Additive roles of PthAs in bacterial growth and pathogenicity associated with nucleotide polymorphisms in effector-binding elements of citrus canker susceptibility genes

Abstract

Citrus canker, caused by Xanthomonas citri, affects most commercial citrus varieties. All X. citri strains possess at least one transcription activator-like effector of the PthA family that activates host disease susceptibility (S) genes. The X. citri strain 306 encodes four PthA effectors; nevertheless, only PthA4 is known to elicit cankers on citrus. As none of the PthAs act as avirulence factors on citrus, we hypothesized that PthAs 1-3 might also contribute to pathogenicity on certain hosts. Here, we show that, although PthA4 is indispensable for canker formation in six Brazilian citrus varieties, PthAs 1 and 3 contribute to canker development in 'Pera' sweet orange, but not in 'Tahiti' lemon. Deletions in two or more pthA genes reduce bacterial growth in planta more pronouncedly than single deletions, suggesting an additive role of PthAs in pathogenicity and bacterial fitness. The contribution of PthAs 1 and 3 in canker formation in 'Pera' plants does not correlate with the activation of the canker S gene, LOB1 (LATERAL ORGAN BOUNDARIES 1), but with the induction of other PthA targets, including LOB2 and citrus dioxygenase (DIOX). LOB1, LOB2 and DIOX show differential PthA-dependent expression between 'Pera' and 'Tahiti' plants that appears to be associated with nucleotide polymorphisms found at or near PthA-binding sites. We also present evidence that LOB1 activation alone is not sufficient to elicit cankers on citrus, and that DIOX acts as a canker S gene in 'Pera', but not 'Tahiti', plants. Our results suggest that the activation of multiple S genes, such as LOB1 and DIOX, is necessary for full canker development.

Keywords: LATERAL ORGAN BOUNDARIES genes; TAL effectors; Xanthomonas aurantifolii; Xanthomonas citri; citrus canker; citrus dioxygenase; pthA.

Figure 1

Host‐dependent effect of single pthA deletions in canker development. Leaves of the sweet orange varieties ‘Hamlin’, ‘Valencia’, ‘Natal’, ‘Pera’ and ‘Sorocaba’, and ‘Tahiti’ lemon, were infiltrated with bacterial suspensions (10⁵ cells/mL) of wild‐type (Wt) Xanthomonas citri and respective single pthA‐deletion mutants (Δ1, Δ3, Δ4). Canker symptoms were evaluated 14 days after bacterial inoculation. pthA4 is essential to elicit cankers in all citrus hosts; nevertheless, pthA1 also contributes significantly to symptom development in ‘Hamlin’, ‘Valencia’ and ‘Pera’, but not in ‘Tahiti’ plants. Similarly, a deletion in pthA3 also reduces canker formation in ‘Pera’ and, to a lesser extent, in ‘Hamlin’ and ‘Sorocaba’, but not in ‘Valencia’, ‘Natal’ or ‘Tahiti’ plants.

Figure 2

Host‐dependent effect of double and triple pthA deletions in canker development. Leaves of ‘Hamlin’, ‘Valencia’, ‘Natal’, ‘Pera’, ‘Sorocaba’ and ‘Tahiti’ plants were infiltrated with bacterial suspensions (10⁵ cells/mL) of wild‐type (Wt) Xanthomonas citri and respective double (Δ1‐3, Δ1‐4, Δ3‐4) and triple (Δ1‐3‐4) pthA‐deletion mutants, and canker symptoms were evaluated 14 days after bacterial inoculation. Although a deletion in pthA4 is sufficient to abolish cankers in all the citrus hosts, an additive effect of pthAs 1 and 3 in disease development is noted in ‘Pera’, ‘Natal’ and ‘Sorocaba’ plants.

Figure 3

Additive effect of PthAs on bacterial growth in planta. Leaves of sweet orange and lemon plants were infiltrated with bacterial suspensions (10⁶ cells/mL) and the growth of wild‐type Xanthomonas citri (WT) and respective pthA‐deletion mutants (Δ) was monitored at 2 and 14 days after bacterial inoculation. The double (Δ1‐4 and Δ3‐4) and triple (Δ1‐3‐4) pthA‐deletion mutants grew significantly less well than the respective single mutants and the WT bacteria in all the citrus hosts tested, suggesting an additive effect of PthAs on bacterial growth in planta. Bacterial growth, expressed in colony‐forming units (CFU)/cm² of leaf, is the mean of three biological replicates. The error bars denote standard deviations, whereas the asterisks above the bars indicate statistically significant differences between X. citri and mutant‐inoculated plants (P < 0.05).

Figure 4

Differential PthA‐dependent induction of LOB (LATERAL ORGAN BOUNDARIES) and DIOX (citrus dioxygenase) genes in sweet orange and lemon plants. Expression levels of LOB1, LOB2, LOB3 and DIOX in ‘Pera’ and ‘Tahiti’ leaves infiltrated with the wild‐type (WT) Xanthomonas citri or the single pthA‐deletion mutants (Δ1, Δ3 and Δ4), 72 h post‐inoculation. (A) PthAs 1, 3 and 4 are required for LOB1 induction in lemon, whereas only PthA4 is important for LOB1 induction in ‘Pera’ sweet orange. However, PthA1 seems to repress LOB1 transcription in this host. (B, C) LOB2 and LOB3 are induced by X. citri in ‘Pera’ and ‘Tahiti’ plants; however, PthAs 1, 3 and 4 do not alter significantly LOB2 expression in ‘Tahiti’ or LOB3 expression in ‘Pera’ plants. LOB2 expression in ‘Pera’ sweet orange appears to require PthA1, but not PthA3 or PthA4. (D) The DIOX gene is also induced by X. citri WT and respective single pthA mutants in ‘Pera’ and ‘Tahiti’ leaves; however, deletion in pthA 1, 3 or 4 significantly reduces DIOX expression in ‘Pera’, but not in ‘Tahiti’, leaves. The error bars denote standard deviations, whereas the asterisks above the bars indicate statistically significant differences between X. citri and mutant‐inoculated plants (P < 0.05).

Figure 5

LOB1 (LATERAL ORGAN BOUNDARIES 1) induction does not always correlate with canker development. (A) Leaves of ‘Pera’ sweet orange and ‘Mexican’ lime were infiltrated with Xanthomonas citri and X. aurantifolii ‘C’, and canker symptoms were evaluated 14 days after bacterial infiltration (10⁶ cells/mL). In contrast with a hypersensitive response (HR) elicited by X. aurantifolii ‘C’ in sweet orange compared with lemon, canker lesions developed in both ‘Pera’ and ‘Mexican’ lime inoculated with X. citri. (B) Expression levels of LOB1, LOB2, LOB3 and DIOX (citrus dioxygenase) genes in ‘Pera’ (P) relative to ‘Mexican’ lime (M) leaves in response to X. citri or X. aurantifolii ‘C’ infection. LOB1 is weakly induced in the compatible interactions in ‘Mexican’ lime and highly induced in both the compatible and incompatible interactions in ‘Pera’. The expression level of DIOX, but not of LOB1, LOB2 or LOB3, is significantly reduced in the incompatible interaction between ‘Pera’ and X. aurantifolii ‘C’. The error bars denote standard deviations, whereas the asterisks above the bars indicate statistically significant differences between X. citri‐ and X. aurantifolii‐inoculated plants (P < 0.05).

Figure 6

The citrus DIOX protein (CsDIOX) is structurally related to Arabidopsis feruloyl‐CoA 6′‐hydroxylase 1 (F6′H1). (A) Superposition of the CsDIOX structural model (cyan) with the Arabidopsis AtF6′H1 crystal structure (sand), depicting the iron atom (yellow) and 2‐oxoglutarate (green) (PDB code 4XAE; Sun et al., 2015), showing that CsDIOX displays the same folding and topology as AtF6′H1. (B) All of the amino acid residues responsible for the binding of the iron atom (H234, D236, H292), 2‐oxoglutarate (N217, Y219, R302, S304) and feruloyl CoA (F150, S152, R213, N215, I237, F308) are structurally conserved in the CsDIOX model.

Figure 7

Psoralen inhibits canker formation in ‘Pera’ sweet orange. Leaves of ‘Pera’ sweet orange and ‘Tahiti’ lemon were infiltrated with a suspension of Xanthomonas citri (10⁶ cells/mL) in the absence or presence of 0.1 or 0.5 mm psoralen (Ps). Canker symptoms (A) and bacterial counts (B) were evaluated 10 days after bacterial inoculation. Psoralen significantly inhibited canker development and bacterial growth in ‘Pera’ sweet orange, but not in ‘Tahiti’ lemon. Bacterial growth, expressed in colony‐forming units (CFU)/cm² of leaf, is the mean of three biological replicates. The error bars denote standard deviations, whereas the asterisks above the bars indicate statistically significant differences between means (P < 0.05).

Figure 8

Promoters of LOB1 (LATERAL ORGAN BOUNDARIES 1), LOB2 and DIOX (citrus dioxygenase) genes show nucleotide polymorphisms at PthA sites. The promoter regions of the citrus LOB1 (A), LOB2 (B) and DIOX (C) genes were amplified from ‘Pera’ and ‘Tahiti’ plants and compared with those of the Citrus sinensis ‘Valencia’ and C. clementine ‘Clemenules’ cultivars (Wu et al., 2014; Xu et al., 2013). The predicted PthA1‐, PthA2‐, PthA3‐ and PthA4‐binding sites are boxed in yellow, green, blue and red, respectively. The Dof‐ and WRKY‐binding sites are coloured in grey and light grey, respectively, whereas the TATA box elements are in purple. Arrows in (C) represent a Dof palindromic sequence. Most of the single nucleotide polymorphisms (SNPs) found in ‘Pera’ relative to ‘Tahiti’ plants are also present in ‘Valencia’ and ‘Clemenules’ cultivars.