6.1. Crops as a factor influencing denitrifiers
The rhizosphere is the volume of soil influenced by plant roots (Hiltner, 1904).
The growth and activity of the root system induce significant modifica- tions in the physicochemical and biological properties of the soil surrounding the roots, which correspond to the so-called rhizosphere effect. It is well known that the major factors regulating denitrification: carbon, oxygen, and NO3 can be modified in the rhizosphere of plants. Thus, carbon compounds, which can be used as electron donor by denitrifiers, are released by plants roots in the surrounding soil through rhizodeposition. The effect of plants on oxygen and NO3 concentration is more complex. Oxygen concentration can be lowered in the rhizosphere by respiration of the roots and microorgan- isms. On the other hand, consumption of water by plant roots increases soil gas exchange and oxygen concentration. Some plants, such as rice, also transport oxygen from the air down to the soil in water-saturated soil. Finally, when roots grow and penetrate the soil, they can modify soil compaction, which affects oxygen diffusion. Nitrate is used by both plants and microorgan- isms and the competition for NO3 is therefore high in the rhizosphere during the growing season. However, plants can also potentially provide NO3 for denitrification when organic matter present in root exudates is mineralized.
Moreover, during plant senescence and litter decomposition in fall and winter, nitrogen becomes bioavailable and can be denitrified. Overall, factors
regulating denitrification in the rhizosphere are strongly interwoven and the stimulating effect of root-derived carbon is only observed under nonlimiting concentrations of NO3 and oxygen. It is therefore not possible to state that plant roots always stimulate denitrification.
6.1.1. Effect of crops on the denitrification activity
Comparison of denitrification rates between planted and nonplanted soil in the field or in incubation experiment has been the most common approach to investigate the influence of crops on this process. Early reports showed enhanced denitrification rates in the rhizosphere compared with bulk soil (Smith and Tiedje, 1979a; Stefanson, 1972; Woldendorp, 1962). The key role of plant on denitrification has later been confirmed in several studies, although the mechanisms responsible for the higher denitrification rates are still not clear. Among the agricultural plants studied, barley (Hordum vulgare) has received the greatest attention so far. Klemedtssonet al. (1987) observed that denitrification rates in pots planted with barley increased with time along with increased root biomass. Stimulation of the denitrification rates in planted pots was 2–22 times compared with the unplanted pots. Similar results were reported by Hjberg et al. (1996) who observed an average NO3 reduction and denitrification rates in the rhizosphere of barley 1.8 times higher than in the bulk soil, with the most pronounced increase of 7 times. By using monoclonal antibodies against the copper nitrite reductase, Metzet al.(2003) clearly showed the presence of active enzymes in the rhizosphere of wheat.
Vinteret al.(1984) demonstrated that this increase of denitrification in the barley rhizosphere was positively correlated with soil NO3 concentration.
Their results showed that for fertilizer applied to barley at 30 kg N ha–1, the denitrification rate increased 2.5 times while a fivefold increase was observed in field plots receiving 150 kg N ha–1. These results were consistent with those of Mahmoodet al.(1997), who carried out a field experiment to examine the effect of maize plants on denitrification. At low soil NO3 levels (1–4mg N g–1 dry soil), the presence of maize plants resulted in a nearly 50% increase in denitrification, whereas at higher NO3 levels (7–19mg N g–1dry soil) the observed increase due to plants was 2.5 times. The combined effect of plant roots and NO3 concentration on denitrification was first pointed out by Smith and Tiedje (1979a). They found that denitrification was lower in planted than in unplanted soil when NO3 concentration was low (0.002 g NO3–N kg1dry soil), while at higher NO3 concentration (0.1 g NO3–N kg1dry soil) the presence of plants increased denitrification. Qianet al.(1997) also reported higher denitrification rates in the unplanted soil compared with planted soil at late maize growth stages when the amount of NO3 was limiting in the planted soil. These neutral or negative effects of plant roots on denitrification were attributed to NO3 depletion around the roots.
It has also been reported that the rhizosphere effect on denitrification was associated with air-filled porosity (Wollersheimet al., 1987). At a low moisture tension, Bakken (1988) observed a tenfold increase in the denitri- fication rate in the planted soil compared with the unplanted soil. At medium or high moisture tension, the plants had no or even a negative effect on denitrification. Similarly, Prade and Trolldenier (1988) reported that the rhizosphere effect on denitrification was confined to air-filled porosity lower than 10–12% (v/v). Thus, the lack of stimulation on deni- trification in the rhizosphere at nonlimiting NO3 concentrations reported by Haider et al.(1985) was attributed to a high air-filled porosity in both planted and unplanted pots.
Carbon, the third factor regulating denitrification, is probably responsi- ble for the stimulating effect of plants on denitrification activity. Several investigators have demonstrated the influence of different organic substrates on denitrification. Denitrification was correlated with soluble organic mat- ter (Bijay-Singh et al., 1988; Burford and Bremner, 1975; Cantazaro and Beauchamp, 1985; McCarty and Bremner, 1993) and easily mineralizable carbon (Bijay-Singh et al., 1988). The release of organic compounds by living roots can directly affect denitrification rates by providing an addi- tional source of electron donor, but also indirectly by increasing microbial activity, which lowers the oxygen concentration. This amount of carbon released by roots into the soil can be up to 20% of photosynthetically fixed carbon during the vegetation period (Hu¨tsch et al., 2002; Nguyen, 2003). The nature of the root-derived carbon is highly variable (mucilage, exudates, root cap cells, and so on). The mucilage is composed of high- molecular-weight polysaccharides, mainly arabinose, galactose, fucose, glucose, and xylose, and up to 6% is proteins. In contrast, exudates are low-molecular-weight compounds released passively from roots such as sugars, amino acids, and organic acids. As expected, daily addition of 70mg C g–1 dry soil of maize mucilage to an agricultural soil increased denitrification 2.8 times compared with water addition (Mounier et al., 2004). Similarly, daily addition at a rate of 150mg C g–1dry soil of different mixtures of amino acids, organic acids, and sugars mimicking maize root exudates greatly stimulated denitrification rates (Henry et al., unpublished data). In addition, several investigations have shown that denitrification rates were also positively related to the distribution of fresh plant residues in the soil profile (Aulakh et al., 1984, 1991; Cantazaro and Beauchamp, 1985; Christensen and Christensen, 1991; Parkin, 1987).
6.1.2. Effect of crop on the denitrifier community
In contrast to denitrification activity, there have been fewer studies of the effect of plant on the denitrifier community. Vintheret al. (1982) reported some early estimates of the diversity and the density of denitrifiers in agricultural soils under continuous barley cultivation. Counts of denitrifiers
performed using the most-probable-number method with NO3 agar broth as growth medium revealed densities ranging between 103and 106bacteria g1of dry soil, which represented less than 1% of total bacteria. In contrast, NO3 reducers counts for less than 10% of total viable count. Identification of denitrifying isolates based on selected physiological and morphological properties showed that numerically predominant denitrifiers belonged to Pseudomonasspp.,Alcaligenessp., andBacillussp. The effect of plant roots on the taxonomic diversity of denitrifiers has further been investigated by isolating denitrifiers from unplanted or maize planted soil in a 3-month incubation experiment (Che`neby et al., 2004). Density of denitrifiers was 2.4106and 1.6107cells g1of dry soil in the unplanted and planted soil, respectively. A total of 3240 NO3-reducing isolates were obtained and 188 of these isolates were identified as denitrifiers based on their ability to reduce at least 70% of the NO3 to N2O or N2. Comparison of the distribution of the denitrifying isolates between planted and unplanted soil showed a difference in the composition of the denitrifier community with an enrichment of phylogenetically Agrobacterium-related denitrifiers in the planted soil. In addition, these predominant Agrobacterium-related isolates from the rhizosphere soil were not able to reduce N2O while dominant isolates from the unplanted soil emit N2as end denitrification product.
Direct molecular approaches have recently been applied to investigate the effect of maize on NO3 reducers community performing the first step of the denitrification pathway. ThenarGgene encoding the membrane-bound NO3 reductase was used as molecular marker to analyze the composition of the NO3 reducers community from planted and unplanted pots after 3 months of repeated maize culture. A shift in the community composition between unplanted and planted soils was reported without significant mod- ification of the diversity indices (Philippot et al., 2002b). Clone library analysis revealed that most of the dominant sequences in the planted soil were related tonarGfrom the Actinomycetes suggesting a specific selection of NO3-reducing actinobacteria by the maize roots. In contrast, Che`neby et al.(2003) detected a reduction of the reciprocal Simpson’s diversity index in the maize planted soil compared with the unplanted soil, but without any major modification of the composition of the NO3-reducing community in another soil type. The results from these two studies suggest that the rhizosphere effect on the structure of the denitrifier community is strongly dependent on the soil type. Several studies aiming at sorting out the relative importance of plant and soil confirmed that these two factors might act simultaneously in determining the composition of the indigenous soil microbial community (Clays-Josserand et al., 1999; Costa et al., 2006;
Marschneret al., 2004; Wielandet al., 2001).
In two studies, effort has been devoted to disentangle the mechanism of the rhizosphere effect by investigating the influence of the two major rhizodeposits, mucilage and exudates, on the genetic structure of denitrifiers
(Henry et al., unpublished data; Mounier et al., 2004). Analysis of the structure of the denitrifier community by direct molecular approaches revealed only minor changes after mucilage amendment (Mounier et al., 2004). Similarly, the addition of sugar, amino acids, and organic acids mimicking maize exudates resulted in minor changes in the structure and the density of the denitrifier community (Henry et al., unpublished data).
Even though root-derived carbon can stimulate denitrification activity, it does not seem to be an important driver of the denitrifier community structure in soil. However, the community structure of the active members of the denitrifying community might be influenced by root exudates, but this has not yet been clarified.
6.1.3. Denitrification provides a selective advantage in the rhizosphere
Since most of denitrifiers are chemoheterotrophs, the increase of denitrifier density together with total microbial density observed in the rhizosphere was mainly attributed to the higher availability of organic substrates in the root vicinity. However, it has been suggested that the ability to grow by respiring nitrogenous compounds when oxygen is limited could be a selective advantage for denitrifiers in the rhizosphere. Thus, using DNA probes for the gene encoding the NO2 and N2O reductase, von Berg and Bothe (1992) found that the denitrifier to other heterotrophic organism ratio was increased near the roots. Such influence of plants on the distribu- tion of denitrifying abilities has also been reported by Clays-Josserandet al.
(1995), who observed that the proportion of denitrifying pseudomonas isolates gradually increased in the root vicinity of tomato. To demonstrate that this selection of denitrifiers in the rhizosphere was due to ability to respire nitrogenous and not to other traits, the competitive abilities of denitrifying strains in the rhizosphere have been compared with those of their isogenic nondenitrifying mutants. Mutants unable to synthesize either the membrane-bound NO3 reductase, the cd1 NO2 reductase, or the copper nitrite reductase were outcompeted by the denitrifying wild- type strains in the rhizosphere of maize demonstrating that denitrification itself could provide an advantage for root colonization (Ghiglione et al., 2000; Philippotet al., 1995).
6.2. Impact of crop species, crop cultivars, and transgenic plants
Because both shoot and root properties, for example, different litter types and roots architecture, and the amount and composition of root exudates are varying among plant species and cultivars (Hu¨tsch et al., 2002), it has been hypothesized that effect of plants on microorganisms differ depending on plant species or cultivars. Therefore, in the last decade many studies were
performed to prove this hypothesis. Most were based on 16S rRNA approaches, which make it impossible to relate any changes in the microbial community structure to functions. Only few attempts were made to use functional genes to measure possible impacts of crop species or cultivars on microorganisms involved in nitrogen cycling.
6.2.1. Rhizosphere effect on denitrification depends on crop species Effects of crop species or cultivars have mainly been investigated on deni- trification activity rather than on the diversity of denitrifiers. Crush (1998) reported a tendency for higher potential denitrification rates in association with bigger root mass in a lysimeters study with various forage plants.
Differences in the denitrification rates between small grains (barley, wheat, and oats) and grasses were also reported by Bakken (1988). Since legume plants associated with nitrogen-fixing bacteria can be used as substi- tute for mineral fertilizers, several authors studied whether their cultivation affect the nitrogen cycle processes. Using the C2H2inhibition technique on intact soil cores sampled during 2 years in a field, Svensson et al. (1991) reported significant differences between plant species with higher denitrifi- cation rates with lucerne (Medicago sativa L.) than with barley (Hordeum disticum) and grass ley (Festuca pretensis Huds.). Larger denitrification rates under legumes than other plants were also reported by other studies (Kilian and Werner, 1996; Scagliaet al., 1985). The higher positive effect of legume on denitrification rates was observed not only with living plants but also during their decomposition process. Aulakh et al. (1991) and McKenney et al. (1993) showed higher denitrification rates in soil amended with legume residues than in soil amended with grass, corn, or wheat residues.
However, lower denitirification rates were observed with clover than with small grains or grasses (Bakken, 1988).
It has been hypothesized that the higher denitrification rates caused by legumes could be due to their symbioses with denitrifying Rhizobiacaea.
Thus, several studies reported that denitrification was very common in rhizobia (Asakawa, 1993; Daniel et al., 1980, 1982; O’Hara and Daniel, 1985; Tiedje, 1988; van Berkum and Keyser, 1985; Zablotowicz et al., 1978) and that many strains can denitrify both as nodule bacteroids and in the free-living state (Arrese-Igoret al., 1992; Garcia-Plazaolaet al., 1995).
Accordingly, Kilian and Werner (1996) showed that mean denitrification was increased fourfold in plots of the nitrogen-fixing beanVicia albacom- pared with nonnodulated V. alba mutant. On the other hand, Garcia- Plazaola et al. (1993) suggested that even with optimal conditions for denitrification and the highest rhizobial populations found in agricultural soils, the contribution ofRhizobiacaeato the total denitrification was virtu- ally neglectable as compared with other soil microorganisms. The fact that different legume plants were analyzed may explain these contrasting results.
Since the symbiosis between rhizobia and legume plants is highly specific,
different rhizobial strains, which can exhibit contrasted denitrification abil- ities, are selected according to the legume species. This hypothesis is supported by the work of Sharma et al. (2005) who studied the diversity of transcripts of the NO2 reductase in the rhizosphere of three different legumes: Vicia faba, Lupinus albus, and Pisum sativum. A significant plant- dependent effect on the transcripts was observed, suggesting that the active denitrifiers were different in the rhizosphere of three legumes. The denitri- fier community structure, based on the DNA analysis ofnirKandnirSgenes, was not as variable between the different plant rhizospheres, indicating a stable denitrifier community. Similar results were also found by Deiglmayr et al. (2004). When investigating the effect ofLolium perenneand Trifolium repenson the NO3 reducer community, based on DNA analysis ofnarG, no plant species effect was observed. In contrast, with a similar approach, Patra et al.(2006) observed an effect of the plant species on both the structure and the activity of the denitrifier community among Arrhenatherum elatius, Dactylis glomerata, andHolcus lanatusin grasslands.
6.2.2. Impact of transgenic crops
Transgenic crops offer agronomic advantages, such as improved yield, improved product quality, herbicide tolerance, or insect resistance, over their corresponding nontransgenic wild-type cultivar. These modifications are mostly obtained by adding a gene in the genome of the parental wild- type crop via genetic manipulation. Plant genetic engineering can be bene- ficial when it improves agronomic features, but ethical concerns and the impact of genetically modified crops on human health and on the environ- ment is under debate. Therefore, quantitative risk assessments have been undertaken to determine the safety of transgenic plants. Such studies were performed on not only insects, earthworms, nematodes, and so on, but also on microorganisms, which dominate soil-borne communities. Like plant developmental stage or genotype can influence microbial diversity and activity in the rhizosphere (Rengelet al., 1998), introduction of a transgene might modify the plant effect on microorganisms, due to altered root rhizodeposition (Kowalchuket al., 2003). For example,Bacillus thuringiensis toxins (Bt) produced by transgenic plants are released in the soil by root exudates (Saxenaet al., 1999), which possibly affects the soil microorgan- isms. Indirect effects of transgenic crops on soil microbes could arise from repeated application of herbicide during cultivation of herbicide-resistant plants (Sessitschet al., 2004).
Most of the studies investigating effects of transgenic crops on soil microorganisms have focused on total bacteria (Baumgarte and Tebbe, 2005; Heuer et al., 2002; Lukow et al., 2000; Milling et al., 2004;
Schmalenberger and Tebbe, 2002). However, Philippotet al.(2006) com- pared the effect of glyphosate-tolerant maize, treated with either glyphosate or atrazine, and two cultivars of pyrale corn pest-resistant maize, treated
with atrazine, on the NO3-reducing community in a field experimented during 8 years. The nitrate reductase activity was higher in the rhizospheric soil than in the bulk soil, but no difference between the three cultivars was observed. A rhizosphere effect was also observed on the NO3-reducer community structure together with a strong influence of the sampling date, but the type of cultivar did not matter. Accordingly, analysis of the NO3-reducing community structure in the rhizosphere of five different cultivars of transgenic maize and the corresponding parental wild-type cultivars in a greenhouse experiment did not reveal any transgene effect (Sarret al., unpublished data).