Auxin influx activity is associated with Frankia infection during actinorhizal nodule formation in Casuarina glauca

Plants from the Casuarinaceae family enter symbiosis with the actinomycete Frankia leading to the formation of nitrogen-fixing root nodules. We observed that application of the auxin influx inhibitor 1-naphtoxy acetic acid (1-NOA) perturbs actinorhizal nodule formation. This suggests a potential role for auxin influx carriers in the infection process. We therefore isolated and characterized homologues of the auxin influx carriers ( AUX1-LAX ) genes in Casuarina glauca . Two members of this family were found to share high levels of deduced protein sequence identity with Arabidopsis AUX-LAX proteins. Complementation of the Arabidopsis aux1 mutant revealed that one of them is functionally equivalent to AUX1 and was named CgAUX1 . The spatial and temporal expression pattern of CgAUX1 promoter:GUS reporter was analyzed in Casuarinaceae. We observed that CgAUX1 was expressed in plant cells infected by Frankia throughout the course of actinorhizal nodule formation. Our data suggest that auxin plays an important role during plant cell infection in actinorhizal symbioses.


INTRODUCTION
Actinorhizal plants, which belong to eight families of angiosperms, can form nitrogen-fixing nodules in symbiosis with the soil actinomycete Frankia (Benson and Silvester, 1993). The symbiotic interaction starts, in condition of nitrogen deprivation, by an exchange of signals between the plant roots and the bacteria. The chemical nature of Frankia nodulation factor(s) is unknown but data suggest that it has different biochemical properties from that of Rhizobium (Cérémonie et al., 1999). During intracellular infection, Frankia signals lead to root hair deformation, some of which become infected. At the same time, limited cell divisions are triggered in the cortex creating a so-called prenodule. Prenodule function is not known but it is an obligatory step of intracellular infection (Laplaze et al., 2000). Concomitently, cell divisions occur in the pericycle in front of xylem pole leading to the formation of a nodule lobe primordium. The growing nodule lobe is infected by Frankia hyphae coming from the prenodule. The structure of the new organ formed upon infection largely differs from Legume nodules even if the infection mechanisms share common features (Pawlowski and Bisseling, 1996). Actinorhizal nodules are considered as modified lateral roots because i) they originate from divisions in the pericycle in front of a xylem pole, ii) they have a lateral root like structure with a central vasculature, infected cells in the cortex and an apical meristem and iii) in some species (e.g. Casuarina sp.) a so-called "nodule root" is produced at the apex (Obertello et al., 2003). Little is known about the mechanisms of actinorhizal nodule development.
The plant hormone auxin is involved in many developmental processes (Tanaka et al., 2006) and is the key signal controlling lateral root development (Casimiro et al., 2003). Auxin transport across the plant is polarized and perturbations of polar auxin transport (PAT) using inhibitors such as NPA (naphthylphthalamic acid) or mutants result in dramatic alteration of the plant developmental pattern (Reed et al., 1998). The existence of auxin transporters has been predicted for a long time to account for PAT (Goldsmith, 1977). Characterization of Arabidopsis thaliana mutants perturbed in auxin transport or sensitivity led to the identification of auxin efflux and influx facilitators encoded by the PIN and AUX-LAX genes respectively (Kramer and Bennett, 2006). The latter are encoded by a small gene family (4 genes) in Arabidopsis (Parry et al., 2001a).

6
Only one member of the AUX-LAX family has been characterized to date : AUX1 is involved in gravitropism (Bennett et al., 1996) and lateral root initiation (Marchant et al., 2002). AUX1 has recently been shown to encode a high affinity auxin influx transporter by heterologous expression in Xenopus oocytes (Yang et al., 2006). The mechanism of transport remains to be elucidated but is predicted to occur by proton symport (Kerr and Bennett, 2007).
Auxin transport is also thought to be involved in the establishment of Legume symbiosis. A local auxin transport inhibition is triggered by spot-inoculation of rhizobia, leading to a subsequent accumulation of auxin at the site of infection as shown by the use of the GH3:gusA auxin response marker in white clover (Mathesius et al., 1998) and in Lotus japonicus (Pacios-Bras et al., 2003). In legumes forming indeterminate nodules, flavonoids are produced as a response to bacterial lipochitin oligosaccharides (Mathesius et al., 2000) and act as inhibitors of auxin efflux transport (Brown et al., 2001) leading to local accumulation of auxin necessary for cell division and subsequent nodule primordium formation (Wasson et al., 2006). Moreover, the expression of auxin influx transporters in Medicago is associated with nodule primordium development and vasculature differentiation (de Billy et al., 2001).
A role of auxin during the actinorhizal symbionts dialog has also been suggested because some Frankia strains can produce different forms of auxin in culture (Gordons et al., 1988;Hammad et al., 2003). However, no link has been made between the production of hormones by Frankia and the establishment of the symbiosis. The symbiotic bacteria Rhizobium produce auxins that were proposed to be involved in establishing the symbiosis with Legume plants (Badenoch-Jones et al., 1983). Indeed, a Bradyrhizobium japonicum mutant producing 30-fold more IAA than the wild-type has a higher nodulation efficiency (Kaneshiro and Kwolek, 1985). Altogether, up to 80% of rhizobacteria are considered to produce auxins (Patten and Glick, 1996). However, nothing is known about the precise role of bacterial auxin during the processes of infection and symbiosis or how and when the plant cell perceives it.
In this study, we show that application of the auxin influx inhibitor 1-naphtoxyacetic acid (1-NOA) perturbs nodule formation. We therefore isolated a small family of AUX-LAX genes homologues in the actinorhizal plant Casuarina glauca. Among this family of genes, we identified CgAUX1, a homologue of AtAUX1, that carries an auxin carrier 7 function as shown by functional complementation of the Arabidopsis aux1 mutant. The expression of CgAUX1 is found in all Frankia infected cells from the root hair to nodule nitrogen-fixing cells. We also bring evidence of differences between the genetic programs of lateral root and actinorhizal nodule primordium based on different patterns of CgAUX1 expression. Altogether, our results shed new light on the role of auxin influx transport during actinorhizal nodule formation.

Inhibition of Auxin Influx Transport Using 1-NOA Perturbs Nodule Formation
We analyzed the effect of 1-naphtoxyacetic acid (1-NOA), a competitive inhibitor of auxin influx, on C. glauca-Frankia interaction. 1-NOA is known to specifically inhibit AtAUX1 (Yang et al., 2006) and to mimic the aux1 mutant phenotype in Arabidopsis (Parry et al., 2001a). C. glauca plants were inoculated and grown in hydroponics in the presence of 25 µM 1-NOA. The number of nodulated plants (i.e. plants bearing prenodules or nodules) was checked every day after 10 days ( Fig. 1A). We found in 3 independent experiments that 1-NOA treatments caused a 2 days delay in nodule appearance. The same effect was observed if the growth medium was changed every three days with fresh 1-NOA to prevent a potential 1-NOA degradation (data not shown). Moreover, 24 days after inoculation plants treated with 1-NOA mainly showed prenodules while control plants showed nodules (Fig 1B and C). This 1-NOA effect on nodulation was not due to a more general effect on root growth as we found no significant differences in shoot or root weight in treated or non treated plants (Student test at P<0.1; NT roots m=0.059g ; NOA treated roots m=0.062g ; NT shoots m=0.136g ; NOA treated shoots m=0.137g ; dry weights ; n=20). Moreover, we also verified that addition of 25 µM 1-NOA had no deleterious effects on Frankia growth (Fig. 1D). We therefore conclude that inhibition of auxin influx transport using 1-NOA partially perturbs actinorhizal nodule formation in C. glauca.

Identification of a Small Family of Auxin Influx Carrier Genes in C. glauca
Our data suggest a role for auxin influx carriers encoded by AUX1 homologues during actinorhizal nodule development. AUX-LAX gene homologues were therefore 8 isolated from C. glauca by amplifying genomic DNA with different sets of degenerate primers (Table I) designed in conserved regions of AUX-LAX proteins of Arabidopsis, Medicago truncatula and Poplar. Seven different PCR products were produced, sequenced and found to correspond to two different genes. The corresponding cDNAs (1440 bp and 1395 bp) were obtained by rapid amplification of cDNA ends (RACE-PCR). They were named CgAUX1 and CgLAX3 according to sequence identity of the predicted proteins to Arabidopsis proteins, 85 % for AUX1 and 87 % for LAX3 respectively.
The genomic sequences corresponding to CgAUX1 and CgLAX3 were amplified by PCR and found to be 2942 bp and 2224 bp long respectively from start to stop codon. suggesting that this gene structure preceded the divergence of monocots and dicots and that a loss of introns is responsible for the observed differences in intron/exon number.
In order to estimate the number of AUX-LAX genes in C. glauca genome, we conducted Southern blot experiments using three different probes : a non-specific probe designed in one of the most conserved region of AUX-LAX genes (exon VII) and two gene-specific probes designed in CgAUX1 and CgLAX3 3' untranslated regions. The conserved probe hybridized with a limited number of genomic DNA fragments in nonstringent conditions that could be assigned to either CgAUX1 or CgLAX3 using the gene-specific hybridizations (Fig. 2B). This together with the fact that we did not recover any other gene by PCR or in a C. glauca EST library (Hocher et al., 2006) suggests that auxin influx carriers are encoded by a small gene family (possibly only 2 genes) in C. glauca.

CgAUX1 Encodes an Auxin Influx Carrier Functionally Equivalent to AtAUX1
Arabidopsis and C. glauca AUX-LAX deduced protein sequences were compared with a representative member of each class of the amino acid transporters family (ATF).
A phylogenetic tree was generated using neighbor-joining distance algorithm showing that AUX-LAX proteins belong to the amino-acid and auxin permease (AAAP; Young et al., 1999) family (Fig. 3A). Among the AUX-LAX proteins, two sub-classes could be defined containing AtAUX1, CgAUX1 and AtLAX1 for the first sub-class and We tested whether CgAUX1 and CgLAX3 encode functional auxin influx carrier proteins equivalent to Arabidopsis AUX1 by carrying out a complementation analysis.
CgAUX1 and CgLAX3 ORF were inserted between AtAUX1 promoter (Pro AtAUX1 ) and terminator sequences in a binary vector and transformed into null aux1-22 mutants. We then analyzed if that was sufficient to restore a gravitropic phenotype in T1 plants 8 days after germination. aux1-22 plants transformed with an empty vector containing AtAUX1 promoter and terminator sequences are agravitropic (Fig. 4A). In contrast, transformation with a vector expressing the AtAUX1 coding sequence under its own promoter and terminator rescued a wild-type gravitropic phenotype (Fig. 4A). In the same conditions, CgAUX1 was able to rescue a gravitropic phenotype to aux1 (Fig. 4A).
However, expressing CgLAX3 under the control of AtAUX1 promoter and terminator in aux1-22 mutant background could not restore a wild-type phenotype (Fig. 4A)

even if
CgLAX3 transcripts were detected in the transgenic plants (Sup Fig. 1 We also checked whether CgAUX1 was sensitive to 1-NOA by attempting to disrupt the complementation of aux1 root gravitropism by CgAUX1. 25 µM 1-NOA treatment leads to a reversion to the mutant agravitropic phenotype (Fig. 4B)  We then focused our expression analysis on CgAUX1 because it encodes a functional auxin influx transporter. We cloned a 1.7 kb promoter fragment and fused it to the βglucuronidase (GUS) reporter gene sequence in a binary vector thus creating the Pro CgAUX1 :GUS construct. This construct was introduced into C. glauca and its close relative Allocasuarina verticillata by Agrobacterium tumefaciens mediated genetic transformation (Franche et al., 1997). Similar patterns of expression were obtained in these two species. CgAUX1 is expressed in root tips (Fig. 6A) and in lateral root primordia ( Fig. 6B and C). Expression was also observed in the root (Fig. 6B) and shoot vasculature (data not shown). This expression pattern is very similar to AtAUX1 expression pattern in Arabidopsis (Marchant et al., 2002). This together with the complementation results suggest that CgAUX1 is orthologous to AtAUX1 and is involved in the same biological processes (gravitropism and lateral root development) as AtAUX1.
We then analyzed CgAUX1 expression during the symbiotic interaction with Frankia. Pro CgAUX1 :GUS expression was studied 2, 7, 10, 14 and 21 days after inoculation (8 transgenic C. glauca plants/time point). All of the plants showed the same expression pattern. CgAUX1 expression was detected very early in very few root hairs from 10 days post-inoculation (Fig. 7A, C-F). Infecting Frankia hyphae were found in CgAUX1 expressing root hairs (Fig. 7F). At the same time, a higher expression level is clearly visible in the vasculature at the site of infection (Fig. 7A, B C and D). At later stages, CgAUX1 expression is associated with the infection process. Nodule sections showing a strong staining in the cortical cells that are infected and no staining in the non-infected cells further confirm this pattern of expression (Fig. 7G, H, I and J).
Surprisingly, CgAUX1 is not expressed in the nodule primordium (Fig. 7B). This lack of expression in nodule primordia is confirmed by the analysis of nodule ramifications (Fig. 7K). We therefore found that CgAUX1 expression was associated with Frankia infection from the first stage of infection but was excluded from nodule primordia.  Fig. 2). We also found no effect of the plant nitrogen status on CgAUX1 expression in response to auxin (Sup Fig. 2).

DISCUSSION
The results presented here suggest that auxin influx activity is important for the symbiotic interaction between C. glauca roots and the soil actinomycete Frankia. We first show that competitive inhibition of auxin influx using 1-NOA delays nodulation and confirms the involvement of auxin carriers in the process. This led us to isolate two members of a small family of auxin influx carrier genes in C. glauca. We found that CgAUX1 can complement the Arabidopsis aux1 mutant while CgLAX3 could not.
AtAUX1 was demonstrated to encode an auxin influx carrier in the Xenopus oocyte (Yang et al., 2006). We therefore conclude that CgAUX1 also encodes for an auxin influx carrier equivalent to AtAUX1.
The actinorhizal nodule is classically regarded as a modified lateral root (Pawlowski and Bisseling, 1996;Obertello et al., 2003). However, we observed that CgAUX1 is expressed in lateral root primordia but not in nodule lobe primordia. These results suggest that these two organs have, at least in part, divergent development programs. This is in agreement with previous observations showing that some heterologous promoters used as molecular markers such as 35S and AtUBQ1 drive different expression patterns in lateral root and nodule primordia in Casuarinaceae plants (Obertello et al., 2005). Nevertheless, since our analysis is only based on a promoter-GUS fusion, we cannot completely rule out that CgAUX1 is expressed in nodule primordia. We cannot exclude either that another AUX-LAX gene, such as CgLAX3, is involved in actinorhizal nodule primordium formation. Further studies will be needed to understand how much of the lateral root developmental program has been recycled during evolution to create the actinorhizal nodule developmental program. By Interestingly, we found that CgAUX1 expression is closely associated with Frankia infection of plant cells during nodulation (summarized in figure 8A). We observed CgAUX1 expression already in Frankia-infected root hairs 10 days after infection. Frankia infected cells would make them more permeable to auxin. Interestingly, some Frankia strains have been shown to produce different forms of auxin in culture, including IAA and PAA (Wheeler et al., 1984;Hammad et al., 2003). This could explain why actinorhizal nodules have been reported to contain more auxin than noninfected roots (Wheeler et al., 1979). We therefore speculate that CgAUX1 expression 13 allows the entry and perception of Frankia-produced auxin and restricts it to infected plant cell (Fig. 8B). Auxin alone or in synergy with a symbiotic signal could induce changes in gene expression, cell metabolism etc, in infected cells to allow the establishment of the intracellular symbiosis (Fig. 8B). For example, the infection process is associated with remodeling of the cell wall to create an infection thread (Berg, 1999a). Many cell wall remodeling genes have been found to be auxin inducible in Arabidopsis (Neuteboom et al., 1999;Overvoorde et al., 2005;Esmon et al., 2006;Osato et al., 2006). Auxin could therefore induce genes encoding cell wall remodeling enzymes necessary for the infection by Frankia. Moreover, infected cells are hypertrophied (Berg, 1999b), a phenotype that has been classically associated to auxin response (Teale et al., 2006). Further experiments will be needed to understand the interaction between cell wall remodeling and auxin transport during nodule formation.

Plant material and growth conditions
Casuarina glauca seeds purchased from Carter Seeds (California, USA) were grown and inoculated by Frankia CcI3 strain as previously described (Franche et al., 1997).

Identification of CgAUX1 and CgLAX3 cDNA and genomic sequences
C. glauca genomic DNA was isolated from young shoot apex using a MATAB extraction method (Ky et al., 2000). Amplification of AUX1 homologues was performed on genomic DNA using different sets of degenerate primers (Table I). Amplified fragments were cloned into pGEM-T easy (Promega) and sequenced. Total RNA was extracted on whole root system by ultracentrifugation (Chirgwin et al., 1979).

Non-quantitative and quantitative RT-PCR
Total RNA was extracted on whole root system, shoot or mature nodules by ultracentrifugation (Chirgwin et al., 1979). Poly(dT) cDNA was prepared out of 1 µg total RNA using Reverse Transcription System (Promega) and 3 independent RT reactions were pooled for quantitative analysis. PCR reactions were carried out at 94°C for 5 min, followed by 40 cycles of denaturation for 30 sec at 94°C, annealing for 30 sec at 60°C and extension for 90 sec at 72°C. Target amplifications were performed with CgAUX1 or CgLAX3 specific primer pairs designed on each side of the last intron

Constructs and generation of transgenic plants
For promoter studies, 1.7 kb genomic DNA fragments upstream of CgAUX1 and CgLAX3 start codon (ATG) were amplified using the Universal GenomeWalker Kit

Microscopy and root sections
GUS assays were performed as previously described (Svistoonoff et al., 2003).
Tissues were cleared in 70% ethanol for 2 d and then immersed in 50% (v/v)