Introduction

Epigenetic regulation is an essential component of the transcriptional expression machinery [1]. Within the different epigenetic mechanisms, DNA methylation and histone modifications have been shown to be essential during plant development as well as under stressful conditions [2,3,4]. For instance, DNA methylation is fundamental for transcriptional repression of transposable elements (TEs) and regulation of gene expression [5, 6]. In plants, DNA methylation occurs at cytosine residue to produce 5-methylcytosine and it is found in CG, CHG, and CHH sequence context (where H represents A, T, or C). While methylation at CHG and CHH sites is predominantly found in TEs, CG methylation occurs in TEs and protein-encoding genes [7, 8].

To identify differentially methylated regions, bisulfite sequencing, methyl-sensitive cut counting (MSCC) using endonuclease digestion followed by sequencing, and enzymatic methyl-seq (EM-seq) have been the traditional strategies. Bisulfite sequencing is the most popular and widely used method in DNA methylation studies [9]. In this method, bisulfite-treated DNA samples undergo deamination of cytosine into uracil. During PCR-amplification and subsequent DNA sequencing analysis, uracil-converted residues will be read as thymine. However, methylated cytosine is resistant to uracil conversion. Thus, it remains read as cytosine. Comparative analysis of treated and untreated DNA samples enables the identification of methylated cytosines. MSCC uses methylation-sensitive endonucleases for initial digestion of unmethylated DNA sites followed by adaptor ligation that contains the site for another digestion enzyme that cuts outside of its recognized site. This strategy allows the generation of small fragments that can be sequenced. Interestingly, this strategy can be implemented using enzymes that cut either methylated or unmethylated DNA. There are a good number of methylation-dependent endonucleases that can use methylated DNA as a substrate. Examples of those include BisI, BlsI, GlaI, GluI, KroI, McrBC, MspI, MteI, PcsI, and PkrI. In plants, McrBC has been widely used in model and crop plant species [2, 10,11,12]. McrBC is an endonuclease able to cleave DNA containing methylcytosine between two half-sites of (G/A) mC that are located within 50–3000 bp from each other [13, 14]. Unlike the previously described techniques, EM-seq selectively uses a first enzymatic step to oxidize methylated cytosines and a second enzymatic step to convert unmethylated cytosines to uracils. After a subsequent PCR amplification step, oxidized methylcytosine forms base pairs with guanine, and uracil forms base pairs with adenine. EM-seq enzymatic reactions are non-destructive. Therefore, a higher yield and accuracy are obtained.

Even though these methods have been the gold standard for epigenetic studies, optical techniques, collectively known as plasmonics, have recently captured the interest of researchers in the epigenetic field. The use of plasmonics has been well implemented in chemistry, physics, and medicine, among other STEM fields. However, the area where these interactions have flourished in the last decade is on biological systems where both bare or functionalized metallic nanoprobes allow detecting and quantifying, targeting specific intracellular processes or as colorimetric biosensors for monitoring or detection objectives [15,16,17,18]. Plasmonics takes advantage of both light scattering and absorption processes and even can use the naked eye as a detector. When a process occurring close to nanoparticles in colloidal suspension causes its aggregation, a change in color and hence a shift of the absorption spectra and scattering profile is observed and can be measured. The heart of the application of all these techniques is the metallic nanostructured surface, either in the form of colloidal suspension, as a solid substrate for SERS spectroscopy, or in colloidal form for dark-field and colorimetric applications. Among all possibilities of bioanalysis, those related to chemical DNA modifications have recently captivated the interest in plasmonics with the additional advantage that its application in this field does not involve nucleotide-based amplifications as in other classical sensing procedures [19,20,21]. This is reinforced by the fact that DNA bases have high affinity towards metallic nanoparticles, mainly gold and silver nanoparticles [22]. Here, we present a comprehensive review of the strategies that exploit the possibilities of techniques based on plasmonics for studying processes associated with chemical modifications of DNA bases or epigenetics.

Fundamentals of Plasmonics

A specific condition is reached when light is confined either in between or at the proximities of a nanostructured metal surface where electromagnetic oscillations or surface plasmon polaritons permanently occur. When these oscillating electrons compelled to the geometry of a nanostructure are irradiated with an excitation wavelength bigger than the nanoparticle size, a resonant interaction starts, causing a localized oscillation with frequency known as localized surface plasmon resonance (LSPR). As a result, unique optical properties arise, being one of the most important the enhancement of the electromagnetic field. As these conditions allow manipulating light at levels below the diffraction limit, the intensity of signals coming from species adsorbed on the metallic surface is drastically intensified. Plasmonics embrace this junction between surface plasmons and light and it is the core in the development of analytical techniques in nano-optics [23].

One of the most interesting plasmonics techniques that have become well known and that take advantage of those optical properties is surface-enhanced Raman spectroscopy (SERS). SERS relies on the strong signal amplification effect of scattered light obtained when a molecule is close, or better, in the gap between two metallic nanoparticles, a region known as hotspots. The irradiation of a chemical species in this condition results in enhancement factors as high as 105–106 times, allowing the acquisition of spectra with a very high signal-to-noise (S/N) ratio. This is perhaps the main reason that has driven the technique from electrochemical and physical chemistry studies to an increasing amount of analytical and quantitative applications, many of them related to biological systems, even though reproducibility of the spectra has been continuously on the spot as its main disadvantage [24, 25].

Scattered light by a metallic nanoparticle acting as a nanoprobe can also be detected through an optical camera and this possibility allows the study of processes happening at nanoparticle or single molecule level in real time. The optical properties that depend on the nanoparticle size, morphology, composition, and refractive index of the surroundings have been the basics for the relatively recent development of the dark-field spectroscopy technique that offers both spectrum and image as output. Using these properties, it is possible to obtain the scattering intensity and its location at the same time [26].

Contrary to the relatively recent development of these techniques, synthesis procedures of noble metal nanostructures have been widely studied, and a myriad of modifications for tuning their physical properties are available in the literature [27]. Nevertheless, solid nanostructured substrates can be prepared by simply dropping a colloidal suspension, or by generating a roughened electrode. They have also gained benefits from modern techniques such as sputtering or electron beam lithography, to mention a few. Syntheses of colloidal nanoparticles are available in all sorts of different procedures, where most of them require simple glassware and involve the use of a reduction agent for taking the metal to its zero-valent form. In the search for a highly reproducible and reliable colloid, many researchers have focused on understanding the mechanisms behind the formation of nanoparticles with a given size and morphology, and various experimental conditions have been studied and evaluated, from modifications in temperature, stirring speed or reaction time, to several reduction agents and stabilizers [24]. In the face of all available morphologies, perhaps the most popular are colloidal nanospheres, since their synthesis is relatively well-established, and reproducible enough, and plenty of applications have been already reported [28, 29].

Plasmonic Techniques

Colorimetric Sensors

Nanoparticles-based colorimetric sensing is perhaps the simpler approach of plasmonic techniques. This methodology is based on the same principle of conventional colorimetric analysis where a given procedure induces the appearance of color (or changes in it), and the detection step is performed by UV–Vis spectrophotometry or even by the naked eye [30, 31]. When introduced to colorimetric detection, the unique plasmon resonance properties of noble metal nanoparticles enable ultrasensitive detection and its absorbance (and hence, their color) can be tuned by modifying their morphology, size, metal, and chemical environment. The extraordinary optical properties come from the free electron oscillation on the metallic surface when stimulated by light. Specifically, the main phenomenon involved is the localized surface plasmon resonance (LSPR), characterized by its high confinement at the nanostructures and enhanced electromagnetic field. The absorption profile shows a peak in the visible and near-infrared regions [32, 33].

Two main experimental strategies can be adopted to perform colorimetric assays, both based on exploiting the intrinsic properties of nanoparticles by modifying interparticle distance and morphology, and size. Changes in the interparticle distance are related to nanoparticle aggregation that induces surface plasmon coupling between particles, accompanying the LSPR shift and color transition at the solution. For instance, the red-to-blue transitions in gold nanoparticles, AuNPs upon aggregation. This strategy is practical for sensing several targets, from small molecules to macromolecules or living cells. More specific modulations are driven by covalent and non-covalent forces such as hydrogen bonding or electrostatic interactions, dependent on pH and temperature. The electrostatic interactions have a direct influence on the nanoparticles assembly aggregation and its redispersion can be tuned by external changes in the surroundings, among them pH alterations, electrical field, temperature, mechanical stress, or presence of specific species (DNA, proteins, for example), and the result will be an LSPR shift, the key step where detections are realized. However, the main disadvantage of relying on electrostatic interactions is a diminished sensitivity, due in part to the low strength of these forces, although several procedures (including polymerization, physicochemical reactions, and catalysis) can be done to improve it [34,35,36].

Another strategy that can be successfully explored lies in the incorporation of functional groups on the nanoparticles surface through a stronger interaction named covalent bonding. This improves sensitivity and facilitates surface functionalization. Hence, more specific analyses can be performed. The hydrophobic character on NPs surfaces or in its surroundings have also an important role in building sensing strategies, even with suitable specificity [37]. Assemblies with highly specific recognition can be obtained by exploiting multiple intermolecular interactions and can be biologically specific. For instance, using nucleobases functionalized NPs to recognize a target species, although the sensitivity of these approaches directly depends on the aggregate size. Another interesting way of reaching specific bindings is based on the concept of host–guest molecules that guarantees, in the first place, high selectivity. It is worth mentioning that non-covalent interactions such as pi-stacking (in aromatic systems) or Van Der Waals forces are useful in plasmonic sensing methodologies. The predominant interaction exploited in its assembly will largely depend on the physical–chemical nature of the whole system [38, 39].

Tuning morphology and size of metal NPs is also an interesting strategy that induces changes in their optical properties, mainly their LSPR, and thus performs recognition. There are numerous factors associated with the NPs synthesis that controls its morphology and size, among them, precursors, capping, and reducing agents. Usually, small changes in it can give origin to quite different types of NPs and thereby shifts in their LSPR are observed. Thus, the species of interest can be used for example as the reducing agent and is the growth process itself that permits its quantification. Another approach uses the molecule target to carry out a controlled etching procedure that starts with an oxidation step of the NPs, until inducing detectable modifications in its composition of morphology. Then, the extent of the etching will be directly related to the number of species of interest leading to their quantification. When the target is used as the reducing agent, its quantification is performed through the controlled growth of NPs, reflected in the LSPR shift [40]. This methodology has advantages; likewise, its short time of analysis makes it suitable for rapid and simple applications. Various applications include the use of enzymes as reducing agents and hence control NPs growth for several applications [41]. Changes in morphology take advantage of the possibility of reshaping NPs in the presence of a given species (the target). This chemical etching process can be mediated by a variety of compounds, making possible the development of sensing procedures of targets such as hydrogen peroxide, heavy metals, and anionic species. Although currently, NPs for plasmonic colorimetric essays are synthesized from mainly noble metals gold and silver, copper NPs are also used, despite their application could be challenging due to the natural reactivity of this metal. The properties of plasmonic surfaces can also be modified by using different coating materials (silica, polymers) and thickness, or even NPs of metal oxides or a variety of composites.

Surface Plasmon Resonance-Based Techniques

Surface plasmon resonance (SPR) techniques rely on the anomalous diffraction obtained upon the excitation of surface plasmon waves, a phenomenon observed for the first time by Wood in 1902 and that boosted the development of various sensing approaches, from microscopy-based techniques to SPR imaging [42, 43].

SPR is defined as the charge-density oscillations that occur at the interface between two media (a metal and a dielectric) having dielectric constants of opposite signs. When the electric field vector from the incident light at a given angle, a part of it matches the free electron movement in the metal (constrained by boundary conditions coming from the material nature), excitation of surface plasma waves occurs, and because of this efficient coupling, a guided electromagnetic wave along the interface is generated. This is called plasmon and it propagates parallel to the metal surface. A typical SPR biosensor configuration includes a high-reflective glass prism as shown in Fig. 1, and it is known as Kretschmann geometry [44, 45].

Fig. 1
figure 1

Kretschmann geometry. A A typical surface plasmon resonance (SPR) setup includes a light source, a metal layer put at the base of the prism separating it from the dielectric medium containing the analytes or species of interest. As a detector, a charged-coupled device camera is usually used. B Functional coating with analyte-biorecognition species on the SPR sensor surface. C Reflected light spectra obtained upon changes in refractive index. D Changes in refractive index due to molecular interactions in the dielectric medium. Before (I) and after (II) binding

Full size image

As the angle at which resonance is generated depends on the refractive index of the material close to the metal surface, when a small change in the reflective index of the sensing medium, there is no plasmon formation. Thus, the changes in the reflected light are measured at the detector, and resonance angle shifts or intensity can also be monitored and associated with quantitative information. In practice, an SPR biosensor works by first immobilizing a probe molecule on the sensor surface, and in a second stage, the analyte is put in contact with the modified surface where affinity interactions should occur. After the probe-analyte binding, there is an increase of the refractive index at the SPR surface, and the resulting signal change is measured in resonance or response units (RU). Each RU unit is equal to a critical angle shift of 10−4 degrees and the first RU value for measurement is the starting critical angle.

The analyte concentration for a layer of thickness h is related to the change in the refractive index delta (n) through Eq. 1 below:

$$\Delta {n}_{d}={(\frac{dn}{dc})}_{vol} \frac{\Delta \Gamma }{h}$$
(1)

where \({(\frac{dn}{dc})}_{vol}\) is the increase of refractive index, n; vol is the volume of the analyte of concentration c. In real time, when the incident light becomes coupled to the propagating surface plasmon, it allows the detection and quantification of the target molecules. The limit of detection (LOD) will depend on the binding affinity of the analyte to the probe molecule, molecular weight, and the coverage surface area of the immobilized molecule. Other particularly useful information as specificity and kinetic parameters can be determined by using SPR approaches. The bioanalytical field has taken great advantage of the information that SPR techniques offer and numerous applications on biological systems including proteins, enzymes, DNA, and virus have been widely studied.

Surface-Enhanced Raman Spectroscopy

The surface-enhanced Raman spectroscopy (SERS) technique is based on the surface-enhanced Raman scattering effect obtained when an incident beam of monochromatic light interacts with a given molecule attached to a noble metal nanostructured surface, mainly fabricated with gold, silver, or copper. The inelastic light scattered by the molecule can be intensified to a large extent (factors up to 108 or higher) being able to reach very low LOD values (up to femtomolar scale) [46]. This effect was accidentally observed when a pyridine molecule was adsorbed on gold electrodes, initially by Fleishmann and co-workers [47] and later by Albrecht and Creighton [48] which associated this phenomenon with plasmon excitation. Through approximately 40 years of SERS research, a remarkably high number of both applied and fundamental studies have been carried out [49,50,51,52] and although quantitative applications were at some point, difficult to perform, today SERS spectroscopy is for sure an exceptionally good option for quantitative purposes [53, 54].

A classical SERS experiment involves in the first step, synthesizing a well-ordered (ideally) nanostructured substrate either as a solid surface or as a colloidal suspension. Although the effect can be obtained on copper, silver, or gold surfaces, while silver offers higher intensification factors, gold nanostructures allow more reproducible signals and are more suitable for biological applications due to their low ability to cause damage to these systems. In the following stage, a procedure for an efficient binding molecule-nanoparticle is realized. For solid surfaces, it is usual to deposit the analyte of interest from a solution either manually or using an automated device.

When nanoparticles are in the form of colloidal suspensions, various possibilities are in the landscape: they could be used as probes itself inside complex systems as cells or tissues (for imaging, for example), or they can be functionalized with a Raman marker to probe a specific bond or to track its route or modifications. The simpler way for using colloidal nanostructures includes a main attaching step where the molecule is adsorbed onto the surface. This procedure can be modified or customized and this is usually done by functionalizing the surface according to the chemical nature of the species of interest and depending on the objective of analysis.

The phenomena behind the high signal enhancement factors that can be obtained with this technique start at nanometer-sized gaps also known as hotspots (typically in the 2–10 nm size range) where the electromagnetic field is intensified. They can be formed between two metallic nanoparticles, or within their junctions and a flat metal surface. Geometrical properties and the gap distance between other characteristics play an important role in the intensification effect. Two main mechanisms are considered responsible for the signal intensification; the first one relies on an enhanced electromagnetic field arising upon plasmon excitation. The second is related to the contributions of Raman scattering not associated with the electromagnetic environment but related to the electrons being transferred between the molecule and the surface or substrate. It is important to remark that in practice, these two effects cannot be rigorously separated [55].

Applications

Colorimetric Sensors

The relative simplicity of measurements using plasmonic colorimetric sensors has allowed the development of several methodologies for assessing epigenetic biomarkers. Studies on DNA methylation modification have shown a notoriously increasing trend over the last ~ 5 years [56]. Among the available metallic nanostructure surfaces, AuNPs have been perhaps the most popular, due to their unique sensitivity associated with changes in size and distance in NPs, high stability, and due to be more chemically inert, although their extinction coefficient is lower than the AgNPs’ [57].

An unamplified genomic DNA (gDNA) nanosensor was developed by Baetsen-Young et al. [58] using dextrin-capped AuNPs. The study was focused on the detection of DNA sequence from Pseudoperonospora cubensis, a primary threat for cucumber production. Another instance is the one-step procedure (dextrin and sodium carbonate were used as capping and reducing agents, respectively) that generates 13 nm glycol-AuNPs [59]. Using gDNA target isolated from Pseudoperonospora cubensis, and non-target DNA, from cucumber leaves, the authors successfully achieved a highly specific methodology, and even differences were visually distinguished. Regarding sensitivity in the detection of extracted and crude (non-extracted) DNA, values in the order of 3 fM for the former and 18 sporangia uL−1 for the latter were reported. Overall, this assay is not expensive, it is rapid (~ 30 min visual assay), does not require specific equipment, and it is becoming a suitable option to currently available diagnostic methods such as ELISA and PCR.

Although so far, studies focused on plant cells are still starting to be developed, an important number of applications in human cells showed how these strategies are very promising for identifying epigenetic modifications in eukaryotic genomes in general [60]. In a recent application, the formaldehyde obtained from HDMs-catalyzed demethylation was used as a probe [61]. This approach explores the fact that the formaldehyde can quantitatively inhibit the oxidation reaction of o-phenylenediamine Ag+-triggered, causing a decrease in fluorescence and absorbance and allowing a dual-mode reading. The associated mechanism behind both signal modes is shown in Fig. 2.

Fig. 2
figure 2

A Demethylation reaction catalyzed by HDMS with the release of formaldehyde. B In absence of formaldehyde, the o-phenylenediamine oxidation reaction (Ag + -triggered) is not inhibited allowing measuring fluorescence. C The formaldehyde inhibits the o-phenylenediamine oxidation and absorbance signals can be obtained

Full size image

SERS

Being one of the most important epigenetic mechanisms, DNA methylation has been extensively studied through techniques such as HPLC–UV or PCR [62]. Large amounts of DNA samples are required for the former and a careful design and selection of primers is mandatory for the latter, not to mention the time-consuming procedures of sample preparation, as a common disadvantage. In these aspects, SERS is showing to be an attractive option, especially because DNA bases have a particularly high affinity towards noble metal nanostructured surfaces, evidenced through their rich-information spectra (Fig. 3) [22, 63]. In addition, SERS is a label-free technique, allowing reaching limit of detection values as low as 10−7 mol L−1 and sample conditioning is typically easy to perform.

Fig. 3
figure 3

A A typical architecture of the DNA nucleobases adenine and thymine attached to gold and silver nanoparticles, along with their macroscopic appearance in colloidal form. B Their SERS spectra with the main Raman frequencies: bands from in-plane symmetric ring breathing modes (665–802 cm−1), -C = O stretching for guanine and thymine (1640–1657 cm−1)

Full size image

A procedure for the detection of DNA methylation combining SERS and chemometrics has been already implemented [64] where a combined Au@Ag core–shell nanopillar structure was used as a plasmonic substrate. Samples of guanine adducts were analyzed in concentrations between 1 nM and 10 µM. Chemometric modeling showed that the methodology allows clearly distinguishing non-methylated (guanine) and methylated guanine bases, at different positions (O6-methylguanine, N7-methylguanine, O6-hydroxyethyl guanine, N7-2hydroxyethyl guanine, O6-benzyl guanine, and O6-methyl-d3- guanine).

Using a plasmonic gold nanohole array and SERS, the methylation of cytosine was studied [65]. As a control system, a 9-base CpG sequence was used. Three main spectral peaks responded to methylation: a lowering in the intensity of ring breathing (785 cm−1), and ring stretching (1225 cm−1) modes of cytosine and increasing of -CH3 group vibration band (1365 cm−1). In the work by Luo et al. [66], a solid gold SERS substrate was used to evaluate the activity of M.SssI (CpG Methyltransferase). Specifically, the sequence 5′-CCGG-3′ contained in a gene fragment was chosen as the target DNA. A probing DNA, carrying both a recognition unit (a target-DNA-recognized sequence) and an extending spacer for triggering the hybridization chain reaction, was included. The probe, containing an assembled -SH at the 3′-end position, was immobilized through an Au–S bond. Moreover, its 5′-end position was labeled with Cy5, chosen as a Raman reporter that is expected to be absent when M.SssI is missing in the sample, becoming an indicator of methylation. Still, its signal intensity increases with M.SssI activity, allowing its quantitative determination (LOD was 2 10−4 U mL−1), a sensitivity comparable to fluorescence-based assays. Another advantage of the methodology reported in this work is the selective detection of M.SssI in human serum, from other MTase interferences, making this essay promising for methylation-related diseases and also plant-related research.

Mirajkar et al. [67] showed how SERS spectroscopy could be used to distinguish genomic DNA (gDNA) from two plants (Brassica juncea and Arabidopsis thaliana), identifying the ratio of adenine and deoxyribose as the sites of the molecule that interacts to the surface of metallic NPs. Extraction of gDNA was carried out following the CTAB extraction method, and Ag- and AuNPs were used as enhancer substrates, synthesized by well-known methods (hydroxylamine, borohydride, and citrate reduction [68, 69]), showing the potential of this technique. Interestingly, histone demethylation in plants has not been extensively studied using SERS. A work showed a relatively easy essay where the activity of two histone demethylases (HDM) could be assessed by taking advantage of the fact that both of them, through different routes, release formaldehyde as a by-product, probed using SERS [19]. Initially, a thiolated probe reacts with formaldehyde forming an adduct, and in a second stage, it covalently binds to a previously peptide-stabilized AuNPs as shown in Fig. 4. This procedure has an important potential as the probed by-product is also formed from the oxidative demethylation of proteins, nucleic acids, and other biological molecules.

Fig. 4
figure 4

SERS strategy for HDM assay. A Formaldehyde obtention from a diMe H3K4 peptide substrate catalyzed by flavin-dependent amine oxidase (LDS1). B The reaction of formaldehyde with Purpald, a thiolated reactive probe, forming an adduct. The adduct binds through covalent interactions to the surface of peptide-stabilized AuNPs for SERS spectrum measurement

Full size image

In another application, duplex SERS nanotags (silica-coated dimer and trimer AuNPs with a 4-mercapto-3-nitrobenzoic acid and 4-mercaptobenzoic acid as Raman reporters) for simultaneous monitoring of DNA methylation levels and quantitative assessment of total gDNA methylation have been used [70]. In this approach, gDNA is enzymatically digested into DNA fragments. Biotin-dNTPs were filled into the ends of the DNA fragments with DNA polymerase, followed by incubation with streptavidin magnetic beads and SERS nanotags. The beads capture all available biotinylated DNA and the streptavidin-SERS nanotags (SERS-SA) represent the DNA contained in the sample. Thus, the SERS-SA works as an internal control of the gDNA amount. At the same time, nanotags labeled with methyl binding domain proteins (SERS-MBD) are used for specific binding methyl groups in DNA. With the information coming from the Raman intensity of the spectra, the % methylation could be quantified.

SPR Sensors

The use of SPR biosensors has shown to be especially suitable for studying biomolecular interactions, and applications in DNA sensing have grown notoriously in the last 15–20 years [71]. In brief, an ssDNA probe is attached to the sensor surface and a fast, real-time monitoring of hybridization with the target DNA is possible. Moreover, this methodology allows tuning the oligonucleotide probe for increased specificity. Although commonly used methods such as the well-known polymerase chain reaction (PCR) offers high selectivity, specificity, and great versatility, it requires a DNA amplification procedure, making it time-consuming and expensive. Progress in SPR methodologies has made the technique simple, fast, and reliable. In this type of application, it is worth mentioning the work of Plácido et al. [72] and Grześkowiak et al. [73] which used biotinylated ssDNA captured probes coupled to SPR biosensor chips used for GMO detection or quantification of gene expression. Plácido et al. [72] targeted the transgenic construct of Roundup Ready (RR) in soybean and the lectin (the taxon-specific soybean gene). For the sensing probe, a universal ssDNA molecule conjugated with streptavidin (SA) binds to the SPR chip with a complementary sequence previously immobilized onto it. The system captures the biotinylated probe, and then the analyte could be studied. In this case, genomic DNA (from soybean bagasse, soybean seeds, soybean extract, textured soybean, and soy protein) was extracted and amplified, and after analysis, the SPR biosensor surface was regenerated and can be reused. Although the amplification procedure is perhaps the most time- and reagent-consuming step, the amplicons do not need further steps of purification or denaturation. In the work by Grześkowiak et al. [73], a gold NPs-based biosensor was used in the detection of foreign nucleic acids in genomic DNA without amplification, in transgenic tobacco plants expressing a Streptococcus mutans antigen. The target DNA was detected following a genomic DNA isolation procedure from roots, stems, and leaves of tobacco transgenic plants. The sensing strategy included the functionalization of AuNPs with streptavidin and a biotinylated DNA probe [74], and a streptavidin-coated SPR sensor where a sixteen-nucleotide-long biotinylated DNA probe was immobilized (Fig. 5).

Fig. 5
figure 5

A General detection procedure of a transgenic plant using AuNPs-based SPR biosensor. B Schematic representation of the transgenic DNA detection using SA SPR sensor and streptavidin-functionalized AuNPs

Full size image

Measurements here were performed using the coated SPR chips, where positive control and DNA samples had sequences complementary to the biotinylated DNA probes, and functionalized AuNPs were injected as a strategy for signal enhancing, achieving a significant increase in sensitivity. In addition, this methodology is shown to be rapid and does not require a DNA amplification step as in conventional diagnostic methods or in similar SPR applications [75]. In the work by Nguyen and Sim [45], a simultaneous detection of both hot-spot mutation and epigenetic changes was carried out in circulating tumor DNA (ctDNA), a promising biomarker for assessing cancer in a noninvasive way. To accomplish this, a biosensing platform or nanoplasmonic biosensor AuNPs-based was built, and immunogold colloids were prepared. The LSPR shifts were analyzed with a dark-field spectrometer.

The immunogold colloids (prepared by treating the gold surface with (1) thiol-oligo ethylene glycol and (2) 5-methylcytosine (5-mC) monoclonal antibody (mAb)), were able to detect methylation at two methylcytosine (mCpG) sites on the ctDNA with increased sensibility when compared to measurements in absence of treatment. When applied to commercial human serum samples, the nanoplasmonic biosensor detected ctDNA and also, this label-free methodology, along with the minimal amount of sample required for analysis, overcomes common issues of PCR or targeted deep sequencing techniques, classic strategies for quantification of tumor-specific mutations.

Imaging

Plasmonic-based imaging is without a doubt a very promising methodology for accessing epigenetic mechanisms. For instance, Wang et al. [76] and Zhao et al. [77, 78] showed the development of an SPR gold chip with a 3D polyrotaxane surface for detecting peptide interactions, then used for profiling histone “modification-reader” pairs in Arabidopsis thaliana, identifying several unique histone “mark-readers” by analyzing nine putative histone reader domains in this species. The comparison and analysis of histone post-transcriptional modifications (PTMs) in this study showed similarities in histone reader conservation between humans and A. thaliana. Still, the identification of plant-specific histone PTMs suggested an epigenetic adaptation or signaling in the studied species during evolution.

Hyperspectral dark-field imaging (HSDFI) has been applied with a quantitative approach, in single human cells (cell line MCF-7, primary glioblastoma multiforme cell line SF767, and cervical cancer cell line HeLa), as described in the work by Wang et al. [79]. In their work, a methodology along with the use of plasmonic nanoprobes was used to identify and locate key cytosine modifications. As plasmonic surfaces, Au- and AgNPs with 30 and 20 nm in size, respectively, were conjugated to secondary antibodies, and incubated with the cell samples. The results showed that it was possible not just to identify and quantify the most abundant form of cytosine modification (5mC), but those that account for < 0.01%, i.e., 5caC e 5fC, even at different stages of the cellular cycle. Additional advantages of applying HSDFI in this kind of sample are the high intensity of signals from the plasmonic nanoprobes and the absence of autofluorescence, a common issue in techniques such as fluorescence microscopy.