Smartphone-assisted urinary methylmalonic acid sensing using an imidazole-based probe via the analyte-triggered aggregation mechanism

Abstract

Methylmalonic acid (MMA) is a key biomarker for vitamin B₁₂ deficiency, and its sensing via a turn-on-based fluorescence mechanism is less explored. The maximum efficiency reported corresponds to only a few-fold increase in the emission. Similarly, the mechanism by which the emission turns on remains unaccounted for. Here, we report an imidazole-based non-fluorescent probe, M1, which turns on blue fluorescence upon the addition of MMA due to the ‘analyte-triggered aggregation’ (ATA) mechanism. An aggregation of probe molecules is triggered by MMA, which is mediated by hydrogen bonding and electrostatic interactions between the probe molecules. The probe-analyte interactions and aggregation formation were corroborated by ¹H NMR, FTIR, DLS, FESEM, zeta potential measurements, and density functional theory (DFT). The analyte triggered a fluorescence enhancement of the probe of 73-fold (the best reported yet), with an observed detection limit of 5.78 µM. M1 demonstrated excellent selectivity toward MMA over biological interfering elements commonly found in human urine. The scope of M1 in practical applications was validated by detecting MMA in real urine samples, achieving recovery rates of 95–108%. This work presents a detailed sensing mechanism with a cost-effective, non-invasive, and field-deployable strategy (smartphone) for monitoring vitamin B₁₂ deficiency viaMMA detection.

---

Introduction

Biomarker detection is crucial for human health, facilitating early disease diagnosis, monitoring, and treatment. Non-invasive methods are increasingly favored over invasive ones for their safety, patient comfort, and accessibility.¹ Although invasive techniques offer in-depth diagnostic information, they are associated with high cost, procedural risk, and discomfort.² Non-invasive approaches mitigate these drawbacks, particularly for vulnerable groups like children, the elderly, and patients with comorbidities, while providing affordability and ease of use for widespread healthcare screening.³

Methylmalonic acid (MMA) is naturally produced in limited quantities in the human body. Elevated MMA levels in body fluids, such as serum and urine, are indicative of cobalamin (vitamin B₁₂) deficiency.⁴ Vitamin B₁₂ serves as a cofactor for methylmalonyl-CoA mutase, which catalyzes the conversion of L-methylmalonyl-CoA to succinyl-CoA in the mitochondrial Krebs cycle. A deficiency in cobalamin disrupts this process, causing the accumulation of methylmalonyl-CoA, which is subsequently converted to MMA and excreted in body fluids.⁵ Chronic vitamin B₁₂ deficiency can lead to irreversible neurological disorders, adverse pregnancy outcomes, and pernicious anaemia.⁶ Therefore, early detection of methylmalonic acidemia and aciduria is essential for timely intervention. Studies have shown that healthy individuals have MMA concentrations ranging from 73 to 270 nM in serum and 3 to 11 µM in urine.⁷

Conventional methods for MMA detection encompass gas chromatography mass spectrometry (GC-MS),^8,9 liquid chromatography mass spectrometry (LC-MS),^10,11 high-performance liquid chromatography (HPLC)¹² and electrochemical techniques.^13,14 Yet, these approaches suffer from drawbacks such as expensive instrumentation, lengthy procedures, and the need for skilled operators. Consequently, there is a pressing demand for rapid, affordable, and real-time alternatives. Fluorescence detection techniques address these needs exceptionally well, delivering high selectivity, sensitivity, and on-the-spot monitoring, making them ideal for biomedical and clinical applications.

Recently, some studies have explored the fluorescence-based detection of MMA. For instance, Miyaji et al. developed an anthracene-based fluorescent colourimetric sensor that exhibited a green emission turn-on response to MMA in urine, although selectivity against other urine components was not demonstrated.¹⁵ Eu(III)-¹⁶ and Zn(II)⁷-based metal–organic frameworks (MOFs) have also been employed for MMA detection using a turn-off fluorescence mechanism. A Cd(II)-based coordination polymer with an amino ligand showed limited fluorescence enhancement and a red shift (∼10 nm) in response to MMA.¹⁷ Despite these advances, selective turn-on fluorescence sensing of MMA remains largely underexplored (Table 3). Additionally, in previous reports, the reason behind the turn-on fluorescence emission has not been thoroughly explained.

Inspired by prior reports^15,17 highlighting the role of hydrogen bonding between a sensing probe and MMA, we designed and synthesised an organic probe, M1, containing an imidazole ring. The M1 probe molecule is capable of hydrogen bonding and electrostatic interactions with MMA, resulting in MMA-triggered aggregation. In this report, we explain the detailed mechanism of turn-on fluorescence via the ‘Analyte-Triggered Aggregation’ (ATA) mechanism^18–20 using different spectroscopic techniques and computational studies. We have also given a thorough explanation for aggregate formation and the most plausible cause of the turn-on emission. M1 is a synthetically simple, efficient, and selective platform for the on-site, smartphone-based turn-on fluorescence detection method for MMA in human urine.

Experimental section

Synthesis and characterization

Synthesis of M1. The synthesis of M1 was performed in two steps (Scheme 1). In the first step, ortho formylation of 4-hydroxybenzophenone occurs via the Duff reaction on both sides, resulting in 5-benzoyl-2-hydroxyisophthalaldehyde (BHP).²¹ In the second step, formylated benzophenone, benzil, and ammonium acetate were refluxed overnight in acetic acid, which yields pale green heterocyclic imidazole M1 (yield 87%).²² Colourless crystals of M1 suitable for X-ray diffraction were grown via slow evaporation of a hexane-layered dichloromethane (DCM) solution (Fig. S7 SI).

Scheme 1 Synthetic routes and molecular structures of M1 [(3,5-bis(4,5-diphenyl-1H-imidazol-2-yl)-4-hydroxyphenyl) (phenyl)methanone].

M1 : ¹H NMR (400 MHz, Methanol-d₄) δ 8.59 (s, 2H), 7.88–7.85 (m, 2H), 7.71 (s, 4H), 7.67–7.61 (m, 1H), 7.58–7.52 (m, 10H), 7.40–7.30 (m, 8H). ¹³C NMR (101 MHz, Chloroform-d) δ 194.87, 158.38, 144.16, 138.02, 133.01, 132.90, 132.60, 130.16, 129.24, 129.12, 129.04, 128.56.128.35, 128.08, 116.19. HRMS calculated for [M1 + H]: m/z 634.2484 found: [M1 + H]: m/z 634.2492 (Fig. S5 SI).

Results and discussion

Photophysical property studies

UV-Vis spectroscopy was performed in methanol, and the results are presented in Fig. 1a. The spectra reveal two distinct bands: a lower-energy band peaking at around 310 nm (4.00 eV) with a shoulder at 375 nm and another high-energy band centred at around 214 nm (5.79 eV). To investigate the excited-state properties, time-dependent DFT (TD-DFT) calculations were performed. Electronic transitions with notable oscillator strengths (f) were identified (Table 1). The first transition was found at 3.92 eV, but with a vanishingly small oscillator strength (f = 10⁻⁴), while the second transition was located at 4.07 eV (f = 0.6192), which is in good agreement with the experimental results (4.00 eV). The NTOs for the second transition are shown in Fig. 1d, including the associated CT percentage. The analysis revealed that the transition is primarily composed of local excitations on both imidazole rings and the phenyl ring of benzophenone, as shown in Fig. 1d. We have also observed an n → π* charge transfer from the hydroxyl and carbonyl groups to the benzophenone's phenyl ring (see 87.08% charge transfer transition in Fig. 1d). A minor contribution from the diphenyl group to the hole orbital is also observed.

Fig. 1 (a) Absorption and (b) emission (λ_ex = 310 nm) spectra of compound M1 in methanol solvent (10⁻⁴ M); (c) optimised structure of M1 using ωb97xd/6-31g(d,p) (functional/basis set); (d) (left) HOMO–LUMO and (right) second transition NTOs of M1 (here, the particle represents the electron density from which the transition occurs, while the hole represents the region from where the electron density decreases during the transition).

Table 1 Excited state as calculated by TDDFT. The energy and oscillator strength are also provided. The first five transitions are considered for M1 and M1–MMA–M1

Transition	Energy (eV)	Oscillator strength (f)
M1
S 1	3.92	0.0001
S 2	4.07	0.6192
S 3	4.32	0.1713
S 4	4.48	0.4440
S 5	4.59	0.3729
M1–MMA–M1
S 1	3.88	0.0345
S 2	3.92	0.0813
S 3	4.01	0.1608
S 4	4.07	0.1101
S 5	4.20	0.0682

---

The emission of the probe was extremely weak. The photoluminescence (PL) measurements were performed in methanol using 310 nm (4.00 eV) excitation. The PL experiment revealed two emission bands: a weak high-energy band near 390 nm and a prominent, broad, structureless band centred around 465 nm (Fig. 1b). The featureless nature of the latter band suggests a CT transition.²³ To validate this, a solvatochromism study was performed, which showed a red shift in emission with increasing solvent polarity, albeit a modest one (Fig. S9, SI). This is consistent with CT behaviour.

MMA sensing

Methylmalonic acid (MMA) detection was carried out using both colorimetric and fluorescence methods, employing compound M1. A 10⁻⁴ M methanol solution of M1 was prepared (lower concentration of M1 (10⁻⁵ M) not giving a turn-on response to MMA (Fig. S18, SI)), to which 0 to 30 µM aqueous MMA was added (30 µL each time), and the mixture was thoroughly mixed. Upon mixing, the solution changed from colourless to pale yellow in daylight and from weak emission to bright blue under a UV lamp (365 nm). This demonstrated a clear turn-on response of MMA to our probe, and we proceeded to investigate the sensing mechanism.

The UV-Vis analysis of MMA dissolved in the probe was performed in methanol (10⁻⁴ M). The spectrum showed the emergence of a new band at 415 nm (Fig. 2a) with an isosbestic point at 380 nm, indicating a ground-state interaction between M1 and MMA. The 214 nm and 310 nm absorption bands decreased, while the 415 nm band increased. Such a ratio-metric plot response is very important for real-life sample detection, as it reduces background interference.²⁴ A change in absorption at 415 nm with increasing concentrations of aqueous MMA was also plotted. An absorption increment was observed with a nearly 0.99 regression coefficient. The DFT calculations were performed with the probe and the analyte (MMA), at the same level of theory as above. NMR corroborated the presence of hydrogen bonding between M1 and MMA (vide infraFig. 6). Based on that, we optimized the M1–MMA–M1 structure by pre-adding the H-bonding interactions between them (Fig. 2c). Surprisingly, after optimization, MMA brought the two M1 molecules face-to-face (Fig. 2c). This suggested that MMA might be inducing aggregation of the molecules. Notably, this significantly impacted the HOMO and LUMO of the molecule, as illustrated in Fig. 2d.

Fig. 2 (a) UV-visible titration of M1 (10⁻⁴ M) with aqueous MMA concentrations (0 to 30 µM each time 30 µL added) in MeOH medium, an insight image of the same in daylight; (b) linear change in the absorption of M1 at 415 nm with the amount of aqueous MMA; (c) optimized and unoptimized structures of M1–MMA–M1 by pre-adding the H-bonding interactions; (d) the HOMO and LUMO of optimized M1–MMA–M1.

Furthermore, TDDFT calculations revealed that the two low-energy transitions were observed at 3.88 eV and 3.92 eV with oscillator strengths of 0.0345 and 0.0813, respectively (Table 1). Both these transitions are 3 orders of magnitude higher than the probe-only transition at 3.92 eV. Furthermore, the other transitions (S₃ = 4.01, f = 0.1608, and S₄ = 4.07, f = 0.1101) showed a lower oscillator strength as compared to the probe-only transition at 4.07, which had f = 0.6192 (Table 1). This corroborates the above ratio-metric observation from the UV-Vis results.

In the study of fluorescence spectra, initially, a weakly emissive blue-green solution exhibited bright blue fluorescence after the addition of aqueous MMA under UV light (365 nm) (Fig. 3a). The PL spectra showed a substantial increase in the emission intensity, approximately 73-fold (in comparison to the pristine), as calculated by the integrated area under the emission curve. The normalized spectra further confirmed a 10 nm blue shift in emission (Fig. 3b), and no new peak was observed, indicating that the interaction does not induce a new emissive state. Since M1 is insoluble in pure water, we made two solutions: one, a (9 : 1) water–methanol solution, and the other, a methanol-only solution, to examine the effect of water on M1. The PL spectra were then recorded (Fig. S11, SI). The decrease in emission intensity of the water–methanol solution demonstrated that water has no role in intensity enhancement.²⁵ Also, the (9 : 1) water–methanol solution does not show intensity enhancement upon the addition of MMA, as confirmed by emission intensity measurements (Fig. S17, SI). The probable reason is MMA's high-water solubility. A high fraction of water makes MMA completely soluble and opposes the formation of the M1–MMA complexation. The comparison of reported probes for MMA sensing is displayed in Table 3. Thus, M1 showed a fast response time of 20 seconds and gives a turn-on emission (Table 3 and Fig. S16, SI).

Fig. 3 (a) PL spectrum (λ_ex = 310 nm) of M1 in methanol solvent (10⁻⁴ M) before and after treatment with aqueous MMA (30 µM), at room temperature (25 °C); inset: image of the same under UV light (365 nm); (b) normalized PL spectrum of M1 before and after treatment with aqueous MMA; (c) PL spectra of the M1 probe solution after titration with different concentrations of aqueous MMA; (d) regression plot of the same.

Table 2 Detection of MMA in real urine samples

Urine samples	MMA added (µM)	MMA found (µM) ± SD	% Recovery
Sample 1	6.00	6.49 ± 0.18	108.16
Sample 2	10.00	9.86 ± 0.38	98.60
Sample 3	17.80	17.08 ± 0.44	95.95
Sample 4	26.50	25.40 ± 0.14	95.84
Sample 5	35.30	33.83 ± 0.67	95.83
Sample 6	40.10	38.25 ± 0.53	95.38

---

Table 3 The comparison between different luminescence sensors of MMA. Probe M1 is synthetically simple, efficient, selective, and highly sensitive turn-on emissive in the presence of MMA

S. no.	Probe molecules [coordination polymer (cp)/MOF/organic molecules]	Sensing behaviour	Limit of detection (LOD)	Medium/solvent	Excitation/emission wavelength (nm)	Quantum yield (%)	Stokes shift (nm)	Response time (seconds)	Application	Ref.
1.	[Cd(bbip)(NH2-BDC) (H2O)]	Red shift with turn-on fluorescence	0.72 µM	EtOH	370/422	—	52	300	Urine	17
2.	Ru/Tb-MOF	Ratiometric fluorescence	3.8 µg mL^−1	—	366/545 and 585	—	79 and 119	60	Urine	37
3.	Zn-MOF	Turn-off fluorescence	1.7 nM	H2O	290/350	—	60	30	Urine	7
4.	Eu-MOF	Turn-off fluorescence	0.107 µM	—	468/525	1.53	57	120	Urine	16
5.	^35 n	Turn-off fluorescence	1.76 µM	H2O	275/400	—	125	—	Urine	35
	{[Co(L)(bimb)]}n		1.03 µM		270/375	—	105	—	Urine
	{[Co(L)(bimb)1/2]}n		0.18 µM		270/380	—	110	—	Urine
6.	{[Co(L)(bibp)]bibp·2H2O}n	Turn-on fluorescence	9.86 µM	—	275/385	—	110	—	Urine	38
7.	Molecular sensor 2	Turn-on fluorescence	1 µM	1% DMSO in ACN	340/512	—	172	—	Urine	15
8.	M1	Turn-on fluorescence	5.78 µM	MeOH	310/465	7	155	20	Urine	This work

---

To evaluate the detection sensitivity, a titration experiment was conducted by adding increasing concentrations of aqueous MMA to a 10⁻⁴ M methanol solution of M1. As shown in Fig. 3c, the emission intensity increased proportionally with the concentration of MMA. Using the standard method based on a signal-to-noise ratio (3σ/S), the LOD for MMA was calculated to be 5.78 × 10⁻⁶ M. The absolute quantum yield of M1 was enhanced from 0.07 to 0.51 after the interaction with MMA.

Selectivity study

Selectivity is a critical factor for the practical application of any sensing system. To assess the selectivity of probe M1 toward MMA in the presence of biologically relevant interfering elements, various analytes commonly found in human urine were tested under identical conditions. The selectivity of the M1 probe was tested in the presence (Fig. S12 SI) and absence of MMA (Fig. 4a). A 10⁻⁴ M methanol solution of M1 (2 mL) was treated with 100 µM of each analyte, including MMA (30 µM) and other species (present in urine) like L-proline, urea, glucose, CaCl₂, KCl, NaHCO₃, NaCl, MgCl₂, Na₂SO₄, creatine, creatinine, uric acid, human serum albumin (HSA), and succinic acid. The PL intensity at 465 nm was used as the response metric using a 310 nm excitation (Fig. 4). The results demonstrated that none of the other analytes caused significant fluorescence enhancement apart from MMA, highlighting the high selectivity and robustness of M1 for MMA detection, even in complex biological matrices such as urine.

Fig. 4 (a) Photograph of the methanol solution of the M1 probe under a UV lamp (365 nm) after the addition of an aqueous solution of different urine constituents like MMA, L-proline, urea, glucose, CaCl₂, KCl, NaHCO₃, NaCl, MgCl₂, Na₂SO₄, creatine, creatinine, uric acid, human serum albumin (HSA), and succinic acid; (b) Bar graph showing the PL Intensity of M1 with different analytes (λ_ex = 310 nm) with (green colour), and without MMA (purple colour).

FTIR and NMR: hydrogen bonding between M1 and MMA

To investigate the interaction between probe M1 and MMA, FT-IR analysis was conducted on M1 and MMA (Fig. 5). The M1 spectrum showed peaks at 1650 and 1603 cm⁻¹, along with a high-energy broad peak at 3200–3500 cm⁻¹ (O–H and N–H). The 1650 cm⁻¹ and 1603 cm⁻¹ peaks were assigned to C=O and C=N stretches, respectively.²⁶ For MMA, a low-energy vibration at 1720 cm⁻¹ (C=O) was observed, along with a high-energy broad signal of OH stretching in the 3300–3400 cm⁻¹ region. FTIR was also performed for the combined probe analyte complex (M1–MMA) (Fig. 5). C=O stretches showed a redshift from 1650 to 1639 (M1) and from 1720 to 1702 cm⁻¹ (MMA). This clearly indicates a weakening of the C=O bond,²⁷ likely due to hydrogen bonding. Furthermore, the N–H and O–H stretches of M1 and MMA disappeared, indicating the involvement of N–H (M1) and O–H (MMA).²⁸ This proves the existence of H-bonding between M1 and MMA.

Fig. 5 FT-IR spectra of compound M1, MMA, and M1 + MMA.

Next, to investigate the exact interaction between hydrogen and oxygen, a ¹H NMR titration was performed in the CH₃OD solvent. First, a ¹H NMR spectrum of the probe (M1) was recorded in the presence of 2 equivalents of M1 to establish the baseline chemical shift under reference conditions. Subsequently, 1 equivalent of MMA was added to the same NMR tube while maintaining a constant volume, mixed thoroughly, and a second ¹H NMR spectrum was acquired to observe the analyte-induced chemical shift. The ¹H NMR spectra (Fig. 6b) showed both shielding and deshielding effects upon the addition of MMA. Notably, the H_d protons (–NH, see Fig. 6a) of M1 were deshielded after the addition of 1 equivalent of MMA, suggesting their direct involvement in binding with the analyte. Furthermore, the aromatic protons of M1 also shifted: H_b and H_c protons were deshielded, while the H_a proton was shielded. These changes are attributed to the altered electronic environment and the distinct ring currents due to the non-coplanar orientation of the phenyl rings.²⁹ The H_OH proton signal was not observed, likely due to the exchange with the trace amounts of moisture present in CD₃OD, which was further confirmed by recording the ¹H NMR spectra in the DMSO-d₆ solvent (Fig. S6 SI).³⁰

Fig. 6 (a) Schematic representation shows the interactions between the M1 compound and MMA; (b) ¹H-NMR spectra of compound M1 (1 equiv.) and MMA (2 equiv.) in the CD₃OD solvent.

Upon addition of aqueous MMA, the analyte appears to participate in intermolecular hydrogen bonding with a nitrogen atom in M1. The stoichiometry of the M1–MMA interaction was further examined using Job's plot (Fig. S14 SI). For this, 10⁻⁴ M methanolic solutions of M1 and MMA were mixed in varying mole fractions while maintaining a constant total volume of 2 mL. The absorption spectra were recorded for each mixture (Fig. S14a SI), and a plot of absorbance versus the molar fraction of M1 showed a maximum at a 0.7 molar fraction (at 415 nm). This indicates a ∼2 : 1 stoichiometric ratio of M1 to MMA, confirming the formation of an M1–MMA complex. Time-resolved fluorescence spectroscopy was also performed for M1 and the M1–MMA mixture. The lifetime of M1 (0.14 ns) decreases (0.11 ns) after the interaction with MMA (Fig. S10, SI).

Thus, corroborating our hypothesis of the presence of H-bonding between M1 and MMA. This was used to carry out M1–MMA–M1 DFT calculations in the previous section and validates the computational results based on H-bonding. Next, we explore the role of this H-bonding in the formation of aggregates in the M1–MMA solution.

Analyte triggered aggregation (ATA): DLS, FESEM and zeta potential study

To investigate the ATA mechanism, we performed dynamic light scattering (DLS), field-emission scanning electron microscopy (FESEM) analysis, and zeta potential measurements of the probe to demonstrate aggregate formation after the addition of the analyte (MMA) (Fig. 7). From the DLS data, it is evident that as MMA (30 µM) is added to the M1 probe solution, the average particle size increases from 667.14 nm to 1534.33 nm. Additionally, DLS measurements exhibit a uniform PDI (<0.4) with a narrow distribution of particles (Fig. 7). FESEM images of M1 and M1–MMA also confirm the aggregate formation in the solid state in the presence of MMA. In the FESEM of M1 (Fig. 7a), the particles are round in shape with an average particle size of 438.8 nm. After the addition of MMA, the particles of M1 aggregated with an average particle size of 2.94 µm (Fig. 7b).

Fig. 7 (a) DLS spectrum (left side) (inset: representation of M1 molecules without MMA) and the FESEM image (right side) of the M1 probe without MMA; (b) DLS spectrum (left side) (inset: representation of M1 with MMA) and the FESEM image (right side) of the M1 probe in the presence of MMA.

Both DLS and FESEM confirmed the formation of analyte-triggered aggregates. One additional experiment was also performed to support the MMA-triggered aggregation formation by diluting the M1–MMA solution from 10⁻⁴ M to 10⁻⁷ M with methanol (Fig. S15, SI). The emission intensity of the solution decreases with dilution, suggesting a decrease in the aggregation.

The zeta potential of M1 before and after the addition of MMA was measured with a narrow zeta potential distribution (Fig. S13 SI). The average zeta potential of M1 and M1–MMA was measured to be +14.4 mV and +6.9 mV, respectively (Fig. 8). Methyl malonic acid contains a –COOH group, which partially has a negative charge. The electrostatic interaction between M1 and MMA decreases the overall positive charge. A change in the positive zeta potential suggests the aggregation of M1 induced by MMA.^31–33 All these experiments demonstrate that MMA is responsible for inducing aggregation mediated by hydrogen bonding with the probe M1.

Fig. 8 Zeta potential of M1 and M1–MMA solutions.

MMA sensing in human urine

We have also evaluated the M1 probe for detecting MMA in human urine samples. To quantify unknown MMA concentrations, a calibration curve was first established by plotting fluorescence intensity against known MMA concentration (Fig. 9b). The spiked recovery method was employed for detection.³⁴ The changes in the emission intensity were observed under a UV lamp (λ_ex = 365 nm) after treating the M1 solution with MMA-spiked urine samples (Fig. 9a).

Fig. 9 (a) Image of the M1 probe solution (10⁻⁴ M) after being treated with an MMA spiked (30 µM) urine sample (under a UV-lamp); (b) calibration curve of the M1 probe after being treated with different concentrations of aqueous MMA; (c) the scheme shows the urine sample and its pre-treatment.

The six urine samples, collected from the BITS Medical Centre, BITS Pilani, were centrifuged at 3000 rpm for 10 minutes to remove suspended particles. The resulting supernatant was spiked with known MMA concentrations. For each measurement, 30 µL of spiked urine was added to 2 mL of a 10⁻⁴ M M1 solution in methanol (Fig. 9c), and the fluorescence emission was recorded. MMA concentrations were then determined using the calibration curve. The recovery results, shown in Table 2, ranged from 95% to 108%, demonstrating the probe's effectiveness in detecting MMA in real urine samples and are comparable to literature reports.^7,35

Low-cost MMA detection on a TLC strip using a smartphone

Based on earlier discussions, it is necessary to accurately measure MMA in urine for monitoring vitamin B₁₂ deficiency. However, advanced path lab facilities are mostly unavailable in remote areas for rapid on-site MMA detection. To overcome this, an alternative method using a “chemically modified TLC strip” was developed, which abolishes the complex instrumentation requirement, tedious sample preparation, and specialized personnel. The M1-coated TLC strip showed no colour in daylight but a light green colour under hand-held UV light (365 nm) (Fig. 10b). The colour changes from green to blue on the addition of spiked MMA urine under a UV lamp. The intensity of blue fluorescence increases with MMA concentration. The control experiment was also performed in the presence of water only, and it was found that there was no significant change in blue colour intensity (Fig. 10b). When the strips were treated with 20 µL of different MMA concentrations, a dependent intensity of blue-coloured fluorescence was observed. The blue colour intensity was captured using a smartphone RGB camera in a black box (Fig. 10a). The colours (red, green, and blue) on the standard RGB scale are represented by whole numbers ranging from 0 to 255. The number [255,255,255] on the scale denotes true white, while [0,0,0] denotes absolute black.³⁶ The result shows that the blue colour intensity increases gradually with increasing MMA concentration. By considering the blue colour intensity as a signal, a linear trend was observed between MMA concentration and blue colour. The blue colour intensity is linear (R² = 0.9766) in the concentration range of 2 to 42 µM (Fig. 10c). These results suggest that the detection of MMA in an “on-site” way is possible using a smartphone.

Fig. 10 (a) The scheme shows the preparation of the M1-coated TLC strip and captures the image from a smartphone RGB camera under a UV lamp after being treated with different spiked concentrations of MMA in urine; (b) TLC strip image (left) along with blue colour intensities (right) captured using a smartphone RGB camera after being treated with different MMA concentrations along with the control experiment (only water); (c) blue colour versusMMA concentration in the range of 2 to 42 µM concentration.

Conclusions

In summary, we have synthesized a new imidazole-functionalized organic fluorescent probe, M1, in a convenient route capable of selectively detecting MMA through a turn-on mechanism. The emission intensity increased by about 73-fold, which is the best reported to date. The sensor operates via ATA formation, supported by hydrogen bonding interactions and electrostatic interactions between the M1 and MMA, as confirmed by FTIR, ¹H NMR spectroscopy, DLS, FESEM, and zeta potential measurements. This ATA (analyte-triggered aggregation) led to the formation of a new lowest excited state, as confirmed by the formation of a new band in the UV-Vis spectrum and TDDFT results. This new state leads to turn-on emission, confirming the ATA-mediated emission. The complex formation displayed ∼2 : 1 (M1: MMA) stoichiometry, as confirmed by Job's plot analysis. M1 exhibited pronounced fluorescence enhancement with a low detection limit of 5.78 µM, enabling trace-level MMA sensing. It also supports low-cost, on-site detection via TLC plate-based imaging on a smartphone. Notably, M1 showed excellent selectivity over common urinary analytes and accurately quantified MMA in spiked human urine (95–108%). This non-invasive, user-friendly method holds strong promise for rapid point-of-care screening for vitamin B₁₂ deficiency-related metabolic disorders.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). The supplementary information contains instrumentation, methodology, NMR, HRMS, crystal structure, experimental data, and additional supporting data. See DOI: https://doi.org/10.1039/d5ma01483b.

CCDC 2455842 contains the supplementary crystallographic data for this paper.³⁹

Acknowledgements

This work was supported by the Department of Biotechnology (DBT), Govt. of India (project number: BT/PR33133/MED/32/704/2019) for financial assistance, and the UGC sponsored Special Assistance Programme (F.540/14/DRS/2007, SAP-I) and DST-FIST (SR/FST/CSI-270/2015) for the HRMS facility in the Department of Chemistry, BITS Pilani. The authors are grateful to the Medical Centre, BITS Pilani, Pilani Campus, for providing the urine samples.

References

1. S. Zakari, N. K. Niels, G. V. Olagunju, P. C. Nnaji, O. Ogunniyi, M. Tebamifor, E. N. Israel, S. E. Atawodi and O. O. Ogunlana, Front. Oncol., 2024, 14, 1405267.
2. J. Hernandez Roman and M. S. Siddiqui, Endocrinol., Diabetes Metab., 2020, 3, e00127.
3. K. Lazaros, S. Adam, M. G. Krokidis, T. Exarchos, P. Vlamos and A. G. Vrahatis, Sensors, 2025, 25, 1396.
4. D. J. Harrington, in Laboratory Assessment of Vitamin Status, ed. D. Harrington, Academic Press, 2019, pp. 265–299 DOI:10.1016/B978-0-12-813050-6.00012-7.
5. W. Zhou, H. Li, C. Wang, X. Wang and M. Gu, Front. Genet., 2019, 9, 726.
6. M. J. Shipton and J. Thachil, Clin. Med., 2015, 15, 145–150.
7. X. Zhang, Y. Tian, J. Shi, X. Kang and Z. Liu, J. Mater. Chem. C, 2022, 10, 12821–12828.
8. Ø. Midttun, A. McCann, O. Aarseth, M. Krokeide, G. Kvalheim, K. Meyer and P. M. Ueland, Anal. Chem., 2016, 88, 10427–10436.
9. M. Aghamohammadi, P. Shahdousti and B. Harooni, Microchem. J., 2016, 124, 188–194.
10. V. M. Carvalho and F. Kok, Anal. Biochem., 2008, 381, 67–73.
11. N. Sriboonvorakul, N. Leepipatpiboon, A. M. Dondorp, T. Pouplin, N. J. White, J. Tarning and N. Lindegardh, J. Chromatogr. B:Anal. Technol. Biomed. Life Sci., 2013, 941, 116–122.
12. T. Bito, Y. Matsunaga, Y. Yabuta, T. Kawano and F. Watanabe, FEBS Open Bio, 2013, 3, 112–117.
13. K. B. Akshaya, V. Anitha, M. Nidhin, Y. N. Sudhakar and G. Louis, Talanta, 2020, 217, 121028.
14. Z. K. Shihabi and M. A. Friedberg, Electrophoresis, 1997, 18, 1724–1732.
15. H. Miyaji, J. Fujimoto, R. Mabuchi, M. Okumura, S. Goto and Y. Honda, Tetrahedron Lett., 2017, 58, 3623–3627.
16. X. Lu, Y. Tang, G. Yang and Y.-Y. Wang, J. Mater. Chem. C, 2023, 11, 2328–2335.
17. W.-H. Xu, S.-L. Yao, X.-C. Xie, P. Wen, J.-Y. Zhong, J.-Y. Ding, R.-H. Wu, J.-L. Lin, H. Liu and S.-J. Liu, J. Mol. Struct., 2025, 1323, 140758.
18. Z. Yao, H. Bai, C. Li and G. Shi, Chem. Commun., 2010, 46, 5094–5096.
19. H. Dong, F. Zou, X. Hu, H. Zhu, K. Koh and H. Chen, Biosens. Bioelectron., 2018, 117, 605–612.
20. J.-H. Lin and W.-L. Tseng, Talanta, 2015, 132, 44–51.
21. F. Aldabbagh, in Comprehensive Organic Functional Group Transformations II, ed. A. R. Katritzky and R. J. K. Taylor, Elsevier, Oxford, 2005, pp. 99–133 DOI:10.1016/B0-08-044655-8/00048-9.
22. N. Farahani, K. Zhu, N. Noujeim and S. J. Loeb, Org. Biomol. Chem., 2014, 12, 4824–4827.
23. J. C. G. Bünzli, in Reference Module in Materials Science and Materials Engineering, Elsevier, 2016 DOI:10.1016/B978-0-12-803581-8.01855-5.
24. Z. Wang, L. Zhang, Y. Hao, W. Dong, Y. Liu, S. Song, S. Shuang, C. Dong and X. Gong, Anal. Chim. Acta, 2021, 1144, 1–13.
25. J. Devasia, F. Joy and A. Nizam, Chem. - Eur. J., 2023, 29, e202203652.
26. L. Ren, J. Wang, X. Meng, C. Liu, Z. Liao and D. Zhang, Inorg. Chim. Acta, 2025, 583, 122705.
27. L. Paoloni, A. Patti and F. Mangano, J. Mol. Struct., 1975, 27, 123–137.
28. K. N. Blodgett, J. L. Fischer, T. S. Zwier and E. L. Sibert, Phys. Chem. Chem. Phys., 2020, 22, 14077–14087.
29. J. Mruk, L. Pazderski, J. Ścianowski and A. Wojtczak, Inorg. Chim. Acta, 2020, 500, 119182.
30. F. Bonaldo, F. Mattivi, D. Catorci, P. Arapitsas and G. Guella, Molecules, 2021, 26, 3544.
31. X. Jiang, H. Jin, Y. Sun and R. Gui, Microchim. Acta, 2019, 186, 580.
32. S. Tanvir, S. Pulvin and W. Anderson, MOJ Toxicol., 2015, 1, 00011.
33. T. Yuan, L. Gao, W. Zhan and D. Dini, Pharm. Res., 2022, 39, 767–781.
34. D. Thakur, N. P. Dubey and R. Singh, Crit. Rev. Anal. Chem., 2024, 54, 2053–2071.
35. R. Luo, C.-G. Xu, D.-M. Zhang, L.-L. Wang, R.-X. Wu, G.-B. Chen, P. Lu, Y.-H. Fan and F. Shao, Talanta, 2023, 265, 124803.
36. X. Liu, R. Gao, L. Han, C. Kan and J. Xu, Talanta, 2023, 252, 123849.
37. Y. Zhang, X. Qu and B. Yan, J. Mater. Chem. C, 2021, 9, 3440–3446.
38. R. Luo, C.-G. Xu, H.-J. Yu, R.-X. Wu, P. Lu, Y.-H. Fan and F. Shao, CrystEngComm, 2023, 25, 4120–4125.
39. CCDC 2455842: Experimental Crystal Structure Determination, 2026 DOI:10.5517/ccdc.csd.cc2nfhqb.

---