{18}O experiments, respectively. These values are unlikely for two reasons: (1) the value for ordinary ice is 0·92 and (2) the values should be identical, and their difference exceeds that allowed for by the errors of our experimental method. It must therefore be assumed that Δ p is larger for cell water than for bulk water. The smallest acceptable value for Δ p is obtained if it is assumed that water has its normal bulk density in the egg cytoplasm. On this assumption equation (13) gives a density of 1·129 for

^{2}H

_{2}O in the cell cytoplasm, as contrasted to the bulk value at 1·105. The corresponding values for H

_{2}

^{18}O are 1·139 and1·124, respectively.

There is one possible way to estimate the density of the cell water, namely, from the regression line in Fig. 4 according to equation (12). We are, of course, aware of that the value of 1·015 th^{us} obtained is highly uncertain. Nevertheless, if it is accepted, the value for cytoplasmic ^{2}H_{2}O is 1·146 and for cytoplasmic H_{2}^{18}O it is 1·156. And it should be recalled that these values are the averages for the total water content, which implies that if our values represent phases of structured water these should have even higher densities.

The only value for the density of intracellular water recorded in the literature is that published by Pocsik (1967). Working with frog muscle, this author estimated the density as a function of the water content, registering an increase from 1·01 at 80% water to 1·33 at 5% water. The former value agrees with our result. However, it should be observed that these experiments were not carried out on living tissue.

A number of observations have been made on the effect of ^{2}H_{2}O on biological systems. For example Ussing (1935) found that the development of frogs eggs was retarded at 20% ^{2}H_{2}O and inhibited at concentrations higher than 30%. Garby & Nordqvist (1955) noted a decrease in the conduction velocity in isolated frog nerves. Gross & Spindel (1960*a, b)* studsied the influence of ^{2}H_{2}O on cell division of seaurchin eggs and found that it was blocked at concentrations higher than 75%. If the cells were not kept too long in the deuterium-rich medium, the block was reversible. In a study of the temperature-tolerance of adult *Drosophila* flies, Pittendrigh & Cosbey (1973) found that ^{2}H_{2}O increased the heat resistance of the flies. All these experiments have been interpreted as the result of a stabilizing effect of ^{2}H_{2}O on various biological macromolecules. If this stabilization follows from the isotopic water being more structured, more strongly bound to cytoplasmic components, and if this structuring is associated with an increase in density, then our result showing a relatively higher density for ^{2}H_{2}O than for H_{2}O constitutes a corroboration of the interpretation of these previous observations.

In several articles (e.g. 1971, 1973, 1976) Drost-Hansen has discussed the structural and functional aspects of water near biological interfaces. The properties of this vicinal water appear to be distinctly different from those of bulk water, reflecting structural differences. He suggests that possible candidates for this vicinal water are clathrate hydrates and high pressure ice polymorphs, all of which have densities higher than 1. In his opinion these vicinal structures extend from a minimum of 0·o01 up to 0·*μm* from the interface.

Many indirect observations thus suggest that cell water has properties different from bulk water and that it is structured. If this is true, it might have a density different from bulk water. We have tested this possibility and found it corroborated. But we are the first to admit that our method as well as our results are open to severe criticism. However, we feel justified in publishing this work since it is the first attempt to measure the density of water in a living cell.

## ACKNOWLEDGEMENTS

We gratefully acknowledge the assistance of Mr André Berglund and Mrs Monica Sandstrom in various aspects of the present work. The work was partly supported by the Royal Swedish Academy of Science.

## REFERENCES

*J. exp. Biol*

*A. Rev. Biophys. Bioeng*

*Physiol. Chem. Phys*

*Chemistry of the Cell Interface B*

*Ann. N.Y. Acad. Sci*

*Physics and Chemistry of Liquids*(Journal, Gordon & Breech, 1977)

*Acta physiol, scand*

*Ann. N.Y. Acad. Sci*

*Ann. N.Y. Acad. Sci*

*Bull. math. Biophys*

*(Rana pipiens)*

*J. cell. Physiol*

*Biochim. biophys. Acta*

*J. exp. Biol*

*J. Membrane Biol*

*J. exp. Biol*

*Ann. N.Y. Acad. Sci*

*Water and Aqueous Solutions, Structure, Thermodynamics and Transport Processes*

*Science, N.Y*

*Mol. cell. Biochem*

*J. theoret. Biol*

_{2}O of the temperature-tolerance of adult

*Drosophila*

*Proc. natn. Acad. Sci. U.S.A*

*Acta biochim. biophys. Acad. Sci. Hung*

*Experimental Embryology*

*Physiologia bohemoslov*

*Biochim. biophys. Acta*

*Scand. Archiv Physiologic*

*Progr. Biophys. mol. Biol*

*J. theor. Biol*