- Top of page
- Materials and Methods
- Supporting Information
Phytoplasmas are bacterial plant pathogens known for causing dramatic symptoms in their plant hosts, including changes in architecture, defense response and volatile production (Hogenhout et al., 2008; Mayer et al., 2008; Hoshi et al., 2009; Strauss, 2009; MacLean et al., 2011; Sugio et al., 2011a,b). These typically include witches' brooms (excessive stem production), phyllody (floral organs that turn into indeterminate leaf-like structures) and virescence (greening of flower organs). The plant phenotypic modulations appear to benefit phytoplasma fitness through the generation of more vegetative tissues and attraction of insect vectors that disperse the phytoplasmas (Hogenhout et al., 2008; Mayer et al., 2008; Hoshi et al., 2009; Sugio et al., 2011b). Phytoplasmas cause dramatic yield losses of crops worldwide, primarily because they interfere with flower, fruit and seed production (Lee et al., 2000). Nonetheless, some phytoplasmas can be beneficial. For example, phytoplasmas are commonly used to induce free branching in commercial Poinsettia cultivars (Euphorbia pulcherrima) in which the phytoplasma-induced proliferation of shorter branches generates a more attractive plant (Lee et al., 1997).
Despite the fact that phytoplasmas cannot be cultured in artificial medium, much progress has been made in the characterization of virulence effector proteins that contribute to the various changes in plant phenotype. These effectors were first discovered through phytoplasma genome sequence analyses and functional genomics approaches (Bai et al., 2009; Hoshi et al., 2009). In the genome of Aster Yellows phytoplasma strain Witches' Broom (AY-WB) more than 50 secreted AY-WB proteins (SAPs) were identified that are candidate virulence effectors (Bai et al., 2009). Phytoplasmas reside in the cytoplasm of sieve cells of the plant phloem. Being intracellular, phytoplasmas secrete effectors via the Sec-dependent secretion pathway in which the signal peptide is cleaved off (Kakizawa et al., 2004). The majority of phytoplasma effector proteins are smaller than 40 kDa and therefore, upon secretion by the phytoplasma, may unload from the sieve cells and migrate to adjacent tissues (Imlau et al., 1999; Bai et al., 2009). SAP11 was shown to predominantly target plant cell nuclei when transiently produced in Nicotiana benthamiana leaves and in AY-WB-infected plants (Bai et al., 2009), and induces stem proliferation, alterations in leaf shape and downregulation of jasmonic acid (JA) production, the latter increasing susceptibility of plants to the AY-WB leafhopper vector Macrosteles quadrilineatus (Sugio et al., 2011b). Another AY-WB effector, SAP54, induces the production of leafy indeterminate flowers that resemble the phyllody symptoms characteristic of AY-WB-infected plants (MacLean et al., 2011). TENGU is an effector characterized from Onion Yellows (OY) phytoplasma and induces dwarfism and witches' brooms in plants (Hoshi et al., 2009; Sugawara et al., 2013).
So far, the plant targets of one phytoplasma effector, SAP11, have been identified. It was found that SAP11 binds and destabilizes class II CIN (CINCINNATA) TCP (TEOSINTE-BRANCHED, CYCLOIDEA, PROLIFERATION FACTOR 1 AND 2) (Sugio et al., 2011b). CIN-TCPs regulate various plant developmental functions, arguably the most obvious of which is leaf morphogenesis (Martin-Trillo & Cubas, 2010). Transgenic Arabidopsis plants that overexpress miR319, which is a negative regulator of five of eight CIN-TCPs, exhibit changes in leaf shape and size due to excess cell division, mainly at the leaf margins, resulting in the production of large crinkly leaves (Palatnik et al., 2003). This Arabidopsis leaf crinkling phenotype is more severe in plants that overexpress miR319 together with an artificial miRNA miR3TCP targeting the remaining three CIN-TCPs (Efroni et al., 2008). SAP11 was previously shown to bind and destabilize all eight CIN class II TCPs leading to the induction of severe leaf crinkling and downregulation of jasmonic acid (JA) that is the direct result of reduced CIN-TCP presence (Schommer et al., 2008; Sugio et al., 2011b; Danisman et al., 2012). Whilst SAP11 may bind class I TCPs, this effector does not appear to destabilize these TCPs (Sugio et al., 2011b). Class I and II TCPs have antagonistic functions in controlling plant development (Kosugi & Ohashi, 2002; Li et al., 2005; Martin-Trillo & Cubas, 2010; Danisman et al., 2012).
SAP11 is c. 10 kDa (90 amino acids), but nonetheless appears to encode at least three distinct activities, which are targeting of nuclei, binding of TCPs and destabilization of TCPs. Nuclear-localized virulence effectors of other bacterial pathogens are often much larger than the size exclusion limit of the nuclear pore complex, which is c. 60 kDa (Gorlich, 1998; Talcott & Moore, 1999) and have distinct domains involved in nuclear targeting and target binding (Rivas, 2012). To better understand how the various functions are accommodated in SAP11, we employed yeast two-hybrid analyses, agroinfiltration assays in N. benthamiana leaves and stable transgenic expression of SAP11 constructs in Arabidopsis to dissect the domains involved in SAP11 nuclear targeting and TCP-binding and destabilization. Surprisingly, SAP11 has a linear modular structure with different parts of the effector being involved in nuclear localization, TCP binding and TCP destabilization. We discuss our findings in the broader context of virulence effector evolution in phytoplasmas.
- Top of page
- Materials and Methods
- Supporting Information
We found that SAP11 mutants lacking the entire N-terminal domain, including the NLS, interacted with TCPs but were impaired in the destabilization of these transcription factors. In addition, SAP11 mutants that lacked the C-terminal domain, including a predicted coiled coil structure, were impaired in both binding and destabilization of TCPs. Unlike wild-type SAP11, these SAP11 mutants did not alter leaf morphogenesis. SAP11 mutants with mutations in the NLS and with a nuclear export signal (NES) had cytoplasmic distributions and both bound and destabilized TCP transcription factors, but instigated weaker changes in Arabidopsis leaf morphogenesis than wild-type SAP11. Whilst the various deletions may affect the three-dimensional structure of SAP11, our data suggest that this structure may not be essential for the three SAP11 activities. First, deletion of the N-terminal 28 amino acid of the mature (without signal peptide) 90-amino acid SAP11 protein did not affect TCP binding of SAP11 and deletion of C-terminal 30 amino acid did not affect SAP11 nuclear localization. Secondly, SAP11 mutants without the C-terminal 15 amino acids bind and destabilize TCPs. Finally, mutations in the NLS or addition of a NES did not affect SAP11 TCP binding and destabilization. Thus, SAP11 appears to have a modular organization in which specific amino acids/domains in different parts of the SAP11 protein are required for nuclear localization, TCP binding or destabilization, and these domains appear to not affect each other's activities. A linear modular structure of SAP11 is in agreement with the hypothesis that additions of GFP, NES and other tags to the SAP11 N- and C-termini do not compromise the SAP11 ability of nuclear localization and TCP binding and destabilization.
We demonstrated that SAP11 interaction with TCPs is not sufficient for TCP destabilization and induction of leaf crinkling, because SAP11ΔN – which binds but not destabilizes CIN-TCPs – does not induce leaf crinkling in Arabidopsis. Therefore, it is unlikely that SAP11 blocks TCP action through steric hindrance, but rather mediates active degradation of CIN-TCPs, possibly by interacting with a plant-specific helper component that has a role in the plant protein degradation pathway. The involvement of a plant helper component is also suggested by the yeast two-hybrid experiments in which the TCPs were not degraded in the presence of wild-type SAP11 and SAP11ΔN. Given that we did not observe a reduction of destabilization in the presence of proteasome inhibitor and protease inhibitor cocktail, it remains unclear how SAP11 mediates destabilization of TCPs complicating studies of where and how SAP11-mediated TCP degradation occurs in the cell. The mechanism may be revealed as soon as a plant helper component is identified. We predict that this helper protein binds to the SAP11 N-terminal 28 amino acids in the mature protein as this domain is required for TCP destabilization but not TCP binding.
Given the important function of CIN-TCPs in leaf morphogenesis, we used the amount of leaf crinkling as a proxy for CIN-TCP presence and function. SAP11ΔNLS-NES transgenic Arabidopsis plants show no or reduced leaf crinkling phenotypes compared to SAP11 transgenic plants, indicating that nuclear localization of SAP11 contributes to TCP destabilization. Indeed, many TCPs have mono- or bipartite NLSs and nuclear localization has been demonstrated experimentally (Baba et al., 2001; Suzuki et al., 2001; Koroleva et al., 2005; Martin-Trillo & Cubas, 2010), although some TCPs target chloroplasts (Baba et al., 2001). Nonetheless, SAP11ΔNLS-NES is still able to destabilize TCPs in co-expression analyses in N. benthamiana leaves. This apparent discrepancy may be due to incomplete exclusion of SAP11ΔNLS-NES from the nucleus. SAP11ΔNLS-NES is a small protein that can passively migrate into the nucleus, but once in the nucleus the NES transports this protein out of the nucleus. In N. benthamiana transient assays, where transgenes are expressed at high levels, sufficient amounts of SAP11ΔNLS-NES may migrate into the nucleus to interact with the nuclear-localized TCPs. It is not yet known if degradation of TCPs occurs inside the nucleus or cytoplasm or in both cell compartments. Therefore, nuclear targeting of SAP11 likely increases the opportunity of SAP11 to bind and destabilize CIN-TCPs.
The coiled coil domain located between residues 91 and 106 of SAP11 protein (Fig. 1a) is required for SAP11 interactions with the CIN-TCPs. We previously provided evidence that SAP11 interacts with the 59-amino acid basic helix-loop-helix motif TCP domain that is involved in the DNA binding and protein–protein interactions of TCP factors (Cubas et al., 1999; Kosugi & Ohashi, 2002; Sugio et al., 2011b). Because coiled coil is a structural motif in proteins that allows 2–7 α-helices to coil together (Liu et al., 2006), it is likely that SAP11 interacts with one or both of the helices in the TCP domain. Helix 1 is less conserved than helix 2 amongst the class I and class II CIN and CYC/TB1 TCPs (Martin-Trillo & Cubas, 2010), allowing for SAP11-binding specificity amongst the different TCPs. SAP11 virulence effectors with highly similar sequences have been identified in diverse phytoplasmas that belong to evolutionary distinct clusters in the phytoplasma phylogenetic tree (Fig. 9) (Hogenhout et al., 2008; Chung et al., 2013). Intriguingly, alignment of the SAP11 protein sequences reveals that the NLS and TCP-binding coiled coil domain are least conserved (Fig. 9) indicating that these regions may be under selection allowing diversity of SAP11 homologs in nuclear vs cytoplasmic targeting and interaction with a different sets of TCPs or other (transcription factor) targets in the plant.
Figure 9. Multiple sequence alignment of AY-WB phytoplasma SAP11 with SAP11 homologs of other phytoplasmas. SAP11 protein sequences obtained from Poinsettia Branch-Inducing Phytoplasma (PBIP), Vaccinium Witches'-Broom Phytoplasma (VWBP), Aster Yellows phytoplasma strain Witches' Broom (AYWB), Maize bushy stunt phytoplasma (MBSP) and Peanut Witches' Broom (PnWB) were aligned using the CLUSTAL 2.1 program. PBIP and VWBP belong to 16SrDNA group III, AYWB and MBSP to groups 1A and 1B, respectively, and PnWB to group II as indicated. Genbank accession numbers: GI:515759334 (PBIP); GI:515761117 (VWBP); GI:85057650 (AYWB); GI:471234556 (PnWB). Signal peptide sequence is underlined, nuclear localization signal (NLS) is indicated in red font, and part of coiled coil structure required for binding TCP in blue font. *, Fully conserved residues, :, conservation of residues with strongly similar properties and ., conservation of residues with weakly similar properties.
Download figure to PowerPoint
It was noticed that 35S:SAP11 transgenic lines have higher levels of variations in transgene expression levels than the transgenic lines expressing mutants of SAP11 (Fig. 8b). A possible explanation is that leaves of SAP11 transgenic lines produce more cells (as evidenced by the curly leaves) and because CIN-TCPs, which are degraded by SAP11, promote cell maturation the cells may also be metabolically more active in the SAP11 transgenic lines than in SAP11 mutant lines. Thus, small changes in initial wild-type SAP11 expression levels may be amplified into larger expression differences during plant growth.
SAP11 is the only virulence effector of phytoplasmas for which plant targets has been identified so far. However, phytoplasmas have multiple virulence effectors; more than 50 candidate virulence effectors has been identified in AY-WB (Bai et al., 2009), many of which are represented in different phytoplasmas (Chen et al., 2012; Saccardo et al., 2012; Chung et al., 2013). The majority of the phytoplasma virulence effector genes lie on genetic islands resembling mobile transposons that may have been derived from ancient prophage attacks and that are likely exchanged between phytoplasmas (Bai et al., 2006; Jomantiene & Davis, 2006; Wei et al., 2008; Chung et al., 2013) providing a possible explanation of why SAP11 homologs of distantly related phytoplasmas are more similar in sequence than homologs of more closely related phytoplasmas (Fig. 9). Other phytoplasma virulence effectors, such as TENGU and SAP54, also have a region with a predicted coiled coil structure. Whilst the targets of these effectors have not yet been described, they both induce dramatic changes in plant development; TENGU induces witches' brooms and dwarfism (Hoshi et al., 2009; Sugawara et al., 2013) and SAP54 the formation of leafy indeterminate flower development (MacLean et al., 2011), indicating that these effectors probably target plant (transcription) factors as well. Thus, phytoplasma virulence effectors may have evolved as versatile linear modular proteins that target a variety of plant (transcription) factors to evoke architectural changes in plant hosts. This effector versatility may be particularly important for phytoplasma success, because these bacteria are dependent on insect vectors for dispersal and hence do not choose their plant hosts. If the changes are beneficial to phytoplasma fitness by, for example, generating more plant tissue for colonization by phytoplasmas and the insect vectors that disperse the phytoplasmas (MacLean et al., 2011; Sugio et al., 2011a,b; Sugio & Hogenhout, 2012), the effector genes are more likely to prevail in phytoplasma populations.