Pathogenicity associated genes in Fusarium oxysporum f . sp . cubense race 4

CITATION: Sutherland, R., Viljoen, A., Myburg, A. A. & Van den Berg, N. 2013. Pathogenicity associated genes in Fusarium oxysporum f. sp. cubense race 4. South African Journal of Science, 109(5/6), Art. #0023, doi: 10.1590/sajs.2013/20120023.


Introduction
The vascular wilt fungus Fusarium oxysporum is a soil-borne facultative parasite that causes disease in more than 100 plant species, including important agricultural crops. 1 The fungus is a morphospecies that is divided into specialised groups (i.e.formae speciales) according to the hosts they attack, and subdivided into races according to the susceptibility of specific host cultivars. 2Host specificity is believed to have evolved independently in F. oxysporum, and does not necessarily reflect phylogenetic relatedness among pathogenic members of the individual hosts. 2 In F. oxysporum, host specificity has been attributed to mutations in avirulence genes and lateral chromosome transfer that overcome defence responses in the host plant. 3,46][7] Certain pathogenicity genes also encode proteins that are involved in the suppression or disruption of host defence mechanisms. 8,9In F. oxysporum, genes that encode cell wall degrading enzymes (CWDEs), such as endo-polygalacturonase (pg1), exo-polygalacturonase (pgx4), pectate lyase (pl1), xylanase and a plant defence detoxifying enzyme like tomatinase, have been identified in F. oxysporum f. sp.6][17] Signalling genes expressed during pathogenesis have also been identified in Fol (e.g.Fusarium mitogen-activated protein kinase (fmk1)) 15 and F. oxysporum f. sp.cucumerinum (e.g.G protein α subunit (fga1) and G protein β subunit (fgb1)). 18,19Several transcription factors that regulate pathogenicity genes during infection have been discovered in F. oxysporum, such as serine/threonine protein kinases (ste12), 20 a Zn(II)2Cys6-type transcription regulator (fow2) 21 and F. oxysporum ste12 homolog (fost12). 22rains of F. oxysporum pathogenic to bananas are known as F. oxysporum f. sp.cubense (Foc).Three races of Foc are recognised based on their ability to cause disease in a set of different banana cultivars, with Foc race 1 affecting Gros Michel, Silk and Pome bananas and Foc race 2 affecting Bluggoe and other cooking bananas. 23Foc race 4 affects Cavendish bananas, which make up 80% of the world's banana export, as well as Foc race 1 and 2 susceptible bananas. 24Foc race 4 is further subdivided into 'tropical' and 'subtropical' strains.Those belonging to Foc 'tropical' race 4 (TR4) are limited to tropical Asia and northern Australia, while Foc 'subtropical' race 4 (STR4) strains are mostly associated with Cavendish bananas in subtropical countries like South Africa, Australia, Taiwan and the Canary Islands.Foc TR4 is more virulent than Foc STR4, and can infect Cavendish bananas under stressed and non-stressed conditions, whereas Foc STR4 typically only infects bananas after the host has been exposed to stressful environments. 23spite the economic importance of Foc, 24 the mechanisms of pathogenesis to banana are still poorly understood.Additionally, non-pathogenic strains of F. oxysporum are known to infect and colonise the cambium tissue of banana roots, but do not enter the xylem to cause Fusarium wilt.Occasionally, the non-pathogens even protect the banana plant from damage caused by Foc 25,26 and nematodes. 27It is not known why non-pathogenic strains of F. oxysporum are unable to cause disease to banana.Therefore, the objective of this study was to identify gene transcripts that are present in Foc TR4 and Foc STR 4 but absent in non-pathogenic F. oxysporum using cDNA-amplified fragment length polymorphism (cDNA-AFLP) analysis.In addition, a reverse transcriptase-quantitative polymerase chain

Fungal isolates and culture conditions
A total of 27 F. oxysporum isolates were selected for this study.These isolates included Foc STR4 from South Africa, Australia and the Canary Islands, Foc TR4 from Malaysia, Indonesia and Northern Australia, and non-pathogenic F. oxysporum obtained from Cavendish banana roots in South Africa (Table 1).The non-pathogenic F. oxysporum isolates were shown to be non-pathogenic as no internal disease symptoms developed after inoculating banana roots with a spore suspension (1x10 spores/mL) in a hydroponic system. 25,27,32,33All isolates were maintained in 15% glycerol at -80 °C at the Department of Plant Pathology, Stellenbosch University.

RNA extraction
RNA was extracted from fungal mycelia grown in vitro rather than in planta, as insufficient genes of fungal origin were previously detected in the roots of tissue-cultured banana plants 14 d after inoculation with Foc race 4 (1x10 5 spore/mL).The F. oxysporum isolates were first grown on half strength potato dextrose agar (PDA) (19.5 g/L PDA and 10 g/L agar) for 5 days at ±25 °C, and then transferred to liquid minimal medium (MM) without a carbon source to enhance the transcript abundance of pathogenicity genes.34OhZ d After culturing the isolates in MM on a rotary shaker set at 90 rpm for 5 days at 25 °C, the medium was filtered through sterile cheesecloth.The mycelial mass was scraped and frozen in liquid nitrogen, ground to a fine powder with a basic analytical mill (IKA A111, United Scientific (Pty) Ltd., San Diego, CA, USA), and stored at -80 °C until RNA was extracted.
RNA of each isolate was extracted from mycelia using Qiazol (Qiagen, Valencia, CA, USA), quantified with a NanoDrop ND-1000

Isolation of polymorphic fragments and sequence data analysis
After polyacrylamide gels were resolved on the LICOR analyser and scanned with the Odyssey ® infrared imaging system (LICOR), unique bands were identified using Quantity One  37 Functional groups were defined according to the Munich Information Center for Protein Sequences (MIPS) 38 and Gene Ontology (GO) 39 databases.
RT-qPCR reactions were performed in 10-μL volumes containing cDNA template (1:10 dilution), 1 μM of each of the forward and reverse primers and 5 μL DNA Master PLUS SYBR Green mix (Roche Diagnostic).The protocol included 10 min at 95 °C followed by 55 cycles of 10 s at 95 °C, 10 s at 57 °C and 10 s at 72 °C.The amplification process was completed by a melting cycle from 55 °C to 95 °C to assess specificity.The fluorescence reading was recorded at 72 °C at the end of the elongation cycles.The PCR products were analysed by electrophoresis on a 2% agarose gel to verify that a single product of the expected size was produced.All reactions were performed in triplicate with three independent biological replicates and a negative control (no template) for all genes.A standard curve was generated by preparing a dilution series (1:10, 1:100 and 1:1000) for each pathogenicity and reference gene.Gene expression stability (M-value) and pairwise variation (V-values) were determined using Genorm. 40Ct values were imported into qbase PLUS (Biogazelle, Ghent, Belgium) for further analysis.The difference in Ct values was determined statistically by one-way analysis of variance, followed by Tukey's post-hoc analysis; p<0.05 was considered statistically significant.
Several different transcript abundance patterns were detected during cDNA-AFLP gel analysis (Table 4).In the first pattern, high transcript abundance was detected in Foc STR4 with no transcripts detected in Foc TR4 or non-pathogenic F. oxysporum.Examples showing high abundance include TDFs corresponding to the aspartyl-tRNA synthetase (TDF52), galactokinase (TDF57) and O-acetylhomoserine (TDF215).The second transcript abundance pattern showed an increase in transcripts in Foc TR4 with no transcripts detected in Foc STR4 or non-pathogenic F. oxysporum.TDFs that exhibited this pattern were 60S ribosomal protein L2 (TDF12), meiosis induction protein (TDF13), L-aminoadipate semialdehyde dehydrogenase large subunit (TDF64), fatty acid synthase subunit alpha reductase (TDF105), small G-protein Gsp1p (TDF174) and glutamine-dependent NAD+ synthetase (TDF190).In the third pattern, transcripts were detected in Foc STR4 and Foc TR4 with no detection in non-pathogenic F. oxysporum.Examples of these transcripts include

Quantitative verification of cDNA-AFLP
Five reference genes (Table 3) were evaluated for stable expression.The average pairwise variation (V-value) calculated for IDH, G6DH and GAPDH was 0.113, with TEF and TUB showing less stable expression levels (V=0.225).As a result, the reference genes IDH, G6DH and GAPDH were used to normalise the data as suggested by Vandesompele et al. 40 The relative transcript abundance of six genes measured by cDNA-AFLP analysis -encoding a MFS multidrug transporter (TDF9), a L-aminoadipate-semialdehyde dehydrogenase large subunit (TDF64), an aspartyl-tRNA synthetase (TDF52), a chsV (TDF107), a ste12 (TDF214) and rhoI (TDF223) -was compared with results obtained by qRT-PCR (Figure 2).Both cDNA-AFLP and qRT-PCR analyses showed an increased abundance of the MFS multidrug transporter gene in Foc STR4 and Foc TR4 when compared with non-pathogenic F. oxysporum (Figure 2a).When the abundance levels of L-aminoadipate-semialdehyde dehydrogenase large subunit were compared, the cDNA-AFLP analysis demonstrated that the transcript was present in Foc TR4 but absent in the transcript in Foc STR4 and the non-pathogenic F. oxysporum.The qRT-PCR data showed similar levels of transcript abundance in Foc TR4, Foc STR4 and the non-pathogenic F. oxysporum (Figure 2b).cDNA-AFLP analyses showed an increased abundance of transcripts of aspartyl-tRNA synthetase in Foc STR4 compared with Foc TR4 and the non-pathogenic F. oxysporum, while qRT-PCR revealed similar transcript levels among the different isolates (Figure 2c).An increase in transcript abundance of chsV was found in Foc STR4 and Foc TR4 when compared with the non-pathogenic F. oxysporum using both cDNA-AFLP analysis and qRT-PCR (Figure 2d).Transcript abundance profiles determined for ste12 by cDNA-AFLP and qRT-PCR were similar, and showed an increase in Foc STR4 and Foc TR4 compared with the non-pathogenic F. oxysporum (Figure 2e).In the case of rhoI, cDNA-AFLP analysis showed an increase in the number of transcripts in Foc race 4 compared with the non-pathogenic F. oxysporum (Figure 2f).However, qRT-PCR showed a higher number of transcripts in Foc STR4 than in Foc TR4 and the non-pathogenic F. oxysporum.Thus, the transcript abundance patterns measured by qRT-PCR were similar to those measured for the corresponding TDFs analysed using cDNA-AFLP.

Transcript abundance of known pathogenicity genes using qRT-PCR
Foc STR4 and Foc TR4 expressed the pathogenicity genes snf (Figure 3a), frp1 (Figure 3b) and cyp55 (Figure 3c) at significantly higher levels than non-pathogenic F. oxysporum.The transcript abundance of snf was 2.6-fold higher in Foc STR4 than in the non-pathogenic F. oxysporum.
The transcript abundance levels of frp1 were lower in non-pathogenic F. oxysporum isolates than in pathogenic Foc STR4 and Foc TR4 isolates, by 3.6-fold and 2.5-fold, respectively.Snf and frp1 are involved in the degradation of plant cell walls. 17,41Cyp55 had a 1.6-fold higher expression in Foc TR4 compared with Foc STR4, but this difference was not statistically significant.Cyp55 is a nitric oxide reductase involved in the nitrogen response pathway, which is fundamental for pathogenicity.
Fmk1 is responsible for maintaining fungal cell wall architecture and signalling.Fmk1 was expressed significantly more in Foc STR4 than in either Foc TR4 or non-pathogenic F. oxysporum (Figure 3d).Fmk1 expression was 2.9-fold higher in Foc STR4 than in the non-pathogenic No significant homology (68)

Fusarium oxysporum hypothetical protein (65)
Fusarium verticillioides hypothetical protein (18)   Fusarium graminearum hypothetical protein (7)   Hypothetical protein from other fungal species (6)   Protein turnover (8)   Metabolism (13)   Cell division and growth (11)   Cell signalling ( 9) Lipid/fatty acid metabolism (5) Transcription and translation factors ( 6) F. oxysporum (Figure 3d).In addition, transcript abundance of fmk1 was a significant 2.1-fold higher in Foc STR4 than in Foc TR4.Expression of the chloride channel (clc) gene, which controls laccase activity, was also significantly higher in Foc STR4 than in the non-pathogenic F. oxysporum (Figure 3e).In contrast, fow2, the fungal gene involved in regulating pathogenicity-related transcription, was expressed significantly more in Foc TR4 than in the non-pathogenic F. oxysporum, but not significantly more than in Foc STR4 (Figure 3f).There were no significant differences observed in the transcript abundance profiles of the arginine biosynthesis gene (arg1) (Figure 3g) or mitochondrial protein gene (fow1) (Figure 3h).

Discussion and conclusion
The transcriptomes of Foc STR4, Foc TR4 and non-pathogenic F. oxysporum isolates on MM (without carbon source) were visually detected with cDNA-AFLP.More than 3000 TDFs were detected, of which 8% showed differential expression patterns.A total of 3% of these TDFs were putatively involved in pathogenicity.Several fungal gene transcripts that have previously been associated with pathogenicity in other fungal organisms have been identified for the first time in the banana pathogen Foc.These genes include chsV, rhoI, MFS multidrug transporter and stel2.In addition, the genes snf, frp1 and cyp55, which result in diseases of crops other than banana, were more abundantly expressed in Foc STR4 and Foc TR4 than in non-pathogenic F. oxysporum.
The genes chsV and rhoI have previously been associated with pathogenicity in Fol on tomato. 42,43ChsV restricts toxic substances produced by the plant for defence against pathogens, 42 whereas rho1 plays a role in preventing the host plant from recognising the pathogen. 43oth genes, therefore, protect the pathogen against the host's defence response.Because chsV and rho1 showed higher transcript abundances in Foc STR4 and Foc TR4 than in the non-pathogen, we hypothesise that Foc expresses these genes when infecting the xylem vessels of Cavendish bananas to avoid the plant's defence responses.
The transcript abundance of the MFS multidrug transporter was fivefold higher in pathogenic Foc than in the non-pathogen.This family of transporters regulate the movement of sugars, Krebs-cycle metabolites, phosphorylated glycolytic intermediates, amino acids, peptides, osmolites, iron-siderophores, nucleosides and organic and inorganic anions and cations. 44In addition, MFS transporters have been linked to fungal pathogenicity by avoiding toxic compounds produced by the pathogen, or by protection against plant defence compounds. 45FS transporter gene in the ascomycete Verticillium dahlia -a vascular pathogen -is essential for pathogenicity on lettuce plants. 46ith a significantly higher transcript abundance of the MFS multidrug transporter, Foc STR4 and Foc TR4 may possibly protect themselves from toxic substances produced by the plant during defence.
The transcription factor ste12 is important during fungal infection of plant roots where it regulates genes involved in the MAPK cascade. 20,47n a study by Garcia-Sanchez et al. 22 , a ste12-like gene, fost12, showed an increased expression after 12-24 h of infection of bean plants by F.   21 a Zn(II)2Cys6-type transcription regulator involved in pathogenicity in F. oxysporum f. sp.melonis, was significantly higher in Foc TR4 than in the non-pathogen, but there was no significant difference between Foc STR4 and the non-pathogen.
Because Foc TR4 is a more virulent pathogen than Foc STR4, fow2 may assist in the more rapid invasion of root tissue or may be differentially regulated in Foc STR4 and Foc TR4.
Two well-studied pathogenicity genes previously isolated from F. oxysporum that regulate the abundance of CWDEs are snf and frp1. 16,17,41Both snf and frp1 were significantly higher in Foc STR4 and Foc TR4 than in the non-pathogen, which suggests that these genes are important for the Fusarium wilt pathogen to enter the host xylem tissue.As an endophyte, the non-pathogenic F. oxysporum isolates are usually restricted to the root cortex, and do not enter the xylem vessels. 48In contrast, Foc STR4 and Foc TR4 both degrade the xylem cell walls to enter the vascular tissue.
Pathogenicity and cell wall degradation are affected by the enhanced expression of MAP kinases in several fungi, for example Fol, 15 Fusarium graminearum, 49 Magnaporthe grisea 50 and Ustilago maydis 51 .In Fol, fmk1 also aids in root attachment, penetration, invasive growth and increased CWDE activity. 15The significant increase in fmk1 in Foc STR4 and non-significant increase in Foc TR4 compared with non-pathogenic F. oxysporum may explain pathogenesis in the banana Fusarium wilt pathogen, that is by accelerating invasive growth as in other Fusarium species. 15,52Pathogenic Foc isolates are able to colonise both the cortex and the xylem tissue, resulting in severe discoloration of the corm and blocking of the vascular bundles.In contrast, the non-pathogenic strains are restricted to the root cortex, which results in no symptoms developing.The reason that fmk1 did not show a significant increase in transcript abundance in Foc TR4 is not certain, but one possible explanation could be that fmk1 transcripts amplified during pathogenicity at earlier time points were not sampled in this study.Genes expressed during the early time points are either translated into proteins or the RNA is degraded as the half-life of RNA is short and therefore the RNA cannot be detected at later time points.
Cyp55 was more abundant in Foc race 4 than in non-pathogenic F. oxysporum.This gene plays a role in the ability to regulate the nitrogen response pathway, which is essential for pathogenicity. 53Cyp55, a cytochrome P450 gene involved in the reduction of nitric oxide in F. oxysporum, was first characterised by Kizawa et al. 54 The cyp55 gene of F. oxysporum f. sp.vasinfectum has been previously reported to be highly expressed in cotton plants following root inoculation. 55ccases serve as virulence factors in fungal pathogens by playing a role in pigmentation, appressorium formation and protection against toxic phytoalexins. 56qRT-PCR analysis in this study revealed a significant increase in clc transcripts in Foc STR4 compared with the non-pathogen.In Fol, mutations of lcc1, lcc3 and lcc5 had no effect on pathogenicity in tomato plants. 57As six lcc genes have been identified in F. oxysporum, Cañero and Roncero 28 suggested that a mutation in one of them may not necessarily prevent laccase activity, as the other isozymes fulfil their role. 57However, clc mutants showed a decrease in laccase activity with a reduction in virulence to tomato seedlings. 28ncreased clc expression and the role of laccases and chloride transport in the banana Fusarium wilt pathogen may be important pathogenicity determinants.
The cDNA-AFLP technique was useful in differentiating the transcript abundance of genes present in Foc race 4 and non-pathogenic F. oxysporum.However, DNA sequence differences could result in the absence or presence of a TDF not necessarily implicating differential expression.DNA and RNA sequencing could provide significantly better results for identifying pathogenicity genes in Foc, both in STR4 and TR4, especially once the full genome sequence of the Fusarium wilt fungus becomes available.Comparison of the Foc genome with that of other forma speciales of F. oxysporum will elucidate the ability of Foc to infect banana roots.Virulence factors can be studied when the genomes of Foc TR4, a more virulent pathogen, are compared with Foc STR4.Furthermore, the function of putative pathogenicity genes during infection should be investigated by gene knockout studies and RNAi silencing.Knockout mutants would help to identify additional genes required for pathogenicity in Foc race 4.
An in-depth understanding of pathogenicity in Foc is required if novel approaches to disease management are to be developed.We have identified several transcripts in Foc race 4 that are more abundant in the pathogenic strains compared with the non-pathogens.Many of these TDFs have been shown to play a role in host infection and colonisation by other Fusarium spp.These TDFs encode for CWDEs and proteins involved in avoiding toxic substances produced during plant defence.
To establish function, knockout mutants of genes underlying these transcripts need to be generated, and the role of genes such as chsV, rhoI, MFS multidrug transporter, ste12, snf, frp1, cyp55 and fmk1 needs further investigation.With the rapid advancement in molecular techniques in recent years, new strategies for increasing plant resistance against specific Fusarium wilt pathogens can be generated by exploiting the molecular and cellular bases of pathogenicity.

Figure 1 : 7 Volume
Figure 1: Classification of differentially accumulated transcript derived fragments (TDFs) after growth of Fusarium oxysporum f. sp.cubense and nonpathogenic F. oxysporum in minimal medium without a carbon source.A total of 229 TDFs were classified based on the BLASTX homology search on the Broad Institute database.Values in parentheses indicate the number of TDFs found in each category.

Table 1 :
Fusarium oxysporum isolates used for cDNA-amplified fragment length polymorphism and reverse transcription real-time quantitative PCR analysis a Number of the isolate in the culture collection of Altus Viljoen b Fusarium oxysporum f. sp.cubense subtropical race 4 c Fusarium oxysporum f. sp.cubense tropical race 4 d Non-pathogenic Fusarium oxysporum e Isolates with the same designation were grouped for DNA and RNA extraction.

Table 2 :
Primers used for the selective amplification of cDNAamplified fragment length polymorphism fragments in Fusarium oxysporum

Table 3 :
Primer sequences of genes from Fusarium spp.used in reverse transcription real-time quantitative PCR analysis Volume 109 | Number 5/6 May/June 2013 South African Journal of Science http://www.sajs.co.za

Table 4 :
Putative identities of selected genes identified in Fusarium oxysporum using cDNA-amplified fragment length polymorphism analysis based on BLAST results obtained from the Broad Institute database