Nitrogen fertilization has minimal influence on rhizosphere effects of smooth crabgrass ( Digitaria ischaemum ) and bermudagrass ( Cynodon dactylon )

Aims Plants generally respond to nitrogen (N) fertilization with increased growth, but N addition can also suppress rhizosphere effects, which consequently alters soil processes. We quantified the influence of N addition on rhizosphere effects of two C4 grasses: smooth crabgrass (Digitaria ischaemum) and bermudagrass (Cynodon dactylon). Methods Plants were grown in nutrient-poor soil for 80 days with either 20 or 120 μg NH4NO3-N g dry soil−1. N mineralization rates, microbial biomass, extracellular enzyme activities and bacterial community structure were measured on both rhizosphere and bulk (unplanted) soils after plant harvest. Important Findings Fertilization showed nominal differences in net N mineralization, extracellular enzyme activity and microbial biomass between the rhizosphere and bulk soils, indicating minimal influence of N on rhizosphere effects. Instead, the presence of plant roots showed the strongest impact (up to 80%) on rates of net N mineralization and activities of three soil enzymes indicative of N release from organic matter. Principal component analysis of terminal restriction fragment length polymorphism (T-RFLP) also reflected these trends by highlighting the importance of plant roots in structuring the soil bacterial community, followed by plant species and N fertilization (to a minor extent). Overall, the results indicate minor contributions of short-term N fertilization to changes in the magnitude of rhizosphere effects for both grass species.

higher biomass of soil microorganisms and fauna (Griffiths 1994) in comparison to bulk soils.The release of low molecular weight organic compounds such as organic acids from roots, and the uptake of anions or cations by roots, can change soil pH (Bardgett et al. 1999;Marschner et al. 1986) and microbial community composition (DeAngelis et al. 2009;Shi et al. 2011).The changes in microbial biomass and community composition can lead to further changes in microbial function in the rhizosphere (Schimel and Schaeffer 2012).Activities of extracellular enzymes for soil organic matter decomposition and nutrient cycling are often higher in the rhizosphere than in the bulk soil (Phillips and Fahey 2006;Priha et al. 1999), and microbial respiration and N mineralization tend to be higher in the rhizosphere (Kuzyakov 2002;Phillips and Fahey 2008;Zhu and Cheng 2011).These plant-microbe interactions in the rhizosphere have significant implications for soil carbon storage and nutrient availability to plants (Cheng et al. 2014;Grayston et al. 1996;Kuzyakov 2002;Sun et al. 2014).
Because rhizodeposition is the main driver of rhizosphere effects (Kuzyakov 2002;Paterson 2003), factors that regulate rhizosphere C fluxes may control the magnitude of rhizosphere effects (Jones et al. 2004).Higher soil fertility due to N fertilization may reduce relative belowground C allocation (Giardina et al. 2004;Phillips and Fahey 2007) and thus lead to lower rhizosphere effects (Ai et al. 2012;Blagodatskaya et al. 2014;Fontaine et al. 2011;Liljeroth et al. 1994;Phillips and Fahey 2008).However, neutral (Cheng et al. 2003) and even positive (Phillips and Fahey 2008) responses of rhizosphere effects to N fertilization have also been reported.These inconsistent results may be caused by differences in plant species and mycorrhizal association, soil type, the amount and chemical composition of fertilizers, and the timing and duration of fertilization in different studies.To better predict rhizosphere effects and to incorporate them into ecosystem process models (e.g.Perveen et al. 2014), we need more mechanistic studies of rhizosphere effects and their responses to environmental factors such as N addition.
In this study, we focus on grasses to better understand how grass rhizospheres impact soil processes under different N levels.Analysis of grass rhizospheres tend to be conducted in the field without a root exclusion treatment.In this study, we examined rhizosphere effects in a controlled growth chamber facility and assessed the role of N fertilization in modifying the responses.Nitrogen (N) addition can restrain a plant's ability to influence soil processes via its suppression on rhizosphere effects in soil.We hypothesized that N addition reduces the magnitude of rhizosphere effects including microbial biomass, microbial extracellular enzyme activity, N mineralization rates (N min ) and microbial community composition.To test this hypothesis, we grew two grass species in nutrient-poor forest soil receiving either low or high N addition from seed to flowering (~80 days) in a growth chamber.We measured plant biomass and a number of soil variables (pH, microbial biomass C and N, extracellular enzyme activity, N min and bacterial community composition) in soils that were occupied by the plants (rhizosphere soil) or kept unplanted (bulk soil).Previous studies of rhizosphere effects focused on trees, crops and unmanaged natural grassland species, whereas these two grasses, annual smooth crabgrass (Digitaria ischaemum Schreb.) and bermudagrass (Cynodon dactylon L.), are important introduced species naturalized in urban and rural landscapes in North America.Here, we report the rhizosphere effects of grasses in comparison to results found in other plant species.

Experimental design
The experiment was conducted in the Weil Hall controlled environment facility at Cornell University.We grew two grass species, annual smooth crabgrass (D. ischaemum Schreb.) and bermudagrass (C.dactylon L.).These species have been introduced to North America and are now naturalized in urban and rural landscapes.We collected surface (0-20 cm A-horizon, with little O-horizon) mineral soils (silt loam, Inceptisols) from a deciduous forest dominated by sugar maple (Acer saccharum) near Mount Pleasant farm of Cornell University, New York.Most urban grasslands are established on former forest and agricultural lands (Raciti et al. 2011).The soil has a pH of 4.61, 30.4 mg g −1 organic C and 2.3 mg g −1 total N.We sieved the soils through a 4-mm screen and filled 24 polyvinyl chloride pots (diameter 8 cm, height 32 cm) with 2146-g soil at a bulk density of 1.42 g cm −3 .Each pot was closed at the bottom with a rubber stopper and has an air inlet and an air outlet consisting of clear plastic tubing.We adjusted soil moisture content in each pot to 70% field capacity and pre-incubated the soils at room temperature (22°C) for 30 days.This preincubation is to minimize the potential impact of initial disturbance (e.g.sieving, packing, rewetting) on measured soil variables, as is commonly used in such experiments (Zhu and Cheng 2012).
All 24 pots were then moved to a growth chamber with 14-h light period (800 µmol m −2 s −1 , 25°C, 60% relative humidity, 400 p.p.m.CO 2 ) and 10-h dark period (20°C).Eight pots were planted with bermudagrass seeds (100 seeds per pot), eight pots were planted with crabgrass seeds (20 seeds per pot) and the other eight pots were kept unplanted.Both plants germinated within 2 weeks.Half of pots in each treatment (n = 4) received 20 (low N addition) and 120 (high N addition) µg N g soil −1 as NH 4 NO 3 solution during the first 2 weeks after germination.The low N addition was used to avoid severe nutrient deficiency of the plants growing in soils that were poor in mineral nutrients (<2 µg g −1 extractable N and P).The high N addition level reflected a typical fertilization concentration added to urban grasslands.The experiment is a complete factorial design with two main factors (planting and N addition) and four replicates.Soil moisture content was maintained at 70% field capacity (or 0.25 g water g −1 dry soil) throughout the experimental period by periodic weighing and watering with deionized water.Anaerobic conditions were avoided by circulating ambient air through each pot for 30 min twice a day using an aquarium pump.We randomly relocated the pots once a week to ensure similar growing conditions for the plants.

Measurements
All pots were destructively harvested after ~80 days of growth after germination.Both plants were at full flowering stage.Shoots were clipped at the soil surface for each planted pot, roots were carefully picked from each pot and then soils from both planted and unplanted pots were homogenized and subsampled for the following analyses.
Harvested plant shoots and roots were washed with deionized water and dried at 60°C to constant weight.Ground plant samples were processed for C (mg g −1 ) and N (mg g −1 ) on a Vario EL III elemental analyzer (GmbH, Hanau, Germany).Soil pH was measured using 10-g soil well mixed in 40-ml water.
Potential N min were determined by extracting NH 4 + and NO 3 − from soils before and after 10-day aerobic incubation at 22°C.Pre-and post-incubation soils (20 g) were extracted with 2 M KCl (50 ml), shaken for 1 h, filtered and frozen until analysis on a SEAL AQ2 analyzer (Seal Analytical Inc., Maquon, WI, USA).N min was calculated as the change in extractable NH 4 + and NO 3 − before and after the 10-day incubation (Phillips and Fahey 2006).Microbial biomass C and N were measured by chloroform fumigation-extraction method (Wu et al. 1990).One subsample (20 g) was extracted with 50-ml 0.05 M K 2 SO 4 solution, another subsample (20 g) was fumigated by ethanol-free chloroform in the dark for 48 h and then extracted with 50-ml 0.05 M K 2 SO 4 solution.The concentration of total organic C and total N in each extract was analyzed using a Shimadzu TOC-V analyzer (Shimadzu Scientific Instruments, Columbia, MD, USA).Microbial biomass C and N were calculated as the difference between fumigated and unfumigated samples, adjusted by a proportionality coefficient (0.45) for both C and N (Jenkinson et al. 2004).
Bacterial community structure was characterized using terminal restriction fragment length polymorphism (T-RFLP).DNA was extracted from soil samples using the PowerSoil DNA isolation kit (Mo Bio Laboratories Inc., Carlsbad, CA, USA).For PCR, 50-µl reactions consisted of 10 µl of 5× GoTaq buffer (Promega Corporation, Middleton, WI, USA), 5 µl of 25-mM MgCl 2 , 1.0 µl of 10-mM dNTP mix (Promega Corporation), 7.5 µl of 10-µM primer 8F, 2.5 µl of 10-µM primer 1492r, 0.6 µl of Go Taq enzyme (Promega Corporation), 1 µl of DNA template and water.The thermocycler settings were programmed as follows: 95°C for 3 min, 35 cycles of (95°C 30 s, 50°C 30 s and 72°C 45 s), final 72°C 12 min and cooled at 12°C.PCR products were cleaned using the QIAEX II gel extraction kit (Qiagen N.V., Netherlands) protocol for desalting and concentrating DNA solutions.The cleaned PCR products (400 ng) were digested in 50-µl reactions with 0.85 µl of HaeIII enzyme and 5 µl of 10× New England Biolabs buffer.The thermocycler program for restriction enzyme digestion was as follows: 6 h at 37°C, 20 min at 80°C, hold at 4°C.The digested samples were then added to 9.85 µl of Hi-Di formamide with 0.15 µl of Liz500 standard and analyzed by the Cornell Life Sciences Core Laboratory.The T-RFLP relative peak area profiles were analyzed using T-REX software (Culman et al. 2009).

Statistical analyses
Rhizosphere effect was calculated as the percent difference between rhizosphere soil (a homogenous sample from the planted pots after picking roots) and bulk soil (a homogenous sample from the unplanted pots) samples for each variable (Phillips and Fahey 2006;Zhu and Cheng 2011) under the same N treatment.A positive (or negative) rhizosphere effect means the variable is higher (or lower) in the rhizosphere than in the bulk soil.Independent samples t-test was used to compare each variable between rhizosphere soil and bulk soil to show the statistical significance of the calculated rhizosphere effect (Fig. 2).It was also used to compare each plant and soil variable between low and high N additions under each of the plant treatment (Fig. 1).For each plant species separately, we used two-way analysis of variance to assess the main effect of rhizosphere (planted vs. unplanted), the main effect of N addition (low N vs. high N) and their interaction on the soil variables (Table 2).For multivariate analysis of T-RFLP data, we performed principal component analysis on the standardized relative peak area of individual terminal restriction fragments (Fig. 3).All statistical analyses were done with PASW Statistics 18.

RESuLTS
Total plant biomass ranged from 27.3 to 41.6 g pot −1 for bermudagrass and from 30.4 to 38.0 g pot −1 for crabgrass (Fig. 1), and root biomass was 10.5-16.0 and 8.6-11.8g pot −1 , respectively.N addition significantly increased bermudagrass total biomass by 43% and crabgrass total biomass by 13% (P < 0.05, Fig. 1).Similar to total biomass, root biomass increased 39% in bermudagrass and 18% in crabgrass with N addition (P < 0.05, Fig. 1), whereas the root biomass: total biomass ratio was not significantly affected by N addition (P > 0.05, Fig. 1).Moreover, N addition strongly enhanced plant N content for both species (P < 0.05, Fig. 1).Compared to the low N treatment, biomass-weighted plant N content (mg g −1 ) in the high N treatment increased 16% for bermudagrass and 93% for crabgrass (P < 0.05, Fig. 1).
Soil pH varied between 4.15 and 4.89 at the end of the experiment (Table 1).Compared to unplanted control soils, pH of the rhizosphere soils of bermudagrass and crabgrass were higher by 0.7 and 0.5 units, respectively (P < 0.001, Fig. 1).However, N addition did not affect pH of planted or unplanted soils (P > 0.05, Fig. 1).
Potential net N mineralization (N min ) was significantly higher in rhizosphere soils of both plants than in unplanted control soils (P < 0.001, Table 2).Rhizosphere effect on N min was 42% (low N) and 39% (high N) for bermudagrass, and was 37% (low N) and 31% (high N) for crabgrass (Fig. 2).N addition decreased N min to a similar extent between rhizosphere soil and bulk soil (Table 1).Therefore, N addition did not affect the rhizosphere effect of bermudagrass on N min (P > 0.10, Fig. 2, Table 2), and only marginally inhibited the rhizosphere effect of crabgrass on N min (P = 0.08, Fig. 2, Table 2).
Microbial biomass carbon (MBC) was significantly higher in rhizosphere soil than in bulk soil for both species (P < 0.01, Table 2) and was not affected by N addition (P > 0.10, Table 2).As a result, the positive rhizosphere effect on MBC was not affected by N addition (P > 0.10, Fig. 2, Table 2).In contrast, microbial biomass nitrogen (MBN) was significantly lower in rhizosphere soil than in bulk soil for both plants (P < 0.01, Table 2) and was not affected by N addition.The negative rhizosphere effect on MBN was not affected by N addition (P > 0.10, Fig. 2, Table 2).
The three extracellular enzymes involved in microbial degradation of soil organic carbon (i.e.β-xylosidase for hemicellulose degradation, β-glucosidase, and β-cellobiohydrolase for cellulose degradation) showed similar responses to rhizosphere effect and N addition.The presence of plants had minor effect on activities of these three enzymes compared to the unplanted control treatment, except that crabgrass showed positive rhizosphere effect on β-glucosidase (Fig. 2, Table 2).N addition did not significantly impact activities of these three enzymes (Table 2), and there was no interaction between rhizosphere and N on these enzymes (Fig. 2, Table 2).
The three extracellular enzymes involved in microbial degradation of soil organic N (i.e.NAG for chitin degradation, LAP for protein degradation and peroxidase for lignin degradation) were significantly higher in the rhizosphere soil of both plants than in the unplanted control soil (P < 0.01, Fig. 2, Table 2).The positive rhizosphere effect was stronger for NAG (54-65%) than for LAP (18-28%) and peroxidase (20-39%, Fig. 2).In addition, N addition significantly inhibited LAP and peroxidase activity (P < 0.05), but it had no effect on NAG activity (P > 0.10, Table 2).For all three enzymes and both plant species, there was no interaction between rhizosphere and N (Table 2), suggesting that the positive rhizosphere effects on these three enzyme activities were not responsive to N addition (Fig. 2).

dISCuSSION
We had hypothesized that the presence of plants would enhance microbial biomass and enzyme activity and lead to faster soil C and N cycling in the rhizosphere, but the magnitude of rhizosphere effects would be reduced by N addition due to reduced belowground C supply.In fact, the results of this study showed that rhizosphere effects were mostly positive (except microbial biomass N) and significant (except C-degrading enzyme activities).N addition stimulated plant growth, suppressed net N mineralization and N-degrading enzyme activities, but had no significant impact on other soil variables.In contrast to our hypothesis, rhizosphere effects were not affected by the type, amount and duration of N addition (i.e.100-µg NH 4 NO 3 -N g dry soil −1 for 80 days) in this study.We also detected shifts in soil bacterial community  composition in the rhizosphere compared to bulk soil, but not with N addition.Taken together, these results suggested that rhizosphere effects of the two grasses on soil microbial biomass, activity and composition are minimally influenced by N addition.

Rhizosphere effects
In this study, we sampled rhizosphere soils in pots fully occupied by living roots of two grass species and bulk soils in unplanted control pots maintained at the same temperature and moisture levels.Many operationally defined methods have been used to sample a representative 'rhizosphere soil' versus a 'bulk soil' in order to study rhizosphere effect (Luster et al. 2009).For example, the commonly used methods include separating soil at difference distance from living roots (e.g.Cheng et al. 1996;Herman et al. 2006;Schenck zu Schweinsberg-Mickan et al. 2012), gentle shaking of freshly sampled soil attached (or not attached) to living roots (e.g.Phillips and Fahey 2006;Yin et al. 2012Yin et al. , 2014) ) and homogenizing soils from pots occupied by living roots or remained unplanted (Bardgett et al. 1999;Cheng et al. 2003;Zhu et al. 2014).We defined homogenized soils from the whole volume of planted pots after removing roots (but returning root-attached soils) as rhizosphere soils, which may dilute the intensity of rhizosphere effect as roots may not have fully occupied the soil volume.Therefore, the actual magnitude of rhizosphere effect in the "hotspot" may be stronger than reported here.
In general, the rhizosphere effects were significant and similar between the two grass species.Soil pH was 0.5-0.7 units higher in the rhizosphere compared to bulk soil, consistent with the results of some previous studies (e.g.Bardgett et al. 1999;Steer and Harris 2000).Root-induced changes in soil pH are affected by the relative uptake of anions (e.g.NO 3 − ) versus cations (e.g. ) by roots (Marschner et al. 1986).The lack of NO 3 − in the rhizosphere soil compared to bulk soil (0.6-0.8 versus 88.6 µg N g soil −1 ) suggests that the release of OH − or HCO 3 − during NO 3 − uptake and reduction may be one of the factors leading to the rise in rhizosphere pH (Steer and Harris 2000).Microbial biomass C was 7-10% higher in the rhizosphere in this study.Generally, microorganisms are more C limited in the bulk soil compared to the rhizosphere (Cheng et al. 1996).Therefore, root-derived carbon input can increase microbial growth and biomass in the rhizosphere as shown in many studies (e.g.Griffiths et al. 1999;Phillips and Fahey 2006).Microbial biomass N, however, was 13-17% lower in the rhizosphere.This result is consistent with some previous studies (e.g.Zhu and Cheng 2012), but not with others (e.g.Herman et al. 2006;Phillips and Fahey 2008;Schenck zu Schweinsberg-Mickan et al. 2012).The lower microbial biomass N and lack of mineral N (ammonium and nitrate) in the rhizosphere at the end of this experiment suggest that the grasses may have outcompeted the microorganisms for the N mineralized from soil organic matter.Although microorganisms often outcompete plants for mineral N in the short term (hours and days, Jackson et al. 1989), plants can compete effectively for N with microorganisms in the long term (weeks and months) because of longer turnover time of roots (Frank and Groffman 2009;Kuzyakov and Xu 2013).Moreover, the higher microbial biomass C:N ratio in the rhizosphere may imply that the community was more dominated by microbial groups with high C:N ratios (e.g.fungi, Strickland and Rousk 2010).Although we did not quantify the relative proportion of fungi versus bacteria in this study, a few recent studies have provided evidence for higher relative abundance of fungi versus bacteria in soils receiving high-substrate loading rate (Griffiths et al. 1999) or in rhizosphere soils (Ai et al. 2012;Bardgett et al. 1999).
The activities of three hydrolytic enzymes degrading soil hemi-cellulose and cellulose showed minor responses to rhizosphere effect (6-27%, mostly not significant).In contrast, many studies have reported higher activities of these enzymes in the rhizosphere of trees and non-woody plants (Ai et al. 2012;Kaiser et al. 2010), and these higher enzyme activities may be linked to higher rates of microbial decomposition of soil organic matter in the rhizosphere (so-called 'rhizosphere priming effect', Cheng et al. 2014;Kuzyakov 2002;Sun et al. 2014).Further studies by labeling plants with enriched or depleted 13 C-CO 2 (e.g.Zhu and Cheng 2011) may help to understand whether the living roots of these two grasses enhanced microbial decomposition of soil organic matter.
In contrast to the three C-degrading enzymes, the three enzymes for N acquisition were significantly higher in the rhizosphere.The presence of roots of both plants elevated NAG activity by 54-65% and LAP and peroxidase activities by 18-39%, respectively.These results are consistent with the idea that root C inputs induce microorganisms to produce N-degrading extracellular enzymes to mine N from N-rich soil organic matter, particularly when soil mineral N concentration is low (Fontaine et al. 2011;Kuzyakov 2002;Phillips and Fahey 2006;Zhu et al. 2014).We also found that potential N min was 31-42% higher in the rhizosphere compared to the bulk soil, generally consistent with the higher activities of N-acquisition enzymes.The positive rhizosphere effects on microbial N-acquisition enzymes (chitin-, protein-and lignin-degrading enzymes) and gross or net N mineralization have also been observed in studies that sampled rhizosphere soils attached to living roots (e.g.Herman et al. 2006;Phillips and Fahey 2006;Yin et al. 2012), or compared planted versus unplanted soils (e.g.Cheng 2009;Dijkstra et al. 2009;Zhu et al. 2014).
We also detected shifts in soil bacterial community composition in the rhizosphere compared to the bulk soil based on T-RFLP analysis.Changes in bacterial community composition between rhizosphere and bulk soil have been shown using many microbial fingerprinting techniques (e.g.DeAngelis et al. 2009;Marschner et al. 2004;Peiffer et al. 2013).The shifts in bacterial community composition between rhizosphere and bulk soil may be associated with the changes in soil pH (and many other factors such as the quantity and quality of substrates, predation pressure), which has been identified as an important controlling factor of soil bacterial and fungal community composition (Lauber et al. 2009;Rousk et al. 2010).Although our T-RFLP results cannot give phylogenetic information of specific bacterial groups that changed in the rhizosphere, the shifts in bacterial (DeAngelis et al. 2009) and potentially fungal (Mouhamadou et al. 2013) community composition may contribute to the changes in enzyme activities and N min .Further work should use more integrated metagenomic and functional analyses of soil microbial communities (e.g.Fierer et al. 2012;Zhou et al. 2012) to better link microbial community composition to function in the rhizosphere (Schimel and Schaeffer 2012).

N effects
N addition (100-µg NH 4 NO 3 -N g soil −1 ) increased soil N availability, and not surprisingly, stimulated plant biomass (13-43%) and tissue N concentration (16-93%), which was expected in this nutrient-limited (<2 µg g −1 extractable N and P) soils.In contrast to the increased plant biomass, many soil variables were not responsive to or even suppressed by N addition.Soil pH, microbial biomass C and N, activities of three C-degrading enzymes and an N-degrading enzyme (NAG) were not significantly impacted by N addition, whereas rate of potential net N mineralization (N min ) and activities of two N-degrading enzymes (LAP and peroxidase) were inhibited by N addition.The lack of impact of N addition on soil pH, microbial biomass and C-degrading enzyme activities is unexpected given that N fertilization often reduces microbial biomass (Phillips and Fahey 2007;Treseder 2008) and increases microbial enzymes degrading hemi-cellulose and cellulose (Keeler et al. 2009;Saiya-Cork et al. 2002).The relatively short duration of the experiment (~80 days) and medium amount of N addition (100-µg NH 4 NO 3 -N g soil −1 ) may prevent us from detecting statistically significant effects (Bardgett et al. 1999), and other nutrients (e.g.phosphorous) may be more limiting to these soil variables than N (Vitousek et al. 2010).Moreover, the suppression of N min and microbial N-acquisition enzyme activities by N addition is consistent with the optimization theory that microorganisms reduce allocation to N-acquisition enzymes when there is freely available N in the soil (Allison and Vitousek 2005).
Although rhizosphere effects were mostly significant and positive for the soil variables measured in this study, they were not significantly affected by N addition.A few studies have reported inconsistent results on the influence of N fertilization on rhizosphere effects in forest, grassland and agricultural soils.Phillips and Fahey (2008) showed that N (as NO 3 − , and together with other micronutrients P, K, Ca and Mg) fertilization had positive, neutral or negative impact on rhizosphere effects, depending on the specific tree species and soil variables.Ai et al. (2012) reported that long-term addition of inorganic N fertilizers to a wheat-maize rotation field reduced rhizosphere effects on most extracellular enzyme activities.Perveen et al. (2014) showed that atmospheric N deposition suppressed rhizosphere priming effects using the SYMPHONY model.Liljeroth et al. (1994) found lower rhizosphere effects of wheat and maize on soil organic matter decomposition with N addition, whereas Cheng et al. (2003) did not find significant impact of N-P-K addition on rhizosphere effects of wheat and soybean.The impact of N fertilization on rhizosphere effects can be difficult to compare among different studies because many factors such as plant species, soil types, the amount and duration of N addition, and the chemical composition of fertilizers (e.g.N versus N-P-K, NH 4 + versus NO 3 − versus urea) can all influence rhizosphere effects (Kuzyakov 2002).In this study, we added 100 µg of NH 4 NO 3 -N g soil −1 over an 80-day experimental period.Root biomass: total biomass ratio was not significantly affected by N addition, whereas both root biomass and total biomass were increased by N addition to a similar extent for each species (39-43% for bermudagrass and 13-18% for crabgrass).Although we did not measure root-derived carbon inputs to soil (e.g.root exudates) in this study, belowground carbon allocation may not be significantly affected by the amount and duration of N addition in this study, which did not lead to a significant shift in rhizosphere effect (Cheng et al. 2003;Phillips and Fahey 2008).
In addition to microbial biomass, enzyme activities and N min , we also measured soil bacterial community composition using T-RFLP technique.The presence of both grasses significantly shifted the bacterial community composition in the rhizosphere compared to unplanted bulk soil, whereas N fertilization had much lower impact on bacterial community composition.Although we could not quantify the rhizosphere effect for bacterial community composition as for other soil variables, it appears that N addition only had minor impact on soil bacterial community composition and its response to rhizosphere.Many studies have found shifts in soil bacterial and fungal community composition in response to N addition (e.g.Bardgett et al. 1999;Fierer et al. 2012;Frey et al. 2004;Marschner et al. 2004), but few studies (e.g.Mouhamadou et al. 2013;Zancarini et al. 2012) have compared the N addition impact on soil microbial community composition between the bulk soil and the rhizosphere.Zancarini et al. (2012) found that N addition changed bacterial community composition in the rhizosphere of Medicago truncatula (but not in the bulk soil), whereas it had no significant impact on fungal community composition.Mouhamadou et al. (2013) detected significant differences in fungal community composition between rhizosphere and bulk soils within two grassland species (Festuca paniculata and Dactylis glomerata), but not between two N fertilization treatments.Although we did not measure soil fungal community composition in this study, the short-term N fertilization many not produce significant impact on fungal community composition (Mouhamadou et al. 2013;Zancarini et al. 2012).

CONCLuSIONS
In conclusion, our data suggest that rhizospheres of smooth crabgrass and bermudagrass had significant impacts on soil pH, microbial biomass, extracellular enzyme activities, N min and bacterial community composition, whereas N fertilization suppressed N min and two microbial N-acquisition enzyme activities and had no significant impact on other variables.Moreover, the direction and magnitude of rhizosphere effect was not significantly influenced by the type, amount and duration of N addition in this study.Taken together, these results suggest that rhizospheres of the two grasses exert a more important control of microbial community composition and function than soil fertility.If these results can be further generalized to more plant-soil combinations and to field conditions, they should improve our understanding of how plant-microbe interactions influence biogeochemical cycling.

Figure 1 :
Figure1: total and root biomass (g pot −1 ), biomass-weighted total N content (mg g −1 ), root biomass: total biomass ratio and soil pH (mean ± standard error, n = 4) measured after final harvesting.In each treatment (unplanted control, bermudagrass or crabgrass), bars with different letters are significantly different at P < 0.05 between low (white bars) and high N (gray bars) levels.

Figure 3 :
Figure 3: principle component analysis of bacterial T-RFLP profiles (means ± standard error, n = 3 or 4) of bermudagrass rhizosphere soils (triangles), crabgrass rhizosphere soils (squares) and unplanted bulk soils (circles).Empty and filled symbols indicate low and high N treatments, respectively.The first and second principal components account for 48% of the variability.

FuNdING
United States Department of Agriculture National Institute of Food and Agriculture Hatch program (NYC-145403), the New York State Turfgrass Association and the US Department of Energy, Office of Science, Office of Biological and Environmental Research Terrestrial Ecosystem Science Program (DE-AC02-05CH11231).

Table 1 :
soil pH, N min (µg N g soil −1 day −1 ), MBC (µg C g soil −1 ), MBN (µg N g soil −1 ) and extracellular enzyme activities The five hydrolytic enzymes (nmol g soil −1 h −1 ) include BX, BG, CB, NAG and LAP.The oxidative enzyme is PER (µmol g soil −1 h −1 ).Values are the mean values of four replicates with standard errors in the brackets.

Table 2 :
two-way analysis of variance results of the main effects of rhizosphere (unplanted vs. planted), N (20 vs. 120 µg NH 4 NO 3 -N g soil −1 ) and their interaction on each of the soil variables