Lignocellulosic biomass is a potential source of fermentable sugars such as glucose. Enzymatic hydrolysis of cellulose is a viable method of solubilizing the glucose from biomass, but the cellulose fraction of native lignocellulosic material is shielded from enzymatic attack by the lignin-hemicellulose matrix surrounding it. Pretreating lignocellulosic biomass with dilute sulfuric acid at high temperatures solubilizes hemicellulose, rendering the cellulose fraction more susceptible to enzymatic hydrolysis. Evaluation of dilute-acid, high-temperature pretreatments depends on polysaccharide analysis of the two fractions resulting from a pretreatment, prehydrolyzed solids(PHS) and prehydrolyzate liquid(PH). The polysaccharide analysis is based on a method described by the National Renewable Energy Laboratory and involves a two-stage sulfuric acid hydrolysis followed by HPLC quantification using ion-moderated partition chromatography and refractive index detection. The subject of this thesis is identifying and quantifying the sources of error associated with the polysaccharide analysis and the error associated with the evaluation of the effects of pretreatment on the polysaccharide fractions of switchgrass and poplar. This was addressed by conducting replicate polysaccharide analyses on single samples of native biomass, PHS, and PH. The variability associated with these measurements was compared to the variability associated with replicate analyses of identically pretreated biomass. It was found that the use of sugar standards to correct for sugar destroyed during the analysis adds error and most likely overestimates the amount of sugar from biomass actually destroyed. It is evident that assuming a volume after neutralization of the hydrolyzed biomass sample is more reproducible than measuring the volume. When using a batch-type reactor and the temperature and acid parameters used in this study,140°C-180°C/ 0.6-1.2 % sulfuric acid (w/w), it is evident that the major source of error in evaluating pretreatment conditions is the pretreatment itself, not the analysis. / Graduation date: 1995

En réponse aux préoccupations environnementales, les procédés industriels comme la production de bioéthanol de deuxième génération sont apparus. Basés sur la conversion enzymatique de la cellulose, ces processus font face à un problème majeur, la réticence de la biomasse lignocellulosique à l'hydrolyse. Afin de résoudre ce problème et celui lié aux coûts d'utilisation de cocktails de cellulases, les recherches se sont axées sur diverses méthodes permettant d'augmenter l'hydrolyse de la cellulose. Les champignons filamenteux sont connus pour être des dégradeurs naturels du bois et, par conséquent, sont utilisés dans de nombreuses applications biotechnologiques. Récemment, quelques études ont révélé l'importance d'enzymes fongiques telles que la CDH et les GH61 dans la dégradation oxydative de la lignocellulose. Les travaux réalisés au cours de cette thèse ont permis de démontrer l'importance de ces enzymes oxydatives dans les phénomènes de déconstruction de la lignocellulose. L'utilisation de ces enzymes oxydatives offre de réelles voies d'amélioration de la production de bioéthanol et de compréhension de la dégradation in vivo des lignocelluloses par les champignons. / In response to environmental concerns, industrial processes such as second generation bioethanol production have emerged. Based on enzymatic cellulose conversion, these processes are confronted with a major problem, the recalcitrance of lignocellulosic biomass. To solve the problem caused by substrate recalcitrance and high cost of cellulase cocktails, research has focused on various methods to enhance cellulose hydrolysis. Fungi are known to be natural degraders of wood and consequently are used in derived biotechnological applications. Recently, several studies have revealed the importance of fungal enzymes such as GH61 and CDH in the oxidative degradation of lignocellulose. During the work done on this thesis, we demonstrated implication of these oxidative enzymes in lignocellulose deconstruction to enhance hydrolysis performed by more classical cellulases. Utilization of oxidative enzymes offers a suitable way for bioethanol processing enhancement and comprehension of the in vivo lignocellulosic degradation by fungi.

A produção de PHAs é limitada devido ao alto custo da fonte de carbono para á produção. No Brasil, o uso de xilose uma fonte de carbono abundante no bagaço de cana é uma alternativa. Neste estudo o catabolismo de xilose em B. sacchari foi estudado para explorar seu potencial para a produção de PHB. Primeiro a organização do operon de xilose foi descrita e foi demostrado que a superexpressão de xylAB melhoro a velocidade máxima de crescimento assim como o teor de acumulo de PHB. Depois foi identificado o fenômeno de repressão catabólica, o qual foi abolido a traves da superexpressão dos genes xylE xylAB. Finalmente foi criado um set de plasmídeos induzíveis para fazer engenharia no consume de xilose em B. sacchari. A superexpressão de xylR permitiu que B. sacchari atinge a velocidade máxima de crescimento mais alta reportada e o melhor fator de conversão de xilose a PHB. Foi concluído então que a superexpressão de xylAB e xylR ajudam a melhorar a velocidade máxima de crescimento e a capacidade de acumulo de PHB usando xylose como fonte de carbono em B. sacchari. / Polyhydroxyalkanoate production is limited by the high production cost of carbon sources. The use of cheap carbon sources like xylose is an alternative to address this issue. In this work we aimed to understand and engineer xylose catabolism in B. sacchari, a bacteria isolated in Brazil to exploit its potential for producing PHB from renewable sources. Initially, we described organization of xylose assimilation genes and demonstrated that xylAB overexpression is an efficient strategy to improve B. sacchari growth rate and production of PHB using xylose as sole carbon source, achieving the highest conversion rate and titer described. Then we identified B. sacchari sequential preference for different sugars (glucose>arabinose>xylose) and overexpress xylE-xylAB to abolish this preference. Finally we created a set of inducible vectors and use them to engineer xylose metabolism. Overexpression of xylR, allowed B. sacchari cells to achieve the highest growth rate and PHB conversion factor and yield reported using xylose as a sole carbon source. Finally, we conclude that overexpression of xylAB and xylR genes improved growth rate, conversion factor and yield when PHB is produced using xylose as carbon source in B. sacchari.

L’extrusion biocatalytique, ou bioextrusion, est une technique d’extrusion réactive utilisant des enzymes comme catalyseurs. Cette technique est considérée en temps qu’étape intermédiaire, subséquente au prétraitement physico-chimique et précédente à l’hydrolyse enzymatique enréacteur fermé. L’utilisation de l’extrusion permet un procédé continu, facilement modulable et adaptable à des conditions de hautes consistances, de nombreuses biomasses et facilement transférable à l’échelle industrielle. Néanmoins, les données bibliographiques font ressortir la complexité des entrants et leurs interactions lors de la bioextrusion de biomasses lignocellulosiques. Les conclusions des bioextrusions de biomasses amidonnées soulignent l’importance de l’étude de l’influence de la concentration en substrat et en enzymes. Les résultats obtenus à partir de la bioextrusion des biomasses lignocellulosiques valident l’existence d’une activité enzymatique en extrudeuse malgré la contrainte thermomécanique et le temps de séjour limité. Lors de cette étape, l’hydrolyse de la fraction cellulosique est favorisée pour des milieux concentrés en substrat et en enzymes. Des modifications significatives des fractions cellulosiques cristallines et amorphes en surface, des réductions des tailles de particules, une dégradation visuelle des structures de la biomasse et l’augmentation de la sensibilité à la décomposition thermique, sont aussi observées sur la fraction solide. L’hydrolyse enzymatique des bioextrudats est prolongée en réacteur fermé. La bioextrusion permet des améliorations significatives des taux et vitesses de conversion des sucres sur le long terme, jusqu’à 48 h. Les gains observés sont relativement constants pour la paille de blé et augmentent avec le temps pour les écorces de bouleau et les résidus de maïs. Post-extrusion, la concentration en substrat influence négativement la conversion des sucres. Cependant, les plus-values de conversion du glucose lié à la bioextrusion de paille de blé sont principalement observables pour des concentrations en substrat et en enzymes élevées. À partir de 4 h, des baisses significatives de la conversion du xylose sont observées après bioextrusion. Les déstructurations de la fraction solide, déjà observées au cours la bioextrusion, se poursuivent en réacteur fermé. Les meilleurs résultats hydrolytiques aux niveaux des hautes charges en enzymes et en substrat sont associables aux bonnes conditions de mélanges caractéristiques des éléments bilobes. L’ensemble enzymatique est probablement réparti de façon plus homogène (mélange distributif) pour cibler plus de sites disponibles. De plus, le mélangé dispersif limite la proximité entre enzymes de même type et les gênes associées. Le procédé d’extrusion permet une agitation efficace, un bon transfert de masse et probablement un meilleur contact entre enzymes et substrat. Les moins bons résultats de conversion du xylose sont probablement à relier à des phénomènes d’adsorption non-spécifique, ou encore de désactivation des hémicellulases, provoqués par l’intensité des contraintes thermomécaniques et les résidus ligneux. Les bons résultats de déstructuration après bioextrusionsont associables à une action synergétique des contraintes mécanique et biochimique. Les analyses d’autofluorescence montrent l’évolution de la fraction ligneuse dans le processus de déconstruction de la fraction solide. Une production progressive de particules très fines,visiblement associée à la fraction ligneuse, est observée. Des complexes lignine-carbohydratessont aussi détectés dans la fraction liquide. Etant peu, voire pas hydrolysable par voie enzymatique, ces fractions hétéropolymériques sont un frein à la déconstruction. Si la déstructuration des lignines est probablement majoritairement liée au prétraitement alcalin, le procédé de bioextrusion provoque une diminution de la teneur en hétéropolymères de plus hautes masses moléculaires. / Biocatalytic extrusion, also named bioextrusion, is a reactive extrusion technique using enzymes as catalysts. Bioextrusion is considered as a link between the previous physico-chemical pretreatment (like alkaline extrusion) and the subsequent enzymatic hydrolysis in batch conditions. The extrusion allows a continuous, flexible and versatile process for high consistency media, easily transferable to the industrial level. However, complexity of both lignocellulosic biomass and lignocellulolytic enzymes and their interactions during the extrusion process are underlined by the literature. Numerous response surface methodology experiments with starchy biomass indicate that bioextrusion efficiency is mainly influenced by substrate and enzymes loading. Enzymatic activity during the bioextrusion process of lignocellulosic biomass is confirmed by the experiments despite the mechanical constraints and the limited residence time. During bioextrusion, best holocellulosic fraction hydrolysis results were obtained with high substrate and enzymes loadings. Significant modifications of the solid fraction like particule size reduction, visual deconstruction of the biomass structure, increased sensibility to thermal decomposition and the evolution of the surface exposure of crystalline and amorphous cellulose were observed. Enzymatic hydrolysis of the bioextrdates is prolonged in batch conditions. Clear improvements of speeds and rates of sugars conversion up to 48 h indicate a long term influence of the bioextrusion. Gain observed are steady for the pretreated wheat straw whereas it increases with time for corn residues and birch barks. Post-extrusion, a negative influence of the substrate loading is measured. However, best enhancements for the glucose conversion of pretreated wheat straw are detected for high substrate and enzymes loadings. From 4 to 48 h, significant losses in xylose conversion are measured with previous bioextrusion. Indicators of the solid fraction deconstruction, observed during the bioextrusion step, indicate a stronger biomass degradation after 48 h. Improvements of glucose conversion rates can be associated with good mixing conditions of the extruder, especially due to the use of kneading elements. Enzymes are probably more homogeneously distributed (distributive mixing) and can access more catalytic sites available. Moreover, dispersive mixing limits the enzyme jamming due to the biocatalysts concentration. Extrusion process permits an better agitation efficiency, good mass transfer conditions and probably a higher contact between substrate and enzymes. Lower xylose conversion results may be attributed to non-specific adsorptions or inactivation phenomena due to mechanical constraints and lignin residues. Good deconstruction results on the solid fraction may be associable with a synergetic action between mechanical and biochemical constraints. Autofluorescent signal analysis of the lignin fraction show its evolution during the deconstruction of the solid residue. During the hydrolysis, a progressive production of very small particles, appearing to be associated with the lignin fraction is observed. Lignin-carbohydrate complexes are also detected in the liquid fraction. These heteropolymeric complexes, difficult or even impossible for the enzymes to hydrolyze, are an obstacle to the biomass valorization. If lignin deconstruction is mainly due to the alkaline pretreatment, bioextrusion process seems to reduce the proportion of these heteropylymers with high molecular weights.

The need to reduce the cost of adsorption technology has led scientists to explore the use of many low cost adsorbents especially those from renewable resources. Lignocellulose and montmorillonite clay have been identified as potentially low cost and efficient adsorbent materials for the removal of toxic heavy metals and organic substances from contaminated water. Montmorillonite clay has good adsorption properties and the potential for ion exchange. Lignocellulose possesses many hydroxyl, carbonyl and phenyl groups and therefore, both montmorillonite and lignocellulose are good candidates for the development of effective and low cost adsorbents in water treatment and purification. The aim of this study was to prepare composite materials based on lignocellulose and montmorillonite clay and subsequently evaluate their efficacy as adsorbents for heavy metal species and organic pollutants in aqueous solution. It was also important to assess the adsorption properties of the modified individual (uncombined) lignocellulose and montmorillonite. Lignocellulose and sodium-exchanged montmorillonite (NaMMT) clay were each separately modified with methyl methacrylate (MMA), methacrylic acid (MAA) and methacryloxypropyl trimethoxysilane (MPS) and used as adsorbents for the removal of heavy metals and dyes from aqueous solution. The lignocellulose and NaMMT were modified with MMA, MAA and MPS through free radical graft polymerisation and/or condensation reactions. NaMMT was also modified through Al-pillaring to give AlpMMT. The materials were characterised by fourier transform infrared spectroscopy (FTIR), thermogravimetric analysis (TGA), scanning electron microscopy (SEM) and small angle X-ray scattering (SAXS) and characterisation results showed that the modification of the montmorillonite with MAA, MMA and MPS was successful. The modified lignocellulose and montmorillonite materials were evaluated for the adsorption of heavy metal ions (Cd2+ and Pb2+) from aqueous solution by the batch method. The adsorption isotherms and kinetics of both Cd2+ and Pb2+ onto the NaMMT clay, AlpMMT and lignocellulose materials are presented. The Langmuir isotherm was found to be the best fit for the adsorption of both heavy metals onto all the adsorbents. AlpMMT showed very poor uptake for heavy metals (both Cd2+ and Pb2+). PMMAgMMT, PMAAgMMT, PMAAgLig and PMPSgLig showed improved adsorption for both heavy metals. The mechanism of heavy metal adsorption onto the adsorbents was best represented by the pseudo second-order kinetic model. PMPSgLig, NaMMT and AlpMMT showed relatively high adsorption capacities for methyl orange, while the adsorption of neutral red was comparable for almost all the adsorbents. Neither the Langmuir model nor the Freundlich model was found to v adequately describe the adsorption process of dyes onto all the adsorbents. The pseudo second-order model was found to be the best fit to describe the adsorption mechanism of both dyes onto all the adsorbents. The modification of lignocellulose and montmorillonite with suitable organic groups can potentially produce highly effective and efficient adsorbents for the removal of both heavy metals and dyes from contaminated water. Novel adsorbent composite materials based on lignocellulose and montmorillonite clay (NaMMT) were also prepared and evaluated for the removal of pollutants (dyes and heavy metals) from aqueous solution. The lignocellulose-montmorillonite composites were prepared by in situ intercalative polymerisation, using methyl methacrylate, methacrylic acid and methacryloxypropyl trimethoxysilane (MPS) as coupling agents. The composite materials were characterised by FTIR, TGA, TEM and SAXS. SAXS diffractograms showed intercalated nanocomposites of PMMAgLig-NaMMT and PMAAgLig-NaMMT, whereas PMPSgLig-NaMMT showed a phase-separated composite and the same results were confirmed by TEM. The lignocellulose-montmorillonite composites were assessed for their adsorption properties for heavy metal ions (Cd2+ and Pb2+) and dyes (methyl orange and neutral red) from aqueous solution. Among these composite materials, only PMAAgLig-NaMMT showed a marked increase in the uptake of both Cd2+ and Pb2+ relative to lignocellulose and montmorillonite when used independently. The adsorption data were fitted to the Langmuir and Freundlich isotherms, as well as to the pseudo first-order and pseudo second-order kinetic models. The data were best described by the Langmuir isotherm and the pseudo second-order kinetic model. On the adsorption of dyes, only PMPSgLig-NaMMT showed enhanced adsorption of methyl orange (MetO) compared with lignocellulose and montmorillonite separately. The enhanced adsorption was attributed to the synergistic adsorption due to the presence of MPS, lignocellulose and NaMMT. Competitive adsorption studies were carried out from binary mixtures of MetO and Cd2+ or Pb2+ in aqueous solution. The adsorption process of MetO onto the composite material was found to follow the Freundlich adsorption model, while the mechanism of adsorption followed both the pseudo first-order and pseudo second-order models. This particular composite can be used for the simultaneous adsorption of both heavy metals and organic dyes from contaminated water. The adsorption of neutral red to the composite materials was comparable and the pseudo second-order kinetic model best described the adsorption mechanism.

L’hydrolyse enzymatique de la biomasse lignocellulosique est une voie prometteuse pour la bioconversion des matières végétales en sucres fermentescibles en vue de la production du bioéthanol de seconde génération. En général, des cocktails enzymatiques contenant différentes familles d’activités, caractérisées par des modes d’action différents, sont utilisés comme biocatalyseurs. L’essentiel des travaux de modélisation de ce procédé abordent la question via des approches cinétiques où les aspects particulaires et dynamique d’évolution des propriétés du substrat/biocatalyseur/système ne sont pas pris en compte. De plus, ce type de modèles, visant à reproduire les cinétiques de production des sucres simples, traite uniquement le cas de la mise en contact simultanée des enzymes et de la matière à hydrolyser (substrat). Dès lors, les questions relatives au design/optimisation du procédé telles que le mode d’alimentation (batch/continu) ou l’ajout séquencé des enzymes et/ou du substrat ne peuvent pas être abordées avec ces modèles. Dans ce travail, une approche de modélisation par le formalisme du bilan de population est proposée. Le modèle est basé sur une hétérogénéité structurale du substrat à savoir la distribution de taille des chaines/particules. Comme première approche numérique, la méthode des classes est utilisée dans le cas de l’hydrolyse de chaines polymères subissant des attaques endoglucanases (rupture aléatoire) et exoglucanases (coupure d’un dimère en bout de chaine). En deuxième lieu, la méthode des moments a été adoptée pour traiter du cas d’un substrat particulaire. Ici la rupture s’opère sous l’effet des contraintes hydrodynamiques tandis que l’hydrolyse enzymatique modifie la cohésion des particules. Par ailleurs, la nécessité de confronter les résultats numériques issus de la méthode des moments avec les distributions expérimentales a motivé un travail sur les méthodes de reconstruction des distributions à partir de leurs moments. Parallèlement à ce travail de modélisation, plusieurs métrologies nécessaires à la caractérisation de ces systèmes ont été mises en oeuvre. Trois techniques granulométriques (Morpho-granulométrie, Focused Beam Reflectance Measurement Technique (FBRM) et Granulométrie laser) ont été utilisées pour accéder à l’évolution de la distribution de taille des particules au cours de l’hydrolyse d’une cellulose microcristalline (Avicel). De plus, les concentrations en sucres réducteurs et en sucres simples (glucose et cellobiose) ont été mesurées. Les modélisations proposées combinent ainsi les aspects de cinétique homogène et de catalyse hétérogène. Intégrées dans une approche de type bilan de population multivariable, elles permettent d’accéder à l’évolution de la distribution de taille des chaines/particules du substrat ainsi qu’aux cinétiques de conversion en sucres simples. / The enzymatic hydrolysis of lignocellulosic biomass is a promising approach for the bioconversion of organic matter into fermentable sugars aiming ultimately to produce second generation biofuel. Globally, enzymatic cocktails, containing different activities which are characterized by their specific mechanisms, are used as biocatalysts. Most of the studies devoted to the modelling of this process address the problem via kinetic approaches in which the particulate aspects and the dynamic evolution of the properties of the substrate/biocatalyst/system are not taken into account. Moreover, such models aim to reproduce the kinetics of release of simple sugars and treat only the case where the enzymes and the substrate are simultaneously mixed at the beginning of the hydrolysis reaction. Therefore, issues related to the design/optimization of the process such as the supplying mode (batch/continuous) and the sequential adding of the enzymes/substrate cannot be addressed with these models. In this work, a population balance formalism is proposed as modelling approach. The model is based on a structural heterogeneity of the substrate namely the chain/particle size distribution. As a first numerical approach, the method of classes is used in the case of polymer chains undergoing endoglucanase (random breakage) and exoglucanase (chain-end scission) attacks. Secondly, the method of moments is adopted to solve the same problem and then adapted to the case of a particulate substrate by introducing the particle cohesion effect which depends on the enzymatic attacks as well as on the hydrodynamic shear stress. Finally, the confrontation of the numerical results from the method of moments to the experimental distributions motivated the development of reconstruction methods in order to restore distributions from a finite sequence of their moments. Alongside this modelling work, the suitable metrology has been developed for the characterization of these systems. Three different granulometric techniques (Morphogranulometry, Focused Beam Reflectance Measurement (FBRM) technique and Laser Diffraction Particle Size Analysis) are used to reach the time-evolution of the particle size distribution of microcrystalline cellulose (Avicel). In this context, numerical tools used for the analysis and the comparison of the different experimental distributions are proposed. In addition, the concentrations of reducing and simple sugars (glucose and cellobiose) are measured all along the hydrolysis reactions. The modeling of enzymatic hydrolysis developed here combines the concepts of homogeneous and heterogeneous catalysis. Integrated into the framework of multivariable population balance model, these allow the chain/particle size distribution evolution during the reaction and the kinetics of simple sugars release to be predicted.

Lecythophora (Coniochaeta) hoffmannii, a soil- and lignocellulose-inhabiting sordariomycete (Ascomycota) that can also live as a facultative tree pathogen causing soft rot, belongs to the family Coniochaetaceae. The strain CBS 245.38 sequenced here was assembled into 869 contigs, has a size of 30.8 Mb, and comprises 10,596 predicted protein-coding genes.

La biomasse végétale est une alternative durable et écologique pour les ressources fossiles. L'exploitation et la valorisation de cette biomasse sont rendues possibles grâce à la capacité naturelle des enzymes fongiques à dégrader et modifier cette biomasse. Parmi ces enzymes, les glyoxal oxydases génératrices de H2O2 (GLOX) restent un groupe peu étudié avec un seul exemple de protéine caractérisée dans la littérature à partir d’un champignon dégradant le bois.Dans cette thèse, trois GLOX, précédemment identifiées dans le génome du champignon dégradant le bois Pycnoporus cinnabarinus (PciGLOX), ont été sélectionnées, produites par voie hétérologue et caractérisées. La caractérisation a révélé des différences entre les trois PciGLOX dans la stabilité des protéines, la spécificité du substrat et l’efficacité catalytique. Les protéines PciGLOX sont produites sous leur forme inactive et leur mécanisme d'activation a été étudié. La capacité des GLOX à catalyser la réaction d'oxydation du 5-hydroxyméthylfurfural (HMF), d’intérêt industriel, a été étudiée pour la première fois dans ce travail. Le HMF a été oxydé par PciGLOX en acide 5-hydroxyméthyl-2-furancarboxylique (HMFCA) comme produit principal. Le HMFCA est difficile à produire par catalyse chimique et est utilisé dans la production de polyesters et de produits pharmaceutiques. PciGLOX ont également été capables de produire l’acide furandicarboxylique (FDCA), qui est un précurseur dans les procédés de production du bioplastique. Ce travail ouvre de nouvelles perspectives pour étudier plus en détail le rôle de GLOX dans la dégradation de la lignocellulose, et dans les applications biotechnologiques. / Plant biomass is a sustainable and eco-friendly alternative for fossil fuels. The exploitation and valorisation of plant biomass is possible through biotechnological processes that rely on the natural ability of fungal enzymes to degrade and modify this biomass. Among these enzymes are H2O2-generating glyoxal oxidases (GLOX), which haven’t been extensively studied with only one example in the literature on GLOX from wood-degrading fungi. In this thesis three GLOX, previously identified in the genome of the wood-degrading fungus Pycnoporus cinnabarinus (PciGLOX), were heterologously produced and characterisation. The three PciGLOX showed differences in their stability, substrate preferences and catalytic properties. The ability of GLOX to catalyse the biotechnologically important oxidation reaction of 5-hydroxymethylfurfural (HMF) was investigated for the first time in this work. PciGLOX oxidized HMF to 5-hydroxymethyl-2-furancarboxylic acid (HMFCA), which is difficult to produce via chemical catalysis and is used in polyesters and pharmaceutical products production. PciGLOX were also able to oxidize HMF derivatives leading to the formation of the final product furandicarboxylic acid (FDCA), which is a bioplastic precursor. PciGlOX proteins are produced in their inactive form and their activation mechanism was investigated in this thesis. This work opens new prospects to investigate more the role of GLOX in plant biomass degradation and biotechnology, and the possible optimization techniques of the catalytic properties of this enzyme.

Acid mine drainage is a well known environmental pollutant, not only in South Africa, but throughout the world, and the use of microbial processes in the treatment of these wastes has been the subject of investigation over past decades. Lignocellulose packed-bed reactors have been used in passive treatment systems, and, although effective initially, they show early decline in performance while the packing material remains largely un-utilized. Little is known about this phenomenon which remains a severe constraint in the development of efficient passive mine water treatment systems. It has been proposed that the degradation pathways of the complex lignocellulose substrate may be limited in some way in these systems during the manifestation of this effect. This study has addressed the problem using a molecular microbial ecology methodology in an attempt to relate trophic functions of the microbial population to the physico-chemical data of the system. A field-scale lignocellulose packed-bed reactor located at Vryheid Coronation Colliery (Northern Kwa-Zulu Natal province, South Africa) was monitored for six years and the results showed the classic profile of performance decline related to a slowdown in sulphate reduction and alkalinity production. The reactor was decommissioned , comprehensive samples were collected along the depth profile and the microbial populations investigated by means of 16S rRNA gene methodology. The population was found to include cellulolytic Clostridia spp., CytophagaIFlavobacterlBacteroidetes, Sphingomonadaceae and as yet uncultured microorganisms related to microbiota identified in the rumen and termite gut. These are all known to be involved as primary fermenters of cellulose. Oesulphosporosinus was present as sulphate reducer. A comparison of substrata sampling and population distribution suggested that spatial and temporal gradients within the system may become established over the course of its operation. Based on these findings, a laboratory-scale reactor was constructed to simulate the performance of the packed-bed reactor under controlled experimental conditions. The laboratory-scale reactor was operated for 273 days and showed comparable performance to that in the field in both biomolecular and physicochemical data. Clearly defined trophic niches were observed. These results suggested that a sequence of events does occur in lignocellulose degradation over time. Based on the spatial and temporal column studies, a descriptive model was proposed to account for these events. It was found that fermentative organisms predominate in the inlet zone of the system using easily extractable compounds from the wood, thus providing feedstock for sulphate reduction occurring in the succeeding compartments. Production of sulphide and alkalinity appears to be involved in the enhancement of lignin degradation and this, in turn, appears to enhance access to the cellulose fraction. However, once the readily extractables are exhausted, the decline in sulphide and alkalinity production leads inexorably to a decline in the overall performance of the system as a sulphate reducing unit operation. These observations led to the proposal that with the addition of a limited amount of a readily available carbon source, such as molasses, in the initial zone of the the reactor, the ongoing generation of sulphide would be sustained and this in turn would sustain the microbial attack on the lignocellulose complex. This proposal was tested in scale-up studies and positive results indicate that the descriptive model may, to some extent, provide an account of events occurring in these systems. The work on sustaining lignocellulose degradation through the maintenance of sulphate reduction in the initial stages of the reactor flow path has led to the development of the Degrading Packed-bed Reactor concept and that, has subsequently been successfully evaluated in the field.

South Africa is renowned for its mining industry. The period over which the polluted waters from the existing and abandoned mines will require treatment has driven research into the development of passive treatment systems. These waters are characterised by a low pH, high concentrations of heavy metals, high levels of sulphate salts and low concentrations of organic material. The biological treatment of these waters has been a subject of increasing focus as an alternative to physicochemical treatment. The utilisation of lignocellulose as a carbon source has been restricted by the amount of reducing equivalents available within the lignocellulose matrix. After a few months of near 100% sulphate reduction, it was found that although there was a large fraction of lignin and cellulose remaining, sulphate reduction was reduced to less than 20%. The present study demonstrated that lignocellulose can be utilised as a carbon source for sulphate reduction. It was established that lignocellulose degradation was enhanced under biosulphidogenic conditions and that lignin could be degraded by a sulphate reducing microbial consortium. It was established using lignin model compounds synthesized in our laboratory, that the bonds within the lignin polymer can be cleaved within the sulphidic environment. The presence of cellulolytic enzymes, using CMCase as a marker enzyme, was detected within the sulphate reducing microbial consortium. Based on the results obtained a descriptive model was formulated for the degradation of lignocellulose under biosulphidogenic conditions. It was determined that the initial reduction in sulphate observed using lignocellulose as a carbon source was due to the easily extractable components. The degradation of which resulted in the production of sulphide, which aided in the degradation of lignin, allowing greater access to cellulose. Once the easily extractable material is exhausted, the cycle is halted, unless the sulphide production can be maintained. This is the focus of an ongoing project, testing the hypothesis that an easy to assimilate carbon source added after exhaustion of the easily extractable material, can maintain the sulphide production.