**Background:** Phytoplasmas are obligate intracellular plant pathogenic bacteria that cause symptoms such as phyllody—abnormal floral organ development. Phyllogens are effector proteins that induce phyllody in eudicots. Previous phylogenetic analyses suggested that phyllogen genes (phyllogens) undergo horizontal gene transfer (HGT) between phytoplasma species, but the mechanism was unknown. Potential mobile units (PMUs) are putative transposable elements found in phytoplasma genomes that contain multicopy genes and have been hypothesized to mediate HGT. This study aimed to determine whether PMUs drive the horizontal transfer of phyllogens by analyzing the synteny and conservation of phyllogen flanking regions across diverse phytoplasma strains.
**Methods:** The authors analyzed phyllogen flanking genomic regions from 17 phytoplasma strains related to six 'Candidatus' species, including three newly sequenced draft genomes (HP, RhY, and PaWB-Japan strains of 'Ca. P. asteris'). PCR was used to determine flanking sequences for the RhY, PaWB-Japan, and PEY strains. Genes were annotated using MetaGeneAnnotator and BLASTp. Phylogenetic trees were constructed using the neighbor-joining method in MEGA X. Pairwise sequence identities were calculated using the Sequence Demarcation Tool. Functional analysis of phyllogens was performed using a tobacco rattle virus (TRV)-based expression system in Arabidopsis thaliana, testing phyl-A group phyllogens from RP166 and PEY strains and phyl-D group phyllogens from NCHU2019 and HP strains.
**Key Results:** Most phyllogens (all phyl-A and phyl-C, and seven of ten phyl-D group phyllogens) were flanked by PMU-associated genes (fliA, ssb, dam, himA, hflB, smc, tmk, dnaB, dnaG, and tra5). PMUs were categorized into two types: type 1 PMUs (found in six strains with phyl-A group phyllogens) contained hp1, hp2, and an hflB-like gene; type 2 PMUs (found in seven strains with phyl-D group phyllogens and one with phyl-C) contained hp4, hp5, hp6, hp7, smc, smc-like, hp8, and hp9 genes. Phyllogen phylogeny correlated perfectly with PMU type. In type 1 PMUs, phyllogens were 87-100% identical between 'Ca. P. asteris' and 'Ca. P. phoenicium', while flanking genes showed identities as low as 57%. In type 2 PMUs, phyllogens were 91-100% identical across species, while flanking genes showed identities as low as 32%. Several PMU-associated genes were truncated by premature stop codons (e.g., hflB in OY strain, smc in HP, RhY, and PaWB-Japan strains). Many PMUs lacked tra5 and other downstream genes. In 'Ca. P. asteris', type 1 PMUs were inserted into different genomic regions in AY-WB versus OY/DY2014/RP166 strains. Type 2 PMUs were inserted into different regions in PaWB-Japan/PaWB-China, MD-China, and HP/RhY strains. The PaWB strains retained intact tra5 genes and inverted repeat sequences at PMU boundaries. All tested phyllogens (PHYL_RP166, PHYL_PEY, PHYL_NCHU2019, PHYL_HP) induced phyllody in A. thaliana, converting sepals, petals, and stamens into leaf-like structures. PHYL_OY showed reduced activity due to a unique lysine at position 33; a K33E mutant restored full phyllody induction.
**Clinical Implications:** This study provides the first mechanistic evidence that PMUs drive the horizontal transfer of phyllogen effector genes among phytoplasma species and strains. The finding that phyllogen sequences and functions are highly conserved while PMU-associated genes deteriorate suggests that phyllogens are critical for phytoplasma fitness, likely because phyllody symptoms facilitate pathogen accumulation and insect vector attraction. The identification of PMU types as predictors of phyllogen presence has practical implications for understanding and potentially managing phytoplasma diseases. The demonstration that PMU-mediated HGT is a major mechanism for sharing virulence genes across phytoplasma species advances our understanding of how plant pathogenic bacteria evolve and acquire new pathogenic capabilities. The observation that 'Ca. P. asteris' strains with phyl-D phyllogens share ecological niches with 'Ca. P. ziziphi' (co-infection in jujube plants, shared insect vectors) suggests that HGT occurs between species occupying the same ecological niche.