Thermolytical techniques to characterize fungal polysaccharides and

Thermolytical techniques to characterize fungal polysaccharides and bacterial lipopolysaccharides. M. C. Ramos-Sanchez, A. Rodriguez-Torres, J. A. Lea...
0 downloads 0 Views 895KB Size
Blotechnol. Prog. 1991, 7,526-533

526

Thermolytical Techniques To Characterize Fungal Polysaccharides and Bacterial Lipopolysaccharides M. C. Ramos-Sinchez’ and A. Rodriguez-Torres Departamento de Microbiologia, Facultad de Medicina, Valladolid, Spain

J. A. Leal Centro de Investigaciones Biolbgicas, CSIC, Madrid, Spain

E’. J. Martin-Gil and J. Martin-Gil’ Laboratorio de Anllisis TBrmico, ETS Ingenieros Industriales, Paseo del Cauce, s/n, 47011 Valladolid, Spain

The present work constitutes an entirely novel contribution in the scope of microbiology and especially in taxonomy, introducing thermolysis curves as a rapid method of characterization of fungal polysaccharides and bacterial lipopolysaccharides. The thermal analysis techniques applied were thermogravimetry and derivative thermogravimetry (TG-DTG), differential thermal analysis (DTA), and differential scanning calorimetry (DSC). Each thermogram of a sample is represented by one or a few temperatures and, in DSC, by complementary enthalpy data. The temperatures of the thermograms from structurally unknown polysaccharides are compared with those used as references, and thus, information on their composition, linkage types, and anomeric configuration can be deduced. The situation is more complicated for bacterial lipopolysaccharides, but in whatever mode, a structural estimation is always possible. In the course of the development and validation of the thermal method, structural findings on relative stabilities of linkage types (valuable in carbohydrate research) have been recognized and are therefore also described in this work.

Introduction Thermal techniques such as thermogravimetry-derivative thermogravimetry (TG-DTG), differential thermal analysis (DTA), and differential scanning calorimetry (DSC) can be used for the rapid analysis of polymeric compounds with only small amounts of solid and liquid samples. The aim of this paper was to assess their use in the characterization of fungal cell-wall polysaccharides and bacterial cell-wall lipopolysaccharides that are considered, in microbiology, to be useful in delineating taxa at various levels (Bartnicki-Garcia, 1968,1970) and in the research of antiviral (Cacabelos, 1988) and antitumor activity (Whistler et al., 1976; Chasseing et al., 1988). Elucidation of the structure of polysaccharides and lipopolysaccharidesinvolvessugar and lipid analysis, linkage analysis, and the determination of anomeric configuration and the sequences of the sugar residues. A wide variety of chemical methods (methylation, partial depolymerization, oxidation by chromium trioxide, ...I and physical techniques (TLC, GLC, HPLC, Py-MS, IR, lH NMR, 13C NMR, ORD, etc.) are applied for these purposes, making the procedure highly laborious. In view of this difficulty, the introduction of a single, almost completely elucidative method like the thermolytical method (TM) is highly desirable. Its all-around applicability is due to the fact that this method is based on integral chemical profiling (“fingerprinting”) of the complex organic material under investigation. Each thermogram of a sample is represented by one or a few temperatures and, in DSC, by complementary enthalpy data.

* To whom correspondence should be addressed. 8756-7938/91/3007-0526$02.50/0

The temperatures of different reference samples are scaled in correspondence to the varying primary structures. The temperatures of the thermograms from structurally unknown polysaccharides are compared with those used as references, and thus, information on their composition, linkage type, and anomeric configuration can be deduced j estimated. The situation is more complicated for bacterial lipopolysaccharides (the main component of Gram-negative cell walls), whose carbohydrate portion (core and 0-chain) frequently contain various sugar residues and multiple linkage types. In whatever mode, a structural estimation is always possible. In the course of the development and validation of TM, structural findings on relative stabilities of linkage types (valuable in carbohydrate research) have been recognized and are therefore also described in this paper.

Materials and Methods Fungal Polysaccharides. Polysaccharides of several species of Aphanoascus (Aphanoascus fulvescens,Aphanoascus reticulisporus, Aphanoascus terreus),Penicillium (Penicillium allahabadense,Penicillium digitatum, Penicillium spinulosum),Eupenicillium (Eupenicillium crustaceum, Eupenicillium stolkiae), Talaromyces (Talaromyces helicus, Talaromyces f l a w s ) , Microsporum (Microsporum gypseum), Trichophyton (Trichophyton ajelloi),Aspergillus (Aspergillusflavipes,Aspergillus cremeus, Aspergillus niger),Botrytis (Botrytis cinerea),Candida (Candida albicans), Gliocladium (Gliocladium viride),and Nomuraea (Nomuraearileyi) were obtained from microorganisms from the Centraalbureau von Schimmelcultures, Baarn, Holland; from the Centro de Investiga-

0 1991 American Chemical Society and American Institute of Chemical Engineers

Biotechnol. Prog., 1991, Vol. 7, No. 6

527

a

1

50.0

*

c

0.0 30

0

n 2 Lu

v

A x

0 Lu

0.0

0

9'0 Id0

ti0

2TO 330

390 450 Si0 570

TEMPERATURE "C Figure 1. TG and DTG thermograms for (a) laminaran from L. digitata (maxima at 243, 330, and 465 " C ) ; (b) mannan from S. cereuisiae (maxima at 260 (sh), 329, 437, and 470 " C ) ; and (c) P-galactomannoglucan (maxima at 266, 300, and 421 "C).

ciones Biokigicas, CSIC, Madrid, Spain; and from the Facultad de Medicina, Reus, Spain. Culture media, cellwall preparation, fractionation of cell-wall material, isolation of polysaccharides, chemical analysis, and methylation analysis have been described elsewhere (GbmezMiranda et al., 1981, 1984, 1988; Prieto et al., 1988). Other Reference Polysaccharides. Chitin, from crab shells (C 3641); laminaran, from Laminaria digitata (L 9634); cellulose, from plant cells; mannans, from Saccharomyces cerevisiae (M 7504 and M 3640); dextran, from Leuconostoc mesenteroides, strain B-512 F (D 9260); nigeran, from Aspergillus awamori (N 8634);pullulan, from Aureobasidiumpullulans (P 4516);amylopectin, from corn ( A 7780); amylose, type HI, from potato; chitosan, from crab shells (C0792); amylose, type 111, from potato; chitosan, from crab shells (C 0792); xylan, from oat spelts (X 0376); xanthan gum, from Xanthomonas campestris (G 0264); inulin, from chicory root (I 2255); fucoidan, from Fucus vesiculosus; glucosamine hydrochloride (G 4875); galactosamine hydrochloride (G 0264); mannosamine hydrochloride (M 4500);and glucuronic acid (G 5269)were supplied from Sigma Chemical Co. Pustulan, from Um-

I

.

'

.

200 240 a 0

.

.

3.u 360

'

.

I b o 440

.

.

4M si0

.

$60

TEMPERATURE

' C

Figure 2. DTA thermograms for (a) pullulan from A . pullulans (effects a t 365,433, and 504 "C);(b) mannans from S . cerevisiae (-) and C. albicans (- - -) (effects at 320-336,437, and 491 "C); and (c) 0-galactoglucan from T. flavus (effects a t 310 and 473 "C).

bilicaria pustulata (Pers. ex Fr.) was obtained by one of the authors (J.A.L.). Lipopolysaccharides. Chromatographically purified extracts from Escherichia coEi serotype Olll:B4 (L 3012), Shigella flexneri 1A (L 9018), Serratia marcescens (L 2512), Salmonella enteritidis (L 2012), Pseudomonas aeruginosa (L 8643), and Salmonella typhimurium (L 2262) were purchased from Sigma Chemical Co. Lipopolysaccharides from Brucella melitensis and Brucella abortus were isolated and purified by Dr. A. Ordufia. Instrumentation. Thermal analysis is a series of techniques that study the effect of temperature on material properties. TG measures any change in mass of a sample as a function of temperature or time. The DTA/DSC techniques measure temperature differences or enthalpy changes during physical transitions and chemical reactions. Either the temperature difference (DTA) between the sample and a reference material or the heat flux difference (DSC) is determined. Our experimental setup of thermal analysis is presented

Biotechnol. frog., 1991, Vol. 7, No. 6

528

Table I. TG-DTG Thermal Effects ( O C ) of Fungal Polysaccharides and Other Polysaccharides Used as References source DTG Deaks linkme polymer P. allahabadense, crab shells 318-313,341-336,520-490 Glci(1-4) chitins 267-243,343-330,483-465 L. digitata Glcfl(1-3) linear laminarans 330 plant cell walls Glcfl(1-4) linear cellulose 331-327 A. flavipes, E. crustaceum Glca(1-3) s-glucans 329-326,450-437,500-470 S. cerevisiae Mana(l-4), br a(1-3) and mannans a(1-2)b L. mesenteroides 206,326-322,444 Glca(l-.6), br a(1-3) dextran 312-302,325-3 16 Glca(1-3) and a(1-4) linear A. cremeus, A. niger nigerans 324.5 Glca(1-6) and a(1-4) linear A . pullulans pullulans 320.5,517 corn Glca(l+4), br a(1-4) amylopectin 319.5 Glca(l44) linear potato amylose A. fulvescensa 319 Man? (1+4)Glcfl( 1-41 glucomannans N. rileyi, B. cinerea 288,322-314,482-441 Glcfl(1-+3),br b(1-4) R-glucans P. allahabadense, A. pullulans, G . viridea 320-314,491-450 Glcfl(1-4) and b(1-3) pustulanoids 212,318-314,455.5 U.pustulata Glcp(1-4) linear pustulans A. reticulisporusa A. terreusa 314-3 11,452-436 Mana(1-4)Glcj3(1-4) (gluco)mannans and fl(1-3) C . albicans Mana(l-4), br 41-2) 313-296,438.5,510-490 mannans T . helicus 310,525 glucan-chitin 8(1-4) 305,603 crab shells chitosan o(1-4) 300,503-495 M. gypseump T. ajelloP Mana(l-4)Glcfl(1+4) glucomannans G . viridea 266, (302)-297.5,421 galactomannoglucans 8 0 - 4 , b(1-6) P. allahabadense, T. helicus 301-298,508-499 glucogalactans 80-4) 294-291,542-474 P. digitatum, T.flavusa galactoglucans 8(1-4,8(1-.6),fl(1-3) P. spinulosum, E . stolkiae 284-278,484-4ai Galfl(1-4) and fl(1-3) galactans oat spelts 283,441 Xylfl(1-4), br 41-21 xylan X . campestris 251,281,462 Glcfl(1-4)/Manfl(1-4) xanthan gum chicory root 242,278 Frufl(2-1) linear inulin 221,245,409,486 F. vesiculosus Fuca(l-2), br a(1-4) fucoidan 210,230 glucosamine hydrochloride 209,512 N-acetylglucosamine 196,477 galactosamine hydrochloride 180 mannosamine hydrochloride 168,203 glucuronic acid ~~

~

a

Unknown fungal polysaccharides. br = branched.

Table 11. DTA Curve Data for Polysaccharides constituent monomers thermal effects Dolvmer and linkage ("C) chitin Glcfl(1-4) 540,385 s-glucans Glca(1+3) 537,433,375 k 5 Glca(1~3),Glca(l+4) 506-491,433 f 10 nigerans Glca(l-4),Glca(l-+4) 504,433,365 pullulans Glcfl(1-3),Glcfl(1-4) 500-497,352 R-glucans pustulanoids Glc~(l-4),Glc~(1+3) 494,334 fl(1-4 492,390,331 glucan-chitin Mana(l-4) 492-490,432 f 5, mannans 336-320 B(1-4) 490-489,324-326 glucogalactans fl(l-4),fl(l43) 479-475,324-322 galactans fl(1-4) 473,310 galactoglucans Glcfl(1+4)Manfl(l--4) 454 xanthan gum 451,425,346 fucoidan Fuca(1-2) N-acetylglucosamine 385 galactosamine 368 hydrochloride mannosamine 348 hydrochloride ~

as follows, In short,the TG-DTG analyzer (Perkin-Elmer TGS-2 System) consists of the following units: the thermobalance, the electronic balance control, the temperature (program) controller, the heater control unit, the first derivative computer (FDC), and the recorder. PerkinElmer DTA 1700 system high-temperature differential thermal analyzer is made up of the 1700 analyzer module (which consists of the high-temperature furnace and the cell base for the furnace) and the system 4 analysis controller. To monitor the temperatures of both sample and reference materials, Pt and Pt-10% Rh thermocouples are used. The TG and DTA thermal data have been recorded in a Perkin-Elmer Model 3700 data station. The DSC 7 differential scanning calorimeter permits, with a PC, the direct calorimetric measurement, charac-

terization, and analysis of thermal properties of materials. The operation theory is based on the power-compensated "null balance" DSC principle, in which energy absorbed or emitted by the sample is compensated by adding or subtracting an equivalent amount of electrical energy to a heater located in the sample holder. The continuous and automatic adjustment of heater power (energy per unit time) necessary to keep the sample holder temperature identical to that of the reference holder provides a varying electrical signal equivalent to the varying thermal behavior of the sample. Operative Conditions. We have used the systematized recording conditions that follow: 2 mg of sample; heating rate of 10 OC/min; temperature range from 25 to 600 "C; atmosphere of dynamic air (60 cm3/min) in TG, static air in DTA, and dynamic Nz (20 cm3/min) in DSC; crucibles of platinum in TG, ceramic in DTA, and sealed capsules of aluminium in DSC, as sample containers; and A1203 (in TG and DTA) and a sealed and empty aluminium capsule (in DSC) as reference materials.

Results Thermal Results of Fungal Polysaccharides. The thermolytical features of the fungal polysaccharides TG curves have been found to be characteristic for each substance studied. Thermograms of some fungal polysaccharides are shown in Figure la-c. Examination of the thermograms reveals a minor effect a t 50 "C, which corresponds to the loss of absorbed water, and three/four major effects a t 255 f 15 "C, 310 f 30 OC, 450 f 30 OC, and/or 510 "C, in agreement with the values of the DTG peaks. The second of these is the most prominent and has the highest value for relative characterization purposes. It appears to involve elimination of polyhydroxy1 groups, accompanied by depolymerization and decomposition

529

Blotechnol. hog.., 1991, Vol. 7, No. 6 0,

4

I

t

'f

I

Figure 3.

'a

TEMPERATURE 'C DSC curves for (a) cellulose, (b) mannan, (c) fucoidan, (d) inulin, (e) amylopectin, and (f) hexosamines.

Table 111. DSC Curve Data for Polysaccharides endothermic effect(s) polymer onset ("C) peak ("C) chitin 146.6 151.8 162.3 164.5 cellulose 103.8 118.0 mannan 138.1 155.8 S-glucan (178.4) 179.9 nigeran pullulan 167.6 171.6 150.6 152.9 amylopectin 104.6 114.2 amylose galactan 166.0 174.7 177.1 xylan 156.7,205.1 162.2,206.6,263.0 inulin 171.6 174.1 fucoidan 187.6 188.8 pustulan 213.0 glucosamine hydrochloride galactosamine hydrochloride 190.0 177.0 mannosamine hydrochloride 154.9 158.0 glucuronic acid

responsible for the ensuing peaks. TG-DTG data of all the fungal polysaccharides studied and other model polysaccharides are described in Table I. The polysaccharides are ordered according the temperature of the main effect in order to show how structurally analogous com-

AH (J g-l) 193.3 84.5 142.2 331.7 (65.9) 345.5 143.0 50.9 237.1 39.9,206.1

exothermic effect(s) onset ("C) peak ("C) AH (J g-1) 390.0,338.5 332.5 344.8 -147.7 320.7 334.4 -172.8 315.6 310.0 304.3 292.0 286.9 287.7 -33.0 286.4 260.9 278.0 -488.7 271.8 201.0 211.7 -283.8

441.3

337.7

221.3

234.3

-203.0

pounds occupied successive or near locations. An approximate elucidation of an unknown polysaccharide can be performed by similitude or interpolation, bearing their temperatures (above all, that of the main peak) in this table.

Biotechnol. Prog., 1991, Vol. 7, No. 6

530

Table IV. Mole Percentages of the Neutral Sugars (as Alditol Acetates). Detected by GLC of the Polysaccharides Extracted with 1 M NaOH from Fungal Wall Material organism fraction xylose mannose galactose alucose G. viride F1S 0.0 26.1 19.1 54.7 F1P 6.2 2.9 0.7 90.1 M. gypseum F1S 0.8 67.8 0.4 4.0 T. ajelloi F1S 0.85 64.4 1.0 3.2 A. fuluescens F1S 4.5 65.7 4.2 9.1 70.6 3.1 8.0 0.6 A . reticulisporus F1S 54.0 7.8 4.6 A. terreus F1S trace 9.0 30.0 41.0 T. flauus F1 17.0 a Hydrolysis was performed with 4 N HzSOl at 100 "C for 5 h. All values are averages of three determinations.

dl

€4

90

L20

L50

in0

ti0

L40

2TO

TEMPERATURE 'C Figure 4. (a) T G and DTG thermograms for LPS S-form from B. melitensis in air at 10 "C min-1. (b) DTA thermogram for LPS from S. flerneri in air at 5 "C mi+. (c) DSC curve from P. aeruginosa at 10 "C mi+.

Apart from its elucidative value, the main DTG peak has shown its utility as a marker of thermal stability. This has allowed us, for the first time, to sequence exhaustively the thermal stability of the polysaccharides: chitin = laminarans > cellulose > s-glucans > a(1-+6)-, a(1+3)-, a(1+2)-branched mannans > dextran > nigerans = pullulans > amylopectin > amylose > Manj3(1-4)GlcB(l-4) glucomannans > R-glucans > P(1-6), P(1-3) glucans > pustulans > Mancu(l-.6)GlcS(l~),P(1-.4) glucomannans > a(l-6)-, a(l+2)-branched mannans > glucanchitin > chitosan > Mancu(l-.G)GlcS(1-4) glucomannans > galactomannoglucans > P-glucogalactans > P-galactoglucans > 8-galactans > xylan > xanthan gum > inulin > fucoidans > glucosamine hydrochloride > N-acetylglu-

cosamine > galactosamine hydrochloride > mannosamine hydrochloride > glucuronic acid. Conscious of the importance of the relationships of structure-stability and reducing the prior sequence to another of bond types, an order for the strength of the different linkage types has been estimated: P(1-4) (*-H--) > P(1-3) > a(1-3) > a(1-6) > a(1-4) > P(1-+6) > P(1-4) > P(2-1) > 41-2). This order is in agreement with the resistance to hydrolysis and is inverse to that of acetolysis. For the studied polysaccharides we can conclude that, independently of their a or /3 configuration, the scale of stabilities is (1-3) > (1-4) > (1-4) > (1+2), always bearing in mind the observation that the abnormally high stability of chitin and cellulose, with @(1-4) bonds, is achieved by hydrogen bonding between their chains. The DTA curves of the same polysaccharides studied by TG-DTG (Table 11) gave two moderate exothermics at about 350 and 430 "C, followed by a pronounced exothermic in the 540-450 "C region (Figure 2). The first and last peaks have value for the characterization of the polysaccharides, but the structural differences caused by the distinct sugar residues and linkage types are better reflected by the temperature differences of the peak which occurs at about 500 "C. These temperatures decreased in the same order that was previously established by DTG: briefly, chitin > S-glucans > nigerans == pullulans > R-glucans > pustulanoids > glucan-chitin 2 mannans > glucogalactans > galactans > galactoglucans > xanthan gum > fucoidan > hexosamines. The DSC standard curve of a polysaccharide shows, in the lower temperature region, an endothermic effect with peak at 150 f 30 "C and, in the far temperature region, an exothermic effect at 280 f 70 "C. This thermal behavior is exemplified by the curves of cellulose (Figure 3a), xylan, glucuronic acid, and mannan (Figure 3b). However, the remaining polysaccharides gave DSC curves less simplified, either by (a) the splitting of the first effect into various peaks, as for nigeran, pustulan, or fucoidan (Figure 3c), (b) the sequentiation in three steps of the decomposition attributed to endothermic effect, as occurs for inulin (Figure 3d), (c) the sudden pyrolysis (sensitized by a very sharp effect in the exothermic region, as for amylopectin (Figure 3e), or (d) the slow and undifferentiated pyrolysis of the hexosamines (Figure 3f). Table 111 summarizes the onset and peak temperature effects and the enthalpy values (AH) associated with such effects (positive/negative values indicating endo-/exothermic effects). The data from the endothermic effect are only useful for characterization aims, but those of exothermic effect additionally have permitted sequencing of thermal stabilities. The order obtained is very close to that reported by the other thermal techniques (DTG and DTA), thus confirming the previous results on bond-type relative strength. As regards the mechanism of the DSC reactions, the endothermic effect at 150 "C involves selective dehydration, whereas the exotherm at 300 "C appears to involve further elimination of the polyhydroxy1 groups, accompanied by depolymerization and early decomposition (with evolution of COz and CO). A t higher temperatures the decomposition is accomplished,with formation of methane and unsaturated hydrocarbons as gas emitted. Periodate and Methylation Results. In this paper, eight applications of the use of the thermolytical method to characterize polysaccharides are presented, viz., the study of two not yet fully chemically characterized fractions

Biotechnol. Prog., 1991, Vol. 7, No. 6

591

Table V. Mole Percentages of the Sugars (as Alditol Acetates). Detected by GLC after Smith Degradation of FlS, FlP, or F1 organism fraction GlY Ery Thr Rib Ara XYl Man Gal Glc 12.7 12.7 0.9 7.0 0.0 12.2 3.3 32.4 G. viride F1S 18.8 0.0 1.5 0.0 F1P 42.5 7.9 5.7 1.9 0.0 40.5 1.0 0.0 0.7 84.8 4.4 0.0 2.4 0.7 6.1 M.gypseum F1S 0.2 0.0 0.0 0.0 T.ajelloi F1S 99.3 0.0 0.1 0.0 0.4 A. fulvescens F1S 34.6 53.2 1.1 0.7 0.0 0.7 0.2 9.4 71.4 11.5 0.0 0.0 A. reticulisporus F1S 0.0 7.8 3.0 6.3 34.6 0.0 FlS 50.3 3.0 1.7 A. terreus 1.4 0.2 8.1 7.2 F1 24.9 12.9 0.9 9.5 T.flavus 18.0 1.8 8.8 16.0 a

Hydrolysis was performed with 4 N H2SO4 at 100 "C for 5 h. All values are averages of three determinations.

Table VI. GLC-MS Data for Methylated Alditol Acetates. from Fractions A and Bb relative mol (%) alditol P A B major mass spectrum fragments (m/z) 0.99 27.0 25.0 87,88,101,102,118,129,145,161,162,205 2,3,4,6-Me4-Mand 1.19 2.8 trace 87,88,101,102,118,129,145,161,162,205 2,3,4,6-Med-Gal 2,3,4-Mes-Glc 2.22 12.5 12.0 87,102,118,129,130,162,173,189,206,233 3,5,6-Me~-Gal 1.90 6.6 6.0 45,59,88,89,130,190,205,306 2,3,6-Me~-Gal 2.22 3.0 3.0 45,102,113,118,130,162,173,233 2,3,4-Mes-Gal 2.89 5.8 7.8 87,102,118,129,130,162,173,189,206,233 7.2 102,117,118,130,173,233 2,3,5-Me~-Gal 2.76 6.3 3,4-Me~-Glc 4.26 15.0 19.0 87,88,129,130,173,174,189,190,233,234 20.0 87,118,129,174,189,234 2,4-Mez-Glc 4.21 21.0

deduced linkage Manp(1Galp(1-6)Gl~p(l-2)Galf(l-4)Galp(1-6)Galp(l-6)Galf(l-2,6)Glcp(l-3,6)Glcp(l-

*

a Hydrolysis with 0.25 M H2S04 for 16 h at 100 "C. All samples were reduced with NaBD4. Retention time relative to that of 1,S-di0-acetyl-2,3,4,6-tetra-O-methylglucitol on OV-225 at 170 "C. 2,3,4,6-Me4-Man = 1,5-di-0-acetyl-2,3,4,6-tetra-O-methylmannitol, etc.

Table VII. DSC Thermal Effects for LPS of Gram-Negative Bacteria onset ("C) peak ("C) organism 199.2 200.9 E. coli 0111 S. flexneri 1A 197.5 200.6 198.6 S. marcescens 196.7 178.9 180.7 S. enteritidis 174.1 176.2 P. aeruginosa S. typhimurium 120.7 139.5 B. melitensis 16M 123.6 139.3 108.7 136.0 B. abortus B19

(J g-9 154.7 (36.5) 150.3 196.7 247.2 161.4 137.5

obtained from the cell-wall material of G. uiride and the study of unknown soluble fractions from M.gypseum, T. ajelloi,A. fulvescens,A. reticulisporus,A. terreus, and T. flauus. Available partial information from structural classical methods (periodate, methylation) is included as support. The neutral sugars released from the different fractions in the species under study are presented in Table IV. The mole percentages of the sugar detected by GLC after Smith degradation of such fractions are shown in Table V. For G. uiride, the GLC-MS data for the methylated alditol acetates from the major fractions (A and B) obtained from F1S are shown in Table VI. Thermal Resultsof Bacterial Lipopolysaccharides. For a series of Gram-negative bacteria, the thermolysis of their LPS showed a TG-DTG effect at 265 f 25 "C (Figure 4a), two DTA effects near 315 and 470 "C (Figure 4b), and a clearly well-defined endothermic effect in DSC a t 160 f 40 "C (Figure 4 4 . As regards the DSC thermal effect, the obtained onset and peak temperatures and the enthalpy change values have been summarized in Table VII. It can easily be seen that the decomposition temperatures of LPS of E. coli, S.flexneri,and S. marcescens are approximately equal and are higher than those of S. enteritidis and P. aeruginosa (also quite closely grouped) and much higher than those of B. melitensis and S. typhimurium (in turn, both very nearly similar). Within same group, the enthalpy changes AH fluctuate considerably due to variations of the supramolecular structure. The order of thermal stabilities is in agreement with that previously obtained for the whole bacteria (Ramos-

Shchez et al., 1988): Morganella morganii > Proteus mirabilis > Neisseria meningitidis, Neisseria lactamica, Neisseria gonorrhoeae > E. coli, Streptococcus pneumoniae, Staphylococcus aureus > Klebsiella pneumoniae, P. aeruginosa > S . typhimurium, Enterobacter cloacae, Streptococcus agalactiae, Streptococcus faecalis. This is to be expected from the fact that the LPS is the major constituent of the cell wall of Gram-negative bacteria. Thus, the thermal behavior of the bacteria seems to be largely independent of the presence of proteins and other components. In searching for a structural reason for the reported differences in thermostability, our attention was directed toward the nature of the polysaccharide residues and, above all, to the predominant linkage and configuration present in the LPS 0-chains of the studied species (Table VIII) (Perry & Bundle, 1990; Wilkinson, 1977). Stability of residues decreased in the order Glc-Glc > Glc-Gal > Glc-Rha > Man-Rha > Rha-Rha. As regards the linkage types, since it has been stated above that the relative strength for linkage types is p(1-3) > a(1-3) > ar(1-6) > a(1-4) > B(1-6) > p(1-4) > p(2-1) > ar(1-2), the presence of an excess of linkagesB(1-3) and (~(1-3) in the 0-polysaccharide structure must contribute to enhance the stability of the LPS, whereas the quantitative presence of (1-2) linkages leads to their lability. Thus, the relatively low stability of B. abortus [with only (1-2) linkages in ita 0-chain] and B. melitensis [one (1-3) and four (1-2) linkages] versus other Gram-negative bacteria [with 0-chains rich in ar//3(1+3), ar/j3(1-4), or ar-(1-6) linkages] becomes justified.

Discussion Systematic studies of the acid hydrolysis of oligo- and polysaccharides have show that the (1-6) linkages are more stable than are the (1-4) linkages and the (1-2) linkages (Swanson & Cori, 1948; Pazur, 1986). At any rate, in the wall fractionation according to the procedure of Mahadevan and Tatum, including solubilization in acid and alkaline solutions (Mahadevan & Tatum, 1965), the P(1-3) resulted the most stable linkages. Our obtaining of this same order of stabilities by thermal

Biotechnol. Rog., 1991, Vol. 7, No. 6

532

Table VIII. Partial Structures for the Repeating Units in the LPS 0-Specific Side Chains of Gram-Negative Bacteria** COlpal C O l ~ l E. coli 0111

t

-

6

D-GIcNAcP( 1

t

4

J)-o-Glcpx(1

-

3)-o-Gal

-4)-~-GlcpB( 1-+3)-~-GalpNAca( 1+2)-~-RhapNAca(1+3)-~-Fucpa(1o-Glcpa(1 + 2)-a-DGkpal

-

-

2)-~-Rhapa(l- P ) - ~ - R h a p ( l - 3 ) - ~ - R h a p ( l

t 4

S. flexneri

-

3)-o-GlcpNAcp(l

S. marcescens 0 2 2 and 010

~-Glcpl

-

E. coli 0157

t 4

Z)-L-Rhapa(l- 2 ) - ~ - R h a p ( l - 3 ) - ~ - R h a p ( l +

DDHpl

3)-~-GlcpNAca(l -D S. enteritidis

~-Glcpl

t

t

3

4%

3)-~--Glcpx(l-3)-~-Rhapa( 1-+2)-~-Rhapa( 1-3)-~-Rhapa(l-

P. aeruginosa

-2)-~-Rhap4NFoa( 142)[-~-Rhap4NFoa(1-2)] e-~-Rhap4NFoa( 1+3)-~-Rhap4NFoa(1-

B. melitensis

-.2)-~-Rhap4NFoa(l+

B. abortus

2 ) - ~ - M a n p r ( l - 4 ) - ~ - R h a p r (-tD

Wilkinson, 1977;Perry & Bundle, 1990. * Col = colitose;DDH = 3,6-dideoxyhexose(paratose, abequose, or tyvelose); NFo = N-formamido.

analysis leads us to equate thermostability with hydrolysis resistance and, therefore, to justify the soundness of the thermal method by parallelism with the hydrolytic method, already well established. In order to probe the capabilities of our techniques for characterization purposes, the thermolytical data obtained of unknown polysaccharides have been interpreted on the basis of comparison with reference compounds (evidence obtained with analytical methods commonly used is also included as security). Two examples are the F1S and F1P fractions of G. uiride. In thermal analysis, the main DTG temperatures of these fractions appear at 297.5 and 316 OC, respectively. The first temperature can be located, in Table I, in the region between the glucomannans with (1-4) and (1-6) linkages and the glucogalactans with 8(1-4) linkages, denoting for F1S an intermediate composition, as really occurs. The second temperature can be interpolated between the peak temperatures of pustulans and R-glucans, thus assuring for F1P their nature as 8-glucan with (1-.6) and (1-3) linkages. Information derived from the classical methods (Tables IV-VI) did not differ from that obtained from the TM since F1S was chemically characterized as a galactomannoglucan (54.7%) and F1P as a glucan (45.3%). For FlS, the results of methylation analysis suggested a (1-6) glucan backbone substituted at C2 and C3 by mannose. For FlP, the periodate oxidation yielded glycerol and glucose, which suggested (1-6) (42.5%) and (1-3) (40%) linkages. Other examples of characterization by the TM are given through the study of the F1S fractions of M. gypseum and T. ajelloi. The TG-DTG curves have shown for both fractions a common maximum at 300 "C, located in the lower decomposition region of the mannans from C. albicans (313-296 "C), near the peak exhibited by chitosan (305 "C) and also neighboring those (1-U-linked 8-glucogalactans from P. allahabadense and T. helicus. For this reason, the presence of Mana(1-6) and Glc/3(1-4) residues in such fractions seems established. The comparative study by chemical analysis of the above fractions indicates that these are glucomannans with (1-6) and (1-4) linkages [predominantly (1-6) linked]. As the TM is considered to be useful in distinguishing among several species of a same genus, it seemed worth-

while to report the characterization of three species of Aphanoascus in a comparative study with chemical methods. The changes in the content of mannose and glucose in the F1S fractions of the cell wall from Aphanoascus species allow differentiation among A. fuluescens, A. reticulisporus, and A. terreus. Also, the GLC analysis of the alditol acetates after Smith degradation has shown for these species very different ratios of (1-6) to (1-4) linkages: 71.4/11.5 (A. reticulisporus), 50.3/ 34.6 (A.terreus),and 34.6/53.2 (A.fuluescens). The DTG curve analyses are suitable to emphasize the earlier differences, because the maxima for A. reticulosporus and A . terreus (314 and 311 "C) appear in Table I in adifferent region than those of A. fuluescens (319 "C). The temperatures of A. reticulisporus and A. terreus are contained, on the one hand, between those of pustulans and pustulanoids, and on the other hand, between those of mannans and glucan-chitin (around 310 "C), suggesting the coexistence of the 8(1-6) linkages of the former with the a(1-6) and 8(1-4) linkages of the latter. The closer proximity of the maximum of A. reticulisporus to those of pustulans indicates a higher content of P(1-6) linkages, whereas for A . terreus, both (1-6) and (1-4) linkage types are present equally. In the case of A. fuluescens, the DTG peak at 319 OC is very near those of amylose (319.5 "C) and U.pustulata (318 "C), suggesting a higher content of glucans with (1-4) and (1-6) linkages. [However, with respect to mannosyl residues, the linkage type is indeterminate, possibly (1-4) linked.] A last example of characterization on the basis of the thermal data is given as follows. For fractions F1 from T . flauus, the DTG maximum takes place at 291 "C, in the galactoderivatives region, between the galactomannoglucans of G. uiride (297.5 "C) and the glucogalactans of P. allahabadense and T. helicus (301-298 "C) at one extreme and the galactans of P. spinulosum and E . stolkiae (284278 "C) a t the other end. Given that the polysaccharides of G. uiride, P. allahabadense, and T . helicus show 8(1-4),(1-6) linkages and those of P. spinulosum and E. stolkiae 8(1-4),(1-+3) linkages, it seems reasonable to deduce the presence of the three linkage types for T. flauus. Partial characterization by classical methods has revealed that the monosaccharides released by 4 N H2S04 hydrolysis of fraction F1 from T. flauus are galactose

Blotechnol. hog., 1991, Vol. 7, No. 6

(30%), glucose (41%), and mannose (9%); the mole percentages of the sugars detected by GLC after Smith degradation of the same fraction are indicative of the presence of various linkage types, (1-4), (1-.6),and (1+3), predominantly (1-6) and (1-4); and the IR absorption band at 890 cm-l is characteristic of @-linked polysaccharides. As shown, these results are in agreement with the above statements. Critical Considerations. We are convinced of the suitability of the new method but not of its absoluteness, all the more because further detailed work is necessary on aspects such as linkage multiplicity, branching, heterogeneity, or polydispersity. (a) Whereas the coexistence of two linkage types in the same polysaccharide is evidenced without difficulty, the occurrence of a third linkage type can only be detected in some cases. (b) No definitive conclusions can be drawn with regard to branching of the polymer chains, as adequate reference compounds were not available. (c) Studies on heterogeneity have not yet been completed for bacterial lipopolysaccharides (for fungal polysaccharides the influence of the heterogeneity of the isolates is little, around 4 "C). (d) In LPS, when the presence of various constituents of different lengths of the 0-chain undesirably widen the thermal effects (the temperature of the maxima are unaffected!), additional purification becomes necessary. In comparison with 13CNMR-FT, our methodology may still fall somewhat short, but it has the advantage of a greater simplicity and rapidness in the interpretation of results. Against other valuable techniques such as microcalorimetry (in which the heat production rate, dQ/dt, is used as a measure of the metabolic activity; Kemp, 1990) and Curie-point pyrolysis mass spectrometry (in which the thermal splitting of macromolecules into building blocks as well as smaller fragments is performed in the vacuum of the mass spectrometer; Haverkamp, 1979,1980), the TG-DTG, DTA, and DSC techniques have the advantage that a mathematical treatment of the data is not required.

Conclusion Thermal analysis seems a valid method for rapid characterization of fungal polysaccharides and bacterial lipopolysaccharides. The method is suitable for identification as well as classification purposes. The resulting highly simplified data handling is of special interest as a screening tool in structural investigation. In addition, it can yield important information pertaining to relating stability-structure-microbiological activity, as both the antiviral and antitumor effects of polysaccharides containing (1-3) glycosidic linkages as well as the influence of the nature of the LPS on the immunobiological properties of bacterial species have been established (Perry & Bundle, 1990).

Acknowledgment We thank M. Shnchez-Valiente for valuable laboratory assistance.

Literature Cited Bartnicki-Garcia, S. Cell wall chemistry, morphogenesis and taxonomy of fungi. Annu. Rev. Microbiol. 1968, 22, 87-108. Bartnicki-Garcia, S. Cell wall composition and other biochemical markers in fungal phylogeny. In Phytochemical Phylogeny; Harborne, J. B., Ed.; Academic Press: New York and London, 1970; pp 81-103.

533 Cacabelos, R. Farmacoloda molecular de la terapia anti-SIDA. Jano. 1988,35,25-26. Chasseing, N. A.;Lederkremer, R. M.; Couto,A.; Mayer,A.; Rumi, L. S. Influencia de la oxidacidn con periodato e hidr6lisis dcida parcial respectivamente, en la accidn antitumoral del polisacdrido PCj3 extraido del hongo Cyttaria-johowii. Znmunologia 1988, 7, 138-142. Gdmez-Miranda, B.; RupBrez, P.; Leal, J. A. Changes in Chemical Composition During Germination of Botrytis cinerea Sclerotia. Curr. Microbiol. 1981, 6, 243-246. Gdmez-Miranda, B.; Guerrero, C.; Leal, J. A. Effect of Culture Age on Cell Wall Polysaccharides of Penicillium allahabadense. Exp. Mycol. 1984,8, 298-303. Gdmez-Miranda, B.; Moya, A.; Leal, J. A. Differences in the Cell Wall Composition in the Type Species of Eupenicillium and Talaromyces. Exp. Mycol. 1988,12, 258-263. Haverkamp, J.; Eshuis, W.; Boerboom, A. J. H.; GuinBe, P. A. M. Pyrolysis Mass Spectrometry as a rapid screening method of biological materials. In Advances in Mass Spectrometry; Quayle, A., Ed.; Heyden and Son Ltd.: London, 1979;pp 983988. Haverkamp, J.;Meuzelaar, H. L. C.; Beuvery, E. C.; Boonekamp, P. M.; Tiesjema, R. H. Characterization of Neisseria meningitidis capsular polysaccharides containing sialic acid by pyrolysis mass spectrometry. Anal. Biochem. 1980,104,407418. Kemp, R. B. In Biological Calorimetry; Wendlandt, W. W., Ed.; Thermochim. Acta 1990,172. Mahadevan, P. R.; Tatum, E. L. Relationship of the major Constituents of the Neurospora crassa cell wall to wild-type and colonial morphology. J . Bacteriol. 1965,90, 1073-1081. Pazur, J. H. Neutral polysaccharides. In Carbohydrate Chemistry: a practical approach; Chaplin, M. F., Kennedy, J. F., Eds.; IRL Press: Oxford, England, and Washington, DC, 1986, p 79. Perry, M. B.; Bundle, D. R. Lipopolysaccharide Antigens and Carbohydrates of Brucella. In Advances in Brucellosis Research; Adams, L. G., Ed.; Texas A&M University Press: College Station, TX, 1990; pp 77-88. Prieto, A.; RupBrez, P.; Hemhdez-Barranco, A.; Leal, J. A. Partial characterization of galactofuranose-containingheteropolysaccharides from the cell walls of Talaromyces helicus. Carbohydr. Res. 1988, 177, 265-272. Ramos-Shchez, M. C.; Martin-Gil, F. J.; MarthGil, J. Estudios termoanaliticos de bacterias y SUB componentes estructurales. I11 Reuni6n del Grupo de Taxonomla Bacteriana de la SEM, Madrid, Spain, September 22-23, 1988. Swanson, M. A.; Cori, C. F. Acid stability of a-Iinked-D-Hexoglycans. J . Biol. Chem. 1948, 172, 797. Whistler, R. L.; Bushway, A. A.; Singh, P. P.; Nakahara, W.; Tokuzen, R. Noncytotoxic, antitumor polysaccharide. Adv. Carbohydr. Chem. Biochem. 1976,32, 235-275. Wilkinson, S. G. Composition of Bacterial Polysaccharides. In Surface Carbohydrates of the Prokaryotic Cell; Sutherland, I. W., Ed.; Academic Press: London and New York, 1977; pp 117-155. Accepted August 14, 1991.

Registry No. Chitin, 1398-61-4; laminaran, 9008-22-4; cellulose, 9004-34-6; S-glucan, 9051-95-0; mannan, 9036-88-8; dextran, 9004-54-0; nigeran, 31799-84-5; pullulan, 9057-02-7; amylopectin, 9037-22-3; amylose, 9005-82-7; glucomannan, 1107831-2; R-glucan, 53238-80-5;pustulan, 37331-28-5;glucomannan, 11078-31-2; chitosan, 9012-76-4; galactomannoglucan, 9040-293; glucogalactan, 39301-05-8; galactan, 9037-55-2; xylan, 901463-5;xanthan gum, 11138-66-2;inulh, 9005-80-5;fucoidan, 907219-9; glucosamine, 3416-24-8; acetylglucosamine, 7512-17-6; galactosamine, 7535-00-4;mannosamine, 14307-02-9;glucuronic acid, 6556-12-3.