Volatile Organic Compounds as a Diagnostic Tool for Detecting Microbial Contamination in Fresh Agricultural Products: Mechanism of Action and Analytical Techniques

Abstract

Volatile organic compounds (VOCs) are secondary metabolites emitted by all living carbon-based organisms. These VOCs are of great importance in the agricultural sector due to their use as biofungicides and biopesticides. In addition, they can also be used as indicators of microbial contamination. The latter has rarely been studied; however, such a role is very relevant because it allows the timely application of corrective treatments that avoid food waste, the development of toxins dangerous to humans, and the design of biosensors. Gas chromatography–mass spectrometry (GC-MS), electronic nose (e-nose), and proton transfer reaction mass spectrometry (PTR-MS) are some of the techniques used to detect VOCs in fruits and vegetables contaminated by microorganisms. Therefore, the objective of this work is to deepen our knowledge of VOCs emitted by microorganisms in terms of their use as an indicator of microbial contamination of fresh agricultural products, as well as the analytical techniques used for their detection.

1. Introduction

According to the Environmental Protection Agency [1], VOCs are molecules with high vapor pressure and low water solubility. They are emitted as gases from certain solids or liquids distributed throughout the environment. VOCs may have short- and long-term adverse human health effects such as altered systolic blood pressure and endothelial dysfunction for acrolein, cancer risk for chloroform, styrene, trichloroethylene, and ethylbenzene, asthma for dimethyl disulfide, ethyl isobutyrate, and ethyl 2-methyl butyrate, and impaired lung function for 3-methyl-1-butanol, 2-methyl-1-butanol, isobutanol, and 1-octen-3-ol [2,3,4]. Large traffic volumes, vehicle engine combustion, industrial processes and exhausts, the decomposition of biological and non-biological waste, and the application of solvent varieties, pesticides, and fungicides all contribute to the release of toxic VOCs, among them benzene, toluene, ethylbenzene, and xylene. On the other hand, other groups of VOCs are produced by plants and emitted into the atmosphere, acting as long-distance modes of cellular communication in living organisms [5,6]. Some of these VOCs do not negatively affect health, have biotechnological applications, and can be used in the food industry to fabricate active packaging combined with matrix polymerics and microcapsules. In addition, such VOCs are sprayed on agricultural products to control microbial growth, toxins, pests, and fungi and are used as spoilage indicators. Depending on the species and the threat, plants can release terpenes, phenylpropanoids, and fatty acids, among others [7]; in turn, these compounds also act as communication devices to warn neighboring plants of danger, stimulating the synthesis of enzymes, phytoalexins, and other defensive substances [8].

Furthermore, VOCs can improve crop yield, soil health, and plant growth and enhance plant resistance against pests and pathogens. For example, Agisha et al. [9] reported the antimicrobial activity of Pseudomonas putida for Phytophthora capsici, Pythium myriotylum, Rhizoctonia solani, C. gloeosporioides, Athelia rolfsii, Gibberella moniliformis, Magnaporthe oryzae, Ralstonia pseudosolanacearum, and Radopholus similis by releasing different pyrazine derivatives, while Taniguchi et al. [10] observed that jasmonic acid induces the accumulation of β-cyclocitral, an apocarotenoid that triggers a defense response against Xanthomonas oryzae in rice plants.

The study of volatile organic compounds (VOCs) is a topic of interest for both the scientific community and society. The Web of Science reports 29,611 manuscripts relating to research carried out in countries such as China (9072 manuscripts), USA (6090 manuscripts), Italy (1747 manuscripts), South Korea (1547 manuscripts), and Germany (1493 manuscripts), amongst others. The main research fields that produce information about VOCs are the environmental sciences (29.9%), environmental and chemical engineering (24.8%), meteorology atmospheric sciences (9.8%), food science technology (3.9%), and agronomy (0.9%). Specifically, in food science technology, 75 manuscripts studied VOCs’ profile as contamination indicators [11].

Fresh agricultural products are often stored for a long time after harvest. They are easily subject to attack, mainly by fungi and bacteria, which causes their spoilage and loss of quality. These microorganisms obtain energy for survival and reproduce through catabolic activities, which release volatile and non-volatile metabolites in low concentrations, making them difficult to detect. Consequently, the analysis of VOCs contributes to our understanding of microorganisms’ metabolic status and serves as an indicator of microbial contamination [12]. Currently, the most popular and sophisticated techniques for detecting the VOCs of microorganisms are gas chromatography-mass spectroscopy (GC-MS), followed by electronic-nose (e-nose) sensors, proton transfer reaction mass spectrometry (PTR-MS), and biosensors [13]. The analysis of quality, differentiation in terms of ripening and cultivars, and the growth stage of plants and fruits are plentiful topics in the literature. This work aimed to collect information about VOCs released by agricultural products contaminated by fungi and bacteria concerning their use as indicators. It also aimed to review the main analytical techniques used for VOC measurement.

2. VOCs Used in the Diagnosis of Fungi

Both plants and fungi produce VOCs that enable them to communicate intra- and inter-specifically. By emitting VOCs, plants defend themselves against herbivores and pathogens, compete with other plants, and feed microbial populations. Fungi emit VOCs to communicate or attack each other to settle on other plants without being detected [14]. These VOCs damage the membranes of the host plant, cell wall, and DNA, modify the intracellular redox balance, and increase the loss of electrolytes, reactive oxygen species levels, and lipid membrane peroxidation. For example, 3-methyl-1-butanol and 2-methyl-1-butanol are responsible for membrane lipid peroxidation. Capric acid disrupts the cell membrane, leading to intracellular content leakage. 1-octen-3-ol causes damage in the DNA without mutagenic and clastogenic effects. Finally, decanoic acid reduces intracellular pH and leads to cellular ATP exhaustion [15]. When the host plant’s cellular structures, cuticle, enzymes, and defense metabolites are weakened, the fungi can enter, colonize, and spread [16]. Guo et al. [17] measured the emission of VOCs at the beginning of the exponential hyphal growth phase of 43 fungal species, using the 15 best predictors to describe each fungal group with an accuracy of more than 80%. They also found that Zygomycota phyla were characterized by the intense emission of the multicomponent volatile profile, Basidiomycota phyla by 10 different compounds, and Ascomycota phyla by relatively strong emission of only 2 compounds (m/z 93.092 is not an identified compound and 4-ethyl resorcinol).

Overall, VOCs are synthesized from different pathways, such as mevalonate, shikimate, the polyketide biosynthetic, and the fatty acid-derived oxylipin pathway, and have distinct odors. Some fungal VOCs referred to as phytotoxins negatively affect the functioning of plants, such as their growth in the presence of 1-octen-3-ol, trans-2-octenal, 1-hexanol, 3-octanone, 3-methyl-1-butanol, and 2-phenyl ethanol. They also inhibit root and cotyledon leaf length due to the effect of 1-octen-3-ol and trans-2-octenal [18]. Conversely, some other fungal VOCs contribute to desirable flavors in the food industry, are identical to produce for the plants, contribute to our understanding of the metabolic status of microorganisms, and potentially can be used as indirect indicators of decay, contamination, and the presence of disease in agricultural products due to alterations in flavors and odors, act as biofilters for use in degrading volatile contaminants, and mediate interactions with other microorganisms in the rhizosphere [19]. Specifically, in the fruit postharvest stage, fungal colonization occurs in three phases: penetration, incubation, and symptom appearance. During the symptom appearance, three lesion grades can be observed on the fruit: the early stage (when the development of the lesion is barely visible), the middle stage (evidenced by lesion expansion and the production of a small number of hyphae on the lesion), and the late stage (when the lesion continues to expand and apparent hyphae and spores are produced on the lesion), which mediate the release of VOCs [20]. Fungal colonization of agricultural products occurs because fungi accelerate the ripening process by promoting ethylene release and increasing the respiratory rate and the enzymatic activity of amylases, lipases, esterases, and proteases. This activity breaks down the fatty acids, amino acids, and sugars, causing their rot. Peris et al. [21] stated that Penicillium fungi secrete approximately 50 enzymes involved in the cell wall degradation of fruit and vegetables, causing the softening of the products and the release of VOCs such as ethers, alcohols, ketones, monoterpenes hydrocarbons, and esters. They also observed that in the ripening stage, orange fruit releases very high levels of monoterpene hydrocarbons through the oil glands in the exocarp, favoring their colonization by Penicillium. This could occur because the fungal growth accelerates the tissue’s deterioration, releasing the fruit’s seed and guaranteeing the species’ preservation.

Monitoring fungal VOCs in environmental and laboratory conditions presents many challenges because of their enormous heterogeneity between strains and species, low concentrations, and innate evaporative properties. Inamdar et al. [22] reported that each species of fungus emits a unique VOC profile that changes qualitatively and quantitatively depending on the age of the fungal colony, the sampling period, nutrient availability, water availability, type of substrate, pH, and temperature, amongst others. Based on the evidence, VOCs act as indicators of fungal contamination and are valuable tools in food sensory analysis, including with regard to fruits and vegetables.

The use of fungal VOCs on a commercial scale is limited because, in most works, authors developed their research in vitro using commercial media (solid or liquid), which do not represent the final plant host, nor are there reports on the profile of VOCs in different stages of the fungus’s growth. The conditions in which quantification and sample collection are carried out are not standardized, and determinations are at the laboratory level. Consequently, these are some of the challenges to be faced. Some contributions on the topic are summarized in

Table 1. VOCs as microbial contamination indicators in agricultural products and in vitro assays.

VOCs	Microorganism	Results	Reference
Ethyl acetate	Fusarium oxysporum, and F. proliferatum were incubated in a liquid onion extract medium at 20 °C for 48 h.	Ethyl acetate increased from 0.02 to 0.79 nmol L−1 in F. oxysporum when the DNA content increased from 0.09 to 1.30 mg. Ethanol, ethyl formate, 2-methyl-1-propanol, methyl thioacetate, n-propyl acetate, and 3-methyl-1-butanol were identified.	[23]
Farmesene, ethyl acetate, butanoic acid ethyl ester, cardinene.	Coletotrichum fragariae was incubated in strawberry (Fragaria × ananassa) stored at 10 °C for 6 days.	The VOCs emitted by the inoculated and non-inoculated strawberries were mainly esters and terpenes.	[24]
Menthol and menthone.	Colletotrichum sp. was incubated in açaí pulp (Euterpe oleracea).	1-propanol, 3-methyl-1-butanol, α-phellandrene, 2-nonanone, 2-phenylethanol, α-cedrene, and α-selinene, were identified.	[25]
2-ethyl hexanol, 1,2,6-hexanetriol, methyl benzoate, 2-methyl-1-butanol, and benzoyl bromide.	Alternaria alternata was incubated in tomato juice (Solanum lycopersicum L.).	The VOCs that indicated fungal growth were 2-ethyl hexanol, 1,2,6-hexanetriol, and methyl benzoate for breaker tomato juice and 2-methyl-1-butanol and benzoyl bromide for red tomato juice.	[26]
α,β-dimethyl-benzenepropanoic acid	Aspergillus flavus was incubated in-shell pistachios (Pistacia vera L.) stored for 4 d at 30 °C.	Trans-2-(2-hydroxy benzyl)-cyclohexanol, methyl formate, cyclohexanone, 2-(hydroxy phenylmethyl), 5-cyano-1-pentene, 1,3-Dioxolane, 2-methyl-2-(4-methyl-3-methylene-pentyl) and o-isopropyl hydroxylamine were detected.	[27]
1-butanol, 3-methyl-, and Cystofilobasidium.	Cystofilobasidium, Vishniacozyma, Aspergillus, and Mucor were incubated in wild mushrooms.	1-butanol, 3-methyl- was positively correlated with Cystofilobasidium (r = 0.96), Vishniacozyma (r = 0.93), Aspergillus (r = 0.93) and Mucor (r = 0.07).	[28]
alpha-guajene.	Penicillium digitatum was incubated in citrus fruit (Citrus unshiu)	Different VOCs were identified, such as 12 for stem-end rot fruit, 9 for green mold rot fruit, and 7 for blue mold rot fruit.	[29]
2,3-butanediol diacetate.	Listeria monocytogenes was incubated in melon cubes (Cucumis melo L.).	2-methyl-1-butanol, propyl cyclopropane, 2,3-butanediol diacetate, benzoic acid, ethyl ester, and 2-methyl-propanoic acid 3-phenyl propyl ester correlated with bacteria growth (<1 log CFU g−1).	[30]
Indole, nonan-2-one, tridecan-2-one, and undecan-2-one.	Escherichia coli O157:H7 was incubated in leaf lysates of spinach and rocket.	VOCs such as indole, phenyl methanol, ethanol, propan-1-ol, decan-1-ol, tridecan-1-ol, nonan-2-one, and tridecan-2-one contributed most to the separation between contaminated and non-contaminated samples.	[31]
HS VOCs not identified with a m/z of 65, 91, 92, 117, 118 and 119.	Escherichia coli O157:H7 was incubated in mediums containing meat extract, vegetable extract, and apple (Malus domestica) extract.	Six peaks were observed after 4 h growth of E. coli O157:H7. The secondary electrospray ionization-mass spectrometry (SESI-MS) technique was to distinguish EC O157:H7 from other non-O157.	[32]
Heneicosane, methyl alcohol, hydrazine carboxamide, and hydrazine.	Erwinia carotovora was incubated in red onions (Allium cepa L.).	During the initial phase of infection, ester and sulfur were at 25 °C, alkane and ketone groups were at 15 °C, and alcohol or aldehyde groups were at 8 °C.	[33]
Aniline and p-toluidine.	Gram-positive and Gram-negative bacteria were incubated in brain heart infusion broth.	Two enzymes were used as substrates, 2-amino-N-phenylpropanamide and 2-amino-N-(4-methylphenyl) propanamide, which liberated the VOCs aniline and p-toluidine, respectively, in the presence of the bacteria.	[34]
Nonanoic acid for enterobacterales and 4-(2,6,6-trimethyl- 2-cyclohexen-1-yl)-2-butanone for Pantoea.	Enterobacterales and Pantoea were incubated in green chili pepper (Capsicum annuum L.)	140 VOCs were determined in fresh-cut chili peppers, of which 22 potential spoilage markers were screened through PLS-DA and VIP analysis.	[35]

.

3. VOCs Used as Indicators of Bacterial Growth

Many bacteria are pathogens to humans but not to agricultural products. The reasons why bacteria produce VOCs are unclear; however, some reports indicate that they favor communication with the host plant and among species and play a critical role in the completion of bacterial pathogenesis by affecting bacterial virulence, promoting stress resistance, acting as signaling agents to warn the host plant encouraging it to modulate its response, have antifungal activity because bacteria often grow slower than fungi, and take part in biofilm formation [36]. According to Büttner and Bonas [37], bacteria are established in the host cell by surface proteins (adhesins), assembled into pilus-like structures, or anchored in the outer membrane. These organism opportunists can invade the horticultural products if cells are disrupted. They produce diverse VOCs to avoid fungal growth and destabilize cellular structures when it occurs. Some of these VOCs are the nonan-2-one, undecan-2-one, butanone, hexan-2-one, octan-2-one, butan-1-ol, propan-1-ol, heptan-1-ol, nonan-1-ol, pentadecan-1-ol, 2-phenyl ethanol, methyl mercaptoethanol, pyrazines, trimethylamine, 2-methyl propylamine, 2-methyl butylamine, 3-methyl butylamine, cyclohexylamine, 2-phenylethylamine, geosmin, and 2-methyl isoborneol, among others [38]. In addition, it has been reported that hexadecane is particularly abundant in cyanobacteria, and long-chain aliphatic alcohols such as 1-octanol, 1-decanol, and 1-dodecanol are produced through the β- and α-oxidation of fatty acid and commonly associated with Enterobacteriaceae. Proteobacteria and Firmicutes produce short-chain alcohols (i.e., 2,3-butanediol) under low-oxygen conditions [36].

Factors such as the density of bacterial growth in the substrate, the bacteria genre, and the substrate type affect the volatilome of agricultural products. With regard to the density of bacterial growth, Sousa et al. [31] observed that the concentration of VOCs is mainly unchanged until the number of bacteria reaches values above the infectious dose (10⁶–10⁸ CFU mL⁻¹). On the other hand, according to Chen et al. [39], the growth time and genre of bacteria could be correlated with the VOCs released. The researchers evaluated the VOCs of five types of bacteria that grew for 16 h in trypticase soy broth at 37 °C with agitation at 150 rpm min⁻¹. They found correlations c.a. 1 between bacteria and a group of VOCs. For example, 1-octanol, 1-decanol, 1-dodecanol, and indole correlated with E. coli O157:H7, 2-undecanone, 1-octanol, 1-decanol, and 1-dodecanol with S. enteritidis and S. flexneri, and 3-hydroxy-2-butanone with L. monocytogenes. Similarly, Gonçalves et al. [40] found differences between Salmonella Typhimurium, Escherichia coli O157:H7, Pseudomonas fluorescens, Pseudomonas aeruginosa, and Staphylococcus aureus using electronic nose (e-nose) sensors based on thin films of an ionogel composite doped with different concentrations of iron oxide. The interaction between VOCs and sensors increased ionogel conductivity. Variations related to each bacterium’s volatile metabolism were noted, leading to bacteria differentiation based on the relative response of each sample.

Additionally, the VOC profile of bacteria is affected by substrate type and growth conditions. Zareian et al. [41] showed a clear difference in the volatilome of Pseudomonas grown in a glucose medium (at 0.5% and 1%) and an egg white protein-enriched medium (0% and 2%). Alcohols, aldehydes, and ketones were mainly detected in the medium enriched with glucose. Conversely, sulfur derivatives such as methanethiol (m/z 49), dimethyl disulfide (m/z 95), and a different array of cyclic-based VOCs such as benzene (m/z 79) were released by bacteria grown in a medium including egg white protein. The variation in VOCs released may be because the ideal carbon sources for Pseudomonas are organic acids and amino acids rather than glucose since it does not essentially affect cyclic adenosine monophosphate that acts as signal transduction levels in Pseudomonas.

On the other hand, Sangiorgio et al. [42] reported a change in the volatilome of four cultivars of raspberry fruit (‘Imara’, ‘Regina’, ‘Anne’, and ‘Enrosadira’) due to the effect of the bacteria established naturally and recolonized after the disinfection process. In the first group, the fruits were naturally colonized by Bacilli, Bacillales, Paenibacilaceae, Enterobacterales, and Rosseberrgiella. In contrast, the second group formed by recolonized fruits had Gammaproteobacteria, Enterobacterales, and Enterobacteriaceae in more than 50% of their microbiome. This explains why, in fruits recolonized by bacteria, it was observed that the alkenes and monoterpenes diminished. At the same time, acids, alcohols, lactones, aldehydes, esters, ketones, and norisoprenoids were unaffected compared to naturally infected ones. After combining the VOCs of four cultivars of raspberry, seventeen compounds were present in all the genotypes analyzed. These were produced by at least one bacterium belonging to the raspberry microbiome, except for γ-terpinene, o-cymene, β-ionone, dihydro-α-ionone, and dihydro-β-ionone. The emitted VOCs could then be used to detect bacteria in raspberry fruits; in fact, they could improve the aroma of the fruit. Some relevant contributions on the topic are summarized in Table 1.

4. Current Analytical Techniques

Due to their low concentration, functional groups, polarity, and physical and chemical properties, VOCs are challenging to detect without precise analytical equipment [12,30]. Gas chromatography and mass spectroscopy (GC-MS) are widely used to analyze a sample’s volatilome based on its polarity and boiling point. Subsequently, these VOCs can be identified and quantified using mass spectra with authentic standards, chemical library spectra, and chromatographic retention indices, as reported by Martinez and Bennett [43].

Table 2. Analytical techniques used to detect phytopathogens through VOCs.

Analytical Technique	Microorganism	Result	Reference
GC-MS and e-nose.	Botrytis sp., Penicillium sp., and Rhizopus sp.	Strawberry was analyzed by GC-MS, which found five compounds common among fungi: hexyl acetate, methyl hexanoate, ethyl hexanoate, hexanoic acid, hexyl ester, and trans-nerolidol. E-nose could detect strawberry fruits’ decay on the second storage day.	[44]
GC-MS and e-nose.	Botrytis cinerea, Monilinia fructicola, and Rhizopus stolonifer.	Changes in VOCs of inoculated peaches were correlated with total amounts and species of fungi. Terpenes and aromatic compounds were the main contributors to e-nose responses. This technique had high discrimination accuracy, among fungal contamination in peaches.	[45]
Headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS).	Candidatus Liberibacter spp.	Orange leaves were inoculated with Diaphorina citri, the vector responsible for Huanglongbing disease. The VOCs generated were analyzed by HS-GC-IMS, reporting 36 compounds released by the leaves during infection, limonene being the main one.	[46]
Selected ion flow tube mass spectrometry (SIFT-MS).	Botrytis cinerea	SIFT-MS was used to detect 1-octen-3-ol in strawberries, which was reported as a biomarker of fungi growth. Mainly alcohols, aldehydes, and terpenes associated with the ions H3O+ 071, H3O+ 111, H3O+ 113, NO+ 063, NO+ 111, NO+ 127, NO+ 141, O2+ 057, O2+ 084, and O2+ 093 were identified.	[47]
E-nose and HS-SPME/GC-MS.	Botryosphaeria dothidea and Dothiorella gregaria.	Ethyl butanoate, ethyl hexanoate, isobutyl butyrate, benzyl benzoate, ethyl benzoate, phenyl ethyl alcohol, and styrene accounted for 93% of the total VOCs when the fruit was severely affected by soft rot disease in kiwifruit. DVB/CAR/PDMS fiber was used for extracting, and the combination of e-nose and GC-MS was feasible for the discrimination of intact and diseased kiwifruits.	[48]
GC-MS	Acidovorax citrulli	Ethyl acetate, 2,3-butanediol, 2-methyl-1-pentanol, 3-methyl-1-pentanol, phenol, 6-methyl-5-hepten-2-one, benzyl alcohol, p-cresol, 2-nonanone, and phenyl ethyl alcohol were detected in watermelon fruit. The sample was recovered after 10 d of inoculation with A. citrulli bacteria. VOCs were extracted by headspace-solid phase microextraction (HS-SPME)	[49]

shows examples of the main analytical techniques used to quantify VOCs emitted by fungi and bacteria in agricultural products.

According to Clarivate [11], as a tool to detect microbial contamination in agricultural products, GC-MS is used in disciplines such as food science technology, applied chemistry, microbiology, agriculture, entomology, and biotechnology. For example, Guo et al. [50] analyzed fresh grapes contaminated with Aspergillus carbonarius using HS-SPME-GC-MS. They found that ochratoxins produced by the fungi can be detected from VOCs such as ethyl acetate (p < 0.01), styrene (p < 0.05), and 1-hexanol (p < 0.05), and VOCs such as (E)-2-hexenal (p < 0.05) and nerolic acid (p < 0.05) are negatively correlated. Similarly, Yu et al. [51] used GC-MS to analyze the volatilome of pathogenic and non-pathogenic E. coli strains to identify if strawberry is a suitable host and to detect if strawberries absorbed the bacterial VOCs. They found that the pathogenic and non-pathogenic E. coli O157:H7 emitted a profile of volatile compounds comprising predominantly indole and methyl ketones. Strawberry was a suitable host for E. coli O157:H7, which either proliferated at relatively low inoculum levels (10⁵ cells mL⁻¹) or remained at a stable population at higher inoculum levels (10⁶–10⁷ cells mL⁻¹). Finally, in terms of inoculating strawberry fruit with E. coli O157-H7 did not detect amounts of indole in headspace vapor samples, showing that strawberries absorbed relatively large quantities of indole and other bacterial volatile compounds from the vapor phase. The behavior observed in the strawberry showed that the absorbed VOCs contribute to controlling microbial growth and consequently act as a technological alternative to antagonist microorganisms.

Another technology is e-nose sensors, which mimic human olfactory perception. These electronic sensors can detect the VOCs in a sample. They incorporate an information-processing unit, pattern recognition software, and reference library databases. The resultant electronic fingerprints express a unique aroma that helps detect odor profiles without separating the mixture into its components [52]. Cellini et al. [53] used the e-nose technique (EOS507C equipment) to distinguish between in vitro apple plants infected by bacteria (Erwinia amylovora, Pseudomonas syringae pv. syringae, and mock) after 0, 2, and 8 days of storage. Only the samples infected by P. syringae on day 8 differed from the remaining samples in terms of odor profiling, even on dormant, asymptomatic material in real conditions. In addition, postharvest sapota contamination by foodborne pathogens has been evaluated by Ezhilan et al. [54]. After analyzing the data, the authors found that the e-nose can separate fresh sapota from contaminated samples and determine the levels of deterioration. It showed that fruits stored on days 1 and 2 displayed fresh sapota, while days 3, 4, and 5 showed half-contaminated sapota, with completely contaminated sapota on day 6.

Finally, proton transfer reaction mass spectrometry (PTR-MS) is an alternative method to GC-MS. It is based on soft chemical ionization through proton transfer reactions. Protonated water generation occurs from pure water’s ion source (a hollow cathode). The reactions are observed in a drift tube operated at low pressure (c.a. 2 mbar) between H₃O⁺ primary ion and VOCs fed directly and continuously into the drift tube from the gas sample. The VOCs with proton affinity higher than water (691 kJ/mol) are protonated, while the components of air (N₂, O₂, and CO₂) are not because of their lower proton affinity, thus allowing the use of air as a carrier gas [55]. The PTR-MS technique detects VOCs in parts per billion (ppb). Researchers in the fields of environmental science, food technology, and medical diagnosis have reported its effectiveness. For instance, in the PTR-MS technique with O₂ ion in operation mode, ethylene was detected in apples, nectarines, apricots fruits, Arundo donax leaves, and Pseudomonas syringae pv. actinidiae bacterium at concentrations as low as one ppb. It has the limitation of a lack of specificity, especially in the presence of complex matrices when other compounds could interfere with the signals for ethylene [56]. Overall, PTR-MS is a powerful tool for evaluating the quality, stage of ripening, and differentiation among cultivars of fruits and vegetables, using their VOCs as indicators rather than for detecting microbial contamination.

Before the analysis, the VOCs are extracted from the sample using two methods. The first method requires solvents, and, in some cases, the sample is heated. It is very efficient for extracting VOCs but can also provoke compound degradation and cause the volatilome profile to be wrong. The second method involves the use of solvent-free techniques. In terms of this method, the “Headspace (HS)” is an alternative procedure known to be fast, universal, sensitive, solvent-free, and economical. It is widely used to trap VOCs in agricultural products. The HS of a sample can be recovered with the use of a tight syringe and the solid-phase micro-extraction (SPME) procedure. Notably, SPME is recognized as a solvent-free procedure that reduces artifact formation and is suitable for recovering VOCs. It is based on the concentration of VOCs onto a fiber coated with the appropriate stationary phase, which is transported to the detector and subsequently desorbed into the hot GC injection port. Equilibrium between the amount of analyte in the sample matrix, in the headspace above the sample, and the stationary phase coated on a fused silica fiber is established in the case of headspace solid-phase microextraction (HS-SPME). The quantity of compounds absorbed depends on the compound partitioning between the HS and the agricultural product matrix, the partitioning between the fiber coating and the HS, and the coating phases in SPME. These coating phases in SPME are of vital importance; PA fiber and DVB fiber are better when polar compounds are predominant in the samples, and PDMS or CAR/PDMS fibers are more suitable for extracting non-polar compounds. Finally, CAR/PDMS or DVB/CARB/PDMS fiber coatings can extract polar and non-polar compounds [4]. Figure 1 summarizes the release processes and analysis of VOCs in the case of horticultural products.

5. Biosensors Developed for Detecting the VOCs of Phytopathogens

The early detection of phytopathogens present in food is of utmost importance to prevent disease and avoid crop losses. It is estimated that between 20% and 40% of crop losses are due to these organisms. In some cases, they can be so severe that they cause total production to be lost [57]. Traditionally, the detection of phytopathogens is carried out by isolation and visual examination techniques based on manuals; however, these approaches are time-consuming and require the use of trained personnel [58].

There are also emerging techniques for detecting phytopathogens that are more specific. These can be categorized into two methods: direct and indirect. The former refers to polymerase chain reaction (PCR) techniques and immunological methods that, despite their precision, are time-consuming and expensive since they require highly sophisticated equipment [59]. Indirect methods use biological indicators for detection in real-time since they perceive physiological and chemical changes in plants during infection or due to stress. These methods use imaging or compound detection techniques such as infrared spectroscopy, fluorescence, X-rays, thermography, and gas chromatography [60]. In this sense, new research uses techniques based on biological indicators to manufacture devices for detecting phytopathogens that meet the speed and ease of use and are of low cost. Such is the case with biosensors. In addition, they have the advantage of being automatable with multi-analysis capacity and portable; moreover, they are highly selective [61].

A biosensor is a highly accurate and compact device that detects an analyte of interest. It includes a bioreceptor formed by a nucleic acid, enzyme, protein, antibody, or cell. It is integrated into a transduction system that detects the signal emitted by the interaction between the bioreceptor and the analyte of interest. This interaction causes changes in the physicochemical properties of the matrix, such as pH, color, and temperature. The transducer perceives and converts these changes into a signal with a known parameter [62,63].

In the agricultural field, biosensors have been developed to detect microorganisms. Regiart et al. [64] designed a biosensor to detect Xanthomonas arboricola in walnut plants by immobilizing a monoclonal antibody on functionalized mesoporous silica SBA15 through covalent bonding. Through the enzymatic interaction between alkaline phosphatase and p-aminophenyl phosphate, the electrical current proportional to the level of X. arboricola in plant samples was measured. Likewise, Lucas-Bautista et al. [65] developed a device capable of detecting the presence of Colletotrichum gloeosporioides in papaya fruits by monitoring the chitin produced during the infection process. However, little research has been reported on detecting VOCs in the interaction between the pathogen and the plant host.

Table 3. Biosensors to detect phytopathogens through VOCs.

Volatile Biomarker	Biosensor	Result	Reference
Carbon dioxide.	The biosensor prototype is based on CuO nanoparticles and nanolayers.	The carbon dioxide released by Aspergillus niger was detected by a biosensor prototype based on layers of CuO nanoparticles deposited on a glass substrate with silver adhesive, onto which the fungus was placed, measuring the resonance with an amperemeter. As the growth rate of the fungus increased, resistance decreased.	[66]
p-ethylphenol	Enzymatic biosensor to detect Phytophthora cactorum in strawberry fruit.	The biosensor contained a carbon nanotube electrode modified with tyrosine, whose interaction with p-ethylphenol resulted in cyclic voltammetry and constant potential amperometry. The biosensor showed a high sensitivity of 4.0 ± 0.5 μA cm−2 μM−1 by constant potential amperometry with a detection limit of 0.10 ± 0.02 μM and a quantification limit of 0.29 ± 0.07 μM.	[67]
Methyl salicylate.	Amperometric biosensor	The biosensor used alcohol oxidase and horseradish peroxidase as receptors. They were placed on a glassy carbon electrode-modified surface with multi-walled carbon nanotubes. The interaction resulted in an amperometric signal proportional to methyl salicylate released by Tetranychus urticae, Fusarium, and Phytophthora capsici during infection. The sensitivity of the biosensor was 112.37 and 282.82 μA cm−2 mM−1 by cyclic amperometry and constant potential amperometry with a detection limit of 2.95 μM and 0.98 μM, respectively.	[68]
Methyl salicylate.	Amperometric biosensor	Esterase, salicylate hydroxylase, and tyrosinase were used as biomarkers for the detection of methyl salicylate in a biosensor based on the reduction in 1,2-benzoquinone to catechol by the action of the enzymes. The resulting reaction was measured amperometrically. The biosensor exhibited a sensitivity of 1.5 μA cm−2 and with a detection limit of 0.7 μM.	[57]
Ethylhexanol, linalool, tetradecane, and phenylacetaldehyde	Biosensor based on single-walled carbon nanotubes functionalized with single-stranded DNA	Detection of Huanglongbing disease in citrus trees through compounds in the asymptomatic stage: ethyl hexanol, linalool, tetradecene, and phenylacetaldehyde. The biosensor managed to detect the compounds in a saturation range between 5 and 100%, having a resistance error in terms of its reproducibility of less than 10%.	[69]
Set of VOCs	Optical biosensor	Penicillium digitatum growth on oranges was detected using an optical biosensor that captured the luminescence emitted by the bacteria due to the changes in VOCs and concentration after infection. The biosensor was able to detect the luminescence of the bacteria on the third day of infection before the first visible symptoms appeared on the fruit.	[70]

shows the relevant works on this.

6. Conclusions

Fresh products emit VOCs in response to compositional changes, environmental conditions, and microbial contamination. Fungi and bacteria are the main causal agents of the contamination and the deterioration of agricultural products, with fungi being the most predominant. These use different modes of action to establish themselves on their host. Some authors have reported one or several volatile biomarkers for detecting microorganisms, while others prefer to associate them with a functional group based on its variability, depending on the substrate and the host. In all cases, the concentration of these VOCs is very low (in the order of ppb), so it is necessary to use precise analytical techniques such as GC-MS and PTR-MS. The e-nose technique has more practical applications because it is based on odor profile detection with sensors facilitating and reducing analysis time. Consequently, its development and application are increasing, and it could replace sensory studies carried out by panels of trained humans. In addition, it is important to consider the sampling method, with solvent-free techniques such as HS and sampling with SPME fiber being increasingly widely accepted. The biosensors also represent an option to detect pathogens by using nucleic acid, enzymes, proteins, antibodies, or cells integrated into a transduction system. However, more research on this topic needs to be conducted in order to obtain more accurate detection methods.