RHIZOBIA AND OTHER LEGUME NODULE BACTERIA RICHNESS IN BRAZILIAN Araucaria angustifolia FOREST

The Araucaria Forest is a sub-type of the Atlantic Forest, dominated by Araucaria angustifolia, which is considered an endangered species. The understory has a high diversity of plant species, including several legumes. Many leguminous plants nodulate with rhizobia and fix atmospheric nitrogen, contributing to forest sustainability. This work aimed at bacteria isolation and phenotypic characterization from the root nodules of legumes occurring in Araucaria Forests, at Campos do Jordão State Park, Brazil. Nodule bacteria were isolated in YMA growth media and the obtained colonies were classified according to their growth characteristics (growth rate, color, extra cellular polysaccharide production and pH change of the medium). Data were analyzed by cluster and principal components analysis (PCA). From a total of eleven collected legume species, nine presented nodules, and this is the first report on nodulation of five of these legume species. Two hundred and twelve bacterial strains were isolated from the nodules, whose nodule shapes varied widely and there was a great phenotypic richness among isolates. This richness was found among legume species, individuals of the same species, different nodule shapes and even among isolates of the same nodule. These isolates could be classified into several groups, two up to six according to each legume, most of them different from the used growth standards Rhizobium tropici, Bradyrhizobium elkanii and Burkholderia sp. There is some evidence that these distinct groups may be related to the presence of Burkholderia spp. in the nodules of these legumes.


INTRODUCTION
The Araucaria Forest, which is characterized by the presence of the Brazilian Pine (Araucaria angustifolia Bert.O. Kuntze), a Gymnosperm, is distributed along the regions with subtropical climate, mainly in the southern states of Brazil.It is a sub-type of the Atlantic Forest, classified as ombrophilous mixed forest (IBGE, 1992).
Being a subtropical forest, quite different from other forests situated in tropical areas, it contains many endemic plant species, including legumes (Robim et al., 1990;Roderjan et al., 1998), which tolerate temperatures down to below 0°C in the coldest months.This specific vegetation and the climatic features may have also contributed to the development of a specific microbial community.A better understanding of this microbiota will be essential for a more adequate management of this type of forest, since Araucaria angustifolia has been considered an endangered species (Brasil, 1992).
Rhizobia are bacteria that can fix nitrogen in symbiosis with legumes, supplying part of the nitrogen input to the forests (Franco & Faria, 1997;Sprent, 2001).Rhizobia may be associated with other bacteria in legume nodules and all of them may play a special role in the ecological balance of Araucaria Forests.A detailed study may support more adequate management practices.
Campos do Jordão State Park has one of the last remnants of well preserved Araucaria Forests, including mature native forests, planted forests and secondary forests regenerating.Little research has been devoted to the survey of the floristic richness of these forest stands (Robim et al., 1990;Mattos & Mattos, 1982).Moreira et al. (2006) reported on mycorriza in the same araucaria forests, but nothing is known about occurrence of rhizobia or other nodule bacteria in this state park.
The objectives of this study were to search for legumes and nodules, to isolate bacteria from these nodule, to test the strain phenotypic richness, and to group the bacteria using phenotypic characteristics and multivariate approaches.

Sampling
Legume and nodule samples were collected in July 2006, in the Campos do Jordão State Park, in Campos do Jordão, SP, Brazil, located in the Mantiqueira mountain range (22º45' S and 45º30' W).The average altitude is 1450 m, although some regions are as high as 2000 m.The climate is classified as Cfb (Seibert et al., 1975), i.e., subtropical of altitude, mesothermal and humid, without a dry period.The mean annual precipitation is 2000 mm, which is relatively well distributed during the year.July is the coldest month, the absolute minimal temperature recorded being -7.2°C.The warmest month is February, with a mean temperature of 17.7°C (Seibert et al., 1975).
Collections were carried out in ecologically diverse sites where Araucaria angustifolia trees were present, legume species being randomly searched throughout the forest.When they were found, vegetative and, when available, reproductive plant samples were separated for further identification and root samples (containing nodules) were obtained by digging around the plant, taking care to check if the roots were connected to the plants.Along the survey, root and nodule samples were placed in plastic bags, and kept inside a thermo-box with ice until they were sent to the laboratory, where they were kept in a refrigerator at 4°C.The nodules were separated from the roots in the laboratory, where the isolation and phenotypic characterization of indigenous rhizobia was performed.Nodule shapes were recorded using a digital camera connected to a stereomicroscope.

Bacteria isolation and phenotypic characterization
The nodules were disinfected for thirty seconds in ethanol (95%) and for one minute in sodium hypochlorite (6%), adapted from Barrett & Parker (2006).They were then washed four times in sterilized water and finally crushed with a flame-sterilized glass rod.A loopful of the crushed nodule was then streaked across the surface of a Petri dish containing yeast mannitol agar (YMA; Vincent, 1970), and incubated at 28°C in the dark.Typical well-isolated colonies were re-isolated on diagnostic media, adapted from Odee et al. (1997): test tubes containing liquid YM (YMA without agar) with 25 mg kg -1 (w/v) brome thymol blue (BTB) as pH indicator; Petri dishes containing YMA with 25 mg kg -1 (w/v) congo red.
The bacteria in test tubes were incubated at 28°C on a shaker (220 rpm) and classified according to their ability to change the pH of the growth medium (alkaline, neutral or acid).
Petri dishes were incubated at 28°C in the dark, until the colonies were evaluated.They were characterized according to color (white, pink, translucid, yellow or white with a pink center), to the amount of extracellular polysaccharides (EPS) production (none to moderate or moderate to copious) and to colony size (the colony diameter measured with a ruler, after 3, 6 and 8 days of incubation).
Color and EPS were only evaluated when the colonies reached the minimum diameter of 3 mm, except for the very slow growing ones, which did not reach this diameter until the eighth day.

Legumes, nodules and isolated bacteria
Two Fabaceae-Faboideae and seven Fabaceae-Mimosoideae legume species containing nodules were found during sampling showing different nodule shapes.Two other Fabaceae-Mimosoideae legumes were collected, however without nodules (Table 1).Mattos & Mattos (1982) did not find any legumes in this same park.Robim et al. (1990) found nineteen legumes while studying the floristic composition of the Campos do Jordão State Park, although only eight of them were collected in habitats associated with Araucaria.Based on Sprent (2001), this is the first report of nodulation in Galactia crassifolia, Collaea speciosa, Mimosa furfuracea, Mimosa pilulifera and Mimosa filipetiola.
Nodule shapes were recorded for each legume species (Figure 1) and, therefore, it was decided to perform bacterial isolation from individual nodules.For each kind of nodule, two were randomly chosen (nodule A and B), for the isolation of bacteria.A third isolation was performed using all other nodules together (group C), and this made it possible to estimate the total richness occurring in the nodules from the same sample.Thus, the bacteria were isolated from nodules of different shapes, which were obtained from different individuals, and from different legume species, resulting in a total of 212 isolates.Some authors use plant taxonomy as a tool for the classification of legume nodules.It is necessary to know some of the developmental history of the nodule to use this criterion (Sprent, 2001).Different shapes may be related to different developmental phases of the nodule.In our study, there was no possibility to find out nodule phenology, and therefore it would not be adequate to try to include our findings in this system.Although accepting the idea that nodule shape is related to plant taxonomy, it was intended to test the hypothesis that the kind of the bacterial isolate might also influence nodule shape.Nevertheless, so far, no correlation was found between the phenotypic classification of the bacteria and nodule shape.

Phenotypic bacterial richness
A great phenotypic bacterial richness was found, and different legumes showed different percentages of strain phenotypes in the nodule bacteria community (Table 2).Cluster analysis was performed for all isolated bacteria together, resulting in six groups, separated by each plant (Table 3).The     *Isolated bacteria codes correspond to the initials of the name of the legume species (Gc -Galactia crassifolia, Cs -Collaea speciosa, Ad -Acacia dealbata, Mf -Mimosa furfuracea, Mp -Mimosa pilulifera, Ms -Mimosa scabrella, Mi -Mimosa filipetiola, Md -Mimosa dolens, M1 -Mimosa sp.1); number of the individual legume; nodule shape (1 -spherical, 2 -small rod, 3 -large rod, 4 -amorphous, 5 -fan-shaped, 6 -agglomerate and 7 -coralloid); nodule extraction (A -strain obtained from one single nodule, B -strain obtained from one single nodule different from "A", C -strain obtained from all the other nodules crushed together), and storage number (the last three numbers).
p s growth standards Bradyrhizobium elkanii, Rhizobium tropici, and Burkholderia sp. are included in groups III, V and VI, respectively, while three groups are quite different from these standards.
PCA analysis was performed for all bacterial isolates and all leguminous species together, but the results are presented separately for each legume species (Figure 2).The interpretation for the axes of the figures is as follows: PCR1 is a combination of "growth" (0.59), "pH change" (0.55) and "color" (0.49), while PCR2 is predominantly formed by "EPS" (0.87).PCR1 explains 47.63% and PCR2 explains 24.2% of the total variation.The variables, represented as vectors, which stand for "growth rate" and "pH reaction" were closely related to each other in the PCA.This had been expected, since, in general, slow growth is related to alkali production and fast growth is related to acid producers (Odee et. al., 1997).Based on the cluster analysis the groups of bacteria were separated for each legume in the PCA.Since the cluster analysis did not provide characteristics about the groups formed, we used this information together with the PCA analysis (Figure 2) to characterize the predominance of bacterial phenotypes in each group.Group I may be characterized by the predominance of bacteria of slow growth, alkali production, white colonies with pink center, and none to moderate EPS production.Group II has similar characteristics of the previous group, except for the intermediate growth rate and moderate to copious EPS production of its predominant bacteria.Group III has predominance of bacteria with a slow growth rate, alkali producers and white colonies.Group IV has the predominance of bacteria with an intermediate growth, and neutral or alkaline pH reaction.Group V presents predominance of fast growing bacteria, acid producers, with translucid colonies.In group VI there is predominance of bacteria with a fast to intermediate growth rate, with acid or neutral reaction, and white colonies.Silva & Döbereiner (1982) reported the presence or absence of nodules in Brazilian legumes, including some samples of legumes growing in Araucaria Forests.The isolated bacteria were classified as Rhizobium sp., but the bacterial richness or diversity was not studied.Similar studies report richness of rhizobia obtained from several legumes from Kenya, where three phenotypic groups were formed, which the authors called "phena", and from Libyan forests, where five distinct clusters were formed (Odee et al., 1997;Mohamed et al., 2000).

Phenotypic bacterial richness in each legume
The richness varied from two groups in Mimosa dolens to six groups in Galactia crassifolia, Mimosa sp.1 and M. furfuracea (Table 3).In general, each legume harbored more than one bacterial group in its nodules.Also, different individuals of the same species generally presented different bacterial communities.In the same manner, most of the nodules were occupied by more than one bacterial group.Using trap plants, Melloni et al. (2006) found a high diversity of rhizobia in Brazilian areas associated to the presence of Mimosa scabrella.In their study six groups were characterized using Phaseolus vulgaris L. as trap plant and eight groups using Vigna unguiculata (L.) Walp.
Since data here shown indicated a great richness of bacterial isolates, many of them quite different from the standards (Rhizobium tropici, Bradyrhizobium elkanii and Burkholderia sp.), we partially sequenced the 16S rRNA gene of some of the strains (unpublished data).Based on this technique it was possible to find out that isolates often present great similarity to Burkholderia, when submitted to the BLAST software in GenBank (http://www.ncbi.nlm.nih.gov/BLAST).Moulin et al. (2001) isolated Burkholderia spp.from legume nodules and recently it was demonstrated that some Burkholderia species form nodules with Mimosa spp., including M. scabrella and other tropical species from Brazil (Chen et. al., 2005;Elliott et al., 2007).In a similar study, Barrett & Parker (2006) suggested that Burkholderia spp.and other βproteobacterial species are more ubiquitous as root nodule symbionts than previously believed.Thus, Burkholderia spp.might also be present in legume nodules from the Araucaria Forest.

CONCLUSIONS
For the first time, nodulation of five legumes (Galactia crassifolia, Colllaea speciosa, Mimosa furfuracea, M. pilulifera and M. filipetiola) is reported.A high phenotypic richness of nodule shapes and bacterial isolates was found in leguminous plants that grow in an Araucaria forest in the Campos do Jordão State Park, Brazil.Several of the isolated bacteria are significantly different in phenotypic characteristics from the bacterial standards.

Table 1 -
Legumes with different growth habits collected in ecologically diverse sites in the Campos do Jordão State Park and their nodule shapes.