Identification of Fusarium species and determination of Sequences Type of TEF-gene and Mating Type of F. oxysporum
Different Fusarium species have been associated with vanilla plants, however, only F. oxysporum is reported as pathogenic at different levels [19, 21], suggesting a high genetic diversity of pathogenic endophyte strains [18, 37]. A total of 46 strains were used in this study. From these, nine were JAGH strains, 12 were NAY strains and all other 25 were isolated from Veracruz, 11 pathogenic and 14 non-pathogenic. BLAST of TEF-1α sequence in the FUSARIUM-ID database allowed the identification of six Fusarium species: F. proliferatum (JAGH1), F. solani (R21), F. fujikuroi (JF62), F. lactis (NAY8), F. pseudocircinatum (V4M, V12M) and F. oxysporum (all other strains). To our knowledge, this is the first report that associates F. lactis to Vanilla sp. 14 ST were identified in F. oxysporum associated to vanilla, with different levels of similarity to reference sequences, this is a relatively high number of ST when compared with other crops such as tomato [14] and cyclamen [15].
Both Mating Type idiomorph was observed, even in strains belonging to the same Sequence Type, demonstrating that the origin of these strains is not clonal. MAT gene has been associated with genetic diversity into a population of asexually reproducing Fusarium [38] and with the differentiation of some physiological races of F. oxysporum [39]. Information about ST, Fusarium species identification, and Mating Type idiomorph are summarized in Table 3.
Table 3
Identification of Fusarium species relationed with Vanilla sp. and determination of Sequence Type and Mating Type of Fusarium oxysporum. ND = Non-Determinated.
Strain | Fusarium ID | Similarity (%) | Mating Type | Sequence Type |
BC1 | F. oxysporum NRRL22550 | 92.93 | MAT-2 | ST1 |
JAGH11 | F. oxysporum NRRL22550 | 93.01 | MAT-1 | ST1 |
NAY12 | F. oxysporum NRRL26449 | 97.88 | MAT-1 | ST2 |
V8M | F. oxysporum NRRL26449 | 99.26 | MAT-2 | ST2 |
JAGH5 | F. oxysporum NRRL28367 | 88.07 | MAT-1 | ST3 |
JAGH5-1 | F. oxysporum NRRL28406 | 97.71 | MAT-1 | ST4 |
NAY22 | F. oxysporum NRRL28406 | 98.06 | MAT-1 | ST4 |
JF22 | F. oxysporum NRRL31418 | 99.83 | MAT-1 | ST5 |
A3 | F. oxysporum NRRL32885 | 99.84 | MAT-1 | ST6 |
28R1 | F. oxysporum NRRL32885 | 99.38 | MAT-1 | ST6 |
A31 | F. oxysporum NRRL32885 | 99.84 | MAT-1 | ST6 |
A32 | F. oxysporum NRRL32885 | 99.69 | MAT-1 | ST6 |
NAY6 | F. oxysporum NRRL32885 | 99.52 | MAT-1 | ST6 |
NAY7 | F. oxysporum NRRL32885 | 99.37 | MAT-2 | ST6 |
V3 | F. oxysporum NRRL36251 | 99.84 | MAT-1 | ST7 |
JAGH4 | F. oxysporum NRRL36284 | 99.17 | MAT-1 | ST8 |
JAGH10 | F. oxysporum NRRL36284 | 96.62 | MAT-1 | ST8 |
JF5 | F. oxysporum NRRL36284 | 97.24 | MAT-1 | ST8 |
V6M | F. oxysporum NRRL36284 | 98.51 | MAT-1 | ST8 |
NAY13 | F. oxysporum NRRL38290 | 97.26 | MAT-2 | ST9 |
NAY20 | F. oxysporum NRRL38290 | 97.64 | MAT-1 | ST9 |
3R1 | F. oxysporum NRRL38290 | 97.96 | MAT-1 | ST9 |
25R2 | F. oxysporum NRRL38290 | 98.12 | MAT-1 | ST9 |
NAY1 | F. oxysporum NRRL38290 | 98.06 | MAT-1 | ST9 |
NAY3 | F. oxysporum NRRL38290 | 98.12 | MAT-1 | ST9 |
NAY9 | F. oxysporum NRRL38290 | 98.07 | MAT-1 | ST9 |
NAY14 | F. oxysporum NRRL38290 | 98.1 | MAT-1 | ST9 |
NAY21 | F. oxysporum NRRL38290 | 98.27 | MAT-2 | ST9 |
R27 | F. oxysporum NRRL38290 | 98.11 | MAT-1 | ST9 |
B3 | F. oxysporum NRRL38514 | 99.68 | MAT-1 | ST10 |
JAGH2 | F. oxysporum NRRL38514 | 99.84 | MAT-1 | ST10 |
A33 | F. oxysporum NRRL38597 | 100 | MAT-1 | ST11 |
BR1 | F. oxysporum NRRL38597 | 99.68 | MAT-1 | ST11 |
JF3 | F. oxysporum NRRL39464 | 95.88 | MAT-1 | ST12 |
JF12 | F. oxysporum NRRL39464 | 100 | MAT-1 | ST12 |
R115 | F. oxysporum NRRL39464 | 100 | MAT-1 | ST12 |
JAGH9 | F. oxysporum NRRL40180 | 100 | MAT-1 | ST13 |
JAGH3 | F. oxysporum NRRL40180 | 98.42 | MAT-1 | ST13 |
A8 | F. oxysporum NRRL45881 | 99.84 | MAT-2 | ST14 |
V9M | F. oxysporum NRRL45881 | 100 | MAT-1 | ST14 |
JAGH1 | F. proliferatum Br-1 | 99.08 | ND | ND |
R21 | F. solani NRRL32755 | 99.24 | ND | ND |
JF62 | F. fujikuroi complex NRRL31418 | 93.3 | ND | ND |
NAY8 | Fusarium lactis NRRL31629 | 98.85 | ND | ND |
V4M | F. pseudocircinatum NRRL31631 | 97.07 | ND | ND |
V12 | F. pseudocircinatum NRRL31631 | 97.39 | ND | ND |
Phylogenetic Analysis Of Tef-gene
Unweighted Parsimony analysis recovered 343 most parsimonious trees (L = 624; CI = 73; RI = 69). To summarize all the information a Strict Consensus Tree was performed and the resultant topology is presented (Fig. 1a). There are some clades composed only by pathogenic strains that are well bootstrap and jackknife supported. A clade composed of five strains, four pathogenic and one non-pathogenic (arrowed in Fig. 1a) suggests a lost pathogenicity event.
Polyphyletic nature of different formae speciales in FOSC has been suggested, some examples are F. oxysporum f. sp. melonis [40], F. oxysporum f. sp. phaseoli [41], F. oxysporum f. sp. apii [42] and many others. Pinaria et al. [20] showed that F. oxysporum f. sp. vanillae is polyphyletic in FOSC, based on TEF and mtSSU genes phylogeny. They observed that F. oxysporum f. sp. vanillae is present in the three clades proposed by O´Donell et al. [43] and in the two phylogenetic species recognized by Laurence et al. [25]. However, they suggest that only Indonesian strains of the pathogen have a polyphyletic origin while Mexican ones are monophyletic. Optimizations based on a parsimony approach allowed us to observe 11 points of pathogenicity origin inside the phylogeny, however, ancestral states are not well resolved by the optimization. A potential pathogenic ancestral state is present in the internode where the separation of F. oxysporum from other species is observed. Some clades are composed of strains from the two states, Veracruz and Nayarit, thus this means that some lineages are present in both places.
The number of clades in our TEF phylogeny was lower than the number of ST. However, a parsimony optimization allowed us to confirm that the pathogenic ability of some endophytes to vanilla, or F. oxysporum f. sp. vanillae, is a polyphyletic trait among FOSC in the Mexican strains of the pathogen [23] and are not monophyletic as proposed by Pinaria et al [20]. Interestingly, some pathogenic strains as A3 surged in a clade where all members and the theoretical ancestor are non-pathogenic. Conversely, BC1 strain is the only non-pathogenic strain in a pathogenic clade (Fig. 1a). This phenomenon can be explained on the basis of Horizontal Gene Transfer, a mechanism that has been demonstrated to play a very important role in the evolution of plant pathogenic fungi [44].
Genetic Differentiation And Diversity
Results from the AMOVA with the two populations (pathogen and non-pathogen) shows that all variation is observed and explained within groups and not among them. A genetic differentiation based on our microsatellite data among pathogen and non-pathogen endophytes to vanilla is not possible. For comparison of diversity among the groups a Shannon index was calculated for every locus and compared with a chi-square, showing a significant difference among the pathogenic and non-pathogenic groups in locus FoFA4 (p = 0.037). The results are summarized in Table 4.
Table 4
Shannon´s Diversity Index comparing between pathogenic and non-pathogenic groups. Significant differences are represented by asterisk (p < 0.05).
| FoAB11 | FoAD12 | FoAG11 | FoAG11 | FoAG11 | FoDC5 | FoDD7 | FoDE7 | FoDF7 | FoFA4 |
Pathogenic | 0.726 | 1.594 | 2.539 | 2.500 | 2.437 | 1.844 | 2.563 | 0.314 | 1.693 | 1.704 |
Non-Pathogenic | 0.994 | 1.871 | 2.600 | 2.831 | 2.636 | 2.033 | 2.242 | 0.567 | 1.989 | 2.186* |
Clustering UPGMA method showed that some pathogenic strains are very close genetically, however, there is not a unique cluster grouping all pathogenic endophytes (Fig. 1B). A comparison among TEF-1α phylogenetic tree and microsatellite data UPGMA revealed that some pathogenic genotypes are similar by convergent evolution. A Correspondence Analysis showed that pathogenic endophytes group are a subgroup inside the non-pathogenic one, supporting conclusions observed in UPGMA dendrogram (Fig. 1B). Different diversity indexes were estimated showing low variation among pathogenic and non-pathogenic groups (Table 5). For determination of the most likely number of genetic groups (k) in our strains collection, a Bayesian approach was applied. Our results showed that the most probable number of groups is k = 3 based in a total variation on logK (∆k = 4.6218).
For the detection of genetic differentiation between pathogenic and non-pathogenic endophytes, molecular data from microsatellite were analyzed. The variation of these markers was explained better by a stepwise model than with a multiple-step model because very low variations were observed in sister strains of F. oxysporum [45]. Dendrogram UPGMA shows that there is not an exclusive cluster that grouped all pathogenic strains. This result is consistent with the polyphyletic distribution of this forma special in the FOSC, because if a group is monophyletic a very similar pattern of repetition of motifs is predicted, as in F. oxysporum f. sp. ciceris [32]. However, a very high diversity grade was observed in contrast to other studies [46] where a very slow diversity of F. oxysporum f. sp. vanillae isolated from India was found.
A comparison between the phylogenetic tree and dendrogram shows that some strains with a genotype very similar are the result of convergent evolution. This has been observed in other systems with other pathogenic fungi to plants, insects and humans and is explained by a co-evolutionary process [47]. The coevolution process of F. oxysporum f. sp. vanillae with vanilla in Mexico is particularly interesting given the ancestral distribution of this plant. All genotypes of vanilla cultivated around the world have their single-origin center in Papantla, Mexico [48], but genotypes in Papantla have multiple origins from southeast Mexico and Mesoamerica [49].
Diversity indexes showed that there is a higher diversity in the non-pathogenic endophytes than in the pathogenic ones. The Shannon index has been used successfully to evaluate diversity among phytopathogenic fungi using molecular markers [50]. The higher diversity in non-pathogenic strains suggests that pathogenic fungi are a small sub-group that has acquired abilities to infect but belong to the non-pathogenic group.
The CA analysis (Fig. 2) supports the inclusion of pathogenic strains in the non-pathogenic group. Inami et al. [51] determined that F. oxysporum f. sp. lycopersici share a common ancestor with non-pathogenic strains obtained from wild tomatoes in Peru, thus, the pathogenic strains are a subgroup of a bigger group of non-pathogenic strains. Evanno et al. [35] method estimated that the most likely number of genetic groups of our collection of endophytes is k = 3, suggesting that pathogenic could be composed by two groups.
In conclusion, genetic differentiation among pathogenic (F. oxysporum f. sp. vanillae) and non-pathogenic endophytes to vanilla are not obtained when microsatellite data are used for determination. However, these endophytes have a great genetic diversity observed in the number of ST of TEF gene, diversity indexes and clustering method of microsatellite variation. Phylogeny and microsatellite variation support the polyphyletic ability to infect vanilla trough the FOSC [20, 23, 32].
Amplification of SIX genes effectors
All pathogenic and non-pathogenic strains were evaluated for the detection of all currently know SIX genes effectors. Proofs for each pair of primers were repeated three times with DNA extractions from different Petri dishes but no amplification was obtained. The positive control gene TEF-1α was always amplified, showing the viability of DNA samples and PCR reactive. For the SIX1 gene, a strain of F. oxysporum f. sp. cubense was used as positive control and a unique product with expected size was obtained, showing the viability of thermal conditions and primer design.
Ma et al. [24] demonstrated that the F. oxysporum genome can be divided into two different parts, the Core Genome consisting of the chromosomes conserved in the Species Complex and inherited vertically; and Specific Lineage Regions that are supernumerary chromosomes with the ability to confer pathogenic features to non-pathogenic strains (Horizontal Gene Transference). In these chromosomes the SIX genes are carried, which have been observed to play an essential role in susceptibility/resistance reaction in the F. oxysporum f. sp. lycopersici – tomato system [27, 52]. The role of SIX genes in the plant-pathogen interaction is of the pathogenicity effector, disrupting the immunity triggered by MAMP´s in the plant and triggering newly the susceptibility, according to with zigzag model [26].
These SIX genes are found in other formae speciales, but their role in the interaction with their hosts are not clarified [53, 54]. In recent studies, evidence for the horizontal transference of these genes was obtained showing that SIX genes are a monophyletic group depending on their host [25, 55]. Contrary to waited, amplification for SIX genes was not obtained from all endophytes to vanilla strains, pathogenic and non-pathogenic. Positive controls (TEF gene and an F. oxysporum f. sp. cubense strain) were amplified with a unique product at the expected size. Primers used in this study have been used in many others studies with different formae speciales [25, 56–59], even in natural ecosystems were they allowed detection of some SIX genes at very low levels directly from the uncultivated soil [60]. SIX genes sequences are highly conserved and few variations at nucleotides are responsible for host specificity [61, 62].
We offer three possible explanations for the lack of SIX genes amplifications: 1) regions to annealing for primers in our strains are very variable and recognition is impossible having no amplification as result, 2) the responsible genes for infection and specificity are different to what were evaluated in this study. Recently, it has been determined that F. oxysporum f. sp. radicis-cucumerinum, a root rotting pathogen, do not have SIX1 in the LS genome, but SIX6 and other less common SIX genes are present [63], moreover, many others putative SIX genes with potential effector activity are found on the basis of the location of a miniature Impala transposon (mimp) associated to SIX genes, with evidence also of horizontal transference [64]; and 3) interaction between plant and pathogen is located at the beginning of the zigzag model and no effectors are needed for establishment of the disease. The presence of a MAMP has been associated with slow and moderate virulent strains of F. oxysporum f. sp. vanillae, while absence is associated with highly virulent ones [21, Luna-Rodríguez et al., submitted for publication]. Moreover, Koyyappurath et al. [18] observed that the damage on vanilla caused by a pathogenic endophyte is limited to the first layers of the root, not to the vascular system, proposing to change the name of the forma special to F. oxysporum f. sp. radicis-vanillae. In fact, resistance to the disease is observed because a high level of lignification on the root hypodermis is present [65], while the activity of SIX genes has been observed when a pathogenic strain colonizes the vascular system only [66].
In conclusion, a genetic comparison among pathogenic and non-pathogenic endophytes using microsatellite data does not allow discrimination of both groups, however, phylogeny and genetic distances suggest that some pathogenic genotypes are similar because of convergent evolution. There is evidence for a great genetic diversity of endophyte F. oxysporum in the vanilla crop, based on microsatellite, TEF-1α sequences types and MT idiomorphs. Amplification of the currently known SIX genes was not possible, further studies are necessary for this field.