1
Expression & affinity analysis of recombinant RX against pathogenic α-synuclein
Isak Simón
Degree project in Biologicals and Immunotherapy, 30 credits Autumn semester 2020
Supervisor: Sofia Stenler Examiner: Sara Mangsbo
Department of Pharmaceutical Biosciences Division for Protein Drug Design
Faculty of Pharmacy
Uppsala University
2
Abstract
Background
In the as of yet uncurable Parkinson´s disease aggregation of α-syn is an accelerator of pathogenesis. Oligomers of α-synuclein is considered to be neurotoxic hence blocking the endocytosis of aggregated α-syn is possibly a way of preventing pathogenesis. With a protein construct of the Receptor X (RX) previously shown to bind α-syn, it can be possible to bind soluble aggregated α-syn and decrease neuron endocytosis.
Aim
The aim of this study was to express, purify and trimerize two different protein constructs of RX to study the binding to α-syn monomers & oligomers and if the proteins have higher affinity to α-syn oligomers.
Methods
In this study two RX constructs was produced with mammalian cell transfection and purified with Strep-Tactin affinity chromatography; D1, D123mut and D123 which affinity to α-syn monomers and oligomers were studied with ELISAs. Indirect ELISAs were optimized and conducted, a competitive ELISA with D123 was tested with poor reliability.
Results
The results show that D1 could not be determined pure enough to examine its α-syn binding ability. D123mut was pure enough for ELISAs but did not show adequate binding to α-syn.
D123 did show binding to α-syn in an indirect ELISA.
Conclusion
The results were not as promising as expected and did not distinctly help strengthen the theory of a recombinant RX protein as a viable drug. Although there is potential, optimization of both protein constructs and methods used is essential for future studies of RX as a therapeutic protein.
3
Table of contents
Abstract ... 2
Background ... 2
Aim ... 2
Methods ... 2
Results ... 2
Conclusion ... 2
Abbreviations ... 4
Keywords ... 4
Populärvetenskaplig sammanfattning ... 5
Introduction ... 6
Background of Parkinson’s disease ... 6
RX as a therapeutic protein ... 8
Aim of study ... 10
Methods ... 10
Design of plasmid ... 10
DNA production ... 11
Plasmid purification ... 12
Cell transfection ... 12
Protein purification ... 13
Purification analysis with SDS-PAGE... 13
Trimerization of RX proteins ... 15
Size Exclusion chromatography ... 17
RX constructs indirect ELISA ... 17
Anti-Strep-tag antibody ELISA ... 19
RX competitive ELISA ... 19
Results ... 20
Purification ... 20
Trimerization of RX proteins ... 26
Indirect ELISA ... 28
Discussion ... 31
Structural differences ... 32
Aggregation of RX ... 32
Quality of α-syn oligomers ... 33
Limits and Strengths ... 33
Future research ... 33
Conclusion ... 35
Acknowledgements ... 36
References ... 37
4
Abbreviations
PD – Parkinson’s disease α-syn – α-synuclein RX – Receptor X
BBB – Blood brain barrier AA – Amino acid
kDa – Kilo Dalton RT – Room temperature CV – Column volumes RPM – Rotations per minute
SDS – Sodium dodecyl sulfate disulfate
ELISA – Enzyme-linked immunosorbent assay PBS – Phosphate buffer saline
TBS – Tris buffer saline
SEC – Size exclusion chromatography HNE – 4-hydroxy-2-nonenal
MWCO – Molecular weight cut-off HRP – Horseradish-peroxidase
Keywords
Parkinson’s disease, α-synuclein, ELISA, Biological drugs, Protein-based drugs
5
Populärvetenskaplig sammanfattning
Den näst mest förekommande neurodegenerative sjukdomen som finns heter Parkinsons sjukdom och är en sjukdom som kraftigt förändrar de drabbades livskvalité till det sämre. Idag finns ingen klockren behandling och en bot är det inte alls på tal omför överskådlig framtid.
Den gamla behandling med Levodopa som används som förstahandsbehandling sedan 1960- talet har endast symptomlindrande effekt vilket gör att patienterna fortsatt blir sämre och sämre.
Med denna bakgrund är det oerhört viktigt att undersöka nya möjligheter för att behandla patienter med denna sjukdom. I denna studie undersökt möjligheten för om ett kroppseget protein RX kan omstruktureras till att användas som möjligt framtida läkemedel mot Parkinsons sjukdom. Tanken är att RX ska binda upp de aggregerade proteinerna i hjärnan kallat α-syn som tros vara anledningen till uppkomst av Parkinsons sjukdom.
Resultaten från denna studie gav inget tydligt besked om möjligheten att motverka uppkomst av Parkinsons sjukdom men potential finns! Ett av de testade RX proteinerna visades binda α- syn. Det gjordes inget test för att se om RX proteinerna binder bättre till aggregerade α-syn än icke-aggregerade vilket behövs för att kunna avgöra möjligheterna för RX som terapeutiskt protein mot Parkinsons sjukdom. Resultaten togs fram genom ELISA som är en metod för att med antikroppsteknik kunna avgöra hur bra ett protein kan binda till ett annat.
För patienter idag så påverkar inte denna studie direkt men med framtida studier på RX mot α-syn och om positiva resultat fås kan en ny grupp av läkemedel möjligen vara på marknaden för nästa generations patienter med Parkinsons sjukdom. Inom neurodegenerativa sjukdomar har biologiska läkemedel inte börjat användas i så stor utsträckning men det är en tidsfråga och när det väl sker kommer det att revolutionera livet för många patienter med bättre effekt och mindre biverkningar.
6
Introduction
Background of Parkinson’s disease
Parkinson’s disease (PD) is next to Alzheimer´s disease the most common neuro degenerative disorder. Prevalence of PD is estimated to 0.3 % for all ages and 1 % for people of at least 60 years of age (1). It is symptomatically characterized by abnormal motor symptoms as
bradykinesia, tremor, and rigidity (2,3). Dopamine secretion due to neuron death leads to absence of function of the dopaminergic pathway causing motoric symptoms. The
characteristic motor symptoms is thought to occur due to down regulation of dopaminergic neurons in substantia nigra and neuron loss in PD is distinctly selective to substantia nigra (4).
Physiologically PD is characterized by dopaminergic cell loss and aggregation of α-syn to insoluble Lewy bodies. Presence of Lewy bodies is a factor for diagnosing all Lewy body diseases, Parkinson’s disease is one of these. (2,5). There is a well examined relationship between PD and lower quality of life where PD mostly affects physical and social factors (6).
There is no definite cure for PD available and the pharmacological treatments have been symptomatic to increase quality of life therefore research in PD treatment is of great importance. Also, in an economic perspective the absence of a cure is in long term more costly for society. Exogenous dopamine in as levodopa combined with carbidopa has been given to patients to counteract the deficiency in endogenous dopamine. Dopamine treatments does not however slow the progression of PD which remains the biggest problem (3,7). With the development of biological drugs with higher target specificity than small molecular drugs, a better treatment for PD can be seen just behind the horizon.
α-syn and its role in PD
α-syn is a 140 aa protein expressed from the SNCA gene and located in presynaptic terminals in the brain, kidney, urinary bladder, bone marrow and lymphoid tissues. It is thought to regulate neurotransmitter release such as calcium release needed for ATP-induced exocytosis (8,9). A schematic view of α-syn is seen in Figure 1. The negatively charged C-terminus of α- syn has been linked to involvement in most interactions of α-syn with other proteins (10,11).
Calcium binding to the C-terminus have in previous studies shown to cause aggregation (11,12).
7
Figure 1. Schematic view of the amino acid sequence of α-syn. The purple part is the amphipathic repeat region, the blue parts are N-terminal repeat sequences of KTKEGV, the white part is the hydrophobic NAC peptide, and
the yellow part is the C-terminus. The most relevant mutations for this study are pointed out with arrows.
Although the complete cause mechanism of aggregated α-syn induced PD is not fully
understood, accumulation of α-synuclein is considered to be one of the main factors impacting PD progression (13). α-syn is a main component of Lewy bodies and is expressed from the SNCA gene which was the first gene associated with PD (13). Point mutations in the SNCA gene have shown correlation to causing PD since mutations in the SNCA gene can cause α- syn aggregation to pre-formed fibrils/oligomers, fibrils, and Lewy bodies (14). Mutational changes in aa sequence change the charge and hydrophobicity of α-syn which changes its ability to aggregate (11,15). Specifically, the mutations A30P, A53T and EK46 have shown to be related to familial Parkinson´s disease and thought to be involved in misfolding of α-syn causing aggregation (14,16). Aggregated α-syn have in several studies shown to cause
neurodegeneration and oligomeric α-syn are considered to be most neurotoxic which has been shown in vitro in several studies while insoluble Lewy bodies possibly is not cytotoxic
(13,17,18). Aggregated α-syn is believed to cause misfolding of monomeric α-syn resulting in more aggregation (19,20). This is believed to explain neuron to neuron cascade effect of aggregated α-syn transmission causing neuronal death.
α-syn transmission and RX
α-syn aggregates are believed to transmit trans-synoptically between neurons and the spread of α-syn oligomers is essential for PD pathogenesis. Thus, blocking neuron transmission of α- syn oligomers should prevent pathogenesis of PD (17,21).
One study suggests transmission of α-syn oligomers is mediated by binding to the cellular protein RX as seen Figure 2. RX consists of four domains, Domain1, Domain2, Domain3 and Domain4. Domain1 has showed best α-syn binding properties and removal of Domain4 does
8
not negatively affect α-syn binding properties of RX. A mutational change has shown to inhibit a regular function of RX which can be used as an advantage to reduce side effects.
Figure 2. Transmission of α-syn in positive (A) and negative (B) neurons and the proposed effect of a RX protein (C). Neuron to neuron spread of pathologic α-syn is delayed in presence of RX protein or in absence of RX gene
when endocytosis of aggregated α-syn is disabled.
It is believed that there are other ways for α-syn transmission than RX mediated endocytosis;
transmission with extracellular vesicles have been shown to contain α-syn indicating that extracellular vesicles can mediate transmission. RX is possibly not the only pathway of α-syn internalization for neurons making RX a less optimal drug target. (22)
RX as a therapeutic protein
Since RX has previously bound α-syn it could when modified act as a therapeutic protein against uptake of aggregated α-syn. Small molecular drugs are less specific than biological drugs, thus a small molecular drug targeted against α-syn would probably target monomeric and oligomeric α-syn. That will decrease the intended effect of binding α-syn oligomers due to the concentration of monomers and increase risk for side effects.
As it is difficult to design a therapeutic protein able to pass the BBB and the cell membrane in neurons, the area where the RX protein is intended to bind the soluble aggregated α-syn is the extracellular space. The RX protein will when present in the extracellular space at adequate
9
concentrations bind α-syn preventing endocytosis thus preventing its neurotoxic effect. The BBB passage is enabled by a transporter which is linked to the RX by the ADA optimized linker v3. The transporter is a single chain fragment binding transferrin on the endothel of the BBB. After binding the protein is transcytosed through the BBB thus entering the brain (23).
One specificity advantage biological drugs have is possibility to bind with avidity. Avidity is commonly referred to as the greater binding strength and synergy achieved when a
multivalent ligand binds to its target via several pharmacophores. Binding specifically to α- syn oligomers is desired and is possible by binding with avidity. In this study the ligand is the RX construct, the target is α-syn, and the pharmacophores on RX protein are the identical aa sequence responsible for α-syn binding. There are several binding sites on an oligomer compared to monomer e.g., if the monomer binding site is aa sequence 140-160 there will only be that one binding site on a monomer but on oligomers several copies of that sequence is present and available to bind in case of no steric obstruction.
By constructing a multivalent protein intended to bind solely oligomeric α-syn, several binding sites on the constructed protein can bind several binding sites on oligomeric α-syn.
When one binding loosens (described by koff in Eq. 1 and Eq. 2), the other bindings may still be intact enabling the disconnected binding sites to rebind faster since the detached binding site on RX constructs are still at close range to the target. This creates in total a stronger binding of the protein to its target. It enables binding of several monomeric targets to one created protein lowering the required dose. The possibility of binding with avidity is the reason for using a SpyTag-SpyCatcher system (further described in Methods) to trimerize RX constructs and acquire better binding properties to α-syn oligomers than monomers (24,25).
kon is the rate constant for forming a RX-α-syn complex and koff is the rate constant for dissociation of that complex. KD is the equilibrium constant used to determine affinity of a drug to its target defining interaction strength seen in Eq. 2 & Eq.3. When at equilibrium, KD
and number of binding sites can be estimated with a Scatchard plot by plotting number of bound ligands (b)/free ligand concentration ([L]) vs number of bound ligands (b) from Eq. 4 where the slope value is equal to -KD and the x-intercept is number of binding sites for the ligand binding molecules (n) (26). If Scatchard plots show non-linearity a Hill plot can be used to estimate KD where Log(v/(1-v) vs Log(Concentration) is plotted. v is fraction bound.
Eq. 1 𝑅𝑋 + α − syn
𝑘𝑜𝑛
⇌ 𝑘𝑜𝑓𝑓
𝑅𝑋 − α − syn
10
Eq. 2 𝐾𝐷 = 𝑘𝑜𝑓𝑓
𝑘𝑜𝑛
Eq. 3 𝐾𝐷 = [𝑅𝑋−α−syn]
[𝑅𝑋]∗[α−syn]
Eq. 4 𝑏
[𝐿]= 𝑛
𝐾𝐷− 𝑏
𝐾𝐷
Aim of study
The aim of this study was to study the affinity to α-syn for two generated recombinant RX proteins with ELISA. Another aim was to study if affinity difference to monomers and oligomers for the generated proteins were in favor of the oligomers. This to achieve
information about recombinant RX proteins as possible novel drugs against PD. Although PD in many cases is not directly fatal, the absence of a cure or adequate treatments strongly motivate studies of novel drugs against PD.
Methods
Design of plasmid
In this study two protein constructs based on RX were designed and purified, one construct only consisting of Domain1, a SpyTag and a Twin-Strep-tag called D1 and the other construct consisting of Domain1, Domain2 Domain3, a SpyTag, a Twin-Strep-tag and the mentioned mutational change called D123mut. An already purified RX construct, D123 was used also used. It consists of Domain1, Domain2 and Domain3 of RX with a Histidine-Tag (His-tag) instead of Strep-tag and contain no trimerization tag. Therefore, D1 and D123 is not able to trimerize by a covalent isopeptide bond to gain avidity. A simplified schematic overview of the RX proteins is seen in Figure 33. Twin-Strep-tag and His-tag are used for purification of the protein constructs in Strep-Tactin® columns while keeping its bioactivity. The twin strep tag (WSHPQFEK-GGGSGGGSGGS-SAWSHPQFEK) contains an internal linker between two binding sites to Strep-Tactin in the column (27), His-tag contain only several histidines.
SpyTag (VPTIVMVDAYKRYK) is used for trimerization of the RX construct to bind with higher avidity to α-syn oligomers. The SpyTag binds to a protein called SpyCatcher rapidly forming a covalent isopeptide bond. In this study two different SpyCatcher proteins of
different SpyCatcher was used to trimerize the RX construct, SpyCatcher-o-tri (20.6 kDa) and TriCatcher-GFP (green fluorescent protein) 60.5 (kDa). The RX constructs are designed in
11
plasmids in Benchling software, which then are ordered to the lab from an external source.
(28)
Figure 3. Schematic view of the 3 RX proteins designed and used in this study. Domain 1-3 are 3 of the 4 domains of RX and the ones shown to bind α-syn. D123mut was the only of the 3 proteins to have the mutation
inhibiting endogenic binding.
DNA production
Transformation is done by the E. coli to absorb the plasmid. The ordered plasmid was
dissolved in 25 µl dH2O and from that 1µl was incubated with 3 µl E. coli cells on ice for 30 min. Then heated to 42 °C for maximum 1 min and cooled on ice for 1 min to induce plasmid transformation. 250 µl warm LB medium was added and centrifuged for 1h at 37 °C 225 rpm.
It was then transferred to LB agar plates and incubated overnight at 37 °C. Expansion of the bacteria is then done to achieve more plasmids to transfect human cells and eventually result in more RX construct expression. A single colony is selected and added to 3 ml LB medium with 350 µl ampicillin diluted 1:1000 in PBS pH 7.4 and incubated overnight at 36 °C
shaking in 200 rpm to ensure bacterial growth of the transformated E. coli cells. It is essential
12
to add appropriate antibiotic to prevent non-relevant bacteria to grow in the medium. The bacteria cells are harvested by centrifuging 5000 x g for 10 min and collecting the pellet.
Plasmid purification
Plasmid purification is used to isolate plasmid DNA from bacterial cell components. Cell transfection requires sterile conditions thus DNA free of bacterial cell components is essential.
The pellet from the cultivated E. coli cells is resuspended in 7 ml resuspension solution to enable lysis and RNA degradation. After 7 ml lysis solution is added, the tube is inverted 4-6 times and incubated at RT for 3 min. 7 ml neutralizing solution is then added and the tube is inverted 5-8 times. 1 ml Endotoxin binding reagent was added, and the tube was inverted 5-8 times. The tube was incubated at RT for 10 min. Endotoxin binding reagent was used to bind endotoxins to avoid endotoxins presence after purification. 4 ml isopropanol was added, and the tube was inverted 5-8 times. The solution was centrifuged at 4500 x g at 10 min (repeated if necessary) and the supernatant was collected to discard the pellet containing cell debris and chromosomal DNA. The supernatant was transferred into a filtration column and centrifuged at 2000 x g for 3 min to filtrate through the plasmids which then was added to a purification column that binds DNA in the plasmid so that all other material than the plasmid DNA runs through the column to be discarded. 8 ml washing solution 1 was added and the column was centrifuged at 3000 x g for 2 min. 16 ml washing solution 2 was added and the tube was centrifuged at 3000 x g for 2 min. After washing the column, the plasmid DNA was eluated first with 1 ml elution buffer and incubated at RT for 2 min, then eluated with 0.5 ml elution buffer and incubated again for 2 min at RT to receive high yield of purificated DNA plasmids.
Cell transfection
Transfection of cells was done to produce the wanted proteins. Cell transfection in cells was done for the plasmid to enter the cells so the cells could express RX constructs. For the protein to achieve the correct post translational modifications to have the proper function of binding α-syn, human cells are required.
The purified plasmids were transfected to human Expi-293 cells using PEI (Polyethylenimine) as per the protocol described in Fang et al (29) but using Invine medium instead of Expi- medium. and incubated for 7-12 days depending on the viability of the cells. The cells were then harvested by centrifuging at 1000g for 10 min at RT and collecting the supernatant followed by centrifuging at 7500g for 60 min at 4 C. The supernatant was then filtered
13
through a 0,22µm filter and stored in -20 C until purification. Two transfections at different time were done, one for D1 & D123mut Sept and another for D123mut Nov.
Protein purification
Twin-Strep-tag purification was done to separate the RX construct from other proteins expressed in the cell medium. The produced protein from the harvested Expi-293 cells was incubated with BioLock to block natural biotin, loaded on to a pre-washed Strep-Tactin®
column by running the cell medium through the column. Retardation of RX proteins occur when Strep-tag binds Strep-Tactin in the column. Samples from medium and flow through were saved as controls for purity analysis with Coomassie staining. The loaded column was applied to an already washed ÄKTA start system (Cytiva) and was washed with 10 CVs of 5 ml 100 mM Tris/HCl, pH 8.0; 150 mM NaCl; 1 mM EDTA as washing buffer at a speed of 5ml/min to wash away other cellular products such as proteases. The protein constructs were then eluated with 100 mM Tris/HCl, pH 8.0; 150 mM NaCl; 1 mM EDTA; 50 mM biotin as elution buffer. The eluated proteins were collected in falcon 50 ml tubes as fractions of 5 ml until the observed UV was balanced at baseline. The column was regenerated by running 4 CVs of 10mM NaOH and equilibrating with 10 CVs washing buffer.
Concentration of the fractions was analyzed at 280 nm on a Nanodrop ND-1000
Spectrophotometer (NanoDrop Technologies) and if needed the fractions were concentrated and pooled. Concentration was made by transferring the fractions to Amicon® Ultra-15 Centrifugal Filter Unit (Merck Millipore) 10 kDa cutoff tubes and centrifuged at 4000g for 10 min. Both the protein in elution buffer and the flow through was analyzed in the nanodrop.
This procedure was done until an absorbance of approximately 0.5 was shown in the nanodrop. Desalting and buffer exchange to PBS was made by transferring the protein
construct fractions to separate 5 ml activated Zeba™ Spin Desalting Columns, 7K MWCO, 5 mL and centrifuged at 1000g for 2 min.
Purification analysis with SDS-PAGE
SDS-PAGE gel electrophoresis was used for Coomassie staining and western blot. Coomassie staining was done to determine the purify of each sample. Western blot analysis was used to assess the yield of the samples compared to the medium sample.
For both Coomassie staining and western blot the proteins are run through a SDS gel with added voltage in a tank connected to a machine with incoming voltage. Since DNA is negatively charged it will run to the positive charged side of the gel (the anode) according to
14
charge and size of the DNA of the protein. Before transfer of protein to the wells in the gels, SDS sample buffer and PBS is added. SDS is a surfactant that makes the proteins added to the gel similarly charged thus the SDS sample buffer makes all the proteins equally negatively charged meaning that the only property differentiating the running distance for the proteins is size (30). A Bolt™ MES SDS Running Buffer (Thermo Fisher Scientific) is added to the setup around the gels for the electricity to run through. After 1-2h of running the machine forcing the electricity a separation of proteins has been made according to size. With a known reference ladder containing several proteins in different sizes in kDa, the band of the produced protein construct can be determined if the protein construct size is known. The samples can be run under reduced and non-reduced conditions. By adding beta-mercaptoethanol as a reducing agent the cysteine disulfide bridges are reduced meaning that the aggregated proteins become separated. Reduced conditions are used to be able to better detect the proteins since if
aggregated the bands might then not appear at expected size on the gels. Purity can be visualized and estimated by the strength of the protein constructs band compared to all other bands for the relevant well. For Coomassie staining, Pageblue protein staining solution (Thermo Fisher Scientific) is added to color the gel to better visualize the separation of DNA and proteins. (30–32)
The Coomassie staining gel was used to estimate purity of the fractions by dividing the colour intensity in Image studio of the band at expected construct kDa with the total colour intensity of all bands for the relevant fraction.
For western blot, the gel from the electrophoresis is put on a transfer membrane with filter papers and sponges on both sides to protect the connection between the gel and the membrane intact. This setup is put in a tank with Bolt™ Transfer Buffer (Thermo Fisher Scientific) and added voltage to enable transfer of the proteins from the gel to the membrane. The membrane is then blocked in blocking buffer of 5 % non-fat dry milk in tris-buffered saline tween TBS- Tween 0.1 %. Primary antibody and enzyme-conjugated secondary antibody are applied to the membrane to bind the targeted protein. When a chemiluminescent HRP substrate is added to the membrane, the fluorescence can be measured of where in the membrane and how much of the targeted protein/DNA there is. Western blots are used for a more sensitive way of
analyzing proteins and DNA than Coomassie staining although it is more time consuming.
(31,32)
15
0.5 µg of protein for western blot and 1 µg of protein for Coomassie staining was loaded to the wells for the fraction samples to Bolt™ 4 to 12 %, Bis-Tris, 1.0 mm, Mini Protein Gel, 15-well (Thermo Fisher Scientific) gels. Transfection medium, Flow through from loading the column, wash from washing the column and cell pellet from transfection together with the eluted fractions from purification was each added to a well in the gel. The volume added to the wells was calculated with the concentrations for each protein. All samples were mixed with 4x bolt sample buffer as a part of 25 % and if needed PBS was used to dilute the protein sample to a total volume of minimum 15 µl depending on the samples´ concentration. For medium, wash, flow through and pellet 30 µl was diluted to 40 µl that was added to the wells.
For the reduced samples, beta-mercaptoethanol 10X (Thermo Fisher Scientific) was added to each sample as a part of 10 %. The gels for Coomassie staining were further developed according to the Coomassie blue staining protocol in Appendix. Transfer from gel to membrane for western blot was made at 100 V for 2 h at 4 °C. Blocking was done for 2 h, primary StrepMAB-Classic 2-1507-001(IBA Lifesciences) and secondary Goat Anti-Mouse IgG Antibody, HRP conjugate (Sigma-Aldrich) antibody was incubated with the membrane for 1 h each and washing between each step was made with TBS-T 0.1 % 3x 10 min. The ECL substrate was prepared from Novex™ ECL Chemiluminescent Substrate Reagent Kit (Thermo Fisher Scientific) and added just before signal development in Odyssey® Fc Dual- Mode Imaging System (LI-COR Biosciences). The signal was analyzed in Image studio Software. For D123mut Nov, a primary monoclonal rabbit anti-RX antibody was used with a Donkey anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Plus 800 (Thermo Fisher Scientific). The anti-RX antibody is specific to domain 1 and domain 2 and can therefore if no binding to the RX proteins is shown give information that the proteins are not fully intact.
Yield from purification was calculated in western blot by dividing the colour intensity of the band for the fractions with the colour intensity of the medium while taking different loading sizes and different concentrations into account.
Trimerization of RX proteins
Trimerization of the constructs was done to enable better binding in form of avidity to α-syn.
Specifically, trimerization was done to get better binding to α-syn oligomers to deviate to the binding to monomers. The trimerization technique of SpyTag-SpyCatcher system is based on sequence tags found in Streptococcus pyogenes known to create spontaneous intramolecular isopeptide bonds. In this study the SpyTag is the RX proteins containing one part required
16
sequence adapted from Streptococcus pyogenes and the SpyCatcher containing the other required sequence. When incubating SpyTag proteins with SpyCatcher proteins the binding sites of the SpyTag and SpyCatcher will bind spontaneously creating the covalent isopeptide bond stabilizing the fused protein of SpyTag & SpyCatcher (33). By having the SpyCatcher as a trimerization platform, the SpyTag protein can be trimerized after binding to the SpyCatcher (28).
After centrifugation of thawed SpyCatchers, the calculated volume of SpyTag proteins and SpyCatcher proteins and PBS was mixed and incubated in RT for 10 min. A molar ratio of 3:1 of SpyTag and SpyCatcher was used since 1 SpyCatcher protein trimerize 3 SpyTag protein.
The result of the trimerization was analyzed with western blot and Coomassie blue staining on NativePAGETM 4-16% Bis-Tris Gels (Thermo Fisher Scientific) that does not contain SDS thus not disbanding aggregates. Instead of Bolt™ MES SDS Running Buffer, NativePAGETM Running Buffer is used with NativePAGETM Cathode Buffer Additive. Signal development for western blot was made for both the trimerization protein and the RX protein. It was made twice since the two proteins have different purification tags. It was also necessary to
determine what bands represented only the trimerization proteins and what bands represented only the RX construct in the lanes for the trimerized protein. For the trimerization proteins that contain a His-tag, a mouse anti-His-tag-HRP conjugated antibody was used. For RX, the proteins containing a strep-tag, StrepMAB-Classic 2-1507-001(IBA Lifesciences) was used as primary and Goat Anti-Mouse IgG Antibody, HRP conjugate (Sigma-Aldrich) was used as secondary. Size exclusion chromatography (SEC) was done after the trimerization to isolate non-aggregated products saving all eluated proteins and up concentrating the later fractions.
The SEC method is described in further detail below.
α-syn oligomerization
Oligomerization of α-syn was done to get oligomeric α-syn from the ordered α-syn monomers. The created oligomers were required for the ELISAs conducted to be able to determine if the RX protein could bind α-syn oligomers thus oligomerization was essential for the purpose of the study.
As described by Almandoz-Gil et al. (34), by adding reactive aldehydes as HNE or 4-oxo-2- nonenal (ONE) stable oligomers can be formed from monomers for use to mimic in vivo oligomers characteristics. These reactive aldehydes covalently modify the structure of α-syn to oligomers rich in beta-sheet structure. In this study, HNE is used instead of ONE to get a
17
better mix of molecular weight of the oligomers to mimic the in vivo environment. HNE and α-syn monomers are added in an Eppendorf tube, incubated for 72h at 37°C and run through a desalting column. Creation of α-syn oligomers is essential to study the RX constructs oligomer affinity in competitive ELISA.
Human Alpha-synuclein monomer RP-001 (Proteos) is dissolved to 500 µl 2mg/ml in phosphate buffer pH 7.4. 465 µl of that solution was added to an Eppendorf tube on ice with 35µl 10 mg/ml HNE (BioNordika). The tube was then incubated at 37°C for 72h and buffer exchange was made in Zeba™ Spin Desalting Columns, 7K MWCO, 5 mL. To determine oligomerization results, a western blot was conducted.
SEC was used to separate created oligomers and possible monomers that did not form to oligomers. Eluated fractions was analyzed by western blot and Coomassie staining.
Size Exclusion chromatography 1
SEC was used to separate regular monomeric from aggregated RX constructs since
aggregated RX constructs might block binding sites complicating binding to α-syn. In SEC, a LC system is used with a column containing small pores only small molecules can enter while the larger molecules cannot thus, bigger molecules will elute faster than smaller molecules.
For all SEC separations of RX proteins, HNE induced oligomers trimerized RX proteins, a Superdex™ 200 Increase 5/150 GL (Cytiva) column was used in ÄKTA purifier 100 (Cytiva) system with PBS as buffer at a constant flow rate of 0.6 ml/ml. Due to a large volume,
D123mut was added to the system not via the sample loop but with the initial tube for adding buffer, the oligomers and trimerized RX proteins were loaded via the sample loop. The
eluated fractions of approximately 0.5 ml per fraction was collected in 1.5 ml Eppendorf tubes and the later fractions containing less aggregates were analyzed and further used. The samples injected to the sample loop were injected according to the column manufacturer’s
recommendations.
RX constructs indirect ELISA
The idea of ELISA is to achieve knowledge about protein binding. A plate containing several wells are used, in this study the plates contain 96 wells. By coating the wells with a protein antigen or antibody, adding another protein and an antibody conjugated with an enzyme that in the presence of a specific substrate catalyzes a reaction producing a colour change for the initially transparent samples. By measuring light absorption, that colour change be quantified with optical density (OD) in a spectrophotometer. The OD is direct proportional to the
18
binding of the proteins as described by Eq. 5 where greater protein binding leads to more colorful samples which have lower transmitted optical intensity (I) since less light passes through darker samples resulting in greater OD values.
Eq. 5 𝑂𝐷 = 𝑙𝑜𝑔10 (𝐼0
𝐼)
With this technique it is possible to know if a constructed possible therapeutic protein binds to its intended target. There are several different ELISA methods regularly used such as direct ELISA, indirect ELISA, sandwich ELISA and competitive ELISA but in this study indirect and competitive ELISA are the ones used. For both indirect and competitive ELISA, the target protein is coated on the surface of the wells and the possible therapeutic protein is added and incubated in the wells, though for competitive ELISA another protein is added as a competitor that may affect the binding of the possible therapeutic protein to the coated
protein. (35,36)
Before proceeding to competitive ELISA, indirect ELISA was used to determine if the produced RX protein binds a α-syn. By doing an indirect ELISA first, material could be saved. The indirect ELISAs were coated with a solution consisting of mostly α-syn oligomers and a minority of α-syn monomers hence the binding studied was to both monomers and oligomers.
For the indirect ELISAs 50 µl 500 ng/ml in PBS HNE induced α-syn oligomers was coated in each well in a high binding Costar half-area 96 well plate (Corning) overnight shaking at 4 °C.
The plate was washed 5 times with PBS/Tween 0.05 % after every following step. Blocking of the wells is necessary to avoid unspecific binding of RX to the surface of the wells and was made by adding 150µl ELISA blocking buffer containing PBS/BSA 1 % to each well and the plate was left shaking at RT for 2h. RX protein was added as a dilution series starting at 1000 nM diluted 1:5 in 7 steps in incubation buffer containing 0.1 % FBS, 0.05 % Tween in PBS.
Syn-O2 antibody was added in a dilution series starting at 1nM diluted 1:5 in 7 steps in incubation buffer. Anti-Human SNCA Therapeutic Antibody (Syn-O2) TAB-0748CLV (Creative Biolabs) was used as a positive control and blank as a negative control. The plate was incubated shaking for 1h at RT before adding primary StrepMAB-Classic 2-1507- 001(IBA Lifesciences) diluted 1:2 000 to all wells containing D123mut. The dilution of anti- strep antibody was determined by an optimization ELISA seen in Appendix Figure A4.
19
Secondary antibody Goat Anti-Mouse IgG Antibody, HRP conjugate (Sigma-Aldrich) diluted 1:10 000 was added and the plate was incubated while shaking for 2 h at RT. 50 µl TMB substrate was added to each well for the enzymatic HRP reaction to happen and the plate was protected from light to avoid degradation of the light sensitive TMB substrate. 50 µl 1M sulfuric acid was added to each well to stop the enzymatic reaction when an expected blue colour gradient could be seen in the wells. The OD of the wells was at wavelength 450 nm analyzed in a FLUOstar Omega plate reader (BMG Labtech) The OD values were blank corrected with the average OD of all 8 wells used as blanks and compared with the positive control Syn-O2 to estimate the binding of D123mut to α-syn. The plate layout and
concentrations used in ELISA-1 is seen in Table A1 in Appendix. To later optimize the concentrations of the RX constructs an optimization ELISA was conducted with three different concentration series for RX F6 but with the same plate layout and Syn-O2 concentration.
The indirect ELISA setup was used with the α-syn oligomer coated to the plate surface, soluble RX constructs incubated with α-syn monomers, a primary StrepMAB-Classic 2-1507- 001(IBA Lifesciences) and a secondary Goat Anti-Mouse IgG Antibody, HRP conjugate (Sigma-Aldrich) added to the well to detect OD. With this setup the light absorption could be measured at 450 nm to determine the OD and affinity of the constructs to their target.
Anti-Strep-tag antibody ELISA
An indirect ELISA was done to determine the optimal dilution of StrepMAB-Classic antibody to be used for the RX indirect ELISA and RX competitive ELISA. D123mut and another protein P101 containing a Strep-tag aa sequence was coated overnight at 4 °C on a high binding Costar half-area 96 well plate (Corning) in duplicates to prevent possible occasional errors. After blocking as in the indirect RX ELISA, anti-Strep-tag antibody was incubated in the wells of the plate shaking for 2 h at RT. StrepMAB-Classic 2-1507-001(IBA Lifesciences) is incubated as a capture antibody a primary antibody is not needed. Secondary antibody, TMB substrate, stop solution and data analysis was made as in the RX indirect ELISA.
RX competitive ELISA
Competitive ELISA is done to see how the constructs bind to oligomers versus monomers. It is essential to know if the RX constructs can bind to both monomers and oligomers or not since they do not have the exact same functions in the human brain.
20
Oligomers are coated at the surface of the plate and monomers incubated with RX constructs are added to the plates. Competitive binding affinities can be studied by OD in a
spectrophotometer. It is essential to know if the constructs do not have higher affinity for oligomers since oligomers are more important to bind to stop cell to cell transmission to prevent further development of PD.
α-syn oligomers were coated to the surface of a high binding Costar half-area 96 well plate (Corning) overnight shaking at 4 °C. The plate was blocked and incubated for 2 h at RT.
Serial dilution of α-syn monomers/oligomers was added to the wells together with RX construct at a constant concentration to a total volume of 100 µl added to each well in a high binding Costar half-area 96 well plate (Corning) and pre-incubated for 1.5 h shaking at RT.
After preincubation of monomers/oligomers and RX construct and blocking the pre-incubated solution was added to the coated plate and incubated for 15 min at RT. Addition of primary antibody, secondary antibody, TMB substrate, stop solution and data analysis was made as in the indirect ELISA described earlier.
Results
Purification
The purification of the RX constructs showed different results for D1, D123mut Sept and D123mut Nov. D1 did only show minimal expression in the eluated fractions and no expression in medium, flow through, wash or pellet thus yield could not be calculated.
D123mut Sept expression and yield is seen in table 1. In the Coomassie staining gels seen in Figure 44, bands can be seen at ~50 kDa for D123mut for reduced and non-reduced conditions though for D1, bands at ~35 kDa is only present under non-reduced conditions indicating the bands do not represent D1 or that D1 has intra cysteine interactions creating unwanted
disulfide bridges. The proteins however are not expected to contain unpaired cysteine residues if the protein is fully intact since two cysteines form a disulfide bond by binding to each other and with an even number of cysteines in a protein no single cysteine is left as a free residue unbound to another cysteine (37). Further investigation of D1 were discontinued due to bad protein expression.
21
Figure 4. Reduced (left) and Non-reduced (right) images of Coomassie staining gels. Expected sizes of D1 and D123mut are 20.9 and 40.4 kDa, respectively. Medium is the cell medium used for transfection, Flow through is
the solution collected after loading the medium on to the purification column, Wash is the solution collected after washing the loaded column, Pellet is the centrifuged medium pellet. Bands believed to be D123mut are
marked with a red circle and bands believed to be D1 are marked with a yellow circle.
Due to high protein expression in the wash fraction and low expression in the fractions compared to medium and wash seen in Figure 55 it was believed that during purification, some amount of RX protein was eluated while washing. The wash was then concentrated with Amicon® Ultra-15 Centrifugal Filter Unit (Merck Millipore) 10 kDa cutoff tubes and is henceforth referred to as F6. The Coomassie staining after concentration can be seen in Appendix Figure A1. It showed a sign of aggregation as a smeary band through the lane. The purity of D123mut F6 can be seen in table 1.
22
Figure 5. Western blot analysis of D1 & D123mut under non-reduced (left) and reduced (right) conditions.
Expected sizes of D1 and D123mut are 20.9 and 40.4 kDa, respectively. Medium is the cell medium used for transfection, Flow through is the solution collected after loading the medium on to the purification column,
Wash is the solution collected after washing the loaded column, Pellet is the centrifuged medium pellet.
Unlike the western blots for D1 & D123mut Sept, the D123mut Nov western blot analysis seen in Figure6 showed negligible amount of construct protein in the wash thus the initial fractions F1 and F2 was the only fractions used further for D123mut Nov. The anti-RX antibody does bind to proteins in the fractions. Aggregation is seen for all lanes containing D123mut to sizes greater than the highest point of the reference ladder making it difficult to determine the highest aggregated sizes. In Figure6, Sept F2 shows aggregation which it did not in the western blot analysis made 2 months earlier seen in Figure 5. Yield for D123mut Nov was calculated as the yield of non-aggregated proteins in the fractions compared to the non-aggregated proteins in the medium to only take the usable RX proteins into account. The total amount usable protein expressed is therefore lower than for D123mut Sept fractions.
23
Figure 6. Western blot analysis of D123mut Nov with its related fractions and D123. Expected size of D123mut is 40.4 kDa. Medium is the cell medium used for transfection, Flow through is the solution collected after loading the medium on to the purification column, Wash is the solution collected after washing the loaded column, Pellet is the centrifuged medium pellet. Rabbit anti-RX antibody was used as primary and Donkey anti-
Rabbit IgG was used as secondary antibody.
The actual concentration is similar for all D123mut Sept fractions and so is the purity differing at most 28 % for Sept. The amount expressed protein for D123mut Sept F6 was several times than other the fractions higher due to the high volume. The purity was also highest for D123mut Sept F6 hence it was the only fraction used for the concentration optimization ELISA.
Table 1. Parameters measured to calculate the total amount protein available for ELISA and total yield from cell medium to after purification. All proteins in this table are fractions of D123mut.
Fraction Total conc (mg/ml)
Purity
%
Actual conc (mg/ml)
Volume (ml)
Expressed Protein
(mg)
Yield
%
Sept F1 0.60 64 0.39 1.7 0.66 14.3
24
Sept F2 0.39 71 0.28 2.0 0.56 47.9
Sept F345 0.36 75 0.27 1.2 0.32 2.2
Sept F6 0.39 82 0.32 8.0 2.58 100
Nov F1 0.38 57 0.22 1.7 0.37 55.6
Nov F12 0.40 62 0.25 2.0 0.46 53.1
α-syn oligomerization
The western blot for oligomerization showed that the used primary antibody Syn-O2 binds both monomeric and oligomeric α-syn. It can be seen in Figure 7 that there was a significant number of monomers left after HNE-oligomerization. At the top of the membrane α-syn is also detected at a size too great to enter the gel expected to represent fibrilized α-syn. After HNE-oligomerization α-syn, dimers, trimers and fibrils were detected but clearly less of a specific oligomeric size than monomers.
Figure 7. Western blot results from α-syn oligomerization. Green light represents detection of Syn-O2 & anti- mouse IgG antibody at 800 nm and red light represents detection of Recombinant Anti-Alpha-synuclein antibody
[MJFR1 (ab138501) and anti-rabbit IgG at 700 nm. Visualizing 2 images of the same western blot. Expected Size of α-syn monomers are 14.5 kDa, dimers 29.0 kDa and trimers 43.5 kDa etc.
25
The Coomassie staining of the HNE-oligomerized α-syn seen in Figure 3 that was made 3 weeks later than the western blot in Figure 7 shows a band at ~60 kDa believed to represent tetramers and another transparent band at ~120 kDa believed to represent octamers. The lowest observed size of HNE induced α-syn oligomers was ~60 kDa in the α-syn
oligomerization Coomassie stained gel. Fractions F22, F66 and F32 had concentrations lower than the limit of detection in both the Coomassie staining and western blot seen in Figure 4.
Figure 3. Oligomerized α-syn after SEC. The used fractions were the fractions that showed highest absorbance at 280 nm. Oligos mix did not go through SEC and was used as a negative control for SEC separation. Expected
size of α-syn monomers is 14.5 kDa, dimers 29 kDa and trimers 43.5 kDa etc.
The western blot analysis of the gel run simultaneously as the α-syn HNE-oligomerization Coomassie staining gel is seen in Figure 4 where the lowest detected size of HNE induced oligomers was at ~15 kDa representing α-syn monomers. A less visible band is seen at ~35 kDa and another band is seen in the top of the membrane. Mostly monomeric α-syn was detected here as well in the samples although dimeric and trimeric were present. The HNE induced oligomers was detected at monomeric size, barely at dimeric size and at fibrilized sizes.
26
Figure 4. Western blot analysis of SEC separation of HNE induced α-syn oligomers. F5-F66 are different fractions of elution after SEC. Oligo/mono mix is the HNE induced α-syn oligomers without SEC separation used as a negative control. Expected size of α-syn monomers is 14.5 kDa, dimers 29 kDa and trimers 43.5 kDa
etc.
Trimerization of RX proteins
The trimerized D123mut showed aggregation in both Coomassie staining and western blot. In Figure 5 the D123mut F2 band is smeary meaning the protein is aggregated which is also seen for the trimerized proteins bands. There is a smeary band in the part of the lanes for the
trimerized proteins where the trimerized protein is expected at kDa ~180 though it is not definite that the bands represent the correctly trimerized RX proteins.
27
Figure 5. Showing results from a Coomassie staining gel of trimerization of D123mut with the trimerization proteins TriCatcher-GFP & SpyCatcher-o-tri. Trimerized GFP and Trimerized o-tri represents the trimerized product of D123mut and TriCatcher-GFP or SpyCatcher-o-tri Expected size for D123mut F2 is 40.4 kDa, for TriCatcher-GFP is 60.5 kDa, for SpyCatcher-o-tri 61.8, for Trimerized TriCatcher-GFP 181.7 kDa and for
Trimerized SpyCatcher-o-tri 183.0 kDa.
Aggregation is also seen for the trimerization protein TriCatcher-GFP in the more sensitive western blot in Figure 6 though aggregation is not seen for D123mut F2. Aggregation is seen for TriCatcher-GFP and a band at kDa ~30 which is lower than expected value at kDa 60.5 in both Coomassie staining and western blot. The band for D123mut F2 is not smeary in the western blot in Figure 11. The bands for the trimerized proteins were smearier after anti-His antibody was added to detect TriCatcher-GFP & SpyCatcher-o-tri.
28
Figure 6. Western blot analysis of Trimerization of D123mut F2. TriCatcher-GFP & SpyCatcher-o-tri are the SpyCatchers used. Trimerized GFP and Trimerized o-tri are the products created after trimerization of D123mut
with TriCatcher-GFP or SpyCatcher-o-tri, respectively. Expected size for D123mut F2 is 40.4 kDa, for TriCatcher-GFP is 60.5 kDa, for SpyCatcher-o-tri 61.8, for Trimerized GFP 181.7 kDa and for Trimerized o- tri 183.0 kDa. The left image is after development with only StrepMAB-Classic 2-1507-001(IBA Lifesciences) as
primary antibody. The right image is after development with both anti-Strep-tag antibody and Goat Anti-Mouse IgG Antibody, HRP conjugate (Sigma-Aldrich).
Indirect ELISA
The first ELISA seen in Figure 7 did not show expected results. The OD for the RX fractions was lower than what has been seen in previous studies. The OD for Syn-O2 was as expected following a sigmoidal curve while the OD for RX proteins did not increase above baseline levels until the last data point. A sigmoidal curve was expected since the x-axis is logarithmic, and the concentration is exponentially increasing by 5x for each data point. Also, the binding is expected to be saturated at the highest concentrations giving at flattened top part of the curve. KD for Syn-O2 in ELISA-1 was 0.003 nM. Since OD seemed to increase at the highest concentrations a concentration optimization ELISA was done seen in Appendix Figure A3.
29
Figure 7. Showing results of α-syn indirect ELISA-1 with 3 different fractions of D123mut Sept and Syn-O2 as positive control on a logarithmic x-axis and non-logarithmic y-axis. KD for Syn-O2 was 0.003 nM.
The OD for the indirect ELISA-2 and ELISA-3 done after optimization of D123mut concentration and anti-Strep-tag antibody dilution are seen in Fel! Hittar inte
referenskälla.3 and Figure 95. Results from α-syn indirect ELISA-3 with SEC separated fractions. D123mut was the RX construct used for all fractions. Nov F1 had not been SEC separated thus acting as a negative control
for SEC separation of D123mut fractions.Figure 95. The OD of Syn-O2 binding to α-syn was as
expected as in Figure 7. KD for Syn-O2 was in ELISA-2 0.003 nM and in ELISA-3 0.004 nM.
The OD for D123mut binding to α-syn however, did not reach significantly higher values than in ELISA-1. D123 had increased OD though not saturated still significantly higher than for D123mut. Due to linear inconsistency of the observed OD/concentration in relation to OD for D123 a Scatchard plot could not be used to determine KD for D123 seen in Figure A5 in Appendix, instead a Hill plot was used to determine KD as the x-intercept to 145.3 nM seen in Figure 14. Increasing concentration for RX proteins & anti-Strep-tag antibody and using size isolated fractions of RX proteins did not increase OD for RX proteins to α-syn.
0 0,5 1 1,5 2 2,5 3
0,00001 0,001 0,1 10 1000
OD450
Conc nM
α-syn binding
F1 F2 F6 Syn-O2
30
Figure 83. Results from α-syn indirect ELISA-2 with fractions of D123mut, D123 and Syn-O2 as positive control on a logarithmic x-axis and non-logarithmic y-axis.
Figure 14. Hill plot of D123. v is the fraction bound α-syn. KD was calculated as the x-intercept to 145.3 nM though the data does not fully fit a linear model thus the KD is not fully accurate.
-0,5 0 0,5 1 1,5 2 2,5 3
0,00001 0,001 0,1 10 1000
OD450
Conc nM
α-syn binding
D123mut Nov D123mut Sept D123 Syn-02
y = 0,5914x - 1,2788 R² = 0,8707
-2,5 -2 -1,5 -1 -0,5 0 0,5 1
-2 -1 0 1 2 3
Log(v/(1-v)
Log conc D123 (nM)
Hill plot
D123 Linear (D123)
31
Figure 95. Results from α-syn indirect ELISA-3 with SEC separated fractions. D123mut was the RX construct used for all fractions. Nov F1 had not been SEC separated thus acting as a negative control for SEC separation
of D123mut fractions.
The results from the conducted competitive ELISA with D123 was poor and unreliable and can be seen in Figure A6 in Appendix.
Discussion
The purpose of this study was to express and trimerize purificated RX proteins to perform ELISAs. Initially the purpose was to do indirect and competitive ELISAs to study the difference in α -syn binding between monomers and oligomers and neurotoxicity studies but due to no α-syn binding of the RX proteins shown and time shortage only indirect ELISA was done reliably. Due to bad protein expression, uncertainty of errors and time shortage α-syn binding of D1 was not tested in ELISA. α-syn binding was not observed except for a small minority of all data points for D123mut. For D123 however binding to α-syn was shown in ELISA-2 in Fel! Hittar inte referenskälla.3 with a observed KD of 145.3 nM. Unfortunately, the results from the ELISAs were too poor for a statistical analysis to be relevant.
In Figure 7 of ELISA-1, the OD rises with the last concentration indicating that D123mut bind α-syn though at higher concentrations than expected. However, when concentration optimizing with D123mut F6 with 2 data points higher than the highest concentration in Figure 7 that showed binding no binding was detected. SEC separation was conducted for D123mut Sept and D123mut Nov for another indirect α-syn ELISA-3 in Figure 95 though no reliable significant OD was detected. If a future continuation of this study shows better
-0,5 0 0,5 1 1,5 2 2,5 3
0,0001 0,01 1 100
OD450
Conc nM
α-syn binding
Nov F1 Nov SEC F20-23 Sept SEC F15-17 Sept SEC F18-29 Syn-O2
32
binding of RX constructs to α-syn in ELISAs a statistical analysis should be made to determine if RX proteins have significant affinity and avidity to α-syn.
Structural differences
Structural differences between D123mut and D123 is not believed to solely be the reason for the binding difference. Aside from the mutational change there is another aa sequence
difference of a linker to a BBB transporter that is not yet included in the sequence. Other than possible steric hindering, the BBB transporter only has relevance in in vivo studies hence the linker does not have relevant effect in this study. The mutational change is believed to be a relevant structural difference that could explain α-syn binding properties. One study showed that RX and specific domains bind α-syn without the change agreeing with the results of D123 binding. D123mut did not show to bind α-syn allowing a theory of the mutational change inhibiting α-syn binding of recombinant RX.
Aggregation of RX
Aggregation of D123mut is believed to be the main issue for why D123mut does not show α- syn binding while D123 does. Aggregation was not initially seen for fractions of September transfections though in Figure 5 it is seen indicating that aggregation could have occurred due to freeze thawing, known to cause protein aggregation. D123mut Nov did however show signs of aggregation in Figure without being frozen between purification and Coomassie staining telling us that freeze thawing is not the only cause of aggregation thus for future experiments with RX proteins, early aliquoting to minimal aliquots to avoid aggregation is strongly recommended. Furthermore, to not risk creating aggregation causing cysteine residues by degradation of proteases from cell medium it is recommended to wash the Strep- Tactin® column immediately after sample loading. In previous studies of El-Agnaf et al (38) and Fagerqvist et al (39) aggregation of the generated antibodies is not mentioned thus it is important to take actions in the methods thought to cause aggregation specific to this study.
Due to aggregation seen in Figure 50, it is not possible to draw the conclusion that the
trimerization was successful. In a theoretical ideal case of trimerization with no aggregation, a clear band at the expected size of 180 kDa would be seen for the trimerized proteins. With aggregation in the lane, it could be possible that the smeary band is mostly representing aggregated monomeric D123mut. Unfortunately, after conducting SEC separation of the trimerized D123mut the dilution by sample injection to the system resulted in too low concentration. Concentrating the samples was tested with Amicon® Ultra-15 Centrifugal
33
Filter Unit (Merck Millipore) 30 kDa cutoff tubes without successful results. Since D123 bind α-syn without trimerization it would be reasonable if D123mut could also bind without
trimerization due to the structure similarity.
Quality of α-syn oligomers
In Figure 7, α-syn was detected at the lowest size of ~15 kDa. In the lane of HNE-induced α- syn, the Coomassie staining gel run 3 weeks later in Figure 3 α-syn was detected as the lowest size of ~60 kDa probably representing tetrameric α-syn indicating occurrence of self-
aggregation. In Figure 4 monomeric α-syn was present in the sample thus the western blot and Coomassie staining analysis show contradictory results. One theory is that MJFR1 antibody is more prone to bind monomeric α-syn than oligomeric previously shown by Graef et al. thus give falsely high signal for monomers (40). The low fraction of oligomeric sizes compared to monomeric seen in Figure 7 &Figure 4 could for future oligomerizations be increased by increasing incubation time to 96 h from 72 h. This oligomerization method used must be optimized to enable creation of pure oligomeric α-syn for competitive ELISAs.
Limits and Strengths
Crucial limitations for this study are low protein expression of D1, the failure of trimerizing the RX proteins, uncertainty of reason for D123mut not show α-syn binding and that no reliable results for competitive ELISA was achieved. Results from a competitive ELISA would give essential information of RX protein affinity to α-syn oligomers and monomers and to which it binds better. This would if achieved have fulfilled the purpose of this study. If results from an indirect ELISA of D1 was achieved, more information about possible systemic errors, if SpyTag or Twin-Strep-tag affects binding properties or if the mutational change inhibits α-syn binding could be known hence why the results from this study are unfortunate.
A Strength of this study is the usage of reliable and well described methods such as
Coomassie staining and ELISA. The results are not believed to be affected by any errors in the mentioned methods used or the implementations of them. The adaptiveness of the study is another strength where the methods needing optimization can be easily optimized and the protein constructs can be modified thus the project can continue without major complications.
Future research
34
The results from this study did not increase the hope of a recombinant RX protein as a
treatment for PD. The possible theory of the mutational change inhibiting α-syn is a theory to continue research on for future RX proteins against α-syn. If the theory is true, this study has more noticeably contributed to PD research.
Future measures for this study would be to re-design the plasmids containing the amino acid sequence for D123mut. The sequence differences mentioned can be tested by doing an indirect ELISA using identical protein constructs of RX except for the changes of adding a SpyTag and or the mutational aa change. It is with the results from this study not possible to abandon the possibility neither D1 nor D123mut having affinity to α-syn. Since the results from a study show domain 1+2+3 having better α-syn affinity than only domain 1 and the in this study difficulty of expressing D1, it is more motivated to continue studies on D123mut and D123 than D1.
RX have previously been described as non-monomeric thus a monomeric RX protein might not have correct binding properties to bind α-syn. Doing a trimerization with TriCatcher-GFP and SpyCatcher-o-tri of SEC separated RX protein and running an ELISA of trimerized RX protein would therefore be good to see if multimerization is required for α-syn binding of D123mut.
If positive results are achieved from indirect and competitive ELISAs further cellular studies to see how endocytosis and neurotoxicity changes in presence of functional RX proteins.
Although the results for D1 and D123mut from this study does not help strengthen the theory of a RX protein as a therapeutic protein drug forward, D123 did show promising results in ELISA-2. Adding a Spy-Tag SpyCatcher system to D123 while keeping the His-tag would be an interesting protein to see if the Twin-Strep-tag is a cause of poor α-syn binding and if trimerized D123 could have improve potency. The natural RX protein has shown to bind extracellular α-syn and constructed RX protein D123 have shown affinity to α-syn thus I believe there is potential for a RX protein as a novel protein drug against PD. Syn-O2 was with a KD of 0.003 nM compared to the unreliable KD of 145.3 for D123 and no KD achieved for D123mut seen to have higher potency for α-syn than the RX proteins. If a RX protein were to be an approved and market-based successful drug it would need to have
characteristics giving it advantage over the Syn-O2 antibody and other antibodies shown to reduce aggregated α-syn levels in mice (38,39). Those characteristics could be easier passage over the BBB with the linked BBB transporter and avidity to α-syn by trimerization. If a RX
35
protein can show significantly better binding to oligomeric α-syn than monomeric a future issue is half-life of the protein that must be sufficiently high for the benefits to outweigh the risks.
Since α-syn is a natural protein present in human in not only brain tissue but also kidney, urinary bladder, bone marrow and lymphoid tissues its other not fully known purposes might be affected if a RX drug is administered. If the RX drug binds monomeric α-syn it will likely bind α-syn wherever α-syn is present. The BBB transporter enables passage to the brain, but the drug will likely still be administered intravenously thus entering the blood system first and circulate through the tissues with α-syn binding soluble monomeric α-syn. Even with a BBB transporter, a clear minority of the administered protein have shown to not reach the brain (23). The possible side effects this could cause is not fully known since the full function of α- syn is not known though the proposed functions of regulating exocytosis and vesicle fusion (9) could be affected. If a reason for D123 bound α-syn and D123mut did not is the
mutational change, it would be needed to keep the RX protein non-mutated to achieve α-syn binding. This would enable unwanted endogenic binding likely interrupting the target from its normal function risking side effects. There are other mutations in mouse RX affecting the same endogenic binding and if the mutational change used in this study is not compatible with α-syn binding the other mutations could be tested instead. For a RX protein to be viable as an authorized drug I believe that a mutational change to inhibit endogenic binding and
multimerization to at least dimeric form is necessary. As biological drugs are important tools against PD, future continuations of this study and research on recombinant RX in general could result in a RX protein being the treatment PD patients are longing for.
Conclusion
To conclude, this study did partly fulfill its purpose of expressing RX proteins and study their binding to α-syn. The results, however, were not as promising as expected. Neither D1 nor D123mut were able show affinity to α-syn. For D1 it was due to bad protein expression and for D123mut it was due to aggregation and undefined binding properties. Further
investigation on trimerized RX protein is needed to include or exclude it as a possible future therapeutic protein. Since D123 have shown affinity to α-syn I believe there is potential for a RX protein as a novel drug against PD but there are many uncertain factors as it is still in very early preclinical phase.
36
Acknowledgements
Firstly, I want to thank my supervisor Sofia Stenler for initially introducing me to the topic of designing protein drugs but mostly for being very helpful at the lab during working hours and quickly answering questions during the off-hours. Without Sofia this project and trying to overcome all difficulties at lab would not have went as smooth as it did.
Also, thanks to all the nice colleagues of ProDDe research group for the help I have got with practical and theoretical issues. I really appreciated being a part of your group this Semester!
