Phytoplasma effector SAP54 hijacks plant reproduction by degrading MADS-box proteins and promotes insect colonization in a RAD23-dependent manner

Abstract

Pathogens that rely upon multiple hosts to complete their life cycles often modify behavior and development of these hosts to coerce them into improving pathogen fitness. However, few studies describe mechanisms underlying host coercion. In this study, we elucidate the mechanism by which an insect-transmitted pathogen of plants alters floral development to convert flowers into vegetative tissues. We find that phytoplasma produce a novel effector protein (SAP54) that interacts with members of the MADS-domain transcription factor (MTF) family, including key regulators SEPALLATA3 and APETALA1, that occupy central positions in the regulation of floral development. SAP54 mediates degradation of MTFs by interacting with proteins of the RADIATION SENSITIVE23 (RAD23) family, eukaryotic proteins that shuttle substrates to the proteasome. Arabidopsis rad23 mutants do not show conversion of flowers into leaf-like tissues in the presence of SAP54 and during phytoplasma infection, emphasizing the importance of RAD23 to the activity of SAP54. Remarkably, plants with SAP54-induced leaf-like flowers are more attractive for colonization by phytoplasma leafhopper vectors and this colonization preference is dependent on RAD23. An effector that targets and suppresses flowering while simultaneously promoting insect herbivore colonization is unprecedented. Moreover, RAD23 proteins have, to our knowledge, no known roles in flower development, nor plant defence mechanisms against insects. Thus SAP54 generates a short circuit between two key pathways of the host to alter development, resulting in sterile plants, and promotes attractiveness of these plants to leafhopper vectors helping the obligate phytoplasmas reproduce and propagate (zombie plants).

Figure 1. Phytoplasma SAP54 interacts specifically with the Keratin-like (K) domain of selected Type II MADS-box transcription factors (MTFs).

(A) A comprehensive yeast two-hybrid screen of 106 Arabidopsis MTFs reveals that SAP54 interacts with members of the Type II subfamily of MTFs (proteins that interact with SAP54 are indicated in red font). For simplicity, not all MTFs are included in the phylogenetic tree. (B) SAP54 interacts primarily with the K domain of AP1. AD, GAL4-activation domain; BD, GAL4-DNA binding domain; EV, empty vector control. (C) Flowers produced from healthy (left) and AY-WB–infected (right) Arabidopsis lines approximately 4 wk postinoculation. (D) SAP54 (indicated by an arrow) co-immunoprecipitates with SEP3–GFP but not FUL–GFP or AG–GFP. Flowers for immunoprecipitation experiments were harvested from transgenic lines pictured in panel C at an early point of infection (approximately 2 wk postinoculation) to minimize MTF loss due to destabilization. Equal loading of samples was confirmed via Bradford assays to quantify protein concentration.

Figure 2. Phytoplasma SAP54 interacts with and destabilizes MADS-box transcription factors in plants.

(A) MTFs AP1 and SEP3 are destabilized in AY-WB–infected Arabidopsis lines. Flowers from healthy and phytoplasma-infected plants were harvested approximately 4 wk postinoculation. (B) MTFs are destabilized when expressed in the presence of SAP54. 10xmyc-tagged MTFs were transiently co-expressed with flag-tagged SAP54 or an RFP control in agroinfiltrated N. benthamiana leaves. (C) SAP54-mediated destabilization of AP1 is inhibited by epoxomicin. Infiltrated tissues were treated with 50 µM epoxomicin (dissolved in DMSO) 8 h prior to harvest. (D) MTFs AP1, SEP3, and SOC1 co-immunoprecipitate with GFP-tagged SAP54. Co-immunoprecipitation experiments of these Type II MTFs were performed alongside Type I MTF AGL50, which was not detected. Proteins were transiently expressed in N. benthamiana in the presence or absence of 50 µM epoxomicin to stabilize MTFs.

Figure 3. Phytoplasma SAP54 interacts with Arabidopsis RAD23 proteins.

(A) SAP54 interacts with Arabidopsis RAD23C and RAD23D but not RAD23A or RAD23B isoforms in yeast two-hybrid assays. (B) RAD23C (44 kDa) and RAD23D (40 kDa) co-immunoprecipitate with GFP–SAP54 in samples obtained from transgenic Arabidopsis expressing 35S:GFP–SAP54. We did not detect RAD23 following immunoprecipitation of GFP (in transgenic Arabidopsis expressing 35S:GFP), nor did we detect an interaction with RAD23A or RAD23B in an Arabidopsis rad23CD double mutant. Equal loading of samples was verified via Bradford assays to confirm protein concentration. (C) Flowers produced from transgenic lines expressing 35S:GFP–SAP54 in wild-type (Col-0) and rad23 mutant Arabidopsis lines indicate that SAP54-induced phyllody requires RAD23C and RAD23D. (D) Western blot analysis reveals GFP–SAP54 expression levels in rosette leaves harvested from plants in panel C. GFP–SAP54 is indicated by an arrow. AD, GAL4-activation domain; BD, GAL4–DNA binding domain; EV, empty vector control.

Figure 4. Arabidopsis rad23BCD triple mutants do not exhibit symptoms of virescence or phyllody when infected with AY-WB.

(A) Flowers produced from AY-WB–infected rad23BD mutants produce leaf-like flowers, whereas infected rad23BCD mutants grow flowers that resemble those of healthy plants. (B) Western blot analysis reveals that SEP3 is destabilized in rad23BD leaf-like flowers but not in rad23BCD flowers. SAP54 was detected in flowers harvested from AY-WB–infected rad23 mutants but not healthy Arabidopsis plants. (C) The infection status of plants in panel A was confirmed using primers specific for AY-WB.

Figure 5. Aster leafhopper Macrosteles quadrilineatus demonstrates oviposition preference for plants with leaf-like flowers.

(A) M. quadrilineatus produces significantly more progeny on AY-WB–infected rad23BD mutants (leaf-like flower phenotype) compared to rad23BCD mutant plants (non-leaf-like flower phenotype) (t ₍₅₎ = 4.7; p = 0.042). Insects do not exhibit a preference between healthy rad23BD and rad23BCD plants (t ₍₅₎ = 0.45; p = 0.694). (B) M. quadrilineatus adults produce more nymphs on transgenic Arabidopsis expressing GFP-tagged SAP54 (leaf-like flowers) compared to GFP control plants (wild-type flowers) (t ₍₇₎ = 6.45; p = 0.008). In these experiments, 10 male and 10 female M. quadrilineatus adults were released in a choice cage containing two test plants for the period of 5 d. After removal of adult insects, plants were bagged individually and incubated for 14 d to allow nymph emergence. The graphs in panel A and B represent the percentage of M. quadrilineatus nymphs found on each test plant within a single choice cage.