Impact of nitrogen application and crop stage on epiphytic microbial communities on silage maize leaf surfaces

Introduction

Maize (Zea mays) silage is a widely used form of roughage due to its high yield, nutritional value, and digestibility (Zhao et al., 2022). In countries like United States and China, millions of hectares are dedicated to silage maize production to meet the demand for animal feed (Bernard & Tao, 2020). Additionally, more than 1.67 million hectares of silage maize were planted in China in 2020 (Zhao et al., 2022). To ensure an adequate supply of feed in different seasons and climatic conditions, fresh maize is often made into silage, which effectively preserves the nutritional value of the maize and improves its palatability to ruminants (Benchaar et al., 2014). Various factors, including water-soluble carbohydrate concentration, buffering energy, and the presence of epiphytic plant microbes, directly influence the fermentation process of silage. Field management practices (e.g., agronomic practices and harvest timing) have been studied with the aim of improving the quality of silage fermentation (Li et al., 2016; Xu et al., 2021). During ensiling, microorganisms (e.g., lactic acid bacteria (LAB), butyrate bacteria, and spoilage bacteria) play a crucial role in the silage fermentation quality. However, their distribution on grass surfaces is often limited due to structural (Vorholt, 2012) and nutritional (Tang et al., 2022) factors.

Common LAB found on grass surface include Enterococcus, Lactococcus, Lactobacillus, Pediococcus, Leuconostoc, and Weissella (Yu, Leveau & Marco, 2020). The population sizes and clustering of LAB are often correlated with morphological characteristics of the leaves (Tang et al., 2022). For instance, glaucous leaves, such as those of rye (Secale cereale), tend to have lower microbial abundance, including lower abundance of aerobic bacteria (AB), LAB, yeasts, and molds, compared to non-glaucous leaves (Tang et al., 2023). Additionally, the composition of microbial communities on different crop surfaces can vary. Silage maize surfaces are mainly populated by AB species, such as Paenibacillus, Flavobacteriaceae, Sphingomonas, Exiguobacterium, Rhizobiaceae, Acinetobacter, and Buchnera (Kadivar & Stapleton, 2003). Wheat (Triticum aestivum) surfaces, on the other hand, are dominated by Pantoea, Weissella, Pseudomonas, Exiguobacterium, and Paenibacillus (Keshri et al., 2019). Moreover, the geographical location can also influence the microbial communities on forage leaf surfaces. For example, in southwest China, silage maize surfaces are dominated by Weissella, Pseudomonas, Lactobacillus, and Leuconostoc (Guan et al., 2018), while in northeast China, Leuconostoc, Weissella, Sphingobacterium, and Stenotrophomonas (Wang et al., 2021) are more prevalent. However, the impact of field management practices, including fertilization and irrigation, on the abundance and species composition of microorganisms on leaf surfaces remains uncertain.

Optimal nitrogen (N) fertilizer application is known to positively affect grass yield. N is very limited in soil, and agricultural systems often receive N inputs from anthropogenic sources (Canfield, Glazer & Falkowski, 2010), even when excessive amounts of N are not required. Recent research has explored the effects of N fertilization on microbial communities of soil (Zhang et al., 2022), but the potential effects of N fertilizer on number and species of microbes on grass surface are not well understood. Therefore, this study conducted a field experiment that aimed to compare the effects of different N application rates and maturity stage on the species, abundance of microbes, and physiochemical properties of silage maize leaf surfaces and on the physiochemical properties of the leaf surfaces. We hypothesized that (a) high N application rates would increase microbial diversity and abundance on silage maize leaf surfaces, (b) high N application rates would enhance nutrient levels on the leaf surfaces of silage maize, and (c) microbial diversity and abundance would be greater during the flowering stage compared to dough stage of silage maize.

Materials and Methods

Results

Microbial numbers

Neither the N application rate alone nor its interaction with maturity stage had any effect on the numbers of AB, LAB, yeasts, or molds (P > 0.05, Table 1). LAB, yeasts, and molds were less abundant on the maize leaf surfaces. At the flowering stage, the populations of AB, LAB, yeasts, and molds populations increased by 0.37, 2.17, 0.78, and 1.52 lg cfu g⁻¹ FM, respectively, compared to those at the dough stage (P < 0.001). However, the coefficient of determination (R²) for the prediction of population sizes from N application rates was relatively small (Fig. 1).

Table 1:

Population sizes of microorganisms on the leaf surfaces for different treatments and maturity stage.

Maturity stage and treatment		Aerobic bacteria (lg cfu g−1 FM)	Lactic acid bacteria (lg cfu g−1 FM)	Molds (lg cfu g−1 FM)	Yeasts (lg cfu g−1 FM)
N application rates (N)	N160	4.14 ± 0.13	2.72 ± 0.34	2.71 ± 0.16	3.14 ± 0.36
	N240	4.29 ± 0.07	2.34 ± 0.63	2.65 ± 0.19	3.06 ± 0.33
	N320	4.31 ± 0.08	2.42 ± 0.63	2.48 ± 0.25	3.23 ± 037
Maturity stage (MS)	Flowering stage	4.43 ± 0.02a	3.58 ± 0.07a	3.00 ± 0.09a	3.90 ± 0.06a
	Dough stage	4.06 ± 0.06b	1.41 ± 0.29b	2.22 ± 0.09b	2.38 ± 0.10b
P	N	0.398	0.877	0.710	0.948
	MS	0.000	0.000	0.000	0.000
	N * MS	0.118	0.183	0.526	0.619

DOI: 10.7717/peerj.16386/table-1

Note:

Different lowercase letters in the same column represent significant difference between N application rates or maturity stage (P < 0.05).

Leaf physical properties

N application rates had significant effects on the transpiration rate, pore conductivity of water vapor, total conductance of water vapor, and total conductivity of carbon dioxide (P < 0.05). Specifically, there were no differences in these factors between the N160 and N240 treatments, but all of them were substantially lower than those in the N320 treatment (Table 3). Additionally, the moisture retention capacity at the flowering stage was higher than that at the dough stage (P < 0.01) (Fig. 2). A similar trend was observed for the stomatal density of maize (Fig. 3A). Moreover, the stomatal densities on the abaxial leaf surfaces in maize were higher than those on the adaxial surfaces at both the flowering and the dough stages (P < 0.001) (Fig. 3B). During the maturation process from flowering to dough stage, there was a gradual loss of water in the corn leaves. The water content of the leaves changed significantly 8 h ago. However, neither the N application rate nor the maturity stage had a significant effect on the rate of water loss (Fig. 2).

Table 3:

Leaf physical properties for different treatments at the flowering stage.

Treatment	Transpiration rate (mmol m−2s−1)	Net photosynthetic rate (µmol m−2s−1)	Intercellular carbon dioxide concentration (µmol mol−1)	Intercellular carbon dioxide partial pressure (Pa)	Pore conductivity of water vapor (mmol m−2s−1)	Total conductance of water vapor (mmol m−2s−1)	Total conductivity of carbon dioxide (mmol m−2s−1)
N160	4.42 ± 0.62b	29.2 ± 0.90	77.8 ± 6.66	9.03 ± 1.05	146 ± 9.30b	133 ± 13.7b	113 ± 18.7b
N240	4.97 ± 0.24b	31.6 ± 1.90	123 ± 20.3	10.6 ± 2.08	186 ± 25.6b	179 ± 23.8b	146 ± 19.9b
N320	7.25 ± 0.53a	38.2 ± 4.49	172 ± 37.0	14.0 ± 3.00	385 ± 19.4a	363 ± 12.8a	220 ± 14.4a
P	0.014	0.148	0.090	0.335	0.000	0.000	0.014

DOI: 10.7717/peerj.16386/table-3

Note:

Different lowercase letters in the same column represent significant difference among N application rates (P < 0.05).

Epiphytic population relationships with physicochemical traits of leaf

The abundance of AB, LAB, yeasts, and molds on the leaf surface of maize was all positively correlated with the concentrations of inorganic phosphorus, soluble sugar, free amino acids, soluble protein, and total phenolics (P < 0.01) (Table 4). However, the water loss rate of leaves showed a negative correlation with the abundance of these microbial species (P < 0.05), while the moisture retention capacity and stomatal density exhibited a positive correlation (P < 0.01). There was no correlation between microbial abundance and transpiration and net photosynthetic rates (P > 0.05).

Table 4:

Correlation coefficients between chemical and physical properties and population sizes of epiphytes (n = 18).

Items		Aerobic bacteria	Lactic acid bacteria	Molds	Yeasts
Chemical properties	Inorganic phosphorus	0.672**	0.732**	0.688**	0.752**
	Soluble sugar	0.706**	0.865**	0.815**	0.893**
	Soluble protein	0.778**	0.782**	0.799**	0.844**
	Free amino acids	0.799**	0.803**	0.807**	0.896**
	Total phenolic	0.742**	0.854**	0.835**	0.840**
	pH	0.133	0.037	−0.217	0.084
Physical properties	Water loss	−0.500*	−0.667**	−0.497*	−0.617**
	Moisture retention capacity	0.695**	0.708**	0.569*	0.825**
	Stomatal density	0.770**	0.684**	0.663**	0.738**
	Transpiration rate	0.349	0.503	0.125	0.566
	Net photosynthetic rate	0.019	0.525	0.253	0.319
	Intercellular carbon dioxide concentration	0.541	0.315	0.067	0.565
	Intercellular carbon dioxide partial pressure	0.541	0.315	0.067	0.565
	Pore conductivity of water vapor	0.393	0.526	0.018	0.633
	Total conductance of water vapor	0.392	0.523	0.023	0.627
	Total conductivity of carbon dioxide	0.392	0.523	0.023	0.628

DOI: 10.7717/peerj.16386/table-4

Note:

Coefficients of non-parameter Spearman’s correlation. Asterisks indicate significant differences at P < 0.05 (*) and P < 0.01 (**), respectively.

Indicators of three N application rates and two maturity stages were plotted on a scatter diagram and analyzed using PCA (Fig. 4), where the first and second principal components explained 81.7% of the total variance in both N application rates and maturity stages. The fertility period has a significant impact on the confidence level of various indicators.

Microbial community composition on leaves

The results of the principal coordinate analysis (PCoA) analysis showed that the different N application rates were closely grouped together, indicating high overlap of OTUs under different N application rates (Fig. 5). The leaf samples obtained from N160, N240, and N320 treatments had a total of 114, 135, and 93 OTUs, respectively. Among these, 87 OTUs were shared between N160 and N240 samples, 75 OTUs were shared between N240 and N320 samples, and 71 OTUs were shared between N160 and N320 samples. The community abundance of Alcaligenaceae, Erwiniaceae, and Pseudomonadaceae families had higher abundance compared to other species (Fig. 6). At the genus level, the abundance percentages of Bordetella, Pantoea, and Pseudomonas were higher than shose of other species (Fig. 7). Notably, the percentage of Pantoea community abundance was significantly lower in the N320 treatment than those of the N160 and N240 treatments. The pH value, free amino acid concentration, and transpiration rate of leaves had a substantial effect on the abundance of microorganisms (Fig. 8). Specifically, the pH value showed a positive correlation with the abundances of Bordetella and Achromobacter abundances (P < 0.05) (Fig. 9).

Discussion

Previous studies have indicated variations in both species and abundance of epiphytic microorganisms (Kasozi, Kaiser & Wilhelmi, 2022). The physiochemical properties of leaves can influence the abundance of epiphyte (Tang et al., 2022). In this study, the N application rates did not have a significant effect on the population sizes of microbial species on the surface of maize leaves (Table 1). This finding contradicts the results of a previous study (Li et al., 2016). The discrepancy might be attributed to the specific crop type or the N application rates. For instance, in a study involving Italian ryegrass, as the N application rate increased from 0 kg ha⁻¹ to 175 kg ha⁻¹, the LAB population initially increased and then decreased, but there were no significant changes in the populations of AB, yeasts, and molds (Chen, Dong & Zhang, 2021). Similarly, when studying wheat, increasing the N application rate from 0 kg ha⁻¹ to 300 kg ha⁻¹ resulted in a gradual decrease in the population size of AB, an initial increase followed by a decrease in LAB, an increase in yeasts, and a fluctuation in molds (Li et al., 2016). Generally, leaf surfaces exhibit lower nutrient levels, and the concentration of these nutrients is closely related to the concentration within the leaf’s internal tissue (Tang et al., 2023). These nutrients, exuded from the leaf interior to the surface, play a crucial role in the colonization of microbial communities on the leaf surface. Tian et al. (2020) found that high N concentrations in leaves were associated with lower diversity of phyllosphere bacterial communities. Notably, the impact of N on microbial diversity is indirect. When plants are stimulated by N fertilization, changes in transmembrane nutrient transport of nutrients on the leaf surface, reservoir/source relationships, and water transport are observed (Darlison et al., 2019). These factors often determine the number and species of microorganisms present on the leaf surface.

Protein metabolism is generally influenced by the rate of N supply, as increased N availability leads to elevated hydrolase activity and accumulation of soluble proteins in plants (Liu et al., 2021). Therefore, the N application rate is closely associated with soluble protein concentration. Interestingly, the present study found no significant effect of N application rates on soluble protein concentration, possibly because silage maize was used, resulting in a low total N per leaf (13 leaves plant⁻¹). Moreover, the relatively small differences in N fertilizer application rates might have contributed to these results. Apart from water, nutrient stress is a crucial factor influencing microbial survival and reproduction on the leaf surface. As the N supply rate increases, plants accumulate more soluble sugars and inorganic phosphorus (Liu, Song & Liu, 2018), regulating their osmotic and water-holding capacities. However, these nutrients must pass through stomata to be excreted onto the plant surface. In this study, the concentration of chemical substances on maize leaf surfaces showed positive correlations with microbial population sizes, consistent with the findings of Tang et al. (2022). Additionally, Mercier & Lindow (2000) concluded that the abundance of carbon sources, N sources, and inorganic molecules required by certain microorganisms on plant surfaces were closely related to microbial population sizes. There is evidence that yeast reduces the abundance of organic N on tall fescue (Festuca arundinacea Schreb.) leaves (Nix-Stohr, Burpee & Buck, 2008). Bacteria have also been shown to utilize sugars on leaf surfaces to increase their population size (Chen, Dong & Zhang, 2021). In the current study, Microbacterium and Rosenbergiella exhibited sensitivity to sugar (Fig. 9). Furthermore, the study demonstrated that only a very small population of LAB was attached to the leaf surface, which is not surprising considering that LAB are facultative anaerobic microorganisms and are therefore less abundant on leaf surfaces. The pH on leaf surfaces showed a positive relationship with the abundances of Bordetella and Achromobacter, likely because Bordetella and Achromobacter are basophilic.

Surprisingly, the present study found a positive relationship between total phenolic concentration and microbial population size, contrary to the results of Yadav, Karamanoli & Vokou (2005). This discrepancy may be explained by certain microorganisms utilizing small molecule compounds. Metabolic activity in the dough stage was weaker compared to the flowering stage, resulting in fewer nutrients both inside and on the surface of leaves. Reduced nutrient availability is not conducive to microbial colonization, explaining the simultaneous occurrence of smaller microbial populations and fewer nutrients on leaf surfaces during the dough stage.

High water content in leaves helps alleviates drastic temperature changes and keeps the leaf surfaces moist, promoting microbial reproduction. Furthermore, increasing environmental humidity raises the water activity on the plant surface, enhancing water and nutrient utilization by microorganisms (Zi et al., 2021). Improved leaf wetting properties facilitate the exudation of nutrients from the interior of leaves to the leaf surface (Knoll & Schreiber, 2000), providing more nutrients for surface microbial colonization. However, regardless of the abundance of leaf nutrients, they must pass through stomata to reach the leaf surface. Higher stomatal density increases leaf epidermal conductivity, facilitating microbial population growth on the leaf surface. In our study, the transpiration rate showed a positive correlation with the abundance of Pseudomonas, Ochrobactrum, Delftia, and Achromobacter in the current study (P < 0.05), possibly because increased stomatal density promotes transpiration and enhances water exchange between plants and the air. Maintaining a consistently higher transpiration rate improves leaf wetting and promotes microbial colonization. This could explain the negative correlation observed in this study between the water loss rate from leaves and the population sizes of the four species, while moisture retention capacity and stomatal density exhibited positive correlations with these microbes (Table 4).

Currently, next -generation sequencing approaches are widely used for analyzing microbial species, effectively revealing the relationships between field management techniques and microorganisms (Chen et al., 2018). According to the PCoA results, the different sample points were closely clustered together (Fig. 5), indicating similar species composition at different N application rates. The abundances of bacteria did not significant increase with higher N application rates in this study (Figs. 6 and 7), suggesting similar species richness of maize leaves under different N application rates. This similarity is closely related to the nutrient deficiency on the surface of maize leaves (Table 2). The larger size and lower density of stomata further explain these findings. Generally, the length of abaxial stomata tends to increase while stomatal density tends to decrease with higher N application rates (Cai et al., 2017). Lower stomatal density hinders the entry of CO₂ into the leaves and reduces plant water vapor conductance (Lawson & Blatt, 2014), resulting in inhibited water exudation from plants. As plants age, the stomatal density on leaves decreases, leading to reduced transpiration and a tendency of decreased water availability on the leaf surface. These factors contribute to the reduction in microbial species and populations. Bordetella (Sukweenadhi et al., 2022), Pantoea, and Pseudomonas (Sun et al., 2021) are common microorganisms found on maize, with the soil being a primary source. The present study found no evidence linking the abundance of these microorganisms to the N application rate. This finding explains the observed similarity in evolutionary relationships among microbial species under different N application rates in our study (Fig. 10). However, these do not hinder producers from using nitrogen fertilizer to regulate the number of lactic acid bacteria on the surface of forage, thereby improving fermentation quality, as previous studies have confirmed the feasibility of this operation (Li et al., 2016).

Conclusion

In conclusion, this study highlights the significant impact of maturity stage on microbial populations, physiological characteristics, and chemical compositions on maize leaf surfaces. The findings indicate that the abundance of nutrients and epiphyte populations is higher during the flowering stage compared to the dough stage. Surprisingly, the N application rate did not have a substantial effect on microbial species or population size. Based on the results, we recommend applying approximately 160 kg ha⁻¹ of N fertilizer for optimal results in silage maize production. Additionally, we advise harvesting at the flowering stage is advised as it not only improves N fertilizer efficiency but also ensures favorable nutrient levels. These findings provide valuable insights for the management of harvest times and N fertilizer application in silage maize cultivation.

Supplemental Information