Role of Microbial Fermentation in the Bio-Production of Food Aroma Compounds from Vegetable Waste

Abstract

Flavour is a key driver of consumer preferences and acceptability of foods, and the food industry has made food aroma compounds a crucial area of research. At present, about 80% of food aroma compounds are produced by chemical synthesis; however, alternative production approaches have been explored to meet consumers’ demand for “clean label” food products and “natural” aromas. Bio-production of food aroma compounds from vegetable wastes through fermentation has emerged as a promising alternative. This review showed that fungi and yeasts, and also lactic acid bacteria, can be used to produce aroma compounds through the fermentation of vegetable waste. The produced compounds were mostly responsible for sweet, fruity, and floral notes. Other molecules imparting cheesy/buttery, creamy, green, herbal, grass notes were also obtained through the fermentation of vegetable food waste. Substrates varied from agricultural waste such as rice bran to by-products and waste from the fruit supply chain, in particular pomace, peels, pods. During the study, challenges and limitations for the scale-up of the process emerged. The production of aromas is still strongly strain and waste dependent. Certain aspects thus still require attention to avoid that a joint occurrence of technical challenges may cause the failure of the process.

1. Introduction

Flavour is a key driver of consumer preferences and acceptability of foods, and sensory science has worked to gain a broader understanding of it, from the generation in food to the human perception during eating [1].

At physiological level, food flavour is affected by the combined and synergistic effects of olfaction, taste, and chemestetic sensations, as well as by the cognitive processing of these inputs by the brain. Aroma and taste are considered to play a prominent role in the multi-modal phenomenon of food flavour perception [2]. Taste sensations are activated by non-volatile, hydrophilic molecules (e.g., soluble sugars, organic acids, triterpenes, etc.), which interact with the receptors located on the tongue and impart five basic sensations: sweet, sour, salty, bitter, and umami [3]. The olfaction sensations are triggered by complex mixtures of volatile compounds (VOCs), belonging to several chemical classes, which are detected by receptors located in the olfactory epithelium inside the nasal cavity. Odors can reach the olfactory epithelium via two pathways: the ortho-nasal and retro-nasal [4]. VOCs are organic molecules with a low molecular weight (≈400 Da) and mostly have a lipophilic character [5]. They can be classified, based on their chemical structure, into several classes: acids, esters, aldehydes, ketones, terpenes, lactones, hydrocarbons, and others. Food aroma compounds are related to specific sensory descriptors (Figure 1).

The role played by food aroma compounds in consumers’ food choices and preferences has made them a crucial area of research for the food industry. Innovation and marketing priorities, technological issues, and food acceptability are the main drivers. It is, in fact, crucial for the food industry to develop new products which have appealing flavours, or to launch new products which are, nevertheless, different to those in the same category which are already available on the market. Food aroma compounds can also change dramatically, depending on agronomical (i.e., ripening, senescence, and decay of fruit and vegetables) and technological factors (e.g., food processing and storage) [3,5]. For this reason, it is necessary to compensate for the loss by adding food aroma compounds. The food flavour market is expected to reach 19.72 billion USD in 2026, with a 5% increase in the annual growth rate with respect to 2018 [6]. Industrial aroma compounds currently find a huge application in the food manufacturing industry, from bakery products (e.g., cookies, pastries, muffins, cakes, etc.) and dairy products (e.g., yoghurts, ice creams, desserts, spreadable creams, etc.), to confectionery (e.g., filled chocolate, chocolate bars, mixes for hot chocolates, etc.), table sauces and seasonings (e.g., flavoured oils, flavoured vinegars, glazes, sauces), and snack seasonings (e.g., chips, crisps, etc.).

When dealing with food aroma compounds, two aspects are of paramount importance: their “natural” identity and their “toxicity”. The concentration of aroma compounds present in natural sources is much lower than in processed food and other products. For this reason, it is crucial that flavouring ingredients are present in food products in concentrations below their toxic level, which is signified by the no-observed-adverse-effect level (NOAEL) [7]. Toxicological data and intended use levels are evaluated by regulatory bodies, such as the European Food Safety Authority (EFSA) in Europe or the U.S. Food and Drug Administration (FDA), that provide approval prior to use. Fruity esters are considered to have a low toxicity; however, they can either irritate the mucosal surfaces if inhaled at concentrations within the range of 50–400 ppm, or can cause skin damage and fatigue, and respiratory irritation and dyspnoea at concentrations above 1000 ppm [7]. Benzaldehyde and cinnamaldehyde, which are the main ingredients in confectionaries and soft drinks with almond and cherry flavours, present a relatively low toxicity; although, at a concentration higher than 500 ppm, skin sensitization and intoxication might occur [7]. Cinnamaldehyde results in severe irritation of the respiratory airways and coughing at concentrations of 100,000 ppm [7]. Vanillin is considered to be a low-toxic compound; however, at high concentrations, it can irritate the eyes and mucous membranes of the respiratory tract, and can also accelerate bile secretion [8]. Vanillin intake via food and beverages is so low that people do not exceed the acceptable daily intake (ADI) of vanillin, which is recommended as being 0–10 mg/kg by the FAO and the World Health Organization [8]. Limonene, which is a key ingredient in citrus flavours, presents no serious toxicity risk in rats, and the NOAEL is up to 250–500 mg/kg/d [7]. Maltol and ethyl maltol are aroma compounds with caramel and sugar-like flavours; they are components of the aroma of freshly baked bread and are therefore used as flavour enhancers in bakery products. Ethyl maltol is less toxic than maltol, for which growth inhibition and kidney damage are reported in rats when concentrations over 1000 mg/kg/d are provided [7].

At present, about 80% of food aroma compounds are produced by chemical synthesis, especially from petroleum-derived feedstocks, such as benzene, xylene, toluene, etc. [9,10]. However, to cope with the current market and social issues, such as the consumers’ increasing interest in “clean label” food products [11] and the threat of environmental issues deriving from the excessive exploitation of fossil fuels, more sustainable production methods of creating food aroma compounds and “natural” compounds have been explored. Within this framework, the bio-production of food aroma compounds via fermentative processes has emerged as an appealing approach/strategy.

Fermentation is carried out by microorganisms which can degrade substrates rich in sugars into secondary metabolites, and this biological process is generally characterized by low costs and low energy consumption. In addition, agri-food waste has increasingly ended up as a valuable resource from which it is possible to recover, at low- or zero-costs, not only compost, livestock feed, energy, biofuels, and biomaterials, but also valuable compounds, such as bioactive compounds, aroma compounds, etc. [12]. The food supply chain produces tons of wastage volumes, depending on the chain stage (i.e., primary production, processing, distribution, etc.) and the food commodity group. Plant-based foods produce the highest percentage of loss and waste. For example, fruit and vegetables produce peelings, pomace, seeds, leaves, trimmings, stones, and stems. Grains produce straw, pods, husks, hulls, skins, bran, germ, broken grains [12]. These vegetable materials are not yet exhausted, and are still rich in nutrients, such as sugars. They can, therefore, become a substrate for fermentative processes.

At present, literature reviews on the analysis and discussion of the bioproduction of aroma compounds through the fermentation of agri-food waste are available [13,14,15,16,17,18,19]. However, these reviews report on the findings of studies carried out from the late years of the last century to the first two decades of the twenty-first century. The aim of this study is, therefore, to provide an up-to-date overview of the latest studies on the bio-production of food aroma compounds via the fermentation of vegetable waste. To this end, a systematic review will be carried out to retrieve only the latest papers. The production of the most common food aroma compounds through fermentation will be analysed, and the most commonly used strains and the applied process conditions will be reported. Limitations and future perspectives will be also discussed.

2. Materials and Methods

Adopted Strategy for Literature Search: Methodological Approach and Selection Criteria

A search of peer-reviewed literature was carried out through the SCOPUS database, in December 2023. A total of 10 queries (

Table 1. List of queries used for the search on SCOPUS.

Search (No.)	Scopus Queries	Results (No. of Documents)
#1	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (waste))	41
#2	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (peeling))	1
#3	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (pomace))	14
#4	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (seed))	49
#5	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (straw))	4
#6	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (bran))	3
#7	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (bagasse))	3
#8	(TITLE-ABS-KEY (aroma AND compound) AND TITLE-ABS-KEY (fermentation) AND TITLE-ABS-KEY (by-products))	49
#9	(TITLE-ABS-KEY (food and waste) AND TITLE-ABS-KEY (solid-state AND fermentation) AND TITLE-ABS-KEY (aroma OR aroma AND compounds))	5
#10	(TITLE-ABS-KEY (submerged AND fermentation) AND TITLE-ABS-KEY (aroma AND compounds))	11
	Total	189

), containing keywords related to the topic under investigation, were used.

For the search, a date range was applied and only papers published from 2020 to present were selected. The application of the filter met the criteria of selecting only up-to-date literature, which represents the novelty of the paper compared to other reviews currently available in the literature.

The search was carried out independently by two authors, who applied the same queries and adopted the same methodological approach. The details (e.g., Author(s), Title, Year, Journal, Volume, Issue, etc.) of the documents retrieved with each query were copied and pasted into an excel spreadsheet, in order to identify and exclude duplicates. Documents were then screened on the basis of title and abstract. As a final step, full texts were analysed: non-accessible papers were excluded, while accessible and relevant studies were deemed as being eligible and were included in the analysis. Possible discrepancies were analysed by all authors and inconsistencies were solved.

3. Results and Discussion

3.1. PRISMA Flow Diagram of the Literature Search

The flow diagram of the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) reported in Figure 2 shows that 189 papers, containing the keywords specified in Table 1 in the title and/or abstract, were published from 2020 to the present. A total of 44 papers were duplicates and were removed from the list. First, the titles and abstracts of the remaining 145 documents were screened and it emerged that the scope of 113 papers fell beyond the scope of this review. A total of seven papers were not accessible, that is, they were published in Journals to which the authors have no access; fifteen more papers were excluded because the analysis of the full text showed they were not relevant to the analysis and discussion of this review. A total of ten papers were finally selected for analysis in the present review (Figure 2).

3.2. Microbial Processes in the Bio-Production of Aroma Compounds: Focus on Solid-State Fermentation and Submerged Fermentation

Microbial processes have recently established themselves as an alternative strategy for the synthesis of natural aroma compounds. The traditional strategies for the production of natural aroma compounds, namely their extraction from plant materials by steam distillation followed by fractional distillation techniques, have, in fact, some drawbacks, such as high costs and poor yield [20]. Compared to chemical synthesis, which is often an environmentally unfriendly process that uses heavy metal catalysts and reagents from limited sources, the bio-production of aroma compounds also presents some advantages for the transition to a more sustainable and circular economy. It uses cheap and easily accessible starting materials and/or substrates which are available from natural and renewable resources and is carried out under mild reaction conditions.

Microbiological methods for the synthesis of natural aroma compounds comprise (i) de novo synthesis and (ii) biotransformation (Figure 3).

The de novo synthesis pertains to the production of aroma compounds from building-block molecules, such as carbohydrates, fats, and proteins, which are converted by microorganisms through a metabolic pathway [21]. In this approach, strains are selected to produce significant amounts of targeted aroma compounds. However, the quantity is also dependent on the culture medium used in the process [20]. It is possible for GRAS microorganisms to be used. This approach commonly enables the production of high amounts of primary metabolites, while aroma compounds are produced at low concentrations. Moreover, complex mixtures of several aroma compounds are obtained [20].

Biotransformation consists of the chemical modification of a precursor into targeted aroma compounds. It is based on single or multistep reactions catalysed by cells or enzymes [21]. Compared to enzymes, cells generally possess the redox system necessary for the synthesis of aroma compounds, as well as the system which enables the regeneration of cofactors. However, this strategy shows some drawbacks: sterile conditions are necessary in all steps and substrates or by-products that may inhibit the cells must be removed [20]. Biotransformation can be obtained by two approaches: solid-state fermentation (SSF) and submerged fermentation (SmF).

3.2.1. Solid-State Fermentation in the Bio-Production of Aroma Compounds

Solid-state fermentation refers to a fermentation process in which microorganisms (bacteria, fungi, and yeasts) grow on a moist, solid, non-soluble organic material without the presence of a free liquid. The material serves as a support and nutrient source [22]. The substrate may be a natural substrate such as agro-industrial residues and agricultural crops or an inert support [20]. When applied for aroma compound production, the natural substrate provides nutrients, including carbon and nitrogen sources, for the growth of microorganisms and the synthesis of metabolites. When the structure of the substrate is not suitable for cell growth, an inert support can be used to enable the microorganisms to properly access their substrate and co-substrate [23]. Cellulose sponge and luffa sponge have been used as inert supports in the production of γ-decalactones by SSF [23].

In SSF, the microorganisms are in close contact with the insoluble substrate and the highest nutrient concentration from the substrate for fermentation is achieved.

Most of the microorganisms used in SSF are GRAS, hence products which are free from toxins and safe for human consumption can be obtained [22]. Bacteria, yeast, and fungi can be used in SSF. Among the bacteria, Bacillus megaterium, Bacillus mycoides, and Lactobacillus spp., including L. acidophilus, L. bulgaricus, L. plantarum, L. rhamnosus, L. delbrueckii, and L. coryniformis, are employed. Yeasts applied in SSF comprise Saccharomyces cerevisiae, Saccharomyces boulardii, and Candida sp. For example, S. cerevisiae strains have been used in the SSF of orange peel to obtain isoamyl acetate, phenyl ethyl acetate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, ethyl dodecanoate [24]. Actinobacteria species, such as Streptomyces thermonitrificans and Streptomyces chattanoogensis, are also employed. Filamentous fungi of the Aspergillus, Fusarium, Penicillium, Rhizopus, and Trichoderma genera have mostly been used in SSF [22,23]. Fungi and yeasts are, in fact, the most suitable microorganisms for SSF because they require only minimal water activity within the substrate [23].

SSF consists of an upstream process followed by a midstream and a downstream process. In the upstream process, substrates and growth media are prepared, and the microorganisms to be used for the fermentation are isolated. In the midstream process, substrates are inoculated, and fermentation takes place. In the downstream process, products are extracted and purified [22]. The recovery of aroma compounds produced by SSF is a challenging step since they are commonly hydrophobic and highly volatile compounds that may be lost if they are released in the headspace, mainly when an aeration flux is applied within the process [22].

The factors affecting the production of aroma compounds by SSF can be biological and physicochemical. The former include the composition of the solid matrix, the selected microorganisms, and the inoculum. The latter encompass moisture content, water activity, temperature, pH, particle size, and aeration.

With regard to the design of the bioreactors used for the production of aroma compounds by SSF, lab-scale, pre-pilot, and pilot-scale bioreactors have been designed [23].

Some of the literature has demonstrated that aroma production through SSF process results in higher yields and in less time compared to SmF processes [23].

3.2.2. Submerged Fermentation in the Bio-Production of Aroma Compounds

Submerged fermentation is a process in which the breakdown of a substrate is undertaken by microorganisms in the presence of plenty of free water. SmF is generally applied in the presence of a moisture content ranging between 80 and 95% [14]. Microbial cultures are inoculated directly into the liquid medium for the production of the desired product(s).

Several operational modes are possible: batch, fed-batch, and continuous [25]. Batch fermentations are versatile and easy to handle, while continuous fermentation is used for the production of enzymes [25]. In industrial production, microorganisms are grown in bioreactors containing nutrient broths.

SmF can be easily scaled up thanks to the ease of automation and does not present the disadvantage of limited heat and mass transfer, which is rather the main drawback of SSF. Due to continuous stirring mechanisms, nutrients, oxygen, and microbial cells are evenly distributed throughout the medium. In addition, online control of major reaction parameters (e.g., oxygen supply in case of aerobic microorganisms, pH, temperature, viscosity, dissolved oxygen, foam formation, aeration, and the formation of desired products) is easily possible [26].

High production costs, complexity of the medium, and low productivity are limitations of the SmF processes. The application of SmF for the production of flavour compounds has generally resulted in low productivity [20]. This has, therefore, limited industrial applications [20].

3.3. Evidence of Aroma Compounds Bio-Production from Agri-Food Waste

The studies which were eligible for the analysis provided an overview of the latest investigations into the production of bio-aromas through fermentative processes. As shown in

Table 2. Principal aroma compounds produced by fermentation of vegetable agri-food waste.

Odor Descriptor	Main Aroma Compounds	Substrate	Strain	Fermentative Process	Reference
Main notes
Sweet	vanillin	banana peels	Enterobacter hormaechei KT385666	SmF	[27]
	vanillin	pomegranate peels	E. hormaechei KT385666	SmF	[28]
	vanillin, guaiacol, hexanoic acid	vanilla pods	LAB co-cultures	lactic acid fermentation	[29]
	vanillin	sugarcane bagasse	E. hormaechei KT385666	SSF	[30]
	3-phenylpropanol, methyl cinnamate	cajá residues	Auriporia aurulenta sp.	SmF	[31]
	γ-octalactone	umbu residues	A. aurulenta sp.	SmF	[31]
	phenylmethanol	apple pomace	Brettanomyces bruxellensis CCT 3469	SmF	[32]
	ethyl dodecanoate	orange pomace	B. bruxellensis CCT 3469	SmF	[32]
	ethanol	rice bran	Rhizopus oligosporus sp.	SSF	[33]
Fruity	2-methyl-1-propanol, 3-methyl-1-butanol, 2-methylpropyl acetate, isoamyl acetate, 2-methylbutanol acetate, (Z)-3-hexenyl acetate, 2-phenylethanol, 2-phenethyl acetate	umbu residues	A. aurulenta sp.	SmF	[31]
	benzaldehyde, hexyl acetate, benzyl alcohol	cajá residues	A. aurulenta sp.	SmF	[31]
	myrtenyl acetate	persimmon residues	A. aurulenta sp.	SmF	[31]
	prenyl acetate, 1-octen-3-ol	plum residues	A. aurulenta sp.	SmF	[31]
	ethyl acetate	apple pomace	B. bruxellensis CCT 3469	SmF	[32]
	3-methylbutyl acetate	apple pomace	B. bruxellensis CCT 3469	SmF	[32]
	ethyl nonanoate	apple pomace	B. bruxellensis CCT 3469	SmF	[32]
	ethyl 2-methylbutanoate	orange pomace	B. bruxellensis CCT 3469	SmF	[32]
	ethyl heptanoate	orange pomace	B. bruxellensis CCT 3469	SmF	[32]
	methyl hexanoate	carrot pomace	B. bruxellensis CCT 3469	SmF	[32]
	3-methylbutyl hexanoate	carrot pomace	B. bruxellensis CCT 3469	SmF	[32]
	methyl pentanoate	carrot pomace	B. bruxellensis CCT 3469	SmF	[32]
	butanoic acid-ethyl ester and butanoic acid, 2-methyl-, ethyl ester	melon by-products	Lacticaseibacillus rhamnosus 1473	-	[34]
Floral	2-Phenylethanol	sugarcane bagasse	Pichia kudriavzevii sp.	batch-SSF	[35]
	2-Phenylethanol	nine agro-industrial wastes (supplemented with l-phenylalanine)	Pichia kudriavzevii CECT 13184.	SSF	[36]
	2-Phenylethyl acetate	apple/orange pomace	B. bruxellensis CCT 3469	SmF	[32]
	Phenylethyl alcohol	orange pomace	B. bruxellensis CCT 3469	SmF	[32]
	Phenylethyl alcohol	carrot pomace	B. bruxellensis CCT 3469	SmF	[32]
	octyl acetate, cinnamil acetate	cajá residues	A. aurulenta sp.	SmF	[31]
	(cis)-rose oxide, β-linalool, 1-nonanol, citronellyl formate, trans- and cis-geraniol, nerolidol	orange pomace	L. rhamnosus 1473	SSF	[34]
	β-damascenone	melon by-products	L. rhamnosus 1473	SSF	[34]
Other notes
Cheesy/buttery	butanoic acid	rice bran	R. oligosporus	SSF	[33]
Citrus	(E)-2-octenol, 1-octanol	persimmon	A. aurulenta sp.	SmF	[31]
	limonene, citral, and valencene	orange pomace	L. rhamnosus 1473	SSF	[34]
Citrus/Lemon	D-limonene	rice bran	R. oligosporus sp.	SSF	[33]
Creamy	2,3-butanediol	rice bran	R. oligosporus sp.	SSF	[33]
Fatty	ethyl oleat, 9,12-octadecadienoic acid (Z,Z), ethyl linoleate	rice bran	R. oligosporus sp.	SSF	[33]
Floral/Green	(3S)-7-hydroxy-3,7-dimethyloctanal	apple pomace	B. bruxellensis CCT 3469	SmF	[32]
Grass	hexanal	rice bran	R. oligosporus sp.	SSF	[33]
Green	2-hexenal	melon by-products	L. rhamnosus 1473	SSF	[34]
Herbal	1-hexanol	orange pomace	L. rhamnosus 1473	SSF	[34]
Mouldy	3-octanone, 3-octanol	cajá residues	A. aurulenta sp.	SmF	[31]
Spicy	γ-terpinene and α-terpineol, 4-terpineol, carveol, cis-carveol, and eugenol	orange pomace	L. rhamnosus 1473	SSF	[34]
Sweet/fruity	ethyl decanoate	apple pomace	B. bruxellensis CCT 3469	SmF	[32]
	ethyl hexanoate	orange pomace	B. bruxellensis CCT 3469	SmF	[32]
Waxy/fatty	1-octanol, (E)-2-decenal	melon by-products	L. rhamnosus 1473	SSF	[34]

SmF: submerged fermentation; SSF: solid-state fermentation; LAB: lactic acid bacteria.

, the aroma compounds produced by fermentation were mostly responsible for sweet, fruity, and floral notes. Other molecules imparting cheesy/buttery, creamy, green, herbal, and grass notes were also obtained. Substrates varied from agricultural waste such as rice bran, to by-products and residues of the fruit supply chain, in particular pomace, peels, pods. Lactic acid bacteria, fungi, and yeasts were used in the different fermentative processes, ranging from submerged to solid-state.

3.3.1. Bio-Production of Aroma Compounds with a Sweet Note

Sweet notes and fragrances are generally imparted by different compounds, including lactones, aldehydes, furanones, and pyranones. Lactones are cyclic esters and potent aroma compounds. γ-Lactones, such as γ-octalactone and γ-decalactone, are the molecules responsible for peachy, creamy, and coconut aromas [37]. Furanones also impart sweet aromas and have very low thresholds. 2,5-Dimethyl-4-hydroxy-3[2H]furanone (furaneol) is likely the most ubiquitous furanone molecule, and 5-ethyl-3-hydroxy-4-methyl-5H-furan-2-one has a powerful, sweet, caramel-like and maple aroma [37]. Maltol (i.e., 3-hydroxy-2-methyl-4(H)-pyran-4-one), which has a relatively high odour threshold, imparts a sweet fruity note; 2-hydroxy-3-methyl-2-cyclopenten-1-one has a strong caramel-like note [37]. Among aldehydes, the most studied aroma compound with a sweet and creamy essence is 4-hydroxy-3-methoxybenzaldehyde, mostly known as vanillin. Because of consumers’ appreciation of this aroma and following the food industry’s use of it in products, such as desserts, ice creams, pastries, confectionery, etc., vanilla’s market value is growing more and more. The global market is estimated to reach 100 million USD in 2025 [29]. The production of natural vanillin from vanilla plants, especially Vanilla planifolia, can meet less than 1% of the over 10,000 tons of vanillin demanded by the global market. To meet the growing demand, vanillin is currently produced by chemical synthesis from different sources, including guaiacol, eugenol, and coniferyl alcohol [38]. The price of synthetic vanillin is about a hundred times lower than that of natural vanillin [39].

However, its chemical synthesis presents operational and environmental issues. For example, the oxidation reaction that occurs during the industrial synthesis implies the use of complex system equipment, abundant energy consumption, inevitable waste materials, and the use of harmful compounds such as toluene and ethyl acetate [40]. These organic solvents are volatile and can generate environmental pollution [40]. Therefore, alternative approaches have been explored, to meet the consumers’ demand for “natural” and “clean” products, and to respond to the definition of “natural” flavouring substances, as laid down by Regulation (EC) No. 1334/2008 [41].

The biotransformation of different substrates, such as ferulic acid, eugenol and isoeugenol, sugars (i.e., glucose), aromatic amino acids, and phenolic stilbenes, for the bio-production of vanillin has been investigated [42]. Ferulic acid is the best-explored substrate. Ferulic acid is a phenylpropanoic acid, which is naturally occurring in plants, and vanillin is a transient intermediate of ferulic acid catabolism in several microorganisms [42]. Five different ferulic acid degradation pathways, proceeding with vanillin as an intermediate, have been identified as occurring in microorganisms (Figure 4) [42].

With regard to the production of bio-vanillin from vegetable food wastes, rice bran oil residues were fermented by Aspergillus niger sp. and/or Penicillium cinnabarium sp. [14], green coconut agro-industrial husks underwent solid state fermentation with Phanerochaete chrysosporium sp. [43] or Pycnoporus cinnabarinus sp. [44,45,46].

More recently, other by-products have been investigated for use in the bioproduction of vanillin and different fermentative conditions have been applied (

Table 3. Experimental conditions of the fermentative process applied in the bio-production of vanillin.

Fermentative Process	Substrate	Substrate Pre-Treatment	Strain(s)	Inoculum	Fermentation Conditions	Reference
SSF (plates)	Exhausted vanilla pods without seeds (10 g for each experiment)	Autoclaving (121 °C, 21 min)	▪ Leuconostoc citreum 4452 ▪ Leuconostoc mesenteroides 2194 ▪ Weissella minor 4451	7 Log CFU/g	RSM conditions: Temp: 25, 30, 35 °C Time: 30, 75, 120 h Glucose: 0, 2.5, 5%	[29]
SSF (plates)	Exhausted vanilla pods without seeds (10 g for each experiment)	Autoclaving (121 °C, 21 min)	▪ LAB co-cultures	7 Log CFU/g	RSM conditions: Temp: 25, 32 °C Time: 30 h Glucose: 0–0.45%	[29]
SSF (plates)	Exhausted vanilla pods with seeds (10 g for each experiment)	Autoclaving (121 °C, 21 min)	▪ Lacticaseibacillus rhamnosus 1473 ▪ Lacticaseibacillus paracasei 4186 ▪ Lacticaseibacillus casei 2240 ▪ L. plantarum 4932 ▪ P. acidilactici 3992	7 Log CFU/g	RSM conditions: Temp: 32, 37, 42 °C Time: 30, 75, 120 h Glucose: 0, 2.5, 5%	[29]
SSF (plates)	Exhausted vanilla pods with seeds (10 g for each experiment)	Autoclaving (121 °C, 21 min)	▪ L. citreum 4452 ▪ L. mesenteroides 2194 ▪ W. minor 4451	7 Log CFU/g	RSM conditions: Temp: 25, 30, 35 °C Time: 30, 75, 120 h Glucose: 0, 2.5, 5%	[29]
SSF (plates)	Exhausted vanilla pods with seeds (10 g for each experiment)	Autoclaving (121 °C, 21 min)	▪ LAB co-cultures	7 Log CFU/g	RSM conditions: Temp: 25, 32 °C Time: 30 h Glucose: 0%	[29]
SSF (Erlenmeyer flasks)	Sugarcane bagasse, rice straw, wheat straw, rice bran, corn cob	▪ Rinsed, chopped into small pieces, oven-dried (60 °C, 48 h), ground, and sieved to a final particle size of 1 mm. ▪ Alkaline extraction of esterified ferulic acid.	▪ E. hormaechei strain KT385666	OD: 0.6	Incubation temperature: 30 °C Incubation time: 48 h pH: 7.0	[30]
SSF (Erlenmeyer flasks)	Sugarcane bagasse	▪ Rinsed, broken into small pieces, oven-dried (60 °C, 48 h), ground, and sieved to a final particle size of 1 mm. ▪ Alkaline extraction of esterified ferulic acid.	▪ E. hormaechei strain KT385666	OD: 0.6 Volume: 2 mL	RSM conditions: Moisture content: 40, 50, 60, 70, 80% pH: 5, 6.5, 7.5, 9, 10 Inoculum: 1, 2, 3, 4, 5 mL Temp: 25, 30, 37.5, 45, 50 °C Incubation time: 12, 24, 32, 48, 60 min	[30]
SmF (shaking flasks)	banana peels	▪ Rinsed, chopped into small pieces, oven-dried (50 °C, 24 h), ground, and sieved to a final particle size of 1 mm. ▪ Extraction of esterified ferulic acid.	▪ E. hormaechei strain KT385666	OD: 0.6 Volume: 1 mL	OFAT conditions: Incubation time: 8, 16, 24, 32, 40 h pH: 5, 6, 7, 8, 9 Temp: 20, 30, 40, 50, 60 °C Agitation speed: 110, 130, 150, 170, 190 rpm	[27]
SSF (shaking flasks)	pomegranate peels	▪ Oven-dried (40 °C, 48 h), ground, and sieved to a final particle size of 1 mm. ▪ Alkaline extraction of esterified ferulic acid.	▪ E. hormaechei strain KT385666	OD: 0.6 Volume: 2 mL	RSM conditions: Ferulic acid concentration: 0.2–1.2% Incubation time: 8–56 h pH: 5–10 Temp: 20–50 °C Agitation speed: 100–200 rpm	[28]

Temp: temperature; OD: optical density; RSM: Response Surface Methodology; OFAT: One Factor At a Time.

).

Hadj Saadoun et al. explored the possibility of fermenting some residues from the vanilla production chain itself [29]. As a matter of fact, the ripened beans from which vanilla is extracted are not completely exhausted pods. They present, in fact, the potential for being repurposed and valorised, because they still contain different amounts of vanillin, trapped in the cellular structures of the plant.

In detail, Hadj Saadoun et al. fermented exhausted vanilla pods, either containing seeds or without seeds, with different lactic acid bacteria strains (e.g., Leuconostoc citreum 4452, Leuconostoc mesenteroides 2194, Weissella minor 4451, L. rhamnosus 1473, L. plantarum 4932, L. paracasei 4186, L. citreum 4452, P. acidilactici 3992, and co-cultures thereof) which have an optimal growth temperature of 30 or 37 °C, depending on the strain [29]. They were used as inoculums for the two substrates under different fermentative conditions, in terms of temperature, incubation time, and glucose percentage in the growth medium (Table 3), as defined by the Design of Experiment tool.

The elaboration of data by Response Surface Methodology (RSM) pointed out that the best fermentation conditions were the minimum levels of all selected conditions. The different LAB strains showed a different behaviour in the two substrates, except for L. rhamnosus 1473 and L. casei 2240 which showed good adaptability to both substrates [29]. However, the major difference observed following the fermentation of exhausted vanilla beans with and without seeds was not represented by the microbial growth, but by the aromatic notes produced and the volatile profile. The fermentation of pods without seeds determined aroma compounds with beany, woody, and phenolic notes, while pods with seeds allowed for the production of sweeter aromas, with vanillin, fatty, pudding notes [29]. Vanillin was not the main aroma compound produced during the fermentation; guaiacol was generally the most abundant compound, followed by hexanoic acid and then vanillin. The production of these compounds was not affected by the LAB strain used, except for the co-cultures which generally determined the highest concentrations [29]. It is possible that the accumulation of these compounds derived from an interaction between the metabolic pathways of the strains, as confirmed by the higher microbial load observed in these samples at the end of fermentation.

The solid-state fermentation of other agricultural wastes, such as sugarcane bagasse, rice straw, wheat straw, rice bran, and corn cob, by the Enterobacter hormaechei strain KT385666, has also been investigated for the production of vanillin [30]. The experiments designed by Mehmood et al. were based on the assumption that vanillin is a temporary intermediate of ferulic acid catabolism in many bacteria, and ferulic acid is found in large amounts in the cell walls of several agricultural crops, including wheat, maize, and sugar beet. Sugarcane bagasse, rice straw, wheat straw, rice bran, and corn cob were thus dried at 60 °C for 48 h, before the alkaline extraction of esterified ferulic acid [30]. A preliminary SSF process carried out with all substrates in Erlenmeyer flasks incubated at 30 °C for 48 h (Table 3) showed that biovanillin production was highest in sugarcane bagasse (0.029 g/100 g), deriving from the fact that bagasse had the highest concentration of ferulic acid (0.94 g/100 g), followed by wheat and rice straw [30]. Upon optimization of the fermentative conditions of sugarcane bagasse by E. hormaechei strain KT385666 through RSM, it emerged that the highest production of biovanillin was obtained when the substrate’s moisture content was 70%, the inoculum volume was 4 mL, and the process conditions were 48 h for the incubation time, 7.5 for pH, and 37.5 °C for temperature [30]. The medium incubation time was related to the fact that vanillin can become toxic for microorganisms when they are exposed for long periods of time and at high concentrations; the incubation time must, therefore, be adequate, so as to have a non-detrimental effect on the production. The temperature was not high because the E. hormaechei strain is mesophilic, and enzymes denature at high temperatures, thus resulting in reduced production of vanillin. The pH of the medium is also a critical aspect in fermentation, and the experiments by Mehmood at al. showed that a pH ranging between 5 and 10 was the most suitable for the process.

Biovanillin production was also obtained from the submerged fermentation of banana peels with the same strain of E. hormaechei KT385666 [27]. After the extraction of ferulic acid from properly treated banana peels (Table 3), it emerged that this vegetable waste is also a cheap source of ferulic acid (60 ± 0.35 mg/100 g of banana peels). The ferulic acid extracted from banana peels was sterilized and added to the fermentation medium in different concentrations (0.3, 0.5, 0.7, 0.9, and 1.1%), in Erlenmeyer flasks, to optimize the substrate concentration for the hyper-production of biovanillin. The optimization, using a One-Factor-At-a-Time (OFAT) approach, of the conditions for the production of biovanillin (i.e., incubation time, pH, temperature, agitation speed) from the ferulic acid crystals showed that the highest production of biovanillin (3.8 ± 0.14 g/L), upon biotransformation of 7 g/L of ferulic acid, occurred at an incubation time of 24 h, pH 7, 30 °C incubation temperature, and a 150 rpm agitation speed.

E. hormaechei strain KT385666 was further used to investigate the potential of pomegranate peels as a source of ferulic acid for biotransformation into biovanillin via submerged fermentation [28]. The pomegranate peels were found to be a rich source of ferulic acid (162.5 ± 0.35 mg/100 g) and the optimization of the fermentation parameters through RSM (Table 3) allowed for a slightly higher production of biovanillin (4.2 g/L) than banana peels. Compared to the fermentation of banana peels, the maximum production of biovanillin was obtained in a shorter incubation time (8 h) and a lower agitation speed (100 rpm) and at the same pH (pH 7) and temperature (30 °C) [28]. Other molecules responsible for sweet notes were also obtained in combination with other aroma compounds from the fermentation of (i) cajá and umbu residues with the fungus Auriporia aurulenta sp. [31], (ii) apple and orange pomace by Brettanomyces bruxellensis CCT 3469 [32], and (iii) rice bran with a strain of Rhizopus oligosporous sp. [33].

3.3.2. Bio-Production of Aroma Compounds with a Fruity Note

Fruity aromas are in high demand in the food industry because of continuous efforts to enhance food taste or diversify food and beverage’s fruity flavour profile.

Esters are fundamental compounds for fruity notes. Ethyl butanoate is, for instance, responsible for the characteristic strawberry aroma; ethyl hexanoate provides the characteristic note of fresh pineapple; 3-methylbutyl acetate is characteristic of pear; and C9 esters are important for a melon aroma [37].

Aldehydes have a fruity or floral aroma with a fresh note [37]. Acetaldehyde imparts fruity ether notes; C3–C5 aldehydes, such as pentanal, butanal, and propanal, tend to have a quite malty/green note [37]. C6 aldehydes (e.g., hexanal) provide a green note, such as the fresh green aroma of green apples. Aldehydes with a chain length longer than C6, have both fruity/floral and fatty descriptors, depending on the concentration [37].

Terpenes, terpenoids, and sesquiterpenes are responsible for the characteristic aroma profile of citrus fruits, herbs, and spices: limonene has a weak orangey citrus-peel aroma, α-valencene imparts classic citrus notes, pinene, myrcene, and ocimene are the major components of a basil aroma [37].

With regard to the bio-production of fruity aromas, this group of compounds has been produced by fermenting citrus pulp, coffee husks, cassava bagasse, and apple pomace with strains of Ceratocystis fimbriata sp., and/or by using wheat bran and cassava and sugarcane bagasse as a substrate of Rhizopus oryzae sp. strains [14].

More recently, other by-products have been investigated for the bioproduction of aroma compounds with fruity notes, and different fermentative conditions have been applied (

Table 4. Experimental conditions of the fermentative process applied in the bio-production of aroma compounds with fruity notes.

Fermentative Process	Substrate	Substrate Pre-Treatment	Strain(s)	Inoculum	Fermentation Conditions	Reference
SmF (shaking flask experiments)	Powder of umbu, cajá, plum, and persimmon waste (6.25 g) + SNL minimum culture medium (125 mL)	Oven-dried (35 °C) until moisture content <5%, ground and sieved (20-mesh sieve)	▪ Auriporia aurulenta	25 mL	Agitation speed: 150 rpm Incubation time: 7 days Temperature: 24 °C In the dark	[31]
SmF (shaking flask experiments)	apple, orange, and carrot pomaces (moisture content > 75%, acid pH)	wet substrate (150 g) was dissolved in 350 mL sterile, distilled water; sterilized by autoclavation (121 °C, 20 min)	▪ B. bruxellensis CCT 3469	OD600: 10 Volume: 10 mL Suspended in saline medium.	Incubation time: 72 h Temperature: 30 °C Agitation speed: 200 rpm	[32]

).

Agro-industrial residues from the processing of umbu (Spondias tuberosa L.), cajá (Spondias mombin L.), plum (Prunus domestica L.), and persimmon (Diospyros kaki L.) fruits have been used as a substrate for a strain of the basidiomycete fungus Auriporia aurulenta sp. in a submerged fermentation process [31].

Basidiomycetes have been largely applied for production of aroma compounds in recent decades. Benzaldehyde was produced by strains of Ischnoderma benzoinum sp., Ischnoderma resinosum sp., and Bjerkandera adusta sp.; the terpene nootkatone was obtained by fermentation with a strain of Pleurotus sapidus sp. [47,48]; and γ-decalactone through fermentation by Rhodotorula aurantiaca sp. and Sporidiobolus ruinenii sp. [49]. Strains belonging to this species possess, in fact, an extracellular enzymatic system, which enable them to synthesize several volatile compounds [50].

In the fermentation tests carried out by Sandes and colleagues, the treated agro-industrial residues were used in shaking flask experiments as substrates in combination with an SNL minimum culture medium, composed of glucose, asparagine monohydrate, yeast extract, potassium hydrogen phosphate, and magnesium sulphate [31]. Following the submerged fermentation of residues of umbu, cajá, plum, and persimmon by the strain of Auriporia aurulenta sp., under the process conditions specified in Table 4, a total of 46 compounds were identified [31]. The most abundant VOCs were twenty-five esters and thirteen alcohols; four terpenoids, two aldehydes, one ketone, and one lactone were also produced.

Depending on the substrate, the aroma compound was more or less abundant. For example, umbu residues allowed for the highest production of compounds related to fruity and sweet odour descriptors: 2-methyl-1-propanol, 3-methyl-1-butanol (banana odour descriptor), 2-methylpropyl acetate (banana odour descriptor), isoamyl acetate (banana odour descriptor), 2-methylbutanol acetate (banana odour descriptor), (Z)-3-hexenyl acetate (banana odour descriptor), 2-phenylethanol (2-PE) (floral odour descriptor), 2-phenethyl acetate (2-PEA) (fruity, honey, floral odour descriptor), and γ-octalactone (sweet, coconut odour descriptor). The fermentation of cajá residues allowed for the production of compounds responsible for fruity and floral odour descriptors, i.e., benzaldehyde (fruity, almond odour descriptor), hexyl acetate (fruity, banana odour descriptor), benzyl alcohol (fruity, floral, candy odour descriptor), octyl acetate (floral odour descriptor), 3-phenylpropanol (spicy, sweet), methyl cinnamate (sweet, balsamic odour descriptor), cinnamyl acetate (floral, honey odour descriptor), but also for mouldy, mushroom odours: 3-octanone and 3-octanol (mouldy/mushroom odour descriptors). The aroma compounds produced with persimmon residues as a substrate were responsible for citrus, green odour descriptors, such as (E)-2-octenol and 1-octanol, and myrtenyl acetate, which has a more fruity, sweet, herbal odour. The fermentation of plum residues allowed for the production of prenyl acetate and 1-octen-3-ol, which have a fruity and green odorous note, respectively [31]. The analysis of the obtained aroma bioproducts and substrates pointed out that the fermentation of umbu by-products allowed for the production of aroma compounds which were not detected in the fermentative processes carried out with the other fruit residues. In detail, (E)-β-damascenone, p-cymen-7-ol, and 1,4-p-menthadien-7-ol are three out of the four terpenoid compounds produced in all substrates [31]. 2-Phenethyl acetate was produced by all the residues, with the highest production occurring with the umbu residue. Since umbu residues were confirmed to be the optimal substrate for production of aroma compounds, the effect of some process parameters on the yield of 2-PE, 2-PEA, and (E)-β-damascenone after SmF of umbu residues was investigated. In detail, the kinetics of the formation of the three compounds were monitored at seven cultivation times (0, 3.5, 7.0, 10.5, 14.0, 17.5, and 21.0 days) and in two medium pH (3 and 6) [31].

Monitoring the kinetics showed that the production of 2-phenethyl acetate was highest after 3.5 days of fermentation in the acidic medium, with a yield of 11.39 mg/L; while the optimal yield (2.28 mg/L) of 2-phenylethanol was observed after 7 days in the basic medium [31]. It is possible to speculate that the A. aurulenta sp. strain preferably synthesizes, with umbu as substrate, 2-PEA in an acidic medium and the 2-PE in a basic medium. The data obtained by RSM showed that the inoculum volume differently affected the concentration of the two compounds. The highest concentration of pre-inoculum determined a reduction in the production of 2-PEA, while an increase in the pre-inoculum determined a higher concentration of 2-PE. The highest production of (E)-β-damascenone (0.25 mg/L) was obtained by fermenting umbu residues in alkaline medium for 10.5 days [31].

Submerged solid-substrate fermentations of apple, orange, and carrot pomace by the yeast Brettanomyces bruxellensis strain CCT 3469 also produced esters with a fruity note (Table 2) [32]. An aliquot of sterile wet substrate was added with sterile distilled water, and inoculated with the pre-inoculum resuspended in 10 mL saline: the 1 L volume Erlenmeyer flasks were then incubated under the conditions specified in Table 4. The fermentation occurred with and without the supplementation of L-phenylalanine (7 g/L) [32]. The production of esters with fruity notes was due to the fact that, during the fermentation, the free reducing sugars available in the apple, orange, and carrot pomaces were used. It was observed that the free reducing sugars were not completely exhausted after the 72-h fermentation, while the yeast strain’s growth stopped, likely because of the lack of other nutrients, such as nitrogen and phosphorus, or growth inhibition by metabolic products which are toxic to the yeast. It was, therefore, speculated that the fermentation duration could be potentially extended upon supplementation of the required nutrients.

Fermentation of melon by-products by Lacticaseibacillus rhamnosus strain 1473 allowed for production of several volatile compounds which had fruity notes: butanoic acid-ethyl ester and butanoic acid, 2-methyl-, ethyl ester (Table 2) [34].

3.3.3. Bio-Production of Aroma Compounds with a Floral Note

Floral aromas consist of a mixture of compounds, including alcohols, terpenoids, phenylpropanoids/benzenoids, fatty acids, and amino acids. Among them, 2-Phenylethanol (2-PE) is an alcohol with a rose-like odour and is the second most widely used flavour after vanillin [10]. It is also the precursor of other aroma compounds such as 2-phenethyl acetate (2-PEA) which also has a floral-like fragrance.

At present, the production of 2-PE and 2-PEA occurs via chemical synthesis and is based on three paths: (i) the Friedel–Craft reaction of ethylene oxide with benzene, (ii) the catalytic reduction of styrene oxide, and (iii) the oxidation of propylene with 2-phenylethyl hydroperoxide. However, these paths involve the use of petrochemical-derived sources and sometimes high-temperature conditions, which impart different off odours in the final product such as chloride-derived species or other undesirable side-products [51]. The extraction of the two aroma compounds from natural products is also deemed as being expensive because of their low content in natural sources and the dependence on external factors. The production of 2-PE and 2-PEA by biotransformation of the precursor L-phenylalanine (Ehrlich pathway) (Figure 5) or by de novo synthesis from simple sugars (Shikimate pathway) is also a possible alternative.

The bio-production of 2-PE and 2-PEA is mainly related to the application of yeasts in solid-state fermentative processes carried out according to the conditions reported in

Table 5. Experimental conditions of the fermentative process applied in the bio-production of aroma compounds with floral notes.

Fermentative Process	Substrate	Substrate Pre-Treatment	Strain(s)	Inoculum	Fermentation Conditions	Reference
SSF (glass reactors, 0.5 L volume)	Sugarcane bagasse supplemented with sugar beet molasses, and L-phe	Oven-dried (60 °C), ground, sieved to a particle size of 0.5–4.75 mm, sterilized	▪ P. kudriavzevii CECT 13184	5·107 CFU g−1	pH: 4 ML: 7.5% (w/w dry basis) L-phe content: 3.8% (w/w dry basis) SAFR: 0.13 L h−1 g−1 MC0: 76% Temperature: 31 °C Time: 40 h	[35]
SSF (glass reactors, 0.5 L volume)	Rice husk, brewer’s spent grain, soy fibre, rice fibre, green and red apple pomace, asparagus tails, orange peels, banana peels	Oven-dried (60 °C), sterilized	▪ P. kudriavzevii strain CECT 13184	5·107 CFU g−1 dry substrate	Substrate: 95 ± 1 g Temperature- controlled water bath: 30 °C Time: 96 h SAFR: 0.13 L h−1 g−1 L-phe (4% dry basis) addition or de novo synthesis without L-phe. With molasses (10%) or without molasses	[36]

M_L: sugar beet molasse; S_AFR: specific air flow rate; MC₀: initial moisture content.

.

In detail, Martínez-Avila et al. explored the possibility of producing 2-PE by fermenting sugarcane bagasse with a strain of Pichia kudriavzevii, isolated from the sugarcane bagasse itself [35]. Sugarcane bagasse, which is the fibrous material generally wasted after the extraction of sugar juice, is, in fact, rich in cellulose (40%), hemicellulose (24.4%), lignin (15%), protein (1.8%), fat and waxes, saccharose (14%), and glucose (1.4%) [52]. In addition, it is, together with sugarcane dry leaves and tops, sugarcane press mud, and molasse, one the main wastes within the sugarcane chain of production, from harvest to the final processing; bagasse represents about 30–40% (w/w) of the sugarcane [52]. Martínez-Avila et al. carried out batch tests to evaluate the ability of a Pichia kudriavzevii sp. strain to produce 2-PE through SSF. In detail, experiments were carried out in 0.5 L glass reactors, filled with 92 ± 1 g of the inoculated substrate, for a total time of 40 h (Table 5). The reactor was equipped with a mass flow controller supplying water-saturated air to the bottom; oxygen concentrations were constantly monitored. The effect of the different process parameters on the 2-PE production was evaluated. It was also observed that some process parameters deeply affected 2-PE production, such as pH and micronutrient addition, while temperature and initial moisture content resulted in being critical parameters for the process. The cultivation of the autochthonous strain at 31 °C, 76% initial moisture content and 0.129 L h⁻¹ g⁻¹ specific airflow rate allowed for the maximum 2-PE content (27.2 ± 0.2 mg per gram of dry substrate) [35].

The same strain (P. kudriavzevii strain CECT 13184) was also used for the production of 2-PE through SSF of nine agro-industrial wastes in glass reactors with mass flow control and a continuous supply of water saturated-air for a total of 96 h (Table 5) [36]. The batch-SSF tests were designed in order to evaluate the effect of L-phe addition or not (de novo synthesis or L-phe biotransformation) and the effect of the addition of a supplementary carbon source, that being sugar beet molasse. Martínez-Avila et al. found that the maximum 2-PE production was obtained when fermenting red apple pomace. A total of 1.7 and 25.2 mg 2-PE per gram of vegetable waste was obtained through de novo biosynthesis and L- phenylalanine biotransformation, respectively [36].

2-Phenylethyl acetate and phenylethyl alcohol were also produced by fermenting the pomace of apples, oranges, and carrots with B. bruxellensis CCT 3469 strain (Table 2) [32]. Other molecules imparting a floral note were also produced in combination with other aroma compounds by fermenting other vegetable residues, such as cajá residues and melon by-products (Table 2).

3.3.4. Bio-Production of Mixtures of Aroma Compounds

Although food aroma compounds have a specific sensory descriptor, which is perceived by the human nasal olfactory apparatus, they are rarely perceived in isolation and their concentration and threshold, as well as the interaction with other compounds, can significantly affect their identification [5].

The systematic search of studies in which vegetable wastes were investigated as a substrate for the bio-production of aroma compounds showed that, in some cases, admixtures of aroma compounds, responsible for diverse aromatic notes and fragrances, were obtained (Table 2).

Admixtures of aroma compounds were produced by SSF of rice bran, obtained from four rice types from Indonesia, with a strain of Rhizopus oligosporus [33]. Rice bran is, in fact, a by-product of rice milling and is generally used as an animal feed ingredient or as raw material for the production of fertilizers [53]. However, it has potential for being valorised because it is a good source of dietary fibre, vitamins, minerals, protein, lipids, and some phytochemicals such as ferulic acid, γ-oryzanol, and tocopherol [54].

In the study by Astuti et al., rice bran was properly treated before fermentation (i.e., milling, sieving, sterilization) and inoculated with 15% (v/w) of the strain of R. oligosporus (

Table 6. Experimental conditions of the fermentative process applied in the bio-production of aroma compounds with multiple aromatic notes.

Fermentative Process	Substrate	Substrate Pre-Treatment	Strain(s)	Inoculum	Fermentation Conditions	Reference
SSF (Petri dishes)	Rice bran (9 g) + distilled water (20% v/w)	Milling, grinding, sieving, autoclaving (121 °C, 15 min)	▪ R. oligosporus	106 spore/mL Volume: 15% (v/w)	Incubation time: 72 h Incubation temperature: 30 °C	[33]
SSF (flasks)	Melon and orange by products	Autoclaving (121 °C for 20 min)	▪ L. rhamnosus 1473	7 Log CFU/g	Incubation time: 72 h Incubation temperature: 37 °C	[34]

) [33]. In SSF, the solid matrix does, in fact, require pretreatments to make the physical structure more suitable for the fermentation process and the chemical constituents more accessible to microorganisms. Upon SSF, under the conditions specified in Table 6, ten alcohols, eight aldehydes, eight ketones, seven acids, seven esters, four phenols, and four terpenes, etc., were produced. Among the alcohols, ethanol and 2,3-butanediol, which are responsible for sweet and creamy notes, respectively, were the most abundant aroma compounds produced during the fermentation. Ethanol was produced upon hydrolyzation of starch to glucose via glycolysis; the formation of 2,3-butanediol was from the catabolism of glucose to pyruvate via the glycolysis pathway. From the degradation of carbohydrates, other alcohol compounds were produced, such as benzyl alcohol and phenyl ethyl alcohol which have a slightly sweet aroma [33]. Hexanal was the most abundant aldehyde produced. Surprisingly, vanilla was not detected. Acids were generally poorly produced; butanoic acid content was highest. Terpenes were also produced in very low amounts; among them, D-limonene was the most abundant. Ethyl oleate, 9,12-octadecadienoic acid (Z,Z), and ethyl linoleate, which are responsible for fatty notes, were the esters which were mostly produced.

Following the submerged solid-substrate fermentation of apple, orange, and carrot pomace by the yeast Brettanomyces bruxellensis strain CCT 3469, under the conditions specified in Table 4, admixtures of aroma compounds were obtained [32]. The highest amount of aroma compounds produced during the fermentative process was observed in carrot pomace: 160 compounds, compared to 142 compounds in orange pomace and 106 in apple pomace. Most aroma compounds were esters and had fruity, floral, or sweet notes. Phenylethyl alcohol, which is an aroma compound which has a rose-like fragrance, was only detected in the headspace of fermented carrot pomace.

Fermentation of orange and melon by products by Lacticaseibacillus rhamnosus strain 1473 allowed for the production of several volatile compounds which had floral, citrus, fruity, green, herbal, spicy, and minty notes [34]. The concentration also varied depending on the recovery method (vacuum versus simple distillation). Simple distillation is, in fact, carried out at about 100 °C, while vacuum occurs at lower temperatures; it is likely that the high temperatures in simple distillation increased the formation and recovery of aroma compounds.

Following the fermentation of orange pomace, different compounds with floral notes were obtained: e.g., (cis)-rose oxide, β-linalool, 1-nonanol, citronellyl formate, trans-and cis-geraniol, and nerolidol [34]. Molecules responsible for citrus notes were also obtained: limonene, citral, and valencene. The concentration of molecules with a fruity note (e.g., benzaldehyde, p-menth-1-en-9-ol, β-cyclocitral, ethyl 3-hydroxyhexanoate, and 2-cyclopentyl-cyclopentanone) was not affected by fermentation, while herbaceous compounds such as 1-hexanol presented a higher concentration in fermented orange pomace recovered by simple distillation. High concentrations of aroma compounds responsible for terpenic and spicy notes, such as γ-terpinene and α-terpineol, were found. It is possible that α-terpineol was produced upon degradation of D-limonene or resulted from the hydrolysis of volatile glycosidic precursors in oranges [34]. Other spicy notes are related to the production of 4-terpineol, carveol, cis-carveol, and eugenol after fermentation.

Fermentation of melon by products allowed for the production of 99 compounds, which mainly had waxy/fatty notes, followed by fruity, floral, herbal, green, etc. [34]. Unlike the orange pomace, for which different aroma compounds were found depending on the recovery method, melon-based distillates presented the same notes but at different concentrations. Molecules responsible for the waxy and fatty notes were 1-octanol and (E)-2-decenal. β-Damascenone was the predominant compound with a floral note. Butanoic acid-ethyl ester and butanoic acid, 2-methyl-, ethyl-ester were the most abundant molecules with fruity notes.

4. Strengths and Limitations of This Study

The systematic analysis of studies on bio-production of food aroma compounds from vegetable waste through fermentation and the setting of a tight time limit allowed obtaining up-to-date information on the current state of research and opportunities in this field. The novelty of the study thus lays in its methodological approach.

However, the studies reviewed in this study lacked a specific overview of the fermentative pathways and the possible toxic effects of the substrate and precursor(s) on the bio-production of aroma compounds. In addition, the paper does not provide perspective speculations on the economic feasibility of the process being scaled up, compared to the traditional methods, which would provide some insights into how competitive the bio-production of aroma compounds might be.

5. Conclusions and Future Trends for the Bioproduction of Aroma Compounds

To conclude, the studies reviewed in this paper mainly reported on the bioproduction of aroma compounds with a sweet note, followed by compounds with floral and fruity notes. This finding echoes the fact that vanillin and 2-PE represent the two flavours of major interest for the food industry and the consumer. Molecules with a citrus aromatic note have been also obtained through the fermentation of vegetable waste.

The analysis showed that SSF was the most commonly applied biotechnology. This is likely due to some operational advantages, in terms of reduced water consumption, no need for sophisticated bioreactors, low energy consumption, and a high efficiency in terms of product yields. SSF was confirmed as the technique of choice for the fermentation of agro-waste residues rich in cellulose, lignin, and hemicellulose, such as sugarcane bagasse and rice straw. The SSF processes were carried out using fungi or yeasts because of their ability to grow on the surface of a solid matrix with a low water content.

Other types of vegetable waste, including peels from bananas and pomegranates, the pomace of apples, carrots, and oranges, and residues from plums, persimmons, umbu, and cajá, were chosen as substrate in SmF. The inoculum was made from bacteria such as E. hormachei, fungi, such as A. aurulenta and R. oligosporus, yeasts, i.e., B. bruxellensis and P. kudriavzevii, and lactobacilli.

The analysis of the selected studies showed that the experiments were carried out at laboratory-scale by using Petri dishes, Erlenmeyer flasks, etc., and the implementation of an industrial bioreactor technology which enables an extensive and stable production of aroma compounds has not been explored. Since the production of aromas is strain and waste dependent, the main challenge may be the standardization of the substrate and the adjustment of the productive process.