Development of easily visualized immunoassays for diagnostic applications

35 Figure 5.1: Fluorescence microarray scan images and intensity profiles of Cy3-tagged AHIgG 100 µg/mL spotted on A BSA 500 µg/mL, B DMOAP and C Protein G 500 µg/mL coated surfaces.. 44

Table of Content

List of Figures IV List of Tables VII List of Equation VII Summary VIII

1 Introduction 1

1.1 Protein arrays 1

1.2 Point of care diagnostics 2

2 Objectives 3

2.1 Protein G Immobilization 3

2.2 Development of microarray 5

2.3 Applying liquid crystal as read out 7

2.4 Tuning liquid crystal sensitivity at low concentration range 8

2.5 Detection of trace proteins 9

2.6 Towards high throughput multiplexing operation 9

3 Literature review 10

3.1 Microarray substrate 10

3.2 Proteins of the microarray assembly 11

3.2.1 Immunoglobulin G (IgG) 12

3.2.1.1 Structure of IgG 12

3.2.1.2 IgG subclasses 13

3.2.2 Protein G 14

3.3 Multiplexing of the protein microarray 16

3.4 Liquid crystals 17

3.4.1 Liquid crystal shapes and phases 18

3.4.2 4-cyano-4’-pentylcyanobiphenyl (5CB) 20

3.4.3 Orientation of liquid crystals on solid substrate 22

3.4.4 Liquid crystal under cross polarizer 23

3.5 Development of protein microarray 25

3.5.1 Labelled techniques of protein detection 25

3.5.1.1 Enzyme-linked immunosorbent assay (ELISA) 25

3.5.1.2 Fluorescence-linked immunosorbent assay (FLISA) and Fluorescent immunoassay (FIA) 27

3.5.1.3 Metallic particle-based scanometry 29

3.5.2 Label free technique of protein detection 30

3.5.2.1 Liquid crystal based microarray 31

3.5.3 Comparing the current microarray to existing method of quantifying IgG 37

4 Materials and Methods 39

4.1 Materials 39

4.2 Cleaning of glass 39

4.3 Cleaning of silicon wafer 40

4.4 Coating of DMOAP on glass and silicon wafer 40

4.5 Protein immobilization 41

4.6 Protein crosslinking 42

4.8 Water contact angle study 43

4.9 Ellipsometry 43

5 Results and Discussion 44

5.1 Anti-human IgG immobilization 44

5.2 Water contact angle measurements 46

5.3 Application of protein G onto DMOAP surface 48

5.4 Crosslinking protein G adsorbed on DMOAP surface 52

5.5 Checking the integrity of Protein G after crosslinking 56

5.6 Verifying the specificity of the immobilized capturing antibody 58

5.7 Quantitative analysis 60

5.8 Adaptation of microarray to liquid crystal-based format 63

5.9 Ellipsometry study of layer thickness 65

6 Conclusion 68

7 Recommended future work 70

7.1 Introducing MPTS/DMOAP mixed SAM and Cys-Protein G 70

7.2 Usage of antibody fragment 70

7.3 Relating solution concentration and solute surface concentration on a substrate upon coating 71

7.4 Harnessing the sensitivity of nematic liquid crystal to magnetism 72

7.5 Harnessing the sensitivity of nematic liquid crystal to electric field 76

8 Reference 78

List of Figures

Figure 2.1: View of birefringent output from 5CB supported on (A) plain

glass and (B) DMOAP coated glass over cross polarized

microscope 4

Figure 2.2: IgG immobilization by (A) physical adsorption (B) chemical bonding (C) protein G affinity binding 4

Figure 3.1: Structure of IgG Dark areas are constant regions while white areas are variable regions 13

Figure 3.2: Sequence of domains in Protein G GA are the albumin-binding domains while B1, B2 and B3 are the IgG-albumin-binding domains 15

Figure 3.3: Chain folds for B1 domains of Protein G 15

Figure 3.4: Positional and orientation order changes from solid crystalline, to liquid crystal and liquid state [38] 17

Figure 3.5: Nematic liquid crystal [38] 18

Figure 3.6: Smectic liquid crystal [38] 19

Figure 3.7: Defining one pitch Left: Chiral nematic (cholesteric) with 1 pitch represented by p; Right: smectic liquid crystal 20

Figure 3.8: Structure of 4-cyano-4’-pentylcyanobiphenyl (5CB) 21

Figure 3.9: Spherical coordinates 22

Figure 3.10: Structure of DMOAP 23

Figure 3.11: Birefringent crystal between cross polarizers [41] 23

Figure 3.12: Enzyme-linked immunosorbent assay (ELISA) 27

Figure 3.13: Fluorescence linked immunosorbent assay 28

Figure 3.16: Effects of phospholipids on alignment of Liquid crystal (LC)

at interface 33 Figure 3.17: Molecular interaction at phospholipids layer 34 Figure 3.18: Structure of DOGS-NTA-Ni [52] 34 Figure 3.19: Dark to bright optical response from sandwiched form liquid

crystal microarray birefringence [12] 35 Figure 5.1: Fluorescence microarray scan images and intensity profiles

of Cy3-tagged AHIgG (100 µg/mL) spotted on (A) BSA (500

µg/mL), (B) DMOAP and (C) Protein G (500 µg/mL) coated

surfaces 44 Figure 5.2: Fluorescence microarray scan images and intensity profiles

of cy3-tagged protein G (300 µg/mL) applied to the entire

DMOAP coated surface using (A) capillary suction method

and (B) spot-and-cover method 48

Figure 5.3: Fluorescence microarray scan images and intensity profiles

of Cy3-tagged protein G of concentration 2 µg/mL and 20

µg/mL applied to DMOAP coated surfaces by capillary

suction (A and C) and spot-and-cover (B and D) method 51

Figure 5.4: Fluorescence microarray scan images of cy3-tagged protein

G coated surfaces 54 Figure 5.5: Graph showing variation in fall off concentration

corresponding to cy3-tagged protein G of 10 – 80 µg/mL

under crosslinked and non-crosslinked situation 56 Figure 5.6: Fluorescence microarray scan of crosslinked protein G (40 57 Figure 5.7: Microarray scan images and intensity profiles of slides with

anti-human IgG (top row) and anti-murine IgG (bottom row)

spotted on protein G crosslinked surface, subjected to (A)

Cy3-tagged human IgG and (B) murine IgG 58

Figure 5.8: Microarray scanned image of anti-human IgG (200 µg/mL)

oriented by crosslinked protein G (40 µg/mL), subjected to

various concentrations of cy3-tagged human IgG, 0.01 µg/mL

to 10 µg/mL 61 Figure 5.9 Quantitative analysis graph showing average fluorescence

intensity of 3 spots of the same cy3-tagged human IgG

concentration for 9 different concentrations plotted against

the concentration of human IgG 62 Figure 5.10: Cross section view of liquid crystal cell showing orientation

change of 5CB during binding event Transition of 5CB from

homeotropic to planar orientation occurs upon the binding of

IgG to anti-IgG at the surface The effect was extended to

the bulk phase and the optical signal from the cross

polarized microscope transit from dark to bright 64 Figure 5.11: Setup of liquid crystal cell 64

Figure 5.12: Appearance of the liquid crystal cell, constructed using the

protein microarray in the absence of analyte, under cross

polarized microscope 65 Figure 7.1: Sandwich immunoassay based protein microarray with

detection antibody conjugated to ferromagnetic nanoparticle

The ferromagnetic nanoparticle is represented by CoPt 74

Figure 7.2: Experimental setup where P is the polarizer, A is the

analyzer, S and N are magnetic poles and C is the liquid

crystal cell [68] 74 Figure 7.3: Birefringent optical output from cross polarized microscope of

a 5CB supported on CoPt coated surface 75 Figure 7.4: Michael Levy Chart of Birefringence 75 Figure 7.5: Birefringent optical output from cross polarized microscope of

a 5CB doped with Fe3O4 in (a) absence and (b) presence of

magnetic field 76 Figure 7.6: Ferroelectric nanoparticles in liquid crystal (A) Particle with

no electric dipole moment, in isotropic phase (B) Particle

with electric dipole moment producing an electric field,

interacting with orientation order of the nematic phase [71] 77

List of Tables

Table 5.1: Signal to noise ratio of cy3-tagged AHIgG spotted on BSA,

DMOAP and protein G surface as shown in Figure 5.1 45 Table 5.2: Water contact angle on (A) Protein G (500 µg/mL), (B)

DMOAP and (C) BSA (500 µg/mL) coated surfaces 47 Table 5.3: Fall off concentrations (Tween 20) of various concentrations

of crosslinked surface bounded cy3-tagged protein G 55 Table 5.4: Fall off concentrations (Tween 20) of various concentrations

of non-crosslinked surface bounded cy3-tagged protein G 56 Table 5.5: Signal to noise ratio analysis of each data point plotted in

Table 5.6: Thickness of each protein layer according to dry state

List of Equation

Equation 3.1: Function that can be averaged to find order parameter,

also known as second Legendre polynomial [39] 18 Figure 3.5: Nematic liquid crystal [38] 18

Summary

Microarray is known to enable multiplex, high throughput protein profiling and quantitative analysis with small sample volume It, therefore, revolutionized the development in functional genomics and functional proteomics studies In addition, it has also enhanced the efficiency of diagnostics

In this study, the ability of adapting a fluorescence based microarray system for detecting human immunoglobulin G (IgG) to a liquid crystal microarray system was investigated The read out system was changed

to enhance the portability and ease of usage so as to attain the point of care application standards

Here, we built an immunoassay on silanized glass slide coated with protein G that was then cross-linked for grafting the capturing protein, anti-immunoglobulin G (anti-IgG), for screening immunoglobulin G (IgG) Upon characterizing this in house built immunoassay, the read out system was changed from fluorescence to liquid crystal, 4-cyano-4’-pentylcyanobiphenyl (5CB), which changes from dark to bright indicating the presence of IgG Adapting the system to liquid crystal read out

crystal works As such, ways of tuning the system such as applying alternative liquid crystal with higher nematic range as well as usage of ferromagnetic and ferroelectric nanoparticles are recommended in this work

1 Introduction

1.1 Protein arrays

Protein arrays are efficient methods commonly sought forth in facilitating processes like comparative proteomics and functional genomics studies of protein-protein, protein-DNA and protein-small molecule interactions, determination of protein expression level, diagnosis as well as cancer prognosis prediction [1-5] The objective behind protein array development is therefore, partly, to enable multiplex and high-throughput protein profiling and quantitative analysis

Unlike DNA fragments, which can be amplified through polymerase chain reaction, the amount of protein sample cannot be increased To exacerbate the situation, biomarkers and metabolites critical in detections exist only in trace concentrations, especially in the early stages of expression [2] Thus, developing

a sensitive protein array with a low limit of detection and low volume requirement remains a bioanalytical challenge

Several processes that have been adapted into microarray format to enhance their test throughput, such as aforementioned antibody screening and protein-protein interaction study, involve labelling of detection antibodies or analytes

Labelling, however, has its drawbacks and hence, various label-free methods have been explored This will be further discussed in Chapter 3

Thermotropic liquid crystals (LCs) exhibit colourful birefringence and can be developed into a label free read out platform for protein detection The presence

of protein can be reflected by the dark to bright change of the optical signal This phenomenon opens the way to development of liquid crystal based microarray for protein detection

1.2 Point of care diagnostics

Point of care diagnostics have been acknowledged as a frontline need and have generated enormous interest in the field of protein array and array reader design [3, 6] This type of applications enables clinical detection with minimal trained hands and instantaneous results obtained can be crucial in monitoring progress

of chronic diseases, such as renal failures and diabetes [7], which helps to improve patients’ quality of living while combating mortality rate The efficiency in life saving conferred by point of care diagnostics is especially significant in pandemics and disasters [8]

2 Objectives

This research aims to develop a portable, zero energy consumption, and label free immunoassay capable of high-throughput quantitative profiling of proteins with low volume requirement The entire study was conducted in a stepwise manner as describe below

2.1 Protein G Immobilization

The first portion of this research concerns immobilization of immunoglobulin G (IgG) on N,N-dimethyl-n-octadecyl-3-aminopropyl-trimethoxysilylchloride (DMOAP) coated glass slide for specific binding to the analyte The DMOAP coating functions to align the nematic liquid crystal 4-cyano-4’-pentylbiphenyl (5CB) for the read out The optical effect of DMOAP with respect to 5CB can be seen from Figure 2.1 DMOAP, however, does not bear any end-functional group for protein immobilization Therefore, using other methods of securing protein on the surface throughout the entire experiment is required Physical adsorption of IgG on DMOAP coated surface, as illustrated in Figure 2.2a, can be easily done, but the poor retention after several rounds of washing in the experiment and random orientation together with the reduction in IgG functionality upon the adsorption may impact the consistency and sensitivity of the immunoassay [9,

Figure 2.1: View of birefringent output from 5CB supported on (A) plain glass and (B)

DMOAP coated glass over cross polarized microscope

Figure 2.2: IgG immobilization by (A) physical adsorption (B) chemical bonding (C) protein

G affinity binding

Silanes bearing end-functional groups can be used instead of DMOAP However, they must possess the ability to align 5CB as DMOAP Alternatively, they can be used together with DMOAP to form a mixed self-assembled monolayer (SAM) that possesses both the ability to align 5CB and to chemically immobilize IgG One such silane is triethoxysilane aldehyde (TEA) It decorates glass surface with aldehyde groups and enables peptide immobilization through Schiff base reaction on the glass surface [11] (Figure 2.2b) Such chemical bonding, however,

risks the denaturation and random orientation of the IgG The aldehyde group on the surface, for instance, may react with the amine group at the proximity of the Fab site and subsequently hinder analyte binding, rendering any detection impossible Therefore, it is preferred that the IgG be immobilized with specific and desirable orientation

The use of protein G is hypothesized to be effective to fill this niche Crosslinked Protein G can adsorb strongly on DMOAP modified surface to bind IgG at Fc site and hence orientating the Fab sites upright as illustrated in Figure 2.2c Binding

in this fashion preserves the structure of Fab sites for specific binding with analytes during detection As such, the signal strength in detection can be improved

Upon immobilization of glutaraldehyde crosslinked protein G, selectivity of protein

G was checked Since the recombinant protein G used has affinity solely towards IgG, strepavidin was used to examine IgG selectivity The ability of protein G layer to withstand the washing steps was also evaluated The first portion of the project probes the strength and selectivity of the first protein layer

2.2 Development of microarray

confirmation read out This phase of the project started with immobilization of anti-IgG on the protein G layer Anti-IgG was used as the capturing antibody that confers the assay specificity Immobilizing too little anti-IgG on protein G layer leaves behind excess IgG-binding domains of protein G capable of binding the analyte IgG, and this leads to false positive Too much anti-IgG, on the other hand, occupies almost all IgG-binding domains of protein G and hence any excess anti-IgG may be physically adsorbed on the first layer of anti-IgG This excess anti-IgG adsorption can be countered by a stringent wash, but still constitute unnecessary use of reagents Therefore, the minimal concentration of anti-IgG at which the microarray functions selectively was probed and the characterization of this second protein layer was necessary

The characterization was performed using fluorescence scanning and ellipsometry Parameters such as minimum concentration of protein immobilized, incubation temperature and duration to ensure even distribution of protein over the entire glass slide were determined Controlling the amount of protein use and limiting the time consumed during the experiment make the operation more economic and efficient A uniform distribution of protein is important in quantitative study as the surface density of immobilized proteins over the glass slide can cause unintended read out signal variation, rendering results inaccurate

or inconsistent

2.3 Applying liquid crystal as read out

Upon optimizing the aforementioned system of its physical parameters, the read out was changed from fluorescent based method to liquid crystal (LC) based method The birefringent read out from LC based system appeared to turn from dark to bright optical signal under cross polarized microscopy when concentration of protein exceeds critical concentration [12] This is the main principle on which the immunoassay was built Thus, it is imperative that the accumulated protein G and anti-IgG must not exceed the critical concentration to allow specific analyte binding on the surface and to induce a dark to bright optical signal transition for read out

In the first portion of this phase, critical concentrations of protein G and anti-IgG were determined When the working concentration established previously falls within the critical concentration, working range of the microarray using these two different read out is tested and compared Their general performance is then evaluated Upon successful conversion to liquid crystal based read out, optimization shall be done to enhance the signal This shall be one step towards developing the system as zero energy consumption, point of care application!

If the working concentration established previously falls beyond the critical concentration of the liquid crystal based system, further optimization of protein

system can be of a higher nematic phase stability (the upper temperature limit to which the nematic phase exist) and a larger nematic range (temperature range which nematic phase occurs) [13] Nematic liquid crystals of higher nematic stability are those with more conjugated core unit and higher polarizability anisotropy For example, 8CB, which is in the same homologous series as 5CB,

is more nematically stable [14] At room temperature, therefore, this liquid crystal will remain more viscous with harder to disturb long range order With this intrinsic property of the supported liquid crystal, the protein critical concentration may be increased As such, a change from fluorescent based sensor to liquid crystal based sensor will then be made possible

2.4 Tuning liquid crystal sensitivity at low concentration range

Liquid crystal (LC) birefringence allows probing of a small difference in concentration of a certain analyte that disrupt their alignment The difference can easily be read off as colour difference [12] However, it is not the strength of LC

to compete with other techniques, especially fluorescence and gold nanoparticle based system, on limit of detection Ways of making the selected working LC sensitive at low concentration will thus be an exploration of great scientific interest Possible ways of enhancing the sensitivity at low limit of detection such

as various methods of doping will be looked at

2.5 Detection of trace proteins

Upon surmounting the hurdle described in Section 2.4, a liquid crystal based microarray for profiling trace protein can be developed The detection for cancer biomarkers, such as epidermal growth factor receptor (EGFR) and Human chorionic gonadotropin (HCG), will be carried out HCG is an extremely sensitive and specific marker for trophoblastic tumors of placental and germ cell origin [15]

On the other hand, EGFR is known for its over-expression on cancer cells and it has been used not just for diagnostics, but also the continual study on cancer prognosis of the cancer patient [16] Developing ability to detect these proteins is also crucial in screening potential cancer drugs, which are largely the derivatives

or conjugates of antibodies to the cancer markers The current way of screening effective anti-EFGR is by gel run, a microarray test will be faster and sample volume requirement will be lower [17, 18] The experiment will involve tests upon anti-HCG and anti-EGFR in selectivity for various biomarkers on the developed platform After that, the limit of detection for each biomarker will be probed Optimization such as discussed in Section 2.4 will be considered A conclusion

on the usability of such platform for biomarkers detection and drug screening will then be drawn

2.6 Towards high throughput multiplexing operation

The next aim, if time permits, is to develop a high throughput, multiplex platform for specific screening of biomolecules by incorporating capabilities from microfluidics It has been known that usage of microfluidic protocol saves time in fabricating the microarray and the time taken for detection step is also known to

be lesser As such, high-throughput operation is thus possible The channel width

of microfluidic system is tiny and thus multiple parallel channels can be built to allow multiplex detection on same substrate This does not only confer convenience and efficiency, it put forth the advantage of required sample volume and batch error reduction

3 Literature review

3.1 Microarray substrate

Various materials can be used in the construction of the solid substrate Commonly used substrates include glass slides, silicon, nylon, nitrocellulose or polyvinylidene fluoride (PVDF) membranes and agarose or glass microbeads [1,

19, 20] Any of these surfaces can be modified chemically to accommodate the binding need of proteins For instance, poly-L-lysine coating and silanization of glass surfaces with amine, carbonyl or carboxylic groups can be done to introduce charges or functional groups for protein binding [9, 21, 22] While analytes delivered onto planar surfaces are the most familiar format, a number of more advanced architectures incorporating developments in microfluidics are

introduced for the purpose of increasing immobilization area and enhancing the efficiency of detections [23]

3.2 Proteins of the microarray assembly

In this section, immunoglobulin G (IgG), which serves as the capturing antibody

in the construction of protein microarray, is reviewed in Section 3.2.1, and protein

G, which binds IgG to the substrate and orientates it upright, is reviewed in Section 3.2.2

3.2.1 Immunoglobulin G (IgG)

Immunoglobulins are glycoproteins produced by plasma cells as antibodies against immunogens Major human immunoglobulin classes include IgA, IgD, IgE, IgG and IgM Immunoglobulin G (IgG) forms 80% of all immunoglobulins in human serum and hence is the most abundant among the 5 major classes of immunoglobulins [24]

3.2.1.1 Structure of IgG

The structure of IgG, as shown in Figure 3.1, consists of 2 identical heavy chains (approximately 50kD each) and 2 identical light chains (approximately 25kD each) These chains are joined together by disulphide bonds to form a Y shape structure Each chain has a variable and a constant region At the amino terminal region, variable heavy (VH) and light chains (VL) form the Fab region for antigen binding The amino acid sequence in this region is highly variable and this confers specificity for antigen recognition and binding On the other hand, constant regions of heavy (CH) and light chains (CL) form Fc region This region accounts for binding to Fc receptors on cell membrane

Figure 3.1: Structure of IgG Dark areas are constant regions while white areas are variable regions

The specificity of IgG can be harnessed in accurate capturing and detection of antigen Although there are other specific biomolecules, such as streptavidin and biotin, the ease of raising immunoglobulins for various types of antigen saves tedious procedures of tagging the analyte for identification Moreover, the versatility of its use in interacting with other biomolecules makes it a better choice

in microarray development

3.2.1.2 IgG subclasses

The structural heterogeneity in both the heavy and light chains constant and variable regions leads to the formation of IgG subclasses A numeric designation

serum The order runs as IgG1 with 5-12 mg/mL in healthy adult serum, IgG2 with 2-6 mg/mL, IgG3 with 0.5-1 mg/mL and IgG4 with 0.2-1mg/mL [25, 26] On an interesting note, occurrence of IgG in other body fluid, such as saliva, urine and cerebrospinal fluid, is lower than in serum For instance, total IgG in cerebrospinal fluid (0.8-7.5 mg/dL) is 100 times lower than in serum [27]

As such, it can be seen that the detection of IgG in serum does not need low limit sensors, but that of other body fluids demands low limit and sensitive sensors This is especially crucial since detection of IgG in such samples is often to do with forensic investigation, diagnosis and disease progression tracking etc [4, 28]

3.2.2 Protein G

Protein G originates from the cell wall of human Streptococcal bacteria strain C and G (G148, G43 and C40) [29, 30] This alphabet protein helps the bacteria to camouflage themselves with host proteins and escape phagocytosis in the host [31] Protein G from strain G148 is 65KD and strain C40 is 58KD They have affinity for both IgG and human serum albumin (HSA) Protein G from strain G43

is 40KD and has affinity solely for IgG This is useful in grafting IgG molecules on microarray surface However, recombinant work is done nowadays to not just ensure that it does not bind to albumin, but also combine capabilities of other alphabet proteins so that it has similar affinity for IgGs from various species

Being a FcγRIII, or commonly known as CD16, receptor, protein G bears domain for binding IgG and albumin molecules GA is the domain for binding albumin and B1, B2, B3 (4 beta sheets and 1 alpha helix structures) are the domains for binding IgG The sequence of these domains is depicted in Figure 3.2 A ribbon diagram of domain B1 is as shown in Figure 3.3 These B domains bind both Fc and Fab fragments of IgG However, at physiological pH, domain B–Fab binding

is weak, Ka=105 M-1 andhence binding is predominately towards Fc, Ka=108 M-1[31] This leads to the fact that protein G orientates IgG molecule almost upright

Figure 3.2: Sequence of domains in Protein G GA are the albumin-binding domains while B1, B2 and B3 are the IgG-binding domains

Figure 3.3: Chain folds for B1 domains of Protein G

Because of its selectivity towards IgG molecules in protein mixture, protein G is usually used in affinity chromatography to purify IgGs On top of this, it can be used in immunoassays to immobilize and orientate IgGs on the substrate surface Hence, protein G was immobilized on the surface of DMOAP coated glass to graft and orientate anti-IgG (which is also an IgG) in this project

3.3 Multiplexing of the protein microarray

Protein spotting in construction of microarray can be done using contact or contact printing robot [1, 32] This allows the amount of protein printed to be in nanoliters, or even 200 picoliters [33] At the same time, printing allows protein spots to be positioned densely and precisely For instance, Gavin and Stuart demonstrated that 10800 protein spots of 150-200 µm in diameter can be printed

non-on non-one standard microscope glass slide for protein functinon-on analysis while Ian and co-workers conducted screening at 18342 spots per slide [34, 35] In a modest scale, Chin and co-workers simultaneously immobilized 360 antibodies for post-translational modification study [1] Amidst the current achievement, microarray can be combined with other techniques such as microfluidics to improve the efficiency of multiplex analysis [36, 37]

3.4 Liquid crystals

Liquid crystals (LCs) are anisotropic molecules in mesophase state, an intermediate phase between crystalline solid phase and liquid phase as illustrated in Figure 3.4 In this phase, LCs possess orientational order but no positional order This means that though the individual molecules may diffuse freely in the bulk, the orientation of the molecules is mostly towards a certain direction

Figure 3.4: Positional and orientation order changes from solid crystalline, to liquid crystal and liquid state [38]

This direction is represented by the director Order parameter quantifies the extent of LC orientation varying from the director Typically, the average of the function below is taken A value of 0 represents absence of orientational order while 1 represents perfect orientational order The usual value will fall between 0.3 and 0.9

(3cos2θ-1)/2

Equation 3.1: Function that can be averaged to find order parameter, also known as

second Legendre polynomial [39]

3.4.1 Liquid crystal shapes and phases

There are several types of LC phases Some materials have more than a single

LC phase transition Hence, at different temperature, they assume characteristics

of the different LC phase The phase transition that LCs undergo depends on the structure of the LCs

Liquid crystal phases:

1 Nematic phase: LCs tend to align with their long axes almost or parallel with the director, assuming a long range orientational order No positional order exists for this phase (Figure 3.5), i.e molecules do not assume lattice position This phase is usually formed by calamitic LCs

2 Smectic phase: This phase displays orientational order and some degree of positional order LCs are arranged in layers perpendicular to the director as in Figure 3.6 When the layer normal is at different angle with the director, different alphabet is used to name the smectic phase For instance, smectic A phase represents the phase where director is parallel to the layer normal while in smectic C phase, the layer normal makes an angle with the director

Figure 3.6: Smectic liquid crystal [38]

3 Chiral phase: This phase is assumed by LCs which are chiral The subsets of this phase are chiral nematic (cholesteric) phase and chiral analogue of smectic C phase, smectic C* phase The main characteristic of this phase is the layer by layer tilting of mesogens from layer normal that is in revolution This is shown in Figure 3.7 As such, a pitch, distance of 1 complete revolution, can be defined

Figure 3.7: Defining one pitch Left: Chiral nematic (cholesteric) with 1 pitch represented

by p; Right: smectic liquid crystal

4 Columnar phase: This phase is usually formed by discotic LCs forming column-like structures

5 Cubic phase which consist of micellar lattice units or complicated interwoven networks Such structures are formed under high concentration by lyotropic LCs

Various types of liquid crystal molecules:

a) Calamitic molecules –rod-like molecules

b) Discotic molecules –disc-like molecules

c) Sanidic molecules – lath-like (board-like) molecules

d) Lyotropic molecules – amphiphilic molecules with polar/ hydrophilic head and non-polar/ hydrophobic tail

3.4.2 4-cyano-4’-pentylcyanobiphenyl (5CB)

The liquid crystal used for this project, 4-cyano-4’-pentylcyanobiphenyl (5CB), is

a nematic liquid crystal The calamitic molecule assumes a rod like structure as shown in Figure 3.8 5CB consists of a 5 carbon length alkyl ending group, 2 benzene rings to confer rigidity and a cyano group to serve as a polar terminal

Figure 3.8: Structure of 4-cyano-4’-pentylcyanobiphenyl (5CB)

Alkyl chain of 3 to 7 carbons allows the molecule to display solely nematic properties At a chain length of 8 and 9, the molecule begins to show smectogenic behaviour and reduced amount of nematogenic behaviour Beyond the chain length of 10, only smectogenic behaviour is observed [14]

Melting temperature of 5CB from crystalline solid state is 297.0K Beyond this, the transition into nematic phase occurs above this temperature At 308.3K, transition into liquid phase takes place [14] Hence, the nematic range or temperature range at which 5CB remains in nematic phase spans 11.3K

3.4.3 Orientation of liquid crystals on solid substrate

Figure 3.9: Spherical coordinates

LCs may align in homeotropic, planar, or tilted orientations at the solid substrate surface Using the spherical coordinate system as illustrated in Figure 3.9, in homeotropic orientation, the director (along y-axis) is perpendicular to the surface (x-z plane) and the polar angle according to the spherical coordinate θ = 90° In planar orientation, the director lies parallel to surface, θ = 0° In the case of tilted orientation, θ is non-zero or 90°, and φ is arbitrary

The substrate that supports LC can be modified chemically or physically to align them For the usage of 5CB, for instance, physisorption of hexadecyltrimethyl-ammonium bromide (HTAB) or chemisorption of dimethyl-n-octadecyl-3-aminopropyltrimehoxysilyl chloride (DMOAP) on the surface of the substrate enable the alignment of LC to assume homeotropic orientation Alternatively, rubbing of the substrate in a single direction can be done This is especially the

case when the surface cannot be chemically modified Surfaces on which rubbing has been done in literature include PVA and polyimide [40]

This project involves the work of constructing protein microarray on DMOAP coated glass surface Doing so allows the detection of protein to be seen through the dark to bright change of optical signal as previously described The structure

of DMOAP is as shown in Figure 3.10

Figure 3.10: Structure of DMOAP

3.4.4 Liquid crystal under cross polarizer

Figure 3.11: Birefringent crystal between cross polarizers [41]

Non-polarized white light from the illuminator enters the polarizer P in Figure 3.11 and is linearly polarized with an orientation in the direction indicated by the arrow (adjacent to the polarizer label) The polarized light is arbitrarily represented by a red sinusoidal light wave Upon entering the liquid crystal cell, the polarized light

is refracted and divided into two separate components vibrating parallel to the crystallographic axes and perpendicular to each other (as represented in the figure by solid red and lined red light waves) This is because liquid crystal is anisotropic when it is not aligned by modified substrate Anisotropy of the LC molecule implies that the refraction index along different axis of the LC molecule

is different Hence, the polarized light is doubly refracted, and this results in birefringence The polarized light waves then travel through the analyzer (whereby polarization position is indicated by the arrow next to the analyzer label), which allows only those components of the light waves that are parallel to the analyzer transmission azimuth to pass The relative retardation of one ray with respect to another is indicated by an equation ∆n·t (thickness of LC cell multiplied by refractive index difference) This retardation leads to the superposition of the analyzer output to be colourful

When the substrate used to sandwich the LC forming an LC cell is modified with DMOAP, LC assumes a homeotropic orientation Hence, it seems isotropic in the direction of polarized light Therefore, double refraction cannot occur and polarized light will pass through the sample propagating in the same plane As such, it will be blocked by the analyzer and the optical output will be dark

It is by this working principle that disruption of the alignment of LC by protein binding on the substrate surface can be transduced into dark to bright optical output

3.5 Development of protein microarray

As earlier mentioned in Chapter 1, several techniques involving labelling of detecting protein or analyte have been adapted into microarray format to enhance their test throughput However, this strategy is not easily applicable to the unlabelled techniques Hence, most unlabelled techniques evolve to microarray format later The following sections review various labelled and unlabelled techniques that have evolved into microarray format A more comprehensive review on the unlabelled microarray applications is written by Yu

et al [22]

3.5.1 Labelled techniques of protein detection

3.5.1.1 Enzyme-linked immunosorbent assay (ELISA)

Protein detection was conventionally performed with enzyme-linked immunosorbent assay (ELISA), which is known for its high specificity ELISA

are then subjected to the bound analytes on the surface Upon washing, only detection antibodies bound specifically to the analytes remain on the surface Colorimetric signal is produced by the surface bound enzyme to indicate presence of the analyte However, the traditional ELISA comes with a few drawbacks These include the consumption of large amounts of plastic microtitre plates and the generation of large amounts of biological waste, low throughput, and the need for a large volume of expensive high-purity antibodies and consumable reagents [42] To ameliorate the situation, microarray platform has been adopted for ELISA In microarray setting, microtitre wells are replaced by spots Each spot is the immobilization area of different capturing antibody Multiple types of screening and higher numbers of repeats can, therefore, be done on a single substrate for a small volume of sample This is illustrated in Figure 3.12 with different spots of capturing antibody represented by different colour The microarray platform enables detection of multiple analytes with

~50µL volume [43] The volume of capturing antibodies required for analysis can

be as low as 200pL/spot when protein is spotted by a print head [33] Due to the reliance of ELISA on enzymatic activity, biological conditions have to be stringently controlled and the detection has to be performed in solution phase The time lag due to the duration needed for enzymatic activity needs to be further improved

Figure 3.12: Enzyme-linked immunosorbent assay (ELISA)

3.5.1.2 Fluorescence-linked immunosorbent assay (FLISA) and Fluorescent

immunoassay (FIA)

The same format of ELISA can be applied to fluorescence-linked immunosorbent assay (FLISA), illustrated in Figure 3.13 In this case, unlabelled analytes captured on the surface by capturing antibodies is detected by fluorescence labelled detection antibodies instead Another format where immobilization of capturing antibodies on the surface of substrates to detect labelled analytes as in Figure 3.14 is involved is called fluorescent immunoassay (FIA)

The ability of these techniques to perform multiplex analysis is represented in the same figures (Figure 3.13 and 3.14) with differently coloured capturing antibodies spotted on the same substrate at designated positions to probe different analytes

in the same sample subjected to the surface Fluorescence detected at that

Y Y

Colourless enzymatic substrate

Coloured product

Figure 3.13: Fluorescence linked immunosorbent assay

Figure 3.14: Fluorescent immunoassay

Fluorophore bleaching is the main drawback of these assay formats This is minimized by adjusting the excitation time, or changing the tag to quantum dots, such as CdSe/ZnS, which resist bleaching and provide superior fluorescence intensity [44] However, the reliance of this method on costly and non-portable instrumentation renders it not possible to be a point of care application

3.5.1.3 Metallic particle-based scanometry

Silver enhancement was used to amplify DNA scanometric signal [45] It is also used to amplify the protein scanometry signal, but the effect is not as good as developing the size of the gold nanoparticle tag itself Mirkin and co-workers demonstrated that a gold nanoparticle-based microarray can detect 300aM of prostate specific antigen using capturing antibody decorated glass slide The process is described in Figure 3.15 This low limit of detection was achieved by gold development, which is a second deposition of gold on the original gold nanoparticle tag, as shown in the figure to increase the original signal strength [46] The amplified signal can be seen directly by naked eyes Such colorimetric methods are superior to fluorescent methods since instrumentation is unnecessary and gold nanoparticle tag does not suffer bleaching However, such procedure is tedious and not suitable for point of care application

Figure 3.15: Basis of silver and gold development [46]

3.5.2 Label free technique of protein detection

Recent progress in label free detection includes efforts in bringing fluorescent immunoassay towards portability by reducing the weight and size of the readout instruments [47] Another worthy effort to push for label free detection is to overcome the drawbacks of existing labelled detection techniques Firstly, chemical labelling of proteins might change their physical characteristics (pI, hydrophobicity, conformation, etc) and impair their native function and activity, especially for the small proteins or peptides containing only a few epitopes Secondly, variation in labelling efficiency for different proteins is very likely to render inaccurate quantitative study Thirdly, the labelling procedure is time

consuming and labour intensive Thus, studies on surface binding events have been conducted using label free methods, such as electrochemical methods, atomic force microscopy (AFM), surface plasmon resonance imaging (SPRi) and mass spectrometry [22] Ellipsometry has also been commented as a feasible system of unlabelled protein detection [48] The disadvantages of using these systems involve the significant amount of capital and operating cost Most importantly, they are currently not suitable for high throughput detections Hence,

a way of detecting unlabelled protein economically and efficiently is necessary

By eradicating the reliance on instrumentation, unlabelled techniques can also be developed into competitive point of care applications

Developing a protein microarray with nematic liquid crystal birefringence as the read out can realizes multifold advantages It is not just label-free, but can also eliminate the reliance on instrumentation and makes the procedure to obtain the readout easy for users without lab training

3.5.2.1 Liquid crystal based microarray

Thermotropic liquid crystals (LCs) exhibit colourful birefringence and can be exploited as a label free read out platform for protein detection The presence of protein can be reflected by the dark to bright change of the optical signal The