Genome Features of the Endophytic Actinobacterium Micromonospora lupini Strain Lupac 08: On the Process of Adaptation to an Endophytic Life Style?

Endophytic microorganisms live inside plants for at least part of their life cycle. According to their life strategies, bacterial endophytes can be classified as “obligate” or “facultative”. Reports that members of the genus Micromonospora, Gram-positive Actinobacteria, are normal occupants of nitrogen-fixing nodules has opened up a question as to what is the ecological role of these bacteria in interactions with nitrogen-fixing plants and whether it is in a process of adaptation from a terrestrial to a facultative endophytic life. The aim of this work was to analyse the genome sequence of Micromonospora lupini Lupac 08 isolated from a nitrogen fixing nodule of the legume Lupinus angustifolius and to identify genomic traits that provide information on this new plant-microbe interaction. The genome of M. lupini contains a diverse array of genes that may help its survival in soil or in plant tissues, while the high number of putative plant degrading enzyme genes identified is quite surprising since this bacterium is not considered a plant-pathogen. Functionality of several of these genes was demonstrated in vitro, showing that Lupac 08 degraded carboxymethylcellulose, starch and xylan. In addition, the production of chitinases detected in vitro, indicates that strain Lupac 08 may also confer protection to the plant. Micromonospora species appears as new candidates in plant-microbe interactions with an important potential in agriculture and biotechnology. The current data strongly suggests that a beneficial effect is produced on the host-plant.


Background
For a long time, it was considered that a healthy plant was a plant without microbes within its tissues. However, this view has started to change with new approaches to allow strains to grow for a longer time upon isolation as well as the use of NGS, which has permitted the identification of several strains present in the tissues of healthy plants, in particular several actinobacteria [1,2].
Endophytic microorganisms live inside plants for at least part of their life cycle. According to their life strategies, bacterial endophytes can be classified as ''obligate'' or ''facultative''. Obligate endophytes are strictly dependent on the host plant for their growth and survival while facultative endophytes have a stage in their life cycle during which they exist outside host plants [3]. These endophytes originate from soil, initially infecting the host plant by colonizing, for instance, the cracks formed at points of emergence of lateral roots from where they quickly spread to the intercellular spaces in the root [4]. Thus, a series of environmental and genetic factors is presumed to have a role in enabling a specific bacterium to become endophytic [5]. Conversely, Marchetti and co-workers [6] recently showed how a pathogen can evolve in a few generations to become a symbiotic endophyte by losing specific transporters and regulators linked to pathogenesis.
Micromonospora is a genus of Gram-positive Actinobacteria that was first isolated from soil [7]. This bacterium has received a lot of attention during natural product screening programs, given its ability to produce a very rich array of secondary metabolites [8,9,10]. The distribution of members of Micromonospora is wideranging since these bacteria have been isolated from different geographical zones. In addition, its habitats are also diverse and include: soil, freshwater and marine sediments, mangrove soils, rocks, and nitrogen fixing nodules of both leguminous and actinorhizal plants [11,12,13]. The recent report [13] that Micromonospora inhabits nitrogen-fixing nodules in a systematic way, has opened up a question as to what is the potential ecological role of this bacterium in the plant and whether this bacterium is in a process of adaptation from a terrestrial to a facultative endophytic life style.
Taxonomically, Micromonospora belongs to the family Micromonosporaceae which currently contains 27 genera and includes aerobic, non-acid fast and mesophilic microorganisms. Many strains produce mycelial carotenoid pigments giving the colonies an orange to red appearance, but blue-green, brown or purple pigmented strains have also been isolated. The family Micromonosporaceae also harbors the genus Salinispora, which is widely distributed in tropical and sub-tropical marine sediments. This taxon was described as the first marine actinomycete given its inability to grow in low salinity medium. Indeed, genomic information obtained from the genomes of Salinispora tropica and Salinispora arenicola provide evidence of marine adaptation of Salinispora species [14]. Thus, it appears that Salinispora evolved from a terrestrial environment to a marine habitat. In the case of some Micromonospora lineages, the question is whether this bacterium has followed a comparable evolution process, changing from a terrestrial to an endophytic lifestyle.
Further examples of closely related actinobacteria with different lifestyles reflected in their genomes include, among others, the genera Frankia, Mycobacterium and Streptomyces. In the case of Frankia, comparative genomic analysis of three representative strains, differing by less than 2% in their 16S rRNA genes revealed significant differences in their genome sizes (5.4-9.0 Mb) suggesting that these differences (e.g. gene deletion, acquisition and duplication, etc.) reflect their rapid adaptation to contrasted host plants and to their environments [15]. Similarly, several mycobacterial genomes were analyzed both at the nucleotide and protein levels. One of the most striking features was lipid metabolism genes with marked expansions of the number of genes related to saturated fatty acid metabolism in the pathogenic mycobacteria compared to the soil-dwelling strains [16].
In an effort to identify the genomic traits which make possible adaptation from a soil dwelling way of life to an endophytic habitat, the aim of this work was to present the genome sequence analysis of a representative strain, Micromonospora lupini Lupac 08, isolated from a nitrogen fixing nodule of the legume Lupinus angustifolius. This strain is part of a collection of more than 2000 strains isolated from nitrogen fixing root nodules of diverse legume [17,18] and actinorhizal species [13]. Strain Lupac 08 was selected as it showed good plant growth promotion, was used previously for in situ localization studies in planta [11] and produced several new secondary metabolites [9,10]. The results presented here show that the genome of M. lupini Lupac 08 contains a diverse array of genes that may help its survival in soils or in plant tissues, while the high number of putative plant degrading enzyme genes identified in its genome is quite surprising since this bacterium is not considered a plant-pathogen and may instead reflect their ability to bind to plant structural compounds.

Results
Phylogenetic position of M. lupini Lupac 08 and general genome features The phylogenetic position based on 16S rRNA gene sequence analysis of strain Lupac 08 with respect to currently described Micromonospora species and other members of the family Micromonosporaceae is presented in Figure 1. Those strains associated with plant/rhizosphere sources are highlighted. Strain Lupac 08 was clearly positioned within the genus Micromonospora and forms a subgroup together with the species Micromonospora saelicesensis, Micromonospora zamorensis and Micromonospora chokoriensis. These strains were isolated from a nitrogen fixing nodule, the rhizosphere of a Pisum sativum plant and a sandy soil, respectively. Nevertheless, a clear picture based on the habitat cannot emerge from this analysis.
M. lupini Lupac 08 was shown to have a circular chromosome of 7,327,024 bp with a GC content of 71.96% and no plasmid. A total of 7158 genomic objects were identified: 7,054 proteincoding, 10 rRNAs, 77 tRNAs, and 12 miscRNAs genes. The average gene length was 964 bp with an average intergenic distance of 126 bp. After manual validation of the automatic annotation, 61.5% (4338 CDSs) of the genes were assigned a biological function while 38.5% were registered as open reading frames (ORFs) with an unknown function. Based on the G+C skew analysis and position of dnaA, the probable origin of replication (oriC), was mapped close to the ribosomal protein rpmH. A circular representation of the M. lupini chromosome is provided in Figure 2 indicating some of the features described above.
The genomic characteristics of strain Lupac 08 and three additional Micromonospora genomes deposited in the public databases including Micromonospora sp. strain L5 isolated from root nodules of Casuarina equisetifolia [19]; M. aurantiaca ATCC 27029 T and Micromonospora sp. ATCC 39149 isolated from soil (Table 1) were compared. An important difference between the four strains was the number of tRNAs identified. M. lupini 08 contained by far the highest number with 77 tRNAs while the other strains had between 51 and 53. At present, M. lupini Lupac 08 contains one of the largest numbers of tRNAs reported among the actinobacteria sequenced. The number of rRNA and tRNA genes in a genome can be seen as an indication of positive selection. A high number of rRNA genes increases ribosome synthesis, which in turn increases the protein synthesis rate [20] and growth rate [21].
The COG distribution of M. lupini was similar to that observed in other bacteria in the family Micromonosporaceae, however various differences were detected such as the abundance of genes related to carbohydrate transport and metabolism. Among the Micromonospora genomes currently available, M. lupini Lupac 08 contained the highest percentage of genes (9.7%, 685) related to this category, followed by Micromonospora sp. L5 (8.9%, 598) and M. aurantiaca ATCC 27029 (8.5%, 576). The gene contents (in the same COG category) of other bacterial genomes classified in the family Micromonosporaceae were lower as in the case of S. tropica CNS-205 (7.4%, 391) and S. arenicola CNH-643 (6.4%, 374) two obligate marine actinomycetes. On the other hand, the overall COG profiles of Verrucosispora maris AB-18-032 T (genome size 6.7 Mb) and M. lupini Lupac 08 were very similar and no clear differences were found. Although V. maris was isolated from a sea sediment, it does not require sea salts for growth and it is not considered an obligate marine microorganism unlike S. tropica and S. arenicola. Thus, its metabolism suggests a terrestrial life style. Micromonosporae are well known for their ability to degrade complex polysaccharides such as cellulose, chitin and lignin [22,23]. In particular, cellulose is frequently utilized as a carbon source [24,25]. Therefore the abundance of these genes in the genome of strain Lupac 08, at first glance may not seem surprising, however, the value of 9.7% is comparable to that of highly active cellulolytic microorganisms such as Cellulomonas Figure 2. Circular representation of Micromonospora lupini Lupac 08. Circles displayed from the outside in: 1. Cellulose-binding genes in black, chitin-binding genes in red, lectin genes in lavender blue; 2. Genome coordinates; 3. MW; 4. GC% (linear range between 65 and 80%); 5. Regions of genome plasticity according to the RGP_Finder method (Mage platform) based on synteny breaks between the query genome (Lupac 08) and close genomes (Micromonospora aurantiaca ATCC 27029 T , Micromonospora sp. L5 and Verrucosispora maris AB-18-032 T ) correlated with HGT features (tRNA hotspot, DNA repeats, mobility genes), and compositional bias and GC deviation computation. C1 to C15 indicate the position of the 15 clusters of genes coding for secondary metabolites of Table 4 flavigena 134 T (9.5%) and Thermobifida fusca XY (7.9%), which are abundant in cellulose enriched environments such as soil, or plant tissues. Synteny. The genome sequence of strain Lupac 08 was aligned with those of Micromonospora sp. L5, M. aurantiaca ATCC 27029 T and Micromonospora sp. ATCC 39149 T (Fig. 3). Although the four genomes share a significant amount of genetic characteristics, they have undergone various inversions and translocations and M. lupini Lupac 08 contains the highest number of non-conserved regions. In addition, this alignment shows a high homology between strains Micromonospora sp. L5 and M. aurantiaca ATCC 27029 T confirming their close phylogenetic relationship as suggested by 16S rRNA gene phylogeny (Fig. 1); nevertheless, strain L5 shows a large inversion event. Thus, although the four Micromonospora genomes share many common features, it is also evident that M. lupini contains unique genomic regions as compared to M. aurantiaca ATCC 27029 T or Micromonospora sp. L5.

Diversity of Micromonosporae: core vs. flexible gene pool
Using the Micromonospora genomes of strains M. lupini Lupac 08, M. aurantiaca ATCC 27029 T and that of Micromonospora sp. L5 available in the NCBI [19], the core genome was calculated using the SiLix software [26]. The core genome was composed of 2294 CDSs, which correspond to approximately 32% of the predicted proteome. In addition, M. lupini Lupac 08 contained the highest number of strain specific CDSs, 4702 (66.6%), which is a very high value when compared to Micromonospora sp. L5 and M. aurantiaca ATCC 27029 T (13-14%, Figure 4), which both share a high gene similarity (85-86%).
Horizontal gene transfer is universally recognized as an efficient mechanism for microorganisms to acquire functions that enable them to adapt to environments with different selective pressures. Therefore insertion elements, transposases, integrated phages, and plasmids can be related to the plasticity of a genome. Strain Lupac 08 contained 49 CDSs (0.7%, of total CDSs) related to gene exchange including eight integrases and eleven recombinases. Except for seven CDSs, most of these genes were grouped into 20 clusters. Interestingly, eight of these mobile element clusters were found near genes related to carbohydrate transport and metabolism.
Metabolic Features. A metabolic pathway reconstruction was performed between the genome of strain Lupac 08 and 20 additional strains among which plant pathogens, symbiotic and saprophytic bacteria were included. The distribution and grouping of the microorganisms analyzed using 798 metabolic routes are presented in Figure 5. A good correlation was obtained between the microorganisms, their life style and phylogeny. Two main groups were obtained, the proteobacteria and actinobacteria. Within the actinobacteria, three clusters were clearly identified: the first one contained strains that belonged to the family Micromonosporaceae, the second cluster corresponded to various streptomycetes and the third cluster included the three Frankia genomes. Surprisingly, Micromonospora lupini Lupac 08 showed a closer metabolic relationship with the three Frankia strains (ACN14a, CcI3 and EAN1pec) than with the other two Micromonospora genomes.

Plant/Soil-associated life style
Transport systems. Organisms living in endophytic associations need to share resources with their host. Membrane transport systems play essential roles in cellular metabolism and activities. Current data suggest a correlation of transporter profiles to both evolutionary history and the overall physiology and lifestyles of organisms [27].
A total of 631 CDSs were located in the genome of M. lupini coding for a large diversity of transporters, representing approximately 8.9% of the genome. The majority of CDSs were related   to ATP-binding (dependent) transporters of which 362 corresponded to ABC transporters; the next most abundant (215 CDSs) coded for secondary transporters, with 105 classified in the Major Facilitator Superfamily -MFS-); 17 transporters belonged to ion channels and 20 were unclassified. The number of transporters determined in M. aurantiaca ATCC 27029 T and Micromonospora sp. L5 were lower with 575 and 587, respectively ( Table 2).
The number of transporters identified in the genome of M. lupini Lupac 08 is correlated with those in other bacteria with a plant/soil associated life style, which requires an efficient nutrient uptake system to obtain nutrients produced by the host plant, in addition to those found in the rhizosphere and the soil (Table 2) [27,28]. However, the number of transporters identified in strain Lupac 08 was lower than those present in other bacteria such as Bradyrhizobium japonicum USDA6 T (1138, 11.8%), Mesorhizobium loti MAF303099 (968, 14.2%), Sinorhizobium meliloti Sm1021 (1024, 16.4%) and Rhizobium leguminosarium bv. trifolii WSM 3125 (1087, 15.5%), which form a very close interaction with legumes. Nevertheless, the overall distribution (types) and the percentages of these values were similar. An additional difference was the absence of phosphotransferase system transporters (PTS) in M. lupini as compared to the strains mentioned above and other soil/plant bacteria included in Table 2. On the other hand, the overall profile of M. lupini Lupac 08 was very similar to those of Frankia sp. ACN14a and Frankia sp. CcI3 which also lack a PTS system.
Secretion systems. Secreted proteins play a number of essential roles in bacteria, including the colonization of niches and host-pathogen interactions. In Gram-positive bacteria, the majority of proteins are exported out of the cytosol by the conserved Sec translocase system or, alternately, by the twinarginine translocation system. In addition, a unique protein export system, the type VII or ESX secretion system also exists in some Gram positive bacteria [29].
The genome of M. lupini Lupac 08 encodes for 537 (7.6%) secreted proteins including several protein secretion systems ( Table 3). All genes related to the Sec-dependent pathway were located and included the SecY and SecE proteins which form the membrane channel and interact with the cytoplasmic membrane protein SecG; the auxiliary proteins SecD, YajC and the ATPase SecA. In addition, the heterodimer Ffh-FtsY (MiLup08_41486 and Milup08_41460) was also present. As in other Gram-positive bacteria, M. lupini Lupac 08 lacks homologs of SecB, the chaperone that targets proteins to the Sec translocon for passage through the cytoplasmic membrane [30].
Genes related to the Sec-independent twin-arginine translocation pathyway (TAT), which exports prefolded proteins across the cytoplasmic membrane using the transmembrane proton gradient as the main driving force for translocation were also located in strain Lupac 08 ( Table 3). Homologs of TatA and TatC were identified, however no homolog for TatB was found. Similar to other actinobacteria (e.g. Frankia sp. ACN14a) the tatA gene was found next to tatC. Only an ORF encoding TatC was located in the genomes of Micromonospora sp. L5 and M. aurantiaca ATCC 27029 T while no copies of tatA or tatB were found.
A set of fifteen genes identified as part of the type VII secretion system were located in M. lupini Lupac 08 (Table 3). These are arranged in three different clusters and included the essential proteins for secretion EccC, EccD, EsxA and EsxB [31]. The first cluster contains eight genes: eccC, esxA, esxB, eccD, eccB, eccE and two copies of mycP, a subtilisin-like serine protease which also appears essential but the function of which is not yet known [32]. The second cluster includes a copy of esxA (MiLup08_40381), esxB (MiLup08_40380) and mycP, annotated as S8 S53 subtilin kexin sedolisin (MiLup08_40382). Finally a third cluster contains the genes eccC (MiLup08_46744), eccD (MiLup08_46743) and mycP (MiLup08_46745).
Gram-negative bacteria use the type II secretion system to transport a large number of secreted proteins from the periplasmic space into the extracellular environment. Many of the secreted proteins are major virulence factors in plants and animals [33]. Type II secretion systems have been found in all completely sequenced plant pathogenic bacterial genomes, except in Agrobacterium tumefaciens. In addition, other bacteria have been shown to use secretion systems for the delivery of toxins, proteases, cellulases and lipases [34][35][36][37]. Genes coding for this system have also been reported for the three symbiotic strains Frankia [38].
Fifteen genes in M. lupini were annotated as components of the Type II secretion system, grouped into clusters of three to five genes ( Table 3). Nine of these genes were annotated as Type II secretion system proteins including protein E and protein F; four were recorded as TadE family proteins and Milup08_40403 was annotated as an uncharacterized protein closest to one found in the Frankia symbiont of Datisca glomerata.
The secretion systems III and IV which are commonly related to plant-associated bacteria transport a wide variety of effector proteins into the extracellular medium or into the cytoplasm of eukaryotic host cells thus affecting the interaction [39]. In addition, a functional type IV system has been described in the plant symbiont M. loti strain R7A [40]. A gene annotated as virB4 and related to secretion system IV was located in Lupac 08 (MiLup08_42651), this ORF is surrounded by proteins with unknown function related to those present in the genomes of Micromonospora sp. L5 and M. aurantiaca ATCC 27029 T .
Survival against plant defenses. Reactive oxygen species (ROS) play a major role in plant defense against pathogens. In response to attempted invasion, plants mount a broad range of defense responses, including the synthesis of ROS. M. lupini needs to survive under an oxidative environment in the rhizosphere before it can colonize plant roots and its genome revealed several genes encoding proteins to neutralize oxidative stress. The following genes were identified: three sod genes (MiLup08_45788, MiLup08_46012 and MiLup08_46604) that code for superoxide dismutases; a catalase HPII katE (MiLup08_44247); a catalaseperoxidase (katG, MiLup08_44435) and a catalase hydroperoxidase (katA, MiLup08_45857); four hydroperoxide reductases (MiLup08_40110, MiLup08_40293, MiLup08_41393, MiLup08_45407); a chloroperoxidase (MiLup08_44157) and a thiol peroxidase (MiLup08_43629).
In addition, a putative organic hydroperoxide resistance protein (Ohr, MiLup08_45098); a 4-hydroxyphenylpyruvate dioxygenase (Hpd, MiLup08_46664) and a homogentisate 1,2-dioxygenase (MiLup08_46677) were identified. Other enzymes include a glutathione peroxidase (MiLup08_45173); two glutathione transferases (MiLup08_46358 and MiLup08_41529) and four glutathione-S-transferases (fdh, MiLup08_42270, MiLup08_42834, MiLup08_44416 and 45648). Experimental data indicated that M. lupini indeed yields a catalase positive reaction [17] confirming the functionality of some of these genes. Therefore, to successfully reach the internal plant tissues, these genes may defend the bacterium against a ROS release by the plant.

Regulation as a means of adaptation
Lifestyle can be viewed as the set of biotopes an organism can thrive into and the relationships that it establishes with other species and its abiotic components. It is one of the driving forces that contribute to the overall characteristics of bacterial genomes [41].  The M. lupini genome shows a strong emphasis on regulation, with 643 proteins (,10%) predicted to have a regulatory function. This value is lower than that reported for the saprophytic strain Streptomyces coelicolor A3(2) with an exclusively terrestrial lifestyle (965 proteins; 12.3%) [42], but higher than the endosymbiotic strains M. loti MAFF303099 (542 proteins, 7.7%) [43]; Frankia alni ACN14a (515 proteins, 7.6%); Frankia sp. EAN1pec (555, 6.1%) and Frankia sp. CcI3 (244 proteins, 4.3%).
The genome codes for various regulator families such as TetR, AraC, LacI, ArsR, MerR, AsnC, MarR, DeoR, GntR and Crp. In addition, thirty-three ECF (extra-cytoplasmic function) sigma factors were located. Furthermore, 147 genes were related to two-component regulatory systems of which 34 were LuxR proteins. These two-component systems appear to play a crucial role in quorum sensing of Gram-positive bacteria and a positive correlation between plant-microbe interactions and the number of LuxR proteins has been suggested [44,45].
Many regulatory genes (,18%) were located near polysaccharide related loci including those involved in plant cell wall degradation. Specifically, 63% of cellulose degradation or cellulose binding genes had a nearby regulator (proximity ranged from 2-4 genes up or downstream). In the case of xylan metabolism, regulators were identified for 50% of the genes, while 43% of pectin metabolism genes also had a regulator nearby. An extended overview of the regulators and their associated carbohydrate genes is presented in Table S2.

An endophytic bacterium highly equipped with an array of plant cell wall degrading enzymes
The ability of M. lupini lupac 08 to assimilate a wide range of sugars was previously reported [19] and this is clearly reflected in its genome. The range of simple and complex saccharides assimilated by this strain include cellobiose, cellulose, glucose, mannitol, starch, sucrose, trehalose, xylan and xylose among others. Genomic analyses confirmed the presence of a large number of genes devoted to the metabolism of carbohydrates, including many compounds of plant origin. Plant-polymer degrading enzymes such as cellulases, xylanases and pectinases have been suspected to play a role in internal plant colonization [46]. In the case of plant pathogenic bacteria and fungi, these gain access by actively degrading plant cell wall compounds using glycoside hydrolases including cellulases and endoglucanases. However, genomic analyses show that non-pathogenic endophytic microorganisms such as Enterobacter sp. 638 [47], Azoarcus BH72 [39] or the symbiotic actinobacterium Frankia sp. [48] have only a reduced set of cell-wall degrading enzymes.
The genome of M. lupini Lupac 08 revealed a significant number of genes encoding enzymes potentially involved in plantpolymer degradation but also an important number of cellulosebinding related genes. Overall, about 10% of the genome coded for genes related to carbohydrate metabolism of which 192 had a hydrolytic function. At least 79 genes putatively involved in interactions with plants and with the potential to hydrolyze plant polymers were identified (Table S1). These genes were placed into the glycosyl hydrolase families GH5, GH6, GH9, GH10, GH11, GH18, GH20, GH43, GH44 and GH62, or into the carbohydrate binding modules CBM2, CBM13, CBM33, CBM3, CBM46, CBM42, CBM5, CBM4, CBM6 and CBM32. The CBM2 family was the most abundant appearing in 46 of the 79 genes identified.
Fourteen genes were further identified as lectins or proteins with lectin binding domains, which presumably bind to and interact with carbohydrates. Some of these loci (e.g. Milup_42969, Milup_42975, Milup_44484, and Milup_44962) appear to be related to cellulases and xylanases, respectively. These proteins are important as they serve as a means of attachment between a bacterium and its host (animal or plant) and are produced by either of the two interacting organisms [48].
Compared to the 45 enzymes predicted to act on oligo-and/or polysaccharides reported for T. fusca XY [49], the number of these enzymes present in the genome of M. lupini is significantly higher.
Cellulose metabolism. Aerobic cellulolytic actinobacteria have been shown to use a system for cellulose degradation consisting of sets of soluble cellulases and hemicellulases. Most of these independent cellulolytic enzymes contain one or more carbohydrate binding domains [50].
A total of 46 genes were found to present a hydrolytic or binding fuction towards cellulose (Table S1). Several endoglucanases were detected in strain Lupac 08 (e.g. C1, C2 C10 and C14), these enzymes hydrolyze internal bonds at random positions of amorphous regions of cellulose and generate chain ends for the processive action of cellobiohydrolases (exoglucanases). A copy of the exoglucanase gene cbhA (C16) was also located in the genome. Exoglucanases act on the ends of cellulose polysaccharide chains, liberating cellobiose as the major product. b-D-glucosidases such as M108 and M109 which would further hydrolyze cellobiose were also identified. In addition, several extracellular cellulase coding genes were identified including celA (C3 and C6), celB (C5) and celD (C13). These results strongly suggest that strain Lupac 08 is potentially capable of completely degrading cellulose.
Strain Lupac 08 was tested for in vitro production of cellulases. Very high cellulase activity was detected in minimal agar supplemented with carboxymethylcellulose (CMC, 0.5%) (Fig. 6A). When the culture medium was supplemented with glucose (1%) similar results were obtained indicating that this sugar did not repress nor derepress the expression of the genes responsible for the production of cellulases.
In vitro xylanase activity was detected in strain Lupac 08 when tested in a minimal medium supplemented with xylan (1%). Production of xylanases was detected after incubation for 4 days increasing significantly after 14 days (Fig. 6D). The substrate was assayed with and without glucose with similar results.
Starch degradation. Starch is a ubiquitous and easily accessible source of energy. In plant cells it is usually deposited as large granules in the cytoplasm. Several genes coding for amyloa-1,6-glucosidases (e.g. M26, M32, M44, M63, M111 and M121; Table S1) were located in addition to two amyE homologs that code for an extracellular a-amylase. Furthermore, strain Lupac 08 was able to degrade this polymer under laboratory conditions (Fig. 6C) and it was previously shown that Lupac 08 can utilize this substrate as a carbon source [17].
Pectin degradation. Pectinolytic enzymes can degrade pectic substances either through hydrolysis (hydrolases) or transelimination (lyases) [51] and are important virulence mechanisms in many soft-rotting and macerating pathogens [52]. Six pectate lyases (P1, P3, P4, P5 and P6; Table S1) were located in the genome of M. lupini, two of which were annotated as virulence factors (P5 and P6). In addition, an extracellular pectin methylesterase gene, pmeA, and a gene coding for a pectate lyase involved in D-galacturonic acid hydrolysis (P7) were also identified. Interestingly, T. fusca XY contains two pectin lyase homologs but does not appear to possess a pectin methylesterase or a pectin acetylesterase gene. Pectinase-encoding genes are reported to be absent in other endophytic microorganisms such as Azoarcus sp. BH72 or Enterobacter sp. 638 [39,47]. Production of pectinases was observed under laboratory conditions and activity was visualized after 8 days of incubation (Fig. 7D).
Expansin-like proteins. Expansins are proteins that were first described from plants [53]. These molecules function as cell wall loosening proteins by disrupting the noncovalent binding of matrix polysaccharides to cellulose [54], resulting in physical effects, such as polymer creep and stress relaxation of extended cell walls [55,56]. Many plant-associated microorganisms including several pathogenic actinobacterial species have been shown to contain proteins with expansin-like domains [57].
Two genes (MiLup_41274 and MiLup_45306) were identified in the genome of strain Lupac 08 that encode for a secreted protein showing 42% and 48% sequence similarity to the corresponding celA genes of Clavibacter michiganensis subsp. michiganensis and Clavibacter michiganensis subsp. sepedonicus, respectively. This gene corresponds to a secreted b-1,4-endoglucanase (CelA) that is required for virulence and contains a Cterminal a-expansin like domain [58,59]. In the case of C. michiganesis subsp. michiganensis CelA, this expansin-like domain is essential for development of wilting symptoms [58]. It is suggested that microbial expansins function to promote microbeplant interactions, both harmful and beneficial ones [60].

Plant growth promotion traits of Micromonospora lupini Lupac 08
Our current knowledge of plant-microbe interactions indicates that populations inhabiting a host plant are not restricted to a single microbial species but comprise several genera and species. Few reports are available regarding the presence of other microorganisms (associated or endophytic) in nitrogen fixing nodules, in spite of the fact that nodules are much richer in nutrients as compared to roots [61]. The recent reports on the isolation of large Micromonospora populations from nitrogen fixing nodules clearly suggest that this bacterium plays an important role which has yet to be defined.
Effect of M. lupini Lupac 08 on Trifolium. Micromonospora lupini Lupac 08 clearly produced a plant growth enhancing effect when it was co-inoculated with Rhizobium sp. E11 under laboratory conditions on clover plantlets. In general, the number of nodules was higher in those plants co-inoculated (18-24 nodules) with both bacteria as compared to the plants inoculated only with Rhizobium sp. E11 (11-15 nodules). Overall, the co-inoculated plants showed better growth and were larger in size as compared to the other two treatments (Fig. 7C). Similar results were previously observed when strain Lupac 08 was inoculated in its original host, Lupinus [62].
Nitrogen fixing capacity. Indirect evidence of nitrogen fixing genes was obtained by partial amplification of nifH-like gene fragments in strains Micromonospora sp. L5 [12] and M. lupini Lupac 08 [11]. In the present work the genomes of Micromonospora lupini Lupac 08 and Micromonospora sp. L5 were screened for the presence of nitrogen fixing genes to confirm this earlier finding. After thorough analysis of the complete genome, no sequences related to this biological process were detected, supporting the results reported for strain Micromonospora sp. L5 [19]. Nitrogenase activity detection by acetylene reduction assays carried out with strain Lupac 08 over a period of two weeks were negative. A positive result was reported for strain L5 [12.].

Trehalose and its role in nodulation and bacteroid
survival. Trehalose is a common reserve disaccharide in the root nodules of legumes, present at high concentrations in bacteroids at the onset of nitrogen fixation [63]. It has been reported that in the interaction between Phaseolus vulgaris and Rhizobium, enhanced germination, quality and grain yield have been correlated with trehalose content, and a higher tolerance to abiotic stress [64,65]. On the other hand, the trehalose content appears to be regulated by trehalase, a nodule stimulated plant enzyme [66,67]. Although trehalose metabolism in leguminous plants is still poorly understood, it has been shown that in senescent nodules, trehalose becomes the most abundant nonstructural carbohydrate [68] and it is proposed that trehalose, a stress protectant accumulated in bacteria, could offset membrane injuries and/or serve as an intermediate energy reserve. Indeed, Müller et al. [68] showed that during terminal senescence of nodules an appreciable part of the bacteria maintained their trehalose pools and survived.
Eight genes related to the metabolism of trehalose were detected in the genome of Lupac 08; seven genes were related with trehalose synthesis (Mlup08_40949, Mlup08_43225, Mlup08_43226, Mlup_45189, Mlup_45758, Mlup08_45759 and Mlup08_45961) and one (treA, Mlup08_45961) with the enzyme trehalase. Barraza et al. [67] proposed that modification of the trehalose content in the nodules could trigger physiological alterations that would enhace carbon and nitrogen metabolism, as well as bacteroid fitness (greater survival) and nitrogen fixation, which in turn would positively impact on symbiotic interactions. Micromonospora may contribute to the survival of rhizobia by helping to maintain high levels of trehalose.
Chitin degradation and protection against pathogens.  Chitinases often work synergistically with chitin-binding proteins (CBPs). The biological roles of bacterial chitinases and carbon binding proteins are easily understood in an environmental context, especially in soil (that harbour fungi and insects) and marine (shellfish) habitats and their impact on chitin cycling. However, there is an increasing amount of direct or indirect evidence suggesting that some chitinases and CBPs additionally serve as virulence factors for bacterial pathogens during infection of non-chitinous substrates [69].
Experimental data confirmed in vitro chitin degradation of strain Lupac 08 ( Figure 6B). As with other endophytic bacteria Micromonospora may produce chitinases to inhibit fungal pathogens, or may produce these molecules to elicit the plant defense mechanism. Either way, it seems that Micromonospora would provide a benefit to its host.

Siderophores (Iron-transport) and other secondary metabolites
Iron is an element essential for every living organism, as a cofactor of numerous proteins. Siderophores produced by plant growth promoting bacteria may reduce the growth of phytophathogens by depriving them of iron. Thus, an efficient iron uptake system can contribute to protect the host plant against pathogens. Interestingly, siderophores can also act as important virulence determinants for both plant and animal pathogens [70].
The genome of strain Lupac 08 revealed several siderophore related genes including specific iron uptake transporters, secretion of different siderophores and synthesis of siderophore receptors. Namely, a zinc/iron permease (MiLup_40258), a ferrous iron permease FTR1 (efeU, MiLup_41076) and eight iron ABC transporters (MiLup_42281-MiLup_42285). The number of the latter transporters is similar to the number of those found in the genome of the endophytic bacterium Enterobacter sp. 638 [47] while the plant pathogen Erwinia amylovora CFBP 1430 presents only three such transporters [71].
Several gene clusters related to the biosynthesis and transport of the siderophores enterobactin (MiLup_44069-MiUp_44071), aerobactin (iucA/iucC family protein, MiLup_44063 and MiLup_44064; MiLup_40326) and alcaligin (MiLup_44065) were also located. In the case of aerobactin, the gene iucA is highly correlated with virulence in avian pathogenic E. coli strains [72].
Two siderophore-interacting proteins (MiLup_40648 and MiLup_45559) were also found. One of these genes (MiLup_40648) was located next to a siderophore transporter of the RhtX/FptX family; RhtX from S. meliloti 2011 and FptX from Pseudomonas aeruginosa appear to be single polypeptide transporters from the major facilitator family for import of siderophores as a means to import iron [73]. In addition, a thiazolinyl imide reductase involved in siderophore biosynthesis was also identified (MiLup_43551).
The genome of Lupac 08 also contained several regulators including an iron-dependent repressor (IdeR, MiLup_41668), two ferric uptake regulation proteins (MiLup_40794 and MiLup_44436) and a putative iron-regulated membrane protein which suggests that these systems are highly regulated. Production of siderophores was detected experimentally (Fig. 7A).
Actinobacteria are well known to be capable of producing a vast diversity of natural secondary metabolite compounds with applications in medicine, agriculture, and other biotechnological areas [74]. Endophytic bacteria are currently of significant interest as an untapped resource of novel bioactive small molecules because their metabolites are speculated to affect the physiological conditions of host plants including growth and disease resistance. Micromonospora strains are well known for their capacity to produce many secondary metabolites and M. lupini Lupac 08 was previously screened for the production of novel compounds with antitumoral activity and the results obtained confirmed the production of a new family of molecules named Lupinacidins A, B and C [9,10].
Fifteen clusters involved in the biosynthesis of secondary metabolites were identified in the genome of M. lupini Lupac 08. These included siderophores (see above), terpenes, butyrolactones, polyketides (PKS), nonribosomal peptides (NRPS), chalcone synthases and bacteriocins (Table 4). A DNA stretch of 544 kb was estimated to code predominantly for secondary metabolites, accounting for about 7.4% of the genome. This percentage is lower than that reported for the marine actinobacterium S. tropica (9.9%, [75]) but it is within the range of other actinobacteria e.g. S. coelicolor (8.2%, [43]). Interestingly, Frankia strains ACN14a and EAN1pec dedicate about 5% of their genomes to natural product assembly while the potential of CcI3, which has the smallest genome of the three Frankia strains, has a much reduced host range and is absent from most soils is significantly smaller (,3%) [76].
Several clusters identified in the genome of M. lupini were also located in other genomes of phylogenetically related bacteria, especially in S. tropica CNB-440, S. arenicola CNS-205 and V. maris AB-18-032. Nevertheless other clusters were unique to M lupini ( Table 4). Eight of the 15 clusters identified were located in the region between coordinates 4,000 kb and 5,000 kb of the genome, close to the terminus of replication. This area of the genome also contains a high density of genes coding for the biosynthesis of various plant cell wall degrading enzymes and several transposases.
Terpene related enzymes present in the genome of M. lupini are involved in the synthesis of carotenoids, sugar-binding lipids and the production of pentalenolactone type antibiotics. Similar molecules have also been predicted from the genomes of the three sequenced Frankia strains [77]. Various polyketide biosynthetic and non-ribosomal peptide synthase pathways were also identified specifically as PKI, PKII (2 clusters), PKIII types, NRPS (2 clusters) and hybrid PKS/NRPS clusters (2 clusters). The presence of these gene clusters suggests that M. lupini is capable of producing a vast diversity of secondary metabolites such as the antitumor anthraquinone derivative lupinacidins reported earlier [9,10]. Some of these metabolites may perform specialized functions in ecological niches and recent studies have reported on the importance of PKS and NRPS molecules and their potential role in communication during root colonization [78,79]. In addition, cluster 10 contains genes that putatively code for the production of granaticin. Granaticins are antibiotics of the benzoisochromanequinone class of aromatic polyketides, the best known member of which is actinorhodin produced by S. coelicolor A3(2). Production of granaticins has mainly been reported from Streptomyces strains [80]. NRPS cluster 11 (see Table 4), appeared to be unique to strain Lupac 08 as this group of genes was not detected in any of the other genomes compared except in S. tropica CNB-440 where it seems to be only partially conserved.
The PKS type III cluster corresponds to several genes that code for the production of naringenin, a central precursor of many flavonoids. It has recently been proposed that flavonoids play an important role in the establishment of plant root endosymbioses. In the case of legume-Rhizobium interactions, flavonoids released by plant roots induce genes involved in nodulation [81]. In a similar way it has also been suggested that these molecules play an important role during the early stages of the symbiotic association between Frankia and actinorhizal plants [82]. Micromonospora flavonoids may contribute to support communication between the nitrogen fixing bacteria and their host plants. Genomic information also revealed that M. lupini has the potential to produce bacteriocins (cluster 12) as suggested by the presence of a putative short-chain dehydrogenase/reductase.

Phytohormones
Indole-3 acetic acid. Diverse bacterial species have the ability to produce auxinic phytohormones such as indole-3-acetic acid (IAA) and a few can also produce phenyl-acetate (PAA) such as Frankia alni [83,84]. Different biosynthesis pathways have been identified and redundancy for IAA biosynthesis is widespread among plant-associated bacteria [83]. Interactions between IAAproducing bacteria and plants may lead to several outcomes, from pathogenesis to phytostimulation [85]. The genome of M. lupini Lupac 08 contains a gene (Milup_45687) potentially involved in the biosynthesis of IAA via the indole-3-acetonitrile pathway. This gene corresponds to the conversion of indole-3-acetonitrile to indole-3-acetic acid. Nitrilases with specificity for indole-3acetonitrile have been reported in Alcaligenes faecalis [86]. In A. tumefaciens and Rhizobium spp. nitrile hydratase and amidase activity could be identified, indicating the conversion of indole-3acetonitrile to indole-3-acetic acid via indole-3-acetamide [87]. Analysis of IAA production by strain Lupac 08 was carried out, yielding a positive result (Fig. 7B).
Acetoin and 2,3-butanediol. The volatile compounds acetoin and 2,3-butanediol produced by bacteria such as Bacillus subtilis GB03 and Bacillus amyloliquefaciens IN937a have been reported as plant growth promoting hormones [88]. Several genes were located in the genome of strain Lupac 08 which could be involved in the production of these compounds. Two copies of the gene pdhB (MiLup_40114 and MiLup_43782) that encode the enzyme pyruvate dehydrogenase were identified. This enzyme transforms pyruvate to acetaldehyde and in this process a small fraction of pyruvate is converted to acetoin as a by-product. In addition, three acetolactate synthases involved in the synthesis of acetolactate from pyruvate are present (MiLup_41336, MiLup_41383 and MiLup_41384). Under aerobic conditions, acetolactate is converted to acetoin by the enzyme acetoin dehydrogenase (MiLup_41670).

Discussion
The most extensively studied bacteria interacting with plants are Gram-negative proteobacteria because they are readily isolated from plant tissues and genetically handled for interaction studies. However, the impact of Gram-positive bacteria on plants should not be underestimated as has been done for many years mainly due to their slow growth. M. lupini Lupac 08, a Gram-positive actinobacterium was isolated from the internal root nodule tissues of Lupinus angustifolius but it is only a representative of a large collection of more than 2000 Micromonospora strains isolated from diverse legumes and actinorhizals from different geographical locations. So far, several genomes of root symbionts and soil saprophytes have been studied; therefore we decided to focus on an intermediary category, that of facultative endophytes.
Lupac 08 was isolated from lupine nodules, and shown to produce the anticancer agents Lupinadicin A, B and C. The genome of strain Lupac 08 was sequenced to obtain information about the potential ecological role of Micromonospora in interaction with legumes and actinorhizal plants. Genomic analysis revealed several strategies which are probably necessary to lead a successful lifestyle as a saprophyte in the rhizosphere, a competitive and harsh environment, and as an endophyte capable of colonizing the internal plant tissues. Micromonospora species have less than 3% distance in their 16S rRNA genes, which can be roughly translated to a time of 150MY according to the equivalence proposed by Ochman and Wilson [89]. In the current phylogeny (Fig. 1), M. lupini has as closest neighbours M. chokoriensis M. saelicesensis and M. zamorensis, isolated from sandy soil, root nodules of L. angustifolius and the rhizosphere of P. sativum respectively. Micromonospora sp. L5 and M. aurantiaca are located further away, with a distance of 1.2% that would translate to 60 MY for the emergence of a group of species that interact with plants, a date that would be close to the postulated time of emergence of Fabaceae and that of many actinorhizal plant families [90]. The separation from the Salinispora and Verrucosispora lineages would constitute two independent events that would have occurred slightly earlier at 160MY and 170MY, while the emergence of the Actinoplanes and that of the Dactylosporangium would have occurred 250MY ago, a time when dicotyledons had not yet appeared but when continents and thus soils had appeared that did permit the growth of primitive plants such as the gymnosperms.
The size of the Micromonospora genomes analyzed is quite uniform, with that of Lupac 08 slightly larger. The chromosome size of M. lupini Lupac 08 appears to reflect a wealth of genes allowing for adaptation to a complex saprophytic/endophytic lifestyle, which means adapting to a wider range of environmental conditions with the ability to metabolize a large variety of nutrient sources. Considering that Micromonospora sp. L5 and M. lupini Lupac 08 were both isolated from nitrogen fixing nodules (actinorhizal and legume plants, respectively), it would be expected that the genomes of these strains be more similar to each other than to M. aurantiaca ATCC 27029 T which was originally isolated from soil. Surprisingly this was not the case as confirmed by the high number of strain specific genes identified in the genome of Lupac 08, suggesting a high capacity of adaptation to a fluctuating environment by this microorganism. On the other hand, Micromonospora sp. L5 and M. aurantiaca ATCC 27029 T share a high number of orthologous genes (86%) suggesting that the niche of origin is not crucial.
An interesting result was the distribution of the metabolic profiles of 20 bacteria representing different living environments (Fig. 5). There was a clear proximity between M. lupini Lupac 08 and the three Frankia genomes. This result suggests that strain Lupac 08 contains metabolic functions similar to those found in Frankia strains that are probably useful for its interaction with plants. This metabolic versatility combined with a diverse transport system make Lupac 08 an organism fit to adapt to a soil/plant environment.
The emergence and evolution of nitrogen fixation ability among the domains Bacteria and Archaea is complex and has not yet been fully elucidated. Although it was previously reported that Micromonospora strains isolated from legume and actinorhizal root nodules contained nifH-like gene fragments [11,12], we could not confirm these results. In a similar approach based on PCRamplification, other authors reported the presence of nif-H like sequences for bacterial isolates obtained from legumes collected in arid zones including Microbacterium, Agromyces, Mycobacterium and Ornithinicoccus [91]. One recurrent problem with the use of a PCR-based approach is that it is limited to a single gene amplified billions of times, which may provide false-positive results [92] and for this reason must always be confirmed by an independent approach.
Plant-polymer degrading enzymes such as cellulases and pectinases have been suspected to play a role in internal colonization. Most plant pathogens secrete cellulases, pectinases, xylanases, or other enzymes to hydrolyze plant cell wall polymers, while a lack of secreted hydrolases has been proposed to be favourable for microorganisms that form beneficial association with plants. Examples of endophytic plant growth promoting bacteria that lack large amounts of cell wall degrading enzymes include Frankia [38], Enterobacter sp. 638 [47], Azoarcus sp. BH72 [60] and Herbaspirillum seropedicae [93]. An Azospirillum sp. that does not colonize root tissues proper, but only the rhizosphere, has a genome containing a large number of putative cellulases similar to soil cellulolytic bacteria with 26-34 glycosyl hydrolases [94], as compared to the 37 present in T. fusca, a highly cellulolytic actinobacterium isolated from soil.
The genome of M. lupini revealed a high number of putative genes that encode for hydrolytic enzymes and specifically cellulolytic, xylanolytic, chitinolytic and pectinolytic activities were confirmed in the laboratory, indicating the capacity of Micromonospora to degrade plant polymers in a way similar to that of plant pathogen microorganisms. In the case of Micromonospora, there seems to be a paradox since strain M. lupini Lupac 08 shows a very high in vitro activity for cellulases and xylanases, however, preliminary inoculation experiments in our laboratory indicate that the microorganism does not behave as a pathogen, on the contrary, Micromonospora appears to interact in a tripartite relationship stimulating nodulation and plant growth (Fig. 7c). Therefore the question arises as to what is the function of these enzymes when Micromonospora interacts with the host plant. Alternatively some of the genes present, especially those related to the metabolism of cellulose may not necessarily imply that the bacterium is involved in plant cell wall degradation but have a different role, yet to be defined [95]. In addition many of the cellulose-related genes contain binding-domains suggesting that these may be related to the adhesion of the bacterium to the plant. These genes could also help Micromonospora digest plant cell walls upon senescence of the nodules.
M. lupini Lupac 08 contains several secondary metabolite gene clusters, many of which appear to be involved in the synthesis of siderophores and also of antibiotics. These would also in all likelihood be involved in the synthesis of the antitumor anthraquinone molecules described previously [9,10]. Micromonospora, like many other endophytic bacteria is a facultative plant colonizer that must compete with other microorganisms in the rhizosphere before entering the plant. In this sense, the NRPS and PKS gene clusters identified in the genome of M. lupini Lupac 08 may be involved in defense as well as in interaction and communication with its host plant. Thus, it will be necessary to identify these compounds and their functional attributes to further expand our knowledge of this plant-microbe interaction.

Conclusions
We have provided experimental data which supports the hypothesis that M. lupini Lupac 08 is a plant growth promoting bacterium. Micromonospora lupini Lupac 08 clearly produces a plant growth enhancing effect as observed in laboratory experiments. The localization of several genes involved in plant growth promotion traits such as the production of siderophores, phytohormones, the degradation of chitin (biocontrol) and the biosynthesis of trehalose may all contribute to the welfare of the host plant. Micromonospora appears as a new candidate in plantmicrobe interactions with an important potential in agriculture and other biotechnological applications. The current data is promising but it is still too early to determine which specific roles are played by this microorganism in interactions with nitrogen fixing plants.

Genome sequencing, annotation and analysis
The genome sequence of M. lupini Lupac 08 was determined using the 454 FLX system and Titanium platform (454 Life Sciences) as previously reported [96]. Sequences were assembled into 50 contigs and four scaffolds ranging from 583 to 7,083,659 nucelotides using the MaGe (Magnifying Genomes) interface [97]. This Whole Genome Shotgun project has been deposited at European Nucleotide Archive under accession number NZ_CAIE00000000.01.

16S rRNA gene phylogeny
Sequences obtained from public databases (Genbank/EMBL) were manually aligned using clustal X software [98]. Phylogenetic distances were calculated with the Kimura 2-parameter model [99] and the tree topologies were inferred using the maximumlikelihood method [100]. All analyses were carried out using the MEGA5 program [101].

Comparative genome analysis
Genome rearrangement of the Micromonospora strains M. lupini Lupac 08, Micromonospora sp. L5, M. aurantiaca ATCC 27029 T and Micromonospora sp. ATCC 39149 were carried out using MAUVE software [102]. The number of shared and unique genes present in the respective genomes were calculated and represented by a Venn-diagram using the EDGAR software [103]. Potential horizontally transferred genes were predicted using the ''Region of Genomic Plasticity Finder'' method implemented on the MicroScope platform. First we selected genomes included in the PkGDB and NCBI RefSeq databases that presented high synteny percentages with the Lupac 08 strain. Automatic results were manually curated according to several features such as base composition, DNA repetitions, presence of near mobile elements and information provided by SIGI and IVON programs [104].

Comparative analysis of metabolic profiles
A bicluster plot of the metabolic profiles for M lupini Lupac 08 and 20 other bacterial genomes was performed with Multibiplot [105]. A comparison of 798 MicroCyc metabolic pathways was made using MaGe. This comparison is based on the calculation of 'pathway completion' values, scaled from 0 to 1, where 0 means that a particular organism does not contain any enzyme for a given pathway and 1 that it contains all the reactions of the pathway. These values were transformed applying row standardization and a JK-Biplot was constructed after performing a PCA (Singular Value Decomposition estimation method). The heatmap was then obtained with the expected values computing the Euclidian distance and average linkage for rows and columns.

Transport proteins identification and classification
Information about transport proteins of genomes was obtained from the TransportDB relational database when available (http:// www.membranetransport.org/). The identification and classification of the transporters of strain Lupac 08 was performed using the TransAAP tool based on TransportDB [106] followed by manual validation.

Cellulose, starch and xylan degradation
Strain Lupac 08 was cultivated on yeast-malt agar for 5 days and subsequently transferred to M3 agar [107] with and without glucose and supplemented with one of the substrates in the following way: carboxymethylcellulose (CMC, 0.5%), starch (1%) and xylan (1%). A bacterial suspension of 10 6 per ml was prepared in saline solution (0.85%) and 200 ml were inoculated on the different plates which were then incubated at 28uC and results were recorded at 4, 7 and 14 days after inoculation. Xylan and CMC plates were stained with Congo red [108] while starch plates were flooded with iodine solution [109].

Pectin degradation
Pectinolytic acitivity was determined as described in Williams et al. [109]. Agar plates supplemented with pectin (0.5%) were streaked with strain Lupac 08 and incubated at 28uC. Hydrolysis zones were detected after 14 days incubation by flooding plates with an aqueous solution of cetyltrimethyl ammonium bromide (CTAB, 1%) and examining them after 30 min.

Siderophore production
Siderophore production was assessed using a modified chrome azurol S (CAS) assay [111]. Strain Lupac 08 was cultured on yeastmalt agar and incubated for 7 days, subsequently it was streaked onto CAS agar plates and incubated at 28uC for 7-10 days. A positive result was indicated by an orange halo around bacterial colonies.

Indole-3-acetic acid production
Production of indole acetic acid was assayed following the method of Glickmann and Dessaux [112]. Strain Lupac 08 was inoculated in 5 ml of yeast-malt medium supplemented with Ltryptophan (0.2%) and incubated at 28uC at 150 rpm during 7 days. The culture was then centrifuged at 12,000 x g for 10 min and 1 ml of the supernatant was mixed with 2 ml of Salkowski's reagent [113] and incubated at room temperature for 30 min. IAA production was measured spectrophotometrically at 530 nm to assess the development of a pink colour.

Plant growth
Surface-sterilized seeds of Trifolium sp. were germinated axenically in Petri dishes on 1.4% w/v agar. Seedlings were transferred to sterile square plastic plates that contained a nitrogen-free nutrient solution [114]. Fifteen plantlets were inoculated in the following manner (5 per treatment): 500 ml (each) of bacterial suspensions (10 6 cells per ml) of M. lupini Lupac 08 and Rhizobium sp. E11 for coinoculation treatment; inoculation with Rhizobium sp. E11; and uninoculated plants as negative controls.

Acetylene reduction activity
Nitrogenase activity was measured using acetylene reduction [115] in sterile 150 ml plasma flasks with a rubber stopper. Cells of Lupac 08 were cultured in liquid minimal glucose medium without nitrogen at 28uC with shaking. The air in flasks was replaced with mixture of air and acetylene (ration 90:10 v/v). One mililiter of mixture was sampled for each measure using gas chromatography with a flame ionization detector (Girdel 30, France). Mesorhizobium melitolti Sm1021 was used as positive control.