Bio-sustainable Control of the Blue Stain Fungi Aureobasidium pullulans on Exterior Wood Coatings

PhD Thesis by Jonas Stenbæk

Supervisor Associated prof. Bo Jensen

Section of Microbiology, Department of Biology, Faculty of Science University of Copenhagen

This thesis has been submitted to the PhD School of The Faculty of Science, University of Copenhagen Cover page: eSEM image of Aureobasidium pullulans (image: Peter Falkman and Jonas Stenbæk).

This is for my mother

Anne

September 22nd 1957 - July 27th 2015 Abstract

Imminent requirements and demands to the composition of protective wood coatings and the use of commercial biocides lead to an increased and necessary interest by the paint industry to find and develop alternative solutions within substitutions of raw materials and new innovative anti-fungal strategies. Besides comprehensive future regulations in the requirements and laws within the use of fossil-based raw materials, companies also have to take the increasing consumer awareness of environmental responsibility into account necessitating the companies to use more environmentally friendly solutions in the production, composition and service life of the protective coatings.

Today’s commercial biocides in the wood coating industry, e.g. IPBC, are proven environmental toxic and exposed to future regulations or indeed complete exclusions so new bio-sustainable alternative solutions are demanded. In this PhD study, a number of alternative and innovative solutions and strategies have been investigated and a number of possible new solutions and reviews are presented. The focus has been on the control of blue stain fungi e.g. Aureobasidium pullulans that lead to discoloration on processed wood and wood protective coatings.

In this study, a number of anionic compounds like fatty acid-based emulsifiers have been screened for anti-fungal properties. The best candidate, a Sodium Caproyl Lactylate, has further been tested in wood panel tests and the mode of action on the fungal cell membrane has been investigated. The compound exhibits promising anti-fungal properties, however, more in-depth investigation on the compound and closely similar compounds are recommended before an actual and useable technology are accessible. Other compounds like silica-encapsulated enzymes and nanoclay have also been investigated and interesting preliminary results makes ground for further studies based on promising anti-fungal properties when added to a standard coating without commercial biocides and tested in environmental mould growth chambers.

As a non-additive biocidal solution the structure of the topography on the surface on the coating has been studied. In this study, manipulation of the micro structure on the coating surface has been performed and revealed that small scratches-structure (between 4-20 µm) had an inhibitory effect on the attachment on fungal conidia from moulds while bigger structure scratches (>50 µm) did not seem to have a notable effect. Abstract

A comprehensive study in the role of an interfacial protein, hydrophobin, in the surface growth of Aureobasidium pullulans resulted in bioinformatic identification of two hydrophobin genes, first ever described in a blue stain fungi. The full genome of Aureobasidium pullulans (De Bary) P268 was sequenced and the bioinformatic information led to molecular gene manipulation in the attempt to clarify the role of hydrophobins in surface growth of Aureobasidium pullulans.

Overall, this PhD thesis presents a range of studies in alternative and bio-sustainable solutions and strategies in controlling attacks of blue stain fungi on wood protective coatings. The results provide a state-of-art platform for further research and development of alternative options in the control of moulds in coatings. Resumé

Stigende krav til sammensætningen af træbeskyttende coatings og kommercielle biocider forventes at føre til en øget interesse hos coating-industrien for både at udvikle alternative løsninger inden for substitution af råmaterialer og for at finde nye, innovative strategier inden for svampebekæmpelse. Udover de stigende krav og regler inden for brugen af fossilt baserede råvarer, er virksomhederne også nødt til at tage hensyn til forbrugernes øgede bevidsthed om miljøansvarlige forbrugsvaner. Dette nødvendiggør, at virksomhederne fremover i højere grad må anvende mere miljøvenlige løsninger inden for produktion, produktindhold og produktets levetid.

Nutidens kommercielt anvendte biocider i coating-industrien, eksempelvis IPBC, er giftige for miljøet og vil forventeligt på et tidspunkt blive udsat for restriktioner eller ultimative forbud. Nye bio-bæredygtige alternative løsninger efterspørges derfor allerede på nuværende tidspunkt af industrien.

Dette ph.d.-studie undersøger og præsenterer en række alternative og innovative løsninger samt strategier inden for bæredygtige og miljøvenlige biocider, som er mulige fremtidige løsninger. Fokus har været på bekæmpelse af blåsplint svampe, som eksempelvis Aureobasidium pullulans, der fører til misfarvning af forarbejdet træ og træbeskyttelse, herunder udendørs træmaling.

I dette studie er en række anioniske forbindelser, som fedtsyre-baserede emulgatorer, blevet screenet for svampehæmmende egenskaber. Den bedste kandidat - en Sodium Caproyl Lactylat (SCL) - er blevet testet yderligere i træpanels-tests. Derudover er virkemåden af SCL på svampens cellemembran blevet undersøgt, og samlet set udviser SCL lovende svampebekæmpende egenskaber. Dog anbefales en udvidet undersøgelse af SCL, og dens nært beslægtede stoffer, før en endelig og brugbar teknologi kan introduceres.

Andre additiver, herunder silica-indkapslede enzymer og nanoclay, er også blevet undersøgt, og foreløbige interessante resultater kan danne grundlag for yderligere studier. Dette er baseret på lovende svampehæmmende egenskaber, når additiverne blandes i en standard coating uden indhold af kommercielle biocider og efterfølgende testes i skimmelvækst-kamre.

Coating-topografien i mikro-skala er blevet undersøgt med henblik på at belyse eventuelle muligheder for svampehæmmende effekter uden brug af additiver. I dette studie er Resumé

mikrostruktur på coating-overfladen blevet manipuleret, så den har fået en præcis ridser- struktur. Det viste sig, at en fin typografi (mellem 4-20 µm) havde en hæmmende virkning på vedhæftningen af konidier fra skimmelsvampe, mens større struktur-ridser (> 50 µm) ikke syntes at have en mærkbar hæmmende virkning.

En omfattende undersøgelse af et overfladeaktivt protein, kaldet hydrofobin, havde fokus på proteinets rolle ved overfladevækst hos blåsplint-svampen Aureobasidium pullulans. Undersøgelsen resulterede i bioinformatisk identifikation af to hydrofobin-gener i Aureobasidium pullulans. Det er de første hydrofobiner, som er beskrevet i en blåsplintssvamp. Aureobasidium pullulans (De Bary) P268 er blevet sekventeret, og det fulde genom er nu tilgængeligt. Bioinformatiske oplysninger førte til molekylærbiologisk genmanipulation i forsøget på at klarlægge hydrofobinernes rolle ved overfladevækst hos Aureobasidium pullulans.

Samlet set præsenterer denne ph.d.-afhandling både en række studier i alternative og bio- bæredygtige løsninger samt strategier for kontrol af svampeangreb fra blåsplints svampe på træbeskyttende coatings. Resultaterne giver en state-of-art platform for yderligere forskning og udvikling af alternative og innovative biocider.

Table of Contents

Table of contents 1

Preface and Acknowledgements 3

List of Manuscripts 5

Conference Proceedings 6

PhD Course Portfolio 9

Teaching and Assistant Supervision 10

Change of Scientific Environment 11

Abbreviations 12

Chapter 1: Background and Aims 15

1.1 Superior Bio Based Coating System for Exterior Wood Applications 1.2 The overall aim: Control of A. pullulans in bio-sustainable coating systems for ex. wood 1.2.1 Time line 1.2.2 Structure of this PhD thesis 1.3 References Chapter 1

Chapter 2: Introduction 21

2.1 Blue Stain and protective coatings 2.2 IPBC 2.3 Aureobasidium pullulans 2.3.1 Morphology and life cyclus. 2.3.2 Pullulan 2.4 Methods and materials 2.4.1 Strains 2.4.2 Full-genome sequences 2.4.3 Growth tests 2.4.4 Wood panel tests in Environmental Chamber 2.4.5 Protoplasts 2.5 Chapter 2 references

Chapter 3: Fatty acid-based emulsifiers: possible biological agents 43

3.1 Introduction 3.2 Candidates 3.2.1 Lactylates 3.3 Mode of action 3.3.1 Supported lipid bilayers (SLB) 3.3.2 Quartz Crystal Microbalance with Dissipation (QCM-D) 3.4 Preliminary results 3.5 Conclusions and perspectives for further research 3.6 References Chapter 3

1 Table of Contents

Chapter 4: Enzymes, Nano clay and other additives 55 4.1 Introduction 4.2 Enzymes 4.2.1 Pullulanase 4.2.2 Glucanex 4.2.3 Aerogel – encapsulation of enzymes. 4.3 Nanoclay 4.4 Materials and Methods 4.4.1 Enzymes 4.4.2 Nanoclay 4.5 Preliminary results and discussions 4.5.1 Enzymes 4.5.1 Nanoclay 4.6 Conclusions and perspectives for further research 4.7 References Chapter 4

Chapter 5: Topography; Surface Structure on Coatings 71

5.1 Introduction 5.2 Materials and Methods 5.3 Conclusions and perspectives for further research

Chapter 6: Hydrophobins; a Unique Protein in Fungal Surface Living 73

6.1 Introduction 6.1.1 Nature of the Hydrophobins 6.1.2 Classification and protein structure 6.1.3 Properties 6.1.4 Aims of the work related to hydrophobins 6.2 Results and discussion 6.2.1 Analysis of the Aureobasidium pullulans strain P268 hydrophobins 6.2.2 Disruption of the hydrophobin encoding genes, hfbA and hfbB 6.2.2.1 Evaluation of potential double gene knockout strategies 6.2.2.2 Prac. workflow and pre. results from chosen gene knockout strategy 6.3 Conclusions and perspectives for further research 6.4 References Chapter 6

Chapter 7: Metagenomics; Soil micro-eukaryotes 105

7.1 Additional work

Manuscript I I

Manuscript II II

Manuscript III III

Manuscript IV IV

Manuscript V V

Poster session VI

Co-Authorship Statement VII

2 Preface and Acknowledgements

The work presented in this PhD thesis concludes three years of research performed at Section of Microbiology, Department of Biology, at the University of Copenhagen. The principal supervision was led by associated professor Bo Jensen.

Most of the practical research has been performed at Section of Microbiology, but additional work was conducted at: University of Tennessee, Malmö University, PPG Industries/Dyrup A/S, Danish Technological Institute, SP Sweden and Department of Chemistry at University of Copenhagen.

The PhD project was funded by The Danish Innovation Fund (previous: The Danish Advanced Technology Foundation) and by the University of Copenhagen.

I would like to thank all the partners in the project “Superior Bio Based Coating System for Exterior Wood Applications” associated with my PhD thesis.

A special thanks goes to David Löf, PPG Industries, who have been a strongly support right from the start of my project. Also, thanks to my two co-PhD fellows in the project, Miroslav Nikolic and Hiep Dinh Nguyen for monthly meetings and scientific discussions.

I would like to thank my supervisor, Bo Jensen, for believing in my qualities from the very beginning and for good scientific supervision. Also, thanks to Leise Riber and Ib Schneider, Coatzymes, for useful inputs and supervision almost every day at Bo’s office. A special thanks to Leise Riber for a lot of help with the hydrophobins.

I would also like to thank:

 Danish Technological Institute, especially Berit, Anne Christine, Morten, Mark, Niels, Thomas, Elisabeth and Trine for always being helpful with scientific issues and practical work in the lab.  Adam Taylor (University of Tennessee) and Jeff Lloyd (Nisus Corp.) for making my visit to Knoxville, USA, a great scientific and personal experience despite the fact that not all of our experiments turned out the way we hoped for.. Thanks to the University of Tennessee in general for letting me do my research at the Center for Renewable Carbon.

3 Preface and Acknowledgements

 All people at Section of Microbiology, University of Copenhagen. Thanks for contributing to a nice social and scientific environment. A special thanks to Karin Vestberg and Anette Løth for helping me every day in the lab. I would also like to thank Samuel Jacquiod for introducing me to metagenomics.  The people at Dyrup A/S in Søborg, Denmark for assistance and help whenever I was in need of it.  Marité Cardenas at Malmö University for patient help and assistance with the nano science and the technical equipment.  SP Sweden for introducing me to different techniques and introductions in their lab.  Last, I would like to thank my family for being patient, indulgent and supportive in my own personal PhD adventure. I look forward to spend more of my time with you.

Jonas Stenbæk, Copenhagen 2015

4 List of manuscripts

Manuscript I

Jonas Stenbæk, David Löf, Peter Falkman, Bo Jensen and Marité Cárdenas (submitted July 2015, PLOS ONE): Alternative Anionic Bio-sustainable Anti-fungal Agents: Investigation of their Modes of Action on the Fungal Cell Membrane.

Manuscript II

Jonas Stenbæk, Miroslav Nikolic, Berit Lindegaard, David Löf, Bo Jensen and Anne Christine Steenkjær Hastrup (2015 ready for submission): The influence of the topography on the fungal conidia adhesion in a protecting coating system.

Manuscript III

Jonas Stenbæk, Bo Jensen and Leise Riber (2015 in prep): Identification and classification of hydrophobin encoding gene sequences in two species of the blue stain fungus Aureobasidium pullulans.

Manuscript IV

Jonas Stenbæk, Leise Riber, Jakob Blæsbjerg Nielsen, Charlotte Møller Hansen and Bo Jensen (2015): The role of hydrophobins in surface growth by the Blue Stain fungi Aureobasidium pullulans. The International Research Group on Wood Protection, IRG/WP 15-10837 (ISSN 2000-8953).

Additional work Manuscript V

Samuel Jacquiod, Jonas Stenbæk, Susana S. Santos, Anne Winding, Søren J. Sørensen, Anders Priemé (ready for submission August 2015 in Molecular Ecology): Metagenomics provide valuable information on soil micro-eukaryotes.

5

roceedings

:

en PRA

: Annual

Conference P PhD PhD day

Wood Wood Protection annual

: 3th month meeting in the Højteknologifond HTF

: Danish Microbiological Sociaty annual meeting.

DMS International Research Group on plan for conference and meeting proceedings during the PhD. : Time IRG meeting. Paint Research Association international conference. meeting of PhD students associated Department of Biology, University of Copenhagen. project (Innovation fond)

6 Conference Proceedings

International conferences:

8th International Woodcoatings Congress, PRA (30 - 31 October 2012), Amsterdam, Holland. 10 minutes of presentation on a workshop.

IRG44, 44th Annual meeting, International Research Group on Wood Protection. (16 - 20 June 2013), Stockholm, Sweden. Two posters.

IRG45, 45th Annual meeting, International Research Group on Wood Protection. (11 – 15 May 2014), St. George, USA.

IRG46, 46th Annual meeting, International Research Group on Wood Protection. (10 – 14 May 2015), Vinã del Mar, Chile. My conference paper was selected to be present i Scientific Main Session, 25 minutes of presentation (picture).

My presentation at the 46th annual meeting of The International Research Group on Wood Protection. Viña del Mar, Chile, 2015. My paper was selected to be presented in the scientific main session at the conference.

Conferences and meetings in Denmark:

Annual meeting 2012, Danish Microbial Society. Copenhagen, Denmark. Poster.

PhD Day 2012. The PhD association, Department of Biology, University of Copenhagen.

7 Conference Proceedings

Annual meeting 2013, Danish Microbial Society. Copenhagen, Denmark. Poster.

PhD Day 2013. The PhD association, Department of Biology, University of Copenhagen. Poster.

PhD Day 2014. The PhD association, Department of Biology, University of Copenhagen. Poster.

Meetings in the Project; “Superior Bio Based Coating System for Exterior Wood Applications”. Innovation Fund Denmark:

Every quarter, 12 times through the PhD project. One meeting is about 1½-2 days. Personal scientific presentation (10-20 minutes) and discussion (30-45 minutes) at every meeting.

8 THE FACULTY OF SCIENCE, UNIVERSITY OF COPENHAGEN PhD School of Science

PhD course portfolio

1.General information

Name of the PhD-student: Jonas Stenbæk

Civil reg. No. (CPR.nr.): 270978 1327 E-mail: [email protected]

Department at Science: Dept. of Biology

Start date of PhD: 01-05-12

Name of principal supervisor: Bo Jensen E-mail: [email protected] X 5+3 Scheme 4+4 Scheme PhD

2. Planned and passed PhD courses

COURSE TYPES 1. Complementary skills courses (max. 10 ECTS credits can be approved) e.g. Introduction to University Pedagogy, scientific writing, dissemination of knowledge, project management. 2. International/specialist courses (min. 10 ECTS credits can be approved) 3. Advanced master’s courses (max. 10 ECTS credits can be approved) 4. In addition, a maximum of 7½ ECTS points may be awarded for participation in journal clubs, self-study etc.

Name of the course Type Organised by Status ECTS (1-4) (institution): (Preapproved or passed) Responsible Conduct of Research 1 UCPH Passed 1

Advanced Methods in Microscopy Module C: 2 UCPH Passed 3 Confocal Laser Scanning Methods Systems Mycology 3 DTU Passed 5

Bioinformatics for Microbiology 2 UCPH Passed 8.5

PhD Day 2014 at Department of Biology 2 UCPH Passed 2

Building Research through Dissemination 2 UCPH Passed 2

Save Time and Improve Your Research 1 UCPH Passed 1

Course in Paint and Coating Systems 1 PPG Industries/ Passed 5 Dyrup A/S 44th IRG Conference 4 International Research Passed 1.5 Group on Wood Protection, Stockholm 45th IRG Conference 4 International Research Passed 1.5 Group on Wood Protection, Utah, USA 46th IRG Conference 4 International Research Passed 1.5 Group on Wood Protection, Chile 32

PhD Progress Assessment Report 9 Teaching and assistant supervision

During the PhD project (2012-2015) the following teaching and supervision activities have been conducted:

Teaching (assistant teacher):

2012: Almen Mikrobiology Bachelor (undergrad.) level

2012, 2013 and 2014: Microbial Biotechnology Master (grad.) level

Assistant supervision:

7 x Bachelor (undergrad.) projects at Section for Microbiology, University of Copenhagen

1 x Bachelor (undergrad.) projects at Center of Renewable Carbon, Uni. of Tennessee

1 x Master (grad.) project at Section for Microbiology, University of Copenhagen

10 Scientific environments

Overview of change of scientific environment through the tree PhD years

11 Abbreviations

aur1 Hydrophobin 1 gene in A. pullulans MUCL 38722 aur2 Hydrophobin 2 gene in A. pullulans MUCL 38722

BPD Biocidal Products Directive

BPR Biocidal Products Regulation

C12TAB Dodecyltrimethylammonium bromide

CLP Classification, Labelling and Packaging regulation

DCOIT 4,5-Dichloro-2-octyl-4-isothiazolin-3-one

EPA Environmental Protections Agency eSEM Environmental Scanning Electron Microscopy

EU The European Union

GRAS Generally Regarded As Safe

HC Holliday Complete Media

HCS Holliday Complete Media with 1M Sorbitol

HFB Hydrophobin

HPLC High Performance Liquid Chromatography hfbA Hydrophobin A gene in A. pullulans P268 hfbB Hydrophobin B gene in A. pullulans P268

IPBC 3-Iodo-2-propynyl butylcarbamate

IRGWP International Research Group on Wood Protection

Milli-Q Milli Q water

PDA Potato Dextrose Agar

12 Abbreviations

PDB Potato Dextrose Broth

POPC 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine

POPS 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine

PVC Polyvinylchlorid

QCM-D Quartz Crystal Microbalance with Dissipation Monitoring

R43 EU Label: "May cause sensitization by skin contact"

RPM Rounds per Minutes

SDS Sodium Dodecyl Sulfate

SEM Scanning Electron Microscopy

SLB Supported Lipid Bilayers

13

14 Chapter 1: Background and Aims

1.1 Superior Bio Based Coating System for Exterior Wood Applications

Coatings, in this PhD thesis used as a generic name for all types of paints, protective coatings, lacquers etc., are an essential part of the majority of products covering solid wood or wood-based materials. Coatings can provide wooden materials with the desired aesthetical properties like color and gloss, but are mostly also of vital importance in the protection of wood against environmental influences like moisture, radiation, biological deterioration or damage from mechanical or chemical origin. This applies to both interior and exterior applications like wooden joinery, claddings or fences. Because the interactions between coating, substrate and external factors are quite different for interior and exterior uses, this project is in principal limited to exterior applications. More than 95% of exterior wood coatings are applied as liquid coatings with either organic solvent or water as the carrier for the other coatings ingredients (de Meijer, 2001; Marrion, 2004).

Today there is an immediate need for a new generation of innovative and highly durable, protective coatings and preservative systems for wood, 100% derived from sustainable and non-fossil bio-derived resources and free of toxic biocides. Wood is versatile and can satisfy coming demands for sustainable construction components. However, protective treatments are needed to prevent microbial degradation of wood exposed outdoors. Present coatings are fossil based and contain toxic biocides that are regulated (Biocidal Products Directive) and cannot be described as “sustainable”. This needs to be addressed, as increased attention is paid to remove dependence on fossil resources for industrial products, and to develop sustainable environmentally friendly alternatives (McCormick and Kautto, 2013).

Therefore, the overall aim of this study is to develop a new kind of bio-sustainable strategy for evolving a non-toxic biocide for a 100% bio sustainable coating system.

This PhD study is a part of a bigger project: “Superior Bio Based Coating System for Exterior Wood Applications” funded both by the Danish Innovation Fund (formerly The Danish National Advanced Technology Foundation) and from partners in the project. It is about substitution of every fossil-based component in a traditional standard coating for exterior wood. There are seven partners who worked together and by themselves on different objectives within the different coating components. The project groups are of high quality and comprise three Danish industrial companies; Dyrup A/S (PPG

15 Chapter 1: Background and Aims

Industries), leaders in protective wood stains who overall managed and coordinated the project; Palsgaard, a supplier of renewable surfactants and Emmelev, a supplier of natural oils. Furthermore the project involves research groups from The Danish Technological Institute (Wood technology), DTU - The Technical University of Denmark (Department of Chemical and Biochemical Engineering with DPC) and two research groups from University of Copenhagen (Section for Forest, Nature and Biomass and Section of Microbiology).

The synergies between, and expertise of, these different partners have ensured an interesting project outcome and key parts of the project are still ongoing. The industry partners were R&D active and comprised a strong supply chain to the commercial market, while research groups underpin the project fundamentals. The projects aim is to create a leading and lasting, high caliber network and expertise base which develops the renewable, durable protective wood coating systems, giving Denmark a cutting edge lead in a highly competitive marked.

The overall aims and objectives for the project Superior Bio Based Coating System for Exterior Wood Applications is to:

 Replace fossil derived additives and materials in water-based wood stains and preservatives, without compromising durability and high performance  Remove harmful biocides, by improving overall paint and wood protective performance.  Build fundamental understanding of, and utilize, the key properties of the wood and new applied protective coating systems, from a “bottom-up” standpoint  Create a strong Danish network for R&D and innovation in renewable coatings

16 Chapter 1: Background and Aims

1.2 The overall aim: Control of Aureobasidium pullulans in bio-sustainable coating systems for exterior wood.

The aim of my PhD study was to investigate alternative and bio-sustainable solutions, strategies and potentials within control of fungal surface attacks on processed wood and coating systems for exterior wood. The aim were narrowed and focused on primary the attacks of the ascomycete fungi, like Aureobasidium pullulans, that lead to blue stain discoloration on the surface of wood or wood coatings. Five main - and one supplementary - approaches were emerged and further reviews and research were performed. All together it ended up in this PhD thesis.

Figure 1.1: The aims of this PhD study. All aims contribute to the overall goal: Control of the blue stain fungi Aureobasidium pullulans in a bio-sustainable protective coating.

17 Chapter 1: Background and Aims

1.2.1 Time line

Figure 1.2 Time line of the three years of scientific work – categorized in the main aims.

1.2.2 Structure of this PhD thesis

This PhD thesis is built on four main manuscripts (plus one supplementary manuscript) and some additional scientific work presented as scientific reports.

This background chapter, Chapter 1, is followed by a broad introductory chapter, Chapter 2, where the fundamental and basic scientific terms and concepts are introduced, besides the basic methods and materials that are relevant throughout the following chapters of the thesis.

Chapter 3 represents a study where fatty acid-based emulsifiers are tested and evaluated for anti-fungal properties in a coating system and the modes of action are examined. This chapter includes Manuscript I. Chapter 4 works with a range of different components as anti-fungal candidates, like enzymes and nanoclays. Chapter 5 contains a very interesting study in the effect of the micro topography structure on fungal attachment on coating surfaces and the chapter contains Manuscript II.

Chapter 6 gives a broad introduction to an important fungal interface protein, the hydrophobin. The chapter contains multiple studies in bioinformatics, Manuscript III, and molecular work, Manuscript IV. Not all work has ended up in manuscripts yet so the chapter also contains a report and review on important non-published laboratory work and preliminary results. The last chapter, Chapter 7, is a additional chapter that contains a concurrent scientific work on metagenomics on soil fungi, Additinal Manuscript V.

18 Chapter 1: Background and Aims

1.3 References Chapter 1 de Meijer, M., 2001. Review on the durability of exterior wood coatings with reduced VOC-content. Prog. Org. Coat. 43, 217–225. Marrion, A., 2004. The chemistry and physics of coatings. Royal Society of Chemistry. McCormick, K., Kautto, N., 2013. The bioeconomy in Europe: An overview. Sustainability 5, 2589–2608.

19

20 2. Introduction

2.1 Blue Stain and protective coatings

Growth of mould fungi with dark-colored hyphae and spores (blue stain fungi) is a common phenomenon, both on coated and uncoated wooden (Bardage, 1997, Bussjaeger et al., 1999; Gobakken and Westin, 2008; Schmidt, 2006). Blue stain is the discoloration caused by a mould attack that colonizes the sapwood, and/or the surface coating, and produces highly colored structures on a surface (figure 2.1). The discoloration is caused by melanin in the fungal hyphae, which gives rise to a dark brown, grey, blue or black color, resulting from a refraction of light in a combination with the substrate (Bardage, 1997; Schmidt, 2006; Zink and Fengel, 1990). Although the blue stain, within the wood, can visually appear in different color tones, the stain seems to be brown when viewed in the light microscope (Stirling and Morris, 2009). The discoloration on the wood surface mainly effects the aesthetic expression and thereby reduces the value of the wood in appearance applications where the natural wood or paint color is desired. However, the blue stain fungi can actually penetrate into the wood structure via the medullary rays (figure 2.2) and colonize the tracheids and fibers in the wood (Schmidt, 2006; Ward, 1897).

21 Chapter 2: Introduction

Blue stain is caused by a range of about 100 to 250 fungi belonging to the Ascomycetes and Deuteromycetes (Kaarik and others, 1980). Some studies suggests a division of the blue stain fungi into three main groups: 1) Ceratocystis, Ophiostoma and Ceratocystiopsis species, – 2) black yeasts such as Hormonema dematioides, Aureobasidium pullulans, Rhinocladiella Atrovirens, and Phialophora species, – 3) dark moulds such as Alternaria alternata, Cladosporium species (Schmidt, 2006).

To protect the wood against blue stain, it is necessary to reduce the water uptake in the wood and use a protective coating for exterior wooden claddings. Beyond the discoloration of blue stain on exterior wood, other external factors have to be taken into account, like physical degradation caused by UV-radiation, temperature, rain, condensation, wind, and high relative humidity leading to other kind of fungal attacks, like decay (de Meijer, 2001; Feist and Hon, 1984; Gobakken and Vestøl, 2012; Williams et al., 2000). The long-time performance of coated wooden claddings depends on the type and quality of the coating

22 Chapter 2: Introduction

products, as well as the wood surface properties, building design and the environmental impacts (Gobakken et al., 2010, 2008; Gobakken and Lebow, 2010; Jermer, 2011).

There is a broad range of different kinds of protective coatings for exterior wood, and they have very different kinds of effects depending on the type of wood substrate for which they are applied. The range of factors regarding the wood is enormous, like variation in the wood species, fiber density, growth ring width, mixture of juvenile wood and mature wood, sapwood and heartwood, angle of grains, and number, distribution, and size of knots etc. (Gobakken and Vestøl, 2012; Sivertsen and Flæte, 2012). A protective coating will not change the equilibrium moister content of the wood, but it will affect the rate at which absorption will occur (Williams et al., 2000). Therefore, a good protective coating system for exterior wood will reduce the water transport significantly (Gobakken and Vestøl, 2012).

Studies indicate that effective protection against microbial colonization of the exterior wood or the topcoat is only possible with the use of biocides (Van Acker et al., 1998; Volkmer et al., 2010; Volkmer and Schwarze, 2008). New environmental legislation and general opinions in North America and the European Union are driving the motivation towards water-based, or even bio-sustainable, coatings and thereby away from solvent- based products. Although solvent-based preservatives penetrate wood more easily, water- based acrylic and hybrid systems are widely used and well-established wood protection products. In addition to azoles, zinc pyritione and octyl isothiazolinone, 3-iodo-2-propynyl butylcarbamate, (IPBC) is a commonly used biocide in water-based coating systems (Volkmer et al., 2010; Volkmer and Schwarze, 2008). IPBC penetrates wood slightly (<250 µm) and non-homogeneously. Studies have also revealed that the total amount of IPBC in the paint is reduced by approximately 70% by leaching following 3 months of exposure to natural weathering (Volkmer and Schwarze, 2008).

2.2 IPBC

3-Iodo-2-propynyl butylcarbamate (IPBC) is a water-based preservative agent originally used in the wood and by paint industries (Künniger et al., 2014a). Use is restricted in some countries due to its toxicity, especially acute inhalation toxicity. IPBC is also becoming recognized as a contact allergen (Badreshia and Marks Jr, 2002).

IPBC is an effective fungicide at low concentrations in products like protective coatings. IPBC was approved in 1996 for use up to 0.1% concentrations in products within coatings and cosmetics. However, this preservative is mostly found in cosmetics at about one-eighth that level (Maier et al., 2009).

23 Chapter 2: Introduction

In wood protective coatings, mainly biocides such as IPBC, propiconazole, tebuconazole, tolylfluanid or dichlofluanid are used (Künniger et al., 2014a). Nevertheless, limited information is available about the release of these biocides from wood coatings during service life, and only little knowledge exists about their fate and behavior in the environment (Künniger et al., 2014a, 2014b). Previous studies have shown that especially IPBC was released very easily from water-soluble coating systems on wood (Bjurman, 1995; Volkmer and Schwarze, 2008). Biocides such as diuron, carbendazim and irgarol were detected in treated waste water and sewage sludge as well as in natural waters in Switzerland (Burkhardt et al., 2009).

The coating industry is using the IPBC in a commercial encapsulated form. The encapsulation initiates a kind of slow release resulting in a reduced release rate of the biocide in the coating and thereby extends their useful life up to 5-10 years. In lots of commercial top-coatings for exterior wood IPBC is the main, or only, fungicide. A mean pure IPBC concentration in a typical wood paint is about 0.3% IPBC due to information from the industry.

There is a continuous discussion in the EU on the potential impact on humans, animals and the environment of various biocides. Opinions are divided among the different Environmental Protections Agencies (EPAs) in the EU about what is considered to be an acceptable health and environmental risk within biocides - in particular about the requirements that may be made to protect the people who handle biocidal products. The discussion is primarily aimed at private consumers. Is it possible or realistic to impose private consumers to wear protective gloves when using these biocidal products? Specifically, there have been discussions about active substances labeled with R43 "May cause sensitization by skin contact". The discussion has been primarily directed to the active substance 4,5-Dichloro-2-octyl-4-isothiazolin-3-one (DCOIT) that has been restricted in some EU member countries because of R43. In the same context, IPBC is also mentioned, and it is also labeled R43 (Directive 98/8/EC, 1998; IA to Directive 98/8/EC, 2008; Montforts and de Knecht, 2002; PA&MRFG-Feb13-Doc.4.5, 2013). Because of the

24 Chapter 2: Introduction

label R43, the biocides have to be CLP labeled on the commercial coating products. CLP, Classification, Labelling and Packaging (also associated with the H317 regulation) represent the EU regulation on labeling commercial products. It is expected to facilitate global trade and the harmonized communication of hazardous information on chemicals, and to promote regulatory efficiency (Fanghella and Catone, 2011; IA to Directive 98/8/EC, 2008; L353, 31.12.2008).

The issue is basically that all active substances, which are effective against moulds in protective coatings, are labeled with R43 according to the CLP regulation. IPBC and DCOIT will presumably not be banned in the very near future, but it could be expected that EU may introduce restrictions in relation to how these active substances work. Therefore, the industry is forced to respond strategically - in the first instance mainly for the private consumers market. It can be difficult to predict future interventions and regulations, but a good guess might be that new restrictions and regulations will be introduced to the industry whenever a new usable, environmentally friendly and bio- sustainable alternative is presented to the market.

2.3 Aureobasidium pullulans

In this PhD study, Aureobasidium pullulans has been chosen to be the main test organism. However, in some tests, assistance of other Ascomycetes; like Aspergillus and Penicillium species has been implemented due to their slightly faster colonization and growth rate.

25 Chapter 2: Introduction

Aureobasidium pullulans is a black fungus belonging to the Ascomycota division (info box 1). It is a ubiquitous mould, or mould-like, fungus that can be found in many different environments like air, soil and water. It is a naturally occurring epiphyte or endophyte of a wide range of plant species, and it is known for the ability to colonize a wide range of these hosts (Andrews et al., 1994; Webb and Mundt, 1978). Aureobasidium pullulans has excellent properties to survive under various environmental conditions (Köhl and Fokkema, 1993). From older available records (Domsch et al., 1980), it is apparently most common in the temperate zones of the world with numerous records from all over Europe, North and South America, Asia, Mediterranean and Africa (Deshpande et al., 1992; Domsch et al., 1980). However, so far it has only rarely been reported in cold environments. This could be caused by the fact that most investigations on the occurrence and diversity of fungi in the cold areas have been limited to frozen Antarctic soils and Siberian permafrost, where basidiomycetous fungi dominates (Abyzov, 1993; Babjeva and Reshetova, 1998; Deegenaars and Watson, 1998; Golubev, 1998; Ma et al., 2005; Vishniac, 2006; Zalar et al., 2008). Nevertheless, studies from the mid 00’s in Svalbard indicate that Aureobasidium-like fungi were found among the dominant ascomycetes (Butinar et al., 2007; Sonjak et al., 2006; Zalar et al., 2008).

Aureobasidium pullulans colonizes anthropogenic and processed surfaces like protective coatings and paints (Bardage et al., 2014; Bardage, 1997; Zabel and Terracina, 1978) and latex surfaces (Zabel and Horner, 1981).

26 Chapter 2: Introduction

27 Chapter 2: Introduction

2.3.1 Morphology and life cyclus

Aureobasidum pullulans was taxonomically (re-)characterized in 1994 (De Hoog and Yurlova, 1994) on the basis of morphology and nutritional physiology. The fungus has both a perfect and imperfect stage. The perfect stage as well as the imperfect stage has a very diverse morphology. The fungus shows polymorphism wherein it can grow as budding yeast or as a mycelium dependent on environmental conditions (Dominguez et al., 1978; Gadd and Mowll, 1985; Ramos and Acha, 1975).

The colonies initially are yellow, cream, light pink or light brown, and after 3-4 days of growth on PDA media they turn black and smooth (figure 2.4) and become covered with slime (Gadd and Mowll, 1985; Gaur et al., 2010). The colonies become blackish due to the chlamydospore production as well as to the melanin production in the hyphae. The hyphae are hyaline smooth and thin-walled (2-16 µm wide) forming a tough and compact mycelium. Aureobasidium pullulans can be recognized by straight asexual spores, conidia, and by the presence of lobed chains of thick-walled chlamydospores (Deshpande et al., 1992; Ramos and Acha, 1975; Yurlova et al., 1999).

28 Chapter 2: Introduction

The life cycle of Aureobasidium pullulans is complex (Kockova-Kratochvilova et al., 1980) (figure 2.5). The imperfect asexual reproduction is most common, whereas perfect sexual reproductions are rarely observed, even in controlled laboratory conditions.

2.3.2 Pullulan

Aureobasidium pullulans produce the extracellular alpha-glucan, pullulan, which is a linear α-D-glucan made mainly of maltotriose repeating units interconnected by α (1→6) linkages, it is a water-soluble homopolysaccharide produced extracellularly by the fungi (Lazaridou et al., 2002; Sutherland, 1998; Wallensfels et al., 1965). The rate of polysaccharide elaboration varies considerably at different points during the growth cycle, the maximum rate appearing to coincide with the production of blastospores (Catley, 1980, 1973). However, other studies indicate that pullulan rather seems to be produced only in swollen cells and chlamydospores (Campbell et al., 2004), but this seems questionable due to studies in pullulans from isolates blastospores (Bardage and Bjurman, 1998; Reeslev et al., 1991).

The pullulan production seems to have an important role in cell protection and adhesion and attachment of the conidia on solid surfaces, like leaf surfaces, wood or wood coatings (Andrews et al., 1994; Bardage et al., 1997; Bardage and Bjurman, 1998; Leathers, 2003), and studies show that pullulan together with uronic acid based polymers is essential when for the conidia to attach to a glass surface (Pouliot et al., 2005). Pullulanase, an enzyme specific for the hydrolysis of α-(1,6) glucosidic bonds in branched amylaceous polysaccharides and derived oligosaccharides (Wallenfels et al., 1961), can completely

29 Chapter 2: Introduction

disconnect the conidial adhesion (Andrews et al., 1994; Pouliot et al., 2005). Although fungal cells often produce polysaccharide for the purpose of energy storage, this does not appear to be the function of pullulan in Aureobasidium pullulans, since the microbe cannot break the polysaccharide down to easily metabolized sugars (Leathers, 2003; Pouliot et al., 2005). The presence of pullulan also seems to facilitate the diffusion of molecules both into and out of the cell (Rehm, 2009).

2.4 Methods and materials

In this section, shared methods and materials in the following chapters (3-6) of this PhD thesis are presented. Specific methods and materials for a single aim or topic are presented in that specific chapter.

2.4.1 Strains

In this PhD study the Aureobasidium pullulans (de Bary) Arnaud, strain P268 (DSM-3497) was selected to be the main test organism. This organism is widely used in European Standard tests (e.g. EN 927-6:2006) of fungal stain attacks, such as blue stain, on exterior wood (Cooke, 1959).

Other Aureobasidium pullulans strains in this PhD study: Aureobasidium pullulans strain MUCL38722.

2.4.2 Full-genome sequences

In this study, the full genome sequence of Aureobasidium pullulans strain P268 was obtained by high-throughput sequencing on the Illumina MiSeq platform at the National High-throughput DNA Sequencing Centre, University of Copenhagen (web site: http://seqcenter.ku.dk/), and sequence reads were annotated with the prediction of open reading frames in CLC Genomic Workbench (version 6.5.1).

Aureobasidium pullulans strain MUCL38722 was previously genome sequenced and annotated to predicted proteins (nucleotides and amino acids, respectively). The genomic sequence data were produced by the Genozymes for Bioproducts and Bioprocesses Development Project (website: http://www.fungalgenomics.ca).

30 Chapter 2: Introduction

2.4.3 Growth tests

A range of solid and liquid growth tests has been performed in the following chapters (chapter 3 and 4). As a standard method in this study, the test organism Aureobasidium pullulans strain P268 was used.

Spore suspension (starter culture) was made for inoculation by 24 hours growth in Potato Dextrose Broth (PDB) at 25 °C, 140 rpm.

For liquid growth tests, 1 ml of the spore suspension was inoculated to 300 ml Erlenmeyer flasks containing 100 ml PDA. The liquid growth tests were set to 3, 5 or 7 days. Harvest was performed by filtration of the media through a triple layer cheese cloth filter. Following 24 hours of incubation at 80 °C of the filter with the fungi, the biomass of the fungi was measured taking the mass of the filter into account.

Solid growth tests were performed on Potato Dextrose Agar (PDA) media in standard Petri dishes (90 mm). 4 µl of the spore suspension was carefully inoculated in a precise circular droplet on three spots on the agar (figure 2.7).

Figure 2.7. Petri dish with three inoculations.

The inoculated petri dishes were incubated for 5 or 7 days at 25 °C. Measurements of the diameter of every colony were performed with a standard ruler. The horizontal and vertical diameters of the colony were measured, and a mean value was calculated. If the deviation

31 Chapter 2: Introduction

between the horizontal and vertical value for a single colony was more than 20%, the colony was discard for further studies due to a presumable failure in the inoculation procedure. A standard growth rate of Aureobasidium pullulans strain P268 can be seen at figure 2.8.

Figure 2.8. Growth rate of Aureobasidium pullulans strain P268 at 25 °C on PDA media (n=28). Bars indicate standard deviations.

All additives in the growth tests in chapter 3 and 4 were added to the PDA and PDB after autoclaving and when the media were cooled down to approximately 50 °C.

2.4.4 Wood panel tests in Environmental Chamber

Wood panel tests were performed in an environmental test chamber made with inspiration from Bardage et al. 2014 (Bardage et al., 2014) but with minor modifications; the soil were Pindstrup Potting Soil (pH 6) and the fungal strains were Aspergillus versicolor (IMI 45554), Penicillium purpurogenum (IMI 178519) and Aureobasidium pullulans (IMI 45533). A chamber was made on request, specific for this PhD study, by Innovations IVS (figure 2.9).

32 Chapter 2: Introduction

The chamber was made of clear PVC. The top roof was detachable and secured with sealing strips. In the bottom are four soil boxes (figure 2.9 A1) 3D printed in PLA plastic. The boxes were placed on legs approximately 5 cm above the PVC bottom and over the 1000W heating element (figure 2.9 A2) that were covered by sterile water (water to a level approximately as shown in figure 2.9 A3 – over the heater but under the bottom of the boxes). The water was heated to 28 °C. The bottom of the soil boxes consists of a thin plastic net (less than 1mm net meshes). The soil was placed up to the edges of the boxes, and the net in the bottom prevents the soil to leak into the water below. At the same time, the net allows the heated evaporating water to flow up through the soil and up to the top roof of the chamber where it condensates on the PVC. The triangle shape of the top roof led the condensated water back to the bottom via the PVC walls of the chamber, without dripping inside the chamber. The water circulation makes a constant and equal air flow inside the chamber, which forces possible fungal spores and conidia to spread homogenously and make them ubiquitous. The fans showed on the figure were not used.

The three fungal strains were grown separately in 100 ml PDB media for 24 hours and then poured evenly on the soil in the four boxes in the chamber. The chamber was then incubated for two weeks. Following incubation, an explicit white/grey fungal growth was observed on the soil as a mouldy sheet. Test agar plates (PDA media) inside the chamber revealed a majority of Aspergillus versicolor and Penicillium purpurogenum.

The wood panels in the chamber test were the same as used in the EN152 European standard test (EN 152, 2011). The wood panels were coated with a natural based alkyd type coating formulation enriched with a test additive. Such alkyd formulation is a standard used for comparisons lacking any biocide. The coating was applied to the panels in two steps (24 hours interval) to a final amount of 8 m2/L. The panels were dried 2 days before being exposed in the environmental chamber for 5 weeks by hanging in the brackets in the chamber.

All wood panels were prepared in minimum three samples and were placed randomly in the chamber so that no identical specimens were on the same bracket. Evaluation of the panel tests were made after 4 weeks by visual examination of the specimens, and a growth coverage evaluation grade was awarded (Bardage et al., 2014):

0: No visible growth (verified by microscope examination) 1: visible < 10% coverage (verified by microscope examination) 2: 10-30% coverage (verified by microscope examination) 3: 30-60% coverage 4: more than 60% coverage 5: 100% coverage

33 Chapter 2: Introduction

A3

A1

A2

Figure 2.9. Environmental growth chamber

34 Chapter 2: Introduction

2.4.5 Protoplasts

For molecular work, protoplasts of Aureobasidium pullulans strain P268 and strain MUCL38722 along with some Aspergillus nidulans strains were prepared based on a previous work (Slightom et al., 2009) but with modifications.

50 ml PDB medium in Erlenmeyer flasks was inoculated with 0.5 ml of an outgrown Aureobasidium pullulans or Aspergillus nidulans culture and grown for 18 to 24 hours at 25 oC at 140 rpm. The content of each flask was centrifuged 4000 rpm for 5 minutes at 19 oC. The pellet was resuspended and washed in 20 ml STC medium (1 M Sorbitol, 25 mM o CaCl2, 50 mM Tris-HCl, pH 7.5) and centrifuged at 4000 rpm for 5 minutes at 19 C. The pellet was resuspended in 6 ml STC-medium and washed and centrifuged at 4000 rpm for 5 minutes at 19 oC. The pellet was resuspended in 6 ml STC medium and mixed with 10 ml STC-medium containing 400 mg Glucanex. The mix was finally incubated in a waterbath at 28 oC and 140 rpm and was checked by microscope examination. Following 2 hours of incubation, the amount of protoplasts (spheroplasts) seemed adequate, and the mixture was passed through 3 layers of sterile gaze. The protoplast suspensions were examined, and the number of cells was counted using a light microscope (figure 2.10).

Figure 2.10. Protoplasts of Aureobasidium pullulans strain P268 in a concentration of approximately 6x107 cells/mL.

35 Chapter 2: Introduction

2.5 Chapter 2 references

Abyzov, S.S., 1993. Microorganisms in the Antarctic ice. Antarctic microbiology 1, 265– 296. Andrews, J.H., Harris, R.F., Spear, R.N., Lau, G.W., Nordheim, E.V., 1994. Morphogenesis and adhesion of Aureobasidium pullulans. Canadian Journal of Microbiology 40, 6–17. Babjeva, I., Reshetova, I., 1998. Yeast resources in natural habitats at polar circle latitude. Food Technology and Biotechnology 36, 1–6. Badreshia, S., Marks Jr, J.G., 2002. Iodopropynyl butylcarbamate. Dermatitis 13, 77–79. Bardage, S.L., 1997. Colonization of painted wood by blue stain fungi. Doctoral thesis. Silvestria 49. Bardage, S.L., Bjurman, J., 1998. Isolation of an Aureobasidium pullulans polysaccharide that promotes adhesion of blastospores to water-borne paints. Canadian Journal of Microbiology 44, 954–958. Bardage, S.L., Bjurman, J., Ruddick, J.N.R., 1997. Adhesion of blastospores of Sclerophoma pityophila to water-borne paints is mediated by soluble extracellular compounds released during early stages of germination. Material und Organismen 31:91-107. Bardage, S., Westin, M., Fogarty, H.A., Trey, S., 2014. The effect of natural product treatment of southern yellow pine on fungi causing blue stain and mold. International Biodeterioration & Biodegradation 86, 54–59. Bjurman, J., 1995. Leaching of components from water-borne paints and fungitoxic effects. Document-the International Research Group on Wood Preservation. Volume: no. 95-20062. Burkhardt, M., Junghans, M., Zuleeg, S., Boller, M., Schoknecht, U., Lamani, X., Bester, K., Vonbank, R., Simmler, H., 2009. Biozide in Gebäudefassaden– ökotoxikologische Effekte, Auswaschung und Belastungsabschätzung für Gewässer. Umweltwissenschaften und Schadstoff-Forschung 21, 36–47. Bussjaeger, S., Daisey, G., Simmons, R., Spindel, S., Williams, S., others, 1999. Mildew and mildew control for wood surfaces. Journal of Coatings Technology 71, 67–69. Butinar, L., Spencer-Martins, I., Gunde-Cimerman, N., 2007. Yeasts in high Arctic glaciers: the discovery of a new habitat for eukaryotic microorganisms. Antonie van Leeuwenhoek 91, 277–289. Campbell, B.S., Siddique, A.-B.M., McDougall, B.M., Seviour, R.J., 2004. Which morphological forms of the fungus Aureobasidium pullulans are responsible for pullulan production? FEMS microbiology letters 232, 225–228.

36 Chapter 2: Introduction

Catley, B.J., 1980. The extracellular polysaccharide, pullulan, produced by Aureobasidium pullulans: a relationship between elaboration rate and morphology. Journal of General Microbiology 120, 265–268. Catley, B.J., 1973. The rate of elaboration of the extracellular polysaccharide, pullulan, during growth of Pullularia pullulans. Journal of general microbiology 78, 33–38. Cooke, W.B., 1959. An ecological life history of Aureobasidium pullulans (de Bary) Arnaud. Mycopathologia et Mycologia applicata 12, 1–45. Deegenaars, M.L., Watson, K., 1998. Heat shock response in psychrophilic and psychrotrophic yeast from Antarctica. Extremophiles 2, 41–50. De Hoog, G.S., Yurlova, N.A., 1994. Conidiogenesis, nutritional physiology and taxonomy of Aureobasidium and Hormonema. Antonie van Leeuwenhoek 65, 41– 54. de Meijer, M., 2001. Review on the durability of exterior wood coatings with reduced VOC-content. Progress in organic coatings 43, 217–225. Deshpande, M.S., Rale, V.B., Lynch, J.M., 1992. Aureobasidium pullulans in applied microbiology: a status report. Enzyme and Microbial Technology 14, 514–527. Directive 98/8/EC, 1998. Directive 98/8/EC of the European Parliament and of the Council of 16 February 1998 concerning the placing of biocidal products on the market (BPD). Dominguez, J.B., Goni, F.M., Uruburu, F., 1978. The transition from yeast-like to chlamydospore cells in Pullularia pullulans. Journal of General Microbiology 108, 111–117. Domsch, K.H., Gams, W., Anderson, T.-H., others, 1980. Compendium of soil fungi. Soc General Microbiol. EN 152, 2011. EN 152:2011, Wood preservatives - Determination of the protective effectiveness of a preservative treatment against blue stain in wood in service - Laboratory method. Fanghella, P.D.P., Catone, T., 2011. The CLP Regulation: origin, scope and evolution. Annali dell’Istituto superiore di sanità 47, 126–131. Feist, W.C., Hon, D.-S., 1984. Chemistry of weathering and protection. Advances in chemistry series 401–451. Gadd, G.M., Mowll, J.L., 1985. Copper uptake by yeast-like cells, hyphae, and chlamydospores of Aureobasidium pullulans. Experimental mycology 9, 0–40. Gaur, R., Singh, R., Gupta, M., Gaur, M.K., 2010. Aureobasidium pullulans, an economically important polymorphic yeast with special reference to pullulan. Afr J Biotechnol 9, 7989–7997.

37 Chapter 2: Introduction

Gobakken, L.R., Høibø, O.A., Solheim, H., 2010. Factors influencing surface mould growth on wooden claddings exposed outdoors. Wood Material Science and Engineering 5, 1–12. Gobakken, L.R., Lebow, P.K., 2010. Modelling mould growth on coated modified and unmodified wood substrates exposed outdoors. Wood science and technology 44, 315–333. Gobakken, L.R., Mattson, J., Alfredsen, G., 2008. In-service performance of wood depends upon the critical in-situ conditions. IRG/WP 08–20382. Gobakken, L.R., Vestøl, G.I., 2012. Surface mould and blue stain fungi on coated Norway spruce cladding. International Biodeterioration & Biodegradation 75, 181–186. Gobakken, L.R., Westin, M., 2008. Surface mould growth on five modified wood substrates coated with three different coating systems when exposed outdoors. International Biodeterioration & Biodegradation 62, 397–402. Golubev, W.I., 1998. New species of basidiomycetous yeasts, Rhodotorula creatinovora and R. yakutica, isolated from permafrost soils of Eastern-Siberian Arctic. Mykologiya i Phytopathologiya 32, 8–13. IA to Directive 98/8/EC, 2008. Directive 98/8/EC concerning the placing of biocidal products on the market. Assessment Report. IPBC, Product-type 8 (Wood preservatives). Jermer, J., 2011. Wood Exter-Service life and performance of exterior wood above ground. Final Report. SP Report. Kaarik, A., others, 1980. Fungi causing sap stain in wood. Rapport Institutionen for Virkeslara, Sveriges Lantbruksuniversitet. Kockova-Kratochvilova, A., Černáková, M., Slavikova, E., 1980. Morphological changes during the life cycle of Aureobasidium pullulans (de Bary) Arnaud. Folia microbiologica 25, 56–67. Köhl, J., Fokkema, N.J., 1993. Fungal interactions on living and necrotic leaves. Ecology of plant pathogens 321–334. Künniger, T., Gerecke, A.C., Ulrich, A., Huch, A., Vonbank, R., Heeb, M., Wichser, A., Haag, R., Kunz, P., Faller, M., 2014a. Release and environmental impact of silver nanoparticles and conventional organic biocides from coated wooden façades. Environmental Pollution 184, 464–471. Künniger, T., Heeb, M., Arnold, M., 2014b. Antimicrobial efficacy of silver nanoparticles in transparent wood coatings. European Journal of Wood and Wood Products 72, 285–288. L353, 31.12.2008, n.d. Regulation on classification, labelling and packaging of substances and mixtures, amending and repealing Directives 67/548/EEC and 1999/45/EC,

38 Chapter 2: Introduction

and amending Regulation (EC) No 1907/2006. Reference L353, 31.12.2008, pp. 1–1355. Lazaridou, A., Roukas, T., Biliaderis, C.G., Vaikousi, H., 2002. Characterization of pullulan produced from beet molasses by Aureobasidium pullulans in a stirred tank reactor under varying agitation. Enzyme and Microbial Technology 31, 122–132. Leathers, T.D., 2003. Biotechnological production and applications of pullulan. Applied Microbiology and Biotechnology 62, 468–473. Maier, L.E., Lampel, H.P., Bhutani, T., Jacob, S.E., 2009. Hand dermatitis: a focus on allergic contact dermatitis to biocides. Dermatologic clinics 27, 251–264. Ma, L.J., Catranis, C.M., Starmer, W.T., Rogers, S.O., 2005. The significance and implications of the discovery of filamentous fungi in glacial ice. Life in Ancient Ice, Princeton University Press, Princeton and Oxford. Montforts, M.H., de Knecht, J.A., 2002. European medicines and feed additives regulation are not in compliance with environmental legislation and policy. Toxicology letters 131, 125–136. PA & MRFG-Feb13-Doc.4.5, 2013. Note for discussion with competent authorities for biocidal products. Pouliot, J.M., Walton, I., Nolen-Parkhouse, M., Abu-Lail, L.I., Camesano, T.A., 2005. Adhesion of Aureobasidium pullulans Is Controlled by Uronic Acid Based Polymers and Pullulan. Biomacromolecules 6, 1122–1131. Ramos, S., Acha, I.G., 1975. A vegetative cycle of Pullularia pullulans. Transactions of the British Mycological Society 64, 129–IN9. Reeslev, M., Nielsen, J.C., Olsen, J., Jensen, B., Jacobsen, T., 1991. Effect of pH and the initial concentration of yeast extract on regulation of dimorphism and exopolysaccharide formation of Aureobasidium pullulans in batch culture. Mycological research 95, 220–226. Rehm, B., 2009. Microbial production of biopolymers and polymer precursors: applications and perspectives. Horizon Scientific Press. Schmidt, O., 2006. Wood and tree fungi. Springer. Sivertsen, M.S., Flæte, P.O., 2012. Water absorption in coated Norway spruce (Picea abies) cladding boards. European Journal of Wood and Wood Products 70, 307– 317. Slightom, J.L., Metzger, B.P., Luu, H.T., Elhammer, A.P., 2009. Cloning and molecular characterization of the gene encoding the aureobasidin A biosynthesis complex in Aureobasidium pullulans BP-1938. Gene 431, 67–79. Sonjak, S., Frisvad, J.C., Gunde-Cimerman, N., 2006. Penicillium mycobiota in Arctic subglacial ice. Microbial ecology 52, 207–216.

39 Chapter 2: Introduction

Stirling, R., Morris, P.I., 2009. Decolorization of blue stain in lodgepole pine sapwood by hypochlorite bleaching and light exposure. Forest products journal 59, 47. Sutherland, I.W., 1998. Novel and established applications of microbial polysaccharides. Trends in biotechnology 16, 41–46. Van Acker, J., Stevens, M., Brauwers, C., Rijckaert, V., Mol, E., 1998. Blue Stain Resistance of Exterior Wood Coatings as a Function of Their Typology. IRG/WP/98e20145. The International Research Group on Wood Preservation, Stockholm. Vishniac, H.S., 2006. Yeast biodiversity in the Antarctic, in: Biodiversity and Ecophysiology of Yeasts. Springer, pp. 419–440. Volkmer, T., Landmesser, H., Genoud, A., Schwarze, F.W., 2010. Penetration of 3-iodo-2- propynyl butylcarbamate (IPBC) in coniferous wood pre-treated with Physisporinus vitreus. Journal of coatings technology and research 7, 721–726. Volkmer, T., Schwarze, F.W., 2008. Diffusionsverhalten von IPBC in wasserbasierten Beschichtungssystemen auf Holzfassaden. Holz als Roh-und Werkstoff 66, 181– 189. Wallenfels, K., Bender, H., Keilich, G., Bechtler, G., 1961. On pullulan, the glucan of the slime coat of Pullularia pullulans. Angew Chem 73, 245–246. Wallensfels, K., Keilich, G., Bechtler, G., Freudenberger, D., 1965. Untersuchungen an Pullulan. IV. Die Klilrung des Strukturproblems mit physikalischen, chemischen, und enzymatischen Methoden. Biochemische Zeitschrifit 433–450. Ward, H.M., 1897. On the biology of Stereum hirsutum (Fr.). Philosophical Transactions of the Royal Society of London. Series B, Containing Papers of a Biological Character 123–134. Webb, T.A., Mundt, J.O., 1978. Molds on vegetables at the time of harvest. Applied and environmental microbiology 35, 655–658. Williams, R.S., Jourdain, C., Daisey, G.I., Springate, R.W., 2000. Wood properties affecting finish service life. Journal of Coatings Technology 72, 35–42. Yurlova, N.A., De Hoog, G.S., Van den Ende, A., 1999. Taxonomy of Aureobasidizum and allied genera. Studies in Mycology 63–69. Zabel, R.A., Horner, W.E., 1981. An accelrated laboratory procedure for growing Aureobasidium pullulans on fresh latex paint films. Journal of coatings technology 53, 33–37. Zabel, R.A., Terracina, F., 1978. Nutrition of saprobic fungi and control strategies for paint mildew caused by Aureobasidium pullulans. Journal of coatings technology 50, 43–47.

40 Chapter 2: Introduction

Zalar, P., Gostinčar, C., De Hoog, G.S., Uršič, V., Sudhadham, M., Gunde-Cimerman, N., 2008. Redefinition of Aureobasidium pullulans and its varieties. Studies in Mycology 61, 21–38. Zink, P., Fengel, D., 1990. Studies on the colouring matter of blue-stain fungi. Part 3. Spectroscopic studies on fungal and synthetic melanins. Holzforschung- International Journal of the Biology, Chemistry, Physics and Technology of Wood 44, 163–168.

41

42 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

3.1 Introduction

It must be considered as improbable to find a single environmental friendly synthetic or natural antimicrobial compound to inhibit the varied biological organisms capable of colonizing wood and wood products. Even a single compound to inhibit only the fungal attacks seems completely unrealistic. Nevertheless, this chapter contains studies of single alternative bio-sustainable compounds that could be candidates to new kind of anti-fungal agents in a protective coating system for exterior wood.

A range of compounds derived from - more or less - natural sources are already known and developed as general biocides, like caprylic acid (Charles et al., 1949), glycolic acid (Britton and Petrie, 1956), lactic and acetic acids (Adams and Hall, 1988), chitosan (El- Ghaouth et al., 2000), antioxidants and metal chelators with an organic biocide (Schultz and Nicholas, 2001), nisin and garlic/phenolic compounds (Adams et al., 2003), sorbates with vanillin or citral (Alzamora et al., 2003), borates and boron (Clausen and Yang, 2007; Lloyd, 1993; Taylor and Lloyd, 2006).

Also the anti-fungal properties and possibilities within saturated fatty acids have been subject for some studies. A conference paper from International Research Group on Wood Protection (IRGWP) evaluated the influence of middle-chain fatty acids on spore germination of basidiomycetes and determined that caprylic (C8), pelargonic (C9), and decanoic (C10) acids (0.01 wv%) destroyed the conidia of the test fungi (Schmidt, 1984). Schmidt also studied longer fatty acids (C12-C16) and concluded these compounds had different impact on different basidiomycetes, where dodecanoic acid (C12) was effective against the brown-rot basidiomycetes but not white-rot basidiomycetes. Also specific concentrations of the compounds had specific impact on different fungi Newer studies of the anti-fungal possibilities of fatty acid-based formulations against ascomycetes has been done and an interesting study by Clausen et al. from 2010 shows that small saturated monocarboxylic acids combined with selected adjuvants could effectively inhibit conidia from moulds. Specifically, formulations containing pentanoic (C5), caproic (C6), heptanoic (C7), octanoic (C8), nonanoic (C9), and/or capric (C10) saturated acid demonstrated efficacy against mould growth for up to 12 weeks in the ASTM D4445 standard laboratory test for mould (Clausen et al., 2010).

43 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

We already know that multiple non-fatty organic acids, like acetic to decanoic acid (C2- C10) and lactic, citric, and malic acids, are classified as generally recognized as safe (GRAS) compounds by the US Food and Drug Administration (FDA) and they are common used worldwide in the food industry as emulsifiers, acidulants and flavor enhancers. Due to their safety it will be common to create strategies and technologies based on these kinds of products for development of antimicrobials that are based on green chemistries. Multiple organic acids, like acetic, citric, and tartaric, are known to inhibit mould (Barbosa-Canovas et al., 1998; Chu et al., 2001; Stratford et al., 2009). The mode of action of these compounds is only indistinctly described in the literature. Chelation may play a beneficial role in biocidal function by altering availability of micronutrients or metallic cofactors necessary for the germination and the hyphal development (Bellion et al., 2006; Goodell et al., 1997). Other studies, including my own, also indicate some interactions with the fungal cell wall, see Manuscript I.

3.2 Candidates

In this PhD study a range of fatty acid-based emulsifiers has been researched for possible properties as anti-fungal candidates. All provided by Palsgaard A/S (table 3.1). The anti- fungal effect has been tested in solid and liquid media on a range of moulds, including Aureobasidium pullulans, Aspergillus niger and Penicillium chrysogenum. The effect has been evaluated and the best candidates selected to be tested in short term panel tests and proper standard outdoor exposures. The best of all candidates was tested further to clarify the mode of action by testing the fatty acid-emulsifier on fungal models made on nano technological instruments, like Quartz Crystal Microbalance with Dissipation (QCM-D).

Compound/candidate Provider Fatty Acid Abbreviation Monocaprin Palsgaard A/S C10 E09 Monolaurin Palsgaard A/S C12 E10 C12 Lactylate Palsgaard A/S C12 E11 Sodium Caproyl Lactylate Palsgaard A/S C10 E15 Lactic Acid Ester Palsgaard A/S C10 E17 Table 3.1: Compounds of fatty acid-based emulsifiers tested in this PhD study.

Among the compounds, the lactylates turned out to be the most efficient. And further studies were performed on these additives.

44 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

3.2.1 Lactylates

The lactylates are organic compounds (some of which are approved by the U.S. Food and Drug Administration, FDA) that find use as food additives and cosmetic ingredients as emulsifiers, but also as a range of other functionalities e.g. anti-foaming agents. These additives are environmentally friendly produced and they are non-toxic to humans (Lamb et al., 2010; Zobel, 1976) and biodegradable in the environment (Baniel et al., 1999; Dahms and Tagawa, 1997).

Due to the safety and versatile functionality of these lactylates, it is interesting to explore their ability as non-toxic bio-sustainable preservatives in – for example – protective coatings such as wood preservative systems. Specifically, lactylates are esters of lactic acid in which the C2-hydroxy group of lactic acid is esterified to a fatty acid. Lactic acids and other fatty acids, the components of the lactylates, are used as antimicrobial components. Lactic acid and various types of lactic acid bacteria (LAB) are used as antifungal agents (Schnürer and Magnusson, 2005) while lactic acid is known to inhibit Salmonella and Campylobacter contamination of broiler carcasses during industrial processing (Byrd et al., 2001). Moreover, fatty acid-based lactylates are used as antifungal agents against moulds and sapstains in wood protection systems (Clausen et al., 2010). Indeed, earlier studies indicate that, specifically, small chain (C2-C8) and middle chain (C9-C14) fatty acid-based emulsifiers have an inhibitory effect on specific fungal germination in asco- and basidiomycetes (Harman et al., 1980; Schmidt, 1984).

Lactylates can be used to provide stability due to their surface activity. They are surfactants and contain hydrophilic polar groups, which interact with water, and non-polar lipophilic groups, which interact with oils. Because of these interactions the lactylates can provide stability to an oil-water system. That results in the formation of an emulsion. Because of that lactylates are often referred to as emulsifiers. The degree of interaction is dependent on the fatty acid, the mole ratio of fatty acid to lactic acid, the degree of neutralization and the nature of the neutralizing base used in the construction and manufacturing of the lactylates (Bennett et al., 1968a, 1968b).

Sodium Capryol Lactylate is a lactylic ester of fatty acids. It is a versatile additive in food and cosmetics (Ash, 2004; Hasenhuettl, 2008). Sodium Capryol Lactylate is non-toxic to animals and biodegradable (Joint et al., 1991; Lamb et al., 2010), and just as important it can be manufactured in bio-sustainable-like ways such as using biorenwable feedstoks (Baniel et al., 1999; Markley, 1961). Sodium Capryol Lactylate are mixed fatty acid esters of lactic acid and its polymers, with minor quantities of free lactic acid, poly(lactic acid),

45 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

and fatty acids. They are dispersible in hot water and are soluble in organic solvents, such as vegetable oils (Baniel et al., 1999; Roby et al., 1998).

3.3 Mode of action

To investigate the mode of action of different compounds on the fungal cell membrane a very useful and interesting method has been selected for this study. A model cell membrane was constructed and exposed for the different compounds while analyzing the impact directly by detecting differences in the lipid layer by measuring the change in the frequency.

3.3.1 Supported lipid bilayers (SLB)

Supported lipid bilayers (SLB) are a model lipid bilayer assembled in vitro, as opposed to the bilayer of natural cell membranes. But unlike the cell membrane in which the lipid bilayer is rolled into an enclosed shell, the SLB are a planar lipid structure attached to a solid substrate (figure 3.1). Because of this structure only the upper lipid layer are exposed to selective solutions. This structure has advantages and drawbacks related to the study of lipid bilayers in cell membranes. Some of the advantages are that the structure makes the SLB very stable for long times and it will remain intact in physical challenged environments like high flow rates of solutions and vibrations. It is very easy to custom make at lipid composition corresponding to the needs.

The SLB is popular as useful and unique model systems for cell membranes (Sackmann, 1996; Salafsky et al., 1996; Watts et al., 1986). It is by now a progressively widespread and well regarded application in biological cellular investigations, diagnostic devices and biomimetics (Ashley et al., 2011; Kung et al., 2000; Picas et al., 2012; Richter et al., 2003). The application was initially developed in the 1980’s (Watts et al., 1984). It is attractive because of its simplicity and reproducibility. Considerable experimental and theoretical work has been devoted to understanding the driving forces involved in the adsorption of vesicles from solutions and in the mechanisms of SLB formation (Johnson et al., 2002; Lind et al., 2014; Reimhult and Höök, 2015; Richter et al., 2003; Zhdanov and Kasemo, 2001).

A SLB can be made with either synthetic or natural lipids. The simplest model systems contain only a single pure synthetic lipid but in this study a dual lipid solution were

46 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

selected based on a lipid study on Candida albicans (Löffler et al., 2000) which was considered as fairly close related to Aureobasidium pullulans. The selected SLB mixture composed of 75 mol% 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 25 mol% 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS), see manuscript I.

Figure 3.1. Dawning of a SLB. Phospholipids (blue and black) double layer formation on a solid substrate (grey) (“Chms.ucdavis.edu,” n.d.).

3.3.2 Quartz Crystal Microbalance with Dissipation (QCM-D)

For this study, one of the major advantages about the SLB is the opportunity to analyze the SLB by mechanical probing techniques which require a direct physical interaction with the sample. So a technique called Quartz Crystal Microbalance with Dissipation (QCM-D) was introduced to analyze the impact and the mode of action of bio-sustainable candidates on the fungal cell membrane by studying binding kinetics at the bilayer surface. The QCM- D is commonly used in the research of various biomaterials, cell adhesion, drug discovery, materials science, and bio- and nanophysics (Dixon, 2008).

QCM-D providing information about the mass and the conformational changes of adsorbed materials. QCM-D has been shown and proved to be a useful tool to study the formation of SLB (Keller and Kasemo, 1998; Reimhult et al., 2004, 2003). It is a common application used to determine a film thickness in a liquid environment (Dixon, 2008).

The principle in the QCM-D is active component, a thin quartz crystal disk sandwiched between a pair of electrodes working real-time (Keller and Kasemo, 1998; Steinem et al., 1996). An AC voltage over the electrodes causes the crystal to oscillate at its acoustic resonance frequency. When the voltage is turned off, the oscillation decays exponentially (Keller and Kasemo, 1998). The decay is detected and the resonance frequency (f) and the energy dissipation factor (D) are extracted by mutable measuring (approximately >200 data point pr. second). When the QCM-D measuring mass changes the instrument has a sensitivity of 0.5 ng/cm2 (Q-SENSE E4 system from Q-Sense, Västra Frölunda, Sweden).

47 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

The dissipation is defined as the loss of energy per oscillation period divided by the total energy stored in the system of crystal and attached compounds. The dissipation is equal to:

resonance bandwidth 퐷푖푠푠푖푝푎푡푖표푛 (퐷) = resonance frequency

The changes of the frequency (Δf) are primarily related to the mass uptake or the release at the crystal sensor surface - in this case the surface of the fungal model SLB. Changes in the dissipation factor (ΔD) are related to the softness of the system (Johannsmann, 2008). The softness is related to the structural changes of the SLB film adhering at the sensor surface (Dixon, 2008). A plot of QCM-D detection of a SLB can be seen in manuscript I, figure 6.

48 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

3.4 Preliminary results

All fatty acid-based emulsifiers (candidates) were tested for anti-fungal properties before the best candidate were analyzed in manuscript I.

Materials and methods: see section 2.4

Figure 3.3. Growth of Aureobasidium pullulans on PDA media containing the candidates (see section 3.2) in different concentrations. Yellow: control, black: 0.01%, grey: 0.03% and white: 0.05% (n=20). PDA containing 0.05% E15 and PDA containing 0.01, 0.03 and 0.05% concentrations of IPBC inhibit all kind of growth. 5 days growth, 25 ºC.

Figure 3.4. Growth rates of Aureobasidium pullulans on PDA media containing: red: no additives, black: 0.01% E15 and yellow: 0.05% E15. Lines are approximate representations of the growth rate (Polynomial lines, Order: 4). Growth conditions: 25 ºC (n=20).

49 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

Based on the preliminary results, E15 (Sodium Capryol Lactylate) showed the best anti- fungal results. Repeating tests on Aspergillus niger and penicillium chrysogenom revialed comparable outcomes indicating that E15 was the best candidate among the tested fatty acid-based emulsifiers. Additional test on the impact on the fungal biomass of the Aureobasidium pullulans colonies and test in liquid media (PDB) didn’t alter the overall tendencies. E15 was selected for further research, see Manuscript I.

3.5 Conclusions and perspectives for further research

Fatty acid-based emulsifiers have in this, and previous, studies revealed possible anti- fungal properties that could qualify them as candidates to future environmental friendly additives in coatings. However, not all fatty acid-based emulsifiers have the same useful properties and it seems that they also have different impact on different kinds of fungi. The Sodium Capryol Lactylate exposed promising properties in the laboratory test where the compound inhibited Aureobasidium pullulans in even small concentrations, but the wood panel tests did not show an effect comparable with the common used and environmental toxic IPBC. Further research could indicate if compounds close to Sodium Capryol Lactylate could have a better effect in the long-term tests. Another interesting issue could be to test the Sodium Capryol Lactylate as in-can preservation where other local environmental factors are important e.g. the humidity which is high in the can compared to the dry coating.

Furthermore, a range of Lactylate with different length of fatty acids could reveal a more beneficial length of the carbon chain as other studies indicate could have an effect.

The introduction of the QCM-D as a method to investigate the impact of compounds on the fungal cell membrane seems very promising and further studies could expand the usability to contain a range of different kinds of fungal species. An extensive study could also involve more compounds from the fungal cell membrane e.g. membrane proteins, making the QCM-D a perfect state-of-art technology for fungal cell membrane investigations.

50 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

3.6 References Chapter 3

Adams, M.R., Hall, C.J., 1988. Growth inhibition of food-borne pathogens by lactic and acetic acids and their mixtures. Int. J. Food Sci. Technol. 23, 287–292. Adams, M., Smid, E., Roller, S., others, 2003. Nisin in multifactorial food preservation. Nat. Antimicrob. Minimal Process. Foods 11–33. Alzamora, S.M., Guerrero, S., López-Malo, A., Palou, E., Roller, S., others, 2003. Plant antimicrobials combined with conventional preservatives for fruit products. Nat. Antimicrob. Minimal Process. Foods 235–249. Ashley, C.E., Carnes, E.C., Phillips, G.K., Padilla, D., Durfee, P.N., Brown, P.A., Hanna, T.N., Liu, J., Phillips, B., Carter, M.B., others, 2011. The targeted delivery of multicomponent cargos to cancer cells by nanoporous particle-supported lipid bilayers. Nat. Mater. 10, 389–397. Ash, M., 2004. Handbook of green chemicals. Synapse Info Resources. Baniel, A.M., Eyal, A.M., Mizrahi, J., Hazan, B., Fisher, R.R., Kolstad, J.J., Stewart, B.F., 1999. Lactic acid production, separation and/or recovery process. US 5892109 (A). Barbosa-Canovas, G., Pothakamury, U., Palou, E., Swanson, B., 1998. Chemicals and biochemicals used in food preservation. Nonthermal Preserv. Foods Marcel Dekker Inc N. Y. 215–233. Bellion, M., Courbot, M., Jacob, C., Blaudez, D., Chalot, M., 2006. Extracellular and cellular mechanisms sustaining metal tolerance in ectomycorrhizal fungi. FEMS Microbiol. Lett. 254, 173–181. Bennett, H., Bishop Jr., J., Wulfinghoff, M., 1968a. “Introduction”. Practical Emulsions Volume 1 Materials and Equipment. Chemical Publishing Company, New York. Bennett, H., Bishop Jr., J., Wulfinghoff, M., 1968b. “Ingredients and Additives”. Practical Emulsions Volume 1 Materials and Equipment. New York: Chemical Publishing Company. pp. 37–52. Chemical Publishing Company, New York. Britton, E.C., Petrie, P.S., 1956. Glycolic acid ethers of polyoxypropylene compounds and method of preparation. Google Patents. Byrd, J.A., Hargis, B.M., Caldwell, D.J., Bailey, R.H., Herron, K.L., McReynolds, J.L., Brewer, R.L., Anderson, R.C., Bischoff, K.M., Callaway, T.R., others, 2001. Effect of lactic acid administration in the drinking water during preslaughter feed withdrawal on Salmonella and Campylobacter contamination of broilers. Poult. Sci. 80, 278–283. Charles, H., Gaston, D., Russ, W.R., Schweitzer, T.R., 1949. Fungicide containing caprylic acid and its salt. Google Patents.

51 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

Chms.ucdavis.edu [WWW Document], URL: https://chms.ucdavis.edu/research/web/faller/surface_interactions.html (accessed 7.21.15). Chu, C.-L., Liu, W.-T., Zhou, T., 2001. Fumigation of sweet cherries with thymol and acetic acid to reduce postharvest brown rot and blue mold rot. Fruits 56, 123–130. Clausen, C.A., Coleman, R.D., Yang, V.W., 2010. Fatty acid-based formulations for wood protection against mold and sapstain. For. Prod. J. 60, 301–304. Clausen, C.A., Yang, V.W., 2007. Multi-component biocide protects wood from fungi and insects in UC2 applications. Proc Am Wood Prot Assoc 103, 31–35. Dahms, G.H., Tagawa, M., 1997. Also containing a lactylate and a lipophilic emulsifier of a fatty alcohol; cosmetics; sunscreen agents; oil in water emulsions; heat resistance; waterproofing; feel; biodegradable; uneffected by ph range. Google Patents. Dixon, M.C., 2008. Quartz Crystal Microbalance with Dissipation Monitoring: Enabling Real-Time Characterization of Biological Materials and Their Interactions. J. Biomol. Tech. JBT 19, 151–158. El-Ghaouth, A., Smilanick, J.L., Wilson, C.L., 2000. Enhancement of the performance of Candidasaitoana by the addition of glycolchitosan for the control of postharvest decay of apple and citrus fruit. Postharvest Biol. Technol. 19, 103–110. Goodell, B., Jellison, J., Liu, J., Daniel, G., Paszczynski, A., Fekete, F., Krishnamurthy, S., Jun, L., Xu, G., 1997. Low molecular weight chelators and phenolic compounds isolated from wood decay fungi and their role in the fungal biodegradation of wood. J. Biotechnol. 53, 133–162. Harman, G.E., Mattick, L.R., Nash, G., Nedrow, B.L., 1980. Stimulation of fungal spore germination and inhibition of sporulation in fungal vegetative thalli by fatty acids and their volatile peroxidation products. Can. J. Bot. 58, 1541–1547. Hasenhuettl, G.L., 2008. Overview of food emulsifiers, in: Food Emulsifiers and Their Applications. Springer, pp. 1–9. Johannsmann, D., 2008. Viscoelastic, mechanical, and dielectric measurements on complex samples with the quartz crystal microbalance. Phys. Chem. Chem. Phys. 10, 4516–4534. Johnson, J.M., Ha, T., Chu, S., Boxer, S.G., 2002. Early steps of supported bilayer formation probed by single vesicle fluorescence assays. Biophys. J. 83, 3371– 3379. doi:10.1016/S0006-3495(02)75337-X Joint, F.A.O., Additives, W. expert C. on F., others, 1991. Evaluation of certain food additives and contaminants: thirty-seventh report of the Joint FA. Keller, C.A., Kasemo, B., 1998. Surface specific kinetics of lipid vesicle adsorption measured with a quartz crystal microbalance. Biophys. J. 75, 1397–1402.

52 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

Kung, L.A., Kam, L., Hovis, J.S., Boxer, S.G., 2000. Patterning hybrid surfaces of proteins and supported lipid bilayers. Langmuir 16, 6773–6776. Lamb, J., Hentz, K., Schmitt, D., Tran, N., Jonker, D., Junker, K., 2010. A one-year oral toxicity study of sodium stearoyl lactylate (SSL) in rats. Food Chem. Toxicol. 48, 2663–2669. doi:10.1016/j.fct.2010.06.037 Lind, T.K., Cárdenas, M., Wacklin, H.P., 2014. Formation of supported lipid bilayers by vesicle fusion: Effect of deposition temperature. Langmuir 30, 7259–7263. Lloyd, J.D., 1993. The mechanisms of action of boron containing wood preservatives. Imperial College London (University of London). Löffler, J., Einsele, H., Hebart, H., Schumacher, U., Hrastnik, C., Daum, G., 2000. Phospholipid and sterol analysis of plasma membranes of azole-resistant Candida albicans strains. FEMS Microbiol. Lett. 185, 59–63. Markley, K.S., 1961. Fatty acids: their chemistry, properties, production, and uses. Interscience Publishers. Picas, L., Rico, F., Scheuring, S., 2012. Direct measurement of the mechanical properties of lipid phases in supported bilayers. Biophys. J. 102, L01–L03. Reimhult, E., Höök, F., 2015. Design of Surface Modifications for Nanoscale Sensor Applications. Sensors 15, 1635–1675. Reimhult, E., Höök, F., Kasemo, B., 2003. Intact vesicle adsorption and supported biomembrane formation from vesicles in solution: influence of surface chemistry, vesicle size, temperature, and osmotic pressure. Langmuir 19, 1681–1691. Reimhult, E., Larsson, C., Kasemo, B., Höök, F., 2004. Simultaneous surface plasmon resonance and quartz crystal microbalance with dissipation monitoring measurements of biomolecular adsorption events involving structural transformations and variations in coupled water. Anal. Chem. 76, 7211–7220. Richter, R., Mukhopadhyay, A., Brisson, A., 2003. Pathways of lipid vesicle deposition on solid surfaces: a combined QCM-D and AFM study. Biophys. J. 85, 3035–3047. Roby, M.S., Jiang, Y., Bobo, J.S., Reinprecht, J.T., 1998. Suture comprising an epsilon- caprolactone copolymer coatings and a salt of lactylate fatty acid ester; lubrication. Google Patents. Sackmann, E., 1996. Supported membranes: scientific and practical applications. Science 271, 43–48. Salafsky, J., Groves, J.T., Boxer, S.G., 1996. Architecture and function of membrane proteins in planar supported bilayers: a study with photosynthetic reaction centers. Biochemistry (Mosc.) 35, 14773–14781. Schmidt, E.L., 1984. Influence of aliphatic acids on spore germination of wood decay fungi. Int Res Group Wood Prot. IRGWP 84–2224.

53 Chapter 3: Fatty acid-based emulsifiers: possible biological agents

Schnürer, J., Magnusson, J., 2005. Antifungal lactic acid bacteria as biopreservatives. Trends Food Sci. Technol. 16, 70–78. Schultz, T.P., Nicholas, D.D., 2001. Mixture of a iodopropargyl compound, triazole compound and a hindered phenol. Google Patents. Steinem, C., Janshoff, A., Ulrich, W.-P., Sieber, M., Galla, H.-J., 1996. Impedance analysis of supported lipid bilayer membranes: a scrutiny of different preparation techniques. Biochim. Biophys. Acta BBA - Biomembr. 1279, 169–180. doi:10.1016/0005-2736(95)00274-X Stratford, M., Plumridge, A., Nebe-von-Caron, G., Archer, D.B., 2009. Inhibition of spoilage mould conidia by acetic acid and sorbic acid involves different modes of action, requiring modification of the classical weak-acid theory. Int. J. Food Microbiol. 136, 37–43. Taylor, A., Lloyd, J., 2006. Potential of near infrared spectroscopy to quantify boron concentration in treated wood. Forest. Watts, T.H., Brian, A.A., Kappler, J.W., Marrack, P., McConnell, H.M., 1984. Antigen presentation by supported planar membranes containing affinity-purified I-Ad. Proc Natl Acad Sci USA 7564–7568. Watts, T.H., Gaub, H.E., McConnell, H.M., 1986. T-cell-mediated association of peptide antigen and major histocompatibility complex protein detected by energy transfer in an evanescent wave-field. Zhdanov, V.P., Kasemo, B., 2001. Comments on Rupture of Adsorbed Vesicles. Langmuir 17, 3518–3521. doi:10.1021/la001512u Zobel, M., 1976. Toxicological Evaluation of Some Food Additives Including Anticaking Agents, Antimicrobials, Antioxidants, Emulsifiers and Thickening Agents. Who Food Additives Series, No. 5. 520 Seiten. Geneva 1974. Preis: Sw. fr. 23, Food/Nahrung 20, 681–682.

54 Chapter 4: Enzymes, Nano clay and other additives

4.1 Introduction

In this PhD study a lot of different kind of alternative bio-sustainable anti-fungal additives has been tested and evaluated. Common for all these additives is that they are intended to be included in some kind of protective system or strategies in a coating system. In this study it has only been possible to go into depth in a few additives and strategies so a selection process has been necessary and some otherwise promising and interesting ideas has been left behind.

In this chapter two short presentations of the most notable types of compound that didn’t past for further in depth studies are presented. These additives have only been researched in screening tests and further studies could be relevant in the future.

4.2 Enzymes

Two enzymes have been selected for screenings in coatings in wood panel tests in environmental growth chamber and outdoor exposures. The enzymes, pullulanase and Glucanex, were encapsulated in silica powder to gain long term effects.

4.2.1 Pullulanase

Pullulan are essential in the living of Aureobasidium pullulans (section 1.5.2). Based on that knowledge, a pullulanase has been selected. Pullulanase with EC 3.2.1.41 or also known as α-dextrin 6-glucanohydrolase, pullulan 6-glucanohydrolase, limit dextrinase, and amylopectin 6-glucanohydrolase. It is derived from various Gram-negative microorganisms such as Bacillus acidopullulyticus and Klebsiella planticola but also a range of other bacteria (Hii et al., 2012; Teague and Brumm, 1992). Pullulanase is a glucanase and a amylolytic exoenzyme that degrade pullulan from Aureobasidium pullulans (Bender and Wallenfels, 1966; Lee and Whelan, 1971; Manners, 1997).

Five groups of pullulanase have been reported, however, only two will be presented here. Pullulanase type I, will hydrolyse efficiently the α-(1,6) glucosidic bonds in pullulan and

55 Chapter 4: Enzymes, Nanoclay and other additives

branched polysaccharides. Pullulanase type II (amylopullulanases) are prominent in starch processing industry due to the specific debranching capacity of hydrolysing either the α- (1,6) or the α-(1,4) glucosidic linkages. This enzyme debranch pullulan and gives maltotriose as outcome product and it will also attacks α-(1,4) bonds in starch, amylose, and amylopectin (Bertoldo et al., 1999; Hii et al., 2012; Roy et al., 2003). Common for both of these types of pullulanases is that they attach the α-1,6 glucosidic linkages in pullulan and producing maltotriose while both of these enzyme types are unable to degrade cyclodextrin (Duffner et al., 2000).

For this study a pullulanase type I was selected and provided by Novozymes Biopharma DK A/S. The idea was to encapsulate the enzyme and add it to a control coating were the enzymes should prevent specific conidial attachment of Aureobasidium pullulans by inhibit the pullulan-based glue on the conidial surface. When the conidia reached the enzyme-rich surface the enzymes simply break the pullulan from the conidia and then the conidia would fall off or at least be inhibited in there succession.

4.2.2 Glucanex

Another enzyme that was tested was the enzyme, or rather an enzyme complex, Glucanex. It facilitates the hydrolysis of gluco-polysaccharides. Glucanex is a multicomponent enzyme preparation that consist of several isoenzymes that all contain β-1,3-glucanase activity (Elvig and Pedersen, 2003; Prieto et al., 2012). It is used commonly for making protoplast of biotechnological genetic manipulation in various fungal strains but also hydrolyzing the oligosaccharides from yeast and yeast-like cell walls in order to obtain β- glucans to control wine spoilage yeasts, protoplast preparation, and as a biocontrol agent against plant pathogenic fungi (Cheng and Bélanger, 2000; De Vries et al., 2004; Enrique et al., 2010; Kim and Yun, 2006; Slightom et al., 2009).

For this study Glucanex was selected and provided by Novozymes Biopharma DK A/S. The idea was to encapsulate the enzyme complex and add it to a control coating where the enzymes should prevent attachment and growth by Aureobasidium pullulans and other molds.

4.2.3 Aerogel – encapsulation of enzymes.

Fungal control agents in various coating systems have to be long lasting due to the customers’ requirements. Typical architectural coatings, like an exterior wood paint, are

56 Chapter 4: Enzymes, Nanoclay and other additives

expected to last 8-15 years. So it doesn’t make any sense to just add enzymes directly to a coating system without protecting the enzymes so it will last for years without just degrade immediately. However, it is important to have a constantly exposure of active enzymes around the surface to prevent microbial attack. For that purpose this study has cooperate with the company, Coatzyme ApS, in encapsulating the enzymes in an aerogel-powder solution. The enzymes will gradually be released, called slow-release, from the powder and long term effects will be achieved.

The aerogels for this project was made by Bjerga AB, Lund, Schweden and the exact method is confidential. But an accessible and similar method to make the encapsulation is described by Orçaire et al. (2006) and the following are an extract of their protocol: The procedure is overall that gels are made by hydrolysis and condensation of the silica precursors TMOS and MTMS. The wet gels were dried to aerogels by the CO2 supercritical method. In the first step, methanol, MTMS and Tetramethyl orthosilicate (TMOS) is mixed together.

Figure 4.1. Tetramethyl orthosilicate (TMOS)

Then a 4% PVA aqueous solution are prepared and mixed to a given mass of either free BCL powder or CALB solution to which an aqueous ammonia solution are added. BCL powder is homogeneously dispersed as a fine suspension in water, in which the true enzymatic component (in this case the pullulanase or the Glucanex), which is water soluble, is actually dissolved. Overall, both the true BCl protein component and the CALB, is homogeneously dispersed in a solution in the silica sol before gelation. The next step is that the two previous solutions were mixed together. Gelation will occur after a time depending of the mole fraction of MTMS to MTMS + TMOS. After homogenization but before gelation, the solution of enzyme in the silica sol which is obtained is distributed in

57 Chapter 4: Enzymes, Nanoclay and other additives

some small Teflon® tubes of internal diameter 5 mm. In each tube, a small piece of the silica fiber felt Quartzel® is previously introduced. The latter of steps is necessary to mechanically reinforce the silica aerogels. Gelation is then occurred while the silica sol is impregnating the fibre felt. After gelation, wet reinforced gels with encapsulated enzyme is extruded out of the Teflon tubes and dried by the CO2 super-critical method (CO2 critical ◦ point Tc = 31.1 C, Pc= 74 bars). The aerogel is then dried out and are now porous. A fine powder (particle size: >40 nm) with encapsulate enzymes can be made by mechanical treatment (Orçaire et al., 2006). The structural size of the aerogel powder can be seen on figure 4.2.

Figure 4.2. Structure of the aerogel dried by the CO2 super-critical method. Left: SEM images of an aerogel powder. Rigth: drawing of the TMOS structure/net where the enzymes can be encapsulated in the small cavities.

The method has previous been used with promising results in marine anti-fouling coatings by Coatzymes ApS. Field tests on a cargo ship revealed enzyme activity after 5 years, however, a significant decrease was detected.

Screening test of the enzyme activity of the encapsulated pullulanase and Glucanex showed positive results indicating that the encapsulating method only decreased the enzyme activity by approximately 35% (pullulanase) and approximately 25% (Glucanex) compared to the solution of the enzymes before adding it to the aerogel. These results have to be repeated before any publication is possible but indicate a useful encapsulation technique.

58 Chapter 4: Enzymes, Nanoclay and other additives

4.3 Nanoclay

Fillers have for long times been used in protective coatings. Typically, as a cheap compounded to reduce the costs of the manufacturing or as an additive to manipulating the stiffness or hardness of the coating properties (Nikolic et al., 2015). In wood coatings the fillers are typically calcium carbonate, barium sulfate, talc and others silica-based compounds (Nikolic et al., 2015; Osman et al., 2004).

Use of small nano fillers (fillers with at least one dimension less than 100 nm) leads to a significant change in the properties of the coating by increasing the interfacial area of the fillers with their surroundings. Coatings containing nanoparticles of different shapes like tubes, rods, spheres give the coating new properties such as improvements in mechanical-, surface-, barrier- and optical properties. The nanoparticles are typically materials of silver, gold, copper, zink oxide and clays (Fufa et al., 2012; Marathe and Kantak, 2008; Nikolic et al., 2015).

Nanoclays, or organically modified layred-silicates, have the recent years become an attractive additive in wide range of applications such as in polymer nanocomposites, rheological modifier in paints, inks, greases and cosmetics, adsorbent for toxic gases, effluent treatment and drug delivery carrier (Patel et al., 2006). Nanoclay can have a unique nano structure with thickness of just 1 nm while the length can be hundreds of nm (Fufa et al., 2012; Patel et al., 2006).

There exist many types of nanoclay. The most commonly used nanoclay is the montmorillonites Si4[Al1–67Mg0–33]O10(OH)2*nH2O x X0–33 = Na, Ca, or K. Montmorillonite clay has a 2:1 layered structure which means that two tetrahedral layers are surrounding an octahedral layer with exchangeable cations between this structure.

59 Chapter 4: Enzymes, Nanoclay and other additives

In this project regarding development of a bio-sustainable wood coating system, some of the cooperators used nanoclay as reinforcement of the alkyd binder, water protection, and influence on rheology. In that manner it became interesting to test the clay for any antimicrobial properties. Previous studies indicate that some nanoparticles in wood coatings could have an inhibiting effect on fungi like molds and wood decaying fungi. Kartal et al. 2009 showed that nanozink oxid moderately inhibit molds but not all kind of wood decaying fungi (Kartal et al., 2009). Another study conclude that nanozinc oxide in two dimensions (1x40 nm and 1x70 nm) had similar and promising inhibiting effect on termites (Clausen et al., 2011).

Three types of nanoclay were introduced to the project. They were all natural montmorillonite modified with a quatermary ammonium salt. They are all commercial available under the label Cloisite (10A, 15A and 30B). They are normally used as an additive for plastics to improve various physical properties such as reinforcement and barrier. There characteristics (provided by Southern Clay Products) are summarized in table 4.1. The modification of the nanoclay is similar to the typical quaternary ammonium based preservatives (Vijayakumar et al., 2012).

60 Chapter 4: Enzymes, Nanoclay and other additives

Modifier Organic modifier concentrations Structure (meq/100 g clay)

Dimethyl, benzyl, Cloisite hydrogenated 10A tallow, quaternary 125 ammonium

Cloisite Dimethyl, 15A dehydrogenated tallow, 125 quaternary ammonium

Cloisite Methyl, tallow, bis-2- 30B hydroxyethyl,quaternary 90 ammonium

Tabel 4.1. Characteristics of the three Cloisite.

4.4 Materials and Methods

4.4.1 Enzymes

All screenings of enzymes were performed directly in wood panel tests by adding the encapsulated enzymes (aerogel powder) to a standard coating and applied on the panels as described in section 2.4.4.

(The aerogel powder containing pullulanase and Glucanex plus an empty control were grinded and added to the standard coating.)

4.4.2 Nanoclay

Growth tests: Cloisite 10A, 15A and 30B were added to PDA media. Method completed in section 2.4. All growth tests were implemented for 5 days.

61 Chapter 4: Enzymes, Nanoclay and other additives

Wood panel tests: A standard coating was prepared for the wood panel tests and the nanoclay was added.

2 (w/v)% of Cloisite 10A was weighted into the 100% solid content alkyd polymer (60% oil length, acid number = 7) and incorporated with the use of SpeedMixer DAC 150.1 FVZ-K. Mixing time was 2x3min at 3000 rpm.

Pure alkyd and alkyd with incorporated Cloisite 10A were emulsified in water. Emulsification was performed at 65 ±5°C and low shear mixing at 200 rpm. The procedure is briefly described as following. Preheated non-ionic and anionic surfactants were first added into 50g of polymer and mixed for 15 minutes then the alkyd was neutralized with the use of 45% solution of potassium hydroxide. Water was then slowly added during 2 hours with an automated syringe pump to final 50% solid content emulsion.

For the coating preparation, defoamer, siccative and thickener were post added on Ultraturrax mixer and the formulation was diluted to final 35% solid content. No other additives were used.

Wood panel tests in the environmental growth chamber were performed as described in section 2.4.4 for 4 weeks.

4.5 Preliminary results and discussions

4.5.1 Enzymes

The results of the wood panel test indicate that both the encapsulated pullulanase and Glucanex has an inhibitory effect on the fungal growth (figure 4.4). However, the specimen with only the empty aerogel showed approximately the same results as the specimen with pullulanase encapsulated in aerogel. That indicates, that the pullulanase didn´t had an effect on the fungal control. Macro- and microscopic examination of the fungal growth on the panels revealed a predominantly majority of Aspergillus versicolor and Penicillium purpurogenum. Further studies of the fungal growth revealed none of the typical elongated shaped Aureobasidium pullulans conidia which leads to the conclusion that the faster-growing Aspergillus versicolor and Penicillium purpurogenum simply has taken over the biodiversity in the chamber. That makes the pullulanase redundant because it is chosen to have a specific effect on the pullulan secreted exclusively by Aureobasidium pullulans.

62 Chapter 4: Enzymes, Nanoclay and other additives

Glucanex seems to have a slightly more inhibitory effect when it is encapsulated in the aerogel compared with the empty aerogel. But the standard deviation gives a blurred picture of the exact effect and further studies has to be done before a useful conclusion can be made and an innovative technology prepared for use in a coating system.

Figure 4.4. Results of the enzyme tests on wood panels. From left 1) the uncoated control panel 2) Coating without any biocides 3) standard coating with a conventional biocide (approx. 0.3% IPBC) control 4) standard coating containing an empty aerogel 5) standard coating containing aerogel with pullulanase 6) standard coating containing aerogel with Glucanex. (n=9).

4.5.1 Nanoclay

Figure 4.5. Aureobasidium pullulans growth test (5 days, 25 ºC) on PDA media containing 0.01% (black), 0.03% (grey) and 0.05% (white) Cloisite 10A, 15A or 30B. (n=9)

63 Chapter 4: Enzymes, Nanoclay and other additives

The results of the growth test (figure 4.5) reviled that Cloisite 10A had a significant better inhibitory effect on Aureobasidium pullulans than Cloisite 15A and 30B. In fact, the screening showed no interesting inhibiting properties for either Cloisite 15A or 30B in any of the tested concentrations. Based on these results Cloisite 10A was selected for further coating tests on wood panels in environmental growth test. The mode of action of Cloisite 10A compared with the two other nanoclays has to be further investigated, but the presences of the benzyl and/or hydrogenated tallow instead of dehydrogenated tallow seems to make a different in the fungal inhibitory properties.

Figure 4.6. Microscope (x100) examination of Aureobasidium pullulans after 5 days growth at 25 ºC on PDA (A) Control, few conidia (red arrow) (B) PDA added 0.01% Cloisite 10A, many conidia (examples in red arrows). Bars indicate 100 µm.

Visual examination of Aureobasidium pullulans exposed for 0.01 (w/v)% Cloiste 10A compared with the control showed also difference in the appearance of the mycelium. The surface of the control were black and leather-like (se section 2.3) while the colonies exposed of Cloisite 10A had white/ligt grey appearances and microscopic examination also revealed at significant higher pretenses of conidia (figure 4.6). The presence of many conidia compared to the control could indicate that the fungal colony are inhibited and in a survival mode (Coley-Smith and Cooke, 1971; Lewis and Papavizas, 1983).

The wood panel tests with Cloisite 10A indicate that Cloisite 10A had a significant inhibitory effect compared with the standard coating without any biocides (figure 4.5). However, the Cloisite 10A didn´t had the same efficient anti-fungal effect as the commercial, toxic and generic biocide, IPBC.

64 Chapter 4: Enzymes, Nanoclay and other additives

Figure 4.5 Wood panel tests with Nanoclay. From left: 1) the uncoated control panel, 2) Coating without any biocides 3) standard coating with a conventional biocide (approx. 0.3% IPBC) control 4) 2% Nanoclay (Cloisite 10A) added to the standard coating (n=8).

A B C D

Figure 4.6 Examples of specimens in the wood panel tests. (A) Wood control [grade: 5] (B) Clear Standard Coating [average grade: 2.3] (C) 2% Cloisite 10A in the Standard Coating [grade average: 0.75] (D) 2% aerogel w. encapsulated pullulanase in the standard coating [grade average: 1.2]. The white color around the holes in the panels is a sealing (Gori 100). The small white dots at specimen D are small grains of aerogel whish was a recurring issue in all specimen with aerogels indicating a necessary change in technology so that all grain is grinded so fine that they cannot be perceived in the coating.

65 Chapter 4: Enzymes, Nanoclay and other additives

4.6 Conclusions and perspectives for further research

Based on the results in chapter 4, further studies of encapsulated Glucanex and Cloisite 10A as anti-fungal agents in wood coatings could be interesting. However, new enzymes and new kind of encapsulations methods should be tested and further analyzed for effects nearly as efficient as commercial biocides like IPBC.

Cloisite 10A showed promising properties and further coating test could be interesting. Potential improved physical contributions to the coating formulation could make Cloisite 10A an interesting dual-effective compound by contributing to physical properties and biocidal effects.

66 Chapter 4: Enzymes, Nanoclay and other additives

4.7 References Chapter 4

Bender, H., Wallenfels, K., 1966. Pullulanase (an amylopectin and glycogen debranching enzyme) from Aerobacter aerogenes. Methods Enzymol. 8, 555–559. Bertoldo, C., Duffner, F., Jorgensen, P.L., Antranikian, G., 1999. Pullulanase type I from Fervidobacterium pennavorans Ven5: cloning, sequencing, and expression of the gene and biochemical characterization of the recombinant enzyme. Appl. Environ. Microbiol. 65, 2084–2091. Cheng, Y., Bélanger, R.R., 2000. Protoplast preparation and regeneration from spores of the biocontrol fungus Pseudozyma flocculosa. FEMS Microbiol. Lett. 190, 287– 291. Clausen, C.A., Kartal, S.N., Arango, R.A., Green III, F., 2011. The role of particle size of particulate nano-zinc oxide wood preservatives on termite mortality and leach resistance. Nanoscale Res. Lett. 6, 1–5. Coley-Smith, J.R., Cooke, R.C., 1971. Survival and germination of fungal sclerotia. Annu. Rev. Phytopathol. 9, 65–92. De Vries, R.P., Burgers, K., van de Vondervoort, P.J., Frisvad, J.C., Samson, R.A., Visser, J., 2004. A new black Aspergillus species, A. vadensis, is a promising host for homologous and heterologous protein production. Appl. Environ. Microbiol. 70, 3954–3959. Duffner, F., Bertoldo, C., Andersen, J.T., Wagner, K., Antranikian, G., 2000. A new thermoactive pullulanase from Desulfurococcus mucosus: cloning, sequencing, purification, and characterization of the recombinant enzyme after expression in Bacillus subtilis. J. Bacteriol. 182, 6331–6338. Elvig, S.G., Pedersen, P.B., 2003. Safety evaluation of a glucanase preparation intended for use in food including a subchronic study in rats and mutagenicity studies. Regul. Toxicol. Pharmacol. 37, 11–19. Enrique, M., Ibáñez, A., Marcos, J.F., Yuste, M., Martínez, M., Vallés, S., Manzanares, P., 2010. β-Glucanases as a Tool for the Control of Wine Spoilage Yeasts. J. Food Sci. 75, M41–M45. Fufa, S.M., Jelle, B.P., Hovde, P.J., Rørvik, P.M., 2012. Impregnated wooden claddings and the influence of nanoparticles on the weathering performance. Wood Mater. Sci. Eng. 7, 186–195. Hii, S.L., Tan, J.S., Ling, T.C., Ariff, A.B., 2012. Pullulanase: Role in Starch Hydrolysis and Potential Industrial Applications. Enzyme Res. 2012. doi:10.1155/2012/921362

67 Chapter 4: Enzymes, Nanoclay and other additives

Kartal, S.N., Green, F., Clausen, C.A., 2009. Do the unique properties of nanometals affect leachability or efficacy against fungi and termites? Int. Biodeterior. Biodegrad. 63, 490–495. Kim, K.S., Yun, H.S., 2006. Production of soluble β-glucan from the cell wall of Saccharomyces cerevisiae. Enzyme Microb. Technol. 39, 496–500. Lee, E.Y.C., Whelan, W.J., 1971. 7 Glycogen and Starch Debranching Enzymes. The enzymes 5, 191–234. Lewis, J.A., Papavizas, G.C., 1983. Production of chlamydospores and conidia by Trichoderma spp in liquid and solid growth media. Soil Biol. Biochem. 15, 351– 357. Manners, D.J., 1997. Observations on the specificity and nomenclature of starch debranching enzymes. Journal of Applied Glycoscience 44, 83–85. Marathe, B., Kantak, A., 2008. Nano Additives: A Review. Paintindia 58. Nikolic, M., Lawther, J.M., Sanadi, A.R., 2015. Use of nanofillers in wood coatings: a scientific review. J. Coat. Technol. Res. 12, 445–461. Orçaire, O., Buisson, P., Pierre, A.C., 2006. Application of silica aerogel encapsulated lipases in the synthesis of biodiesel by transesterification reactions. J. Mol. Catal. B Enzym. 42, 106–113. Osman, M.A., Atallah, A., Suter, U.W., 2004. Influence of excessive filler coating on the tensile properties of LDPE–calcium carbonate composites. Polymer 45, 1177– 1183. Patel, H.A., Somani, R.S., Bajaj, H.C., Jasra, R.V., 2006. Nanoclays for polymer nanocomposites, paints, inks, greases and cosmetics formulations, drug delivery vehicle and waste water treatment. Bull. Mater. Sci. 29, 133–145. Peter, S.L., Woldesenbet, E., 2009. Nanoclay and Microballoons Wall Thickness Effect on Dynamic Properties of Syntactic Foam. J. Eng. Mater. Technol. 131, 021007. Prieto, M.A., Vázquez, J.A., Murado, M.A., 2012. Comparison of several mathematical models for describing the joint effect of temperature and ph on glucanex activity. Biotechnol. Prog. 28, 372–381. doi:10.1002/btpr.733 Roy, A., Messaoud, E.B., Bejar, S., 2003. Isolation and purification of an acidic pullulanase type II from newly isolated Bacillus sp. US149. Enzyme Microb. Technol. 33, 720–724. Slightom, J.L., Metzger, B.P., Luu, H.T., Elhammer, A.P., 2009. Cloning and molecular characterization of the gene encoding the aureobasidin A biosynthesis complex in Aureobasidium pullulans BP-1938. Gene 431, 67–79. Teague, W.., Brumm, P.J., 1992. Commercial enzymes for starch hydrolysis products,” in Starch Hydrolysis Products. Prod. Appl. N. Y. 45–79.

68 Chapter 4: Enzymes, Nanoclay and other additives

Vijayakumar, R., Kannan, V.V., Sandle, T., Manoharan, C., 2012. In vitro Antifungal Efficacy of Biguanides and Quaternary Ammonium Compounds against Cleanroom Fungal Isolates. PDA J. Pharm. Sci. Technol. PDA 66, 236–242. doi:10.5731/pdajpst.2012.00866

69

70 Chapter 5: Topography; Surface structure on coatings

5.1 Introduction

Imminent changes in the restrictions on the use of commercial biocides in the coating industry leads to an increased interest in alternative biological control methods and an increased interest in the mechanisms of fungal adhesion. The increased knowledge on fungal attachment mechanisms may lead to new forms of bio-renewable or bio-sustainable mechanisms to replace biocides in a near future. An interesting and innovative topic would be to research the coating topography effects on fungal conidial adhesion. In this study (see Manuscript II), six different microstructures (parallel scratches between 4-100 µm) in a standard alkyd coating were examined for the impact on the conidial attachment of Aspergillus versicolor, Penicillium purpurogenum and Aureobasidium pullulans. The results revealed that an increase in the size of the scratches led to an increase of attached conidia on the coating. Significant decreases in conidial adhesion were observed when sizes of micro scratches went below 20 µm indicating that a small topographic structure could have an inhibiting effect even when the scratches are relatively broader and deeper than the size of the conidia. The results elucidates possibilities in integrating the microsurface structures of the coatings as a part of the fungal control properties and further studies could expand the knowledge to a level where new innovative technologies can be designed.

5.2 Materials and Methods

All relevant methods and materials are included in Manuscript II.

5.2 Conclusions and perspectives for further research

The results of this study was interesting because it introduced an innovative way of thinking the micro structure in the coating as a kind of non-additive strategy for reducing the conidial attachment to the coating surface. The results indeed call for further research with new prerequisites e.g. different patterns separating from just long parallel micro scratches to other kind of structures there could be relevant for a coating technology. Also the compounds in the coating could have an influence on the results, so different kinds of

71 Chapter 5: Topography; Surface structure on coatings

coating with different kinds of topography could be scientifically important to study further.

The results indicate that it is certain that the micro topography does matter. So it is obvious to make further research based on this study.

72 Chapter 6: Hydrophobins; a unique protein in fungal surface living

6.1 Introduction

Hydrophobins are small amphiphilic proteins first described in the early 1990´s by Joseph Gerard Hubert Wessels and co-workers (Schuren and Wessels, 1990; Wessels et al., 1991). Wessels is now considered as the grandfather of hydrophobins and has published more than 35 peer-reviewed papers concerning this protein. The name “hydrophobins” was inspired by the hydrophobic structures found in and around cell walls of many prokaryotic and eukaryotic microorganisms (Wessels et al., 1991). The first nucleotide analysis of the protein was made by Shuren et al. (Schuren and Wessels, 1990) who sequenced two genes, SC3 and SC4, from Schizophyllum commune. These two genes showed similar nucleotide sequences and hydrophobicity patterns, when translated into amino acid sequences. Later the same pattern was identified in a third hydrophobin gene, SC1, in Schizophyllum commune. All three hydrophobins were excreted by S. commune in culture medium and found in the fungal walls of the aerial hyphae (SC3) and fruit body hyphae (SC1 and SC4) (J. G. Wessels et al., 1991; J. G. H. Wessels et al., 1991).

Additional studies describe a gene (rodA) in Aspergillus nidulans with similarities to the S. commune hydrophobin genes (Stringer et al., 1991). Like SC1, SC3 and SC4, the rodA gene encoded a protein located and identified in and close to the fungal cell wall. The discovery and detailed analyses of the hydrophobins resulted in a range of reclassified fungal proteins from earlier studies. Comparing their amino acid sequences with the available hydrophobins, led to the identification of the hydrophobin cerato-ulmin (CU) from Ophiostoma ulmi (Stringer and Timberlake, 1993). The biological role of the CU hydrophobin would affect the transmission of Dutch elm disease, and therefore the CU hydrophobin acts as a parasitic fitness factor against the Dutch elm (Temple et al., 1997).

Following the early 1990’s studies by Wessels, Schuren and Stringers, de Vries et al (de Vries et al., 1993) showed that hydrophobins were found in several other filamentous fungi primarily from the division of the ascomycetes but also in some basidiomycetes. As a result of the exponential progress in the sequencing of fungal genomes, the number of identified hydrophobins has significantly increased, and Linder et al. claimed in 2005 that hydrophobins can be found in all filamentous fungi (Linder et al., 2005). Despite the fact that the protein is present in all the filamentous fungi, it may seem peculiar that

73 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

they are only described and studied in relatively few fungi and the majority of these studies are in relation to industrial production fungi, like Schizophyllum commune, Aspergillus nidulans and Thricoderma reesei. No one had ever isolated, described or classified any hydrophobins in Aureobasidium pullulans, and the roles of the hydrophobins in fungi like the mould-like Aureobasidium pullulans, therefore, are unclear.

6.1.1 Nature of the Hydrophobins

The hydrophobins are proteins secreted exclusively by Filamentous fungi. They are small, approximately 100-150 amino acids in size (Wessels, 1997). Due to the surface activity of the hydrophobins and their ability to form amphiphilic protein self-assembled films at hydrophilic-hydrophobic interfaces, like surface-water or water-air (Linder et al., 2005; Wessels, 1997; Wösten and de Vocht, 2000; Wösten and Wessels, 1997), these proteins are considered as microbial surfactants. The interfacial self-assembled layer can occur e.g. between water and air, a hydrophilic cell wall and air or water and a hydrophobic solid surface (figure 6.1).

Hydrophobins have been identified and localized on the surface of aerial fungal structures such as long branching threads (the hypha), specialized spore-producing structures (the fruiting bodies) and spores (Wessels, 1997; Wösten, 2001). The localization is very diverse, like in Cladosporium fulvum where the localization of six hydrophobins was studied (Lacroix et al., 2008; Whiteford et al., 2004). Four of these hydrophobins (HCf-1, 2, 3, 4) were found on conidia and aerial hyphae. One (HCf-5) only appeared on the early aerial hyphae, and one (HCf-6) was secreted to the surrounding environment and coated the nearby solid surfaces. Like the HCf-6 hydrophobin, many hydrophobins contain a signal sequence and thereby have the ability to be secreted into the surroundings, but some of these may also be retained in the fungal structures, like the hyphae (Linder, 2009; Wessels, 1997; Wösten, 2001). Unfortunately, no bioinformatic method is currently available to predict whether a hydrophobin is secreted to the surroundings (Linder, 2009).

Hydrophobins have a unique amino acid sequence containing eight cysteine residues organized in a characteristic pattern. In spite of this very conservative cysteine pattern, the similarity among the remaining amino acids of hydrophobins is rather low (Linder et al., 2005; Wessels, 1997, 1994). Many fungi have several genes encoding hydrophobins, thus different hydrophobins are probably expressed at different times during the fungal life cycle, under different environmental conditions and may serve individual functions (Linder et al., 2005). The general role of hydrophobins in the fungal physiology and living fulfil a

74 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

variety of functions in fungal growth and development. Later in this thesis, some of these functions will be described and some new functions will be suggested.

6.1.2 Classification and protein structure

Wessels divided the hydrophobins into two classes, Class I and Class II (Wessels, 1994). This main division is still applied in most publications, although Jensen et al. indicates that not all hydrophobins in Aspergillus species can clearly be divided into those two classes (Jensen et al., 2010). Nevertheless, Class I hydrophobins have been found in Ascomycetes and Basidiomycetes, whereas Class II hydrophobins occur only in Ascomycetes (Linder et al., 2005; Wösten, 2001).

75 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

The division into Class I and Class II is based on hydropathy plots and solubility characteristics. Both kinds of hydrophobins form, or self-assemble into, aggregates. The aggregates of Class I are found to be highly insoluble and dissociated by reagents like formic acid or trifluoroacetic acid (de Vries et al., 1993). The Class II hydrophobin aggregates dissociate more readily and are soluble in e.g. 60% ethanol and 2% sodium dodecyl sulphate (SDS) (Carpenter et al., 1992; Russo et al., 1982; Wösten, 2001). If we look deeper into the overall amino acid sequence of hydrophobins, we can see that the conservation is pretty low, but all originally classified hydrophobins contain eight cysteine residues organized in a characteristic pattern including two cysteine pairs (figure 6.2 and 6.3) (Wessels, 1997). de Vocht et al. suggests that the conserved cysteine patterns are important to the structure and role of the hydrophobins (de Vocht et al., 2000). In the Class I hydrophobins the number of amino acids between cysteine residues is more variable than in Class II hydrophobins. The spacing between the third and the fourth cysteine residue is shorter in Class II hydrophobins than in the Class I hydrophobins. The two classes follow

76 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

this amino acid pattern (Hakanpää et al., 2006b) where C indicates a cysteine residue, and N represents any other amino acid than cysteine.

Class I: N{25-158}CN{5-9}CCN{4-44}CN{7-23}CN{5-7}CCN {6-18}C N{2-13}

Class II: N{17-165}CN{7-10}CCN{11}CN{15-16}CN{6-9}CCN {10-11}C N{3-8}

The eight cysteines form four disulfide bonds in the pattern Cys1-Cys6, Cys2-Cys5, Cys3- Cys4, Cys7-Cys8. In particular the Cys3-Cys4 loop can vary considerably in length compared with the other loops (Kwan et al., 2006; Wessels, 1997).

In Class I hydrophobins, the cysteine pairs are followed by hydrophilic amino acids, while the cysteine pairs are followed by hydrophobic amino acids in Class II hydrophobins (figure 6.4) (Wessels, 1997). However, all in all, the cysteine pattern is similar in both hydrophobin classes: the second and the third cysteine residue as well as the sixth and the seventh are located next to each other (figure 6.2 and 6.3).

All hydrophobins have a core fold of an irregular β-barrel consisting of four antiparallel β- sheets. In hydrophobin Class I two of the four disulphide bonds between the cysteins can be found in the center of the barrel, and the remaining two bonds is connected to the surface of the barrel with a nearby loop and an additional antiparallel β-sheet (Hakanpää et al., 2006b; Kwan et al., 2006). Class II hydrophobins contain a α-helix instead of the additional structures found in Class I. Two disulphide bonds can be found in the center of the barrel, one bridge connects the β-barrel and α-helix and one bridge attaches the β-barrel and the N-terminal loop (Hakanpää et al., 2006a, 2006b, 2004).

At low concentrations, the proteins can exist as monomers, but the hydrophobic effect drives hydrophobins to form assemblies at higher concentrations (Szilvay et al., 2006; Xiaoqin Wang et al., 2004; X. Wang et al., 2004).

77 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

6.1.3 Properties

The ability to decrease the surface tension of water has been reported for Class I hydrophobins ABH3, SC3 and SC4 and Class II hydrophobins CFTH1, CRP and HFBII, indicating that these proteins migrate into the water-air interface. With maximal reduction of the water surface tension from 72 to 24 mJ m-2, SC3 is the most surface-active protein known (Askolin, 2006; Wösten et al., 1999; Wösten and de Vocht, 2000).

78 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

79 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

6.1.4 Aims of the work related to hydrophobins

In this PhD thesis, several aims were raised in order to investigate the role of hydrophobins in surface growth of Aureobasidium pullulans. First, are these proteins of particular importance to the fungus during growth on wood or wood-coating surfaces? Second, do they influence on the fungal formations of discoloration, the blue stain? Finally, will an extensive overview and detailed understanding of the role of the protein in fungal surface growth assist in further research focusing on the development of novel approaches to control the fungi?

Several studies indicate a range of functions of the hydrophobins, and many suggestions on conventional potentials have been stated. However, no work, so far, reports the importance of the hydrophobins during fungal surface attacks on processed wood and on wood coatings, like paints.

In this study, several bioinformatics tools were applied to perform a thorough analysis of the complete genome sequence of Aureobasidium pullulans strain P268 in relation to the identification of gene sequences encoding possible hydrophobins (Manuscript III). Attempts to identify and to explore the phenotypic effects regarding surface-related growth of a hydrophobin double null-mutant in Aureobasidium pullulans strain P268 are described in Manuscript IV and section 6.2.

To support this molecular biological work, a complete genome sequencing of Aureobasidium pullulans strain P268 was initially performed. The genome sequence may be obtained by contacting Jonas Stenbæk.

6.2 Results and discussion

6.2.1 Analysis of the Aureobasidium pullulans strain P268 hydrophobins

An in-depth bioinformatics analysis of two potential hydrophobin encoding gene sequences has been conducted based on the complete genome sequence of Aureobasidium pullulans strain P268 and has been described in detail in Manuscript III. Briefly, in total four hydrophobin encoding gene sequences were found. Two in Aureobasidium pullulans strain MUCL 38722 (aur1 and aur2) and two in Aureobasidium pullulans strain (De Bary) Arnaud, P268 (hfbA and hfbB). The study in Manuscript III indicates that aur1 and hfbA and aur2 and hfbB, respectively, represent gene sequences that translate into approximately

80 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

90% identical hydrophobin amino acid sequences in the two strains. The sequences are summarized in table 6.1.

6.2.2 Disruption of the hydrophobin encoding genes, hfbA and hfbB

In order to explore the function of the hfbA and hfbB genes and the influence of hfbA and hfbB encoded hydrophobin proteins on the attachment related phenotype of Aureobasidium pullulans strain P268, a double gene knockout strategy was designed and applied, which is a natural entry point when knowing the full genome sequence of the specific strain.

6.2.2.1 Evaluation of potential double gene knockout strategies

In recent years, the need for improving the capacities of industrial production of primary and secondary metabolites from filamentous fungi has resulted in major advances within recombinant technologies with the development of several molecular tools that allow the design and construction of genetically manipulated strains. The concept of genomic mining (Gross, 2009) together with information made available for genome-wide sequencing projects have increased the understanding of fundamental genetic processes, which

81 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

improves strategies to manipulate gene expression and function in diverse filamentous fungi.

For a functional gene analysis, homologous recombination by DNA-mediated transformation is necessary in order to substitute the corresponding target in both pro- or eukaryotes. Targeted gene replacement is, however, a major challenge as most genome editing methods have low specificity and might result in unexpected mutant phenotypes or even fail to generate mutations (Jiang et al., 2013). In general, DNA-mediated transformation experiments have two distinct differences in prokaryotic bacteria and eukaryotic filamentous fungi. First, DNA integration in filamentous fungi is often mediated by non-homologous end-joining, whereas site-specific recombination occurs with low frequencies (less than 1%). Second, self-replicating vectors are somewhat rare in filamentous fungi, and often the transferred DNA is integrated into the chromosomal DNA, which makes the recombinant DNA hard to recover from transgenic fungal strains (Kück and Hoff, 2010).

Nevertheless, homologous recombination is highly required for studying various gene functions in filamentous fungi. Generally, two types of fungal transformants exist: 1) knock-in transformants in which a foreign DNA sequence is inserted into the target gene resulting in truncated transcripts, and 2) knock-out transformants in which the target gene is substituted completely by a marker gene resulting in no expression from the targeted gene. The most frequently used fungal markers are typically resistance genes against chemical drugs, such as hph (Hygromycin B resistance), ble (Phleomycin resistance), neo (Kanamycin resistance) and bar (Glyphosate resistance) (Furukawa et al., 2012; Weld et al., 2006).

In this study, an approach to construct knock-out transformants was chosen in order to completely remove the function of the hydrophobin encoding genes, hfbA and hfbB, using the hph and/or ble resistance genes for selection of transformants.

To apply this approach for fungal gene deletion, a linear DNA fragment consisting of the chosen selectable marker gene flanked on both sides by short stretches of DNA that target the gene of interest needs to be introduced (Wirsel et al., 1996). To facilitate this deletion construction synthesis, the split-marker deletion approach, previously developed for Saccharomyces cerevisiae (Fairhead et al., 1996), and later successfully applied to diverse filamentous fungi (Kück and Hoff, 2010), was at first chosen for this study.

The split-marker deletion technology requires a mixture of two linear DNA constructs per transformation, each comprising a fusion product of one of the flanking sequences of the targeted gene together with roughly two-thirds of the selectable marker gene cassette

82 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

(Catlett et al., 2003; Gravelat et al., 2012). These multisegment fragments are typically amplified and assembled by fusion PCR (Catlett et al., 2003). Following transformation, three crossing-over events are necessary to generate a functional and intact marker gene that substitutes the target gene by homologous recombination. In general, crossing-over occurs between the flank regions and their genome counterparts as well as between the overlap regions of the selectable marker gene (figure 6.7). Typically, the selectable marker gene sequences overlap for about 600 bp (Gravelat et al., 2012), and the system is designed so that neither marker gene construct on its own encodes sufficient sequence to reconstitute the functional resistance protein. Generation of a functional resistance marker gene, therefore, most likely takes place when the two DNA multisegment fragments are in close proximity to one another, such as during integration at the target locus, which promotes a triple recombination event to replace the target gene and re-establish the selectable marker gene, and minimizes the frequency of integration in random and undesired insertion locations within the genome. The risk of non-homologous end-joining is thereby significantly reduced (Kück and Hoff, 2010).

Figure 6.7: Illustration of the split-marker approach. Substitution of the wild-type target gene requires three crossing-over events in order to generate a functional marker gene. The target gene is flanked by homologous gene sequences (dark grey) that allow homologous recombination between genomic DNA and the split-marker DNA fragments to substitute the target gene by the selectable marker gene (Kuck and Hoff, 2010).

One study using a complex PCR fusion protocol reports this approach to be successful in disrupting six target genes in the two plant pathogens Cochliobolus heterostrophus and Gibberella Zeae. With hygromycin as resistance marker and using flanking sequences of only 285-761 bp in size, extremely high integration efficiencies of sometimes up to 100%

83 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

were obtained (Cattlet et al., 2003). Usually, sizes of flanking sequences are recommended to be 500-1000 bp in size (Gravelat et al., 2012). However, the split-marker system is reported to yield lower transformation rates compared to classical transformation methods (Kück and Hoff, 2010). As an example, the transformation frequency was reduced by around 85-96% in Magnaporthe grisea (Jeong et al., 2007).

In this work, successful homologous integrants were never recovered, and the split-marker approach was consequently discarded. The decision was also based on the fact that small sizes of the selected resistance marker genes (ble, 371 bp; hph, 1016 bp) would complicate the design of the overlap regions between the truncated parts of the marker gene for which a certain number of nucleotide bases are required for successful homologous recombination between the gene parts, and for which only two-thirds of the marker gene should be comprised by each integrating DNA fragment to ensure that a functional marker gene is only expressed when homologous recombination occurs between the two fragments. Especially, the small size of the ble resistance gene was expected to complicate this approach.

A second method for knock-out of the targeted gene using only one integrating linear DNA fragment, carrying the complete selectable marker gene flanked on both sides by stretches of DNA being homologous to sequences flanking the targeted chromosomal gene, was then chosen, despite increasing the risk of non-homologous end-joining and unspecific genomic integration (figure 6.8). Previously, one study reported the successful deletion of different amounts (2.5 kb, 5.7 kb and 6.3 kb) of DNA at the mating-type (MAT) locus of the Cochliobolus heterostrophus with a high integration frequency of about 90% of the transformants when introducing a linearized plasmid consisting of a selectable marker gene flanked on both sides by DNA sequences being homologous to the targeted gene locus. When using a circular plasmid, the integration efficiency was decreased to around 15% of the transformants (Wirsel et al., 1996).

84 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Figure 6.8: Simplified illustration of a gene knockout procedure using double homologous recombination (two crossing-over events) for genomic integration of one single linear DNA fragment carrying a marker gene cassette flanked by stretches of DNA that are homologous to DNA sequences flanking the chromosomal gene targeted for disruption. Figure has been applied from (Betapudi et al., 2004).

For linearized single DNA multi-segment fragments requiring only two crossing-over events (figure 6.8), the integration efficiency therefore seems appealing, and potential non- homologous end-joining might be accounted for by subsequent PCR screening techniques.

Additionally, these linearized DNA multisegment fragments for targeted chromosomal integration and heterologous expression of the selectable marker gene can easily and rapidly be simultaneously fused together using the USER (uracil specific excision reagent) cloning technology (Nisson et al., 1991). The USER cloning method is a ligation- independent cloning technique that was conceived in the early 1990s to rival conventional cloning (Nisson et al., 1991), the latter being highly dependent on the introduction of suitable restriction enzyme sites for the ligation based cut-and-paste reactions. Briefly, the power of USER cloning lies in the ability to generate long, complementary overhangs in multiple PCR products and fuse these together prior to directional insertion of the resulting stable hybridization product into specific USER cloning adapted plasmids (Hansen et al., 2011; Nour-Eldin et al., 2010, 2006). For the generation of PCR product overhangs, a custom-made nucleotide sequence of around 8 bp that ends in a single deoxyuridine residue is included as upstream extension in each USER primer. Subsequently, a proof reading polymerase that can read through uracil residues without stalling is required for amplification of the target DNA. Finally, the resulting PCR products are treated with a commercial USER enzyme mix of uracil DNA glycosylase (Smith et al., 1993) and DNA

85 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

glycosylase-lyase Endo VIII (Jiang et al., 1997) to remove single deoxyuridine residues and thereby enable the series of treated PCR products with generated long complementary single-stranded overhangs to fuse together (Nour-Eldin et al., 2010). In addition, the destination USER-compatible vector is digested with a restriction and a nicking enzyme to create the desired overhangs that are compatible to selected PCR products, treated as described above.

Figure 6.9: Overview of the USER cloning technique. The USER cassette (upper left corner) located in the USER- compatible vector contains the PacI restriction site (light blue) in the middle, flanked by oppositely oriented nicking sites, Nt.BbvCI (light pink). When the USER vector is digested with PacI and Nt.BbvCI, overhangs of 8 bp are generated. A PCR fragment (upper right corner) is amplified with deoxyuridine-containing primers adding overhangs to the PCR product that are compatible to the digested USER vector ends by an uracil-tolerating proof reading DNA polymerase. When mixed together with the digested USER-compatible vector and the USER enzyme mix, the deoxyuridine residues are excised (pink), generating single-stranded compatible overhangs, and a stable hybridization product is fused together (Nour-Eldin et al., 2010).

The long overhangs on the treated PCR fragments together with the digested vector eventually form a stable hybridization product that can be used for transformation of competent Escherichia coli cells. The USER cloning technique is illustrated in figure 6.9.

Previously, a single USER cloning step was successfully applied to insert single PCR fragments into USER-compatible plasmids that were subsequently digested into linear targeting-expression gene cassettes for rapid genomic integration by double crossing-over events in order to genetically modify multiple filamentous fungi such as Aspergillus carbonarius (Hansen et al., 2014) and Aspergillus nidulans (Hansen et al., 2011). So far, no literature describes the use of USER cloning for generating linear multisegment

86 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

integration gene cassettes for the genetic modification of the Aureobasidium pullulans strain P268. However, studies report the successful gene knockout procedure for other strains of Aureobasidium pullulans using single linear DNA integration fragments carrying a selectable marker gene flanked by stretches of DNA being homologous to sequences flanking the targeted chromosomal gene that are constructed by traditional fusion PCR methods. In one study, the gene, aba1, encoding the NRPS complex responsible for the synthesis of the cyclic peptide antibiotic Aureobasidin (AbA) in Aureobasidium pullulans strain BP-1938, was characterized. Among others, a linear aba1 gene knockout vector, carrying the bacterial hygromycin B phosphotransferase encoding gene (hph) cassette flanked by the Aureobasidium pullulans TEF1 gene promoter and a poly(A) signal region, was constructed using fusion PCR and successfully transformed into Aureobasidium pullulans BP-1938 cells to insert into the module 4 sequence of the aba1 gene, resulting in a cessation of AbA production (Slightom et al., 2009). In another study, the same approach was applied to efficiently disrupt the sidA gene in Aureobasidium pullulans strain HN6.2 using site-specific genomic integration mediated by double homologous recombination with a linear knockout vector carrying the hph gene as selectable marker (Chi et al., 2012).

In this work, it was decided, based upon the studies described above, to apply the USER cloning technology for generating linear DNA multisegment fragments for transformation into cells of Aureobasidium pullulans strain P268 with subsequent knockout of the hydrophobin encoding genes, hfbA and hfbB, by double homologous recombination. For this approach (figure 6.10), the linear knockout vector cassette consisted of a selectable marker gene, encoding resistance towards either hygromycin B (hph) or phleomycin (ble), transcriptionally fused to the thymidine kinase gene, TK, under the control of the constitutive glyceraldehyde-3-phosphate promoter, PgpdA, and followed by the trpC gene terminator from Aspergillus nidulans. In addition, this construct was flanked by 290 bp direct repeat (DR) sequences, and finally both DR regions were flanked by stretches of DNA of size 1000-2000 bp being homologous to the genomic DNA sequences representing the up- and downstream regions, respectively, of the target gene. Following selection of potential transformants with subsequent PCR-based verification of insert location, a counter-selectable strategy was applied by adding the successful integrants to a medium containing Fluoro-deoxyuridine (FUDR) that during replication would be incorporated into the DNA by the thymidine kinase gene (TK), resulting in no growth for TK carrying cells. Only integrants, for which the resistance marker gene-TK cassette was lost by homologous recombination between the flanking direct repeat regions, would prevail, which ensures an efficient counter-selectable method that allows the selectable marker gene to be recycled. This strategy further allows up to multiple rounds of gene targeting to be performed in the same mutant strain because the marker can be repeatedly

87 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

excised via direct-repeat recombination and recycled (Nielsen et al., 2006), which is a convenient method for the construction of a hfbA/hfbB double knockout mutant strain (figure 6.10).

Figure 6.10: Illustration of the knockout strategy chosen for disrupting the hfbA and hfbB genes in the Aureobasidium pullulans strain P268. The targeting linear vector cassette carries either the ble or the hph genes (here shown with hph) as selectable markers encoding resistance towards either phleomycin or hygromycin, respectively. The resistance marker gene is transcriptionally fused to the thymidine kinase encoding gene, TK, for the counter-selectable approach. The hph-TK cassette is flanked by direct repeat regions of 290 bp and these regions are finally flanked by stretches of DNA of 1000-2000 bp being homologous to the DNA sequences flanking the targeted gene, hfbA. A double homologous recombination process (two crossing-over events) between these latter regions and the genomic P268 DNA will lead to the disruption of the chromosomal hfbA gene, and potential integrants are selected for resistance towards hygromycin. The selectable marker, fused to TK, is subsequently lost when plating potential integrants on Fluoro-deoxyuridine containing medium, which is mediated by a second homologous recombination event between the direct repeat regions. The entire process can thereby be repeated for knockout of a second target gene using the same, but recycled, selectable marker.

Finally, the above described homologous recombination based insertion strategy was performed by applying the same transformation protocol as previously described (Slightom et al., 2009). This transformation protocol was originally developed by Wang et al. for the transformation of the phytopathogenic fungus Ustilago maydis with the hygromycin B resistance marker gene (hph) using polyethylene glycol-induced fusion of spheroplasts. Transformation frequencies of 50 and 1000 transformants per μg of DNA per 2x107 spheroplasts were reported for circular and linear vector DNA, respectively (Wang et al., 1988). This protocol was later modified by Cullen et al. for the transformation of Aureobasidium pullulans strain Y117 with a plasmid encoded hygromycin B resistance as

88 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

selectable marker and using Holliday’s complete medium (Holliday, 1974) supplemented with around 50-60 μg hygromycin per mL. Putative transformants were reported to appear at low frequency (approximately 1 per μg of plasmid DNA), and in contrast to Wang et al. (Wang et al., 1988) the transformation efficiency did not improve when linearizing the plasmid. The low transformation frequency was, however, partly explained by a decreased functionality of the transformed glaA promoter in Aureobasidium pullulans (Cullen et al., 1991). Finally, Slightom et al. (2009) modified the procedures published by Wang et al. (1988) and Cullen et al. (1991) for transformation of Aureobasidium pullulans strain BP- 1938 without the use of polyethylene glycol-induced fusion of protoplasts and with using at least 1 µg of circular or linear plasmid DNA (Slightom et al., 2009).

A more detailed description and discussion of the molecular methods applied for the practical workflow as well as of the preliminary results of this study is given in the following section.

6.2.2.2 Practical workflow and preliminary results from chosen gene knockout strategy

The practical workflow applied for the chosen strategy for disrupting the hfbA and hfbB genes in Aureobasidium pullulans strain P268, as well as the optimizations performed according to preliminary results, are outlined in the three flow diagrams, A-C. Briefly, to summarize the chosen approach, a linear vector cassette carrying a selectable resistance marker gene was constructed using USER cloning. The marker gene was transcriptionally fused to the thymidine kinase encoding gene, TK, and flanked by direct repeat regions, for subsequent counter-selection and recycling of the selectable marker (flow diagram B). Finally, this DNA fragment was flanked by stretches of DNA of size 1000-2000 bp being homologous to DNA sequences representing the up- and downstream regions of the chromosomal target genes, hfbA/hfbB, respectively (flow diagram A+B). Following transformation into protoplasts of Aureobasidium pullulans strain P268, potential integrants generated by homologous recombination between the linear vector cassette and the genomic DNA could be verified for correct insert location by PCR screening (flow diagram C). See also figure 6.10.

In the following, a more detailed summary of the molecular methods, optimizations and preliminary results, described by each flow diagram (A-C, next three pages), is given.

89 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

90 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

91 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

92 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

A. PCR amplification of DNA sequences flanking the up –and downstream regions of the chromosomal target genes, hfbA and hfbB. In this section, the design and optimization of USER primers was performed in order to amplify DNA sequences of around 1000-2000 bp in size flanking the up –and downstream regions of the targeted genes. The products were in all cases amplified with deoxyuridine- containing primers adding overhangs that were compatible to the nucleotide ends of the digested USER vector and of the amplified cassette carrying the selectable marker gene (see flow diagram B) by using an uracil-tolerating proof reading Hotstart DNA polymerase, PfuX7 (Nørholm, 2010). Initial primers were designed to generate up –and downstream products of 2020 bp. However, for most products, several bands were obtained with an exception of the downstream region of the hfbA gene (HfbA-down), for which only one single band of size 2300 bp was observed. Several optimization steps were then applied to generate single bands of the expected length for all regions. These included general adjustments of the PCR settings, such as lowering/increasing the annealing temperature, titrating the

concentration of several buffers like DMSO and MgCl2, changing the amount of genomic DNA template and comparing to other types of DNA polymerases. Finally, for most up –and downstream regions, new USER primers targeting slightly different genomic positions were designed in the CLC genomic workbench software (version 6.5.1), taking hairpin looping, primer-dimers etc. into account. Following these optimizations, single bands were obtained for the upstream region of the hfbA gene (HfbA-up) and the hfbB gene (HfbB-up), though with an actual size of the HfbB-up fragment larger than expected (2000 bp versus 1850 bp). For the downstream region of the hfbB gene, several smeared bands including faint bands of assumed correct size were still obtained using the new primer designs. For all regions, the purified PCR products were sequenced in order to verify the location of the amplified fragments on the Aureobasidium pullulans strain P268 genome. In brief, sequencing results revealed a completely correct sequence for the HfbA- up region. For the HfbA-down and the HfbB-up regions, both resulting in single PCR bands of larger size than expected, correct sequences according to the assembled genome sequence were found as well, but these actual sequences seemed to contain additional nucleotides of around 180-250 bp compared to the genome sequence, which might indicate that the assembled Aureobasidium pullulans strain P268 genome suffers from a lack of sequence reads in these regions, perhaps due to repetitive sequences. For the HfbB-down region a readable

93 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

sequence was never obtained for any of the gel-purified PCR bands, which might indicate that the genome sequence of strain P268 suffers from a biased assembly of sequence reads in this particular region. Consequently, it was decided to re- sequence the region downstream the hfbB gene using the primer walking method as previously described (Guo and Xiong, 2006). Primer walking did, however, not reveal the DNA sequence of this region, and therefore it was decided initially to continue the gene knockout strategy using only the USER PCR products amplified for the up –and downstream regions of the hfbA gene. The successful amplification of the DNA regions flanking the hfbA gene as well as the designed USER primers are described in detail in section 2.4 of Manuscript IV.

B. Construction of the molecular resistance marker gene cassettes followed by USER cloning of the linear vector cassette for homologous recombination-mediated integration into the genomic DNA of Aureobasidium pullulans strain P268 with subsequent disruption of the hfbA gene. First, the vector cassettes carrying the selectable marker gene were constructed. In these vector cassettes the selectable resistance marker gene, encoding either phleomycin or hygromycin resistance (ble or hph, respectively), was transcriptionally fused to the thymidine kinase gene, TK, under the control of the constitutive glyceraldehyde-3-phosphate promoter, PgpdA, and followed by the trpC gene terminator from Aspergillus nidulans. These gene cassettes were inserted into the high copy number pUC18-derived pDEL1 vector, previously described (Nielsen et al., 2006), resulting in the plasmids, pDEL1-bleTK and pDEL1-hphTK, which are pDEL1 vector based backbone plasmids carrying either the PgpdA-bleTK-TrpC or the PgpdA-hphTK-TrpC fragments, respectively, flanked by 290 bp direct repeat sequences. The latter are used for subsequent excision and recycling of the selectable marker via direct-repeat recombination following chromosomal integration and counter-selection on fluoro-deoxyuridine (FUDR) containing medium as described previously (Nielsen et al., 2006). The vector molecules were successfully constructed using either traditional ‘cut- and-paste’ cloning (pDEL1-bleTK) or USER-based cloning (pDEL1-hphTK). Traditional cloning of the pDEL1-bleTK vector has been described in detail in section 2.3 of Manuscript IV. The pDEL1-hphTK vector was constructed by amplifying a 1010 bp fragment of the hph gene using plasmid pAN7.1 (Punt et al., 1987) as template. The fragment was amplified with deoxyuridine-containing primers adding overhangs that were compatible to the USER nucleotide ends of the vector backbone of the pDEL1-bleTK plasmid, PCR amplified without the ble

94 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

gene, by using an uracil-tolerating proof reading Hotstart DNA polymerase, PfuX7 (Nørholm, 2010). Following USER cloning of the amplified fragments, the ble gene of the pDEL1-bleTK vector was directly substituted with the amplified hph gene in such a way that the hph gene was transcriptionally fused to the ble startcodon and to the start of the TK gene so that the stopcodon of the hph gene and the startcodon of the TK gene would be removed. The resulting vector molecule was verified by restriction enzyme analysis (figure 6.11a). Second, the linear vector cassettes for homologous recombination-mediated chromosomal integration were constructed. These vector cassettes would carry the stretches of DNA being homologous to the up –and downstream DNA sequences flanking the chromosomal hfbA gene (see flow diagram A) on both sides of the selectable marker gene flanked by direct repeat regions. The latter fragment was directly obtained by USER PCR amplification from either the pDEL1-bleTK or the pDEL1-hphTK plasmid as template, which resulted in a 3.6 Kbp DR-PgpdA- bleTK-TrpC-DR fragment or a 4.3 Kbp DR-PgpdA-hphTK-TrpC-DR fragment, respectively. The amplification procedure with USER primers has been described in more detail in section 2.4.1 of Manuscript IV. The primers were designed to add 8-10 bp uracil containing overhangs that were complementary to the overhangs of the reverse primer used for amplification of the DNA fragment upstream of the hfbA gene and of the forward primer used for amplification of the DNA fragment downstream the hfbA gene, respectively, as described in flow diagram A. The DNA fragments representing the up –and downstream regions of the chromosomal hfbA gene (HfbA-up and HfbA-down, respectively) would additionally carry uracil containing overhangs that were complementary to the PacI insertion sites of the pU0002 vector (Hansen et al., 2011), which following one single step of USER cloning resulted in this vector cassette carrying the complete hfbA(up)-DR- PgpdA-ble/hphTK-TrpC-DR-hfbA(down) fragment; pUSER-hfbA-bleTK and pUSER-hfbA-hphTK, respectively (figure 6.11b). Following SwaI enzyme digestion of these constructed USER cloning plasmids, the above described linear DNA fragments were isolated by gel purification and ready for homologous recombination-mediated integration into the genome of Aureobasidium pullulans strain P268 with subsequent disruption of the hfbA gene, as described in Manuscript IV.

95 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Figure 6.11: Graphic illustration of the constructed vector molecules used for the knockout procedure chosen for disrupting the hfbA gene in the Aureobasidium pullulans strain P268. A. The pDEL1-bleTK vector was constructed by traditional ‘cut-and-paste’ cloning. To construct the pDEL1-hphTK vector, the hph gene was PCR amplified from pAN7.1. A pDEL1-bleTK backbone fragment was then amplified without the ble gene and both fragments were fused together with USER cloning. B. From both pDEL1 vectors, the indicated fragments (flanked by red primer arrows) were amplified with USER primers, fused together with USER amplified DNA fragments representing the up- and downstream regions of the hfbA gene (HfbA-up and HfbA-down, respectively), and cloned into the PacI restriction enzyme sites of the USER vector pU0002. Linear integration cassettes for knockout of the hfbA gene were obtained by SwaI digestion of the resulting vector molecules.

C. Optimization of a suitable protocol for genetic transformation of Aureobasidium pullulans strain P268 protoplasts with linear vector cassettes (see flow diagram B) for chromosomal integration and disruption of the hfbA gene. Several protocols were tested for this part of the knockout strategy. First, a quick protocol, originally developed for Aspergillus nidulans (Nødvig et al., 2015) was applied. This protocol has been described in detail in section 2.4.3 of Manuscript IV, but was discarded after several tests using 1-4 µg of circular control plasmid DNA, pAN8.1 (bleR; (Mattern et al., 1988) as well as linearized integration vector cassette DNA carrying the phleomycin resistance marker gene. Selection was initially performed on non-selective PDA (potato dextrose agar) plates containing 1M sorbitol allowing colonies to appear, followed by an overlay with selective topagar allowing only true transformants to continue growing. However, non-specific growth continuously appeared also on the negative control plates after addition of selective topagar, indicating either an inappropriate

96 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

selection strategy or simply that the transformation procedure was not suitable for strains of Aureobasidium pullulans. Second, a transformation protocol modified from Wang et al. (Wang et al., 1988) and optimized for strains of Aureobasidium pullulans (Chi et al., 2012; Slightom et al., 2009) was applied. This protocol, based on stepwise addition of PTC buffer (50% w/v PEG in STC buffer, β-mercaptoethanol), has previously been reported to result in successful integrants using at least 1 μg of DNA (circular or linear) for homologous recombination between regions (integration cassette and chromosomal regions flanking the targeted gene) of only 500-800 bp in size. Selection was performed on two layer agar plates of Holliday complete medium containing 1M sorbitol, with a selective bottom layer and a non-selective top layer (Slightom et al., 2009). Despite using 2-5 μg of circular control plasmid DNA, pAN7.1 (hphR, (Punt et al., 1987) and pAN8.1 (bleR; (Mattern et al., 1988) as well as linearized integration vector cassette DNA carrying the phleomycin resistance marker gene, successful transformants were never obtained. Results displayed either most plates negative with a few potential transformants or non-specific growth on most plates, also on the negative control plates, which at least indicated that transformation efficiency might occur at a very low frequency. Despite the protocol being developed for other strains of Aureobasidium pullulans (BP-1938; Slightom et al., 2009 and HN6.2; Chi et al., 2012), this was surprising as some optimizations of the applied protocol were already implemented, such as using increased amounts of DNA for genomic integration and having increased flanking chromosomal regions for homologous recombination of 1000-2000 bp in size. Also, protoplasts of Aureobasidium pullulans strain P268 were prepared according to the protocol by Slightom et al., which resulted in high-quality protoplasts with an increased concentration of approximately 5x107 cells/mL (see section 2.4.5). A PCR-based screening method was designed in order to verify whether potential transformants were true positives. For this setup either purified genomic DNA or conidia from potential transformants were directly applied as template, the latter making this method a fast tool for evaluating the transformation success during optimization. The screening method included; (1) primers targeting the resistance gene of the transformed plasmid (either pAN7.1 or pAN8.1), resulting in a ‘yes/no’ answer based on the presence of plasmid (extra-chromosomal or integrated), and (2) primers flanking the chromosomal hfbA gene region, which in all cases resulted in PCR bands of various sizes depending on whether or not an integration at this specific genomic location had occurred. In figure 6.12 an example of potential transformants being tested for the presence of the pAN7.1 plasmid as well as for integration of the linear vector cassette carrying the

97 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

phleomycin resistance marker gene is shown together with appropriate control samples (figure 6.12). Unfortunately, PCR screening results never revealed the presence of true positive transformants, which made the hfbA/hfbB gene knockout procedure an on-going process.

Figure 6.12: Gel picture illustrating the results from the PCR-based screening method developed in order to verify whether transformants were true positives. For this example transformants no. 1-3 had been introduced to the hygromycin resistant pAN7.1 control plasmid, whereas transformants no. 4-6 had been introduced to the linear vector cassette carrying the phleomycin resistance marker gene. From these transformants, genomic (incl. plasmid) DNA had been purified to be directly applied as template in the PCR reaction. Controls were a non- template control (NTC), the plasmid pAN7.1 and chromosomal DNA purified from strain P268. The screening method included the following primers: Left part of gel; primers flanking the genomic hfbA region (hfbA_up and hfbA_down). Without an insert in this region, a band of approximately 2500 bp is expected, whereas this band is increased to around 6500 bp if the hfbA gene becomes substituted with the phleomycin resistance marker cassette. Right part of gel; primers targeting the hph gene of plasmid pAN7.1 (hph_AN71_up and hph_AN71_down). With the plasmid present (extra-chromosomal or integrated), a band of around 1000 bp is expected, and without the plasmid, no bands are obtained.

In a previous study, the transformation frequency was also reported to be somewhat low (1-10 transformants per μg of transformed DNA; (Cullen et al., 1991), which implies that increasing the concentration of DNA (control plasmids or linear integration cassettes) for transformation might be crucial. Other steps for future optimization might include; (1) further increasing the concentration of Aureobasidium pullulans protoplasts, (2) adjusting the concentration of phleomycin and hygromycin for selection of transformants, and (3) adjusting general parameters such as incubation time and incubation temperature. In general an incubation temperature of 25oC was applied, whereas other studies used incubation temperatures of 30oC (Slightom et al., 2009; Chi et al., 2012). Finally, it was speculated whether the bleTK gene cassette was less efficient in expressing resistance towards phleomycin compared to the hphTK gene cassette, which is also more frequently used in fungal genetic recombineering studies (Catlett et al., 2003; Chi et al., 2012; Cullen et al., 1991; Hansen et al., 2014; Slightom et al.,

98 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

2009). Further optimizations might, therefore, include only the hygromycin resistance marker gene. So far, transformation tests with selection for hygromycin resistance included only the control plasmid, pAN7.1. However, previous research has reported an increased transformation frequency for linear DNA templates compared to circular ones (Wang et al., 1988), which indicates the need for applying also the linearized vector cassette, carrying the hfbA(up)-DR-PgpdA- hphTK-TrpC-DR-hfbA(down) fragment, for homologous recombination-mediated integration following transformation into protoplasts of Aureobasidium pullulans strain P268.

6.3 Conclusions and perspectives for further research

The studies in chapter 6 resulted in several findings regarding the hydrophobins in relation to Aureobasidium pullulans. For the first time, gene sequences likely to encode hydrophobins were identified in the genome of a blue stain fungus using various bioinformatics tools, and the properties of the expected proteins were described. In addition, a lot of effort was put into the research of different molecular strategies for constructing a double hfbA hfbB knockout Aureobasidium pullulans P268 strain, which was expected to reveal the phenotypic impact of the hydrophobins on the abilities of this particular strain to grow on wood or wood-coating surfaces. During this process, however, several issues were detected. In particular, these included problems relating to both the amplification of genomic regions flanking the target genes, hfbA and hfbB, respectively, as well as to problems regarding the selected transformation protocol. Hopefully, ongoing work will at some point solve such technical problems.

Considering the eventual success in generating the double knockout strain supplemented with the identification of the actual secretion of hydrophobins, the obtained knowledge and characterization of the two hydrophobins in Aureobasidium pullulans will lead to further investigations regarding the role of the hydrophobins in surface growth of the fungi, which eventually might also enhance the development of novel approaches to control the fungi.

99 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

6.4 References Chapter 6

Askolin, S., 2006. Characterization of the Trichoderma reesei hydrophobins HFBI and HFBII. VTT Technical Research Centre of Finland. Betapudi, V., Shoebotham, K., Egelhoff, T.T., 2004. Generation of double gene disruptions in Dictyostelium discoideum using a single antibiotic marker selection. Biotechniques 36, 106–113. Carpenter, C.E., Mueller, R.J., Kazmierczak, P., Zhang, L., Villalon, D.K., Van Alfen, N.K., 1992. Effect of a virus on accumulation of a tissue-specific cell-surface protein of the fungus Cryphonectria (Endothia) parasitica. Mol. Plant-Microbe Interact 5, 55–61. Catlett, N.L., Lee, B.-N., Yoder, O.C., Turgeon, B.G., 2003. Split-marker recombination for efficient targeted deletion of fungal genes. Fungal Genetics Newsletter 9–11. Chi, Z., Wang, X.-X., Ma, Z.-C., Buzdar, M.A., Chi, Z.-M., 2012. The unique role of siderophore in marine-derived Aureobasidium pullulans HN6. 2. Biometals 25, 219–230. Cullen, D., Yang, V., Jeffries, T., Bolduc, J., Andrews, J.H., 1991. Genetic transformation of Aureobasidium pullulans. Journal of biotechnology 21, 283–288. de Vocht, M.L., Reviakine, I., Wösten, H.A., Brisson, A., Wessels, J.G., Robillard, G.T., 2000. Structural and functional role of the disulfide bridges in the hydrophobin SC3. Journal of Biological Chemistry 275, 28428–28432. de Vries, O.M., Fekkes, M.P., Wösten, H.A., Wessels, J.G., 1993. Insoluble hydrophobin complexes in the walls of Schizophyllum commune and other filamentous fungi. Archives of Microbiology 159, 330–335. Fairhead, C., Llorente, B., Denis, F., Soler, M., Dujon, B., 1996. New vectors for combinatorial deletions in yeast chromosomes and for gap-repair cloning using “split-marker”recombination. Yeast 12, 1439–1457. Furukawa, K., Randhawa, A., Kaur, H., Mondal, A.K., Hohmann, S., 2012. Fungal fludioxonil sensitivity is diminished by a constitutively active form of the group III histidine kinase. FEBS letters 586, 2417–2422. Gravelat, F.N., Askew, D.S., Sheppard, D.C., 2012. Targeted gene deletion in Aspergillus fumigatus using the hygromycin-resistance split-marker approach, in: Host-Fungus Interactions. Springer, pp. 119–130. Gross, H., 2009. Genomic mining–a concept for the discovery of new bioactive natural products. Current opinion in drug discovery & development 12, 207–219. Guo, H., Xiong, J., 2006. A specific and versatile genome walking technique. Gene 381, 18–23.

100 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Hakanpää, J., Linder, M., Popov, A., Schmidt, A., Rouvinen, J., 2006a. Hydrophobin HFBII in detail: ultrahigh-resolution structure at 0.75 \AA. Acta Crystallographica Section D: Biological Crystallography 62, 356–367. Hakanpää, J., Paananen, A., Askolin, S., Nakari-Setälä, T., Parkkinen, T., Penttilä, M., Linder, M.B., Rouvinen, J., 2004. Atomic resolution structure of the HFBII hydrophobin, a self-assembling amphiphile. Journal of Biological Chemistry 279, 534–539. Hakanpää, J., Szilvay, G.R., Kaljunen, H., Maksimainen, M., Linder, M., Rouvinen, J., 2006b. Two crystal structures of Trichoderma reesei hydrophobin HFBI—the structure of a protein amphiphile with and without detergent interaction. Protein science 15, 2129–2140. Hansen, B.G., Salomonsen, B., Nielsen, M.T., Nielsen, J.B., Hansen, N.B., Nielsen, K.F., Regueira, T.B., Nielsen, J., Patil, K.R., Mortensen, U.H., 2011. Versatile enzyme expression and characterization system for Aspergillus nidulans, with the Penicillium brevicompactum polyketide synthase gene from the mycophenolic acid gene cluster as a test case. Applied and environmental microbiology 77, 3044–3051. Hansen, N.B., Lübeck, M., Lübeck, P.S., 2014. Advancing USER cloning into simple USER and nicking cloning. Journal of microbiological methods 96, 42–49. Holliday, R., 1974. Ustilago maydis, in: Bacteria, Bacteriophages, and Fungi. Springer, pp. 575–595. Jensen, B.G., Andersen, M.R., Pedersen, M.H., Frisvad, J.C., Søndergaard, I., 2010. Hydrophobins from Aspergillus species cannot be clearly divided into two classes. BMC Research Notes 3, 344. Jeong, J.S., Mitchell, T.K., Dean, R.A., 2007. The Magnaporthe grisea snodprot1 homolog, MSP1, is required for virulence. FEMS microbiology letters 273, 157–165. Jiang, D., Hatahet, Z., Melamede, R.J., Kow, Y.W., Wallace, S.S., 1997. Characterization of Escherichia coli endonuclease VIII. Journal of Biological Chemistry 272, 32230–32239. Jiang, D., Zhu, W., Wang, Y., Sun, C., Zhang, K.-Q., Yang, J., 2013. Molecular tools for functional genomics in filamentous fungi: recent advances and new strategies. Biotechnology advances 31, 1562–1574. Kück, U., Hoff, B., 2010. New tools for the genetic manipulation of filamentous fungi. Applied microbiology and biotechnology 86, 51–62. Kwan, A.H.Y., Winefield, R.D., Sunde, M., Matthews, J.M., Haverkamp, R.G., Templeton, M.D., Mackay, J.P., 2006. Structural basis for rodlet assembly in fungal hydrophobins. Proceedings of the National Academy of Sciences of the United States of America 103, 3621–3626.

101 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Lacroix, H., Whiteford, J.R., Spanu, P.D., 2008. Localization of Cladosporium fulvum hydrophobins reveals a role for HCf-6 in adhesion. FEMS microbiology letters 286, 136–144. Linder, M.B., 2009. Hydrophobins: proteins that self assemble at interfaces. Current Opinion in Colloid & Interface Science 14, 356–363. Linder, M.B., Szilvay, G.R., Nakari-Setälä, T., Penttilä, M.E., 2005. Hydrophobins: the protein-amphiphiles of filamentous fungi. FEMS microbiology reviews 29, 877– 896. Mattern, I.E., Punt, P.J., Van den Hondel, C., 1988. A vector of Aspergillus transformation conferring phleomycin resistance. Fungal Genet. Newsl 35, 25. Nielsen, M.L., Albertsen, L., Lettier, G., Nielsen, J.B., Mortensen, U.H., 2006. Efficient PCR-based gene targeting with a recyclable marker for Aspergillus nidulans. Fungal Genetics and Biology 43, 54–64. Nisson, P.E., Rashtchian, A., Watkins, P.C., 1991. Rapid and efficient cloning of Alu-PCR products using uracil DNA glycosylase. Genome Research 1, 120–123. Nødvig, C.S., Nielsen, J.B., Kogle, M.E., Mortensen, U.H., 2015. A CRISPR-Cas9 System for Genetic Engineering of Filamentous Fungi. PloS one 10, e0133085. Nørholm, M.H., 2010. A mutant Pfu DNA polymerase designed for advanced uracil- excision DNA engineering. BMC biotechnology 10, 21. Nour-Eldin, H.H., Geu-Flores, F., Halkier, B.A., 2010. USER cloning and USER fusion: the ideal cloning techniques for small and big laboratories, in: Plant Secondary Metabolism Engineering. Springer, pp. 185–200. Nour-Eldin, H.H., Hansen, B.G., Nørholm, M.H., Jensen, J.K., Halkier, B.A., 2006. Advancing uracil-excision based cloning towards an ideal technique for cloning PCR fragments. Nucleic acids research 34, e122–e122. Punt, P.J., Oliver, R.P., Dingemanse, M.A., Pouwels, P.H., van den Hondel, C.A., 1987. Transformation of Aspergillus based on the hygromycin B resistance marker from Escherichia coli. Gene 56, 117–124. Russo, P.S., Blum, F.D., Ipsen, J.D., Abul-Hajj, Y.J., Miller, W.G., 1982. The surface activity of the phytotoxin cerato-ulmin. Canadian Journal of botany 60, 1414– 1422. Schuren, F.H., Wessels, J.G., 1990. Two genes specifically expressed in fruiting dikaryons of Schizophyllum commune: homologies with a gene not regulated by mating-type genes. Gene 90, 199–205. Slightom, J.L., Metzger, B.P., Luu, H.T., Elhammer, A.P., 2009. Cloning and molecular characterization of the gene encoding the aureobasidin A biosynthesis complex in Aureobasidium pullulans BP-1938. Gene 431, 67–79.

102 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Smith, C., Day, P.J., Walker, M.R., 1993. Generation of cohesive ends on PCR products by UDG-mediated excision of dU, and application for cloning into restriction digest-linearized vectors. Genome Research 2, 328–332. Stringer, M.A., Dean, R.A., Sewall, T.C., Timberlake, W.E., 1991. Rodletless, a new Aspergillus developmental mutant induced by directed gene inactivation. Genes & Development 5, 1161–1171. Stringer, M.A., Timberlake, W.E., 1993. Cerato-ulmin, a toxin involved in Dutch elm disease, is a fungal hydrophobin. The Plant Cell 5, 145. Szilvay, G.R., Nakari-Setälä, T., Linder, M.B., 2006. Behavior of Trichoderma reesei hydrophobins in solution: interactions, dynamics, and multimer formation. Biochemistry 45, 8590–8598. Temple, B., Horgen, P.A., Bernier, L., Hintz, W.E., 1997. Cerato-ulmin, a hydrophobin secreted by the causal agents of Dutch elm disease, is a parasitic fitness factor. Fungal Genetics and Biology 22, 39–53. Wang, J., Holden, D.W., Leong, S.A., 1988. Gene transfer system for the phytopathogenic fungus Ustilago maydis. Proceedings of the National Academy of Sciences 85, 865–869. Wang, X., Graveland-Bikker, J.F., De Kruif, C.G., Robillard, G.T., 2004. Oligomerization of hydrophobin SC3 in solution: From soluble state to self-assembly. Protein Science 13, 810–821. Wang, X., Permentier, H.P., Rink, R., Kruijtzer, J.A.W., Liskamp, R.M.J., Wösten, H.A.B., Poolman, B., Robillard, G.T., 2004. Probing the self-assembly and the accompanying structural changes of hydrophobin SC3 on a hydrophobic surface by mass spectrometry. Biophysical journal 87, 1919–1928. Weld, R.J., Plummer, K.M., Carpenter, M.A., Ridgway, H.J., 2006. Approaches to functional genomics in filamentous fungi. Cell research 16, 31–44. Wessels, J.G., 1997. Hydrophobins: proteins that change the nature of the fungal surface. Advances in microbial physiology 38, 1–45. Wessels, J.G., De Vries, O.M., Asgeirsdottir, S.A., Schuren, F.H., 1991. Hydrophobin genes involved in formation of aerial hyphae and fruit bodies in Schizophyllum. The Plant Cell 3, 793–799. Wessels, J.G.H., 1994. Developmental regulation of fungal cell wall formation. Annual review of phytopathology 32, 413–437. Whiteford, J.R., Lacroix, H., Talbot, N.J., Spanu, P.D., 2004. Stage-specific cellular localisation of two hydrophobins during plant infection by the pathogenic fungus Cladosporium fulvum. Fungal Genetics and Biology 41, 624–634.

103 Chapter 6: Hydrophobins; a unique protein protein in fungal surface living

Wirsel, S., Turgeon, B.G., Yoder, O.C., 1996. Deletion of the Cochliobolus heterostrophus mating-type (MAT) locus promotes the function of MAT transgenes. Current genetics 29, 241–249. Wösten, H.A., 2001. Hydrophobins: multipurpose proteins. Annual Reviews in Microbiology 55, 625–646. Wösten, H.A., de Vocht, M.L., 2000. Hydrophobins, the fungal coat unravelled. Biochimica et Biophysica Acta (BBA)-Reviews on Biomembranes 1469, 79–86. Wösten, H.A., van Wetter, M.-A., Lugones, L.G., van der Mei, H.C., Busscher, H.J., Wessels, J.G., 1999. How a fungus escapes the water to grow into the air. Current Biology 9, 85–88. Wösten, H.A., Wessels, J.G., 1997. Hydrophobins, from molecular structure to multiple functions in fungal development. Mycoscience 38, 363–374.

104

Chapter 7: Supplementary work Metagenomics; Soil micro-eukaryotes

7.1 Introduction

This study has been performed in collaboration with Samuel Jacquiod and co-authors. My contribution involves the fungal part of the analysis. The paper is attached to this PhD thesis and is considered as additional work because it has relevance to the PhD subject. However, it does not directly contribute to a solution on the aims and problems stated by the PhD study. Knowledge, as well as the application of different identification techniques, within the fungal diversity in specific environments is of crucial importance in order to understand the various types of biological attacks that need to be taken into account and to initiate the development of effective strategies for controlling these organisms.

7.2 Metagenomics provide valuable information on soil micro-eukaryotes

Despite the critical ecological roles of micro-eukaryotes in terrestrial ecosystems, most descriptive studies of soil microbes published so far focus on bacteria, fungi, or specific groups of microorganisms. Meanwhile, the fast development of metagenome shotgun sequencing results in considerable data accumulation in public repositories, providing microbiologists with substantial amounts of easily available information. We took advantage of publicly available shotgun metagenomic data in order to investigate micro- eukaryote communities in a grassland soil. The data gathered allowed the evaluation of several factors impacting the micro-eukaryote community structure, including sampling year, soil depth, and DNA extraction procedure. While most studies on soil micro- eukaryotes involve sequencing of PCR amplified taxonomic genetic markers (e.g. 18S, ITS), this work represents, to our knowledge, the first ecological report based solely on metagenomic micro-eukaryote DNA. In contrast to sampling year and depth, the data revealed a significant and pronounced effect of the DNA extraction procedure on the taxonomic composition of soil micro eukaryotes. Some DNA extraction procedures favored the detection of specific taxonomic groups, e.g. density gradient based methods enrich for protists, algae and pico-eukaryote related sequences. Our analyses suggest that publicly available metagenome data provide valuable quantitative information on soil micro-eukaryotes, complementing current qualitative ribosomal marker amplicon sequencing methods – see Manuscript V.

105

106

Manuscript I

1 Alternative Anionic Bio-sustainable Anti-fungal Agents: Investigation of 2 their Modes of Action on the Fungal Cell Membrane.

3 Jonas Stenbæk1*, David Löf2, Peter Falkman3, Bo Jensen1 & Marité Cárdenas3,4* (2015).

4

5 1: Section of Microbiology, University of Copenhagen, Copenhagen, Denmark

6 2: PPG Industries, Gladsaxevej 300, Søborg, Denmark

7 3: Department of Biomedical Science, Biofilm Research Center for Biointerfaces, Malmö University, Malmö, Sweden

8 4: Department of Chemistry, University of Copenhagen, Copenhagen, Denmark

9

10 *corresponding authors

11 E-mail: [email protected]; [email protected] (MC) and Jonas Stenbæk (JS)

12

13

14 Full title: Alternative Anionic Bio-sustainable Anti-fungal Agents: Investigation of their Modes of 15 Action on the Fungal Cell Membrane.

16

17 Short title: Antifungal Effects of Anionic Agents.

18 Abstract 19 The potential of a lactylate (the sodium caproyl lactylate or C10 Lactylate) as an anionic environmental 20 friendly anti-fungal additive was tested in growth medium and formulated in a protective coating for exterior 21 wood. Different laboratory growth tests on the blue stain fungus Aureobasidium pullulans were performed and 22 the mode of action on a model fungal cell membrane was studied. Promising short term anti-fungal effects in 23 growth tests were observed, however, less dramatic effects were observed in coating test on wood panels. 24 Scanning electron microscope analysis shows clear differences in the amount of fungal slime on the mycelium of 25 Aureobasidium pullulans when the fungus was exposed of C10 Lactylate. This could indicate an effect on the 26 pullulan production by the fungus. Our interaction studies on model fungal cell membranes show that C10 27 Lactylate affects the phospholipid bilayer in a similar manner to other negative charged detergents.

1 28 Introduction 29 Today there is an immediate need for a new generation of innovative and highly durable, protective 30 coatings and preservative systems for wood among other surfaces, especially for processed materials used in 31 construction. These protective treatments are needed to prevent microbial wood degradation when wood is 32 exposed outdoors. Current wood coatings are fully or at least partially fossil based, containing toxic biocides 33 such as 3-Iodoprop-2-yn-1-yl butylcarbamate (IPBC) that are regulated by the Biocidal Products Directive 34 (BPD). The coating industry dependence on fossil resources must be eliminated and replaced by sustainable, 35 non-fossil biosources that - at the same time - constitute environmentally friendly alternatives as described in EU 36 White paper, European Bioeconomy 2030 [1]. Many efforts are now placed into developing a bio-sustainable 37 strategy for non-toxic biocides that can be incorporated into 100% bio sustainable coating systems including 38 coatings for wood.

39 Lactylates are organic compounds (some of which are approved by the U.S. Food and Drug Administration, 40 FDA) that find use as food additives (emulsifiers) and cosmetic ingredients, among a long range of other 41 functionalities. These additives are environmentally friendly produced and they are non-toxic to humans [2,3] 42 and biodegradable in the environment [4,5]. Due to the safety and versatile functionality of these lactylates, it is 43 interesting to explore their ability as non-toxic bio-sustainable preservatives in – for example – protective 44 coatings such as wood preservative systems. Specifically, lactylates are esters of lactic acid in which the C2- 45 hydroxy group of lactic acid is esterified to a fatty acid. Lactic acids and fatty acids (FAL), the components of 46 the lactylates, are used as antimicrobial components. Lactic acid and various types of lactic acid bacteria (LAB) 47 are used as antifungal agents [6] while lactic acid is known to inhibit Salmonella and Campylobacter 48 contamination of broiler carcasses during industrial processing [7]. Moreover, FAL are used as antifungal agents 49 against molds and sapstains in wood protection systems [8]. Indeed, earlier studies indicate that, specifically, 50 small chain (C2-C8) and middle chain (C9-C14) FAL have a inhibitory effect on specific fungal germination in 51 asco- and basidiomycetes [9,10].

52 In this work, we perform a systematic study of the antimicrobial activity of a specific lactylate: the sodium 53 caproyl lactylate (C10 Lactylate), see Fig. 1. Since it contains a C10 fatty acid, it finds use as emulsifier or foam 54 booster in the cosmetic industries. This is in part due to its non-irritant character to humans and lack of toxicity 55 by ingestion. As test organism, we selected Aureobasidium pullulans (de Bary) Arnaud, strain P since this 56 organism is widely used in European Standard tests (e.g. EN 927-6:2006) of fungal stain attacks, such as Blue 57 Stain, on exterior wood [11]. This fungus is known to produce the polysaccharide pullulan to e.g. aid adhesion of 58 conidia to surfaces [12,13] and melanin to protect itself against free radicals [14,15]. Melanin and pullulan are 59 ones of the major compounds responsible for the slime and black color of the mycelium [16,17]. We perform 60 growth tests to estimate the lactylates impact on the fungal growth on agar plates and on wood panels. In order to 61 identify a possible mechanism of action for C10 Lactylate in terms of its antimicrobial activity, we investigate 62 their effect not only on the agar plates by Scanning Electron Microscopy (SEM) but also on simple model 63 cellular membranes by Quartz Crystal Microbalance with Dissipation measurements (QCM-D). Supported lipid 64 bilayers (SLB) are commonly used as simple model systems for cell membranes for instance in relation to 65 possible targets for anti-microbial agents (18). Here, we used a simple lipid mixture composed of 75 mol% 1- 66 palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) and 25 mol% 1-palmitoyl-2-oleoyl-sn-glycero-3- 67 phospho-L-serine (POPS), see Fig. 1. This mixture matches the fluidity and charge of the phospholipid

2 68 composition in ascomycetes (Candida albicans) cellular membrane [19]. The interaction between C10 Lactylate 69 and SLB was investigated using a mass sensitive technique commonly used in this type of studies: QCM-D 70 [20,21]. Finally, we compared the results to two other amphiphiles typically used in industry as emulsifiers and

71 antimicrobial agents (Sodium dodecyl sulfate, SDS, and dodecyltrimethylammonium bromide, C12TAB). Our 72 results suggest that C10 Lactylate is a promising anionic, biosustainable anti-fungal agent that attacks both the 73 integrity of the cellular membrane in a detergent like manner and the production of pullulan.

74

75 76 Fig. 1. Structures of Sodium Caproyl Lactylate and phospholipids. (A) Chemical structure of Sodium Capryl 77 Lactylate. (B) POPC. (C) POPS.

78

79

3 80 Methods 81

82 Materials 83 Pure powder (>99%) of POPC and POPS was purchased from Avanti Polar Lipids Inc and used without

84 further purification. SDS, MgCl2, C12TAB (99% Purity), potato dextrose agar/broth (PDA/PDB, respectively) 85 were purchased from Sigma and used without further purification. The Fatty acid-based C10 Lactylate was 86 kindly provided by Palsgaard A/S, which is based on an ester between lactic acid and sodium lactate and capric 87 acid. MilliQ purified water (MilliQ) was used in all experiments. Strains of Aureobasidium pullulans (De Bary) 88 Arnaud, P268 were provided by the Danish Technological Institute.

89

90 Supported Lipid Bilayer (SLB) 91 SLB were formed by vesicle fusion, a simple method in which small unilamellar vesicles (SUV) are 92 introduced onto a ultra-clean surface [22–24] like the silica sensor. Upon contact with the silica surface the 93 vesicles fuse and break forming the SLB. First, the SUV suspension was prepared following standard 94 procedures: Briefly, POPC and POPS were dissolved to a concentration of 20 mg/mL in chloroform. Lipid films

95 were then formed by mixing the lipids to the desired molar ratio, and by subsequent drying under N2 flow and 96 then under vacuum for 30 min. The films were then stored at -18 °C until use. Prior to use, the films were 97 suspended in 2 mL MilliQ to a final concentration of 0.25 mg/mL. The suspension was then bath sonicated for 98 60 min at room temperature, followed by 60 sec tip sonication (50% duty cycle 3 seconds on/off). Prior to

99 injection into the QCM-D flow cells, 2 mL of 4 mM MgCl2 were then added to the lipid suspension.

100

101 Antifungal additives

102 SDS and C12TAB were dissolved in MilliQ to concentrations (w/w %) of 0.01%, 0.02%, 0.03%, 0.04%, 103 0.05%, 0.1% and 1%. C-10 Lactylate was dissolved in MilliQ at similar concentrations but heated to 60 °C 104 under stirring until the solution was homogenous. Additives for agar test were dissolved in 1 mL sterile MilliQ 105 and heated to 60 °C. The solution was added to the autoclaved agar before clotting, in an equivalent amount to a 106 final additive-agar concentration range as described above.

107

108 Scanning electron microscopy (SEM) 109 Aureobasidium pullulans colonies were cut from the agar growth medium and attached to SEM 110 aluminum sample stubs using carbon adhesive discs (Agar Scientific). A droplet of 4% formalin was added to 111 fixate the fungi and the samples were allowed to rest on ice for 1 h. Formalin was then rinsed off using MilliQ, 112 excess water dried off in air and samples were frozenby subjecting to -80ºC freezer. After overnight freezing the 113 samples were freeze dried. Dry samples were sputter coated with gold using an Agar automatic sputter coater 114 (Oxford Instruments) at 30 mA, 0.08 mbar pressure and with a sputtering time of 40 s prior to analysis. SEM 115 micrographs were obtained using a Zeiss EVO LS10 scanning electron microscope equipped with a LaB6

4 116 filament. Imaging was done in high vacuum mode using a secondary electron detector, at 15 kV accelerating 117 voltage, 50 pA probe current and 5-6 mm working distance.

118

119 Quartz Crystal Microbalance with Dissipation (QCM-D) 120 QCM-D measurements were performed with a Q-SENSE E4 system (Q-Sense, Västra Frölunda, 121 Sweden). Sensor crystals coated with 50 nm silicon oxide were purchased from Q-Sense. Sensor surfaces were 122 cleaned by thorough rinsing in absolute ethanol followed by placing them in 2 (v/v) % Hellmanex (Hellma 123 Analytics, Germany) under bath sonication for 10 min. Sensors were then placed in MilliQ and bath sonicated

124 for additional 10 min. The surfaces were dried under N2 flow and oxidized in a UV-ozone chamber (BioForce 125 Nanosciences, Inc., Ames, IA) for 20 min in order to remove all contamination. O-Rings were placed in 2

126 (v/v)% Hellmanex for 10 min followed by careful rinsing in MilliQ and drying under N2 flow. The QCM-D 127 instrument temperature was set to 25 °C. The fundamental frequency (5 MHz) and six overtones (3rd, 5th, 7th, 128 9th, 11th, and 13th) were recorded. 4 mL of the SUV suspension was then pumped into the cells at a constant 129 flow rate of 100 mL/min and until a stable signal was obtained, typically around -24 Hz. After supported lipid 130 bilayer formation, the membranes were rinsed with excess MilliQ. Then, additives were pumped into the cells 131 until a stable signal was obtained. Finally, rinsing with excess MilliQ was performed. Results were analyzed in 132 the Q-Tools software (Q-Sense) using the 7th frequency overtone due to the lack of spreading of the various 133 overtones.

134

135 Assays of fungal growth 136 All fungal and media work were made under aseptic conditions. Agar plates were made to measure the 137 impact of the antifungal additives effect on Aureobasidium pullulans growing on a carbon rich surface. The 138 additives were prepared as described above and added to the PDA. Inoculations cultures of Aureobasidium 139 pullulans for agar test were grown in PDB for 3 days to obtain a mixture of conidia, yeast-like cells and hyphae 140 in an exponential growing state. The mixtures were used without any treatment and were inoculated by pipetting 141 5 µL of the culture to the agar plate and careful dispensed in a single droplet forming a perfect circle. The agar 142 plates were incubated for 7 days at 25 °C. The areal of the mycelium was calculated by measuring the diameter 143 of the colony. Wood panel tests were performed in an environmental test chamber as described in Bardage et al. 144 2014 [25] with minor modifications; the soil were Pindstrup Potting Soil (pH 6) and the fungal strains were 145 Aspergillus versicolor (IMI 45554), Penicillium purpurogenum (IMI 178519) and Aureobasidium pullulans (IMI 146 45533). The panels were coated with a natural based alkyd type coating formulation enriched with 2 (w/v)%

147 C10-Lactylate or C12TAB. Such alkyd formulation is a standard used for comparisons lacking any biocide. The 148 coating was applied to the panels in two steps (24 hours interval) to a final amount of 8 m2/L. The panels were 149 dried 2 days before exposed in the environmental chamber in 4 weeks.

150

5 151 Critical Micelle Concentration (CMC) measurements 152 A Lauda Drop Volume Tensiometer TVT 2 was used to measure the surface tension of C10 Lactylate 153 solutions in MilliQ. Triplicate measurements were performed in the concentration range of 0 to 0.08 (w/v) %.

6 154 Results & Discussions

155 In this work, we investigated the effects of C10 Lactylate, SDS and C12TAB as a biocidal additive in its 156 pure form in agar, and in a model protective coating for wood (paint formulation). As expected, the fungal 157 growth tests containing specific amounts of C10 Lactylate showed that C10 Lactylate had an impact on 158 Aureobasidium pullulans, see Fig. 2. The growth tests on agar plates indicated that the critical C10 Lactylate 159 concentration to induce total fungal control ranged between 0.03 and 0.05 (w/v)%. Test in liquid media support 160 these results (results not shown).

161

162 163 Fig. 2. Growth tests. Agar (PDA) screening after 7 days of growth of Aureobasidium pullulans P368 in the

164 presence of additives (A) Fatty Acid-based Lactylate (C10 Lactylate), (B) SDS and (C) C12TAB. n=9. Error bars 165 are standard deviations. The control colonies had an area of 14 cm2 (+/- 0.5 cm2).

166

167

168 Visual observations and SEM micrograph (Fig. 3) of the fungus on the agar media (PDA) containing 0.01 169 (w/v)% C10 Lactylate indicated a change in the fungal growth and morphology of Aureobasidium pullulans. 170 Besides the significant reduction of the size of the colonies (~40% size of the control), the colonies were 171 significantly paler, softer and possesed thinner and more fragile mycelium (Fig. 3 A) as compared to the normal 172 dark, thick and leather-like mycelium of the control fungi (Fig. 3 B). This indicates reduced growth and a 173 reduced production of slime (pullulan, etc) by the fungus in the presence of C10 Lactylate.

174

7 175 176 Fig. 3. Change in morphology of Aureobasidium pullulans. Aureobasidium pullulans P268 mycelium after 5 177 days of exposure to 0.01 (w/v) % C10 Lactylate in PDA media (A) and a PDA control (B): SEM micrographs of 178 the mycelium (A1 and B1) and macroscopic photographs of the colonies, bars indicate 1 cm, (A2 and B2) are 179 shown.

180

181

182 Laboratory tests in a fungal environmental chamber after 4 weeks of C10 Lactylate exposure showed that some 183 fungal growth occurred in a standard basic coating on wood, see Fig. 4. However, C10 Lactylate had promising

184 effects compared with the controls that included the standard coating and one containing C12TAB at 185 approximately the same net concentration (this value corresponds to 66 and 22 C/CMC for C10 Lactylate and

186 C12TAB respectively, see Fig. 5). The critical C10 Lactylate concentration for total fungal control was then much 187 higher in the coating formulation than in its pure form (2 relative to 0,03 (w/v) %), due in part to the more static 188 and rigid structures in the coating compared with agar structure [26–28] and also to a potential decrease in the 189 free surfactant concentration due to interactions with other components in the coating and wood.

190

8 191 192 Fig 4. Wood panel test. Results after 4 weeks exposure in an environmental growth chamber for wood panels 193 that used a standard coating (8 m2/L) without any biocide (A), or including 2 (w/v)% of the additives C12TAB 194 (B) and C10 Lactylate (C). The result of a panel is assembled in one qualitative “evaluation grade” and an 195 average of the grades was calculated (n=3). Standard deviations are included in parenthesis. Evaluation grades: 196 0: No visible growth; 1: visible < 10%; 2: 10-30%; 3: 30-60%; 4: more than 60%; 5: 100%. Each showed panel 197 is representative for the repetitive specimens.

198

199

200 Our hypothesis is that the amphiphilicity and detergency properties of C10 Lactylate are one of FAL main 201 modes of actions as biocides. Thus, we proceeded to measure the CMC of C10 Lactylate using surface tension 202 measurements and we obtained a CMC value of approximately 0.03 (w/v)% in MilliQ at room temperature (Fig.

203 5). This value can now be compared to two well-known, common detergents: C12TAB (CMC = 0.09 (w/v)% in

204 water at T =25 °C [29,30]) and SDS (CMC = 0.17 (w/v)% in water at T =25 °C [31–33]). C12TAB is used as 205 biocide due to its net positive charge while SDS carries a net negative charge in a similar manner to C10 206 Lactylate.

207

9 208 209 Fig. 5. Critical Micelle Concentration (CMC) measurements. Surface tension measurements for C10 210 Lactylate solutions in MilliQ. The Critical Micelle Concentration (CMC) was estimated to be 0.03 +/- Standard 211 deviation (w/v) % (n = 3) by the point at which the minimum surface tension (red line) was reached. Lines are 212 added as guide to the eye.

213

214

215 After identifying the capacity of C10 Lactylate to act as a biocide as well as its CMC, we proceeded to 216 investigate a possible mode of action of the additives in the fungal model cell membrane. As a simple model, a

217 SLB was made, composed of POPC (75 mol%) and POPS (25 mol%), by vesicle deposition on a SiO2 sensor in 218 the QCM-D, see Fig. 6. The traces obtained are similar to those previously reported for other lipid compositions 219 [22,24,34,35].

220

10 221 222 Fig. 6. Supported Lipid Bilayer (SLB). Typical QCM-D signal for SLB formation by vesicle fusion from 0.1

223 mg/mL lipid suspension composed of 75% POPC and 25 mol% POPS in 4 mM MgCl2 at 25 °C. For clarity only, 224 7th (blue) and 9th (red) overtones for F (full line) and D (broken lines) are shown. Steady state values upon SLB 225 formation are -23 Hz +/- 2 Hz (n =20), which were stable against extensive with MilliQ.

226

227

228 Then we proceeded to expose the SLB to C10 Lactylate solutions at various concentrations above and below its 229 CMC, see Fig. 7. QCM-D experiments showed that low C10 Lactylate concentrations (≈0.01 (w/v)%) had no 230 significant effect on the SLB, but as the C10 Lactylate concentration was increased (0.01 (w/v)% to 0.03 231 (w/v)%) there was a slight decrease in the frequency and therefore an increase of adsorbed mass to the SLB on

232 the SiO2 sensor. This indicates that the C10 Lactylate molecules interact with the SLB, co-adsorbing at the 233 interface. Interestingly, the same concentration range produced some inhibition of the fungal growth in the agar 234 test, Fig. 2. C10 Lactylate concentrations above the CMC (≈0.03 (w/v)%) resulted in a significant and

235 immediately increase of the frequency. Such change implies massive decrease of adsorbed mass to the SiO2 236 sensor comparable to the completely destruction and leaching of the SLB (a perfect SLB gives a ΔF≈-23 Hz, see

11 237 Fig. 6). Here, the results in the model cell membrane are similar to those obtained by the growth tests for which 238 C10 Lactylate concentrations above the CMC inhibited, or completely prevented, the fungi growth.

239

240

th 241 Fig. 7. QCM-D measurements. Relative change in the 7 overtone for frequency (ΔFadditive-SLB) for steady state 242 values after influx of additives at different concentrations to SLB and upon extensive rinsing with MilliQ relative 243 to the traces for SLB formation (-23.5 +/- 2 Hz). The grey zone indicates the expected threshold (including the 244 standard deviation, +/- 2 Hz) for completely removal of the SLB. (A) Fatty Acid-based Lactylate (C10

245 Lactylate), n=4 (B) SDS, n=3 and (C) C12TAB, n=3. For negatively charged surfactants, complete removal of

246 the SLB is observed at concentrations close or above the surfactant CMC. For the cationic C12TAB, on the other

247 hand, only approximately +5 Hz relative change was observed at the CMC. Thus, C12TAB was not able to fully

248 remove the SLB as SDS or C10 Lactylate did, but rather a mixed C12TAB/lipid layer remains on the surface.

249

250

251 Our QCM-D control experiments showed that low SDS concentrations well below its CMC (0-0.03 (w/v)%) 252 produced a slight increase in frequency (Fig. 7). Thus, SDS did not incorporate into the SLB but rather interacts 253 by partially destroying the SLB. When the SDS concentration comes closer to its CMC (above 0.05 (w/v)%), it 254 induced a frequency increase that compared to the increase gained by the SLB deposition (Fig. 5). This suggests 255 that SDS completely destroys and leach the SLB. Growth tests indicates the same pattern with inhibition of the 256 fungal growth with low SDS concentrations and a completely inhibition above a SDS concentration threshold of 257 0.06 (w/v)% (Fig. 2). Thus, C10-Lactylate behaves in a similar manner than SDS against model cell membranes.

258 Our second control, the cationic C12TAB showed excellent inhibitory properties in the growth tests and a critical

259 fungal growth control above concentrations of 0.01 (w/v)%, Fig. 2. Our QCM-D experiments show that C12TAB

260 did interact with the SLB at quite low concentrations since, upon C12TAB addition, the frequency increased as 261 compared to the SLB. However, the adsorbed amount at the interface remained stable above 0.03 (w/v)% with a

12 262 net value well below the corresponding for leaching of the SLB. The final frequency change values (∆f) after

263 exposure of 0.03 (w/v)% (0.33 C/CMC) C12TAB to SLB were quite small (~5 Hz) implying that C12TAB was

264 not able to fully remove the SLB and probably a mixed C12TAB/lipid admicellar layer remained on the surface

265 of the sensor. Due to the cationic charge of C12TAB, a monolayer is typically obtained on negatively charged 266 surfaces [36,37]. This is not the case for C10 Lactylate or SDS, where electrostatic repulsions between the

267 similarly charged molecules and SiO2 surface impair adsorption to SiO2.

268 In our case, the cationic C12TAB seems to have a more efficient inhibitory effect in the agar screenings than C10 269 Lactylate and SDS since both C10 Lactylate and SDS tests did not show a completely inhibition at a 0.03

270 (w/v)% concentration while C12TAB did (see Fig 5). Moreover, all three candidates inhibit the Aureobasidium 271 pullulans completely at concentrations of 0.05 (w/v)% which is above the CMC only for C10-Lactylate. The 272 QCM-D tests revealed that C10 Lactylate and SDS interact with the lipid bilayer leading to co-adsorption at 273 concentrations well below their CMC while they completely removed the SLB at concentrations above 0.04 274 (w/v)%. Together the growth tests on agar and QCM-D data show that the mode of action for C10 Lactylate can 275 be linked to its detergency and the impact on the fungal cell membrane is similar to other amphiphiles such as

276 SDS and C12TAB, although other mechanism might be also relevant including a decreased production of slimes

277 such as pullulan, etc. The fact that mixed C12TAB remains on the surface is just an effect of the overall charge of 278 the system (surface, surfactant, lipid) but the same micellar solubilization mechanism applies in solution: even if 279 QCM-D does not sense the mixed SDS or C10 Lactylate/lipid layers on the surface, mixed SDS or C10 280 Lactylate/lipid micelles exist in solution above their critical aggregation concentration, which should be at a 281 lower concentration than the CMC of the corresponding surfactant [38].

282

283

13 284 Fig. 8. Mode of Action. Schematics for effects of additive studied on model cell membrane. The negative 285 additives (SDS or C10 Lactylate) leads to integration into the lipid bilayer and its eventual total re-solubilisation

286 into mixed lipid/additive micelles. The cationic additive (C12TAB) interacts both with the anionic phospholipids 287 and the SiO2 surface, leading to formation of mixed lipid/additive aggregates in solution and the surface.

288

289

290 In this work we aimed at investigating the potential of C10 Lactylate, a FAL, as a sustainable antifungal for 291 industrial use. The long term tests in the environmental chamber (Fig. 4) showed that C10 Lactylate presented

292 similar biocide capacity than the common biocide C12TAB when iadded on the standard coating without biocide. 293 However, longer term field tests need to be done to clarify the long term effects of this FAL candidate. Finally,

294 our C10 Lactylate showed better compatibility in the coating system as compared to C12TAB: the cationic nature

295 of C12TAB complicates its incorporation into a coating system such as the wood paint.

296

297 Conclusion 298 In this work, we have shown that the FAL candidate, sodium caproyl lactylate or C10 Lactylate, had 299 promising antifungal properties. In agar growth tests C10 Lactylate inhibited the fungal growth completely at 300 concentrations as low as 0.05 (w/v)% and affected the morphology of the fungi producing a thinner and more 301 fragile mycelium. Test coating formulations on wood showed reduced fungi growth for C10 Lactylate although

302 it did not present the same antifungal impact as the cationic C12TAB. On the model fungal cell membrane

303 system, C10 Lactylate as well as the two controls used (SDS and C12TAB) incorporated in the lipid bilayer and 304 affected its integrity close to the surfactant’s CMC. Further longer exposure test studies must be done to get the 305 fully picture of the C10 Lactylate’s capabilities within new kind of bio-sustainable biocides.

14 306 References 307

308 1. McCormick K, Kautto N. The bioeconomy in Europe: An overview. Sustainability. 2013;5(6):2589–608. 309

310 2. Lamb J, Hentz K, Schmitt D, Tran N, Jonker D, Junker K. A one-year oral toxicity study of sodium 311 stearoyl lactylate (SSL) in rats. Food and Chemical Toxicology. Oktober 2010;48(10):2663–9. 312

313 3. Zobel M. Toxicological Evaluation of Some Food Additives Including Anticaking Agents, Antimicrobials, 314 Antioxidants, Emulsifiers and Thickening Agents. Who Food Additives Series, No. 5. 520 Seiten. Geneva 315 1974. Food/Nahrung. 1976;20(6):681–2. 316

317 4. Dahms GH, Tagawa M. Also containing a lactylate and a lipophilic emulsifier of a fatty alcohol; 318 cosmetics; sunscreen agents; oil in water emulsions; heat resistance; waterproofing; feel; biodegradable; 319 uneffected by ph range. Google Patents; 1997. Available: http://www.google.com/patents/US5674475 320

321 5. Baniel AM, Eyal AM, Mizrahi J, Hazan B, Fisher RR, Kolstad JJ, et al. Lactic acid production, separation 322 and/or recovery process. US 5892109 (A), 1999. Available:: https://www.google.com/patents/US5892109 323

324 6. Schnürer J, Magnusson J. Antifungal lactic acid bacteria as biopreservatives. Trends in Food Science & 325 Technology. 2005;16(1):70–8. 326

327 7. Byrd JA, Hargis BM, Caldwell DJ, Bailey RH, Herron KL, McReynolds JL, et al. Effect of lactic acid 328 administration in the drinking water during preslaughter feed withdrawal on Salmonella and 329 Campylobacter contamination of broilers. Poultry Science. 2001;80(3):278–83. 330

331 8. Clausen CA, Coleman RD, Yang VW. Fatty acid-based formulations for wood protection against mold and 332 sapstain. Forest Products Journal. 2010;60(3):301–4. 333

334 9. Harman GE, Mattick LR, Nash G, Nedrow BL. Stimulation of fungal spore germination and inhibition of 335 sporulation in fungal vegetative thalli by fatty acids and their volatile peroxidation products. Canadian 336 Journal of Botany. 1980;58(14):1541–7. 337

15 338 10. Schmidt EL. Influence of aliphatic acids on spore germination of wood decay fungi. International Research 339 Group on Wood Protection IRG/WP. 1984;84–2224. 340

341 11. Cooke WB. An ecological life history of Aureobasidium pullulans (de Bary) Arnaud. Mycopathologia et 342 Mycologia applicata. 1959;12(1):1–45. 343

344 12. Bardage SL, Bjurman J. Isolation of an Aureobasidium pullulans polysaccharide that promotes adhesion of 345 blastospores to water-borne paints. Canadian journal of microbiology. 1998;44(10):954–8. 346

347 13. Andrews JH, Harris RF, Spear RN, Lau GW, Nordheim EV. Morphogenesis and adhesion of 348 Aureobasidium pullulans. Canadian Journal of Microbiology. 1994;40(1):6–17. 349

350 14. Hung Y-C, Huang GS, Sava VM, Blagodarsky VA, Hong M-Y. Protective effects of tea melanin against 2, 351 3, 7, 8-tetrachlorodibenzo-p-dioxin-induced toxicity: antioxidant activity and aryl hydrocarbon receptor 352 suppressive effect. Biological and Pharmaceutical Bulletin. 2006;29(11):2284–91. 353

354 15. Zheng W, Campbell BS, McDougall BM, Seviour RJ. Effects of melanin on the accumulation of 355 exopolysaccharides by Aureobasidium pullulans grown on nitrate. Bioresource Technology. 356 2008;99(16):7480–6. 357

358 16. Gadd GM. Melanin production and differentiation in batch cultures of the polymorphic fungus 359 Aureobasidium pullulans. FEMS Microbiology Letters. 1980;9(3):237–40. 360

361 17. Gaur R, Singh R, Gupta M, Gaur MK. Aureobasidium pullulans, an economically important polymorphic 362 yeast with special reference to pullulan. African Journal of Biotechnology. 2010;9(47):7989–97. 363

364 18. Kagan BL, Selsted ME, Ganz T, Lehrer RI. Antimicrobial defensin peptides form voltage-dependent ion- 365 permeable channels in planar lipid bilayer membranes. Proceedings of the National Academy of Sciences. 366 1990;87(1):210–4. 367

368 19. Löffler J, Einsele H, Hebart H, Schumacher U, Hrastnik C, Daum G. Phospholipid and sterol analysis of 369 plasma membranes of azole-resistant Candida albicans strains. FEMS microbiology letters. 370 2000;185(1):59–63. 371

16 372 20. Lind TK, Zielinska P, Wacklin HP, Urbanczyk-Lipkowska Z, Cárdenas M. Continuous flow atomic force 373 microscopy imaging reveals fluidity and time-dependent interactions of antimicrobial dendrimer with 374 model lipid membranes. ACS nano. 2013;8(1):396–408. 375

376 21. Lind TK, Polcyn P, Zielinska P, Cárdenas M, Urbanczyk-Lipkowska Z. On the Antimicrobial Activity of 377 Various Peptide-Based Dendrimers of Similar Architecture. Molecules. 2015;20(1):738–53. 378

379 22. Åkesson A, Lind T, Ehrlich N, Stamou D, Wacklin H, Cárdenas M. Composition and structure of mixed 380 phospholipid supported bilayers formed by POPC and DPPC. Soft Matter. 2012;8(20):5658–65. 381

382 23. Reviakine I, Brisson A. Formation of supported phospholipid bilayers from unilamellar vesicles 383 investigated by atomic force microscopy. Langmuir. 2000;16(4):1806–15. 384

385 24. Richter RP, Bérat R, Brisson AR. Formation of solid-supported lipid bilayers: an integrated view. 386 Langmuir. 2006;22(8):3497–505. 387

388 25. Bardage S, Westin M, Fogarty HA, Trey S. The effect of natural product treatment of southern yellow pine 389 on fungi causing blue stain and mold. International Biodeterioration & Biodegradation. 2014;86:54–9. 390

391 26. Duckworth M, Yaphe W. The structure of agar: Part I. Fractionation of a complex mixture of 392 polysaccharides. Carbohydrate Research. 1971;16(1):189–97. 393

394 27. Perez C, Collazo A, Izquierdo M, Merino P, Novoa XR. Characterisation of the barrier properties of 395 different paint systems: Part I. Experimental set-up and ideal Fickian diffusion. Progress in Organic 396 Coatings. 1999;36(1):102–8. 397

398 28. Perez C, Collazo A, Izquierdo M, Merino P, Nóvoa XR. Characterisation of the barrier properties of 399 different paint systems: Part II. Non-ideal diffusion and water uptake kinetics. Progress in Organic 400 Coatings. 1999;37(3):169–77. 401

402 29. Monteux C, Williams CE, Meunier J, Anthony O, Bergeron V. Adsorption of oppositely charged 403 polyelectrolyte/surfactant complexes at the air/water interface: formation of interfacial gels. Langmuir. 404 2004;20(1):57–63. 405

17 406 30. Ray GB, Chakraborty I, Moulik SP. Pyrene absorption can be a convenient method for probing critical 407 micellar concentration (cmc) and indexing micellar polarity. Journal of colloid and interface science. 408 2006;294(1):248–54. 409

410 31. Sanchez-Camazano M, Sanchez-Martin MJ, Rodriguez-Cruz MS. Sodium dodecyl sulphate-enhanced 411 desorption of atrazine: Effect of surfactant concentration and of organic matter content of soils. 412 Chemosphere. 2000;41(8):1301–5. 413

414 32. Mukerjee P, Mysels KJ. Critical Micelle Concentrations of Aqueous Surfactant Systems. National Bureau 415 of Standards; 1971. Accession Number: ADD095344. Available: 416 http://oai.dtic.mil/oai/oai?verb=getRecord&metadataPrefix=html&identifier=ADD095344 417

418 33. Hayase K, Hayano S. Effect of alcohols on the critical micelle concentration decrease in the aqueous 419 sodium dodecyl sulfate solution. Journal of Colloid and Interface Science. 1978;63(3):446–51. 420

421 34. Jurkiewicz P, Cwiklik L, Vojtíšková A, Jungwirth P, Hof M. Structure, dynamics, and hydration of 422 POPC/POPS bilayers suspended in NaCl, KCl, and CsCl solutions. Biochimica et Biophysica Acta (BBA)- 423 Biomembranes. 2012;1818(3):609–16. 424

425 35. Lind TK, Cárdenas M, Wacklin HP. Formation of supported lipid bilayers by vesicle fusion: Effect of 426 deposition temperature. Langmuir. 2014;30(25):7259–63. 427

428 36. Decher G, Hong J-D. Buildup of ultrathin multilayer films by a self-assembly process: II. consecutive 429 adsorption of anionic and cationic bipolar amphiphiles on charged surfaces. Makromolekulare Chemie 430 Macromolecular Symposia. Wiley Online Library; 1991. s. 321–7. Available: 431 http://onlinelibrary.wiley.com/doi/10.1002/masy.19910460145/full 432

433 37. Kügler R, Bouloussa O, Rondelez F. Evidence of a charge-density threshold for optimum efficiency of 434 biocidal cationic surfaces. Microbiology. 2005;151(5):1341–8. 435

436 38. Kronberg B, Holmberg K, Lindman B. Surface chemistry of surfactants and polymers. First Edition. John 437 Wiley & Sons, Ltd. Published 2014, by John Wiley Online Library. DOI: 10.1002/9781118695968.ch9

18

Manuscript II

Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

The influence of the topography on the fungal conidia adhesion in a protecting coating system

Jonas Stenbæk1*, Miroslav Nikolic2, Berit Lindegaard3, David Löf4, Bo Jensen1 and Anne Christine Steenkjær Hastrup3*

1: Section of Microbiology, University of Copenhagen, 2100, Copenhagen, Denmark 2: Department of Geosciences and Natural resource Management, IGN, Copenhagen University, Rolighedsvej 23, Frederiksberg C, Denmark 3: Centre for Wood Technology and Biobased Materials, Danish Technological Institute, Gregersensvej 1, Taastrup, Denmark 4: PPG Industries, Gladsaxevej 300, Søborg, Denmark

Abstract

Imminent changes in the restrictions on the use of commercial biocides in the coating industry leads to an increased interest in alternative biological control methods and an increased interest in the mechanisms of fungal adhesion. The increased knowledge on fungal attachment mechanisms may lead to new forms of bio- renewable or bio-sustainable mechanisms to replace biocides in a near future. An interesting and innovative topic would be to research the coating topography effects on fungal conidial adhesion. In this study, six different microstructures (parallel scratches between 4-100 µm) in a standard alkyd coating were examined for the impact on the conidial attachment of Aspergillus versicolor, Penicillium purpurogenum and Aureobasidium pullulans. The results revealed that an increase in the size of the scratches led to an increase of attached conidia on the coating. Significant decreases in conidial adhesion were observed when sizes of micro scratches went below 20 µm indicating that a small topographic structure could have an inhibiting effect even when the scratches are relatively broader and deeper than the size of the conidia. The results elucidates possibilities in integrating the microsurface structures of the coatings as a part of the fungal control properties and further studies could expand the knowledge to a level where new innovative technologies can be designed.

Introduction rethink the formulation of their products. More bio- sustainable component solutions and production New requirements to the creation of paint- and methods are demanded by customers, as well as coating have now forced the coating industry to required by national and international

* Corresponding: (JS) [email protected] and (ACHA) [email protected] 1 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

regulations. During the course of a nearby future a Deuteromycota (e.g. Aspergillus fumigatus) is a substantial substitution of raw materials must be division of fungi for which a sexual cycle has not expected by the industry and new kind of been observed and its members produce only ingredients, recipes and strategies must be conidia. Conidia spread through the air and every developed. The production and use of biocides in single fungal organism continuously produces protective coatings have always been subject millions of conidia. Every conidium have the to strict directives, e.g. The Biocidal Products potential to give rise to a new fungal colony if they Regulation (BPR), but it must be expected that even are able to attach to a conquered surface where quite more stringent restrictions will be introduced in the simple growth requirements, like water and future. Therefore, antifungal strategies must be nutrients, are present. Germinating conidia reconsidered and new ideas and go through a significant escalation in their adhesive technologies developed to substitute today’s use of properties. It is a process that includes the initial environmental inculpatory biocides, like IPBC, adhesion initiated by the pre-existing glycoprotein DCOIT. In this study, we focused on the fungal layer or different components in the conidial cell control in protective coatings for exterior wood. wall (Osherov and May, 2001). In the blue stain fungus, Aureobasidium pullulans, it is the major One attacking point for fungal control in a effects of pullulan that act as an adhesion protective exterior coating could be the adhesion of glue (Bardage, 1996; Bardage and Bjurman, 1998; the asexual fungal spores, the conidia. The fungal Catley, 1980; Lazaridou et al., 2002). After the first attacks on woods and wood coating surfaces can adhesion to a surface, tighter adhesion appears as a lead to a range of different unwanted results like result of the fungal metabolism and fungal secreted discolorations, e.g. blue stain, algae growth or proteins. The adhesion process happens fast and actual decay and weakening of the wood. some species, like Blumeria graminis, responds The fungal life cycles are complex and diverse, but immediately i.e. within seconds of contact to a one critical point within most of the wood attacking surface, by releasing an extracellular matrix that fungi is the dispersion of the fungal reproductive bind the conidium strong and effective to its host cells. Fungi from the divisions Ascomycota (e.g. plant surface (Bailey et al., 1992). During the Penicillium, Aspergillus and Aureobasidium), conidial germination some fungi use mucilage, Basidiomycota (Gloeophyllum and Serpula), and which is a thick, viscous substance and a Zygomycota (e.g. Rhizopus and Mucorales) produce exopolysaccharide glycoprotein (Braun and both sexual spores called ascospores, basidiospores Howard, 1994). Osherov and May, and zygospores respectively, and their very 2001, described a number of other conidial adhesion common asexual spores, called conidia. mechanisms; Colletotrichum species where conidia

2 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

are embedded in a pre-formed, water-soluble An study by Gobakken et al. conclude that soft extracellular matrix composed of glycoproteins coating structure had better anti-mould properties (Sela-Buurlage et al., 1991). Conidia of than hard paints but the study also called up for Magnaporthe grisea adheres to the repellent surface further studies in surface structure to add important of its host, the rice plant, by releasing an adhesive knowledge to the field (Gobakken et al., 2010). from apex of the conidium and it is triggered by Gobakken et al. emphasize the importance in wetting of the condium on the conidiogenous cells bringing attention to the construction design, the (Mims et al., 1995). Tronchin et al. (1995) conclude specific climatic factors and interactions involved that the surface of conidia from the powdery mildew when evaluating the different parameters in the fungus Uncinuliella Australiana is coated with a research of anti-fungal attacks on coatings thin network of extracellular mucilage, which, on (Gobakken et al., 2008). contact, with a wet surface, spreads instantly to There is not much literature available concerning form sticky adhesion pads. The adhesion the effect of the surface charge or the conidial mechanisms of numerous Aspergillus species has surface charge in the matter of attachment. A study been studied and two examples can be mentioned; on conidial adhesion of Aspergillus Sp. conclude Aspergillus fumigatus has a laminin-binding protein that conidial adherence to fibronectin and other in its cell wall that adhere through the hosts basal-lamina proteins is mediated via negatively extracellular matrix proteins; fibronectin, collagen charged carbohydrates on the conidial surface and laminins (Thau et al., 1994). Hydrophobins, (Wasylnka and Moore, 2000). secreted by all filamentous fungi including Aspergillus sp. (Linder et al., 2005; Wösten, 2001), With the more strict regulations regarding the use of are indicated to play an important role in the biocides an increased focus on preventive protection conidial adhesion during the formation of is essential and new ways of fungal control agents amphiphilic layers on the conidial wall (Bayry et and strategies has to be explored and developed. In al., 2012; Bell-Pedersen et al., 1992; Linder, 2009; this study we will investigate the influence of the Paris et al., 2003). A deletion of the hydrophobin surface topography of a standard protective coating, genes rodA or dewA, responsible for the outer containing no biocides, on the adhesion of a mixture rodlett-layer on the conidia of A. fumigatus and A. of the three fungal species Aspergillus versicolor, nidulans decreased adherence to e.g. collagens on Penicillium purpurogenum and Aureobasidium the host surfaces (Bayry et al., 2012; Stringer and pullulans. Timberlake, 1995; Sundstrom, 1999). The hydrophobicity of the surface could be a factor in the conidial adhesion.

3 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

Methods and Materials glass slides were prepared for scanning electron microscopy (SEM) by placing a piece of the glass Preparation of the coating slide on a metal SEM-stub. The stub was transferred to a Thermo Scientific FEI Quanta 400 scanning The coating was based on 50% solid content alkyd electron microscope and the specimen was emulsion (60% oil length, acid number = 7) examined. The pictures were taken and saved obtained from Dyrup/PPG Industries together with digitally. the other additives. A defoamer, iron based siccative and an associative thickener were added under The size of the scratches was estimated based on the medium shear on a dissolver. During the mixing, the SEM analysis. The grain size of the sandpaper formulation was diluted to final 35% solid content. resulted in grooves with the following widths: P No other additives were used. 100: (specimen 1.1, 1.2, 2.2): approx. 80-100 µm wide, P 120: (specimen 3.1, 3.2, 4.2): approx. 40-50 µm wide, P 150: (specimen 5.1, 5.2, 6.2): approx. Making the topography 15-20 µm wide, P 180: (specimen 7.1, 7.2, 8.1): 10 µm wide, P 280: (specimen 2.1, 4.1, 9.1): 5-8 µm Microscopy Glass slides (VWR) were weighed and wide, P 400: (specimen 6.1, 8.2, 9.2): 4-5 µm wide, coated by brushing. The weight of the applied coated, non-sandpaper treated (control): (specimen amount of coating was recorded. The topography 10.1, 10.2): no grooves (figure 1). The contact was constructed by making light transverse strokes angle (Ɵ), which describes how a liquid and the with sandpaper in the longitudinal direction three surface influence each other, was measures. This times over the surface. Microscopy Glass slides was done by adding a droplet of water to the surface with no coating and coated without sand paper and evaluation the angle between the surface and treatment were used as controls. Three replicates the tangent to the water surface where the three were prepared for each treatment. The microscope phases (air, water and surface) meet (figure 2).

4 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

Figure 1. SEM images of the surface scratches in the coatings (A) Specimen 2.2: Sandpaper grain size P 100 (B) Specimen 4.2: Sandpaper grain size P 120 (C) Specimen 6.2: Sandpaper grain size P 150 (D) Specimen 8.1: Sandpaper grain size P 180 (E) Specimen 9.1: Sandpaper grain size P 280 (F) Specimen 8.2: Sandpaper grain size P 400 (G) Specimen Control, no scratches.

Figure 2. Measurements of contact angel (Ɵ). A: Droplet on coated surface. B: Placement of the tangents for measuring the contact angle..

5 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

Fungal treatment Growth chamber at 23° C, 95% RH and were incubated for at least 2 weeks before use. A fungal mixture was made of Aspergillus versicolor (IMI 45554), Penicillium purpurogenum The growth chamber was a rectangular container (IMI 178519) and Aureobasidium pullulans (IMI made of PVC with a triangle top, making it “house- 45533), by placing fungal inoculated agar plugs looking”. The bottom of the growth chamber was from each species on the agar surface of the same covered with 2-3 cm demineralized water (27o C). Petri dish. The fungi were grown on potato dextrose Over the water, four trades with net-bottoms (1 mm agar (PDA) for 4 days and a spore suspension was pores) were placed and autoclaved soil was added. prepared from the agar plate by adding 250 ml of The airflow caused by the water flow from the distilled water. The suspensions was added to the heated water made the conidia dispersed soil (Pindstrup Potting Soil, pH 6) in a growth homogenously in the chamber (Bardage et al., 2014) see figure 3.

Figure 3. Environmental growth chamber.

The coated specimen and untreated (coated and moisture was detected by visual inspection or non-coated) control samples were mounted on wood magnification (10x). panels (cedar tree) with two-sided adhesive tape. The specimens were placed on brackets in the top of the chamber (figure 3) for 4 weeks. The wood Data collecting panels were removed from the chamber and the Every glass specimen was examined in a phase glass slides carefully collected by removing the contrast light microscope (Leica DM 2000 LED glass slides from the wood panels. No surface

6 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

with a Leica MC120 HD camera). The location and Results amounts of attached conidia and hyphae were estimated by evaluation of the overall trends. The specimens were examined for conidia (figure 4) Ultimately a rating was performed based on the and conidial ratings were performed concurrent with number of conidia in and around the scratches in the the measurements of the contact angels (Ɵ) (table coating films (table 1). 2).

Table 2. Surface angles (Ɵ) and grade of conidial adhesion compared to the sizes of the scratches on the coating surfaces (uncoated control: n=1, coated control and 40-50 µm: n=2, rest: Table 1: Rating of the conidial adhesions on the specimens n=3). Details and evaluation for every specimen can be seen in (conidial rate). supplementary table S1. Only one sanding group (P180) were subject to a standard deviation (0.9), all other triplet samples were rated equally relative to the conidial rate.

7 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

A B C

D E F

G H I

Figure 4. Light microscopy (phase contrast) of the coatings after exposures in the environmental chamber. A: Control (x50, specimen 10.1). Many conidia, red circle B: (x200, specimen 10.1) C: P100 (x100, specimen 2.2). Many conidia chains, red arrow D: P120 (x100, specimen 4.2). Conidia chains, red arrow E: P150 (x100, specimen 6.2), not many conidia in the scratches, but a few in between the scratches F: P180 (x200, specimen 8.1), only a few conidia, red arrow G: P280 (x200, specimen 9.1), few conidia but some observed between the scratches, red arrow H: 280 (x400, specimen 9.1), an approximately 5-7 µm scratch with no conidia attached despite of multiple weeks of fungal exposures I: P400 (x200, specimen 8.2). No conidia.

The alkyd coating had a 56° contact angle with contact angle was measured to be 64° with the least water allowing partial wetting of the surface. The rough sandpaper used in this study (P400). It two coarsest sandpapers reduced the apparent appears that the increase in contact angle is in contact angle, while the other sandpaper grades accordance with the reduction of the amount of resulted in an increase in contact angle. The highest fungal growth (figure 5).

8 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

Figure 5. Contact angles (red) and conidial rates (blue) as a function of the topography (size of the scratches). The lines indicate the linear trend lines. The trend line for contact angels does not contain the contact angle for the coating with no sanding (red cross)).

As an additional observation, strong hyphal growth Discussion was observed along the edges of the glass slides We observed a significant higher amount of outside the scratched area (Figure 6). The fungal dispersed and attached conidia in the two controls conidia germinate on the wooden panel and the (coated and uncoated). Germinated conidia and hyphae grow up on the glass slides bringing the hyphae were also observed in substantial amounts. nutrition from the wood. That supports the fact that The dispersion of the conidia was homogenous and the coating in itself is not anti-fungal. the attachment was strong.

The presence of topography was initially expected to initiate easier conidial attachment due to the larger surface area and a more rough structure where the conidia could get trapped and attach. The more rough topography was also expected to facilitate the wetting of the surface and thereby to create a more attractive environment than the smooth controls with no scratches or just pure glass. This corresponds well to the observation that the size of the scratches had a significant influence on Figure 6. Hyphal growth along the edges of the glass slides. the adhesion of the conidia. The smallest scratches

9 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

(< 10 nm) had a substantial inhibitory effect on the adding a component in the coating formulation. We conidial adhesion. An increase in the size of the show that it can also be the physical change of the scratches leads to an increase of attached conidia surface roughness that may influences the level of (table 2 and figure 4 and 5). fungal protection. However we show that this probably occurs due to a change in surface tension In relation to the size of the created microscratches, associated with the surface roughness level. It can higher conidial grade was expected in the specimen be very important if this effect can be used in actual with bigger scratches, because of the deeper micro outdoor conditions to increase the protection level ravines and clefts in the coating structure, resulting and it could be a route to pursue in the future. With in more hiding places and protection from the largest micro scratches the contact angle environmental changes. The increase in the decreased compared to the smooth alkyd coating. topography/level differences was found to indicate This usually happens when roughness of a an increased adhesion of the spores and homogeneous surface is increased and it is conidia. This is still regarding to the described by a model proposed by Wenzel (Wenzel, microscopic level, it could be significant 1936). With decrease of scratch size, the contact different on the macroscopic level angle was measured to actually increase by a small for example planed and sawn wood. However, it level. This probably means that the equilibrium state was unexpected to find so few conidia in the changed to another, which now has a lower free smallest scratches. The small scratches are still energy minimum. One should remember that the broader and deeper than the single conidia, so we pure coating also has a certain level of surface expected conidia to hide in the ravines. roughness. Pure polymer binder with additives In an extensive study the smallest amount in respect typically has surface roughness below 200 to mould growth on paints were obtained, when a nm (Dhoke et al., 2009, 2008; Poaty et al., 2014; matting agent was used (Gobakken et al., 2010). Veigel et al., 2014). When nano roughness is Matting agents are typically fillers, like amorphous combined with micro roughness a range of wetting silica, that reduce the gloss level. Gloss is reduced modes is possible to occur depending on whether due to a change in way light reflects from the the droplet/air fills the micro and/or nano surface as the filler has increased the surface voids (Nosonovsky and Bhushan, 2012a). roughness. Particle size of matting agents is in the Nosonovsky and Bhushan, 2012 (based on work of range of 3-12 µm and this corresponds well with the Bhushan and Her, 2010) showed that at similar significant decrease in conidial adhesion in our contact angles, with different wetting modes, study when size of micro scratches went below 20 adhesion forces can be very different in relation to µm. In their study, roughness was achieved by

10 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

frequency and size of bumps that are influencing surface roughness. It is possible that this is occurring when conidial spores try to adhere to the alkyd surface. At certain range of roughness level combinations, spore adhesion becomes much more difficult leading to fewer spores attached on the coating surface.

Concerning the hydrophobic rodlet monolayer of hydrophobins that covers the conidial surface, the tendency of more conidial adhesion on decreasing contact angels makes sense. In the matter of lower contact angels the surface becomes more hydrophilic-like, which facilitate a hydrophobic/hydrophilic interaction between the surfaces of the coating and the conidia. Multiple parameters influence the adhesion of fungal spores (conidia) to the surface. Surface roughness and surface tension are two facets that may also be taken into account when formulating paints to avoid or reduce the addition of biocides.

11 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

References Catley, B.J., 1980. The extracellular polysaccharide, pullulan, produced by Aureobasidium Bailey, J.A., Jeger, M.J., others, 1992. pullulans: a relationship between Colletotrichum: biology, pathology and elaboration rate and morphology. Journal of control. Colletotrichum: biology, pathology General Microbiology 120, 265–268. and control. Dhoke, S.K., Khanna, A.S., Sinha, T.J.M., 2009. Bardage, S.L., 1996. Colonisation of painted wood Effect of nano-ZnO particles on the by Aureobasidium pullulans-analysis of corrosion behavior of alkyd-based features and consequences for failure in waterborne coatings. Progress in Organic service. Document-the International Coatings 64, 371–382. Research Group on Wood Preservation Dhoke, S.K., Sinha, T.M., Dutta, P., Khanna, A.S., (Sweden). 2008. Formulation and performance study Bardage, S.L., Bjurman, J., 1998. Isolation of an of low molecular weight, alkyd-based Aureobasidium pullulans polysaccharide waterborne anticorrosive coating on mild that promotes adhesion of blastospores to steel. Progress in Organic Coatings 62, water-borne paints. Canadian journal of 183–192. microbiology 44, 954–958. Gobakken, L.R., Høibø, O.A., Solheim, H., 2010. Bardage, S., Westin, M., Fogarty, H.A., Trey, S., Mould growth on paints with different 2014. The effect of natural product surface structures when applied on wooden treatment of southern yellow pine on fungi claddings exposed outdoors. International causing blue stain and mold. International Biodeterioration & Biodegradation 64, 339– Biodeterioration & Biodegradation 86, 54– 345. 59. Gobakken, L.R., Mattson, J., Alfredsen, G., 2008. Bayry, J., Aimanianda, V., Guijarro, J.I., Sunde, M., In-service performance of wood depends Latge, J.-P., 2012. Hydrophobins—unique upon the critical in-situ conditions. IRG/WP fungal proteins. PLoS pathogens 8, 08–20382. e1002700. Lazaridou, A., Roukas, T., Biliaderis, C.G., Bell-Pedersen, D., Dunlap, J.C., Loros, J.J., 1992. Vaikousi, H., 2002. Characterization of The Neurospora circadian clock-controlled pullulan produced from beet molasses by gene, ccg-2, is allelic to eas and encodes a Aureobasidium pullulans in a stirred tank fungal hydrophobin required for formation reactor under varying agitation. Enzyme of the conidial rodlet layer. Genes & and Microbial Technology 31, 122–132. development 6, 2382–2394. Linder, M.B., 2009. Hydrophobins: proteins that Bhushan, B., Her, E.K., 2010. Fabrication of self assemble at interfaces. Current Opinion superhydrophobic surfaces with high and in Colloid & Interface Science 14, 356–363. low adhesion inspired from rose petal. Linder, M.B., Szilvay, G.R., Nakari-Setälä, T., Langmuir 26, 8207–8217. Penttilä, M.E., 2005. Hydrophobins: the Braun, E.J., Howard, R.J., 1994. Adhesion of fungal protein-amphiphiles of filamentous fungi. spores and germlings to host plant surfaces, FEMS microbiology reviews 29, 877–896. in: The Protistan Cell Surface. Springer, pp. Mims, C.W., Liljebjelke, K.A., Richardson, E.A., 202–212. 1995. Surface morphology, wall structure, and initial adhesion of conidia of the

12 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

powdery mildew fungus Uncinuliella of the Aspergillus spore wall. Molecular australiana. Phytopathology 85, 352–358. microbiology 16, 33–44. Nosonovsky, M., Bhushan, B., 2012a. Green Sundstrom, P., 1999. Adhesins in Candida albicans. Tribology, its History, Challenges, and Current opinion in microbiology 2, 353– Perspectives, in: Nosonovsky, M., Bhushan, 357. B.(Eds.), Green Tribology, Green Energy Thau, N., Monod, M., Crestani, B., Rolland, C., and Technology. Springer Berlin Tronchin, G., Latge, J.-P., Paris, S., 1994. Heidelberg, pp. 3–22. rodletless mutants of Aspergillus fumigatus. Nosonovsky, M., Bhushan, B. (Eds.), 2012b. Green Infection and immunity 62, 4380–4388. Tribology, Green Energy and Technology. Tronchin, G., Bouchara, J.P., Ferron, M., Larcher, Springer Berlin Heidelberg, Berlin, G., Chabasse, D., 1995. Cell surface Heidelberg. properties of Aspergillus fumigatus conidia: Osherov, N., May, G.S., 2001. The molecular correlation between adherence, mechanisms of conidial germination. FEMS agglutination, and rearrangements of the Microbiology Letters 199, 153–160. cell wall. Canadian journal of microbiology Paris, S., Debeaupuis, J.-P., Crameri, R., Carey, M., 41, 714–721. Charlès, F., Prévost, M.C., Schmitt, C., Veigel, S., Grüll, G., Pinkl, S., Obersriebnig, M., Philippe, B., Latgé, J.P., 2003. Conidial Müller, U., Gindl-Altmutter, W., 2014. hydrophobins of Aspergillus fumigatus. Improving the mechanical resistance of Applied and Environmental Microbiology waterborne wood coatings by adding 69, 1581–1588. cellulose nanofibres. Reactive and Poaty, B., Vardanyan, V., Wilczak, L., Chauve, G., Functional Polymers 85, 214–220. Riedl, B., 2014. Modification of cellulose doi:10.1016/j.reactfunctpolym.2014.07.020 nanocrystals as reinforcement derivatives Wasylnka, J.A., Moore, M.M., 2000. Adhesion of for wood coatings. Progress in Organic Aspergillus species to extracellular matrix Coatings 77, 813–820. proteins: evidence for involvement of doi:10.1016/j.porgcoat.2014.01.009 negatively charged carbohydrates on the Sela-Buurlage, M.B., Epstein, L., Rodriguez, R.J., conidial surface. Infection and immunity 1991. Adhesion of ungerminated 68, 3377–3384. Colletotrichum musae conidia. Wenzel, R.N., 1936. Resistance of solid surfaces to Physiological and molecular plant wetting by water. Industrial & Engineering pathology 39, 345–352. Chemistry 28, 988–994. Stringer, M.A., Timberlake, W.E., 1995. dewA Wösten, H.A., 2001. Hydrophobins: multipurpose encodes a fungal hydrophobin component proteins. Annual Reviews in Microbiology 55, 625–646.

13 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

S1 supplementary materials

Specimen Sand Scratches Conidia Rating paper size Control Control 0 Conidia: Some, evenly dispersed, many conidia chains. 4 Glass Hyphae: Many hyphae, One – two forks, seems to be in a different layer than the spores 1.1 P100 80-100 µm Conidia: Spores in scratches, seemed to be more here than in even surface. Chains of 3 spores observed. Hyphae: Few germinations / growth from spores 1.2 P100 80-100 µm Conidia: Spores in scratches, seemed to be more here than in even surface. Chains of 3 spores observed. Hyphae: Few germinations / growth from spores 2.2 P100 80-100 µm Conidia: Many spores in and between scratches, spore chains. 3 Hyphae: few 3.1 P120 40-50 µm Glass broken N/A 3.2 P120 40-50 µm Conidia: Some, evenly dispersed, seemed to be less in scratches 3 Hyphae: None to very few observed 4.2 P120 40-50 µm Conidia: Some, evenly dispersed, seemed to be less in scratches 3 Hyphae: None observed 5.1 P150 15-20 µm Conidia: few spores 2 Hyphae: None observed 5.2 P150 15-20 µm Conidia: Few spores, single chains observed 2 Hyphae: None observed 6.2 P150 15-20 µm Conidia: Few in the scratches – but some in between the scratches-zones 2 Hyphae: None observed 7.1 P180 10 µm Conidia: Very few 2 Hyphae: None observed 7.2 P180 10 µm Conidia: Few spores 2 Hyphae: None observed 8.1 P180 10 µm Conidia: Few spores 2 Hyphae: None observed 2.1 P280 5-8 µm Conidia: Many spores in and between scratches, spore chains 3 Hyphae: None observed 4.1 P280 5-8 µm Conidia: Not many spores, but many hyphae in the sides. 1 Hyphae: None observed 9.1 P280 5-8 µm Conidia: Few spores, 1 Hyphae: None observed 6.1 P400 4-5 µm Conidia: Very few spores in and around the scratches, but many spores between 1 scratches-zones (doesn’t count in the rating because these area must be considered as “no scratches”) Hyphae: Some hyphae found

14 Stenbæk, Nikolic, Lindegaard, Löf, Jensen and Hastrup (2015) Ready for submission in a Scientific Journal (2015/7)

8.2 P400 4-5 µm Conidia: Very few spores 1 Hyphae: None observed 9.2 P400 4-5 µm Conidia: Very few spore chains 1 Hyphae: A single observed 10.1 No 0 Conidia: Two types of spores 4 Hyphae: None observed 10.2* No 0 Conidia: Some / many, evenly dispersed, many spores chains, often located in bobbles 4 Hyphae: Long linear hyphae with few forks,

Table S1: The results from every observed specimens. *) had fallen off during the exposure. The colour codes indicate the size of the topographical scratches. The rating is based on an over-all evaluation of the area in and around the scratches of every specimen.

15

Manuscript III

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Identification and classification of hydrophobin encoding gene sequences in two species of the blue stain fungus Aureobasidium pullulans

Jonas Stenbæk*, Bo Jensen & Leise Riber

Section of Microbiology, University of Copenhagen, Copenhagen, Denmark

Hydrophobins constitute a group of small cysteine-rich proteins that are expressed by filamentous fungi only. They are known for their ability to form hydrophobic-hydrophilic interfaces around the hyphae and spores during surface growth. Thereby, the fungus generates a protein membrane that keeps water close to its hyphae and spores, even in dry conditions. In this study, the full genome sequence of the black fungus Aureobasidium pullulans strain P268 was obtained by high-throughput sequencing on the Illumina MiSeq platform, and sequence reads were annotated with the prediction of open reading frames. Aureobasidium pullulans is known to attack and colonize anthropogenic and processed surfaces such as protective coatings and paints causing discoloration by blue stain. Gene sequences encoding hydrophobins were identified based on a comparative alignment to the annotated genome sequence of the previously characterized Aureobasidium pullulans strain MUCL38722. Two hydrophobin sequences were identified in both strains, which represent the first identified and described hydrophobins in a Blue Stain fungi. The results of this work will lead to future studies describing the phenotypical effects of hydrophobins during surface growth based on the construction and characterization of hydrophobin null-mutants of Aureobasidium pullulans strain P268.

* Corresponding: (JS) [email protected] May 2015

Introduction

Microorganisms are often covered by a fungi, the rodlet layer (Wessels et al., 1991). The proteinaceous surface layer that serves as a sieve for self-assembly properties and remarkable structural external molecular influx, a shield to protect and physicochemical characteristics of hydrophobin microorganisms from external influence or as an aid proteins are based on the multiple roles played by to help microbial dispersion to the environment. It these unique proteins in fungal biology. Although could, however, also constitute an important studies in hydrophobins have been conducted since function regarding physiology and metabolism. In the start of the 1990’s and more and more functions bacteria, the latter is called the S-layer, in have been associated to this small protein and more Actinomycetes, the rod-like fibrillar layer, and in work is needed to completely elucidate the role of

1

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

the hydrophobins (Aimanianda et al., 2009; Bayry variation is seen in the cysteine spacing of class I et al., 2012; Khalesi et al., 2013). hydrophobins, while less variation is seen for class II hydrophobins (Kershaw and Talbot, 1998). Hydrophobins are a family of small proteins found exclusively in filamentous fungi (Linder et al., Members of class I consist of 100-125 amino acids 2005; Wösten, 2001). The currently characterized in contrast to members of class II that usually only hydrophobins span approximately 100 amino acids span 50-100 amino acids (Hektor and Scholtmeijer, (AA) and display very little AA sequence homology 2005). The possibility to vary the spacing of the except from eight conserved cysteines organized in cysteine residues is also more likely in class I a characteristic motif (Wessels, 1997, 1994; Wösten hydrophobins. Two consensus cysteine spacing and Wessels, 1997). The eight cysteines form four patterns were noticed by comparative analyses of disulfide bonds in the pattern Cys1-Cys6, Cys2- know hydrophobins of class I and II. Cys5, Cys3-Cys4, Cys7-Cys8 (figure 1) and especially the Cys3-Cys4 loop can vary These are: considerably in length (Kwan et al., 2006). Class I: CN{5–7}CCN{19–39}CN{8–23}CN{5}CCN{6–18} CN{2–13}

Class II: CN{9–10}CCN{11}CN{16}CN{8–9}CCN{10} CN{6–7} Figure 1. The eight cysteines motif. The disulfide bonds between the cysteines are marked with dashed lines (Askolin, 2006). where C=cysteine residue and N=any amino acid Based on their distinct hydropathy patterns and except cysteine (Kershaw and Talbot, 1998). As physical properties, hydrophobins are traditionally more hydrophobins are discovered, these spacing divided into two classes (Wessels, 1994). Class I patterns seem to slightly change (Kershaw et al., hydrophobins forms highly insoluble membranes in 2005). Finally, class I hydrophobins are more water, organic solvents and 2% sodium dedocyl diverse than class II hydrophobins, as they have sulfate (SDS), whereas the membranes formed by been identified in fungi from two phyla groups; class II hydrophobins easily are dissolved in Ascomycetes and Basidiomycetes. So far, class II aqueous ethanol (60%) or 2% SDS (Wessels, 1997). hydrophobins have been discovered only in Typically, a single fungal species expresses either Ascomycetes. This observation has led to the class I or class II hydrophobins, however, previous parting of class I hydrophobins into class Ia studies have reported that few species have the (Ascomycetes) and class Ib (Basidiomycetes) ability to constitutively express both class I and (Lunkenbein et al., 2011). class II hydrophobins (Linder et al., 2005; Segers et Aureobasidium pullulans is a black fungus al., 1999). One single study even suggests that belonging to the Ascomycota division. It is an interclassical hydrophobins might occur in some ubiquitous mould that can be found in many Aspergillii species (Jensen et al., 2010). In class I different environments like air, soil and water. It is a hydrophobins the cysteine doublets are followed by naturally occurring epiphyte or endophyte of a wide hydrophilic amino acids, whereas hydrophobic range of plant species and it is well-known for the amino acids are observed after the cysteine doublets ability to colonize a wide range of these hosts in class II hydrophobins (Wessels, 1997; Wösten (Andrews et al., 1994; Webb and Mundt, 1978). and Wessels, 1997). Furthermore, considerable

2

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Aureobasidium pullulans has excellent properties to candidates. The annotated and accessible genome of survive under various environmental conditions strain MUCL38722 was applied in a comparative (Köhl and Fokkema, 1993). It is apparently most alignment study with the full-genome sequence of common in the temperate zones of the world with strain P268 to describe, for the first time, the numerous records from all Europe, North and South identification of two hydrophobin encoding genes Amerika, Asia, Mediterranean and Africa harbored by a blue stain fungus. (Deshpande et al., 1992; Domsch et al., 1980).

Aureobasidium pullulans colonize anthropogenic and processed surfaces like protective coatings and Materials and Methods paints (Bardage et al., 2014; Bardage, 1997; Zabel and Terracina, 1978) and latex surfaces (Zabel and Strains and genome sequences Horner, 1981). Attack of Aureobasidium pullulans For this study we used Aureobasidium pullulans (de may cause Blue Stain, the discoloration of sapwood Bary) Arnaud, strain P268 (ATCC 48434), and surfaces and/or the surface of protective coatings. Aureobasidium pullulans strain MUCL38722. The P268 strain is used for European standard test for wood- and coating trials, whereas strain MUCL38722 is a common laboratory strain.

The genomic data for Aureobasidium pullulans P268 has been obtained by high-throughput Illumina MiSeq sequencing at The Danish National High-Throughput DNA-Sequencing Centre (Copenhagen, Denmark), and raw reads were filtered, trimmed and quality checked before

Figure 2. Blue Stain on posts. Aureobasidium pullulans. Foto: assembly into contigs at the Section of http://www.sustainablelumberco.com. Microbiology, University of Copenhagen.

Strain MUCL38722 was previously genome sequenced and annotated to predicted proteins Like all other filamentous fungi, Aureobasidium (nucleotides and amino acids, respectively). The pullulans are expected to contain genomic genomic sequence data were produced by the sequences encoding for hydrophobins. Nevertheless, Genozymes for Bioproducts and Bioprocesses the number of expected hydrophobin genes to Development Project (Website: identify remains un-known. Previous studies http://www.fungalgenomics.ca). indicate that different fungi harbor a different number of sequences related to this protein (Jensen et al., 2010). As a recent study by Gostin et al. reports a significant general genomic variation within the species of Aureobasidium pullulans (Gostin et al., 2014), two different strains of Aureobasidium pullulans, P268 and MUCL38722, were used in this study in order to obtain a reliable indication of potential hydrophobin encoding gene

3

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Hydrophobin sequences in Aureobasidium Protein caracterization pullulans Kite Doolitle (Kyte and Doolittle, 1982) plot A script was composed on the basis of Kershaw and diagrams were generated with online tools from Talbot’s definition (Kershaw and Talbot, 1998) to Swiss Prot (Gasteiger et al., others, 2005) in order aim the search of the 8-cysteine pattern. In order to to reveal whether the identified hydrophobins increase the probability of identification of both belonged to Class I or Class II based on the location classes of hydrophobins, as well as of hydrophobins of hydrophobic and hydrophilic amino acids expanding the known pattern, the class definitions (Wessels, 1997). were merged and the number of amino acids between the cysteines was further expanded with To scan the full genome of strain P268, all ORFs approximately 10% to an expanded uni-class were identified by using the free mEMBOSS hydrophobin definition: software TRANSEQ (Version: EMBOSS:6.6.0.0). All six reading frames were analyzed through the CN{3–14}CCN{7–45}CN{5–27}CN{5-14}CCN{5–25} C advanced search in GrepWin with the same script as used with the MUCL38722 strain (see above). This

analysis resulted in the identification of 43 potential The script for hydrophobin search was made: hydrophobin candidates, however, the results did C[^C]{4,14}[\n|\r]*C[\n|\r]*C[\n|\r]*[^C]{7,20}[\n|\ not account for quality parameters of the ORFs, r]*C[\n|\r]*[^C]{5,20}[\n|\r]*C[\n|\r]*[^C]{5,14}[\n such as promoter regions, TATAA box sequences, |\r]*C[\n|\r]*C[\n|\r]*[^C]{5,15}[\n|\r]*C STOP codons, and with reference to hydrophobins, the size of the predicted protein, which was between 90 and 120 amino acids (Wessels, 1994). Therefore, By using this script in the free software GrepWin each candidate sequence was manually evaluated. (version 1.6.10) as advanced search tool in the The best candidates to meet the criteria from each genome with predicted proteins (amino acids) of the search were chosen for a pairwise alignment using Aureobasidium pullulans strain MUCL38722, two EMBOSS Needle (“EMBOSS Needle < Pairwise hydrophobin encoding gene candidates were Sequence Alignment < EMBL-EBI,” n.d.) and revealed, which fulfilled the expanded uni-class candidates displaying a short sequence similarity hydrophobin definition. These were named Aur1 were also used for EMBOSS Water (“EMBOSS and Aur2. Water < Pairwise Sequence Alignment < EMBL- CLC Genomic Workbench (version 6.5.1) was used EBI,”). EMBOSS Needle is used for global to BLAST the amino acid sequences of the alignments and applies the Neddleman-Wunsch identified hydrophobin candidates from strain algorithm, whereas EMBOSS Water is used for MUCL38722 against the assembled full genome of local alignments and applies the Smith-Waterman strain P268. Although open reading frames (ORFs) algorithm. were not found in strain P268, the CLC software Hydrophobin candidate genes that passed the above corrected for this and included all six reading stated criteria were aligned against Class II frames and did - to some extent - take possible hydrophobins from other ascomycetes (appendix 1) intron sequences into account. chosen from National Center for Biotechnology Information (NCBI) using Clustal Omega (McWilliam et al., 2013), and different trees were

4

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

made using Neighbor-Joining (NL) and Maximum Likelihood (ML) principles (Tateno et al., 1994).

Finally, an automatic data processing was provided in the The Metagenomics RAST (MG-RAST) pipeline (Glass et al., 2010), which is a server that generates automated analysis platforms for metagenomes providing quantitative insights into

microbial populations based on sequence data. The server primarily provides upload, quality control, automated annotation and analysis for prokaryotic metagenomic shotgun samples.

Figure 3. Hydrophobins from Aureobasidium pullulans strain MUCL (aur1 and aur2) and strain P268 (hfbA and hfbB). Clustal Omega alignment; (A) aur1 vs. hfbA and (B) aur2 vs. hfbB. The color codes: Green = identical amino acid (except cysteine), yellow = cysteine.

5

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Results and Discussion Classification and characterization of identified hydrophobins

Kite Doolitle plot diagrams within the 8-Cysteine Identification of gene sequences likely to encode pattern of the hydrophobin candidate molecules hydrophobins in Aureobasidium pullulans strain show us that the hydrophobin genes hfbA and hfbB P268 could most likely belong to Class II (figure 4) due to Two gene sequences, likely to encode its hydrophobic amino acids after the cysteine hydrophobins, were located by the script in the doublets (Wessels, 1997). An additional Kite reference Aureobasidium pullulans strain Doolitle plot diagram for the complete hydrophobin MUCL38722. These two hydrophobins were named sequences is shown in appendix 4. Aur1 and Aur2.

By using a BLAST analysis of Aur1 and Aur2 against the genome sequence of Aureobasidium pullulans strain P268 in the CLC Genomic Workbench, two similar hydrophobins in

Aureobasidium pullulans strain P268 were identified. These two hydrophobins were named Figure 4: Kite Doolitle diagrams of hfbA and hfbB. Only within the 8 Cysteine pattern. Blue line indicates a Cysteine. The hydrophobic part hfbA and hfbB. A local Clustal Omega alignment in the middle of the sequences indicates a Class II hydrophobin. between aur1/hfbA and aur2/hfbB are presented in Detailed figures are placed in appendix 4. figure 3. The aberrant parts in the two alignments A neighbor joining tree (figure 5) of known are presumably caused by the presence of introns in hydrophobins originating from other filamentous the sequence of Aureobasidium pullulans strain fungi (NCBI database) as well as of the P268. An alignment in which all possible introns hydrophobins identified for both Aureobasidium have been removed is included in appendix 2. pullulans strains (MUCL38722 and P268, respectively) indicates, not surprisingly, a close The local positions of hfbA and hfbB in the relation between aur1 (Aureobasidium A) and hfbA sequence of Aureobasidium pullulans strain P268 (Aureobasidium1). Interestingly, the tree further are shown in Appendix 3 and can subsequently be indicates that hfbB (Aureobasidium2) seems closely applied for identifying up- and downstream regions related to hydrophobins of Trichoderma ssp. for primer design used for further studies involving (Trichoderma, Trichoderma2, Trichoderma3). As the construction of hydrophobin knock-out mutant Trichoderma and Aureobasidium represent closely strains. related fungal species, it is not surprising that their The identified hydrophobin encoding genes, hfbA hydrophobin genes are highly alike. The amino acid and hfbB, were submitted for a BLAST analysis in sequences of all hydrophobins displayed in the tree the NCBI database and both showed a significant are presented in appendix 1. A Maximum likelihood query coverage (>90%) with other hydrophobins tree was also made and revealed the same tendency originating from filamentous fungi. They were all (results not shown). placed to be included in the Hydrophobin Superfamily.

6

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

100 AureobasidiumA The positions of β-barrel and loop in HfbA and Aureobasidium1 HfbB are comparable with HFBII and could Aureobasidium2 77 96 Trichoderma2 indicate a proteomic similarity. Trichoderma 81 Trichoderma3 Finally, 43 potential hydrophobin candidates 95 Polyporus 88 Grifolafrondosa|133907722| (including the hfbA and hfbB genes) were identified 60 Laccariabi Nomuraea in the annotated sequence of Aureobasidium 100 Aspergillus pullulans strain P268 by using the script in

0.1 GrebWin. Every candidate gene was subsequently analyzed for promotor regions, START and STOP Figure 5: Neighbor-Joining Tree. hfbA (Aureobasidium1), aur1 (AureobasidiumA), hfbB (Aureobasidium2). codons as well as for the presence of the 8-cysteines pattern between the START and STOP codons. Two All the above stated indications led to the candidate genes passed the above stated conclusion that hfbA and hfbB represent gene requirements and turned out to be similar to hfbA sequences coding for hydrophobins in and hfbB that we previously found by BLAST in Aureobasidium pullulans strain P268. The first ever CLC. No other candidate genes passed the criteria described and identified hydrophobins in a blue and were therefore discarded. stain fungus. Molecular weight and pI values are shown in appendix 1. Full amino acid and DNA sequences (incl. introns) of HfbA and HfbB are found in appendix 1 and 2. Protein structures of HfbA and HfbB molecules are made and predicted by “most-likely-principle” in Conclusion Phyre2 (figure 6) to compare the location of β-barrel and loop with the class II hydrophobin HFBII, To summarize, two hydrophobin encoding gene described by Kallio et al. (Kallio et al., 2007) . sequences were identified in Aureobasidium pullulans strain P268 and strain MUCL38722. The A genomic positions of the annotated contigs were located. It is likely that more hydrophobin genes could be present in the fungal genome, but this seems doubtful because of good sequencing quality, so if so, they would thereby have to be evasive from common genomic definitions of hydrophobins.

B

Acknowledgements

We would kindly like to thank the Danish Innovation Foundation and collaborators in the project. Figure 6: 3D protein structure of (A) HfbA (010869) and HfbB (001745). Made in Phyre2 (Kelley et al., 2015). β-barrel (yellow) and loop (pink) (B) The Class II HFBII with β-barrel and loop made by Kallio et al. 2007.

7

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

References EMBOSS Water < Pairwise Sequence Alignment < EMBL-EBI [WWW Document], n.d. URL http://www.ebi.ac.uk/Tools/psa/emboss_wa ter/ (accessed 7.2.15). Aimanianda, V., Bayry, J., Bozza, S., Kniemeyer, O., Perruccio, K., Elluru, S.R., Clavaud, C., Gasteiger, E., Hoogland, C., Gattiker, A., Wilkins, Paris, S., Brakhage, A.A., Kaveri, S.V., M.R., Appel, R.D., Bairoch, A., others, 2005. Protein identification and analysis others, 2009. Surface hydrophobin prevents tools on the ExPASy server. Springer. immune recognition of airborne fungal spores. Nature 460, 1117–1121. Glass, E.M., Wilkening, J., Wilke, A., Andrews, J.H., Harris, R.F., Spear, R.N., Lau, Antonopoulos, D., Meyer, F., 2010. Using G.W., Nordheim, E.V., 1994. the metagenomics RAST server (MG- RAST) for analyzing shotgun Morphogenesis and adhesion of metagenomes. Cold Spring Harbor Aureobasidium pullulans. Canadian Journal of Microbiology 40, 6–17. Protocols 2010, pdb–prot5368. Askolin, S., 2006. Characterization of the Gostin, C., Ohm, R.A., Kogej, T., Sonjak, S., Turk, Trichoderma reesei hydrophobins HFBI and M., Zajc, J., Zalar, P., Grube, M., Sun, H., Han, J., others, 2014. Genome sequencing HFBII. VTT Technical Research Centre of of four Aureobasidium pullulans varieties: Finland. Bardage, S.L., 1997. Colonization of painted wood biotechnological potential, stress tolerance, by blue stain fungi. and description of new species. BMC genomics 15, 549. Bardage, S., Westin, M., Fogarty, H.A., Trey, S., Hektor, H.J., Scholtmeijer, K., 2005. Hydrophobins: 2014. The effect of natural product treatment of southern yellow pine on fungi proteins with potential. Current Opinion in causing blue stain and mold. International Biotechnology 16, 434–439. Biodeterioration & Biodegradation 86, 54– Jensen, B.G., Andersen, M.R., Pedersen, M.H., Frisvad, J.C., Søndergaard, I., 2010. 59. Hydrophobins from Aspergillus species Bayry, J., Aimanianda, V., Guijarro, J.I., Sunde, M., Latge, J.-P., 2012. Hydrophobins—unique cannot be clearly divided into two classes. fungal proteins. PLoS pathogens 8, BMC Research Notes 3, 344. Kallio, J.M., Linder, M.B., Rouvinen, J., 2007. e1002700. Crystal structures of hydrophobin HFBII in Deshpande, M.S., Rale, V.B., Lynch, J.M., 1992. Aureobasidium pullulans in applied the presence of detergent implicate the microbiology: a status report. Enzyme and formation of fibrils and monolayer films. Microbial Technology 14, 514–527. Journal of Biological Chemistry 282, 28733–28739. Domsch, K.H., Gams, W., Anderson, T.-H., others, Kelley, L.A., Mezulis, S., Yates, C.M., Wass, M.N., 1980. Compendium of soil fungi. Soc General Microbiol. Sternberg, M.J., 2015. The Phyre2 web EMBOSS Needle < Pairwise Sequence Alignment < portal for protein modeling, prediction and analysis. Nature protocols 10, 845–858. EMBL-EBI [WWW Document], n.d. URL Kershaw, M.J., Talbot, N.J., 1998. Hydrophobins http://www.ebi.ac.uk/Tools/psa/emboss_nee dle/ (accessed 7.2.15). and repellents: proteins with fundamental

8

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

roles in fungal morphogenesis. Fungal the EMBL-EBI. Nucleic acids research 41, Genetics and Biology 23, 18–33. W597–W600. Kershaw, M.J., Thornton, C.R., Wakley, G.E., MG-RAST - Home [WWW Document], n.d. URL Talbot, N.J., 2005. Four conserved https://metagenomics.anl.gov/ (accessed intramolecular disulphide linkages are 7.2.15). required for secretion and cell wall Segers, G.C., Hamada, W., Oliver, R.P., Spanu, localization of a hydrophobin during fungal P.D., 1999. Isolation and characterisation of morphogenesis. Molecular microbiology five different hydrophobin-encoding 56, 117–125. cDNAs from the fungal tomato pathogen Khalesi, M., Venken, T., Deckers, S., Winterburn, Cladosporium fulvum. Molecular and J., Shokribousjein, Z., Gebruers, K., General Genetics MGG 261, 644–652. Verachtert, H., Delcour, J., Martin, P., Tateno, Y., Takezaki, N., Nei, M., 1994. Relative Derdelinckx, G., 2013. A novel method for efficiencies of the maximum-likelihood, hydrophobin extraction using CO 2 foam neighbor-joining, and maximum-parsimony fractionation system. Industrial Crops and methods when substitution rate varies with Products 43, 372–377. site. Molecular Biology and Evolution 11, Köhl, J., Fokkema, N.J., 1993. Fungal interactions 261–277. on living and necrotic leaves. Ecology of Webb, T.A., Mundt, J.O., 1978. Molds on plant pathogens 321–334. vegetables at the time of harvest. Applied Kwan, A.H.Y., Winefield, R.D., Sunde, M., and environmental microbiology 35, 655– Matthews, J.M., Haverkamp, R.G., 658. Templeton, M.D., Mackay, J.P., 2006. Wessels, J.G., 1997. Hydrophobins: proteins that Structural basis for rodlet assembly in change the nature of the fungal surface. fungal hydrophobins. Proceedings of the Advances in microbial physiology 38, 1–45. National Academy of Sciences of the Wessels, J.G., De Vries, O.M., Asgeirsdottir, S.A., United States of America 103, 3621–3626. Schuren, F.H., 1991. Hydrophobin genes Kyte, J., Doolittle, R.F., 1982. A simple method for involved in formation of aerial hyphae and displaying the hydropathic character of a fruit bodies in Schizophyllum. The Plant protein. Journal of molecular biology 157, Cell 3, 793–799. 105–132. Wessels, J.G.H., 1994. Developmental regulation of Linder, M.B., Szilvay, G.R., Nakari-Setälä, T., fungal cell wall formation. Annual review Penttilä, M.E., 2005. Hydrophobins: the of phytopathology 32, 413–437. protein-amphiphiles of filamentous fungi. Wösten, H.A., 2001. Hydrophobins: multipurpose FEMS microbiology reviews 29, 877–896. proteins. Annual Reviews in Microbiology Lunkenbein, S., Takenberg, M., Nimtz, M., Berger, 55, 625–646. R.G., 2011. Characterization of a Wösten, H.A., Wessels, J.G., 1997. Hydrophobins, hydrophobin of the ascomycete from molecular structure to multiple Paecilomyces farinosus. Journal of basic functions in fungal development. microbiology 51, 404–414. Mycoscience 38, 363–374. McWilliam, H., Li, W., Uludag, M., Squizzato, S., Zabel, R.A., Horner, W.E., 1981. An accelrated Park, Y.M., Buso, N., Cowley, A.P., Lopez, laboratory procedure for growing R., 2013. Analysis tool web services from Aureobasidium pullulans on fresh latex

9

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

paint films. Journal of coatings technology paint mildew caused by< i> Aureobasidium 53, 33–37. pullulans. Journal of coatings technology Zabel, R.A., Terracina, F., 1978. Nutrition of 50, 43–47. saprobic fungi and control strategies for

10

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Appendix 1

Hydrophobin candidate genes that passed stated criteria were aligned against Class II hydrophobins from other ascomycetes. Data from NCBI.

>Species – gi/gb# – protein – [species]

AA sequence

>Trichodermareesei gi|572280373|gb|ETS03470.1| hydrophobin [Trichoderma reesei RUT C-30]

MKFFAIAALFAAAAVAQPLEDRSNGNGNVCPPGLFSNPQCCATQVLGLIGLDCKVPSQNVYDGTDFRNVC

AKTGAQPLCCVAPVAGQALLCQTAVGA

>Grifolafrondosa gi|133907722|gb|ABO42329.1| hydrophobin [Grifola frondosa]

MFSKLAIFATAAFAVLAAATPVRRQQCTTGQLQCCESTSTANDPATSELLGLIGVVISDVDALVGLTCSP

ISVIGVGSGSACTANPVCCDSSPIGGLVSIGCVPVNV

>Polyporusumbellatus gi|441032734|gb|AGC26950.1| hydrophobin [Polyporus umbellatus]

MFSRAIVVTALTLPLLAAATPVELEARQSCSTGSIQCCNTVEDAKSASASLLLGLLGIVVGDITGLIGLN

CSPLNVVGVGSGNACSANAVCCQNNNVGGLISIGCVPVIL

>Nomuraearileyi gi|375273598|gb|AFA43698.1| hydrophobin [Nomuraea rileyi]

MAFFKVLVAAATLATALALPSGGSGNGNGNGHGHSVGDAAAQCGNHQQLSCCNKGNSAGTLLDGLLGGNC

SPLDLSIIGVGVPLSTACSNQVACCTGDQNGLLNLACTNLNL

>Laccariabicolor S238N-H82 gi|164635101|gb|EDQ99414.1| hydrophobin [Laccaria bicolor S238N-H82]

MFSKVLLVAATLVTFVAATPVPGGVDNSCNTGTLQCCNQTFSSTSGTATLLAALLNLNLSQLTGQIGLSC

TPISVIGLGQGASCTQQPVCCSGNTYNGLINVGCSPINL

11

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

>TrichodermareeseiQM6a gi|340521353|gb|EGR51587.1| hydrophobin [Trichoderma reesei QM6a]

MQFFAVALFATSALAAVCPTGLFSNPLCCATNVLDLIGVDCKTPTIAVDTGAIFQAHCASKGSKPLCCVA

PVADQALLCQKAIGTF

>TrichodermareeseiQM6a2 gi|340514533|gb|EGR44794.1| hydrophobin [Trichoderma reesei QM6a]

MQFLAVAALLFTTALAAPSSDVNGIIRRANAFCPEGLLYTNPLCCDLDVLGVADVDCVVPPAKPSSCKSF

GSVCASIGRKPRCCAVPVAGVALLCTDPIPAI

>Aspergillusclavatus gi|3046650|gb|AAC13521.1| hydrophobin, partial [Aspergillus clavatus]

KGNAXVXFPVXEEMTIKQGTEKCGDQAQLSCCNKATYAGDSTDIDSGILSGTLKNLIGGGSGHQGLGLFD

QCSKLDLQIPIIGIPIQDLINQQCKQNIACCQNPSLA

>Aureobasidiumpullulansaur1 MUCL38722 aur1

MKVFASILAIAAIASAAATPDVKRRSAVLASRDASDLCGPLDTPMCCGTDVLGLADLSCSSVPSDVTTTDDFTAYCGAEGLSSHCCVTSLLGSLGVACAAA

>Aureobasidiumpullulansaur2 MUCL38722 aur2

MLAATIITAFIGSSIVAAAPADVSARQLTICSGTYSNAQCCATDVLGLADLNCANPPTTPTSQEDFIDICATEGQQARCCALPILGQALLCGSPLG

>Aureobasidiumpullulans P268

MKVFASILAIATIAAAAATPNQKRTGTLEMRQSTSLCGPLDTPMCCQTDVLGVADLSCTAGMFPPSLHLA*YQ*Y*CVVFKVPNTVTTDANFTSYCAAEGKTAECCVTQLVSQL

12

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Appendix 2

AA alignment of the Aur1/HfbA and Aur2/HfbB without any introns

+ pI and molecular weights

Amino acid sequences for HfbA, HfbB, Aur1 and Aur2.

Aur1: MKVFASILAIAAIASAAATPDVKRRSAVLASRDASDLCGPLDTPMCCGTDVLGLADLSCSSVPSDVTTTDDFTAYCGAEGLSSHCCVTSLLGSLGVACAAA HfbA: MKVFASILAIATIAAAAATPNQKRT-GTLEMRQSTSLCGPLDTPMCCQTDVLGVADLSCTAVPNTVTTDANFTSYCAAEGKTAECCVTQLVSQLGWCLQS FGENESH: MKVFASILAIATIAAAAATPNQKRT-GTLEMRQSTSLCGPLDTPMCCQTDVLGVADLSCTAVPNTVTTDANFTSYCAAEGKTAECCVTQLLDSVGLLCSAAA Contig 9: pI: 4,66. Molecular weight: 10,282 kDa. Signal peptid = yes FGENESH: pI: 4,46. Molecular weight: 10,281 kDa. Signal peptid = yes

Aur2: MLAATIITAFIGSSIVAAAPADVSARQLTICSGTYSNAQCCATDVLGLADLNCANPPTTPTSQEDFIDICATEGQQARCCALPILGQALLCGSPLG HfbB: MLAATIITALIGSSIVAAAPADVSARQVTICAGTYSNAQCCATDVLGLADLNCANRESYTTNQTDFISICSAEGQQARCCALPILGQALLCGSPL FGENESH: MLAATIITALIGSSIVAAAPADVSARQVTICAGTYSNAQCCATDVLGLADLNCANPPTVPVNQTDFISICSAEGQQARCCALPILGQALLCGSPL Contig 20: pI: 4,23. Molecular weight: 9,641 kDa. Signal peptid = yes FGENESH: pI: 4,04. Molecular weight: 9,494 kDa. Signal peptid = yes

AA + DNA of the Aur1/HfbA and Aur2/HfbB

DNA sequences for HfbA and HfbB incl introns (underlined) and AA sequence for Aur1, Aur2, HfbA and HFBB:

13

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Appendix 3

Hfb1 and Hfb2 with up- and downstreams

Two identified hydrophobins in the new sequence Aureobasidium pullulans P268.

#1: Identified via aur 1 from ATCC 62921. In CLC Main Workbench.

The sequence, HfbA:

Nucleotides

Where in P268 sequence: contig 9.

Upstream and downstream Black: The P268 HfbA nt sequence

GAGCCGATGTGGCCAGACGACACTGAGGATCATCTTCATGAACAAGTCCTATCTTGTGGTCCTCATCGCCCCAGTATCGATGGACTGGTCTCTGATCGGTATTGGCCCAGGACGCGGGTGTTACCGTTGACGTCAAAGTCTCTCTCAACATCGTATGCAAATTCGCAAAGGAACTTCCCTCTGCTCATCGAAGGCATTCTGATGGTT GACCGAGGTGTACGGTTGTGTTCTGGAGGAGGACATACACCAAGCCGAGGCAGCAGTGGTGGAAGTGAATCGCGGCAGCATTCCGTCAGCGAGATCAGCGAGATCAGCGAGATCAGCGCTGTGCGAAATACTTCTGTTGACATCACTCAAATCATGGTGCATGATGGCAACAGCAATTGAGTTGCCAACACCACGTGTTAGCAAAGC TGCATATTTAGGGTTCCCTCGGTTTTCCCGGCAGACGGAGTGACTCACCGGCTGCTATAATGTTGCGCAATGTGCATCACGTTTCGCTGTCCAGATCGCAATTCCATATCCAAAATCATCATCCTCATCTGCCCGCCATCACCAACAACGTCATGCGACAAGGTATAACACAGCGGGAGATTCCAGCGGGTTGACCCGGGAGCGCAC CAGCGTGCCGGATCTGGGCGGCTTGACTTGCCCTAACTCCCCATGTCTCCTGTCTACACGGTAACTCTCGTGCGTGATTATTTTGTATCCCGAGATCCACCTATGTGTCAAAGACAGGCATATGCGCGTAGATAAGCCTGATCGCTCTATCAGAGCCAGATCTAGCAAGGATCTTTGCTTCAGGAGGCCGGAGGTTGAAGATCACCT AACTTAATTTTCAATGGTCTCGGTTCGCTCATCATCCGTTTGGGAAAGAGCTCGTATAAGACACGAACAGTCAAGTACAACCGAAAAGACAGGATCTTTGAGAAACACCGATGATGATCGTAGTGTCAATCCTTGGTGAGGATCAATGTCTCCGCGTATCTATGGACTTGCAATTGACTTCTGATGTGTCGATCTCGATCCGTTGCG TCTCTGATTTGCATACATGCAGCTTCAACTTACGTTTCTTGCGCCGGTGAGCATATCGCAGGACGTGCTCTACTAGGTATACCGGAGGACAGTGACGGTTCAAGCACATCGAGGTGAGATCGACATAGATAGGATGACTCTATGGCCATTCGACCGAGGTCCATGTTGACGACGGTATTAGAGGAAAAGGTATAAGTAGGAGGTGTT CAGTCGCATTGAAGACTCGACAGAACAATCAGAACAATCACAAGCATCTCTCCTCAGCTTTACATCAACAAACTTCCATCTCTTTCTTTTTCAAGAACATCAAATCATCCACCGTCAACATGAAGGTCTTCGCCAGCATCCTCGCTATCGCAACCATC GCCGCTGCTGCTGCAACTCCCAACCAGAAGCGCACTGGCACCCTCGAGATGCGCCAGTCAACCAGCCTCTGCGGCCCCTTGGACACTCCCAT GTGCTGTCAGACCGACGTCCTTGGAGTTGCTGACCTGTCATGCACTGCTGGTATGTTTCCTCCCTCTCTCCACTTGGCTTAGTACCAGTGAT ACTGATGTGTGGTTTTCAAAGTCCCCAACACCGTCACCACCGACGCCAACTTCACCTCCTACTGCGCTGCCGAGGGCAAGACTGCTGAGTGC TGTGTTACCCAGCTGGTAAGTCAACTCATGGGAATGTCTCAAGATTTCGATATTGACTCTGCTTTGCACAGCTCGACAGTGTCGGTCTCCTCTGCTCTGCTGCCGCATAAACGATGGCACTCGACACGA CAACATGACTTCTACGCTCTTTTTGGAATTATTAGTACTATATATAGACTTTAGTATCGACATTCAACATTGTCTTTTTCTATAACATTGTTTCCAGTTGACCATCTTATATTGCTCTCTACAGACTCTGCAGTGGCAATAAAGTGACAAGTCTGCTTGTGTGTAT TGTCTCTGAAAAGTGAAGTCGTCCTCACATCTGCATGGAAGTCTGAACCTGCGAACGACCTGGGCTGGTGCCTCCAAAGTTGATCGCGTTGACTTCACGCTGATGTCAACATAGATGACAGCAACGATGAGAAAGCAACGATAGGCATTGTTGCCAACGGCACGCC AAGTGTGGCACAGTAGGATGAGATAGTTGCTGCCTAGCTGTAAGCATATATAGCTTGATGCGATTACTTCCTCATCGCCATCATCGTTCAGCAGAAAGCGCAATCTCATCCACCATGACAGAGCATGAATCCACCCAGTCTTTACACTGCTCGTTCAGAGTCCCAT CACGTCACACAACAACAGTACAAGTCACTCTGGAGCCAAACAAGCATCCAGCAAAAGTGGGAAGCTGCCTGATCTTCCCAGAACAACCTATCGACGACTTTGTAGACTTTCCAGTCTGCAATGCTAAGATCAAAGCAACCGAGGACCGTGGTTACGCAGCGATATA CGGATGGATACAAATGGTCCGCGAGGCTCCTTTGAGCTCGCAGTCGACGACGACCCAGACCTCGGATTTTTGGGAAATGGATCCCATTCCTATGACAGCCGATCTTGAAAATCCTTTCATCTTGTTCGGTCCAGAAGCGCAACTATTTGATGCTCCATTCAGATCT AACAGAACGGATATGGATTGGACCTGCTGGAGCTTCTTGACATACATCAAAGACTCTCTGATGTCAAAGTCTGTACGTCCAATTCTGGTCATCGAGTGGGGGTTCCAGATCGACGCAGGCAATGTCACCATCAAGACATTGAGACTGATCGATGTTCAAGAAGCTT GGGAGCAGCAGCGAGAGATGCTAGAACAGAAGTTCGGTAGCTGGTCTTTCGAGCCAGCAGACGAGCATCTCGAAGTCAGTCAAATATCAGGACAATCAGATTAGGCATACTAGCTGTATGAGTCAAGAGCATTGCTTTGAAGCTTCAGGAACCGCATTTGAAGATT GGCAACCGGCTATTGGCGCTTGGAAGCGTTGCTTGTGATACGAGCCTTTGTAGCATTGCGCAAACAGGACAGATGCAGTAAGCATAGGTGGTACTTACAATCAAGAAGTACAAAATTGCAGGCTGAAGGATGATAATGCCAAGAAAGGGTCACGGATCATATCTGT ATGGAGCACAAGGTTGTTTATCCTTCCTCAAAGCAGCACTCAGATTATGGTATCTTATGTACAGTCGTGACGCATTGATATTGATATACTACACACAAAACAGAGCTGCACAAATCTGAATTGTCAGAACACCTGCACCTTTTCGATACACACCACGCGATCTCGT ATCCTTTAAGCAGCATAAGGAGATTTGCAGCTGAAGGAGGTAGTGACACTACCACCCTTGA

#2: Identified via aur 2 from ATCC 62921. In CLC Main Workbench.

The sequence, HfbB:

Nucleotides

Where in P268 sequence: Contig 20 (nt: 305155-305550)

Upstream and downstream Black: The P268 HfbB nt sequence

TCTCAAATCCTTGTGGCTGCTGGTCATGATCAATGATTGACATCGACGCAAGCAATGACGCTCGTTCAACCAAGACACCGTGAGATTGAATCATGTTGAGCGAAGGGAGACAAGGTATAAAAGAGCATACATTTCGTCCAGAGACTCTGCTTGGAGCCATAGAGAC TTCATCCATTCAATCTTCCATCCAAATTTCTCACAACTCCCAAACACATCAATTACCTTCCAAGATGCTCGCCGCAACCATCATCACCGCCCTCATTGGCAGCTCCATCGTCGCCGCTGC CCCTGCTGATGTATCAGCCCGTCAAGTCACCATCTGCGCCGGCACGTACAGCAATGCCCAATGTTGCGCTACCGACGTCCTTGGTCTTGCCG ACTTGAACTGCGCAAACCCCTGTTAACCAAACCGACTTCATCAGCATCTGCTCGGCTGAGGGCCAGCAAGCTCGTTGCTGCGCATTGCCTAT CTTGGCCAAGCTCTGCTGTGCGGCTCGCCTCTGTAAACGTCTCTGACGAAAGGATGCAACGTCTCTGACGAAAGGATGCACCTCCTTGAGACCGATCAAGACGAGGCAGGACCTCTTCTGTAGCCGTCAGGGAAGTGTT CTTGGGCCTGTAATCATACTCTTTACAAAAAGTTCGCAGAAGATCCCAAACATTTGTAATTGAGCTGAGCCGTGTTATGTTGACTTTGTTGTTTTTATCTGATTTGCCTTTCAAACCCCTGCCACAAAGAGTATTAATCAACATGTCACTTTGCTCCACTTTACGC TCTAGCATCGCAGTTACTCTGAGTTGAGTGTTTGCGTGGTAGAATCAATGTTTCAATGCCTCACGTTTTCTTCGTTTGTTCTTCCTCGTGGCTAGGG

14

Stenbæk, Jensen & Riber (2015) Manuscript for a Scientific Journal 2015/5

Appendix 4

Kyle Doolitle diagrams of HfbA and HfbB

Kyte-Doolittle:

Hydrophobicity; AA sequence:

HfbA: HfbB:

Hydrophobicity; AA sequence ONLY between first and last Cysteine:

Between 1-8 cysteine:

15

Manuscript IV

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

IRG/WP 15-10837

THE INTERNATIONAL RESEARCH GROUP ON WOOD PROTECTION

Section 1 Biology

The role of hydrophobins in surface growth by the Blue Stain fungi Aureobasidium pullulans

Jonas Stenbæk1, Leise Riber1, Jakob Blæsbjerg Nielsen2, Charlotte Møller Hansen1 & Bo Jensen1

1 Section for Microbiology Department of Biology University of Copenhagen Universitetsparken 15, Building 1 2100 Copenhagen, Denmark

2 Center for Microbial Biotechnology Department of Systems Biology Technical University of Denmark Søltofts Plads 2800 Kgs. Lyngby, Denmark

Paper prepared for the 46th IRG Annual Meeting Viña del Mar, Chile 10-14 May 2015

Disclaimer The opinions expressed in this document are those of the author(s) and are not necessarily the opinions or policy of the IRG Organization.

IRG SECRETARIAT Box 5609 SE-114 86 Stockholm Sweden www.irg-wp.com Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

The role of hydrophobins in surface growth by the Blue Stain fungi Aureobasidium pullulans

Jonas Stenbæk1, Leise Riber1, Jakob Blæsbjerg Nielsen2, Charlotte Møller Hansen1 & Bo Jensen1

1 Section for Microbiology, Department of Biology, University of Copenhagen, Universitetsparken 15, Building 1, 2100 Copenhagen, Denmark

2 Center for Microbial Biotechnology, Department of Systems Biology, Technical University of Denmark, Søltofts Plads, 2800, Kgs. Lyngby, Denmark

Abstract

Hydrophobins are small proteins found exclusively in filamentous fungi. These proteins are able to assemble spontaneously into amphiphilic monolayers at hydrophobic–hydrophilic interfaces. Hydrophobins have a diverse role, e.g. allowing the fungi to break through interfaces during aerial hyphae formation, sporulation, fruit body production and cell penetration. In this ongoing study, we are looking at the role of hydrophobins in the fungal ability to adhere to different substrates. We hypothesize that hydrophobins have a positive effect on the fungi by reducing the surface tension and assisting fungal growth of aerial hyphae and colonization of the fungi between humid and dry areas.

The Blue Stain filamentous fungus, Aureobasidium pullulans, represents a common mold in the nature, commonly known to have a huge role in the attack of wood treated and untreated surfaces like commercial wood products. As Aureobasidium pullulans is also known to produce hydrophobins, a specific aim of this study includes the analysis of the impacts of hydrophobins on mediating surface growth of Aureobasidium pullulans on exterior wood coatings (e.g. wood paints). At least two gene sequences encoding hydrophobins in Aureobasidium pullulans were identified, both to be successively disrupted by integrating a recycled selectable marker into the open reading frames of the genes, creating a double knock-out mutant strain, using homologues recombination.

In this study we will present our work and strategies for identification and clarification of the role of hydrophobins, which future wise can lead to new kind of Blue Stain controllers.

Key words: hydrophobins, Aureobasidium pullulans, surface growth, blue stain, knock-out mutant, transformation

1. Introduction

Hydrophobins are known as small proteins (~100 amino acids) that are produced by filamentous fungi on the surface of spores and hyphae or secreted to the surrounding environment (Bayry et al., 2012; Wessels, 1997). Hydrophobins are highly surface-active molecules, characterized by having a hydrophilic side facing the spore or hyphae surface and a hydrophobic side facing the environment. Thus, these molecules are capable of self-assembly at a hydrophilic-hydrophobic

2 Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

interface, resulting in the formation of a polymeric amphipathic monolayer membrane with the abilities to reverse a surface polarity in such a way that hydrophobic surfaces of a liquid or a solid material become wettable, whereas a hydrophilic surface can be turned into a hydrophobic one (Linder et al., 2005; Yang et al., 2013). Therefore the properties of these proteins and protein assemblies play multiple roles in the development, reproduction and morphogenesis of filamentous fungi, including fungal life mediation at air:water and fungal:host interfaces (Fig. 1), fungal attachment to surfaces, stabilization of fruiting bodies and aerial hyphae, as well as protective coatings for fungal structures (Bayry et al., 2012; Kershaw and Talbot, 1998; Wösten, 2001).

Figure 1: The emerge of an aerial hyphae. Hydrophobin monomers are secreted from the hyphae (grey dots) and forming a monolayer (black/white dots. Black: hydrophobic, white: hydrophilic) in the interface between water and air (Wessels, 1999)

All hydrophobins are characterized by the presence of eight cysteine residues including two pairs of adjacent cysteines and a relatively high content of hydrophobic amino acids (Fig. 2). However, based on differences in sequence, the spacing between the cysteine residues and the distribution of hydrophobic and hydrophilic residues along the protein chain, hydrophobins are divided into two distinct families, known as class I and class II hydrophobins (Wessels, 1994). This classification also corresponds with the differences in the physiochemical properties of the amphipathic monolayers assembled by the members of these two hydrophobin families (Wösten, 2001). Class I hydrophobins form rodlets, monolayers with amyloid fibrillar substructures, that are extremely stable and can be depolymerized only by treatment with strong acids, such as TFA (Wösten et al., 1993). The amphipathic monolayer formed by the class II hydrophobins is, on the other hand, less robust, not amyloid-like as in the case of class I hydrophobins and responds to a variety of stimuli, including denaturing agents, surfactants and organic solvents (Russo et al., 1982). However, a study on hydrophobins from Aspergillus terrerus shows that some hydrophobins cannot clearly be allocated to neither class I nor class II (Jensen et al., 2010).

3 Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Figure 2: The eight cysteines (C) motif. The disulfide bonds between the Cysteines are marked with dashed lines (Askolin, 2006). The exact numbers of amino acids between the Cysteines are shown in Appendix A.

Due to the remarkable abilities of self-assembly and surface modification, e.g. reversal of surface wettability, a wide range of technical, biomedical and personal care applications for hydrophobins have been suggested. For hydrophobin-coated surfaces, one notable application includes the provision of hydrophilic, improved biocompatibility and ultralow relative friction coefficients highly useful for biomedical implants, such as catheters and guide-wires (Misra et al., 2006). Also, hydrophobins provide an efficient method to increase the solubility and biocompatibility of drug delivery vehicles, which may enhance the bioavailability through oral administration (Akanbi et al., 2010). Other applications include the coating of nanodevices for use as emulsifiers and foam stabilizers in the food and beverage industry (Stübner et al., 2010), antifouling or antibacterial coating agents and finally emulsion stabilization aids for personal care products (Hektor and Scholtmeijer, 2005). In addition, recent studies indicate that hydrophobins even have fluorosurfactant-like properties (Milani et al., 2013).

Aureobasidium pullulans (formerly Pullularia pullulans) is an ubiquitous saprophyte that commonly occurs in the phyllosphere of many crop plants and on various tropical fruits (Domsch et al., 1980). However, it is also a common mold in the Scandinavian nature where it remains one of the first fungi to attack wood treated and untreated surfaces like commercial wood products, such as posts and boards.

Like all other filamentous fungi, Aureobasidium pullulans is expected to produce hydrophobins. In this study, the aim is to investigate the role of hydrophobins when Aureobasidium pullulans cultivates and grows on exterior wood or coating surfaces. First, the genome of relevant strains will be analyzed and potential hydrophobin candidates will be isolated. Second, the identified hydrophobin genes will be successively disrupted by a homologue recombination-based gene- knockout procedure, and the resulting Aureobasidium pullulans P268 hydrophobin null mutant strains will be tested in order to clarify the function of the identified hydrophobin candidates during surface growth. The role of the hydrophobins will be measured by the ability of the fungi to produce blue stain in an EN 152 test.

4 Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

2. Materials and methods

2.1 Strains and media

Cells of Escherichia coli K-12 (GeneHogs®, Invitrogen, Life Technologies, Carlsbad, CA) were used as host for plasmid propagation. Plasmid purification was carried out using the Mini AX Plasmid preparation kit according to manufacturer’s instructions (A&A Biotechnology, Gdynia, Poland). Bacterial growth conditions were mediated at 37 oC using LB (Luria-Bertani) medium containing ampicillin to a final concentration of 100 μg/mL.

Aureobasidium pullulans (de Bary) Arnaud, strain P 2681 from the European Standard test 152 (EN 152, 2011) was used as test fungi. Fungal growth was performed on PDA (Potato Dextrose Agar, Difco) medium for 7 days at 25 oC. As a bioinformatical reference the full annotated genome of Aureobasidium pullulans strain ATCC62921 was available from The Gonozyme Project organized by Genome Canada, Genome Quebec, and Genome Alberta. Genomic DNA purification was performed with Fast DNA Spin Kit for Soil (MB Bio) according to manufacturer’s instructions. A full genome sequencing of Aureobasidium pullulans strain P 268 was performed on the Illumina MiSeq platform and sequencing data were produced by The Danish National High-Throughput DNA-Sequencing Centre (Copenhagen, Denmark). The genome was collected and annotated to contigs.

2.2 Bioinformatical comparative analysis of hydrophobin gene sequences in the Aureobasidium pullulans strain ATCC62921 and P268

Search and identification of possible hydrophobins in the Aureobasidium pullulans strain ATCC62921 was made by searching for Cysteine patterns. The search was performed by the Open Access program grepWin (version 1.6.3.546) and the criteria for the Cysteine pattern were:

N(25-40)CN(6-9)CCN(10-12)CN(15-17)CN(7-9)CCN(9-12)CN(2-9) where C=cysteine and N=any amino acid except cysteine.

The corresponding hydrophobin candidates in the Aureobasidium pullulans strain P 268 were identified by blasting the DNA sequence of the hydrophobins from strain ATCC62921 against the genome sequence of strain P 268 using the function BLAST in the CLC genomic workbench software (version 7). Additional search was made with no further results by searching for hydrophobin-related Cysteine patterns. The two identified hydrophobins in strain ATCC62921 were named aur1 and aur2, whereas these hydrophobins were named hfbA and hfbB in strain P 268.

1 Identical to strain no. IMI 269 216 of culture deposited at CABI Bioscience, Egham.

5 Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

2.3 Construction of the pDEL1-bleTK resistance plasmid

A 2920 bp PCR fragment carrying the phleomycin resistance encoding gene, ble, fused to the thymidine kinase gene, TK, under the control of the constitutive promoter, PgpdA, and followed by the trpC gene terminator was amplified with primers BleTK-bw-NotI and BleTK-fw-KpnI using the vector, pSK341 (Krappmann et al., 2005), as template. Both primers carry overhangs containing the recognition sites for the NotI and KpnI restriction enzymes, respectively, and are listed in Table 1. The purified PCR product was subsequently digested with the restriction endonuclease enzymes, NotI and KpnI, and ligated with the 3 Kbp backbone fragment of the high copy number pUC18-derived pDEL1 vector (Nielsen et al., 2006), cut with the same enzymes, thereby removing the pyr-4 marker gene. Ligated products were transformed into electrocompetent E. coli GeneHogs® cells (Invitrogen, Life Technologies, Carlsbad, CA) according to manufacturer’s instructions and selected for resistance towards 100 μg/mL ampicillin. The resulting plasmid; pDEL1-bleTK, a pDEL1 vector based backbone carrying the PgpdA-bleTK-TrpC fragment flanked by 290 bp direct repeat sequences, was verified using restriction enzyme analysis and PCR.

2.4 DNA manipulation methods for disruption of the hfbA gene

2.4.1 PCR and construction of the USER cloning plasmid, pU6000-hfbAΔ

The high copy number one-step PacI USER vector, pU0002 (Hansen et al., 2011) was used as backbone for creating the simple USER cloning plasmids. The plasmid contains both the constitutive glyceraldehyde-3-phosphate (gpdA) promoter and the trpC gene terminator from Aspergillus nidulans.

Primers for USER (uracil-specific excision reagent) cloning of single PCR fragments were designed using the CLC genomics workbench software (version 7). For knockout of the hfbA gene, in total six primers were designed; (1) Two primers, R1-BleTK-USER and R2-BleTK-USER, were designed for the amplification of the 3.6 Kbp PgpdA-bleTK-TrpC gene cassette flanked by 290 bp direct repeat sequences using the constructed pDEL1-bleTK plasmid as template (this study, see above). Both primers carry 8-10 bp uracil containing overhangs (5’-ATCCCCTGAU and 5’- ACGGCCAGU, respectively) that each are complementary to either the uracil containing overhang of the reverse primer used for amplification of the upstream region flanking the hfbA gene (5’- ATCAGGGGAU), or to the uracil containing overhang of the forward primer used for amplification of the downstream region flanking the hfbA gene (5’-ACTGGCCGU), respectively. (2) Two primers, HydA-P1-USER and HydA-P2-USER, were designed for the amplification of a 955 bp ‘Up’ fragment positioned 107 bp upstream the startcodon of the hfbA gene using genomic DNA of the Aureobasidium pullulans P268 strain as template. Both primers carry 8-10 bp uracil containing overhangs (5’-GGGTTTAAU and ATCAGGGGAU, respectively), complementary either to the PacI insertion site of the pU0002 vector (Hansen et al., 2011), or to the overhang of the forward primer for amplification of the bleTK cassette (see above), respectively. (3) Two primers, D1-HfbA-USER and D2-HfbA-USER, were designed for the amplification of a 2300 bp ‘Down’ fragment positioned 272 bp downstream the stopcodon of the hfbA gene using genomic DNA of the Aureobasidium pullulans P268 strain as template. Both primers carry 8-10 bp uracil containing overhangs (5’- ACTGGCCGU and 5’-GGTCTTAAU, respectively), complementary either to the overhang of the reverse primer for amplification of the bleTK cassette (see above), or to the PacI insertion site of the pU0002 vector, respectively. Primers for amplification of the 900-2300 bp regions flanking the

6

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

hfbA gene were designed and optimized (annealing temperature, annealing position on template DNA) according to obtaining the best amplification results.

All primers are shown in Table 1.

For primers (final concentration 0.4 μM) containing uracil bases, the proof reading PfuX7 Hotstart DNA polymerase (Nørholm, 2010) was used together with the 5x High Fidelty Phusion PCR reaction buffer (Thermo Fisher Scientific, Waltham, MA) according to manufacturer’s instructions.

For all PCR reactions additional DMSO and MgCl2 were added to a final concentration of 2% and 2mM, respectively. For amplification of the up –and downstream regions of the hfbA gene, genomic DNA (gDNA) of the Aureobasidium pullulans strain P268 strain was used, whereas a tenfold dilution of the constructed pDEL1-bleTK plasmid (see above) was used as template for the amplification of the bleTK gene cassette. All PCR reactions were carried out using a 2720 Thermal Cycler (Applied Biosystems, Life Technologies, Carlsbad, CA) with the following program: an initial denaturation step at 98 oC for 3 minutes followed by 35 cycles of a denaturation step at 98 oC for 10 seconds, an annealing step at 60 oC for 10 seconds and an elongation step at 72 oC for 4 minutes. Finally, the program was terminated with an elongation step at 72 oC for 10 minutes.

Prior to USER plasmid cloning, the three generated USER PCR products were purified using either the Qiaquick PCR purification kit (Qiagen, Hilden, Germany) or the QiaxII gel extraction kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. Preparation of the PacI USER vector pU0002 for USER cloning and the subsequent fusing of PCR fragments in this vector were as previously described (Hansen et al., 2011). The resulting plasmid carrying the USER cassette for disruption of the hfbA gene was named pU6000-hfbAΔ.

2.4.2 Preparation of the cloned USER gene cassettes for transformation

Purified plasmid DNA of the constructed USER cloning plasmid containing the hfbA(up)-bleTK- hfbA(down) gene cassette for knockout of the hfbA gene was digested with the SwaI restriction endonuclease enzyme (New England Biolabs) according to manufacturer’s instructions. The enzyme digest resulted in two bands of 3 Kbp, representing the plasmid backbone, and 7-8 Kbp, representing the USER cloning gene cassette, respectively, of which the latter was subsequently purified by gel extraction using the QiaxII gel extraction kit (Qiagen, Hilden, Germany).

2.4.3 Transformation of Aureobasidium pullulans strain P268

A fresh cell culture was prepared by inoculating 50 mL of HC (Holliday complete) medium (10 g/L glucose, 3 g/L yeast extract, 1 g/L beef extract, 10 g/L peptone, 3 g/L malt extract, pH 5.7) in a 250 mL baffled flask with 2 mL of an outgrown culture of Aureobasidium pullulans strain P268, allowing the organism to grow for 18-24 hours at 25 oC. Cells were harvested by centrifugation at 4000 rpm for 5 min at 19 oC. The cell pellet, except for a small white layer, was resuspended and washed in

20 mL of SCT buffer (1 M sorbitol, 25 mM CaCl2, 50 mM Tris-HCl, pH 7.5) and centrifuged at 4000 rpm for 5 min at 19 oC. The cell pellet was resuspended in 6 mL of SCT buffer and centrifuged at 4000 rpm for 5 min at 19 oC. Finally, the cell pellet was resuspended in 6 mL of SCT buffer and mixed with another 10 mL of SCT buffer containing 400 mg glucanex (Novozymes, Switzerland). The mixture was incubated in a waterbath for 1 hour at 28 oC during which the amount and quality of protoplasts was examined using microscopy. The mixture was passed through 3 layers of gaze and washed twice with 5 mL of SCT buffer. Finally, the washed protoplasts were adjusted to a

7

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

final concentration of ~5x107 protoplasts/mL and stored in SCT buffer. For genetic transformation, SwaI digested, unpurified USER plasmid DNA was used as well as SwaI digested, gel extracted USER cloning plasmid DNA (7-8 Kb band containing the USER cloning gene cassette). The plasmid, pAN8.1 (Mattern et al., 1988) was used as a positive control. Briefly, 2-5 μg of DNA was mixed with 200 μL of freshly prepared protoplasts. Subsequently, 50 μL of PCT (50 % w/v PEG8000,

50 mM CaCl2, 20 mM Tris and 0.6 M KCl, pH adjusted to 7.5) buffer was added and mixed carefully with protoplasts and DNA prior to incubation at room temperature for 10 minutes. Finally,

250 μL of ATB (1.2 M Sorbitol, 50 mM CaCl2·2 H2O, 20 mM Tris and 0.6 M KCl, pH adjusted with 2 N HCl to 7.2) solution was added followed by the addition of 650 μL sterile PDB (potato dextrose broth, Difco) media. Cells were then incubated for 16 hours at room temperature on a Grant-bio PS-3D platform rotator (Grant Instruments Ltd, Cambridgeshire, England). Different dilutions of transformants were finally spread plated on selective PDA plates to which phleomycin (InvivoGen, San Diego, USA) had been added to a final concentration of 50 μg/mL.

Growing colonies were cut out from the transformation agar plates and regrown on selective media with a final phleomycin concentration of 100 μg/mL. To isolate successful single spore transformants, spores were streaked on selection media and germinating spores were isolated.

Disruption of the hfbB gene will follow the same principles as above.

Table 1: Oligonucleotides used in this study.

All sequences are shown in 5’-3’ direction. Bases shown in capitals anneal to the template DNA, whereas sections of the oligonucleotides in lower case characters represent fusion tags (either restriction enzyme sites or uracil containing overhangs for USER cloning, the latter marked with bold).

3. Results and discussion

3.1 Bioinformatical identification of Hydrophobins in Aureobasidium pullulans

Search and bioinformatical identification of hydrophobins in Aureobasidium pullulans strain ATCC62921 and strain P268 resulted in two hypothetical hydrophobins in ATCC62921 (aur1 and aur2) and two associated hydrophobins in P268 (hfbA and hfbB) (Fig. 3).

8

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Figure 3: Amino acid sequences and alignments of pair A) aur1 and hfbA [70% similarity], pair B) aur2 and hfbB [85% similarity]. Green: Identical amino acids between the pairs, yellow: the eight Cysteines. All noncoding regions/introns were removed by FGENESH (Softberry).

Kyte-Doolitle plots (Fig. S1) reveal that the hydrophobins in Aureobasidium pullulans strain P268 are closest to the class II classification.

3.2 Generating the hfbAΔ knockout mutant strain

Amplification of the up- and downstream DNA regions of the hfbA gene was performed as described in Materials and Methods and resulted in fragments of 955 bp (band 1, Fig. 4) and approx. 2300 bp (band 2, Fig. 4), respectively. Amplification of the DNA fragment carrying the phleomycin resistance marker gene, ble, transcriptionally fused with the thymidine kinase encoding gene, TK, resulted in a 3600 bp fragment (band 3, Fig. 4).

The three amplified PCR products were purified and successfully fused via USER cloning to assemble the 7-8 Kbp hfbA(up)-PgpdA-bleTK-TrpC-hfbA(down) USER gene cassette that was subsequently integrated into the PacI USER vector, pU0002, to generate the resulting USER cloning plasmid, pU6000-hfbAΔ, for disruption of the hfbA gene. Prior to transforming the hfbAΔ USER gene cassette into protoplasts of Aureobasidium pullulans strain P268, isolation of this USER fragment (7-8 Kbp) was performed by a SwaI restriction endonuclease digest of the plasmid, pU6000-hfbAΔ, with subsequent gelelectrophorese mediated separation from the plasmid backbone band of 3 Kbp, followed by gelextraction and purification (band 4, Fig. 4).

9

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Figure 4: Gel electrophorese of DNA fragments. Marker: 1 kb DNA ladder (Fermentas Life Science).

Next, the disruption of the hfbA gene will be made using homologue recombination with the isolated and purified hfbAΔ USER gene cassette, and hereafter the hfbB gene will be knocked out as well using the same procedure. This double knockout mutant strain will be generated either by using integration into the hfbA and hfbB genes with two different resistance marker genes, or by using an optimized version of the described method that allows the selectable marker to be recycled. Briefly, for the latter method, the selectable marker gene is integrated into one target gene using the above described procedure, but will be flanked by direct repeat regions, which allows multiple rounds of gene targeting to be performed in the same mutant strain because the marker can then be repeatedly excised via direct-repeat recombination as previously described (Nielsen et al., 2006).

Controlled laboratory tests will then clarify the role of hydrophobins associated to surface growth and Blue Stain discoloring of Aureobasidium pullulans.

4. Conclusion

Two hydrophobin genes were identified in Aureobasidium pullulans strain P 268. Using other different bioinformatical search tools resulted in no further hydrophobin candidates, meaning that a complete focus will be given to the two identified hydrophobin genes, hfbA and hfbB. The generation of ΔhfbAΔhfbB double knockout mutant strains is in progress and future tests of these mutants will expose the effects of hydrophobins on surface growth by Aureobasidium pullulans. Future antifungal agents and/or strategies could benefit from an acquired deeper knowledge of the role of hydrophobins.

10

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

5. Acknowledgements

Thanks to Ib Schneider, CEO, Coatzyme Aps for weekly discussions and critical views on the scientific work and the perspectives of future results. Also thanks to the IRGWP for enabling a presentation of our study at the RG46, the 46th annual meeting of the International Research Group on Wood Protection. Additional we would like to thank The Danish National Advanced Technology Foundation for founding the Ph.d scholarship associated with this study.

6. References

Akanbi, M.H.J., Post, E., Meter-Arkema, A., Rink, R., Robillard, G.T., Wang, X., Wösten, H.A., Scholtmeijer, K., 2010. Use of hydrophobins in formulation of water insoluble drugs for oral administration. Colloids Surf. B Biointerfaces 75, 526–531. Askolin, S., 2006. Characterization of the Trichoderma reesei hydrophobins HFBI and HFBII. VTT Technical Research Centre of Finland. Bayry, J., Aimanianda, V., Guijarro, J.I., Sunde, M., Latge, J.-P., 2012. Hydrophobins—unique fungal proteins. PLoS Pathog. 8, e1002700. Domsch, K.H., Games, W., Anderson, T.H., 1980. Compendium of Soil Fungi. EN 152, 2011. EN 152:2011, Wood preservatives - Determination of the protective effectiveness of a preservative treatment against blue stain in wood in service - Laboratory method. Hansen, B.G., Salomonsen, B., Nielsen, M.T., Nielsen, J.B., Hansen, N.B., Nielsen, K.F., Regueira, T.B., Nielsen, J., Patil, K.R., Mortensen, U.H., 2011. Versatile enzyme expression and characterization system for Aspergillus nidulans, with the Penicillium brevicompactum polyketide synthase gene from the mycophenolic acid gene cluster as a test case. Appl. Environ. Microbiol. 77, 3044–3051. Hektor, H.J., Scholtmeijer, K., 2005. Hydrophobins: proteins with potential. Curr. Opin. Biotechnol. 16, 434–439. Jensen, B.G., Andersen, M.R., Pedersen, M.H., Frisvad, J.C., Søndergaard, I., 2010. Hydrophobins from Aspergillus species cannot be clearly divided into two classes. BMC Res. Notes 3, 344. Kershaw, M.J., Talbot, N.J., 1998. Hydrophobins and repellents: proteins with fundamental roles in fungal morphogenesis. Fungal Genet. Biol. 23, 18–33. Kershaw, M.J., Thornton, C.R., Wakley, G.E., Talbot, N.J., 2005. Four conserved intramolecular disulphide linkages are required for secretion and cell wall localization of a hydrophobin during fungal morphogenesis. Mol. Microbiol. 56, 117–125. Krappmann, S., Bayram, Ö., Braus, G.H., 2005. Deletion and allelic exchange of the Aspergillus fumigatus veA locus via a novel recyclable marker module. Eukaryot. Cell 4, 1298–1307. Linder, M.B., Szilvay, G.R., Nakari-Setälä, T., Penttilä, M.E., 2005. Hydrophobins: the protein- amphiphiles of filamentous fungi. FEMS Microbiol. Rev. 29, 877–896. Lunkenbein, S., Takenberg, M., Nimtz, M., Berger, R.G., 2011. Characterization of a hydrophobin of the ascomycete Paecilomyces farinosus. J. Basic Microbiol. 51, 404–414. Mattern, I.E., Punt, P.J., Van den Hondel, C., 1988. A vector of Aspergillus transformation conferring phleomycin resistance. Fungal Genet Newsl 35, 25. Milani, R., Monogioudi, E., Baldrighi, M., Cavallo, G., Arima, V., Marra, L., Zizzari, A., Rinaldi, R., Linder, M., Resnati, G., others, 2013. Hydrophobin: fluorosurfactant-like properties without fluorine. Soft Matter 9, 6505–6514. Misra, R., Li, J., Cannon, G.C., Morgan, S.E., 2006. Nanoscale reduction in surface friction of polymer surfaces modified with Sc3 hydrophobin from Schizophyllum commune. Biomacromolecules 7, 1463–1470.

11

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Nielsen, M.L., Albertsen, L., Lettier, G., Nielsen, J.B., Mortensen, U.H., 2006. Efficient PCR-based gene targeting with a recyclable marker for Aspergillus nidulans. Fungal Genet. Biol. 43, 54–64. Nørholm, M.H., 2010. A mutant Pfu DNA polymerase designed for advanced uracil-excision DNA engineering. BMC Biotechnol. 10, 21. Russo, P.S., Blum, F.D., Ipsen, J.D., Abul-Hajj, Y.J., Miller, W.G., 1982. The surface activity of the phytotoxin cerato-ulmin. Can. J. Bot. 60, 1414–1422. Stübner, M., Lutterschmid, G., Vogel, R.F., Niessen, L., 2010. Heterologous expression of the hydrophobin FcHyd5p from Fusarium culmorum in Pichia pastoris and evaluation of its surface activity and contribution to gushing of carbonated beverages. Int. J. Food Microbiol. 141, 110–115. Wessels, J.G., 1997. Hydrophobins: proteins that change the nature of the fungal surface. Adv. Microb. Physiol. 38, 1–45. Wessels, J.G., 1999. Fungi in their own right. Fungal Genet. Biol. 27, 134–145. Wessels, J.G.H., 1994. Developmental regulation of fungal cell wall formation. Annu. Rev. Phytopathol. 32, 413–437. Wösten, H.A., 2001. Hydrophobins: multipurpose proteins. Annu. Rev. Microbiol. 55, 625–646. Wösten, H.A., De Vries, O.M., Wessels, J.G., 1993. Interfacial self-assembly of a fungal hydrophobin into a hydrophobic rodlet layer. Plant Cell Online 5, 1567–1574. Yang, W., Ren, Q., Wu, Y.-N., Morris, V.K., Rey, A.A., Braet, F., Kwan, A.H., Sunde, M., 2013. Surface functionalization of carbon nanomaterials by self-assembling hydrophobin proteins. Biopolymers 99, 84–94.

12

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Supplementary material

Figure S1: Kyle Doolitle diagrams of hfbA and hfbB:

Kyte-Doolittle: hfbA: hfbB:

Only between the 1-8 cysteine:

13

Proceedings IRG Annual Meeting (ISSN 2000-8953) © 2015 The International Research Group on Wood Protection IRG/WP 15-10837

Appendix A Members of class I consist of 100-125 amino acids and are usually longer than class II that only consist of 50-100 amino acids (Hektor and Scholtmeijer, 2005). The possibility to vary the spacing of the cysteine residues is also biggest in class I. Two consensus cysteine spacing patterns were noticed by comparison of know hydrophobins of class I and II. These are:

Class I CN(5–7)CCN(19–39)CN(8–23)CN(5)CCN(6–18)CN(2–13) Class II CN(9–10)CCN(11)CN(16)CN(8–9)CCN(10)CN(6–7) where C=cysteine residue and N=any amino acid except cysteine (Kershaw and Talbot, 1998).

As more hydrophobins are discovered, these spacing patterns seem to change a bit (Kershaw et al., 2005). Class I hydrophobins are also more diverse than class II, as they are found in fungi from two phyla, Ascomycetes and Basidiomycetes. Class II has only been discovered in Ascomycetes. This has led to the parting of class I hydrophobins into class Ia (Ascomycetes) and class Ib (Basidiomycetes) (Lunkenbein et al., 2011).

References Appendix A:

Hektor, H.J., Scholtmeijer, K., 2005. Hydrophobins: proteins with potential. Curr. Opin. Biotechnol. 16, 434–439. Kershaw, M.J., Talbot, N.J., 1998. Hydrophobins and repellents: proteins with fundamental roles in fungal morphogenesis. Fungal Genet. Biol. 23, 18–33. Kershaw, M.J., Thornton, C.R., Wakley, G.E., Talbot, N.J., 2005. Four conserved intramolecular disulphide linkages are required for secretion and cell wall localization of a hydrophobin during fungal morphogenesis. Mol. Microbiol. 56, 117–125. Lunkenbein, S., Takenberg, M., Nimtz, M., Berger, R.G., 2011. Characterization of a hydrophobin of the ascomycete Paecilomyces farinosus. J. Basic Microbiol. 51, 404–414.

14

Additional work

Manuscript V

Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

1 Metagenomics provide valuable information on soil micro‐eukaryotes 2 3 Samuel Jacquiod1, Jonas Stenbæk1, Susana S. Santos2, Anne Winding2, Søren J. Sørensen1, Anders Priemé1 4 5 1Section of Microbiology, University of Copenhagen, 2100, Copenhagen, Denmark 6 2Department of Environmental Science, Aarhus University, Roskilde, Denmark 7 8 Abstract 9 Despite the critical ecological roles of micro‐eukaryotes in terrestrial ecosystems, most descriptive studies 10 of soil microbes published so far focus on bacteria, fungi, or specific groups of microorganisms. Meanwhile, 11 the fast development of metagenome shotgun sequencing results in considerable data accumulation in 12 public repositories, providing microbiologists with substantial amounts of easily available information. We 13 took advantage of publicly available shotgun metagenomic data in order to investigate micro‐eukaryote 14 communities in a grassland soil. The data gathered allowed the evaluation of several factors impacting the 15 micro‐eukaryote community structure, including sampling year, soil depth, and DNA extraction procedure. 16 While most studies on soil micro‐eukaryotes involve sequencing of PCR amplified taxonomic genetic 17 markers (e.g. 18S, ITS), this work represents, to our knowledge, the first ecological report based solely on 18 metagenomic micro‐eukaryote DNA. In contrast to sampling year and depth, the data revealed a significant 19 and pronounced effect of the DNA extraction procedure on the taxonomic composition of soil micro‐ 20 eukaryotes. Some DNA extraction procedures favored the detection of specific taxonomic groups, e.g. 21 density gradient based methods enrich for protists, algae and pico‐eukaryote related sequences. Our 22 analyses suggest that publicly available metagenome data provide valuable quantitative information on soil 23 micro‐eukaryotes, complementing current qualitative ribosomal marker amplicon sequencing methods. 24 25 Introduction 26 Microbial community composition is a key determinant in understanding the functioning of soil ecosystems 27 (Strickland et al., 2009), but most descriptive studies of soil microorganisms have focused so far on bacteria 28 or fungi (Gilbert et al., 2010; Rachid et al., 2015). Therefore, dedicated and integrated studies on micro‐ 29 eukaryotes are becoming a necessity in order to complement our current knowledge on soil microbial 30 communities. Soil micro‐eukaryotes are defined sensus stricto on the basis of their size, mostly below 1 31 mm. This includes a wide distribution of microscopic members with very divergent taxonomical origins, 32 ranging from unikonts (fungi, amoebozoa, and some members of holozoa) to bikonts 33 (stramenopiles/alveolates/rhizaria [SAR], excavates, apusozoa, archaeplastida, hacrobia and non‐SAR 34 chromalveolates). These microorganisms form an important part of soil microbial communities (Adl and 35 Gupta, 2006; Coleman and Whitman, 2005) and are known to be involved in many crucial aspects of 36 terrestrial ecology. For instance, soil fungi represent a large fraction of the microbial biomass (Bailey et al., 37 2002), and are key players in organic matter decomposition (Dinghton, 2003; Treseder and Lennon, 2015), 38 accumulation (Clemmensen et al., 2013), plant nutrition (Liu et al., 2015) and disease (Luo et al., 2015). Soil 39 protists (Adl et al., 2007) are known to be important predators of other microorganisms via their grazing 40 activities (Bonkowsky, 2004), which was reported to be correlated to carbon and nitrogen cycling 41 (Bonkowski and Schaefer, 1997; Esteban et al., 2006). In addition, soil protists may be pathogen/parasites 42 of animals or plants (Siński and Behnke, 2004). A recent study suggested that soil is indeed a potential 43 reservoir of parasite protists (Geisen et al., 2015). Soil is also known to be the major non‐aqueous habitat 44 on Earth for micro‐algae (Zenova et al., 1995). Micro‐algae are obvious players in primary production in soil 45 ecosystems through photosynthesis, but are also involved in many other important aspects, like acting as 46 pioneers in soil formation, participating in soil aggregate stability, producing biologically active compounds, 47 and representing a food source for many heterotrophic soil organisms (Santina et al., 2005). Finally, the 48 metazoan fraction of micro‐eukaryotes is typically dominated by nematodes, which are included at several 49 trophic levels, as different species are known to feed on bacteria, fungi, protists, rotifers, plant roots and 50 other nematodes (Neher, 2010; Ferris, 2010).

1 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

51 52 Culture‐based methods and microscopic observations have been fundamental for our understanding of soil 53 micro‐eukaryotes, but are known to target only a fraction of the diversity (Ekelund and Rønn, 1994). For 54 instance, protist species inventories are overlooking the abundant cryptic encysted forms in soils, as they 55 evade microscopic observations (Esteban et al., 2006) and the fraction of fungal species described so far is 56 considered to be extremely low (Hawksworth, 2001; Buée et al., 2009). However, the fast development of 57 next generation DNA sequencing technologies in terms of data output volume, combined with cost 58 reductions, has opened new possibilities to analyze the diversity of microbial communities by means of 59 molecular approaches. So far, the most successful and widely used method is PCR amplicon sequencing of 60 genetic markers, mainly regions within the ribosomal operon (18S rRNA genes and ITS for micro‐ 61 eukaryotes) (Boenigk et al., 2005; Von der Heyden and Cavalier‐Smith, 2005; Lentendu et al., 2014; Baldrian 62 et al., 2012). While these approaches are widely used to investigate the bacterial component of 63 environmental microbial communities, the proportion of studies targeting micro‐eukaryotes is still 64 relatively small, mostly because of primer specificity issues (Lentendu et al., 2014). This lack of focus and 65 knowledge has direct repercussions, for instance, in studies on complex soil fungal communities, where the 66 proportion of unclassified and unknown ITS sequences may dominate (Buée et al., 2009). As a 67 consequence, this observation is directly limiting further advancement in our understanding of the 68 ecological roles of fungi in soils (McGuire and Treseder, 2010). Furthermore, despite its widespread use by 69 environmental microbiologists, amplicon sequencing still has quantitative biases due to our inability to 70 define primers encompassing all organisms within a specified group, and also because of differential 71 amplification efficiencies between targets (Klindworth et al., 2013). In addition, the quantitative issue of 72 gene copy number per cell within a given species when targeting ribosomal genetic markers is still 73 significantly hampering the methodology (Klappenbach et al., 2000). Although some studies have tried to 74 address and correct this problem for bacteria using information from sequenced genomes (Klappenbach et 75 al., 2001; Větrovský and Baldrian, 2013), it is still hard to evaluate the impact of this bias for quantification 76 of micro‐eukaryotes, as the number of ribosomal genes per genome in different species may range from a 77 few to several hundred copies (Inqa et al., 1998; Kibe et al., 1994). 78 79 Besides the traditional ribosomal marker amplicons approaches, direct shotgun sequencing of soil 80 metagenomes may also be used to investigate soil micro‐eukaryotes. Deep sequencing of total extracted 81 DNA has become popular to investigate soil microbial communities (Delmont et al., 2012; Howe et al., 82 2014). However, this approach should be clearly distinguished from amplicon‐based studies using the term 83 “metagenomic” in their title (Esposito and Kirschberg, 2014). In the present study, we refer to 84 metagenomic as the analysis of shotgun sequencing data generated from the total pool of extracted 85 environmental DNA. This method is based on annotation of short (200‐450bp) DNA sequences in public 86 databases, followed by taxonomic and/or functional affiliation for further quantitative diversity analyses 87 (Delmont et al., 2012; Jacquiod et al., 2013). However, the method has obvious limitations when it comes 88 to analysis of micro‐eukaryotes in soil due to their relative low abundance in shotgun datasets. Despite 89 much larger genome sizes (e.g. 6‐10 times larger for fungi compared to the average 5 mega‐bases bacterial 90 genome), most groups of unicellular eukaryotic soil microorganisms are found in very low abundance, e.g. 91 0.5‐2 % of total biomass (Frostegard et al., 1997; Moore‐Kucera and Dick, 2008), and bacterial sequences 92 typically represent the largest fraction of soil shotgun metagenomes (Delmont et al., 2011; Tveit et al., 93 2013). This observation may partly stem from i) over representation of bacterial sequences in databases, 94 resulting in higher chances of assigning bacterial origin to a given short DNA sequence with the standard 95 similarity cut‐off applied, and ii) morphological aspects of eukaryotic microorganisms, including resistance 96 to chemical and mechanical cell lysis steps, iii) the multiple nuclear states of fungi (haploid, diploid, 97 dikaryotic) resulting in a lower DNA to biomass ratio. Low abundance at the metagenomic level 98 considerably limits the possibilities, including the inability to assemble the reads and extend the sequence 99 lengths for better annotations. In addition, the constitutive presence of introns in the genomes of micro‐ 100 eukaryotes is also preventing the functional affiliation of genes from metagenomes (Lindahl and Kuske,

2 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

101 2014). Finally, due to the considerable lack of representative genomes sequenced in databases (del Campo 102 et al., 2014), correct assignment of reads remains extremely challenging, as most of the sequences simply 103 won’t have any match, but also because false positive detection is very likely to occur, where reads will be 104 annotated toward the wrong taxon. As a consequence, the use of metagenomic data on eukaryotic 105 microorganisms have been dismissed in most soil microbial ecology studies in favor of metranscriptomic 106 approaches, bypassing the bias of intron sequences for accessing their functional potentials (Bailly et al., 107 2007; Geisen et al., 2015). However, with time, more metagenome sequencing projects are completed, and 108 sequence data accumulate rapidly in public repositories, offering microbial ecologists enough information 109 to address novel questions and hypotheses without having to redo the time‐consuming and costly DNA 110 sequencing effort (Nesme et al., 2014; Xu et al., 2014). 111 112 In this study, we took advantage of the growing number of metagenomic sequences in public repositories 113 and the increased available proportion of micro‐eukaryote related sequences for further analysis. 114 Approximately 15.46 million sequences were retrieved from a selected set of publicly available soil 115 metagenomes from an experimental grassland site in England. This particular site was already deeply 116 investigated by metagenomic approaches targeting bacteria (Delmont et al., 2012, Jacquiod et al., 2013; 117 Jacquiod et al., 2014), but no attempts to characterize the micro‐eukaryotic community on this site have 118 been reported so far. Here, we focused on DNA sequences assigned to micro‐eukaryotes after taxonomical 119 annotation in public databases, and performed diversity analyses of the micro‐eukaryote community. 120 Furthermore, in order to circumvent the expected bias of the methodology related to expected miss‐ 121 assignment, we used the Lowest Common Ancestor (LCA) method, and compared it to the classical Best Hit 122 Annotation (BHA). BHA is expected to maximize the amount of read annotation and provide a very sensitive 123 view at the data. However, the amount of false positives obtained is also expected to be high due to wrong 124 taxon assignment. On the other hand, the LCA approach will trade most of the sensitivity for a very 125 conservative annotation rooted at the closest common ancestor in case of multiple hits toward 126 taxonomically divergent targets. In both cases, the taxonomical affiliation was done at the family level as an 127 additional security in order to minimize wrong interpretation of the data. In addition, we also screened the 128 metagenomes in search for rRNA sequences in order to benchmark and validate the overall approach with 129 qualitative molecular evidence of our findings. Our study provides insights into a soil micro‐eukaryote 130 community with regards to changes attributed to different factors including the sampling years, soil depths, 131 and DNA extraction procedures applied. 132 133 Material and methods

134 i. Soil metagenomes retrieval and annotation

135 Thirteen publicly available soil metagenomes generated by 454 Titanium pyrosequencing (Roche, 136 Genoscope, France) were retrieved from the MG‐RAST public repository 137 (https://metagenomics.anl.gov/?page=MetagenomeSelect, METASOIL project) and used in this study (Table 138 1). The metagenomes originates from a park grass soil at Rothamsted Research Station (Harpenden, UK) 139 (Silvertown et al., 2006). A meteorological station at the research center provides public access to weather 140 summaries (http://www.rothamsted.ac.uk/aen/ecn/YEARLYSUMMARY.htm). Additional meteorological 141 information corresponding to the soil sampling years is provided in supporting data (S1). The soil and its 142 related metagenomic shotgun database was originally generated for providing the international scientific 143 community with a publicly available reference soil metagenome for further investigation and comparative 144 purposes (Vogel et al., 2009). The soil metagenomes were generated to cover variation due to sampling 145 year, soil depth, and DNA extraction procedure (Delmont et al., 2011). The average read length is 385 bp 146 (+/‐128 bp) and the average GC content is 61.2 % (+/‐ 7.2 %). 147

3 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

148 The metagenomes were generated using direct and indirect DNA extraction protocols. The direct DNA 149 extraction protocols relied on in situ lysis of cells by chemical and physical bead‐beating treatment and 150 recover the total DNA pool from the soil sample. The indirect DNA extraction approach involved a 151 segregation step by means of a centrifugation gradient in Nycodenz to recover cells separated from the soil 152 matrix (Bakken and Lindahl, 1995; Jacquiod et al., 2014). The metagenomic datasets were analyzed through 153 read affiliation using the MG‐RAST platform (Meyer et al., 2008). The taxonomical affiliation of reads was 154 done using two different methods: the Best Hit Annotation (BHA) and the Lowest Common Ancestor (LCA). 155 BHA was done against the M5NR protein database. For the LCA method, the annotation was rooted to the 156 last common ancestor in case of several matches against taxonomically divergent targets in the M5NR 157 protein database. In both cases, the stringency of the BLAST input parameters was increased compared to 158 the standard default setting of MG‐RAST, from 15 bp minimal alignment lengths (5 AA residues for protein 159 match) up to 45 bp (15 AA) and from 1.00E‐05 e‐value threshold down 1.00E‐08. The choice of stringency 160 was determined after several trials in order to limit the amount of poor quality hits while still preserving 161 enough working material for the downstream statistical analysis. Full taxonomical annotation tables were 162 set down to the familly level, and abundance tables were generated from sequence counts. For the specific 163 survey of micro‐eukaryotes, only the OTUs defined at the family level and belonging to the following 164 taxonomical groups were retained: archaeplastides (chlorophytes, glaucophytes, rhodophytes, bryophyte), 165 non‐SAR chromalveolates (cryptophytes, haptophytes), apusozoa, excavates, SAR, amoebozoa, mycetozoa, 166 fungi, animalia (chaetognatha, cnidaria, nematodes, placozoa, platyhelminthes, rhombozoa, rotifera, 167 tardigrada), filozoa, mesomycetozoea and unclassified sequences affiliated to micro‐eukaryotes at higher 168 taxonomical level (S2, S3). 169 170 In addition, presence of metagenomic sequences matching ribosomal databases where checked. For that 171 purpose, sequences where annotated through different rRNA databases, including SILVA SSU and LSU 172 (Pruesse et al., 2007), Greengenes (DeSantis et al., 2006) and RDP (Cole et al., 2003) and M5RNA (MGRAST 173 database which includes all the others). While several hits were found, no sequences matching the ITS 174 could be identified in this study. The affiliation was done at the family level, and the BLAST cut‐off 175 parameters were set as follow: (95 % minimal identity, 50 bp minimal alignment lengths, and 1.00E‐20 e‐ 176 value threshold). Ribosomal matches were used to benchmark the presence of the dominant groups 177 detected with the conservative LCA annotation method only, respectively for the fungi (S4), the SAR (S5), 178 the micro‐metazoans (S6), the archaeplastids (S7), the amoebozoa (S8) and the excavates (S9).

179 ii. Sample grouping for testing effect of environmental parameters

180 The soil database used in this study contains 13 individual metagenomes generated through different 181 combinations of environmental and technical parameters (Table 1), including the DNA extraction protocols 182 (direct and indirect approaches), the sampling year (2009 and 2010), and the soil sampling depth (0‐10 cm 183 below the rhizosphere [denoted “top soil”] and 0‐20 cm [denoted “full core”]). In order to successfully 184 address each effect, samples were clustered in four groups for the downstream statistical analysis: “Direct‐ 185 2009” (metagenomes F1, J1, J7), “Direct‐2010” (J1a, J1b, J2),”Indirect‐top soil” (F3, F4, F5, J4), and 186 “Indirect‐full core” (F2a, F2b, F6). Due to lack of samples for this particular soil metagenome collection, 187 investigating the combination effects of the sampling year with indirect extraction, as well as sampling 188 depth together with direct extraction was not possible.

189 iii. Alpha‐diversity analysis

190 LCA and BHA contingency tables were assessed by means of rarefaction curves in order to estimate the 191 sequencing depth and sample completeness and Venn diagrams were established to display the richness 192 distribution among the different tested conditions (S2 and S10). The rest of the analysis focuses then on the 193 results obtained with the LCA method. The metagenomic taxonomical distribution of the metagenomic 194 sequences is presented in the pie charts (Figure 1). The percentage of total sequences annotated toward

4 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

195 eukaryotes and micro‐eukaryotes against the total sequence numbers per metagenome were calculated 196 (Table 2). The alpha‐diversity of the 13 metagenomes was computed at family level. The following indices 197 were calculated on the datasets using the Past software (Hammer et al., 2001): total sample richness (S), 198 Simpson index (1‐D), Shannon index (H) and Chao‐1 index (C). Shannon’s Equitability index (E) was 199 calculated as: E = H/ln(Sr); where E ranges between 0 (no evenness) to 1 (complete evenness). The effect of 200 tested factors (DNA extraction approach, year and soil depth) on alpha‐diversity indices was tested with the 201 bilateral Student test after verification of normality (Shapiro test) and variance homoscedasticity (Bartlett 202 test).

203 iv. Beta‐diversity analysis

204 The multivariate analysis was done with the R software version 3.0.2 (R core Team, 2013) with the package 205 vegan (Oksanen et al., 2015), ade4 (Dray and Dufour, 2007), and made4 (Culhane et al., 2002). A Principal 206 Component Analysis (PCA) was performed after center‐scaling normalization. A pattern search was applied 207 to the PCA by grouping the samples according to the DNA extraction method, sampling year and soil depth 208 as described above using the Between Group Analysis method (BGA). The statistical significance of the 209 selected grouping factor was tested with a Monte‐Carlo simulation involving 10,000 permutations (Figure 210 2). The taxonomical profiles were also investigated by means of cluster dendrograms using Jaccard’s 211 similarity and Bray‐Curtis dissimilarity indices (S11). SIMPER analysis (Similarity Percentage) was performed 212 with the Past software to extract each taxa contribution to the overall Bray‐Curtis dissimilarity profile 213 (Hammer et al., 2001). In order to respect the degree of freedom requirement, individual one‐way 214 PERMANOVA tests were performed on the Bray‐Curtis dissimilarity profiles using 10,000 permutations 215 (Table 3). In addition, individual one‐way ANOVA tests were performed to investigate the effect of tested 216 conditions with the manylm function from the package mvabund (Wang et al., 2012) involving multiple 217 linear model regression with 10,000 resampling iterations using residual variance (Table 3). Generalized 218 heatmaps were generated using centered and scaled counts to represent the taxa that were significantly 219 impacted between the two annotation methods (LCA and BHA, S3), and also to focus on the effect of the 220 DNA extraction procedure on the results obtain with the LCA method (Figure 3, manylm ANOVA, p‐value < 221 0.05).

222 v. Spearman’s rank correlations and network analysis

223 Prior to correlations analysis, contingency tables were trimmed in order to avoid generation of too complex 224 networks due to the presence of unnecessary, non‐informative data. Only taxa occurring at least three 225 times in the 13 individual datasets were considered for computing the correlation index (Berry and Widder, 226 2014). As several orders of magnitudes were observed for the OTU counts, data were normalized with log10 227 transformation to account for uneven distribution. Correlations between taxa with different taxonomical 228 origins were investigated using the non‐parametric monotonic Spearman´s rank correlation coefficient for 229 studying co‐occurrences and exclusions. Prior to the network analysis, significance of the Spearman’s rank 230 correlation coefficient was tested with the rcorr function from the Hmisc package in R. Only strong and 231 significant correlation coefficients were considered for this analysis (r >|0.6|, p‐value < 0.05). A network 232 analysis was performed to investigate potential correlation between taxa occurrence with the R package 233 igraph (Csardi and Nepusz, 2006). A sub‐community based approach was applied to capture clusters of 234 OTUs displaying high internal co‐occurrence connectivity using the spinglass.community function of igraph. 235 A null‐model comparison against a simulated random graph with similar vertex‐to‐degree levels was 236 established, implementing negative coefficient for preserving original exclusions correlations. Poorly 237 connected nodes with a degree level below three were trimmed in order to focus only on high centrality 238 sub‐communities. In order to remove the underlying effect of DNA extraction methods, two independent 239 Spearman´s rank correlation matrices were built using separately the direct and indirect extraction 240 metagenomes (r >|0.6|, p‐value < 0.05). Only strong positive and negative correlations verified in both

5 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

241 matrices were kept in order to establish an intersection network between the two DNA extraction methods 242 (Figure 4). 243 244 Results

245 i. The annotation procedure: LCA vs BHA

246 The 13 individual datasets within the Rothamsted soil metagenomic database represent a total of 247 12,575,135 high quality read sequences after trimming. A total of 159,014 and 44,859 sequences (1.26 % 248 and 0.36 % of the total) could be affiliated to the eukaryote domain using the BHA and LCA methods 249 respectively. After filtering for micro‐eukaryotes related reads, a total of 85,276 and 24,323 sequences 250 were considered for analysis from each method respectively, corresponding to 53.63 % and 54.22 % of the 251 total eukaryotic sequences, or 0.68 % and 0.19 % of the original sequences in the database. After 252 taxonomical affiliation, a total of 324 families and 84 rooted unclassified clusters were identified amongst 253 48 phyla. 110 families (34 %) were commonly found by both methods, while the BHA and LCA respectively 254 yielded 159 (49 %) and 55 (17 %) unique families (S2, A). In spite of a higher richness found with the BHA, 255 the rarefaction analysis revealed no significant differences between the observation rates of the two 256 methods for an identical given number of recruited sequences (S2, B, 99% confidence interval). It was 257 worth noticing that the BHA method resulted in equivalent richness discovery between the two extraction 258 methods (S2, C and D) while the LCA clearly resulted in significant differences (S2, E and F). 259 260 The taxonomical composition (S3) revealed a clear dichotomy between the two methods, with enrichment 261 toward unclassified Eukaryota sequences when using the LCA method (22.49 % of total Eukaryotic 262 sequences) compared to the BHA (0.31 %). In addition, some groups were enriched using the LCA method, 263 including members from fungi (ascomycota, basidiomycota, Mortierellaceae from zygomycota and 264 unclassified microsporidia), SAR (Pelagomonadales, apicomplexans, the dinoflagellate Kareniaceae, 265 oomycetes Peronosporomycetes and cercozoa Plasmodiophoridae), chlorophytes Chaetopeltidaceae and 266 Mamiellaceae, choanoflagellate unclassified capsaspora‐like (choanoflagellate), chromalveolates 267 (Blastocystidae and unclassified diatoms) and unclassified Chromadorea from nematode. On the other 268 hand, the BHA resulted in a dominance of sequences affiliated to specific groups of fungi (ascomycota, 269 basidiomycota, zygomycota, entomophthoromycotina and neocallimastigomycota), excavates (e.g. 270 euglenozoa and metamonada), SAR (oomycetes, dinoflagellates, cercozoa Paulinellidae), animalia 271 (cnidarians and rotifer), amoebozoa (Mastigamoebidae and Physaraceae), algae (chlorophytes, 272 rhodophytes and glaucophyte Cyanophoraceae) and chromalveolates (cryptophytes, haptophytes and 273 bacillariophyta).

274 ii. Taxonomical composition of the soil micro‐eukaryote community

275 For the rest of the analysis, we will only focus on the results obtained based on the conservative LCA 276 affiliation method. Half of the Eukaryote‐related sequences were annotated toward micro‐Eukaryote in this 277 soil metagenome database (Figure 1, A). In total, 165 families were identified, in addition to 52 rooted 278 unclassified clusters. The presence of reads matching rRNA databases was verified for 96 of these families 279 (58.18 %), as well as 14 unclassified clusters. Most of these sequences were attributed to Fungi (Figure 1, B 280 and C), especially from Ascomycota (Dothideomycetes, Eurotiomycetes, Sordariomycetes and Leomycetes) 281 and Basidiomycota (Agaromycetes). Other rare phyla were observed at low diversity and abundance, 282 including zygomycota, microsporidia and glomeromycota. A list of dominant fungal families found is 283 provided in supporting information (S4). The remaining sequences were respectively distributed amongst 284 archaeplastida, animalia, SAR, excavate and amoebozoa. The dominant archaeplastida were mainly 285 members from chlorophyte, streptophyta (bryopsida), rhodophyta and diatoms (Figure 1, D). Few 286 sequences could be affiliated to the chromalveolate cryptophytes, Perkinsea and Blastocystidae. A list of 287 dominant algal families found is provided in supporting information (S5). Micro‐animalia related sequences

6 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

288 were mostly affiliated to nematodes and choanoflagellates, but surprisingly also to cnidarians. To some 289 extent, sequences could also be affiliated to rotifers, placozoa, platyhelminthes, tardigrada and porifera 290 (Figure 1, E). A list of dominant micro‐animalia families found is provided in supporting information (S6). 291 The SARs were dominated by members from Oomycetes, Pelagophyceae, Apicomplexa and Ciliophora 292 (Figure 1, F). A list of dominant SAR families found is provided in supporting information (S7). Excavates and 293 amoebozoa were found in very small proportion, and the sequences were mostly affiliated to families 294 Vahlkampfiidae and Trypanosomatidae (excavates), and Dictyosteliida and Hartmannellidae (amoebozoa) 295 (Figure 1, G and H). Lists of dominant excavate and amoebozoa families found are provided in supporting 296 information (S8 and S9). In most of the cases, rRNA sequences matching all the dominant observed groups 297 were retrieved, except for most of the amoebozoa, the ascomycota family Nectriaceae, the diatom 298 Phaeodactylaceae, the family Blastocystidae, the excavate Trypanosomatidae, the choanoflagellate 299 Salpingoecidae, the nematode order Trichinellidae and the ciliophoran Parameciidae.

300 iii. Alpha‐diversity analysis

301 The alpha‐diversity of the 13 micro‐eukaryote metagenomic profiles was analyzed to determine the impact 302 of DNA extraction method, the sampling year and soil depth (Table 2). The comparison between the 303 richness and Chao‐1 gave an average ratio of 0.77 ± 0.11, indicating a relatively minor contribution of 304 singleton and rare taxa to the overall sequence volume. In the case of the metagenomes extracted with the 305 direct approach, the rarefaction curve nearly reached a plateau, indicating that the sequencing depth was 306 enough to estimate the accessible core diversity of the most abundant micro‐eukaryote members using this 307 extraction method. However, in the case of the indirect extraction method, the rarefaction curve isn’t 308 displaying any plateau trend and is significantly showing higher level diversity compared to the other 309 method (S2, F, 99% confidence interval). It is noteworthy mentioning that this trend doesn’t appear with 310 the BHA method. While no significant differences were observed for the total richness, an effect of 311 extraction procedure was seen on the relative proportion of total Eukaryotic and micro‐Eukaryotic DNA 312 obtained in favor of the direct method. As a complement, Simpson, Shannon and Shannon’s equitability 313 indices were all significantly lower in direct extraction samples compared to indirect ones, pointing toward 314 a higher evenness distribution within samples generated with the indirect method. No significant 315 differences on alpha‐diversity indices were observed for sampling year and soil depth. 316 317 Taxonomical distribution of the 217 families/unclassified clusters observed with LCA method revealed a 318 diversity core accounting for 44.2 % of the richness (96/217) (S2, panel E). 56 unique taxa were detected 319 only in samples generated with direct extraction methods. Deeper investigation using the Jaccard similarity 320 index on the effect of sampling year also revealed a core of OTUs accounting for 48.8 % of the richness 321 observed within the direct extraction samples (79/162) (S10 and S11, A). 41 unique taxa were observed 322 only for the year 2009 while 34 unique ones were detected in samples from the year 2010. On the other 323 hand, 65 unique taxa were detected only in samples generated with indirect extraction method. Deeper 324 investigation of the sampling depth effect also revealed also a core of OTUs accounting for 47.8 % of the 325 observed diversity within the indirect extraction samples (77/161). 30 unique taxa were found at 0 – 20 cm 326 soil depth, while 54 unique ones were observed only in the top 10 cm soil were detected (S10).

327 iv. Beta‐diversity analysis

328 The between‐group analysis (BGA) performed on the original Principal Component Analysis (PCA) revealed 329 a statistically significant non‐random distribution of the taxonomic profiles according to the DNA extraction 330 method, sampling year and soil depth (Figure 2, Monte‐Carlo simulation, p < 9.99E‐05). The effect of DNA 331 extraction was seen on the first component of the BGA (51.82 %), while the combination of the second 332 (27.36 %) and third component (20.82 %) was segregating each extraction method according to soil 333 sampling depth and season. The clustering of direct extraction samples based on the sampling year was 334 more obvious for samples from 2010, while the samples from 2009 displayed a higher dispersion as they

7 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

335 also reflect variation due to the milder DNA extraction protocol applied to sample J7. Also, the clustering of 336 indirect extraction samples based on the sampling depth revealed a much larger fluctuation of the top soil 337 samples compared to the full‐core ones. 338 339 These observations were reinforced by the SIMPER analysis based on the Bray‐Curtis dissimilarity index, 340 revealing an overall average divergence of 54.6 % in the taxonomic profiles between DNA extraction 341 approaches (S11, panel B). The direct extraction cluster showed high similarity profiles between the 342 samples, with a SIMPER divergence level of 25 %. The indirect extraction cluster showed a larger variation, 343 with an overall level of dissimilarity of 28 % between samples. The Bray‐Curtis dissimilarity profiles were 344 analyzed with individual one‐way PERMANOVA in order to investigate independently the effect of DNA 345 extraction, year and soil depth (Table 3). The PERMANOVA confirmed the general pattern observed with 346 the multivariate analysis, as the DNA extraction procedure was found to be the only factor significantly 347 impacting the overall dissimilarity profiles (r2 = 0.60, p < 9.99E‐05). A tendency was detected for the effect 348 of depth, but not supported statistically (r2 = 0.30, p = 0.06).

349 v. Linear model regression and analysis of variance

350 The effect of DNA extraction approach, sampling year, and soil depth on taxa abundance was 351 independently investigated using linear model regression and subsequent analysis of variance. The effect of 352 each factor was tested with one‐way ANOVA (Table 3). As identified with the beta‐diversity analysis, DNA 353 extraction approach was again the most impacting factor (p = 2.00E‐04), while the two other parameters 354 did not reveal statistically significant differences. 355 356 The community composition changes were investigated by considering only the taxa displaying statistically 357 significant trends for each of the individual factors (p < 0.05). A generalized heatmap was designed to 358 highlight the predominant effect of the DNA extraction procedure (Figure 3). The impact of the DNA 359 extraction approach resulted in 51 clades from different taxonomical levels displaying significant 360 differences. Direct in situ lysis resulted in a significant enrichment in metagenomic DNA affiliated to fungi in 361 general, mostly from ascomycota and basidiomycota, as well as sequences related to the nematode family 362 Heteroderidae and the chlorophyte family Volvocaceae. On the other hand, the indirect extraction 363 metagenomes were found to be specifically enriched with a very wide diversity of micro‐eukaryote groups. 364 Many micro‐animalia groups were increased, including the Hydridae from cnidarians, the placozoa‐related 365 sequences and the Meloidogynidae from nematode. The fraction of choanoflagelattes (mainly unclassified 366 Capsaspora‐related sequences) and non‐algae protists were also increased, including alveolates like 367 apicomplexans, Aconoidasida and Monocystidae. The fraction of excavates and some archaeplastida was 368 also enhanced, including euglenozoa, trypanosomatidae, Cyanidiaceae (rhodophyta), Chaetopelidaceae 369 (chlorophyte) and Bryopsida‐related sequences. Very specific families of Basidiomycota and Ascomycota 370 were also enriched, including for instance the fraction of unclassified Saccharomycetales and 371 Ustilaginaceae, as well as the fraction of unclassified microsporidia‐related sequences. The proportion of 372 unclassified sequences related to Eukaryota also appeared to be much higher with the indirect approach. 373 374 Regarding the effect of sampling depth and year, the ANOVA revealed much less significant effect on 375 specific taxa. The colder year 2010 (S12, panel A) was found to be enriched with specific groups of 376 unclassified sequences rooted from Basidiomycota/Agaricomycetes, and also the Magnaporthaceae family 377 from the Sordariomycetes (ascomycota). On the other hand, 2010 revealed less members of the order 378 Chaetothyriales and the family Herpotrichiellaceae from Eurotiomycetes (ascomycota), as well as the family 379 Heteroderidae from nematodes. The effect of depth was almost nonexistent (S12, panel B), with only 380 detection of the family Meloidogynidae and the order Tylenchida from nematode that was apparently 381 more abundant it the top 10 cm soil together with the fraction of unclassified Eukaryote sequences, while 382 the fraction of Funariaceae from Bryopsida decreased in the top soil.

8 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

383 vi. Spearman’s rank correlation and network analysis

384 After trimming the low abundance taxa, 46 % of the taxa (100/217) were retained for correlation analysis. 385 The investigation of strong correlations (r >|0.6|, p < 0.05) between the 100 selected taxa across the 13 386 metagenomes revealed the presence of two highly interconnected sub‐communities showing high co‐ 387 occurrence centrality, mostly revealing the differential occurrence patterns found between the two DNA 388 extraction methods applied (Figure 4). Network “A” contained families, which were mostly enriched by 389 direct extraction method, while network “B” contained families, which were conversely selected with the 390 indirect extraction method. Both clusters displayed a very strong inter‐exclusion profile, with a total 391 amount of 59 exclusion edges versus only six co‐occurrences. Network “A” contained the lowest diversity of 392 co‐occurring OTUs, with a dominance of fungi, mostly ascomycota, basidiomycota and zygomycota, 393 together with nematodes and amoebozoa, and a single diatom and alveolate. On the other hand, the sub‐ 394 community B) contains a large diversity panel ranging from fungi to algae, stramenopiles, excavates, 395 amoebozoa, choanoflagelates and other groups. 396 397 Discussion

398 i. Comparison between LCA and BHA

399 So far, the diversity of soil micro‐eukaryotes has been mostly studied through high‐throughput sequencing 400 using universal primer sets to target the 18S rRNA gene (Urich et al., 2008; Meadow and Zabinski, 2012; 401 Baldwin et al., 2013; Bates et al., 2013; Lentendu et al., 2014) or through cDNA libraries derived from 402 environmental RNA in order to by‐pass the intron bias (Grant et al., 2006; Bailly et al., 2007; Damon et al., 403 2012; Geisen et al., 2015). Here, we used shotgun metagenomic data retrieved from a public repository as 404 an alternative to investigate a soil micro‐eukaryote community using the conservative LCA method, 405 benchmarked by comparison with the BHA approach and also detection of rRNA hits. 406 407 We found that the annotation procedure had a tremendous effect upon the qualitative information held by 408 our contingency results. As expected, the BHA method resulted in a significant increase in the amount of 409 sequences being affiliated at the family level, resulting in an overall higher level of observed richness (S2, 410 A). The LCA results in less richness detection due to the conservative assignment method, as seen by the 411 significant increase in the proportion of rooted unclassified Eukaryotic sequences, from 0.31 % with the 412 BHA to 22.5 % (S3). However, this effect doesn’t compromise richness detection, as the rarefaction analysis 413 clearly revealed that both methods provide the same level of richness for a fixed number of recruited 414 metagenomic sequences (S2, B). On the other hand, the fact that much more sequences got rooted back to 415 unclassified Eukaryote using LCA indicates miss‐assignment using the BHA method. Indeed, when looking at 416 the taxonomical divergence found between the two methods, a clear over‐representation of specific 417 groups was detected using the BHA method compared to the LCA (S3). While some of these groups could 418 perfectly make sense in the frame of a soil study (e.g. ascomycota, basidiomycota, oomycetes, and 419 amoebozoa), others don’t, like aquatic associated taxa (e.g. fresh water cercozoa family Paulinellidae, all 420 members of the cnidarians, the glauchophyte Cyanophoraceae, and the diatom family Thalassiosiraceae). 421 On a related note, it was also unexpected to detect neocallimastigomycota‐related DNA in this soil. 422 Neocallimastigomycota related species are known to be associated to warm (≃37°C) and anoxic rumens of 423 the herbivore digestive tracks. Indeed, DNA affiliated to this fungal phylum has been found in non‐gut 424 environments, including soil (Lockhart et al., 2006), and their inter‐host dispersal ability through 425 aerotolerant propagules has been reported as a possible survival strategy outside guts (Gruninger et al., 426 2014). Nevertheless, the fact that no sequences affiliated to this phylum were retrieved using the LCA 427 methods indicates a potential source of incorrect assignment from the BHA method. 428

9 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

429 While the LCA approach gave much more trust‐worthy results from a qualitative point of view by removing 430 and/or reducing significantly most of the wrong assignments, it was still surprising to observe some specific 431 groups, like the cnidarian classes Anthozoa and Hydrozoa as well as Placozoa‐related DNA only in indirect 432 metagenomes (containing only the Trichoplax adhaerens species). The presence of cnidaria‐related 433 organisms in soil has been confirmed by previous molecular approaches, especially in tropical forest soils 434 (Wu et al., 2011), but on the other hand, further recent bioinformatic analyses have also revealed similarity 435 in shared gene content and synteny degree between these organisms and soil nematodes (Wolenski et al., 436 2013). However, as far as we know, Placozoa‐related evidences haven’t been recovered yet from soil. 437 Interestingly, when looking at the presence of hits matching rRNA genes, we could affiliate only one 438 sequence to Tricoplax LSU rRNA (e‐value = 1.00E‐22, 100% identity, 53 bp). Ribosomal sequences matching 439 cnidarians were also retrieved, including seven SSU and one LSU rRNA sequences (e‐values 1.00E‐27/1.00E‐ 440 44, identities 98.4/100 %, length 61/91 bp). While this indicates our current limitations in our attempts to 441 resolve accurate taxonomical affiliation of metagenomic sequences toward these two phyla, our results 442 also may suggest the presence of unknown soil micro‐eukaryote organisms related to these phyla.

443 ii. Assessing a soil micro‐eukaryote community with shotgun metagenomic data

444 The rest of the discussion focuses only on the results obtained with the conservative LCA approach. A total 445 of 217 micro‐eukaryotic OTU were detected (165 families and 52 unclassified rooted OTUs), belonging to a 446 wide phylogenetic distribution covering fungi, protists (including algae) and metazoan (Figure 1). Compared 447 to a recent 18S rRNA gene amplicon study using multiple bar‐coding approach on a grassland soil (Lentendu 448 et al., 2013), our result show that metagenomic DNA results in detection of much less diversity compared 449 to the overall richness detected with the six datasets generated with different group‐specific primers (Table 450 4). However, when compared to a single dataset generated with Eukaryota primers, metagenomic data 451 assigned with LCA resulted in almost exactly the same number of detected families (162 Eukaryotic families 452 detected from 18S rRNA gene amplicons compared to 165 for the soil metagenomes). 453 454 The most abundant group was found to be the fungi, dominated by the phyla ascomycota 455 (Dothideomycetes, Eurotiomycetes and Sordariomycetes) and basidiomycota (Agaricomycetes, 456 Tremellomycetes), which is coherent with other studies (Lentendu et al., 2014; Baldwin et al., 2013). The 457 recovery of metagenomic DNA affiliated to the elusive glomeromycota phylum, which are responsible for 458 the development of arbuscular mycorrhiza (Colombo et al., 2014), was interesting. Indeed, accessing the 459 diversity of these microorganisms was defined as a crucial challenge in mycorrhizal ecology studies (Öpik et 460 al., 2009), has they often evade ribosomal amplicon sequencing due to primer limitations (e.g. ITS2 461 primers, Ihrmark et al., 2013). Nevertheless, our study shows that shotgun metagenomic analysis opens 462 new possibilities for looking at these ecologically important microorganisms. However, only few sequences 463 were detected, probably due to the exclusion of the rhizosphere during sampling and the lack of 464 representative glomeromycota genomes, as only one is currently available in public databases 465 (Rhizophagus irregularis, Börstler et al. 2010, Tisserant et al. 2013). 466 467 The remaining micro‐eukaryote groups mainly contained OTUs known from soil, but also some that were 468 only described from aquatic environments. This is the case, for instance, of members from Mamiellaceae 469 (Demir‐Hilton et al., 2011), which was supported by the detection of three LSU 28S and one SSU 18S rRNA 470 gene hits rooted to the Mamiellophyceae (e‐values 1.00E‐39/‐40, identity 100%, length 81/83 bp). Another 471 example was the detection of metagenomic reads and two amplicon LSU 28S rRNA gene hits (e‐values 472 1.00E‐21/‐40, identity 100/98.8%, length 51/85 bp) matching the family Perkinsidae from Chromerida, 473 containing typical aquatic and parasitic protozoans, causing severe lethal disease in molluscs. On the other 474 hand, some groups, like the stramenopiles Pelagomonadales (pico‐planktonic algae) and choanoflagellates 475 were found. The presence of typical active marine‐related protists in soil could be expected, but not

10 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

476 impossible, as recently shown in a metatranscriptomic study (Geisen et al., 2015). Although we could not 477 identify any rRNA hits matching these groups, this result coupled to the fact that these organisms were 478 significantly enriched with the indirect centrifugation‐based extraction method remains worth noticeable 479 (Figure 3). 480 481 Our finding that ciliophora (one of the three major groups within alveolates along with dinoflagellates and 482 apicomplexa) is a common member of eukaryotic soil microbial community is supported by direct 483 observation and 18S rRNA gene based surveys of other soil protist communities (Adl and Gupta, 2006; Chao 484 et al., 2006; Urich et al., 2008; Bates et al., 2013). Detection of DNA related to parasitic protist groups like 485 apicomplexans and oomycetes is supported by recent metatranscriptomic findings revealing that these 486 groups are indeed active in soils (Geisen et al., 2015). Indeed, members of apicomplexan (e.g. Toxoplasma 487 and Cryptosporidium)are known to include an environmental maturation step under non‐infectious oocysts 488 life form (Siński and Behnke, 2004). However, the low abundance of cercozoa (diverse flagellates and 489 amoebae within rhizaria) contrasts previously published results reporting their dominance within the soil 490 food web (Baldwin et al., 2013; Lentendu et al., 2014), reflecting probable discrepancies between 491 molecular methodologies. However, other rare eukaryotic microorganisms such as members of the 492 apusomonada, euglenozoa, glaucophyta, heterolobosea, and choanomonada were also detected in our 493 study. These rare groups, as well as some stramenopiles, remain difficult to access with metatranscriptomic 494 methods (Bailly et al., 2007, Geisen et al., 2015), indicating the potential of metagenomic data for the 495 genetic detection and analysis of specific rare members that could be difficults to target with other 496 approaches.

497 iii. Evaluating the effect of DNA extraction

498 It is now acknowledged that environmental microbial community analyses based on molecular techniques 499 are strongly linked to the chosen DNA extraction method (Gabor et al., 2003; Inceoglu et al., 2010; Delmont 500 et al., 2011, Lentendu et al., 2013). DNA extraction techniques can be classified in two categories, (1) direct 501 extraction, where the lysis step is directly applied in situ, on the soil samples (Griffiths et al., 2000, Jacquiod 502 et al., 2013), or (2) indirect extraction, where cells are separated from the sample matrix before lysis 503 (Lentendu et al., 2013; Jacquiod et al., 2014). Direct DNA extraction is less time consuming, and yields more 504 DNA compared to indirect extraction. Among direct extraction techniques, bead beating protocols are 505 among the fastest and most effective strategies (Griffiths et al., 2000; Santos et al., 2015), even though 506 shearing of DNA increases with lysis time (Van Elsas et al., 1997; Burgmann et al., 2001). Despite lower DNA 507 yields and longer processing time, indirect DNA extraction techniques can overcome limitations of direct 508 methods by isolating different cell size fractions through a centrifugation density gradient. This approach 509 can be applied to enrich for prokaryotic cells (Bertrand et al., 2005; Jacquiod et al., 2014), and can also be 510 adapted for enrichment of unicellular eukaryotic organisms like algae and yeasts (Lentendu et al., 2013). 511 However, co‐extraction of low density eukaryote cells such as fungal spores and pico‐eukaryotes (Moreira 512 and Lopez‐Garcia, 2002) is a possible source of contamination. Another main bias of the indirect DNA 513 extraction approach is the exclusion of microbes tightly associated to the matrix (Demaneche et al., 2001). 514 515 In this study, DNA extraction approach was the major driving force separating the micro‐eukaryotic 516 communities. Based on our observations (Table 2), the proportion of micro‐eukaryote DNA was significantly 517 higher after applying direct extraction. This is primarily due to an over‐representation of sequences 518 affiliated to ascomycota and basidiomycota (Figure 3). On the other hand, it was very interesting to observe 519 the specific enrichment with a wide range of unexpected OTUs with different taxonomical origin from 520 metagenomes generated with the indirect approach, which correlates with the high diversity level found in 521 these datasets, especially in terms of rarely detected OTUs (Chao‐1, no difference to direct extraction, 522 Table 2). This enrichment encompasses members from protist like apicomplexans, euglenozoa and 523 trypanosomatidae, as well as placozoan‐like and cnidarian‐like sequences. Some fungi taxa were also

11 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

524 increased in the indirect metagenomes, including microsporidia, as well as the unicellular 525 saccharomycetales from ascomycota. This observation was confirmed by the correlation and network 526 analysis, revealing a clear dichotomy between the occurrence patterns of specific members of ascomycota 527 and basidiomycota, compared to all the other members of the micro‐eukaryote community (Figure 4), and 528 is coherent with previous work aiming at isolating unicellular eukaryotes based on centrifugation methods 529 (Lentendu et al., 2013). 530 531 Considering that a centrifugation step was applied to remove large and dense eukaryotic cells prior to lysis, 532 it was unexpected to find this high diversity, equaling the one observed in direct extraction samples in 533 terms of richness. This resulted in an increased resolution for detection of micro‐eukaryotic groups that 534 may otherwise be masked by dominant and easily accessible groups like ascomycota and basidiomycota. 535 For instance, we observed a significant enrichment toward the so‐called pico‐sized eukaryotes in the 536 indirectly extracted metagenomes (Figure 3). These microorganisms have so far mainly been described 537 from aquatic environments (Moreira and Lopez‐Garcia, 2002), but pico‐eukaryotes like bicosoecophyceae 538 and chrysophyceae (stramenopiles) have been previously described in terrestrial ecosystems (e.g. peat bog, 539 Kim et al., 2010). Other pico‐sized groups detected here included prasinophyceae from chlorophytes 540 (Worden, 2006), haptophytes (Cuvelier et al., 2010), dinoflagellates, stramenopiles (oomycota like 541 Lagenidium, Phytopthora and Achlya, bicosoecida like Syluania and Cafeteria, labyrinthulea like 542 labyrinthuloides and thraustochytrids), choanoflagellates and acantharea (Moon‐van der Staay et al., 543 2001). Based on our observations, we conclude that indirect extraction of metagenomic DNA based on cell 544 density segregation results in the specific targeting of an unexpected unknown fraction of the soil micro‐ 545 eukaryotes, potentially including elusive and poorly known pico‐sized soil eukaryotes.

546 iv. Evaluating the effect of soil sampling depth and year

547 Our analysis was sensitive enough to reveal faint fluctuations in the micro‐eukaryote community linked to 548 sampling depth and year, indicating that the method is relevant for investigating effects related to 549 environmental fluctuations in soil. Metagenomic approaches are extremely powerful for detecting changes 550 in complex communities, for pinpointing specific taxa behaving in a predictable ways, and for the 551 development of bioindicators. 552 553 Despite no significant effects detected based on alpha‐ and beta‐diversity, a significant non‐random pattern 554 could be attributed to soil sampling depth and year. Metagenomes from the full soil cores were found to be 555 much more homogenous in terms of variance compared to the highly dispersed top soil samples (S11). 556 Specific abundance increases of particular groups of fungi in different soil layers was expected, as soil 557 fungal communities are stratified along soil depth due to carbon nutrient availability (Voříšková et al., 2014; 558 Rosling et al., 2003). Previous studies on soil protists (Tsyganov et al., 2011; Vincke et al., 2006) revealed 559 high vertical heterogeneity, characterized by higher population density, species richness, and dominance of 560 small‐sized testate amoebae (≤60 μm) in upper soil layers relative to the deeper layers. However, in our 561 study, we found only one difference between the two soil depths, enrichment in the top soil of the 562 nematode family Meloidogynidae known to contain species living in the rhizosphere (Bent et al., 2008). 563 564 The observed effects related to the sampling year may be due to 2010 being colder and drier compared to 565 2009 (S1). 2010 showed particular enrichment of taxa from fungi members, especially Agaricomycetes and 566 Magnaporthaceae. The seasonal impact on soil micro‐eukaryote communities has already been 567 investigated together with water, nutrient and temperature fluctuations (Moore‐Kucera and Dick, 2008, 568 Tsyganov et al., 2011; Voříšková et al., 2014). The colder conditions in 2010 compared to 2009 may have 569 enriched for fungal saprotrophic species adapted to degrade organic matter at lower temperature. For 570 instance, members of Agaricomycetes are indeed known contain organisms involved in plant litter 571 degradation (Kuramae et al., 2013).

12 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

572 573 Conclusion 574 Despite increased accessibility in public repositories, the use of shotgun metagenomic sequencing is still 575 extremely marginal in the field of environmental micro‐eukaryote studies. This is mostly due to the well‐ 576 established gold‐standard amplicon sequencing methods targeting ribosomal markers for accurate 577 taxonomical investigation of these microorganisms, despite the known quantitative biases linked to the 578 methodology. The estimation of taxa abundance based on their count proportion within metagenomes 579 could easily circumvent these biases, and thus complements qualitative analyses brought by ribosomal 580 marker amplicon sequencing. With the constant increase and accessibility of metagenomic data, shotgun 581 sequencing data are on the verge of becoming a relevant source of information for helping deciphering the 582 complex ecology of micro‐eukaryotes. Our study showed that shotgun metagenomic data, when handled 583 correctly with conservative annotation methods, is sensitive enough to link specific fluctuations in the 584 micro‐eukaryote community to environmental factors, like soil depth and weather fluctuations, but also 585 DNA extraction methods. As the current trend is orientated toward more awareness on deciphering the 586 ecological role of soil micro‐eukaryotes, it is expected that metagenomic data will take a prominent role, 587 alongside with already well‐established and emerging techniques. 588 589 Acknowledgements 590 This research was funded by the ITN Marie‐Curie project Trainbiodiverse (SJS, SD, and SSS), by the Villum 591 Center of Excellence CREAM2 Consortium (SJS), and Innovation Fund Denmark (JS). The authors declare no 592 conflict of interests. 593 594 References

595 ‐ Adl, M.S. and Gupta, V.S. (2006) Protists in soil ecology and forest nutrient cycling. Can J For Res 36: 596 1805‐1817. 597 ‐ Adl SM, Leander BS, Simpson AG, Archibald JM, Anderson OR, Bass D, et al. (2007) Diversity, 598 nomenclature, and taxonomy of protists. Syst Biol. 56: 684‐689. 599 ‐ Bailey, V.L., Smith, J.L., Bolton, H. (2002) Fungal‐to‐bacterial ratios in soils investigated for 600 enhanced C sequestration. Soil Biol Biochem 34: 997‐1007. 601 ‐ Bailly, J., Fraissinet‐Tachet, L.M.C., Verner, G.C., Debaud, M., Lemaire, M., We´solowski‐Louvel, M., 602 Marmeisse, R. (2007) Soil eukaryotic functional diversity, a metatranscriptomic approach. ISME J 1: 603 632‐642. 604 ‐ Baldrian, P., Kolařík, M., Stursová, M., Kopecký, J., Valášková, V., Větrovský, T. et al. (2012) Active 605 and total microbial communities in forest soil are largely different and highly stratified during 606 decomposition. ISME J 6: 248‐258. 607 ‐ Baldwin, D.S., Colloff, M.J., Rees, G.N., Chariton, A.A., Watson, G.O., Court, L.N. et al. (2013) 608 Impacts of inundation and drought on eukaryote biodiversity in semi‐arid floodplain soils. Mol Ecol 609 22: 1746‐1758. 610 ‐ Bakken, L.R. and Lindahl, V. (1995) Recovery of bacterial cells from soil. In: Nucleic Acids in the 611 Environment: Methods and Applications. Van Elsas, J.D. and Trevors, J.T. (eds). Springer, Berlin, 612 Germany. pp. 9‐27. 613 ‐ Bates, S.T., Clemente, J.C., Flores, G.E., Walters, W.A., Parfrey, L.W., Knight, R., Fierer, N. (2013) 614 Global biogeography of highly diverse protistan communities in soil. ISME J 7: 652‐659. 615 ‐ Bent, E., Loffredo, A., McKenry, M.V., Becker, J.O., Borneman, J. (2008) Detection and Investigation 616 of Soil Biological Activity against Meloidogyne incognita. J Nematol 40: 109–118. 617 ‐ Berry, D. and Widder, S. (2014) Deciphering microbial interactions and detecting keystone species 618 with co‐occurrence networks. Front Microbiol 5: 219.

13 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

619 ‐ Bertrand, H., Poly, F., Van, V.T., Lombard, N., Nalin, R., Vogel, T.M., Simonet, P. (2005) High 620 molecular weight DNA recovery from soils prerequisite for biotechnological metagenomic library 621 construction. J Microbiol Methods 62: 1‐11. 622 ‐ Boenigk, J., Pfandl, K., Stadler, P., Chatzinotas, A. (2005) High diversity of the 'Spumella‐like' 623 flagellates: an investigation based on the SSU rRNA gene sequences of isolates from habitats 624 located in six different geographic regions. Environ Microbiol 7: 685‐697. 625 ‐ Bonkowski, M. and Schaefer, M. (1997) Interactions between earthworms and soil protozoa: a 626 trophic component in soil food web. Soil Biol Biochem 29: 499‐502. 627 ‐ Bonkowski, M. (2004) Protozoa and plant growth: the microbial loop in soil revisited. New Phytol 628 162: 617‐631. 629 ‐ Börstler, B., Thiéry, O., Sýkorová, Z., Berner, A., Redecker, D. (2010) Diversity of mitochondrial large 630 subunit rDNA haplotypes of Glomus intraradices in two agricultural field experiments and two 631 semi‐natural grasslands. Mol Ecol 19: 1497‐1511. 632 ‐ Burgmann, H., Pesaro, M., Widmer, F., Zeyer, J. (2001) A strategy for optimizing quality and 633 quantity of DNA extracted from soil. J Microbiol Methods 45: 7‐20. 634 ‐ Buée, M., Reich, M., Murat, C., Morin, E., Nilsson, R.H., Uroz, S., Martin, F. (2009) 454 635 Pyrosequencing analyses of forest soils reveal an unexpectedly high fungal diversity. New Phytol 636 184: 449‐456. 637 ‐ Chao, A., Li, P.C., Agatha, S., Foissner, W. (2006) A statistical approach to estimate soil ciliate 638 diversity and distribution based on data from five continents. Oikos 114: 479‐493. 639 ‐ Clemmensen, K.E., Bahr, A., Ovaskainen, O., Dahlberg, A., Ekblad, A., Wallander, H. et al. (2013) 640 Roots and associated fungi drive long‐term carbon sequestration in boreal forest. Science 339: 641 1615‐1618. 642 ‐ Cole, J.R., Chai, B., Marsh, T.L., Farris, R.J., Wang, Q., Kulam, S.A. et al. (2003) The Ribosomal 643 Database Project (RDP‐II): previewing a new autoaligner that allows regular updates and the new 644 prokaryotic taxonomy. Nucleic Acids Res 31: 442‐443. 645 ‐ Coleman, D.C. and Whitman, W.B. (2005) Linking species richness, biodiversity and ecosystem 646 function in soil systems. Pedobiologia 49:479‐497. 647 ‐ Colombo, R.P., Fernández, B.L., Silvani, V.A., Carbonetto, M.B., Rascovan, N., Bompadre, M.J. et al. 648 (2014) Diversity of arbuscular mycorrhizal fungi in soil from the Pampa Ondulada, Argentina, 649 assessed by pyrosequencing and morphological techniques. Can J Microbiol 60: 819‐827. 650 ‐ Csardi, G. and Nepusz, T. (2006) The igraph software package for complex network research. 651 InterJournal, Complex Systems: 1695. http://igraph.org. Accessed 2015, April 27. 652 ‐ Culhane, A.C., Perrière, G., Considine, E.C., Cotter, T.G., Higgins, D.G. (2002) Between‐group 653 analysis of microarray data. Bioinformatics 18: 1600‐1608. 654 ‐ Cuvelier, M.L., Allen, A.E., Monier, A., McCrow, J.P., Messié, M., Tringe, S.G. et al. (2010) Targeted 655 metagenomics and ecology of globally important uncultured eukaryotic phytoplankton. Proc Natl 656 Acad Sci U S A 107: 14679‐14684. 657 ‐ Damon, C., Lehembre, F., Oger‐Desfeux, C., Luis, P., Ranger, J., Fraissinet‐Tachet, L., Marmeisse, R. 658 (2012) Metatranscriptomics reveals the diversity of genes expressed by eukaryotes in forest soils. 659 PLoS One. 7: e28967. 660 ‐ del Campo, J., Sieracki, M.E., Molestina, R., Keeling, P., Massana, R., Ruiz‐Trillo, I. (2014) The others: 661 our biased perspective of eukaryotic genomes. Trends Ecol Evol 29:252‐259. 662 ‐ Delmont, T.O., Prestat, E., Keegan, K.P., Faubladier, M., Robe, P., Clark, I.M. et al. (2012) Structure, 663 fluctuation and magnitude of a natural grassland soil metagenome. ISME J 6:1677‐1687. 664 ‐ Delmont, T.O., Robe, P., Clark, I., Simonet, P., Vogel, T.M. (2011) Metagenomic comparison of direct 665 and indirect soil DNA extraction approaches. J Microbiol Methods 86:397‐400. 666 ‐ Demaneche, S., Jocteur‐Monrozier, L., Quiquampoix, H., Simonet, P. (2001) Evaluation of Biological 667 and Physical Protection against Nuclease Degradation of Clay‐Bound Plasmid DNA. Appl Environ 668 Microbiol 67: 293‐299.

14 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

669 ‐ Demir‐Hilton, E., Sudek, S., Cuvelier, M.L., Gentemann, C.L., Zehr, J.P., Worden, A.Z. (2011) Global 670 distribution patterns of distinct clades of the photosynthetic picoeukaryote Ostreococcus. ISME J 5: 671 1095‐1107. 672 ‐ DeSantis, T.Z., Hugenholtz, P., Larsen, N., Rojas, M., Brodie, E.L., Keller, K. et al. (2006) Greengenes, 673 a chimera‐checked 16S rRNA gene database and workbench compatible with ARB. Appl Environ 674 Microbiol 72: 5069‐5072. 675 ‐ Dinghton, J. (2003) Fungi in ecosystem processes. Vol. 17. Marcel Dekker Inc., Basel, New York. 676 ‐ Dray, S. and Dufour, A.B. (2007) The ade4 package: implementing the duality diagram for 677 ecologists. J Stat Softw 22: 1‐20. 678 ‐ Ekelund, F. and Rønn, R. (1994) Notes on protozoa in agricultural soil with emphasis on 679 heterotrophic flagellates and naked amoebae and their ecology. FEMS Microbiol Rev 15:321‐353. 680 ‐ Esposito, A. and Kirschberg, M. (2014) How many 16S‐based studies should be included in a 681 metagenomic conference? It may be a matter of etymology. FEMS Microbiol Lett 351: 145‐146. 682 ‐ Esteban, G.F., Clarke, K.J., Olmo, J.L., Finlay, B.J. (2006) Soil protozoa ‐ An intensive study of 683 population dynamics and community structure in an upland grassland. Appl Soil Ecol 33: 137‐151. 684 ‐ Ferris, H. (2010) Form and function: metabolic footprints of nematodes in the soil food web. Eur J 685 Soil Biol 46: 97‐104. 686 ‐ Frostegård, A., Petersen, S.O., Bååth, E., Nielsen, T.H. (1997) Dynamics of a microbial community 687 associated with manure hot spots as revealed by phospholipid fatty acid analyses. Appl Environ 688 Microbiol 63: 2224‐2231. 689 ‐ Gabor, E.M., de Vries, E.J., Janssen, D.B. (2003) Efficient recovery of environmental DNA for 690 expression cloning by indirect extraction methods. FEMS Microbiol Ecol 44: 153‐163. 691 ‐ Geisen, S., Tveit, A.T., Clark, I.M., Richter, A., Svenning, M.M., Bonkowski, M. et al. (2015) 692 Metatranscriptomic census of active protists in soils. ISME J. doi: 10.1038/ismej.2015.30. [Epub 693 ahead of print] 694 ‐ Gilbert, J.A., Field, D., Swift, P., Thomas, S., Cummings, D., Temperton, B. et al. (2010) The 695 taxonomic and functional diversity of microbes at a temperate coastal site: a 'multi‐omic' study of 696 seasonal and diel temporal variation. PLoS One 5: e15545. 697 ‐ Grant, S., Grant, W.D., Cowan, D.A., Jones, B.E., Ma, Y., Ventosa, A., Heaphy, S. (2006) Identification 698 of eukaryotic open reading frames in metagenomic cDNA libraries made from environmental 699 samples. Appl Environ Microbiol 72: 135‐43. 700 ‐ Griffiths, R.I., Whiteley, A.S., O'Donnell, A.G., Bailey, M.J. (2000) Rapid Method for Coextraction of 701 DNA and RNA from Natural Environments for Analysis of Ribosomal DNA‐ and rRNA‐Based 702 Microbial Community Composition. Appl Environ Microbiol 66: 5488‐5491. 703 ‐ Gruninger, R.J., Puniya, A.K., Callaghan, T.M., Edwards, J.E., Youssef, N., Dagar, S.S. et al. (2014) 704 Anaerobic fungi (phylum Neocallimastigomycota): advances in understanding their taxonomy, life 705 cycle, ecology, role and biotechnological potential. FEMS Microbiol Ecol 90: 1‐17. 706 ‐ Hammer, Ø., Harper, D.A.T., Ryan, P.D. (2001) PAST: Paleontological statistics software package for 707 education and data analysis. Palaeontologia Electronica 4:9. 708 ‐ Hawksworth, D.L. (2001) The magnitude of fungal diversity: the 1.5 million species estimate 709 revisited. Mycol Res 105: 1422‐1432. 710 ‐ Howe, A.C., Jansson, J.K., Malfatti, S.A., Tringe, S.G., Tiedje, J.M., Brown, C.T. (2014) Tackling soil 711 diversity with the assembly of large, complex metagenomes. Proc Natl Acad Sci U S A. 111: 4904‐ 712 4909. 713 ‐ Ihrmark, K., Bödeker, I.T.M., Cruz‐Martinez, K., Friberg, H., Kubartova, A., Schenck, J. et al. (2012) 714 New primers to amplify the fungal ITS2 region ‐ evaluation by 454‐sequencing of artificial and 715 natural communities. FEMS Microbiol Ecol 82: 666‐677. 716 ‐ Inceoglu, Ö., Hoogwout, E.F., Hill, P., Van Elsas, J.D. (2010) Effect of DNA extraction method on the 717 apparent microbial diversity of soil. App Env Micr 76: 3378‐3382.

15 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

718 ‐ Inga, R., De Doncker, S., Gomez, J., Lopez, M., Garcia, R., Le Ray, D. et al. (1998). Relation between 719 variation in copy number of ribosomal RNA encoding genes and size of harbouring chromosomes in 720 Leishmania of subgenus Viannia. Mol Biochem Parasitol 92: 219‐228. 721 ‐ Jacquiod, S., Demanèche, S., Franqueville, L., Ausec, L., Xu, Z., Delmont, T.O. et al. (2014) 722 Characterization of new bacterial catabolic genes and mobile genetic elements by high throughput 723 genetic screening of a soil metagenomic library. J Biotechnol 190: 18‐29. 724 ‐ Jacquiod, S., Franqueville, L., Cécillon, S., Vogel, T.M., Simonet, P. (2013) Soil bacterial community 725 shifts after chitin enrichment: an integrative metagenomic approach. PLoS One 8: e79699. 726 ‐ Kibe, M.K., ole‐MoiYoi, O.K., Nene, V., Khan, B., Allsopp, B.A., Collins, N.E. et al. (1994) Evidence for 727 two single copy units in Theileria parva ribosomal RNA genes. Mol Biochem Parasitol 66: 249‐259. 728 ‐ Kim, E., Yubuki, N., Leander, B.S., Graham, L.E. (2010) Ultrastructure and 18S rDNA phylogeny of 729 Apoikia lindahlii comb. nov. (Chrysophyceae) and its epibiontic protists, Filos agilis gen. et sp. nov. 730 (Bicosoecida) and Nanos amicus gen. et sp. nov. (Bicosoecida). Protist 161: 177‐196. 731 ‐ Klappenbach, J.A., Dunbar, J.M., Schmidt, T.M. (2000) rRNA operon copy number reflects ecological 732 strategies of bacteria. Appl Environ Microbiol 66: 1328‐1333. 733 ‐ Klappenbach, J.A., Saxman, P.R., Cole, J.R., Schmidt, T.M. (2001) rrndb: the Ribosomal RNA Operon 734 Copy Number Database. Nucleic Acids Res 29: 181‐184. 735 ‐ Klindworth, A., Pruesse, E., Schweer, T., Peplies, J., Quast, C., Horn M, Glöckner, F.O. (2013) 736 Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next‐generation 737 sequencing‐based diversity studies. Nucleic Acids Res 41: e1. 738 ‐ Kuramae, E.E., Hillekens, R.H., de Hollander, M., van der Heijden, M.G., van den Berg, M., van 739 Straalen, N.M., Kowalchuk, G.A. (2013) Structural and functional variation in soil fungal 740 communities associated with litter bags containing maize leaf. FEMS Microbiol Ecol 84: 519‐531. 741 ‐ Lentendu, G., Wubet, T., Chatzinotas, A., Wilhelm, C., Buscot, F., Schlegel, M. (2014) Effects of long‐ 742 term differential fertilization on eukaryotic microbial communities in an arable soil: a multiple 743 barcoding approach. Mol Ecol. 23: 3341‐3355. 744 ‐ Lentendu, G., Hübschmann, T., Müller, S., Dunker, S., Buscot, F., Wilhelm, C. (2013) Recovery of soil 745 unicellular eukaryotes: an efficiency and activity analysis on the single cell level. J Microbiol 746 Methods. 95: 463‐469. 747 ‐ Lindahl, B.D. and Kuske, C.R. (2014) Metagenomics for study of fungal ecology. In: The Ecological 748 Genomics of Fungi. Martin, F. (eds). Hoboken, NJ: John Wiley & Sons, Inc. pp. 281‐303. 749 ‐ Liu, T., Li, C.M., Han, Y.L., Chiang, T.Y., Chiang, Y.C., Sung, H.M. (2015) Highly diversified fungi are 750 associated with the achlorophyllous orchid Gastrodia flavilabella. BMC Genomics. 16:185. 751 ‐ Lockhart, R.J., Van Dyke, M.I., Beadle, I.R., Humphreys, P., McCarthy, A.J. (2006) Molecular 752 biological detection of anaerobic gut fungi (Neocallimastigales) from landfill sites. Appl Environ 753 Microbiol 72: 5659‐5661. 754 ‐ Luo, J., Qiu, H., Cai, G., Wagner, N.E., Bhattacharya, D., Zhang, N. (2015). Phylogenomic analysis 755 uncovers the evolutionary history of nutrition and infection mode in rice blast fungus and other 756 Magnaporthales. Sci Rep. 5: 9448. 757 ‐ McGuire, K.L., Treseder, K.K. (2010) Microbial communities and their relevance for ecosystem 758 models: decomposition as a case study. Soil Biol Biochem 42: 529‐535. 759 ‐ Meadow, J.F. and Zabinski, C.A. (2012) Spatial heterogeneity of eukaryotic microbial communities 760 in an unstudied geothermal diatomaceous biological soil crust: Yellowstone National Park, WY, 761 USA. FEMS Microbiol Ecol 82: 182‐191. 762 ‐ Meyer, F., Paarmann, D., D'Souza, M., Olson, R., Glass, E.M. (2008) The metagenomics RAST server‐ 763 a public resource for the automatic phylogenetic and functional analysis of metagenomes. BMC 764 Bioinformatics 9: 386. 765 ‐ Moon‐van der Staay, S., De Wachter, R., Vaulot, D. (2001) Oceanic 18S rDNA sequences from 766 picoplankton reveal unsuspected eukaryotic diversity. Nature 409: 607‐610.

16 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

767 ‐ Moore‐Kucera, J. and Dick, R.P. (2008) PLFA profiling of microbial community structure and 768 seasonal shifts in soils of a Douglas‐fir chronosequence. Microb Ecol 55: 500‐511. 769 ‐ Moreira, D. and Lopez‐Garcia, P. (2002) The molecular ecology of microbial eukaryotes unveils a 770 hidden world. Trends Microbiol 10: 31‐38. 771 ‐ Neher, D.A. (2010) Ecology of plant and free‐living nematodes in natural and agricultural soil. Annu 772 Rev Phytopathol 48: 371‐394. 773 ‐ Nesme, J., Cécillon, S., Delmont, T.O., Monier, J.M., Vogel, T.M., Simonet, P. (2014) Large‐scale 774 metagenomic‐based study of antibiotic resistance in the environment. Curr Biol 24:1096‐1100. 775 ‐ Oksanen, J., Kindt, R., Legendre, P., O'Hara, B. Package "Vegan". University of Oulu, Oulu 2015. 776 Open URL: http://cran.r‐project.org/web/packages/vegan/vegan.pdf. Accessed 2015 April 27. 777 ‐ Opik, M., Metsis, M., Daniell, T.J., Zobel, M., Moora, M. (2009) Large‐scale parallel 454 sequencing 778 reveals host ecological group specificity of arbuscular mycorrhizal fungi in a boreonemoral forest. 779 New Phytol 184: 424‐437. 780 ‐ Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J. et al. (2007) SILVA: a 781 comprehensive online resource for quality checked and aligned ribosomal RNA sequence data 782 compatible with ARB. Nucleic Acids Res 35: 7188‐7196. 783 ‐ Rachid, C.T., Balieiro, F.C., Fonseca, E.S., Peixoto, R.S., Chaer, G.M., Tiedje, J.M., Rosado, A.S. (2015) 784 Intercropped silviculture systems, a key to achieving soil fungal community management in 785 eucalyptus plantations. PLoS One 10: e0118515. 786 ‐ R Development Core Team (2013) R: A language and environment for statistical computing. R 787 Foundation for Statistical Computing, Vienna, Austria. ISBN 3‐900051‐07‐0, http://www.R‐ 788 project.org/. Accessed 2013 Jun 3. 789 ‐ Rosling, A., Landeweert, R., Lindahl, B.D., Larsson, K.H., Kuyper, T.W., Taylor, A.F.S., Finlay, R.D. 790 (2003) Vertical distribution of ectomycorrhizal fungal taxa in a podzol soil profile. New Phytol 159: 791 775‐783. 792 ‐ Santina, Z., Renata, T., Maurizio, G.P. (2005) Soil algae composition under different agro‐ 793 ecosystemsin North‐Eastern Italy. Agric Ecosyst Environ 112: 1‐12. 794 ‐ Santos, S.S., Nielsen, T.K., Hansen, L., Winding, A. (2015) Comparison of three DNA extraction 795 methods for recovery of soil protist DNA. J Microbiol Methods 115: 13‐9. 796 ‐ Silvertown, J., Poulton, P., Johnston, E., Edwards, G., Heard, M., Biss, P.M. (2006) The Park Grass 797 Experiment 1856‐2006: its contribution to ecology. J Ecol 94: 801‐814. 798 ‐ Siński, E. and Behnke, J.M. (2004) Apicomplexan parasites: environmental contamination and 799 transmission. Pol J Microbiol 53:67‐73. 800 ‐ Strickland, M.S., Lauber, C., Fierer, N., Bradford, M.A. (2009) Testing the functional significance of 801 microbial community composition. Ecology 90: 441‐451. 802 ‐ Tisserant, E., Malbreil, M., Kuo, A., Kohler, A., Symeonidi, A., Balestrini, R. et al. (2013) Genome of 803 an arbuscular mycorrhizal fungus provides insight into the oldest plant symbiosis. Proc Nat Acad Sci 804 USA 110: 20117‐20122. 805 ‐ Treseder, K.K., Lennon, J. (2015) Fungal traits that drive ecosystem dynamics on land. Microb Mol 806 Biol Rev 79: 243‐262. 807 ‐ Tsyganov, A.N., Nijs, I., Beyens, L. (2011) Does climate warming stimulate or inhibit soil protist 808 communities? A test on testate amoebae in high‐arctic tundra with free‐air temperature increase. 809 Protist 162: 237‐248. 810 ‐ Tveit, A., Schwacke, R., Svenning, M.M., Urich, T. (2013) Organic carbon transformations in high‐ 811 Arctic peat soils: key functions and microorganisms. ISME J 7: 299‐311. 812 ‐ Urich, T., Lanzén, A., Qi, J., Huson, D.H., Schleper, C., Schuster, S.C. (2008) Simultaneous assessment 813 of soil microbial community structure and function through analysis of the meta‐transcriptome. 814 PLoS One 3: e2527.

17 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

815 ‐ Van Elsas, J.D., Mantynen, V., Wolters, A.C. (1997) Soil DNA extraction and assessment of the fate 816 of Mycobacterium chlorophenolicum strain PCP‐1 in different soils by 16S ribosomal RNA gene 817 sequence based most‐probable‐number PCR and immunofluorescence. Biol Fertil soils 24: 188‐195. 818 ‐ Větrovský, T. and Baldrian, P. (2013) The variability of the 16S rRNA gene in bacterial genomes and 819 its consequences for bacterial community analyses. PLoS One 8: e57923. 820 ‐ Vincke, S., van de Vijver, B., Nijs, I., Beyens, L. (2006) Changes in the testacean community structure 821 along small soil profiles. Acta Protozool 45: 395‐406. 822 ‐ Vogel, T.M., Simonet, P., Jansson, J.K., Hirsch, P.R., Tiedje, J.M., Van Elsas, J.D., et al. (2009) 823 TerraGenome: a consortium for the sequencing of a soil metagenome. Nat Rev Microbiol 7: 252. 824 ‐ Von der Heyden, S. and Cavalier‐Smith, T. (2005) Culturing and environmental DNA sequencing 825 uncover hidden kinetoplastid biodiversity and a major marine clade within ancestrally freshwater 826 Neobodo designis. Int J Syst Evol Microbiol 55: 2605‐2621. 827 ‐ Voříšková, J., Brabcová, V., Cajthaml, T., Baldrian, P. (2014) Seasonal dynamics of fungal 828 communities in a temperate oak forest soil. New Phytol 201:269‐278. 829 ‐ Wang, Y., Naumann, U., Wright, S.T., Warton, D.I. (2012) mvabund ‐ an R package for model‐based 830 analysis of multivariate abundance data. Methods Ecol Evol 3: 471‐474. 831 ‐ Wolenski, F.S., Layden, M.J., Martindale, M.Q., Gilmore, T.D., Finnerty, J.R. (2013) Characterizing 832 the spatiotemporal expression of RNAs and proteins in the starlet sea anemone, Nematostella 833 vectensis. Nat Protoc. 8: 900‐915. 834 ‐ Worden, A.Z. (2006) Picoeukaryote diversity in coastal waters of the Pacific Ocean. Aquat Microb 835 Ecol. 43: 165‐175. 836 ‐ Wu, T., Ayres, E., Bardgett, R.D., Wall, D.H., Garey, J.R. (2011) Molecular study of worldwide 837 distribution and diversity of soil animals. Proc Natl Acad Sci U S A 108: 17720‐17725. 838 ‐ Xu, Z., Hansen, M.A., Hansen, L.H., Jacquiod, S., Sørensen, S.J. (2014) Bioinformatic approaches 839 reveal metagenomic characterization of soil microbial community. PLoS One 9: e93445. 840 ‐ Zenova, G.M., Shtina, E.A., Dedysh, S.N., Glagoleva, O.B., Likhacheva, A.A., Gracheva, T.A. (1995) 841 Ecological relations of algae in biocenoses. Mikrobiologiya 64: 121‐133.

842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860

18 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

861 Figures list: 862 863 Figure 1: Taxonomical distribution of micro‐eukaryotes in the Rothamsted metagenomic database. 864 Figure 2: Between Group Analysis (BGA) of the micro‐eukaryote soil communities. 865 Figure 3: Generalized heatmap of the DNA extraction effect on soil micro‐eukaryote community. 866 Figure 4: Network of micro‐eukaryote families based on strong Spearman’s rank correlations. 867 868 Tables list: 869 870 Table 1: Description of the metagenomes used in this study. 871 Table 2: Alpha‐diversity analysis of the soil micro‐eukaryote community. 872 Table 3: One‐way analysis of variance for the tested factors on the soil micro‐eukaryote community. 873 Table 4: Comparison of OTU richness between 18S rRNA gene amplicon and metagenomic data sets. 874 875 Supporting data list: 876 877 S1: Meteorological records from the Rothamsted research station. 878 S2: Comparative richness analysis between LCA and BHA methods. 879 S3: Generalized heatmap of the most significant taxonomical differences between LCA and BHA methods. 880 S4: List of dominant fungal OTUs detected in this study. 881 S5: List of dominant archaeplastids/algae OTUs detected in this study. 882 S6: List of dominant hmicro‐metazoan OTUs detected in this study. 883 S7: List of dominant SAR OTUs detected in this study. 884 S8: List of dominant amoebozoa OTUs detected in this study. 885 S9: List of dominant excavates OTUs detected in this study. 886 S10: Venn diagrams of the micro‐eukaryote community richness after LCA annotation. 887 S11: Cluster dendrograms of the micro‐eukaryote metagenomic profiles. 888 S12: Barcharts of the significant changes attributed to the sampling year and soil depth 889 891 892 Figure 1: Taxonomical distribution of micro‐eukaryotes in the Rothamsted metagenomic database. Pie‐ 890 893 charts display the relative proportion of the main taxonomical groups identified (A: Eukaryote, B: Micro‐ 894 Eukaryote, C: Fungi, D: archaeplastida and hacrobia “AH”, E: Micro‐Animalia, F: stramenopiles, alveolata 895 and rhizaria “SAR”, G: Excavata and H: Amoebozoa) 896

897

19 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

898

899 900 901 Figure 2: Between Group Analysis (BGA) of the micro‐eukaryote soil communities. The figure shows a 902 principal component analysis of the taxonomic profiles after applying sample grouping, according to DNA 903 extraction method, sampling year, and soil depth (BGA ratio = 0.22). Non‐random distribution of the BGA 904 was tested using a Monte‐Carlo simulation with 10.000 permutations (p < 9.99E‐05). 905 906

907 908 Figure 3: Generalized heatmap representing the effect of the DNA extraction procedure on soil micro‐ 909 eukaryote community. The color key represents the centered and scaled counts of each taxon. Only 910 phylogenetic groups showing significant abundance differences are shown (ANOVA, p < 0.05). The color 911 keys at the beginning of each row and column are, respectively, showing the sample origins and the 912 taxonomic affiliations of displayed taxa. 913 914

20 Soil DNA Metagenomics provide valuable information on soil micro‐eukaryotes Number of Archaea SubmittedBacteria Eukaryota (Molecular Viruses OtherEcology ) Metagenome Year depth extraction Protocol MGRAST sequences (%) (%) (%) (%) (%) (cm) procedure F1 2009 0‐20 Direct MP Bio1O1 976268 0.74 85.58 2.03 0.04 11.61 4453246.3 J1 2009 0‐20 Direct MP Bio1O1 1130719 0.76 85.65 2.10 0.02 11.47 4453261.3 J7 2009 0‐20 Direct PowerSoil (MO BIO®) 938860 0.77 84.00 1.41 0.03 13.80 4453406.3 J1a 2010 0‐20 Direct MP Bio1O1 1137813 0.72 84.88 1.81 0.03 12.56 4453434.3 J1b 2010 0‐20 Direct MP Bio1O1 919406 0.75 85.27 1.94 0.03 12.01 4453436.3 J2 2010 0‐20 Direct MP Bio1O1 1068770 0.63 86.74 1.69 0.05 10.90 4453435.3 F3 2009 0‐10 Indirect Lysis in agarose plugs 754829 0.81 86.60 1.65 0.02 10.90 4453407.3 F4 2009 0‐10 Indirect DNA Tissue 946839 0.81 86.21 1.94 0.03 11.02 4453257.3 F5 2009 0‐10 Indirect Gram‐positive 754185 0.61 87.28 0.79 0.02 11.30 4453274.3 J4 2009 0‐10 Indirect DNA Tissue 1135084 0.78 86.82 1.99 0.03 10.38 4453256.3 F2a 2009 0‐20 Indirect MP Bio1O1 1094883 0.72 85.58 1.58 0.02 12.10 4453247.3 F2b 2009 0‐20 Indirect MP Bio1O1 890966 0.75 86.03 1.45 0.02 11.75 4453254.3 F6 2009 0‐20 Indirect Lysis in agarose plugs 826513 0.90 86.91 1.25 0.02 10.93 4453433.3 925 926 Table 1: Description of the metagenomes used in this study. All metagenomes are publicly available on the MG‐RAST database website 927 (https://metagenomics.anl.gov/). 928 929

915 916 917 918 Figure 4: Network of micro‐eukaryote families based on strong Spearman’s rank correlations (r > |0.6|, p < 919 0.05). Vertices (nodes) show the families, while edges (links) show the correlation types. The two clusters (A 920 and B) of highly co‐occurring OTUs (green edges) show dominant inter‐cluster exclusion boundaries 921 simplified by the thick red edge (59 negative correlations). Vertices are colored according to the 922 taxonomical color‐key. 923 924

21 930 Metagenomics provide valuable information on soil micro‐eukaryotes 937 Submitted (Molecular Ecology ) Sequence Micro‐ Equitability Eukaryotes (%) Richness (S) Simpson (1‐D) Shannon (H) Chao‐1 (C) A. PERMANOVA on Bray‐Curtis B. Linear Regression Model counts (n) eukaryotes (%) (Eq) 2 Factor p r Significance p Significance Rothamsted metagenomic 2647 (±1553) 0.53 (±0.27) 0.70 (±0.28) 65 (±12) 0.91 (±0.04) 3.09 (±0.29) 0.36 (±0.12) 87 (±24.28) database DNA extraction 9.99E‐05 0.60 *** 2.00E‐04 *** 0.32 0.28 ‐ 0.12 ‐ Direct (n = 6) 3042 (±1305) 0.74 (±0.25) 0.91 (±0.27) 70 (±12) 0.88 (±0.04) 2.89 (±0.24) 0.68 (±0.07) 87 (±24) Sampling year DNA Soil sampling depth 0.06 0.30 . 0.16 ‐ extraction Indirect (n = 7) 867 (±454) 0.36 (±0.11) 0.51 (±0.12) 62 (±12) 0.93 (±0.02) 3.27 (±0.19) 0.80 (±0.03) 87 (±29) method 938 t‐test ** ** ** ‐ ** ** ** ‐ 939 Table 3: One‐way analysis of variance for the tested factors on the soil micro‐eukaryote community. The analysis was done on the beta‐diversity by 2009 (n = 3) 3454 (±2187) 0.65 (±0.33) 0.86 (±0.39) 70 (±14) 0.88 (±0.05) 2.92 (±0.34) 0.69 (±0.11) 83 (±19) 940 means of PERMANOVA based on the Bray‐Curtis dissimilarity (A) and also by means of a linear model regression. The PERMANOVA was done with Sampling year 2010 (n = 3) 4103 (±313) 0.82 (±0.16) 0.96 (±0.17) 69 (±14) 0.87 (±0.02) 2.86 (±0.17) 0.68 (±0.01) 90 (±24) 941 10.000 permutations, while the ANOVA with linear model was done with 10.000 iterations by residual variance resampling. **: p < 0.01; ‐: p > 0.1. t‐test ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ 942 0‐10 cm (n = 4) 1893 (±1016) 0.41 (±0.11) 0.57 (±0.13) 65 (±14) 0.93 (±0.02) 3.28 (±0.25) 0.79 (±0.04) 94 (±37) 943 Soil sampling 0‐20 cm (n = 3) 1390 (±617) 0.28 (±0.06) 0.44 (±0.09) 58 (±10) 0.93 (±0.01) 3.25 (±0.13) 0.80 (±0.02) 79 (±20) depth t‐test ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ 931 932 Table 2: Alpha‐diversity analysis of the soil micro‐eukaryote community (average ± standard deviation). Sequence counts: total sequence counts; S: 933 Sample richness (family); D‐1: Simpson index; H: Shannon index; E: Shannon’s Equitability index; C: Chao‐1 index. The effect of tested variables on the 934 indices was assessed with a bilateral Student´s t‐test. **: p < 0.01; *: 0.01 < p < 0.05; . : 0.05 < p < 0.1; ‐: p > 0.05. 935 936

22 944 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology ) Reference Datasets OTUs (n) Families (n) Unclassified (n) Families (%) Eukaryota 467 162 305 34.69 Fungi 359 204 155 56.82 Cercozoa 187 128 59 68.45 Lentendu et al., 2014 950 Photoautotrophs 117 27 90 23.08 Chryso/Synurophyceae 56 47 9 83.93 Kinetoplastida 46 43 3 93.48 This study Soil metagenomic DNA 217 165 52 76.04 945 946 Table 4: Comparison of OTU richness of sequences obtained from 18S rRNA gene amplicon and metagenomic data sets. 947

948

949

951 952 953 S1: A) Meteorological records from the Rothamsted research station. The year 2010 was a colder year 954 when compared to the average for 1981‐2010, and also to 2008, 2009 and 2011. Data are publicly available 955 at http://www.rothamsted.ac.uk/aen/ecn/YEARLYSUMMARY.htm. B) 5‐years average temperature records 956 at the Rothamsted research station, from 1880 to 2010. The 5‐years means (dots) show that by the year 957 2000 the mean air temperature was approximately 1°C higher than the long‐term mean (1880 to 1985, 958 green line). http://www.rothamsted.ac.uk/environmental‐change‐network/climate). 959 960

23 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

961 970

962 971 963 S2: Comparative richness analysis between LCA and BHA methods. Venn diagrams represent the 972 S3: Generalized heatmap of the most significant taxonomical differences between the LCA and the BHA 964 distribution of shared and unique families and rooted unclassified clusters between A: the LCA and BHA 973 methods. The color key represents the centered and scaled counts of each taxon. Only phylogenetic groups 965 methods, C: the two extraction methods in the case of the BHA, E: the two extraction methods in the case 974 showing significant abundance differences are shown (ANOVA, p < 0.05). The color keys at the beginning of 966 of LCA. The rarefaction curves are showing the family and rooted unclassified clusters diversity detection 975 each row and column are, respectively, showing the sample origins and the taxonomic affiliations of 967 upon sequence recruitment between, B: the LCA and BHA methods, D: the two extraction methods in the 976 displayed taxa. 968 case of the BHA and F: the two extraction methods in the case of the LCA. 977 969

24 978 Metagenomics provide valuable information on soil micro‐eukaryotes 981 S4: List of dominant fungi OTUs found in this study. The count column is indicating theSubmitted number of metagenomic (Molecular sequence Ecology hits while the) others are 979 982 showing the number of matches in different ribosomal databases. Phylum Class Order Family Counts SILVA SSU SILVA LSU Greengenes RDP M5RNA 983 SILVA SILVA Ascomycota Dothideomycetes Capnodiales Mycosphaerellaceae* 6847 2 3 0 0 5 Phylum Class Order Family Counts Greengenes RDP M5RNA Ascomycota Dothideomycetes Pleosporales Phaeosphaeriaceae* 216 1 1 0 0 2 SSU LSU Ascomycota Dothideomycetes Pleosporales Pleosporaceae* 160 1 0 0 0 1 Chlorophyta Chlorophyceae* Chaetopeltidales Chaetopeltidaceae 50 0 9 0 0 9 Ascomycota Dothideomycetes Pleosporales* Unclassified 349 3 1 0 0 4 Chlorophyta Chlorophyceae Chlamydomonadales Chlamydomonadaceae* 65 0 3 0 0 3 Ascomycota Eurotiomycetes Eurotiales Trichocomaceae* 1876 8 9 2 0 21 Chlorophyta Chlorophyceae Chlamydomonadales* Volvocaceae 105 0 6 0 0 6 Ascomycota Eurotiomycetes* Unclassified ‐ 1228 25 31 2 0 60 Chlorophyta Mamiellophyceae* Mamiellales Mamiellaceae 178 1 3 0 0 4 Ascomycota Eurotiomycetes Onygenales Ajellomycetaceae* 118 3 9 0 0 12 Chlorophyta Trebouxiophyceae Chlorellales Chlorellaceae* 139 3 0 1 1 3 Ascomycota Eurotiomycetes Onygenales Arthrodermataceae* 150 1 3 0 0 4 Rhodophyta Bangiophyceae Cyanidiales Cyanidiaceae* 180 0 1 0 0 1 Ascomycota Eurotiomycetes Onygenales* Unclassified 489 7 19 0 0 26 Streptophyta Bryopsida Funariales Funariaceae* 335 1 1 0 0 2 Ascomycota Leotiomycetes Helotiales Sclerotiniaceae* 1326 4 6 0 0 10 Bacillariophyta Bacillariophyceae Naviculales Phaeodactylaceae 56 0 0 0 0 0 Ascomycota Sordariomycetes Glomerellales* ‐ 244 21 0 0 26 25 Bacillariophyta Coscinodiscophyceae* Thalassiosirales Thalassiosiraceae 45 0 1 0 0 1 Ascomycota Sordariomycetes Hypocreales Clavicipitaceae* 723 8 5 0 0 12 984 Ascomycota Sordariomycetes Hypocreales* Unclassified 136 15 19 0 0 30 985 S5: List of dominant archaeplastids OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the Ascomycota Sordariomycetes Hypocreales Nectriaceae 499 0 0 0 0 0 986 others are showing the number of matches in different ribosomal databases. Ascomycota Sordariomycetes Sordariales Chaetomiaceae* 351 0 2 0 0 2 987 Ascomycota Sordariomycetes Sordariales* Lasiosphaeriaceae 178 0 70 0 0 70 Phylum Class Order Family Counts SILVA SSU SILVA LSU Greengenes RDP M5RNA Ascomycota Sordariomycetes Sordariales Sordariaceae* 405 0 68 0 0 68 Choanozoa Choanoflagellatea Choanoflagellida Salpingoecidae 51 0 0 0 0 0 Ascomycota Sordariomycetes* Unclassified ‐ 766 28 90 0 26 128 Cnidaria Anthozoa Actiniaria* Edwardsiidae 54 1 0 0 0 1 Basidiomycota* Unclassified ‐ ‐ 435 62 126 0 0 187 Cnidaria Hydrozoa Hydroida Hydridae* 102 4 0 0 0 4 Basidiomycota Agaricomycetes Agaricales* Unclassified 1440 34 71 0 0 105 Filasterea Capsaspora Unclassified ‐ 238 0 0 0 0 0 Basidiomycota Agaricomycetes Agaricales Marasmiaceae* 226 0 1 0 0 1 Nematoda Chromadorea* Unclassified ‐ 93 13 20 0 0 33 Basidiomycota Agaricomycetes Agaricales Psathyrellaceae* 115 1 2 0 0 3 Nematoda Rhabditida Rhabditidae* ‐ 108 2 7 0 0 9 Basidiomycota Agaricomycetes Agaricales Tricholomataceae* 813 10 27 0 0 37 Nematoda Spirurida Onchocercidae* ‐ 27 0 3 0 0 3 Basidiomycota Agaricomycetes* Unclassified ‐ 483 52 114 0 0 165 Nematoda* Trichocephalida Trichinellidae ‐ 19 22 23 0 0 45 Basidiomycota Agaricomycetes Polyporales Coriolaceae* 108 2 6 0 0 8 Placozoa ‐ ‐ ‐ 43 0 1 0 0 1 Basidiomycota Tremellomycetes Tremellales Tremellaceae* 306 3 5 0 0 8 Rotifera Bdelloidea Adinetida Adinetidae* 53 0 1 0 0 1 980 Rotifera Bdelloidea Philodinida Philodinidae* 14 1 0 0 0 1 988

25 989 S6: List of dominantMetagenomics micro‐metazoan provide OTUs found valuable in this study. informationThe count column onis indicating soil micro the number‐eukaryotes of metagenomic sequence hits while the Amoebozoa Tubulinea Euamoebida Hartmannellidae 27 Submitted0 0 (Molecular0 Ecology0 )0 990 others are showing the number of matches in different ribosomal databases. Mycetozoa Dictyostelia Dictyosteliida Unclassified 185 0 0 0 0 0 991 Mycetozoa Myxogastria Physariida Physaraceae 5 0 0 0 0 0 SILVA SILVA 996 SAR Phylum Class Order Family Counts Greengenes RDP M5RNA SSU LSU 997 S8: List of dominant amoebozoa OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others Alveolata Apicomplexa Aconoidasida Haemosporida* Unclassified 12 0 6 0 6 0 998 are showing the number of matches in different ribosomal databases. Alveolata Apicomplexa* Unclassified ‐ ‐ 11 31 16 0 0 47 999 Alveolata Apicomplexa Conoidasida Eimeriorina Sarcocystidae* 31 1 1 0 0 2 1000 Alveolata Apicomplexa Conoidasida Eucoccidiorida* Unclassified 13 5 9 0 0 14 1001 Alveolata Apicomplexa Conoidasida Cryptosporidium Cryptosporidiidae* 10 4 0 0 0 4 SILVA SILVA Phylum Class Order Family Counts Greengenes RDP M5RNA Alveolata Apicomplexa Gregarinia Eugregarinida* ‐ 13 19 6 0 0 25 SSU LSU Alveolata Chromerida Perkinsea Perkinsida Perkinsidae* 21 0 2 0 0 2 Euglenozoa* Kinetoplastida Trypanosomatida Trypanosomatidae 66 1 2 0 0 3 Alveolata Ciliophora Intramacronucleata Hymenostomatida Tetrahymenidae* 14 0 4 0 0 4 Metamonada Parabasalia Trichomonadida Trichomonadidae* 29 10 0 0 0 10 Alveolata Ciliophora Intramacronucleata* Peniculia Parameciidae 45 1 9 0 0 10 Percolozoa Heterolobosea Schizopyrenida Vahlkampfiidae* 289 4 0 0 0 4 Alveolata Dinoflagellata Dinophyceae Amphidinium Kareniaceae* 12 0 1 0 0 1 1002 Rhizaria Cercozoa Monadofilosa Euglyphida Paulinellidae* 4 0 2 0 0 2 1003 S9: List of dominant excavates OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others Rhizaria Cercozoa Plasmodiophoromycetes Plasmodiophorida Plasmodiophoridae* 5 3 6 0 0 8 1004 are showing the number of matches in different ribosomal databases. Rhizaria Cercozoa Sarcomonadea Cercomonadida* ‐ 3 2 3 0 0 5 1005 Stramenopiles* Heterokontophyta Pelagophyceae Pelagomonadales Unclassified 237 2 7 0 0 9 1006 Stramenopiles Heterokontophyta Phaeophyceae Ectocarpales Ectocarpaceae* 27 0 2 0 0 2 Stramenopiles Heterokontophyta Oomycota Peronosporales* Peronosporomycetes 236 0 1 0 0 1 Stramenopiles Heterokontophyta Blastocystae Blastocystida Blastocystidae 29 0 0 0 0 0 992 993 S7: List of dominant SAR OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others are 994 showing the number of matches in different ribosomal databases. 995 SILVA SILVA Phylum Class Order Family Counts Greengenes RDP M5RNA SSU LSU Amoebozoa Incertae sedis Incertae sedis Breviatea* 1 1 0 0 0 1 Amoebozoa Archamoebae Entamoebida Entamoebidae 4 0 0 0 0 0 Amoebozoa Archamoebae Mastigamoebida Mastigamoebaea 2 0 0 0 0 0 Amoebozoa Discosea Centramoebida Acanthamoebidae* 1 1 0 0 0 1

26 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

1007 1014 1008

1015 1016 1017 1009 1018 S11: Cluster dendrograms of the micro‐eukaryote metagenomic profiles. Dendrogram A is showing the 1010 1019 distribution of the profiles according the Jaccard similarity index, while dendrogram shows the distribution 1011 S10: Venn diagrams of the micro‐eukaryote community richness after LCA annotation. 1020 according to the Bray‐Curtis dissimilarity index. 1012 1021 1013 1022

27 Metagenomics provide valuable information on soil micro‐eukaryotes Submitted (Molecular Ecology )

1023

1024 1025 S12: Barcharts of the significant changes attributed to the sampling A) year and B) soil depth (Linear model, 1026 ANOVA, p < 0.05). 1027

28 861 Figures list: 862 863 Figure 1: Taxonomical distribution of micro‐eukaryotes in the Rothamsted metagenomic database. 864 Figure 2: Between Group Analysis (BGA) of the micro‐eukaryote soil communities. 865 Figure 3: Generalized heatmap of the DNA extraction effect on soil micro‐eukaryote community. 866 Figure 4: Network of micro‐eukaryote families based on strong Spearman’s rank correlations. 867 868 Tables list: 869 870 Table 1: Description of the metagenomes used in this study. 871 Table 2: Alpha‐diversity analysis of the soil micro‐eukaryote community. 872 Table 3: One‐way analysis of variance for the tested factors on the soil micro‐eukaryote community. 873 Table 4: Comparison of OTU richness between 18S rRNA gene amplicon and metagenomic data sets. 874 875 Supporting data list: 876 877 S1: Meteorological records from the Rothamsted research station. 878 S2: Comparative richness analysis between LCA and BHA methods. 879 S3: Generalized heatmap of the most significant taxonomical differences between LCA and BHA methods. 880 S4: List of dominant fungal OTUs detected in this study. 881 S5: List of dominant archaeplastids/algae OTUs detected in this study. 882 S6: List of dominant hmicro‐metazoan OTUs detected in this study. 883 S7: List of dominant SAR OTUs detected in this study. 884 S8: List of dominant amoebozoa OTUs detected in this study. 885 S9: List of dominant excavates OTUs detected in this study. 886 S10: Venn diagrams of the micro‐eukaryote community richness after LCA annotation. 887 S11: Cluster dendrograms of the micro‐eukaryote metagenomic profiles. 888 S12: Barcharts of the significant changes attributed to the sampling year and soil depth 889 891 892 Figure 1: Taxonomical distribution of micro‐eukaryotes in the Rothamsted metagenomic database. Pie‐ 890 893 charts display the relative proportion of the main taxonomical groups identified (A: Eukaryote, B: Micro‐ 894 Eukaryote, C: Fungi, D: archaeplastida and hacrobia “AH”, E: Micro‐Animalia, F: stramenopiles, alveolata 895 and rhizaria “SAR”, G: Excavata and H: Amoebozoa) 896

897 898

899 900 901 Figure 2: Between Group Analysis (BGA) of the micro‐eukaryote soil communities. The figure shows a 902 principal component analysis of the taxonomic profiles after applying sample grouping, according to DNA 903 extraction method, sampling year, and soil depth (BGA ratio = 0.22). Non‐random distribution of the BGA 904 was tested using a Monte‐Carlo simulation with 10.000 permutations (p < 9.99E‐05). 905 906

907 908 Figure 3: Generalized heatmap representing the effect of the DNA extraction procedure on soil micro‐ 909 eukaryote community. The color key represents the centered and scaled counts of each taxon. Only 910 phylogenetic groups showing significant abundance differences are shown (ANOVA, p < 0.05). The color 911 keys at the beginning of each row and column are, respectively, showing the sample origins and the 912 taxonomic affiliations of displayed taxa. 913 914 Soil DNA Number of Archaea Bacteria Eukaryota Viruses Other Metagenome Year depth extraction Protocol MGRAST sequences (%) (%) (%) (%) (%) (cm) procedure F1 2009 0‐20 Direct MP Bio1O1 976268 0.74 85.58 2.03 0.04 11.61 4453246.3 J1 2009 0‐20 Direct MP Bio1O1 1130719 0.76 85.65 2.10 0.02 11.47 4453261.3 J7 2009 0‐20 Direct PowerSoil (MO BIO®) 938860 0.77 84.00 1.41 0.03 13.80 4453406.3 J1a 2010 0‐20 Direct MP Bio1O1 1137813 0.72 84.88 1.81 0.03 12.56 4453434.3 J1b 2010 0‐20 Direct MP Bio1O1 919406 0.75 85.27 1.94 0.03 12.01 4453436.3 J2 2010 0‐20 Direct MP Bio1O1 1068770 0.63 86.74 1.69 0.05 10.90 4453435.3 F3 2009 0‐10 Indirect Lysis in agarose plugs 754829 0.81 86.60 1.65 0.02 10.90 4453407.3 F4 2009 0‐10 Indirect DNA Tissue 946839 0.81 86.21 1.94 0.03 11.02 4453257.3 F5 2009 0‐10 Indirect Gram‐positive 754185 0.61 87.28 0.79 0.02 11.30 4453274.3 J4 2009 0‐10 Indirect DNA Tissue 1135084 0.78 86.82 1.99 0.03 10.38 4453256.3 F2a 2009 0‐20 Indirect MP Bio1O1 1094883 0.72 85.58 1.58 0.02 12.10 4453247.3 F2b 2009 0‐20 Indirect MP Bio1O1 890966 0.75 86.03 1.45 0.02 11.75 4453254.3 F6 2009 0‐20 Indirect Lysis in agarose plugs 826513 0.90 86.91 1.25 0.02 10.93 4453433.3 925 926 Table 1: Description of the metagenomes used in this study. All metagenomes are publicly available on the MG‐RAST database website 927 (https://metagenomics.anl.gov/). 928 929

915 916 917 918 Figure 4: Network of micro‐eukaryote families based on strong Spearman’s rank correlations (r > |0.6|, p < 919 0.05). Vertices (nodes) show the families, while edges (links) show the correlation types. The two clusters (A 920 and B) of highly co‐occurring OTUs (green edges) show dominant inter‐cluster exclusion boundaries 921 simplified by the thick red edge (59 negative correlations). Vertices are colored according to the 922 taxonomical color‐key. 923 924 930 937 Sequence Micro‐ Equitability Eukaryotes (%) Richness (S) Simpson (1‐D) Shannon (H) Chao‐1 (C) A. PERMANOVA on Bray‐Curtis B. Linear Regression Model counts (n) eukaryotes (%) (Eq) 2 Factor p r Significance p Significance Rothamsted metagenomic 2647 (±1553) 0.53 (±0.27) 0.70 (±0.28) 65 (±12) 0.91 (±0.04) 3.09 (±0.29) 0.36 (±0.12) 87 (±24.28) database DNA extraction 9.99E‐05 0.60 *** 2.00E‐04 *** 0.32 0.28 ‐ 0.12 ‐ Direct (n = 6) 3042 (±1305) 0.74 (±0.25) 0.91 (±0.27) 70 (±12) 0.88 (±0.04) 2.89 (±0.24) 0.68 (±0.07) 87 (±24) Sampling year DNA Soil sampling depth 0.06 0.30 . 0.16 ‐ extraction Indirect (n = 7) 867 (±454) 0.36 (±0.11) 0.51 (±0.12) 62 (±12) 0.93 (±0.02) 3.27 (±0.19) 0.80 (±0.03) 87 (±29) method 938 t‐test ** ** ** ‐ ** ** ** ‐ 939 Table 3: One‐way analysis of variance for the tested factors on the soil micro‐eukaryote community. The analysis was done on the beta‐diversity by 2009 (n = 3) 3454 (±2187) 0.65 (±0.33) 0.86 (±0.39) 70 (±14) 0.88 (±0.05) 2.92 (±0.34) 0.69 (±0.11) 83 (±19) 940 means of PERMANOVA based on the Bray‐Curtis dissimilarity (A) and also by means of a linear model regression. The PERMANOVA was done with Sampling year 2010 (n = 3) 4103 (±313) 0.82 (±0.16) 0.96 (±0.17) 69 (±14) 0.87 (±0.02) 2.86 (±0.17) 0.68 (±0.01) 90 (±24) 941 10.000 permutations, while the ANOVA with linear model was done with 10.000 iterations by residual variance resampling. **: p < 0.01; ‐: p > 0.1. t‐test ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ 942 0‐10 cm (n = 4) 1893 (±1016) 0.41 (±0.11) 0.57 (±0.13) 65 (±14) 0.93 (±0.02) 3.28 (±0.25) 0.79 (±0.04) 94 (±37) 943 Soil sampling 0‐20 cm (n = 3) 1390 (±617) 0.28 (±0.06) 0.44 (±0.09) 58 (±10) 0.93 (±0.01) 3.25 (±0.13) 0.80 (±0.02) 79 (±20) depth t‐test ‐ ‐ ‐ ‐ ‐ ‐ ‐ ‐ 931 932 Table 2: Alpha‐diversity analysis of the soil micro‐eukaryote community (average ± standard deviation). Sequence counts: total sequence counts; S: 933 Sample richness (family); D‐1: Simpson index; H: Shannon index; E: Shannon’s Equitability index; C: Chao‐1 index. The effect of tested variables on the 934 indices was assessed with a bilateral Student´s t‐test. **: p < 0.01; *: 0.01 < p < 0.05; . : 0.05 < p < 0.1; ‐: p > 0.05. 935 936 944 Reference Datasets OTUs (n) Families (n) Unclassified (n) Families (%) Eukaryota 467 162 305 34.69 Fungi 359 204 155 56.82 Cercozoa 187 128 59 68.45 Lentendu et al., 2014 950 Photoautotrophs 117 27 90 23.08 Chryso/Synurophyceae 56 47 9 83.93 Kinetoplastida 46 43 3 93.48 This study Soil metagenomic DNA 217 165 52 76.04 945 946 Table 4: Comparison of OTU richness of sequences obtained from 18S rRNA gene amplicon and metagenomic data sets. 947

948

949

951 952 953 S1: A) Meteorological records from the Rothamsted research station. The year 2010 was a colder year 954 when compared to the average for 1981‐2010, and also to 2008, 2009 and 2011. Data are publicly available 955 at http://www.rothamsted.ac.uk/aen/ecn/YEARLYSUMMARY.htm. B) 5‐years average temperature records 956 at the Rothamsted research station, from 1880 to 2010. The 5‐years means (dots) show that by the year 957 2000 the mean air temperature was approximately 1°C higher than the long‐term mean (1880 to 1985, 958 green line). http://www.rothamsted.ac.uk/environmental‐change‐network/climate). 959 960 961 970

962 971 963 S2: Comparative richness analysis between LCA and BHA methods. Venn diagrams represent the 972 S3: Generalized heatmap of the most significant taxonomical differences between the LCA and the BHA 964 distribution of shared and unique families and rooted unclassified clusters between A: the LCA and BHA 973 methods. The color key represents the centered and scaled counts of each taxon. Only phylogenetic groups 965 methods, C: the two extraction methods in the case of the BHA, E: the two extraction methods in the case 974 showing significant abundance differences are shown (ANOVA, p < 0.05). The color keys at the beginning of 966 of LCA. The rarefaction curves are showing the family and rooted unclassified clusters diversity detection 975 each row and column are, respectively, showing the sample origins and the taxonomic affiliations of 967 upon sequence recruitment between, B: the LCA and BHA methods, D: the two extraction methods in the 976 displayed taxa. 968 case of the BHA and F: the two extraction methods in the case of the LCA. 977 969 978 981 S4: List of dominant fungi OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others are 979 982 showing the number of matches in different ribosomal databases. Phylum Class Order Family Counts SILVA SSU SILVA LSU Greengenes RDP M5RNA 983 SILVA SILVA Ascomycota Dothideomycetes Capnodiales Mycosphaerellaceae* 6847 2 3 0 0 5 Phylum Class Order Family Counts Greengenes RDP M5RNA Ascomycota Dothideomycetes Pleosporales Phaeosphaeriaceae* 216 1 1 0 0 2 SSU LSU Ascomycota Dothideomycetes Pleosporales Pleosporaceae* 160 1 0 0 0 1 Chlorophyta Chlorophyceae* Chaetopeltidales Chaetopeltidaceae 50 0 9 0 0 9 Ascomycota Dothideomycetes Pleosporales* Unclassified 349 3 1 0 0 4 Chlorophyta Chlorophyceae Chlamydomonadales Chlamydomonadaceae* 65 0 3 0 0 3 Ascomycota Eurotiomycetes Eurotiales Trichocomaceae* 1876 8 9 2 0 21 Chlorophyta Chlorophyceae Chlamydomonadales* Volvocaceae 105 0 6 0 0 6 Ascomycota Eurotiomycetes* Unclassified ‐ 1228 25 31 2 0 60 Chlorophyta Mamiellophyceae* Mamiellales Mamiellaceae 178 1 3 0 0 4 Ascomycota Eurotiomycetes Onygenales Ajellomycetaceae* 118 3 9 0 0 12 Chlorophyta Trebouxiophyceae Chlorellales Chlorellaceae* 139 3 0 1 1 3 Ascomycota Eurotiomycetes Onygenales Arthrodermataceae* 150 1 3 0 0 4 Rhodophyta Bangiophyceae Cyanidiales Cyanidiaceae* 180 0 1 0 0 1 Ascomycota Eurotiomycetes Onygenales* Unclassified 489 7 19 0 0 26 Streptophyta Bryopsida Funariales Funariaceae* 335 1 1 0 0 2 Ascomycota Leotiomycetes Helotiales Sclerotiniaceae* 1326 4 6 0 0 10 Bacillariophyta Bacillariophyceae Naviculales Phaeodactylaceae 56 0 0 0 0 0 Ascomycota Sordariomycetes Glomerellales* ‐ 244 21 0 0 26 25 Bacillariophyta Coscinodiscophyceae* Thalassiosirales Thalassiosiraceae 45 0 1 0 0 1 Ascomycota Sordariomycetes Hypocreales Clavicipitaceae* 723 8 5 0 0 12 984 Ascomycota Sordariomycetes Hypocreales* Unclassified 136 15 19 0 0 30 985 S5: List of dominant archaeplastids OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the Ascomycota Sordariomycetes Hypocreales Nectriaceae 499 0 0 0 0 0 986 others are showing the number of matches in different ribosomal databases. Ascomycota Sordariomycetes Sordariales Chaetomiaceae* 351 0 2 0 0 2 987 Ascomycota Sordariomycetes Sordariales* Lasiosphaeriaceae 178 0 70 0 0 70 Phylum Class Order Family Counts SILVA SSU SILVA LSU Greengenes RDP M5RNA Ascomycota Sordariomycetes Sordariales Sordariaceae* 405 0 68 0 0 68 Choanozoa Choanoflagellatea Choanoflagellida Salpingoecidae 51 0 0 0 0 0 Ascomycota Sordariomycetes* Unclassified ‐ 766 28 90 0 26 128 Cnidaria Anthozoa Actiniaria* Edwardsiidae 54 1 0 0 0 1 Basidiomycota* Unclassified ‐ ‐ 435 62 126 0 0 187 Cnidaria Hydrozoa Hydroida Hydridae* 102 4 0 0 0 4 Basidiomycota Agaricomycetes Agaricales* Unclassified 1440 34 71 0 0 105 Filasterea Capsaspora Unclassified ‐ 238 0 0 0 0 0 Basidiomycota Agaricomycetes Agaricales Marasmiaceae* 226 0 1 0 0 1 Nematoda Chromadorea* Unclassified ‐ 93 13 20 0 0 33 Basidiomycota Agaricomycetes Agaricales Psathyrellaceae* 115 1 2 0 0 3 Nematoda Rhabditida Rhabditidae* ‐ 108 2 7 0 0 9 Basidiomycota Agaricomycetes Agaricales Tricholomataceae* 813 10 27 0 0 37 Nematoda Spirurida Onchocercidae* ‐ 27 0 3 0 0 3 Basidiomycota Agaricomycetes* Unclassified ‐ 483 52 114 0 0 165 Nematoda* Trichocephalida Trichinellidae ‐ 19 22 23 0 0 45 Basidiomycota Agaricomycetes Polyporales Coriolaceae* 108 2 6 0 0 8 Placozoa ‐ ‐ ‐ 43 0 1 0 0 1 Basidiomycota Tremellomycetes Tremellales Tremellaceae* 306 3 5 0 0 8 Rotifera Bdelloidea Adinetida Adinetidae* 53 0 1 0 0 1 980 Rotifera Bdelloidea Philodinida Philodinidae* 14 1 0 0 0 1 988 989 S6: List of dominant micro‐metazoan OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the Amoebozoa Tubulinea Euamoebida Hartmannellidae 27 0 0 0 0 0 990 others are showing the number of matches in different ribosomal databases. Mycetozoa Dictyostelia Dictyosteliida Unclassified 185 0 0 0 0 0 991 Mycetozoa Myxogastria Physariida Physaraceae 5 0 0 0 0 0 SILVA SILVA 996 SAR Phylum Class Order Family Counts Greengenes RDP M5RNA SSU LSU 997 S8: List of dominant amoebozoa OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others Alveolata Apicomplexa Aconoidasida Haemosporida* Unclassified 12 0 6 0 6 0 998 are showing the number of matches in different ribosomal databases. Alveolata Apicomplexa* Unclassified ‐ ‐ 11 31 16 0 0 47 999 Alveolata Apicomplexa Conoidasida Eimeriorina Sarcocystidae* 31 1 1 0 0 2 1000 Alveolata Apicomplexa Conoidasida Eucoccidiorida* Unclassified 13 5 9 0 0 14 1001 Alveolata Apicomplexa Conoidasida Cryptosporidium Cryptosporidiidae* 10 4 0 0 0 4 SILVA SILVA Phylum Class Order Family Counts Greengenes RDP M5RNA Alveolata Apicomplexa Gregarinia Eugregarinida* ‐ 13 19 6 0 0 25 SSU LSU Alveolata Chromerida Perkinsea Perkinsida Perkinsidae* 21 0 2 0 0 2 Euglenozoa* Kinetoplastida Trypanosomatida Trypanosomatidae 66 1 2 0 0 3 Alveolata Ciliophora Intramacronucleata Hymenostomatida Tetrahymenidae* 14 0 4 0 0 4 Metamonada Parabasalia Trichomonadida Trichomonadidae* 29 10 0 0 0 10 Alveolata Ciliophora Intramacronucleata* Peniculia Parameciidae 45 1 9 0 0 10 Percolozoa Heterolobosea Schizopyrenida Vahlkampfiidae* 289 4 0 0 0 4 Alveolata Dinoflagellata Dinophyceae Amphidinium Kareniaceae* 12 0 1 0 0 1 1002 Rhizaria Cercozoa Monadofilosa Euglyphida Paulinellidae* 4 0 2 0 0 2 1003 S9: List of dominant excavates OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others Rhizaria Cercozoa Plasmodiophoromycetes Plasmodiophorida Plasmodiophoridae* 5 3 6 0 0 8 1004 are showing the number of matches in different ribosomal databases. Rhizaria Cercozoa Sarcomonadea Cercomonadida* ‐ 3 2 3 0 0 5 1005 Stramenopiles* Heterokontophyta Pelagophyceae Pelagomonadales Unclassified 237 2 7 0 0 9 1006 Stramenopiles Heterokontophyta Phaeophyceae Ectocarpales Ectocarpaceae* 27 0 2 0 0 2 Stramenopiles Heterokontophyta Oomycota Peronosporales* Peronosporomycetes 236 0 1 0 0 1 Stramenopiles Heterokontophyta Blastocystae Blastocystida Blastocystidae 29 0 0 0 0 0 992 993 S7: List of dominant SAR OTUs found in this study. The count column is indicating the number of metagenomic sequence hits while the others are 994 showing the number of matches in different ribosomal databases. 995 SILVA SILVA Phylum Class Order Family Counts Greengenes RDP M5RNA SSU LSU Amoebozoa Incertae sedis Incertae sedis Breviatea* 1 1 0 0 0 1 Amoebozoa Archamoebae Entamoebida Entamoebidae 4 0 0 0 0 0 Amoebozoa Archamoebae Mastigamoebida Mastigamoebaea 2 0 0 0 0 0 Amoebozoa Discosea Centramoebida Acanthamoebidae* 1 1 0 0 0 1 1007 1014 1008

1015 1016 1017 1009 1018 S11: Cluster dendrograms of the micro‐eukaryote metagenomic profiles. Dendrogram A is showing the 1010 1019 distribution of the profiles according the Jaccard similarity index, while dendrogram shows the distribution 1011 S10: Venn diagrams of the micro‐eukaryote community richness after LCA annotation. 1020 according to the Bray‐Curtis dissimilarity index. 1012 1021 1013 1022 1023

1024 1025 S12: Barcharts of the significant changes attributed to the sampling A) year and B) soil depth (Linear model, 1026 ANOVA, p < 0.05). 1027

Poster Session

QUICK DESIGN GUIDE Control of fungal growth on wood coang with sustainable QUICK TIPS (--THIS SECTION DOES NOT PRINT--) (--THIS SECTION DOES NOT PRINT--) non-synthec anmicrobial agents This PowerPoint 2007 template produces a 70cm x This PowerPoint template requires basic PowerPoint 100cm professional poster. You can use it to create your Frederik Lyduch, Jonas Stenbæk & Bo Jensen (version 2007 or newer) skills. Below is a list of research poster and save valuable time placing titles, Secon for Microbiology, Instute of Biology, commonly asked questions specific to this template. subtitles, text, and graphics. If you are using an older version of PowerPoint some University of Copenhagen, Denmark template features may not work properly. We provide a series of online tutorials that will guide you through the poster design process and answer your ABSTRACT Template FAQs poster production questions. Table 2: Screening of FFA on Aureobasidium pullulans As part of a greater exterior wood coating project, the objective of this study was to Inoculated with spores Inoculated with mycelium reveal if selected emulsifiers and Aliphatic Short Chained Free Fatty Acids (FFA) possess Conc. → 0,01M 0,002M 0,01M + B 0,002M + B 0,01M 0,002M 0,01M + B 0,002M + B FFA↓ Verifying the quality of your graphics To view our template tutorials, go online to potential as substitutes for traditional synthetic fungicides in exterior wood coating. C3 - + + + + + + + C5 - + - - - + + + Go to the VIEW menu and click on ZOOM to set your PosterPresentations.com and click on HELP DESK. The results revealed that the antimicrobial effect of the tested emulsifiers, didn’t C6 + E - - - - - + + + preferred magnification. This template is at 100% the depend on specific pH-values, as it did for the FFA’s. On the other hand the FFA’s was better as antimicrobial at the right pH-values. C7 + E - - - - - + + + size of the final poster. All text and graphics will be When you are ready to print your poster, go online to C8 + E ------+ + printed at 100% their size. To see what your poster will PosterPresentations.com. OBJECTIVES C9 + E - - - - - + - + look like when printed, set the zoom to 100% and The general theory that this study is based on, states that FFA’s in their protonated C10 + E - - - - + + + + form can penetrate fungal cell membranes due to their low polarity, and then Table 3: Screening of FFA on Alternaria alternata evaluate the quality of all your graphics before you subsequently become deprotonated once inside the cell. The deprotonation is then Inoculated with spores Inoculated with mycelium Need Assistance? Call us at 1.866.649.3004 Conc. → 0,01M 0,002M 0,01M + B 0,002M + B 0,01M 0,002M 0,01M + B 0,002M + B submit your poster for printing. believed to lead to a drop in intracellular pH (figure 1). In order to maintain the FFA↓ desired pH-value of the fungal cell, the cells proton pump is activated, resulting in C3 - + + + + + + + exhaustion of ATP-reserves (figure 2). Furthermore the free radicals of the FFA’s are C5 - + + + - + + + Object Placeholders believed to react with internal macromolecules, causing cell damage [1]. By adding C6 + E - - + + - + + + FFA’s, which are weak acids, to a formulation the pH will drop to around 3-4 depending Modifying the layout on the specific FFA and the amount added. The low polarity of the acid comes from the C7 + E - - + + - + + + This template has four different Using the placeholders protonated form, which constitutes around ≈99 % at pH 3-4. C8 + E - - - + - + + + column layouts. Right-click your To add text, click inside a placeholder on the poster and C9 + E - - - - - + + +

Figure 1: Protonated FFA’s can C10 + E - - - - + + + + mouse on the background and type or paste your text. To move a placeholder, click it penetrate the fungal cell membrane, with a subsequent Table 4: Screening of FFA on Ulocladium atrum click on LAYOUT to see the layout once (to select it). Place your cursor on its frame, and Inoculated with spores Inoculated with mycelium deprotonation inside the cell. Conc. → 0,01M 0,002M 0,01M + B 0,002M + B 0,01M 0,002M 0,01M + B 0,002M + B options. The columns in the provided layouts are fixed your cursor will change to this symbol . Click once FFA↓ and drag it to a new location where you can resize it. C3 - + + + + + + + and cannot be moved but advanced users can modify C5 - - + + - + + + any layout by going to VIEW and then SLIDE MASTER. C6 + E - - + + - + + + Section Header placeholder Figure 2: In order to maintain the C7 + E - - + + - + + + cells desired intracellular pH- Click and drag this preformatted section header C8 + E - - - + - + + + value, the protons is pumped out Importing text and graphics from external sources placeholder to the poster area to add another section at the expense of ATP. C9 + E - - - - + + + + header. Use section headers to separate topics or C10 + E - - - - + + + + TEXT: Paste or type your text into a pre-existing concepts within your presentation. placeholder or drag in a new placeholder from the left side of the template. Move it anywhere as needed. Table 2, 3 & 4: Screening of FFA added to PDA on the molds Aureobasidium pullulans The binders in the coating, currently under development, only work at pH 8-9. By (Table 2), Alternaria alternata (Table 3), and Ulocladium atrum (Table 4). Left side adding a pH 8 phosphate buffer to a formulation containing FFA’s, the pH is right for of the table is the effect on spores, right side is the effect on mycelium. B = 0,5M pH the binders to work, but the FFA’s have now switched from being around ≈99 % 8 phosphate buffer, E = emulsifier. PHOTOS: Drag in a picture placeholder, size it first, protonated to around ≈99 % deprotonated (Figure 3). Text placeholder click in it and insert a photo from the menu.

Move this preformatted text placeholder to the poster CONCLUSIONS Figure 3: The objective of the TABLES: You can copy and paste a table from an to add a new body of text. study was to test if deprotonated • FFA efficiently inhibits growth of both spores and mycelium of the tested molds in its FFA’s still have an antimicrobial non-polar, protonated form. external document onto this poster template. To adjust effect. the way the text fits within the cells of a table that has • Relative to the protonated form, FFA doesn’t inhibit growth of either spores or been pasted, right-click on the table, click FORMAT mycelium in its deprotonated form very well. SHAPE then click on TEXT BOX and change the • The tested emulsifier, E09 was the most efficient antimicrobial agent. It wasn’t INTERNAL MARGIN values to 0.25. Picture placeholder MATERIALS AND METHODS particularly affected by changes in pH and inhibited growth of spores, but didn’t Move this graphic placeholder onto your poster, size it inhibit growth of mycelium. first, and then click it to add a picture to the poster. The fungi used for the experiments was Aureobasidium pullulans. Alternaria alternata and Ulocladium atrum. They are all known to cause damage on exterior wood products • A molecule with a combination of low polarity and immunity to changes in pH, might Modifying the color scheme [2]. They have been cultivated on potato dextrose agar (PDA), added FFA´s and prove promising as an antimicrobial agent. emulsifiers, and stored in darkness at 25 °C. The diameter of the colonies was To change the color scheme of this template go to the measured once per day, and the radial growth rate was calculated. REFERENCES DESIGN menu and click on COLORS. You can choose from the provided color combinations or create your own. RESULTS [1] R. Coleman, V. Yang, B. Woodward, P. Lebow, C. Clausen. Efficacy of Fatty Acid Table51:5Screening5of5emulsifiers Chemistry: Candidate Mold and Decay Fungicides, One Hundred Sixth Annual Meeting of Aureobasidium)pullulans Table 1: The effect of the selected Inoculated5with5spores Inoculated5with5mycelium the American Wood Protection Association (2010) p. 287-297. Emulsifier 300ppm 500ppm 1000ppm 1500ppm 300ppm 500ppm 1000ppm 1500ppm - Growth inhibited emulsifiers added to PDA, on spores and [2] M. A. Shirakawa, C. C. Gaylarde, P. M. Gaylarde, V. John, W, Gambale. Fungal E09 0 0 0 0 + + + + mycelium. colonization and succession on newly painted buildings and the effect of biocide, FEMS + Growth of spores E10 + 0 + + + + + +

+ Growth of mycelium E11 + 0 0 0 + + + + Microbiology Ecology 39 (2002) p. 165-173. Alternaria)alternata Inoculated5with5spores Inoculated5with5mycelium Emulsifier 300ppm 500ppm 1000ppm 1500ppm 300ppm 500ppm 1000ppm 1500ppm

E09 0 0 0 0 + + + +

E10 + + + + + + + + CONTACT ! Growth(inhibited E11 + 0 0 0 + + + +

Ulocladium)atrum + Growth(of(spores Inoculated5with5spores Inoculated5with5mycelium Frederik Lyduch: [email protected] Emulsifier 300ppm 500ppm 1000ppm 1500ppm 300ppm 500ppm 1000ppm 1500ppm Jonas Stenbæk: [email protected] + Growth(of(mycelium E09 0 0 0 0 + + + + © 2013 PosterPresentaons.com E10 + + + + + + + + Bo Jensen: [email protected] Explanaon for table 1, 2, 3 & 4 E11 + 0 0 0 + + + + 2117 Fourth Street , Unit C

Berkeley CA 94710 Student discounts are available on our Facebook page.

RESEARCH POSTER PRESENTATION DESIGN © 2012 [email protected] Go to PosterPresentations.com and click on the FB icon. www.PosterPresentations.com

Co-Authorship Statement