**Toward a Sustainable Agriculture Through Plant Biostimulants: From Experimental Data to Practical Applications**

Editors

**Youssef Rouphael Giuseppe Colla**

MDPI • Basel • Beijing • Wuhan • Barcelona • Belgrade • Manchester • Tokyo • Cluj • Tianjin

*Editors* Youssef Rouphael University of Naples Federico II Italy

Giuseppe Colla University of Tuscia Italy

*Editorial Office* MDPI St. Alban-Anlage 66 4052 Basel, Switzerland

This is a reprint of articles from the Special Issue published online in the open access journal *Agronomy* (ISSN 2073-4395) (available at: https://www.mdpi.com/journal/agronomy/special issues/plant biostimulants).

For citation purposes, cite each article independently as indicated on the article page online and as indicated below:

LastName, A.A.; LastName, B.B.; LastName, C.C. Article Title. *Journal Name* **Year**, *Volume Number*, Page Range.

**ISBN 978-3-0365-0028-7 (Hbk) ISBN 978-3-0365-0029-4 (PDF)**

Cover image courtesy of Giuseppe Colla.

© 2020 by the authors. Articles in this book are Open Access and distributed under the Creative Commons Attribution (CC BY) license, which allows users to download, copy and build upon published articles, as long as the author and publisher are properly credited, which ensures maximum dissemination and a wider impact of our publications.

The book as a whole is distributed by MDPI under the terms and conditions of the Creative Commons license CC BY-NC-ND.

### **Contents**


Reprinted from: *Agronomy* **2019**, *9*, 306, doi:10.3390/agronomy9060306 ............... **143**




Reprinted from: *Agronomy* **2020**, *10*, 387, doi:10.3390/agronomy10030387 ............. **615**


### **About the Editors**

**Youssef Rouphael** is Associate Professor at the University of Naples Federico II, Italy. He is Editor-in-Chief of *Agronomy* MDPI and has served as Guest Editor of numerous international journals based on his expertise in biostimulants (*Scientia Horticulturae*, *Frontiers in Plant Science*, and *Agronomy* MDPI). He is a member of the Biostimulant.com scientific committee. He is internationally recognized for his research in horticultural science.

**Giuseppe Colla** is Professor of Vegetable Production, Floriculture and Greenhouse Crop Management at the University of Tuscia, Italy. He serves on the editorial boards of several leading journals in horticultural science. Professor Colla has an international reputation for his work on plant nutrition and plant biostimulants and has coordinated numerous projects to improve horticultural production.

### *Editorial* **Toward a Sustainable Agriculture Through Plant Biostimulants: From Experimental Data to Practical Applications**

### **Youssef Rouphael 1,\* and Giuseppe Colla <sup>2</sup>**


Received: 18 September 2020; Accepted: 22 September 2020; Published: 24 September 2020

**Abstract:** Modern agriculture increasingly demands an alternative to synthetic chemicals (fertilizers and pesticides) in order to respond to the changes in international law and regulations, but also consumers' needs for food without potentially toxic residues. Microbial (arbuscular mycorrhizal and plant growth promoting rhizobacteria: *Azotobacter*, *Azospirillum* and *Rizhobium* spp.) and non-microbial (humic substances, silicon, animal- and vegetal-based protein hydrolysate and macro- and micro-algal extracts) biostimulants represent a sustainable and effective alternative or complement for their synthetic counterparts, bringing benefits to the environment, biodiversity, human health and economy. The Special Issue "Toward a sustainable agriculture through plant biostimulants: from experimental data to practical applications" compiles 34 original research articles, 4 review papers and 1 brief report covering the implications of microbial and non-microbial biostimulants for improving seedling growth and crop performance, nutrient use efficiency and quality of the produce as well as enhancing the tolerance/resistance to a wide range of abiotic stresses in particular salinity, drought, nutrient deficiency and high temperature. The present compilation of high standard scientific papers on principles and practices of plant biostimulants will foster knowledge transfer among researchers, fertilizer and biostimulant industries, stakeholders, extension specialists and farmers, and it will enable a better understanding of the physiological and molecular mechanisms and application procedure of biostimulants in different cropping systems.

**Keywords:** humic substances; protein hydrolysates; silicon; arbuscular mycorrhiza; plant growth promoting rhizobacteria; macroalgae; microalgae; abiotic stresses; nutrient use efficiency; physiological mechanisms

#### **1. Biostimulants in Agriculture: Rationale**

Modern agriculture needs to review and broaden its practices and business models, by integrating opportunities coming from different adjacent sectors and value chains, including the biobased industry, in a fully circular economy strategy [1–3]. Farmers need to operate as managers of the countryside, valorizing their own by-products and using agricultural products with improved environmental profile. Therefore, searching for new technologies and approaches to boost crop productivity under optimal and sub-optimal conditions and to improve resources use efficiency (water and fertilizers) is crucial to ensure food security, while preserving soil quality and providing opportunities of business for farmers [4]. Biobased products such as biostimulants represent a sustainable, efficient technology or complement to their synthetic counterparts (i.e., agrochemicals) to improve nutrient use efficiency and secure yield stability of agricultural and horticulture crops under optimal and sub-optimal conditions [5,6]. Recently, under the new Regulation (EU) 2019/1009, plant biostimulants were defined based on four agricultural functional claims as follow: "*EU fertilising product the function of which is to stimulate plant nutrition processes independently of the product's nutrient content with the sole aim of improving*

*one or more of the following characteristics of the plant and*/*or the plant rhizosphere: (1) nutrient use e*ffi*ciency, (2) tolerance resistance to (a)biotic stress, (3) quality characteristics, or (4) availability of confined nutrients in the soil or rhizosphere*" [7]. Many diverse natural substances and chemical derivatives of natural or synthetic compounds as well as beneficial microorganisms are catalogued as plant biostimulants including: (i) humic substances; (ii) vegetal- or animal-based protein hydrolysates; (iii) macro- and micro-algal extracts; (iv) silicon; (v) arbuscular mycorrhizal fungi (AMF); and (vi) plant growth promoting rhizobacteria (PGPR) belonging to the genus *Azotobacter*, *Azospirillum* and *Rizhobium* spp. [8–16].

Plant biostimulants were initially used in organic production, but now they are adopted in several cropping systems such as conventional and integrated crop production [17]. Microbial and non-microbial plant biostimulants are usually used for open field and greenhouse crops including fruit trees, berry crops, grapevines, vegetables, ornamentals, cereals and turfs [18–21]. The biostimulants market is increasing year by year; as a matter of fact, the market of active ingredient biostimulants (amino acids, seaweed extracts, humic substances and microbial amendments) is estimated to account for 2.6 billion dollars in 2019 and is projected to reach almost 5 billion dollars by 2025, at a compound annual growth rate of 11.2% during the forecast period [7,22]. Moreover, more than 1000 scientific papers published in the last 10 years (2010–2020) were found by searching the term "plant biostimulants" and many more articles are available on the Scopus database using related words/terms (i.e., humic substances, seaweed extracts, microalgae, silicon, AMF or PGPR) (www.scopus.com).

The current Special Issue collects 39 scientific contributions (34 research papers, 4 reviews and 1 brief report) covering the different aspects of the agronomic and horticultural crops response to microbial and non-microbial biostimulants application. We highly believe that the current Special Issue: (i) will foster knowledge transfer among scientists, commercial enterprises, stakeholders and farmers; and (ii) will shed light on the cellular, molecular and physiological mechanisms as well as the application procedure of biostimulants in different cropping systems including organic farming.

#### **2. The Role of Non-Microbial and Microbial Biostimulants in Morpho-Anatomical, Biochemical and Physiological Traits of Crops**

Applications of non-microbial and microbial plant biostimulants have been shown to enhance plant growth and development, as well as macro- and micronutrient uptake and translocation in several agronomic and horticultural crops resulting in increased biomass production and yield [3]. The stimulation of seedling growth and crop productivity in response to application of non-microbial and microbial plant biostimulants is attributed to the action of bioactive substances on the primary and/or secondary metabolisms, leading to a wide array of biochemical, physiological and molecular responses [3]. Seven combinations of soy flour, diatomaceous earth, concentrated vermicompost extract (liquid) and micronized vermicompost were investigated in laboratory experiments to assess their potential biostimulant action to improve cover crops (red clover and perennial ryegrass) germination and seedling growth [23]. In their research, the authors reported that coated treatments affected in a species-specific manner the germination rate and uniformity, with a significant improvement in total germination rate recorded in red clover, while a reduction was observed in perennial ryegrass. Interestingly, the application of soy flour:diatomaceous earth at a rate of 30:70 boosted the seedlings performance in terms of shoot and root growth as well as dry matter percentage in both tested species. The authors concluded that soy flour provided a sustained source of key amino acids, thus positively influencing N uptake and transplant quality. Furthermore, Ben-Jabeur et al. [24] conducted a three-year experiment on durum wheat aiming to assess the effect of coating wheat seeds with thyme essential oil or *Paraburkholderia phytofirmans* PsJN strain on yield and resistance/tolerance to spetoria leaf botch. The two tested biostimulants were able to alleviate the Septoria leaf botch and to enhance yield in terms of number of spikes per square meter as well as straw and grain yields. The dual beneficial effect (i.e., biocontrol and biostimulant action) was also observed on tomato, where the application of four commercial biostimulants: neem seed cake, sesame oil, quillay extract and seaweeds significantly mitigated the parasitism of root-knot nematodes by reducing eggs and galls on tomato roots with the

best results recorded on neem seed cake and sesame oil treatments [25]. The authors also demonstrated that the four tested biostimulants triggered shoot and root biomass production compared to untreated control. The dual beneficial effect was also recorded on tomato, since Allaga et al. [26] reported that a composite bioinoculant containing beneficial fungi and bacteria (*Trichoderma*, *Azotobacter* and *Streptomyces*) was an efficient biocontrol agent, as well as an efficient biostimulant able to improve growth and photosynthetic activity of tomato.

Ertani et al. [27] carried out a short-term trial on hydroponically grownmaize to assess the physiological responses to leonardite-humate- and lignosulfonate-based biostimulants. The biostimulants application in particular lignosulfonates boosted root and leaf growth by 51–140% and 5–35%, respectively. The authors concluded that a putative mechanism involved in the biostimulant action of these products might be the stimulation of N metabolism in the belowground organs (i.e., roots) according to the increased activity of key enzymes such as glutamine synthetase and glutamate synthase [27]. Moreover, Kim et al. [28], elucidated the hormonal effects of a commercial vegetal-based biostimulants containing amino acids, lateral root promoting peptide, lignosulfonates and micronutrients on cuttings of basil, tomato and chrysanthemum, characterized by different relative root ability: easy, moderate and difficult, respectively. Thanks to the combination of morphological, biochemical and metabolomics approaches, the authors demonstrated that the vegetal-based biostimulant exerted similar effects to the synthetic hormone (i.e., auxin) by improving adventitious rooting responses. Finally, the authors shed light for the first time onto hormonal regulation of vegetal-based biostimulant and the crucial role of brassinosteroids in adventitious root formation.

Different amino acids (L-methionine, L-glycine and L-tryptophan at 20, 210 and 220mg/L, respectively) were applied separately on hydroponically grown butterhead lettuce to assess their stimulators role [29]. In their study, L-methionine boosted lettuce growth parameters, whereas a negative effect was observed when L-glycine and L-tryptophan were applied. Based on the results of the first experiment, Khan and co-workers conducted a second experiment with five increasing concentrations of L-methionine (0.02, 0.2, 2.2, 22, 220 and 2220 mg/L). The authors concluded that L-methionine at a concentration of 0.2 mg/L exhibited the best effect of lettuce growth parameters. In fact, it is well established that key amino acids are rapidly absorbed by the crops and act as a stable source of molecule precursors to be integrated into plant metabolism [30]. This was demonstrated by the former authors, who reported that foliar application of glutamate to creeping bentgrass foliage was rapidly absorbed and directly utilized as a precursor to synthesize gamma-aminobutyric acid and proline, two important metabolites with well-known roles in plant stress adaptation.

Bákonyi et al. [31] and Kisvarga et al. [32] reported that alfalfa brown juice could be considered a potential growth stimulator. In their studies, Celosia seedlings where sprayed at five increasing rates of fermented brown juice (0.5%, 1.0%, 1.5%, 2.0% or 2.5%), while basil was sprayed at three different increasing doses (0.5%, 1.0% or 2.5%). Water was adopted in both experiments as an untreated control. The application of alfalfa brown juice at a rate of 0.5% boosted plant growth parameters in both tested species due to the modulation of the anatomical and biochemical responses, in particular increasing the antioxidant activity of key enzymes (catalase and peroxidase) and photosynthetic pigments (chlorophyll a and b) as well as reducing the content of malondialdehyde. Moreover, Niewiadomska et al. [33] carried out a three-year experiment on white lupine cultivation, where two commercial biostimulants and six foliar fertilizers were tested. The commercial biostimulants and fertilizers were able to boost some of the biochemical activity of the soil. The authors attributed the better performance of treated-white lupine to a higher uptake, translocation and assimilation of macroand microelements.

Seaweed extracts, also known as macroalgae, are considered an important category of non-microbial plant biostimulants due to their use on several agronomic and horticultural crops under both conventional and organic farming systems [34]. Several authors reported that macroalgae such as *Ascophyllum nodosum*, *Ecklonia maxima* or *Pterocladia capillacea* can: (i) improve the agronomic performance of soybean and bean [35,36], potato [37], and Jew's mallow [38]; and (ii) enhance fruit setting in eggplant [39]. In addition

to seaweed extracts, the use of PGPR such as *Bacillus thuringiensis* was also considered an efficient approach to boost yield in a sustainable manner. Jo and co-workers [40] inoculation of *Bacillus thuringiensis* KNU-07 incurred a significant increase of total growth biomass of pepper seedlings. The beneficial effect recorded on inoculated pepper plants was associated with a strong modulation of the soil bacterial community even quantitatively or qualitatively.

#### **3. The Role of Non-Microbial and Microbial Biostimulants in Enhancing Nutrient Uptake and E**ffi**ciency**

Non-microbial and microbial plant biostimulants may positively influence nutrient use efficiency (NUE), in particular nitrogen (N) by enhancing root system architecture and soil exploration as well as increasing macro- and micronutrient solubilization that can result in an increase in NUE [17,41]. Di Mola et al. [42] demonstrated that foliar application of vegetal- (protein hydrolysates or tropical plant extract) and seaweed extract-based biostimulants (brown macroalgae: *Ecklonia maxima*) is considered a sustainable approach to increase greenhouse baby lettuce productivity and NUE in low-input cropping systems. In their study, the authors reported that the application of legume-derived protein hydrolysates and especially seaweed extract elicited important increases in fresh yield under sub-optimal and optimal N conditions (0 and 10 kg ha<sup>−</sup>1) compared to the untreated and tropical plant extract-treated plants, but the beneficial effect of plant biostimulants was not apparent under luxurious N fertilization conditions (20 and 30 kg ha<sup>−</sup>1). Similar results were also observed by the same research group [43] on two other important greenhouse leafy vegetables, namely baby spinach and lamb's lettuce, treated with a legume-derived protein hydrolysates and grown under optimal and sub-optimal N regimes. Interestingly, the foliar application of vegetal-based biostimulants incurred a significant increase in N uptake and N use efficiencies in both leafy vegetables (19% and 18%, respectively, for baby spinach and 50% and 73%, respectively, for lamb's lettuce). The authors concluded that improved agronomical performance and use efficiency of baby lettuce, baby spinach and lamb's lettuce was associated with a better photosynthetic activity and biochemical status (higher content of chlorophyll a, b and total and carotenoids) [42,43]. The synergistic biostimulant action through the application of microbial (*Trichoderma virens*) and non-microbial biostimulant (vegetal biopolymer containing amino acids, peptides and vitamins) was demonstrated on greenhouse lettuce grown with three N conditions: sub-optimal, optimal and supra-optimal (0, 70 and 140 kg ha−1) [44]. Lettuce grown under non-fertilized conditions showed an increase in marketable yield when inoculated with *T. virens* alone (45%) and a greater increase with both microbial and non-microbial biostimulant (67%). The beneficial effect of plant biostimulant was less pronounced under optimal N condition and absent under luxurious N conditions. Rouphael and co-workers concluded that, based on the improved fresh yield and NUE in greenhouse lettuce plants, treatment with plant biostimulants improved not only the chlorophyll synthesis and mineral status but also the synthesis and accumulation of antioxidant metabolites that were responsible for reactivating the photosynthetic activity and consequently the agronomic performance.

Concerning floricultural species, Leoni et al. [45] investigated the application of chemical fertilization and integrated nutrient management on yield, quality attributes and NUE of two chrysanthemum cut flower cultivars. Integrated nutrient management based on 50% synthetic fertilizers plus seaweed extract (*A. nodosum*) and microbial consortium (*Glomus* sp. and *Bacillus* sp.) was able to boost yield, quality parameters and NUE compared to the untreated control treatment.

#### **4. The Role of Non-Microbial and Microbial Biostimulants in Abiotic Stresses Tolerance**/**Resistance**

Abiotic stresses, in particular drought, salinity, heat stress, hypoxia and nutrient deficiency, are responsible for 60–70% of yield gap, dictated by global climate changes [46]. To overcome the detrimental effects of sub-optimal conditions on agronomic and horticultural crops, plant biostimulants have been proposed as an efficient agronomic tool to improve tolerance/resistance to unfavorable

environment and soil conditions [47]. In their review paper, Bulgari and co-workers summarized the biostimulants literature (humic substances, seaweed extracts, protein hydrolysates, amino acids and beneficial microorganisms) regarding their use on vegetables, focusing on their application and mode of actions to counteract the most common abiotic stresses: cold/chilling stress, heat, salinity, drought stress and nutrient deficiency. In addition to the categorized plant biostimulants, Arnao and Hernández-Ruiz [48] proposed the dual use of melatonin (N-acetyl-5-methoxytryptamine) as plant protector and biostimulant. In their review paper, they discussed the different legal aspects to categorize this natural substance as potential biostimulant at the European level. Arnao and Hernández-Ruiz [48] summarized studies of different responses of melatonin in different plant species and under diverse stress conditions by reporting the observed effects/mechanisms.

The application of four commercial biostimulants containing protein hydrolysates, humic acid and especially brown seaweed extracts (*A. nodosum*) were found to mitigate the negative effects of water stress (70% or 50% of the container substrate capacity) on potted mint by increasing the antioxidant activity of key enzymes such as catalase and superoxide dismutase and by reducing the H2O2 accumulation in leaf tissue [49]. The physiological and biochemical effects of β-(1,3)-glucan (paramylon) purified from the microalga *Euglena gracilis* on water-stress Micro-Tom were also assessed by an Italian research group [50]. The eco-physiological approach adopted in this study allowed the identification of several physiological and biochemical mechanisms of improved water stress tolerance, following the application of paramylon nanofibers, for example: (i) increasing of the photosynthetic rate; and (ii) reducing the sensitivity of photosystem II to potential dehydration damages. Moreover, Petropoulos et al. [51] showed that the application of four commercial microbial biostimulants containing AMF, *Trichoderma* and rhizosphere symbiotic bacteria enriched with amino acids or seaweed extracts were able to increase the pods and seeds yield as well as nutritional value and chemical composition of common bean under both optimal and sub-optimal water regimes. In the study by Mannino et al. [52], the impacts of four microbial biostimulants, namely AMF mono fungal inoculum, AMF multi fungal inoculum, PGPB and AMF + PGPB, on molecular and physiological responses of water-stressed tomato were evaluated. Different physiological and molecular responses of tomato to water limitation were recorded depending on microbial inocula, confirming the importance to characterize the optimal plant/beneficial microorganism genotype combination(s) to enhance plant resilience to water stress condition. Non-microbial plant biostimulants such as amino acids/peptides-based product and protein hydrolysates can also be considered an effective tools to improve the tolerance to a wide range of abiotic stresses: heat, hypotoxic, nutrient and salt stresses as well as combined environmental stresses [53,54]. The application of biostimulant based on plant and yeast extracts and containing amino acids, soluble peptides and vitamins improved the heat stress tolerance of four tomato landraces grown under Mediterranean conditions. The biostimulant effects were associated to physiological and biochemical mode of actions, for example: (i) stronger antioxidant defense system; and (ii) maximal photochemical efficiency (Fv/Fm) in leaves of the four tested tomato landraces [53]. Finally, Trevisan et al. [54] demonstrated in a short-term trial that the application of a protein hydrolysates-based biostimulant was able to mitigate the detrimental effects of single (hypoxia, salt or nutrient deficiency) and multiple (nutrient stress + hypoxia or nutrient stress + salinity) stresses of hydroponically grown maize. Root development in terms of biomass and architecture (length and density) was strongly influenced by protein hydrolysates, by upregulating the expression of key genes involved in nitrate transport and reactive oxygen species detoxification and consequently inducing a significant boost of shoot biomass.

#### **5. The Role of Non-Microbial and Microbial Biostimulants in Improving Quality Traits**

Pre-clinical and clinical studies have demonstrated the functional (i.e., health-promoting) effects of fruit and vegetables consumption in supporting human health and longevity [55]. In their review paper, Drobek et al. [56] gave an overview on how the application of microbial and non-microbial plant biostimulants can modulate the primary and secondary metabolisms of horticultural species, leading to the synthesis and accumulation of lipophilic and hydrophilic antioxidant molecules also

known as phytochemicals [3,15]. The application of vegetal-based biostimulants, in particular tropical plant extract and legume-derived protein hydrolysates, in two important leafy and fruit vegetables induced significant increase in lettuce and tomato nutritional and functional quality [57,58]. Weekly foliar application of tropical plant extract incurred a significant increase of hydrophilic antioxidant activity and total ascorbic acid in lettuce compared to untreated control [57]. Similar results were also recorded in tomato fruits, where tropical plant extract and protein hydrolysates resulted in higher bioactive compounds (total phenols and vitamin C) and lipophilic antioxidant activity than those observed in the non-treated control [58]. Concerning berry fruits, Soppelsa et al. [59] investigated the application of ten commercial biostimulants belonging to almost all the categories including: alfalfa hydrolysate, humic acids, macro-seaweed, extract and microalgal hydrolysate, amino acids alone or in combination with micronutrient (zinc), B-group vitamins, chitosan and a commercial product containing silicon. Biostimulant products based on chitosan had a major impact on strawberry pulp firmness, whereas biostimulant products based on alfalfa hydrolysate, macro-seaweed extract and microalgal hydrolysate induced an improvement in phenolic compounds compared to the remaining treatments. Moreover, in three varieties of winter rape, the application of three biostimulants with the following active substances improved the content of crude fiber and fat: titanium, sodium orto nitrophenol, sodium para nitrophenol, sodium 5-nitroguaiacolate and silicon [60].

Concerning the implications of microbial plant biostimulants on improving produce quality, Chandrasekaran et al. [61] reported that the inoculation of PGPR strain, *Bacillus subtilis* CBR05 induced a significant increase in tomato quality in terms of carotenoids profile (β-carotene and lycopene). Finally, Caser et al. [62,63] showed that the inoculation of soilless-grown saffron with *Rhizophagus intraradices* and to a lesser extent with a mixture of *R. intraradices* and *Funneliformis mosseae* boosted significantly the synthesis and accumulation of health-promoting molecules such as anthocyanins, polyphenols and vitamin C; antioxidant activity; and important bioactive compounds in saffron, such as crocin II, picrocrocin and quercitrin.

#### **6. Conclusions and Looking Forward**

In the coming few years, we can expect that plant biostimulants including both natural and synthetic substances, as well as microbial inoculants, will not only make a significant contribution to ecologically and economically sustainable crop production systems within more resilient agro-ecosystems, but will also lay the cornerstone for a future large-scale sustainable agriculture catalyzed by the biobased industry. Although plant biostimulants appear to be a novel and potential category of agricultural inputs complementing synthetic fertilizers, there is an urgent need among the research community and fertilizer industries to elucidate the molecular and physiological mechanisms which will definitely facilitate the diffusion of these bio-products in the agricultural sector. Briglia et al. [64] demonstrated that the combination of phenomic (high-throughput plant phenotyping) and genomic (Next Generation Sequencing) tools opens new perspectives to release effective biostimulant formulations to meet the emerging needs of crops. Finally, Giovannini et al. [65] suggested that, in the near future, transcriptomics research should be adopted as an integrated tool to identify the best synergistic combinations of AMF and associated bacterial communities able to enhance resources use efficiency, plant resilience and boosting nutraceutical compounds in plant species.

**Author Contributions:** Conceptualization, Y.R. and G.C.; writing—original draft preparation, Y.R. and G.C.; and writing—review and editing, Y.R. and G.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

1. Colla, G.; Rouphael, Y. Biostimulants in horticulture. *Sci. Hortic.* **2015**, *196*, 1–2. [CrossRef]


iostimulant-market-1081.html?gclid=CjwKCAjw4\_H6BRALEiwAvgfzq1LVX47L4C4O0v0leN5GfYGuk0xW2 oF25JDZhWGs03E3I2rL1kEwGxoCnsAQAvD\_BwE (accessed on 10 September 2020).


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Saffron Cultivation in Marginal Alpine Environments: How AMF Inoculation Modulates Yield and Bioactive Compounds**

**Matteo Caser 1, Íris Marisa Maxaieie Victorino 2,3,4, Sonia Demasi 1, Andrea Berruti 2, Dario Donno 1, Erica Lumini 2, Valeria Bianciotto <sup>2</sup> and Valentina Scariot 1,\***


Received: 3 December 2018; Accepted: 30 December 2018; Published: 31 December 2018

**Abstract:** Arbuscular mycorrhizal fungi (AMF) establish mutualistic symbiotic associations with plant roots and act as biofertilizers by enhancing plant nutrient and water uptake. Moreover, AMF colonization may influence the biosynthesis of plant bioactive compounds in medicinal and aromatic plants. There is limited information on AMF associations with *Crocus sativus* L. (saffron) roots and their effect on crop performances and spice quality. In the present work we verified the efficiency of root mycorrhization in potted conditions, and then we evaluated the yield and quality of the saffron produced in two Alpine sites during two cultivation cycles with the application of AMF. Two inocula were applied, either a single-species (*Rhizophagus intraradices*) or a multispecies mixture (*R. intraradices* and *Funneliformis mosseae*). The trial conducted in potted conditions confirmed that both AMF commercial inocula established symbiotic relationships with saffron roots. The multispecies inoculation yielded the highest content of arbuscules in colonized portions of the root (100%), while the single-species was slightly less (82.9%) and no AMF were recorded in untreated control corms. In open-field conditions, AMF colonization of the root systems, flower production, and saffron yields were monitored, and bioactive compounds contents and antioxidant activity in the dried spice were analyzed using spectrophotometry and high performance liquid chromatography. Overall, the saffron produced was high quality (ISO category) and had high contents of bioactive compounds, with very high total polyphenol content and elevated antioxidant activity. The use of arbuscular mycorrhizal symbionts as biostimulants positively affected saffron cultivation, improving the crop performances and the content of important nutraceutical compounds. In particular, the inoculum composed by *R. intraradices* and *F. mosseae* increased flower production and the saffron yield. *R. intraradices* alone enhanced the spice antioxidant activity and the content of bioactive compounds such as picrocrocin, crocin II, and quercitrin. Since saffron is the world's highest priced spice, the increases in yield and quality obtained using AMF suggests that farms in marginal areas such as alpine sites can increase profitability by inoculating saffron fields with arbuscular mycorrhiza.

**Keywords:** *Crocus sativus* L.; biofertilization; arbuscular mycorrhizal fungi; antioxidant activity; crocin; picrocrocin; polyphenols; safranal

#### **1. Introduction**

Saffron (*Crocus sativus* L.) is a triploid herbaceous geophyte that is reproduced by means of replacement corms and is cultivated in environments with very different soil characteristics [1–3] for its red scarlet stigmas that are used worldwide as a spice and natural dye [4]. Origin, abiotic stresses, agronomical practices, and processing methods (stigma separation, drying, and storage) can influence both the plant and the saffron spice yield, composition, and quality [5,6]. The spice's organoleptic properties are ascribed to the relative percentage of peculiar secondary metabolites—crocin, picrocrocin, and safranal—which provide the unique color, bitter taste, and aroma, respectively. The concentrations of these constituents combine to determine the saffron spice quality, as defined by the International Organization for Standardization [7]. Studies related to saffron quality are expanding mainly due to the antioxidant properties of this spice and their positive influence on human health [8]. Antitumor and cancer-preventive properties are mainly attributed to the high carotenoids content [9].

Reproductive, vegetative, and dormancy are the main phenological stages [10]. Saffron flower induction is a very complicated mechanism directly related to ecological conditions and field management [11,12]. As in most geophyte plants, both seasonal and daily thermoperiodism are involved as the main environmental factors [11]. Flower induction requires an incubation of the corms at high temperature (23–27 ◦C), followed by a period of exposure at moderately low temperature (17 ◦C) for flower emergence. In Mediterranean environments, flower induction occurs from early spring to mid summer, while flower emergence occurs from early- to late-autumn. Differences in the time required for flower initiation have mostly been attributed to the corm size [13]. In addition, Molina et al. [14] reported that air and soil temperatures might be responsible for differential flower induction and duration of up to two months. Flowering is followed by a vegetative stage throughout the winter and formation of replacement corms at the base of shoots. At the end of spring, the leaves reach the highest length, start to senesce, and wither, and the bulbs go into dormancy [14].

Due to its unique biological, physiological, and agronomic traits, saffron is able to exploit marginal land and is included in low-input cropping systems, even if high amount of skilled labour is required [11]. In Italy, saffron cultivation is gaining increasing attention as an alternative crop for sustainable agriculture systems [11,15], where it could represent a valid mean for increasing incomes of multifunctional farms, with a positive impact on the recovery and economy of these areas [15,16]. Since saffron is the world's highest-priced spice due to the intensive hand labour required for daily flower picking and stigma separation [14], small increases in the yield and/or quality can connote a large increase in profitability. In this context, the adoption of sustainable cultivation techniques such as the use of biostimulants may represent further help in both the increase in spice yield and active ingredients accumulation [17].

Recent research has focused on the benefits of soil organisms to crops, especially to promote plant nutrient uptake and assimilation [18,19]. Indeed the soil is not only the location of plant life cycle stages, but also the main reservoir for a wide range of plant biostimulants (PBs), including arbuscular mycorrhizal fungi (AMF) [19–21]. Ubiquitous and abundant, AMF are obligate endosymbionts living inside most plant roots present in diverse environments, including productive agricultural systems [22–25]. When colonizing roots, hyphae extend root limits, improving water and inorganic nutrient acquisition from the soil, mainly phosphorus (P) and other minerals, in exchange for photosynthetic products. The use of AMF has a demonstrated economic impact on agriculture and horticulture and they may also confer pathogen protection by altering plant physiological parameters, and improving soil nutrition and aggregation under different growing conditions [26–28].

Mounting evidence indicates that AMF may induce changes in primary and secondary metabolism of host plants, increasing polyphenols, flavonoids, and phytohormone dynamics [29,30]. Such metabolic changes may be ascribed to a transient activation of host defence reactions in colonized roots [20,31]. The role of AMF symbiosis in flowering date and flower production is fragmented [32].

In medicinal and aromatic plants (MAPs), such as *Arnica montana* L., *Coriandrum sativum* L., and *Anethum graveolens* L., AMF colonization influenced bioactive compound biosynthesis such

as ascorbic acid, flavonoids, polyphenols, carotenoids, and vitamins [33–36]. Inoculation with *Funneliformis mosseae* Gerd. & Trappe and *G. versiforme* P. Karst. improved plant growth and enhanced the glycyrrhizin concentration in *Glycyrrhiza uralensis* Fisch plants [26]. Moreover, under low P availability, a mix of AMF increased the production of root biomass and of pseudohypericin and hypericin content in flowers of *Hypericum perforatum* L. [32]. Although widely applied, evidence for AMF symbiosis efficacy and persistence is scant, incomplete, or lacking [37,38] and the use of AMF in crop production is facing some limitations due to product costs, producer awareness levels, and variability in mycorrhizal inoculum quality [21,27]. Many factors can affect the success of inoculation and AMF persistence, including environmental and cultivation conditions, species compatibility, degree of spatial competition with other soil organisms, and the time of inoculation. However, once AMF inoculation is restored and well established in soil, the AMF community will persist through time. If detrimental practices are minimized before and after cultivation, biodiverse mycorrhizal hyphal networks will remain unaltered and infective in the field [27]. Hence, it is important to assess the effects of AMF on crop traits both as early application and as residual persistence in the following crop cultivation seasons.

Incidence of AMF, alone or in combination with plant growth promoting bacteria (PGPB), was reported in corms of *C. sativus* [39–44]. Different authors report that well-established AMF colonization of saffron roots results in increased corm P content, chlorophyll, fresh and dry corm mass, and leaf matter, and greater soil P and nitrogen assimilation [43–45]. Shajari et al. [44] indicated a significant effect of AMF in corm growth and mineral assimilation during the second cultivation season, supporting their effective residual effects in saffron cultivation. However, little is known about the effects of AMF on spice yield, and phytochemical profiles in open field cultivation [46,47].

The possibility that AMF can enhance the economic value of saffron by increasing yield and quality is even more interesting if we consider the worldwide increase in use of biocompounds in the food and pharmaceutical industries. Thus, the aims of the present study were (1) to preliminarily verify the constitutive association of AMF with saffron roots in sterile pot conditions, and (2) to assess the AMF symbiosis in open field conditions and its effects on saffron plant growth, productivity, and bioactive compounds content in Alpine open field conditions.

#### **2. Materials and Methods**

#### *2.1. AMF Inoculation in Pot*

Saffron corms with horizontal diameters of 1.3 to 2.8 cm were sown in pots (4 L; 1 corm per pot) in the last ten days of August 2016. Pots were filled with sterile quartz sand (3 L per pot) on a layer of sterilized expanded clay (1 L per pot). Corms were treated with two inocula (MycAgro Lab, Breteniére, FR), one composed of a single fungus *Rhizophagus intraradices* (Ri) and one of *R. intraradices* and *Funneliformis mosseae* (Ri + Fm). Ten grams of each inoculum were placed under each corm in order to guarantee the contact between the inoculum and the roots and therefore to favor the symbiosis between AMF and roots. Saffron corms used as controls were not inoculated (AMF-). Corms were not treated against fungal pathogens. A randomized block design was used with a total of 48 pots displayed in two experimental plot units (24 pots per unit) and three treatments (8 pots per treatment). Cultivation lasted for one cycle (August 2016–April 2017) in a heated glasshouse of the Department of Agricultural Forest and Food Sciences (DISAFA) of the University of Torino (Italy, 45◦06 23.21" N Lat, 7◦57 82.83" E Long; 293 m a.s.l.), with an average temperature of 22 ◦C during the day and 16 ◦C in the night. Irrigation water (pH 7.4, EC 505 μS cm) was added weekly (250 mL per pot) with a drip system. The corms were fertilized by fertigation (VIGORFLOR, AL.FE. srl, MN, Italy) every two weeks starting from the emergence of the spate, in quantities of 1.5 g L−<sup>1</sup> of water. No flowering occurred because of the small size of the corms.

#### *2.2. AMF Inoculation in Open Field*

Saffron corms with horizontal diameters of 2.5 to 3.5 cm were planted in the last ten days of August 2016 in two Alpine experimental sites located in the municipality of Morgex (45◦45 35.1" N; 7◦02 37.3" E; 1000 m a.s.l.) and Saint Cristophe (45◦45 06.9" N; 7◦20 37.0" E; 700 m a.s.l.) in Italy and cultivation lasted for two cycles (2016–2017 and 2017–2018). Both sites were cultivated with saffron for at least the previous three years. Before starting the experiment both fields were milled. To assess the effects of AMF inocula on saffron cultivation and production, the same treatments used in the pot trial were applied (Ri, Ri + Fm or AMF-). A randomized block design was used, with three experimental plot units (blocks). Each plot unit consisted of 56 corms, planted in a 1.44 m<sup>2</sup> area (39 corms m<sup>−</sup>2). Inter-row planting distance was of 7 cm, while between-row distance was 25 cm. Plots were separated from each other with at least 4 m distance. Before planting, 10 g of inoculum was placed under the corms to ensure contact between plant and the treatment. Irrigation was provided when needed and hand weeding control was conducted during cultivation, while no preplanting fertilization, tillage, or treatments against pathogens were applied. The two Alpine sites were characterized by semicontinental climate, with a long and cold winter (Supplementary Figure S1). In general, both sites had a sandy-loam texture according to the USDA classification and similar chemical characteristics (Supplementary Table S1).

#### *2.3. AMF Evaluation*

At the end of the vegetative phase in both pot (February 2017) and open field experiments (April 2017 and 2018), saffron roots were harvested, rid of topsoil, cleaned and stained with 0.1% (*w/v*) cotton blue in 80% lactic acid overnight, then destained 3 times with lactic acid for 18 h, cut into 1-cm-long segments and placed on microscope slides for further morphological analysis. Approximately 25 fragments were observed under light microscope for each replicate for a total of 300 root fragments. Fungal colonization was determined and calculated as described by Trouvelot et al. [48].

#### *2.4. Plant Performance and Saffron Yield in Open Field*

The daily number of picked flowers per corm (Supplementary Figure S2) and the yield of spice (i.e., stigmas dried at 40 ◦C for 8 h in an oven) were measured at flowering (November 2016 and 2017). When leaves were fully expanded (April 2017 and 2018), 50 mg of fresh leaves per treatment were used to determine chlorophyll and carotenoids content as described by Caser et al. [49]. Simultaneous with leaf sampling, the Chlorophyll Meter SPAD-502 (Konica Minolta Sensing Inc., Osaka, Japan) was used to determine the relative quantity of chlorophyll present in 27 randomly selected plants per treatment in the field.

At the end of full plant development (April 2017 and 2018), the leaves length of all corms was measured. Then, 27 plants per treatment were lifted, and corms rid of topsoil, cleaned, and detunicated. The wilted rate as the ratio between the number of wilted corms and the total number of sown corms, the shoot caliber size, and the number, the size and the weight of replacement corms were determined.

#### *2.5. Saffron Extract Preparation and Quality*

The saffron aqueous extracts were prepared according to Gresta et al. [11]. Fifty mg of powdered saffron from each treatment and both cultivation years were put into 5 mL of deionized water. After stirring for 1 h at room temperature (circa 21 ◦C) in the dark, the solution was filtered with polytetrafluoroethylene (PTFE, VWR international, Milano, Italy) filters of 25 mm diameter and 0.45 μm pore size. The saffron extract obtained was diluted 1:10 with deionized water (1 mg mL−1). Saffron extracts were analyzed with a spectrophotometer (Ultrospec 2100 Pro, GE Healthcare, UK Ltd., Little Chalfont, Buckinghamshire, UK) to determine the amount of picrocrocin, crocin, and safranal, according to ISO 3632 [7].

#### *2.6. Total Phenols*

The content of total phenols (TPC) was measured by using the Folin–Ciocalteu's phenolic method and determined as reported by Donno et al. [50]. Five hundred μl of saffron extract was added and mixed with 30 mL of deionized water, 2.5 mL of Folin–Ciocalteu's reagent (diluted 1:10), and, after eight minutes, 10 mL of 7.5% (*w/v*) saturated sodium carbonate solution. The solution was incubated at room temperature for 2 h in the dark and the absorbance was detected at 765 nm with a spectrophotometer (Ultrospec 2100 Pro, GE Healthcare, UK Ltd., Little Chalfont, Buckinghamshire,UK). The results were expressed as mg of gallic acid equivalents (GAE) per 100 g of fresh weight (FW).

#### *2.7. Total Anthocyanins*

The total anthocyanins content (TAC) was determined using the pH-differential method [50]. Saffron extracts were added to a pH 1 and pH 4.5 buffer solutions. Absorbance of samples was determined at 515 nm and 700 nm after a 15 min equilibration. The formula for calculating TAC is as follows

$$\text{TAC (mg L}^{-1}) = \text{(A } \times \text{ sample dilution factor} \times 1000) / (\text{molar absorption} \times 1) \tag{1}$$

where A is (Absorbance 515 nm—Absorbance 700 nm) at pH 1.0—(Absorbance 515 nm Absorbance 700 nm) at pH 4.5. The results were expressed as milligrams of cyanidin 3-O-glucoside (C3G) per 100 gof fresh weight (mg of C3G 100 g−<sup>1</sup> FW).

#### *2.8. Antioxidant Activity*

The antioxidant activity (AOA) was determined using the ferric reducing antioxidant power (FRAP) method as reported by Caser et al. [51] and the 2 -azinobis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) method as described by Urbani et al. [52].

For the FRAP method, a total of 30 μL of saffron extract was added and mixed with 90 μL of deionized water and 900 μL of the FRAP reagent. After incubation at 37 ◦C for 30 min, the absorbance of the solutions was measured at 595 nm using a spectrophotometer (Ultrospec 2100 Pro, GE Healthcare, UK Ltd., Little Chalfont, Buckinghamshire, UK). Results were expressed as millimoles of ferrous iron (Fe2+) equivalents per kilogram of fresh weight.

The ABTS radical cation (ABTS +) was obtained by the reaction of 7.0 mM ABTS stock solution with 2.45 mM potassium persulfate solution. After the incubation for 12–16 h before use in the dark and at room temperature, the solution was diluted with distilled water to obtain an absorbance of 0.70 (±0.02) at 734 nm. After addition of 0.6 mL of diluted ABTS·+ solution to 180 μL of saffron extract, the reaction was left in the dark at room temperature for six min. The absorbance was then measured at 734 nm using a spectrophotometer (Ultrospec 2100 Pro, GE Healthcare, UK Ltd., Little Chalfont, Buckinghamshire, UK). The antioxidant activity was expressed as μmol of Trolox equivalents per gram of dry weight (μmol TE g−<sup>1</sup> DW).

All analyses were performed in three replicates.

#### *2.9. Identification and Quantification of Bioactive Compounds*

The chromatographic analysis of saffron extracts (Supplementary Table S2) was conducted with an Agilent 1200 high-performance liquid chromatograph coupled to a diode array detector (HPLC-DAD; Agilent Technologies, Santa Clara, CA, USA), according to established methods [53]. Different chromatographic methods were used for analysis: benzoic acids (ellagic and gallic acids), catechins ((+) catechin and (−) epicatechin), cinnamic acids (caffeic, chlorogenic, coumaric, and ferulic acids), flavonols (hyperoside, isoquercitrin, quercetin, quercitrin, and rutin), carotenoids (crocin I and II and safranal), and vitamin C (ascorbic + dehydroascorbic acids).

Four chromatographic methods were used to separate the bioactive molecules on a Kinetex C18 column (4.6 × 150 mm, 5 μm, Phenomenex, Torrance, CA, USA). Several mobile phases were used for bioactive compound identification and ultraviolet (UV) spectra were recorded at different wavelengths, based on HPLC methods, previously tested and validated [4], with some modifications. UV spectra were recorded at 330 nm (α), 280 nm (β), 310 and 441 nm (χ), and 261 and 348 nm (δ).

All single compounds were identified in samples by comparison and combination of their retention times and UV spectra with those of authentic standards analyzed with the same chromatographic conditions.

#### *2.10. Chemicals and Reagents*

All the chemicals and reagents used for the AMF evaluation, phenols, anthocyanins, FRAP, and ABTS assays and bioactive quantification were purchased from Sigma-Aldrich (Saint Louis, MO, USA).

#### *2.11. Statistical Analysis*

Arcsin transformation was performed on all percent incidence data before statistical analysis in order to improve homogeneity of variance (Levene test). All the analyzed data were checked for normality of variance. For all indices analyzed in the greenhouse assay, mean differences were computed using a one-way analysis of variance (ANOVA) with Tukey's post hoc test (*p* ≤ 0.05). Data from open field were analyzed by means of a linear mixed effect models considering AMF treatments as a fixed factor, year as a repeated measure, and sites and blocks as random factors. The following interactions (year × AMF treatment) were included in the model. Pairwise comparisons (according to sequential Bonferroni post hoc tests) were used to separate means when a treatment was significantly affecting the variable at a *p* ≤ 0.05. All presented values are means of untransformed data. All computations were conducted with SPSS statistical package (version 25.0; SPSS Inc., Chicago, IL, USA).

#### **3. Results**

#### *3.1. Assessment of Saffron Mycorrhization at Pot and Open Field Scale*

Values concerning intensity of colonization in the root system and abundance of arbuscules or coils in the saffron roots in potted conditions are shown in Table 1. Corms treated with Ri + Fm reached the highest level of mycorrhization (M%) (93.33%), however, high levels were also obtained with Ri inoculum (71.37%). The Ri + Fm treatment also had the highest occurrence of arbuscules (a%) in the mycorrhizal portions (100%), being significantly higher than Ri (82.99%) and AMF-(0%).

**Table 1.** AMF colonization indices (intensity in the whole root system, M; intensity of the mycorrhizal portions, m; presence of arbuscules in the whole root system, A; presence of arbuscules in the mycorrhizal portions, a) of *Crocus sativus* L. roots treated with the inoculum composed by *Rhizophagus intraradices* and *Funneliformis mosseae* (Ri + Fm), *R. intraradices* alone (Ri), or the control (AMF-) in the saffron pot cultivation.


Mean values with the same letter are not statistically different at *p* ≤ 0.05 according to Tukey's post-hoc tests. The statistical relevance is provided (\*\*\* *p* < 0.001).

In open field conditions, the AMF root colonization measurements in *C. sativus* treated with Ri + Fm or with Ri alone during the two cultivation cycles are presented in Table 2. In general both the presence of arbuscules in the mycorrhizal portions (a%) and in the whole root system (A%) indices were affected by the inoculum composition only in the first cultivation year, while control plants

(AMF-) were not colonized. In the second year, low root colonization was observed and no differences among the treated and untreated corms were detected.

**Table 2.** AMF colonization intensity in open field conditions after the first and second cultivation year of the whole root system (M) and of the mycorrhizal portions (m), and presence of arbuscules in the whole root system (A) and in the mycorrhizal portions (a) of *Crocus sativus* roots treated with inoculum composed of *Rhizophagus intraradices* and *Funneliformis mosseae* (Ri + Fm), *R. intraradices* alone (Ri), or the control (AMF-).


Values with the same letter denote no significant differences. The statistical relevance is provided (ns, not significant; \* *p* < 0.05; \*\*\* *p* < 0.001).

#### *3.2. Impact of AMF Symbiosis on Saffron Productivity and Qualitative Traits in Open Field*

Significant differences between the two cultivation years emerged for several studied parameters. In general, the wilting rate, all the main productivity traits (number of flowers m<sup>−</sup>2, number of flowers per corm, mg of saffron m−2, saffron per flower, and the number of replacement corms), and the content of leaf chlorophyll and carotenoids significantly increased after the second year of cultivation (Table 3); a reduction in leaf length, SPAD unit, and shoot size was also observed.

**Table 3.** Effects of cultivation seasons (Year 1 and Year 2), AMF treatments (Ri + Fm was composed of *Rhizophagus intraradices* and *Funneliformis mosseae*, Ri of *R. intraradices* alone, and AMF-was the uninoculated control), and their interaction (Year × AMF treatment) on saffron plant growth and productivity based on linear mixed-effects models considering AMF treatments as a fixed factor, year as a repeated measure, and sites and blocks as random factors.


Values with the same letter denote no significant differences. The statistical relevance is provided (ns, not significant; \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001).

In Table 3, the productivity and growth traits influenced by the AMF treatments are also reported. Particularly, the inoculum composed of the mix of *R. intraradices* and *F. mosseae* significantly increased the number of flowers m−<sup>2</sup> (+138.2%), the number of flowers corm−<sup>1</sup> (+130.8%), and the mg of saffron m−<sup>2</sup> (+135.2%) in comparison to other treatments. In contrast, the mg of saffron flower−<sup>1</sup> and the shoot size were significantly reduced (−20% and −40%, respectively) by the inoculum of *R. intraradices* alone in comparison to Ri + Fm and AMF-. Significant interaction between cultivation year and AMF treatments resulted for the number of flowers m-2, the number of flowers corm−1, saffron yield, saffron flower<sup>−</sup>1, and shoot size.

Regarding the synthesis of bioactive molecules in the studied saffron spice, differences between the two cultivation seasons occurred (Table 4). Overall, the saffron produced at the two experimental sites belonged to the quality category I for the picrocrocin, safranal, and crocins analysis [7] with a significant increase after the second cultivation year. On the contrary, different bioactive compounds (isoquercitirin, quercitrin, ellagic acid, safranal, and total vitamin C) were significantly reduced. Very few differences were observed among AMF treatments (Table 4). Both Ri + Fm and Ri positively affected the antioxidant activity (FRAP assay) of the saffron produced. While, the effect of the Ri inoculum significantly increased the absorbance value of picrocrocin (ISO 3632) and the content of quercitrin in comparison to Ri + Fm, and the content of crocin II compared to AMF- (Table 4). A significant interaction between cultivation seasons and AMF treatments resulted for picrocrocin (ISO), quercitrin, crocin II, and antioxidant activity (FRAP assay).

**Table 4.** Effects of cultivation seasons (Year 1 and Year 2), AMF treatments (Ri + Fm was composed of *Rhizophagus intraradices* and *Funneliformis mosseae*, Ri of *R. intraradices* alone, and AMF- was the uninoculated control), and their interaction (year × AMF treatment) on bioactive compounds, total polyphenol content (TPC), anthocyanins, quality traits as defined by ISO 3632 [7], and antioxidant activity of the produced saffron based on liner mixed-effects models considering AMF treatments as a fixed factor, year as a repeated measure, and sites and blocks as random factors.


Values with the same letter denote no significant differences. The statistical relevance is provided (ns, not significant; \* *p* < 0.05; \*\* *p* < 0.01; \*\*\* *p* < 0.001).

#### **4. Discussion**

#### *4.1. AMF Colonization*

In the literature only some studies report AMF colonization of *C. sativus* roots. In the present study, their presence in potted cultivation was detected in *C. sativus* roots subjected to both AMF treatments (Ri + Fm and Ri). Saffron root fragments showed extensive AM fungal colonization, characterized by a moderate to high intensity of colonization and arbuscule formation. Saffron root colonization in the present pot cultivation trial was markedly superior to the results obtained in the open field test. This could be due to the antagonistic action of the naturally occurring fungi in the soil that compete with the AMF and by the different cultivation substrate used. However, to the best of our knowledge, this is the first report clearly indicating and measuring successful symbiosis between *C. sativus* roots and AMF under pot cultivation conditions.

Our open field data are equal or lower than findings obtained in other open field trials as reported by Aimo et al. [40] and Lone et al. [43]. Applying the percent colonization method for root AMF evaluation, these authors reached a maximum of 30% and 60% mycorrhizal colonization in saffron roots in Italy and Kashmir cultivation fields when using a mix of AMF belonging to the genus Glomus, respectively. In a similar study conducted in Iranian fields, the percentage of root colonization of saffron was of 39% [42], while in a field in Kashmir, ranged between 15 and 90% on the basis of the season [43]. As reported in Supplementary Table S1, P Olsen values measured at the experimental sites are high (>69.2 mg kg−1), indicating the potential for a detrimental effect of P on AMF colonization in our experiment. As the cost of the symbiosis to the plant outweighs the benefit of access to P via the fungal pathway, plants reduce fungal access to carbohydrate [54]. Similar data were reported also in other species such as *Zea mays* L. in which the AMF root colonization was reduced with a soil P content of 90 mg kg−<sup>1</sup> [55]. In other geophyte plants, such as *Allium tricoccum* Aiton., a low level of AMF symbiosis was observed in the absence of leaves and photosynthetic activity. However, once leaves elongate in early spring, root colonization increases rapidly. This is similar to the pattern of *Maianthemum racemosum* L., where AMF colonization peaked during vegetative growth [56]. Here, AMF sampling was performed during maximum leaf elongation, and therefore, the detection of low colonization is likely more related to soil characteristics than to other physiological or biochemical parameters.

Taken together, all these findings indicated that under open field conditions in alpine environments, AMF colonization was substantially lower than under pot conditions as already indicated in literature. This is in agreement with the meta-analysis of Berruti et al. [27], in which successful outcomes of AMF inoculation were more often found in controlled (greenhouse and growth chamber) conditions. In this condition, environmental extremes and variation are minimized or absent [38]. Moreover, one of the most important confounding factors in pot or field experiments is the effect of root temperature on the AMF growth [57]. The higher temperatures typical of greenhouse conditions favor greater growth and superior colonization by AMF [58].

#### *4.2. AMF Modulate Crop Performance and Spice Quality*

Flower yield is a difficult parameter to forecast in saffron since it is influenced by a combination of agronomic, biological, and environmental factors [11]. Generally, a saffron field may produce from 200 to 3000 mg m−<sup>2</sup> of spice, depending on the cultivation factors [11] and obviously, by the planting density, which may vary considerably. By planting at a 55 corms m−<sup>2</sup> density in southern Italy (Sicily), Gresta et al. [3] obtained more than 1200 mg m−2. In the area of Navelli (central Italy) [59], with a similar corm density, the average yield ranged between 1000 and 1600 mg m−2. In Iranian fields with a density of 150 and 100 corms m−2, Mollafilabi et al. [60] and Koocheki et al. [61] obtained 740 and 370 mg m−<sup>2</sup> of saffron, respectively. Recently, the path coefficient analysis conducted by Bayat et al. [62] highlighted that fresh stigma weight, flower number, dry stigma and flower weight, leaf size, and number and size of replacement corms have the highest positive correlation with saffron yield.

Arbuscular mycorrhizal fungi are known to be beneficial to several important plants, including some medicinal plants [30]. Unfortunately, very scarce reports of the effective role of AMF in saffron yield are available. Only, Aimo et al. [40] indicate an increase in flower production m−<sup>2</sup> (equal to 68%, compared to control) using a mix of AMF species belonging to the genus Glomus. Our results are generally more supportive of the benefits of AMF inoculation with an increase of flower production m−<sup>2</sup> of circa 140%. Taken together, these findings suggest a beneficial effect of AMF inoculation with a mixture of *R. intraradices* and *F. mosseae* on saffron yield performance.

Few spices are able to provide the combination of color, taste, and aroma to the foods and possess several nutraceutical properties for human health as saffron. Most of the beneficial effects of saffron, recognized since ancient times, are due mainly to its total phenolic content (TPC) and antioxidant activity (FRAP and ABTS assays). *R. intraradices* alone was found to induce an increase in secondary metabolite contents, such as terpenes and phenolics, in *Salvia officinalis* L. [63] and *Echinacea purpurea* L. [64]. Overall, the saffron produced in the studied alpine areas had very high TPC (ranges between 1340.7 and 2355.5 mg GAE 100 g−<sup>1</sup> DW), which was more than saffron cultivated in different areas of Lebanon (160 mgGAE 100 g−<sup>1</sup> DW) [65], and is much greater when compared with other common food additives and spices, such as *Eugenia caryophylate* (Thunb.), *Lavandula* spp., *Curcuma domestica* Val, and *Curcuma longa* L. (0.26, 0.22, 23, and 36 mg GAE 100 g−<sup>1</sup> DW, respectively) (Table 4) [66,67]. Results of ABTS and FRAP assays also demonstrated elevated antioxidant activity (Table 4). ABTS assay values were comparable to what was found in Greek saffron by Ordoudi et al. [68]. FRAP assay values (between 408.9 and 1937.1 mmol Fe2+ kg<sup>−</sup>1) were generally higher in comparison to the Iranian samples (circa 570 mmol Fe2+ kg−1) analyzed by Karimi et al. [69]. The saffron produced in the west Italian Alps also had different bioactive compounds (Table 4) known for their health-promoting activity, that is, cinnamic acids, flavonols, benzoic acids, catechins, and carotenoids [50]. Other studies report that water-soluble carotenoids such as crocins have antioxidant effects superior to α-tocopherol [67]. It was recently observed in a clinical study that high crocin I and crocin II contents (4000 and 1000 mg, respectively) inhibit β-amyloid and tau aggregation [70]. Apart from crocins, Asdaq and Inamdar [71] suggest that flavonols are responsible for the synergistic antihyperlipidemic and antioxidant potential of saffron. Amin et al. [72] indicated that a concentration of 1 mg of safranal attenuated the behavioral symptoms of neuropathic pain. Our data indicate that the saffron produced presented high crocin II content (27.7−38.8 mg 100 g−<sup>1</sup> DW), almost in line with the saffron produced in Sardinia (Italy, DOP Zafferano di Sardegna) [73], while also presenting a higher content of gallic acid compared to what was found in Iranian and Greek saffron (2 mg and 1.2 100 g−<sup>1</sup> DW) by Karimi et al. [69] and Proestos et al. [74], respectively. Thus, the saffron obtained could be of particular interest for its elevated antioxidant properties.

#### **5. Conclusions**

Saffron quality may vary greatly by site on the basis of several factors, among which are climatic conditions and cultivation techniques. We hereby provide data indicating the production of high quality saffron in marginal alpine areas, thus confirming that this crop is a strategic resource and good alternative for mountainous areas building multifunctional economies. Besides the phytochemical profile highlighted, the crop had many bioactive compounds. The use of arbuscular mycorrhizal symbionts as biostimulants positively affected saffron cultivation, mainly by increasing crop productivity, and partially by increasing the content of important nutraceutical compounds. Specifically, the inoculum composed by *R. intraradices* and *F. mosseae* was particularly effective in increasing flower production and saffron yield, while *R. intraradices* alone increased the content of some bioactive compounds—picrocrocin, quercitrin, crocin II—as well as antioxidant activity. Since saffron is the world's highest priced spice, the increases in yield and quality obtained using AMF should allow for an increase in profitability.

Furthermore, a new perspective can be envisaged. Since AMF symbiosis was more effective under soilless pot cultivation, this system may be a valuable alternative for saffron production and further work is underway to assess the potential of AMF inocula in saffron soilless cultivation.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4395/9/1/12/s1, Figure S1: Climatic conditions of the Alpine experimental sites, Figure S2: Effects of AMF inoculum composed by *Rhizophagus intraradices* and *Funneliformis mosseae* (Ri + Fm), *R. intraradices* alone (Ri), or control (AMF-) on flower production m-2 during the first (a) and second (b) cultivation cycle, Table S1: Physical and chemical properties of the soils collected in the three saffron experimental fields located in the municipality of Saint Cristophe and Morgex (north west Italy), Table S2: Characteristics of the HPLC methods applied to analyse the bioactive compounds present in the studied saffron samples.

**Author Contributions:** M.C., A.B., E.L., V.B., and V.S. contributed to the experimental design. M.C., Í.M.M.V., S.D., D.D., A.B., and V.S. acquired and interpreted data. M.C. drafted the manuscript. V.B. and V.S. conceived, coordinated the work, and critically revised the manuscript.

**Funding:** This research was funded by the project titled 'SaffronALP—Lo zafferano di montagna: tecniche sostenibili per una produzione di qualità'—Fondazione Cassa di Risparmio di Torino (RF = 2017.1966) and by the program Interreg V-A Francia Italia Alcotra "Attività innovative per lo sviluppo della filiera transfrontaliera del fiore edule—Antea' n. 1139.

**Acknowledgments:** The authors acknowledge Dario Sacco for statistical assistance, Alessandro Putzolu for technical assistance, and Azienda Agricola La Branche di Diego Bovard and Azienda Agricola Rosset for providing plant material and hosting cultivation.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **A Novel Biostimulant, Belonging to Protein Hydrolysates, Mitigates Abiotic Stress Effects on Maize Seedlings Grown in Hydroponics**

#### **Sara Trevisan 1,2,\*, Alessandro Manoli <sup>1</sup> and Silvia Quaggiotti <sup>1</sup>**


Received: 28 November 2018; Accepted: 7 January 2019; Published: 9 January 2019

**Abstract:** The main challenge to agriculture worldwide is feeding a rapidly growing human population, developing more sustainable agricultural practices that do not threaten human and ecosystem health. An innovative solution relies on the use of biostimulants, as a tool to enhance nutrient use efficiency and crop performances under sub-optimal conditions. In this work a novel biostimulant (APR®, ILSA S.p.A., Arzigano VI, Italy), belonging to the group of protein hydrolysates, was supplied to maize seedlings in hydroponic and its effects were assessed in control conditions and in the presence of three different kinds of stresses (hypoxia, salt and nutrient deficiency) and of their combination. Our results indicate that APR® is soluble and is able to influence root and shoot growth depending on its concentration. Furthermore, its effectiveness is clearly increased in condition of single or combination of abiotic stresses, thus confirming the previously hypothesised action of this substance as enhancer of the response to environmental adversities. Moreover, it also regulates the transcription of a set of genes involved in nitrate transport and ROS metabolism. Further work will be needed to try to transfer this basic knowledge in field experiments.

**Keywords:** Maize; biostimulant; root; stress; growth; gene expression

#### **1. Introduction**

Global population is expecting to increase to nine billion by 2050 [1] and agriculture will need to push crop production accordingly in order to sustain the greater demand for food. This is especially true in developing countries where high rates of population growth are associated to an increased urbanization, leading to changes in income levels and food preferences [2].

Moreover, climate change leading to abiotic pressures, such as rising droughts and other stresses correlated to higher temperature, are predicted to escalate in their severity and frequency [3,4] thus seriously compromising crop productivity [5]. In fact, abiotic stress can reduce crop yields by more than 60% for major crops [6–8].

New crop protection solutions able to mitigate the main abiotic stresses represent a substantial opportunity to contribute to secure, higher and more stable yields. These innovations span across conventional breeding to biotechnology solutions [9] and also encompass new generations of agrochemicals [10]. The global crop protection market attained US \$56.7 billion in 2014. However, there are only limited solutions currently available to mitigate abiotic stresses.

In recent years, the use of natural-derived biostimulants is proposed as an innovative solution to address the challenges of sustainable agriculture, by ensuring optimal nutrient uptake, crop yield, quality, and tolerance to abiotic stress [11].

An innovative technology with promising application potential entails the use of a particular class of biostimulants, the protein hydrolysates (PHs). PHs are mixtures of polypeptides, oligopeptides, and free amino acids derived by chemical or enzymatic hydrolysis of plant residues or animal connective tissues. The protein hydrolysates have been demonstrated to stimulate root growth and leaf biomass of several crops. Du Jardin [11] reviewed various effects resulting from the application of these compounds to crops and Van Oosten [12] reported several studies demonstrating the role of PHs in abiotic stress response.

Although the effects of protein hydrolysates on crop performance have been documented by several studies [11,12] the scientific basis of their action has only partially been elucidated mainly due to the complex nature of these products [13]. However, the synthesis of the enzymatic hydrolysis of protein has been an advantageous, ecologically safe strategy to produce biostimulant [13], and more studies are needed to improve protein hydrolysates production techniques and to ensure a low- cost product for consumption and a high use efficiency [14].

In an earlier study we demonstrated a role for a new-synthetized PH (APR®, ILSA S.p.A.) in regulating the expression level of a thousand of genes in maize roots, and hypothesised that it could act by improving the plant responses to various environmental stresses [15]. Based on the results therein obtained APR® has been proposed to enhance plant response to stress. However, this preliminary work has tested APR® on plants grown in not adverse conditions and APR® was applied directly to the soil mixture as solid granules. The chemical composition of this compound (identified also as AA309) is reported in subsequent study by Ertani et al. [16] which also performed Fourier transform infrared (FTIR). The chemical analysis revealed the presence of several amino acids, as lysine, phenylalanine, glycine, aspartate, and isoleucine. The present substance is still under study and its dossier is expected to be definitively completed by the next three months.

Due to the economic importance of maize and the limitations in fertilizer applications imposed during its development, it is important to dissect the effects of the biostimulant on the initial growth of maize seedlings, at both morphological and transcriptomic levels. For this reason, most of the present works on APR® are focused on its effects on this species.

Our previous work [15] suggested that APR® is at least in part soluble and reach root through the soil solution. Furthermore, it seems to act as a stress tolerance enhancer, by modulating the transcription of a wide set of genes involved in ROS detoxification and nutrient acquisition. However, no results on its effects in abiotic stress conditions were gained until now. Our various results with this species indicate that maize is able to sense and rapidly respond to nutritional fluctuations already after hours or minutes of treatment [17–20]. Therefore in the present work, we tried to deepen the effects of APR®, supplied in in hydroponic, in affecting the early response of maize seedlings to abiotic stresses. To this aim we first aimed to assess the APR® activity by measuring its effects on plant growth and identified the optimal concentration to be used in further experiments. Subsequently, to study the effectiveness of APR® as an enhancer of plant tolerance to abiotic stress we grown maize seedlings in the presence of different single and combined abiotic stresses and supplying them with APR®. Our results on root and shoot growth and on the expression profiles of a number of previously identified genes [16] provide further evidence of the APR® biostimulant activity, which early induce root to elongate and affects gene expression, especially increased in conditions of environmental limitations.

#### **2. Materials and Methods**

#### *2.1. Maize Seedlings Growth*

Seeds of maize (*Zea mays* L.), inbred line B73, were washed in distilled water and germinated on wet filter paper at 25 ◦C in the dark. After three days, maize seedlings were transferred in a controlled environmental chamber in 500 ml tanks containing a nutrient medium which was constantly aerated and composed as previously described in Quaggiotti [21] and changed every two days. Plants were grown in a growth chamber with an 8-h photoperiod under 200 μmol m−<sup>2</sup> s−<sup>1</sup>

of photosynthetically active radiation (PAR; daylight and warm white 1:1, LF-40W) at day/night temperatures of 21/18 ◦C [21]. The pH of the medium was checked during the growth period and remained at a stable level of around pH 6.5. For analysis of RNA root samples were frozen in liquid nitrogen and stored at −80 ◦C.

#### *2.2. Set up of the Novel Biostimulant Concentration to be Supplied to Stressed Maize Plants*

Maize seedlings were hydroponically grown for three days in distilled water containing different APR® concentrations, resulting in a variable nitrogen content which ranged from 1% to 10% of the amount of nitrogen supplied by the Hoagland-modified nutrient solution previously described. APR® granules were added to tanks 2 h before putting plants into the water and constantly stirred until all product has dissolved.

This series of concentrations was selected basing on their relative content of nitrogen, paying attention to keep it to a sub-nutritional level. To evaluate their effects the root length, root and shoot fresh weight were measured. Data are expressed as the average of three replicates (*n* = 10) ± standard error. For statistical analysis, we compared morphological data derived from the corresponding four different APR® concentrations with those of control plants.

#### *2.3. Stress Application and Morphological Analyses*

To try to assess the effect of APR® on maize tolerance to abiotic stress three single stress (hypoxia, salt and nutrient starvation) and two stress combination (hypoxia plus nutrient starvation, salt plus nutrient starvation) were imposed to seedlings for three days. Comparisons were made among non-stressed and stressed plants, which were then compared with plants supplied also with 5% of APR®.

Hypoxic stress conditions were achieved by not bubbling air through the liquid solution for the entire experiment. For salt stress, a 25 mM NaCl concentration, which corresponds to mild salt stress in maize was employed [22–24]. For nutritional stress, seedlings were grown in distilled water only. Each treatment was performed in three biological replicates.

After 3 days, roots and shoots of control and APR® treated plants were harvested. For the morphological analyses, 10 randomly selected seedlings for biological replicate were used.

The remaining plants were immediately frozen in liquid nitrogen and kept at −80 ◦C for subsequent RNA extraction.

#### *2.4. RNA Extraction, and cDNA Synthesis*

Total RNA was extracted from root tissues using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) as previously described by Trevisan et al., [17]. DNAse digestion was performed whit RQ1 RNAse-free DNAse (Promega, Madison, WI, USA) on an aliquot of total RNA as described by Trevisan et al. [17]. RNA was quantified using a Nanodrop 1000 (Thermo Scientific, Nanodrop Products, Wilmington, DE, USA) and its quality further validated by sterile agarose gel electrophoresis. cDNA was synthesized from 500 ng of total RNA mixed with 1 μl of 10 μM oligo-dT, as described by Manoli et al. [25].

#### *2.5. Real Time qPCR*

Relative quantification of transcripts by RT-qPCR was performed in a StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). Reactions were performed using SYBR Green chemistry (Applied Biosystems fast SYBR Green Master Mix, Thermo Fisher Scientific, Waltham, MA, USA), following the manufacturer's instructions. Reverse-transcribed RNA (2.5 ng) was used as template in each reaction as indicated by Manoli [26]. Three technical replicates were performed for each thesis using the conditions described by Trevisan et al. [17]. Melting-curve analysis confirmed the absence of multiple products and primer dimers. Data were exported and analysed according to the method of Livak and Schmittgen [27] and MIQE guidelines [28]. Only transcripts showing amplification with quantification cycle (Cq) < 35 were selected for subsequent gene expression analysis. All of the primers used in these assays are listed in Table S1.

#### *2.6. Data Analysis*

The expression levels of the analysed genes were normalized via comparison to the expression of the internal reference gene (MEP, membrane proteinPB1A10.07c, primers: forward 5'-TGTACTCGGCAATGCTCTTG-3' and reverse 5'-TTTGATGCTCCAGGCTTACC-3'), as the reference gene [25]. The standard error was calculated from the standard deviation and the variation coefficient of the reference gene and of the genes under assessment.

For statistical analysis, we compared stress condition plants with its own control. Data represent means ± SD of 3 independent experiments performed in triplicate. For the gene expression levels analyses and the choice of the APR® concentration, multiple comparison statistics were calculated using the software RStudio (https://www.rstudio.com/) Version 1.1.453. differences among samples were verified with either ANOVA (normality and homogeneous variances) or Welch's one-way ANOVA (normality and non-homogeneous variances) followed by post hoc LSD or Waller-Duncan test, respectively, and with Kruskal-Wallis (non-normality and homogeneous variances) or Friedman test (non-normality and nonhomogeneous variances). For all statistics a *p*-value threshold of 0.05 was adopted. For the subsequent growth analyses, one-way ANOVA test followed by Tukey's HSD test was performed. Asterisks indicate significant differences (\*: *p* < 0.05; \*\*: *p* < 0.02). One-way ANOVA, Tukey's HSD test.

#### **3. Results**

#### *3.1. Choice of the Novel Biostimulant Concentration to be Used for Subsequent Treatments in Hydroponics*

In order to assess the most effective APR® concentration four different APR® concentrations (1%, 2%, 5% and 10%) were used and their effects were observed in comparison to those measured for seedlings grown in distilled water for the same period (Ctrl) (Figure 1).

APR® application induced a significant increment of root length when supplied in hydroponic at a concentration of 5%The 10% dose showed a reduction on root length respect to the 5%.

In the case of root and shoot weight no statistically significant differences were observed (Figure 1B,C). According to these data, we decided to use a 5% concentration of APR® for all subsequent analyses.

**Figure 1.** Effects of different APR® concentrations on the physiological growth parameters of *Zea mays* L. seedlings. The graphs represent the root length, the root weight and the shoot weight of maize seedlings hydroponically grown for three days at increasing concentrations of APR®. The values represented in the graphs were calculated from three independent experiments (*n* = 10) and represent the mean ± the standard error. Significantly different values (*p* < 0.05) are evidenced by different letters (One-way Anova, LSD post-hoc test).

#### *3.2. Biostimulant Effects on Root Length in the Presence of Different Stress*

When maize seedlings were subjected to hypoxic stress (H) the root length showed values 12% lower if compared to control plants (Figure 2). However, when APR® was supplied to the nutrient solution a significant increase of root elongation was measured, with values 10% higher than those observed for hypoxic plants and similar to those noticed for control plants.

A similar pattern was observed when plants were subjected to salt stress (S), which triggered a visible reduction in primary root length. However, the provision of APR® triggered a significant increment of root length, thus restoring the phenotype of control plants.

Also in the case of nutrient deprivation (N) the supply of APR® significantly induced the primary root to elongate.

The positive effect of APR® provision was even more marked in the case of combined stress. In fact, when hypoxia was associated to nutritional stress (N/H) primary root length was visibly in comparison to the control, but the presence of APR® markedly and significantly restrained this negative effect leading to a phenotype comparable with that observed in not stressed plants. A positive influence of the biostimulant was observed also in the case of the combination between salt stress and nutritional stress (N/S), which inhibited the primary root growth, but the provision of APR® led to a root elongation 40% higher than that measured for stressed plants.

The provision of APR® to control plants did not induce significant effects on root elongation.

**Figure 2.** APR® counteracts the negative effects of single and combined abiotic stress on root length. Maize seedlings were subjected for 3 days to several abiotic stresses in absence (grey bars) or in presence (black bars) of 5% APR®. The applied single stresses were: hypoxic stress (H), salt stress (S), nutritional stress (N). The single stresses were combined as: hypoxic stress plus nutritional stress (N/H) and salt stress plus nutritional stress (N/S). The values of root length (cm) are represented in the graphs (mean ± SE) and were calculated from three independent experiments (*n* = 10). Significantly different values are evidenced by \* (\*: *p* < 0.05; \*\*: *p* < 0.02; One-way ANOVA, Tukey's HSD test).

#### *3.3. Biostimulant Effects on Root and Shoot Weight in the Presence of Different Stress*

To verify if the increments observed in terms of primary root length were associated to an increase of total root weight, these parameters was measured in the same conditions described above (Figure 3). When maize seedlings were subjected singularly to the three different stresses (hypoxia, salt, nutritional deficiency) or to the hypoxia and nutritional deficiency (N/H) combination the root weight did not evidence significant differences in comparison to the control plants nor in response to the biostimulant. On the contrary, a significant increase of root weight in response to APR® provision was measured when plants were subjected to the combination of nutritional and salt stress (N/S).

**Figure 3.** Effects of APR® on root weight in response to different single and combined abiotic stresses. Maize seedlings were subjected for 3 days to several abiotic stresses in absence (grey bars) or in presence (black bars) of 5% APR®. The applied single stresses were: hypoxic stress (H), salt stress (S), nutritional stress (N). The single stresses were combined as: hypoxic stress plus nutritional stress (N/H) and salt stress plus nutritional stress (N/S). The values of root weight (g) are represented in the graphs (mean ± SE) and were calculated from three independent experiments (*n* = 10). Significantly different values are evidenced by \* (\*: *p* < 0.05; \*\*: *p* < 0.02; One-way ANOVA, Tukey's HSD test).

As far as the shoot weight was concerned (Figure 4) no differences were observed in response to stress, nor providing APR®, except in the case of the contemporary presence of nutritional starvation and salt stress (N/S). In fact, the co-presence of these two stresses highly inhibited shoot weight accumulation, which was significantly induced in response to APR®.

**Figure 4.** Effects of APR® on shoot weight in response to different single and combined abiotic stresses. Maize seedlings were subjected for 3 days to several abiotic stresses in absence (grey bars) or in presence (black bars) of 5% APR®. The applied single stresses were: hypoxic stress (H), salt stress (S), nutritional stress (N). The single stresses were combined as: hypoxic stress plus nutritional stress (N/H) and salt stress plus nutritional stress (N/S). The values of root weight (g) are represented in the graphs (mean ± SE) and were calculated from three independent experiments (*n* = 10). Significantly different values are evidenced by \* (\*: *p* < 0.05; \*\*: *p* < 0.02; One-way ANOVA, Tukey's HSD test).

#### *3.4. Biostimulant Regulation of Gene Expression*

A number of genes belonging to the group of nitrate transporters and of ROS metabolism were selected basing both on previous results (Trevisan et al. 2017 [20]) and on their putative physiological role (Table 1).


**Table 1.** Description and classification of the targets genes studied in qPCR expression analysis. The expression levels of genes belonging to nitrate transport system (HATS and LATS) and related to reactive oxygen species (ROS) generation and homeostasis were analyzed.

The nitrate transporters genes include five genes encoding putative high affinity components of nitrate transport (three *ZmNRT2* and two *ZmNAR2* genes respectively, Figure 5) and five encoding members of the NRT1 gene family which is involved in the low affinity nitrate transport system (Figure 6). As far as the ROS genes were concerned this group comprise four genes encoding NADPHoxidase (*ZmRBOH a*, *b*, *c* and *d*), one encoding Catalase2 (*ZmCAT2*) and a gene encoding a Cu-Zn Superoxide dismutase (*ZmSOD*).

The specific effects of the biostimulant on the different groups of genes in conditions of different stress are discussed below.

**Figure 5.** The gene expression patterns of nitrate transporters belonging to the High Affinity Transport Systems (HATS) maize gene family in response to different single and combined abiotic stress are influenced by the presence of APR®. Q-PCR analyses were carried out on root of stressed (H, S, N, N/H, N/S) or unstressed (optimal) maize seedlings grown for 3 days in absence (grey bars) or in presence (black bars) of 5% APR®. Relative mRNA level represents data normalized to control (Optimal = 1). The values shown are means of three biological replicates ± SE. Significantly different values (*p* < 0.05) are evidenced by different letters (One-way Anova, LSD post-hoc test).

**Figure 6.** The gene expression patterns of nitrate transporters belonging to the Low Affinity Transport Systems (LATS) maize gene family in response to different single and combined abiotic stress are influenced by the presence of APR ®. Q-PCR analyses were carried out on root of stressed (H, S, N, N/H, N/S) or unstressed (optimal) maize seedlings grown for 3 days in absence (grey bars) or in presence (black bars) of 5% APR®. Relative mRNA level represents data normalized to control (Optimal = 1). The values shown are means of three biological replicates ± SE. Significantly different values (*p* < 0.05) are evidenced by different letters (One-way Anova, LSD post-hoc test).

#### *3.5. Biostimulant Effects on ZmNRT2 Genes Expression*

In general, the most striking effects of APR® on gene transcription regulation in roots were observed in conditions of stress and for the group of genes operating in the high affinity nitrate transport system. In hypoxic conditions, the expression of *ZmNRT2.1*, *ZmNRT2.2, ZmNRT2.3* and *ZmNAR2.1* was down-regulated in response, but when APR® was supplied a significant increase of their transcription was observed (Figure 5).

The same group of genes were, on the contrary, up-regulated in response to salt stress, but the provision of APR® significantly counteracted this effect, leading to restore the phenotype of un-stressed roots (Figure 5).

In the case of nutrient starvation, instead, unique behaviours were observed for each gene belonging to the high affinity nitrate transport group, with *ZmNRT2.1* and *ZmNRT2.2* being significantly up-regulated and, *ZmNRT2.3* and *ZmNAR2.2*, being down-regulated as a consequence of APR® provision (Figure 5).

When seedlings were subjected to a combination of hypoxia and nutritional stress the transcription of *ZmNRT2.1* and *ZmNRT2.3* were clearly induced, whilst *ZmNRT2.2* expression was down-regulated (Figure 5).

The co-presence of nutritional deficiency and salt triggered for all these genes significant dysregulation of transcription. However, APR® provision restrained this outcome for all of them, except for *ZmNAR2.3* which was further induced by APR®. Except for this situation, the transcription of *ZmNAR2.3* evidenced always minor changes in response to both stress conditions and APR® provision (Figure 5).

In the case of *ZmNRT2.1*, *ZmNRT2.3* and *ZmNAR2.2* a significant up-regulation of expression was noticed also in control condition (un-stressed plants),

#### *3.6. Biostimulant Effects on NRT1 Genes Expression*

The transcription of genes implicated in the low affinity transport apparatus was less affected by both stress conditions and APR®, if compared to that of high affinity constituents. Hypoxic conditions induced an increase of the transcription of *ZmNRT1.5* which was significantly counteracted when APR® was provided to the solution (Figure 6). Salt stress triggered an increased transcription of *ZmNRT1a*, which was then inhibited by APR®. The supply of APR® to nutritional starved roots triggered significant change of transcription for *ZmNRT1.1*, *ZmNRT1b*, *ZmNRT* (Figure 6).

APR® significantly affected the expression of these genes, except for *ZmNRT1.5* in plants subjected to a combination of hypoxia and nutrient deficiency (Figure 6). When the combination of nutritional starvation and salt was applied to plants and APR® was supplied to roots significant changes of transcription were noticed, except for *ZmNRT1a* and *ZmNRT1.5* (Figure 6).

#### *3.7. Biostimulant Effects on ROS Genes Expression*

As observed for *NRT1* genes also in this case no regulation of expression was noticed upon APR® supply on unstressed seedlings (Figure 7). A more appreciable effect of the biostimulant was observed upon stress conditions. As reported in Figure 7 *ZmSOD1A* transcription was induced upon APR® treatment in hypoxia, nutritional deficiency, association between hypoxia and nutritional stress and also in the case of nutritional and salt combined stresses. On the contrary no evident alterations of expression were measured for *ZmCAT2* neither in response to stress nor in response to APR® (Figure 7). As far as *ZmRboh* genes were concerned their expression was regulated by APR® in response to single and combined stress, even if to a lower extent in compared to *ZmSOD1A* (Figure 7).

**Figure 7.** The expression patterns of ROS-related genes in response to different single and combined abiotic stress are influenced by the presence of APR ®. Q-PCR analyses were carried out on root of stressed (H, S, N, N/H, N/S) or unstressed (optimal) maize seedlings grown for 3 days in absence (grey bars) or in presence (black bars) of 5% APR®. Relative mRNA level represents data normalized to control (Optimal = 1). The values shown are means of three biological replicates ± SE. Significantly different values (*p* < 0.05) are evidenced by different letters (One-way Anova, LSD post-hoc test).

#### **4. Discussion**

Protein hydrolysates are defined as 'mixtures of polypeptides, oligopeptides and amino acids that are manufactured from protein sources using partial hydrolysis' [29]. Their positive effects on plant performance have encouraged an increasing interest for their use in a more sustainable model of agriculture [30], this leading likewise to a promising solution to the issue of waste disposal [29–33].

Recently a trascriptomic approach was used to study the molecular effects of a collagen derived protein thermal hydrolysate (APR® ) produced by Ilsa S.p.A. (Arzignano) on maize roots grown in a solid medium and supplied with localized patches of APR® [16]. Globally the results allowed to recognize a complex APR® action on physiological pathways involved in the stress response and in nutrient acquisition, which seems likely to prime the plant to better tolerate environmental adversities. In the present work the effectiveness of this biostimulant in modulating and improving the maize tolerance to environmental constrains was tested by growing seedlings in different specific abiotic stress conditions and supplying APR® in hydroponics. Overall our data indicate that this compound is soluble in an aqueous solution, suggesting the idea that in soil it can likely move toward roots through mass flow and diffusion and not only being intercepted as a nutrient patch by root growth.

To choose the most effective APR® concentration on plant development, the root and shoot growth were assessed by determining their fresh weights and the root length upon four different concentrations, chosen on the basis of our previous results [16]. The most accepted scientific definition for biostimulants is: "a plant biostimulant is any substance or microorganism applied to plants with the aim to enhance nutrition efficiency, abiotic stress tolerance and/or crop quality traits, regardless of its nutrients content", as reported in du Jardin [11]. According to this, we tested different solutions containing four APR® amounts to which corresponded four different N sub-nutritional concentrations (1%, 2%, 5% and 10% respect to the control Hoagland solution). The most remarkable effect was observed for primary root growth which was stimulated in response to APR® concentrations ranging from 1 to 5% and then inhibited in the presence of a 10% concentration. Detrimental effects of high concentrations of various protein hydrolysates have been observed also by other authors depending on the crop, the typology of biostimulant and the conditions of application [34].

Furthermore, the present results showed that APR® affects root elongation and gene expression in particular when seedlings were subjected to different kind of stresses, confirming the hypothesis put forward by Trevisan and co-authors [16] and thus supporting the suggestion that biostimulants could act as plant protectors able to improve stress tolerance [11], likely by activating the main signalling pathways underlying the response to adverse conditions. Other reports showed that protein hydrolysates modulate plant growth, increase yield and alleviate the impact of abiotic stress on crops [35,36]. The present results, together with those of Trevisan [16] further suggest that this action could involve the molecular regulation of definite genes.

In general, the combination of two or more abiotic stresses has a detrimental impact on crops that is not predictable from that of each of the stresses composing the combination if applied individually. In recent years stress combination has been acknowledged as a novel state of stress and as a major cause of crop loss worldwide [37–39]. For this reason, we decided to assess the APR® potentiality in alleviating stress impact also in condition of stress combination.

As expected the most striking effect of APR® on growth re-establishment in conditions of abiotic stress was observed for roots, which are the main target for hypoxia, salt and nutrient deprivation stresses. The plastic control of the root development throughout time and in response to endogenous and exogenous stimuli allows plants to efficiently adapt to environmental constraints [40,41]. Root apex is highly responsive to external stimuli and rapidly adjusts its growth to efficiently adapt to environmental constraints and resources availability [19,26,42–47]. In this work a clear induction of primary root growth upon APR® treatment was noticed in all the conditions examined, with the most prominent effect in the case of combination of stresses. The simultaneous presence of nutritional deficiency and salt stress led to the most relevant arrest of growth which was, however, at least partially prevented when plants were supplied with APR®. In this case a similar behaviour was observed also in shoot, leading to hypothesise that APR® is able to act also as a systemic clue, firstly perceived by root cells, but likewise triggering a phenotypic response in shoots. This systemic action could be the outcome of the already hypothesised function of APR® as activator of the stress tolerance [16]. Moreover, it could depend on the protein hydrolysates ability to interfere with hormonal signaling, due to the presence of bioactive peptides (for a review Colla [48]) or aminoacids, as confirmed by the chemical composition of this same compound described by Ertani et al., [15]. Recent transcriptomic findings which highlighted the regulation of hormonal key elements by APR® [16] and a different study aimed to characterize the metabolomic regulation by biostimulant [49] reinforces this hypothesis.

Protein hydrolysates seem to improve nutrient uptake through modifications of root architecture (density, length and number of lateral roots), as well as through complexation of nutrients by peptides and amino acids, and also enhancing microbial activity thus increasing the nutrient availability in soil [11,34]. Moreover, a recent paper [50]. demonstrated that protein hydrolysates modulate plant growth and the expression of key genes in N assimilation (including Nitrate and ammonia transporters) in tomato. However only few information has been obtained on protein hydrolysates regulation of nutrient transport system. To better decipher this last aspect, a number of previously identified by Trevisan et al. [16] target genes involved in nitrate transport were chosen as markers for evaluating the transcriptional effects of the treatment.

Our results evidenced a marked regulation of the transcription of genes encoding members of the high affinity nitrate transport system (HATS, NRT2 and NAR genes), which was particularly relevant in condition of abiotic stresses. The impact of APR® supply on the molecular regulation of the Low Affinity Transport System was less evident, leading to suppose that the provision of APR® mainly affects the functioning of the uptake of nitrate in the range of the High Affinity System, which are recognised to play a crucial role in determining the global Nitrogen Use Efficiency (NUE) in condition of limited nutritional inputs [51].

Trevisan et al. [16] also hypothesised that APR® could activate tolerance pathways, by mimicking the plant responses to environmental stresses, thus priming them against unfavourable conditions through the regulation of enzymes involved in the pathway governing the response to oxidative stress. To deepen this hypothesis the analyses of the expression of six genes involved in ROS signalling and defence was assessed, in condition of stress and in the presence of APR®. Only *SOD1A* showed a clear regulation in response to APR® which almost in all the conditions analysed induced its expression, whilst for the other five genes no significant differences were evidenced upon APR® supply.

Superoxide dismutases (SODs) are key enzymes functioning as the first line of antioxidant defence by virtue of the ability to catalyse the enzymatic dismutation of superoxide to H2O2 [52]. The present result reinforces the hypothesis that APR might preventively prepare plants to oxidative stresses, by enhancing their own detoxifying tools.

In conclusion, basing on the more acknowledged definition of biostimulant [11], our results confirm the effectiveness of APR® as an enhancer of abiotic stress tolerance, thus allowing to definitely include it among the category of biostimulants (Figure 8). Moreover, present results strengthen the importance of root as a target for APR®, which has been proven to affect both root development and transcription of genes involved in Nitrogen Use Efficiency and ROS detoxification. Both these actions could lead to an improved tolerance to abiotic stresses, as nutritional starvation, salt and hypoxia which take place in the soil environment.

These preliminary knowledges should be in the future transferred in field experiment to further assess the APR® usefulness in agriculture.

**Figure 8.** Direct and indirect effects of APR on plant growth in response to single and combined abiotic stresses. APR is perceived by the roots and modulates the root length by balancing the expression of genes involved in nitrate transport and ROS detoxification. APR could be systemically transported to the upper part of the plant, inducing a growth response in the shoot.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4395/9/1/28/s1, Table S1: List of primers used in gene expression analysis.

**Author Contributions:** Conceptualization, S.Q. and A.M.; Methodology, S.T. and A.M.; Data Curation, S.T.; Writing—Original Draft Preparation, S.Q. and S. T.; Writing—Review & Editing, S.Q. and S.T.; Supervision, S.Q.; Project Administration, S.Q.; Funding Acquisition, S.Q.

**Funding:** This work has been supported by ILSA S.p.A and by University of Padova (DOR 2017). S.T. was financed by a grant from ILSA S.p.A.

**Acknowledgments:** The authors are thankful to F. Favaron (Department of Land, Environment, Agriculture and Forestry, University of Padova) for the constructive comments in improving this study. Special thanks to C. Franceschi (ILSA S.p.A.) and C. Manoli (ILSA S.p.A.) who assisted in this work.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

#### *Article*

### **Vegetal-Derived Biostimulant Enhances Adventitious Rooting in Cuttings of Basil, Tomato, and Chrysanthemum via Brassinosteroid-Mediated Processes**

### **Hye-Ji Kim 1,\*, Kang-Mo Ku 2, Seunghyun Choi <sup>1</sup> and Mariateresa Cardarelli <sup>3</sup>**


Received: 23 December 2018; Accepted: 5 February 2019; Published: 10 February 2019

**Abstract:** Plant-derived protein biostimulants exhibit hormone-like activities promoting plant growth and yield, yet detailed investigations on hormonal function have remained limited. This study was conducted to investigate the effects of vegetal-derived-biostimulant on morphological and metabolic changes in cuttings of three herbaceous species demonstrating different rooting ability, basil (*Ocimum basilicum* L.), tomato (*Solanum lycopersicum* L.), and chrysanthemum (*Chrysanthemum indicum* L.), in comparison to auxin. Unrooted cuttings were applied with or without biostimulant (100, 1000, 5000, and 10,000 mg L−1) or auxin [1% indole-3-butyric acid (IBA) plus 0.5% 1-naphthaleneacetic acid (NAA); 100, 200, 300, and 500 mg L<sup>−</sup>1] as a basal quick-dip, stuck into inert media, and evaluated at 20 days after placement under intermittent mist. Both compounds increased adventitious rooting in all cuttings. Biostimulant required a significantly higher threshold for a series of adventitious rooting responses than auxin, and the maximum effectiveness was achieved at 5000 mg L−<sup>1</sup> for biostimulant and 100, 200, and 300 mg L−<sup>1</sup> for auxin in basil, tomato, and chrysanthemum, respectively. Adventitious rooting responses (dry mass and length) to biostimulant showed a gradual logarithmic rise as a function of increasing dosages, which was not in agreement with biphasic dose-response of auxin. Biostimulant significantly increased or tended to increase fine roots in all tested cuttings, which was not consistent with auxin. Relatively high levels of endogenous brassinosteroids (BRs) were present in non-treated cuttings of basil, tomato, and chrysanthemum in decreasing order. Both compounds had no effects or concomitantly increased or decreased BR levels in plant tissues, with fewer effects on basil and tomato, containing high BR levels, but more prominent effects on chrysanthemum, containing relatively low BR levels. Contrasting effects of biostimulant and auxin were found in antioxidant activities, which were promoted by biostimulant but inhibited by auxin either in roots or shoots. These results indicate that the hormonal effects of vegetal-derived biostimulant are primarily exerted by BR-mediated processes while involving interaction with auxin. Both the biostimulant-derived BRs and auxin were suggested to modulate endogenous BR pool via overlapping and interdependent regulatory functions, inducing morphological and metabolic changes during adventitious rooting of cuttings in a plant species-specific manner.

**Keywords:** stem cuttings; propagation; root morphology traits; indole-3-acetic acid (IAA); indole-3-butyric acid (IBA); gibberellins; phenolic compounds

#### **1. Introduction**

Plant-derived biostimulants represent a well-known group of biostimulants and have been proposed as an innovative tool to address the sustainability challenges facing horticulture and to ensure high yield and quality of horticultural commodities [1–3]. Manufactured from plant protein sources using partial hydrolysis, plant biostimulants are considered as a subgroup of growth regulators and bioregulators which are composed of a mixture of polypeptides, oligopeptides, and amino acids [4]. Plant-derived biostimulants are reported to be more effective than animal-derived biostimulants as they contain a higher concentration of amino acids and soluble peptides, with peptides being the principal active compounds [5–7]. Plant-derived biostimulants are defined as materials other than fertilizers that promote plant growth when applied in small quantities or metabolic enhancers [8]. They are available on the market as various forms, including liquid products, soluble powder or in granular form, and were demonstrated to be effective as a seed treatment, foliar spray, and soil drench for crop production [9–11]. When applied as a foliar spray or soil-drench, biostimulants can induce a series of physiological responses in crops changing their phenotypic characteristics and promoting plant growth [10,12].

It has been proposed that such plant responses induced by biostimulants are derived from hormone-like activities and the production of secondary metabolites [13]. Auxin- and gibberellin-like activities were demonstrated in corn coleoptiles and tomato cuttings [5,6], particularly due to the presence of bioactive peptides [14,15]. Peptides are known to be involved in cell differentiation, protease inhibitor induction, cell division, and pollen self-incompatibility response [16,17]. Similar results were reported in degraded soybean meal products, which had promotive effects on root hairs in *Brassica rapa* and tomato cuttings [7].

The positive effects of biostimulant on plant growth and yield have been demonstrated in many studies. The application of biostimulant not only enhanced the growth of corn seedlings [5,6,18] and stem cuttings of tomato [5], but also improved nutrient status, yield, and quality of herbaceous and woody plants, including corn, bean, tomato, sweet yellow pepper, strawberry, banana, papaya, and red grape [5,13,19–22]. It also enhanced tolerance to a wide range of abiotic stresses, such as drought [23], salinity [9], extreme temperatures [24], nutrient deficiency [25], and adverse soil pH [26]. The application of biostimulant increased root morphology, such as root dry mass, total root length, and root surface area, which was associated with improved nitrogen status [5,6]. However, it is not clear how such morphological and physiological changes are induced by the biostimulant.

Adventitious rooting involves significant cellular metabolic activities, leading to the formation of new roots at the base of stem cuttings. Auxin plays a pivotal role in promoting cell growth, cell division, and adventitious root formation in cuttings [27–29] and its mode of action on adventitious rooting is well-elucidated [27,30–32]. Rooting compounds commonly contain indole-3-butyric acid (IBA), 1-naphthaleneacetic acid (NAA), or a combination of the two compounds. The application of auxin to unrooted cuttings promotes adventitious rooting at relatively low doses.

Meanwhile, the mode of action of plant-derived biostimulants on adventitious rooting is largely unknown. An auxin-signaling mediated pathway was proposed to be involved in adventitious rooting of tomato cuttings and their improved nitrogen status, as represented by higher Soil Plant Analysis Development (SPAD) values, following the biostimulant Trainer® treatment [5]. They also found that biostimulant applications increased the shoot elongation rate of dwarf pea plants, prompting the idea of gibberellin involvement in regulating their shoot growth. Unlike auxins, gibberellins are known to inhibit the production of adventitious roots [33–35]. Therefore, stem cuttings provide an ideal experimental system with which mechanistic investigations on hormonal regulations associated with plant-derived-biostimulant can be undertaken. The system eliminates: (1) gibberellic acid as a potential candidate for biostimulant effects due to their antagonistic nature on adventitious rooting and (2) nutritional effects of biostimulant because nutrients are not required for initial stages of adventitious root formation. Nevertheless, carbohydrates play important roles in adventitious rooting, not only by providing energy and carbon chains for biosynthetic processes of new meristems and roots, but also by affecting gene expression, in collaboration with auxin [32].

Recent metabolomic investigations of the hormonal profile on greenhouse melon demonstrated that the application of biostimulant induced upregulation of metabolites related to brassinosteroids (BRs) and their interactions with other phytohormones were postulated to play a critical role in plant growth responses [36]. Similarly, transcriptomic profiles of lateral roots of maize seedlings demonstrated the involvement of BR signal transduction when treated with biostimulant [37]. Meanwhile, the effects of biostimulants were varied by plant species and/or cultivars, growing seasons, and the application method and concentration of the product [38] although the causes for these variations are not clear.

The objectives of the present study were: (i) to examine the hormonal effects of a plant-derived-biostimulant on adventitious rooting in cuttings of three herbaceous plant species with different rooting ability, (ii) to determine dose responses of stem cuttings to biostimulant and auxin, and (iii) to characterize morphological and metabolic changes induced by biostimulant. Cuttings of basil, tomato, and chrysanthemum were treated with biostimulant by a basal quick-dip, and morphological, physiological, and metabolic changes were evaluated to elucidate the hormonal regulation of biostimulant involved in adventitious rooting formation.

#### **2. Materials and Methods**

#### *2.1. Plant Materials*

Based on our preliminary observations on adventitious root formation of herbaceous plants, cuttings of basil (*Ocimum basilicum* L. cv. Genovese), tomato (*Solanum lycopersicum* L. cv. Washington Cherry), and chrysanthemum (*Chrysanthemum indicum* L. cv. Hollister) were chosen in this study for differences in their relative rooting ability: easy-to-root, moderate-to-root, and difficult-to-root, respectively. In general, herbaceous plant species can produce adventitious roots without application of exogenous auxin; however, auxin application is of commercial importance in cutting propagation, because the endogenous level of auxin is critical to increase the ease during root induction period [34]. The experiment was carried out in summer 2017 to spring 2018, in a glass greenhouse situated at Purdue University, West Lafayette, IN (lat. 40N, long. 86W; altitude 188m above sea level).

Seeds of tomato and basil were acquired from a commercial source (Johnny's Selected Seeds, Albion, ME, USA). The seeds were sown and grown in a growth room for 2 to 3 weeks at Plant Growth Facilities. Meanwhile, unrooted cuttings of chrysanthemum 'Hollister' were obtained from a commercial source (Syngenta Flowers, LLC., Gilroy, CA, USA). Immediately upon receipt, a box of stem cuttings was kept in a refrigerator maintained at 5 ◦C. The cuttings were applied with auxin within three days and then stuck into the media and propagated as described below to produce stock plants. Uniform seedlings of basil and tomato and rooted cuttings of chrysanthemum were randomly chosen and transplanted into 2 L plastic containers filled with a commercial potting mix (Fafard 2P Mix; Conrad Fafard, Agawam, MA, USA). Plants were fertigated with acidified water supplemented with a combination of two water-soluble fertilizers (3:1 mixture of 15N–2.2P–12.5K and 21N–2.2P–16.6K, respectively; Everris NA Inc., Dublin, OH, USA) to provide the following (in mg L−1): 150 nitrogen (N), 20 phosphorous (P), 122 potassium (K), 38 calcium (Ca), 15 magnesium (Mg), 0.8 iron (Fe), 0.4 manganese (Mn) and zinc (Zn), 0.2 copper (Cu) and boron (B), and 0.1 molybdenum (Mo). Nitrate form was 76% of nitrogen provided. Irrigation water was supplemented with 93% sulfuric acid (Brenntag, Reading, PA, USA) at 0.08 mL L–1 to reduce alkalinity to 100 mg L–1 calcium carbonate (CaCO3) and pH to a range of 5.8 to 6.2. The stock plants were grown in a glass-glazed greenhouse with exhaust fan and evaporative-pad cooling, radiant hot water heating, and retractable shade curtains controlled by an environmental computer (Maximizer Precision 10; Priva Computers, Vineland Station, ON, Canada). The average day and night temperatures were 23.8 ± 0.8 and 20.3 ± 0.9 ◦C, respectively. The photoperiod was 14-h (0800 to 2200 HR) consisting of natural day lengths with supplemental lighting

using high-pressure sodium (HPS) lamps. A supplemental photosynthetic photon flux (*PPF*) was measured using a quantum sensor (LI-250A light meter; LI-COR Biosciences, Lincoln, NE, USA) and was approximately 150 μmol m−<sup>2</sup> s−<sup>1</sup> at canopy height. The relative humidity inside the greenhouse ranged from 50% to 70% during the study.

Cuttings were taken from the tips of mature stock plants grown in a greenhouse for 2 to 3 months. The cuttings were prepared to have four apical leaves by removing extra leaves from the basal node and trimmed to be uniform in length.

#### *2.2. Biostimulant and Auxin Treatments and Propagation Conditions*

The commercial plant-derived biostimulant Quik-link® (Italpollina S.p.a, Rivoli Veronese, Italy) was used in this study. It contains trace elements (10 g kg−<sup>1</sup> Fe; 7 g kg−<sup>1</sup> Mn; 3 g kg−<sup>1</sup> Zn; 1 g kg−<sup>1</sup> Cu; 0.2 g kg−<sup>1</sup> Mo) and organic compounds biologically active like vegetal amino acids and peptides. The aminogram (expressed as percentage of the total amino acids) is: Ala(4.5), Arg(6.7), Asp(12.7), Cys(1.1), Glu(20.2), Gly(4.5), His(3.0), Ile(4.9), Leu(8.3), Lys(6.8), Met (1.5), Phe (5.6), Pro (5.6), Thr (4.1), Trp (1.1), Tyr (4.1), Val (5.3). The product also contains the Root Hair Promoting Peptide (RHPP) which is a signaling peptide stimulating root growth [36].

Quik-link is allowed in organic agriculture according to the Council Regulation (EC) No. 834/2007 of 28 June 2007 [36], and is manufactured by Italpollina USA Inc. (Anderson, IN, USA). The biostimulant was prepared in five concentrations of 0 (control), 1000, 3000, 5000, and 10,000 mg L<sup>−</sup>1. Meanwhile, a commercial formulation of indole-3-butyric acid (IBA) and 1-naphthaleneacetic acid (NAA) (Dip'N Grow, Inc., Clackamas, OR, USA) was used for auxin treatment, since indole-3-acetic acid (IAA), a naturally occurring compounds, can be easily degraded in the presence of light and is susceptible to destruction in the plant by IAA-oxidase [27–29]. IBA and NAA are more effective than the naturally occurring or synthetic IAA for rooting, and therefore, are the most widely used auxins for rooting stem cuttings [34]. The formulation was prepared in five concentrations of 0, 100, 200, 300, and 500 mg L−1, providing IBA and NAA concentrations at 0, 492 μM IBA + 537 μM NAA, 984 μM IBA + 1074 μM NAA, 1476 μM IBA + 1611 μM NAA, and 2460 μM IBA + 2685 μM NAA, respectively.

Stem base of unrooted cuttings were dipped into a solution of either biostimulant or auxin using a basal quick dip method for 3 s to a depth of 2 cm. The stems were quickly stuck into polystyrene cell packs (300 cm<sup>3</sup> soil volume per cell) filled with inert media (1:1 (v/v) perlite and vermiculite mixture). The cell packs were then placed into polystyrene trays and placed under an intermittent mist, providing bottom heat and overhead mist for 10 s every 20 min during daylight hours with 76 to 98% relative humidity at canopy height for a rooting period of 21 days. The photoperiod was 14-h (0800 to 2200 HR) consisting of natural day lengths with supplemental lighting using high-pressure sodium (HPS) lamps. A supplemental photosynthetic photon flux (*PPF*) was approximately 75 μmol m−<sup>2</sup> s−<sup>1</sup> at canopy height and daily maximum/minimum temperatures in the greenhouse were 23.3 ± 0.8 /22.6 ± 0.7 ◦C.

#### *2.3. Plant Growth Measurements*

When the maximum rooting was observed at day 20, stem length was measured from the stem end to the apical growing point and the number of leaves were recorded. The Soil Plant Analysis Development (SPAD) value, an index of chlorophyll content per unit leaf area, was measured using the SPAD chlorophyll meter (Minolta Corporation, Ltd., Osaka, Japan) on three newly expanded leaves and three fully matured leaves separately, and averaged at each group.

At the end of the rooting experiment, plant parts were separated into leaves, stems and roots. The fresh mass of each part was determined immediately after harvest and were dried in a forced-convection oven at 70 ◦C (Heratherm OMH400, Thermo Scientific Inc., Waltham, MA, USA) for 3 days until a constant weight was reached. Shoot dry mass was calculated as the sum of aerial vegetative parts, and total dry mass was calculated as the shoot and root dry mass. The root-to-shoot ratio was calculated based on the dry mass of roots and shoots. Total plant dry mass was calculated by adding the dry mass of each plant part. Shoot and root dry mass were analyzed by regression as a

continuous response to log [concentration] and by analysis of variance to the treatment. The dried plant tissues were ground in a Wiley mill to pass through a 20-mesh screen, and 0.1 mg samples were weighed and subsequently analyzed for the nitrogen content using a Flash EA elemental analyzer (Thermo Scientific, Waltham, MA, USA).

#### *2.4. Measurements of Root Morphological Traits*

Cuttings were subjected to root morphological analysis at day 20. The number of adventitious roots were counted manually at harvest when the roots were separated from the stems using a razor blade. Entire roots were carefully rinsed and scanned using the Epson Expression 11000XL scanner (Epson America Inc., Long Beach, CA, USA). The debris removal filter was set to discount objects less than 1 cm2 with a length/width ratio less than 4. The scanned images were then used to determine root morphological traits, such as total root length, root surface area, average root diameter, and root volume, using WinRHIZO Pro software (Regent Instrument Inc., Quebec City, QC, Canada). After root images were taken, the roots were weighed and dried in an oven set at 75 ◦C until the samples were completely dry to weigh dry mass. Diameter class length (root length within a diameter class) were generated in the images of adventitious roots acquired from WinRHIZO. The roots were divided into 26 diameter classes at 0.25 mm intervals and root length per each root diameter class was calculated. The root diameter class distribution was computed based on the proportion of the root length in each root diameter class compared to the total root length.

#### *2.5. Primary Metabolite Extraction and Qualitative Analysis from Biostimulant*

Primary metabolites were extracted following published protocols with modifications of extraction solvent volume. Quik-link (0.5 mL) were weighed into 2 mL microcentrifuge tubes, followed by the addition of 0.2 mL of water. To fractionate non-polar compounds, 0.375 mL of cold chloroform (−20 ◦C) and 0.7 mL methanol were added. After vigorous up-and-down mixing by hand (50 times), the extracts were centrifuged at 12,000 × *g* for 4 min, 100 μL supernatant (water soluble metabolites) and organic phase (lipid soluble metabolites) were transferred to 1.5 mL microcentrifuge tubes, respectively. The extracts were dried using Vacufuge concentrator (Eppendorf, Thermo Fisher Scientific, Waltham, MA, USA) with 20 μL of methanol to facilitate water evaporation. For water soluble metabolites, dried extracts were derivatized with 50 μL methoxyamine hydrochloride (40 mg ml–1 in pyridine) for 90 min at 37 ◦C, then with 100 μL MSTFA + 1% TMCS at 50 ◦C for 20 min. For lipid soluble metabolites, dried extracts were derivatized with 200 μL n,o-bis(trimethylsilyl)trifluoroacetamide with 1% of trimethylchlorosilane at 75 ◦C for 30 min. Metabolites were analyzed using a gas chromatography-mass spectrometry (GC-MS) (Trace 1310 GC, Thermo Fisher Scientific, Waltham, MA, USA) coupled to an MS detector system (ISQ QD, Thermo Fisher Scientific, Waltham, MA, USA) and an autosampler (Triplus RSH, Thermo Fisher Scientific, Waltham, MA, USA). A capillary column (Rxi-5Sil MS, Restek, Bellefonte, PA, USA; 30 m × 0.25 mm × 0.25 μm capillary column w/10 m Integra-Guard Column) was used to detect polar metabolites. For water-soluble metabolite analysis, after an initial temperature hold at 80 ◦C for 2 min, the oven temperature was increased to 330 ◦C at 15 ◦C min–1 and held for 5 min. For lipid-soluble metabolite analysis, after an initial temperature hold at 150 ◦C for 1 min, the oven temperature was increased to 320 ◦C at 12 ◦C min−<sup>1</sup> and held for 7 min. Injector and detector temperatures were set at 250 ◦C and 250 ◦C, respectively. An aliquot of 1 μL was injected with the split ratio of 70:1. The helium carrier gas was kept at a constant flow rate of 1.2 mL min−1. The mass spectrometer was operated in positive electron impact mode (EI) at 70.0 eV ionization energy at m/z 40–500 scan range. Metabolite identification was based on the National Institute of Standards and Technology (NIST) library.

#### *2.6. Quantification of BRs in Plant-Derived Biostimulant and Plant Samples*

Campesterol, stigmasterol, and beta-sitosterol were quantified based on GC-MS. Quik-link (0.5 mL) were weighed into 2 mL microcentrifuge tubes, followed by the addition of 0.2 mL of water. To fractionate non-polar compounds, 0.375 mL of cold chloroform (–20 ◦C) and 0.7 mL methanol were added. After vigorous up-and-down mixing by hand (50 times), the extracts were centrifuged at 12,000× *g* for 4 min, and 187.5 μL chloroform layer were transferred to 1.5 mL microcentrifuge tubes. For the quantification of BRs from plant samples, 100 μL of organic phase from the primary metabolite analysis above session was used. The extracts were dried using Vacufuge concentrator (Eppendorf, Thermo Fisher Scientific, Waltham, MA, USA). Dried extracts were derivatized with 200 μL n,o-bis(trimethylsilyl)trifluoroacetamide with 1% of trimethylchlorosilane at 75 ◦C for 30 min. BRs were analyzed using a GC-MS (Trace 1310 GC, Thermo Fisher Scientific, Waltham, MA, USA) coupled to an MS detector system (ISQ QD, Thermo Fisher Scientific, Waltham, MA, USA) and an autosampler (Triplus RSH, Thermo Fisher Scientific, Waltham, MA, USA). A capillary column (Rxi-5Sil MS, Restek, Bellefonte, PA, USA; 30 m × 0.25 mm × 0.25 μm capillary column w/10 m Integra-Guard Column) was used to detect polar metabolites. After an initial temperature hold at 150 ◦C for 1 min, the oven temperature was increased to 320 ◦C at 12 ◦C min−<sup>1</sup> and held for 7 min. Injector, MS detector temperatures were set at 250 ◦C, 250 ◦C, and 300 ◦C respectively. An aliquot of 1 μL was injected with the splitless mode. The helium carrier gas was kept at a constant flow rate of 1.2 mL min−1. The mass spectrometer was operated in positive electron impact mode (EI) at 70.0 eV ionization energy at m/z 45–600 scan range. Metabolite identification was based on standard compounds in comparison with the mass spectra and retention time. The standard BRs were injected from 25 ng mL−<sup>1</sup> to 1000 ng mL−<sup>1</sup> concentrations.

#### *2.7. Amino Acid Quantification of Vegetal-Biostimulant*

To quantify the free amino acid content in the sample, EZ:faast free amino acid for GC-MS kit (Phenomenex, Torrance, CA, USA) was utilized to extract and measure the amino acid concentration. 75 mg of the sample was incubated with 1.5 mL water overnight to extract the free amino acid from the sample. After the 24 h incubation, samples were centrifuged at 12,000 × g for 3 min. Amino acid purification and derivatization were conducted on EZ:faast instruction. The analysis of amino acid was carried out in a gas chromatograph (Trace 1310 GC, Thermo Fisher Scientific, Waltham, MA, USA) coupled to a flame ionization detector (FID), and an autosampler (Triplus RSH, Thermo Fisher Scientific, Waltham, MA, USA). A capillary column (ZebronTM EZ-AAA amino acid GC, Phenomenex, Torrance, CA, USA; 10 m, 0.25 mm) was used. The injection ratio was set at 1:15 and the injection temperature was 250 ◦C. The injection volume was 1.5 μL. The carrier gas was helium and the flow rate was 1.1 mL min<sup>−</sup>1. The column oven was set at 110 ◦C and increased 30 ◦C per minute to 320 ◦C. FID temperature was set at 220 ◦C and the air flow 450 mL min−<sup>1</sup> and the hydrogen flow was 45 mL min<sup>−</sup>1.

#### *2.8. Total Phenolic Content and Antioxidant Capacity*

Total phenolic content and antioxidant capacity were analyzed using methanol extracts that described above primary metabolite analysis based on the published methods [39]. Freeze-dried samples (20 mg) were extracted in 1.4 mL of 100% methanol at 60 ◦C for 10 min. After centrifuge the supernatants were used for the total phenolic content, 2,2 -azino-bis(3-ethylbenzothiazoline-6 sulphonic acid) (ABTS), and 2,2-diphenyl- 1-picryl-hydrazyl-hydrate (DPPH) antioxidant capacity analyses. Various concentrations of vitamin C were used as standard curves for ABTS and DPPH assays [39]. For the DPPH assay, reaction mixtures containing test samples (10 μL) and 190 μL of a 200 μM DPPH in ethanol were incubated at room temperature for 30 min in 96-well plates. The absorbance of the DPPH free radical was measured at 515 nm using an Epoch 2 plate reader (Biotek Instruments Inc., Winooski, VT, USA). Antioxidant data were expressed as vitamin C equivalent concentration (μg g−<sup>1</sup> DW). For the ABTS assay, 7 mM ABTS ammonium salt was dissolved in a potassium phosphate buffer (pH 7.4) and treated with 2.45 mM potassium persulfate. The mixture was then allowed to stand at room temperature for 12–16 h for full color development (dark blue). The solution was then diluted with potassium phosphate buffer until absorbance reached 1.0 ± 0.02 at 735 nm using an Epoch 2 plate reader (Biotek Instruments Inc., Power Wave XS, Winooski, VT, USA). Subsequently, 190 μL of this solution was mixed with 10 μL of the sample extracts. The absorbance was recorded at room temperature after 6 min. Antioxidant data were expressed as vitamin C equivalent concentration (μg g−<sup>1</sup> DW). For total phenolic content, Folin-ciocalteu reagent was used to determine total phenolic content [39]. Each sample (10 μL) was mixed with (100 μL) of Folin-Ciocalteu reagent (0.2 N) followed by 3 min of incubation at room temperature. Then, 90 μL of sodium carbonate (7.5%) was added. After 60 min of incubation in the dark at room temperature, absorbance was obtained at 735 nm. The total phenolic concentration was determined based on a standard curve of gallic acid.

#### *2.9. Experimental Design and Statistical Analysis*

Treatments were arranged in a completely randomized block design. The procedure was repeated at three different time blocks, and each block consisted of 9 treatments and 10 replicates per treatment, amounting to a total of 270 cuttings (90 samples per each plant species) per each time block. All data were subjected to analysis of variance using JMP for Windows, Version 13.2 (SAS Institute Inc., Cary, NC, USA). Polynomial contrasts were used to compare the treatment effects of biostimulant and auxin. Mean separation within each measured parameter was performed by Tukey's honestly significant difference (HSD) test at *p* < 0.05. Regression analysis was carried out to look for trends in response to the concentration for each treatment. Results from the three experiments showed similar trends and the data sets were consistent with each other. However, because the error variance was not homogeneous between experiments, statistical analyses were conducted separately for each experiment, and data from the two trials were pooled and presented here.

#### **3. Results**

#### *3.1. The Effects of Biostimulant on Adventitious Rooting in Cuttings of Basil, Tomato, and Chrysanthemum*

All cuttings achieved 100% rooting regardless of plant species and treatment. The average number of adventitious roots in untreated cuttings of basil, tomato, and chrysanthemum were 39, 22, and 16, respectively, demonstrating genetic variations in rooting ability. Total root length was higher in the order of tomato, basil, and chrysanthemum (Table 1).

Meanwhile, both biostimulant and auxin increased adventitious rooting in a dose-dependent manner: The number of adventitious roots, root dry mass, and total root length in all plant species increased or showed an increasing trend by higher concentrations of biostimulant and auxin, with exception of overdoses (IBA + NAA0.5) in auxin (Table 1). However, the response level of plant species varied significantly by the treatments. Rooting response increased more prominently by auxin than biostimulant. An optimal level of auxin to induce rooting was highly plant-species specific and maximum root length was achieved nearly at concentrations of 100, 200, and 300 mg L−<sup>1</sup> in basil, tomato, and chrysanthemum, respectively (Figure 1f). When compared to auxin, biostimulant was required approximately 15- to 50-time higher concentrations to induce the onset of adventitious root formation. In general, the application of biostimulant at a concentration of 5000 mg L−<sup>1</sup> increased total root length in all tested cuttings (Table 1). An overdose of auxin tended to negatively affect the dry mass of adventitious roots in basil and tomato to the levels of unrooted cuttings but not in chrysanthemum. Such response was contrasting to biostimulant, where higher concentrations of biostimulant tended to increase or gradually increased adventitious rooting, and even the highest concentration at 10,000 mg L−<sup>1</sup> did not negatively affect root morphological characteristics (Table 1). The root dry mass was positively correlated with the total root length of cuttings treated with either biostimulant or auxin (Figure S1); however, the relationship between dry mass and total root length slightly varied among plant species and between the treatments (Figure S1), indicating that dry mass and/or total root length do not precisely predict response levels of cuttings to biostimulant and auxin applications.

**Figure 1.** Dose response curves showing the effects of biostimulant or auxin applications on shoot and root dry mass, and total length of adventitious roots in cuttings of basil, tomato, and chrysanthemum. Each data point is the mean ± SE of 20 replicates.

**Table 1.** Effects of biostimulant (B) or auxin (IBA + NAA) applications on adventitious root characteristics in cuttings of basil, tomato, and chrysanthemum. Cuttings were treated with or without biostimulant (100, 1000, 5000 (B5), and 10,000 (B10) mg L−1) or auxin (100, 200 (IBA + NAA0.2), 300, and 500 (IBA + NAA0.5) mg L<sup>−</sup>1) at the stem base as quick-dip, stuck into inert media, and evaluated at day 20 after placement under intermittent mist. Note that two concentrations per each treatment were presented here.


ns, \*, \*\*, and \*\*\* indicate non-significant, or significant at *p* < 0.05, 0.01, and 0.001, respectively. Different letters within each column indicate significant differences according to Tukey's HSD test (*p* = 0.05). Data are means of 20 replicates.

#### *3.2. The Effects of Biostimulant on Root Diameter Class Distribution*

Average diameters of adventitious roots varied among plant species. Tomato cuttings had relatively fine roots with an average root diameter of 0.38 mm, while the roots of basil and chrysanthemum composed of coarser roots with average root diameters of 0.55 and 0.63 mm, respectively (Table 1). In tomato, most of the roots were in the finer root class ranging from 0.0 to 0.50 mm, accounting for 75% of total root length (Table 2). The roots of basil consisted of a mixture of finer root diameter classes: about 43% of the total roots were in the finer root class (0.0 to 0.50 mm) and about 41% were intermediate root class (0.50 to 0.75 mm) (Table 2). On the other hand, chrysanthemum produced a wide range of root diameter classes (0.0 to 3.0 mm). About 40% of the total roots were in the finer root class, while the rest of the roots were composed of coarser roots (>0.5 mm). Unlike basil and tomato, where the roots thicker than 1 mm were only a small fraction among the roots (less than 3 and 0.5%, respectively) and were primarily proximal near the stem, more than 10% of the total roots of chrysanthemum were in the coarse root class (>1.0 mm).

Root diameter class distribution analyses revealed treatment differences even within the same plant species (Table 2). In basil and tomato, increasing biostimulant concentrations promoted fine roots (0.0 to 0.25 mm). These results were contrasting to auxin-treated cuttings, in which higher auxin concentrations had an increasing trend of promoting coarser roots (0.50 to 1.00 mm). The response of chrysanthemum roots was quite different from those of basil and tomato: Auxin had more pronounced effects on changing root morphological traits in chrysanthemum, and an optimal auxin concentration (IBA + NAA0.2) significantly promoted the formation of fine roots (0 to 0.25 mm) while decreasing coarser roots (>0.75 mm). Likewise, vegetal-biostimulant tended to promote finer roots in chrysanthemum, but to a lesser degree than auxin.

**Table 2.** Root diameter class (mm) and relative diameter class length (%) of cuttings of basil, tomato, and chrysanthemum. Cuttings were treated with or without biostimulant (100, 1000, 5000 (B5), and 10,000 (B10) mg L−1) or auxin (100, 200 (IBA + NAA0.2), 300, and 500 (IBA + NAA0.5) mg L–1) at the stem base as quick-dip, stuck into inert media, and evaluated at day 20 after placement under intermittent mist. Note that two concentrations per each treatment were presented here. Percentage values at each diameter class are given.


ns, \*, \*\*, and \*\*\* indicate non-significant, or significant at *p* < 0.05, 0.01, and 0.001, respectively. Different letters within each column indicate significant differences according to Tukey's HSD test (*p* = 0.05). Data are means of 20 replicates.

#### *3.3. The Effects of Biostimulant on Shoot Growth*

Consistently with adventitious rooting, response levels of shoots to biostimulant and auxin slightly varied among plant species (Table 3; Tables S1–S3). Increasing concentrations of biostimulant increased or showed an increasing trend of shoot dry mass (Table 3). Biostimulant increased shoot dry mass in cuttings by 10 to 20% at a concentration of 5000 mg L<sup>−</sup>1, which was somewhat associated with the increase in leaf or stem dry mass of the cuttings. Contrarily, auxin did not affect shoot dry mass of cuttings in basil and tomato with exception of chrysanthemum. SPAD index measured on newly expanded leaves and three fully matured leaves were not significantly different among the treatments, and therefore, pooled for comparisons. The results showed that SPAD index increased only in chrysanthemum when applied with auxin at an optimum level (IBA + NAA0.2), but there were no differences in total nitrogen (N) concentration among treatments (Table 3).

**Table 3.** Effects of biostimulant (B) or auxin (IBA + NAA) applications on stem length, leaves, stems, and shoot dry mass, the Soil Plant Analysis Development (SPAD) index, total nitrogen (N), and root-to-shoot ratio of basil, tomato, and chrysanthemum cuttings. Cuttings were treated with or without biostimulant (100, 1000, 5000 (B5), and 10,000 (B10) mg L<sup>−</sup>1) or auxin (100, 200 (IBA + NAA0.2), 300, and 500 (IBA + NAA0.5) mg L−1) at the stem base as quick-dip, stuck into inert media, and evaluated at day 20 after placement under intermittent mist. Note that two concentrations per each treatment were presented here.


ns, \*, \*\*, and \*\*\* indicate non-significant, or significant at *p* < 0.05, 0.01, and 0.001, respectively. Different letters within each column indicate significant differences according to Tukey's HSD test (*p* = 0.05). Data are means of 20 replicates.

#### *3.4. Tissue Responsiveness to Biostimulant and Auxin*

Cuttings of basil, tomato, and chrysanthemum differently responded to biostimulant in comparison to auxin. Biostimulant induced gradual and progressive changes on shoot and root dry mass, as shown in the logarithmic curves (Figure 1a,c), and there were no detrimental effects or phytotoxicity caused by higher doses of biostimulant. This was contradictory to auxin-treated cuttings where higher doses had negative effects on rooting responses, especially in basil and tomato (Figure 1b,d). The rooting responses, as expressed as root dry mass and total root length, were less dramatically influenced by biostimulant than by auxin in all plant species tested.

Basil cuttings were more responsive to a lower concentration of auxin compared to tomato and chrysanthemum, rapidly increasing adventitious roots (Figure 1d,f). At an optimal concentration, auxin-treated cuttings produced similar or higher root biomass relative to biostimulant-treated cuttings (Figure 1c,d). Regression analyses showed that auxin responsiveness of plant species increased in a biphasic manner with increasing concentrations (Figure 1d,f). The response pattern of adventitious rooting to auxin showed that basil and tomato were highly responsive. Basil responds to a lower threshold for rooting followed by a rapid polynomial decay, while tomato required a higher threshold than basil (Figure 1d). Chrysanthemum responded to a lower threshold for rooting, but displayed a gradual polynomial rise to a wide range of auxin, possibly followed by a gradual polynomial fall to a higher concentration of auxin. This rooting response was associated with increased total root length in all the plant species tested (Figure S1). However, a universal scenario of increased root dry mass and/total root length accompanied by biostimulant treatment does not explain the root morphological changes as represented by the proliferation of fine roots, as such subtle changes contribute less to dry mass. Shoots were slightly less responsive to biostimulant applications than roots as characterized by a gentle slope in a logarithmic plot (Figure 1a). Shoot dry mass of basil and tomato did not increase even when a wide range of auxin was applied; however, that of chrysanthemum increased with higher doses of auxin, indicating that chrysanthemum had different responsiveness to auxin from basil and tomato (Figure 1b).

#### *3.5. BRs in Roots and Shoots of Cuttings*

Metabolic analyses demonstrated that biostimulant contained precursors of BRs, such as campesterol and stigmasterol (or β-sitosterol) at 11.87 and 28.86 ng mL−<sup>1</sup> in 5000 mg L−<sup>1</sup> solution. In addition, a large profile of various compounds, including sugars (ribofuranose, arabinose, and galactose), organic acids (lactic, oxalic acid, glycolic acid, butanoic acid, tartaric acid, and gluconic acid), glucono-1,4-lactone, and fatty acids (palmitic acid and stearic acid) were found to be present as major compounds in the biostimulant (data not shown).

Metabolic profiling of cuttings elucidated that relatively high levels of BRs were present in non-treated cuttings of basil, tomato, and chrysanthemum in decreasing order. Total sterol levels were higher in roots than shoots by 3.4-, 5.3-, and 1.4-times in basil, tomato, and chrysanthemum, respectively) with the highest concentration in roots of basil, tomato, and chrysanthemum in decreasing order, which averaged at 1126, 397, and 213 μg g−<sup>1</sup> dry weight, respectively (Table 4). There were three major phytosterols present in these plant species: stigmasterol, beta-sitosterol, and campesterol (Table 4). Overall, the combined proportions of stigmasterol and sitosterol were more than 80% of the total sterols, and the proportion of campesterol was less than 20%.

Notably, biostimulant and auxin treatments concomitantly increased or decreased BR levels in plant tissues or had no effects on the levels. Stigmasterol levels in roots tended to be affected by the treatments in all the tested crops (*p* ≤ 0.12); however, in a different manner. For example, in basil and chrysanthemum, the optimum level of biostimulant tended to increase stigmasterol levels in roots, while, biostimulant significantly (*p <* 0.01) reduced the levels in tomato. Auxin had similar effects as biostimulant on stigmasterol levels in roots. Correlation relationships were determined between BRs and growth parameters of basil, tomato, and chrysanthemum, i.e., root dry mass, total root length, length of fine roots (0.00 to 0.25 mm) and shoot dry mass. Overall, total sterol levels were not correlated or weakly correlated with root growth parameters in basil and tomato, but moderately correlated (e.g., root dry mass: *r*<sup>2</sup> = 0.51, *p* < 0.001; root length: *r*<sup>2</sup> = 0.27, *p* < 0.01) in chrysanthemum.

Total sterol levels in shoots were the highest in basil, chrysanthemum, and tomato in decreasing order, and averaged at 327, 155, and 75 μg g−<sup>1</sup> dry weight, respectively (Table 4). Both biostimulant and auxin treatments appeared to have similar increasing or decreasing effects on the levels of BRs as observed in roots. Sitosterol levels were significantly increased in shoots of tomato and chrysanthemum cuttings by biostimulant. Auxin had similar increasing effects on the levels of sitosterol in shoots of those cuttings.



ns, \*, \*\*, and \*\*\* indicate non-significant, or significant at *p* < 0.05, 0.01, and 0.001, respectively. Different letters within each column indicate significant differences according to Tukey's HSD test (*p* = 0.05). Data shown are means ± SE of five replicates. a Significant at *p* ≤ 0.12. a Significant at *p* ≤ 0.1.

#### *3.6. Antioxidant Capacities and Total Phenolic Content of Cuttings*

The radical scavenging activities of roots and shoots in cuttings of basil, tomato, and chrysanthemum were examined as estimated by the DPPH (Figure 2a,d,g) and ABTS assays (Figure 2b,e,h). The DPPH method is one of the most frequently used and inexpensive antioxidant assays; however, pH sensitivity is a major disadvantage of the assay [40]. In order to generate robust results, we used the two different scavenging radical assays in this study. The antioxidant capacities obtained from DPPH assay were in accordance with those obtained from ABTS assays regardless of plant species and tissue type (basil: roots *r*<sup>2</sup> = 0.94, shoots *r*<sup>2</sup> = 0.79; tomato: roots *r*<sup>2</sup> = 0.75, shoots *r*<sup>2</sup> = 0.82; chrysanthemum: roots *r*<sup>2</sup> = 0.94, shoots *r*<sup>2</sup> = 0.91).

The results showed large variations in antioxidant capacities among plant species and tissues. Roots of basil showed the highest antioxidant activities (3.6 mg ascorbic acid equivalent per g DW) followed by chrysanthemum (2.3 mg) and tomato (0.4 mg) (Figure 2). In basil, antioxidant capacities were three-times higher in roots than shoots and were less affected by the treatment (Figure 2a,b),

while tomato and chrysanthemum was approximately two-times lower in roots than shoots and were either positively or negatively affected by the treatment (Figure 2d,e,g,h).

Biostimulant significantly increased the scavenging activities of roots of chrysanthemum, and such increases were strongly correlated with the concentrations of total phenolic acids (*r*<sup>2</sup> = 0.77 for DPPH and *r*<sup>2</sup> = 0.78 for ABTS) (Figure 2g–i). Meanwhile, shoots of chrysanthemum behaved differently and demonstrated significantly higher antioxidant activities concomitantly by both biostimulant and auxin compared to control. There was a trend of concentration-dependent increase in radical scavenging activities by biostimulant in shoots and roots of basil and chrysanthemum (Figure 2). Overall, biostimulant had stimulatory effects on antioxidant activities of adventitious roots in cuttings (based on DPPH and ABTS assays), and such results were contradictory to those induced by auxin.

**Figure 2.** Antioxidant capacity and total phenolic compounds of roots and shoots in cuttings of basil (**a**–**c**), tomato (**d**–**f**) and chrysanthemum (**g**–**i**) at day 20 after treatment with either control (Cont), biostimulant at 5000 (B5) or 10,000 mg L−<sup>1</sup> (B10), or IAA+NAA at 200 mg L−<sup>1</sup> (A0.2). Antioxidant capacity was estimated by the DPPH and ABTS assays. The antioxidant capacity and total phenolic compounds of the aqueous extracts are equivalent to indicated concentrations of water-soluble standard antioxidant ascorbic acid (mg g−<sup>1</sup> DW) and gallic acid (mg g−<sup>1</sup> DW), respectively. Different letters indicate significant differences within each plant part (roots or shoots) according to Tukey's HSD test (*p* = 0.05). Data shown are means of five replicates.

#### **4. Discussion**

#### *4.1. Biostimulant Promotes Adventitious Rooting Responses of Stem Cuttings Similar to Auxin, but to a Lesser Extent*

It is well established that auxin promotes the formation of adventitious roots [41,42] and lateral roots [43–45]. Previous studies have reported on root morphological changes induced by biostimulant applications; however, the changes have been focused primarily on the increases in root biomass, total root length, and root surface area [5,6] without sufficient information on root characteristics. Further, no detailed investigations have been made on hormonal effects of biostimulant in promoting plant growth and yield although these aspects have been demonstrated in many studies.

First, we measured morphological responses of stem cuttings, including the number of adventitious roots, root dry mass, total root length, and average root diameter, as a function of the exogenous concentration of either biostimulant or auxin. While most of these variables exhibited plant species-specific responses, partly due to the genetics of differential tissue responsiveness, it was clear that biostimulant was effective in promoting adventitious rooting of basil and tomato, easy-to-root types, as well as chrysanthemum, moderate-to-root type.

Dose-response analyses were used to evaluate the relationship between compound dosage and plant response (Figure 1), and it was found that rooting responses to biostimulant not only were considerably less compared to auxin but also did not fully in agreement with those to auxin. Biostimulant promoted adventitious rooting leading to a gradual logarithmic rise as a function of increasing dosages, while auxin induced a biphasic dose response characterized by rapid polynomial rise and fall in a plant species-specific manner. A significantly higher threshold was required for biostimulant to induce a series of responses compared to auxin. One of the reasons for the mild changes over a wide range of concentrations induced by biostimulant in cuttings is partly due to a basal quick dip method employed in this study. It was confirmed that such an approach eliminates the possibility of biostimulant as a nutrient source, since there were no differences in total nitrogen level regardless of treatments (Table 3). Changes in endogenous auxin pool may be another possibility because biostimulant Quik-link we used in this study contained about 4.1% tryptophan, as well as other amino acids. As a precursor for auxin biosynthesis pathways in plants, tryptophan might have exerted a weak auxin-induced process, which was postulated in maize seedlings treated with animal-based biostimulant [6]. We did not quantify IAA and other auxin derivatives from plant samples, and therefore, it was not possible to determine how biostimulants interact with endogenous auxin in adventitious rooting formation. Nevertheless, a similar but somewhat unique behavior of biostimulant-treated cuttings observed in our study cannot be justified solely by auxin-mediated activities, opening the possibility of other hormonal regulation in this process.

Variations in adventitious rooting responses were also observed in root architectural traits. An adventitious root system has two major components of root: long, relatively thick roots arising either from the cut stem end or the lower part of the stem that forms its framework and shorter, fine lateral roots arising either directly from these framework roots or indirectly as higher-order lateral roots. Since the complete physical separation of lateral roots from adventitious roots was not possible, particularly in basil and tomato due to fibrous nature of their roots, we performed root diameter class distribution analyses to differentiate these root components by carefully manipulating parameters. This method has been proven to be effective in separating different root types [46], and commonly used in root studies. The results revealed that the roots examined in our study are actually classified into very fine (<0.5) to fine (0.5–2 mm) [47] and we further classified them into multiple categories within the range. The adventitious roots of untreated tomato cuttings were composed primarily of finer roots (average root diameter: 0.38 mm) with about 76% of the total roots within 0 to 0.50 mm diameter class, whereas those of basil cuttings consisted primarily of fine to intermediate roots with about 70% of the roots within 0.25 to 0.75 mm diameter class (average root diameter: 0.55 mm). More than 30% of

the total roots in chrysanthemum were composed of coarse roots (>0.75 mm) (average root diameter: 0.63 mm) with a wide range of root diameter classes (Table 2).

Interestingly, biostimulant applications significantly increased or tended to increase finer root classes in these plant species, providing direct evidence that biostimulant stimulates proliferation of lateral roots in cuttings. This is in agreement with a recent study in maize seedlings, in which protein hydrolysates increased length and surface area of lateral roots by about 7 and 1.5 times compared to inorganic nitrogen and free amino acids, respectively [48]. Fine roots are considered to be the most permeable part of a root system and play the key role in the acquisition of water and nutrients and root adaptation to extreme environments, particularly in herbaceous plants [49] and such developmental changes may confer significant advantages on long-term plant growth and survival, particularly under suboptimal water and nutrient conditions.

#### *4.2. Biostimulant Induces Adventitious Rooting of Stem Cuttings Primarily via BR-Mediated Processes*

In this study, we measured metabolic responses, including BR levels in plant tissues, antioxidant capacities, and total phenolic compounds in roots and shoots of basil, tomato, and chrysanthemum. We found that biostimulant negatively or positively affects BR biosynthesis in plant tissues and increases antioxidant activities and total phenolic compounds in both roots and shoots of cuttings. Consistently with morphological traits, these metabolic responses were not fully in agreement with auxin.

As a group of steroidal plant hormones, BRs are known to mediate modulation of various components of the antioxidant defense system in plants under abiotic stresses, including drought, salinity, and temperature extremes [50]. BRs were reported to be involved in mitigating the adverse effect of high temperature stress on snap bean plants by increasing total free amino acids in leaves and total phenolic acids in the pod [51]. Increases in antioxidative capacities and phenolic compounds were also found in BR-treated *Brassica junica* seedlings under lead toxicity [52]. The BR-mediated antioxidant system was also demonstrated to modulate root growth as the *Arabidopsis det2-9* mutant defective in BR biosynthesis exhibited inhibited root growth and accumulated more reactive oxygen species than the wild type [53]. We found that stigmasterol, sitosterol, and campesterol were the major phytosterols in cuttings of plant species tested. These phytosterols serve as precursors for BR biosynthesis and are integral membrane components which regulate the permeability and fluidity of membranes [54], and phytoserol composition in the plasma membrane affects the proper functioning of auxin transporters [55]. Campesterol influences the level of active BR, and regulates a number of physiological activities in plant development, such as cell elongation, xylem differentiation, and stress tolerance [54].

Herein, we postulate that biostimulant induces adventitious root formation primarily via BR-mediated processes while interacting with auxin-mediated mechanisms and that native BR pool in plant tissues influences adventitious rooting responses to biostimulant. There are at least six pieces of evidence to support this view: (1) endogenous auxin plays the key role in adventitious rooting formation in these cuttings as adventitious roots were produced in cuttings that did not receive any treatment, (2) endogenous BRs also play a critical role in adventitious rooting formation in these cuttings as relatively high levels of native BRs were present in cuttings that did not receive any treatment, (3) both biostimulant and auxin influenced endogenous BR levels in most cuttings, (4) biostimulant exerted weaker effects on adventitious rooting of cuttings than did auxin treatment, (5) adventitious rooting responses to biostimulant was most prominent in chrysanthemum cuttings that have relatively low native BR levels and are less responsiveness to auxin, and (6) antioxidant activities in adventitious roots tended to be increased by biostimulant but decreased by auxin.

As discussed earlier, the extent to which the increased induction and formation of adventitious rooting varied greatly in response to the compound, with more prominent effects by auxin than biostimulant (Figure 2). For example, the optimal levels of auxin and biostimulant increased the dry mass of adventitious roots by 54% and 20% in chrysanthemum, 67% and 26% in tomato, and 42% and 26% in basil, compared to untreated cuttings. Further, major differences between auxin and biostimulant existed not only in the patterns of dose response curve, but also in the absolute amount of the compounds required for promoting adventitious rooting (Figure 1).

Clouse et al. [56] noted that measurable effects on cell elongation induced by BR required much longer treatment time compared with the rapid effects caused by auxin. Nemhauser et al. [57] elucidated that auxin-response element ARFAT is the crucial intersection point of BR and auxin pathways, which is BR responsive and requires BR biosynthesis for normal expression. These findings are consistent with our observations and support our interpretations that the hormonal effects induced by biostimulant is more likely to be related to BRs than auxin and that auxin and BRs interact in controlling BR pool in plant tissues and work coordinately in fine-tuning adventitious rooting responses of cuttings.

#### *4.3. Biostimulant-Induced BRs and Auxin Have Overlapping Functions in Adventitious Root Formation*

We found that roots and shoots of cuttings produce relatively high levels of endogenous BRs which were increased or decreased concomitantly by biostimulant and auxin. This similar effect of both compounds demonstrates that their overlapping role in BR biosynthesis. Although auxin was not quantified in this study, there is no doubt that auxin plays an important role in adventitious rooting formation in these tested plant species. Auxin and BRs are two important phytohormones and are known to exert some similar physiological effects exclusively or through their functional interaction, which include cell division and expansion, vascular differentiation, root growth, and senescence [58]. It was reported that a shared auxin and BR pathway is required for seedling growth, and response from one pathway requires the function of the other, and this interdependence occurs at gene expression level [57]. Consistently, auxin-treated cuttings in our study showed increased levels of BRs, indicating that auxin treatments in cuttings also involve BR biosynthesis. The extent of increased response levels and more remarkable effects on rooting responses induced by auxin indicate that auxin triggers cellular and molecular responses of adventitious rooting synergistically and interdependently from BRs. While such synergistic and interdependent interactions of auxin and BRs have been demonstrated in other plant systems and was well reviewed by Tian et al. [59], this is the first time demonstrating the interactions between biostimulant-induced BRs and auxin in adventitious rooting responses of cuttings. The interaction also includes the lateral root formation of an adventitious root system. BRs are required for lateral root development in Arabidopsis and act synergistically with auxin to promote lateral root formation by increasing acropetal auxin transport [58,60]. BRs mainly function at the lateral root primordia initiation while auxin is required for both initiation and emergence stages of lateral root formation [43,58].

Based on this view, various responses of plant species to biostimulant and auxin can be explained by endogenous BR pools of plant species. The application of biostimulant and auxin had negative effects on BR levels in basil and tomato, highly responsive plant species containing a higher level of native BRs, but had positive effects on BR levels in chrysanthemum, less responsive plant species containing lower BR levels (Table 4). We also demonstrated that a high level of auxin has an inhibitory effect on antioxidant capacities and phenolic compounds in chrysanthemum cuttings (Figure 2). A similar inhibitory effect of increased auxin levels on BR-induced growth responses was observed in auxin-overproducing *yucca* mutants [57]. Thus, it is likely that different plant species have a different level of BR-pool which restricts plant growth response to these compounds, and that increased auxin levels saturate the BR-pool, significantly reducing BR-effects on regulatory changes.

The induction phase in cuttings or detached organs, such as leaves, is generally marked by the immediate consequences of the wounding response caused by severance. It encompasses the first hours after cutting removal, with a local increase in jasmonate, phenolic compounds and auxin at the cutting base [32]. Phenolic compounds exert antioxidant properties against oxidative stress [61], and were demonstrated to promote adventitious roots of stem slices from apple microshoots by protecting IAA from decarboxylation and the tissue from oxidative stress caused by wounding [62], contributing to the auxin stability for adventitious root induction [62]. The high positive correlation (*p* < 0.001) between antioxidant capacities and total phenolic content indicates that phenolic compounds are a major contributor to the antioxidant activities of these plants. Phenolic compounds act as antioxidants protecting auxins from decarboxylation and the tissue from oxidative stress, allowing more auxin is available to induce roots [62].

In addition to BR-related proteins, other plant hormone related proteins were identified when maize seedlings were exposed to protein hydrolysate [37]. Metabolomics studies of greenhouse melons treated with biopolymer-based biostimulant as substrate drench demonstrated that BRs interact with other hormones in the leaves, possibly via translocation from roots, as the compounds related to other hormones were observed in the leaves [36], and this translocation may explain the lower level of BRs in shoot of cuttings observed in our study. These findings suggest that there are cross-talks among hormones during the adventitious rooting process. BRs positively regulate lateral root formation whereas cytokinin and abscisic acid negatively regulate the event, and ethylene has positive and negative roles during lateral root formation [60]. On the contrary, the root growth-stimulating effect of BRs was proposed to be independent of auxin and gibberellin action, in which processes genes related to other phytohormones did not show changes, suggesting that the stimulatory effect of BRs on root growth is an autonomous effect rather than cross-talks with other phytohormones [63].

Relatively little is known about the effects of BRs on growth and development of adventitious roots. There are only a few reports showing that BRs mainly inhibit adventitious root development in cuttings of tomato and mung bean at low concentrations (0.1 μM), but the effects mainly occur on the shoot [64,65]. BRs have shown to be involved in jasmonate signaling and exert a mild negative regulation of jasmonate-induced inhibition of root growth [66]. Increase in adventitious root formation in geranium stem cuttings were observed in the treatment with BRs which also improved shoot growth of coleus cuttings [67,68].

Our results provide evidence that adventitious rooting responses of cuttings treated with biostimulant involve BR biosynthesis and their overlapping function with auxin, leading to the morphological and metabolic changes occurring during adventitious root formation. Due to the short-term investigations on adventitious rooting processes, we did not find subsequent effects of morphological changes occurring in roots and shoots. Yet, it is expected that such developmental changes improve crop performance and resource acquisition under suboptimal water and nutrient environment and confer significant advantages on long-term plant growth and survival, particularly under abiotic stresses.

#### **5. Conclusions**

To elucidate the hormonal effects of plant-derived-biostimulant, adventitious rooting responses of cuttings were examined after a basal quick-dip treatment with various concentrations of biostimulant in comparison to auxin. This approach allows detailed investigations on the hormonal function of biostimulant as auxin is known to play a key role in adventitious rooting process and eliminates potential nutrient effects of the compound. Biostimulant exerted similar effects as auxin increasing adventitious rooting responses. Dose-response analyses revealed that biostimulant showed a gradual logarithmic rise as a function of increasing dosages, contrary to a typical biphasic dose response of auxin, and required a significantly higher threshold than auxin. Metabolic profiles showed that BRs were highly present in non-treated cuttings of basil, tomato, and chrysanthemum in decreasing order, and both biostimulant and auxin had fewer effects in basil and tomato, high BR producers, and greater effects in chrysanthemum, less BR producer, indicating that native BR-pools of plant species influence adventitious rooting responses to biostimulant, as well as auxin. Biostimulant promoted antioxidant activities and phenolic compounds in cuttings, particularly in chrysanthemum, while auxin inhibited these metabolic responses. The inhibitory effect of auxin is likely due to the saturation of BR-pool, significantly reducing BR-effects. These provide evidence that biostimulant has overlapping functions with auxin in adventitious root formation, while exerting distinctive and independent contributions. We demonstrate for the first time that biostimulant induces adventitious rooting responses of cutting

via BR-mediated processes while interacting with auxin and that there are interdependent effects of BRs and auxin on antioxidant activities of cuttings. Our results provide new insight into the hormonal regulation of biostimulant and a fine-tuning role of BRs in adventitious root formation.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4395/9/2/74/s1, Figure S1. The relationship between root dry mass and total root length of basil, tomato, and chrysanthemum cuttings as affected by biostimulant and auxin applications, Table S1. Polynomial contrasts on the means of adventitious root number, root dry mass, total root length, total root surface area, root volume and average root diameter of basil, tomato, and chrysanthemum cuttings as affected by biostimulant and auxin applications, Table S2. Polynomial contrasts on the means of root diameter class (mm) and relative diameter class length (%) of basil, tomato, and chrysanthemum cuttings as affected by biostimulant and auxin applications. Percentage values at each diameter class are given, Table S3. Polynomial contrasts on the means of on stem length, leaves, stems, and shoot dry mass, SPAD index, and root-to-shoot ratio of basil, tomato, and chrysanthemum cuttings as affected by biostimulant and auxin applications. Percentage values at each diameter class are given.

**Author Contributions:** H.K. coordinated and supervised the research, provided intellectual inputs for defining the experimental design, data analysis, and interpretation, and wrote the entire manuscript. K.K. performed metabolic analysis and contributed on metabolomics results, S.C. conducted experiments, collected data, assisted in data analysis and interpretation, and implemented the manuscript. M.C. gave support in the experimental design, data analysis, and interpretation, and implemented the manuscript.

**Funding:** This research was supported by the USDA National Institute of Food and Agriculture, Hatch/Multi State project NE-1335 Resource Management in Commercial Greenhouse Production; Purdue University Research Funds.

**Acknowledgments:** We are grateful to Rob Eddy, Michael Russell, Nathan Deppe, and Dan Little for their technical assistance; Hye Su Lee and Gaotian Zhu for their help with crop management and data collection; and Hélène Reynaud for the helpful discussions and support on the research.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
