\chapter{Introduction} \section{Rhizobia} \label{s-rhizobia-intro} In this thesis `rhizobia' are defined as bacteria capable of forming root nodules on legumes, mediated by \emph{nod} genes. This term describes the phenotype (causing root nodules), but has no taxonomic relevance and should not be confused with the genus name \emph{Rhizobium} (although `rhizobia' has been used by others for the plural form of \emph{Rhizobium}). An equivalent term used by other researchers is `root nodule bacteria' (RNB) \citep{Zakhia04,Howieson05}. Rhizobia are soil-inhabiting bacteria with the potential for forming specific root structures called nodules. In effective nodules the bacteria fix nitrogen gas (N$_2$) from the atmosphere into ammonia \citep{OGara76}, which is assimilated by the plant and supports growth---particularly in nutrient deficient soils. In return the rhizobia are supplied with nutrients (predominantly dicarboxylic acids \citep{Lodwig03}), and are protected inside the nodule structure \citep{VanRhijn95}. In ineffective nodules no nitrogen is fixed, yet rhizobia are still supplied with nutrients, and in this situation the rhizobia could be considered parasitic \citep{Denison04b}. The nitrogen-fixing symbiotic relationship has been exploited in agriculture to enhance crop and pasture growth without the addition of nitrogen fertilisers. For this reason, the majority of research in this field has focused on herbaceous crop and forage legumes of agricultural significance. In contrast, few studies have been made of rhizobial associations among non-crop legumes, despite the fact that they may be ecologically important in the natural landscape \citep{Boring88}. Worldwide, there are an estimated 17\,000--19\,000 legume species \citep{Martinez96}, although nodulating bacterial species have only been identified for a small proportion of these. To date (September 2006), 55 rhizobial species have been identified in twelve genera (Table \ref{t-Rhizobia-species}). Most of the species are in the genera \emph{Rhizobium} (from the Latin `root living'), \emph{Bradyrhizobium}, \emph{Mesorhizobium} and \emph{Ensifer} (\emph{Sinorhizobium}). A detailed discussion of rhizobial systematics is presented in Section \ref{s-systematics}, but important taxonomic distinctions are noted below. All currently known rhizobia are in the phylum Proteobacteria, most in the class Alphaproteobacteria, which contains six rhizobial families in a single order---Rhizobiales, as listed in the hierarchy below \citep{Bergeys-outline}. \singlespacing \begin{tabbing} 0000 \= 0000 \= 0000 \= 0000 \kill % \> for next tab, \\ for new line... Rhizobiales \\ \> Rhizobiaceae \\ \> \> \emph{Rhizobium} \\ \> \> \emph{Ensifer} (\emph{Sinorhizobium}) \\ \> Brucellaceae \\ \> \> \emph{Ochrobactrum} \\ \> Phyllobacteriaceae \\ \> \> \emph{Phyllobacterium} \\ \> \> \emph{Mesorhizobium} \\ \> Bradyrhizobiaceae \\ \> \> \emph{Bradyrhizobium} \\ \> Hyphomicrobiaceae \\ \> \> \emph{Azorhizobium} \\ \> \> \emph{Devosia} \\ \> Methylobacteriaceae \\ \> \> \emph{Methylobacterium} \\ \end{tabbing} \onehalfspacing %\doublespacing There are also three rhizobial species in two families in the Betaproteobacteria, all of which are in the Burkholderiales order, as listed below \citep{Bergeys-outline}. \singlespacing \begin{tabbing} 0000 \= 0000 \= 0000 \= 0000 \kill % \> for next tab, \\ for new line... Burkholderiales \\ \> Burkholderiaceae \\ \> \> \emph{Burkholderia} \\ \> \> \emph{Cupriavidus} \\ \> Oxalobacteraceae \\ \> \> \emph{Herbaspirillum} \\ \end{tabbing} \onehalfspacing %\doublespacing Although it remains to be confirmed, it is possible that some Gammaproteobacteria also nodulate legumes \citep{Benhizia04}. %\vspace{2cm} \singlespacing % \afterpage{\clearpage \input{rhizobia-sp.tex}} {\small \begin{longtable}{ll} \caption{List of rhizobial species} \\ \toprule \textbf{Binomial Name} & \textbf{Authority}$^a$ \\ \midrule \emph{Rhizobium daejeonense} & \citealp{Quan05} \\ \emph{Rhizobium etli} & \citealp{Segovia93} \\ \emph{Rhizobium galegae} & \citealp{Lindstrom89} \\ \emph{Rhizobium gallicum} & \citealp{Amarger97} \\ \emph{Rhizobium giardinii} & \citealp{Amarger97} \\ \emph{Rhizobium hainanense} & \citealp{Chen97} \\ \emph{Rhizobium huautlense} & \citealp{Wang98} \\ \emph{Rhizobium indigoferae} & \citealp{Wei02} \\ \emph{Rhizobium leguminosarum}$^{\mathrm{T}}$ & \citep{Frank79} \citealp{Frank89} \\ \emph{Rhizobium loessense} & \citealp{Wei03} \\ \emph{Rhizobium mongolense} & \citealp{vanBerkum98} \\ \emph{Rhizobium sullae} & \citealp{Squartini02} \\ \emph{Rhizobium tropici} & \citealp{Martinez91} \\ \emph{Rhizobium undicola} & \citep{deLajudie98a} \citealp{Young01a} \\ \emph{Rhizobium yanglingense} & \citealp{Tan01a} \\ \cmidrule{1-2} \emph{Ensifer} (\emph{Sinorhizobium}) \emph{abri} & \citealp{Ogasawara03} \\ \emph{Ensifer adhaerens} & \citep{Wang02a} \citealp{Young03b}\\ \emph{Ensifer} (\emph{Sinorhizobium}) \emph{americanum} & \citealp{Toledo03} \\ \emph{Ensifer arboris} & \citep{Nick99} \citealp{Young03b} \\ \emph{Ensifer fredii}$^{\mathrm{T}}$ & \citep{Scholla84} \citealp{Young03b} \\ \emph{Ensifer} (\emph{Sinorhizobium}) \emph{indiaense} & \citealp{Ogasawara03} \\ \emph{Ensifer kostiensis} & \citep{Nick99} \citealp{Young03b} \\ \emph{Ensifer kummerowiae} & \citep{Wei02} \citealp{Young03b} \\ \emph{Ensifer medicae} & \citep{Rome96} \citealp{Young03b} \\ \emph{Ensifer meliloti} & \citep{Dangeard26} \citealp{Young03b} \\ \emph{Ensifer saheli} & \citep{deLajudie94} \citealp{Young03b} \\ \emph{Ensifer terangae} & \citep{deLajudie94} \citealp{Young03b} \\ \emph{Ensifer xinjiangense} & \citep{Chen88} \citealp{Young03b} \\ \cmidrule{1-2} \emph{Mesorhizobium amorphae} & \citealp{Wang99a} \\ \emph{Mesorhizobium chacoense} & \citealp{Velazquez01a} \\ \emph{Mesorhizobium ciceri} & \citep{Nour94a} \citealp{Jarvis97} \\ \emph{Mesorhizobium huakuii} & \citep{Chen91} \citealp{Jarvis97} \\ \emph{Mesorhizobium loti}$^{\mathrm{T}}$ & \citep{Jarvis82} \citealp{Jarvis97} \\ \emph{Mesorhizobium mediterraneum} & \citep{Nour95} \citealp{Jarvis97} \\ \emph{Mesorhizobium plurifarium} & \citealp{deLajudie98b} \\ \emph{Mesorhizobium septentrionale} & \citealp{Gao04b} \\ \emph{Mesorhizobium temperatum} & \citealp{Gao04b} \\ \emph{Mesorhizobium tianshanense} & \citep{Chen95} \citealp{Jarvis97} \\ \cmidrule{1-2} \emph{Bradyrhizobium canariense} & \citealp{Vinuesa05b} \\ \emph{Bradyrhizobium elkanii} & \citealp{Kuykendall93} \\ \emph{Bradyrhizobium japonicum}$^{\mathrm{T}}$ & \citep{Kirchner96} \citealp{Jordan82} \\ \emph{Bradyrhizobium liaoningense} & \citealp{Xu95} \\ \emph{Bradyrhizobium yuanmingense} & \citealp{Yao02} \\ \cmidrule{1-2} \emph{Burkholderia caribensis} & \citealp{Vandamme02} \\ \emph{Burkholderia cepacia} & \citealp{Vandamme02} \\ \emph{Burkholderia phymatum} & \citealp{Vandamme02} \\ \emph{Burkholderia tuberum} & \citealp{Vandamme02} \\ \cmidrule{1-2} \emph{Azorhizobium caulinodans}$^{\mathrm{T}}$ & \citealp{Dreyfus88} \\ \emph{Azorhizobium doebereinerae} & \citealp{Souza06} \\ \cmidrule{1-2} \emph{Cupriavidus taiwanensis} & {\footnotesize \citep{Chen01x} \citealp{Vandamme04}} \\ \cmidrule{1-2} \emph{Devosia neptuniae} & \citealp{Rivas03} \\ \cmidrule{1-2} \emph{Herbaspirillum lusitanum} & \citealp{Valverde03} \\ \cmidrule{1-2} \emph{Phyllobacterium trifolii} & \citealp{Valverde03} \\ \cmidrule{1-2} \emph{Methylobacterium nodulans} & \citealp{Jourand04} \\ \cmidrule{1-2} \emph{Ochrobactrum lupini} & \citealp{Trujillo05} \\ \bottomrule \label{t-Rhizobia-species} \end{longtable} } \vspace{-0.9cm} {\footnotesize \noindent $^{\mathrm{T}}$ Type species \\ $^a$ Parenthesis indicate original publication, following reference is subsequent reclassification \\ } \medskip \onehalfspacing There are a number of species present in these rhizobial genera that have not been observed to form nodules, and therefore do not fit the functional definition of rhizobia. These include many of the species that were formerly known as \emph{Agrobacterium} (e.g. \emph{R.~larrymoorei}, \emph{R.~rubi}, and \emph{R.~vitis}; \citep{Young01a,Young04a}). However, there is recent evidence that other species formerly classified as \emph{Agrobacterium} are capable of nodulation. For example \emph{R.~radiobacter}\footnote{As \emph{Agrobacterium tumefaciens} in the publication.} nodulates \emph{Phaseolus vulgaris}, \emph{Campylotropis} spp., \emph{Cassia} spp.~\citep{Han05}, and \emph{Wisteria sinensis} \citep{Lui05}. Both nodules and tumours were formed on \emph{Phaseolus vulgaris} by \emph{R.~rhizogenes} strains containing a Sym plasmid \citep{Velazquez05}. There are also other species, although classified within genera commonly considered to be represented entirely by nodulating strains, in fact include strains apparently devoid of nodulation ability. For example \emph{Bradyrhizobium betae} forms tumours on \emph{Beta vulgaris} (Beetroot) but is not known to fix N$_2$ \citep{Rivas04a}. \emph{Mesorhizobium thiogangeticum} is a sulfur-oxidising bacterium, and does not nodulate the tested legumes of \emph{Clitoria ternatea}, \emph{Pisum sativum}, and \emph{Cicer arietinum} \citep{Gosh06}. There are also non-symbiotic strains of \emph{Mesorhizobium} (and other genera) that can become nodulating species by acquiring symbiosis genes \citep{Sullivan95}. The genus \emph{Sinorhizobium} was recently reclassified to \emph{Ensifer} on the basis of similarity of DNA sequences and priority of publication \citep{Willems03,Young03b}. \emph{Ensifer adhaerens} is a soil bacterium that attaches to other bacteria and may cause cell lysis \citep{Casida82}. Although wild type \emph{E.~adhaerens} did not nodulate \emph{Phaseolus vulgaris} nor \emph{Leucaena leucocephala}, it did so when transformed with a symbiotic plasmid from \emph{Rhizobium tropici} \citep{Rogel01}, demonstrating its capacity to become a rhizobial species. Other \emph{E.~adhaerens} strains were subsequently isolated that nodulated legumes naturally. These form a single clade with \emph{Sinorhizobium} in 16S rRNA and \emph{recA} phylogenies leading \citet{Willems03} to suggest that these strains be reclassified as \emph{Sinorhizobium adhaerens}. However, \emph{Ensifer} \citep{Casida82} is the senior heterotypic synonym and thus takes priority \citep{Young03b}. This means that all \emph{Sinorhizobium} spp.~must be renamed as \emph{Ensifer} spp.~according to the Bacteriological code \citep{Lapage90}. In this thesis \emph{Ensifer} is used exclusively. \emph{Cupriavidus} species have recently undergone several taxonomic revisions, being formerly known as both \emph{Wautersia} and \emph{Ralstonia}. This genus currently contains a single rhizobial species, and ten other non-symbiotic species \citep{Euzeby97}. Rhizobial systematics is rapidly changing, and recently many new species have been recognised. Novel species may also be associated with the native legumes of New Zealand. %-------------------------------------------------------------------------------- \section{New Zealand native legumes} \label{s-NZ-natives} \subsection{Introduction} New Zealand has 33 species of legumes that are native. These are comprised of four genera: \emph{Sophora}, \emph{Carmichaelia}, \emph{Clianthus}, and \emph{Montigena}. %------------------------------------------------------------------------------------ \subsection{\emph{Sophora}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Sophora} \caption[\emph{Sophora chathamica}]{\emph{Sophora chathamica} showing yellow bell-shaped flowers and mature seed pod (right of centre).} \label{p-Sophora} \end{figure} \emph{Sophora} L.~(1753) was named after \emph{sufayra}, the arabic name for the tree. The M\=aori name (and the vernacular) for the endemic \emph{Sophora} is `k\=owhai', from the word for yellow---which describes the colour of the flowers (Fig.~\ref{p-Sophora}). There are eight species native to New Zealand: \emph{S.~chathamica}, \emph{S.~fulvida}, \emph{S.~godley}, \emph{S.~longicarinata}, \emph{S.~microphylla}, \emph{S.~molloyi}, \emph{S.~prostrata} and \emph{S.~tetraptera} \citep{Heenan01c}. There are another 49 species in the genus \emph{Sophora} that are not native to New Zealand. Species endemic to the Southern Hemisphere are in the \emph{Edwardsia} sector of \emph{Sophora}. \emph{Edwardsia} members other than the New Zealand natives are from South America (\emph{S.~macrocarpa}), Lord Howe Island (\emph{S.~howinsula}), Hawaii (\emph{S.~chrysophylla}), La R\'eunion (\emph{S.~denudata}), Easter Island (\emph{S.~toromiro}), and Raivavae Island (\emph{S.~raivavae}). \emph{S.~microphylla} was considered to occur in Chile and on Gough Island in the south Atlantic \citep{Markham72}, however \citet{Heenan01a} considers these species to be \emph{Sophora cassioides}, distinct from the New Zealand species. The type species of the genus is \emph{S.~tomentosa}$^{\mathrm{T}}$ which is closely related to sect.~\emph{Edwardsia} \citep{Heenan04,ILDIS}. %------------------------------------------------------------------------------------ \subsection{\emph{Carmichaelia}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Carmichaelia} \caption[\emph{Carmichaelia australis}]{\emph{Carmichaelia australis} showing the cladodes and mature seed pods. Insert: 2$\times$ magnification of flowers from the same plant.} \label{p-Carmichaelia} \end{figure} \emph{Carmichaelia} R.Br.~(1825) was named after Captain Dugald Carmichael, a Scottish army officer and botanist who collected plants in New Zealand \citep{Allen81}. The English vernacular name is `New Zealand broom', and in M\=aori is variably known as tawao, m\=akaka, maukoro, and tainoka \citep{NZPlant} (illustrated in Fig.~\ref{p-Carmichaelia}). The taxonomic history of this genus is complex, and has been confused by inadequate collections and intraspecific variation \citep{Heenan95b}. The formerly recognised genera of \emph{Chordospartium}, \emph{Corallospartium}, \emph{Notospartium}, and \emph{Huttonella} are now included in \emph{Carmichaelia} \citep{Heenan95a,Heenan98c,Heenan98a}. In the most recent treatment \citep{Heenan95b,Heenan96b}, there are 22 species of \emph{Carmichaelia} native to New Zealand (\emph{C.~appressa}, \emph{C.~arborea}, \emph{C.~astonii}, \emph{C.~australis}$^{\mathrm{T}}$, \emph{C.~carmichaeliae}, \emph{C.~compacta}, \emph{C.~corrugata}, \emph{C.~crassicaule}, \emph{C.~curta}, \emph{C.~glabrescens}, \emph{C.~hollowayi}, \emph{C.~juncea}, \emph{C.~kirkii}, \emph{C.~monroi}, \emph{C.~muritai}, \emph{C.~nana}, \emph{C.~odorata}, \emph{C.~petriei}, \emph{C.~stevensonii}, \emph{C.~torulosa}, \emph{C.~uniflora}, \emph{C.~vexillata}, and \emph{C.~williamsii}). An additional species, \emph{C.~exsul}, is found on Lord Howe Island in the Tasman Sea, 600 km east from Australia. The species exhibit remarkable diversity, from trees to prostrate forms a few centimetres high. \emph{Carmichaelia} is distributed throughout New Zealand, although most species are restricted to certain localities. Most of the diversity (15 species) is in the eastern South Island. They typically invade disturbed habitats on shallow poor soils, drought and frost prone areas, and alluvial soils \citep{Wagstaff99}. %------------------------------------------------------------------------------------ \subsection{\emph{Clianthus}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Clianthus} \caption[\emph{Clianthus puniceus}]{\emph{Clianthus puniceus} showing distinctive beak-shaped flowers.} \label{p-Clianthus} \end{figure} \emph{Clianthus} Soland. ex Lindl. was named from the Greek \emph{kleos} `glory' and \emph{anthos} `flower' \citep{Allen81}. The English vernacular name is `kakabeak' after its distinctive flowers shaped like a native parrot's (k\=ak\=a) beak (Fig.~\ref{p-Clianthus}), it is known in M\=aori as `k\=owhai ngutuk\=ak\=a' \citep{Shaw97}. Once considered monotypic, in the most recent treatment \citep{Heenan95c,Heenan00}, there are now two species (\emph{C.~maximus} and \emph{C.~puniceus}$^{\mathrm{T}}$) native to New Zealand. It is found naturally only in isolated refuges in the eastern North Island. Formerly some Australian and Asian legumes were classified as \emph{Clianthus}, these are now known as \emph{Swainsona} and \emph{Sarcodum} \citep{ILDIS}. %---------------------------------------------------------------------------- \subsection{\emph{Montigena}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Montigena} \caption[\emph{Montigena novae-zelandiae}]{\emph{Montigena novae-zelandiae}, growing on a scree slope, with mature seed pods. Photo \copyright\ Peter Heenan.} \label{p-Montigena} \end{figure} \emph{Montigena} (Hook.f.) Heenan, is named from `mountain-born' referring to its habitat. \citep{Heenan98d}. The English vernacular name is `scree pea' (Fig.~\ref{p-Montigena}). \emph{Montigena novae-zelandiae}$^{\mathrm{T}}$ is the only species in the \emph{Montigena} genus. It was known as \emph{Swainsona novae-zelandiae} until \citet{Heenan98d} reclassified it based on morphological features. There are currently 55 \emph{Swainsona} species, mostly in Australia \citep{ILDIS}. \emph{Montigena} has a distinctly different ITS sequence from other New Zealand legumes, but forms a clade with the Australian \emph{Swainsona} \citep{Wagstaff99} (See Fig.~\ref{p-legume-tree}). \emph{Montigena} is endemic to the dry eastern mountains of the South Island of New Zealand, where it grows on partially stable scree slopes. %----------------History--------------------------------------- \section{Evolution and history of New Zealand native legumes} \label{legume-history} \subsection{Geology and palaeobotany} The archipelago of New Zealand began to split away from the larger landmass of Gondwana about 80 million years ago (mya) due to continental drift, although was still relatively close for another 10 to 20 million years \citep{Cooper93,Stevens95}. The start of this separation coincides approximately with the date of the evolution of legumes \citep{Sprent94}, although legumes were not abundant until 35--54 million years ago \citep{Doyle03}. Hence all legumes must have arrived in New Zealand after its separation from Gondwana. The historical presence of legumes in New Zealand is largely inferred from fossil pollen records. Fossils of \emph{Carmichaelia} were detected in the ``late Pliocene Waipaoa series''of soils dating from less than 5 mya \citep{Oliver28}, but \emph{Sophora} is not common in fossil pollen records until the Pleistocene (1.81 mya) \citep{Hurr99}. Fossil pollen records also show that before the last Ice Age ended, 10\,000 years ago, New Zealand had an indigenous population of \emph{Acacia} spp.~\citep{Mildenhall72,Lee01}. \begin{table} [tbp] \caption{Native legume taxonomic hierarchy} \centering \begin{tabular}{|c|c|c|c|c|} \hline \textbf{Kingdom} & \multicolumn{4}{c|}{Plantae} \\ \hline \textbf{Division} & \multicolumn{4}{c|}{Magnoliophyta} \\ \hline \textbf{Class} & \multicolumn{4}{c|}{Magnoliopsida} \\ \hline \textbf{Order} & \multicolumn{4}{c|}{Fabales} \\ \hline \textbf{Family} & \multicolumn{4}{c|}{Fabaceae} \\ \hline \textbf{Subfamily} & \multicolumn{4}{c|}{Faboideae} \\ \hline \textbf{Tribe} & Sophoreae & \multicolumn{2}{c|}{Carmichaeliaeae$^a$} & Galegeae$^a$ \\ \hline \textbf{SubTribe} & & \multicolumn{3}{c|}{Carmichaelinae$^b$} \\ \hline \textbf{Genus} & \emph{Sophora} & \emph{Carmichaelia} & \emph{Montigena} & \emph{Clianthus} \\ \hline \end{tabular} \label{taxbox-Native} \medskip \raggedright {\footnotesize \hskip 1cm $^a$ After \citet{Pohill81-Carm}. \\ \hskip 1cm $^b$ After \citet{Wagstaff99}. \\} \end{table} \subsection{The Carmichaelinae} The original classification of native legumes placed \emph{Carmichaelia} and \emph{Montigena} in the tribe Carmichaelieae, and \emph{Clianthus} in the diverse tribe Galegeae \citep{Pohill81-Carm}; however this classification is polyphyletic \citep{Wagstaff99}, and recent evidence has suggested that \emph{Carmichaelia}, \emph{Clianthus}, \emph{Montigena}, and the Australian genus \emph{Swainsona}, form a single clade called Carmichaelinae at the subtribe rank \citep{Wagstaff99} (Table~\ref{taxbox-Native}). \citet{Wagstaff99} used ITS sequences of 39 species of \emph{Carmichaelia}, \emph{Clianthus}, \emph{Montigena}, \emph{Swainsona} and related legumes, to determine the classification and origins of New Zealand legumes. Most species of \emph{Carmichaelia} had nearly identical ITS sequences, indicating recent radiation. The results suggested that Carmichaelinae were derived from the Northern Hemisphere Astragalinae, and confirmed an earlier study of \citet{Heenan98a} using 47 phenotypic characters. \begin{figure} [tbp] \centering \includegraphics[width=12cm]{legume-tree} \caption[Carmichaelinae phylogenetic tree]{Phylogenetic tree of the New Zealand Carmichaelinae clade (in bold) and related genera (Galegeae tribe) from Australia and other countries, using ITS sequences. This penalised likelihood rate-smoothed Bayesian consensus phylogeny and the estimated ages were derived from the data and information provided in \citet{Wagstaff99}. Carmichaelinae are mainly of Australia but with two independent lines in New Zealand. Figure modified from \citet{Lavin04}, misspelling of `\emph{Swainsona}' in original. Symbols: * -- \emph{Clianthus puniceus.}} \label{p-legume-tree} \end{figure} \citet{Lavin04} extended the work of \citet{Wagstaff99} by reanalysing the data using Bayesian methods to estimate the age of divergence of each clade (Fig.~\ref{p-legume-tree}). From these data it appears that the New Zealand Carmichaelinae, including all \emph{Carmichaelia} species and \emph{Clianthus} (marked on the tree by `*') diverged 5.3$\pm$1.1 mya, and all Carmichaelinae have a common origin 7.5$\pm$0.8 mya. \emph{Carmichaelia} shares a common ancestor with \emph{Sutherlandia} (found in Australia, Africa, and Mauritius \citep{ILDIS}) 10.4$\pm$2.0 mya. These dates agree with those from fossil pollen. In a large study of 235 genera using \emph{matK} sequence data, \citet{Wojciechowski04} included \emph{Clianthus} and \emph{Carmichaelia} in a larger ``Astragalean clade'' including \emph{Swainsona}, \emph{Colutea}, \emph{Sutherlandia}, \emph{Oxytropis}, and \emph{Astragalus}. These publications show that the radiation of Carmichaelinae legume species into New Zealand was quite recent (compared to the diversification of legumes in the Northern Hemisphere). The ancestor of the Carmichaelinae derived from a Northern Hemisphere lineage and arrived (probably in Australia) between 10 and 7.5 mya. \subsection{New Zealand \emph{Sophora}} \emph{Sophora} is distinct from the other legume genera of New Zealand, being a member of the Sophoreae tribe (Table \ref{taxbox-Native}). \emph{Sophora} is a diverse genus that has about 80 members spread throughout the world. Molecular analyses indicate that the genus is polyphyletic, and and comprises three distinct and unrelated lineages \citep{Kass95,Kass96,Crisp00,Pennington01}. New Zealand \emph{Sophora} belong to a subset known as ``\emph{Sophora} sect.~\emph{Edwardsia}'' \citep{Kass97,Heenan04}. This sector is one of the largest groups in \emph{Sophora}, and includes about 19 species whose distribution is centred on islands in the southern Pacific Ocean. Most species of sect.~\emph{Edwardsia} have identical ITS sequences, indicating a recent and rapid radiation \citep{Mitchell02}. There are competing theories on the origin of \emph{Sophora} sect.~\emph{Edwardsia}. Some believe that they originated in Chile from a North American ancestor \citep{Sykes68,Pena00}. Molecular genetics indicates the likely origin is from the North Western Pacific, from an Eurasian ancestor, in the last 2--5 million years, and dispersal occurred around the pacific \emph{via} the buoyant saltwater-resistant seeds \citep{Hurr99,Mitchell02,Heenan04}. In summary it is proposed that New Zealand \emph{Sophora} spp.~derived from a separate legume lineage and geographical origin than the Carmichaelinae, and were dispersed to New Zealand perhaps a few million years later. \section{Exotic weed legumes in New Zealand} \subsection{Introduction} The indigenous people of New Zealand---the M\=aori---arrived in the mid 13th century from Eastern Polynesia \citep{Irwin05}. They brought with them new species, such as mammals (rats, dogs) and tuber plants (k\=umera, taro, yam), but there is no evidence that they brought any legumes \citep{Bellich96}. The first exotic legumes were introduced into New Zealand by settlers from Europe in the early 19th century. The settlers brought many plants and animals to establish familiar industries, and to remind them of their previous homelands. In their endemic habitats, these shrubs are in equilibrium with their natural flora, but in New Zealand, some have become serious invasive noxious weeds. Legumes have several properties that make them successful invaders. They have a high seed production, often with many seeds per pod, and many pods per tree. Most legume seeds are able to survive long periods in soil banks due to their thick impervious testa \citep{Lee01}. The mature plant generally lives for many years, and high-density seedling success allows rapid coverage of large areas. Possibly the success of legumes as invasive weeds is augmented by their ability to grow in nutrient deficient soils, in association with nitrogen-fixing rhizobia. There are now over 100 naturalised legume species in New Zealand \citep{NZPlant}. A small number of these have become common weeds and include: \emph{Chamaecytisus palmensis} (tree lucerne), \emph{Cytisus scoparius} (broom), \emph{Galega officinalis} (goat's rue), \emph{Lathyrus latifolius} (everlasting pea), \emph{Lotus pedunculatus} (lotus), \emph{Lotus suaveolens} (hairy birdsfoot trefoil), \emph{lupinus arboreus} (tree lupin), \emph{Medicago lupulina} (black medick), \emph{Medicago sativa} (lucerne), \emph{Melilotus indicus} (King Island melilot), \emph{Ornithopus perpusillus} (wild serradella), various wattles (\emph{Acacia} spp.), \emph{Psoralea pinnata} (dally pine), various \emph{Trifolium} spp.~(clover), \emph{Ulex europaeus} (gorse), \emph{Vicia hirsuta} (hairy vetch), \emph{Vicia sativa} (vetch) \citep{Roy04}. The woody species of \emph{Ulex}, \emph{Cytisus}, and \emph{Acacia} are the most invasive, and are the three of this study. \begin{table} \caption{Weed legume taxonomic hierarchy} \centering \begin{tabular}{|c|c|c|c|} \hline \textbf{Kingdom} & \multicolumn{3}{c|}{Plantae} \\ \hline \textbf{Division} & \multicolumn{3}{c|}{Magnoliophyta} \\ \hline \textbf{Class} & \multicolumn{3}{c|}{Magnoliopsida} \\ \hline \textbf{Order} & \multicolumn{3}{c|}{Fabales} \\ \hline \textbf{Family} & \multicolumn{3}{c|}{Fabaceae} \\ \hline \textbf{Subfamily} & \multicolumn{2}{c|}{Faboideae} & Mimosoideae\\ \hline \textbf{Tribe} & \multicolumn{2}{c|}{Genisteae} & Acacieae\\ \hline \textbf{Genus} & \emph{Ulex} & \emph{Cytisus} & \emph{Acacia} \\ \hline \end{tabular} \label{taxbox-Weed} \medskip \raggedright {\footnotesize \hskip 3cm Note: Information from \citet{ILDIS}} \end{table} \emph{Ulex} and \emph{Cytisus} are classified in the Genisteae tribe, and \emph{Acacia} is in the Acacieae tribe (Table~\ref{taxbox-Weed}). These are distinct from the woody native New Zealand legumes, which belong to the tribes Sophoreae and Carmichaelinae. %-------------------------------------------------------------------------------- \subsection{\emph{Ulex europaeus}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Ulex} \caption[\emph{Ulex europaeus}]{\emph{Ulex europaeus} showing spines and flowers beginning to open.} \label{p-Ulex} \end{figure} \emph{Ulex europaeus} L.~is known in the vernacular as whin, furze, or more commonly in New Zealand---gorse (Fig~\ref{p-Ulex}). There are some eleven \emph{Ulex} species but only \emph{U.~europaeus} is important in New Zealand \citep{ILDIS}. Its habitat is mostly disturbed and modified ecosystems, including river-beds, pasture, scrubland, forest margins and wasteland. Gorse is native to Western Europe and was naturalised in New Zealand in 1867 \citep{Bellingham04}, although \citeauthor{Darwin09} recorded it at Waimate some thirty years earlier in December 1835. It was introduced as a `living fence', but outgrew its usefulness and was soon classified as a weed. It is now considered New Zealand's worst weed, and millions of dollars are spent annually in control \citep{Hill86}. Gorse is also a problem in parts of Spain, Portugal, Chile, Hawaii, Ireland, coastal Oregon, and Southern Australia \citep{Roy04,Leary06} %Recently however some biocontrol agents, the gorse spider mite and %... ask Chris. Gorse in some situations may be beneficial in the %long term (25+ years) as mature gorse strands thin out and die %allowing other vegetation to grow. However burning resets this %cycle. %-------------------------------------------------------------------------------- \subsection{\emph{Cytisus scoparius}} \begin{figure} [tb] \centering \includegraphics[width=12cm]{Cytisus} \caption[\emph{Cytisus scoparius}]{\emph{Cytisus scoparius}. Photo \cc Jon J. Sullivan.} \label{p-Cytisus} \end{figure} \emph{Cytisus scoparius}~(L.)~Link, is also classified as \emph{Sarothamnus scoparius} (L.) W.D.J. Koch\footnote{Occasionally incorrectly spelt as \emph{Sarathamnus}.}. It is commonly called `broom' or `scotch broom'. \emph{Cytisus} has some 51 taxa \citep{ILDIS}, but only \emph{C.~scoparius} is of importance in New Zealand (Fig.~\ref{p-Cytisus}). Broom is common throughout New Zealand, particularly on the drier eastern side of the South Island, and the central North Island \citep{Fowler00}. Its habitat is mostly river-beds, hedgerows, low-fertility hill country, scrubland, coastal areas, and waste land. It was originally from Europe, Asia, and Russia. In New Zealand it grows more vigorously than in its native range, with a greater maximum age and larger size. It was naturalised in New Zealand in 1872 \citep{Bellingham04}. Broom causes economic losses to agricultural and forestry operations, and occupies 0.92\% of South Island farmable land \citep{Fowler00}. %-------------------------------------------------------------------------------- \subsection{\emph{Acacia}} \emph{Acacia} (commonly called wattle) is a large genus with over 950 species \citep{ILDIS}. Recent studies have shown that \emph{Acacia} is polyphyletic and should be split into five genera \citep{Luckow05}, although there are competing proposals for this. `Proposal 1584' would retypify \emph{Acacia}: The type of the Australian taxon (\emph{A.~penninervis}) would be conserved over the current lectotype (\emph{A.~scorpioides}) of an African taxon \citep{Orchard05}. Alternate proposals keep the lectotype, and reclassify some \emph{Acacia} species as `\emph{Racosperma}' \citep{Luckow05}. `\emph{Acacia}' will be used for the Australian species in this thesis\footnote{A summary of the events relating to the renaming of \emph{Acacia} can be found at \url{http://www.worldwidewattle.com/infogallery/nameissue/chronology.php}}. \begin{figure} [tb] \centering \includegraphics[width=12cm]{Acacia} \caption[\emph{Acacia longifolia}]{\emph{Acacia longifolia}. Photo \cc Brenda Foran.} \label{p-Acacia} \end{figure} \emph{Acacia longifolia} (Andrews) Willd. (Sydney golden wattle) is investigated in this study. It is a serious invasive weed in Northland, where it was introduced to control sand dune erosion, but has now spread widely and invaded wetlands \citep{Hicks01}. \emph{A.~longifolia} is also a significant problem in South Africa endangering the floristically unique Cape Floral Kingdom \citep{Dennill99,VanWilgen04}. \section{Previous research on New Zealand rhizobia} Work on this project started in early 2002. At this time there were few reports of rhizobia nodulating native legumes apart from an Honours dissertation using a small number of strains \citep{McCallum96}, and work in the 1960's--70's \citep{Greenwood69,Greenwood78a,Greenwood78b}. Likewise, there were no investigations using molecular techniques of rhizobia nodulating gorse and broom in any country, although historical research lumped the strains into the inaccurately described `cowpea rhizobia' \citep{Pieters27,Wilson39a}. It is not until recently that molecular techniques allowed affordable and accurate assessment of the phylogeny of bacterial strains. In early studies, many host-range experiments were done \citep{Greenwood69,Jarvis77,Crow81}, but interpretation of the data was difficult, as then the molecular mechanisms behind nodulation were not known, nor was it known that symbiosis genes were transmissible. A more comprehensive account of previous \emph{Rhizobium}--legume research in New Zealand is presented in Section \ref{s-NZ-rhizobia-history}. This thesis will build on this previous work, assisted by modern techniques and knowledge. \section{Research objectives} This thesis aims to establish a better understanding of the nature (taxonomy, diversity, host-range, and distribution) of the associations between rhizobia and New Zealand's endemic and weed legume flora. It is assumed that the native legume genera have co-evolved with nitrogen-fixing bacterial symbionts for millions of years, potentially leading to new species. In contrast the origin of rhizobia nodulating the recently introduced exotic legumes is unknown. Previous studies overseas \citep[reviewed by][]{Perret00} have shown that rhizobial strains differ in host-range specificity. Some (e.g. \emph{Ensifer fredii} NGR234) are promiscuous, while others appear to be host specific (e.g. \emph{Rhizobium leguminosarum} bv. \emph{trifolii}). Based on this, there are three possibilities that could explain exotic legume nodulation: 1) Introduced legumes are promiscuous and use the same rhizobia as native legumes. 2) Introduced legumes use specific rhizobia that were recently introduced---perhaps in conjunction with exotic legumes. 3) Introduced legumes use specific rhizobia that were already present in New Zealand (possibly cosmopolitan). %Either bradyrhizobia exist as ubiquitous free-living autochthons, or %they are carried with the plants when introduced into pristine %environments, perhaps with seed. Specific objectives for this thesis are: \begin{itemize} \item To establish the identity and diversity of the rhizobial species associated with New Zealand's endemic legume species. \item To establish the identity and presumptive origins of the rhizobial species associated with the woody legume weeds introduced into New Zealand. \item To determine the specificity and efficacy of the symbiotic associations of rhizobial species with both plant groups, endemic and woody weeds, by an investigation of their nodulation and nitrogen-fixing capacity. \item To investigate possible exchange of transmissible genetic elements between rhizobial species associated with endemic and introduced legumes. \end{itemize} \section{Research strategy} To investigate the identity and diversity of rhizobia, a polyphasic strategy employing both phenotypic and phylogenetic characteristics was used \citep{Vandamme96}. Phylogenetic analyses were based on the sequencing of three protein-coding conserved `housekeeping' genes (\emph{atpD}, \emph{glnII}, \emph{recA}), and one ribosomal RNA gene (16S). Phenotypic characteristics included metabolic fingerprints based on substrate utilisation (Biolog), and whole cell fatty acid methyl ester profiles (FAME). The symbiosis genes were investigated by sequencing a protein-coding gene (\emph{nodA}) involved in \emph{Rhizobium}--plant signalling, which is usually carried on a transmissible genetic element (plasmid or symbiosis island). The efficacy of the symbiotic combinations was tested by inoculating legumes with rhizobial strains in host-range experiments. The potential to fix nitrogen was determined by acetylene reduction, and roots were visually examined for nodulation. %%----------------------FIN-------------------------------------------