Editing of the rice importin gene IMPα1b results in sequestration of TAL effectors from plant cell nuclei

Abstract

Transcription activator-like effectors (TALEs) produced by plant pathogenic bacteria mainly belonging to the genus Xanthomonas cause plant diseases through activation of host susceptibility genes in plant cell nuclei. How TALEs enter plant cell nuclei was not clear until recent studies about PthXo1 and TALI, two TALEs produced by Xanthomonas oryzae pv. oryzae (Xoo) and X. oryzae pv. oryzicola (Xoc), the rice (Oryza sativa) pathogens that cause bacterial blight and bacterial leaf streak, respectively. Here, we report that rice importin IMPα1b serves as a nuclear transport receptor in rice plants to facilitate the nuclear import of PthXo1 and TALI from Xoo and Xoc, respectively. While wild-type (WT) rice plants support the nuclear import of PthXo1 and TALI, nuclear trafficking is defective in OsIMPα1b loss-of-function mutants generated by clustered regularly interspaced short palindromic repeats (CRISPR)–CRISPR-associated 9 (Cas9)-based gene editing. In the edited plants infected by Xoo, OsIMPα1b sequesters PthXo1 from the nucleus, the PthXo1-targeted rice susceptibility gene is no longer activated, and bacterial virulence and blight disease are alleviated as a result. In the edited plants infected by Xoc, OsIMPα1b sequesters TALI from the nucleus, the role of TALI in suppressing rice defense responses is nullified, and rice defense responses are in turn activated to inhibit bacterial virulence and alleviate bacterial leaf streak severity.

Background

Plant-pathogenic gram-negative bacteria use the type III (T3) secretion system to secrete a large number of pathogenicity determinants called T3 effectors (Büttner 2012; Bundalovic-Torma et al. 2022). T3 effectors can play a role in the virulence of bacteria, resulting in diseases in susceptible varieties of host plants (Chen et al. 2021), and they can be an avirulent factor, inducing defense responses in nonhost plants or resistant varieties of host plants (Wang et al. 2018). To exert these distinct functions, T3 effectors must be translocated from bacteria into plant cells (Büttner 2012; Li et al. 2019; Chen et al. 2021; Hajra et al. 2021; Bergeron and Marlovits 2022). Depending on their locations in plant cells, T3 effectors are translocated via a shorter trafficking route toward the cell cytosol or a longer route toward the cell nucleus.

The longer and shorter trafficking routes apply to bacterial transcription activator-like effectors (TALEs) and non-TAL effectors (nTALEs), respectively (Deng et al. 2012; Hui et al. 2019). nTALE translocation may occur via a T3 translocon (Picking and Barta 2019), which is thought to be formed by molecular interactions of bacterial proteinic translocators with each other and with their receptors situated in plasma membranes (PMs) of plant cells (Büttner 2012; Ji and Dong 2015; Zhang et al. 2019a). Due to variations in the translocator repertoire present in different bacteria, endocytosis of effectors may occur by direct interactions with recognition compounds in plant PMs (Santi-Rocca and Blanchard 2017; Gaytán et al. 2018; Wagner et al. 2018; Shanmugam and Dalbey 2019). Thus, a translocon-dependent route or a translocon-independent route may be used to transport nTALEs into the cytosol of plant cells (Zhang et al. 2019a). In contrast, TALEs need to be first translocated into the cytosol by a translocon or via endocytosis and then delivered into the nucleus (Bogdanove and Voytas 2011; Wang et al. 2017; Hui et al. 2019) by molecular trafficking with the aid of plant importins (Wirthmueller et al. 2013; Roth et al. 2017; Zhao et al. 2019; Xiong et al. 2021). Characterizing plant importins that transport bacterial TALEs into plant cell nuclei has been an important subject in phytopathology research (Wirthmueller et al. 2015; Jo et al. 2017; Roth et al. 2017; Schäfer et al. 2017; Hui et al. 2019).

The gram-negative bacteria Xanthomonas oryzae pv. oryzae (Xoo) and X. oryzae pv. oryzicola (Xoc) are devastating bacterial pathogens of rice (Oryza sativa), causing bacterial blight (BB) and bacterial leaf streak (BLS), respectively, via virulence functions of either nTALEs known as Xanthomonas outer proteins (Xops) or TALEs (White et al. 2009). Both Xops and TALEs are translocated via bacterial T3 translocators and their receptors in plant PMs (Büttner et al. 2002; Büttner and Bonas 2002; Li et al. 2011; Ji and Dong 2015; Bian et al. 2019; Li et al. 2019; Mo et al. 2020; Chen et al. 2021). T3 translocators already identified in Xoo and Xoc include Hpa1, HrpF, and XopN (Li et al. 2011; Wang et al. 2018; Mo et al. 2020; Chen et al. 2021). XopN interacts with a receptor-like protein kinase located in plant PMs (Kim et al. 2009) and facilitates Xoo effector translocation into rice cells (Mo et al. 2020). Bacterial HrpF facilitates T3 effector translocation, possibly by interacting with PM lipids (Büttner et al. 2002; Li et al. 2011). In rice, OsPIP1;3, an aquaporin of the PM intrinsic protein (PIP) family, has been identified as the receptor for bacterial Hpa1 (Li et al. 2019; Chen et al. 2021).

In Xoo-infected rice plants, Hpa1 is secreted via the bacterial T3 secretion system and serves as a translocator of the TALE PthXo1 secreted through the same pathway (Wang et al. 2018). PthXo1 determines the virulence of the standard Xoo strain PXO99^A on the susceptible rice variety Nipponbare (Yang et al. 2006). To infect Nipponbare plants, PXO99^A secretes Hpa1 and delivers it to the plant–bacterium interface, where Hpa1 interacts with OsPIP1;3 to expedite translocation of the subsequently secreted PthXo1 (Bian et al. 2019; Li et al. 2019; Chen et al. 2021). Thereafter, PthXo1 induces virulence through activating the host susceptibility gene OsSWEET11 (Yang et al. 2006) in an OsPIP1;3-dependent manner (Bian et al. 2019; Li et al. 2019). Similarly, Hpa1–OsPIP1;3 interactions are also critical for TALI translocation to induce BLS in Nipponbare plants (Chen et al. 2021). TALI is a TALE that was identified in the Xoc hypervirulent Chinese strain GD41 and determines virulence in the susceptible rice variety Nipponbare (Long et al. 2021). Extensive studies have shown that bacterial T3 translocators and their receptors on plant PMs function cooperatively to facilitate TALE translocation into the cytosol of plant cells (Büttner et al. 2002; Li et al. 2011; Bian et al. 2019; Li et al. 2019; Mo et al. 2020; Chen et al. 2021). However, these findings also raise questions about the next step of TALE trafficking in plant cells. How do bacterial TALEs move into plant cell nuclei? Do any plant importins play a role in nuclear trafficking? Do TALE-transporting importins affect plant growth and development?

Importins present in eukaryotic cells are cargo-transporting nucleocytoplasmic trafficking proteins that serve as nuclear transport receptors or as carriers for cytoplasmic proteins that have a nuclear localization signal (NLS). Typically, importins have a cargo-binding α-subunit (IMPα) that perceives NLS-containing cytosolic proteins and a cargo-navigating β-subunit (IMPβ) that facilitates transport of cytosolic proteins into plant nuclei (Merkle 2003). In general, IMPα recognizes a short positively charged NLS, while IMPβ mediates interactions with nuclear protein complexes and RanGTP to complete cargo transport (Xiong et al. 2021). In particular, RanGTP binds to IMPs (α and β) to induce cargo release into the nucleus and then brings IMPs back to the cytosol, where IMPs prepare for the next round of nuclear trafficking (Stewart 2007; Ems-McClung et al. 2020). IMPα has three key conserved structural features (Goldfarb et al. 2004): an N-terminal IMPβ-binding (IBB) domain, a series of armadillo (Arm) repeats, and a C-terminal acidic cluster (AC) motif. The Arm repeats serve as internal cargo NLS-binding sites, whereas the AC motif functions as a binding site for the nuclear export of IMPα in conjunction with RanGTP (Miyamoto et al. 2016). These features enable IMPα to effectively function in nucleocytoplasmic trafficking in a recycling manner (Lüdke et al. 2021).

In plants, both IMPα and IMPβ exist as multiple isoforms that are often associated with responses to pathogen attack (Lüdke et al. 2021; Xiong et al. 2021) and environmental stress (Merkle 2003). In rice, there are at least 16 importins, each designated as an OsIMP (Additional file 1: Table S1), although only OsIMPα1a and OsIMPα1b have been implicated in bacterial virulence in rice (Hui et al. 2019). OsIMPα1a and OsIMPα1b have been shown to transport NLS-containing proteins (Jiang et al. 2001; Chang et al. 2012, 2014). Moreover, Hui et al. (2019) found that both OsIMPα1a and OsIMPα1b interact with PthXo1 in yeast, tobacco, and rice. They further demonstrated that, in a susceptible rice variety, silencing of OsIMPα1a or OsIMPα1b considerably enhances resistance to both Xoo and Xoc, along with marked alleviation of the disease severities of BB and BLS (Hui et al. 2019). Thus, OsIMPα1a and OsIMPα1b might facilitate bacterial virulence through mediating the nuclear trafficking of bacterial TALEs within plant cells.

To verify this hypothesis, we analyzed the effects of clustered regularly interspaced short palindromic repeats (CRISPR)–CRISPR-associated 9 (Cas9)-edited OsIMPα1a and OsIMPα1b on rice growth and bacterial virulence, and the results showed that OsIMPα1a is vital to rice, while OsIMPα1b has a crucial role in plant growth and development. We present evidence that OsIMPα1b is required for PthXo1 and TALI to move into rice cell nuclei to fulfil their virulence functions.

Results

OsIMPα1a and OsIMPα1b interact with PthXo1 through the AC motif

We initiated this study in 2015, at which point no plant importins had been determined to be related to bacterial TALEs. We retrieved the sequences of 16 OsIMP genes from rice genome annotation databases (Additional file 1: Table S1) but cloned only 10 OsIMPs from the rice variety Nipponbare (Fig. 1a). The OsIMPs were subjected to split-ubiquitin-based yeast two-hybrid (SUB-Y2H) assays together with PthXo1ΔN152 which lacks the first 152 amino acids at N-terminal (Fig. 1a). This modification was shown to prevent PthXo1 from autoactivation in SUB-Y2H assays performed previously (Yuan et al. 2016) and in the present study (Fig. 1b inset).

Fig. 1

OsIMPα1a and OsIMPα1b interact with PthXo1. a OsIMPα1a and OsIMPα1b were identified as interactors of PthXo1 in split-ubiquitin-based yeast two-hybrid (SUB-Y2H) assays of PthXo1 in combination with each of the indicated IMPs from rice. The protein‒protein interactions were detected by an X-Gal assay of colonies grown on synthetic dropout (SD)-amino acid tryptophan-leucine nutrient (SD-WL) media. b Diagram of PthXo1 structural components, including the 152-amino acid-long N-terminal sequence, which is responsible for autoactivation of the protein. The inset shows the presence and absence of colonies growing on SD-tryptophan-leucine-alanine-histidine (SD-WLAH) medium from yeast cultures that produce full-length PthXo1 protein or the PthXo1Δ152 variant lacking the first 152 amino acids. c An additional SUB-Y2H assay revealed acidic cluster (AC) motif, instead of IMPβ-binding (IBB) domain, as a recognition motif driving the interactions of OsIMPα1a and OsIMPα1b with PthXo1. In a and c, each colony represents 9 colonies investigated in 3 experimental replicates. d–f Luciferin (Luc) living fluorescence imaging (LLFI) assays performed on Nicotiana benthamiana leaves transformed with each pair of the indicated proteins. Luc^N and Luc^C represent the N-terminal and C-terminal halves of Luc fusion constructs of each tested proteins. Each leaf image represents 12 leaves of 6 plants tested in 3 experimental replicates

Our SUB-Y2H assays clearly revealed direct interactions of PthXo1ΔN152 with OsIMPα1a or OsIMPα1b (Fig. 1a) but not for multiple pairs of control proteins (Fig. 1a and Additional file 2: Figure S1a). PthXo1 has been thoroughly demonstrated to have structural features (Fig. 1b) that meet functional requirements (Boch et al. 2009; Deng et al. 2012; Mak et al. 2012), especially the C-terminal NLS that determines its interactions with OsIMPα1a and OsIMPα1b (Hui et al. 2019). The results of our SUB-Y2H assays indicated that the interactions of OsIMPα1a and OsIMPα1b with PthXo1 were determined by the AC motif present in the C-terminal region of both OsIMPα1 sequences (Fig. 1c and Additional file 2: Figure S1a). In contrast, the N-terminal IBB domain was not involved in the interactions between OsIMPα1a/OsIMPα1b and PthXo1 (Fig. 1c and Additional file 2: Figure S1a).

These results were confirmed by luciferin (Luc) living fluorescence imaging (LLFI) assays performed on Nicotiana benthamiana leaves (Fig. 1d–f and Additional file 2: Figure S1b). LLFI assay confirmed the direct interaction between OsIMPα1a and PthXo1 (Additional file 2: Figure S1b). It further revealed that PthXo1ΔN152 resembled PthXo1 in terms of its ability to bind to OsIMPα1b (Fig. 1d) and that the ability of OsIMPα1bΔIBB (OsIMPα1b lacking the IBB motif) to interact with PthXo1 was similar to that of OsIMPα1b (Fig. 1e). Conversely, OsIMPα1bΔAC (OsIMPα1b lacking the AC motif) failed to interact with PthXo1 or PthXo1ΔN152 (Fig. 1f). Taken together, these results suggest that OsIMPα1a and OsIMPα1b interact with PthXo1 in a manner dependent on the AC motif, which is important for IMPα to efficiently function in nucleocytoplasmic trafficking (Miyamoto et al. 2016; Lüdke et al. 2021).

CRISPR/Cas9-based disruption of the OsIMPα1a gene in rice plants is embryo-lethal

To determine whether OsIMPα1a and OsIMPα1b are related to bacterial TALE translocation, both OsIMPα1 genes were edited by the use of the CRISPR/Cas9-based gene editing technique. A single guide RNA (sgRNA) motif linked to the Cas9-recognizable protospacer adjacent motif (PAM) consisting of NGG was designed to target the OsIMPα1a complementary DNA strand at sites 212–234 and 315–336 in the first exon (Additional file 2: Figure S2a). The sgRNA, PAM, and NGG fusion constructs were introduced into the Nipponbare rice genome through Agrobacterium-mediated genetic transformation of embryo-derived callus with the recombinant plant binary vector pYLCRISPR/Cas9P_Ubi-H, which harbors Cas9 and the hygromycin resistance gene Hyg, in addition to sgRNA (Ma et al. 2015). In total, 41 transgenic rice lines designated #1 to #41 were obtained from independent calli. According to Sanger sequencing-based genotyping, 7 transgenic lines were identified as having nucleotide mutations, including deletions and additions. Among them, 2 mutants were heterozygous and 5 mutants were homozygous (Additional file 1: Table S2).

These transgenic lines were designated OsIMPα1a-edited (OsIMPα1a-ED) lines OsIMPα1a-ED#2, OsIMPα1a-ED#4, OsIMPα1a-ED#14, OsIMPα1a-ED#17, OsIMPα1a-ED#21, OsIMPα1a-ED#32, and OsIMPα1a-ED#39 in accordance with the original codes of the independent T0 callus lines (Additional file 1: Table S2). The two homozygous T0 callus lines (OsIMPα1a-ED#2 and OsIMPα1a-ED#4) were lethal, as the embryos ultimately failed to produce progeny. T1 progeny of the three heterozygous T0 callus lines (OsIMPα1a-ED#14, OsIMPα1a-ED#21, and OsIMPα1a-ED#39) were further investigated and compared with the wild-type (WT) plants. We found that all the OsIMPα1a-ED#14, OsIMPα1a-ED#21, and OsIMPα1a-ED#39 lines were inferior to the WT plants in terms of seed germination and seedling growth (Additional file 2: Figure S2b, c). Although the WT seedlings grew well, the T1 embryos of the OsIMPα1a-ED#14, OsIMPα1a-ED#21, and OsIMPα1a-ED#39 seedlings did not survive during germination (Additional file 2: Figure S2c). Indeed, the T1 seedlings of these three lines died quickly after cotyledon stratification from the soil (Additional file 2: Figure S2c). Clearly, OsIMPα1a is critical for plant viability.

OsIMPα1b is an integral component of plant growth and development

The Nipponbare rice OsIMPα1b gene was also edited with the method similar to that used for editing of OsIMPα1a. A sgRNA motif linked to the PAM was designed to target the OsIMPα1b complementary genomic DNA strand at sites 203–225 and 505–527 in the first and second exons, respectively (Fig. 2a). The sgRNA-PAM fusion construct and the Cas9 and Hyg genes were concomitantly introduced into the Nipponbare genome with the aid of a plant binary vector based on T-DNA construction (Ma et al. 2015). In total, 45 transgenic rice lines (#1 to #45) were obtained from independent calli. Based on genotyping by Sanger sequencing, five independent T0 mutant lines (OsIMPα1b-ED#6, OsIMPα1b-ED#17, OsIMPα1b-ED#26, OsIMPα1b-ED#29, and OsIMPα1b-ED#42) were identified as having nucleotide changes, including deletions and additions (Fig. 2b and Additional file 1: Table S2). All the five mutant lines were heterogeneous and had no random insertions or deletions. Despite evidently impaired seed germination (Additional file 2: Figure S2b), all these mutant lines were able to produce progeny (Fig. 3 and Additional file 2: Figures S2–S4).

Fig. 2

CRISPR/Cas9-based OsIMPα1b gene editing and removal of transgenes by genetic segregation. a OsIMPα1b structural features along with the Cas9-targeting sites based on the T-DNA construction of an sgRNA motif linked to the Cas9-recognizable PAM consisting of NGG (CGG here). b Nucleotide changes in the targeting sequences of the OsIMPα1b-edited (OsIMPα1b-ED) T0 lines OsIMPα1b-ED#6 and OsIMPα1b-ED#17, both of which were further investigated. c Partial fragment sequences of OsIMPα1b in WT plants (rice variety Nipponbare) and T1 progenies of the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutants. d Genetic segregation in which the transgenes were removed during the generation of T1 and T2 progenies of both mutants tested in comparison with the WT plants. The transgenes were analyzed via PCR using genomic DNA extracted from the corresponding plant templates, and the housekeeping gene OsActin was used as a reference. Each electrophoretic agarose image is representative of 3 experimental replicates

Fig. 3

Imaging and quantitative data showing growth and development differences between the WT plants and OsIMPα1b-edited (OsIMPα1b-ED) lines. a Plants in a growth chamber and in a specific pond within the experimental field. The image represents plant populations in approximately 1/15 hectare (1 Mu or 667 m²) for each genotype. b Plant height and fresh weight of 56-day-old and 90-day-old plants and tiller number at the heading stage. The data shown are mean values ± standard deviations (SD) (n = 3). c Three spikelets in a row excised from different rice genotypes and representing grain production in approximately 1/15 hectare (1 Mu or 667 m²) for each genotype. d Grain weight empirically determined in three experimental ponds and grain yield surveyed in 1/15 hectare (1 Mu or 667 m²) divided into the three experimental ponds. The data shown are mean values ± standard deviations (SD) (n = 3). In b and d, different letters on bar graphs indicate significant differences among different plants based on Duncan’s new multiple range test (P < 0.05)

T1 progenies of the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutant lines were genotyped to identify the individuals that segregated away from the T-DNA (Fig. 2c). Plant DNA amplification by PCR in which primers specific to Cas9, sgRNA, or Hyg were used to identify T1 individuals that lack the transgene owing to genetic segregation (Fig. 2d). Moreover, Sanger sequencing revealed that some plants (OsIMPα1b-ED#6 and OsIMPα1b-ED#17) were homozygous (T2 progenies similar to their T1 counterparts) (Fig. 2d), which were used for further studies.

We determined that OsIMPα1b played a critical role in rice growth and development. Both the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutant lines were much inferior to the WT plants in terms of embryo development, as indicated by impaired seed germination and seedling growth (Additional file 2: Figure S2b). These two mutants were also inferior to the WT plants in terms of vegetative growth and grain production (Fig. 3a and Additional file 2: Figure S3). In contrast to the WT, OsIMPα1b-ED#6 and OsIMPα1b-ED#17 presented considerable reductions in all quantitative growth parameters, such as plant height, plant fresh weight, and tiller number (Fig. 3b). Similarly, grain production was also impaired in both mutants (Fig. 3c and Additional file 2: Figure S4). In the end, grain yield was substantially decreased in both OsIMPα1b-ED#6 and OsIMPα1b-ED#17 compared with that of the WT plants (Fig. 3d). Clearly, OsIMPα1b gene disruption impaired rice growth, development, and grain production, suggesting that OsIMPα1b is an integral component of plant growth and development.

CRISPR/Cas9-based disruption of the OsIMPα1b gene results in the sequestration of TALEs from rice cell nuclei

PthXo1 and TALI produced by Xoo and Xoc, respectively, share highly conserved NLS (Fig. 4a), which determines TALE interactions with OsIMPα1a or OsIMPα1b (Szurek et al. 2001; Hui et al. 2019). Consistently, TALI was similar to PthXo1 in terms of the ability to bind to OsIMPα1a (Additional file 2: Figures S1b, S5) or OsIMPα1b (Figs. 1d, 4b), suggesting that the nuclear import of both TALEs may be subject to similar regulation by OsIMPα1b. To infer whether OsIMPα1b is related to the nuclear import of PthXo1 in the susceptible rice variety Nipponbare, PthXo1-RFP and OsIMPα1b-CFP fusion constructs were produced independently and concomitantly in Nipponbare protoplasts by transformation with the individual genes or both fusion constructs. Using laser scanning confocal microscopy (LSCM), we found the independently produced PthXo1-RFP and OsIMPα1b-CFP in the nucleus and cytoplasm, respectively (Additional file 2: Figure S6). However, when PthXo1-RFP and OsIMPα1b-CFP were expressed together, both proteins were colocalized to the nucleus (Additional file 2: Figure S6). Similarly, TALI-RFP and OsIMPα1b-CFP were localized to the nucleus and cytoplasm, respectively, when they were produced independently, but were colocalized to the nucleus when coexpressed (Additional file 2: Figure S7).

Fig. 4

Alignment of PthXo1 and TALI C-terminal sequences and luciferin (Luc) living fluorescence imaging (LLFI) assay of rice OsIMPα1a and Xoc TALI expressed in tobacco. a Sequence alignment highlighting the conserved consensus nuclear localization signals (NLSs) (taller orange letters) and highlighting differences in residues (in red). The conserved consensus NLSs were identified using the Clustal Omega program (https://www.ebi.ac.uk/Tools/msa/clustalo/). b LLFI assay representing 12 leaves of 6 tobacco plants

This hypothesis was verified by determining the subcellular localization of PthXo1 and TALI in the protoplasts of the WT and OsIMPα1b-ED#6 plants. Based on LSCM, the PthXo1-RFP fusion protein was localized to the nucleus of WT but in the cytoplasm of OsIMPα1b-ED#6 (Fig. 5a). Immunoblotting (IB) analyses revealed that the PathXo1-HA fusion protein was abundant in the nuclear fraction isolated from the WT plant protoplasts but was present at low levels in the nucleus of OsIMPα1b-ED#6 (Fig. 5b). Similarly, the TALI-RFP fusion protein was localized to the nucleus of WT but in the cytoplasm of OsIMPα1b-ED#6 (Fig. 5c). IB analyses revealed that the TALI-HA fusion protein was abundant in the nuclear fraction isolated from the WT plant protoplasts but was present at low levels in the nucleus of OsIMPα1b-ED#6 (Fig. 5d). In all the plants, neither PthXo1-HA nor TALI-HA was detected in the PM fractions (Fig. 5b, d). These results indicate that disruption of OsIMPα1b resulted in the sequestration of PthXo1 and TALI away from rice cell nuclei, suggesting that OsIMPα1b is required for both TALEs to be translocated into the nuclei of rice cells.

Fig. 5

Protoplast fluorescence images and Immunoblotting (IB) analyses showing differences in subcellular localization of the TALEs between the WT (Nipponbare) and OsIMPα1b-ED#6 plants. a Laser scanning confocal microscopy (LSCM) images of the plant protoplasts at 16 h after transformation with the pthXo1-RFP fusion gene. Each image is representative of 200–300 protoplasts. b IB of proteins extracted from protoplasts equivalent to those in a. c LSCM images of the plant protoplasts at 16 h after transformation with the TALI-RFP fusion gene. In a and c, each image is representative of 200–300 protoplasts; scale bars = 10 µm. d IB of proteins extracted from protoplasts equivalent to those in c. The asterisks and vertical lines indicate unidentified cross-reactive proteins

CRISPR/Cas9-based disruption of the OsIMPα1b gene suppresses Xoo virulence and promotes rice resistance to Xoc

In Nipponbare rice plants infected with the Xoo strain PXO99^A, the bacterial TALE PthXo1 induces virulence by activating its target gene OsSWEET11 (Li et al. 2019; Chen et al. 2021) that determines host susceptibility and causes disease (Yang et al. 2006). Therefore, we hypothesized that the translocation of PthXo1 into rice cell nuclei is a prerequisite for OsSWEET11 expression and bacterial virulence. In good agreement with this hypothesis, a considerable amount of PthXo1 was translocated from Xoo (PXO99^A) bacteria into the nuclei of rice cells (Fig. 6a), which is consistent with the strong expression of OsSWEET11 (Fig. 6b) in the WT plants. In this case, Xoo bacteria multiplied sufficiently, attaining high population numbers within three days in rice leaves (Fig. 6c). This resulted in the full development of BB in the following 10 days, manifested as a long leaf blight lesion (Fig. 6d) and severe symptoms (Fig. 6e). In contrast to the WT plants, the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutants coincidently presented greater reductions in PthXo1 translocation (Fig. 6a), OsSWEET11 expression levels (Fig. 6b), and in planta bacterial populations (Fig. 6c). In the end, leaf blight lesion length (Fig. 6d) and symptoms (Fig. 6e) were considerably less severe in both mutants compared with the WT plants. Based on these results and those of the analyses described above, we propose that disruption of OsIMPα1b resulted in the suppression of Xoo virulence through sequestration of PthXo1 from cell nuclei of rice plants.

Fig. 6

Effects of PthXo1 translocation on Xoo virulence. a–d Comparisons of the Xoo virulence on the WT plants and OsIMPα1b-edited (OsIMPα1b-ED) mutants. PthXo1 translocation at 9 h post-inoculation (hpi) (a), expression of the PthXo1-targeted OsSWEET11 gene at 24 hpi (b), in planta bacterial populations at 0 and 72 hpi (c), and disease severity at 14 days post-inoculation (dpi) (d) were determined correspondingly. The data shown are mean values ± SDs (n = 3). The P values are from two-sided unpaired Student’s t tests (P < 0.01). e Fragments of Xoo-inoculated leaves imaged at 10 dpi. The leaf images are representative of 30 leaves of 15 plants tested in 3 experimental replications per genotype (WT, OsIMPα1b-ED#6, or OsIMPα1b-ED#17)

In Nipponbare plants infected with the Xoc strain GD41, the bacterial effector TALI induces virulence by suppressing plant defense responses, which essentially involve the expression of defense response genes (Chen et al. 2021; Long et al. 2021). Therefore, we hypothesized that the translocation of TALI into rice cell nuclei is a prerequisite for suppressing the expression of genes involved in the defense response and for rice resistance to BLS. In support of this hypothesis, substantial transport of TALI from Xoc (GD41) bacterial cells into rice cell nuclei (Fig. 7a) concurrent with weak expression of the rice defense response genes PR1a, PR5, PR10, and GSL12 (Fig. 7b) was detected in the WT plants. Consistently, Xoc bacteria sufficiently multiplied, resulting in a large bacterial population within three days in rice leaves (Fig. 7c). In the subsequent 7 days, severe BLS signs were observed on leaves of the WT plants (Fig. 7d). In contrast to the WT plants, the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutants presented considerably less TALI translocation (Fig. 7a) and smaller in planta bacterial populations (Fig. 7c). In addition, in contrast to the WT plants, the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutants exhibited greatly increased expression levels of PR1a, PR5, PR10, and GSL12 (Fig. 7b). Ultimately, BLS symptom severity and lesion length were considerably alleviated in both the mutants compared with the WT plants (Fig. 7d). Evidently, disruption of OsIMPα1b resulted in the activation rather than suppression of rice defense responses to inhibit Xoc virulence.

Fig. 7

Effects of TALI translocation into the nuclei of rice cells on the expression of genes involved in the rice defense response and Xoc virulence. TALI localization to rice cell nuclei at 9 h post-inoculation (hpi) (a), expression of rice defense response genes at 24 hpi (b), in planta bacterial populations at 0 and 72 hpi (c), and disease severity at 10 days post-inoculation (dpi) (d) were determined. The data shown are mean values ± SDs (n = 3). The P values are from two-sided unpaired Student’s t tests (P < 0.01). In d, the leaf images are representative of 30 leaves of 15 plants tested in 3 experimental replications per genotype (WT, OsIMPα1b-ED#6, or OsIMPα1b-ED#17)

OsIMPα1b disruption results in the promotion of pattern-triggered immunity (PTI)

As one of the OsIMPα1b-inhibited defense response genes (Fig. 7b), GSL12 encodes the β-glucan synthetase-like (GSL) protein GSL12, which contributes to callose deposition on plant surfaces (Fu et al. 2014; Malinovsky et al. 2014; Zhai et al. 2017). Callose deposition is a conserved defense response against pathogen, insect attack, and other biotic stresses (Fu et al. 2014; Malinovsky et al. 2014; Zhai et al. 2017). In particular, GSL12 expression and callose deposition also occur as downstream responses of the PTI pathway (Malinovsky et al. 2014; Chen et al. 2021). Based on these results, we hypothesized that OsIMPα1b might be related to PTI, the innate immunity activated by pathogen-associated molecular patterns (PAMPs) such as bacterial flagellin (Zipfel et al. 2004).

In support of this hypothesis, after editing the OsIMPα1b gene, we determined that critical PTI responses were promoted considerably (Fig. 8). The OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutant plants greatly exceeded the WT plants in terms of the extent of callose deposition on the surface of their leaves following exogenous application of flg22 (Fig. 8a), a PAMP functional fragment (Asai et al. 2002). Both mutants were more effective than the WT in supporting the induction effect of flg22 on the expression of the PTI regulatory genes MPK3 and MPK6 (Fig. 8b). Following flg22 application, moreover, several PTI response genes (WRKY70, NHL10, PHI1-1, and GSL12) were expressed at higher levels in both the OsIMPα1b-ED#6 and OsIMPα1b-ED#17 mutants than in the WT plants (Fig. 8b). Evidently, OsIMPα1b gene editing resulted in the promotion of the PTI pathway.

Fig. 8

Essential pattern-triggered immunity (PTI) responses occurred in leaves of the WT and OsIMPα1b-edited plants treated by spraying the tops of the plants with sterile water (control) or an aqueous solution of flg22. a Callose deposition on leaf surfaces of the different plants at 24 h after treatment. Each leaf image is representative of 15 leaves of 3 plants tested in 3 experimental replications. b Flg22-induced fold increases in foliar expression of rice PTI regulatory genes (MPK3 and MPK6) and rice PTI response genes (WRKY70, NHL10, PHI1-1, and GSL12). The expression levels of the PTI regulatory and response genes were determined at 1 and 24 h after treatment. The data shown are mean values ± SDs (n = 3). The different letters on bar graphs indicate significant differences among different plants based on Duncan’s new multiple range test (P < 0.01)

Discussion

This study was conceived to bridge the pivotal research gap concerning the in planta nucleocytoplasmic linkage found in the process of bacterial T3 effector translocation (Zhang et al. 2019a). We chose to test the specific hypothesis that OsIMPα1a and OsIMPα1b might facilitate bacterial virulence by mediating the nuclear trafficking of bacterial TALEs within plant cells (Hui et al. 2019). The importance of verifying this hypothesis has been increasingly highlighted by recent studies on bacterial T3 effector translocation (Ji and Dong 2015; Zhang et al. 2019a; Chen et al. 2021). It has been demonstrated that T3 effectors, including TALEs and nTALEs, secreted by plant pathogenic bacteria are translocated into plant cells to execute their virulence or avirulence functions, which depends on plant species and variety (Wang et al. 2018; Bian et al. 2019; Li et al. 2019; Mo et al. 2020; Chen et al. 2021). Both TALEs and nTALEs are translocated by bacterial T3 translocators (Ji and Dong 2015; Zhang et al. 2019a, b) and their receptors in the PMs of host plants (Wang et al. 2018; Bian et al. 2019; Li et al. 2019; Mo et al. 2020; Chen et al. 2021). Nevertheless, how TALEs further move into the nucleus has been not known until the present study.

In this study, we investigated the rice importins OsIMPα1a and OsIMPα1b, and three major results were discovered. First, both OsIMPα1a and OsIMPα1b interact with the TALEs PthXo1 (Fig. 1 and Additional file 2: Figure S1) and TALI (Fig. 4b and Additional file 2: Figure S5) from Xoo and Xoc, respectively, on the basis of the identical NLSs present in both TALEs (Fig. 4a). We found that OsIMPα1b interacts with PthXo1 via the AC motif present in the C-terminal region (Fig. 1). Similarly, a recent study showed that PthXo1 depends on its NLS domain to interact with OsIMPα1a or OsIMPα1b and that both IMPs affect bacterial virulence (Hui et al. 2019). Nevertheless, whether OsIMPα1a and OsIMPα1b also affect bacterial effector translocation into rice cell nuclei were not determined. This possibility was found with respect to OsIMPα1–TALE interactions, for which there is a solid molecular basis concerning the interacting motifs (Fig. 1), and helps establish a molecular basis for bacterial effectors trafficking into rice cell nuclei to fulfil their virulence functions.

Second, we determined that OsIMPα1b is an integral component of rice growth and development and contributes considerably to rice grain production (Fig. 3 and Additional file 2: Figures S2–S4). Unfortunately, we have to reject further studies about OsIMPα1a owing to the embryo lethality of the OsIMPα1a-ED mutants (Additional file 2: Figure S2). At present, we do not know if and how the innate OsIMPα1a gene affects rice growth and development. We also do not know whether OsIMPα1a is related to bacterial virulence and effector translocation. These questions remain on hold until the embryo lethality of OsIMPα1a-ED mutants can be rescued.

Third, we presented evidence that, in rice, OsIMPα1b facilitates the translocation of the bacterial TALEs PthXo1 and TALI into the cell nucleus (Fig. 5 and Additional file 2: Figures S6, S7). Following nuclear import, both effectors execute their virulence functions (Chen et al. 2021; Long et al. 2021), possibly via different mechanisms, which were partially elucidated in the present study. In Xoo-infected OsIMPα1b-edited rice plants, OsIMPα1b sequesters PthXo1 from the nucleus, the PthXo1-targeted host susceptibility rice gene (OsSWEET11) is no longer activated, and bacterial virulence and severity of BB disease are alleviated (Fig. 6). In Xoc-infected OsIMPα1b-edited rice plants, OsIMPα1b sequesters TALI from the nucleus, the function of TALI in suppressing rice defense responses (Long et al. 2021) is nullified, and rice defense responses are in turn activated to inhibit bacterial virulence and attenuate BLS severity (Fig. 7). These findings suggest that OsIMPα1b is a physiologically and pathologically relevant nuclear transport receptor with dual effects on plant growth and development and bacterial effector translocation.

In rice, the pathological function of OsIMPα1b affects at least two immunity pathways associated with the Xoc effector TALI (Figs. 7, 8). TALI has been shown to confer virulence to its host by suppressing systemic acquired resistance (SAR) and the expression of genes that are involved in the defense response and are characteristic of the SAR pathway in the susceptible rice variety Nipponbare (Long et al. 2021). As the defense response genes tested in the present study represent different immunity pathways, the expression of these genes indicated the activation of the corresponding immunity pathways. PR1a, PR5, and PR10 encode the corresponding pathogenesis-related (PR) proteins that are associated with SAR (Boatwright and Pajerowska-Mukhtar 2013). Therefore, the OsIMPα1b-dependent PR1a, PR5, and PR10 expression indicated that OsIMPα1b is related to the SAR pathway (Fig. 7b). Moreover, OsIMPα1b plays an important role in the PTI pathway (Fig. 8). Evidently, OsIMPα1b is associated with at least two cellular pathways that regulate innate immunity in rice. Presumably, some transcription factors involved in the innate immunity pathways are modulated by OsIMPα1b such that defense responses are regulated at the transcriptional level. Overall, characterizing the molecular mechanisms assumed to govern the functions of OsIMPα1b in the SAR and PTI pathways will be the subjects of future studies.

Conclusions

We have presented evidence demonstrating that OsIMPα1b is a physiologically and pathologically relevant nuclear transport receptor that has dual effects on plant growth and bacterial virulence. Rice OsIMPα1b interacting with the bacterial effectors PthXo1 and TALI represents a molecular basis on which plant importins transport bacterial effectors (TALEs) into the cell nucleus. This notion was confirmed by the LSCM and IB results showing that TALEs colocalized with OsIMPα1b in the nuclei of the WT plants, in contrast to the distribution of TALEs throughout the cytoplasm of the OsIMPα1b loss-of-function mutants. In Xoo-infected OsIMPα1b-edited rice plants, OsIMPα1b sequesters PthXo1 from the nucleus, the PthXo1 target gene OsSWEET11 is no longer activated, and bacterial virulence and BB disease symptoms are ultimately alleviated. In Xoc-infected OsIMPα1b-edited rice plants, OsIMPα1b sequesters TALI from the nucleus, rice defense responses are no longer suppressed, and rice defense responses are therefore activated to inhibit bacterial virulence and reduce BLS severity.

Methods

Microbial materials and culture conditions

The Xoo strain PXO99^A and Xoc strain GD41 were cultured on nutrient broth (NB) or nutrient agar (NA) medium (Li et al. 2011). The Agrobacterium tumefaciens strain GV3101 was used in plant transformations and cultured on NB medium supplemented with kanamycin of 50 μg/mL and rifampicin of 20 μg/mL. The yeast (Saccharomyces cerevisiae) strain AH109 was used in SUB-Y2H (Dualsystems) assays and grown on synthetic dropout (SD)-amino acid nutrient medium. These microbes were cultured in incubators at 28 °C.

Plant materials and growth conditions

Seeds of rice (O. sativa) and tobacco (N. benthamiana) were maintained in our lab. Tobacco plants were grown in a greenhouse under regular conditions and used for LLFI assays. Rice seeds were germinated in flat plastic trays filled with the PINDSTRUP substrate (Pindstrup group). Three days later, the germinating seedlings were moved into 12-L pots (2–5 plants per pot) filled with the PINDSTRUP substrate. Seeds were germinated and plants were grown in environment-controlled chambers at 26 ± 1 °C, 12-h light at 250 ± 50 μmol quanta/(m².s) and a relative humidity of 85% except specified elsewhere. These rice plants were used in bacterial infection assessments and protoplast preparation. For growth and development assessments, nursery seedlings grown for 30 days in the plant growth chamber were transplanted into rice-growth ponds located at Panhe Campus of Shandong Agricultural University, Taian, Shandong Province, China.

Molecular interaction assays

SUB-Y2H (Li et al. 2015) and LLFI (Zhou et al. 2018) assays were performed as previously described.

Plant gene editing and mutant generation

Editing of rice importin genes (OsIMPα1a and OsIMPα1b) and subsequent generation of the OsIMPα1a or OsIMPα1b loss-of-function rice (Nipponbare) mutants were conducted by using the protocols described previously (Kuang et al. 2020). The pYLCRISPR/Cas9P_Ubi-H binary vector was constructed as previously described (Ma et al. 2015). In brief, the nucleotide oligos (Additional file 1: Table S3) corresponding to the identified mutation regions of both OsIMPα1s were chosen and incorporated into the sgRNA expression vector (pYLgRNA). The resulting sgRNA expression cassettes were then shuttled into pYLCRISPR/Cas9P_Ubi-H through Gateway recombination (Ren et al. 2018). These constructs were introduced into Nipponbare genome by the conventional Agrobacterium-mediated transformation procedure. Sanger sequencing with gene-specific sequencing primers were performed to identify the editing-caused mutations and the homozygosity of the mutation (nucleotide change). T1 and T2 progenies of the mutant lines OsIMPα1b-ED#6 and OsIMPα1b-ED#17 were genotyped to identify the individuals in which the T-DNA was segregated out, and these plants were used in further studies.

Plant growth and development assessments

Rice growth and development were investigated in small-scale field trials conducted in rice-growing ponds. In China, the term ‘Mu’ has been conventionally used to specify the area of cultivated land, and 1 Mu is equivalent to 666.7 m² or 1/15 hectare. In Taian, a single field trial occupied 1/10 hectare (1.5 Mu) and was divided equally into three experimental plots. The WT, OsIMPα1b-ED#6, and OsIMPα1b-ED#17 plants were planted randomly into three plots. Rice growth was monitored. Plant weight and grain yield were surveyed.

Subcellular localization of proteins

Rice protoplasts were prepared from aerial parts of 20-day-old rice seedlings using a previously described protocol (Yoo et al. 2007). Protoplasts were transformed with OsIMPα1b-CFP alone or in combination with pthXo1-RFP and TALI-RFP, respectively, using the polyethyleneglyco-mediated method (Zhang et al. 2011). Approximate 16 h later, transformed protocols were observed under a laser scanning confocal microscope (Zeiss LSM880). Excitation laser wavelengths of 470 nm and 563 nm were used to capture fluorescence signals of CFP and RFP, respectively. To carry out IB, proteins were isolated from the rice protoplasts at 16 h after transformation. Proteins present in the nucleus and associated with PMs were isolated using the Minute Plasma Membrane Protein Isolation Kit for Plants (Invent). Nuclear and membrane proteins were quantified using a BCA protein assay kit (TransGen Biotech) and then separated by electrophoresis using 10% SDS-PAGE gels. The target proteins were detected by hybridization with the corresponding antibodies, with rat-produced α-HA (TransGen Biotech) at a 5000-fold dilution, both rabbit-produced α-Histone H3 (PhytoAB Inc) and α-Na⁺/K⁺-ATPase (PhytoAB Inc) at a 2000-fold dilution, and secondary peroxidase-conjugated anti-mouse or anti-rabbit antibody (TransGen Biotech) at a 5000-fold dilution.

Bacterial infection assessments

Overnight bacterial cultures of Xoo and Xoc incubated on NA medium were washed twice and resuspended in sterile water to generate inoculum suspensions with an optical density of OD₆₀₀ = 0.3. The Xoo suspension was amended with 0.03% v/v Silwet L-77, and the mixture was inoculated on leaves of 30-day-old rice plants by the leaf-clipping method (Kauffman et al. 1973). The Xoc suspension was directly inoculated on leaves of 2-week-old rice seedlings by infiltration with needleless syringes (Reimers et al. 1992) approximately at the central site. Infection degrees were evaluated by determining populations of bacteria multiplied in leaves and quantifying the subsequently formed symptoms. Bacteria multiplied in leaves were recovered, incubated on NA medium, quantified as colony formation unit (CFU), and presented as logarithmic CFU values per leaf. BB and BLS symptoms were scored by photographing or measuring lesion lengths and sizes in leaves at 14 and 10 dpi, respectively.

Gene expression analysis

Total RNA was isolated at 24 h after treatment with an RNAiso Plus reagent (Takara). RNA quantity and quality were evaluated using NanoDrop 8000 (Thermo Scientific). cDNA was prepared from equal amounts of total RNA per sample with the HiScript® II 1st Strand cDNA Synthesis Kit (Vazyme). Reverse transcription-quantitative PCR (RT-qPCR) experiments were conducted on ABI QuantStudio 3 (Thermo Fisher) using ChamQ™ Universal SYBR® qPCR Master Mix (Vazyme). All the RT-qPCR experiments were performed using specific primers (Additional file 1: Table S3) with the constitutively expressed OsActin gene as a reference (Chen et al. 2021). Relative expression level of a gene tested was quantified as the transcript quantity ratio to OsActin.

Callose deposition assay

Leaves of 2-week-old rice seedlings were sprayed with flg22 (GeneScript) or sterile H₂O containing 0.03% v/v Silwet L-77. Callose deposition was detected at 24 h after treatment as described previously (Yang et al. 2019).

Statistical analysis

Significance in differences between (among) data from different plants (genotypes) or treatments were estimated by analysis of variance and student’s t-tests or Duncan’s new multiple range tests using GraphPad Prism 8.0 (https://www.graphpad.com/).