Heterologous expression and purification of Cellulose synthase-like B (NbCslB)

(1)

IN

DEGREE PROJECT BIOTECHNOLOGY, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2020,

Heterologous expression

and purification of

Cellulose

Nicotiana benthamiana

synthase-like B (NbCslB)

OLIVIA STÅHL

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF ENGINEERING SCIENCES IN CHEMISTRY, BIOTECHNOLOGY AND HEALTH

(2)

Heterologous expression and purification of Nicotiana

benthamiana Cellulose synthase-like B (NbCslB)

Degree Project in Biotechnology Second cycle, 30 credits

Author: Olivia Ståhl

Supervisor: Sara Díaz-Moreno Examiner: Yves Hsieh

KTH Royal Institute of Technology

School of Engineering Sciences in Chemistry, Biotechnology and Health SE-100 44 Stockholm, Sweden

(3)

1

Abstract

Hemicelluloses are synthesized by proteins encoded by genes from the cellulose synthase gene superfamily. One subgroup of this gene family is the cellulose synthase-like B, which is largely uncharacterized and unexplored. The common model organism Nicotiana benthamiana has one such gene in its genome, NbCslB, encoding a membrane protein. The expression of this gene has previously been studied in vivo, but in order to study the protein in vitro a viable solubilization and purification protocol is required. This study evaluated the use of the detergent n-Dodecyl β-D-maltoside (DDM) for solubilization, followed by purification using immobilized metal ion affinity chromatography (IMAC), and thereafter reconstitution of the protein into proteoliposomes. SDS-PAGE as well as Western blot analyses showed that the purification was successful and provided a pure sample of protein. Throughout the analyses performed, an anti-FLAG antibody was discovered to bind well to the protein, and thereby be especially useful for analysis. An activity assay was performed on the purified protein, to characterize its function and evaluate whether the protein had maintained its activity and conformation after the steps of purification and reconstitution. No activity could be detected in the enzymatic assay, which indicated that the purification protocol may have been too rough on the protein, that the reconstitution was not successful, or that the assay conditions were not optimal. These results can be used as a base for future research, where the protocols for solubilization, purification, and reconstitution should be further refined in order to obtain an end result where the purified protein is active. When an active and pure protein sample is achieved, it will be possible to perform further attempts at characterizing the function of the protein using enzymatic activity assays. Additionally, the results showed that the choice of antibody can be crucial for proper analysis of this protein.

Keywords: Biotechnology, Cellulose Synthase-Like B, Glycoscience, Glycosyltransferase, Hemicellulose, Nicotiana benthamiana

(4)

2

Sammanfattning

Hemicellulosa syntetiseras av proteiner vars gener återfinns i genfamiljen cellulosasyntas. En undergrupp till denna genfamilj är cellulosasyntasliknande B, en grupp som till stor del är okarakteriserad och outforskad. Den vanliga modellorganismen Nicotiana benthamiana har en sådan gen i sitt genom, NbCslB, som kodar för ett membranprotein. Hur denna gen uttrycks har tidigare studerats in vivo, men for att kunna studera proteinet in vitro krävs ett hållbart protokoll för solubilisering och rening. Denna studie utvärderade användningen av lösningsmedlet n-Dodecyl β-D-maltoside (DDM) för solubilisering, följt av rening med immobiliserad metalljon-affinitetskromatografi (IMAC), och efter det rekonstitution av proteinet till proteoliposomer. SDS-PAGE och Western blot analyser visade att reningen var lyckad, och att ett rent proteinprov erhållits. När analyserna genomfördes upptäcktes att en anti-FLAG antikropp band särskilt väl till proteinet, och därmed var mycket användbar vid analys. En aktivitetsanalys genomfördes med det renade proteinet för att karakterisera dess funktion och utvärdera huruvida proteinet hade bevarat sin aktivitet och konformation efter rening och rekonstitution. Ingen aktivitet kunde detekteras i den enzymatiska aktivitetsanalysen, vilket indikerade att reningen eventuellt var för hård mot proteinet, alternativt att rekonstitutionen inte var lyckad, eller att förhållandena för analysen inte var optimala. Dessa resultat kan användas som en bas för framtida forskning om proteinet, där protokollen för solubilisering, rening och rekonstitution bör vidareutvecklas för att uppnå ett slutresultat där det renade proteinet är aktivt. När ett aktivt och rent proteinprov uppnåtts är det möjligt att genomföra ytterligare försök att karakterisera proteinets funktion med enzymatiska aktivitetsanalyser. Resultaten visade också att valet av antikropp kan vara avgörande för att ordentligt kunna analysera detta protein.

Nyckelord: Bioteknik, Cellulosasyntasliknande B, Glykosyltransferas, Glykovetenskap, Hemicellulosa, Nicotiana benthamiana

(5)

3

1. Introduction

When purifying proteins, it is most often desirable for the protein to maintain its native conformation and activity. The preserved activity of the protein is essential for studying its functions and characteristics. This is especially difficult to achieve in the case of membrane proteins, where detergents are often needed in order to solubilize the protein, before using any common purification protocol. Detergents are amphipathic molecules with a hydrophilic head and a hydrophobic tail, and can be used to disintegrate the lipid bilayer that is the membrane, and incorporate the membrane lipids in micelles. The same principle applies to the proteins, which have an at least partly hydrophobic surface, which can be incorporated in micelles as well. However, some membrane proteins lose their conformation and activity when they lose contact with their native lipids [1]. The native lipids that are essential for the function of the membrane protein need to, where applicable, be incorporated in the detergent micelles together with the protein itself. It is therefore important to make sure that the detergent and the methods used for solubilization and purification are suitable for the specific protein in each case. After purification it is common to implement a reconstitution step, where the detergent is removed and replaced with lipids of appropriate origin, to form proteoliposomes containing the protein that can be used for future storage or functional assays [2].

This thesis aims to evaluate one strategy for solubilization and subsequent purification of a Nicotiana benthamiana glycosyltransferase (NbCslB), using the detergent n-Dodecyl β-D- maltoside (DDM) (see molecular structure in Figure 1), for solubilization, and purification involving the detergent LysoFos Choline Ether 14 (LFCE14) (see molecular structure in Figure 2), followed by reconstitution of the protein by exchanging the detergent to proteoliposomes [3]. The molecular structures for two of the most abundant yeast lipids in the yeast total lipid extract used can be seen in Figures 3 and 4. Prior to this, the initial goal of the thesis is to screen for expression of NbCslB in two different strains of Saccharomyces cerevisiae (FGY217 and LoGSA), using two different genetic constructs. Based on analysis using Western blot, the strain and construct best suited for expression of the protein are then selected for further experiments. The final goal of this thesis is to use the sample of pure protein received from purification to perform an in vitro enzymatic assay, in order to characterize the catalytic activity of the protein and investigate what substrate NbCslB utilizes. The main objective of this thesis is to contribute to and develop the current knowledge about the NbCslB protein, and facilitate future research on this, as well as other similar proteins.

Figure 1. The molecular structure of n-Dodecyl β-D- maltoside (DDM). Image from [4]

Figure 2. The molecular structure of LysoFos Choline Ether 14 (LFCE14). Image from [5]

(7)

5 Figure 3. The molecular structure of phosphatidylcholine, one of the most abundant yeast lipids in the yeast total lipid extract used. Image from [6]

Figure 4. The molecular structure of phosphoinositol, one of the most abundant yeast lipids in the yeast total lipid extract used. Image from[7]

1.1. Background

The cell wall is a very important structural component in plant cells, which provides strength and resilience. The cell wall contains several different polysaccharide components that all contribute to its function. Cellulose is the most abundant polysaccharide, but in between the cellulose microfibrils of the cell wall are also, among many other components, hemicelluloses [3, 8]. Hemicelluloses are the second most abundant compound in the cell wall, present in the vast majority of land living plants [9], and consist of (1,3)- and (1,4)-β-linked sugar monomer residue backbones, using monomers such as glucose, mannose, xylose, and more. The hemicellulose molecule contains a mix of different monomers, and the composition of sugar monomers can vary greatly between organisms and organs [8, 9]. In contrast to the linear cellulose molecule, hemicelluloses are branched molecules, its branches consisting of sugar monomers and polymers. These branches give the hemicellulose molecule a noncrystalline structure. Hemicelluloses serve several purposes in the cell wall, among other functions they act as a support matrix for the cellulose, and absorb large amounts of water to maintain moisture [10].

The genes for proteins that synthesize cellulose are found in what is called the cellulose synthase (CesA) gene superfamily, which also includes the genes that encode proteins for synthesizing hemicellulose backbones, found in cellulose synthase-like (Csl) groups [8].

Other proteins are involved in the synthesis of the hemicellulose branches. Several of the Csl sub-groups are largely uncharacterized, and their functions are unknown. The sub-group CslB is suspected to be specific to dicots, and the N. benthamiana genome has been found to only contain one CslB gene. The phylogenetically closely related sub-group CslH is in turn suspected to be specific for monocots [8, 11]. CslHs are known to play a role in the synthesis of mixed-linkage β-glucans [11]. The functions of CslBs, however, have been largely unknown, and have therefore been investigated by researchers at the Division of Glycoscience at KTH. The expression of the N. benthamiana CslB gene in the different organs of the plant was studied, and the by far highest expression was found in fully developed flowers, followed by a moderate amount in the plant leaves. The vast majority of the expression in the flower was found in the anthers, but expression was also found in several of the other organs of the flower. Díaz-Moreno et al. [11] found that when the NbCslB gene was silenced, the phenotype showed an altered cell wall composition in the leaves, and a lack of flowers, as

(8)

6

well as poorly developed growth. Coupled to the silencing of the NbCslB gene, changes in the expression of other genes could also be observed. Several genes, for example related to cellulose biosynthesis and cell wall remodeling, were up-regulated. In contrast, other genes related to for example cell wall remodeling and lignin biosynthesis were found to be down- regulated, together with many others [11].

The NbCslB protein has been predicted to have two transmembrane domains at its N- terminal, a large cytosolic domain including a glycosyltransferase domain, followed by six more transmembrane domains at its C-terminal, as seen in Figure 5 [11]. The protein has a molecular weight of approximately 85 kDa [12], but because it is a membrane protein it is expected to migrate faster and thus appear smaller on gels than its theoretical molecular weight. Typically, this phenomenon will cause a membrane protein to migrate up to 20 % faster [13], which implies that the NbCslB protein will be found at approximately 70 kDa in SDS-PAGE gels and Western blot.

Figure 5. Predicted domain organization of NbCslB, as shown by Díaz-Moreno et al.[11]

Because NbCslB is a membrane protein, the methods for purification often require detergents.

However, the use of detergents for purifying membrane proteins makes it difficult to keep the proteins in an active form. The lack of appropriate purification protocols for keeping the protein in active form complicates the process of characterizing its catalytic activity [14].

Most methods for processing proteins are adapted to soluble proteins, and membrane proteins therefore first have to be solubilized in an aqueous solution, before these methods can be applied. Detergents are commonly used to achieve this solubilization, but this can be harsh on the proteins and can cause them to denature or aggregate, or lose contact with native lipids without which they cannot function. The difficulties linked to purification of membrane proteins are a main reason to why these proteins are generally less studied than soluble proteins, even though they play vital roles in all cells [2]. A common method for finding an appropriate purification process for a membrane protein is to screen many different detergents in several concentrations to determine the most suitable circumstances for the protein in

(9)

7

question. An in vitro assay can be applied to monitor the protein activity, either before or after reconstitution of the protein, depending on the specific case. This can show whether the protein has maintained its structure through the purification process [1].

A previous study [15] showed that using the detergent DDM for solubilization treatment of membrane proteins provided a much higher yield, and enabled solubilization of a wider variety of membrane proteins, than several other screened detergents. The same study [15]

showed that the presence of detergent LFCE14, instead of DDM, during purification and elution could prevent the protein from aggregating, which was an issue in the presence of DDM, while also helping to stabilize the protein. DDM and LFCE14 will therefore be used in this study.

The natural abundance of membrane proteins is in most cases low, which contributes to the low understanding of these proteins, in comparison to soluble proteins. It is also difficult to overexpress membrane proteins, as this can result in high levels of inactive protein or protein that lacks the proper posttranslational modifications, among other issues. In later years it has become easier to work with membrane proteins, as many generic protocols for membrane protein expression and purification have been developed. These can oftentimes be used as a base, which researchers can then optimize to fit their own needs [2]. S. cerevisiae is a common heterologous overexpression system for producing eukaryotic membrane proteins, which is otherwise often difficult [16]. S. cerevisiae will therefore be used to express the NbCslB membrane protein in this study. Two different gene constructs will be expressed and evaluated in two strains of S. cerevisiae, namely FGY217 and LoGSA. The strain FGY217 carries a deletion in the pep4 gene, meaning that it does not express the protease encoded by this gene [16]. Because proteases degrade proteins, it is advantageous for the expression host to carry this deletion, to minimize possible deterioration of the product. The LoGSA strain has a reduced glucan synthase activity, and was generated to minimize background levels of glucosyltransferase activity, when used as an expression system [14]. While FGY217 is a more commonly used strain, LoGSA has the potential of giving clearer results when assessing the expressed CslB protein activity.

The two gene constructs that will be used are pYES-DEST52-NbCslB-6xHis and pDD- FLAG-NbCslB-12xHis. Both include His-tags, which can be utilized for detection by antibody, and for purification using affinity chromatography. It is common to use a tag consisting of six histidine units, but a longer tag can be useful especially when working with membrane proteins to get a stronger binding to the chromatography resin, which can in turn improve the purification yield [2]. When purifying membrane proteins using immobilized metal ion affinity chromatography (IMAC), detergent molecules associated to the protein can obstruct the histidine tag, and thereby weaken the binding to the resin. Also for this reason, a longer histidine tag can be useful to improve binding [2]. IMAC purification takes advantage of the affinity between the His-tag attached to the protein and Ni²⁺ ions. The metal ions are immobilized on the surface of the chromatography resin, and tagged proteins selectively bind to it as they are loaded onto the chromatography column. Other proteins can be washed out, while bound proteins are eluted using imidazole [2].

(10)

8

2. Materials and Methods

2.1. Screening of Expression

Transformants were screened for expression using the methods described by Drew et al. [16]

with a few modifications. All buffer and media recipes are found in Appendix 3. The cells were cultivated in 50 mL aerated tubes with 10 mL -URA medium containing 2 % glucose overnight in 30 °C and 230 rpm. New tubes containing 10 mL -URA media with 0.1 % glucose were then inoculated to an OD600 of 0.12. The cells were incubated at 30 °C and 230 rpm until they reached an OD600 of 0.6, at which point expression was induced by adding 1 mL of 20 % galactose, to a final concentration of 2 %. Approximately 20-22 hours after induction the cells were harvested by centrifugation at 3000 × g for five minutes to form a pellet, which was then resuspended in 200 μL of YSB buffer. Of this 200 μL were saved for further processing. Glass beads and an additional 500 μL of YSB buffer were added to the cell suspension, to a total volume of 1 mL, and the cells were then lysed using a Tissue Lyzer at 30 Hz for two minutes at a time, repeated four to five times. Cells were kept on ice between runs. The tubes were briefly centrifuged and 500 μL of supernatant were transferred to new tubes. 500 μL of YSB buffer were added to the remaining unbroken cells, and these were once again processed using the Tissue Lyzer in the same manner. The tubes were briefly centrifuged, and 500 μL of supernatant were transferred to the same tubes as after the first round. These tubes containing 1 mL of cell breakage were then centrifuged at 17,000 × g for 90 minutes, and the pellet was resuspended in 30-50 μL of YSB buffer.

A Bradford assay (see protocol in Appendix 2) was performed on the samples in order to determine their protein concentration. After this a Western blot (see chapter 2.4.) was performed to screen for expression of the desired protein.

2.2. Expression, Solubilization, and Purification of NbCslB

Expression, solubilization, and purification were carried out using the methods described by Purushotham et al. [3], adapted to fit the specific circumstances. All buffer and media recipes are found in Appendix 3. Conical 250 mL flasks containing 25 mL of -URA medium with 2

% glucose were inoculated from a fresh plate of cells. The cells were cultivated overnight at 30 °C and 230 rpm. They were then diluted to an OD600 of 0.12 in 2.5 L flasks containing 1 L of -URA medium with 0.1 % glucose. The cells were then cultivated to an OD600 of 0.6, at 30

°C and 230 rpm. Expression was induced by adding 20 % galactose to a final concentration of 2 %, and the cells were then incubated at 30 °C and 230 rpm. Approximately 22 hours after induction the cells were harvested by centrifugation at 3,500 × g for five minutes. The cell pellet was then stored at -80 °C.

The harvested cells were resuspended in 25 mL CRB buffer per liter of original cell culture, in 50 mL tubes. The cells were homogenized using a French press at 1000 psi for three passes.

The lysate was centrifuged for ten minutes at 10,000 × g at 4 °C in order to remove unbroken cells and cell debris. The supernatant containing the membranes was filtered through one

(11)

9

layer of Miracloth (Millipore, Merck KGaA #475855) and collected in ultracentrifuge tubes.

The supernatant was then centrifuged for two hours at 150,000 × g at 4 °C in an ultracentrifuge, to pellet the membranes. The membrane pellet was resuspended in 8 mL of membrane resuspension buffer (MRB) per liter of original cell culture. The protein concentration was determined using a Bradford assay (see protocol in Appendix 2), and was then adjusted to 3 mg/mL with MRB. N-Dodecyl β-D-maltoside (DDM) powder was added to 2 w/v% final concentration and incubated in an ice bath with gentle agitation for one hour.

Insoluble material was removed by centrifugation in an ultracentrifuge at 150,000 × g for 30 minutes at 4 °C. The supernatant was then incubated overnight with 5 mL of preequilibrated Ni metal affinity resin (Ni-NTA His●Bind Resin, Millipore, Merck KGaA #70666) at 4 °C.

The resin was packed into a gravity flow column and was then washed with 60 mL (12 bed volumes) of MRB containing 1 mM of LysoFos Choline Ether 14 (LFCE14) instead of DDM, and 30 mM imidazole. The protein was then eluted with 300 mM imidazole. An Amicon Ultra 15 mL centrifugal filter (Millipore, Merck KGaA #UFC905024) was used to concentrate the eluted protein in a centrifuge at 4,000 × g for approximately one hour. Using the same centrifugal filter, a buffer exchange was done using 15 mL MRB containing 1 mM LFCE14. The centrifugal tube was centrifuged at 3,400 × g for approximately one hour. The remaining sample volume was then adjusted to 1 mL using MRB containing 1 mM LFCE14.

Samples were taken at multiple times during the purification, and were saved for analysis with SDS-PAGE and Western blot (see chapter 2.4.).

2.3. NbCslB Reconstitution into Proteoliposomes

Reconstitution of the protein was done according to the methods described by Purushotham et al. [3]. To reconstitute the protein, 1 mL of the concentrated protein was mixed with solubilized yeast total lipid extract at a final concentration of 4 mg/mL, and was incubated for one hour at 4 °C. The protein-lipid mix was added to a 15 mL tube containing 0.6 mL of prewashed beads (Bio-Beads SM-2 Absorbent Media, Bio-Rad #152-3920), and was incubated for one hour at 4 °C with gentle end-over-end rotation. The supernatant was then transferred to a new 15 mL tube containing 0.6 mL of prewashed beads, and incubated for one hour under the same conditions. The supernatant containing the proteoliposomes was then stored in aliquots at -80 °C. The reconstituted protein sample was analyzed using SDS-PAGE, Western blot (see chapter 2.4.), and mass spectrometry (MS) (see chapter 2.5.).

2.4. Western Blot Analysis

SDS-PAGE was carried out using either Mini-protean TGX gel (Bio-Rad #456-1083) or Novex WedgeWell 10 % Tris-Glycine Gel (Thermo Fisher Scientific #XP00100BOX) at 150 V for approximately 60 minutes. From each sample, including any positive and negative controls, 20-30 μg of protein were loaded on the gel, as well as 5 μL of PageRuler Plus Prestained Protein Ladder (Thermo Scientific #26619).

(12)

10

An Immun-blot PVDF membrane (Bio-Rad #1620177) was submerged in methanol until it became translucent. It was then equilibrated in transfer buffer (Western blot buffer recipes are found in Appendix 4) for three minutes. All parts of the blot were equilibrated in transfer buffer, and were then mounted on the Trans-Blot SD, semi-dry transfer cell (Bio-Rad). First, a pre-soaked piece of extra thick blot filter paper (Bio-Rad #1703965) was placed on the transfer cell surface, and the equilibrated membrane was then carefully placed on top of it.

The gel, which had been equilibrated in transfer buffer for 10-20 minutes, was placed next, followed by a second piece of pre-soaked extra thick blot paper. The semi-dry cell was then run at 25 V for 45 minutes. When finished, the blot was disassembled, and the membrane was washed in water three times for five minutes each. Blocking of the membrane was performed using 5 % non-fat milk in TBS with 0.05 % TWEEN 20 at room temperature for at least one hour, or overnight.

The next morning the membrane was washed in TBS with 0.05 % TWEEN 20 two times for five minutes each. It was then incubated with the proper antibody, titered to 1:5000 or 1:3000 in TBS with 0.05 % TWEEN 20, at room temperature for at least one hour. Three different antibodies were used in this study; 6x-His Tag Monoclonal Antibody (HIS.H8) HRP (Thermo Scientific #MA1-21315-HRP), Anti-His(C-term)-HRP (Thermo Scientific #46-0707), and Monoclonal ANTI-FLAG M2-Peroxidase (HRP) antibody produced in mouse (Sigma-Aldrich

#A8592). The membrane was washed with TBS with 0.05 % TWEEN 20 two to four more times for five minutes each. The membrane was then treated with Amersham ECL Prime Western Blotting Detection Reagent (GE Healthcare #RPN2232).

This was done by mixing solutions A (luminol) and B (peroxide) at a ratio of 1:1 to a working solution of 4 mL per membrane. Excess wash buffer was drained from the membrane, which was placed on a plastic sheet. The detection reagent was added to the membrane, completely covering it, and was left to incubate for five minutes. Excess detection reagent was drained from the membrane. The plastic sheet with the membrane was placed in a CCD camera set to detect chemiluminescence, where results could then be evaluated.

2.5. Mass Spectrometry

40 μL of the purified, reconstituted, and concentrated sample of NbCslB protein was loaded on a gel (Novex WedgeWell 10 % Tris-Glycine Gel, Thermo Fisher Scientific

#XP00100BOX), which was run and stained with Coomassie Brilliant Blue. After destaining of the gel the band that corresponded to the NbCslB protein was cut out and divided into millimeter size pieces, and placed in a 1.5 mL tube.

In-gel trypsin digestion was done in preparation for mass spectrometry. The digestion and mass spectrometry were both performed by another lab worker. The in-gel trypsin digestion was done according to a protocol by Srivastava [17]. The gel pieces were incubated with 200 μL of 50 % 0.1 M ammonium carbonate and 50 % acetonitrile at pH 10 for 60 minutes at 37

°C. The solution was discarded and the gel pieces were incubated with 100 % acetonitrile for five minutes to dehydrate. The gel pieces were then incubated with 100 μL of a solution with 10 mM DTT in 50 mM ammonium bicarbonate for 30 minutes at 60 °C. The supernatant was

(13)

11

discarded, and 30 μL of 1 % iodoethanol solution in 50 mM ammonium bicarbonate was added and incubated for 15 minutes at 37 °C. Once more 100 % acetonitrile was used to dehydrate the gel pieces. After that 30 μL of 10 ng/μL trypsin in 50 mM ammonium bicarbonate was added to the gel pieces, and the tube was kept on ice for 15 minutes. The solution was then removed, and 50 μL of 50 mM ammonium bicarbonate was added and incubated overnight at 37 °C. After incubation the supernatant containing the protein was collected in a clean tube, and additional protein was extracted from the remaining gel pieces with 50 μL formic acid solution, containing 5 % formic acid and 50 % acetonitrile. This was incubated for five minutes at room temperature. The proteins were then dried using a SpeedVac Vacuum Concentrator, and stored at -20 °C before they were analyzed using a reversed-phase liquid chromatography electrospray ionization mass spectrometer (LC-ESI- MS/MS), according to a method described by Leijon et al. [18].

2.6. Enzymatic assay

Assays for cellulose synthase, callose synthase, glucomannan synthase, and xylane synthase were carried out. Master reaction mixes were prepared for the four different assays according to the recipes in Appendix 5. 200 μL reactions were set in 1.5 mL tubes in triplicates for the purified protein, the mock purification, and blank. The reactions were incubated at room temperature overnight. The next morning 400 μL of 100 % ethanol were added to each of the 36 reaction tubes, and the reactions were incubated for two hours at -20 °C. After incubation, the reactions were filtered using a Millipore 1225 sampling manifold and Whatman GF/C glass microfiber filters. This was done by first starting the vacuum, and then rinsing the filters with water and ethanol. The reactions were vortexed and added to the filters. The tubes were washed twice with water, which was also added to the filter. The filter was then washed twice with approximately 4 mL of 70 % ethanol. When the filters had dried the manifold was disassembled and the filters were placed in scintillation tubes, where they were left to dry for two hours. When the filters had dried, 3.5 mL of Ultima Gold F (Perkin Elmer #6013179) was added to each scintillation tube. The tubes were placed in the scintillation counter, and scintillations were counted during one minute.

(14)

12

3. Results

3.1. Screening of Expression

The expression of two gene constructs, pYES-DEST52-NbCslB-6xHis and pDD-FLAG- NbCslB-12xHis (henceforth referred to as pYES and pDD, respectively), was examined using two different strains of S. cerevisiae (FGY217 and LoGSA) as expression hosts. Both constructs were expressed in both strains, and thus four different combinations were evaluated. In addition to each of the combinations initially being evaluated in quadruplicate, positive and negative controls were also included. Cells of the LoGSA strain expressing 8xHis-GFP were used as a positive control of expression, as well as one with a fluorescent tagged protein of approximately 150 kDa, with GFP at its N-terminal. As negative control an empty vector in each of the respective strains was used.

Initial repetitions and optimizations of the Western blot method can be seen in Figures 6-7 and 11-12. The membranes shown in Figure 6 were the first attempts at using this method, and showed little usable results. The membranes shown in Figure 7 were clearer in terms of signal, but the semi-dry transfer was only partly successful, and the images were smudged.

However, the left membrane in Figure 7 had a clear signal and good transfer at the site around approximately 70 kDa where the NbCslB protein should have shown. From the image, no distinction could be seen at this site between samples and negative control, and it could therefore not be shown that the protein was expressed and present in this case.

Figure 6. Western blot membranes with samples from cultivated FGY217 cells with the two gene constructs. The numbered lanes contained the following samples: 1. pDD(1); 2. pDD(2); 3. pYES(1);

4. pYES(2); 5. Positive control (8xHis-GFP); 6. Negative control; 7. pDD(3); 8. pDD(4); 9. pYES(3);

10. pYES(4); 11. Positive control (8xHis-GFP); 12. Negative control. The antibody used was 6x-His Tag Monoclonal Antibody (HIS.H8) HRP (Thermo Scientific #MA1-21315-HRP)

(15)

13

Figure 7. Western blot membranes with samples from cultivated LoGSA cells with the two gene constructs. The numbered lanes contained the following samples: 1. pDD(1); 2. pDD(2); 3. pYES(1);

4. pYES(2); 5. Positive control (8xHis-GFP); 6. Negative control; 7. Positive control (fluorescent tagged protein); 8. pDD(3); 9. pDD(4); 10. pYES(3); 11. pYES(4); 12. Positive control (8xHis-GFP);

13. Negative control; 14. Positive control (fluorescent tagged protein). The antibody used was 6x-His Tag Monoclonal Antibody (HIS.H8) HRP (Thermo Scientific #MA1-21315-HRP)

In order to control that the positive controls were properly expressing their fluorescent proteins, the samples used in Figure 7 were analyzed by looking at the in-gel fluorescence.

This is displayed in Figure 8. Only a very faint signal could be seen from the positive control at size 45 kDa, in lane one, in this case, which indicated a low expression. This low expression could explain why neither the positive control, nor the target protein was visible in Figure 7. In Figure 8, as well as in Figures 9 and 10, one protein band could be seen in all of the samples, at a size of approximately 65 kDa. This corresponded to a fluorescent protein endogenous to the S. cerevisiae cells [16]. A new cultivation was done to see if better expression could be achieved, and the samples were analyzed by looking at the in-gel fluorescence again, the results of which can be seen in Figure 9. Also in Figure 9, the signal at 45 kDa in lane five was faint, and no good expression could be confirmed from the cell cultivation. Because no good expression could be determined from the plated transformants that had thus far been used, fresh newly transformed cells were plated and used going forward. From this point on, only the fluorescent tagged 150 kDa protein was used as a positive control. After a new cultivation and membrane preparation, in-gel fluorescence was used to confirm the expression of the fluorescent tagged protein by the positive control cells.

This can be seen in Figure 10, which confirmed that the freshly plated cells expressed the plasmid gene, by the strong bands at approximately 150 kDa in lanes one and seven.

(16)

14

Figure 8. In-gel fluorescence of samples from cultivated LoGSA cells with the two gene constructs.

The numbered lanes contained the following samples: 1. Positive control (8xHis-GFP); 2. Positive control (fluorescent tagged protein); 3. Negative control; 4. pDD(1); 5. pDD(2); 6. pDD(3); 7.

pYES(1); 8. pYES(2). The band that is visible in all of the samples represents a fluorescent protein endogenous to S. cerevisiae

Figure 9. In-gel fluorescence of samples from cultivated FGY217 cells with only one of the gene constructs. The numbered lanes contained the following samples: 1. pDD(1); 2. pDD(2); 3. pDD(3);

4. Negative control; 5. Positive control (8xHis-GFP); 6. pDD(4); 7. pDD(5); 8. pDD(6). The band that is visible in all of the samples represents a fluorescent protein endogenous to S. cerevisiae

Figure 10. In-gel fluorescence of samples from cultivated FGY217 cells with the two gene constructs.

The numbered lanes contained the following samples: 1. Positive control (fluorescent tagged protein);

2. Negative control; 3. pDD(1); 4. pDD(2); 5. pYES(1); 6. pYES(2); 7. Positive control (fluorescent tagged protein). The band that is visible in all of the samples represents a fluorescent protein endogenous to S. cerevisiae

(17)

15

After expression of the positive controls had been confirmed by the in-gel fluorescence, Western blot was done on the same samples as in Figure 10, using two different antibodies on two separate membranes, to confirm expression of the non-fluorescent tagged NbCslB. Figure 11 displays a membrane with good transfer. Despite the good transfer it was not possible to distinguish the NbCslB protein. However, it was possible to see weak bands slightly above 130 kDa in lanes one and seven, which most likely represented the 150 kDa fluorescent tagged positive control protein. A reason for the NbCslB not being visible could be the high level of unspecific binding of the antibody, which could be seen by the strong signals at the bottom of each lane. A different antibody was used for the second membrane, namely Anti- His(C-term)-HRP (Thermo Scientific #R931-25), which only binds to His-tags on the C- terminal of the protein. This antibody should thus be more specific than the first, although it cannot bind to, and thus not visualize, the positive control, since its His-tag is located at the N-terminal. The result of this can be seen in Figure 12. The transfer was slightly smudged, especially in lanes three and five, but some conclusions could still be drawn. Lanes two and six showed clear bands at approximately 70 kDa, where the negative control showed no signal. This was most likely the NbCslB protein expressed. However, there was still a fair amount of unspecific binding and the entirety of the membrane could not be clearly evaluated.

Figure 11. Western blot with samples from cultivated FGY217 cells with the two gene constructs. The numbered lanes contained the following samples: 1. Positive control (fluorescent tagged protein); 2.

Negative control; 3. pDD(1); 4. pDD(2); 5. pYES(1); 6. pYES(2); 7. Positive control (fluorescent tagged protein). The antibody used was 6x-His Tag Monoclonal Antibody (HIS.H8) HRP (Thermo Scientific #MA1-21315-HRP)

(18)

16

Figure 12. Western blot with samples from cultivated FGY217 cells with the two gene constructs. The numbered lanes contained the following samples: 1. Negative control; 2. pDD(1); 3. pDD(2); 4.

Empty well; 5. pYES(1); 6. pYES(2). The antibody used was Anti-His(C-term)-HRP (Thermo Scientific

#R931-25)

Further attempts were made to produce a clear image that proved expression of the protein.

Another Western blot was carried out using the same samples as in Figures 11 and 12, which had been saved at -20 °C. These protein samples had, however, been degraded, most likely due to repeated thawing and freezing. This could be seen from most of the signal being accumulated at the bottom of the membrane. Fresh samples were produced for the strain FGY217, one for each of the two gene constructs pDD and pYES, and one negative control.

These were analyzed by Western blot, as seen in Figure 13. There it could be observed that the cells transformed with the pDD construct did express the protein, while the signal was not as clear for the pYES construct. The negative control was ambiguous and could be interpreted as having some expression. The third lane was, however, slightly skewed, which might have caused this unclear appearance. From Figures 12 and 13 it was concluded that the FGY217 cells transformed with the pDD construct did in fact express the NbCslB protein, and thus were the most suitable to use for larger scale cultivation and subsequent purification of the NbCslB protein.

(19)

17

Figure 13. Western blot with samples from cultivated FGY217 cells with the two gene constructs. The numbered lanes contained the following samples: 1. pDD; 2. pYES; 3. Negative control. The antibody used was Anti-His(C-term)-HRP (Thermo Scientific #R931-25)

3.2. Solubilization and Purification of NbCslB

One construct and strain was chosen for this step, based on the results of the expression screening. The S. cerevisiae strain FGY217 was used, with the gene construct pDD-FLAG- NbCslB-12xHis (the specific amino acid sequence for this construct can be found in Appendix 1), to express the protein for subsequent purification. A negative control in the form of an empty vector transformant was also included.

Samples were taken at various times during the purification process, as well as after reconstitution of the protein. These samples were analyzed using SDS-PAGE stained with Coomassie Brilliant Blue, which can be seen in Figures 14 and 15. Figure 14 shows samples taken at different stages of IMAC, from flowthrough to elution. One faint band could be observed at approximately 70 kDa in the sixth lane that had no equivalent in the fifth lane.

This most likely represented the expressed NbCslB protein. The other samples on the same gel contained a far larger variety of proteins, which was to be expected, as these samples represented volumes that were discarded, i.e., contained all proteins that did not bind to the resin. Bands could be seen around 70 kDa also for these samples, for both the samples containing the actual enzyme and for their corresponding negative controls. This indicated that the S. cerevisiae cells have some native proteins around this size as well, although these are not capable of binding to the resin, as they could not be seen in the mock purification eluate. In Figure 15 a more prominent band could be seen at 70 kDa in lane six, which represented the purified, concentrated, and reconstituted protein. This showed that the protein was present in the final sample, and also showed a clear distinction to the negative control in lane five, which showed no bands at all. This pointed to the purification being successful, as the only distinction between the two was the band at 70 kDa, which corresponded well to

(20)

18

where the NbCslB enzyme was expected to show. Lanes one through four represented samples that were expected to contain far more contaminations, which was also confirmed by their somewhat smudged appearances. These samples were taken before purification, and therefore contained all the cells’ native proteins.

Figure 14. SDS-PAGE gel 1 stained with Coomassie Brilliant Blue. The numbered lanes contained the following samples from the IMAC purification process: 1. Mock purification flowthrough; 2. Real purification flowthrough; 3. Mock purification first wash; 4. Real purification first wash; 5. Mock purification eluate; 6. Real purification eluate

Figure 15. SDS-PAGE gel 2 stained with Coomassie Brilliant Blue. The numbered lanes contained the following samples: 1. Empty vector solubilized membranes; 2. pDD vector solubilized membranes; 3.

Empty vector after incubation with DDM and centrifugation; 4. pDD vector after incubation with DDM and centrifugation; 5. Mock purified final reconstituted sample; 6. Purified final reconstituted protein sample

The same samples that were analyzed using SDS-PAGE in Figures 14 and 15 were also analyzed using Western blot, utilizing an Anti-His(C-term)-HRP (Thermo Scientific #R931- 25) antibody. However, some unknown factor caused this analysis to fail, and no signal could be seen in either of the two membranes. The transfer seemed successful, based on the ladders being properly displayed on both membranes, as could be seen by the naked eye.

(21)

19

Two additional gels were run, where one was dyed and the protein band at approximately 70 kDa was cut and analyzed using MS, and the other one was used for Western blot with a new type of antibody. Figure 16 displays a result consistent with what could be seen in Figures 14 and 15, i.e., a distinct band coherent with the estimated size of the NbCslB protein. This band was cut and digested for MS analysis, the full results of which can be seen in Appendix 6. The MS gave a protein sequence coverage of 29 %, which is high enough to be considered a confirmation of the protein identity.

Figure 17 shows a Western blot of samples from two points in time, the eluted protein and the reconstituted, concentrated protein, as well as their respective mock-purified negative controls. A different type of antibody was used for this membrane, than had been used before, which targeted the FLAG-tag on the protein instead of the His-tag. This was done because of the several failed and unclear attempts that had previously been done using two different anti- His antibodies. Figure 17 gave a clear signal in both the sample of the eluted protein, and the reconstituted and concentrated sample of NbCslB, and a considerably stronger signal in the latter, as expected. This clearly indicated that the FLAG-tagged protein was present in the samples. It was also clear that no such protein was present in the negative controls, where no signal could be observed for either of the samples. The protein concentration of the final reconstituted and concentrated sample was measured using a Bradford assay (see protocol in Appendix 2). The final sample had a protein concentration of 0.405 μg/μL, which indicated a total process yield of at least 0.8 mg of pure protein per liter of culture.

Figure 16. SDS-PAGE gel stained with Coomassie Brilliant Blue. The numbered lanes contained the following samples: 1. Mock purified final reconstituted sample; 2. Purified final reconstituted protein sample

(22)

20

Figure 17. Western blot of the purified protein samples. The numbered lanes contained the following samples: 1. Mock purification eluate; 2. Real purification eluate; 3. Mock purified final reconstituted sample; 4. Purified final reconstituted protein sample. The antibody used was Monoclonal ANTI- FLAG M2-Peroxidase (HRP) antibody produced in mouse (Sigma-Aldrich #A8592)

3.3. Enzymatic Assay

Enzymatic in vitro assays were done in order to evaluate the activity of the purified protein.

This included assays for cellulose, callose, glucomannan, and xylane synthase activity, the recipes of which are described in Appendix 5. Scintillations for each reaction were counted during one minute in a scintillation counter. In addition to the purified protein, reactions were also carried out for the mock-purified sample, as well as a reaction where only buffer was added, to measure the background signal to use as a blank. The measurements from the enzymatic assays can be found in Appendix 7. No activity could be detected in any of the four assays. There was no significant difference between the scintillation counts for the NbCslB enzyme, the mock-purified sample, and the blank, which indicated that there was no enzymatic activity corresponding to these four assays.

(23)

21

4. Discussion

The first few Western blots (Figures 6-7 and 11) were moderately successful. Several of these lacked a clear signal at the expected location of the protein, and instead had a lot of unspecific binding at the bottom of the membrane. The antibody used for these membranes (6x-His Tag Monoclonal Antibody (HIS.H8) HRP (Thermo Scientific #MA1-21315-HRP)) was rather unspecific, and might have been able to bind to any sequential His. In order to mitigate this issue of unspecificity, another antibody was used for the next membranes, as seen in Figures 12 and 13. This antibody was only able to bind to His-tags on the C-terminal of the protein, which was the location of the His-tag for both the genetic constructs that were expressed and screened. This antibody gave a much clearer result, and gave distinguishable signals from the protein bands expected to be the NbCslB protein. Although this was an improvement, there was still a fair amount of unspecific binding of the antibody, especially at the bottom of the membrane. When analyzing the samples taken during and after purification of the enzyme, this antibody was used again, but rendered no visible results at all. A third antibody was used in a new attempt at identifying the purified protein, this time taking advantage of the FLAG- tag attached to the protein. As could be seen in Figure 17, this attempt was successful, and the protein could clearly be identified.

A possible reason for the anti-His antibodies not providing consistently good results could be that the His-tag on the protein was somehow obscured. It could be that the protein is folded in such a way that the antibody was not able to access the tag and bind to it properly. Difficulties in binding, in combination with a lot of unspecific binding of the antibody, could easily have prevented the protein band from being visible on the Western blot. A slightly more specific binding, as provided by the Anti-His(C-term)-HRP antibody (Thermo Scientific #R931-25), gave a better signal, albeit not very robust over multiple analyses. It is also unknown how the protein conformation is affected by the purification procedure. The steps of the purification themselves could have affected the protein in such a way that the His-tag became inaccessible for the antibody. An example of this could be conformational changes, due to the continuously changing surrounding environment of the protein, where the tag could be folded into the protein, thus making it impossible for an antibody to bind. This would cause it to not visualize the protein properly. However, based on there not being any visible signal at all in the first Western blot performed after purification, using the Anti-His(C-term)-HRP antibody (Thermo Scientific #R931-25), neither for the target protein nor for anything else, it is likely that something else went wrong during the detection of the protein.

For the last Western blot analysis of the purified samples, an anti-FLAG antibody (Monoclonal ANTI-FLAG M2-Peroxidase (HRP) antibody produced in mouse (Sigma- Aldrich #A8592)) was used. This was done based on the previous varying degrees of success of using the His-tag for visualization of the protein. Based on the successful outcome of using this antibody for analysis, the N-terminal FLAG-tag is likely better situated on the protein surface than the His-tag, to allow an antibody to bind. A few faint bands could be observed in Figure 17 in addition to the approximately 70 kDa band. These likely represented both aggregated and degraded protein units, which are likely to be present in a purified sample

(24)

22

along with the proper size protein. Because the analysis in Figure 17 was done on the purified samples, no other proteins were present, and it was therefore not possible to determine whether using the anti-FLAG antibody provided less unspecific binding than the previously used antibodies. Regardless, using the anti-FLAG antibody to detect the NbCslB enzyme in future experiments is recommended, based on the strong binding and clear signal it provided compared to the anti-His antibodies. Although the different antibodies were used under different circumstances, it is reasonable to conclude that the anti-His antibodies were not consistently providing usable results, and using the anti-FLAG antibody in the future could be a safer choice for obtaining clear analyses.

Using an anti-FLAG antibody requires that the expressed protein has an attached FLAG-tag, which is only the case for one of the two gene constructs that were screened. The choice to express the pDD-FLAG-NbCslB-12xHis (pDD) construct over the pYES-DEST52-NbCslB- 6xHis (pYES) construct was based partly on the length of the His-tag that was to be used for IMAC purification. During screening it was indicated that the pDD construct was possibly better expressed than the pYES, judging from Figure 13. This, in combination with the larger size of the His-tag on the pDD construct making it more likely to be successfully purified, motivated the use of the pDD construct for expression and purification. In hindsight, based on the difficulties in getting antibodies to bind to the His-tag, and this indicating that it might be slightly obscured, choosing the longer His-tag appears to have been a good strategy. A shorter His-tag might have been more inaccessible, which could potentially have caused problems already in the IMAC purification, where the His-tag was utilized to bind the protein to the chromatography resin. The pYES construct also lacks the FLAG-tag that later turned out to be very useful for Western blot analysis. Based on the FLAG-tag’s usability during analysis, and the additional length of the His-tag, the pDD would be recommended for future experiments, over the alternative presented here.

For future experiments, it is likely that either of the two strains of S. cerevisiae could be used.

In this study the FGY217 strain was chosen for expression and purification, however, this choice was not extensively supported by experimental results. As the expression of the protein was initially somewhat difficult to confirm, the first strain to successfully display expression was used going forward. As could be seen in Figure 12, the first display of successful expression was seen in FGY217. After that, no additional attempts at expressing the protein in LoGSA were made. The initially suspected advantage of using LoGSA was its lower background levels of glycosyltransferase, which could provide clearer results in an activity assay. However, since the purification was successful in removing contaminants, the real difference between using the FGY217 or the LoGSA strain, was likely minimal. For future experiments it could be interesting to express the NbCslB using the LoGSA strain instead, to see whether this could provide a different result with respect to purity, yield, or activity.

Based on the SDS-PAGE and Western blot results that can be seen in Figures 14-17, as well as the MS results that can be seen in Appendix 6, the solubilization and purification appear to have been successful. A single band could be seen for the purified samples that were analyzed using SDS-PAGE (Figures 14-16), and no contaminations appeared to be present in the samples. It was clear that the final purified sample was more concentrated than the sample

(25)

23

from elution, which was to be expected considering the buffer exchange and concentration steps that followed the elution of the protein. Only a minor amount of protein appeared to be lost during buffer exchange, concentration, and reconstitution, based on the strong bands that were seen in the analyses of the final sample. Based on the calculated protein concentration of the final sample, the yield was roughly 0.8 mg of pure protein per liter of culture. Similar protein yields for other eukaryotic membrane proteins were found by Newstead et al. [13].

The overall protein yield achieved in this study can thus be considered good, although yields can nearly always be further improved. The identity was confirmed by the Western blot analysis seen in Figure 17, where it was verified that the protein band seen in fact also had the FLAG-tag. The IMAC purification step takes advantage of the affinity between the His-tag and Ni²⁺, which heavily indicated that the protein purified by this method also had a His-tag, and thus was the expressed NbCslB protein. Moreover, the MS analysis showed a 29 % protein sequence coverage, which further confirmed the protein identity. The analyses collectively confirmed that the final purified sample was a pure solution of the NbCslB protein. The solubilization thus also appears to have been successful, although it would be interesting to perform more experiments, where for example detergent concentrations could be varied, to investigate whether a higher yield or activity could be achieved.

One aspect that could improve the process and increase the final yield of purified protein is to improve the overexpression of the protein in the S. cerevisiae expression host. Figures 7 and 11-13 gave an overview of the protein expression, and it could be seen that the expressed target protein was not highly abundant in comparison to the cells’ native proteins.

Overexpression of eukaryotic membrane proteins is notoriously difficult, and it is not a trivial task to develop a better expression host than the one used here. Nevertheless, it would be interesting to attempt to achieve a higher expression of the protein than was reached in this study.

The assays that were done to detect the enzymatic activity of the protein did not show any activity for the protein. This lack of activity could have several explanations. One potential explanation is that the protein lost its activity somewhere along the way during purification and reconstitution. It was rather harshly treated, and could easily have lost some aspect essential to the activity during these steps. It could also be that the reconstitution did not fully work as intended. As some membrane proteins require contact with their native lipids for function, the reconstitution can be essential for the protein to be active after solubilization and purification. Activity assays can be used to confirm that a reconstitution has been successful and that the protein is truly in its active form after purification. In this case it may have proven the opposite, i.e., that the reconstitution may have failed. Another possibility is that the enzyme in question has some unknown function that was not tested by the assays, or that the assay conditions were not optimal for the enzyme. It is difficult to determine how and when the protein activity might have been lost, and more experiments would have to be conducted in order to decide this. Evaluating other reconstitution protocols or attempting to optimize the used protocol to achieve a satisfactory level of activity for the protein would be a necessary step forward. It would also be relevant to look closer at the solubilization and purification to see if it can be done more gently in order to better preserve the protein activity.

(26)

24

5. Future perspectives

Due to the prevailing circumstances of the spring of 2020, during the COVID-19 pandemic, much of the lab work intended for this study had to be cut short. One experiment that would likely have been prioritized, if the possibility had been given, was to try solubilization and purification using a saposin-lipoprotein nanoparticle system, called Salipro [19]. As described by Frauenfeld et al. [19] saposin A and lipids can be mixed with membrane proteins to self- assemble to nanoparticles that incorporate the membrane proteins. Frauenfeld et al. deemed this system easy to use and adapt to the user’s specific needs. It would therefore be interesting for whoever may continue this research to start by trying this protocol, and compare it to the results from the solubilization and purification protocol used in this study. Although the protocol involving DDM, LFCE14, and lipids that was used in this study was successful in providing a pure sample of the NbCslB protein, another protocol could provide an even better purity, a better yield, or a higher activity in the purified sample. The Salipro protocol could be a viable alternative, but it would also be interesting to look further into the protocol used in this study. Different concentrations of DDM could be screened, to see which concentration is the optimal for solubilization of this specific protein. It could naturally also be an option to screen other detergents.

The extensive use of Western blot in this study provided some clarity to what antibody should be used when trying to detect the NbCslB protein. As the two anti-His antibodies used provided bad or moderate quality results, there is reason to believe that the His-tag may be obscured in such a way that the antibody is not able to bind properly. The anti-FLAG antibody that was later used provided a much clearer picture of the results achieved, and clearly bound better to the NbCslB protein. This indicates that the FLAG-tag is more accessible to the antibody by being better situated on the protein. A takeaway from this thesis should thus be to utilize the FLAG-tag for analysis of this protein, when possible. Only one of the genetic constructs that were screened in the initial phase of this study included a FLAG- tag, which could in turn constitute an additional argument for choosing to use this construct over the alternative one also in future research on this protein. One aspect that could be of interest in the future is to make a 3D kinetic image (kinemage) of the protein using software such as KiNG, in order to determine some key characteristics of the protein. One aspect to look at could be the location and environment of the His-tag, to determine its theoretical accessibility for potential antibodies and IMAC.

Using detergents to solubilize membrane proteins often causes them to lose their catalytic activity, which can be cumbersome, especially if the activity is to be characterized, or if the protein’s activity is meant to be utilized. Different approaches for solubilizing and purifying membrane proteins could simplify the studying of these proteins for researchers. This study has provided a potentially viable method of solubilization for a specific type of glycosyltransferase, but the results can likely be applied for similar proteins, or slightly adapted to fit them.

Much is still unknown about the proteins that manufacture cell wall components in plants.

The composition of the cell wall plays an important role for the properties of wood based

(27)

25

materials [20]. By characterizing more and more of the proteins responsible for this composition, the opportunity to modify these proteins may arise in the future. Some modifications could potentially enhance desired properties of the cell wall, in order to further improve the quality or productivity of wood based materials, or improve the yield or nutritional values of food crops. Improving food quality, wood material quality, or the amount of product possible to extract, could increase financial profit, have great societal effects, and could also contribute to a lesser need for deforestation. Many other modifications could possibly be made in the future, maybe in order to increase plants’ uptake of CO2, or increase their growth speed. One can imagine many types of modifications that could have positive environmental implications, as well as financial and societal ones. Further research on plant cell wall properties also has great potential to contribute to the understanding of plant diseases, how these affect the plant and possibly how they can be mitigated by making plants more resistant to these diseases. Additionally, it can lead to a better understanding of plant development, where many key aspects are still unknown.

Hopefully this study can constitute a contribution to future research about cellulose synthase- like proteins in general, and NbCslB in particular, done at the Division of Glycoscience at KTH, or elsewhere. There is much more to discover on the subject that this study has touched upon, and further experiments and optimizations can be made. This study has highlighted a few issues that can contribute to a base for future research on the subject, and also provided a plausible solubilization and purification protocol that can potentially be used in future studies.

(28)

26

Acknowledgements

First of all I want to thank my splendid supervisor Sara Díaz-Moreno for being incredibly patient and kind throughout this project. Thank you for guiding me through lab work, writing, and unforeseen circumstances.

A great thank you to the Division of Glycoscience at KTH for having me in your lab, for helping out when machines were acting up, and for helping out with any issues in general.

Thank you to Vaibhav Srivastava for helping out with mass spectrometry analysis, and the associated preparations.

Thank you to my examiner, Yves Hsieh, who made sure that the project stayed on track throughout its course.

Thank you to Ulrika von Otter, who has worked in parallel to me during these months, for being a great friend and colleague, and someone to bounce ideas off of about thesis, lab work, and anything else. Thank you also for sharing some of the tasks in the lab, and for being a good opponent.

Thank you to my family who has believed in me from day one, and never doubted I would graduate from KTH.

Lastly, thank you to my wonderful fiancé for always being there for me, for helping out and giving advice on my thesis even without any particular knowledge in the subject, and for having the patience to spend this spring writing your own thesis by the desk next to mine.

(29)

27

References

[1] Merck KGaA. Solubilization of Membrane Proteins, [Website], Sigma-Aldrich. Accessed:

April 16, 2020. URL: https://www.sigmaaldrich.com/technical- documents/protocols/biology/purifying-challenging-proteins/solubilization.html

[2] GE Healthcare. 2007. Purifying Challenging Proteins – Principles and Methods. GE Healthcare Bio-Sciences AB. URL: https://www.sigmaaldrich.com/content/dam/sigmaaldrich/docs/Sigma-Aldrich/General_Information/1/ge-purifying-challenging-proteins.pdf [3] Purushotham, P., Hyun Cho, S., Díaz-Moreno, S.M., Kumar, M., Nixon, B.T., Bulone, V.

& Zimmer, J. 2016. A single heterologously expressed plant cellulose synthase isoform is sufficient for cellulose microfibril formation in vitro, PNAS, vol. 113, no. 40.

[4] Merck KGaA. n-Dodecyl β-D-maltoside, [Website], Sigma-Aldrich. Accessed: May 18,

2020. URL:

https://www.sigmaaldrich.com/catalog/product/sigma/d4641?lang=en&region=SE&gclid=Cj wKCAjw5Ij2BRBdEiwA0Frc9aPBbEgDzV13MMBcaydDn2fkjqMmNkAKDmC1OdSs1Qm 0yo-tjCHoPRoCdlEQAvD_BwE

[5] Antrace. L414 - LysoFos Choline Ether 14, Anagrade, [Website], Antrace Products LLC.

Accessed: May 18, 2020. URL: https://www.anatrace.com/Products/Lipids/LYSOFOS/L414 [6] Merck KGaA. L-α-Phosphatidylcholine, [Website], Sigma-Aldrich. Accessed: May 18, 2020. URL: https://www.sigmaaldrich.com/catalog/product/sial/44924?lang=en&region=SE [7] Merck KGaA. 18:1 PI(3,4,5)P3, [Website], Sigma-Aldrich. Accessed: May 18, 2020.

URL: https://www.sigmaaldrich.com/catalog/product/avanti/850156p?lang=en&region=SE [8] Little, A., Schwerdt, J.G., Shirley, N.J., Khor, S.F., Neumann, K., O’Donovan, L.A., Lahnstein, J., Collins, H.M., Henderson, M., Fincher, G.B. & Burton, R.A. 2018. Revised Phylogeny of the Cellulose Synthase Gene Superfamily: Insights into Cell Wall Evolution, Plant Physiology, vol. 177, pp.1124-1141.

[9] Li, L., Rowbotham, J.S., Greenwell, H.C., Dyer, P.W. 2013. ‘Chapter 8 - An Introduction to Pyrolysis and Catalytic Pyrolysis: Versatile Techniques for Biomass Conversion’, In Suib, S.L. (Ed), New and Future Developments in Catalysis, Elsevier, pp. 183-184.

[10] Patel, J.P., Parsania, P.H. 2018. ‘3 - Characterization, testing, and reinforcing materials of biodegradable composites’, In Shimpi, N.G. (Ed), Biodegradable and Biocompatible Polymer Composites, Woodhead Publishing, pp. 55-79.

[11] Díaz-Moreno, S.M., Ekengren, S., Malm, E., Xing, X. & Bulone, V. Silencing of the only Cellulose Synthase-like B member in Nicotiana benthamiana affects plant growth, cell wall composition and flower development. 2020, Manuscript in preparation.

[12] The Sequence Manipulation Suite. Protein Molecular Weight. Accessed: April 25, 2020.

URL: https://www.bioinformatics.org/sms/prot_mw.html