Daniel Bravo-Pérez, Ma. Teresa Santillán-Galicia, Roberto M. Johansen-Naime, Héctor González-Hernández, Obdulia L. Segura-León, Daniel L. Ochoa-Martínez, Stephanie Guzman-Valencia

1 Posgrado en Fitosanidad-Entomología y Acarología, Colegio de Postgraduados, Km 36.5 Carretera México-Texcoco, Montecillo,Texcoco, Estado de México 56230, México

2 Instituto de Biología de la Universidad Nacional Autónoma de México (IBUNAM), Circuito exterior s/n, Ciudad Universitaria,Coyoacán, Ciudad de México 04510, México

3 Posgrado en Fitosanidad-Fitopatología, Colegio de Postgraduados, Km 36.5 Carretera México-Texcoco, Montecillo, Texcoco,Estado de México 56230, México

Abstract Avocado is one of the most important crops in the world, and Mexico is the largest producer of this fruit. Several insect pests affect its production, and thrips are amongst the most important. A key step in the design of control methods is accurate species identification. Despite this, formal reports on species diversity of thrips in Mexico are very scarce. Morphological identification can sometimes be time-consuming and inconclusive. Therefore, we explored the species diversity of thrips in Mexican avocado orchards (Michoacan state) based on partial sequences of the mitochondrial gene cytochrome oxidase subunit I (COI). Forty-four specimens were analysed, which represented approximately 8% of all individuals collected from five localities distributed in three Municipalities. All specimens were analysed using the COI marker, and specimens within the genera Frankliniella were also analysed using a marker within the D2 domain of the 28S (28SD2) nuclear ribosomal DNA.Molecular identifications were confirmed using morphological taxonomy. Overall, six genera were found (Neohydatothrips,Scirtothrips, Frankliniella, Arorathrips, Caliothrips and Leptothrips). All genera contained only one species, except Frankliniella,for which there were six species. Data from the two molecular markers suggest the existence of cryptic species within Mexican F. occidentalis populations.

Keywords: molecular taxonomy, morphological identification, cytochrome oxidase subunit I, 28S nuclear ribosomal DNA.

1. Introduction

Mexico is the world’s main producer of avocados (FAO 2014). Annually, avocado producers spend more than 973 USD ha-1on the control of avocado pests (APROAM 2014). Amongst the most important insect pests are thrips(Ramírez-Dávila et al. 2013). Thrips are sucking pests,that remove the cellular content from various plant tissues and damage epidermal cells (Ascención et al. 1999).Consequently, necrotic areas appear and, when these occur on the fruit, they cause scars and the fruit becomes unmarketable (Marroquín 1999). When thrips feed on very young avocado fruits they can cause malformation and early fruit drop (Johansen et al. 2007). In California, infestation by Scirtothrips perseae Nakahara in avocado orchards causes economic losses of 7.6-13.4 million USD every year (Hoddle et al. 1999). As a consequence, a number of different strategies have been used to control thrips populations in the field and reduce the chances of avocado fruits becoming infested (Johansen et al. 2007). Although thrips are economically important for avocado production in Mexico, formal descriptions of the species composition and abundance of thrips are limited.

Some reports suggest that more than 80 species of thrips can be found on avocado in Mexico, mainly from the genera Scirtothrips, Neohydatothrips and Frankliniella(Johansen et al. 2007). However, it is possible that some of these species do not use avocado as a host plant and were just randomly dispersed there by wind from their true host plant (Mound 2005). This could lead to a potential overestimation of the actual number of species present in avocado orchards. Understanding the precise host plant and host range of an insect is important for a number of studies in ecology, speciation and coevolution of the insect-host plant association (Jurado-Rivera et al. 2009).

Accurate identification of insects is fundamental for any study, particularly for the development of pest management strategies, where correct identification is essential to understanding the biology and distribution of a pest over time (Danks 1988). Identification of thrips to species level requires great expertise and, in many cases, it is only possible using adults (Brunner et al. 2002; Przybylska et al. 2015). Despite reports of successful identification of larval stages of thrips (Skarlinsky and Funderburk 2016),the use of larval stages for accurate identification is still not widely taken up (Mound 2013). In addition, intraspecific morphological variation in some thrips species, such as S. persea (Hoddle et al. 2008) and F. occidentalis Pergande(Rugman-Jones et al. 2010) is large, making identification based solely on morphological characters questionable;accurate identification should, therefore, include the use of molecular techniques (Rugman-Jones et al. 2006).

Here we present a formal report of the species composition of thrips collected from avocado orchards in the state of Michoacan, Mexico, the main avocado (var. Hass)production area in Mexico. We identified the specimens based on molecular data using the sequence information from the mitochondrial cytochrome oxidase subunit I (COI),and, for species within the genus Frankliniella, we also used the D2 domain of the 28S (28SD2) nuclear ribosomal DNA (rDNA); we then confirmed these identifications from morphological attributes. In this study, we aimed to provide an accurate list of species on avocado that can be used as a basis for pest management decision making. Furthermore,all sequence information we produced can be used as a reference for molecular identification of larvae or eggs.

2. Materials and methods

2.1. Thrips sampling

Thrips were collected from four commercial avocado orchards (approx. 1 ha area/orchard), each from a different locality within the three municipalities of Michoacan State,Mexico (Table 1). The samples were taken between August 2013 and March 2014 and only from trees with visible signs of thrips damage on the leaves or fruits. Although thrips were collected from randomly selected infested trees, we did try to ensure they were representative of the complete sampled area. From each tree, four young leaflets 1.5 m above the ground were sampled at each cardinal point in the canopy. Each leaflet was sprayed with a 10% soap solution inside a square plastic container (40 cm×40 cm×20 cm) to dislodge the thrips. Later, thrips from all four leaflets were combined to form a single sample from each tree. Therefore,each tree was considered as the sampling unit. The soap solution containing the thrips was filtered through an antiinsect mesh inside the plastic container. All thrips that were trapped in the mesh were collected using a fine brush and deposited into a 1.5-mL Eppendorf tube containing 96%ethanol. The collection procedure was done in the field, and the tubes containing the thrips were taken to the laboratory and maintained at -20°C until required. Successful DNA extractions were obtained with samples stored at -20°C for up to one and a half years. A total of 66 samples (trees)were processed from all locations (Table 1).

2.2. Phylogenetic analysis of thrips species

Phylogenetic placement was made from DNA extracted from individual adult thrips specimens, followed by morphological identification of that specimen to confirm taxonomic status.Overall, we obtained approximately 600 specimens from which we selected 44 specimens for genetic analyses, based on visual observations using a stereomicroscope.

2.3. DNA extraction

For DNA extraction, thrips were retrieved from the 96%ethanol and placed onto sterile filter paper for 5 min to remove the excess ethanol. Each insect was then placed onto a sterile glass slide and pierced in the abdominal region with a sterile size (0.25 mm×40 mm) entomological needle(Fine Science Tools, Inc., Foster City, CA, USA) (Rugman-Jones et al. 2006). DNA was extracted using a DNeasy Blood & Tissue Kit (Qiagen?). Immediately after being pierced, the insect was deposited into a 1.5-mL Eppendorf tube and incubated at 60°C for 18 h in a mixture of ATL buffer and proteinase K (both provided with the kit); thereafter we followed the manufacturer’s instructions to extract the DNA.At the final step of the extraction process, the DNA was eluted from the membrane of the DNeasy Mini spin column provided with the kit. The membrane and the remains of the specimen were carefully removed from the spin column and deposited into a Syracuse watch glass (25 mm×8 mm)containing absolute ethanol; under a stereomicroscope,the specimen was carefully removed and mounted in Canada balsam for morphological identification. Briefly,each specimen was first dehydrated in 100% ethanol for 15 min, followed by a clearing step in which each specimen was immersed in a xylene and clove oil (40:1) solution for 5 min. Cleared thrips were mounted on glass slides with Canada balsam. Genus and species confirmation was made using the taxonomic keys of Mound and Marullo(1996) and Johansen and Mojica (1998). For Leptothrips species, the taxonomic keys of Johansen (1987) and Mound and O′Donnell (2017) were used. Species within the genus Neohydatothrips were identified based on Nakahara (1988)and Lima and Mound (2016).

Table 1 Thrips collection sites in the state of Michoacan, Mexico

2.4. PCR and sequencing

Partial sequences of the COI gene were obtained using the primers LCO1490 (GGTCAACAAATCATAAAGATATTGG)and HCO2198 (TAAACTTCAGGGTGACCAAAAAATCA)(Folmer et al. 1994). For the D2 domain of the 28S nuclear ribosomal DNA, amplifications were made using the primers 28sF3633 (TACCGTGAGGGAAAGTTGAAA)and 28sR4076 (AGACTCCTTGGTCCGTGTTT) (Rugman-Jones et al. 2010). PCRs were done in reaction volumes of 20 and 25 μL for the COI and 28S markers, respectively.Each reaction contained 1×PCR buffer (Tris-Cl, KCl,(NH4)2SO4,15 mmol L-1MgCl2; pH 8.7), 0.2 mmol L-1of each primer, 0.8 mmol L-1of dNTPs, 0.5 U of TaqDNA polymerase (Qiagen?, GmbH, Hilden, Germany), 1.5 mmol L-1of MgCl2(Qiagen) and 2 μL (approx. 40 ng) of DNA.PCR amplifications were made using a MyCyclerTM(Bio-Rad Laboratories Inc., Hercules, CA, USA). Thermal conditions for the COI gene were one cycle of 60 s at 94°C, four cycles of 30 s at 94°C, 90 s at 45°C and 60 s at 72°C, followed by 35 cycles of 30 s at 94°C, 90 s at 51°C and 60 s at 72°C with a final extension at 72°C for 5 min. Conditions for the 28S rDNA were one cycle of 2 min at 94°C, followed by 38 cycles of 30 s at 94°C, 50 s at 58°C and 75 s at 72°C with a final extension at 72°C for 10 min. PCR products were visualized on 1% agarose gels in 1×TAE. GelPilot?50 and 100 bp Plus (Qiagen?, GmbH, Hilden, Germany)size markers were used. The gels were stained with ethidium bromide (0.1 μg mL-1) and photographed. The PCR products were sent to Macrogen Inc. (South Korea)for purification and direct sequencing.

2.5. Data analysis

Sequence traces were manually assembled using BioEdit(Hall 1999). Multiple sequence alignments were made using the Clustal W programme (Thompson et al. 1994).

After alignment and trimming, the final length of the COI sequences was 366 bp for the 44 specimens (Table 2).For the Frankliniella specimens, the final length of the 28S sequences was 466 bp (Table 2).

Phylogenetic relationships amongst the species were assessed using maximum parsimony (MP) and maximum likelihood (ML), both with the nearest neighbour interchange algorithm, and neighbour-joining (NJ) using the p-distance method. All analyses were done in the Molecular Evolutionary Genetic Analysis (MEGA) Software ver. 5.0 for Windows (Tamura et al. 2011). The robustness of branches was estimated by bootstrap analysis with 1 000 repeated samples from the data (Felsenstein 1985). If available,sequences were obtained from GenBank and used as reference in all analyses.

Genetic differences amongst haplotypes within F. occidentalis, the most abundant species, were detected in a maximum parsimony network (Templeton et al. 1992),using TCS v.1.21 (Clement et al. 2000) separately for the COI and 28S partial sequences obtained for this species. When analysing the 28S sequences, six additional sequences retrieved from GenBank were included, three from F. occidentalis specimens collected in the USA (‘wilderness populations’; GU147942, GU147943 and GU147944), andthree collected in New Zealand (‘glasshouse populations’;GU148000, GU148001 and GU148002), representing two genetically different lineages (Rugman-Jonas et al. 2010).No COI sequences targeting the same region that we amplified could be found in the reference specimens. The connection limit amongst haplotypes (limits of parsimony)was set to the default value of 95%.

Table 2 List of specimens sequenced for phylogenetic analysis1)

3. Results

Fig. 1 Dendrogram inferred from maximum parsimony (MP), maximum likelihood (ML) and neighbour-joining (NJ) analyses of mitochondrial cytochrome oxidase subunit I (COI) data on thrips collected in avocado trees. Sequences used as references are in bold and labelled according to their GenBank accession numbers. Only bootstrap values above 80% for the three analyses are shown.

Phylogenetic analyses performed with the COI sequences successfully separated specimens from six genera, all with bootstrap values above 80% (Fig. 1). All phylogenetic placements were confirmed by morphological identification.The following six genera were identified: Scirtothrips,Neohydatothrips (Sericothripini), Frankliniella (Thripini),Leptothrips (Dendrothripini), Arorathrips (Chirothripini)and Caliothrips (subfamilily Panchaetothripinae). Each genus contained only one species except for the genus Frankliniella. The most abundant specimens were those within the genera Scirtothrips and Frankliniella (Table 2).The following species within each genus were identified:Scirtothhrips perseae, Neohydatothrips signifer Priesner,Arorathrips mexicanus Priesner, Caliothrips marginipennis Hood. Only one predatory species was found: Leptothrips mcconnelli (Crawford). Six species were found in the Frankliniella genus: F. occidentalis (Pergande), F. gardeniae(Moulton), F. borinquen (Hood), F. brunnea (Priesner),F. rostrata (Priesner) and F. insularis (Franklin). The presence of these species within the genus Frankliniella was confirmed by all three analyses, with bootstrap values above 80% for both COI sequences (Fig. 1) and 28S sequences (Fig. 2).

Fig. 2 Dendogram of Frankliniella (Thysanoptera: Thripidae) species inferred from maximum parsimony (MP), maximum likelihood(ML) and neighbour-joining (NJ) analyses of the 28S sequence data. Sequences used as references are labelled according to their GenBank accession numbers. Only bootstrap values above 80% are shown.

Haplotype network analysis using COI sequences from F. occidentalis, revealed the existence of five haplotypes.The most common haplotype contained specimens D13,D56, D49 and D50, followed by the haplotype containing specimens D36 and D16, and haplotypes with specimens D55 and D15. In addition, one independent haplotype was obtained containing specimen D40 (Fig. 3-A). When haplotype network analysis was done using the 28S sequences, three haplotypes were obtained. The most common haplotype contained specimens D36, D15, D55,D13, D56, D49, D50 and the three sequences used as a reference from F. occidentalis specimens collected in the USA (‘wilderness populations’). The second most abundant haplotype contained specimen D16 and the three sequences used as a reference from F. occidentalis specimens collected in New Zealand (‘glasshouse populations’), followed by one haplotype with specimen D40 (Fig. 3-B).

4. Discussion

Despite their pest status on avocado, formal reports on the species diversity of thrips in Mexico are surprisingly scarce, or difficult to access. Here we present information regarding the species diversity of thrips on avocado in Mexico that are based on collections of samples from five localities in Michoacan state (Table 1). Overall, the key species found in avocado orchards using sequence information were S. perseae,F. occidentalis and F. gardeniae with more than five specimens for each (Table 2). All other species were represented by only one or two specimens (Table 2). Only one potentially predatory species, L. mcconnelli, was found.However, experimental evidence for its predatory feeding behaviour requires confirmation; previous publications on Leptothrips species have generally just ‘a(chǎn)ssumed’ them to be predatory (Mound and O’Donnell 2017). It is difficult to guarantee that all species we found are the only species present in our samples, but we certainly believe that we selected the most abundant and therefore the most important in the avocado orchards we sampled.

When sequences from genera other than Scirtothrips or Frankliniella were analyzed, species separation was evident with significant bootstrap values (Fig. 1), and in accordance with morphological identification. Sequences identified morphologically as N. signifer (Table 2), were allocated to the same group with a highly significant bootstrap value;however, they were also grouped with the sequence of a specimen identified as N. burungae retrieved from GenBank.Neohydatothrips burungae was considered as a synonym of N. signifer by Mound and Marullo (1996) while other researchers consider them as separate species, and that N. signifier is only found in Mexico (Hoddle et al. 2012).Based on our results we consider that N. signifer and N. burungae are likely to be the same species. However,we also believe there is a need for more morphologicallyidentified specimens from both species to be studied at the molecular level to obtain more robust results, as suggested by Lima and Mould (2016).

Previous reports of Scirtothrips species in Mexican avocado orchards, have reported at least 20 different species(Johansen et al. 2007), which is different to our results where we only found S. perseae. We believe that this difference may be because the report by Johansen et al. (2007) was based on several other non-published reports that included the analysis of hundreds of individuals. It is likely that some of these species may have just been randomly dispersed by the wind from their true host plants onto the non-host plant, avocado (Mound 2005); this would overestimate the species diversity. Furthermore, most of these non-published reports were based only on morphological identification and, as reported previously for the genus Scirtothrips, some discrepancies in morphological identification can result in overestimations in species diversity (Hoddle et al. 2008); a greater number of specimens need to be analysed in order to confirm this. In addition, Rebijith et al. (2014) reported the possibility of cryptic species within S. perseae; we found no evidence for this, but it is very likely that this was because the number of specimens we studied was small(15 samples) compared with the number (47) analysed by Rebijith et al. (2014).

Fig. 3 Most parsimonious haplotype network for Frankliniella occidentalis (Pergande 1895) (Thysanoptera: Thripidae) specimens for the five haplotypes found using mitochondrial cytochrome oxidase subunit I (COI) sequences (A) and the three haplotypes using 28S sequences (B). Haplotypes are connected with a 95% confidence limit. Each line in the network represents a single mutational change. Small black circles indicate missing haplotypes.

Six species were found within the genus Frankliniella as confirmed by both molecular markers (Figs. 1 and 2). Species identification was very clear in F. borinquen, F. brunnea,F. rostrata and F. insularis, which were all confirmed by the morphological identification. Significant genetic variation was found amongst the F. occidentalis specimens,with both markers used (Figs. 1 and 2), suggesting the existence of cryptic species within F. occidentalis in Mexican avocado orchards. Similar results have been reported previously for F. occidentalis (Rugman-Jones et al. 2010;Rebijith et al. 2014). Although only nine specimens from F. occidentalis were analysed in our study, the haplotype network analysis using sequences from both markers,showed the existence of great genetic variation amongst individuals. This was particularly evident when the COI sequences were analysed (Fig. 3-A). Even when only the 28S sequences were analysed, a very clear separation was apparent with the majority of specimens falling into one haplotype which also contained the three reference sequences from thrips collected in the USA (‘wilderness populations’). Specimen D16 was grouped with the other reference sequences from New Zealand (‘glasshouse populations’) (Rugman-Jones et al. 2010). Rugman-Jones et al. (2010) suggested that such separation could be evidence of cryptic speciation within the F. occidentalis complex. In addition, specimen D40 was not included in either the group containing the USA sequences or the New Zealand sequences, suggesting that this specimen may be part of a completely different genetic lineage within the F. occidentalis complex. This was confirmed by analysis of both the 28S and the COI sequences (Fig. 3-A and B);in fact, using the COI sequences, specimen D40 formed an independent haplotype (Fig. 3-A). The use of nuclear and mitochondrial markers as an accurate method to infer speciation has been reported previously (Bensch et al.2004; Gunawardana et al. 2017), using COI sequences successfully separated F. panamensis from F. occidentalis,which are two morphological similar species; moreover,these authors also found two different clades within their F. occidentalis specimens, which was similar to our results.We need to study and analyse more specimens from the species F. occidentalis, including specimens collected from other plant hosts, as F. occidentalis is considered polyphagous (Lewis 1997). Considering that F. occidentalis is one of the most economically important species worldwide(Kirk and Terry 2003), including the avocado system in Mexico (Johansen et al. 2007), the large genetic variation within this species warrants further research, as each haplotype may have different biological attributes such as reproductive capacity, severity of damage or insecticide resistance. Information regarding each haplotype could ultimately help to develop more effective control strategies.

5. Conclusion

Eleven species distributed in five genera were genetically identified in the avocado orchards sampled in our study. All species were phytophagous, except L. mcconnelli, which is thought to be a predator. The genus Frankliniella, with six species, had the greatest species diversity of all the genera.We confirmed the existence of cryptic species within the F. occidentalis complex. Our results represent the first attempt to provide a formal report of the species diversity of thrips in Mexican avocado orchards using a combination of both, genetic and morphological data.

Acknowledgements

Daniel Bravo-Pérez received a scholarship from Consejo Nacional de Ciencia y Tenología (CONACyT), Mexico for his Master’s degree.