**1. Introduction**

Eggplant (*Solanum melongena* L.) is a major solanaceous vegetable crop cultivated worldwide [1]. Eggplant originated in the Old World, evolved from *S. insanum*, and it was self-sufficiently domesticated in India and China [2,3]. With a global production around 49.5 Mt, predominantly in Asia [4], eggplant is a relevant source of minerals and vitamins, and in terms of total nutritional value, it has been compared to tomato [1]. *S. macrocarpon* L., also known as African eggplant, and *S. aethiopicum* L., also known as Aethiopicum eggplant, are closely related to *S. melongena* and are mainly distributed in the African continent. In the eggplant cultivation scenario, the Mediterranean Basin represents an important area and Sicily is counted as a secondary eggplant diversification zone [5]. In Italy,

eggplant is grown under either greenhouse or outdoor conditions, and due to the intensive cropping systems commonly used in the vegetable production farms, soilborne diseases and pests widely spread, limiting the yield and growth traits in eggplant [6]. At present, the absence of resistant genotypes, together with the ban on methyl bromide [7], has increased interest in the use of grafted eggplant. There are reports underlining the advantages derived from adopting grafting both in terms of plant biotic/abiotic stress tolerance and yield stability [8–10]. *Solanum torvum* Sw. is the most used rootstock for eggplant as it allows several soilborne diseases to be overcome, such as *Verticilium dahliae* Klebahn, *Ralstonia solanacearum* (Smith) Yabuuchiet al., *Fusarium oxysporum* (Schlechtend:Fr.) f. sp. *Melongenae* Matuoand Ishigami, and *Meloidogyne* spp. root-knot nematodes [6,11]. However, both eggplant's wild and allied relatives and interspecific hybrids are proposed as alternative eggplant rootstocks [12–14]. The use of tomato rootstock is also suggested to improve yield, fruit visual quality and plant vigor in eggplant [15].

Crop rotations and biological diversity long have been the key factors in successful traditional agricultural production systems. Rotation is essential for minimizing the build up of pest and soil borne disease problems. However, intensive greenhouse production often precludes vegetable growers from applying rotation. The lack of rotation and tendency towards monocropping in intensive protected vegetable production systems not only increase the prevalence of insects, soil plant pathogens and nematodes but, along with high nutrient levels, can suppress arbuscular mycorrhizal fungi (AMF). AMF are obligate symbionts between plants and fungi belonging to the monophyletic phylum *Glomeromycota* [16]. AMF symbiosis is extremely important for enhancing the assimilation of key macroelements and microelements (P, Cu and Zn) and improving nutrient uptake and efficiency due to its ability to develop an extended external hypha up to 40–50 times their length [17]. Another prominent and sustainable means of boosting yield is the inoculation of beneficial AMF in specific environments such as soil greenhouses where these fungi are generally suppressed.

Moreover, recent findings demonstrated that AM may modulate plant secondary metabolism, enhancing antioxidant activity as well as the accumulation of antioxidant molecules known as phytochemicals (i.e., carotenoids, flavonoids, glucosinolates and phenolic acids) with health-promoting properties [18]. AM fungi have been reported to not only enhance nutritional status and the quality of the produce but also to be able to reduce several forms of plant distress such as thermal stress [19], salinity stress [20], drought stress [21] and soil heavy metal stress [22–25]. Furthermore, there are reports showing an increased resistance to certain diseases in mycorrhized plants [26]. Although there are several studies on the synergistic effects of grafting and other agronomical or chemical means commonly applied in vegetable crop production [27,28], few experiments have been conducted on the interactive effects of grafting and AM on plant performance and the nutritive value of vegetables. In particular, Kumar et al. [24], studying the combined role of grafting and AM on Cd stress tolerance in tomato, observed that vigorous rootstock such as Maxifort successfully alleviate Cd stress symptoms via several physiological and biochemical mechanisms, such as (i) better nutrient absorption and translocation, (ii) higher synthesis of pigments linked to photosynthetic activity and (iii) better capability of producing enzymes (CAT and APX), proline and key metabolites (phytochelatin, fructans and inulins). Furthermore, Oztekin et al. [29], after studying the effect of grafting and AM on the performance of tomato plants cultivated under different salinity conditions in two growing seasons, conclude that grafting and AM have synergistic effects on tomato plant salinity tolerance. Nevertheless, no specified research has been conducted to study the co-operative effects between using wild and allied eggplant relatives as rootstocks and AM application in improving eggplant crop performance and nutritive value. On the aforesaid basis, the aims of our investigation were to appraise the concerted action of eggplant wild/allied relatives' rootstocks and AM inoculation on the yield, fruit quality and nitrogen use efficiency of "Birgah" eggplant grown under greenhouse conditions. The outcome of this study should provide useful information on the performance of new eggplant rootstocks and on their response to AMF inoculation.

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

#### *2.1. Trial Site, Nursery Production and Growing Settings*

The investigation was conducted in 2016 at the experimental farm of the University of Palermo. Seeds of the *Solanum torvum*, *Solanum macrocarpon* and *Solanum paniculatum* rootstocks were sown in 44-cell seedling trays, containing pasteurized peat moss (FAP, Padova, Italia), in a protected environment with a day/night temperature cycle of 25 ◦C/18 ◦C. Twenty days after rootstock sowing, seeds of the "Birgah" eggplant cultivar were seeded in 104-cell trays and maintained under the same climatic conditions and with the same planting method as the rootstocks. "Birgah" is an eggplant F1 hybrid belonging to the violet round shape group, and it is one of the most cultivated eggplants, both in open fields and in protected environments, in Sicily. Seventy-five days after planting, all seedlings attained a hypocotyl diameter of at least 2 mm (the minimum recommended for the tube grafting method). The grafting of "Birgah" scions was performed as reported by Sabatino et al. [12,13]. Briefly, the rootstock was cut off, 0.5 cm below the cotyledons, at a 45◦ angle, and a similar cut was performed on the scion. Care was taken to ensure that the cut surfaces fitted perfectly. To complete the grafting procedure, a plastic clip was attached at the grafting point to guarantee that the correct amount of pressure was applied. The grafted plantlets were placed in a greenhouse equipped with a fog system in order to maintain a humidity of 95% and an air temperature of 20 ◦C for 7 days. After that, the grafted transplants were subjected, for 3 days, to a slow dropping of the humidity, useful for plant acclimatization. Then, they were ready for transplantation. Ungrafted and self-grafted eggplants were also included.

Prior to transplanting, half of the ungrafted, self-grafted and grafted plantlets were treated with 10 g plant−<sup>1</sup> of micorrhizal inoculum carrying 40 spores g−<sup>1</sup> of *Rhizophagus irregularis* (formerly *Glomus intraradices*). All plantlets were transplanted on 9 February, 2016 and maintained till the end of May, 2016 in an unheated greenhouse in Typic Rhodoxeralf soil, characterized by the following soil texture: 46.5% sand, 22.3% silt and 31.2% clay. The soil pH was 7.2. Before the experiment, a brassica crop was cultivated. Mulching with a black polyethylene film of 20 μm was installed. Eggplant plantlets were spaced in order to obtain a plant density of 2 plants m<sup>−</sup>2. Plants were periodically drip irrigated, receiving 250 kg of nitrogen ha−1, 150 kg of phosphorous pentoxide ha−<sup>1</sup> and 250 kg of potassium oxide ha−1. The cultivation practices reported by Baixauli [30], suggested for eggplant growth in Mediterranean conditions, were applied regularly subject to crop requirements.

#### *2.2. Yield, Nitrogen Use E*ffi*ciency, Nutritional Traits and Functional Compounds*

Immediately after harvest, the fruits were weighed and separated into marketable and waste production categories. The number of marketable fruits per plant was recorded, and average fruit weight was also calculated. In total, ten harvests were realized, starting on 21 March, 2016.

A digital penetrometer (Trsnc, Forlì, Italy) was used to determine fruit firmness. Fruit firmness was measured based on three replicates of five fruits per scion/rootstock combination. Each fruit was perforated, using a 6 mm diameter stainless steel cylinder probe, on two opposite sides of the fruit equatorial zone. Firmness was expressed in newtons (N).

Soluble solids content (SSC) was recorded based on three replicates of five fruits and was determined using a refractometer (MTD-045nD, Three-In-OneEnterprises Co. Ltd. Taiwan). Prior to the SSC determinations, the fruit juice was adequately clarified.

Three to five commercially mature fruits for each replicate, from the second and third harvests, were used for the analytical determinations. Each sample consisted of the same quantity of apical, equatorial and distal parts of the fruits. Nutritional and functional determinations were performed on fruits harvested from labelled flowers at the fruit set stage, and all the fruits were harvested 35 days after labelling (fruit commercial maturity stage).

Fruit dry weight was measured via the dehydration of the sample in a heater at 80 ◦C until a constant weight was achieved. The fruit dry matter percentage was calculated using the fruit fresh and dry weights. For protein determination, the Kjeldal method was applied. Specifically, sample was exposed to acid-catalyzed mineralization to convert the organic nitrogen into ammoniacal nitrogen. Subsequently, the ammoniacal nitrogen was distilled under alkaline pH. The ammonia produced via the distillation was collected in a boric acid solution and measured through titrimetric dosage. The ammoniacal nitrogen value was multiplied by 6.25.

For Ca, Mg, K and Fe determinations, atomic absorption spectroscopy following wet mineralization was adopted as reported by Morand and Gullo [31]. Phosphorus values were assessed using colorimetry as reported by Fogg and Wilkinson [32].

The nitrogen use efficiency (NUE = yield/N application rate) was calculated and expressed as t·kg<sup>−</sup>1.

The ascorbic acid content was determined as described by Sabatino et al. [33] for tomato. Therefore, the determinations were made with a reflectometer (Merck RQflex∗ 10 m) using Reflectoquant Ascorbic Acid Test Strips. Thus, one gram of eggplant fruit juice was mixed with distilled water to a 10 mL total volume. Subsequently, an appropriate test strip was immersed into the prepared sample then inserted into the meter. The results were expressed as mg of ascorbic acid per 100 g of fresh weight.

The anthocyanin extraction procedure applied was that described by Mennella et al. [34]. The determination was conducted on a lyophilized and powdered sample of 200 mg. The flow rate and absorbance units full scale adopted were 0.8 mL min−<sup>1</sup> and at 0.1, respectively. For RP-HPLC analyses, purified delphinidin-3-rutinoside (D3R, Polyphenols Laboratories AS, Sandnes, Norway) was used as an external standard with a dissimilar retention time (23.9 min) compared to delphinidin-3-(p-coumaroyl rutinoside)-5-glucoside (nasunin), which was eluted with a longer retention time (25.8 min for cis-nasunin and 26.1 min for trans-nasunin, respectively). Regarding the nasunin determination, in agreement with Lo Scalzo et al. [35], a partially purified standard was used. Total anthocyanins were expressed as mg·100 g−<sup>1</sup> dry weight (dw), highlighting that the detection limit was 2.00 mg 100 g−<sup>1</sup> of dw.

For chlorogenic acid determination, the extraction and analysis procedure described by Stommel and Whitaker [36], with slight variations, was applied. Therefore, a binary mobile phase gradient of methanol in 0.01% aqueous phosphoric acid was provided according to this procedure: 0–15 min, linear increase from 5% to 25% methanol; 15–28 min, linear increase from 25% to 50% methanol; 28–30 min, linear increase from 50% to 100% methanol; 30–32 min, 100% methanol; 32–36 min, linear decrease from 100% to 5% methanol; 36–43 min, 5% methanol. The flow rate was 0.8 mL·min<sup>−</sup>1. The chlorogenic acid quantification, subsequently conducted by RP-HPLC separation, was based on the absorbance at 325 nm relative to the external standard of chlorogenic acid (Sigma-Aldrich, St.Louis, MO) and expressed as mg·100 g−<sup>1</sup> of dw.

For glycoalkaloids, the extraction method adopted was that reported by Birner [37] with some adjustments. Thus, 0.5 g of lyophilized and powdered flesh tissue sample was extracted with 95% ethanol. For glycoalkaloids (expressed as mg·100 g−1dw) analyses, the method described by Kuronen et al. [38] was applied. Therefore, the analyses were conducted by means of RP-HPLC using purified solasonine and solamargine as the external standards. The detection limit was 0.03 mg 100 g−<sup>1</sup> of dw.

#### *2.3. Root Colonization Assessment*

To evaluate mycorrhizal colonization, the method described by Phillips and Haymann [39] and revised by Torta et al. [40] was adopted. Briefly, three lateral root samples from mychorrhized plants were collected and marked with acid fuchsin. As reported by Kormanik and McGraw [41], mycorrhizal colonization (Mycorrhization Index (MI = % of marked tissue, with respect to the hyaline portion, on the unit of the length of the root)) was evaluated on three fragments, attaining the average value.

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

The trial was organized in a randomized complete block design with 3 replicates per treatment. Each replicate consisted of 10 plants. Consequently, the total number of plants was 300 (10 treatments × 30 plants per treatment = 300 plants). The data were subjected to GLM (General Linear Model)

analysis using the SPSS software package version 20.0. For data expressed in percentages, an arcsin transformation before ANOVA analysis (Ø = arcsin(p/100)1/2) was performed. Tukey's HSD test (*p* < 0.05) was used to compare means. Principal component analysis (PCA) was conducted on the whole yield, fruit quality and nitrogen use efficiency data set.

#### **3. Results**

#### *3.1. Root Colonization by AM Fungi*

AM fungi root colonization was significantly affected by mycorrhizal inoculation (M), whereas the grafting combination (G) did not significantly influence AM root colonization. ANOVA showed no significant interaction (G × M) (Figure S1). Irrespective of the grafting combination, the percentage of root colonization was higher with the inoculated treatment (71.8%) compared to with the non-inoculated one (2.5%) (Figure S1).

### *3.2. E*ff*ect of Grafting Combination and Mycorrhizal Inoculation on Yield, Yield Components and NUE*

The marketable yield, number of marketable fruit per plant and percentage of unmarketable production were significantly influenced by the G × M interaction (Table 1).

**Table 1.** Main effects of the grafting combination and arbuscular mycorrhiza (AM) inoculation on yield and the yield components of greenhouse eggplant.


NS, \*\*\* non-significant or significant at *p* ≤ 0.001. Data represent mean values of three replicates. Values within a column followed by the same letter are not significantly different at *p* ≤ 0.05 according to Tukey's HSD Test. NS = not significant. +AM, −AM = mycorrhizal and non-mycorrhizal eggplants, respectively.

On the other hand, neither the grafting combination nor mycorrhizal treatment had a significant effect on fruit mean weight (avg. 500.6 g fruit−1). When averaged across mycorrhizal treatments, the marketable yield production reached a maximum value with Brigah grafted onto *S. torvum* (B/T), followed by both the B/M and B/P grafting combinations, while the lowest crop productivity was recorded for ungrafted and self-grafted crops (Table 1). Interestingly, the higher marketable fresh yield observed for eggplants grafted onto *S. torvum* and to a lesser extent onto *S. macrocarpon* and *S. paniculatum*, in comparison to both the ungrafted and self-grafted plants, was associated with an increase in the number of eggplant fruits per plant and not with a change in the mean fruit weight (Table 1). Moreover, the highest discarded production was observed with the B/M grafting combination (Table 1). A significant effect of mycorrhizal treatment on the yield and yield components was also observed (Table 1). Irrespective of grafting combinations, the inoculation of eggplant at the transplantation stage with *R. irregularis* elicited significant increases in the marketable yield and number of fruits per plant of 13.3% and 12.6%, respectively, compared to for the non-inoculated control (Table 1).

Similarly to the effects on the yield and yield components, the nitrogen use efficiency (NUE) for the B/T grafting combination (0.242 t kg<sup>−</sup>1) was significantly higher, by 31.2%, than that obtained from the ungrafted and self-grafted eggplant (avg. 0.184 t kg<sup>−</sup>1; Table 1). Finally, the inoculation of eggplant with *R. irregularis*, when averaging across grafting combinations, induced a significant increase in the NUE, which was 13.3% higher than for the non-inoculated control treatment (Figure 1).

**Figure 1.** Main effects of grafting combination and AM inoculation on the nitrogen use efficiency (NUE) of greenhouse eggplant. (B): Birgah ungrafted; (B/B): Birgah self-grafted; (B/T): Birgah/*S. torvum*; (B/M): Birgah/*S. macrocarpon*; (B/P): Birgah/*S. paniculatum*.
