I. INSTRUMENTAL CALIBRATION AND ANALYSIS OFRANDOMLY GROWING CULTURES
ALICE WARLEY, J. STEPHENDepartment of Microbiology, South West Campus, University of Birmingham BI5 5TT,U.K.
A. HOCKADAY AND T. C. APPLETONPhysiological Laboratory, Downing Street, Cambridge, U.K.
SUMMARY
Cryo-ultramicrotomy and X-ray microanalysis were used to study the elemental composition ofHeLa S3 cells. Quantitation was achieved by reference to elemental standards of known concentra-tion made up in 25 % gelatin. Analysis of standards showed linear calibration for each of theelements studied: Na, P, S, Cl, K. Standardization was validated by comparing flame-photometricanalysis of gelatin containing sodium potassium tartrate with that of X-ray microanalysis. Freeze-dried sections of cells showed good morphology and analysis of whole sections of the cells showedthat K/Na varied in individual cells. Low K/Na could not be ascribed to cell damage or to thesequestering of Na in any particular subcompartment of the cells. Treatment with ouabain causedchanges in levels of all the elements studied and resulted in a low K/Na ratio in all cells.
INTRODUCTION
X-ray microanalysis of biological tissue is becoming an established method for inves-tigating the inorganic content of cells and tissues (for a review, see Moreton, 1981).Studies involving quench-freezing, cryo-ultramicrotomy and subsequent X-raymicroanalysis of the frozen-hydrated or freeze-dried specimens have shown that thesetechniques can be applied to the investigation of the distribution of diffusible elements.Several detailed studies have been published on electrolyte distribution in transportingepithelia, e.g. in molluscan mantle epithelium (Appleton & Newell, 1977; Appleton,Newell & Machin, 1979), toad urinary bladder (Ricked al. 1978a; Civan, Hall &Gupta, 1980), frog skin (Rick, Dorge & Von Arnim, 19786, c) and rabbit ileum(Gupta, Hall & Naftalin, 1978). For such tissues knowledge of ion distribution aids inunderstanding mechanisms of salt and water transport. Another tissue that has beenwidely studied is muscle. Saetersdal, Myklebust, Justesen&Engedal (1977) suggestedthat calcium — important in heart muscle contraction — is sequestered in mitochondria.However, subsequent studies in both vascular smooth muscle tissue (Somlyo, Somlyo& Shuman, 1979) and cultured aortic muscle (James-KrackeeZ al. 1980) indicated thathigh levels of calcium in mitochondria are characteristic of damaged cells. Laterstudies on frog striated muscle (Somlyo et al. 1981) showed calcium to be localized interminal cisternae of the sarcoplasmic reticulum when this muscle is in the restingstate. These studies show that it is now possible to study in detail subcellular locationof ions and hence to relate these more clearly to specific functions.
15 CEL60
218 A. Warley, J. Stephen, A. Hockaday and T. C. Apple ton
Apart from erythrocytes (Kirk, Bronner, Barba & Tosteson, 1978; Kirk, Lee &Tosteson, 1978; Jones, Johnson, Gupta & Hall, 1979) few investigations have beenreported on the elemental composition of cell suspensions, as opposed to cells ascomponents of tissues. The lack of studies on cells in tissue culture is perhaps surpris-ing in view of the widespread use of tissue culture cells as experimental systems. Inthis paper we report the intracellular concentrations of P, S, Cl, K and Na in HeLaS3 tissue culture cells as revealed by X-ray microanalysis.
MATERIALS AND METHODS
CellsHeLa S3 cells were routinely grown as monolayers in plastic flasks (Falcon) in suspension
medium supplemented with 10% foetal calf serum, 1 % antibiotic/antimycotic, 15mM-Hepes, 1%non-essential amino acids and 2 mM-glutamine. For experimental purposes cells were removed fromthe monolayer using disodium EDTA and grown in suspension on an orbital shaker. Cell viabilitywas checked using Trypan Blue. All cell culture material was obtained from Gibco Europe Ltd,Paisley, Scotland. For details see Nome, Wolstenholme, Howcroft & Stephen (1982). Humanerythrocytes suspended in 150mM-NaCl, 15mM-Mops, 5 miu-glucose (pH7-5) were supplied byMr D. Fincham, Physiological Laboratory, Cambridge.
Ouabain treatment of cells
HeLa S3 cells growing in suspension at 5 X 10s cells/ml were treated with 0-5 miu-ouabain(Sigma); samples were taken at 2-5 h and 6h for microanalysis.
Freezing and sectioning
Cells grown in suspension were centrifuged (200 g for 5 min, at room temperature). Drops of thethick pellet were frozen on silver stubs in liquid nitrogen slush. Frozen sections approximately300 nm thick were cut at —68 to —70 °C, mounted on Formvar-coated nickel grids and freeze-driedin a nitrogen atmosphere at the same temperature for 2-3 h. After allowing them to warm to roomtemperature in sealed vials in an atmosphere of dry nitrogen the grids were coated with a thick layerof carbon under a vacuum as described in detail elsewhere (Appleton, 1977, 1978).
AnalysisCarbon-coated grids were examined in an AEI EMMA 4 analytical electron microscope using
60 kV accelerating voltage and a beam current of 4nA measured with a Faraday cage. For studiesat a subcellular level a probe size of 0-5 jim was used; when average values for the whole section ofa cell were sought, then probe size was increased to cover the whole cell (approximately 10/ttn).Analysis proceeded for a live time of 100 s. Spectra were obtained using an energy-dispersive systemconsisting of a Kevex detector and a Link system 290 multichannel analyser. Spectra, stored on disc,were processed using the Quantem/FLS program (on loan from Link Systems Ltd, HighWycombe, England) for deconvolution and background subtraction by least-squares fitting ofprefiltered spectra (Statham, 1977). This program was also used for quantification of data by thecontinuum method of Hall (Hall, 1979; Hall & Gupta, 1979). The concentration of the elementpresent is proportional to the ratio of characteristic counts (peak—background (P—b)) to the numberof counts in the continuum (W) contributed by the specimen, after the contribution from the Ni gridand Formvar film had been removed. To make a good estimate of continuum, counts were collectedover a wide range 4-16 KeV. Quantification was achieved by comparison with standards of knowncomposition. Correction was made for absorption of Na X-rays within both specimen and Standardusing the absorption correction facility in the Quantem program.
StandardsStandards consisting of solutions of inorganic salts in 25 % gelatin were prepared. Concentrations
of sodium and potassium in the standards were checked by flame photometry (Instrument: Corning
X-ray microanalysis ofHeLa S3 cells 219
455). Drops of standard solutions were then frozen, sectioned and analysed using methods identicalto those used for the cells. Calibration curves for each element to be studied were constructed.Gelatin (25 %), without any additions, was also analysed to check for contaminating elements; itwas found to contain measurable amounts of sulphur and calcium only. Because of this, peak/continuum ratios for standards containing sulphur were corrected for the amount of this elementpresent in the blank. Calcium standards were prepared in gelatin that had been dialysed againstdistilled water. Validation of the quantitation was performed by analysing standard sodiumpotassium tartrate/gelatin preparations treated identically to the other standards.
Electron microscopyTransmission electron micrographs were taken on a Philips 300 electron microscope and used to
facilitate analysis on EMMA.
RESULTS
Morphology
The appearance of freeze-dried sections of HeLa S3 cells is shown in Figs 1—3.Detail was apparent despite the fact that this material was unfixed and unstained, andhad been carbon-coated. Major subcellular compartments were clearly visible. Thesecells possess a prominent dense nucleolus (nu, Figs 1, 3), which could be easilyidentified. The nucleus (Figs 1-3) appears granular; it was not easy to distinguishbetween the light and dark areas of the nucleus that are seen in conventional electronmicroscopy. The nuclear membrane (me, Fig. 3) could be distinguished. Thecytoplasm had fewer paniculate inclusions and detail of cytoplasmic organelles, whichis normally apparent in conventional electron microscopy, could not be seen. How-ever, mitochondria (m) were distinguished as dense organelles within the cytoplasm.Small areas of ice-crystal damage could sometimes be seen, as in the nucleus in Fig.3, and as gaps between mitochondria and the surrounding cytoplasm. Gross ice-crystal damage was observed in cells that occupied the more central regions of thepellet; such damage was manifest as large holes. Damaged cells appeared electron-dense, lacked underlying morphological detail and were not used for analysis.
Analysis of standards
Plots of (P—b)/W versus concentration of P, S, Cl, K, Na are shown in Fig. 4.Linear calibration was obtained for each of the elements. Points in Fig. 4 are meansof at least ten determinations. For elements of atomic number greater than that ofsodium, the standard deviation of the mean was usually between 3 and 5 % of thevalue of the mean, but occasionally it was as high as 10%. Such values are in agree-ment with those found by other authors for analysis of similar standards (Sauberman,Beeuwkes & Peters, 1981). Standard deviations of the mean for sodium were usually20-30 % of the mean. This is due to the low counting efficiency for sodium, causedby absorption of the low-energy X-rays in the beryllium window of the detector. (Thishas been discussed elsewhere; Somlyo, Shuman & Somlyo, 1977.) Estimations of[Na] in samples were therefore subject to greater error than estimations for otherelements. The efficiency of detection of Mg is lowered, but to a lesser extent, byabsorption of the X-rays, but since the HeLa S3 cells were grown in medium low in
220 A. Warley, J. Stephen, A. Hockaday and T. C. Appleton
n
-^iPIRLifc^ ^^H
Figs 1 and 2. For legend see p. 222.
X-ray microanalysis of HeLa S3 cells
r221
Fig. 3. For legend see p. 222.
222 A. Warley, J. Stephen, A. Hockaday and T. C. Appleton10 T
0-8--
0-6 - -
W
0-4 +
0-2--
100 200 300 400Electrolyte (mmol/kg dry weight)
600
Fig. 4. Calibrated standards for Na(#), P(A); S(») , Cl( V) and K(O), obtained fromanalysis of freeze-dried sections of 25% gelatin with known electrolyte composition.
Table 1. Measurement of [Na] [hQ by X-ray analysis in sodium potassium tartratestandards
Measured [Na](mmol/kg dry weight)
Measured [K] Calculated value(mmol/kg dry weight) n (mmol/kg dry weight) Mean K/Na
(a)(b)(c)
168136348 1 88169145
218153433 1 82195133
111118
184368182
1-2710-26l-22±0-161-1710-16
Standard solutions in gelatin were prepared, frozen, sectioned and then analysed in EMMA.Calculated values of K and Na were based on composition by weight of standard added to gelatin(a) and (b), and flame photometry of same solutions (c). Mean K/Na is the mean of individual ratiosobtained in these analyses. Concentrations are given as mmol/kg dry weight ± s.E. of the mean.n, number of observations.
divalent cations, Mg concentrations were found to be below the levels of detectabilityby our system, even though linear calibration was obtained for this element and lowlevels of Mg have been detected in other tissues (Dr M. Kendall, personal commun-ication). Calcium, also, was not detected in our cells despite removal of the overlap-ping potassium K/3 peak by the Quantem program during processing of the spectra.
The analysis of gelatin containing sodium potassium tartrate is shown in Table 1.
Figs 1-3. Transmission images of freeze-dried sections of HeLa S3 cells. Sections of300 nm thickness were cut at —68 to — 70°C freeze-dried and coated with a thick layer ofcarbon, c, cytoplasm; n, nucleus; nu, nucleolus; me, nuclear membrane; m, mitochon-dria. Figs 1, 2, X20 000; Fig. 3, X6000.
Fig. 5. Elemental composition of: A, randomly growing; and B,C, ouabain-treated (B, 2h;c, 6h) HeLa S3 cells. Each point represents a value from one cell.
X-ray microanalysis of HeLa S3 cells
1000
so0
800
I, 700-- .- al 3 600-- t 'O 500-- 0) Y
3 400-- E E 300--
200
loo
P S CI K N a K/ Na
Fig. 5
- C .. --
-- . . z ..
. . .. .. f? --
-- --
9. I I I I I
- 4
-- 3
9 z
-- 2 >
1
224 A. Warley, jf. Stephen, A. Hockaday and T. C. Appleton
Table 2. Elemental composition of whole sections ofHeLa S3 cells
n
50
P(mmol/kg)
315 ± 19
Ouabain treatmentt •
10t-
10
= 2-5h462 ±46
= 6h517 ±41
S(mmol/kg)
100 ±11
109 ±13
243 ± 19
Cl(mmol/kg)
317±23
596 ± 29
676 ± 55
K(mmol/kg)
322 ±17
151 ±15
100 ± 7
Na(mmol/kg)
331 ±26
662 ± 38
868 ±58
MeanK/Na
1-24
0-22
0-12
Values are means of all values from stated number of observations and are expressed as mmol/kgdry weight ± S.E. of the mean, n, number of observations; t, time (h) of ouabain treatment. Themean K/Na values are the means of K/Na for individual cells. In the case of the normal cells thevalue 1 -24 reflects the fact that 27/50 had K/Na ^ 1, and excludes the bias of one very high Na valuefor one cell, which 19 reflected in the value 331 ± 26.
In all cases the [Na] was lower than the calculated value, the difference between theanalysed value and the expected value being between 5 and 9 %. Mean [K] was higherthan the calculated value, the mean difference in this cases being between 7 and 18 %.In individual analyses [Na] and [K] varied together, i.e. higher [Na] was accom-panied by higher [K], thus some of the variation observed may be due to unevendistribution of the salt in the gelatin. Overall K / N a ratios were between 17 and 27 %of the expected, i.e. calculated, value.
Analysis of human erythrocytes
Human erythrocytes suspended in 150mM-NaCl, 15mM-Mops, 5 mM-glucose(pH7-5) were frozen, sectioned and analysed as a control for the effect of freezingprocedures. Twenty cells were analysed using a probe size of 0-5 fim. Mean [K](mmol/kg dry weight) was 199 ± 25 ( S . D . ) , and the mean [Na] 27 ± 17 ( S . D . ) . T h e K /Na ratio of 7-5: 1 compares favourably with results that have been previously obtainedfor human erythrocytes (Appleton, 1979).
Analysis ofHeLa S3 cells
Results for the analyses of whole sections of HeLa S3 cells are shown in Fig. 5A andTable 2. Concentrations of all elements studied varied from cell to cell. If the resultsare dealt with as a whole (Table 2) sodium and potassium levels appear to be equal.On an individual cell basis it is seen (Fig. 5A), that 27 out of 50 (54%) cells studiedshowed a K / N a ratio of 1—4, whereas in 23 cells this ratio was 0-5—1. Potassiumconcentrations were usually between 200 and 450 mmol/kg dry weight, but sodiumconcentrations showed a wider spread, the majority being between 100 and 500 mmol/kg dry weight.
Cells from the same population were analysed after treatment with the drugouabain, an inhibitor of the membrane-associated sodium pump. Treatment with
X-ray microanalysis ofHeLa S3 cells 225
ouabain resulted in an increase in intracellular sodium and a decrease in potassiumconcentrations (Table 2, Fig. 5B). After 2-5 h of drug treatment the intracellularpotassium concentration had fallen to half, and after 6h to one third, of the initialaverage concentrations, respectively. These changes resulted in a mean K/Na ratiofor all the cells analysed of 0-23 ± 0-05 after 2-5 h, and 0-12 ± 0-02 after 6 h treatment,respectively. These ratios are well below those recorded for cells with low K/Na ratiosin the untreated cells. Mean Cl concentrations also rose after ouabain treatment.Concentrations of both P and S were also higher after 6 h, presumably due to leakageof medium into the cells.
To investigate whether the low K/Na ratio in untreated normal cells was due tosequestering of sodium in any particular sub-compartment of the cells, concentrationsof elements were investigated in the compartments of a limited number of cells. Thecompartments studied were cytoplasm, nucleus and nucleolus. Although mitochon-dria were frequently seen in the sections, these were not included in the study ofcellular compartments, as often the diameter of the mitochondrion was below that ofthe minimum probe size used (0*5 [im). For this reason areas of cytoplasm containingmitochondria were avoided, as they could differ substantially in ion content from thesurrounding cytoplasm. Results for subcellular compartments of HeLa S3 cells areshown in Table 3 and Fig. 6A-C. With the exception of some nucleoli each point isthe mean of at least three determinations from the different areas of each cell; in somesections nucleoli were small and in others absent. All elements showed variation in thethree compartments studied. Concentrations of all elements were similar in thecytoplasm and the nucleolus, but lower in the nucleus. The range of values forelemental concentrations was similar in the three compartments. High sodium levelswere not confined to any particular one of these; thus the low K/Na ratios found inthe analysis of whole sections were not due to the contribution of any one particularcompartment.
DISCUSSION
In the method of quantitation used here the concentration of an element present inthe area analysed is proportional to the area under the characteristic peak minus thebackground over the number of counts in the continuum area. It is recognized (Hall,1979; Hall & Gupta, 1979) that one of the problems in using this method is that ofensuring the reliable estimation of counts in the continuum in thin biological
Table 3. Elemental composition of compartments of Hela S3 cells
Compartment n P S Cl K Na
348 ±34 421± 45274 ±28 360 ± 47339 ±57 471 ±70
Values are means from stated numbers of observations and are expressed as mmol/kg dryweight ± s.E. of mean, n, number of observations.
CytoplasmNucleusNucleolus
303111
423 ± 32324 ± 27385 ± 50
185147171
±13±18±26
383 ± 42308 + 31449 ±51
226 A. Warley, J . Stephen, A. Hockaday and T. C. Appleton
looO I A
n
1000 - 900 --
a00 -- C) r, 700-- .- 0
600-- t
500-- (9 x
3 400 -- E E 300 --
200 --
loo --
P S CI K N a K I N ~
Fig. 6
X-ray microanalysis ofHeLa S3 cells 227
specimens. To overcome this problem we have used a wide energy band for thecollection of continuum counts coupled with a probe area wider than the minimumwidth of the instrument, to increase the number of counts collected. The results (Fig.4) show a linear relationship between peak/continuum and concentration, indicatingthat the method established in our laboratory can be used for quantitation of data frombiological specimens. Independent analysis by flame photometry of the sodiumpotassium tartrate standard further confirms this, allowing us to have confidence inour results (Table 1).
The results presented in this paper for the analysis of HeLa S3 cells grown insuspension show that individual cells have a variable K/Na ratio. This cannot beascribed to any particular aspect of quantitation as discussed above. Also, the K/Na ratio should be independent of the problems associated with estimation ofbackground, since the latter must be the same for both elements in each spotanalysed.
Low K/Na ratios in cells are usually associated with cell damage (James-Kracke etal. 1980), although erythrocytes in some species do show such ratios under naturalconditions (Williams, 1970). Cells in this study were 99 % viable before freezing, anddoubled in suspension over 30—36 h. The number of cells with low K/Na ratios istherefore not due to lack of viability in the cell population under study. Ionicredistribution could occur during the freezing procedure but the excellent morpholog-ical preservation seen in the sections suggests that damage did not occur at this step.Further evidence that cells were not damaged is obtained by comparing the resultsfrom randomly growing HeLa S3 cells with those from members of the same popula-tion treated with ouabain. The K/Na ratios exhibited by all the cells after treatmentwith this drug were lower than the values noted in the randomly growing cells, whichagain suggests that these latter cells were not damaged. These conclusions on cellintegrity and hence the validity of the analysis are enhanced by the analysis of humanerythrocytes. The latter were handled in exactly the same way as the HeLa cells andgave the expected K/Na ratio (7-5: 1).
James-Cracke et al. (1980), studying cryosections of cultured aortic muscle cells,noted variable concentrations of sodium in these cells and the range quoted (66—509mmol/kg dry weight) compares favourably with the range of values reported here. Itmay be that variable sodium levels are characteristic of cells in culture. Cameron &Smith (1980) reported that high levels of sodium and chlorine are characteristic ofrapidly growing or transformed cells. However, as their results are reported as meanand standard deviation rather than on an individual cell basis it is not clear from theirstudies whether low K/Na ratios are found in all cells or whether the ratio variesbetween individual cells. A study of rat vaginal epithelium by this latter group(Cameron, Pool & Smith, 1980) showed that K/Na ratios changed when cells of thistissue were stimulated to divide by administration of oestradiol. It is thus clear from
Fig. 6. Elemental composition of subcellular compartments of HeLa S3 cells. A.Cytoplasm; B nucleus; c, nucleolus. Except in the case of the nucleolus, each point is themean of at least three determinations.
228 A. Warley, J. Stephen, A. Hockaday and T. C. Appleton
these other studies of tissues by X-ray microanalysis that K/Na ratios may changeduring the cell-division cycle.
Norrie et al. (1982) reported a K/Na ratio of approximately 2: 1 for the same lineof HeLa cells as used in the work described here. This value was observed in cells afterthey were removed from monolayers in EDTA and held in suspension for a few hours— a handling regime comparable to that described in this paper. Their values wereobtained by flame photometry and are a measure of the average intracellular value of106 cells taken for each analysis. It is not easy to explain the difference between thevalues obtained by Norrie et al. (1982) and those reported here for K/Na ratios.There was, however, one potentially important point of difference between thepresent work and that reported by Norrie et al. (1982). During the execution of thelatter experiments difficulty was experienced in observing an increase in cell numbersin suspension (this problem arises in many laboratories, often without any obviousexplanation; it is not unique to us). In contrast, during the period over which thepresent work was done the HeLa cells did divide. As argued above, the evidenceobtained from standards, ouabain-treated cells and erythrocytes strongly suggests thatthe data obtained from individual cells must be meaningful and not artefactual. In ourview this is most likely to reflect the fact that cells in random culture are in differentstages of the cell cycle. We are currently testing this hypothesis by analysing cellcultures that have been deliberately synchronized. Preliminary data indicate thatK/Na ratios may well vary throughout stages in the cell cycle; detailed results will bepresented in a subsequent paper.
A. W. is in receipt of a M.R.C. Research Fellowship. The authors would like to thank LinkSystems Limited for the loan of the Quantem program and associated equipment.
REFERENCES
APPLETON, T. C. (1977). The use of ultrathin frozen sections for X-ray microanalysis of diffusibleelements. In Analytical and Quantitative Methods in Microscopy (ed. G. A. Meek & H. Y.Elder), pp. 247-268. Cambridge University Press.
APPLETON, T. C. (1978). The contribution of cryo-ultramicrotomy to X-ray microanalysis inbiology. In Electron Probe Microanalysis in Biology (ed. D. A. Erasmus), pp. 148—182. London:Chapman & Hall.
APPLETON, T. C. (1979). The localization of diffusible substances, experience in the electronmicroscope. J. Histochem. Cytochem. 27, 1518—1519.
APPLETON, T. C. & NEWELL, P. F. (1977). X-ray microanalysis of freeze-dried ultra thin frozensections of a regulating epithelium from the snail Otala. Nature, Land. 266, 854—855.
APPLETON, T. C , NEWELL, P. F. & MACHIN, J. (1979). Ionic gradients within mantle-collarepithelial cells of the land snail Otala lactea. Cell Tiss. Res. 199, 83-98.
CAMERON, I. L., POOL, T. B. & SMITH, N. K. R. (1980). Intracellular concentration of potassiumand other elements in vaginal epithelial cells stimulated by eatradiol administration. J. cell.Physiol. KM, 121-125.
CAMERON, I. L. & SMITH, N. K. R. (1980). Energy dispersive x-ray microanalysis of the concentra-tion of elements in relation to cell reproduction in normal and in cancer cells. In ScanningElectron Microscopy (part 2), pp. 463—471. Chicago: Scanning Electron Microscopy Inc.
CIVAN, M. M., HALL, T. A. & GUPTA, B. L. (1980). Microprobe study of toad urinary bladderin absence of serosal K+. J. membr. Biol. 55, 187-202.
GUPTA, B. L., HALL, T. A. & NAFTALIN, R. J. (1978). Microprobe measurement of sodium,
X-ray microanalysis of HeLa S3 cells 229
potassium and chloride concentration profiles in epithelial cells and intercellular spaces of rabbitileum. Nature, Lond. 272, 70-73.
HALL, T. A. (1979). Problems of the continuum-normalization method for the quantitative analysisof sections of soft tissue. In Microbeam Analysis in Biology (ed. C. P. Lechene&R. Warner), pp.185-208. London, New York: Academic Press.
HALL, T. A. &GUPTA, B. L. (1979). EDS quantitation and application to biology. In Introductionto Analytical Electron Microscopy (ed. J. J. Hren, J. I. Goldstein &D. C.Joy), pp. 169-197. NewYork: Plenum.
JAMES-KRACKE, M. R., SLOANE, B. F., SHUMAN, H., KARP, R. & SOMLYO, A. P. (1980).Electron probe analysis of cultured vascular smooth muscle, jf. cell. Physiol. 103, 313—322.
JONES, R. T., JOHNSON, R. T., GUPTA, B. L. & HALL, T. A. (1979). The quantitative measure-ment of electrolyte elements in nuclei of maturing erythrocytes of chick embryo using electron-probe X-ray microanalysis. J. Cell Sci. 35, 67-85.
KIRK, R. G., BRONNER, C., BARBA, W. & TOSTESON, D. C. (1978). Electron probe microanalysisof red blood cells. I. Method and evaluation. Am.J. Physiol. 235, C245-C250.
KIRK, G. R., LEE, P. & TOSTESON, D. C. (1978). Electron probe microanalysis of red blood cells.II. Cation changes during maturation. Am.J. Physiol. 235, C251-C255.
MORETON, R. B. (1981). Electron probe X-ray microanalysis: techniques and recent applicationsin biology. Biol. Rev. 56, 409-461.
NORRIE, D. H., WOLSTENHOLME, J., HOWCROFT, H. & STEPHEN, J. (1982). Vaccinia virus-induced changes in [Na+] and [K+] in HeLa cells. J . gen. Virol. 62, 127-136.
RICK, R., DORGE, A., MACKNIGHT, A. D. C , LEAF, A. & THURAU, K. (1978a). Electronmicroprobe analysis of the different epithelial cells of toad urinary bladder: electrolyte concentra-tions at different functional states of transepithelial sodium transport. J. membr. Biol. 39,257-272.
RICK, R., DORGE, A. & VON ARMIN, E. (19786). X-ray microanalysis of frog skin epithelium:evidence for a syncitial Na transport compartment. \n Microscopica Acta, suppl. 2. MicroprobeAnalysis in Biology and Medicine (ed. P. Echlin & R. Kaufmann), pp. 156-165. Stuttgart:Verlag.
RICK, R., DORGE, A., VON ARMIN, E. & THURAU, K. (1978C). Electron microprobe analysis of frogskin epithelium: evidence for a syncytial sodium transport compartment. J. membr. Biol. 39,313-331.
SAETERSDAL, T. S., MVKLEBUST, R., JUSTESEN, N. P. B. &ENGEDAL, H. (1977). Calcium contain-ing particles in mitochondria of heart muscle cells as shown by cryo-ultra microtomy and x-raymicroanalysis. Cell Tiss. Res. 182, 17-31.
SAUBERMANN, A. J., BEEUWKES, R. III . & PETERS, P. D. (1981). Application of scanning electronmicroscopy to x-ray analysis of frozen hydrated sections. II. Analysis of standard solutions andartificial electrolyte gradients. J. Cell Biol. 88, 268-273.
SOMLYO, A. V., GONZALEZ-SERRATOS, H., SHUMAN, H., MCCLELLAN, G. & SOMLYO, A. P.(1981). Calcium release and ionic changes in the sarcoplasmic reticulum of tetanised muscle: Anelectronprobe study. J. Cell Biol. 90, 577-594.
SOMLYO, A. P., SOMLYO, A. V. & SHUMAN, H. (1979). Electron probe analysis of vascular smoothmuscle. Composition of mitochondria, nuclei and cytoplasm. J. Cell Biol. 81, 316-335.
SOMLYO, A. V., SHUMAN, H. & SOMLYO, A. P. (1977). Elemental distribution in striated muscleand the effects of hypertonicity: electron probe analysis of cryosections.J. Cell Biol. 74, 828—854.
STATHAM, P. J. (1977). Deconvolution and background subtraction by least-squares fitting ofprefiltered spectra. Analyt. Chem. 49, 2149-2154.
WILLIAMS, R. J. P. (1970). The biochemistry of sodium, potassium magnesium and calcium.Chem. Soc. Q. Rev. 24, 331-365.
