Adsorption of Cationic (Basic) Dyes byFixed Yeast Cells'
CHARLES H. GILES AND ROBERT B. McKAY'Department of Pure and Applied Chemistry, The University of Strathclyde, Glasgow, Scotland
Received for publication 19 September 1964
ABSTRACTGILES, CHARLES H. (The University of Strathclyde, Glasgow, Scotland), AND ROB-
ERT B. McKAY. Adsorption of cationic (basic) dyes by fixed yeast cells. J. Bacteriol.89:390-397. 1965.-The adsorption of 10 typical cationic dyes on formalin-fixed yeastcells has been studied by determining isotherms, and the results are consistent with anion-exchange mechanism. The adsorption on this complex substrate is similar to thaton the simpler substrate, alumina. The dyes are probably aggregated when adsorbed,and the size of the aggregates increases with increase in the molecular weight of thedye ion. After considering the possible adsorption sites, and comparing the data withadsorption on simpler substrates, we suggest that the most important adsorptionsites may be phosphate or other strongly acidic groups.
The solution adsorption mechanisms of avariety of organic solutes on organic and inorganicsurfaces have been investigated here in recentyears by determining adsorption isotherms. (Ad-sorption from solution is quite distinct fromadsorption of gases or vapors. A solution is aphase of at least two components, whereas thegases or vapors used in adsorption tests areusually phases of one component.) An adsorptionisotherm is the plot at a given temperature of theweight of solute adsorbed on unit weight ofsubstrate (millimoles per kilogram) against theweight of solute remaining in solution whenequilibrium has been established (millimoles perliter).The substrates so far used have been essentially
homogeneous. Some are highly porous, e.g.,anodic alumina (Giles et al., 1959), carbons(Giles and Nakhwa, 1962), and natural fibrousmaterials (Chipalkatti, Giles, and Vallance,1954; Giles and Hassan, 1958; Giles et al., 1958),and others are nonporous, e.g., powdered alumina,silica, and graphite (see Giles and Nakhwa, 1962,and earlier papers in the present series). In allthese cases, the same general principles relatingthe nature of adsorption isotherm to the adsorp-tion mechanism have been found to apply.Usually the nonporous substrates reach equi-librium more rapidly than the porous ones, but,except in a very few cases, the nature of the
'Part 22 in the series "Studies in Adsorption."2 Present address: Research Laboratory, Kodak
Ltd., Harrow, England.
isotherm appears to be independent of the degreeof porosity of the surface.A system of classification was proposed (Giles
et al., 1960) whereby all solution adsorptionisotherms are divided into four main classes,according to the shape of the initial part of thecurve, and then further into several subgroups.This classification enables the adsorption mecha-nism to be identified, in many cases, from theisotherm shape. In this paper we shall be con-cerned with S, L ("Langmuir"), and H ("highaffinity") classes of isotherm (Fig. 1). For furtherdetails and references, see Giles et al. (1960).The aim of the present work, which is a quanti-
tative investigation of the adsorption of cationicdyes on formalin-fixed yeast cells, is to determinewhether the general principles established foressentially homogeneous substrates can be usedto interpret adsorption mechanisms on biologicalsubstrates. Adsorption of dyes is the basis of mosthistochemical techniques; cationic dyes, inparticular, are especially irnportant for stainingmicroorganisms.
Briefly, in an aqueous suspension of fixed yeastcells, each cell may be regarded as a highlyporous solid particle, ovoid in shape, and ofaverage diameter of about 4 u. The cell wall, acompletely permeable mucopeptide framework,on which are supported other constituents, suchas teichoic acids, is about 20%o of the dry cellweight (Trevelyan, 1958); the cell membrane, apolysaccharide-protein-lipid complex, is probably<1% of the cell weight and completely per-
meable in fixed cells; the cytoplasm and nucleus,rich in nucleoprotein, together comprise about80%NO of the dry cell weight. Clearly, the cytoplasmand nucleus will have the greatest effect on theadsorption properties, whereas the effect of thecell membrane will be almost negligible.
It has long been known that cationic dyes arestrongly adsorbed by the dead cells of micro-organisms, including yeast. Yeast is a gram-positive organism, and, since cationic dyes areused in the Gram staining test, their adsorptionproperties in this case have special interest. Thereis evidence to suggest adsorption depends on theliquid-solid ratio (Borzani and Vairo, 1958,1959; Vairo and Borzani, 1960; Finkelstein andBartholomew, 1960) and that the mechanism ofadsorption is ion exchange (McCalla, 1940, 1941a,b; Bartholomew, Roberts, and Evans, 1950;James, 1957).
MATERIALS AND METHODSPurification of dyes. The dyes used were Crystal
Violet (C.I. 42555), Ethyl Violet (C.I. 42600),Magenta P (C.I. 42510), Malachite Green (C.I.42000), Methylene Blue BP (C.I. 52015), Rhoda-mine B (C.I. 45170), Rhodamine 3B (C.I. 45175),Safranine (C.I. 50240), Victoria Blue BN (C.I.44045), and Victoria Pure Blue BO (C.I. 42595).Normal types of cationic dye were used, and allexcept Victoria Blue BN, Victoria Pure Blue BO,and Rhodamine 3B were purified by leaching dilu-ent-free "batch" grade samples with hot 10%HCl solution, filtering the hot liquors, and al-lowing the dyes to crystallize from them. Thecrystallization was repeated at least once, andthe crystals were dried first at 40 to 50 C. forseveral hours and then in a vacuum desiccatorover potassium hydroxide overnight (high-tem-perature drying was considered to be undesirable).Victoria Blue BN and Victoria Pure Blue BOcould not be purified satisfactorily, because oftheir low solubility in the acid solution; Rhoda-mine 3B formed an unfilterable gel with ethanol(probably the best alternative solvent; recrystal-lization from hydrochloric acid was considered tobe undesirable on account of possible hydrolysisof the-CO2C2H5 group); and untreated batchgrade (diluent-free) samples were therefore used.The purity of the dyes was determined by com-bustion microanalysis for C, H, and N; by flameanalysis (flame spectrophotometer; Evans Elec-troselenium Ltd.) for Na (a rough measure ofsalt impurity), and, in the case of Victoria PureBlue BO and Victoria Blue BN, by potentiometrictitration of chloride ion with silver nitrate inneutral solution. All the dyes had only traceamounts of Na; Rhodamine B had no detectableamount (confirmed by ashing; 0.04% ash). Theelementary analyses suggested that the impuritypresent in the recrystallized samples was mainlywater, with a little HCI in some cases.
F- czw
' cr.~Iin4(
EQUILIBRIUM SOLUTIONCONCENTRATION
FIG. 1. Three typical solution adsorption iso-therms, classified (Giles et al., 1960) as S.2, L.2,and H.2.
Preparation of dye solutions. Distilled or de-mineralized water was used in all experiments.All the dyes were dissolved cold, except VictoriaBlue BN and Victoria Pure Blue BO, which weredissolved with careful warming. In all cases, thesolution concentrations were corrected for purityof the samples. All dye solutions were preparedimmediately before use, and exposure to light wasminimized.
Pretreatment of glassware for dye solutions.Basic dyes are strongly adsorbed by glass. Tominimize this effect, all glassware to be used incontact with dye solutions was steeped beforeuse for at least 1 hr (usually overnight) in asolution (1 to 2 g per liter) of a cationic surface-active agent, cetyl trimethylammonium bromide,which is preferentially adsorbed. The glass sur-faces were then thoroughly rinsed with waterbefore use.Analysis of dye solutions. The dye solutions were
analyzed at the long-wavelength absorption peakon a Unicam SP 600 spectrophotometer before andafter the adsorption tests.
Tests of stability of basic dyes at elevated tempera-tures. Some basic dyes in solution are susceptibleto decomposition on standing at elevated tempera-tures (greater than ca. 50 C.). It was found, byobserving the optical densities of the presentdye solutions before and after rotation in sealedglass tubes at 50 C (highest temperature used inthe adsorption tests) in a thermostatically con-trolled water bath, that no significant changeoccurred over a period of 2 hr. It was concludedthat no significant decomposition had occurred.
Substrate. The substrate used in all the adsorp-tion experiments was brewer's yeast, a strain ofSaccharomyces cerevisiae, obtained from the Dis-tillers Co., Ltd. The yeast was fixed and storedin a 4% formaldehyde solution, such that theyeast to liquid ratio was ca. 1:4 by weight.Preliminary tests showed that the adsorptioncharacteristics of basic dyes on this substratestored for 18 months were the same as on freshlyfixed yeast.
For all the experiments, a small sample of yeastwas removed from storage, washed three times,and suspended in distilled or demineralized water.
suspension (5 ml) was pipetted into a silica cruci-ble previously dried to constant weight. The sam-ple was carefully evaporated to dryness at 120 C.The dry weight of yeast, and hence the concen-tration of the suspension, was then found. Theexperiment was performed in triplicate. A series ofdilute suspensions was prepared from the original,and their optical densities were measured at a suit-able wavelength (5,000 A, arbitrarily chosen). Acalibration graph of optical density against yeastsuspension concentration (dry weight basis) wasthen drawn for use in subsequent experiments.
Adsorption procedure. For determining adsorp-tion isotherms, a series of standard dye solutionswas prepared and 5-ml samples of each were mixedwith 5-ml samples of a standard yeast suispensionin oven-dried test tubes. The tubes were sealedin a Bunsen flame and rotated mechanically at
TABLE 1. Preliminary isotherm determinationswith yeast cells subjected to variouts treatments,and at various suspension concentrations
Dye
CrystalViolet
Methyl-ene BlueBP
Safranine
Rhoda-mine B
V'ictoriaPureBlueBO
Yeasttreat-menta
FlFlFlFlU
U
U
BB
F2FlF2FlFl
FlBB
Fl
U
U
Yeastsuspen-sionconcn
g/liter
0.1100.4400.3840.0830.1600.2790.1580.2410.343
0.3120.1840.2220.2080.213
0.3240.3910.475
0.151
0.1750.208
SolTem- Iso- Maximal
ventb pera- therm adsorp-tue typec tion
w
w
w
B9w
w
w
w
W
w
w
w
w
B9
w
w
w
w
w
w
C
223950201916402050
1721494323
181749
14
1940
L2L2L2LlL2L2L2L2L2
L2L2L2L2L2
L2L2L2
Si
H2H2
mmoleslliter
420
283335400d
370288364230230
168308272312534
194133171
70d
478483
a The following symbols are used: Fl, formalin-fixed yeast cells (sample 1); F2, formalin-fixedyeast cells (sample 2); U, untreated yeast cells;B, boiled (20 min) yeast cells.
b W, water; B9, buffer solution at pH 9.See Giles et al., 1960.
d Maximal adsorption not reached.
ca. 35 rev/min in a thermostatically controlledwater bath for 2 hr. The tubes were then removed,broken, and the contents were centrifuged. Thesupernatant solution was decanted off for analysis.
Preliminary experiments (Table 1). The adsorp-tion of the basic dyes used was found to be 80 to90% complete in 2 to 3 min; the remaining adsorp-tion in some cases required up to 1 hr (rate curvesgiven by McKay, 1963). In all the isotherm de-terminations, the solutions were therefore agitatedfor at least 2 hr.
Isotherms for several of the dyes used weredetermined (i) over a wide range of solid-liquidratios, i.e., yeast suspension concentrations,unbuffered; (ii) at several temperatures, unbuff-ered; and (iii) in systems buffered at pH 9. Al-though the amount of dye adsorbed was dependenton the yeast-liquid ratio (being higher at lowvalues), on the temperature, and on the pH ofthe test solution, the shapes of the isotherms were,in all cases, found to be independent of thesevariables.
Further, a few control tests were performed byuse of unfixed yeast cells (the live proportion ofthe cells being killed during the experiment by thedyes) and on yeast cells fixed by boiling a con-centrated aqueous suspension for 20 min. Theshapes of the isotherms obtained for the adsorp-tion of the dyes on these substrates were essen-tially the same as those on the formalin-fixedcells.The above experiments show that the type of
isotherm obtained for the adsorption of a basicdye on yeast cells is a real characteristic of theyeast and is unaffected by the conditions of theexperiment and by the pretreatment givein tothe yeast.pH values of adsorption solutions. The pH values
of the equilibrium solutions corresponding to thebeginning of the isotherm plateau lay within thenarrow range 4.42 to 5.10, with one exception:Rhodamine B, pH 3.63. The concentration ofthis dye was higher than that of the others (Fig.2). It was therefore not necessary to buffer thesolutions. Buffering would be undesirable becausebuffer salts would tend to alter the aggregationproperties of the dyes in solution, and so perhapsinterfere with the adsorption. Moreover, in bio-logical staining, pH control is not normally used.Some preliminary tests (Table 1) showed that,when the pH is raised as high as 9.0, adsorption isincreased, as expected for adsorption at pre-dominantly ainionic sites.
Selected conditions for the main series of experi-ment.r. The main series of experiments, the resultsof which are discussed below, were performedunder the following conditions. The solid-liquidratio was maintained at the same value for allthe isothernm determinations on yeast. This wasachieved by using yeast suspensions of concen-tration 0.246 ± 0.006 g per liter in all the testmixtures. The tests at room temperature werecontrolled at 20 + 0.5 C; those at elevated tem-peratures were controlled at 50 i 1 C. As stated
previously, all the tests on yeast were performedwithin the pH limits of 4.42 to 5.10 (except withRhodamine B).
Reproducibility of results. The isotherms de-termined at room temperature by use of suspen-
sions prepared from any one formalin-fixed yeastsample could be reproduced to within ca. +3%accuracy (estimated at the plateau). The marginof error in the isotherms at 50 C was found to begreater than this, probably owing to a fall intemperature between the removal of the tubesfrom the thermostat bath and the end of thecentrifuging step. There are two main types oferror: (i) those affecting the whole isotherm, i.e.,making all points too high or too low, and (ii)those affecting isolated points. The formner in-cludes errors in the suspension analysis. whichinvolves reading from an experimentally deter-mined graph; the latter includes errors in pipettingsamples of yeast suspensions and errors due toincomplete separation in centrifuging. Bad centri-fuging, however, was found to produce veryserious and obvious errors which could be elimi-nated by careful manipulation.The adsorption properties of formalin-fixed
samples of yeast remained constant over longperiods of time.
Adsorption experiments on other substrates. Afew adsorption tests were made on other sub-strates, i.e., silk and wool [cleaned and scouredfabrics; wool, specific surface area, 56.6 m2/g,by p-nitrophenol from water (Giles and Nakhwa,1962); silk, not determined], graphite (AchesonColloids Ltd., 0.28% ash, specific surface area bynitrogen adsorption, 125 m2/g), chromatographicalumina (May and Baker, specific surface area,4.8 mn2/g), and deoxyribonucleic acid (DNA).In the latter case, 5-ml portions of a 4% aqueoussolution of the sodium salt of DNA (L. Light andCo.) were acidified with 0.1 ml of 50% (v/v)hydrochloric acid. The free DNA was thus pre-
cipitated out; the final pH of the suspensionwas 2.5. To each 5-mI portion thus treated, a
5-ml portion of standard dye solution was added,and the mixtures were rotated mechanically for24 hr in the thermostatic bath at 20 C. The super-
natant liquor was separated by centrifuging andanalyzed spectrophotometrically.
RESULTS AND DISCUSSION
The results of the main series of experiments are
given in Table 2 and Fig. 2. The extremely highrate of adsorption is in accordance with the ion-exchange mechanism suggested by other authors(McCalla, 1940, 1941a, b; Bartholomew, Roberts,and Evans, 1950; James, 1957). Also, it is wellknown that yeast cells are stained throughout bycationic dyes, and the present results are con-
sistent with this fact.Adsorption isotherms on yeast (Fig. 2). Crystal
Violet, Magenta 1', Malachite Green, MethyleneBlue BP, Rhodamine 3B, and Safranine giveisotherms of type L (Giles et al., 1960) and thusshow "normal" Langmuir adsorption (see Borzaniand Vairo, 1958, 1959; Vairo and Borzani, 1960;James, 1957). Victoria Pure Blue BO, VictoriaBlue BN, and Ethyl Violet give isotherms of typeH. These three dyes are completely removed fromdilute solutions by the yeast and clearly have a
high affinity for it. Rhodamine B on yeast givesan isotherm of type S (Fig. 2H); this anomalousbehavior is discussed below.
Adsorption of the rhodamines on yeast. Theadsorption of Rhodamir.e B (Fig. 2H) andRhodamine 313 (Fig. 2G) is similar to theiradsorption on chromatographic alumina, wherethey were also adsorbed by ion exchange andgave S- and L-type isotherms (Fig. 4E), respec-
TABLE 2. Adsorption data for cationic dyes on yeast
Isotherm Probable Maximal Projected Calculated Coverage CationicDve type* orientation equilibrium area of dye specific surface factort wtadsorption cationt area
mmole/kg A2 n121gCrystal Violet L2 Aggregated 264 224 356 5.41 372Ethyl Violet H2 Aggregated 316 266 509 7.75 456Magenta P L2 Aggregated 163 168 165 2.51 302Malachite Green L2 Flat 61 181 66.5 1.01 329Methylene Blue BP L2 Aggregated 153 120 111 1.69 284Rhodamine B 82 Edg.e-on 88 124 65.7 1.00Rhodamine 3B L2 Aggregated ca. 130 184 144 2.19 471Safranine L2 Aggregated 176 148 157 2.39 315Victoria Blue BN H2 Aggregated 300 251 453 6.89 470Victoria Pure Blue BO H2 Aggregated 368 268 595 9.06 478
* See Giles et al. (1960).t Area of smallest enclosing rectangle (in most probable orientation).t Ratio of maximal amount of dye adsorbed to calculated monolayer capacity (for stated probable
FIG. 2. Equilibrium adsorption isotcationic dyes on formalin-fixed yeastconcentration of 0.245 i 0.006 g (dry tliter in the test mixture. Symbols: X an
0.5 C; 0, 50 4- 1 C. (A) Crystal Violet, (IPure Blue BO, (C) Methylene Blue BP,nine, (E) Magenta P, (F) MalachiteRhodamine SB, (H) Rhodamine B, (I) El(J) Victoria Blue BN.
tively (Giles, Easton, and McKay, 196mine B was adsorbed on alumina as a clmonolayer of molecules oriented verticsurface; indeed, this orientation and paalways been found when S-type isothbeen obtained (Giles et al., 1960). Rhodhowever, behaved differently. Reasoidifferences in behavior have been discusEaston, and McKay, 1964). Briefly, thare identical in structure except that IB has a free carboxyl group, whereas
(C2H5)2N 0 N(2H5).
C
v- COOH
mine 3B this group is fully ethylated. IB is therefore to some degree dipolicharacter, although overall it is predcationic, as would be expected and hasfirmed by electrophoresis on paper in a
buffer at pH 7.1. Nevertheless, the carbwill be repelled to some extent by thsubstrate, thereby reducing the affindye for the substrate and causing the r
approach with a vertical orientation. At very low400 concentrations, the dye is scarcely adsorbed, but,
as the concentration is increased, the dye isforced to the surface of the alumira by theinternal pressure of the solution, and the adsorp-tion is aided by interaction between adjacentvertically oriented adsorbed molecules. The
200 fully esterified Rhodamine 3B, which is notdipolar ionic in character, has a higher affinity forthe substrate and the adsorption follows the
° normal L-type isotherm. The fact that these two50 dyes, indeed, also the other dyes, behave the same
way on yeast as on alumina suggests that theK complex structure of yeast and the dipolar ionic
200 character of many of the constituent macro-molecules have little effect on the adsorption
0I5 0.2 mechanism, and that the isotherms on yeast maybe interpreted in the same way as has been de-
therms for scribed for the isotherms on alumina. The follow-cells at a ing section lends further support to this view.weight) per Aggregation of cationic dyes adsorbed by iond 0, 20 i exchange. It has previously been shown thatB) Victoria cationic dyes are aggregated when adsorbed by(D) Safra- ion exchange on chromatographic alumina (Giles,Green, (G) Easton, and McKay, 1964). An estimate of thethyl Violet, size of the aggregates of each dye was made by
calculating the coverage factor: the ratio of thespecific surface area of the substrate estimated
4). Rhoda- from the dye adsorption (the product of thecse-packed amount of dye adsorbed at the isotherm plateau,ally to the the projected area of the dye molecule at thecking have surface, and the Avogadro number) to the trueierms have specific surface area (estimated by nitrogenlamine 3B, adsorption and by the adsorption of well-char-ns for the acterized solutes which give S isotherms). Forssed (Giles, simplicity, the dye molecules were assumed to bee two dyes in their most probable orientation, i.e., lying flatIhodamine on the surface, and the projected area of eachin Rhoda- molecule was estimated from Catalin molecular- scale models. It was shown that the coverage
2factor was a linear function of the logarithm ofthe molecular weight of the dye cation. [Thisrelationship has been dealt with more fully inrecent work (Easton, Giles, and McKay, 1964).Coverage factors of monovalent dyes (both ani-onic and cationic) on a variety of simple sub-strates vary linearly with the logarithm of themolecular weight of the dye ion. Further, all the
Ihodamine plots can be superimposed to give a line of high3r ionic in statistical significance. The present results withlominantly yeast conform to this general relationship.] Ins been con- other words, the size of the adsorbed aggregatesphosphate increased with the size of the dye cation.ooxyl group With yeast, although the concept of a specifice basophil surface area is difficult to comprehend fully, it istity of the significant that a similar relationship can benolecule to demonstrated (Table 2, Fig. 3). The "specific
surfacestimewhichalumi]of verand Aisothedata fmay rare alpossibhas be
IO dye cations are clearly subjected to many types ofattractive forces. Nevertheless, the present results
O and those of other authors (McCalla, 1940, 1941ab; Bartholomew, Roberts, and Evans, 1950;
7s5 - James, 1957) suggest that the adsorption mecha-O nism is predominantly ionic in character. This
implies that the most probable adsorption sitesare carboxyl groups in polypeptide side chains
° phosphoric acid residues in nucleic acids (and5.0/ perhaps teichoic acids), and perhaps even sul-
furyl residues in polysaccharides. It seemsreasonable to suppose that the powerful attrac-tion of the strongly acidic phosphoric acid residues
2-5 O will predominate over that of the weakly acidiccarboxyl groups.
There is convincing experimental evidence fromhistochemical studies of sectioned animal cells
0 I (Baker, 1962) that the nucleic acids in the cyto-200 300 400 500 plasm (RNA) and especially the nuclei (DNA)are strongly stained by basic dyes, much more
CATIONIC WEIGHT so than the other cellular parts when dilute dye3. Relation between "coverage factor" and solutions are used (with concentrated solutions,
ic weight of dyes adsorbed on fixed yeast cells. other substances such as acidic proteins and acidicmucopolysaccharides are often strongly stained
e area" of the yeast (65.7 m2/g, Table 2) is as well). If this is so in yeast cells (it should beLted from the adsorption of Rhodamine B, remembered that in yeast cells, the nucleic acid-gives an S-type isotherm on yeast, as on rich nucleus and cytoplasm together comprise
na, where it forms a close-packed monolayer about 80% of the dry cell weight), then the effectrtically oriented molecules (Giles, Easton, of the phosphoric acid residues will have thedcKay, 1964), as do all solutes giving an greatest influence on the overall adsorptionrm of this type (Giles et al., 1960). The characteristics and thus on the shapes of thefor the other dyes are given in Table 2. It adsorption isotherms. To give experimentaleasonably be concluded that cationic dyes support to this view, a study of the adsorption ofggregated when adsorbed on yeast. The Rhodamine B on DNA and various other selected)le effect of aggregation on staining processes substrates was made (Fig. 4).een discussed in a previous paper which is Adsorption of Rhodamine B on other substrates.
intended for physical chemists (McKay, 1963).Malachite Green and Rhodamine 3B are
exceptional in that they give much lower coveragefactors than would be expected from the molecu-lar weights of their cations. They are not includedin Fig. 3. They are also exceptional, however, inthat the ionic charge distribution (Lewis et al.,1943) is highly unsymmetrical about an axisperpendicular to the plane of the molecule, andit may be that this reduces their ability to formaggregates. It remains possible, though, that theyare involved in some specific interaction withyeast.The dyes which have the largest coverage
factors, i.e., Ethyl Violet, Victoria Pure Blue BO,and Victoria Blue BN, give isotherms of type H2which have been observed in several other systemswhere large ionic micelles are adsorbed (Giles etal., 1960).
Nature of the adsorption sites. On a substrate ofsuch complex chemical constitution as yeast, the
EOIJIRIUM SOLUTION CONCENTRATION (MMOLE/A)
FIG. 4. Equilibrium adsorption isotherms forRhodamine B on various substrates (at room tem-perature). (A) Silk, (B) wool (unacidified), (C)wool (initial bath pH 2.5), (D) graphite, (E)chromatographic alumina, (F) deoxyribonucleicacid (initial bath pH 2.5; weight of substrate refersto that of initial Na salt).
On wool and silk, isotherms of type L2 are ob-tained (Fig. 4A, B, C), and, on graphite, one oftype H2. The amount of adsorption on graphitecorresponds very closely to a monolayer of dyecations oriented parallel to the surface (area of thedye molecule flat is ca. 230 A2, as estimated fromthe model). On chromatographic alumina, how-ever, an S2-type isotherm (Fig. 4E) is obtained.This indicates a vertical orientation of the dyecations (cross-sectional area, 124 A2), and theamount of adsorption does, in fact, correspond toa close-packed monolayer so oriented.On substrates with a high proportion of hydro-
carbon residues, e.g., protein fibers, or with con-densed aromatic systems, e.g., graphite, adsorp-tion is due largely to van der Waal's attractionbetween the mainly hydrophobic surface and thearomatic ring system of the dye ions (Chipalkatti,Giles, and Vallance, 1954; Vickerstaff, 1954).This mechanism favors flatwise orientation(Fig. 5), and so accounts for the L-type isotherm.On chromatographic alumina, which has a nega-tively charged ionic surface and no hydrocarbonresidues or aromatic systems, ion-ion attractionis predominant, and the vertical orientation isfavored, as discussed above and by Giles, Easton,and McKay (1964).On the above argument it is indeed unlikely
that the principal adsorption sites in yeast cellsare located in the protein constituents; they aremore likely to be in the strongly acidic nucleicacids, or in teichoic acids. To check this view, theisotherm for Rhodamine B on solid DNA wasdetermined. This isotherm (Fig. 4F) is, in fact, ofthe S-type, like that on yeast, and quite unlikethat obtained with the protein fibers. (Isothermsfor wool were determined by use of both un-acidified dye solutions, for comparison with theyeast experiment, and solutions acidified to pH2.5 for comparison with the DNA experiment.There is no difference in isotherm type, though;as expected, adsorption is less from the acid solu-tion, because of the increased positive charge onthe fibers. Note that the scales on the X and Yaxes of the isotherm on DNA are both IOOX
RZN 1C4pH
FIG. 5. Suggested orientation of Rhodamine Bcations on graphite or protein surfaces. Here thesurface attracts preferentially the aromatic nucleiof the dye cations, and thus there is flatwise orienta-tion.
less than those of the isotherm on yeast; hence,the isotherms are indeed truly comparable.)
There is evidence that RNA is less permeableto cationic dyes than DNA (Goldstein, 1961).With a mixture of two cationic dyes, one of lowand the other of high molecular weight, DNA isstained by both dyes, but RNA is preferentiallystained by the dye of low cationic weight. In thepresent experiments with yeast, however, thedyes are applied separately and competition forthe substrate does not occur. There is thereforeno reason to suppose that the lower permeabilityof RNA will prevent its being stained here.Indeed, there is evidence that the very largeanionic dye, MIethyl Blue (anionic weight, 745),which stains chromatin strongly (Baker, 1962),reacts with both the DNA and RNA constituents(by a nonionic mechanism) in preference to theprotein constituent ('McKay, 1962). It is clearthat this form of RNA at least is permeable tomuch larger molecules than those used in thepresent study.
Effect of temperature. Adsorption is essentiallyan exothermic process and hence should befavored by low temperatures. If, however, theheat of adsorption is low (e.g., in ion exchange orvan der Waal's adsorption), then the effect oftemperature is less marked. For example, it hasbeen shown that the ion-exchange adsorption ofanionic dyes by acidified alumina (Cummings etal., 1959) and by proteins (Vickerstaff, 1954;Klotz and Urquart, 1949) is virtually unaffectedby substantial changes in the ambient tempera-ture.
Isothermis for several of the dyes used havebeen determined at two temperatures, i.e., 20and 50 C (Fig. 3). In each case, the adsorption isanomalous in being greater at high than at lowtemperature. This type of anomalous adsorptionhas been observed with some inorganic substratesand appears to be due to aggregation of the dyesin solution (Giles, Nakhwa, and Greezek, 1961;Giles, Easton, and McKay, 1964).
It is well known that high temperatures favorthe disaggregation of dye molecules in solution(Vickerstaff, 1954), and it has been suggested(Giles, Nakhwa, and Greezek, 1961) that the dyemolecules are reaggregated on adsorption. In thepresent case, in which the adsorption is extremelyrapid and occurs throughout the highly poroussubstrate, the monodisperse dye cations willtend to be preferentially withdrawn from solution,since their smaller size will enable them to passmore easily and more quickly than micellesthrough the pores. The action of high temperaturewill accentuate this effect by greatly increasingthe proportion of the monodisperse form in
solution. Nevertheless, the high coverage factorsshow that at equilibrium the species present atthe adsorption sites is a micellar form. It istherefore suggested that aggregation of themonodisperse dye cations to form ionic micellesoccurs at the moment of close approach to theadsorption sites. Such aggregation will be favoredby the increased dye concentration near the sites,and perhaps by loss of solvated water.
ACESNOWLEDGMENTSWe thank P. D. Ritchie for his interest; Allied
Colloids (Bradford) Ltd., Badische Anilin-undSoda-Fabrik A. G., L. B. Holliday and Co. Ltd.,and Imperial Chemical Industries Ltd., DyestuffsDivision, for the gift of dyes; The Distillers Co.Ltd., for gifts of yeast; the Department of Scien-tific and Industrial Research for a Scholarship(to R. B. McK.); A. C. Syme for analyzing thepurified dye samples; and I. A. Easton and A. H.Tolia for some of the tests illustrated in Fig. 4.
LITERATURE CITEDBAKER, J. R. 1962. Experiments on the action ofmordants. 2. Aluminium-haematein. Quart. J.Microscop. Sci. 103 :493-517.
BARTHOLOMEW, J. W., M. A. ROBERTS, AND E. E.EVANS. 1950. Dye exchange in bacterial cells,and the theory of staining. Stain Technol.25:181-186.
BORZANI, W., AND M. L. R. VAIRO. 1958. Quanti-tative adsorption of methylene blue by deadyeast cells. J. Bacteriol. 76: 251-255.
BORZANI, W., AND M. L. R. VAIRO. 1959. Quanti-tative adsorption of Crystal Violet by deadnmicroorganisms. J. Bacteriol. 77:757-759.
CHIPALKATTI, H. R., C. H. GILES, AND D. G. M.I. Adsorption of organic compounds by poly-amide and protein fibres from aqueous and non-aqueous solutions. J. Chem. Soc., p. 4375-4390.
CUMMINGS, T., H. C. GARVEN, C. H. GILES,S. M. K. RAHMAN, J. G. SNEDDON, AND C. E.STEWART. 1959. Adsorption at inorganic sur-faces. IV. Mechanism of adsorption of organicsolutes by chromatographic alumina. J. Chem.Soc., p. 535-544.
EASTON, I. A., C. H. GILES, AND R. B. MCKAY.1964. Association of adsorbed aromatic ions.Chem. Ind. (London), p. 1863-1864.
FINKELSTEIN, H., AND J. W. BARTHOLOMEW. 1960.Dye uptake by gram-positive and gram-nega-tive cells as related to adsorption laws. J. Bac-teriol. 80:14-17.
GILES, C. H., I. A.. EASTON, AND R. B. MCKAY.1964. Mechanism of adsorption of cationic dyesby alumina, and a note on heat changes insolution adsorption. J. Chem. Soc., p. 4495-4503.
GILES, C. H., AND A. S. A. HASSAN. 1958. Adsorp-tion at organic surfaces. V. A study of theadsorption of dyes and other organic solutes by
cellulose and chitin. J. Soc. Dyers Colourists74:846-857.
GILES, C. H., A. S. A. HASSAN, M. LAIDLAW,AND R. V. R. SUBRAMANIAN. 1958. Adsorptionat organic surfaces. III. Some observations onthe constitution of chitin and on its adsorptionof inorganic and organic acids from aqueoussolution. J. Soc. Dyers Colourists 74:647-654.
GILES, C. H., T. H. MACEWAN, S. N. NA1KHWA,AND D. SMITH. 1960. Studies in adsorption. XI.A system of classification of solution adsorptionisotherms, etc. J. Chem. Soc., p. 3973-3993.
GILES, C. H., H. V. MEHTA, S. M. K. RAHMAN,AND C. E. STEWART. 1959. Adsorption at inor-ganic surfaces. V. Adsorption of sulphonateddyes by the anodic film on aluminium. J. Appl.Chem. London 9:457-466.
GILES, C. H., AND S. N. NAKHWA. 1962. Studiesin adsorption. XVI. The measurement of specificsurface areas of finely divided solids by solutionadsorption. J. Appl. Chem. 12:266-273.
GmEs, C. H., J. J. GRECZEK, AND S. N. NAKHWA.1961. Studies in adsorption. XIII. Anomalous(endothermic) effects of adsorption on inorganicsolids. J. Chem. Soc., p. 93-95.
GOLDSTEIN, D. J. 1961. Mechanism of differentialstaining of nucleic acids. Nature 191:407-408.
JAMES, A. M. 1957. The identification of surfacecomponents on the bacterial cell wall. Proc.Intern. Congr. Surface Activity, 2nd, London.4:254-261.
KLOTZ, I. M., AND J. M. URQUART. 1949. The bind-ing of organic ions by proteins. Effect of tem-perature. J. Amer. Chem. Soc. 71:847-851.
LEWIs, G. N., 0. GOLDSCHMID, T. T. MAGEL, ANDJ. BIGELEISEN. 1943. Dimeric and other formsof Methylene Blue: absorption and fluorescenceof pure monomer. J. Amer. Chem. Soc. 65:1150-1154.
MCCALLA, T. M. 1940. Cation adsorption by bac-teria. J. Bacteriol. 40:23-32.
MCCALLA, T. M. 1941a. The adsorption of hy-drogen ions by bacteria as measured by theglass electrode. J. Bacteriol. 41:775-783.
MCCALLA, T. M. 1941b. The reaction of certainstains with bacteria. Stain Technol. 16:27-32.
McKAY, R. B. 1962. An investigation of theanomalous staining of chromatin by the aciddyes, methyl blue arid aniline blue. Quart. J.Microseop. Sci. 103:519-530.
McKAY, R. B. 1963. The dye-staining of micro-organisms: a review. Textile Res. J. 33:381-391.
TREVELYAN, W. E. 1958. Synthesis and degrada-tion of cellular carbohydrates by yeast. InA. H. Cook [ed.], The chemistry and biology ofyeasts. Academic Press, Inc., New York.
VAIRO, M. L. R., AND W. BORZANI. 1960. Quanti-tative study of Crystal Violet adsorption bydead gram-negative bacteria. J. Bacteriol.80:572.
VICKERSTAFF, T. 1954. The physical chemistryof dyeing, 2nd ed. Oliver and Boyd, Edinburgh.
