Silencing of the Mitogen-Activated Protein Kinases (MAPK) Fus3 and Slt2 in Pseudocercospora fijiensis Reduces Growth and Virulence on Host Plants

Pseudocercospora fijiensis, causal agent of the black Sigatoka disease (BSD) of Musa spp., has spread globally since its discovery in Fiji 1963 to all the banana and plantain growing areas across the globe. It is becoming the most damaging and economically important disease of this crop. The identification and characterization of genes that regulate infection processes and pathogenicity in P. fijiensis will provide important knowledge for the development of disease-resistant cultivars. In many fungal plant pathogens, the Fus3 and Slt2 are reported to be essential for pathogenicity. Fus3 regulates filamentous-invasion pathways including the formation of infection structures, sporulation, virulence, and invasive and filamentous growth, whereas Slt2 is involved in the cell-wall integrity pathway, virulence, invasive growth, and colonization in host tissues. Here, we used RNAi-mediated gene silencing to investigate the role of the Slt2 and Fus3 homologs in P. fijiensis in pathogen invasiveness, growth and pathogenicity. The PfSlt2 and PfFus3 silenced P. fijiensis transformants showed significantly lower gene expression and reduced virulence, invasive growth, and lower biomass in infected leaf tissues of East African Highland Banana (EAHB). This study suggests that Slt2 and Fus3 MAPK signaling pathways play important roles in plant infection and pathogenic growth of fungal pathogens. The silencing of these vital fungal genes through host-induced gene silencing (HIG) could be an alternative strategy for developing transgenic banana and plantain resistant to BSD.


INTRODUCTION
Pseudocercospora fijiensis, causal agent of black Sigatoka disease (BSD) in Musa spp. (banana and plantain), was first recognized in 1963 in the South-Eastern Coast of Viti Levu in Fiji (Mourichon et al., 1997;Marin et al., 2003;Churchill, 2011). Almost 30 years later, BSD was reported in Honduras from where it spread to Guatemala, Southern Mexico, Panama, Ecuador, and Peru. In Southeast Asia, it is found in the Philippines, Taiwan, and Indonesia. In Africa, BSD was first reported about 25-30 years ago in Gabon and Zambia (Churchill, 2011). Since then, BSD has spread to sub-Saharan countries in the West African coast and to the Eastern African countries (Mourichon et al., 1997;Ploetz, 2001;Marin et al., 2003;Churchill, 2011).
Pseudocercospora fijiensis is one of the most damaging and economically important pathogens of Musa spp. worldwide (Farr et al., 1995;Stewart et al., 1999;Carlier et al., 2000). Attempts to control BSD include frequent application of fungicides and cultural practices such as the removal of infected leaves and proper spacing and drainage in plantations (Ploetz, 2001). Fungicide control of BSD in Central America is responsible for approximately 27% of the retail price of bananas (Stover, 1980(Stover, , 1986Stover and Simmons, 1987). The cost of chemical control of BSD is estimated about US$400 to US$1,400 per hectare (Stover, 1986;Pasberg-Gauhl et al., 2000;Ploetz, 2000;Arias et al., 2003;Churchill, 2011), and smallholder farmers cannot afford fungicides and are therefore more prone to losses due to BSD. Also, P. fijiensis develops resistance to fungicides after many sprays and therefore, better strategies are needed to efficiently control this disease (Stover, 1990).
Pseudocercospora fijiensis reproduces sexually by means of ascospores that are mainly produced during the later stages of disease and asexually through conidia produced during the early stages. Ascospores are the main way of long-distance dispersal of P. fijiensis between plantations and into new areas (Ploetz, 2001;Marin et al., 2003;Agrios, 2005). The spores germinate within 2-3 h under high humidity and temperatures greater than 20 • C, and then enter the host through the stomata openings within 48-72 h (Stover, 1980). The fungal hyphae then grow inside the leaf, colonizing the intercellular spaces and killing plant cells. After infection, the pathogen emerges from the stomata and will develop conidiophores that can start the new cycle of infections (Churchill, 2011). Streaks that usually appear first near the leaf apex and along the leaf margin are a sign of infection. Initial symptoms are seen only at 10-30 days after infection. Diseased leaves will become sources of inoculum for new infections (Meredith, 1970;Carlier et al., 2000;Marin et al., 2003). Aggressiveness of P. fijiensis is directly related to environmental conditions; BSD is more pronounced when relative humidity is greater than 80% and when temperature is above 23 • C (Fouré, 1994;Gauhl, 1994;Torrado-Jaime and Castaño-Zapata, 2008).
Though efforts are underway to develop BSD host resistance in banana and plantain through conventional breeding, genetic engineering could be an alternative approach for developing resistant cultivars. One possible means might be to use RNA interference (RNAi) to target fungal genes responsible for regulating plant infection, invasive growth, and pathogenicity of P. fijiensis. Transgenic banana resistant to fusarium wilt disease were developed through posttranscriptional silencing of fungal genes velvet and Fusarium transcriptional factor 1 (Ghag et al., 2014).
A family of serine/threonine protein kinases known as mitogen-activated protein kinases (MAPKs) are involved in the transduction of a variety of extracellular signals and the regulation of different developmental processes. The yeast extra cellular signal-regulated kinase (YERK1) is the most thoroughly investigated MAPK. The MAPK cascade in Saccharomyces cerevisiae has three protein kinases that act in series; a MAP kinase kinase kinase (MAPKKK or MEKK), a MAP kinase kinase (MAPKK or MEK) and finally MAP kinase (MAPK) (Marshall, 1994;Cobb and Goldsmith, 1995). Upon activation of the cascade, MAPKKK phosphorylates the MAPKK, which in turn phosphorylates MAPK (Figure 1). The MAPK cascades in fungi regulate transcription factors by MAPK mediated phosphorylation (Errede et al., 1993(Errede et al., , 1995. The MAPKs, Fus3, and Slt2 appear to be involved in pathogenicity of fungi (Mayorga and Gold, 1999;Xu, 2000). Fus3 regulates pheromone response and invasion pathways, while Slt2 is involved in the cell-wall integrity pathway (Mayorga and Gold, 1999;Xu, 2000).
A number of YERK1 proteins have been shown to be involved in the formation of infection structures such as appressoria and the invasive growth of fungal plant pathogens, such as the maize pathogen, Ustilago maydis (Mayorga and Gold, 1999;Muller et al., 1999). In addition, MAPKs take part in signal transduction pathways that are activated in regulation of growth and development (Alonso-Monge et al., 2001). The Fus3/Kss1type related gene MaMk1, identified in Metarhizium acridum, encodes a member of the YERK1 subfamily, which is known for regulating appressorium formation and insect cuticle penetration (Jin et al., 2014). In several fungal plant pathogens including Zymoseptoria tritici, Puccinia striiformis f. sp. tritici, Fusarium oxysporum, and F. proliferatum, Fus3/Kss1 homologs are shown to be responsible for colonization in mesophyll tissue, growth, spore formation, penetration, and virulence (Mendgen et al., 1996;Cousin et al., 2007;Guo et al., 2011;Zhao et al., 2012).
Likewise, MAPK Slt2 has been well studied in S. cerevisiae and is known to be required for cell-wall integrity (Xu, 2000). In Z. tritici and the entomopathogenic fungus Beauveria bassiana, Slt2 homologs are well known for their roles in invasive growth and virulence (Mehrabi et al., 2006;Luo et al., 2012) and in Alternaria alternata, it is crucial for conidial formation, hyphal elongation and fungal pathogenicity (Yago et al., 2011). Silencing of PsMpk1, a Slt2 type MAPK in the oomycete Phytophthora sojae, showed loss in pathogenicity on susceptible soybean host plants, with triggered enhanced cell death (Li et al., 2014). Defects in fungal growth, zoosporogenesis, and increased hypersensitivity to cell-wall degrading enzymes were also reported (Li et al., 2014).
Pseudocercospora fijiensis and Z. tritici are phylogenetically related, and since Fus3 and Slt2 are known to be responsible for regulating the host penetration, invasive growth, and pathogenicity of Z. tritici, it is possible that Fus3 and Slt2 are important pathogenicity factors for P. fijiensis as well. Therefore, to study the roles of Fus3 and Slt2 in pathogenicity of P. fijiensis, we silenced these genes and tested the transformants on young potted tissue-culture plants of East African Highland Banana (EAHB) cultivar 'Nakitembe' for disease development. This study confirmed that the Slt2 and Fus3 MAP kinase signaling pathways are important for plant infection, invasive growth and FIGURE 1 | Schematic diagram showing pheromone, filamentous invasion, and cell-wall integrity pathways of the MAP kinase cascade of Saccharomyces cerevisiae. Diagram is adopted from (Errede et al., 1995;Gustin et al., 1998). pathogenicity of P. fijiensis in EAHB. Therefore, targeting PfFus3 and PfSlt2 could contribute to developing resistant varieties of banana and plantain against P. fijiensis causal agent of BSD.

Pseudocercospora fijiensis Culture Isolation and Confirmation
The P. fijiensis culture used in this study was isolated from infected leaves of the banana cultivar 'Nakitembe' as described by Onyilo et al. (2017). Genomic DNA was extracted from mycelia following the protocol described by Mahuku (2004) with some modifications. The P. fijiensis isolate was confirmed by PCR using P. fijiensis-specific primers (MF137 GGCGCCCCCGGAGGCCGTCTA and R635 GGTCCGTGTTTCAAGACGG) based on ITS region (Johanson and Jeger, 1993). The cultures of P. fijiensis, P. musae, and P. eumusae collected from CBS-KNAW, the Fungal Biodiversity Centre in Netherlands, were used as controls.
The PCR reaction mixture contained 10 µM each of the forward and reverse primers (0.5 µl), AmpliTaq R DNA polymerase (0.25 µl), (Applied Biosystems, United States), 10× buffer with 15 mM MgCl2 (2.5 µl), (Applied Biosystems, United States), 10 µM deoxyribonucleotides (dNTP) (0.5 µl), 1 µl (100 µg) of genomic DNA of P. fijiensis, adjusted with water to 25 µl final volume. The conditions used were the following: initial denaturation at 95 • C for 5 min, 34 cycles of denaturation at 95 • C for 30 s, annealing temperature at 55 • C for Fus3 and 52 • C for Slt2 for 30 s, and extension at 72 • C for 1 min, followed by a final extension at 72 • C for 5 min and storage at 12 • C. Amplicons were separated by electrophoresis on 0.8% (w/v) agarose gels. PCR products of Fus3 (358 bp) and for Slt2 (264 bp) were isolated from the gel and purified using Zymoclean TM gel DNA recovery kit following the manufacturer's protocol.
The purified Fus3 (358 bp) and Slt2 (264 bp) fragments were ligated into pKOIISD1 plasmid at the EcoRI site and constructs were named as pKOIISD1-PfFus3 and pKOIISD1-PfSlt2 (Figures 2A,B). Plasmid constructs pKOIISD1-PfFus3 and pKOIISD1-PfSlt2 were confirmed for the presence and orientation of inserts by PCR and sequencing, respectively.

Transfer of Plasmid Constructs to Agrobacterium tumefaciens
After validation, the plasmids pKOIISD1-PfFus3 and pKOIISD1-PfSlt2 were transferred to Agrobacterium tumefaciens strain AGL1 according to Onyilo et al. (2017). Colony PCR was then performed to validate the presence of PfFus3 and PfSlt2 in transformed AGL1, using primer pairs, pSD15 (CTTTAAGTTCGCCCTTCCTC) and pSD13 (GTTGACAA GGTCGTTGCGT) designed from the pKOIISD1 vector. PCR reaction mixtures contained 10 µM primers (0.5 µl each of pSD15 and pSD13 ), AmpliTaq R DNA polymerase (0.25 µl, Applied Biosystems, United States), 10× Buffer with 15 mM MgCl2 (Applied Biosystems, United States), 10 µM deoxyribonucleotides (dNTP), AGL1, and water in a total volume of 25 µl. PCR program cycles used were the following: initial denaturation at 95 • C for 5 min and then 34 cycles of denaturation at 95 • C for 30 s, annealing at 55 • C for 30 s and extension at 72 • C for 1 min, followed by final extension at 72 • C for 5 min and storage at 12 • C. The amplified PCR product was separated by electrophoresis on agarose gels.

Transformation of P. fijiensis and Molecular Characterization
The PCR-positive colonies of transformed Agrobacterium tumefaciens strain AGL1 containing pKOIISD1-PfFus3 or pKOIISD1-PfSlt2 were maintained and used to transform P. fijiensis through Agrobacterium tumefaciens-mediated transformation as described by Onyilo et al. (2017).
Genomic DNA was isolated from plugs of mycelia of transformed P. fijiensis grown in V8 juice medium at 25 • C following extraction protocol as described above. The transformed P. fijiensis was validated for presence of PfSlt2 and PfFus3 by PCR using primer pairs pSD15 and pSD13 . Wild-type (WT) untransformed P. fijiensis was used as control.

Evaluation of Fus3 and Slt2 Expression in P. fijiensis Transformants
Total RNA was extracted from wild-type and transformed P. fijiensis using TRIzol Reagent following the protocol provided by Ambion RNA life technologies. Total RNA was purified by RNA clean and concentrator TM kit according to the Zymo research Corp manual. cDNA was prepared using Maxima first strand cDNA kit (Thermo Fishers Scientific, Inc.).

Virulence Assays for PfFus3 and PfSlt2 Transformants
Mycelial fragments from wild-type P. fijiensis and silenced P. fijiensis strains (i.e.,  were cultured in 200 ml of rich medium (yeast extract 10 g, glucose 30 g in 1 L of doubledistilled water) containing 100 µg/ml ampicillin and incubated at room temperature for 5 to 10 days at 150 rpm. Equal amounts of mycelia were macerated in a sterile mortar and filtered using double-layered cheesecloth. Mycelial fragments were resuspended in 10% rich medium containing 1% tween 20, counted using a hemocytometer, and adjusted to 10 4 mycelia fragments/mL. The inoculum was applied on the abaxial side of leaves of 3-month-old potted tissue culture plants of EAHB cultivar 'Nakitembe' with the help of a fine paintbrush. The inoculated plants were incubated under high humidity in a humid chamber covered with clear polythene. Misting of plants was done three times every day for a period of 3 days in order to maintain 80 to 90% humidity in the humid chamber. Disease severity was estimated by counting the number of necrotic lesions per 2 cm 2 areas at four different points on the inoculated leaf. These counts were used to determine the mean necrosis and the disease stage scores were assessed according to Onyilo et al. (2017). Three plant replicates with three leaves per plant were used in each experiment and repeated three times.
To determine the fungal mycelia growth in leaf tissues ca. 3 cm leaf disks were neatly cut with sterile surgical blades at 45 days post inoculation and stained as described by Onyilo et al. (2017). The slides were observed for fungal plant infections, invasive growth, and colonization using a COSLAB light microscope, and pictures were taken using a digital camera MDCE-5C (ISO 9001 Co) and analyzed in Optika Vision Lite 2.1 Software.

Biomass Quantification of P. fijiensis in Leaf Tissue
Infected leaves were harvested at 45 days post inoculation. Genomic DNA was extracted from 1 g of samples of pure P. fijiensis culture, plant leaves infected with P. fijiensis transformants (i.e., PfFus3-5, PfFus3-11, PfTFus3-12, PfSlt2-1, PfSlt2-11 and PfSlt2-12), wild-type P. fijiensis, and non-infected banana leaf samples according to the protocol described by Mahuku (2004). The DNA samples were treated with 1 µl of 10 mg/µl RNase A in a total volume of 50 µl at 37 • C for 45 min. The reaction was terminated at 65 • C for 10 min. DNA was re-precipitated by adding 150 µl of 100% absolute ethanol and resuspended in 50 µl nuclease-free water.
Detection of P. fijiensis and biomass estimation from samples (i.e., pure culture, non-inoculated, and inoculated) were determined by qPCR and calculating sample DNA (Threshold cycle) Ct mean values, to generate an equation Y = −0.265x + 6.0582 from linear regression curve. Y is defined as concentration and X is the Ct values. The reaction mixture and conditions for qPCR remained as described in gene expression assay above, except here we used βtubulin primers for P. fijiensis detection and 100 ng/µl DNA in 12 µl total reaction volume.
In this experiment, three silenced strains of PfFus3 and PfSlt2 were selected as biological replicates including three technical replicates and experiments were repeated thrice.

Statistical Data Analysis
The data were analyzed using GenStat 7th Edition statistical software package employing ANOVA to test significance differences, comparison of means and total mean necrosis in the silenced transformants and wild type.

Confirmation of Pseudocercospora fijiensis Isolates
The pure cultures of P. fijiensis isolated from infected leaves of the banana cultivar 'Nakitembe' were confirmed by PCR analysis. An amplicon of the expected size of 1000 bp was obtained similar to control P. fijiensis collected from CBS-KNAW Fungal Biodiversity Centre in Netherlands, confirming their identity to be P. fijiensis (Supplementary Figure S1). However, P. musae, P. eumusae and non-template used as controls did not show any amplification.

Plasmid Construct Preparation
Two RNAi plasmid constructs were prepared by cloning independently PfFus3 or PfSlt2 into dual promoter pKOIISD1, a silencing vector for fungal pathogen. The promoter Ptrpc drives sense and Pgpd drives antisense sequences of the target genes (i.e., PfFus3 and PfSlt2) to generate dsRNA (Figures 2A,B). Both the plasmid constructs were confirmed by PCR for presence of insert and sequencing for orientation of insert. The orientation of both PfSlt2 and PfFus3 genes was confirmed to be as 5 to 3 (sense strand) for the Ptrpc promoter and 3 to 5 (antisense) for the Pgpd promoter.

Generation and Molecular
Characterisation of Transformed P. fijiensis The P. fijiensis transformants were generated through Agrobacterium tumefaciens mediated transformation and validated by PCR analysis. PCR analysis revealed presence of insert with an expected size (based on plasmid construct map) of 613 bp amplicon in all the PfFus3 transformants tested ( Figure 3A). As seen in Figure 3B, the PfSlt2 transformants showed a product of expected size of 520 bp except for two transformants (lane S5 and S6). The non-transformed control P. fijiensis (WT) did not show any amplified products.
These results confirmed that silencing of PfFus3 and PfSlt2 in P. fijiensis reduced expression by more than 95% (Fus3 range from 96.54 to 99.2%, Slt2 from 98.77 to 99.915%) in comparison to wild-type control. The expression of PfFus3 and PfSlt2 genes was nearly undetectable in most of the transformed P. fijiensis.

Virulence Assays for PfFus3 and PfSlt2 P. fijiensis Transformants on Host Plants
The effect of silencing of PfFus3 and PfSlt2 on pathogenicity of P. fijiensis was determined by inoculating leaves of banana plants of the EAHB cultivar 'Nakitembe' with mycelia of PfFus3 and PfSlt2 transformants and the wild-type control. Plants inoculated with the wild-type control strain developed disease symptoms between 9 and 10 days post inoculation (dpi). However, development of disease symptoms in plants inoculated with PfSlt2 transformants was delayed and was apparent at 15 dpi, while plants infected with PfFus3 transformants showed disease symptoms between 18 and 19 dpi (Figure 5).
Symptom development and disease progression was faster in plants inoculated with wild-type P. fijiensis compared to the PfFus3 and PfSlt2 transformants. This is shown by the higher levels of necrosis in plants infected with wild type at 25 and 45 dpi (Figures 5A,B). There was a significant difference (P < 0.001) in total mean necrosis between the plants infected with wild type and transformants. The percentage mean necrosis at 45 dpi in plant leaves inoculated with wild type strain was 79.6% (166.1), which was three times higher than the necrosis in plant leaves inoculated with transformed strains of PfFus3 10.5% (35.3) and PfSlt2 12.1% (39.4) ( Figure 5B). The PfFus3 transformants were significantly (p < 0.001) less virulent than the PfSlt2 transformants.

Role of PfFus3 and PfSlt2 in Invasive Growth of P. fijiensis
To confirm the role of PfFus3 and PfSlt2 for invasive growth of P. fijiensis, leaf tissues inoculated with the PfSlt2 and PfFus3 transformants along with non-inoculated tissues as negative control and leaf tissues inoculated with wild-type P. fijiensis as a positive control were stained with Lacto phenol cotton blue. No mycelium was observed in non-inoculated tissues ( Figure 6A). However, the wild-type P. fijiensis colonized the leaf tissue as shown by staining with the lacto phenol cotton blue ( Figure 6B). Leaf tissues inoculated with PfFus3 transformants revealed aggregation of mycelia in necrotic leaf tissues without any growth in intercellular spaces ( Figure 6C). Similarly, PfSlt2 transformant-inoculated tissues showed deformed swollen knob mycelia structure with no invasive growth in the intercellular spaces ( Figure 6D).

Measurement of Mycelia Growth of P. fijiensis in Leaf Tissues
The presence of P. fijiensis in the leaf tissue inoculated with wild-type and transformants was confirmed by PCR amplification using primers specific to β-tubulin gene and based on the ITS region of the fungus. The expected size of the amplicons were observed to be similar to the pure culture of P. fijiensis collected from CBS-KNAW Fungal Biodiversity Centre in Netherlands, confirming their identity to be P. fijiensis (Supplementary Figures S3A-C). No amplification was noticed in the non-inoculated leaf tissue.
Colonization or invasive growth of P. fijiensis in tissue to evaluate pathogenicity was further confirmed through biomass quantification. The leaves inoculated with wild-type P. fijiensis showed a high amount of fungal DNA (1.718 ng/g), confirming the presence of high fungal biomass. However, leaf tissues inoculated with silenced PfFus3 and PfSlt2 transformants showed Leaves inoculated with 10% rich medium containing 1% tween 20 acted as non-inoculated control (NIC). Leaves were inoculated with wild-type strain of P. fijiensis (WT) and three independent PfFus3 transformants (PfFus3-5, PfFus3-11, PfFus3-12) and PfSlt2 transformants (PfSlt2-1, PfSlt2-11, PfSlt2-12). (B) Summary representation of total mean necrosis in plant leaves inoculated with WT, transformants PfFus3 and PfSlt2 across different days post inoculation. extremely low P. fijiensis DNA concentrations; this varied between 0.000727 to 0.0250 ng/g and 0.0098 to 0.0322 ng/g, respectively (Figures 7A,B and Table 1). As expected, no fungal DNA was detected in non-inoculated leaf tissue controls while positive control pure cultures of P. fijiensis gave high biomass.
The relationship between the amount of biomass estimate per sample and threshold cycle (Ct) values were also estimated. High Ct values were observed in non-inoculated and silenced PfSlt2, PfFus3 infected samples. However, P. fijiensis pure culture and samples infected with wild-type P. fijiensis showed very low Ct values. This confirms that high Ct values relate to low fungal biomass and low Ct values is an indication of high P. fijiensis biomass ( Table 1).

DISCUSSION
MAP kinase Fus3 and Slt2 pathways are known to be responsible for regulating host penetration, infectious, invasive growth, and pathogenicity of several fungal pathogens including Mycosphaerella graminicola (Mehrabi et al., 2006;Cousin et al., 2007). A key interest of this study was to assess the importance of Fus3 and Slt2 homolog genes of P. fijiensis in infection processes and pathogenicity.
The fragments of MAP kinase genes PfSlt2 and PfFus3 were cloned into the RNAi vector pKOIISD1. The transformed P. fijiensis carrying Slt2 and Fus3 was confirmed by end point PCR assays. We further showed decreased gene expression effects by Relative Quantitation (RQ) using qRT-PCR. The expression of Fus3 and Slt2 genes in P. fijiensis transformants was significantly reduced compared to expression in the wild-type strain. RNAinterference-mediated gene silencing proved to be highly efficient as demonstrated by the nearly undetectable expression of Fus3 and Slt2 in the PfFus3 and PfSlt2 silenced strains. This concurred with previous studies which showed that silencing of endogenous gene Mpg1 and polyketide synthase-like gene in Magnaporthe oryzae using pSilent-1 led to reduction in expression level by 70-90% (Nakayashiki et al., 2005). This study demonstrated that RNAi-mediated gene silencing being a good tool for the study of gene functions in fungal pathogens including P. fijiensis.
The molecular mechanism behind virulence has been studied in Candida glabrata (Miyazaki et al., 2010) and a few fungal plant pathogens such as U. maydis (Mayorga and Gold, 1999), F. proliferatum (Zhao et al., 2012), and Colletotrichum higginsianum (Wei et al., 2016). However, in the ascomycete P. fijiensis the perception of host signaling, penetration, and colonization of the host plant tissues is unknown. Most especially the role of MAP kinase encoding genes Fus3 and Slt2 as pathogenicity factors in P. fijiensis are not known.
Several previous studies identified Fus3 and Slt2 as important genes in regulating pathogenesis factors and pathogenicity in other fungal pathogens (Miyazaki et al., 2010;Luo et al., 2012;Li et al., 2014). Here, we investigated the role of Fus3 and Slt2 in the pathogenicity of P. fijiensis on banana plants, and to the best of our knowledge, this is the first study reporting the importance of Fus3 and Slt2 in invasive growth and pathogenicity of P. fijiensis. The silenced strains of PfFus3 and PfSlt2 showed less virulence characterized by reduced efficiency of plant infection, reduced invasive growth, and fewer necrotic symptoms on "susceptible" EAHB cultivar 'Nakitembe'. Symptom development in plants inoculated with silenced strains of PfFus3 and PfSlt2 was delayed by only few days as compared to plants inoculated with wild-type strain. However, progression of symptoms and colonization of fungi were significantly impaired. This implies that these genes play minor roles in initial symptom development but are critical in regulating fungal development processes like invasive growth of P. fijiensis and plant infection. Previous studies also reported that MAPKs pathways regulate growth and development of other fungal pathogens (Alonso-Monge et al., 2001). Interestingly, there was no evidence of invasive growth in leaf tissues inoculated with PfSlt2 and PfFus3 silenced transformants when tissues were examined microscopically.
Our findings are supported by earlier research which showed that MAP kinase Slt2 and their homologs contribute to invasive growth of Mycosphaerella graminicola (Mehrabi et al., 2006), colonization in host tissue (Rui and Hahn, 2007), and virulence   and PfFus3 transformants  in Phytophthora sojae, Beauveria bassiana, Candida glabrata, Colletotrichum higginsianum (Miyazaki et al., 2010;Luo et al., 2012;Li et al., 2014;Wei et al., 2016). Furthermore, PfSlt2 transformants failed to invasively grow and colonize cells but remained as swollen dormant thalli. It is possible that PfSlt2 transformants failed to adapt to the intercellular environment leading to the failure to develop sufficient turgor pressure to penetrate cells from intercellular space. MAP kinase Slt2 is well known to be responsible for the fungal cell wall integrity (Mehrabi et al., 2006), thus PfSlt2 transformants could have lost the ability to maintain cell wall integrity in the intercellular space. Similarly, Slt2 homolog in Alternaria alternata (AaSlt2) was shown to be critical for cell-wall integrity, responsible for hyphal elongation. AaSlt2 transformants produced globose, swollen hyphae, and failed to elongate (Yago et al., 2011). The formation of swollen hyphal structures in leaves inoculated with PfSlt2 transformants is an indication of retarded growth and hyphal deformation. It is an indication that PfSlt2 regulates key factors critical for pathogenicity of P. fijiensis. The PfFus3 transformants formed undifferentiated massive aggregation of mycelia in necrotic tissues suggesting that fungal development was arrested at an early stage, thereby impairing intercellular hyphal growth. Homologs of Fus3 are known to be essential for infection processes like formation of infection structures, sporulation, invasive and filamentous growth and virulence in other fungal pathogens such as Metarhizium acridum, Colletotrichum higginsianum, and F. proliferatum (Zhao et al., 2012;Jin et al., 2014;Wei et al., 2016). Similar observations were demonstrated by silenced PfFus3, confirming the role of Fus3 in infection processes and pathogenicity of P. fijiensis.
Lastly, low fungal-biomass estimates in leaf tissue infected with PfFus3 and PfSlt2 mutant strains demonstrated that colonization and invasive growth was at least partly regulated by PfFus3 and PfSlt2 in P. fijiensis. There was also a positive correlation between Ct value and biomass estimate, meaning high Ct values indicate low biomass while a low Ct value is an indication of high biomass. This greatly complements screen-house visual pathogenicity assay assessments and Lacto phenol cotton blue staining assays for estimating fungal biomass. A similar study was used to quantify growth of F. graminearum and Magnaporthe oryzae in planta fungal pathogenicity (Qi and Yang, 2002;Horevaj et al., 2011). This study clearly demonstrates that MAP kinase Fus3 and Slt2 pathways are significant contributors to plant infection and growth of P. fijiensis in infected plant tissues. Similarly, several previous studies showed that the MAP kinases Fus3 and Slt2 are involved in the formation of infection structure, spore formation, pathogenic growth or colonization, and pathogenicity. For example Fus3/Kss1 homolog FPK1 in F. proliferatum is involved in hyphal growth, conidiation, and plant infection (Zhao et al., 2012). In Phytophthora sojae PsMPK1 a homolog of Slt2 is known to be required for hyphal growth, zoosporogenesis, cell-wall integrity, and pathogenicity (Li et al., 2014).
Despite efforts to investigate the role of PfSlt2 and PfFus3 in the pathogenicity of P. fijiensis in this study, the molecular interaction between banana and P. fijiensis is not yet well understood. However, previous studies showed that some Musa accessions are highly resistant to the BSD, as the mortality of the host cells occurs fast after infection avoiding and preventing the spread of the pathogen into the rest of the plant (Lepoivre et al., 2003). The resistance in Musa against P. fijiensis is obtained after stomata penetration due to hypersensitive reaction or antifungal activity of phytoalexins and structural analogs (Hoss et al., 2000;Quiñones et al., 2000;Lepoivre et al., 2003). The wild diploid accession 'Calcutta 4' is known to be resistant to BSD because of the expression of pathogenesis related proteins especially during the infection process by P. fijiensis (Rodriguez et al., 2016). 'Calcutta 4' also seems to have some unknown resistance genes that recognizes the PfAVR4 protein, which resulted in hypersensitive reaction upon infiltration of PfAVR4 protein into the banana leaves (Arango Isaza et al., 2016). The disease resistance genes from resistant Musa accessions could be transferred to susceptible bananas and plantain cultivars through conventional breeding or genetic engineering. However, a deeper understanding of the genes involved in fungal resistance process is required.
In summary, the role of MAP kinase Fus3 and Slt2 genes in the pathogenicity and growth in the host plant has been demonstrated in several fungal pathogens, now including P. fijiensis. It has been reported that fungal genes responsible for pathogenicity could be silenced through host-induced gene silencing (HIGs) to develop disease-resistant plants. Example transgenic wheat with resistance against Blumeria graminis and transgenic banana with resistance to fusarium wilt disease (Nowara et al., 2010;Ghag et al., 2014). Therefore, findings from this study suggest that BSD might be controlled by developing transgenic banana-targeting silencing of PfFus3 and PfSlt2 in P. fijiensis through host induced gene silencing.

AUTHOR CONTRIBUTIONS
FO developed the research concept, conducted the experiments, collected and analyzed the data, and wrote the manuscript. GT shaped the research concept and guided and supervised the experiments. L-HC supported vector design, gene cloning, and transformation of fungi. BF and IS shaped the research concept and guided and supervised experiments. JT supported gene expression assay and microscopy. WT and JK provided research supervision and LT shaped the research concept, guided and supervised the experiments, and wrote the manuscript. All authors reviewed and edited the manuscript.

FUNDING
This work was supported by a research grant from Norman E. Borlaug Leadership Enhancement in Agriculture Program (LEAP) to University of California, Davis, United States and Agricultural Biotechnology Support Project II-USAID feed the future to National Banana Research Program, NARO, Uganda.