Survey of Chickpea Rhizobia Diversity in Portugal Reveals the Predominance of Species Distinct from Mesorhizobium ciceri and Mesorhizobium mediterraneum

PLANT MICROBE INTERACTIONS

Survey of Chickpea Rhizobia Diversity in Portugal Revealsthe Predominance of Species Distinct from Mesorhizobiumciceri and Mesorhizobium mediterraneum

Ana Alexandre & Clarisse Brígido & Marta Laranjo &

Sérgio Rodrigues & Solange Oliveira

Received: 12 February 2009 /Accepted: 6 May 2009 /Published online: 26 May 2009# Springer Science + Business Media, LLC 2009

Abstract Several Mesorhizobium species are able toinduce effective nodules in chickpea, one of the mostimportant legumes worldwide. Our aims were to examinethe biogeography of chickpea rhizobia, to search for apredominant species, and to identify the most efficientmicrosymbiont, considering Portugal as a case study. Onehundred and ten isolates were obtained from continentalPortugal and Madeira Island. The 16S ribosomal RNA genephylogeny revealed that isolates are highly diverse, groupingwith mostMesorhizobium type strains, in four main clusters(A–D). Interestingly, only 33% of the isolates groupedwith Mesorhizobium ciceri (cluster B) or Mesorhizobiummediterraneum (cluster D), the formerly described specificchickpea microsymbionts. Most isolates belong to clusterA, showing higher sequence similarity with Mesorhizobiumhuakuii and Mesorhizobium amorphae. The associationfound between the province of origin and species cluster ofthe isolates suggests biogeography patterns: most isolatesfrom the north, center, and south belong to clusters B, A,and D, respectively. Most of the highly efficient isolates(symbiotic effectiveness >75%) belong to cluster B. Acorrelation was found between species cluster and originsoil pH of the isolates, suggesting that pH is a key environ-

mental factor, which influences the species geographicdistribution. To our knowledge, this is one of the fewsurveys on chickpea rhizobia and the first systematicassessment of indigenous rhizobia in Portugal.

Introduction

Legumes are able to establish nitrogen-fixing symbioseswith bacterial microsymbionts (rhizobia), thus reducing theneed for chemical fertilizers. This association furtherprovides a nitrogen supplement for the subsequent crops [1].

Chickpea (Cicer arietinum L.) is the third most impor-tant legume crop worldwide, after dry bean and pea [2].This legume has been used as an alternative crop to cerealsunder dry land conditions, namely in Portugal [3]. Thesymbiotic relationship between rhizobia and chickpea hasnot been extensively analyzed, and there are few studiesaddressing the genetic diversity of chickpea rhizobia [4–7].In south Portugal, native chickpea rhizobia from an agri-cultural region have been studied [8–12].

Rhizobia that nodulate chickpea were described for the firsttime by Cadahía and colleagues [13]. Later, Jarvis and co-workers [14] included them in the genus Mesorhizobium.Two species were first identified as specific chickpeamicrosymbionts: Mesorhizobium ciceri [15] and Mesorhi-zobium mediterraneum [16].

The promiscuity of a given legume is related to thenumber of Nod factors it can interact with, rather than thediversity of rhizobia, which are able to nodulate suchlegume [17]. Chickpea has been considered a narrow-hostrange legume [18], mainly because it cannot be nodulatedby broad-host range rhizobia, such as Rhizobium sp. NGR234 [19]. Nevertheless, recent studies have shown thatchickpea is able to establish symbioses with several species

Microb Ecol (2009) 58:930–941DOI 10.1007/s00248-009-9536-6

A. Alexandre : C. Brígido :M. Laranjo : S. RodriguesLaboratório de Microbiologia do Solo,Instituto de Ciências Agrárias Mediterrânicas (I.C.A.M.),Universidade de Évora,Évora, Portugal

S. Oliveira (*)Departamento de Biologia, Universidade de Évora,Apartado 94,7002-554 Évora, Portugale-mail: [email protected]

of Mesorhizobium, namely Mesorhizobium amorphae,Mesorhizobium loti, and Mesorhizobium tianshanense;however, these carry symbiosis genes (nodC and nifH)identical to those carried by M. ciceri and M. mediterra-neum, which may lead to the production of similar Nodfactors [12, 20, 21]. Similar results have been reportedfor Mesorhizobium species that nodulate Anagyris latifolia[22] and Bradyrhizobium species nodulating Lupinusspp. and Ornithopus spp. [23]. On the other hand, legumessuch as Phaseolus vulgaris are non-selective hosts fornodulation [24].

Bacterial phylogeny has relied on the sequence analysisof single core genes, such as 16S ribosomal RNA (rRNA)gene [25], dnaJ [26], atpD, and recA [27], among others.However, several studies have shown that single-gene treesmay not adequately reflect phylogenetic relationships, somultilocus approaches have been proposed for speciesidentification [28–31]. Still, in the last decade, analysis ofthe 16S rRNA gene has been, by far, the most widely usedapproach to define molecular phylogeny and taxonomy ofbacteria [32, 33]. The 16S rRNA gene is the only sequenceavailable for most bacterial species, including type strains.Thus, the 16S rRNA gene is a useful tool for placing anynew isolate among its closer taxonomic relatives.

Our aims were to examine the biogeography of rhizobiaable to nodulate chickpea, to investigate the presence of apredominant chickpea rhizobia species, and to identify themost efficient species in the symbiosis, considering Portugalas a case study. A survey on chickpea rhizobia was carriedout in continental and insular Portugal. Genetic diversity ofnative isolates was examined through molecular phylogenybased on 16S rRNA gene sequences and by plasmid profilesanalysis. Symbiotic effectiveness of chickpea native isolateswas estimated.

Methods

Isolates Collection

Rhizospheric soil samples were collected from more than40 sites of the 11 provinces of continental Portugal (Minho,Trás-os-Montes e Alto Douro, Douro Litoral, Beira Litoral,Beira Alta, Beira Baixa, Estremadura, Ribatejo, AltoAlentejo, Baixo Alentejo, and Algarve) and from theMadeira and Azores Islands, covering the entire country(total area of 92,000 km2). Soil samples were collectedfrom fields not used for chickpea cultivation, with theexception of the Elvas-ENMP site, which is an experi-mental agricultural field.

Chickpea seeds (Chk 3226) were surface-sterilized withcalcium hypochlorite 14%, washed with sterile distilledwater, and pre-germinated in water–agar. Seeds were sown

in sterilized pots containing the soil samples. Plants weremaintained in the plant growth chamber under controlledconditions for 8 weeks. Nodules were harvested andisolates were obtained as described by Somasegaran andHoben [34]. Isolates were re-inoculated, under sterileand controlled conditions, in order to confirm their abilityto nodulate chickpea.

Amplification of the 16S rRNA Gene

The 16S rRNA gene was amplified for each isolate usingprimers Y1 [35] and Y3 [12], corresponding to positions 20to 1507 in Escherichia coli. Amplification reaction wascarried out as previously reported [12]. Polymerase chainreaction (PCR) products were purified using GFXTM PCRDNA and Gel Band Purification kit (GE Healthcare) orExoSAP-IT®T (usb) following the manufacturer’s instruc-tions. Two additional internal primers, namely IntF andIntR [12], were used to obtain double-stranded and nearlycomplete sequences.

Phylogeny Based on the 16S rRNA GeneSequence Analysis

Nucleotide sequences were analyzed and edited usingBioEdit Sequence Alignment Editor (version 7.0.4.1) [36].Alignments were generated using Clustal W [37].

The 16S rRNA gene sequences of the isolates were com-pared to those of the type strains of the following species:Mesorhizobium albiziae (DQ100066), M. amorphae(AF041442), Mesorhizobium chacoense (AJ278249), M.ciceri (DQ444456), Mesorhizobium huakuii (FJ491264), M.loti (X67229), M. mediterraneum (AM181745), Mesorhi-zobium plurifarium (Y14158), Mesorhizobium septentrionale(AF508207), Mesorhizobium temperatum (AF508208),Mesorhizobium thiogangeticum (AJ864462),M. tianshanense(AF041447), Rhizobium etli (U28916), Rhizobium legumino-sarum bv. viciae (U29386), Sinorhizobium medicae(L39882), and Sinorhizobium meliloti (X67222). Azorhi-zobium caulinodans (X67221) and Bradyrhizobium japoni-cum (U69638) were included as outgroups.

Molecular Evolutionary Genetics Analysis 4 (version 3.1)software [38] was used to infer the molecular phylogeny bythe neighbor-joining method based on a distance matrixwith the distance correction calculated by Kimura’s two-parameter model, with 1,000 resamplings in the bootstrapanalysis.

Plasmid Profiles

Plasmid profiles were analyzed by horizontal agarose gelusing a two-comb system that allows the in-well lysismethod, as described previously [9].

Survey of Chickpea Rhizobia in Portugal 931

Symbiotic Effectiveness

Plant growth chamber trials were performed under con-trolled conditions in order to evaluate the symbioticeffectiveness (SE) of the isolates [11]. Pre-germinatedchickpea seeds, obtained as described before, were sownin sterilized vermiculite and inoculated with a bacterialsuspension of each isolate grown in MLA [39]. Uninocu-lated plants were used as negative control and uninoculatedplants supplemented with nitrogen (140 ppm nitrogen asKNO3, in the nutrient solution) were used as positivecontrol. Three replicates were used for each treatment.Plants were collected after 8 weeks and several parameterswere measured, such as shoot dry weight, root dry weight,number of nodules, and nodules dry weight. Symbioticeffectiveness was estimated according to Gibson [40], asthe ratio between the positive control and each treatment,using shoot dry weight values. The value of the negativecontrol was subtracted from both treatment and positivecontrol values.

Statistical Analysis

Statistical analysis was performed using SPSS 15.0 software(SPSS Inc., Chicago, IL, USA). Relationships betweencategorical variables were determined using the chi-squaretest of association. Relationships between a continuousvariable and an unordered categorical variable were testedusing analysis of variance (one-way ANOVA). Results arepresented as the test statistic (χ2), degrees of freedom (df),and probability of equal or greater deviation (P). Forsamples not satisfying Cochran’s criteria (some categorieswere represented by only one isolate, and more than 20% ofthe categories were represented by less than five isolates),the exact value of P, the critical probability, was computedrather than the asymptotic P value, which is an approxi-mation reserved for large samples [41]. Correspondenceanalysis (CA) was used as an explorative method to studyassociations and to reveal interdependencies between twovariables [42]. Visualization using CA is based onrepresenting χ2 distances among variables.

Results

Isolates Collection

Most of the sampled sites harbored rhizobia able tonodulate chickpea. Nevertheless, no nodules were obtainedwith the several soil samples collected from Minhoprovince and Azores Islands. A total of 110 chickpearhizobia isolates, from 23 sites in ten provinces ofcontinental Portugal and Madeira Island, were used for

further studies. Soil characteristics of each site are shown inTable 1.

Phylogeny Based on the 16S rRNA Gene Analysis

GenBank accession numbers for the 16S rRNA genesequences of all isolates are shown in Table 2. Analysisof molecular diversity was performed using full-length 16SrRNA gene sequence of a set of 43 rhizobia isolates. Adendrogram was generated by the neighbor-joining methodfrom a 1,357-bp-long alignment (257 variable sites).According to the 16S rRNA gene molecular phylogeny(Fig. 1), all native isolates were assigned to the genusMesorhizobium. Isolates form a large cluster together withthe Mesorhizobium type strains, which received 99%bootstrap support. Four main clusters (A to D) can beidentified, each cluster including isolates from at least threedifferent provinces. The largest cluster (A) comprises thetype strains of M. amorphae, M. huakuii, M. plurifarium,and M. septentrionale, as well as 25 isolates from sevenprovinces (Beira Litoral, Beira Alta, Beira Baixa, Ribatejo,Alto Alentejo, Baixo Alentejo, and Algarve) and MadeiraIsland, and received 76% bootstrap support. AlthoughM. plurifarium and M. septentrionale type strains areincluded in cluster A, no isolate was found to group closelyto these strains. In terms of 16S rRNA gene sequence,isolate C-14-Coimbra is 100% identical to M. huakuiiand isolate STR-2-Santarém is 100% identical to M. amor-phae. Isolate L-19-Leiria shares the same sequence simi-larity (99.7%) with both M. amorphae and M. huakuii.Isolates A-8b.-Aveiro and SA-4-Serra d’Água are moresimilar to M. amorphae (99.5% similarity). All otherisolates share higher sequence similarity with M. huakuii(99.7%) and are divided in subgroups apart from anyMesorhizobium type strain; some of them might be differentenough to represent new Mesorhizobium species. Cluster Bincludes eight isolates from four provinces (Douro Litoral,Estremadura, Alto Alentejo, and Baixo Alentejo), togetherwith M. loti and M. ciceri type strains. Six of these isolatesare closer to M. ciceri (99.9–100%) and two isolates arecloser to M. loti (99.8%). Cluster C includes four isolatesfrom three different provinces (Estremadura, Ribatejo, andAlto Alentejo) that share 99.8–99.9% of sequence similaritywith M. tianshanense, the only type strain included inthis cluster. Cluster D comprises the type strains of M.mediterraneum and M. temperatum and includes six isolatesfrom three provinces (Alto Alentejo, Baixo Alentejo, andAlgarve), all identical in sequence.

In order to extend the analysis of species diversity, thecomplete set of 110 Portuguese isolates was used. Takinginto account that the complete and the partial 16S rRNAgene sequence analysis, for the set of 43 isolates, generatedthe same four main clusters (data not shown), the analysis

932 A. Alexandre et al.

of the total set of 110 isolates was performed using theirpartial 16S rRNA gene sequence (Table 2). A dendrogramwas generated from a 578-bp-long alignment with 104variable sites (data not shown). The 110 isolates arecomprised in the previously described four main clusters:cluster A (56 isolates), cluster B (32 isolates), cluster C (13isolates), and cluster D (nine isolates). This analysis revealedthat chickpea rhizobia isolates are highly diverse and groupwith nine Mesorhizobium type strains. No isolate groupswith M. chacoense, M. albiziae, or M. thiogangeticum.

Figure 2 shows the geographical distribution of isolatesby provinces, according to their 16S rRNA gene clusters. Inthe north of Portugal (Trás-os-Montes e Alto Douro andDouro Litoral), isolates from cluster B prevail; in the center(Beira Alta, Beira Litoral, Beira Baixa, Ribatejo, and AltoAlentejo), most isolates are from cluster A; and in the south(Baixo Alentejo and Algarve), isolates mainly belong tocluster D. All isolates from Madeira belong to cluster A.Moreover, Estremadura is the only province where isolatesfrom cluster C predominate and is the single centerprovince with no isolates from cluster A. Isolates fromcluster C are found only in three provinces of the center ofPortugal (Estremadura, Ribatejo, and Alto Alentejo).

The geographic distribution of isolates according to theirspecies cluster is not random given that an association wasfound between species cluster and province of origin of

individual isolates (χ2=126.382, df=30, P<0.001). TheCA biplot (data not shown) revealed the existence of threeclasses of sites, consistent with the distribution of isolatesobserved in Fig. 2. One class, which includes Beira Litoral,Alto Alentejo, and Madeira, is associated with cluster A.A second class, Estremadura, is mainly associated withcluster C. Finally, a class including Trás-os-Montes e AltoDouro, Douro Litoral, Baixo Alentejo, and Algarve isassociated with clusters B and D.

Plasmid Profiles

Plasmid profiles were analyzed, and for most chickpeanative rhizobia, at least one plasmid was detected (Table 2).Plasmid number ranges from zero to six, though no isolatewith five plasmids was found. For 39% of the isolates, oneplasmid was detected. Only in about 9% of the isolates,three or more plasmids were detected. An association wasfound between plasmid number and province (χ2=99.295,df=45, P<0.001). Alto Alentejo is the province with isolatesmore variable in terms of plasmid number, harboring zero tofour plasmids, while isolates from Douro Litoral, Beira Alta,Ribatejo, Algarve, and Madeira show the least variability inplasmid number. There is also an association betweenplasmid number and species clusters (χ2=59.740, df=15,P<0.001). Figure 3 shows the distribution of isolates in

Soils Texture pH Phosphorus (ppm) Nitrogen (ppm) Potassium (ppm)

Bragança Medium 6.69 444 681 710

Lamego Medium 6.58 563 53 580

Porto II Medium 6.37 40 35 340

Aveiro Medium 7.08 44 9 66

Aveiro II Medium 6.07 200 68 740

Coimbra Medium 5.66 108 66 224

Leiria Medium 8.19 182 57 356

Guarda Coarse 5.10 476 74 1,060

Viseu Medium 5.94 12 11 126

Castelo Branco Medium 6.46 248 132 216

Telhado Medium 7.32 210 86 178

Caldas da Rainha Medium 6.83 82 131 170

Setúbal Medium 8.06 631 18 350

Sintra Medium 7.83 808 352 910

Santarém Medium 7.83 140 101 530

Elvas-CV Medium 6.20 48 64 143

Elvas-ENMP Medium 7.90 529 12 137

Évora Coarse 5.10 56 47 55

Portalegre Medium 5.25 26 87 152

Beja Medium 8.20 108 28 197

Portimão Medium 8.66 744 194 1,600

Portimão I Medium 7.20 318 153 180

Serra d’Água Medium 7.63 74 42 560

Table 1 Characteristics of soilsused to obtain chickpearhizobia isolates

Soil sample analyses wereperformed in the LaboratórioQuímico Agrícola of theUniversity of Évora, Portugal

Survey of Chickpea Rhizobia in Portugal 933

Table 2 Rhizobia isolates used in the present study

Origin Isolate 16S rRNA geneaccession number

Speciescluster

Plasmidnumber

SE (%)

Trás-os Montes e Alto Douro BR-8-Bragança EU652123 B 2 45

BR-9-Bragança EU652124 B 1 43

BR-15-Bragança EU652125 B 2 21

BR-16-Bragança EU652126 B 2 35

BR-28-Bragança EU652127 B 0 48

LM-1-Lamego EU652128 A 1 14

LM-9-Lamego EU652129 A 1 55

LM-13-Lamego EU652130 A 1 11

LM-18-Lamego EU652131 B 1 61

LM-21-Lamego EU652132 A 1 22

Douro Litoral PII-1-Porto EU652133 B 3 58

PII-2-Porto EU652134 B 2 71

PII-3-Porto EU652106a B 2 47

PII-4-Porto EU652135 B 2 31

Beira Litoral A-3-Aveiro EU652136 A 0 36

A-8b.-Aveiro EU652107a A 0 0

AII-5-Aveiro EU652137 A 2 26

AII-7-Aveiro EU652138 A 2 32

C-1-Coimbra EF504313a A 1 [48] 47

C-3-Coimbra EU652108a A 1 15

C-7-Coimbra EU652139 A 1 14

C-9-Coimbra EU652140 A 2 20

C-13-Coimbra EU652109a A 1 49

C-14-Coimbra EU652110a A 2 32

C-15-Coimbra EU652141 A 1 20

C-23-Coimbra EU652142 A 1 23

C-24-Coimbra EU652143 A 1 39

C-25-Coimbra EU652144 A 2 21

C-27b.-Coimbra EF504314a A 1 62

L-19-Leiria EU652111a A 0 48

Beira Alta G-1-Guarda EU652145 B 0 34

G-4-Guarda EU652146 B 0 41

G-10-Guarda EU652147 B 0 48

G-24-Guarda EU652148 B 0 58

G-55-Guarda EU652149 B 0 88

V-5b.-Viseu EU652112a A 1 65

V-15b.-Viseu EF504315a A 0 23

V-18-Viseu EF504316a A 0 67

V-20-Viseu EF504317a A 1 67

V-25b.-Viseu EU652113a A 1 69

Beira Baixa CB-10-Castelo Branco EU652150 B 0 56

CB-19-Castelo Branco EU652151 B 0 30

CB-23-Castelo Branco EU652152 B 4 52

CB-30-Castelo Branco EU652153 B 4 45

CB-38-Castelo Branco EU652154 B 4 61

CB-75-Castelo Branco EU652155 B 0 38

T-1-Telhado EU652156 A 1 100

T-3-Telhado EU652157 A 1 32

934 A. Alexandre et al.

Table 2 (continued)

Origin Isolate 16S rRNA geneaccession number

Speciescluster

Plasmidnumber

SE (%)

T-4-Telhado EU652158 A 1 86

T-5-Telhado EU652159 A 0 56

T-7-Telhado EU652160 A 1 54

T-8-Telhado EU652114a A 1 31

Estremadura CR-3-Caldas da Rainha EU652161 C 0 77

CR-16-Caldas da Rainha EU652162 C 0 79

CR-18-Caldas da Rainha EU652163 C 0 41

CR-29-Caldas da Rainha EU652164 C 0 55

CR-32-Caldas da Rainha EU652115a C 0 [48] 57

ST-2-Setúbal AY225401a C 0 4 [21]

ST-5-Setúbal EU652165 C 0 21

ST-8-Setúbal EU652166 C 0 7

ST-20-Setúbal EU652167 C 0 43

ST-33-Setúbal EU652168 C 0 44

S-1-Sintra EU652169 D 3 53

S-8-Sintra EU652116a B 1 83

S-15-Sintra EU652170 B 0 79

S-26-Sintra EU652171 B 0 68

Ribatejo STR-2-Santarém EU652117a A 1 40

STR-4-Santarém EU652172 A 1 50

STR-10-Santarém EU652173 A 1 28

STR-14-Santarém EU652118a C 1 64

STR-16-Santarém EU652174 C 2 49

Alto Alentejo 75-Elvas AY225386a B 1 [9] 35 [21]

78-Elvas AY225387a A 1 [9] 63 [21]

79-Elvas DQ787130 B 1 [9] 47 [9]

83-Elvas DQ787131 A 1 [9] 49 [9]

85-Elvas AY225388a A 1 [9] 60 [21]

CV-1-Elvas DQ787132 A 0 [11] 28 [11]

CV-11-Elvas DQ787133 A 0 [11] 21 [11]

CV-16-Elvas AY225389a B 1 [11] 42 [21]

CV-18-Elvas AY225390a A 1 [11] 72 [21]

EE-2-ENMP AY225396a D 2 [11] 36 [21]

EE-7-ENMP AY225397a B 0 [11] 84 [21]

EE-12-ENMP AY225398a B 1 10 [21]

EE-14-ENMP AY225399a D 4 [11] 32 [21]

EE-29-ENMP AY225400a D 4 21 [21]

87-Évora DQ787134 A 1 n.d.

89a.-Évora DQ787135 A 4 n.d.

90-Évora AY225391a A 1 49 [21]

92-Évora DQ787136 A 2 42 [9]

93-Évora AY225392a C 0 27 [21]

94-Évora AY225393a A 1 33 [21]

96-Évora DQ787137 A 2 n.d.

98-Évora AY225394a A 2 72 [21]

101-Évora DQ787138 A 2 n.d.

102-Évora AY225395a A 0 [9] 54 [21]

PT-35-Portalegre EU652119a A 1 [48] 56

Survey of Chickpea Rhizobia in Portugal 935

each cluster, according to their plasmid number. Withincluster A, most isolates have one plasmid, while in clustersB and C, isolates with no plasmids predominate. Cluster Donly includes isolates with two or more plasmids.

Symbiotic Effectiveness

Evaluation of SE was performed for all 110 isolates(Table 2). SE values range from 0% to 100%. SE trialsrevealed that 40% of the isolates show a SE above 50%.Eleven isolates, which represent about 10% of the totalnumber of isolates, were found to be highly effective infixing N2 in symbiosis with chickpea (SE values above75%); most of these isolates belong to cluster B. The mosteffective isolates were T-1-Telhado from cluster A andG-55-Guarda from cluster B (SE values of 100% and 88%,respectively). Although isolates closer to M. ciceri/M. loti(cluster B) have the highest mean SE (51%), no correlationwas found between SE and species clusters. No correlationwas found between SE and plasmid number, contrary to aprevious study with a smaller set of isolates from Alentejoprovinces [11].

Discussion

The present study is the first survey of chickpea rhizobianative populations covering the Portuguese territory. Onehundred and ten isolates were confirmed as chickpea

symbionts and identified as Mesorhizobium sp., forming amonophyletic cluster with all Mesorhizobium type strains inthe 16S rRNA gene phylogeny.

The four main clusters of the complete 16S rRNA gene-based phylogeny (Fig. 1) show that isolates positioning isscattered within the Mesorhizobium genus. Isolates fromcluster A, which are more related to M. huakuii and M.amorphae than to any other type strain, are the mostabundant chickpea microsymbionts found in Portuguesesoils. This was unexpected since the M. huakuii and M.amorphae type strains are unable to nodulate chickpea. M.huakuii was originally isolated from Astragalus sinicus [43]that does not exist in Portugal. M. amorphae was originallyisolated from Amorpha fruticosa [50], a plant unrelated toC. arietinum, which is uncommon and considered invasivein Portugal. However, it is probable that most isolates fromcluster A belong to a new species. M. ciceri and M.mediterraneum species groups (clusters B and D, respec-tively) could be expected to include the majority of nativeisolates, as these species were described as the specificchickpea microsymbionts [15, 16]. However, only 33% ofthe isolates grouped with these two type strains. Isolatesrelated to M. amorphae (cluster A), M. loti (cluster B), andM. tianshanense (cluster C) were found, as in previousstudies on chickpea rhizobia isolated from Portugal andSpain [12, 20]. The present work screened the entirePortuguese territory, confirmed the high diversity of nativerhizobia, and revealed an unexpected high proportion ofisolates unrelated to M. ciceri and M. mediterraneum.

Table 2 (continued)

Origin Isolate 16S rRNA geneaccession number

Speciescluster

Plasmidnumber

SE (%)

Baixo Alentejo 6b.-Beja AY225381a D 2 [9] 76 [21]

7a.-Beja AY225382a B 2 39 [21]

27-Beja AY225383a B 1 [9] 41 [21]

29-Beja AY225384a D 6 [48] 71 [21]

64b.-Beja AY225385a A 1 [9] 70 [21]

Algarve PM-1-Portimão EU652175 D 2 51

PM-14-Portimão EU652176 D 2 33

PM-17-Portimão EU652120a D 2 84

PMI-1-Portimão EU652177 B 1 80

PMI-6-Portimão EU652121a A 1 81

Madeira SA-4-Serra d’Água EU652122a A 0 63

SA-9-Serra d’Água EU652178 A 0 36

SA-12-Serra d’Água EU652179 A 0 56

SA-17-Serra d’Água EU652180 A 3 16

Province of origin, 16S rRNA gene accession number, species cluster defined from the 16S rRNA gene sequence analysis, plasmid number, andsymbiotic effectiveness (SE) values are indicated for each isolate. All data in bold are results from this study

n.d. not determinedaComplete 16S rRNA gene sequence

936 A. Alexandre et al.

C-1-Coimbra

V-15b-Viseu

V-20-Viseu

C-13-Coimbra

V-5b-Viseu

V-18-Viseu

V-25b-Viseu

C-27b-Coimbra

78-Elvas

85-Elvas

102-Évora

64b-Beja

98-Évora

CV-18-Elvas

PMI-6-Portimão

PT-35-Portalegre

94-Évora

90-Évora

T-8-Telhado

A-8b-Aveiro

SA-4-Serra d’Água

L-19-Leiria

C-3-Coimbra

M. huakuii

C-14-Coimbra

M. plurifarium

M. amorphae

M. septentrionale

STR-2-Santarém

7a-Beja

EE-12-ENMP

EE-7-ENMP

27-Beja

M. ciceri

CV-16-Elvas

S-8-Sintra

M. loti

75-Elvas

PII-3-Porto

M. chacoense

M. albiziae

M. tianshanense

CR-32-Caldas da Rainha

ST-2-Setúbal

93-Évora

STR-14-Santarém

29-Beja

PM-17-Portimão

6b-Beja

EE-14-ENMP

M. mediterraneum

EE-2-ENMP

M. temperatum

EE-29-ENMP

M. thiogangeticum

R. leguminosarum

R. etli

S. medicae

S. meliloti A. caulinodans

B. japonicum

99

99

71

99

61

64

75

44

47

99

99

98

64

95

16

99

24

43

35

90

39

80

45

28

87

72

18

34

76

54

6159

70

51

0.01

A

B

C

D

Figure 1 Dendrogram showingthe phylogenetic relationships ofisolates and type strains, basedon 16S rRNA gene analysis(alignment length 1,357 bp).Neighbor-joining method wasused. Bootstrap values are listedat the nodes. The four mainclusters generated are markedwith letters A to D. The scalebar indicates 1% substitutionsper site

Survey of Chickpea Rhizobia in Portugal 937

To our knowledge, there is only one study addressing thediversity of chickpea rhizobia covering an entire country,which was performed in Morocco [44]. Using PCR-RFLPanalysis of the 16S rRNA gene, the authors found a lowerdiversity of chickpea rhizobia than the one revealed by the

present study, since most isolates were described as close toM. ciceri, M. loti, and M. mediterraneum. However, fourisolates were described as close to Sinorhizobium species.More recently, L’taief and co-workers [45] isolated chick-pea native rhizobia from several regions of Tunisia and

(a)

(b)

Figure 2 Distribution of the110 chickpea rhizobia isolatesby species clusters, as definedfrom the 16S rRNA gene-basedphylogeny. a Map of Portugalshowing the distribution of iso-lates in each province, accordingto their cluster. Pie chartsizes are proportional to thenumber of isolates in eachprovince. b Cluster A—isolatesclose to the type strains ofM. huakuii and M. amorphae,M. plurifarium, and M. septen-trionale (white area). ClusterB—isolates close to the typestrains of M. ciceri andM. loti (light gray area). ClusterC—isolates close to the typestrain of M. tianshanense (darkgray area). Cluster D—isolatesclose to the type strains ofM. mediterraneum andM. temperatum (black area)

plasmidnumber

0

5

10

15

20

25

30

35

Cluster A Cluster B Cluster C Cluster D

num

ber

of is

olat

es

0

1

2

3

4

6

Figure 3 Number of isolates ineach cluster according to theirplasmid number

938 A. Alexandre et al.

found isolates only belonging to either M. ciceri or M.mediterraneum. Probably, the low diversity found inTunisia is related to the history of chickpea cultivation onthe sampled sites, as supported by several studies reportinga decrease in rhizobia diversity associated with the presenceof the host plant [46]. Accordingly, the high diversity foundin Portuguese soils could be explained by the absence ofchickpea crop in Portugal [3] and the non-existence ofchickpea wild relatives [47]; furthermore, there are norecords of the use of commercial inoculants that couldreduce the natural chickpea rhizobia diversity. Interestingly,isolates from the single site where chickpea has beencultivated (Elvas-ENMP) group with M. ciceri or M.mediterraneum.

A previous study on chickpea rhizobia diversity in twoPortuguese provinces revealed a group of isolates apartfrom any type strain, still closer to M. huakuii that couldrepresent a new Mesorhizobium species [12]. Sequenceanalysis of other housekeeping genes confirmed theseparate position of this group of isolates (unpublishedresults). The present study further supports this putativenew species, since 17 new isolates from five differentprovinces were included into this group. Thus, isolates fromcluster A should be further studied in order to investigatetheir species affiliation.

Considering all isolates, an association was foundbetween the province of origin and species clusters ofisolates, suggesting that the geographical distribution is notrandom and that some species are typically found in acertain region (Fig. 2). For instance, most isolates found inthe center of Portugal belong to cluster A. A higherdiversity of isolates species was found in the center–southand south of the country.

A correlation was found between isolates species clusterand origin soil pH (P<0.001), which confirms our previousresults obtained with a smaller set of isolates [48]. Forexample, all isolates assigned to the M. mediterraneum/M.temperatum species cluster D were obtained from the soilswith higher pH values. This may indicate that geneticdeterminants, which allow rhizobia survival in alkaline soilconditions, are species-specific.

Considering the correlation found between speciescluster and soil pH, it is likely that pH is a key environ-mental parameter determining the species geographicdistribution. This hypothesis is supported by wider studiesaddressing soil bacterial communities, which indicate soilpH as the variable that best explains the populationdiversity and overall community composition [49]. Severalstudies in rhizobia showed that the pH affects both survivaland competitiveness in soil, as well as the nodulationprocess [1]. The effect of pH in chickpea rhizobia growthhas been addressed in our previous studies [8, 48]. Theseshowed a positive correlation between maximum growth

pH and origin soil pH, for isolates belonging to the fourspecies clusters (C-1-Coimbra, PT-35-Portalegre, and 64b.-Beja from cluster A; 75-Elvas from cluster B; CR-32-Caldas da Rainha from cluster C; 29-Beja from cluster D)[48]. In addition, using three isolates, a higher symbioticeffectiveness was achieved using watering solution at a pHvalue closest to the bacterial maximum growth pH [48].Altogether these studies suggest that pH is a key environ-mental factor for rhizobia population composition, actingon bacteria, both free-living and in symbiosis.

In each 16S rRNA gene-based cluster, isolates with highand low symbiotic effectiveness were found. A large set ofisolates with very high SE values (above 75%) are goodcandidates for field inoculation. Many of these isolates arefrom the M. ciceri cluster (B). Interestingly, the mosteffective isolate (T-1-Telhado) is closer to M. huakuii(cluster A), which is not a chickpea microsymbiont. About70% of the isolates from cluster B present a SE value abovethe corresponding type strain M. ciceri, which showed aSE of 41%, estimated in a previous study [21]. In theM. mediterraneum/M. temperatum cluster (D), 56% of theisolates showed a SE above 39%, which is the valuedescribed for the type strain of M. mediterraneum [21].

The plasmid number of rhizobia isolates was found to beassociated with species cluster, suggesting that this featuremight be species constrained. In most isolates from clusterA, one plasmid was detected, similar to M. amorphae [50].In cluster D, both isolates and M. mediterraneum [13]showed more than one plasmid. All cluster C isolates,except for two, lack plasmids, similar to M. tianshanense[51]. Isolates from cluster B seem to be more diverse inplasmid number, including isolates with zero, one, and twoplasmids. The type strain of M. ciceri (cluster B) harborsone plasmid [13].

Contrary to previous studies, which suggested that mostrhizobia-nodulating chickpea are M. ciceri and M. medi-terraneum, this wider survey shows a predominance ofother species. The obtained collection of isolates, highlydiverse in terms of species, as well as SE, provides animportant source of rhizobia strains to be used, namely aspotential inoculants. The present study is the first system-atic assessment of C. arietinum microsymbionts in Portugaland contributes to clarify the biogeography of chickpearhizobia, providing a global picture of how species aredistributed across the country.

Acknowledgments This work was supported by the programPOCTI (POCTI/BME/44140/2002) from FCT (Fundação para aCiência e a Tecnologia) and co-financed by EU-FEDER. A. Alexandreand C. Brígido acknowledge Ph.D. fellowships from FCT (SFRH/BD/18162/2004 and SFRH/BD/30680/2006). M. Laranjo acknowledges apost-doctoral fellowship from FCT (SFRH/BPD/27008/2006). Theauthors thank L. S. Dias for suggestions on the statistical analysis andG. Mariano for technical assistance.

Survey of Chickpea Rhizobia in Portugal 939

References

1. Zahran HH (1999) Rhizobium–legume symbiosis and nitrogenfixation under severe conditions and in an arid climate. MicrobiolMol Biol Rev 63:968–989

2. FAO (2003) Production year book, 2002. FAO, Rome, Italy3. Duarte Maçãs IMV (2003) Selecção de linhas de grão de bico

(Cicer arietinum L.) adaptadas ao ambiente Mediterrânico-critérios morfológicos e fisiológicos. Universidade de Évora,Évora, p 171

4. Aouani ME, Mhamdi R, Jebara M, Amarger N (2001) Charac-terization of rhizobia nodulating chickpea in Tunisia. Agronomie21:577–581

5. Kuykendall LD, Gaur YD, Dutta SK (1993) Genetic diversityamong Rhizobium strains from Cicer arietinum L. Lett ApplMicrobiol 17:259–263

6. Maâtallah J, Berraho EB, Muñoz S, Sanjuan J, Lluch C (2002)Phenotypic and molecular characterization of chickpea rhizobiaisolated from different areas of Morocco. J Appl Microbiol 93:531–540

7. Nour SM, Cleyet-Marel JC, Beck D, Effosse A, Fernandez MP(1994) Genotypic and phenotypic diversity of Rhizobium isolatedfrom chickpea (Cicer arietinum L.). Can J Microbiol 40:345–354

8. Rodrigues C, Laranjo M, Oliveira S (2006) Effect of heat and pHstress in the growth of chickpea mesorhizobia. Curr Microbiol53:1–7

9. Laranjo M, Rodrigues R, Alho L, Oliveira S (2001) Rhizobia ofchickpea from southern Portugal: symbiotic efficiency and geneticdiversity. J Appl Microbiol 90:662–667

10. Alexandre A, Laranjo M, Oliveira S (2006) Natural populations ofchickpea rhizobia evaluated by antibiotic resistance profiles andmolecular methods. Microb Ecol 51:128–136

11. Laranjo M, Branco C, Soares R, Alho L, Carvalho M, Oliveira S(2002) Comparison of chickpea rhizobia isolates from diversePortuguese natural populations based on symbiotic effectivenessand DNA fingerprint. J Appl Microbiol 92:1043–1050

12. Laranjo M, Machado J, Young JPW, Oliveira S (2004) Highdiversity of chickpea Mesorhizobium species isolated in aPortuguese agricultural region. FEMS Microbiol Ecol 48:101–107

13. Cadahía E, Leyva A, Ruiz-Argüeso T (1986) Indigenous plasmidsand cultural characteristics of rhizobia nodulating chickpeas(Cicer arietinum L.). Arch Microbiol 146:239–244

14. Jarvis BDW, van Berkum P, Chen WX, Nour SM, Fernandez MP,Cleyet-Marel JC, Gillis M (1997) Transfer of Rhizobium loti,Rhizobium huakuii, Rhizobium ciceri, Rhizobium mediterraneum,and Rhizobium tianshanense to Mesorhizobium gen. nov. Int JSyst Bacteriol 47:895–898

15. Nour SM, Fernandez MP, Normand P, Cleyet-Marel J-C (1994)Rhizobium ciceri sp. nov., consisting of strains that nodulatechickpeas (Cicer arietinum L.). Int J Syst Bacteriol 44:511–522

16. Nour SM, Cleyet-Marel J-C, Normand P, Fernandez MP (1995)Genomic heterogeneity of strains nodulating chickpeas (Cicerarietinum L.) and description of Rhizobium mediterraneumsp. nov. Int J Syst Bacteriol 45:640–648

17. Downie JA (1998) Functions of rhizobial nodulation genes. In:Spaink HP, Kondorosi A, Hooykaas PJJ (eds) The Rhizobiaceaemolecular biology of model plant-associated bacteria. KluwerAcademic, Dordrecht, pp 387–402

18. Broughton WJ, Perret X (1999) Genealogy of legume-Rhizobiumsymbioses. Curr Opin Plant Biol 2:305–311

19. Perret X, Staehelin C, Broughton WJ (2000) Molecular basis ofsymbiotic promiscuity. Microbiol Mol Biol Rev 64:180–201

20. Rivas R, Laranjo M, Mateos PF, Oliveira S, Martinez-Molina E,Velázquez E (2007) Strains of Mesorhizobium amorphaeand Mesorhizobium tianshanense, carrying symbiotic genes of

common chickpea endosymbiotic species, constitute a novelbiovar (ciceri) capable of nodulating Cicer arietinum. Lett ApplMicrobiol 44:412–418

21. Laranjo M, Alexandre A, Rivas R, Velázquez E, Young JPW,Oliveira S (2008) Chickpea rhizobia symbiosis genes are highlyconserved across multipleMesorhizobium species. FEMS MicrobiolEcol 66:391–400

22. Donate-Correa J, León-Barrios M, Hernández M, Pérez-GaldonaR, del Arco-Aguilar M (2007) Different Mesorhizobium speciessharing the same symbiotic genes nodulate the shrub legumeAnagyris latifolia. Syst Appl Microbiol 30:615–623

23. Jarabo-Lorenzo A, Perez-Galdona R, Donate-Correa J, Rivas R,Velázquez E, Hernandez M, Temprano F, Martinez-Molina E,Ruíz-Argüeso T, Leon-Barrios M (2003) Genetic diversity ofbradyrhizobial populations from diverse geographic origins thatnodulate Lupinus spp. and Ornithopus spp. Syst Appl Microbiol26:611–623

24. Michiels J, Dombrecht B, Vermeiren N, Xi C, Luyten E,Vanderleyden J (1998) Phaseolus vulgaris is a non-selective hostfor nodulation. FEMS Microbiol Ecol 26:193–205

25. Menna P, Hungria M, Barcellos FG, Bangel EV, Hess PN,Martínez-Romero E (2006) Molecular phylogeny based on the16S rRNA gene of elite rhizobial strains. Syst Appl Microbiol29:315–332

26. Alexandre A, Laranjo M, Young JPW, Oliveira S (2008) dnaJ is auseful phylogenetic marker for alphaproteobacteria. Int J SystEvol Microbiol 58:2839–2849

27. Young JM, Park DC (2007) Relationships of plant pathogenicenterobacteria based on partial atpD, carA, and recA as individualand concatenated nucleotide and peptide sequences. Syst ApplMicrobiol 30:343–354

28. Martens M, Delaere M, Coopman R, De Vos P, Gillis M, WillemsA (2007) Multilocus sequence analysis of Ensifer and related taxa.Int J Syst Evol Microbiol 57:489–503

29. Coenye T, Gevers D, Peer YV, Vandamme P, Swings J (2005)Towards a prokaryotic genomic taxonomy. FEMS Microbiol Rev29:147–167

30. Clayton R, Sutton G, Hinkle P Jr, Bult C, Fields C (1995)Intraspecific variation in small-subunit rRNA sequences inGenBank: why single sequences may not adequately representprokaryotic taxa. Int J Syst Bacteriol 45:595–599

31. van Berkum P, Elia P, Eardly BD (2006) Multilocus sequencetyping as an approach for population analysis of Medicago-nodulating rhizobia. J Bacteriol 188:5570–5577

32. Sun L, Qiu FB, Zhang XX, Dai X, Dong XZ, Song W (2008)Endophytic bacterial diversity in rice (Oryza sativa L.) rootsestimated by 16S rDNA sequence analysis. Microb Ecol 55:415–424

33. Gevers D, Coenye T (2007) Phylogenetic and genomic analysis.In: Hurst CJ, Crawford RL, Garland JL, Lipson DA, Milles AL,Stetzenbach LD (eds) Manual of environmental microbiology, volII. ASM, Washington, p 1352

34. Somasegaran P, Hoben HJ (1994) Handbook for rhizobia.Springer, New York

35. Young JP, Downer HL, Eardly BD (1991) Phylogeny of thephototrophic Rhizobium strain BTAi1 by polymerase chainreaction-based sequencing of a 16S rRNA gene segment. JBacteriol 173:2271–2277

36. Hall TA (1999) BioEdit: a user-friendly biological sequencealignment editor and analysis program for Windows 95/98/NT.Nucleic Acids Symp Ser 41:95–98

37. Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W:improving the sensitivity of progressive multiple sequence.Nucleic Acids Res 22:4673–4680

38. Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: molecularevolutionary genetics analysis (MEGA) software version 4.0. MolBiol Evol 24:1596–1599

940 A. Alexandre et al.

39. Vincent JM (1970) A manual for the practical study of root-nodule bacteria. Blackwell Scientific, Oxford

40. Gibson AH (1987) Evaluation of nitrogen fixation by legumes in thegreenhouse and growth chamber. In: Elkan GH (ed) Symbioticnitrogen fixation technology.Marcel Dekker, NewYork, pp 321–363

41. Louvrier P, Laguerre G, Amarger N (1996) Distribution ofsymbiotic genotypes in Rhizobium leguminosarum biovar viciaepopulations isolated directly from soils. Appl Environ Microbiol62:4202–4205

42. Benzécri JP (1973) Analyse des données. Tome I: Analyse descorrespondances. Tome II: La Classification. Dunod, Paris

43. Chen WX, Li GS, Qi YL, Wang ET, Yuan HL, Li JL (1991)Rhizobium huakuii sp. nov. isolated from the root-nodules ofAstragalus sinicus. Int J Syst Bacteriol 41:275–280

44. Maâtallah J, Berraho E, Sanjuan J, Lluch C (2002) Phenotypiccharacterization of rhizobia isolated from chickpea (Cicerarietinum) growing in Moroccan soils. Agronomie 22:321–329

45. L'Taief B, Sifi B, Gtari M, Zaman-Allah M, Lachaal M (2007)Phenotypic and molecular characterization of chickpea rhizobiaisolated from different areas of Tunisia. Can J Microbiol 53:427–434

46. Coutinho HLD, Kay HE, Manfio GP, Neves MCP, Ribeiro JRA,Rumjanek NG, Beringer JE (1999) Molecular evidence for shifts

in polysaccharide composition associated with adaptation ofsoybean Bradyrhizobium strains to the Brazilian Cerrado soils.Environ Microbiol 1:401–408

47. Talavera S, Aedo C, Castroviejo S, Romero Zarco C, Sáez L,Salgueiro FJ, Velayos M (1999) Flora iberica-Plantas vascularesde la Península Ibérica e Islas Baleares. Real Jardín Botánico,CSIC, Madrid, Spain

48. Brígido C, Alexandre A, Laranjo M, Oliveira S (2007) Moder-ately acidophilic mesorhizobia isolated from chickpea. Lett ApplMicrobiol 44:168–174

49. Fierer N, Jackson RB (2006) The diversity and biogeography ofsoil bacterial communities. Proc Natl Acad Sci U S A 103:626–631

50. Wang ET, van Berkum P, Sui XH, Beyene D, Chen WX,Martínez-Romero E (1999) Diversity of rhizobia associatedwith Amorpha fruticosa isolated from Chinese soils and descrip-tion of Mesorhizobium amorphae sp. nov. Int J Syst Bacteriol49:51–65

51. Chen WX, Wang E, Wang SY, Li YB, Chen XQ (1995)Characteristics of Rhizobium tianshanense sp. nov., a moderatelyand slowly growing root-nodule bacterium isolated from an aridsaline environment in Xinjiang, People’s Republic of China. Int JSyst Bacteriol 45:153–159

Survey of Chickpea Rhizobia in Portugal 941