Mutational analysis of a predicted double β-propeller domain of the DspA/E effector of Erwinia amylovora


  • Sabrina Siamer,

    1. INRA, UMR217, LIPP, Paris Cedex 05, France
    2. UPMC, Université Paris VI, UMR217, LIPP, Paris Cedex 05, France
    3. AgroParisTech, UMR217, LIPP, Paris Cedex 05, France
    Search for more papers by this author
  • Stéphane Gaubert,

    1. INRA, UMR217, LIPP, Paris Cedex 05, France
    2. UPMC, Université Paris VI, UMR217, LIPP, Paris Cedex 05, France
    3. AgroParisTech, UMR217, LIPP, Paris Cedex 05, France
    Current affiliation:
    1. Insitut Micalis (UMR 1319/INRA-Agroparistech) INRA, Jouy en Josas Cedex, France
    Search for more papers by this author
  • Tristan Boureau,

    1. INRA, UMR217, LIPP, Paris Cedex 05, France
    2. UPMC, Université Paris VI, UMR217, LIPP, Paris Cedex 05, France
    3. AgroParisTech, UMR217, LIPP, Paris Cedex 05, France
    Current affiliation:
    1. Université d'Angers, UMR1345, Institut de Recherches en Horticulture et Semences, Beaucouzé Cedex, France
    Search for more papers by this author
  • Marie-Noëlle Brisset,

    1. INRA, UMR 1345, Institut de Recherche en Horticulture et Semence, Beaucouzé Cedex, France
    Search for more papers by this author
  • Marie-Anne Barny

    Corresponding author
    1. UPMC, Université Paris VI, UMR217, LIPP, Paris Cedex 05, France
    2. AgroParisTech, UMR217, LIPP, Paris Cedex 05, France
    • INRA, UMR217, LIPP, Paris Cedex 05, France
    Search for more papers by this author

Correspondence: Marie-Anne Barny, AgroParisTech, UMR217, LIPP, 16 rue Claude Bernard, 75231 Paris cedex 05, France. Tel.: +33 1 44 08 17 02; fax: +33 1 44 08 16 98; e-mail :


The bacterium Erwinia amylovora causes fire blight, an invasive disease that threatens apple trees, pear trees and other plants of the Rosaceae family. Erwinia amylovora pathogenicity relies on a type III secretion system and on a single effector DspA/E. This effector belongs to the widespread AvrE family of effectors whose biological function is unknown. In this manuscript, we performed a bioinformatic analysis of DspA/E- and AvrE-related effectors. Motif search identified nuclear localization signals, peroxisome targeting signals, endoplasmic reticulum membrane retention signals and leucine zipper motifs, but none of these motifs were present in all the AvrE-related effectors analysed. Protein threading analysis, however, predicted a conserved double β-propeller domain in the N-terminal part of all the analysed effector sequences. We then performed a random pentapeptide mutagenesis of DspA/E, which led to the characterization of 13 new altered proteins with a five amino acids insertion. Eight harboured the insertion inside the predicted β-propeller domain and six of these eight insertions impaired DspA/E stability or function. Conversely, the two remaining insertions generated proteins that were functional and abundantly secreted in the supernatant suggesting that these two insertions stabilized the protein.


Many Gram-negative bacterial pathogens of plants and animals use a type III secretion system (T3SS) to inject virulence proteins, called type III effectors (T3E), into the cytosol of eukaryotic cells. T3Es subvert the metabolism of the eukaryotic cell for the benefit of the pathogen. Although the exact functions of most T3Es are not fully understood, it is known that collectively T3Es suppress host defences and allow nutrient release from the host cells to support bacterial growth within the host (Chen et al., 2010; Block & Alfano, 2011).

The AvrE- family effectors (synonyms include DspE, DspA/E, WtsE) are important T3Es of plant-pathogenic bacteria and are widespread in the genera Pseudomonas, Pantoea, Erwinia, Dickeya and Pectobacterium (Kvitko et al., 2009). They are also found in Gram-negative bacteria such as Erwinia tasmaniensis and Marinomonas mediterranea that are associated but not pathogenic on plants (Espinosa et al., 2010; Kube et al., 2010). These effectors promote bacterial growth following infection. They are required for pathogenicity of Erwinia amylovora, Pantoea stewartii subsp. stewartii and Pantoea agglomerans pv. gypsophylae (Gaudriault et al., 1997; ogdanove et al., 1998a, b; Frederick et al., 2001; Mor et al., 2001) and are important virulence factors for Pseudomonas syringae and Pectobacterium sp. (Holeva et al., 2004; Badel et al., 2006). Several members of the AvrE family of T3Es are able to block callose deposition, a plant basal defence response which strengthens the plant cell wall at the site of infection (DebRoy et al., 2004; Ham et al., 2009; Boureau et al., 2011). Furthermore, ectopic expression of several members indicated that these effectors are toxic to plant and yeast cells (Boureau et al., 2006; Oh et al., 2007; Ham et al., 2008), whereas yeast toxicity has been associated with strong perturbation of cellular trafficking (Siamer et al., 2011). Probably because of this toxicity, attempts to localize these ectopically expressed effectors inside eukaryotic cells remained unsuccessful (Boureau et al., 2006; Ham et al., 2009; Siamer et al., 2011). Furthermore, yeast toxicity has precluded the use of the yeast two-hybrid technology to identify eukaryotic interactors of the full-length proteins. Finally, effectors of the AvrE family are very large proteins, which lack overall sequence similarity with proteins of known function, and the molecular mechanism by which they carry out their functions remains unsolved.

Effectors of the AvrE family can complement each other indicating that they perform similar functions, but their sequences are fairly divergent. For example, although functional cross-complementation was demonstrated between DspA/E of E. amylovora and AvrE of P. syringae, these two proteins only share 21% identity (Bogdanove et al., 1998b). Motifs such as nuclear localization signal (NLS) or endoplasmic reticulum membrane retention signal (ERMRS) were identified in some effectors of the AvrE family and, the importance of the ERMRS motif for the function of the WtsE effector of P. stewartii was tested experimentally (Ham et al., 2009).

In this study, we performed in silico motif searches for DspA/E of E. amylovora and eleven additional members of the AvrE-effector family. Motif searches identified nuclear localization signals (NLS), endoplasmic reticulum membrane retention signals (ERMRS), leucine zipper motifs (LZ) and peroxisome targeting signals (PTS) but none of these motifs were present in all the AvrE-related effectors analysed. As during evolution, protein folds are known to be more conserved than sequence similarity; we also analysed the twelve sequences with protein threading softwares dedicated to predict protein structure. This approach led to the identification of a predicted conserved double β-propeller domain in the N-terminal portion of all the analysed sequences. We then performed a random pentapeptide mutagenesis that led to the isolation of 13 altered versions of DspA/E bearing a five amino acids insertion, eight of which were located inside the β-propeller domain. All the obtained altered versions of dspA/E were introduced into a dspA/E nonpathogenic mutant of E. amylovora, and their ability to be secreted and to restore the pathogenicity of the dspA/E mutant was tested.

Materials and methods

Bacterial strains and culture conditions

Bacterial strains and plasmids are listed in Table 1 or in the text. The Escherichia coli strain used for plasmid construction was DH5-α (Sambrook et al., 1989). Luria–Bertani Broth medium (LB) and M9 minimal medium were routinely used for bacterial growth (Sambrook et al., 1989). When necessary, the following antibiotics were added to the medium: ampicillin (Ap) 100 μg mL−1, kanamycin (Km) 50 μg mL−1or streptomycin (Sm) 100 μg mL−1.

Table 1. Plasmids and strains used in this study
NameRelevant characteristicsaReferences
  1. a

    The numbers before and after the 15 nucleotides insertion refer to the insertion site inside the dspA/E gene.

pSG52pHrpL-dspA/E without Kpn1 restriction sites-dspB cloned into pUC19This work
pDam1pSG52 with a 201-GGGGTACCCCCAGAA-202insertion into dspA/EThis work
pDam5pSG52 with a 1454-GGGGTACCCCTATGC-1455 insertion into dspA/EThis work
pDam8pSG52 with a 894-GGGGTACCCCTGAAA-895 insertion into dspA/EThis work
pDam10pSG52 with a 1461-GGGGTACCCCTGAAA-1462 insertion into dspA/EThis work
pDam12pSG52 with a 2081-GGGGTACCCCCAGGA-2082 insertion into dspA/EThis work
pDam13pSG52 with a 1507-GGGGTACCCCAGAAA-1508 insertion into dspA/EThis work
pDam14pSG52 with a 1286-GGGGTACCCCTTATT-1287 insertion into dspA/EThis work
pDam17pSG52 with a 141-GGGGTACCCCCAGAA-142 insertion into dspA/EThis work
pDam18pSG52 with a 1152-GGGGTACCCCTAAAG-1153 insertion into dspA/EThis work
pDam21pSG52 with a 875-GGGGTACCCCCTTAA-876 insertion into dspA/EThis work
pDam24pSG52 with a 2147-GGGGTACCCCAAGCA-2148 insertion into dspA/EThis work
pDam25pSG52 with a 1878-GGGGTACCCCATGAC-1879 insertion into dspA/EThis work
pDam27pSG52 with a 5162-GGGGTACCCCATGGA-5163 insertion into dspA/EThis work
M52CFBP1430, dspA/E::uidA-Kan, nonpathogenicGaudriault et al. (1997)
PA406M52 (pSG52), AmpR, KmR, pathogenicThis work
PA349M52 (pDam1), AmpR, KmR, pathogenicThis work
PA351M52 (pDam5), AmpR, KmR, reduced virulenceThis work
PA352M52 (pDam8), AmpR, KmR, pathogenicThis work
PA353M52 (pDam10), AmpR, KmR, pathogenicThis work
PA355M52 (pDam12), AmpR, KmR, pathogenicThis work
PA356M52 (pDam13), AmpR, KmR, reduced virulenceThis work
PA357M52 (pDam14), AmpR, KmR, non pathogenicThis work
PA358M52 (pDam17), AmpR, KmR, reduced virulenceThis work
PA359M52 (pDam18), AmpR, KmR non pathogenicThis work
PA360M52 (pDam21), AmpR, KmR, pathogenicThis work
PA361M52 (pDam24), AmpR, KmR, reduced virulenceThis work
PA362M52 (pDam25), AmpR, KmR, reduced virulenceThis work
PA364M52 (pDam27), AmpR, KmR pathogenicThis work

Plasmids constructions

The dspA/E gene with silent mutations, removing the three KpnI sites from dspA/E sequence, was generated by PCR and cloned into plasmid pSG52. Details of templates, oligonucleotides and construction procedures are available on supplementary Data S1.

Pentapeptide mutagenesis

Pentapeptide mutagenesis is a technique whereby a five amino acids insertion is introduced at random in a target protein (Hallet et al., 1997). Briefly, a donor strain containing the target plasmid pSG52 (ApR), pHT385 and a conjugative Tn4430 (KmR) delivery vector is mated with the DH5-α SmR plasmid-free recipient strain. By plating the mating mix simultaneously on antibiotics selecting the recipient (SmR), the target plasmid (ApR) and Tn4430 (KmR), transconjugants containing pHT385::target plasmid co-integrates are isolated. These co-integrates resolve in vivo, regenerating pHT385 and pSG52 into which a copy of Tn4430 has been inserted. Tn4430 contains KpnI sites located 5 bp from both ends of the transposon and duplicates 5 bp of target site sequence during transposition. By digesting the pSG52::Tn4430 hybrid with KpnI and relegating the digested DNA, the bulk of the transposon is deleted to generate a derivative target plasmid containing a 15-bp insertion.

Three independent conjugations were performed. Single colonies of donor and recipient grown on selective media were resuspended separately in 100 μL of LB, and 25 μL of these cell suspensions was mixed on an LB plate. Following incubation for 3 h at 37 °C, the conjugation mix was resuspended in 100 μL of LB and spread on LB plates containing Sm, Ap and Km to select for transconjugants that appeared following overnight incubation at 37 °C. The co-integrates resolved rapidly in 4% of the transconjugants (Mahillon & Lereclus, 1988); therefore, to eliminate the unresolved co-integrates, DNA was extracted in bulk from the ApR-KmR-SmR transconjugant colonies and used to transform DH5-α selecting on Km and Ap. As the resolved co-integrate is a plasmid of 14.5 kb and the unresolved co-integrate is a very large plasmid of 72 kb, all the transformants obtained correspond to resolved co-integrates. The plasmid DNA was extracted from the 100 transformants and restriction analysis with XbaI-PfLFI determined that the transposition had occurred inside the dspA/E gene for 27 plasmids. To delete Tn4430 from these 27 plasmids, a restriction with KpnI followed by a ligation and transformation of DH5-α was performed. The DNA extracted from these transformants was then digested with XbaI-KpnI to roughly delineate the localization of the 15 pb insertions. Depending on the localization of the insertion, the plasmids were sequenced with different dspA/E primers.

Pathogenicity and secretion tests

Pathogenicity tests were performed on young (six to eight leaves) actively growing apple seedlings issued from open-pollinated cv. ‘Golden delicious’. The upper part of the youngest fully expanded leaf was cut-off with scissors previously soaked in the bacterial inoculum (109 cfu mL−1 in distilled water). Symptoms were assessed 2 weeks after inoculation. A necrosis reaching the petiole was considered as a positive result. A minimum of 20 seedlings were inoculated for each strain tested in two independent experiments. Secretion tests were performed as already described (Gaudriault et al., 1997). Proteins were separated on a 7% polyacrylamide gel in sodium dodecyl sulfate and stained by rapid silver staining procedure. Immunoblot analysis was performed as described in (Gaudriault et al., 2002).

Bioinformatic analysis

Amino acids sequences of DspA/E orthologues representative of the diversity of AvrE T3Es family were retrieved from NCBI. Sequences were aligned using MUSCLE as implemented in mega 5.0 (Tamura et al., 2011) or using cobalt (Papadopoulos & Agarwala, 2007) and alignments were then checked manually. A phylogeny was reconstructed by maximum likelihood with default parameters [equal substitution rates, Jones Thomson Taylor (JTT) model] using mega 5.0 (Tamura et al., 2011). Robustness of the phylogeny was tested with 1000 bootstrap replicates. Sequences were divided into two parts (residues 1 or 200 to 1000 and residues 1001 to the end) prior to submission to fugue (, phyre2 ( and hhpred (


Bioinformatic analysis of DspA/E sequence and comparison with that of other T3Es of the AvrE family

Bioinformatic analysis was performed with the DspA/E sequence and eleven sequences representative of the AvrE family members identified in Erwinia pyrifoliae, Erwinia tasmaniensis, Pantoea.stewartii subsp stewartii, Dickeya Dadantii, Dickeya zeae, Pectobacterium atrosepticum, Pectobacterium carotovorum, Pseudomonas syringae, Pseudomonas viridiflava, Pseudomonas cichorii and Marinomonas mediterranea (Fig. 1a). Motif search using psort ii software (Nakai & Horton, 1999) identified an ERMS, a NLS, a LZ and a PTS motif in at least one effector analysed. None of these motifs are present in the entire family (Fig. 1b). The ERMRS signal was found in 9 of the 12 sequences analysed. A NLS and a LZ motif were, respectively, observed in seven and four sequences, but their localization inside these sequences were not conserved. Finally, only one sequence harboured a PTS motif. We then performed a protein threading analysis of the twelve sequences with fugue, hhpred and phyre2 softwares (Shi et al., 2001; Soding, 2005; Kelley & Sternberg, 2009), which are dedicated to predict protein structures. In the N-terminal part of DspA/E protein, threading softwares highlighted a domain between amino acid 300 and amino acid 950. Indeed, fugue, hhpred and phyre2 identified structural similarity with WD40 repeat proteins that usually fold in β-propeller domains. fugue predicted similarities between the N-terminal domain of DspA/E and the following domains: (1) the Caenorhabditis elegans homologue of yeast actin interacting protein 1 (AIP1), which folds in a double seven-bladed β-propeller domains (Z score 11.06; Mohri et al., 2004), (2) the NHL β-propeller domain of the Brain tumour Protein BRAT (Z score 12.41; Edwards et al., 2003) and (3) the β-propeller domain of RCC1 (Z score 11.15; Renault et al., 1998) (Table 2). In the N-terminal part of DspA/E and its orthologues, WD40 repeat type of β-propeller fold was also predicted with two other softwares, hhpred (Table S1) and phyre2 (data not shown). When the same type of analysis was performed with the C-terminal part of DspA/E and its orthologues, no significant hits were detected.

Table 2. Z scores obtained with the FUGUE analysis for DspA/E orthologuesa
  1. a

    Accession numbers are indicated in Fig. 1.

  2. nf: not found among the ten best scores obtained following FUGUE analysis Z SCORE ≥ 6.0 (CERTAIN 99% confidence).

DspE D. dadantiinfnf19.6
DspE D. zeae20.86nfnf
DspE P. carotovorumnf7.22nf
DspE P. Atrosepticum13.62nfnf
WtsE P. stewartiinf11.67nf
DspE E. tasmaniensisnf8.46nf
DspE E. pyrifoliae5.359.47nf
Hyp Protein M. mediteranneanf8.4510.35
AvrE P. syringae5.5nf17.42
AvrE P. cichorii16.41nfnf
AvrE P. viridiflava23.43nfnf
Figure 1.

Phylogenetic tree of DspA/E orthologues and motif localization among orthologues. (a) The accession number of the sequences used to establish this phylogenetic tree is the following: (1) DspA/E protein of Erwinia amylovora strain CFBP1430 (CAA74156.1), (2) DspE of E. pyrifoliae strain WT3 (AAS45452.1), (3) DspE of E. tasmaniensis strain Et1/99 (YP_001906490.1), (4) WtsE of Pantoea stewartii subsp. stewartii strain SS104 (AAG01467.2), (5) DspE of D. dadantii strain 3937 (ADM98561.1), (6) putative avirulence protein of Dickeya zeae strain Ech 1591 (YP_003004347.1), (7) DspE of P. atrosepticum strain SRI1039 (AAS20351.1), (8) DspE protein of P. carotovorum strain 3-2 (CBI45053.1), (9) AvrE1 of P. syringae strain DC3000 (NP_791204.1), (10) AvrE of P. viridiflava strain SP15.1a (AAX58437), (11) AvrE of P. cichorii strain 83-1 (ABA47296.1), (12) hypothetical protein of Marinomonas mediterranea strain MMB1 (YP_004312170.1). (b) The sequence of the different orthologues is represented with the different motifs found along each sequence: β-propeller (green), ERMRS (orange), PTS (black), LZ (white) and NLS (red). Orthologues are represented in the same order as in the phylogenetic tree: D. d (D. dadantii) D. z (D. zeae), Pe. c (P. carotovorum), Pe. a (P. atrosepticum), Pa. s (P. stewartii), E. t (E. tasmaniensis), E. a (E. amylovora), E. p (E. pyrifoliae), M. m (M. mediterranea), P. s (P. syringae), P. v (P. viridiflava), P. c (P. cichorii). The number on the right hand side of each represented protein indicates its length in amino acids.

Random pentapeptide mutagenesis of DspA/E

Pentapeptide mutagenesis is a technique whereby a five amino acid insertion is introduced at random in a target protein (Hallet et al., 1997). To perform this mutagenesis, we first engineered a dspA/E gene that did not contain KpnI restriction sites. This gene was cloned in plasmid pSG52. In this plasmid, dspA/E is under the control of its own promoter, pHrpL, and is followed by its chaperone gene, dspB/F. This plasmid was introduced in the dspA/E mutant. The resulting strain was pathogenic indicating that the introduced silent mutations had no effect on DspA/E function (Fig. 2b). Plasmid pSG52 was then used as a target plasmid to perform the pentapeptide mutagenesis as described in Materials and Methods. Following this mutagenesis, we analysed 100 transposition events into pSG52. The restriction analysis indicated that among these 100 transposition events, 27 occurred into the dspA/E sequence, while the remaining 73 transpositions events occurred elsewhere in the plasmid. Given the relative big size of the dspA/E gene (5.5 kb) compared with the rest of the plasmid (3.7 kb), this result indicated that Tn4430 did not insert randomly into the target plasmid pSG52. Tn4430 was then deleted from the 27 selected plasmids, and the dspA/E insertion sites were sequenced. This revealed siblings so that finally 13 new altered versions of DspA/E, each containing an additional stretch of five amino acids at different locations within the sequence, were identified (Fig 2a and Table 3). Twelve of the introduced insertions were located in the first half of the DspA/E sequence (between amino acid 47 and 716) and only one insertion was introduced in the C-terminal part of the DspA/E sequence (between amino acids 1720 and 1721). Interestingly, eight of the thirteen generated insertions were located inside the predicted β-propeller fold.

Table 3. Characteristics of the DspA/E pentapeptide mutants
  1. a

    The numbers before and after the five amino acid insertion refer to the insertion site inside the DspA/E protein.

  2. S, secretion; P, pathogenicity.

Dam 1I67-RGTPI-S68N-term++
Dam21L291-KGYPL-K292Border propeller++
Dam8K298-GVPLK-G299Border propeller++
Figure 2.

Analysis of the DspA/E pentapeptide-generated proteins. (a) Schematic view of the DspA/E protein with localization of the insertions generated following the pentapeptide mutagenesis. Insertions indicated in green, red or blue have, respectively, no impact, abolished or impaired DspA/E function. (b) Pathogenicity tests were performed with the dspA/E mutant bearing the following plasmids (from left to right): pSG52 (appropriate WT control without KpnI site); pDam10; pDam12; pDam27; pDam21; pDam8; pDam1; pDam13; pDam17; pDam5; pDam24; pDam25; pDam14; pDam18; M52: no plasmid. Results of pathogenicity tests were expressed as percentage of seedlings showing symptoms. (c) Secretion tests were performed with the dspA/E mutant bearing the indicated plasmids (from left to right) pDam24, pDam17, pDam13, pDam1, pDam10, pDam25, pDam18, pDam8, pDam14, pDam21, pDam27, pDam5, pSG52 (appropriate WT control without KpnI sites) and M52 (no plasmid). As already observed by Gaudriault et al. (1997), a cleavage product of DspA/E, indicated as DspA/E’, is often observed in the supernatant.

Analysis of the 13 altered DspA/E proteins generated following pentapeptide mutagenesis

DspA/E is required for E. amylovora pathogenicity on host plants with the dspA/E mutant being unable to produce symptom (Gaudriault et al., 1997). This allowed us to test rapidly the functionality of the 13 pentapeptide-generated versions of DspA/E.

We transformed the dspA/E mutant with each plasmid bearing one of the 13 altered versions of the protein and tested whether these altered proteins were able to complement the nonpathogenic phenotype of the dspA/E mutant (Fig. 2b). According to this criterion, the 13 altered proteins fell into three classes. One contained six proteins (Dam 1, Dam 8, Dam 10, Dam 12, Dam 27 and Dam 21) fully able to complement the dspA/E mutant for pathogenicity as more than 60% of inoculated seedlings developed symptoms. A second class of five proteins (Dam 5, Dam 13, Dam 17, Dam 24 and Dam 25) contained insertions that impaired the protein function as < 25% of inoculated seedlings developed symptoms. Finally, two proteins (Dam 14 and Dam 18) were totally unable to complement the dspA/E mutant for pathogenicity.

We then asked whether these proteins were secreted by the bacteria. Not surprisingly, the Dam 14, Dam 18 and Dam 5 proteins, which were totally or partially unable to complement the dspA/E mutant phenotype, were not detected in the supernatant (Fig. 2c). Neither could these three proteins be detected in association with the bacterial cells (data not shown). The remaining 10 pentapeptide proteins were secreted in the supernatant. Surprisingly, the Dam 10 and 12 proteins were secreted in a very abundant way suggesting that the introduced insertions likely stabilized these proteins (Fig. 2c).


DspA/E and its orthologues are very large T3Es proteins whose functions are unknown.

Our motif search identified several motifs but none was conserved in the twelve analysed sequences. The most frequent motif found was an ERMRS motif present in nine sequences. The importance of the ERMRS motif for the function of the WtsE effector of P. stewartii was tested experimentally and it was shown that a WtsE mutant that was altered in this motif is less virulent than the wild-type strain on sweet corn seedlings suggesting that this protein may associate with endoplasmic reticulum inside the eukaryotic cell (Ham et al., 2009). The presence of NLS motifs in seven sequences suggests that proteins of the AvrE family may localize inside the nucleus. Unfortunately, due to their toxicity, the localization of AvrE-like effectors is still unknown. These two motifs, nevertheless, suggest two possible localizations inside the eukaryotic cell once the effector has been injected by the bacteria. Other motifs like LZ or PTS were poorly distributed among the effectors of the AvrE family. The PTS was only observed once, and the LZ motifs were observed in four orthologues but always at different localizations questioning their relevance.

In contrast to the motif search, analysis of the sequence of DspA/E and its orthologues with three distinct protein threading softwares (fugue, hhpred and phyre2) predicted that the N-terminal part of all AvrE family proteins studied likely adopt a double seven β-propeller fold. This predicted β-propeller appears, thus, to be a hallmark of the AvrE-effector family. Proteins with β-propeller domains display a huge functional diversity, the β-propeller domain being often described as protein binding domain (Chen et al., 2011). The N-terminal domain of DspA/E is known to interact with four similar putative leucine-rich repeat receptor-like serine/threonine kinases from apple (Meng et al., 2006). It is therefore possible that the predicted β-propeller structure acts as a binding domain for these kinases. Interestingly, most of the DspA/E pentapeptide-generated proteins that we studied were altered in the predicted β-propeller structure. Three insertions likely induced misfolding and degradation of the generated proteins because we could not detect them (Dam 5, Dam 14 and Dam 18). The insertions introduced in the Dam 13, Dam 24 and Dam 25 proteins were also deleterious to the proteins because these proteins were only weakly complementing the nonpathogenic phenotype of the dspA/E mutant. These latter proteins were secreted and it remains to be determined how these insertions affect DspA/E function. Conversely, the two remaining insertions introduced in this region seem to stabilize the generated proteins (Dam 10 and Dam 12) as these proteins were secreted very abundantly and were fully complementing the nonpathogenic phenotype of the dspA/E mutant. It would be interesting to obtain the structures of the β-propeller domains of these proteins and the wild-type DspA/E protein to understand how these two insertions stabilize the protein.

We found only five insertions outside the predicted β-propeller. This indicates that the Tn4430 transposon does not insert into the plasmid sequence randomly. Among the five insertions generated outside the predicted β-propeller structure, only one, generated a weakly secreted DspA/E mutant protein, Dam 17, impaired in its ability to complement the dspA/E mutant for pathogenicity. This insertion is located between amino acids 47 and 48 of the DspA/E sequence, which is in agreement with the fact that the first 50 amino acids of DspA/E are dedicated to DspA/E secretion and translocation (Triplett et al., 2009; Oh et al., 2010). The four remaining insertions had no effect on DspA/E secretion and function.

In summary, our work identifies a predicted double β-propeller structure in the N-terminal part of the protein in all AvrE family members. Our mutational analysis indicates that introducing insertions in this domain has consequences on the stability of the protein suggesting that this domain is likely to be subject to structural constraints. All the pentapeptide-generated versions of DspA/E might represent interesting tools to study the DspA/E function in the future.


We thank D. Pichard for his help in generating pentapeptide mutants, R. Chartier for performing pathogenicity tests on apple seedlings and F. Van Gisjegem and C. Kunz for critical reading of the manuscript. We thank C. Colomb-Boureau for help in the editing of English and J. Pedron for his help in figure editing.