Diversity of Colletotrichum Species Associated with Olive Anthracnose Worldwide

Olive anthracnose caused by Colletotrichum species causes dramatic losses of fruit yield and oil quality worldwide. A total of 185 Colletotrichum isolates obtained from olives and other hosts showing anthracnose symptoms in Spain and other olive-growing countries over the world were characterized. Colony and conidial morphology, benomyl-sensitive, and casein-hydrolysis activity were recorded. Multilocus alignments of ITS, TUB2, ACT, CHS-1, HIS3, and/or GAPDH were conducted for their molecular identification. The pathogenicity of the most representative Colletotrichum species was tested to olive fruits and to other hosts, such as almonds, apples, oleander, sweet oranges, and strawberries. In general, the phenotypic characters recorded were not useful to identify all species, although they allowed the separation of some species or species complexes. ITS and TUB2 were enough to infer Colletotrichum species within C. acutatum and C. boninense complexes, whereas ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH regions were necessary to discriminate within the C. gloesporioides complex. Twelve Colletotrichum species belonging to C. acutatum, C. boninense, and C. gloeosporioides complexes were identified, with C. godetiae being dominant in Spain, Italy, Greece, and Tunisia, C. nymphaeae in Portugal, and C. fioriniae in California. The highest diversity with eight Colletotrichum spp. was found in Australia. Significant differences in virulence to olives were observed between isolates depending on the Colletotrichum species and host origin. When other hosts were inoculated, most of the Colletotrichum isolates tested were pathogenic in all the hosts evaluated, except for C. siamense to apple and sweet orange fruits, and C. godetiae to oleander leaves.


Introduction
The olive (Olea europaea L. subsp. europaea) is the most important tree crop worldwide, covering over 11 million hectares, more than the whole of stone fruit species [1]. Most olives are grown near the Mediterranean Sea, especially Spain, Italy, Greece, Tunisia, and Portugal. The excellent adaptation of the olive plant to different conditions has prompted a spread of olive farming to countries where it is not a traditional crop, such as Australia, Brazil, or China [2,3]. Due to this expansion through different areas, the olive plant has been gradually exposed to new pathogens. This situation is particularly striking in olive anthracnose, the most important disease of the fruit.
Olive anthracnose caused by numerous Colletotrichum species causes dramatic losses of fruit yield and oil quality during epidemic years [4][5][6][7]. The pathogen infects through the seasons, but disease symptoms appear at the beginning of ripening when the color of the fruit changes from green to black [6,8]. Typical symptoms are depressed, round, and ochre or brown lesions leading to fruit rot with great orange conidial masses (Figure 1a,b), the "soapy olive" syndrome that gives its name to this disease in Spanish [9]. Subsequently, fruit are mummified (Figure 1c,d) when the temperature falls, relative humidity increases in late autumn-winter, and most of them fall to the soil [6,10]. The pathogen also causes the dieback of olive branches via phytotoxins (Aspergillomarasmine A) produced by the fungus in the rotten fruit (Figure 1e,f) [8,11,12]. Likewise, the pathogen can cause the blight of olive inflorescences, mainly when mummies remain attached to the tree canopy during the flowering [8,12,13]. In addition, the pathogen may act as a secondary invader of injured tissue and can also survive as endophyte or saprophyte. The ability to survive and multiply in the absence of symptoms may explain why anthracnose fungi often cause unexpected crop losses in olives [12,14]. plant has been gradually exposed to new pathogens. This situation is particularly in olive anthracnose, the most important disease of the fruit.
Olive anthracnose caused by numerous Colletotrichum species causes drama of fruit yield and oil quality during epidemic years [4][5][6][7]. The pathogen infects the seasons, but disease symptoms appear at the beginning of ripening when the the fruit changes from green to black [6,8]. Typical symptoms are depressed, rou ochre or brown lesions leading to fruit rot with great orange conidial masses (Figu the "soapy olive" syndrome that gives its name to this disease in Spanish [9 quently, fruit are mummified (Figure 1c,d) when the temperature falls, relative h increases in late autumn-winter, and most of them fall to the soil [6,10]. The patho causes the dieback of olive branches via phytotoxins (Aspergillomarasmine A) p by the fungus in the rotten fruit (Figure 1e,f) [8,11,12]. Likewise, the pathogen c the blight of olive inflorescences, mainly when mummies remain attached to the opy during the flowering [8,12,13]. In addition, the pathogen may act as a secon vader of injured tissue and can also survive as endophyte or saprophyte. The a survive and multiply in the absence of symptoms may explain why anthracno often cause unexpected crop losses in olives [12,14]. The causal agent of olive anthracnose was described for the first time in Por de Almeida [15] as Gloeosporium olivarum. Subsequently, this species was reclas Colletotrichum gloeosporioides (anamorph of Glomerella cingulata) after reviewing t sporium genus by von Arx [16]. Later, the species C. gloeosporioides was consider erogeneous species complex affecting about 300 plant species [17]. Currently, o epithets are listed in Mycobank [18] under Colletotrichum, which comprises 248 The causal agent of olive anthracnose was described for the first time in Portugal by de Almeida [15] as Gloeosporium olivarum. Subsequently, this species was reclassified as Colletotrichum gloeosporioides (anamorph of Glomerella cingulata) after reviewing the Gloeosporium genus by von Arx [16]. Later, the species C. gloeosporioides was considered a heterogeneous species complex affecting about 300 plant species [17]. Currently, over 1000 epithets are listed in Mycobank [18] under Colletotrichum, which comprises 248 accepted species, most of them grouped into 14 species complexes [19].
As with many other crops, olives can be affected by a wide range of Colletotrichum species [20][21][22]. To date, a total of 14 Colletotrichum spp. have been associated with olive anthracnose over the world. These species belong to three Colletotrichum complexes: C. acutatum, C. boninense, and C. gloeosporioides [5,13,23,24]. Among them, the species C. acutatum sensu stricto (from now on C. acutatum), C. fioriniae, C. godetiae, C. nymphaeae, C. rhombiforme, and C. simmondsii, all of them belonging to the C. acutatum species complex, are currently considered the major pathogens of this genus [5,13,24]. While several Colletotrichum species can be found in an olive-growing area, there is usually one dominant one and some secondary [7]. For example, the species C. nymphaeae is dominant in the olive orchards of Portugal [4,25], while C. godetiae is the prevalent species in several Mediterranean countries such as Greece, Montenegro, and Spain [6,7,24,26].
Furthermore, other species belonging to the C. gloeosporioides species complex have been described in Australia (C. siamense and C. theobromicola), Montenegro (C. queenslandicum), and Uruguay (C. alienum and C. theobromicola) as associated with olive fruit [13,23]. Nevertheless, the pathogenicity of several of these species (C. aenigma, C. cigarro, C. karstii, C. queenslandicum, and C. siamense) on olive fruit is still uncertain, suggesting that their role in the fruit infection could be secondary [5].
From the first descriptions of C. acutatum sensu lato and C. gloeosporioides s. l. as the causal agents of the olive anthracnose in 1999 [33], many studies focused on aetiology have been conducted mainly in Italy and Portugal, generating relevant knowledge about the diversity of Colletotrichum species associated with the disease [4,23,24,28,34]. In the case of Spain, the etiological knowledge about olive anthracnose is much more limited and suggests that there are two prevalent species, with C. godetiae being dominant [7]. Therefore, more etiological studies are necessary to elucidate the diversity of Colletotrichum species involved in the olive anthracnose. Furthermore, differences in virulence among Colletotrichum species, and isolates of the same species, have also been described in different woody crops [5,[35][36][37][38]. However, the number of species tested is still slight, and broader studies on the pathogenicity of Colletotrichum spp. causing olive anthracnose are necessary.
During these last two decades, many Colletotrichum isolates from olive fruit showing anthracnose symptoms in orchards located in the Iberian Peninsula, Spain, and Portugal were studied in our laboratory. Both Spain and Portugal produce around 65% of the global supply of olive oil [1]. In parallel, many Colletotrichum isolates from olives or other susceptible hosts to anthracnose from Australia, Brazil, California, Greece, Portugal, Italy, Tunisia, and Uruguay have been studied in close collaboration with different international research groups. Thus, the main goal of the present study was to characterize a vast collection of Colletotrichum isolates obtained from olives and other hosts showing anthracnose symptoms in Spain and other olive-growing countries. To this end, in the present study, we combined different techniques for characterization of the Colletotrichum population affecting olive fruit around the world, including morphological characteristics [4,5,16,26,31], physiological traits including tolerance to fungicides and enzymatic activity [4,5,7,33], and molecular tools [4,19,23,27]. Therefore, the specific objectives of this study were (i) to obtain a wide collection of Colletotrichum isolates representative of the geographic origin described above and from different hosts showing anthracnose symptoms; (ii) to characterize them based on phenotypic (colony and conidial morphology; and benomyl-sensitive and casein-hydrolysis tests) and molecular characters (multilocus alignments and phylogenetic analyses); (iii) to determine the pathogenicity of Colletotrichum isolates to olive fruit; and (iv) to evaluate cross-pathogenicity on different Colletotrichum hosts. Elucidating the biodiversity of Colletotrichum species causing olive anthracnose is essential for a better understanding of the aetiology and epidemiology of the most critical fruit disease of this legendary crop.

Collection of Fungal Isolates
Olive fruit samples showing symptoms of anthracnose were collected from many commercial orchards from 1998 to 2016. Symptomatic fruit were collected from different provinces across Spain, emphasizing the orchards located in the Andalusia region of southern Spain, the world's leading olive-producing region. Many other samples were collected from commercial orchards situated in southern Portugal, where olive anthracnose is endemic [34,39]. Isolations were made from affected fruit with the typical anthracnose lesions. Diseased fruit were surface disinfested with commercial bleach (Cl at 50 g L -1 ) at 10% (v/v) in sterile water for 1 min, and air-dried on sterile filter paper for 30 min. Affected tissues were cut with a sterile scalpel and plated on potato dextrose agar (PDA) (Difco Laboratories ® , Detroit, MI, USA) acidified with lactic acid (2.5 mL of 25% [v/v] per liter of medium) to minimize bacterial growth (APDA). When the affected fruit tissues showed abundant pathogen sporulation, masses of conidia were removed using a sterile needle and cultured in Petri dishes on APDA. Petri dishes were incubated at 23 ± 2 • C under a 12-h daily photoperiod of cool fluorescent light (350 µmol m -2 s -1 ) for 5 days. Singlespore isolates were prepared before use in further experiments using serial dilutions [40]. Moreover, Colletotrichum isolates recovered from olive fruit showing anthracnose symptoms in Australia, Brazil, California (the USA), Greece, Italy, Tunisia, and Uruguay were also included in this study as collaboration with several international research groups from those major olive-growing regions of the world (Table 1). All the isolates were maintained on colonized PDA into sterile plastic tubes with sterile paraffin oil (Panreac Química SA, Barcelona, Spain) at 4 • C in darkness for long-term storage in the fungal collection of the Department of Agronomy at the University of Cordoba (UCO; Spain).

Colonies and Conidial Morphology
Thirty-eight representative isolates belonging to C. acutatum (27 isolates), C. boninense (two isolates), and C. gloeosporioides (nine isolates) species complexes (Table 1) were used to study mycelium colony and conidium morphology. To this end, all the isolates were grown on PDA (Difco ® Laboratories, Detroit, MI, USA) for two weeks at the same incubation conditions described above. There were three replicated Petri dishes per isolate.
Characteristics of mycelia (texture, density, color, and zonation) were recorded by visual observations on 7-day-old colonies [41]. For all isolates, color was determined using a color scale [42]. For conidial measures, conidial masses removed from the margin of 10-day-old colonies were placed on slides with a drop of 0.005% acid fuchsine in lactoglycerol (1:1:1 lactic acid, glycerol, and water) and covered with a coverslip. For each isolate and replicate Petri dish, the size, and the shape of 50 conidia were measured utilizing a Nikon Eclipse 80i microscope (Nikon Corp., Tokyo, Japan) at 400× magnification. The conidial size was determined by measuring its length and width, and the length/width ratio was calculated. According to their shape, conidia were classified into three categories: (0) Conidia with two rounded ends (ellipsoid); (1) Conidia with one rounded end and the other acute (clavate); and (2), Conidia with two acute (sharp) ends (fusiform). Data were expressed as a percentage (%) of each type of conidium.   Panama d,e,f Representative Colletotrichum spp. isolates selected for morphological characterization with regards on mycelium and conidium characteristics; in vitro sensitivity tests to determine sensitivity against benomyl and ability to hydrolyse casein; and pathogenicity test to olives or to other hosts, respectively.

Benomyl-Sensitive Assay
Twenty-seven representative isolates belonging to C. acutatum (16 isolates), C. boninense (two isolates), and C. gloeosporioides (nine isolates) species complexes were used to evaluate their sensitivity to benomyl by in vitro sensitivity assay (Table 1). Based on previous studies [4,43], and our preliminary trials using isolates of C. acutatum and C. gloeosporioides species complexes, we determined a threshold of 5 µg of benomyl per milliliter to differentiate sensitive and tolerant isolates to this fungicide. Thus, mycelial plugs (5 mm in diameter) obtained from the margins of 7-day old actively growing colonies on PDA were transferred to Petri dishes with PDA amended with 5 µg mL −1 of benomyl (benomyl 50%, WP, Adama Agriculture, Madrid, Spain). Mycelial plugs of each isolate were plated on non-amended PDA as control. All Petri dishes were incubated under the described conditions. There were three replicated Petri dishes per isolate and treatment (benomyl and control), and the experiment was conducted twice.
The evaluation was performed at 7 days, measuring the largest and smallest diameters of each colony. For each isolate, the inhibition percentage (%) was calculated by comparing the growth on PDA and on PDA amended with benomyl. Data from repetitions of the experiment were combined after checking for homogeneity of variances of the experimental error of the two replicated experiments by the F test. Subsequently, analysis of variance (ANOVA) was conducted using a randomized complete block design with the two repetitions of the experiment as blocks, fungal isolate as the independent variable, and inhibition percentage as the dependent variable. Mean comparisons were made using Tukey's honestly significant difference (HSD) test [44]. Data were analyzed using Statistix 10 [45].

Casein-Hydrolysis Assay
The 27 Colletotrichum isolates, previously studied according to their sensitive/tolerance benomyl fungicide, were also characterized according to their ability to hydrolyse the casein. Thus, the 27 Colletotrichum isolates were transferred, as described previously, to hydrolyse casein medium (CHM). We formulated the CHM media using a 15% milk powder solution (Sveltesse Nestle ® , Esplugues de Llobregat, Barcelona, Sapin) in deionized water sterilized at 120 • C for 15 min, and 20 mL of the sterile milk solution was added to 980 mL of sterilized Water Agar (WA; Biokar-Diagnostics, Allonne, France) before solidification (around 50 • C) and homogenized for 2 min using a magnetic rotor (Agimatic-N, JP-Selecta, Barcelona, Spain).
A 5-mm diameter plug of each Colletotrichum isolate was plated to Petri dishes with CHM and incubation for 5 days as described for the benomyl sensitivity assay. For each isolate, mycelial plugs were plated on non-amended PDA as control. We visually determined the presence or absence of the hydrolysis halo surrounding the Colletotrichum colony growing on CHM media. There were three replicated Petri dishes per isolate and treatment (milk powder suspension and control), and the experiment was conducted twice. Six genomic areas, 5.8S nuclear ribosomal gene with two flanking internal transcribed spacers (ITS), beta-tubulin (TUB2), actin (ACT), partial sequences of the chitin synthase 1 (CHS-1), histone 3 (HIS3), and a 200-bp intron of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH), were amplified and sequenced. For that, the following primer pairs were correspondingly used: ITS4 and ITS5 [46], Bt-2a and Bt-2b [47], ACT-512F and ACT-783R [48], CHS-354R and CHS-79F [48], CYLH3F and CYLH3R [49], and GDF1 and GDR1 [50]. Additionally, to infer the identity of fungal isolates belonging to the C. gloeosporioides complex, the intergenic region between Apn2 and Mat1-2 genes (ApMat) was also amplified and sequenced with the primer pair AMF1 and AMR1 [51].
PCR amplifications were performed in a MyCycler TM Thermal Cycler (BIO-RAD) in a total volume of 25 µL. All PCRs mixture contained 5 µL of 5×MyTaq reaction buffer, 0.13 µL of MyTaq DNA polymerase (Bioline), and 20 ng of genomic DNA template. Additionally, 0.2 µM of each primer was added for the ACT, CHS-1, HIS3, and GADPH PCRs, and 0.4 µM of each primer for ITS, TUB2, and ApMat PCRs. Negative control was included in all PCRs using ultrapure water instead of DNA. The PCRs cycling programs were conducted as follows: an initial denaturation at 95 • C for 5 min, followed by 40 cycles of 95 • C for 15 s, annealing for 15 s and 72 • C for 10 s, and a final extension at 72 • C for 7 min. The annealing temperatures used were: 48 • C for ITS, 52 • C for GAPDH, CHS-1, HIS3, TUB2, and ACT, and 55 • C for ApMat. All PCRs were stopped at 4 • C.
Amplification products were checked by electrophoresis in 1.7% (wt/vol) agarose gel stained with RedSafe (Intron Biotechnology, Sagimakgol-ro Joongwon-gu Seongnam-Si Korea, Republic of South) and visualized under ultraviolet light. DNA gTP-Ladder (gTPbio) was used for electrophoresis as DNA size markers. Single-band products were purified using MEGAquick-spin TM Total Fragment DNA Purification kit (INTRON Biotechnology), following the manufacturer's instructions. Subsequently, purified PCR products were sequenced in both forward and reverse directions by the Central Service Support Research (SCAI) at the University of Córdoba (Spain).

Phylogenetic Analyses and Species Delimitation
All consensus sequences were blasted against the NCBIs GenBank nucleotide database to determine the closest relative species of Colletotrichum to our isolates. In total, 185 isolates of Colletotrichum were included in the molecular phylogenetic analyses. Additionally, sequences from 70 species of Colletotrichum (90 isolates in total) were downloaded from GenBank and included in the analysis as reference sequences or outgroups (Table 1). Reference sequences were selected based on their high similarity with our query sequences using MegaBLAST and were added to the data set and aligned using CLUSTAL W v. 2.0.11 [52].
A Neighbour-Joining (NJ) analysis was performed individually for each genomic area using the Maximum Composite Likelihood method and 2000 bootstrap replications to determine whether the sequence datasets were congruent and combinable. Tree topologies of 70% reciprocal bootstrap generated individually for each locus were compared visually. Because no supported nodes were in conflict, the data of different loci were combined into single concatenated datasets. Three different datasets were analyzed to compare and identify our Colletotrichum isolates correctly. For a first identity approach, one phylogeny was constructed using a combination of ITS and TUB2 sequences (dataset I). This phylogeny consisted of 40 taxa of the Colletotrichum genus including species belonging to C. acutatum, C. boninense, and C. gloeosporioides species complexes, among other Colletotrichum spp., with C. dracaenophilum (CBS 118199) as an outgroup. Subsequently, a second phylogeny was performed by multilocus alignment of ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH sequences (dataset II) to identify our isolates [27]. This second phylogeny consisted of 41 taxa of the Colletotrichum genus including species belonging to C. acutatum, C. boninense, and C. gloeosporioides species complexes, with C. dracaenophilum (CBS 118199) again as an outgroup. A third multilocus alignment combining the ITS, TUB2, and ApMat sequences (dataset III) was performed for inferring organismal phylogeny of 14 isolates belonging to the C. gloeosporioides species complex [53,54].
For multilocus alignments, phylogenetic analyses were conducted by Bayesian Inference (BI) and Maximum Parsimony (MP). The MP trees were obtained using the Tree-Bisection-Regrafting (TBR) algorithm with search level 1, in which the initial trees were obtained by the random addition of sequences (10 replicates). All positions containing gaps and missing data were eliminated. A set of 2000 bootstrap replications evaluated the robustness of the generated trees. Tree length (TL), consistency index (CI), retention index (RI), rescaled consistency index (RC), and homoplasy index (HI) were recorded. BI analyses were performed with MrBayes v.3.2.6 [55], which uses Markov Chain Monte Carlo to approximate the posterior probability of trees. Two analyses with four chains each were run at the same time, for 1 × 10 7 generations, sampled every 100 generations, and starting from a random tree topology. The "temperature" parameter was set to 0.2. For the consensus tree, the first 25% of the saved trees were discarded as the burn-in phase of the analysis. Each of the individual genes and a combined data set were aligned, adjusted manually, and analyzed by NJ or MP using MEGA v.7 [56]. In BI and NJ analyses, the best evolutionary model for each gene partition was also determined by MEGA v.7. The genes were concatenated in a single nucleotide alignment using Phylogenetic Data Editor (PhyDE-1).

Pathogenicity on Olive Fruit
The following Colletotrichum isolates were evaluated according to their pathogenicity on olive fruit: C. acutatum isolates from olive fruit (Col-193 and Col-256) and almond fruit (Col-536); C. fioriniae isolate from olive fruit (Col-172); C. gloeosporioides from olive fruit (Col-41) and sweet orange fruit (Col-69); C. godetiae from olive fruit (Col-30, Col-57, Col-88, Col-508, Col-515, and Col-519) and almond fruit (Col-522); C. karstii from sweet orange fruit (Col-79); and C. nymphaeae isolates from olive fruit (Col-42 and Col-506) and strawberry fruit (Col-84 and Col-86) ( Table 1). Violet (color class 3) olive fruit of the highly susceptible cv. Hojiblanca were collected from olives growing in the World Olive Germplasm Bank (WOGB), belonging to the IFAPA located in the Córdoba province [57]. Before inoculation, the olive fruits were washed and surface-disinfested according to Moral et al. [9]. Surfacedisinfected olive fruits were placed in moist chambers (plastic containers, 22 × 16 × 10 cm) at 100% relative humidity (RH) and inoculated by spraying them up run-off with a conidial suspension adjusted with a haemocytometer to 10 5 conidia mL −1 . After inoculation, humid chambers were incubated at 23 ± 2 • C with a 12-h photoperiod. Additionally, olive fruit sprayed with sterile distilled water were included as a control. There were three replicated humid chambers per isolate and 20 fruits per humid chamber, and the experiment was conducted twice. A completely randomized design was used with fungal isolates as the independent variable and moist chambers as replications. The pathogens were re-isolated from the olive fruit as described above.

Pathogenicity on Other Hosts
Colletotrichum godetiae isolates from olive fruit (Col-9 and Col-57), C. gloeosporioides from sweet orange fruit (Col-69), C. karstii from sweet orange fruit (Col-79), C. nymphaeae from olive fruit (Col-42), and C. siamense from strawberry fruit (Col-44) were selected to evaluate their pathogenicity on different hosts (Table 1). Fruits of almond (Prunus dulcis (Mill.) D.A.Webb) cv. Guara, apple (Malus domestica Borkh.) cv. Golden Delicious, sweet orange (Citrus sinensis L.) cv. Lanelate, and strawberry (Fragaria × Ananassa L.) cv. Camarosa, as well as leaves of oleander (Nerium oleander L.) were selected for this assay. Plant material was washed, and surface disinfested as described above. The pathogenicity of the six Colletotrichum isolates was evaluated by independent inoculation on the different hosts. Thus, almond, apple, olive, and strawberry fruits were inoculated by surface deposition of one mycelial plug (9 mm in diameter) per fruit pierced with a sterile needle, according to Moral et al. [9]. Oleander leaves were inoculated by the same method, but in this case, three mycelial plugs (7 mm in diameter) were deposited per leaf. Inoculated fruits and leaves were incubated in moisture chambers at 23 ± 2 • C with a 12-h photoperiod. Additionally, non-inoculated fruits or leaves treated with PDA plugs were included as a control. There were three replicated humid chambers per isolate-host combination, 10 fruits or leaves per humid chamber, and the experiment was conducted twice. A completely randomized design was used with fungal isolates and host as the independent variable and moist chambers as replications. The pathogens were re-isolated from the fruits and leaves as described above.

Disease Assessment and Data Analysis
Disease severity (DS) in inoculated olive and almond fruits was evaluated weekly until most of the fruit achieved the maximum value (approx. 14 and 21 days for olive and almond fruits, respectively). DS was assessed using a 0-5 rating scale: (0) no symptoms; (1) 1-25% of the fruit surface affected; (2) 26-50%; (3) 51-75%; (4) >75%; and (5) 100% [9]. A disease severity index (DSI) was calculated in each replication using the following formula: DSI = [(Σn i × i)/(N × 5)] × 100, where i represents a severity (zero to five), n i is the number of fruits with severity i, N is the total number of fruits, and five is the highest value of the severity rating scale. For the rest of the hosts, the largest and smallest diameters of lesions were measured weekly, and mean data were converted to the radial growth rate (mm day −1 ). DS of the inoculated fruits of apple, sweet orange, and strawberry, and in leaves of oleander was evaluated weekly until most of the fruits or leaves reached 90-100% of their surface affected (approx. 21, 41, 12, and 18 days for apple, sweet orange, strawberry, and oleander, respectively). In all cases, relative areas under the disease progress curve (RAUDPC) were calculated using the trapezoidal integration of DSI values over time. RAUDPC data from the two runs of the experiment were subjected to analysis of variance (ANOVA). The non-pathogenic isolates were excluded from the statistical analysis. The RAUDPC data were logarithmically transformed when necessary to the homogeneity of variances or normality. When ANOVA showed significant differences for each host, means were compared according to Tukey's honestly significant difference (HSD) test at p = 0.05 [44]. Data were analyzed using Statistix 10 [45].

Collection of Fungal Isolates
In total, 137 Colletotrichum isolates were obtained from different hosts across the Iberian Peninsula: 83 of them isolated from olive trees in Portugal, and 54 of them isolated from olives and other hosts in Spain. Forty-six of the Spanish isolates were obtained from olives located in the four major olive-producing regions (Andalusia, Extremadura, Catalonia, and Valencia; located at Southern, South-western, North-eastern, and Eastern Spain, respectively). The other eight Spanish isolates were recovered from almond (two isolates), Citrus (two isolates), Pistacia terebinthus (one isolate), and strawberry (three isolates). In addition to Iberian isolates, we included 16 isolates from Australia, 1 isolate from Brazil, 5 isolates from California, 6 isolates from Greece, 12 isolates from Italy, 5 isolates from Tunisia, and 3 isolates from Uruguay obtained from affected olive fruit (Table 1).

Colonies and Conidial Morphology
Most Colletotrichum colonies were similar regarding texture and density characteristics with abundant aerial mycelium with regular margins. However, the Australian isolates Col-166, Col-200, Col-152, and Col-214 had colonies with lobulated margins. The growth pattern of all colonies was radial with concentric circles. Nevertheless, the colonies showed a broad variation in color, mainly white, whitish to dark gray, and pinkish-orange being the most common colors observed. Thus, colony color was helpful to discriminate color sub-groups. In general, colonies of all C. godetiae isolates were gray (from dark to light gray), colonies of C. acutatum isolates showed pinkish-orange tones, and C. fioriniae and C. gloeosporioides isolates were light gray. In particular, the isolate C. siamense Col-44 from strawberry showed a distinctive greenish-gray colony color. However, the rest of the Colletotrichum spp. isolates showed colonies with many variations in color within the same species, so it was impossible to establish a relationship between species and the color of their colonies (Table 2; Figure 2).   The average length of the conidia ranged between 8.3 and 14.8 µm for C. acutatum isolate Col-175 and C. gloeosporioides Col-41, respectively. The average width varied from 2.7 to 5.1 µm for C. acutatum isolate Col-175 and for the isolates C. godetiae Col-30 and C. gloeosporioides Col-69, respectively. In general, the conidia were hyaline, varying in type (ellipsoid, clavate, and fusiform) between isolates within species complex or even within the same fungal species. Isolates belonging to C. acutatum species complex had the three types of conidia. Isolates of C. fioriniae and C. nymphaeae showed fusiform and clavate conidia, respectively. Most isolates identified as C. godetiae showed clavate conidia, except isolates Col-88 and Col-522, which showed ellipsoid and fusiform conidia, respectively. Colletotrichum simmondsii isolate Col-169 showed fusiform conidia. Concerning the isolates belonging to the C. boninense complex, differences in the type of conidia were also observed between species. For example, C. boninense isolate Col-178 showed clavate conidia, while C. karstii isolate Col-79 showed ellipsoid conidia. Finally, most isolates belonging to the C. gloeosporioides complex showed ellipsoid conidia, except two isolates identified as C. siamense (isolates Col-44 and Col-184), which showed clavate conidia ( Table 2).

Benomyl Sensitive Assay
Wide variability in mycelial growth rate was observed among the Colletotrichum isolates grown on PDA amended with 5 µg mL −1 of benomyl. In general, Colletotrichum isolates developed lower aerial mycelium and greater conidial production than those in the presence of the fungicide. There were significant differences (p < 0.001) for mycelial growth inhibition between isolates. According to their sensitivity to benomyl, the Colletotrichum isolates could be grouped into two groups (moderately and highly sensitive). The moderately sensitive group only included isolates belonging to C. acutatum species complex, whose percentages of inhibition ranged from 33.5% to 71.1% for C. acutatum isolate Col-166 and C. fioriniae isolate Col-172, respectively, both from olive trees in Australia ( Table 2). The highly sensitive group was formed by isolates belonging to C. boninense and C. gloeosporioides species complexes, whose percentages of mycelial growth inhibition ranged from 93.8% to 100% for C. siamense isolate Col-160 (from olive fruit, Australia) and C. gloeosporioides isolate Col-41 (from olive fruit, Spain), respectively. However, no benomyl-resistant Colletotrichum isolates were observed in any case.

Hydrolysis-Casein Assay
Seventeen out of the 26 tested Colletotrichum isolates caused a casein hydrolysis halo surrounding their colonies in CHM that was observable at 1 day of incubation. At 5 days of incubation, Colletotrichum isolates were classified as able (+ or ++ for the width of halo ≤ 2 or > 2 mm, respectively) or not able (-) to hydrolyze casein. This phenotypic characteristic was also helpful to discriminate isolates between Colletotrichum species complexes, but with some exceptions. Thus, all the isolates belonging to the C. acutatum species complex could hydrolyze casein, except C. fioriniae isolate Col-172 from olive trees in Australia. Most isolates belonging to C. boninense and C. gloeosporioides species complexes could not hydrolyze casein, except three isolates within the C. gloeosporioides species complex (C. alienum isolate Col-211, C. siamense isolates Col-184, and Col-187, all of them from olive trees in Australia) ( Table 2).

Molecular Characterization. Phylogenetic Analyses
Our Colletotrichum isolates were initially identified based on the combined data of ITS and TUB2 sequences alignment. This first analysis (dataset I) included 186 taxa from which 109 were sequences of our isolates, and 77 were reference sequences from GenBank including the outgroup C. dracaenophylum isolate CBS 118199. A total of 867 characters, including gaps, were analyzed (ITS from 1 to 500, and TUB2 from 501 to 867 position). For BI analysis, a K2 + G model was used to combine both regions, and the phylogenetic tree is shown in Figure 3. In the MP analysis of the ITS and TUB2 regions, there were 811 positions in the final dataset, from which 220 characters were parsimony-informative and 591 conserved sites. The five most parsimonious trees were retained (TL = 555 steps, CI = 0.520, RI = 0.956, RC = 0.539, and HI = 0.480). The consensus tree obtained by MP analysis confirmed the topology obtained with BI, and bootstrap supports agreed with Bayesian probability values. Our isolates were grouped into three well-supported clades in this first phylogenetic tree according to the three Colletotrichum species complexes. Likewise, 93 isolates (from different countries and hosts) were grouped into the C. acutatum species complex, two isolates (Col-178 and Col-79, from Australia and Spain, and olive and sweet orange fruits, respectively) were grouped into the C. boninense species complex, and 14 isolates (from Australia, Spain and Tunisia, and most of them from olive trees) were grouped into the C. gloeosporioides species complex. These three different clades were well supported with a Bayesian posterior probability (PP) value of 1.0 for all of them, and with bootstrap support (MP (BS); %) values of 99%, 98%, and 100% for C. acutatum, C. boninense, and C. gloeosporioides species complexes, respectively. Most of the isolates belonging to the C. acutatum species complex clustered in four clades: (i) 45 isolates clustered together with reference isolates of C. godetiae (PP/BS(%):1/99), (ii) 11 isolates clustered with reference isolates of C. acutatum (1/99), (iii) 29 isolates clustered with reference isolates of C. nymphaeae (<0.90/70), and (iv) 7 isolates clustered with reference isolates of C. fioriniae (1/99). The isolate Col-169 from olive fruit (Australia) could not be well-identified with this phylogenetic analysis due to the fact that it clustered between reference sequences of C. paxtonii (IMI 165753) and C. simmondsii (CBS 122122) within the C. acutatum species complex.
Concerning the C. boninense species complex, the isolate Col-178 (from olive fruit, Australia) clustered together with reference isolates of C. boninense (1/99), and the isolate Col-79 (from sweet orange, Spain) clustered together with reference sequences of C. karstii (0.99/93). ITS and TUB2 multilocus alignment was not helpful to distinguish between species belonging to the C. gloeosporioides species complex, which formed a unique clade (1/100) (Figure 3).
A second multigene analysis (dataset II) was performed based on ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH regions with a total of 188 taxa from which 126 were sequences of our isolates, and 62 were reference sequences from GenBank, including the outgroup C. dracaenophylum isolate CBS 118199. A total of 2143 characters, including gaps, were analyzed. The gene boundaries in the multialignment were ITS (from 1 to 518 positions), TUB2 (519-902), ACT (903-1186), CHS-1 (1187-1464), HIS-3 (1465-1853), and GADPH (1854-2143). For Bayesian analysis, a K2 + G model was selected for ITS, a K2 + I model for TUB2 and ACT, a TN93 + G model for CHS-1 and HIS-3, and a K2 + G+I model for GADPH, and they were incorporated in the analysis. The tree obtained with Bayesian PP values is shown in Figure 4.     Regarding MP analysis, there were 1845 positions in the final dataset, from which 693 characters were parsimony-informative, 1266 conserved sites, and 114 parsimonyuninformative. Two most parsimonious trees were retained (TL = 1673 steps, CI = 0.525, RI = 0.944, RC = 0.529, HI = 0.475). The consensus tree obtained by MP analysis confirmed the topology obtained with Bayesian inference, and BS values agreed with Bayesian probability values. This second phylogenetic improved the identification of isolates belonging to the C. acutatum species complex. Regarding our isolates, 119 were grouped as C. acutatum species complex, one isolate (Col-79) was grouped as C. boninense complex, and 14 isolates were grouped as C. gloeosporioides species complex. These three clades were well supported with a PP value of 1.0 and BS values of ≥ 99%. The isolates belonging to the C. acutatum species complex clustered in five well-supported clades: (i) 84 isolates (one from Italy, 73 from Portugal, and 10 from Spain, most of them from olive trees) clustered together with three reference isolates of C. nymphaeae (1/98), (ii) 25 isolates (20 from Spain, and 5 from Portugal, all of them from olive trees except Col-522 from almond trees) clustered with seven reference isolates of C. godetiae (1/99), (iii) 6 isolates from olive trees (five from California and one from Australia) clustered with six reference isolates of C. fioriniae (1/100), and (iv) 3 isolates with different origins (Australia, Tunisia, and Spain) clustered with the reference isolates of C. acutatum (1/100). In this case, the isolate Col-169 from olive trees (Australia), which could not be identified before based on ITS and TUB2-combined alignment, clustered consistently (1/94) with the reference ex-type isolate of C. simmondsii CBS 122122. The combined alignment of ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH regions was insufficient to distinguish between species belonging to the C. gloeosporioides species complex (Figure 4).
Finally, an additional multilocus alignment combining ITS, TUB2, and ApMat gene sequences (dataset III) was performed for inferring organismal phylogeny of the isolates belonging to the C. gloeosporioides species complex. It included 42 taxa, from which 14 were sequences of our isolates, and 29 were reference sequences from GenBank including the outgroup C. xanthorrhoeae ICMP17903. A total of 1749 characters, including gaps, were processed. The gene boundaries in the multialignment were ITS (1-484), TUB2 (485-842), and ApMat (843-1749). For Bayesian analysis, a K2 + G model was selected for ITS and ApMat, while a K2 model was used for TUB2. The tree obtained with Bayesian posterior probability values is shown in Figure 5. Regarding MP analysis, there were a total of 1576 positions in the final dataset, from which 363 characters were parsimony-informative, 1097 conserved sites, and 116 parsimony-uninformative. The four most parsimonious trees were retained (TL = 869 steps, CI = 0.764, RI = 0.894, RC = 0.683, and HI = 0.236). MP analysis confirmed the tree obtained by BI and BS agreed with PP values. The 14 isolates from this study classified within the C. gloeosporioides complex were identified as C. alienum (two isolates from olive trees, Australia; 1/99), C. fructicola (one isolate from olive trees, Spain; 1/99), C. gloeosporioides (four isolates, two of them from Tunisia (Col-251 and Col-295); and the other two from Spanish olive (Col-41) and sweet orange trees (Col-69); 1/100), C. perseae (one isolate from olive trees, Australia; 1/100), C. siamense (five isolates, four of them from olive trees Australia, and one from strawberry, Spain (Col-44; 1/83), and C. theobromicola syn. C. fragariae (one isolate from olive trees, Australia; 1/99) ( Figure 5).
Considering just the new olive tree isolates in this study, we molecularly identified 177 isolates belonging to 12 species of Colletotrichum. Most of these were Portuguese and Spanish isolates, 83 and 45 isolates, respectively. The species C. nymphaeae and C. godetiae were the most frequent (87 and 59 isolates, respectively) followed by C. acutatum, C. fioriniae, and C. gloeosporioides with 10, 7, and 3 isolates, respectively. The rest of the species (C. boninense, C. fructicola, C. perseae, C. simmondsii, and C. theobromicola) were represented by just one. At the same time, two and four Australian isolates were identified as C. alienum and C. siamense, respectively. Overall, C. godetiae was the dominant species in all the European countries (Greece, Italy, and Spain) except in Portugal, where the C. nymphaeae (89%) was the predominant species followed by C. godetiae (11%) (Figure 6). ienum (two isolates from olive trees, Australia; 1/99), C. fructicola (one isolate from olive trees, Spain; 1/99), C. gloeosporioides (four isolates, two of them from Tunisia (Col-251 and Col-295); and the other two from Spanish olive (Col-41) and sweet orange trees (Col-69); 1/100), C. perseae (one isolate from olive trees, Australia; 1/100), C. siamense (five isolates, four of them from olive trees Australia, and one from strawberry, Spain (Col-44; 1/83), and C. theobromicola syn.C. fragariae (one isolate from olive trees, Australia; 1/99) ( Figure 5). Considering just the new olive tree isolates in this study, we molecularly identified 177 isolates belonging to 12 species of Colletotrichum. Most of these were Portuguese and Besides the isolates from olive trees, we have included eight Colletotrichum isolates obtained from other hosts in Spain, such as almonds, sweet orange trees, strawberries, and terebinth (Pistacia terebinthus). Among these isolates, six species were identified: C. acutatum from an almond tree, C. godetiae from almond and terebinth trees, C. nymphaeae (two isolates) and C. siamense from strawberries, and C. gloeosporioides and C. karstii from a sweet orange tree (Table 1).

Pathogenicity to Olive Fruit
Significant differences in virulence (p ≤ 0.001) were observed between isolates depending on the Colletotrichum species and the original host. Most of the isolates tested were pathogenic in olive fruit except for C. nymphaeae isolates Col-84 and Col-86, and both originated from strawberries. Among the pathogenic isolates, RAUDPC values varied from 4.7% to 83.1% for C. karstii isolate Col-79 and C. nymphaeae isolate Col-506. The olive tree isolates Col-508 and Col-515 of C. godetiae, together with the isolates Col-506 of C. nymphaeae, were the most virulent olive fruit (RAUDPC > 64%) (Figures 7 and 8). Overall, olive tree isolates belonging to C. godetiae and C. nymphaeae caused the typical "soapy rot", i.e., rot covering the totality of the fruit surface with abundant conidia in mucilaginous orange masses.
Besides the isolates from olive trees, we have included eight Colletotrichum isolates obtained from other hosts in Spain, such as almonds, sweet orange trees, strawberries, and terebinth (Pistacia terebinthus). Among these isolates, six species were identified: C. acutatum from an almond tree, C. godetiae from almond and terebinth trees, C. nymphaeae (two isolates) and C. siamense from strawberries, and C. gloeosporioides and C. karstii from a sweet orange tree (Table 1).

Pathogenicity Tests
3.4.1. Pathogenicity to Olive Fruit Significant differences in virulence (p ≤ 0.001) were observed between isolates depending on the Colletotrichum species and the original host. Most of the isolates tested were pathogenic in olive fruit except for C. nymphaeae isolates Col-84 and Col-86, and both originated from strawberries. Among the pathogenic isolates, RAUDPC values varied from 4.7% to 83.1% for C. karstii isolate Col-79 and C. nymphaeae isolate Col-506. The olive tree isolates Col-508 and Col-515 of C. godetiae, together with the isolates Col-506 of C. nymphaeae, were the most virulent olive fruit (RAUDPC > 64%) (Figures 7 and 8). Overall, olive tree isolates belonging to C. godetiae and C. nymphaeae caused the typical "soapy rot", i.e., rot covering the totality of the fruit surface with abundant conidia in mucilaginous orange masses.

Discussion
Fungal species belonging to the Colletotrichum genus are characterized by a global distribution associated with anthracnose diseases affecting a wide range of hosts, including many tree crops [5,[58][59][60]. Although numerous species of Colletotrichum have been associated with the olive crop, these studies have focused on specific producing regions and lack an overall view [4,5,23,24,31]. Interestingly, the diversity of Colletotrichum species affecting olive trees in Spain, the leading olive oil producer globally, is very little known since the main study was conducted before dividing the Colletotrichum species complex into species to molecular profiles [33]. The present work focused on elucidating the biodiversity of the Colletotrichum species, causing olive anthracnose worldwide, emphasizing the fungi population from the Iberian Peninsula, i.e., Spain and Portugal. To this end, a vast collection of Colletotrichum isolates obtained mainly from olives in the main olive-growing regions of the world (Australia, Brazil, California, Greece, Italy, Portugal, Spain, Tunisia, and Uruguay) were characterized based on morphological, molecular, and pathogenic characters.
Several Colletotrichum species can produce infections in a single host, showing high pathogenic specialization; much more frequent, however, are the Colletotrichum species with the ability to infect multiple hosts [58,60]. These antecedents suggest that correct taxonomic identification of Colletotrichum species is essential to avoid etiological ambiguities. Therefore, determining the aetiology of Colletotrichum diseases will be crucial to develop studies on the epidemiology and control of the disease in the future [10].
By tradition, the taxonomic identification of the species of Colletotrichum genus has been mainly based on phenotypic differences of colony morphology and conidium shape and size [59,61,62]. Several authors considered the curvature of the ends of the conidia as the essential morphological character to distinguish between Colletotrichum species [16,62,63]. This conidial character has been traditionally used to discriminate between C. acutatum (sharp conidium ends, fusiform) and C. gloeosporioides (rounded conidium ends, ellipsoid) [58,64]. Nevertheless, conidia of the isolates characterized morphologically in this study varied in form (ellipsoid, clavate, or fusiform) between fungal isolates within the same species complexes and, even, the same fungal species. For example, within the C. acutatum species complex, the three shapes of conidia were observed for C. acutatum isolates, whereas C. godetiae isolates showed clavate conidia except for the isolates Col-88 and Col-522, which showed ellipsoid and fusiform conidia, respectively. Because of this morphological characteristic (clavate conidia), Faedda et al. [26] described the new species C. clavatum as the most common associated with olive anthracnose in Italy. However, this new species did not show molecular and morphological differences with the previous one, C. godetiae [27]. Likewise, in this study, and previous ones, different types of conidia were also observed between species belonging to C. boninense species complex (clavate or ellipsoid conidia) as well as within the C. gloeosporioides species complex (ellipsoid or clavate conidia).
Regarding the colony color, there were no differences that allowed their specific identification. Usually, colonies of C. godetiae isolates were gray, C. acutatum isolates showed pink tones, and C. fioriniae and C. gloeosporioides isolates showed light gray ones. The only exception was C. siamense from strawberries (Col-44), which showed a distinctive greenish-gray colony color. Similar results about differences between Colletotrichum species affecting almond trees depending on colony color-subpopulations were described by López-Moral et al. [36], who indicated that C. acutatum, C. godetiae, and C. nymphaeae isolates were associated with pinkish-orange, dark gray, and light gray subpopulations, respectively. Despite these specific differences in colony color between Colletotrichum species, it does not correctly identify Colletotrichum species since environmental factors could significantly influence the stability of morphological traits becoming in intermediate forms [58,65,66].
Regarding the sensitivity to the benomyl, all the isolates included within C. boninense and C. gloeosporioides species complexes were highly sensitive to the fungicide with inhibition percentages of mycelial growth higher than 93%. Conversely, the isolates belonging to the C. acutatum species complex were more tolerant (from 33.5% to 71.1%) to the fungi-cide than those of the C. gloeosporioides species complex. However, there was an unclear association between fungicide tolerance and pathogen species or geographic or host origin. Our results are concordant with those obtained by several authors who indicated that C. gloeosporioides isolates are highly sensitive to benomyl while C. acutatum isolates are moderately tolerant independent of the host origin [4,43,58,67,68]. Species belonging to the C. acutatum complex are predominant in the olive-growing areas in the Andalusia region [7]. Thus, Andalusian C. acutatum isolates show higher tolerance to benomyl than those of C. gloeosporioides and C. boninense. These differences could be a consequence of the use of this fungicide in olive orchards. However, benomyl had not been traditionally used by olive growers in Andalusia to prevent olive diseases and is currently not registered for use [7].
The ability of Colletotrichum isolates to hydrolyze casein was also helpful to discriminate isolates between Colletotrichum species complexes, but with some exceptions. Thus, most isolates belonging to the C. acutatum species complex hydrolyzed the casein, whereas those belonging to C. boninense and C. gloeosporioides species complexes did not. These results are in concordance with those obtained by Martín et al. [69].
Molecular techniques, such as phylogenetic analyses of ribosomal genes (i.e., ITS, 28S, etc.) and functional protein regions (i.e., actin, β-tubulin, calmodulin, etc.) have been set up during these last few years, improving the identification of Colletotrichum species within this genus [27,50,54,66,70]. The combined alignment of ITS and TUB2 helped identify the isolates into the three Colletotrichum species complexes: C. acutatum, C. boninense, and C. gloeosporioides. In general, ITS and TUB2 were enough to infer Colletotrichum species within C. acutatum and C. boninense complexes except for the isolate Col-169, which was identified as C. simmondsii based on the ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH regions.
In corroboration with previous studies [54,[70][71][72], we used an additional alignment combining ITS, TUB2, and ApMat gene sequences for inferring the phylogeny of the isolates previously grouped as C. gloeosporioides species complex. Phylogenetic studies conducted to determine the provided information by ApMat and glutamine synthetase (GS) showed that regions offer similar information, but ApMat discriminates more species in the C. gloeosporioides species complex [70,72].
All the aspects discussed above are in agreement with the ideal polyphasic approach for Colletotrichum systematics described by Cai et al. [73], who suggested that the identification of Colletotrichum species should be based on multi-gene phylogenetic analysis together with recognizable phenotypic characters, such as morphology, physiology, pathogenicity, or cultural characteristics, among others.
Concerning the global distribution of our Colletotrichum isolates, most of those collected from olive trees in Spain were classified within the C. acutatum species complex, with C. godetiae being the most common species, followed by C. nymphaeae. The Spanish isolates of C. godetiae were collected in the Andalusia region, whereas C. nymphaeae isolates showed more diversity regarding the country's geographic origin. We previously observed that olive anthracnose is caused by the C. acutatum species complex in the olive-growing areas of southern Spain [7]. Still, the molecular identification of these isolates has not been conducted until the present study. The species C. gloeosporioides and C. fructicola were also isolated from olive trees in Valencia (Eastern Spain) and Catalonia (North-Eastern Spain). Remarkably, the species C. fructicola (Col-82) was isolated from olive leaves showing necrotic lesions (an unusual symptom for anthracnose) from plants in a nursery, probably due to cross-contamination with citrus plants. The infection/contamination of olive stock with Colletotrichum could influence the long-distance spread of these pathogens.
In our study, most olive isolates from Greece, Italy, and Tunisia were identified as C. godetiae. These results are in concordance with those obtained by several authors who indicate that the species belonging to the C. acutatum complex are considered the most important ones associated with olive anthracnose in European countries [6,24,26,32]. Conversely, and in concordance to the previous studies [4,25], our study confirmed that the most prevalent species associated with olive anthracnose in Portugal is C. nymphaeae. In initial studies, we observed that the Spanish C. godetiae isolates, coming from olive-growing regions where copper-based fungicides are frequently used by farmers, are more tolerant to copper than C. nymphaeae isolates, while in Portugal, the opposite is true. However, the adaptation to weather and agronomic conditions, including the potential specialization in the local olive cultivars, could explain these differences [7,74]. In addition, in previous studies we occasionally detected interactions between olive cultivar-Colletotrichum spp., but none so important so as to explain such a different species distribution [7][8][9]57].
Although we identified eight Colletotrichum species among the Australian isolates, neither C. godetiae nor C. nymphaeae were found. This substantial variability of species associated with olive anthracnose in Australia was influenced by the fact that the 16 studied isolates were previously selected from a higher search to maximize the variability. Besides, previous studies hypothesized that the center of origin of Colletotrichum could be in Oceania since the highest level of variability and strains of the species complexes occurred mainly in Australia and New Zealand [59]. The species C. siamense and C. theobromicola have been previously described in olive trees in Australia [23]. However, the species C. perseae (Col-205) was identified for the first time as associated with olive anthracnose. The species C. alienum has been identified in a broad diversity of hosts, including olive trees [13,53,75], while C. perseae has been described as novel species associated with avocado anthracnose in Israel [21].
All of the isolates from olive trees from California were identified as C. fioriniae. Nevertheless, the etiological studies of olive anthracnose in this state have not been conducted yet. However, C. fioriniae is a common pathogen of nut trees in California [27,76].
Finally, among all our isolates, species belonging to the C. gloeosporioides complex were only identified for the isolates collected from Australia, Tunisia, and Eastern Spain. These results agree with those described by Talhinhas et al. [5], who indicated that the C. gloeosporioides complex occurs in several countries presenting lower frequency than other species. On the other hand, these authors described C. acutatum as the prevalent species complex associated with olive anthracnose in the Southern Hemisphere.
Regarding the six Colletotrichum species obtained from hosts other than the olive tree (i.e., C. acutatum, C. gloeosporioides, C. godetiae, C. karstii, C. nymphaeae and C. siamense), it is interesting to note that all of them are new reports from the respective hosts in Spain, except C. gloeosporioides from sweet orange and C. acutatum from almond trees [36].
Among the Colletotrichum isolates from almonds, olives, sweet oranges, and strawberries tested for pathogenicity on olive fruit, only the isolates from strawberries were not pathogenic. Overall, the isolates from olive trees were more virulent in olive fruit than those from other hosts [35]. These results agree with López-Moral et al. [36], who observed that Colletotrichum isolates from olives (Col-506 and Col-508) were more virulent than ones from almonds (Col-522 and Col-536) on olive fruit. Overall, the pathogenicity in olive fruit has been confirmed in eight species, which differ in their virulence [5,35]. These pathogenicity tests have demonstrated that C. acutatum and C. nymphaeae are the most virulent species, C. godeatiae and C. fioriniae resulted in an intermediate virulence, and C. gloeosporioides is less virulent [4,23,24]. When cross inoculations were conducted using different isolates and hosts, a notable pathogenic specialization was observed in some cases. For example, C. siamense isolate Col-44 from strawberries resulted as non-pathogenic to apple and sweet orange fruit [7]. Although we can find many differences in virulence between isolates and host combinations, our results demonstrated the pathogenic specialization of Colletotrichum isolates on their host. This characteristic has been used to identify specific or intraspecific taxa in this genus [16,61]. However, further research is needed to determine the pathogenic specialization of Colletotrichum isolates on olive trees.
In conclusion, in the present study, the largest so far, we recorded 12 species of the pathogen affecting the olive tree, C. acutatum, C. alienum, C. boninense, C. fioriniae, C. fructicola, C. gloeosporioides, C. godetiae, C. nymphaeae, C. perseae, C. siamense, C. simmondsii, and C. theobromicola. According to our knowledge, this study is the first report of C. boninense, C. fructicola, and C. perseae affecting olive trees. Other studies have described another six Colletotrichum species associated with this crop, C. aenigma, C. cigarro, C. karstii, C. queenslandicum, C. lupini, and C. rhombiborme [4,5,23,77,78]. Although many other woody crops are affected by numerous species of Colletotrichum [21,54,79,80], the olive tree is one of the plant species affected by the most remarkable diversity of taxa of this fungal genus with 18 species. This fact may be associated with the enormous expansion capacity of the olive tree in the last 30 years, which has led it to be the main woody crop in the world [1]. Our results also showed that the dominant species in Spain, Italy, and Greece is C. godetiae, while C. nymphaeae is the dominant species in Portugal. Interestingly, neither of these two species have been described in Australia, where we have found the highest diversity with eight Colletotrichum spp. These results reinforce the hypothesis that native species of Colletotrichum to each place jumped from other hosts to the olive tree when it colonized new growing areas, rather than the pathogen having moved with the crop.

Conclusions
This study aimed to elucidate the biodiversity of Colletotrichum species causing olive anthracnose worldwide. Our results demonstrated that the phenotypic characters (colony and conidium morphology, benomyl-sensitivity, and casein-hydrolyse ability) are not helpful enough to identify Colletotrichum species, although they allow for the separation of some species complexes. For instance, conidia of the Colletotrichum isolates characterized morphologically in this study varied in form (ellipsoid, clavate, or fusiform) among fungal isolates within the same species complexes and even the same fungal species. Thus, molecular tools are essential to infer phylogenetic species within the Colletotrichum genus. In this respect, ITS and TUB2 are enough to distinguish Colletotrichum species within the C. acutatum and C. boninense species complexes. In contrast, ITS, TUB2, ACT, CHS-1, HIS-3, and GADPH regions were necessary to discriminate within the C. gloeosporioides complex. Consequently, our results reinforce the hypothesis based on the ideal polyphasic approach for Colletotrichum systematics, suggesting that the identification of Colletotrichum species should be based on multi-gene phylogenetic analysis together with recognizable phenotypic characters. Pathogenicity tests in olive showed significant differences in virulence to this host between isolates depending on the Colletotrichum species and host origin. When cross-pathogenicity was conducted, most of the Colletotrichum isolates tested were pathogenic in all the hosts evaluated, except for C. siamense to apple and sweet orange fruits, and C. godetiae to oleander leaves. Finally, regarding the diversity of Colletotrichum species causing olive anthracnose worldwide, among the 177 Colletotrichum isolates from olive included in this study, 12 Colletotrichum species belonging to C. acutatum, C. boninense, and C. gloeosporioides complexes were identified. The species C. godetiae was dominant in Spain, Italy, and Greece. The highest diversity was in Australia, where eight Colletotrichum species were identified. Altogether, this study also reinforces the hypothesis that native species of Colletotrichum to each place jumped from other hosts to the olive tree when it colonized new growing areas, rather than the pathogen having moved with the crop.

Data Availability Statement:
The data that support the findings of this study are available from the corresponding author upon reasonable request.