A novel ultrasensitive electrochemical genosensor based on PtNPs/EEGO-AGs/AuNPs nanoassembly for rapid detection of HCMV miRNA

Abstract

MicroRNAs (miRNAs) are stable and accessible biomarkers for disease diagnosis. Elevated levels of human cytomegalovirus (HCMV)-encoded miRNA-UL22A-5p (miR-UL22A-5p) are associated with virologic recurrence in transplant patients with HCMV infection and can serve as a prognostic marker. Here, we introduce the first electrochemical genosensing platform for the sensitive and accurate detection of miR-UL22A-5p in serum samples. This platform employs a ratiometric strategy to improve probe specificity within complex serum matrices. A new nanocomposite consisting of platinum nanoparticles, effectively oxidised graphene oxide aerogels, and gold nanoparticles (PtNPs/EEGO-AGs/AuNPs) was electrodeposited onto a glassy carbon electrode (GCE). This design boosts sensor performance: PtNPs provide high electrocatalytic activity for signal amplification, EEGO-AGs offer antifouling properties to reduce interference from biological matrices while serving as a substrate for further modification, and AuNPs facilitate stable thiol-mediated attachment of single-stranded DNA probes. Under optimal conditions, the nanobiosensor demonstrated a linear detection range from 10^− 12 to 10^− 5 M, with a detection limit (LOD) of 6.1 × 10^− 13 M and a quantification limit (LOQ) of 2 × 10^− 12 M in human serum. Reproducibility was acceptable, with a standard deviation of 1.8%. To assess its quantitative ability and detection sensitivity, serum samples spiked with miR-UL22A-5p were analysed simultaneously with the biosensor and qRT-PCR, showing that the biosensor has a lower detection limit. This biosensor holds significant potential for clinical use as a sensitive tool for early HCMV detection in transplant patients. However, further validation through prospective clinical trials is necessary to confirm its effectiveness and impact on patient outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13065-025-01699-5.

Introduction

Human cytomegalovirus (HCMV) infection remains a major public health concern, especially among immunocompromised groups such as organ transplant recipients [1]. While HCMV typically remains asymptomatic in healthy people, it can lead to severe, life-threatening issues in immunocompromised individuals. Rapid and accurate detection of active HCMV infection is essential for proper clinical treatment. However, traditional diagnostic methods like serological tests and quantitative nucleic acid amplification tests (QNATs) have notable drawbacks. Serology is limited by the widespread presence of IgG, which indicates prior infection; a narrow detection window for IgM, which indicates acute infection; and reduced antibody production in immunocompromised patients. Likewise, QNATs are affected by variable viral load thresholds due to differences in DNA extraction, gene targets, and sample types, and are costly and lack a standardised protocol [2]. In recent years, miRNAs have become promising biomarkers for disease diagnosis and prognosis because of their high specificity, stability, and ease of access in body fluids [3]. In particular, HCMV-encoded miR-UL22A-5p has demonstrated potential as a diagnostic and prognostic biomarker, as it is closely linked to CMV infection and can predict virologic recurrence in transplant patients [3]. Conventional miRNA detection methods, such as qRT-PCR, microarrays, northern blotting, and sequencing, are constrained by their complexity, high cost, variable sensitivity, and lengthy assay durations [4]. Electrochemical biosensors have emerged as a quick, sensitive, and affordable option for detecting miRNA, addressing these limitations [5]. Electrochemical biosensors are highly advantageous for point-of-care (POC) use because they can be easily miniaturised. Electronic components such as potentiostats are now available in pen-drive-sized formats, enabling portable, affordable diagnostics. When paired with mass-produced, low-cost electrodes and compatible with smartphones or Iot platforms, these systems facilitate real-time, accessible, and decentralised disease monitoring. Adding nanomaterials enhances their sensitivity and selectivity, enabling the development of highly durable diagnostic tools [6, 7]. Various nanomaterials have been employed in electrochemical nanobiosensors, including gold nanoparticles (AuNPs) [8], graphene oxide (GO) [9, 10], GO/AuNPs [11], graphitic carbon nitride (g-C₃N₄) [12], reduced graphene oxide with AuNPs (rGO/AuNPs) [13], platinum nanoparticles with rGO (PtNPs/rGO) [14], efficiently oxidised graphene oxide aerogels (EEGO-AGs) [15], etc. These nanomaterials, primarily when integrated, can provide synergistic effects that enhance biosensor performance. PtNPs offer high conductivity and electrocatalytic activity [16], while graphene-based materials provide exceptional electrical, mechanical, and thermal properties [17]. Combining graphene and gold nanostructures in nanomaterials enhances biosensor performance by leveraging the strengths of both. Graphene oxide, with its many functional groups and large surface area, serves as an effective platform for further modifications. At the same time, gold nanostructures have a strong affinity for thiol groups, enabling stable attachment of thiolated biomolecules, while their excellent catalytic properties facilitate efficient electron transfer [18, 19]. GO aerogels offer additional advantages, including a high specific surface area, low bulk density, superior electrical conductivity, and a three-dimensional interconnected network, making them ideal scaffolds for nanoparticle integration [20, 21].

Despite these advancements, the practical use of biosensors in real biological settings continues to face significant hurdles, especially biofouling or the buildup of biological substances on the probe (sensor) surface. This unwanted process weakens sensor performance by diminishing the target analyte signal and increasing background noise. As a result, it causes a lower signal-to-noise ratio (SNR), decreased sensitivity and specificity, longer response times, and a shorter operational life. To address this issue, various chemical and physical strategies have been developed and summarised in our group’s review article [7].

In this context, EEGO-AGs exhibit antifouling or low-fouling properties. Their hydrophilic surfaces, enriched with oxygen-containing functional groups (-COOH, -OH, epoxide), promote the formation of a hydrated layer that inhibits the adhesion of proteins, bacteria, and biocontaminants via hydrated resistance. The negative surface charge from intensive oxidation induces electrostatic repulsion against anionic biomolecules (such as many proteins and bacterial membranes). Moreover, the three-dimensional porous architecture of aerogels ensures uniform distribution of antifouling functional groups, provides a physical barrier against biofouling, reduces direct interactions between biological materials and the electrode surface, and enhances electrochemical responses by increasing the effective surface area. Due to the mechanical and chemical stability of oxidised aerogels in aqueous environments, their antifouling properties are maintained over time [22, 23]. Collectively, these characteristics make EEGO-AGs highly suitable for integration into ultrasensitive electrochemical biosensors operating in complex biological matrices.

Another significant challenge in analysing target analytes in biological environments is the matrix effect, which arises from the interference of coexisting substances [24]. To address or minimise this problem, strategies such as the use of internal standards in separation-based methods or the application of ratiometric approaches in other analytical techniques are commonly employed. In ratiometric detection, the signal intensity of the target analyte is normalised against a reference, blank, or a non-overlapping peak. The ratiometric strategy not only reduces or eliminates background interference but also significantly enhances the accuracy, sensitivity, and reproducibility of electrochemical sensing by minimising the impact of intrinsic and systematic errors arising from microenvironmental variations and external factors.

Considering these benefits, we used a ratiometric electrochemical genosensor to detect HCMV-encoded miR-UL22A-5p. So far, few studies have reported biosensors that target HCMV miRNAs. For example, Chang et al. [25] reported an SPR biosensor for the detection of miR-UL112-3p and miR-UL22A-5p, while Lee et al. [26] presented a fluorometric biosensor for miR-UL112-3p and miR-US25-1-5p (analytical performance details in Table 1). To the best of our knowledge, this study introduces the first electrochemical platform for direct detection of Hcmvmir-UL22A-5p, utilising a ratiometric approach to reduce matrix interferences in real serum samples, without the need for preprocessing or enzymatic steps.

Table 1.

Reported biosensing for the detection of HCMV and comparison with the developed probe

Method	Materials	Analyte	LOD	Dynamic range	Refs.
Electrochemical	AuNPs	406-bp HCMV-amplified DNA	5 pM	5-5000 pM	[48]
Electrochemical	Zn-Ag nanoblooms	HHV-5 DNA	97 copies/mL	113–103 and 3 × 105-106 copies/mL	[49]
Piezoelectric	Au surface	DNA	–	–	[50]
Piezoelectric	QCM with thiol/poly L-lysine SAMs	CMV glycoprotein B	1 µg/mL	2.5–5 µg/mL	[51]
Colorimetric	PSS-modified paper, AuNPs decorated with anti-gB antibodies	CMV glycoprotein B	0.03 ng/mL	–	[52]
Electrochemical	SPEs, AuNPs, AgNPs	CMV glycoprotein B	3.2 ng/mL	5–15 ng/mL	[43]
Electrochemical	Pt-PdNPs@SWCNHs with HRP	CMV pp65 antigen	30 pg/mL	0.1 to 80 ng/mL	[53]
Electrochemical	MWCNT-GS-Chit-[BMIM][PF6], AuNPs, Thionine	CMV PP65 antigen	30 fg/ mL	0.12 to 300 pg/mL	[1]
Electrochemiluminescence	SnS2 QDs/Ag NFs	anti-CMV pp65	0.33 fM	1 fM to 100 nM	[54]
Imaging Ellipsometry	Silicon wafer/CMV-3 A	CMV IgG	0.01 IU/mL	0.011–2.725 IU/mL	[55]
SPR	Magnetic nanoparticles (MNPs), poly(A) extension	miR-UL22A-5p and miR-UL112-3p	108 fM, 24 fM	100 fM to 1 nM	[25]
Fluorometric	DGON, fluorescent PNA probe	miR-US25-1-5p and miR-UL112-3p	502.7 pM 68.5 pM	488.3–7812.5 pM, 61.0–488.3 pM	[26]
Electrochemical	EEGO-AGs/ AuNPs/PtNPs/	miRNA	0.61 pM	10− 12 to 10− 5 M	This work

We designed a novel nanocomposite consisting of EEGO-AGs/AuNPs/PtNPs, electrodeposited onto a glassy carbon electrode. This unique architecture integrates the high conductivity and electrocatalytic activity of PtNPs, the antifouling and high surface area properties of EEGO-AGs, and the strong probe immobilisation capacity of AuNPs. A thiolated single-stranded DNA probe (ss-capture DNA) was covalently bound to AuNPs via self-assembly. The hybridisation process between the ss-capture DNA and miR-UL22A-5p was investigated using SWV, enabling the detection of complementary, non-complementary, and mismatched RNA sequences. The ratio of the signal in the presence of the analyte to that in its absence was taken as the analytical signal. Alongside its outstanding sensitivity and selectivity, this nanobiosensor exhibited excellent reproducibility and remarkable performance in the analysis of real serum samples. A comparative evaluation of RT-qPCR and the developed biosensor demonstrated the biosensor’s high-sensitivity detection of miR-UL22A-5p in serum samples. The proposed electrochemical biosensor enables the detection of miR-UL22A-5p in human serum samples, showing significant potential for clinical diagnostics and POC monitoring of HCMV, particularly in transplant recipients, pending further validation.

Experimental

Oligonucleotides and materials

Graphite powder (99%, 200 mesh), sodium nitrate (NaNO₃), potassium ferricyanide (K₃[Fe(CN)₆]), potassium ferrocyanide (K₄[Fe(CN)₆]), hydrogen tetrachloroaurate (III) hydrate (HAuCl₄·3 H₂O), potassium permanganate (KMnO₄), platinum solution (1000 mg L⁻¹ suspension), phosphorus pentoxide (P₂O₅), and ascorbic acid (AA) were obtained from Sigma-Aldrich (Darmstadt, Germany). Hydrogen peroxide solution (H₂O₂, 30%), hydrochloric acid (HCl), acetone (CH₃COCH₃), sulfuric acid (H₂SO₄, 98%), mercaptoethanol (HS(CH₂) ₆OH), dithiothreitol (DTT), and other required chemicals were obtained from Merck Chemicals Co. (NJ, USA). All solvents were of analytical grade and used without further purification.

Double-distilled (DD) water was used throughout this work. Phosphate-buffered saline (PBS) solution was prepared using sodium dihydrogen phosphate (NaH₂PO₄) and disodium hydrogen phosphate (Na₂HPO₄). Furthermore, EEGO aerogel and gold nanoparticles were uniformly dispersed in PBS via sonication. All solutions were prepared using triple-deionised water. For electrochemical characterisation, a ferrocyanide/ferricyanide solution containing 5 mM of [Fe (CN)₆]^3−/4− (pH 6.5) was prepared. As well, ss-capture DNA and other synthetic oligonucleotides used for biosensor design were obtained from Bioneer Corporation (Seoul, South Korea) as lyophilised powders. The sequences are listed in Table S1. A stock solution of each oligonucleotide (100 µM) was prepared using DD water and kept at – 20 °C until use.

Human serum was isolated from healthy human whole blood samples by centrifuging at 3000 rpm for 10 min. The isolated serum was then diluted 1:10 with PBS (pH 7.4). All samples were obtained with ethical approval from the Research Ethics Committee of Iran University of Medical Sciences (IRB number: IR.IUMS.FMD.REC.1402.315) and with informed consent from the participants.

Apparatus

All electrochemical experiments were conducted using a Galvanostat/Potentiostat Autolab PGSTAT 30 equipped with NOVA 1.11 software at laboratory temperature (25 ± 1 °C). A conventional three-electrode system was employed, consisting of an Ag/AgCl electrode (Methrom, the Netherlands) as the reference electrode, a platinum electrode as the counter electrode, and a glassy carbon electrode (GCE, diameter = 2 mm, Azar Electrode Company, Urmia, Iran) as the working electrode. Materials were dispersed via an ultrasonic homogeniser (SonoPlus HD 3200, Bandelin, Germany). pH measurements were performed with a Metrohm digital pH-meter (Herisau, Switzerland). The surface morphology, elemental composition, and crystallographic structure of the nanomaterials and the electrodes were determined using Field Emission Scanning Electron Microscopy (FESEM) with a FEG-SEM MIRA3 TESCAN (Brno, Czech Republic) equipped with Energy-Dispersive X-ray Spectroscopy (EDS); X-ray Diffraction (XRD) was facilitated by a D500 Siemens Diffractometer (Germany). The microstructure and topography of the EEGO-AGs/AuNPs nanocomposite were examined using Transmission Electron Microscopy (TEM) with a Carl Zeiss LEO 906 (Oberkochen, Germany), Atomic Force Microscopy (AFM) from Nanosurf AG (Liestal, Switzerland) and a dynamic light scattering (DLS, Malvern, UK). Additionally, Fourier-transform infrared (FTIR) spectra were acquired using a Bruker TENSOR 27 FT-IR spectrometer (Bruker, Ettlingen, Germany). At the same time, the specific surface area of the nanocomposite was determined through the Brunauer–Emmett–Teller (BET)/Barrett-Joyner-Halenda (BJH) method, with nitrogen adsorption-desorption isotherms measured on a Belsorp-mini II analyser (MicrotracBEL, Japan). Furthermore, electrochemical measurements, including CV and SWV, were conducted to assess the electrochemical behaviour and signal response of the modified electrode.

Synthesis of EEGO-AGs/AuNPs nanocomposite

The synthesis of EEGO and EEGO-AGs was performed in accordance with previously published methods. Detailed procedures are provided in Supplementary Section S1 [15]. The EEGO-AGs/AuNPs nanocomposite was prepared by dispersing freshly synthesised EEGO-AGs in 0.1 M PBS (pH 9.18) via ultrasonication to obtain a yellow-brown solution. Then, 5.0 mL of EEGO-AGs dispersed solution (250 mg/L) was added with 0.2 mL of AuCl₄·3 H₂O solution (0.05 mol/L) and sonicated for 15 min to form a uniform suspension. The nanocomposite was subsequently electrodeposited onto the PtNPs/GCE surface using the CV technique.

Genosensor fabrication

Electrode cleaning

The GCE was polished with a gentle lapping pad for 5 min, followed by immersion in a piranha solution comprising 0.1 M H₂SO₄ and 0.1 M HNO₃ for 10 min. Afterwards, it was rinsed with a 1:1 water-to-acetone mixture for 15 min, washed with DI water, and dried at room temperature. Finally, the GCE surface underwent electrochemical activation by potential cycling between − 1.5 and 1.5 V in 0.1 M H₂SO₄ until stable, reproducible cyclic voltammograms were achieved (30 cycles).

Electrodeposition of PtNPs on the GCE

The cleaned GCE was immersed in a Pt precursor solution (1000 mg L-1 suspension). PtNPs were then electrodeposited onto the electrode surface using CV within a potential window from − 0.5 to 1.5 V (vs. Ag/AgCl) at a scan rate of 100 mV s-1 for 20 cycles. Figure S1 shows the consecutive cyclic voltammograms recorded during the electrodeposition of PtNPs on the GCE surface. The resulting electrode, named PtNPs/GCE, was thoroughly rinsed with DI water, dried at ambient conditions, and then used as the substrate for further surface modification.

Electrodeposition of EEGO-AGs/AuNPs on the surface of PtNPs/GCE

The PtNPs/GCE electrode was placed in an electrochemical cell containing a dispersion of the EEGO-AGs and Au ions. The nanocomposite was subsequently electrodeposited onto the PtNPs/GCE surface using CV within the potential range of − 1.5 to 1.5 V (vs. Ag/AgCl) at a scan rate of 100 mV s⁻¹ for 10 cycles. Figure S2 presents the cyclic voltammograms obtained during electrodeposition. After deposition, the modified electrode was rinsed with DI water to remove unbound residues and then dried at ambient temperature. The resulting EEGO-AGs/AuNPs/PtNPs/GCE was subsequently employed for the immobilisation of thiolated ss-capture DNA.

Immobilisation of ss-capture DNA and target miR-UL22A-5p hybridisation

The thiolated ss-capture DNA was immobilised on the modified electrode surface via self-assembly. Before immobilisation, the ss-capture DNA was treated with DTT to protect the thiol groups from oxidation, which is critical for maintaining the integrity of the capture probe. In the presence of oxygen, thiol groups can dimerise, reducing the efficiency of immobilisation. To prevent this, 10 µL of DTT was added to 10 µL of ss-capture DNA (1.0 × 10^− 6 mol L^− 1), followed by vortexing and incubation at 25 °C for 30 min. After activation, the ss-capture DNA solution was dropped onto the surface of the EEGO-AGs/AuNPs/PtNPs/GCE and incubated at 4 °C for 1 hour to allow the thiol groups at the 5’ end of the ss-capture DNA to form dative bonds with the AuNPs on the electrode surface (Au–S binding). The modified electrode was rinsed with DI water to remove any unattached probes and then treated with mercaptohexanol (MCH) at 4 °C for 20 min to block nonspecific binding and passivate residual active sites. Finally, the electrode was washed three times with DI water to remove excess reagents. For the hybridisation step, a tenfold-diluted serum sample spiked with the target miR-UL22A-5p was cast onto the modified electrode surface and incubated at 25 °C for 50 min. During this process, the immobilised ss-capture DNA hybridised with its complementary target sequence, forming stable DNA/miRNA duplexes. A schematic representation of the biosensor fabrication process is provided in Scheme 1.

Scheme 1.

Schematic representation of the sensor fabrication and detection of HCMV-miR-UL22A-5p using the electrochemical method

Electrochemical measurements

After fabricating the ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE electrode, different concentrations of target miR-UL22A-5p prepared in a tenfold-diluted serum sample were incubated on the electrode surface. The hybridisation of ss-capture DNA with target miR-UL22A-5p produced concentration-dependent electrochemical signals, as measured by SWV and CV.

Results and discussions

Characterisation tests

FESEM analysis

The surface morphology of the synthesised materials and modified electrodes was characterised using the FESEM technique (Fig. 1). Specifically, Fig. 1a presents the FESEM image of EEGO-AGs, which reveals a distinctive sheet-like, wrinkled structure characterised by a highly porous morphology. This unique architecture not only increases the effective surface area but also enhances interactions with analytes, thereby improving the biosensor’s sensitivity and overall performance. The porous nature of EEGO-AGs provides numerous active sites for molecular binding and facilitates efficient mass transport, thereby further enhancing the biosensor’s effectiveness [27]. Figure 1b depicts the FESEM image of EEGO-AGs after decoration with AuNPs, which manifest as bright, grain-like structures distributed across the EEGO-AGs sheets, resulting from their high electron density. The uniform distribution of AuNPs on the EEGO-AGs surface significantly enhances the composite’s surface area, creating an effective substrate for immobilising ss-capture DNA and improving the biosensor’s sensitivity [28].

Fig. 1.

FE-SEM images of a EEGO-AGs, b EEGO-AGs/AuNPs, c PtNPs-modified GCE, d EEGO-AGs/AuNPs/PtNPs-modified GCE, and e ss-capture DNA/EEGO-AGs/AuNPs/PtNPs-modified GCE at different magnifications

The electrochemical modification of the GCE surface was evaluated using FESEM, as shown in Fig. 1c, where the successful electrodeposition of PtNPs is evident. The PtNPs are well-dispersed, quasi-spherical, and uniformly distributed, providing high surface coverage and a conducive surface for further modification. The subsequent electrodeposition of EEGO-AGs/AuNPs on the PtNPs-modified electrode resulted in a partial accumulation of nanomaterials, as observed in Fig. 1d. The observed partial aggregation is primarily ascribed to the pronounced Van der Waals forces acting between the nanoparticles.

Upon immobilisation of ss-capture DNA, the surface appearance of the electrode changes, as shown in Fig. 1e. The FESEM image reveals a smoother, less porous surface, indicating successful DNA binding and the formation of a uniform monolayer.

AFM and EDS analysis

The morphological characteristics of the EEGO-AG/AuNPs hybrid were analysed using AFM. The AFM topography map and 3D surface view of the EEGO-AG/AuNPs hybrid are presented in Fig. S3a, b. As observed, the surface displays a uniform distribution of AuNPs across the EEGO-AG matrix, confirming the effective anchoring and dispersion of the nanoparticles. The peak-to-valley roughness, the root mean square (RMS) roughness, and the average roughness are 8.347 nm, 1.079 nm, and 809 pm, respectively. The peak-to-valley roughness value indicates that the surface height variations are relatively low, indicating good smoothness. Also, the RMS roughness value suggests the surface is relatively smooth, confirming the peak-to-valley roughness result. Both results are confirmed by the average roughness value obtained.

Energy-dispersive X-ray spectroscopy (EDS) was employed to confirm the elemental composition of the synthesised EEGO-AGs/Au nanocomposite and EEGO-AGs/AuNPs/PtNPs modified GCE.

For the EEGO-AGs/Au nanocomposite, the EDS spectrum (Fig. S3c; Table S2) confirmed the presence of carbon (C), oxygen (O), and gold (Au) in the nanocomposite, validating its successful synthesis. Quantitative analysis revealed elemental composition of 5.58% C, 31.9% O, and 14.69% Au, along with trace elements likely originating from the PBS solution used during GO preparation. These results clearly demonstrate the successful formation of the EEGO-AGs/AuNPs nanocomposite, highlighting its potential for use in electrochemical biosensing applications.

The EDS analysis of the EEGO-AGs/AuNPs/PtNPs/GCE electrode (Fig. S3d; Table S3) revealed the elemental composition of the electrode surface. The presence of carbon (C) and oxygen (O) in the spectrum confirms the successful incorporation of EEGO-AGs, graphene sheets functionalized with oxygen-containing groups, including hydroxyl, epoxy, and carboxyl. The simultaneous detection of C and O is characteristic of the EEGO-AGs material, confirming its presence on the electrode surface. Furthermore, the EDS spectrum showed the presence of gold (Au) and platinum (Pt), with weight percentages of 2.33% Pt and 3.83% Au, indicating the successful deposition of PtNPs and AuNPs on the electrode surface. Although the atomic percentages of Pt and Au are low compared to carbon, this is expected, as both PtNPs and AuNPs are deposited as thin layers or dispersed nanoparticles on the EEGO-AGs surface. The EDS signal is predominantly derived from the surface layer (~ 1–3 μm depth), and the porous structure of the EEGO-AGs matrix results in a lower apparent contribution of metals relative to carbon. The higher weight% of Au compared to Pt (3.83% vs. 2.33%) could indicate a higher surface density of AuNPs, primarily because PtNPs are positioned in the deeper layer. Also, the results of elemental mapping analysis showed that AuNPs and PtNPs were uniformly distributed on the electrode surface, with no local aggregation or significant heterogeneity observed (Fig. S3f). This uniform distribution confirms the nanocomposite’s homogeneity and the stability of the modified electrode.

X-ray diffraction

The crystallinity and structural properties of the EEGO-AG/AuNPs nanocomposite were thoroughly investigated using the XRD technique, as depicted in Fig. S3d. The XRD analysis of the EEGO-AG/AuNPs nanocomposite exhibited prominent peaks at 2θ values of 38.12° and 44.33°, which correspond to the (111) and (200) crystal planes of cubic crystallites of AuNPs, respectively. Furthermore, a broad peak at 24.85° was observed, associated with reduced GO sheets, indicating a decrease in interlayer spacing due to the removal of oxygen functional groups.

These peaks confirmed the successful synthesis of the EEGO-AG/AuNPs nanocomposite and provided valuable information on its crystallographic structure. In particular, the intense diffraction peak at 38.12° demonstrated the presence of pure crystalline Au, indicating the high crystallinity of the AuNPs with face-centred cubic (fcc) structure [29].

TEM analysis results

The morphological structure of the synthesised EEGO-AG/AuNPs nanocomposite was investigated using TEM. The TEM images reveal a slightly wrinkled, transparent, and few-layered sheet morphology of EEGO (Fig. 2), with small spherical AuNPs uniformly distributed across its surface [30]. A size histogram indicates that the majority of AuNPs decorating the EEGO-AGs surface have diameters ranging from 16 to 19 nm (inset). These findings are consistent with previous reports on phyto- and chemosynthesized rGO-AuNP nanocomposites [31].

Fig. 2.

TEM images of EEGO-AGs/AuNPs at different magnifications (inset: Size distributions of AuNPs on the EEGO-AGs/AuNPs composite)

FTIR analysis

FTIR analysis was employed to characterise the EEGO-AGs/AuNPs nanocomposite, as it is widely used to identify functional groups and study interactions between compounds [32]. As shown in Fig. S4(a), a broad and strong peak at 3440.2 cm^− 1 corresponds to the stretching vibrations of the hydroxyl moiety in carboxyl groups. Additionally, 1723 cm^− 1 and 1245 cm^− 1 peaks are attributed to the stretching vibrations of –C = O (carbonyl) and C–O (epoxy) groups, respectively [33]. These results support the presence of functional groups in the EEGO-AGs/AuNPs nanocomposite.

BET surface area analysis

The N₂ adsorption–desorption isotherms and pore size distribution analysis of EEGO-AGs/AuNPs are shown in Fig. S4. A typical type-IV curve with an H2 hysteresis loop was observed between relative pressures (p/p₀) of 0.4 and 1.0, demonstrating the existence of a mesoporous 3D structure [34]. The specific surface area of EEGO-AGs/AuNPs, calculated using the BET method, was 6.424 m² g^− 1. BJH analysis revealed pore-size distributions centred at approximately 18.94 nm and 39.676 nm, respectively, further confirming the mesoporous nature of the material. The total pore volume is calculated to be 0.063377 cm3/g, suggesting favourable porosity of the materials. These results demonstrate the porous structure of EEGO-AGs/AuNPs, which enhances their adsorption capacity and makes them highly suitable for sensor fabrication.

DLS analysis

Dynamic Light Scattering (DLS) was employed to analyse the size distribution and aggregation behaviour of the EEGO-AGs/AuNPs nanocomposite (Fig. S5). The DLS data revealed three distinct peaks: the first at 3.224 nm (intensity 12.7%, standard deviation 0.75 nm), the second at 612 nm (intensity 55.3%, standard deviation 168 nm), and the third at 5.224 μm (intensity 32%, standard deviation 462 nm). The first peak is most likely related to individual gold nanoparticles (core size ~ 3–5 nm), which is consistent with conventional citrate-reduced or seed-grown gold nanoparticles. The low intensity indicates that this population contributes little to the light scattering. The dominant peak at 612 nm corresponds to EEGO-AGs sheets decorated with gold nanoparticles or small gold clusters. The high intensity of this peak indicates significant light scattering due to larger structures. This peak represents an active nanocomposite in which gold nanoparticles are densely immobilised on the EEGO-AGs substrate, facilitating efficient electron tunnelling and probe attachment. The third peak at 5.224 μm suggests the formation of large aggregates, likely due to the strong aggregation of the EEGO-AGs/AuNPs nanocomposite. This phenomenon may result from π–π interactions between graphene domains, electrostatic bridging, or the effects of drying and redispersion processes.

Electrochemical investigation of the prepared biosensing platform

CV and SWV techniques were used to evaluate the stepwise construction of the developed biosensor [35]. CV measurements were performed at 5.0 mM of [Fe (CN)₆]^{3 –/4–} in the potential window from − 1.0 V to + 1.0 V (rate, 100 mV s⁻¹). Fig. 3a, b show a CV curve with weak redox peaks for the bare GCE. After electrodeposition of PtNPs, the [Fe(CN)₆]^{3 –/4–} peak current increased significantly, demonstrating the excellent electrical conductivity of PtNPs [36]. This increase in current is directly related to the reduction in charge-transfer resistance at the bare GCE electrode and to the increase in surface conductivity provided by PtNPs. The modified EEGO-AGs/AuNPs/PtNPs/GCE shows a lower peak current than PtNPs/GCE due to the resistance of EEGO-AGs, which provides essential antifouling functionality, while AuNPs restore and enhance the signal and enable efficient probe immobilisation [37]. Current recovery can occur through the formation of conductive contact points between AuNPs and underlying PtNPs, as well as through mechanisms that increase the effective electrode surface area. This means that the sensitivity of the fabricated probe is maintained while the low-fouling property is also enhanced. Immobilisation of a capture probe monolayer on the modified electrode led to a decrease in peak current, likely due to electrostatic repulsion between the negatively charged phosphate groups of ss-capture DNA and [Fe(CN)6]3–/4–, which restricts charge transfer to the electrode surface [38]. MCH treatment slightly increased the probe current, and repositioning the probe to a vertical orientation on the AuNPs may increase the surface area available for charge transfer. Subsequent hybridisation with miR-UL22A-5p decreased the redox peak currents due to steric hindrance and increased surface density, which impeded electron transfer [39, 40]. These observations, supported by the SWV responses shown in Fig. 4c, confirm that the sensor has been successfully designed and can be applied to measure miR-UL22A-5p.

Fig. 3.

a CV of bare GCE (I), PtNPs/GCE (II), EEGO-AGs/AuNPs/PtNPs/GCE (III), b CV of EEGO-AGs/AuNPs/PtNPs/GCE (I), ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (II), MCH/ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (III), AmiR-UL22A-5p/MCH/ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (VI), and c SWV of EEGO-AGs/AuNPs/PtNPs/GCE (I), ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (II), MCH/ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (III), and miR-UL22A-5p/MCH/ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE (IV) in 5.0 mM [Fe(CN)₆] ^{3 –/4–} in 0.05 M PBS

Fig. 4.

The time optimization of different parameters. (a) Immobilization time of ss-capture DNA, (b) the incubation time of MCH, and (c) hybridization time of ss-capture DNA with miR-UL22A-5p sequences spiked in diluted serum sample (1.0 × 10^− 6 mol L^-1) in 5.0 mM of [Fe(CN)₆] ^3–/4– containing 0.05 M PBS solution

Optimisation of biosensor performance

Optimisation of the concentration and the electrodeposition cycles of PtNPs and EEGO-AGs/AuNPs

To enhance the performance of the nanobiosensor, the concentration and number of electrodeposition cycles for PtNPs and EEGO-AGs/AuNPs were optimised. All measurements were performed in PBS (pH 6.5) containing five mM [Fe(CN)₆]³⁻/⁴⁻ at ambient temperature. Fig. S6(a) depicts the graph of peak currents of GCE at different PtNPs concentrations of 250, 500, 100, and 2000 mg L^− 1. The peak current increased with PtNPs concentration up to 1000 mg L⁻¹, which was selected as the optimal value for the biosensor. The number of PtNPs electrodeposition cycles on the pre-cleaned GCE was optimised between 10 and 25, with 20 cycles yielding the highest peak current (Fig. S6b); accordingly, the potential was scanned 20 times during fabrication to deposit PtNPs on the GCE surface.

To optimise the concentration of EEGO-AGs/AuNPs, various amounts (25–500 mg L⁻¹) were evaluated by CV. As shown in Fig. S6(c), deposition of EEGO-AGs/AuNPs increased the current due to the enlarged electrode surface, reaching a maximum at 250 mg L⁻¹. At higher concentrations, the current decreased, likely owing to repulsion of the negatively charged [Fe (CN)₆] ³⁻/⁴⁻. Therefore, 250 mg L⁻¹ was chosen as the optimal concentration. Also, the number of cycles for electrodepositing EEGO-AGs/AuNPs on the PtNPs-modified GCE was optimised between 5 and 15, with 5 cycles providing the best electrochemical response, as higher numbers likely led to agglomeration of the nanocomposite.

Time optimisation for ss-capture DNA immobilisation, MCH incubation, and target MiRNA hybridisation

To improve the selectivity and sensitivity of the designed genosensor, the parameters under study, including the incubation time of ss-capture DNA on the sensor surface, the incubation time of MCH, and the hybridisation time of ss-capture DNA with target miR-UL22A-5p, were optimised using SWV.

In the present study, the effect of ss-capture DNA immobilisation time on the biosensor’s performance was systematically evaluated over a range of 30–100 min. As shown in Fig. 4a, increasing the immobilisation time from 30 to 60 min resulted in a pronounced increase in the signal ratio. However, further prolongation of the immobilisation period resulted in a gradual decline in the electrochemical response. Therefore, 60 min is the optimal incubation time for immobilising ss-capture DNA. As previously indicated, MCH was employed to block and deactivate the free sites on the electrode surfaces. Figure 4b presents the optimisation results for the MCH incubation time. For this purpose, 1 mM MCH was added to the capture probe-modified electrode surface and incubated for 10–30 min. As illustrated in Fig. 4b, an incubation time of 20 min was optimal, yielding the highest signal ratio. Moreover, optimising the target hybridisation time is essential for achieving optimal genosensor performance. Therefore, a 10-µL (1.0 × 10^− 6 mol L^− 1) solution of the target miR-UL22A-5p has been dropped onto the surface of the modified electrode. The duration of hybridisation can play a crucial role in the genosensor response [41]. Thus, optimal conditions have been chosen to investigate incubation times for various target miR-UL22A-5p sequences. The effect of hybridisation time on the sensor response to target miR-UL22A-5p (1.0 × 10 − 6 M) was studied, as illustrated in Fig. 4c. The current response ratio increased with hybridisation time, reaching a maximum at 50 min; therefore, this duration was selected for miR-UL22A-5p on ss-capture DNA. These values were adopted in the following study.

Kinetic investigation

The association between scan rate and peak current typically provides valuable information about electrochemical mechanisms. Therefore, the CV technique was utilised to assess the kinetics of the proposed biosensing approach in the presence of [Fe (CN)₆] ^{3 –/4–} at low scan rates from 5.0 0 to 200 mV s^− 1. Figure 5a demonstrates that increasing the scan rate increases both the anodic and cathodic peaks and the peak separation in the voltammograms. This observation indicates that electron transfer at the modified electrode surface increases progressively at lower scan rates, resulting in narrower voltammograms with reduced peak currents. As shown in Fig. 5b, it is evident that the current (Ip) increases proportionally with the square root of the potential scan rate (v¹/²). This linear correlation further indicates that the process is diffusion-controlled [42]. The Randles-Sevcik equation (Eq. 1) is proposed to estimate the effective surface area of a modified electrode as follows.

Fig. 5.

a CV voltammograms of EEGO-AGs/AuNPs/PtNPs modified GCE electrode in the solution containing [Fe(CN)₆] ^{3 –/4–} in various scan rates, b plot of I_p.a. versus ν^0.5, (c) ln (I_p.a.) versus ln (ν), and I_p.a. versus ν of the bare GCE, GCE/PtNPs, and GCE/PtNPs/EEGO-AGs/AuNPs electrode

1

In which, Ip (µA), A (cm²), D (cm² s^− 1), n, C (mol/cm³), and ν (V s^− 1) represent the peak current, electroactive surface area, molecular diffusion coefficient in solution, the number of electron transferred in an electrochemical reaction, of [Fe (CN)₆] ^{3 –/4–} concentration, and the scan rate, respectively. For 5 mM of [Fe (CN)₆] ^{3 –/4–}, D equals 7.6 × 10^− 5 cm² s^− 1.

The results show that the ln Ip and ln ʋ have a very high linear correlation with a coefficient of determination of 0.9957 (Fig. 5c). This strong correlation provides further evidence for the dominance of the diffusion-controlled mechanism in the electrochemical reaction kinetics. In contrast, the Ip vs. ʋ plot exhibits nonlinear behaviour (Fig. 5d), which strongly confirms the diffusion-controlled nature of the process. Overall, mass transfer at the GCE/PtNPs/EEGO-AGs/AuNPs electrode is mainly governed by diffusion rather than adsorption. The effective electroactive surface area was determined from the slope of the Ip vs. ν^1/2 plot. The specific surface area of ​​the bare GCE electrode is 0.06 cm², which is almost equivalent to its geometric surface area (3 mm in diameter) and has limited active sites due to the smooth surface. In contrast, the GCE/PtNPs electrode with a specific surface area of ​​0.72 cm² shows about a 12-fold increase in active surface area due to the nanostructured platinum nanoparticles and high catalytic sites, making it the best choice for high-efficiency electrochemical reactions. The EEGO-AGs/AuNPs/PtNPs/GCE electrode also has a specific surface area of ​​0.67 cm² (11.2 times that of Bare GCE), but 7% less than that of GCE/PtNPs, which may be due to the possible coverage of active sites by additional layers such as graphene oxide or gold nanoparticles. However, this multilayer structure is better suited to selective, stable applications.

Calibration curve and analytical features

The analytical performance of the fabricated genosensor was assessed by incubating varying concentrations of synthetic miR-UL22A-5p spiked in diluted serum (10⁻¹²–10⁻⁵ M) at 25 °C for 50 min. As depicted in Fig. 6a, the SWV peak current exhibited a concentration-dependent decrease, attributable to electrostatic repulsion between the [Fe (CN)₆] ³⁻/⁴⁻ redox couple and the negatively charged ss-capture DNA sequences. Following hybridisation with the target miRNA, the increased negative charge density on the electrode surface amplified this repulsion, resulting in a further decrease in the electrochemical response. The calibration plot, shown in Fig. 6b, demonstrates a linear relationship between the normalised current responses. The normalised current was calculated by dividing the post-hybridisation current (I) by the pre-hybridisation blank current (I₀), which is measured immediately after the blocking step and before incubation with the target sequence. The logarithm of target concentration in serum samples over the range of 10⁻¹² to 10⁻⁵ M. The limit of quantification (LOQ) and limit od detection (LOD) of the present sensor with a linear equation of y = 0.0604 log[miR-UL22A-5p] + 0.7093 and the linearity correlation coefficient of R² = 0.9986 were calculated to be 2 × 10^− 12 M and 6.1 × 10^− 13 M, respectively. Notably, the LOQ and LOD of the developed genosensing probe are determined using the criteria of 10 δ/m and 3 δ/m, respectively. δ and m denote the blank standard deviation and the calibration curve slope. Triplicate measurements demonstrated excellent reproducibility, with an overall mean RSD of 1.8%. Table 1 presents analytical data for previously reported HCMV biosensors. As shown, traditional antibody- or antigen-based sensors (e.g., targeting glycoprotein B or pp65) can detect biomarkers at very low concentrations. Still, their clinical applicability is often limited by factors such as dependence on infection stage, immune variability, and potential cross-reactivity with other herpesviruses [43, 44]. Additionally, prior miRNA-based HCMV biosensors, such as SPR or fluorometric systems, offer enhanced specificity for miRNA targets and demonstrate ultralow detection capabilities. However, these systems require labelled probes, complex instrumentation, or additional amplification steps, all of which limit portability and increase assay complexity [25, 26]. Despite the impressive detection limits achieved by many electrochemical sensors targeting miRNA biomarkers, typically in the picomolar to femtomolar range, their practical utility is often constrained by narrow dynamic ranges, susceptibility to nonspecific adsorption, matrix interferences, and reliance on enzymatic amplification or labelling techniques. In contrast, the genosensor proposed in this work integrates a ratiometric detection approach, a label-free assay format, and a tailored nanoarchitecture of EEGO-AGs/AuNPs/PtNPs. These combined features enhance analytical reliability, minimise matrix interferences in real serum samples, and streamline the detection workflow, setting this platform apart from existing biosensors. The ratiometric detection strategy inherently mitigates experimental fluctuations, nonspecific interferences, and matrix effects, introducing a self-calibration mechanism that enhances assay reliability and robustness. By leveraging the electrochemical signal ratio between the target’s presence and absence, the system ensures accurate, reproducible results in complex biological samples such as serum. The ratiometric detection strategy intrinsically suppresses errors caused by experimental fluctuations, nonspecific interferences, and matrix effects in complex biological samples such as serum. By leveraging the ratio of electrochemical signals obtained in the presence and absence of the target analyte, this approach affords an inherent self-calibration mechanism, thereby bolstering assay reliability and robustness. Furthermore, the antifouling properties of EEGO-AGs, characterised by their hydrophilic, negatively charged, three-dimensional porous structure, reduce biofouling and nonspecific adsorption. The surfactant-free electrodeposition protocol employed to assemble the EEGO-AGs/AuNPs/PtNPs nanoarchitecture ensures superior electroactivity, electrode stability, and inter-assay reproducibility. Operating in a truly label-free manner through direct hybridisation detection, this platform eliminates the need for enzymatic amplification or labelling agents, thereby avoiding potential false-positive or false-negative signals. Consequently, this system demonstrates outstanding selectivity while enabling the ultrasensitive detection of HCMV-encoded miR-UL22A-5p at concentrations well below the detection limits of conventional qRT-PCR. Clinically, this capability facilitates timely diagnosis and prognostic evaluation, particularly for transplant recipients, where rapid diagnosis is essential for early intervention and management. However, while the system demonstrates considerable potential for clinical deployment in the monitoring and management of HCMV infection dynamics, we emphasise that further comprehensive validation through additional clinical studies is necessary to confirm its clinical readiness.

Fig. 6.

a SWV responses of the genosensor for different concentrations of target miR-UL22A-5p in diluted serum sample in the presence of 5.0 mM of [Fe (CN)₆] ^{3 –/4–}solution containing 0.05 M PBS and b calibration curve in serum sample. (x is the log concentration target miR-UL22A-5p (µM) and y is ratio)

Method validation

Reproducibility, accuracy, and intra- and inter-day repeatability of the genosensing bio-assay

Repeatability reflects variation under identical conditions, while reproducibility indicates variation among different operators and systems [45]. Reproducibility was assessed by measuring the target miRNA at 1.0 × 10 − 6 M using five independently fabricated biosensors under optimal hybridisation conditions, and recording the corresponding signal intensities. The relative standard deviation (RSD%) of the five modified electrodes was 8.9%. The relatively high RSD% for repeatability between sensors may be attributed to slight, inherent differences in the electrode deposition process and nanocomposite formation during electrode fabrication.

The precision of proposed approaches is measured by their inter-day and intra-day repeatabilities (Table 2). For intra-day repeatability, three different concentrations of miR-UL22A-5p (10.0, 0.1, and 0.001 µM) were analysed with the modified electrode on the same day, yielding an RSD% of 0.8%. Furthermore, the method showed satisfactory inter-day repeatability, as three miRNA concentrations (1.0, 0.01, 0.00001 µM) were measured on different days, with an RSD% of 2.4%. According to analytical method validation guidelines, such as those from the Food and Drug Administration (FDA), %RSD values below 15% for all concentrations and 20% for the LLOQ are acceptable and indicate high method reproducibility [46]. Also, the probe’s accuracy was assessed by calculating relative errors (%) for intraday and interday tests. The results showed that the developed probe detected the added concentrations, with absolute RE% values of 11.2% and 11.1% for intraday and interday assays, respectively. These findings confirm that the fabricated electrochemical biosensor has excellent reproducibility, accuracy, and acceptable repeatability. The accuracy and precision of this method were fully verified through recovery experiments and by comparing measured concentrations with theoretical values. The results showed acceptable recovery values ​​in the range of 85.6 to 114.7% with a low relative error (about 11%). For a definitive statistical evaluation, a two-tailed t-test was used, which showed no significant difference between the expected and experimental values. This was confirmed by the t-values being less than the critical value and the p-value exceeding 0.05. Additionally, the recovery range falls entirely within the accepted criteria set by authoritative guidelines, such as the FDA, which range from 80 to 120% [46]. Finally, this method demonstrates high statistical accuracy and reliability and is well-suited for analysing real samples.

Table 2.

Accuracy and repeatability of the EEGO-AGs/AuNPs/PtNPs/GCE for CMV detection

Concentration (µM)	Intraday
	Precision	Accuracy
	RSD%	Recovery	(RE%)
10.0	1.06	86.0	-14.0
0.1	0.64	109.72	+ 9.7
0.001	0.96	90.0	-10.0
Interday
1.0	1.1	114.7	+ 14.7
0.01	1.1	85.6	-14.4
0.00001	4.95	104.2	+ 4.2

Selectivity investigation

The limited length of mature miRNA sequences, combined with their substantial sequence homology across family members, renders selective hybridisation between the probe DNA and the target miRNA a critical determinant of assay performance. The selectivity of the proposed biosensor towards different target miR sequences (1.0 × 10^− 6 mol L^− 1), including complementary, single-base, two-base, three-base mismatched, and non-complementary sequences, was investigated using SWV, and their voltammograms are presented in Fig. S7(a). The mismatch sequences are shown in Table S1.

There was a significant decrease in peak intensity after genosensor incubation with the complementary target, indicating effective hybridisation. The analysis revealed that the electrochemical signal generated by the 1-base mismatch was comparable to that of the complementary target in the genosensor assay, yet lower than the responses observed for other mismatch types. This behaviour can be attributed to the close resemblance between the 1-base mismatch sequence and the target miRNA, which likely allows for partial hybridisation with the ss-capture DNA. As a result of the reason mentioned above, the electrochemical response to 2- and 3-base mismatches has been enhanced, and their signals are in proximity to the ss-capture DNA-related signal. Incubation of the non-complementary sequence resulted in no substantial change in the genosensor’s voltammogram, indicating that no hybridisation occurred. These outcomes uncover the selectivity and specificity of the fabricated genosensing bioassay.

The stability tests

The stability of a genosensor is a key parameter for its practical application and commercialisation [47]. To study the long-term stability of the genosensor platform, four ss-capture DNA/EEGO-AGs/AuNPs/PtNPs/GCE electrodes were prepared according to the previously outlined procedure. The electrodes were stored at 4 °C. Then, on days 1, 7, 15, and 30, one of them was tested via SWV to measure the response to the 1.0 × 10^− 6 M target miRNA, and the results were compared with those from the first day. No significant decline in sensor response was observed after 1 month of storage, suggesting that the immobilised probes retained their recognition capability under these conditions (Fig. S7b).

Consecutive CV scans studied the electrochemical stability of the EEGO-AGs/AuNPs/PtNPs-modified GCE at a scan rate of 100 mV s^− 1 for 1, 2, 5, 10, 15, 20, 25, and 30 cycles. As displayed in Fig. S7c, the modified GCE exhibited outstanding stability with almost unchanged peak current and potential after 30 cycles of scanning.

Application

The biosensor’s accuracy was assessed using a recovery assay. Three concentrations of synthetic miR-UL22A-5p were spiked into the healthy serum and then tested by the fabricated genosensor without any preparation. Then, the recovery percentage was calculated using the formula: (calculated concentration/added concentration) × 100. The results are summarised in Table 3. For spiked concentrations of 1.0 pM, 10.0 pM, and 100 pM of miR-UL22A, the recovery percentages were 113.0%, 104.2%, and 85.5%, respectively, confirming minimal matrix interference due to the physical antifouling filtration provided by EEGO-AGs.

Table 3.

Results of performance of the sensing method in real sample

Added concentrations (pM)	Detected concentrations (pM)	RSD%	Recovery (%)
1.0	1.13	3.6	113.0
10.0	10.42	5.3	104.2
100.0	85.5	2.2	85.5

To evaluate the biosensor’s functional capacity, serial dilutions of healthy human serum spiked with synthetic miR-UL22A-5p (initial concentration: 100 µM) were prepared at dilution factors ranging from 1:10 to 1:108, corresponding to final concentrations between 10 µM and 1 pM, and then analysed by the developed biosensor. In parallel, identical dilutions were subjected to total RNA extraction using TRIzol reagent (Yekta Tajhiz Azma, Iran). The concentration and quality of the extracted RNA were assessed using a NanoDrop™ spectrophotometer (Thermo Scientific, Wilmington, MA), with A260/A280 ratios ranging from 1.9 to 2.1 across all samples. Subsequently, cDNA synthesis was performed with the BioTech Rabbit™ cDNA Synthesis Kit (Germany) employing miRNA-specific stem-loop reverse-transcription primers, followed by analysis via SYBR Green-based qRT-PCR. Stem-loop RT and qPCR primers were synthesised and purchased from Metabion Company (sequences listed in Table S3). The detailed stem-loop RT-qPCR protocol is provided in the supplementary file (Section S2). The genosensor successfully detected miR-UL22A-5p down to the 1:10⁸ dilution (1 pM), whereas RT-qPCR reached its detection limit at the 1:10⁴ dilution (10 nM) and reported further dilutions as negative (Fig. S8). Consequently, the LOD of the proposed biosensor is 10⁴ times lower than that of RT-qPCR, highlighting the platform’s superior sensitivity for early detection in complex biological matrices.

Limitations

This study developed a ratiometric electrochemical genosensor based on a novel EEGO-AGs/AuNPs/PtNPs nanocomposite for the detection of HCMV-miR-UL22A-5p. The platform demonstrated excellent sensitivity, specificity, and robustness in a controlled serum matrix spiked with synthetic miRNA, confirming its strong analytical performance in a proof-of-concept setting. However, these results do not fully reflect the biological and clinical complexity encountered in authentic samples from HCMV-infected transplant recipients. Clinical specimens may involve additional factors, such as fluctuations in endogenous miRNA expression, matrix interference associated with immunosuppressive therapy, and co-existing medical conditions that could affect assay performance. These aspects were not addressed in the present study due to limited access to clinical cohorts and the focus on early-stage platform development and analytical validation. To bridge this gap, future studies will involve prospective validation using authentic clinical samples collected from HCMV-positive transplant recipients. Parallel qRT-PCR and ROC curve analyses will be performed to establish diagnostic performance parameters, including sensitivity, specificity, and AUC. Addressing these aspects is essential to confirm the biosensor’s diagnostic reliability and facilitate its translation into POC applications for HCMV monitoring in high-risk populations.

Beyond clinical validation, practical and technical aspects must also be considered to ensure the successful translation of this platform into real-world applications. Although the fabrication procedure is efficient and straightforward at the laboratory scale, the use of high-purity nanomaterials and surface modification steps may pose challenges when transitioning to large-scale production. The current work primarily focused on demonstrating analytical performance, and a comprehensive cost analysis was not conducted. Moreover, reproducibility was evaluated across different fabrication batches on separate days and demonstrated consistent analytical performance, supporting the stability of the electrode modification process. However, future work with larger fabrication series and extended time intervals could provide further confirmation of long-term robustness.

Conclusions and future perspectives

The identification of miR-UL22A-5p as a biomarker for predicting virologic recurrence in transplant patients with CMV emphasises the urgent need for quick, sensitive, and user-friendly detection methods. This research introduces a new electrochemical DNA biosensor based on a ratiometric EEGO-AGs/AuNPs/PtNPs nanocomposite for detecting HCMV-miR-UL22A-5p through hybridisation. The ratiometric approach improves accuracy by reducing matrix effects, supported by the nanocomposite’s synergistic properties, which offer a large active surface and enhanced electron transfer. Additionally, EEGO-AGS’ antifouling properties reduce nonspecific interactions, thereby increasing selectivity in complex serum samples. A thiolated short ss-DNA probe attached to gold nanoparticles ensures specific hybridisation with the target miRNA. Results show the biosensor’s dynamic range from 10 to 12 to 10 − 5 M, with a detection limit of 6.1 × 10–13 M, indicating high sensitivity and selectivity. The complete assay, including target addition, hybridisation (50 min), washing (2 min), and electrochemical readout (5–10 min), takes about an hour. The biosensor demonstrated good stability, reproducibility, and accuracy in serum samples spiked with synthetic HCMV-miR-UL22A-5p. Future validation with clinical samples from HCMV-positive transplant recipients will be essential to evaluate diagnostic metrics, such as sensitivity, specificity, and AUC, using parallel qRT-PCR and ROC analysis. Although this study focused on a single miRNA, the platform is adaptable for multiplexing, enabling the simultaneous detection of multiple miRNAs. Future efforts will involve modifying probe design, electrode patterning, and signal differentiation to support multiplexed detection. Overall, this work lays a solid foundation for creating a rapid, cost-effective, and highly sensitive biosensing platform with promising potential for point-of-care HCMV monitoring in high-risk patients.

To facilitate translating this system into POC applications, its scalable design can be easily adapted for disposable screen-printed electrodes (SPEs) and microfluidic cartridges, supporting the creation of a compact, automated, and user-friendly platform. It can also be integrated with portable potentiostats linked to smartphones for rapid detection within an hour. The antifouling EEGO-AGs interface, combined with a ratiometric detection strategy, ensures consistent, reliable performance in unprocessed serum samples, eliminating the need for complex pretreatment. These features collectively improve the system’s usability and portability, making it a strong candidate for POC applications. Still, further validation and refinement are necessary to confirm its clinical potential fully.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

R.N.S., Methodology, Validation, Investigation, Writing-the first draft, J.S.N; Investigation, Validation, Writing-Editing the final version, P.S.D.; Investigation, S.J.K, K.K. and S.H.M.; Investigation, Resources, A.T.; Methodology, J.S. and F.B.S.; Conceptualization, Methodology, Resources, Writing-Editing the final version.

Funding

This work was supported by the Research Vice-Chancellor of Iran University of Medical Sciences (IUMS), Tehran, Iran, with grant numbers 26655 and 27262.

Data availability

All data generated or analysed during this study are included in this published article and its supplementary information files. The sequence for HCMV miR-UL22A-5p is publicly accessible in miRBase (accession MIMAT0001574; https://www.mirbase.org/hairpin/MI0001678). No new datasets requiring deposition in external repositories were produced. Data are available on request by R.N.S.

Declarations

Ethics approval and consent to participate

The present survey was approved by the ethics committee of the IUMS, Tehran, Iran (ethical code: IR.IUMS.FMD.REC.1402.315), in accordance with the Second Declaration of Helsinki. All participants were informed about the study, and written consent was obtained from all participants before enrolment.

Consent for publication

All authors approved the publication of the current study.

Competing interests

The authors declare no competing interests.