Functional Characterization of Calcineurin Homologs PsCNA1/PsCNB1 in Puccinia striiformis f. sp. tritici Using a Host-Induced RNAi System

Calcineurin plays a key role in morphogenesis, pathogenesis and drug resistance in most fungi. However, the function of calcineurin genes in Puccinia striiformis f. sp. tritici (Pst) is unclear. We identified and characterized the calcineurin genes PsCNA1 and PsCNB1 in Pst. Phylogenetic analyses indicate that PsCNA1 and PsCNB1 form a calcium/calmodulin regulated protein phosphatase belonging to the calcineurin heterodimers composed of subunits A and B. Quantitative RT-PCR analyses revealed that both PsCNA1 and PsCNB1 expression reached their maximum in the stage of haustorium formation, which is one day after inoculation. Using barely stripe mosaic virus (BSMV) as a transient expression vector in wheat, the expression of PsCNA1 and PsCNB1 in Pst was suppressed, leading to slower extension of fungal hyphae and reduced production of urediospores. The immune-suppressive drugs cyclosporin A and FK506 markedly reduced the germination rates of urediospores, and when germination did occur, more than two germtubes were produced. These results suggest that the calcineurin signaling pathway participates in stripe rust morphogenetic differentiation, especially the formation of haustoria during the early stage of infection and during the production of urediospores. Therefore PsCNA1 and PsCNB1 can be considered important pathogenicity genes involved in the wheat-Pst interaction.


Introduction
Calcineurin, a serine-threonine-specific calcium/calmodulindependent protein phosphatase with two subunits (CNA and CNB), regulates a variety of physiological processes, such as growth, morphogenesis, pathogenicity, and membrane stress responses through the calcium signaling pathway in eukaryotes [1,2]. The first fungal calcineurin genes were reported in 1991 from the budding yeast Saccharomyces cerevisiae [3,4] and the filamentous fungus Neurospora crassa [5]. Many homologs of CNA or/and CNB have been found in medicinal fungi [6] and plant pathogens such as Botrytis cinerea [7] and Magnaporthe oryzae [8,9]. Recent studies have confirmed that calcineurin controls virulence, hyphal elongation and multiple stress responses in the human pathogens Candida dubliniensis [10], Cryptococcus neoformans [11,12], Candida albicans [13] and Aspergillus fumigatus [14,15]. Similar findings have also been reported for the phytopathogens Ustilago maydis [16], Cochliobolus miyabeanus [17], and Sclerotinia sclerotiorum [18]. The calcineurin pathway also plays a role in drug resistance to azoles in C. albicans [19,20], and in C. dubliniensis [10]. Inhibition of calcineurin can decrease fungal growth and arrest tissue invasion [21]. This opens possibilities to develop new antifungal agents targeting the calcineurin pathway in fungi [6]. RNA induced gene silencing or RNA interference (RNAi) is a complex natural phenomenon and a powerful reverse genetics tool for the analysis of gene function in eukaryotes [22][23][24][25][26]. In plants, virus-induced gene silencing (VIGS) was developed for rapid functional analysis of plant genes using viruses to deliver silencing constructs [27][28][29]. It has widely been applied in dicots such as Arabidopsis [30], tobacco [31] and tomato [32][33], and monocots such as barley [34] and wheat [35][36][37][38]. In fungi, RNAi technology has been deployed in more than 40 species including plant and human pathogens [39]. Nguyen et al. [40] developed a highthroughput RNA-silencing vector for M. oryzae to identify an involvement of calcineurin genes in colony pigmentation, sporulation, appressorium formation, and pathogenicity. However, for there is still no applicable transformation system available there are currently no techniques on hand for silencing genes in obligate biotrophic fungi directly. Host-induced gene silencing (HIGS) is a newly developed RNAi technology to indirectly silence parasite genes by expressing an RNAi construct in vivo in the host [41]. Host induced RNAi of three target genes suppressed their expression in the planthopper Nilaparvata lugens after feeding on rice plants [42]. Recent studies confirm the hypothesis that fungal genes can be suppressed in planta during interaction of the fungus with the host. Tinoco et al. [43] silenced the reporter gene GUS in Fusarium verticillioides by expressing GUS dsRNA in tobacco. HIGS was also successfully used in obligate biotrophic fungi. Using a BSMV-VIGS system expressing the target dsRNA in wheat, Nowara et al. [44] showed that the fungal genes GTF1 and GTF2 in Blumeria graminis play a role in haustorium formation and elongation of secondary hyphae. Yin et al. [45] also developed a BSMV-based HIGS approach to identify gene function in the biotrophic rust fungus Puccinia striiformis f. sp. tritici (Pst).
Wheat stripe rust, caused by the Basidiomycete Pst, is an important disease in wheat worldwide. As an obligate biotrophic pathogen infecting wheat leaves, Pst undergoes a high degree of morphological and physiological differentiation from urediospore to germ tube, invasive hypha and haustorium, a special structure for nutrient uptake from the host [46][47][48]. A few studies reported that the calcium signaling pathway is involved in the initial infection and biotrophic growth of rust fungi [49,50]. Some homologs involved in calcium signaling such as CDPK were identified in Pst [49]. Recently, the PST-130 genome has been sequenced [51]. The sequences provided information to clone Pst genes involved in calcium signaling. In this study, we describe cloning, sequencing and transcription analysis of two calcineurin subunits from Pst, designated PsCNA1 and PsCNB1. HIGS analysis using the BSMV-VIGS system and drug tests indicate a vital function in rust growth, development and sporulation.

PsCNA1 and PsCNB1 encode calcineurin homologs
One of the expressed sequence tags (ESTs) from a full-length cDNA library of Pst [52] was found to be highly similar to PtCNA from Puccinia triticina (PTTG_07903) and PgCNA from Puccinia graminis f. sp. tritici (PGTG_14891). Another two ESTs from cDNA libraries contructed by Zhang et al. [49] and Ma et al. [53] are almost identical to the CNB genes from the other two wheat rusts (PTTG_02210 and PGTG_04308). Further sequencing of these clones from the Chinese Pst race CYR31, provided full-length cDNA sequences of PsCNA1 and PsCNB1 (Genbank accession numbers JX424819 and JX424820, respectively). The full length cDNA sequence for PsCNA1 is 2,680 bp with an open reading frame (ORF) of 2,097 bp encoding a 698 amino acid (AA) protein, which consists of two calcineurin A domains and six Serine/ threonine-protein phosphatase domains with a calculated molecular mass of 76.74 kDa (Fig. S1). The PsCNB1 cDNA is 770 bp in length with an ORF of 528 bp encoding a 175 AA protein, with a molecular mass of 19.77 kDa, which has four calcium-binding EFhand motifs and a N-myristoylation site (Fig. S2).
The levels of conservation of PsCNA1 and PsCNB1 are indicated in comparison with homologs from other fungi and some model organisms in Figure S1 and Figure S2, respectively. PsCNA1 is 90% identical to the calcineurin A subunit of P. triticina and 84% identical to P. graminis f. sp. tritici CNA. PsCNA1 is conserved among other fungi analyzed with 54% to 75%. PsCNB1 exhibits strong similarity to calcineurin B proteins from other organisms. It is 100% identical to PtCNB, 99% identical to PgCNB, and 60% to 83% identical to CNB genes from other fungi analyzed.
Phylogenetic analysis revealed that PsCNA1 and PsCNB1 cluster with other Basidiomycete fungi. Especially the three Puccinia sp. were most close to each other (Fig. 1). However, PsCNA1 was closer to P. graminis f. sp. tritici than to P. triticina while PsCNB1 was closer to P. triticina than to P. graminis f. sp. tritici.

Suppressors block Pst germination
The immuno-suppressants cyclosporin A (CsA) and FK506 inhibit calcineurin activity and affect its function just like mutants of CNA or/and CNB in several fungi [6]. In order to test whether these two drugs affect the function of PsCNA1 and PsCNB1, stripe rust urediospores were incubated with these drugs and germination was monitored. After 10 hours, microscopic analyses indicated that germination rate was reduced to 40.5% for FK506 (3 mM) and 66.5% for CsA (3 mM) treatment, compared to water (Table 1). Germ tubes of Pst were limited in their elongation by treatment with of FK506 (1 mM) or CsA (0.1 mM) compared to the control (Fig. 2C, 2B, 2A). Urediospores also frequently produced two or three more irregular germ tubes than the control (Fig. 2D, 2E, 2F).

Expression profiles of PsCNA1 and PsCNB1
To gain insight into the possible function of PsCNA1 and PsCNB1 in Pst, we investigated the expression of PsCNA1 and PsCNB1 (mRNA abundance) in different stages of Pst using quantitative PCR (qRT-PCR). PsCNA1 and PsCNB1 had similar expression profiles (Fig. 3). Transcript levels of both genes were drastically increased at 1.0 dpi, but quickly decreased to 3-fold, and 0.5-fold, respectively, at 11 dpi (Fig. 3). The maximum accumulation of transcript was 58 fold for PsCNA1 and 38 fold for PsCNB1. However, kinetics of transcript accumulation differed between the two genes. The transcripts for PsCNB1 drastically increased over time up to 1 dpi, whereas transcript level of PsCNA1 only showed a dramatic increase at 1 dpi.

HIGS for PsCNA1 and PsCNB1
In order to determine the best inoculation day for Pst inoculation, virus symptoms were scored by visual assessment at four time point (8, 10, 12, 14 dpi) for the BSMV:c:0-as vector (data not shown). Only three out of eighteen seedlings showed slight virus symptoms at 8 dpi. Symptoms increased at 10 dpi, and almost all seedlings showed 100% virus infection at 12 dpi. At 14 dpi leaves showed heavy symptoms with large yellow areas.Therefore, 12 dpi with BSMV was chosen for rust inoculation.
In order to identify HIGS efficiency (knockdown rates), transcript levels of PsCNA1 and PsCNB1 were scored in inoculated silenced plants at 8 dpi by qRT-PCR. Results showed that silencing was detected for both BSMV vectors. PsCNA1 transcript level exhibited an average of 24% expression in BSMV:c:PsCNA1as infected plants. However, PsCNB1 transcript level showed only an average of 18% reduction. HIGS for both genes lost most effectiveness at 16 dpi (PsCNA1: 49% expression, PsCNB1: 110% expression) (Fig. 4).

Silencing calcineurin blocks Pst growth and development in wheat leaves
To determine cytological changes associated with fungal growth on plants, silenced for PsCNA1 or PsCNB1 wheat leaves inoculated with race CYR31 were examined microscopically. Two time points (2 and 5 dpi) were compared. No significant differences in Pst development and hyphal growth were observed between control plants and plants carrying PsCNA1 or PsCNB1 knock-down constructs at 2 dpi (Table 2; Fig. 5A, 5B, 5C). However, at 5 dpi hyphal length on average in BSMV:c:PsCNA1-as and BSMV:c:PsCNB1-as infected wheat leaves were much shorter than those observed in controls (Table 2).

Reduction in rust sporulation after silencing Pst calicineurin
We also scored sporulation and found that the number of uredia was reduced on silencing plants (Fig. 6). Sporulation of Pst on silencing plants occurred two days later (12 dpi) than on the control plants (10 dpi). Statistical analyses 16 dpi determined an average number of 99 uredia for control plants, 60 uredia for BSMV:c:PsCNA1-as, and 68 uredia for BSMV:c:PsCNB1-as ( Table 2). Uredia in silencing plants were smaller in size with open cavities that were shorter, and contained fewer spores (Fig. 7).

Discussion
In this study we describe the isolation and characterization of two calcineurin genes from the wheat stripe rust fungus Pst. Phylogenetic analyses of eukaryotic CNA and CNB genes clearly show that PsCNA1 and PsCNB1 are closely related to calcineurin genes from other basidiomycetes. The calcineurin A/B protein family appears to be conserved in size and structure. Most CNA genes encode a protein of more than 500 amino acids and most CNB proteins contain about 175 amino acids [54]. All CNA proteins share high homology within the N-terminal 50 to 420 amino acid residues, but with significant variations in the C-   terminal part. With a size of almost 700 aa PsCNA1 is considerably larger than other CNA proteins (Fig. S1). In contrast, PsCNB1 shares length, four conserved EF-hands, and a Nmyristoylation site with other CNB proteins (Fig. S2). Especially the three Puccinia sp. have conserved amino acid residues at the Nmyristoylation site and the first three EF motifs. Juvvadi et al. [15] found that CnaB is required for localization of CnaA to the septum and that the two calcineurin subunits are required to control hyphal growth and septation in A. fumigatus. PsCNB1 is upregulated earlier than PsCNA1, which might indicate that expression of PsCNB1 is necessary for expression or stability of PsCNA1. Whether PsCNB1 is indeed necessary for stabilization of PsCNA1 and whether the two gene products co-operate still needs more research.
Harel et al. [18] reported that the transcript level of CNA in S. sclerotiorum is 2.5-fold higher in sclerotia than in infection hyphae. In contrast, our results from qRT-PCR show that transcript levels of PsCNA1 and PsCNB1 were much higher at 1 dpi, which corresponds to an early infection stage such as invasive hypha and initial haustorium formation. These differences in expression may be due to the fact that RNAs from S. sclerotiorum were prepared from mycelium grown on artificial medium, while RNAs of Pst were prepared from infected host plants (except germinated urediospores). Another reason may be different functions of calcineurin in the basidiomycete Pst and the ascomycete S. sclerotiorum.
Complexes of CsA with Cyclophilin A or/and FK506 with FK506-binding protein 12 (FKBP12) interfere with calcineurin binding to other phosphoprotein substrates in eukaryotes [1,6]. The two inhibitors have been applied in numerous fungi to illustrate functions of calcineurin. CsA (10 mg/ml) induced celldivision patterns indistinguishable from mutants in calcineurin in U. maydis [16]. Both CsA and FK506 suppress growth of C. neoformans in vitro at 37uC [6,11], and a similar phenomenon is seen in C. dubliniensis [10]. With FK506 and CsA treatment, A. fumigatus reveals stunted hyphae, more branches and defects of the conidiophore [55]. The immuno-suppressants CsA and FK506 led to germination defects in Pst. More than two or branched germ tubes appeared, which is a similar phenotypes as inhibitor treatment in U. maydis [16], or mutants of calcineurin in Ustilago hordei [56] or A. fumigatus [15].
Host induced gene silencing (HIGS) has succeeded to identify functions of parasite genes as an efficient reverse genetic tool in several fungi, insects and nematodes [41]. HIGS has emerged as parasite-derived resistance (PDR) to develop durable resistance in agricultural industry [41]. Virus induced gene silencing (VIGS) mediated by the barley stripe mosaic virus (BSMV) has been successfully developed in wheat [35,36], and recently applied to HIGS with B. graminis [44] and Pst [45]. The BSMV-VIGS system is developed to express double stranded RNA (dsRNA) of targets from Blumeria and Puccinia genes in plants to trigger RNA silencing. We use BSMV-VIGS as viral vector to deliver Pst silencing constructs for calcineurin genes in order to silence them in Pst through the host. The suppression was almost 76% for PsCNA1, but only 18% for PsCNB1 at 8 dpi (Fig. 4). Effective silencing was evident by the reduced number of uredia (Table 2. and Fig. 6), and smaller uredia with less open areas (Fig. 7). Our results illustrate different knock-down efficiencies for Pst HIGS by BSMV vectors with the two calcineurin genes. Silencing efficiency is variable for different genes or even the same gene. This has been shown for thirty-seven genes in M. oryzae [40], eleven genes of Pst [45], inf1 in Phytophthora infestans [57] and GUS in F. verticillioides [43]. However, the phenotype of PsCNB1 was similar to that of PsCNA1 knock downs. An explanation might be that PsCNA1 or/and PsCNB1 act   Baulcombe [26] reported that plants have feedback mechanisms in RNA silencing. Fungi seem to exhibit a similar regulatory phenomenon. GUS expression and activity resumed to normal levels after seven generations in F. verticillioides [39]. Expressions of gfp and inf1 are partially revcoverd in the Oomycete P. infestans [57]. Our results for Pst also showed recovery of expression of target genes after RNA silencing. The expression of silenced PsCNA1/PsCNB1 genes increased up to almost normal expression levels during Pst development in planta (Fig. 4). This might be due to self-repair or self-protection mechanisms for an RNAi dynamic balance. In developing hyphae and haustoria increasing normal RNA might break the balance of RNA silencing. These phenomena must be considered as drawback of RNAi [39,58].
The silencing signal can be propagated to the offspring in plants and fungi. VIGS by BSMV vectors can be transferred to the next generation in wheat and barley [38]. Other VIGS vectors still also transmit silencing signals to next generation seedlings. Gene silencing by VIGS-ALSV (Apple Latent Spherical Virus) is from 33% of first progeny seedlings to 55% of subsequent progeny in soybean [27]. Silencing gfp was maintained in subsequent generations of Moniliophthora perniciosa [59]. We did not directly show BSMV heredity to next generation Pst urediospores. But after HIGS to PsCNA1 and PsCNB1, we observed that the new urediospores also had similar defects with strange branched tubes (Fig. 8B, 8C) to that by the immuno-suppressors FK506 and CsA (Fig. 2E, 2F). These results illustrated that PsCNA1 and PsCNB1 take important function in the morphodifferentiation, hyphal development and sporulation in wheat stripe rust.
Although HIGS can be a valuable tool in identifying functions of fungal genes, the question is how the transfer of the silencing signals takes place. Fungal haustoria and the extrahaustorial matrix between host and fungus are specialized places to exchange    nutrients and information [47,48]. Nowara et al. [44] regarded the exosomal pathway as the means of secreted dsRNA or siRNA tranfer from the host plant wheat into B. graminis. By this way multivesicular bodies and exosomes transport RNAs to the extrhaustorial matrix. Yin et al [45] postulated that RNA silencing signals also extend from the expressing host cell into haustoria of Pst. Another possibility would be that the VIGS-BSMV vectors can cross the extrahaustorial matrix into haustoria in Pst. In this case the rust fungus needs a complete silencing machinery to accomplish degradation of RNAs. Although Argonaute-like proteins (AGO) have been identified in Pst (GenBank accession: AEM61140.1) [51], future work could answer these questions.

Strains and RNA isolation
Chinese Pst race CYR31 was inoculated and propagated on wheat cultivar Suwon 11 as described previously [53]. For isolating RNA from infected plants, infected wheat leaves were harvested at 0, 0.5, 1, 3, 7, 9 and 11 days post inoculation (dpi). All material was frozen in liquid nitrogen and stored at 280uC.
Total RNA was extracted using the Qiagen Plant RNeasy kit (Qiagen, Hilden, Germany) from sampled urediospores, germinated urediospores, and infected wheat leaves. First-strand cDNA was synthesized from 1 mg total RNA of each sample using the SMART TM reverse transcription Kit (Clontech Laboratories, Inc., Mountain View, CA) according to the manufacturer's instructions.

Isolation and sequence analysis of PsCNA1 and PsCNB1
Screening the Pst cDNA library constructed by Ling et al. [52], one EST clone (480 bp) was found to be homologous to CNA genes. Another two EST clones (zyh1090 and mjb959; GenBank accession: ES322265 and GR305110, respectively) with homology to CNB genes were obtained from a Pst cDNA library constructed by Zhang et al. [49] and a cDNA library from wheat leaves inoculated with Pst constructed by Ma et al. [53]. Two primers PsCNA-S/AS and PsCNB-S/AS (Table 3) were designed to get the full length cDNA sequence of PsCNA1 and PsCNB1 from Pst. The clones of PsCNA1 and PsCNB1 were sequenced on an ABI PRISM 3130XL Genetic Analyzer (Applied Biosystems, Carlsbad CA). Sequences were analyzed using NCBI BLAST (http://blast. ncbi.nlm.nih.gov/Blast.cgi), BLAST searches against the database of P. graminis f. sp. tritici and P. triticina (http://www.broadinstitute. org/annotation/genome/puccinia_group/GenomesIndex.html), and ORF Finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). The alignments of the deduced protein sequences and phylogenetic trees were computed using MEGA4 and ClustalX version 1.83 as described by Guo et al. [60]. PsCNA1 and PsCNB1 sequences have been deposited in GenBank (GenBank accession number JX424819 and JX424820, respectively).

Quantitative RT-PCR
To analyze the transcript levels of the two subunits of Pst calcineurin, relative quantification of gene expression was performed using quantitative RT-PCR (qRT-PCR) on an ABI prism 7500 sequence detection system (Applied Biosystems, Carlsbad, CA). Transcript abundance was assessed with three independent biological replicates. Amplification was performed as follows: 95uC for 1 min, followed by 40 cycles of 10 s at 95uC, 20 s at 60uC and 40 s at 72uC. This was followed by melting curve analysis. The transcript levels of PsCNA1 and PsCNB1 were calculated by the 2 2DDCT method with the EF1 gene of Pst as endogenous reference for normalization as described by Guo et al. [60]. The following primers were used for qRT-PCR PsCNA1 Table 3), PsCNB1 (PsCNB-1RT S/AS, Table 3). Relative quantification of PsCNA1 and PsCNB1 was computed for the different stages in comparison to that at zero hour incolulation with Pst urediospores.

Construction of BSMV-based VIGS vectors and VIGS assay
BSMV-VIGS vectors are based on the constructs by Holzberg et al. [35]. To avoid non-specific silencing of wheat genes, the target regions of VIGS vectors were blasted for homologs to all wheat sequences in the NCBI database and designed to be rustspecific. Selected PsCNA1 and PsCNB1 gene fragments were amplified by PCR from Pst cDNA using primers with restriction enzymes NotI and PacI sites (Primers: PsCNA-1VS/AS and PsCNB-1VS/AS, respectively; Table 3). Amplicons were ligated into the BSMV c vector generating BSMV:c:PsCNA1-as and BSMV:c:PsCNB1-as. The native BSMV:c:0-as was used as negative control.
Two-leaf wheat seedlings were used for virus inoculation by rubbing the first leaf as described by Yin et al [45]. Seedlings were incubated in the greenhouse after spraying with water (25uC for 16 hours light and 20uC for 8 hours dark). After inoculation with Pst urediospores, plants were incubated at 20uC for 16 hours light and 16uC for 8 hours dark. Primers (PsCNA-1RT S/AS and PsCNB-2RT S/AS, respectively; Table 3) were used for assaying the transcript levels of PsCNA1 and PsCNB1. Control seedlings were infected with the BSMV:c:0-as vector and also inoculated with Pst. Total RNA was extracted from leaves of 18 wheat seedlings at two time points (8 and 16 days) after rust inoculation.

Histological obervtion of Pst growth in wheat leaves
Wheat leaves were sampled at 2 and 5 days and stained with Calcofluor White. The infected leaves were examined with the microscope (Olympus BX-51) to observe Pst haustoria and infection hyphae under UV light. Wheat leaves infected Pst (16 dpi) were observed by scanning electron microscopy (JEM1230). Inhibitor assays using CsA or FK506 CsA (Sigma, USA) was diluted to be 1 mM mother solution in 95% ethyl alcohol. FK506 (Sigma, USA) was diluted to be 1 mM mother solution in DMSO. CsA and FK506 were added to 5 ml of sterile water with Pst urediospores to reach final concentrations of 3 mM and 0.1 mM for CsA, or 3 mM and 1 mM for FK506. Spore suspensions were incubated at 4uC for germination with sterile water-incubated urediospores as control. Germination was examined under the microscope after 10 hours.