Litter Mass Loss of the Invasive Rhus typhina L. and Native Koelreuteria paniculata Laxm. Trees Alters Soil N-Fixing Bacterial Community Composition under Different N Forms

Abstract

Soil N-fixing bacterial (NFB) community may facilitate the successful establishment and invasion of exotic non-nitrogen (N) fixing plants. Invasive plants can negatively affect the NFB community by releasing N during litter decomposition, especially where N input from atmospheric N deposition is high. This study aimed to quantitatively compare the effects of the invasive Rhus typhina L. and native Koelreuteria paniculata Laxm. trees on the litter mass loss, soil physicochemical properties, soil enzyme activities, and the NFB. Following N supplementation at 5 g N m⁻² yr⁻¹ in four forms (including ammonium, nitrate, urea, and mixed N with an equal mixture of the three individual N forms), a litterbag-experiment was conducted indoors to simulate the litter decomposition of the two trees. After four months of decomposition, the litter cumulative mass losses of R. typhina under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 57.93%, 57.38%, 58.69%, 63.66%, and 57.57%, respectively. The litter cumulative mass losses of K. paniculata under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 54.98%, 57.99%, 48.14%, 49.02%, and 56.83%, respectively. The litter cumulative mass losses of equally mixed litter from both trees under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 42.95%, 42.29%, 50.42%, 46.18%, and 43.71%, respectively. There were antagonistic responses to the co-decomposition of the two trees. The litter mass loss of the two trees was mainly associated with the taxonomic richness of NFB. The form of N was not significantly associated with the litter mass loss in either species, the mixing effect intensity of the litter co-decomposition of the two species, and NFB alpha diversity. Litter mass loss of R. typhina was significantly higher than that of K. paniculata under urea. The litter mass loss of the two trees under the control and N in four forms mainly affected the relative abundance of numerous NFB taxa, rather than NFB alpha diversity.

1. Introduction

Invasive plants can establish successfully if they can acquire soil nitrogen (N) more efficiently than native plants [1,2,3,4]. The invasiveness and the invasion intensity of several invasive plants are also significantly associated with the level of soil N available [5,6,7,8]. The abundance and diversity of the nitrogenase reductase gene nifH (encoding the nitrogenase reductase subunit) are closely related to the level of soil N available [9,10,11,12]. Numerous invasive plants, such as Trifolium spp. [13], Ageratina adenophora [7], and Amaranthus retroflexus [14], Cenchrus spinifex [15], and multiple invasive plants (including Paspalum notatum, Sphagneticola trilobata, Alternanthera philoxeroides, and Hydrocotyle vulgaris) [8] can significantly increase the abundance and diversity of the nifH gene. Therefore, the soil N-fixing bacterial community (NFB) may be an important contributor to the successful invasion of invasive plants [6,7,16,17]. In addition, invasive plants can produce more litter and decompose more effectively and rapidly than native plants [18,19,20,21]. More importantly, invasive plants can also affect the NFB community via the decomposition process [7,22]. Thus, assessing the key mechanisms that govern the interactions between invasive plants-NFB via the decomposition process is crucial for elucidating the mechanisms that drive the successful invasion of invasive plants.

Atmospheric N deposition consists of several N components globally, including nitrate, ammonium, urea, and a mixture of several individual N forms. Additionally, the proportion of these different N components is dynamic and is expected to change as the frequency and intensity of human activities increase in the future [23,24,25,26]. In general, the positive effects of the mixture of several individual N forms on the ecological functions (e.g., the plant litter decomposition, soil enzyme activities, and the soil bacterial community metabolic activities) appear to be greater than those of the individual N forms [27,28]. More importantly, invasive plants may be more competitive than native plants under atmospheric N deposition via the altered NFB [14,29,30] or litter decomposition [28,31,32,33]. Consequently, it is essential to further elucidate the soil micro-ecological mechanisms underlying the invasion process of invasive plants from the perspective of the interactions between invasive plants and NFB, especially under different forms of atmospheric N deposition.

This study aimed to assess the effects of the invasive Rhus typhina L. and native Koelreuteria paniculata Laxm. trees on the litter mass loss, soil physicochemical properties, soil enzyme activities, and the NFB community in combination with N addition at 5 g N m⁻² yr⁻¹ in four forms (including ammonium, nitrate, urea, and an equal mixture of the three individual N forms) in southern Jiangsu, China. As two Sapindales trees, R. typhina and K. paniculata can coexist in the same habitat, the two trees have similar growing seasons (e.g., the growing season is usually from ~April to ~August in southern Jiangsu, China), growing environments (e.g., the parks, urban green spaces, and the areas near major roads in southern Jiangsu, China, etc.), and lifestyles (i.e., deciduous broadleaf trees). The two trees have similar stem heights and there is no significant difference in the space occupied by healthy mature individuals. In addition, both trees are commonly used in urban ornamentation in China. Rhus typhina originated from the Americas and was introduced to China in 1959 as an ornamental and landscape plant species [34,35]. However, R. typhina has caused biodiversity loss in China, especially in northern China, and has been classified as a destructive invasive tree [34,36]. The geographical range where R. typhina and K. paniculata are found in China is one of the most affected by atmospheric N deposition [24,37,38,39].

This study tested the following hypotheses: (1) Rhus typhina litter decomposes more easily than K. paniculata litter. (2) A synergistic effect might exist between the two trees’ co-decomposition. (3) Rhus typhina increases soil enzyme activities and NFB alpha diversity compared to K. paniculata. (4) N addition increases the litter mass loss of the two trees and NFB alpha diversity. (5) The magnitude of influence of N addition on the litter mass loss of R. typhina may be greater than that of K. paniculata. (6) The magnitude of influence of the addition of mixed N forms on the litter mass loss of the two trees and NFB alpha diversity may be greater compared to the addition of the three individual N forms, individually.

2. Materials and Methods

2.1. Experimental Design

Leaf litter from R. typhina and K. paniculata was randomly collected from an urban ecosystem of Zhenjiang (32.21 °N, 119.51 °E), southern Jiangsu, China. Figure 1 defines the geographical location of the sampled area in this study. Figure S1 defines the image of the environment in which the two trees grow. The annual mean temperature of Zhenjiang was ~17.1 °C, and the monthly mean temperature reached a maximum of ~28.1 °C in July and decreased to a minimum of ~3.7 °C in January [40]. The annual precipitation of Zhenjiang was ~1164.1 mm, and the monthly mean precipitation reached a maximum of ~432.1 mm in July and decreased to a minimum of ~2.7 mm in December [40]. The annual sunshine duration of Zhenjiang was ~1909.0 h, and the monthly mean sunshine duration reached a maximum of ~208.2 h in December and decreased to a minimum of ~125.9 h in August [40].

The litterbag experiment was conducted indoors from 15 April to 15 August 2021 (experimental period: ~4 months) to simulate the litter decomposition process. The air-dried leaf litter of the two trees was loaded into litterbags (10 × 15 cm; mesh size: 0.425 mm). Specifically, 6 g of R. typhina leaf litter, 6 g of K. paniculata leaf litter, or 6 g equally mixed R. typhina and K. paniculata leaf litter, were used per litterbag. The litterbags were then buried into the store-bought pasture soil (pH: ~6.3), at a depth of approximately 2 cm in planting pots (height: ~16.5 cm; top diameter: ~25 cm) with one litterbag per planting pot. The pasture yellow soil (pH value: ≈6.3; organic content: ≥30%; soil electrical conductivity: ≤3 ms/cm) was purchased from Zhong-Fang Agriculture & Livestock Co. Ltd., Taizhou, Jiangsu, China. The reason for using the pasture yellow soil as the culture medium was to minimize or even eliminate the possibility of the presence of invasive plants and the possible previous N deposition in the natural soil. The pasture yellow soil was not sterilized to ensure the natural occurrence of soil microorganisms. N was added to the litterbags in four forms: ammonium (ammonium chloride), nitrate (potassium nitrate), urea, and mixed N (an equal mixture of the three individual N forms). The ratio of the three individual N forms in mixed N was set at 1:1:1 to match the actual proportion of the different N forms deposited in the soil through the natural atmospheric N deposition in the region (viz., southern Jiangsu, China) [23,26,41,42]. All four forms of N were added at 5 g N m⁻² yr⁻¹, with sterile distilled water as the control (0 g N L⁻¹). The levels of the four N forms were similar to the actual concentration of N forms naturally deposited in the study area (viz., southern Jiangsu, China) [24,37,38,39].

The litterbag-experiment included 20 treatments: (1) Control (distilled water). (2) Ammonium, ammonium chloride. (3) Nitrate, potassium nitrate. (4) Urea, urea. (5) MixN, mixed N. (6) Rt, R. typhina litter under the control. (7) RtAmmonium, R. typhina litter under ammonium chloride. (8) RtNitrate, R. typhina litter under potassium nitrate. (9) RtUrea, R. typhina litter under urea. (10) RtMixN, R. typhina litter under mixed N. (11) Kp, K. paniculata litter under the control. (12) KpAmmonium, K. paniculata litter under ammonium chloride. (13) KpNitrate, K. paniculata litter under potassium nitrate. (14) KpUrea, K. paniculata litter under urea. (15) KpMixN, K. paniculata litter under mixed N. (16) RK, equally mixed litter from both trees under the control. (17) RKAmmonium, equally mixed litter from both trees under ammonium chloride. (18) RKNitrate, equally mixed litter from both trees under potassium nitrate. (19) RKUrea, equally mixed litter from both trees under urea. (20) RKMixN, equally mixed litter from both trees under mixed N. Each treatment was performed in three planting pots.

Litterbags were collected after ~120 d. Litter samples of the two trees in the litterbags were moderately cleaned and completely air-dried to a constant weight to evaluate the decomposition variables. Soil samples within 1 cm around the litterbags were also collected and passed through a 2 mm sieve and were used to estimate soil physicochemical properties, soil enzyme activities, and NFB.

2.2. Determination of the Decomposition Variables

The litter mass loss of the two trees was calculated as the ratio between the initial litter dry weight and the dry weight after time t to the initial litter dry weight [20,43,44].

The expected litter decomposition coefficient of the equally mixed litter from the two trees was calculated as follows [45,46]:

$$\mathrm{E}\mathrm{x}\mathrm{p}\mathrm{e}\mathrm{c}\mathrm{t}\mathrm{e}\mathrm{d} \mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \mathrm{c}\mathrm{o}\mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{t}=\frac{x+y}{2}$$

where x and y correspond to the observed litter decomposition coefficient of R. typhina and that of K. paniculata, respectively. The litter decomposition coefficient (k) of the two trees was determined as follows [47]:

$$X_t=X_o\times e^{-kt}$$

where X_o and X_t correspond to the initial litter dry weight and the litter dry weight after time t, respectively.

The mixing effect intensity of the litter co-decomposition of the two trees was calculated as follows [20,45,46]:

$$\mathrm{M}\mathrm{i}\mathrm{x}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{e}\mathrm{c}\mathrm{t} \mathrm{i}\mathrm{n}\mathrm{t}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y} \mathrm{o}\mathrm{f} \mathrm{t}\mathrm{h}\mathrm{e} \mathrm{l}\mathrm{i}\mathrm{t}\mathrm{t}\mathrm{e}\mathrm{r} \mathrm{c}\mathrm{o}-\mathrm{d}\mathrm{e}\mathrm{c}\mathrm{o}\mathrm{m}\mathrm{p}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}=\frac{O}{E}-1$$

where O and E correspond to the observed litter decomposition coefficient and the expected litter decomposition coefficient of equally mixed litter from both trees, respectively. A stronger response corresponds to a greater deviation from zero. In the presence of synergistic effects, the intensity is greater than zero, while the intensity is lower than zero in the presence of antagonistic effects.

2.3. Determination of Soil Physicochemical Properties and Soil Enzyme Activities

Soil pH and moisture were determined in situ using a digital soil acidity-moisture meter (ZD-06; ZD Instrument Co., Ltd., Taizhou, China) [28,48,49].

Soil enzyme activities, closely related to the soil nutrient cycles, were estimated: (1) urease (E.C. 3.5.1.5) activity was assessed using the sodium phenolate–sodium hypochlorite method with a colorimetric assay at 578 nm [50,51]; (2) protease (E.C. 3.4.11.4) activity was measured using the tyrosine method with a colorimetric assay at 700 nm [52]; (3) polyphenol oxidase (E.C. 1.10.3.1) activity was measured using the pyrocatechol method with a colorimetric assay at 410 nm [53]; (4) catalase (E.C. 1.11.1.6) activity was measured using the pyrogallol method with a colorimetric assay at 430 nm [51].

2.4. Determination of NFB

To estimate the composition and structure of the NFB community in this study, DNA sequences of the nifH gene were amplified. The nifH gene is highly conserved among NFB taxa and is widely used as a marker for phylogenetic analyses of NFB [12,54,55,56,57]. The primers for the amplification of the nifH gene are PolF and PolR [57]. The methods for subsequent sequencing data analysis can be found in earlier studies [14]. The sequence data of NFB did not submit to the NCBI database.

2.5. Statistical Analysis

Shapiro–Wilk’s test and Bartlett’s test were used to assess the deviations from normality and the homogeneity of the assessed variances, respectively. Differences in the values of the decomposition variables, soil physicochemical properties, soil enzyme activities, and NFB alpha diversity among different treatments were evaluated using a one-way analysis of variance (ANOVA; Tukey’s test). Path analysis was used to assess the contribution intensity of soil physicochemical properties, soil enzyme activities, and NFB alpha diversity to the litter mass loss of the two trees, according to the absolute values of the direct path coefficients. Statistical analyses were performed using IBM SPSS Statistics 26.0 (IBM, Inc., Armonk, NY, USA).

3. Results

3.1. Differences in the Decomposition Variables

After four months of decomposition, the litter cumulative mass losses of R. typhina under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 57.93%, 57.38%, 58.69%, 63.66%, and 57.57%, respectively. The litter cumulative mass losses of K. paniculata under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 54.98%, 57.99%, 48.14%, 49.02%, and 56.83%, respectively. The litter cumulative mass losses of equally mixed litter from both trees under the control, ammonium chloride, potassium nitrate, urea, and mixed N were 42.95%, 42.29%, 50.42%, 46.18%, and 43.71%, respectively (Figure 2). The litter mass loss of equally mixed litter from both trees was lower than that of R. typhina under the control (p < 0.05; Figure 2). The litter mass loss of equally mixed litter from both trees was lower than that of R. typhina and K. paniculata under ammonium chloride (p < 0.05; Figure 2). The litter mass loss of equally mixed litter from both trees and that of K. paniculata was lower than that of R. typhina under urea (p < 0.05; Figure 2). The litter mass loss of equally mixed litter from both trees was also lower than that of R. typhina under mixed N (p < 0.05; Figure 2). There was no significant difference in the litter mass loss of R. typhina and that of K. paniculata under the control (p > 0.05; Figure 2). The form of N did not significantly affect the litter mass loss of the two trees (p > 0.05; Figure 2).

The observed decomposition coefficient of equally mixed litter from both trees was lower than its expected decomposition coefficient under the control, ammonium chloride, and mixed N (p < 0.05; Figure 3a). The mixing effect intensity of the litter co-decomposition of the two trees was less than zero under all treatments (Figure 3b). The form of N did not significantly affect the mixing effect intensity of the litter co-decomposition of the two trees (p > 0.05; Figure 3b).

3.2. Differences in Soil Physicochemical Properties and Soil Enzyme Activities

Urea and mixed N decreased soil pH and polyphenol oxidase activity, but increased soil total N content compared to the control (p < 0.05;

Table 1. Soil physicochemical properties and soil enzyme activities. Data (means ± SE; n = 3) with different letters in a vertical column indicate significant differences (p < 0.05). Data (including soil moisture and soil catalase activity) without significant differences (p > 0.05) did not show in this table. Abbreviations have the same meanings as presented in Figure 2.

	Soil pH	Soil Total N Content (g/kg)	Soil Urease Activity (μg/g soil/d)	Soil Protease Activity (μg/g soil/d)	Soil Polyphenol Oxidase Activity (μmol/g soil/h)
Control	6.54 ± 0.03 a–c	8.31 ± 0.41 de	76.67 ± 5.48 b	486.78 ± 53.12 b–e	79.91 ± 6.96 ab
Ammonium	6.53 ± 0.07 a–c	8.94 ± 0.41 de	73.28 ± 2.64 b	838.24 ± 66.48 a	86.67 ± 7.22 a
Nitrate	6.47 ± 0.03 a–d	9.33 ± 0.44 de	98.08 ± 15.08 b	251.47 ± 13.20 e–h	50.21 ± 4.05 b–d
Urea	6.13 ± 0.09 de	15.92 ± 0.37 c	95.50 ± 25.05 b	501.94 ± 41.44 b–e	30.30 ± 2.79 d
MixN	6.10 ± 0.06 e	15.42 ± 0.50 c	78.59 ± 21.97 b	731.19 ± 76.38 ab	31.79 ± 6.59 d
Rt	6.67 ± 0.07 ab	9.20 ± 0.33 de	86.16 ± 8.89 b	653.43 ± 48.04 a–c	36.37 ± 1.85 cd
RtAmmonium	6.33 ± 0.07 b–e	9.22 ± 0.20 de	85.36 ± 14.53 b	133.31 ± 36.15 gh	27.74 ± 2.24 d
RtNitrate	6.50 ± 0.10 a–c	9.40 ± 0.41 de	79.08 ± 21.52 b	553.44 ± 71.11 a–d	32.86 ± 6.22 d
RtUrea	6.40 ± 0.00 a–e	9.38 ± 0.27 de	82.78 ± 12.34 b	89.882 ± 11.91 h	45.06 ± 17.21 d
RtMixN	6.533 ± 0.07 a–c	9.39 ± 0.11 de	79.56 ± 10.38 b	430.23 ± 54.23 c–f	44.74 ± 5.0 cd
Kp	6.67 ± 0.07 ab	10.06 ± 0.40 d	87.45 ± 8.18 b	112.10 ± 24.30 gh	65.15 ± 2.61 a–c
KpAmmonium	6.60 ± 0.00 a–c	8.33 ± 0.52 de	85.84 ± 8.14 b	359.54 ± 23.23 d–h	50.30 ± 1.86 b–d
KpNitrate	6.27 ± 0.07 c–e	9.31 ± 0.34 de	207.76 ± 19.63 a	456.49 ± 67.90 b–f	50.14 ± 6.75 b–d
KpUrea	6.43 ± 0.03 a–e	9.25 ± 0.48 de	86.32 ± 14.70 b	248.44 ± 42.42 e–h	45.62 ± 1.39 cd
KpMixN	6.53 ± 0.03 a–c	9.23 ± 0.04 de	47.67 ± 7.91 b	199.972 ± 71.46 f–h	47.74 ± 1.53 cd
RK	6.40 ± 0.06 a–e	7.82 ± 0.62 e	76.02 ± 10.19 b	258.54 ± 47.60 e–h	51.47 ± 6.31 b–d
RKAmmonium	6.43 ± 0.09 a–e	8.35 ± 0.44 de	94.86 ± 9.13 b	274.70 ± 46.49 d–h	46.10 ± 7.40 cd
RKNitrate	6.630.12 ab	21.73 ± 0.23 a	64.90 ± 14.08 b	276.72 ± 86.13 d–h	29.26 ± 2.016 d
RKUrea	6.70 ± 0.06 a	19.30 ± 0.53 b	85.04 ± 5.32 b	483.76 ± 55.60 b–f	26.82 ± 2.58 d
RKMixN	6.60 ± 0.10 a–c	9.78 ± 0.52 de	87.45 ± 7.86 b	391.85 ± 51.37 c–g	41.23 ± 2.93 cd

). Ammonium chloride increased soil protease activity compared to the control (p < 0.05; Table 1).

The equally mixed litter from both trees under potassium nitrate and urea increased soil total N content compared to the control (p < 0.05; Table 1). K. paniculata litter under potassium nitrate increased soil urease activity compared to the control (p < 0.05; Table 1). R. typhina litter under ammonium chloride and urea, and K. paniculata litter under the control decreased soil protease activity compared to the control (p < 0.05; Table 1). The litter of the two trees, whether mixed or not under N, regardless of the form of N, decreased soil polyphenol oxidase activity compared to the control (p < 0.05; Table 1).

Soil pH treated with equally mixed litter from both trees was higher than that treated with K. paniculata litter under potassium nitrate (p < 0.05; Table 1). Soil pH treated with K. paniculata litter under the control was greater than that treated with K. paniculata litter under potassium nitrate (p < 0.05; Table 1). Soil total N content treated with K. paniculata litter was higher than that treated with equally mixed litter from both trees under the control (p < 0.05; Table 1). Soil total N content treated with R. typhina litter and that treated with K. paniculata litter was lower than that treated with equally mixed litter from both trees under potassium nitrate and urea (p < 0.05; Table 1). Soil total N content treated with equally mixed litter from both trees following the addition of N in four forms decreased in the following order: potassium nitrate > urea > ammonium chloride > MixN and the control (p < 0.05; Table 1). Soil urease activity treated with R. typhina litter, and that treated with equally mixed litter from both trees, was lower than that treated with K. paniculata litter under potassium nitrate (p < 0.05; Table 1). Soil urease activity treated with K. paniculata litter under potassium nitrate was greater than that treated with other forms of N (p < 0.05; Table 1). Soil protease activity treated with R. typhina litter was greater than that treated with K. paniculata litter under the control and that treated with equally mixed litter from both trees under the control (p < 0.05; Table 1). Soil protease activity treated with R. typhina litter was lower than that treated with equally mixed litter from both trees under urea (p < 0.05; Table 1).

3.3. Differences in NFB Alpha Diversity

The form of N did not significantly affect NFB alpha diversity (p > 0.05;

Table 2. Alpha diversity of soil N-fixing bacterial communities. Data (means ± SE; n = 3) with different letters in a vertical column indicate significant differences (p < 0.05). Data (i.e., Shannon’s diversity index) without significant differences (p > 0.05) did not show in this table. Abbreviations have the same meanings as presented in Figure 2.

	OTU’s Species Index	Simpson’s Dominance Index	Chao1′s Richness Index	ACE’s Richness Index
Control	941.67 ± 25.56 a	0.81 ± 0.01 a–c	1597.13 ± 52.36 a	1591.61 ± 33.56 a–d
Ammonium	897.33 ± 19.62 ab	0.79 ± 0.00 a–c	1489.10 ± 3.04 ab	1534.02 ± 44.11 a–e
Nitrate	941.00 ± 35.57 a	0.83 ± 0.02 a–c	1588.74 ± 62.94 a	1619.08 ± 68.79 ab
Urea	904.33 ± 22.81 ab	0.85 ± 0.02 ab	1583.79 ± 57.64 a	1656.02 ± 38.10 a
MixN	823.67 ± 21.17 a–c	0.86 ± 0.02 a	1435.62 ± 107.20 ab	1487.79 ± 59.44 a–f
Rt	819.67 ± 0.67 a–c	0.81 ± 0.01 a–c	1385.54 ± 37.03 ab	1417.76 ± 32.54 b–f
RtAmmonium	921.00 ± 20.31 ab	0.79 ± 0.01 a–c	1560.81 ± 22.94 ab	1603.92 ± 17.34 a–c
RtNitrate	895.00 ± 9.45 ab	0.79 ± 0.01 a–c	1486.11 ± 25.69 ab	1530.61 ± 20.43 a–e
RtUrea	755.33 ± 31.57 c	0.81 ± 0.01 a–c	1325.73 ± 23.99 ab	1396.96 ± 35.28 c–f
RtMixN	861.67 ± 37.24 a–c	0.77 ± 0.00 a–c	1370.22 ± 24.60 ab	1444.76 ± 35.28 a–f
Kp	856.33 ± 19.70 a–c	0.75 ± 0.02 a–c	1383.04 ± 73.72 ab	1362.96 ± 54.94 ef
KpAmmonium	849.00 ± 15.72 a–c	0.74 ± 0.02 bc	1385.36 ± 21.76 ab	1375.59 ± 24.04 ef
KpNitrate	834.67 ± 56.16 a–c	0.73 ± 0.03 c	1368.83 ± 130.51 ab	1379.77 ± 72.14 d–f
KpUrea	876.67 ± 13.25 a–c	0.74 ± 0.02 a–c	1430.61 ± 25.78 ab	1428.08 ± 12.92 b–f
KpMixN	844.67 ± 21.94 a–c	0.79 ± 0.04 a–c	1288.93 ± 55.25 b	1341.60 ± 39.35 ef
RK	821.67 ± 12.17 a–c	0.76 ± 0.02 a–c	1312.43 ± 32.68 ab	1323.73 ± 26.17 ef
RKAmmonium	875.00 ± 20.11 a–c	0.78 ± 0.04 a–c	1443.83 ± 34.18 ab	1448.28 ± 29.61 a–f
RKNitrate	806.67 ± 8.41 bc	0.78 ± 0.03 a–c	1317.81 ± 14.34 ab	1306.66 ± 7.45 f
RKUrea	795.00 ± 28.43 bc	0.74 ± 0.03 bc	1412.85 ± 40.51 ab	1350.84 ± 38.09 ef
RKMixN	832.00 ± 15.63 a–c	0.73 ± 0.03 c	1383.86 ± 44.21 ab	1356.12 ± 25.61 ef

).

R. typhina litter under urea, and equally mixed litter from both trees under potassium nitrate and urea, decreased OTU’s species index of NFB compared to the control (p < 0.05; Table 2). K. paniculata litter under mixed N decreased Chao1′s richness index of NFB compared to the control (p < 0.05; Table 2). K. paniculata litter under the control, ammonium chloride, and mixed N, and equally mixed litter from both trees under the control, potassium nitrate, urea, and mixed N, decreased ACE’s richness index of NFB compared to the control (p < 0.05; Table 2).

The OTU’s species index of NFB treated with R. typhina litter under ammonium chloride and potassium nitrate was greater than that treated with R. typhina litter under urea (p < 0.05; Table 2). The ACE’s richness index of NFB treated with R. typhina litter under potassium nitrate was greater than that treated with equally mixed litter from both trees under potassium nitrate (p < 0.05; Table 2).

3.4. The Contribution Intensity of Soil Physicochemical Properties, Soil Enzyme Activities, and NFB Alpha Diversity to the Litter Mass Loss of the Two Trees

The absolute values of the direct path coefficient of Simpson’s dominance index (~0.6358), Chao1’s richness index (~0.5922), and ACE’s richness index (~0.8369) of NFB on the litter mass loss of the two trees were significantly higher than those of other variables (<0.400) (Figure 4).

3.5. Differences in the NFB Community Structure among Different Treatments

Based on the results of LEfSe analyses, Desulfovibrio and Methylomonas methanica were primarily changed for NFB treated with the control and mixed N, respectively (Figure 5a). Rhodocyclales and Thioploca were primarily changed for NFB treated with R. typhina litter under the control (Figure 5b). Thiohalocapsa and Chromatiaceae were primarily changed for NFB treated with K. paniculata litter under the control (Figure 5b). Xanthobacter_sp_91 was primarily changed for NFB treated with equally mixed litter from both trees under ammonium chloride (Figure 5c). Desulfotomaculum arcticum was primarily changed for NFB treated with R. typhina litter under potassium nitrate (Figure 5d). Desulfovibrio longus, Thioalbus denitrificans, and Methylomicrobium were primarily changed for NFB treated with K. paniculata litter under potassium nitrate (Figure 5d). Propionibacterium, Propionibacteriaceae, Propionibacteriales, Actinobacteria, Niveispirillum, and Geoalkalibacter were primarily changed for NFB treated with equally mixed litter from both trees under potassium nitrate (Figure 5d). Marinospirillum, Oceanospirillaceae, and Oceanospirillales were primarily changed for NFB treated with R. typhina litter under urea (Figure 5e). Sphingomonadaceae, Sphingomonadales, and Geopsychrobacter were primarily changed for NFB treated with K. paniculata litter under urea (Figure 5e). Geothermobacter, Methylomonas, Thioflexothrix, Thiotrichales, and unclassified Thiotrichales were primarily changed for NFB treated with equally mixed litter from both trees under urea (Figure 5e). Celerinatantimonas diazotrophica, Celerinatantimonas, Celerinatantimonadaceae, and unclassified Gammaproteobacteria were primarily changed for NFB treated with R. typhina litter under mixed N (Figure 5f). Azoarcus communis was primarily changed for NFB treated with equally mixed litter from both trees under mixed N (Figure 5f).

4. Discussion

The litter mass loss of R. typhina was similar to that of K. paniculata (Figure 2). Thus, contrary to the first hypothesis, R. typhina litter did not degrade more easily than K. paniculata litter. These observations are also inconsistent with the results of previous research which has shown that invasive plants either degrade more rapidly [18,19,20,21], or significantly more slowly than native plants [58,59,60]. This phenomenon may be due to the similar proportions of soluble and recalcitrant components in the litter of the two trees [28]. In other words, the two trees had similar litter quality, probably because they coexist in the same habitat and have similar growing seasons, growing environments, and lifestyles. This phenomenon may also be attributed to the relatively short time frame of this study, in which the rapid decomposition of R. typhina could not be manifested.

In nature, the two trees typically co-exist [21,42]. As a result, the litter decomposition may be altered when the litter from the two trees is mixed. Non-additive responses in the litter co-decomposition of invasive plants and native plants are often observed, and generally, the litter decomposition of invasive plants can accelerate that of native plants [21,49,61,62]. In this study, the decomposition of the mixed litter from both trees was slower than that of the individual litter of either species (Figure 2). In addition, the observed decomposition rate for equally mixed litter from both trees was also significantly lower than its expected decomposition rate (Figure 3a). More importantly, the mixing effect intensity of the litter co-decomposition of the two trees was less than zero (Figure 3b). Thus, contrary to the second hypothesis, antagonistic responses were observed in the litter co-decomposition of the two trees. Accordingly, there may be an interspecific interference during the litter co-decomposition process for the co-decomposition of the two trees. In addition, some recalcitrant components (difficult to decompose) may be formed during the litter co-decomposition process for the co-decomposition of the two trees. Studies on the non-additive responses for the co-decomposition of two plant species have shown that only 30% additive, 50% synergistic, and 50% antagonistic responses were observed [63]. The reason for these differences may be closely related to the differences in the relatedness of plant species, the initial quality and quantity of plant litter, the type of culture medium, the initial soil physicochemical properties, the initial (decomposer) microbial communities in culture medium, the ratio of the two plant litters mixed, and the duration of the experiment.

Invasive plants can mediate changes in soil enzyme activity [64,65,66,67] by releasing nutrients, such as carbon- and N-containing substances, during the litter co-decomposition process. Surprisingly, R. typhina litter significantly increased soil protease activity under the control, but significantly decreased soil urease activity under potassium nitrate compared to that of K. paniculata in this study (Table 1). Hence, R. typhina can increase the level of urea hydrolysis but decrease the level of protein hydrolysis. Previous studies also showed that invasive plants can increase [65,66,68,69] or decrease soil enzyme activities [48,70,71,72], or have no significant on soil enzyme activities [67,73,74,75]. Thus, the effects of invasive plants on soil enzyme activities may be species-dependent [28,76] and N-form-dependent [48,77], mainly due to the differences in soil physicochemical properties and the level of available soil nutrients under different plant species and/or different forms of N.

ACE’s richness of NFB under K. paniculata litter and equally mixed litter from both trees was lower than that under the control (Table 2). Thus, the litter decomposition of K. paniculata and that of equally mixed litter from both trees under the control may be slower, mainly due to the reduced richness of NFB. Previous studies have shown that the abundance and diversity of the nifH gene are strongly associated with the soil available N levels [9,10,11,12]. At the same time, the contribution intensity of the richness of NFB to the litter mass loss was obviously greater than other variables based on the results of path analysis (Figure 4). However, contrary to the third hypothesis, the litter decomposition of R. typhina under the control did not significantly affect NFB alpha diversity (Table 2). Interestingly, the litter mass loss of the two trees under the control appeared to result in significant variations in the relative abundance of various NFB taxa, i.e., Rhodocyclales and Thioploca were abundant during the litter decomposition of R. typhina under the control, and Thiohalocapsa and Chromatiaceae were abundant during the litter decomposition of K. paniculata under the control (Figure 5). Thus, the litter decomposition of the two trees under the control results in the presence of numerous dominant biomarkers of NFB. The main reason may be that there is still some difference in the litter quality between the two trees, leading to the species differentiation of NFB. Accordingly, the litter decomposition of the two trees under the control mainly affected the composition of NFB, rather than their alpha diversity. Earlier studies have also verified that plant species, particularly the invasive plants, mainly affected the composition of soil microbial communities rather than NFB alpha diversity [78,79].

Generally, N addition can trigger soil acidification mainly due to the release and accumulation of free H⁺ via nitrification [14,80,81]. The same results were observed in this study (Table 1). In addition, ammonium chloride increased soil protease activity, but urea and mixed N decreased soil polyphenol oxidase activity (Table 1). Thus, the impacts of N addition on soil enzyme activities could vary and depend on the form of N. However, the form of N did not significantly affect the litter mass loss of the two trees and NFB alpha diversity (Figure 2 and Table 2). Thus, the fourth hypothesis could not be supported based on this finding. Nonetheless, mixed N triggered a significant variation in the abundance of Methylomonas methanica (Figure 5). Thus, Methylomonas methanica may be used as a dominant biomarker under mixed N.

In addition, contrary to the fifth and sixth hypotheses, the form of N did not significantly affect the litter mass loss of the two trees (Figure 2), the mixing effect intensity of the litter co-decomposition of the two trees (Figure 3b), or NFB alpha diversity (Table 2). Thus, the degree of influence of N addition on the litter mass loss of R. typhina and NFB alpha diversity was similar to that of K. paniculata. However, the litter mass loss of R. typhina was significantly higher than that of K. paniculata under urea (Figure 2). Thus, R. typhina litter may decompose more effectively and rapidly than K. paniculate litter under urea. This finding is consistent with the previous studies, i.e., invasive plants decompose more easily and rapidly than native plants [18,19,20,21]. This phenomenon may be due to the higher percentage of soluble components and lower percentage of recalcitrant components in R. typhina litter compared to those of K. paniculate litter, and/or the compounds contained in R. typhina litter may be more readily released into the soil in the presence of urea. This may also be due to the fact that the limitation of the level of N utilization for the metabolic activity of soil bacterial community in relation to the decomposition process is alleviated to a greater extent in R. typhina litter, compared to K. paniculate litter under urea [14,82]. In this study, soil total N content under urea was significantly higher than that under ammonium chloride and potassium nitrate (Table 1). Previous studies also showed that soil bacterial community is more likely to utilize organic N than inorganic N [14,83]. Thus, the nutrient cycling rate during the decomposition process of R. typhina may be obviously higher than that of K. paniculate under urea. Consequently, an increase in the relative proportion of urea in the atmospheric N deposition may be beneficial to the invasion of R. typhina via the increased nutrient cycling rate mediated by the accelerated litter mass loss.

The LEfSe analysis revealed that the litter decomposition of the two trees following the addition of N in four forms resulted in significant variations in the relative abundance of various NFB taxa, i.e., Xanthobacter_sp_91 under the litter decomposition of equally mixed litter from both trees treated with: ammonium chloride; Desulfotomaculum arcticum under the litter decomposition of R. typhina treated with potassium nitrate; Desulfovibrio longus, Thioalbus denitrificans, and Methylomicrobium under the litter decomposition of K. paniculata treated with potassium nitrate; Propionibacterium, Propionibacteriaceae, Propionibacteriales, Actinobacteria, Niveispirillum, and Geoalkalibacter under the litter decomposition of equally mixed litter from both trees treated with potassium nitrate; Marinospirillum, Oceanospirillaceae, and Oceanospirillales under the litter decomposition of R. typhina treated with urea; Sphingomonadaceae, Sphingomonadales, and Geopsychrobacter under the litter decomposition of K. paniculata treated with urea; Geothermobacter, Methylomonas, Thioflexothrix, Thiotrichales, and unclassified Thiotrichales under the litter decomposition of equally mixed litter from both trees treated with urea; Celerinatantimonas diazotrophica, Celerinatantimonas, Celerinatantimonadaceae, and unclassified Gammaproteobacteria under the litter decomposition of R. typhina treated with mixed N; Azoarcus communis under the litter decomposition of equally mixed litter from both trees treated with mixed N (Figure 5). Thus, the litter decomposition of the two trees following the addition of N in four forms causes a substantial effect on certain NFB taxa. The differential shifts in NFB composition in response to the litter decomposition of the two trees following the addition of N in four forms could be primarily due to the differences in the level of nitrogenophilic ability of those NFB taxa. Thus, the addition of N in different forms can exert different intensities of selective pressure on different NFB taxa, leading to increases (e.g., the nitrogenophilic NFB taxa) or decreases (e.g., the NFB taxa that are poorly tolerant to N addition) in the proportion of specific NFB taxa.

5. Conclusions

This study is the first attempt to elucidate the effects of the invasive R. typhina and native K. paniculata trees on the litter mass loss, soil physicochemical properties, soil enzyme activities, and the NFB community under different forms of N deposition. The main conclusions are that: (1) There were antagonistic responses to the litter co-decomposition of the two trees based on the values of the mixing effect intensity of the litter co-decomposition of the two trees; (2) the litter mass loss of the two trees was mainly affected by the richness of NFB based on the results of path analysis; (3) the form of N did not significantly affect the litter mass loss of the two trees, the mixing effect intensity of the litter co-decomposition of the two trees, or NFB alpha diversity based on the results of one-way ANOVA; (4) the litter mass loss of R. typhina was significantly higher than that of K. paniculata under urea based on the results of one-way ANOVA; (5) The litter decomposition of the two trees under the control, and following the addition of N in four forms, mainly affected the composition of NFB, i.e., resulted in significant variations in the relative abundance of various NFB taxa, rather than NFB alpha diversity, based on the results of LEfSe analyses.