Cryphonectriaceae associated with rust-infected Syzygium jambos in Hawaii

Abstract Syzygium jambos (Myrtales, Myrtaceae) trees in Hawaii are severely affected by a rust disease caused by Austropuccinia psidii (Pucciniales, Sphaerophragmiaceae), but they are commonly co-infected with species of Cryphonectriaceae (Diaporthales). In this study, S. jambos and other trees in the Myrtales were examined on three Hawaiian Islands for the presence of Cryphonectriaceae. Bark samples with fruiting bodies were collected from infected trees and fungi were isolated directly from these structures. Pure cultures were produced and the fungi were identified using DNA sequence data for the internal transcribed spacer (ITS) region, part of the β-tubulin (BT1) gene and the transcription elongation factor-1α (TEF1) gene. Five species in three genera of Cryphonectriaceae were identified from Myrtaceae tree samples. These included Chrysoporthe deuterocubensis, Microthia havanensis and three previously-unknown taxa described here as Celoporthe hauoliensis sp. nov., Cel. hawaiiensis sp. nov. and Cel. paradisiaca sp. nov. Representative isolates of Cel. hauoliensis, Cel. hawaiiensis, Cel. paradisiaca, Chr. deuterocubensis and Mic. havanensis were used in artificial inoculation studies to consider their pathogenicity on S. jambos. Celoporthe hawaiiensis, Cel. paradisiaca and Chr. deuterocubensis produced lesions on young S. jambos trees in inoculation trials, suggesting that, together with A. psidii, they may contribute to the death of trees. Microsatellite markers were subsequently used to consider the diversity of Chr. deuterocubensis on the Islands and thus to gain insights into its possible origin in Hawaii. Isolates of this important Myrtaceae and particularly Eucalyptus pathogen were found to be clonal. This provides evidence that Chr. deuterocubensis was introduced to the Hawaiian Islands as a single introduction, from a currently unknown source.


Introduction
Fungi in the Cryphonectriaceae (Diaporthales) include at least twenty-three genera of bark-, wood-and leaf-infecting fungi (Gryzenhout et al. 2009Begoude et al. 2010;Vermeulen et al. 2011Vermeulen et al. , 2013Crous et al. 2012;Chen et al. 2013aChen et al. , b, 2016Chen et al. , 2018Crane and Burgess 2013;Beier et al. 2015;Ali et al. 2018;Ferreira et al. 2019;Jiang et al. 2020;Wang et al. 2020). They occur on trees and shrubs in various parts of the world and include saprophytes, facultative parasites and important pathogens of woody plants (Gryzenhout et al. 2009). Pathogens in the family reside mainly in the genera Cryphonectria and Chrysoporthe and include important agents of tree disease, both in natural forest ecosystems, as well as in intensively-managed plantations (Wingfield 2003;Gryzenhout et al. 2009;Wang et al. 2020). These fungi generally have yellow to orange or brown stromata and these structures turn purple in 3% potassium hydroxide (KOH) or yellow in lactic acid (Gryzenhout et al. 2006c(Gryzenhout et al. , 2009Jiang et al. 2020).
The Cryphonectriaceae infect trees and shrubs residing in more than 100 species in at least 26 families and 16 orders of plants worldwide (Gryzenhout et al. 2009;Jiang et al. 2020;Wang et al. 2020). The chestnut blight pathogen, Cryphonectria parasitica (Murrill) M.E. Barr is the best-known tree-killing pathogen in the family. It is native to Asia and outbreaks of the disease in North America and Europe have caused the virtual extinction of endemic populations of chestnut trees on these two continents (Anagnostakis 1987;Heiniger and Rigling 1994;Gryzenhout et al. 2009). Other important pathogens in the Cryphonectriaceae include: Chrysoporthe cubensis (Bruner) Gryzenh. & M.J. Wingf., which is native to South and Central America and causes a canker disease of Eucalyptus species in West Africa and South America (Alfenas et al. 1983;Gryzenhout et al. 2004Gryzenhout et al. , 2009Roux and Apetorgbor 2010); Chrysoporthe deuterocubensis Gryzenh. & M.J. Wingf., native to Southeast Asia and causal agent of a canker disease of Eucalyptus species in Africa, Australia, China and Hawaii (Davison and Coates 1991;Roux et al. 2005;Nakabonge et al. 2006;Zhou et al. 2008;Chen et al. 2010; Van der Merwe et al. 2010;Wang et al. 2020); and Chrysoporthe austroafricana Gryzenh. & M.J. Wingf., endemic to Africa and causal agent of a canker disease of Eucalyptus, Syzygium and Tibouchina species in southern and eastern Africa (Wingfield et al. 1989;Myburg et al. 2002a;Gryzenhout et al. 2004;Roux et al. 2005;Nakabonge et al. 2006).
Hawaii, in the central Pacific Ocean, is comprised entirely of islands and is the northernmost island group in Polynesia (Little and Skolmen 1989). The vegetation is multivariate including many forest types that cover more than 41% of the State's total land area (Anonymous 2003). Hawaii's forests broadly comprise native forest and plantations of non-native trees, interspersed with stands of non-native, invasive tree species. Native forests are dominated by Metrosideros polymorpha Gaudich. (Myrtaceae, Myrtales) and Acacia koa A. Gray (Fabaceae, Fabales) trees, whereas plantations of nonnative trees include various conifers and many tree species (mostly Eucalyptus) that reside in the Myrtaceae (Anonymous 2003). Eight species of indigenous Myrtaceae and more than 200 non-native Myrtaceae have been recorded from the Islands (Loope 2010).
In April 2005, a rust disease caused by Austropuccinia psidii G. Winter (syn. Puccinia psidii, Sphaerophragmiaceae, Pucciniales), was detected on the Island of O'ahu (Uchida et al. 2006;Loope 2010). The pathogen spread rapidly and, consistent with its broad host range in the Myrtaceae (Coutinho et al. 1998;Glen et al. 2007;Carnegie et al. 2016), has been reported to cause disease on at least five native and fifteen non-native Hawaiian species. Of these, the non-native and invasive Syzygium jambos (rose apple) has been especially severely affected by the disease (Loope 2010). Instances of crown death of S. jambos are common and, in some cases, large areas of trees have died (Loope 2010).
During a casual inspection of rust-infected S. jambos in Hawaii by M.J. Wingfield during August 2011 (unpublished data), fruiting bodies of fungi resembling species in the Cryphonectriaceae were observed on the stems and branches of dying trees. This raised interest as very little was known regarding the diversity and distribution of the Cryphonectriaceae infecting Myrtaceae on the Hawaiian Islands. Two species are known to occur and these include, Chr. deuterocubensis, collected from cankers on Eucalyptus species on the Islands of Kauai and Hawaii (Gryzenhout et al. 2006a(Gryzenhout et al. , 2009; Van der Merwe et al. 2010) and Microthia havanensis (Bruner) Gryzenh. & M.J. Wingf., first found on Eucalyptus species grown on the same Islands (Gryzenhout et al. 2006a).
The dramatic death of S. jambos in Hawaii could be caused solely by A. psidii, but the extent of the rapid die-back of branches and stems raised the question as to whether other pathogens, such as the Cryphonectriaceae, might be involved. The aim of this study was, thus, to identify species of Cryphonectriaceae on rust-infected S. jambos, as well as on some other species of Myrtaceae. Furthermore, pathogenicity tests were used to consider the possibility that species in the Cryphonectriaceae might contribute to the death of trees that had become infected and were consequently stressed by A. psidii. The genetic diversity of a collection of the most commonly isolated Cryphonectriaceae species was also characterised to gain insight into its possible origin in Hawaii.

Collection of samples and fungal isolation
Surveys for Cryphonectriaceae were conducted in Hawaii during July 2012. Samples were collected mainly from non-native S. jambos trees infected by A. psidii, but also from various native and non-native Myrtaceae, on the Islands of Maui, O'ahu and Hawaii. Samples were collected using an unstructured approach where the areas sampled were determined by the time available for collections to be made on the three selected Islands. On each of the Islands, two to three sites, where rust-infected trees had previously been found, were selected and surveyed during the course of a single day. As much as possible of each Island was also covered by driving along main roads and sampling at regular intervals where S. jambos plants were observed.
The presence on samples of fruiting structures (ascostromata, conidiomata), typical of the Cryphonectriaceae, was ascertained using a 10× magnification hand lens. Pieces of bark bearing these fruiting structures were excised from infected stems and branches and placed in separate brown paper bags for each tree sampled. Samples from each Island were labelled and placed in plastic bags to prevent desiccation and to promote sporulation of the fungi. Isolations and purification of the Cryphonectriaceae from the wood samples followed the technique described by Chen et al. (2011). All isolates used in this study were deposited in the culture collection (CMW) of the Forestry and Agricultural Biotechnology Institute (FABI, www.fabinet.up.ac.za), University of Pretoria, Pretoria, South Africa. Representative isolates, including ex-type cultures, were deposited in the culture collection (CBS) of the Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands. Dried specimens of cultures were deposited in the National Collection of Fungi (PREM), Roodeplaat, Pretoria, South Africa.

DNA extraction, PCR amplification and sequencing
DNA was extracted from all isolates using PrepMan Ultra Sample Preparation Reagent kits (Applied Biosystems, California, USA) following the manufacturer's instructions. An Eppendorf Mastercycler (Merck, Germany) was used for PCR amplification of the nuclear rDNA region encompassing the internal transcribed spacer regions (ITS1, ITS2) and 5.8S gene of the ribosomal RNA (ITS) operon, part of the β-tubulin gene (BT1) and the transcription elongation factor-1α gene (TEF1). The ITS was amplified using primers ITS1 and ITS4 (White et al. 1990), the BT1 using primers βt1a and βt1b (Glass and Donaldson 1995) and TEF1 using primers EF728F and EF986R (Carbone and Kohn 1999). The PCR reaction mixtures and thermal cycling conditions were the same as described previously for the ITS, BT1 (Chen et al. 2011(Chen et al. , 2013a and TEF1 gene regions (Vermeulen et al. 2013).
A 5 µl aliquot of the PCR products was pre-stained with GelRed TM Nucleic Acid Gel stain (Biotium, Hayward, USA), separated on 1% agarose gels and visualised under UV light. PCR products were purified using Sephadex G-50 Gel (Sigma-Aldrich), following the manufacturer's instructions. The concentrations of the purified PCR products were determined using a Nanodrop ND-1000 Spectrophotometer (Nanodrop Technologies, Rockland, USA). Sequencing reactions were performed using the Big Dye cycle sequencing kit with Amplitaq DNA polymerase FS (Perkin-Elmer, Warrington, UK), following the manufacturer's protocols, on an ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). Protocols for sequencing PCR amplicons were the same as those described by Chen et al. (2011) and both DNA strands were sequenced for each gene region. Sequences of both DNA strands for each isolate were examined visually and combined using the programme Sequence Navigator v. 1.01 (ABI PRISM, Perkin Elmer). The ITS and BT1 gene regions were sequenced for all isolates used in this study. The TEF1 gene region was sequenced for selected isolates in genera for which this region was required for species-level identification.

Phylogenetic analyses
A preliminary identification of the isolates was obtained by performing a similarity search (standard nucleotide BLAST) against the GenBank database (http://www.ncbi. nlm.nih.gov) using the ITS and BT1 sequences. The BLAST results showed that the isolates obtained in the current study grouped in the genera Celoporthe, Chrysoporthe and Microthia.
For analyses of the ITS and BT1 sequences of isolates from Hawaii, the datasets of Wang et al. (2020) were used as templates. Sequences of the ITS and BT1 gene regions were analysed separately and in combination. A partition homogeneity test (PHT), as implemented in PAUP (Phylogenetic Analysis Using Parsimony) v. 4.0b10 (Swofford 2003), was used to determine whether there was conflict between the datasets, prior to performing combined analyses in PAUP (Farris et al. 1995;Huelsenbeck et al. 1996). Two isolates of Diaporthe ambigua (CMW5288 and CMW5587), residing in the Diaporthaceae (Diaporthales), were used as outgroups.
For isolates that grouped in Celoporthe, based on ITS and BT1 gene sequences, TEF1 sequences were required to obtain accurate species-level identifications (Chen et al. 2011;Vermeulen et al. 2013). The ITS, BT1 and TEF1 gene regions were analysed separately and in combination. This made it possible to determine the phylogenetic relationships amongst the isolates from Hawaii and all 10 previously described Celoporthe species (Nakabonge et al. 2006;Chen et al. 2011;Vermeulen et al. 2013;Ali et al. 2018;Wang et al. 2018). A PHT was used to determine if conflict existed amongst the ITS, BT1 and TEF1 datasets (Farris et al. 1995;Huelsenbeck et al. 1996). Two isolates of Holocryphia capensis (CMW37329 and CMW37887) were used as outgroups.
The sequences for each of the single gene datasets, as well as for a combined dataset consisting of two or three gene regions, were aligned using MAFFT online v. 7 (http://mafft.cbrc.jp/alignment/server/) (Katoh and Standley 2013) and applying the iterative refinement method (FFT-NS-i setting). The alignments were edited manually with MEGA4 (Tamura et al. 2007). For each dataset, Maximum Parsimony (MP) and Maximum Likelihood (ML) analyses were performed.
PAUP v. 4.0 b10 (Swofford 2003) was used for MP analyses, with gaps treated as the fifth character. Uninformative characters were excluded and informative characters were unordered and of equal weight with 1000 random addition replicates. The most parsimonious trees were obtained using the heuristic search function with stepwise addition, tree bisection and reconstruction branch swapping. Maxtrees were set to 5000 and zero-length branches were collapsed. A bootstrap analysis (50% majority rule, 1000 replicates) was undertaken to determine statistical support for the internal nodes in the trees. Tree length (TL), consistency index (CI), retention index (RI) and homoplasy index (HI) were used to assess the trees (Hillis and Huelsenbeck 1992).
PhyML v. 3.1 was used for the ML analyses for each dataset (Guindon and Gascuel 2003). The best nucleotide substitution model for each dataset was determined using the software package jModeltest v. 1.2.5 (Posada 2008). In PhyML, the maximum number of retained trees was set to 1000 and nodal support was determined by non-parametric bootstrapping with 1000 replicates. The phylogenetic trees for both MP and ML analyses were viewed in MEGA4 and edited in Microsoft Office PowerPoint version 2013.

Morphology
Isolates of the Cryphonectriaceae were grown at 25 °C on 2% malt extract agar (MEA: 20 g/l malt extract and 15 g/l agar, Biolab, Midrand, South Africa and 1000 ml sterile deionised water) containing 0.05 g/l of the antibiotic streptomycin sulphate (Sigma-Aldrich, Steinheim, Germany). Where no sporulation was obtained on agar media, six isolates, representing the putative new species, were inoculated on water agar medium on to which ~ 5 cm long sterilised Eucalyptus stem sections had been placed. These were kept at room temperature (~ 25 °C) in the dark until fruiting structures were observed. For each new taxon, micro-morphological structures were studied using Nikon microscopes (Eclipse Ni, SMZ18, Tokyo, Japan) and a mounted Nikon DS-Ri2 camera. The structures were initially mounted in water, later being replaced with 85% lactic acid on glass microscope slides. In order to study the morphology of fruiting structures and stromatic tissues, pieces of bark, bearing fungal fruiting structures, were mounted on discs in Leica Tissue Freezing Medium and dissected to 12-16 µm thickness using a Leica CM1520 cryostat (Wetzler, Germany). The cut sections were mounted in 85% lactic acid for observation. Whenever available, up to 50 measurements of characteristic features were made for isolates chosen to represent the types of putative new species. Measurements were recorded as minimum-maximum, except for spore dimensions for which supplementary information (mean ± standard deviation) was added.
Growth in culture was examined for two isolates of each putative new species identified. The protocols used to assess growth in culture were the same as those described by Chen et al. (2011). The growth rate at optimum temperature was repeated twice for ex-holotype isolates and the average was presented.

Pathogenicity tests
Syzygium jambos seeds were collected from a garden in Pretoria, South Africa and germinated to produce seedlings for artificial inoculation studies under quarantine greenhouse conditions. These seedlings were grown for one year, until their stem diameters had reached at least 0.5 cm. Ten seedlings (~ 0.5-1 cm diam. × 30 cm high), were inoculated with each test strain and ten seedlings of the same size were inoculated with a sterile agar disc to serve as controls. Inoculations were made using the same technique as that described by Chen et al. (2011). Four weeks (28 days) after inoculation, the lengths of the lesions in the cambium on each plant were measured. The JMP version 5.0 of SAS software (SAS Institute Inc. 2002) was used for statistical analysis of the lesion length data. One way ANOVA was used to test statistical differences between the means of the lesion lengths. Re-isolations were made from the lesions to confirm that they had resulted from the effects of the inoculated fungi.

Genetic diversity of Chr. deuterocubensis isolates
The genetic diversity of the most commonly encountered and globally important species in the Cryphonectriaceae from Myrtales on the Hawaiian Islands was analysed using microsatellite markers. DNA was extracted from all isolates of freshly-prepared cultures using PrepMan Ultra Sample Preparation Reagent (Applied Biosystems, California, USA), following the manufacturer's instructions. A set of ten microsatellite markers (Suppl. material 1: Table S1), that had been developed and used in previous studies (Van der Merwe et al. 2003, 2010, was tested on ten randomly-selected isolates. The PCR reaction mixes and thermal cycling were the same as those described by Van der Merwe et al. (2003, 2010. PCR aliquots of 5 µl were pre-stained with GelRed Nucleic Acid Gel stain (Biotium, Hayward, USA) and amplicons were separated on 1% (w/v) agarose gels and visualised under UV light to confirm successful amplification. Primer pairs that did not amplify the target loci successfully after several repetitions were discarded. Those that were successful were used to amplify target loci from all the isolates of the available population.
PCR products for each isolate were multiplexed for GeneScan analysis. The composition of each sample mix was the same as that described by Kamgan Nkuekam et al. (2009). Sample mixes were separated on a 36-cm capillary column with POPTM4 polymer on an ABI Prism 3500 sequencer (Perkin-Elmer, Warrington, UK). Allele sizes were determined by comparing the mobility of the PCR products with that of a LIZ 500 size standard (Applied Biosystems, Foster City, California). Microsatellite size data were analysed using the software GeneMapper version 3.0 (Applied Biosystems, Foster City, California).
The allele size for each of the seven loci was scored for each isolate from the collection. These data were used to generate a multilocus haplotype profile for each isolate. Isolates that had identical alleles for each of the seven loci were treated as clones. The frequency of each allele within the collection was calculated by taking the number of times the allele was present in the population and dividing it by the population sample size. This was then used to calculate gene diversity using the formula (Nei 1973), where x k is the frequency of the k th allele.

Collection of samples and fungal isolation
A total of 139 Cryphonectriaceae isolates were obtained from 106 trees sampled on three Hawaiian Islands (Table 1). Trees, from which the fungi were obtained, included a single specimen of the native species, Ohia (Metrosideros polymorpha) and multiple specimens of four non-native Myrtaceae hosts, including an unknown Metrosideros sp., Psidium cattleianum, S. cumini and S. jambos. The majority of trees sampled were those of S. jambos (66 trees), since this was the main tree of focus in the study and it also displayed the most evident examples of rust infection and death at the time of the survey. Samples were obtained from dead sapling trees (~ 0.5 cm or more diameter) or from older dying/dead trees and cankers on living trees. In the case of M. polymorpha, a species of Cryphonectriaceae was obtained from the surface of a single cut stump. On older trees, signs and symptoms of the Cryphonectriaceae could be found on dead branches and stem cankers, including on trees with no obvious infection by the myrtle rust pathogen.

Phylogenetic analyses
For the isolates selected for sequencing, the PCR fragments were approximately 550, 450 and 260 bp for the ITS, BT1 and TEF1 regions, respectively. All sequences obtained in this study were deposited in GenBank ( Table 2). The alignments of each of the datasets were deposited in TreeBASE (http://treebase.org, study ID: S19035). The number of taxa and characters in each of the datasets and a summary of the most important parameters applied in the Maximum Parsimony (MP) and Maximum Likelihood (ML) analyses are presented in Suppl. material 2: Table S2.
For the ITS and BT1 datasets, the PHT generated a value of P = 0.001, indicating that the accuracy of the combined data had not suffered relative to the individual partitions (Cunningham 1997). Sequences of the two regions were combined for analyses. For each of the ITS, BT1 and ITS+BT1 datasets, the ML and MP analyses generated   trees with generally consistent topologies and phylogenetic relationships amongst taxa. Based on the phylogenetic analyses of the ITS, BT1 and combined datasets, the isolates obtained in this study were grouped in three Clusters, referred to as Clusters A-C ( Fig. 1; ITS and BT1 trees not presented). Isolates in Cluster A grouped in the genus Chrysoporthe and they all resided in the same phylogenetic clade as Chrysoporthe deuterocubensis. Isolates in Cluster B grouped in the genus Microthia and were phylogenetically closely related to Microthia havanensis. Isolates in Cluster C grouped with species of Celoporthe. They formed three distinct Clades (Clades a-c) within Celoporthe based on the ITS+BT1 tree (Fig. 1).
In the ITS, BT1 and TEF1 datasets for Celoporthe isolates, the PHT generated a value of P = 0.001, showing that the accuracy of the combined data were unaffected relative to the individual partitions (Cunningham 1997) and the three gene regions were thus combined in the analyses. Other than the ITS tree ( Fig. 2A), Hawaiian isolates formed distinct lineages (Clades a-c) that differentiated them from other Celoporthe species (Fig. 2B-D). In the combined analyses of ITS, BT1 and TEF1 gene sequences, isolates in each of Clades a, b and c formed independent lineages, supported by high bootstrap values (Clade a: ML/MP = 98%/98%; Clade b: ML/MP = 88%/79%; Clade c: ML/MP = 99%/100%) (Fig. 2D). These three clades were consequently recognised as representing three undescribed species. Isolates in Clades a and b were most closely related to Celoporthe guangdongensis and those in Clade c were all most closely related to Cel. eucalypti and Cel. cerciana (Fig. 2D).

Morphology
Fruiting bodies developed for all six isolates grown on Eucalyptus stem sections on water agar after two months of incubation at room temperature. Other than some minor differences, all fungal isolates, obtained in this study, were morphologically similar. This was consistent with the fact that fungi in the Cryphonectriaceae are mostly indistinguishable on artificial media (Gryzenhout et al. 2009).
Colonies on 2% MEA were fluffy and white when young, turning yellow or greenish-grey to greenish when old. The optimal growth temperatures for novel species was 30 °C, at which colonies reached 59-80 mm within 4 days.

Taxonomy
Based on phylogenetic analyses of sequence data for the three gene regions, three previously unknown Cryphonectriaceae species are recognised from non-native Myrtaceae on the Hawaiian Islands. The three fungi reside in the genus Celoporthe and are distinct from described Celoporthe species, based on sequence data. Since limited numbers of fruiting bodies were available from the originally-collected plant material for these three species and mostly conidia were obtained under laboratory conditions, they are defined primarily based on multiple gene DNA sequence data. Morphological descriptions are provided for colonies on MEA and fruiting structures produced on Eucalyptus stem sections.
Culture characteristics. Colonies on 2% MEA, when young showing circular growth with smooth margins, above white with tint of yellow (30 °C) or orange (25 °C) towards the edge of Petri dish, reverse yellow, except for at 30 °C becoming brown towards the edge; with age above becoming brown, except for 30 °C at which each colony showing variable yellow with white mycelial clumps, reverse dark brown at all temperatures; optimal growth at 30 °C (9.4 mm/d), followed by 25 °C (7.9 mm/d) and 20 °C (4.8 mm/d), minimal growth at 35 °C (0.2 mm/d), no growth at 5 °C; mycelia fluffy, density sparce in centre becoming thicker towards the edge.

Celoporthe hawaiiensis
Culture characteristics. Colonies on 2% MEA, when young showing circular growth with smooth margins, above white with yellow tint towards edge (25 °C), reverse pale brown, becoming darker in centre at 25 °C and 30 °C; with age above becoming darker yellow to brown, reverse dark brown, except at 20 °C, 25 °C having yellow with dark brown patches; optimal growth at 30 °C (6.6 mm/d), followed by 25 °C (6.0 mm/d) and 20 °C (4.1 mm/d), minimal growth at 35 °C (0.1 mm/d), growth at 5 °C restricted to mycelial plug; mycelia fluffy, density sparse in centre becoming thicker towards the edge.
Habitat. On/in bark of Psidium cattleianum and Syzygium jambos Distribution. Hawaii, USA Notes. Celoporthe paradisiaca is morphologically similar to its phylogenetically closest relatives, Cel. hawaiiensis and Cel. guangdongensis, but can be differentiated from them by DNA sequences. In the ITS, BT1 and TEF1 datasets, Cel. paradisiaca differs from Cel. hawaiiensis by 6, 3 and 3 base pairs and from Cel. guangdongensis by 7, 4 and 2 base pairs, respectively (Tables 6, 7).

Pathogenicity tests
Inoculation with two isolates each of Chr. deuterocubensis (CMW38375, CMW38549), Mic. havanensis (CMW38563, CMW38585), Cel. hawaiiensis (CMW38553, CMW38610), Cel. hauoliensis (CMW38373, CMW38389) and Cel. paradisiaca (CMW38360, CMW38384) resulted in lesions on the cambium of one-year-old S. jambos trees. There were no significant differences between the means for Cel. hauoliensis and Mic. havanensis when compared to the negative control (Fig. 6). There were significant differences in the means for Chr. deuterocubensis and Cel. hawaiiensis when compared with one another, as well as with the negative control. A strain (CMW38610) of Cel. hawaiiensis was the most pathogenic (Mean = 23.4 mm) of all the fungi tested and it resulted in a mean lesion length that was statistically different when compared to the means for other test strains and the negative control (Fig. 6). The inoculated fungi were re-isolated from the treated plants and not from the controls, thus fulfilling the requirements of Koch's Postulates.

Genetic Diversity of Chr. deuterocubensis isolates
Chrysoporthe deuterocubensis was the most commonly isolated fungus from Myrtales in this study (Table 1). Due to its known importance as a plantation tree pathogen, isolates obtained were subjected to a genetic diversity test using previously-developed microsatellite markers for this fungus. Seven of the 10 microsatellite primers amplified the desired target loci in 93 isolates obtained from four tree species on three Islands of Hawaii  (Table 1). Allele sizes at each locus were estimated and these were within the size ranges for each marker (Van der Merwe et al. 2003). A total of seven alleles (one allele at each locus) and one haplotype were identified in the collection. The gene diversity was zero and the Chr. deuterocubensis collection from Hawaii was determined as 100% clonal.

Discussion
Five species of Cryphonectriaceae, residing in the genera Celoporthe, Chrysoporthe and Microthia, were identified from native and non-native Myrtaceae from three of the Hawaiian Islands (USA). Of these, only Chr. deuterocubensis and Mic. havanensis have previously been found in Hawaii (Gryzenhout et al. 2006a(Gryzenhout et al. , 2009Van der Merwe et al. 2010). In addition, three new species of Celoporthe were discovered and described.
Chrysoporthe deuterocubensis is known to occur in Hawaii where it has been previously recorded as a pathogen of Eucalyptus trees from the Islands of Kauai and Hawaii (Hodges et al. 1979;Gryzenhout et al. 2009;M.J. Wingfield, unpubl.). This fungus, originally known as Chr. cubensis and later recognised as distinct from that species (Van der Merwe et al. 2010), is well-known from many south-eastern Asian countries where it is believed to have originated (Zhou et al. 2008;Gryzenhout et al. 2009;Chen et al. 2010; Van der Merwe et al. 2010;Wang et al. 2020). It exclusively infects trees in the Myrtaceae and is an important pathogen of Eucalyptus outside the native range of this tree (Gryzenhout et al. 2009;Van der Merwe et al. 2010).
The occurrence of Chr. deuterocubensis on native Ohia (M. polymorpha) in Hawaii could be of concern given its importance as a tree pathogen. This prompted us to investigate the population diversity of the fungus in Hawaii and, thus, to gain insights into its possible origin and movement in the region. The seven microsatellite markers, used to study the population diversity of Chr. Deuterocubensis, amplified target loci in ninety-three isolates of the fungus. The trees from which isolates were obtained represented three genera and four different species. The single isolate of the fungus from native M. polymorpha was also included. All isolates, irrespective of the host or island on which they were collected, represented a single genotype of Chr. deuterocubensis and further comparisons were not justified. Overall, the results of this study provide convincing evidence that Chr. deuterocubensis has been introduced into Hawaii.
The occurrence of a single clone of Chr. deuterocubensis in Hawaii is consistent with that of an introduced pathogen that would be expected to have low gene diversity. This is in contrast to native pathogens that are typically genetically diverse in their areas of origin (Gordon et al. 1996;McDonald 1997). The area of origin of Chr. deuterocubensis in Hawaii is unknown, but it is most likely some part of Asia where the pathogen is found on native, as well as non-native, Myrtales (Van der Merwe et al. 2010). The discovery of only a single genotype of Chr. deuterocubensis in Hawaii was surprising and unexpected. This is especially because the isolates were collected from a wide range of different trees spanning three genera and four species and occurring on three different Islands.
Chrysoporthe deuterocubensis has been known on Eucalyptus in Kauai (as Cryphonectria cubensis) for many years (Hodges et al. 1979;Gryzenhout et al. 2004) and this could be the area where it was first introduced. The pathogen also occurs on highly sought-after ornamental trees/shrubs, such as Tibouchina species (Myrtales: Melastomataceae) (Myburg et al. 2003;Gryzenhout et al. 2009) and it is believed to have been moved on cuttings of this tree (Myburg et al. 2003;Gryzenhout et al. 2009). Tibouchina is commonly grown in Hawaii and these trees could also represent a source of a first introduction. This would be in contrast to other Myrtales, such as Eucalyptus spp., that are more commonly moved as seed.
Chrysoporthe deuterocubensis is an aggressive and important pathogen of trees in the Myrtales. It is clearly widespread in Hawaii and it has most likely been present in the state for many years. It appears that the population of the pathogen has increased substantially where it infects S. jambos, apparently being pre-disposed to the development of the canker pathogen by rust caused by A. psidii. Once large populations of a pathogen, such as Chr. Deuterocubensis, develop in an area, the chance of their moving to new environments is heightened by what has been termed a "bridgehead effect" and for which there are numerous examples in Eucalyptus forestry (Wingfield et al. , 2015. Microthia havanens, found in this study on P. cattleianum, S. jambos and S. cumini, was first described as a saprobe on Eucalyptus trees and other trees such as Mango  (Bruner 1916). Other hosts and areas of occurrence for this fungus include Eucalyptus in Mexico and Hawaii, Myrica faya Ait (Myricaceae, Fagales) trees in Madeira and the Azores (Gryzenhout et al. 2006a) and Eucalyptus grandis Hill: Maiden trees in Florida (USA) (Barnard et al. 1987). Microthia havanensis is considered a weakly pathogenic bark-infecting fungus. This was also confirmed in our pathogenicity studies on S. jambos, where the two isolates tested produced lesions that did not differ significantly from the controls.
Preliminary pathogenicity trials on S. jambos showed that some of the isolates of Chrysoporthe and Celoporthe, tested under greenhouse conditions, can result in significant lesions on inoculated plants within a short period of time. Both isolates of Cel. paradisiaca caused distinct lesions, while one isolate each of Cel. hawaiiensis and Chr. deuterocubensis resulted in lesions that were significantly larger than those of the controls. One of the Cel. hawaiiensis isolates was the most aggressive fungus tested and surprisingly more so than the well-recognised pathogen Chr. deuterocubensis. This fungus clearly deserves further study.
Austropuccinia psidii infects mostly young, actively growing leaves and shoots, as well as fruits and sepals (Coutinho et al. 1998;Alfenas et al. 2004;Glen et al. 2017). Infections of leaves and meristems are severe on susceptible seedlings, cuttings, young trees and coppice, causing plants to be stunted and multi branched, inhibiting normal growth and development and sometimes causing death to young seedlings (Booth et al. 2000;Rayachhetry et al. 2001). This is in contrast to species of the Cryphonectriaceae that infect the bark of trees and shrubs (Gryzenthout et al. 2009). Chrysoporthe species, for example, infect the bark and cambium of trees, giving rise to rapidly-expanding cankers on the stems (Gryzenhout et al. 2009). These cankers often girdle the stems, killing the cambium and leading to rapid tree death (Hodges et al. 1976;Wingfield et al. 1989;Gryzenhout et al. 2009).
In the surveys conducted in this study, samples with symptoms of the Cryphonectriaceae were obtained from various parts of trees, including dead branches, stem cankers and also on trees with no obvious infection by the myrtle rust pathogen, A. psidii. We believe that the rapid die-back of S. jambos trees and other nonnative myrtles in Hawaii is, at least in part, due to infection by one or more Cryphonectriaceae species that apparently proliferate in tissue stressed by the Myrtle rust fungus.