OCCURRENCE AND FATE OF CARBOHYDRATES IN RECENT AND ANCIENT SEDIMENTS FROM DIFFERENT ENVIRONMENTS OF DEPOSITION

^ ( l el-^

!

?'V

OCCURRENCE AND FATE OF CARBOHYDRATES IN RECENT AND ANCIENT SEDIMENTS FROM DIFFERENT ENVIRONMENTS OF DEPOSITION

OCCURRENCE AND FATE OF CARBOHYDRATES IN RECENT AND ANCIENT SEDIMENTS FROM DIFFERENT ENVIRONMENTS OF DEPOSITION PROEFSCHRIFT ter verkrijging van de graad van doctor aan de Technische Universiteit Delft op gezag van de Rector Magnificus, Prof. drs. P.A. Schenck in het openbaar te verdedigen ten overstaan van een commissie door het College van Dekanen daartoe aangewezen op donderdag 2 maart 1989 te 16.00 uur door

Maria Elisabeth Catharina Moers

geboren te Zoetermeer doctorandus in de chemie en in de geologie

f TR diss 1702

Dit proefschrift is goedgekeurd door de promotor Prof. drs. P.A. Schenck Toegevoegd promotor Dr. J.W. de Leeuw

The author gratefully acknowledges the Koninklijke/Shell Exploratie en Produktie Laboratorium (Shell Research BV), Rijswijk, The Netherlands, for financial support of the investigations described in this thesis (agreement no 59643/Mkp/4).

voor mijn ouders voor Ron voor Debru

cover design: Jan Moers

CONTENTS

SUMMARY SAMENVATTING

page 13 17

1. INTRODUCTION 1.1 Natural occurrence of carbohydrates Occurrence of saccharide containing polymers Monosaccharides and their occurrence in polymers Organisms contributing to the organic matter in various depositional environments 1.2 Diagenesis of organic matter and of carbohydrates in particular 1.3 Objective and framework of the thesis Objective Methods Samples

21

25 27

2. OCCURRENCE AND ORIGIN OF CARBOHYDRATES IN PEAT SAMPLES FROM A MANGROVE ENVIRONMENT AS REFLECTED BY ABUNDANCES OF NEUTRAL SACCHARIDES 2.1 Abstract 2.2 Introduction 2.3 Experimental 2.4 Results 2.5 Discussion 2.6 Conclusions

31 32 34 36 43 47

3. CARBOHYDRATE SPECIATION AND PY-MS MAPPING OF PEAT SAMPLES FROM A SUBTROPICAL OPEN MARSH ENVIRONMENT 3.1 Abstract 3.2 Introduction 3.3 Experimental Analysis of carbohydrates as alditol acetates Analysis by Curie point pyrolysis-mass spectrometry 3.4 Results and Discussion Analysis of alditol acetates Py-MS mapping General discussion 3.5 Conclusions

49 50 51

54

67

4. CHARACTERIZATION OF TOTAL ORGANIC MATTER AND CARBOHYDRATES IN PEAT SAMPLES FROM A CYPRESS SWAMP BY PYROLYSIS-MASS SPECTROMETRY AND WET-CHEMICAL METHODS 4.1 Abstract 4.2 Introduction 4.3 Experimental Analysis of alditol acetates Analysis by pyrolysis-mass spectrometry Multivariate techniques 4.4 Results and Discussion

69 70 71

73

Analysis of carbohydrates as alditol acetates Comparison of sugar data with literature data Py-MS mapping Additional observations 4.5 Conclusions

87

5. ANALYSIS OF NEUTRAL SACCHARIDES IN FEN PEAT AND BOG PEAT SAMPLES FROM THE ASSENDELVER POLDERS (THE NETHERLANDS) 5.1 Abstract 5.2 Introduction 5.3 Experimental 5.4 Results and Discussion Total sugar yields Sugar yields in relation to peat type Factor analysis of minor sugars Special cases 5.5 Conclusions

89 90 91 94

105

6. ANALYSIS OF NEUTRAL SACCHARIDES IN MARINE SEDIMENTS FROM THE EQUATORIAL EASTERN ATLANTIC (KANE GAP) 6.1 Abstract 6.2 Introduction 6.3 Experimental 6.4 Results and Discussion 6.5 Conclusions

107 108 108 111

7. ANALYSIS OF CARBOHYDRATES IN QUATERNARY SAPROPEL SAMPLES FROM THE EASTERN MEDITERRANEAN 7.1 Abstract 7.2 Introduction 7.3 Experimental 7.4 Results and Discussion Sugar yields Sugar-sugar correlations Sugar-lipid correlations Sugar-stable carbon isotope correlations Sugar-pollen correlations Comparison with other sediments

119 120 121 122

7.5 Conclusions

132

8. ORIGIN AND DIAGENESIS OF CARBOHYDRATES IN ANCIENT SEDIMENTS 8.1 Abstract 8.2 Introduction 8.3 Samples 8.4 Experimental 8.5 Results and Discussion Mahakam Delta Brandon Lignite Messel Shale DSDP-samples Ancient marine (oil) shales Comparison of 'terrestrial' and 'marine' samples

135 136 137 141 142

9. INTERACTION OF GLUCOSE AND CELLULOSE WITH HYDROGEN SULPHIDE AND POLYSULPHIDES 9.1 Abstract 9.2 Introduction 9.3 Methods Simulation experiments Analysis of Standard Control experiments 9.4 Results 9.5 Discussion and Conclusions

APPENDICES REFERENCES DANKWOORD CURRICULUM VITAE

159 160 161

162 164

13 SUMMARY

The objective of the present study was to determine the contents of saccha­ rides in sediments from various depositional environments, to investigate to which diagenetical stage more or less intact saccharides can be encountered and to which extent they contribute to the organic carbon content. To this end neutral saccharides were determined in sediments from different depositional environments, geological ages and maturity. The carbohydrates were hydrolyzed by sulphuric acid. The neutral monosaccharides formed were converted to alditol acetates by reduction of the aldoses to alditols foliowed by acetylation of the alditols to alditol acetates. Gas chromatography and gas chromatography-mass spectrometry were used as analytical techniques to identify the alditol acetates and to quantify their yields. The occurrence of neutral saccharides in peat samples is described in chapters 2, 3, 4 and 5. In chapters 3 and 4 the results of the characterization of the total organic matter by Curie point pyrolysis-mass spectrometry are given as well. The occurrence of saccharides in recent marine sediments is described in chapters 6 and 7. Chapter 8 deals with the occurrence of neutral saccharides in ancient sediments from various depositional environments and with different diagenetic histories. Results of an investigation on the interaction of glucose and cellulose with sulphur compounds are comprised in chapter 9. In the recent sediment samples with ages up to ca. 850,000 y. B.P. (chapters 2 through 7) about 40 to 50 different neutral saccharides could be identified. The monosaccharides concerned are hexoses (glucose, galactose, mannose, allose, altrose, idose), pentoses (xylose, arabinose, ribose), 6-deoxyhexoses (rhamnose, fucose), tetroses (erythrose, threose), glycerol, several heptoses and amino sugars, and partially methylated derivatives of the heptoses, hexoses, pentoses and 6-deoxy-hexoses mentioned. The fraction of the total organic carbon that can be ascribed to carbohydrate carbon ranges from ca. 1 to 15 wt% in the recent sediments investigated. This is a very conservative estimate of the amounts of carbohydrate present.

14 The distribution patterns of monosaccharides in recent sediments can be used to differentiate between different kinds of input of organic material. A major input of vascular plants can be deduced from relatively high glucose contributions, i.e. higher than ca. 40 wt% of the total sugar yield and from the total yields of saccharides. A major input of microorganisms to the organic matter is assumed when the relative contributions of partially methylated saccharides, glycerol, tetroses, allose, altrose, heptoses and amino sugars -collectively called 'minor sugars'- are together higher than 5 to 10 wt% of the total sugar yield. Variations in the contributions of different types of 'higher' plants to peat sediments are in general reflected by variations in the relative contributions of glucose, xylose, arabinose and sometimes rhamnose. This can be explained by variations in the composition of hemicelluloses and pectins. The compositions of these polymers are often characteristic for certain groups of plants. Variations in the contributions of different types of microorganisms on the other hand are better reflected by variations in the contributions of the partially methylated saccharides, heptoses, amino sugars, tetroses, allose and sometimes altrose and glycerol. This is probably due to the specificity of compositions of cell walls of the various bacteria, algae, protozoa and fungi. The absence of overall decreasing trends in total saccharide yields with increasing depth in the cases of the recent sediments studied indicates that no serious degradation of carbohydrates is taking place in recent immature sediments during the first million years or so after deposition. Most carbohydrate degradation appears to be limited to the upper cm's -if not mm'sof the sediment. The data in chapters 2, 3 and 4 are derived from fractionated peat samples. The fine grained samples show rather low total saccharide yields and the saccharides are for the major part derived from microorganisms. The coarse grained samples show comparatively high total saccharide concentrations and the saccharides are for the major part derived from vascular plants. Multivariate analysis of the minor sugar data suggests that the remains of different microbial populations are present at different levels in the peats. Interpretation of these minor sugar data is seriously hampered by a general lack of knowledge of specific compositions in terms of partially methylated saccharides of the cell walls and membranes of the various microorganisms.

15 Characterization of fractionated peat samples by Curie point pyrolysis-mass spectrometry (Py-MS) (chapters 3 and 4) and factor-discriminant analysis of the Py-MS data reveal that the coarse grained samples show most of the markers for lignins and carbohydrates, which pointe to the presence of vascular plant material. A relative accumulation of lignified vascular plant material is observed with increasing depth. The fine grained samples show markers indicative for the presence of microorganisms and refractory matter. Chapter 5 shows the results of analysis of neutral saccharides in fen and bog peats. Eutrophic fen peats and oligotrophic bog peats can be distinguished on basis of total saccharide yields and the relative contributions (in wt% of the total saccharide yield) of glucose, xylose, arabinose and rhamnose. Factor analysis of the 'minor sugar' data shows that some minor sugars are associated with the presence of fungi living in symbiosis with heather rootlets. Another set of minor sugars is correlated with peats rich in cotton-grass or reed. In chapters 6 and 7 the occurrence of neutral saccharides in marine sediments is described. The carbohydrates are for the major part derived from marine microorganisms. Comparison of saccharide data with lipid data show that the results support each other in general. Increased primary production, presumably of cyanobacteria and/or diatoms, is correlated with increased contributions of specific mono-O-methyl-saccharides. Chapter 8 shows that intact saccharides can be identified and quantified in ancient sediments, but increase in temperature has a pronounced effect on the quantity and quality of the saccharides present. In marine and lacustrine oil shales and in humic coals less than 0.1 wt% of the total organic carbon can be attributed to carbohydrate carbon. Glucose and mannose are selectively preserved with respect to the other monosaccharides, irrespective as to whether the original organic matter was predominantly 'marine' or 'terrestrial'. Data on saccharides from Tertiary DSDP samples that have not been subjected to increased temperatures do not reveal any symptoms of diagenesis. Simulation experiments in which glucose and cellulose were reacted with hydrogen sulphide and polysulphides (chapter 9) reveal that organic sulphur containing compounds are formed that yield several thiophenes upon

16 pyrolysis/evaporation. It appears that interaction of carbohydrates with sulphur compounds is a possible way for carbohydrates to react in very recent sediments; so the resulting products become less susceptible to microbial attack.

17

SAMENVATTING

Het onderzoek waarvan de resultaten in dit proefschrift zijn beschreven, had tot doel om het voorkomen van sacchariden in sedimenten van verschillende herkomst te bepalen, om na te gaan tot welk diagenetisch stadium er nog min of meer intacte sacchariden in sedimenten kunnen worden aangetoond en om te bepalen in welke mate deze bijdragen tot het organisch koolstofgehalte. Ter realisatie van deze doelstelling werden sacchariden bepaald in een aantal sedimenten afkomstig uit verschillende afzettingsmilieus en van veschillende geologische ouderdom en rijpheid. Om de sacchariden te analyseren werden zij allereerst met zwavelzuur gehydrolyseerd. De hierbij gevormde neutrale monosacchariden werden omgezet in alditol-acetaten door reductie van de aldoses tot alditolen, gevolgd door acetylering van de alditolen tot alditol-acetaten. Gaschromatografie en gaschromatografie-massaspectrometrie werden gebruikt als analysemethoden om de alditol-acetaten te identificeren en hun opbrengsten te kwantificeren. Het voorkomen van neutrale sacchariden in veenmonsters wordt beschreven in de hoofdstukken 2, 3, 4 en 5. In de hoofdstukken 3 en 4 worden tevens de resultaten van de karakterisering van het totale organische materiaal met Curie punt pyrolyse-massaspectrometrie (Py-MS) gegeven. Het voorkomen van sacchariden in recente mariene sedimenten wordt uiteengezet in de hoofdstukken 6 en 7. Hoofdstuk 8 heeft als onderwerp het voorkomen van neutrale sacchariden in oude sedimenten afkomstig van verschillende afzettingsmilieus en met verschillende diagenese achtergronden. De resultaten van een onderzoek naar de interactie van glucose en cellulose met zwavelverbindingen zijn samengevat in hoofdstuk 9. In monsters van recente sedimenten met leeftijden tot ca. 850.000 jaar (hoofdstukken 2 tot en met 7) konden ongeveer 40 tot 50 verschillende neutrale sacchariden worden geidentificeerd. De betreffende monosacchariden zijn hexoses (glucose, galactose, mannose, allose, altrose, idose), pentoses (xylose, arabinose, ribose), 6-deoxy-hexoses (rhamnose, fucose), tetroses (erythrose, threose), glycerol, verscheidene heptoses en aminosuikers, en partieel gemethyleerde derivaten van de bovengenoemde heptoses, hexoses, pentoses en 6-deoxy-hexoses. De fractie van de totale hoeveelheid organisch

18 koolstof die kan worden toeschreven aan koolhydraatkoolstof ligt tussen de 1 en 15 gewichtsprocenten in de onderzochte recente sedimenten. Dit zijn voorzichtige ramingen van de totale hoeveelheden koohydraten en koolhydraat­ koolstof die in sedimenten aanwezig zijn. De verdelingspatronen van monosacchariden in recente sedimenten kunnen worden gebruikt om de bijdragen van verschillende soorten organisch materiaal te onderscheiden. Een belangrijke aanvoer van vaatplanten in het sediment kan worden afgeleid uit relatief hoge bijdragen van glucose, d.w.z. hoger dan ca. 40 gew%, en aan de opbrengst van de total sacchariden. Een hoofdbijdrage van microorganismen aan het organisch materiaal wordt verondersteld wanneer de gesommeerde relatieve bijdragen van partieel gemethyleerde sacchariden, glycerol, tetroses, allose, altrose, heptoses en aminosuikers -gezamenlijk 'minor sugars' genoemd- hoger zijn dan 5 è 10 gew% van de total saccharide opbrengst. Variaties in de bijdragen van verschillende typen 'hogere' planten aan venen worden meestal weerspiegeld in variaties in de relatieve bijdragen van glucose, xylose, arabinose en soms rhamnose. Dit kan worden verklaard als een gevolg van variaties in de samenstelling van hemicelluloses en pectinen. De samenstelling van deze biopolymeren is namelijk vaak karakteristiek voor bepaalde groepen planten. Variaties in de bijdragen van verschillende typen microorganismen uiten zich echter in variaties in de bijdragen van partieel gemethyleerde sacchariden, heptoses, aminosuikers, tetroses, allose en soms altrose en glycerol. Dit is waarschijnlijk een gevolg van de specificiteit van de samenstelling van de celwanden en celmembranen van de verschillende bacteria, algen, protozoae en schimmels. De afwezigheid van een globale dalende trend aan saccharide opbrengsten in hun totaliteit met toenemende diepte in de gevallen van de bestudeerde recente sedimenten, is een aanwijzing dat in recente, onrijpe sedimenten geen belangrijke afbraak van koolhydraten plaatsvindt gedurende de eerste millioen jaar na afzetting. De sterkste afbraak van koolhydraten lijkt te zijn beperkt tot de bovenste centimeters -misschien zelfs millimeters- van het sediment. De saccharide data zoals beschreven in de hoofdstukken 2, 3 en 4 zijn het resultaat van bepalingen aan gefractioneerde venen. De fijnkorrelige monsters vertonen tamelijk lage opbrengsten aan sacchariden en deze sacchariden zijn voor het grootste gedeelte afkomstig van microorganismen. De grofkorrelige monsters vertonen relatief hoge concentraties aan sacchariden die voor het

19 grootste gedeelte afkomstig zijn van vaatplanten. Multivariantie analyse van de 'minor sugar' resultaten laat zien dat overblijfselen van verschillende microbiele populaties aanwezig zijn op verschillende niveaus in de venen. Interpretatie van deze 'minor sugar" gegevens wordt ernstig bemoeilijkt door een algemeen gebrek aan kennis van de specifieke samenstellingen in termen van partieel gemethyleerde suikers, van celwanden en celmembranen van de verschillende microorganismen. Karakterisering van gefractioneerde veenmonsters met Curie punt pyrolysemassaspectrometrie (hoofdstukken 3 en 4) and factor-discriminant analyse van de Py-MS gegevens laten zien dat de grofkorrelige monsters de meeste markers voor ligninen en koolhydraten bevatten. Dit wijst op de aanwezigheid van materiaal afkomstig van vaatplanten. Met toenemende diepte wordt het organisch materiaal relatief rijker aan gelignificeerd materiaal afkomstig van vaatplanten. De fijnkorrelige monsters bevatten 'markers' die kenmerkend zijn voor de aanwezigheid van microorganismen en van refractair materiaal. De resultaten van de analyse van neutrale sacchariden in laagvenen en hoogvenen worden behandeld in hoofdstuk 5. Eutrofe laagvenen en oligotrofe hoogvenen kunnen van elkaar worden onderscheiden op basis van de totale opbrengsten aan sacchariden en de relatieve bijdragen van glucose, xylose, arabinose en rhamnose daaraan. Factor analyse van de 'minor sugar' resultaten laat zien dat de aanwezigheid van sommige 'minor sugars' is geassocieerd met de aanwezigheid van schimmels die in symbiose leven met heidewortels. Een andere groep 'minor sugars' is gecorreleerd met venen die rijk zijn aan pijpestrootje of riet. In de hoofdstukken 6 en 7 wordt het voorkomen van neutrale sacchariden in mariene sedimenten beschreven. De koolhydraten zijn voor het grootste gedeelte afkomstig van mariene microorganismen. Toegenomen primaire productie, naar verwachting van cyanobacteriën en/of diatomeën, is gecorreleerd met toegenomen bijdragen van enkele specifieke mono-O-methyl-sacchariden. Ook in oude sedimenten kunnen intacte koolhydraten worden geïdentificeerd en gekwantificeerd (Hoofdstuk 8). Toename van de temperatuur waaraan de sedimenten zijn blootgesteld tijdens diagenese, heeft een uitgesproken effect op de kwantiteit en kwaliteit van de aanwezige sacchariden: Minder dan 0,1 gewichtsprocent van de totale hoeveelheid organisch koolstof kan worden

20

toegeschreven aan koolhydraatkoolstof in mariene en lacustriene olie schalies en in humuskolen. Glucose en mannose worden selectief gepreserveerd in vergelijking met de andere sacchariden. Dit is onafhankelijk van een oorspronkelijk mariene of terrestrische samenstelling van het organische materiaal. Saccharide resultaten van Tertiaire DSDP monsters die niet aan verhoogde temperaturen blootgesteld zijn geweest, vertonen geen kenmerken van diagenese. Met simulatie-experimenten waarin glucose en cellulose werden behandeld met waterstofsulfide en polysulfides (hoofdstuk 9) is aangetoond dat organische zwavelbevattende verbindingen worden gevormd die tijdens pyrolyse verschillende thiofenen opleveren. Het blijkt dat interactie van koolhydraten met zwavelverbindingen een mogelijke manier is waarop koolhydraten kunnen reageren in recent sedimenten; op deze wijze ontstaan verbindingen die minder gevoelig zijn voor microbiele afbraak.

21

1. INTRODUCTION

1.1 NATURAL OCCURRENCE OF CARBOHYDRATES Occurrence of saccharide containing polvmers Carbohydrates are important and omnipresent constituents of the living biomass. They are. present as cell constituents in all organisms from protists to vascular plants and mammals. In bacteria they comprise 20 to 40 percent of the dry weight and in vascular plants ca. 75 wt% of the biomass is attributable to carbohydrates (e.g. Kirk, 1973; Sjöström, 1981; Aspinall, 1983). In living organisms carbohydrates perform various different functions; some are metabolically active and are used for energy storage (e.g. starch, inulin, glycogen, laminaran, mannitol). Other saccharides are associated with cell walls and membranes where they may provide protection, stability and strength. This concerns polymers like cellulose, hemicelluloses, pectins, chitin, agar, alginic acid, peptidoglycans, lipopolysaccharides, teichoic acids, glycoproteins and glycolipids. DNA and RNA constitute still another category of biopolymers which contain saccharide moieties. The kind of polymers present is often specific for a certain group of organisms, for example: Cellulose, hemicelluloses, pectin and tannin are widespread in 'higher' plants. Peptidoglycans are important constituents of bacterial cell walls. Lipopolysaccharides are typical constituents of the cell membranes of Gram-negative bacteria and cyanobacteria. Teichoic acids are present in cell membranes of Gram-positive bacteria. Agar and alginic acid are structural polysaccharides in certain algae. Chitin is an important constituent of many fungal cell walls and of the exoskeletons of insects and crustaceans. Acid mucopolysaccharides, glycoproteins and glycolipids are of normal occurrence in cell coats of animal cells.

22

Monosaccharides and their occurrence in polymers The sacchande monomers contained in the polymers mentioned above vary widely both quantitatively and qualitatively. The monomers present may be specific for certain kinds of polymers and hence for certain groups of organisms or may be even specific for a certain species. Fig. 1.1 shows structures of monosaccharides that occur frequently in organisms. Glucose is a very abundant, omnipresent monomer and occurs for instance in starch, glycogen, cellulose and hemicellulose. Other monomers frequently encountered in various polymers are galactose, mannose, xylose, arabinose, ribose, fucose, rhamnose and 2-deoxy-ribose. Partially methylated derivatives of the monosaccharides mentioned above are reported to occur in cell membranes of various microorganisms, for instance in their lipopolysaccharides (e.g. Laskin and Lechevalier, 1982; Aspinall, 1983). Many of them are thought to be species specific (Weckesser et al., 1979). Apart trom neutral sugars, a number of sugar alcohols (e.g. glycerol, ribitol, mannitol and inositol), sugar acids (uronic acids, aldonic acids, aldaric acid) and amino sugars (e.g. glucosamine and galactosamine) occur in nature. The monomers may occur in homopolysaccharides like mannans, xylans, galactans etc, but they may also occur in heteropolysaccharides like hemicelluloses and acid mucopolysaccharides or they may form polymers with non-sugar constituents like proteins or lipids (e.g. steroids) and form for example glycolipids or glycoproteins. For instance, the major hemicelluloses are polymers of xylose, mannose, galactose, glucose, arabinose and 4-O-methyl-glucuronic acid. Acid muco­ polysaccharides in the cell coats of 'higher' animals usually contain two types of alternating monosaccharide units, often a uronic acid and an amino sugar. Teichoic acids contain either glycerol or ribitol in their backbones. Partially methylated galacturonic acid forms the backbone of pectin and mannuronic acid is an important constituent of alginic acid. The backbone of the rigid peptidoglycan framework of bacterial cell walls consists of Nacetyl-glucosamine and muramic acid. Chitin is also built up of N-acetylglucosamine units.

23

C no. HCO 1 2 HCOH 3 HOCH 4 HCOH 5 HCOH I 6 H2COH glucose

HCO HCOCH3 HOCH HCOH HCOH H2COH

2-0-methylglucose

HCO HCOH HOCH HCOH HCOH COOH glucuronic acid

HCO HCOH HOCH HOCH HCOH I H2COH

HCO HCOH HCOH HCOH

HCO HCOH HOCH HCOH

H2COH

H2COH

galactose

ribose

xylose

HCO HCOH HCOH HCOH HOCH CH3

H2C0H

H2COH

HCOH

OH

HèoH

H,6OH

fucose

ribitol

HCO HCNH2 HOCH HOCH HCOH H,COH galactosamine

HCOH H,COH glycerol

HCO HOOC HC-NH-C(0)-CH, HC-O-CH H3C HCOH HCOH HoC0H N—acetylmuramic acid

Fig. 1.1 Structures of monosaccharides occurring frequently in organisms.

24

Organisms contributing to the organic matter in various depositional environments It was mentioned above that the various saccharide containing polymers and their monomeric building blocks may be specific for various groups of organisms. For investigations of this kind it is necessary to know which organisms occur in which environments. Depositional environments important in the present context are swamps, marshes, lakes, seas and oceans. In the open marine environment planktonic organisms can be considered as the major primary producers of organic material and consequently of carbohydrates (e.g. Degens and Mopper, 1976). Planktonic organisms concerned are protozoa (e.g. foraminifera, silicoflagellates), algae (e.g. diatoms, coccolithophores, dinoflagellates, radiolaria) and cyanobacteria. During transport of the remains to the sea floor and incorporation into the sediment, carbohydrates are partly degraded due to bacterial and benthic activity. Hence, bacteria and benthic organisms can be regarded as an additional source of carbohydrates in the sediment. The contribution of terrestrially derived organic matter is in general rather small. In lacustrine environments fresh water counterparts of the marine organisms mentioned above may cause an important contribution to the organic material. Depending on the specific circumstances, remains of vascular plants may also contribute to the organic material in the sediment. In terrestrial environments like swamps and marshes, remains of land plants form the major source of carbohydrates. Additional contributions may be derived from algae or cyanobacteria, depending on the specific ecological conditions. Degradation of the primary organic material due to fungal and bacterial activity results in contributions of carbohydrates derived from these organisms to peats as well.

25

13, DIAGENESIS OF ORGANIC MATTER AND OF CARBOHYDRATES IN PARTICULAR After death of an organism (and in some cases during senescence), its molecular structures are for the greater part degraded, either by autolyse, or by grazing and predation, or by the activity of bacteria and fungi, or abiotically. Metabolic activity results in the formation of new biopolymers which in their turn are also prone to degradation. The result is that the more labile organic components participate in cycles of synthesis-degradationresynfhesis as part of the foodchain in the biosphere. A small proportion, generally less than 1%, consisting of the more recalcitrant organic components accumulates in sediments and participates in cycles associated with the formation of kerogen, oil and coals in the geosphere (e.g. Tissot and Welte, 1984). It appears that compounds with a protective function in a living organism are often rather recalcitrant after death of the organism as well. Compounds with a metabolic function, such as storage saccharides, on the other hand are often destroyed during senescence or very shortly after death of the organism and do not contribute to the organic matter in sediments (e.g. Ittekkot et al, 1982; Klok et al., 1984b). Non-storage organic compounds may be preserved in the biosphere and geosphere long after the death of the organisms from which they originated. This recalcitrant behaviour of certain organic compounds is determined by several factors: 1) Some compounds -e.g. lignins and structural lipids- show recalcitrant behaviour and are selectively preserved in sediments as a result of their natural molecular composition, molecular configuration and degree of crystallinity in living organisms. In many cases recalcitrance appears to be a consequence of inaccessibility for enzymatic decomposition. In a living plant recalcitrance may be aimed at as a means of protection against invasions from outside or as protection against stress due to the environment. Examples are lignification, cutinization, suberinization of leave, stem and root. 2) Natural compounds that are labile by themselves may show recalcitrant behaviour due to covering by or incorporation into recalcitrant compounds in living organisms. For instance, carbohydrates become more recalcitrant when they are protected by lignin in lignocellulosic complexes (e.g. Boon et al., 1988), when they are incorporated into the highly aliphatic biopolymer that is

26 present in cuticulae (e.g. Nip et al., 1986a,b, 1987), when they are encapsulated in a mineral matrix (e.g. Liebezeit et al., 1983). 3) Microbially initiated (stereo)chemical transformations may also render a compound recalcitrant. For example the transformation of cholesterol into 50(H)-cholestanol (coprostanol) in the intestinal tract of mammals (Eyssen et al., 1973). This latter compound is less degradable than its precursor and it is used as a tracer for (polution by) sewage. It should be kept in mind that recalcitrance is a relative concept. The preservation of organic material in general and of carbohydrates in particular will also depend on a number of environmental factors, such as primary production, sedimentation rate and anoxicity of the water column and sediment. Increased primary production, higher sedimentation rates and lack of oxygen in the water column and sediment all favour enhanced preservation of organic material and hence of carbohydrates as well. Structural saccharides are often assumed to be relatively rapidly degraded upon burial in sediments, despite their protective function in the living organism. Hence their contribution to the organic matter in mature sediments and kerogen is often assumed to be negligible (e.g. Tissot and Welte, 1984). Nevertheless, literature data show that hydrolyzable saccharides may account for 2 to 20 wt% of the total organic carbon in recent sediments (e.g. Degens and Mopper, 1975, 1976; Mopper, 1977; Mopper et al, 1978; Ittekkot et al., 1982; Cowie and Hedges, 1984; Klok et al., 1984a,b,c; Liebezeit, 1986; Steinberg et al., 1987; Hamilton and Hedges, 1988). Total' lipids (i.e. free plus hydrolyzable by acid and base) on the other hand constitute generally less than 10 wt% of the total organic carbon (e.g. Boon, 1978; Hunt, 1979, Klok et al., 1984c). Acid extractable and identifiable saccharides have also been analysed in ancient sediments: Cretaceous black shales appeared to contain between 0.2 and 0.0003 g carbohydrate carbon per 100 g total organic carbon (Michaelis et al., 1986), and averages of ca. 1 and 3 ng sugar/g sediment could be extracted from Devonian shales and (Pre)cambrian rocks respectively (Swain and Rogers, 1966; Swain et al., 1970). These literature data indicate that it may be somewhat premature to dispatch the presence of carbohydrates in sediments as insignificant and negligible.

27 1.3 OBJECTTVE AND FRAMEWORK OF THE THESIS Objective The objective of the present study was to determine the contents of saccharides in sediments from varying depositional environments, to investigate to which diagenetical stage more or less intact sacchatides can be detected and to which extent they contribute to the organic carbon. To this end a variety of natural sediment samples from different depositional environments, geological ages and maturities was studied. Apart from this it was investigated whether carbohydrate carbon could be preserved in other ways. To this end simulation were carried out. Methods The methods used comprise hydrolysis by sulphuric acid of the carbohydrates in the sediment samples to their monomeric building blocks, reduction of the neutral monosaccharides released to alditols and derivatization of the alditols to alditol acetates. Identification and quantification of the alditol acetates was achieved by gas chromatography and by gas chromatography-mass spectrometry (chapters 2 through 8). Details on the procedures foliowed during the simulation experiments can be found in chapter 9. During acid hydrolysis the various glycosidic bonds behave differently and the various monosaccharides are degraded by acid at different rates (e.g. Cheshire and Mundie, 1966; Dutton, 1973; Mopper, 1977; Mopper et al., 1978; Moers, unpublished results). It is important to ensure that hydrolysis of the carbohydrates is as complete as possible, although this might cause partial decomposition of small amounts of the less stable monosaccharides. The results of quantitative determinations of the released monosaccharides are also influenced by the reduction and acetylation procedures (e.g. Albersheim et al., 1967; Torello et al, 1980). Therefore it is important to ensure maximum reproducibility during treatment of the samples in order to obtain results that can be mutually compared. The hydrolysis and derivatization conditions for the analysis of the sediment samples have therefore been kept as much identical as possible. The only exception are the experiments with the samples

28

from the Mediterranean (chapter 7). Control experiments have shown, however, that this will not lead to misinterpretation of the results. The sacchandes analysed by the method used in the present case comprise only the neutral sacchandes. This implicates that the carbohydrate yields reported constitute only a fraction of the total amount of carbohydrates and carbo­ hydrate carbon present in the sediments. Uronic acids for instance may be present in ca. equal quantities as the neutral monosaccharides (e.g. Steinberg et al., 1987). Moreover, in the present case sediment samples were treated with dilute hydrochloric acid and cold water for the removal of carbonates and other inorganic salts, prior to analysis for carbohydrates. This procedure inevitably leads to the removal of small amounts of soluble carbohydrates as well. The methods used for determining concentrations of identifiable saccharides do not take into account the fraction of carbohydrate carbon that is not immediately recognizable as such due to alterations by which carbohydrates moved out of the analytical windows. This implies that the concentration of carbohydrate carbon in sediments may be much higher than would be deduced from the data of identifiable saccharides. Samples Chapters 2, 3, 4 and 5 deal with the occurrence of neutral saccharides in various swamp and marsh systems. Chapters 2, 3 and 4 describe the occurrence of neutral saccharides in fractionated peat samples from cores procured from sub-tropical environments in the Florida Everglades and the Okefenokee Swamp, both in the southeastern USA. It was investigated how differences in input of vascular plant species and differences in input of microorganisms were reflected in the contributions of the individual monosaccharides, their partially methylated derivatives and the total sugar yields. The total organic matter in the peats was characterized by Curie point pyrolysis-mass spectrometry (Py-MS) and multivariate analysis of the pyrolysis data. Both saccharide and Py-MS data were examined for trends with increasing depth, c.q. age, of the peat layers. Chapter sequence sequence foliowed

5 covers the results of analysis of neutral saccharides in a peat from a temperate climate. The samples represent a terrestrialization during which the environment changed from eutrophic to oligotrophic, by deposition of oligotrophic raised bogs. It was studied how these

29 changes in environment and associated changes in the flora were reflected by the saccharide data. In chapters 6 and 7 the results of analysis of neutral saccharides in sediment samples from marine environments are presented. Chapter 6 is concerned with the results of saccharides from sediment samples ranging in age from ca. 9,000 to 845,000 y. B.P. from the equatorial eastern Atlantic Ocean and chapter 7 reports on the saccharide results from sediment samples ranging in age from ca. 6,500 to 200,000 y. B.P. from the eastern Mediterranean. The saccharide data were used to study relative inputs of marine versus terrestrial organic matter, to investigate variations in the marine microbial assemblages and to explore changes in saccharide abundances attributable to prolonged times of burial c.q. diagenesis. Chapter 8 deals more specifically with the effects of abiotic diagenesis on the contributions of saccharides to sediments. To this end neutral saccharides were determined in sediment samples ranging in age from Tertiary to Jurassic, from different depositional environments and with different diagenetic histories, namely Tertiary marine samples from DSDP-cores, a Miocene oil shale from the Monterey Formation (USA), Tertiary humic coals from the Mahakam Delta (Indonesia), an Oligocene lignitic wood sample from the Brandon Lignite (USA), Eocene oil shales from Messel (FRG), Cretaceous oil shales from Julia Creek (Australia) and from Jurf ed Darawish (Jordan) and Jurassic samples from the Paris Basin (France) and from the Posidonia shale (FRG). It was investigated how saccharide contents were affected by increasing maturity and age of the sediment and whether original sources of input of organic material could still be recognized. Chapter 9 deals with simulation experiments investigating the interaction of glucose and cellulose with hydrogen sulphide and polysulphides. In this way the feasibility was explored of carbohydrates escaping biomineralization by formation of organic sulphur containing compounds that are more resistant to microbial attack and hence have a greater potential for survival during diagenesis than their carbohydrate precursors.

31

2. OCCURRENCE AND ORIGIN OF CARBOHYDRATES IN PEAT SAMPLES FROM A MANGROVE ENVIRONMENT AS REFLECTED BY ABUNDANCES OF NEUTRAL SACCHARIDES

M.E.C. Moers, M. Baas, J.W de Leeuw, JJ. Boon and P.A. Schenck Submitted

2.1 ABSTRACT Acid hydrolysates of fractionated peat samples from Jewfish Key in the Florida Everglades were analysed for neutral saccharides. Two major sources of carbo­ hydrates could be determined: 1) vascular plant carbohydrates derived from Rhizophora mangle and 2) microbially derived carbohydrates. Significant correlations exist between the relative contributions of most sugars and the total carbohydrate concentration: Low total carbohydrate yields with high microbial contributions in the fine grained samples and high total carbo­ hydrate yields with high vascular plant contributions in the coarse grained samples. It is estimated that the greater part of the sugars analysed in the fine grained samples originates from microorganisms. The absence of a trend in total carbohydrate concentrations with depth suggests that microbial degradation is limited to the upper levels of the peat and that the microbial sugars determined at lower peat levels are derived from nonviable or dormant microorganisms. Results from factor analysis indicate differences in microbial populations in the various peat samples.

32

12INTRODUCTION Jewfish Key is located in the Florida Everglades at the mouth of the Shark River. The present vegetation on this small island consists of a forest of Rhizophora mangle (red mangrove) with occasional trees of Avicennia nitida (black mangrove) and Laguncularia racemosa (white mangrove). The area is under the influence of fresh water currents due to outflow firom the Shark River and of strong tidal action because of the proximity of the Gulf of Mexico. The peat at Jewfish Key is ca. 3.4 m thick and is underlain by a Pleistocene limestone (Miami oolite). The basal peat layer is dated at ca. 4000 y. B.P. (Spackman et al., 1966). The peats in the Florida Everglades have been deposited under the influence of a transgressing sea, so at several locations peat cores consist of upper layers deposited under saline water conditions while the deeper layers were deposited under progressively less saline (from brackish to fresh) water conditions (Spackman et al., 1966). Changes in water conditions and consequently in vegetation were also observed in the peat core under investigation; the upper four samples consist of material derived from Rhizophora mangle (red mangrove) and the peat layers at these depths have been deposited under saline water conditions. The lower two samples have been deposited under brackish to saline water conditions with Mariscus jamaicensis (sawgrass) as constituent in the sample from 245-254 cm. The deepest sample shows contributions of fresh Rhizophora rootlets. These are attributed to the presently living Rhizophora vegetation at the site (Spackman et al., 1981; Given et al., 1983). The core under investigation has been studied before by Rhoads (1985) and by Ryan (1985). They took samples from different depths in the core and fractionated these samples into coarse and fine grained fractions. The coarse grained fractions ( + 80 and +20 mesh) consisted of plant tissues and their fragments and the fine grained fraction (-80 mesh) consisted of amorphous matter (Ryan, 1985). Microscopic observations of the coarse grained fractions by Ryan (1985) showed that the samples derived from tissues and tissue fragments of Rhizophora were in different stages of degradation. The uppermost level (0-10 cm) con­ sisted of unaltered and slightly altered roots and rootlets. The sample from 64-74 cm consisted of remnants of roots and rootlets with most tissues largely

33

degraded. Many cell inclusions were present. The samples from 122-132 and 183188 cm showed small amounts of highly disrupted rootlets, no roots were observed. In the sample from 244-254 cm, derived from Mariscus, virtually no identifiable tissues or organs were noted. The peat at this level consisted predominantly of cell inclusions. A large amount of slightly disrupted rootlet tissue from Rhizophora embedded in a mineral rich matrix was observed in the sample from 305-310 cm. For explanation of the terminology used in the microscopic descriptions see Cohen and Spackman (1977,1980). Rhoads (1985) and Ryan (1985) analysed the peat samples by Fourier-transformed infrared spectroscopy (FT-IR) and by Curie point pyrolysis-mass spectrometry and Curie point pyrolysis-gas chromatography-mass spectrometry (Py-MS and PyGC-MS), respectively. These methods provide a general view of the occurrence of cellulose, hemicellulose and lignin in the peat samples. They noticed a strong decrease in the abundance of carbohydrates with increasing depth when compared to the abundance of lignin. This is in agreement with the generally held opinion that carbohydrates are labile compounds that are degraded rapidly during and after incorporation into sediments (e.g. Hatcher et al., 1981; Benner et al., 1984a,b; Given et al, 1984; Hedges et al., 1985). In contrast to former studies, the present study focusses in a quantitative fashion and in much more detail on the abundance of individual sugars in peats. This should give more insight into the molecular aspects and occurrence of carbohydrates than is possible with the more general wet chemical and spectroscopie methods used previously for the analysis of carbohydrates in peats. Jewfish Key is a very interesting site for studying the occurrence of carbo­ hydrates in peats, because there seems to be only one major source of vascular plant organic material to the peat at nearly all levels, namely Rhizophora mangle roots. This precludes the occurrence of differences in carbohydrate abundances due to differences in vascular plant input in the coarse grained fractions. The mangrove root system is so devised that fine grained detritus is easily trapped between the roots. This implies that the fine grained material in these peats may also contain contributions from trapped allochthonous material in addition to those from decayed mangrove tissues and the degrading organisms themselves.

34

2.3 EXPERIMENTAL Twelve samples, six coarse grained and six fine grained, taken from different depths in the peat (see Table 2.1) were analysed for neutral saccharides. Four samples were analysed in duplicate to obtain an indication of the reproducibility of the carbohydrate determinations. The samples had been fractionated by Rhoads (1985). For information on the fractionation procedures see Given et al. (1984) or Rhoads (1985). The results from analysis by Py-MS, Py-GC-MS and FT-IR showed that the two coarse grained fractions (+20 and +80 mesh) behaved very similarly (Rhoads, 1985; Ryan, 1985). Therefore in the present investigation only the +20 mesh and -80 mesh fractions were studied. The organic carbon content of the peat fractions on a dry weight basis is ca. 60 wt%, with a carboh/nitrogen ratio of ca. 22 (Rhoads, 1985). The samples of the fine grained fraction had been washed with 0.1 M HC1 for removal of carbonates and all samples had been Soxhlet extracted by benzene/ ethanol 2/1 v/v for 48 h at 100° C for removal of soluble lipids, phenols and pigments, also as part of the investigation by Rhoads (1985). About 6-7 % of the material on a dry weight basis appeared to be extractable by benzene/ ethanol (Rhoads, 1985). For the present study 100 mg of Soxhlet extracted, dry and pulverized peat was mixed with 5 mi's of 12 M H 2 S0 4 at room temperature for 2 h. The acid was diluted to 1 M and the polysaccharides were hydrolyzed for 4.5 h at 100°C. Myo-inositol was added as an internal Standard. The monosaccharides released by hydrolysis were reduced to alditols by NaBH4 (16 h, room temperature) and acetylated to alditol acetates by acetic anhydride/pyridine (3h, 100°C). The alditol acetates were analysed by gas chromatography (GC) on a Carlo Erba Fractovap 4160 gas chromatograph and by gas chromatography - mass spectrometry (GC-MS) on a Hewlett-Packard 5890 gas chromatograph coupled to a VG 70-250SE mass spectrometer. Gas chromatographic separations were performed on a CPsil88 fused silica capillary column (1=25 m, i.d.=0.32 mm, df=0.12 /im; Chrompack, Middelburg, The Netherlands). Samples were injected at 50°C (GCMS) or 70°C (GC). The temperature was rapidly raised to 150°C and from thereon further programmed at 3°C/min to 230°C and held at this tempera­ ture for 40 min. The mass spectrometer was operated in the electron impact mode at 70 eV, with a source temperature of 250°C.

35 The alditol acetates were identified on the basis of known relative retention times established by the analysis of Standard mixtures of partially methylated alditol acetates (Klok et al., 1982) and on the basis of mass spectra (Stoffel and Hanfland, 1973; Schwarzman and Jeanloz, 1974; Jansson et al., 1976; Radziejewsky-Lebrecht et al, 1979; Wong et al., 1980; Klok et al., 1982). In the case of two heptoses and two amino sugars no retention times were available. Therefore identification was based on mass spectra only. Small amounts of mono-O-methyl-heptoses were present in nearly all samples as were partially formylated alditol acetates. Quantification of the alditol acetates identified was achieved by peak area integration using a Maxima Chromatography Workstation (Dynamic Solutions Company, Ventura, CA, USA) coupled to the gas chromatograph. The responses of all alditol acetates were presumed to be equal on a weight basis, so that absolute quantifications became possible on the basis of the amount of internal Standard (myo-inositol) added. In some cases when peaks in the gas chromatogram overlapped, quantification was performed on GC-MS results by peak area integration of the responses of selected characteristic mass fragments with the aid of the software available with the mass spectrometer. The duplicate analyses indicate that the errors in the determinations of the individual major monosaccharides are 10% or less. The errors in the deter­ minations of the minor components are larger: up to 50% for components present in quantities smaller than 0.02 mg/g. Multivariate analysis was applied to normalized yields of 33 minor components -listed in Table 2.2- and to a set consisting of normalized yields of all sugars analysed plus the total carbohydrate yield. The data were subjected to factor analysis using a modified ARTHUR computer package (Infometrix, Seattle, WA, USA). The principles and the application of this procedure are described by Windig et al. (1982).

36 2.4 RESULTS Results of the analyses of neutral saccharides are listed in Table 2.1. From this table it can be seen that the samples in the coarse grained fraction yield systematically more carbohydrates upon hydrolysis than those from the fine grained fraction, with exception of the coarse grained sample from 244254 cm. It seems that the samples from neither coarse nor fine grained fraction show a significant decrease in total carbohydrate yields with increasing depth. The coarse grained sample from 0-10 cm yields, however, a markedly greater amount of neutral sugars than the other samples. The sugars collectively grouped as "minor components" in Table 2.1 are listed in full in Table 2.2. The yields of these minor sugars are listed in Appendix 1. Glucose, Table 2.1 Sugar y i e l d s In mg per gram dry peat fraction. -80 mesh Depth (cm) Rhamnose Fucose Rlbose Arabinose Xylose Mannose Galactose Glucose Mln.comp. SUM

0 -10 3.3 0.8 0.5 8.0 4.0 3.6 4.8 12.3 2.9 40.2

COHC/TOC*

2.7

64 122 -74 -132 2.6 2.2 0.6 0.8 0.3 0.4 2.0 2.1 2.5 6.1 3.6 5.2 4.1 5.3 9.8 17.2 2.9 4.3 28.4 43.6

183 -188 2.5 1.6 0.7 7.3 10.2 4.6 5.4 23.3 3.6 59.2

2.9

3.9

1.9

+20 mesh 0 64 Depth -10 -74 (cm) Rhamnose 5.4 5.4 1.1 1.0 Fucose 0.4 0.5 Rlbose Arabinose 23.8 20.3 24.4 16.9 Xylose Mannose 6.4 4.9 Galactose 9.5 6.9 Glucose 50.5 38.2 2.1 Mln.comD. 1.7 SUM 123.3 96.1

122 183 -132 -188 3.0 4.2 1.0 1.3 0.6 0.6 13.2 18.4 18.0 15.0 6.0 6.0 7.4 7.5 30.3 42.8 3,3 2.9 82.8 98.7

244 305 -254 -310 1.8 1.3 0.9 0.7 0.3 0.7 0.9 1.5 3.5 2.4 4.0 2.6 3.9 2.8 11.7 5.7 3.4 2.9 30.4 20.9 2.0

1.4

244 305 -254 -310 2.6 2.8 1.1 1.0 0.6 0.4 5.2 15.4 6.5 14.6 4.9 4.8 6.3 6.0 12.5 42.6 1,9 2.7 42.4 89.5

COHC/TOC* 8.2 6.4 5.5 6.6 2.8 6.0 *: grams carbohydrate carbon per 100 g total

37

galactose, mannose, xylose, arabinose, ribose, fucose regarded as "major" sugars.

and rhamnose are

The relative yields of major and minor sugars as percentages of the total carbohydrate yield are depicted in Fig. 2.1. They are plotted versus the total carbohydrate yield (in mg/g dry peat fraction). This way of presentation illustrates that in most cases definite relationships can be discerned between the relative contributions of sugar monomers, the total carbohydrate concentration and the grain size fraction. Galactose, mannose, fucose, rhamnose,

Table 2.2 Minor sugars identified and their loadings on first flve factors obtained from factor analvsls. variance preserved 26% 19% 13% 11% Fl F2 F3 F4 1) a heptose +0.68 -0.03 -0.20 +0.09 2) glucoheptose +0.83 -0.15 -0.18 +0.44 3) an amino sugar +0.81 +0.11 +0.22 +0.29 4) 6-O-methyl-mannose +0.70 -0.32 +0.37 -0.12 5) 6-O-methyl-galactose +0.51 -0.46 +0.19 -0.28 6) 2-O-methyl-arabinose +0.09 -0.79 -0.16 +0.23 7) 4-0-methyl-arabinose +0.18 -0.73 -0.13 +0.14 8) 2/5-O-methyl-mannose+0.22 -0.87 -0.07 -0.06 9) 2/5-O-methyl-galactose- +0.01 -0.66 -0.26 -0.35 10) 2/4-0-methyl-ribose~ -0.37 -0.52 -0.22 +0.08 11) 2/4-0-methyl-xylose~ -0.56 -0.74 -0.02 +0.23 12) 3-O-methyl-arabinose -0.54 -0.46 +0.57 -0.15 13) 3-O-methyl-xylose -0.59 -0.39 +0.31 -0.57 14) 4-0-inethyl-rhamnose -0.70 -0.15 +0.35 -0.07 15) erythrose -0.77 +0.06 -0.18 -0.14 16) threose -0.85 -0.05 -0.12 +0.20 17) 6-0-methyl-glucose -0.77 +0.02 +0.04 +0.22 18) a heptose -0.77 -0.04 +0.02 +0.20 19) 3-O-methyl-fucose -0.69 +0.15 +0.29 +0.36 20) glycerol -0.82 +0.29 -0.09 +0.25 21) 3-O-methyl-rhamnose +0.22 +0.71 +0.27 +0.19 22) glucosamine +0.42 +0.56 -0.11 +0.56 23) 3/4-O-methyl-mannose-0.03 +0.37 +0.36 -0.29 24) 3/4-0-methyl-galactose~ -0.16 +0.42 +0.75 -0.18 25) 3-O-methyl-glucose -0.04 +0.38 +0.56 -0.55 26) 2-O-methyl-rhamnose -0.23 +0.42 -0.69 +0.04 27) 2-O-methyl-fucose -0.39 +0.13 -0.64 -0.30 28) allose +0.21 -0.04 -0.62 -0.66 29) altrose +0.05 +0.06 -0.56 -0.46 30) 4-O-methyl-glucose +0.51 +0.30 +0.23 -0.75 31) 2-O-methyl-glucose -0.19 +0.29 -0.62 -0.42 32) 4-0-methyl-fucose +0.12 -0.55 +0.54 -0.01 33) an amino sugar -0.01 -0.47 +0.12 -0.24

the 9% F5 +0.01 -0.02 +0.00 +0.29 +0.39 +0.34 +0.47 +0.24 -0.43 +0.57 +0.25 -0.09 +0.00 -0.28 -0.21 -0.20 +0.32 +0.14 +0.09 +0.02 -0.06 -0.02 +0.31 +0.18 +0.30 +0.20 -0.08 -0.01 -0.09 -0.03 +0.18 -0.47 -0.58

-: Enantinttieric alditol acetates are not separated on CPsil88.

38

ribose, and the minor components (allose, altrose, the heptoses, amino sugars, tetroses, glycerol and partially methylated aldoses) show an increase in their relative contributions with decreasing total saccharide concentration irrespective of sample depth. In Fig. 2.1 only the relative contribution of the sum of the minor components is shown, because the individual minor components all show the same negative correlation between their relative contributions and total sugar yield. Arabinose and xylose tend to show increased contributions with increasing total carbohydrate concentration. A trend in contributions of glucose with increasing total carbohydrate concentration is not clear. The results of the coarse grained sample from 244-254 cm are remarkable: This sample has a low overall carbohydrate yield and shows at the same time large contributions of the minor components and of mannose, galactose, ribose, MINOR COMPONENTS

FUCOSE

RHAMNOSE • -2

•-5

-1

'-« —2 . - 1 ■-3

■~5 »+S

s O)

> -t-j

rham fuc rib ara xyl man gal gluc

V -V -

rham fuc rib ara xyl man gal gluc

rham fuc rib ara xyl man gal gluc

_o

P B

rham fuc rib ara xyl man gal gluc

Fig. 3.2 Distribution of the eight "major" sugars (in wt%) with the total yield of the eight sugars taken as 100%. A: Peat samples. B: Mariscus rhizome. Abbreviations: Rhamnose (rham), fucose (fïïc), ribose (rib), arabinose (ara), xylose (xyl), mannose (man), galactose (gal) and glucose (gluc).

59 The presence of a microbial population at various depths in the peat at Rookery Branch and in others has been established from microbiological studies (Given, 1972; Dickinson et al., 1974; Given and Dickinson, 1975). Microbial contributions to peat material have also been deduced from the occurrence of muramic acid and glucosamine (Casagrande and Park, 1978; Casagrande et al., 1985). Given and Dickinson (1975) mention the occurrence of blooms of cyanobacteria in fresh water environments in the Everglades. So some of the microbial saccharides analysed might be derived from these organisms. The attribution of minor components to the presence of microorganisms on the basis of literature data is supported by the absence of minor components in the results of analysis by GC-MS of the Mariscus rhizome (see Table 3.1). A similar absence of minor components has been observed in analysis of Alnus 0 - 1 0 cm. -SOmaiti

5? % _c

lil B

v D > '■*-»

o V

Fig. 3.3 Distribution patterns of the "minor" components (in wt%) in the peat samples. The total yield of all "major" and "minor" sugars is taken as 100%. Key to the numbers: 1} glycerol, 2) erythrose, 3) threose, 4) allose, 5) altrose, 6),8) heptoses, 7) glucoheptose, 9),11) amino sugars, 10) glucosamine, 12) 2-O-methylrhamnose, 13) 2-O-methyl-fucose, 14) 4-O-methyl-rhamnose, 15) 4-O-methylfucose, 16) 3-O-methyl-rhamnose, 17) 3-O-methyl-fucose, 18) 2/4-O-methylribose-, 19) 2-O-methyl arabinose, 20) 4-O-methyl-arabinose, 21) 2/4-O-methyl xylose-, 22) 3-O-methyl arabinose, 23) 3-O-methyl xylose, 24) 6-O-methylmannose, 25) 6-O-methyl-galactose, 26) 6-O-methyl-glucose, 27) 2/5-O-methylmannose-, 28) 2/5-0-methyl-galactose~, 29) 2-O-methyl-élucose, 30) 3/4-0methyl-mannose-, 31) 3/4-O-methyl-galactose-, 32) 3-O-methyl-glucose, 33) 4-0methyl-glucose. ~: Enantiomeric alditol acetates are not separated on CPsil88.

60 rootlets procured from an alder marsh and in the analysis of Myrica gale rootlets (Moers, unpublished results). Absence of minor components -with the exception of allose and a heptose, which are speculated to be derived from haustoria- was also observed in analyses of a leave, root, rhizome and periderm of a fresh Nymphaea alba specimen (Moers, unpublished results). Py-MS mapping Pyrolysis mass spectra are shown in Fig.'s 3.4 and 3.5. The mass fragments herein are attributed to various sources, like lignins, carbohydrates, proteins, aliphatic hydrocarbons and sulphur and nitrogen compounds on the basis of pyrolysis mass spectra of likely precursor materials and of reference compounds (van der Kaaden et al., 1984; Helleur et al., 1985; Boon and de Leeuw, 1987; Genuit et al., 1987; Helleur, 1987; Pouwels et al., 1987; SaizJimenez et al., 1987). It should be kept in mind that the mass spectrometer was operated in the low resolution mode so that the assignment of mass fragments to precursors is not unambiguous. The series of mass peaks at m/z 124, 138, 150, 152, 154, 164, 166, 178, 180, 182, 194, 196, 208 and 210 is indicative of guaiacyl-syringyl lignin (Genuit et al., 1987). The relative distribution of lignin markers contains information on the quality of the lignin: Shifts to lower m/z-values in the pyrolysis mass spectra probably due to demethoxylation and demethylation of the phenol derivatives point to increasing decomposition of the organic material and modification of the lignin (Stout et al., 1988). The series at m/z 94, 107, 108, 110, and 122 may indicate this modification (Boon, 1988). The high contribution of m/z 120 (vinylphenol) in some pyrolysates points to a contribution of monocotyledon plants. This compound may originate from pcoumaryl acid occurring esterified to xylose and arabinose in vascular plant tissues (Mueller-Harvey et al., 1986). Vinylphenol is formed readily during pyrolysis by decarboxylation of paracoumaryl acid (Alborn and Stenhagen, 1987). The mass peaks from polysaccharides in Py-MS are fragment-ion peaks and molecular ions of small pyrolysis products (Genuit and Boon, 1985; Pouwels et al., 1987). The mass peaks at m/z 114, 126, 128 and 144 are molecular ions of monomer specific sugar pyrolysis products, mainly anhydrosugars. The peaks at m/z 85, 86 and 114 are attributable to xylose and arabinose, m/z 128 originates from rhamnose, and m/z 126 and 144 are indicative for the presence of

61 glucose or other hexose isomers. These peaks can be used for tentative chemical interpretations. Hexosan and pentosan polymers also produce anhydrosugars which are evident in the spectra by their fragment-ions at m/z 57, 60 and 73 (e.g. Pouwels et al, 1987). The presence of fragments with m/z 43, 57, 69, 71, 83, 85, 97, 99, 111 and 113 is thought to be indicative of aliphatic hydrocarbons. Proteins are inferred from mass fragments at m/z 34, 48, 56, 64, 69, 80, 81, 83, 92, 97, 100 and 117 (Meuzelaar et al, 1982). Fig. 3.4 shows the pyrolysis mass spectra of two +20 mesh peat samples (from 0-10 cm and 137-147 cm) and of a Mariscus jamaicensis rhizome from 0-10 cm. The spectrum of the coarse grained sample from 0-10 cm shows mass peaks of polysaccharides and lignin and also marker peaks for proteins and fragment-ion series attributed to aliphatic hydrocarbon chains. The intensities of the lignin peaks are relatively low. A contribution of monocotyledon tissues, as from sawgrass, to this peat sample can be inferred from the intensity of m/z 120. The coarse grained peat sample from 137-147 cm shows high intensity peaks for lignin with a different nature than the shallow sample. The higher end of this spectrum shows similarities with the Mariscus rhizome spectrum, but the high intensities of peaks at m/z 110, 124, 138 and 150 are unusual for natural samples. These peaks are tracers for modifications of angiosperm lignin (Stout et al., 1988). The in this context remarkably high intensities of m/z 114, 57, 60, 73 and 126, indicate that intact polysaccharide must be present. The Py-MS spectrum of the Mariscus rhizome displays relatively high peaks for lignin, which points to heavy lignification and protection of the polysaccharides in this sample. Compared to other native plant material, the contribution of mass peaks from polysaccharides is remarkably low (Boon, 1988). This could be an indication for extreme senescence or even early stages of degradation. The relative abundance of m/z 34 from the pyrolysis product H 2 S in the spectra of Fig. 3.4 is much higher than commonly present in vascular plant material (Stout et al., 1988) and is also too high to be solely derived from proteins. It is therefore suggested that part of the sulphur originates from sulphides that are formed in the sediment due to activity of anaerobic bacteria.

40

EO

80 Hoss nuaber

40

GO

80

100 Moss number

100!

43

90

SAWGRASS RHIZOME 0 - 1 0 CM

180

Jao c T) 70 c a 60 O • so

ï» 20 10 0

40

GO

80

100

Moss number

63 Multivariate analysis yielded two discriminant functions that describe virtually all the important variance in the data. The first discriminant function (Dl) explains 69% of the characteristic variance (42% of the total variance) and the second discriminant function (D2) explains 31% of the characteristic variance (25% of the total variance). The chemical significance of the discri­ minant functions is deduced from their spectra shown in Fig. 3.5, indicating the sets of correlated mass peaks which differ either positively or negatively from the overall average spectrum (zero point spectrum).

Ui Fig. 3.5 The zero point spectrum and the pyrolysis-mass spectra of the positive and negative sides of the first and the second discriminant functions. For explanation see text. (opposite page) Fig. 3.4 Py-MS spectra of two +20 mesh peat samples from 0-10 and 137-147 cm and of the Mariscus rhizome from 0-10 cm depth.

64 The mass peaks in the negative Dl spectrum in Fig. 3.5 can mostly be attributed to methoxylated phenols and represent intact and modified lignin, whereas the mass peaks in the positive Dl spectrum represent polysaccharides, proteins and perhaps aliphatic hydrocarbons. It is known from the literature that the presence of protein markers is an indication of microbial activity (Boon and Haverkamp, 1982). The distribution pattern of the polysaccharide markers differs from the ones commonly seen in vascular plants: The hexose characteristics are relatively low and the rhamnose (m/z 128) and pentose (m/z 112 and 114) characteristics are comparatively high in the present case. Mass fragments attributed to aliphatic hydrocarbon chains on the positive side of the Dl function might be derived from resistant aliphatic biopolymers as occurring in cell walls of certain algae (Derenne et al., 1988), and in plant cuticles (Nip et al., 1986) and probably rootlets (van Smeerdijk and Boon, 1987). The first discriminant function is therefore interpreted as differentiating between a mixture of microbial remains and refractory fine grained matter on the positive side versus vascular plant input on the negative side. The positive side of the D2 spectrum shows mass peaks predominantly derived from mono- and non-methoxylated phenols and represents modifications in the lignin polymers (compare Stout et al., 1988). Markers for nitrogen compounds (m/z 67, 79 and 81) and for sulphur compounds (m/z 34 and 64) also load on the positive side of this second discriminant function. This points to incorporation of organic nitrogen and to anaerobic conditions. The negative side of the D2 spectrum represents intact vascular plant polysaccharides composed of glucose, xylose and arabinose. The relatively high m/z 45 may be related to acetylation of this polymer system. This second discriminant function is interpreted as differentiating between the extent of degradation (as indicated by the extent of lignin modification) on the positive side versus the presence of intact (living) vascular plant material on the negative side. Figures 3.6A and B show depth profiles for the scores of the samples on the Dl and D2 discriminant functions. The fine grained samples score higher on the positive side of the Dl function and the samples from the two coarse grained fractions score higher on the negative side. According to the interpretation mentioned above, this means that the fine grained samples show more indications of microbial activity while the consecutive coarser grained samples show more indications of vascular plant contributions. The depth profile should

65 therefore be interpreted as a relative accumulation of refractory lignified vascular plant materials. The plot of the second discriminant function versus depth (Fig. 3.6B) shows that the fine grained samples exhibit the most pronounced characteristics of modified lignin, while the consecutive coarser grained samples show increasingly more characteristics of intact vascular plant polysaccharides. This means that the fine grained samples reveal more indications of degradation than the coarser grained samples. The trend with increasing depth can be interpreted as an initial increase in degradation characteristics going from 0-10 to 51-61 cm in the peat. The relative contributions of modified lignin and intact vascular plant polysaccharide seem to remain rather constant going from 51-61 to 137-147 cm in case of the +80 and -80 mesh samples. The +20 mesh sample from 137-147 cm, however, shows increased characteristics of intact vascular plant polysaccharides. B first discriminant function

second discriminant function

O

50

100

1501 i i i i i i i i i i i i i i—r—r-1 0 1

score ■ +20 mesh

+ +80 mesh

150

1

-4

-2

1

i

r

0 score

O - 8 0 mesh

> rhizome

Fig. 3.6 A: Plot of the first discriminant function versus depth. Markers for carbohydrates, proteins and aliphatic hydrocarbons load on the positive side, and markers for (guaiacyl-syringyl) lignin load on the negative side. B: Plot of the second discriminant function versus depth. Markers for modified lignin load on the positive side, and markers tor intact vascular plant polysaccharides load on the negative side.

66 The plotting of the rhizome spectrum at highly negative Dl and D2 scores is a reflection of its relatively high lignin content and of its relatively low degree of degradation and consequently good preservation of vascular plant polysaccharides. General discussion The trend from positive to negative Dl scores with increasing depth in Fig. 3.6A indicates a relative loss of polysaccharide, protein and aliphatic hydrocarbon characteristics and a relative gain in lignin derived markers. These results indicate a relative decrease in microbial activity. This is in agreement with the results of wet chemical analysis, where the decrease in total carbohydrate concentrations going from 0-10 to 51-61 cm in the peat (see Table 3.1) seems to be caused by degradation of vascular plant and microbial sugars (e.g. Casagrande and Park, 1978; Hatcher et al, 1981; Benner et al., 1984a,b; Given et al, 1984; Casagrande et al., 1985; Hedges et al, 1985). The initial trend from negative D2 scores to more positive ones on increasing depth is also in agreement with this decrease in total carbohydrate concentra­ tions, because the trend in the D2 scores exemplifies increasing degradation of the organic material. The comparatively high total carbohydrate concentration in the +20 mesh sample from 137-147 cm (see Table 3.1) and its relatively high score on the negative side of the second discriminant function (see Fig. 3.6) can be explained by the presence of living and senescent rootlets, presumably from Meniscus and maybe Persea and Myrica in the peat at this depth. A high concentration of living rootlets in the deepest sample may be caused by the fact that the rootlets are forced to grow horizontally when they reach the underlaying limestone, thus giving rise to a higher rootlet density. The high proportion of glucose in this sample (see Table 3.1 and Fig. 3.1) could then be partly derived from energy storage glucans. The relatively high contribution of arabinose may be connected with the presence of tissues derived from Mariscus. This link is suggested by the high contribution of arabinose in the rhizome sample from 0-10 cm (see Table 3.1). It is unlikely that the arabinose is derived from the common arabinogalactans in pectins, seeing the distribution patterns of the two constituting monomers in the rhizome and peat samples. It is more likely that the high arabinose concentration is related to

67 the presence of protective materials in (tropical) grasses in which arabinoxylans are linked to phenolic acids (Mueller-Harvey et al., 1986; Boon, 1988). The absolute yields (in mg/g) of glucose and arabinose from the +20 mesh sample from 137-147 cm might be "corrected" for this contribution of living and senescent rootlets. Based on the sugar concentrations in the +20 mesh samples from 0-10 and 51-61 cm and on the concentrations in the -80 mesh samples from 51-61 cm and 137-147 cm, it seems reasonable to subtract 30 and 7 mg/g from the contributions of glucose and arabinose respectively. This procedure yields both a total carbohydrate yield (mg/g) and relative carbohydrate contributions (wt%) that are in better correspondence with the sugar results of the other peat samples (see Table 3.2). Table 3.2 Sugar yields of +20 mesh samples, corrected for contributions of glucose and arabinose attributed to fresh rootlets in the sample from 137-147 cm. 0-10 cm 51-61 cm 137-147 cm mg/p wt% wt% wt% 2.3 rham 3.8 3.3 1.5 fuc 2.4 1.0 2.3 0.7 rib 0.8 0.5 1.2 0.9 16.0 ara 10.3 11.0 15.9 xyl 16.8 17.9 11.6 15.9 man 5.2 3.4 8.7 5.4 gal 11.0 9.3 6.0 9.9 gluc 34.8 40.9 43.6 28.2 minor 2.4 comron. 10.3 5.2 3.7 64.6 SUM

3.5 CONCLUSIONS 1) The results of wet chemical carbohydrate analysis agree very well with those of Py-MS and multivariate analysis of the Py-MS data. 2) It appears that two sources of organic material can be recognized in the peat samples: Vascular plants and microorganisms. 3) The coarse grained samples show most of the characteristics attributed to vascular plants: Relatively high total carbohydrate yields with relatively high contributions of glucose, xylose and arabinose, and comparatively high scores on Dl- (loaded with lignin markers) and on D2- (loaded with markers indicative for intact vascular plant polysaccharides).

68 4) The fine grained samples show most of the characteristics related to microbial activity: Relatively low total carbohydrate yields with relatively high contributions of galactose, mannose, ribose, fucose, rhamnose and the minor components, and comparatively high scores on D1+ (with loadings of markers for proteins, carbohydrates and aliphatic hydrocarbons) and on D2+ (with loadings of markers indicative of modified lignin). 5) With increasing depth, there is an initial decrease in total carbohydrate concentrations and the samples show characteristics pointing to a relative accumulation of refractory lignified vascular plant material. 6) Both Py-MS and wet chemical analysis indicate that living and/or senescing rootlets may be present at the deepest level in the peat.

Acknowledgements. The financial support for this investigation of the Koninklijke/Shell Exploratie en Produktie Laboratorium (Shell Research BV), Rijswijk, is gratefully acknowledged. The authors thank Drs. P.H. Given and NJ. Ryan for providing the samples. Mrs. B. Brandt-de Boer is thanked for performing the Py-MS and multivariate analysis at the FOM Institute AMOLF, Amsterdam. Dr. H.C. Cox is thanked for critically reading the manuscript.

69

4. CHARACTERIZATTON OF TOTAL ORGANIC MATTER AND CARBOHYDRATES IN PEAT SAMPLES FROM A CYPRESS SWAMP BY PYROLYSIS-MASS SPECTROMETRY AND WET-CHEMICAL METHODS

M.E.C. Moers, J.J. Boon and J.W. de Leeuw Submitted

4.1 ABSTRACT Organic matter present in coarse and fine grained sample preparations of a peat core from Minnie's Lake, Okefenokee Swamp (Georgia) was analysed on a molecular level by means of gas chromatography-mass spectrometry of alditol acetates obtained from polysaccharides after acid hydrolysis and derivatization and by means of pyrolysis-mass spectrometry. The results from analysis of alditol acetates indicate that: 1) The coarse grained fraction is dominated by sugars derived from vascular plants whilst the fine grained fraction is dominated by sugars derived from microorganisms. 2) Relatively high contributions of xylose discriminate the Taxodium derived peat from the underlying peats derived from Carex and Nymphaea which show relatively low contributions of this sugar. Factor-discriminant analysis of the Py-MS data indicates that: 1) Most carbohydrate degradation is restricted to the upper peat levels and that guaiacyl and syringyl type lignins are better preserved than carbohydrates. 2) Resinous organic matter -presumably derived from Taxodium- is selectively preserved in the fine grained fraction. Deviant sugar concentrations and Py-MS results observed for the top sample of the Carex peat were explained by oxygenation of the peat, a concomitantly high microbial activity and subsequent preservation of the organic material due to the action of phytotoxins derived from Taxodium.

70

42 INTRODUCTION The major aims of this investigation were firstly to determine how decomposition and degradation of the organic material in peats are reflected in the carbohydrate abundances and in the lignin and carbohydrate patterns and secondly to establish in which way carbohydrates in peats can be attributed to their precursors. To this end coarse (+20 mesh) and fine grained (-80 mesh) peat sample preparations from several depths were analysed by pyrolysis-mass spectrometry to evaluate qualitatively the nature of the input material and changes therein during peatification, as well as by wet chemical methods to determine the different kinds of sugar monomers present and their concentrations. Minnie's Lake -a cypress swamp in the Okefenokee Swamp in southern Georgia (USA)- was selected for this purpose because the peat core procured from this location represents different environments of peat deposition and so information may be obtained on the pathways by which different vascular plants are decomposed. Another advantage is that the upper three samples from this core are mainly composed of the same vascular plant species, i.e. Taxodium, and therefore give insight into the effects of degradation of one particular species. The Okefenokee swamp/marsh system consists of swamp forests dominated by Taxodium (cypress), of glades or island fringes inhabited by emergent aquatic plants like Carex (sedge), Panicum (maidencane) and Woodwardia (chainfern) and of open water environments dominated by Nymphaea (waterlily) (Cohen, 1973a; Spackman et al. 1981). The swamp is underlain by a Pleistocene marine terrace, which in most cases is a relatively pure sand. Peat accumulation in this area did only start about 6500 y. B.P., so it was concluded that the swamp developed in a fresh water environment from the start onward (Cohen 1973b). The basal peat at Minnie's Lake has been dated at 4000 y. B.P. (Casagrande et al., 1985) and a total of 3.5-3.7 m of peat accumulated at this site (Cohen, 1973b). The present vegetation at Minnie's Lake consists of a dense Taxodium swamp (Cohen, 1973b; Spackman et al., 1981). Components observed microscopically by Ryan (1985) in the same peat core were: 0-10 cm: Partially disrupted Taxodium twigs, leaves and rootlets. 71-81 cm: Fragments of cortical tissue from Taxodium roots and rootlets, though few recognizable organs. 112-122

71 cm: Similar to previous level, but mainly cell fragments. 152-163 cm: Tissues from leaves, roots and rhizomes of Carex. 183-193 cm: Cortical tissue from rhizomes and roots of Carex. 244-254 cm: Mixture of Nymphaea and Carex peat, and some fusinite. It appears from these observations that the Minnie's Lake area was previously an open marsh which has since been overgrown by swamp vegetation (Cohen 1973b). The terminology used for description of the microscopic observations is explained in Cohen and Spackman (1977). The Py-MS approach has been shown to be an appropriate method to monitor the overall organic matter in natural samples and changes therein (e.g. Boon et al., 1984; Saiz-Jimenez and de Leeuw, 1986; Pouwels et al., 1987). The analysis of carbohydrates by wet chemical methods has been shown to be succesful in discriminating microbial and vascular plant sugars in peat systems (chapters 2 and 3).

4.3 EXPERIMENTAL For the analyses peat samples from six different depths within the core (see Table 4.1) were used. During previous investigations by Rhoads (1985) and by Ryan (1985) each sample had been separated into coarse grained (+20 and +80 mesh) fractions consisting of vascular plant tissues and tissue fragments, and into a fine grained (-80 mesh) fraction consisting of amorphous matter. For information on the fractionation procedures the reader is referred to Given et al. (1984) or to Roads (1985). The -80 mesh fraction had been treated with 0.1 M HC1 to remove carbonates and all peat fractions had been Soxhlet extracted with benzene/ethanol 2/1 v/v for 48 h. by Rhoads (1985). This extraction solubilized ca. 12 wt% of the dry material. The organic carbon content of the peat fractions is ca. 63 wt% with a carbon/ nitrogen ratio of ca. 21 (Rhoads, 1985). Analvsis of alditol acetates For carbohydrate determinations +20 mesh (coarse grained) and -80 mesh (fine grained) samples were analysed. Half of the samples was analysed in duplicate as a check on the reproducibility.

72

From each fractionated, freeze dried and pulverized peat sample ca. 100 mg was pretreated with 5 ml of 12 M H 2 S0 4 for 2 h. at room temperature after which the acid was diluted to 1 M and the sample hydrolysed at 100° C for 4.5 h. Myo-inositol was added as an internal Standard. The monosaccharides released were reduced to alditols by NaBH« (room temperature, 16 h) and then acetylated to alditol acetates by acetic anhydride/pyridine (100°C, 3 h). The reaction mixtures were analysed by gas chromatography (GC) on a Carlo Erba 4160 gas chromatograph and by gas chromatography - mass spectrometry (GC-MS) on a Hewlett-Packard 5890 gas chromatograph coupled to a VG 70-250SE mass spectrometer. In both cases a CPsil88 fused silica capillary column (1=25 m, i.d.=0.32 mm, df=0.12 /im; Chrompack, Middelburg, The Netherlands) was used. The samples were injected at 50°C (GC-MS) or 70°C (GC). The tempera­ ture was rapidly raised to 1500C and from thereon programmed at 3° C/min to 230°C and maintained at this temperature for 40 min. The mass spectrometer was operated in the electron impact mode at 70 eV, with a source temperature of250°C. For identification of the alditol acetates combined use was made of mass spectral data and of relative retention times, except in the cases of some heptoses and amino sugars where only mass spectral data were available (Stoffel and Hanfland, 1973; Schwarzmann and Jeanloz, 1974; Jansson et al., 1976; Radziejewsky-Lebrecht et al., 1979; Wong et al., 1980; Klok et al., 1982). Quantification of the compounds analysed was achieved by peak area integration of the GC results with the aid of a Maxima Chromatography Workstation (Dynamic Solutions Corporation, Ventura, CA, USA) and comparing the responses with the one of myo-inositol hexaacetate (the internal Standard). The flame ionization detector responses of all alditol acetates were assumed to be equal on weight basis. In a few cases in which compounds coeluted, quantification was achieved by peak area integration of the responses of selected characteristic ion fragments from GC-MS analyses with the aid of software available with the mass spectrometer. From the duplicate analyses it was estimated that the error in the determinations of the "major" components is 10% or less and that the error in the determinations of the "minor" components ranges from ca. 50% for the compounds present in less than 0.02 mg/g to 10% for the others.

73 Analvsis by pvrolysis-mass spectrometry Aliquots of peat samples suspended in water were applied to ferromagnetic wires and dried in vacuo. Curie point pyrolysis mass spectra were obtained with the FOMautoPYMS (Meuzelaar et al., 1982). The conditions were as follows: Curie temperature, 610°C; heating time, 0.1 sec; total pyrolysis time, 0.8 sec; heated pyrolysis chamber, ca. 180°C; temperature of the expansion chamber, 210° C; ionization by electron impact, 16 eV; mass range, m/z 25240; scan speed, 10 scan/sec; total number of averaged spectra, 200. The samples were analysed in triplicate. Multivariate techniques Comparison of the Py-MS spectra was carried out by a factor and discriminant analysis procedure using a modified ARTHUR computer program package, adapted to accept Py-MS data (Hoogerbrugge et al., 1983; Boon et al., 1984). The output consists of plots which show the relative differences in composition between the samples. The variables which describe the discrimination are given by so-called discriminant functions. Factor analysis using the same package as described above was also applied to the data of the 33 "minor" components (see below). The output consists of plots and maps which show the scores, i.e. the relative differences in minor sugar distributions, of the samples on the various factors.

4.3 RESULTS AND DISCUSSION Analysis of carbohvdrates as alditol acetates The hydrolysates of the peat samples yield a great variety of neutral saccharides: 1) Glucose, galactose, mannose, xylose, arabinose, ribose, fucose and rhamnose (see Table 4.1). This group will be called 'major' saccharides and comprises 98 to 99 wt% of the total sugar yield in the coarse grained fraction and 86 to 93 wt% in the fine grained fraction. 2) Partially methylated sugars, amino sugars, heptoses, allose, altrose, tetroses and glycerol, which are collectively called 'minor' sugars in this paper. Table 4.2 lists all minor

74 Table 4.1 Carbohydrate y ields in mg per gram dry peat fractlon. -80 mesh Depth 0 71 112 152 183 244 (cnO -10 -81 -122 -163 -193 -254 1.3 Rhamnose 3.3 3.7 1.7 3.1 2.1 0.4 0.8 0.4 1.6 Fucose 0.7 0.6 Rlbose 0.5 0.2 0.2 0.4 0.3 0.2 2.3 0.6 0.6 1.9 Arablnose 0.7 1.1 6.0 2.1 Xylose 2.4 7.2 3.0 3.3 3.1 8.0 2.9 7.6 4.8 Mannose 3.7 7.0 2.3 2.4 6.7 2.9 Galactose 5.1 36.4 23.6 21.3 43.2 Glucose 30.7 19.3 Min.comp. 5.9 3.1 3.2 7.0 3.5 6.3 SUM 70.6 37.0 34.8 77.8 47.4 45.1 2.3

2.2

4.9

3.0

2.9

+20 mesh Depth 0 71 (cm) -10 -81 Rhamnose 4.2 0.9 0.6 Fucose 0.3 0.6 Rlbose 0.4 3.8 Arabinose 2.0 25.9 Xylose 17.5 Mannose 16.9 4.8 Galactose 12.2 4.5 Glucose 119.8 91.5 Min.comp. 2.1 1.7 SUM 185.7 124.0

112 -122 1.2 0.2 0.4 1.7 12.9 6.1 3.9 75.9 1.3 103.6

152 -163 0.8 0.2 0.4 0.9 7.5 3.5 2.7 79.3 1.1 96.4

183 -193 1.4 0.3 0.3 0.9 8.1 4.5 3.3 77.3 1.6 97.7

244 -254 2.0 0.9 0.4 1.2 7.6 5.2 5.3 67.7 2.2 92.5

COHC/TOC*

4.5

COHC/TOC* 11.8 7.8 6.6 6.1 6.2 5.9 *: gram carbohydrate carbon (C0HC) per 100 g total orpanlc carbon (T0O .

sugars identified. The yields of the minor sugars are listed in Appendix 4. The occurrence of the so-called major sugars is reported in many papers dealing with carbohydrate analyses of natural samples, i.e. recent sediments with contributions of organic material from 'marine' and 'terrestrial' sources and their floral and faunal precursors (e.g. Myklestad, 1974; Sjöström, 1981; Cowie and Hedges, 1984; Klok et al, 1984a,b; Michaelis et al., 1986; Tanoue and Handa, 1987; Hamilton and Hedges, 1988). The occurrence of the minor sugars is often attributed more specifically to microorganisms (Cheshire et al., 1969; Given, 1972; Given and Dickinson, 1975; Stewart, 1974; Casagrande and Park, 1978; Cheshire 1979; Aspinall, 1983; Klok et al, 1984a,b; Casa­ grande et al., 1985). Some of the partially methylated saccharides and amino sugars are also reported to occur in vascular plants (Lowe, 1978; Cheshire,

75 Table 4.2 Minor components and loadings on the flrst five factors obtalned from factor analvsls. variance preserved 30% 25% 10% 15% factors F2 F4 Fl F3 -0.30 -0.82 +0.42 +0.12 1) glycerol -0.38 -0.84 -0.19 -0.17 2) erythrose -0.45 -0.54 -0.39 -0.04 3) threose -0.30 +0.07 -0,53 -0.25 4) allose +0.03 -0.11 -0.39 -0.73 5) altrose +0.38 -Q.62 -0.00 -0.04 6) a heptose +0.78 -0.30 -0.31 +0.13 7) glucoheptose +0.76 +0.10 +0.43 -0.20 8) a heptose +0.62 +0.47 -0.00 -0.27 9) an amino sugar +0.72 -0.35 +0.47 -0.11 10) glucosamlne +0.79 +0.40 +0.22 -0.08 11) an amino sugar -0.60 -0.26 -0.45 +0.34 12) 2-0-methyl-rhamnose -0.64 -0.28 -0.39 +0.09 13) 2-0-methyl-fucose -0.79 +0.52 -0.13 -0.22 14) 4 -0-methyl-rhamnose +0.06 +0.12 -0.89 +0.23 15) 4-0-methyl-fucose +0.61 +0.43 -0.50 -0.25 16) 3-0-methyl-rhamnose +0.29 +0.24 -0.81 +0.06 17) 3-0-methyl-fucose -0.79 +0.35 +0.14 +0.02 18) 2/4-0-methyl-ribose-0.23 -0.52 -0.65 -0.38 19) 2-0-methyl-arablnose 20) 4-0-methyl-arabinose +0.10 -0.58 -0.62 +0.20 -0.82 +0.26 +0.41 +0.05 21) 2/4-0-methyl-xylose-0.09 +0.86 -0.30 -0.10 22) 3-0-methyl-arabinose 23) 3-0-methyl-xylose -0.21 +0.84 -0.17 -0.13 +0.72 +0.31 -0.21 +0.47 24) 6-0-methyl-mannose 25) 6-0-methyl-galactose +0.75 +0.29 -0.09 +0.55 +0.31 -0.28 +0.02 +0.81 26) 6-0-methyl-glucose -0.73 +0.31 +0.49 +0.02 27) 2/5-0-methyl-mannose28) 2/5-0-methyl-galac tose- -0.98 -0.00 +0.12 -0.06 -0.58 -0.32 +0.08 +0.18 29) 2-0-methyl-glucose -0.20 +0.79 +0.04 +0.30 30) 3/4-0-methyl-mannose31) 3/4-0-methyl-galac tose- -0.08 +0.79 -0.13 -0.11 -0.26 +0.69 -0.02 +0.14 32) 3-0-methyl-glucose -0.39 +0.75 -0.23 +0.39 33) 4-0-methyl-glucose

7%

F5 -0.04 -0.16 -0.49 +0.72 -0.22 +0.22 +0.28 +0.02 -0.20 +0.18 +0.17 -0.26 -0.18 +0.06 +0.26 -0.05 -0.13 +0.25 +0.22 +0.04 +0.16 +0.14 +0.14 +0.18 -0.00 -0.09 +0.16 +0.03 +0.64 +0.00 -0.05 -0.64 +0.02

Enantiomeric alditol acetates are not separated on CPsll88

1979; Aspinall, 1983), but their contributions are too low for detection with the GC-MS technique used at present (chapters 2 and 3). This justifies the exclusive attribution of the minor sugars to microorganisms, i.e. bacteria, algae and fungi, in the present case. From Table 4.1 it can be seen that the coarse grained samples yield higher total amounts of carbohydrates than the fine grained samples, irrespective of depth and composition of the peat. It appears that the sugar concentrations in

76 the coarse grained samples initially decrease with increasing depth foliowed by stabilization to ca. 95 mg/g. The fine grained samples also show an initial decrease in total sugar concentration, foliowed by an unexpected üregular increase in total sugar yield in the sample from 152-163 cm (see Table 4.1). In Fig. 4.1 the sum of all individual sugar yields (= the total sugar yield in mg sugar per g peat fraction) is plotted on the horizontal axes of the plots, while the vertical axes are formed by the relative yields of individual sugars in wt%. Relative means the weight fraction of an individual sugar with respect to the total sugar yield. This figure shows that a relationship exists between the relative sugar contributions, the grain size and the total sugar yields: 1) The minor sugars together with rhamnose, fucose, ribose, mannose and galactose tend to be relatively depleted in the coarse grained fraction and enriched in the fine grained fraction. 2) Glucose tends to be enriched in the coarse grained fraction compared to the fine grained fraction. 3) The coarse rhamnose

nrjinor components

fucose

10 •-5 •-2

&?

+4/-+J

o 1 —■ ■ . . 7

V) CD CD

>

'-t

-,

^

A

ribose

arabinose

5 +3 •

»+2

.+6

3

_o

galactose

mannose

«-1

.+4 •+5

r-t

10

alucose 3

10'

•-4 -2

+ 1»

+6

70

+ 1.

•-3

•-3 -4

5

5

50

100

150

50

100

. •-

50

^ 5i «+3 .-f2 •+4

•-6

150

50

yield total sugars

(mg/g)

100

150

Fig. 4.1 Relative sugar yields in wt% of the total yield versus the total sugar yield (mg/g). The numbers next to the dots within the figures indicate the depth of the peat sample: 1= 0-10 cm, 2= 71-81 cm, 3= 112-122 cm, 4= 152163 cm, 5= 183-193 cm, 6= 244-254 cm. "+" and "-" indicate coarse grained (+20 mesh) and fine grained (-80 mesh) samples respectively.

77 grained fraction tends to show higher total sugar yields than the fine grained fraction as was already mentioned above. On the basis of literature data (see above) and data reported here, it can be concluded that the comparatively high contributions of minor sugars and at the same time of rhamnose, fucose, ribose, mannose, and galactose to the fine grained samples suggest a predominant microbial origin of all sugars in these samples. This is in agreement with the reported total sugar yields of the two fractions: Microbial activity will certainly lead to a reduced carbohydrate yield, not only because vascular plant tissues themselves yield more carbohydrates than bacteria do (Cowie and Hedges, 1984), but the more so because the biomass of newly synthesized microbial polysaccharides represents only a small fraction of the biomass of -microbially degraded- vascular plant sugars that was present originally. Carbohydrate compositions are different for different vascular plant species (e.g. Kirk, 1973; Cowie and Hedges, 1984). It is therefore to be expected that the monosaccharide compositions of the coarse grained samples reflect their vascular plant precursors. Only arabinose and xylose seem to give clues in this respect in the present case: The Taxodium peat shows higher contribu­ tions of these sugars than either the Carex or Nymphaea peats (see Fig. 4.1). This fits with observations from the literature (Lowe, 1978) that polysaccharides of vascular plants are commonly composed predominantly of only one to three monomeric constituents, while polysaccharides in soils, which are derived from a great number of different organisms, are usually characterized by six or more different monomers. Factor analysis of the minor sugars was carried out in order to investigate possible relationships between the contributions of certain minor sugars and sample grain size, sample depth, microbial communities etc. This technique permits the grouping of components that are covariant. Table 4.2 lists the loadings of the minor sugars on the first three factors. These factors describe 30 %, 25 % and 15 % of the total variance respectively. Analyses of minor sugars in cell walls and membranes of several photosynthetic prokaryotes by Weckesser et al. (1979) and Schmidt et al. (1980a,b) have shown that the composition of the polysaccharide moiety of lipopolysaccharides is rather species specific. We hypothesize that the observation of species (or

78

family or class) specificity of the composition of structural polysaccharides probably can be expanded for microorganisms in general. This implicates that the differences and similarities in minor sugar contributions in the peat samples probably reflect differences and similarities in microbial populations. It should be stressed, however, that interpretation of minor sugar contributions in terms of differences in microbial populations is seriously hampered by a general lack of knowledge about systematic occurrences of partially methylated sugars, heptoses, amino sugars etc. in microorganisms (apart from the above mentioned studies of Weckesser et al. (1979) and of Schmidt et al. (1980a,b). Consequently it is impossible to speculate on any detailed level about the biological meaning of the differences in minor sugar distributions. Fig. 4.2 shows depth profïles of the scores of the samples on the first three factors. It appears that the fine grained samples do not show much differentiation in their contributions of minor sugars. This is interpreted as indicating rather similar microbial populations in these samples, irrespective of depth. The coarse grained samples show much more variation, which is inter­ preted as a consequence of the presence of different microbial populations in

A. f i r s t

factor

B. second f a c t o r

■ - 8 0 mesh

+

C. t h i r d

factor

+ 2 0 mesh

Fig. 4.2 Depth profiles of the scores on the first three factors that describe the variance in the contributions of the minor components. The identities of the sugars that load on these factors are listed in Table 4.2.

79 the various coarse grained samples. No relationship seems to exist between the distributions of minor sugars in the coarse and fine grained samples. This might point to a similar lack of relationship between the microbial populations in the two fractions. The coarse grained samples tend to score higher on the negative side of Fl than the fine grained samples, apparently showing more characteristics of 2-0methyl-sugars while the fine grained samples show more characteristics of amino sugars, heptoses, and 6-O-methyl-mannose and -galactose (see Fig. 4.2 and Table 4.2). It is speculated that this first factor differentiates between bacterial sugars (F1 + ) and sugars from fungi and algae (F1-), because especially the lipopolysaccharides of Gram-negative bacteria and cyanobacteria are known to contain substantial amounts of heptoses and aminosugars (Carr and Whitton, 1982; Laskin and Lechevalier, 1982; Aspinall, 1983). Part of the amino sugars, of course, must also be derived from peptidoglycans in bacterial cell walls. The coarse grained samples from the Taxodiwn peat show the largest variation in their scores on the first and second factor as can be seen from Fig. 4.2a and 4.2b. The third factor clearly distinguishes the coarse grained Taxodiwn samples from the coarse grained Carex/Nymphaea samples. The latter group is characterized by relatively high contributions of 3- and 4-O-methyl-fucose and of 2- and 4-O-methyl-arabinose. Again it is not clear how this should be interpreted in terms of differences in microbial populations associated with different plant species. Comparison of sugar data with literature data Cowie and Hedges (1984) used xylose and mannose contributions -on a glucose free basis- to distinguish angiosperms from gymnosperms: They report xylose contributions to be relatively high in angiosperms and low in gymnosperms, while the opposite is true for mannose. They also state that nonwoody tissue shows higher arabinose + galactose contributions -on a glucose free basisthan woody tissue. Application of these parameters in the present investigation would strongly indicate a non-woody angiosperm source of the organic matter (i.e. relatively high xylose values, low mannose values and high arabinose + galactose values -all on a glucose free basis- and no trends with depth in any of the cases). This is in contradiction with the results from microscopic determinations (Cohen, 1973a; Spackman et al., 1981; Ryan, 1985),

80

which have shown that the upper three coarse grained samples are derived from woody and nonwoody gymnosperm (roots/rhizomes and leaves of Taxodium), while the lower three samples are predominantly derived from nonwoody angiosperm (roots/rhizomes of Carex and Nymphaea). From other sources (e.g. Lomax et al., (1983), Chesson et al., (1983, 1985); chapter 5) it is known, however, that especially grasses (nonwoody angiosperms) contain relatively high contributions of xylose. A second contradiction between our data and literature data concerns the specificity of vascular plant versus microbial sugars: Hamilton and Hedges (1988) report, among other things, results of carbohydrate analyses in very recent sediments from a coastal marine environment. They were able to distinguish a vascular plant input -indicated by high contributions of glucose, lyxose, mannose and xylose-, and input from marine algae and presumably bacteria -indicated by high contributions of arabinose, galactose, rhamnose and fucose-. So their explanation of the origin of high contributions mannose (from vascular plants) opposes the one we offer in our case (from microorganisms). The two examples mentioned above illustrate that great care should be taken when comparing carbohydrate results of samples from such different backgrounds as intact plant tissues (Cowie and Hedges, 1984), sediment samples from a coastal marine environment (Hamilton and Hedges, 1988) and fractionated peat samples (our case). Several fresh water algae are known to contain appreciable amounts of structural mannans, in contrast to marine algae (Stewart, 1974; Aspinall, 1983). This could explain why Hamilton and Hedges consider mannose as derived from vascular plants, while it appears to be derived from microorganisms in our case. From further comparisons of the results of several studies in which comparable hydrolysis methods were used (e.g Ittekkot et al., 1982; Degens and Mopper, 1976; Cowie and Hedges, 1984; Hedges et al., 1985; Tanoue and Handa, 1987; Hamilton and Hedges, 1988; chapters 2 and 3), it appears that the environment of sedimentation and the microbial processes taking place in the watercolumn and/or sediment leave complicated imprints with respect to sugar concentrations. This makes it impossible to make generally applicable state­ ments concerning origins of sugar monomers in sediments or of their relatively high or low contributions. Liebezeit (1987) reached similar conclusions concerning attempts to use individual sugars as source indicators from investigations of carbohydrates in the water column.

81 In summary on might conclude that within a reasonably well defined ecosystem carbohydrate profiles appear to be very useful and rather straightforward. However, generalizations for different environments of the significance of individual sugars as markers for specific sources should be made with great care. Pv-MS mapping Fig. 4.3 shows representative pyrolysis mass spectra of fine grained peat samples. The mass peaks in these spectra are tentatively attributed to various sources on the basis of literature data obtained for pure biopolymers and sedimentary organic matter: The presence of guaiacyl-syringyl lignin is deduced from mass peaks at m/z 124, 138, 150, 152, 154, 164, 166, 178, 180, 182, 194, 196, 208, 210 (Genuit et al., 1987; Saiz-Jimenez et al., 1987). The peaks at m/z 94, 107, 108, 110, 122 are indicative of modification of the lignin, because increasing decomposition of the organic material and modification of lignin is attended by demethoxylation and demethylation of the phenol derivatives and consequently by shifts to lower m/z values in the Py-MS spectra (Stout et al., 1988). Polysaccharides are indicated by several series of mass peaks (e.g. van der Kaaden et al., 1984; Genuit and Boon, 1985; Helleur et al., 1985; Pouwels et al., 1987): The peaks at m/z 85, 86, 114 are attributable to xylose and arabinose. The peak at m/z 128 represents a characteristic pyrolysis product of rhamnose. Hexoses are indicated by peaks at m/z 126, 144. Hexosan and pentosan polymers also produce anhydrosugars which are evident in the spectra by fragment ions at m/z 57,60,73 (e.g. Helleur, 1987). Mass fragments at m/z 34, 48, 56, 64, 69, 80, 81, 83, 92, 97, 100, 117 are interpreted as being derived from proteins (Meuzelaar et al., 1982; Broek et al., 1985). Markers for polysaccharides and lignin constitute the major message in the pyrolysis mass spectra of the peat samples as is shown in Fig. 4.3 for some fine grained samples. Apart from polysaccharide and lignin contributions there are mass peaks at m/z 42, 43, 56, 57, 70, 71, 82, 84, 85, 97, 98 which might point to the presence of aliphatic hydrocarbons. The fine grained samples from 0-10 and 152-162 cm and the coarse grained one from 0-10 cm mainly show saccharide characteristics. The pyrolysis mass spectra of the samples from 0-

R*latlv« obundanc»

5 S 8 S B S c * 8 8 g

Ralollv*

obundanc*

8 S S 8 8 8 8 §

R* 1o 11v* obundonc•

8 8 S 8 8 Ü 8 8 §

R o l o t l v * obundonc»

° S 8 8 S 8 8 8 8 8 ê

00

83 10 cm resemble spectra of Taxodium peats from different locations (Stout et al., 1988; J.J. Boon, priv. comm.). The pyrolysis mass spectra of the peat samples from 183-193 and 244-254 cm resemble those recorded for a Mariscus/ Nymphaea peat from a different location (chapter 3). The markers for lignin are for the greater part of the guaiacyl type (m/z 124, 138, 150, 164, 180). Saiz-Jimenez and de Leeuw (1986) and Saiz-Jimenez et al. (1987) report the occurrence of lignin exclusively of the guaiacyl type in pyrolysates of modern and buried spruce, and generalize this phenomenon to be diagnostic for gymnosperms. This conclusion seems to be supported by the present results. Curie point pyrolysis of lignin from several angiosperms has revealed the presence of both syringyl and guaiacyl derivatives, but grasses appear to yield mainly non-methylated phenol derivatives (e.g. Saiz-Jimenez and de Leeuw, 1986; Saiz-Jimenez et al., 1987). The presence of markers for guaiacyl/syringyl lignin in the pyrolysis mass spectra is in agreement with the microscopic observations of Carex and Nymphaea tissues in the lower

I J,„ II ,1. h

. u jIUL

.i. ,t

N

i ,U .

UI

OJ

.ii.m A.1J .iLiilji iiln,iLink

'M

I)

«

mj .'.Mi

Fig. 4.4 Mass spectra of the positive and negative sides of the first (Dl) and second (D2) discriminant functions that describe the variance in pyrolysis mass spectra. (opposite page) Fig. 4.3 Pyrolysis mass spectra of four fine grained samples. A: 0-10 cm. B: 112-122 cm. C: 152-163 cm. D: 244-254 cm.

84

three peat samples. The presence of markers for syringyl moyeties (m/z 154, 168, 194, 210) in the upper three samples must be caused by admixture of angiosperm material in the cypress swamp and hence in the peat. So these samples do not represent a pure Taxodium swamp. Modification of angiosperm lignin by defunctionalization would result in a shift from syringyl (=dimethoxy-phenol derivatives) to guaiacyl (=monomethoxyphenol derivatives) moieties (Stout et al., 1988). This is observed indeed with increasing depth of the samples. Modification of lignin together with a change in vascular plant precursors might also explain the change with depth of the relative contributions of the guaiacyl lignin markers at m/z 110, 124, 138, 150, 164, 180; at shallower depth peaks at m/z 110, 138, 164 tend to be rather prominent, while the deepest samples show higher contributions of the peaks at m/z 124,150,180 (see Fig. 4.3). With increasing depth in the peat the pyrolysis mass spectra of coarse and fine grained samples show increasing relative contributions of lignin markers and decreasing relative contributions of polysaccharide markers. This indicates that saccharides are preferentially degraded when compared to lignins. The fine grained sample from 152-162 cm shows a deviant behaviour in several aspects and will be discussed later. Multivariate analysis yielded several discriminant functions, of which the first two are depicted in Fig. 4.4. The first discriminant function (Dl) describes 63 % of the characteristic variance (35 % of the total variance) and the second discriminant function (D2) describes 22 % of the characteristic variance (18 % of the total variance). Fig. 4.4 shows the sets of correlated mass peaks that differ either positively or negatively from the overall average spectrum (zero point spectrum). The positive side of the Dl spectrum shows contributions of polysaccharides interpreted as being derived from both microorganisms and from vascular plants- and the negative side shows contributions of guaiacyl/syringyl lignin, which is typically derived from vascular plants. The pattern of the mass peaks in the Dl- spectrum indicates modification of the lignin (Stout et al, 1988). The positive spectrum of the second discriminant function represents markers for both polysaccharides and lignin, and is interpreted as indicating a more or less intact lignin-carbohydrate complex of vascular plants. The negative side of the D2 spectrum contains mass peaks that are tentatively interpreted as markers for (terpenoid) hydrocarbons (m/z 67, 69, 79, 81, 82, 93, 95, 97,

85

109, 111, 121, 135, 136, 191). These markers might point to the presence of resins in the peat samples. Fig.'s 4.5a and 4.5b show depth profiles for the first two discriminant functions. Plots of the first dicriminant function with depth (see Fig. 4.5a) show that the coarse grained samples tend to show more guaiacyl/syringyl characteristics and that the fine grained samples tend to show more carbohydrate characteristics. The coarse grained samples are thought to contain a much greater proportion of vascular plant material than the fine grained samples (see discussion above) which explains why the coarse grained ones plot more on the negative side of the Dl function. This is not in contradiction with the wet chemical carbohydrate data, because the fact that the fine grained samples show more carbohydrate characteristics than the coarse grained samples does not say anything about the absolute amounts of carbohydrates present in these fractions. The general trends with depth probably reflect degradation: Polysaccharides degrade more easily than lignins and a relative

A. DF1 J

I

1 I

B. DF2 1 I

L—I

1 I

■ -80

I

I

I

I

L_l

rnesti

I

■

+

I

I

I

L

L_J

L

+20 mesh

Fig. 4.5 Depth profiles of scores on the the Dl and D2 functions. D1+ is loaded with with carbohydrate markers and Dl- with markers for guaiacyl/ syringyl lignin. D2+ is loaded with markers for a lignin-carbohydrate complex and D2- with markers for resinous material.

86 loss of polysaccharides from the peat is therefore accompanied by a relative preservation of lignins (Hatcher et al., 1981; Hedges et al., 1985; Wilson et al., 1987). Fig. 4.5b shows that the coarse and fine grained samples from the Taxodium peat -i.e. the samples from the upper three levels- show increasingly more terpenoid characteristics with increasing depth. This can be explained by a relative (i.e. in comparison to the carbohydrate-lignin complex) accumulation of recalcitrant resinous material with increasing depth. Relative resistance to degradation (or selective preservation) of resins is in agreement with higher scores of the fine grained samples on the negative side of the D2 function and with a generally more degraded character of the material in the fine grained samples. Vascular plants like Carex and Nymphaea do not contain resins, which explains why the peat samples from the lower three levels hardly show any differentation. Additional observations The trends with depth for total sugar yields (see Table 4.1) from the peat samples agree well with the trends observed for the first discriminant function (see Fig. 4.5). An initial sharp decrease with depth in polysaccharide content in the coarse grained fraction foliowed by a much slower decrease has been observed by Rhoads (1985) in Fourier-transformed infrared (FT-IR) spectra. This author also observed an initial decrease with depth in polysaccharide content in the fine grained fraction foliowed by an unexpected high carbohydrate yield in the sample from 152-163 cm. The results from these three different analytical techniques appear to be in good agreement. The fine grained sample from 152-163 cm behaves strangely when compared with the other samples: It shows a high total sugar yield and at the same time rather "normal" relative contributions of microbial sugars (see Table 4.1 and Fig. 4.1). The pyrolysis mass spectrum of this sample (see Fig. 4.3) shows relatively high contributions of polysaccharide markers and low contributions of lignin markers, which is confirmed by the high score of this sample on the positive side of the first discriminant function (see Fig. 4.5a). At this depth the peat also contains twice as much sulphate (from 35 to 73 ppm) as at other depths, while the concentrations of other sulphur species remain low over the whole length of the core (Casagrande et al., 1977).

87

These results may be related to the change in environment from open marsh to Taxodium swamp. It is speculated that such a transition is accompanied by oxygenation of the surface peat layers at that time and subsequently higher microbial activity. This high activity could result in degradation of deeper peat layers so that the overall result would be a highly decomposed top peat layer with a high amount of fine grained material and at the same time a high contribution of microbial sugars. On top of this material Taxodium peat would be deposited. Taxodium is rich in phytotoxins (Given and Dickinson, 1975), which could protect the peat layer in question from further extensive microbial degradation.

4.5 CONCLUSIONS 1) The coarse and fine grained sample preparations show contributions of vascular plants, indicated by high contributions of glucose and xylose, and of microorganisms, indicated by high contributions of partially methylated sugars, amino sugars, heptoses, tetroses and rhamnose, fucose, ribose, mannose and galactose. The contribution of sugars from microorganisms and consequently of microorganisms to the fine grained fraction is much greater than to the coarse grained fraction. This is connected a.o. with higher overall sugar concentrations in the coarse grained fraction. 2) In the coarse grained samples Taxodium can be distinguished from Carex and Nymphaea on basis of higher relative contributions of xylose to the former. From Py-MS data it appears that the Taxodium peat is not pure but shows an angiosperm contribution. 3) Factor analysis of the minor sugars shows that differences exist between the microbial populations of the fine and coarse grained samples. The fine grained samples show very little variation among themselves. 4) Most of the degradation of polysaccharides appears to be restricted to the upper levels of the peat. 5) Factor discriminant analysis of the Py-MS data yielded a first discriminant function that describes the relative accumulation of lignin when compared to polysaccharides with increasing depth in the peat. The second discriminant function differentiates between terpenoid hydrocarbons and a lignin- carbohydrate complex. This is interpreted as selective preservation of resinous material derived from Taxodium, in the fine grained samples.

88 6) The change in environment from open marsh to Taxodium swamp is thought to be reflected by deviant carbohydrate concentrations and Py-MS results of the fine grained sample at the depth of 152-163 cm.

Acknowledgements. Financial support of the first author by the Koninklijke/ Shell Exploratie en Produktie Laboratorium (Shell Research BV), Rijswijk, is gratefully acknowledged. The peat samples were made available by Drs NJ. Ryan and P.H. Given, whom we hereby wish to thank. Mrs B. Brandt-de Boer and Mr G. Eijkel are thanked for performing Py-MS and multivariate analyses at the FOM Institute AMOLF, Amsterdam. Dr. H.C. Cox is thanked for critically reading the manuscript.

89 5. ANALYSIS OF NEUTRAL SACCHARIDES IN FEN PEAT AND BOG PEAT SAMPLES FROM THE ASSENDELVER POLDERS (THE NETHERLANDS)

M.E.C. Moers, J.W. de Leeuw, J.J. Boon, D. van Smeerdijk and F. Hans-Hans

5.1 ABSTRACT This paper describes the analysis of neutral sugars in acid hydrolysates of nine peat samples from the Assendelver Polders. The samples were taken from different depths in a core and comprise different peat types. Variations in the total sugar yield and in the contributions of xylose, arabinose, glucose and rhamnose allowed differentiation between the various plant communities and consequently between various peat types: Eutrophic fen peats with high contributions of monocotyledons such as grasses show rather low total sugar yields with relatively high contributions of xylose and arabinose and generally low contributions of glucose. A peat sample rich in Eriophorum (cotton grass) showed very high contributions of arabi­ nose. Oligotrophic bog peats, on the other hand, which are rich in dicotyledons such as heathers and bog myrtle and may contain mosses show rather high total sugar yields with a relatively high contribution of glucose and relatively low contributions of xylose and arabinose. The moss containing bog peats showed relatively high contributions of rhamnose and mannose. The presence of a number of sugars such as partially methylated sugars, heptoses, amino sugars, tetroses and a few others -collectively called minor sugars- is attributed to the presence of microorganisms. Variations in the the contributions of these minor sugars as revealed by factor analysis allowed differentiation between various microbial communities in the peat samples: One group of minor sugars was correlated with the presence of hyphae of fungi living in symbiosis with heather rootlets. Another set of minor sugars was correlated with peats rich in Eriophorum or Phragmites. Other groups of minor sugars are thought to reflect various bacterial communities.

90 5.2 INTRODUCTION This paper describes the occurrence of neutral sugars in hydrolysates of peat samples from the Assendelver Polders, a location in North-Holland, the western Netherlands. The core from which the samples were taken had a length of ca. 110 cm, with the oldest strata dated at 2346 y. B.P. and the youngest strata dated at 1133 y. B.P. as determined by 14C-dating in combination with calibration against tree rings (van Smeerdijk, 1989). The core under consideration has been studied before by pollen analysis, macrobotanical analysis, conventional element analysis and by pyrolysis-mass spectrometry, pyrolysis-gas chromatography and pyrolysis-gas chromatographymass spectrometry (van Smeerdijk and Boon, 1987; van Smeerdijk, 1989; van Smeerdijk and Boon, in prep.). From botanical studies (van Smeerdijk, 1989) it foliowed that the peat in the lower half of the core, i.e. from 305 to 255 cm NAP (see Table 5.1), represents a terrestrialization sequence during which the environment became increasingly oligotrophic. The peat consists of material derived mainly from remains of Phragmites australis (reed) at the base and of Molinia caerulea (purple moor grass), Myrica gale (bog myrtle) and Ericaceae (heathers) at higher levels. The peat in the upper half of the core, i.e. from 255 to 195 cm NAP (see Table 5.1), has a more homogeneous composition and can be described as a raised bog with major contributions of remains of Ericaceae (heather), Sphagnum (peat moss), Eriophorum vaginatum (cotton-grass) and Myrica gale (van Smeerdijk, 1989). Py-MS, Py-GC and Py-GC-MS studies (van Smeerdijk and Boon, 1987; van Smeerdijk and Boon, in prep.) have shown that the different peat types give different carbohydrate and lignin responses. These pyrolysis data revealed that the carbohydrate/lignin signatures of the various vegetational types down the core predominated over the decomposition pattern. Quantitative analysis and speciation of carbohydrates provides a good opportunity to study the polysaccharide material in this core on a more detailed level. In the present study it will be investigated how the various vegeta­ tional communities are reflected by concentrations and by distribution patterns of neutral saccharides and to which extent transformation and/or mineralization of carbohydrates is taking place.

91 The samples for the present investigation were selected with care: Based on the results of the paleobotanical and pyrolysis studies mentioned above, fen peat and bog peat samples representing different paleo-ecological environments were selected. Table 5.1 lists data relevant to the peat sections analysed in the present study.

5.3 EXPERIMENTAL Nine deep frozen peat sample sections of ca. 1 cm thickness from several depths in the core and handpicked remains of Eriophorum vaginatum (cottongrass) from 216 cm below NAP were provided by D.G. van Smeerdijk. They were freeze dried and pulverized. Each sample was analysed in duplicate. For analysis of neutral saccharides subsamples of 100 mg material were immersed in 5 ml 12 M H 2 S0 4 for 2 h at room temperature. The acid was diluted to 1 M and the samples were hydrolysed for 4.5 h at 100° C. Next myo-inositol was added as an internal Standard and the acid supernatant with released monosaccharides was neutralised by BaC0 3 . Reduction of aldoses to alditols was achieved by NaBH4 (16 h, room temperature). Boric acid was removed by methanol and the alditols were acetylated to alditol acetates by addition of acetic anhydride and pyridine and heating the mixture for 3 h at 100°C. The alditol acetates were analysed by gas chromatography (GC) on a Carlo Erba Fractovap 4160 gas chromatograph and by gas chromatography-mass spectrometry (GC-MS) on a Hewlett-Packard 5890 gas chromatograph coupled to a VG 70-250SE mass spectrometer. A fused silica capillary column coated with CPsil88 (1=25 m, i.d.=0.32 mm, df=0.12 /jm; Chrompack, Middelburg, The Netherlands) was used for the gas chromatographic separations. For details concerning the derivatization procedures and the GC and GC-MS conditions the reader is referred to chapter 3. Quantification of the alditol acetate yields was performed with the aid of a Maxima Chromatography Workstation (Dynamic Solutions Corporation, Ventura, C USA), by peak area integration of responses of the flame ionization detector that was coupled to the gas chromatograph. Responses of alditol acetates were compared to the response of myo-inositol hexaacetate (the internal Standard) on an equal weight basis. In some cases when components coeluted quantifi­ cation was carried out on GC-MS results by integration of the responses of

92 selected diagnostic mass fragments (m/z-values) in mass chromatograms using the software available with the mass spectrometer.

Table 5 . 1 Data r e l e v a n t to the peat samples analysed i n the present wt% C

Depth

Botanical composition

cm

of recognizable remains

-NAP

In peat fraction

Hyphae age org of dry m/cms years * peat wet B.P.

wt% of dry peat >

> 190 urn Ericaceae, (sub)aerial Sühapnum Mvrica. leaves. wood others monocot roots unldentified

45% 20% 15% 8% 7% 5%

47

40

1107

190 urn' 30

215

Ericaceae, (sub)aerial Erior>horum. stem bases unldentified Mvrlca leaves

65% 25% 5% 5%

47

111

1234

15

235

Ericaceae, roots, wood unldentified Sphapnum others

65% 25% 5% 5%

47

170

1470

18

255

unldentified Ericaceae, roots Mollnla. epidermis monocot roots charcoal

30% 20% 20% 20% 10%

49

52

1704

23

276

monocot roots Molinia epidermis unldentified Mvrica. leaves Ericaceae, roots

50% 20% 15% 10% 5%

51

45

1943

17

204

tieat

93

Continuation of table 5.1 Depth

Botanical composition

wt% C org

Hyphae Jr

age 6

cm

of recognizable remains

of dry

m/cms

years

-NAP

wtX

*

in peat fraction

peat

wet

of dry

B.P.

peat >

> 190 urn Ericaceae, (sub)aerial monocot roots Aulacomnium unidentlfied others

60% 15% 10% 10% 5%

50

68

2043

190 urn 15

289

monocot roots Molinia epidermis Ericaceae, roots unidentified others

40% 30% 15% 10% 5%

47

95

2101

31

301

Phrapmites epidermis monocot roots.. unidentified

40% 40% 20%

42

10

2241

304

Phragmites epidermis monocot roots

50% 40%

42

284

peat

2277

12

10% unidentified *** C14-age calibrated against tree rings. The wt% fraction >190 pm is determined on the basis of dry, ash-free peat. ***• These two peat samples contain seeds of ca. seven different dicot species. It is assumed therefore that a substantial part of the 'unidentified' material is derived from dicot material.

*: **•

94 Identification of the alditol acetates was achieved by comparing the mass spectra with those known from the literature (Stoffel and Hanfland, 1973; Schwarzmann and Jeanloz, 1974; Jansson et al., 1976; Radziejewsky-Lebrecht et al, 1979; Wong et al, 1980; Klok et al., 1982) and in nearly all cases also by comparison of the relative retention times with those of Standard mixtures of (partially methylated) alditol acetates (Klok et al., 1982,1984a). From the duplicate analyses it was estimated that the errors in the determinations of the yields of the major sugars are 10% or less. However, errors in the determinations of some minor components that were present in quantities smaller than 0.02 mg/g may be as large as 50%. Normalized data of the 34 so-called minor sugars were subjected to factor analysis using a modified ARTHUR computer package (Infometrix, Seattle, WA, USA). The principles and application of this procedure are described by Windig et al. (1982).

5.4 RESULTS AND DISCUSSION Total sugar yields Table 5.2 shows results of the analyses of neutral saccharides. From this table and from Fig. 5.1 it can be seen that there is a general decrease in total carbohydrate yield (both in mg/g and corrected for variation in the organic carbon content) with increasing depth in the sediment. This picture of the total sugar concentrations is largely determined by the glucose yields which constitute from 35 to 60 wt% of the total sugar yields (see Fig. 5.2). The other sugars show quite different patterns as can be seen from Fig. 5.2. This suggests that the downward trend in Fig. 5.1 is caused by differences in input of organic matter rather than by increased degradation due to longer residence times (see also below). The quantitative sugar data of the present study can be correlated with the paleo ecology of the peat (Table 5.1): It appears that the upper samples rich in Ericaceae designating bog peat yield rather high amounts of sugars, while the deeper samples rich in grasses indicating fen peat show the lower sugar

95 Table 5.2 Ylelds of neutral saccharldes In mg/g and ylelds carbohydrate carbon as fractlon of total organlc carbon. Depth in cm below NAP: 204 215 235 255 276 284 289 6.4 4.2 1.6 1.2 4.5 3.1 2.6 rhamnose 0.6 0.9 1.3 0.4 1.1 1.0 0.7 fucose 0.5 0.6 0.5 0.4 0.4 0.5 0.4 ribose 5.9 13.6 3.7 3.6 5.4 4.4 2.7 arablnose 14.8 14.3 19.6 21.1 22.2 11.1 25.5 xylose 11.8 7.4 10.9 3.2 10.1 12.2 6.1 mannose 14.8 10.5 12.0 7.8 5.2 14.1 18.6 galactose 77.4 84.7 56.1 47.9 50.8 93.0 70.7 glucose 6.0 5.4 3.9 2.3 6.8 2.8 2.7 minor conro. 132.2 170.3 141.7 118.8 116.6 83.2 93.3 SUM Carbohydrate carbon as wt% of total organlc carbon: 11.3 14.5 12.1 9.7 9.1 6.7 7.9

of

301 1.7 0.7 0.4 5.0 21.9 4.5 7.2 30.6 4.1 76.1

304 1.9 1.1 0.4 6.2 25.8 5.2 8.1 26.7 3.8 79.3

7.3

7.6

total sugars p.

O* ^

'/X

'' // 150-

X

'7'7

X. ^N

X x>> ^

' /

>■

X

■/

X

// +/

N

s X. X. s

X. \ vs X. V X. N

^y~^--^

.

^"-~-ï

100-

1

200

1

220

1

1

240

I

1

260

I

T

280

1

1

300

depth in c m Fig. 5.1 Total sugar yields in mg/g dry peat (H -4-) and total sugar yields corrected for variation in organic carbon contents in (mg x 50)/(g x wt% Corg) (O- - - -Q). '50' is an arbitrary scaling factor.

96 concentrations (see Fig. 5.1). Given and Dickinson (1975) also found higher polysacchande contents for oligotrophic bog peats than for eutrophic fen peats and forest peats. Analysis of Assendelver peats by pyrolysis-mass spectrometry (Py-MS) (van Smeerdijk and Boon, in prep.) revealed that peats from eutrophic environments (i.e. fen peats) are relatively rich in lignin and poor in carbohydrates. Peats from oligotrophic environments (i.e. bog peats) on the other hand are relatively poor in lignin and rich in carbohydrates. These observations are in agreement with the present quantitative sugar data.

fucose

220

240

260

280

300

200

220

240

260

280

20

240

260

280

300

220

00

240

260

280

300

glucose

h\

220

00

\

*

200

300

qqlactose

220

240

260

280

ƒ 300

200

220

240

260

280

300

depth in cm below NAP Fig. 5.2 Sugar yjelds in (mg x 50)/(g x wt% Corg) (O- - - -rj). Relative sugar concentrations in wt% (H O with the total yield of all sugars taken as 100%.

97 Sugar vields in relation to peat type Xvlose and arabinose Inspection of Fig. 5.2 shows that both xylose and arabinose are present in rather large quantities (both absolute and relative) in the lower two samples, i.e. in the reed peat. This is in good agreement with literature data which report relatively high contributions of xylose and arabinose in grasses (e.g. Lomax et al., 1983; Chesson et al., 1985). Linear regression shows a positive correlation between the relative contributions of xylose and arabinose: r=0.85 when the outlier from 215 cm is omitted and r=0.49 when this outlier is included. This suggests the presence of arabinoxylans which are known to be of common occurrence in the hemicelluloses of grasses (Aspinall, 1983). The outlier position of the peat sample from 215 cm suggests a different origin for the arabinose in this sample. The high concentration of arabinose, coupled with a high total sugar yield and high contributions of altrose, erythrose and threose (see discussion below) in the sample from 215 cm below NAP is probably a reflection of the presence of Eriophorum remains in this sample, because analysis of Eriophorum fragments selectively handpicked out of a peat sample from 216 cm below NAP, also showed a high total carbohydrate yield (146 mg/g), a high contribution of arabinose (7.5 wt%) and not insignificant contributions of altrose, erythrose and threose. Another way of relating sugar yields with peat type is by comparing these yields with the abundances of recognizable botanical remains. Fen peats are generally rich in monocotyledons (e.g. grasses), whereas bog peats are often rich in dicotyledons (e.g. heathers) and mosses. The recognizable botanical remains in Table 5.1 were determined in the peat fraction >190 /Jtn. This fraction constitutes only 9 to 31 wt% of the total peat on a dry and ash free basis (see Table 5.1). It is not always legitimate to extend the botancial composition of the fraction > 190 ^m to the whole peat; especially reed peats are known to contain a great variety of dicotyledons -seeds of seven different dicotyledon species could be identified in the reed peats from 301 and 304 cm by van Smeerdijk (1989)-, although no recognizable remains of the plants themselves could be identified. Consequently one has to recalculate the botanical data in the first column of Table 5.1 to the basis of a total, ash free peat before one can make comparisons. This is done in Fig. 5.3a where the relative xylose yields (as wt% of the total sugar yield) are plotted versus the percentage of identifiable monocotyledon remains in the total, ash free

98 peat. A similar figure is obtained when the absolute xylose yield (in mg sugar/g dry peat, corrected for differences in organic carbon content) is plotted (not shown). These results reveal a positive relationship between the xylose yield and the monocot contribution, which is in agreement with the results of the previous alinea. Difference in peat type -fen peat versus bog peat- is illustrated by the deviant position of the two reed peat samples from 301 and 304 cm.

■ 304 3 0

* %

-

■ 301

■

289i

co O

■o 2 0 a>

■ 276

rel.

■ 255 ■ 284

10J

,

235

.20i215 i r i 1 i 1 1 4 8 12 16 % monocot in ash free peat

1

o-

i

i

20

• 204

~ 4a> in O

c b _ o 3i

eld

L.

o ,

C

235

'

2 8 4

■215 ■ 301

«276 "304

rel.

>N 2 "

■ 255

()

4 8 12 16 % monocot in ash free peat

2891 20

Fig. 5.3 The relative yields (in wt% of the total sugar yield) of xylose and rhamnose versus the percentage of identifiable monocotyledon remains in the total, ash free peat.

99 Glucose Table 5.1 and Fig. 5.2 show that the two reed peat samples contain the smallest amounts of glucose. This can be explained in several ways: First, it can be argued that these samples are the oldest ones and that the glucose yields are a reflection of prolonged degradation of all carbohydrate material. This explanation seems unlikely, because other studies (e.g. chapters 2 and 3) have already shown that a time span of a few thousand years does not have a significant effect on sugar concentrations in peats. Moreover, the other sugars show different concentration patterns quite different from the one of glucose as was already mentioned above (see Fig. 5.2). A second explanation might be that glucose is degraded preferentially over the other saccharides. This explanation seems unlikely as well, because it is well known from the literature that glucose is often selectively preserved during diagenesis (e.g. Hatcher et al., 1981; Hedges et al., 1985; Wilson et al., 1987; Stout et al., 1988; chapter 8). A third explanation is that the low glucose yields are inherent to the type of peat deposit, i.e. eutrophic Phragmites peat. Support from literature data comes from Cheshire and Mundie (1966) who analysed sugars in acid hydrolysates of a Phragmites peat (horizon 130/180 cm) and found ca. equal quantities of pentoses and glucose (78 mg/g and 74 mg/g respectively). Our analyses yield lower values (ca. 30 mg/g for both glucose and the pentoses), but in the same ratio. It should be noticed that this type of peat deposit is relatively susceptible to degradation. This can a.o. be deduced from Fig. 5.4, which shows a rela­ tively high contribution of minor sugars and a relatively small fraction of coarse grained materal in the two reed peat samples (see also discussion below). The absence of identifiable remains of dicotyledon plants and the low total sugar yield are related indications of the susceptibility of eutrophic reed peat for degradation. Rhamnose The negative trend in Fig. 5.3b where the relative rhamnose yield (in wt% of the total sugar yield) is plotted versus the percentage identifiable monocot remains in the total peat (on an ash free basis) suggests that (some) plants contributing to bog peats are richer in rhamnose than the plants contributing to fen peats. This is probably partly due to the presence of moss in some of the bog peats, because samples with high relative rhamnose contributions are

100 also the ones containing moss (see Table 5.1 and Fig. 5.2). Moreover, Py-MS studies have already shown that Sphagnum contains considerable amounts of rhamnose (van Smeerdijk and Boon, 1987). It is not likely that the elevated rhamnose levels in the moss/heather-rich peats are due to the presence of hyphae associated with Ericaceae roots, because no correlation was observed between the contribution of rhamnose (or of any other major sugar) and the presence of hyphae. Fig. 5.2 suggests that the contributions of mannose are also influenced by the presence of mosses in the peat samples. Microbial sugars In former studies (chapters 2 and 3) it was noticed that mechanical fractionation of peat samples yielded a coarse grained fraction consisting of less degraded material and a fine grained fraction consisting of more heavily degraded material. The relative contributions of microbially derived sugars were enhanced in fine grained peat fractions. This was explained by degradation of vascular plant material, yielding a more heavily degraded fraction

'301

r=0.93 s=-0.15 s.d=0.02

^ ■304 "276

284

W

o

■235 ■215

tn 4 O

c

'Ë 2v 3 _. ï

"

28Ï "

"255

204 2 -\

8

1

1

12

1

1

16

1

1

20

1

1

1

24

1

28

1

32

peat fraction > 190/urn (wt%)

Fig. 5.4 The relative yield of the minor sugars (in wt% of the total sugar yield) versus the relative amount of peat fraction > 190 /mi (in wt% of the total, ash free peat).

101 predominantly composed of microbially derived sugars with some resistant vascular plant polysaccharides and a less degraded fraction predominantly composed of vascular plant material with a smaller contribution of microbially derived sugars. In those former studies it was concluded that partially methylated sugars, amino sugars, heptoses, allose, altrose, threose, erythrose and glycerol were almost exclusively derived from microorganisms, that galactose, mannose, rhamnose, fucose and ribose were of a mixed microbial and vascular plant source and that arabinose, xylose and glucose were predomi­ nantly derived from vascular plants. In the present study no sugars were separately analysed in mechanically fractionated peat samples, but the principles of plant degradation probably remain the same: The extent of degradation of the peat material in the present case is reflected in the relative amount of coarse grained (>190 /im) material and the relative contributions of microbially derived sugars. Fig. 5.4 shows the presence of a negative correlation between these parameters. This points to a predominantly microbial origin of the 'minor' sugars in the present case as well. Mannose and rhamnose show patterns resembling those of the abovementioned sugars only for the lower five samples (see Fig. 5.2), which suggests that these sugars are derived from mixed sources. The presence of these two sugars in mosses was already mentioned above. Factor analysis of minor sugars Investigation of the minor sugar contributions by factor analysis yielded a number of factors with the first five describing 30, 20, 16, 12 and 9 % of the total variance respectively. Table 5.3 lists the major loadings of the minor sugars on these factors. Sugars loading high on the same factor tended to show similar depth profiles. Therefore the relative contributions of all sugars at a certain depth loading highly on the same factor (see Table 5.3) were averaged. The resulting depth profiles of these averaged contributions are shown in Fig. 5.5. Differences in depth profiles between the groups suggest that they represent different microbial populations. However, little is known about the contributions of these minor components to the various microbial sources, so that it is not (yet) possible to link the various factors to specific populations. Some general remarks can be made, however. For instance, there is the group of

102 Table 5.3 Major loadings of minor components on the first five factors obtained from factor analvsis. var. preserved 30% 20% 16% 12% 9% Fl F2 F3 F4 F5 A a heptose +0.74 +0.10 +0.37 -0.34 -0.32 4-OMe-rham +0.58 -0.14 +0.13 -0.27 -0.55 3/4-OMe-man+0.63 +0.53 +0.13 -0.26 +0.32 3/4-OMe-gal+0.79 +0.15 -0.02 -0.12 +0.40 3-OMe-gluc +0.84 -0.10 +0.38 -0.21 +0.27 4-OMe-gluc +0.53 -0.16 +0.37 -0.62 +0.07 6-OMe-gal +0.42 -0.22 -0.02 +0.41 -0.36 B erythrose threose 2-OMe-fuc 2-OMe-ara 4-OMe-ara allose altrose 2-OMe-rham 2/4-OMe-rib4-OMe-fuc

-0.75 -0.94 -0.75 -0.90 -0.78 -0.74 -0.62 -0.70 -0.82 -0.53

-0.24 -0.14 +0.44 +0.34 -0.02 -0.12 -0.51 +0.41 +0.13 +0.48

+0.12 +0.12 +0.00 -0.19 -0.34 +0.19 -0.26 +0.55 +0.19 +0.54

-0.18 -0.09 -0.25 -0.18 -0.40 -0.26 +0.37 +0.00 +0.13 +0.22

-0.54 -0.27 -0.41 +0.04 -0.04 +0.52 +0.25 +0.11 +0.13 -0.22

C glucoheptose a heptose glucosamine 3-0Me-xyl

-0.12 +0.51 +0.11 -0.47

+0.58 +0.35 +0.35 +0.73 -0.20 -0.37 +0.65 -0.37 +0.62 +0.56 +0.49 +0.00

+0.44 -0.13 -0.18 -0.28

D 6-OMe-gluc 2/5-OMe-man2/5-OMe-gal2-OMe-gluc

-0.07 +0.16 -0.09 -0.27

-0.77 +0.14 -0.78 +0.00 -0.82 +0.19 -0.90 +0.17

E 2/4-OMe-xyl3-OMe-rham an amino sugar

-0.47 +0.34 +0.71 +0.31 +0.35 -0.33 +0.78 +0.14 +0.25 -0.22 +0.63 -0.19

F glycerol 6-OMe-man an amino sugar

-0.23 -0.40 +0.22 +0.05 +0.12 +0.53

-0.65 -0.24 +0.16 -0.68 +0.44 -0.43 -0.75 +0.02 -0.01

G a heptose

+0.17

+0.34

+0.46

+0.76

H 3-OMe-ara 3-OMe-fuc

-0.25 -0.44

-0.14 -0.17

+0.47 -0.35

-0.57 +0.35 -0.52 -0.42

+0.31 +0.29 +0.50 +0.06 +0.05 -0.20 +0.08 +0.02 +0.21 -0.34 -0.49

-0.19

~: Enantiomeric a l d i t o l acetates are not separated on CPsil88.

103 sugars showing high loadings on the negative side of the fïrst factor (group B in Table 5.3 and Fig. 5.5). These sugars show rather high contributions in the two reed peat samples (301 and 304 cm), and together with arabinose also in the sample from 215 cm which shows a specific contribution of Eriophorum (Fig.'s 5.2 and 5.5). This suggests the presence of a microbial population especially related to these peats. Another feature worth mentioning is the positive correlation observed between the presence of Ericaceae and the length of fungal hyphae in the peat (van Smeerdijk, 1989) and the contributions of sugars loading highly on the negative side of the second factor (group D in Table 5.3 and Fig. 5.5). The positive correlation between the presence of heather rootlets and the amount of fungal hyphae suggests a mycorrhizal origin (Hawker and Iinton, 1972; van Smeerdijk, 1989). It seems likely therefore that relatively high contributions of 6-O-methyl-hexoses and 2- and 5-O-methyl-hexoses (i.e. sugars loading highly on F2-, group D in Table 5.3 and Fig. 5.5) are related to the presence of these mycorrhizae. Bacteria are known to contain a large variety of partially methylated sugars, heptoses, amino sugars, and also fucose, galactose, mannose, rhamnose, tetroses and glycerol (e.g. Weckesser et al., 1979; Laskin and Lechevalier, 1982; Aspinall, 1983; Klok et al, 1984a,b). It is suggested therefore that the minor sugars showing high loadings on F1+, F2+, F3+ etc. reflect different bacterial populations associated with the various vegetational communities.

fc^0.15 A -1—'

§

A,

0.1

^ 0.05

/

2(10

A/V

V

240

O.B

A

c

0.4 0.2

/ ' \

200

280

240

/AlV 280

2Ó0

ï

1 -A-V

0.05 200

240

280

\ \

»

240

0.1

0.15

\

/

/

0.05 2
0.1

9-s\

0.15

/

A

/ V1/

V

280

F

0.06

0.1

0.04

0.05

0.02

2J)0

240

280

f/ \ /

2^X)

\ 240

V/

V

r

v

280

depth in cm below NAP Fig. 5.5 Averaged relative yields (in wt%) of groups of minor sugars versus depth. The sugars are divided into eight groups according to their major loadings on the factors obtained from factor analysis as listed in Table 5.3. The letters A through H in the figure correspond to those in Table 5.3.

104 Special cases The carbohydrate yields of the sample from 289 cm below NAP are rather peculiar; compared with the other fen peats samples this sample shows the expected contnbutions of vascular plant derived sugars (glucose, arabinose, xylose), but surprisingly low contnbutions of most minor sugars, rhamnose, fucose, mannose and galactose, i.e. of sugars whose occurrences are attributed predominantly to microorganisms (see Fig.'s 5.2 and 5.5). These rather low contnbutions of 'microbial' sugars might be related to the low nitrogen content (ca. 0.9 wt%) and high C/N-ratio (ca. 54, data after van Smeerdijk, 1989) of the peat at this depth. C/N-ratios are used to determine rates of decomposition (van Smeerdijk and Boon, in prep.). It has been suggested (Given and Dickinson, 1975) that nitrogen can be a limiting nutriënt in peat forming environments, which indicates that a lack of nitrogen may have hampered the development of a "normal" bacterial community (Clymo, 1983). The low yields of microbial sugars, the low organic nitrogen content and other peculiarities such as a low concentration of pollen and rather high concentrations of fungal hyphae and spores (JJ. Boon, priv. comm.) might all be a reflection of the habitat of Molinia. These phenomena would then be of very local occurrence as the present sample probably is taken from a Molinia tussock in an otherwise 'normal' fen peat. Another case worth of extra attention is the sample from 284 cm below NAP. This sample is taken from a peat in the transition from a Molinia peat to an oligotrophic bog peat and more specifically from a local depression between the tussocks. This sample shows a relatively low total carbohydrate con­ centration (see Fig. 5.1), which is even more conspicuous when one takes into account that remains of heathers constitute a substantial part of the identifiable botanical remains in the fraction >190 pm (see Table 5.1). The low total sugar yield, together with the relatively high contribution of microbial sugars and the relatively small fraction of material >190 nm (see Fig.'s 5.2 and 5.4) are indications of enhanced microbial degradation. This might be explained in the context of the depositional environment which suggests that the organic material in this peat sample is derived from de-grading plant remains that have been washed in from the tussocks.

105 5.5 CONCLUSIONS 1) Variation in the total carbohydrate yield and in the relative contributions of glucose, xylose and arabinose make it possible to distinguish between the various peat types: Bog peats rich in mosses and in dicotyledons such as heathers show rather high total sugar concentrations with high contributions of glucose and rather low contributions of xylose and arabinose. Fen peats which are rich in grasses on the other hand yield lower amounts of total sugar with high contributions of xylose and arabinose and often low contributions of glucose. More detailed inspection shows that a peat rich in Eriophorum (cotton-grass) yields very high amounts of arabinose. 2) Relatively high contributions of rhamnose (and mannose) are indicative for the presence of mosses in the peats. 3) The decrease in total sugar yields with depth can for the most part be explained by variation in input of vascular plants. Oligotrophic reed peats seem to be more susceptible to microbial degradation than the other peat types. 5) Microbial input to the peat is reflected by high contributions of amino sugars, heptoses, partially methylated sugars, altrose, allose, fucose, ribose, galactose, tetroses, and glycerol. High contributions of mannose and rhamnose in the present case are not only due to microorganisms but also to mosses. 6) Factor analysis applied to the minor sugar data yields several factors, with each factor presumably describing a different microbial population: High contributions of 6-O-methyl-glucose and 2- and 5-O-methyl-hexoses are correlated with the presence of fungi living in symbiosis with heather rootlets. A microbial population with high contributions of altrose, allose, erythrose, threose, 2-O-methyl-rhamnose, 2- and 4-O-methyl-fucose, 2- and 4-O-methylarabinose and 2- and/or 4-O-methyl-ribose is correlated with the presence of Eriophorum and Phragmites in the peat. The other factors obtained from factor analysis probably reflect various different bacterial populations.

Acknowledgements. The Koninklijke/Shell Exploratie en Produktie Laboratorium (Shell Research BV), Rijswijk is gratefully acknowledged for financial support of the first author. The authors thank Mr. G. Eijkel (FOM-Institute AMOLF, Amsterdam) for performing the factor analysis.

107 6. ANALYSIS OF NEUTRAL SACCHARIDES IN MARINE SEDIMENTS FROM THE EQUATORIAL EASTERN ATLANTIC (KANE GAP)

M.E.C. Moers, M. Baas, J.W. de Leeuw and P.A. Schenck

6.1 ABSTRACT Sediment samples from a core spanning the last 850,000 y. procured from the equatorial east Atlantic were analysed for neutral saccharides. The results show that carbohydrate carbon constitutes from 3.0 to 5.8 wt% of the total organic carbon. No apparent degradation of carbohydrates is observed with increasing depth in the sediment. The great variety of saccharides present, the rather uniform relative contributions of the major sugars and the substantial contribution of minor sugars indicate that the major part of the carbohydrates is derived from marine organisms, i.e. algae, protozoa and bacteria. The contribution of carbohydrates derived from terrestrial plants is negligible. Variations in relative saccharide contributions with depth indicate changes in the composition of the microbial assemblages. Covariance of TOC contents and the relative contributions of xylose, 6-O-methyl-mannose and 3- and/or 4-0methyl-mannose indicate that periods of increased primary productivity are associated with the presence of marine organisms, presumably cyanobacteria or marine algae, characterized by relatively high contributions of xylose and of 6-O-methyl-mannose and 3- and/or 4-O-methyl-mannoses. Correlations between the alkenone unsaturation index, 5 1 8 0 and the relative contribution of glucoheptose indicate that higher sea-surface temperatures are associated with a restricted contribution of bacterial saccharides.

108 62 INTRODUCTION This paper describes the occurrence of neutral saccharides in samples from Kane Gap, which is located in the Atlantic Ocean off the coast of Guinea (approximate position 9°N,19°W). The aim of this investigation was to study the occurrence of neutral sac­ charides in sediments deposited in a marine environment and to investigate how differences in input of organic material and time of burial are reflected in the absolute and relative sugar concentrations. Table 6.1 summarizes sample depths, ages and other relevant data. The organic matter in the sediments under investigation is derived from several sources like algae, protozoa, bacteria and vascular plants (e.g. Muller et al., 1983; Brassell et al., 1986b). The results of lipid analyses indicate that the major input of primary organic material is marine in the form of algae (e.g. Prymnesiales and Dinoflagellates), protozoa (Foraminifera) and cyanobacteria (Brassell et al., 1986a,b). This material is possibly partly reworked by bacteria, either in the water column or in the sediment, so that these may constitute an additional source of carbohydrates. The distance of the sample location from the shore precludes a contribution to the sediment of terrestrial organic matter transported by rivers. Therefore, any terrestrially derived material present in the sediment, will represent eolian fall-out (Muller et al., 1983). The samples in this set have been taken from shallow depths in the sediment (max. 12 m). Consequently, they have not been subjected to raised temperatures and thus form an excellent series to study the occurrence of carbohydrates in recent sediments with a mixed source of organic material, while the effect of abiotic diagenesis can be neglected.

63 EXPERIMENTAL The samples were taken from core M16415-2, collected during Meteor cruise 65 in 1983 (Sarnthein et al, 1984; Brassell et al., 1986a). The samples used for carbohydrate analysis in the present case had been extracted ultrasonically with dichloromethane and methanol and dried in an oven at 60°C in the course of a previous investigation (Marlowe, 1984; Brassell et al., 1986a). This extraction does not seem to influence the sugar yields, because no differences

109 in yields of neutral sugars (except glycerol) were observed when, as a check, marine samples from the Mediterranean were analysed for neutral saccharides one time with and one time without previous extraction and subsequent oven drying. The sediment samples were first treated with 1 M HC1 for removal of carbonates. The carbonate-free samples were washed with bidistilled water till neutral and freeze dried. By this procedure soluble salts were removed as well. The carbohydrate analyses were carried out in duplicate. Aliquots of 1 to 2 grams of dry decarbonated and desalted sediment were immersed in 30 ml 12 M H 2 S0 4 at room temperature for 2 h. The acid was diluted to 1 M and the polysaccharides were hydrolysed at 100° C for 4.5 h. Myo-inositol was added as an internal Standard. The acid supernatant with released monosaccharides (aldoses) was neutralized with BaCOs and centrifuged. The pH was raised to 9 by addition of tri-ethylamine. The solution was left to stand for 30 min to allow hydrolysis of lactones. The neutral aldoses were reduced to alditols by NaBH4 (16 h, room temperature). Boric acid was removed with methanol (5 min, 100° C, repeated five times with fresh methanol) and remaining water was removed with acetic anhydride (1 h, 100°C). The alditols were acetylated to alditol acetates with acetic anhydride/pyridine (1/1 v/v; 3 h, 100°C). The reaction mixture with alditol acetates was first dried by rotary evaporation and then in vacuo over KOH and P 2 O s . A few mi's of water were added to the reaction mixture and it was extracted three times with dichloromethane. The combined dichloromethane extracts were dried on anhydrous Na 2 S0 4 and by rotary evaporation under reduced pressure. The alditol acetate mixture was taken up in ethyl acetate and purified over a small Si02-column. The alditol acetates were analysed by gas chromatography (GC) on a Carlo Erba 4160 gas chromatograph and by gas chromatography - mass spectrometry (GC-MS) on a Varian 3700 gas chromatograph coupled to a Finnigan-MAT 44 mass spectrometer or on a Hewlett-Packard 5890 gas chromatograph coupled to a VG 70-250SE mass spectrometer. Gas chromatographic separations were carried out on a CPsil88 fused silica capillary column (1=25 m; i.d.=0.32 mm; df—0.12 /mi; Chrompack, Middelburg, The Netherlands). All three gas chromatographs were equipped with a septumless on-column injector. For GC hydrogen was used as

110 carrier gas. The samples were injected at 70° C, the oven temperature was rapidly raised to 150° C and further programmed at 3° C/min to 230° C and maintained at this temperature for 40 to 50 min. A flame ionization detector was used for detection. For GC-MS helium was used as carrier gas. The temperature was approximately programmed as described above for GC. The mass spectrometer was operated at 70 eV in the electron impact mode and the source temperature was 250°C. Quantification of the alditol acetates was performed by peak area integration of the GC-responses with the aid of a Maxima Chromatography Workstation (Dynamic Solutions Corporation, Ventura, CA USA). Responses were compared to the response of the internal Standard (myo-inositol hexaacetate). The responses of all alditol acetates were assumed to be equal on a weight basis. In a few cases when compounds coeluted quantification was achieved by peak area integration of GC-MS-responses of selected characteristic mass fragments, using the VG-software available in case the VG was used, or software developed "in-house" in case the Finnigan-MAT was used. Identification of the alditol acetates was achieved by comparison of the mass spectra with those known from the literature (Stoffel and Hanfland, 1973; Schwarzmann and Jeanloz, 1984; Jansson et al., 1976; Radziejewsky-Lebrecht et al., 1979; Wong et al, 1980; Klok et al., 1982) and from comparison of the relative retention times with those of Standard mixtures of (partially methylated) alditol acetates (Klok et al, 1982,1984a). The results of the duplicate analyses indicate that the error in the determination of the sugars is 15% or less for the main components and 20% or less for the minor components. The yields of glycerol showed variations up to 200% in the duplicate analyses. These data have therefore been excluded from the discussion. It is speculated that this lack of reproducibility is connected with degradation of glycerol containing compounds (like triglycerides) during the methanol/dichloromethane extraction and non-quantitative release of glycerol from these degraded compounds during acid hydrolysis.

111 6.4 RESULTS AND DISCUSSION In all samples about 40 different alditol acetates could be identified. These include non-methylated, mono-O-methylated and di-O-methylated neutral monosaccharides of varying carbon number, including glucosamine. Uronic acids are not detected by the method used. The sugar yields are summarized in Table 6.1. It follows from this table that the non-methylated hexoses, pentoses and 6-deoxyhexoses form the main components in the hydrolysates. They comprise from 87 to 89 wt% of the total Table 6 . 1 General Informatlon and sugar y i e l d s (mg/g) of t h e samples from Kane Gap, c o r e M16415-2. depth (cm) 15255435136375596925124517 138 257 437 598 377 927 1247 a age 3 (10 y. BP.) 9 TOC (wt%)

b

108

205

284

317

417

653

848

0.42

0.38

0.99

0.67

0.69

0.54

0.55

0.50

5 13 C ( % o ) b CaC03 + salt (wt%) c

-19.0

-19.1

-17.8

-18.2

-18.6

-20.4

-18.5

-18.9

66

27

25

27

50

24

38

20

_ d stage

1

5

6

8

9

11

16

24

0.036 0.048 0.053 0.095 0.117 0.105 0.188 0.141 0.101 0.883

0.040 0.043 0.041 0.064 0.076 0.084 0.120 0.100 0.084 0.652

0.025 0.029 0.033 0.053 0.064 0.077 0.100 0.076 0.055 0.511

0.019 0.026 0.030 0.049 0.047 0.053 0.087 0.072 0.048 0.428

0.027 0.034 0.030 0.053 0.066 0.056 0.112 0.078 0.066 0.521

0.023 0.026 0.018 0.043 0.041 0.044 0.078 0.060 0.044 0.376

c sugar yields in mg/g rhamnose 0.032 0.028 fucose 0.053 0.034 rlbose 0.030 0.035 arablnose 0.062 0.054 xylose 0.074 0.054 mannose 0.085 0.061 0.111 0.113 galactose glucose 0.096 0.091 minor comp. 0.066 0.055 SOM 0.609 0.524

3.0 3.2 3.8 wt% COHC/TOC6 5.8 5.5 3.6 3.9 3.0 a : The ages a r e t a k e n from Marlowe (1984). They a r e based on t h e "Carpor" t i m e s c a l e developed from d a t a of c o r e M13519 from the nearby S i e r r a Leone Rise ( S a r n t h e i n e t a l . , 1984). b : Determined on d r y , decarbonated and d e s a l t e d samples. 5 13 C measured w i t h r e f e r e n c e t o PDB. c : Determined on dry sediment. d: Oxygen i s o t o p e s t r a t i g r a p h y b a s e d on d a t a of core M13519, S i e r r a Leone Rise ( S a r n t h e i n e t a l . , 1984). e : Gram c a r b o h v d r a t e carbon (COHO per 100 g t o t a l o r g a n i c carbon (TOC)

112 Table 6.2 Minor supars 1 erythrose 2 chreose 3 2-OMe-rhamnose 4 2-OMe-fucose 5 3-OMe-rhamnose 6 3-OMe-fucose 7 4-OMe-rhamnose 8 4-OMe-fucose 9 2/4-0Me-rlbose10 2-OMe-arablnose 11 4-OMe-arablnose 12 2/4-OMe-xylose13 3-OMe-arablnose 14 3-OMe-xylose

Identlfled. 15 6-OMe-mannose 16 6-OMe-galactose 17 6-OMe-glucose 18 2/5 - OMe -mannose19 2/5-OMe-galactose20 2-OMe-glucose 21 3/4-OMe-mannose22 3/4-OMe-galactose23 3-OMe-glucose 24 4-OMe-glucose 25 allose 26 altrose 27 a heptose 28 glucoheptose 22- plucosamlne Enantlomerlc aldltol acetates are not separated on a CPsll88.

minor comDonents

fucosc

15 X l

O O

en

10

\

V \

/

\

depth in m Fig. 6.1 Sugar yields versus sample depth. o o = relative yield in g sugar per 100 g total sugars. +- - -+ = mg sugar per g organic carbon.

113 Table 6.3 Di-0-methyl sugars and two i n samples from Kane Gap. Sample depth (cm) Name 15136- 25517 138 257 + + 2,3-di-OMe-rham 2,3-di-OMe-fuc + 2,4-di-0Me-fuc 3,4-di-0Me-fuc 2,3-di-OMe-ara 2,4-di-OMe-ara + 3,4-di-0Me-ara + + 1,4-di-OMe-xyl + + 2,4-di-OMe-man + + 2.5-di-OMe-eluc + 3,6-dideoxyhex + 3-OMe-heptose +

o t h e r minor s u g a r s a n a l y s e d 375377

435437

+ + + + +

596598 +

925927

+

+ + + +

12451247 + + +

+

+ + +

sugar yield. The mono-O-methylated sugars, tetroses (erythrose, threose), uncommon hexoses (allose, altrose), (gluco)heptose and glucosamine, collectively called minor sugars, constitute the remaining 11 to 13 wt% of the total sugar yield. They are listed in Table 6.2. The yields of the individual minor sugars are summarized in Appendix 6. The di-O-methyl sugars identified in the hydrolysates are listed in Table 6.3. These latter compounds were not present in all samples and their low abundances prevented quantification. Figure 6.1 shows depth profiles of relative sugar contributions in wt% and of sugar yields with respect to the total organic carbon content of the sample in mg sugar per gram organic carbon. The variations in the relative sugar contributions probably reflect variations in types and abundances of marine microorganisms (see discussion below). A predominantly marine origin of the organic matter was deduced from stable isotope, sedimentologic and climatologic evidence by Muller et al. (1983) and from the presence of long-chain alkenones and dinosterol as reported by Brassell et al. (1986a,b). The present sugar data also indicate a predominant input of marine organisms. This is deduced 1) from the rather uniform contribution of the eight major sugars in Table 6.1 (e.g. Degens and Mopper, 1975; Klok et al., 1984 a,b). 2) from the relatively high contributions of the minor sugars (Klok et al., 1984a,b; chapters 2 and 3). 3) from the glucose/ribose ratio's which range from 2.3 to 3.3, values typical for marine derived organic matter (Degens and Mopper, 1975).

114 The S1SC values (see Table 6.1) are also in a range commonly found for marine microorganisms (Degens, 1969; Fontugne and Duplessy, 1978, 1981). However, 513C values are known to be affected by variations in planktonic assemblages and by sea-surface temperatures (e.g. Fontugne and Duplessy, 1978, 1981). Variations both in planktonic assemblages and in sea-surface tempera­ tures are known to be reflected in the organic material of the sediment samples studied at present, so the use of s 13 C data for the assessment of input of organic material in the present case is not unambiguous and will therefore not be attempted. The total sugar yields (indicated as 'SUM' in Table 6.1) range between 0.376 and 0.883 mg saccharides per gram dry sediment. Table 6.1 shows that no significant decreasing trend with depth is observed for this figure. This suggests that only small amounts of carbohydrates are degraded once they have become incorporated in the sediment. Degradation of carbohydrates due to bacterial activity in the sediment is apparently so slow that no major effects on the sugar yields are observed over a period of 800,000 years. The organic carbon data (TOC) in Table 6.1 also do not show a decreasing trend with depth. A plot of these parameters (see Fig. 6.2) indicatès that no significant correlation exists between them. The total carbohydrate carbon content (COHC) of the samples ranges from 3.0 to 5.8 wt% of the total organic carbon (TOC), with the upper two samples showing the higher figures (Table 6.1). In line with the results discussed in the previous alinea it is suggested that the observed variations with depth reflect chiefly variation in the input of organic material and not preferential decay of carbohydrates in the lower six samples. This interpretation is in agreement with results of Iiebezeit (1986) who did not observe any preferential decay of saccharides in sediments from the Baltic and from a location off NW-Africa. Similar results were obtained from analysis of neutral saccha­ rides in sapropels from the Mediterranean, ranging in age from ca. 6,500 to 200,000 y. B.P. (chapter 7). Comparison of the present COHC/TOC data (see Table 6.1) with data from the literature of saccharide concentrations determined in the water column and in sediments shows that the values do not differ much: Carbohydrate contributions to the total particulate organic carbon pool in the water column are reported to range between 3 and 15 wt% (Ittekkot et al., 1982; Liebezeit, 1984). Tanoue and Handa (1987) report 7 to 10 wt% carbohydrate carbon/total organic carbon

115 in sinking particles in the water column. These results also suggest that no extensive microbial degradation of carbohydrates is taking place at the sediment/water interface or deeper in the sediment. Muller et al. (1983) report increased total organic carbon concentrations associated with the onset and during glacial periods in sediment samples off NW-Africa not far from where the present samples were taken. This was explained by increased upwelling giving rise to increased productivity of marine organisms in the euphotic zone and hence increased levels of organic material in the sediment. The samples showing the highest TOC values in the present case, i.e. the samples from stages 6, 8 and 9 (see Table 6.1 or Fig. 6.2), are also the ones deposited shortly before the onset or during glacial periods (Muller et al, 1983; Marlowe, 1984). However, no significant correlations are observed between the alkenone unsaturation indices or « l s O signals, which are used to determine cyclicities related to glacial and

en

255

£?0.8H-

c

■ 375 ■ 15

0.6

O

en

■ 136

D (/)

■ 925

■ 435

■ 596 ■ 1245

o 1

0

— i

1

0.5

—

1

0.7

T

1

0.9

TOC in wt% Fig. 6.2 Total sugar yield (in mg per g dry sediment) versus the total organic carbon content (in wt% of the dry sediment).

116 Table 6.4 Correlation coefficients of relative suear vlelds with varlous parameters, 51SC °/oo

TOC wt%

total sugars

U*7

6 18 0* % o

SÊZÊ

threose +0.70 +0.85 glucoheptose -0.74 +0.84 glucosamine -0.78 2-OMe-ara +0.86 4-OMe-ara +0.88 6-OMe-man +0.66 +0.87 6-OMe-gal +0.82 2/5-OMe-gal +0.71 2-OMe-gluc -0.65 3/4-OMe-man +0.72 +0.81 3/4-OMe-gal +0.71 fucose -0.64 xylose +0.64 +0.70 +0.67 total minor +0.89 All sugars were investlgated, but only entrles with r > 0.60 and 2c , < slope are listed. slope *: The coefficients with 5 18 0 are based on data of the upper slx samples only. The U 3 7 and 5 18 0 values are after Brassell et al. (1986a.b).

interglacial periods, and either TOC or total sugar data. So it seems that increased TOC values in the present case are related more strictly to periods of increased primary productivity which do not necessarily coincide with lower sea-surface temperatures during glacial periods. Closer examination of the relative sugar yields reveals that the contributions of some sugars are positively correlated with the TOC and total sugar data (see Table 6.4). Especially the covariance of xylose, 3- and/or 4-O-methylmannose and 6-O-methyl-mannose with both parameters and consequently the association of these sugars with increased primary productivity is noteworthy. The presence of xylose in marine sediments is probably derived from structural polysaccharides of marine algae (e.g. Stewart, 1974; Liebezeit et al., 1983) or from cyanobacteria (Klok et al., 1984a). These results together suggest that xylose, 3- and/or 4-O-methyl-mannose and 6-O-methyl-mannose are derived from marine organisms abundant during periods of high primary productivity.

117 Changes in sea-surface temperatures give rise to different populations of foraminifera (Diester-Haass and Schrader, 1979). As saccharide compositions of microbial cell walls are thought to be specific for the different species (Klok et al., 1984a), one might expect correlations between the 5 1 8 0 values of foraminiferal carbonate tests, which are a measure for the sea-surface temperatures, and the relative saccharide contributions. Data in Table 6.4 suggest that the paleo sea-surface temperature is reflected in the relative contributions of some of the minor components: Lower temperatures, i.e. less negative SiaO, seem to be associated with higher relative contributions of minor sugars and specifically glucoheptose, threose, and 2- and 4-O-methylarabinose. Heptoses are thought to be characteristic constituents of the lipopolysaccharide complexes of cell-walls of Gram-negative bacteria and cyanobacteria (Weckesser et al., 1979; Schmidt et al, 1980a,b). High heptose contributions are therefore regarded as indicative for a high contribution of these bacteria to the organic matter in the sediment. It seems therefore that higher sea-surface temperatures are associated with a lower contribution of (certain kinds of) bacteria in the water column. The composition of the assemblages reflected by these correlations is not known. In view of the correlation existing between paleo sea-surface temperature and the alkenone unsaturation index (Brassell et al., 1986a,b), it would be expected that the sugars correlating with s 1 8 0 would correlate with the alkenone unsaturation index as well, because the alkenone unsaturation index is a measure of the fluidity of cell membranes and hence a measure of paleo sea-surface temperatures in the same way as the oxygen isotope fractionation. It is remarkable that only glucoheptose is consistent in this respect as can be seen from Table 6.4. The significance of the lack of correlation between the relative contributions of threose and 2- and 4-O-methyl-arabinose with the alkenone unsaturation index data (see Table 6.4) is not completely clear, though it should be emphasized that correlation between s 1 8 0 data and the alkenone unsaturation index is rather low (r=-55) in the present samples. So it is speculated that the correlations between the various partially methylated sugars and 6 1 8 0 reflect variations in microbial assemblages not necessarily associated with variations in sea-surface temperature. Moreover, it should be kept in mind that the correlations of the sugar contributions with the 5 l s O data are based on six observations and that all other cor­ relations are based on eight observations. This implies that only tentative statements can be made.

118 6.5 CONCLUSIONS 1) Analysis of neutral saccharides yields about 40 different sugars. The relative abundances of the sugars together with the great variety of monomers present show that microorganisms are the most important contributors to the organic material and that the contribution of terrestrially derived organic matter from vascular plants is negligible in a quantitative sense. 2) No significant correlations of total sugar yields and of total organic carbon contents with increasing depth are observed. This suggests that only small amounts of organic material and of carbohydrates are degraded in the sediment and that no preferential decay of carbohydrates with respect to other organic constituents is taking place. 3) Differences in relative saccharide contributions with depth reflect variations in the composition of the microbial assemblages. These differences are due to variations in the relative contributions of algae, protozoa and bacteria. 4) Periods of increased primary productivity are reflected by relatively high total organic carbon contents. Covariance of the relative contributions of xylose, 6-O-methyl-mannose and 3- and/or 4-O-methyl-mannose with this parameter suggests that these sugars are associated with the presence of marine organisms, probably cyanobacteria or algae, relatively abundant during periods of increased primary productivity. 5) Interglacial periods are associated with decreased contributions of glucoheptose which probably points to a decreased contribution of bacteria.

Acknowledgements. Koninklijke/Shell Exploratie en Produktie Laboratorium (Shell Research BV), Rijswijk is gratefully acknowledged for financial support of this investigation. Prof. Dr. S.C. Brassell and Prof. Dr. M. Sarnthein are thanked for providing the samples and additional information. We thank Mr. J. Meesterburrie (Vening Meinesz Laboratory, State University Utrecht) for determination of the organic carbon and 5 13 C values. Dr. H.C. Cox is thanked for critically reading the manuscript.

119 7. ANALYSIS OF CARBOHYDRATES IN QUATERNARY SAPROPEL SAMPLES FROM THE EASTERN MEDITERRANEAN.

M.E.C. Moers, M. Baas, J.W. de Leeuw, H.C. Cox and P.A. Schenck

7.1 ABSTRACT This paper describes the presence of neutral saccharides in acid hydrolysates of a series of recent sapropels (SI, S5, S6, S7), varying in age from ca. 6500 to 200,000 y. B.P., recovered from the eastern Mediterranean. In the hydro­ lysates of all samples over 50 different neutral saccharides could be identified. The total sugar yields vary between 0.7 and 4.7 mg per gram dry sediment and no overall trends with age of the sediment are observed. The rather uniform relative contributions of major sugars (glucose, galactose, mannose, xylose, arabinose, ribose, rhamnose and fucose) and the relatively high contribution of minor sugars (a.o. partially methylated sugars, heptoses, amino sugars and tetroses) indicate that the majority of the organic matter is derived from marine organisms, such as algae, protozoa and bacteria. High contributions of 3- and 4-O-methyl-hexoses, 2-O-methyl-glucose, 2- and/or 5-O-methyl-mannose and 2-O-methyl-arabinose are associated with the original presence of cyanobacteria. These sugars are presumably partly derived from lipopolysaccharides. Both total sugar yields and total organic carbon concentrations are positively correlated with the relative contribution of diatoms. This is interpreted as being the result of increased primary productivity and preservation of diatoms presumably in relation to upwelling.

i

120 12 INTRODUCTION This paper describes the occurrence of neutral saccharides in Quaternary sapropels from the eastern Mediterranean. In late Quaternary sediments from this area at least twelve organic rich layers (sapropels) are known to occur. In the case of the Mediterranean, organic rich layers are called sapropels when their organic carbon content is greater than 2 wt% and their thickness greater than 1 cm (Kidd et al., 1978). The sapropel code used in the present case is SI, S2, S3 etc. (McCoy, 1974). The occurrence of sapropels is a reflection of the preservation of relatively large amounts of organic matter due to special circumstances such as restriction of water circulation which causes oxygen depletion and consequently anoxic conditions in the water column and/or water/sediment-interface, or to high productivity of organic matter due to upwelling of cold nutriënt rich waters, or to high productivity due to increased inflow of nutriënt rich waters from the Nile. For more extensive covering of models of sapropel formation the reader is referred to ten Haven et al. (1987a,c) and references cited therein. Mediterranean sapropels have been the subject of a number of studies in order to investigate their occurrence and origin (for references, see ten Haven et al., 1987a). A systematic organic geochemical study of lipids in SI, S5, S6 and S7 sapropels was undertaken by ten Haven (1986). This study showed that the sapropel layers contain markers for marine organisms such as algae and protozoa, for bacteria and for terrestrial plants (ten Haven et al., 1987a,b,c) and that the relative abundances of these markers are different in the different layers. In the present case an SI, S5, S6 and an S7 layer taken from cores collected during a 1983 expedition of the R.V. Tyro were studied. These are the same sapropel layers as those studied by the Haven (1986). The S6 layer was subdivided into three sub-layers and the S7 layer into two sub-layers on the basis of pollen analysis (Ganssen and Troelstra, 1987; ten Haven et al., 1987a,b). Pollen analysis was carried out as a means of assessing variations in the climates prevailing during deposition of the sediments under investigation. This gives insight into the variations in sea-surface temperatures as well. It should be noted that pollen and other palynomorphs are relatively rare in the sapropels under investigation (ten Haven et al., 1987c), so that their biomass is negligible in a quantitative sense.

121 The aim of the present study is to investigate how neutral saccharides reflect the differences in input of organic matter and to see whether this correlates with the lipid data.

7.3 EXPERIMENTAL The samples were first treated with 0.1 M HCl for removal of carbonates. Next the sediments were washed with water till neutral and then they were freeze dried. For analysis of carbohydrates about 20 g of dry sediment was hydrolyzed in ca. 2500 ml 0.25 M H 2 S0 4 for 18 h at 100°C. Myo-inositol was added as an internal Standard. The monosaccharides released by the hydrolysis were reduced to alditols by NaBH4 (room temperature, 16 h) and then acetylated to alditol acetates by acetic anhydride/pyridine (100°C, 3 h). For more com­ plete descriptions of the reduction and acetylation procedures the reader is referred to chapter 3. The reaction mixtures were analysed by gas chromatography (GC) on a Carlo Erba 4160 gas chromatograph and by gas chromatography-(electron impact) mass spectrometry (GC-(EI)MS) on a Varian 3700 gas chromatograph coupled to a Varian-MAT-44 mass spectrometer. The samples were also analysed by gas chromatography-(chemical ionization) mass spectrometry (GC-(CI)MS) either on a Varian 3700 gas chromatograph coupled to a Varian-MAT-44 mass spectrometer or on a Packard 438 gas chromatograph coupled to a Jeol DX300 mass spectrometer and DA500 mass data system. Isobutane was used as ionization gas. For the gas chromatographic separations a CPsil88 fused silica capillary column was used (1=25 m, i.d.=0.32 mm, df=0.12 nm; Chrompack, Middelburg, The Netherlands). The samples were injected at 70° C in ethyl acetate. The temperature was then rapidly raised to 150°C and from thereon programmed at 3° C/min to 230° C and kept isothermal at this temperature for 15 min. Identification of the alditol acetates was achieved by comparison of the relative retention times with those of Standard mixtures and by comparison of the mass spectra with those of known compounds (see references in chapters 2 and 3). Quantification of the compounds analysed was achieved by peak area integration of the GC results using a Maxima Chromatography Workstation (Dynamic Solutions

122 Corporation, Ventura CA, USA) and comparing the responses with the one of myoinositol hexaacetate (the internal Standard). The FTD-responses of all alditol acetates, including myo-inositol, were assumed to be equal on a weight basis. In a few cases in which compounds coeluted, quantification was achieved by peak area integration of the responses of selected characteristic mass fragments from the GC-(CI)MS analyses. Aliquots of an S6 and an S7 sub-sample were analysed with the use of NaBD< as reducing agent. This step was taken to investigate the natural occurrence of alditols such as ribitol or mannitol in the sediment samples. The remaining parts of the procedure were the same as described above. The minor sugars were not investigated on their natural occurrences as alditols. Aliquots of the S5, S6 and S7 samples were analysed without the use of myoinositol as internal Standard. This step was taken to investigate the natural occurrence of myo-inositol in the sediments.

7.4 RESULTS AND DISCUSSION Sugar yields Table 7.1 summarizes the yields of the individual major sugars. Apart from these major sugars a number of minor sugars could be identified. The minor sugars that could be identified but not quantified due to the low concentrations in which they were present are listed in Table 7.2. The minor sugars that could be quantified are listed in Table 7.3. This latter group of minor sugars will be called 'collective minor sugars' and the summed sugar yields of this group are listed in Table 7.1 under 'minor sugars'. The yields of the individual minor sugars are collated in Appendix 7. Investigation of the natural occurrence of alditols revealed that none of the major sugars appeared to be present as natural alditols in the sediment samples studied. This precludes for example that ribitol is derived from teichoic acid, an important cell wall constituent of Gram-positive bacteria. Analysis of sediment samples without myo-inositol as an internal Standard showed that the amounts in which this sugar is naturally present in the sediments are below the detection limit of the analytical techniques used.

123 Table 7.1 General information and yields of neutral saccharldes in mg per pram drv sediment. ^___ sapropel-layer code SI S5 S6 S6 S6 S7 S7 core no. 25 30 7 7 7 7 7 depth In cm 23269- 176- 192- 218- 283- 29830 317 188 216 232 296 318 5 13 C in % o -20.5 -22.1 -24.0 -23.6 -22.7 -22.0 -22.0 age in 10' y. B.P. 6.5 125 185 185 185 200 200 TOC in wtZ 2 10 6 3 4 5 7 sugar yields in mg/g rhamnose 0.149 0.353 0.060 0.037 0.099 0.043 0.170 fucose 0.154 0.304 0.064 0.027 0.133 0.028 0.179 ribose 0.071 0.235 0.045 0.014 0.119 0.008 0.096 arabinose 0.190 0.302 0.096 0.060 0.141 0.046 0.163 xylose 0.224 0.460 0.122 0.575 0.219 0.074 0.267 mannose 0.250 0.591 0.105 0.142 0.233 0.111 0.197 galactose 0.321 0.743 0.139 0.101 0.344 0.101 0.352 glucose 0.353 0.922 0.166 0.206 0.423 0.163 0.383 minor suears SUH

0.552 0.802 0.171 0.080 0.273 2.264 4.711 0.968 0.725 1.984

0.120 0.364 0.694 2.170

COHC/TOC

4.5

0.6

in wtX

1.9

0.6

1.0

2.0

1,2

*: The general information i s taken from ten Haven e t a l . (1987b). **: Gram carbohydrate carbon (COHC) per 100 gram t o t a l organic carbon (TOC) .

The total sugar yields range between ca. 0.7 and 4.7 mg/g sed. and it appèars that between 0.6 and 4.5 wt% of the total organic carbon (TOC) can be attributed to carbohydrate carbon (COHC) (see Table 7.1). The data in this table suggest that the sugar yields are not noticeably effected by the prolonged burial times of the S5, S6 and S7 sapropels, though it should be mentioned that the COHC/TOC value is highest for the youngest (SI) sapropel sample. The differences in the sugar yields seem to stem therefore mainly from differences in contributing organic matter (see discussion below). Variation in the contribution of organic matter was already deduced from studies of lipid compounds and from microscopic studies by ten Haven et al. (1987a,b,c): The S5 sapropel contained abundant diatoms next to contributions of prymnesiophyte algae and planktonic cyanobacteria. The S6 and S7 sapropels were characterized by rather high contributions of prymnesiophyte algae and planktonic cyanobacteria. Dinoflagellates were minor constituents in all samples analysed. Fig. 7.1 shows relative sugar yields as wt% of the total sugar yields. The samples show distribution patterns typical for samples in which the organic matter is for the major part derived from organisms such as algae, protozoa

124 and bacteria. This can be deduced for instance from the comparatively low contribution of glucose (16 to 28 wt%) and from the comparatively high contributions of galactose (14 to 17 wt%) and of the collective minor sugars (11 to 25 wt%) (see e.g. Klok et al., 1984a,b; chapters 2, 3, 4 and 5). A major marine origin of the organic matter in the sapropel layers under investigation was also deduced from studies of lipid markers (e.g. ten Haven et al., 1987a,b). Table 7.2 Partially tdentifled. but not sapropellayer code core no. depth in cm 2,3,4-tri-OMe-rham 2,3,4-tri-OMe-fuc 2,3,4-tri-OMe-ara 2,3,4-tri-OMe-xyl 2,3-di-OMe-rham 2,4-di-0Me-rham 3,4-dl-0Me-rham 1,3-di-OMe-fuc 2,3-di-OMe-fuc 2,4-di-0Me-fuc 3,4-di-0Me-fuc 2,3-di-OMe-rib 2,3-di-OMe-ara 2,4-di-OMe-ara 3,4-di-OMe-ara 2,3-di-OMe-xyl 2,4-di-OMe-xyl 2,3-di-OMe-man 2,4-di-OMe-man 3,6-di-OMe-man 2,3-di-OMe-gal 2,4-di-OMe-gal 2,5-di-OMe-gal 3,4-di-OMe-gal 2,6-di-OMe-gluc 3,4-di-OMe-gluc 3,6-di-OMe-gluc 3,6-dideoxy-hex 3,6-dideoxy-hex 3,6-dideoxy-hex 6-deoxy-gluc 6-deoxy-hex 2-deoxy-gal glucoheptose amino sugar glucosamine

methylated auantified SI S5 25 30 23- 26930 317

and other minor sugars

+

+

+

+

+

+

S6 7 192216 + + + + + + +

+ + +

+ + + +

+ + + +

+ + + +

+ + + +

+ + + +

S6 7 176188

+

+ + + + + +

S6 7 218218

+ +

S7 7 298318

+ +

+ + + +

+ + + + + +

+ + + +

+ + + +

+ + + +

+ + + +

+ + + +

+

+ + + + + + + + + +

+ + + + + + + + +

+ +

+ +

+ + + +

+

+

+

+ +

+ +

+ +

+

+

+

+

+ +

+

+ + +

S7 7 283296 + + + + +

+ +

+ + + +

+

+

+ + +

+ + + +

+ + +

+

125 Sugar-sugar correlations Additional information is obtained on examination of a correlation diagram of the relative sugar yields (see Table 7.4). It appears that the relative contributions of glucose and mannose are negatively correlated with especially fucose and xylose. A high contribution of glucose is thought to indicate an input of terrestrially derived organic matter, namely as the hydrolysis product of cellulose (see references above). Xylose on the other hand is reported to occur in appreciable amounts in hemicelluloses of many vascular plants (e.g. chapters 2, 3 and 4 and references cited therein) and especially in grasses (e.g. Lomax et al., 1983; Chesson et al., 1985). Xylose is, however, also an important constituent of structural polysaccharides of certain (green) algae (e.g. Stewart, 1974; Liebezeit et al., 1983) and of benthic cyanobacteria (Klok et al., 1984b). Hemicelluloses are thought to be much more rapidly degraded than cellulose (e.g. Hatcher et al., 1981; Hedges et al., 1985; Wilson et al, 1987; Stout et al., 1988) and transportation of vascular plant remains to the sea either by river run-off or by eolian trans­ port is likely to cause extensive degradation of the saccharides contained. Consequently it seems plausible that the xylose analysed in the present sapropel samples is predominantly derived from algal or bacterial matter and not from vascular plant remains. The negative correlation between the glucose and xylose contributions and the positive correlation between xylose and fucose (see Table 7.4) confirm this, because fucose is often considered as a sugar typical for microorganisms (e.g. Myklestad, 1974; Degens and Mopper, 1975; Klok et al., 1984a,b; chapters 2 and 3). Other correlations worth mentioning are the positive correlation between glucose and mannose and the negative correlation between the relative con­ tributions of glucose and of the collective minor sugars. Concomittant high contributions of glucose and mannose have been observed in other sediments as well, for instance in an Oligocene sample from the Falkland Plateau (chapter 8). High contributions of mannose in peats have been attributed to micro­ organisms (chapters 2, 3 and 4), but high contributions of mannose in marine sediments have been attributed to vascular plants (e.g. Michaelis et al., 1986). This sugar is also reported as constituent of structural polysaccharides in algae and bacteria (e.g. Stewart, 1974; Aspinall, 1983). So the origin of mannose in sediments is not unambiguous. Therefore, the correlation observed between glucose and mannose can be explained in several

126 ways. Firstly it can be argued that both mannose and glucose represent resistant remains of allochtonous terrestrial plants. The observed negative correlation with the contributions of the minor sugars would then be the result of dilution of the autochtonous organic matter, which is predominantly derived from microorganisms. Another explanation is that diagenesis of autochtonous marine organic matter itself leads to selective preservation of resistant sugars, namely glucose and mannose. This latter possibility might explain the rather low total sugar yields in the S6 sample from 192-216 cm and in the S7 sample from 283-296 cm. However, no correlations were observed between the relative contributions of mannose and glucose and the total sugar yield. Moreover, ten Haven et al. (1987b) concluded from the absence of a decreasing trend with depth in the relative phytol concentration that diagenetic effects are probably minimal, which makes the second explanation less likely. Sugar-Iipid correlations Interpretation of the sugar-sugar correlations mentioned above is partly confirmed when the present data are compared with the results of analysis of lipid biomarkers, pollen analysis and 513C data as reported by ten Haven et al. (1987b). Correlations of the contributions of major and minor sugar with the contributions of lipid biomarkers and with results of pollen analysis are listed in Table 7.3. It should be noted that the correlations between the lipid data and the sugar results are based on the data of the S5, S6 and S7 samples. The correlations with the pollen record are based on the data of the S6 and S7 samples. The small number of samples in all cases allows only tentative statements to be made. Table 7.3 shows that no correlation is observed between relative sugar con­ tributions (except rhamnose) and variations in the contributions of marine and terrestrial lipid biomarkers. This lack of correlation seems remarkable in light of the presumed predominant terrestrial origin of glucose and mannose and marine origin of e.g. xylose and fucose mentioned above. The discrepancy (opposite page) Fig. 7.1 Relative sugar yields as wt% of the total sugar yield. 'Minor' refers to the collective minor sugars. The identities of the sugars comprised by this term are listed in Table 7.3.

127 S5

SI

20-

7

' Atyzx rhom fuc

rib

gH § ara

xyl

I

I

10-

man gal

."7"

gluc minor

A M4 / ^ 4 fuc

rib

aro

I

man gal

Vmmn rib

ara

xyt

/

man gal gluc minor

S7. 283-296 cm

In n oi L4 n

n

1 /

0 g xyl

/

fuc

S6. 176-188 cm

y

s

20

S2L!=Pmam fuc rib ara

gluc minor

S6. 192-216 cm

/

*

xyl man gal gluc minor

S7. 298-318 cm 20-

rhom fuc

rte

ara

xyl

man got

gluo minor

S6. 218-232 cm

20

y00

mam fuc

rib

ara

1 ISi xyl

man gal

Ld

gluc minor

1

0 0 H 3 ifl H^ 0.03 X » o c