Y4lO of Rhizobium sp. Strain NGR234 Is a Symbiotic Determinant Required for Symbiosome Differentiation

Abstract

Type 3 (T3) effector proteins, secreted by nitrogen-fixing rhizobia with a bacterial T3 secretion system, affect the nodulation of certain host legumes. The open reading frame y4lO of Rhizobium sp. strain NGR234 encodes a protein with sequence similarities to T3 effectors from pathogenic bacteria (the YopJ effector family). Transcription studies showed that the promoter activity of y4lO depended on the transcriptional activator TtsI. Recombinant Y4lO protein expressed in Escherichia coli did not acetylate two representative mitogen-activated protein kinase kinases (human MKK6 and MKK1 from Medicago truncatula), indicating that YopJ-like proteins differ with respect to their substrate specificities. The y4lO gene was mutated in NGR234 (strain NGRΩy4lO) and in NGRΩnopL, a mutant that does not produce the T3 effector NopL (strain NGRΩnopLΩy4lO). When used as inoculants, the symbiotic properties of the mutants differed. Tephrosia vogelii, Phaseolus vulgaris cv. Yudou No. 1, and Vigna unguiculata cv. Sui Qing Dou Jiao formed pink effective nodules with NGR234 and NGRΩnopLΩy4lO. Nodules induced by NGRΩy4lO were first pink but rapidly turned greenish (ineffective nodules), indicating premature senescence. An ultrastructural analysis of the nodules induced by NGRΩy4lO revealed abnormal formation of enlarged infection droplets in ineffective nodules, whereas symbiosomes harboring a single bacteroid were frequently observed in effective nodules induced by NGR234 or NGRΩnopLΩy4lO. It is concluded that Y4lO is a symbiotic determinant involved in the differentiation of symbiosomes. Y4lO mitigated senescence-inducing effects caused by the T3 effector NopL, suggesting synergistic effects for Y4lO and NopL in nitrogen-fixing nodules.

Legumes establish nodule symbioses with nitrogen-fixing bacteria (called rhizobia). The formation of effective nodules is a complex developmental process which relies on various rhizobial signals and determinants from both symbiotic partners. During the infection process, flavonoids exuded by host plants act as primary signals to induce rhizobial genes required for symbiosis. The activation of these genes depends on NodD transcriptional regulators, which bind to conserved DNA sequences, so-called nod boxes (26, 41). Flavonoids stimulate the synthesis of rhizobial nodulation factors (Nod factors), which induce plant responses, such as root hair deformation, the expression of early nodulation genes, and the induction of cortical cell divisions. Rhizobia vary in their ability to enter into symbiosis with host species and certain strains induce ineffective associations, which are detrimental for both partners. Symbiotically relevant molecules have been found to be structurally different from strain to strain. In addition to Nod factors, the host-specific determinants of rhizobia include exopolysaccharides (or oligosaccharides derived from them), lipopolysaccharides, cyclic β-glucans, K antigens, and type 3 (T3) effector proteins secreted by the bacterial T3 secretion system (T3SS) (14, 29, 41).

Pathogenic bacteria use T3SSs to deliver T3 effector proteins into host cells where they play a key role in pathogenic attack, e.g., to overcome defense reactions. On the other hand, many plants developed resistance mechanisms that target T3 effectors within the host cell. The direct or indirect recognition of T3 effectors often results in the induction of resistance responses (e.g., a rapid hypersensitive reaction) that block bacterial multiplication, bacterial spread, and the development of disease symptoms (10, 22, 34). Thus, T3 effectors are “two-edged swords” that either act as virulence or as avirulence factors in plant-pathogen interactions. Similarly, Nops (nodulation outer proteins—extracellular components of rhizobial T3SSs and proteins secreted by T3SSs) may play either a positive or negative role in nodule formation (29). Although not proven, it is hypothesized that rhizobial T3 effectors are delivered into legume host cells. Rhizobial T3 effectors are able to induce striking alterations in plant cells. NopL, a T3 effector of Rhizobium sp. NGR234, interfered with plant defense reactions. When expressed in tobacco or Lotus japonicus plants, NopL suppressed the expression of pathogen-related defense genes (6). NopT, another T3 effector of NGR234, elicited a hypersensitive reaction when expressed in tobacco cells (16).

Rhizobium sp. strain NGR234 possesses a T3SS, which plays a role in the nodulation of certain host plants, such as Tephrosia vogelii, Phaseolus vulgaris, Pachyrhizus tuberosus, Crotalaria juncea, and Flemingia congesta. In NGR234, the expression of genes encoding the T3SS depends on host flavonoids, NodD1, and the transcriptional activator TtsI, which has a nod box in the promoter region (26, 30, 55). So far, only three bona fide T3 effectors of NGR234 have been characterized, namely NopL (5, 6, 31), NopP (4, 46), and NopT (16). The open reading frame y4lO on the symbiotic plasmid pNGR234a is predicted to encode another putative T3 effector of NGR234. The Y4lO sequence displays similarities to proteins belonging to the YopJ effector family (13, 19). Representatives of this T3 effector family have been identified and characterized in strains of Yersinia sp. (YopJ/YopP [20, 32]), Salmonella enterica serovar Typhimurium (AvrA [24]), Vibrio parahaemolyticus (VopA [51]), Xanthomonas campestris pv. vesicatoria (AvrRxv [13, 56]; AvrBsT [40]; AvrXv4 [3]; and XopJ [37]), Erwinia amylovora (ORFB [38]), Pseudomonas syringae (ORF5 in B728a [1] and AvrPpiG [2]), and Ralstonia solanacearum (PopP1 [28] and PopP2 [17]). These T3 effectors seem to target specific signal transduction pathways of the eukaryotic host. YopJ in mammalian host cells inhibits mitogen-activated protein (MAP) kinase and nuclear factor κB (NF-κB) signaling pathways (35) as well as the TLR3-mediated interferon response (48). In plant-pathogen interactions, resistance mechanisms (hypersensitive response) have been reported for certain nonhosts that seem to recognize a specific YopJ-like effector as avirulence protein (3, 8, 40, 45, 49, 56, 57). In Arabidopsis thaliana, ecotype Nd-1 plants carrying the RRS1-R resistance gene are able to recognize PopP2 (17), and the AvrBsT resistance of ecotype Pi-0 is caused by a mutation in a carboxylesterase, which is predicted to hydrolyze lysophospholipids and acylated proteins (15).

For many years, YopJ-like T3 effectors were thought to function as cysteine proteases with conserved active site residues (H, D/E, Q, C). Transient expression studies in host cells provided certain indications that YopJ and AvrA are deubiquitinating enzymes (40, 48, 59, 60), whereas SUMO (small ubiquitin-like modifier) protease activity has been proposed for AvrXv4 (45). Furthermore, in vitro experiments suggested that YopJ cleaved the artificial substrate ubiquitin-7-amino-4-methylcoumarin (60) and that AvrA exhibited deubiquitinase activity on ubiquitinated IκBα (59). On the other hand, recent biochemical analysis revealed clear evidence that YopJ and VopA are bacterial acetyltransferases. YopJ uses acetyl-coenzyme A (acetyl-CoA) to acetylate human MAP kinase kinase 6 (MKK6) and the NF-κB pathway kinase IKKβ. Acetylation prevented the phosphorylation of these kinases and thus seems to be a refined strategy to inactivate protein kinases in the host cell (23, 33, 35, 36, 51).

In this study, we investigated the symbiotic role of y4lO from Rhizobium sp. strain NGR234. Inoculation experiments with a mutant strain and ultrastructural analysis of infected nodule cells indicated that y4lO is a symbiotic determinant involved in the differentiation of symbiosomes.

MATERIALS AND METHODS

Bioinformatic analysis.

The promoter region of y4lO was analyzed with the NNPP (Neural Network Promoter Prediction) version 2.2 program (http://www.fruitfly.org/seq_tools/promoter.html). GeneMark version 2.5 (http://opal.biology.gatech.edu/GeneMark/) was used for the identification of alternative start codons. The molecular weights of predicted proteins were calculated with the Compute pI/M_w tool (http://ca.expasy.org/tools/pi_tool.html). Sequence comparisons with databases were performed with the BLAST program (http://ncbi.nlm.nih.gov/BLAST/). Y4lO homologues (amino acid sequences) were aligned with the Clustal W algorithm. The unrooted radial tree (Fig. 1B) was constructed with the MEGA3.1 program using the neighbor-joining method (27). The amino acid sequence of the short ORF72 upstream of y4lO (Fig. 1C) was analyzed with the MYR prediction server (http://mendel.imp.ac.at/myristate/SUPLpredictor.htm).

FIG. 1.

Analysis of the y4lO sequence of Rhizobium sp. strain NGR234. (A) Genetic map of the y4lO gene and flanking sequences in pNGR234a. The promoter sequence contains a putative tts box. The gene was mutated by the insertion of an Ω interposon at the indicated position (resulting in strain NGRΩy4lO). (B) Unrooted phylogenic tree of predicted Y4lO protein and related T3 effectors (accession numbers: YopJ, ABX88765; YopP, AAK69256, AvrA, AAL21745; AvrRxv, CAJ22102; AvrBsT, AAD39255; XopJ, CAJ23833; AvrXv4, AAG39033; PopP1, CAD14528; PopP2, CAD14570; and VopA, AAT08443). The horizontal bar represents a distance of 0.3 substitution per site. (C) Amino acid sequence alignment of the short ORF upstream of y4lO with the N-terminal sequence of XopJ from X. campestris pv. vesicatoria (CAJ23833). Identical residues are marked with asterisks. Similar residues are marked with single or double dots.

Transcriptional and translational fusions.

A 870-bp fragment containing the putative promoter region of y4lO (accession number U00090) was inserted upstream of the promoterless gusA gene of vector pRG960 (52), yielding plasmid pRG-y4lOp (transcriptional fusion; Fig. 2A). For the construction of the pRG960 derivative pRG-891, an 891-bp fragment containing the promoter region of y4lO and 138 bp of the coding region was fused to gusA without the ATG start codon (translational fusion; Fig. 2A). Plasmid pRG-893 with an additional insertion of two nucleotides at position 72 of the y4lO coding region, was generated in a similar way (Fig. 2A).

FIG. 2.

(A) Constructs of transcriptional and translational fusions used in this study. P, PstI; B, BamHI. The black area in pRG-y4lOp is the translation initiation site of gusA. Cultures of NGR234 carrying the indicated plasmids were treated with 1 μM apigenin, and GUS activity was fluorometrically determined 24 h later. (B) The transcriptional activation of y4lO depends on ttsI. Cultures of NGR234 pRG-y4lOp and NGRΩttsI pRG-y4lOp were induced with 1 μM apigenin or left untreated. GUS activity was measured after 24 h. Data represent means (± standard errors) of the results from three independent cultures.

The plasmids were verified by sequencing and then mobilized into NGR234 (50) or NGRΩttsI (53) using a triparental mating procedure with the helper plasmid pRK2013 (18). Where indicated, rhizobia grown in liquid TY (tryptone-yeast extract) medium (7) were induced with 1 μM apigenin and incubated for 24 h. Rhizobial cells (equal amounts adjusted to an optical density at 600 nm of 0.25) were then transferred to 2-ml test tubes. After centrifugation (13,000 rpm for 3 min), the pellets were resuspended in 1 ml extraction buffer (50 mM phosphate buffer [pH 7.0], containing 0.1% Triton X-100, 0.1% sarcosyl, 10 mM EDTA, 10 mM 2-mercaptoethanol) and kept on ice for 10 min. β-Glucuronidase (GUS) activity was fluorometrically measured at 37°C with 2 mM 4-methylumbelliferyl-β-d-glucuronide as the substrate.

Expression of proteins in Escherichia coli and Western blots.

The DNA of y4lO (Fig. 1A) from Rhizobium sp. strain NGR234 (accession number U00090) was cloned into pET28b using NdeI and XhoI (yielding plasmid pET-y4lO; Y4lO with an N-terminal six-His tag). A similar plasmid (pET-y4lOΔN) was generated with y4lO lacking the first 117 N-terminal nucleotides. Furthermore, the DNA of y4lO was cloned into the EcoRI and XhoI sites of pGEX-4T-1, an expression vector for glutathione S transferase (GST) fusion proteins (generating plasmid pGEX-y4lO).

A DNA sequence encoding a MAP kinase kinase of Medicago truncatula, named in this study MtMKK1 (accession number AC144503; position 62783 to 61683), was amplified by a PCR-based technique from genomic DNA from M. truncatula cv. Jemalong A17 with the primers 5′-TTGAAATCATATGAGGCCGATTCAGCTTCC-3′ and 5′-TAACTCGAGCAACAAAAATTACATTGACGAAC-3′. The amplification product digested with NdeI and XhoI was then cloned into the pET28b vector, generating plasmid pET-MtMKK1.

Cells of E. coli BL21(DE3) with plasmids verified by sequencing were grown in Luria-Bertani (LB) medium containing 50 μg ml⁻¹ kanamycin at 37°C until cells reached an optical density at 600 nm of ∼0.6. Protein expression was then induced with 0.2 mM isopropyl-β-d-thiogalactopyranoside (IPTG), and the cells were cultured for 12 h at 25°C. The cells were collected by centrifugation and lysed with 2× sodium dodecyl sulfate (SDS) loading buffer, and the proteins were separated on SDS-polyacrylamide gels. The proteins were visualized by Coomassie brilliant blue R-250 staining. For further characterization of the His-tagged Y4lO, the cells were lysed by sonication and the proteins were purified by nickel affinity chromatography under denaturing conditions according to the manufacturer's instructions (Qiagen).

Purified His-tagged Y4lO protein [from BL21(DE3) pET-y4lOΔN] was used to immunize a New Zealand rabbit. For the Western blots, the proteins were separated on SDS-polyacrylamide gels and then transferred onto nitrocellulose membranes by electroblotting. To visualize the Y4lO protein, the membranes were incubated with the antiserum raised against Y4lO (1:4,000 dilution) and then with a horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G antiserum (Boster, Wuhan, China). The blots were developed either with 3,3′-diamino-benzidine (Boster) or with ECL detection reagents (Amersham Biosciences/GE Healthcare). Secreted proteins from apigenin-induced rhizobial cultures were concentrated according to a published procedure (31) and finally desalted by dialysis.

Acetylation assay.

For the acetylation assay, His-tagged Y4lO, His-tagged human MKK6 [from E. coli BL21(DE3) carrying pET28a-His-MKK6 (36)], and His-tagged MtMKK1 were purified by nickel affinity chromatography under native conditions according to the manufacturer's instructions (Qiagen). Y4lO without a six-His tag was obtained from the digestion of His-tagged Y4lO with thrombin. GST beads from Sigma were used for the purification of GST-Y4lO and GST-YopJ [from E. coli BL21(DE3) carrying pGEX-TEV-YopJ (36)]. The eluted proteins were concentrated with Microcon centrifugal filter devices (Amicon; Millipore). Protein quantification was performed on polyacrylamide gels with Coomassie brilliant blue R-250 and bovine serum albumin as a standard. For the acetylation assays (36), purified enzymes (0.5 to 1 μg) and MAP kinase kinases (4 to 6 μg) were incubated at 30°C for 1 h in the following reaction mixture (20 μl): 50 mM HEPES buffer (pH 7.4), 10% glycerol, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 2 μl [¹⁴C]-acetyl-CoA (56 μCi μM⁻¹) from Sigma. The proteins were then separated on SDS-polyacrylamide gels and fixed with 50% methanol and 10% glacial acetic acid for 1 h. The gels were dried and radioactivity was visualized by a Typhoon imaging system (Typhoon 8600 scanner; Amersham Biosciences/GE Healthcare).

Construction of mutant strains.

The strains and plasmids used for the construction of NGRΩy4lO and NGRΩnopLΩy4lO are listed in Table 1. A 1.5-kb fragment containing y4lO (accession number U00090) was cloned into pBluescript II SK(+) yielding plasmid pSK-y4lO. A point mutation construct with an EcoRI site (at position 136 of y4lO [Fig. 1A]) was then generated by PCR-based site-directed mutagenesis using pSK-y4lO as a template and DpnI for digestion of the amplification products. A spectinomycin resistance Ω interposon was then excised from pHP45Ω (42) with EcoRI and ligated into the EcoRI site of the mutated pSK-y4lO. Finally, the excised 3.7-kb SacI-XhoI fragment was cloned into the suicide vector pJQ200SK (44). The construct was mobilized from E. coli DH5α into NGR234 and NGRΩnopL (31) by a triparental mating procedure with the helper plasmid pRK2013 (18). The replacement of the mutated gene was forced by selecting for the resistance of the interposon marker (Sp^r) and for resistance to sucrose (5% wt/vol). Double-crossover events at homologous sites were confirmed by Southern blot analysis with rhizobial DNA (12) using the DIG DNA labeling and detection kit as specified by the supplier (Roche).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Characteristicsa	Reference or source
Strains
Rhizobium sp.
NGR234	Rhizobium sp. strain NGR234 isolated from Lablab purpureus (Rifr derivative)	50
NGRΩy4lO	NGR234 carrying an Ω insertion in y4lO (Rifr, Spr) (Fig. 1A)	This work
NGRΩnopL	NGR234 carrying an Ω insertion in nopL (Rifr, Kmr)	31
NGRΩnopLΩy4lO	NGRΩnopL carrying an Ω insertion in y4lO (Rifr, Spr) (Fig. 1A)	This work
NGRΩrhcN	NGR234 carrying an Ω insertion in rhcN (Rifr, Spr)	53
NGRΩttsI	NGR234 carrying an Ω insertion in ttsI, formerly named NGRΩy4xI (Rifr, Spr)	53
E. coli
DH5α	supE44 ΔlacU169 (φ80 lacZΔM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1	Promega Corp.
BL21(DE3)	F−ompT hsdSB (rB− mB−) gal dcm (DE3)	Novagen (Merck Chemicals Ltd.)
Plasmids
pBluescript II SK(+)	High-copy-no. ColE1-based phagemid (Apr)	Stratagene (Agilent Technologies)
pSK-y4lO	1.5-kb fragment containing y4lO cloned into pBluescript II SK(+) (Apr)	This work
pRG960	Broad-host-range vector containing a promoterless gusA gene with start codon (Spr) (Fig. 2A)	52
pRG-y4lOp	870-bp fragment containing the y4lO promoter cloned into pRG960 (Spr) (Fig. 2A)	This work
pRG-891	pRG960 derivative with an 891-bp fragment containing the y4lO promoter and 138 bp of the coding region fused to gusA lacking the start codon (Spr) (Fig. 2A)	This work
pRG-893	pRG-891 derivative with an insertion of two nucleotides (GC) at position 72 of the y4lO coding region (Spr) (Fig. 2A)	This work
pET28b	Expression vector for His-tagged proteins (Kmr)	Novagen (Merck Chemicals Ltd.)
pET-y4lO	y4lO cloned into pET28b (N-terminal His-tagged Y4lO) (Kmr)	This work
pET-y4lOΔN	y4lO lacking the first 117 N-terminal nucleotides cloned into pET28b (Kmr)	This work
pET28a-His-MKK6	Human MKK6 (MAPKK6) cloned into pET28a (Kmr)	36
pET-MtMKK1	MtMKK1 from M. truncatula (AC144503; position 62783-61683) cloned into pET28b (Kmr)	This work
pGEX-4T-1	Expression vector for GST fusion proteins (Apr)	Amersham Biosciences/GE Healthcare
pGEX-y4lO	y4lO cloned into pGEX-4T-1 (Apr)	This work
pGEX-TEV-YopJ	YopJ cloned into pGEX-TEV (Apr)	36
pJQ200SK	Suicide vector used in directed mutagenesis (Ger)	44
pRK2013	Tra+ helper plasmid for mobilization (Kmr)	18
pHP45Ω	Vector containing an Ω interposon (Apr, Spr)	42
pFAJ1815	Mini-Tn5 transposon derivative containing gfp-gusAp (constitutively expressed gusA) (Kmr)	58

^aAp^r, resistance against ampicillin; Ge^r, resistance against gentamicin; Km^r, resistance against kanamycin; Rif^r, resistance against rifampin; Sp^r, resistance against spectinomycin.

For the construction of NGRΩy4lO constitutively expressing gusA, the mini-Tn5 transposon derivative pFAJ1815 containing gfp-gusAp (58) was introduced into NGRΩy4lO using a biparental mating procedure. The selection of tagged colonies constitutively expressing gusA was performed on agar plates supplemented with 0.5 mg ml⁻¹ 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid. An inoculation experiment with P. vulgaris cv. Yudou No. 1 indicated that the mTn5gfp-pgusA insertion had no effect on the symbiotic phenotype. For the visualization of NGRΩy4lO expressing gusA, nodules were stained with 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid as described previously (47).

Nodulation tests.

Nodulation tests with Crotalaria juncea, Pachyrhizus tuberosus, Tephrosia vogelii, Phaseolus vulgaris cv. Yudou No. 1, and Vigna unguiculata cv. Sui Qing Dou Jiao were performed according to previously described procedures (47). Briefly, seeds were surface-sterilized and left to germinate on agar plates. The plantlets (1 plant per jar) were then transferred to sterilized 300-ml plastic jar units linked with a cotton wick (3:1 [vol/vol] mixture of vermiculite and expanded clay in the upper vessel; nitrogen-free nutrient solution in the lower vessel). The plants were inoculated with ∼10⁹ bacteria (grown in TY medium, centrifuged, and resuspended in 10 mM MgSO₄). The plants were grown in a plant growth room or air-conditioned greenhouse at 26 ± 2°C. If not otherwise indicated, P. vulgaris cv. Yudou No. 1 plants were harvested 4 weeks postinoculation. The other plants were harvested 7 to 8 weeks postinoculation. Biomass (dry weight) data were obtained from lyophilized plant material. The nitrogen contents of lyophilized and pulverized total plant material were determined with a CHNS analyzer (Elementar Analysensysteme, Hanau, Germany).

Electron and light microscopy.

Nodules were excised from roots, sliced manually, and immediately fixed in 67 mM sodium phosphate buffer (pH 7.4) containing 2.5% (vol/vol) glutaraldehyde and 2% (vol/vol) paraformaldehyde (4°C, overnight). After three washing steps, the samples were postfixed for 90 min in 67 mM phosphate buffer (pH 7.4) containing 1% (wt/vol) OsO₄. The samples were then rinsed in the same buffer (three times for 15 min) and dehydrated with increasing volumes of ethanol (30%, 50%, 70%, 90%, and 95% for 15 min each and 100% three times for 30 min each). The samples were then embedded in Poly/Bed 812 resin as specified by the supplier (Polyscience, Warrington, PA). An ultramicrotome equipped with a diamond knife (Leica UC6; Austria) was used to obtain ultrathin sections (∼85 nm thick). The sections were stained with uranyl acetate and lead citrate and finally observed in a transmission electron microscope. For the light microscopy observations, semithin sections (1 μm thick) from the same samples were stained with 0.6% (wt/vol) toluidine blue.

RESULTS

Sequence analysis of y4lO.

The ORF y4lO in the plasmid pNGR234a of Rhizobium sp. strain NGR234 (19) is flanked by several ORFs with sequence similarities to insertion sequence homologs, suggesting transposon-related insertion events (Fig. 1A). A putative promoter is predicted by the NNPP version 2.2 program with the transcription start sequence CTTGCATATG. The promoter region of y4lO contains a tts box, a putative binding site for the transcriptional activator TtsI (30, 55) (Fig. 1A). Comparisons with sequence databases showed similarities with YopJ family T3 effectors from bacterial pathogens. A phylogenetic analysis indicated that the Y4lO sequence of NGR234 is most closely related to XopJ (27) (Fig. 1B). Other rhizobial strains with T3SSs, e.g., Bradyrhizobium japonicum USDA110, lack genes homologous to y4lO. Similarly to related T3 effectors, four conserved amino acids, predicted to play a role in enzymatic activity, are present in the Y4lO sequence (residues H123, E143, Q179, and C185). The annotated Y4lO protein (29.1 kDa) is encoded by a DNA sequence with the unusual translation start codon TTG. The program GeneMark Version 2.5 (e.g., with the model organism Sinorhizobium meliloti plasmid pSymA as a reference) predicts a protein (24.7 kDa) with the N-terminal sequence MSSSL. The annotated Y4lO protein (29.1 kDa) is shorter than the coding sequences of related T3 effectors, such as XopJ. Sequence alignment revealed a short ORF of 72 bp (ORF72) upstream of the ORF y4lO (Fig. 1A); the corresponding amino acid sequence shows similarities with the N-terminal amino acid residues of XopJ (Fig. 1C).

Transcriptional activation of y4lO.

To study the transcriptional activation of y4lO, a 870-bp fragment containing the putative promoter region of y4lO was inserted upstream of a promoterless gusA gene of vector pRG960. The resulting transcriptional fusion (pRG-y4lOp [Fig. 2A]) was then mobilized into NGR234 and NGRΩttsI, a strain with a mutated ttsI gene (53). As transcription of ttsI is dependent on NodD1 and flavonoids (26), rhizobial cultures were treated with apigenin and GUS activity was measured 24 h later. As shown in Fig. 2B, GUS activity in NGR234 pRG-y4lOp was considerably higher than in NGRΩttsI pRG-y4lOp, indicating that the transcriptional activation of y4lO depended on ttsI. As expected, the GUS activity of the NGR234 derivative was stimulated by the treatment with apigenin, whereas no elevated values were obtained with the NGRΩttsI derivative. These results together with those from a recent study (55) confirm the bioinformatic prediction that y4lO with a tts box in the promoter region is regulated by the transcriptional activator TtsI.

Two translational fusions were constructed to examine whether the TTG start codon or the ATG codon (at position 118 of y4lO) is the translation initiation site of Y4lO. One construct (pRG-891 [Fig. 2A]) was generated by the in-frame fusion of the first 138 nucleotides of y4lO with gusA lacking the ATG start codon. The other construct with a frameshift mutation was obtained by the insertion of two additional nucleotides at position 72 of y4lO (pRG-893 [Fig. 2A]). NGR234 derivatives containing these plasmids were grown in the presence of apigenin, and GUS activity was determined 24 h later. As shown in Fig. 2A, GUS activity for NGR234 pRG-891 was high (comparable to pRG-y4lOp), whereas only background activity was measured with NGR234 pRG-893. These findings indicate that the frameshift mutation in pRG-893 abolished the formation of a functional GUS protein and that the TTG start codon is likely the translation initiation site of Y4lO.

Expression of y4lO in E. coli.

To characterize the Y4lO protein, the y4lO sequence was cloned into vector pET28b. E. coli BL21(DE3) containing this plasmid synthesized recombinant Y4lO protein with an N-terminal six-His tag. Under denaturing conditions, the purification of the protein by nickel affinity chromatography resulted in a single band (apparent molecular weight ≈ 31 kDa) (Fig. 3A). An antiserum raised against the purified His-Y4lO protein [from BL21(DE3) pET-y4lOΔN] specifically recognized low amounts of Y4lO protein on Western blots (detection of ≥10 ng protein) (Fig. 3C). Similarly, Y4lO expressed as a GST fusion protein (GST-Y4lO) (apparent molecular weight ≈ 54.7 kDa) was obtained from E. coli cells carrying plasmid pGEX-y4lO (data not shown).

FIG. 3.

Expression of Y4lO in E. coli BL21(DE3) and acetylation assay with MAP kinase kinases. (A) Proteins were separated on a SDS-polyacrylamide gel and visualized by staining with Coomassie brilliant blue R-250. Lane 1, BL21(DE3) pET-y4lO; lane 2, BL21(DE3) carrying the empty vector pET28b; lane 3, purified His-tagged Y4lO protein. (B) Acetylation assay with [¹⁴C]-acetyl-CoA and the indicated purified proteins. Reaction mixtures were separated on a SDS-polyacrylamide gel, and radioactivity was visualized with a Typhoon imaging system. (C) Western blot analysis with a rabbit serum raised against recombinant Y4lO. Lane 1, proteins from E. coli with the empty vector pET28b; lane 2, purified His-tagged Y4lO from BL21(DE3) pET-y4lO; lane 3, purified His-tagged Y4lO from BL21(DE3) pET-y4lOΔN.

His-tagged Y4lO, Y4lO (six-His tag removed by thrombin) and GST-Y4lO proteins purified as native proteins (or affinity beads with immobilized enzyme) were then used for in vitro acetylation assays with [¹⁴C]-acetyl-CoA and two representative His-tagged MAP kinase kinases: human MKK6 (36) and MtMKK1 from the legume Medicago truncatula, which is homologous to SIMKK of Medicago sativa (25). For comparison, the YopJ protein expressed as a GST fusion protein (GST-YopJ) was also purified (36). As shown in Fig. 3B, both MAP kinase kinases were not acetylated by Y4lO, whereas GST-YopJ acetylated MKK6 but not MtMKK1.

Is Y4lO a secreted protein?

The T3 effectors of NGR234 from bacterial culture supernatants (i.e., NopL, NopP, and NopT) could be detected on immune blots with specific antisera. Secreted proteins from mutant derivatives lacking a functional T3SS (e.g., strain NGRΩrhcN) served as controls to demonstrate T3SS-dependent secretion (4, 16, 31). When similar experiments were performed with secreted proteins from rhizobial cultures and the anti-Y4lO antibodies, Y4lO was not detected on immune blots. In contrast to Y4lO, NopL and NopT secreted by NGR234 were recognized by anti-NopL and anti-NopT antibodies, indicating that the sample contained T3SS-related nodulation outer proteins. When proteins from the total bacteria of NGR234 (pellets) or proteins from nodules were probed with the anti-Y4lO antibodies, Y4lO was also not detected (not shown), suggesting that Y4lO is produced at low levels.

Y4lO is a determinant of symbiosis.

To explore the effects of Y4lO during symbiosis with legumes, inoculation tests were performed with NGR234 and constructed mutant derivatives, namely NGRΩy4lO and the double mutant NGRΩnopLΩy4lO (mutation of y4lO in NGRΩnopL, which is mutated in the T3 effector gene nopL [31]). For comparison, the T3SS null mutant NGRΩrhcN with a nonfunctional T3SS (53) was included into the nodulation analysis. C. juncea and P. tuberosus were chosen for inoculation tests, as the effective nodulation of these legumes was blocked by nondefined T3SS proteins (31, 53). The symbiotic phenotypes of NGRΩy4lO and NGRΩnopLΩy4lO on these plants were not significantly different compared to those of NGR234, however (Table 2).

TABLE 2.

Nodulation of legumes with NGR234 and the indicated mutant strains

Planta	Straina	No. of nodules/plantb	Plant biomass (mg dry weight)b	Phenotype at harvest time
C. juncea	NGR234	79 ± 11	548 ± 200	Fix+/−
	NGRΩy4lO	14 ± 2	161 ± 22	Fix−
	NGRΩnopLΩy4lO	20 ± 3	228 ± 44	Fix−
	NGRΩrhcN	110 ± 6	1,505 ± 220	Fix+
P. tuberosus	NGR234	0.4 ± 0.2	291 ± 31	Fix−
	NGRΩy4lO	0.2 ± 0.2	286 ± 35	Fix−
	NGRΩnopLΩy4lO	0.2 ± 0.2	341 ± 45	Fix−
	NGRΩrhcN	31 ± 2	417 ± 92	Fix+
T. vogelii	NGR234	33 ± 4	631 ± 51	Fix+
	NGRΩy4lO	22 ± 2	224 ± 8	Fix−
	NGRΩnopLΩy4lO	33 ± 2	547 ± 54	Fix+
	NGRΩrhcN	18 ± 2	512 ± 45	Fix+
P. vulgaris cv.Yudou No. 1	NGR234	25 ± 4	894 ± 92	Fix+
	NGRΩy4lO	28 ± 2	771 ± 50	Fix−
	NGRΩnopLΩy4lO	30 ± 4	1,132 ± 94	Fix+
	NGRΩrhcN	31 ± 3	962 ± 78	Fix+
V. unguiculata cv. Sui Qing Dou Jiao	NGR234	71 ± 12	2,052 ± 485	Fix+
	NGRΩy4lO	134 ± 15	555 ± 72	Fix−
	NGRΩnopLΩy4lO	69 ± 8	1,443 ± 153	Fix+
	NGRΩrhcN	75 ± 12	1,368 ± 148	Fix+

^aThe plants (one plant per jar) were inoculated with the indicated strains. P. vulgaris cv. Yudou No. 1 plants were harvested 4 weeks postinoculation. The other plants were harvested 7 to 8 weeks postinoculation.

^bNumber of nodules and total plant biomass (dry weight) was determined for each plant at the time of harvest. Data indicate means ± standard errors (n ≥ 6).

The nodulation of T. vogelii, another host plant of NGR234, is positively affected by the T3SS (53). A recent mutant analysis showed that the mutation of nopL did not affect symbiosis with this host plant. A double mutant with deleted nopL and nopP genes induced fewer nodules than the parent strain, however (31, 46). Interestingly, T. vogelii challenged with NGRΩy4lO had only ineffective nodules (Fix⁻ phenotype) at the time of harvest. The plants inoculated with NGR234, NGRΩnopLΩy4lO, or NGRΩrhcN formed effective nodules that promoted plant growth (Table 2; Fig. 4A). Ineffective nodules of T. vogelii induced by NGRΩy4lO were greenish in the infected zone, whereas effective nodules induced by the other strains were pink (Fig. 4B to E). The nitrogen contents of T. vogelii plants infected with NGRΩy4lO were similar to those of noninfected control plants. Plants nodulated by the other strains exhibited considerably higher nitrogen contents (Fig. 4L).

FIG. 4.

Symbiotic phenotype of the mutants NGRΩy4lO and NGRΩnopLΩy4lO. (A) Aerial part of nodulated T. vogelii plants 8 weeks postinoculation. 1, NGRΩy4lO; 2, NGRΩrhcN; 3, NGRΩnopLΩy4lO; 4, NGR234. (B) Corresponding ineffective nodule of T. vogelii induced by NGRΩy4lO. (C) Effective nodule of T. vogelii induced by NGRΩrhcN. (D) Effective nodule of T. vogelii induced by NGRΩnopLΩy4lO. (E) Effective nodule of T. vogelii induced by NGR234. (F) Aerial part of nodulated P. vulgaris cv. Yudou No. 1 plants 4 weeks postinoculation. 1, NGR234; 2, NGRΩy4lO; 3, NGRΩrhcN; 4, NGRΩnopLΩy4lO. (G) Corresponding nodule of P. vulgaris cv. Yudou No. 1 induced by NGR234. (H) Ineffective nodule of P. vulgaris cv. Yudou No. 1 induced by NGRΩy4lO. (I) Effective nodule of P. vulgaris cv. Yudou No. 1 induced by NGRΩrhcN. (J) Effective nodule of P. vulgaris cv. Yudou No. 1 induced by NGRΩnopLΩy4lO. (K) Nodules of P. vulgaris cv. Yudou No. 1 after coinoculation with NGR234 and NGRΩy4lO constitutively expressing gusA (ratio, 1:1). Nodules were stained with 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid. (L) Nitrogen contents of harvested T. vogelii plants infected with the indicated strains and noninfected control plants (black column). Data indicate means ± standard errors (n = 3). (M) Nitrogen contents of harvested P. vulgaris cv. Yudou No. 1 plants infected with the indicated strains and noninfected control plants (black column). Data indicate means ± standard errors (n = 3). DW, dry weight. Scale bars = 1 mm.

Further nodulation assays were performed with P. vulgaris cv. Yudou No. 1 (Table 2; Fig. 4F). In this host plant of NGR234, nodule formation was modulated by the T3 effector NopT (16). Plants inoculated with NGR234 formed pink effective nodules (Fig. 4G), resulting in well-developed plants. When inoculated with NGRΩy4lO, the nodules were pink at first but turned greenish later (Fig. 4H). At the time of harvest, the leaves were yellow, plant growth was reduced, and nitrogen contents were low (Fig. 4 M). NGRΩrhcN and the double mutant NGRΩnopLΩy4lO induced effective nodules (Fig. 4I and J) that promoted plant growth (Fig. 4F) and increased nitrogen contents in harvested plants (Fig. 4M). These data indicate that the symbiotic phenotypes of the tested strains on P. vulgaris cv. Yudou No. 1 were similar to those on T. vogelii.

Finally, nodulation studies were performed with the host plant Vigna unguiculata cv. Sui Qing Dou Jiao. Previous inoculation tests with NGRΩrhcN suggested that the nodulation of V. unguiculata cv. Red Caloona is independent of a functional T3SS system (53). NGRΩy4lO and NGRΩnopLΩy4lO induced pink nodules on V. unguiculata at a frequency similar to that of the parent strain NGR234 when the plants were harvested 4 weeks postinoculation. Differences between NGRΩy4lO and NGR234 were seen later, however. The leaves of plants inoculated with NGRΩy4lO turned yellow and nodules greenish. At 7 weeks postinoculation, the biomass accumulation was low and numerous small ineffective nodules were observed (Table 2).

Mutation in y4lO does not affect early symbiotic stages.

The symbiotic capacity of NGRΩy4lO to induce nodules on P. vulgaris cv. Yudou No. 1 roots was further analyzed in coinoculation experiments with NGR234. To visualize NGRΩy4lO bacteria in nodules, a derivative of NGRΩy4lO expressing gusA was constructed. All nodules induced by this strain turned blue when stained with 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid. The coinoculation of NGR234 with NGRΩy4lO expressing gusA resulted in the blue coloration of ∼50% of the nodules (Fig. 4K). A few blue nodules were also seen when NGRΩy4lO expressing gusA was diluted 100-fold in the coinoculation experiment (not shown). Hence, the mutation in y4lO did not affect the capacity of NGR234 to infect P. vulgaris cv. Yudou No. 1. These observations are corroborated by light microscopic analysis. Nodules of P. vulgaris cv. Yudou No. 1 induced by NGR234 and NGRΩy4lO showed infected cells, which were deeply stained with toluidine blue. Cells infected by NGR234 had a granular matrix, indicating that the cells were completely filled with symbiosomes (Fig. 5A). In infected cells harboring NGRΩy4lO, however, the dark blue granular matrix appeared to be aggregated, suggesting differences in the structure of symbiosomes (Fig. 5B).

FIG. 5.

Light and electron microscopic analysis of infected nodule cells from P. vulgaris cv. Yudou No. 1 at 30 days postinoculation. (A) Semithin section of nodule cells infected with NGR234 after toluidine blue staining. (B) Nodule cells infected with NGRΩy4lO stained with toluidine blue. (C) Ultrathin section of an infected cell with NGR234 bacteroids. (D) Ultrathin section of infected cells with NGRΩy4lO bacteroids. IC, infected cell; NC, noninfected cell; S, symbiosome; ID, infection droplet; CW, cell wall. Scale bars in panels A and B, 10 μm. Scale bars in panels C and D, 2 μm.

Ultrastructural analysis of infected cells suggests a role of Y4lO in symbiosome differentiation.

To investigate ultrastructural differences in infected nodule cells of P. vulgaris cv. Yudou No. 1, nodules induced by NGR234 and NGRΩy4lO were examined by electron microscopy. In effective nodules 30 days postinoculation, cells infected by NGR234 were filled with many symbiosomes. In most cases, symbiosomes consisted of one single bacteroid surrounded by a symbiosome (peribacteroid) membrane. Few bacteroids were also seen in vacuole-like structures, which can be considered as infection droplets derived from infection threads (Fig. 5C). Ineffective nodules induced by NGRΩy4lO also contained many bacteroids, which varied in size and shape. In contrast to NGR234, however, bacteroids of NGRΩy4lO were mainly present in enlarged infection droplets with up to 30 bacteroids. Bacteroids of NGRΩy4lO seem to rapidly divide within the abnormal infection droplets (Fig. 5D).

Under the tested conditions, nodules induced by NGR234 were pink, whereas those induced by NGRΩy4lO were pink at first and later (∼30 days postinoculation) greenish, indicating the degradation of leghemoglobin. To examine symbiosome differentiation during nodule development in more detail, nodules of P. vulgaris cv. Yudou No. 1 were investigated in a time-course experiment (Fig. 6). In young nodules induced by NGR234 (24 and 26 days postinoculation [Fig. 6A and B]), many bacteroids were present within infection droplets. Some bacteroids already differentiated into symbiosomes harboring a single bacteroid. Nodules induced by NGRΩy4lO (24 and 26 days postinoculation) were similar, but most bacteroids were seen in infection droplets (Fig. 6D and E). In the later stages of symbiosis, ultrastructural differences between NGR234 and NGRΩy4lO nodules were more pronounced (Fig. 6C and F to L). Bacteroids of NGR234 localized in the periphery of infection droplets were released and rapidly differentiated into symbiosomes. The infection droplets contained few or finally no bacteroids, resulting in vacuole-like structures (Fig. 6I). The differentiated symbiosomes with bacteroids of NGR234 further divided, thereby increasing the number of symbiosomes within the infected cell (Fig. 6G and H). In contrast, most bacteroids of NGRΩy4lO divided within persistent infection droplets (Fig. 6E and J), whereas bacteroids surrounded by individual symbiosome membranes were only randomly observed. Bacteroids of NGRΩy4lO frequently contained white poly-β-hydroxybutyrate granules. Bacteroids in infection droplets were imbedded in a gray matrix, which differed from the halo-like bright material surrounding each bacteroid (Fig. 6J and K). Highly glycosylated nodule extensin is likely an essential component of the gray matrix material, which is a typical characteristic of infection threads and infection droplets (9). At 34 days postinoculation, all nodules induced by NGRΩy4lO were greenish, indicating premature senescence. “Tight junctions” formed by adjacent membranes were observed at this stage, a typical feature of senescent nodules (11, 43) (Fig. 6L). Infection droplets in old nodules induced by NGRΩy4lO were occasionally deteriorated, suggesting that bacteroids had direct contact with the host cytoplasm (not shown).

FIG. 6.

Morphological differentiation of symbiosomes in nodules from P. vulgaris cv. Yudou No. 1. (A to C) Bacteroids in nodules induced by NGR234 (at 24, 26, and 28 days postinoculation, respectively). (D to F) Bacteroids in nodules induced by NGRΩy4lO (at 24, 26, and 28 days postinoculation, respectively). (G to I) Bacteroids in nodules induced by NGR234 (at 30, 32, and 34 days postinoculation, respectively). (J to L) Bacteroids in nodules induced by NGRΩy4lO (at 30, 32, and 34 days postinoculation, respectively). S, symbiosome; ID, infection droplet; white arrow, symbiosome membrane; black arrow, halo-like matrix surrounding bacteroids; white arrowhead, tight junction formed by adjacent membranes; white asterisk, dividing bacteroids; black asterisk, poly-β-hydroxybutyrate granules. Scale bars = 2 μm.

In addition to P. vulgaris cv. Yudou No. 1, infected nodule cells of T. vogelii were analyzed (Fig. 7A and B). In pink nodules induced by NGR234 (harvested 8 weeks postinoculation), most symbiosomes harbored one single bacteroid. In contrast, persistent infection droplets were frequently observed in the ineffective greenish nodules induced by NGRΩy4lO. Some of the bacteroids in these atypical structures appear degraded. Hence, mutation in y4lO also negatively affected symbiosome differentiation in the interaction with T. vogelii.

FIG. 7.

Symbiosomes in nodules of T. vogelii and P. vulgaris cv. Yudou No. 1 induced by the double mutant NGRΩnopLΩy4lO. (A) Symbiosomes in effective T. vogelii nodules induced by NGR234. (B) Infection droplets in ineffective T. vogelii nodules induced by NGRΩy4lO. (C) Symbiosomes in effective T. vogelii nodules induced by NGRΩnopLΩy4lO. (D) Differentiation into symbiosomes in young P. vulgaris cv. Yudou No. 1 nodules induced by NGRΩnopLΩy4lO. S, symbiosome; ID, infection droplet; white arrow, symbiosome membrane; black asterisk, poly-β-hydroxybutyrate granules. Scale bars = 2 μm.

Finally, pictures were taken for nodules from P. vulgaris cv. Yudou No. 1 and T. vogelii plants, which have been inoculated with the double mutant NGRΩnopLΩy4lO. Similarly to NGR234, strain NGRΩnopLΩy4lO released bacteroids from infection droplets, resulting in bacteroids surrounded by symbiosome membranes under the conditions tested (Fig. 7C and D).

DISCUSSION

In many pathogen-host interactions, proteins belonging to the YopJ T3 effector family play an important role as virulence or avirulence factors. The y4lO gene of the symbiotic bacterium Rhizobium sp. NGR234 shows sequence similarities with T3 effectors belonging to the YopJ effector family. In this study, we provide genetic evidence that the y4lO gene plays an important role in the symbiosis with various leguminous host plants. Nodulation tests with a constructed y4lO mutant derivative (strain NGRΩy4lO) and ultrastructural analysis of nodules suggest that y4lO affects symbiosome differentiation in infected cells of P. vulgaris cv. Yudou No. 1 and T. vogelii. In the interaction with these hosts, Y4lO can be considered as a symbiotic determinant required for the formation of functional nitrogen-fixing nodules. This is reminiscent to NopL, NopP, and NopT, T3 effectors of NGR234 that play a role in host-specific nodulation (16, 31, 46). As our immunological approach failed to detect Y4lO protein, it remains an open question whether Y4lO is a protein secreted by the rhizobial T3SS.

In this study, we tested whether MAP kinase kinases are substrates for Y4lO protein. The previously characterized YopJ protein and human MKK6 served as a positive control for the acetylation assay (23, 33, 35, 36). In contrast to YopJ, Y4lO proteins expressed in E. coli lacked acetyltransferase activity under the tested in vitro conditions, suggesting that YopJ-like proteins differ in their substrate specificity. All proteins tested did not incorporate radioactivity into MtMKK1, a representative plant MAP kinase kinase with the sequence TMDPCNS in the activation loop (T and S are predicted phosphorylation sites). Thus, further attempts are required to identify target proteins of Y4lO and to test the possibility that the enzyme requires host cell factors for activation.

The symbiotic role of Y4lO extends the function of YopJ-like proteins from pathogens to a symbiotic Rhizobium strain. Genes homologous to y4lO of NGR234 have not been identified in other rhizobial strains, so far. We therefore suggest that y4lO is an imported gene which derived from a phytopathogenic bacterium. The possibility of horizontal gene transfer is supported by the genetic organization of the y4lO region in the symbiotic plasmid pNGR234a. Various ORFs flanking y4lO show similarities to insertion sequences, suggesting transposon-related sequence rearrangements during evolution (Fig. 1A). Southern blot analysis indicated a single copy of y4lO in the genome of NGR234 (not shown). The promoter region of y4lO contains a predicted tts box, a putative binding site for the transcriptional activator TtsI (30, 55). Thus, during evolution, the promoter region of y4lO apparently acquired typical features of a rhizobial promoter with a tts box. Compared to the T3 effector XopJ from X. campestris pv. vesicatoria (Fig. 1B), the predicted Y4lO protein is 46 amino acid residues shorter at the N terminus. T3 effectors from pathogenic bacteria possess an N-terminal export-signal pattern, which is required for T3SS-dependent secretion (21). Sequence analysis revealed indeed a short DNA fragment (ORF72) upstream of y4lO with sequence similarities to the N-terminal region of XopJ (Fig. 1C). We therefore suggest that ORF72 is a remnant from an ancestral XopJ-like protein (with the N terminus MGGCISRLS), which NGR234 acquired from a pathogenic bacterium. The corresponding amino acid sequence of ORF72 has characteristic features of an N-myristoylation signal, suggesting that the G2 residue of the ancestral T3 effector was myristoylated in eukaryotic host cells (by the MYR prediction server, probability of a false positive prediction for Y4lO, 1.34e⁻⁴). A recent report showed that the localization of a XopJ-GFP fusion protein within plant cells depended on the G2 residue of XopJ, indicating that the myristoylation of XopJ is required for protein targeting to the plasma membrane (49). Our mutant analysis provides evidence for a symbiotic role for Y4lO, which lacks a glycine residue at the N terminus. Future work is needed to characterize the spatial distribution of Y4lO within infected host cells and to elucidate the function of an ORF72-Y4lO fusion protein during symbiosis.

The nodulation tests of this study suggest that Y4lO mitigated the deleterious effects induced by NopL, which are most pronounced in P. vulgaris cv. Yudou No. 1 plants. The nodulation outer protein NopL is secreted via the rhizobial T3SS and a bona fide T3 effector, whose expression in plant cells caused the suppression of plant defense reactions in tobacco and Lotus japonicus (6) as well as disease-like symptoms in other plants (L. Zhang, M. Li, and C. Staehelin, unpublished results). Thus, it seems that Y4lO suppressed the deleterious effects caused by NopL. In contrast to P. vulgaris cv. Yudou No. 1, mutant analysis with NGRΩnopL revealed no deleterious effects of NopL in the interaction with the host plant T. vogelii (46), and we confirmed these findings under our nodulation test conditions (data not shown). In combination with NopP, however, NopL seems to be required for the efficient nodule formation of T. vogelii, indicating synergistic symbiosis-promoting effects (46). Surprisingly, our nodulation data suggest that NopL secreted by NGRΩy4lO may have deleterious effects on T. vogelii. Thus, NopL seems to trigger multiple responses in T. vogelii, and it is tempting to speculate that Y4lO thwarts the cytotoxic effects induced by NopL.

An ultrastructural analysis of the nodules from P. vulgaris cv. Yudou No. 1 and T. vogelii suggests that Y4lO promoted differentiation into symbiosomes harboring a single bacteroid. Cells infected by NGRΩy4lO contained enlarged infection droplets. Our electron microscopy analysis of cells infected by NGR234 revealed a rapid differentiation of symbiosomes harboring a single bacteroid, whereas bacteroids of NGRΩy4lO remained entrapped in persistent infection droplets. Thus, Y4lO apparently promoted the endocytotic release of bacteroids from infection droplets in young nodules. Interestingly, an ineffective pea mutant (sym40) displayed a similar symbiotic phenotype with enlarged infection droplets and abnormal endocytosis, indicating that host genes significantly contribute to symbiosome differentiation (reference 54 and references therein). Factors involved in symbiosome differentiation have been poorly characterized (39). Efficient synthesis of membranes might play a key role in symbiosome differentiation. Any cellular changes in the host cell affecting membrane synthesis could have negative effects on symbiosome differentiation. Furthermore, nitrogen starvation and free oxygen could affect the formation of abnormal symbiosomes in ineffective nodules. For example, young nodules of P. vulgaris cv. Negro Jamapa infected by nifA or nifH mutants of Rhizobium etli developed “multiple-occupancy symbiosomes” (11).

An analysis of the nodules induced by the double mutant NGRΩnopLΩy4lO revealed a normal differentiation into symbiosomes, suggesting that NopL negatively affected this process. It is tempting to speculate that plant defense responses induced by NopL action negatively affected symbiosome differentiation. In other words, Y4lO could function as a suppressor of a host defense response.

In conclusion, the findings of this work shed light on the symbiotic role of Y4lO. In the interaction with P. vulgaris cv. Yudou No. 1 and T. vogelii, Y4lO is crucial for the formation of effective nodules and seems to mitigate the deleterious effects of the T3 effector NopL. Future experiments are required to elucidate the molecular function of Y4lO and to investigate whether Y4lO is a T3 effector delivered into legume host cells.