Department of Chemistry

Umeå University, Sweden 2013

Biorefining of lignocellulose Detoxification of inhibitory hydrolysates and potential utilization of residual streams for production of enzymes

Adnan Cavka

Biorefining of lignocellulose Detoxification of inhibitory hydrolysates and potential utilization of residual streams for production of enzymes

Adnan Cavka

Copyright © Adnan Cavka 2013

This work is protected by the Swedish Copyright Legislation (Act 1960:729)

Cover page: Environmental issues, purchased from iStockphoto.com

ISBN: 978-91-7459-759-2

Digital version available at http://umu.diva-portal.org/

Printed by: VMC-KBC Umeå

Umeå, Sweden 2013

WHAT IS SUCCESS? I THINK IT IS A MIXTURE OF HAVING A FLAIR FOR THE

THING THAT YOU ARE DOING; KNOWING THAT IT IS NOT ENOUGH, THAT YOU

HAVE GOT TO HAVE HARD WORK AND A CERTAIN SENSE OF PURPOSE.

THE RT. HON. MARGARET H. THATCHER

i

Abstract

Lignocellulosic biomass is a renewable resource that can be utilized for the

production of biofuels, chemicals, and bio-based materials. Biochemical

conversion of lignocellulose to advanced biofuels, such as cellulosic ethanol,

is generally performed through microbial fermentation of sugars generated

by thermochemical pretreatment of the biomass followed by an enzymatic

hydrolysis of the cellulose. The aims of the research presented in this thesis

were to address problems associated with pretreatment by-products that

inhibit microbial and enzymatic biocatalysts, and to investigate the potential

of utilizing residual streams from pulp mills and biorefineries to produce

hydrolytic enzymes.

A novel method to detoxify lignocellulosic hydrolysates to improve the

fermentability was investigated in experiments with the yeast

Saccharomyces cerevisiae. The method is based on treatment of

lignocellulosic slurries and hydrolysates with reducing agents, such as

sodium dithionite and sodium sulfite. The effects of treatment with sodium

borohydride were also investigated. Treatment of a hydrolysate of Norway

spruce by addition of 10 mM dithionite resulted in an increase of the

balanced ethanol yield from 0.03 to 0.35 g/g. Similarly, the balanced ethanol

yield of a hydrolysate of sugarcane bagasse increased from 0.06 to 0.28 g/g

after treatment with 10 mM dithionite. In another study with a hydrolysate

of Norway spruce, addition of 34 mM borohydride increased the balanced

ethanol yield from 0.02 to 0.30 g/g, while the ethanol productivity increased

from 0.05 to 0.57 g/(L×h). While treatment with sulfur oxyanions had a

positive effect on microbial fermentation and enzymatic hydrolysis,

treatment with borohydride resulted in an improvement only for the

microbial fermentation. The chemical effects of treatments of hydrolysates

with sodium dithionite, sodium sulfite, and sodium borohydride were

investigated using liquid chromatography-mass spectrometry (LC-MS).

Treatments with dithionite and sulfite were found to rapidly sulfonate

inhibitors already at room temperature and at a pH that is compatible with

enzymatic hydrolysis and microbial fermentation. Treatment with

borohydride reduced inhibitory compounds, but the products were less

hydrophilic than the products obtained in the reactions with the sulfur

oxyanions.

The potential of on-site enzyme production using low-value residual

streams, such as stillage, was investigated utilizing recombinant Aspergillus

niger producing xylanase and cellulase. A xylanase activity of 8,400 nkat/ml

and a cellulase activity of 2,700 nkat/ml were reached using stillages from

ii

processes based on waste fiber sludge. The fungus consumed a large part of

the xylose, the acetic acid, and the oligosaccharides that were left in the

stillages after fermentation with S. cerevisiae. In another study, the

capability of two filamentous fungi (A. niger and Trichoderma reesei) and

three yeasts (S. cerevisiae, Pichia pastoris, and Yarrowia lipolytica) to grow

on inhibitory lignocellulosic media were compared. The results indicate that

the two filamentous fungi had the best capability to utilize different nutrients

in the media, while the S. cerevisiae strain exhibited the best tolerance

against the inhibitors. Utilization of different nutrients would be especially

important in enzyme production using residual streams, while tolerance

against inhibitors is desirable in a consolidated bio-process in which the

fermenting microorganism also contributes by producing enzymes.

iii

Table of Contents

Abstract i List of Papers v List of abbreviations vii Populärvetenskaplig sammanfattning ix Introduction 1

A brief history of ethanol as a fuel 5 Incentives for ethanol fuel 6

Bioethanol production from renewable and sustainable raw materials 9 Current bioethanol production 9 Conventional bioethanol 9 Cellulosic bioethanol 10 Lignocellulosic feedstocks 11 Pretreatment of lignocellulosic feedstocks 15 Fermentation processes 19

Fermentation inhibition 22 Small aliphatic acids 22 Furan aldehydes 23 Aromatic compounds 23 Strategies for dealing with fermentation inhibition problems 24 Adaptation and strain selection 24 Genetic engineering of microorganisms 25 Process design 26

Detoxification 28 Detoxification with reducing agents 29

Analysis of inhibitory compounds 36 Mass Spectrometry 36 Ionization methods 37 Mass analyzers 38 Detectors 39 Mechanism behind treatment with reducing agents 40

Enzymes in biorefining 44 Cellulases 44 Hemicellulases 45 Lytic polysaccharide monooxygenases 45 Assay of enzymatic activity 46 Filamentous fungi 47 Aspergillus niger 48 Trichoderma reesei (Hypocrea jecorina) 49

Potential of on-site enzyme production 51

iv

Enzyme expression and gene regulation 55 Inhibition of enzyme-producing microorganisms 56 Conclusions 59 Acknowledgments 61 References 63

v

List of Papers

This thesis is based on the following publications, which are referred to in

the text by their Roman numerals.

I Alriksson B, Cavka A, and Jönsson LJ. (2011) Improving the

fermentability of enzymatic hydrolysates of lignocellulose

through chemical in-situ detoxification with reducing agents.

Bioresource Technol. 102:1254-1263.

II Cavka A, Alriksson B, Ahnlund M, and Jönsson LJ. (2011)

Effect of sulfur oxyanions on lignocellulose-derived

fermentation inhibitors. Biotechnol. Bioeng. 108:2592-2599.

III Cavka A and Jönsson LJ. (2013) Detoxification of

lignocellulosic hydrolysates using sodium borohydride.

Bioresource Technol. 136:368-376.

IV Cavka A, Alriksson B, Rose SH, van Zyl WH, and Jönsson LJ.

(2011) Biorefining of wood: combined production of ethanol

and xylanase from waste fiber sludge. J. Ind. Microbiol. Biot.

38:891-899.

V Cavka A, Alriksson B, Rose SH, van Zyl WH, and Jönsson LJ.

Production of cellulosic ethanol and enzyme from waste fiber

sludge using SSF, recycling of hydrolytic enzymes and yeast,

and recombinant cellulase-producing Aspergillus niger.

(Submitted)

VI Cavka A and Jönsson LJ. Comparison of the growth of

filamentous fungi and yeasts in lignocellulose-derived media.

(Submitted)

vi

Additional publications by the author

Cavka A, Guo X, Tang S-J, Winestrand S, Jönsson LJ, and Hong F. (2013)

Production of bacterial cellulose and enzyme from waste fiber sludge.

Biotechnol. Biofuels 6:25.

Guo X, Cavka A, Jönsson LJ, and Hong F. (2013) Comparison of methods

for detoxification of spruce hydrolysate for bacterial cellulose production.

Microb. Cell Fact. 12:93.

vii

List of abbreviations

ADH Alcohol Dehydrogenase

AFEX Ammonia Fiber Explosion

A. niger Aspergillus niger

APCI Atmospheric Pressure Chemical Ionization

ARP Ammonia Recycled Percolation

ATP Adenosine Triphosphate

ATPase Adenosinetriphosphatase

BAFF BioAlcohol Fuel Foundation

CBP Consolidated Bioprocessing

Cel7B Endo-1,4-β-glucanase I Protein

CI Chemical Ionization

CPB Consolidated Bioprocess

DAD Diode Array Detector

DART Direct Analysis in Real Time

DIOS Desorption/Ionization On Silicon

DME Dimethyl Ether

DW Dry Weight

E10 Automobile fuel containing 90 percent gasoline and 10

percent ethanol

E85 Automobile fuel containing 85 percent ethanol and 15 percent

gasoline

ED Entner–Doudoroff Pathway

ED95 Diesel substitute with 95 percent ethanol and 5 percent water

and ignition improver

EgI Endo-1,4-β-glucanase I Gene

EI Electron Ionization

EM Embden-Meyerhof Pathway

ESI Electrospray Ionization

FAB Fast Atom Bombardment

FD Field Desorption

FPU Filter Paper Units

GHG Greenhouse Gas

HMF 5-(Hydroxymethyl)furfural

HPAEC High-Performance Anion-Exchange Chromatography

HPLC High-Performance Liquid Chromatography

ICR Fourier-Transform Ion-Cyclotron Resonance

IEA International Energy Agency

LC Liquid Chromatography

LPMO Lytic Polysaccharide Monooxygenases

viii

LTQ Linear Trap Quadrupole

MALDI Matrix-Assisted Laser Desorption/Ionization

MS Mass Spectrometry

MS/MS Tandem Mass Spectrometry

M/Z Mass-to-Charge

NAD Nicotinamide Adenine Dinucleotide

NADP Nicotinamide Adenine Dinucleotide Phosphate

NEV Net Energy Value

OAPEC Organization of Arab Petroleum Exporting Countries

PAD Pulsed Amperometric Detector

P. pastoris Pichia pastoris

RFA Renewable Fuels Association

RID Refractory Index Detector

S. cerevisiae Saccharomyces cerevisiae

SHF Separate Hydrolysis and Fermentation

SIMS Secondary Ion Mass Spectrometry

SSF Simultaneous Saccharification and Fermentation

SSL Spent Sulfite Liquor

TOF Time-Of-Flight

T. reesei Trichoderma reesei

UHPLC Ultra-High-Performance Liquid Chromatography

UV/VIS Ultraviolet/Visible Spectroscopy

VOC Volatile Organic Compounds

XynII Endo-1,4-β-xylanase II Gene

Y. lipolytica Yarrowia lipolytica

YPD Yeast Extract Peptone Dextrose

Yx-ccs Yield of Biomass on Consumed Carbon Source

Yx-ics Yield of Biomass on Initial Carbon Source

ix

Populärvetenskaplig sammanfattning

Innovation och tekniska framsteg har genom historien varit de främsta

drivkrafterna bakom civilisationers kulturella, tekniska och ekonomiska

utveckling. Vi lever nu i en tid av extraordinär ekonomisk och teknisk

utveckling som omfattar hela vår värld. Råoljan som varit det bränsle många

ekonomier har byggt på kommer inte att bytas ut utan att uppenbara

fördelar med alternativen kan uppvisas. Den framtida långsiktiga

ekonomiska tillväxt som vi förväntas nå, kan bara drivas av ännu större

innovation och tekniska framsteg än vad som tidigare uppnåtts. En av dessa

innovationer som kan bidra till stabilare miljömässig och ekonomisk

utveckling är tillverkning av kemikalier, material och drivmedel från

förnyelsebara råvaror.

Denna avhandling presenterar forskning om tekniker med vilka en mer

innovativ och effektiv konvertering av vedråvara och jordbruksrester till

flytande biobränslen kan uppnås. En metod är baserad på användning av

natriumditionit, natriumsulfit och natriumborhydrid, kemikalier som kan

fås till relativt låg kostnad och som idag används inom massa- och

pappersindustrin samt inom textiltillverkning. Denna nya metod ger stora

förbättringar i jäsningssteget vid tillverkning av cellulosabaserad etanol, en

effekt som uppnås redan vid låga temperaturer, låg kemikaliedosering och

utan behov av separata processteg. Avhandlingen presenterar och diskuterar

också möjligheten att utnyttja kvarvarande restströmmar från massabruk

och bioraffinaderier för att producera enzymer på samma plats där dessa

enzymer ska användas för produktion av cellulosabaserad etanol. Våra

resultat visar att det är möjligt att utnyttja kvarvarande industriella

restströmmar av lågt värde, t.ex. fiberslam och drank, för att producera

relativt höga halter av enzymer. Dessa resultat är exempel på tekniska

innovationer och framsteg som kan tillåta oss att ta ett steg närmare

kommersiell produktion av förnyelsebara flytande biobränslen som kan vara

en del av en framtida bio-baserad ekonomi.

1

Introduction

'The Stone Age did not end for lack of stone, and the oil age will end long

before the world runs out of oil.'

Sheik Ahmed Zaki Yamani, 2nd Saudi Oil Minister (in office 1962-1986)

Throughout history innovation and technological advancements have been

the main driving forces behind human, cultural and economic development.

We are now living in an age of extraordinary economic and technological

development across the globe. Oil-fueled economies will not exchange their

fuel for something else, unless there are highly evident advantages with the

new fuel. Future long-term economic growth has become even more

dependent on innovation and technological advances than ever before. One

of these technological advances involves a more environmentally and

economically sustainable production of fuels.

The technique to produce ethanol by fermentation of naturally abundant

sugar is one of the earliest biocatalytic reactions employed by humanity.

Archeological evidence suggests that the earliest production of ethanol by

means of fermentation occurred as early as 9,000 B.C. in modern day

Georgia (Cavalieri et al., 2003). In more recent times, ethanol has gained

increasing importance as a fuel in internal combustion engines and as a

precursor chemical for production of materials. Environmental awareness,

high crude oil prices, and economic potential have all contributed to this

development.

Ethanol is currently the most commonly used bio-fuel in internal

combustion engines and the most widely spread renewable transportation

fuel in the world (Leboreiro and Hilaly, 2013). However, conventional

bioethanol, which is produced from starch or sucrose derived from crops

such as corn, wheat or sugarcane, currently represents the largest portion of

bioethanol used as a transportation fuel (Leboreiro and Hilaly, 2013). Low-

cost lignocelluloses represent an alternative raw material for production of

bioethanol and other commodities in biorefineries (Saddler and Mabee,

2007; Lynd et al., 2008). Biorefineries may either be thermochemical, based

on gasification and syngas, or biochemical, based on enzymatic and

biological conversion of lignocellulose to value-added products. The use of

lignocellulose for production of bioethanol does however present us with

several difficult challenges. Lignocellulose is a complex substrate mainly

composed of lignin, cellulose and hemicelluloses that form a recalcitrant

material that is relatively difficult to convert to fermentable sugars. Before

2

the rise of biotechnology, conversion of lignocellulose to fermentable sugar

was typically performed by using strong mineral acid and high temperature

in an acid-hydrolysis process.

Acid hydrolysis results in the formation of by-products, which in sufficiently

high concentrations are inhibitory to the fermenting microorganism

(Jönsson et al., 2013). Alternatively, hydrolysis of lignocellulose

polysaccharides can be achieved by using enzymes. In that case, highly

specific hydrolytic enzymes, cellulases and hemicellulases, are used to

depolymerize cellulose and hemicelluloses to sugars without the formation of

toxic and inhibitory compounds. However, prior to enzymatic hydrolysis

lignocellulosic biomass needs to be pretreated to make the cellulose more

susceptible to the action of cellulolytic enzymes. Pretreatment is commonly

carried out using steam under acidic conditions in a process that degrades

the hemicelluloses through acid hydrolysis and leaves the cellulose for the

subsequent enzymatic step. The main challenge that arises with enzymatic

hydrolysis is the high cost of hydrolytic enzymes. It is considered to be one of

the major bottlenecks for commercialization of cellulosic ethanol. From a

life-cycle assessment point of view, conventional enzyme production is

associated with high negative environmental impacts such as high energy

consumption, greenhouse gas (GHG) emissions, acidification, and, most

importantly, eutrophication due to high nitrogen utilization (Fu et al., 2003).

Filamentous fungi, such as Aspergillus niger and Trichoderma reesei, are

commonly used as industrial workhorses in several different industrial

fermentation processes, among them the production of commercial

hydrolytic enzymes. These fungi are metabolically diverse and have the

natural ability to produce and secrete a wide array of hydrolytic enzymes

which are naturally involved in the breakdown of cellulose and

hemicelluloses to sugars. Utilization of these fungi in an on-site enzyme

production facility based on residual streams originating from the bioethanol

process is an intriguing opportunity to lower the cost of hydrolytic enzymes

and to fully utilize the different components of the raw material.

Residual streams, such as stillage, do, however, not only contain

unconventional carbon sources, such as pentose sugars and acetic acid, but

also degradation products of sugar and lignin. Toxic inhibitory compounds

are a problem for both the fermenting microorganism and microorganisms

used in potential subsequent cultivation steps. Dealing with toxic and

inhibitory compounds and lowering the cost of hydrolytic enzymes are both

of vital importance for the commercialization of cellulosic bioethanol. This

thesis will present research on the potential of on-site enzyme production

3

from residual streams as well as on novel methods for dealing with toxic and

inhibitory compounds that affect fermenting microorganisms.

4

Biorefining of lignocellulose

Conventional oil refineries utilize physical and chemical processes to refine

crude oil to different fractions that are used for the production of fuels,

chemicals and materials. The products have one more thing in common

besides the raw material; they are produced because of their economic value.

Biorefineries are not very different from traditional oil refineries in the sense

that they operate on the same principles; raw material goes into the

biorefinery and different types of value-added products come out of it. The

difference between the two types of refining lies in the raw material used, the

conversion techniques, as well as in the products produced. However, some

of the products that are today produced in oil refineries may potentially also

be produced in biorefineries. Biorefineries aim to create and produce

sustainable substitutes for fuels, chemicals and materials currently produced

in oil refineries, i.e. create similar values for the end user of oil-derived

products through a distinctively different process.

The advantage with biorefineries compared to oil refineries lies in the costs

of the raw materials; lignocellulosic biomass has a much bigger potential

than crude oil. Lynd et al. (2008) and Lynd (2011) state that the raw material

cost of crude oil at a price of 50-75 USD per barrel is about 8.7-13.5 USD per

gigajoule. The value added in the refining of oil is around 25 percent from

raw material to finished product (Lynd, 2011), i.e. an increase in value from

75 USD before refining to 100 USD after refining. This means that the cost of

the raw material constitutes around three quarters (75 percent) of the total

value of the refined products, and process costs are thus one quarter (25

percent). In a future scenario, lignocellulosic biomass may cost around 50 or

60 USD per dry ton, which corresponds to 3-4 USD per gigajoule (Lynd et

al., 2008; Lynd, 2011). The raw material price for biorefining is thus around

one third of the price of the raw material for a conventional oil refinery.

The difference in the cost of the raw material results in that the process costs

of biomass refining may be three times as high as for oil refining and

biomass refining would still result in competitive prices for the end products.

Alternatively, it can be interpreted so that the potential of biorefining is

much greater than that of oil refining, especially since the process costs are

likely to become lower for the biorefinery through technical development

and since it is unlikely that crude oil will be cheaper in the future. Either way

these results show the potential of biorefining for production of fuels,

materials or chemicals that are today produced from oil refining. Further

innovation and technical development is needed to realize this potential and

5

allow processing of lignocellulosic materials to fuels, materials and other

commodities to really become competitive against oil-derived products of the

same kind.

A brief history of ethanol as a fuel

The use of ethanol as a fuel in internal combustion engines is not an entirely

novel idea. In fact, the earliest use of ethanol as a fuel in combustion engines

was back in the 1820s when Samuel Morey constructed the first American

prototype internal combustion engine, running it on a mix of turpentine and

ethanol (Soni, 2007, p.277). Despite filing for a patent regarding the idea in

1826, the invention never rendered any great attention, mainly due to

Morey’s lack of funding and his greater interest in steam engines.

It was not until 50 years later that ethanol would be used as fuel again, when

the German engineer and inventor Nicholas Otto used ethanol to power

some of his four-stroke-piston-chamber engines (Soni, 2007, p.277). This

type of engine, which was the first efficient alternative to the established

steam engine, would become increasingly popular, and would later on be

considered to be the very prototype of combustion engines built ever since.

The first use of ethanol as a transportation fuel in larger scale did not come

about until Henry Ford I entered the automobile stage. The engines in his

early Ford Model T built in 1908, commonly known as T-Ford, were built to

run on either ethanol, gasoline or a combination of the two (DiPardo, 2000).

Henry Ford was also quite optimistic regarding the potential of ethanol as an

automobile fuel, expressing in an interview with the New York Times in 1925

that ethanol is “the fuel of the future” which “is going to come from fruit like

that sumach out by the road, or from apples, weeds, sawdust -- almost

anything. There is fuel in every bit of vegetable matter that can be

fermented. There's enough alcohol in one year's yield of an acre of potatoes

to drive the machinery necessary to cultivate the fields for a hundred years”

(New York Times, 1925, p.24). Turn of events would however not fulfill

Henry Ford’s prediction regarding the use of ethanol as the main automobile

fuel in the future, as gasoline would displace ethanol due to the lower

production cost and due to its abundance.

Ethanol as an alternative to gasoline for use as automobile fuel would

however make a strong comeback as it received increasing attention in the

1970s. The 1973 oil embargo imposed by the Organization of Arab Petroleum

Exporting Countries (OAPEC) as a result of the Yom Kippur war made the

gasoline price skyrocket and forced governments around the world to look

6

for alternatives. The US president at the time, Richard M. Nixon, proclaimed

the famed “Energy Independence” in an address to the nation in Nov. 1973

saying: “Let us set as our national goal, in the spirit of Apollo, with the

determination of the Manhattan Project, that by the end of this decade we

will have developed the potential to meet our own energy needs without

depending on any foreign energy sources” (Williams, 2005, p. 268).

Nixon’s aim to make America energy independent by the end of the decade

was, however, never met, but a second oil crisis sparked by the Iranian

revolution in 1979 went to prove that the idea of energy security through

independence was indeed a highly desirable target. The Brazilian

government that had initiated the National Alcohol Program, also known as

PROALCOOL, in 1974 was more successful with their aim of energy

independence. This program resulted in the mid-1980s in that 85 percent of

all cars in Brazil ran exclusively on domestically produced ethanol (Andrietta

et al., 2007). This figure had, however, dropped to around 40 percent in

2008 (Goldemberg, 2008) as a result of the crude oil price dropping sharply

during the mid-1980s and being kept low until the beginning of the 21st

century.

A Swedish ethanol program was also initiated as a result of the oil crises, but

focused mainly on research and development around softwood as a feedstock

for production of bioethanol. During the period from 2001 to 2011 Sweden

would experience a large increase in the number of cars using E85 (a

combination of 85 percent ethanol and 15 percent gasoline) as a fuel, as the

E85 car fleet increased from 717 to 229,400 cars (BAFF, 2012), which

comprises around 5 percent of the total car fleet. Ethanol in Sweden is,

however, not only used as blend with gasoline in E10 or E85 but can also be

used for buses as ED95, a diesel substitute consisting of 95 percent

anhydrous ethanol and 5% ignition improver (Wikström et al., 2011).

Incentives for ethanol fuel

The main driving force behind any change or transition of any kind is usually

a perception that the change will lead to some type of improvement in

relation to the original starting point (Nordlund et al., 2012, p.85-96). This

was clearly evident for the energy and transportation fuel sector during the

oil crisis in the 1970s. The shift of interest towards ethanol during that time

was mainly an attempt to secure the wellbeing of domestic economies and

realize a change for the better. Since then, many different options of

alternative energy have been proposed and developed, also in the

transportation sector. Some commonly mentioned driving forces behind

7

these changes are energy security and independence from oil imports,

economic stability, and higher environmental standards.

Not everyone does, however, agree that ethanol is a good option as a

transportation fuel. One objection which is usually raised against the current

production of conventional bioethanol as a transportation fuel is the claim of

a possible negative net energy value (NEV) (Pimentel, 2001, p.159-171 and

2003; Pimentel and Patzek, 2005). Different definitions of NEV may be

found in the literature; either as the difference between the energy contained

in the produced ethanol fuel and the total energy input that have been used

to produce the fuel, both fossil and renewable (Khatiwada and Silveria,

2009), or as the difference between the energy contained in the produced

ethanol fuel and the fossil energy inputs used (Shapouri et al., 2002;

Pimentel 2001, p.159-171 and 2003). What is included in the calculations of

NEV may also differ between studies, factors that need to be accounted for

when comparing NEV values from different studies. For instance,

calculations of NEV may include (Pimentel and Patzek, 2005) or exclude

(Shapouri et al., 2002) the energy required to distribute the fuels. The

discussion of NEV is thus somewhat complicated but is important to take

into account while considering incentives for bioethanol as a fuel. Regardless

of factors that are included or excluded from the calculations, values lower

than 1.00 NEV are considered negative, as it requires more fossil energy to

produce the fuel than the total energy of the produced fuel. This statement

of negative NEV for ethanol should be seen in relation to the NEV’s of fossil

gasoline and diesel. Gasoline and diesel do indeed have negative NEV's of

0.81 and 0.83, respectively, when considered on a life-cycle basis (Sheehan

et al., 1998; Hammerschlag, 2006). Oil companies do evidently produce

these fuels despite their negative NEV, mainly because the fuels are part of a

bigger picture in which they are co-products with other more valuable

products. The statement of negative NEV for ethanol does, however, hold

true only if ethanol production itself is considered without any other co-

products in the value chain being produced. When co-products are included

and considered, the contention of a negative NEV does not hold true for

ethanol, while the NEVs of gasoline and diesel are still negative, according to

several different studies (Wang et al., 1999; Shapouri et al., 2002; Kim and

Dale, 2005). A report from the United States Department of Agriculture

(USDA), based on corn as the feedstock, shows that the production of

ethanol from corn may indeed have a positive net energy value as high as

1.67 (Shapouri et al., 2002), i.e. yielding 67 percent more energy than it takes

to produce it. Production of advanced bioethanol from sugarcane bagasse

has an even higher NEV surplus, with output:input ratios being as high as

11.2 (Macedo et al., 2004). When the NEV is taken into consideration, there

is thus an incentive for ethanol as a fuel in the transportation sector.

8

Environmental benefits with the use of ethanol compared to the use of fossil

fuels, such as gasoline and diesel, should also be considered when the

incentives for using ethanol as a transportation fuel are discussed. A report

from the Argonne National Laboratory suggests that a significant reduction

(18-29 percent) of the GHG emissions can be achieved with the use of corn

ethanol blends (E10 or E85) as transportation fuel (Wang, 2005). If

cellulosic ethanol is used the reduction is calculated to be between 86-91

percent (Wang et al., 2007). Another important environmental benefit with

the use of ethanol blends is the reduction of potentially toxic and

carcinogenic aromatic hydrocarbons, such as benzene, in the transportation

fuels (Moreira and Goldemberg, 1999). Furthermore, it has also been

suggested that the use of ethanol blends (both E10 and E85) can

substantially reduce the emissions of hydrocarbons, carbon monoxide (CO),

particulate matter and volatile organic compounds (VOC) in comparison to

the emissions caused by gasoline and diesel (Niven, 2005). The use of

ethanol blends may potentially increase the emissions of acetaldehyde with

up to 100 percent (E10) (Poulopoulos et al., 2001) or up to 27 times (E85)

(Niven, 2005). The tailpipe emissions of acetaldehyde can, however, be

significantly reduced with the use of a catalytic converter (Poulopoulos et al.,

2001).

Additional benefits with ethanol as a transportation fuel include the

possibility to utilize already existing internal combustion engine technology

with a few basic modifications (DiPardo, 2000). This would allow already

existing cars to be modified and converted to ethanol use. The strongest

incentive for ethanol as a transportation fuel is, however, the possibility to

produce transportation fuels which in contrast to fossil fuels are not

produced from a finite raw material (Ragauskas et al., 2006; Saddler and

Mabee, 2007; Lynd et al., 2008). This incentive is obviously shared with

many other renewable fuels, such as DME, butanol, pentanol, biogas, and

biodiesel. Even though the production of fuels from renewable resources

may be beneficial compared to fuels from fossil raw material in the long

term, the incentive is only true if the raw material is produced in a

sustainable manner. Sustainable forestry and agriculture are for that reason

vital for the long-term prospect of ethanol production and sustainable

biofuels in general, as the Brazilian alcohol program clearly shows

(Goldemberg, 2007). Improved sustainable production of ethanol will not

only require further technological development with regard to conventional

bioethanol, but most certainly also a strong development of large-scale

production of cellulosic ethanol from lignocellulosic feedstocks (Farrell et al.,

2006; Goldemberg, 2007).

9

Bioethanol production from renewable and sustainable raw materials

Current bioethanol production

The largest producers of bioethanol on the world market are currently the

United States, Brazil, and the European Union, with a combined production

that amounts to 78.2 million metric tons of fuel-grade ethanol,

corresponding to 92.4% of the total world production. These figures can be

put in relation to those of the year 2000 when the total production of

bioethanol was 29.9 million metric tons, of which fuel ethanol made up

approximately two thirds (RFA, 2012). The total world production of

bioethanol based on starch, mainly from corn, and sugar cane has

skyrocketed in the last decade, reaching a total production of 84.6 million

metric tons in 2011 (RFA, 2012).

Conventional bioethanol

When sugarcane is used for conventional ethanol production, washing,

cutting, shredding and crushing the raw material is enough for releasing

sugar for fermentation. The use of corn requires either dry- or wet-milling of

the crop. These processes involve either grinding the corn to flour and

addition of chemicals, or soaking in water and dilute sulfurous acid to

release sugars for fermentation (Gupta and Demirbas, 2010, p.79-81). As

previously mentioned, the production of conventional bioethanol is

dominating among industrial ethanol producers in the world. The

dominating feedstock for ethanol production in the U.S. is corn, which

unfairly has been subject to strong criticism over the years. Critics accuse

corn ethanol of being a waste of resources. There are claims that corn is one

of the most energy- and water-intensive crops to grow and that its use as a

feedstock for biofuel will result in a low net energy balance and a high carbon

footprint (Mubako and Lant, 2008). However, the values look far better if

estimates of energy balances also take into consideration that fertilizers are

produced by modern processing plants, that the corn is converted in modern

ethanol facilities, and that farmers achieve average corn yields (Shapouri et

al., 2002). Positive net energy values and energy output/input ratios are vital

for the future of corn ethanol. As of 2013, there are only a small number of

biorefineries and ethanol plants in the U.S.A. that are based on other

feedstocks than corn. The total capacity of the conventional bioethanol

10

production plants was well over 90% of the total production capacity of the

American ethanol production in 2011 (RFA, 2012).

As of 2011, sugarcane ethanol represented the second largest contribution to

the world production of bioethanol. In Brazil, the sugarcane ethanol industry

was originally developed from the existing sugar industries, gaining

increased importance during the oil embargos in the 1970s. Since then,

Brazil has grown into the second largest producer of bioethanol and is an

unchallenged leader in the production of sugarcane ethanol. Production of

sugarcane ethanol has not been subjected to as strong criticism as corn

ethanol but several issues have been addressed concerning production, land

use, and the potential effects on Brazilian rainforests. However, the Amazon

region does not offer favorable conditions for commercial sugarcane

production (Goldemberg, 2008). Most of the sugarcane (around 90 percent)

used for production of ethanol in Brazil is grown and harvested in the

southeastern region around São Paulo, some 2,500 km from the Amazon

rainforest. The remaining 10 percent are grown in the northeastern parts,

which are located approximately the same distance from the Amazon

rainforest as the São Paulo region in the southeast. Only 0.2 percent of the

overall production is produced on soil belonging to the Amazon region with

mills built in the 1980s to ensure local market supplies (Goldemberg, 2008).

The benefits of sugarcane ethanol include a favorable energy balance

compared to corn ethanol. Cane husk can be burned at the ethanol plants to

provide heat for distillation and to provide electricity to run the facility. The

production process allows sugarcane ethanol plants to be energetically self-

sufficient, which lowers the overall production costs and renders sugarcane

ethanol more competitive to fossil fuels. Conventional ethanol production in

Sweden is confined to one production facility operated by Agroetanol AB,

producing ethanol from wheat (Watanabe 2013, p.4).

Cellulosic bioethanol

Production of cellulosic bioethanol on a large scale in general consists of four

subsequent steps: pretreatment, hydrolysis, fermentation and

distillation/purification (Mosier et al., 2005) (Figure 1). Production

processes are usually raw-material dependent. Cellulosic ethanol, also

known as second generation ethanol, is an advanced biofuel that can be

produced from a wide range of lignocellulosic feedstocks and represents an

essentially untapped source of bioethanol (IEA, 2011). Production of

bioethanol from lignocellulosic materials is considerably more complicated

and challenging than production of conventional bioethanol utilizing

feedstocks containing sugar, such as sugarcane, and starch, such as corn. The

11

benefits of cellulosic ethanol include raw material flexibility, lower raw

material costs, higher net energy values, no major effects on the world

commodity prices of food and animal feed, and no competition for

agricultural land or water used for food and fiber production (Fargione et al.,

2008; Searchinger et al., 2008; Sims et al., 2010). Despite the benefits of

cellulosic ethanol compared to conventional bioethanol, the technology has

not yet become commercial to the same extent as conventional bioethanol.

The main reasons for this are connected to capital costs and technological

issues. The main challenge with cellulosic ethanol is the conversion of the

recalcitrant raw material to fermentable sugars in an efficient and

economically viable way. Some commercial production facilities do,

however, exist and notable commercial European cellulosic ethanol plants

(in November 2013) include Dong Inbicon in Denmark and Beta

Renewables' Crescentino plant in Italy, which operate on agricultural

residues and energy crops as the raw material. Full-scale cellulosic ethanol

production from woody biomass in Europe includes ethanol produced

mainly from spent sulfite liquor (SSL) at Domsjö Fabriker in Sweden and at

Borregaard in Norway, both of which operate on softwood (Sànchez I Nogué

et al., 2012).

Lignocellulosic feedstocks

Lignocellulosic biomass has a complex structure and composition. The main

constituents are cellulose (38–50 %), hemicellulose (23–32 %), and lignin

(10–25 %) (Hu and Ragauskas, 2012). Cellulose is considered to be the most

abundant organic polymer in nature, making it an almost unlimited resource

for production of biofuels and biomaterials. The cellulose polymer is linear

and consists of β-D-glucopyranosyl subunits linked by 1→4 glycosidic bonds

(Pu et al., 2007). The extensive cellulose chains are in turn bundled together

in larger microstructures, i.e. microfibrils, which in turn are bundled

together to create cellulose fibers. The cellulose fibers are stabilized and held

together mainly by hydrogen bonds arising from interaction between

hydroxyl groups on the cellulose chains (Klemm et al., 2005). The hydrogen

bonds contribute to create a crystalline structure of the cellulose

macromolecule (Pu et al., 2007; Hu and Ragauskas 2012). The highly

crystalline structure of cellulose is suggested to have a negative effect on

enzymatic hydrolysis of cellulose to fermentable sugars (Hu and Ragauskas,

2012). The cellulose macromolecule does, however, also consist of areas with

lower order of aggregation, i.e. amorphous regions (Klemm et al., 2005). The

amorphous regions of the cellulose molecule are considered to be more

reactive and thus less difficult for enzymes to hydrolyze to fermentable

sugars (Hu and Ragauskas, 2012).

12

Figure 1. Schematic overview of an advanced biofuel SHF process coupled with potential on-site enzyme

production, as envisaged in this thesis work.

In contrast to cellulose, hemicellulose is not a homogeneous macromolecule.

The hemicellulose backbone consists mainly of different types of sugar

residues of five- or six-carbon sugars, i.e. arabinose, xylose, galactose,

glucose, and mannose. Hemicellulose is also frequently acetylated and

employs side-chain groups, such as uronic acids and 4-O-methyl esters (Hu

and Ragauskas, 2012). The chemical composition of hemicellulose and the

ratios of sugar residues are strongly correlated to the type of lignocellulosic

material. In hardwood and herbaceous plants, the hemicellulose is mainly

composed of glucuronoxylan and to a lower degree of glucomannan.

Hemicelluloses in softwoods mainly consist of galacto-glucomannan and

arabino-glucuronoxylan (Vogel and Jung, 2001; Pu et al., 2007). In contrast

to cellulose, hemicellulose is mainly amorphous, which results in higher

reactivity, i.e. it is easier to hydrolyze than cellulose (Hu and Ragauskas,

2012).

Lignin is the third main component of lignocellulosic materials. In contrast

to cellulose and hemicellulose, it does not consist of sugar residues. Lignin is

mainly an amorphous phenolic polymer that consists of cross-linkages

between three different types of phenylpropane units, i.e. guaiacyl (G),

syringyl (S), and p-hydroxyphenyl (H) units. The monolignol alcohol

13

precursors that give rise to G, S and H units are coniferyl alcohol, sinapyl

alcohol, and p-coumaryl alcohol (Pu et al., 2007). In similarity with

hemicellulose, the composition of the lignin macromolecule is raw-material

dependent. Hardwood lignin consists mainly of guaiacyl and syringyl units

and a smaller portion of p-hydroxyphenyl units, while softwood lignin is

mainly composed of guaiacyl units, with a smaller quantity of p-

hydroxyphenyl units. In contrast to woody materials, lignin in herbaceous

plants is composed of all three types of lignin units, i.e. guaiacyl, p-

hydroxyphenyl, and syringyl units, as well as p-hydroxycinnamic acid units

(p-coumaric acid, ferulic acid, and sinapic acid) (Pu et al., 2007; Hu and

Ragauskas, 2012).

The chemical bonds between cellulose, hemicelluloses and lignin that make

up the recalcitrant structure of lignocelluloses consist of both hydrogen

bonds and covalent bonds. The bonding between unbranched hemicellulose

and linear cellulose microfibrils is believed to arise from hydrogen bonds,

while the side chains of branched hemicelluloses are believed to form

covalent bonds with lignin, i.e. lignin-carbohydrate complexes (LCCs)

through phenyl-glycoside bonds, esters, and benzyl-ether bonds (Hu and

Ragauskas, 2012).

In addition to cellulose, hemicellulose and lignin, lignocellulosic materials

contain also smaller amounts of extractives and inorganic ash (Eom et al.,

2011). Extractives are different groups of compounds that include among

them; alkanes, fatty alcohols, fatty acids, free and conjugated sterols,

terpenoids, triglycerides, and waxes (Marques et al., 2010). The exact

functions of extractives in lignocellulosic materials are diverse and include

participation in the defense system of the plant and serving as precursors of

certain plant compounds and metabolites (Rowell et al., 2005, p.35-72). The

content of inorganic ash components, mainly comprised of inorganic metals,

is also raw-material dependent and may range from less than 1% in woody

biomass up to 15% in some forest residues and herbaceous plants (Eom et

al., 2011).

Table I summarizes the composition of different lignocellulosic materials

included in the studies (Papers I-VI). The different materials include

softwood (Norway spruce), agricultural residues (sugarcane bagasse), and

fiber residues from softwood (sulfite pulp) and from a mixture of spruce and

hardwood (mainly birch) (Kraft pulp) (Table I). As evident from the data in

Table I, the cellulose content was rather similar for Norway spruce and

sugarcane bagasse, although these materials are biologically diverse and

were harvested at different time periods and in different regions. The main

differences was the hemicellulose content (Table I). Hemicellulose of

14

sugarcane bagasse mainly consists of xylan (≥20%), while spruce

hemicellulose mainly consists of mannan (≥11%). As expected, Norway

spruce contained more lignin than sugarcane bagasse (Table I). As

previously mentioned, the composition of lignin from different types of

plants may also differ.

a Compositional analysis was performed by MoRe Research, Örnsköldsvik, Sweden. b Not

detected (N.D.).

The waste fiber sludges that were used originated from sulfite pulping of

spruce (Figure 2) and kraft pulping of a mixture of spruce (85%) and birch

(15%). This difference was quite evident in the composition of the waste fiber

sludges, as the sulfite-process sludge contained mainly glucan (90%), while

the kraft-process sludge contained lower amounts of glucan (≥66%) but also

xylan (≥15%) originating mainly from the birch that was used in the process.

The lignin content was much lower in the waste fiber sludges than in spruce

Table I. Compositional content of lignocellulosic materials used for studies

presented in this thesis.a

Lignocellulosic

material

Glucan

(%)

Xylan

(%)

Mannan

(%)

Galactan

(%)

Arabinan

(%)

Lignin

(%) Reference

Norway

spruce 41 5 11 2 1 29 Paper I

Sugarcane

bagasse 38 20 1 1 2 24 Paper I

Norway

spruce 42 5 12 2 1 28 Papers

III, VI

Sugarcane

bagasse 42 22 1 1 2 21 Paper III

Waste fiber

sludge

Kraft pulp

66 17 N.D.b N.D. N.D. 1.2 Paper IV

Waste fiber

sludge

Sulfite pulp

90 2 N.D. N.D. N.D. 0.8 Paper V

Waste fiber

sludge

Kraft pulp

69 15 N.D. N.D. N.D. 3.5 Paper V

15

and sugarcane bagasse. This is due to that they originate from pulping

processes in which the lignin is removed. These raw material compositions

correlate well to ratios reported in reviews of the literature of the area (Sun

and Cheng, 2002)

Figure 2. Waste fiber sludge (used in the study of Paper V) at Domsjö Fabriker AB, Örnsköldsvik, Sweden.

Pretreatment of lignocellulosic feedstocks

Efficient hydrolysis of cellulose requires the use of a pretreatment step in

order to open the structure of the lignocellulose and make the material

accessible to hydrolytic enzymes. Pretreatment typically includes both

mechanical (i.e. cutting, chipping or milling of the biomass to reduce particle

size) and thermochemical steps. Table II summarizes commonly used

pretreatment methods for lignocellulosic materials and their primary mode

of action. Alkaline pretreatment involves different types of strong alkali, such

as sodium hydroxide, potassium hydroxide, or ammonium hydroxide. The

raw material is soaked in alkali and then heated (Park and Kim, 2011; Chen

et al., 2013 ). Pretreatment with alkali is believed to solubilize lignin and

hemicellulose, and break the bonds between lignin and carbohydrate

polymers, thus making the raw material more accessible to further

degradation (Galbe and Zacchi, 2007). Ammonia is used for pretreatment at

elevated temperatures as in ammonia recycled percolation (ARP) or as in

16

ammonia fiber explosion (AFEX) (Dale et al., 1996; Kim et al., 2003). The

mechanisms of AFEX and ARP are, however, distinctively different as ARP

solubilizes lignin while AFEX does not. AFEX is believed to disrupt and alter

the structures of both lignin and hemicellulose, thus making the

lignocellulose more accessible to further degradation (Galbe and Zacchi,

2007).

Biological pretreatment may involve either microorganisms, such as white-

rot or brown-rot fungi, or enzymes from microorganisms that are involved in

breakdown of lignin and disrupt the structure of the lignocellulosic material

(Lee et al., 2008; Ray et al., 2010). The rate of biological processes is,

however, quite slow and the microorganisms have a tendency to consume

parts of the cellulose and hemicellulose that could instead be used for

ethanol production (Galbe and Zacchi, 2007). Carbon dioxide treatment,

which involves the use of supercritical CO2, and steam explosion are similar

to AFEX from a technical point of view, as they are performed at elevated

Table II. Overview of pretreatment methods used within cellulosic

bioethanol production.

Pretreatment method Primary mode of action Reference(s)

Alkali Delignification and hemicellulose

removal

Park and Kim, 2011; Chen et

al., 2013

Ammonia (ARP

and AFEX)

Delignification and decrystallization

of cellulose

Dale et al., 1996; Kim et al.,

2003

Biological Delignification Lee et al., 2008; Ray et al.,

2010

CO2 Explosion Hemicellulose removal and

decrystallization of cellulose Narayanaswamy et al., 2011

Dilute Acid Hemicellulose removal Larsson et al., 1999a; Saha et

al., 2005

Hydrothermolysis Hemicellulose removal Jung et al., 2013; Castro et

al., 2013

Ionic Liquids Carbohydrate and lignin dissolution Sun et al., 2009; Karatzos et

al., 2012

Organosolv Delignification and hemicellulose

removal

Pan et al., 2006; Martín et

al., 2011

Steam Explosion Hemicellulose removal Chandra et al., 2007; Horn et

al., 2012a

17

temperatures and pressures (Narayanaswamy et al., 2011; Chandra et al.,

2007; Horn et al., 2012a). In contrast to AFEX, both CO2 explosion and

steam explosion tend to hydrolyze the hemicellulose while also disrupting

the structure of lignin (Galbe and Zacchi, 2007; Narayanaswamy et al.,

2011). Hydrothermolysis, also known as hot water treatment, utilizes water

at elevated temperatures and is rather similar to dilute-acid pretreatment, as

it primarily removes the hemicellulose from the raw material, while the

structure of lignin is only modified (Larsson et al., 1999a; Saha et al., 2005;

Jung et al., 2013; Castro et al., 2013).

Organosolv treatment involves the utilization of alcohols (usually primary

alcohols, which could be methanol, ethanol, or butanol) or acetone to

solubilize lignin and hydrolyze hemicellulose, leaving the cellulose exposed

for further hydrolysis (Pan et al., 2006; Martín et al., 2011). Ionic liquids

have the capability to solubilize both lignin and cellulose and allow the

cellulose to be extracted with the addition of water to the ionic liquid (Sun et

al., 2009; Karatzos et al., 2012). High costs and difficulty to recover the ionic

liquids are currently limiting factors for the use of ionic liquids in large scale.

The suitability of the different pretreatment methods is to some extent

related to the physical and chemical structure of the raw materials.

Somewhat simplified, it could be stated that acidic methods and high

pressure methods are suitable for recalcitrant materials such as softwood

and hardwood. Alkaline methods and more neutral methods, such as

hydrothermolysis, are more suitable for materials with lower lignin contents,

such as agricultural residues and herbaceous plants (Galbe and Zacchi,

2007). Methods that are potentially cost-effective and could be included in

bioethanol plants for production of cellulosic ethanol include steam

explosion, dilute-acid pretreatment, alkali pretreatment, and organosolv

pulping (Mosier et al., 2005). Hydrothermolysis is also currently employed

in large-scale facilities, such as Dong Inbicon in Denmark and Beta

Renewables' Crescentino plant in Italy.

Most part of the cellulose, which is the predominant polysaccharide in

lignocellulose, is left intact after the pretreatment. Degradation of cellulose

to glucose can be achieved through treatment with the use of concentrated

acid, diluted acid, or hydrolytic enzymes. Several of the previously described

pretreatment techniques aim to open up the structure of cellulose and make

it accessible for further degradation. Cellulose is not structurally a

homogeneous polymer, as it consists of both crystalline and amorphous

regions, which has an impact on hydrolysis of the cellulose by enzymes or

chemicals (Den Haan et al., 2007; Ciolacu et al., 2011). The crystallinity of

cellulose is a major obstacle for hydrolysis to glucose. The crystalline

structures arise from the formation of hydrogen bonds between long

18

cellulose polymer chains (Ciolacu et al., 2011). Addition of concentrated

sulfuric acid (70 percent) to cellulose at low temperature and atmospheric

pressure disrupts the hydrogen bonds and produces cellulose in non-

crystalline state. Wyman et al. (2005, pp.995–1033) suggest that

concentrated sulfuric acid at concentrations of around 75% used at moderate

temperatures may yield as high sugar yields as those achieved with

hydrolytic cellulase enzymes. The use of low temperature and atmospheric

pressure minimizes degradation of the fermentable sugar while allowing the

use of low-cost material for containers and piping. Challenges to make the

technology economically viable include long hydrolysis times and costs

connected with separation of the acid from the sugar stream.

Dilute-acid hydrolysis involves high pressure and high temperature with

addition of one or a few percent of sulfuric acid. The process can either be

carried out in two-stages or in a single stage in a continuous-flow reactor.

However, the high temperatures (often above 200°C) which are required to

achieve cellulosic hydrolysis also lead to decomposition of sugars and lignin

(Wyman, 2005). Along with high pressure it may also result in equipment

corrosion, which is known to increase with increased pressure and

temperature. The sugar degradation reduces the sugar yield, but also leads to

formation of inhibitory substances and other degradation products that may

decrease the efficiency of the fermenting microorganisms. Furthermore,

overall glucose yields with dilute-acid hydrolysis are considerably lower than

with concentrated-acid hydrolysis and do usually not exceed 50-60 percent

of the theoretical value (Wyman et al., 2005).

Enzymatic hydrolysis is a synergetic multi-step reaction, which involves

several distinctly different hydrolytic cellulases and hemicellulases as well as

non-hydrolytic oxidative enzymes. These enzymes include mainly different

types of endoglucanases, exoglucanases/cellobiohydrolases, β-glucosidases,

xylanases, mannanases, and lytic polysaccharide monooxygenases (LPMO)

(Yang et al., 2011). There are several advantages with enzymatic hydrolysis

compared to chemical hydrolysis. The technique may be conducted under

mild conditions, leading to low degradation of sugars to toxic compounds,

low utility costs, no corrosion of equipment, no lignin breakdown, and high

reaction rates (Duff and Murray, 1996). The primary benefit with utilization

of hydrolytic enzymes is, however, the high monosaccharide yields (Yang et

al., 2011). Commercial enzyme preparations, which catalyze the

biodegradation of cellulose and hemicelluloses, usually consist of an

assortment of the different hydrolytic and oxidative enzymes previously

mentioned. The amount of lignin in the raw material is important for the

enzyme activity, as lignin contributes to prevent adsorption of cellulases to

cellulose by decreasing the accessibility and by causing catalytically

19

unproductive binding of cellulolytic enzymes. Several different factors affect

hydrolysis of lignocellulosic material by enzymes; temperature, pH,

pretreatment method, enzymatic activity, and substrate concentration. The

cellulase activity is inhibited by cellobiose and, to a lesser extent, by glucose

(Sun and Chang, 2002). This problem can, however, be dealt with by several

approaches. These include the use of high concentrations of enzymes in the

process, simultaneous addition of β-glucosidase during ongoing hydrolysis,

and removal of glucose during hydrolysis (Sun and Chang, 2002).

Fermentation processes

The fermentation process, in which sugars are converted to ethanol, is

carried out by microorganisms such as bacteria, yeast or filamentous fungi

(Lin and Tanaka, 2006). Many different types of bacteria have been shown to

possess the ability to produce ethanol as their primary fermentation product,

such as Escherichia coli (Geddes et al., 2011), Zymomonas mobilis

(Yamashita et al., 2008), several Clostridium spp., among them C.

thermocellum (Ellis et al., 2012), C. acetobutylicum (Nölling et al., 2010), C.

phytofermentans (Jin et al., 2012), C. cellulolyticum (Li et al., 2012), as well

as several Spirochaeta spp. and Klebsiella spp. (Lin and Tanaka, 2006). The

perceived high efficiency displayed by Z. mobilis is due to an inherent

deficiency that the microorganism possesses; the ability to utilize glucose for

production of ethanol anaerobically through the Entner–Doudoroff (ED)

pathway. As the ED pathway yields half the amount of ATP per glucose of the

commonly employed Embden-Meyerhof (EM) pathway, it usually results in

that Z. mobilis tends to produce ethanol rather than other metabolites,

resulting in the perception of high ethanol yields (Conway, 1992; Lin and

Tanaka, 2006). Bacteria that naturally possess the ability to produce ethanol

are usually metabolically limited to glucose, unless genetically engineered to

utilize other carbon sources derived from lignocellulose (Lin and Tanaka,

2006). This has resulted in increasing interest to genetically modify bacteria

to utilize other sugars than glucose, most notably xylose (Dien et al., 2003).

Bacterial fermentation is not of major use in current industrial production of

ethanol. Fermentation in industrial scale makes use of yeasts, primarily

Saccharomyces cerevisiae. Other yeast species have also been used for

ethanol production, e.g. Scheffersomyces (Pichia) stipitis (Lee et al., 2011)

and Hansenula polymorpha (Grabek-Lejko et al., 2011). Yeasts are usually

more metabolically diverse when it comes to ethanol fermentation. For

instance, wild-type S. cerevisiae has the capability to ferment hexoses, such

as fructose, glucose, mannose, and galactose, as well as di-saccharides, such

as sucrose and maltose, to ethanol (van Maris et al., 2006). The theoretical

net reaction in an anaerobic fermentation performed by yeast involves the

20

production of two ethanol molecules for each glucose molecule entering the

Embden-Meyerhof pathway. Theoretically, 100 grams of glucose will give a

yield of 51.1 g of ethanol and 48.9 g of carbon dioxide. In practice, the actual

yield will be less since the microorganism will use some of the glucose for

biomass production (Badger, 2002).

Industrial fermentation for production of bioethanol is usually performed

either as a separate hydrolysis and fermentation (SHF), together with

hydrolysis in a simultaneous saccharification and fermentation (SSF)

[sometimes denoted as simultaneous saccharification and co-fermentation

(SSCF)] or as Consolidated Bioprocessing (CBP). In SHF, the separation of

the hydrolysis and the fermentation steps allows the use of optimal

conditions for both enzymes and microorganisms. Benefits of SHF also

include the possibility to separate the liquid and the solid fractions before

the fermentation, which may facilitate recirculation of the fermenting

microorganism and thereby lower the costs. One major issue with SHF is the

accumulation of sugar during the hydrolysis (Sun and Chang, 2002). As

monomeric, dimeric and oligomeric sugars inhibit enzymatic hydrolysis, the

hydrolysis rate will slow down with increasing sugar concentrations. This

problem has, however, become the focal point of many enzyme producers,

who are developing enzymes which tolerate higher product concentrations

and higher optimal temperature. One additional drawback with SHF

compared to SSF can also be found with regard to compounds that inhibit

the fermentation, as this seems to affect an SHF design more than when the

hydrolysis is performed together with the fermentation (Öhgren et al., 2007).

In SSF, enzymes are mixed together with yeast and pretreated lignocellulose

(Olofsson et al., 2008). There are several advantages with SSF compared to

SHF. The simultaneous presence of yeast and enzymes potentially results in

higher ethanol yields due to lowered feedback inhibition of enzymes by

sugars. The ethanol that is produced in the vessel also lowers the risk of

contamination from other microorganisms. There are, however, also

drawbacks with SSF. It is not possible to have optimal temperature and pH

for both enzymes and microorganisms. Enzymatic hydrolysis is usually

carried out at temperatures between 45 and 60°C. Yeast performs best at 30-

37°C, while higher temperatures may render the yeast inactive. SSF is thus

usually performed at temperatures of around 30-35°C. Higher

concentrations of ethanol may also be inhibitory to hydrolytic enzymes,

which make it desirable to remove the ethanol continuously from the

bioreactor.

Consolidated Bioprocessing (CBP) is a more recent technique for conversion

of biomass to ethanol. CBP combines enzyme production, hydrolysis and

21

fermentation into one single step instead of two or three (Lynd et al., 2002

and 2005; Olson et al., 2012). The aim is to develop genetically engineered

microorganisms that produce cellulases and that have the capability to

ferment the pentose sugars xylose and arabinose to ethanol (Olson et al.,

2012). This allows for a more efficient conversion of the lignocellulosic

material with higher product yields, rates and improved stability of cultures

and strains, while also saving on costs by allowing the hydrolysis and

fermentation step to be performed by the same microorganism (Lynd et al.,

2002; Olson et al., 2012). The microorganisms used for CBP do not yet

supply all the enzymes that are necessary for the hydrolysis of cellulose, so

there is still a need for supplementing the process with externally produced

enzymes.

22

Fermentation inhibition

Formation of compounds that inhibit enzymes or the growth of

microorganisms is a major problem associated with production of cellulosic

bioethanol. Inhibitors cause problems for enzymatic hydrolysis (Ximenes et

al., 2011) and for fermentation processes based on various types of

microorganisms including bacteria, yeasts, and filamentous fungi (Pienkos

and Zhang, 2009; Paper VI). The inhibitory compounds can be divided into

aromatic compounds, small aliphatic acids, furan aldehydes, and inorganic

compounds (Jönsson et al., 2013). Dilution, removal or conversion of these

compounds is needed to obtain efficient hydrolysis and fermentation. To

make cellulosic ethanol production economically and environmentally

sustainable, it is preferred to work with undiluted systems that give high

product concentrations and to recirculate process water, which may

potentially lead to accumulation of inhibitory compounds if they are not

converted or removed.

Small aliphatic acids

Influx of acids into the cell causes a drop in the intracellular pH, which is

neutralized by the plasma membrane protein ATPase. The plasma

membrane ATPase pumps protons out of the cell in a process that requires

ATP. In order to maintain the intracellular pH additional ATP must be

generated, which under anaerobic conditions is achieved by increased

ethanol production at the expense of biomass formation. The anion

accumulation theory suggests that equilibrium-driven diffusion of

undissociated acids across the cell membrane will occur when weak aliphatic

acids are present in the fermentation medium. After diffusion across the cell

membrane the weak acids dissociate in the higher pH environment of the

cytosol, resulting in generation of protons and acid anions (Piper et al.,

2001). As the equilibrium is concentration dependent, the diffusion of the

undissociated acid will be a function of the pH and will occur until

equilibrium is reached. Furthermore, accumulation of acids may also result

in free-radical production, which results in severe oxidative stress in S.

cerevisiae (Piper et al., 2001). Studies on the effects of acetic acid on S.

cerevisiae showed that the ethanol yield increased at low concentrations of

acids (<100 mM) (Larsson et al., 1999a). Higher concentrations (roughly

>1oo mM) were, however, shown to have inhibitory effects (Larsson et al.,

1999a). Several other organic acids, which occur in lower amounts, have also

been detected and quantified. These include malonic acid, maleic acid, cis-

aconitic acid, succinic acid, glutaric acid, and itaconic acid (Du et al., 2010).

23

Furan aldehydes

Furfural is formed from pentose sugars, such as xylose and arabinose, under

harsh pretreatment/hydrolysis conditions with high pressure and high

temperatures (Jönsson et al., 2013). An NADH-dependent alcohol

dehydrogenase (ADH) is responsible for reduction of furfural to furfuryl

alcohol by yeast (Wahlbom and Hahn-Hägerdal, 2002). The production of

glycerol, which is usually produced by yeast in order to regenerate NADH,

was found to be significantly lower in the presence of furfural (Palmqvist et

al., 1999). This indicates that the inhibition mechanism involving furfural is

affecting the regeneration of NADH, and suggests that reduction of furfural

to furfuryl alcohol by an NADH-dependent ADH has a high priority

(Palmqvist et al., 1999). The same study suggested that moderate amounts of

furfural may actually be beneficial for ethanol production by S. cerevisiae. As

furfural reduction is preferred to glycerol production as a redox sink, this

would leave more glucose to be available for ethanol production.

5-(Hydroxymethyl)furfural (HMF), which is formed by dehydration of

hexose sugars at high temperatures, is suggested to have similar effects on

the fermentation as furfural (Liu et al., 2004). Biconversion of HMF is

considered to be catalyzed by an NADPH-dependent enzyme rather than an

NADH-dependent, which leads to stronger inactivation of cell replication

than is caused by furfural (Wahlbom and Hahn-Hägerdal, 2002).

Aromatic compounds

During some forms of pretreatment, lignin is partially degraded to aromatic

compounds of smaller size, many of which are phenolics. In contrast to

aliphatic acids and furan aldehydes, where a few compounds are

predominant, there is a variety of aromatic compounds with different

functional groups in lignocellulosic hydrolysates. Studies of specific aromatic

compounds have shown that they have widely different inhibitory effect on

yeast growth, ethanol productivity, and ethanol yield (Larsson et al., 2000;

Paper II). Due to the variety of compounds within this group of inhibitors, it

is difficult to suggest one general mechanism of action. However, their

toxicity for the fermenting microorganism may be related to specific

functional groups of the molecules (Jönsson et al., 2013). Larsson et al.

(2000) showed that compounds that structurally are very similar, such as

coniferyl alcohol and coniferyl aldehyde as well as benzoquinone and

hydroquinone, had very different toxicity for yeast. The molecular structures

of some of these compounds are shown in Figure 3. The results presented in

24

Larsson et al., (2000) support the notion that functional groups of aromatic

compounds are important for their toxicity. Several studies have indicated

that aromatic compounds of low molecular weight are more toxic to yeast

than larger aromatic compounds (Jönsson et al., 1998; Larsson et al., 2000).

Inhibitory effects of some aromatic compounds commonly found in

lignocellulosic hydrolysates are suspected to be connected to interference

with mitochondrial function and cell respiration, particularly interference

with the proton transport chain across the inner mitochondrial membrane

(Jönsson et al., 2013).

Strategies for dealing with fermentation inhibition problems

Many different approaches to deal with fermentation inhibition have been

developed over the years. These include the use of chemical additives, liquid-

liquid extraction, microbial treatments, enzymatic treatments, heating and

vaporization, and liquid-solid extraction (Jönsson et al., 2013). Some have

been found to be more efficient than others, and some are more realistic

from an economical point of view for use in a larger scale. One approach is to

reduce the severity of the pretreatment, since a more severe pretreatment

usually results in formation of more inhibitors. However, a more severe

pretreatment usually opens up the raw material more efficiently, allowing for

a more efficient hydrolysis of the raw material to fermentable sugars.

Furthermore, if cellulosic bioethanol is produced in an industrial scale, the

sugar yield must be prioritized in order to achieve a high overall bioethanol

yield. It is therefore difficult to lower the severity of the pretreatment and

maintain a high product yield.

There are, however, other approaches than reducing inhibitor formation that

can be employed to deal with inhibition problems. These approaches include

process design, strain adaption, genetic engineering and detoxification or

conditioning of lignocellulosic hydrolysates.

Adaptation and strain selection

Adaption of microbial strains to industrial media has been studied

extensively in the past. This is usually performed by exposing the microbe to

one or a few of the fermentation inhibitors that are usually present in

lignocellulosic hydrolysates or directly to the hydrolysates themselves

(Keating et al., 2006; Martín et al., 2007; Landaeta et al., 2013). The method

is efficient but time consuming and there are issues connected to lower

25

ethanol yields and strain stability. Strain selection is also an integral part of

overcoming inhibition of lignocellulosic hydrolysates as different strains of

the same microorganism can have very different tolerance towards inhibitory

compounds (Martín and Jönsson, 2003).

Genetic engineering of microorganisms

While strain adaption results in more tolerant strains over time, it is time

consuming, perhaps also medium specific, and in part restricted by the

genetic and physiological constitution of the microorganism. Strain adaption

may work well if the medium which is used in the process is similar over

time, which was the case with S. cerevisiae strains isolated from a spent

sulfite liquor fermentation plant (Lindén et al., 1992; Sànchez I Nogué et al.,

2012). While adaption of strains may require several years of intensive work

and may lead to lower overall bioethanol yields, the introduction of certain

genes through genetic engineering is usually much faster and may lead to a

more fundamental change of the properties of the cell.

Genetic modifications may be introduced by heterologous or homologous

expression of selected genes. Heterologous expression aims to introduce a

gene from a foreign species. The new gene may encode a protein with a

function that is completely new for the engineered microorganism, which

can gain a resistance mechanism that it does not naturally possess.

Homologous expression aims to overexpress an already existing gene of the

microorganism, which results in an amplification of the tolerance through a

mechanism that the microorganism already possesses. Homologous

expression has been used to increase tolerance against acetic acid

(Hasunuma et al., 2011a; Wright et al., 2011), formic acid (Hasunuma et al.,

2011a and 2011b) as well as tolerance for several different aromatic

compounds (Larsson et al., 2001a). Heterologous expression of Trametes

versicolor laccase has been shown to improve the fermentation rate of dilute

acid hydrolysates compared to a reference strain (Larsson et al., 2001b). It is

important to note when considering studies of strain engineering if the

engineered strains achieve tolerance towards model compounds in synthetic

media which contain the particular compound that the organism has been

engineered to withstand (Hasunuma et al., 2011a ; Wright et al., 2011) or if

tolerance was actually achieved in lignocellulosic hydrolysates (Larsson et

al., 2001a and 2001b). A potential drawback connected with genetically

engineered microorganisms is that the method may work well against a

specific inhibitor against which the microorganism has been made more

resistant but not as well for other inhibitory compounds which usually also

26

are present in lignocellulosic hydrolysate. Additional potential drawbacks

include lower bioethanol yields and strain stability over time.

Process design

There are several ways through which a process can be modified so that it

becomes less sensitive to problems with inhibition. One modification, which

is possible in laboratory experiments but seldom included in the design of

large-scale processes, is dilution of the medium. By adding a solvent to the

lignocellulosic hydrolysate, the concentration of inhibitory compounds will

be lower per volumetric unit. Diluting the media will, however, also dilute

the concentration of fermentable sugars, which is associated with higher

operating costs due to increased costs for distillation (Lin and Tanaka,

2006). An alternative to dilution that gives a similar effect is to increase the

size of the inoculum of the fermenting microorganism. As inhibition is a

matter of concentration, an increase of the number of cells which can be

exposed to the inhibitory compound will lead to that the overall inhibitory

effect per cell will be lower. This technique can be economically feasible if it

is possible to recirculate and reuse the same microorganism within the

process, as with SHF. However, if a process such as CBP or SSF is used, the

amount of solids in the fermentation vessel will render recirculation

problematic due to the difficulty to separate the solids from the

microorganism. Furthermore, increasing the inoculum to as high final

concentrations as 10 g/L (DW) (which in large scale would not be

economically attractive) may not be enough to ferment strongly inhibitory

hydrolysates (unpublished results).

Additional possibilities include using fed-batch or continuous processes.

During a fed-batch process, the fermenting microbes will at least initially be

exposed to lower inhibitor concentrations through a slow feed of the

hydrolysate to the fermentation vessel. In this manner, the medium is being

fed to the fermenting microorganism over a longer period of time and it is

allowed to consume the sugars and overcome or adapt to the inhibitors

present in the medium before additional medium is fed into the bioreactor.

It is, however, important to note that some inhibitors may be converted by

the microorganisms, such as furans and aromatic aldehydes, while small

aliphatic and aromatic acids are not and will persist and accumulate in the

fermentation process. This process may result in an efficient utilization of

fermentable sugars and can be beneficial if a pentose-fermenting

microorganism is used in the process, as even the presence of low amounts

of glucose will inhibit pentose fermentation pathways in yeast (Subtil and

Boles, 2012). By employing a fed-batch process, the glucose in the medium

27

may be depleted and allow pentose fermentation to commence before

additional sugars are supplemented. The main drawback with a fed-batch

process is that it is a more time-consuming process than batch fermentation

and that an accumulation of inhibitors, which are not converted by the

fermenting microorganism or that are chemically stable, may occur.

28

Detoxification

Detoxification of lignocellulosic hydrolysates, also referred to as

conditioning, is frequently included in both scientific studies and industrial

process schemes, but this usually depends on several different factors, such

as the type of methods used for pretreatment and hydrolysis, the

recalcitrance of the feedstock, the type and concentration of the

microorganism used in the fermentation step, and whether issues such as

economically viable product concentrations and process water recirculation

at all have been taken into consideration. In contrast to other previously

described methods for overcoming inhibition problems, detoxification aims

to remove inhibitory compounds or to alter their chemical structure.

One of the most efficient methods for detoxification of lignocelluosic

hydrolysates is treatment with anion exchangers (Larsson et al., 1999b;

Nilvebrant et al., 2001; Sárvári Horváth et al., 2004). The treatment is often

performed at a relatively high pH and may be performed in combination

with alkaline treatment, which, however, is a detoxification method in itself.

Treatment with alkali has most commonly been performed as overliming, i.e.

treatment with calcium hydroxide (Martínez et al., 2000; Sárvári Horváth et

al., 2005; Alriksson et al., 2005 and 2006; Mohagheghi et al., 2006).

However, alkali detoxification can be performed with several different forms

of alkali, such as Ca(OH)2, NaOH, and NH4OH. The pH is raised to at least 9

or 10 but possibly as high as pH 12, where it is kept for at least a few hours

with moderate heating, before it is adjusted to a pH that is suitable for

fermentation (pH 5-6). The mechanism behind alkali detoxification has not

been fully explained. Suggested mechanisms behind the resulting

improvements in fermentability after alkali detoxification include

precipitation of inhibitors with Ca(OH)2, as well as conversion of inhibitors

to less toxic compounds without precipitation, which is achieved with alkali

detoxification employing NaOH and NH4OH (Jönsson et al., 2013). Potential

drawbacks with alkali detoxification include precipitation of CaSO4 (gypsum)

when Ca(OH)2 is used, as well as potential breakdown of sugar in the

hydrolysates. The breakdown is related to the severity of the treatment, e.g.

high temperature, long treatment time and high pH (Nilvebrant et al., 2003).

When the treatment was performed at 80oC, pH 12 and at time intervals of 1

h, 4 h and 7 h, a complete degradation of sugars occurred no matter the

treatment time. Optimization of alkali detoxification has been performed in

experimental series in which different types of alkali were used and the

temperature, pH and duration of the treatments were varied (Alriksson et

al., 2006). The treatments were evaluated on basis of both maximum ethanol

yield and removal of inhibitory compounds. In this study (Alriksson et al.,

29

2006), the best results were obtained with treatment with NH4OH at 55oC

and pH 9 during 3 h. The sugar degradation was below 10%, while a major

part of the inhibitory compounds was removed (Alriksson et al., 2006).

Much lower sugar degradation (about 1%) has, however, also been reported

with considerable improvements of the fermentability when the treatment

was performed at room temperature with Ca(OH)2 at pH 10 for 1 h (Martín

et al., 2002a). Nilvebrant et al. (2003) later reported that alkali treatment of

lignocellulosic hydrolysates may be performed at temperatures below 30oC

and at a pH between 9 and 12 without the risk of sugar degradation or

formation of inhibitory compounds. These results indicate that the risk of

sugar degradation depends on variable experimental parameters and that

sugar degradation, ethanol yields, and ethanol productivity should be taken

into consideration in relation to improvements of fermentability. Sugar

degradation should therefore not be raised as a general objection against the

utilization of alkali detoxification. Alkali detoxification was chosen as the

benchmark detoxification method in studies of detoxification of

lignocellulosic hydrolysates conducted with reducing agents, as a part of the

investigations presented in this thesis. For several reasons alkali

detoxification was chosen as the benchmark method, but most importantly

for its simplicity and overall efficiency. Alkali detoxification has been shown

to result in ethanol yields as good as or better than reference fermentations

without inhibitors (Alriksson et al., 2006).

Except ion exchange and alkali treatment, other efficient detoxification

methods have been reported. These include the use of activated charcoal

(Mussatto et al., 2004) and treatment with enzymes such as laccases and

peroxidases (Jönsson et al., 1998; Larsson et al., 1999b; Martín et al.,

2002b). Of the different detoxification methods, overliming has been

estimated to be the most economically viable (Ranatunga et al., 2000).

Concerns that have been raised regarding the inclusion of alkali

detoxification in large-scale bioethanol processes include formation of

gypsum if overliming is used and that the requirement of an additional

process step would lead to high costs (von Sivers et al. 1994; Hamelinck et

al., 2005).

Detoxification with reducing agents A novel method for detoxification, based on addition of reducing agents, has

recently been developed. The results of these studies show that detoxification

of lignocellulosic hydrolysates, both softwood and sugarcane bagasse, can be

performed as a one-step treatment at low temperature, slightly acidic pH

and short reaction times, such as 25oC, pH 5.5 and 10-20 min (Papers I, II

and III). This new method is based on the addition of low concentrations of

30

inexpensive industrial chemicals, i.e. sodium dithionite, sodium sulfite and

sodium borohydride. These chemicals, which are today used in pulp and

paper industry as well as in textile manufacturing, can be added directly to

the bioreactor without any need for an additional process step (Papers I, II

and III).

Industrial applications of dithionite, which is unstable in aqueous solutions,

include bleaching of pulp. It is usually applied through dissolution in water

in which it under the presence of oxygen forms bisulfite and bisulfate,

written as an overall reaction in Equation 1. Equation 1 is the combination of

two possible reactions, written as Equations 2 and 3 (Tao et al., 2008).

Chemical loadings of dithionite in bleaching may vary, but are usually in the

range of around ten kg of chemical per ton of fiber, i.e. 1% (Suess 2010, p.

230). Bleaching with dithionite is temperature and time dependent; it may

be performed at temperatures as low as 30oC at which the bleaching process

is slow, thus the process is usually performed around 70-80oC in order to

reduce the bleaching time to 1-3 h (Friman et al., 2004; Suess 2010, p. 234).

The technical applications of sodium borohydride are numerous reduction

reactions but include among them the generation of dithionite through the

reaction of sodium borohydride, sodium bisulfate and sulfur dioxide, a

technique that is occasionally used on-site by some pulp mills (Suess 2010,

pp. 31-32). The use of sodium sulfite in pulp and paper industries has long

been used in alkaline sulfite pulping to produce cellulose pulp (Thompson et

al., 2013).

S2O42- + O2 + H2O = HSO4

- + HSO3- (1)

2 S2O42- + O2 + 2 H2O = 4 HSO3

- (2)

2 S2O42- + 3 O2 + 2 H2O = 4 HSO4

- (3)

The potential of using the reducing agents sodium dithionite and sodium

sulfite for detoxification was investigated using enzymatic hydrolysates of

Norway spruce and sugarcane bagasse (Paper I). The detoxification effects

were investigated with S. cerevisiae in both SHF and SSF configurations and

were compared to alkali detoxification as well as to synthetic sugar

references without inhibitors. The results of these experiments indicate that

additions of no more than 5-15 mM of the reducing agents are sufficient to

significantly improve the fermentability of lignocellulosic hydrolysates

(Figures 1 and 2 in Paper I). The best improvements of fermentability were

reached in experiments performed in SHF configuration with 10 mM sodium

dithionite, which resulted in ethanol productivities of 2.5 g × L-1 × h-1 for the

spruce hydrolysate and of 3.9 g × L-1 × h-1 for the bagasse hydrolysate

31

compared to 0.2 and 0.9 g × L-1 × h-1 for untreated spruce and bagasse

hydrolysates, respectively. These productivity values were higher than those

achieved with fermentations of reference sugar solutions containing similar

sugar concentrations as the hydrolysates but without any inhibitors. The

ethanol productivities of the reference fermentations were 2.0 g × L-1 × h-1

for spruce and 2.6 g × L-1 × h-1 for bagasse. Furthermore, the results were

comparable to the ethanol productivities achieved with alkali detoxification,

which were 2.8 g × L-1 × h-1 for the spruce hydrolysate and 3.5 g × L-1 × h-1 for

the bagasse hydrolysate. These results show that additions of 10 mM

dithionite 10 min prior to addition of yeast were equally efficient as

optimized alkali detoxification, e.g. treatment with NH4OH at 55oC and pH 9

during 3 h.

Treatments with 10 mM sodium sulfite also resulted in improved

fermentability compared to the untreated hydrolysates but did not reach as

high improvements as alkali or 10 mM dithionite treatments. For 10 mM

sodium sulfite, ethanol productivities of 1.2 and 2.9 g × L-1 × h-1 were

reached for the spruce and bagasse hydrolysates, respectively. The

productivity of the bagasse hydrolysate was thus higher than that of the

reference medium, which did not contain any fermentation inhibitors (Table

2 in Paper I). Further support of the effects of detoxification with reducing

agents was found in the balanced ethanol yields. The best results were

achieved with additions of 10 mM sodium dithionite, which resulted in a

balanced ethanol yield of 0.35 g ethanol per g consumed glucose and

mannose for the spruce hydrolysate and 0.28 g/g for the bagasse

hydrolysate. These results stand in a strong contrast to the results obtained

with untreated hydrolysates, which were 0.03 and 0.06 g/g for spruce and

bagasse hydrolysates, respectively. Alkali detoxification was again similar to

treatment with 10 mM sodium dithionite and the balanced ethanol yields

were 0.37 g/g for the spruce hydrolysate and 0.25 g/g for the bagasse

hydrolysate. The results from this study also show that both sodium

dithionite and sodium sulfite may be added to SSF experiments even after

the yeast has been added and still result in improved fermentability. These

improvements were evident up to 105 minutes after inoculation of the

untreated hydrolysates with yeast (Figure 5 in Paper I).

The mechanism behind the improvements that were observed in the study

presented in Paper I was investigated with the use of mass spectrometry, and

are commented below in the section on Mechanism behind treatment with

reducing agents. Before that, a general overview of the techniques and

instruments used for mass spectroscopy is presented to facilitate the

understanding of the analytical aspects of the investigation of the

mechanism.

32

In the study presented in Paper III, the effects of treatment with sodium

borohydride on the fermentability of enzymatic hydrolysates of Norway

spruce and sugarcane bagasse and the yield of the enzymatic hydrolysis were

investigated. The study also compared the effects of sodium borohydride

treatment with treatments with sodium dithionite and sodium sulfite. For

the spruce hydrolysate, improvements were observed for a concentration of

15 mM borohydride and there was no negative impact on the fermentability

at borohydride concentrations up to 40 mM, which was the highest

concentration that was added to the spruce hydrolysate. The bagasse

hydrolysate required additions of at least 23 mM borohydride and no

negative impact was found up to 47 mM, while an addition of 55 mM showed

a negative impact on the fermentability. Improvements in ethanol

productivity and balanced ethanol yields were rather similar in the range

from 16 to 40 mM for the spruce hydrolysate and from 31 to 47 mM for the

bagasse hydrolysate. The highest ethanol productivity for the spruce

hydrolysate was obtained after addition of 34 mM sodium borohydride,

where the ethanol productivity reached 0.57 g × L-1 × h-1. The ethanol

productivity of the untreated hydrolysate was only 0.05 g × L-1 × h-1. In the

experiments with the bagasse hydrolysate, the highest ethanol productivity

(0.41 g × L-1 × h-1) was obtained after addition of 39 mM sodium

borohydride. The ethanol productivity of the untreated bagasse hydrolysate

reached 0.02 g × L-1 × h-1. The highest balanced ethanol yield for the spruce

hydrolysate (0.30 g/g) was reached after addition of 22 and 34 mM sodium

borohydride. The balanced ethanol yield of untreated spruce hydrolysate was

0.02 g/g. The bagasse hydrolysate reached the highest balanced ethanol

yield (0.32 g/g) after addition of 39 mM sodium borohydride. In

comparison, the balanced ethanol yield of untreated bagasse hydrolysate was

0.02 g/g. Conditioning in YPD medium for 2 h prior to inoculation improved

the performance of the yeast.

The results presented in Paper III suggest that higher concentrations of

sodium borohydride than sulfur oxyanions may be required to obtain

satisfactory results, at least if the yeast is not conditioned (dissolved in YPD

medium for 2 h) or pre-grown before inoculation. Using conditioned yeast,

treatment with sodium borohydride was compared with treatments with

sulfur oxyanions in experiments in which 15 mM of the reducing agent was

added to the spruce hydrolysate. The ethanol productivities in these

experiments were 1.35 g × L-1 × h-1 for borohydride, 1.31 g × L-1 × h-1 for

dithionite, 0.90 g × L-1 × h-1 for sulfite and 0.45 g × L-1 × h-1 for the untreated

hydrolysate. The balanced ethanol yields were 0.29 g/g for borohydride,

0.28 g/g for dithionite, 0.19 g/g for sulfite, and o.10 g/g for untreated

hydrolysate. Thus, experiments with conditioned yeast and addition of 15

mM reducing agent indicate that the efficiency of sodium borohydride

33

treatment was in par with that of sodium dithionite and greater than that of

sodium sulfite.

The effects of treatments with sulfur oxyanions and sodium borohydride on

enzymatic hydrolysis were also investigated (Paper III). The results from

these experiments show that sulfur oxyanions had positive effects on the

enzymatic hydrolysis in the presence of pretreatment liquid from pretreated

lignocellulose. Sodium borohydride had no negative effect in the presence of

liquid from pretreated lignocellulose, but did show a negative effect on

enzymatic hydrolysis in buffered medium without inhibitors (Figure 4 in

Paper III). A negative effect in the presence of pretreated lignocellulose

would have rendered the reducing agents unsuitable for use in SSF or CBP

processes. The results also indicate that sodium dithionite and sodium sulfite

are suitable to use for both SHF, SSF and CBP processes, and that the

suspended solids concentration may be at least as high as 12.5%. This is

because sodium dithionite and sodium sulfite are less reactive chemicals

than sodium borohydride, which is mainly suitable for use in SHF processes

due to its strong reactivity (Paper III).

All of the reducing agents that were used in these studies (Papers I and III)

potentially have an anti-microbial effect, but there were no negative effects

on the fermenting microorganism at the concentrations that showed

significant improvements when used for detoxification. Sodium dithionite

and sodium sulfite could be added in concentrations up to 20 mM without

any negative effects on 2 g/L (DW) yeast (Paper I). Sodium borohydride

could be added without negative effects in concentrations of up to 40 mM

when spruce hydrolysate was used, and up to 50 mM when a slightly more

toxic bagasse hydrolysate was used (Paper III). While sodium dithionite and

sodium borohydride showed greater improvements than what was reached

with similar concentrations of sodium sulfite, it is important to note that in

aqueous solution dithionite may dissociate into bisulfite (Equation 2). The

stoichiometry of the equation shows that in aqueous solutions and in the

presence of oxygen one dithionite molecule can give a yield of up to two

bisulfite molecules. This could possibly contribute to the higher

fermentability improvements observed for dithionite compared to sulfite.

There are several aspects to be taken into consideration when discussing and

comparing detoxification methods. Besides efficiency, these should also

include safety as well as economic and technical viability in the large scale.

Several alternative methods for overcoming inhibitory conditions have

previously been mentioned: alteration of pretreatment conditions, process

design, strain adaptation, and genetic engineering. Efficient detoxification

methods, such as detoxification with reducing agents, possess several

34

advantages over the previously discussed methods. The means of altering

pretreatment conditions in order to decrease the formation of inhibitory

compounds and thereby avoid inhibitor problems can instead result in

decreased process efficiency as a result of lower sugar yields, less efficient

conversion of biomass, lesser ethanol yields and insufficient productivity.

Efficient conversions of lignocellulosic biomass through high sugar and

ethanol yields as well as high productivity are vital for a cost-efficient large-

scale production of second generation fuel-grade ethanol. Detoxification

based on reducing agents indicates advantages which allow proper

pretreatment conditions to be conducted as the method efficiently deals with

the resulting fermentation inhibitors downstream of the pretreatment.

Strain selection, adaption and genetic modification of the microorganisms

that are used in the fermentation step may to some extent be viable options.

However, results obtained with these methods that show improvements

comparable to those reached by using chemical detoxification remain to be

achieved. Many studies in this area suffer from a lack of a benchmark. The

efficiency of the approach taken can be related to the yield and productivity

obtained in reference fermentation without inhibitors. Results showing

similar ethanol yields and ethanol productivity as reference fermentations

without inhibitors have been achieved for both alkali detoxification and

detoxification with reducing agents, as reported in the studies included in

this thesis.

Treatment with reducing agents has several advantages over alkaline

detoxification, such as: (i) treatment with reducing agents is performed at

the fermentation pH, which results in cost savings due to use of less

chemicals, (ii) treatments with reducing agents are efficient already at room

temperature and elevated temperatures are not necessary as they are for

alkali detoxification (if the aim is to perform the treatment within a few

hours rather than in days), (iii) the reaction of reducing agents with

inhibitory compounds is instant and there is no need for longer reaction

times as may be required with alkali detoxification, (iv) additions of reducing

agents to hydrolysates do not result in even the slightest sugar degradation,

(v) the nature of reducing agents allows them to be added even after the

fermenting organism has been added, in the case that the fermentation

would show signs of inhibition, and (vi) the treatment with reducing agents

does not result in any formation of precipitates which may cause problems

downstream in the production process. A potential drawback with the

employment of reducing agents in detoxification of lignocellulosic

hydrolysates is the addition of sulfur-containing compounds to the

production process. Sulfur-containing compounds would, however, be added

also prior to detoxification in many pretreatment methods, such as in dilute-

acid hydrolysis with sulfuric acid or as in impregnation of the raw material

35

with sulfuric acid or sulfur dioxide before steam-explosion pretreatment.

While these pretreatments (in the low chemical loading range) employ

concentrations of 1% sulfur-containing compound per dry weight of the raw

material, the addition of sulfur oxyanions is not necessarily more than 10 to

15 mM per total content of solids in SSF experiments, i.e. 0.5-1% (Paper I).

The concentrations of sulfur oxyanions added to lignocellulosic slurries to

achieve detoxification are thus no higher than those commonly used during

some pretreatment processes. Scrubbing is a technique that can be employed

to efficiently remove sulfur-containing compounds from gases and liquids. It

is a technique that is well established in industrial settings. This can be

exemplified with efficient removal of sulfurous compounds in gases and

liquid phases from geothermal wells on Iceland (Hauksson et al., 2013).

Scrubbing techniques are also employed in the pulp and paper industry, and

may possibly be applied in biofuel production when sulfurous compounds

are added to the production process.

The main objection which is raised against detoxification in general and

alkali detoxification in particular is the necessity of introducing a separate

treatment step to carry out detoxification. This would lead to increased

construction and process costs. The objection is based on the perception that

detoxification cannot be accomplished instantly or be performed in direct

connection with the fermentation step (von Sivers et al. 1994; Hamelinck et

al., 2005). The results of the studies on detoxification with reducing agents,

which are presented in this thesis, show that this assumption no longer holds

truth. Detoxification with reducing agents can be performed as a one-step

treatment at the fermentation pH, at moderate temperature, and render an

instant improvement for the fermenting microorganism. The method can

even be performed as an in-situ detoxification in the bioreactor with both

enzymes and the fermenting microorganism present.

36

Analysis of inhibitory compounds

Analysis of monomeric sugars, ethanol, fermentation by-products as well as

inhibitory sugar- and lignin-degradation products may be performed with

different types of analytical instrumentation. High-performance liquid

chromatography (HPLC) systems equipped with refractory index detectors

(RID), UV/VIS detectors or diode array detectors (DAD) are suitable for

analysis of all the mentioned compounds (Larsson et al., 1999b; Persson et

al., 2002; Sluiter et al., 2006). The robustness and simplicity in combination

with the array of compounds that can be quantified with HPLC instruments

have made the technique widespread in laboratories working with

lignocellulose and bioethanol. High-performance anion-exchange

chromatography (HPAEC) coupled with a pulsed amperometric detector

(PAD) can be used for analysis of monomeric sugars (Nilvebrant et al., 2001;

Pedersen et al., 2011). HPAEC may also be coupled with a conductivity

detector for analysis of small aliphatic acids (Guo et al., 2013). Some of the

detectors mentioned may also be suitable for analysis of aromatic inhibitory

compounds commonly found in lignocellulosic hydrolysates, such as UV or

UV-DAD (Persson et al., 2002). The limitation with such detection methods

is that the detection and identification of compounds is based on retention

times and (with UV-DAD) absorption spectra. The analysis thus becomes

rather non-specific and the risks of overlapping peaks that absorb light in the

same wavelength are significant. Mass spectrometry (MS) is a detection

method that may overcome such limitations. MS is discussed in more detail

below, as it was extensively used in the studies presented in Papers II and

III.

Mass Spectrometry

Mass spectrometry instruments are analytical tools that provide information

regarding mass, elemental composition and molecular structure. These

instruments consist of an ion source that creates charged molecules or

molecular fragments (i.e. ions), a mass analyzer that separates the charged

ions according to their mass-to-charge ratio (m/z), and a detector that

registers the number of molecules that are present at each m/z value

(Aebersold and Mann, 2003; El-Aneed et al., 2009). A general overview of

the most commonly used instruments for MS analysis is provided as a

complement to the analytical part of the investigations of the mechanisms

behind treatment with reducing agents for detoxification of lignocellulosic

hydrolysates (Papers I and III).

37

Ionization methods

The two most well established methods of ionization in MS instruments are

chemical ionization (CI) and electron ionization (EI). Both methods were

originally developed for gas chromatography systems coupled with mass

spectrometers (GC-MS). EI creates charged ions through collisions between

analytes and a concentrated electron beam. Upon impact the molecules are

fragmented and electrically charged, i.e. ionized, and subsequently

introduced into the mass analyzer (Guilhaus et al., 2000; de Hoffmann and

Stroobant, 2007, pp. 15-17). EI does not enable analysis of intact molecules

but rather results in molecule fragments, making it unsuitable for some types

of analytical applications. Chemical ionization (CI) is considered to be softer

than electron ionization, and in contrast to EI it results in little or no

fragmentation of the analyzed molecules (El-Aneed et al., 2009). CI occurs

through reaction of sample molecules and ionization plasma that is created

inside the ion source (Guilhaus et al., 2000; de Hoffmann and Stroobant,

2007; pp. 17-18).

Electrospray ionization (ESI) is one of the most commonly used ionization

techniques for liquid chromatography systems (LC). The nature of ESI

makes it suitable for analysis of large and/or instable as well as intact

molecules, which are usually also non-volatile, an attribute that made ESI

suitable for investigations of the mechanisms behind treatments with

reducing agents (Papers II and III). ESI operates at atmospheric pressure

and with the use of heated bath gas, usually nitrogen. The sample, which is

in a liquid state, is sprayed (with the help of the bath gas from a capillary

syringe) into a strong electric field maintained in vacuum, in which the

molecules are charged and sent into the mass analyzer by the flow of the bath

gas (Aebersold and Mann, 2003; de Hoffmann and Stroobant, 2007, pp. 43-

45). The main drawback is that it does not allow very much structural

information to be obtained from the rather simple mass spectra that are

produced using this ionization method.

An alternative method to ESI for analysis of large and fragile molecules is

matrix-assisted laser desorption/ionization (MALDI). While also being

suitable for large molecules, the mechanism behind MALDI allows for the

molecules to carry a single charge rather than multiple, as with ESI. This is

beneficial, as increased charge also may result in increased instability of the

molecules. MALDI is for that reason suitable when complex mixtures of

proteins, peptides, DNA, RNA or highly labile compounds are to be analyzed

(Guilhaus et al., 2000; Aebersold and Mann, 2003; de Hoffmann and

Stroobant, 2007, pp. 33-35).

38

There are also several other methods that may be used for ionization in mass

spectrometry. These include fast atom bombardment (FAB), field desorption

(FD), surface-activated laser desorption ionization (SALDI),

desorption/ionization on silicon (DIOS), direct analysis in real time (DART),

and secondary ion mass spectrometry (SIMS) (de Hoffmann and Stroobant,

2007). These methods are, however, not as common as the ones previously

mentioned, and are thus not discussed in further detail here.

Mass analyzers

While ionization and ion sources may take on numerous different designs

and operational mechanisms, there are four general types of mass analyzers:

time-of-flight (TOF) instruments, sector instruments, quadrupole mass

filters, and ion traps (Guilhaus et al., 2000; de Hoffmann and Stroobant,

2007, pp. 85-87). A TOF instrument makes use of the principle that ions

with identical charges that are accelerated across a constant distance by the

same repulsive force will have identical kinetic energies. The length of the

mass analyzer flight tube is also constant, meaning that the time it takes for

the molecules to travel the length of the tube will only depend on the velocity

of the ion, which is directly proportional to the ion's mass-to-charge ratio

(m/z) by the definition of kinetic energy (Wollnik, 1993; Guilhaus et al.,

2000). One of the instruments used for analyses in Papers II and III was an

ESI-TOF instrument.

Sector instruments are rather similar to TOF instruments in the sense that

sector instruments also use a flight tube to separate ions based on their

charge-to-mass ratios (m/z). The main difference is that sector instruments

utilize electric or magnetic fields to separate the ions. In sector instruments,

the path radius for a certain m/z value is unique as long as the magnetic or

electric field magnitude and potential difference used for acceleration are

held constant. The flight tube of a sector instrument is usually bent rather

than straight, as it is in TOF instruments. This allows for analysis of narrow

m/z intervals of interest in a sample as both the magnetic field magnitude

and the voltage can be varied. These operational properties allow sector

instruments to be used for screening/identification/quantification of a

narrow mass interval in complex samples, where the molecule of interest

may be present in trace concentrations in relation to other compounds in the

sample matrix (de Hoffmann and Stroobant, 2007, pp. 143-146).

Quadrupole instruments are the most common type of mass analyzers today

due to their simple and inexpensive design. These instruments are

constructed with four parallel rods/electrodes of opposite and variable

potentials. This variable potential creates oscillating electrical fields in the

39

space between the electrodes. Ions, which are within a desired mass interval,

are sent through the space between the electrodes and will be stabilized and

allowed to reach the detector. Ions outside of the desired interval become

instable and will collide with the electrodes or the walls of the quadrupole.

Thus, the mass analyzer operates as a mass-selective filter. As the potential

and frequency of the oscillation can be varied and adjusted, the instrument

may either scan for a single m/z value or scan through several different m/z

values during the course of the analysis (de Hoffmann and Stroobant, 2007,

pp. 88-96).

Ion-trap instruments are perhaps the most elegant of all the mass analyzers

when it comes to the general principle behind the separation of ions. As the

name suggests, the principle is based on trapping ions within a confined

space either with the use of electric or magnetic fields. In quadrupole trap

instruments the frequency and potential of the electrodes can be adjusted to

trap ions within a desired m/z interval. The trap usually employs a selective

ejection of the ions into the detector, allowing for very accurate mass

measurements to be performed (de Hoffmann and Stroobant, 2007, pp. 122-

126). Ion traps based on magnetic fields, i.e. Fourier-transform ion-cyclotron

resonance (ICR) instruments, employ magnetic fields rather than electric

and do not operate on basis of separation of ions. In contrast to quadrupole

ion traps, ICR instruments use a magnetic field to trap all the entering ions

in an orbit. The instrument then generates a signal through measurements of

relaxing atoms that are excited through the application of an external electric

field to all of the trapped ions (de Hoffmann and Stroobant, 2007, pp. 157-

159).

Mass spectrometers may consist of either a single mass analyzer or two in

tandem, i.e. MS or MS/MS. Tandem mass spectrometers may consist of two

mass analyzers of the same type or two that are distinctively different. One of

the instruments that were used for analysis of elemental composition of

compounds that were detected as products of treatments with sulfur

oxyanions (Paper II) was a linear trap quadrupole-orbitrap tandem mass

spectrometry instrument, i.e. a UHPLC-LTQ/Orbitrap-MS/MS.

Detectors

The final stage of mass spectrometry is to detect the ionized and separated

ions and quantify their presence. Detection of ions can either be destructive

or non-destructive, and this can be performed through a variety of designs

and operational principles, but these are usually based either on measuring

the charge or the momentum. The Faraday cup is the traditional way of

40

detecting ion signals. It may be used when the signal/amount of ions that is

quantified is considered to be large. This metal cup is built to measure the

resulting current which occurs when the charged ions enter the vacuum in

the middle of the cup. As the ions usually carry a single charge, the current

which is measured can thus be related to the number of ions entering the cup

(de Hoffmann and Stroobant, 2007, pp. 176-177).

Other ways of detecting ions include different types of electron and photon

multipliers, which are vacuum tubes operating on the basis of secondary

electron or photon emission. In principle, when a particle, such as an

electron or a photon, impact a secondary emissive material, the emission of

1-3 particles is induced. These emitted electrons or photons may in turn be

further accelerated between thin plates to induce the emission of yet more

particles. If this is repeated over and over again, the initial impact may result

in an avalanche of particles which may be registered by either metal

electrodes or photon counters as a signal (de Hoffmann and Stroobant,

2007, pp. 175-181).

Mechanism behind treatment with reducing agents

The improvement of the fermentability observed in the study presented in

Paper I raised questions about the molecular mechanism behind treatment

with sulfur oxyanions. Larsson et al. (2000) had previously used a number of

different aromatic model compounds to investigate the toxicity to yeast and

had found major differences between the toxic effects of the studied

compounds. Several of the compounds investigated by Larsson et al. (2000)

were chosen for the studies of the mechanism behind the fermentation

improvements observed in the study presented in Paper I. The compounds

that were selected (Figure 3) were dissolved in citrate buffer at pH 5.5 and at

a concentration of 5 mM. Then, the samples were treated with the

concentration of sodium dithionite and sodium sulfite used in the study of

Paper I. The analysis of the treated samples was performed using a Waters

Ultra-High-Performance Liquid Chromatography (UHPLC) system coupled

with a time-of-flight mass spectrometer (UHPLC-ESI-TOF-MS). The

elemental compositions of selected ions were later analyzed using Ultra-

High-Performance Liquid Chromatography coupled with a linear trap

quadrupole and orbitrap tandem mass spectrometer (UHPLC-

LTQ/Orbitrap-MS/MS).

41

Figure 3. Furan aldehydes and aromatic compounds studied in Papers I and II.

UHPLC was used mainly due to the nature of the model compounds that

were chosen for the mechanistic experiments. None of the chosen aromatic

compounds were volatile enough to render gas chromatography suitable for

analysis. Furthermore, the C18 capillary column that was used was known to

have an excellent chromatographic resolution of rather similar compounds.

The ESI of the mass spectrometer used was suitable for these studies due to

the nature of the ionization method, since it allowed the compounds to be

ionized while being kept intact without the risk of fragmentation and thus

lowering the risk of losing important information. The time-of-flight mass

analyzer in the mass spectrometer also allowed the compounds to be kept

intact while being separated on basis of their mass-to-charge ratio, without

risking any significant occurrence of fragmentation in the mass analyzer. The

Orbitrap-MS/MS used for elemental analysis was mainly chosen due to its

powerful resolution and mass accuracy. This is evident in the result of the

studies (Paper II).

The analyses performed with UHPLC-ESI-TOF-MS revealed the formation

of new compounds in the samples that were treated with sodium dithionite

and sodium sulfite. The results also indicated that these compounds had

42

increased mass and shorter retention times on the C18 column than the

original model compounds. These initial analyses indicated that phenolics

with aldehyde side chains, ferulic acid, and furan aldehydes (Figure 3)

reacted with dithionite and sulfite, creating new compounds in the analyzed

samples. Phenolics with an alcohol side chain, cinnamic acid and p-coumaric

acid (Figure 3) did not seem to react as no formation of new compounds was

observed. It was suggested that this shift was due to the addition of a sulfur-

based compound to the treated model inhibitors. Samples from the

experiments with coniferyl aldehyde, vanillin, HMF, and ferulic acid were

chosen for further analysis using the high-mass-accuracy UHPLC-

LTQ/Orbitrap. The isotopic pattern which appeared in the MS spectra that

were studied supported the presence of a sulfur atom in the reaction

products. All analyzed isotopes in the treated samples analyzed with the

UHPLC-LTQ/Orbitrap resulted in small differences between the detected

and theoretical masses, supported by the minimal deviation in mass

accuracy, presented as mass deviation in milli-mass units (mmu) (Table II in

Paper II). The addition of a sulfonate group [SO3H2] to the studied

fermentation inhibitors could explain the net mass increase of the

compounds and agrees with what could be expected for a reaction with

sulfur oxyanions. This would also explain why the inhibitors became strongly

hydrophilized after treatment with dithionite and sulfite at pH 5.5. The

sulfonated compounds resulting from treatment of model compounds

resemble the structures of sulfonic acids, i.e. [R-SO3H] where R is an alkyl or

an aryl group. The pKa values of sulfonic acids strongly depend on the R-

substituent, but as a group sulfonic acids are considered to have very low

pKa values in aqueous solutions, as low as sulfuric acid (Remington, 2006

p.394). This would indicate that the compounds resulting after treatment

with dithionite and sulfite should be deprotonated at fermentation pH 5.5

and thus strongly hydrophilic. Some of the similarities of the compounds

resulting after treatments with dithionite and sulfite could tentatively be

explained by the instability of dithionite in aqueous solutions, in which it to

some extent would dissociate into sulfite (Equation 2).

Sulfonation appears to result in that the inhibitors become less reactive and

more hydrophilic, and thereby less toxic to the fermenting microorganism.

The reaction evidently occurs with different types of side-chains in phenolic

compounds, as both aldehydes and some compounds with unsaturated side-

chains with carboxylic acid groups were found to react. Ferulic acid, which

has the same side chain as cinnamic acid and p-coumaric acid but a more

substituted aromatic ring, was reactive while the other acids were not.

Alcohols were also studied, but no sulfonation was detected when they were

mixed with sulfite or dithionite. However, as previously reported (Larsson et

43

al., 2000), the alcohols appear to be far less toxic than the corresponding

aldehydes or carboxylic acids that were affected by the treatment.

After the promising results presented in Papers I and II, studies on the

effects and mechanisms of a different reducing agent, sodium borohydride

were conducted (Paper III). The experiments were performed in a similar

fashion as those with sodium dithionite and sodium sulfite: 5 mM of the

model compound was dissolved in citrate buffer at pH 5.5 and it was then

treated with 5 mM of sodium borohydride. Also this time, analysis was

performed with UHPLC-ESI-TOF-MS. As expected, treatment with sodium

borohydride resulted in different types of products compared to the

treatments with sulfur oxyanions. The reaction between fermentation

inhibitors and borohydride resulted in reduction of aldehydes and

benzoquinones to the corresponding alcohols. This was observed for

coniferyl aldehyde, furfural and two benzoquinones (Figure 3). The

carboxylic acids did not react. Another important conclusion that can be

drawn from these studies concerns the effect on enzymatic hydrolysis. While

addition of sodium dithionite improved enzymatic hydrolysis in the presence

of inhibitors, addition of sodium borohydride did not. In contrast to the

strongly hydrophilic sulfonated products which occurred in treatments with

sulfur oxyanions, the corresponding alcohols were just slightly more

hydrophilic than the original aldehydes used as model compounds for

treatment with sodium borohydride. The results suggest that increased

hydrophilicity is important for overcoming inhibition of enzymes, while

decreased reactivity is important primarily for overcoming inhibition of the

fermenting microorganism.

44

Enzymes in biorefining

Detoxification contributes mainly to the improvement of the fermentability

of lignocellulosic hydrolysates and thus to faster fermentations and higher

product yields, which are important aspects for commercialization of

cellulosic bioethanol. Enzymes are also very important in this respect, as

they have been identified as one of the major bottlenecks for

commercialization of cellulosic bioethanol (Wingren et al., 2003). One

approach to making enzymes less costly and thus increase the potential for

commercialization is through on-site enzyme production. This could be done

using glucose-based media. However, it could potentially also be achieved

through utilization of residual streams arising in the biorefinery, such as

stillages left after distillation, which would be even more advantageous than

on-site enzyme production using glucose-based media. Before on-site

enzyme production is discussed in more detail, an overview of enzymes and

of microorganisms commonly used for enzyme production is given.

Cellulases

During the last few decades, microbial enzymes have gained increased

attention as catalysts in different types of industrial applications. Cellulases

are enzymes that in nature are produced mainly by fungi and aerobic

bacteria during their growth and utilization of cellulose from plant matter

(Kubicek et al., 2008; Kubicek, 2013). There is a wide variety of cellulases,

but they can be classified in three main different families; endo-glucanases,

exo-glucanases and β-glucosidases (Himmel et al., 1999; Horn et al., 2012b).

The difference between the families lies mainly in the type of reaction that

the enzymes catalyze. The endo-glucanases act randomly along the cellulose

polymer chains mainly on the internal O-glycosidic bonds, producing glucan

chains of varying and random lengths as well as exposing the ends of

individual cellulose chains. Exo-glucanases act on the exposed ends of the

cellulose chain left behind by the endoglucanases and release β-cellobiose as

the product. β-Glucosidases act specifically on the β-cellobiose produced by

exo-glucanases and yield glucose as the end product (Percival Zhang et al.,

1999; Horn et al., 2012b). The system of cellulose degradation of anaerobic

bacteria is somewhat different than that of aerobic bacteria and fungi.

Anaerobic bacteria utilize membrane-bound multienzyme complexes called

cellulosomes to degrade cellulose. The complex has two main purposes, first

to bind to the cellulose and then reconfigure or rearrange its structure to be

able to hydrolyze the cellulose chains into sugars, which the bacteria then

may use as nutrient (Bayer et al., 2004). The enzymatic activity and thus the

45

rate of conversion of the substrate is known to increase with elevating

temperatures, up until a temperature is reached where denaturation and

inactivation of enzymes start to occur (Fullbrook, 1996). Thermostable

cellulases from fungi and bacteria are gaining increased attention in

bioconversion of cellulose due to that their activity is retained even after

exposure to heating for a prolonged period of time (Viikari et al., 2007).

Thermostability is, however, a complex term to discuss as different types of

enzymes from different types of organisms may display completely different

tolerance towards prolonged heating. Viikari et al. (2007), who reviewed this

field extensively, concluded that some enzymes retained their activity even

after several days of heating, such as one endoglucanase from Clostridium

stercorarium, while other enzymes could only withstand heating for a few

minutes or one hour before losing half of their activity.

Hemicellulases

In contrast to cellulose, hemicelluloses are not homogeneous repeated

polymers of glucose residues, but are mainly composed of a variety of hexose

and pentose sugar residues derived from xylose, mannose, arabinose, and

galactose. Due to this heterogeneity, hemicellulases comprise a complex

group of enzymes that hydrolyze different types of linkages in

hemicelluloses. They are classified according to the type of substrate that

they hydrolyze (Shallom and Shoham, 2003). Hydrolysis of xylan to xylose

involves several different types of xylanases, the most important being endo-

1,4-β-xylanase (Figure 4), exo-β-1,4-xylosidase and β-xylosidase (Shallom

and Shoham, 2003; Collins et al., 2005). In the same manner, hydrolysis of

mannan to mannose involves mainly endo-β-1,4-mannanase, exo-β-1,4-

mannanase and β-mannosidase (Shallom and Shoham, 2003; van Zyl et al.,

2010). The complex nature of hemicellulose makes the involvement of other

enzymes necessary, for instance acetyl xylan esterase and acetyl mannan

esterase, which hydrolyze bonds between acetyl groups on xylan or mannan.

Different types of arabinases and galactosidases are also involved (Shallom

and Shoham, 2003).

Lytic polysaccharide monooxygenases

About two decades ago, a new type of cellulose-binding enzyme was

discovered, namely CelI from Agaricus bisporus. It was classified in group

61 of the glycoside hydrolases and was thus named GH-61 (Lo Leggio et al.,

2012). This group of enzymes did, however, not show any strong hydrolytic

activity on cellulose and their exact function was unknown at the time. It was

46

recently discovered that these enzymes have the capability to catalyze

oxidative cleavage of the cellulose polymer (Horn et al., 2012b; Lo Leggio et

al., 2012). Thus, in 2013 this enzyme was reclassified in accordance with its

functionality, i.e. as a copper-dependent lytic polysaccharide monooxygenase

(LPMO) (Levasseur et al., 2013). This enzyme disrupts the homogeneous

structure of cellulose and makes the substrate more accessible to hydrolytic

cellulases (Horn et al., 2012b). The discovery of the synergistic effects of GH-

61 enzymes and hydrolytic cellulases is one of the most important

discoveries in recent time in the field of cellulose biodegradation.

Figure 4. 3D image of endo-1,4-β-xylanase from T. reesei.

Assay of enzymatic activity

The current SI unit for catalytic activity of enzymes is katal (kat), which is

defined as mole of conversion per second (mol/s) (Dybkær, 2001). However,

47

for characterization of cellulases and other hydrolytic enzymes, other units

are also used; particularly filter paper units (FPU) and units (U) (Eveleigh et

al., 2009). Filter paper units (FPU) were originally referred to as the amount

of reducing sugar released from a strip of Whatman no.1 filter paper by 1 ml

of undiluted enzyme preparation. This was later redefined as the amount of

enzyme that releases 2 mg of glucose/ml from a strip of Whatman no.1 filter

paper and is currently the used definition of FPU (Ghose, 1987; Adney and

Baker, 1996). Units (U) are defined as the amount of enzyme required to

release 1 µmol reducing sugar from an appropriate substrate per min (Ghose,

1987). Different temperatures and pH values are occasionally used in these

enzymatic assays, which renders comparisons of different studies rather

complicated. Enzymatic activity is strongly temperature dependent, i.e.

higher activities are reached with higher temperatures until denaturation

and inactivation starts to occur (Fullbrook, 1996). Enzyme kinetics is thus

strongly temperature dependent and even a few degrees in temperature

difference may give a significant difference of the rate of the conversion

catalyzed by the studied enzyme. For these reasons (the use of different

enzymatic activity units, different substrates, and different temperature and

pH used in the determination of enzymatic activity), it is occasionally very

difficult to compare enzymatic activities reported in different studies, even

though the enzyme itself may be identical.

Filamentous fungi

Fungi are truly roving packs, who steal whatever they find left behind,

after Flora has set their winter tents in fall.

Carl von Linnaeus

The kingdom of Fungi is a very special one within the world of eukaryotic

organisms. It includes yeasts, molds, and the more well-known mushrooms.

Von Linnaeus was evidently not very fond of fungi, and their versatility did

not render them alike anything else that he had categorized before. His

dislike of these organisms was shared by many others long before him and

after his own time. The Greek poet and physician Nicander of Colophon [ca.

185 B.C.] described these organisms with the words: “Let not the evil

ferment of the earth which often causes swelling in the belly or strictures in

the throat distress a man … an evil ferment is that, men generally call the

ferment by the name of fungus” (Rolfe and Rolfe, 1925, p. 227).

48

Despite their soiled reputation, fungi have nevertheless found their place in

science and modern society. They have become true industrial workhorses

for production of food ingredients, specialty chemicals, pharmaceuticals,

heterologous proteins and industrial enzymes (Pel et al., 2007; Rördam

Andersen et al., 2008). In science, they are mostly studied for the

understanding of fundamental cellular processes but also for their ability to

cause problems and interfere with the affairs of humans, plants and animals

alike. They can be found in essentially every ecosystem on the planet and

with their versatile nutritional ability they play a vital role within these

ecosystems (Borkovich and Ebbole, 2010, p. 3). Approximately 100,000

different species of fungi have so far been isolated, categorized and studied.

Not nearly all of these are extensively studied or utilized in industrial

processes. Genera and species of filamentous fungi that have become

industrial workhorses and therefore also have become very well studied

include the aspergilli and Trichoderma reesei (Hypocrea jecorina).

Aspergillus niger

A member of the Trichocomaceae family and the genus Aspergillus, A. niger

(Figure 5) was first characterized by the French botanist Philippe Édouard

Léon Van Tieghem in 1867. Together with its 15 close relatives in the Nigri

section, A. niger has become very well studied and an important industrial

fungus. A. niger and its relatives have a natural ability to secrete large

amounts of enzymes in order to utilize nutrition from biopolymers in their

near environment. A. niger is one of the most utilized enzyme producers in

industry. Enzymes produced by A. niger are today used in a wide array of

fields including the food and beverages industry, the pharmaceutical

industry, and for production of animal feed, chemicals and fuels (Pel et. al.,

2007; Rördam Andersen et al., 2008). Features that contribute to making A.

niger interesting include its versatile nutritional metabolism, its high

secretory capacity, its GRAS (generally regarded as safe) status, and its well-

studied genomic and metabolomic profile (Rose and van Zyl, 2002).

49

Figure 5. Submerged cultivation of A. niger in synthetic medium (above) and in waste fiber sludge

stillage (below).

Trichoderma reesei (Hypocrea jecorina)

Trichoderma reesei was first isolated during World War II in the Salomon

Islands and is known to originate from a single isolate (QM6a) (Kuhls et al.,

1996). It is mainly utilized in industry for production of hydrolytic enzymes

and for production of antibiotics (Ghisalberti and Sivasithamparam, 1991;

Penttilä et al., 1991, pp. 85-102). T. reesei is known to be a strong producer

of cellulases and hemicellulases, such as endoglucanases, exoglucanases and

xylanases, while producing limited activities of β-glucosidase (Duff et al.,

1986). The results of Duff et al. (1986) also revealed that the enzyme-

expression pattern of T. reesei is complementary to that of Aspergillus spp.,

which have a strong expression of β-glucosidase. Enzyme production by T.

50

reesei is strongly dependent on gene induction and regulation. The fungus

has a tendency to produce an array of different enzymes when cultivated. For

that reason, T. reesei has undergone extensive scientific studies regarding

gene regulation and gene induction, with focus on understanding the

genome and protein secretion of the fungus. The complete sequencing of the

T. reesei genome revealed, somewhat surprisingly, that its genome encodes

fewer hydrolytic enzymes (such as cellulases and hemicellulases) than any

other sequenced fungus that has the capability to hydrolyze plant cell wall

polysaccharides (Martínez et al., 2008). These results indicate that the

cellulose-degrading enzymes produced by T. reesei are highly relevant for

fundamental research and applications within the field.

51

Potential of on-site enzyme production

Production of cellulosic bioethanol is dependent on efficient hydrolytic

enzymes for breakdown of long chains of sugar polymers to fermentable

sugars. Hydrolysis can be achieved with either concentrated or dilute acid,

but, as discussed previously, those processes have significant drawbacks.

Enzymes are strongly selective and highly efficient catalysts. The major

drawback with enzymes is the production cost, which is also considered as

one of the major bottlenecks for commercialization of advanced biofuels.

Achieving an efficient production of enzymes on-site at a lignocellulosic

biorefinery as a complement to production of cellulosic ethanol may lower

the cost for production of ethanol and have additional benefits, such as more

efficient utilization of the lignocellulosic biomass used in the process and less

environmental impact of enzyme production.

The concept of on-site enzyme production with utilization of residual

streams from a cellulosic bioethanol process was introduced by Alriksson et

al. (2009) who studied utilization of stillage from lignocellulosic

hydrolysates for enzyme production with A. niger. This concept was based

on sequential fermentation processes with yeast and the filamentous fungus.

The idea behind the concept was to allow the yeast to consume glucose and

mannose for ethanol production in the first fermentation step. Rejected

monosaccharides, oligosaccharides, organic acids, furan aldehydes and

fermentation metabolites, such as glycerol, would be left in the stillage, i.e.

the liquid residue after the distillation. The stillage could then be utilized in a

second fermentation step in which enzymes are produced using a

filamentous fungus.

An investigation was performed on the potential of using waste fiber sludge,

a low value residual stream from pulp mills and biorefineries, for combined

production of bioethanol and endo-1,4-β-beta-xylanase (Paper IV). Waste

fiber sludge contains mainly cellulose and in some cases a smaller portion of

hemicellulose. The lignin content is usually low (Table II). The

microorganism that was used for enzyme production was a recombinant A.

niger, harboring the xyn2 gene from T. reesei. One of the benefits of using a

recombinant A. niger with the inserted gene under the control of a

constitutive promotor is that heterologous expression of the enzyme is not

dependent on gene induction (discussed in detail in a later section). This

allowed the enzymatic activities to reach levels of around 8200 nkat/ml after

12 days of fermentation. In Table III, the results of this study are compared

to other attempts to produce endo-1,4-β-beta-xylanase from T. reesei.

52

Table III. Comparisons of enzyme activities reached with expression of T.

reesei endo-1,4-β-beta-xylanase.

Organism Nutrient source Xylanase activity

(nkat/ml)

Level of activity vs.

Paper IV

A. niger[xyn2] Stillage 8,000a 1:1

S. cerevisiae YPD medium 1,200b 1:7

P. pastoris YPD medium 4,300c 1:2

T. reesei Birch glucuronoxylan 5,400d 2:3

A. niger[xyn2] 20% (v/v) molasses 5,000e 3:5

A. niger[xyn2] 10% glucose 8,000e 1:1

References for enzymatic activities: a Paper IV (Cavka et al., 2011); b La Grange et al., 1996; c

He et al., 2009; d Bailey et al., 1993; e Rose and van Zyl, 2002.

As can be seen in Table III, the xylanase activity reached in the study

reported in Paper IV was equal to those reached by cultivation of A. niger in

rich medium with 10% glucose and significantly better than those studies

that were performed using other expression hosts or native T. reesei.

In a further development of this concept (Paper V), waste fiber sludge was

used in an SSF configuration rather than in an SHF, which was used

previously (Paper IV). This was done to investigate cellulase production

using fiber sludge from different sources and also included attempts to

recycle enzymes added to the SSF. For that purpose, there was a second SSF

step which included recirculation of hydrolytic enzymes from an initial SSF

step. The attempt to recycle hydrolytic enzymes from the first SSF to the

second was made through transfer of residual solids from the first SSF. The

results of this study showed that it was possible to recycle a fraction

corresponding to around 30 percent of the enzymic activity that was added

in the initial SSF step to the second, to which new solvent and raw material

were added, but no fresh enzymes. Thus, the results indicate that a

substantial part of the hydrolytic enzymes that are added to an SSF can be

recirculated to a subsequent step using centrifugation of the medium. The

results also show that the average ethanol concentration in the two SSF steps

was 40 g/L (Paper V). In this study an attempt was also made to produce

53

enzyme using spent fiber sludge hydrolysate, i.e. the stillage that was left

after distillation of the fermented waste fiber sludge hydrolysate.

Recombinant A. niger harboring the endo-glucanase I (EgI) gene from T.

reesei was used for enzyme production. The differences between the studies

of Paper IV and Paper V include not only production of different enzymes,

enzyme recirculation and different types of processes (SHF vs. SSF), but

Paper V also reports an attempt to produce enzyme directly from waste fiber

sludge rather than doing it only from stillage. This attempt was, however,

unsuccessful. A. niger did display growth on the waste fiber sludge, but no

notable activity was detected in the liquid fraction. This was probably due to

binding of the produced endo-glucanase I to cellulose fibers in the raw

material used in the cultivation. For that reason it was concluded that even

though the waste fiber sludge did serve as nutrient for A. niger, the binding

of the produced protein to the fibers made it unsuitable for enzyme

production. Using the raw material for combined production of biofuels and

enzymes was more efficient. In Table IV, the enzyme activity obtained in this

study (Paper V) is compared to other attempts to produce the same enzyme

using the same gene introduced into other expression hosts or using A .

niger cultivated on other media.

The enzymatic activity that was reached in the study of Paper V, 2,500

nkat/ml, is not much lower than the activity reached with T. reesei and five

percent bleached pulp as nutrient source (3,800 nkat/ml) (Table IV). The

activity was almost twice as high as that achieved with the same A. niger

strain and molasses (1,400 nkat/ml), and it was equal to that obtained with

an optimized medium with 10% glucose (2,500 nkat/ml) (Table IV).

Similarly to the results presented in Paper IV (Cavka et al., 2011), the

enzymatic activities produced by A. niger using residual streams appear to

be equal to or better than those achieved with other expression hosts and

media. Thus, the results indicate that utilization of stillage for production of

enzymes with A. niger results in high enzymatic activities as well as in

efficient utilization of the components of the raw material. This is achieved

as the glucose and mannose in the raw materials are utilized for ethanol

production by S. cerevisiae in the first fermentation step. A. niger is then

allowed to utilize monosaccharides and oligosaccharides that S. cerevisiae

was unable to utilize for ethanol fermentation, as well as fermentation

inhibitors, such as acetic acid, for production of hydrolytic enzymes in a

subsequent second fermentation step.

54

Table IV. Comparisons of enzyme activities reached with expression of T.

reesei endo-1,4-β-glucanase I (Cel7B).

Organism Nutrient source CMCase activity

(nkat/ml)

Level of activity vs.

Paper V

A. niger[EgI] Stillage 2,700a 1:1

S. cerevisiae YPD medium 120b 1:25

Y. lipolytica YPD medium 370c 1:8

T. reesei 5% bleached pulp 3,800d 1:2

A. niger[EgI] 20% (v/v) molasses 1,400e 1:2

A. niger[EgI] 10% glucose

(2×MM) 2,500e 1:1

References for enzymatic activities: a Paper V; b La Grange et al., 1996; c Park et al., 2000; d

Montenecourt et al., 1983; e Rose and van Zyl, 2002.

Enzyme production with recombinant A. niger as an expression host is

suitable for production of specific selected enzymes, as the constituent

promoter that is used allows a strong overexpression of the cloned gene. This

was evident by the results reported in Papers IV and V, as A. niger produced

very high enzymatic activities on stillage. Enzymes that may be produced in

this manner do not necessarily need to be useful for production of liquid

biofuels. For example, enzymes that may be used for producing animal feed

or for bleaching of pulp may be produced in this manner. The endo-1,4-β-

beta-xylanase that was produced in Paper IV is an example of an enzyme

that may be useful for other processes than production of cellulosic ethanol.

The production of such enzymes could improve the overall process economy

of a biorefinery and thereby assist commercialization of production of

cellulosic biofuels. Production of enzymes using residual streams and A.

niger could also be suitable for supplementing a CBP with enzymes that are

not produced in sufficient amounts by the microorganism that is utilized for

the CBP process.

Compared to enzyme production with A. niger, enzyme production with T.

reesei has both advantages and drawbacks. T. reesei is a far better producer

of endoglucanase, exoglucanase and xylanase than A. niger, but it produces

55

relatively small amounts of β-glucosidase. A. niger is a strong producer of β-

glucosidase, which is important for the yield of monomeric sugars (Lima et

al., 2013). One possible solution to such a limitation could be to perform

mixed cultivations of aspergilli and T. reesei. These cultures could give a

mixture of all enzymes required for an efficient hydrolysis of biomass to

fermentable sugars. Such experiments have also successfully been performed

in the past (Gutiérrez-Correa and Villena, 2102). The limitation with that

approach is that it becomes quite difficult to optimize the cultivation

conditions for two different microorganisms that need to co-exist during the

cultivation. A further drawback with this approach is that most organisms in

nature, bacteria, fungi and vertebrates, compete with other organisms for

living space and nutrients in their near vicinity. Performing a mixed

cultivation of two microorganisms will inevitably create a competition for the

available nutrients.

Enzyme expression and gene regulation

Through evolution most organisms have developed genetic regulatory

systems that allow them to respond to a shifting environment. For organisms

such as fungi, it is vital that the enzymes that are used to release nutrients

from their surroundings are regulated so that they are secreted only when

suitable nutrients are available in their environment. Gene regulation in

Aspergillus spp. and T. reesei has been investigated thoroughly due to the

industrial and economic importance of these fungi (Ruijter and Visser, 1997;

Ilmén et al. 1997; David et al., 2003; Kubicek et al., 2008). The type of

carbon source that is available is vital for gene repression in T. reesei

(Margolles-Clark et al., 1997). The mechanism of the gene regulation of

hydrolytic enzymes in these fungi is rather similar. The repressor protein

CreA plays an important role in aspergilli, while T. reesei has a similar

protein called CRE1. Both these proteins regulate gene expression through

transcriptional inhibition of genes by direct binding to specific binding

regions of the promoters (Ruijter and Visser, 1997; Martínez et al., 2008). It

appears that the targets for these transcriptional inhibitors are genes that

encode permeases, i.e. membrane transport proteins. This indicates that

these fungi regulate their gene expression depending on the presence of

carbon sources which are available to them through a mechanism which

affects substrate transportation through the cell membrane and into the cell

(Portnoy et al., 2011). Several different carbon sources, including mono- and

polysaccharides as well as oligosaccharides, are known to induce these

repressor proteins (Ruijter and Visser, 1997). Through this type of

regulation, the fungus establishes a preference for some carbon sources over

56

others, and carbon sources that are easily metabolized are transported into

the fungal cells.

The natural gene regulation mechanisms that control expression can be

overcome through the use of a constituent promotor, as was done with the A.

niger strains developed by Rose and van Zyl (2002) and used by Alriksson et

al. (2009) and in the studies of Papers IV and V. T. reesei Rut C-30, which is

of great interest for industrial production of hydrolytic enzymes, is a CRE1

mutant with a defect and shortened cre1 gene. This allows it to become a

highly efficient enzyme producer even in the presence of high concentrations

of easily available carbon source (Ilmén et al., 1996; Portnoy et al., 2011).

Even though T. reesei Rut C-30 is an efficient producer of cellulases and

hemicellulases, it is still dependent on the composition of the carbon source

that is available in the cultivation media.

Inhibition of enzyme-producing microorganisms

Lignocellulose-derived inhibitory compounds are not only a problem in

ethanolic fermentations but are also potential problems for microorganisms

used for enzyme production. To investigate the sensitivity of different

potential enzyme producers, a study based on five different microorganisms

that are today commonly utilized for enzyme production or for CBP was

conducted (Paper VI). The five microorganisms were A. niger D15, T. reesei

Rut C-30, Pichia pastoris X-33, Yarrowia lipolytica ATCC 8661 and

common baker’s yeast S. cerevisiae. The microorganisms were exposed to

pretreatment liquid and hydrolysate from spruce in concentrations of 25%,

50%, 75%, and 100% (v/v) and the results were compared to reference

fermentations without inhibitory compounds. Pretreatment liquid and

enzymatic hydrolysates of pretreated Norway spruce were used rather than

stillage from ethanol production. This was due to the similarities of inhibitor

composition between pretreatment liquid and hydrolysate compared to

stillage, in which some of the inhibitors may have been converted by S.

cerevisiae in the first fermentation step. The nutrient composition of the

pretreatment liquid and the hydrolysate was also different, which allowed

studies of nutrient utilization in media with different composition of carbon

source (Table 1 in Paper VI). Pretreatment liquid mainly consists of

monosaccharides from hemicellulose degradation, i.e. mainly xylose and

mannose, but with low concentrations of glucose (around 1%). Hydrolysates

usually contain higher concentrations of glucose (between 5-10%)

originating from the hydrolysis of cellulose. The microorganisms were added

in concentrations of around 0.25 g/L (DW) and were allowed to grow until

the maximum biomass concentration was reached.

57

Knowledge of the tolerance to inhibitory compounds that are present in

lignocellulose-derived media, such as pretreatment liquid and hydrolysate

from pretreated spruce, is of importance for assessing the suitability of

microorganisms for on-site enzyme production in lignocellulosic

biorefineries. The results presented in Paper VI show that the five tested

microbial strains exhibited variable tolerance to lignocellulose-derived

inhibitors. All microorganisms displayed growth in concentrations of 25%

pretreatment liquid and 25% hydrolysate, but the results were different at

higher concentrations (Table 2 in Paper VI). S. cerevisiae was the only

microorganism that displayed growth in 50% pretreatment liquid. The

results from hydrolysate experiments were somewhat different than the

results obtained with pretreatment liquid, as three out of five

microorganisms (A. niger, P. pastoris and S. cerevisiae) displayed growth in

50%, hydrolysate. S. cerevisiae was the only microbe that grew in 75%

hydrolysate. The results presented in Paper VI indicate that the common

baker's yeast strain of S. cerevisiae was most tolerant to inhibitors among

the five microbial strains studied. The results also suggest that T. reesei and

Y. lipolytica may be more sensitive to inhibitors in lignocellulose-derived

media than A. niger, P. pastoris and S. cerevisiae.

In the assessment of the suitability of microorganisms for on-site enzyme

production it is also important to take their ability to utilize nutrients into

consideration. On-site enzyme production should preferably be based on

carbon sources other than glucose and mannose, as these monosaccharides

could be used for bioethanol production instead. For that reason, the five

micoorganisms included in the study of Paper VI were also evaluated on

basis of their ability to utilize different lignocellulose-derived nutrients in the

media. The results show that the filamentous fungi A. niger and T. reesei

possess the ability to consume a wider array of nutrients than P. pastoris, S.

cerevisiae and Y. lipolytica (Table 1 in Paper VI). The fungi had the ability to

consume all the measured monosaccharides, i.e. glucose, mannose, xylose,

arabinose and galactose, as well as acetic acid and oligomeric sugars. The

extent of consumption was, however, dependent on the concentration of

glucose present in the media. When glucose was readily available, as in the

reference medium, then no or little (i.e. only a few percent) of the other

monosaccharides were consumed (Table 1 in Paper VI). The yeasts were less

metabolically diverse and consumed mainly glucose and mannose. The

results also indicate that both P. pastoris and S. cerevisiae utilized a major

portion of the consumed carbon source for production of ethanol. In

medium with 50% hydrolysate, P. pastoris produced 0.30 g of ethanol per g

of consumed carbon source, while S. cerevisiae produced 0.33 g/g (Table 3

in Paper VI).

58

Besides displaying differences in inhibitor tolerance and nutrient utilization,

biomass formation varied between the studied microorganisms. The ability

to grow to high biomass concentrations is important for on-site enzyme

production as high biomass concentrations may also allow high protein

expression levels. The results show that all of the studied microorganisms

exhibited the highest biomass yields, both with respect to initial carbon

source (Yx-ICS) and consumed carbon source (Yx-CCS), in medium with 25%

pretreatment liquid. The highest yields were reached with A. niger, which

exhibited a yield of 0.46 g/g both with regard to Yx-ICS and Yx-CCS. Y.

yarrowia reached a Yx-ICS of 0.32 g/g and a Yx-CCS of 0.43 g/g, slightly lower

than A. niger. T. reesei was not far behind and reached a Yx-ICS of 0.30 g/g

and a Yx-CCS of 0.41 g/g. P. pastoris reached a Yx-ICS of 0.20 g/g and a Yx-CCS

of 0.26 g/g, while S. cerevisiae reached a Yx-ICS of 0.22 g/g and a Yx-CCS 0.31

g/g. These results suggest that if high biomass formation is preferred, a

medium with relatively low concentration of glucose should be used. This is

an aspect that would be beneficial if stillage would be used for on-site

enzyme production. That would, however, require that the microorganism

that is used does possess the ability to utilize other carbon sources than

glucose and mannose. The filamentous fungi A. niger and T. reesei should

therefore be prime targets for on-site enzyme production in terms of their

metabolic diversity, their ability to efficiently utilize carbon sources for

biomass formation, as well as their efficiency as enzyme producers. A. niger

has been shown to possess the ability to utilize or convert common

fermentation and growth inhibitors when grown in stillage from bagasse,

spruce and waste fiber sludge (Alriksson et al., 2009, Paper IV). These

inhibitory compounds include small aliphatic acids, furan aldehydes, and to

some extent also phenolic compounds.

59

Conclusions

High concentrations of inhibitors formed during pretreatment cause

problems for efficient biocatalytical conversion of lignocellulosic feedstocks

to advanced biofuels and other products. Detoxification of lignocellulosic

hydrolysates by treatment with sodium dithionite, sodium sulfite and

sodium borohydrate, reducing agents used in industrial-scale processes, was

investigated in experiments with the yeast S. cerevisiae. The results clearly

show that reducing agents are very efficient for improving the fermentability

of hydrolysates. In some cases, it was possible to reach an ethanol

productivity and an ethanol yield that was equal to that obtained with a

reference medium without any inhibitors and to that obtained using alkali

detoxification. The benefits with treatment with reducing agents include that

a separate process step for detoxification is not needed, as detoxification can

be performed in situ directly in the fermentation vessel. Furthermore, there

was no degradation of fermentable sugars, and the treatment can be

performed at room temperature, with short treatment time, and using low

concentrations of low-cost industrial chemicals. The mechanism behind the

improvements effected by sulfur oxyanions, such as dithionite and sulfite,

was investigated and was shown to be sulfonation of common fermentation

inhibitors, such as furan aldehydes, aromatic aldehydes, and some carboxylic

acids. Treatment with sodium borohydride instead resulted in the reduction

of aromatic and furan aldehydes to their corresponding alcohols. In both

cases, the fermentability was greatly enhanced. While the treatments with

sulfur oxyanions had the additional benefit that enzyme inhibition was

alleviated, the treatment with sodium borohydride did not have that effect,

which indicates a mechanistic difference between fermentation inhibition

and enzyme inhibition.

Other problems with the bioprocessing of lignocellulosic feedstocks are the

environmental impact and the high costs that are associated with the

production of enzymes used for saccharification of the cellulose. These

problems were addressed in attempts to utilize residual process streams

from pulp mills and biorefineries to produce enzymes. The results show that

it was possible to utilize the filamentous fungus A. niger and obtain

relatively high enzyme activity using residual streams, such as fiber sludge

and stillages from the production of cellulosic ethanol. In some of the

studies, the levels of enzyme activity were equal to or better than what has

previously been obtained using conventional feedstocks and the same

microorganism.

60

The commercialization of advanced biofuels may benefit from the research

described in this thesis. Further research and development is, however,

needed. The economic feasibility of detoxification with reducing agents

needs to be investigated and compared to other detoxification methods. The

possibility to lower the loadings of biocatalysts and shorten the fermentation

time should be investigated further before an evaluation of the feasibility of

detoxification with reducing agents can be performed. Studies with

filamentous fungi other than A. niger, for instance T. reesei, need to be

performed and evaluated. The possibilities to supplement enzyme produced

by microorganisms used for consolidated bio-processing with enzyme

produced on-site by using filamentous fungi cultivated in residual streams

appear as an interesting area for further research.

61

Acknowledgments

I wish to take this opportunity to thank all those around me which have

made contributions to this thesis and the works which it is based on.

First of all I would like to express my deepest gratitude to my academic

supervisor and closest colleague, Prof. Leif Jönsson for his invaluable

support and for sharing his vast knowledge and experience. Thank you for

always finding the time to guide me in the correct direction even if your own

time was always greatly limited.

Dr. Björn Alriksson for sharing your laboratory skills and for the training

provided within the field of industrial microbiology. Your experience and

knowledge transfer of fermentation and detoxification was vital for the

successful completion of this thesis.

I would also like to thank all the people at the biorefinery industries in

Örnsköldsvik, especially SEKAB E-Technology for co-funding this project

and providing me with materials from the demonstration plant to base this

thesis on. To all the people at Processum who extended their help and made

me feel welcome from day one. To MoRe Research for their help with some

of the analytical work which this thesis is based on.

I want to extend a special thanks to Prof. Petter Gustafsson and to Benkt

Wiklund for their remarkable work with the Industrial Graduate School and

for giving me the opportunity to be part of such a great group of people, and

all the skills and experiences that I have acquired thanks to this

participation.

A big thanks goes to Prof. Feng Hong at Donghua University, Shanghai,

China, for giving me the opportunity to join your group and laboratory to

learn from your skills of bacterial cellulose and fungal fermentation. I would

also like to thank all the students and colleagues who put up with me and

Sara during our stays in Shanghai in both 2010 and 2011 and who

introduced us to the Chinese culture and cuisine.

Prof. Willem H. (Emile) van Zyl is greatly acknowledged for the contribution

to the enzyme production papers as well as for great discussions about

science and determination.

All my colleagues within our research group who always welcomed me and

helped me when I needed to perform experiments outside of my own

62

expertise. These thanks are especially extended to Dr. Sandra Winestrand for

her help with the HPAEC analyses and to Dr. He Jun for his help with

protein separation tools and analysis as well as for inviting me to Sichuan

Agricultural University, in Chengdu and Ya’an in 2011.

Dr. Maria Ahnlund is greatly acknowledged for help with mass spectrometry

analysis, work which was invaluable for the deeper understanding of our

work on detoxification.

Finally, and no words are enough to express the gratitude which goes to my

wife Sara who has been with me on this journey in highs and lows, in the

middle of the nights and in early Saturday and Sunday mornings. Thank you

for your greatest support!

The research has been supported by the Kempe Foundations, the

Bio4Energy research initiative (www.bio4energy.se), the Biorefinery of the

Future (www.bioraffinaderi.se), the Swedish Energy Agency, and the

Swedish Research Council.

Jedno veliko hvala pružam mojoj familiji za vašu veliku podršku i za svu

pomoć u mokrom i suhom. Hvala vam puno za svu pomoć u životu.

63

References

Adney B and Baker J. (1996) Measurement of cellulase activities. NREL

Laboratory Analytical Procedure 6. Accessed October 2013

<http://www.nrel.gov/biomass/pdfs/42628.pdf>

Aebersold R and Mann R. (2003) Mass spectrometry-based proteomics.

Nature 422:198-207.

Alriksson A, Horváth IS, Sjöde A, Nilvebrant N-O and Jönsson LJ. (2005)

Ammonium hydroxide detoxification of spruce acid hydrolysates. Appl

Biochem Biotechnol 121-124:911-922.

Alriksson B, Sjöde A, Nilvebrant N-O and Jönsson LJ. (2006) Optimal

conditions for alkaline detoxification of dilute-acid lignocellulose

hydrolysates. Appl Biochem Biotechnol 129-132: 599-611.

Alriksson B, Rose SH, van Zyl WH, Sjöde A, Nilvebrant N-O and Jönsson

LJ (2009) Cellulase production from spent lignocellulose hydrolysates by

recombinant Aspergillus niger. Appl Environ Microbiol 75:2366-2374.

Andrietta MGS, Andrietta SR, Steckelberg C and Stupiello ENA. (2007)

Bioethanol: Brazil, 30 years of Proálcool. Int Sugar J 109:195-200.

Badger PC. (2002) Ethanol from cellulose: A general review. In Janick J

and Whipkey A (eds.) Trends in new crops and new uses. Alexandria, VA,

USA: ASHS Press.

Bailey M, Buchert JJ and Viikari L. (1993) Effect of pH on production of

xylanase by Trichoderma reesei on xylan- and cellulose-based media. Appl

Microbiol Biotechnol 40:224-229.

Bayer EA, Belaich JP, Shoham Y and Lamed R. (2004) The cellulosomes:

multienzyme machines for degradation of plant cell wall polysaccharides.

Annu Rev Microbiol 58:521-554.

BioAlcohol Fuel Foundation (BAFF) Bought ethanol cars graph from

<http://www.baff.info/english/> Accessed October 2013.

Borkovich KA and Ebbole DJ. (2010) Cellular and molecular biology of

filamentous fungi. Washington D.C., USA: ASM Press.

64

Castro JF, Parra C, Yáñez-S M, Rojas J, Teixeira Mendonça R, Baeza J and

Freer J. (2013) Optimal pretreatment of Eucalyptus globulus by

hydrothermolysis and alkaline extraction for microbial production of

ethanol and xylitol. Ind Eng Chem Res 52:5713–5720.

Cavalieri D, McGovern PE, Hartl DL, Mortimer R and Polsinelli M. (2003)

Evidence for S. cerevisiae fermentation of ancient wine. J Mol Evol 57:226-

232.

Chandra RP, Bura R, Mabee WE, Berlin A, Pan X and Saddler JN (2007)

Substrate pretreatment: The key to effective enzymatic hydrolysis of

lignocellulosics? Adv Biochem Eng Biotechnol 108:67-93.

Chen Y, Stevens MA, Zhu Y, Holmes J and Xu H. (2013) Understanding of

alkaline pretreatment parameters for corn stover enzymatic

saccharification. Biotechnol Biofuels 6:8.

Ciolacu D, Ciolacu F and Popa VI. (2011) Amorphous cellulose – structure

and characterization. Cell Chem Technol 45: 13-21.

Collins T, Gerday C and Feller G. (2005) Xylanases, xylanase families

and extremophilic xylanases. FEMS Microbiol Rev 29:3-23.

Conway T. (1992) The Entner-Doudoroff pathway: history, physiology and

molecular biology. FEMS Microbiol Rev 103:1-28.

Dale BE, Leong CK, Pham TK, Esquivel VM, Rios I and Latimer VM.

(1996) Hydrolysis of lignocellulosics at low enzyme levels: application of

the AFEX process. Bioresour Technol 56: 111–116.

David H, Åkesson M and Nielsen J. (2003) Reconstruction of the central

carbon metabolism of Aspergillus niger. Eur J Biochem 270:4243–4253.

de Hoffmann E and Stroobant V. (2007) Mass Spectrometry: Principles

and Applications. 3rd ed. Chichester: John Wiley and Sons Ltd.

Den Haan R, Rose SH, Lynd LR and van Zyl WH. (2007) Hydrolysis and

fermentation of amorphous cellulose by recombinant Saccharomyces

cerevisiae. Metab Eng 9:87-94.

Dien BS, Cotta MA and Jeffries TW. (2003) Bacteria engineered for fuel

ethanol production: current status. Appl Microbiol Biotechnol 63:258-266.

65

DiPardo J. (2000) Outlook of biomass ethanol production and demand,

Washington, DC: U.S. Department of Energy, Energy Information

Administration. <ftp://tonto.eia.doe.gov/features/biomass.pdf> Accessed

July 2012.

Du B, Sharma LN, Becker C, Chen SF, Mowery RA, van Walsum GP and

Chambliss CK. (2010) Effect of varying feedstock-pretreatment chemistry

combinations on the formation and accumulation of potentially inhibitory

degradation products in biomass hydrolysates. Biotechnol Bioeng 107:430-

440.

Duff SJB and Murray WD. (1996). Bioconversion of forest products

industry waste cellulosics to fuel ethanol: a review. Bioresour Technol 55:1-

33.

Duff CJB, Cooper DG and Fuller OM. (1986) Evaluation of hydrolytic

potential of a crude cellulase from mixed cultivations of Trichoderma reesei

and Aspergillus phoenicis. Enzyme Microb Technol 8:305-308.

Dybkær R. (2001) Unit “katal” for catalytic activity. Pure Appl Chem

73:927–931.

El-Aneed A, Cohen A and Banoub J. (2009) Mass spectrometry, Review

of the basics: electrospray, MALDI, and commonly used mass analyzers.

Appl Spectrosc Rev 44:210-230.

Ellis LD, Holwerda EK, Hogsett D, Rogers S, Shao X, Tschaplinski T,

Thorne P and Lynd LR. (2012) Closing the carbon balance for fermentation

by Clostridium thermocellum (ATCC 27405). Bioresour Technol 103:293-

299.

Eom IY, Kim KH, Kim JY, Lee SM, Yeo HM, Choi IG and Choi JW. (2011)

Characterization of primary thermal degradation features of

lignocellulosic biomass after removal of inorganic metals by diverse

solvents. Bioresour Technol 102:3437-3444.

Eveleigh DE, Mandels M, Andretti R and Roche C. (2009) Measurement

of saccharifying cellulose. Biotechnol Biofuels 2:21.

Fargione J, Hill J, Tilman D, Polasky S and Hawthorne P. (2008) Land

clearing and the biofuel carbon debt. Science 319:1235-1238.

66

Farrell A, Plevin RJ, Turner BT, Jones AD, O’Hare M and Kammen D.

(2006) Ethanol can contribute to energy and environmental goals. Science

311: 506-508.

Friman L, Höglund H, Högberg H and Agnemo R. (2004) Tannin-Fe

impregnated mechanical pulp: Part II: Bleachability and yellowing. Nord

Pulp Pap Res J 19:525-531.

Fu GZ, Chan AW and Minns DE. (2003) Life cycle assessment of bio-

ethanol derived from cellulose. Int J Life Cycle Assess 8:137-141.

Fullbrook PD. (1996) Practical limits and prospects (kinetics). In Godfrey

T and West S. (eds) Industrial enzymology. 2nd ed. London, UK: MacMillan.

Galbe M and Zacchi G. (2007) Pretreatment of lignocellulosic materials

for efficient bioethanol production. Adv Biochem Eng Biotechnol 108:41-65.

Geddes CC, Mullinnix MT, Nieves IU, Peterson JJ, Hoffman RW, York

SW, Yomano LP, Miller EN, Shanmugam KT and Ingram LO. (2011)

Simplified process for ethanol production from sugarcane bagasse using

hydrolysate-resistant Escherichia coli strain MM160. Bioresour Technol

102:2702-2711.

Ghisalberti EL and Sivasithamparam K. (1991) Antifungal antibiotics

produced by Trichoderma sp. Soil Biol Biochem 23:1011-1020.

Ghose TK (1987) Measurement of cellulase activities. Pure Appl

Chem 59:257-268.

Goldemberg J (2007) Ethanol for a sustainable energy future. Science

315: 808-810.

Goldemberg J (2008) The Brazilian biofuels industry. Biotechnol Biofuels

1:6.

Grabek-Lejko D, Kurylenko OO, Sibirny VA, Ubiyvovk VM, Penninckx M

and Sibirny AA. (2011) Alcoholic fermentation by wild-type Hansenula

polymorpha and Saccharomyces cerevisiae versus recombinant strains with

an elevated level of intracellular glutathione. J Ind Microbiol Biotechnol

38:1853-1859.

Guilhaus M, Selby D and Mlynski V. (2000) Orthogonal acceleration

time-of-flight mass spectrometry. Mass Spectrom Rev 19:65-107.

67

Guo X, Cavka A, Jönsson LJ and Hong F. (2013) Comparisons of methods

for detoxification of spruce hydrolysate for bacterial cellulose production.

Microb Cell Fact 12:93.

Gupta RB and Demirbas A. (2010) Conventional ethanol production from

corn and sugar cane. In Gasoline, diesel and ethanol biofuels from grasses

and plants. New York, NY, USA, Cambridge University Press.

Gutiérrez-Correa M and Villena GK. (2012) Batch and repeated batch

cellulase production by mixed cultures of Trichoderma reesei and

Aspergillus niger or Aspergillus phoenicis. J Microbiol Biotechnol Res

2:929-935.

Hamelinck CN, van Hooijdonk G and Faaij APC. (2005). Ethanol from

lignocellulosic biomass: Techno-economic performance in short-, middle-

and long-term. Biomass Bioenerg 28:384–410.

Hammerschlag R. (2006) Ethanol's energy return on investment: A

survey of the literature 1999-present. Environ Sci Technol 40:1744-1750.

Hasunuma T, Sanda T, Yamada R, Yoshimura K, Ishii J and Kondo A.

(2011a) Metabolic pathway engineering based on metabolomics confers

acetic and formic acid tolerance to a recombinant xylose-fermenting strain

of Saccharomyces cerevisiae. Microb Cell Fact 10:2.

Hasunuma T, Sung KM, Sanda T, Yoshimura K, Matsuda F and Kondo A.

(2011b) Efficient fermentation of xylose to ethanol at high formic acid

concentrations by metabolically engineered Saccharomyces cerevisiae. Appl

Microbiol Biotechnol 90:997–1004.

Hauksson T, Markusson S, Einarsson K, Karlsdóttir SN, Einarsson Á,

Möller A and Sigmarsson Þ. (2013) Pilot testing of handling the fluids from

the IDDP-1 exploratory geothermal well, Krafla, NE Iceland. Geoth In

Press.

He J, Yu B, Zhang K, Ding X and Chen D. (2009) Expression of endo-1, 4-

beta-xylanase from Trichoderma reesei in Pichia pastoris and functional

characterization of the produced enzyme. BMC Biotechnol 9:56-66.

Himmel ME, Ruth MF and Wyman CE. (1999) Cellulase for commodity

products from cellulosic biomass. Curr Opin Biotechnol 10:358–364.

68

Horn SJ, Nguyen QD, Westereng B, Nilsen PJ and Eijsink VGH. (2012a)

Screening of steam explosion conditions for glucose production from non-

impregnated wheat straw. Biomass Bioenerg 35:4879-4886.

Horn SJ, Vaaje-Kolstad G, Westereng B and Eijsink VGH. (2012b) Novel

enzymes for the degradation of cellulose. Biotechnol Biofuels 5:45.

Hu F and Ragauskas AJ. (2012) Pretreatment and lignocellulosic

chemistry. Bioenerg Res 5:1043–1066.

Ilmén M, Thrane C and Penttilä M. (1996) The glucose repressor gene

cre1 of Trichoderma: isolation and expression a full-length and a truncated

mutant form. Mol Gen Genet 251:451-460.

Ilmén M, Saloheimo A, Onnela M-L and Penttilä M. (1997) Regulation of

cellulase gene expression in the filamentous fungus Trichoderma reesei.

Appl Environ Microbiol 63:1298-1306.

International Energy Agency (2011) Technology roadmap: Biofuels for

transportation. OECD/IEA, Paris, France.

Jin M, Gunawan C, Balan V and Dale BE. (2012) Consolidated

bioprocessing (CBP) of AFEX™-pretreated corn stover for ethanol

production using Clostridium phytofermentans at a high solids loading.

Biotechnol Bioeng 109:1929-1936.

Jönsson LJ, Palmqvist E, Nilvebrant N-O and Hahn-Hägerdal B. (1998)

Detoxification of wood hydrolysates with laccase and peroxidase from the

white-rot fungus Trametes versicolor. Appl Microbiol Biotechnol 49: 691-

697.

Jönsson LJ, Alriksson B and Nilvebrant N-O. (2013) Bioconversion of

lignocellulose: inhibitors and detoxification. Biotechnol Biofuels 6:16.

Jung CD, Yu JH, Eom IY and Hong KS. (2013) Sugar yields from

sunflower stalks treated by hydrothermolysis and subsequent enzymatic

hydrolysis. Bioresour Technol 138:1-7.

Karatzos SK, Edye LA and Doherty WO. (2012) Sugarcane bagasse

pretreatment using three imidazolium-based ionic liquids; mass balances

and enzyme kinetics. Biotechnol Biofuels 5:62.

69

Keating JD, Panganiban C and Mansfield SD. (2006) Tolerance and

adaptation of ethanologenic yeasts to lignocellulosic inhibitory compounds.

Biotechnol Bioeng 93:1196-1206.

Khatiwada D and Silveira S. (2009) Net energy balance of molasses-

based ethanol: The case of Nepal. Renew Sust Energ Rev 13:2515-2524.

Kim S and Dale BE. (2005). Environmental aspects of ethanol derived

from no-tilled corn grain: Nonrenewable energy consumption and

greenhouse gas emission. Biomass Bioenerg 28:475–489.

Kim TH, Kim JS, Sunwoo C and Lee YY. (2003) Pretreatment of corn

stover by aqueous ammonia. Bioresour Technol 90:39-47.

Klemm D, Heublein B, Fink HP and Bohn A. (2005) Cellulose: fascinating

biopolymer and sustainable raw material. Angew Chem Int Ed Engl

44:3358-3393.

Kubicek CP. (2013) Systems biological approaches towards

understanding cellulase production by Trichoderma reesei. J Biotechnol

163:133-142.

Kubicek CP, Komon-Zelazowska M and Druzhinina IS. (2008) Fungal

genus Hypocrea/Trichoderma: from barcodes to biodiversity. J Zhejiang

Univ Sci B 9:753-763.

Kuhls K, Lieckfeldt E, Samuels GJ, Kovacs W, Meyer W, Petrini O, Gams

W, Börner T and Kubicek CP. (1996) Molecular evidence that the asexual

industrial fungus Trichoderma reesei is a clonal derivative of the

ascomycete Hypocrea jecorina. Proc Natl Acad Sci 93:7755–7760.

La Grange DC, Pretorius IS, and van Zyl WH. (1996) Expression of a

Trichoderma reesei beta-xylanase gene (xyn2) in Saccharomyces cerevisiae.

Appl Environ Microbiol 62:1036–1044.

Landaeta R, Aroca G, Acevedo F, Teixeira JA and Mussatto SI. (2013)

Adaptation of a flocculent Saccharomyces cerevisiae strain to lignocellulosic

inhibitors by cell recycle batch fermentation. Appl Energ 102:124-130.

Larsson S, Palmqvist E, Hahn-Hägerdal B, Tengborg C, Stenberg K,

Zacchi G and Nilvebrant N-O. (1999a). The generation of fermentation

inhibitors during dilute acid hydrolysis of softwood. Enzyme Microb

Technol 24:151-159.

70

Larsson S, Reimann A, Nilvebrant N-O and Jönsson LJ. (1999b)

Comparison of different methods for the detoxification of lignocellulose

hydrolyzates of spruce. Appl Biochem Biotechnol 77–79:91–103.

Larsson S, Quintana-Sáinz A, Reimann A, Nilvebrant N-O and Jönsson

LJ. (2000) The influence of lignocellulose-derived aromatic compounds on

oxygen-limited growth and ethanolic fermentation by Saccharomyces

cerevisiae. Appl Biochem Biotechnol 84-86: 617-632.

Larsson S, Nilvebrant N-O and Jönsson LJ. (2001a) Effect of

overexpression of Saccharomyces cerevisiae Pad1p on the resistance to

phenylacrylic acids and lignocellulose hydrolysates under aerobic and

oxygen-limited conditions. Appl Microbiol Biotechnol 57:167-174.

Larsson S, Cassland P, and Jönsson LJ. (2001b) Development of

Saccharomyces cerevisiae with enhanced resistance to phenolic

fermentation inhibitors in lignocellulose hydrolysates by heterologous

expression of laccase. Appl Environ Microbiol 67:1163-1170.

Leboreiro J and Hilaly AK. (2013) Analysis of supply chain, scale factor,

and optimum plant capacity for the production of ethanol from corn stover.

Biomass Bioenerg 54:158-169.

Lee JW, Kim HY, Koo BW, Choi DH, Kwon M and Choi IG. (2008)

Enzymatic saccharification of biologically pretreated Pinus densiflora using

enzymes from brown rot fungi. J Biosci Bioeng 106:162-167.

Lee JW, Zhu JY, Scordia D and Jeffries TW. (2011) Evaluation of ethanol

production from corncob using Scheffersomyces (Pichia) stipitis CBS 6054

by volumetric scale-up. Appl Biochem Biotechnol 165:814-822.

Levasseur A, Drula E, Lombard V, Coutinho PM and Henrissat B. (2013)

Expansion of the enzymatic repertoire of the CAZy database to integrate

auxiliary redox enzymes. Biotechnol Biofuels 6:41.

Li Y, Tschaplinski TJ, Engle NL, Hamilton CY, Rodriguez M Jr, Liao JC,

Schadt CW, Guss AM, Yang Y and Graham DE. (2012) Combined

inactivation of the Clostridium cellulolyticum lactate and malate

dehydrogenase genes substantially increases ethanol yield from cellulose

and switchgrass fermentations. Biotechnol Biofuels 5:2.

Lima M A, Oliveira-Neto M, Kadowaki MAS, Rosseto FR, Prates ET,

Squina FM, Leme AFP, Skaf MS and Polikarpov I. (2013) Aspergillus niger

71

beta-glucosidase has a cellulase-like tadpole molecular shape: insights into

GH3 beta-glucosidases structure and function. J Biol Chem In Press.

Lin Y and Tanaka S. (2006) Ethanol fermentation from biomass

resources: current state and prospects. Appl Microbiol Biotechnol 69:627-

642.

Lindén T, Peetre J and Hahn-Hägerdal B. (1992) Isolation and

characterization of acetic acid-tolerant galactose-fermenting strains of

Saccharomyces cerevisiae from a spent sulphite liquor fermentation plant.

Appl Environ Microb 58:1661-1669.

Liu ZL, Slininger PJ, Dien BS, Berhow MA, Kurtzman CP and Gorsich SW.

(2004) Adaptive response of yeasts to furfural and 5-

hydroxymethylfurfural and new chemical evidence for HMF conversion to

2,5-bis-hydroxymethylfuran. J Ind Microbiol Biotechnol 31:345–352.

Lo Leggio L, Welner D and De Maria L. (2012) A structural overview of

GH61 proteins – fungal cellulose degrading polysaccharide

monooxygenases. Comput Struct Biotechnol J 2:e201209019.

Lynd L Reflecting on progress in the bioenergy sector. (2011) Nat

Biotechnol 29:196.

Lynd LR, Weimer PJ, Van Zyl WH and Pretorius IS. (2002) Microbial

cellulose utilization: fundamentals and biotechnology. Microbiol Mol Biol

Rev 66:506-577.

Lynd LR, Laser MS, Bransby D, Dale BE, Davison D, Hamilton R, Himmel

M, Keller M, McMillan JD, Sheehan J and Wyman CE. (2008) How biotech

can transform biofuels. Nat Biotechnol 26:169-172.

Lynd LR, Zyl WH, McBride JE and Laser M. (2005) Consolidated

bioprocessing of cellulosic biomass: an update. Curr Opin Biotechnol

16:577-583.

Macedo IC, Leal MRLV and da Silva JEAR. (2004) Assessment of

greenhouse gas emissions in the production and use of fuel ethanol in

Brazil. São Paulo State Environment Secretariat

<http://www.wilsoncenter.org/events/docs/brazil.unicamp.macedo.greenh

ousegas.pdf>

Accessed July 2013.

72

Margolles-Clark E, Ilmén M and Penttilä M. (1997) Expression patterns of

ten hemicellulase genes of the fungus Trichoderma reesei on various carbon

sources. J Biotechnol 57:167-179.

Marques G, del Río JC and Gutiérrez A. (2010) Lipophilic extractives

from several nonwoody lignocellulosic crops (flax, hemp, sisal, abaca) and

their fate during alkaline pulping and TCF/ECF bleaching. Bioresour

Technol 101:260-267.

Martín C and Jönsson L.J. (2003) Comparison of the resistance of

industrial and laboratory strains of Saccharomyces and

Zygosaccharomyces to lignocellulosic-derived fermentation inhibitors.

Enzyme Microbial Technol 32:386–395.

Martín C, Galbe M, Wahlbom CF, Hahn-Hägerdal B and Jönsson LJ.

(2002a) Ethanol production from enzymatic hydrolysates of sugarcane

bagasse using recombinant xylose-utilising Saccharomyces cerevisiae.

Enzyme Microb Technol 31:274-282.

Martín C, Galbe M, Nilvebrant N-O and Jönsson LJ. (2002b) Comparison

of the fermentability of enzymatic hydrolyzates of sugarcane bagasse

pretreated by steam explosion using different impregnating agents. Appl

Biochem Biotechnol 98-100:699-716.

Martín C, Marcet M, Almazán O, and Jönsson LJ. (2007) Adaptation of a

recombinant xylose-utilizing Saccharomyces cerevisiae strain to a

sugarcane bagasse hydrolysate with high content of fermentation

inhibitors. Bioresour Technol 98:1767-1773.

Martín C, Puls J, Saake B and Schreiber A. (2011) Effect of glycerol

pretreatment on component recovery and enzymatic hydrolysis of

sugarcane bagasse. Cellulose Chem Technol 45:487–494.

Martínez A, Rodriguez ME, York SW, Preston JF and Ingram LO. (2000)

Effects of Ca(OH)2 treatments (“overliming”) on the composition and

toxicity of bagasse hemicellulose hydrolysates. Biotech Bioeng 69:526-536.

Martínez D, Berka RM, Henrissat B, Saloheimo M, Arvas M, Baker SE,

Chapman J, Chertkov O, Coutinho PM, Cullen D, Danchin EGJ, Grigoriev IV,

Harris P, Jackson M, Kubicek CP, Han CS, Ho I, Larrondo LF, Lopez de Leon

A, Magnuson JK, Merino S, Misra M, Nelson B, Putnam N, Robbertse B,

Salamov AA, Schmoll M, Terry A, Thayer N, Westerholm-Parvinen A, Schoch

CL, Yao J, Barabote R, Nelson MA, Detter C, Bruce D, Kuske CR, Xie G,

73

Richardson P, Rokhsar DS, Lucas SM, Rubin EM, Dunn-Coleman N, Ward

M and Brettin TS. (2008) Genome sequencing and analysis of the biomass-

degrading fungus Trichoderma reesei (syn. Hypocrea jecorina). Nat

Biotechnol 26:553-560.

Mohagheghi A, Ruth M and Schnell DJ. (2006) Conditioning

hemicellulose hydrolysates for fermentation: effects of overliming pH on

sugar and ethanol yields. Proc Biochem 41:1806-1811.

Montenecourt BS. (1983) Trichoderma reesei cellulases. Trends

Biotechnol 1:156-161.

Moreira JR and Goldemberg J. (1999) The alcohol program. Energ Policy

27:227-229.

Mosier N, Wyman C, Dale BE, Elander R, Lee YY, Holtzapple M and

Ladisch M. (2005) Features of promising technologies for pretreatment of

lignocellulosic biomass. Bioresour Technol 96: 673-686.

Mubako S and Lant C. (2008) Water resource requirements of

corn‐based ethanol. Water Resour Res 44:W00A02.

Mussatto SI, Santos JC and Roberto IC. (2004) Effect of pH and activated

charcoal adsorption on hemicellulosic hydrolysate detoxification for xylitol

production. J Chem Technol Biotechnol 79:590-596.

Narayanaswamy N, Faik A, Goetz DJ and Gu T. (2011) Supercritical

carbon dioxide pretreatment of corn stover and switchgrass for

lignocellulosic ethanol production. Bioresour Technol 102:6995-7000.

New York Times, Sept. 20, 1925, Ford Predicts Fuel from Vegetation.

Nilvebrant NO, Reimann A, Larsson S and Jönsson LJ. (2001)

Detoxification of lignocellulose hydrolysates with ion-exchange resins. Appl

Biochem Biotechnol 91-93:35-49.

Nilvebrant NO, Persson P, Reimann A, De Sousa F, Gorton L and Jönsson

LJ. (2003) Limits for alkaline detoxification of dilute-acid lignocellulose

hydrolysates. Appl Biochem Biotechnol 107:615-628.

Niven RK. (2005) Ethanol in gasoline: environmental impacts and

sustainability. Renew Sust Energ Rev 9:535–555.

74

Nölling J, Breton G, Omelchenko MV, Makarova KS, Zeng Q, Gibson R,

Lee HM, Dubois J, Qiu D, Hitti J, Wolf YI, Tatusov RL, Sabathe F, Doucette-

Stamm L, Soucaille P, Daly MJ, Bennett GN, Koonin EV and Smith DR.

(2010) Genome sequence and comparative analysis of the solvent-

producing bacterium Clostridium acetobutylicum. J Bacteriol 183:4823-

4838.

Nordlund C, Mårald E and Rosvall O. (2012) Forests, Technoscience and

the Future: Some Thoughts about Visions and Challenges. In Nyström M.

(ed) Being and acting in times of (un)certainty. Uppsala, Sweden: Cemus.

Öhgren K, Bura R, Lesnicki G, Saddler J and Zacchi G. (2007) A

comparison between simultaneous saccharification and fermentation and

separate hydrolysis and fermentation using steam-pretreated corn stover.

Process Biochem 42:834-839.

Olofsson K, Bertilsson M and Lidén G. (2008) A short review on SSF - an

interesting process option for ethanol production from lignocellulosic

feedstocks. Biotechnol Biofuels 1:7.

Olson DG, McBride JE, Shaw JA and Lynd LR. (2012) Recent progress in

consolidated bioprocessing. Curr Opin Biotechnol 23:396-405.

Palmqvist E, Almeida J and Hahn-Hägerdal B. (1999) Influence of

furfural on anaerobic glycolytic kinetics of Saccharomyces cerevisiae in

batch culture. Biotechnol Bioeng 62: 447–454.

Pan X, Gilkes N, Kadla J, Pye K, Saka S, Gregg D, Ehara K, Xie D, Lam D

and Saddler J. (2006) Bioconversion of hybrid poplar to ethanol and co-

products using an organosolv fractionation process: optimization of

process yields. Biotechnol Bioeng 94:851-861.

Park YC and Kim JS (2011) Comparison of various alkaline pretreatment

methods of lignocellulosic biomass. Energy 47:31-35.

Park CS, Chang CC and Ryu DDY. (2000) Expression and high-level

secretion of Trichoderma reesei endoglucanase I in Yarrowia lipolytica. Appl

Biochem Biotechnol 87:1-15.

Pedersen M, Johansen KS and Meyer AS. (2011). Low temperature

lignocellulose pretreatment: effects and interactions of pretreatment pH

are critical for maximizing enzymatic monosaccharide yields from wheat

straw. Biotechnol Biofuels 4:19.

75

Pel HJ, de Winde JH, Archer DB, Dyer PS, Hofmann G, Schaap PJ, Turner

G, de Vries RP, Albang R, Albermann K, Andersen MR, Bendtsen JD, Benen

JA, van den Berg M, Breestraat S, Caddick MX, Contreras R, Cornell M,

Coutinho PM, Danchin EG, Debets AJ, Dekker P, van Dijck PW, van Dijk A,

Dijkhuizen L, Driessen AJ, d'Enfert C, Geysens S, Goosen C, Groot GS, de

Groot PW, Guillemette T, Henrissat B, Herweijer M, van den Hombergh JP,

van den Hondel CA, van der Heijden RT, van der Kaaij RM, Klis FM, Kools

HJ, Kubicek CP, van Kuyk PA, Lauber J, Lu X, van der Maarel MJ,

Meulenberg R, Menke H, Mortimer MA, Nielsen J, Oliver SG, Olsthoorn M,

Pal K, van Peij NN, Ram AF, Rinas U, Roubos JA, Sagt CM, Schmoll M, Sun

J, Ussery D, Varga J, Vervecken W, van de Vondervoort PJ, Wedler H,

Wösten HA, Zeng AP, van Ooyen AJ, Visser J and Stam H. (2007) Genome

sequencing and analysis of the versatile cell factory Aspergillus niger CBS

513.88. Nat Biotechnol 27:221-231.

Penttilä M, Teeri TT, Nevalainen H and Knowles JKC. (1991) The

molecular biology of Trichoderma reesei and its application to

biotechnology. In Applied Molecular Genetics of Fungi. New York, NY, USA:

Cambridge University Press.

Percival Zhang YH, Himmel ME and Mielenz JR. (1999) Outlook for

cellulase improvement: screening and selection strategies. Biotechnol Adv

24:452-481.

Persson P, Andersson J, Gorton L, Larsson S, Nilvebrant NO and Jönsson

LJ. (2002) Effect of different forms of alkali treatment on specific

fermentation inhibitors and on the fermentability of lignocellulose

hydrolysates for production of fuel ethanol. J Agric Food Chem 50:5318-

5325.

Pienkos PT and Zhang M. (2009) Role of pretreatment and conditioning

processes on toxicity of lignocellulosic biomass hydrolysates. Cellulose

16:743-762.

Pimentel D. (2001) Limits of Biomass Utilization. In Encyclopedia of

Physical Science and Technology 3rd ed. San Diego, CA, USA: Academic

Press.

Pimentel D. (2003) Ethanol fuels: Energy balance, economics, and

environmental impacts are negative. Nat Resour Res 12:127-134.

76

Pimentel D and Patzek TW. (2005) Ethanol production using corn,

switchgrass, and wood; Biodiesel production using soybean and sunflower”

Nat Resour Res 14:65-76.

Piper P, Calderon CO, Hatzixanthis K and Mollapour M. (2001) Weak acid

adaptation: the stress response that confers yeasts with resistance to

organic acid food preservatives. Microbiology 147:2635-2642.

Portnoy T, Margeot A, Linke R, Atanasova L, Fekete E, Sándor E, Hart L,

Karaffa L, Druzhinina IS, Seiboth B, Le Crom S and Kubicek CP. (2011) The

CRE1 carbon catabolite repressor of the fungus Trichoderma reesei a

master regulator of carbon assimilation. BMC Genomics 12:269-271.

Poulopoulos SG, Samaras DP and C. Philippopoulos. (2001) Regulated

and unregulated emissions from an internal combustion engine operating

on ethanol-containing fuels. Atmos Environ 35:4399–4406.

Pu Y, Zhang D, Singh PM and Ragauskas AJ. (2007) The new forestry

biofuels sector. Biofuels Bioprod Biorefin 2:58-73.

Ragauskas AJ, Williams CK, Davison BH, Britovsek G, Cairney J, Eckert

CA, Frederick WJ Jr, Hallett JP, Leak DJ, Liotta CL, Mielenz JR, Murphy R,

Templer R and Tschaplinski T. (2006) The path forward for biofuels and

biomaterials. Science 311:484–489.

Ranatunga TD, Jervis J, Helm RF, McMillan JD and Wooley RJ. (2000)

The effect of overliming on the toxicity of dilute acid pretreated

lignocellulosics: the role of inorganics, uronic acid and ether-soluble

organics. Enzyme Microb Technol 27:240-247.

Ray MJ, Leak DJ, Spanu PD and Murphy RJ. (2010) Brown rot fungal

early stage decay mechanism as a biological pretreatment for softwood

biomass in biofuel production. Biomass Bioenerg 34:1257-1262.

Remington JP. (2006) Remington: The science and practice of

pharmacy. Troy DB and Beringer P. (Eds.). Baltimore, MD, USA: Lippincott

Williams and Williams.

Renewable Fuels Association (RFA), (2012) Acelerating Industry

Innovation – 2012 Ethanol Industry Outlook.

<[http://ethanolrfa.3cdn.net/d4ad995ffb7ae8fbfe_1vm62ypzd.pdf]>

Accessed June 2013.

77

Rolfe RT and Rolfe FW. (1925) The romance of the fungus world: an

account of fungus life in its numerous guises, both real and imaginary. New

York, NY, USA: Dover Publications.

Rördam Andersen M, Lynge Nielsen M and Nielsen J. (2008) Metabolic

model integration of the bibliome, genome, metabolome and reactome of

Aspergillus niger. Mol Sys Biol 4:178-191.

Rose SH and van Zyl WH. (2002) Constitutive expression of the

Trichoderma reesei β-1,4-xylanase gene (xyn2) and the β-1,4-

endoglucanase gene (egI) in Aspergillus niger in molasses and defined

glucose media. Appl Microbiol Biotechnol 58:461–468.

Rowell RM, Pettersen R, Han JS, Rowell JS and Tshabalala MA. (2005)

Cell wall chemistry. In Handbook of wood chemistry and wood composites.

Rowell RM. (ed) Boca Raton, FL, USA: CRC Press.

Ruijter GJG and Visser J. (1997) Carbon repression in Aspergilli. FEMS

Microbiol Lett 151:103-114.

Saddler J and Mabee W. (2007) Choosing biorefining platforms for the

commercialisation of the biomass-to-ethanol process. Biomass Bioenerg

31:1-5.

Saha BC, Itena LB, Cotta MA and Wub YV (2005) Dilute acid

pretreatment, enzymatic saccharification and fermentation of wheat straw

to ethanol. Process Biochem 40:3693–3700.

Sànchez I Nogué V, Bettiga M and Gorwa-Grauslund MF. (2012) Isolation

and characterization of a resident tolerant Saccharomyces cerevisiae strain

from a spent sulfite liquor fermentation plant. AMB Express 2:68.

Sárvári Horváth I, Sjöde A, Nilvebrant N-O, Zagorodni A and Jönsson LJ.

(2004) Selection of anion exchangers for detoxification of dilute-acid

hydrolysates from spruce. Appl Biochem Biotechnol 114:525-538.

Sárvári Horváth I, Sjöde A, Alriksson, Jönsson LJ and Nilvebrant NO.

(2005) Critical conditions for improved fermentability during overliming

of acid hydrolysates from spruce. Appl Biochem Biotechnol 121-124:1031-

1044.

Searchinger T, Heimlich R, Houghton RA, Dong F, Elobeid A, Fabiosa J,

Tokgoz S, Hayes D and Yu TH. (2008) Use of U.S. croplands for biofuels

78

increases greenhouse gases through emissions from land-use change.

Science 319:1238-1240.

Shallom D and Shoham Y. (2003) Microbial hemicellulases. Curr Opin

Microbiol 6:219-228.

Shapouri H, Duffield JA and Wang M. (2002), The energy balance of corn

ethanol: an update. U.S. Department of Agriculture, Office of Chief

Economist, Office of Energy Policy and New Use, Agricultural Economic

Report No. 814, Washington, D.C., USA.

Sheehan J, Camobreco V, Duffield J, Graboski M and Shapouri H. (1998)

Life‐cycle inventory of biodiesel and petroleum diesel for use in an urban

bus. NREL/SR‐580‐24089. Golden, CO, USA.: National Renewable Energy

Laboratory.

Sims REH, Mabee W, Saddler JN and Taylor M. (2010) An overview of

second generation biofuel technologies. Bioresour Technol 101:1570-1580.

Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J and Templeton D. (2006)

Determination of sugars, byproducts, and degradation products in liquid

fraction process samples. Golden, CO, USA: National Renewable Energy

Laboratory.

Soni SK (2007) Microbes: A Source Of Energy For 21st Century. New

Delhi, India: New India Publishing Agency.

Subtil T and Boles E. (2012) Competition between pentoses and glucose

during uptake and catabolism in recombinant Saccharomyces cerevisiae.

Biotechnol Biofuels 5:14.

Suess HS (2010) Pulp Bleaching Today. Berlin, Germany: Walter de

Gruyter GmbH & Co. KG.

Sun Y and Chang J. (2002) Hydrolysis of lignocellulosic materials for

ethanol production: a review. Bioresource Technol 83:1-11.

Sun N, Rahman M, Qin Y, Maxim ML, Rodríguez H and Rogers RD

(2009) Complete dissolution and partial delignification of wood in the ionic

liquid 1-ethyl-3-methylimidazolium acetate. Green Chem 11:646–655.

79

Tao Z, Goodisman J and Souid A-K. (2008) Oxygen measurement via

phosphorescence: Reaction of sodium dithionite with dissolved oxygen. J

Phys Chem A 112:1511-1518.

Thompson DN, Campbell T, Bals B, Runge T, Teymouri F and Ovard LP.

(2013) Chemical preconversion: application of low-severity pretreatment

chemistries for commoditization of lignocellulosic feedstock. Biofuels 4:323-

340.

van Maris AJ, Abbott DA, Bellissimi E, van den Brink J, Kuyper M, Luttik

MA, Wisselink HW, Scheffers WA, van Dijken JP and Pronk JT. (2006)

Alcoholic fermentation of carbon sources in biomass hydrolysates by

Saccharomyces cerevisiae: current status. A Van Leeuw J Microb 90:391-

418.

van Zyl WH, Rose SH, Trollope K, and Görgens JF. (2010) Fungal β-

mannanases: Mannan hydrolysis, heterologous production and

biotechnological applications. Process Biochem 45: 1203-1213.

Viikari L, Alapuranen M, Puranen T, Vehmaanperä J and Siika-Aho M.

(2007) Thermostable enzymes in lignocellulose hydrolysis. Adv Biochem

Eng Biotechnol 108:121-145.

Vogel KP and Jung HJG. (2001) Genetic modification of herbaceous

plants for feed and fuel. Crit Rev Plant Sci 20:15-49.

von Sivers M, Zacchi G, Olsson L and Hahn-Hägerdal B. (1994) Cost

analysis of ethanol production from willow using recombinant Escherichia

coli. Biotechnol Prog 10:556–560.

Wahlbom CF and Hahn-Hägerdal B. (2002) Furfural, 5-hydroxymethyl

furfural, and acetoin act as external electron acceptors during anaerobic

fermentation of xylose in recombinant Saccharomyces cerevisiae.

Biotechnol Bioeng 78:172-178.

Wang M (2005) Updated energy and greenhouse gas emissions results of

fuel ethanol. Proceedings of the 15th International Symposium on Alcohol

Fuels: 26–28 September 2005; San Diego, CA.

Wang M, Saricks C and Santini D. (1999). Effects of fuel ethanol use on

fuel-cycle energy and greenhouse gas emissions. U.S. Department of

Energy, Argonne National Laboratory, Center for Transportation Research,

Argonne, IL, USA..

80

Wang M, Wu M and Hou H (2007) Life-cycle energy and greenhouse gas

emission impact of different corn ethanol plant types. Environ Res Lett

2:024001.

Watanabe T. (2013) Introduction: Potential of cellulosic ethanol. In

Faracao V. (ed) Lignocellulose conversion. Berlin, Germany: Springer.

Wikström M, Folkesson A and Alvfors P (2011) First experiences of

ethanol hybrid buses operating in public transport. Sust Transport 13:3653-

3668.

Williams N. (2005) Chasing the sun: Solar adventures around the world.

Gabriola Island, Canada: New Society Publishers.

Wingren A, Galbe M and Zacchi G. (2003) Techno-economic evaluation of

producing ethanol from softwood: Comparison of SSF and SHF and

identification of bottlenecks. Biotechnol Prog 19:1109-1117.

Wollnik H. (1993) Time-of-flight mass analyzers. Mass Spectrom Rev

12:89–114.

Wright J, Bellissimi E, de Hulster E, Wagner A, Pronk JT and van Maris

AJ. (2011) Batch and continuous culture-based selection strategies for

acetic acid tolerance in xylose-fermenting Saccharomyces cerevisiae. FEMS

Yeast Res 11:299-306.

Wyman CE, Decker SR, Himmel ME, Brady JW, Skopec CE and Viikari L.

(2005) Hydrolysis of cellulose and hemicellulose. In Polysaccharides:

Structural diversity and functional versatility. New York, NY, USA: INC.

Ximenes E, Kim Y, Mosier N, Dien B and Ladisch M. (2011) Deactivation

of cellulases by phenols. Enzyme Microb Technol 48:54-60.

Yamashita Y, Kurosumi A, Sasaki C and Nakamura Y. (2008) Ethanol

production from paper sludge by immobilized Zymomonas mobilis.

Biochem Eng J 42:314-319.

Yang B, Dai Z, Ding S-Y and Wyman CE. (2011) Enzymatic hydrolysis of

cellulosic biomass. Biofuels 2:421-450.