Introduction

Maize (Zea mays L.) is one of the world’s most important food crops (Shiferaw et al. 2011). In Southern Brazil, the crop is grown mainly in smallholder farms (CONAB 2020), which conserve a high number of landraces in the region, a maize diversity micro center (Costa et al. 2017). Maize benefits from the association with arbuscular mycorrhizal fungi (AMF), and inoculation can double grain yield in low-P soils (Stoffel et al. 2020).

Plant breeding has resulted in maize varieties and hybrids that are highly productive, but require large amounts of fertilizers and agrochemicals (Machado et al. 2008; Brzozowski and Mazourek 2018). Landrace and modern hybrids, either genetically modified or non-modified, are grown with diverse management, input use, plant densities, as well as different methods of weed control, soil tillage, and fertilizer application. The association between plants and AMF can also be affected by genotypes and environmental factors, including temperature and humidity (Walter et al. 2016; Püschel et al. 2020), and soil attributes, such as aeration, organic matter, phosphorus (P), nitrogen (N), and pH (Mello et al. 2006; Moreira et al. 2007). Current management for high-yielding genotypes uses various pesticides, which can harm some groups of microorganisms that possess the product's target enzymes (Busse et al. 2001).

Some studies have shown that improved modern maize genotypes have lower AMF colonization rates than landrace genotypes (Aquino 2003; Morales-Londoño 2019). Kaeppler et al. (2000) found differences in response to AMF among maize genotypes in a soil with low P and responsiveness to AMF, suggesting that plant breeding may produce cultivars that are less responsive to association with those fungi. AMF fungi are dependent on the symbiont plant to function and reproduce (Liu 2010), and the effect of different maize genotypes on arbuscular mycorrhizal fungi (AMF) is still little understood (Cheeke et al. 2014), it is necessary to study the interaction of AMF and different maize genotypes.

We aimed to characterize the mycorrhizal association in the most widely used maize landraces and modern genotypes in a maize diversity micro center in Southern Brazil. The genotypes used in the experiment are the most widely used in the region, correspond to 22% of maize sown area of western Santa Catarina (Santa Catarina State Department of Agriculture and Fisheries, 2016; unpublished data). Landraces and modern genotypes were tested in a preliminary assay in greenhouse conditions and subsequently in maize fields for two consecutive years. We evaluated root mycorrhizal colonization and the effect of different plant and soil management procedures on the association.

Materials and methods

Greenhouse study

The experiment was carried out in a greenhouse in Florianópolis, Brazil (27°35′54.1"S 48°30′56.4"W) using an Aquic Quartzipsamment soil (Soil Survey Staff 1999), with the following attributes: organic matter 45 g dm−3, pH in water 4.73, resin-extracted P 6.8 mg kg−1, exchangeable K 16.0 mg kg−1 (Tedesco et al. 1995). The AMF community had 6124 spores 50 cm−3 (same soil used in Morales-Londoño et al. 2020). The soil was limed to reach pH 5.7 using 8.5 g limestone kg soil−1, homogenized, and placed in 5-L pots. Maize seeds were washed, disinfected with 70% alcohol (30 s), 2% sodium hypochlorite (2 min), and rinsed with sterile water (Sauer and Burroughs 1986). Three maize seeds were sown in each pot, and plants were thinned to one per pot one week after emergence.

Treatments included ten maize genotypes: five landraces, and five conventional modern genotypes (four hybrids, and one open pollination variety) (Table 1). The experimental design was completely randomized with five replicates. Pots periodically received distilled, non-sterilized water to maintain 70–100% water retention capacity.

Table 1 Morphological groups and thousand-grain weight of maize modern genotypes and landraces
Full size table

Each pot received 0.42 g of P and 2.83 g of K, split into two applications (at sowing and 30 days later), 0.45 g of N, split into three applications (at sowing and at 20 and 40 days), and a micronutrient solution (Hoagland and Arnon 1950) at sowing. The micronutrient solution contained 2.86 mg L−1 H3BO3, 1.81 mg L−1 MnCl2, 0.10 mg L−1 ZnCl2, 0.04 mg L−1 CuCl2, 0.02 mg L−1 H2MoO4.H2O, and Fe-EDTA (24.9 mg L−1 FeSO4.7H2O and 33.2 mg L−1 of EDTA).

At 60 days, plants were collected and we measured height, stem diameter, shoot and root biomass, mycorrhizal colonization (arbuscules, vesicles, and total), and soil AMF spore number. Root dry biomass was separated as structural roots (primary, nodal, and seminal roots) and lateral roots (York et al. 2013). Shoots and roots were dried at 60 °C until constant mass and weighted.

On-farm study location and data collection

Plant and soil were sampled in maize-producing farms in the western region of Santa Catarina State. The region has a humid subtropical climate (Cfa), according to Köeppen, with 18.1 °C mean annual temperature and 1959-mm annual rainfall, distributed throughout the year (Thomé et al. 1999). Meteorological stations in São Miguel do Oeste and São José do Cedro (Figure S1 and Table S1) provided temperature and rainfall data for the growing seasons.

Technicians from a local cooperative (OesteBio, São Miguel do Oeste) helped to select fifteen farms in six municipalities (Anchieta, Bandeirante, Barra Bonita, Palma Sola, Romelândia, and São Miguel do Oeste) (Figure S1). Five farms had landrace maize, five grew non-modified corn, and five had genetically modified (GM) maize.

Plant samples were taken in two consecutive years (February 2016 and January 2017) (Figure S1), at the reproductive stage (70–90 days after seeding (DAS)) (Tables S2 and S3). Soil samples for chemical analysis (about 500 g of soil) were taken at 0–20-cm depth in five random points from each maize field. At each point, an undisturbed sample (0–10-cm depth) was taken to assess apparent density, and we estimated mycorrhizal colonization in thin roots from each collected plant. The roots were placed in plastic bags, transported, and stored at 4 °C until processing. Two leaves (one above and one below the cob) from the corresponding maize plant were collected to assess phosphorus concentration.

Plant tissue analyses

Shoot and root samples were dried (60 ºC) and ground, submitted to sulfuric digestion (Tedesco et al. 1995), and P concentration was measured by colorimetry (Murphy and Riley 1962). P accumulated in plant tissues in the greenhouse experiment was calculated multiplying the concentration in shoots and roots by the respective dry mass.

AMF spore count and mycorrhizal colonization measurement

AMF spores were extracted from 50-cm3 soil samples by wet sieving, followed by centrifugation in a sucrose gradient (Gerdemann and Nicolson 1963) and counted with a stereomicroscope.

Fine-root samples (approximately 1 g) from each plant were washed in tap water, cleared in KOH 10% at 80 °C for 60 min, acidified in HCl 5.0%, and stained with trypan blue (Koske and Gemma 1989). Root colonization rates were quantified by slide intersection method (McGonigle et al. 1990) at 200× magnification. A total of 100 intersections per sample were examined, recording arbuscules, vesicles, and total colonization.

Determination of soil physical and chemical attributes

Soil pH (H2O), organic carbon (Walkley and Black 1934), and potassium and phosphorus (Mehlich-I) were quantified according to Tedesco et al. (1995) and Claessen (1997). P was also determined after extraction with anion-exchange resin (Tedesco et al. 1995). Exchangeable aluminum, calcium, and magnesium extracted with 1-mol L−1 KCl were determined by atomic spectrometry. Cation exchange capacity (CEC pH 7.0) was estimated as the sum of cations [Ca + Mg + (H + Al) + K]. Apparent density was determined by the cylinder method (Tedesco et al. 1995).

Statistical analyses

The Bartlett test analyzed data variance homogeneity. Since variances were homogeneous, we did analyses of variance (ANOVA), and when there were significant effects, we separated the means using the Skott–Knott test (p ≤ 0.05) for the greenhouse experiment and the confidence interval (p ≤ 0.05) for the field samples. Sigma-Plot v. 12.5 software (Systat Corp., San Jose, USA) generated the graphs with the mean standard error bars. Reference lines with the data of Betancur-Agudelo (2016) were added to the mycorrhizal colonization graphs.

The Vegan package (Oksanen et al. 2013) was used for the Redundancy Analysis (RDA) using the mycorrhizal variables (colonization by arbuscules, vesicles, and total) and soil physical–chemical attributes. PERMANOVA (p ≤ 0.05) was used to verify the significance of the model RDA and the effect of soil physical–chemical variables on mycorrhizal variables. Log (x + 1) transformation was applied for data standardization before the RDA was performed. The RDA graph variables were selected using the “vif” command of the Vegan package (Oksanen et al. 2013).

Results

Greenhouse experiment

In the greenhouse experiment, two maize conventional modern genotypes and four landraces grew taller than the other genotypes (Table 2). Genotypes Morgan 20A55, SCS 155, Taquara, Pixurum 07, Língua-de-papagaio, and Amarelão had the highest height. Morgan 20A55, SHS 5050, SCS 155, Taquara, Pixurum 07, Língua-de-papagaio, and Amarelão had the highest shoot dry biomass (SDM). SHS 5050, Morgan 20A55, SHS 5070, SCS 155, and Língua-de-papagaio had higher root dry biomass (RDM) than the other treatments. Lateral roots were the most important contributors to root biomass, and total dry mass (TDM) followed the same group separation pattern as SDM (Table 2).

Table 2 Plant height, shoot dry mass (SDM), root dry mass (RDM) sorted as structural (primary, nodal, and seminal roots), lateral, and total roots, and total dry mass (TDM) of ten maize genotypes grown in a greenhouse
Full size table

Phosphorus concentration in Morgan 20A55 genotype, which had the highest SDM, was lower than in all other genotypes (Table 3), but there were no differences among genotypes in total P accumulation. Mycorrhizal root colonization rates, as arbuscules, vesicles, and total colonization, did not differ among genotypes or maize genotype groups (Table 3), and there were no differences in AMF spore number.

Table 3 Phosphorus shoot concentration and accumulation, mycorrhizal colonization, and number of arbuscular mycorrhizal fungal (AMF) spores in ten maize genotypes grown in a greenhouse
Full size table

On-farm study

The farms with genetically modified maize (GM) were concentrated around São Miguel do Oeste meteorological station, while those with conventional modern genotypes maize (CO) and landrace (LR) were near the São José do Cedro meteorological station (Figure S1). Accumulated rainfall in the first year (2015/16) differed between the meteorological stations; the São José do Cedro station recorded 41% more rainfall than the station at São Miguel do Oeste. In the second year (2016/17), rainfall was similar at both meteorological stations (Figure S1 and Table S1).

In both years, root colonization rates were highly variable within the genotype groups (GM, CO, or LR) (Fig. 1). In the first year, the percentages of arbuscules varied between 6 and 34%, while it ranged between 13 and 63% in the second year. Vesicle rates varied between 1 and 12% in the first year, and between 2 and 40% in the second year. Total colonization had its highest value in the GM2 field (67%) and the lowest in the GM5 field (27%) in the first year; in the second year, the highest rate occurred in the CO5 (87%) and the lowest in the LR5 (43%), with no difference among genotype groups.

Fig. 1
figure 1

available at https://repositorio.ufsc.br/xmlui/handle/123456789/167888

Percentages of arbuscules, vesicles, total root colonization, in genetically modified (GM), conventional modern genotypes (CO), and landrace (LR) maize fields in two consecutive growing seasons: 2015/2016 (A, C, and E) and 2016/2017 (B, D, and F). Bars represent the standard error bar of the mean, n = 5. Horizontal lines represent means for maize fields (GM, CO, and LR) found by Betancur-Agudelo (2016), in the same study areas,

Full size image

Root colonization in genotype groups had distinct patterns in each year. In the first year, mycorrhizal colonization rates were arranged, in descending order: CO > GM > LR for arbuscules, CO > LR > GM for vesicles, and CO > LR > GM (Fig. 1) for total colonization. In the second year, the descending order was CO > GM > LR for arbuscules, vesicles, and total colonization.

Shoot P concentration did not differ among maize genotypes in the first year, with a mean value of 1.97 ± 0.69. In the second year, shoot P concentrations in GM mean was 0.97 ± 0.47, 37% lower than the LR and CO genotypes (1.53 ± 0.49).

Soil pH, clay, exchangeable Al, H + Al, and AD did not differ among genotype groups in both years (Table 3). Organic matter was lower in CO, as compared with LR and GM in the first year, while in the second year, there were no differences among genotype groups. Resin-extracted P was higher in LR, intermediate in GM and lower in CO maize in the first year, while in the second year, GM cornfields had higher levels of resin-extracted P than the other two treatments. Resin-P, levels, as well as K, were higher in the second year than in the first year. In both years, exchangeable Ca was higher in CO than in GM fields, while LR fields had intermediate values, not differing from the other groups. CO fields had higher exchangeable Mg than LR and GM in both seasons. Cation exchange capacity (CEC) showed differences only in the second season, when CO fields had higher levels than GM, and LR had intermediate values, that did not differ from the other genotype groups (Table 3).

The multivariate analysis showed that some patterns in mycorrhiza-related traits varied according to environmental conditions. The axes explain 26% of the total variation, with the RDA1 axis explaining 20.5% while the RDA2 axis explains 5.5% (Fig. 2). On the other hand, PERMANOVA, which tested the effect of environmental variables on mycorrhizae, showed no significance at a 5% probability of error.

Fig. 2
figure 2

Redundancy analysis (RDA) of mycorrhizal colonization rates and soil physical–chemical attributes in maize fields in two growing seasons. AD Apparent density; OM = Organic matter; Resin-P Resin-extracted phosphorus; pH pH in H2O; K exchangeable potassium; Mg exchangeable magnesium; Al exchangeable aluminum; H + Al Potential acidity; Coln Total colonization; Arbu Arbuscule colonization; Vesi Vesicle colonization. Legend: LR landrace; CO conventional modern genotypes maize; GM genetically modified maize

Full size image

When the mycorrhizal variables were compared between years, there were significant differences, in higher total colonization and arbuscule and vesicle frequencies in the second year, and those differences appear in the RDA graph. Analyses of differences in soil and climate variables between years show less rainfall, with a more uniform distribution in the second year, as compared with the first year, at both weather stations (Figure S2; Table S1). There was also P and K higher concentration f in the soil in the second year than in the first growth season (Table 3).

Discussion

The genotypes showed different growth strategies with low soil P. The first experiment showed that maize genotypes differ in shoot and root biomass ratio, and in tissue P concentration. Increased root production by some genotypes may be a strategy towards better use of soil resources, by increasing uptake of nutrients, such as phosphorus. The Morgan 20A55 genotype is an example, as it accumulated biomass with an 18% lower P concentration than the other maize genotypes, which indicates higher efficiency in the use of P (Table 4).

Table 4 Soil attributes of fields with genetically modified (GM), conventional modern genotypes (CO), and landrace (LR) maize in two consecutive years
Full size table

Mycorrhiza-related traits did not differ among genotypes nor genotype groups in the low-P soil used in the greenhouse experiment. Mycorrhizal colonization rates are more responsive to soil P levels than to plant genotypes (Kaeppler et al. 2000). Lehmann et al. (2012) performed a meta-analysis, encompassing 320 different crop plant genotypes, including maize. Although modern genotypes showed lower colonization, there was no evidence that modern genotypes have lost their ability to respond to arbuscular mycorrhizae. On the other hand, Cobb et al. (2016) evaluated two landraces and two modern sorghum hybrids and found modern hybrids less responsive to mycorrhization and more responsive to mineral fertilizer application. They argue that plant breeding seems to have selected plants that are less responsive to mycorrhizas. However, as they analyzed a limited number of genotypes, it is not possible to infer a generalized pattern from their work. A previous study in the same region in which we worked (Betancur-Agudelo 2016) found 50% higher arbuscular colonization and 37% higher total colonization in landrace than in GM maize, a pattern that did not occur in the present study. That suggests that changes in climate conditions affect mycorrhizal colonization to a significant extent.

In our on-farm assay, mycorrhizal root colonization in maize is within the range found in previous works, such as those by Miranda et al. (2005), with 84%, and Barboza (2016), with 69% colonization rates (ranging from 20 to 100%) in various soil P levels. Those results are possibly due to soil and plant management, including fertilizer and agrochemical applications, weeding, presence and diversity of spontaneous plants, and the use of tillage or no-tillage systems. Such procedures may be more important for AMF root colonization than plant genotype (Carrenho et al. 2010).

Some factors related to management may have affected the results of our study. In the previous year (Betancur-Agudelo 2016) and in the first year of our study, farmers growing landrace maize used primarily organic fertilizers (animal manure and green manure), while in the second growth season, they changed to chemical fertilizers (Table S2 and S3). The immediate availability of nutrients from mineral fertilizers affects the mycorrhizal symbiosis in a different way from organic fertilizers (Smith and Read 2008; Moura 2015), which release nutrients gradually as materials decompose (Busato et al. 2009).

Root colonization rates occurred in the second growing season were 22% higher than in the first year, when there were lower P and K and higher rainfall. Since abiotic conditions such as water availability affect the mycorrhizal association, the higher rainfall in the first growth season in the CO and LR fields than in the GM areas, may explain their higher vesicle production in the first year than in the second year (Fig. 1). Vesicle formation increases under stressful environmental conditions (Cooke et al. 1993; Smith and Read 2008), and differently from the first year, vesicle intensity rates in the second season were similar among genotype groups, as were rainfall values. There are other possible reasons for the differences, such as a change in the pattern of AMF species that predominate in different climate conditions, or AMF which produce more vesicles (Pereira 2013) or do not produce vesicles (Morton and Benny 1990). Some management procedures do not favor formation of large spores (Varela-Cervero et al. 2015), as observed with AMF in the Gigasporaceae family (Douds et al. 1993; Cuenca et al. 1998; Picone 2000).

The low tissue P concentration found in GM genotypes may result from a strategy to increase maize yield. As genotypes are selected for higher yields, soil P (Table 3) may not have been sufficient to supply the entire plant. Phosphorus is translocated from older leaves to plant organs with higher demand, such as grains (Raghothama 1999), which results in plants having lower foliar P while maintaining growth and yield.

None of the soil attributes correlated with mycorrhiza-related variables, although studies show that soil characteristics affect mycorrhiza establishment. According to Joner and Jakobsen (1995), higher organic matter levels increase soil porosity and facilitate hyphal growth, affecting fungal structures and root colonization. Vieira et al. (2018) showed that soil pH, MO, Al, Mg, and S differently affect the abundance of several AMF species in dry and rainy seasons. However, as stated by Stürmer and Siqueira (2008), it is difficult to establish a clear relationship of AMF occurrence with soil and climate variables.

Phosphorus is the best-known modulator of mycorrhizal symbiosis. Availability of this nutrient significantly affects mycorrhizal root colonization, as low P levels stimulate the symbiosis, while high levels tend to impair it (Smith and Read 2008). Studies with different P doses and inoculation with phosphate-solubilizing bacteria and AMF have shown that controlled reductions in P fertilization favor the association, and in general, plant yield is equivalent to plants receiving full P fertilizer application (Bressan and Vansconcellos 2002; Bressan et al. 2001; Suri et al. 2011; Pereira et al. 2014). Since symbioses are seldom considered in soil fertility management, many farmers add high P doses, thus reducing the potential for association of plants with beneficial microorganisms. A promising strategy would be establishing adequate fertilizer doses for each type of soil, aiming at high crop yield and better use of benefits promoted by soil microorganisms (Suri et al. 2006). Improved soil management, selection of plant species and variety, crop rotation, and adjusted fertilization (Bonfim et al. 2010) may promote AMF growth. That would avoid the negative influence on the AMF caused by application of high amounts of chemical fertilizers and other associated practices (Verbruggen et al. 2012; Roy et al. 2017). The use of highly soluble fertilizers and agrochemicals on crops leads to strong ecological and evolutionary selection in agroecosystems (Verbruggen and Toby-Kiers 2010). Plant genotypes, and specifically maize varieties and hybrids, need to be investigated in a broader way, aiming to better understand plant-microbial symbioses.

Conclusions

Maize genotypes—landrace and conventional modern genotypes—have diverse strategies for shoot and root growth at low soil P, and the Morgan 20A55 genotype can yield high plant dry matter under these conditions.

There are no differences in mycorrhizal root colonization and spore production among landrace maize and modern maize genotypes in low-P soils.

Soil and climate conditions have a stronger effect on mycorrhizal root colonization than maize genotype groups.