The role of rhizosphere pH in regulating the rhizosphere priming effect and implications for the availability of soil-derived nitrogen to plants

Abstract

Background and Aims

A comprehensive understanding of the rhizosphere priming effect (RPE) on the decomposition of soil organic carbon (SOC) requires an integration of many factors. It is unclear how N form-induced change in soil pH affects the RPE and SOC sequestration.

Methods

This study compared the change in the RPE under supply of NO₃-N and NH₄-N. The effect of the RPE on the mineralization of soil N and hence its availability to plant and microbes was also examined using a ¹⁵N-labelled N source.

Key Results

The supply of NH₄-N decreased rhizosphere pH by 0.16–0.38 units, and resulted in a decreased or negative RPE. In contrast, NO₃-N nutrition increased rhizosphere pH by 0.19–0.78 units, and led to a persistently positive RPE. The amounts of rhizosphere-primed C were positively correlated with rhizosphere pH. Rhizosphere pH affected the RPE mainly through influencing microbial biomass, activity and utilization of root exudates, and the availability of SOC to microbes. Furthermore, the amount of rhizosphere primed C correlated negatively with microbial biomass atom% ¹⁵N (R² 0.77–0.98, n = 12), suggesting that microbes in the rhizosphere acted as the immediate sink for N released from enhanced SOC decomposition via the RPE.

Conclusion

N form was an important factor affecting the magnitude and direction of the RPE via its effect on rhizosphere pH. Rhizosphere pH needs to be considered in SOC and RPE modelling.

INTRODUCTION

Rhizosphere effects can induce an increase or decrease in the decomposition of soil organic carbon (SOC), which is frequently referred to as a positive or negative rhizosphere priming effect (RPE). These rhizosphere effects include root release of organic C and N substances, depletion of nutrients and water, or root-induced chemical changes such as soil pH (Kuzyakov, 2002; Hinsinger et al., 2003). In particular, root exudates can stimulate microbial growth and activity, generally leading to an increase in SOC turnover in the rhizosphere (Cheng and Kuzyakov, 2005; Dijkstra and Cheng, 2007; Phillips et al., 2011). Plant traits such as plant phenology and shoot and root biomass (Fu and Cheng, 2002; Cheng et al., 2003), and environmental factors such as light intensity (Kuzyakov and Cheng, 2001), temperature (Zhu and Cheng, 2011) and CO₂ concentration (Carney et al., 2007) could have a great impact on the RPE through affecting the quantity or quality of root exudates.

Root-mediated changes in rhizosphere pH are a well-documented interaction between soil and the root interface (Hinsinger et al., 2003; Tang et al., 2004), but its impact on the RPE has seldom been studied. The release of protons is believed to be largely influenced by plant N nutrition. In general, ammonium (NH₄-N) addition leads to rhizosphere acidification because of excess uptake of cations over anions, while nitrate (NO₃-N) nutrition causes rhizosphere alkalization (Hinsinger et al., 2003). Ramirez et al. (2010) found that soil microbial respiration was closely correlated with changes in soil pH following application of different N sources. Under field conditions, long-term SOC sequestration could be favoured by the acidity generated during NH₄-N oxidation (Malhi et al., 1997; Zhang and Wang, 2012). Soil pH can affect the decomposition of SOC directly through affecting SOC solubility or indirectly via changes in microbial activity (Andersson et al., 2000; Briedis et al., 2012). This then suggests that the N-form-induced change in rhizosphere pH may also have an impact on the root-induced decomposition of SOC, namely the RPE. Addition of N either increases (Cheng and Johnson, 1998), decreases (Liljeroth et al., 1994) or has no effect on the RPE (Cheng et al., 2003). Contradictory mechanisms have been proposed to explain the observed effects of N on the RPE (Kuzyakov et al., 2000; Fontaine et al., 2003). It is not clear whether the influence of N addition on the RPE also depends on the source of N applied due to its role in regulating rhizosphere pH.

While factors affecting the RPE have been extensively examined, few studies have looked at how the RPE may affect the mineralization of soil N and hence its availability for plant uptake. Owing to the close link between SOC decomposition and N mineralization, increased N availability has been expected in the rhizosphere following enhanced RPE (Murphy et al., 2015; Rousk et al., 2016). Under N-limited conditions, positive RPEs have been proposed as an adaptive behaviour of plants to gain N through microbial mineralization of N-rich SOC (Kuzyakov, 2002; Luo et al., 2006; Dijkstra et al., 2013). However, enhanced N immobilization has also been detected in the rhizosphere characterized by continuous supply of labile C via rhizodeposits (Zak et al., 2000; Dijkstra et al., 2013). Therefore, positive RPEs have resulted in both increased and decreased net N mineralization (Dijkstra et al., 2009; Phillips et al., 2011; Bengtson et al., 2012). Even in the case of increased N mineralization, rhizosphere-primed C did not correlate with rhizosphere-primed N (Cheng, 2009; Dijkstra et al., 2009). The effect of the RPE on the contribution of soil-derived N to plant uptake remains largely unknown.

This study aimed to examine the change in the RPE under supply of NO₃-N and NH₄-N. The application of different N sources was expected to produce a major difference in rhizosphere pH. The study also used a ¹⁵N-labelled N source to examine the fate of soil-derived N following an enhanced RPE. We hypothesized that the plants supplied with NH₄-N would have lower RPEs than those fed with NO₃-N due to NH₄-induced rhizosphere acidification, in comparison to the alkalization caused by NO₃-N supply.

MATERIALS AND METHODS

Soil

Topsoil (0–10 cm) was collected from a C₄ Kangaroo grassland (Themeda triandra) in Gulgong, New South Wales (32°11′S, 149°33′E). The soil was air-dried and sieved through a 2-mm sieve. The soil was a granite-derived sandy loam. It had the following basic properties: pH 5.0 (0.01 m CaCl₂), organic C 27 g kg⁻¹, total N 1.6 g kg⁻¹, clay 130 g kg⁻¹ and soil pH buffer capacity 30 mmol_c kg⁻¹ unit pH⁻¹. The δ¹³C value of the SOC was −18.0 ‰.

Experiment set up

The experiment consisted of three plant treatments, two N forms and three replicates. The three plant treatments were wheat, white lupin and the no-plant control. Wheat (Triticum aestivum L. ‘Yitpi’) and white lupin (Lupinus albus L. ‘Kiev’) were grown in PVC columns (diameter 10 cm, height 40 cm) containing 2.8 kg air-dried C₄ soil. To prevent anaerobic conditions, a 3-cm layer of plastic beads sealed in nylon bags (mesh size 50 µm) was placed at the bottom portion of each column. Prior to planting, basal nutrients were mixed into the soil at following rates (mg kg⁻¹): KH₂PO₄, 180; K₂SO₄, 120; CaCl₂.2H₂O, 180; MgSO₄.7H₂O, 50; MnSO₄.H₂O, 15; ZnSO₄.7H₂O, 9; CuSO₄.5H₂O, 6; Na₂MoO₄.2H₂O, 0.4; Na[FeEDTA].3H₂O, 5.5. The two N forms were NH₄-N and NO₃-N, applied as Ca(¹⁵NO₃)₂ (2 atom% ¹⁵N) and (¹⁵NH₄)₂SO₄ (2 atom% ¹⁵N), respectively, at a rate of 30 mg N kg⁻¹ at sowing. To minimize possible nitrification and maximize NH₄-N uptake by plants as the NH₄-N form, all N sources were applied twice weekly at 15 mg N kg⁻¹ from week 4. The total amount of N supplied was 90 and 240 mg N kg⁻¹ for the harvests at day 42 and 80, respectively. They were applied 1 d before the total CO₂-trapping, and 3 d before the destructive harvest. Soil columns without plants but with the two N forms (30 mg N kg⁻¹ once at the start of the experiment) were included as no-plant controls. To test whether possible nitrification affected soil pH and consequently total amount of soil-derived CO₂ released from the control columns with N applied once, compared with planted columns with regular N additions, another set of control columns were set up to include five rates of NH₄-N (10, 30, 50, 70 and 100 mg N kg⁻¹). The columns were destructively harvested at the vegetative (day 42) and flowering stage (day 80) (three replicates at each time). An additional eight pots (two species × two N forms × two replicates) were filled with washed fine sand plus 10 g of C₃ soil as an inoculant, to account for isotopic fractionation between the root tissue and root-respired CO₂.

Planting and CO₂ trapping

The experiment was conducted in a controlled environmental room with 20 °C and 18 °C during the day and night, respectively, a light intensity of 400 µmol m⁻² s⁻¹ and a day length of 14 h. Twelve germinated seeds of wheat and eight seeds of white lupin were sown in a row in the centre of each column. After emergence, plants were thinned to six per column for wheat, and four for white lupin. All the columns were weighed and maintained at 80 % water-holding capacity by watering with reverse osmosis water for the soil columns and with Hoagland solution (Hoagland and Arnon, 1950) for the sand columns. The sand columns had been watered with Hoagland solution twice daily to meet the requirement for water and nutrient by plants. A modified CO₂-trapping system (Wang et al., 2016) was used to trap total below-ground CO₂ produced in each column 3 d before each harvest. Briefly, CO₂-free air was introduced through an air inlet at the top of each column after being pumped through 1 m NaOH via an air stone. The outlet, on the opposite side of the soil column, was fitted with a vacuum line passing through two CO₂ traps containing 150 mL of 0.5 m NaOH. The simultaneous operation of the pump and vacuum facilitated airflow through the soil columns, reduced the pressure within the headspace and stopped possible leaking around the plant stems. The CO₂ was trapped for 48 h in 150 mL of 0.5 m NaOH at each harvest.

Harvest and measurements

At harvest, plant shoots were cut off at the soil surface one column at a time. The rhizosphere soil was collected by gently shaking off soil adhering to the roots. All rhizosphere soil was sieved through a 2-mm sieve, and sub-samples were immediately put into a plastic bag and stored at 4 °C for determination of soil microbial biomass C (MBC), microbial biomass N (MBN) and soil respiration. Another set of samples were air-dried for determination of soil pH, total organic C and N. Plant roots were carefully washed with water, and root length was quantified using WinRHIZO Pro 2003b (Regent Instruments, Quebec City, Canada) and an EPSON EU-35 scanner (Seiko Epson Corp., Suwa, Japan). All plant materials were oven-dried at 70 °C, weighed and ground in a ball mill.

The δ¹³C in the shoot and root samples was analysed using an isotope ratio mass spectrometer (Sercon 20–22, Crewe, UK). The C and N concentrations of shoot, root and rhizosphere soil were determined by dry combustion using a CHNS analyser (PerkinElmer EA2400, Shelton, CT, USA). Soil pH was measured in 0.01 m CaCl₂ (1: 5 soil solution ratio) after end-over-end shaking for 1 h and centrifuging at 700 g for 10 min.

Microbial respiration in the rhizosphere soil was determined after 12 h of incubation at 25 °C to give an indication of the quantity of labile C substrates (e.g. root exudates) and microbial activity (Wang et al., 2016). The CO₂ concentration within the headspace of jars was measured using an infrared gas analyser (Servomex 4210 Industrial Gas Analyser, Cowborough, UK) (Rukshana et al., 2012). MBC and MBN were determined following the chloroform fumigation-extraction procedure according to Vance et al. (1987a). Briefly, 8 g of fumigated or non-fumigated moist soil was extracted using 40 mL of 0.5 m K₂SO₄. Total SOC in the extracts was determined colorimetrically following wet digestion with dichromate-sulfuric acid at 135 °C for 30 min (Cai et al., 2011). Total soluble N was determined by persulphate oxidation of both fumigated and non-fumigated extracts (Cabrera and Beare, 1993), and the final concentrations of inorganic N (NH₄⁺ + NO₃⁻) were measured using a QuickChem 8500 flow injection analyser (Lachat Instruments, Loveland, CO, USA). MBC and MBN were calculated as the difference in concentrations of total organic C and total N in extracts between fumigated and non-fumigated soils, adjusted by a proportionality coefficient of 0.45 (Jenkinson et al., 2004). Soil inorganic N was measured on the non-fumigated and non-oxidized 0.5 m K₂SO₄ extracts using the flow injection analyser. Before determination of microbial biomass ¹³C and ¹⁵N by mass spectrometry, both fumigated and un-fumigated K₂SO₄ extracts were freeze-dried, and the dried salts were finely ground using a mortar and pestle (Dijkstra et al., 2006b).

The amount of CO₂ trapped in 0.5 m NaOH was determined by titrating 10 mL of NaOH solution with 0.5 m HCl after addition of 5 mL 1 m BaCl₂. To produce SrCO₃ precipitates for δ¹³C analysis, 5 mL of 1 m SrCl₂ was added to another 10 mL of NaOH solution. The suspension pH was then adjusted to 7.0 using 0.3 m HCl. The δ¹³C of the SrCO₃ was analysed by mass spectrometry.

Calculations and statistical analysis

To partition the total CO₂ efflux using the ¹³C natural abundance method, the following equations were applied (Mary et al., 1992):

$${\text{C}}_4={\text{C}}_t\times \left({\delta }_{\text{t}}–{\delta }_{\text{3}}\right)/\left({\delta }_{\text{4}}–{\delta }_{\text{3}}\right)$$

$${\text{C}}_{\text{primed}}={\text{ C}}_{\text{4}}–{\text{ C}}_{\text{control}}$$

where C_t is the total C from below-ground CO₂, C₄ is the amount of CO₂-C derived from C₄ soil, δ_t is the δ¹³C value of the C_t, δ₃ is the δ¹³C value of the root-derived C (CO₂ trapped from the sand columns) and δ₄ is the δ¹³C value of the C₄ soil C (CO₂ trapped from the no-plant control columns). C_control is the amount of CO₂-C evolved from the column without plants. C_primed is the total primed C.

The isotopic signature of microbial biomass was calculated as follows (Dijkstra et al., 2006b):

$${\delta }^{\text{13}}{\text{C}}_{\text{MB}}=\left({\delta }^{\text{13}}{\text{C}}_F*{\text{C}}_F–{\delta }^{\text{13}}{\text{C}}_{\text{UF}}*{\text{C}}_{\text{UF}}\right)/{\text{C}}_{\text{MB}}$$

atom%¹⁵N_MB = (atom% ¹⁵N_F*N_F – atom% ¹⁵N_UF*N_UF)/N_MB

where MB, F and UF are microbial biomass, fumigated and un-fumigated fractions, respectively. C_F, N_F, C_UF and N_UF are total dissolved organic C and N in the fumigated and un-fumigated extracts, respectively.

A two-way ANOVA was conducted using Genstat (11^th Version) (VSN International, Hemel Hempstead, UK) to assess the effect of N form and plant species at each harvest. Where necessary, square-root transformation of the data was performed to obtain a normal distribution. Significant (P ≤ 0.05) differences between means were identified using the Tukey’s honest significant difference test. Single linear regression analyses were used to determine the relationship between rhizosphere-primed C and soil pH or atom% ¹⁵N in microbial biomass.

RESULTS

Nitrogen form did not affect the shoot dry weight of either species. However, compared with NO₃-N, NH₄-N nutrition decreased root dry weight and total root length of wheat by 16 and 28 %, respectively, at day 80 (Table 1). The concentration of N in the shoot or root did not differ between the two N forms for either species. The δ¹³C value of CO₂ trapped from the sand column (root-derived C) receiving NO₃-N was 1.2–2.7 ‰ more enriched than that of the root tissues, indicating an isotopic fractionation between the root tissue and root-respired CO₂ (Table 1). Given that the δ¹³C value of root tissue was not affected by N forms, the δ¹³C value of CO₂ trapped from the sand columns with NO₃-N addition was used to represent the root-derived C for both N treatments. Shoot biomass of both species in the sand columns was lower when N was applied as NH₄-N compared with NO₃-N (data not shown), probably due to low rhizosphere pH in this poorly buffered system.

Table 1.

Dry biomass and N concentration of shoot and roots, root length and ¹³C abundance in roots of wheat and white lupin supplied with NO₃-N and NH₄-N at days 42 and 80

Harvest time	Species	N form	Biomass (g per column)		N concentration (%)		Root length (m per column)	13C abundance (‰)
			Shoot	Root	Shoot	Root		Root	Root-derived CO2
Day 42	Wheat	NO3/NH4-N	4.04 b	0.87 b	5.22 b	2.90 a	183 b	−31.89 a	−30.6 a
	White lupin	NO3/NH4-N	3.45 a	0.55 a	4.70 a	2.98 a	22 a	−30.30 b	−28.7 b
Day 80	Wheat	NO3-N	15.47 b	2.50 c	3.41 a	2.01 a	437 c	−32.24 a	−30.7 a
		NH4-N	14.90 b	2.11 b	3.32 a	1.93 a	314 b	−32.26 a
	White lupin	NO3-N	11.23 a	1.05 a	4.19 b	3.00 b	30 a	−31.74 b	−29.0 b
		NH4-N	10.19 a	1.13 a	4.23 b	3.17 b	32 a	−31.62 b
Two-way ANOVA
Day 42	Species		**	**	*	n.s.	***	***	***
	N form		n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
	Species × N form		n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
Day 80	Species		***	***	***	***	***	**	***
	N form		*	*	n.s.	n.s.	***	n.s.
	Species × N form		n.s.	*	n.s.	n.s.	***	n.s.

n.s., *, ** and *** represent P > 0.05, P < 0.05, P < 0.01 and P < 0.001, respectively. For each column, different letters indicate significant differences between means for a given harvest time (two-way ANOVA, Tukey’s test, P < 0.05). At day 42, the effect of N form and interactions between N from and species are not significant and thus only the main effects of species (average of two N forms) are presented.

When compared with the no-plant control, rhizosphere pH was 0.37 units lower at both harvests of wheat fed with NH₄-N, but was 0.19 and 0.78 units higher for wheat fed with NO₃-N at days 42 and 80, respectively (Fig. 1). Supply of NH₄-N decreased soil pH by 0.16–0.39 units in the rhizosphere of white lupin, relative to the no-plant control, in contrast to alkalization of 0.2–0.4 units under NO₃-N supply. Neither soil pH nor total CO₂ production from the control columns were affected by the rate of NH₄-N (data not shown) or N form (Figs 1 and 2A). The little change in soil pH following addition of NH₄-N at different rates was mainly attributed to the relatively high pH buffering capacity of the C₄ soil used in this study.

Fig. 1.

Soil pH of the no-plant control and the rhizosphere of wheat and white lupin supplied with NO₃-N or NH₄-N at days 42 and 80. Error bars represent ±s.e.m. of three replicates. Different italic letters above the bars indicate significant differences among treatments at each harvest time (Tukey’s test, P < 0.05).

Fig. 2.

Total below-ground CO₂ efflux (A), δ¹³C values of CO₂ (B) and primed soil C (C) under wheat and white lupin fed with NO₃-N and NH₄-N or no-plant control at days 42 and 80. Error bars represent s.e.m. of four replicates. Different italic letters above the bars indicate significant differences among treatments at each harvest time (Tukey’s test, P < 0.05).

Total below-ground CO₂ efflux was 17–18 % lower under NH₄-N than NO₃-N addition in all planted columns, except that N form did not affect the CO₂ efflux under wheat at day 42 (Fig. 2A). The δ¹³C value of CO₂ showed no difference between the two N sources for white lupin and the no-plant control, but decreased by 0.9 ‰ for wheat fed with NH₄-N relative to NO₃-N (Fig. 2B). The CO₂ collected from the control column had a consistent δ¹³C value of −13.2 ‰ during the whole experiment. Regardless of plant species, N form had a significant impact on the decomposition of SOC and the RPE. A decreased or negative RPE was observed for the plants fed with NH₄-N while a consistently higher and positive RPE was shown for the plants fed with NO₃-N (Fig. 2C). The amounts of rhizosphere-primed C were positively correlated with rhizosphere pH (P < 0.05) (Fig. 3A), and negatively correlated with the atom% ¹⁵N in the microbial biomass (P < 0.05) (Fig. 3B, C). Other soil properties such as extractable organic C, soil respiration and microbial C and N were not correlated with primed C, when all data were included in the analysis. The SOC concentration in the rhizosphere of plants fed with NH₄-N was 3–5 % higher (P < 0.05) than that of NO₃-N (Table 2).

Fig. 3.

Relationships between rhizosphere-primed soil C and rhizosphere pH (A) and microbial biomass atom% ¹⁵N for wheat (B) and white lupin (C) fed with NO₃-N and NH₄-N at days 42 and 80. Open and closed circles denote data for wheat and white lupin, respectively. Open and closed triangles denote the data for urea-fed wheat and N₂-fixing white lupin, respectively, from Wang et al. (2016).

Table 2.

Extractable soil organic C, soil respiration, rhizosphere soil organic C (SOC), microbial C and N in the no-plant control and the rhizosphere of wheat and white lupin supplied with NO₃-N or NH₄-N at days 42 and 80

Harvest time	Treatment	N form	Extractable organic C (µg C g−1 soil)	Soil respiration (µg CO2 g−1 soil)	Rhizosphere SOC (mg g–1)	Microbial C (µg C g−1 soil)	Microbial N (µg N g−1 soil)
Day 42	Wheat	NO3-N	84 a	161 b	27.6 ab	256 b	20.6 b
		NH4-N	96 a	186 b	28.2 b	216 a	15.2 a
	White lupin	NO3-N	107 a	387 c	26.8 a	376 c	41.1 c
		NH4-N	151 b	363 c	28.1 b	234 ab	24.6 b
	Control	NO3-N	90 a	88 a	27.4 ab	248 ab	24.1 b
		NH4-N	86 a	79 a	27.3 ab	266 b	23.3 b
Day 80	Wheat	NO3-N	39 a	247 c	27.4 ab	366 bc	33.8 b
		NH4-N	87 b	166 b	28.2 c	256 a	23.3 a
	White lupin	NO3-N	105 b	520 e	26.5 a	416 c	46.0 c
		NH4-N	146 c	371 d	27.9 bc	305 ab	29.0 b
	Control	NO3-N	77 b	69 a	27.0 a	283 a	29.3 b
		NH4-N	82 b	76 a	27.1 a	280 a	27.7 b
Two-way ANOVA
Day 42	Treatment		**	**	n.s.	*	**
	N form		*	n.s.	*	*	*
	Treatment × N form		**	n.s.	*	**	*
Day 80	Treatment		***	***	*	*	***
	N form		***	***	*	***	***

n.s., *, ** and *** represent P > 0.05, P < 0.05, P < 0.01 and P < 0.001, respectively. Different letters indicate significant differences among treatments at each harvest time (Tukey’s test, P < 0.05).

The K₂SO₄-extractable organic C (EOC) was higher in the rhizosphere of plants fed with NH₄-N than NO₃-N, except for wheat at day 42 (Table 2). Noticeably, EOC in the rhizosphere of white lupin was 72 and 83 % higher under NH₄-N nutrition than the no-plant control at days 42 and 80, respectively. Wheat showed little effect on EOC, except a 51 % decrease at day 80 when fed with NO₃-N. The rhizosphere soil respiration during 12 h of incubation was invariably higher in the planted columns than the no-plant controls, and in the rhizosphere of white lupin than wheat. When compared between the two N forms, soil respiration was lower under NH₄-N than NO₃-N nutrition at day 80 for both species (Table 2).

The effect of N form on MBC and MBN in the rhizosphere soil differed between wheat and white lupin (Table 2). Microbial biomass C and MBN in the rhizosphere of white lupin supplied with NO₃-N increased by 46 and 68 %, respectively, relative to the no-plant control, but remained less changed for wheat. The addition of NH₄-N showed no effect on MBC, but decreased MBN in the rhizosphere of wheat, compared to the control. For both species, MBC and MBN were invariably lower under NH₄-N than NO₃-N supply.

¹⁵N abundance in the extractable inorganic N in the rhizosphere did not differ between the two N sources except for white lupin at day 80 (Fig. 4A). Nitrogen form greatly affected the ¹⁵N enrichment in microbial biomass and plants (Fig. 4B, C). Microbial biomass was less enriched with ¹⁵N when NO₃-N was applied, in comparison with NH₄-N (Fig. 4B). Nevertheless, an opposite trend was shown in the plant shoot at day 42, although no effect of N form was detected at day 80 (Fig. 4C). In the no-plant control and bulk soil, N form did not affect microbial atom%¹⁵N.

Fig. 4.

Atom% ¹⁵N in the rhizosphere inorganic N (A), plant shoots (B) and rhizosphere microbial biomass (C) of wheat and white lupin supplied with NO₃-N or NH₄-N at days 42 and 80. Error bars represent s.e.m. of three replicates. Different italic letters above the bars indicate significant differences among treatments at each harvest time (Tukey’s test, P < 0.05).

DISCUSSION

This study revealed that the supply of NO₃-N led to enhanced RPEs in contrast to decreased or negative RPEs caused by NH₄-N addition. The higher SOC concentration in the rhizosphere of plants fed with NH₄-N than NO₃-N provided further evidence of reduced SOC decomposition by NH₄-N nutrition. This was also consistent with the enhancement of SOC sequestration due to NH₄-N fertilization reported by other studies (McAndrew and Malhi, 1992; Malhi et al., 1997; Zhang and Wang, 2012). Furthermore, the C₄ soil used in this study was a mixture of different SOC pools with different δ¹³C values; for example, the easily decomposable C was more ¹³C-enriched (-13.2 ‰) than non-labile C fractions associated with soil minerals (−18.8 ‰) (Wang et al., 2016). If the persistent positive RPE resulted in the decomposition of more stable SOC at the late stage, the NO₃-induced increase in RPE might have been underestimated due to the use of δ¹³C values of soil-derived CO₂ for the non-planted control columns (−13.2 ‰). A decreased or negative RPE following NH₄-N addition should be less affected by the non-uniformity of the ¹³C signature among different SOC fractions. As indicated by the constant δ¹³C values of CO₂ evolved from the control columns with time, decomposition of SOC in the absence of plants was mainly from labile C pools. Low or negative RPEs mean that the quantity of soil-derived CO₂ from the planted columns was similar to or less than that from the control. Overall, N form had a large impact on both the magnitude and the direction of the RPE, and future evaluation of the effect of N addition on the RPE should also consider the form of N sources.

The effects of N form on plant biomass and root exudation (indicated by EOC and rhizosphere microbial respiration) could not account for the significant changes in the RPE. Many plant species had shown slower growth when N was applied as NH₄-N relative to NO₃-N (Raab and Terry, 1994; Britto and Kronzucker, 2002; Guo et al., 2002). In this study, the negative effects of NH₄-N on plant growth were confirmed by decreases in root biomass and length for wheat plants at day 80. There is evidence to suggest that plants with higher shoot or root biomass generally produce greater RPEs (Fu and Cheng, 2002; Dijkstra et al., 2006a), although RPE showed no correlation with shoot and root biomass in this study. For instance, restricted root growth by NH₄-N nutrition was detected only for wheat at day 80, but a decreased or negative RPE occurred for both species at both harvests. This was in line with our previous finding that plant biomass played a minor role in accounting for the RPE of species with contrasting effect on soil acidification (Wang et al., 2016). On the other hand, NH₄-N rather than NO₃-N nutrition could increase root exudation by plant roots, due to NH₄-induced changes in root architecture such as increased branching (Martins-Loucao et al., 2000), cell death in root cortex (Brown and Hornby, 1987) or increased membrane permeability by low pH (Yin and Raven, 1998). If N form-induced changes in the quantity of root exudates was mainly responsible for the differences in the RPE between the two N forms, the RPE would be higher under NH₄-N than NO₃-N nutrition. Finally, no differences were detected in either 12-h soil respiration or 48-h total CO₂ efflux between NH₄-N- and NO₃-N-amended controls with no plants, suggesting N form per se had less impact on microbial decomposition of SOC.

It is evident that rhizosphere pH was positively correlated with the amount of rhizosphere-primed C. As expected, NH₄-N nutrition caused rhizosphere acidification with pH decreasing by up to 0.38 units. This resulted from H⁺ production during excess uptake of cations over anions by plant roots and/or, to a lesser extent, nitrification (Malhi et al., 1991). In contrast, NO₃-N nutrition led to rhizosphere alkalization (increasing pH by up to 0.78 units), due to excess uptake of anions over cations (Tang et al., 1999). In decomposition studies, increasing soil pH increased the magnitude of the positive priming effect (Perelo and Munch, 2005; Luo et al., 2011), and the optimum pH was between 6 and 8 (Blagodatskaya and Kuzyakov, 2008; Nang et al., 2017). Our previous study also found that rhizosphere pH was mainly responsible for the variation in the RPE among crop species differing in rhizosphere acidification (from pH 4.1 to 5.2) (Wang et al., 2016). It is expected that the addition of NH₄-N could decrease the RPE more in acid soils than in neutral or alkaline soils.

The N source-induced changes in rhizosphere pH could affect the RPE in the following ways. Firstly, low rhizosphere pH greatly decreased microbial activity and growth (Fig. 5), as indicated by lower soil respiration and microbial biomass C and N under NH₄-N than NO₃-N supply. In other studies, the effect of soil pH on SOC decomposition was mainly via its effect on microbial growth and activity (Andersson et al., 2000; Briedis et al., 2012), and the turnover of SOC correlated positively with soil pH in the range of 4.3–5.3 (Leifeld et al., 2008). Secondly, decreased microbial biomass and activity at low pH possibly led to inefficient use of root exudates or soluble SOC, which accounted for the greater accumulation of EOC in the rhizosphere of plants fed with NH₄-N than NO₃-N (Fig. 5). Negative relationships between soil pH and EOC were also observed in other studies (Vance et al., 1987b; Pietri and Brookes, 2008). Finally, addition of NO₃-N might have favoured microbial utilization of SOC due to increased availability of SOC at higher pH. Increases in the solubilization or desorption of humic substances from the surface of mineral colloids were often detected following elevated soil pH by liming (Andersson et al., 2000; Garbuio et al., 2011). The activation of microbes by root exudates was believed to be the main mechanism responsible for the enhanced RPE (Kuzyakov et al., 2000), but low soil pH following NH₄-N addition decreased microbial activity regardless of the quantity of root exudates.

Fig. 5.

Conceptual diagram of the effects of N form (NO₃-N vs. NH₄-N)-induced pH changes on the rhizosphere priming effect (the RPE) and involved N immobilization mechanisms.

Microbes in the rhizosphere acted as an immediate sink for N released from enhanced SOC decomposition via the RPE. The increase in both MBN and MBC indicated that microbial immobilization of N was enhanced by increased microbial growth. Moreover, decreased ¹⁵N abundance in microbes under NO₃-N than NH₄-N supply possibly reflected a greater dilution of ¹⁵N at higher RPE and an enhanced contribution of soil-derived N, relative to fertilizer-derived N. This was directly supported by the negative correlation between primed C and microbial biomass ¹⁵N abundance in the rhizosphere of the two species. The recovery of ¹⁵N in microbial biomass in the unplanted columns (Fig. 4C) and in the bulk soil of the planted columns (data not shown) did not differ between two N sources, indicating a lack of microbial discrimination between the two N forms. Norton and Firestone (1996) and Inselsbacher et al. (2010) also found that N form per se did not affect the amount of N assimilated by microbes. Immediate and rapid utilization of soil-derived N following a positive RPE by microbes was possible (Fig. 5), given the greater substrate affinities, larger specific surface area and greater spatial distribution of microbes than plant roots (Lipson and Näsholm, 2001; Kaštovská and Šantrůčková, 2011). Moreover, amino acids released during the decomposition of SOC could be more efficiently acquired by microbes before they were mineralized and came into soil solution for plant uptake (Kuzyakov and Xu, 2013). Little difference in atom% ¹⁵N in the extractable inorganic N in the rhizosphere between the two N forms, in most cases, might also reflect a rapid immobilization of soil-derived N by microbes in the case of an enhanced RPE.

An enhanced RPE under NO₃-N nutrition did not result in a greater contribution of soil-derived N to plant uptake. Higher atom% ¹⁵N in white lupin fed with NO₃-N than NH₄-N at day 42 might indicate a faster uptake of NO₃-N because of its higher mobility in the soil (Norton and Firestone, 1996; Burger and Jackson, 2003; Song et al., 2007). Also, plants fed with NH₄-N might have fixed more N₂ than those fed with NO₃-N, due to the slower uptake or lower effectiveness of NH₄-N at suppressing nitrogenase activity (Silsbury et al., 1986; Svenning et al., 1996). On the other hand, if N mineralized via the RPE under NO₃-N nutrition was rapidly captured by microbes in the first place, the availability of soil-derived N to plants would decrease. Other studies also found that an enhanced RPE did not result in apparent increases in plant N uptake due to short-term N immobilization (Cheng, 2009; Dijkstra et al., 2011; Kuzyakov and Xu, 2013). Nevertheless, N immobilized by microbes with a short turnover time (days to weeks) could eventually be transferred to the forms that can be utilized by plants (Hodge et al., 2000; Cheng and Kuzyakov, 2005; Harrison et al., 2007). Overall, the contribution of the RPE to plant N uptake in this study was complicated and may have been regulated by many factors. Further studies are required to quantify the partitioning of SOC-derived N between microbes and plants in response to an enhanced RPE during short- and long-term periods.

CONCLUSION

Our results revealed that N form affected the RPE mainly through its impact on rhizosphere pH. Rhizosphere acidification under NH₄-N resulted in decreased or negative RPEs and hence favoured SOC accumulation, while alkalization in the supply of NO₃-N enhanced the RPE and hence SOC mineralization. Noticeably, microbes in the rhizosphere acted as the immediate sink for N released during SOC decomposition. Enhanced microbial N immobilization under NO₃-N supply might delay plant uptake of soil N under field conditions. The role of N fertilization on SOC sequestration should take into account the apparent effect of N form-induced pH change on the RPE. Under field conditions, the addition of organic amendments or crop residue return is essential to replenish SOC lost via the RPE under NO₃-N supply. This study also suggests that rhizosphere pH needs to be considered in SOC and RPE modelling.