WO2016139479A1 - Sers active nanoparticles - Google Patents

Abstract

The present invention relates to functionalised metal surfaces, such as nanoparticles for use in methods of detecting analytes, as well as detecting analytes in a sample, using the functionalised sufaces/nanoparticles. In particular the invention relates to a method of detecting a target analyte in a sample using Raman spectroscopy, the method comprising: mixing an analyte binding agent functionalised metal surface with a sample, in order to allow any analyte which may be present in the sample to specifically bind to the binding agent, such that upon binding the analyte is positioned so as to be within 0 nm - 10 nm of the surface, such that an enhancement in Raman scattering of the analyte can be obtained; and directly detecting any bound analyte by way of Raman spectroscopy. The invention further relates to a SERS enhancing metal surface (such as the surface of a nanoparticle),functionalised with at least one analyte specific binding agent, wherein the analyte binding agent is bound to the surface, such as the surface of the nanoparticle such than when said analyte binds the analyte binding agent, the analyte is orientated to be in close proximity to the surface, such as the surface of the nanoparticle, wherein the binding agent is an antibody or antibody fragment and wherein at least one free side chain amine group on the antibody or antibody fragment is modified to comprise one or more reactive group(s), such as an isothiocyanate group(s), for reaction with the surface, such as the surface of the nanoparticle.

Description

SERS Active Nanoparticles

Field of the invention The present invention relates to functionalised metal surfaces, such as nanoparticles for use in methods of detecting analytes, as well as detecting analytes in a sample, using the functionalised sufaces/nanoparticles.

Background to the invention

Metal nanoparticles (NPs) are an ideal surface for conjugation of biomolecules due to the ease of functionalisation. Biomolecule functionalised NPs allow for the specific detection of a target by plasmon enhanced spectroscopies. Commonly used capture biomolecules which have been conjugated to NPs include nucleic acids,¹ peptide aptamers,² and antibodies,³ demonstrating very high selectivity and specificity.⁴ However, the detection of small molecules is not as straightforward.

US2005/0089901 describes a Raman-active reagent comprising a Raman-active reporter molecule which includes a reactive group, a binding molecule and a surface enhancing particle capable of causing SERS. However, detection is based on the SERS signal generated by the reporter molecule rather than any target analyte.

US2006/0050268 describes the use of nanosensors for detecting target analytes by binding the analytes to the surface of a SERS active nanoparticle. However, the nanoparticles are described as being modified so as to include relatively non-specific chemistries which may bind a variety of target analytes and so may result in a variety of SERS spectrums being detected. This may be problematic, particularly where low amounts of analyte are to be detected.

WO201 1/053247 describes the use of nanoparticles which have functionalised by aryl boronic acid and related derivatives to be capable of binding glucose and subsequent detection of the bound glucose using SERS

WO201 1/078794 describes the use of antibody functionalised nanoparticles for use in detecting relatively large molecules which may be present in a sample. The antibodies are bound to the surface of the nanoparticles through a terminal carboxy group and the antibodies are oriented away from the surface of the nanoparticles.

WO2012/161683 describes a method of detecting analytes using nanorods and associated nanoparticles which are capable of binding the nanorods. The nanoparticles are functionalised to include a Raman signal generator, such as a dye or the like. The analyte to be detected may bind to the surface of the nanoparticle through use of an antibody which is specific for the analyte and the nanorod may include a further antibody designed to bind to the analyte through a different region. TNT has recently been detected using fluorescence resonance energy transfer (FRET) in a sandwich immunoassay.⁵ FITC labelled TNT specific DNA aptamer was coated onto a poly-L-lysine microtitre plate allowing for capture of TNT followed by the introduction of a fluorescently labelled antibody specific for TNT. When TNT was present, FRET occurred between the labelled aptamer and the labelled antibody, resulting in a change in fluorescence emission compared to when TNT was absent. However, fluorescence and chemiluminescence spectroscopy has the disadvantage that a broad spectrum is obtained. Surface enhanced Raman scattering (SERS) produces a unique 'fingerprint' spectrum in which the peaks are narrow, this has advantages for multiplexing, i.e. the detection of multiple target analytes simultaneously. Moreover, SERS is a sensitive technique that has the ability to detect single molecules.⁶ Capturing and then detecting small target analytes using functionalised NPs in combination with SERS is very challenging and little has been reported in the literature, and often labelling/indirect detection is required. TNT has been detected directly by utilising molecular imprinted polymers (MIPs).⁷ MIPs contain artificial recognition molecules for which target analytes of interest can be detected. In this case, the commercially available gold SERS substrate, klarite was used to immobilise the amine rich MIP to which target TNT binds, forming a meisenheimer complex, resulting in a SERS spectrum of TNT specific peaks. This method was very powerful, highly specific and non-invasive, however the sensitivity of the assay was only in the μΜ range.

Detection of small molecules using SERS and functionalised metallic NPs has the potential to provide highly selective and sensitive detection. However, a key challenge is the immobilisation of a biosensor on the surface of the NPs, as biomolecules usually carry several reactive groups. Many immobilisation methods have been reported, for example Johnson et al. used protein G encapsulated NPs for subsequent Ab binding,⁸ and El Sayed et al. reported the conjugation of an Ab directly onto NPs surface by using a pH correction method to conjugate the multiple amines to the NP.⁹ However, these methods result in the Ab being orientated on the NPs surface, such that the target molecules are either held several nm away from the metal surface or that the binding site is unavailable to the target. This orientation is not ideal for detection by SERS as the SERS response is distance dependant.¹⁰

It is amongst the objects of the present invention to obviate and/or mitigate at least one or more of the aforementioned disadvantages.

Summary of the invention

The present invention is based on the development of a method of capturing target analytes so they are close to a SERS active surface and their Raman signal is enhanced, such that very low levels of the target analyte in a sample can be detected. In a first aspect there is provided a method of detecting a target analyte in a sample using Raman spectroscopy, the method comprising: a) mixing an analyte binding agent functionalised metal surface, such as the surface of a nanoparticle with a sample, in order to allow any analyte which may be present in the sample to specifically bind to the binding agent, such that upon binding the analyte is positioned so as to be in close proximity to the surface such that an enhancement in Raman scattering of the analyte can be obtained; and b) directly detecting any bound analyte by way of Raman spectroscopy.

The analyte to be detected may be any analyte which can be captured by a suitable binding agent, such as a nucleic acid, protein, carbohydrate, aldehyde, thiol, amine, explosive, drug of abuse, therapeutic agent, metabolite, environmental pollutant and the like. However, the present invention is particularly suited to the detection of analytes which typically display a relatively weak Raman Scattering signal and/or need to be present in relatively high concentrations, in a sample, in order to be detected. The present invention also allows for the detection of one or more analytes which may be present in complex mixtures. For example, a sample, such as waste water may comprise a great many different chemical entities and hence any particular analyte or analytes may be present in a complex mixture of chemicals. The present invention however, allows for such analyte(s) to be detected, even when present in such complex mixtures. As described in more detail hereinafter, the present inventors have been able to detect the explosives 2-methyl-1 ,3,5-trinitrobenzene (TNT) and Research Department explosive (RDX), formally cyclotrimethylenetrinitramine, also called cyclonite, hexogen, or T₄ and [3-Nitrooxy-2,2-bis(nitrooxymethyl)propyl]nitrate (PETN). An non explosive example is also demonstrated using 2-Aminomethyl-1 ,4-benzo is a dioxane which is very toxic to the environment and are formed in many manufacturing processes. Both the explosives and dioxane are small in size and molecular weight and obtaining Raman scattering spectra of small molecules such as these would be conventionally difficult and/or require high concentrations of the molecule, such as at least 2μΜ to be present in a sample. This of course may not be practical in many instances. The present inventors have however developed a method whereby small molecules, typically less than 1000 g/mol, such as less than 750, 500, 400 or even 300 g/mol, but greater than 50g/mol. can be detected in concentrations of less than 1 μΜ, such as less than 100, 75, 50, 25, 20, 10 or even 1 nM, but typically greater than 20, 50, 75 or 100 fM.

Without wishing to be bound by theory, this has been achieved by capturing an analyte such that it is held in close proximity to a metal surface, such that an enhancement in Raman scattering of the analyte can be obtained. Desirably, the analyte, such as a small molecule as described above, is positioned, when bound to the binding agent, within 0 nm - 10 nm, such as 0 - 5 nm from the surface. The binding agent may be substantially parallel to the surface. It may also be desirable for efficient enhancement of Raman scattering for any analyte to be orientated to be substantially perpendicular to the surface of the nanoparticle. An enhanced Raman signal of the analyte is directed detected, rather than the binding agent or change in its conformation, label or dye which is often applied in the art to SERS active surfaces, such as nanoparticles in order to provide a detectable signal.

The sample containing the analyte to be detected may be an environmental sample, for example water sample, air sample, or soil sample. The sample may also be, for example a swab or other sample taken from a surface, or from an individual (e.g. fingerprint or sample of hair or bodily fluid) or an individual's clothing or footwear. The present invention can also be used to detect an analyte in a sample from an animal, for example a tissue sample, a hair or wool sample, a urine sample, a blood sample, saliva sample or a faecal sample. The sample may comprise more than one analyte to be detected and/or more than one analyte binding agent may be provided in order that more than one analyte is capable o? being detected. Thus, the present invention may in some embodiments be considered as capable of multiplexing. In accordance with a further aspect there is provided a SERS enhancing metal surface such as a metal nanoparticle functionalised with at least one analyte specific binding agent, wherein the analyte binding agent is bound to the surface, such as the surface of the nanoparticle such than when said analyte binds the analyte binding agent, the analyte is orientated to be in close proximity to the surface such as the surface of the nanoparticle.

Suitable SERS active surfaces are known to the skilled addressee, such as described in "Principles of Surface-Enhanced Raman Spectroscopy", Chapter 7 (2009) Authors Le Ru and Etchegoin.

Separate metal nanoparticles may be provided for the detection of different analytes. Alternatively, the metal nanoparticle may be functionalised with more than one analyte specific binding agent in order to be able to bind more than one analyte if present in a sample.

The binding agents of the present invention may be specific for the analyte or nonspecific and hence capable of binding a variety of analytes. In one preferred embodiment the binding agent has the property of specifically binding to one or more analytes of interest. The binding agent may be either a member of a pair of specific binding molecules, e.g., an antibody-hapten pair or a receptor-ligand pair. Examples of such binding agents include, but are not limited to, antibodies, integrins, adhesins, cell surface markers, T cell receptors, MHC proteins, and the like.

In a preferred embodiment the binding agent is an antibody. The antibody may include any polypeptide or protein comprising an antibody antigen-binding site, including Fab, Fab2, Fab3, diabodies, t iabodies, tetrabodies, minibodies and single-domain antibodies, including nanobodies, as well as whole antibodies of any isotype or subclass. Antibody molecules and methods for their construction and use are described, in for example Hoiiiger & Hudson, Nature Biotechnology 23(9) :1 126-1 136 (2005).

In some preferred embodiments, the antibody molecule may be a whole antibody. For example, the antibody molecule may be an IgG, IgA, IgE or IgM or any of the isotype sub-classes, particularly igG1 and igG4. The antibody molecules may be monoclonal or polyclonal antibodies. In other preferred embodiments, the antibody molecule may be an antibody fragment.

Antibodies may be obtained using techniques which are standard in the art. Methods of producing antibodies include immunizing a mammal (e.g. mouse, rat, rabbit, horse, goat, sheep or monkey) with the analyte of interest in order to raise an antibody thereto. Antibodies may be obtained from immunised animals using any of a variety of techniques known in the art, and screened, preferably using binding of the antibody to analyte of interest. However, many antibodies are commercially available, see for example, http://www.antibodyresource.com/oniinecomp.htmi. in a further embodiment, the binding agent may be an aptamer which is specific for the analyte of interest. The term "aptamer" as described herein is intended to mean a single strand of RNA or DNA that specifically binds to a particular analyte. The term "aptamer" relates to polynucleotide or oligonucleotide sequences. The terms "polynucleotide" or "oligonucleotide" may be used interchangeably and is a term commonly used and understood within the art. Those skilled in the art will readily understand that variation in the sequence code of the aptamer may be varied by standard methodology without substantially affecting the binding of the analyte to the aptamer. Aptamers are single-stranded nucleic acids (ssRNA, ssDNA), which unlike traditional nucleic acids, possess unique binding characteristics to specific targets with high affinity and specificity analogous to antibodies [Tuerk, C. Gold, L, Science, 1990, 249(4968), 505-510; Ellington, A.D., Szostak, J.W., Nature, 1990, 346(6287), 818-822.] Aptamers are isolated in vitro from combinatorial oligonucleotide libraries, typically containing 1012 to 1015 oligonucleotides, and are chemically synthesised by a process known as SELEX. The oligonucleotides are subjected to a selection process for their ability to bind a specified target and over a number of selection rounds (typically 8-16 rounds): the most specific nucleic acid sequences are isolated. Depending on the techniques used in SELEX, the process might take from days to months [Cho, E. J., Lee, J.W., Ellington, A.D., Ann. Rev. Anal. Chem., 2009, 2(1 ), 241 -264; Ellington, A. D., Ann, Rev. Anal. Chem., 2009, 2(1 ), 241 - 264.] Aptamers have been generated for a wide range of targets, ranging from ions to entire ceils. The use of an in vitro process enables the generation and selection of aptamers that can bind, for example, toxic targets, which are not possible by immunologically initiated recognition elements, such as antibodies. in addition to nucleic acid based aptamers, peptide aptamers are known in the art and may also be used in accordance with the present invention. A discussion of peptide apatmers and how they may be obtained may be found in Mascini et ai 20 1 (Angew. Chem. int. Ed., 50, 2 -19), to which the skilled reader is directed. in accordance with the teaching of the present invention it is possible to bind the various binding agents to the metal surface, such that when the analyte is bound thereto, the analyte is in close proximity to the SERS active surface. This may in some embodiments be through binding of a side group present on a nucleic acid or peptide molecule, for example, rather than a C or N terminal group.

The nanoparticles can be made of, or comprise at least one metal, such as a noble metal and/or a transition metal. Some suitable exemplary metals include gold, silver, tantalum, platinum, palladium, rhodium, copper, and mixtures thereof. The nanoparticle surface may be a naked metal or may comprise a metal oxide layer on a metal surface.

Nanoparticles may, for example, comprise or consist essentially of silver, gold or copper surfaces, especially silver, are particularly preferred for use in the present invention. However, metal alloys can also be used. The nanoparticles can be any shape and hence do not necessary have to be spherical in nature. Thus, rods, stars, cubes, raspberries or any other shape may also be envisaged. Suitable nanoparticles can range from about 1 nm to about 1000 nm in diameter. In certain embodiments, the diameter of the nanoparticles can range from about 1 nm to about 500 nm. In some embodiments, the diameter of the nanoparticles can range from about 10 nm to about 200 nm. In some other embodiments, the diameter of the nanoparticles can range from about 20 nm to about 80 nm.

The nanoparticles may be formed, for instance, by the reduction of a metal salt (eg. silver nitrate) with a reducing agent such as citrate, to form a stable nanoparticle suspension (see P. C. Lee & D. Meisel, J. Phys, Chem. (1982), 86, p. 3. 391 ); the use of EDTA (Lee, N. S.; Sheng, R. S.; Morris, M. D.; Schopfer, L. M., The active species in surface-enhanced Raman scattering of flavins on silver colloids. J. Am. Chem. Soc. 1986, 108 (20), 6179-6183); or hydroxylamine (Brown, K. R.; Natan, M. J., Hydroxylamine Seeding of Colloidal Au Nanoparticles in Solution and on Surfaces. Langmuir 1998, 14 (4), 726-728 and N. Leopold, B. Lendl, J. Phys. Chem. B 2003, 107, 5723-5727). Unlike other prior art techniques, This "stock" suspension need not be aggregated prior to use.

In accordance with the present invention, the analyte binding agent is attached, conjugated, bonded or otherwise adhered to the surface of the nanoparticles. Preferably the analyte binding agent is bonded to the surface of the nanoparticle, such as by way of ionic, covalent, hydrogen bonding, van der Waals forces, or mechanical bonding, in order to provide stability to the functionalised nanoparticles.

The analyte binding agent may comprise a reactive group or groups specifically designed to adhere or react and form a bond with the surface, such as the surface of the nanoparticle. Such a group or groups may simply be naturally present on the binding agent, or the binding agent may be modified so as to include such a group or groups. Examples of potentially suitable groups include but are not limited to a thiol, primary or secondary amines, pyridyl, imidazolium, thiophene, selenophene, isothiocyanate, multi-sulphur organic moiety (that is a molecule having two or more sulfur atoms), multi-heterosulphur organic moiety (that is a molecule having two or more heterocyclic rings each incorporating sulfur atoms), benzotriazole group, and/or combinations thereof.

In one embodiment, the analyte binding agent, such as an antibody or antibody fragment, is modified to comprise one or more reactive group(s), such as an isothiocyanate group(s) for reaction with the surface of a nanoparticie. Said one or more reactive/isothiocyanate group(s) may be provided by a moiety which specifically reacts or interacts with a group or groups present on the binding agent and in such a way as to substantially not affect binding of any analyte to the analyte binding agent. Moreover, the moiety may be chosen, such that upon reaction/interaction with the analyte binding agent and subsequent binding of the analyte binding agent to the surface of the nanoparticie, through the one or more reactive group(s) described above, the portion of the binding agent which is capable of binding to the analyte is suitably positioned and optionally oriented in relation to the surface o? the nanoparticie, such that surface enhancement of the analyte Raman signal is effected, once the analyte is bound to the analyte binding agent.

In a particularly preferred embodiment the analyte binding agent is an antibody or fragment thereof which is modified to comprise one or more reactive groups, such as isothiocyanate groups. The moiety which comprises such a reactive isothiocyanate group may for example be a multicyclic moiety which in addition to the reactive isothiocyanate group comprises a carboxylic acid group capable of reacting with a free amine moiety on the antibody. An example of such a multicyclic moiety is fluorescein isothiocyanate. This may be allowed to react and form an amide bond between the free carboxylic acid group present on the fluorescein molecule and an exposed side chain amine group present on the antibody. Such a free amine group is found in lysine groups, for example. Typically the moiety should not react with the free amine present at the amino-terminus of the antibody. By reacting with free side chain amines, one or more, such as three, four, five or more fluorescein, or other suitable molecules can be attached along the length of the antibody, rather than at the N- or C- terminus. Without wishing to be bound by theory, when the antibody is subsequently bound to the surface of the nanoparticie, it is expected to do so in a manner whereby the antigen binding portion of the antibody remains in close proximity to the surface of the nanoparticie. To better appreciate this, if an antibody were to bind to the surface of a nanoparticie through the C-terminus of the antibody, the rest of the antibody would be directed away from the surface of the nanoparticie and the antigen binding site would expected to be at least 10 -15 nm away from the surface of the nanoparticie. Such a distance is not conducive to a surface enhancement effect which is observed in accordance with the present invention. For the avoidance of doubt when a full length antibody, rather than an antibody fragment which is still capable of specifically binding an analyte to be detected, in accordance with the present invention, the antibody is not bound to the surface of a nanoparticie through its C-terminal carboxy moiety. Antibody fragments, by virtue of being smaller, may be bound to the surface of a nanoparticie through an non- analyte binding portion of the antibody fragment, as the binding portion may still be in close proximity to the surface of the nanoparticie.

In accordance with the present invention significant enhancements in Raman signal can be observed. For example, a Raman signal enhancement of at least 10², for example at least 10³, 10⁴, 0⁵, 0⁶, 10⁸ 10¹⁰, 10¹² or even higher, for one or more analytes, when bound in close proximity to the surface of a nanoparticie of the present invention, may be observed. As would be readily apparent to one of ordinary skill in the art, Raman enhancement represents the degree o? signal amplification that can be achieved during detection of a particular material. For example, Raman enhancements may allow a vibrational spectrum, or the chemical fingerprint, of low concentrations o? analytes to be detected, as described herein.

This surface enhancement effect in Raman scattering is well known in the art

Briefly, Raman scattering occurs when a light source irradiates a sample and scattered light is given off. Most of the light is scattered with the same frequency as that of the incident light but a weak component is scattered one vibrational unit different. The weak component is Raman scattering. By subtracting the frequency of the Raman scattered light from the frequency of the incident light, a vibrational spectrum characteristic of the molecule can be obtained. The light can then be detected in a suitable spectrometer, many of which are commercially available.

The detection of Raman scattering is attractive since it uses visible or near infrared radiation to provide the excitation. Moreover, flexible and effective optics can be designed and water gives a weak signal so that detection in aqueous solution is possible. Further, the set of signals obtained gives a unique pattern from which a particular analyte can be identified. However, the main disadvantage of Raman scattering is that it is not sufficiently sensitive and is not therefore generally suitable for detecting analytes at extremely low concentrations, and fluorescence can interfere with detection. The sensitivity of Raman scattering may however be improved. Firstly, if the analyte is adsorbed onto a suitably roughened metal surface as in accordance with the present invention, then there is an interaction between the analyte and the surface electron waves on the metal (plasmons) which provide an enhancement in the intensity of the Raman scattering by a factor claimed to be 10⁶. This technique is known as surface enhanced Raman scattering (SERS). The method for obtaining the Raman or SERS spectrum, may be conventional. However, the following might apply to the spectroscopic measurements: Typically, the methods of the invention will be carried out using incident light from a laser, having a frequency in the visible spectrum ie.380 nm-785 nm, particularly between 400 nm-650 nm (the exact frequency chosen will generally depend on the dye used in each case - frequencies in the red area of the visible spectrum tend, on the whole, to give rise to better surface enhancement effects). However, it is possible to envisage situations in which other frequencies, for instance in the ultraviolet (ie. 200 nm-400 nm) or the near- infrared ranges (700 nm-1300 nm) or up to 1600 nm, might be used. Thus, Raman/SERS detection may be conducted between about 200 nm-1300 nm. The selection and, if necessary, tuning of an appropriate light source, with an appropriate frequency and power, will be well within the capabilities of one of ordinary skill in the art, particularly on referring to the available SERS literature. To achieve highly sensitive detection, using SERS, a coherent light source is needed with a frequency at or close to the absorption maximum for the dye. If lower sensitivities are required, the light source need not be coherent or of high intensity and so lamps may be used in combination with a monochromator grating or prism to select an appropriate excitation frequency.

Several devices are suitable for collecting Raman/SERS signals, including wavelength selective mirrors, holographic optical elements for scattered light detection and fibre- optic waveguides. The intensity of a Raman/SERS signal can be measured for example using a charge coupled device (CCD), a silicon photodiode, or photomultiplier tubes arranged either singly or in series for cascade amplification of the signal. Photon counting electronics can be used for sensitive detection. The choice of detector will largely depend on the sensitivity of detection required to carry out a particular assay. Note that the methods of the invention may involve either obtaining a full Raman or SERS spectrum across a range of wavelengths, or selecting a peak and scanning only at the wavelength of that peak (ie. Raman "imaging"). It is also possible to detect all Raman scattering using only a filter to remove reflected light, Raleigh scattering etc and a detector such as photodiode. Apparatus for obtaining and/or analysing a Raman or SERS spectrum will almost certainly include some form of data processor such as a computer. Raman signals consist of a series of discrete spectral lines of varying intensity. The frequencies and the relative intensities of the lines are specific to the SERRS active dye being detected and the Raman signal is therefore a "fingerprint" of the analyte. If a Raman/SERS analyser is being used selectively to detect a mixture of analytes then it will be necessary to detect the "fingerprint" spectrum for identification purposes. Once the Raman/ SERS signal has been captured by an appropriate detector, its frequency and intensity data will typically be passed to a computer for analysis. Either the fingerprint Raman spectrum will be compared to reference spectra for identification of the detected Raman active compound or the signal intensity at the measured frequencies will be used to calculate the amount of Raman active compound detected.

A commercial Raman/SERS analyser of use in carrying out the invention would be expected to comprise the following components : a laser light source, the appropriate optics for carrying the light to the SERS active surface, a stage for mounting the sample for analysis, optics for receiving the Raman signal, a detector for converting the Raman signal into a series of intensities at certain wavelengths and a data processor for interpreting the wavelength/intensity data and/or multiple signal differentiation and providing an analytical output. The analyser may also comprise a database of wavelength signals representative of particular analytes in order to be able to detect and indicate to the user the analyte detected based upon the database of representative signals. Detailed description of the invention

The invention will now be further described by way of non-limiting example and with reference to the figures which show: Figure 1 :(a) SERS spectra of 80 nM (bottom), 120 nM (middle) and 2.6 μΜ (top, 250 μΙ) of TNT adsorbed onto silver nanoparticles surface (250 μΙ) after the addition of NaCI (8.8 mM, 1 μΙ), 532 nm laser excitation was used with a of 10 s accumulation time for 5 x 3 replicate samples, (b) schematic representation of low concentrations TNT on nanoparticles surface, (c) schematic representation of high concentrations of TNT on nanoparticles surface. Figure 2: SERS spectra of 150 nM DNT (top) in acetonitrile and solvent acetonitrile (bottom) only, both adsorbed on nanoparticles with the addition of NaCI (8.8 mM) using 532 nm laser excitation for 10 s for 5 x 3 replicate samples.

Figure 3: (a) SERS spectra of 0 (top), 10 (bottom), 50 (2^nd bottom) and 150 (middle) nM of TNT binding to antibody conjugated nanoparticles, 150 nM of DNT (2^nd top) was used as a control. SERS spectra was obtained by using 532 nm laser excitation for 0.5 s accumulations, spectra shown is the average of 5 x 3 replicate samples, (b) Bar chart of the SERS intensity of the TNT peak present at 1066 cm^"1 for TNT concentrations of 1 , 10, 50, 75 and 150 nM in the assay and the control samples acetonitrile (no TNT) and DNT. The error bars represent the standard deviation of 5 x 3 replicate samples, (c) SERS spectrum of 10 (2^nd bottom), 50 (2^nd top) and 150 (top) nM of TNT in the assay, FITC background subtracted and SERS of 120 nM TNT (bottom) adsorbed onto silver nanoparticles with NaCI (8.8 mM) in the absence of antibodies.

Figure 4: SERS enhancement observed at 1066 cm^"1 over time Figure 5: SERS spectra of unlabelled antibody conjugated to silver nanoparticles with 60 (2^nd bottom) and 100 (2^nd top) nM of TNT, and 100 nM TNT (top) in the presence of NaCI (8.8 mM) using 532 nm laser excitation for 0.5 s for 5 x 3 replicate samples.

Figure 6: 3D AFM images of a gold surface with (a) FITC modified antibody and (b) antibody with no ITC group present, (c) and (d) are examples of line profiles across a feature on (a) and (b), respectively. AFM images were obtained in air at room temperature using tapping mode AFM.

Figure 7: (a) SERS spectra of 500 nM RDX (bottom), silver nanoparticles conjugated to antibodies with no RDX present (top) and silver nanoparticles conjugated to antibodies with 1 .5 nM RDX present (middle), (b) Bar chart of peak intensity at 1271 cm^"1 for no RDX, 30, 150, 300, 600, 900, 1200, 1500 and 3000 pM RDX and control samples TNT and RDX and (c) SERS spectrum of 60, 90, 120, 150 and 300 pM of RDX in the assay, FITC background subtracted and SERS of 100 nM RDX adsorbed onto AgNPs and 8.8 mM NaCI. Figure 8: (a) shows SERS spectra of 170 nM of dioxane (bottom), silver nanoparticle conjugated to antibody with no dioxane present (top) and 170 nM dioxane present in assay (middle), (b) SERS spectrum of 170 nM dioxane (bottom) adsorbed onto silver nanoparticles and SERS spectrum of 150 nM RDX bound to antibody conjugated silver nanoparticles with the FITC spectrum subtracted (top).

Figure 9: (a) SERS spectra of 50 nM PETN (top) and no PETN (bottom) in silver nanoparticles conjugated to antibodies. Spectra were obtained using 532 nm laser excitation for 0.5 s accumulation time. Spectra shown is the average of 3 x 4 replicate samples (b) PCA plot of PC1 vs PC2 for no PETN (a), 10 nM PETN (b), 30 nM PETN (c), 50 nM PETN (d) and 500 nM PETN (e).

Figure 10: (a) PCA plot of PC1 vs PC2 for 50nM PETN (a), 50 nM RDX (b), 50 nM TNT (c), and 50 nM of all three samples (d) present in the multiplex assay, (b) Loading plot for PC1 (48.86 %), (c) Loading plot for PC2 (16.29%).

Experimental details All chemicals were purchased from Sigma Aldrich unless otherwise stated.

SERS of all samples was obtained using a Renishaw plate reader, RenDX, with 532 nm laser excitation from a diode laser. A 96 well plate was placed onto the stage and the instrument's software was used to move the stage automatically so that a spectrum could be recorded from each well. 3 replicate 150 μΙ samples were analysed for each sample using a 96 well plate. The accumulation time was 0.5 s for 5 x 3 replicate samples using 0.15 mW laser power, unless otherwise stated. All SERS spectra were normalised against a standard solution of ethanol. Each spectrum was baseline corrected using the asymmetrical least squared smoothing method and then each spectrum was scaled. Synthesis of Citrate Reduced Silver Nanoparticles

500 ml of distilled water was heated to 45 °C with constant stirring. Silver nitrate (90 mg) was dissolved in 10 ml distilled water and added. Heating was continued 5 until the temperature reached 98°C, then 100 ml of a 1% aqueous solution of sodium citrate was added. The solution was stirred for 90 mins, maintaining a temperature of 98 °C throughout. The colloid was then cooled to room temperature.

SERS of TNT/ RDX/ PETN

TNT, RDX and PETN were purchased from LGC standards, UK (4.4 μΜ diluted in acetonitrile). Final concentrations of 120 nM, 720 nM, 1.3 μΜ, 2 μΜ, 3μΜ and 3.5 μΜ (100 μΙ) of TNT/RDX/PETN were diluted with acetonitrile and added to 150 μΙ of AgNP solution. NaCI (8.8 mM, 1 μΙ) was then added as an aggregating agent in order to obtain SERS spectra. SERS spectra of TNT/RDX/PETN were obtained by increasing the accumulation time to 10 s.

SERS of Control Samples

As TNT, RDX and PETN were dissolved in acetonitrile solvent, the SERS spectrum of acetonitrile was obtained. The spectrum of 2, 6 - dinitrotoluene (DNT), also purchased from LGC standards, UK (4.4 μΜ dissolved in acetonitrile) to check the specificity of the antibody. SERS of DNT and acetonitrile were also obtained by increasing the accumulation time to 10 s.

Conjugation of Silver Nanoparticles and TNT Specific Antibody

FITC labelled TNT/RDX specific antibodies were purchased from Fitzgerald, USA. The antibody was modified with 4.5 fluorescein isothiocyanate (FITC) groups per an antibody. 20 μΙ of 2 mg/ml Ab were added to 1 ml of citrate reduced 5 nanoparticles and kept at 4°C for 48 hours. The silver nanoparticle-antibody conjugates (AgNP-Ab) were then centrifuged (6000 rpm, 20 mins) and resuspended in tris buffer (pH 7.4). Addition of Target TNT/RDX/Dioxane to silver nanoparticle-antibody conjugate (final concentrations of 1 , 10, 50, 75 and 120 nM) were diluted in acetonitrile and added to 150 μΙ of AgNP-Ab. The binding of the small molecule to the Ab proceeded instantaneously and therefore analysis via SERS was carried out directly after addition.

Enhancement of SERS assay over time

The samples were analysed by SERS every 10 minutes for 100 minutes and the enhancement observed at 1066 cm^"1 was then plotted against time. Addition of controls to silver nanoparticle-antibody conjugate samples were analysed to demonstrate the specificity of the assay, therefore the highest concentration of 150 nM DNT (100 μΙ) diluted in acetonitrile was added to 150 μΙ AgNP-Ab. Also a negative control in which 100 μΙ of acetonitrile was added to 150 μΙ of AgNP-Ab was also carried out. To demonstrate that the binding of TNT to the AgNPs-Ab conjugate and hence the subsequent instantaneous detection of TNT by SERS, the samples were analysed every 10 minutes for 100 minutes.

SERS of assay using 633 nm laser excitation

The assay was also analysed using 633 nm laser excitation. Final concentrations of 1 , 10, 50, 75 and 120 nM of RDX (100 μΙ) were diluted in acetonitrile and added to 150 μΙ of AgNP-Ab. The reaction proceeded instantaneously and therefore analysis via SERS was carried out directly after addition.

SERS spectra were obtained using a WITec Alpha 300 R confocal microscope (WITec, Ulm, Germany) with 633 nm laser excitation. A 96 well plate was placed onto the stage, ethanol was used to calibrate the system. 3 x 150 μΙ of each sample were analysed. The accumulation time was 0.5 s using 0.15 mW laser power and each well was scanned 5 times from the 3 samples using a 20 x long distance objective lens. Au SERS assay

The assay was also analysed with AuNPs instead of AgNPs. 20 μΙ of 2 mg/ml FITC labelled Ab were added to 1 ml of citrate reduced nanoparticles and kept at 4°C for 48 hours. The gold nanoparticle-antibody conjugate (AgNP-Ab) was then centrifuged (6000 rpm, 20 mins) and resuspended in tris buffer (pH 7.4). Reaction proceeded instantaneously and therefore analysis via SERS was carried out directly after. Conjugation of Unlabelled Antibody to Silver Nanoparticle

TNT specific antibody (unlabelled) was purchased from BBI solutions, UK. 1 ml of citrate reduced nanoparticles were diluted in HEPES buffer (pH 7.8). 80 μΙ of 1 mg/ml Ab were added to 2 ml of diluted citrate reduced nanoparticles and kept at 4°C for 48 hours. The silver nanoparticle-antibody conjugate (AgNP-Ab) was then centrifuged (6000 rpm, 20 mins) and resuspended in 1 ml tris buffer (pH 7.4).

AFM AFM images were obtained on a DPN 5000™ nanofabrication system using ACT silicon probes with a spring constant of 37 N/m and a resonance frequency of 300 kHz. All images were collected in close-contact (or tapping) mode under ambient conditions. The surface was then washed 3 times with H₂0. Linewise levelling was performed on the images using SPIP™ software and line profiles were taken from the resulting images. SPIP™ was also used to generate 3D images. Antibodies were immobilised on the surface by spotting 2.5 μΙ of HEPES buffer (10 mM, pH 7.4) for 30 minutes. HEPES buffer was removed from the surface and 2.5 μΙ of antibody was spotted and left to dry for 30 min. The surface was then washed 3 times with d.H₂0 to remove any unbound antibodies.

SERS of Dioxane

SERS of 2-aminomethyl-1 , 4-benzodioxane was obtained. 2 μΙ of dioxane (150 nM) was added to 150 μΙ of AgNP. The suspension aggregated and a SERS spectrum was obtained using an acquisition time of 13 seconds.

In-house FITC Modification of PETN Antibody

The FITC is conjugated to the antibody by carbodiimide chemistry. The carboxylic acid on the FITC was conjugated to the free amine groups on the PETN antibody by (1 - ethyl-3-(3-dimethylaminopropyl)-carbodiimide) EDC and N-hydroxysuccinimide (sNHS) coupling. Different concentrations of FITC were prepared (10, 30, 50 and 100 μΙ of 2 x10^"4 M) and therefore increasing concentrations of EDC (1 mg/ml, 3mg/ml, 5mg/ml and 10 mg/ml) was diluted in PBS buffer (10 mM). 435 μΙ of each EDC solution was added to the corresponding FITC solution and left to stir for 5 mins. sNHS was then dissolved (1 mg/ml) in PBS buffer. The antibody was dissolved (20 μΙ in 80 μΙ d.H₂0) and added to 435 μΙ of sNHS solution, and left to stir for 5 mins. The EDC/FITC mixture was added very slowly to the sNHS/antibody mixture, with stirring. This solution was left for 3 hours shaking. The antibody conjugates were then cleaned up using a cartridge centrifugation (10,000 Da), and HPLC (C4 column, a gradient mobile phase was applied from 4.9% acetonitrile , 1 % trifluoroacetic acid, 95% d.H₂0 to 99% acetonitrile, 1 % trifluoroacetic acid.) Multiplexing RDX, TNT and PETN Specific Antibodies

To prove the selectivity of the assay 3 x 75 μΙ of RDX specific Ab AgNP was added to 75 μΙ of TNT specific Ab AgNP solution and 75 μΙ of PETN specific An AgNP solution was added. Then only RDX or TNT or PETN (50 nM, 100 μΙ) was added to the nanoparticle suspension. In the final sample all three target analytes were added. Also a negative control in which 100 μΙ acetonitrile was added to 150 μΙ AgNP-Ab was also carried out.

PCA

As mentioned previously, the data set was normalised and scaled before performing PCA,¹¹' ¹² using Matlab software version R2013a (The MathWorks, Natick, MA, USA). Principal component 1 was then plotted against principal component 2. The loading plots were also plotted to determine correlation.

Results and Discussion:

To demonstrate the applicability of the present invention to small molecules, the inventors have developed an assay for detecting specificity of this assay, the analytes 2, 4, 6-trinitrotoluene (TNT), hexahydro-1 , 3, 5-trinitro-1 , 3, 5-triazine (RDX), [3- Nitrooxy-2, 2-bis(nitrooxymethyl)propyl] nitrate (PETN) and 2-Aminomethyl-1 ,4-benzo (dioxane). TNT, RDX and PETN are small molecule targets which are of high interest as there remains a continuing threat of terrorist/insurgent attack on military/civilian personnel and key strategic infrastructures. TNT, RDX and PETN are nitro based molecules which are commonly used as concealed explosives. Dioxanes are highly unstable due to their peroxide characteristics. Dioxanes along with furanes are one of the most toxic chemicals known. In general dioxanes are an unintentional by product of many industrial processes. Therefore, a simple, rapid test for these compounds which has an unambiguous result, is affordable and can be widely deployed is highly desirable.

Synthesis of Citrate Reduced Silver Nanoparticles

The nanoparticles obtained by citrate reduction were measured to be 64 ± 4.26 nm in diameter using dynamic light scattering (DLS). The zeta potential of the nanoparticles was also measured in order to determine the stability of the nanoparticle and surface charge. For the AgNP synthesised, the zeta potential was recorded to be -38.8 ± 1 1 .1 10 mV. This confirms that the nanoparticles were stable and the negative charge is to be expected as the nanoparticles were capped with a negatively charged citrate layer. The extinction profile, shows the extinction band was present at 410 nm which is typical for this size of silver nanoparticle.

SERS of TNT

A SERS spectrum of TNT was obtained by adding a range of concentrations of TNT to silver nanoparticles in the presence of an aggregation agent. A range of TNT concentrations (50 nM - 3.5 μΜ) were added to silver nanoparticles (0.326 nM) and NaCI (8.8 mM) was added to aggregate the nanoparticles. Analysis by SERS was then performed by interrogating the sample, 532 nm laser excitation was used, with an accumulation time of 10 s. Figure 1 demonstrates the SERS spectra of three different concentrations of TNT adsorbed onto silver nanoparticles. The black spectrum (Figure 1 (a)) represents 80 nM of TNT adsorbed onto silver nanoparticles. Raman bands were observed at 923 and 1376 cm^"1 , which are due to the presence of the acetonitrile in which the TNT was dissolved. No TNT specific peaks were observed at this low concentration. However, when the concentration of TNT was increased to 120 nM (red spectrum) a strong band at 1066 cm^"1 was observed. This band was assigned to being due to the asymmetric staggering of the CH₃.¹³ There were also peaks present at 1560 and 1618 cm^"1 , which are due to the asymmetric N0₂ vibrations. The band present at 1178 cm^"1 is due to the stretching of the phenyl ring of TNT. Furthermore, Clarkson et al. have previously performed DFT calculations on TNT, and the peaks present in the spectrum shown in Figure 1 are in agreement, within ± 5 cm^"1, to the theoretical values reported.¹³ At 2.6 μΜ of TNT (Figure 1 (a), blue spectrum) there were peaks observed at 1 148 and 1269 cm^"1, which were assigned to the symmetrical vibrations of CN and the 2- and 6- N0₂ groups. The change in the SERS spectrum, with increasing TNT concentrations suggests that the TNT is changing orientation, to allow for more molecules to be adsorbed onto the surface. It is suggested that the TNT molecules are change from a 'flat' orientation to a 'standing' orientation (Figure 1 (c)). This standing orientation would result in less SERS peaks, as the C-C stretches in benzene ring would now be perpendicular to the nanoparticle surface. These C-C stretches are not as Raman active as N0₂ vibrations as they are not as polarisable. Furthermore, it was found at higher concentrations of TNT, strong enhancements were also observed for acetonitrile. This was thought to be due to the 'upright' orientation of the TNT molecules. When the TNT coverage in this 'upright' orientation on the nanoparticle is below monolayer, there is space for acetonitrile to be adsorbed onto the surface, resulting in the enhancement of acetonitrile peaks.

Table 1 : Summary of TNT peak assignments

Wavenumber (cm ¹) Peak Assignment

1066 CHA asyrornetricat

1178 Phenyl ring 1560 NO_≤ asymmetrical

stretch ina

1596 C-C ring vibration 1618 N(¾ asymmetricgl

stretching

SERS of Control Samples Since TNT, RDX and PETN was dissolved in acetonitrile, SERS was also obtained of acetonitrile adsorbed onto nanoparticles under the same conditions as for TNT. DNT was used as a control to ensure the specificity of the Ab towards TNT, therefore the SERS spectra of DNT was also obtained as shown in Figure 2. In Figure 2, the red spectrum denotes the SERS spectrum of acetonitrile, there are peaks present at 923 and 1374 cm^"1, from C-C stretching and CH₃ deformation respectively. As DNT is also dissolved in acetonitrile these peaks are also present in the DNT spectrum. The main peaks associated with DNT are 1310 cm^"1 due to the N0₂ stretching and the peak present at 986 cm^"1 is due to the aromatic ring. DNT has a different spectrum from TNT due to DNT having two strong electron withdrawing groups opposed to three in TNT. Therefore, the methyl group has much less steric hindrance, and hence there is no staggering of molecules, therefore making DNT a planar molecule.

Conjugation of Silver Nanoparticles and TNT Specific Antibody

The specific antibody for TNT was purchased conjugated to the fluorophore, fluorescein isothiocyanate (FITC). Characterisation of the AgNPs and the AgNP-Ab was carried out using UV-Vis spectroscopy to ensure that the Ab was bound to the surface of the nanoparticles. As determined, upon the addition of Ab there is a slight red shift of 2 nm from 409 to 41 1 nm in the extinction peak. The peak has also very slightly broadened in size indicating the conjugation of Ab to AgNP. It can be seen in the spectrum a small shoulder at 495 nm which is due to the presence of the FITC conjugated Ab which has an absorbance maxima at 495 nm. This technique suggests that the antibody is on the surface of AgNP. Figure 3 represents the SERS analysis, when no TNT was present, a background spectrum from the FITC modification on the antibody was observed. However, when three different concentrations of TNT were present in the assay, shown in Figure 3(a) (10, 50 and 150 nM TNT) it was observed that there was an enhancement in the peak at 1066 cm^"1. Interestingly, as the concentration of TNT was increased in Figure 3(a), from 0 to 150 nM of TNT, the peak counts at 1066 cm^"1 also increased. This was further demonstrated in the bar chart in Figure 3(b), in which the peak intensity at 1066 cm^"1 was plotted against concentration. As the concentration of TNT was increased in Figure 3(b), the peak present at 1066 cm^"1 was also enhanced respectively. Surprisingly, this peak at 1066 cm^"1 was attributed to TNT, assigned to the asymmetric staggering of CH₃ which can be observed in the spectrum of TNT alone (Figure 1 ).¹³ This highly suggests, that the antibody orientation on the nanoparticle surface is allowing for TNT to come into close proximity to the nanoparticle surface, resulting in the direct detection of TNT by SERS. To further investigate whether the TNT spectrum could be observed in the overall spectrum, the FITC background spectrum was subtracted. The spectra obtained from the 10, 50 and 150 nM of TNT in the assay had the FITC background spectrum subtracted (Figure 3(c)). This was achieved by directly subtracting the control acetonitrile spectrum, from the spectra shown in Figure 3, as each was scaled to the background FITC spectra. Interestingly, in Figure 3(c), the same peaks were present at 720, 1066, 1 179 and 1558 cm^"1, which were assigned to be from TNT in Table 1 . Thus, indicating the spectrum of TNT is clearly observed in this detection assay. This result is highly significant as the molecular SERS vibrations from the TNT target was very strongly present in the spectrum, opposed to being masked by the fluorescent modification on the antibody.

Furthermore, DNT was used as a control as it is chemically very similar to TNT. Therefore, it can be used to prove the specificity of the assay. DNT was added into the assay at the highest concentration of TNT explored (150 nM). In Figure 3(a), pink spectrum, it can be seen that there was only peaks present which represents FITC. However, in the bar graph in Figure 3(b) it can be seen that DNT (150 nM) did have a small enhancement at 1066 cm^"1. This could be due to the antibody having a slight cross reactivity with DNT. Furthermore, by using 2, 6 - DNT as a control over 2, 4 - DNT, 2, 6 - DNT has a rotated out of plane methyl group due to the steric hindrance of the two N0₂ groups, the peak at 1066 cm^"1 was attributed to. However, this enhancement is miniscule in comparison to the 10 nM TNT enhancement. More specifically, at 10 nM TNT there was over a 7 times enhancement in signal in comparison to no TNT present. Furthermore, at 150 nM of TNT there was a large enhancement of over 17 times in comparison to no TNT present.

This suggests that the use of the ITC group on the antibody plays an essential role as it allows for the antibody to be conjugated to the surface of the nanoparticles in an unusual orientation. The antibody was immobilised on the nanoparticle in a 'flat' orientation, allowing for the target molecule to come in close proximity to the nanoparticles surface and crucially, that the binding of the target to the antibody conjugated silver nanoparticles resulted in a SERS spectrum being obtained specifically for the target analyte. If the antibody was immobilised to the nanoparticle by the Fc chain, in the 'standing' position, the TNT would be orientated too far from the metal surface for a spectrum of TNT itself to be obtained.

SERS Enhancement Over Time

To demonstrate the binding of TNT to the modified antibody conjugated silver nanoparticles and hence the subsequent detection of TNT by SERS, the samples were analysed over time (every 10 min for 100 min) to observe the optimum time required for maximum TNT binding. Concentrations ranging from 0-150 nM of TNT was added to the antibody conjugated silver nanoparticles (150 μΙ).

Figure 4 shows a line graph of the SERS enhancement observed due to the peak at 1066 cm^"1 over time. Control samples were also analysed to determine if the spectra changes over time. The control samples explored consisted of: antibody conjugated silver nanoparticles, no TNT, shown in blue in Figure 4. Acetonitrile was added to the antibody conjugated silver nanoparticles, black spectrum in Figure 4. And finally, DNT was used as a control, shown red in Figure 4. The control samples showed no discrimination at the peak at 1066 cm^"1 over time, hence no TNT was detected. However, for the 1 nM of TNT (pink), a very slight enhancement at around 60 min after addition of TNT was observed. Interestingly, at 10 nM TNT (green), there was an enhancement observed directly after the addition of TNT, which steadily increases and levels out after approximately 70 mins. Concentrations higher than 10 nM of TNT demonstrate a similar trend, in that there is an enhancement observed immediately after addition of TNT into the assay, and that this enhancement is optimum after about 70 mins. However, for explosive detection, real-time sensing is very highly desirable; therefore all the results shown in this thesis are taken immediately after addition of TNT to the assay. However, if the application of this assay allows for time, it was demonstrated in Figure 4, that the optimum time to allow for the TNT molecule to bind to the antibody was 70 min.

SERS of assay using 633 nm laser excitation

The assay was also analysed using 633 nm laser excitation. When TNT is absent there are only peaks present which can be assigned to acetonitrile. When TNT was present in the assay no TNT specific peaks were observed FITC maximum occurs at 495 nm therefore FITC is not in resonance with 633 nm laser excitation. Hence, FITC does not contribute to the enhancement of the SERS spectrum in the TNT/RDX assay in comparison to using 532 nm laser excitation wavelength. Furthermore, another control was carried out using gold nanoparticles (AuNP) instead of AgNPs. There are only acetonitrile peaks present in both samples. This is due to AgNPs having a higher scattering to absorption ratio and hence making gold a weaker Raman scatterer.

Conjugation of Unlabelled Antibody to Silver Nanoparticle

In Figure 5, the black spectrum represents no TNT present in the assay, only the antibody bound silver nanoparticles. The peaks present at 1 100 and 1128 cm^"1 were thought to be from the presence of the antibody, which could be detonated to be C-C aliphatic stretching.¹⁴ However, when TNT was present at 60 and 100 nM (red and blue spectra) there were no TNT peaks present, only acetonitrile peaks. This could be due to random orientation of the antibodies on the nanoparticle surface, resulting in TNT being too far from the metallic surface to allow for a SERS spectrum. Another possibility is that the random orientated antibodies could be immobilising 'head on' on the nanoparticle surface. Therefore, the binding site on the antibody is not available for TNT, hence no detection of TNT was observed. However, upon the addition of NaCI solution to the assay, the nanoparticles aggregated and a small shoulder may be observed at 1066 cm^"1 , which has been assigned to the out of plane methyl stretching of TNT. This could be due to the nanoparticle coming in close proximity to each other and forming 'hotspots'. These 'hotspots' can then enhance molecular vibrations, i.e. the bound TNT to the antibody conjugated nanoparticle, hence a small shoulder at 1066 cm^"1.

This suggests that the antibody was successfully conjugated to the surface of the nanoparticles in a random orientation. However, no SERS spectrum of TNT was obtained using this adsorption method. This result demonstrates that the ITC group is essential for direct detection of small molecules by SERS as the ITC group allows for the antibody to orientate 'flat' on the surface of silver nanoparticles, so that the target molecule is in closer proximity hence an surface enhanced Raman spectrum of the target can be observed.

AFM In Figure 6 (a), numerous small features can be observed which have an average height of 4.73 ± 1 .14 nm. This average height was taken from the height profiles (Figure 6(c)) of 36 features across the area where the FITC labelled antibody had been spotted. Considering the dimensions of an antibody (14.5 nm x 8.5 nm x 4.0 nm), this 4.73 nm height strongly indicates that the antibody is orientated 'flat' on the surface. This orientation is consistent with the SERS data, as a target TNT spectrum would only be observed if upon binding to the antibody, the TNT was brought into close proximity to the metallic surface. Figure 6(b) demonstrates the AFM image of the antibody with no ITC modification. Several features around 14.06 ± 1.80 nm can be observed, (Figure 6(d)) which is close to the estimated height of an antibody in the 'tail on' or 'head on' orientation. However, there are also smaller features present, at around 4 nm, and a larger feature around 28 nm, which is indicative that the antibodies are randomly orientated on the metal surface. The average height calculated was taken from 12 features across the area where the antibody had been spotted. It should be noted that significantly less antibodies were observed that were conjugated to the gold surface when using the pH correction method compared to the ITC modified antibody. This is because the pH method results in much weaker binding to gold surface than the formation of the ITC-metal bond. This is because the pH method resulted in binding to the gold surface that is much weaker than formation of a thiol-metal bond. This resulted in more antibodies being bound to the surface in a specific orientation when the ITC group was used.

RDX Assay

Figure 7(a) shows that there was FITC peaks present in the SERS spectrum at 1170 cm^"1 due to the C-OH stretch,¹⁵ 1414 cm^"1 due to CH₂ stretching,¹⁶ 1476 and 1555 cm^"1 due to C-C ring deformation and finally 1640 cm^"1 due to C=N, which was to be expected from the previous TNT antibody assay. When RDX was present in the assay, there was enhancement of peaks observed at 1271 and 1500 cm^"1 (Figure 7(a), blue box) which were not present when RDX was absent. These bands were assigned to N- O stretching of RDX.¹⁷ This was further demonstrated in the bar chart in Figure 7 (b). As the concentration of RDX was increased in the assay, the peak intensity at 1271 cm^"1 also increased.

However, in the RDX assay we observe cross reactivity with other explosive molecules such as TNT and DNT (Figure 7(b)), which were used as controls to demonstrate the specificity. TNT and DNT were added (150 nM) to the RDX specific antibody conjugated silver nanoparticles and the SERS response was also observed. It was shown in Figure 7(b) that DNT had a high cross reactivity, as did TNT for the antibody designed specifically to detect RDX, as an enhancement at 1271 cm^"1 was observed. This was thought to be from the fact that the explosive molecules are very similar, therefore when a single peak is analysed in this case at 1271 cm^"1 , due to the N0₂ groups, both TNT and DNT possess N0₂ groups, therefore some cross reactivity was observed. This is not ideal, however the antibody was more selective towards RDX. For example, at a concentration of 150 nM of TNT, the same enhancement as 300 pM of RDX was observed, and 150 nM DNT had the same enhancement as 600 pM of RDX. Therefore, the antibody was still preferentially binding to RDX, however analysis of more than one peak should be performed in order to distinguish between specific explosive materials. Despite this cross reactivity, this result strongly agrees with the results for TNT, as RDX can be detected directly using this method. Furthermore, in Figure 7(c) the background FITC spectrum was directly subtracted from the RDX assay spectra (Figure 7 (a)) in order to determine if other RDX specific peaks were observed. Whilst there are still some background FITC peaks present, there are also peaks present at 1271 and 1500 cm^"1 which were also enhanced and showed concentration dependence.

This suggests that RDX was binding to the antibody and being held close to the nanoparticles surface, allowing for detection by SERS due to the orientation of the antibody. An observed limit of detection of 600 pM RDX was obtained which produced a SERS response of 5 times greater than when RDX was absent. This result is very exciting as it demonstrates that both the small molecule targets bound in a similar manner to the FITC labelled antibody and could be detected at very low concentrations.

SERS of Dioxane

The SERS of dioxane shown in Figure 8(a), shows there were peaks present at 732 cm^"1 assigned to C-C-C deformation, 879 cm^"1 due to ring bending, 1 1 19 cm^"1 from C-C stretching and 1256 cm^"1 assigned to C-H bending.¹⁸

The blue spectrum in Figure 8(a), showed the antibody functionalised silver nanoparticles with no target, dioxane. As described previously, only FITC peaks were observed due to the modification on the antibody. However, when 170 nM of dioxane was added to the antibody conjugated suspension, peaks were observed 732, 879, 11 19 and 1256 cm^"1 which were assigned to dioxane. Furthermore, Figure 8(b) shows that when the FITC background was subtracted from 170 nM dioxane in the assay, dioxane specific peaks were clearly present (blue box).

Therefore, the assay developed showed that the ITC group modification on the antibody is essential for allowing the target molecule to selectively come in close proximity to the nanoparticles surface. Furthermore, this assay has been shown to work for three small molecule examples, supporting that with changing the specific antibody, this assay can potentially be used for the detection any small target analyte.

FITC Modification of PETN Antibody Assay

The FITC modified antibodies specific for TNT and RDX were commercially available. However, the FITC modified antibody specific for PETN was made in-house. This involved the use of carbodiimide chemistry to conjugate the COOH group on FITC to the free amine groups present in an antibody by making an amide bond. This should result in the antibody being modified the same way as the commercially available TNT and RDX antibodies, such that the antibody should bind to the silver nanoparticle surface in the novel, 'flat' orientation. It can be seen in Figure 9, that when no PETN is present in the assay (black), that a FITC spectrum could be observed. This was expected, illustrating the successful modification of the antibody. When PETN was present in the assay, a small shoulder at 643 cm^"1 , assigned to N0₂ rocking of PETN. Furthermore, multivariate analysis in the form of PCA was performed. PCA assesses the reproducibility of the data set and reduces the dimensionalities of the data, making it easier to identify any variations in the spectra. After PCA was performed principal component 1 was plotted against principal component 2. The PCA plot for no PETN (blue) shows a distinct grouping from when 10 nM PETN (green), 30 nM PETN (red), 50 nM PETN (yellow) and 500 nM PETN (purple) was present, demonstrating that the spectrum obtained was different, and hence PETN can be detected.

Multiplexing RDX, TNT and PETN Specific Antibodies

Moreover, the specificity of this assay was demonstrated by adding TNT, RDX and PETN specific AgNP-Ab conjugates together in order to determine if each explosive could be detected and differentiated in the presence of both Ab conjugates. Each target was introduced into the multiplex AgNP-Ab solution individually and finally all three targets were added and the SERS spectrum was obtained. RDX was detected through the presence of the peaks at 1271 and 1500 cm^"1. TNT was also detected by positive identification of the peaks present at 1066 cm^"1 and 1558 cm^"1. PETN was identified by the peaks at 643, 756 and 981 cm^"1. Furthermore, multivariate analysis was performed on the multiplex data in the form of principle component analysis (PCA), Figure 10.

In Figure 10, the PCA plot of PC1 vs PC2 demonstrates individual groupings for each of the targets. 50 nM PETN (yellow) can therefore be said to be spectrally different from 50 nM RDX (green) which is spectrally different to 50 nM TNT (blue). Finally, 50 nM of all three samples (red) present in the multiplex assay is also grouped separately. This demonstrates that all targets can be detected specifically in this multiplex assay. Figure 10(b) demonstrates were the spectral differences arising in Figure 10(a). In conclusion, current sensitive detection methods on trace explosive materials involve an indirect or a labelling approach, which is not ideal for multiplex detection of several targets simultaneously. We have designed a new class of direct SERS detection immunoassay which offers quantitative results, is fast, simple, highly specific and selective towards targeting small molecules. The use of the ITC group on the Ab plays an essential role in direct detection of target molecules as it allows for the Ab to conjugate to the NPs surface in a unique flat orientation which we have characterized using AFM. This orientation allows for the target molecule to come in close proximity to the NPs surface and it can hence be detected by SERS. The assay was used for the detection of small molecule targets, TNT, RDX and PETN, which demonstrates the power of the technique as it was possible to detect the SERS molecular signature directly from these small molecules due to their orientation on the surface. The assay designed resulted in an observed limit of detection of 10 nM of TNT and 600 pM of RDX. Furthermore, with the use of multivariate analysis, we can also detect and distinguish both explosive targets from within a complex matrix i.e. when both Ab AgNP solutions are present. This approach demonstrates the use of sensors for fast, sensitive and direct detection of small target analytes. Notes and references

K. Grade, E. Correa, S. Mabbott, J. A. Dougan, D. Graham, R. Goodacre and K. Faulds, Chemical Science, 2014, 5, 1030-1040.

Z. Kuang, S. N. Kim, W. J. Crookes-Goodson, B. L. Farmer and R. R. Naik, ACS nano, 2009, 4, 452-458.

M. Sanles-Sobrido, L. Rodriguez-Lorenzo, S. Lorenzo-Abalde, A. Gonzalez- Fernandez, M. A. Correa-Duarte, R. A. Alvarez-Puebla and L. M. Liz-Marzan, Nanoscale, 2009, 1 , 153-158.

S. Schlucker, Angewandte Chemie International Edition, 2014, 53, 4756-4795.

P. Sabherwal, M. Shorie, P. Pathania, S. Chaudhary, K. K. Bhasin, V. Bhalla and C. R. Suri, Analytical Chemistry, 2014, 86, 7200-7204.

S. Nie and S. R. Emory, science, 1997, 275, 1 102-1 106.

E. L. Holthoff, D. N. Stratis-Cullum and M. E. Hankus, Sensors, 201 1 , 11 , 2700-

2714.

B. N. Johnson and R. Mutharasan, Langmuir, 2012, 28, 6928-6934.

I. H. El-Sayed, X. Huang and M. A. El-Sayed, Cancer Letters, 2006, 239, 129-

135.

B. J. Kennedy, S. Spaeth, M. Dickey and K. T. Carron, The Journal of Physical Chemistry B, 1999, 103, 3640-3646.

E. M. Timmins, S. A. Howell, B. K. Alsberg, W. C. Noble and R. Goodacre, Journal of Clinical Microbiology, 1998, 36, 367-374.

R. Goodacre, R. Burton, N. Kaderbhai, A. M. Woodward, D. B. Kell and P. J. Rooney, Microbiology, 1998, 144, 1 157-1 170.

J. Clarkson, W. E. Smith, D. N. Batchelder, D. A. Smith and A. M. Coats, Journal of Molecular Structure, 2003, 648, 203-214.

E. Smith and G. Dent, Modern Raman spectroscopy: a practical approach, John Wiley & Sons, 2013.

P. Hildebrandt and M. Stockburger, Journal of Raman Spectroscopy, 1986, 17, 55-58.

A. Matschulat, D. Drescher and J. Kneipp, ACS Nano, 2010, 4, 3259-3269. P. Torres, L. Mercado, I. Cotte, S. P. Hernandez, N. Mina, A. Santana, R. T. Chamberlain, R. Lareau and M. E. Castro, The Journal of Physical Chemistry B, 2004, 108, 8799-8805.

T. Teslova, C. Corredor, R. Livingstone, T. Spataru, R. L. Birke, J. R. Lombardi, M. Canamares and M. Leona, Journal of Raman Spectroscopy, 2007, 38, 802- 818.

Claims

1. A method of detecting a target analyte in a sample using Raman spectroscopy, the method comprising: mixing an analyte binding agent functionalised metal surface with a sample, in order to allow any analyte which may be present in the sample to specifically bind to the binding agent, such that upon binding the analyte is positioned so as to be within 0 nm - 10 nm of the surface, such that an enhancement in Raman scattering of the analyte can be obtained; and directly detecting any bound analyte by way of Raman spectroscopy.

2. The method according to claim 1 wherein the metal surface is the surface of a metal nanoparticle or nanostructured metal, such as noble metal (e.g. silver or gold) surface.

3. The method according to either of claims 1 or 2 wherein the analyte is a nucleic acid, protein, carbohydrate, aldehyde, thiol, amine, explosive, drug of abuse, therapeutic agent, metabolite, environmental pollutant and the like.

4. The method according to claim 3 wherein the analyte is an explosive, such as 2-methyl-1 ,3,5-trinitrobenzene (TNT), Research Department eXplosive (RDX) and/or [3-Nitrooxy-2,2-bis(nitrooxymethyl)propyl]nitrate (PETN)..

5. The method according to claim 3 wherein the analyte is a dioxane, such as 2- Aminomethyl-1 ,4-benzodioxane

6. The method according to any preceding claim for use in detecting one or more analytes.

7. The method according to any preceding claim wherein the analyte to be detected is smaller than1000 g/mol, such as less than 750, 500, 400 or even 300 g/mol, but greater than 5, 10 , 20 or 50g/mol.

8. The method according to any preceding claim wherein the analyte can be detected in concentrations of less than 1 mM, such as less than 100, 75, 50, 25, 20, 10 or even 1 nM. 9. The method according to any preceding claim wherein the binding agent is a member of a pair of specific binding molecules, e.g., an antibody-hapten pair or a receptor-ligand pair.

The method according to any preceding claim wherein the binding agent is an antibody or antibody fragment, integrin, adhesin, nucleic acid or peptide aptamer, cell surface marker, T cell receptor, MHC protein, and the like.

1 1 . The method according to any preceding claim wherein the binding agent is attached, conjugated, bonded or otherwise adhered to the surface

12. The method according to claim 1 1 wherein the analyte binding agent is bonded to the surface, such as by way of ionic, covalent, hydrogen bonding, van der Waals forces, or mechanical bonding. 13. The method according to either of claims 1 1 or 12 wherein the analyte binding agent comprises a reactive group or groups specifically designed to adhere or react and form a bond with the surface

14. The method according to claim 13 wherein the group or groups is naturally present on the binding agent, or the binding agent may be modified so as to include such a group or groups.

15. The method according to claim 14 wherein the group or groups includes a thiol, primary or secondary amines, pyridyl, imidazolium, thiophene, selenophene, isothiocyanate, multi-sulphur organic moiety (that is a molecule having two or more sulfur atoms), multi-heterosulphur organic moiety (that is a molecule having two or more heterocyclic rings each incorporating sulfur atoms), benzotriazole group, and/or combinations thereof.

16. The method according to claim 14 wherein the analyte binding agent is modified to comprise one or more reactive group(s), such as an isothiocyanate group(s) for reaction with the surface.

The method according to claims 15 or 16 wherein said one or more reactive/isothiocyanate group(s) is provided by a moiety which specifically reacts or interacts with a group or groups present on the binding agent and in such a way as to substantially not affect binding of any analyte to the analyte binding agent.

The method according to claim 17 wherein the moiety which comprises a reactive/ isothiocyanate group is a multicyclic moiety which in addition to the reactive isothiocyanate group comprises a carboxylic acid group capable of reacting with a free amine moiety on the antibody, such as fluorescein isothiocyanate.

The method according to any of claims 1 1 - 18 wherein upon reaction/interaction with the analyte binding agent and subsequent binding of the analyte binding agent to the surface, the portion of the binding agent which is capable of binding to the analyte is suitably positioned and optionally oriented in relation to the surface of the nanoparticle, such that surface enhancement of the analyte Raman signal is effected, once the analyte is bound to the analyte binding agent.

The method according to claim 19 wherein the analyte binding agent orientated to be substantially parallel to the surface.

21. The method according to claims 1 1 -20 wherein the analyte when bound to the binding agent is 0 - 5 nm from the surface of the nanoparticle.

22. A SERS enhancing metal surface (such as the surface of a nanoparticle),functionalised with at least one analyte specific binding agent, wherein the analyte binding agent is bound to the surface, such as the surface of the nanoparticle such than when said analyte binds the analyte binding agent, the analyte is orientated to be in close proximity to the surface, such as the surface of the nanoparticle,

wherein the binding agent is an antibody or antibody fragment and wherein at least one free side chain amine group on the antibody or antibody fragment is modified to comprise one or more reactive group(s), such as an isothiocyanate group(s), for reaction with the surface, such as the surface of the nanoparticle.

The SERS enhancing metal surface according to claim 22 wherein the reactive/isothiocyanate group is a multicyclic moiety which in addition to the reactive isothiocyanate group comprises a carboxylic acid group capable of reacting with a free amine moiety on the antibody, such as fluorescein isothiocyanate.

The SERS enhancing metal surface according to claims 22 - 23 wherein analyte binding agent is orientated to be substantially parallel to the surface.

At least two different SERS enhancing metal surfaces according to any of claims 22 - 24 for the detection of different analytes. 26. The SERS enhancing metal surface according to any of claims 22 - 24 which is functionalised with more than one analyte specific binding agent in order to be able to bind more than one analyte if present in a sample.