Genomics and transcriptomics of the hydrogen producing extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus

Marcel R. A. Verhaart

Thesis committee

Thesis supervisor Prof. Dr. Alfons J.M. Stams Personal Chair at the Laboratory of Microbiology Wageningen University

Thesis co‐supervisor Dr. Servé W.M. Kengen Assistant Professor at the Laboratory of Microbiology Wageningen University

Other members Prof. dr. ir. C.J.N. Buisman Wageningen University

Prof. Dr. W.R Hagen Delft University of Technology

Dr. Astrid E. Mars Wageningen UR Food & Biobased Research

Dr. ir. E.W.J van Niel Lund University, Sweden

This research was conducted under the auspices of the Graduate School WIMEK

Genomics and transcriptomics of the hydrogen producing extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus

Marcel R. A. Verhaart

Thesis submitted in fulfilment of the requirements for the degree of doctor at Wageningen University by the authority of the Rector Magnificus Prof. dr. M.J. Kropff, in the presence of the Thesis Committee appointed by the Academic Board to be defended in public on Friday 19 November2010 at 4 p.m. in the Aula.

Marcel R.A. Verhaart Genomics and transcriptomics of the hydrogen producing extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus, 178 pages.

Thesis, Wageningen University, Wageningen, NL (2010) With references, with summaries in Dutch and English

ISBN 978‐90‐8585‐796‐9

Abstract

As fossil fuels are depleting, there is a clear need for alternative sustainable fuel sources. One of the interesting alternatives is hydrogen, which can be produced from biomass by bacteria and archaea. To make the application feasible, organisms are needed which have high hydrogen productivities as well as the capacity to handle a wide variety of substrates. Caldicellulosiruptor saccharolyticus has these abilities and this was the reason to select this organism to study hydrogen production at elevated temperatures. The aim of this study was to clarify the pathways involved in hydrogen production as well as studying the regulation of the hydrogen production on different levels. Especially the regulation of the disposal of reducing equivalents (NADH, reduced ferredoxin, NADPH), was an important aspect. The genome sequence was annotated and compared with transcriptome data. This revealed the presence of the Emden‐Meyerhof pathway, and the non‐oxidative part of the pentose phosphate pathway for the breakdown of glucose and xylose. The genomic data explained the high hydrolyzing capacity of this organism by the presence of a large number of carbohydrate active enzymes as well as a large variety of ABC transporters. With transcriptomics the substrate specificity of several of these transporters was predicted. The absence from the genome of some components involved in classical carbon catabolite repression explained the capacity of C. saccharolyticus to grow simultaneously on C5 and C6 sugars. When C. saccharolyticus was cultured on rhamnose, 1.2 propanediol production was observed. This observation could be explained by the presence of a rhamnose pathway, which could be unraveled using the genome sequence as well as the transcriptome data. As hydrogen pressure seems to be an important factor for hydrogen production by C. saccharolyticus, the effect of a high P(H2) was studied in a continuous culture. This showed a

clear shift to lactate and ethanol production at higher P(H2). By comparing the

transcriptome for the high and low P(H2) situation, this shift could be explained by the upregulation of the genes encoding an alcohol and a lactate dehydrogenase. This suggested that the electrons of NADH are most likely disposed of via lactate production, while the electrons of reduced ferredoxin might be shuttled via an oxidoreductase to produce ethanol.Overall this study revealed several new insights in the regulation of hydrogen production by thermophilic anaerobic organisms.

Table of contents

Chapter 1 Introduction and outline Page 7

Chapter 2 Hydrogen production by hyper‐ and extremely thermophilic Page 23 bacteria and archaea: Mechanisms for reductant disposal

Chapter 3 Hydrogenomics of the extremely thermophilic bacterium Page 41 Caldicellulosiruptor saccharolyticus

Chapter 4 Carbohydrate utilization patterns by the extremely thermophilic Page 77 bacterium Caldicellulosiruptor saccharolyticus reveal broad growth substrate preferences

Chapter 5 Analysis of the rhamnose pathway in Caldicellulosiruptor Page 95 saccharolyticus

Chapter 6 The effect of hydrogen pressure on the metabolism of the Page 109 thermophilic bacterium Caldicellulosiruptor saccharolyticus

Chapter 7 General discussion Page 125

References Page 140 Summary Page 165 Samenvatting Page 168 List of publications Page 171 About the Author Page 172 Acknowledgements Page 173 Overview of completed training activities Page 176

6

Chapter 1

Introduction and outline Chapter 1

Introduction

Fossil fuels

To place the research described in this thesis in context, the fossil fuel problem should be addressed. These fuels are named “fossil” as these are the remains of thick layers of organic material of terrestrial or marine origin, which have been exposed for millions of years to heat and high pressure deep down in the earth. This resulted in coal, oil and natural gas, which are rich in hydrocarbons.

Fossil fuel depletion

Not only at the moment, these fossil fuels are the most important energy sources, but as predicted by the U.S. Energy Information, they still will be for the coming decades (200). It is also clear that the total energy demand is still increasing and will be twice as high in 2035 compared with 1990. As fossil fuels need millions of years to form, it is obvious that there will be a problem if energy demands increase. Although estimation of the depletion of the fossil fuel sources is difficult, all experts agree that there will be a need for alternatives in the future. At a certain point in time a situation will be reached where it is no longer economically feasible or physically possible to supply our energy demands with the fossil fuels.

Environmental effects

Another point of concern when using fossil fuels is the production of CO2 when combusted.

For instance, in the US, 90% of all the CO2 produced is coming from fossil fuel combustion

(201). CO2, together with some other gasses (e.g. water vapor, methane and ozone), is a so called greenhouse gas, which contributes to the greenhouse effect. This effect occurs when radiation from the sun, which is reflected by the earth’s surface, is absorbed by greenhouse gas in the atmosphere. Due to this absorption, the energy is trapped in our atmosphere and as a consequence makes the earth warm enough to live on. On the other hand if there is too much greenhouse gas present, the temperature will rise too much and can have very dramatic effects on the climate and biodiversity. Although it receives a lot of attention, the

8

Introduction and outline

problems with these greenhouse gasses are far from solved. Predictions show that the CO2 production from fossil fuels only, will double in 2030 compared to 1990.

Renewable energy sources

As described above these negative effects of the use of fossil fuels, show the need to find alternative energy sources. Renewable energy is the term which is used for these alternatives, and it means that the energy is generated from naturally replenishable resources, such as wind, rain, tides, geothermal heat and sunlight. Sunlight in this case is not only the use of solar panels but also the generation of biomass which can be used as energy source. Renewable energy gained a lot of interest as shown by the investments in this area, which has increased fourfold from 2004 to 2008 to $120 billion. Also on the production side the interest is increasing as can be seen in table 1.1.

Table 1.1 Increased interest in renewable energy shown by numbers; USD: US dollars; GW: gigawatt; GWth: gigawatts thermal; PV: photovoltaic. (161) SELECTED INDICATORS 2006 2007 2008 Investment in new renewable capacity 63 104 120 billion USD (annual) Renewables power capacity (existing, excl. 207 240 280 GW large hydro) Renewables power capacity (existing, incl. 1020 1070 1140 GW large hydro) Wind power capacity (existing) 74 94 121 GW Grid‐connected solar PV capacity (existing) 5.1 7.5 13 GW Solar PV production (annual) 2.5 3.7 6.9 GW Solar hot water capacity (existing) 105 126 145 GWth Ethanol production (annual) 39 50 67 billion liters Biodiesel production (annual) 6 9 12 billion liters

Although these numbers show a clear increase, a lot more is needed to really replace fossil fuels, especially whenit is compared with the total power consumption, which was 18.8

9

Chapter 1

Terawatt in 2008 worldwide, and this will only increase in the future (200). If taken into account that the use of nuclear power has been quite stable in recent years and, because of political and environmental reasons, will most likely not increase in the near future, more investments and research on renewable resources are needed.

Wind energy

Although there is a lot of wind energy available, it is not always easy to harvest. This energy can be used in different ways, like for example, using wind turbines (electricity) or wind mills (mechanical power). In 2008 there was a existing wind power capacity of 121 GW (Table 1) which is a very small portion of the total renewable power capacity. Wind energy is sustainable and wind is widely available, but it also has some clear drawbacks. There is a visual impact of the large turbines and possible negative environmental effects, for example on birds and sealife. On the other hand there is also the problem that wind is not storable and the energy needs to be transformed to other energy carriers like hydropower. This will make wind energy a costly energy source as large investments need to be done if the proportions increase. So far wind energy seems to be more suited for local usage, when the installation costs are low and energy can be used immediately.

Solar energy

Solar energy is a very promising alternative to fossil fuels. If only a small portion of the total radiation (3,850,000 exoJoule per year)(186) of the sun could be captured, it would be enough to fulfill the current energy demand (474 exojoule per year in 2008)(200). There are three main ways of capturing the energy from solar radiation. The most used method is to use the solar power as thermal source to convert solar radiation into heat directly. This happens in so called solar boilers, which in most cases contain water which will be heated by the sun. Another way is to immediately convert solar radiation directly into electricity which happens in photovoltaics. The main barriers for these two methods are the high costs of making solar cells and the dependence on weather, and day and night situations. But advances in the technology and economies of scale, will increase the use of solar energy as energy source.

10

Introduction and outline

A third method in which solar power is used, is different compared to the two others as it is using the solar energy indirectly. In this case biomass, which captured the solar energy, is used an energy source. This topic will be addressed in more detail later in this chapter.

Hydropower

Hydropower is another alternative energy source and it uses the power of moving water by gravity. At the moment this is the largest supplier of renewable energy (Table 1) and this will not change soon. At the moment most of the used hydroelectric power is coming from dammed rivers or lakes coupled to a water driven water turbine and generator. The capacity of such a hydropower plant is depending on the volume of the flow of water as well as the height difference between the source and the outflow. The possibility to adapt the power capacity of such a plant by managing the water flow is a very large advantage of hydropower. Making use of tidal changes in the seas and oceans is another method that can be very promising. In this case the tides are very predictable and with the construction of reservoirs even peaks of high electricity demands can be counterbalanced.

Geothermal power

The heat stored in the earth can also be a renewable energy source and is called geothermal power. Geothermal power is very reliable, sustainable and can be very cost effective, but is so far limited to countries close to tectonic plate boundaries. Recent technological developments make it possible to reach deeper into the earth and will make geothermal energy an option for other regions as well, although energy available will never fulfill the demands.

Energy from biomass

An indirect way of using solar energy is to use biomass as energy source. Biomass on itself is biological material originating from living organisms, having a short carbon cycle. This is the main difference with fossil fuels, as the carbon in this fuels have been taken out of the carbon cycle millions of years ago. This organic material can be used as an energy source in different ways, resulting in different forms of stored energy. The most well known way to

11

Chapter 1

use biomass as energy source is not new, because ever since mankind invented fire, wood has been used to produce heat. Still this is the most used form of renewable energy. Combustion of biomass is the easiest method available, but the energy loss and the problem that it is not usable as a gaseous or liquid fuel asks for conversions of the biomass into different energy carriers. Nowadays a variety of methods are being developed and some of them are already widely used. The methods of converting biomass to other energy carriers can be divided into thermal and biochemical conversions. During thermal conversion, heat is the main mechanism to convert the biomass into other biofuels, and commonly used methods are combustion, pyrolysis and gasification. Biochemical conversions are all the processes in which the enzymes of micro‐organisms or the organisms itself, are used to convert the biomass into other forms of biofuel like biogas, bioethanol, biodiesel or biohydrogen. A recent very important discussion, the food versus fuel debate, indicates that the use of biomass for energy also has its drawbacks. Diverting farmland or crops for biofuel production should not harm the food supply, especially in developing countries. Another

point of attention is that the total CO2 reduction of the overall process should be calculated. This is the reason why nowadays biofuels are divided into three different groups, first, second and third generation biofuels. Although multiple definitions are used for these groups, in general biofuel generated from crops which are only grown for fuel purposes are called first generation (202). Some examples of these first generation crops are, different grain species, corn and sugar cane. The use of these resources will affect the food supply and is an important factor in this food for fuel debate . The second generation biofuels are made from biomass which is non edible, like stems, leaves and husks. Non‐food crops which are grown only for fuel production, like several species of grasses and trees, are used to produce second generation biofuels (202) as long as they are not competing with food production. The use of fast growing crops which need less water and nutrients can increase the biomass yield per hectare of land and reduces the production costs. Technological breakthroughs are still needed to make these second generation biofuels cheaper than fossil fuels, but this is only a matter of time (202).

12

Introduction and outline

All biofuels coming from algae are called third generation biofuels. This field gained a lot of interest in the recent years. The advantage of algae is that they grow relatively fast, but as direct sunlight is needed, sufficient mixing is needed which can be hard and expensive. Although the results for algae to produce biofuel are promising, more research is needed to improve this technology.

Types of biofuel from biomass

As is the case for fossil fuels, different types of fuels are needed. There are options needed for replacing liquid car fuels (e.g. biodiesels, bioalcohols) as well as for instance gaseous fuels to replace natural gas. The most common types of biofuels will be shortly described here.

Bioalcohols

The most commonly produced biofuel worldwide at the moment is bioethanol. Bioethanol , together with propanol and butanol are called the bioalcohols. Although bioethanol is the most common biofuel, especially because of high production in Brazil, biobutanol is claimed to be the best replacement for gasoline, as it can be easily used by normal gasoline engines present in most cars and will produce more energy when combusted than ethanol. In most cases bioalcohol is produced by fermentation by micro organisms or by enzymes derived from them, and in general is derived from wheat, corn and sugar cane. Of course this opens the discussion again about the food for fuel competition, and that is why more and more research is done on the use of biomass from waste streams (39).

Biodiesel

As bioethanol is worldwide the most common biofuel, within Europe this is biodiesel. Biodiesel is quite similar in composition as fossil diesel and consists mainly out of fatty acid (m)ethyl esters (FAMEs). It is produced by a chemical or a enzymatic transesterification of vegetable oils or animal fats (56). In most situations these FAMEs are mixed with fossil diesel in different ratios and are already available at gas stations. The cleanest variant would be pure biodiesel but this needs a larger production capacity than nowadays available, and car engines need some adaptations to drive on it. Mostly used as feedstock for biodiesel are

13

Chapter 1

animal fats and vegetable oils such as the oil of soybean and rapeseed. Also the production of biodiesel from algae is gaining interest nowadays.

Biogas

Biogas is produced by anaerobic digestion of biomass, and mainly consists of methane and carbon dioxide. This gas can be used by a combined heat and gas system (CHP) to produce heat and electricity, or can be used directly as car fuel, or for cooking and heating. For this latter applications the biogas can be mixed into the natural gas network, though the biogas needs to be purified first. An interesting possibility is to mix animal manure and crop residues to produce biogas on farms.

Solid biofuels

As most biomass is available in solid form, the easiest way of using it is to burn it. For example wood, sawdust and agricultural waste can be easily combusted yielding heat. The only treatment needed in most cases is to densify it by grinding. A drawback of combusting solid biomass directly is that a considerable amount of pollutants can be formed.

Biohydrogen

Hydrogen is not only a fuel, it is also widely used in industry, mainly for hydrogenation of unsaturated substrates . Hydrogen is the most abundant element available, although almost always trapped in other molecules like water. Hydrogen is the lightest gas and has the highest combustion energy release per mass unit. A drawback of hydrogen as a fuel is that a very low temperature or a very high pressure is needed to store it as a liquid. During combustion no carbon dioxide is formed which makes hydrogen a very clean fuel. This also makes hydrogen a very useful energy storage material, if non storable energy is produced from e.g. wind or solar energy. A often used technique for this is steam reforming of natural gas. At high temperatures the water vapor reacts with methane yielding carbon monoxide and hydrogen (Reaction 1).

Reaction 1 CH4 + H2O → CO + 3 H2

14

Introduction and outline

In a second stage, more hydrogen is formed by a water shift reaction performed at about 130°C (Reaction2)

Reaction 2 CO + H2O → CO2 + H2

This mixture of CO and H2 is called syngas, and can be used as a fuel on its own or cleaned to pure hydrogen. A comparable technique which is also resulting in syngas, is the gasification of biomass or coal. During this gasification oxygen and/or steam is used at high temperatures (>700°C) to react with the raw material, resulting in the production of carbon monoxide and hydrogen. The hydrogen produced in this way from biomass can also be considered a biofuel. Another way of producing hydrogen is by water splitting, which is a reaction to separate water into hydrogen and oxygen by applying electricity (Reaction 3).

Reaction 3 2 H2O → O2 + 2 H2

The electricity needed to perform this reaction can come from different resources. The amount of energy to perform this reaction is high and so far hydrogen produced from water cannot compete with hydrogen produced from fossil fuels. In the future this can change, especially if the electricity used can originate from biofuels. It is also possible to produce hydrogen chemically, in which several materials can react with water or acids to produce hydrogen. The biggest problem with this method is the reduced metals which are used in considerable amounts for their production and are in general to expensive compared to the value of hydrogen.

Biological hydrogen production

Hydrogen produced by gasification can be called a biofuel, but in general the energy needed for heat production is coming from fossil fuel resources. Hydrogen which is produced by microorganisms from biomass is a more sustainable method. Although the results of current research are promising, large scale applications are not yet applied. Fermentative hydrogen

15

Chapter 1

production in general can be divided into dark and photo fermentation. The main difference between these two forms of hydrogen production is the need of light as an energy source for the light fermentation .

Dark fermentation

In the case of dark fermentation, anaerobic and facultative aerobic bacteria are used to convert the biomass or the hydrolysates of biomass into hydrogen, carbon dioxide and organic acids. This fermentation can be performed by different groups of organisms at different temperatures. Mesophilic fermentation takes place between 15 and 40 °C, while thermophilic hydrogen production takes place at temperatures between 45 and 80 °C. Thermophilic hydrogen production has the general advantages of processes that take place at higher temperatures, like lower viscosity, less chance of contamination, better mixing, no need for cooling and higher reaction rates. In addition, a lot of thermophilic organisms have a wide variety of hydrolytic enzymes, and also the thermodynamics of the reactions is more favorable at higher temperatures. For these last two advantages I would like to refer to chapter 2 where these aspects are described in more detail.

Photofermentation

By performing a dark fermentation a maximum of 4 moles of hydrogen can be produced from one mole of glucose, provided that the temperature is relatively high and the hydrogen concentration is kept low (Chapter 2). The rest of the potentially available energy will stay trapped in the organic acids as it is thermodynamically not possible to oxidize these anaerobically without input of extra energy (Reaction 4).

‐1 Reaction 4 2 CH3COOH + 4 H2O → 4 CO2 +8 H2 ΔG’0= +209.2 kJ mol

During photofermentation, light is used as an extra energy source to convert organic acids into hydrogen and carbon dioxide. To capture the energy available in light, photosynthesis is used by photofermentative organisms, like algae, cyanobacteria and species of the purple non‐sulphur bacteria genus Rhodobacter. Especially, the latter holds a significant promise as photofementative hydrogen producer (16). To perform optimally the construction of the

16

Introduction and outline

photofermenter is very important. This means that a large surface is needed and proper mixing should occur inside the fermenter. So far the production rates in photofermenters are much lower than in dark fermentation fermenters.

Hyvolution

The research described here is part of a larger project, called Hyvolution. This European project has the aim to develop a blueprint for decentral hydrogen production from biomass. Within this project the basis is a combination of a dark‐ and a photofermentation in which biomass is converted to pure biohydrogen (Figure 1.1) (37). The biomass will be hydrolyzed and pumped into the dark fermenter in which C. saccharolyticus or comparable organisms

will degrade the sugars into hydrogen, CO2 and organic acids. The effluent containing the acids will be pumped into the photofermenter in which with light as energy source the

organic acids are further degraded to hydrogen and CO2. The H2 produced will be separated from the gas mixture, and very pure hydrogen can be obtained. Within this project research is being performed on all steps of this process, from biomass pretreatment till gas separation. The fermentations are not only tested on lab scale but also upgraded to for instance a 600‐liter fermenter for the dark fermentation.

Gas separator

BIOMASS Gas separator CO2

H2 + CO2 H2 + CO2

carbohydrates organic acids H 2

organic acids

LIGHT

C6 H12 O6 + 2H2 O CH3COOH + 2H2O 2CO2 + 2CH3COOH +4H 2 2CO2 +4 H2

Figure 1.1 The Hyvolution process Besides the technical and microbiological part of the process also models are made to simulate the performance of the process and to do the cost estimations. Within the project there is also focus on the societal integration of the process.

17

Chapter 1

Caldicellulosiruptor saccharolyticus

The research described in this thesis covers the dark fermentation step of the Hyvolution system. For this research Caldicellulosiruptor saccharolyticus (Figure 1.2) was chosen as model organism, because of its excellent features for hydrogen production (207).

Figure 1.2 Electron microscopy picture of Caldicellulosiruptor saccharolyticus (A. Pereira and M. Verhaart)

C. saccharolyticus was isolated from the thermal area Taupo in New Zealand and first experiments showed immediately the high hydrolytic capacities of this organism (162, 185). In 1993 Rainey et al (158) proposed that this organism was part of a new branch within the Bacillus / Clostridium subphylum of the Gram‐positive bacteria and the same group named it C. saccharolyticus (157). C. saccharolyticus is an asporogenous bacterium with a growth optimum of 70 °C and a pH optimum of 7.0. The cells are straight rods which under normal conditions appear as single and double cells. Since the isolation of C. saccharolyticus 9 Caldicellulosiruptor other species have been described (Figure 1.3)(28, 61, 74, 125, 127, 141). ‘Anaerocellum thermophilum’ has been recently renamed to Caldicellulosiruptor bescii (227).

18

Introduction and outline

Figure 1.3 Phylogenetic tree containing of the Caldicellulosiruptor species (61)

Besides the general thermophilic advantages, C. saccharolyticus has other features why it was chosen as model organism. C. saccharolyticus has the capacity to grow on a wide variety of monosugars as well as directly on cellulose and other complex substrates like agarose and alginic acid, hemicellulose‐rich pine tree wood shavings, Miscanthus hydrolysates and paper sludge (43, 69, 78, 86, 157). Another useful feature is that it can simultaneously use C5 and C6 sugars (Chapter 3). Also the availability of the genome sequence and the possibilities to do transcriptomics makes this organism a ideal candidate for a model organism to study anaerobic thermophilic hydrogen production (Chapter 3). On the other hand enzymes responsible for the breakdown of complex substrates, like cellulose and xylan, by C. saccharolyticus were cloned, purified and characterized (7, 48, 71, 72, 83, 110, 146). The most important feature of a hydrogen producer is the hydrogen yield obtained from biomass. Where most mesophiles do not reach higher yields than 2 moles of hydrogen per mole of hexose, C. saccharolyticus has been shown to produce up to 3.6 moles of hydrogen per mole of hexose (43), which is close to the theoretical maximum of 4 (91, 195). Higher values are hard to reach as also electrons are used for biomass production. C. saccharolyticus employs the Embden Meyerhof pathway to oxidize sugars to pyruvate, followed by further oxidation to acetate. These pathways result in the formation of the

19

Chapter 1

reducing equivalents, NADH and reduced ferredoxin, which can be oxidized by two hydrogenases to produce hydrogen. The hydrogenases present in C. saccharolyticus showed high homology with the hydrogenases present in Thermoanaerobacter tengcongensis, which were described to be NADH‐ and ferredoxin‐dependent(188). The proposed NADH‐ dependent hydrogenase also showed homology to a bifurcating hydrogenase in Thermotoga maritima, which uses both NADH and reduced ferredoxin in a 1:1 ratio (180). So far, this bifurcating behavior has not been shown for the C. saccharolyticus hydrogenases. As in most organisms the main energy carrier in C. saccharolyticus is ATP, but recently pyrophosphate (PPi) has been described to be an additional energy carrier for this organism (20). This research showed that C. saccharolyticus can utilize a PPi‐dependent 6‐phosphofructokinase as an alternative for the ATP‐dependent 6‐phosphofructokinase as well as a PPi‐dependent pyruvate phosphate dikinase instead of the ADP‐dependent pyruvate kinase (Figure 1.4). During the exponential growth phase the PPi levels were high and the ATP levels were relatively low, whereas in the stationary phase the PPi/ATP ratio decreased 13‐fold. Willquist and van Niel (219) also showed that this PPi is a strong modulator of the lactate dehydrogenase, which indicates that PPi has a role in the acetate to lactate switch often observed in batch cultures. Altogether this suggests that PPi has an important role in the central metabolism of C. saccharolyticus.

20

Introduction and outline

Figure 1.4 The central glycolytic pathway of Caldicellulosiruptor saccharolyticus consisting of the enzymes: (1) glucokinase; (2) glucose‐6‐phosphate isomerase; (3a) ATP‐dependent 6‐phosphofructokinase; (3b) PPi‐ dependent 6‐phosphofructokinase; (4) Fructose 1,6‐bisphosphate aldolase (4a) Triose phosphoisomerase; (5) Glycerol‐3‐phosphate dehydrogenase; (6) phosphoglycerate kinase; (7) phosphoglycerate mutase; (8) phosphopyruvate hydratase (enolase); (9a) Pyruvate kinase; and (9b) pyruvate, phosphate dikinase. (20)

The research described in this thesis and future developments like a recently described proteomics set up (130) and a to be designed genetic system will increase the knowledge and the possibilities of using C. saccharolyticus for sustainable hydrogen production as a alternative to fossil fuels.

21

Chapter 1

Outline of this thesis

Chapter 2 is a review on the thermodynamics of anaerobic thermophilic hydrogen production. Different mechanisms of reductant transposal are described for three different groups of organisms. In chapter 3 the annotated genome sequence of C. saccharolyticus is described and the glycolytic enzymes are compared with other thermophilic hydrogen producers. In addition, the transcriptome data of C. saccharolyticus when grown on rhamnose , glucose, xylose and a mixture of the last two are discussed . In this paper we also show the capacity of C. saccharolyticus to grow on xylose and glucose simultaneously. In chapter 4 we show with transcriptome data, the capacity of C. saccharolyticus to grow on a wide variety of single and more complex sugars. In this chapter the main focus was to describe the substrate preferences of the large number of transporters present. Next to this a bioinformatical analysis is described to show the absence of a classical catabolite repression system in C. saccharolyticus. In chapter 5 a proposed rhamnose pathway is described in which the transcriptome data are coupled to product pattern analysis. In addition, experiments were done to show the different mechanisms of reducing equivalent disposal in C. saccharolyticus when grown on rhamnose compared with other sugars. In chapter 6 the effect of a high hydrogen pressure was analyzed in a chemostat. With the combination of product pattern analysis and transcriptome date insight is obtained in the regulation of reducing equivalent disposal under high P(H2) conditions. Chapter 7 is the general discussion and chapter 8 is a summary of this thesis.

22

Chapter 2

Hydrogen production by hyper‐ and extremely thermophilic bacteria and archaea: Mechanisms for reductant disposal

Verhaart, M. R. A., Bielen, A. A. M., v. d. Oost, J. , Stams, A. J. M. and Kengen, S. W. M. (2010) Environmental technology 31:993 ‐ 1003. Chapter 2

Abstract

Hydrogen produced from biomass by bacteria and archaea is an attractive renewable energy source. However, to make its application more feasible, micro‐organisms are needed with high hydrogen productivities. For several reasons, hyper‐ and extremely thermophilic bacteria and archaea are promising is this respect. In addition to the high polysaccharide hydrolyzing capacities of many of these organisms, an important advantage is their ability to use most of the reducing equivalents (e.g. NADH, reduced ferredoxin) formed during

glycolysis for the production of hydrogen, enabling H2/hexose ratios between 3 and 4. So, despite the fact that the hydrogen yielding reactions, especially the one from NADH, are thermodynamically unfavourable, high hydrogen yields are obtained. In this review we focus on three different mechanisms that are employed by a few model organisms viz. Caldicellulosiruptor saccharolyticus and Thermoanaerobacter tengcongensis, Thermotoga maritima and Pyrococcus furiosus, to efficiently produce hydrogen. In addition, recent developments to improve hydrogen production by hyper‐ and extremely thermophilic bacteria and archaea are discussed.

Introduction

With the declining availability of fossil fuels and the unfavourable environmental aspects coupled to their use, there is an obvious need for alternative and cleaner energy sources. Hydrogen gas is an attractive candidate as a future fuel, for several reasons. Hydrogen gas is non‐toxic and its use in a fuel cell or combustion engine only yields pure water without

production of the greenhouse gas CO2. However, hydrogen is a sustainable fuel only when it is produced from renewable resources like wind, sunlight, geothermal energy or biomass. Presently, a wide variety of microorganisms is known to form hydrogen from biomass in light‐dependent and light‐independent processes. The micro‐organisms capable of producing hydrogen can be roughly divided into three groups: phototrophs, facultative anaerobes, and strict anaerobes. In this review we will focus on the strict anaerobes and more specifically on the hyper‐ and extremely thermophilic bacteria and archaea, which are excellent hydrogen producers (91). To produce the maximum amount of hydrogen, the biomass derived carbohydrates need to be oxidized entirely to acetate (or even further to

24

Review: Hydrogen production by thermophilic organisms

CO2) and all the reducing equivalents (e.g. NADH, reduced ferredoxin) should be recycled producing hydrogen and not result in branched pathways leading to other end products (e.g. lactate, ethanol). Although hydrogen production from NADH and reduced ferredoxin is energetically not very favourable, these organisms are still capable of producing high amounts of hydrogen. By describing the catabolic pathways that have been elucidated for some model organisms, the mechanisms of recycling reducing equivalents will be explained. In addition, the future perspectives for improving biohydrogen production are discussed.

Why thermophilic hydrogen production?

Thermophilic hydrogen production benefits from some general advantages of performing processes at elevated temperatures, like a lower viscosity, better mixing, less risk of contamination, higher reaction rates and no need for reactor cooling (217). In addition, the hydrolytic capacity of especially thermophilic bacteria and the thermodynamics of the reactions at elevated temperatures play an important role, and therefore will be discussed here in more detail.

Hydrolytic capacity of extremophiles

The organisms described in this review are all extremely thermophilic or hyperthermophilic anaerobes, with temperature optima above 65°C and 80°C, respectively. Many representatives of this diverse group of microorganisms are capable of performing fermentative hydrogen production from a wide variety of substrates. To make hydrogen production sustainable, organisms are needed which can produce hydrogen directly or indirectly from biomass. The main substrates for fermentative hydrogen producers are carbohydrates, and because cellulose and hemicellulose are the most abundant polysaccharides in nature, xylose and glucose are the predominant monomeric sugars available (89)]. In addition, starch and sucrose can be present in high amounts in plants as they are used as storage material. Especially the bacteria described in this review have the capacity to hydrolyze a wide range of substrates derived from biomass. For instance, by analyzing the functional categories of Clusters of Orthologous Groups of proteins (COGs) responsible for carbohydrate metabolism, van de Werken et al. showed that (204)

25

Chapter 2

Caldicellulosiruptor saccharolyticus has a very high number of transporters for a very broad range of substrates (e.g. 177 ABC‐transporter genes). Also a comparative study in three model extreme thermophiles (Pyrococcus furiosus, Thermotoga maritima and C. saccharolyticus ) showed a wide variety of glycoside hydrolases and transferases (208).

Thermodynamics of Hydrogen Formation

The hydrogen yield is strongly determined by the thermodynamics of the reactions leading to hydrogen. Under standard conditions (1 Molar concentrations of the reactants, 25 °C, pH

7) the complete oxidation of glucose to CO2 and H2 has a positive Gibbs energy change and, thus, will not proceed without input of extra energy.

‐ + 0 Equation 1: glucose + 12 H2O  6 HCO3 + 6H + 12 H2 ΔG ’= +3.2 kJ/mol

Numerous fermentation experiments have shown that, under optimal conditions, the oxidation of glucose will at best result in the formation of 4 molecules of hydrogen per

molecule of hexose in addition to acetate and CO2 (Table 2.1). The production of acetate yields the maximum amount of ATP for the organism, but this can only occur if all the reducing equivalents are transferred to hydrogen. Theoretically this is possible, but in natural environments this can only be reached when the hydrogen pressure is kept low by hydrogen consuming microorganisms (177)

‐ ‐ + 0 Equation 2: glucose + 4 H2O  2 acetate + 2 HCO3 + 4H + 4 H2 ΔG ’= ‐206.1 kJ/mol

In most fermentative hydrogen producers catabolism (e.g. via the Embden‐Meyerhof pathway) yields reducing equivalents in the form of NADH (GAP dehydrogenase reaction) and reduced ferredoxin (pyruvate:ferredoxin oxidoreductase reaction). The midpoint redox potentials of NAD+/NADH and oxidized ferredoxin/reduced ferredoxin are ‐320 and ‐398 mV, respectively (195). Recycling of these electron carriers can be accomplished by a variety of reactions, like the production of lactate by reduction of pyruvate by NADH, but the likelihood of these reactions to occur is determined by the standard Gibbs free energy change (ΔG0’) of the individual conversions (Table 2.2). From this Table it can be concluded that under standard conditions the reaction resulting in hydrogen by the reduction of

26

Review: Hydrogen production by thermophilic organisms protons with NADH is thermodynamically unfavourable, the midpoint redox potential of the + couple H /H2 being ‐414 mV. The situation for ferredoxin is more favourable but formation of other products like ethanol and lactate is thermodynamically more feasible. Nevertheless, for certain thermophiles (Table 2.1) the amount of hydrogen produced has been shown to reach 4 molecules per molecule of glucose, suggesting that despite the more favourable

production of lactate and ethanol, NADH is used for proton reduction to form H2 . There are several reasons that can explain these observations. Firstly, ΔG0’ values are calculated for standard conditions and these can differ considerably from the physiological conditions. The actual Gibbs energy change can be calculated using the following equation in which A and B are the substrates and C and D the products of this reaction:

Equation 3: ΔG’ = ΔG0’ + RT ln ([C][D]/ [A][B])

‐2 If the H2 concentration is kept low (e.g. P(H2) of 10 kPa) even proton reduction with NADH becomes exergonic (‐4.7 kJ/mol). Secondly, the temperature at which the reaction takes place affects the thermodynamics, according to ΔG0 = ΔH –TΔS0 (189). Making use of reported thermodynamic data (10, 195), it can be calculated that at higher temperatures the Gibbs free energy change for the overall reaction from glucose to acetate (Equation 2)

becomes more favourable . Figure 2.1 shows the combined effect of temperature and P(H2) (according to Equation 3). If we then assume that per mole of ATP approximately 70‐75 kJ is needed (177, 195), and that for the total reaction 4 moles of ATP will be produced, the total amount of free energy has to be lower than ‐280 to ‐300 kJ (indicated as grey area in Figure

2.1). The indicated arrows then show that at 25°C the P(H2) needs to be as low as 0.05 kPa to make the reaction feasible, in contrast to only 5 kPa at 100°C. Table 2.1. Overview of thermophilic and hyper‐/extremely thermophilic hydrogen producing microorganisms.

27

Chapter 2

Table 2.1. Overview of thermophilic, hyperthermophilic and extremely thermophilic hydrogen‐producing microorganisms.

Organism Domain T‐grown Culturing Substrate Reported end H2/hexose Reference type products Thermophiles

Thermoanaerobacterium B 55 batch cellobiose acetate, lactate, 0.87 (182) saccharolyticum YS485 ethanol

Thermoanaerobacterium B 60 batch starch acetate, 2.8 (139) thermosaccharolyticum ethanol, PSU‐2 butyrate Clostridium thermocellum B 60 chemostat α‐cellulose acetate, lactate, 1.65 (118) ATCC 27405 ethanol, formate Hyper‐ and extreme thermophiles Thermotoga elfii DSM B 65 controlled sucrose acetate 3.3 (207) 9442 batch 65 batch glucose acetate 3.3 (44) Thermotoga neapolitana B 80 batch glucose acetate, lactate 2.4 (50) DSM 4359 80 controlled glucose/ acetate, lactate 3.3 (43) batch xylose 85 batch glucose acetate, lactate 3.8 (132) 77 batch glucose acetate, 3.2 (136) butyrate Thermotoga maritima B 80 batch glucose acetate 4 (179) DSM 3109 Caldicellulosiruptor B 70 controlled sucrose acetate, lactate 3.3 (207) saccharolyticus DSM 8903 batch 72 controlled glucose/ acetate, lactate 3.4 (43) batch xylose Thermoanaerobacter B 75 batch starch acetate, ethanol 2.8 (188) tengcongensis JCM11007 75 batch glucose acetate 4 (188) Thermococcus A 85 chemostat starch acetate, alanine 3.3 (88) kodakaraensis TSF100 Pyrococcus furiosus DSM A chemostat maltose acetate, 2.9 (176) 3638 butyrate 90 batch cellobiose acetate, alanine 2.8 (92) 90 batch maltose acetate, alanine 3.5 (92) 90 chemostat maltose acetate, alanine 2.6 (36) 90 chemostat cellobiose acetate, 3.8 (36) alanine, ethanol

28

Review: Hydrogen production by thermophilic organisms

‐ + Figure 2.1 ∆G’ for the reaction: glucose + 4 H2O  2 acetate‐ + 2 HCO3 + 4 H + 4 H2 at different temperatures

and different P(H2) (10, 195). Grey area indicates ∆G’ needed to produce 4 moles of ATP from one mole of glucose (≈ 280‐300 kJ). The arrows indicate the corresponding P(H2) values at 100°C and 25°C.

Table 2.2. Gibbs free energy values for different fermentative reactions (91). Data were calculated using (195); (9); (10) ΔG0’ Fermentative reaction kJ/reaction

NADH +H+ + pyruvate‐  NAD+ + lactate‐ ‐25.0 2NADH +2H+ + acetyl‐CoA  2NAD+ + ethanol + CoA ‐27.5 + ‐ + + NADH +H + pyruvate + NH4  NAD + alanine + H2O ‐36.7 + + NADH +H  NAD + H2 +18.1 + 2 Ferredoxin(red) + 2H  2 Ferredoxin(ox) + H2 +3.1

A third explanation for hydrogen formation from NADH can be that the reaction might be pushed in the right direction by reversed electron transport. This can only take place if the NAD‐dependent hydrogenase resides in the membrane. This phenomenon of energy‐driven uphill oxidation of NADH has been described for Clostridium tetanomorphum, in which a

29

Chapter 2

sodium gradient is used by a NADH dependent oxidoreductase to accomplish the endergonic reduction of ferredoxin. The reduced ferredoxin is then used for hydrogen production (23). So far this has not been described for thermophiles and needs further research. Finally, a fourth mechanism was recently presented, which entails that the responsible hydrogenase is able to use the energetically more favourable oxidation of ferredoxin to push the less favourable production of hydrogen from NADH in T. maritima (180) . This process will be explained in more detail below.

Mechanisms of hydrogen production

The enzymes responsible for hydrogen production are the hydrogenases (EC 1.12.99.6, EC 1.12.7.2), which catalyze the reversible oxidation of molecular hydrogen. In one direction hydrogen uptake is coupled to the reduction of electron acceptors, while in the other direction proton reduction by reduced electron carriers will result in hydrogen production. Hydrogenases can be divided into different groups based on their metal content. Two main types of hydrogenases often found in anaerobic thermophiles are the Fe‐only and the [Ni‐ Fe]‐ hydrogenases (49, 65, 66, 80). Hydrogenases can use different types of electron carriers (e.g. NAD, NADP, FAD, ferredoxin) which are reduced during the oxidation steps in the central metabolic pathways. The most important oxidation steps in the glycolytic pathway are the conversion of glyceraldehyde‐3‐P to 3‐P‐glycerate and the conversion of pyruvate to acetyl‐CoA. In most fermentative hydrogen producers NADH and reduced ferredoxin are generated in these steps and need to be recycled to keep the metabolism going. The situation in some hyperthermophilic archaea (Pyrococcus, Thermococcus) is strikingly different as a modified Embden Meyerhof pathway, resulting in only ferredoxin as reducing equivalent, is employed. Among the different (hyper)thermophilic hydrogen producers essentially different mechanisms are used to dispose of the reductant which is generated during the oxidation of the substrates. Here, these mechanisms will be addressed by discussing the model organisms C. saccharolyticus and Thermoanaerobacter tengcongensis, T. maritima and P.furiosus

30

Review: Hydrogen production by thermophilic organisms

Caldicellulosiruptor saccharolyticus and Thermoanaerobacter tengcongensis

The first mechanism concerns the proposed situation in C. saccharolyticus (Figure 2.2) and T. tengcongensis. According to the currently available data, these two bacterial species are similar in their hydrogen production machinery and will therefore be treated together. C. saccharolyticus is an asporogenous thermophilic bacterium which is capable of using a wide variety of substrates, including cellulose, and which yields high hydrogen levels (78, 79, 157, 209). As fermentation end products acetic acid, lactic acid and ethanol have been reported, in addition to carbon dioxide and hydrogen. The production of lactic acid and ethanol is normally low resulting in relatively high hydrogen levels. Because of its favourable fermentation pattern and its cellulolytic capacity, C. saccharolyticus has been selected for further research on hydrogen production from biomass (207) (45). Recently also the genome sequence of this organism has been obtained which gave new opportunities for research (204). T. tengcongensis is also able to produce hydrogen from a wide variety of substrates. The main differences compared to C. saccharolyticus are that T. tengcongensis can respire with thiosulfate and sulphur, cannot grow on cellulose, and may form endospores (190). During the glycolysis via the Embden Meyerhof pathway, both species produce NADH as well as reduced ferredoxin. It is well shown that reoxidation of reduced ferredoxin can be

catalyzed by a H2 producing membrane bound NiFe hydrogenase (e.g. T. tengcongensis (188)) or by a cytoplasmic monomeric Fe‐only hydrogenase as described for Clostridium pasteurianum (4). Hydrogen production directly from NADH is, however, unusual and for most mesophilic organisms an NADH:ferredoxin oxidoreductase (Nfo) is proposed that transfers the reducing equivalents from NADH to ferredoxin (23). These Nfo encoding gene clusters have been identified in the genomes of several thermophilic hydrogen producers (e.g. Clostridium thermocellum and T. maritima), but seem to be missing in the genomes of C. saccharolyticus and T. tengcongensis (204). Experiments done in T. tengcongensis have revealed that this organism uses a NADH‐dependent hydrogenase which at low hydrogen concentrations directly uses NADH to produce hydrogen (188). In the genome of C. saccharolyticus a hydrogenase complex with high homology to the enzyme complex of T. tengcongensis has been identified suggesting that also this organism is capable of directly

31

Chapter 2

using NADH for hydrogen production (204). A fact that has to be kept in mind when directly

using NADH to produce hydrogen (E0’= ‐414) is the unfavourable midpoint potential of

NADH compared to reduced ferredoxin (E0’= ‐320 mV for NADH, E0’= <‐398 mV for ferredoxin).

Figure 2.2 Scheme of typical carbon (grey)‐ and electron (black) flow during glucose fermentation by C. saccharolyticus at low and high P(H2). 1, Glyceraldehyde‐3‐P dehydrogenase (GAPDH); 2, Pyruvate: ferredoxin oxidoreductase (POR); 3, NADH‐dependent hydrogenase; 4, Ferredoxin‐dependent hydrogenase; 5, Lactate dehydrogenase (LDH); 6, NADPH:ferredoxin oxidoreductase; 7, Alcohol dehydrogenase and acetaldehyde dehydrogenase (45, 188, 204)

As discussed above, the obtained yield of almost 4 molecules of hydrogen per molecule of glucose can only be reached at elevated temperatures and at very low hydrogen pressure (195). As a result, when facing high hydrogen pressures, the NADH produced by this organism will be used by lactate dehydrogenase to produce lactate instead of acetate and

hydrogen. We were able to show an increase in ethanol production at higher P(H2) (M. Verhaart, unpublished results) which suggests that C. saccharolyticus somehow converts the NADH and/or the reduced ferredoxin to NADPH which is most likely the cofactor for its alcohol dehydrogenase and acetaldehyde dehydrogenase (188). In the case of T.

32

Review: Hydrogen production by thermophilic organisms tengcongensis it seems that mainly ethanol, instead of lactate, is produced as an alternative to hydrogen (188). However, this suggests that T. tengcongensis has the capacity to convert the generated NADH to NADPH. (188). But so far this has not been clarified and needs further study.

Thermotoga maritima

The situation in the anaerobic thermophilic bacterium T. maritima (75) is different from the other previously described species. Although this organism also uses the Embden Meyerhof pathway for glycolysis resulting in acetate, lactate, ethanol, carbon dioxide and hydrogen, recycling of the reducing equivalents seems to differ. Recently, a trimeric Fe‐only hydrogenase from this organism has been described that uses the reducing equivalents from both NADH and reduced ferredoxin in a 1:1 ratio to produce hydrogen (Equation 4) (180). This so‐called bifurcating enzyme apparently uses the exergonic oxidation of ferredoxin to drive the unfavourable oxidation of NADH (Figure 2.3). To maintain this 1:1 ratio of reducing equivalents from Fd and NADH, T. maritima seems to use a Fd:NADH oxidoreductase to supply the bifurcating hydrogenase. But so far this oxidoreductase has not been identified.

+ + Equation 4: NADH + 2 Ferredoxin(red) + 3H  NAD + 2 Ferredoxin(ox) + 2 H2

In the case of higher hydrogen pressures a switch to lactate production has been observed (75). This, however, does not affect the bifurcating mechanism of the hydrogenase because in all situations the 1:1 ratio of reducing equivalents from NADH to ferredoxin is maintained.

33

Chapter 2

Figure 2.3 Scheme of typical carbon (grey)‐ and electron (black) flow during glucose fermentation by T. maritima at low and high P(H2). 1 Glyceraldehyde‐3‐P dehydrogenase (GAPDH); 2, Pyruvate:ferredoxin oxidoreductase (POR); 3, Bifurcating NADH/Ferredoxin dependent hydrogenase ; 4, NADH:ferredoxin oxidoreductase (Nfo); 5, Lactate dehydrogenase(LDH); 6, Ferredoxin‐dependent hydrogenase

Although homologs of the bifurcating hydrogenase for T. maritima have been identified in genomes of other organisms more research needs to be done to verify the proposed mechanism. For instance, the NADH‐dependent hydrogenase from T. tengcongensis, which is a homolog of the bifurcating enzyme, showed hydrogenase activity with NADH in the absence of ferredoxin (188). The same type of experiments performed with cell free extract of T. maritima showed that hydrogenase activity could only be shown in the presence of reduced ferredoxin. Whether T. maritima can also use reduced ferredoxin solely by a non‐ bifurcating hydrogenase still has to be shown. A ferredoxin‐dependent hydrogenase as described for T. tengcongensis (188) could not be identified in T. maritima. On the other hand this organism contains a gene cluster (Mbx operon; 13 genes) which shows high homology to gene clusters, present in several archaea like Pyrococcus furiosus, and which encodes a ferredoxin‐dependent membrane bound hydrogenase complex (172, 184). Whether this mbx operon in T. maritima is also coding for a similar type hydrogenase needs further research.

34

Review: Hydrogen production by thermophilic organisms

Pyrococcus furiosus

The final mechanism discussed in this review for recycling reducing equivalents is present in the archaeon, Pyrococcus furiosus from the order of Thermococcales (53). This hyperthermophilic sulphur reducing heterotrophe is the organism with the highest optimum temperature discussed here, and can ferment a variety of sugar‐ or amino acid based oligomers and polymers, however, cellulose and xylan cannot be used as carbon source. In the metabolism of this organism a modified Embden Meyerhof pathway is used to produce acetate, (92, 213). In addition to modifications at the level of the kinases, which use ADP instead of ATP, this pathway is essentially different from the bacterial type in that 8 moles of reduced ferredoxin are produced instead of the more common 4 moles of reduced ferredoxin and 2 moles of NADH (figure 2.4). This is the result of an exchange of the classical glyceraldehyde‐3‐ phosphate dehydrogenase (GAPDH) by a glyceraldehyde‐3‐phosphate oxidoreductase (GAPOR). Consequently the oxidation of glyceraldehyde‐3‐phosphate (GAP)

is not Pi dependent and does not involve the intermediate 1,3‐biphosphoglycerate and no ATP is formed by substrate level phosphorylation (93). P. furiosus, has two cytoplasmic hydrogenases and a membrane bound hydrogenase complex (184). This latter hydrogenase complex is proposed to accept electrons from ferredoxin directly and is encoded by the mbh operon and was shown to have a very high hydrogen evolving activity compared to other hydrogenases. The genes of the mbh operon code for a 14 subunit transmembrane hydrogenase complex (184), which also can function as a proton pump to generate a proton motive force that drives the synthesis of ATP (172). This complex is not able to use NADPH

as an electron donor and cannot catalyze the reduction of elemental sulphur to H2S, in contrast to the cytoplasmic hydrogenases. These cytoplasmic hydrogenases normally reduce sulphur but can switch to hydrogen production when sulphur is not available. Since the latter are not able to use reduced ferredoxin directly, they are probably not involved in hydrogen production. In the case of a high hydrogen pressure a ferredoxin:NADP oxidoreductase (or sulphide dehydrogenase) can transfer electrons from ferredoxin to NADP or polysulfide. Subsequently, the generated NADPH can be reoxidized to produce alanine by

glutamate dehydrogenase and alanine aminotransferase (92). In contrast to the earlier

35

Chapter 2

mentioned bacteria, a ferredoxin:NADP oxidoreductase has been identified and characterized for P. furiosus (117).

Figure 2.4 Scheme of typical carbon (grey)‐ and electron (black) flow during glucose fermentation for

Thermococcales at low and high P(H2). 1, Glyceraldehyde‐3‐P:ferredoxin oxidoreductase (GAPOR); 2, Pyruvate:ferredoxin oxidoreductase; 3, Ferredoxin dependent hydrogenase; 4, NADPH:ferredoxin oxidoreductase; 5, Glutamate dehydrogenase and alanine aminotransferase.

Improving biohydrogen production, a future perspective

The hyperthermophiles in general (Table 2.1), and the discussed model organisms in particular, show that hydrogen production at elevated temperatures is more feasible compared to more moderate temperatures. Although different types of enzymes or enzyme complexes are employed by the different model organisms, all are approaching the apparent limit of 4 H2 per hexose. Several strategies can be adapted for further optimizing

fermentative hydrogen production with respect to H2 yield (mol/C6 sugar equivalent) and H2 production rate (volumetric productivity (mmol l‐1 h‐1) or specific production rate (mmol gCDW‐1 h‐1)). Particularly, experimental and reactor design helps to increase these values (59, 134, 215). In view of the complex biomass composition, degradation by designed microbial consortia (12, 29, 230) instead of pure cultures might be of particular interest, potentially making use of the broad hydrolytic capabilities of different hyperthermophilic

36

Review: Hydrogen production by thermophilic organisms organisms (22, 59). However, the hydrogen yield is limited by the organism’s existing metabolic pathways and under fermentative conditions constrained to 4 molecules of H2 per molecule of glucose (eq.2). But for biohydrogen production to be economically feasible, yields should surpass this theoretical limit (221). Thermodynamic and metabolic constraints suggest that it would be impossible to find a single organism capable of the complete conversion of sugar‐based substrates to hydrogen by fermentation yielding 12 H2/glucose (eq. 1), because there is no free energy available to drive ATP synthesis.

In vitro systems

To increase the total energy extraction (H2 yield) of the overall substrate conversion one could make us of designed in vitro systems. Building up an entire hydrogen producing pathway in a reaction vessel, while making use of the large enzyme arsenal available for substrate degradation to hydrogen. A simple two‐component system demonstrated the proof of principle of in vitro hydrogen production (221). By combining Thermoplasma acidophilum’s glucose dehydrogenase and a NADPH‐dependent hydrogenase from P. furiosus, glucose could be converted to glucuronic acid and hydrogen, concomitant with the recycling of the co‐factor. This research was expanded to include a broader range of substrates including different renewable resources (222). In vitro system research was taken to a next level when the enzymes of the oxidative pentose phosphate cycle were combined with the NADPH‐ dependent sulfhydrogenase of P. furiosus (228). In this system each mole of glucose‐6‐phosphate could be converted into 11.6 moles of hydrogen. The in vitro hydrogen production from cellobiose, by combining 13 enzyme components, resulted in an

overall yield of 11.2 moles H2 (93.1% of the theoretical yield) (210). The high cost and slow reaction rates of in vitro systems limit applications in the near future. Securing stability of involved enzymes, cofactors and generated intermediates but also the upscaling form major challenges. The usage of (hyper)thermophilic enzymes is suggested to increase conversion rates and decrease cost due to their stability and activity at elevated temperatures.

Metabolic Engineering, redirecting metabolic fluxes and introducing non native pathways

Improvements of H2 yield through metabolic engineering has been demonstrated for several organisms, but should be separated into two categories based on their focus. One,

37

Chapter 2

engineering of the native pathways and two, introducing non‐native pathways leading to H2 formation. Native pathway engineering focuses mainly on redirecting metabolic fluxes by the blocking of pathways that compete with H2 formation, overexpressing native hydrogenase and inactivating alternative electron acceptors. Increased yields have been demonstrated for E.coli and Clostridium and Enterobacter species, and an overview of several advances is given by Vardar‐Schara et al. (152). A computational framework for designing pathway modification for the overproduction of targeted compounds has been presented by Pharkya et al. (35). Again, these changes can not elevate yields above the networks potential. This can only be done by introducing non‐native pathways. For instance by expressing non native hydrogenases. The overexpression of a non‐native [Fe]‐

hydrogenase from E. cloacae in a non‐hydrogen producing E. coli strain resulted in a H2 yield exceeding the yield of wild strain E. cloacae (45). Introducing the pentose phosphate pathway, supplemented with NADPH dependent hydrogenases or a NADPH:ferredoxin

oxidoreductase, could in theory lead to a yield of 8 moles H2/glucose concomitant with 2 moles of ATP and 1 mole of acetate (171, 174). For thermophiles the reduced end products formed are limited and yields are relatively high (Table 2.2). Although sequenced genomes are available for several hyper‐ and extremely thermophilic hydrogen producers, the majority of them are at present not accessible for genetic engineering. No genetic systems are available except for Thermococcus kodakaraensis (149) and Thermoanaerobacter ethanolicus JW200 (199). Genetic systems for the thermophilic hydrogen producers Clostridium thermocellum (47, 182) and Thermoanaerobacterium saccharolyticum (51) are also available.

Genome scale models

Since annotated genomes are available for some thermophilic H2‐producers, the road to reconstructing biochemical networks appears to be open (82). When combining genome scale models with high throughput data sets derived from proteome, transcriptome and metabolome analysis (90, 106) a deeper understanding of the system as a whole can be achieved (Systems Biology). Flux balance analysis (FBA), although limited in regulatory and kinetic details, provides a useful tool in analyzing steady state fluxes through metabolic networks (81). The solution space of a FBA model, defined by the investigated biochemical

38

Review: Hydrogen production by thermophilic organisms network and constrained by the stoichiometry of the network’s reactions, represent the feasible network fluxes. Additional physiochemical constrains further restrict these possible fluxes, thus defining the organisms network capability. By optimizing the steady‐state

performance, with respect to a defined objective function like maximizing H2 yield, the total system’s behaviour can be investigated. FBA performed on the E.coli iJR904 model with

introduced non‐native NAD(P)H:H2 pathway predicted increased H2 yields, surpassing the

enteric‐type fermentation limit of 2 moles H2/glucose (211). However when an almost

identical NAD(P)H pathway was introduced in E.coli the H2 yields were substantially lower (97). Although Kim et al. presented a metabolic flux analysis of metabolically engineered E.coli strains (120), thus far no genome scale models have been presented for thermophilic

H2‐producers. In general, constructed genome scale models can be used as drawing boards for the design of new non‐native pathway in existing hydrogen producers of for developing newly constructed systems based on easily accessible organisms (Synthetic Biology). Additionally, computer models allow in silico testing prior to experimental work and could be used to reveal possible bottlenecks in the central metabolism limiting hydrogen production. However, comprehension of the central metabolic processes is necessary. To be able to exploit the introduction of non‐native pathways one has to safeguard the organism’s energy supply and recycling mechanisms.

Acknowledgements

This research of M. Verhaart was supported by the European Commission (HYVOLUTION project; Reference nr. 19825). The research of A. Bielen was supported by the IP/OP programme “Systems Biology”, subsidized by Wageningen University.

39

Chapter 2

40

Chapter 3

Hydrogenomics of the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus

Verhaart MRA*, Van de Werken HJG*, VanFossen AL*, Willquist K, Lewis DL, Nichols JD, Goorissen HP, Mongodin EF, Nelson KE, van Niel EWJ, Stams AJM, Ward DE, de Vos WM, van der Oost J, Kelly RM, Kengen SWM (2008) Applied and Environonmental Microbiology. 74:6720‐6729

* These authors contributed equally to this work. Chapter 3

Abstract

Caldicellulosiruptor saccharolyticus is an extremely thermophilic, Gram‐positive anaerobe, which ferments cellulose‐, hemicellulose‐ and pectin‐containing biomass to acetate, CO2 and hydrogen. Its broad substrate range, high hydrogen‐producing capacity, and ability to co‐ utilize glucose and xylose make this bacterium an attractive candidate for microbial bioenergy production. Here, the complete genome sequence of C. saccharolyticus, consisting of a 2,970,275 base pair circular chromosome encoding 2,679 predicted proteins, is described. The genome reveals an extensive polysaccharide hydrolyzing capacity for cellulose, hemicellulose, pectin and starch, coupled to a large number of ABC transporters for monomeric and oligomeric sugar uptake. Components of the Embden‐Meyerhof and the non‐oxidative pentose phosphate pathways are all present, however, no evidence exists for an Entner‐Doudoroff pathway. Catabolic pathways for a range of sugars, including rhamnose, fucose, arabinose, glucuronate, fructose, and galactose, were identified. These pathways lead to the production of NADH and reduced ferredoxin. NADH and reduced ferredoxin are subsequently used by two distinct hydrogenases to generate hydrogen. Whole‐genome transcriptome analysis revealed significant upregulation of the glycolytic pathway and an ABC‐type sugar transporter during growth on glucose and xylose, indicating that C. saccharolyticus co‐ferments these sugars unimpeded by glucose‐based catabolite repression. The capacity to simultaneously process and utilize a range of carbohydrates associated with biomass feedstocks represents a highly desirable feature of this lignocellulose‐utilizing, biofuel‐producing bacterium.

Introduction

Microbial hydrogen production from biomass has been recognized as an important route for renewable energy (1, 2). For biohydrogen production from plant polysaccharides, high temperature microorganisms are well‐suited, as anaerobic fermentation is thermodynamically favored at elevated temperature (60, 189). The extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus DSM 8903, a fermentative anaerobe initially isolated from wood in the flow of a thermal spring in New Zealand, first received attention

42

Hydrogenomics of Caldicellulosiruptor saccharolyticus

for its capacity to utilize cellulose at its optimal growth temperature of 70 °C (157). Further work showed that C. saccharolyticus: (1) can utilize a wide range of plant materials including

cellulose, hemicellulose, starch and pectin, (2) has a very high hydrogen yield (almost 4 H2 per mole of glucose) (45, 86, 207), and (3) can ferment C5 and C6 sugars simultaneously. These features led to the development of bioprocessing schemes based on C. saccharolyticus. For example, H2 production is now being investigated using a two‐step

process in which H2 and acetate from biomass hydrolyzates are generated in one bioreactor, with the acetate fed to a second bioreactor to be used by phototrophic organisms

(Rhodobacter spp.) to produce additional H2 at the expense of light (37). To provide a basis to fully exploit the biohydrogen producing capacity of C. saccharolyticus, its complete genome was sequenced and analyzed in conjunction with transcriptome information for this bacterium grown on glucose and xylose. The comparative and functional genomics approach that is employed here is referred to as “hydrogenomics”. Insights arising from this effort reveal that C. saccharolyticus has the capacity to process and utilize a broad range of sugars, ultimately forming hydrogen from their catabolism.

Materials and methods

Cultivation and DNA isolation

C. saccharolyticus (DSM 8903/ATCC 43494) was cultured overnight on DSMZ 640 medium at 70 ºC with glucose (50 mM) as carbon and energy source. Cells were harvested and genomic DNA was isolated according to the method of (154) using guanidinium thiocyanate. Residual protein was removed in an additional purification step with SDS and proteinase K, followed by chloroform/isoamylalcohol extraction and isopropanol precipitation.

Genome Sequencing and assembly

High molecular weight genomic DNA was provided to the US Department of Energy Joint Genome Institute (http://www.jgi.doe.gov/) for cloning and shotgun sequencing. A combination of small (average insert sizes: 3 and 8 kb) and large (40 kb, fosmid) insert libraries were prepared and used for analysis as indicated at http://www.jgi.doe.gov/.

43

Chapter 3

The complete final assembly was released on 8‐May‐2007 and listed under GenBank accession #CP000679 (http://genome.jgi‐psf.org/finished_microbes/calsa/calsa.home.html).

Genome annotation and comparative analysis

Critica (11) and Glimmer (46) software programs were used for coding region detection and gene identification. TMHMM 2.0 (104) was used to predict transmembrane helices in translated sequences. SignalP v2.0b2 (137) was used to predict the presence and location of N‐terminal signal peptides. All automatic gene and function predictions were manually checked using BLAST programs (8), InterPro (131) and The Integrated Microbial Genomes (IMG) system (120) and corrected if necessary. Protein functions were checked with Carbohydrate‐Active enzymes (CAZy; http://www.cazy.org (41)) classification. The comparative analysis was conducted based on the assignment and classification of Clusters of Orthologous Groups of proteins (COG) (192) by the IMG system.

Growth experiments and RNA isolation

C. saccharolyticus was subcultured (overnight) 3 times on the substrate of interest in modified DSMZ 640 medium before inoculating a pH‐controlled (pH = 7) 1‐liter fermentor containing 4 gram substrate per liter. Cells were grown at 70 °C until mid‐logarithmic phase

(~OD660 = 0.3‐0.4) and harvested by centrifugation and rapid cooling to 4 °C and stored at ‐ 80 °C. Total RNA was isolated using a modified Trizol (Invitrogen) protocol in combination with an RNA easy kit (Qiagen). Quality was tested with the Experion Bioanalyzer (Biorad) and cDNA was constructed with Superscript III reverse transcriptase (Invitrogen).

Whole genome oligonucleotide DNA microarray design and construction

A DNA microarray was designed and constructed based on 2695 protein‐coding sequences in the C. saccharolyticus genome obtained from the Department on Energy’s Joint Genome Institute (http://genome.ornl.gov/microbial/csac). OligoArray 2.0 (167) was used to generate one 60‐mer oligonucleotide probe sequence for each open reading frame. The probes were synthesized (Integrated DNA Technologies, IA), re‐suspended in 50% DMSO, and printed onto Ultragap microarray slides (Corning, NY) using a QArrayMini arrayer

44

Hydrogenomics of Caldicellulosiruptor saccharolyticus

(Genetix, UK). Each probe was spotted five times onto each array to fortify statistical analysis.

Microarray hybridization

The cDNA samples were processed using the Qiaquick purification kit (Qiagen, CA) with the cDNA samples eluted using phosphate buffer. The quantity and quality of the recovered cDNA samples were subsequently analyzed with absorbance at 260/280 nm. Cyanine‐3 and Cyanine‐5 dye (Amersham, UK) labeling and sample hybridizations were done following instructions from TIGR (http://www.tigr.org/tdb/microarray/protocolsTIGR.shtml), with minor adjustments to accommodate long‐oligonucleotide platforms. Samples were hybridized in a 4‐slide loop (Supplemental Figure. 3.1).

Data collection and analysis.

After incubation, slides were washed to remove non‐specifically bound material, and scanned with a ScanArray Lite microarray scanner (Perkin Elmer, MA). Data acquisition and spot quantitation were performed with the ScanArray Express software. Once all the slides were quantitated, data from the loop was analyzed with JMP Genomics 3.0 (SAS, NC), as described previously (156) using a mixed effects ANOVA model (220). The gene expression data have been deposited at the Gene Expression Omnibus database (GEO: http://www.ncbi.nlm.nih.gov/projects/geo/) under accession no. GSE11153.

Results

General features and comparative genomics of the genome of C. saccharolyticus

The genome of Caldicellulosiruptor saccharolyticus DSM 8903/ATCC 43494 consists of one circular chromosome of 2,970,275 base pairs (bp), which has a G+C‐content of 35.3% (Table 3.1). The gene locations of the 2679 predicted coding sequences on the two strands reflect the correlation between the direction of transcription and replication, and show a chromosome with two unequal replichores (Figure 3.1). In addition, the GC‐skew analysis confirms the huge size difference of both replication arms, which might be attributed to a recent major inversion event. Apart from protein‐coding genes, which are classified

45

Chapter 3

according to the COG system (Table 3.2), the chromosome harbors three ribosomal RNA operons and 46 tRNA genes with 41 different anticodons. These anticodons encode for all the 20 canonical amino acids. Like in many prokaryotes, the chromosome of C. saccharolyticus contains Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR). CRISPR are DNA repeats separated by highly variable intervening sequences (spacers) and accompanied by CRISPR‐associated (CAS) genes. The CRISPR and CAS proteins have been proposed to function as a defense mechanism against bacteriophages (14). With nine CRISPR loci and three different CAS genes, C. saccharolyticus is well‐equipped to fend off bacteriophages. Table 3.1. General features of Caldicellulosiruptor saccharolyticus genome

Length chromosome (bp) 2,970,275

G+C content (%) 35.3

Coding density (%) 86.3

Total number of protein‐coding genes 2679

Average length of the protein‐coding genes (b) 958

Total number of pseudogenes 92

Total number of tRNA genes 46

Total number of rRNA genes 9 (3 operons)

CRISPR‐loci 9

46

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Figure 3.1. Circular representation of the Caldicellulosiruptor saccharolyticus chromosome. From the outer circle to the inner circle (1) genomic position in kilobases (kb) (2) coding sequences on the positive and (3) negative strand, which are colored according to the Clusters of Orthologous Groups of proteins (COG) functional categories, (4) tRNA genes (5) GC% (blue) (5) GC‐skew (red). The Microbial Genome Viewer was used to make the circular chromosome wheel (94).

The complete genome sequence confirms the phylogenetic position of C. saccharolyticus as member of the class Clostridia and reveals Thermoanaerobacter tengcongensis (whose genome sequence has also been completed) as closest relative (Supplemental Table 3.1). According to recently updated small subunit rRNA databases C. saccharolyticus, as well as T. tengcongensis, is assigned to the order of the Thermoanaerobacterales. The diversity within this order is, however, very large and of limited biological significance. Reclassification is expected in the near future (114). The genome of C. saccharolyticus was compared to two thermophilic relatives: Clostridium thermocellum and T. tengcongensis (13), as well as to distantly related hyperthermophiles, the bacterium Thermotoga maritima (135) and the archaeon Pyrococcus furiosus (164). These microorganisms have small to moderate genome

47

Chapter 3 sizes and also produce hydrogen while growing on a range of carbohydrates (Supplemental Table 3.2). The genomic distribution of proteins into COG categories is comparable for this group of species (Table 3.2). However, one major difference is that both C. thermocellum and C. saccharolyticus have many more transposases and transposase derivatives than the other species, namely, 99 and 93, respectively. The genomes of T. maritima, P. furiosus and T. tengcongensis, harbor 11, 17 and 43 transposases and transposase derivatives, respectively. Therefore, C. saccharolyticus and C. thermocellum have a high number of proteins, 222 and 252, respectively, in the category ‘Replication, recombination and repair‘ (L).

48

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Table 3.2. Functional categories of predicted open reading frames in the genomes of hydrogen producing organisms

COG functional categoriesa Number of proteins in genomesb Caldicellulo‐ Thermo‐ Clostridium Thermotoga Pyrococcus siruptor anaerobacter thermocellum maritima furiosus saccharolyticus tengcongensis Information storage and processing B Chromatin structure and dynamics 2 1 2 1 4 Replication, recombination and L 222 252 149 89 109 repair A RNA processing and modification 0 0 0 0 2 K Transcription 134 174 141 82 80 Translation, ribosomal structure J 147 165 150 135 166 and biogenesis Cellular processes and signaling Cell cycle control, cell division, D 35 38 40 21 18 chromosome partitioning N Cell motility 71 95 67 58 13 Cell wall/membrane/envelope M 107 172 110 76 61 biogenesis Z Cytoskeleton 3 1 0 0 0 V Defense mechanisms 48 40 42 27 29 Intracellular trafficking, secretion, U 42 60 46 39 20 and vesicular transport Posttranslational modification, O 59 89 81 55 55 protein turnover, chaperones T Signal transduction mechanisms 125 170 122 73 19 Metabolism Amino acid transport and E 166 166 206 181 158 metabolism Carbohydrate transport and G 213 144 160 166 93 metabolism Coenzyme transport and H 101 102 67 62 89 metabolism C Energy production and conversion 111 115 130 119 128 Inorganic ion transport and P 73 91 96 113 95 metabolism I Lipid transport and metabolism 34 46 55 31 25 Nucleotide transport and F 56 62 62 54 52 metabolism Secondary metabolites Q biosynthesis, transport and 14 18 27 15 11 catabolism Poorly characterized S Function unknown 177 187 173 137 189 R General function prediction only 228 261 249 204 275 Not in COG 655 985 615 305 322 aGene classification according to Integrated Microbial Genomes (IMG) system (120) using the functional classification of Clusters of Orthologous Groups of proteins (COG) (192) bNumber of protein‐coding genes in each category without pseudogenes.

49

Chapter 3

Furthermore, the C. saccharolyticus genome harbors the largest number of carbohydrate transport and metabolism genes in this group. In fact, the C. saccharolyticus genome contains at least 177 ABC‐transporter genes, outnumbering the 165 identified in T. maritima (40, 135). The C. saccharolyticus genome contains a reverse gyrase gene (Csac_1580), the product of which induces positive supercoiling of DNA (95). Reverse gyrase is regarded as a molecular marker of hyperthermophilicity, and therefore distinguishes C. saccharolyticus from C. thermocellum, which lacks reverse gyrase. Despite the fact that C. saccharolyticus was described as a non‐motile organism, a set of flagella structure, biogenesis and chemotaxis genes were detected; it is not clear whether these genes are functional, since one is interrupted by a stop codon (pseudogene Csac_1277). C. saccharolyticus has a nitrogen‐fixation cluster (Csac_2461‐2466) and many sporulation genes; neither of these phenotypic properties have been described for this bacterium.

Central carbon metabolism

C. saccharolyticus is able to metabolize a wide variety of carbohydrates, including the monosaccharides D‐glucose, D‐xylose, D‐fructose, D‐galactose, D‐/L‐arabinose, D‐mannose, L‐

rhamnose and L‐fucose, but also α‐ and β‐linked di‐ and polysaccharides, including maltose, starch, pullulan, sucrose, trehalose, amorphous and micro‐crystalline cellulose, xylan, locust bean gum and pectin (157). Once hydrolyzed, sugars are channeled to the central catabolic pathways (Figure 3.2). The genome sequence reveals components of a complete Embden‐Meyerhof (EM) pathway, including a ROK family glucokinase (Csac_0778), 6‐ phosphofructokinase, (Csac_2366/1830), a bifunctional phosphoglucose/phosphomannose isomerase (Csac_1187), fructose‐1,6‐bisphosphate aldolase (Csac_1189), pyruvate kinase (Csac_1831) as well as pyruvate‐phosphate dikinase (PPDK) (Csac_1955). Also a gapA operon is evident, consisting of glyceraldehyde‐3‐phosphate (GAP) dehydrogenase (Csac_1953), the fusion protein phosphoglycerate kinase/triose‐phosphate isomerase (Csac_1952), phosphoglycerate mutase (Csac_1951), and enolase (Csac_1950) (Figure 3.2). However, the oxidative branch (ox) of the Pentose Phosphate Pathway (PPP) and the Entner‐Doudoroff (ED) pathway were not detected, which is consistent with previous reports using 13C‐NMR (45). The absence of the ox‐PPP, however, raises questions about how NADPH is produced for biosynthesis. The only other obvious NADPH‐producing

50

Hydrogenomics of Caldicellulosiruptor saccharolyticus

reaction is isocitrate dehydrogenase (Csac_0751). However, based on its sequence homology the isocitrate dehydrogenase is likely to produce NADH instead of NADPH. Also, no obvious homolog to an NADPH‐producing glyceraldehyde‐3‐phosphate dehydrogenase can be identified, as has been reported for Streptococcus species and some clostridia (26). NADPH can also be synthesized from NADH by a transhydrogenase, which is either membrane‐bound or soluble. In the genome no orthologs exist of subunits of the membrane‐bound type, but several homologous genes coding for soluble FAD‐dependent pyridine nucleotide‐disulfide oxidoreductases are present (Csac_0759, Csac_2199, Csac_0402). There exact physiological role remains to be investigated. Furthermore, no ferredoxin:NADPH reductase homolog is present, although such activity has been measured in some Thermoanaerobacter spp. (77). Xylose, a major constituent of hemicellulose, is funneled by a putative xylose isomerase (Csac_1154) and xylulokinase (Csac_0798) into the non‐oxidative branch (nox) of the PPP. The nox‐PPP uses ribulose‐phosphate 3‐epimerase (Csac_2074), ribose‐5‐phosphate isomerase (Csac_1200), the N‐terminal (Csac_1351) and C‐terminal transketolase (Csac_1352) and transaldolase (Csac_2036) to produce the EM intermediates fructose‐6‐ phosphate and glyceraldehyde‐3‐phosphate. Galactose also enters the EM via the Leloir‐ pathway, which includes galactokinase (Csac_1511), galactose‐1‐phosphate uridylyltransferase (Csac_1510), UDP‐glucose 4‐epimerase (Csac_1512) and phosphoglucomutase (Csac_2295). Strikingly, none of the established types of fructose‐ bisphosphatase (Class I to IV; (175)) are evident in the C. saccharolyticus genome. Since fructose‐bisphosphatase is an essential enzyme of the gluconeogenesis, C. saccharolyticus presumably uses a novel phosphatase. Moreover, a gene for the gluconeogenic PEP synthetase is also missing, although the conversion of pyruvate to PEP could occur via the reversible PPDK (Csac_1955) or via oxaloacetate. Pyruvate, the end product of the EM‐ pathway, is subsequently decarboxylated to acetyl‐CoA by pyruvate:ferredoxin oxidoreductase (POR). C. saccharolyticus contains three 2‐oxoacid:ferredoxin oxidoreductase enzyme complexes (Csac_2248‐2249, 1458‐1461 and Csac_1548‐1551). According to transcriptional response information (vide infra), the true POR is probably encoded by Csac_1458‐1461. Acetyl‐CoA is used to generate acetate and ATP

51

Chapter 3

Csac_2040/2041), or it enters the tricarboxylic acid (TCA) cycle for biosynthetic purposes. The TCA cycle in C. saccharolyticus is incomplete, with an oxidative branch to succinyl‐CoA catalyzed by a citrate (Re)‐synthase (Csac_0746), aconitate hydratase (Csac_0750), isocitrate dehydrogenase (Csac_0751) and the 2‐oxoglutarate:ferredoxin oxidoreductase complex (1548‐1551). In the reductive direction, only orthologs of the subunits of fumarate hydratase were detected with a high level of confidence (Csac_2759/Csac_0738). Malate dehydrogenase (oxaloacetate‐decarboxylating) (Csac_2059) may be used to generate malate directly from pyruvate instead from oxaloacetate. Fumarate reductase, however, could not be identified, which is in agreement with the lack of this enzyme in related clostridia. Besides the malate dehydrogenase, TCA metabolites could be replenished by a putative sodium pump oxaloacetate decarboxylase enzyme complex (Csac_2482‐2485).

52

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Figure 3.2. An overview of the carbon metabolism and transport systems in Caldicellulosiruptor saccharolyticus. The identity of the various ABC‐type sugar transporters is not known. Secondary transport systems may be involved as well.

Polysaccharide degrading enzymes

The capacity of C. saccharolyticus to hydrolyze a broad range of polysaccharides prior to fermentation differentiates this bacterium from many thermophilic anaerobes. Indeed, the genome of C. saccharolyticus encodes a wide range of carbohydrate active enzymes (Supplemental Table 3.3). These carbohydrate‐utilizing enzymes are often clustered on the chromosome and can be assigned to substrate specific catabolic pathways for cellulose, hemicellulose and, to a lesser extent, starch and pectin. The α‐1,4‐glucan polymers, for instance, can be transported into the cell using the maltodextrin ABC‐transport system proteins (Csac_0427‐0428/0431). An intracellular α‐amylase (Csac_0426) and a 1,4‐α‐glucan phosphorylase (Csac_0429) further degrade the intracellular maltodextrins, releasing

53

Chapter 3

glucose‐1‐phosphate. Remarkably, a transcriptional regulator of the LacI family (Csac_0430) is also in this maltodextrin cluster and is, therefore, a good candidate for controlling expression of this maltodextrin‐degrading pathway at the transcriptional level. In addition, a GCAAACGTTTGC consensus sequence was found in upstream sequences of this transport cluster and several starch‐degrading enzymes, such as an α‐amylase precursor (Csac_0408), an oligo‐1,6‐glucosidase (Csac_2428), a pullulanase (Csac_0689), a 4‐α‐glucanotransferase (Csac_0203), and a putative glucan 1,4‐α‐glucosidase (Csac_0130). The consensus sequence resembles the binding site (CGCAAACGTTTGCGT) of the maltose/maltodextrin transcriptional repressor MalR from the Gram‐positive Streptococcus pneumoniae (138). Besides this putative starch‐degrading regulon, C. saccharolyticus has a glycogen metabolic cluster (Csac_0780‐0784), a maltose ABC‐transport system (Csac_2491‐2493), and a second pullulanase (Csac_0671). Taken together, C. saccharolyticus is well‐equipped for starch utilization.

An important feature of C. saccharolyticus is its ability to produce H2 not only from α‐linked polymers, but also from complex β‐linked glycans, such as cellulose, hemicellulose, laminarin and galactomannan. Growth on cellulosic substrates is rare among (hyper)thermophilic microorganisms. C. saccharolyticus does not metabolize cellulose by means of a cellulosome (193). For example, typical molecular components of a cellulosome, i.e., dockerin domains and scaffolding proteins, were not identified in the genome. Nevertheless, a gene cluster (Csac_1076‐1081) containing cellulase precursors is present. These cellulases are potentially capable of degrading this plant polysaccharide (19) (Supplemental Table 3.3). Moreover, another gene cluster (Csac_1089‐1091) and an extracellular cellulase (Csac_0678) may assist in completely hydrolyzing cellulose to glucose. A remarkable aspect of the β‐glycanases in Caldicellulosiruptor species is their bifunctional domain architecture, consisting of central cellulose binding domains which are bordered by distinct catalytic domains (19). The C. saccharolyticus genome sequence confirmed the presence of these multidomain hydrolases viz. Csac_1076‐1079 and Csac_2411, which comprise cellulase, mannanase, xylanase and arabinofuranosidase domains (19, 115, 129, 193). No additional bifunctional β‐glycanases were found in the genome. The bifunctional Csac_2411 is part of another large gene cluster (Csac_2404‐2411) coding for enzymes

54

Hydrogenomics of Caldicellulosiruptor saccharolyticus

involved in the hydrolysis of the plant polysaccharide xylan (hemicellulose). These mostly extracellular enzymes might be co‐expressed with a smaller putative xylan‐utilizing cluster (Csac_0203‐0205). This latter cluster was not significantly up‐regulated on xylose, in contrast to genes in the former cluster. Furthermore, putative genes that encode enzymes to degrade galactomannan (Csac_0663‐0664), galactoarabinan (Csac_1560‐1562) and laminarin (Csac_2548) can be identified.

The plant cell wall component pectin consists of α‐1,4‐linked D‐galacturonic acid backbone,

sometimes interspersed by L‐rhamnose, and side chains made of monosaccharides, such as

D‐galactose, D‐xylose and L‐arabinose (163). Degradation of the main pectin component, D‐ galacturonate, requires a galacturonate isomerase, a tagaturonate reductase, and an altronate dehydratase to form 2‐keto‐3‐deoxygluconate (KDG). Galacturonate isomerization may occur by glucuronate isomerase (Csac_1949). However, tagaturonate reductase and altronate dehydratase were not detected in the genome of C. saccharolyticus. Apparently, novel enzymes or a novel pathway are responsible for the degradation of galacturonate. In contrast, a gene cluster for the conversion of glucuronic acid to KDG (Csac_2686‐2689) can be identified, and includes fructuronate reductase, mannonate dehydratase, a putative β‐ galactosidase/β‐glucuronidase, and an α‐glucuronidase. Glucuronic acid is a common substituent of xylan. Enzymes for the subsequent conversion of KDG to pyruvate and GAP, viz. KDG kinase (Csac_0355 or Csac_2720) and KDG‐6‐phosphate aldolase (Csac_0354) are present as well. The encoding genes of these last two steps are clustered with genes (Csac_0356‐0357 and Csac_2718‐2719) that both metabolize 5‐keto‐4‐deoxyuronate (DK‐I), an unsaturated cleavage product from pectate, to KDG. The enzymes that are able to hydrolyze the pectate backbone and the side chains (e.g., unsaturated rhamnogalacturonyl hydrolase (Csac_0360), galacturan 1,4‐ α ‐galacturonidase (Csac_0361), β‐galactosidase (Csac_0362) and a glycoside hydrolase with unknown substrate specificity (Csac_0363)) are in proximity to these KDG metabolic enzymes as well. However, neither a pectate lyase nor a methylesterase could be definitively identified in the genome; although Csac_2721/2728 might be candidates for a pectate lyase based on distant homology to known lyases.

C. saccharolyticus is also able to grow on L‐rhamnose and on L‐fucose, thereby producing 1,2‐propanediol as end product (unpublished data). A putative rhamnose catabolic pathway

55

Chapter 3

can be assigned that generates dihydroxy‐acetone phosphate and 1,2‐propanediol catalyzed

by a L‐rhamnose isomerase (Csac_0876), a putative L‐rhamnulokinase (Csac_0989), a L‐ rhamnulose‐1‐phosphate aldolase (Csac_0865) and a putative lactaldehyde reductase (Csac_0407). Fucose can be processed by a similar pathway, using the aforementioned lactaldehyde reductase, and yet to be identified versions of L‐fuculokinase, a bifunctional L‐

fucose isomerase/D‐arabinose isomerase (Csac_1339) and fuculose‐1‐phosphate aldolase (Csac_0425).

Fermentation products

Reducing equivalents are produced at the level of NAD and ferredoxin (Csac_0737). Since C. saccharolyticus can produce almost 4 H2 per mole of glucose (45), both NADH and reduced ferredoxin should ultimately be able to transfer their reducing equivalents to protons to form hydrogen. In the genome, two hydrogenase gene clusters could be identified, which are very similar to the two related clusters in Thermoanaerobacter tengcongensis (188). The first cluster (Csac_1534‐1539) encodes subunits of a Ni‐Fe hydrogenase (EchA‐F) and various genes required for maturation of the hydrogenase complex (HypA‐F; Csac_1540‐1545). For T. tengcongensis, this Ni‐Fe hydrogenase is ferredoxin‐dependent, membrane‐bound, and may act as a proton pump to generate a proton motive force. The second cluster (Csac_1860‐1864) codes for a Fe‐only hydrogenase (HydA‐D), which is NAD‐dependent and located in the cytoplasm, similar to the case for T. tengcongensis (188). Hydrogenases that

form H2 directly from NADH are unusual, and make an NAD:Fd oxidoreductase (Nfo) redundant. Nfo’s (also known as Rnf) are membrane‐bound multi‐subunit complexes that use or create a Na+‐gradient coupled to the transfer of reducing equivalents between NADH and ferredoxin (23). An Nfo‐cluster has been identified in the genomes of C. thermocellum, T. maritima and T. ethanolicus, but not in T. tengcongensis and C. saccharolyticus. The

absence of an Nfo in C. saccharolyticus also implies that under elevated levels of H2, reduced ferredoxin may either not be used to produce NADH or that a novel type of enzyme (complex) performs this reaction. Altogether, information available suggests that C. saccharolyticus is able to produce hydrogen from ferredoxin, but can also do this directly from NADH. Production of hydrogen would seem to be preferable, because under these

56

Hydrogenomics of Caldicellulosiruptor saccharolyticus

conditions all pyruvate is converted to acetate (and CO2), which is coupled to the synthesis of ATP.

When the hydrogen partial pressure (pH2) becomes too high, hydrogen formation from NADH is no longer thermodynamically favorable. In that case, NADH is oxidized through the formation of lactate or ethanol. A gene for a lactate dehydrogenase can be identified (Csac_1027), but genes for acetaldehyde dehydrogenase and alcohol dehydrogenase were not obvious. In T. tengcongensis and T. ethanolicus, ethanol formation is NADPH‐dependent and catalyzed by a bifunctional ADH acetyl‐CoA thioesterase; this enzyme also has a homolog in C. saccharolyticus (Csac_0395). A third small hydrogenase‐like cluster could be detected in the C. saccharolyticus genome, composed of four genes encoding two NADH‐binding proteins (Csac_0619‐0620), a molybdopterin oxidoreductase containing NAD and 4Fe‐4S binding regions (Csac_0621), and an iron‐containing alcohol dehydrogenase (Csac_0622). The function of this cluster is yet unknown.

Transport systems

As mentioned earlier, there are a number of genes involved in ABC transporters found in the C. saccharolyticus genome including the previously noted carbohydrate specific maltodextrin ABC‐transport system (Csac_0427‐0428/0431) and the maltose ABC‐transport system (Csac_2491‐3). As is the case with certain T. maritima maltose transporters, both sets of these transport proteins lack ATP‐binding subunits. For many bacteria, the intracellular ATPase used in the system is not encoded within the same operon. Both of these set of ABC transporters are located downstream from a two‐component system (TCSs) of a sensor histidine protein kinase and a response regulator. In C. saccharolyticus ~50% of the ABC carbohydrate transport systems are located near TCSs on the chromosome. Comparative analysis of C. saccharolyticus sugar binding proteins (SBPs) revealed that about 2/3 belong to COG1653. This category includes the CUT1 subfamily members TM0432, TM0595 and TM1855 that transport a variety of di‐ and oligosaccharides, such as maltose. More than half of COG1653 members are proximate in genomes to glycoside hydrolases, supporting their designation as ABC transporters involved in carbohydrate utilization. Putative SBPs Csac_0242, Csac_0391, Csac_2326 and Csac_2507 belong to COG1879.

57

Chapter 3

Csac_2506 and Csac_2510 are associated with the xylose transport specific COG4213. As is the case with T. maritima, a few putative SBPs (Csac_0261 and Csac_4166) annotate as peptide transporters, although their actual function is unknown. Components of phosphotransferase systems (PTSs) have been identified in C. saccharolyticus (although only one set of carbohydrate‐specific EII), along with a few putative members of the major facilitator superfamily (Csac_0685, Csac_0786, Csac_1100, Csac1170 and Csac_2298). However, it is likely that carbohydrate utilization proceeds mainly through ABC transporters.

Transcriptional regulation

The ability of C. saccharolyticus to utilize many different carbohydrates suggests a tight regulation among the pathways. Many carbohydrate utilization pathways in the genome appear to be regulated at the transcriptional level. Apart from the RNA polymerase core enzyme subunits (Csac_2259/2085/0951/0952), this Gram‐positive species has 12 different σ‐factors to construct the RNA polymerase holoenzyme. In addition, many of the sugar transcriptional regulators are present in multiple copies in the genome, e.g. nine proteins from the LacI family, six proteins from the DeoR family and eight from the GntR family, as well as 19 receiver proteins from a two‐component system with a helix‐turn‐helix AraC domain. The latter are always clustered with sugar transporters and sugar hydrolytic enzymes. Carbon catabolite repression (CCR) by glucose was not observed in C. saccharolyticus (Figure 3.3). Nevertheless, some indicators of a Carbon Control Protein A (CcpA) dependent CCR in Gram‐positives are present in the genome: (i) a histidine‐containing phosphocarrier (HPr) (Csac_2438) that is in proximity to the only phosphoenolpyruvate‐dependent phosphotransferase system (PTS), which is fructose specific; (ii) A HPr(Ser) kinase (Csac_1186); (iii) a catabolite repression HPr (CrH) (Csac_1163) and nine members of the CcpA‐containing LacI family. Binding sites for a putative CcpA, the catabolite‐responsive element (cre), could not be identified. The Bacillus subtilis consensus sequence WWTGNAARCGNWWWCAWW (126) is, for instance, detected only twice: (1) In the upstream region of the aforementioned α‐amylase precursor (Csac_0408) where it overlaps with the putative MalR binding site. (2) In the middle of the gene encoding the fumarate

58

Hydrogenomics of Caldicellulosiruptor saccharolyticus

hydratase subunit α (Csac_2759). Nevertheless, in C. saccharolyticus, CCR is probably present, although the metabolite that induces this repression is unknown.

Figure 3.3. Figure 3. Growth of Caldicellulosiruptor saccharolyticus on a xylose:glucose mixture (1:1 (w/w)). ○ = optical density at 660 nm (OD660); ▲= hydrogen; ∆ = acetate; ■ = glucose; ♦ = xylose; □ = lactate.

Besides global regulation through CCR, many local transcriptional regulators control the expression of carbohydrate metabolic pathways. Several orthologous transcriptional regulators were identified in C. saccharolyticus. The central glycolytic genes regulator (CggR) (Csac_1954), for instance, represses the transcription of the gapA operon (113), while the FruR (Csac_2442) controls the fructose operon (15). Based on the fact that many transcriptional regulators are in proximity to their target operons, putative functions could be assigned to: an α‐linked glucan transcriptional regulator (Csac_0430), a regulator of the

oxidative‐branch of the TCA cycle to oxoglutarate (Csac_0752), a repressor of the L‐ arabinose metabolism (Csac_0722), and a putative response regulator receiver protein of the glucuronate degradation (Csac_2690).

59

Chapter 3

Transcriptome analysis Caldicellulosiruptor saccharolyticus

One of the beneficial features of C. saccharolyticus for hydrogen production is its ability to degrade cellulosic substrates as well as hemicellulose. Moreover, mixtures of glucose and xylose can be fermented simultaneously (Figure 3.3) suggesting that classical CCR by glucose does not occur. To elucidate the central carbon metabolic pathways and their regulation, transcriptome analysis was performed after growth on glucose, xylose and a 1:1 mixture of both substrates. l‐Rhamnose, which was likely to follow another pathway, was used as a reference substrate. The transcriptional data clearly show that glucose, xylose and the glucose:xylose mixture all trigger up‐regulation of genes in the EM pathway, when compared to rhamnose (Figure 3.4; Supplemental Table 3.4). In particular, the fructose‐ bisphosphate aldolase, GAP dehydrogenase, PPDK and POR are significantly stimulated. The ultimate acetate‐forming acetate kinase is also highly up‐regulated. A catabolic role for PPDK is intriguing, since it normally is associated with gluconeogenesis (as in propionic acid bacteria and plants), and PEP is usually converted by pyruvate kinase. However, homologs of PPDK are also present in related clostridia and Thermoanearobacter species

60

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Figure 3.4. Intensity ratio of transcript levels of selected genes that responded to growth on glucose (white), xylose (grey) or a mixture of glucose and xylose (black), compared to rhamnose. Ratios are expressed as log2 value. GAPDH, glyceraldehyde‐3‐phosphate dehydrogenase; POR, pyruvate:ferredoxin oxidoreductase; NdH, NAD‐dependent hydrogenase; FdH, ferredoxin‐dependent hydrogenase.

It is worth noting that growth on glucose, xylose or both sugars all trigger transcription of the gene encoding a xylose‐specific ABC transport system (Csac_2504‐2506) (Figure 3.4), suggesting that glucose and xylose are transported by the same uptake system. Moreover,

61

Chapter 3 none of the identified putative CCR genes (vide infra) were differentially transcribed, confirming the fact that catabolite repression by glucose was not a factor. Transcriptional response to growth on monosaccharides enabled the identification of genes and groups of adjacent genes (gene clusters), that were specifically up‐regulated in response to either glucose or xylose. On glucose, several genes coding for ‐glucan hydrolases responded. The most striking observation, however, was the up‐regulation of an entire gene cluster (Csac_1991‐2000) involved in purine synthesis, that was not observed for xylose. Up‐ regulation of purine biosynthesis genes was also detected in the transcriptome of Escherichia coli growing on glucose compared to xylose (57). On xylose, several gene clusters required for xylan or xylose conversion were up‐regulated (Csac0692‐0696; Csac0240‐0242; Csac2416‐2419). These clusters encode ABC transport systems, transcriptional regulators and endo‐xylanases. In addition, genes specifically required for growth on rhamnose, were highly up‐regulated during growth on rhamnose, thus indicating the utilization pathway for this sugar.

62

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Discussion

C. saccharolyticus has been shown to be an excellent candidate for biohydrogen production (45, 86, 207). In contrast to mesophilic fermentative anaerobes, it produces almost no reduced end products, such as lactate or ethanol, and the amount of hydrogen approaches

the “Thauer limit” of 4 H2/glucose (60, 194). Moreover, C. saccharolyticus hydrolyzes various biomass‐derived polymers, such as cellulose, hemicellulose, starch, and pectin, and ferments corresponding sugar monomers, including glucose and xylose. The complete genome sequence of C. saccharolyticus provides new insights into the exceptional capacity of the bacterium to degrade a variety of plant polysaccharides and further reveals its high plasticity with many transposases, CRISPRs and two uneven replication arms. A large number of sugar hydrolases and transferases could be identified, outnumbering those in the hyperthermophile Thermotoga maritima (34). Metabolic pathways for the degradation of residual components of cellulose, hemicellulose, starch and pectin could be assigned. Reducing equivalents are produced as NADH or reduced ferredoxin, which are apparently used directly to produce hydrogen by a soluble NADH‐dependent Fe‐only hydrogenase and a membrane‐bound ferredoxin‐dependent Ni‐Fe hydrogenase. The ability to produce hydrogen directly from NADH is not known from mesophilic anaerobes and may be responsible for the relatively high hydrogen production rates by C. saccharolyticus. In mesophiles, reducing equivalents from NADH first have to be transferred to ferredoxin, which requires input of energy, either by a sodium gradient (23) or by coupling to an exergonic reaction (109). In extreme thermophiles, such as C. saccharolyticus, this is apparently not necessary. The absence of catabolite repression by glucose is an important characteristic for biohydrogen producers since it allows them to process an array of biomass‐derived substrates simultaneously. Whereas glucose is generally known to repress the use of xylose by CCR (76), this was not observed in C. saccharolyticus. Moreover, the transcriptome showed that the various components of a CCR system, present in the genome (CcpA homologs, HPr, HPr kinase), were not differentially transcribed under the conditions examined, suggesting that this type of regulation is not triggered by glucose or xylose. No

63

Chapter 3

obvious differences were noted in the transcriptome for central metabolic pathways during growth on either glucose or xylose or a mixture of both. The EM pathway was not affected by the hexose or pentose substrate, which is in contrast to the transcriptome analysis of E. coli for growth on the same substrates (57). Remarkably, also the same specific ABC transporter is upregulated on both substrates, which is also in line with the non‐preferential behavior of C. saccharolyticus towards these two monomeric sugars. Detailed knowledge on the metabolic pathways leading to hydrogen production enables one to identify key enzymes that may be targets for improving the H2 yield by metabolic engineering. Currently, a genetic system for C. saccharolyticus is under development which will initially target the dehydrogenases involved in lactate and ethanol formation.

Alternatively, genes could be introduced to constitute an ox‐PPP, to achieve higher H2 yields greater than 4 per mole of glucose (232). In any case, the C. saccharolyticus genome provides new insights into the metabolic features of a versatile biohydrogen producer, which can inspire efforts to optimize microbial bioenergy systems.

Acknowledgements The authors gratefully acknowledge the US Department of Energy and the Joint Genome Institute for providing the genome sequence for Caldicellulosiruptor saccharolyticus. This research is supported by the Earth and Life Sciences Foundation (ALW) and Chemische Wetenschappen (CW) project 050.50.206 (subsidized by the Netherlands Organization for

Scientific Research (NWO)) and by the European Commission (HYVOLUTION project;

Reference nr. 19825). Microarray fabrication was financially supported in part by Danisco (Genencor International, Inc., Palo Alto, USA). RMK acknowledges support from the US NSF Biotechnology Program. ALV and DLL acknowledge support from a US Department of Education GAANN Fellowship.

64

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Supplementary information

Supplementary Table 3.1 Taxonomic distribution of species with more than 100 best blast hits, based on Blastp against the KEGG database. Species Thermoanaerobacter tengcongensis 549 Clostridium thermocellum 512 Carboxydothermus hydrogenoformans 145 Clostridium acetobutylicum 140

Supplementary Table 3.2 Genome properties of organisms used for comparative genomics based on IMG data(41). Size GC‐ content Organism Chromosome Proteinsa (kbp) (%) Caldicellulosiruptor saccharolyticus DSM 8903 1 2,970 2,679 32.3 Clostridium thermocellum ATCC 27405 1 3,843 3,191 39.0 Thermoanaerobacter tengcongensis MB4 1 2,689 2,588 37.6 Thermotoga maritima MSB8 1 1,861 1,858 46.3 Pyrococcus furiosus DSM 3638 1 1,908 1,983 40.8 aNumber of predicted proteins without pseudogenes.

Supplementary Table 3.3 Carbohydrate‐active enzymes encoded by the genome of Caldicellulosiruptor saccharolyticus Locus id Protein name E.C.a Familyb CBMb Csac_1089 β‐glucosidase 3.2.1.21 GH1 Csac_0129 β‐mannosidase 3.2.1.25 GH2 Csac_0362 β‐galactosidase 3.2.1.23 GH2 Csac_2686 putative β‐galactosidase/β‐glucuronidase GH2 Csac_2734 β‐galactosidase 3.2.1.23 GH2 3.2.1.21/3. Csac_0586 β‐glucosidase/xylan 1,4‐β‐xylosidase GH3 2.1.37 Csac_1102 β‐glucosidase 3.2.1.21 GH3 Csac_1562 α‐galactosidase (melibiase) 3.2.1.22 GH4 Csac_2748 putative α‐galactosidase GH4 Csac_0137 cellulase 3.2.1.4 GH5

65

Chapter 3

Csac_0678 cellulase precursor 3.2.1.4 GH5 Csac_1080* mannan endo‐1,4‐β‐mannosidase 3.2.1.78 GH5 Csac_2528 glycoside hydrolase, family 5 GH5 3.2.1.4/3.2. Csac_1078 cellulase/cellulose 1,4‐β‐cellobiosidase precursor GH5/10 CBM3 1.91 3.2.1.78/3. Csac_1077 mannan endo‐1,4‐β‐mannosidase/cellulase precursor GH5/44 CBM3/3 2.1.4 Csac_1079 cellulase/mannan endo‐1,4‐β‐mannosidase precursor 3.2.1.4 GH9 CBM3/3 Csac_1076 cellulase precursor 3.2.1.4 GH9/48 CBM3/3 Csac_0204 putative endo‐1,4‐β‐xylanase GH10 Csac_2405 endo‐1,4‐β‐xylanase precursor 3.2.1.8 GH10 Csac_2408 endo‐1,4‐β‐xylanase precursor 3.2.1.8 GH10 Csac_0696 endo‐1,4‐β‐xylanase precursor 3.2.1.8 GH10 CBM22/22 Csac_2410 endo‐1,4‐β‐xylanase precursor 3.2.1.8 GH10 CBM22/22 Csac_0203 4‐α‐glucanotransferase 2.4.1.25 GH13 Csac_0408 α‐amylase precursor 3.2.1.1 GH13 Csac_0426 α‐amylase 3.2.1.1 GH13 Csac_2428 oligo‐1,6‐glucosidase 3.2.1.10 GH13 Csac_0689 pullulanase 3.2.1.41 GH13 CBM20/41 Csac_0671 pullulanase 3.2.1.41 GH13 CBM48 Csac_0784 1,4‐α‐glucan branching enzyme 2.4.1.18 GH13 CBM48 Csac_0130 putative glucan 1,4‐α‐glucosidase GH15 CBM22/22/22 Csac_2548* putative endo‐1,3(4)‐β‐glucanase GH16 /22/22 Csac_2539 putative β‐N‐acetylhexosaminidase GH20 Csac_1720 protein containing lytic transglycosylase SLT domain GH23 Csac_1986 protein containing lytic transglycosylase SLT domain GH23 Csac_0663 mannan endo‐1,4‐β‐mannosidase 3.2.1.78 GH26 Csac_0361 galacturan 1,4‐α‐galacturonidase 3.2.1.67 GH28 Csac_0664 putative galacturan 1,4‐α‐galacturonidase GH28

Csac_1340 α‐L‐fucosidase 3.2.1.51 GH29 Csac_2513 glycoside hydrolase, family 30 GH30 Csac_1354 α‐xylosidase 3.2.1.‐ GH31 Csac_1118 α‐galactosidase 3.2.1.22 GH36 Csac_2404 xylan 1,4‐β‐xylosidase 3.2.1.37 GH39 Csac_2409 xylan 1,4‐β‐xylosidase 3.2.1.37 GH39 Csac_1018 β‐galactosidase 3.2.1.23 GH42 Csac_0359 glycoside hydrolase, family 43 GH43

Csac_1560 arabinan endo‐1,5‐α‐L‐arabinosidase 3.2.1.99 GH43 Csac_0437 glycoside hydrolase, family 43 GH43

66

Hydrogenomics of Caldicellulosiruptor saccharolyticus

3.2.1.37/3. Csac_2411 xylan 1,4‐β‐xylosidase/α‐N‐arabinofuranosidase precursor GH43/43 CBM22/6 2.1.55 Csac_1561 α‐N‐arabinofuranosidase 3.2.1.55 GH51 Csac_0439 kojibiose phosphorylase 2.4.1.230 GH65 putative trehalose/maltose hydrolase (possible Csac_0444 GH65 phosphorylases) Csac_2689 α‐glucuronidase 3.2.1.139 GH67 glycoside hydrolase, family 74 with carbohydrate‐binding Csac_1085 GH74 CBM3 module protein precursor

Csac_1105/ α‐L‐rhamnosidase N‐terminal domain/C‐terminal domain 3.2.1.40 GH78 Csac_1107 protein Csac_2730 unsaturated glucuronyl hydrolase GH88 Csac_1090 putative cellobiose phosphorylase GH94 Csac_1091 cellobiose phosphorylase 2.4.1.20 GH94 Csac_0360 unsaturated rhamnogalacturonyl hydrolase 3.2.1.‐ GH105 Csac_0206 putative acetylesterase IPR005181 Csac_0258 glycosidase, PH1107‐related IPR007184 Csac_0259 glycosidase, PH1107‐related IPR007184 Csac_0296 glycosidase, PH1107‐related IPR007184 Csac_0762 glycosidase, PH1107‐related IPR007184 Csac_0853 glycosidase, PH1107‐related IPR007184 Csac_2527 glycosidase, PH1107‐related IPR007184 putative glycoside hydrolase precursor with carbohydrate‐ Csac_2519 IPR013781 CBM28 binding module Csac_0268 glycosyltransferase, family 2 GT2 Csac_1057 glycosyltransferase, family 2 GT2 Csac_1681 glycosyltransferase, family 2 GT2 Csac_1877 glycosyltransferase, family 2 GT2 Csac_2350 glycosyltransferase, family 2 GT2 Csac_2426 glycosyltransferase, family 2 GT2 Csac_2567 glycosyltransferase, family 2 GT2 Csac_2631 glycosyltransferase, family 2 GT2 glycosyltransferase WecB/TagA/CpsF family and Csac_1349 GT26 Polysaccharide pyruvyl transferase domain protein undecaprenyldiphospho‐muramoylpentapeptide β‐N‐ Csac_0925 2.4.1.227 GT28 acetylglucosaminyltransferase Csac_2337 1,2‐diacylglycerol 3‐glucosyltransferase homolog GT28 Csac_0429 α‐1,4‐glucan phosphorylase 2.4.1.1 GT35 Csac_0780 glycogen phosphorylase 2.4.1.1 GT35 Csac_1081 glycosyl transferase, family 39 GT39

67

Chapter 3

Csac_0134 glycosyltransferase, family 4 GT4 Csac_0194 glycosyltransferase, family 4 GT4 Csac_1092 glycosyltransferase, family 4 GT4 Csac_1346 glycosyltransferase, family 4 GT4 Csac_1682 glycosyltransferase, family 4 GT4 Csac_1729 glycosyltransferase, family 4 GT4 Csac_1745 glycosyltransferase, family 4 GT4 Csac_1808 glycosyltransferase, family 4 GT4 Csac_2361 glycosyltransferase, family 4 GT4 Csac_2568 glycosyltransferase, family 4 GT4 Csac_2569 glycosyltransferase, family 4 GT4 Csac_2570 glycosyltransferase, family 4 GT4 Csac_2572 glycosyltransferase, family 4 GT4 Csac_0781 glycogen synthase 2.4.1.21 GT5 Csac_0371 membrane carboxypeptidase (penicillin‐binding protein) GT51 Csac_0490 membrane carboxypeptidase (penicillin‐binding protein) GT51 Csac_2407 acetylesterase 3.1.1.6 CE1 Csac_0205 polysaccharide deacetylase family protein CE4 Csac_0719 polysaccharide deacetylase family protein CE4 Csac_2009 polysaccharide deacetylase family protein CE4 Csac_2371 polysaccharide deacetylase family protein CE4 Csac_2436 acetylxylan esterase 3.1.1.72 CE7 Csac_0213 putative amidohydrolase CE9 Csac_2538 N‐acetylglucosamine‐6‐phosphate deacetylase 3.5.1.25 CE9 Truncated genes are indicated by an asterisk and may not be functional. a EC enzyme commission number; bProteins are grouped by Carbohydrate‐Active enzymes (CAZy; http://www.cazy.org(131)) classification: Glycoside Hydrolase (GH), Glycosyltransferase (GT), Carbohydrate Esterases (CE), Carbohydrate‐Binding Modules (CBM), no Polysaccharide lyases were detected. The list was extended with InterPro (IPR) GH‐ families(111) , which are not in the CAZy database.

68

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Supplementary Table 3.4 Relative expression of ORFs in cells grown on different carbon sources. Only ORFs whose expression is dramatically up‐ or downregulated are shown (log2‐value > 2 or < ‐2, respectively).

GLUCOSE versus RHAMNOSE

Intensit Change in ORF Protein name Function y ratio expressio (log2) n (fold) Csac_2506 xylose/glucose ABC transporter, periplasmic component ABC transporter 2.45 5.47 Csac_2505 xylose/glucose ABC transporter, ATPase component ABC transporter 2.12 4.33 Csac_0243 conserved hypothetical protein 2.50 5.64 Csac_0346 ATP‐binding region, ATPase‐like protein 2.01 4.04 Csac_0431 putative maltodextrin ABC transport system, periplasmic maltosedextrin 2.45 5.47 component utilization Csac_0622 iron‐containing alcohol dehydrogenase 3.56 11.80 Csac_0880 ATP‐dependent Clp protease, ATP‐binding subunit ClpX 2.07 4.19 Csac_1189 fructose‐1,6‐bisphosphate aldolase, class II glycolysis 2.37 5.18 Csac_1461 pyruvate:ferredoxin oxidoreductase subunit beta glycolysis 2.10 4.28 Csac_1606 Acyl carrier protein (ACP) 2.03 4.07 Csac_1628 Conserved hypothetical protein 3.30 9.86 Csac_1823 Predicted metal‐binding, possibly nucleic acid‐binding protein 2.05 4.13 Csac_1824 ribosomal protein L32 ribosome complex 2.15 4.44 Csac_1846 hypothetical protein 2.20 4.61 Csac_1864 NADH‐dependent Fe‐only hydrogenase subunit A hydrogenase (NADH) 3.23 9.35 Csac_1955 pyruvate, phosphate dikinase glycolysis 2.83 7.13 Csac_1990 Rubredoxin 2.01 4.04 Csac_1991 phosphoribosylamine‐glycine ligase purine synthesis 2.26 4.78 Csac_1997 phosphoribosylformylglycinamidine synthase I purine synthesis 2.06 4.16 Csac_2000 putative purine permease purine salvage 3.91 15.07 Csac_2039 desulfoferrodoxin 2.06 4.18 Csac_2040 acetate kinase glycolysis 2.29 4.90 Csac_2044 Hfq protein RNA binding 2.35 5.09 Csac_2204 50S ribosomal protein L7/L12 ribosome complex 2.80 6.98 Csac_2205 Ribosomal protein L10 homolog 2.13 4.39 Csac_2450 conserved secreted protein (prefoldin like domain) 2.26 4.78 Csac_2488 Putative carbamoyl‐phosphate synthase large chain 2.00 4.01

Csac_0375 Circadian clock protein kinase kaiC (EC 2.7.1.37)., putative ‐2.10 4.28 Csac_0407 putative lactaldehyde reductase ‐3.19 9.10 Csac_0476 DNA binding protein, putative transcriptional regulator ‐2.05 4.13 Csac_0792 3,4‐dihydroxy‐2‐butanone‐4‐phosphate synthase/GTP Flavin biosynthesis ‐2.49 5.63

69

Chapter 3

cyclohydrolase II Csac_0793 riboflavin synthase, alpha subunit Flavin biosynthesis ‐2.05 4.15 Csac_0865 rhamnulose‐1‐phosphate aldolase rhamnose pathway ‐3.94 15.39 Csac_0866 Class II Aldolase and Adducin N‐terminal domain protein rhamnose pathway ‐4.66 25.28 Csac_0868 sorbitol‐6‐phosphate 2‐dehydrogenase ‐2.71 6.55 Csac_0870 Lipoate‐protein ligase B keto‐acid ‐2.87 7.30 dehydrogenase Csac_0871 Lipoic acid synthetase keto‐acid ‐4.37 20.71 dehydrogenase Csac_0872 Dihydrolipoamide S‐acetyltransferase (E2 component) keto‐acid ‐5.17 36.00 dehydrogenase Csac_0873 Dihydrolipoamide dehydrogenase (E3 component) keto‐acid ‐4.12 17.34 dehydrogenase Csac_0874 Acetoin/Pyruvate/2‐oxoglutarate dehydrogenase complex (E1 keto‐acid ‐3.01 8.03 component) dehydrogenase Csac_0875 transcriptional regulator, DeoR family keto‐acid ‐4.56 23.55 dehydrogenase Csac_0876 L‐rhamnose isomerase rhamnose pathway ‐2.57 5.96 Csac_1146 2‐isopropylmalate synthase/homocitrate synthase family ‐2.94 7.67 protein Csac_1164 Methionine synthase (vitamin‐B12 dependent), ‐2.15 4.44 Csac_1224 Formate‐‐tetrahydrofolate ligase (Formyltetrahydrofolate ‐2.00 4.00 synthetase) Csac_1633 Putative binding‐protein‐dependent transport systems inner ‐2.02 4.07 membrane comp. Csac_1635 Molybdopterin synthase sulfurylase ‐2.25 4.75 Csac_2698 phospho‐2‐dehydro‐3‐deoxyheptonate aldolase amino acid biosynthesis ‐2.76 6.77 Csac_2699 Prephenate dehydrogenase amino acid biosynthesis ‐2.20 4.58

GLUCOSE versus XYLOSE

Intensity Change in ORF Protein name Function ratio expression (log2) (fold) Csac_0431 putative maltodextrin ABC transport system, periplasmic maltodextrin 2.05 4.13 component utilization Csac_1627 Hypothetical protein purine biosynthesis 3.33 10.07 Csac_1628 Conserved hypothetical protein purine biosynthesis 4.10 17.10 Csac_1992 phosphoribosylaminoimidazolecarboxamide purine biosynthesis 2.60 6.08 formyltransferase/IMP cyclohydrolase

70

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Csac_1993 phosphoribosylglycinamide formyltransferase purine biosynthesis 2.39 5.25 Csac_1994 phosphoribosylformylglycinamidine cyclo‐ligase purine biosynthesis 2.13 4.37 Csac_1995 amidophosphoribosyltransferase purine biosynthesis 2.90 7.45 Csac_1996 phosphoribosylformylglycinamidine synthase II purine biosynthesis 2.07 4.20 Csac_1997 phosphoribosylformylglycinamidine synthase I purine biosynthesis 2.16 4.48 Csac_2000 putative purine permease purine salvage 3.32 10.02

Csac_0240 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐4.73 0.04 systems, ATPase component Csac_0241 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐4.31 0.05 systems, permease component Csac_0242 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐4.68 0.04 systems, periplasmic component Csac_0692 ABC‐type sugar transport system, periplasmic component xylan/xylose ‐4.86 0.03 utilization Csac_0693 ABC‐type sugar transport systems, permease component xylan/xylose ‐2.69 0.16 utilization Csac_0694 ABC‐type sugar transport system, permease component xylan/xylose ‐2.06 0.24 utilization Csac_0695 putative xylose repressor xylan/xylose ‐2.50 0.18 utilization Csac_0696 endo‐1,4‐β‐xylanase precursor xylan/xylose ‐2.63 0.16 utilization Csac_1154 putative xylose isomerase xylan/xylose ‐2.91 0.13 utilization

GLUCOSE versus MIXTURE

Intens Change in ity ORF Protein name Function expression ratio (fold) (log2) Csac_1627 Hypothetical protein purine biosynthesis 2.91 7.50 Csac_1628 Conserved hypothetical protein purine biosynthesis 4.13 17.46 Csac_1992 phosphoribosylaminoimidazolecarboxamide purine biosynthesis 2.52 5.72 formyltransferase/IMP cyclohydrolase Csac_1993 phosphoribosylglycinamide formyltransferase purine biosynthesis 2.23 4.69 Csac_1995 amidophosphoribosyltransferase purine biosynthesis 2.51 5.70 Csac_1996 phosphoribosylformylglycinamidine synthase II purine biosynthesis 2.11 4.32 Csac_1997 phosphoribosylformylglycinamidine synthase I purine biosynthesis 2.19 4.57

71

Chapter 3

Csac_2000 putative purine permease purine salvage 2.01 4.02

Csac_0240 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐3.06 8.35 systems, ATPase component Csac_0241 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐3.89 14.81 systems, permease component Csac_0242 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter ‐4.00 16.01 systems, periplasmic component Csac_0692 ABC‐type sugar transport system, periplasmic component xylan/xylose utilization ‐5.26 38.34 Csac_0693 ABC‐type sugar transport systems, permease component xylan/xylose utilization ‐2.06 4.18 Csac_0694 ABC‐type sugar transport system, permease component xylan/xylose utilization ‐2.14 4.39 Csac_0695 putative xylose repressor xylan/xylose utilization ‐2.19 4.55 Csac_0696 endo‐1,4‐β‐xylanase precursor xylan/xylose utilization ‐2.72 6.59 Csac_0792 3,4‐dihydroxy‐2‐butanone‐4‐phosphate synthase/GTP flavin biosynthesis ‐3.31 9.94 cyclohydrolase II Csac_0793 riboflavin synthase, alpha subunit flavin biosynthesis ‐2.67 6.36 Csac_0794 riboflavin biosynthesis protein RibD flavin biosynthesis ‐2.17 4.49 Csac_0870 Lipoate‐protein ligase B keto‐acid ‐2.01 4.04 dehydrogenase Csac_0871 Lipoic acid synthetase keto‐acid ‐2.84 7.14 dehydrogenase Csac_0873 Dihydrolipoamide dehydrogenase (E3 component) keto‐acid ‐1.97 3.93 dehydrogenase Csac_0875 transcriptional regulator, DeoR family rhamnose pathway ‐2.30 4.94 Csac_1154 putative xylose isomerase xylan/xylose utilization ‐2.55 5.85

XYLOSE versus MIXTURE

Intensity Change in ORF Protein name Function ratio expressio (log2) n (fold) Csac_0792 3,4‐dihydroxy‐2‐butanone‐4‐phosphate synthase/GTP flavin biosynthesis ‐2.82 7.04 cyclohydrolase II Csac_0793 riboflavin synthase, alpha subunit flavin biosynthesis ‐2.52 5.74 Csac_0794 riboflavin biosynthesis protein RibD flavin biosynthesis ‐2.09 4.27 Csac_1499 FMN‐dependent NADH‐azoreductase ‐2.04 4.10

MIXTURE versus RHAMNOSE

72

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Change Intensity in ORF Protein name ratio (log2) expressi on (fold) Csac_0053 Transposase, Mutator family 2.20 4.59 Csac_0240 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 3.74 13.38 systems, ATPase component Csac_0241 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 4.73 26.59 systems, permease component Csac_0242 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 4.66 25.30 systems, periplasmic component Csac_0243 conserved hypothetical protein 2.13 4.38 Csac_0346 ATP‐binding region, ATPase‐like protein 2.43 5.38 Csac_0622 iron‐containing alcohol dehydrogenase 2.74 6.69 Csac_0692 ABC‐type sugar transport system, periplasmic component xylan/xylose utilization 5.72 52.86 Csac_0693 ABC‐type sugar transport systems, permease component xylan/xylose utilization 2.31 4.95 Csac_0694 ABC‐type sugar transport system, permease component xylan/xylose utilization 2.01 4.03 Csac_0695 putative xylose repressor xylan/xylose utilization 2.14 4.41 Csac_0696 endo‐1,4‐β‐xylanase precursor xylan/xylose utilization 2.94 7.66 Csac_1211 glutamine synthetase, type 1 2.52 5.73 Csac_1302 conserved hypothetical protein 2.18 4.52 Csac_1305 Protein of unknown function (DUF1015) 2.35 5.11 Csac_1324 Ribosome‐associated protein Y ribosome complex 2.05 4.14 Csac_1335 30S ribosomal protein S6 ribosome complex 2.17 4.51 Csac_1520 ribosomal protein L31 ribosome complex 2.45 5.47 Csac_1606 Acyl carrier protein (ACP) 2.09 4.25 Csac_1846 hypothetical protein 2.35 5.09 Csac_1864 NADH‐dependent Fe‐only hydrogenase subunit A hydrogenase (NADH) 3.90 14.97 Csac_1953 glyceraldehyde‐3‐phosphate dehydrogenase glycolysis 2.02 4.07 Csac_1955 pyruvate, phosphate dikinase glycolysis 2.97 7.83 Csac_2039 desulfoferrodoxin 2.02 4.05 Csac_2040 acetate kinase glycolysis 2.04 4.12 Csac_2044 Hfq protein RNA binding 2.83 7.13 Csac_2109 CoA binding domain 2.33 5.01 Csac_2204 50S ribosomal protein L7/L12 ribosome complex 3.21 9.26 Csac_2205 Ribosomal protein L10 homolog ribosome complex 2.59 6.02 Csac_2450 conserved secreted protein (prefoldin like domain) 2.25 4.75 Csac_2488 Putative carbamoyl‐phosphate synthase large chain 2.28 4.86 Csac_2504 xylose/glucose ABC transporter, permease component ABC transporter 2.26 4.80 Csac_2505 xylose/glucose ABC transporter, ATPase component ABC transporter 2.28 4.85

73

Chapter 3

Csac_2506 xylose/glucose ABC transporter, periplasmic component ABC transporter 2.60 6.05

Csac_0407 putative lactaldehyde reductase ‐1.96 3.89 Csac_0865 rhamnulose‐1‐phosphate aldolase rhamnose pathway ‐3.95 15.42 Csac_0866 Class II Aldolase and Adducin N‐terminal domain protein rhamnose pathway ‐4.14 17.69 Csac_0872 Dihydrolipoamide S‐acetyltransferase (E2 component) keto‐acid ‐3.46 10.99 dehydrogenase Csac_0873 Dihydrolipoamide dehydrogenase (E3 component) keto‐acid ‐2.14 4.41 dehydrogenase Csac_0875 transcriptional regulator, DeoR family keto‐acid ‐2.25 4.77 dehydrogenase

XYLOSE versus RHAMNOSE

Intens Change in ity ORF Protein name expression ratio (fold) (log2) Csac_0053 Transposase, Mutator family 2.33 5.02 Csac_0240 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 5.42 42.69 systems, ATPase component Csac_0241 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 5.16 35.67 systems, permease component Csac_0242 ribose/xylose/arabinose/galactoside ABC‐type transport ABC transporter 5.34 40.62 systems, periplasmic component Csac_0243 conserved hypothetical protein 3.18 9.09 Csac_0622 iron‐containing alcohol dehydrogenase 2.71 6.55 Csac_0650 hypothetical protein 2.02 4.06 Csac_0692 ABC‐type sugar transport system, periplasmic component xylan/xylose utilization 5.32 40.06 Csac_0693 ABC‐type sugar transport systems, permease component xylan/xylose utilization 2.93 7.64 Csac_0695 putative xylose repressor xylan/xylose utilization 2.46 5.49 Csac_0696 endo‐1,4‐β‐xylanase precursor xylan/xylose utilization 2.84 7.17 Csac_0798 xylulokinase xylan/xylose utilization 2.10 4.28 Csac_0930 cell division protein FtsZ 1.96 3.90 Csac_1154 putative xylose isomerase xylan/xylose utilization 2.09 4.24 Csac_1211 glutamine synthetase, type I 2.43 5.38 Csac_1335 30S ribosomal protein S6 ribosome complex 2.50 5.66 Csac_1460 pyruvate:ferredoxin oxidoreductase subunit alpha 2.17 4.51 Csac_1461 pyruvate:ferredoxin oxidoreductase subunit beta glycolysis 2.07 4.19 Csac_1520 ribosomal protein L31 ribosome complex 2.30 4.91

74

Hydrogenomics of Caldicellulosiruptor saccharolyticus

Csac_1580 reverse gyrase 2.04 4.12 Csac_1824 ribosomal protein L32 ribosome complex 2.24 4.73 Csac_1864 NADH‐dependent Fe‐only hydrogenase subunit A hydrogenase (NADH) 3.17 9.01 Csac_1955 pyruvate, phosphate dikinase glycolysis 2.68 6.41 Csac_2040 acetate kinase glycolysis 2.30 4.93 Csac_2044 Hfq protein RNA binding 3.13 8.78 Csac_2204 50S ribosomal protein L7/L12 ribosome complex 2.79 6.94 Csac_2205 Ribosomal protein L10 homolog ribosome complex 2.45 5.45 Csac_2450 conserved secreted protein (prefoldin like domain) 2.47 5.55 Csac_2488 Putative carbamoyl‐phosphate synthase large chain 2.68 6.40 Csac_2505 xylose/glucose ABC transporter, ATPase component ABC transporter 2.27 4.81 Csac_2506 xylose/glucose ABC transporter, periplasmic component ABC transporter 2.65 6.26

Csac_0407 putative lactaldehyde reductase ‐2.51 5.70 Csac_0432 Hypothetical protein ‐2.06 4.16 Csac_0792 3,4‐dihydroxy‐2‐butanone‐4‐phosphate synthase/GTP riboflavin biosynthesis ‐2.00 3.99 cyclohydrolase II Csac_0865 rhamnulose‐1‐phosphate aldolase rhamnose pathway ‐3.74 13.39 Csac_0866 Class II Aldolase and Adducin N‐terminal domain protein rhamnose pathway ‐4.77 27.31 Csac_0870 Lipoate‐protein ligase B keto‐acid ‐2.15 4.45 dehydrogenase Csac_0871 Lipoic acid synthetase keto‐acid ‐2.61 6.13 dehydrogenase Csac_0872 Dihydrolipoamide S‐acetyltransferase (E2 component) keto‐acid ‐4.19 18.23 dehydrogenase Csac_0873 Dihydrolipoamide dehydrogenase (E3 component) keto‐acid ‐2.86 7.28 dehydrogenase Csac_0874 Acetoin/Pyruvate/2‐oxoglutarate dehydrogenase complex (E1 keto‐acid ‐2.08 4.23 component) dehydrogenase Csac_0875 transcriptional regulator, DeoR family rhamnose pathway ‐3.27 9.63 Csac_0876 L‐rhamnose isomerase rhamnose pathway ‐2.15 4.44 Csac_1030 Oligopeptide transport ATP‐binding protein oppD ‐2.12 4.34 Csac_1146 2‐isopropylmalate synthase/homocitrate synthase family ‐2.33 5.03 protein Csac_1224 Formate‐‐tetrahydrofolate ligase (Formyltetrahydrofolate ‐2.35 5.09 synthetase) Csac_1635 Molybdopterin synthase sulfurylase ‐2.00 4.01 Csac_2590 LemA family protein ‐2.41 5.30 Csac_2698 phospho‐2‐dehydro‐3‐deoxyheptonate aldolase amino acid biosynthesis ‐2.28 4.85

75

Chapter 3

Rhamno se Gluc ose 1

4 2

3 Xylose-Glucose Xylose

Supplementary Figure 3.1. The experimental 4‐loop design of C. saccharolyticus grown on L‐rhamnose, glucose, xylose and a mix of xylose and glucose.

76

Chapter 4

Carbohydrate utilization patterns by the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus reveal broad growth substrate preferences

M. R. A. Verhaart*, A. L. VanFossen*, S. M. W. Kengen, and R. M. Kelly. (2009) Appl. Environ. Microbiol. 75:7718‐7724.

* These authors contributed equally to this work Chapter 4

Abstract Co‐utilization of hexoses and pentoses derived from lignocellulose is an attractive trait in microorganisms considered for consolidated biomass processing to biofuels. This issue was

examined for the H2‐producing, extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus growing on individual monosaccharides (arabinose, fructose, galactose, glucose, mannose and xylose), mixtures of these sugars, as well as on xylan and

xyloglucooligosacchrides. C. saccharolyticus grew at approximately the same rate (td ~ 95 min) and to the same final cell density (1‐3 x 108 cells/ml) on all sugars and sugar mixtures tested. In the monosaccharide mixture, while simultaneous consumption of all monosaccharides was observed, not all were utilized to the same extent (fructose > xylose/arabinose > mannose/glucose/galactose). Transcriptome contrasts for monosaccharide growth showed that minimal changes were observed in some cases (e.g., 32 ORFs changed  2‐fold for glucose/galactose) while substantial changes occurred for cases involving mannose (e.g., 353 ORFs  2‐fold for glucose/mannose). No evidence for catabolite repression was noted for either growth on multi‐sugar mixtures or in the corresponding transcriptomes. Based on the whole‐genome transcriptional response analysis and comparative genomics, carbohydrate specificities for transport systems could be proposed for most of the 24 putative carbohydrate ATP‐binding cassette (ABC) transporters and one phosphotransferase system (PTS) identified in C. saccharolyticus. While most transporter genes responded to individual monosacchrides and polysaccharides, Csac_0692‐0694 was up‐regulated only in the monosaccharide mixture. The results here affirm the oligotrophic growth physiology of C. saccharolyticus on carbohydrates representative of lignocellulosic biomass and suggest that this bacterium holds promise for biofuels applications.

Introduction Cellulose and hemicellulose comprise approximately 70 percent (dry basis) of plant biomass (170) and, as such, these complex carbohydrates are promising targets for alternative energy feedstocks. For biofuels production, lignocellulose must be deconstructed into fermentable sugars by various combinations of physical, physical‐chemical, chemical, and

78

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

biological processing steps (116, 151, 191, 223). The more efficiently and economically this is accomplished, the more attractive this material becomes for alternative fuels (96). For this reason, there is great interest in identifying microorganisms that can minimize or eliminate the need for physical and chemical pre‐treatment of plant‐based biomass, while at the same time being capable of simultaneously fermenting the resulting pentoses and hexoses to biofuels (102, 108). In environmental settings, microbial survival can be based on the capacity to utilize a variety of available carbohydrates as carbon and energy sources (102, 173). When complex polysaccharides are involved, as is the case for plant‐based biomass, extracellular glycoside hydrolases depolymerize potential growth substrates, rendering them into oligosaccharides and monosaccharides that can be transported into the cytoplasm for further processing. Metabolism of these sugars is thus mediated by the existence and regulation of membrane‐ based molecular transporters and impacted by regulatory processes, such as carbon catabolite repression (CCR) (173). The number, specificity and control of transporters presumably reflect the nutritional needs and preferences for a given microorganism. From a biofuels perspective, microorganisms that have a broad growth substrate range (i.e., oligotrophs) are attractive, since rapid and efficient metabolism of the spectrum of sugars comprising both cellulose and hemicellulose fractions is critical for economic lignocellulosic biomass conversion (101, 133). Microbial uptake mechanisms for sugar transport have been studied in some detail and include phosphoenolpyruvate‐dependent phophotransferase systems (PTS), ATP‐binding cassette (ABC) transporters, and proton‐linked transport systems (40, 135). The distribution and number of these transport mechanisms varies from microorganism to microorganism. The genomes for several anaerobic, extremely thermophilic bacteria, such as Thermotoga maritima (224) and Carboxydothermus hydrogenoformans (6, 40, 224), indicate a lack of a PTS in lieu of ABC transporters for sugar uptake (216). Alternatively, other thermophiles, such as Thermoanaerobacter tengcongensis, contain one PTS in addition to multiple ABC transporters (185).

Caldicellulosiruptor saccharolyticus is an extremely thermophilic (growth Topt 70‐75°C), gram‐positive, fermentative anaerobe initially isolated from wood in the flow of a thermal

79

Chapter 4

spring in New Zealand (45, 86, 157, 207). C. saccharolyticus can utilize a variety of simple (e.g., glucose, xylose, arabinose, mannose) and complex carbohydrates (e.g., cellulose, xylan, mannan) associated with lignocellulose as carbon and energy sources, producing hydrogen with yields approaching the Thauer limit (e.g., 4 H2 of per mole of glucose consumed) (204). Not surprisingly, of the 2,679 predicted coding sequences in the C. saccharolyticus genome, over 177 ORFs encode ABC transporter genes and at least one PTS, which together help to sequester a broad spectrum of potential growth substrates (85, 204). Studies to date show that C. saccharolyticus does not exhibit glucose‐based carbon catabolite repression (CCR) and can co‐ferment glucose and xylose (204). To understand more about the oligotrophic characteristics of C. saccharolyticus, transcriptional response analysis was used to examine growth on monosaccharide and polysaccharides as this relates to specific regulatory and functional features of resident sugar transport systems in this bacterium.

Materials and methods

Growth of the microorganism on sugar substrates. C. saccharolyticus DSM 8903 (also, ATCC 43494) was passed seven times in serum bottles shaken at 100 rpm under anaerobic conditions (N2 headspace) on the substrate of interest in modified DSMZ 640 medium before inoculating a 400 mL screw top, batch culture at 70°C, shaken at 100 rpm, containing either 0.5 g substrate per liter or 0.5 g of each substrate

per liter. Growth was followed for 18 hours. Growth substrates included D‐glucose, D‐

mannose, L‐arabinose, D‐xylose, D‐fructose, D‐galactose, xylan (oatspelt) (Sigma‐Aldrich, St. Louis, MO), xyloglucan and xyloglucan oligosaccharides (provided by H. Brumer of the Royal Institute of Technology, Stockholm, Sweden).

Analysis of residual sugar and fermentation products.

The cultures grown on the various sugars were harvested by centrifugation at 9,000 x g, following which the supernatants were clarified by passage through 0.22 μm PVDF filters (FisherSci, Hampton, NH). Monosaccharide and organic acid concentrations were determined by HPLC, using a Waters 1525 binary pump with Breeze control software (Waters, Milford, MA). Sugar concentrations were determined using a SP0810 sugar column

80

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

(Showa Denko, New York, NY) operating at 800C with water as the mobile phase. Organic acid byproducts were also determined by HPLC, using an Aminex HPX‐87H column (Bio‐Rad 0 Laboratories, Hercules, CA), with a mobile phase of 5 mM H2SO4 at 65 C. Peak heights for all samples, analyzed with an RI detector (Waters 2414), were integrated and the values were compared to standard curves for quantification.

Isolation of RNA.

Isolation of total RNA from C. saccharolyticus was performed on cells that were grown at 70°C until early‐ to mid‐logarithmic phase (mid‐107 cells/mL) on the various growth substrates. Cells were rapidly cooled to 4oC and then harvested by centrifugation at 5,000 x g, before storage at ‐80°C until they were processed further. Total RNA was isolated using a protocol described previously, with the additional step of sonicating the cells after re‐ suspension in TRIzol reagent (Invitrogen, San Diego, CA) (64). cDNA was produced from the collected RNA using Superscript III reverse transcriptase (Invitrogen), random primers (Invitrogen), and the incorporation of 5‐(3‐aminoallyl)‐2'‐deoxyuridine‐5'‐triphosphate (Ambion, Austen, TX), as described elsewhere (204).

Microarray protocols.

A spotted whole‐genome C. saccharolyticus microarray was developed from 60‐mer oligonucleotides based on at least 2679 open reading frames (ORFs), as discussed previously (155). cDNA was labeled with either Cy3 or Cy5 dye (GE Healthcare, Little Chalfont, United Kingdom) and hybridized to Corning Ultra‐Gap II slides (Corning, Acton, MA). Samples were hybridized according to either a ten slide (Supplementary Figure 1) or five slide loop experimental design (Supplementary Figure 2). Microarray slides were scanned using a Packard BioChip Scanarray 4000 scanner. ScanArray Express (v2.1.8, Perkin‐Elmer, Waltham, MA) was used to quantitate signal intensities, prior to analysis with JMP Genomics 3.0 (SAS, Cary, NC), as described previously (220), using a mixed‐effects analysis of variance model (220). ORFs that were differentially transcribed 2‐fold or more and met the Bonferroni statistical criterion were considered to be up‐ or down‐regulated (160).

81

Chapter 4

Microarray accession numbers.

The microarray platform used in this study is available in the Gene Expression Omnibus database (http://www.ncbi.nlm.nih.gov/geo/) under accession number GPL6681. The raw

data along with the final log2 fold changes have been deposited in the same database under accession number GSE17436.

Bioinformatic identification of sugar transporter genes.

Protein sequences of putative substrate binding proteins found in the C. saccharolyticus genome (http://www.ncbi.nlm.nih.gov/nuccore/CP000679) were subjected to similarity searches (BLASTP) at the National Center for Biotechnology Information (NCBI) website, based on the Thermotoga maritima MSB8 genome (http://www.ncbi.nlm.nih.gov/nuccore/AE000512), to identify ABC transporter systems involved in carbohydrate import. Similarity was defined as ≥ 25% identity and Eval ≤ 10‐5. Candidate ORFs were then compared to carbohydrate transporter predictions by TransportDB (85, 204). Assignment of substrate(s) for given transporters was based on the following criteria: (i) the substrate binding protein (SBP) was more than 30% identical to a previously studied protein in Escherichia coli, Bacillus subtilis or T. maritima; (ii) more than one gene of the operon was conserved; or, most importantly, (iii) functional genomics analysis of the ORFs encoding proteins in the transporter responded by 2‐fold or more to C. saccharolyticus growth on specific sugars. To identify CcpA homologs, sequences from Bacillus subtillis 168 and Lactobacillus brevis ATCC 367 were used, with a threshold similarity of ≥ 35% identity and Eval ≤ 10‐5.

Supplemental material

Supplemental material for this chapter may be found at: http://aem.asm.org/cgi/content/full/75/24/7718/DC1

82

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

Results and discussion

Carbohydrate utilization patterns

C. saccharolyticus was grown on a variety of monosaccharides (i.e., glucose, xylose, galactose, arabinose, mannose, fructose and mixtures of the monosaccharides) commonly found in lignocellulose. On all sugars and sugar mixtures tested, cultures typically grew to final cell densities of 1‐3 x 108 cells/mL at 700C, with doubling times of ~ 95 min. While previous studies of C. saccharolyticus monosaccharide co‐fermentation showed that xylose was used to a greater extent than glucose (67), its capacity to co‐utilize other biomass‐ related sugars has not been reported. Figure 4.1 shows the extent of utilization for individual sugars present in a mixture of pentoses and hexoses. Several monosaccharides were utilized simultaneously; fructose was consumed to the greatest extent followed by the pentoses, xylose and arabinose (fructose > arabinose > xylose > mannose > glucose > galactose). While fructose has not been a specific focus as a biofuels substrate, it is the main constituent of fructans, which are found in many plant families including wheat, barley, onion and chicory (85, 204). The possible source of fructans in thermal springs, the natural habitat of C. saccharolyticus, however, is not clear. To examine sugar substrate preferences more carefully, cultures were grown on arabinose + glucose and on a mixture of the hexoses (fructose, galactose, mannose and glucose). C. saccharolyticus growth on arabinose + glucose showed that the pentose was consumed to a greater extent (data not shown); a similar result was reported for growth on xylose + glucose (204). Preference for pentoses may be related to the fact that C. saccharolyticus lacks an oxidative pentose‐phosphate pathway to convert hexoses to pentoses (204). Growth on a mixture of hexoses changed the utilization pattern to fructose > galactose > glucose > mannose (data not shown) from fructose > mannose > glucose > galactose when xylose and arabinose are present.

83

Chapter 4

Figure 4.1. Simultaneous utilization of sugars during growth of C. saccharolyticus in a batch culture. Equal amounts of each of the six monosaccharides were added to cultures at time of inoculation. The remaining concentration of each sugar is shown at various times during the growth of C. saccharolyticus in batch culture. Average values from triplicate experiments are shown; standard deviation for each sugar measured by HPLC was < 0.02 g/L.

Monosaccharide transcriptomes.

RNA from cultures grown on the six monosaccharides, both individual sugars and a mixture, was collected to examine the associated transcriptomes (see Tables S1‐S7). Table 4.1 summarizes the number of ORFs differentially transcribed 2‐fold or more for these experiments. Typically, fewer than ~200 ORFs were differentially transcribed (up‐regulated + down‐regulated) when comparing one monosaccharide to another, with the exception of glucose:mannose (353 ORFs).

84

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

Table 4.1. Transcriptional response contrasts for C. saccharolyticus growth on monosaccharides

Sugar 1 Sugar 2 ORFs Up‐regulated  2‐fold on ORFs Up‐regulated  2‐fold on Sugar 1 [transporter ORFs] Sugar 2 [transporter ORFs] ARABINOSE Xylose 20 [0] 138 [12] Fructose 22 [4] 80 [11] Galactose 10 [3] 48 [8] Glucose 34 [3] 45 [5] Mannose 32 [3] 25 [5] Sugar Mix 5 [0] 84 [15]

XYLOSE Fructose 19 [4] 22 [8] Galactose 70 [8] 25 [3] Glucose 59 [11] 23 [0] Mannose 113 [9] 56 [3] Sugar Mix 62 [17] 69 [11] FRUCTOSE Galactose 36 [6] 25 [4] Glucose 32 [8] 29 [0] Mannose 92 [10] 29 [1] Sugar Mix 27 [5] 29 [6] GALACTOSE Glucose 21 [10] 11 [3] Mannose 57 [9] 51 [4] Sugar Mix 12 [6] 27 [8] GLUCOSE Mannose 174 [6] 179 [11] Sugar Mix 13 [5] 69 [18] MANNOSE Sugar Mix 58 [3] 91 [11]

In the complete genome, approximately 25% of all ORFs (~600) were annotated as either “conserved hypothetical protein” or “protein of unknown function” (204) ‐ of these, the mannose:glucose contrast implicated over a tenth of them. Growth on mannose triggered transcription of two CcpA homologs, Csac_0389, and Csac_2494 (Figure S3). These two CcpA homologs were transcribed higher on mannose than during growth on the other individual monosaccharides. CcpA, coupled with a serine‐phosphorylated Crh, can repress the expression of genes involved in the catabolism of alternative sugars by binding to the

85

Chapter 4

upstream catabolic repression element (cre) sites. However, searches for possible cre sites near gene promoters, based on cre sites found in other microorganisms such as Bacillus subtilis, were unproductive. In the presence of certain monosaccharides, several genes related to carbohydrate metabolism were differentially transcribed, as expected (Figure 4.2 and 4.3).

Figure 4.2. Differential transcription of genes encoding enzymes involved in non‐oxidative pentose phosphate pathway in C. sacharolyticus. Abbreviations: A, arabinose; F, fructose; Gal, galacose; Glu, glucose; M, mannose; X, xylose; SM, sugar mix; L‐Ru‐5‐P, L‐ribulose‐5‐phosphate; D‐Ru‐5‐P, D‐ribulose‐5‐phosphate; D‐Ri‐5‐P, D‐ ribose‐5‐phosphate; D‐Xu‐5‐P, D‐xylulose‐5‐phosphate; GAP, glyceraldehydes‐3‐phosphate; S‐7‐P, sedoheptulose‐7‐phosphate; F‐6‐P, fructose‐6‐phosphate; E‐4‐P, erythrose‐4‐phosphate. Least squares mean (lsm) estimates (see Methods) of transcript level shown for selected genes (ORF # in box). Black and white denote transcript levels above (black) and below (white) the mean across all genes.

These genes were often, also up‐regulated in the presence of the sugar mixture. For example, three genes in the Leloir pathway involved in galactose metabolism, galactose‐1‐ phosphate uridylyltransferase (Csac_1510, EC 2.7.7.12), galactokinase (Csac_1511, EC 2.7.1.6) and UDP‐galactose 4‐epimerase (Csac_1512, EC 5.1.3.2), were highly transcribed compared to the other five monosaccharides but not the sugar mixture. It was interesting

86

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

that the monosaccharides did not stimulate transcription of many of the 53 glycoside hydrolases identifiable in the C. saccharolyticus genome, perhaps suggesting that these genes are triggered by oligo‐ and polysaccharide substrates.

Figure 4.3. Differential transcription of genes encoding enzyme involved in glycolysis in C. saccharolyticus. Abbreviations: A, arabinose; F, fructose; Gal, galacose; Glu, glucose; M, mannose; X, xylose; SM, sugar mix; M‐ 6‐P, mannose‐6‐phosphate; F‐6‐P, fructose‐6‐phosphate; Gal‐1‐P, galactose‐1‐phosphate; G‐1‐P, glucose‐1‐ phosphate; G‐6‐P, glucose‐6‐phosphate; F‐1,6‐bisP, fructose‐1,6‐bisphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehydes‐3‐phosphate. Least squares mean (lsm) estimates (see Methods) of transcript level shown for selected genes (ORF # in box). Black and white denote transcript levels above (black) and below (white) the mean across all genes.

ABC‐type carbohydrate transporter identification in C. saccharolyticus genome.

At least 177 putative ABC transporter genes could be identified in the C. saccharolyticus genome (40, 135). These comprise 24 putative carbohydrate transporters, based on their substrate binding protein similarity to proteins identified in the genome of T. maritima (160) and/or prediction by TransportDB (40). Of these, two belong to the carbohydrate uptake (CUT) 2 family, two are members of the di/oligopeptide transporter family (Dpp/Opp), and the rest are CUT1‐types (Table 4.2). In T. maritima, many of the transporters are co‐localized with glycoside hydrolases (204), which also seems to be the case for C. saccharolyticus. For

87

Chapter 4

these transporters in C. saccharolyticus, predictions for sugar substrate specificity were based on one or more of the following criteria: (i) functional genomics analysis of the ORFs in the transporter showed a 2‐fold change for a given sugar contrast; (ii) the substrate binding protein identity was > 30% compared to the homologous protein in Escherichia coli, Bacillus subtilis or T. maritima; or, (iii) genes located near the ABC transporter are conserved in the same relative location in another well‐studied genome. As mentioned earlier, C. saccharolyticus genome also encodes a phosphoenolpyruvate‐dependent phophotransferase system (PTS), which is annotated as fructose‐specific (101, 178).

Proposed annotation of monosaccharide transporters in C. saccharolyticus.

Table 4.2 shows the proposed assignment of substrates to ABC sugar transporters in the C. saccharolyticus genome, based on functional genomics and bioinformatics analysis. CUT2 family transporters consist of an extracellular substrate binding protein, one permease and one ATPase; the CUT2 ATPase with two sets of Walker A and B sites is, in general, longer than other ATPases (101). CUT2 family transporters only import monosaccharides. There are only two CUT2 family transporters identifiable in the C. saccharolyticus genome, Csac_0238, 0240‐0242 and Csac_2504‐2506. These may have broad specificity for monosaccharides: Csac_0238‐0242 was up‐regulated on arabinose, galactose and xylose, while Csac_2504‐ 2506 responded to xylose, glucose, fructose and galactose. The CUT1 family ABC transporters have one substrate binding protein, two permeases and one ATPase; they transport both monosaccharides and di/oligosaccharides, in addition to glycerol phosphate (63, 198). Similar to T. maritima, the C. saccharolyticus CUT1 ATPases utilized are not necessarily located in the same operon as the permeases and substrate binding protein. Three CUT1 family transporters were differentially transcribed in the C. saccharolyticus genome during growth on monosaccharides. Csac_0440‐0442 was up‐regulated only on galactose. The two other CUT1 transporters, Csac_2321‐2322,2324,2326 and Csac_2491‐ 2493, both responded to glucose, xylose and fructose. It is interesting that the CUT1 family transporter encoded in Csac_0692‐0694 was up‐ regulated when C. saccharolyticus was grown on the six monosaccharide mixture. Directly downstream from Csac_0692‐0694 is a ROK family transcriptional regulator (Csac_0695) with the highly conserved consensus sequences 1 and 2, plus the conserved DNA‐binding

88

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

(helix‐turn‐helix) motif (42, 165). Csac_0695 was highly transcribed only in the monosaccharide mixture. This was the only transcription factor in the genome that responded in this way, suggesting it regulates this transporter. The protein encoded in Csac_0695 has 40% aa sequence identity to the ROK family xylose repressors found in two strains of B. subtilis (XylR1, P94490; XylR2, P16557), which negatively regulate the xyl operon by binding to the xyl operators in the absence of the inducer xylose (101). A similar role in C. saccharolyticus seems likely. Members of the Dpp/Opp family are known to transport di‐ and oligopeptides, simple and complex carbohydrates, nickel and heme into the cells through an extracellular binding protein, two permeases and two ATPases (3). The Dpp/Opp family transporter encoded in Csac_1028‐1032 was highly transcribed on every monosaccharide and on the sugar mixture. The other Dpp/Opp transporter, encoded in Csac_0261‐0265, was highly transcribed only on fructose. Upstream of this gene cluster is a LacI family regulator (Csac_0260), and two putative glycoside hydrolases, which were also clearly up‐regulated on fructose. Csac_0258, a putative intracellular β‐fructosidase, may be an invertase that hydrolyzes sucrose to fructose and glucose. This suggests that Csac_0261‐ 0265 could be a sucrose transporter; in fact, C. saccharolyticus grows as well on sucrose as it does on the monosaccharides examined here (unpublished results). In the presence of fructose, the ORFs encoding a putative PTS (Csac_2437‐2440) were induced 1.8‐ to 35‐fold, compared to glucose, xylose and arabinose. These ORFs were also up‐regulated in the presence of mannose. Note that a PTS can transport different sugars, with each EII being sugar‐specific (5). An inducible mannose transporter occurs in the fructose/mannose PTS in Streptococcus mutans, importing fructose and mannose into the cytoplasm (204). EIIA (Csac_2440) and EIIBC (Csac_2439) in the putative C. saccharolyticus PTS were also up‐ regulated on galactose, but this was not the case for EI (Csac_2437) or HPr (Csac_2438). Fructose and galactose appear to have ABC transporters (Csac_0261‐0265 and Csac_0440‐ 0442) (see below), but for mannose, the only identifiable transporter is this PTS. The unusual transcriptome for mannose merits further examination. All of the transporters discussed above that were up‐regulated on the individual monosaccharides were also up‐ regulated during growth on the six sugar mixture, with the lone exception of the galactose‐ specific transporter (Csac_0440‐0442).

89

Chapter 4

Table 4.2: Putative substrates for sugar ABC transporters in C. saccharolyticus ABC GH GH SBP1 Proposed Transporter Group Neighbor neighbor SBP Homolog in T. maritima (Csac_) substrate(s) (Csac_) (Csac_) family2 TM0810 (N‐acetylglucosamine 0126‐0128 CUT1 0126 0129, 0130 2 , 15 unknown or Nag sacch) 0238, 0240‐ arabinose, CUT2 0242 ‐ ‐ TM0114 (monosaccharides) 42 galactose, xylose 0261‐0265 Dpp/Opp 0261 0258, 0259 ? TM1223 (cellobiose, barley) fructose, sucrose 0297‐0299 CUT1 0297 0296 43 None unknown None ‐ proximate to two xylose 0391‐0394 CUT1 0391 ‐ ‐ isomerase‐like TIM barrel unknown proteins TM1839 (maltose, maltotriose, 0427‐28,31 CUT1 0431 0426 13 maltodextrin (17) trehalose) 0440‐0442 CUT1 0440 0439, 0444 65, 65 TM1220 (cellobiose, barley) galactose TM0810 0679‐0682 CUT1 0682 0678 5 (N‐acetylglucosamine or Nag xyloglucan sacch) 0692‐0694 CUT1 0692 0689, 0696 13, 10 TM1120 (glycerol‐3‐phosphate) monosaccharides 1028‐1032 Dpp/Opp 1032 ‐ ‐ TM1746 (β‐mannans) monosaccharides 1357‐1359 CUT1 1359 1354 None xyloglucan 1560, 1557‐1559 CUT1 1557 43, 51, 4 TM1120 (glycerol‐3‐phosphate) xyloglucan 1561, 1562 2321‐2322, glucose, xylose, CUT1 2326 ‐ ‐ TM0102 (CUT2 unknown) 2324, 2326 fructose 2408, 10, 39, 10, TM1204 (maltose, maltotriose, xylooligosacchari 2412‐2414 CUT1 2412 2409, 43 mannotetraose) des 2410, 2411 TM1839 (maltose, maltotriose, xylooligosacchari 2417‐2419 CUT1 2419 ‐ ‐ trehalose) des xylose, glucose, 2491‐2493 CUT1 2493 ‐ ‐ None fructose xylose, glucose, 2504‐2506 CUT2 2506 ‐ ‐ TM0114 (monosaccharides) fructose, galactose TM0595 (N‐acetylglucosamine glucooligosacchar 2514‐2516 CUT1 2516 2513 30 or Nag sacch) ides None ‐ proximate to multi‐ 2529‐2531 CUT1 2529 2527, 2528 43,5 unknown domain GHs N‐ acetylglucosamin TM0810 (N‐acetylglucosamine 2542‐2544 CUT1 2544 2539 20 e or GlcNAc polysaccharide) GlcNAc polysaccharide 2552‐2554 CUT1 2552 2548 16f3 None unknown 2692‐2694 CUT1 2694 2689 67 None unknown 2730,2731, 88,39f3, None – proximate to multi‐ 2724‐2726 CUT1 2726 unknown 2732 39f3 domain GHs 2740‐2742 CUT1 2742 2734, 2748 2, 4 TM1120 (glycerol‐3‐phosphate) unknown 1. SBP: substrate binding protein 2. GH family indicates glycosides hydrolases family (www.cazy.org) 3. “f” indicates fragment of the glycoside hydrolase family

90

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

Complex carbohydrate response

C. saccharolyticus was also grown on xylan (oat spelt), xyloglucan, xylogluco‐ oligosaccharides (19, 115) to facilitate transcriptional response comparisons to growth on the monosaccharides glucose and xylose. Final cell densities and growth rates with the complex carbohydrate substrates were comparable to the monosaccharides, although there were longer lag times on xyloglucan and xylan. The transporter encoded in Csac_0692‐0694, that was up‐regulated only during growth on the monosaccharide mixture, was not induced on the complex carbohydrates, suggesting it is involved in general monosaccharide transport into the cell. ABC transporters encoded in Csac_0679‐0682 and Csac_1357‐1359 were highly transcribed on the complex carbohydrates. Both are located near putative glycoside hydrolases (Csac_0678, GH family 5; Csac_1354, GH family 31) involved in the breakdown of glucans or xylan, suggesting a role in the uptake of degradation products of these carbohydrates. The xylan utilization cluster in C. saccharolyticus (105) contains the ABC transporters (Csac_2412‐2414 and Csac_2417‐2419) and, as expected, the components were transcribed at higher levels on xylan compared to the other sugars. Growth on the xyloglucooligosaccharides mixture triggered transcription of components of the transporter Csac_2514‐2516, indicating a role in the uptake of this substrate.

Transporters with unknown function

Several putative transporters could not be assigned specific substrates, based on transcriptional response and bioinformatic analysis. One of these putative transporters, Csac_1557‐1559, did not respond during growth on the simple or complex carbohydrates used here. However, this transporter is co‐localized in the genome with glycoside hydrolase family 4, 43 and 51 enzymes, and may transport a different type of glucan‐ and/or xylan‐ based saccharide. Further examination of the genome showed that the putative sugar transporter encoded in Csac_2542‐2544 is downstream of a N‐acetylglucosamine‐6‐ phosphate deacetylase (Csac_2538) and a N‐acetyl‐β‐hexosaminidase (Csac_2539). Based on this, and its homology to TM0810, this transporter could be associated with N‐ acetylglucosamine‐based polysaccharides. The other loci encoding ABC sugar transporters (see Table 4.2) that did not respond to any of the substrates tested and were, furthermore,

91

Chapter 4

consistently transcribed at very low levels. These transporters (Csac_0126‐0128, Csac_0261‐ 0265, Csac_0391‐0394, Csac_2529‐2531, Csac_2552‐2554, Csac_2692‐2694, Csac_2724‐ 2726 and Csac_2740‐2742) are located in the genome near GH families 2, 4, 5, 15, 16, 39, 43, 67 and 88 or near carbohydrate metabolism genes, so a role in carbohydrate acquisition cannot be ruled out, especially given C. saccharolyticus’ varied carbohydrate appetite.

No evidence of carbon catabolite repression (CCR)

Here, there was no indication of CCR. This physiological characteristic, as observed in a low G+C content, gram‐positive bacteria, typically involves three components: cis‐acting sequences or catabolite responsive elements (CREs), catabolite control protein A (CcpA), which binds to the CRE element, and the Crh (catabolite repression HPr) (68). CCR has been noted in various gram‐positive bacteria, including Bacillus subtilis (24), Lactococcus lactis (73) and Streptococcus pneumoniae (197) and is mediated by transcriptional control by a CcpA/HPr‐Ser‐P complex binding to cre (103). In the standard low G+C, gram‐positive model of CCR, the primary sugar triggers the transcriptional repression of genes encoding saccharolytic components, related to other sugars, through the serine‐phosphorylated HPr. With C. saccharolyticus, there was specific induction of carbohydrate catabolism genes for growth on individual sugars, although the carbohydrate uptake systems were, in general, differentially transcribed during growth on individual sugars. When the organism was presented a mixture of monosaccharides, including glucose, genes within these inducible loci were not impacted. Examination of the C. saccharolyticus genome identified an HPr homolog (Csac_1163), with the active His‐15 site replaced with a serine residue. Thus, it cannot be involved in PTS transport. It could, however, be phosphorylated at the active site Ser‐46 and participate in CCR in C. saccharolyticus as Crh. A number of putative CcpA homologs were found, but these candidates contain most of the residues that are typical for LacI‐type transcription factors but are missing residues which are typical for CcpAs (Figure S3) (37). As yet, their role in C. saccharolyticus is unclear, since no cre sequences in promoter‐operator regions are present. All these results point to no involvement of a global regulatory system in C. saccharolyticus for carbohydrate utilization.

92

Carbohydrate utilization patterns by Caldicellulosiruptor saccharolyticus

Summary

The results here clearly demonstrate that C. saccharolyticus can metabolize a number of monosaccharides simultaneously, a process apparently mediated by a system of transporters with low specificity and an absence of CCR. This phenotype is advantageous to bacteria coping in natural environments where substrate availability and composition is variable. Given its broad appetite for plant‐based carbohydrates, C. saccharolyticus is an attractive candidate for fermentation of biomass‐derived substrates to biofuels, such as hydrogen.

Acknowledgments

R.M.K. acknowledges support from the US National Science Foundation (CBT0617272) and the Bioenergy Science Center (BESC), a U.S. DOE Bioenergy Research Center supported by the Office of Biological and Environmental Research. A.L.V. acknowledges support from a US Department of Education GAANN Fellowship. M.R.A.V and S.M.W.K. acknowledge support from the Commission of the European Communities, Sixth Framework Programme, Priority 6, Sustainable Energy Systems (019825 HYVOLUTION). Microarray fabrication was financially supported in part by Danisco (Genencor International, Inc., Palo Alto, CA). The authors thank Harry Brumer (Royal Institute of Technology, Stockholm, Sweden) for providing samples of xyloglucans.

93

Chapter 4

94

Chapter 5

Analysis of the rhamnose pathway in Caldicellulosiruptor saccharolyticus.

Marcel R. A. Verhaart, Vincent van der Vlist, Dirk‐Jan de Groot, Alfons J.M. Stams. Servé M. W. Kengen

Manuscript in preparation Chapter 5

Abstract

Caldicellulosiruptor saccharolyticus is an anaerobic thermophilic bacterium that is able to produce hydrogen at high rates on a wide variety of substrates, including rhamnose. The pathway for rhamnose utilization was reconstructed based on genomic, transcriptomic and fermentation data. In agreement with the anticipated pathway, fermentation of rhamnose results in the production of 1,2‐propanediol in addition to the common products acetate,

CO2 and hydrogen in a 1:1:1:1 ratio. This stoichiometry demands that all NADH is used for lactaldehyde reduction and not for forming hydrogen. A considerable upregulation of all genes of the rhamnose pathway was observed during growth on rhamnose, in comparison to glucose. Growth on rhamnose also led to a highly upregulated gene cluster, coding for an α‐keto acid dehydrogenase with unknown substrate specificity. In agreement with the special redox situation, addition of carbon monoxide (CO) inhibited hydrogen and acetate formation, but not the NADH‐dependent formation of 1,2‐propanediol. Analysis of the carbon balance suggests that in the presence of CO some unknown reduced end product is produced, presumably derived from ferredoxin. The CO‐dependent increase in cell lysis could not completely account for the missing carbon.

Introduction

Caldicellulosiruptor saccharolyticus was selected as a model organism for fermentative hydrogen production at elevated temperatures (157). C. saccharolyticus is an extremely thermophilic (growth temperature optimum 70‐75°C), gram‐positive, fermentative anaerobe, initially isolated from a piece of wood in the flow of a thermal spring in New Zealand (89, 204, 209). Besides the advantages of growing at a higher temperature, the capacity to use complex substrates like cellulose and xylan and is able to use C5 and C6 sugars (e.g. glucose and xylose) simultaneously. The latter aspect is important in view of the abundant sugar types present in the renewable resources which are intended to be used (45, 86, 204, 209). Due to these advantages C. saccharolyticus has been shown to produce up to 3.6 moles of hydrogen per moleof hexose, which is close to the theoretical limit of 4 moles per mole of hexose (45, 204). Another advantage of this organism for research purposes is that the genome has been sequenced and DNA microarrays are available (188).

96

Rhamnose pathway in Caldicellulosiruptor saccharolyticus

In the metabolism of C. saccharolyticus the Embden‐Meyerhof pathway is the most important route for the degradation of sugars, sometimes in combination with the non‐ oxidative part of the pentose phosphate pathway (e.g. xylose), which under optimal

conditions results in acetate, hydrogen and CO2 as products (212, 219). Hydrogen is formed by two hydrogenases which use NADH and reduced ferredoxin, formed during the glycolysis, as electron donor. The hydrogenases show a high sequence homology with the NADH‐ and the ferredoxin‐dependent hydrogenase described for Thermoanaerobacter tengcongensis (219). When certain environmental conditions make hydrogen production less favorable (e.g. high hydrogen pressure), the NADH can no longer be used by the NADH‐dependent hydrogenase, and a switch to lactate production (157). It has been described that the lactate dehydrogenase in C. saccharolyticus is regulated by the NAD/NADH ratio and the levels of PPi in the cell (204) also indicating that it plays a key role in the oxidization of NADH. C. saccharolyticus is also able to use rhamnose as sole source for carbon and energy (25, 166). Rhamnose is a deoxysugar which can be present in relatively high amounts in some fractions of biomass, like pectin and hemicelluloses. The pathway for rhamnose catabolism has recently been proposed, and yields 1,2‐propanediol in addition to the common fermentation end products (107, 196, 214). This additional pathway, which is comparable with the anaerobic pathway in Escherichia coli ((207), has important implications for the redox balance of the organism. Rhamnose is converted to rhamnulose by a rhamnose isomerase (Csac_0875), phosphorylated by a rhamnulose kinase (Csac_0989) and an aldolase (Csac 0865 and/or Csac_0866) catalyses the formation of L‐lactaldehyde and dihydroxyacetone phosphate (DHAP). The latter will then enter the Embden‐Meyerhof pathway, while L‐lactaldehyde is reduced to 1,2‐propanediol by a propanediol oxidoreducatase (Csac_0407) with NADH as cofactor. The stoichiometry of the rhamnose pathway dictates that all NADH which is formed during the step from glyceraldehyde 3‐ phosphate (GAP) to D‐1,3‐bisphosphoglycerate (1,3BPG), is used for the reduction of lactaldehyde to 1,2‐propanediol. This means that there is no NADH available for the hydrogenase reaction and that all hydrogen is derived from reduced ferredoxin. Here the rhamnose pathway is analyzed in more detail. The transcriptome of C. saccharolyticus on different substrates combined with product analysis and enzymatic data

97

Chapter 5

are described. The special redox situation during growth on rhamnose was further investigated by making use of CO, which is known to be a competitive inhibitor of NiFe‐ and Fe‐only hydrogenases. (45).

Material & methods

Growth of the microorganism

C. saccharolyticus DSM 8903 (also ATCC 43494) was purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen and was grown in a 2‐liter Applikon fermentor containing 1 liter modified DSMZ 640 medium containing NH4Cl (0.9 g/L), KH2PO4

(0.75 g/L), K2HPO4 (1.5 g/L), yeast extract (1 g/L), trace element solution SL‐10 (1 ml/L; (204), resazurine (1 mg/L), Glucose (4 g/L), MgCl2 (0.4 g/L), and cysteine‐HCL (0.5 g/L). The glucose‐ limited continuous culture was run at a dilution rate of 0.1 h‐1 at 70°C and the pH was

controlled at 7.0 . To achieve a low P(H2) condition, the growth medium was continuously sparged with N2 (4 l/h) at a stirring speed of 300 rpm. A high P(H2) condition was established, by only flushing the headspace of the fermentor and the flushing gas was pure

H2. Moreover, a lower stirring speed of 50 rpm was applied, to facilitate supersaturation of the medium with hydrogen.

Analysis of residual sugar and fermentation products.

During the chemostat run samples were taken, OD660 was determined. After centrifugation. residual sugars and organic acids in the supernatant were analyzed by HPLC containing a Varian Metacarb 67H 300mm column. Peak sizes were measured with an RI

detector and values were compared with standard curves for quantification.

Carbon balances

Carbon balances were calculated as a percentage of the total amount of moles of carbon present in glucose. The dry weight of the cells was calculated from the optical density by a calibration graph based on cell dry weight versus optical density (dry weight (mg/l) = OD660*422.5). The amount of biomass in moles of carbon could be calculated from these

values with the following formula for the composition of biomass, CH1.62O0.46

98

Rhamnose pathway in Caldicellulosiruptor saccharolyticus

N0.23S0.0052P0.0071 for C.saccharolyticus (220). CO2 concentrations were calculated assuming

that per mole of acetatate or ethanol 1 mole of CO2 is produced.

ATP and growth yield calculations

The ATP yield for different situation was calculated from the ATP obtained per mole of product. Oxidation of glucose to acetate will result in 6 ATP per glucose. Assuming that 2 ATP are consumed in the first part of the glycolysis and one is needed for transport of the substrate the overall yield is 3 ATP per glucose. As 2 acetate are formed, 1.5 mole of ATP is formed per acetate. As 1 of these ATP is formed during acetate kinase step, it means 0.5 ATP are formed per lactate and ethanol. In the case of growth on rhamnose, only 3 ATP are maximally formed. The overall ATP yield will be 1 ATP per rhamnose, as most likely 1ATP is used for transport and 1 is used during the rhamnose pathway. This would mean that per acetate only 1 ATP is formed and none for ethanol and lactate. Growth yield was calculated from the dry weight divided by the amount of moles of substrate.

Transcriptome analysis

For the transcriptome analysis of C. saccharolyticus grown on rhamnose, micro array data were reanalyzed, which were described previously. (70). ORFs that were differentially transcribed twofold or more and met the Bonferroni statistical criterion were considered to be up‐ or downregulated (27) .

Pyruvate and oxoglutarate dehydrogenase activity assay

Pyruvate and/or oxoglutarate dehydrogenase activity assays were performed by a standard NADH assay as well as an INT/PMS coupled assay adapted from Hinman et al (27) . For the standard NADH assay the reaction volume contained 50 mM Tris‐HCl buffer (pH 7.5), containing 12 mM NAD+, 0.2 mM thiamin pyrophosphate (TPP), 1,2 mM coenzyme A, 0.3 mM dithiothreitol, 1 mM MgCl2 and 10 mM pyruvate or oxaloacetate. Formation of NADH was followed at 350 nm at 50 °C. As these assays were performed with crude cell‐free extracts NADH may be directly used by lactate dehydrogenase. Therefore an INT/PMS coupled assay was also performed. The reaction mixture was as described above except that

99

Chapter 5

in these assays INT (2‐(4‐iodophenyl)‐3‐(4‐nitrophenyl)‐5‐phenyl tetrazolium chloride; 0.6 mM) and PMS (phenazinemethosulfate; 6.5 µM) were added and absorption was measured at 500 nm at several temperatures (25°C, 37°C and 50 °C). The same assays were also performed with 12 mM NADP+ instead of NAD+. Protein levels were determined according to Bradford (204) using bovine serum albumin as the standard.

Protein concentration

To determine cell lysis during growth, cells were harvested at different stages during the growth curve. Cells were centrifuged and supernatant was analyzed for protein concentrations according to Bradford (219) using bovine serum albumin as the standard.

Results and discussion

Growth on rhamnose

Batch cultures of C. saccharolyticus showed rapid growth on L‐rhamnose as sole carbon and energy source although doubling times on rhamnose were approximately 1.5 times longer than on glucose. Analysis of the medium at the end of growth showed the production of 1,2‐ propanediol, acetate and hydrogen in a 1:1:1 stoichiometry (Figure 5.1). This is in line with the pathway as proposed previously (204) based on genome information. Under these conditions no other products were observed (Table 5.1). This differs from the situation during growth on glucose, where in most batch experiments considerable amounts of lactate were produced. Formation of lactate acts as an alternative sink for NADH when the disposal via hydrogen formation becomes difficult (147). During rhamnose fermentation, formation of lactate is not necessary because all NADH is already used for the formation of 1,2‐propanediol. The observed stoichiometry between propanediol, acetate and hydrogen of 1:1:1 also demands that all hydrogen comes from reduced ferredoxin.

100

Rhamnose pathway in Caldicellulosiruptor saccharolyticus

Figure 5.1 Growth of C. saccharolyticus on rhamnose in batch.

Table 5.1 Product formation, growth rate and overall carbon balance for C. saccharolyticus. Substrates:

Glucose (19.4 mM); Rhamnose ( 23.5 mM); At CO conditions: N2 headspace replaced by CO.

H2 Acetate 1,2‐ Lactate Ethanol Maximum Biomass yield Maximum Carbon (mM) (mM) Propanediol (mM) (mM) OD660 (Dry weight growth rate balance (mM) (mg/l)/ mole of (h‐1) substrate)

Rhamnose 19.6 17.1 18.5 ND ND 1.01 18.2 0.22 92%

+N2

Rhamnose 6.3 12.5 18.4 1.5 2.5 0.61 11.0 0.10 77% +CO

Glucose 40.9 18.2 ND 11.1 ND 1.20 26.1 0.37 94%

+N2

Glucose 11.3 5.7 ND 25.0 3.2 0.49 10.7 0.03 81% +CO

101

Chapter 5

Growth on rhamnose results in a lower yield (18.2 (mg/l dry weight) / mole of substrate)) compared to glucose (26.1 (mg/l dry weight) / mole of substrate). The lower yield obtained on rhamnose can be explained by the estimated ATP yield. For rhamnose conditions the overall ATP yield is 17.1 moles and for glucose 32.85 moles based on the products formed. This would indicate that the ATP yield on rhamnose is 52% of the yield on glucose, as the growth yield is 69%. The higher optical density is unclear so far, but might indicate that rhamnose transport is performed in a different way without, or less ATP consumption.

The overall carbon balances (glucose 94%, rhamnose 92%) for growth under N2 conditions shows that there is a small gap, indicating that some carbon compound is missing.

Transcriptomics

The differential expression data that were obtained after growth on rhamnose compared to glucose, xylose or a mixture thereof, as described earlier (147), were further analyzed. During growth on rhamnose a clear upregulation was observed for all the genes that code for the components of the proposed pathway (Figure 5.2). Among the highest hits within the entire data sets of these experiments were the two proposed aldolases (Csac_0865, Csac_0866) which were upregulated more than 15‐fold for the rhamnose conditions when compared to glucose. The upregulation of the genes coding for this pathway is in big contrast to the downregulation of the genes coding for the enzymes of the Embden‐ Meyerhof pathway, like for instance the acetate kinase (Csac_2040) which is down‐ regulated almost five‐fold. It is interesting that the lactate dehydrogenase (Csac_1027) was downregulated two‐fold when grown on rhamnose which fits with the hypothesis that all the NADH is oxidized by the propanediol dehydrogenase for the formation of 1,2‐ propanediol. Production of lactate is no longer a necessary switch under unfavorable hydrogen producing conditions (e.g. high hydrogen pressure, stationary phase). The downregulation of one gene (Csac_1864) coding for HydA, which is part of the NADH‐ dependent hydrogenase and the absence of an effect on the expression level of the ferredoxin‐dependent hydrogenase, is in line with the role of NADH in 1,2‐propanediol production, and its decreased role in hydrogen formation. It is, however, questionable when only one subunit of an obvious gene cluster is differentially expressed. Separate expression

102

Rhamnose pathway in Caldicellulosiruptor saccharolyticus of hydA would require that promoter elements are present in the intergenic region between Csac_1863 and Csac_1864. This is, however not the case; only a RBS can be identified in the 17 bp intergenic region. These data suggest that an individual expression of hydA is rather unlikely.

Figure 5.2 Transcriptome data for genes coding for the rhamnose pathway and the Embden Meyerhof pathway. Four digit numbers indicate locus tag of the different genes in C. saccharolyticus. 2 digit numbers indicate the log2 ratios for the up‐ or downregulation of these genes. Green lines indicate upregulation of the responsible genes, red lines indicate downregulation, orange line indicates no clear up‐ or downregulation.

103

Chapter 5

The two earlier mentioned aldolases (Csac_0865 and Csac_0866) as well as the proposed rhamnose isomerase (Csac_0876) are located within the same region on the genome (Figure 5.3). This entire region is clearly upregulated but the function of all the clustered components of this region is not clear. In addition to the aforementioned enzymes of the rhamnose pathway, also 3 components of an annotated alpha‐keto acid dehydrogenase complex were highly upregulated. This type of dehydrogenase complexes, with pyruvate dehydrogenase (PDH) as best studied example, generally consists of three subunits, indicated by E1, E2, and E3. The current gene cluster in C. saccharolyticus (Csac_0872,Csac_0873, Csac_0874) also shows a clear homology to the canonical E1/E2/E3 composition (100). This family of dehydrogenases consists of related complexes, which can function at different points in the carbon metabolism (169). The presence of genes coding for a lipoate‐protein ligase B (Csac_0870) and a lipoic acid synthase (Csac_0870) in this cluster can be explained by the fact that lipoic acid is an important component of the complex (121). The substrate specificity of these dehydrogenases is difficult to predict by annotation, but the most promising candidates in C. saccharolyticus are pyruvate (glycolysis) and 2‐oxoglutarate (citric acid cycle). Although others were able to predict specificities by specific sequences of the E2 sequence, for C. saccharolyticus we were not able to show this by sequence comparison (87). Pyruvate dehydrogenase is commonly not observed in strictly anaerobic microorganisms, as its reaction produces reduced pyridine nucleotides (NADH and NADPH) which might cause problems to maintain the redox balance. Strict anaerobes, like C. saccharolyticus, use instead the oxygen‐sensitive pyruvate: ferredoxin oxidoreductase

for the oxidative decarboxylation of pyruvate to acetyl‐CoA and CO2. The use of an NADH‐ dependent keto‐acid dehydrogenase instead of a ferredoxin‐dependent oxidoreductase is unusual finding in a strict anaerobes. Although, under certain conditions it has been shown that PDH plays a role in anaerobic pyruvate metabolism in Klebsiella pneumoniae (33), Escherichia coli (187), Streptococcus mutans (219) and Enterococcus faecalis (92). Unfortunately, various attempts to demonstrate pyruvate dehydrogenase activity (or 2‐ oxoglutarate dehydrogenase activity) in anoxically prepared cell‐free extracts of cells grown on rhamnose, were not successful. Thus, a physiological role for the keto‐acid dehydrogenase could not be established.

104

Rhamnose pathway in Caldicellulosiruptor saccharolyticus

Figure 5.3. Upregulation of a gene cluster (Csac_0865‐Csac_0876) of C. saccharolyticus when grown rhamnose. Arrows indicate genes. Four digit numbers indicate locus tags of the different genes in C. saccharolyticus. 2 digit numbers indicate the log2 ratios for the up‐ or downregulation of these genes when grown on rhamnose compared to glucose.

Effect of carbon monoxide on product formation.

When C. saccharolyticus was grown on rhamnose with CO in the headspace clear differences were found in the fermentation pattern (Figure 5.4). Hydrogen production as well as optical density were lower, but only a small amount of ethanol and lactate was produced. As lactate dehydrogenase is probably using NADH as cofactor, it seems that a small amount of NADH is used for lactate production. As most of the NADH is used for 1.2‐ propanediol production, the lower hydrogen concentrations indicate, an inhibiting effect of CO on the ferredoxin dependent hydrogenase. This would also mean that most likely the ethanol production is coming from the disposal of reduced ferredoxin. Ethanol production in C. saccharolyticus is most likely performed by one or more alcohol dehydrogenases (ADHs) using NADPH, as has been shown by activity measurements (219). This would mean that a ferredoxin:NAD(P) oxidoreductase is needed to transfer the electrons from ferredoxin to NAD(P). But these oxidoreductase have been often proposed but never characterized. The 1,2‐propanediol production was comparable with that under normal growth conditions (Table 5.1). These results are in line with the envisaged pathway and related reductant disposal, meaning that all NADH is recycled by a propanediol dehydrogenase instead of the

105

Chapter 5

NADH‐dependent hydrogenase, which is used when grown on most other sugars. When C. saccharolyticus is growing on glucose in the presence of CO, lactate production increased dramatically. Furthermore, the growth rate on rhamnose was half as high compared to the conditions without CO. For glucose this effect was even more obvious as the growth rate was ~10‐fold lower compared to the conditions without CO. The data show that the disposal of reducing equivalents in the presence of CO, is more difficult during growth on glucose than on rhamnose. In addition to an effect of CO on the product pattern, we also observed that in the presence of CO the overall carbon balance was much lower than in the absence. This suggests that also here a reduced end product, whose formation is even more triggered by CO, is missing. Unfortunately this missing product could not be explained by common fermentation end products as measured by HPLC using an organic acid column. As alanine was shown to be a sink for reducing equivalents in Pyrococcus furiosus (188) the medium was also analyzed on an amino acid column. However, no significant amounts of alanine or other amino acids were detected.

Figure 5.4 Growth of C.saccharolyticus on rhamnose in batch with CO as gas phase.

To explain the gap in the carbon balance the experiment was repeated and during time samples were taken to analyze the amount of proteins in the supernatant after removal of

106

Rhamnose pathway in Caldicellulosiruptor saccharolyticus the cells, the product pattern was comparable although highest measured optical density for growth without CO was lower (0.81 versus 1.0) indicating that most likely the optical density was already decreasing at t=25 h. These protein concentrations were compared with the OD of the culture (Figure 5.5). It was found that the maximal OD decreases when CO is present (0.8 versus 0.6). Furthermore, it could also be observed that after ~25 h cells started to lyse and that this process was much faster when CO was present. In the growth curve on rhamnose without CO, 14 mg/l of proteins are measured at the peak of the growth (t=25 h, OD660=0.81, dry weight 342.2 mg/ml). If assumed that intact cells consist for 50% of

proteins, this would mean that at this point only 28 mg/l cells are lysed. Based on the biomass composition, this would mean that these lysed cells would contribute for less than 2%, to the missing carbon in the carbon balances. As these calculations are based on assumptions and a calibration curve for cells grown under non stress conditions, further research is needed.

Figure 5.5 Optical density and protein concentration in supernatant during growth of C. saccharolyticus on rhamnose, with and without CO as gas in the headspace.

Conclusions

In this study, we examined the metabolism of C. saccharolyticus with respect to growth on rhamnose. We could that a novel pathway is used resulting not only in common products

like acetate, hydrogen and CO2, but also 1,2‐ propanediol. This means that NADH which is

107

Chapter 5

normally used by an NADH‐dependent hydrogenase to form hydrogen, will now be oxidized by a propanediol dehydrogenase The lower hydrogen and acetate production is in line with the predicted pathway. By analyzing the transcriptome data all the proposed components of the rhamnose pathway were clearly upregulated when grown on rhamnose compared to growth on glucose or xylose. The components of the Emden‐Meyerhof pathway as well as the NADH‐dependent hydrogenase were mostly downregulated, also indicating that NADH is not used for hydrogen production when grown on rhamnose. Another interesting result of the transcriptome analysis was the clear upregulation of a big cluster of genes (Csac_0865‐ Csac_0876) coding for components of the rhamnose pathway as well as a complete alpha‐ keto acid dehydrogenase complex. We were, however, not able to show the specificity of this complex by activity assays. The special redox state of the rhamnose pathway prompted us to investigate the effect of CO. CO did not affect the NADH‐dependent formation of 1,2‐ propanediol and did not lead to lactate formation. This shows that there is no need for an alternative way to oxidize NADH as all is used to produce 1.2 propanediol. However, we did observe an inhibition of hydrogen and acetate production in the presence of CO. As all NADH was consumed by the 1,2‐propandiol fermentation, the lowered formation of hydrogen must have been caused by inhibition of the ferredoxin‐dependent hydrogenase. Unfortunately we could not explain the gap in the carbon balance. It seems that even more reduced end product is missing under CO conditions. This unknown reduced product is most likely produced from reduced ferredoxin, possibly via NADP as intermediate carrier. It has been shown that NADPH is the preferred substrate for the alcohol dehydrogenase, and this would indicate that a ferredoxin: NADP oxidoreductase must be present (37). So far this oxidoreductase could not be detected in C. saccharolyticus, although also for comparable organisms there are clear indications of its existence (157). Overall we were able to show a new pathway for C. saccharolyticus with a clear effect on the pathways for reductant disposal. Acknowledgements M.R.A.V, S.M.W.K. and A.J.M.S. acknowledge support from the Commission of the European Communities, Sixth Framework Programme, Priority 6, Sustainable Energy Systems (019825 HYVOLUTION)

108

Chapter 6

The effect of hydrogen pressure on the metabolism of the thermophilic bacterium Caldicellulosiruptor saccharolyticus

Marcel R. A. Verhaart, Amy L. VanFossen, Robert M. Kelly, Alfons, J.M. Stams, Servé M. W. Kengen

Manuscript in preparation Chapter 6

Abstract

Caldicellulosiruptor saccharolyticus is an extremely thermophilic grampositive bacterium which is able to ferment a large variety of substrates to acetate, CO2 and hydrogen. Therefore, it is a good candidate to study fermentative hydrogen production. To understand the pathways involved in biohydrogen production it is important to know the mechanisms to dispose of reducing equivalents. C. saccharolyticus was exposed to a low and a high

hydrogen partial pressure (PH2) in a chemostat, resulting in a clear shift from acetate to lactate and ethanol production. Transcriptome data experiment showed upregulation of several genes involved in the disposal of NADH and reduced ferredoxin, including both hydrogenases, a lactate dehydrogenase and an alcohol dehydrogenase. Upregulation of a cluster coding for a pyruvate carboxylase is so far not understood. A possible explanation could be that the reductive citric acid cycle might act as a sink for reducing equivalents as well. Another option could be that malate is used to transfer electrons from NADH to NADPH via a malic enzym and a malate dehydrogenase. The data described here indicate that the P(H2) dependent change in fermentation pattern in C. saccharolyticus is also regulated at gene level. Various downregulated genes had no direct redox related function,

but may represent redundant energy consuming metabolic steps at high P(H2) condition.

Introduction:

With the decline of fossil fuels there is an obvious need for clean biofuels. Biohydrogen is one of the most promising alternatives as it only yields water during combustion. One way to produce hydrogen is by dark fermentation. Caldicellulosiruptor saccharolyticus was selected as a model organism for thermophilic anaerobic fermentative hydrogen production (204, 209). C. saccharolyticus is an extremely thermophilic (growth temperature optimum 70‐75°C), gram‐positive, fermentative anaerobe, initially isolated from wood in the flow of a thermal spring in New Zealand (91, 212). The advantages of this organism are its capacity to utilize a wide variety of substrates, varying from simple (e.g. glucose, xylose) to more complex carbohydrates, as well as its capacity to use C5 and C6 sugars simultaneously (188, 204). In addition, this organism is able to produce close to 4 moles of hydrogen per mole of glucose, which is the theoretical limit (206). To make the application of hydrogen

110

Effect of hydrogen pressure on the metabolism production by thermophilic micro‐organisms like C. saccharolyticus feasible, more knowledge is necessary on the metabolism of such bacteria. The Embden‐Meyerhof

pathway is used by C. saccharolyticus to degrade the substrates to acetate, CO2 and hydrogen. In this pathway NADH is formed in the step from glyceraldehyde‐3‐phosphate to 1,3‐diphoshoglycerate catalyzed by the triose phosphate dehydrogenase. Reduced ferredoxin is produced by pyruvate:ferredoxin oxidoreductase converting pyruvate to acetyl

CoA and CO2. This results in a of 1:1 ratio of reducing equivalents available from NADH and reduced ferredoxin. As the production of acetate is coupled to the production of ATP, reductant disposal via hydrogen is energetically most advantageous for the organism. However, these hydrogen yielding reactions are thermodynamically rather unfavorable, especially from NADH. At lower temperatures and higher hydrogen partial pressures, this leads to the production of other reduced end products, like lactate or ethanol (207). Genome annotation showed the presence of a NADH‐ and a ferredoxin‐dependent hydrogenase comparable to the situation in Thermoanaerobacter tengcongensis (45). Although, some research has been done on the effect of the hydrogen partial pressure on the fermentation pattern, nothing is known about the effect on gene expression level (204, 209). In this study, C. saccharolyticus was grown in a glucose‐limited chemostat and conditions with a high and a low hydrogen partial pressure were compared. The main focus of this research was to investigate the effect of the hydrogen pressure on the pathways for reductant disposal, and their regulation on transcript level.

Material and methods

Growth of the microorganism C.saccharolyticus DSM 8903 (also ATCC 43494) was purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen and was grown in a 2‐liter Applikon fermentor with 1 liter modified DSMZ 640 medium containing NH4Cl (0.9 g/L), KH2PO4 (0.75 g/L), K2HPO4 (1.5 g/L), yeast extract (1 g/L), trace element solution SL‐10 (1 ml/L; (204), resazurine

(1 mg/L), Glucose (4 g/L), MgCl2 (0.4 g/L), and cysteine‐HCL (0.5 g/L).The glucose‐limited continuous culture was run at a dilution rate of 0.1 h‐1 at 70°C and the pH was controlled at

7.0 . To achieve a low P(H2) condition, the growth medium was continuously sparged with

111

Chapter 6

N2 (4 l/h) at a stirring speed of 300 rpm. A high P(H2) condition was established by flushing

theheadspace with pure H2 at a lower stirring speed of 50 rpm. The lower stirring speed was used to still secure mixing and avoid supersaturation.

Analysis of residual sugar and fermentation products.

During the chemostat run samples were taken, OD660 was determined and after centrifugation, residual sugars and organic acids in the supernatant were analyzed by HPLC containing a Varian Metacarb 67H 300mm column. Peak sizes were measured with an RI

detector and values were compared with standard curves for quantification.

Carbon balances

Carbon balances were calculated as a percentage of the total amount of moles of carbon present from glucose. The dry weight of the cells was calculated from the optical density by a calibration graph based on cell dry weight versus optical density (dry weight (mg/l)=OD660*422.5). The amount of biomass in moles of carbon could be calculated from these values with the composition of biomass,

CH1.62O0.46 N0.23S0.0052P0.0071 in C.saccharolyticus(155). CO2 concentrations were calculated

assuming that per mole of acetatate or ethanol 1 mole of CO2 is produced.

ATP and growth yield calculations The ATP yield was calculated from the ATP obtained per mole of product. Oxidation of glucose to acetate will result in 6 ATP per glucose. Assuming that 2 ATP are consumed in the first part of the glycolysis and one is needed for substrate transport, via an ABC transporter, the overall yield is 3 ATP per glucose. As 2 acetate are formed, 1.5 mole of ATP is formed per acetate. As 1 of these ATP is formed during acetate kinase step, it means 0.5 ATP are formed per lactate or ethanol.

RNA isolation Cells were harvested by centrifugation and rapid cooling to 4°C and then stored at ‐80°C. Total RNA was isolated using a modified Trizol (Invitrogen) protocol in combination with an RNAeasy kit (Qiagen) as described before (220). Quality was determined with an

112

Effect of hydrogen pressure on the metabolism

Experion bioanalyzer (Bio‐Rad), and cDNA was constructed with Superscript III reverse transcriptase (Invitrogen) random primers (Invitrogen), and the incorporation of 5‐(3‐ aminoallyl)‐2'‐deoxyuridine‐5'‐triphosphate (Ambion, Austin, TX).

Microarray protocols A spotted whole‐genome C. saccharolyticus microarray was developed from 60‐mer

oligonucleotides based on at least 2,679 ORFs, as discussed previously (220). cDNA was labeled with either Cy3 or Cy5 dye (GE Healthcare, Little Chalfont, United Kingdom) and hybridized to Corning Ultra‐Gap II slides (Corning, Acton, MA). Microarray slides were scanned by using a Packard BioChip Scanarray 4000 scanner. ScanArray Express (v2.1.8; Perkin‐Elmer, Waltham, MA) was used to quantitate signal intensities, prior to analysis with JMP Genomics 3.0 (SAS, Cary, NC), as described previously (180), using a mixed‐effects analysis of variance model (219). ORFs that were differentially transcribed twofold or more and met the Bonferroni statistical criterion, were considered to be up‐ or downregulated (204) .

Results and discussion

Effect of the hydrogen pressure on the growth pattern of C. saccharolyticus

When C. saccharolyticus is grown in continuous culture at a low hydrogen partial pressure,

all glucose is converted into acetate, CO2 and hydrogen, resulting in an OD660 of around 1.2 (Figure 6.1).When switched to a high hydrogen pressure situation a clear shift takes place.

The OD660 drops to about 0.65, acetate production declines and ethanol and lactate are produced. As the dephosphorylation of acetyl phosphate to acetate is yielding an extra ATP, the lower acetate concentration is coupled to a decrease of the optical density. If the ATP

yield is calculated for both situations (low P(H2)= 51 ATP, High P(H2)=36.6 ATP) a decrease of 30% in the optical density can be expected. As the decrease of optical density is even

around 40% in this experiment this could indicate that at high P(H2) more ATP is needed for maintenance.

Switching back to the original situation with a low P(H2), C. saccharolyticus recovered to

produce acetate CO2 and hydrogen, whereas ethanol and lactate were not produced

113

Chapter 6

anymore. At low P(H2) the reducing equivalents NADH and reduced ferredoxin are probably used by the two hydrogenases to produce hydrogen. On the other hand the annotated NADH dependent hydrogenase shows homology to a recently described bifurcating hydrogenase in Thermotoga maritima (219). In this case reducing equivalents of reduced ferredoxin and NADH are used in a 1:1 ratio by one hydrogenase. The thermodynamically unfavorable production of hydrogen from NADH is in this way pushed by the use of reduced ferredoxin. For C. saccharolyticus so far the bifurcating function has not been shown in this

organism. When the P(H2) increases these reactions to hydrogen become less favourable and will result in a shift in the disposal of the reductants. NADH is most likely used to produce lactate. It has been described that the lactate dehydrogenase in C. saccharolyticus is regulated by the NAD/NADH ratio and the levels of PPi in the cell (84) also indicating that it plays a key role in the oxidization of NADH. In the genome of C. saccharolyticus only one candidate gene was indentified for a lactate dehydrogenase (204, 209). On the other hand the produced ethanol could originate from reduced ferredoxin as most

likely also the ferredoxin‐dependent hydrogenase activity is affected by the high P(H2). Enzyme assays have shown that NADPH is the preferred substrate for alcohol dehydrogenase activity in cell extract of C. saccharolyticus (220). In the case NADPH is used for ethanol production, and is formed by oxidation of reduced ferredoxin, an oxidoreductase is necessary to transfer the electrons from reduced ferredoxin to NADP. These ferredoxin‐NAD(P)H oxidoreductases have been proposed for decades and enzyme activities could be measured, but a molecular characterization has never been described for strict anaerobes (30, 188).

114

Effect of hydrogen pressure on the metabolism

Figure 6.1 C. saccharolyticus growing in continuous culture on glucose under low P(H2) and high P(H2). N2 = medium flushed at 4 l/h N2, Stirring = 300 rpm. H2 = headspace flushed with 4 l/h H2, stirring 50 rpm. Arrows indicate sample points for transcriptome analysis.

From the calculated carbon balances (Figure 6.2) for the different sample times, it can be concluded that there is a reduced end product missing. Especially the total amount of carbon found back in the measured products is dropping for the high hydrogen pressure situation. This indicates that formation of this missing product is increasing or another reduced product is missing in our C‐balance. However, no unidentified peaks were observed in the HPLC systems used here to detect organic acids, alcohols and amino acids. Analysis of the spend medium also did not reveal elevated protein levels, which indicates that there was no increased cell lysis. Further research is needed to clarify the gap in the carbon balance. As a continuously flushed system is used for this experiment a loss of ethanol by evaporation cannot be excluded.

115

Chapter 6

Figure 6.2. Carbon balance of C. saccharolyticus growing in continuous culture on glucose at low P(H2) and high

P(H2). Sample points are the same as shown in figure 6.1. Products are shown in percentages of the carbon

input (glucose, 20 mM = 120 mM C) N2 = medium flushed with 4 l/h N2, Stirring= 300 rpm. H2 = headspace

flushed with 4 l/h H2, stirring 50 rpm.

Transcriptomics

To analyze the effect of the hydrogen pressure on the level of gene regulation, cells were harvested at different time points (arrows in figure 6.1) during the experiment. As an additional biological duplicate for low P(H2) conditions, cells were harvested from another chemostat run under identical conditions and with a comparable product pattern. Unfortunately the differences in gene expression are not as large as what was seen in the controlled batch experiments (30, 31), and a very low number of genes were considered to be up‐ or downregulated according to the Bonferroni statistical criterion (21). Because of this reason we also took the genetic region into account for which all genes were upregulated, although the 2‐fold upregulation was not met. Expression data for the genes involved in glycolysis showed almost no differences (log 2 ratio ‐0.5 to 0.5) except for the enolase (Csac_1950) and the acetate kinase (Csac_2040) (Figure 6.3). In agreement with the

116

Effect of hydrogen pressure on the metabolism observed product pattern, an alcohol dehydrogenase (Csac_0407) and the only annotated lactate dehydrogenase (Csac_1027) were upregulated. This alcohol dehydrogenase shows high homology with one of the alcohol dehydrogenase described for T. tengcongensis (TTE0696)(204) and Thermoanaerobacter ethanolicus. This so‐called primary alcohol dehydrogenase in Thermoanaerobacter ethanolicus, was shown to use both NADH and NADPH (124). Interestingly, the genomes of C. saccharolyticus and T. tengcongenis did not reveal any obvious aldehyde dehydrogenase (ALDH) . For T. tengcongensis it was suggested that another alcohol dehydrogenase (TTE0695) would be bifunctional, catalyzing both ADH and ALDH activity. The closest homolog to this so called secondary dehydrogenase in C.

saccharolyticus (Csac_0395), however, was not upregulated at high P(H2). So, the ALDH activity in this organism still needs further research.

Figure 6.3 Expression profile for genes involved in glycolyis and reductant disposal. (orange arrows: no clear

up‐ or downregulation (log2 ratio ‐0.5 to 0.5), green arrows show upregulation. 4 digit numbers indicate locus tag in C. saccharolyticus. 2 number digits show log2 ratios for H2 vs N2 conditions.

117

Chapter 6

The pyruvate carboxylase (Csac_2482‐2485) that catalyzes the reversible carboxylation of pyruvate to form oxaloacetate (OAA) is also upregulated. Although the reason for this upregulation is still unclear, two possible explanations are suggested here. Oxaloacetate is an intermediate of the citric acid cycle. The incomplete citric acid cycle could act as a sink for reducing equivalents, as was shown for instance for Escherichia coli which produced succinate (99, 225, 226). Succinate formation was not observed and the other necessary genes involved in the reductive citric acid cycle were not present in the genome or not upregulated in this experiment. So far only Willquist et al (personal communication) were able to show trace amounts of succinate by NMR. During the annotation of the genome no genes were identified coding for a fumarate reductase, as was the case so far for any of the related Clostridia (209). Also no clear candidate could be identified for a malate dehydrogenase. Another explanation could be that a malate dehydrogenase is used in combination with a malic enzyme to transfer electrons from NADH to NADPH. Malate dehydrogenases catalyze the reversible reaction from oxaloacetate to malate using NADH or NADP (52). Malic enzyme has been described to catalyze the decarboxylation of malate to pyruvate resulting in NADPH or NADH (98, 229). The gene with highest homology to a malic

enzyme in C. saccharolyticus is Csac_2059 and was not upregulated at high P(H2) conditions. Besides the genes mentioned before, also both hydrogenases, were upregulated at the high

P(H2) condition. At first sight this may seem counter intuitive, as hydrogen formation is expected to be lower under high hydrogen pressure. However, the observed upregulation of all genes involved in reductant disposal (dehydrogenases and hydrogenases) suggests that a general regulatory response takes place, as reaction to the elevated hydrogen pressure. Whether this is a result of elevated levels of reduced electron carriers like NADH or reduced ferredoxin remains is not clear. Besides the genes involved in the glycolysis, also other genes were up‐ (Table 6.1) or downregulated (Table 6.2). As the differences in gene expression were not that large, only clusters of genes showing a similar profile are discussed here. The first cluster of genes (Csac_0261‐Csac_0265) which is upregulated, is annotated as an oligopeptide ABC transporter. This cluster was also upregulated for fructose in C. saccharolyticus and because of this, predicted to have a function in sucrose transportation (98). As glucose is also a

118

Effect of hydrogen pressure on the metabolism constituent of sucrose,this might explain this upregulation. The next upregulated cluster (Csac_1451‐1454) contains three hypothetical proteins and RadC. This RadC has been characterized in Escherichia coli and mutant studies showed that it is involved in DNA repair

(122). The reason for upregulation of this cluster under high P(H2) remains unclear. Csac_1601 and Csac_1602 are also strongly upregulated and are coding for the fatty acid/phospholipid synthesis protein plsX and its regulator fapR. PlsX is a key enzyme that coordinates the production of fatty acids and membrane phospholipids (159) and was proposed to be a phosphate acyltransferase that catalyzes the formation of acyl phosphate from the acyl‐acyl carrier protein (ACP). So far, the specific biochemical function needs further research (209). A cluster coding for such an ACP is located on the genome next to the plsX and fapR genes (Csac_1603‐Csac_1608) and is also upregulated. ACP has a key role in fatty acid synthesis, as it has been shown to interact with 12 or more proteins involved in this process (204). The upregulation of this total gene cluster, containing plsX and ACP, might indicate that C. saccharolyticus is in need of extra fatty acids and/or phospholipids in case of stress.The next upregulated cluster is involved in ferrous ion uptake, which corresponds with the upregulation of both hydrogenases. Ferrous ion is an important component of both complexes. The upregulated propionyl‐CoA carboxylase and the pyruvate carboxylase (Csac_1881‐1883) also have a role in oxaloacetate production which has been described earlier as an alternative sink of reducing equivalents. The last upregulated cluster discussed here, is an ABC transporter for cobalt, situated next to a gene involved in cobalamin synthesis (Csac_2682‐Csac_2685). Cobalamin or vitamin B12, is a coenzyme which is used in bacteria in a wide spectrum of metabolic processes (209). Because of this diverse role the upregulation of this cluster remains unexplained. In addition to the upregulated clusters, also some clusters were clearly downregulated. First of all there were multiple clusters coding for annotated ABC transporter complexes (Csac_0140‐Csac_0145)(Csac_1631‐1633)(Csac_2417‐2419)(Csac_2491‐2493). These ABC transporters were most likely involved in transport of substrates which are not available during this experiment. Synthesis of these transporters can be energy consuming and the downregulation of these clusters could be expected under stress conditions. In general, identifying substrate specificity is difficult, although for some of these transporters

119

Chapter 6

substrate specificity was predicted by analyzing transcriptome data on specific sugars and comparison with homologs in other organisms (119, 181). For C. saccharolyticus at least 177 putative ABC transporter genes were identified in the genome (38, 203). Based on the similarity with proteins present in Thermotoga maritima and prediction by TransportDB, 24 putative carbohydrate transporters were suggested. The first two clusters (Csac_0140‐ Csac_0145) (Csac_1631‐1633) were not part of this list, and no substrate specificity was predicted for these transporters. Whether these two clusters code for functional ABC transporters needs further research and their role in C. saccharolyticus remain unclear. The next clusters, were predicted to have xylooligosaccharides (Csac_2417‐2419) and xylose, glucose or fructose (Csac_2491‐2493) as substrates (38). From the data here, it can be concluded that the latter cluster is definitely not coding for the main glucose transporter, as all the glucose was consumed even under high hydrogen pressure conditions. Besides these transporters a gene cluster (Csac_1142‐1146) was downregulated containing an acetolactate synthase and a 2‐isopropylmalate synthase. Acetolactate synthase is used to form acetolactate from pyruvate and 2‐isopropylmalate synthase uses acetyl‐CoA to form isopropylmalate. The downregulation of these two enzymes might be explained by the fact that both enzymes are involved in amino acid synthesis, which might be less needed under

high P(H2) conditions. Three genes (Csac_1340,Csac_1339 and Csac_0425) coding for components of the fucose catabolic pathway were downregulated. This is in line with the expectation that under stress conditions, some of the unnecessary pathways are downregulated.. The next downregulated cluster (Csac_2051‐2054) to be discussed is proposed to code for a indolepyruvate ferredoxin oxidoreductase. This enzyme is described to play a role in archaea in peptide fermentation. It catalyzes the oxidative decarboxylation of aryl pyruvates, which are generated by the transamination of aromatic amino acids, to the corresponding aryl acetyl‐CoA (144). The reason for the downregulation in under these conditions in C. saccharolyticus remains unclear.

120

Effect of hydrogen pressure on the metabolism

Table 6.1 Upregulated gene clusters Locus tag Annotation Log 2 ratio Csac0260 transcriptional regulator, LacI family 0.53 Csac0261 homolog oligopeptide‐binding protein appA 0.78 Csac0262 homolog oligopeptide transporter system, permease 1.00 appB Csac0263 homolog oligopeptide transport system permease appC 0.63 Csac0264 oligopeptide transport ATP‐binding protein appD 0.69 Csac0265 oligopeptide transport ATP‐binding protein appF 0.48 Csac0266 hypothetical protein 0.54 Csac1451 hypothetical protein 1.78 Csac1452 hypothetical protein 1.00 Csac1453 DNA repair protein RadC 1.67 Csac1454 hypothetical protein 0.83 Csac1601 Transcription factor fapR (Fatty acid and phospholipid 1.04 biosynthesis regulator). Csac1602 Fatty acid/phospholipid synthesis protein plsX 0.74 Csac1603 3‐oxoacyl‐[acyl‐carrier‐protein] synthase (Beta‐ketoacyl‐ 0.19 ACP synthase III) (KAS III). Csac1604 Malonyl CoA‐acyl carrier protein transacylase 0.96 Csac1605 3‐oxoacyl‐[acyl‐carrier‐protein] reductase (3‐ketoacyl‐ 0.41 acyl carrier protein reductase) Csac1606 Acyl carrier protein (ACP) 0.92 Csac1607 3‐oxoacyl‐[acyl‐carrier‐protein] synthase 2 (Beta‐ 1.57 ketoacyl‐ACPsynthase II) Csac1608 Ribonuclease III (RNase III) 0.88 Csac1737 KaiC protein 0.98 Csac1738 putative circadian clock protein, KaiC 0.51 Csac1739 ferrous ion uptake protein A, domain 1.64 Csac1740 ferrous ion uptake protein B, homolog 0.76 Csac1741 ferric uptake regulation protein 0.36 Csac1881 Propionyl‐CoA carboxylase beta chain 0.68 Csac1882 putative biotin‐dependent acetyl/propionyl‐CoA 0.90 carboxylase alpha subunit Csac1883 Pyruvate/oxaloacetate carboxytransferase 1.07 Csac2682 ABC transporter, ATPase 0.81 Csac2683 cobalt transport protein 0.56 Csac2684 putative cobalt transport protein 1.68 Csac2685 cobalamin biosynthesis protein CbiM 2.02

121

Chapter 6

Table 6.2 downregulated gene clusters

Locus tag Annotation Log 2 ratio Csac0140 putative HTH‐type transcriptional regulator ‐0.95 Csac0141 ABC transporter, ATP‐binding protein ‐1.32 Csac0142 hypothetical protein ‐1.39 Csac0143 hypothetical protein ‐1.75 Csac0144 ABC transporter, ATP‐binding protein ‐1.50 Csac0145 putative ABC transporter permease ‐0.75 Csac1142 acetolactate synthase, large subunit, biosynthetic ‐1.19 type Csac1143 acetolactate synthase, small subunit ‐0.95 Csac1144 Ketol‐acid reductoisomerase ‐0.74 Csac1145 2‐isopropylmalate synthase ‐0.48 Csac1146 2‐isopropylmalate synthase/homocitrate synthase ‐0.93 family protein Csac0425 L‐fuculose phosphate aldolase homolog ‐1.10 Csac1339 L‐fucose isomerase ‐0.53 Csac1340 Alpha‐L‐fucosidase (GH fam 29) ‐0.87 Csac1631 ABC‐type probable sulfate transporter, substrate‐ ‐0.78 binding protein Csac1632 ABC transporter domain protein ‐0.13 Csac1633 putative binding‐protein‐dependent transporter, ‐0.86 permease Csac2051 Coenzyme F390 synthetase II ‐0.81 Csac2052 Indolepyruvate ferredoxin oxidoreductase, subunit ‐0.56 beta Csac2053 indolepyruvate ferredoxin oxidoreductase, chain ‐1.30 alpha Csac2054 Coenzyme F390 synthetase ‐1.08 Csac2417 putative multiple sugar transporter, permease ‐1.18 Csac2418 probable ABC transporter permease (YesP) ‐0.87 Csac2419 similar to ABC‐type sugar transporter, sugar‐ ‐0.63 binding protein Csac2491 ABC transporter permease ‐0.80 Csac2492 ABC transporter permease ‐1.01 Csac2493 putative ABC transporter ‐1.74

122

Effect of hydrogen pressure on the metabolism

Conclusions

With the research described here we show that under high hydrogen conditions in a chemostat, C. saccharolyticus partly shifts its metabolism. As hydrogen production becomes

less favorable at this high P(H2), the reducing equivalents might accumulate and the organism is in need of alternative reducing equivalent sinks. As expected this also resulted in a lower optical density as less ATP is produced. The transcriptome data obtained from this chemostat experiment showed that both hydrogenases, lactate dehydrogenase and an

alocohol dehydrogenase were upregulated under high P(H2) conditions. This suggests that C.

saccharolyticus on high P(H2) reacts by upregulation of genes involved in multiple reactions for the disposal of reducing equivalents. The downregulation of several ABC transporters and the almost complete pathway to metabolize fuculose indicated that C. saccharolyticus is able to switch off unnecessary pathways by regulation on gene expression level. As lactate

dehydrogenase has been shown to have high activities on NADH, under high P(H2), most

likely NADH is used to produce lactate. At high P(H2) it seems that next to NADH, also all reduced ferredoxin cannot be used for hydrogen production. We propose that this reduced ferredoxin is indirectly used via NADPH for ethanol production. The function of the upregulation of cluster coding for a pyruvate carboxylase remains unclear, but might also contribute to the disposal of reducing equivalents. Overall this research contributes to the understanding of the different mechanisms to dispose of reducing equivalents, not only in C. saccharolyticus but also for anaerobic fermentative bacteria in general.

Acknowledgements

M.R.A.V and S.M.W.K. acknowledge support from the Commission of the European Communities, Sixth Framework Programme, Priority 6, Sustainable Energy Systems (019825 HYVOLUTION). R.M.K. acknowledges support from the US National Science Foundation (CBT0617272) and the Bioenergy Science Center (BESC), a U.S. DOE Bioenergy Research Center supported by the Office of Biological and Environmental Research. A.L.V. acknowledges support from a US Department of Education GAANN Fellowship. Microarray fabrication was financially supported in part by Danisco (Genencor International, Inc., Palo Alto, CA).

123

Chapter 6

124

Chapter 7

General discussion Chapter 7

General discussion

The aim of the research described in this thesis was to better understand and improve hydrogen production by the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus. This was achieved by a combination of bioinformatics and transcriptomics approaches and analysis of products and enzyme activities. This resulted in the identification of several sugar transport systems, the elucidation of some of the major catabolic pathways, their regulation as well as the mechanisms for reductant disposal. In this general discussion the research will be placed in a more general context, including future perspectives for hydrogen research.

Biohydrogen

So far, most hydrogen is produced by steam reforming of natural gas or gasification of coal

or biomass in combination with CO2 storage. Only the latter method can be called sustainable. An alternative method to produce sustainable biohydrogen is fermentative hydrogen production. To make this method profitable and sustainable, one should not consider the dark fermentation on its own. During the dark fermentation only part of the energy available in the biomass can be liberated as hydrogen. Theoretical considerations have shown that 4 moles of hydrogen per mole of hexose is the maximum that can be achieved. Complete oxidation of the substrate is thermodynamically not feasible. To enable

hydrogen production beyond the limit of 4 H2/hexose, investment of extra energy is needed to overcome the thermodynamic barrier. There are different options available to achieve this energy‐dependent complete oxidation of substrate, like coupling it to a photofermentative step or microbial electrolyisis.

Hyvolution project

As mentioned in Chapter 1, within the Hyvolution project the dark fermentation step is coupled to photofermentation, in which light is used as an extra energy source. Recently a special issue of the Journal of Cleaner Production became available online in which the last developments of the project were described (145). Within the overall project (Figure 7.1) there are several points which I would like to discuss here shortly.

126

General discussion

Figure 7.1 HYVOLUTION: An integrated approach for non‐thermal hydrogen production, which covers the whole chain from biomass to hydrogen, including societal integration (205).

Biomass pretreatment

The first step in this process is the pretreatment of the biomass. As the project aims on local hydrogen production, simple procedures are needed, which make it possible to use low cost biomass. The pretreated biomass should also be suited for the dark fermentation, and within the project this has been tested for a variety of biomass substrates (231). These experiments showed that most of the used substrates can be converted into hydrogen by C. saccharolyticus, but every feedstock needs to be analyzed carefully as sometimes inhibitory products can be released in the feedstock. As C. saccharolyticus has high hydrolytic capacities, it might be possible that for some of the feedstocks no pretreatment is needed, which would save costs. Recent case studies on sugar beet and sugar beet pulp indicated this possibility and even showed that some nutrients which are normally added to the growth medium of C. saccharolyticus, are already present in these substrates (144).

127

Chapter 7

Dark fermentation reactor design

As most of the research so far on C. saccharolyticus was done in batch or continuous fermentations in 250 ml flasks or 2 l fermenters, upscaling of the process was also an important topic within the project. As shown in chapter 6, a high P(H2) results in a shift of products formed, so keeping the hydrogen pressure low in the reactors is very important and can be a problem in large scale fermenters (148). Research within the project showed, that for upscaling another species of the Caldicellulosiruptor genus, Caldicellulosiruptor owensensis was performing better in the selected reactor types ( W. Schnitzhofer, personal communication), probably because of its capacity to form biofilms. These biofilms result in higher cell densities and more stable performance in the bioreactor. C. owensensis was shown to have a clear preference for glucose over xylose (218). This is different to C. saccharolyticus which grows on both C5 and C6 sugars simultaneously, but uses C5 sugars faster. An alternative approach could be to use co‐cultures of C. saccharolyticus with C. owensensis to convertC5 and C6 simultaneously and reach higher cell densities and more stable performance due to the biofilms formed (218). Both the trickling bed and the fluidized bed systems showed advantages and disadvantages (205). Both systems showed similar productivities although the yield was higher for the trickling bed system. On the other hand the application of stripping gas was more important for this set up than for the fluidized bed. The fluidized bed also offers the advantage of more compact dimensions, as less dead volume is present compared to the trickling bed. Further research is needed to improve the performances in these reactors. As mentioned before an important factor of both reactors is the capacity to remove hydrogen from the medium broth. As shown in chapter 6, high hydrogen pressures have a negative effect on the hydrogen production by C. saccharolyticus. In most laboratory scale experiments N2 was used as stripping gas, but this has the disadvantage that it is difficult to separate from

hydrogen with generally used gas upgrading systems. CO2 would be a better candidate as stripping gas from the point of view of gas separation, but it has been shown that at least for C. saccharolyticus, it has a negative effect on growth and hydrogen production (55).

When flushed with CO2, bicarbonate is formed in the medium, which results in acidification and increased osmotic pressure. Especially this osmotic pressure seems to have a high

128

General discussion

negative effect on the performance of C. saccharolyticus (142). An alternative to gas

stripping to remove hydrogen, would be absorption of H2 in metals. The feasibility of this method is largely depending on the costs and the absorption properties of the used metals. Also the effect of the recovery of the metals to recycle plays an important role in choosing the right metal. Another option for gas removal would be applying an underpressure on the system to

create boiling conditions. The steam bubbles formed strip the H2 from the liquid phase. The problems with this method are to control the boiling conditions, and also the negative effect the boiling can have on the survival of the bacteria. Also the heat loss via steam does not make this a preferable method (143). Another point of discussion is that the use of a single species in the fermenter could have high risks of contamination. So far in laboratory scale experiments, contaminations hardly occur, most likely because of the high temperature and the fast growth of C. saccharolyticus. Of course when a continuous set up would be used, the chance of infections in the influent vessel would be higher and is a point of attention.

Photofermentation

The theoretical yield that can be obtained in total from the envisaged two step process is now 12 moles of hydrogen per mole of hexose. Although in the dark fermentation the maximum yield is nearly reached, the photo fermentation is so far less successful. To improve this yield, several problems still need to be solved. The reactor design should be in a way, that the maximum amount of light can reach the cells, and hydrogen is removed from the medium. Within the project two different systems were tested with the purple non‐sulfur bacterium Rhodobacter capsulatus (18). The conclusions from this research were that both the tubular reactor and the flat panel reactor show similar hydrogen productivities. The choice of one or the other should be location dependent. The horizontal tubular panel is using a large surface, but has a better cooling capacity and is more suitable for southern Europe. The flat panel reactor is better applicable for western Europe as it is using less surface. The composition of the effluent of the dark fermentation used for the photo fermentation can also be a problem. The advantage of an integrated project as Hyvolution makes it possible to test different effluents from the dark fermenter, originating

129

Chapter 7

from a variety of biomass feedstocks, in the photo fermenter. Recent developments within the project showed that the hydrogen productivity could be increased by adding buffer and Fe and Mo to the effluent of the dark fermentation (54, 128). As this might also be dependent on the biomass which is used for the dark fermentation, a lot of testing still needs to be done. This research also showed that the yield of this process is dependent on the amount of acetate in the effluent. Undiluted effluent for instance only yielded 2% of the

theoretical maximum of 8 moles of H2 per mole of hexose, while 4 times diluted effluent yielded 34% . Research done on a sequential dark en photo fermentation on sugar beet molasses, showed an overall hydrogen yield of 57% of the theoretical maximum of 24 moles of hydrogen per mole of sucrose (128). This is quite promising but is not yet reaching the target of the project which is 75% of the maximum yield. To reach this target more research is needed, not only on the performance of the photofermentative bacteria, but also on the biomass pretreatment, the effluent from the dark fermentation, and the possibilities to omit certain compounds from the media.

Gas upgrading

As the gaseous product stream coming from both bioreactors only contains a small percentage of hydrogen, gas upgrading is also an important step in the overall progress. Within the project a novel gas upgrading system was developed, based on absorption and

desorption of CO2 employing membrane contactors with dense membranes (123). Testing of

the system showed a very good efficiency of separating CO2 from the gas mixture. As mentioned before N2 should not be used as a flushing gas in the reactors.

System and societal integration

To make the set up of Hyvolution a sustainable and economically feasible process analysis of the system integration and societal integration are needed. Within Hyvolution both these topics are integrated, to avoid processes within the set up which will have no economic future and/or are not environmental friendly. An exergy analysis and simulation of the mass and energy balances showed that heat integration and recirculation of fermenter effluents reduce process irreversibilites and the amount of exergy leaving the process in waste streams (123). Nevertheless, depending on

130

General discussion

the used feedstock, a large amount of exergy leaves the process via by‐products. This exergy loss could be largely decreased if tail gas from the gas upgrading is used for flushing of the fermenters, and recycling of the produced mass of the bacterial cells (140). During this simulations options to reduce heat and water demands were assigned and the osmolality level in the thermophilic fermenter was determined to be a critical parameter. On the other side, recirculation of the effluent from the photofermenter to both reactors will significantly reduce the water and heat demand. By process simulation of all the process routes, using steam reforming of natural gas as a reference, several biomass based technologies were compared with the 2 stage process (168). Although the efficiency of the process is highly dependent on the choice of raw material, the Hyvolution process is comparable in efficiency and energy usage to biogas production combined with steam reforming of methane (112). A first analysis of the environmental impact and the economic consequences of the process was performed by using a life cycle approach. This showed that in the case of potato steam peels, the small non‐thermal decentralized hydrogen production has 5.7 times higher environmental impact than large scale centralized steam reforming. The largest portion of this impact is caused by the inputs (mainly phosphate, base and steam) (168). Of course this process is still in development, whereas for instance steam reforming and biogas technologies are already used on a large scale. This means that more research is needed on all the different processes within Hyvolution, but that the first steps toward application of the 2 stage biohydrogen production set up are made.

Microbial electrolysis cell

An alternative for the photofermentative production of hydrogen could be microbial electrolysis (62). With this technique the organic acids produced in the dark fermentation, are converted by electrochemically active micro‐organisms into CO2, protons and electrons. The electrons are then transported via the anode, though an electrical circuit to the cathode. The protons go through a cation membrane to also enter the cathodic department. In this compartment, driven by a small input of electricity, these protons react with the

electrons and H2 is produced. This input of energy is very small compared to the overall energy which can be derived from the hydrogen produced. With this technique 82% of the

131

Chapter 7

maximum hydrogen yield from acetate were already obtained (204). Combining this technique with the dark fermentation step will result in higher hydrogen yields, but also solve the problems of the large surface which is needed for the reactors needed for the photofermentative steps. Further research should be done on the applicability of the dark fermentation effluents in this microbial electrolysis cells.

Figure 7.2 Schematic representation of hydrogen production through biocatalyzed electrolysis of acetate (153).

Methane formation

A third option would be to couple the hydrogen production of the dark fermenter to methane formation. In this case the organic acids formed during the hydrogen production are transported to a UASB reactor containing methanogenic communities (32). With this method methane and hydrogen are formed which both can be used as a biofuel.

132

General discussion

This research

Reproducible results

Culturing of C. saccharolyticus was sometimes rather unpredictable. This hampered the reproducibility of certain experiments. Although care was taken that all conditions were exactly the same, still the results could be different. For instance, the lag phase of this organism can be variable, but also the moment the shift to lactate takes place during growth on glucose can differ. This unpredictable behavior might be related to its capacity to adapt in different ways to certain conditions. It not only has very high numbers of hydrolytic enzymes, but also a large number of transposases and transposase derivatives compared to other organisms, indicating that C. saccharolyticus is able to modify its genetic content (150).

Genome annotation

For the annotation of a genome, databases with genomic information are used to predict gene products. This is the fastest and easiest way to handle the huge amount of information available in the genome, but it also risks that genes are wrongly annotated. As most of the hits in the genomic databases are not based on characterization of the proteins, but on homology to characterized proteins, it is obvious that mistakes are easily made. Especially nowadays, with the low costs and high speed of sequencing projects, not many genomes are manually corrected after computational annotation. The genome annotation of C. saccharolyticus as described in chapter 3, was based on a computational annotation and then manually curated.

Transcriptomics

With the use of microarrays, expression patterns of genes of a complete genome can be compared. Within this research for several experiments transcriptome data were used. This showed to be a very useful method to predict pathways and substrate specificities of transporters. On the other hand, it should be kept in mind that a high number of processes in an organism are not regulated at gene expression level. A good example can be seen in

133

Chapter 7

chapter 6 in which at a high hydrogen pressure the expression of the genes coding for the NADH and reduced ferredoxin dependent hydrogenases are both upregulated. This does not mean that there is more hydrogen produced as this is not possible because of the thermodynamics (Chapter 2). The analysis of transcriptome data can also lead to discussions. In most cases only the highly up‐ or downregulated genes are taken into account. This can make the analysis of the data difficult especially in cases where the differences in expression are not that large. For instance in chapter 6 not many of the up‐ or downregulated genes met the criteria of 2 fold change. In this case we looked not only at the genes meeting this 2 fold change rule, but also at clusters of genes which were up‐ or down regulated for all its components just below 2 fold up‐ or downregulation. As the analysis of the transcriptome data is always dependent on the quality of the annotation of the genome, conclusions should be drawn carefully. Enzyme activity assays can be a useful tool to verify hypotheses obtained from transcriptome data. Also the combination of genome data, transcriptomics and product analysis as described in chapter 5, can lead to plausible conclusions.

Pathways in Caldicellulosiruptor saccharolyticus

We studied the different pathways involved in hydrogen production and these were directly linked to mechanisms of reductant disposal (Figure 7.3). In the Embden‐Meyerhof pathway, these reducing equivalents are formed as NADH and reduced ferredoxin. Under optimal hydrogen producing conditions these are reoxidized to NAD+ and oxidized ferredoxin by the two hydrogenases (NADH:Csac_1860, Csac‐1862‐1864; reduced ferredoxin: Csac_1543‐ 1539). The hydrogenases in general can be dived into 3 groups, the NiFe‐hydrogenases, FeFe‐ hydrogenases and Fe‐hydrogenases. The latter has only been described in methanogenic archaea (32). The annotated NADH dependent hydrogenase in C. saccharolyticus is FeFe‐hydrogenase and the ferredoxin dependent hydrogenase a NiFe‐ hydrogenase. The annotated ferredoxin dependent hydrogenase (Csac_1543‐1539) belongs to the group of Ech hydrogenases. This group consists out of multisubunit membrane

associated H2‐evolving NiFe‐hydrogenases. This hydrogenase consists of six subunits encoded by the echABCDEF operon (32) and it has been shown to produce hydrogen with ferredoxin as a redox partner (180) .

134

General discussion

A comparison of FeFe‐hydrogenases in the genus Clostridium, showed that these enzymes showed a wide variety of these hydrogenases. Besides monomeric hydrogenases, many clostridial FeFe‐hydrogenases are putatively dimeric, trimeric or even tetrameric enzymes (180). The presence and the pattern of conserved cystein residues within the catalytic and accessory subunits, is an indication that a high number of FeS clusters can be accommodated, and make interaction with a variety of redox partners possible. In a comparative genomic study it was shown that the diversity Fe‐only hydrogenase are severely under‐reported (188). Based on their sequences the clostridial hydrogenases were divided in 4 groups. The annotated FeFe‐hydrogenase in C. saccharolyticus (Csac_1860, Csac‐1862‐1864) was categorized into group A, as a heterotrimeric hydrogenase with NADH‐ binding domains in the catalytic subunit. Recently it was shown that a hydrogenase of Thermotoga maritima was capable to use NADH and reduced ferredoxin simultaneously (188). This hydrogenase showed homology with the FeFe‐hydrogenase (Csac_1860, Csac‐1862‐1864), which was annotated to be NADH dependent in C. saccharolyticus. Although, so far, there are no indications that this hydrogenase in C. saccharolyticus is also bifurcating, this possibility should be kept in mind. The presence of this bifurcating hydrogenase would have a strong effect on the disposal of reducing equivalents in this organism. So far, I think that just based on homology it is hard to predict this bifurcating activity of hydrogenases. Schut and Adams also predicted that the FeFe‐hydrogenase present in T. tengcongensis might have these capacities (219), but earlier activity assays with this organism showed that only NADH could be used as substrate in cell extracts (58, 84, 183). This hydrogenase shows higher homology to the Fe‐only hydrogenase of C. saccharolyticus than the described bifurcating hydrogenase in T. maritima.

135

Chapter 7

Glucose

Glucose-6-P L-Rhamnose

Fructose-6-P

L-Rhamnulose Frustose-1.6-bisP

Rhu-1-P DHAP GAP

Lactaldehyde

NAD(H2)

1.2-Propanediol H2 ? ? Succinate Pyruvate Lactate

Fd(H2)

?

Acetyl-CoA Acetaldehyde Ethanol NADP(H2) ?

Acetate

Figure 7.3 Scheme of typical carbon (grey)‐ and electron (black) flows in C. saccharolyticus described in this thesis during growth on glucose and rhamnose. ? indicate reactions which need further study.

Another pathway that needs more research is the production of ethanol in C. saccharolyticus. Under high hydrogen pressure (Chapter 6) and under carbon monoxide (Chapter 5) stress conditions, ethanol is produced. Under high hydrogen pressure conditions only one of the identified alcohol dehydrogenases (ADH) was upregulated and no candidate could be found for an acetaldehyde dehydrogenase (ALDH) in the genome of C. saccharolyticus. In T. tengcongensis, also no ALDH could be identified in the genome, but activity could be measured (130). Based on results obtained for Thermoanaerobacter ethanolicus, one dehydrogenase (TTE0695) was proposed to be bifunctional, performing both reactions from acetyl‐CoA to ethanol via acetaldehyde. A homolog of this

136

General discussion

dehydrogenase was identified in the genome of C. saccharolyticus (Csac_0395), but was not

upregulated under high P(H2) conditions. On the other hand, another ADH (Csac_0407) was upregulated under these conditions, and showed highest homology to another dehydrogenase in T. tengcongensis (TTE0696). In T. tengcongensis this dehydrogenase was proposed to have no ALDH activity and was shown to use both NADH and NADPH. The ADH activity in C. saccharolyticus cell extract was shown to be the highest on NADPH compared to NADH and direct use of reduced ferredoxin seems unlikely (232). In this case it would mean that a ferredoxin:NADP oxidoreductase should be present, but so far this protein could not be identified in C. saccharolyticus or any comparable organism. Although activities of both ferredoxin‐NAD+ and ferredoxin‐NADP+ oxidoreductase have been shown and suggested for several organisms, a molecular characterization has never been performed for thermophilic anaerobic bacteria . This characterization would really contribute to the work described in this thesis, as it will help to understand the disposal of the reducing equivalents. Under high P(H2) we saw upregulation of a cluster coding for a pyruvate carboxylase. An explanation of this upregulation could be that succinate can also function as a reducing equivalent sink via the incomplete reductive citric acid cycle. So far succinate production was not measured for C. saccharolyticus, and other genes involved in this pathway leading to succinate, were not upregulated. A fumarate reductase and a malate dehydrogenase could even not be identified in C. saccharolyticus. This indicates further research on this topic is needed to draw conclusions on the role of the incomplete reductive citric acid cycle as reductant disposal in C. saccharolyticus. Another explanation of the upregulation of this cluster of genes coding for a pyruvate carboxylase, could be that via malate protons can be transferred from NADH to NADPH. In this case also a malate dehydrogenase is needed to convert oxaloacetate to malate, which then is decarboxylated to pyruvate by a malic enzyme. As mentioned before, this dehydrogenase could not be identified in the genome of C. saccharolyticus. With the research described in this thesis, some pathways in C. saccharolyticus could be unraveled but the whole picture is not clear yet.

137

Chapter 7

Future perspectives

Combination of results of different forms of –omics will hopefully in the end lead to a mathematical model to predict hydrogen production under a variety of conditions. An important step has already been made with the development of a proteomics set up which hopefully will lead to better understanding of the regulation at protein level . During the research of this thesis, cells were also harvested for proteomics but unfortunately due to logistic problems the quality of the proteins could not be guaranteed and the data were not used for comparison with the transcriptome data. The study of the metabolites present in C. saccharolyticus would be the next step in the research leading to a systems biology approach of this organism. With metabolomics the complete set of small‐molecule metabolites (such as metabolic intermediates, signaling molecules, and secondary metabolites) can be analyzed for a certain condition and compared with another. So far there is no set up for C. saccharolyticus for metabolomics, but the fast developments in this field will hopefully lead to a reproducible protocol. A the moment C. saccharolyticus is capable to obtain almost the maximum hydrogen yield of 4 moles of hydrogen per mole of glucose using its Embden‐Meyerhof pathway. This value is strongly coupled to the amount of ATP which is needed for growth. The overall glycolysis from glucose results in the net production of 4 ATP per mole of glucose. As glucose is most likely transported into the cell by an ABC transporter, this will cost another ATP, leaving 3 ATP for growth and maintenance. With recombinant techniques it might be possible to redirect the pathways in C. saccharolyticus to obtain a higher yield of hydrogen from hexoses. To reach this the substrate should be further oxidized than acetate. An option could be to introduce the oxidative part of the pentose phosphate pathway as it is not present in C. saccharolyticus. The pathway that needs to be constructed, has already been shown to produce higher hydrogen yields in vitro by using a synthetic enzymatic pathway . In this case 13 purified enzymes were used from different origin to produce hydrogen from starch which showed to produce up to 8.35 moles of hydrogen per mole of glucose‐6‐ phosphate. This is more than twice as much as normally produced by thermophilic fermentation. Although this research is still under development, and the stability of the process is far from ideal, this can be an important method in the future. By introducing this

138

General discussion

pathway into C. saccharolyticus there might be a better stability of the hydrogen production compared to the in vitro system. On the other hand the effects of introducing a novel pathway are hard to predict.

Although we worked in the beginning of this thesis project on the development of a genetic system, this was not successful. Both in our group as well as thorough research performed by the group of Professor Kovacs in Szeged, showed that several methods to transform C. saccharolyticus were unsuccessful. Apparently this organism contains a strong defense system to block, or destroy all incoming foreign DNA material. As nowadays commercial genetic systems are available for a wide variety of organisms, I expect that in the future a working system will also be developed for C. saccharolyticus or any of the related Caldicellulosiruptor species

139

References

References

1. 2002. Future needs and challenges for non‐nuclear energy research in the European Union. Discussion Paper. 2. 2002. A National Vision of America’s Transition to a Hydrogen Economy—to 2030 and Beyond 3. Aboulwafa, M., R. Hvorup, and M. H. Saier, Jr. 2004. Dependency of sugar transport and phosphorylation by the phosphoenolpyruvate‐dependent phosphotransferase system on membranous phosphatidylethanolamine in Escherichia coli: studies with a pssA mutant lacking phosphatidylserine synthase. Archives of Microbiology 181:26‐ 34. 4. Adams, M. W., E. Eccleston, and J. B. Howard. 1989. Iron‐sulfur clusters of hydrogenase I and hydrogenase II of Clostridium pasteurianum. National Academy of Sciences of the United States of America 86:4932‐6. 5. Ajdic, D., and V. T. Pham. 2007. Global transcriptional analysis of Streptococcus mutans sugar transporters using microarrays. Journal of Bacteriology 189:5049‐59. 6. Albers, S. V., and A. J. Driessen. 2008. Membranes and transport proteins of thermophilic microorganisms, p. 39‐54. In F. T. Robb, G. Antranikian, D. W. Grogan, and A. J. Driessen (ed.), Thermophiles: Biology and technology at high temperatures. CRC Press. 7. Albertson, G. D., R. H. McHale, M. D. Gibbs, and P. L. Bergquist. 1997. Cloning and sequence of a type I pullulanase from an extremely thermophilic anaerobic bacterium, Caldicellulosiruptor saccharolyticus. Biochimica et Biophysica Acta 1354:35‐9. 8. Altschul, S. F., W. Gish, W. Miller, E. W. Myers, and D. J. Lipman. 1990. Basic local alignment search tool. Journal of Molecular Biology 215:403‐10. 9. Amend, J., P., and E. Shock, L. 2001. Energetics of overall metabolic reactions of thermophilic and hyperthermophilic Archaea and Bacteria. FEMS Microbiology Reviews 25:175‐243.

140

References

10. Amend, J. P., and A. V. Plyasunov. 2001. Carbohydrates in thermophile metabolism: calculation of the standard molal thermodynamic properties of aqueous pentoses and hexoses at elevated temperatures and pressures. Geochimica et Cosmochimica Acta 65:3901‐17. 11. Badger, J. H., and G. J. Olsen. 1999. CRITICA: coding region identification tool invoking comparative analysis. Molecular Biology and Evolution 16:512‐24. 12. Bagi, Z., N. Acs, B. Balint, L. Horvath, K. Dobo, K. R. Perei, G. Rakhely, and K. L. Kovacs. 2007. Biotechnological intensification of biogas production. Applied Microbiology and Biotechnology 76:473‐82. 13. Bao, Q., Y. Tian, W. Li, Z. Xu, Z. Xuan, S. Hu, W. Dong, J. Yang, Y. Chen, Y. Xue, Y. Xu, X. Lai, L. Huang, X. Dong, Y. Ma, L. Ling, H. Tan, R. Chen, J. Wang, J. Yu, and H. Yang. 2002. A complete sequence of the T. tengcongensis genome. Genome Research 12:689‐700. 14. Barrangou, R., C. Fremaux, H. Deveau, M. Richards, P. Boyaval, S. Moineau, D. A. Romero, and P. Horvath. 2007. CRISPR provides acquired resistance against viruses in prokaryotes. Science 315:1709‐12. 15. Barriere, C., M. Veiga‐da‐Cunha, N. Pons, E. Guedon, S. A. van Hijum, J. Kok, O. P. Kuipers, D. S. Ehrlich, and P. Renault. 2005. Fructose utilization in Lactococcus lactis as a model for low‐GC gram‐positive bacteria: its regulator, signal, and DNA‐binding site. Journal of Bacteriology 187:3752‐61. 16. Basak, N., and D. Das. 2007. The Prospect of Purple Non‐Sulfur (PNS) Photosynthetic Bacteria for Hydrogen Production: The Present State of the Art. World Journal of Microbiology and Biotechnology 23:31‐42. 17. Baumann, M. J., J. M. Eklof, G. Michel, A. M. Kallas, T. T. Teeri, M. Czjzek, and H. Brumer, 3rd. 2007. Structural evidence for the evolution of xyloglucanase activity from xyloglucan endo‐transglycosylases: biological implications for cell wall metabolism. Plant Cell 19:1947‐63. 18. Beggel, F., M. Modigell, I. J. Nowik, M. G. Shalygin, V. V. Teplyakov, and V. B. Zenkevitch. 2010. A novel gas purification system for biologically produced gases. Journal of Cleaner Production In Press, Accepted Manuscript.

141

References

19. Bergquist, P. L., M. D. Gibbs, D. D. Morris, V. S. J. Te'o, D. J. Saul, and H. W. Morgan. 1999. Molecular diversity of thermophilic cellulolytic and hemicellulolytic bacteria. FEMS Microbiology Ecology 28:99‐110. 20. Bielen, A. A. M., K. Willquist, J. Engman, J. van der Oost, E. W. J. van Niel, and S. W. M. Kengen. 2010. Pyrophosphate as a central energy carrier in the hydrogen‐ producing extremely thermophilic Caldicellulosiruptor saccharolyticus. FEMS Microbiology Letters 307:48‐54. 21. Blackwood, A. C., A. C. Neish, and G. A. Ledingham. 1956. Dissimilation of gucose at controlled pH Values by pigmented and non‐pigmented strains of Escherichia coli. Journal of Bacteriology 72:497‐9. 22. Blumer‐Schuette, S. E., I. Kataeva, J. Westpheling, M. W. Adams, and R. M. Kelly. 2008. Extremely thermophilic microorganisms for biomass conversion: status and prospects. Current Opinion in Biotechnology 19:210‐7. 23. Boiangiu, C. D., E. Jayamani, D. Brugel, G. Herrmann, J. Kim, L. Forzi, R. Hedderich, I. Vgenopoulou, A. J. Pierik, J. Steuber, and W. Buckel. 2005. Sodium ion pumps and hydrogen production in glutamate fermenting anaerobic bacteria. Journal of Molecular Microbiology and Biotechnology 10:105‐19. 24. Bolotin, A., P. Wincker, S. Mauger, O. Jaillon, K. Malarme, J. Weissenbach, S. D. Ehrlich, and A. Sorokin. 2001. The complete genome sequence of the lactic acid bacterium Lactococcus lactis ssp. lactis IL1403. Genome Research 11:731‐53. 25. Boronat, A., and J. Aguilar. 1979. Rhamnose‐induced propanediol oxidoreductase in Escherichia coli: purification, properties, and comparison with the fucose‐induced enzyme. Journal of Bacteriology 140:320‐6. 26. Boyd, D. A., D. G. Cvitkovitch, and I. R. Hamilton. 1995. Sequence, expression, and function of the gene for the nonphosphorylating, NADP‐dependent glyceraldehyde‐ 3‐phosphate dehydrogenase of Streptococcus mutans. Journal of Bacteriology 177:2622‐7. 27. Bradford, M. M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein‐dye binding. Analytical Biochemistry 72:248‐54.

142

References

28. Bredholt, S., J. Sonne‐Hansen, P. Nielsen, I. M. Mathrani, and B. K. Ahring. 1999. Caldicellulosiruptor kristjanssonii sp. nov., a cellulolytic, extremely thermophilic, anaerobic bacterium. International Journal of Systematic and Evolutionary Microbiology 49 991‐6. 29. Brenner, K., L. C. You, and F. H. Arnold. 2008. Engineering microbial consortia: a new frontier in synthetic biology. Trends in Biotechnology 26:483‐9. 30. Burdette, D., and J. G. Zeikus. 1994. Purification of acetaldehyde dehydrogenase and alcohol dehydrogenases from Thermoanaerobacter ethanolicus 39E and characterization of the secondary‐alcohol dehydrogenase (2 degrees Adh) as a bifunctional alcohol dehydrogenase‐‐acetyl‐CoA reductive thioesterase. Biochemical Journal 302 163‐70. 31. Burdette, D. S., C. Vieille, and J. G. Zeikus. 1996. Cloning and expression of the gene encoding the Thermoanaerobacter ethanolicus 39E secondary‐alcohol dehydrogenase and biochemical characterization of the enzyme. Biochemical Journal 316 115‐22. 32. Calusinska, M., T. Happe, B. Joris, and A. Wilmotte. 2010. The surprising diversity of clostridial hydrogenases: a comparative genomic perspective. Microbiology 156:1575‐88. 33. Carlsson, J., U. Kujala, and M. B. Edlund. 1985. Pyruvate dehydrogenase activity in Streptococcus mutans. Infection and Immunity 49:674‐8. 34. Chhabra, S. R., K. R. Shockley, S. B. Conners, K. L. Scott, R. D. Wolfinger, and R. M. Kelly. 2003. Carbohydrate‐induced differential gene expression patterns in the hyperthermophilic bacterium Thermotoga maritima. J Biol Chem 278:7540‐52. 35. Chittibabu, G., K. Nath, and D. Das. 2006. Feasibility studies on the fermentative hydrogen production by recombinant Escherichia coli BL‐21. Process Biochemistry 41:682‐8. 36. Chou, C. J., K. R. Shockley, S. B. Conners, D. L. Lewis, D. A. Comfort, M. W. W. Adams, and R. M. Kelly. 2007. Impact of substrate glycoside linkage. and elemental sulfur on bioenergetics, of and hydrogen production by the hyperthermophilic Archaeon Pyrococcus furiosus. Applied and Environmental Microbiology 73:6842‐53.

143

References

37. Claassen, P. A. M., and T. de Vrije. 2006. Non‐thermal production of pure hydrogen from biomass: HYVOLUTION. International Journal of Hydrogen Energy 31:1416‐23. 38. Claassen, P. A. M., T. de Vrije, E. Koukios, E. van Niel, I. Eroglu, M. Modigell, A. Friedl, W. Wukovits, and W. Ahrer. 2010. Non‐thermal production of pure hydrogen from biomass: HYVOLUTION. Journal of Cleaner Production In Press, Corrected Proof. 39. Clean Fuels Development Coalition. 2007. The ethanol fact book. 40. Conners, S. B., C. I. Montero, D. A. Comfort, K. R. Shockley, M. R. Johnson, S. R. Chhabra, and R. M. Kelly. 2005. An expression‐driven approach to the prediction of carbohydrate transport and utilization regulons in the hyperthermophilic bacterium Thermotoga maritima. Journal of Bacteriology 187:7267‐82. 41. Coutinho, P. M., and B. Henrissat. 1999. Carbohydrate‐active enzymes: an integrated database approach, p. 3‐12. In H.J. Gilbert, G. Davies, B. Henrissat, and B. Svensson (ed.), Recent Advances in Carbohydrate Bioengineering. The Royal Society of Chemistry Cambridge. 42. Dahl, M. K., D. Schmiedel, and W. Hillen. 1995. Glucose and glucose‐6‐phosphate interaction with Xyl repressor proteins from Bacillus spp. may contribute to regulation of xylose utilization. Journal of Bacteriology 177:5467‐72. 43. de Vrije, T., R. R. Bakker, M. A. Budde, M. H. Lai, A. E. Mars, and P. A. Claassen. 2009. Efficient hydrogen production from the lignocellulosic energy crop Miscanthus by the extreme thermophilic bacteria Caldicellulosiruptor saccharolyticus and Thermotoga neapolitana. Biotechnology for Biofuels 2:12. 44. de Vrije, T., G. G. de Haas, G. B. Tan, E. R. Keijsers, and P. A. Claassen. 2002. Pretreatment of Miscanthus for hydrogen production by Thermotoga elfii. Interational Journal of Hydrogen Energy 27:1381‐90. 45. de Vrije, T., A. E. Mars, M. A. W. Budde, M. H. Lai, C. Dijkema, P. de Waard, and P. A. M. Claassen. 2007. Glycolytic pathway and hydrogen yield studies of the extreme thermophile Caldicellulosiruptor saccharolyticus. Applied Microbiology and Biotechnology 74:1358‐67.

144

References

46. Delcher, A. L., D. Harmon, S. Kasif, O. White, and S. L. Salzberg. 1999. Improved microbial gene identification with GLIMMER. Nucleic Acids Res 27:4636‐41. 47. Desai, S. G., M. L. Guerinot, and L. R. Lynd. 2004. Cloning of L‐lactate dehydrogenase and elimination of lactic acid production via gene knockout in Thermoanaerobacterium saccharolyticum JW/SL‐YS485. Applied Microbiology and Biotechnology 65:600‐5. 48. Dwivedi, P. P., M. D. Gibbs, D. J. Saul, and P. L. Bergquist. 1996. Cloning, sequencing and overexpression in Escherichia coli of a xylanase gene, xynA from the thermophilic bacterium Rt8B.4 genus Caldicellulosiruptor. Applied Microbiology and Biotechnology 45:86‐93. 49. Eberly, J. O., and R. L. Ely. 2008. Thermotolerant Hydrogenases: Biological Diversity, Properties, and Biotechnological Applications. Critical Reviews in Microbiology 34:117‐30. 50. Eriksen, N. T., T. M. Nielsen, and N. Iversen. 2008. Hydrogen production in anaerobic and microaerobic Thermotoga neapolitana. Biotechnology Letters 30:103‐ 9. 51. Feist, A. M., M. J. Herrgard, I. Thiele, J. L. Reed, and B. O. Palsson. 2009. Reconstruction of biochemical networks in microorganisms. Nature Reviews Microbiology 7:129‐43. 52. Felzenszwalb, I., N. J. Sargentini, and K. C. Smith. 1986. Escherichia coli Radc Is Deficient in the Reca‐Dependent Repair of X‐Ray‐Induced DNA Strand Breaks. Radiation Research 106:166‐70 53. Fiala, G., and K. O. Stetter. 1986. Pyrococcus‐Furiosus Sp‐Nov Represents a Novel Genus of Marine Heterotrophic Archaebacteria Growing Optimally at 100‐Degrees C. Archives of Microbiology 145:56‐61. 54. Foglia, D., M. Ljunggren, W. Wukovits, A. Friedl, G. Zacchi, K. Urbaniec, and M. Markowski. 2010. Integration studies on a two‐stage fermentation process for the production of biohydrogen. Journal of Cleaner Production In Press, Accepted Manuscript.

145

References

55. Gebicki, J., M. Modigell, M. Schumacher, J. van der Bourg, and E. Roebroeck.

Comparison of two reactor concepts for anoxygenic H2 production by Rhodobacter capsulatus. Journal of Cleaner Production In Press, Corrected Proof. 56. Gerpen, J. V. 2005. Biodiesel processing and production. Fuel Processing Technology 86:1097‐107. 57. Gonzalez, R., H. Tao, K. T. Shanmugam, S. W. York, and L. O. Ingram. 2002. Global gene expression differences associated with changes in glycolytic flux and growth rate in Escherichia coli during the fermentation of glucose and xylose. Biotechnology Progress 18:6‐20. 58. Guedon, E., S. Payot, M. Desvaux, and H. Petitdemange. 1999. Carbon and electron flow in Clostridium cellulolyticum grown in chemostat culture on synthetic medium. Journal of Bacteriology 181:3262‐9. 59. Hallenbeck, P. C. 2009. Fermentative hydrogen production: Principles, progress, and prognosis. International Journal of Hydrogen Energy 34:7379‐89. 60. Hallenbeck, P. C. 2005. Fundamentals of the fermentative production of hydrogen Water Science & Technology 52:21–29 61. Hamilton‐Brehm, S. D., J. J. Mosher, T. Vishnivetskaya, M. Podar, S. Carroll, S. Allman, T. J. Phelps, M. Keller, and J. G. Elkins. 2010. Caldicellulosiruptor obsidiansis sp. nov., an anaerobic, extremely thermophilic, cellulolytic bacterium isolated from Obsidian Pool, Yellowstone National Park. Applied and Environmental Microbiology 76:1014‐20. 62. Han, S. K., and H. S. Shin. 2004. Performance of an innovative two‐stage process converting food waste to hydrogen and methane. Journal of the Air & Waste Management Association 54:242‐9. 63. Hansen, T., B. Reichstein, R. Schmid, and P. Schonheit. 2002. The first archaeal ATP‐ dependent glucokinase, from the hyperthermophilic crenarchaeon Aeropyrum pernix, represents a monomeric, extremely thermophilic ROK glucokinase with broad hexose specificity. Journal of Bacteriology 184:5955‐65. 64. Hasseman, J. 2001. TIGR Microarray Protocols, http://www.tigr.org/tdb/microarray/protocolsTIGR.shtml.

146

References

65. Hedderich, R., and L. Forzi. 2005. Energy‐converting [NiFe] hydrogenases: More than

just H2 activation. Journal of Molecular Microbiology and Biotechnology 10:92‐104. 66. Heinekey, D. M. 2009. Hydrogenase enzymes: Recent structural studies and active site models. Journal of Organometallic Chemistry 694:2671‐80. 67. Hendry, G. A. F., and R. K. Wallace. 1993. The origin, distribution, and evolutionary, significance of fructans., p. 119‐39. In M. Suzuki and N. J. Chatterton (ed.), Science and technology of fructans. CRC Press, London. 68. Henkin, T. M. 1996. The role of CcpA transcriptional regulator in carbon metabolism in Bacillus subtilis. FEMS Microbiology Letters 135:9‐15. 69. Herbel, Z., G. Rakhely, Z. Bagi, G. Ivanova, N. Acs, E. Kovacs, and K. L. Kovacs. 2010. Exploitation of the extremely thermophilic Caldicellulosiruptor saccharolyticus in hydrogen and biogas production from biomasses. Environmental Technology 31:1017‐24. 70. Hinman, L. M., and J. P. Blass. 1981. An NADH‐linked spectrophotometric assay for pyruvate‐dehydrogenase complex in crude tissue‐homogenates. Journal of Biological Chemistry 256:6583‐6. 71. Hong, M. R., Y. S. Kim, C. S. Park, J. K. Lee, Y. S. Kim, and D. K. Oh. 2009. Characterization of a recombinant beta‐glucosidase from the thermophilic bacterium Caldicellulosiruptor saccharolyticus. Journal of Bioscience and Bioengineering 108:36‐40. 72. Hong, M. R., C. S. Park, and D. K. Oh. 2009. Characterization of a thermostable endo‐1,5‐alpha‐L‐arabinanase from Caldicellulosiruptor saccharolyticus. Biotechnology Letters 31:1439‐43. 73. Hoskins, J., W. E. Alborn, Jr., J. Arnold, L. C. Blaszczak, S. Burgett, B. S. DeHoff, S. T. Estrem, L. Fritz, D. J. Fu, W. Fuller, C. Geringer, R. Gilmour, J. S. Glass, H. Khoja, A. R. Kraft, R. E. Lagace, D. J. LeBlanc, L. N. Lee, E. J. Lefkowitz, J. Lu, P. Matsushima, S. M. McAhren, M. McHenney, K. McLeaster, C. W. Mundy, T. I. Nicas, F. H. Norris, M. O'Gara, R. B. Peery, G. T. Robertson, P. Rockey, P. M. Sun, M. E. Winkler, Y. Yang, M. Young‐Bellido, G. Zhao, C. A. Zook, R. H. Baltz, S. R. Jaskunas, P. R. Rosteck, Jr.,

147

References

P. L. Skatrud, and J. I. Glass. 2001. Genome of the bacterium Streptococcus pneumoniae strain R6. Journal of Bacteriology 183:5709‐17. 74. Huang, C. Y., B. K. Patel, R. A. Mah, and L. Baresi. 1998. Caldicellulosiruptor owensensis sp. nov., an anaerobic, extremely thermophilic, xylanolytic bacterium. International Journal of Systematic and Evolutionary Microbiology 48 Pt 1:91‐7. 75. Huber, R., T. A. Langworthy, H. König, M. Thomm, C. R. Woese, U. B. Sleytr, and K. O. Stetter. 1986. Thermotoga maritima sp. nov. represents a new genus of unique extremely thermophilic eubacteria growing up to 90°C. Archives of Microbiology 144:324‐33. 76. Hueck, C. J., and W. Hillen. 1995. Catabolite repression in Bacillus subtilis: a global regulatory mechanism for the gram‐positive bacteria? Molecular Microbiology 15:395‐401. 77. Hyun, H. H., G. J. Shen, and J. G. Zeikus. 1985. Differential amylosaccharide metabolism of Clostridium thermosulfurogenes and Clostridium thermohydrosulfuricum. J Bacteriol 164:1153‐61. 78. Ivanova, G., G. Rakhely, and K. L. Kovacs. 2008. Hydrogen production from biopolymers by Caldicellulosiruptor saccharolyticus and stabilization of the system by immobilization. International Journal of Hydrogen Energy 33:6953‐61. 79. Ivanova, G., G. Rákhely, and K. L. Kovács. 2009. Thermophilic biohydrogen production from energy plants by Caldicellulosiruptor saccharolyticus and comparison with related studies. International Journal of Hydrogen Energy 34:3659‐ 70. 80. Jenney, F. E., and M. W. W. Adams. 2008. Hydrogenases of the model hyperthermophiles. Annals of the New York Academy of Sciences 1125:252‐66.

81. Jones, P. R. 2008. Improving fermentative biomass‐derived H2‐production by engineering microbial metabolism. International Journal of Hydrogen Energy 33:5122 ‐30. 82. Joyce, A. R., and B. O. Palsson. 2006. The model organism as a system: integrating 'omics' data sets. Nature Reviews Molecular Cell Biology 7:198‐210.

148

References

83. Ju, Y. H., and D. K. Oh. 2010. Characterization of a recombinant L‐fucose isomerase from Caldicellulosiruptor saccharolyticus that isomerizes L‐fucose, D‐arabinose, D‐ altrose, and L‐galactose. Biotechnology Letters 32:299‐304. 84. Jungermann, K., R. K. Thauer, G. Leimenstoll, and K. Decker. 1973. Function of reduced pyridine nucleotide‐ferredoxin oxidoreductases in saccharolytic Clostridia. Biochimica Biophysica Acta 305:268‐80. 85. Kadar, Z., T. De Vrije, M. A. Budde, Z. Szengyel, K. Reczey, and P. A. Claassen. 2003. Hydrogen production from paper sludge hydrolysate. Applied Biochemistry and Biotechnology 105 ‐108:557‐66. 86. Kadar, Z., T. de Vrije, G. E. van Noorden, M. A. Budde, Z. Szengyel, K. Reczey, and P. A. Claassen. 2004. Yields from glucose, xylose, and paper sludge hydrolysate during hydrogen production by the extreme thermophile Caldicellulosiruptor saccharolyticus. Applied Biochemistry and Biotechnology 113‐116:497‐508. 87. Kaiser, M., and G. Sawers. 1994. Pyruvate formate‐lyase is not essential for nitrate respiration by Escherichia coli. FEMS Microbiology Letters 117:163‐8. 88. Kanai, T., H. Imanaka, A. Nakajima, K. Uwamori, Y. Omori, T. Fukui, H. Atomi, and T. Imanaka. 2005. Continuous hydrogen production by the hyperthermophilic archaeon, Thermococcus kodakaraensis KOD1. Journal of Biotechnology 116:271‐82. 89. Kapdan, I. K., and F. Kargi. 2006. Bio‐hydrogen production from waste materials. Enzyme and Microbial Technology 38:569‐82. 90. Kauffman, K. J., P. Prakash, and J. S. Edwards. 2003. Advances in flux balance analysis. Current Opinion in Biotechnology 14:491‐6. 91. Kengen, S. W. M., H. P. Goorissen, M. R. A. Verhaart, A. J. M. Stams, E. W. J. van Niel, and P. A. M. Claassen. 2009. Biological Hydrogen Production by Anaerobic Microorganisms, p. 197‐221. In W. Soetaert and E. J. Vandamme (ed.), Biofuels. John Wiley & Sons ltd, Chichester. 92. Kengen, S. W. M., and A. J. M. Stams. 1994. Growth and energy conservation in batch cultures of Pyrococcus furiosus. FEMS Microbiology Letters 117:305‐9. 93. Kengen, S. W. M., A. J. M. Stams, and W. M. deVos. 1996. Sugar metabolism of hyperthermophiles. FEMS Microbiology Reviews 18:119‐37.

149

References

94. Kerkhoven, R., F. H. van Enckevort, J. Boekhorst, D. Molenaar, and R. J. Siezen. 2004. Visualization for genomics: the Microbial Genome Viewer. Bioinformatics 20:1812‐4. 95. Kikuchi, A., and K. Asai. 1984. Reverse gyrase‐‐a topoisomerase which introduces positive superhelical turns into DNA. Nature 309:677‐81. 96. Kim, J. H., S. P. Shoemaker, and D. A. Mills. 2009. Relaxed control of sugar utilization in Lactobacillus brevis. Microbiology 155:1351‐9. 97. Kim, S., E. Seol, Y. K. Oh, G. Y. Wang, and S. Park. 2009. Hydrogen production and metabolic flux analysis of metabolically engineered Escherichia coli strains. International Journal of Hydrogen Energy 34:7417‐27. 98. Kim, Y., H. Li, T. A. Binkowski, D. Holzle, and A. Joachimiak. 2009. Crystal structure of fatty acid/phospholipid synthesis protein PlsX from Enterococcus faecalis. Journal of Structural and Functional Genomics 10:157‐63. 99. Kobayashi, K., S. Doi, S. Negoro, I. Urabe, and H. Okada. 1989. Structure and properties of malic enzyme from Bacillus stearothermophilus. The Journal of Biological Chemistry 264:3200‐5. 100. Koike, M., and L. J. Reed. 1960. Alpha‐keto acid dehydrogenation complexes .2. Role of protein‐bound lipoic acid and flavin adenine dinucleotide. Journal of Biological Chemistry 235:1931‐8. 101. Koning, S. M., S. V. Albers, W. N. Konings, and A. J. Driessen. 2002. Sugar transport in (hyper)thermophilic archaea. Research in Microbiology 153:61‐7. 102. Kovarova‐Kovar, K., and T. Egli. 1998. Growth kinetics of suspended microbial cells: from single‐substrate‐controlled growth to mixed‐substrate kinetics. Microbiology and Molecular Biology Reviews 62:646‐66. 103. Kraus, A., E. Kuster, A. Wagner, K. Hoffmann, and W. Hillen. 1998. Identification of a co‐repressor binding site in catabolite control protein CcpA. Molecular Microbiology 30:955‐63. 104. Krogh, A., B. Larsson, G. von Heijne, and E. L. Sonnhammer. 2001. Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. Journal of Molecular Biology 305:567‐80.

150

References

105. Kuster‐Schock, E., A. Wagner, U. Volker, and W. Hillen. 1999. Mutations in catabolite control protein CcpA showing glucose‐independent regulation in Bacillus megaterium. Journal of Bacteriology 181:7634‐8. 106. Lee, J. M., E. P. Gianchandani, and J. A. Papin. 2006. Flux balance analysis in the era of metabolomics. Briefings in Bioinformatics 7:140‐150. 107. Lemon, B. J., and J. W. Peters. 1999. Binding of exogenously added carbon monoxide at the active site of the iron‐only hydrogenase (CpI) from Clostridium pasteurianum. Biochemistry 38:12969‐73. 108. Lendenmann, U., M. Snozzi, and T. Egli. 1996. Kinetics of the simultaneous utilization of sugar mixtures by Escherichia coli in continuous culture. Applied and Environmental Microbiology 62:1493‐9. 109. Li, F., J. Hinderberger, H. Seedorf, J. Zhang, W. Buckel, and R. K. Thauer. 2007. Coupled ferredoxin‐ and crotonyl‐CoA reduction with NADH catalyzed by the butyryl‐ CoA dehydrogenase/Etf complex from Clostridium kluyveri. J Bacteriol. 110. Lim, Y. R., R. Y. Yoon, E. S. Seo, Y. S. Kim, C. S. Park, and D. K. Oh. 2010. Hydrolytic properties of a thermostable alpha‐l‐arabinofuranosidase from Caldicellulosiruptor saccharolyticus. Journal of Applied Microbiology In Press, Corrected Proof. 111. Liu, C. F., J. L. Ren, F. Xu, J. J. Liu, J. X. Sun, and R. C. Sun. 2006. Isolation and characterization of cellulose obtained from ultrasonic irradiated sugarcane bagasse. Journal of Agricultural and Food Chemistry 54:5742‐8. 112. Logan, B. E., D. Call, S. Cheng, H. V. M. Hamelers, T. H. J. A. Sleutels, A. W. Jeremiasse, and R. A. Rozendal. 2008. Microbial electrolysis cells for high yield hydrogen gas production from organic matter. Environmental Science & Technology 42:8630‐40. 113. Ludwig, H., G. Homuth, M. Schmalisch, F. M. Dyka, M. Hecker, and J. Stulke. 2001. Transcription of glycolytic genes and operons in Bacillus subtilis: evidence for the presence of multiple levels of control of the gapA operon. Molecular Microbiology 41:409‐22. 114. Ludwig, W., K.‐H. Schleifer, and W. B. Whitman. "Revised road map to the phylum Firmicutes". In P. De Vos, G. Garrity, D. Jones, N. R. Krieg, W. Ludwig, F. A. Rainey, S.

151

References

K.‐H., and W. B. Whitman (ed.), Bergey's Manual of Systematic Bacteriology, 2nd ed., vol. 3 (The Firmicutes), in press. Springer‐Verlag, New York. 115. Luthi, E., D. R. Love, J. McAnulty, C. Wallace, P. A. Caughey, D. Saul, and P. L. Bergquist. 1990. Cloning, sequence analysis, and expression of genes encoding xylan‐degrading enzymes from the thermophile "Caldocellum saccharolyticum". Applied and Environmental Microbiology 56:1017‐24. 116. Lynd, L. R. 1996. Overview and evaluation of fuel ethanol from cellulosic biomass: Technology, Economics, the Environment, and Policy. Annual Review of Energy and Environment 21:403‐65. 117. Ma, K. S., Z. H. Zhou, and M. W. W. Adams. 1994. Hydrogen‐Production from Pyruvate by Enzymes Purified from the Hyperthermophilic Archaeon, Pyrococcus‐ Furiosus ‐ a Key Role for Nadph. FEMS Microbiology Letters 122:245‐50. 118. Magnusson, L., N. Cicek, R. Sparling, and D. Levin. 2009. Continuous Hydrogen Production During Fermentation of alpha‐Cellulose by the Thermophillic Bacterium Clostridium thermocellum. Biotechnology and Bioengineering 102:759‐66. 119. Mai, X., and M. W. Adams. 1994. Indolepyruvate ferredoxin oxidoreductase from the hyperthermophilic archaeon Pyrococcus furiosus. A new enzyme involved in peptide fermentation. Journal of Biological Chemistry 269:16726‐32. 120. Markowitz, V. M., F. Korzeniewski, K. Palaniappan, E. Szeto, G. Werner, A. Padki, X. Zhao, I. Dubchak, P. Hugenholtz, I. Anderson, A. Lykidis, K. Mavromatis, N. Ivanova, and N. C. Kyrpides. 2006. The integrated microbial genomes (IMG) system. Nucleic Acids Research 34:D344‐8. 121. Menzel, K., A. P. Zeng, and W. D. Deckwer. 1997. Enzymatic evidence for an involvement of pyruvate dehydrogenase in the anaerobic glycerol metabolism of Klebsiella pneumoniae. Journal of Biotechnology 56:135‐42. 122. Mercer, A. C., and M. D. Burkart. 2007. The ubiquitous carrier protein ‐ a window to metabolite biosynthesis. Natural Product Reports 24:750‐73. 123. Miltner, A., W. Wukovits, T. Pröll, and A. Friedl. Renewable hydrogen production: A technical evaluation based on process simulation. Journal of Cleaner Production In Press, Corrected Proof.

152

References

124. Minarik, P., N. Tomaskova, M. Kollarova, and M. Antalik. 2002. Malate dehydrogenases ‐ Structure and function. General Physiology and Biophysics 21:257‐ 65. 125. Miroshnichenko, M. L., I. V. Kublanov, N. A. Kostrikina, T. P. Tourova, T. V. Kolganova, N. K. Birkeland, and E. A. Bonch‐Osmolovskaya. 2008. Caldicellulosiruptor kronotskyensis sp. nov. and Caldicellulosiruptor hydrothermalis sp. nov., two extremely thermophilic, cellulolytic, anaerobic bacteria from Kamchatka thermal springs. International Journal of Systematic and Evolutionary Microbiology 58:1492‐6. 126. Miwa, Y., A. Nakata, A. Ogiwara, M. Yamamoto, and Y. Fujita. 2000. Evaluation and characterization of catabolite‐responsive elements (cre) of Bacillus subtilis. Nucleic Acids Res 28:1206‐10. 127. Mladenovska, Z., I. M. Mathrani, and B. K. Ahring. 1995. Isolation and characterization of Caldicellulosiruptor lactoaceticus sp‐nov, an extremely thermophilic, cellulolytic, anaerobic bacterium. Archives of Microbiology 163:223‐30. 128. Modarresi, A., W. Wukovits, D. Foglia, and A. Friedl. 2010. Effect of process integration on the exergy balance of a two‐stage process for fermentative hydrogen production. Journal of Cleaner Production In Press, Accepted Manuscript. 129. Morris, D. D., R. A. Reeves, M. D. Gibbs, D. J. Saul, and P. L. Bergquist. 1995. Correction of the beta‐mannanase domain of the celC pseudogene from Caldocellulosiruptor saccharolyticus and activity of the gene product on kraft pulp. Appl Environ Microbiol 61:2262‐9. 130. Muddiman, D., G. Andrews, D. Lewis, J. Notey, and R. Kelly. 2010. Part II: defining and quantifying individual and co‐cultured intracellular proteomes of two thermophilic microorganisms by GeLC‐MS(2) and spectral counting. Analytical and Bioanalytical Chemistry In Press, Accepted Manuscript. 131. Mulder, N. J., R. Apweiler, T. K. Attwood, A. Bairoch, A. Bateman, D. Binns, P. Bork, V. Buillard, L. Cerutti, R. Copley, E. Courcelle, U. Das, L. Daugherty, M. Dibley, R. Finn, W. Fleischmann, J. Gough, D. Haft, N. Hulo, S. Hunter, D. Kahn, A. Kanapin, A. Kejariwal, A. Labarga, P. S. Langendijk‐Genevaux, D. Lonsdale, R. Lopez, I. Letunic,

153

References

M. Madera, J. Maslen, C. McAnulla, J. McDowall, J. Mistry, A. Mitchell, A. N. Nikolskaya, S. Orchard, C. Orengo, R. Petryszak, J. D. Selengut, C. J. Sigrist, P. D. Thomas, F. Valentin, D. Wilson, C. H. Wu, and C. Yeats. 2007. New developments in the InterPro database. Nucleic Acids Research 35:D224‐8. 132. Munro, S. A., S. H. Zinder, and L. P. Walker. 2009. The fermentation stoichiometry of Thermotoga neapolitana and influence of temperature, oxygen, and pH on hydrogen production. Biotechnology Progress 25:1035‐42. 133. Nataf, Y., S. Yaron, F. Stahl, R. Lamed, E. A. Bayer, T. H. Scheper, A. L. Sonenshein, and Y. Shoham. 2009. Cellodextrin and laminaribiose ABC transporters in Clostridium thermocellum. Journal of Bacteriology 191:203‐9. 134. Nath, K., and D. Das. 2004. Improvement of fermentative hydrogen production: various approaches. Applied Microbiology and Biotechnology 65:520‐9. 135. Nelson, K. E., R. A. Clayton, S. R. Gill, M. L. Gwinn, R. J. Dodson, D. H. Haft, E. K. Hickey, J. D. Peterson, W. C. Nelson, K. A. Ketchum, L. McDonald, T. R. Utterback, J. A. Malek, K. D. Linher, M. M. Garrett, A. M. Stewart, M. D. Cotton, M. S. Pratt, C. A. Phillips, D. Richardson, J. Heidelberg, G. G. Sutton, R. D. Fleischmann, J. A. Eisen, O. White, S. L. Salzberg, H. O. Smith, J. C. Venter, and C. M. Fraser. 1999. Evidence for lateral gene transfer between Archaea and bacteria from genome sequence of Thermotoga maritima. Nature 399:323‐9. 136. Nguyen, T. A. D., S. J. Han, J. P. Kim, M. S. Kim, and S. J. Sim. 2010. Hydrogen production of the hyperthermophilic eubacterium, Thermotoga neapolitana under

N2 sparging condition. Bioresource Technology 101:S38‐S41. 137. Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Engineering 10:1‐6. 138. Nieto, C., M. Espinosa, and A. Puyet. 1997. The maltose/maltodextrin regulon of Streptococcus pneumoniae. Differential promoter regulation by the transcriptional repressor MalR. Journal of Biological Chemistry 272:30860‐5. 139. O‐Thong, S., P. Prasertsan, D. Karakashev, and I. Angelidaki. 2008. 16S rRNA‐ targeted probes for specific detection of Thermoanaerobacterium spp.,

154

References

Thermoanaerobacterium thermosaccharolyticum, and Caldicellulosiruptor spp. by fluorescent in situ hybridization in biohydrogen producing systems. International Journal of Hydrogen Energy 33:6082‐91. 140. Ochs, D., W. Wukovits, and W. Ahrer. Life cycle inventory analysis of biological hydrogen production by thermophilic and photo fermentation of potato steam peels (PSP). Journal of Cleaner Production In Press, Corrected Proof. 141. Onyenwoke, R. U., Y. J. Lee, S. Dabrowski, B. K. Ahring, and J. Wiegel. 2006. Reclassification of Thermoanaerobium acetigenum as Caldicellulosiruptor acetigenus comb. nov. and emendation of the genus description. International Journal of Systematic and Evolutionary Microbiology 56:1391‐5. 142. Özgür, E., N. Afsar, T. de Vrije, M. Yücel, U. Gündüz, P. A. M. Claassen, and I. Eroglu. Potential use of thermophilic dark fermentation effluents in photofermentative hydrogen production by Rhodobacter capsulatus. Journal of Cleaner Production In Press, Corrected Proof. 143. Özgür, E., A. E. Mars, B. Peksel, A. Louwerse, M. Yücel, U. Gündüz, P. A. M. Claassen, and I. Eroglu. 2009. Biohydrogen production from beet molasses by sequential dark and photofermentation. International Journal of Hydrogen Energy 35:511‐7. 144. Panagiotopoulos, I. A., R. R. Bakker, M. A. Budde, T. de Vrije, P. A. Claassen, and E. G. Koukios. 2009. Fermentative hydrogen production from pretreated biomass: a comparative study. Bioresource Technology 100:6331‐8. 145. Panagiotopoulos, J. A., R. R. Bakker, T. de Vrije, K. Urbaniec, E. G. Koukios, and P. A. M. Claassen. 2010. Prospects of utilization of sugar beet carbohydrates for biological hydrogen production in the EU. Journal of Cleaner Production In Press, Corrected Proof. 146. Park, A. R., and D. K. Oh. 2010. Effects of galactose and glucose on the hydrolysis reaction of a thermostable beta‐galactosidase from Caldicellulosiruptor saccharolyticus. Applied Microbiology and Biotechnology 85:1427‐35. 147. Patel, M. S., R. T.E., and R. A. Harris. 1996. Alpha‐keto acid dehydrogenase complexes. Birkhäuser, Basel.

155

References

148. Peintner, C., A. A. Zeidan, and W. Schnitzhofer. Bioreactor systems for thermophilic fermentative hydrogen production: evaluation and comparison of appropriate systems. Journal of Cleaner Production In Press, Corrected Proof. 149. Peng, H., B. L. Fu, Z. G. Mao, and W. L. Shao. 2006. Electrotransformation of Thermoanaerobacter ethanolicus JW200. Biotechnology Letters 28:1913‐7. 150. Pereira, P. M., Q. He, F. M. Valente, A. V. Xavier, J. Zhou, I. A. Pereira, and R. O. Louro. 2008. Energy metabolism in Desulfovibrio vulgaris Hildenborough: insights from transcriptome analysis. Antonie Van Leeuwenhoek 93:347‐62. 151. Peterson, A., M. H. Thomsen, H. Hauggaard‐Nielsen, and A. Thomsen. 2007. Potential bioethanol and biogas production using lignocellulosic biomass from winter rye, oilseed rape and faba bean. Biomass and Bioenergy 31:812‐9. 152. Pharkya, P., A. P. Burgard, and C. D. Maranas. 2004. OptStrain: A computational framework for redesign of microbial production systems. Genome Research 14:2367‐ 2376. 153. Pilak, O., B. Mamat, S. Vogt, C. H. Hagemeier, R. K. Thauer, S. Shima, C. Vonrhein, E. Warkentin, and U. Ermler. 2006. The crystal structure of the apoenzyme of the iron‐sulphur cluster‐free hydrogenase. Journal of Molecular Biology 358:798‐809. 154. Pitcher, D. G., N. A. Saunders, and R. J. Owen. 1989. Rapid extraction of bacterial genomic DNA with guanidium thiocyanate. Letters in Applied Microbiology 8:151‐ 156. 155. Pysz, M. A., S. B. Conners, C. I. Montero, K. R. Shockley, M. R. Johnson, D. E. Ward, and R. M. Kelly. 2004. Transcriptional analysis of biofilm formation processes in the anaerobic, hyperthermophilic bacterium Thermotoga maritima. Applied and Environmental Microbiology 70:6098‐112. 156. Pysz, M. A., D. E. Ward, K. R. Shockley, C. I. Montero, S. B. Conners, M. R. Johnson, and R. M. Kelly. 2004. Transcriptional analysis of dynamic heat‐shock response by the hyperthermophilic bacterium Thermotoga maritima. Extremophiles 8:209‐217. 157. Rainey, F. A., A. M. Donnison, P. H. Janssen, D. Saul, A. Rodrigo, P. L. Bergquist, R. M. Daniel, E. Stackebrandt, and H. W. Morgan. 1994. Description of Caldicellulosiruptor saccharolyticus gen‐nov, sp‐nov ‐ an obligately anaerobic,

156

References

extremely thermophilic, cellulolytic bacterium. FEMS Microbiology Letters 120:263‐ 6. 158. Rainey, F. A., N. L. Ward, H. W. Morgan, R. Toalster, and E. Stackebrandt. 1993. Phylogenetic analysis of anaerobic thermophilic bacteria: aid for their reclassification. Journal of Bacteriology 175:4772‐9. 159. Raux, E., H. L. Schubert, and M. J. Warren. 2000. Biosynthesis of cobalamin (vitamin B12): a bacterial conundrum. Cellular and Molecular Life Sciences 57:1880‐93. 160. Ren, Q., K. H. Kang, and I. T. Paulsen. 2004. TransportDB: a relational database of cellular membrane transport systems. Nucleic Acids Research 32:D284‐8. 161. Renewable Energy Policy Network for the 21st Century. 2009. Renewables Global Status Report: 2009 Update. 162. Reynolds, P. H. S., C. H. Sissons, R. M. Daniel, and H. W. Morgan. 1986. Comparison of cellulolytic activities in Clostridium thermocellum and three thermophilic, cellulolytic anaerobes. 51:12‐7. 163. Ridley, B. L., M. A. O'Neill, and D. Mohnen. 2001. Pectins: structure, biosynthesis, and oligogalacturonide‐related signaling. Phytochemistry 57:929‐67. 164. Robb, F. T., D. L. Maeder, J. R. Brown, J. DiRuggiero, M. D. Stump, R. K. Yeh, R. B. Weiss, and D. M. Dunn. 2001. Genomic sequence of hyperthermophile, Pyrococcus furiosus: implications for physiology and enzymology. Methods of Enzymology 330:134‐57. 165. Rodionov, D. A., A. A. Mironov, and M. S. Gelfand. 2001. Transcriptional regulation of pentose utilisation systems in the Bacillus/Clostridium group of bacteria. FEMS Microbiology Letters 205:305‐14. 166. Ros, J., and J. Aguilar. 1985. Propanediol oxidoreductases of Escherichia coli, Klebsiella pneumoniae and Salmonella typhimurium. Aspects of interspecies structural and regulatory differentiation. Biochemical Journal 231:145‐9. 167. Rouillard, J. M., M. Zuker, and E. Gulari. 2003. OligoArray 2.0: design of oligonucleotide probes for DNA microarrays using a thermodynamic approach. Nucleic Acids Res 31:3057‐62.

157

References

168. Rozendal, R. A., H. V. M. Hamelers, G. J. W. Euverink, S. J. Metz, and C. J. N. Buisman. 2006. Principle and perspectives of hydrogen production through biocatalyzed electrolysis. International Journal of Hydrogen Energy 31:1632‐40. 169. Russell, G. C., and J. R. Guest. 1991. Sequence similarities within the family of dihydrolipoamide acyltransferases and discovery of a previously unidentified fungal enzyme. Biochimica Et Biophysica Acta 1076:225‐32. 170. Sanchez, O. J., and C. A. Cardona. 2008. Trends in biotechnological production of fuel ethanol from different feedstocks. Bioresource Technology 99:5270‐95. 171. Santangelo, T. J., L. Cubonova, and J. N. Reeve. 2008. Shuttle vector expression in Thermococcus kodakaraensis: Contributions of cis elements to protein synthesis in a hyperthermophilic archaeon. Applied and Environmental Microbiology 74:3099‐104. 172. Sapra, R., K. Bagramyan, and M. W. Adams. 2003. A simple energy‐conserving system: proton reduction coupled to proton translocation. Proceedings of the National Academy of Sciences of the United States of America 100:7545‐50. 173. Sasaki, M., T. Jojima, M. Inui, and H. Yukawa. 2008. Simultaneous utilization of D‐ cellobiose, D‐glucose, and D‐xylose by recombinant Corynebacterium glutamicum under oxygen‐deprived conditions. Applied Microbiology and Biotechnology 81:691‐ 9. 174. Sato, T., T. Fukui, H. Atomi, and T. Imanaka. 2005. Improved and versatile transformation system allowing multiple genetic manipulations of the hyperthermophilic archaeon Thermococcus kodakaraensis. Applied and Environmental Microbiology 71:3889‐99. 175. Sato, T., H. Imanaka, N. Rashid, T. Fukui, H. Atomi, and T. Imanaka. 2004. Genetic evidence identifying the true gluconeogenic fructose‐1,6‐bisphosphatase in Thermococcus kodakaraensis and other hyperthermophiles. Journal of Bacteriology 186:5799‐807. 176. Schicho, R. N., K. Ma, M. W. W. Adams, and R. M. Kelly. 1993. Bioenergetics of sulfur reduction in the hyperthermophilic archaeon Pyrococcus furiosus. Journal of Bacteriology 175:1823‐30.

158

References

177. Schink, B., and A. Stams. 2006. Syntrophism among prokaryotes, p. 309‐335. In M. Dworkin, S. Falkow, E. Rosenberg, K.‐H. Schleifer, and E. Stackebrandt (ed.), The Prokaryotes. Springer Verlag, New York. 178. Schneider, E. 2001. ABC transporters catalyzing carbohydrate uptake. Research in Microbiology 152:303‐10.

179. Schroder, C., M. Selig, and P. Schonheit. 1994. Glucose fermentation to acetate, CO2

and H2 in the anaerobic hyperthermophilic eubacterium Thermotoga maritima ‐ Involvement of the Embden‐Meyerhof Pathway. Archives of Microbiology 161:460‐ 70. 180. Schut, G. J., and M. W. Adams. 2009. The iron‐hydrogenase of Thermotoga maritima utilizes ferredoxin and NADH synergistically: a new perspective on anaerobic hydrogen production. Journal of Bacteriology 191:4451‐7. 181. Schut, G. J., A. L. Menon, and M. W. W. Adams. 2001. 2‐keto acid oxidoreductases from Pyrococcus furiosus and Thermococcus litoralis. Hyperthermophilic Enzymes, Pt B 331:144‐58. 182. Shaw, A. J., D. A. Hogsett, and L. R. Lynd. 2009. Identification of the [FeFe]‐ Hydrogenase responsible for hydrogen generation in Thermoanaerobacterium saccharolyticum and demonstration of increased ethanol yield via hydrogenase knockout. Journal of Bacteriology 191:6457‐64. 183. Shaw, A. J., F. E. Jenney, M. W. W. Adams, and L. R. Lynd. 2008. End‐product pathways in the xylose fermenting bacterium, Thermoanaerobacterium saccharolyticum. Enzyme and Microbial Technology 42:453‐8. 184. Silva, P. J., E. C. D. van den Ban, H. Wassink, H. Haaker, B. de Castro, F. T. Robb, and W. R. Hagen. 2000. Enzymes of hydrogen metabolism in Pyrococcus furiosus. European Journal of Biochemistry 267:6541‐51. 185. Sissons, C. H., K. R. Sharrock, R. M. Daniel, and H. W. Morgan. 1987. Isolation of cellulolytic anaerobic extreme thermophiles from New Zealand thermal sites. Applied and Environmental Microbiology 53:832‐38. 186. Smil, V. 2006. Energy at the Crossroads, Global Science Forum Conference on Scientific Challenges for Energy Research, Paris.

159

References

187. Snoep, J. L., M. J. Teixeira de Mattos, P. W. Postma, and O. M. Neijssel. 1990. Involvement of pyruvate dehydrogenase in product formation in pyruvate‐limited anaerobic chemostat cultures of Enterococcus faecalis NCTC 775. Archives of Microbiology 154:50‐5. 188. Soboh, B., D. Linder, and R. Hedderich. 2004. A multisubunit membrane‐bound [NiFe] hydrogenase and an NADH‐dependent Fe‐only hydrogenase in the fermenting bacterium Thermoanaerobacter tengcongensis. Microbiology 150:2451‐63. 189. Stams, A. J. 1994. Metabolic interactions between anaerobic bacteria in methanogenic environments. Antonie Van Leeuwenhoek 66:271‐94. 190. Subbotina, I. V., N. A. Chernyh, T. G. Sokolova, I. V. Kublanov, E. A. Bonch‐ Osmolovskaya, and A. V. Lebedinsky. 2003. Oligonucleotide probes for the detection of representatives of the genus Thermoanaerobacter. Microbiology 72:331‐9. 191. Sun, Y., and J. Cheng. 2002. Hydrolysis of lignocellulosic materials for ethanol production: a review. Bioresource Technology 83:1‐11. 192. Tatusov, R. L., N. D. Fedorova, J. D. Jackson, A. R. Jacobs, B. Kiryutin, E. V. Koonin, D. M. Krylov, R. Mazumder, S. L. Mekhedov, A. N. Nikolskaya, B. S. Rao, S. Smirnov, A. V. Sverdlov, S. Vasudevan, Y. I. Wolf, J. J. Yin, and D. A. Natale. 2003. The COG database: an updated version includes eukaryotes. BMC Bioinformatics 4:41. 193. Te'o, V. S., D. J. Saul, and P. L. Bergquist. 1995. celA, another gene coding for a multidomain cellulase from the extreme thermophile Caldocellum saccharolyticum. Applied Microbiology and Biotechnology 43:291‐6. 194. Thauer, R. 1976. Limitation of microbial hydrogen formation via fermentation, p. 201‐294. In H. G. Schlegel and J. Barnea (ed.), Microbial Energy Conversion. Erich Goltze KG, Göttingen. 195. Thauer, R. K., K. Jungermann, and K. Decker. 1977. Energy conservation in chemotrophic anaerobic bacteria. Bacteriological Reviews 41:100‐80. 196. Thauer, R. K., B. Kaufer, M. Zahringer, and K. Jungermann. 1974. The reaction of the iron‐sulfur protein hydrogenase with carbon monoxide. European Journal of Biochemistry 42:447‐52.

160

References

197. Titgemeyer, F., and W. Hillen. 2002. Global control of sugar metabolism: a gram‐ positive solution. Antonie Van Leeuwenhoek 82:59‐71. 198. Titgemeyer, F., J. Reizer, A. Reizer, and M. H. Saier, Jr. 1994. Evolutionary relationships between sugar kinases and transcriptional repressors in bacteria. Microbiology 140 ( Pt 9):2349‐54. 199. Tyurin, M. V., S. G. Desai, and L. R. Lynd. 2004. Electrotransformation of Clostridium thermocellum. Applied and Environmental Microbiology 70:883‐90. 200. U.S. Energy Information Administration. 2010. International Energy Outlook 2010. 201. U.S. Environmental Protection Agency. 2010. U.S. Greenhouse Gas Inventory Report 2010. 202. United Nations. 2007. Sustainable bioenergy: a framework for decision makers. 203. Urbaniec, K., A. Friedl, D. Huisingh, and P. Claassen. 2010. Hydrogen for a sustainable global economy. Journal of Cleaner Production In Press, Corrected Proof. 204. van de Werken, H. J. G., M. R. A. Verhaart, A. L. VanFossen, K. Willquist, D. L. Lewis, J. D. Nichols, H. P. Goorissen, E. F. Mongodin, K. E. Nelson, E. W. J. van Niel, A. J. M. Stams, D. E. Ward, W. M. de Vos, J. van der Oost, R. M. Kelly, and S. W. M. Kengen. 2008. Hydrogenomics of the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus. Applied and Environmental Microbiology 74:6720‐9. 205. Van Groenestijn, J. W., J. H. O. Hazewinkel, M. Nienoord, and P. J. T. Bussmann. 2002. Energy aspects of biological hydrogen production in high rate bioreactors operated in the thermophilic temperature range. International Journal of Hydrogen Energy 27:1141‐47. 206. van Niel, E. W., P. A. Claassen, and A. J. Stams. 2003. Substrate and product inhibition of hydrogen production by the extreme thermophile, Caldicellulsiruptor saccharolyticus. Biotechnology and Bioengineering 81:255‐62. 207. van Niel, E. W. J., M. A. W. Budde, G. G. de Haas, F. J. van der Wal, P. A. M. Claasen, and A. J. M. Stams. 2002. Distinctive properties of high hydrogen producing extreme thermophiles, Caldicellulosiruptor saccharolyticus and Thermotoga elfii. International Journal of Hydrogen Energy 27:1391‐8.

161

References

208. Vanfossen, A. L., D. L. Lewis, J. D. Nichols, and R. M. Kelly. 2008. Polysaccharide degradation and synthesis by extremely thermophilic anaerobes. Annals of the New York Academy of Sciences 1125:322‐37. 209. VanFossen, A. L., M. R. A. Verhaart, S. M. W. Kengen, and R. M. Kelly. 2009. Carbohydrate utilization patterns for the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus reveal broad growth substrate preferences. Applied and Environmental Microbiology 75:7718‐24. 210. Vardar‐Schara, G., T. Maeda, and T. Wood, K. 2008. Metabolically engineered bacteria for producing hydrogen via fermentation. Microbial Biotechnology 1:107‐ 25. 211. Veit, A., M. Kalim Akhtar, T. Mizutani, and P. Jones, R. 2008. Constructing and

testing the thermodynamic limits of synthetic NAD(P)H:H2 pathways. Microbial Biotechnology 1:382‐94. 212. Verhaart, M. R. A., A. A. M. Bielen, J. v. d. Oost, A. J. M. Stams, and S. W. M. Kengen. 2010. Hydrogen production by hyperthermophilic and extremely thermophilic bacteria and archaea: mechanisms for reductant disposal. Environmental Technology 31:993 ‐ 1003. 213. Verhees, C. H., S. W. Kengen, J. E. Tuininga, G. J. Schut, M. W. Adams, W. M. De Vos, and J. Van Der Oost. 2003. The unique features of glycolytic pathways in Archaea. Biochemical Journal 375:231‐46. 214. Vignais, P. M., and A. Colbeau. 2004. Molecular biology of microbial hydrogenases. Current Issues in Molecular Biology 6:159‐88. 215. Wang, J., and W. Wan. 2009. Experimental design methods for fermentative hydrogen production: A review. International Journal of Hydrogen Energy 34:235‐44. 216. Wang, J., Y. Xue, X. Feng, X. Li, H. Wang, W. Li, C. Zhao, X. Cheng, Y. Ma, P. Zhou, J. Yin, A. Bhatnagar, R. Wang, and S. Liu. 2004. An analysis of the proteomic profile for Thermoanaerobacter tengcongensis under optimal culture conditions. Proteomics 4:136‐50. 217. Wiegel, J., L. G. Ljungdahl, and A. L. Demain. 1985. The Importance of Thermophilic Bacteria in Biotechnology. Critical Reviews in Biotechnology 3:39‐108.

162

References

218. Willquist, K., P. A. M. Claassen, and E. W. J. van Niel. 2009. Evaluation of the

influence of CO2 on hydrogen production by Caldicellulosiruptor saccharolyticus. International Journal of Hydrogen Energy 34:4718‐26. 219. Willquist, K., and E. W. J. van Niel. 2010. Lactate formation in Caldicellulosiruptor saccharolyticus is regulated by the energy carriers pyrophosphate and ATP. Metabolic Engineering 12:282‐90. 220. Wolfinger, R. D., G. Gibson, E. D. Wolfinger, L. Bennett, H. Hamadeh, P. Bushel, C. Afshari, and R. S. Paules. 2001. Assessing gene significance from cDNA microarray expression data via mixed models. Journal of Computational Biology 8:625‐37. 221. Woodward, J., S. M. Mattingly, M. Danson, D. Hough, N. Ward, and M. Adams. 1996. In vitro hydrogen production by glucose dehydrogenase and hydrogenase. Nature Biotechnology 14:872‐74. 222. Woodward, J., M. Orr, K. Cordray, and E. Greenbaum. 2000. Biotechnology ‐ Enzymatic production of biohydrogen. Nature 405:1014‐1015. 223. Wooley, R., M. Ruth, D. Glassner, and J. Sheehan. 1999. Process design and costing of bioethanol technology: a tool for determining the status and direction of research and development. Biotechnology Progress 15:794‐803. 224. Wu, M., Q. Ren, A. S. Durkin, S. C. Daugherty, L. M. Brinkac, R. J. Dodson, R. Madupu, S. A. Sullivan, J. F. Kolonay, D. H. Haft, W. C. Nelson, L. J. Tallon, K. M. Jones, L. E. Ulrich, J. M. Gonzalez, I. B. Zhulin, F. T. Robb, and J. A. Eisen. 2005. Life in hot carbon monoxide: the complete genome sequence of Carboxydothermus hydrogenoformans Z‐2901. PLoS Genetics 1:e65. 225. Wynn, J. P., A. B. A. Hamid, and C. Ratledge. 1999. The role of malic enzyme in the regulation of lipid accumulation in filamentous fungi. Microbiology‐Uk 145:1911‐7. 226. Wynn, J. P., and C. Ratledge. 1997. Malic enzyme is a major source of NADPH for lipid accumulation by Aspergillus nidulans. Microbiology‐UK 143:253‐7. 227. Yang, S. J., I. Kataeva, J. Wiegel, Y. Yin, P. Dam, Y. Xu, J. Westpheling, and M. W. Adams. 2009. Reclassification of 'Anaerocellum thermophilum' as Caldicellulosiruptor bescii strain DSM 6725T sp. nov. International Journal of Systematic and Evolutionary Microbiology In Press, Corrected Proof:‐.

163

References

228. Ye, X. H., Y. R. Wang, R. C. Hopkins, M. W. W. Adams, B. R. Evans, J. R. Mielenz, and Y. H. P. Zhang. 2009. Spontaneous high‐yield production of hydrogen from cellulosic materials and water catalyzed by enzyme cocktails. Chemsuschem 2:149‐52. 229. Yoshimura, M., T. Oshima, and N. Ogasawara. 2007. Involvement of the YneS/YgiH and PlsX proteins in phospholipid biosynthesis in both Bacillus subtilis and Escherichia coli. Bmc Microbiology 7:‐. 230. Zeidan, A. A., and E. W. J. Van Niel. 2009. Developing a thermophilic hydrogen‐ producing co‐culture for efficient utilization of mixed sugars. International Journal of Hydrogen Energy 34:4524‐8. 231. Zeidan, A. A., and E. W. J. van Niel. 2010. A quantitative analysis of hydrogen production efficiency of the extreme thermophile Caldicellulosiruptor owensensis OLT. International Journal of Hydrogen Energy 35:1128‐37. 232. Zhang, Y. H., B. R. Evans, J. R. Mielenz, R. C. Hopkins, and M. W. Adams. 2007. High‐ yield hydrogen production from starch and water by a synthetic enzymatic pathway. PLoS ONE 2:e456.

164

Summary

Summary

As fossil fuels are depleting and oil prices are rising the need for alternative biofuels is increasing. Microbial hydrogen production by anaerobic thermophilic microorganisms is one of the promising options that are currently being investigated. During the fermentation

process, sugar substrates are converted to hydrogen, CO2 and organic acids. In a second light‐dependent step, the organic acids are further converted to even more hydrogen and

CO2. The ongoing research in this field can lead to a biotechnological application. Within the research described in this thesis, Caldicellulosiruptor saccharolyticus was chosen as a model organism to study anaerobic thermophilic hydrogen production, because of its thermophilic advantages in combination with its capacity to break down a large variety of substrates. To understand hydrogen production by thermophilic organisms, research was done on the different pathways involved with the main focus on the disposal of reduction equivalents (e.g. NADH, NADPH and reduced ferredoxin). Chapter 2 is a review in which different mechanisms for the disposal of reducing equivalents are described for three different groups of thermophilic hydrogen producers. In C. saccharolyticus a NADH‐ and a ferredoxin‐dependent hydrogenase are involved to produce hydrogen from the respective electron carriers. In case of less optimal hydrogen production conditions this organism switches to lactate production from NADH and ethanol production form reduced ferredoxin, most likely via NADPH. To understand the pathways of hydrogen production, the genome of C. saccharolyticus was sequenced and the results of the manual curation of this genome is described in chapter 3. This showed that a large variety of genes coding for hydrolytic enzymes are present, and this clearly supports its capacity to grow on the variety of substrates, including more complex substrates like cellulose, hemicellulose, pectin and starch. In combination with transcriptomic data it could be concluded that an Embden Meyerhoff pathway is used for glycolysis in combination with the non‐oxidative pentose phosphate pathway for xylose. Chapter 3 also shows the capacity of C. saccharolyticus to co‐utilize glucose and xylose which is an important advantage when producing hydrogen from biomass.

165

Summary

This capacity to grow on a wide variety of sugars and to co‐utilize them is also shown by transcriptomic data in combination with product analysis of Caldicellulosiruptor saccharolyticus growing on individual monosaccharides (arabinose, fructose, galactose, glucose, mannose and xylose), mixtures of these sugars, as well as on xylan and xyloglucooligosacchrides (Chapter 4). In this chapter a bioinformatical analysis and the transcriptome data showed the absence of classical catabolite repression, which is in line with the observation that in the mixture of 6 monomeric sugars, all sugars were simultaneously consumed. The transcriptome data also made possible to propose the specificity of most of the 24 putative carbohydrate ATP‐binding cassette (ABC) transporters and one phosphotransferase system (PTS) identified in C. saccharolyticus. In chapter 5 an alternative pathway is described for growth on rhamnose by C. saccharolyticus. In this case, the NADH formed in the glycolysis, is entirely used to produce 1,2‐propanediol which has a large effect on the disposal of reducing equivalents. This results in a lower hydrogen production as only reduced ferredoxin is available for the hydrogenases. As expected, with changing the gas phase to carbon monoxide, to inhibit the NADH and ferredoxin dependent hydrogenases, there was no shift to lactate production, also indicating that all the NADH is used for 1,2‐propanediol production. With transcriptomics it could be shown that all proposed genes for this pathway were upregulated during growth on rhamnose. In chapter 6 is shown that an increase in the hydrogen pressure of a continuous culture of C. saccharolyticus, results in a clear metabolic shift in which the optical density and acetate production dropped, and lactate and ethanol production increased. Analysis of the transcriptome data of this experiment revealed that the responsible lactate dehydrogenase and an alcohol dehydrogenase were upregulated under high hydrogen pressure conditions. Besides this also both hydrogenases, the acetate kinase and a gene cluster coding for a pyruvate carboxylase were upregulated. On the other hand, several gene clusters coding for sugar transporters were downregulated as well as parts of pathways necessary for converting sugars other than glucose. This suggests that C. saccharolyticus under stress conditions regulates gene expression in a way that unnecessary pathways are switched off to save energy, and all options to recycle reducing equivalents are activated.

166

Summary

With the research described in this thesis we gained more insight in the fermentative behavior of anaerobic thermophilic hydrogen producing organisms and in particular of C. saccharolyticus. This will contribute to our general understanding of microbial hydrogen production, and hopefully will lead to sustainable hydrogen production in the future.

167

Samenvatting

Samenvatting

Nu fossiele brandstofvoorraaden aan het slinken zijn en de olieprijzen stijgen, neemt de interesse in alternatieve biobrandstoffen toe. Microbiële productie van waterstof door anaërobe thermofiele micro‐organismen is een van de veelbelovende mogelijkheden die momenteel worden onderzocht. Tijdens het fermentatieproces, worden suiker substraten

omgezet naar waterstof, CO2 en organische zuren. In een tweede licht‐afhankelijke stap

worden de organische zuren verder omgezet in nog meer waterstof en CO2. Het onderzoek op dit gebied kan uiteindelijk leiden tot een biotechnologische toepassing. Binnen het onderzoek beschreven in dit proefschrift, was Caldicellulosiruptor saccharolyticus gekozen als model organisme voor een sutdie naar anaërobe thermofiele productie van waterstof, vanwege de thermofiele voordelen in combinatie met de capaciteit om een grote verscheidenheid vaan substraten af te breken. Om productie van waterstof door thermofiele organismen beter te begrijpen, werd onderzoek gedaan op de verschillende pathways betrokken bij waterstofproductie met als belangrijkste focus de afvoer van reductie‐equivalenten (bijvoorbeeld NADH, NADPH en gereduceerd ferredoxine). Hoofdstuk 2 bevat een review waarbij verschillende mechanismen voor de verwijdering van reducerende equivalenten zijn beschreven voor drie verschillende groepen van thermofiele waterstof producenten. In C. saccharolyticus zijn een NADH‐en een ferredoxine‐afhankelijke hydrogenase betrokken om waterstof te produceren uit de respectieve elektronen dragers. In het geval van minder optimale condities om waterstof te prodcueren kan dit organisme omschakelen naar lactaat productie via NADH en de productie van ethanol via gereduceerd ferredoxine, hoogstwaarschijnlijk via NADPH. Om de pathways betroken bij de productie van waterstof beter te begrijpen, werd het genoom van C. saccharolyticus gesequenced en werden de aanwezige genen vergeleken met reeds eerder beschreven genomen. Dit deel van het onderzoek is beschreven in hoofdstuk 3. Dit toonde aan dat een grote verscheidenheid aan genen die coderen voor hydrolytische enzymen aanwezig zijn, en dit ondersteund haar capaciteit om te groeien op de verscheidenheid van substraten, waaronder meer complexe substraten zoals cellulose, hemicellulose, pectine en zetmeel. In combinatie met transcriptoom data kon worden geconcludeerd dat een Embden Meyerhof

168

Samenvatting

route wordt gebruikt voor de glycolyse in combinatie met de niet‐oxidatieve pentosefosfaat route voor xylose. Hoofdstuk 3 toont ook de capaciteit van C. saccharolyticus om tegelijkertijd glucose en xylose te gebruiken, wat een belangrijk voordeel is bij de productie van waterstof uit biomassa. Deze capaciteit om te groeien op een breed scala van suikers en het simultaan gebruik van C5 en C6 suikers is ook aangetoond door transcriptoom data in combinatie met product analyse van Caldicellulosiruptor saccharolyticu, waarbij individuele monosachariden (arabinose, fructose, galactose, glucose, mannose en xylose), mengsels van deze suikers en xylaan en xyloglucooligosacchrides (hoofdstuk 4) werden gebruikt als substraat. Een bioinformatica analyse en transcriptoom gegevens, toonden het ontbreken van de klassieke katabolische repressie aan, wat in overeenstemming is met de observatie dat in het mengsel van 6 monomere suikers, alle suikers gelijktijdig werden geconsumeerd. De transcriptoom data maakte het ook mogelijk om de specificiteit van de meeste van de 24 vermoedelijke ATP‐bindende cassette (ABC) transporters en een fosfotransferase systeem (PTS) geïdentificeerd in C. saccharolyticus te voorspellen. In hoofdstuk 5 wordt een alternatieve route beschreven voor groei op rhamnose door C. saccharolyticus. In dit geval is de NADH gevormd in de glycolyse, geheel gebruikt om 1,2‐propaandiol te produceren, wat een groot effect heeft op de afvoer van gereduceerde equivalenten. Dit resulteert in een lagere productie van waterstof doordat alleen gereduceerd ferredoxine beschikbaar is voor de hydrogenases. Zoals verwacht kon met het veranderen van de gasfase naar koolstofmonoxide, de NADH en ferredoxine afhankelijk hydrogenases geremd woden, en was er geen shift naar lactaat productie. Ook dit bevestigd dat alle NADH wordt gebruikt voor 1,2‐propaandiol productie. Met transcriptomics kon worden aangetoond dat alle voorgestelde genen voor deze route werden gereguleerd tijdens de groei op rhamnose. In hoofdstuk 6 wordt aangetoond dat een verhoging van de waterstof druk tijdens een continu cultuur van C. saccharolyticus, resulteert in een duidelijke metabole verschuiving waardoor de optische dichtheid en acetaat productie daalt, en lactaat en ethanol producte toenemen. Na analyse van de transcriptoom data van dit experiment is gebleken dat de genexpressie van verantwoordelijke lactaatdehydrogenase en een alcohol dehydrogenase naar boven gingen onder hoge waterstofdruk. Naast deze genen ging ook de expressie van zowel beide hydrogenases, het acetaat kinase en genen die coderen voor een pyruvaat carboxylase

169

Samenvatting

cluster omhoog. Aan de andere kant werd de genexpressie voor verschillende genenclusters coderent voor suiker transporters lager evenals delen van de pathways die nodig zijn voor het omzetten van suikers, andere dan glucose. Dit suggereert dat C. saccharolyticus onder stress‐omstandigheden genexpressie regelt op een manier die onnodige paden uitschakelt om energie te besparen, en alle opties om gereduceerde equivalenten te recyclen zijn geactiveerd.

Met het onderzoek beschreven in dit proefschrift hebben we meer inzicht gekregen in de fermentatieve gedrag van anaërobe thermofiele waterstof producerende organismen en in het bijzonder van C. saccharolyticus. Dit zal bijdragen aan de algemene kennis van de microbiële productie van waterstof, en hopelijk leiden tot een duurzame productie van waterstof in de toekomst.

170

List of publications

List of publications

Van de Werken HJG*, Verhaart MRA*, VanFossen AL*, Willquist K, Lewis DL, Nichols JD, Goorissen HP, Mongodin EF, Nelson KE, van Niel EWJ, Stams AJM, Ward DE, de Vos WM, van der Oost J, Kelly RM, Kengen SWM Hydrogenomics of the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus. (2008) Applied and Environonmental Microbiology. 74:6720‐6729.

A. L. VanFossen*, M. R. A. Verhaart*, S. M. W. Kengen, and R. M. Kelly. Carbohydrate utilization patterns by the extremely thermophilic bacterium Caldicellulosiruptor saccharolyticus reveal broad growth substrate preferences. (2009) Applied and Environmental Microbiology. 75:7718‐7724.

Verhaart, M. R. A., Bielen, A. A. M., v. d. Oost, J. , Stams, A. J. M. and Kengen, S. W. M. Hydrogen production by hyper‐ and extremely thermophilic bacteria and archaea: Mechanisms for reductant disposal. (2010) Environmental technology 31:993 ‐ 1003.

Verhaart, M.R.A, van der Vlist, V, de Groot, D., Stams A. J.M. Kengen, S. M.W. Analysis of the rhamnose pathway in Caldicellulosiruptor saccharolyticus. Manuscript in preparation.

Verhaart M.R.A., VanFossen A.L., Kelly, R.M., Stams, A.J.M., Kengen, S.M.W. The effect of hydrogen pressure on the metabolism of the thermophilic bacterium Caldicellulosiruptor saccharolyticus. Manuscript in preparation

*contributed equally

171

About the author

About the author

Marcel Reinier Antonius Verhaart was born on the 27th of November 1981 in Eindhoven the Netherlands. After graduating in 1999 for his VWO diploma at the Dorenweerd College in Doorwerth, he started studying Biotechnology at the Wageningen University. As part of this study he performed a internship in Uppsala at the Department of Molecular Evolution., doing research on the evolution of several strains of Bartonella graham. Next to this he did a thesis at Department of Food Microbiology in Wageningen on the swimming and swarming behavior of Bacillus cereus. His final thesis was at the department of Molecular Biology in Wageningen where he studied nodule formation in Medicago truncatula. After obtaining his Master degree end of 2005, he started his PhD at the Laboratory of Microbiology in February 2006. Since September 2010 he is working as a Project consultant at BioGast Sustainable Energy in Haarlem. At BioGast he is involved in several projects concerning the use of biomass to produce biogas, and upgrading it to green gas (natural gas equivalent).

172

Acknowledgements

Acknowledgements

Hier voor je ligt hij dan, mijn proefschrift, het resultaat van mijn promotieonderzoek. Dit was mij echter nooit gelukt zonder de hulp van de vele mensen die mij direct of indirect geholpen hebben met het behalen van dit resultaat.

Ten eerste moet ik Servé natuurlijk bedanken, als dagelijks begeleider heb ik enorm veel gehad aan jouw altijd parate kennis, maar zeker ook aan het feit dat je altijd tijd had voor discussies en wij als AIO’s altijd bij jouw aan konden kloppen. Je was niet alleen mijn directe begeleider maar ook een aangenaam reisgezelschap bij all reisjes die wij gemaakt hebben binnen Hyvolution. Ook wil ik je enorm bedanken voor de correcties die je hebt gedaan aan mijn proefschrift toen de deadline begon te naderen. De volgende die ik wil bedanken is Fons, ook jij hebt mij enorm geholpen tijdens het afronden van mij leesversie. Ook heb ik onze meetings erg gewaardeerd, waarbij je mij vaak door je enorme kennis in andere richtingen liet denken dan ik op dat moment deed. Dat ik 4 jaar onder je heb mogen werken bij Microbiologie heb altijd als heel plezierig ervaren. Willem de Vos wil ik bedanken als hoofd van onze vakgroep maar ook zeker voor de kritische vragen en zijn commentaar op het presenteren tijdens de PhD‐meetings. Tijdens de DB meetings heb ik ook veel van je kunnen leren op het gebeid van efficiënt managen John van der Oost wil ik bedanken voor het attenderen op deze functie bij Microbiologie maar ook voor de discussies die we over mijn project gehad hebben. Naast dit natuurlijk ook zeker voor de vogelkijktripjes en vooral de meesterlijke schaatstocht op de Oostvaarders plassen. Ook Stan en Edze wil ik hierbij bedanken voor hun bijdrage aan deze activiteiten.

Here I would also like to acknowledge all the Hyvolution people! It was very nice to meet you twice a year and I think we had very nice discussions especially with our WP2 partners. Karin it was really nice to get to know you during this four years, and I hope we stay in touch! Een extra bedankje gaat ook uit naar de A&F mensen, niet alleen waren jullie gezellige reispartners, maar de meetings bij ons of bij jullie heb ik altijd enorm gewaardeerd.

173

Acknowledgements

I would also like to acknowledge Robert Kelly and Amy Vanfossen for the very nice collaboration on the micro‐array experiments and writing the papers afterwards. Harmen wil ik hierbij bedanken voor de samenwerking aan het Caldi project, ik heb altijd enorm met je gelachen, maar ook zeker een heleboel van je geleerd. Ook wil ik Bram bedanken voor zijn bijdrages aan de discussies over Caldi, het was erg relax om ook op ons kantoor af en toe te brainstormen over alle pathways die mogelijk betrokken waren.

Heel erg dankbaar ben ik de 4 studenten die mij een afstudeervak hebben gedaan, jullie hebben daardoor stuk voor stuk een grote bijdrage geleverd aan mijn onderzoek. Niet alleen vond ik het superleuk om jullie te begeleiden en te proberen jullie iets te leren, ook andersom heeft het mij veel geleerd en waardeer ik jullie inzet enorm. Dus Dirk‐Jan, Bas, Ruben en Vincent, bij deze hartstikke bedankt.

Natuurlijk kan ik mijn kantoorgenoten ook niet vergeten, ik heb het altijd enorm gewaardeerd om bij jullie op 1 kamer te zitten en ik denk dat ik wel durf te stellen dat jullie (John, Mark en Marco) echte vrienden van mij zijn geworden. Suzanne, ondanks dat je tussen de mannen zat heb je je toch altijd aardig geweerd, ook jij bedankt voor je bijdrage aan de gezelligheid bij ons in het hok. Andere collega’s die ik graag wil bedanken die voor mij ook zeker meer waren dan gewoon collega’s waren Christian en Bart. Christian, bedankt voor al de gezellige uurtjes bij Micro, maar misschien nog wel meer erbuiten. Ik heb altijd enorm met je gelachen en ik hoop dat er nog veel meer van dit soort avondjes zullen komen. Bart ik hoop dat we onze spelletjes avond traditie in ere kunnen houden, ook al is de afstand nu wel meer dan 200km geworden. I also would like to thank my International Club mates, Teun, Martin and Jose. I always had a lot of fun with you guys going to parties, and I hope we can continue this tradition.

One of the things I will never forget about my PhD‐time was organizing the PhD‐trip. This would have not been as nice if we were not with a very nice group of people doing this. Thomas, Fabian, Matthijs and John thanks a lot for that.

174

Acknowledgements

Here I also would like to thank the other people of BacGen (Marke, Wouter,Vincent, Katrin, Nicolas, Pierpaolo, Rie, Mark M, Fabian, Magnus, Jasper A) and MicPhys (Caroline, Petra, Marjet, Rozelin, Janine, Melike, Anne Meint, Bart, Miriam, Gosse, Alona, Flavia) and of course the people of MolEco and FunGen De volgende groep mensen in het lijstje die ik zeker niet wil vergeten zijn de mensen die mij altijd bij hebben gestaan als ik vragen had over methodes of apparatuur of gewoon om te zeggen waar ik welk potje kon vinden. Bij deze bedankt, Sjon, Ton, Hans en Ans! Ook het vaste team wat alle andere zaken rond Microbiologie regelt mag natuurlijk niet vergeten worden. Anja, Francis, Wim, Reneé, Jannie en natuurlijk Nees, zonder jullie was het leven voor een AIO bij Microbiologie een stuk lastiger geweest! Bij deze wil ook Joop ook bedanken, voor het schoonmaken rond mijn bureau, helaas was bovenop niet altijd mogelijk vanwege het aanwezige papierwerk. Buiten het werk om was het ook erg relax om mij geregeld op het vogelen te kunnen storten. Hierbij wil ik dan ook Daan en Leendert van harte bedanken, om mij geregeld af te leiden van dit proefschrift, niet alleen op zoek naar vogels, maar ook de gezellige avondjes er tussen door. Verder kan ik het ook niet maken om mijn “Rijnsteeg maatjes” Sjoerd en Allart te vergeten, ik hoop dat we nog lang maatjes zullen blijven.

Een andere belangrijk vorm van afleiding was voor mij het badmintonnen en bij deze wil ik alle Lobbers van harte bedanken voor de toptijd die ik bij deze club gehad heb.

Ook mijn zusje mag natuurlijk niet ontbreken in dit lijstje. Carla bedankt voor de gezellige telefoongesprekken die we geregeld hebben, en we moeten echt proberen vaker in real‐life af te spreken. Ten slotte wil ik mijn ouders enorm bedanken voor al de support, zowel financieel als mentaal, die zij mij mijn hele leven hebben gegeven. Zonder jullie was ik nooit zover gekomen en had dit proefschrift er waarschijnlijk niet gelegen.

175

176

177

The research described in this thesis was supported by the Commission of the European Communities, Sixth Framework Programme, Priority 6, Sustainable Energy Systems (019825 HYVOLUTION).

178