Advertisement

Analytical and Bioanalytical Chemistry

, Volume 410, Issue 3, pp 1111–1121 | Cite as

Functional electrospun nanofibers for multimodal sensitive detection of biogenic amines in food via a simple dipstick assay

  • Nadezhda S. Yurova
  • Alexandra Danchuk
  • Sarah N. Mobarez
  • Nongnoot Wongkaew
  • Tatiana Rusanova
  • Antje J. Baeumner
  • Axel Duerkop
Research Paper
Part of the following topical collections:
  1. ABCs 16th Anniversary

Abstract

Electrospun nanofibers (ENFs) are promising materials for rapid diagnostic tests like lateral flow assays and dipsticks because they offer an immense surface area while excluding minimal volume, a variety of functional surface groups, and can entrap functional additives within their interior. Here, we show that ENFs on sample pads are superior in comparison to standard polymer membranes for the optical detection of biogenic amines (BAs) in food using a dipstick format. Specifically, cellulose acetate (CA) fibers doped with 2 mg/mL of the chromogenic and fluorogenic amine-reactive chameleon dye Py-1 were electrospun into uniform anionic mats. Those extract cationic BAs from real samples and Py-1 transduces BA concentrations into a change of color, reflectance, and fluorescence. Dropping a BA sample onto the nanofiber mat converts the weakly fluorescent pyrylium dye Py-1 into a strongly red emitting pyridinium dye. For the first time, a simple UV lamp excites fluorescence and a digital camera acts as detector. The intensity ratio of the red to the blue channel of the digital image is dependent on the concentration of most relevant BAs indicating food spoilage from 10 to 250 μM. This matches the permitted limits for BAs in foods and no false positive signals arise from secondary and tertiary amines. BA detection in seafood samples was also demonstrated successfully. The nanofiber mat dipsticks were up to sixfold more sensitive than those using a polymer membrane with the same dye embedded. Hence, nanofiber-based tests are not only superior to polymer-based dipstick assays, but will also improve the performance of established tests related to food safety, medical diagnostics, and environmental testing.

Graphical Absract

Keywords

Biogenic amines Electrospun nanofibers Rapid diagnostic Food analysis Fluorescence Dipstick 

Introduction

Rapid testing of biogenic amines (BAs) is of increasing interest because they are potential indicators of the different stages of freshness of protein-rich food, cheese, and fermented food. BAs can have adverse effects on human health even in concentrations that cannot be recognized by the human nose [1, 2, 3]. It is therefore not surprising that many instrumental methods for the analysis of BAs have been proposed. Among them are chromatography [4], capillary electrophoresis [5], sensors, flow-injection analysis [6], and ELISA [7, 8]. All of them are suitable for a precise determination of individual concentrations of BAs in a laboratory environment with well-educated staff and with costs per sample being less important, which makes translation into “in-field” operations impossible.

In-field BA determination requires inexpensive and simple on-site methods for quantitation, i.e., a dipstick [9]. Here, the aim is not to obtain a detailed concentration of each individual BA in the sample but rather an overall concentration level of all BAs potentially contained within to judge on the freshness or potential danger of a food sample [10]. Therefore, rapid tests such as dipsticks are advantageous because they can be easily operated by an inexperienced worker in the field and because they are also cheap enough to be widely available. There are many visual tests for determination BAs in food. It has been reported that a nanoporous colorimetric sensor array for trimethylamine (TMA) detection was developed [11]. In this work, a sol–gel method was used to obtain a TiO2 nanoporous film as substrate material to improve the sensitivity and stability of the colorimetric sensor. Several other optical-based sensors have been described for the detection of BAs using various chromogenic reagents, including Meldrum’s activated furan (MAF) for the determination of amines in solution, on solid supports, and in the vapor phase [12]; simultaneous using of GJM-492 and Remazol Brilliant Blue (RBBR) for detection of ammonia and biogenic amines [13]; or immobilizing of the indicator dye ETH4001 for the development of a spermidine and spermine sensing system [14]. Moreover, there are further colorimetric chemical sensors based on various classes of indicator dyes, such as phthalocyanines [15], porphyrins [16], calixarenes [17], and others. Recently, the application of gold nanoparticles (AuNPs) has acquired much attention for the development of different sensors. Tyramine-protected AuNPs were used as a probe for colorimetric and fluorescence turn-on detection of spermine and spermidine based on AuNP aggregation in the presence of diamines [18]. An optical method relying on the application of a Cu2+ complex of organic nanoparticles for the simultaneous quantification of spermidine and spermine in vapors and aqueous phase [19] and a sensor based on AuNPs for rapid detection of histidine and histamine in meat samples [20] were proposed. The aforementioned sensors commonly suffer from the fact that some require a large number of different reagents and only one of them permits both colorimetric and luminescence readout [18]. As luminescence is in general more sensitive than a colorimetric readout, the creation of dipsticks offering luminescent readout with additional colorimetric detection of BAs can provide a dual response by providing both a yes/no answer and quantitative information [21].

Electrospun nanofibers are an emerging field in biosensing and chemosensing because they have a high porosity and immense surface area, are easy to handle, can be mass-produced, and can be reusable. Furthermore, electrospun nanofibers provide excellent loading capacities for immobilization of recognition molecules that can introduce an optical response into a fiber net [22]. Commonly, polymeric materials are used for the production of fibers like polymethyl methacrylate (PMMA), polyacrylamide (PAM), polystyrene (PS), polyvinylpyrrolidone (PVP), polyvinyl alcohol (PVA), polycarbonate (PC), or cellulose acetate [23]. By spinning a charged polymer, additional functionality can be introduced into the resulting fiber mat because the charged surface of the resulting nanofibers can serve to enrich a counter-charged analyte. That means that charged electrospun nanofibers could adopt functionality similar to a solid-phase extraction material. Hence, for the quantitation of BAs (which are commonly cationic in aqueous solution of neutral pH), a negatively charged polymer should be chosen. By additional doping the fibers with an appropriate chromogenic fluorescent probe for BAs, a new material for analyte extraction and optical readout of dipsticks with three different methods could be obtained.

We therefore embedded a blue dye (Py-1) [24] that is responsive to BAs into electrospun nanofibers made of cellulose acetate (CA) to obtain uniform, flexible, blue colored anionic nanofibers that respond to the presence of various BAs in three different ways. First, a color change from blue to red can be seen by naked eye. Secondly, the reflectance of the fibers at 611 nm changes upon dipping into a liquid sample containing BAs. Finally, there is a pronounced increase in fluorescence at 588 nm in proportion to the concentration of BAs. Fluorescence emission was acquired from dipsticks as a digital image upon illumination with a simple handheld UV-lamp. From the images acquired in RAW-format, the intensities of the red, green, and blue channels are extracted with public-domain software and the red-to-blue intensity ratio was used for the calibration of BA concentrations. The fibers on the dipsticks selectively respond to primary amines but not to secondary or tertiary amines and were successfully tested with real samples.

Materials and methods

Materials

Py-1 was from ActiveMotif Chromeon (www.chromeon.com). The buffer N-cyclohexyl-2-amino ethanesulfonic acid (CHES) was from Roth (www.carlroth.de). Spermidine, putrescine, cadaverine, 2-aminoethylmethacrylate, dimethylamine were purchased from Sigma-Aldrich (www.sigmaaldrich.com) and histamine from Fluka, all as hydrochloride salts. Tyramine and trimethylamine, each as free base, were from Sigma. All amines were of analytical grade. Poly-styrene-co-acrylonitrile (SAN) (Mw 185,000 Da), poly-methyl methacrylate (PMMA) (Mw 996,000 Da), cellulose acetate (CA) (Mw 30,000 Da, 39.8 wt% acetyl content), and poly-vinylpyrrolidone (PVP) (Mw 1,300,000 Da) were obtained from Sigma-Aldrich. The polyurethane polymer (HydroMed D4) was obtained from AdvanSource Biomaterials (www.advbiomaterials.com). Indium tin oxide (ITO) coated on polyethylene terephthalate (PET) with a surface resistivity 60 Ω/sq was purchased from Sigma-Aldrich.

Stock solutions of BA (10.0 mM) were prepared in CHES buffer (pH 9.5). Working standard solutions of compounds were freshly prepared by diluting stock solutions with CHES. CHES buffer (5.00 mM) was prepared by dissolving of solid CHES (0.1036 g) in 100 mL of deionized water. The pH of CHES was adjusted with sodium hydroxide (1.00 M).

Apparatus

The electrospinning was performed using a commercial electrospinning machine (Spraybase® power supply unit PLS000048 and Spraybase® syringe pump module PLS000004). Fiber mat thickness, fiber diameter, and images for pore size determination were obtained by an Olympus LEXT OLS4000 3D Measuring Laser Microscope with 10 nm minimum z-resolution. pH was checked with a pH meter CG 842 from Schott (www. si-analytics.com). Fluorescence spectra of dipsticks were acquired in a solid sample holder (with 30° incident angle) with a Jasco FP-6300 luminescence spectrometer with 520 nm excitation and 5 nm slits for excitation and emission monochromator, respectively, 500 nm scan speed, 0.5 nm data pitch, and a PMT voltage adjusted to “medium” in the software. All spectra are corrected. Reflectance spectra were acquired with an Ocean Optics Reflectance measurement kit, containing of a white light source, a y-shaped bifurcated optical fiber, a fiber holder, and a Flame-S VIS-NIR spectrometer.

Electrospinning of fibers containing Py-1

The polymer CA (0.720 g) and Py-1 (8.00 mg) were dissolved in a mixture of 3.00 mL of acetic acid and of 1.00 mL of acetone. Then, this spinning dope should be stirred for about 48 h until Py-1 is completely dissolved and the mixture is homogeneous. The spinning dope should be protected from light while stirring. The CA-Py-1 electrospun nanofibers were fabricated using the electrospinning machine with the following parameters: plastic syringe 5 mL (covered with aluminum foil); voltage, 17 kV; flow rate, 0.002 mL/min; tip-to-collector distance, 11 cm. An ITO film (size 7 × 5 cm) was used as a supporting material. The resulting materials were prepared with different electrospinning times (15, 30, 60 min). Electrospun fiber mats on ITO should be stored dry and dark.

Preparation of dipsticks and BA determination

The dipsticks with CA-Py-1-nanofibers (Ø = 8 mm) were cut from the ITO sheets with a hole puncher and mounted into a multi stick holder which can house several sticks in a row, circumvented by black, solvent resistant plastic. After that, 5.00 μL of BA (in CHES buffer, pH 9.5) was added on the circles with a micropipette. The multi stick holder was immersed into an ethanol chamber (4 mL of ethanol) for spreading of the drop of analyte to the full size of the stick. After spreading, the spots were dried at ambient air (20 min). Then again, the holder was placed into the ethanol chamber for color development (20 min) of the sticks.

Acquisition of images and evaluation

For the fluorimetric determination of BA, pictures were taken in a dark room using a UV lamp (254 nm) for excitation. The lamp was positioned under a 50° illumination angle and fixed with respect to an adjusted position of the multi stick holder. The digital camera was mounted on top over the multi stick holder on a Novoflex MagicStudio Repro-Stand which permits reproducible distances between the objective of the camera and the stick holder. Images were acquired by means of a Canon EOS 550D camera equipped with a 67 mm UV filter using the following preset parameters f = 5.5; 1/13 s exposure time, ISO 6400, 57 mm distance, and custom white balance. The resulting photos were processed with the use of Photoshop CS6. In this regard, the spot was highlighted with the Lasso tool, the color was averaged over the area of the whole stick and the colorimetric parameters such as R, G, and B were determined by the software.

ImageJ software was used to determine pore size of the fiber mats. Here, the microscopic image was changed to 8-bit and the threshold was set at 35%. A pore of the fiber mat was treated as a particle in which Feret sizes larger than 0.5 μm2 were taken into account.

Preparation of real samples

Shrimp samples were purchased from a local discounter supermarket and stored at −22 °C. A 10.0 g portion was mixed with 100 mL of methanol in a beaker and homogenized in a blender at high speed for 2 min. The homogenisate was transferred into a conical flask and placed in a 60 °C water bath for 30 min. The extract was filtered through a porcelain Buchner funnel with blue ribbon filter paper (Schleicher und Schüll: 5893, www.whatman.com) for three times to yield particle-free samples. Then, 25.0 μL aliquots were taken and histamine solution was added to reach concentrations of 0–140 μM of added histamine after dilution to 500 μL overall volume with CHES buffer (5.00 mM; pH 9.5). After that, 5.00 μL of aliquot solutions were then added on dipsticks mounted into a multi stick holder and developed and imaged as described in the sections above.

Results and discussion

Choice of materials, conditions of spinning, and fiber morphology

The main idea (Fig. 1) of this research was to spin functional anionic nanofibers that could act as a solid-phase extraction material for BAs (that are cationic in aqueous samples) on the sample pads of dipstick assays. By designing a spinning dope with an appropriate chromogenic fluorescent probe (Py-1) for BAs, new nanofibers for both analyte extraction and optical readout of dipsticks should be obtained. Those nanofibers were tested with respect to their response to biogenic amines with various optical detection methods and to their applicability for BA monitoring during aging in real samples. The amine reactive probe Py-1 was chosen because of its chromogenic and fluorogenic properties. It is blue (\( {\uplambda}_{\mathrm{abs}}^{\mathrm{max}}=605\ \mathrm{nm}\Big) \) and virtually non-fluorescent (ϕ = 0.01) in its non-conjugated form in solution but shows a dramatic color change to red (\( {\uplambda}_{\mathrm{abs}}^{\mathrm{max}}=503\ \mathrm{nm}\Big) \) accompanied by a strong increase in fluorescence intensity (ϕ = 0.5) when covalently reacted with primary amino groups [21]. This enables a fluorescence readout of the reacted dye even in presence of unreacted Py-1. The optical response of the dye to various amines in solution was studied in detail in earlier work [24].
Fig. 1

Schematic rendering the nanofiber-based BA detection dipstick assay

Prior to electrospinning, both dye and polymer are dissolved in a solvent mixture to yield the so-called spinning dope. Solutions of poly-styrene-acrylonitrile (SAN), poly-vinylpyrrolidone (PVP), poly-methyl methacrylate (PMMA), HydroMed D4, and cellulose acetate (CА) were tested for their ability to dissolve the Py-1 dye, the stability of the dye embedded in the polymer, and for their optical response toward BAs. The polymer solutions (SAN 35%, PVP 14%, PMMA 5%, D4 20%, and CA 18%) and the dye solution (Py-1 1 g/L) were mixed on a microscope glass slide and after drying on air, a drop of BA solution (tyramine, 10 mM) was applied to evaluate the response of the respective spinning dope to the analyte. The tested polymers should have a threshold polarity as to enable a quick access of the polar BAs into the polymer and show a distinct change of color upon reaction with a BA.

As can be seen from Table 1, only three polymers (2, 4, and 5) enable a response of Py-1 to the amine. In the case of poly-styrene-acrylonitrile, the polymer is highly hydrophobic, which makes the analyte penetration into the polymer very difficult. This is also obvious from the fact that no color change is observed upon addition of the amine solution. Also, PMMA does not respond to tyramine and the dye is unstable in the polymer. PVP and D4 are not suitable for long-term use due to decomposition of the dye when mixed within. In the case of PVP, the crosslinking reaction at 165 °C (to render the polymer water-insoluble) leads to the decomposition of the dye. Only CA retains the color of the unreacted dye for a long time and displays a distinct color change upon reaction with a BA. Accordingly, CA was chosen for the spinning of nanofibers.
Table 1

Stability of polymers with embedded Py-1 and visual response to tyramine

Polymer

Characterization

Response to tyramine solution

1

Poly-styrene-acrylonitrile

High hydrophobicity, stability of color for 1 month

No

2

Poly-vinyl-pyrrolidone

Green to brown after 2 days at room temperature/after 1 h at 165 °C

Changing color from green to red

3

Poly-methyl-methacrylate

Disappearance of color after 2 h at room temperature

No

4

HydroMed D4

Disappearance of color after 1 h at room temperature

Changing color from green to pink

5

Cellulose acetate

Stability of color for 3 weeks

Changing color from blue-green to pink

For the optimization of the electrospinning process, Py-1 was dissolved in the CA spinning dope solution in 1 or 2 mg/mL, respectively. The latter concentration showed a brighter color of the BA-Py-1 conjugate, and this concentration was selected for further study. Spinning for 30 min onto indium tin oxide (ITO) slides yields homogeneous, well-formed nanofibers (Fig. 2). A number of different collector substrates were studied, including aluminum foil, glass slides, and filter paper which all lead to non-uniform distributions of nanofibers on the surface, upon spinning times > 15 min. In contrast, spinning onto ITO resulted in homogeneously distributed nanofibers (and color). This is presumably due to the high planarity of the ITO sheets which yields a more homogenous electric field on the collector plate during the electrospinning process.
Fig. 2

Microscopic image of CA-Py-1 nanofibers on ITO with 30 min electrospinning time (c(Py-1) = 2 mg/mL). The microscopic pictures were taken by a 3D laser scanning microscope with ×2160 magnification (×100 objective lens)

Characterizations of fiber morphology with respect to fiber mat thickness, pore size, and diameter were performed by a 3D measuring laser microscope. CA-Py-1 nanofibers were then prepared on ITО with different electrospinning times (15, 30, 60 min). The thickness of the resulting fiber mats were determined to be 0.65 ± 0.01, 6.9 ± 1.6, and 11.2 ± 0.4 μm and the respective fiber mats are displayed in the Electronic Supplementary Material in Figs. S1a–c. For the mats that are spun for 30 min, this is about the same thickness as of common polymeric sensor membranes. The related pore sizes derived from the 2D-microscopic image are 2.6 ± 1.3 μm as determined by the Feret diameter (example shown in Fig. S1d) from 1896 pores (n = 3). In addition, within a specified area, the density of pores is very high (~ 1.48 × 105 pores/mm2), enabling a huge accessible sensing area for BAs. The fibers have an average diameter of 168 ± 18 nm, as to determination with the 3D measuring laser microscope with ×2160 magnification.

Optical properties of nanofibers and acquisition of images

The CA fiber mat containing Py-1 as amine reactive probe is blue and virtually non-fluorescent (see Fig. 3(A2)), and reacts with primary amine groups in aqueous solution of pH 8–9.5 at room temperature to give a covalent red conjugate (see Fig. 3(B2)) [21]. This color change is visible at 210 μM which is close to the BA concentrations (0.3–1 mM) in foods that can induce serious health problems [21]. A visible red color of the dipsticks upon testing a sample therefore should be understood as a warning that the sample might be suspicious. The visual color change can also be derived from the reflectance spectra of the dipsticks monitored in the absence and presence of TYR, respectively (see Fig. S2). The reflectivity at 611 nm significantly increases in the presence of TYR because the blue form of Py-1 is no longer present and converted to the red one. This means that more light is reflected in the yellow/orange range of the spectrum thus leading to the faint red color of the dipstick shown in Fig. 3(B2). Hence, reflectance spectroscopy could be used for evaluation of the dipsticks, as well, albeit its sensitivity will not be very high.
Fig. 3

Left: images of CA-Py-1 nanofiber mats before (A1 and A2) and after (B1 and B2) reaction with tyramine under excitation with UV light 254 nm (A1 and B1) and under visible light (A2 and B2). Right: fluorescence spectra of unreacted and reacted (with 210 μM tyramine) CA-Py-1 nanofiber mats, respectively (λ em  = 588 nm for excitation and λ exc  = 517 nm for emission spectra; the excitation spectrum of the unreacted dipstick is omitted for clarity)

While the change of reflectance and visual color is not too pronounced due to the small amount of dye doped into the CA nanofibers, the dye concentration is sufficient for fluorescence analysis. The dye amount is smaller than in a plain polymer sensor membrane because the porosity of a fiber mat is much higher than that of a layer of knife-coated polymer. Doping of electrospun nanofibers with a fluorophore therefore helps reducing assay costs in comparison to polymer membranes. The fluorescence change from unreacted Py-1 (A1) in the fiber mat to orange red (B1) upon formation of the conjugate between Py-1 and tyramine when the fiber mat is excited at 254 nm with a UV lamp is highly pronounced. The blue color of the spot in A1 is the common background luminescence from the PET support of the ITO sheet and stronger than the weak red emission from Py-1. This is confirmed by the fluorescence spectra of Py-1 in the nanofiber mat where the initially barely detectable emission of unreacted Py-1 (\( {\uplambda}_{\mathrm{em}}^{\mathrm{max}} \)= 663 nm) in the fiber mat increases by a factor of about 10 upon formation of the conjugate between Py-1 and tyramine (\( {\uplambda}_{\mathrm{em}}^{\mathrm{max}} \)= 588 nm). Compared to the emission spectra in solution [25], the emission maximum of unreacted Py-1 remains almost unchanged, whereas the emission maximum of reacted Py-1 shifts 15 nm shortwave in the mat. The excitation maximum of the reacted form of Py-1 embedded in the nanofibers is found at 517 nm.

Assay procedure for quantitation of BAs

A simple assay protocol was developed (Fig. S3) using the optimized nanofiber mats in a dipstick format for the fluorimetric detection of BA. Nanofiber mat circles (Ø = 8 mm) were cut and fit into a multi stick holder to enable the readout of multiple dipsticks in a row under the same conditions for fluorescence excitation and detection. The mats are surrounded by black solvent-resistant plastic to suppress stray light and ensure an even dark background when acquiring the digital images. As the extracts of BAs from real samples are typically delivered in an aqueous/alcoholic mixture and this environment assists in the even distribution of the small BA volume over the fiber mat, the stick holder was located in an ethanol chamber after addition of BA (in CHES buffer, pH 9.5). After spreading, the spots were dried at ambient air (20 min). Then again, the holder was placed into the ethanol chamber for color development (20 min). It should also be noted that it is known from earlier work that the BA conjugation with Py-1 works more rapid and reproducible in an alcoholic environment and the emission of the conjugate is brighter when derivatized in an alcohol [26].

The incubation time of the BA determination was studied in more detail using fiber mats with three different electrospinning times (15, 30, 60 min). Digital images in RAW-format were taken each 5 min under fluorescence excitation by a UV-lamp at 254 nm under a 50° illumination angle. The intensity of the red, green, and blue channel was extracted via Adobe Photoshop CS6 software. Data from the red channel contained the emission intensity information of the amine-Py-1 conjugate. The blue channel contained the low emission intensity information from the PET support of the ITO sheet. Finally, the intensity of the red channel was divided by one of the blue channel and the red-to-blue intensity ratio was used as a parameter for BA detection. Using excitation by a UV-lamp for dipstick assays containing Py-1 is new and was never described in earlier publications, so far. This omission of a noncommercial excitation source further simplifies the assay procedure and makes it amenable to less-qualified users which is beneficial for rapid diagnostic test procedures.

The ratio of the luminescence found in the red to the blue channel of the digital image of replicate tests was followed over time. As can be seen from Fig. 4, the development of the change of the fluorescence takes 35–40 min after addition of the BA. Over the first 20 min after BA addition, the drop is spreading and the fluorescence intensity ratio is almost constant over the remaining time. After 20 min, there is a pronounced increase of the R/B signal for about additional 20 min. At incubation times longer than 40 min (after addition of the BA drop), no further significant increase in signal is observed. Further analyses were therefore done using a 40-min overall incubation time. Furthermore, a significant increase in the fluorescence intensity ratio with development in the ethanol vs. ambient air chamber is noted, as expected, based on earlier results using polymer films [26]. Electrospinning time only modestly affects the increase of the luminescence intensity ratio of the dipstick tests over the whole incubation time and the highest increase is found for the fiber mats with 30-min spinning time. Small variations in spinning time will therefore not negatively affect the reproducibility of the assay.
Fig. 4

Effect of incubation time on luminescence response (expressed as R/B-ratio) of dipstick assays upon addition of tyramine (c(TYR) = 1 mM); 0–20 min: drying of sensing spot after BA addition; 20–50 min: color development (n = 3)

Calibration and sensitivity

The effect of different BA concentrations on the luminescence intensity ratio of the dipstick assay was studied with tyramine as a model. The calibration curves for fiber mats with three different electrospinning times (15, 30, 60 min) are shown in Fig. 5. We used eight concentrations of tyramine (0.01; 0.025; 0.05; 0.1; 0.25; 0.5; 0.75; 1.0 mM) and three replicates for each concentration. The R/B fluorescence intensity ratio is well linear proportional to the concentration of tyramine in the range from 10.0 to 100 μM of tyramine, as shown in the inset of Fig. 5. This is lower by about one order of magnitude compared to dipsticks consisting of a knife-coated polyurethane membrane with embedded Py-1 and a reference dye [21] and fluorescence RGB imaging via a digital camera. The signal increase reaches saturation at 250 μM of tyramine. The LOD is defined as the analyte concentration yielding a red/blue (R/B) signal equal or higher than the average value produced by the blank sample plus three standard deviations. The LOD of tyramine is 0.009 mM using a mat with 15 min of electrospinning, 0.003 mM for 30 min, and 0.006 mM for 60 min of electrospinning, respectively. These LODs are lower by up to a factor of 6 compared to earlier work using knife-coated polymer membrane dipsticks with Py-1 [21] where a LOD of 0.02 mM was found. Parameters of the linear calibration plots for tyramine detection are shown in Table 2. The correlation coefficients are on a comparatively high level considering that rapid diagnostics tests such as dipstick assays always show less reproducibility than sophisticated instrumental detection methods. Compared to earlier linear calibration data [21] with R 2 from 0.92 to 0.95, the dipstick assays based on electrospun nanofibers show considerable better performance. The slope of the calibration plots slightly increase with electrospinning time. This behavior could be expected because the thickness and density of the nanofiber mat increase with spinning time. This means that more reactive Py-1 is available per surface area and therefore can lead to a higher change of the luminescence intensity ratio. As to their luminescence response, the mats are stable for more than 3 months (longer times could not be tested, yet).
Fig. 5

Response of CA-Py-1 nanofibers based dipsticks to different tyramine concentrations. The insert shows the calculated calibration curves (n = 3)

Table 2

Parameters of calibration curves of tyramine detection

Electrospinning time, min

Equation of calibration plot

R 2

15

y = 11.26× + 0.416

0.98

30

y = 11.803× + 0.522

0.96

60

y = 12.027× + 0.503

0.98

The concentration range of BAs relevant to induce health risks is between 0.3 and 1.0 mM. BAs in concentrations below 1.0 mM are not detectable by the human nose in most cases but may be indicative of food spoilage and, hence, represent a health risk [21]. The linear range of the determination of tyramine with the dipsticks using luminescent electrospun nanofibers therefore is just below this critical range of interest. This is advantageous because low but uncritical concentrations as, e.g., naturally occurring in shrimp [21] can be detected as well as increasing concentrations over prolonged storage. The latter case might require appropriate dilution during sample preparation, as described in the experimental section.

Selectivity

The selectivity of the new CA-Py-1-nanofiber-based dipstick assays were tested with the following amines: tyramine, histamine, cadaverine, spermidine, putrescine, 2-aminoethylmethacrylate, dimethylamine, triethylamine. These particular amines were chosen to cover monoamines and polyamines as well as aliphatic, aromatic, and heteroaromatic amines. We added BAs in an excess concentration (10 mM) and acquired the fluorescence intensity ratio readout of digital images for the mats obtained with different spinning times (i.e., thickness). As can be seen from Fig. 6, the nanofiber mats selectively respond to primary amines but not to secondary or tertiary amines. This is comparable to earlier findings with Py-1 embedded in other polymers [25, 27]. Earlier dipsticks [21] using the same dye showed that there can be certain interference of ammonia, cysteine, and dithiothreitol in 1–10 mM concentrations which decrease the fluorescence response of the dipstick to the BAs. This behavior cannot be ruled out for the dipsticks containing CA nanofibers. On comparing the average R/B signals of tyramine, histamine, cadaverine, spermidine, putrescine, 2-aminoethylmethacrylate, respectively, at the same spinning times, it is noted that the fiber mats with a 30-min spinning time respond most comparably to the presence of different BAs. This is further confirmed from the average relative standard deviations (RSDs) of the R/B values calculated for those BAs for the same spinning time which are 21.5, 16.5, and 17.3% for 15, 30, and 60 min, accordingly. This suggests the applicability of the mats to detect the overall concentration of various BAs in real samples and that they can be used as a pre-screening tool to determine the total amine content (TAC) in real samples before more advanced analytical methods are applied to determine individual concentration of each BA in a sample.
Fig. 6

Selectivity of the dipstick assays (obtained with various spinning time) toward various types of amines in 10 mM concentration (n = 3)

Quantitation of BAs in real samples

We finally tested the response of the dipsticks to BAs in shrimp samples over a 6-day storage period at 4 °C. The extraction of the shrimp samples follows the AOAC method 35.1.32 [28] and a standard addition method was applied. We choose histamine as BA to be added because it commonly accounts for the major fraction of all BAs contained in real meat or seafood samples [29]. As the fluorescence response of fiber mats on dipsticks spun for 30 min is very similar to both, tyramine and histamine (Fig. 6) and still similar to cadaverine and putrescine, the overall variation of the R/B response of the dipsticks induced by the presence of other BAs in a real sample was regarded acceptable. The equal fluorescence response of the Py-1 dye to tyramine and histamine was also found, recently in a sensor microplate [30]. Hence, all BAs in the sample will contribute to the fluorescence response of the dipstick and are expressed in equivalents of histamine, similarly as done in earlier work [31] and as validated against GC-MS data [10]. We found concentrations of 14.1 ± 0.3, 16.4 ± 1.5, and 38.8 ± 2.6 μmol/g (n = 4, each) of histamine, on days 0, 3, and 6, respectively. The slow increase of the BA concentration over 3 days and a distinctly higher content after 6 days of storage correspond well with aging profiles of real samples found in earlier work [31] and show the suitability of the dipsticks for BA quantitation in real samples. With its suitability for detection of real samples, the new dipsticks based on CA nanofibers fit well into the performance profiles of other optical sensors for the detection of BAs (Table 3).
Table 3

Comparative table of the merits of representative optical sensors for BAs

Composition of sensor layer/response time

Analyte

Detected signal

Analytical range, LOD

Real sample

Comment

Ref.

Combination of 2-fluoro-4-[4-(2-hydroxyethanesulfonyl)-phenylazo]-6-methoxyphenol and Remazol Brilliant Blue R, immobilized on cellulose microparticles/1.5 h

Methylamine, dimethylamine, putrescine, cadaverine, histamine, tyramine, tryptamine

Reflectometry

0.3–30 ppm

Chicken, pork

[32]

Paper strips based on Py 1 and fluorescein incorporated in Hypan HN 80 polymer/15 min

β-Alanine, Spermidine, Ethanolamine, Putrescine, Histamine, Tyramine

Fluorescence via RGB image from digital camera

0.5–2.0 mM, 0.75 mM

0.1–1.0 mM, 0.05 mM

0.5–2.0 mM, 0.25 mM

0.05–1.0 mM, 0.04 mM

0.25–2.0 mM, 0.10 mM

0.04–1.0 mM, 0.02 mM

Meat, fish, shrimp

 

[21]

Microtiterplate with film based on Py-1 embedded in Hypan HN 80/10 min

Histamine

Fluorescence in microtiterplate

0.00–70.0 μg/mL, 0.165 μg/mL

Meat, cheese (10 real samples)

Total BA content, validation results by GC/MS

[10]

Cellulose acetate nanofibers embedded with Py-1/20 min

Tyramine

Fluorescence via RGB image from digital camera

10–100 μM

Shrimp

Total BA content

This work

Bis(2,4,6-trichlorophenyl)-oxalate(TCPO)–H2O2 system modified with MgAl(CO3)2-LDH nanosheet colloids

Histamine

Chemiluminescence

0.1–100 μM (3.2 nM)

Fish, pork, meat

Total BA content

[33]

Glass support with hydroxyethyl cellulose membrane included luminol and covered with Co(II) and enzyme (putrescine oxidase or diamine oxidase)

Putrescine

Chemiluminescence in microplate

1–2 mg/L, 0.8 mg/L

Beef, pork, chicken, turkey and fish meat

Validation with HPLC

[34]

Gold nanoparticles (11–19 nm)

Histamine

Photometry or luminescence

0.6–12 μM

Poultry meat samples

 

[20]

Inkjet paper with monolayers/array of hollow AuNPs received by reversal nanoimprint lithography/30 min

Putrescine

Shift of reflectance spectra

0–200 ppm, 13.8 ppm

Spiked fish

 

[35]

Conclusion

New dipsticks using multimodal electrospun nanofibers for the determination of BAs with inexpensive instrumentation for chromogenic, fluorescence, or reflectance detection are presented. The nanofibers are based on cellulose acetate doped with the amine-reactive Py-1 dye. The readout can be done fluorimetrically upon excitation with a simple UV-lamp with a digital camera using the intensity ratio of the red versus the blue channel. The dipstick assay selectively responds to primary amines and is more sensitive by about one order of magnitude toward tyramine with respect to linear range and LOD than those based on a polymer film. On the one hand, the dipstick assay can be used for simple yes/no analysis by the naked eye or it allows semi-quantitative evaluation via comparison to a calibration color scale or using reflectance spectrometry. These sensitive tests can therefore be used for rapid determination of the total content of biogenic amines detection in food. This work demonstrates that the main advantage of using electrospun nanofibers as an immobilization matrix in rapid tests is a higher sensitivity and a lower consumption of the reagent for BAs due to the high porosity and surface area of the fiber mat that can improve the optical response of dipsticks.

Notes

Acknowledgements

NY and AD gratefully acknowledge support by the DFG and the Russian Ministry of Education for project 16.674.2016/DAAD and 16.751.2016/DAAD.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Supplementary material

216_2017_696_MOESM1_ESM.pdf (613 kb)
ESM 1 (PDF 613 kb)

References

  1. 1.
    Erim FB. Recent analytical approaches to the analysis of biogenic amines in food samples. Trends Anal Chem. 2013;  https://doi.org/10.1016/j.trac.2013.05.018.
  2. 2.
    Kalac P, Gloria MBA. Biogenic amines in cheeses, wines, beers and sauerkraut. In: Dandrifosse G, editor. Biological Aspects of Biogenic Amines, Polyamines and Conjugates. Scarborough: Research Signpost; 2009. pp. 267–310. http://www.ressign.com/UserBookDetail.aspx?bkid=873&catid=208
  3. 3.
    Fogel WA, Lewinski A, Jochem J. Histamine in food: is there anything to worry about? Biochem Soc Trans. 2007;  https://doi.org/10.1042/BST0350349.
  4. 4.
    Özdestan Ö, Üren A. A method for benzoyl chloride derivatization of biogenic amines for high performance liquid chromatography. Talanta. 2009;  https://doi.org/10.1016/j.talanta.2009.02.001.
  5. 5.
    He L, Xu Z, Hirokawa T, Shena L. Simultaneous determination of aliphatic, aromatic and heterocyclic biogenic amines without derivatization by capillary electrophoresis and application in beer analysis. J Chromatogr A. 2017;  https://doi.org/10.1016/j.chroma.2016.12.067.
  6. 6.
    Önal AA. Review: current analytical methods for the determination of biogenic amines in foods. Food Chem. 2007;  https://doi.org/10.1016/j.foodchem.2006.08.028.
  7. 7.
    Marcobal A, Polo MC, Martín-Alvarez PJ, Moreno-Arribas MV. Biogenic amine content of red Spanish wines: comparison of a direct ELISA and an HPLC method for the determination of histamine in wines. Food Res Int. 2005;  https://doi.org/10.1016/j.foodres.2004.10.008.
  8. 8.
    Luo L, ZL X, Yang JY, Xiao ZL, Li YJ, Beier RC, et al. Synthesis of novel haptens and development of an enzyme-linked immunosorbent assay for quantification of histamine in foods. J Agric Food Chem. 2014;  https://doi.org/10.1021/jf504689x.
  9. 9.
    Bidmanova S, Steiner MS, Stepan M, Vymazalova K, Michael A, Gruber M, et al. Enzyme-based test strips for visual or photographic detection and quantitation of gaseous sulfur mustard. Anal Chem. 2016;  https://doi.org/10.1021/acs.analchem.6b01272.
  10. 10.
    Khairy GM, Azab HA, El-Korashy SA, Steiner MS, Duerkop A. Validation of a fluorescence sensor Microtiterplate for biogenic amines in meat and cheese. J Fluoresc. 2016;  https://doi.org/10.1007/s10895-016-1885-1.
  11. 11.
    Huang X, Li Z, Zou X, Shi J, Mao H, Zhao J, et al. Detection of meat-borne trimethylamine based on nanoporous colorimetric sensor arrays. Food Chem. 2016;  https://doi.org/10.1016/j.foodchem.2015.11.041.
  12. 12.
    Diaz YJ, Page ZA, Knight AS, Treat NJ, Hemmer JR, Hawker CJ, et al. A versatile and highly selective colorimetric sensor for the detection of amines. Chem Eur J. 2017;  https://doi.org/10.1002/chem.201700368.
  13. 13.
    Schaude C, Meindl C, Fröhlich E, Attard J, Mohr GJ. Developing a sensor layer for the optical detection of amines during food spoilage. Talanta. 2017;  https://doi.org/10.1016/j.talanta.2017.04.029.
  14. 14.
    Nedeljko P, Turel M, Lobnik A. Hybrid sol-gel based sensor layers for optical determination of biogenic amines. Sens Actuat B. 2017;  https://doi.org/10.1016/j.snb.2017.02.011.
  15. 15.
    Sutarlie L, Yang KL. Colorimetric responses of transparent polymers doped with metal phthalocyanine for detecting vaporous amines. Sens Actuat B. 2008;  https://doi.org/10.1016/j.snb.2008.07.011.
  16. 16.
    Roales J, Pedrosa JM, Guillén MG, Lopes-Costa T, Pinto SMA, Calvete MJF, et al. Optical detection of amine vapors using ZnTriad porphyrin thin films. Sens Actuat B. 2015;  https://doi.org/10.1016/j.snb.2014.12.080.
  17. 17.
    Wöllner K, Vollprecht M, Leopold N, Kasper M, Busche S, Gauglitz G. Interaction behaviour of a PDMS-calixarene system and polar analytes characterised by microcalorimetry and spectroscopic methods. Anal Bioanal Chem. 2007;  https://doi.org/10.1007/s00216-007-1600-9.
  18. 18.
    Rawat KA, Bhamore JR, Singhal RK, Kailas SK. Microwave assisted synthesis of tyrosine protected gold nanoparticles for dual (colorimetric and fluorimetric) detection of spermine and spermidine in biological samples. Biosens Bioelectron. 2017;  https://doi.org/10.1016/j.bios.2016.07.069.
  19. 19.
    Chopra S, Singh A, Venugopalan P, Singh N, Kaur N. Organic nanoparticles for visual detection of Spermidine and Spermine in vapors and aqueous phase. ACS Sustain Chem Eng. 2016;  https://doi.org/10.1021/acssuschemeng.6b01295.
  20. 20.
    El-Nour KMA, Salam ETA, Soliman HM, Orabi AS. Gold nanoparticles as a direct and rapid sensor for sensitive analytical detection of biogenic amines. Nanoscale Res Lett. 2017;  https://doi.org/10.1186/s11671-017-2014-z.
  21. 21.
    Steiner MS, Meier RJ, Duerkop A, Wolfbeis OS. Chromogenic sensing of biogenic amines using a chameleon probe and the red–green–blue readout of digital camera images. Anal Chem. 2010;  https://doi.org/10.1021/ac102029j.
  22. 22.
    Matlock-Colangelo L, Baeumner AJ. Biologically inspired Nanofibers for use in translational bioanalytical systems. Ann Rev Anal Chem. 2014;  https://doi.org/10.1146/annurev-anchem-071213-020035.
  23. 23.
    Mercante LA, Scagion VP, Migliorini FL, Mattoso LHC, Correa DS. Electrospinning-based (bio)sensors for food and agricultural applications: a review. Trends Anal Chem. 2017;  https://doi.org/10.1016/j.trac.2017.04.004.
  24. 24.
    Azab HA, El-Korashy SA, Anwar ZM, Khairy GM, Duerkop A. Reactivity of a luminescent “off-on” pyrylium dye towards various classes of amines and its use in a fluorescence sensor microtiter plate for environmental samples. J Photochem Photobiol A. 2012;  https://doi.org/10.1016/j.jphotochem.2012.05.029.
  25. 25.
    Wetzl BK, Yarmoluk SM, Craig DB, Wolfbeis OS. Chameleon labels for staining and quantifying proteins. Angew Chem Int Ed. 2004;  https://doi.org/10.1002/anie.200460508.
  26. 26.
    Steiner MS, Meier RJ, Spangler C, Duerkop A, Wolfbeis OS. Determination of biogenic amines by capillary electrophoresis using a chameleon-type of fluorescent stain. Microchim Acta. 2009;  https://doi.org/10.1007/s00604-009-0247-y.
  27. 27.
    Caro B, Guen-Robin FL, Salmain M, Jaouen G. 4-Benchrotrenyl pyrylium salts as protein organometallic labelling reagents. Tetrahedron. 2000;  https://doi.org/10.1016/S0040-4020(99)00947-3.
  28. 28.
    AOAC. AOAC Official Methods of Analysis. 16th ed. Washington; AOAC, 1995. Method 35.1.32.Google Scholar
  29. 29.
    Stadnik J, Dolatowski ZJ. Biogenic amines in meat and fermented meat products. Acta Sci Pol Technol Aliment. 2010;9:251–63.Google Scholar
  30. 30.
    Azab HA, El-Korashy SA, Anwar ZM, Khairy GM, Steiner MS, Duerkop A. High-throughput sensor microtiter plate for determination of biogenic amines in sea food using fluorescence or eye-vision. Analyst. 2011;  https://doi.org/10.1039/C1AN15049A.
  31. 31.
    Alonso-Lomillo MA, Dominguez-Renedo O, Matos P, Arcos-Martinez MJ. Disposable biosensors for determination of biogenic amines. Anal Chim Acta. 2010;  https://doi.org/10.1016/j.aca.2010.03.012.
  32. 32.
    Schaude C, Meindl C, Fröhlich E, Attard J, Mohr GJ. Developing a sensor layer for the optical detection of amines during food spoilage. Talanta [Internet]. 2017;170(February):481–7. Available from: http://linkinghub.elsevier.com/retrieve/pii/S0039914017304551 CrossRefGoogle Scholar
  33. 33.
    Wang Z, Liu F, Evolution LC. Of biogenic amine concentrations in foods through their induced chemiluminescence inactivation of layered double hydroxide nanosheet colloids. Biosens Bioelectron. 2014;  https://doi.org/10.1016/j.bios.2014.04.013.
  34. 34.
    Omanovic-Miklicanin E, Valzacchi S. Development of new chemiluminescence biosensors for determination of biogenic amines in meat. Food Chem. 2017;  https://doi.org/10.1016/j.foodchem.2017.05.031.
  35. 35.
    Tseng S, Li S, Yi S, Sun AY, Gao D, Wan D. Food quality monitor: paper-based plasmonic sensors prepared through reversal nanoimprinting for rapid detection of biogenic amine odorants. ACS Appl Mater Interfaces. 2017;  https://doi.org/10.1021/acsami.7b00115.

Copyright information

© Springer-Verlag GmbH Germany 2017

Authors and Affiliations

  1. 1.Institute of Analytical Chemistry, Chemo- and BiosensorsUniversity of RegensburgRegensburgGermany
  2. 2.Institute of ChemistryNational Research Saratov State UniversitySaratovRussian Federation

Personalised recommendations