An Assessment of Infrared Spectra As Indicators of Fungal Cell Wall Composition

AN ASSESSMENT OF INFRARED SPECTRA AS INDICATORS OF FUNGAL CELL WALL COMPOSITION

By A. J. MICHELL* and G. SOURFIELD*

[Manuscript received August 22, 1969J

Summary Infrared spectroscopy is assessed as a technique for identifying polymers derived from fungal cell walls, both as isolated materials and in mixtures with one another. The technique is then applied to a study of the composition of fungal cell walls and the conclusion reached that infrared spectra provide a rapid and valuable indication of the major components of such walls. They can also be used to follow the effect of chemical treatments designed to separate major wall components.

I. INTRODUOTION Methods for studying the chemical composition of fungal cell walls have been summarized by Aronson (1965) and the structural chemistry offungal polysaccharides has been reviewed by Gorin and Spencer (1968b). Recent methods have included analysis of the products of enzymic and chemical hydrolysis of cell walls isolated either by mechanical means or chemical treatment (Bartnicki-Garcia 1966; Aronson, Cooper, and Fuller 1967) and identification of cell wall components by means of X-ray diffraction (Kreger 1967) and nuclear magnetic resonance spectroscopy (Gorin and Spencer 1968a). Infrared spectroscopy has been used to demonstrate the presence of cellulose and chitin in the extracted walls of a number of fungi (Michell and Scurfield 1967) and as "corroborative evidence for glucan heterogeneity in oomycete walls" (Aronson, Cooper, and Fuller 1967). This paper is concerned with the extent to which infrared spectroscopy can be used to identify other fungal cell wall polymers. The results are relevant to the recent conclusion of Bartnicki-Garcia (1968) that "Powder X-ray diffraction continues to be the most reliable single technique for identifying cell wall polymers ... Infrared spectroscopy may be used as an auxiliary technique for distinguishing chitinous walls from non-chitinous ones."

II. MATERIALS AND METHODS Infrared spectra of the following reference substances were determined:

(i) Mannans These were homopolymers derived from the cell walls of the yeasts Saccharomyces cerevisiae (Peat, Whelan, and Edwards 1961), S. carlsbergensis (Behrens and Cabib 1968), Candida albicans (Yu et al. 1967), and Trichosporon aculeatum (Gorin, Mazurek, and Spencer 1968). An extracellular mann an produced by the yeast Rhodotorula glutinis (Gorin, Horitsu, and Spencer 1965) was also examined.

(ii) Glucans (I)

* Division of Forest Products, CSIRO, P.O. Box 310, South Melbourne, Vic. 3205.

Aust. J. bioI. Sci., 1970, 23, 345-60 346 A. J. MICHELL AND G. SCURFIELD

(2) f3-Glucans.-The nature and orrgm of most of these are listed by Clarke and Stone (1963). They included crown gall and Mycoplasma glucans, paramylon, cellulose I from wood, amorphous cellulose produced by fine grinding of cellulose I, pustulan, laminarans from Laminaria digitata, L. cloustoni, and Eisenia bicyclis, oat and barley glucans, lichenin, lutean, and glucans derived from Saccharomyces cerevisiae cell walls, sclerotia of Poria cocos, Polyporus myllitae, and Lentinus tuber-regium and the culture medium of Claviceps fusiformis (Buck et al. 1968).

(iii)

cellulose I + mannan (1 : 1) glycogen (1 : 1)

The glucan and mannan referred to in the above list were derived from Saccharomyces cerevisiae cell walls (Peat, Whelan,. and Edwards 1958, 1961). Spectra of the following compounds were examined for evidence of the effect of chemical linkage between two of the components: a galactomannan derived from Trichosporon fermentans (Gorin and Spencer 1968a), two glycoprotein fractions (2A and 2B) from the cell walls of Pithomyces chartarum (Sturgeon 1964), and a chitin-protein complex from the scraped larval cuticle of Oalliphora sp. (Rudall 1967). They were compared with the spectra of 1 : 1 mixtures of the two components in each case, and of the components taken singly. The spectra of various residues derived from fungi were examined in the light of the results for the reference materials with a view to obtaining some indication of their chemical composition. The residues were derived from Oeratocystis ulmi, Phytophthora cinnamomi, and species of Pythium and Epicoccum, following extraction of their mycelia successively with KOH, glacial acetic acid-hydrogen peroxide, and H 2S04 . The spectra of the residues derived from the last three species had been examined by us on a previous occasion (Michell and Scurfield 1967). The spectra of water-washed cells of the yeasts Saccharomyces cerevisiae, Pichia membranaeformis, Schizo saccharomyces malidevorans, Oandida mycoderma, and Rhodotorula rubra were also examined. The water-washed cells of Saccharomyces cerevisiae were then subjected to the extraction pro cedures outlined by Aronson and Preston (1960a, 1960b) and Bell and N orthcote (1950). The spectra of residues obtained after each stage of these procedures were examined for evidence of the efficacy of these treatments. A sample of cell wall material isolated mechanically (Phaff, personal communication) served as a standard. Materials, other than those derived from the cell walls of Phytophthora, Pythium, and Epicoccum (see above), were enclosed in potassium bromide disks. The spectra (4000-250 cm-1 ) were determined using a Perkin-Ehner model 457 spectrophotometer.

III. RESULTS AND DISCUSSION (a) Mannans The spectra of mannans derived from yeasts are shown in Figure 1. Mannans from Saccharomyces cerevisiae and S. carlsbergensis gave identical spectra, a result in accord with their structural similarity. IDENTIFICATION OF FUNGAL CELL WALL POLYMERS 347

Of the features common to all the yeast spectra the band in the region 812-805 cm-1 is the most interesting. Barker, Bourne, and Whiffen (1956) claimed that it was characteristic of both a;- and ,8-mannans and Marchessault (1962) regarded

RhodotoTula

t c: o .;;;., .~ c: ~

3'20

4000 3'iOO 3000 1800 1600 1400 1200 1000 800 600 400 250

Wavenumber (em-l ) Fig. I.-Spectra of mannans derived from yeasts and of mixtures involving the mannan from Saccharomyces cerevisiae. its presence in the spectra of polysaccharide mixtures as indicative of mannan. Marchessault's conclusion is confirmed by the occurrence of this band in the spectra of mixtures of Saccharomyces mannan with cellulose I, oc-chitin, laminaran, or casein 348 A. J. MICHELL AND G. SCURFIELD

{Fig. 1). It likewise occurs in the spectrum of Trichosporon fer mentans galactomannan (Fig. 1) and in that of the glycoprotein fraction 2A (Fig. 5) isolated from Pithomyces chartarum by Sturgeon (1964). Some differences between the yeast mannan spectra are evident, but these do not appear to reflect the nature of the linkages or the degree of branching of the mannans. For example, the Rhodotorula glutinis extracellular mannan is linear and contains fi-l,3- and fi-l,4.linked mannopyranose residues (Gorin, Horitsu, and Spencer 1965), whereas the mannan from Saccharomyces cerevisiae is highly branched and seems to be built of a backbone chain of oc-l,6-linked mannose units with side chains of oc-l,2 and oc-l,3 linkages (Stewart, Mendershausen, and Ballou 1966).

I § ]

4000 3500 3000 1800 1600 1400 1200 1000 800 600 400 250 Wavenumber (em-I)

Fig. 2.-Speotra of some referenoe ",.gluoans alone and in admixture.

The broad band centred at 1530 cm-1 in the spectrum for Oandida albicans mannan suggests the presence of protein impurity (see below), such impurity possibly being responsible also for the covering up of a band in the region of 900 cm-1 . The spectrum of Trichosporon aculeatum mannan is distinguished by sharp peaks at 1260 cm-1 and 617 cm-1 and that of the Rhodotorula mannan by a band at 740 cm-1, a broad band centred on 885 cm-I, and the absence of a band in the region 975-965 cm-1. ------,

IDENTIFICATION OF FUNGAL CELL WALL POLYMERS 349

(b) Ghwans (i) rx-Glucans The spectra in Figure 2 show that both glycogen and nigeran absorb at c. 930 and 850 cm-i . In addition, a band occurs at 760 cm-i in the glycogen spectrum, and at 790 cm-i in the nigeran spectrum. According to Barker, Bourne, and Whiffen (1956), a band at 850 cm-i is characteristic of the spectra of polysaccharides having the rx-configuration, bands at 930 and 760 cm-i are features of the spectra of a-1,4- linked glucans, and a band at 787 cm-i is a feature of the spectra of rx-1,3-linked glucans. It is noteworthy, therefore, that the spectrum of nigeran-a linear glucan in which a-1,4-linkages alternate with a-1,3-linkages (Barker, Bourne, and Stacey 1953)-contains no band at 760 cm-i . The spectrum of a 1 : 1 mixture of glycogen and cellulose I (Fig. 2) largely resembles that of cellulose I (Fig. 3). However, in this spectrum the characteristic cellulose band at 895 cm-i is overlain and absorption takes place over a broad region centred on 850 cm-i , as well as at 760 and 702 cm-i . Similarly, the spectrum of a 1 : 1 mixture of nigeran and rx-chitin (Fig. 2) is very much like that of a-chitin (Fig. 5) save for the occurrence of bands near 850 and 787 cm-i indicative of the presence of nigeran in the mixture.

(ii) f3-Glucans The spectra of some linear f3-glucans having only one type of linkage between residues and of some others having more than one type of linkage are shown in Figure 3. All the spectra show a band at 890 cm-i . This band was listed by Barker, Bourne, and Whiffen (1956) as a characteristic feature of f3-glucan spectra. Apart from this, the cellulose I spectrum is distinguished by strong absorption at 1428 cm-l, a doublet at 1335 and 1315 cm-l, bands at 1058 and 1025 cm-l, as well as no fewer than nine bands in the region between 700 and 250 cm-i . Absorption at or near 2920, 1370, 1248, and 1200 cm-i is noteworthy in the spectra of paramylon and laminaran. The band pattern in the region 1170-890 cm-i appears to be the most useful for distinguishing f3-1,6-glucans such as pustulan and lutean. The spectrum of crown gall glucan is perhaps least distinctive, though the band at 1175 cm-i appears to be significant. This spectrum does not differ significantly from that of the glucan derived from Mycoplasma mycoides. As regards the spectra of glucans which contain more than one type of linkage, those of barley glucan and lichenin resemble the spectrum of oat glucan. The spectra of laminarans derived from Laminaria cloustoni, L. digitata, and Eisenia bicyclis are indistinguishable from one another. In other words, spectra give little indication of the differences known to exist between the cereal glucans and lichenin in respect to the relative proportions of f3-1,3- and f3-1,4-linkages in their chains (references cited by Clarke and Stone 1963). Nor is there any indication ofthe fact that Eisenia laminaran is linear, has f3-1,3- and f3-1,6-linkages in the ratio of 2 : 1, and is mannitol-free, while laminarans from Laminaria spp. consist of mannitol- or glucose-terminated chains of f3-1,3- and possibly a few f3-1,6-linked residues joined by occasional f3-1,6-interchain 350 A. J. MICHELL AND G. SCURFIELD linkages (references cited by Bull and Chesters 1966). The situation is thus similar in some respects to that shown above to exist in the case of mannans.

530 5'0 Amorphous cellulose ~(1-4) 360 Cellulose I (1--4)

t o= 6eo .;; 620 ·s

060"" Lutean 1-3,1-4)

4000 3500 3000 1800 1600 1400 1200 1000 800 600 400

Wavenumber (cm-1)

Fig. 3.-Spectra of some reference ,B-glucans.

It is possible, however, to obtain some indication of the predominant type of linkage present in ,S-glucans which contain more than one type of linkage from an examination of their spectra. Thus, the spectrum of laminaran (a predominantly ,8-1,3-glucan) resembles that of paramylon, the spectrum of lutean (a predominantly ,S-1,6-glucan) resembles that of pustulan, and the spectrum of oat glucan (a pre dominantly ,8-1,4-glucan) resembles that of amorphous cellulose. Mention of amorphous cellulose rather than cellulose I emphasizes that isolated glucans are likely IDENTIFICATION OF FUNGAL OELL WALL POLYMERS 351

to differ in crystallinity as well as in the type of linkage which they contain. The effects of such differences in crystallinity might be gauged from a comparison of the spectra of amorphous cellulose and cellulose I shown in Figure 3. The intensities of the bands at 1428 and 1110 cm-I are much less, and the intensity of the band at 895 cm-1 is greater, in the spectrum of amorphous cellulose. This spectrum also differs from that of cellulose I in the region 700-300 cm-I . These results confirm those of Nelson and O'Connor (1964). They also suggest that differences between the spectra of the laminarans and paramylon could be explained in terms of the lower crystallinity of the laminarans, since bands at 1410, 1l05, and 1000 cm-I in the paramylon spectrum are either very much weaker in, or absent from, the spectra of the laminarans.

Paehyman ex Porio eoeus

ct .2 ·s::l '"c ...os E-<

fuber·regium sclerotia

Clavieeps sp. gluean

4000 3500 3000 1800 1600 1400 1200 1000 800 600 400 Wavenumber (em-I)

Fig. 4.-Spectra of some j3·g1ucans derived from fungi.

The spectra, shown in Figure 4, of certain f1-glucans derived from fungi can be interpreted in terms of the above conclusions. Without exception, the spectra suggest that f1-1,3-linkages are likely to be the predominant ones in the glucans. This is especially true of yeast glucan and of the residue (yeast hydroglucan) obtained from 352 A. J. MICHELL AND G. SCURFIELD it by treatment with boiling 2% hydrochloric acid (Houwink and Kreger 1953). The spectra of these compounds duplicate the spectrum of paramylon more closely than any other, a result which is consistent with the finding of identical X-ray patterns for yeast "hydroglucan" and paramylon (Kreger and Meeuse 1952). The spectrum of Poria cocos glucan is intermediate between the spectra of paramylon and laminaran, whilst those of the other glucans appear to resemble the laminaran spectrum more closely. The differences probably reflect differences in crystallinity of the glucans. Varying intensities of absorption in the region 1550-1535 cm-1 in several of the spectra are believed to be related to the presence of protein as an impurity (see below). The spectra of cellulose I + mannan and of laminaran + mannan are shown in Figure 1. Nearly all the peaks characteristic of cellulose I persist in the former spectrum, although some are less clear; the region 560-398 cm-1, for example, is relatively featureless. The spectrum of laminaran + mannan is very much like that of laminaran, but bands at 915 and 810 cm-I indicate mannan in the mixture. The spectrum of cellulose I + laminaran leaves no doubt as to the presence of cellulose in the mixture. However, it gives little indication of the nature of the additional component since this merely produces changes in the intensities of characteristic cellulose bands in the spectrum. The same is true of the laminaran + chitin spectrum: only when the amount of laminaran in the mixture considerably exceeds that of chitin is any indication obtained of its presence.

(c) rx-Chitin, Chitosan, and Protein Spectra of chitin and chitosan have been published previously (Michell and Scurfield 1967). In Figure 5 they are extended down to 250 cm-I and, together with a spectrum of casein as a representative protein spectrum, are included for the sake of ease of comparison with the spectra of mixtures. Bands near 3265, 3105, 1655, 1620, 1550, 953, and 392 cm-I are characteristic of chitin spectra; the band at 1590 cm-I appears to be the most distinctive one for chitosan. In the region 4000-1450 cm -1 the casein spectrum differs only quantitatively from that of chitin except in having a single band centred near 1645 cm-1, but the spectrum is quite different in the region 1450-250 cm-I where many of the chitin bands arise from vibrations of the sugar backbone. In the spectrum of a 1 : 1 mixture of chitin and cellulose I (Fig. 5), bands indicative of the presence of chitin occur at 3265,3105,1655,1620, 1550, and 953 cm-I, though their intensity is in all cases reduced compared with those in the spectrum of chitin alone. The presence of a band at 1428 cm-I and of several bands characteristic of cellulose in the region 690-250 cm-I is sufficient evidence that the mixture includes cellulose I also. However, cellulose becomes increasingly difficult to detect in this way as its degree of crystallinity decreases. The spectrum of casein + chitin is similar to that of chitin (Fig. 5). However, the former has distinctive bands at 1440 and 1235 cm-I . That these differences are sufficient to distinguish a spectrum of chitin + protein from one of chitin alone is confirmed by their expression in the spectrum of the larval cuticle chitin-protein complex (Rudall 1967). IDENTIFICATION OF FUNGAL CELL WALL POLYMERS 353

I c o :§ ~ ~

"" 4000 3500 3000 1800 1600 1400 1200 1000 800 600 400 250 Wavenumber (em-I)

Fig. 5.-Spectra of some known fungal cell wall constituents, alone and in admixture, and of a chitin-protein complex of insect origin. 354 A. J. MICHELL AND G. SCURFIELD

Bands at 3080, 1530, and 1440 cm-I in the spectrum of casein + mannan are sufficient to indicate the presence of a protein as one component in the mixture. It has already been noted that a band at 810 cm-I plus, in this case, bands at 970 and 910 cm-I, point to mannan as the second component. In the spectrum of casein + laminaran, bands at 1530 and 1440 cm-I again indicate the presence of casein as a component of the mixture. The spectrum otherwise resembles very closely that of a poorly crystalline f3-1,3-glucan or a glucan in which f3-1,3-linkages are the pre dominant ones. The spectra of the Pithomyces glycoproteins 2A and 2B (Sturgeon 1964) resemble the spectra of the casein mixtures with laminaran and mannan. Evidently, the mode of linkage between the glycan and protein (if, indeed, they are linked) has little effect on their infrared spectra. These results are wholly in accord with the isolation of glucose as a major product of the hydrolysis of the glycoproteins, the occurrence of mannose in hydrolysates of 2A but not 2B, and the considerable amount of protein nitrogen which they contain (Sturgeon 1964).

(d) Fungal Cell Wall Composition Spectra of residues remaining after various chemical treatments had been applied to cell wall material of Phytophthora cinnamomi, Pythium spp., and Epicoccum spp. have been published previously (Michell and Scurfield 1967). Redetermined spectra are presented in Figure 6 along with that of similarly treated wall material from Ceratocystis ulmi. In the case of both Phytophthora and Pythium the spectra were interpreted previously as indicating the presence in their walls of cellulose I, albeit of uncertain crystallinity. However, it will be evident from a comparison of the spectra in Figures 3 and 6 that Phytophthora and Pythium walls contain, not only cellulose I, but a laminaran-like glucan also. This conclusion is in accord with the finding of cellobiose, laminarabiose, and gentiobiose amongst the products of hydrolysis of the cell walls of Phytophthora and Pythium spp. (Aronson, Cooper, and Fuller 1967; Bartnicki-Garcia and Lippman 1967). In the case of Epicoccum, the spectrum resembles most closely that of a mixture of chitin + laminaran (Fig. 6). Our previous tentative suggestion (Michell and Scurfield 1967) that cellulose I might be a component of Epicoccum walls has not been substantiated. It has been suggested, on the basis of X-ray diffraction data, that the walls of Ceratocystis ulmi and C. olivacea contain both chitin and cellulose (Rosinski and Campana 1964; Rosinski 1965; Smith, Patik, and Rosinski 1967). The spectrum (Fig. 6) of the residue obtained after extracting the walls of C. ulmi successively with potassium hydroxide, glacial acetic acid-hydrogen peroxide, and sulphuric acid confirms that ex-chitin is indeed a wall component. A mann an (indicated by the band at 815 cm-I ), and ex- and f3-glucans may also be present. The indications are that the f3-glucan is a 13-1,3- rather than a f3-1,4-glucan (cf. the spectrum of C. ulmi in Figure 6 with the spectra of the mixtures ex-chitin + laminaran and ex-chitin + cellulose I). While the presence of a f3-1,3-glucan might be anticipated in view of the almost universal occurrence of such glucans in fungal cell walls, the presence of IDENTIFICATION OF FUNGAL CELL WALL POLYMERS 355 cellulose cannot be precluded on this evidence alone. If present, however, it must be of low crystallinity or small in amount or both. This matter is being investigated further.

Cellulose I + 1: 1

1 c:: o ·ii ·1 ~

4000 3500 3000 1800 1600 1400 1200 1000 800 600 400 250 Wavenumber (em-I)

Fig. 6.-Spectra of cell walls of species of Phytophthora, Pythium, Epicoccum, and GeratocY8tis and of the mixtures chitin + laminaran and cellulose I + laminaran. Cell wall materials from the first three of these species were dispersed in potassium chloride which absorbs below 400 cm-I .

Spectra of water-washed cells of Rhodotorula rubra, Schizosaccharomyces malidevorans, Oandida mycoderma, and Pichia membranaeformis are almost identical with the spectrum of similar cells of Saccharomyces cerevisiae shown in Figure 7. Most of the bands are fairly broad. Despite this, bands centred on 970 and 810 cm-1 are indicative of the presence of mannan; bands centred on 1640, 1535, 1450, 1400, and 1235 cm-1 are suggestive of the presence of protein, at least in the presumed absence of chitin or chitosan in amounts sufficient to exert any influence on the spectra; and bands centred on 1370, ll20, 1040, and 900 cm-1 are indicative of a 356 A. J. MICHELL AND G. SCURFIELD

1s:: o :; j

'OJ!)

4000 3500 3000 1800 1600 1400 1200 1000 800 600 400 250 Wavenumber (em-I)

Fig. 7.-Spectra of the fungus Saccharomyces cerevisiae before and after various chemical treat· ments: 1, water-washed walls; 2, after extraction with boiling methanol; 3, after treatment with 6% sodium hydroxide; 4, final residue after treatment with hot 6% sodium hydroxide and hot water (Bell and Northcote 1950); 5, final residue after treatment with boiling water, methanol, 5% potassium hydroxide, 5% acetic acid, 2% potassium permanganate, 2% oxalic acid, and 0·5N hydrochloric acid (Aronson and Preston 1960a); 6, final residue after treatment with 2% potassium hydroxide and 1 : 1 (v/v) mixture of glacial acetic acid and 30% hydrogen peroxide (Aronson and Preston 1960b); 7, mechanically isolated cell walls (Phaff, personal communication); 8, synthetic mixture casein + mannan + glucan (1 : 2 : 2). ------~------.------

IDENTIFICATION OF FUNGAL OELL WALL POLYMERS 357

,8-glucan. Bands centred on 2920, 2850, and 1740 cm-I may arise, in some part, from lipids. This interpretation of the spectrum is in accord with the following facts: (1) The spectrum of washed cells of Saccharomyces cerevisiae resembles that of mechanically isolated cell walls. Differences such as the greater intensity of the 1530 cm-I band in the former spectrum could be accounted for if the water-washed cells contained a larger amount of protein. (2) Bands at 2850 and 1740 cm-I in the spectra of water-washed cells are weaker in the spectra of such material after extraction with boiling methanol. (3) The spectrum of the cell wall material is almost identical with that of an artificial mixture of yeast glucan, yeast mannan, and casein in a ratio of 2 : 2 : 1. This ratio is an approximation to the reported composition of the cell walls of S. cerevisiae (Northcote and Horne 1952). (4) The residues remaining after the extraction with alkali of the washed cells of S. cerevisiae, and of the yeasts listed above, give spectra almost identical with that of a ,8-1,3-glucan such as paramylon or laminaran. The major difference is the occurrence in the spectra of the residues of bands at 2850 and 1740 cm-I . The treatment, in other words, was successful in removing protein and mannan, but possibly not lipid. It follows that spectra can provide not only information about the chemical composition of cells prior to chemical treatment, but can also be used to monitor the effect of treatment. For example, the spectra suggest that stages following the first extraction with 6% sodium hydroxide at 60°C in the procedure used by Bell and Northcote (1950) to isolate yeast glucan produce little further purification of the material. The spectrum of the final residue obtained by this procedure suggests the persistence in it of protein and possibly lipid or chitin or both. Similarly, the complex extraction procedure involving the use of potassium permanganate (Aronson and Preston 1960a) achieves, at least as far as yeast cell walls are concerned, very little more than can be achieved by extraction with 6% potassium hydroxide at 60°C.

IV. CONOLUSIONS (1) Infrared spectra can be used for indicating the presence of mannans in admixture with ,8-glucans, such as cellulose I or laminaran, or with O(-chitin or protein, and also in heteropolymers such as galactomannans. The spectra cannot, as yet, be interpreted in terms of the type of linkage present or of the degree of branching of the mannans. (2) Infrared spectra can be used for indicating the presence of O(-glucans in admixture with O(-chitin and, with less certainty, cellulose 1. (3) Infrared spectra can be used as aids in characterizing ,8-1,2-, ,8-1,3-, ,8-1,4-, and ,8-1,6-glucans, and for indicating the nature of the predominant type of linkage, but not the degree of branching, in ,8-glucans containing more than one type of linkage. They can also be used for detecting the presence of ,8-glucans in admixture with certain other substances, but with less certainty than in the cases of mannans in similar mixtures. Some indication might 358 A. J. MICHELL AND G. SCURFIELD

be obtained, for example, of cellulose I in the presence of IX-chitin, but laminaran in admixture with IX-chitin would be much more difficult to detect. In the case of mixtures of two f1-glucans, an indication of the nature of one glucan might be obtained, but rarely the nature of both glucans. Cellulose I, for example, might be detected in the presence of laminaran, but it is doubtful iflaminaran could be detected in the presence of cellulose I. (4) Infrared spectra can be used to distinguish between IX- and f1-chitin (Hackman and Goldberg 1965; Michell and Scurfield 1967), and to detect IX-chitin in admixture with mannans, IX- or f1-glucans, protein, or lipid. (5) Infrared spectra can be used as a means for recognizing the presence of protein in admixture with either mannan or f1-glucan. The detection of protein in admixture with IX-chitin is less certain~ The validity of the above conclusions regarding the value of infrared spectro scopy as a means for indicating the chemical composition of fungal cell walls is dependent on at least three factors. These are the relative proportions of the wall components, their nature, and degree of crystallinity. Laminaran, for example, is detectable in mixtures with chitin only when the ratio of laminaran : chitin exceeds about 6 : 1. Protein can be detected in the presence of glucans with more certainty than it can in the presence of chitin. Similarly, highly crystalline cellulose I is more readily detected than amorphous cellulose in mixtures with other components. The effect of differences in crystallinity, indeed, is such as to render the extrapolation of conclusions about the limits of detection of components in artificial mixtures to cell walls a doubtful procedure. Possible differences in the grinding properties of the polymers have also to be borne in mind. Despite these limitations, however, it is evident that infrared spectroscopy affords a relatively rapid and easy means of obtaining an indication of the nature of the major components of fungal cell walls. Its usefulness clearly extends beyond merely distinguishing between species with chitinous and non-chitinous walls. Furthermore, it offers a means of judging the efficacy of chemical treatments designed to separate major wall components.

V. AOKNOWLEDGMENTS The authors are grateful to Dr. B. A. Stone, University of Melbourne, for helpful discussions and for supplying many of the reference compounds. They would also like to thank Dr. N. H. Behrens, Dr. C. T. Bishop, Mr. S. H. Buttery, Dr. P. A. J. Gorin, Dr. I. R. Johnston, Dr. N. B. Madsen, Dr. A. Misaki, Professor S. Peat, Professor H. J. Phaff, Dr. B. C. Rankine, Professor M. A. Rosinski, Dr. K. M. Rudall, and Dr. R. J. Sturgeon for kindly supplying others.

VI. REFERENOES ARONSON, J. M. (1965).-In "The Fungi". (Eds. G. C. Ainsworth and A. S. Sussman.) Vol. l. pp. 49-76. (Academic Press, Inc.: New York.) ARONSON, J. M., COOPER, B. A., and FULLER, M. S. (1967).-Glucans of oomycete cell walls. Science, N. Y. 155, 332-5. IDENTIFICATION OF FUNGAL CELL WALL POLYMERS 359

ARONSON, J. M., and PRESTON, R. D. (1960a).-An electron microscopic and X.ray analYsis of the walls of selected lower Phycomycetes. Proc. R. Soc. B 152, 346-56. ARONSON, J. M., and PRESTON, R. D. (1960b).-The microfibrillar structlll"e of the cell wall of the filamentous fungus Allomyces. J. biophys. biochem. Oytol. 8, 247-56. BARKER, S. A., BOURNE. E. J., and STACEY, M. (1953).-Studies of Aspergillus niger. I. The structlll"e of the polyglucosan synthesised by Aspergillus niger 152. J. chem. Soc. 1953, 3084--90. BARKER, S.A.,BoURNE, E. J.,and WHIFFEN, D. H. (1956).-In "Methods of Biochemical Analysis". (Ed. D. Glick.) Vol. III. pp.213-45. (Interscience Publishers: New York.) BARTNICKI· GARCIA, S. (1966).-Chemistry of hyphal walls of Phytophthora. J. gen. Microbiol. 42,57-69. BARTNICKI-GARCIA, S. (1968).-Cell wall chemistry, morphogenesis and taxonomy of fungi. A. Rev. Microbiol. 22, 87-108. BARTNICKI-GARCIA, S., and LIPPMAN, E. (1967).-Enzymic digestion and glucan structure ofhyphal walls of Phytophthora cinnamomi. Biochim. biophys. Acta 136, 533--43. BEHRENS, N. H., and CABIB, E. (1968).-The biosynthesis of mannan in Saccharomyce8 carla bergensi8. J. biol. Ohern. 243, 502-9. BELL, D. J., and NORTHCOTE, D. H. (1950).-Structure of a cell wall polysaccharide of bakers' yeast. J. chern. Soc. 1950, 1944--7. BUCK, K. W., CHEN, A. W., DICKERSON, A. G., and CHAIN, E. B. (1968).-Formation and structure of extracellular glucans produced by Olaviceps species. J. gen. Microbiol. 51, 337-52. BULL, A. T., and CHESTERS, C. G. C. (1966).-The biochemistry of laminaran and the nature of laminaranase. Adv. Enzymol. 28, 325-64. CLA.RKE, A. E., and STONE, B. A.(1963).-Chemistry and biochemistry of ,8-1,3-g1ucans. Rev. pure appl. Ohem. 13, 134--56. GORIN, P. A. J., HORITSU, K., and SPENCER, J. F. T. (1965).-An exocellular mannan, alternately linked 1,3-,8 and 1,4-,8 from Rhodotorula glutinis. Oan. J. Ohem. 43, 950--4. GORIN, P. A. J., MAZUREK, M., and SPENCER, J. F. T. (1968).-Proton magnetic resonance spectra of Trichosporon aculeatum mannan and its borate complex and their relationship to ohemical structure. Oan. J. Ohem. 46, 2305-10. GORIN, P. A. J., and SPENCER, J. F. T. (1968a).-Galactomannans of Tricho8poron jermentans and other yeasts; proton magnetic resonance and chemical studies. Oan. J. Ohem. 46, 2299-304. GORIN, P. A. J., and SPENCER, J. F. T. (1968b).-Structural chemistry of fungal polysaccharides. Adv. Oarbohyd. Ohem. 23, 367-417. HACKMAN, R. H., and GOLDBERG, M. (1965).-Studies on chitin. VI. The nature of ex- and ,8-chitins. Aust. J. biol. Sci. 18, 935-46. HOUWINK, A. L., and KREGER, D. R. (1953).-Cell wall of yeast~lectron miorosoope and X-ray diffraotion. Antonie van Leeuwenhoek 19, 1-24. JOHNSTON, I. R. (1965).-The oomposition of the oell wall of A8pergillus niger. Biochem. J. 96,651-8. KREGER, D. R. (1967).-X-ray aspeots of yeast oell walls with partioular reference to the gluoan components. Abh. dt. Akad. Wiss. Berl. 1967, 81. KREGER, D. R., and MEEUSE, B. J. D. (1952).-X-ray diagrams of Euglena paramylon, of the acid-insoluble glucan of yeast cell walls and of laminaran. Biochim. biophys. Acta 9,699--700. MARCHESSAULT, R. H. (1962).-Application of infrared spectroscopy to cellulose and wood polysaccharides. Pure appl. Ohem. 5, 107-129. MICHELL, A. J., and SCURFIELD, G. (1967).~omposition of extracted fungal cell walls as indioated by infrared spectroscopy. Archs Biochem. BiophY8. 120, 628--37. NELSON, M. L., and O'CONNOR, R. T. (1964).-Relation of certain bands to oellulose crystallinity and lattioe type. I. Spectra of lattice types I, II, III and amorphous cellulose. J. appl. Polym. Sci. 8, 1311-24. NORTHCOTE, D. H., and HORNE, R. W. (1952).-Chemical composition and structure of yeast oell walls. Biochem J. 51, 232-6. 360 A. J. MICHELL AND G. SOURFIELD

PEAT, S., WHELAN, W. J., and EDWARDS, T. E. (1955).-Polysaccharides of bakers' yeast. I. Glycogen. J. chem. Soc. 1955, 355-9. PEAT, S., WHELAN, W. J., and EDWARDS, T. E. (1958).-Polysaccharides of bakers' yeast. II. Yeast glucan. J. chem. Soc. 1958, 3862-8. PEAT, S., WHELAN, W. J., and EDWARDS, T. E. (1961).-Polysaccharides of bakers' yeast. IV. Mannan. J. chem. Soc. 1961, 29-34. ROSINSKI, M. A. (1965).-Further confirmation of the occurrence of cellulose in Oeratocystis ulmi. Mycologia 57, 668. ROSINSKI, M. A., and OAMPANA, R. J. (1964).-Ohemical analysis of the cell wall of Oeratocystis ulmi. M ycologia 56, 738-44. RUDALL, K. M. (1967).-In "Oonformation of Biopolymers". (Ed. G. N. Ramachandran.) Vol. 2. pp.751-65 (Academic Press, Inc.: London.) SMITH, M. J., PATIK, O. M., and ROSINSKI, M. A. (1967).-A comparison of cellulose production in the genus Oeratocystis. Mycologia 59, 965-9. STEWART, T. S., MENDERSHAUSEN, P., and BALLOU, O. E. (1966).-Structure of yeast mannan. Fedn Proc. Fedn Am. Socs. expo Biol. 25, 410. STURGEON, R. J. (1964).-Oomponents of the cell wall of Pithomyces chartarum. Biochem. J. 92, 60P-61P. Yu, R. J., BISHOP, O. T., OOOPER, F. P., HASENCLEVER, H. F., and BLANK, F. (1967).-Structural studies ofmannans from Oandida albicans (serotypes A and B), O. parapsilosis, O. stellatoidea and O. tropicalis. Oan. J. Ohem. 45, 2205-11.