Arbuscular mycorrhizal fungi (AMF) are also an important microbial group in soil, since they can form symbiosis with most of the plants, contributing to plant health and nutrition. AMF is beneficial to tropical plants and presents potential influence on soil processes and plant diversity, increasing the interest For studying this group this group, especially in Amazon where little is known about them (Sturmer & Siqueira, 2010).

Most of AMF studies consist on identification of its spores from soil samples. Since AMF produce spores significantly bigger than the other fungi species, it is possible to separate them from soil samples by sieve and centrifugation in a sucrose gradient. Up to now, the studies in Brazilian Amazon were made using this approach (Leal et al., 2009; Mescolotti et al. 2010; Sturmer and Siqueira, 2010).

In Southwestern Amazon an AMF study compared three land uses: native vegetation, soybean fields and pastures, in two regions: Sinop (Forest) and Campo Verde (Cerrado), both in Mato Grosso State, Brazil (Mescolotti et al., 2010). Comparing Forest with Cerrado different patterns were observed. The largest amount of spores was found in soybean fields in the Forest region, and the number of spores was the same for the three land uses in the Cerrado region. Glomus spp. was the most common specie found (Fig. 8.).

Glomus sp1 Glomus sp2 Glomus sp3

In Western Amazon different AMF patterns were observed in different land uses (Sturmer & Siqueira, 2010). A total of 61 AMF morphotypes were recovered and 30% could not be classified as known species. Acaulospora and Glomus were the most common genera identified in the sites and higher AMF richness values were found in agriculture and pasture sites, than in the pristine areas. AMF patterns were also influenced by land use in a survey using different trap cultures in the same region (Leal et al., 2009). Among all trap plants and land uses, a higher number of spores were found in pasture and young secondary forest. In total 24 AMF species were recovered. Acaulospora spp. (10 species) was the most common genera followed by Glomus spp (5 species). Both studies showed that in Amazon soils the land use change from pristine vegetation to pasture and crops did not reduce the AMF diversity and probably new AMF species were found.

The two most common methods to measure the utilization of substrates by microorganisms are Biolog (Garland & Mills, 1991; Zak et al., 1994) and the respiratory response to addition of substrates, known as substrate induced respiration (SIR) (Degens & Harris, 1997; Degens et al., 2001). The authors claim that these techniques are sensitive enough to distinguish changes in the catabolic diversity that occur over short periods of time, as well as large differences that occur in the soil after a few years (Graham & Haynes, 2005). The main substrates used for SIR analysis are shown in Table 2. The diverse substrates are dissolved in 2 ml of solution for each equivalent of 1g dry soil and incubated in sealed bottles. The flow of CO2 for each sample is usually measured in an Infra-Red Gas Analyser (IRGA), after incubation of bottles for 4 hours at 25oC.

Few studies have been carried out in the Amazon region. Among these is the work of Mazzetto et al. 2008. This research evaluated the possibility to check whether there are

catabolic patterns in the Amazon soils under agricultural cultivation, native forest and pasture. A total of 60 areas were chosen distributed as: 20 native forest, 20 agricultural lands and 20 pasture sites in the regions of Mato Grosso and Rondonia, which are part of the Brazilian Amazon.

At first analyses were performed only in the native areas, which could be separated in Amazon rainforest, Cerrado and Cerradao. The low catabolic response obtained in the Cerrado soils may be linked to the frequent firing process that this biome suffers (Fig. 9). According to Arocena & Opio (2003), fire has a major impact on the physical (aggregate stability, clay content) and chemical (pH) soil properties, with significant influence on the microbial biomass. According to Hart (2005) fire alters the structure of microbial biomass, this being a selection factor in areas exposed to periodic events. Campbell et al. (2008) demonstrated in their studies that the use of carbonated substrates decreases with burning of area, suggesting a lower resistance/resilience of the microbial community. Among the substrates that can be influenced by burning of vegetation is arginine, which has a low response in Cerrado and Cerradao soils. The use of arginine in the microbial metabolism requires the presence of deaminase arginine enzyme, which is inhibited by fire.

Regarding the disturbed areas analysis were realized aiming to characterize the diversity of soil microbial biomass at these sites (Fig. 10), and to check the possible separation of the areas through multivariate statistical analysis (Fig. 11).

Soils under pasture had significant catabolic responses to amine and carbohydrate, and individually to the substrates glutamic acid, glutamine, glucose, mannose, serine and fumaric acid. In contrast soils under native vegetation had significant responses to malonic acid, malic acid and succinic acid. Soils under agriculture use did not show significant responses to any substrate examined, however they showed expressive responses to the aminoacids group, but not statistically different from the pasture soil (Fig. 10).

The canonical analysis showed that datasets related to CDP had great success in distinguishing the three land uses analyzed (Fig. 11). CV1 explained 67.5% of the variability observed, separating pastures from native areas and agriculture. Averages of native and agriculture areas were negative (-1.38 and -0.58, respectively) for CV1, while the average of pasture was positive (1.96). Asparagine, histidine and quinic acid with highly negative values were closely tied to native areas and agriculture, while glutamic acid and glucosamine had great representation in relation to pasture. CV2 explained 32.5% of the variability observed, separating native areas from agriculture and pastures. The average of native areas for the second axis was positive (1.34), while those of agriculture and pastures were negative (-1.02 and -0.32, respectively). The main substrates that provided this separation were serine and quinic acid, which showed negative values (linked to pasture and agriculture), and the tartaric acid, considered the more representative substrate related to native areas.

Among the major substrates involved, serine is documented as present in root exsudates (Bolton et al., 1992), quinic acid is a component of plant tissues (Gebre & Tchaplinski, 2002), and tartaric acid is one of main intermediary compounds of the Krebs cycle, in the basic metabolism of aerobic microorganisms (Tortora et al., 2005).

When only one ecoregion (Alto Xingu) was selected for analysis results of the CDP approach was even more significant (Fig. 12). CV1 explained 66.5% of the variability, separating native areas (-7.87 — negative score) of areas under agriculture and pasture (4.33 and 0.49 — positive scores, respectively). The main substrates involved in such axis were: succinic acid and malonic acid, with negative values. With positive values quinic acid and glucose also contributed to the separation observed. CV2 explained the remaining 33.5% of the variability, separating areas under pasture (4.84 — positive score) of native and agricultural areas (-2.04 and -2.65 — negative scores, respectively). Among the major substrates in this axis are highlighted asparagine and tartaric acid showing negative values, while lysine and pantothenic acid had positive values (Fig. 12).

Taking into account only data corresponding to the agricultural areas present in the database, we could distinguish areas under perennial crops, tillage and conventional tillage. By means of discriminant analysis the reallocation of data was performed in order to observe if datasets was homogeneous among the land uses analyzed. Data from areas under conventional tillage were relocated with 70% success, while data from conventional tillage and perennial cultivation showed higher percentage (98% and 100%, respectively). The same analysis was performed for pasture data that could be reallocated according to the following classification: typical pasture (100% success), improved pasture (95% success) and degraded pasture (91% success). This high percentage of reallocation of data shows that the microbial communities analyzed by CDP have high correlation with the use of land deployed. According to Mazzetto et al. 2008 the application of substrate induced respiration was efficient in distinguishing the land uses. The composition of microbial community revealed, through CDP approach, a close relationship with vegetation cover, regardless of climatic factors or the soil type.

As highlighted by Totola & Chaer (2002) and San Miguel et al. (2007), the importance of functional and catabolic diversity lies in the fact that only based on changes in the genetic diversity it is not possible infer whether some functions of soil were lost or not. The physiological profile of microbial community allows accessing the metabolic capacity of the microbial biomass as a whole, through tests realized with specific carbon sources defined in the laboratory.

2. Conclusion

Soil microbial diversity is still a difficult field to study, since 95-99% of organisms cannot be cultivated by culturing methodologies. The most popular techniques for soil microbial communities fingerprinting are DGGE and T-RFLP, which should be complemented by sequencing information to provide an overview of the study areas, especially those with high spatial variability that requires the collection of a high number of samples and replicates. New DNA and RNA sequencing provide high resolution information especially using depth sequencing of metagenomic samples.

Using DGGE, T-RFLP and other approaches, it has been clear that land use changes influenced significantly the diversity and structure of microbial communities in the Amazonian soils. Data available of DNA sequencing provided a high resolution view pointing changes of specific microbial groups and also the high quantities of unknown microorganisms. Catabolic diversity profile was efficient in distinguishing the land uses. The composition of microbial community revealed, through CDP approach, a close relationship with vegetation cover, regardless of climatic factors or the soil type.

Land use changes modify the genetic structure of microbial communities in the Amazonian soils, but they do not reduce the diversity in the areas affected by deforestation and conversion for pasture and crops, in comparison with the native areas. Also many new species are to be discovered in such areas.