This page is a quick conversion to text format, for figures, tables, and references, see the pdf file.
Collection and isolation of bacterial strains
Collection
Bacterial strains used in this study (Tables t-Complete-native, t-Complete-exotic, t-Complete-type) were either directly isolated from the root nodules of wild plants, or obtained from the extensive collection in the ICMP (International Collection of Micro-organisms from Plants, Landcare Research, Auckland, New Zealand)A searchable catalog of strains is available at: The ICMP
Nodules of native legumes were obtained from throughout the country from pristine sites on conservation lands that were distant from agricultural plantings. Introduced legume nodules were obtained from arable lands or from conservation lands. Sample locations are shown in Figure p-Isolates. Young seedlings were preferred as the nodules were easier to locate than on mature plants, and older nodules are possibly contaminated by bacteria invading the nodule. Plant roots containing several nodules from each plant were sealed in plastic bags for transport (each plant in a separate bag). Bacteria were generally isolated the same day, if not, bags were stored at 4 .
figure [width=12cm]isolates-group [Geographical distribution of rhizobial strains]Map of New Zealand showing geographical distribution of strains used in this study. The genus of the isolate is indicated by the shape of marker. Letter inside marker indicates genomic group. p-Isolates figure
sidewaystable [p] Strains isolated from native legumes and experiments performed tabularlllccccccc ICMP & Genus &Isolated & 7cExperiment performed 4-10 Number & & From & 16S & atpD & glnII & recA & nodA & Biolog & FAME 11727 & Rhizobium & Carmichaelia australis & & --- & --- & --- & & --- & --- 12687 & Rhizobium & Carmichaelia australis & & --- & --- & --- & & --- & 13190 & Mesorhizobium & Carmichaelia australis & & & & & & &
15054 & Mesorhizobium & Carmichaelia australis & & & --- & & & & --- 14324 & Mesorhizobium & Carmichaelia crassicaule & --- & --- & --- & --- & --- & --- & 11708 & Mesorhizobium & Carmichaelia nana & & & & & & & 11722 & Mesorhizobium & Carmichaelia nana & & & & & & --- & 14319 & Mesorhizobium & Carmichaelia odorata & & & & & & & 12635 & Mesorhizobium & Carmichaelia petriei & & & & & --- & & --- 12649 & Mesorhizobium & Carmichaelia petriei & & --- & --- & --- & & --- & 11541 & Mesorhizobium & Clianthus puniceus & & & & & & & 11542 & Rhizobium & Clianthus puniceus & & --- & --- & --- & & --- & 11720 & Mesorhizobium & Clianthus puniceus & & --- & --- & --- & & --- & --- 11721 & Mesorhizobium & Clianthus puniceus & & & --- & --- & & & 11726 & Mesorhizobium & Clianthus puniceus & & & --- & --- & & & 12680 & Mesorhizobium & Clianthus puniceus & & --- & --- & --- & & & --- 12685 & Mesorhizobium & Montigena novae-zelandiae & & & & & & &
12690 & Mesorhizobium & Montigena novae-zelandiae & & --- & --- & --- & & & 14642 & Rhizobium & Sophora chathamica & & & & & & & 11736 & Mesorhizobium & Sophora microphylla & & & --- & & & &
12637 & Mesorhizobium & Sophora microphylla & & --- & --- & --- & --- & --- & 14330 & Mesorhizobium & Sophora microphylla & & & & & & & 11719 & Mesorhizobium & Sophora tetraptera & & & & & & & 12642 & Mesorhizobium & Soil & & --- & --- & --- & --- & --- & --- tabular
t-Complete-native sidewaystable
sidewaystable [p] Strains isolated from exotic legumes and experiments performed tabularlllccccccc ICMP & Genus &Isolated & 7cExperiment performed 4-10 Number & & From & 16S & atpD & glnII & recA & nodA & Biolog & FAME 12835 & Bradyrhizobium sp. & Acacia dealbata & & & & & & --- & --- 14754 & Bradyrhizobium sp. & Acacia longifolia & & & & & & --- & --- 14755 & Bradyrhizobium sp. & Acacia longifolia & & & & & & --- & 14752 & Bradyrhizobium sp. & Albizia julibrissin & & & & & & --- & --- 14753 & Bradyrhizobium sp. & Albizia julibrissin & & & & & & --- & --- 12624 & Bradyrhizobium sp. & Cytisus scoparius & & --- & --- & --- & & --- & --- 14291 & Bradyrhizobium sp. & Cytisus scoparius & & & & & & &
14309 & Bradyrhizobium sp. & Cytisus scoparius & & --- & --- & --- & --- & --- & --- 14310 & Bradyrhizobium sp. & Cytisus scoparius & & --- & --- & --- & --- & --- & --- 14328 & Bradyrhizobium sp. & Cytisus scoparius & & --- & --- & --- & --- & --- & --- 12674 & Bradyrhizobium sp. & Ulex europaeus & & --- & --- & --- & & --- & 14292 & Bradyrhizobium sp. & Ulex europaeus & & --- & --- & --- & --- & --- & --- 14304 & Bradyrhizobium sp. & Ulex europaeus & & --- & --- & --- & --- & --- & 14306 & Bradyrhizobium sp. & Ulex europaeus & & --- & --- & --- & & --- & --- 14320 & Bradyrhizobium sp. & Ulex europaeus & & --- & --- & --- & --- & --- & --- 14533 & Bradyrhizobium sp. & Ulex europaeus & & & & & & --- &
tabular
t-Complete-exotic sidewaystable
table Rhizobial type strains and experiments performed tabularllccccccc ICMP & Type & 7cExperiment performed 3-9 Number & Strain & 16S & atpD & glnII & recA & nodA & Biolog & FAME
15022 & M. amorphae & --- & & & & --- & & 14587 & M. chacoense & --- & & & & --- & & 13641 & M. ciceri & --- & --- & --- & --- & --- & & 11069 & M. huakuii & --- & --- & --- & --- & --- & & 4682 & M. loti & --- & --- & --- & --- & --- & & 13644 & M. mediterraneum & --- & --- & --- & --- & --- & --- & 13640 & M. plurifarium & --- & & & & & & --- & M. temperatum & --- & --- & --- & --- & --- & --- & --- 13645 & M. tianshanense & --- & --- & --- & --- & --- & & --- & M. septentrionale & --- & --- & --- & --- & --- & --- & ---
13642 & R. etli & --- & --- & --- & --- & --- & & 13643 & R. galegae & --- & --- & --- & --- & --- & --- & 13690 & R. hainanense & --- & --- & --- & --- & --- & --- & 13551 & R. huautlense & --- & --- & --- & --- & --- & --- & 13689 & R. leguminosarum & --- & --- & --- & --- & --- & & --- 13688 & R. mongolense & --- & --- & --- & --- & --- & --- & 13646 & R. tropici & --- & --- & --- & --- & --- & & 11139 & E. fredii & --- & --- & --- & --- & --- & & 13798 & E. medicae & --- & --- & --- & --- & --- & --- & 12623 & E. meliloti & --- & --- & --- & --- & --- & & 13648 & E. saheli & --- & --- & --- & --- & --- & & --- 13649 & E. terangae & --- & --- & --- & --- & --- & --- & --- & B. canariense & --- & --- & --- & --- & --- & --- & --- 13638 & B. elkanii & --- & & & --- & --- & & 2864 & B. japonicum & --- & --- & --- & --- & --- & & --- 13639 & B. liaoningense & --- & & & & --- & & --- --- & B. yuanmingense & --- & --- & --- & --- & --- & --- & --- tabular
t-Complete-type table
Bacterial isolation
Root material was washed in running tap water to remove adherent soil. Individual nodules were dissected from the roots, using a flame sterilised scalpel and tweezers---if nodules were very small, a little root tissue was left attached either side of the nodule. Nodules were washed thoroughly in RO water and the non-ionic surfactant `Tween 80' (100 LL) to remove all traces of soil. The nodules were then transferred to a sterile petri dish and surface sterilised by immersion in 10 mL of a 5 solution of commercial sodium hypochlorite (final concentration: 0. 25 gL NaOCl) and Tween 80 (10 LL), for 10--30 minutes depending on the nodule size.
Individual nodules were rinsed once in sterile mQ water and crushed with the flattened end of a flamed glass rod. The exudate was aseptically streaked onto surface-dried Yeast Mannitol Agar (YMA) plates (Table t-medium).
table [b] Bacterial medium composition
tabularll Name & Composition per litre YMA & 10 g yeast, 10 g mannitol, 2. 5 g peptone, 15 g agar YMA+Ca & YMA + 3 g Calcium carbonate, 1. 5 g Calcium gluconate R2A & 18 g R2A powder (Difco) TY-M & 5 g Tryptone, 3 g yeast extract, 0. 87 g CaCl. 2HO, 15 g agar tabular t-medium table
Purification and storage
Agar plates were incubated at 26 for between three and ten days. Individual colonies appearing over this period were re-streaked onto YMA plates, and sub-cultured onto YMA+Ca (Table t-medium) slopes in test tubes for short-term storage at 8 . Strains used and reported in this study were deposited in the ICMP (where cultures are permanently stored under liquid nitrogen).
Several agar formulations were tested for the ability to grow rhizobia. The best media for growth were YMA, YMA+Ca, and R2A. YMA was subsequently used for isolation, while YMA+Ca was used for growth prior to storage, as the additions to this medium prolonged the life of the culture by neutralising acid production. R2A medium (Difco) was used for preparation of subcultures for DNA extraction and inoculant preparation, as organisms grown on this medium produced less extracellular polysaccharide.
DNA extraction
DNA was extracted from bacterial cultures using a SDS/CTAB lysis and phenol/chloroform extraction method Ausubel87. Bacterial cultures were grown on R2A plates, at 26 until 2--3 mm colonies were visible. A small amount (10 L) of fresh cell mass was taken from multiple colonies to minimise the possible influence of mutation in a single colony, and placed in a 1. 5mL microfuge tube containing 567 L of TE, 30 L of SDS (10 w/v), and 3 L of 20 mgmL proteinase K. The mixture was mixed by vortexing and incubated for one hour at 37 . 100 L of 5 M NaCl was then added to the tube and mixed by inversion. Then 80 L of CTAB/NaCl solution (10 CTAB in 0. 7 M NaCl) was added, mixed by inversion, and incubated for 10 minutes at 65 . This solution was extracted with an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1) and centrifuged for 5 minutes at 19000 RCF. The aqueous phase was transferred to a fresh tube and extracted with chloroform/isoamyl alcohol (25:1), then centrifuged for 5 minutes at 19000 RCF. The supernatant was transferred to a fresh tube and precipitated with 0. 6 volumes of 100 isopropanol for 15 minutes, then the DNA pelleted by centrifugation for 5 minutes at 19000 RCF. The supernatant was removed, and the DNA pellet washed with 250 L of 70 ethanol and centrifuged again. All supernatant was removed by aspiration and the DNA pellet air-dried for 30 minutes at room temperature. The pellet was resuspended in 100 L of TE and stored at --20 .
The quality of the DNA was assessed by analysis on a Biospec-mini spectrophotometer (Shimadzu) at wavelengths of 260 and 280 nm to determine concentration and degree of any protein contamination Rodriguez83. Genomic DNA was run on a 0. 5 agarose gel to assess any shearing. If protein contamination or DNA shearing was significant, DNA was re-extracted from fresh culture.
Primers for PCR amplification
Six genes were targeted for PCR (Polymerase Chain Reaction) amplification and DNA sequencing. These were the small subunit rRNA gene (16S rRNA or rrn), ATP synthaseSynthase enzymes are also known as `synthetase' enzymes. `synthase' sensu lato is used in this thesis IUPAC-IUB. beta-subunit (atpD), glutamine synthase I (glnI), glutamine synthase II (glnII), DNA recombinase A (recA), and acyl transferase nodulation protein A (nodA). glnI sequences were unable to be amplified from a number of isolates and therefore DNA sequence data were not obtained for this gene.
sidewaystable PCR primers for `housekeeping' genes tabularllcl Primer & Sequence & Target gene & Reference 16S-1F & AGCGGCGACGGGTGAGTAATG & & Normand95 16S-485F & CAGCAGCCGCGGTAA & & Normand95 16S-1100R & GGGTTGCGCTCGTTG & 16S rRNA & Normand95 16S-1509R & AAGGAGGGGATCCAGCCGCA & & Normand95 16S-PB36 & AGRGTTTGATCMTGGCTCAG & & Bell02 1-4 atpD-273F & SCTGGGSCGYATCMTGAACGT & ATP synthase & Gaunt01 atpD-771R & GCCGACACTTCCGAACCNGCCTG & beta-subunit & Gaunt01 atpD-294F & ATCGGCGAGCCGGTCGACGA & (atpD) & Gaunt01 1-4 GSII-1 & AACGCAGATCAAGGAATTCG & & Turner00 GSII-2 & ATGCCCGAGCCGTTCCAGTC & Glutamine & Turner00 GSII-3 & AGRTYTTCGGCAAGGGYTC & synthase II & Turner00 GSII-4 & GCGAACGATCTGGTAGGGGT & (glnII)& Turner00 1-4 GSI-1 & AAGGGCGGCTAYTTCCCGGT & & Turner00 GSI-2 & GTCGAGACCGGCCATCAGCA & Glutamine & Turner00 GSI-3 & GAYCTGCGYTTYACCGACC & synthase I & Turner00 GSI-4 & CTTCRTGGTGRTGCTTTTC & (glnI) & Turner00 1-4 recA-6F & CGKCTSGTAGAGGAYAAATCGGTGGA & & Gaunt01 recA-555R & CGRATCTGGTTGATGAAGATCACCAT & DNA & Gaunt01 recA-63F & ATCGAGCGGTCGTTCGGCAAGGG & recombinase A & Gaunt01 recA-504R & TTGCGCAGCGCCTGGCTCAT & (recA)& Gaunt01 recA-BF & CGTACTGTCGAAGGTTCTTCCATGGA & & This study tabular
t-Primers Symbols: A, C, G, T -- standard nucleotides; Y: C,T; R: A,G; W: A,T; S: G,C; K: T,G; M: C,A; D: A,T,G; V: A,C,G; B: T,C,G; N: all. sidewaystable
Most primers used were taken from previous studies, eliminating the need to design and optimise new primers. Additionally this provided sequences in the GenBank public database that were useful to compare with the New Zealand strains.
All oligonucleotide primers were manufactured by Invitrogen (Auckland), dry lyophilised DNA pellets were received at the laboratory and were reconstituted to a stock concentration of 100 M. The stock was diluted 10 fold to the working primer solution of 10 M. All primer solutions were stored at --20 .
Oligonucleotide primer sequences and sources are shown in Tables t-Primers and t-nodA-Primers.
16S rRNA gene amplification and sequencing
The 16S rRNA gene was PCR-amplified using primer pairs 16S-1F/16S-1509R or 16S-PB36/16S-1509R, yielding 1300-bp and 1400-bp products respectively. PCR products were sequenced using the same forward and reverse primers and internal primers 16S-485F and 16S-1100R (Table t-Primers).
atpD amplification and sequencing
The atpD gene was amplified using primer pairs atpD-273F/atpD-771R, yielding a 540-bp product. PCR products were sequenced using atpD-294F/atpD-771R (Table t-Primers).
glnI amplification and sequencing
The glnI (glnA) gene was amplified in two sections. Primer pairs GSI-1/GSI-2 amplified the first section yielding a 612 bp product, and primer pairs GSI-3/GSI-4 amplified the second section yielding a 586-bp product. Sections overlapped by 96 base pairs. PCR products were sequenced using GSI-1/GSI-2 and GSI-3/GSI-4 respectively (Table t-Primers).
Amplification was not initially successful, except for strain ICMP 15054. This was sequenced and deposited in GenBank as AY941195. There are no results for other strains. Later experiments revealed that the final PCR conditions for the glnII gene amplified most glnI sequences.
glnII amplification and sequencing
The glnII gene was amplified in two sections. Primer pairs GSII-1/GSII-2 amplified the first section yielding a 618-bp product, and primer pairs GSII-3/GSII-4 amplified the second section yielding a 440-bp product. Sections overlapped by 109 base pairs. PCR products were sequenced using GSII-1/GSII-2 and GSII-3/GSII-4 respectively (Table t-Primers).
recA amplification and sequencing
The recA gene was PCR amplified using primer pairs recA-6F/recA-555R or recA-BF/recA-555R, yielding a 602-bp product. PCR products were sequenced using recA-63F/recA-504R (Table t-Primers).
recA PCR products were unable to be amplified for Bradyrhizobium strains using the recA-6F/recA-555R primer pair. After aligning the sequences, it was found that the forward primer (recA-6F) was located in a variable region of the gene with poor homology to Bradyrhizobium sequences. A new forward primer, recA-BF, was designed based on the recA sequence from the related bacterium Rhodopseudomonas palustris (GenBank accession D84467). R. palustris is an Alphaproteobacterium, in the Bradyrhizobiaceae family along with Bradyrhizobium. At the time of this experiment this was the closest match in the database and the full genome sequence (including recA) of R. palustris was available on GenBank Larimer04. With the new primer (recA-BF), recA PCR products were obtained for all New Zealand Bradyrhizobium strains and the type strain of B. liaoningense, but no PCR product was obtained from B. elkanii or B. japonicum type strains.
nodA amplification and sequencing
Selecting appropriate primers for the nodA gene proved to be a significant obstacle, twelve different primers were trialed, using six different PCR conditions, in many different combinations, of which only three yielded PCR products suitable for sequencing (shown in Table t-nodA-Primers). The optimal PCR conditions are shown in Table PCR-nodAcycle.
table [b] [nodA PCR primers]Primers for PCR amplification of the nodA gene tabularlll Primer & Sequence & Reference nodA1 & TGCRGTGGAARNTRNNCTGGGAAA & Haukka98 nodA2 & GGNCCGTCRTCRAAWGTCARGTA & Haukka98 nodA3 & TCATAGCTCYGRACCGTTCCG & Zhang00 TSnodD1-1a & CAGATCNAGDCCBTTGAARCGCA & Moulin04 TSnodB1 & AGGATAYCCGTCGTGCAGGAGCA & Moulin04 TSnodA2 & GCTGATTCCAAGBCCYTCVAGATC & Moulin04 TSnodA3 & AGYTGGTCYGGTGCDMGRCCNGA & Moulin04 tabular
t-nodA-Primers 0. 5cm For symbols see Table t-Primers table
A possible explanation is that the nodA gene is involved in the symbiosis process, and is under direct positive selection, thus is less conserved than housekeeping genes. Additionally there is recombination around this area of the genome (see section s-nodA-intro) which meant that it was difficult to design flanking primers.
The primer pair nodA1 and nodA2 amplified type 3, 4 and 5 nodA genes from Mesorhizobium and Rhizobium sequences yielding a 600-bp product (see Fig. p-nodA-MB). Primer pair nodA1 and nodA3 amplified type 1 and 6 nodA genes yielding a 600-bp product. Primer pair TSnodD1-1a and TSnodB1 amplified Bradyrhizobium nodA types 7 and 8, yielding a 1600--2000-bp product.
For cycle sequencing of Mesorhizobium and Rhizobium strains, the PCR primer was used as the cycle sequencing primer. For Bradyrhizobium sequences, TSnodA2 and TSnodA3 were used as internal cycle sequencing primers.
PCR conditions
The Polymerase Chain Reaction was used to amplify gene fragments for sequencing.
All PCRs were performed using the Applied Biosystems AmpliTaq Gold DNA polymerase kit, and Roche dNTPs. Although published protocols were available for most primer pairs, these needed to be optimised for use with this amplification kit. In particular longer extension times were required for sufficient yield of larger products. Optimised PCR conditions are listed in Tables PCR-16Scycle--PCR-nodAcycle.
Each PCR was set up according to Table t-Mastermix, in a total individual volume of 25 L (multiple reactions were done with a master mix). Each set of reactions included a negative control consisting of the same reagent as Table t-Mastermix, but the genomic DNA was replaced with sterile mQ water. PCR amplifications were performed with an Applied Biosystems 9700 thermal cycler.
table PCR Components for a single reaction tabularlc Component & Amount (L) GeneAmp 10 PCR Buffer II & 2. 5 2 mM dNTPs & 2. 5 25 mM MgCl & 2. 0 10 M forward primer & 0. 5 10 M reverse primer & 0. 5 50 ng genomic DNA & 1. 0 AmpliTaq Gold & 0. 3 Sterile mQ HO & 15. 7 2-2 & 25 tabular
t-Mastermix Buffer, MgCl, and AmpliTaq are from the Applied Biosystems AmpliTaq Gold DNA polymerase kit. table
table 16S rRNA PCR cycle tabularccc Temperature & Time & 95 & 5 mins & 1 Hold 95 & 45 s &
6353 & 45 s & 20 Cycles
72 & 90 s & 95 & 45 s &
53 & 45 s & 15 Cycles
72 & 90 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-16Scycle table
table atpD PCR cycle tabularccc Temperature & Time & 95 & 5 mins & 1 Hold 95 & 45 s &
60 & 60 s & 15 Cycles
72 & 45 s & 95 & 45 s &
6050 & 60 s & 20 Cycles
72 & 45 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-atpDcycle table
table glnII PCR cycle tabularccc Temperature & Time & 95 & 4 mins & 1 Hold 95 & 45 s &
63 & 30 s & 35 Cycles
72 & 40 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-glnIIcycle table
table recA PCR cycle tabularccc Temperature & Time & 95 & 5 mins & 1 Hold 95 & 30 s &
55 & 20 s & 35 Cycles
72 & 40 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-recAcycle table
table nodA PCR cycle tabularccc Temperature & Time & 95 & 4 mins & 1 Hold 95 & 45 s &
49 & 60 s & 35 Cycles
72 & 120 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-nodAcycle table
Gel electrophoresis
DNA was resolved by agarose gel electrophoresis in TAE buffer followed by staining with ethidium bromide, and visualisation under UV light. PCR products were resolved in 1 gels, and 0. 5 gels were used for analysing genomic DNA.
5 L of PCR product or 2 L genomic DNA was mixed with 2 L of loading dye (0. 25 Bromophenol blue, 0. 25 xylene cyanol, 25 ficoll (type 400)), and pipetted into the wells. 300 ng of 1kb Plus DNA ladder (Invitrogen) was used to determine the size of the PCR products. All gels were run at 5 Vcm for 40 minutes which allowed good separation of the bands. Gels were stained with ethidium bromide (120 gL) for 10 minutes, and then destained for 20 minutes in RO water to reduce background fluorescence. The stained gel was visualised under UV light in a Bio-Rad Gel Doc 2000 transilluminator. Images were captured, cropped, and printed with Gel Doc software version 4. 3. 1.
Sequencing of PCR products
PCR products were column-purified with a Roche High Pure PCR Product Purification Kit, following the manufactures instructions HighPure. Purified products were quantified on a Shimadzu Biospec-mini spectrophotometer.
Purified PCR products were cycle sequenced in both directions with the appropriate primers using BigDye Terminator Ready Reaction Mix (ABI) (version 3. 0 or 3. 1) and an ABI PRISM 310 (or 3100) Avant Genetic Analyzer located at landcare Research, Auckland. Sequences were assembled and edited with Sequencher 3. 11 (Gene Codes Corp. ).
Phylogenetic analysis
Alignments
Nucleotide alignments were initially constructed with ClustalX 1. 83 Thompson97 using the default gap opening and extension penalties (GO:10 GE:0. 2). Alignments were then manually edited with GeneDoc 2. 6. 02 Nicholas97 to ensure that alignment gaps in protein coding genes did not cause errors in the amino acid sequence.
The four primers for glnII amplify two overlapping sections. In cases where one of the two sequences successfully amplified, missing data were replaced with the symbol `?'. All sequences were checked for possible chimeras using the Bellerophon server Huber04This server may be accessed on the internet at: http://foo.maths.uq.edu.au/huber/bellerophon.pl , although none were found.
GenBank sequences from the type strains of representative species from Mesorhizobium, Bradyrhizobium, Rhizobium and Ensifer were also included for comparison (Table t-GenBank-type). The outgroup for each alignment of 16S rRNA, atpD and recA was the appropriate homologous sequence from Caulobacter crescentus strain CB15, obtained from the complete genome (GenBank accession AE005673). This species was selected for its evolutionary distance from the Rhizobiales order, yet is still situated within the Alphaproteobacteria class. Addition of this outgroup to phylogenies had a neutral effect on the position of ingroup taxa. The outgroup for the nodA analysis was Azorhizobium sp. SD02 (GenBank accession AJ300262). There was no outgroup for glnII because there is too little homology between the glutamine synthase II gene of rhizobia and other taxa that could act as an appropriate outgroup.
Model selection
Phylogenetic analysis of aligned DNA sequences requires an assumed model of DNA evolution. The simplest model is Jukes Cantor (JC) which assumes that DNA sequences have equal base frequencies and equal mutation rates Jukes69. This model is unlikely to be accurate with real data. The most complex model of DNA evolution is the general time reversible (GTR+I+) model which allows base frequencies, substitution rates, proportion of invariant sites (I), and the gamma distribution shape parameter () to vary independently. Although it may seem wise to choose the most complex model, over parametrisation can lead to incorrect conclusions Posada03. Hence different models of evolution were tested to determine the most appropriate to use for each particular alignment.
For Maximum Likelihood (ML) DNA trees, model parameters were selected with Modeltest 3. 7 Posada98. This computer program tests nucleotide alignments against 56 different models of DNA evolution using ML. The resultant negative log likelihood (--lnL) scores and associated parameters were subjected to a hierarchical likelihood ratio test (hLRT), the Akaike Information Criterion test (AIC), and the Bayesian Information Criterion test (BIC) to determine which model best fitted the sequence data. In cases where the hLRT, AIC, and BIC disagreed in model selection, the AIC choice was used, as this can ``simultaneously compare multiple nested or nonnested models, assess model selection uncertainty, and allow for the estimation of phylogenies and model parameters using all available models'' Posada04.
Bayesian analyses were conducted with MrBayes 3. 11 Huelsenbeck01,Ronquist03. For Bayesian analyses, exact specification of model parameters is not possible. The equivalent of the GTR+I+ model was used with a `flat' Dirichlet distribution (all distribution parameters set to 1), as the prior probability for both the stationary state frequencies and the substitution rates.
For protein coding genes, DNA sequences were translated to protein sequences using GeneDoc. Protein sequence model parameters were selected among 72 possibilities with ProtTest 1. 2. 6 Abascal05,Guindon03,Drummond01. The models were tested using hLRTs, AIC, and BIC on the `slow' method, optimising branch lengths, model, and topology. Protein sequences analysed using Bayesian methods, used the amino acid model prior derived from the ProtTest analysis.
Tree construction
Preliminary analyses were performed with ClustalX and PAUP* Neighbour Joining (NJ) methods, with 1000 bootstrap replicates. The HKY85 model of DNA evolution was used HKY85. NJ trees were used for a quick evaluation of tree shape and taxa position, but were not used for phylogenetic inference, as the model is too simple to reflect true evolutionary processes accurately. The commands used for running analyses are presented in Appendix s-PAUP-block.
Maximum Likelihood, although computationally demanding, was the preferred method of analysis because the assumptions of this model are more rigorous than NJ Baxevanis04. Model parameters (base frequencies, proportion of invariable sites, gamma distribution shape parameter, and substitution rate matrix) derived from Modeltest were specified in PAUP* 4. 0b10 Swofford02 to build phylograms using Tree-Bisection-Reconnection (TBR) heuristics. Ten replicates were run for reproducibility and to examine the tree-island profile.
To complement the ML analyses, sequence data were also analysed with Bayesian inference methods. These have the advantage of providing a quantitative measure of clade support through posterior probabilities. Bayesian MCMCMC analyses were performed with two independent runs, each with four chains, for 10 million generations. This was sufficient such that measures of convergence (average standard deviation of split frequencies, potential scale reduction factors) were at acceptable levels. The first 25 of the 10000 trees generated were discarded to eliminate data not in the stationary phase, and a 50 majority rule consensus tree built with the remainder.
Protein Bayesian analyses were run for 2 million generations, until convergence diagnostics were acceptable. The first 25 of the 20000 trees generated were discarded, and a 50 majority rule consensus tree built with the remainder.
ML protein trees were generated internally by the ProtTest program. All trees were viewed in Treeview 1. 6. 6 Page96, saved as enhanced metafiles (. emf), and edited as vector objects in CorelDraw.
Biolog phenotypic profiles
The Biolog GN2 microplate system uses carbon source substrate metabolism to create a `metabolic fingerprint' of bacterial strains. The system consists of a 96 well plate, with 95 different carbon sources and a blank well (Table Biolog-grid). A standardised suspension of bacteria is added to each well and the plate incubated. If bacteria are able to metabolise the substrate, a redox dye (tetrazolium violet) is reduced, changing from clear to purple.
In general the manufacturer's instructions were followed. Preliminary testing revealed, however, that rhizobia grew very poorly on the recommended `Bug' agar, thus R2A medium was substituted.
Fresh bacterial cultures were subcultured twice, grown on R2A agar plates for 24 hours, then suspended in sterile `gelling' inoculation fluid (0. 40 NaCl, 0. 03 Pluronic F-68, 0. 02 Gellan gum) to a concentration of 522 transmittance. 150 L of this suspension was added to each of the 96 wells in a GN2 microplate and incubated at 30 . Microplates were analysed at 4, 24, 36, and 48 hours on a Biolog microstation plate reader using MicroLog3 software.
Analysis
The results from Biolog plate reads were available as raw absorbance values for each well, or expressed qualitatively as positive, negative and partial reactions (Fig. p-Biolog). The software decided the state of a well by comparison to the control well (A1). These values were visually verified. The results for each strain at each time interval were exported from the software (MicroLog3 4. 01C) as a CSV file, and converted to numerical values (1: positive, 0: negative). Partial wells were scored as another state (0. 5) in some analyses and as positives (1) in other analyses to assess the difference on the dendrogram.
figure [tbp] [width=10cm]Biolog [Biolog software screen]Biolog software screen after the 96 well plate has been scanned. The strain shown is Mesorhizobium sp. ICMP 14319. The grid in the upper right section indicates positive (dark circles), partial (half circles), and negative (empty circles) reactions. The + and = symbols refer to identification and are not relevant. p-Biolog figure
The data matrices were analysed with GenStat 6th edition (VSNI) and MrBayes 3. 11. In GenStat, a hierarchical clustering dendrogram was generated with simple matching and average linkage. In MrBayes, Bayesian inference was used under the standard datatype, analyses were run for 10 million generations such that measures of convergence were at excellent levels.
Fatty acid methyl ester (FAME) profiles
A total of 50 strains, comprising 25 types, and 25 New Zealand isolates, were selected and sent to Central Science Laboratories in York, England for FAME analysis with the following method.
Strains were grown on TY-M medium (Table t-medium) for 48 hours at 28 . Bacterial suspensions were saponified in NaOH/methanol (45g NaOH, 150 mL methanol, 150 mL mQ water), then methylated at 80 in hydrochloric acid (6 M) and methanol. The organic phase was extracted in hexane and methyl tert-butyl ether, and analysed by gas chromatography.
The data were analysed with unweighted pair match grouping (UPGMA) and expressed in Euclidean distances. Results received included a hardcopy of the dendrogram and raw GC reads.
Host-range experiments
Legume seed
Seed from the legumes used in this study (Table t-legume-seed) was sourced from field collections and commercial suppliers.
Sophora seed was collected from the Canterbury region and purchased from Proseed (Amberley, NZ). Clianthus, Acacia, and Carmichaelia stevensonii seed was also purchased from Proseed. Carmichaelia australis seed was collected from Mt Albert and Bethells beach in AucklandMany of the seed pods found at Bethells were empty of seed, but contained a native weevil Peristoreus sudus. The adults of this species feed on pollen, and the larvae on seeds [Identified by Stephen][]Thorpe-pc. . Ulex and Cytisus seed was collected from farmland in Rotorua. Trifolium seeds were purchased from Newton Seed Produce (Auckland, NZ). Lotus, Cicer, and Astragalus were purchased from Kings seeds (Katikati, NZ). Phaseolus and Pisum were purchased from Carnival seeds (Auckland). Styphnolobium was purchased from New Zealand Tree Seeds (Rangiora, NZ). Montigena seeds were not available for study.
table Legume plants used in this study tabularllP3. 1 Full name & Authority & 1cSeed mass (mg) Sophora microphylla & Aiton (1789) & 70. 9 Sophora tetraptera & J. S. Mill. (1780) & 59. 1 Carmichaelia australis & R. Br. (1825) & 7. 0 Carmichaelia stevensonii & (Cheeseman) Heenan (1998) & 2. 7 Clianthus puniceus & (G. Don) Sol. ex Lindl. & 9. 6 Montigena novae-zelandiae & (Hook. f. ) Heenan (1998) & NA^d Cytisus scoparius & (Linnaeus) Link (1822) & 9. 1 Ulex europaeus & (Linnaeus) Link & 6. 2 Acacia longifolia & (Andrews) Willd. (1806) & 14. 4 Trifolium repens & Linnaeus & 0. 6 Phaseolus vulgaris & Linnaeus & 276. 6 Pisum sativum & Linnaeus & 261. 4 Astragalus membranaceus & Schischkin (1933) & 3. 5 Lotus tetragonolobus & Linnaeus & 37. 9 Cicer arietinum & Linnaeus & 485. 1 Glycine max & (Linnaeus) Merr. & 215. 5 Styphnolobium japonicum & (Linnaeus) Schott & 136. 7 tabular
t-legume-seed Sections are respectively: native New Zealand legumes, exotic weed legumes, Rhizobium leguminosarum hosts, Mesorhizobium spp. hosts. Authority from ILDIS. Average of 100. No seed was available for this species. table
Seeds from the different species used in this experiment required slightly different protocols for germination. For most species, the seed was pierced with the tip of a sterile scalpel, such that the point just penetrated the testa. For the harder seeds of Acacia, Sophora, Lotus, and Styphnolobium a chip was made in the testa away from the embryonic axis. Seeds were soaked overnight in sterile RO water to leach possible inhibitory compounds.
Seeds were surface sterilised in a 5 solution of commercial bleach and Tween 80 detergent (10 LL) for 10--30 min, then transferred to water agar plates (7. 5 g agar per litre of mQ HO, autoclaved) and germinated at room temperature.
Growth of seedlings
Trial legumes were grown in clear PET (Polyethylene terephthalate) plastic screw top jars (Fig. p-jar). Most seeds were planted in 400mL jars, however 750mL jars were used for the larger Phaseolus and Pisum seedlings. PET jars do not tolerate autoclaving, therefore they were sterilised by submersing in 5 bleach solution for one hour, then drying in an oven at 45 .
The jars were filled with 150 mL of dry, twice-autoclaved, fine grade vermiculite (300 mL for the large jars), and the inoculant added (see Section s-inoculant).
Seedlings were selected for evenness of growth, and a single seedling planted in the vermiculite of each jar. All operations (vermiculite filling, inoculation, planting) were carried out aseptically in a laminar flow cabinet, and seedings were handled with flame-sterilised forceps.
figure [h] [width=5cm]jar [Plant inoculation jar]400 mL PET plastic jar used for inoculation and growth of legumes in this study. This jar contains a single Sophora microphylla seedling, inoculated with strain ICMP 5943. p-jar figure
Inoculation of seedlings
Bacteria for inoculation was grown as a lawn on R2A agar plates. After 24--48 hours growth, cells were resuspended in decanted Jensen's nitrogen-free plant growth medium (CaHPO 1. 0g, KHPO. 3HO 0. 262g, MgSO. 7HO, 0. 2g, NaCl 0. 2g, FeCl 0. 1g, per litre. Always used at half strength) to a concentration of 110 cells per millilitre (an absorbance of 0. 167 at 600 nm in a 1 cm cuvette, previously calibrated by plate counts). A 5mL aliquot of this suspension was diluted 10-fold in 0. 5 Jensen's medium, and the resultant 50mL added evenly over the surface of the vermiculite, giving a final number of rhizobia (cfu) per jar of approximately 510, or 3. 310 per mL of vermiculite. The inoculant volume was doubled for the 750 mL jars. In uninoculated control jars, an equivalent amount of sterile 0. 5 Jensen's medium was added in place of the inoculant. Jars were covered with perforated plastic wrap and placed in the greenhouse.
Greenhouse facilities
During the course of these experiments, different trials were held in different locations due to the research facility relocating. Fully operational greenhouse units were not available at the new location forcing improvisation by keeping some plants indoors on a window ledge. Edge effects were mitigated by pseudo-random rearrangement of jars during watering.
Initial experiments were done in a South-facing greenhouse, at the Mt. Albert Research Centre in Auckland, New Zealand [53'28. 93''S, 43' 36. 4'' E]. The temperature was maintained between 18 and 25 . Sodium vapour lamps were on a timer to provide for a 12 hour light--dark cycle.
In April 2004, Landcare Research relocated 11 kilometres east to the suburb of Tamaki [ 53' 6. 71'' S, 50' 54. 7'' E]. No temperature- or lighting-controlled greenhouse units were available. Several experiments were performed here, before extremes of temperature necessitated a move to the windowsill of a temperature-controlled (25 ) PC2 containment laboratory on site.
In all facilities a data logger recorded temperature and humidity every hour.
Assessment of nitrogen fixation
The ability of a plant to fix N was indirectly assessed by acetylene reduction to determine the presence of an effective nitrogenase enzyme. Although nitrogen fixation can be assessed by the colour of leaves or cotyledons Mears59, with large-seeded species such assessment is difficult, as the embryo has large reserves of nitrogen in the endosperm (see seed mass, Table t-legume-seed). Acetylene reduction is a much more accurate method.
Acetylene reduction was performed by a protocol modified from Silvester83. Acetylene made directly from calcium carbide chips immersed in water was injected into each jar to give a final concentration of 10 v/v. Jars were incubated on the lab bench and analysed for ethylene (CH) after approximately one hour. Ethylene was analysed by standard flame ionisation gas chromatography (Shimadzu GC8A) standardised with pure ethylene and results expressed as pmol of CH produced per jar per minute.
Positive results were hundreds to thousands of pmoljarmin. Since these experiments were not designed to be quantitative, the values were converted to qualitative values (Fix for ethylene production well above background, and Fix for background levels).
Assessment of nodulation
Nodulation was assessed by uprooting the plant, washing away adhering vermiculite, and counting the number of nodules present. The presence of nodules was scored as Nod and absence of nodules as Nod. Nodule characteristics such as colour and shape were also recorded. Some nodules were also cut open to check for the presence of red pigment (leghaemoglobin) (see Fig. p-Nodules), and were crushed, the exudate spread on a slide, heat fixed (or wet mounted), and observed for bacteroids.
figure [width=12cm]Nodules [Cross section of effective and ineffective nodules] Cross section of ineffective (A), and effective (B) nodules from C. puniceus showing absence and presence respectively of leghaemoglobin (red pigment). Bar indicates 1 mm length. p-Nodules figure
Verification of isolate identification UARR
To confirm the identity of the nodulating strain, bacteria were isolated from a single nodule from each replicate for each experiment and the DNA extracted as previously described (Section DNAex). A 400-bp segment of the 16S rRNA gene was PCR amplified with UARR universal primers Rivas04b, U1F: CTYAAAKRAATTGRCGGRRRSSC and U1R: CGGGCGGTGTGTRCAARRSSC, using the recommended PCR conditions (Table PCR-UARRcycle). PCR products were sequenced and compared to the inoculum strain.
table [h] UARR PCR cycle tabularccc Temperature & Time & 95 & 5 mins & 1 Hold 95 & 60 s &
55 & 120 s & 35 Cycles
72 & 60 s & 72 & 7 mins & 1 Hold 10 & & 1 Hold tabular PCR-UARRcycle table
Pristine soil experiments
These experiments used soil from pristine areas in the National Parks of New Zealand. Legume species were grown in these soils as `trap hosts' to elicit nodulation. As a positive control, an appropriate inoculant for native legumes (ICMP 15054) or exotic legumes (ICMP 14291) was added to some jars from each soil sample to ensure nodulation was possible under these conditions.
Soil was collected by workers from the Department of Conservation (DoC) or local scientists who were instructed to collect in pristine areas away from farmland, and walking tracks, and to collect in places where legumes are absent. Spades were cleaned with bleach before digging from the top 30 cm of soil.
The soil samples were posted to Auckland at ambient temperature and kept refrigerated at 4 for a maximum of one month before use. 150 cm of soil was aseptically placed into bleach-sterilised 400 mL PET plastic jars as previously described. Four germinated seedlings from surface-sterilised seed were planted per jar. Jars were initially watered with Jensen's nitrogen free medium then covered with perforated plastic wrap and grown in the greenhouse for 10 weeks, watering as necessary with sterile RO water.
After 10 weeks the plants were harvested by flushing with water to remove the soil. Nodules were counted and removed for isolation of bacteria as previously described.