The Hemileia vastatrix effector HvEC‐016 suppresses bacterial blight symptoms in coffee genotypes with the SH1 rust resistance gene

Summary A number of genes that confer resistance to coffee leaf rust (SH1–SH9) have been identified within the genus Coffea, but despite many years of research on this pathosystem, the complementary avirulence genes of Hemileia vastatrix have not been reported. After identification of H. vastatrix effector candidate genes (HvECs) expressed at different stages of its lifecycle, we established an assay to characterize HvEC proteins by delivering them into coffee cells via the type‐three secretion system (T3SS) of Pseudomonas syringae pv. garcae (Psgc). Employing a calmodulin‐dependent adenylate cyclase assay, we demonstrate that Psgc recognizes a heterologous P. syringae T3SS secretion signal which enables us to translocate HvECs into the cytoplasm of coffee cells. Using this Psgc‐adapted effector detector vector (EDV) system, we found that HvEC‐016 suppresses the growth of Psgc on coffee genotypes with the SH1 resistance gene. Suppression of bacterial blight symptoms in SH1 plants was associated with reduced bacterial multiplication. By contrast, HvEC‐016 enhanced bacterial multiplication in SH1‐lacking plants. Our findings suggest that HvEC‐016 may be recognized by the plant immune system in a SH1‐dependent manner. Thus, our experimental approach is an effective tool for the characterization of effector/avirulence proteins of this important pathogen.

Methods S2 HEVA-Sanger library construction and sequence analysis.

Methods S3 CAHV-Sanger library construction and sequence analysis.
Methods S4 CAHV-454 library construction and sequence analysis.
Methods S11 Plant material and assays.
Notes S1 Sequencing statistics for cDNA libraries from Hemileia vastatrix prebiotrophic and biotrophic phases of the infection cycle.
Notes S2 Fungal origin and expression of candidate Hemileia vastatrix effectors.

Notes S3 Number of introns of a select group of HvEC genes.
Notes S4 Alignment of genomic and cDNA nucleotide sequences of candidate Hemileia vastatrix effector genes.
Notes S5 Coffee genotypes carrying S H 1 used to determine the effect of HvEC-016 on multiplication of Pseudomonas syringae pv. garcae 1202.   Methods S1 Urediniospores germination.
Single pustule H. vastatrix isolate Hv-01, previously characterized as race II in our laboratory, was multiplied on leaves of C. arabica var. Catuaí Vermelho IAC 44 (carrying S H 5). Two grams of fresh urediniospores were washed in sterile distilled water containing 0.05% Tween 80 (Sigma-Aldrich, St Louis, MO, USA). Urediniospores were homogeneously distributed in polystyrene Petri dishes containing a film of sterile distilled water and incubated at 22°C under darkness. Germination was confirmed by light microscopy at 16 hours of incubation. About 80% of urediniospores produced long germ tubes (Fig. S1). Germinated spores were collected, frozen in liquid nitrogen, and stored at -80°C until total RNA extraction.
Methods S2 HEVA-Sanger library construction and sequence analysis.
Total RNA was extracted from germinated urediniospores according to Bilgin et al. (2009). mRNA was purified from total RNA with the NucleoTrap® mRNA Midi Kit (Macherey-Nagel, Düren, Germany), according to the manufacturer's instructions. A unidirectional cDNA library was made using the Creator TM SMART TM cDNA library construction kit (Clontech, Mountain View, CA, USA) using 0.5 µg de polyA RNA. cDNA was digested with SfiI (New England Biolabs, Beverly, MA, USA) and size-fractionated by electrophoresis. Digested fragments larger than 500 bp were ligated into pDNR-LIB (Clontech) and transformed into Escherichia coli EC-100 (Epicentre, Madison, WI, USA). A total of 7,200 recombinant clones were sequenced from the 5' end using the M13 universal primer. Sequencing reactions were carried out using the DYEnamicTM ET Dye Terminators kit (GE Healthcare, Pittsburgh, PA, USA), and readings were performed in an automatic MegaBACE 1000 Sequencing System (GE Healthcare). EST sequences with a Phred (Ewing & Green, 1998) score higher than 20 and longer than 70 nucleotides were retained for further analysis. Quality-trimmed sequences were assembled with the CAP3 assembler (Huang & Madan, 1999). A customized Perl script was used to predict ORFs.

Methods S3 CAHV-Sanger library construction and sequence analysis.
Hemileia vastatrix inoculum was prepared as previously described (Maia et al., 2013). Sixmonth-old Catuaí Vermelho IAC 44 plants were inoculated with fresh H. vastatrix Hv-01 urediniospores and samples of infected leaves were collected at 12 dpi. Unidirectional cDNA library construction and sequence analysis were performed essentially as described for the HEVA-Sanger library. A total of 9,828 recombinant clones were sequenced and 4,305 unigenes were obtained. Unigenes were compared with 254,546 ESTs of Coffea spp. deposited in nonredundant NCBI databases (www.ncbi.nlm.nih.gov/genbank) using BlastN (Altschul et al., 1997). Sequences with no significant homology (E-value < 10-5) with Coffea spp. ESTs were predicted their ORFs using a customized Perl script and included in further analysis.
Methods S4 CAHV-454 library construction and sequence analysis.
Total RNA was isolated from three biological replicates of Catuaí leaves infected with H.
vastatrix Hv-01 at 48 and 72 hpi, and 9 dpi according to Bilgin et al. (2009). An equimolar mixture of RNA from all samples was precipitated and sent to Eurofins MWG Operon (Huntsville, AL, USA) for construction of a normalized random-primed cDNA library and pyrosequencing using the 454 GS-FLX Titanium platform (Roche, Branford, CT, USA). A total of 629,890 reads were grouped into 43,763 contigs using the MIRA assembler v3.4 (Chevreux et al., 2004). Singlets were not included in further analysis. The plant or fungal origin of the contigs was determined according to the procedures described for the CAHV-Sanger library. Six-frame ORF predictions were conducted using a customized Perl script.

Methods S5 Secretome prediction.
A bioinformatics pipeline was designed in order to predict secreted proteins of unknown function, rich in cysteine, and specific to H. vastatrix. Predicted amino acid sequences from each library (HEVA-Sanger, CAHV-454 and CAHV-Sanger) were initially subjected to SignalP v4.0 (Petersen et al., 2011) and then to TargetP v.1.1 (Emanuelsson et al., 2000) in order to select those carrying a signal peptide and lacking a mitochondrial targeting peptide. A customized Perl script was used to remove the signal peptide and the sequences corresponding to the mature peptide were subjected to TMHMM v2.0 (Krogh et al., 2001) to filter out proteins containing transmembrane domains. Then, predicted secreted proteins from each library were queried against the nr-NCBI protein database using BlastP (Altschul et al., 1997) and those showing no significant similarity (E-value higher than 10 -5 ) were retained. Candidates containing full-length sequences were subjected to WolfPsort (Horton et al., 2007) and only sequences with extracellular and/or nuclear localization predictions were selected. Predicted proteins from CAHV-454 and CAHV-Sanger libraries were compared with each other using BlastP and those showing 100% identity were reduced to a single one. In addition, predicted proteins from CAHV-Sanger and CAHV-454 libraries showing significant similarity (E-value < 10 -10 ) with those from HEVA-Sanger were removed from the final set. The number of cysteine residues in each predicted mature peptide was determined using a customized Perl script. Significant Pfam domains (E-values lower than 10 -5 ) were scanned using the PFAM search server (Finn et al., 2010). ORF sequences from the final set of predicted effectors were compared with sequences previously reported (Fernandez et al., 2012;Cristancho et al., 2014;Talhinhas et al., 2014) using BlastN (Altschul et al., 1997).

Methods S7 Identification of HvECs introns.
Genomic DNA from H. vastatrix Hv-01 was extracted from fresh urediniospores according to Maia et al. (2013). Forward and reverse HvEC-YST primers (Table S5) were used to amplify HvECs by PCR, and amplicons cloned into pGEM-T Easy (Promega, Madison, WI, USA) and sequenced using universal M13 primers. Intron-exon junctions were identified by aligning the genomic DNA and cDNA sequences using ClustalW (Thompson et al., 1994).

Methods S8 Nonquantitative RT-PCR.
Total RNA was isolated from dormant and germinated urediniospores, and from non-inoculated and inoculated coffee leaves at 24 hpi and 12 dpi (Bilgin et al., 2009). RNA was treated with RNase-free DNase I (Qiagen, Valencia, CA, USA) and purified using RNeasy CleanUp Kit (Qiagen). Ten micrograms of total RNA were used to synthesize cDNA using the Oligo(dT)12-18 primer with the SuperScript First-Strand Synthesis System for RT-PCR kit (Invitrogen, Carlsbad, CA, USA) in a total volume of 40 μl. The efficiency of cDNA synthesis was estimated using the endogenous genes ubiquitin and β-tubulin for coffee and H. vastatrix, respectively.
Each RT-PCR reaction consisted of 94°C for 2 min, 30 cycles of 94°C for 30 s, 60°C for 30 s and 72°C for 30 s, followed by a final extension at 72°C for 5 min. Primer sequences and amplicon sizes are shown in Table S6.
RNA isolation and cDNA synthesis were conducted essentially as described for RT-PCR with the same primer sets. Gene expression was quantified according to the comparative method 2 -ΔΔCt (Livak & Schmittgen, 2001). For quantification of gene expression at each time point (24, 48, and 72 h, and 9 and 15 d after inoculation) the mean Ct value resulting from three biological replicates and two technical replicates at each sampling time was used. The expression of each candidate effector gene was then normalized with the three endogenous fungal genes glyceraldehyde-3-phosphate dehydrogenase (HvGAPDH), cytochrome c oxidase subunit III (HvCytIII) and β-tubulin (Hvβ-tub), which have been validated for quantification of coffee rust gene expression (Vieira et al., 2011). qPCR reactions were carried out in a 7500 Real Time PCR Systems (Applied Biosystems, Grand Island, NY, USA), programed for initial denaturation at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 30 s. Each reaction was run in a total volume of 20 μl containing 40 ng cDNA, 8 mM of each primer and 12 µl SYBR ® Green PCR Master Mix (Applied Biosystems). The specificity of the amplification was determined using a dissociation curve with a gradient of temperature ranging from 60°C to 95°C at a rate of 1°C per 30 s.

Methods S10 Bacterial strains and EDV constructs.
HvECs without secretion sequences were cloned into pENTR D-TOPO (Invitrogen, Carlsbad, CA, USA) and then recombined into pEDV6 (Badel et al., 2013;Fabro et al., 2011) using the Gateway ® LR Clonase TM II Enzyme Mix kit (Invitrogen) according to the manufacturer's instructions. In-frame fusions with the secretion signal of AvrRps4 and integrity of the effector sequence of all pEDV6::HvEC constructs were confirmed by sequencing with universal M13 primers. pEDV6 plasmids expressing HvECs were mobilized from E. coli DH5α to Psgc 1202 by standard triparental mating using E. coli HB101 (pRK2013) as helper strain. Empty pEDV6 vector was maintained in E. coli DB3.1. Transformed Psgc 1202 cells were selected on solid King's B medium containing rifampicin (100 µg ml -1 ) and gentamicin (25 µg ml -1 ).

Methods S11 Plant material and assays.
Plants of coffee accessions differential for H. vastatrix physiological races used in infiltration assays were obtained by vegetative propagation by stake rooting and/or grafting using Catuaí Vermelho IAC 44 as rootstock. After 4 months of nursery, plants were transferred to the glasshouse when the coffee seedlings had two or three pairs of leaves and grown in 5-l plastic pots containing a mix of soil and manure supplemented by liquid fertilization. Psgc 1202 wild type and Psgc 1202 carrying pEDV6::HvEC constructs or empty vector were grown on solid King's B medium containing appropriate antibiotics at 28°C overnight, washed once and resuspended in sterile 10 mM MgCl 2 . In order to test whether expression of HvECs in Psgc 1202 could alter its ability to cause disease on coffee plants expressing a subset of S H resistance genes, fully expanded tender leaves were infiltrated with bacterial suspensions at 2 × 10 7 CFU ml -1 using a needless syringe and symptoms evaluated daily until 10 dpi. The experiment was repeated at least three times with similar results. To determine bacterial multiplication in planta bacterial suspensions at 1 × 10 4 CFU ml -1 were infiltrated into fully expanded tender leaves of plants carrying or not the S H 1 resistance gene. Leaves from three independent plants were infiltrated with each bacterial strain. Four 1-cm diameter disks were collected from each infiltrated leaf at several time points after inoculation using a cork borer. Leaf disks were ground in 10 mM MgCl 2 , the macerate serially diluted and plated on solid King's B medium containing appropriate antibiotics and the number of colony forming units (CFU) counted.
Plasmid pNR527 (expressing AvrRpm1 T3SS::Cya) (Upadhyaya et al., 2014) was transformed into Psgc 1202 by electroporation. Psgc 1202 (pNR527) was grown on solid King's B medium at 28°C for 24 h, the cells resuspended in 10 mM MgCl 2 and the suspension adjusted to 4 × 10 7 CFU ml -1 . The bacterial suspension was infiltrated into young fully expanded leaves of Catuaí Vermelho IAC 44 plants using a needless syringe. Leaves inoculated with 10 mM MgCl 2 were used as controls. One-centimetre diameter leaf discs were collected from the inoculated leaves at 0, 18, and 24 h after infiltration. The discs were frozen in liquid nitrogen and kept at -80°C until processing. Three independent samples, each consisting of four leaf discs, from different inoculated plants were analysed. cAMP extraction and protein quantification were conducted as previously described (Carper-Lindley et al., 2002;Upadhyaya et al., 2014) with some modifications. Briefly, leaf discs were macerated in liquid nitrogen, 325 µl 1.1 M HClO4 was added, the suspension vortexed and centrifuged at maximum speed for 10 min. Three hundred microlitres of supernatant was neutralized with 40 µl 6M K2CO3. The pellet was used for protein quantification using the Bio-Rad Protein Assay kit II (Bio-Rad, Hercules, CA, USA).
The supernatant was centrifuged for 8 min at maximum speed and 200 µl were collected. Ten-microlitre aliquots were used for cAMP quantification using the Cyclic AMP EIA kit (Cayman Chemical Company, Ann Arbor, MI, USA).
Notes S1 Sequencing statistics for cDNA libraries from Hemileia vastatrix pre-biotrophic and biotrophic phases of the infection cycle. Red font indicates the canonical GT/AG splicing sites in exon-intron junctions. Alignments were done using ClustalW (Thompson et al., 1994).