Ambrosia beetles in the Euwallacea nr. fornicatus complex (Coleoptera: Curculionidae) vector Fusarium spp. fungi pathogenic to susceptible hosts, including avocado, Persea americana Mill., (Lauraceae). Previous survey traps in Florida avocado groves indicated significant beetle populations in several groves with minimal observed beetle activity, suggesting an external beetle source. A natural area near one such grove revealed E. nr. fornicatus colonization of wild tamarind, Lysiloma latisiliquum (L.) Bentham (Fabaceae). A survey of the natural area was conducted to understand the role that natural areas might play in E. nr. fornicatus ecology in southern Florida. Headspace volatiles from rasped avocado and L. latisiliquum bark were analyzed by gas chromatography-mass spectroscopy (GC-MS) to identify potential attractants. Genetic analysis confirmed that these beetles and their symbiotic fungi are of the same complex that attacks Florida avocado. Gas chromatography-mass spectroscopy analysis indicated that avocado is high in α-copaene (a known attractant of E. nr. fornicatus), but this kairomone is lacking in L. latisiliquum. Host diam and ht were examined for potential influence on colonization behavior. Albizia lebbeck (L.) Bentham (Fabaceae) and an unknown shrub also were observed to be suitable hosts. Concurrent with this study, a nearby grove of soursop, Annona muricata L. (Annonaceae), was found to have infestations of E. nr. fornicatus. Euwallacea nr. fornicatus populations are increasing in Florida and other cultivated and native trees are potentially at risk. Further research is warranted to better understand the ecology of this emerging pest and the chemical cues used for host location.
The introduction of exotic ambrosia beetles (Coleoptera: Curculionidae: Scolytinae) poses a serious threat to U.S. forest health and several agricultural tree fruits. Ambrosia beetles culture symbiotic fungi inside galleries that they excavate in host trees as their sole food source (Ploetz et al. 2013; Hulcr & Stelinski 2017). Although most ambrosia beetles are benign, some introduced species attack and even cause the death of apparently healthy trees (Ploetz et al. 2013). In 2002, the redbay ambrosia beetle, Xyleborus glabratus Eichhoff (Coleoptera: Curculionidae), was discovered in Georgia (Rabaglia et al. 2008). Its primary nutritional symbiont, Raffaelea lauricola T.C. Harr., Fraedrich & Aghayeva (Ophiostomataceae) is a devastating, systemic pathogen of North American Lauraceae, including avocado, Persea americana Mill. (Lauraceae) (Mayfield et al. 2008; Kendra et al. 2014a; Ploetz et al. 2017). Unlike many ambrosia beetles, X. glabratus is not attracted to standard ethanol lures and is, in North America, a specialist on the Lauraceae family. Extensive research examining host preferences among the Lauraceae and their volatile kairomones was instrumental in developing sensitive lures for detection programs (Hanula & Sullivan 2008; Kendra et al. 2011, 2014a, b, 2016a, b).
In 2002 and 2003, exotic ambrosia beetles identified as the tea shot hole borer, Euwallacea fornicatus Eichhoff (Coleoptera: Curculionidae), were discovered in Florida and California (CABI 2017). Subsequent molecular analysis of these beetle populations indicated the presence of multiple cryptic species in the United States, and all species groups in North America are referred to as Euwallacea near fornicatus (O'Donnell et al. 2015; Stouthamer et al. 2017). The primary nutritional symbionts are members of the Ambrosia Fusarium Clade, some of which destroy functional xylem around the galleries, resulting in Fusarium dieback disease (Freeman et al. 2012; Eskalen et al. 2013; Kasson et al. 2013; O'Donnell et al. 2015). In its native southeast Asia range, E. fornicatus attacks dozens of hosts in 35 plant families, including economically important fruit trees (Danthanarayana 1968). The cryptic species in California also are extremely polyphagous, attacking hundreds of host species (Eskalen et al. 2013). Members of the complex attack avocado in the US and Israel (Carrillo et al. 2012, 2016; Eskalen et al. 2012, 2017; Mendel et al. 2012; Kendra et al. 2017).
Since its discovery in Florida avocado in 2010, E. nr. fornicatus has spread throughout the commercial avocado production region centered around Homestead, Florida, USA (Carrillo et al. 2012, 2015). In early 2016, the beetle first caused extensive damage to avocado (cv. ‘Donnie,’ planted in 2004) in a single grove, killing branches up to 12.7 cm diam. This outbreak prompted a survey for E. nr. fornicatus and revealed abundant beetle populations in several groves that did not exhibit signs of significant beetle injury (sawdust sticks, sugary exudate, dead twigs and branches), suggesting that beetles were originating from either visually inaccessible locations in the groves or migrating from non-avocado hosts (Carrillo et al. 2016). One such grove with large beetle numbers captured in traps was located near a natural area. Very few hosts from Florida have been recorded, despite the growing populations and potential for extreme polyphagy. Previous host records include mango (Mangifera indica L., Anacardiaceae), swamp bay (Persea palustris [Raf.] Sarg. [Lauaraceae]), and royal poinciana (Delonix regia [Boj. ex Hook] Raf [Fabaceae]) (Rabaglia et al. 2006; Carrillo et al. 2012, 2016). We hypothesized that tree species in the natural area may serve as a source of beetles flying to the trap located in the avocado grove. Furthermore, if additional, acceptable hosts were discovered, analysis of their volatile profiles might help guide further bioassays to develop improved lures, as was the case with X. glabratus.
Materials and Methods
HOST SURVEY
The diam of all trees greater than 2.5 cm (measured at a ht of 1.5 m) within the 0.37 hectare natural area (located at the UF/IFAS Tropical Research and Education Center, 25.5073°N, 80.5038°W) were measured and examined for signs of ambrosia beetle activity during Dec 2016. When beetle damage was observed, galleries were examined for the presence of E. nr. fornicatus. In May 2017, E. nr. fornicatus damage was identified by the owner of a soursop grove [Annona muricata L. (Annonaceae)]. The grove was visited to assess damage and symptoms.
BEETLE IDENTIFICATION AND COMMUNITY COMPOSITION
A total of 5 beetles and 5 samples of xylem tissue around the galleries were collected from 2 Lysiloma latisiliquum (L.) Bentham (Fabaceae) trees. Fungal isolation followed methods by Carrillo et al. (2016). Beetles were surface-disinfected with 70% ethanol, and xylem tissue was disinfected with 10% bleach and 70% ethanol prior to Fusarium isolation. The beetle head macerate and sapwood tissue were plated on potato dextrose agar supplemented with 0.1g per L streptomycin (PDA+). Fusarium-like colonies were counted on the agar plates for the beetle head isolates, and distinct morphologies were individually cultured on PDA+ for both wood tissue and beetle isolates. Pure isolates were obtained by monosporic cultures.
Total genomic DNA was extracted from the beetle bodies and from the mycelia of fungal isolates following a modified cetyl trimethylammonium bromide (CTAB) protocol (Doyle and Doyle 1987). For Fusarium spp. identification, portions of the EF-1α and DNA-directed RNA polymerase II largest and second largest subunits RBP1 and RBP2 were amplified using primers EF1 and EF2 (O'Donnell et al. 1998), R8/R8 (O'Donnell et al. 2010) and 5f2 / 7cr and 7cf / 11ar (Liu et al. 1999). For beetle identification, portions of the cytochrome oxidase subunit (COI), elongation factor-1α (EF-1α), CAD protein, 16s mtDNA, and 28s rDNA were amplified by PCR using primers sets LCO1490/ HCO2198, ets149/efa754, apCADforB2/ apCADrevlmod, forB2/rev and B2 LRJ-12961/LR-N-13398 and D2F1/D3R2; 3665/4048 according to Dole et al. (2010). PCR products were purified using ExoSAP-IT (Affymetrix, Inc., Santa Clara, California, USA) and sequenced in both directions. Nucleotide alignments were performed in the NCBI Basic Local Alignment Search Tool (BLAST) for sequence identification.
The ambrosia beetle species composition was estimated from 4 dead L. latisiliquum trees and from 4 avocado branches for comparison. Polypropylene micro-centrifuge tubes (1.7 mL; Thomas Scientific, Swedesboro, New Jersey, USA) were used to trap ambrosia beetles emerging from their natal galleries. The ends of centrifuge tubes were cut and covered with screening material. Tanglefoot® trap (The Scotts Co., Marysville, Ohio, USA) coated the inside of the centrifuge tube and tubes were glued onto the bark of infested trees to enclose gallery entrance holes. Galleries were selected according to their size; galleries that were obviously too small for E. nr. fornicatus were not sampled. Very few small gallery entrances were present. Sampled galleries were located 0.3 to 6 m above-ground. Thirty tubes were deployed per L. latisiliquum tree. One-hundred-four vials were attached to avocado branches. Branches were removed from trees and placed in plastic emergence containers (167 liter Brute® containers, Rubbermaid Commercial Products, Winchester, Virginia, USA) prior to trap tube attachment. Intercepted scolytines were identified to species according to Rabaglia et al. (2006) and Atkinson et al. (2013).
DENSITY OF ENTRANCE HOLES ON LYSILOMA LATISILIQUUM
Gallery density of 4 dead L. latisiliquum with active beetle infestation was estimated every 1.5 m above the ground on the main trunk and at several points on secondary branches until section diam was less than 2.5 cm. At each location, galleries were counted from sections measuring either 929 cm2 (basal trunk locations), 465 cm2 (upper trunk locations), or 232 cm2 (small branches).
CHEMICAL ANALYSIS
Headspace volatiles were collected and analyzed from 5 samples of rasped outer bark of apparently healthy, non-infested L. latisiliquum and 1 sample of avocado (c.v. ‘Donnie’) using methods similar to Niogret et al. (2011). Ten grams of the outer bark was rasped with a micro plane. Bark shavings were transferred to a glass cylinder (10 cm diam × 44 cm length) for headspace collection. A purified air stream flowed over the wood shavings at a rate of 1 L per min and was pulled through a super-Q filter by vacuum for 1 h. Volatiles then were eluted from the filter with 200 µl of methylene chloride (99.8% pure, Avantor Performance Materials, Inc., Center Valley, Pennsylvania, USA). For quantification, a hexadecane (CAS Registry No. 544-76-3, Sigma-Aldrich, St. Louis, Missouri, USA) internal standard was added to each sample to make a solution of 2.5 µg hexadecane per µl of sample.
Samples were injected into a gas chromatograph-flame ionization detector (GC-FID, Trace GC2000, Thermoquest Corporation, Austin, Texas, USA) and a gas chromatograph-mass spectrometer (Agilent 5975B, Agilent Technologies, Santa Clara, California, USA). The column of both instruments was a DB5-MS capillary column (25 m × 0.25 mm × 0.25 µm, Agilent Technologies). Oven temperature was programmed at 45 °C for 1 min, then to 94 °C at a rate of 4 °C per min, then to 180 °C at a rate of 2 °C per min, and then to 240 °C at a rate of 20 °C per min. Gas chromatograph-mass spectrometer chromatograms were analyzed with MassHunter software (B.07.02, Agilent Technologies) and compounds were identified by correlating mass spectra to the NIST (2011) database, the Adams Library (Adams 2007), and our own “SHRS Essential Oil Constituents” library, authentic samples, and the components of known essential oils (Kendra et al. 2011; Ali et al. 2014; Blythe et al. 2016; Kendra et al. 2017). The authentic compounds [cyclosativene (CAS Registry No. 22469-52-9); β-elemene (CAS Registry No. 515-13-9); β-caryophyllene (CAS Registry No. 87-44-5); alloaromadendrene (CAS Registry No. 25246-27-9); (Z)-3-hexen-1-ol (CAS Registry No. 928-96-1); eucalyptol (CAS Registry No. 470-82-6)] were purchased from Sigma-Aldrich, St. Louis, Missouri, USA, and α-copaene (CAS Registry No. 3856-25-5) was from Fluka Chemical Co. (Buchs, SG, Switzerland). Retention indices (RI) were calculated according to the method of van Den Dool & Kratz (1963) using a homogenous series of n-alkanes (C8-C40). Chemical quantity was calculated by comparing gas chromatograph-flame ionization detector peak area to the area of the hexadecane standard (2.5 µg per µl).
DATA ANALYSIS
The number of E. nr. fornicatus galleries in the 4 measured trees was estimated by multiplying the counted gallery entrances in a sample by the area of a cylinder for the corresponding trunk section. This estimate then was multiplied by the tree's proportion of E. nr. fornicatus-identified galleries from the trap samples. The effect of section diam and ht of the tree on the number of galleries was analyzed with analysis of variance and Tukey's Test for mean separation (Systat Software 2010).
Results
HOST SURVEY
Of the 536 trees measured in the natural area, 68 could not be reliably identified. Lysiloma latisiliquum was the dominant species present, comprising 55.4% of the stand (Table 1). Thirty-nine dead or dying L. latisiliquum trees exhibited ambrosia beetle attack symptoms consistent with trees infested by E. nr. fornicatus. In living branches and on small dead trees and branches, galleries often were concentrated at nodes. On large dead trees, high densities of beetle galleries were present on the main trunk and branches (Fig. 1). Dried sap flow was observed on the trunks of several live or recently killed trees infested with L. latisiliquum. The nearest L. latisiliquum to a survey trap in the adjacent avocado orchard was 15 m.
A single Albizia lebbeck (L.) Benth. (Fabaceae) was damaged by E. nr. fornicatus, and another dead A. lebbeck had E. nr. fornicatus galleries. One shrubby plant that could not be identified had several stems with E. nr. fornicatus galleries. Only 2 leaf clusters remained on a single stem when the plant was discovered (Table 1).
There were 2 dead branches on a single gumbo limbo tree, Bursera simaruba (L.) Sargent (Burseraceae) with ambrosia beetle galleries. Although E. nr. fornicatus was not recovered from the galleries, 2 specimens were removed from sap flowing from galleries located at the base of the branch. Ambrosia beetle injury, if caused by E. nr. fornicatus, was similar to avocado damage, where beetles concentrate on secondary branches without attacking the main trunk.
The soursop grove infested by E. nr. fornicatus was located 2.0 km southwest from the infested natural area. Damage was light, and limited to the lower, deeply shaded branches. Trees had a dense canopy and were planted about 2 m apart. Euwallacea nr. fornicatus was observed in several small branches that began dying. Gallery entrances mostly were located at nodes. The grove owner practiced regular sanitation, pruning infested branches and destroying them.
BEETLE AND FUNGI IDENTIFICATION
All 5 beetles removed from galleries for genotyping matched previous sequences for E. nr. fornicatus sp. 2, also identified as the type species with the common name tea shot hole borer (for nomenclature discussion, see O'Donnell et al. 2015 and Stouthamer et al. 2017; Table 2). Fusarium sp. AF-8 was isolated from 4 beetles with a mean of 680 ± 143 (SE) colony forming units per beetle and from galleries in both sampled trees. Fusarium sp. AF-9 was recovered from a single E. nr. fornicatus gallery in 1 tree, but was not recovered from E. nr. fornicatus.
The percentage of E. nr. fornicatus galleries sampled with the modified centrifuge tubes ranged from 28.6 to 83.9% on the 4 dead L. latisiliquum trees. Ambrosia beetles were not recovered from 15 of the 120 vials, and 4 vials had fallen from the tree and were not recovered. Theoborus ricini Wood & Bright (Coleoptera: Curculionidae) accounted for between 12.9 to 60.7% of the galleries. Single Xyleborus ferrugineus Fab. and Xyleborus bispinatus Eichhoff (Coleoptera: Curculionidae: Scolytinae) galleries were identified from a single dead tree. In avocado, the percentage of confirmed E. nr. fornicatus galleries ranged from 41.7 to 75.6% (mean 60.6%). No emergence or beetle activity was observed from 16 galleries, and ambrosia beetle activity was detected in 17 vials (heavy sawdust deposition), but beetles either escaped or failed to be captured. Only 8 vials captured T. ricini.
ATTACK DENSITY
Galleries were counted at 36 sites (diam of trunk or branches > 2.5 cm) combined from the 4 trees. Trunk diam ranged between 5.3 and 25.1 cm. Average gallery density was 0.11 galleries per cm2. Extrapolating over the entire tree, suitable L. latisiliquum can support a significant E. nr. fornicatus population (Table 3). Gallery density was influenced by tree section diam, with the smallest diam having fewer attacks (F = 4.88; df = 6, 29; P = 0.001; Fig. 2a). Gallery density was not influenced by ht (F = 0.27; df = 5, 30; P = 0.926; Fig. 2b).
HEADSPACE ANALYSIS
The avocado sample contained numerous terpenoids, 12 of which were identified (Table 4). The most abundant was α-copaene. Four chemicals were identified from L. latisiliquum (Table 4). The green leaf volatiles (Z)-3-hexen-1-ol and 1-hexanol were emitted in large quantity. Anisol also was detected in 4 of the 5 samples. Eucalyptol, a monoterpene ether, was the only terpenoid detected from L. latisiliquum, and in low quantity.
Table 1.
Plant species present in the Florida natural area in which Euwallacea nr. fornicatus was discovered in Lysiloma latisiliquum. Hosts infested with Euwallacea nr. fornicatus are marked with an asterisk.
![t01_229.gif](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/t01_229.gif)
Discussion
The discovery of infested native trees indicates that natural areas in Florida may function as reservoirs for pest E. nr. fornicatus populations capable of dispersing to avocado and other cultivated fruit trees. This is the first time that the beetle has been reported from L. latisiliquum and Annona muricata. Lysiloma latisiliquum is widely distributed throughout Central America, the Caribbean, and southern Florida. It is a pioneer species, capable of growing on calcareous soils common in Miami-Dade County (Brown & Cooprider 2013). Euwallacea nr. fornicatus may become an increasingly serious threat to native forest stands, as E. nr. fornicatus populations have in California (Eskalen et al. 2013; Boland 2016). At the time of sampling, 13% of the L. latisiliquum trees in the natural area were infested by E. nr. fornicatus, and gallery counts indicated that infested L. latisiliquum can support significant E. nr. fornicatus populations. It is highly likely that beetles from this natural area were captured by the nearby survey trap in the avocado grove.
Other Albizia spp. and members of the Burseraceae family have been recorded as hosts for the tea shot hole borer in Asia (Danthanarayana 1968). The polyphagous shot hole borer (Euwallacea sp.) in California also attacks Bursera spp. (Eskalen et al. 2013). Although E. nr. fornicatus was not recovered from galleries in B. simaruba, beetles were removed from sap flow at the branch collar. This suggests that either B. simaruba is a host in Florida, it contains attractive kairomones, or E. nr. fornicatus was responding to fungal volatiles produced by galleries of previous ambrosia beetles colonizing the wood (X. glabratus is attracted to symbiont fungal volatiles [Kuhns et al. 2014a]). A third species in the natural area that was colonized by E. nr. fornicatus could not be identified, and it is very likely that other native plant species are hosts for E. nr. fornicatus. Euwallacea nr. fornicatus has been recovered only from a few plant species in Florida, while in California and Asia, hundreds of species are attacked by the E. nr. fornicatus complex (Danthanarayana 1968; Amarasinghe & Devy 2003; Eskalen et al. 2013). It is interesting to note also that Fusarium species AF-9 previously has been identified only from beetles in Costa Rica (Kasson et al. 2013). It is not known how different Fusarium species affect host selection and colonization by E. nr. fornicatus.
Fig. 1.
Ambrosia beetle gallery entrances in the trunk of a Lysiloma latisiliquum. Euwallacea nr. fornicatus and Theoborus ricini were the two most abundant species of ambrosia beetle recovered from L. latisiliquum.
![f01_229.jpg](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/f01_229.jpg)
In avocado, beetle attack usually is confined to the secondary branches and activity often does not extend past the branch collar, but in L. latisiliquum, beetles readily colonized the main trunk. Euwallacea nr. fornicatus galleries were present only on soursop branches, not the trunk, and galleries were concentrated at nodes. On L. latisiliquum, there were differences in beetle colonization at locations of differing diam but not ht. Differences in colonization behavior within a host and among host species could be due to either visual or semiochemical cues. Xyleborus glabratus prefers larger diam hosts (Mayfield & Brownie 2013; Kendra et al. 2013). At the same time, Niogret et al. (2013) discovered terpenoid gradients among avocado trunks, branches, leaf petioles, and leaves, with some being produced in greater quantity in the trunk and others by smaller branches and leaves. The lack of terpenoid emissions from L. latisiliquum indicates that E. nr. fornicatus is responding to other volatiles. Eucalyptol was detected in low quantity from rasped bark samples. This chemical has been identified previously as an attractant for another primary ambrosia beetle colonizer, X. glabratus (Kuhns et al. 2014b). Subsequent studies found eucalyptol to be less attractive than essential oil lures high in sesquiterpenes (Kendra et al. 2016b). The most abundant volatile from avocado was α-copaene, which was consistent with previous reports. This chemical is also a known attractant for both E. nr. fornicatus and X. glabratus (Kendra et al. 2016a, 2017). Both β-caryophyllene and germacrene-D, identified from avocado, are potential attractants for E. nr. fornicatus (James 2007). However, lures containing pure β-caryophyllene, presented alone or in combination with essential oil lures (Kendra et al. 2016a), were not attractive to X. glabratus.
Table 2.
Genotyping results of Euwallacea nr. fornicatus and associated Fusarium collected from Lysiloma latisiliquum in southern Florida.
![t02_229.gif](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/t02_229.gif)
Fig. 2.
Relationships among host tree diameter, height above ground, and site of attack by Euwallacea nr. fornicatus. Density of beetle entrance holes versus the trunk or branch diameter (A) and the trunk or branch height (B) of host Lysiloma latisiliquum from 4 trees. Mean values topped by the same letter are not significantly different (Tukey's test, P = 0.05).
![f02_229.jpg](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/f02_229.jpg)
Table 3.
Proportion and estimated number of Euwallacea nr. fornicatus gallery entrances on four infested, dead Lysiloma latisiliquum trees.
![t03_229.gif](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/t03_229.gif)
It is not known what physiological condition the L. latisiliquum trees were in at the time of initial E. nr. fornicatus colonization, although Nov and Dec 2016 weather was unusually hot and dry. Trees under severe drought stress produce ethanol (Kimmerer & Kozlowski 1982), and ethanol in combination with other host volatiles is attractive to E. fornicatus in Asia (Karunaratne et al. 2008). However, there are indications that the presence of ethanol decreases trap captures of the E. nr. fornicatus species present in California (Dodge et al. 2017). Analytical chemistry techniques sensitive to low-molecular weight compounds, such as solid phase microextraction, may further aid in identification of attractive kairomones.
Table 4.
Headspace volatile identification from 10 g rasped outer bark volatiles collected onto super-Q for one hour.
![t04_229.gif](ContentImages/Journals/flen/101/2/024.101.0212/graphic/WebImages/t04_229.gif)
To date, the only commercial attractants for E. nr. fornicatus are quercivorol (Carrillo et al. 2015) and a proprietary essential oil enriched in α-copaene (Kendra et al. 2017). The combination of both lures results in additive or synergistic increase in beetle capture, depending on population levels (Kendra et al. 2017). Identification of additional hosts is an important step toward the identification of other attractive terpenoids that could improve attractiveness of field lures for E. nr. fornicatus. With X. glabratus, it has been hypothesized that optimal host location is achieved by detection of a complex terpenoid mixture, a ‘signature bouquet’ of the Lauraceae (Kendra et al. 2014a). In Sri Lanka, 7 plant species are reported to be more attractive to E. nr. fornicatus than tea (Amarasinghe & Devy 2003). One of these species, Jacaranda mimosifolia Don (Bignoniaceae), is widely planted in southern Florida. Further studies comparing the volatile emissions from reported and observed hosts, including avocado and Annona spp., may identify unique attractants that potentially could improve detection of E. nr. fornicatus in Florida.
Acknowledgments
Cliff Martin (UF TREC) assisted with tree identification. Carlos de la Torre provided avocado samples and Carlos Ugalde with Shangri La Farms identified the damage to soursop. David Jenkins (SC Forestry Commission) and Gregg Nuessly (UF) provided additional reviews that helped improve this manuscript. This research was supported in part by an appointment to the Agricultural Research Service (ARS) Research Participation agreement between the U.S. Department of Energy (DOE) and the USDA. ORISE is managed by ORAU under DOE contract number DE-SC0014664. All opinions expressed in this paper are the authors' and do not necessarily reflect the policies and views of USDA, ARS, DOE, or ORAU/ORISE.