Phylogenomics clarifies repeated evolutionary origins of inbreeding and fungus farming in bark beetles (Curculionidae, Scolytinae)

Molecular Phylogenetics and Evolution 127 (2018) 229–238

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev

Phylogenomics clarifies repeated evolutionary origins of inbreeding and fungus farming in bark beetles (Curculionidae, Scolytinae)

T

Andrew J. Johnsona, Duane D. McKennab, Bjarte H. Jordalc, Anthony I. Cognatod, ⁎ Sarah M. Smithd, Alan R. Lemmone, Emily Moriarty Lemmonf, Jiri Hulcra,g, a

School of Forest Resources and Conservation, University of Florida, Gainesville FL 32611, United States Department of Biological Sciences, University of Memphis, Memphis, TN 38152, United States c University Museum of Bergen, University of Bergen, PO 7800, 5020 Bergen, Norway d Department of Entomology, Michigan State University, East Lansing, MI 48824, United States e Department of Scientific Computing, Florida State University, Dirac Science Library, Tallahassee, FL, United States f Department of Biological Science, Florida State University, Tallahassee, FL, United States g Department of Entomology, University of Florida, Gainesville, FL 32611, United States b

A B S T R A C T

Bark and ambrosia beetles (Curculionidae, Scolytinae) display a conspicuous diversity of unusual genetic and ecological attributes and behaviors. Reconstructing the evolution of Scolytinae, particularly the large and ecologically significant tribe Cryphalini (pygmy borers), has long been problematic. These challenges have not adequately been addressed using morphological characters, and previous research has used only DNA sequence data from small numbers of genes. Through a combination of anchored hybrid enrichment, low-coverage draft genomes, and transcriptomes, we addressed these challenges by amassing a large molecular phylogenetic dataset for bark and ambrosia beetles. The resulting DNA sequence data from 251 protein coding genes (114,276 bp of nucleotide sequence data) support inference of the first robust phylogeny of Scolytinae, with a special focus on the species rich tribe Cryphalini and its close relatives. Key strategies, including inbreeding mating systems and fungus farming, evolved repeatedly across Scolytinae. We confirm 12 of 16 hypothesized origins of fungus farming, 6 of 8 origins of inbreeding polygyny and at least 11 independent origins of a super-generalist host range. These three innovations are statistically correlated, but their appearance within lineages was not necessarily simultaneous. Additionally, the evolution of extreme host plant generalism often preceded, rather than succeeded, fungus farming. Of the high-diversity tribes of Scolytinae, only Xyleborini is monophyletic, Corthylini is paraphyletic and Cryphalini is highly polyphyletic. Cryphalini sensu stricto is part of a clade containing the genera Hypothenemus, Cryphalus and Trypophloeus, and the tribe Xyloterini. Stegomerus and Cryptocarenus (Cryphalini) are part of a clade otherwise containing all Corthylini. Several other genera, including Ernoporus and Scolytogenes (Cryphalini), make up a distantly related clade. Several of the genera of Cryphalini are also intermixed. For example, Cryphalus and Hypocryphalus are intermingled, as well as Ernoporicus, Ptilopodius and Scolytogenes. Our data are consistent with widespread polyphyly and paraphyly across Scolytinae and within Cryphalini, and provides new insights into the evolution of inbreeding mating systems and fungus farming in the species rich and ecologically significant weevil subfamily Scolytinae.

1. Introduction 1.1. Background To understand the life history of an organism, it is necessary to

⁎

integrate knowledge of both its biology and its evolutionary past. Scolytinae (Coleoptera: Curculionidae), the bark and ambrosia beetles (excluding Platypodinae), are an example of a group where such integration is long overdue. This subfamily of beetles display a diversity of genetic, ecological, and behavioral strategies, such as the evolution

Corresponding author at: 361 Newins-Ziegler Hall, Gainesville, FL 32611-0410, United States. E-mail addresses: ajj@ufl.edu (A.J. Johnson), [email protected] (D.D. McKenna), [email protected] (B.H. Jordal), [email protected] (A.I. Cognato), [email protected] (S.M. Smith), [email protected] (A.R. Lemmon), [email protected] (E.M. Lemmon), hulcr@ufl.edu (J. Hulcr). https://doi.org/10.1016/j.ympev.2018.05.028 Received 19 September 2017; Received in revised form 11 May 2018; Accepted 21 May 2018 Available online 31 May 2018 1055-7903/ © 2018 Elsevier Inc. All rights reserved.

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

within Scolytinae (Wood, 1986). It is traditionally defined by only two putative morphological characters: the posterior end of the metanepisternum (as metepisternum in Wood, 1986) covered by the elytra, and no more than five antennal funicle segments (Wood, 1986). However, many species of Cryphalini species do not possess these characters (Johnson, unpublished) and these characters are homoplastic, occurring in other tribes of Scolytinae. Current phylogenetic evidence based on 5–18 genes (Gohli et al., 2017; Jordal and Cognato, 2012; Pistone et al., 2017) suggests that Cryphalini may be composed of at least three separate clades, and contains species belonging to other tribes, including some Diamerini and all of Xyloterini. The abovementioned previously published phylogenies share surprising results. For example, the inclusion of genera from groups outside of Cryphalini (e.g. Carphoborus, Polygraphini) with genera of Cryphalini despite morphological differences, and the deep polyphyly of several genera of Cryphalini. In summary, it remains unclear whether Cryphalini is monophyletic as currently defined, and the phylogenetic relationships among the genera of Cryphalini are unclear based on previously published phylogenies.

of fungus farming and inbreeding mating systems, which rival the diversity present in entire insect orders. Scolytinae are also ecologically and economically important. Many destructive species are the harbingers of anthropogenic global change in forest ecosystems and industries, facilitated by the changing climate, introductions of non-native species, and management practices (Hulcr et al., 2015). Their economic importance has led to the accumulation of an extensive body of scientific literature, including biological data (Wood and Bright, 1987, 1992). Bark and ambrosia beetles are also becoming prominent as genomic models: two of the first four published beetle genomes were scolytines (Keeling et al., 2013; McKenna, 2018; Vega et al., 2015). A satisfactory understanding of Scolytinae has been hindered by the absence of a robust phylogeny. The subcortical (under bark) lifestyle has imposed significant constraints on the morphology of Scolytinae, which has led to the convergent evolution of many morphological characters that were relied upon by taxonomists to serve as the basis for the current classification system (Hulcr et al., 2015; Hulcr et al., 2007a). Entire distantly related groups have been mistakenly classified as clades due to convergent morphological characters, mating systems, or their fungus farming habits (e.g. Xyloterini, Xyleborini, and Premnobius Kirkendall, 1983; Wood, 1982). Only with the advent of molecular systematic analyses, has it been demonstrated that these groups are not closely related, and that their similarities result from convergent evolution (Cognato, 2013; Farrell et al., 2001; Normark et al., 1999). To understand the evolutionary history of Scolytinae, a well-resolved phylogeny is needed. Recent attempts have recovered the placement of Scolytinae within weevils (family Curculionidae) using molecular phylogenetic data from mitochondrial genomes (Gillett et al., 2014) and multi-marker nuclear DNA sequence data sets (Jordal et al., 2011; McKenna, 2011; McKenna et al., 2009, 2015; Shin et al., 2017) but the analyses lacked broad sampling across Scolytinae. Conversely, analyses that include larger numbers of Scolytinae have been based on smaller numbers of DNA markers and do not resolve all relationships between tribes (Gohli et al., 2017; Jordal and Cognato, 2012; Jordal et al., 2011; Pistone et al., 2017). The present study had three complementary objectives. First, we developed and tested methods to infer robust phylogenetic relationships of Scolytinae using phylogenomic data. Second, we applied these methods to several of the most diverse tribes of bark and ambrosia beetles, including Cryphalini, Corthylini and Xyleborini. These tribes alone comprise more than half of the ∼6000 species of Scolytinae (Hulcr et al., 2015). They have often been considered the “advanced” bark and ambrosia beetles and have been grouped together in various arrangements by traditional taxonomists (Hopkins, 1915; Wood, 1986). Third, we used our well-resolved phylogeny to study the evolution of the diverse trophic and reproductive strategies observed across Scolytinae, placing an emphasis on Cryphalini and related groups whose biology and trophic habits are especially variable (Kirkendall, 1983). The third analysis focused on traits that had previously been proposed as key evolutionary innovations underlying the species diversity and ecological success of Scolytinae (Gohli et al., 2017). These traits include: inbreeding polygyny, fungus farming, and host taxon generalism (Gohli et al., 2017).

1.3. Inbreeding polygyny Scolytinae contains examples of several different mating systems. One remarkable aspect of this is the routine inbreeding of species, termed as inbreeding polygyny (Hamilton, 1967; Kirkendall, 1983, 1993), characterized by a collection of traits shared among several scolytine genera (or in some cases, just species). These traits include: (a) extreme sexual dimorphism, particularly males with non-functional wings and eyes; (b) a skewed sex ratio with very few males; (c) mating between siblings in the natal gallery; and (d) female dispersal after mating, and lone female initiation of the gallery (Kirkendall, 1983). This phenotype is in stark contrast to other Scolytinae, in which both males and females can fly and will disperse to their host plant material and mate on or inside the host. This suite of characters seems inextricably linked, and always occurs simultaneously in species of Scolytinae, with the exception of some Dendroctonus species which appear intermediate, and Pityophthorus puberulus which appears to reproduce through parthenogenesis (Deyrup and Kirkendall, 1983). Inbreeding polygyny is used synonymously with other terms such as “consanguineous polygyny” (Wood, 1986), “sibling mating” (Jordal and Cognato, 2012) or “permanent inbreeding” (Gohli et al., 2017). It should be emphasized that these terms are imprecise; outbreeding does occur based on observations of males outside the gallery presumably seeking mating opportunities (Cooperband et al., 2016; Johnson et al., 2016b), as well as genetic signatures of outbreeding (Holzman et al., 2009; Keller et al., 2011; Storer et al., 2017). Mating between siblings is, however, the norm, and galleries are always founded by a single female. The mechanism behind the syndrome of inbreeding polygyny is largely unknown. Where it has been studied, there is an unusual sex determination system. For example, in the clade containing the genus Coccotrypes and the tribe Xyleborini, for some species, the mechanism for sex determination is haplodiploidy, with limited evidence from karyology, flow cytometry and the ability to rear an all-male brood from a virgin female (Herfs, 1950; Kirkendall, 1993; Ueda, 1997). Some members of a different inbreeding clade, the genus Hypothenemus, are known to have a similar pattern with a different mechanism, paternal genome elimination (=pseudoarrhenotoky), in which the paternal genome is present in all somatic nuclei, but is neither inherited nor expressed.

1.2. The pygmy borers: Cryphalini We densely sampled the pygmy borers, Cryphalini sensu Wood (1986). This tribe contains 25 genera and hundreds of species with unusually diverse habits and considerable economic and ecological importance. Cryphalini is a suitable phylogenetic model for the remainder of the bark and ambrosia beetles because it includes species that represent a diversity of behavioral and ecological traits observed in other Scolytinae, and preliminary taxonomic and phylogenetic assessment suggests that these traits have evolved multiple times within the tribe. Cryphalini has been the most formidable taxonomic challenge

1.4. Fungus farming Farming fungi is a conspicuous feature of the biology of many species of Scolytinae. The interactions observed between scolytines and fungi can be antagonistic, commensal, asymmetrical, beneficial, or 230

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

fungal growth on the walls of larval galleries. The placement of a gallery in xylem or phloem was recorded, but not considered decisive for fungus farming. Host specificity data were based on an extensive literature search and explicit count of records from plant families; specialists included species feeding on fewer than three plant families, and super-generalists include species feeding on three or more plant families. This cutoff was chosen as a convenient point on the distinctly bimodal distribution of specializations; most bark and ambrosia beetles live on one host plant family, or live on many. The three summarized binary traits, as well as the observations which support these traits are presented in Table S1. Taxa were selected to capture the variation in these traits, with an emphasis placed on certain species-rich groups, namely the tribe Cryphalini.

mutualistic. Dependence on fungi varies widely among the phloem-infesting Scolytinae (Harrington, 2005), whereas most xylem-inhabiting species of Scolytinae are strictly dependent on fungi (Skelton et al., 2018). The farming of nutritional fungi in plant tissue, termed “the ambrosia symbiosis”, evolved many times among unrelated groups of Scolytinae (Jordal and Cognato, 2012). There are remarkable similarities among fungus farming species of Scolytinae from different taxonomic groups, such as the presence of mycangia, the fungus transport organs (Hulcr and Stelinski, 2017). Phloem-inhabiting fungus farming (phloeomycetophagous) Scolytinae are usually excluded from the definition of true ambrosia beetles following a forest entomology tradition (e.g. Gohli et al., 2017). However, because they farm fungi and have evolved mycangia (e.g. Pityoborus spp., Dendroctonus spp.), they are classified as fungus farmers here. Many additional beetles are known to have mycangia, but their ecology is poorly known. Cryphalini also engage with fungi in diverse ways. Species of Trischidias and some Hypothenemus colonize fungus-infested decaying wood (Deyrup, 1987), and at least two Hypothenemus species are recorded as being able to farm fungi and also possess mycangia (H. concolor by Schedl (1961), and H. curtipennis, by Beaver (1986)). The taxonomic diversity of Scolytinae that farm fungi suggests that there are multiple transitions to or from this feeding strategy. This has been corroborated via molecular phylogenetic studies (Jordal and Cognato, 2012; Pistone et al., 2017).

2.2. Genomic data sources The new sequences used for molecular phylogenetics came from three sources: DNA sequences for hundreds of nuclear genes obtained via targeted enrichment. Genomic DNA from 119 individuals (summarized in Table S1) was obtained by preparing whole DNA extracts. Each specimen was photographed, then crushed whole, and DNA was extracted using the OmniPrep genomic DNA extraction kit (G-biosciences, USA). Photographs of all specimens are available at http://www. ambrosiasymbiosis.com and deposited in Dryad (doi:10.5061/dryad. 5r6d5kc). Barcoded libraries were prepared and enriched at the Center for Anchored Phylogenomics (http://www.anchoredphylogeny.com) following methods described in Lemmon et al. (2012), Prum et al. (2015), Haddad et al. (2017) and Shin et al. (2017). Briefly, the genomic extracts were fragmented using a Covaris E220 focused-ultrasonicator (Covaris, USA), and prepared as single-indexed libraries following methods described Meyer and Kercher (2010) with an additional size selecting step. The enrichment kit contained probes targeting 941 loci representing 641 target genes. The enriched targets were then pooled and sequenced over four separate batches on a HiSeq 2500 (Florida State University, Tallahassee, Florida, USA), giving paired 150 bp reads with an approximate insert size of 400 bp. Four samples were discarded and repeated due to apparently failed library preparation or high levels of contamination. Whole RNA Transcripts. RNA was extracted from seven individual adult beetles (Table S1) with the RNAqueous RNA extraction kit (New England Biosciences Ltd.) and libraries were prepared using the NEB RNAseq library prep kit with dual indexed barcodes. Each RNA library had approximately 1/25 of a pooled Illumina NextSeq high output run (150 × 2) (ICBR, University of Florida, Florida, USA). Whole DNA sequencing without enrichment. Genomic DNA from two individuals (Table S1) was extracted using the OmniPrep DNA extraction kit (G Biosciences Ltd). Libraries were prepared using the low input NEB library prep kit (New England Bioscience Ltd.), with a target of 4/ 25 of a pooled run on an Illumina NextSeq high output run (150 × 2) (ICBR, University of Florida). Existing published genomes and transcriptomes were downloaded from sources summarized in Table S2. The genome for Dendroctonus ponderosae (Keeling et al., 2013) was modified to break up scaffolds by removing long stretches of ambiguous bases connecting the scaffolds, which greatly improved speed of further steps for searching and annotation. The transcriptome of Tomicus yunnanensis (Zhu et al., 2012) was initially included but subsequently removed due to unexpected level of sequence heterogeneity, likely a consequence of pooled sequencing of genetically divergent individuals.

1.5. Host specificity Host specificity of most insect herbivores follows a continuum from reliance on a single plant host to indiscriminate feeding (Novotny et al., 2002). In Scolytinae, the distribution of feeding habits is less continuous (Hulcr et al., 2007b; Novotny et al., 2010). Species ingesting plant tissues, such as most bark inhabiting species, are typically monophages or oligophages. Scolytine species farming and ingesting fungi, the ambrosia beetles, will often be found on a much broader range of hosts, regardless of the taxonomic identity of the host plant (Beaver, 1979; Hulcr et al., 2007b). Many scolytine species deviate from this pattern, however, because even though they ingest plant tissue, they feed on dozens of plant families. For example, the species Hypothenemus birmanus is recorded from 29 different plant families (Atkinson, 2017). Such bark beetles were excluded from some previous analyses based on “difficult species delimitation” (Hulcr et al., 2007b). Whether the “super-generalists” (Wood, 2007) are an evolutionary intermediary to fungus farming, or an unrelated condition, is unknown, and requires assessment in a phylogenetic framework. 2. Materials and methods 2.1. Trait inference Ecological and genetic traits were inferred by a combination of literature searches (Atkinson, 2017; Beaver et al., 1989; Kirkendall, 1983; Wood, 1982, 2007; Wood and Bright, 1987, 1992), and direct observations in the field by the authors. The studied traits were reduced to three binary characters: Inbreeding polygyny versus regular outbreeding, fungus farming versus plant tissue feeding, and host plant generalists versus host plant specialists. Inbreeding polygyny was inferred based on three characters: (1) morphology of the male (smaller and flightless), (2) lone foundresses indicating pre-dispersal mating, and (3) a skewed sex ratio. This information was complete for most species, and all of those that were complete followed the expected dichotomy. Therefore, species with some incomplete information were still scored, even if, for example, a male was absent. The taxa that are known anomalies, Dendroctonus micans and related species, were not available for study. Fungus farming was inferred based on the presence of mycangia and obvious

2.3. Bioinformatics All analyses were implemented on the High-Performance computer 231

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

overlapping remaining sequences). Manual validation of methods was done by assessing gene alignments and gene trees estimated with RAxML v8.2.3 (Stamatakis, 2014). Alignments were reduced by deleting regions represented by less than 60% of taxa, minimizing the number of missing characters, and thus the computational load when inferring phylogenies. All genes were concatenated, and the third codon positions were discarded, following empirical evidence by Breinholt and Kawahara (2013). Samples and their genes remaining in the final data matrix are presented in Table S4.

system at the University of Florida (HiPerGator 2.0). The methods described below are graphically illustrated in Figure S1, and were written as bash and Python 3 scripts interacting with other software (detailed below) and are available with the in the Dryad repository for this project (doi:10.5061/dryad.5r6d5kc). 2.3.1. Assembly and reference preparation For each library, all reads were evaluated to remove low quality regions of sequences, short sequences and stray sequencing adapters, using Trimmomatic (Bolger et al., 2014). Numbers quality filtered reads are presented in Table S3. RNAseq samples and targeted enrichment samples were separately assembled using Bridger (Chang et al., 2015) with a minimum kmer depth (and subsequently coverage) of 10. Two low-coverage genomes were assembled with Soapdenovo2 (Luo et al., 2012), with parameter guidance by kmergenie (Chikhi and Medvedev, 2013), set to kmer length of 97 and 117, and a minimum coverage of 2 reads. The assembly software used, settings and sizes of assembled contigs are indicated in Table S3. The transcript assemblies were treated as intermediate references to test orthology of the targeted genes and as a template for annotation of the enriched samples. Orthology and homology of the 641 target genes were evaluated using an all-against-all BLAST search within each of the reference samples, and against Tribolium. Genes for which potential paralogs were recovered (defined by similarity of two peptide sequences, observed in Tribolium or the samples of Scolytinae), were identified and eliminated using Usearch (Edgar, 2010) and custom Python scripts which sorted and summarized the similarity of peptide sequences. Of the surviving target genes, reference sequences from transcriptomic and whole genome samples were prepared using Usearch, Exonerate (Slater and Birney, 2005) and custom Python scripts, producing complete coding DNA sequences (CDSs) and peptide sequences of 397 nuclear genes for at least some of the reference samples.

2.3.4. Phylogenetic inference Phylogenetic trees were inferred using Bayesian inference (Exabayes v1.5; Aberer et al., 2014), partitioned by genome source (i.e., mitochondrial versus nuclear), with two independent runs of four chains each, run for a minimum of 2.5 million generations, which was sufficient time for convergence (average standard deviation of split frequencies < 0.005), sufficient effective sample sizes (> 500, checked using Tracer; Rambaut et al., 2015), and sufficient topological convergence (using RWTY; Warren et al., 2017). A 50% majority rules tree was estimated (following 25% burn-in). This was repeated using a maximum likelihood strategy with RAxML (function “d”, model “GTRGAMMA”, 100 searches), which yielded a similar topology for the best scoring tree except for the placement of the root. One thousand bootstrapped alignments were generated, and their phylogenies inferred with RAxML, using the same methodology, except for executing fewer independent tree searches per bootstrap (n = 10). Trees were rooted with Scolytini as an a priori outgroup. This morphologically distinct tribe is frequently found in molecular phylogenies to be distant from all other Scolytinae, and has even been suggested as a separate weevil lineage (Gillett et al., 2014; Pistone et al., 2017; Shin et al., 2017). The use of an additional nearby outgroup from Curculionidae was avoided due to uncertain relationships between Scolytinae and other Curculionidae.

2.3.2. Extraction and annotation of target genes Contigs of putative homologs in all targeted enrichment assemblies and shallow genomes were gathered and annotated based on the peptide sequence of the closest matching references using Usearch to identify potential matches, Exonerate to accurately annotate the coding regions, and custom Bash and Python scripts to choose appropriate references, check the quality of the alignment and change file types. Non-coding regions of the assemblies such as introns, were excluding from further analyses, but are available as part of the annotated sequences in the supplementary files. All annotated sequences for each gene were assessed for contamination. For each gene, we gathered the number of other samples with identical copies (≥98% identical nucleotide sequence), nonidentical copies in the same sample, and number of copies of the other samples. After summarizing this information for all genes, we predicted and eliminated sequences that were (a) likely contaminants, and (b) sequences that could not be ruled out as contaminants in because no matching assembly was present in a known donor sample. This filtering was implemented using Usearch to find matches and give similarity statistics, and custom Python scripts to summarize information and filter annotated assemblies. The rigorous decontamination step lead to some samples to be subsequently unusable in further steps. The numbers of assemblies and genes surviving decontamination for each sample are presented in Table S3.

2.4. Testing taxonomic hypotheses

2.3.3. Sequence alignment Surviving sequences were aligned by their peptide sequence, connected if appropriate, and filtered to remove highly divergent, likely poorly aligned individual sequences, using MAFFT (Katoh and Standley, 2013), EMBOSS distmat (Rice et al., 2000), and custom Python scripts (using methods similar to TranslatorX (Abascal et al., 2010) to translate, align, then merge or filter out

2.5. Ancestral state reconstruction and character state correlation analysis

Following phylogeny construction, hypotheses of species relationships that were published previously and conflicted with the relationships presented here were tested using three metrics: (1) the proportion of sampled posterior output trees in which the hypothesis is supported in the sampled Bayesian trees, (2) proportion of the same in the bootstrapped alignment trees, and (3) a comparison of the likelihoods of the tree presented and constrained ML tree built in RAxML, using the Shimodaira-Hasegawa test (implemented in RAxML) to test if the conflicting hypotheses are significantly worse than the tree presented. The sources and overview of the topological hypotheses are presented in Table 1. Constrained trees did not include genera not included/mentioned by the source. Miocryphalus pennatus and Cosmoderes imitatrix were not included in constraints because their current placement in the respective genera is erroneous. The former is known to be genetically and morphologically very similar to Cosmoderes madagascarensis Schedl and bears no resemblance to any Miocryphalus, and Cosmoderes imitatrix bears no similarity to other Cosmoderes. Dubious generic placements by Schedl, as in these two cases, are a known phenomenon in scolytine taxonomy (Hulcr and Cognato, 2013). This is a conservative approach, since inclusion as their current generic assignment suggests, yielded significantly worse constrained trees for all taxonomy-based hypotheses.

The evolution of the three key ecological traits (breeding system, fungus farming and host breadth) was investigated for taxa where all three states were known. The correlation between the three traits was investigated using Pagel’s test for correlation between the pairs of traits using Mesquite 232

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

(caption on next page)

rates were not used based on the advice of Pagel (1994), who suggested inevitably higher confidence based on a posteriori observations). Pagel’s test can lead to false positive results (Maddison and FitzJohn, 2014) but these are uncommon in trees with multiple origins of each character

V3.2 (Maddison and Maddison, 2017). This analysis tests the hypothesis that two traits, for example fungus farming and inbreeding polygyny, evolved more often together than independently in the history of bark beetles. An asymmetrical model of rate evolution was used (irreversible 233

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

Fig. 1. Phylogenetic tree estimated via Bayesian inference using 114,276 bp of DNA sequence data from 251 protein coding genes. Numbers above nodes indicate bootstrap support in maximum likelihood analyses, while numbers below represent posterior probabilities. Full circles indicate presence (colored) or absence (white) of evolutionary innovations: inbreeding polygyny, fungus farming and host generalism. Color scheme follows Gohli et al (2017). Pie charts on some tips (circled in bold) and all internal nodes represent predicted characters based on the ancestral state reconstruction. Tribes are indicated following the color scheme of Pistone et al. (2017). Colored boxes indicate examples of evolutionary innovations (photos by AJJ unless indicated). Inbreeding polygyny: (A) Coccotrypes distinctus (Dryocoetini); (B) Premnobius cavipennis (Ipini); (C) Hypothenemus eruditus (Cryphalini); (D) Sueus niisimai (Hyorrhynchini). Fungus farming: (A) Xyloterinus politus (Xyloterini) (Photo by You Li); (B) Corthylus sp. (Corthylini) (Photo by JH); (C) Cnestus mutilatus (Xyleborini) (Photo by C. Bateman); (D) Ambrosiodmus rubricollis (Xyleborini) (Photo by You Li); (E) Premnobius cavipennis (Ipini) (Photo by You Li); (F) Sueus niisimai (Hyorrhynchini) (Photo by You Li). Host generalist; Cryptocarenus heveae (Cryphalini), showing the diversity of plant species from which it has been collected. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

in the bootstrapped alignments, and constrained trees were not significantly worse. Therefore, the specific hypothesis of Xyloterini (and Coriacephilus) sister to Cryphalus, Hypocryphalus and Margadillius cannot be rejected. The major Cryphalini + Xyloterini clade is sister to a clade containing Strombophorus, Hapalogenius and Xyloctonini, corroborating the results of Jordal and Cognato (2012). The large tribe Corthylini is monophyletic only with inclusion of the genus Cryptocarenus, which is currently classified within the tribe Cryphalini. Within Corthylini, the fungus farming Corthylina are also monophyletic, while the phloemfeeding Pityophthorina are paraphyletic. Dryocoetini, specifically Coccotrypes, are paraphyletic with respect to the mega-diverse clade Xyleborini. Among the strictly fungus farming Xyleborini, there is a strongly supported group containing the monophyletic Xylosandrus and several other genera with a mesothoracic mycangium. Our sampling of other tribes was not as extensive, but still provides a basis for re-evaluation of many current taxonomic groups. Most tribes are supported, with the exception of the following: Phloeotribus (Phloeotribini) and Chramesus (Phloeosinini) are sister taxa, and the clade they form is sister to the clade that includes Scolytodes, corroborating the topology of Jordal and Cognato (2012) and Pistone et al. (2017). Hyorrhynchini, represented here by Sueus, are sister to the genus Sphaerotrypes (currently in Diamerini); both share completely divided eyes. Carphoborus and Polygraphus were recovered as sister taxa, which supports traditional morphological classification but was not recovered in previous molecular phylogenies. Both of these clades show extensive divergence and low support.

state, such as the one estimated here. Character states on ancestral nodes and on taxa with missing values were calculated using the re-rooting method of Yang et al. (1995), and implemented in the R package Phytools (Revell, 2011). This method allows ambiguous character states of taxa to be predicted (root and ambiguous tips were initially coded with a prior of 0.5).

3. Results 3.1. Bark beetle relationships We present the first robust and well-resolved phylogeny of bark and ambrosia beetles with an emphasis on Cryphalini (Fig. 1). The phylogeny suggests that many groupings in the current taxonomic classification of Scolytinae do not reflect true evolutionary relationships. Among the clades with the highest diversity, only Xyleborini are monophyletic; Corthylini are paraphyletic and Cryphalini are highly polyphyletic. The largest Cryphalini clade, including the type genus Cryphalus, has Xyloterini nested within it. Furthermore, there are two genera of Cryphalini in a clade containing the Corthylini, and six genera of Cryphalini formed a distantly related clade, provisionally termed the Ernoporus clade. Trees constrained to retain the monophyly of Cryphalini were significantly worse than the relationships presented (Table 1). Among the major clades of Cryphalini, the genera are also highly polyphyletic. Hypocryphalus, Margadillius and Cryphalus species are intermixed. Among the Ernoporus clade, species of Ernoporicus, Scolytogenes and Ptilopodius are highly intermixed. The fungus farming Xyloterini have evolved from within the Cryphalini clade, sister to the genus Coriacephilus. This clade was found to be sister to the genera Hypothenemus, Trypophloeus and Cosmoderes, contradicting the sister relationship to Cryphalus found in previous molecular studies. The alternative topology was also recovered during the Bayesian search and

3.2. Evolution of inbreeding polygyny The results of our phylogenetic analyses suggest there are at least five origins of inbreeding polygyny among the included taxa. It is

Table 1 Testing of conflicting phylogenetic hypotheses. Taxonomic hypothesis (references describing/using the hypothesis in perentheses)

% Occurrence in Bayesian inference

% Occurrence in bootstrapped alignments

Constrained tree statistically plausible? (S-H test)

(a) Monophyletic tribe Cryphalini sensu Wood 1986 (Alonso-Zarazaga and Lyal, 2009; Wood, 1986; Wood and Bright, 1992)

0.0

0.0

No (p < 0.01)

(b) Monophyletic Hypocryphalus and monophyletic Cryphalus (Wood, 1986; Wood and Bright, 1992)

0.0

0.0

No (p < 0.01)

(c) Trypophloeus sister to Carphoborus (Gohli et al., 2017).

0.0

0.0

No (p < 0.01)

(d) Xyloterini sister to Cryphalus and Hypocryphlaus (Gohli et al., 2017; Hulcr et al., 2015; Jordal and Cognato, 2012).

9.0

34.0

Yes

(e) Hypothenemus and Cosmoderes sister to Corthylini + Cryptocarenus (Gohli et al., 2017).

0.0

0.0

No (p < 0.01)

(f) Ernoporus clade (Ernoporus, Ernoporicus, Scolytogenes and Ptilopodius) sister to Micracidini (Gohli et al., 2017).

0.0

0.0

No (p < 0.01)

(g) Corthylina paraphyletic with respect to Gnathotrichus (Gohli et al., 2017).

0.0

1.0

No (p < 0.05)

234

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

unlikely that any of these patterns are a consequence of secondary loss of the habit (Fig. 1; all ancestral nodes with < 0.95 probability); in all cases the most recent common ancestor is predicted to reproduce by regular outbreeding. Cryptocarenus represents an independent origin of the inbreeding polygyny. The paraphyletic relationship with Araptus is notable, since some species of Araptus are recorded as being inbreeding (Kirkendall, 1983; Wood, 2007).

Table 2 Correlation analyses.

3.3. Evolution of fungus farming

Correlated traits

Pagel’s test 1000 simulations (p)

Pairwise comparisons (p)

Mating system and Fungus farming Mating system and Host breadth Fungus farming and host breadth

0.031

0.0625

0.000

0.015625

0.033

0.0625

Tests where p < 0.5 are highlighted.

We found 12 clades of fungus farming Scolytinae in the sampled taxa. Our analysis included 13 of the 22 putative independent associations between beetle clades and major fungal clades (Hulcr and Stelinski, 2017), suggesting that the majority of beetle-fungus associations arise de novo. Where multiple independent associations exist within a fungus farming clade, only one was paraphyletic (Gnathotrichus + Monarthrum), suggesting that a beetle clade, which was already ambrosial, adopted a new fungus. Among the Xyleborini, where subclades are associated with different fungal groups, we found no evidence of any of the associations being ancestral to others (i.e. no evidence of switching fungus symbiont). The phloem-dwelling fungus farming Scolytinae (phloeomycetophages) are widely dispersed across the tree and are not related to xylomycetophagous taxa. We confirm a strong pattern of asymmetrical evolutionary association between clades of ambrosia beetles and their fungal mutualists. Most clades of ambrosia beetles are associated primarily with a single genus of fungus, while fungal genera can associate with multiple beetle clades. The monophyly of Corthylina (including Gnathotrichus) suggests that the apparent independent evolution of fungus farming in Gnathotrichus indicated by the tree published by Gohli et al. (2017) is not supported, which is consistent with the results of Pistone et al. (2017). Also, contradictory to Gohli et al. (2017), Sueus and Dactylipalpus were not monophyletic, suggesting that the two evolved fungus farming independently. We did not include the fungus farming groups Scolytoplatypodini, Scolytodes unipuctatus, Bothrosternus + Eupagiocerus, and Pachycotes, which represent a further three likely origins of fungus farming, bringing the total known number of independent origins to 16.

many inbreeding scolytines exist with narrow host ranges. 4. Discussion 4.1. Strengths and weaknesses of target enrichment Analyses of additional loci produced a scolytine phylogeny that strongly contradicts the current classification (which is based on morphology), consistent with other recent studies that analyzed smaller molecular data sets and (typically) fewer taxa. The tree presented here is more strongly supported than trees resulting from earlier molecular studies, undoubtedly as a consequence of the increased number of loci analyzed. A similar pattern has been observed in analyses of large target enrichment data sets involving other groups of organisms (Haddad et al., 2017; Prum et al., 2015). Although the library preparation and sequencing pipeline required to generate large numbers of loci has the potential for generating contaminated sequences (i.e. through the automation of library preparation or the pooling of indexed samples for seqeuncing), our analytical pipeline ensured that any contaminants that may have been derived from this process were filtered out. The downside to this stringent filtering is the exclusion of some of the taxa that were initially intended for this study. 4.2. Phylogeny of Scolytinae The phylogeny recovered herein highlights the lack of congruence between the current morphology-based classification of Scolytinae and their actual evolutionary relationships. This is especially true of Cryphalini, the most densely sampled tribe. This should be no surprise, since the diagnostic characters of Cryphalini are poorly defined, and have not been systematically reviewed. The few described synapomorphies (Hopkins, 1915; Wood, 1986) are unreliable, and exceptions within and outside of Cryphalini are common. The genus-level classification within Cryphalini is also in need of a complete revision. The genera of Cryphalini were originally described and diagnosed using (in most cases) characters of the antennae. For example, Hypocryphalus and Cryphalus are defined as differing by a single funicle segment, which is generally an unreliable character, and our topology confirms that these genera are not monophyletic. It is clear that the current definitions of genera are not necessarily indicative of actual relationships in this group, and that taxonomic changes are needed. Cryptocarenus, currently placed in the Cryphalini, is robustly nested within Corthylini, specifically within the diverse and understudied genus Araptus. This is consistent with the results of (Gohli et al., 2017). Pistone et al (2017) recovered Cryptocarenus sister to Dacnophthorus. While Dacnophthorus was not included in our dataset, it is likely that it too is nested among Araptus. The placement of Cryptocarenus in Corthylini is supported by the presence of an oblique locking groove in the absence of an interlocking spine, overlooked by S. L. Wood who described it as a unique character for the tribe (Wood, 1978). The genera Stegomerus and Dendroterus lack this character, suggesting that it had

3.4. Evolution of plant host generalism The phylogeny includes eleven clades of the species defined as host plant generalists. The finding that Cryptocarenus is unrelated to Hypothenemus identifies another independent example of the generalist habit. Ancestral state reconstruction suggests the generalist habit is derived, and the specialist habit is ancestral. One example of a reversal to monophagy is observed in Hypothenemus hampei, a coffee seed feeder. Additional examples of specialist species nested among generalist genera are known within Corthylini and Xyleborini (based on records from Atkinson, 2017), but exemplars were not included in our dataset. 3.5. Correlation of traits The various evolutionary innovations identified here show a correlation in their occurrences in the tree. All traits, especially mating system and host specificity, appear significantly correlated when analyzed using Pagel’s test for correlation of binary traits (Table 2). For example, the majority of clades that have broad host ranges are also inbreeding. Only mating system and host specificity are correlated when analyzed using pairwise comparisons. Equally interesting are cases where no correlation appears. Fungus farming does not correlate with broad host range; on the contrary, many generalist scolytine clades are able to use numerous substrates unaided by fungi. Similarly, even though inbreeding polygyny and plant host generalism are correlated, 235

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

of Xyleborini is predicted to be fungus farming, but we do not understand if, how, or when any transitions between fungal symbiont and mycangia location occurred. It remains a viable hypothesis that the different associations are independently evolved in a series of Coccotrypes-like ancestors. Only 12 of the 36 Xyleborini genera have any known information about their symbionts. A broader understanding of the symbiosis and the phylogeny of Xyleborini will be needed to resolve the evolutionary history of this group.

evolved within the clade, and is a synapomorphy for a subclade. The subtribe Corthylina is nested among Pityophthorina in our trees. The paraphyly of Dryocoetini with respect to Xyleborini has been demonstrated before (Jordal and Cognato, 2012; Normark et al., 1999), but the relationship to Coccotrypes was poorly resolved. Normark et al (1999) recovered Xyleborini as paraphyletic with respect to Coccotrypes, but with very low nodal support. Our results confirm the suggestion by Kirkendall (1993) that Xyleborini is derived from a Coccotrypes ancestor, as the genus is currently characterized. This produces a challenge reconciling traditional hierarchical classification with monophyly, since an entire tribe is nested within a single genus. We can speculate that the evolution of fungus farming led to a rapid radiation of Xyleborini. Within Xyleborini, the genus Xylosandrus is monophyletic. This is contrary to previous studies focusing on the relationships of Xylosandrus (Dole et al., 2010). While our taxon sampling was much more limited, all major lineages of Xylosandrus and similar genera discovered by Dole et al. (2010) were included in our study.

4.5. Evolution of plant host generalism Most origins of scolytine polyphagy coincide with some degree of reliance on fungi. Of the eleven host plant generalist clades, seven contain fungus farming species, and six of those are exclusively mycetophagous. Several taxa included appear to be host plant generalist without feeding on fungi (e.g. Cnesinus strigicollis, Cryptocarenus heveae). Interestingly, while the generalist species included in our phylogeny do not farm fungi, other species in the same genera do: (a) two Hypothenemus species have been recorded farming fungi (H. concolor and H. curtipennis), (b) fungus-lined galleries of Cryptocarenus sp. have been observed (Johnson, unpublished), and (c) a fungus farming Cnesinus is known (Kolařík and Kirkendall, 2010; Jordal unpublished; Wood, 1982). The habit of host generalism precedes fungus farming in these cases. Therefore, while mycophagy has been proposed as the cause of host plant generalism (Beaver et al., 1989; Gohli et al., 2017), future work should also test an alternative hypothesis: evolution of bark beetles towards host plant generalism facilitates the capacity to feed on fungi, and a subsequent switch to fungi as the sole nutrition. Additionally, many of the generalist species are primarily pith feeding, or feed on older material which is expected to contain fewer plant allelochemicals. Older material is also likely to have already been colonized by fungi, perhaps facilitating an intermediate step to a dietary dependence on fungi. Unlike inbreeding and fungus farming, the evolution towards polyphagy does include examples of reversal back to specialization. Within the Hypothenemus clade, where most species colonize dozens of plant families, H. hampei has specialized in Coffea spp. seeds. Several species within the fungus farming generalist tribe Xyleborini also specialize on tree genera or families, such as Xyleborus pubescens and X. glabratus (not included in the phylogeny). The genus Coccotrypes contains a mixture of species, including some known from a single host plant species, and others collected from dozens of plant families. The number of taxa in each of these clades is not sufficient to explore finer scale specialist-generalist patterns of host association.

4.3. Evolution of inbreeding polygyny Inbreeding polygyny is rare in Coleoptera, yet in Scolytinae it has evolved at least six times. The addition of Cryptocarenus, which does not share an inbreeding most common recent ancestor, is confirmed here. We conclude that the presence of inbreeding polygyny among extant scolytine groups is correlated with the presence of fungus farming, in agreement with Gohli et al. (2017). Five of the six inbreeding lineages contain fungus farming taxa, and in two cases all the contained species are inbreeding as well. However, while the occurrences of two features are correlated, their origins appear asynchronous. For some tribes, the inbreeding apparently preceded fungus farming (e.g. Xyleborini among Dryocoetini + Xyleborini, Hypothenemus dolosus among Hypothenemus + Trischidias). In others, fungus farming preceded inbreeding polygyny (e.g. Xyloterinus among Xyloterini). Some genera of Scolytinae, not included in the analyses, may represent additional origins of inbreeding polygyny. Bothrosternus has been described as inbreeding (Wood, 1982), although this observation has not been verified. Two Araptus species are known to be inbreeding (Kirkendall, 1983), which is likely the same clade as the Araptus included here and Cryptocarenus. Periocryphalus is likely to be sister to or nested within Hypothenemus (Johnson et al., 2016a). Margadillius loranthus, Ptilopodius shoreae, and P. squamosus are noted as inbreeding (Browne, 1961), but these have been determined as misdiagnosed species of Hypothenemus based on morphological analyses (Johnson, unpublished).

5. Conclusions 4.4. Evolution of fungus farming Our phylogenetic analyses produced a uniquely robust phylogeny and reconstructed evolutionary history for the species-rich and ecologically significant weevil subfamily Scolytinae. While the estimated numbers of origins of fungus farming and inbreeding polygyny have not changed substantially since Jordal and Cognato (2012), Gohli et al. (2017) and Pistone et al. (2017), we now have increased confidence in the estimated number and direction of major events associated with the evolution of scolytine trophic habits and life histories. Most importantly, we provide a new path towards a robust reclassification of some of the most diverse and important groups of bark beetles, particularly the previously intractable tribe Cryphalini. The most significant limitation to the study of the evolution of bark beetles is the lack of accurate, published information on their behavior and ecology. The apparent complexity of the evolutionary history of Scolytinae leaves many unanswered questions, even in ecologically and economically important groups. Several of the independently evolved innovations are represented here only by undescribed species with no published notes on their biology and ecology. While we applaud the recent advances in evolutionary inference, we urge researchers to continue studies on the

Twelve clades of fungus farming beetles are predicted from the dataset. All but one of the described associations of beetles and specific fungi (sensu Hulcr and Stelinski, 2017) are monophyletic. This suggests that all but one origin of the capacity to farm fungi in Scolytinae might be independent. Most lineages have also evolved a new type of mycangium, the fungus carrying structure, and many have evolved specificity to only one or a few fungal groups, often also evolved de novo. The group containing Anisandrus, Cnestus, Diuncus, Eccoptopterus and Xylosandrus was found to be monophyletic, contradicting some recent phylogenies focusing on the group (Cognato et al., 2011; Dole et al., 2010) and supporting others (Gohli et al., 2017; Jordal, 2002). This conforms to the morphology and symbiotic association of beetles in this clade, all of which possess a large mesonotal mycangium and are highly specific to their symbiotic fungus, Ambrosiella sp. (Mayers et al., 2015). Among Xyleborini clades with different fungal farming systems, there is no specific evidence of one symbiosis being ancestral. This is corroborated by (Gohli et al., 2017). The most recent common ancestor 236

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

group’s natural history.

Farrell, B.D., Sequeira, A.S., O'Meara, B.C., Normark, B.B., Chung, J.H., Jordal, B.H., 2001. The evolution of agriculture in beetles (Curculionidae: Scolytinae and Platypodinae). Evolution 55, 2011–2027. Gillett, C.P.D.T., Crampton-Platt, A., Timmermans, M.J.T.N., Jordal, B.H., Emerson, B.C., Vogler, A.P., 2014. Bulk De Novo Mitogenome Assembly from Pooled Total DNA Elucidates the Phylogeny of Weevils (Coleoptera: Curculionoidea). Mol. Biol. Evol. 31, 2223–2237. Gohli, J., Kirkendall, L.R., Smith, S.M., Cognato, A.I., Hulcr, J., Jordal, B.H., 2017. Biological factors contributing to bark and ambrosia beetle species diversification. Evolution 71, 1258–1272. Haddad, S., Shin, S., Lemmon, A.R., Lemmon, E.M., Svacha, P., Farrell, B., Ślipiński, A., Windsor, D., McKenna, D.D., 2017. Anchored hybrid enrichment provides new insights into the phylogeny and evolution of longhorned beetles (Cerambycidae). Syst. Entomol. 43, 68–89. Hamilton, W.D., 1967. Extraordinary sex ratios. Science 156, 477–488. Harrington, T.C., 2005. Ecology and evolution of mycophagous bark beetles and their fungal partners. Insect-Fungal Assoc.: Ecol. Evol. 1, 22. Herfs, A., 1950. Studien an dem Steinnussborkenkafer Coccotrypes tanganus Eggers, 2 Die Soziologie. Hofchen-Briefe fur Wissenschaft und Praxis 3, 3–31. Holzman, J.P., Bohonak, A.J., Kirkendall, L.R., Gottlieb, D., Harari, A.R., Kelley, S.T., 2009. Inbreeding variability and population structure in the invasive haplodiploid palm-seed borer (Coccotrypes dactyliperda). J. Evol. Biol. 22, 1076–1087. Hopkins, A.D., 1915. Classification of the Cryphalinae with Descriptions of New Genera and Species. US Government Printing Office. Hulcr, J., Atkinson, T.H., Cognato, A.I., Jordal, B.H., McKenna, D.D., 2015. Morphology, Taxonomy, and Phylogenetics of Bark Beetles. In: Vega, F.E., Hofstetter, R.W. (Eds.), Bark Beetles: Biology and Ecology of Native and Invasive Species. Elsevier, pp. 41–84. Hulcr, J., Cognato, A.I., 2013. Xyleborini of New Guinea, a Taxonomic Monograph (Coleoptera: Curculionidae: Scolytinae). Entomological Society of America. Hulcr, J., Dole, S.A., Beaver, R.A., Cognato, A.I., 2007a. Cladistic review of generic taxonomic characters in Xyleborina (Coleoptera: Curculionidae: Scolytinae). Syst. Entomol. 32, 568–584. Hulcr, J., Mogia, M., Isua, B., Novotny, V., 2007b. Host specificity of ambrosia and bark beetles (Col., Curculionidae: Scolytinae and Platypodinae) in a New Guinea rainforest. Ecol. Entomol. 32, 762–772. Hulcr, J., Stelinski, L.L., 2017. The ambrosia symbiosis: from evolutionary ecology to practical management. Annu. Rev. Entomol. 62, 285–303. Johnson, A.J., Atkinson, T.H., Hulcr, J., 2016a. Two remarkable new species of Hypothenemus Westwood (Curculionidae: Scolytinae) from Southeastern USA. Zootaxa 4200, 417–425. Johnson, A.J., Kendra, P.E., Skelton, J., Hulcr, J., 2016b. Species diversity, phenology, and temporal flight patterns of Hypothenemus pygmy borers (Coleoptera: Curculionidae: Scolytinae) in South Florida. Environ. Entomol., nvw039. Jordal, B.H., 2002. Elongation Factor 1 α resolves the monophyly of the haplodiploid ambrosia beetles Xyleborini (Coleoptera: Curculionidae). Insect Molec. Biol. 11, 453–465. Jordal, B.H., Cognato, A.I., 2012. Molecular phylogeny of bark and ambrosia beetles reveals multiple origins of fungus farming during periods of global warming. BMC Evol. Biol. 12, 133. Jordal, B.H., Sequeira, A.S., Cognato, A.I., 2011. The age and phylogeny of wood boring weevils and the origin of subsociality. Molec. Phylogenet. Evol. 59, 708–724. Katoh, K., Standley, D.M., 2013. MAFFT multiple sequence alignment software version 7: Improvements in performance and usability. Mol. Biol. Evol. 30, 772–780. Keeling, C.I., Yuen, M.M.S., Liao, N.Y., Docking, T.R., Chan, S.K., Taylor, G.A., Palmquist, D.L., Jackman, S.D., Nguyen, A., Li, M., 2013. Draft genome of the mountain pine beetle, Dendroctonus ponderosae Hopkins, a major forest pest. Gen. Biol. 14, R27. Keller, L., Peer, K., Bernasconi, C., Taborsky, M., Shuker, D.M., 2011. Inbreeding and selection on sex ratio in the bark beetle Xylosandrus germanus. BMC Evolution. Biol. 11, 359. Kirkendall, L.R., 1983. The evolution of mating systems in bark and ambrosia beetles (Coleoptera: Scolytidae and Platypodidae). Zool. J. Lin. Soc. 77, 293–352. Kirkendall, L.R., 1993. Ecology and evolution of biased sex ratios in bark and ambrosia beetles (Scolytidae). In: Wrensch, D.L., Ebbert, M.A. (Eds.), Evolution and Diversity of Sex Ratio: Insects and Mites. Chapman and Hall, New York, pp. 235–345. Kolařík, M., Kirkendall, L.R., 2010. Evidence for a new lineage of primary ambrosia fungi in Geosmithia Pitt (Ascomycota: Hypocreales). Fungal Biol. 114, 676–689. Lemmon, A.R., Emme, S.A., Lemmon, E.M., 2012. Anchored Hybrid Enrichment for Massively High-Throughput Phylogenomics. Syst. Biol. 61, 727–744. Luo, R., Liu, B., Xie, Y., Li, Z., Huang, W., Yuan, J., He, G., Chen, Y., Pan, Q., Liu, Y., Tang, J., Wu, G., Zhang, H., Shi, Y., Liu, Y., Yu, C., Wang, B., Lu, Y., Han, C., Cheung, D.W., Yiu, S.-M., Peng, S., Xiaoqian, Z., Liu, G., Liao, X., Li, Y., Yang, H., Wang, J., Lam, T.W., Wang, J., 2012. SOAPdenovo2: an empirically improved memory-efficient shortread de novo assembler. GigaScience 1. Maddison, W.P., FitzJohn, R.G., 2014. The unsolved challenge to phylogenetic correlation tests for categorical characters. Syst. Biol. 64, 127–136. Maddison, W.P., Maddison, D.R., 2017. Mesquite: A Modular System for Evolutionary Analysis. Version 3.2. Mayers, C.G., McNew, D.L., Harrington, T.C., Roeper, R.A., Fraedrich, S.W., Biedermann, P.H.W., Castrillo, L.A., Reed, S.E., 2015. Three genera in the Ceratocystidaceae are the respective symbionts of three independent lineages of ambrosia beetles with large, complex mycangia. Fungal Biol. 119, 1075–1092. McKenna, D.D., 2011. Temporal lags and overlap in the diversification of weevils and flowering plants: recent advances and prospects for additional resolution. Am. Entomol. 57, 54–55. McKenna, D.D., 2018. Beetle genomics in the 21st century: prospects, progress and priorities. Curr. Opin. Insect Sci. 25, 76–82.

Acknowledgements We thank Michail Mandelshtam, Matthew Kasson, and numerous other collectors worldwide for generously sharing specimens for this study. We thank the Natural History Museum in Vienna, the Natural History Museum, London, the United States National Museum of Natural History, and the Zoological Museum of Moscow State University for providing access to type material to facilitate accurate identification. We also thank Michelle Kortyna, Sean Holland, and Kirby Birch at the Center for Anchored Phylogenetics for their assistance with AHE data collection and analysis. We also thank three reviewers for constructive comments and suggestions for improving the manuscript. This work was supported by National Science Foundation DEB 1256663, DEB 1256968, DEB 1556283, and DEB 1355169, USDAAPHIS agreement 15-8130-0547-CA, USDA Forest Service agreement 12-CA-11420004-042, and the Research Council of Norway award 170565/V40. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.ympev.2018.05.028. References Abascal, F., Zardoya, R., Telford, M.J., 2010. TranslatorX: multiple alignment of nucleotide sequences guided by amino acid translations. Nucl. Acids Res. 38, 7–13. Aberer, A.J., Kobert, K., Stamatakis, A., 2014. ExaBayes: massively parallel bayesian tree inference for the whole-genome era. Molec. Biol. Evol. 31, 2553–2556. Alonso-Zarazaga, M.A., Lyal, C.H., 2009. A catalogue of family and genus group names in Scolytinae and Platypodinae with nomenclatural remarks (Coleoptera: Curculionidae). Zootaxa 2258, 112–127. Atkinson, T.H., 2017. Bark and Ambrosia Beetles < http://www.barkbeetles.info/ > . Beaver, R.A., 1979. Host specificity of temperate and tropical animals. Nature 281, 139–141. Beaver, R.A., 1986. The taxonomy, mycangia and biology of Hypothenemus curtipennis (Schedl), the first known cryphaline ambrosia beetle (Coleoptera: Scolytidae). Insect System. Evol. 17, 131–135. Beaver, R.A., Wilding, N., Collins, N., Hammond, P., Webber, J., 1989. Insect-fungus relationships in the bark and ambrosia beetles. Insect-Fungus Interact. 121–143. Bolger, A.M., Lohse, M., Usadel, B., 2014. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120. Breinholt, J.W., Kawahara, A.Y., 2013. Phylotranscriptomics: saturated third codon positions radically influence the estimation of trees based on next-gen data. Gen. Biol. Evol. 5, 2082–2092. Browne, F.G., 1961. The biology of Malayan Scolytidae and Platypodidae. Mal. Forest Rec. 22, 255. Chikhi, R., Medvedev, P., 2013. Informed and automated k-mer size selection for genome assembly. Bioinformatics 30, 31–37. Chang, Z., Li, G., Liu, J., Zhang, Y., Ashby, C., Liu, D., Cramer, C.L., Huang, X., 2015. Bridger: a new framework for de novo transcriptome assembly using RNA-seq data. Genome Biol. 16, 30. Cognato, A.I., 2013. Molecular phylogeny and taxonomic review of Premnobiini Browne, 1962 (Coleoptera: Curculionidae: Scolytinae). Front. Ecol. Evol. 1, 1. Cognato, A.I., Hulcr, J., Dole, S.A., Jordal, B.H., 2011. Phylogeny of haplo-diploid, fungus-growing ambrosia beetles (Curculionidae: Scolytinae: Xyleborini) inferred from molecular and morphological data. Zoolog. Scr. 40, 174–186. Cooperband, M.F., Stouthamer, R., Carrillo, D., Eskalen, A., Thibault, T., Cossé, A.A., Castrillo, L.A., Vandenberg, J.D., Rugman-Jones, P.F., 2016. Biology of two members of the Euwallacea fornicatus species complex (Coleoptera: Curculionidae: Scolytinae), recently invasive in the USA, reared on an ambrosia beetle artificial diet. Agric. For. Entomol. 18, 223–237. Deyrup, M., 1987. Trischidias exigua Wood, new to the United States, with notes on the biology of the genus (Coleoptera: Scolytidae). Coleopt. Bull. 339–343. Deyrup, M., Kirkendall, L.R., 1983. Apparent Parthenogenesis in Pityophthorus puberulus (Coleoptera: Scolytidae). Ann. Entomol. Soc. Am. 76, 400–402. Dole, S.A., Jordal, B.H., Cognato, A.I., 2010. Polyphyly of Xylosandrus Reitter inferred from nuclear and mitochondrial genes (Coleoptera: Curculionidae: Scolytinae). Mol. Phylogenet. Evol. 54, 773–782. Edgar, R.C., 2010. Search and clustering orders of magnitude faster than BLAST. Bioinformatics 26, 2460–2461.

237

Molecular Phylogenetics and Evolution 127 (2018) 229–238

A.J. Johnson et al.

Biol. Evol. 35, 823–836. Skelton, J., Jusino, M.A., Li, Y., Bateman, C., Thai, P.H., Wu, C., Lindner, D.L., Hulcr, J., 2018. Detecting symbioses in complex communities: the fungal symbionts of bark and ambrosia beetles within Asian pines. Microb. Ecol. 1–12. Slater, G.S.C., Birney, E., 2005. Automated generation of heuristics for biological sequence comparison. BMC Bioinform. 6, 31. Stamatakis, A., 2014. RAxML version 8: a tool for phylogenetic analysis and post-analysis of large phylogenies. Bioinformatics 30, 1312–1313. Storer, C., Payton, A., McDaniel, S., Jordal, B., Hulcr, J., 2017. Cryptic genetic variation in an inbreeding and cosmopolitan pest, Xylosandrus crassiusculus, revealed using ddRADseq. Ecol. Evol. 7 (24), 10974–10986. Ueda, A., 1997. Brood development of an inbreeding spermatophagous scolytid beetle, Coccotrypes graniceps (Eichhoff)(Coleoptera: Scolytidae). Japan. J. Entomol. 65, 677–687. Vega, F.E., Brown, S.M., Chen, H., Shen, E., Nair, M.B., Ceja-Navarro, J.A., Brodie, E.L., Infante, F., Dowd, P.F., Pain, A., 2015. Draft genome of the most devastating insect pest of coffee worldwide: the coffee berry borer, Hypothenemus hampei. Scient. Rep. 5, 12525. Warren, D.L., Geneva, A.J., Lanfear, R., 2017. RWTY (R We There Yet): an R package for examining convergence of Bayesian phylogenetic analyses. Molec. Biol. Evol. 34, 1016–1020. Wood, S.L., 1978. A reclassification of the subfamilies and tribes of Scolytidae (Coleoptera). Annales de la Société Entomologique de France 14. Wood, S.L., 1982. The bark and ambrosia beetles of North and Central America (Coleoptera: Scolytidae), a taxonomic monograph. Great Basin Natural. Mem. 6, 1–1356. Wood, S.L., 1986. A reclassification of the genera of Scolytidae (Coleoptera). Great Basin Natural. Mem. 10, 2. Wood, S.L., 2007. Bark and Ambrosia Beetles of South America (Coleoptera, Scolytidae). Monte L, Bean Life Science Museum, Provo, Utah. Wood, S.L., Bright, D.E., 1987. A catalog of Scolytidae and Platypodidae (Coleoptera), Part 1: Bibliography. Great Basin Naturalist Memoirs 11, 1–685. Wood, S.L., Bright, D.E., 1992. A catalog of Scolytidae and Platypodidae (Coleoptera), Part 2: taxonomic index. Vols. A and B. Brigham Young University. Great Basin Natural. Mem. 13, 1553. Yang, Z., Kumar, S., Nei, M., 1995. A new method of inference of ancestral nucleotide and amino acid sequences. Genetics 141, 1641–1650. Zhu, J.-Y., Zhao, N., Yang, B., 2012. Global transcriptome profiling of the pine shoot beetle, Tomicus yunnanensis (Coleoptera: Scolytinae). PloS One 7, e32291.

McKenna, D.D., Sequeira, A.S., Marvaldi, A.E., Farrell, B.D., 2009. Temporal lags and overlap in the diversification of weevils and flowering plants. Proc. Natl. Acad. Sci. 106, 7083–7088. McKenna, D.D., Wild, A.L., Kanda, K., Bellamy, C.L., Beutel, R.G., Caterino, M.S., Farnum, C.W., Hawks, D.C., Ivie, M.A., Jameson, M.L., 2015. The beetle tree of life reveals that Coleoptera survived end-Permian mass extinction to diversify during the Cretaceous terrestrial revolution. Syst. Entomol. 40, 835–880. Meyer, M., Kircher, M., 2010. Illumina sequencing library preparation for highly multiplexed target capture and sequencing. Cold Spring Harbor Protocols 2010, pdbprot5448. Normark, B.B., Jordal, B.H., Farrell, B.D., 1999. Origin of a haplodiploid beetle lineage. Proc. R. Soc. B: Biol. Sci. 266, 2253. Novotny, V., Basset, Y., Miller, S.E., Weiblen, G.D., Bremer, B., Cizek, L., Drozd, P., 2002. Low host specificity of herbivorous insects in a tropical forest. Nature 416, 841–844. Novotny, V., Miller, S.E., Baje, L., Balagawi, S., Basset, Y., Cizek, L., Craft, K.J., Dem, F., Drew, R.A.I., Hulcr, J., Leps, J., Lewis, O.T., Pokon, R., Stewart, A.J.A., Allan Samuelson, G., Weiblen, G.D., 2010. Guild-specific patterns of species richness and host specialization in plant-herbivore food webs from a tropical forest. J. Anim. Ecol. 79, 1193–1203. Pagel, M., 1994. Detecting correlated evolution on phylogenies: a general method for the comparative analysis of discrete characters. Proc. R. Soc. Lond. B: Biol. Sci. 255, 37–45. Pistone, D., Gohli, J., Jordal, B.H., 2017. Molecular phylogeny of bark and ambrosia beetles (Curculionidae: Scolytinae) based on 18 molecular markers. Syst. Entomol. 43, 387–406. Prum, R.O., Berv, J.S., Dornburg, A., Field, D.J., Townsend, J.P., Lemmon, E.M., Lemmon, A.R., 2015. A comprehensive phylogeny of birds (Aves) using targeted next-generation DNA sequencing. Nature 526, 569–573. Rambaut, A., Suchard, M.A., Xie, D., Drummond, A.J., 2015. Tracer v1.6 < http://tree. bio.ed.ac.uk/software/tracer/ > . Revell, L.J., 2011. phytools: an R package for phylogenetic comparative biology (and other things). Meth. Ecol. Evol. 3, 217–223. Rice, P., Longden, I., Bleasby, A., 2000. EMBOSS: the European molecular biology open software suite. Trends Genet. 16, 276–277. Schedl, K.E., 1961. Scolytidae und Platypodidae Afrikas. Band 1 (part). Familie Scolytidae. Revista de Entomologia de Moçambique 4, 335–742. Shin, S., Clarke, D.J., Lemmon, A.R., Lemmon, E.M., Aitken, A.L., Haddad, S., Farrell, B.D., Marvaldi, A.E., Oberprieler, R.G., McKenna, D.D., 2017. Phylogenomic data yield new and robust insights into the phylogeny and evolution of weevils. Molec.

238