Plant-soil feedback of two legume species in semi-arid Brazil

Abstract

Positive feedback between arbuscular mycorrhizal fungal (AMF) and vascular plants can contribute to plant species establishment, but how this feedback affects plant invasion by Prosopis juliflora SW. (DC.), or resistance to invasion by Mimosa tenuiflora (Willd.) Poir in Brazilian semi-arid region is not well known. In this work, we tested how modified and native AMF communities affect the establishment of P. juliflora and M. tenuiflora plants. We examined the effects of inoculation with modified and native AMF communities on number of AMF spores, root colonization, number of N-fixing nodules, plant dry biomass, plant phosphorous concentration, and plant responsiveness to mycorrhizas of P. juliflora and M. tenuiflora. We found that the modified AMF community enhanced the root colonization, plant dry biomass, and plant phosphorous concentration of invasive P. juliflora, whereas native AMF enhanced M. tenuiflora. Our results demonstrate that the invasive P. juliflora alters soil AMF community composition, and this change generates positive feedback to the invasive P. juliflora itself and decreases AMF associations with native M. tenuiflora.

Introduction

Arbuscular mycorrhizal fungi (AMF) are obligate symbionts which have different levels of host-plant specificity. These symbionts form an efficient root association with vascular plants [1] and they are considered to play a key role in mediating the biological invasion process [2, 3]. Many works demonstrated that the exotic plant species can have their growth improved by certain AMF species [4, 5] and the invader can also display a more positive, plant-soil feedback than native plant species, in their new environmental range [6]. So, understanding the ecological processes that regulate AMF-plant interaction in the Brazilian tropical seasonal dry forest is essential to explain the success of some exotic plant species, such as Prosopis juliflora SW. (DC.) in colonizing disturbed and abandoned agricultural areas in the Caatinga ecoregion [7, 8].

Usually, exotic plant species are adapted to high resource availability in their natural habitats [9]. Once introduced in a new habitat that is characterized by soils with low fertility as Regosols from Brazilian semi-arid region, these plant species invest strongly in increasing resource availability (e.g., Island of fertility hypothesis proposed by Richardson et al. [2] and Lekberg et al. [3]) and thus, contribute to their establishment and spread. Changes in soil chemical properties may have direct impacts on AMF community composition, where a native AMF community (e.g., AMF community that occurs in the root zone of native plant species) that evolved during many years in a native root zone and was adapted to natural conditions like low fertility, become a modified AMF community that has all its community structure altered in benefit of the invader and adapted to the new soil conditions like pH and P availability. There are many works that demonstrate changes in AMF communities during exotic plant invasion [5, 10–13], but the mechanisms underlying these changes and their consequences on host’s growth have not been well elucidated.

Biological invasion is recognized as one of the most important ecological disturbances affecting the dynamics of plant communities in the Brazilian semi-arid region [14]. There is evidence that P. juliflora invades and dominates degraded areas in this ecoregion [7]. Increasing biological invasion by this exotic plant species is characterized by the invader removing competitively superior native plant species and increase nutrient availability [15]. But biological invasion can also modify the mutualistic communities, like arbuscular mycorrhizal fungi, altering their structure and functioning [11, 12], which may influence the direction or magnitude of plant-soil feedback on P. juliflora and Mimosa tenuiflora (Willd.) Poir. To our knowledge, it is not known if changes in soil communities induced by biological invasion differentially alter arbuscular mycorrhizal fungi interactions on P. juliflora (an exotic plant species) relative to M. tenuiflora (a native plant species).

The main objective of this study was to determine whether biological invasion disturbance differentially influences plant-soil interaction in exotic vs native plant individuals. Also, we investigated whether the influence of inoculum soil sources from invaded (modified AMF community) and non-invaded (native AMF community) soils on net soil effects (plant-soil feedback) changed after the invasion by P. juliflora upon a biological invasion process that occurred 20 years before. Exotic plant species are known to change soil chemical and microbial processes, which is suggested to generate positive plant-soil feedback [11, 16, 17]. The subsequent invasion and dominance of a plant such as P. juliflora after its introduction in a new geographical range may affect soil mutualistic community causing positive feedback and favoring its own growth. Conversely, it may lead to growth reductions of native plants by supporting to a lesser degree its associated mutualistic communities [18]. Our work addressed the following question: how P. juliflora and M. tenuiflora respond to the inoculation of AMF community from invaded (modified AMF community) and native (native AMF community) soils. We expected P. juliflora to show higher mycorrhizal colonization, rhizobial colonization, biomass production, and plant responsiveness to mycorrhizas after inoculation with modified AMF communities. This expectation is based on the disturbance-contingent niche creation model proposed by Kulmatiski and Kardol [9]. To accomplish this, we combined field sampling of two root zone types (e.g., P. juliflora and M. tenuiflora) and four different sites of Brazilian semi-arid region, a trap culture bioassay to produce AMF inoculum, and a bioassay with host plants. The bioassays allowed us to investigate soil feedback effects on plant dry biomass, root colonization, plant P concentration, and plant responsiveness to mycorrhizas.

Materials and methods

Plant species and studied sites

The invasive P. juliflora, and the native M. tenuiflora, which co-occur in the Brazilian semi-arid region, were selected for this study. P. juliflora has been invasive in this ecoregion for 20 years [7, 14] and M. tenuiflora is a common native tree from Brazilian xeric shrubland [19]. Both P. juliflora and M. tenuiflora are highly dependent on AMF in drought conditions [20, 21]. The seeds of both species were collected directly from the studied sites.

We collected samples from the root zone of both model plants in four different sites of Brazilian semi-arid region: S1—Algodão de Jandaría, PB, Brazil (06° 46′ 17.3″ S 36° 01′ 55.3″ W, 493 m); S2—Carnaúba dos Dantas, RN, Brazil (06° 35′ 09.47″ S 36° 29′ 41.99″ W, 682 m); S3—Esperança, PB, Brazil (06° 56′ 45.7″ S 35° 54′ 06.8″ W, 588 m); and S4—Poçinhos, PB, Brazil (07° 10′ 23.3″ S 36° 12′ 41.8″ W, 522 m). The climatic conditions of the studied sites are classified as hot semi-arid climate (Type Bsh following Köppen-Geiger climate classification) with hot summers and mild to warm winters, annual precipitation, and temperature of 600 mm and 30 °C respectively. Rainfall in these sites is scanty, unpredictable, and irregular [ 22]. The soil type of the four studied sites was classified as a Regosol [23].

Field sampling

We sampled in each of the four sites 40 plots of 100 m [2] according to Fortin and Dale [24]. Within each plot, we sampled 40 individuals of P. juliflora and 40 individuals of M. tenuiflora. So, in total, 80 individuals were sampled according to the following criteria: (a) the stem has a diameter near soil surface > 3 cm; (b) plant individual height varying from 1 to 3 m; (c) no individual from a different plant individuals were growing in a 3-m radius to the sampling point in all directions [24–28]. Root zone samples (including soil and root fragments) were collected near the drip line and beyond (0–20 cm deep), during the dry season (i.e., at the beginning of September 2014). By sampling during the dry season, we guaranteed to sample the largest number of AMF spores because fungal sporulation is expected to be higher at this time of the year in semi-arid ecoregion [20]. Samples from each studied condition were bulked, mixed, and stored at 4 °C until both soil chemical characterization and host-plant bioassays. Later, each soil sample collected from the field was divided into portions intended for chemical soil characterization, AMF community assessment, and mounting trap cultures.

Soil characterization

For chemical characterization, we analyzed soil pH, total organic carbon, total nitrogen, and available phosphorus. Soil pH was measured in a suspension of soil and distilled water (1:2.5 v:v, soil: water suspension) [29]. Total organic carbon was determined by rapid dichromate oxidation method according to the methodology described by Okalebo [30]. To quantify total nitrogen, soil samples were first digested with sulfuric acid plus potassium sulfate, and we then followed the protocol described in Kjeldahl [29]. Available phosphorus (Olsen’s P) was determined colorimetrically using a spectrophotometer at 882 nm by extraction with sodium bicarbonate for 30 min [31]. The chemical characterization of the soil from each studied plant individual and site are given in Table 1.

Table 1.

Chemical properties of soil by each studied plant individual (Prosopis and Mimosa) in S1 (Algodão de Jandaíra), S2 (Carnaúba dos Dantas), S3 (Esperança), and (S4) Poçinhos, in the Brazilian xeric shrubland (mean ± SD, N = 320)

Soil properties	S1—Algodão de Jandaíra		S2—Carnaúba dos Dantas		S3—Esperança		S4—Poçinhos
	P. juliflora	M. tenuiflora	P. juliflora	M. tenuiflora	P. juliflora	M. tenuiflora	P. juliflora	M. tenuiflora
pH (1:2.5)	6.5 ± 0.3	4.9 ± 0.3	6.6 ± 0.2	5.6 ± 0.5	6.7 ± 0.2	5.2 ± 0.1	7.7 ± 0.4	6.7 ± 0.3
TOC (g kg−1)	8.9 ± 0.7	9.3 ± 2.3	7.4 ± 0.9	2.6 ± 0.4	8.4 ± 0.3	2.0 ± 0.1	14.7 ± 4.1	6.9 ± 1.7
N (g kg−1)	0.9 ± 0.1	0.3 ± 0.1	0.9 ± 0.2	0.2 ± 0.1	0.8 ± 0.1	0.2 ± 0.1	1.1 ± 0.5	0.4 ± 0.2
P (mg kg−1)	9.7 ± 0.3	3.1 ± 0.8	9.8 ± 1.1	3.6 ± 0.4	8.4 ± 0.3	2.0 ± 0.1	12.3 ± 2.3	8.9 ± 1.0

Bioassay experiment and factorial design

AMF communities extracted from P. juliflora and M. tenuiflora root zones were classified as “modified AMF community” if they occurred in the root zone of invasive plant species only, and as “native AMF community” if they occurred in the root zone of native plant species only. These were used to establish trap culture of each AMF community for each studied condition in sterilized 2-L plastic pots to later be propagated on a common host plant (Z. mays L.). To establish the initial fungal inoculum (i.e., soil, roots, and spores), we used a portion of soil (500 g) from the field. We sterilized the seeds of Z. mays in 10% sodium hypochlorite for 10 min, and then these were thoroughly rinsed with deionized water. This species is a standard host plant used for mycorrhizal trap cultures and inoculum potential assays according to the INVAM (http://invam.caf.wvu.edu/). Z. mays plants were grown in a mixture 3:1 (m:m) of sterilized sand to field soil during 4 months in a glasshouse at the University of Coimbra. This was the period required until sporulation occurred. The average temperature in the glasshouse was 28 °C, ranging from 20 to 35 °C, an irradiance of up to 70% of full sun, relative humidity ranging between 65 and 75%, and a photoperiod of 16 h/8 h light/dark. The host plants received weekly amounts of a modified nutrient solution [32] containing 554.0 mg L⁻¹ KCl, 200.0 mg L⁻¹ NaH₂PO₄·H₂O, 2.24 mg L⁻¹ MgSO₄, 520.0 mg L⁻¹ CaCl₂·H₂0, 1.7 mg L⁻¹ MnSO₄, 0.25 mg L⁻¹ CuSo₄·5H₂O, 0.30 mg L⁻¹ ZnSO₄·7H₂0, 5.0 mg L⁻¹ NaCl, 3.0 mg L⁻¹ H₃O₃, 0.09 mg L⁻¹ (NH₄)₆Mo₇O₂₄·4H₂O, and 32.9 mg L⁻¹ Na-Fe EDTA. The nutrient solution was previously sterilized by UV radiation to avoid external contamination.

The bioassay experiment involved a fully factorial design with two soil inoculum origins (modified and native AMF communities), four sources of soil inoculum (S1, S2, S3, and S4), two soil inoculum treatments (non-sterilized and sterilized), and two host plants (P. juliflora and M. tenuiflora) with twenty repetitions per treatment combination. The sterilized treatment received 500 mL of filtrate from 500 g of AMF inoculum with no mycorrhizal spores obtained by sieving through a 25-μm mesh and 500 g of sterilized mixed AMF inoculum (a mix of all AMF inoculum by host plant), thereby controlling for potential mineral and non-mycorrhizal microbial components of the AMF inoculum.

The experiment was conducted in a glasshouse with average temperatures of 25/16 °C (day/night). Seeds of P. juliflora and M. tenuiflora were surface sterilized with 10% sodium hypochlorite for 10 min, and then thoroughly rinsed with deionized water. Seeds were then germinated in sterilized trays containing sterilized sand (sterilized twice at 121 °C for 20 min each in two consecutive days). Seven days after emergence, seedlings (plant height varied between 2 and 4 cm) were selected and individually transferred to sterilized plastics pots containing 2000 g of a substrate obtained by mixing 1500 g of sterilized sand mixed with 500 g of containing fungal inoculum. The number of infective propagules ranged between 8.52 and 9.23 propagules per gram of soil. All pots were covered with aluminum wrap around the seedling to prevent dehydration and external contamination, and plants were watered with sterilized water as necessary and fertilized with the same nutrient solution as in the trap culture [32].

AMF community characterization

Spores and sporocarps from field and trap cultures were extracted by the wet sieving technique [33] followed by sucrose centrifugation [34]. Initially, the extracted spores were examined in water under a dissecting microscope and they were separated based on morphological characteristics. Subsequently, they were mounted in polyvinyl alcohol in lacto-glycerol (PVLG) with and without the addition of Melzer’s reagent [35]. Species identification was based on the descriptions provided by Schenck and Perez [36], and by consulting the on-line AMF collection of the Department of Plant Pathology, University of Agriculture in Szczecin, Poland (http://www.agro.ar.szczecin.pl/~jblasxkowski/) and the International culture collection of arbuscular mycorrhizal fungi database—INVAM (http://invam.caf.wvu.edu). In this work, we adopted the classification proposed by Oehl et al. [37], including recently described new taxa [38–40].

Data collection

Five months after planting, host plants were harvested. Roots were separated from shoots, and the fresh roots were weighed immediately. A total of 1 g of each fresh root sample was used for determination of root colonization following the gridline intersect method [41, 42]. To estimate shoot and root dry biomass, the remaining root material and shoots were oven-dried at 72 °C for 48 h. Roots of P. juliflora and M. tenuiflora were examined for quantification of N-fixing nodules. After quantification, the N-fixing nodules were washed carefully to remove soil debris. Dry weight of nodules was recorded. The plant responsiveness to mycorrhizas (PRM)—“response”—was calculated using the values of dry biomass of P. juliflora and M. tenuiflora from the AMF-inoculum treatment and the control, by using the formula: PRM = [(B_AMF − B_control)/B_AMF] × 100, where B_AMF is the dry biomass of host plant in the AMF-inoculum treatment and B_control is the dry biomass of the control treatment [43]. To estimate the total plant P content, a known mass of the dry material was first digested in a 10:1:4 mixture of HNO₃:H₂SO₄:HCl (60%), and then analyzed using the vanadate molybdate colorimetric method [44].

Statistical analyses

The Shapiro-Wilk test was applied to assess the normality of the data distribution. Two-way ANOVA (AMF communities vs studied sites) was used to test for the effect of fungal inoculum and inoculum source on plant dry biomass, PRM, number of N-fixing nodules, and plant P concentration for each studied host plant in the bioassay. In the control treatments, spore number, root colonization, and PRM index were zero and consequently, they were excluded from the statistical analyses. Data sets were arcsin square root transformed for percentage variables and log₁₀(x) for the remaining. When necessary, Tukey’s LSD post hoc comparison tests were conducted. All statistical analyses were conducted using SAS 9.1.3 Portable.

Results

Effect of biological invasion by P. juliflora on AMF community composition

Significant differences between modified and native AMF communities were found for AMF frequency of occurrence in all studied sites (Table 2). In total, we identified 19 different AMF species corresponding to 11 genera: Acaulospora [2], Claroideoglomus [1], Dentiscutata [2], Entrophospora [1], Funneliformis [3], Gigaspora [3], Glomus [1], Quatunica [1], Racocetra [1], Rhizoglomus [2], and Scutellospora [1]. The most abundant taxa were species from the Order Glomerales in all studied sites from the modified AMF community, whereas AMF species from Order Diversisporales were mostly found in the native AMF community. The modified AMF community considering all studied sites was composed of 13 AMF species of those, six from the Order Diversisporales and seven from the Order Glomerales, whereas native AMF community was composed of 19 AMF species, 13 similar to modified AMF community. We also found higher dominance of Funneliformis mosseae in all studied modified AMF communities (AMF frequency of occurrence of 51.4, 64.3, 51.0, and 51.0% in S1, S2, S3, and S4, respectively), when compared with native AMF communities (1.5, 4.7, 2.8, and 5.0%, respectively) (Table 2).

Table 2.

AMF frequency of occurrence of each studied condition observed from spores extracted from field soil and trap culture. FO (%) observed in the P. juliflora and M. tenuiflora root zone, and their AMF classification (Cla)

AMF species	S1—Algodão de Jandaíra				S2—Carnaúba dos Dantas				S3—Esperança				S4—Poçinhos
	Modified AMF		Native AMF		Modified AMF		Native AMF		Modified AMF		Native AMF		Modified AMF		Native AMF
	FOa	Clab	FO	Cla	FO	Cla	FO	Cla	FO	Cla	FO	Cla	FO	Cla	FO	Cla
Order Diversisporales
Family Acaulosporaceae
Acaulospora denticulata Sieverd. & S. Toro	–	–	5.0	R	–	–	8.1	R	–	–	5.0	R	–	–	2.6	R
Acaulospora tuberculata Janos & Trappe	–	–	1.6	R	–	–	2.3	R	–	–	2.5	R	–	–	–	–
Family Dentiscutataceae
Dentiscutata cerradensis (Spain & J. Miranda) Sieverd., F.A. Souza & Oehl	1.0	R	1.8	R	1.48	R	3.1	R	0.8	R	1.5	R	0.5	R	2.0	R
Dentiscutata heterograma (T.H. Nicolson & Gerd.) Sieverd., F.A. Souza & Oehl	6.5	R	7.1	R	1.78	R	11.7	C	7.6	R	7.5	R	2.0	R	8.9	R
Quatunica erythropus (Koske & C. Walker) C. Walker & D. Redecker	2.5	R	0.5	R	–	–	–	–	2.4	R	0.6	R	–	–	–	–
Family Gigasporaceae
Gigaspora albida N.C. Schenck & G.S. Sm.	–	–	10.5	C	–	–	10.9	C	–	–	12.5	C	–	–	6.8	C
Gigaspora decipiens I.R. Hall & L.K. Abbott	–	–	5.5	R	–	–	8.3	R	–	–	5.5	R	–	–	5.8	R
Gigaspora gigantea (T.H. Nicolson & Gerd.) Gerd. & Trappe	–	–	5.5	R	–	–	6.2	R	–	–	7.0	R	–	–	6.6	R
Family Racocetracea
Racocetra coralloidea (Trappe, Gerd. & I. Ho) Oehl, F.A. Souza & Sieverd.	5.3	R	3.0	R	8.63	R	–	–	4.4	R	2.1	R	–	–	–	–
Family Scutellosporaceae
Scutellospora calospora (T.H. Nicolson & Gerd.) C. Walker & F.E. Sanders	1.0	R	1.0	R	–	–	–	–	0.7	R	1.0	R	0.5	R	2.0	R
Order Glomerales
Family Entrophosporaceae
Entrophospora infrequens (I.R. Hall) R.N. Ames & R.W. Schneid.	–	–	0.6	R	–	–	5.3	R	–	–	0.6	R	–	–	10.8	C
Claroideoglomus etunicatum (W.N. Becker & Gerd.) C. Walker & Schuessler	–	–	13.5	C	–	–	–	–	3.0	R	10.5	C	13.2	C	10.7	C
Family Glomeraceae
Funneliformis caledonium (T.H. Nicolson & Gerd.) C. Walker & Schuessler	5.7	R	12.6	C	2.97	R	11.1	C	3.0	R	13.0	C	10.9	C	10.0	C
Funneliformis geosporum (T.H. Nicolson & Gerd.) C. Walker & Schuessler	3.0	R	12.8	C	–	–	–	–	3.0	R	12.5	C	7.1	R	10.5	C
Funneliformis mosseae (T.H. Nicolson & Gerd.) C. Walker & Schuessler	51.4	D	1.5	R	64.32	D	4.7	R	51.0	D	2.8	R	51.0	D	5.0	R
Glomus multicaule Gerd. & B.K. Bakshi	5.3	R	5.0	R	3.86	R	8.2	R	4.8	R	4.9	R	2.9	R	8.5	R
Rhizoglomus aggregatum (N.C. Schenck & G.S. Sm.) Sieverd., G.A. Silva & Oehl					2.08	R	12.3	C	–	–	–	–	–	–	–	–
Rhizoglomus clarum (T.H. Nicolson & N.C. Schenck) Sieverd., G.A. Silva & Oehl	18.3	C	12.5	C	14.88	C	7.8	R	19.3	C	10.5	C	11.9	C	9.8	R

^aFO_i = n_i/N, where n_i is the number of times an AMF species was observed and N is the total of AMF spores observed from each studied condition; ^bclassification of AMF frequency of occurrence described by Souza et al. [13]: R, rare (FO < 10%); C, common (10 ≤ FO ≤ 30%); MC, most common (31 ≤ FO ≤ 50); and D, dominant (FO > 50%)

Fungal inoculum effect on P. juliflora and M. tenuiflora performance

Significant differences between AMF inoculum (F_{1, 152} = 106.75, p < 0.001; F_{1, 152} = 23.83, p < 0.001; F_{1, 190} = 68.3, p < 0.001), AMF origins (F_{3, 152} = 41,92, p < 0.001; F_{3, 152} = 18.91, p < 0.001; F_{4, 190} = 15.60, p < 0.001), and their interaction (F_{3, 152} = 34.93, p < 0.001; F_{3, 152} = 15.03, p < 0.001; F_{4, 190} = 13;50, p < 0.001) were found for the number of AMF spores, root colonization, and N-fixing nodules biomass from P. juliflora plants respectively. For M. tenuiflora plants, the two-way ANOVA showed a significant effect of the AMF inoculum (F_{1, 152} = 62.63, p < 0.001; F_{1, 152} = 95.62, p < 0.001; F_{1, 190} = 42.9, p < 0.001), AMF origins (F_{3, 152} = 107.06, p < 0.001; F_{3, 152} = 37.27, p < 0.001; F_{4, 190} = 46.9, p < 0.001), and the interaction between these two factors (F_{3, 152} = 2.08, p = 0.105; F_{3, 152} = 21.23, p < 0.001; F_{4, 190} = 21.9, p < 0.001) on number of AMF spores, root colonization, and the N-fixing nodules biomass. The number of AMF spores, the root colonization, and the N-fixing nodules biomass varied significantly among the two AMF inoculums and the four AMF origins for both P. juliflora and M. tenuiflora plants. For the AMF-inoculum effects, the highest values of these variables in P. juliflora plants were found with modified AMF inoculum, whereas for M. tenuiflora plants, the highest values of the number of spores, root colonization, and the N-fixing nodules biomass were found with native AMF inoculum (Table 3).

Table 3.

Results from the number of spores, root colonization, and N-fixing nodules biomass by Prosopis juliflora and M. tenuiflora plants, AMF origin (S1, S2, S3, S4, and control), and AMF inoculum (modified and native AMF communities) (mean ± SD)

Studied plant/AMF origin	Number of spores (g soil−1)		Root colonization (%)		N-fixing nodules biomass (g plant−1)
	Modified	Native	Modified	Native	Modified	Native
P. juliflora
S1	5.9 ± 0.1 aA	5.2 ± 0.2 bB	40.2 ± 3.1 aC	17.7 ± 4.7 bC	0.14 ± 0.04 aC	0.09 ± 0.02 bB
S2	6.1 ± 0.1 aA	5.6 ± 0.2 bA	43.7 ± 1.1 aB	23.8 ± 6.1 bA	0.32 ± 0.02 aA	0.17 ± 0.04 bA
S3	5.0 ± 0.1 aB	4.8 ± 0.1 aC	38.0 ± 1.5 aC	20.9 ± 3.4 bB	0.32 ± 0.01 aA	0.15 ± 0.04 bA
S4	6.3 ± 0.4 aA	5.1 ± 0.1 bB	47.8 ± 2.7 aA	19.6 ± 3.1 bB	0.22 ± 0.07 aB	0.11 ± 0.06 bB
Control*	–	–	–	–	0.09 ± 0.02 aD	0.05 ± 0.01 bC
M. tenuiflora
S1	3.83 ± 0.40 bC	5.52 ± 0.17 aA	22.7 ± 3.2 bB	51.9 ± 3.7 aB	0.12 ± 0.01 bB	0.31 ± 0.01 aA
S2	4.31 ± 0.31 bA	5.07 ± 0.21 aC	20.5 ± 1.8 bC	53.9 ± 2.9 aB	0.11 ± 0.02 bB	0.30 ± 00.3 aA
S3	4.23 ± 0.27 bB	5.38 ± 0.15 aB	31.0 ± 3.8 bA	57.0 ± 5.2 aA	0.18 ± 0.03 bA	0.28 ± 0.03 aA
S4	3.91 ± 0.16 bC	5.28 ± 0.18 aB	23.3 ± 2.3 bB	61.4 ± 4.3 aA	0.16 ± 0.01 bA	0.27 ± 0.02 aA
Control*	–	–	–	–	0.15 ± 0.01 aA	0.15 ± 0.03 aB
Exotic vs nativea	8.71**	2.31ns	19.33**	21.20**	9.33**	9.13**

Within AMF inoculum (modified and native), means with different lower case letters are significantly different by LSD test at the 5% significance level. Within each studied site, means with different capital letters are significantly different by LSD test at the 5% significance level. *We did not include the control treatments for spore number, and root colonization when LSD test was used to compare these variables between AMF origin and AMF inoculum; ^aindependent sample t test comparing P. julifora and M. tenuiflora groups; ns, not-significant; and **p ≤ 0.01; S1, Algodão de Jandaíra; S2, Carnaúba dos Dantas; S3, Esperança; S4, Poçinhos

There was a significant effect of the AMF inoculum (F_{19, 172} = 66.89, p < 0.001), AMF origins (F_{19, 172} = 36.73, p < 0.001), and the interaction between these two factors (F_{19, 172} = 20.59, p < 0.001) on plant dry biomass from studied plants. P. juliflora performed well in the modified AMF inoculum, whereas M. tenuiflora performed well in the native AMF inoculum. No significant differences in plant dry biomass for both P. juliflora and M. tenuiflora plants were found among the four origins of both AMF inoculum, but these origins were different from the control (Fig. 1).

Fig. 1.

Plant individuals (P. juliflora and M. tenuiflora), AMF-inoculum (modified and native), and AMF-origins (S1, S2, S3, S4, and control) interaction effect on plant dry biomass (mg plant⁻¹) from P. juliflora and M. tenuiflora plants (means ± SD). Within AMF inoculum, means with different lower case letters are significantly different by LSD test at the 5% significance level. Within each studied site, means with different capital letters are significantly different by LSD test at the 5% significance level. S1, Algodão de Jandaíra; S2, Carnaúba dos Dantas; S3, Esperança; S4, Poçinhos

The two-way ANOVA showed a significant effect of the AMF inoculum (F_{19, 172} = 90.89, p < 0.001), AMF origins (F_{19, 172} = 34.19, p < 0.001), and the interaction between these two factors (F_{19, 172} = 14.81, p < 0.01) on plant phosphorous concentration from studied plants. There is a higher positive effect of modified AMF inoculum on plant phosphorous concentration in P. juliflora plants, which had lower values than the native AMF inoculation. On the other hand, there is a higher significant positive effect of native AMF inoculum on plant phosphorous concentration for M. tenuiflora plants (Fig. 2).

Fig. 2.

Plant individuals (P. juliflora and M. tenuiflora), AMF-inoculum (modified and native), and AMF-origins (S1, S2, S3, S4, and control) interaction effect on plant phosphorous concentration (g kg⁻¹) for P. juliflora and M. tenuiflora plants (means ± SD). Within AMF inoculum (modified and native), means with different lower case letters are significantly different by LSD test at the 5% significance level. Within each studied site, means with different capital letters are significantly different by LSD test at the 5% significance level. S1, Algodão de Jandaíra; S2, Carnaúba dos Dantas; S3, Esperança; S4, Poçinhos

For P. juliflora plants, response to AMF inoculum ranged between 60 and 65% for modified AMF inoculum, and 35 and 50% for native AMF inoculum, whereas for M. tenuiflora, it ranged between 20 and 23% for modified AMF inoculum, and 58 and 59% for native AMF inoculum. The two-way ANOVA showed a significant effect of the AMF inoculum (F_{15, 140} = 59.93, p < 0.001) on dependency index from M. tenuiflora plants; no main effects were found for AMF origins (F_{15, 140} = 4.48, p < 0.05), and the interaction between AMF inoculum and AFM origin (F_{15, 140} = 5.95, p < 0.05) (Fig. 3).

Fig. 3.

Plant individuals (P. juliflora and M. tenuiflora), AMF-inoculum (modified and native), and AMF-origins (S1, S2, S3, S4, and control) interaction effect on response to AMF inoculum (%) for P. juliflora and M. tenuiflora plants (means ± SD). Within AMF-inoculum (changed and unaltered), means with different lower case letters are significantly different by LSD test at the 5% significance level. Within each studied site, means with different capital letters are significantly different by LSD test at the 5% significance level. S1, Algodão de Jandaíra; S2, Carnaúba dos Dantas; S3, Esperança; S4, Poçinhos

Discussion

In our experiment, arbuscular mycorrhizal fungi were able to form a mycorrhizal association with P. juliflora and M. tenuiflora, independently of the AMF inoculum or AMF origin. Both AMF inoculums (e.g., modified and native AMF inoculum) increased the number of spores, the root colonization, and the N-fixing nodules biomass of both P. juliflora and M. tenuiflora. We found evidences that the performance of P. juliflora on the pots with modified AMF community was higher than its performance on pots with the native AMF community. Our results indicated that the biological invasion by P. juliflora alters the original native AMF community, that in turns the modified AMF community promotes the invader’s growth. These results support our hypothesis that P. juliflora may experience a stronger positive effect from modified AMF community inoculation and agree with the work done by Zhang et al. [11] that found positive feedback between modified AMF community and Solidago canadensis L. (an invasive plant species from Northeastern China). They also reported that the invader altered the spore composition of the initial AMF community in that AMF species that were most beneficial for invader growth were promoted at the expense of AMF species that were most beneficial to the native plant species. We found that the Order Diversisporales were positively correlated with the rhizosphere of M. tenuiflora in all studied sites, whereas the Order Glomerales and Funneliformis mosseae were positively correlated with the rhizosphere of P. juliflora. These results agree with the work done by Zhang et al. [11], Souza et al. [13], and El-Keblawy and Al-Rawai [15] that reported high abundance of AMF species from Order Diversisporales in native plant species and F. mosseae as a dominant AMF species in the rhizosphere of P. juliflora.

P. juliflora promoted AMF species from Glomerales order like genera Funneliformis (F. mosseae) and Rhizoglomus (Rhizoglomus clarum), while the same genera of the Diversisporales order were absent (Acaulospora and Gigaspora) or had their frequency decreased (Quatunica, Dentiscutata, and Scutellospora) in the modified AMF community. In this condition, we can assume that the inoculation of a modified AMF community where there is a high abundance of AMF species from Order Glomerales affects positively plant invasion by P. juliflora [15].

On the other hand, M. tenuiflora promoted all AMF species from both Diversisporales and Glomerales order, without any dominant AMF species within these orders. In the field conditions, where there was a well-preserved natural ecosystem with a native AMF community under influence of several native plant species, such as M. tenuiflora, the resistance to invasion can be promoted [13]. This is an important result for biodiversity and conservation of native flora of Brazilian semi-arid region. In this ecoregion, well-preserved natural ecosystems are scant [8], and our results highlight the importance to consider the maintenance and management of both natural ecosystems and native AMF community in the Brazilian semi-arid region as a tool against the biological invasion process.

In the native AMF community, we found that the Diversisporales order was composed of genera Acaulospora, Dentiscutata, Quatunica, Gigaspora, and Scutellospora. These genera are common in semi-arid sites [20, 45, 46] and acid soils with low phosphorous availability [47]. Several previous studies have shown the presence of these genera in the root zone of native plant species [46, 48, 49], and there are no reports about dominant genera in the M. tenuiflora rhizosphere [19]. Our results support this lack of dominance of specific genera or species of AMF associated with native AMF community. Our results for native AMF community, showing a higher abundance of Gigasporales, favor the hypothesis proposed by Goto et al. [38] and Marinho et al. [50] that Brazilian sites under natural conditions represent a diversification and dispersion center of this order of arbuscular mycorrhizal fungi. These results are important to further studies considering the ecological process and the positive feedback between AMF community and native plant species on resistance to invasion by M. tenuiflora in Brazilian semi-arid region.

The differences between the AMF communities from the root zone of both P. juliflora and M. tenuiflora may explain why the mutualism formed between the modified AMF community and P. juliflora was more effective when compared with M. tenuiflora inoculated with the modified AMF community and vice-versa. Prior experiments have shown that the AMF colonization differs significantly among vegetation types [12, 13] and invasive plants can reduce overall AMF abundance [51] and degraded AMF relationship with native plant species [18, 52]. In our study, the invasive P. juliflora is a highly mycorrhizal host that decreased the total richness and diversity of AMF community, and increases the dominance of the same AMF species from Glomerales order in the modified AMF community composition. Based on these results, we suggest that the decrease in mutualism between the modified AMF community and M. tenuiflora was due to the change of diversity and dominance in the AMF community that had been changed by P. juliflora. Other studies also showed that the presence of invasive plants changed AMF assemblages in the roots of native plant species [53].

These results support our second hypothesis that changes in soil resource availability in a degraded or abandoned agricultural area can create a window of opportunity, where the invader can be more responsive to modified AMF community than native species. For example, in our experiment, we observed that the inoculation with modified AMF community resulted in a lower plant dry biomass and plant P concentration for M. tenuiflora plants, whereas resulted in higher values of these variables for P. juliflora. Many works have demonstrated a positive effect of the inoculation with AMF species (autochthonous or allochthonous) from Glomerales order on plant growth and nutrient uptake of introduced plants, like inoculation of Funneliformis mosseae, Rhizoglomus aggregatum (described as Glomus aggregatum), Rhizoglomus intraradices (described as Rhizophagus intraradices), and Septoglomus constrictum (described as Glomus constrictum) [54–60].

Our data on the response to AMF inoculum demonstrated that P. juliflora performed well in a modified AMF community, while M. tenuiflora performed well in a native AMF community. Moreover, the effect of modified AMF inoculation detected in our experiment further demonstrated that this reduction was due to a change in the diversity of AMF community, which could be most effective for M. tenuiflora before P. juliflora invasion. It is well documented that invasion by exotic species can reduce AMF abundance and disrupt mutualism of native plants [18, 52, 61]. By using P uptake in the current study, we demonstrated that an exotic plant species can reduce the nutrient-acquiring functions of mutualisms by degrading the preferred mutualisms of native plant species [17].

Here we demonstrate that (i) there is a positive feedback between the AMF communities and both studied plants; (ii) Prosopis juliflora experienced a positive feedback with the modified AMF inoculum, which indicates that this kind of AMF community promotes the biological invasion processes in the Brazilian semi-arid region; (iii) M. tenuiflora performed well in a native AMF community, which indicates resistance to invasion in well-preserved ecosystems; and (iv) the biological invasion promoted by Prosopis juliflora changed the outcome of modified AMF interaction with M. tenuiflora plants, and the modified AMF community may promote the invasion success of the invader in field conditions. These results indicated that the effects of modified AMF community responded more positively to the exotic plant species than the native plant species [12, 62].

The competition between some both invasive and native plant species can be mediate by AMF species that affect differently the growth of the hosts by different levels of exchange carbon with host [63, 64]. We found that modified AMF community by P. juliflora in which Order Glomerales became dominant, and AMF genera from Diversisporales order like Acaulospora and Gigaspora were absent, the response to modified AMF inoculum of M. tenuiflora was lower than P. juliflora. These results suggest that by inoculating the modified AMF community, higher positive feedback is generated to the invasive host with increased of its growth, and lower positive feedback to the native host when decreased the AMF species that could favor own growth as described by Bever [65].

Our work increases the understanding of the ecological mechanisms underlying how plant-AMF interactions can participate in feedback affecting biologic invasion by Prosopis juliflora. Exotic plant species may promote certain AMF species that facilitate their establishment [66], or the invader may change AMF community, that in turns reduce the performance of native plant species [18]. Our results indicate that, under greenhouse conditions, the performance of P. juliflora on the pots where the modified AMF community was inoculated generated better results to plant growth and plant phosphorus concentration to the exotic plant species than the native plant species (M. tenuiflora in our study). This finding highlights the importance to consider a modified AMF community as a key factor promoting the biological invasion process as described by Mummey and Rillig [4].

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest.