| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Environmental Science Policy and Management, Univ. of California, Berkeley, CA 94720; E. Schwartz, current address: Northern Arizona State Univ., Flagstaff, AZ
* Corresponding author (skyhawk{at}nature.berkeley.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AM, arbuscular mycorrhizal • CFDE, chloroform fumigation-direct extraction • PCR, polymerase chain reaction • TRFLP, terminal restriction fragment length polymorphism
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The interactions of plant roots and microbes in the rhizosphere have been studied extensively because of their broad-ranging importance in nutrient availability, pathology, and soil C dynamics (Pinton et al., 2000). Plant roots exude large amounts (Minchin and Pate, 1973; Norton et al., 1990) and complex arrays of organic compounds into the nearby soil (Juma and McGill, 1986; Paul and Clark, 1996; Kennedy, 1998). Primarily in response to elevated C availability, bacterial, fungal, and protozoal numbers are generally higher in rhizosphere soil than in bulk soil (Kennedy, 1998). Some types of microbial activity have also been found to be higher in rhizosphere than bulk soil (Hojberg and Sorensen, 1993; Sorensen, 1997; Naseby and Lynch, 1997; Yang and Crowley, 2000).
The quality and quantity of root exudates vary temporally and spatially along the root (Klein et al., 1990; Newman, 1985; Jaeger et al., 1999; Bringhurst et al., 2001). We have mapped sugar and amino acid exudate patterns in soil adjacent to Avena barbata roots using engineered bacterial reporter gene systems (two different strains of Erwinia herbicola 299r) and found that sucrose/fructose availability was highest near the root tip and declined with distance from the tip (Jaeger et al., 1999).
Increased numbers of microorganisms in rhizosphere soil can represent a potentially labile stock of organic N near plant roots. There are several ways that N contained in microbial biomass can become available to plants. If the supply of labile C is high near young roots and declines substantially in older root sections, then C-limited heterotrophs in a mature rhizosphere would mineralize NH4 during catabolism of N-rich cell components (Myrold, 1998). Such a spatial pattern of C-availability along roots (high C availability near root tips and low C availability near mature roots) could in itself result in N mineralization. Alternatively, root-C enhancement of microbial numbers and activity may attract bacterivores, which on consumption of low C/N microbial biomass, release N as NH4 into the rhizosphere. Protozoa and other soil fauna excrete an estimated 30% of consumed bacterial N into the rhizosphere (Griffiths et al., 1992), where it is available for plant uptake (Elliott et al., 1984; Clarholm, 1985). Infection of rhizosphere bacteria by bacteriophage would also result in cell lysis and biomass N mineralization. Finally, rhizosphere soil is a zone of water potential fluctuation as a result of evapotranspiration during the day followed by re-equilibration with surrounding soil water during the night (Papendick and Campbell, 1975). Such relatively rapid fluctuations in soil water potential could also result in N mineralization from the rhizosphere microbial biomass as N-rich cellular materials are released during cell water potential equilibration with the surrounding soil solution (Bottner, 1985; Kieft et al., 1987; Halverson et al., 2000).
The current study was designed to quantify rates of gross N mineralization in rhizosphere soil and relate these rates to spatial patterns of root-microbial interaction. The experimental design allowed testing of the first two mechanisms discussed in the preceding paragraph: N mineralization resulting from C-starvation in older root zones and protozoal grazing of bacterial cells. We grew Avena barbata in microcosms designed for calibrated temporal and spatial access to roots and rhizosphere soil (here within 2 mm of root). Using this simplified system we have: (i) quantified gross N mineralization and nitrification rates along the root and in bulk soil using a micro–15N pool dilution technique; (ii) mapped the density of native bacteria in soil around the root using direct microscopic counts; (iii) characterized the total bacterial community through DNA-based techniques; and (iv) mapped the biomass of protozoal grazers around the root using a standard MPN technique. These data are analyzed to determine if rates of gross N transformation are directly impacted by the presence of roots and if so, how.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Mapping Native Bacteria and Protozoa
Roots were allowed to grow through the thin compartment for 9 to 13 d, until new roots were within 2 to 3 cm of the bottom of the thin compartment. After removal of the faceplate, two roots and the 2-mm radius of rhizosphere soil surrounding each root were excised. Each root was divided into four sections: 0 to 4, 4 to 8, 8 to 12, and 12 to 16 cm from the root tip. For each section, the soil from the two roots was removed by gentle agitation in a preweighed tube containing 0.2 M Trizma buffer. This process was replicated for a total of 10 microcosms (n = 10). Bulk soil control samples were taken from each microcosm at a distance >1.5 cm from any root. The samples were stored at 5°C or frozen at –80°C depending on analysis to be performed. A large (
16 g) soil sample was composited from multiple different areas in the microcosm, weighed, dried at 105°C to calculate the gravimetric soil water content for each microcosm. Soil samples were diluted to yield a 10-fold dilution series to 1:10–5 and used for assays of bacterial and protozoal numbers. Bacterial numbers were determined using the two-part stain BacLight Bacterial Viability Kit (Molecular Probes Inc., Eugene OR), and counted using epiflourescence microscopy to determine live and dead bacterial numbers. Serial dilutions of soil were made in phosphate buffered saline (0.14 M NaCl + 9 mM PO4), sonicated, stained, and viewed within 48 h. At least two slides were prepared per sample, one for each dilution, and ten fields of view were counted per slide. Total bacterial counts are a sum of live plus dead counts. Protozoal populations were determined by a most probable number method adapted from that of Ingham (1994). Protozoal biomass was then calculated based on the average biomass of the three major groups of protozoa (ciliates, amoebae, and flagellates) as described by Ingham (1994).
Community Analysis
DNA was extracted from five frozen rhizosphere soil samples, five frozen bulk soil samples and two control soil samples. The control samples consisted of DNA extracted from the soil before it was added to the growth chamber. DNA was extracted from all soil samples using the BIO 101 soil DNA extraction kit (Qbiogene, Carlsbad, CA) according the protocol provided by the manufacturer.
TRFLP patterns of the total bacterial community were constructed using the following PCR reaction: 10 ng of soil DNA, 0.2 µM of primers 6FAM- 27F (5' 6-FAM- AGAGTTTGATCCTGGCTCAG 3') and 1492R (5' TACGGYTACCTTGTTACGACTT 3'), 2.5 mM MgCl2, 1 x Taq buffer and 5 units of Taq polymerase enzyme combined in a total volume of 50µL. The PCR reaction was initiated with a hot start for 3 min at 92°C followed by 30 cycles of 30 s of 92°C, 30 s of 53°C, and 60 s of 72°C. The PCR reaction was cleaned with a QIAGEN PCR cleanup kit (Qiagen Inc., Valencia CA), digested with 10 units of MspI and analyzed on a ABI 3700 automatic DNA sequencing system (Applied Biosystems, Foster City CA).
Gross Nitrogen-Mineralization Rates
Nitrogen mineralization rates were determined using 15N pool dilution experiments. After new roots had colonized the thin compartments of the microcosms, the faceplates were removed from five microcosms at a time, and 60 mL of 15N label-containing solution was sprayed onto the soil surface of the five microcosms. The label solution was 400 mg N L–1 as (NH4)2SO4 at 29.7 atom% 15N, and was applied at a rate of 14 mg N kg–1 soil. This application rate was necessary to minimize the effect of the variability of background soil NH4, and the resulting concentration was comparable with a fertilization event. After labeling all microcosms, the faceplates were replaced and the microcosms returned to the growth chamber.
Since homogeneity of label application is critical in pool dilution experiments (Davidson et al., 1991), we tested the uniformity of label application within and among microcosms. We applied NH4 to the surfaces of five unplanted microcosms, and collected a total of five 3-g soil samples from each microcosm. We found no differences among microcosms (data not shown), although the top third of the microcosms were lower (P < 0.05) than the middle and bottom thirds (24 vs. 29 mg NH4–N kg–1). With respect to penetration of label applied to a soil surface, Murphy et al. (2003) suggest the soil depth be restricted to <2 cm; the 5-mm depth of the experimental compartment and the 2-mm radius of rhizosphere soil we harvest is well within that limit.
It has been reported that nonuniform exploitation of indigenous and applied N pools occurs in short-term pool dilutions (Watson et al., 2000), and it has therefore been suggested that at least 24 h should elapse after label application before initial sampling (Murphy et al., 2003). We performed pool dilution trials up to 31 h long, but found that beyond 7 h, standard deviations exceeding 5 atom% 15N were typical in rhizosphere soil. This rendered rate calculations uninterpretable. For this reason, and to avoid overestimation of gross rates due to nonlinear changes in pool sizes imparted by root assimilation, we used incubation times not exceeding 3 h. At 2.3 h, for example, sufficient label remained in the soil (Table 1) as to avoid problems of label exhaustion and excessively high variability.
|
Gross Nitrification Rates
In a separate experiment, gross rates of nitrification were determined using methods analogous to gross mineralization. However, K15NO3 was substituted for 15NH4, the incubation period was 6 h, and root sections were 8 cm long, rather than 4 cm. Rates of gross nitrification were calculated using standard isotope dilution equations (Hart et al., 1994).
Nitrification Potential
Soil nitrification potentials were assayed according to the method developed by Schmidt and Belser (1982) and modified by Hart et al. (1994). Roots were divided into 4-cm sections, and soil from two or three roots was composited for each replicate. Soil was washed from roots into buffered solution containing NH4 in phosphate buffer (Hart et al., 1994). Aliquots were taken from shaken slurry incubation over a 24-h period and analyzed for NO3 content using flow-injection analysis (Lachat QC8000 flow injection analyzer, Lachat Instruments, Milwaukee, WI). A regression of nitrate concentration against time was used to estimate potential nitrification rates.
Biomass Nitrogen
We estimated bulk soil microbial biomass N by chloroform fumigation-direct extraction (CFDE; Brookes et al., 1985) as well as from direct bacterial counts. For the CFDE method, we collected approximately 5 g oven-dry equivalent of bulk soil from each microcosm at the 3-h harvest, as well as from the planted-unlabeled microcosms, using half of each sample as the nonfumigated control.
We also estimated biomass N in individual root sections and bulk soil from direct cell counts of total (live + dead) bacteria. We assumed rod-shaped cells of 0.5-µm diameter and 1.5-µm length, a bacterial cell density of 1.1 mg mm–3, a solids content of 0.4, and a carbon content of 0.45 g g–1 (Paul and Clark, 1996). Then we multiplied biomass C per cell by the number of cells per gram of soil to estimate biomass C on a unit soil weight basis. Finally, we estimated biomass N assuming a C/N ratio of 6:1.
Statistical Analysis
For background NH4 pools, we tested differences between bulk and rhizosphere soil by pooling rhizosphere sections, and performing a Student's t test. We tested for differences in bacterial cell numbers and protozoa biomass using randomized block designs; rhizosphere sections and bulk soil were blocked by microcosm. We analyzed gross mineralization results by first pooling all four rhizosphere sections, but not bulk soil; we then tested differences between rhizosphere and bulk soil using a factorial design, with experiment (1 or 2) and soil source (rhizosphere or bulk) as main effects. Bacterial numbers, protozoa biomass, mineralization rates, and NH4 pools were log-normally distributed, and so were log10–transformed before analysis. Statistics were performed using Statistix 7 software (Analytical Software, Tallahassee, FL). Principal component analysis of TRFLP patterns was performed using JUMP software (SAS Institute Inc. Cary, NC) TRFLP fragments were binned if they were within one base-pair in size of each other. Only fragments with intensities higher than 0.5% of the total fluorescence were used in the analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Calculated values for bacterial biomass N in rhizosphere soil are shown in Table 2. The bacterial biomass N value for bulk soil (3.0 mg N kg–1) is about 25% of the value for microbial biomass N found by chloroform fumigation direct extraction [12.8 (± 1.0) mg N kg–1]. If the standing stock of microbial biomass N decreases over time (that is, along the root), then that change in biomass N should result in net N mineralization. The calculated values for bacterial biomass N shown in Table 2 can thus be used to estimate net rates of N mineralization that could result from the changes in the standing stock of microbial biomass N. The greatest decrease in bacterial biomass N (0.3 mg N kg–1) occurs between the 8- to 12-cm and 12- to 16-cm root sections. Under these experimental conditions, A. barbata roots were growing at an average of about 2 cm d–1 (data not shown). This translates to a net rate of N mineralization of 0.3 mg N kg–1/2 d = 0.15 mg N kg–1 d–1.
Alternatively, the flux of N through the microbial biomass can be calculated as the biomass N turnover time, by dividing the standing stock of biomass N by the gross rate of N mineralization. This approach assumes that most of the N mineralized comes through bacterial biomass N but does not necessarily originate from the microbial biomass. The calculated values for N-turnover are shown in Table 2. The turnover time for N in the microbial biomass in bulk soil is 3 d. The turnover time for microbial biomass N in rhizosphere soil is much shorter and very rapid, ranging from 0.3 to 0.6 d. To the extent that direct microscopic counts underestimate bacterial numbers and CFDE underestimates total microbial biomass, the actual turnover rates of the biomass in these soils would be slower than those calculated here.
Bacterial Community Analysis
The TRFLP patterns indicate there were few differences between the bacterial communities of the bulk soil and rhizosphere samples. Principle component analysis of the TRFLP patterns (Fig. 3
) showed no discernable differences in the first principle component (44% variability). However, six terminal fragments (out of the total 109 TRFLP peaks) constituted significantly different portions of the total TRFLP pool in rhizosphere vs. bulk soil as shown in Fig. 4
. Five new fragments appeared or increased significantly in rhizosphere soil while one fragment (71 bp) declined in rhizosphere soils.
|
|
|
Spatial Pattern of NH4 in Soil
The distribution of NH4 in the microcosm soil before the addition of labeled-N substrate is shown in Table 1. Before the spray application of 15N-label, average ammonium concentrations in the rhizosphere zones were slightly higher than in the bulk soil (p < 0.05). All ammonium concentrations were low (<1 mg N kg–1soil) before spraying on the labeled NH4 solution. During the first 30 min following label application, there was a substantial decline in the NH4 pool surrounding the 8- to 12- and 12- to 16-cm root sections; this did not occur in the 0- to 4- and 4- to 8-cm sections (data not shown). From 30 min to 2.3 h, the NH4 pool surrounding the 8- to 12- and 12- to 16-cm sections appeared to recover somewhat, while NH4 surrounding the 0- to 4- and 4- to 8-cm sections remained relatively unchanged (Table 1). These changes in pool size reflect simultaneous mineralization and NH4 consumption. While we have determined the gross rate of mineralization during the 0.5- to 2.3-h interval, we cannot parse out gross consumption into microbial and plant assimilation. It would be interesting to determine if short-term dynamics of root assimilation changed in response to the NH4 application.
Patterns of Gross Nitrogen Consumption
The 15N pool dilution method for determining N mineralization also yields values for total N-consumption from the soil (Fig. 5). Nitrogen-consumption measured by this method includes all possible fates of NH4 including nitrification, microbial assimilation, and plant uptake. In the area near the root tip (0–8 cm), the rates of nitrification (about 11 mg N kg–1 d–1) were similar in magnitude to the gross N consumption shown in Fig. 5. Although the 15NH4 label addition may stimulate root assimilation of NH4, gross nitrification rates were measured in a separate experiment using 15NO3, and so were unaffected by the NH4 addition. In the root zones 8 to 16 cm from the root tip, the high rates of gross N consumption (60 to 90 mg N kg–1d–1) exceeded even nitrification potential rates, and hence could not have resulted from nitrifier consumption of NH4. These high rates of consumption most likely reflected root uptake of NH4. It appears that rapid NH4 uptake from the root zones 8 to 16 cm from the root tip severely depressed the rates of gross nitrification by depleting soil solution NH4.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Increased rates of gross N mineralization demonstrated increased flux of N into the NH4 pool. Decreases in the standing stock of bacterial biomass N were too small to account for the increased rates, so bacterial biomass was not the ultimate source of the N being mineralized. Mobilization of organic N must have increased to enable the increased rates of N mineralization.
The direct microscopic counts used in this study only estimated bacteria; we have no data on fungal numbers or biomass. In these mesocosms however, the fact that the roots under study ranged from 2 to 12 d old and that the soil and sand had been recently mixed and added to the experimental compartments, likely minimized the importance of fungi. Annual grasses, including Avena barbata, form arbuscular mycorrhizal (AM) associations (Hawkes et al., 2006). It has been suggested that AM hyphae may accelerate rates of decomposition of N-containing organic materials by stimulating soil bacteria (Hodge et al., 2001); thus the impacts of AM associations on rhizosphere N mineralization are potentially quite important.
We found only small differences in bacterial community composition by TRFLP analysis. By its nature however, TRFLP analysis is not a particularly sensitive method for assessing subtle differences in microbial communities. The results of TRFLP analysis did show that several bacterial populations were enriched in the rhizosphere. Whether the changes in bacterial community found between rhizosphere and bulk soil can explain differences in N-dynamics in the rhizosphere is as yet unknown.
Rates of N mineralization are significantly enhanced in rhizosphere soil; yet potential rates of N uptake by roots are greater than the rates of N supplied by N mineralization in rhizosphere soils. While the total N-demand of the roots would likely be supplied by advection/diffusion processes in combination with mycorrhizal N supply to the plants, rapid uptake of N by roots likely depletes NH4 in soil near roots actively involved in N acquisition. Gross rates of nitrification were effectively zero in soil adjacent to root zones most rapidly consuming NH4. While nitrification potential rates indicated a functional nitrifying community all along the root, the measured potential rates varied by more than two-fold over the period of root growth. If nitrification potential is an accurate index of nitrifier number, then nitrifier numbers (as well as gross nitrification) were depressed relatively rapidly by root competition for NH4. Previous work by Norton and Firestone (1996) investigating the N dynamics associated with Ponderosa pine roots, also demonstrated significant interaction between root uptake and rhizosphere nitrification; however in that case, the nitrifiers held their own against root competition.
Avena barbata is a common species in California annual grasslands. In the shallow rooting zone of California annual grasslands (0–15 cm), root density is commonly very high (900 g m–2, Jackson et al., 1988). A relatively large proportion of soil would then be expected to be functioning as rhizosphere soil (within 2 mm of a root) as plant biomass approaches peak standing crop.
In current models of ecosystem N dynamics, plant roots function primarily as passive assimilators of mineral (and amino acid) N diffusing/advecting into the root zone (Schimel and Bennett, 2004). Nitrogen mineralization has been viewed as a steady state microbial process in which C-limited soil microbes mineralize organic matter and release excess inorganic N into the soil (Paul and Clark, 1996). Both plants and microbes are seen as passive recipients of products from each other (as in MIT models), rather than as strong interactors. Ecosystem models based on this view commonly determine plant N availability and N cycling rates simply on the basis of C/N ratio of decomposing organic material and soil temperature and moisture. The underlying assumption is that mechanisms of N mineralization and uptake are highly complex and not really important at the ecosystem level. There is mounting evidence however, that this description is inadequate. Numerous studies have found that plants assimilate more N during a growing season than is available from net N mineralization (Chapin et al., 1988; Nadelhoffer et al., 1991; Shaver and Chapin, 1991; Fisk et al., 1998).
Plants and free-living soil microorganisms appear to interact more strongly in the cycling of N than has generally been recognized. If root influences on N mineralization are significant, then net N mineralization incubations conducted in the absence of live roots underestimate the actual N cycling rates in terrestrial ecosystems. Measurements of net N mineralization in the absence of active plant roots are the basis of most estimates of N cycling in terrestrial ecosystems. Gross rates of N mineralization are not only substantially higher than net rates (Booth et al., 2005), but show the direct effects of living-root processes. Given the interaction of plant roots and soil microorganisms that we have demonstrated in rhizosphere N mineralization, it is not surprising that environmental perturbations that affect plant C balance and root exudation can have substantial effect on ecosystem N cycling (Hu et al., 2001).
|
ACKNOWLEDGMENTS |
|---|
Received for publication April 7, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Crop Science | |||
| Journal of Natural Resources and Life Sciences Education |
Vadose Zone Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
