Vapyrin, a gene essential for intracellular progression of arbuscular mycorrhizal symbiosis, is also essential for infection by rhizobia in the nodule symbiosis of Medicago truncatula


(fax +1 580 224 6692; e-mail


Intracellular invasion of root cells is required for the establishment of successful endosymbioses in legumes of both arbuscular mycorrhizal (AM) fungi and rhizobial bacteria. In both interactions a requirement for successful entry is the activation of a common signalling pathway that includes five genes required to generate calcium oscillations and two genes required for the perception of the calcium response. Recently, it has been discovered that in Medicago truncatula, the Vapyrin (VPY) gene is essential for the establishment of the arbuscular mycorrhizal symbiosis. Here, we show by analyses of mutants that the same gene is also required for rhizobial colonization and nodulation. VPY encodes a protein featuring a Major Sperm Protein domain, typically featured on proteins involved in membrane trafficking and biogenesis, and a series of ankyrin repeats. Plants mutated in this gene have abnormal rhizobial infection threads and fewer nodules, and in the case of interactions with AM fungi, epidermal penetration defects and aborted arbuscule formation. Calcium spiking in root hairs in response to supplied Nod factors is intact in the vpy-1 mutant. This, and the elevation of VPY transcripts upon application of Nod factors which we show to be dependent on NFP, DMI1, and DMI3, indicates that VPY acts downstream of the common signalling pathway.


Legumes establish mutually beneficial symbioses with soil bacteria and fungi in which photosynthetically derived carbon is provided to the microbes by the plant in return for other nutrients such as nitrogen and phosphorus that otherwise limit plant growth. The best studied and probably the most agriculturally important of these symbioses are the interactions between legumes and rhizobial bacteria, which provide the plant with reduced nitrogen, and between most terrestrial plants and arbuscular mycorrhizal (AM) fungi that provide the plant with phosphate and other nutrients. Despite obvious developmental differences, it is believed that nodulation evolved from the much more ancient AM interaction, and genetic studies have revealed that these symbioses share many features (Parniske, 2008).

Arbuscular mycorrhizal fungi enter plant roots either by intracellular penetration of an epidermal cell or less frequently by bypassing the epidermal cells (Genre et al., 2005, 2008). During intracellular passage the hyphae are always surrounded by the host plasma membrane and a thin layer of cell wall-like material (Bonfante and Genre, 2008). Subsequent to passage through the epidermis the fungal hyphae explore the root cortex apoplastically before establishing intracellular branched structures, called arbuscules, within inner cortical cells that are the site of nutrient exchange between the plant and fungus.

In a similar fashion, rhizobial bacteria are permitted entry into epidermal root cells via an invagination of the plant plasma membrane called an infection thread (IT), which is bounded by cell wall-like material (Vandenbosch et al., 1989). The IT typically progresses though an epidermal root hair cell and then descends into the underlying cortical cells until it reaches the rapidly dividing cells of the nascent nodule. At this point, the IT repeatedly branches and the bacteria are released from unwalled IT tips into the cytoplasm of the host cell by an endocytotic process, resulting in plant membrane-enclosed structures called symbiosomes. The bacteria within the symbiosomes divide, differentiate and then begin to fix atmospheric nitrogen into ammonia, which is exported for use by the plant. The intracellular phases of both AM and rhizobial symbioses involve cytoskeletal rearrangements which begin prior to cellular invasion, and are called the pre-penetration apparatus and the pre-infection thread, respectively. These apparatuses are associated with repositioning of the plant cell nucleus to a central position and eventually to a position adjacent to the cell wall near the point of entry. At this stage, a cytoplasmic bridge is formed across the cell which predicts the path of passage of the fungal hyphae or formation of the rhizobial infection thread (van Brussel et al., 1992; Genre et al., 2005). During this passage the nucleus closely associates with the tip of the hyphae or infection thread. Several genes controlling the early signalling events immediately preceding rhizobial and AM invasion have been identified (Endre et al., 2002; Ane et al., 2004; Levy et al., 2004; Messinese et al., 2007) as well as several transcription factors downstream of signalling that are required for rhizobial but not AM entry (Oldroyd and Long, 2003; Kalo et al., 2005; Smit et al., 2005; Marsh et al., 2007, Middleton et al., 2007). Recently, two new genes required for the formation of normal rhizobial infection threads, but having normal AM interactions, were identified: CERBERUS/LIN (Kiss et al., 2009; Yano et al., 2009), a WD40/UBOX protein, and RPG (Arrighi et al., 2008), a coiled-coil protein. In addition putative members of lipid rafts, flotillins and a remorin have been implicated in the infection process during nodulation (Haney and Long, 2010; Lefebvre et al., 2010). Here we describe a Medicago truncatula mutant exhibiting defects in entry and progression of both AM and rhizobial infections and identify the underlying gene as VAPYRIN (VPY) which has recently been implicated in AM symbiosis using an RNA interference (RNAi)-based approach (Pumplin et al., 2010).


Identification and characterization of nodulation mutants

Four M. truncatula mutants (FNB2, FNB4, FNB10 and FNB15) having impaired nodule development when infected with Sinorhizobium meliloti were identified from a single pool of bulked seeds from a fast-neutron mutant population (Wang et al., 2006). The phenotype with rhizobial infection was examined in M4 and backcrossed lines. The four mutants exhibited identical phenotypes in response to rhizobial inoculation. At 21 days post-inoculation (dpi) the mutants formed a few small white nodule primordia (Figure 1a,b) that were associated with abnormal ITs. These ITs varied greatly in their width (ranging from 1 to 10 μm compared with 1–2 μm in wild-type (WT) root hairs) and exhibited rough edges and blebbing compared with the WT ITs which were uniformly narrow (Figure 1c,d). Microcolony formation and root hair curling ranged from normal in appearance, to slightly enlarged microcolonies accompanied by exaggerated shepherd’s crook curls (Figure 1e,f). The observed infection threads were always blocked within the root hair cell and never penetrated into cortical cells. FNB4 was chosen for further study. Observations at 8 dpi, showed that the number of infection threads and uninfected nodule primordia on FNB4 were increased relative to WT; no infected nodules were found at 8 dpi (Figure 1g). At 42 dpi, zero to three large elongated slightly pink nodules were present on the mutants and the plants exhibited symptoms of severe nitrogen deprivation including yellowing of the leaves, stunting and anthocyanin accumulation. Plants developing these pink nodules grew noticeably larger than those that didn’t, but were still severely stressed, suggesting that some limited nitrogen fixation occurred.

Figure 1.

 Nodulation phenotypes of FNB4 and wild-type (WT) control roots at 21 days post-inoculation (dpi) with Sinorhizobium meliloti 1021.
(a) Wild-type nodules and (b) uninfected small nodules on FNB4. Scale bars = 1 mm.
(c, e) Histochemical staining of lacZ-tagged S. meliloti in root hairs in the WT.
(d, f) Histochemical staining of lacZ-tagged S. meliloti in root hairs in FNB4. FNB4 infections have normal root hair curling and microcolonies and infection threads (ITs) have non-uniform width and blebbing at the margins. Scale bars in (c–f) = 10 μm.
(g) Nodulation phenotypes (events per plant) using staining of lacZ-tagged rhizobia at 8 dpi. White bars = WT; grey bars = FNB4; NOD = bacteria inside nodule; UNP = uninfected nodule primordia; IE = infection events. Symbiotic phenotypes were scored as follows: all infection events (IE) were counted including infection pockets and very short infection threads. Nodules were classified as either infected nodules (NOD) or uninfected nodule primordia (UNP) depending on whether they had blue staining inside the nodule. Superficial infections in outer cell layers were scored as uninfected nodule primordia.
(h) Representative calcium spiking responses of M. truncatula roots in response to application of nodulation factor (NF). For WT and FNB4, 7 out of 10 cells from four plants and 10 of 12 cells from three plants exhibited spiking, respectively.

FNB4 has normal calcium oscillations in response to Nod factors

In order to position the mutant relative to known symbiotic mutants, the ability of FNB4 to initiate calcium oscillations in response to nod factors (NF) was tested, as previously described (Sun et al., 2006), and was found to be normal (Figure 1h) indicating that early nodulation signalling is intact in this mutant.

The FNB4 mutant has reduced penetration and truncated arbuscule structures

The FNB4 mutant was inoculated with Glomus intraradices (DAOM 181602) or a mixture of G. intraradices and Glomus mossae. Quantitative mycorrhizal colonization assays indicated that FNB4 and the WT had similar overall levels of colonization except for the absence of arbuscules and an increased number of intraradical hyphae on the mutant (Figure 2). Fungi were able to grow on the root surface but were unable to produce normal appressoria, the hyphal structures formed on epidermal cells at the site of penetration (Genre et al., 2005). Instead, complex deformed appressoria that were often septate were formed on mutant roots, indicating that the fungus was not able to readily gain entry into plant epidermal cells (Figure S1a,b). The few hyphae that penetrated epidermal cells became bloated and did not form the epidermal coils typically seen in the WT before passing into outer cortical cell files (Figure S1c). Rarely, hyphae proceeded through the epidermal cells and colonized a substantial portion of the mutant root cortex with longitudinal intercellular hyphae (Figure S1d). The intercellular growth phase of the fungus in the root cortex of FNB4 resembled that of the WT in terms of production of lateral branches and formation of anastomoses, except that hyphae were septate in the mutant indicating growth retardation of the fungus (Figure S1d). Furthermore, hyphae were not able to form the highly branched arbuscules characteristic of successful mycorrhizal symbiosis within cortical cells of the mutant. Instead, penetration of the mutant cortical cells by fungal hyphae resulted in small lateral protrusions (Figure S1e). Despite having no arbuscules, vesicles were still able to form on the mutant (Figure S1f). Similar results were obtained when plants were inoculated using nurse plants.

Figure 2.

 Colonization of M. truncatula in the presence of leek plants colonized with Glomus intraradices.
Intraradical fungal hyphae (light grey bars), vesicles (white bars) and arbuscules (dark grey bars) were counted in plants grown in two separate pots and harvested 14, 28 and 42 days after inoculation. Standard errors of the mean for five to eight individual plants are shown.
(a) Wild type (WT) (A17).
(b) Mutant line FNB4.
H = hyphae only, without vesicles or arbuscules; V = vesicles; A = arbuscules.

Microarray-based cloning of the VPY gene

Considering that these mutants originate from fast neutron bombardment that can lead to DNA deletions, we attempted to detect deletions using microarrays in the hope of identifying the gene responsible for the mutant symbiotic phenotypes. A microarray (Affymetrix Medicago GeneChip) experiment was conducted using RNA isolated from S. meliloti-inoculated roots at 24 h post-inoculation for FNB2 and FNB4. Hundreds of genes were down-regulated relative to the WT control, making it difficult to identify missing transcripts corresponding to deleted genes in FNB2 and FNB4. In order to increase the specificity of the analysis, genomic DNA (gDNA) from all four mutants, FNB2, FNB4, FNB10 and FNB15, was used separately as a template to generate a biotinylated DNA probe for microarray hybridization. The distribution of mutant/WT hybridization signal ratios for individual probesets on the GeneChip revealed between 29 and 54 putative genomic DNA deletions in these lines (mutant/WT signal ratio <0.30; Figure S2a; Table S1). A comparison of mRNA and gDNA signals relative to the WT revealed that a number of the putatively deleted genes had correspondingly low transcript levels (Figure S2a). Notably, genes corresponding to 14 probesets appeared to be deleted in all four fast neutron lines, which strongly suggested that these lines shared at least one large deletion and were descended from a common M1 plant (Figure S2b; Table S1). The 14 deletions were confirmed by PCR analysis of all four mutant lines. A set of eight PCR-based markers corresponding to deletions co-segregated with both the AM and rhizobial symbiotic phenotypes in F2 segregating populations from backcrosses of FNB10 and FNB15, while another set of six markers co-segregated with one another but not with the symbiotic phenotypes, indicating at least one linked and one unlinked deletion. The observed frequency of mutants deviated significantly from a 1:3 ratio in both populations (FNB10, 15:123, χ2 = 14.696, P = 0.001; FNB15, 22:143, χ2 = 11.978, P = 0.0005) suggesting a recessive inheritance with slightly reduced survival of homozygous individuals. Examination of the expression profiles of the genes corresponding to the eight probesets linked to the mutant phenotypes, using data from the M. truncatula Gene Expression Atlas (MtGEA; Benedito et al., 2008) which includes data from nodulated as well as mycorrhized plants (Gomez et al., 2009), indicated that two probesets within the deletion were root-specific and were upregulated in nodules and mycorrhizal roots (Figure S2c). blast analysis of public databases failed to find matches to the Medicago genomic sequence for any of the eight probesets, indicating that the region has not been sequenced, but matches to the NCBI expressed sequence tag (EST) database indicated that the two probesets that detected symbiosis-enhanced gene expression (Mtr.39050.1.S1_at and Mtr.42828.1.S1_at) probably corresponded to a single gene. Polymerase chain reaction from gDNA was used to amplify a complete 1.7 kb open reading frame (ORF) that lacked introns and which confirmed that the ESTs used to design the two probesets originated from the same mRNA.

VPY encodes a major sperm domain-containing protein

The cloned gene encodes a putative 541 amino acid protein with an N-terminal Major Sperm Protein (MSP) domain and a C-terminal domain containing several ankyrin repeats (Figure S3a). The complete ORF was cloned downstream (3′) of the CaMV 35S promoter and the resulting construct was transformed into the FNB4 mutant using Agrobacterium rhizogenes (ARqua 1 strain)-mediated hairy root transformation. The ability to form WT nodules and arbuscules was restored in transformed roots (= 17 independent transformed roots for each treatment) demonstrating the function of the gene in both symbioses (Figure 3). The gene had been previously named VAPYRIN (VPY) (Pumplin et al., 2010). A PCR analysis indicated that this gene was entirely deleted in FNB4; this allele was designated as vpy-1. To further confirm the function of VPY, a collection of Tnt1-transposon mutagenized lines of M. truncatula were screened using PCR, which yielded two independent lines (NF6898 and NF4489) harbouring Tnt1-insertions in VPY at positions 660 and 1500 bp downstream of the start codon (Figure S3a). These mutant lines were designated as vpy-2 (NF6898) and vpy-3 (NF4489). Several plants with homozygous insertions in VPY (two for NF6898 and five for NF4489) were identified for each line and were found to have AM and rhizobial phenotypes identical to vpy-1 at 21 dpi (representative phenotypes are shown in Figure S3c–f). Progeny of selfed homozygous plants were grown and VPY gene expression levels in the roots were shown to be significantly reduced in these mutants (Figure S3b), possibly a result of nonsense-mediated decay resulting from Tnt1 transposon insertion. The MSP domain-containing proteins (MDPs) are present in plants, animals and fungi. One of the best studied is Scs2p from Saccharomyces cerevisiae, which is involved in membrane biogenesis. The Scs2p MSP domain contains a FFAT (two phenylalanines in an acidic tract) binding domain that mediates interactions with proteins involved in sterol regulation (Loewen and Levine, 2005), as well as conserved residues required for binding of phosphoinositides in an in vitro assay (Kagiwada and Hashimoto, 2007). Both of these domains are absent in the VPY family members but present in other classes of MDPs from Medicago and Arabidopsis (Figure S4).

Figure 3.

 Complementation of nodulation and arbuscular mycorrhizal (AM) phenotypes of the FNB4 (vpy-1) mutant with using Agrobacterium rhizogenes-mediated hairy root transformation.
(a) Pink elongated nodules in the vpy-1 mutant transformed with 35S::VPY infected with Sinorhizobium meliloti. Scale bar = 1 mm.
(b) Arbuscule formation (stars) in the vpy-1 mutant transformed with 35S::VPY and infected with Glomus intraradices. Scale bar = 50 μm.

VPY is induced during symbiotic interactions and encodes a protein that localizes to the nucleus and cytoplasm

In order to validate the microarray results and confirm the expression of VPY in the different plant organs, quantitative (q)RT-PCR was carried out using primers designed from the 3′ untranslated region (UTR) on cDNA derived from the same material that was used for the MtGEA microarray experiments as shown in Figure S2 (c). The qRT-PCR results were highly correlated (R2 = 0.88) with the original Medicago GeneChip results and show that VPY is more highly expressed in mycorrhizal roots and nodules than in roots, and shows minimal expression in aboveground tissues (Figure S5).

VPY is expressed preferentially in the root epidermis

Since the most overt rhizobial phenotype of the vpy mutants is abnormal ITs, root hairs were isolated from plants infected with rhizobia to examine VPY expression in this key cell type. Mtexp7, the Medicago orthologue of the root hair-specific Arabidopsis expansin7 (Cho and Cosgrove, 2002), was used as a marker for root hair enrichment. Using gene-specific primers in qRT-PCR assays, Mtexp7 transcripts were found to be more highly expressed in isolated root hairs relative to roots stripped of root hairs in the absence of rhizobia, indicating that root hair enrichment was successful (Table 1). Similarly VPY transcripts were also increased in root hairs (Table 1). VPY transcript levels increased in both the root hair-enriched fraction and the stripped root fraction following inoculation of intact roots with rhizobia (Table 1).

Table 1.   Relative expression of Vapyrin and Expansin7 in response to infection by Sinorhizobium meliloti strain ABS7
GeneTissueTreatmentRelative expression levelSEM
  1. *P < 0.05 relative to no rhizobia treatment (Student’s t-test).

Expansin7Stripped root0.00050.00008
Expansin7Stripped rootRhizobia0.00070.00028
Expansin7Root hairs0.16090.00768
Expansin7Root hairsRhizobia0.23530.04477
VapyrinStripped root0.00090.00010
VapyrinStripped rootRhizobia0.0033*0.00080
VapyrinRoot hairs0.11760.02625
VapyrinRoot hairsRhizobia0.2079*0.02224

VPY is induced by NF via the common symbiosis signalling pathway

The Medicago GeneChip was used to measure the effect of NF application on VPY expression in WT and mutant plants. VPY transcript levels increased substantially in response to NF application to roots of the WT but not of the nod factor receptor mutant nfp, or the signalling mutants dmi1 and dmi3 (Figure 4). Interestingly, NF induced the expression of VPY in the nsp1, nsp2, ern1 and nin mutants, which are defective in transcription factors downstream of the NF signalling pathway that are required for nodulation but not AM invasion (Figure 4). Notably the induction of VPY gene expression was reduced in the nsp1 and nsp2 mutants.

Figure 4.

VPY expression with 0.1 nm nodulation factor (NF) in the wild type (WT) (Jemalong A17) and the mutants nfp, dmi1, dmi3, nsp1, nsp2, ern1 and nin.
*P < 0.05, Student’s t-test, n = 3 biological replicates.


vpy is a new common symbiotic mutant with a unique phenotype

The vpy mutant phenotype is unique among common symbiotic mutants. Like ccamk and cyclops, vpy mutants are competent for Ca2+ spiking in response to NF, but ccamk mutants never form ITs (Catoira et al., 2000; Mitra and Long, 2004). Furthermore, the CCaMK dependence of VPY induction by NF suggests that VPY acts downstream. On the other hand, cyclops mutants occasionally form infection threads, but they are reduced in number, abort very early and the nodules never become infected (Yano et al., 2006, 2008). Notably, vpy mutants form more ITs than WT plants and these ITs progress further than those seen in cyclops, and eventually vpy mutants develop some infected cells in abnormal nodules. The hyperinfection of vpy roots may be a result of a feedback mechanism originating from failed infections, such as is seen in pea sym38 (Tsyganov et al., 2002).

The AM phenotypes of impeded epidermal penetration and inability to form arbuscules that we describe for the stable vpy mutant alleles are very similar to those described by the RNAi-based approach (Pumplin et al., 2010) and in the Petunia mutant pam1 (Sekhara et al., 2007; Feddermann et al., 2010). However, truncated intracellular structures were observed in vpy and pam1 mutants (Figure S1e; Feddermann et al., 2010). This suggests that the initiation of intracellular structures, while possibly impaired, was still possible. This differs from the results of Pumplin et al. (2010) who observed short hyphal projections that did not penetrate cortical cells. This could be due to differences in the arbuscular mycorrhizal species used (Gao et al., 2001). Pumplin et al. (2010) report the presence of rare arbuscule formation in one experiment using the Vapyrin RNAi construct. They suggest that this may have been the result of the presence of some non-transgenic roots. In our experiments no arbuscules were observed on any vpy mutants in any of the conditions used. A recently characterized mutant, rpg, displays abnormally thick and slowly progressing infection threads and abnormal root hair curling, but has normal AM interactions (Arrighi et al., 2008). The absence of an AM phenotype in the rpg mutants indicates that VPY and RPG may have separate roles in IT development. Similarly, a predicted E3 ubiquitin ligase (LIN/CERBERUS), two flotillins and a remorin, are required for IT progression but not for the AM symbiosis (Lombardo et al., 2006; Kiss et al., 2009; Yano et al., 2009; Haney and Long, 2010; Lefebvre et al., 2010). This apparent delay in formation and progression of rhizobial ITs is analogous to the AM phenotype of impeded entry, suggesting that vpy is required for a process that is common to both interactions.

MDPs associate with proteins that interact with membranes

While the role of the VPY protein remains to be determined, the presence of the MSP domain provides some clues to its function. The MSP of Caenorhabditis elegans, for which MSP is named, consists of an MSP domain alone and can polymerize to form fibril structures analogous to actin which mediate sperm motility (Roberts and Stewart, 2000). Major Sperm Protein is also found in bi-layered membrane vesicles that form in sperm cells which have no endoplasmic reticulum (ER) or Golgi and appears to drive vesicle budding (Kosinski et al., 2005). The MDPs are not well studied in plants. The MSP domain is found in at least 10 Arabidopsis proteins (Figure S4) all having different overall domain structures from VPY; four of these have been shown to be membrane associated (Oufattole et al., 2005; Marmagne et al., 2007; Petersen et al., 2009), including PVA12 which binds to ORP3a, an ER-localized sterol-binding protein (Saravanan et al., 2009), and VAP27-1 which binds to ACDH1, a putative glycolipid-binding protein (Petersen et al., 2009). Notably, VPY lacks the conserved lysines required for binding of phosphoinositides by the yeast MDP Scs2p (Loewen and Levine, 2005; Figure S4).

VPY might be involved in membrane trafficking processes

The subcellular localization of VPY is also unique. CYCLOPS/IPD3 and CCaMK are found exclusively in the nucleus where they interact with one another (Kalo et al., 2005; Messinese et al., 2007; Yano et al., 2008), while VPY is found both in the cytoplasm and the nucleus and is found in prominent mobile subcellular compartments (Feddermann et al., 2010; Pumplin et al., 2010). The significance of the nuclear localization of VPY is unclear. Pumplin et al. (2010) report seeing nuclear localization of VPY in cells invaded by arbuscules and hyphae, which leaves open the possibility that VPY could interact with other nuclear-localized elements of the common symbiosis pathway.

The localization of VPY-GFP to mobile puncta (Feddermann et al., 2010; Pumplin et al., 2010) in mycorrhizal roots suggests it may be associated with vesicle transport. Intracellular ITs and arbuscules require large amounts of membrane to grow. One possible role for VPY is in exocytosis-driven polar tip growth in response to external signals (i.e. NFs or Myc factors). It seems possible that in addition to the common signalling pathway, some genes involved in membrane-trafficking processes have also been recruited into the nodulation programme from the much older AM symbioses. Identification of proteins that bind to the MSP and/or ankyrin domains of VPY will provide further insight into the commonalities of these widespread and agriculturally important plant–microbe interactions.

Experimental procedures

Rhizobial staining

Plants infected with S. meliloti 1021 pXLD4 expressing HemA-lacZ were fixed in a solution of 1.25% gluteraldehyde and 0.1 m sodium phosphate (pH 7.2) and stained for β-galactosidase activity overnight in 0.1 m sodium phosphate (pH 7.2), 5 mm K3Fe(CN)6, 5 mm K4Fe(CN)6 and 0.02 m 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-GAL).

Mycorrhizal staining

M. truncatula plants were inoculated with G. intraradices (Glomus intraradices DAOM 181602) for microscopic analysis of FNB2 and FNB4 (vpy-1) or a mixture of G. intraradices and G. mossae for phenotyping of the segregating F2 population and analysed as previously described (Sekhara et al., 2007). For quantification of mycorrhizal structures, seeds of M. truncatula A17 and mutants FNB2 and FNB4 were scarified with sandpaper and pre-germinated on wet filter paper for 2 days. The seedlings were then transferred to 10-cm pots (four plants per pot) containing 50:50 sand and TerraGreen. A 5-month-old leek plant colonized with G. intraradices had previously been placed in the centre of the pot. The plants were grown under a 16-h daylight regime and samples taken at 2 and 4 weeks. The roots were washed in distilled water and 5 cm long sections cut 1 cm below the hypocotyl were taken for staining. The root sections were split longitudinally into two halves, half were ink stained and half stained with nitroblue tetrazolium (NBT). Ink staining was performed using a method based on that of Vierheilig et al. (1998). The roots were cleared in 10% KOH for 4 days at room temperature, washed in distilled water and stained for 10 min in 10% Pelikan black ink in 25% acetic acid. The roots were rinsed in distilled water and destained for 15 min in 25% acetic acid before a further rinsing in distilled water and mounting on slides. The NBT staining was carried out using the method as described by Schaffer and Peterson (1993). The clearing method used was 5% (w/v) KOH for 48 h at 55°C. The roots were cut into sections of approximately 1 cm and mounted on glass slides in distilled water. Counting was done as described by Mcgonigle et al. (1990) using a Nikon Optiphot microscope (

Generation of VPY constructs

For the complementation construct, a VPY-coding DNA sequence with appropriate primer modifications was cloned into pENTR as described in Karimi et al. (2002). Subsequently an LR reaction was performed to transfer the VPY sequence into the destination vector (pB7WGF2) by recombination. The plasmids were then transformed by electroporation into A. rhizogenes (Arqua 1).

Agrobacterium rhizogenes hairy root transformation

Seedlings were scarified for 10 min using concentrated H2SO4 followed by three washes with sterile water. Seeds were then sterilized for 10 min in 30% commercial bleach solution with 0.1% Tween, washed five times with sterile water and then plated on sterile wet filter paper on Petri dishes and wrapped with Parafilm. The plates were then covered completely with aluminium foil and placed upside down at 4°C for 3 days. The plates were then exposed to light and incubated at room temperature for 1 day until roots were elongated. Using sterile technique, approximately five seedlings were transferred to each square Petri dish which contained 0.9% agarose and minimal nutrient medium prepared as follows: CaCl2 (0.9 mm), MgSO4 (0.5 mm), KH2PO4 (20 μm), Na2HPO4 (10 μm), ferric citrate (20 μm), NH4NO3 (1.0 mm), MnCl2 (33 μg L−1), CuSO4 (33 μg L−1), ZnCl2 (7 μg L−1), H3BO3 (100 μg L−1), Na2MoO4 (33 μg L−1), 200 mg L−1 2-(N-morpholine)-ethanesulphonic acid (MES) and the pH was adjusted to 7.4. The bottoms of the plates were wrapped in aluminium foil to minimise exposure of the roots to light. Plates were then grown vertically in an incubator with a day/night regime (16 h light/8 h dark) at 20°C for about a week. The ARqua1 strain of A. rhizogenes was transformed with the pB7WG2D plasmid or pB7WG2D containing the 35S::VPY gene and then grown on plates containing 2 mg L−1 BASTA. Using an insulin needle (27.8–28 G) fresh ARqua1 was inoculated three to five times into the hypocotyls of the vpy-1 mutant and the WT controls just below the cotyledons, trying to not to penetrate the stele. Plants were then grown for 10–12 days in a vertical position in a 20°C growth chamber until the first hairy roots emerged from the hypocotyls. At this point the primary root was removed and the seedlings were transferred to minimal media plates as described above containing 2 mg L−1 BASTA and grown another 7–10 days vertically with roots covered in a 20°C growth chamber. Plants were then transferred to a Turface/vermiculite mixture (6:1) and inoculated with rhizobia (Sm1021) 1 week after transfer. Plants were watered with low-nitrogen fertilizer as previously described (Benedito et al., 2008). Confocal imaging was carried out with a Leica TCS SP2 AOBS laser confocal scanning microscope (Leica Microsystems, using standard settings for GFP.

RNA isolation from root hairs and qRT-PCR

For the root hair enrichment experiment 100 Jemalong (A17) plants for each treatment (control, ABS7 infected) were grown vertically for 2 days on Petri dishes as described above, and then inoculated with 200 μl of a suspension of S. meliloti (ABS7) in sterile distilled water having an absorbance of 0.02 (OD 600 nm). The plants were grown for an additional 4 days with the roots shielded from light. The roots were then harvested, the roots tips were removed and the remaining roots were placed in a 50 ml Falcon tube with liquid nitrogen and put through three cycles of vortexing to remove the root hairs, being careful to keep the roots frozen. The root hairs were then passed twice through a 150 μm wire filter. For the root hair fraction the RNA was prepared using RNeasy Micro kit (Qiagen, and the DNAse treatment was done on a column. For the stripped roots TRIzol reagent (Invitrogen, was used as per the manufacturer’s directions. For the expression of Vapyrin in WT and vpy-1 and vpy-2 roots seedlings were grown in Turface for 1 week and roots were harvested. The RNA was prepared using TRIzol as described above. All quality of RNA samples was checked using a Bioanalyzer (Agilent, For the root hair and stripped roots, 200 ng and 3 μg of RNA, respectively, was used for cDNA synthesis. All qRT-PCR was carried out using a 7900HT Fast Real-Time PCR System (Applied Biosystems, Gene-specific primers used for Vapyrin and the Medicago orthologue of Expansin7 were designed from the 3′ UTR sequence: Vapyrin: 5′-GGGTTTAAATTGCTGTTACAATCG-3′, 5′-AGCCAAAGACATCAGACAACC-3′ and MTEXP7A: 5′-GAAATGTGGCCTTCACCACT-3′, 5′-TGAACTCGTGGTGGGTTACA-3′. The quality of the cDNA was checked using a qRT-PCR based comparison of 5′ and 3′ UTR amplicons of MtUbiquitin 3′UTR: 5′-GGCCCTAGAACATTTCCTGTGG-3′, 5′-TTGGCAACCAAAATGTTCCC-3′ and 5′UTR: 5′-TTGGAGACGGATTCCATTGCT-3′, 5′-GCCAATTCCTTCCCTTCGAA-3′. For qRT-PCR analysis three technical replicates and four reference genes were used: MtUBC, 5′-CTGACAGCCCACTGAATTGTGA-3′, 5′-TTTTGGCATTGCTGCAAGC-3′; Mt-PDF2 5′-GTGTTTTGCTTCCGCCGTT-3′, 5′-CCAAATCTTGCTCCCTCATCTG-3′; MtUBC9 5′-GGTTGATTGCTCTTCTCTCCCC-3′, 5′-AAGTGATTGCTCGTCCAACCC-3′; MtHelicase 5′-GTACGAGGTCGGTGCTCTTGAA-3′, 5′-GCAACCGAAAATTGCACCATAC-3′. The PCR efficiencies were determined using the linregpcr program (Ramakers et al., 2003) and efficiency corrections between primer sets were made.

Sequence data

Gene sequences were deposited in GenBank under the following accession numbers: VAPYRIN: FJ795648; Abelmoschus esculentus: HM193247; Chamaecrista fasciculata: HM193248.

Additional experimental procedures

Experimental procedures for the identification of Tnt1 insertion mutants and the Affymetrix GeneChip hybridization are provided in the Supporting Information (Appendix S1).


Our work is supported by the National Science Foundation (NSF) and the Samuel Roberts Noble Foundation. The authors would like to thank Steven Cannon for the Chamaecrista fasciculata sequence, Jie Wang for assistance with plant hybridizations, Heather Cross, Andrew Potter and Xinbin Li for technical assistance, and Cuc Ly and Scott McNeil for help in preparing figures.