Cell-Free Supernatants of Plant Growth-Promoting Bacteria: A Review of Their Use as Biostimulant and Microbial Biocontrol Agents in Sustainable Agriculture

Plant growth-promoting bacteria (PGPB) afford plants several advantages (i.e., improvement of nutrient acquisition, growth, and development; induction of abiotic and biotic stress tolerance). Numerous PGPB strains have been isolated and studied over the years. However, only a few of them are available on the market, mainly due to the failed bacterial survival within the formulations and after application inside agroecosystems. PGPB strains with these challenging limitations can be used for the formulation of cell-free supernatants (CFSs), broth cultures processed through several mechanical and physical processes for cell removal. In the scientific literature there are diverse reviews and updates on PGPB in agriculture. However, no review deals with CFSs and the CFS metabolites obtainable by PGPB. The main objective of this review is to provide useful information for future research on CFSs as biostimulant and biocontrol agents in sustainable agriculture. Studies on CFS agricultural applications, both for biostimulant and biocontrol applications, have been reviewed, presenting limitations and advantages. Among the 109 articles selected and examined, the Bacillus genus seems to be the most promising due to the numerous articles that support its biostimulant and biocontrol potentialities. The present review underlines that research about this topic needs to be encouraged; evidence so far obtained has demonstrated that PGPB could be a valid source of secondary metabolites useful in sustainable agriculture.


Introduction
Plant growth-promoting bacteria (PGPB) are a widespread group of bacteria generally living in association with plants, having several beneficial effects related to (i) improvement of plant nutrient acquisition [1], (ii) promotion of plant growth and development [2], and (iii) induction of tolerance towards abiotic and biotic stress [3]. Although the mechanisms behind these effects are complex and not fully known, most of the effects can be ascribed to the bacterial ability to produce metabolites with stimulant and/or protective effects.
Among stimulant molecules, a meaningful role is played by phytohormones (i.e., abscisic acid, auxins, cytokinins, ethylene, and gibberellins). These substances regulate plant growth at all stages of development, by stimulating growth, coordination between cells, tissues and organs, and by

Methods
To find relevant publications on CFSs and their metabolites an online literature search was conducted. The following databases were employed in the search: Several combinations of search terms were attempted in each database. The terms "cell-free supernatant", "spent supernatant", "bacterial broth", "bacterial culture", and "bacterial metabolites" were combined with "biostimulant", "biocontrol", "phytopathogens", "fungi", "bacteria" "oomycetes", and "sustainable agriculture". The search was extended to all manuscript sections.
The online literature search produced a large collection of articles that have been screened according to Title and Abstract contents (Initial check). Then, articles were read completely and related papers were included in the collection if they were not already present (Related paper check). The reading and screening allowed us to discard irrelevant papers from the collection and to find a total of 109 relevant articles. The complete reading of the articles also allowed the organization of the collection based on two main categories: "biostimulant" and "biocontrol". The Biostimulant category was organized based on details about PGPB strain, compound, production technique utilized to obtain CFS/metabolites (C, centrifugation; F, Filtration; E, solvent extraction; DP, several downstream processes), crop, and experiment (P, in vitro growth; G, greenhouse growth; O, open field growth). The Biocontrol category was organized depending on the type of phytopathogen (i.e., bacteria, fungi and oomycetes) and based on details about PGPB strain, pathogen, compound, production technique utilized to obtain CFS/metabolites (C, centrifugation; F, Filtration; E, solvent extraction; DP, several downstream processes), and experiment (V, in vitro antagonism; X, ex vivo antagonism; P, in vitro growth; G, greenhouse/pot growth). For each category, tables were prepared to provide these details per reference.

CFSs as Biostimulant Agents
Over the years, the application of synthetic fertilizers in agriculture has increased to the maximum requested by global demand-crop yield [10]. Continuous fertilization campaigns repeated over the years involve considerable production costs, environmental pollution and soil degradation [11,12]. The use of PGPB-CFSs and isolated metabolites can represent an alternative sustainable technique to synthetic fertilizers. Table 1  The capability of CFSs to stimulate in vitro growth of seedlings has been reported for Medicago polymorpha [13], Oryza sativa [14], Glycine max [15,16], Zea mays [17], Lemna minor [18], Solanum lycopersicum [19], Glycine max, and Triticum aestivum [20]. The CFS obtained from A. brasilense Cd strain has been reported to be able to promote growth in an M. polymorpha seedling inoculated with Rhizobium meliloti RT1 early nodulation and changes in root morphology and function by ethylene production [13]. An 8% (v/v) CFS-based formulation obtained from A. brasilense Cd strain showed a good capability to increase in vitro O. sativa growth. In particular, the presence of CFS in the culture medium promoted better elongation, root surface area, root dry matter, and development of lateral roots of O. sativa seedlings than those grown on culture media without CFSs addition [14]. Idris et al. also described concentration-related positive effects of Bacillus spp. CFSs in Z. mays L. in a coleoptiles cylinder test [17] and in L. minor in 48-well microtiter plates growth [18]. Bacillus amyloliquefaciens KPS46 CFS metabolites positively affected growth and development of G. max under gnotobiotic condition [15]. The CFSs obtained from Burkholderia seminalis (an isolated strain selected for high levels of Indole-3-Acetic Acid (IAA) production) showed a positive impact on in vitro germination of tomato seeds [19]. Ethyl acetate extract of Methylobacterium spp. CFSs, composed mostly of cytokinins, demonstrated positive effects on Triticum aestivum L. seed germination and seedling growth [20]. To assess the actual capability of a certain compound to stimulate plant growth, in vitro experiments should be followed by in planta ones. However, among the above mentioned reports, only a few studies [13,15] [21]. Posada et al. [22] reported that CFSs of Bacillus subtilis EA-CB0575, either from vegetative cells or from spores, significantly increased shoot length and total dry weight of Musa plants compared with control. CFSs of Streptomyces acidiscabies, containing siderophores and auxins, were able to promote growth and alleviate metal toxicity in Vigna unguiculata L. [23]. Rhizobium leguminosarum bv. viciae CFSs rich in LCOs were able to ameliorate Pisum sativum and Vicia villosa growth [24]. G. max was positively affected by treatment with A. brasilense Sp7 CFSs, inducing better root growth than experimental condition treated with the bacterial inoculum [25]. For this plant, the enhancement of biostimulant effectiveness has been reported when a combination of different treatments was tested. The application of CFSs of A. brasilense strains Ab-V5 (CNPSo 2083) and Ab-V6 (CNPSo 2084) via seeds improved root morphology and nodulation in G. max inoculated with Bradyrhizobium spp. [16]. However, the efficacy was lower than co-inoculation with Bradyrhizobium spp. single strains. Positive effects on G. max were reported by Moretti et al. [26]. In their work the best results were obtained with a combination of (i) Bradyrhizobium diazoefficiens (USDA 110) and Rhizobium tropici (CIAT 889) metabolites enriched in LCO seed treatment, (ii) Bradyrhizobium japonicum (SEMIA 5079) and B. diazoefficiens (SEMIA 5080) inoculation; and (iii) A. brasilense (Ab-V5 and Ab-V6) foliar application. Efficient combination was also reported by Morel et al. [27]. These authors indicated that hydroponic solution added with bacterial and root-secreted molecules (i.e., flavonoids, phytohormones, and lipophilic chitin oligosaccharides obtained during a co-inoculation of Medicago sativa L. with Sinorhizobium and Delftia strains) increased growth of M. sativa. Overall, this combination was the most effective in terms of root development, activity (i.e., greater exploitation of the soil), nodulation, and crop grain yield (+10%) compared with plants inoculated only with Bradyrhizobium strains and other formulations.
The final confirmation of the effectiveness of a formulation can be reached in open-field experiments, where the environmental conditions are extremely variable. Open-field studies of CFS biostimulant activity are few. Marks et al. [28] reported the enhancement of grain yields of Glycine max L. and Zea mays L. when rhizobial metabolites (exopolysaccharides, phytohormones, and LCOs) were co-inoculated with both Bradyrhizobium spp. and Azospirillum spp. Similar trends were also obtained by adding Bacillus subtilis QST 713 to this combination within the foliar application. The recent article by Tewari et al. [29] indicated that a combined formulation of Bradyrhizobium sp. IC-4059, its CFSs, and exopolysaccharides (EPS) increased the productivity and nodulation of Cajanus cajan in the field, compared to both bacterial inoculum and CFS applied alone.
From all these reports it is evident that further processing of CFSs provides several metabolites with interesting stimulant properties. Among these metabolites, LCOs are the most tested. Lesueur et al. [30] summarize the effective applications of different LCOs on legume-rhizobia symbiosis, with positive outcomes on plant growth. Positive LCO application effects have also been recorded for non-leguminous plants, e.g., Zea mais, Solanum lycopersicum, Picea abies, Daucus carota, Arabidopsis thaliana [31]. Biostimulant PGPB metabolites can also be obtained from lactic acid bacteria (LABs). In addition to their probiotic properties, metabolites of these strains showed interesting biostimulant and biocontrol potential in agriculture [32]. Rodríguez-Morgado et al. [33] reported that L-lactic acid obtained from Lactobacillus rhamnosus whey-waste stimulated soil microbial activity and release of soluble phosphates. PGPB inoculation enriched with lactic acid was also involved in shaping the composition of soil bacterial communities. In a second study, the same research team published similar results on metabolites isolated by L. rhamnosus whey fermentation and separated by physicochemical processes [34]. The protein hydrolysates and the lactic acid-induces soil microbial activity. Lactic acid also positively influenced microbial biodiversity, favoring some plant growth promoter families (i.e., Bacilliaceae and Veillonellaceae family). Several PGPB strains can also be exploited to produce biosurfactants (BFs) and bacteriocins. Positive outcomes on soil quality and plant growth promotion have been extensively reviewed both for BFs [35-37] and for bacteriocins [38,39].

CFSs as Biocontrol Agents
Beyond biostimulant activity, CFSs and metabolites of PGPB can be used for the inhibition of microbial soil-borne pathogens. The strategies behind this antagonistic activity are mainly related to antibiosis and induction of plant defense response (i.e., induced systemic resistance -ISR) mechanisms [40]. The use of bioformulations in agriculture can be interesting, as it offers a valid tool for phytopathogen control whilst safeguarding ecosystems [40]. Pathogen control is a major concern in agriculture. Nowadays, the most effective strategy against plant pathogens is the use of resistant cultivars. However, due to its high costs, the application of agrochemicals remains one of the most utilized techniques [41]. Agrochemicals cause environmental pollution, with serious consequences for human health. These issues force agriculture towards effective and sustainable techniques to manage bacterial, fungal, and oomycete pathogens.

Bacterial Pathogen Control
Among soil-borne pathogens, phytopathogenic bacteria are one of the major threats for agriculture, due to the deficiency of effective agrochemicals, the absence of host plants' resistance or immunity, and the accidental and undetected spread or latency [42]. Plant bacterial diseases cause devastating damage to cultivation with huge economic losses [43].
Studies of CFSs and PGPB useful to counteract this risk are limited. In Table 2 details of the studies concerning the application of CFSs or their metabolites against bacterial phytopathogens are summarized. Literature on bacterial biocontrol by CFSs/metabolites is mainly on tests carried out in vitro against pathogens belonging to Bacillus, Clavibacter, Ralstonia, Erwinia, Micrococcus, Agrobacterium, Pectobacterium and Xanthomonas genera. Several CFS/metabolites obtained from Bacillus spp. demonstrated activity against these pathogens. In particular, the B. amyloliquefaciens species is one of the most promising. The antagonistic capability of B. amyloliquefaciens CFSs was first reported by Yoshida et al. [45], who described good inhibition of Agrobacterium tumefaciens and Xanthomonas campestris pv. campestris in ex situ Morus alba leaves. B. amyloliquefaciens Bk7, together with Bacillus laterosporus spp. (B4, S5), and Alcaligenes faecalis spp. (Bk1, P1), showed good in vitro biocontrol capabilities against Xanthomonas oryzae pv. oryzae [49]. Interesting results in planta biocontrol of X. oryzae pv. oryzae were reported by Kong et al. for CFS extracts (i.e., surfactin, iturin, and acid precipitate with a concentration of 500 µg mL −1 ) obtained from Bacillus licheniformis N1 [50]. Among several PGPB strains isolated from the rhizosphere of three horticultural and tree crops (i.e., apple, apricot, and strawberry), biocontrol capabilities were showed by B. amyloliquefaciens KM658175 CFSs against Clavibacter michiganensis ssp. michiganensis [46]; best in vitro inhibition was achieved utilizing 1% (v/v) concentration of the CFS of this strain. Extracts of B. subtilis ATCC 6633 and BBG100 CFSs inhibited in vitro growth of Erwinia chrysanthemi, Pseudomonas aeruginosa, and Micrococcus luteus due to the presence of mycosubtilin, surfactin, subtilin, subtilosin, and rhizocticins [55]. CFS of B. subtilis 14B was able to reduce the Agrobacterium tumefaciens infection both in vitro and in planta in Solanum lycopersicum [54].
The main active compounds identified in Bacillus CFSs are iturins. Iturins extracted from Bacillus sp. SS12.9 CFSs showed effective antagonism against X. oryzae pv. oryzae in in vitro experiment [51]. Iturins were also found in CFSs successfully applied in Beta vulgaris, Oryza sativa, and Cucumis sativus, in which they were able to inhibit several bacterial phytopathogens. CFSs of B. amyloliquefaciens and Bacillus pumilus inhibited Pseudomonas syringae pv. apta pathogenic activity in B. vulgaris in vitro cultivation [48]. CFS of B. subtilis NB22 and UB24 counteracted infections of X. oryzae and Pseudomonas lachrymans in O. sativa and C. sativus, respectively, during ex vivo and in planta experiments [53].
Other studies demonstrated the capability of different compounds to counteract several bacterial diseases. The ability of B. amyloliquefaciens CFSs to decrease Glycine max pustule disease severity caused by Xanthomonas axonopodis pv. glycines in a greenhouse experiment to surfactin has been ascribed [47]. Inhibition capabilities of Bacillus brevis, B. subtilis, Paenibacillus granivorans, and M. luteus strains to amylocyclicin isolated by B. amyloliquefaciens FZB42 has been recognized [44]. The ability of a lipopeptide mixture from Bacillus sp. EA-CB0959 to decrease the incidence of R. solanacearum disease in Musa plants to fengycin, and in a lesser extent to surfactin and iturin, has been ascribed [52]. In vitro antibacterial properties against A. tumefaciens to the bacteriocin BAC IH7, isolated from B. subtilis IH7, have been recognized [56].
In addition to the Bacillus genus, several CFSs obtained by LABs, showed significant in vitro inhibition against P. syringae pv. actinidiae, Xanthomonas arboricola pv. pruni and Xanthomonas. fragariae [57], thanks to the presence of organic acids. Antibacterial effects have been inactivated by pH neutralization of CFS. CFSs containing siderophores produced by P. aeruginosa RZS3 and Alcaligenes sp. STC1 strains efficiently inhibited in vitro growth of Pseudomonas solanacerum [58]. Metabolites present in the culture supernatant of Ochrobactrum lupini KUDC1013 were able to elicit ISR against Pectobacterium carotovorum ssp. carotovorum in Nicotiana leaves [59]. Several CFSs of bacterial strains isolated from suppressive soils showed in vitro antagonistic activity against X. campestris. Among them, CFSs from Peanibacillus polymyxa also revealed a strong in vivo inhibition activity against this black rot causal agent [60]. Interesting results were also reported for the purified CFS of Paenibacillus sp. strain B2; superdex-purified CFS, constituted mainly by polymyxin B, inhibited in vitro growth of Pseudomonas viridiflava and Erwinia carotovora pathogens with minimal inhibitory concentrations (MICs) of 0.6 and 6.7 µg mL −1 , respectively [61].

Fungal Pathogens Control
In addition to bacteria, phytopathogenic fungi are one of the other major microbial soil-borne pathogens that threaten productive landscapes. Fungal plant pathogens cause enormous losses in yield and quality of plants [62]. A broad-spectrum antifungal activity has been observed for diverse CFSs against the genera Fusarium, Rhizoctonia, Botrytis, Sclerotinia, Colletrotrichum, and Ralstonia. However, the majority of the studies report results on in vitro assays. Most of the studies are on Bacillus. Table 3 summarizes studies on CFS and extracted metabolites from this genus.  Bacillus mycoides (+ Pichia guilermondii) Botrytis cinerea -C -V+R (Fragaria × ananassa) [101] PT, production technique utilized to obtain CFS/metabolites; C, centrifugation; F, Filtration; E, solvent extraction; DP, several downstream processes; V, in vitro antagonism; X, ex vivo antagonism; R, antagonism on fruit; P, in vitro growth; G, greenhouse/pot growth; O, open field growth.
In addition to Bacillus genus, other genera can be valid sources of CFSs and metabolites for the biocontrol of fungal phytopathogens. In Table 4 the details of studies of species belonging to these other genera are shown.  PT, production technique utilized to obtain CFS/metabolites; C, centrifugation; F, Filtration; E, solvent extraction; DP, several downstream processes; V, in vitro antagonism; X, ex vivo antagonism; P, in vitro growth; G, greenhouse/pot growth.
One of the first studies available in the literature reports the Erwinia herbicola CFS in planta biocontrol capability against Puccinia recondita f. sp. tritici in a Triticum aestivum greenhouse experiment, thanks to the presence of herbicolin A [104]. However, no other reports can be found on this species. In the recent literature, there are many studies of the in vitro biocontrol potential of Pseudomonas spp. CFSs [58,107] and CFS metabolites [72,100], thanks to the presence of siderophores, phenazines, and 2-hexyl 5-propyl resorcinol N-Butylbenzenesulphonamide [108][109][110]. The in vitro inhibition of fungal pathogens has also been demonstrated for the CFSs and metabolites of other species of Alcaligenes [58,102], Chryseobacterium [103], and Paenibacillus [61] genera.
Actinomycetes are also a source of formulates for the management of fungal plant diseases. However, only a few studies have evaluated CFSs or metabolites obtainable by these microorganisms [124] and dealing exclusively with the Streptomyces genus [112][113][114][115][116]119]. Noteworthy are the studies of Kaur et al. and Jacob et al., who reported good in planta biocontrol capabilities of CFS on Fusarium moniliforme on S. lycopersicum [117] and Sclerotium rolfsii on Arachis hypogaea [118], respectively. LABs are capable of producing several bioactive metabolites that effectively counteracted several plant diseases [32,105]. El-Mabrok et al., for example, reported L. plantarum CFS' effective inhibition of Colletotrichum capsici, both in vitro and during a Capsicum annum seed germination experiment under sterile conditions [106]. Several works report the capability of CFSs of Xenorhabdus spp. to inhibit some fungal phytopathogens in vitro [120,122,123]. For this genus, relevant is the study of Fang et al., who reported that the extracted metabolites from X. nematophila TB CFS can inhibit B. cinerea under in vitro S. lycopersicum cultivation [121].

Oomycete Phytopathogens
Oomycetes are endemic phytopathogens responsible for destructive outcomes in several crop plants. There are only a few anti-oomycete compounds for the control of their diseases. These pathogens are spreading severely and developing resistant strains [125]. In Table 5, details of the studies concerning the application of CFSs and their metabolites against oomycetes phytopathogens are summarized. PT, production technique utilized to obtain CFS/metabolites; C, centrifugation; F, Filtration; E, solvent extraction; DP, several downstream processes; V, in vitro antagonism; X, ex vivo antagonism; P, in vitro growth; G, greenhouse/pot growth. SID, racemic 3-indoleethyl(3'-methyl-2'-oxo)pentanamide.
Biocontrol of bacterial, oomycetes, and fungal pathogens can also be achieved by bacterial BFs, bacteriocins, and hydrolytic enzymes. Several formulations of these molecules have great potential for use in agriculture. Mode of action and inhibition effectiveness have been extensively reviewed for BFs [35][36][37], bacteriocins [38,39], and hydrolytic enzymes [138].

CFSs and Metabolites -Limitations and Advantages
Data on the use of CFS in agriculture are extremely limited and their application in agriculture has been completely ignored in recent decades. No published studies have investigated formulation and shelf life of CFSs; thus, the limitations are mainly related to the downstream processes for their production. According to Doran et al. [9] downstream processes can often be technically challenging due to: • Metabolites' lability: these compounds are sensitive to temperature, high salt concentrations, and addition of chemicals (i.e., solvents, strong acids and bases). • the complexity of the broth mixture. • contamination susceptibility.
These factors limit the operation units that can be applied, lowering the purity and stability of final products. Concerning the use of CFSs as fertilizers, other possible limitations are similar to those found for other biofertilizers, namely [139]: • lower nutrient content that may be inadequate for maximum crop growth. • slower nutrient release rate. • highly variable nutrient composition.
On the other hand, CFSs have more advantages than synthetic fertilizers that can overcome these negative aspects [139]: • a more balanced nutrient supply. • soil biological and fertility status enhancement. • soil structure improvement.
These advantages sustain crop production whilst safeguarding agroecosystem health. Concerning bacteriocins, purified metabolites, hydrolytic enzymes, and BFs, currently large-scale application and production are limited mostly due to the high cost of production [31,140,141].

Perspectives
Our literature survey underlined that studies of CFSs and their metabolites should be encouraged. This resource from bacteria is in our opinion very interesting both from the scientific and commercial point of view. The metabolites present in CFS-based formulations have demonstrated effectiveness against a certain number of species. The biocontrol potential against fungi, bacteria and actinomycetes has also been demonstrated. The biostimulant market is in constant increase, with an annual growth rate of 10.4% in 2016-2021. Thus, the formulation of new products by biostimulant producers could be a valid financial investment in such a lucrative market. However, the formulation of new products ready to be commercialized would require new scientific and industrial scale-up studies. This request would challenge the scientific world as a not yet fully explored field. New studies should deal with the: (i) identification of PGPB species with interesting metabolite profile; (ii) selection of procedures to obtain cost-effective formulation; (iii) chemical characterization of formulates; (iv) modes of action; (v) effectiveness studies under different environmental conditions; (vi) studies on formulation stabilities (vii) product registration and commercialization. Even if this process is long and challenging, we think that these formulations could be one of the new tools useful for sustainable agriculture, equal to the biostimulants present on the market. Our literature survey shows that Bacillus is the most promising genus for the isolation of CFSs and/or their metabolites. Moreover, several Bacillus strains are already commercialized in biostimulant/biocontrol products. Thus, the scale-up procedures for reaching the formulation stage should also be less challenging. The collaboration of different field specialists (i.e., academics, industrial and commercial fields, farmers) should be activated to explore the CFS field and obtain new biostimulant products. We believe that the formulation of natural products for agriculture is not only important at the scientific and economic level but also for our planet. To cope with an increasingly global food demand, agriculture is maximizing production by excessive use of chemicals. The development of new fields of study and the publication of scientific reports can lead to the awareness of farmers and companies engaged in food production.

Conclusions
From the data reported, it is evident that the literature contains only a few reports useful for the creation of valid scientific evidence to support the development of CFS formulations. The majority of the reports deal with environmental controlled biostimulant and in vitro microbial biocontrol experiments. Among the 109 articles selected and examined, the Bacillus genus seems to be the most promising due to the numerous articles that support its biostimulant and biocontrol potentialities. Several CFSs and CFS metabolites of Bacillus strains demonstrated activity against a broad spectrum of bacterial, fungal, and oomycete pathogens, under different cultivation conditions. The present review underlined that research on this topic needs to be encouraged; evidence so far obtained has demonstrated that PGPB could be a valid source of secondary metabolites useful in sustainable agriculture. For the production of CFS-based formulations useful for agriculture, new PGPB strains/metabolites should be studied and obtained. Moreover, through advanced biotechnologies, standardized formulations and shelf life investigations should be carried out. To introduce these formulations in agriculture, future studies of CFSs should include effectiveness tests with trials in greenhouse and field experiments. The present review creates the first literature summary of CFSs and their metabolites as plant growth-promoting bacteria. Data organization provided details of their use as biostimulant and microbial biocontrol agents in agriculture. This review can also be used as a starting point for drawing up new reviews regarding the use of CFSs and their metabolites. These formulations can be exploited for other purposes in agriculture (e.g., biocontrol of nematodes, insects, protozoa).