**Popular Biofortified Cassava Cultivars Are Heavily Impacted by Plant Parasitic Nematodes, Especially** *Meloidogyne* **Spp.**

#### **Aminat Akinsanya 1,2,**†**, \*, Steve Afolami 1,**†**, Peter Kulakow 2, Elizabeth Parkes <sup>2</sup> and Danny Coyne 2,3**


Received: 5 May 2020; Accepted: 23 June 2020; Published: 26 June 2020

**Abstract:** The development of new biofortified cassava cultivars, with higher micronutrient contents, offers great potential to enhance food and nutrition security prospects. Among the various constraints affecting cassava production are plant parasitic nematodes (PPN), especially root-knot nematodes. In this study, six popular biofortified cultivars were field-evaluated for their response to PPN in Nigeria. A field naturally infested with a diversity of PPN but dominated by root-knot nematodes was used. Application of the nematicide carbofuran significantly reduced PPN densities, and at harvest, no root galling damage was observed, compared with untreated plots, which had heavy galling damage. Plant height, stem girth, plant weight, marketable storage root number and weight were significantly lower for most cultivars in untreated plots. Percentage yield losses in the range of 21.3–63.7% were recorded from two separate trials conducted for 12 months each. Lower total carotenoid and dry matter contents were associated with higher PPN densities in some biofortified cultivars, resulting in a loss of as much as 63% of total carotenoid and 52% of dry matter contents. The number and weight of rotted storage roots were significantly greater in untreated plots across cultivars, reducing in-field and post-harvest storability. This study demonstrates that natural field populations of PPN can substantially affect yield, quality and nutritional value of released biofortified cassava cultivars.

**Keywords:** carotenoid content; *Manihot esculenta*; nutrition; root-knot nematodes; storability

#### **1. Introduction**

Cassava is a major staple food crop in tropical and subtropical Africa, Asia, and Latin America, where approximately 500 million people depend on it as a major carbohydrate (energy) source [1]. It is an important crop for food security in these regions, partly because it yields more energy per hectare than many other major crop. In Africa, cassava is the most important of all root and tuber crops as a source of calories for human and livestock needs, ranking 4th, after rice, sugarcane and maize [2]. To meet rising demands for cassava but also to improve its nutritional value as a food, there has been considerable investment to breed improved cultivars, including for higher mineral and vitamin contents, a process referred to as biofortification [3]. The specific enhancement of nutritional elements through genetic improvement is referred to as biofortification [4]. The Bill and Melinda Gates Foundation have supported a global effort to develop cassava germplasm enriched with bioavailable nutrients since 2005. The BioCassava Plus initiative has six major objectives, including reducing cyanogen content, delaying postharvest deterioration and developing disease-resistant cultivars [5]. Using hybridization and selective breeding, researchers in Nigeria have developed new yellow cultivars of cassava that naturally produce a higher level of beta-carotene, which will help in reducing malnutrition caused by vitamin A deficiency in the region [6]. A total of seven biofortified cassava cultivars with total carotenoid content in the range of 8–12 μg/g fresh weight and 30–33% dry matter have been released [5], which compare with white cultivar total carotenoid content and dry matter in the range of 0.05–0.09 μg/g fresh weight and 35–37%, respectively. The dry matter content for provitamin A cultivars, therefore, is relatively lower compared to locally used cultivars and is a priority for improvement [5]. Cassava biofortification has largely been aimed at addressing vitamin A deficiency [7], an important public health problem in sub-Saharan Africa.

However, while raising the nutritional value of storage roots is a worthy objective, with great prospects for impacting the lives of millions, it must be additionally coordinated and associated with other valuable traits, such as pest and disease resistance. A number of biotic constraints affect the production of cassava, especially diseases and in particular virus diseases [8]. Major efforts have focused on breeding resistance against these threats into new, improved cultivars [8]. Less recognized threats, however, such as PPN have received much less attention. Although not well recognized, there is considerable, and growing, evidence of the damage that PPN inflict on cassava production [9–12]. In many cases, however, nematode damage often goes unnoticed. Traditionally, cassava is considered a hardy crop and generally viewed as immune to PPN. The naturally 'knobbly' and rough texture of the roots, which can disguise nematode damage to casual observation, partly aids this perception, while there may also be few roots present at harvest, especially if affected by nematode infection and have become necrotic and died off. Among the PPN that infect cassava, root-knot nematodes (*Meloidogyne* spp.) are the most prominent [13,14]. A number of studies have demonstrated the damaging impact of root-knot nematode infection, including their association with an increased incidence of rots [14]. Other PPN are also associated with cassava losses, such as *Pratylenchus brachyurus* and *Scutellonema bradys*, but the root-knot nematodes *M. incognita* and *M. javanica* are the most commonly reported and the most important nematode pests [12]. Some reports have documented almost total losses due to *Meloidogyne* spp. [15], although the association is sometimes not always so clear [16]. The association with rots can also disguise nematode damage and losses attributable to PPN but is an important aspect. This indirect consequence can lead to greater losses due to secondary fungal and bacterial rots and indeed has been shown to be strongly associated [17]. Given that improving the storability of cassava storage roots is a key breeding objective, the appropriate management of PPN will undoubtedly contribute to improving this objective [15,17]. Earlier studies have illustrated the variable susceptibility of cassava cultivars to *Meloidogyne* spp. [14,16], including biofortified cultivars [14,18,19]. The current study was undertaken to evaluate the effect of PPN on officially released biofortified cultivars in the field. These cultivars, released in Kenya and Nigeria, had been reported in a previous study to be susceptible to *M. incognita* in pots [19]. Our study builds on the pot evaluation study; using a field naturally infested with PPN, we assessed the impact of PPN infection on these cultivars and the implication of this for cassava farmers and consumers in the region.

#### **2. Results**

#### *2.1. Plant-Parasitic Nematode Identification and Densities*

Eleven genera of PPN were recorded: Meloidogyne, Pratylenchus, Helicotylenchus, Scutellonema, Hoplolaimus, Tylenchus, Longidorus, Aphelenchus, Xiphinema, Rotylenchulus and Radopholus (Figure 1). The initial population densities (Pi) were relatively low for seven nematode genera, while the more prominent genera were Meloidogyne, Pratylenchus, Helicotylenchus and Scutellonema (Table 1). Carbofuran application significantly (*p* ≤ 0.05) suppressed PPN densities in all treated plots, compared with the untreated plots (Figure 1). The genera Meloidogyne, Pratylenchus, Helicotylenchus

and Scutellonema remained prominent over the duration of the trials with P*f*'s significantly (*p* ≤ 0.05) higher than other genera; Meloidogyne spp. had higher soil densities than all other genera (Figure 1).

**Figure 1.** Effect of carbofuran on the final population of four major plant-parasitic nematodes encountered in field trials in Nigeria. P*f* = final density at harvest—12 months after planting; treatments: treated = 3G Carbofuran at 60 g/plot twice; untreated = untreated control. Error bars = Least Significant Difference (*p* ≤ 0.05).

**Table 1.** Population of four major plant-parasitic nematodes encountered at planting in field trials in Nigeria 1.


<sup>1</sup> <sup>n</sup> <sup>=</sup> 8: means of four replications x two experiments; <sup>2</sup> 3G Carbofuran was applied at a rate of 60 g/plot twice; <sup>3</sup> soil densities from 250 g/soil; for each treatment group values within a column followed by a different letter are significantly (*p* ≤ 0.05) different.

#### *2.2. Root Galling Damage and Host Suitability of Biofortified Cassava Cultivars Due to Meloidogyne Spp., 12 Months after Planting*

All cultivars in the untreated plots reacted to *Meloidogyne* species infection with varying intensity, ranging from a gall index of 3.6–5.0 (Table 2). No galling of feeder roots was recorded in carbofuran-treated plots, but a low *Meloidogyne* population was recorded in the soil. A gall index of 5.0 was recorded on the check cultivar only, IITA-TMS-IBA30572, which also recorded the highest number of galls. Of the biofortified cultivars, IITA-TMS-IBA011368, IITA-TMS-IBA011371 and IITA-TMS-IBA070593 recorded the highest number of galls. In the untreated plots, all biofortified cassava cultivars reacted to *Meloidogyne* spp. infection and were all rated as good hosts, based upon a reproduction factor (RF) greater than 5.0 (Table 2).


**Table 2.** Root galling damage and host suitability of six biofortified cassava cultivars due to *Meloidogyne* spp, 12 months after planting in field trials in Nigeria 1.

<sup>1</sup> <sup>n</sup> <sup>=</sup> 8: means of four replications x two experiments; <sup>2</sup> 3G Carbofuran was applied at a rate of 60 g/plot twice; <sup>3</sup> root and soil densities combined from 10 g/roots and 250 g/soil; <sup>4</sup> treated plots recorded no galling; gall index: 1 = 1–2 galls, 2 = 3–10 galls, 3 = 11–30 galls, 4 = 31–100 galls, 5 = > 100 galls [20]; <sup>5</sup> RF = nematode reproduction factor [21]; <sup>6</sup> host status was categorized as good (G) when P*f*/P*i* > 5.0, fair (F) if 5.0 ≥ P*f*/P*i* > 1, poor (P) if 1 ≥ P*f*/P*i* > 0, and nonhost (N) when P*f*/P*i* = 0 [22]; for each treatment group values within a column followed by a different letter are significantly (*p* ≤ 0.05) different.

#### *2.3. Growth and Development of Biofortified Cassava Cultivars*

The analysis of the data showed that there was no cultivar or treatment interaction per year at 6, 9 and 12 months after planting (MAP) (Table 3a). Therefore, the data from the experiments were pooled together for analysis (Table 3b).


**Table3.**(**a**).MeansquaresforgrowthanddevelopmentofsixbiofortifiedcassavaandonewhitecultivarinfieldtrialsinNigeria.(**b**)Growthand

Carbofuran was applied at a rate of 60 g/plot twice; MAP = months after planting; SE = standard error; for each treatment group values within a column followed by a different letter are

significantly (*<sup>p</sup>* ≤ 0.05) different.

The application of carbofuran improved the growth of all cassava cultivars at some point during the growing cycle of the experiments; treated plants were generally significantly (*p* ≤ 0.05) taller and sturdier (Table 3b). Generally, stunting of aerial growth was observed on untreated plants at 3 MAP, which became more pronounced at 6, 9 and 12 MAP, when compared with treated plots. Cultivars were significantly (*p* ≤ 0.05) shorter in untreated plots compared with treated plots, except for IITA-TMS-IBA070593. Significant (*p* ≤ 0.05) reduction was also recorded in the stem girth of untreated plots in some of the cultivars, when compared with the treated plots. The overall mean showed that the growth and development of all cultivars in the untreated plots were significantly (*p* ≤ 0.05) suppressed at 3, 6, 9 and 12 MAP, when compared with treated plots, and the standard error (SE) was mostly higher in the untreated plots when compared with treated plots and increased at 3, 6, 9 and 12 MAP (Table 3b).

#### *2.4. Yield Evaluation of Biofortified Cassava Cultivars*

The results showed that there was cultivar and treatment interaction per year (Table 4a). The non-marketable storage yields showed no interaction per year (Table 4b), and these data were pooled together for analysis (Table 4e).






**Table 4.** *Cont.* (**a**) \*, \*\*, \*\*\*squares significant at *p* ≤ 0.05, 0.01 and 0.0001 probability levels, respectively; fresh storage roots number and weight in 5 plants/plot. (**c**) n = 4: means of four replications; 3G Carbofuran was applied at a rate of 60 g/plot twice; 3 fresh storage roots number and weight in 5 plants/plot; SE = standard error; for each treatment group values within a column followed by a different letter are significantly (*<sup>p</sup>* ≤ 0.05) different. (**d**) 1 n = 4: means of four replications; 2 3G Carbofuran was applied at a rate of 60 g/plot twice; 3 fresh storage roots number and weight in 5 plants/plot; SE = standard error; for each treatment group values within a column followed by a different letter are significantly (*<sup>p</sup>* ≤ 0.05) different. (**e**) 1 n = 8: means of four replications x two experiments; 2 3G Carbofuran was applied at a rate of 60 g/plot twice; 3 fresh storage roots number and weight in 5 plants/plot; SE = standard error; for each treatment group values within a column followed by a different letter are significantly (*<sup>P</sup>* ≤ 0.05) different.

*Plants* **2020**, *9*, 802

Plant weight and storage root yields were largely improved (*p* ≤ 0.001) across cassava cultivars in the two trials (Table 4a,c,d). Nematicide treatment led to higher (*p* ≤ 0.05) numbers and fresh weights of marketable storage roots in most cultivars, compared with untreated plots (Figure 2). The number and weight of non-marketable storage roots were significantly (*p* ≤ 0.05) lowered in cultivars IITA-TMS-IBA011368 and IITA-TMS-IBA011412 when compared with untreated plots in the first trial (Table 4c), while in the second trial (Table 4d), these cultivars in addition to NR 07/0220 were significantly (*p* ≤ 0.05) lowered. A significant (*p* ≤ 0.05) reduction in the number of rotted storage roots was also observed in treated plots, compared with untreated, with lower numbers for cultivars IITA-TMS-IBA011368 and NR 07/0220 (Table 4e). Fresh weights of rotted storage roots were similarly lower in treated plots, with cultivars IITA-TMS-IBA011368, NR 07/0220 and IITA-TMS-IBS30572 having significantly (*p* ≤ 0.05) less. The number and weight of deformed storage roots of cultivar IITA-TMS-IBA011412 were also less (*p* ≤ 0.05) in the treated plots, compared with the untreated (Table 4e). The overall mean showed that the aerial fresh weight of plants and the number and weight of marketable storage roots in the untreated plots were significantly (*p* ≤ 0.05) lower, when compared with treated plots in the two trials, while SE rates were higher in the untreated plots when compared with treated plots (Table 4c,d), while the number and weight of non-marketable (rotted and deformed) storage roots were significantly (*p* ≤ 0.05) lower in treated plots, with higher SE in the untreated plots when compared with treated plots (Table 4e).

**Figure 2.** Storage roots of treated and untreated biofortified cassava cultivar IITA-TMS- IBA011368 at 12 months after planting; (**a**) treated with 3G Carbofuran; (**b**) untreated storage roots.

Results showed that two biofortified cassava cultivars, IITA-TMS-IBA070593 and IITA-TMS-IBA070539, had significantly (*p* ≤ 0.05) lower total carotenoid contents of roots from untreated plants when compared with treated plants (Table 5). Similarly, dry matter content from untreated plants was lower in cultivars IITA-TMS-IBA011368 and IITA-TMS-IBA070593. When assessing total carotenoid and dry matter contents at the plot scale, however, taking into consideration the contents and yields, all the biofortified cultivars had significantly (*p* ≤ 0.05) lower values, compared with treated plots. The overall mean showed that total carotenoids per plot and dry matter per plant and per plot were significantly (*p* ≤ 0.05) lower in untreated plots, while SE rates were higher in the untreated plots when compared with treated plots (Table 5).


**Table 5.** Nutritional quality of six biofortified cassava cultivars in field trial in Nigeria 1.

<sup>1</sup> n = 4: means of four replications; <sup>2</sup> 3G Carbofuran was applied at a rate of 60 g/plot twice; <sup>3</sup> fresh storage weight in 5 plants/plot; SE = standard error; for each treatment group values within a column followed by a different letter are significantly (*p* ≤ 0.05) different.

#### **3. Discussion**

A pot study using the same six biofortified cultivars as in the current study found them all to be good hosts to the root-knot nematode *M. incognita*, which reduced growth and development after six months in the screenhouse [19]. Although a number of other PPN genera were encountered, the majority were in relatively low densities and likely posed little threat to the cassava. Four genera were more prominent, of which *Meloidogyne*, the most important nematode genera attacking cassava, dominated the PPN community. The focus of the current study therefore centered on *Meloidogyne* spp., although it is understood that *Pratylenchus, Helicotylenchus* and *Scutellonema* spp. could have had some influence on cassava growth, which can become important when they are present in large densities [23]. The effect of *M. incognita* on the nutritional content of these biofortified cassava was less conclusive, but the study provided an indication that *M. incognita* infection can negatively impact cassava quality. The current study clearly supports the pot study findings but now also demonstrates that *Meloidogyne* spp. infection will reduce the nutritional value of improved, biofortified cassava under field conditions. Although the effect varied across cultivar and quality was not consistently reduced proportionally (per unit weight), by taking the yield impact into account, the overall damaging effect of *Meloidogyne* spp. can be better appreciated. All the tested cassava cultivars were susceptible to *Meloidogyne* spp. infection, resulting in significant (*p* ≤ 0.05) root galling damage and a reduction in plant growth and storage root yield of all but one of the six biofortified cultivars. Along with PPN densities, rotted storage roots were also much reduced in nematicide-treated plots. Rotting of storage roots directly affects their in-field storability, as well as their post-harvest longevity. Therefore, placing more emphasis on the management of PPN may be well justified, especially given the emphasis placed on nutritional biofortification and that increasing the storability and longevity of storage roots is a key breeding trait [16,24]. Carbofuran, however, is a toxic carbamate pesticide, which affects a wider range of pests and diseases than PPN alone. The reduction of rot-causing pathogens therefore is likely an additional effect, which would additionally reduce potential rots of cassava roots and should be considered. The pesticide did, however, enable a suitable comparison of PPN field densities, creating a differential against which to assess their impact on cassava. Other studies that have sought to assess the effect of PPN on cassava yield have used similar techniques, in addition to other methods, such as solarization [12,18,25]. From these studies some sizeable yield reductions have been associated with PPN, in particular *Meloidogyne* spp., demonstrating the need for their management if cassava production systems are to be sustainably intensified. Low yields have consistently characterized cassava production in Nigeria and other sub-Saharan countries. Nematode management may provide a major way forward in improving yields in farmers' fields. The association between root rot incidence and *Meloidogyne* spp. infection has also been well demonstrated on cassava [18,25], as well as other root and tuber crops [16,26]. There is no doubting therefore the value of investing in PPN management and root-knot nematodes in particular, towards improving cassava productivity [12,14,15].

The current study showed that high PPN densities were associated with reduced crop performance following treatment with carbofuran, resulting in significant (*p* ≤ 0.001) yield loss of biofortified cassava. *Meloidogyne* spp. in the untreated plots caused galling on feeder roots of all biofortified cultivars. In Nigeria, *Meloidogyne* spp. infection caused significant (*p* ≤ 0.05) suppression in the growth and yield of elite cassava cultivars after 12 months in the field, despite relatively low observed levels of the nematode [14]. The loss in cassava yield was, however, mainly attributed to direct damage of the root system by the feeding activities of *Meloidogyne* spp. Although the current study was conducted at the International Institute of Tropical Agriculture (IITA) station, no inoculation was undertaken, natural PPN infestation levels were used and the trials were managed relative to farmer conditions. It is assumed, therefore, that these results provide a relatively fair reflection of the likely losses that farmers would experience. Elsewhere in Nigeria, significant cassava yield losses have also been recorded in farmer field trials naturally infested with *Meloidogyne* spp. Up to 200% yield increases were observed following the reduction of *Meloidogyne* spp. using solarization [25]. In Uganda, severe galling due to *Meloidogyne* spp. was reported in farmers' fields [27]. Separately, 94% of fields examined in Uganda presented galling damage, with 17% severely affected, indicating substantial yield losses [28]. The impact of *Meloidogyne* spp. on cassava production is a threat to production that is likely to become increasingly acute and more intense under more intensified cropping conditions [11,14,15]. Besides reducing crop growth, vigor and productivity, PPN can reduce the quality and nutritional value of crop products. This is not surprising as PPN infect the root system, disrupting nutrient uptake and reducing their distribution within the plant [29–31]. PPN parasitize plants, changing the nutrient apportioning and cause disturbance in water and nutrient relations necessary for optimal plant growth [32]. Although a number of studies on various crops have indicated or demonstrated this, the empirical evidence is relatively limited. In our study, the total carotenoid and dry matter contents per plot of all biofortified cultivars were significantly (*p* ≤ 0.05) lower in untreated plots with higher PPN densities than treated plots. The current study and the preliminary pot study [15] now clearly show the impact on nutrition that *Meloidogyne* spp. can have, both on an individual plant, but especially when multiplied at scale. For example, the carotenoid content of cultivar IITA-TMS-IBA070539 was less by 0.19 kg per plot in untreated plots. This equates to a loss of 63% of total carotenoid content in the yield and quality of biofortified cassava, seriously undermining the efforts and investment to develop these high content biofortified cultivars.

Our study further confirms earlier reports that *Meloidogyne* spp. are the most prevalent and abundant PPN affecting cassava in Southwestern Nigeria. In the current study, the *Meloidogyne* spp. were not identified to species level, although it is likely that *M. javanica* and/or *M. incognita* were present, both of which are common to the region [33] and are the two most commonly occurring *Meloidogyne* spp. found infecting cassava [23]. As resistance against *Meloidogyne* spp. can be bred for in cassava, it would appear a useful mechanism for improving cassava for more intensive cultivation. The presence and infection of cassava by *Meloidogyne* spp. will reduce the yield and quality of cassava, including the nutritional content of biofortified cassava. Furthermore, *Meloidogyne* spp. infection is additionally associated with higher levels and incidence of rots, reducing the storability of cassava. PPN infection and damage to cassava has largely been overshadowed by other pests and diseases but is, however, a considerable threat to both yield, quality and storability. Breeding or actively selecting for nematode resistance during the evaluation process may therefore be more warranted than

generally acknowledged or appreciated. In addition to creating more durable cultivars, more suitable to intensified cropping conditions, indirectly, this is likely to improve in-ground storability.

#### **4. Materials and Methods**

#### *4.1. Experimental Details and Layout*

Two field trials were planted in June 2017 and May 2018 in a well-drained sandy loam soil after ploughing and harrowing once each, at the IITA Ibadan, Nigeria (120 km north of Lagos). Cassava stems ~15 cm long were planted at an angle into the ground, spaced 1 × 1 m in a line 7 m long for each cultivar, representing a plot of 8 plants. Trials were maintained for 12 months after planting (MAP) before harvesting. The study consisted of two factors, cassava genotype (seven cultivars) and nematicide treatment (two levels). Six biofortified cassava cultivars (IITA-TMS-IBA011368, IITA-TMS-IBA011412, IITA-TMS-IBA011371, IITA-TMS-IBA070593, IITA-TMS-IBA070539 and NR 07/0220) and a check cultivar of white cassava (IITA-TMS- IBA30572) were obtained from the IITA. The nematicide 3G Carbofuran was applied at the rate of 3 kg a.i./ha (60 g/plot) at planting and repeated at 3 MAP and compared with a control receiving no nematicide. The experiment was laid out in a randomized complete block design with four replicates each per cultivar per treatment.

#### *4.2. Assessment of Nematode Population Density and Damage*

Soil samples were collected from 8 points per plot using a soil auger to a depth of 30 cm at planting to obtain initial nematode population density (P*i*) and at harvest to obtain final nematode population density (P*f*). Nematodes were extracted from 250 g soil sub-samples using the tray method [34], after removing all stones and debris and thoroughly mixing the bulked soil from each plot. At harvest, roots from 5 plants per plot were combined, gently tapped free of soil, chopped finely, thoroughly mixed and a 10 g sub-sample removed for nematode extraction using the same method as for soil. Nematode extracts were removed after 24 h, allowed to settle for 5 h and the volume adjusted to 30 mL by siphoning off the excess [35]. The mean nematode density was assessed under a compound microscope from 5 x 1 mL aliquots pipetted into a Doncaster counting dish [36]. Nematodes were identified to genus level using Bell's Key [37] and a multiple tally counter used to count the different nematode genera. Total number of nematodes per plot from soil and root data was used to calculate the nematode reproduction factor (RF) [21]:

$$\frac{\text{Pf} \times 250 \text{ g/soil} + \text{Pf} \times 10 \text{ g/root}}{\text{Pi}} \tag{1}$$

At harvest, the number of galls on 5 cm feeder roots per plant, removed randomly from 5 plants per plot, was counted and galling index (GI) per plant root assessed using the 1–5 gall index scale [20] (1 = 1–2 galls, 2 = 3–10 galls, 3 = 11–30 galls, 4 = 31–100 galls, 5 = > 100 galls).

#### Host Status

Host suitability was categorized as good when P*f*/P*i* > 5.0, fair if 5.0 ≥ P*f*/P*i* > 1, poor if 1 ≥ P*f*/P*i* > 0 and nonhost when P*f*/P*i* = 0 based on a study method [22].

#### *4.3. Measurement of Crop Growth Parameters*

Crop growth parameters were collected at 3, 6, 9 and 12 MAP for plant height and girth from five randomly selected cassava plants per plot. At harvest, the five selected plants per plot were additionally assessed as a bulk (plot) for aerial plant weight, number and weight of marketable and non-marketable storage roots. Plant height was measured to the tallest point of pre-harvested plants using a wooden ruler; girth was measured at 10 cm above the soil surface using a Vernier caliper. Stem and leaf material per plant was weighed together per plot and recorded as plant fresh weight. Harvested storage roots were sorted into non-marketable (small) and marketable storage roots. Deformed storage roots

(physically twisted) and those affected by root rot were counted and weighed separately. Total yield was computed from all harvested marketable and non-marketable storage roots per plot.

#### *4.4. Carotenoid and Dry Matter Analysis*

The nutritional content of storage roots was assessed using total carotenoid nutrient and dry matter content following the procedure outlined in [15]. Cassava storage roots from each plot were randomly divided, one for fresh and the other for dried analysis for total carotenoid and dry matter content, respectively. The roots were chopped into ~0.5 cm<sup>3</sup> cubes and 100 g sub-samples for each plot were randomly removed to determine the total carotenoid content using the iCheck™ method (BioAnalyt GmbH, Teltow, Germany). Total carotenoid content and dry matter were conducted for the first trial only due to the high cost of this procedure. For dry matter analysis, the 100 g fresh storage root cubes were oven-dried at 70 ◦C for 72 h, then milled to obtain a homogeneous powder, stored in moisture-free plastic containers and dry matter calculated for each cultivar [38]:

$$\text{Dry matter} \left( \% \right) = \frac{\text{Final weight}}{\text{Fresh weight}} \times 100 \tag{2}$$

#### *4.5. Statistical Analysis of Data*

Data were subjected to a factorial analysis of variance (ANOVA) using SAS 9.4 [39] statistical package and means separated using least significant difference (LSD) at 5% level of probability. The data from the two experiments were pooled for analysis for those that recorded no cultivar or treatment interaction per year.

#### **5. Conclusions**

It is abundantly clear from the results that nematodes are a major constraint to cassava production. Root-knot nematode *Meloidogyne* spp. and the lesion nematode *Pratylenchus* spp. were the most common and important nematodes encountered from the study while *Helicotylenchus* spp., *Scutellonema* spp. and *Hoplolaimus* spp. could also become important when present in large numbers. All the biofortified cassava cultivars were susceptible and reacted to *Meloidogyne* spp. with varying intensity of root galling, ranging between 3.50 and 5.00 index. This was associated with a significant (*p* ≤ 0.05) reduction in above-ground fresh weight, plant height, stem girth, marketable storage root weight and number in most cultivars. The nutrient analysis clearly demonstrates the negative impact of PPN on the nutrient quality of biofortified cassava. Therefore, breeding and/or selecting for resistance against PPN, especially *Meloidogyne* spp., is here highlighted as highly necessary to achieve good yields and maintain nutrient quality in biofortified cassava. This has particular relevance under more intensified cropping conditions, which exaggerate soil and root borne constraints. Furthermore, the effect of root-knot nematode infection on the reduction of total carotenoid and dry matter contents should be investigated. Carbofuran was used to effectively manage PPN densities in the field in the current study, but it is an environmentally hazardous product that has been removed from the market in many places, even if it is systemic and not toxic to plants [40,41]. Synthetic pesticides are also often out of reach for resource-poor African farmers due to their high cost. Consequently, there is the need to work out effective and sustainable nematode control strategies in order to improve growth, yield and quality of biofortified cassava. Root-knot nematodes are highly pervasive pests, which are becoming increasingly problematic across tropical cropping systems and as such require particular attention from breeders.

**Author Contributions:** Conceptualization, A.A. and S.A.; methodology, A.A.; software, A.A.; validation, A.A., S.A. and D.C.; formal analysis, A.A.; investigation, A.A.; resources, D.C., P.K. and E.P.; data curation, A.A.; writing—original draft preparation, A.A.; writing—review and editing, S.A., D.C. and P.K.; visualization, A.A., S.A. and D.C.; supervision, S.A., D.C. and P.K.; project administration, A.A. and D.C.; funding acquisition, D.C. and P.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by donor contributions to the CGIAR Fund (https://www.cgiar.org/funders/), in particular to the CGIAR Research Program for Roots, Tubers and Bananas (CRP-RTB), HarvestPlus and by the Bill and Melinda Gates Foundation and UKAID (Grant 1048542).

**Acknowledgments:** The authors wish to appreciate the technical support of the nematology, cassava breeding and yam barn units of IITA, Ibadan, for providing assistance in planting, managing, harvesting and processing of the cassava roots.

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

#### **References**


© 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* **Changes in the Plant** β**-Sitosterol/Stigmasterol Ratio Caused by the Plant Parasitic Nematode** *Meloidogyne incognita*

**Alessandro Cabianca 1, Laurin Müller 1, Katharina Pawlowski <sup>2</sup> and Paul Dahlin 1,\***


**Abstract:** Sterols play a key role in various physiological processes of plants. Commonly, stigmasterol, β-sitosterol and campesterol represent the main plant sterols, and cholesterol is often reported as a trace sterol. Changes in plant sterols, especially in β-sitosterol/stigmasterol levels, can be induced by different biotic and abiotic factors. Plant parasitic nematodes, such as the root-knot nematode *Meloidogyne incognita,* are devastating pathogens known to circumvent plant defense mechanisms. In this study, we investigated the changes in sterols of agricultural important crops, *Brassica juncea* (brown mustard), *Cucumis sativus* (cucumber), *Glycine max* (soybean), *Solanum lycopersicum* (tomato) and *Zea mays* (corn), 21 days post inoculation (dpi) with *M. incognita*. The main changes affected the β-sitosterol/stigmasterol ratio, with an increase of β-sitosterol and a decrease of stigmasterol in *S. lycopersicum*, *G. max*, *C. sativus* and *Z. mays*. Furthermore, cholesterol levels increased in tomato, cucumber and corn, while cholesterol levels often were below the detection limit in the respective uninfected plants. To better understand the changes in the β-sitosterol/stigmasterol ratio, gene expression analysis was conducted in tomato cv. Moneymaker for the sterol 22C-desaturase gene *CYP710A11*, responsible for the conversion of β-sitosterol to stigmasterol. Our results showed that the expression of *CYP710A11* was in line with the sterol profile of tomato after *M. incognita* infection. Since sterols play a key role in plant-pathogen interactions, this finding opens novel insights in plant nematode interactions.

**Keywords:** sterol; β-sitosterol; stigmasterol; plant parasitic nematode; *CYP710A*; 22C-desaturase

#### **1. Introduction**

Plants are consistently exposed to numerous pests and pathogens, which leads to variations in plant metabolism, including sterol profiles. Sterols are biomolecules which play important roles in various biological processes. Besides their essential function in cell membrane support and fluidity, they are also important as hormone precursors and are involved in biotic and abiotic stress responses [1–5]. Sterols belong to the large group of isoprenoid synthesized via the lanosterol (animals and fungi) or cycloartenol (plants) pathway (Figure 1), sharing a basic structure with a four-cyclic hydrocarbon ring, called gonane, and a hydroxyl group at position C-3. Depending on the organism, sterols are differently modified in the ring structure or in the side chain at position C-17, by methylations or double bonds [4,6]. Cholesterol, arguably the most studied sterol, is mainly synthesized in animals. In contrast, plants largely contain a mixture of C-24 sterols, such as β-sitosterol, campesterol and stigmasterol (collectively known as phytosterols). Nevertheless, they also synthesize minor amounts of cholesterol (Figure 1).

**Citation:** Cabianca, A.; Müller, L.; Pawlowski, K.; Dahlin, P. Changes in the Plant β-Sitosterol/Stigmasterol Ratio Caused by the Plant Parasitic Nematode *Meloidogyne incognita*. *Plants* **2021**, *10*, 292. https:// doi.org/10.3390/plants10020292

Academic Editor: Carla Maleita Received: 30 December 2020 Accepted: 29 January 2021 Published: 4 February 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 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 (https:// creativecommons.org/licenses/by/ 4.0/).

**Figure 1.** Plant sterol synthesis pathway starting with the conversion of 2,3 oxidosqualene to cycloartenol by oxidosqualene cyclase (OSC). OSC enzymes are classed as cycloartenol synthase (CAS) or lanosterol synthase (LAS) depending on their first cyclic product. The main sterol synthesis pathway in plants is indicated by multiple arrows representing several enzymatic steps with detailed information on β-sitosterol conversion to stigmasterol by a C22-desaturase. The most common end sterols in plants are highlighted in gray. The lanosterol synthesis pathway, well known for animal and fungi is indicated by dotted lines as lanosterol synthesis has been reported in plants, although lanosterol was not detected in this study.

Remarkably not all multicellular organisms that require sterols for growth and reproduction are able to synthesize these molecules de novo [7]. Plant parasitic nematodes (PPN), for instance, are among the sterol auxotrophic parasites that rely on host plant sterols for growth and reproduction [7–9]. Several PPN are sedentary endoparasites that burrow inside plant roots and induce the formation of feeding sites, such as the root-knot nematodes, *Meloidogyne* spp. These nematodes induce the formation of giant cells in the

differentiating vascular tissue that act as nutrient sinks, for example, sterols, which the nematode feeds on [10,11].

Biotic and abiotic factors have been reported to cause changes in plant sterol levels. Metabolic changes in β-sitosterol and stigmasterol levels have also been associated with fungal or bacterial infection and were related to the induction of signaling pathways leading to the synthesis of antimicrobial molecules and changes in membrane permeability [2,5,12–15]. Differences in the β-sitosterol/stigmasterol ratio have also been associated with resistance and susceptibility of tomato plants to *Meloidogyne incognita* [16]. Furthermore, studies of Hedin et al. [17] show changes in β-sitosterol/stigmasterol levels after *M. incognita* infection of cotton plant roots. Besides these biotic factors, abiotic stresses, such as drought and temperature, have also been reported to affect plant β-sitosterol and stigmasterol levels [5,18].

Stigmasterol is synthesized from β-sitosterol by a single desaturase reaction that occurs at position C22 of the sterol side chain, catalyzed by the enzyme sterol C22-desaturase that belongs to the cytochrome P450 710 family (EC 1.14.19.41; Figure 1) [19,20]. Little is known about the regulation of β-sitosterol and stigmasterol levels in roots during plant defense against PPN. Thus, understanding how plant sterols change after PPN infection and how these changes influence plant defense might help designing nematode-resistant or tolerant crops, possibly with an altered sterol profile. In this way, to better understand the role of plant sterol composition during nematode infection, we investigated the sterol composition of *Brassica juncea* (brown mustard), *Cucumis sativus* (cucumber)*, Glycine max* (soybean)*, Solanum lycopersicum* (tomato cv. Moneymaker and cv. Oskar) and *Zea mays* (corn), after infection with *M. incognita*. Furthermore, changes in sterol composition were tracked over time and expression levels of sterol C22-desaturase gene followed in tomato cv. Moneymaker.

#### **2. Results and Discussion**

#### *2.1. Plant Sterol Composition*

First, we investigated the profiles of free sterols in the roots of five different agricultural crops, brown mustard, corn, cucumber, soybean and two tomato cultivars (cv. Moneymaker and cv. Oskar) (Figure 2, Table 1). Notably, the cholesterol levels were significantly higher in both tomato cultivars than in the other four crop species. Brown mustard (*B. juncea*) had higher levels of β-sitosterol and lower levels of stigmasterol than all the other species. Significant sterol variations among vegetables, fruits, berries and medicinal plants have been reported [21–23]. However, data available for comparisons of plant root sterol composition are limited. With 80.7% stigmasterol in corn root systems, our data are in agreement with previous reports of Bladocha and Benveniste [24], which showed that sterol composition of corn roots and leaves differed strongly in the ratio of β-sitosterol to stigmasterol. Stigmasterol was the most abundant root sterol and β-sitosterol the most abundant sterol in leaves. In the medicinal plant Cannabis, significant differences in campesterol, β-sitosterol and stigmasterol have been observed between organs, with βsitosterol as the most abundant sterol in stem bark and roots and stigmasterol being most abundant in leaves. Campesterol had the lowest concentration in roots and stem bark compared to β-sitosterol and stigmasterol [23].

**Figure 2.** Free sterol composition in percentage of *Brassica juncea, Cucumis sativus, Glycine max, Solanum lycopersicum* cv. Moneymaker (M) and cv. Oskar (O) and *Zea mays*.



Student's *t*-test was used for comparisons of uninfected vs. infected root systems. \*\*\*, *p* < 0.001; \*\*, *p* < 0.01; \*, *p* < 0.05. ND = not detected. n = minimum of 3 samples.

> Similar to our study where *B. juncea* sterols were composed of 94.1 % β-sitosterol (Figure 2; Table 1), the sterol composition from roots and leaves of the close relative *Brassica*

*napus* is dominated by β-sitosterol [25]. On the other hand, Surjus and Durand [26] reported that β-sitosterol is the prominent plant sterol in roots of soybean cv. Hodgson, which does not match our findings where stigmasterol is the most abundant sterol with 62.4% in soybean cv. Aveline Bio.

*C. sativus* was the only species in this study where no campesterol was detected in the root sterol fraction, which was mainly composed of stigmasterol (Figure 2; Table 1). A study on the sterol composition of selected grains, legumes and seeds has shown that campesterol was also not detected in pumpkin seeds [27], whose sterols were mainly made up of βsitosterol. In another study, neither campesterol, stigmasterol nor β-sitosterol were detected in *C. sativus* fruits, however other sterols were present [21]. Altogether, sterol compositions differ between organs of a plant, and even the same organs of different cultivars of the same species can differ significantly in their sterol composition and abundance [28].

Within plants, conjugated sterols are ubiquitous. However, their profile and relative content can differ among organs, plant developmental stage and environmental signals [29]. The analysis of total sterols (sterol ester and free sterols) and free sterol fraction is included in Table 1. When comparing the total sterol fraction to the free sterol fraction, the abundance of cholesterol and β-sitosterol increased, campesterol maintained a similar relative abundance, while the abundance of stigmasterol decreased. These results indicate that more cholesterol and β-sitosterol are present as steryl esters compared to stigmasterol. Overall, sterol profile changes have been reported for different tissues and conjugated forms [29] and even if a plant sterol, such as cholesterol represents a minor amount of the total sterol fraction of the plant, it can be the most abundant phytosterol in some tissue. For example, the sterol fraction of the phloem exudate of bean and tobacco plants contains over 88% of cholesterol [30].

#### *2.2. Plant Sterol Composition after Meloidogyne Incognita Infection*

The sterol compositions of *M. incognita*-infected *B. juncea, C. sativus*, *G. max*, *S. lycopersicum* cvs. Oskar and Moneymaker, and *Z. mays* roots were determined 21 dpi (Table 1), to allow nematodes to establish and expand feeding sites [10]. Compared to uninfected tomato roots, sterols of cv. Moneymaker and cv. Oskar were composed of 6.5% and 6.1% free cholesterol, 86.7% and 84.7% stigmasterol, 5.0% and 8.0% β-sitosterol and 1.9% and 1.1% campesterol, respectively (Table 1). That means, infection with *M. incognita* led to an overall increase in cholesterol and β-sitosterol and a decrease in stigmasterol. Cholesterol levels increased up to 7.5% in cv. Moneymaker roots and up to 8.2% in cv. Oskar. The highest contribution of cholesterol to the sterol pool was determined in the galls, i.e., the nematode feeding sites, with 12.3% (cv. Oskar) and 10.3% (cv. Moneymaker; Table S2). Yet, the most pronounced sterol change observed 21 days post *M. incognita* inoculation was in the relative abundance of β-sitosterol and stigmasterol. In both tomato cultivars, levels of free β-sitosterol increased from 5.0% to 15.6% and 8.0% to 11.6% in cv. Moneymaker and cv. Oskar, respectively. At the same time, stigmasterol levels decreased from 86.7% to 75% and from 84.7% to 78.7% in infected roots of cv. Moneymaker and cv. Oskar, respectively. These changes in the β-sitosterol/stigmasterol ratio were even more pronounced when the sterol composition of the galls was evaluated (Figure 3; Table S2).

**Figure 3.** Relative stigmasterol to β-sitosterol abundance of uninfected (green), *M. incognita* infected (yellow) and galls (red) for one generation (oneg) and (brown) for second generation (twog) of *M. incognita*, samples of the plants: *Solanum lycopersicum* cv. Moneymaker (**A**) and *Cucumis sativus* (**B**). For *S. lycopersicum* cv. Oskar (**C**), *Glycine max* (**D**), *Zea mays* (**E**) and *Brassica juncea* (**F**) the results are presented as the mean of the 3 replicates. For A and B the average is marked by X. n = ≥3 replicates.

The analysis of the free and the total sterol fraction of *C. sativus*, *G. max* and *Z. mays* roots infected by *M. incognita* showed similar β-sitosterol, stigmasterol and cholesterol changes compared to the control plants (Table 1). Infection with *M. incognita* resulted in a relative increase in cholesterol and β-sitosterol combined with a relative decrease in stigmasterol levels compared to uninfected plants. However, such changes were not observed in *B. juncea*, where infection resulted in an increase in campesterol.

It seems plausible that the observed changes in the sterol pool are linked to a metabolic reaction against the infection by *M. incognita*. For example, in solanaceous plants, cholesterol can make up a significant portion of the overall sterol pool and has been suggested as a precursor of toxic steroidal alkaloids and glycoalkaloids [31]. Campesterol is used in numerous plants as precursor for the synthesis of brassinosteroid phytohormones, essential for the regulation of numerous plant processes, such as cell expansion and elongation, senescence and protection against drought and chilling [32]. The conversion of β-sitosterol to stigmasterol has been linked to biotic and abiotic stress [2,19,20] and has previously been linked to resistance against *M. incognita* in tomato cultivars [16].

As *M. incognita* induces a formation of giant cells, the galls sterol composition might be influenced by the lipid bilayer reorganization of these cells. Studies on the lipid bilayer revealed that β-sitosterol is slightly more efficient in ordering a fluid membrane of 2 dipalmitoyl-sn-glycero-3-phosphocholine than stigmasterol, resulting in a more packed membrane liquid ordered phase [33]. Furthermore, simulations have shown that cholesterol was slightly more efficient in packing the lipid bilayer than β-sitosterol [34].

Since the β-sitosterol to stigmasterol ratio is regulated by a single C22 desaturation step and strong changes in this ratio were observed, scatter plots were prepared to compare β-sitosterol/stigmasterol changes after nematode infection in the different plant species (Figure 3). All plant species analyzed displayed an increase of β-sitosterol and a decrease in stigmasterol after nematode infection, with the exception of *B. juncea*, which showed a decrease of β-sitosterol levels. β-Sitosterol accounted for 94.1% and stigmasterol for only 1.7% of free sterols in non-infected *B. juncea* plants (Table 1). This might be the reason why *B. juncea* displayed a completely different alteration on the sterol profile in response to nematode infection than the other plant species investigated (Table 1, Figures 2 and 3). Anyhow, similar β-sitosterol/stigmasterol observations can be seen for other sterol analyses, e.g., of two cotton cultivars, cv. ST-213 and cv. 81-249 where the β-sitosterol/stigmasterol ratio changed from 32.6/53.1% (cv. St213) and 30.0/43.8% (cv. 81-249) to 36.8/43.8% (cv. ST-213) and 33.8/47.3% (cv. 81-249) after *M. incognita* infection [17].

A reason for the different sterol response in *B. juncea* compared to the other plant species might be that *Brassica* species contain a particular sterol, brassicasterol. Brassicasterol synthesis belongs to the same sterol branch as campesterol (Figure 1). The campesterol precursor 24-methyldesmosterol is converted to 24-epi-campesterol and then to brassicasterol. This final enzymatic step described in *Arabidopsis thaliana* is catalyzed by a C22 desaturase [19]. In this context, it is also important to note that *Brassica* species can produce isothiocyanates (ITCs) the glycosides of which are hydrolyzed by myrosinases in response to herbivory [35]. ITCs are highly toxic, leading to a suppressive effect of *Brassica* species on soil-borne pathogens and herbivores [36]. Therefore, *Brassica* species including *B. juncea* are used as cover crops in PPN management via so-called bio-fumigation [37,38]. Nevertheless, *B. juncea* is a host of *M. incognita* [39].

#### *2.3. β-Sitosterol/Stigmasterol Conversion in Tomato after Meloidogyne Incognita Infection*

The β-sitosterol to stigmasterol conversion requires the creation of a double bond at position C22, which is catalyzed by a monooxygenase of the Cytochrome P450 enzyme family 710A (CYP710A), the only family in the CYP710 clan (Figure 1) [19,40]. The observed increase of β-sitosterol and decrease of stigmasterol led us to investigate the expression of the tomato gene *SlCYP710A11* during *M. incognita* infection. This gene encodes the enzyme previously characterized as a C22 desaturase in tomato sterol biosynthesis [19]. Temporal gene expression analysis of the *SlCYP710A11* gene in uninfected tomato cv. Moneymaker showed only small variations in gene expression levels during a time course of 21 days (Figure 4A). However, in tomato plants of the same developmental stage infected by *M. incognita*, the expression of *SlCYP710A11* was downregulated significantly in the samples taken at 14 and 21 dpi (Figure 4B). At the same time, the tomato sterol profile of β-sitosterol and stigmasterol reflected the gene expression levels (Figure 4C) in that the β-sitosterol/stigmasterol ratio gradually increased over the course of 21 days due to a relative increase of β-sitosterol and a corresponding decrease of stigmasterol (Figure 4C). The β-sitosterol/stigmasterol change was most pronounced at 21 dpi, confirming the previous results on plants infected with *M. incognita* that displayed reduced relative levels of stigmasterol and increased levels of β-sitosterol compared to the uninfected plants, most easily to explain by a decrease in C22 desaturase activity (Figure 4B). Interestingly, the change in the β-sitosterol/stigmasterol ratio was already visible at 6 dpi when transcriptional repression was not apparent yet, suggesting additional regulatory mechanisms (Figure 4B). Altogether, both gene expression and sterol profile data support the finding that the synthesis of stigmasterol from β-sitosterol is downregulated as an effect of *M. incognita* infection in *S. lycopersicum*.

**Figure 4.** Temporal gene expression analysis of the C-22 desaturase gene *CYP710A11* and changes in the β-sitosterol/stigmasterol ratio in *Solanum lycopersicum* cv. Moneymaker 2, 6, 14 and 21 days post inoculation (dpi). *SlCYP710A11* gene expression is presented as fold change. (**A**) Data on uninfected roots are marked in blue. (**B**) Data on *M. incognita*-infected roots (Inf. roots) in orange. N = 4 biological replicates of 2 pooled plants per analysis. (**C**) Changes in β-sitosterol/stigmasterol ratios are displayed as percentage of total sterols extracted. ANOVA was used for comparisons of gene expression levels in uninfected vs. infected root systems. \*, *p* < 0.05.

Since the reaction to *M. incognita* infection is a modulation of C22 desaturase activity on behalf of the plants, it is important to note that the enzyme responsible for the conversion of 24-epi-campesterol to brassicasterol also represents a C22 desaturase; indeed, it was found for *Arabidopsis* that the enzyme encoded by *CYP710A2* was responsible for both brassicasterol and stigmasterol production [19]. However, *M. incognita* infection did not lead to a significant change in the sterol pattern of *B. juncea* (Table 1). Hence, in spite of the fact that brassicasterol was not analyzed, we can conclude that it is unlikely that the expression of the *CYP710A2* orthologue was affected.

Changes in the β-sitosterol/stigmasterol equilibrium might represent a general plant response to environmental cues as reviewed by Zhang et al. [28], and not a specific response to *M. incognita*. For example, an increase in stigmasterol levels has generally been observed as response to cold stress [5,41]. An increase in C22 desaturase expression levels has been reported as response of *Arabidopsis thaliana* plants to biotic and abiotic factors: to inducers of PAMP-triggered immunity like flagellin and lipopolysaccharides, to reactive oxygen species (ROS) and osmotic stress as well as to infections with bacterial and fungal pathogens [3,5,14,15,42]. Other than in *Arabidopsis*, a relative increase in stigmasterol has also been observed in leaves of *Triticum aestivum* infected by a biotrophic fungus, and in *Z. mays* leaves infected by a necrotrophic fungus ([43,44].

While our results seemed to show *CYP710A* gene induction at the first two time points of *M. incognita* infection, these changes were not significant. However, the repression of *SlCYP710A11* expression at 14 and 21 dpi, and the corresponding changes in the βsitosterol/stigmasterol ratio, contrast with the abovementioned studies, where *CYP710A* expression was induced, β-sitosterol levels decreased and stigmasterol levels increased. It has to be kept in mind that most previous studies on plant sterol abundance during plant defense focused on shorter time intervals after exposure to pathogens, above-ground plant organs and were conducted mainly on *Arabidopsis* plants, where β-sitosterol is the most abundant sterol and brassicasterols make up part of the end sterols [14,15,43,44]. Furthermore, *Arabidopsis*, like *B. juncea*, is a member of the Brassicaceae and can produce nematocidal ITCs, which might affect its additional responses to PPN [45]. Altogether, the finding of an increase in the β-sitosterol/stigmasterol ratio in response to PPN infection in a diverse group of plants that do not produce nematocidal toxins might indeed represent a specific response. However, given that this response takes some time to establish, it is possible that it is not part of the defense against PPN, but of the supply of PPN with suitable sterols by the plant.

Given that plant-pathogen interactions are processes with different stages, in which gene expression levels often vary, it is not surprising to see changes in profiles of metabolites, such as sterols that could play a critical role in a plant-nematode interaction. It would also not be surprising that different pathogens/herbivores trigger similar or different plant responses. At this point, additional investigations have to be conducted to (a) compare the effects of PPN vs. other root pathogens/herbivores, and (b) evaluate the impact of the initial plant sterol composition on sterol changes after pathogen attack. After all, in the current study, *B. juncea* had the highest β-sitosterol abundance and was the only outlier in the sterol response to *M. incognita*, presumably due to the fact that Brassicaceae have particular sterol profiles including brassicasterol.

#### *2.4. CYP710A*

CYP710A represents the plant cytochrome P450 monooxygenase family encoding the sterol C22 desaturase, which is converting β-sitosterol to stigmasterol [40]. Like plants, fungi possess C22 desaturase enzymes known as CYP61 family of P450 enzymes, which are experimentally characterized and phylogenetically represent orthologues of the plant CYP710 protein family. Phylogenetic analysis of P450 diversity suggests that the CYP710 family is conserved from green algae to higher plants throughout evolution [46] and that the biochemical function can be traced back to plant-fungal divergence but was lost in animals [40]. During evolution, sterol 14-demethylase (CYP51) gene is assumed to have given rise to the *CYP710/CYP61* genes as their function in sterol biosynthesis is downstream of that of CYP51 [40]. CYP51 enzymes are present in plants, fungi and animals synthesizing sterols.

While the phylogeny of P450 monooxygenases is well researched, only limited phylogenetic information is available for CYP710 [28,40]. Overall, CYP710 enzyme activity and/or gene expression has only been studied in few plants, such as *A. thaliana* [2,14,19], *S. lycopersicum* [19], *Physcomitrella patens* [47] and *Calotropis procera* [48]. Therefore, we conducted a phylogenetic analysis of our studied tomato SlCYP710A11 protein and other

plant CYP710 enzymes (Figure 5; Table S3). The well-studied AtCYP710A1 (*A. thaliana*) and SlCYP710A11 (*S. lycopersicum*) amino acid sequences were used as queries to mine for plant homologues. Four hits were scored in *A. thaliana*: Cytochrome P450 proteins 710A1, 710A2, 710A3 and 710A4 (NCBI accessions NP\_180997.1, NP\_180996.1, NP\_180451.1 and NP\_180452.1). It is worth mentioning that in *A. thaliana* both 710A1 and 710A2 can convert β-sitosterol to stigmasterol [19]. For *Z. mays*, two protein sequences were found in the NCBI database, from two different studies, one annotated as 'uncharacterized protein' and one as CYP710A11 (NP\_001307723.1 and PWZ33314.1, respectively). For *G. max*, two proteins were identified, one annotated as CYP710A1 (XP\_003542931.1) and one as CYP710A11 (XP\_003546088.1). Only one homologous protein was found in *C. sativus* (XP\_004134602.1), also annotated as CYP710A11 (Table S3). Since *B. juncea* sequences were not present in the NCBI or UniProt databases, *Brassica rapa* was used as a close relative.

**Figure 5.** Phylogenetic maximum likelihood tree of the CYP710 enzyme family. The tree is rooted at ERG5, which is the *Saccharomyces cerevisiae* protein from which all CYP710 proteins originated [20]. Multiple sequences of the same plant species are numbered, and accession numbers of all selected proteins are reported in Table S2. Tree branches are colored and grouped by taxon. External ring shows the eudicots apart from the monocots. \* Tomato CYP710A11 enzyme.

> During the blast search, multiple gene duplication events were observed, mostly at the species level (data not shown). The only duplication observed at the family level was found in the Brassicaceae family (whole genome duplication [49]). The phylogenetic analysis showed the divergence of eudicot and monocot CYP710 enzymes and basically followed plant phylogeny (Figure 5).

Based on the sterol analysis of the selected plants, the phylogenetic analysis, and recent studies (e.g., where *C. procera CYP710A* gene expression did not respond to abiotic factors [48]), we cannot conclude that in all plants C22 desaturase gene expression responds the same way to PPN infection. Moreover, not all CYP710A enzymes function the same way in sterol biosynthesis, and there might be undiscovered members of the CYP710A family catalyzing the same, or a different reaction (like the desaturation of 24-epi-campesterol to brassicasterol as reviewed by Zhang et al. [28]). Generally, among plant sterol synthesis enzymes, sterol methyl transferase (SMT), delta (24)-sterol reductase (DWF1) and CYP710A are assumed to adjust end sterol composition [28]. Altogether, further studies are required to address the questions if the observed β-sitosterol/stigmasterol changes are speciesspecific and how additional sterol related genes are involved in the activation of *CYP710A* and changes of the β-sitosterol/stigmasterol equilibrium, and to evaluate their impact on nematode performance. These data might help to develop new nematode-resistant cultivars able to maintain a sterol equilibrium that is not suitable for nematode development.

#### **3. Materials and Methods**

#### *3.1. Nematode Inoculation and Plant Material*

The root-knot nematodes, *Meloidogyne incognita* (isolate Reichenau 2, R2) were maintained at Agroscope (Wädenswil, Switzerland) on *S. lycopersicum* cv. Oskar. Greenhouse conditions were set at 22 ± 2 ◦C, 60% relative humidity (RH) and 16 h/8 h light/dark rhythm. Second-stage juveniles (J2) were extracted from heavily galled root systems using a mist chamber (PM 7/119). J2 were stored at 6 ◦C prior to use [50]. For sterol profiling a minimum of three biological replicates were used per treatment (negative and positive controls) and species:, *Brassica juncea* cv. Sareptasenf (P. H. Petersen), *Cucumis sativus* cv. Landgurken (Bigler Samen) *Glycine max* cv. Aveline Bio (UFA), *Solanum lycopersicum* cultivars (cvs.) Moneymaker (HILDA) and Oskar (Syngenta) and *Zea mays* cv. Grünschnittmais (UFA) were used. Seeds were pre-germinated (*B. juncea* 3–5 days, *C. sativus* 2–3 days, *G. max* 4–6 days, *S. lycopersicum* 4–6 days and *Z. mays* 5–6 days) in Petri dishes with 5 mm of tap water and then planted into 14 cm diameter plastic pots, using a 3:1 (vol/vol) silver sand:steamed soil mixture (sieved field soil from Cadenazzo, Switzerland). Greenhouse conditions were set to 22 ± 4 ◦C, 60% RH and 16 h:8 h light:dark rhythm. Three four-weekold plants of each species/cultivar were inoculated with 10,000 *M. incognita* (R2) J2 per pot.

#### *3.2. Sterol Extraction and GC-MS Analysis*

Infected and uninfected (control) plant roots were washed free of soil 21 days post inoculation (dpi). For "galls" sterol analysis, galled uproot systems were manually separated with a scalpel. Roots and galls were washed and the separated materials shock-frozen in liquid nitrogen, and ground to powder using mortar and pestle. Sterols were extracted according to Bligh and Dyer [51]. Each root-powder sample (1 g) was separated into two equal parts and total lipids were extracted in chloroform:methanol (2:1 *v*/*v*) for 1 h at 60 ◦C. One of the two lipid fractions was further saponified for extraction of free and esterified sterols. Saponification was performed as described by Dahlin et al. [52] (alkaline saponification with 2M KOH in 95% ethanol). Both lipid fractions (saponified and total lipid extract) of each root sample were dried under nitrogen and processed for sterol separation by suspending the dried samples in hexane and using a silica solid phase extraction (SPE) column (6 mL SiOH columns, Chromabond, Macherey Nagel, Düren, Germany) as described by Azadmard-Damirchi and Dutta [53]. Eluted sterols were dried under nitrogen and suspended in chloroform for sterol analysis on the Varian 450-GC coupled to a Varian 240-MS Ion Trap (GC-MS) (Darmstadt, Germany). The software VARIAN MS Workstation v. 6.9.3 was used for instrument control and data acquisition. A VARIANT FactorFour Capillary column VF-5 ms of 30 m length, 0.25 mm inner diameter, and 0.25 μm film thickness was used as stationary phase. Helium was used as carrier gas at a flow rate of 1.0 mL/min. Inlet temperature was set at 320 ◦C. 10 μL of the chloroform sample were

injected. Initial GC temperature was set at 225 ◦C and ramped up to 300 ◦C at 1.5 ◦C/min. Temperature was maintained at 300 ◦C for 10 min before ramping to 320 ◦C with 5 ◦C/min, and finally remaining stable at 320 ◦C for 6 min. Transfer line was set to 270 ◦C and ion trap temperature was 150 ◦C. Ion trap was operated with electron ionization (EI) set at an ionization energy of 70 eV and scan mode selection (*m*/*z* 50–900) started after 5 min solvent delay. Sterol standards (cholesterol, campesterol, β-sitosterol and stigmasterol) were obtained from Sigma-Aldrich (St. Louis, MO, USA) and used to compare retention times, sterol fragmentation and for relative sterol quantification. The software R (v. 3.6.2; R core team, 2018) was used to perform Student's *t*-tests (*t*-tests) and ANOVA (analysis of variance) tests on the data obtained to investigate the statistical differences between samples. *T*-tests were used when only infected and uninfected samples were compared, ANOVA was performed when gall samples were included in the comparison.

#### *3.3. CYP710A11 Temporal Gene Expression Analysis*

Tomato cv. Moneymaker plants were grown as described above. 4000 *M. incognita* J2/plant were inoculated by pipetting equal amounts of nematodes into four 5 cm deep holes next to three-week-old tomato plants. 8 Plants were used per time point and pooled in 4 groups of 2 plants each. Plant roots were harvested from infected and uninfected plants at 2, 6, 14 and 21 dpi, frozen in liquid nitrogen and stored at −20 ◦C before RNA extraction in liquid nitrogen using the Thermo Scientific GeneJET Plant RNA Purification Mini Kit (Waltham, MA, USA). Genomic DNA was removed from the isolated RNA using iScript DNase, followed by RNA quality testing by agarose gel electrophoresis and NanoDrop One One/OneC Microvolume UV-Vis Spectrophotometer measurements (Thermo Fisher Scientific, Reinach, Switzerland). cDNA synthesis was performed using the iScript cDNA synthesis kit (Bio-Rad, Hercules, CA, USA). The tomato gene coding sequence of *SlCYP710A11* was used to design qPCR primers with the online tool Primer3 (v. 4.1.0, Whitehead Institute for Biomedical Research), with the setting of 20 nt primer sequence length, 110 to 130 bases of amplified fragments, 50% GC content and 60 ◦C melting temperature. Primer sequences (Table S1) were BLASTed against WormBase and NCBI databases to check target specificity. The same parameters were used to design qPCR primers for the reference genes. NormFinder statistical algorithms were used to evaluate the housekeeping gene stability of actin, α-tubulin, SlCBL1, GADPH and eEF1-α. Primer efficiency was determined using the program Real-time PCR Miner [54]. qPCR analyses were carried out according to the 480 SYBR Green 1 Master mix (Roche, Basel, Switzerland) protocol and optimized to the primer melting temperature of 60 ◦C on the Roch LightCycler 480. For each qPCR run, the Roche LightCycler 480 program was used for melting peak and temperature evaluation. Each experiment was normalized according to the reference gene expression of actin and α-tubulin. Relative fold-changes in expression levels were analysed in Excel using 2(−ΔΔCt) [55].

#### *3.4. Phylogenetic Analysis of Cytochrome P450 Proteins*

The protein sequences of *A. thaliana* AtCYP710A1 and *S. lycopersicum* SlCYP710A11, retrieved from the UniProtKB (UniProt) database, were used as queries in a sequence similarity search, performed on the UniProt and National Center for Biotechnology Information (NCBI) databases. The number of CYP710A1 proteins and their accession numbers were recorded for the plant species used in the sterol analysis. Protein sequences were searched for conserved protein domains using the Pfam (v. 32, European Bioinformatics Institute) and PANTHER protein databases. AtCYP710A1 was also used as query in a BLAST on Phytozome database (v12.1.5) [56]. Retrieved cytochrome P450 710 protein sequences were aligned using MUSCLE with the software MegaX (Molecular Evolutionary Genetics Analysis X). Aligned sequences were used in MegaX for phylogenetic analysis using the Maximum Likelihood approach, with 1000 bootstraps. The online tool iTOL (interactive Tree Of Life, v. 5.6) was used to finalize the phylogenetic tree.

#### **4. Conclusions**

In this study, we report changes in plant sterol profiles, in response to infection by the plant parasitic nematode *M. incognita*. The β-sitosterol/stigmasterol ratio in *C. sativus*, *G. max*, *S. lycopersicum* cv. Moneymaker and cv. Oskar and *Z. mays* were strongly affected by *M. incognita*. Interestingly, *B. juncea* revealed a sterol response different from that in the other plants examined. Since the conversion of β-sitosterol to stigmasterol is mediated by a single desaturation reaction at position C22 of the sterol side chain catalyzed by CYP710A, we investigated the transcriptional response of tomato *SlCYP710A11*. Infection of *S. lycopersicum* cv. Moneymaker with *M. incognita* led to repression of *SlCYP710A11* transcription that paralleled the change in the β-sitosterol/stigmasterol ratio. However, a detailed comparison indicates that the change in expression levels was not the only factor changing the sterol profile. Further studies are required to investigate whether the changes in plant sterol composition were specific to the response to *M. incognita* infection, if other nematode species generate the same changes in plant sterol composition, and whether they can represent a resistance mechanism.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/2223-7 747/10/2/292/s1, Table S1: Primer pairs used for qPCR analysis of tomato (*Solanum lycopersicum*), Table S2: Sterol composition (%) of tomato (*Solanum lycopersicum*) and cucumber (*Cucumis sativus*) galls caused by *Meloidogyne incognita*, Table S3: List of CYP710 enzyme sequences used for the phylogenetic analysis.

**Author Contributions:** Conceptualization, P.D., A.C., K.P., L.M.; methodology, P.D., A.C., L.M. and K.P.; A.C. and L.M. performed the experiments, with input from P.D. and K.P.; data curation, A.C., P.D., L.M. and K.P.; writing—original draft preparation, A.C.; manuscript finalized by A.C., P.D. and K.P. with input from L.M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research did not receive any specific funding from granting agencies in the public, commercial, or nonprofit sector.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Data is contained within the article or supplementary material.

**Acknowledgments:** We thank the nematology team at Agroscope for their consistent support in the laboratory and greenhouse. The authors also acknowledge Thomas Eppler for his technical support on the GC-MS and Andrea Caroline Ruthes for their helpful comments, discussions, and corrections throughout the study.

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

#### **References**


## *Article* **Bacterial Microbiota Isolated from Cysts of** *Globodera rostochiensis* **(Nematoda: Heteroderidae)**

#### **Violeta Oro 1,\*, Magdalena Knezevic 2, Zoran Dinic <sup>2</sup> and Dusica Delic <sup>2</sup>**


Received: 16 July 2020; Accepted: 1 September 2020; Published: 4 September 2020

**Abstract:** The potato cyst nematode (PCN) *Globodera rostochiensis* is a plant parasite of potato classified into a group of quarantine organisms causing high economic losses worldwide. Due to the long persistence of the parasite in soil, cysts harbor numerous bacteria whose presence can lead to cyst death and population decline. The cysts of *G. rostochiensis* found in two potato fields were used as a source of bacteria. The universal procedure was applied to extract DNA from bacteria which was then sequenced with 16S primers. The aims of the study were to identify bacterial microbiota associated with the PCN populations and to infer their phylogenetic relationships based on the maximum likelihood and Bayesian phylogeny of the 16S sequences. In addition, the impact of the most significant climate and edaphic factors on bacterial diversity were evaluated. Regarding the higher taxonomy, our results indicate that the prevalent bacterial classes were Bacilli, Actinobacteria and Alphaproteobacteria. Phylogenetic analyses clustered *Brevibacterium frigoritolerans* within the family Bacillaceae, confirming its recent reclassification. Long-term climate factors, such as air temperature, insolation hours, humidity and precipitation, as well as the content of soil organic matter, affected the bacterial diversity. The ability of cyst nematodes to persist in soil for a long time qualifies them as a significant natural source to explore the soil bacterial microbiota.

**Keywords:** potato cyst nematodes; Bacilli; Actinobacteria; Alphaproteobacteria; 16S; maximum likelihood; Bayesian inference; climate and edaphic factors

#### **1. Introduction**

Bacteria are ubiquitous organisms, inhabiting even the most extreme environments like polar snow [1], volcanoes and acidic hot springs [2,3]. The natural soil environment, aside from other microorganisms, harbors as many as 106–108 bacterial cells and 106–107 actinomycete cells per 1 g and around 107 nematodes per 1 m<sup>2</sup> [4].

The potato cyst nematodes (PCNs) *Globodera rostochiensis* and *G. pallida* are plant parasites of potatoes and other Solanaceae plants, classified as quarantine organisms. PCN females are sedentary organisms living inside potato roots with numerous eggs within their enlarged spherical bodies called cysts. The nematodes develop within the eggs to first and second stage juveniles. The latter is the invasive stage, searching for the appropriate host plant. When they find a target host, they start to invade roots, penetrating the host tissue with their stylets and move inside it. Inside the root tissue, they develop into females and males. After mating and fertilization, new eggs and juveniles are produced within the cysts, so the parasitic cycle continues. Some juveniles do not hatch until the following season or favorable conditions, remaining in soil for a long time [5]. The potato cyst nematodes cause up to GBP 300M worth of damage to the potato crop in the EU each year [6].

Both *Globodera* species were brought to Europe with the introduction of potato from South America [7]. Because the PCNs persist in soil, the external and internal areas of cysts harbor numerous microorganisms whose presence can lead to cyst death and population decline, suggesting that they can be potential candidates for use in biocontrol. Microscopic counts using 5-(4,6-dichlorotriazine-2-yl) aminofluorescein staining and in situ hybridization (EUB 338) revealed that cysts contain 2.6 <sup>×</sup> 105 bacteria [8].

Diverse bacterial species have been reported as nematode antagonists. *Streptomyces avermitilis* and *Pseudomonas fluorescens* were found to possess anthelmintic properties [9]. Nine isolates belonging to *Pseudomonas* and *Streptomyces* species were found to control both fungal pathogens and *Meloidogyne incognita* and were considered as promising biological control agents [10]. Bacterial isolates that inhibited egg hatching of the potato cyst nematodes were mostly from the genus *Bacillus* [11]. Bacterial species of the genus *Pasteuria* were found to be parasites of *Meloidogyne*, *Belonolaimus*, *Pratylenchus*, *Heterodera*, and *Globodera* spp. [12]. The Gram-negative bacterium *Stenotrophomonas (Xanthomonas) maltophilia* G2 was found to have a high nematotoxic activity against the free-living nematode *Panagrellus redivivus*, and the plant parasitic nematode *Bursaphelenchus xylophilus* [13]. *Serratia*, *Curtobacterium*, *Pseudomonas*, *Pantoea*, and *Rhanella* species were nematotoxic toward *B. xylophilus* [14]. Treatment with *B. cereus* strain S2 had a lethal effect on *Caenorhabditis elegans* and *M. incognita* [15].

This study aims to: (i) identify bacterial species associated with two PCN populations, (ii) infer phylogenetic relationships of the bacteria based on the maximum likelihood (ML) and Bayesian inference (BI) of 16S sequences rRNA genes, (iii) evaluate the influence of some microclimate and edaphic factors on bacterial diversity.

#### **2. Results and Discussion**

The results revealed that bacterial microbiota from the locations of Pozega and Krupanj (the Republic of Serbia) generally contain similar species with varying abundance. The cysts obtained from Pozega have more diverse bacterial microbiota (Figure 1) with the presence of 74.0% of members of the class Bacilli and the order Bacillales divided into the families Bacillaceae and Paenibacillaceae. Furthermore, there are 14.0% of members of Proteobacteria, whereas Actinobacteria are present in the lowest percentage (6.0%). The Alphaproteobacteria are represented by the order Rhizobiales and the family Hyphomicrobiaceae (*Devosia* sp.), while Actinobacteria are represented by the order Micrococcales and the family Brevibacteriaceae i.e., *Brevibacterium* sp. The bacterial microbiota of Krupanj (Figure 2) is less diverse, containing the majority of the class Bacilli (40.0%), represented by the families Bacillaceae and Paenibacillaceae as well. The next group is Actinobacteria (28.0%) with the family Micrococcaceae and *Arthrobacter* spp., while the lowest percentage (20.0%) pertains to Alphaproteobacteria represented by the family Hyphomicrobiaceae and *Devosia* sp.

**Figure 1.** Bacterial microbiota found in cysts from Pozega.

**Figure 2.** Bacterial microbiota found in cysts from Krupanj.

The genus *Bacillus* was the principal genus in Pozega, which was similar to Costa et al. [16] (p. 718), who observed that *Bacillus* was present in 80% of the isolates of the bacterial microbiota of *M. exigua* egg masses in coffee plantations.

*Bacillus* was found not only to be prevalent in the rhizosphere, but also in the phyllosphere. Maximum colonization was shown by the genus *Bacillus* isolated from carrot, cabbage and turnip phyllosphere bacteria [17]. Members of the order Bacillales (*B. pumilus* and *P. xylanexedens*) were found in both locations. In contrast, more Actinobacteria were detected in Krupanj, suggesting that this location was probably more polluted with organic contaminants and the processes of natural bioremediation occurred. In Krupanj, *Arthrobacter* spp. corresponded to 28% of the total bacterial microbiota; likewise, the genus *Arthrobacter* comprised more than 21% of the total soil community of the burned holmoak forest [18].

In comparison with two soil samples from Spain, analyzed by the denaturing gradient gel electrophoresis of bacteria isolated from *M. incognita* and *P. penetrans*, in which the most abundant bacterial classes were Betaproteobacteria, Bacilli and Actinobacteria [19], in our study, the prevalent classes were Bacilli, Actinobacteria and Alphaproteobacteria. The dominance of the order Bacillales was evident in both locations with 80% in Pozega and twice less in Krupanj. In contrast, more Actinobacteria and Alphaproteobacteria (*Arthrobacter* spp. and *Devosia* sp., respectively) were detected in Krupanj.

The phylogenetic analyses based on 16S sequences are shown in the Figures 3 and 4. Both ML and BI trees are in agreement and generated three distinct clades. Within the first clade, there are subclades composed of *Bacillus cereus*, *B. megaterium*, *B. flexus*, *B. subtilis*, *B. pumilus* and a *Psychrobacillus* species, representing the family Bacillaceae. The other subclade with *Paenibacillus* spp. represents the family Paenibacillaceae, which, together with the family Bacillaceae, are affiliated to the order Bacillales and the phylum Firmicutes. The difference is that *Devosia* spp. are independent in the ML tree (Figure 3). The *Devosia* species clade represents the family Hyphomicrobiaceae and Alphaproteobacteria linked with the two subclades of Actinobacteria, the subclade of *Arthrobacter* spp. and the subclade of *Brevibacterium* species in the BI tree, because the Bayesian inference considers all the species to be monophyletic (Figure 4). The sequences of *Brevibacterium frigoritolerans* were not clustered with other *Brevibacterium* species. Instead, they were grouped with *Bacillus cereus* species as the closest relatives, suggesting their affiliation to the family Bacillaceae.

**Figure 3.** Maximum likelihood phylogenetic tree of bacterial microbiota isolated from *G. rostochiensis* cysts based on 16S sequence region using General Time Reversible (GTR), invariable sites and gamma distribution (GTR + I + G) nucleotide evolution model.

**Figure 4.** Bayesian phylogenetic tree of bacterial microbiota isolated from *G. rostochiensis* cysts and derived from consensus 50% majority rule based on 16S sequence region using GTR + I + G nucleotide evolution model.

Similar observations were reported by other authors. *Brevibacterium frigoritolerans* was in the same group with other *Bacillus* spp., i.e., *B. simplex*, *B. muralis*, *B. psychrosaccharolyticus* [20–22]. This bacterium can biosynthesize silver nanoparticles and tolerate silver as some *Bacillus* species can tolerate salt [23]. In addition, *B. frigoritolerans* has the ability to sporulate, thereby providing evidence that this strain is actually a misidentified *Bacillus* sp. [20]. Recently, based on the phenotypic, chemotaxonomic, phylogenetic and genomic characteristics, it has been demonstrated that *B. frigoritolerans* DSM 8801T should belong to the genus *Bacillus*, and to be reclassified as *Bacillus frigoritolerans* [24]. Our study confirms its reclassification and genetic closeness to *B. cereus*. On the other hand, the other species of *Brevibacterium* were clustered together with *Arthrobacter* spp. within Actinobacteria. Apart from *G. rostochiensis*, this species was isolated from juveniles of *B. xylophilus* [25]. Under in vitro bioassay conditions, the isolate of *Brevibacterium frigoritolerans* exhibited bacteremia-like symptoms and induced mortality of the Coleopteran larvae of *Anomala dimidiata* and *Holotrichia longipennis* [26], suggesting its possible use in biocontrol.

Comparisons based on climate factors during the 28-year period (1990–2018) revealed differences between the two locations. Pozega shows the lower values of the air temperatures (optimum, minimum and maximum) and insolation, and the higher values of relative humidity, and cloudiness (Table 1). On the contrary, the values of temperatures, insolation hours and precipitation are higher in Krupanj, whereas the values of relative humidity are lower and there are fewer cloudy days (Table 1).


**Table 1.** Comparison of annual means and honest significant difference (HSD) of climate factors for two observed locations during the 28-year period (1990–2018).

\* statistically significant.

The honest significant difference (HSD) test demonstrates that there are statistically significant differences between insolation hours and precipitation values. The difference in insolation between locations is almost 500 h with more variation of this factor in Pozega. In contrast, the precipitation sum was higher in Krupanj throughout the year. The insolation itself has a direct impact on the air temperature, making the distinction of this factor between the two locations. A decrease in air temperature causes the decrease in soil temperature, which, in combination with higher relative humidity, favors the environment suitable for cold tolerant species. This fact was confirmed by the presence of *Bacillus frigoritolerans* and a *Psychrobacillus* species in Pozega. Despite the fact that there are no significant differences in air temperature at two locations, the lower annual temperatures in Pozega favored the development of psychrotolerant species. In climate studies, statistical significance does not always provide an adequate basis for decision making; for example, a rise in temperature by two degrees Celsius may not be statistically significant but it can adversely affect the vegetation growth and lead to ecological imbalances [27].

All of the physicochemical properties of the soils, except the content of potassium, were similar in both locations (Table 2). However, the content of soil organic matter in Pozega is higher than in Krupanj, which may explain the more diverse bacterial microbiota in Pozega. Soil with a higher content of organic matter is generally associated with high microbial abundance and diversity [28].


**Table 2.** Comparison of soil physicochemical parameters and HSD for two observed locations.

\* statistically significant.

The HSD test demonstrates that there is a significant difference in the amount of potassium between the two locations. Since K<sup>+</sup> is a major nutritional element for plants, enrichment of K<sup>+</sup> in the exchange sites due to fertilizer practice can be expected [29], which may indicate high potassium fertilizer inputs in Krupanj.

Regarding the granulometric content of the two examined soils, the smallest clay and silt particles (0.002–0.02 mm) are dominant: 77.7 versus 80.9%. Pozega has a higher content of silt, whereas Krupanj has a higher content of clay. With decreasing particle size, there is an increase in particle number and in the surface area per gram of soil. It is clear that the interfacial area enlarges with an increase in the proportion of the clay–size fraction and, consequently, the opportunities for sorptive interactions between microorganisms and soil particles should increase [30]. The dominance of silt and clay in both soil samples enables good interaction between bacteria and soil.

All found species of the family Bacillaceae have been reported to have high potential as biocontrol agents, which resulted in the development of commercial bionematicidal agents [12]. *Bacillus cereus* strain S2 can produce sphingosine to induce reactive oxygen accumulation, destroy the genital area in nematodes, and inhibit nematode reproduction [15].

*Bacillus pumilus* demonstrated its ability as a potential biocontrol agent against *M. arenaria*, causing 39.8 and 92.8% J2 mortality after three days of exposure to 2.5 and 10% concentrations of bacterial culture, respectively [31]. *Bacillus subtilis* and *B. pumilus* caused the highest reduction (82% and 81.8%, respectively) in *M. incognita* on cowpea [32]. An isolate of *Bacillus megaterium* reduced the root penetration and migration of *M. graminicola* to between 40 and 60% compared with non–treated roots of rice plants [33].

*Paenibacillus nematophilus* has been found to hamper more than 98% of the dispersal of the beneficial nematode *Heterorhabditis megidis* and reduce its infectivity in moth larvae [34].

*Psychrobacillus* species play a role in biodegradation and as antimicrobial agents. *Psychrobacillus soli* could degrade around 72% of oil components at an initial oil concentration of 1500 ppm [35]. Among ten endophytic bacteria, *Psychrobacillus insolitus* and *Curtobacterium oceanosedimentum* showed the highest anticandidal effect against *Candida albicans* and *C. glabrata* [36], while two strains of *P. insolitus* (Mam2 and Ame3) exhibited an inhibitory action against staphylococcal strains isolated from food [37].

*Devosia* and *Arthrobacter* species are best characterized for their bioremediation potential. *Devosia* are well known for their dominance in soil habitats contaminated with various toxins. The uptake and utilization of nutrients for growth and survival was found to be the dominant function of the genus along with the detoxification and degradation of organic pollutants [38].

*Arthrobacter* species were involved in biodegrading a wide variety of compounds, e.g., nicotine, organosilicon compounds, fluorene, the herbicide atrazine [39], and *m*-chlorobenzoate, the central molecule in many pesticides [40]. The majority of the selected strains exhibited a great ability to degrade organic polymers in vitro. Moreover, they possibly present a direct mechanism for plant growth promotion [18]. One of the strains of *A. nicotianae* showed 100% nematicidal activity against *C. elegans* and 91–97% nematicidal activity against *M. incognita* [41].

The higher presence of bioremediators in our samples may indicate the higher presence of pollutants in Krupanj and explain the reduced diversity of bacterial microbiota.

#### **3. Materials and Methods**

#### *3.1. Isolation of Bacteria*

The cysts of *G. rostochiensis* found in potato fields near the locations of Pozega (44◦04 N 20◦14 E) and Krupanj (44◦18 N 19◦20 E) were used as a source for screening bacterial microbiota. During the growing season, the soil samples were taken as 50 subsamples/ha in a systematic sampling pattern in order to make approximately one kilogram of composite sample [42].The cyst extraction was done with the Spears apparatus [43] and collected on a 150-μm sieve.

Fifty randomly selected cysts of different ages from each location were surface sterilized with 96% ethanol, 1.5% NaOCl and washed with sterile water according to the procedure applied for *Globodera* juveniles [44]. The cysts were placed on potato dextrose agar (PDA) and maintained for seven days at 25 ◦C. After the emergence of bacteria on PDA, single bacterial colonies were used to obtain pure cultures by the streakplate method [45].

#### *3.2. Molecular Study*

The extraction of DNA from bacteria was performed according to a previously described procedure [46]. The PCR reaction mixture consisted of 25 μL 2× PCR Mastermix, 0.5 μL of forward and reverse primers (10 μM), 1 μL of DNA template and PCR-grade water to a total volume of 50 μL. Amplification of the DNA region coding for 16S rRNA was performed by using P0 (5 -GAGAGTTTGATCCTGGCTCAG-3 ) and P6 (5 -CTACGGCTACCTTGTTACGA-3 ) primers. The temperature profile for the PCR reaction was as follows: 95 ◦C for 90 s followed by 35 cycles consisting of 95 ◦C for 30 s, the annealing temperature (60 ◦C for the first 5 cycles, 55 ◦C for the next 5 cycles, and 50 ◦C for the last 25 cycles) for 30 s, and 72 ◦C for 4 min. The reaction mixture was then incubated at 72 ◦C for 10 min and at 60 ◦C for 10 min. The obtained PCR products were purified and sequenced [47]. Phylogenetic analyses were performed with sequences of the isolated bacterial species deposited under accession numbers MT394477-MT394483 (Pozega) and MT410635-MT410639 (Krupanj) and related species from the GenBank nucleotide sequence database, using maximum likelihood (ML) and Bayesian inference (BI) phylogenetic methods. The ML and BI were calculated with the help of PhyML 3.1 [48], and MrBayes 3.1.2 [49] computer programs, respectively. The sequence alignment was done with ClustalW in Mega 4 [50].

The ML tree was obtained with the General Time Reversible model (GTR), invariable sites and gamma distribution (GTR + I + G). The dendrogram obtained by Bayesian inference was created by 2.2 <sup>×</sup> 10<sup>6</sup> generations of Markov Chain Monte Carlo, with a sample frequency of 100, and burning function of 20%. The nucleotide evolution model was GTR + I + G as well. Branch supports higher than 70% were shown next to the node.

#### *3.3. Statistical Data Analysis*

The annual values of climate factors of Pozega and Krupanj were obtained from the official site of the Republic Hydrometeorological Institute of Serbia. The 28-year period (1990–2018) was used for calculating the means of the optimum, maximum, and minimum air temperature, the relative humidity, insolation, cloudiness and precipitation.

The units for the air temperatures were presented in degrees Celsius, the relative air humidity was expressed in percentages, while the duration of the solar radiation (insolation) was expressed in hours. Values of the cloudiness parameter lower than 2 were considered as clear days, while values higher than 6 were considered as cloudy days. The precipitation was expressed in millimeters (Table 1). Soil pH, the content of organic matter, the amount of nitrogen, phosphorus and potassium, as well as the soil granulometric composition (Table 2), were determined according to standard methods and those from the literature [51–55]. The values were compared with a post-ANOVA Tukey's honest significant difference (HSD) test using DSAASTAT computer program [56], at the 95% confidence interval. Values with the same letter were not significantly different from each other.

#### **4. Conclusions**

Regarding the higher bacterial taxonomy, our results indicate that the observed locations have similar microbiota, but with a different abundance and species identity. The dominant bacterial phyla are Firmicutes, Actinobacteria and Proteobacteria. Based on 16S sequences, the maximum likelihood and the Bayesian phylogeny clustered the members of the genus *Bacillus, Psychrobacillus* and *Paenibacillus* within the family Bacillaceae. *Brevibacterium frigoritolerans* belonged to the same group with *B. cereus, B. megaterium* and *B. flexus* within the family Bacillaceae, confirming its recent reclassification. Other clades were occupied by *Devosia* and *Arthrobacter* species known for their function in environmental detoxification and the degradation of pesticides. The lower values of air temperatures, insolation, and precipitation and the higher values of relative humidity and cloudiness created conditions for the development of psychrophilic species. The location of Pozega is characterized by psychrotolerant representatives of *Bacillus frigoritolerans,* and a *Psychrobacillus* species. In contrast, Krupanj is characterized by the higher content of potassium, the lower content of organic matter and the presence of bioremediators such as *Devosia* and *Arthrobacter* species. In other words, bacterial species perform as specific indicators of microclimate properties and environmental pollution.

As efforts have been moved towards expanding the source of microorganisms involving the more complex systems in nature [57], nematodes and their related bacterial microbiota present the next biological system to explore the taxonomic diversity of soil bacteria. Nematodes, especially cyst nematodes, are a significant natural source of microorganisms due to their long persistence in soil and the specific environmental conditions inside and outside of the closed area of cysts, in which diverse bacteria are hidden.

**Author Contributions:** Conceptualization, V.O. and D.D.; data curation, M.K. and Z.D.; formal analysis, V.O., M.K. and Z.D.; investigation, V.O., M.K. and Z.D.; methodology, D.D., V.O. and Z.D.; supervision, D.D.; writing—original draft, V.O., M.K. and Z.D.; writing—review and editing, D.D. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by the SerbianMinistry of Education, Science and Technological Development.

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

#### **References**


© 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/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Plants* Editorial Office E-mail: plants@mdpi.com www.mdpi.com/journal/plants

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel: +41 61 683 77 34

www.mdpi.com

ISBN 978-3-0365-5464-8